Hans W. Gottinger, Essays on Technology, Markets and Development

<strong>Hans</strong> W. <strong>Gottinger</strong> 

ESSAYS ON 

TECHNOLOGY, 

MARKETS AND 

DEVELOPMENT


ESSAYS ON 

TECHNOLOGY, 

MARKETS AND 

DEVELOPMENT 


STRATEC Munich, Germany 

www.stratec-con.net 

stratec_c@yahoo.com 

gottingerhans@gmail.com

Aggregate Keywords (10): Investment, Technological Development, Stochastic Models, 

Network Economics, Supply Chain Competition/Coopetition, Strategic Alliances, 

Microeconomics of Economic Growth, Economic Growth and Catching-Up, Development Paths, 

Economic Growth under Environmental Constraints

CONTENTS 

HANS W. GOTTINGER 

CONTENTS 

ESSAYS ON TECHNOLOGY, MARKETS 

AND DEVELOPMENT 

Contents ................................................................................................ I 

Foreword. .............................................................................................. III 

Preface and Introduction ............................................................................... V 

1. Innovation and Technology ............................................................ 1 

1.1 Stochastics of Innovation Processes .............................................................. 3 

1.2 Stochastic Models for Capital Growth ........................................................... 17 

1.3 Modeling stochastic innovation races ........................................................... 33 

1.4 Global technological races. ..................................................................... 53 

1.5 Stochastic racing and competition in network markets .......................................... 67 

1.6 High speed technology competition. ............................................................ 85 

1.7 Sequential Technology Choice and R&D Racing. ................................................ 103 

1.8 Internet of Things Economic and Industrial Dimensions ......................................... 123 

2. Markets and Competition ........................................................... 133 

2.1 Innovation, Dynamics of Competition and Market Dynamics ................................... 135 

2.2 Network Economics for Open Source Technologies ............................................. 151 

2.3 Vertical Competition and Outsourcing ......................................................... 169 

2.4 Supply-Chain Coopetition ..................................................................... 177 

2.5 A tale on Chinese Innovation and Competition ................................................. 183 

2.6 Chinese Innovation and Competition .......................................................... 189 

2.7 Competitive Positioning through Strategic Alliances ............................................ 199 

3. Growth and Development ........................................................... 217 

3.1 Microeconomic Foundations of Economic Growth (Review) ..................................... 219 

3.2 Economic Growth, Catching-Up, Falling Behind and Getting Ahead ............................. 301 

3.3 Catch-Up and Convergence. ................................................................... 315 

3.4 Escaping from Development Traps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345 

3.5 Some Applications of a Result in Control Theory to Economic Planning Models .................. 355 

3.6 Optimal Economic Growth when CO2 Constraints are Critical ................................... 377 

3.7 Economic Choice and Technology Diffusion in New Product Markets ............................ 385 

I


CONTENTS 

II

SOURCES OF ARTICLES CORRESPONDING TO SECTIONS 


Sources of articles corresponding to Sections 1.1 – 3.7 

1 Technology 

1.1 Journal of Economics 49(2), 1989, 123-138 

1.2 Journal of Economics 54(3), 1991, 267-281 

1.3 Technological Forecasting & Social Change 69, 2002, 607-624 

1.4 Japan and the World Economy 18, 2006, 181-193 

1.5 Int. J. Technology, Policy and Management 4(3), 2004, 240-256 

1.6 Int. J. Business and Systems Research 10(1), 85-102 

1.7 Journal of Policy Studies (Japan) 15(9), 2003, 43-62 

1.8 Int. J. Bus. and Economics Research 6(5), 2017, 115-123 

2 Markets 

2.1 Archives of Bus.Research 4(1), 2015, 1-16 

2.2 Adv. Soc.Sci. Research 4(7), 2017, 164-181 



2.5 Research Memo. Abstract, STRATEC Munich , 2012 


2.7 Int. J. Revenue Management 8(1), 2008, 1-17 

3 Growth and Development 

3.1 Unpublished Supplement to: <strong>Hans</strong> W. <strong>Gottinger</strong> and Mattheus F.A. Goosen , eds 

Strategies of Economic Growth and Catch-Up, NovaScience, New York 2012 

3.2 Journal of Policy Studies (Japan) 21(11), 2005, 1-14 

3.3 J.Appl.Economics and Business 2(5), 2014, 67-95 

3.4 (with C. Umali), J.Appl.Economics and Business 3(1), 2015, 5-13 

3.5 Weltwirtschaftliches Archiv (Review of Economics) 111. 1975, 728-751 

3.6 Energy Economics, July 1992, 192-199 

3.7 Weltwirtschaftliches Archiv (Review of Economics) 123, 1987, 93-120 

III


SOURCES OF ARTICLES CORRESPONDING TO SECTIONS 

IV

FOREWORD 


FOREWORD 

Yesterday (or was it Thursday) Apple was crowned as the first “trillion dollar company”, willing a supposed 

race with Alphabet (Google), Amazon and Microsoft. Many guru hours among financial journalists have 

been, and more will be spent on recounting the high (and a few low) points of Apple’s rise. This book has 

little to say on the dynamics of that particular competition, except perhaps to note that insights from the 

theory of games may play a role in the story. 

Speaking for myself, I find it more fascinating that Apple’s path to success, is littered with the corpses 

or dying remains of former corporate super-stars. Why, one wonders, did the inventors and innovators 

behind the key inventions and innovations, gain so little of the wealth they created? For that matter, why 

did formerly dominant tech firms, such as GE, IBM, Xerox, Kodak and HP, lose their way on the yellow 

brick road? Why did the current crop of winners get so powerful and why did the losers lose out? Moreover, 

can we expect Amazon and Alphabet to fall by the wayside in their turn? 

These questions are of interest to inventors and venture capitalists, of course. They are also of interest 

to long-term institutional investors, such as pension funds, endowment funds and insurance companies. 

At another level, they are potentially important to our understanding of the economic system as a whole. 

Economists have long been at a loss to understand the underlying drivers of booms and busts, along with 

their accompanying changes in wealth distribution and consequences for political power. 

Economic theory relies heavily on linear models. But the “boom-bust” pattern is evidently non-linear. 

Let me indulge in a bit of theorizing of my own. I postulate that both booms and busts are likely to reflect 

two sides of the same phemomenon, the phenomenon of “lock-in/lock-out”. What this means is that a first 

choice among competing bidders or technologies or business models can become effectively irreversible 

once initial advantages (e.g. learning or experience) kick in. For example, the “experience curve” is a 

particularly important component of increasing productivity Returns to scale create (temporary) advantages 

for incumbents (booms) and, consequently, disadvantages to challengers. This phenomenon can (and does) 

often result in sub-optimal (or even bad) long-term outcomes (busts). 

The obvious example today is the fact that our whole industrial system (and urban development, in particular) 

is “locked in” to dependence on a road-based transport system requiring private automobiles (and trucks) 

using liquid hydrocarbon fuels. Electric and steam-powered vehicles were still competitive in 1900, but they 

had a short range, due to storage limitations. Steam power overcame that limit by introducing a condenser 

in the 1920s (with the Doble), but it was much too late to compete. Electrification of cars only began again 

after the year 2000, thanks to the invention of lithium-ion batteries. Range and battery re-charge rate are 

still limitations, but performance in those domains is improving fast. What if the lithium-ion battery had 

been invented a century earlier? 

The “lock-in” of uranium-based nuclear power technology (and the “lock-out” of a potentially much superior 

thorium-based technology) is another example. It seems that uranium (specifically the isotope U-235) became 

the fuel of choice precisely because of the military uses of its by-products, neptunium and plutonium. Now, 

in stationary power applications plutonium is an unwanted (except by terrorists) pollutant, not a valuable 

by-product. Yet uranium technology still gets all the investment. 

Positive returns-to-scale create a short-term advantage – hence a barrier to challenge by competitors – 

resulting from positive returns to scale. The assumption of constant or declining returns to scale is wrong 

V


FOREWORD 

because “network” service industries (like the telephone system) demonstrably exhibit increasing returns 

to scale. (See pp 151-160 of this book). The simple explanation for positive returns in network sectors is that 

the larger the network the greater the number of people connected to each node, resulting in increased value 

to each customer. This phenomenon, in more general terms, is also responsible for population clustering 

(cities), the success of department stores, browsers, the success of search engines (like Google) and a variety 

of other economically important activities. All of these phenomena are impossible to explain in terms of 

economic theories based on constant or decreasing returns to scale. 

Positive feedbacks in a deterministic system do not necessarily lead the system to a unique equilibrium 

state. The dynamic equations for the economy in disequilibrium are non-linear, and for non-linear systems 

there can be multiple alternative equilibria. At a bifurcation point the system then “chooses” one path, 

often for short-term reasons. There is no guarantee whatsoever that the path adopted at a bifurcation point 

leads to the “best” long-term solution. Indeed, as, the path selected is often sub-optimal, and may lead 

the economy “over a cliff” if there is no scope for remedial action. It is a priori unlikely that the optimum 

short-term choice is also the optimum long-term choice. 

This book will, hopefully, provide some clues to these mysteries. 

Robert Ayres, Novartis Professor of Economics and Technology, INSEAD, Paris 

VI

PREFACE AND INTRODUCTION 


<strong>Essays</strong> on Decisions, Technology, Markets and Development 


This collection of essays highlights the interface topics of innovation and technology, markets and competition, 

growth and development in an evolutionary dynamic economy of a dominantely capitalistic market economy 

in a globalized context. They span a time frame from the 1980s to current developments. 

Previews 

In Part 1 the core of Secs. 1.1. to 1.8 emphasize the linkage between high speed technological competition 

through technology races in firms, industries and national economies. We start with observations how 

stochastic innovation processes emerge in a firm and are carried through by capital investments and 

how the capital allocation activity is governed by a stochastic process constrained by optimal stopping 

of productivity enhancing investments (Secs.1.1, 1.2). Further on, the remaining Secs. 1.3 to 1.7 rely on 

(stochastic) differential game analysis in symmetric and asymmetric rivalrous situations among firms that 

extends the modeling of R&D based competition beyond conventional models found in the economics and 

technology literature. Many high technology industries are characterized by positive network externalities. 

Firms essentially compete and cooperate on R&D and the production of goods and services that share a 

network. Some models of competition contain special features that apply equally well to network markets. 

One is the uncertainty in technological development or uncertainty in the realization of a firm’s R&D effort. 

The other is the dynamic nature of price competition between firms in the presence of network effects. 

Firms compete with each other over an extended period of time and must therefore strategically choose 

prices as the market shares of the firms evolve. 

Most existing models of network markets focus on only one of the two features. Further, almost all the 

models have the commonality that one firm captures the market instantaneously and sells to all consumers 

from then on (the “winner-takes-all” market) . However, the history of high technology industry abounds 

with instances where rival firms (and technologies) have had extended battles for providing the industry 

standard. 

In network markets evolution of market share is a very interesting phenomenon especially in the face of 

uncertainty about future product quality. Sec.1.5 attempts to provide a framework that could capture the 

richness of market share evolution in the presence of network externalities. In a specific situation of a high 

-tech market if you are not the indisputable leader you have only little hope to achieve leadership unless 

you come up with a major innovation replacing the leading firm. Incremental innovation – making slight 

improvements in the leaders’ products – will not enable a small firm to overtake a leader that enjoys the 

benefits of network economies. Therefore, it is not atypical for a fringe firm that invests heavily to displace 

the leader by leapfrogging the leader’s technology. 

We are encountering the problem of tradeoff between “network dominance” and “radical innovation” 

that could tip the market the other way, with a significant caveat added that breakthrough R&D is highly 

uncertain. From a strategic perspective, in this environment, for any two firms of asymmetric size, both 

compete dynamically over prices to win market share. In this dynamic process there are two ways to achieve 

(temporary) monopolistic status. The ‘smaller’ firm can use dynamic pricing competition to delay the time 

in which the ‘larger’ firm wins a critical market share in the hope to hit the innovation first and displace 

it. If the innovator is “patient”, the probability of innovation and the discount factor are sufficiently high, 

there is an equilibrium in which duopoly persists (no firm achieves a critical market share) until one of the 

two firms wins the race for innovation. 

In Sec.1.7 we model the R&D process as one that is composed of a series of stages, at each of which some 

set of product characteristics are determined. At each stage, the firm exercises discretion as to the extent 

VII



to which the particular stage characteristics will be developed. Further, when the firm’s R&D resource 

endowment is fixed and stage-specific, the firm faces a tradeoff between the extent to which it will develop 

the stage characteristics and the time that it will spend in doing so. We will refer to the degree to which 

the stage characteristic is developed as the characteristic’s (targeted) level of strategic positioning. In this 

sense, the more “forward-looking” the product characteristics, the greater the profits the firm can expect 

to get from the product but the longer the expected time to completion of R&D. We will speak of this as 

the speed–leadership tradeoff in each stage. When the completion of R&D requires many such tradeoffs to 

be resolved sequentially, firms will, at the beginning of each stage, want to make their choices in light of 

the choices made by all firms in previous stages and their stage completion records to that time. To carry 

technological innovation through networks across firm/industry boundaries creates a new paradigm of 

network centered innovation that integrate through intelligent software tools as outlined and reviewed on 

the present state of affairs in Sec. 1.8. 

In highly competitive technological industries new challenges and opportunities are arising in the new 

product development arena. Driven by global markets, global competition, the global dispersion of scientific/ 

engineering talent, and the advent of new information and communication technologies (ICT) a new vision 

of product development is that of a highly disaggregated , distributive process with people and organizations 

spread throughout the world. At the same time products are becoming increasingly complex requiring 

numerous engineering decisions to bring them to market. Competitive pressures mean that “time to market” 

has become a key to new product success. However, at the same time it is important to keep the innovation 

and quality dimensions of the new product at their optimal level. 

A central question to address is how should firms invest in innovation and what are the implications of such 

investments for competitive advantage? Understanding why a firm benefits from investments in innovation 

and quality illuminates issues of competitive strategy and industrial organization. In the field of competitive 

strategy, much attention has been devoted to the concept of core capabilities. Understanding how firms make 

optimal investments in the face of competition reveals the nature of competition and provides theoretical 

and managerial implications for developing core competence and dynamic capabilities. 

In major parts of competitive analysis involving R&D decisions the focus is on breakthrough innovations 

which could create entirely new markets, for example, in studies featuring patent racing between competing 

firms. In more common competitive situations we observe firms, however, competing by investing in 

incremental improvements of products. It is an important aspect when innovation is considered to be 

manifested in product quality, process improvements and in the overall quality culture of an organization 

. For example, after product launch, incremental improvement of different aspects of product quality, 

improvements in various business processes and an incremental adoption of a quality culture are quite realworld 

phenomena. Thus innovation is conceptualized as being manifested in terms of quality. Moreover, 

the quality dimension extends to pursue continuous total quality improvement for the entire product life 

cycle. Some firms operate in a simultaneous product launch situation while others compete sequentially 

by adopting the role of leader or follower. The strategic implications in these diverse circumstances can be 

treated within a unified framework of dynamic stochastic differential games. 

Part 2 , Competition and Markets, is devoted to look into the interaction of innovation culture, competitive 

dynamics and shaping of markets with continuous emergence and proliferation of open source technologies 

dominating in ICTs (Secs. 2.1, 2.2). In Sec. 2.2 we focus on a contestable market with network externalities 

engaging an incumbent and an entrant. The incumbent, unlike the entrant, already has an installed base 

of consumers. We look at decision situations of firms regarding how proprietary they want to make their 

technology , either through patent protection or through development in open source systems (OSS). 

We explicitly model the direct and indirect effects of network externalities. For example, more software 

companies are willing to produce programs for an operating system (OS) if it has a larger consumer base. 

This increased competition could lead to an improvement of the quality of the OS. The model predicts that 

using open source technologies is likely to enhance the rate of R&D , and consequently the quality of the 

product.. An incumbent that would choose this strategy is likely to deter entrance of a newcomer because 

it can play out its advantage of a larger network. 

Competitive and cooperative forces, termed “coopetition”, are likely to synchronize in extending to supply 

VIII



chain networks (Secs. 2.3, 2.4) . This opens the door for strategic cooperation as focussed on a timely case 

of a Chinese industry study (Secs. 2.5, 2.6). The degree of competitive positioning in view of forming 

strategic alliances (Sec. 2.7) is explored. This preview concludes by emphasizing the analytical , strategic 

and practical implications and providing directions for future industry analysis. 

In recent years with the emergence of e-business and a supply chain view for product development processes, 

multiple firms with varying and at times conflicting objectives enter into collaborative arrangements. In 

such situations, the competitive strategy based on quality and innovation could potentially permeate in 

those collaborative setups. A recent example is the collaborative venture of Sony and Samsung to build 

a cutting edge plant for LCD flat screen TV though displaying ongoing fierce rivalry in new product 

launches in exactly the same product categories. When innovation and quality levels form the core of a 

firm’s capabilities, each member in the supply chain would have an incentive to invest and improve their 

dynamic capabilities. This leads to tacit competition among collaborative product development partners 

by means of active investment in innovation and quality. 

Although the forces of innovation are central to competition in young, technically dynamic industries, they 

also affect mature industries where life cycles historically were relatively strong, technologies mature, and 

demands stable. 

A strategy for technology must confront primarily what the focus of technical development will be. The 

question is what technologies are critical to the firm’s competitive advantage. In this context, technology 

must include the know-how the firm needs to create, produce and market its products and deliver them 

to customers. As a major step in creating a technology strategy it has to define those capabilities where 

the firm seeks to achieve a distinctive advantage relative to competitors. For most firms, there are a large 

number of important areas of technological know-how but only a handful where the firm will seek to create 

truly superior capability. 

Having determined the focus of technical development and the source of capability, the firm must establish 

the timing and frequency for innovation efforts. Part of the timing issue involves developing technical 

capabilities, and the rest involves introducing technology into the market. The frequency of implementation 

and associated risks will depend in part on the nature of the technology and the markets involved (e.g. disk 

drive vs. automotive technology), but in part on strategic choice. At the extreme, a firm may adopt a rapid 

incremental strategy, that is, frequent, small changes in technology that cumulatively lead to continuous 

performance improvement. The polar opposite might be termed the great leap forward strategy. In this 

approach, a firm chooses to make infrequent but large-scale changes in technology that substantially 

advance the state of the art . 

Sec. 2.3 uncovers the importance of asymmetries by investigating the difference between firms in terms of 

key parameters. Moreover, for the examination of inter-firm competition through product development, the 

context of simultaneous entry and sequential entry are treated separately. This allows a deeper understanding 

of the implications of information asymmetry and commitment which have been regarded as important 

determinants in the context of game-theoretic studies. 

The study of innovation based competition has often considered aspects related to patent races and incremental 

product-process innovation to achieve distinctive advantage. However, recently innovation-based competition 

has become an aspect of buyer-supplier relationships. There are many instances in manufacturing where one 

finds situations of lock-ins created by innovative suppliers. For example, in the computer industry Intel and 

Microsoft as suppliers of microprocessor and operating systems, respectively, to desktop manufacturers like 

IBM , Hewlett-Packard and Dell illustrate such innovation-based lock-ins. Indeed, there exists an evolving 

power structure (dubbed “channel power”) in a supply chain driven by innovation competence of its members. 

A business context is envisaged in which, at any given point in time of the relationship, both the buyer and 

the supplier could be pursuing innovation simultaneously. We recognize that the primary motivation for 

such investments in innovation by members of the supply chain is to increase their differential or relative 

IX



channel power in the supply chain. Also in a situation where the buyer is locked in by a supplier, the 

buyer may actively pursue the creation of a substitute technology by investing in innovation. The primary 

motivation for the buyer in this case would be to eliminate the technology lock-in and become independent. 

Another contribution, Sec.2.4, presents the competitive role of innovation among collaborating firms. 

The models provide reasoning for inter-firm incentives in forming collaborative arrangements for product 

development. The dynamics of relationship among supply chain partners is viewed in terms of their 

respective innovation competence. It is emphasized that varying power arrangements in a supply chain leads 

to different implications for investments in innovation by buyer and supplier. It highlights the incentive for 

buyer and supplier to strategically maneuver their overall innovation levels by appropriate investments. 

The purpose is to provide a strategic framework and insights regarding power and competition in a 

collaborative supply chain setup. 

Managerial dimensions 

The analytical results provide insights into managerial implications regarding strategic issues concerned 

with product development and supply chain planning. The conceptual framework provides a foundation for 

managers to view competitive advantage emanating from product development in a more integrated manner. 

The results of Sec.1.6 point towards a time variant innovation investment strategy. It suggests that firms 

competing by means of new product development must choose instantaneous investment in innovation 

such that it increases with time. The quality manifestations exhibit an increasing trajectory. In a symmetric 

competition, the investment profile of the leader is sigmoidal while that of the follower is convex increasing. 

‘T’he impact of such an investment profile is aptly reflected in the quality improvement trajectory. 

The results provide implications for a firm for innovation investments by considering their relative strengths 

and weaknesses. The exploration of firm asymmetries provide some interesting implication that a firm 

should consider when competing via new product development. The results can be translated into executable 

decisions for investments. These results provide insights to evaluate different scenarios for pursuing strategies 

which would lead to best market outcome. 

Secs. 2.3 and 2.4 provide a way of thinking about business strategy and operational alignment in a 

collaborative network. The results provide directions for strategies to be adopted in a supply chain context. 

These are becoming increasingly important in the present business environment, where many firms are 

joined together in a collaborative network. In industries driven by innovation ,e,g. semi-conductors or 

biotechnology, having the control on the overall innovation levels that drive the technology landscape could 

mean a strong strategic advantage for one partner over other contributing firms. 

Innovation competence plays an important role in this approach on a supply chain relationship. In Sec. 2.4 

the specific scenario modeled is representative of practical situations. As an example, it can be observed 

that IBM is actively involved in creating quantum computers. One reasoning for investments in quantum 

computers could be an extension of a technology landscape. However, equally important is the realization 

that a success of such an endeavour could lead to a lock-out of Intel from this newly formed market for 

quantum courputers. The simplistic model used in Sec. 2.4 allows for an exploration of such a situation. 

The stationary Markov perfect Nash equilibrium investment strategy of the buyer and supplier is found to 

be time invariant and is characterized by the parameters in the model. An interesting insight obtained as 

a result of analysis of the model is that the supplier can influence the motivation of the buyer to invest in 

substitute technology. The underlying mechanism can be translated into pricing strategy that results in a 

long-term buyer-supplier relationship. 

The analysis of a model of buyer-supplier relationship allows to investigate a more generalized setup of 

buyer-supplier innovation-based competition. The aspect of channel power is tightly integrated in the 

analysis to understand the complex behavioral aspects using a simple framework. These results indicate 

that structural risk plays an important role in innovation investment decisions. Moreover, the variance in 

wealth formation is an important indicator of how much to invest. These results are quite intuitive and 

X



enable managers to objectively resolve some of these strategic decisions. 

The propositions provide insights into the dynamics of relationship between a buyer and a supplier firm. 

These propositions present the role and responsibility of buyer in creating motivation for the supplier to 

collaborate. The importance of critical assets is aptly amplified in the results. The possession of critical 

assets, often observed with the upstream partner gives the supplier a potential to achieve relative market 

closure through a position of dominance over competitors. It is likely that a firm in possession of such a 

critical asset also has the potential to achieve effective leverage over collaborating partner-, and suppliers. 

The responsibility of the buyer is to create a delicate balance between managing supplier and customers. 

Similar to the conceptual development the results suggest that for a buyer that wishes to be successful an 

understanding must be developed about how to own and control critical assets that provide opportunities to 

create customer dependency and “lock-in” .A competence in procurement management forms the cornerstone 

for success of long term business strategy. 

Extensions 

Several extensions can be considered for supply chain modeling and they establish behavioral features for 

the Microfoundations of Economic Growth, Sec. 3.1. 

First; within the framework of Sec.1.6 the model can be enriched by evaluating the improvement in product 

quality with learning effect. The learning effect is very well documented in the literature and the extension 

to incorporate learning effects is straightforward. Specifically, learning can be used to characterize the 

dynamics of evolution of product quality and in the characterization of costs associated with innovation 

investments. 

For analytical simplicity the revenue function is treated as a salvage value. As an extension the incorporation 

of the dynamics of evolution of revenue in the analysis of the model can be explored. In the sequential play 

game, the state dynamics of the follower can be considered to be dependent on the state of the leader at 

previous time instant. This modification would allow for analysis of a model in which the leader uses openloop 

Nash equilibrium strategy for innovation investments whereas the follower would adopt a Markovian 

Nash equilibrium strategy. Additional insights can be gained by this modification for leader-follower 

competitive dynamics in new product development. 

Firm asymmetries can be explored by considering multiple parameters at the same time. This would enable 

a richer understanding of the strategies that a leader and a follower should adopt based on their strengths 

and weaknesses. 

The competitive dynamics of buyers and suppliers are explored. The essays present the competitive role 

of innovation among collaborating firms. The research asserts that the buyer and the supplier could be 

in a competitive relationship due to efforts in innovation. Existing literature in operations management 

doesn’t explicitly consider the implications of innovation based competitive strategies in a collaborative 

supply chain context. Models are presented that provide motivation to explicitly consider such competitive 

situations. The notion of lock-in and channel power is implicit in the model and presents an approach for 

theory-building activities in this area. 

We know that the model presented derives from a simplified version of the dynamics between a buyer and 

a supplier. The constraints placed on different parameters and the context of a monopoly supplier and a 

buyer facing perfect competition is an approximation for analytical convenience. However, this enables a 

theoretical investigation of some important aspects underlying buyer-supplier relationships. Specifically, 

the research provides a way of thinking about business strategy in collaborative networks. These are 

becoming increasingly important in the present business environment, where many firms are part of a 

collaborative network. In industries driven by innovation , the industrial internet, advanced manufacturing, 

semiconductors and biotechnology, having control over the overall innovation levels could mean a strong 

strategic advantage for one partner over other contributing firms. 

At the same time, the firm already having the channel power can take some actions that would enable a 

XI



long-term relationship built on trust based governance. 

Several extensions could be considered to better represent such “coopetitive” relationships between 

collaborating firm. Given the objective interests, it is necessary for collaborating partners to know how far 

the other side is prepared to concede before it is no longer profitable to be in a relationship. Information about 

the cost structures of the collaborating partners and their relative utility from the exchange relationship is 

critical to an understanding of power in exchange relationship. Hence, one of the extension of the problem 

is to consider aspects of information asymmetry. The two key problems generated by private information 

and imperfect observability are – adverse selection and moral hazard. The adverse selection is a condition 

of supplier opportunism that occurs prior to a signing of a contract. Moral hazard on the other hand refers 

to supplier opportunism that occurs once the buyer has signed a contract.. A rational buyer tries to possess 

information that counters the above motives on the part of suppliers. Aspects related to detailed price 

comparison between the supplier; its competitors and the range of substitutes provides means to investigate 

buyer strategies. An investigation of the implications of threat by buyer of placing the contract elsewhere 

to get additional price concessions from its preferred supplier can be carried out. 

The assumptions made in the models and the constraints put on variables can be relaxed to obtain a better 

representation. A scenario with multiple suppliers/buyers enables enrichment of the findings . The result 

regarding guidelines for price bandwidth to create long-term relationships can be explored in further detail. 

The implications on profits by adopting a strategy for long-term relationship can be compared and contrasted 

with those of adopting the Markov-perfect Nash equilibrium strategies. The insights obtained from this 

approach also provide a background for future studies in supply chain contracts. The problem context and 

results clearly demonstrate the competitive role of innovation between the buyer and the supplier. It provides 

a rationale for investigation of supply chain contracts by explicitly considering the evolution of innovation 

competence and critical assets of collaborating partners. 

In Sec.2.4 the stochastic differential. game model considers the Brownian motion for the evolution of 

wealth for supplier and buyer. These Brownian motions can be treated to be autocorrelated to investigate 

the associated implications of investment strategy. The model can be extended to examine a sequential 

game and the associated investment strategy for buyer and supplier. Another extension to the model is 

consideration of spill-over effects. The spill-over of knowledge created by investments in innovation is 

common in literature pertaining to economics and management science. Such a spill-over of knowledge 

may provide some interesting results regarding buyer-supplier power structure and investment strategy. 

In Part 3, Economic Growth and Development, Secs. 3.1 – 3.7 we start with a broad survey in Sec. 3.1 on 

the microfoundations of economic growth where many model results on firm and industrial racing behavior 

in Parts 1 and 2 are sourced to built the structural basis for industrial growth and in the aggregate for 

macroeconomic growth. In Sec.3.2 the link micro-macro is the focus of a modified classical growth model 

(the Solow model) which at the center is built upon catch-up behavior in the aggregate. Secs. 3.3 and 3.4 

turn the page on development , Sec. 3.3 on catch-up behavior and development paths between nations in 

a global and historical context from the first industrial revolution on convergence and divergence of paths 

and changing positions on various performance criteria in economic growth and prosperity. While Sec. 3.3 

is more concerned with “racing to the top” in a more closely related field Sec. 3.4 emphasizes development 

paths from the bottom to overcome or break out of development traps or poverty traps. Secs. 3.5 and 

3.6 partially depart from the main theme, Sec. 3.5 portrays a a control theoretic model of development 

planning, more a traditional macro model either to be superimposed on or interacting cooperatively with 

more decentralized market and trade oriented industrial structures. Sec. 3.6 emphasizes another aspect 

of opimal economic growth under possible fixed/variable constaints trough climate change restrictions. 

In what follows we restrict comments on the main theme of Part 3 . 

On economic growth and development we assume that a key driver of economic growth among nations 

or regional economic entities is an intrinsic, ongoing, perpetual and historically observable rivalry to 

propel a state‘s standing, prestige, power and economic performance through getting ahead of or not to 

fall too much behind its rivals in a pecking order. As we observe in economic history through the age of 

industrialization from the mid 18th century, the cumulative process of industrial development in a dynamic 

time context can be correlated with some countries emerging as early leaders being challenged by others, 

XII



later adopters that strived for new leadership positions. This is in contrast to those that were not part of the 

techno-economic race, falling behind in economic prosperity, and inevitably not being able to catch-up 

over decades or even centuries. 

The design mechanism for strategies of economic growth essentially embodies a science and technology 

competition between nations and its industries as the creator of value added products and processes through 

competition as reflected in the level of gross domestic product (GDP). This is in the Schumpeterian tradition. 

It follows from exploring the varying implications of the notion, clearly advanced by Schumpeter (1942) that 

it is the expectation of supernormal profits from a temporary monopoly position following an innovation 

that is the chief driver of R&D investment. 

The success of economic growth due to diffusion of advanced technology or the possibility of leapfrogging 

is mainly attributable to how social capability evolves, (i.e., which effects become more influential, growing 

responsiveness to competition or growing obstacles to it on account of vested interests and established 

positions). In a time of overcoming a major financial crisis with a slowing pace of deregulation, privatization, 

liberalization and lifting of trade barriers it remains an interesting question regarding an international 

competitive order whether industry racing patterns are sufficiently controlled by open world-wide markets 

or whether complementary international agreements (e.g. regulations, controls) are needed to eliminate or 

mitigate negative externalities without compromising the positive externalities that come with industry racing. 

We investigate multiple kinds of races, frontier races among leaders and would be’ leaders on a global, 

regional and local scale corresponding to different leagues like in major sports events, as there are in each 

category catchup races among laggards and imitators. Technological frontiers at the firm and industry race 

levels offer a powerful tool through which to view evolving technologies within an industry. By providing 

benchmarking roadmaps that show where an individual firm is relative to other firms in the industry, they 

highlight the importance of strategic interactions in the firm’s technology decisions. 

In an economic history perspective getting ahead, catching up and falling behind processes are taking 

place toward leaders and followers, trailers and laggards within a group of industrialized, industrializing 

and developing countries in pursuit of higher levels of power, welfare and productivity. Moses Abramovitz 

(1986) explains the central idea of the catch-up hypothesis as the trailing countries adopting behaviour 

of a backlog of unexploited technology. Supposing that the level of labor productivity were governed 

entirely by the level of technology embodied in capital stock, one may consider that the differentials in 

productivities among countries are caused by the “technological age” of the stock used by a country relative 

to its “chronological age”. 

The technological age of capital is an age of technology at the time of investment plus years elapsing from 

that time. Since a leading country may be supposed to be furnished with the capital stock embodying, in 

each vintage, technology which was at the very frontier at the time of investment, the technological age 

of the stock is, so to speak, the same as its chronological age. While a leader is restricted in increasing 

its productivity by the advance of new technology, trailing countries, along Abramovitz’ line, have the 

potential to make a larger leap as they are provided with the privilege of exploiting the backlog in addition 

to the newly developed technology. Hence, followers being behind with a larger gap in technology will 

have a stronger potential for growth in productivity. The potential, however, will be reduced as the catchup 

process goes on because the unexploited stock of technology becomes smaller and smaller. 

This hypothesis explains the diffusion process of best practice technology and gives the same sort of 

S-curve change in productivity rise of catching-up countries among a group of industrialized countries as 

that of followers to the leader in an industry. Although this view can explain the tendency to convergence 

of productivity levels of follower countries, it fails to answer the historical puzzles why a country, such as 

the United States, has preserved the standing of the technological leader (taking exceptions for two world 

wars) for a long time since taking over leadership from Britain and Germany before the First World War 

and why the shifts have taken place in the ranks of follower countries in their relative levels of productivity, 

i.e., technological gaps between them and the leader. 

Abramovitz poses some extensions and qualificatios on this simple catchup hypothesis in an attempt to 

XIII



explain these facts. Among other factors than technological backwardness he lays stress on a country’s 

social capability, i.e., years of education as a proxy of technical competence in its political, commercial, 

industrial, and financial institutions. The social capability of a country may become stronger or weaker 

as technological gaps close and thus, he states, the actual catch-up process does not lend itself to simple 

formulation. This view has a common understanding to what Mancur Olson (1982) expresses to be public 

policies and institutions’ as his explanation of the great differences in per capita income across countries, 

stating that any poorer countries that adopt relatively good economic policies and institutions enjoy rapid 

catchup growth. This view should be taken seriously when we wish to understand the technological catching 

up to American leadership in some key industries by Japan, in particular, during the post-war period and 

explore the possibility of a shift in standing between these two countries. This consideration will directly 

bear on the future trend of the state of the art which exerts a crucial influence on the development of the 

world economy. 

On a macro scale this racing paradigm would suggest that the U.S. would mobilize all its technology/ 

science/entrepreneurial resources to increase or keep its distance to other large emerging economies (such 

as China or India). With science/ technology being an evolutionary cumulative enterprise, for a dominating 

country with a portfolio of major increasing returns industries, the odds of leapfrogging oneself are higher 

than being leapfrogged by close followers, thus this asymmetry could play a distinctive role. Abramovitz’ 

advantage of backwardness may hold on up to a certain limit but with decreasing returns. 

In an interesting paper, under some seemingly reasonable assumptions on linear technology investment 

and dynamic equilibrium path of capital accumulation in a neoclassical type model Lau and Wan (1993) 

obtain the following results which they argue are fully consistent with empirical growth economics: These 

are: (a) Not all economies converge in growth with each other. (b) Economies with an initial technical 

capability will converge in growth with the advanced economies. The difference in per capita output grows 

exponentially, if the developing economy engages only in imitation (and not innovation). (c) With an initial 

technical capability there is a ‘high growth’ period, preceded (followed) by a phase of ‘trend acceleration’ 

(‘trend deceleration’). (d) With an initial technical capability the the technology gap widens forever. The 

previous analysis indicates that the possibilities opened through competitive industrial racing are far richer 

and more surprising than they would emerge from the macro-scale models. 

A dynamic competitive national economy with its underlying institutional factors is fed by a consistent 

industrial policy intrinsically supported by a strategic innovation system. Although this view can explain the 

tendency to convergence of productivity levels of follower countries, it fails to answer the historical puzzle 

why a country, the United States, has preserved the standing of the technological leader for a long time this 

century, and why it has intermittently been threatened by Germany, Japan and lately by emerging China and 

to a lesser degree by India. We show the empirical proliferation of comparative growth and development 

over the main period of industrialization. Why have the shifts taken place in the ranks of follower countries 

in their relative levels of productivity, i.e., technological gaps between them and the leader? 

XIV

1. 

INNOVATION 

AND 

TECHNOLOGY

1.1 STOCHASTICS OF INNOVATION PROCESSES 


3



4



5



6



7



8



9



10



11



12



13



14



15



16

1.2 STOCHASTIC MODELS FOR CAPITAL GROWTH 


17



18



19



20



21



22



23



24



25



26



27



28



29



30



31



32

1.3 MODELING STOCHASTIC INNOVATION RACES 


33



34



35



36



37



38



39



40



41



42



43



44



45



46



47



48



49



50



51



52

1.4 GLOBAL TECHNOLOGICAL RACES 


53



54



55



56



57



58



59



60



61



62



63



64



65



66

240 Int. J. Technology, Policy and Management, Vol. 4, No. 3, 2004 

1.5 STOCHASTIC RACING AND COMPETITION IN NETWORK MARKETS 

Stochastic racing and competition in network 

markets 


<strong>Hans</strong>-Werner <strong>Gottinger</strong> 

Institute of Management Sciences, Maastricht, 

Kwansei Gakuin University (KGU), 

Kobe-Sanda Campus, Japan 

E-mail: bvz28020@ksc.kwansei.ac.jp 

Abstract: We look at a competitive situation in network markets where there is 

uncertain technological development in product/process technologies. Firms 

‘price’ compete in those markets to gain market share before any of them 

succeeds in getting an innovation to move ahead of its rival(s). If the firms are 

in a technology race and the probability of innovation success is small, then a 

firm with a bigger network advantage is likely to attract more customers in the 

absence of innovation. If, however, any of those firms expect innovation with a 

high probability, and none of them have a big advantage, then they keep on 

sharing the market until the innovation occurs. 

Keywords: network economics; market structure and pricing; firm strategy; 

market performance. 

Reference to this paper should be made as follows: <strong>Gottinger</strong>, H-W. (2004) 

‘Stochastic racing and competition in network markets’, Int. J. Technology, 

Policy and Management, Vol. 4, No. 3, pp.240–256. 

Biographical notes: Dr <strong>Gottinger</strong> is Professor of Economics at the University 

of Osaka, and with the Institute of Management Science, Maastricht. His fields 

of interest include network economics, strategy and competition issues, energy 

and environmental economics. 

1 Introduction 

Many high technology industries are characterised by positive network externalities. 

Firms essentially compete against each other for goods that share a network. 

Our model of competition contains two special features that apply equally well to 

network markets. One is the uncertainty in technological development or uncertainty in 

the realisation of a firm’s R&D effort. The other is the dynamic nature of price 

competition between firms in the presence of network effects. 

Firms compete with each other over an extended period of time and must, therefore, 

strategically choose prices as the market shares of the firms evolve. 

Most existing models of network markets focus on only one of the two features. 

Further, almost all the models have the feature that one firm captures the market 

instantaneously and sells to all consumers from then on (the ‘winner-takes-all’ market). 

However, the history of the high technology industry abounds with instances where rival 

Copyright © 2004 Inderscience Enterprises Ltd. 

67



Stochastic racing and competition in network markets 241 

firms (and technologies) have had extended battles for providing the industry standard 

(Uttenback, 1994). 

In network markets, evolution of market share is a very interesting phenomenon 

especially in the face of uncertainty about the future product quality. This paper also 

attempts to provide a framework that could capture the richness of the market share 

evolution in the presence of network externalities. 

This paper is a follow-up to a stochastic racing model (<strong>Gottinger</strong>, 2002a): it also takes 

some cues on racing from the seminal contribution by C. Futia (1980). 

In its strategic dimension it is closely related to Katz and Shapiro (1992) and Harris 

and Vickers (1987), particularly, in their ‘tug-of-war’ presentation and multi-stage racing. 

Overall our approach is emerging from new ideas on anti-trust analysis in 

dynamically competitive industries (Evans and Schmalensee, 2001). 

It is illuminating to take a full quote out of their comprehensive analysis: 

“Firms that are not leaders in network industries generally have little hope of 

reaching that status unless they come up with a major innovation – one that can 

defeat the natural advantage that network effects bestow on the industry 

leaders. Incremental innovation – making slight improvements in the leaders’ 

products – will not enable a small firm to overtake a leader that enjoys the 

benefits of network economies. Similarly, the possibility of being displaced by 

a major innovation will shape leaders’ research agendas. If there is a chance 

that today’s products will be replaced by a major innovation, a leader’s survival 

depends on bringing this innovation to market and thereby replacing itself 

before others do. As a result, competition in network industries often involves 

intense R&D efforts aimed at capturing or retaining market leadership. It is not 

atypical for a fringe firm that invests heavily to displace the leader by 

leapfrogging the leader’s technology…” (Evans and Schmalensee, 2001, 

pp.10–12) 

We are exploring this problem of trade-off between ‘network dominance’ and ‘radical 

innovation’ that could tip the market the other way, with the significant complexity added 

that breakthrough R&D is highly uncertain. From a strategic perspective, in this 

environment, for any two firms of asymmetric size, both compete dynamically over 

prices to win market share. In this dynamic process there are two ways to achieve 

(temporary) monopolistic status. The ‘smaller’ firm can use dynamic pricing competition 

to delay the time in which the ‘larger’ firm wins a critical market share in the hope to hit 

the innovation first and displace it. If the probability of innovation and the discount factor 

are sufficiently high, there is an equilibrium in which duopoly persists (no firm achieves 

a critical market share) until one of the two firms wins the race for innovation. 

We take a historical example from markets for IT products. In the personal computer 

(PC) operating system market. Microsoft products (MS DOS and Windows) have been 

dominating the market since the mid 1980s. An operating system is the fundamental 

programme that controls the allocation and use of computer resources. Thus, the utility 

that operating systems provide to consumers depends on the number of compatible 

applications. As a general rule, an application that relies on a specific operating system 

will not function on another operating system unless it is ported to that specific operating 

system. Therefore, because of its dominance, the majority of applications have been 

written to run on Microsoft operating systems (MS DOS). The domination of MS DOS 

has become even stronger since the arrival of Windows 95 in the PC operating system 

market. This, in turn, has provided a great indirect positive network externality to PC 

owners who adopted Windows 95 as an operating system. Many other firms, such as IBM 

68



242 H-W. <strong>Gottinger</strong> 

and Be Inc., introduced their own operating systems and tried to compete with Windows. 

These products, however, lacked sufficient compatible applications to efficiently compete 

with Microsoft products. The lack of compatible applications prevented enough 

application developers and consumers from regarding OS/2 Warp or BeOS as a viable 

alternative to the dominant incumbent, Windows. This obstacle prevented these potential 

entrants from obtaining a sizeable market share. Their failure to enter the market 

successfully, however, was not due to the inferior quality of their operating systems. 

In fact, OS/2 Warp is reported to be at least as good as Windows, and BeOS offers 

superior support for multimedia applications. If consumers who use multimedia 

applications frequently adopt Windows at the expense of BeOS, they have to give up the 

convenience that is provided by BeOS. Thus, for multimedia specific users, adopting 

BeOS as their operating system at the expense of another operating system might provide 

the highest utility. Nevertheless, the lack of compatible applications, which in turn 

implies the lack of positive network externality, has prevented consumers from adopting 

OS/2 Warp or BeOS. As a conclusion, the positive network externality for the dominant 

incumbent (Windows) has worked as an entry barrier against entrants (OS/2 Warp or 

BeOS), which do not have network externality. Such an entry barrier can only be 

overcome by a radical innovation, virtually leapfrogging the dominant incumbent. 

2 Market histories 

We substantiate our claim on a natural tradeoff between network size and quality change 

(through significant technological change) by looking independently on four market 

histories for the IT industry. 

2.1 Internet explorer 

The first version of Internet Explorer was introduced in the web browser market in July 

1995. This version of Internet Explorer was reportedly inferior to Netscape’s Navigator. 

Microsoft made many improvements in its next version, Internet Explorer 3.0, in August 

1996. Internet Explorer 3.0 matched nearly all the features in Navigator. However, 

Microsoft was not able to draw significant market share at the expense of Navigator with 

this enhanced version. Internet Explorer 4.0, released in late 1997, finally managed to 

obtain roughly the same number of reviewers who preferred it to Navigator as the number 

of those who preferred Navigator. In 1996, most consumers were much more familiar 

with Navigator, the dominant web browser than with Internet Explorer. Thus although the 

quality of Internet Explorer was much improved since its release into the market, more 

than 80% of consumers were still using Navigator until late 1996. It was evident that 

Internet Explorer could not draw a significant percentage of consumers away from 

Navigator by simply improving the quality of its web browser. Senior executives within 

Microsoft concluded that Internet Explorer could not catch up with Navigator’s market 

share by itself. J. Allchin, senior executive within Microsoft, wrote to P. Martiz, in late 

December 1996: 

“I don’t understand how IE (Internet Explorer) is going to win. The current 

path is simply to copy everything that Netscape does, packaging and product 

wise. Let’s {suppose} IE is as good as Navigator/Communicator. Who wins? 

The one with 80% market share. Maybe being free helps us, but once people 

are used to a product it is hard to change them. Consider Office. We are most 

expensive today, but we’re still winning.” 

69




From 1996 to 1998, Internet Explorer’s market share increased dramatically. 

The estimated market share of Navigator fell from 80% in January 1996 to 55% in 

November 1997, while that of Internet Explorer increased from 5% to 36% over the same 

period. These changes in market shares, however, were mainly due to the fact that 

Microsoft leveraged Windows to increase the usage of Internet Explorer, rather than 

constantly improving the quality of its browser. 

2.2 Server operating system 

The environment of the server operating system market is different from that of the 

PC-operating system in the sense that a server is maintained by a group of computer 

experts (not by novices as in the case of a PC) and is intended for specific purposes, 

rather than for general uses. Moreover, a choice of server operating systems requires 

many technical features, such as functionality, system management, performance, 

reliability (stability) and security. Functionality and performance are important technical 

features for both PCs and servers, but other aspects are not very important for PCs. This 

difference comes from the fact that while a PC is designed for use by one person at a 

time, a server is designed to provide data, services, and functionality through a digital 

network to multiple users. Most individual PC owners do not have an incentive to buy a 

server-operating system, since server operating systems generally cannot function 

efficiently on PC hardware. In other words, the consumers in the two markets are 

basically separated. According to Data Quest, the Unix System is currently the most 

widely used operating system platform. It expanded its market share from 36% in 1996 to 

43% in 1998. The newcomer in the server operating system market, Windows NT, also 

achieved great success during that period; its market share rose from 10% in 1996 to 16% 

in 1998. Linux licence shipments grew at a rate of 212.5%, accounting for more than 

17% of all server operating system units shipped, Other than Unix servers and Windows 

NT, Open VMS and Novel. 

NetWare and OS/2 are the major competitors in the server-operating system market. 

A small product differentiation in a server-operating system enables it to create and to 

maintain its own demand. In sum, a dominant product cannot eliminate other products 

with its prevailing positive network externalities when substitutability for products are 

low. 

2.3 OS/2 warp 

In late 1994, IBM introduced its Intel-only compatible operating system, OS/2 Warp, as 

an alternative to Windows. OS/2 was originally designed to be optimised for Intel 

microprocessors. This resulted in a lack of portability. Because the Intel compatible PC 

dominated the PC market, the lack of portability did not prevent consumers from 

adopting OS/2 as an operating system. Furthermore, OS/2 was not reported to be inferior 

to Windows. The greatest disadvantage of adopting OS/2 as an operating sytem was the 

lack of compatible software; approximately 2,500 applications were available. In 1995, 

IBM still attempted to compete with Windows 95 in the PC-operating system market 

with its release of OS/2 2.0. However, IBM figured that it would lose between 70% and 

90% of its sales volume if it installed OS/2 to the exclusion of Windows 95. In 1996, 

IBM withdrew from the PC-operating system market and, instead, has been repositioning 

OS/2 for the server-operating system market. The failure of OS/2 Warp to gain enough 

70




market share to survive is mainly due to its lack of applications rather than inferiority to 

Windows. By the time OS/2 Warp was released, Windows was already the predominant 

operating system and ran an overwhelming number of applications. Most consumers used 

Windows not because it was superior in quality to its competitors, but rather because it 

provided the greatest positive network externality among the available PC-operating 

systems. 

In other words, Microsoft successfully deterred entry into the PC-operating system 

market by using its predominating network externality. 

2.4 BeOS 

Be, Inc., founded in 1990 by a former executive of Apple’s product division, released its 

cross-platform operating system, BeOS. BeOS functions virtually identically on both 

Intel-compatible PCs and Macintosh systems. In addition, BeOS is optimised for 

multiprocessor media-based systems. These two features provide great convenience for 

multimedia-specific users. Be has stated from the beginning that it intends to coexist 

within the Windows and Apple World as a specialised operating system. Even taking this 

specific purpose into account, the market share of BeOS is too small compared to the 

number of Windows users. The most outstanding disadvantage is the lack of compatible 

applications (approx. 1,000). Be, Inc. tried to overcome this problem by maintaining a 

website which provided a list of freeware and shareware applications. Be, Inc. failed to 

gain sizeable market share, and was eventually sold to Palm. However, Palm continues to 

produce BeOS and to maintain its BeWare website. BeOS still exists as a niche operating 

system, in spite of its superior support for multimedia applications. The failure of BeOS 

to obtain a significant market share is evidence in support of the proposition that the 

incumbent can deter entry by using its network externality. 

3 Some issues of technological racing 

Katz and Shapiro (1992) look at whether there is too much or too little technological 

innovation in a market with network externalities, compared to the social optimum. In 

particular, they explore whether a new product, which embodies technological progress, 

is introduced too early or too late compared to the social optimum. In their model, 

technological progress is deterministic. Only the entrant gets the benefit of improving 

technology, but once it enters the market its product is fixed. Therefore, in any 

equilibrium of their model, either the new firm captures the market as soon as it enters or 

it does not enter at all. Their model is thus not able to capture the richness of pricing 

strategies and market evolution as a function of market shares. For instance, it is unable 

to explain the phenomenon of two competing technologies fighting for superiority in a 

market when both are making sales until one firm finally leaves the market. Harris and 

Vickers (1987) present two models of races in which there is both technological 

uncertainty and strategic interaction between competitors as the race unfolds. 

Their aim is to see how the equilibrium effort levels of competitors vary with the 

intensity of rivalry between them. They show that under certain conditions the leader in 

the race makes greater efforts than the follower, and efforts increase as the gap between 

competitors in their progress decreases. The model presented here can be seen along 

71




similar lines to that of Harris and Vickers, but more realistically, with the network 

advantage as the state variable instead of progress, and prices providing strategic 

interaction between firms instead of the amount of effort put into R&D. While there is 

one ultimate goal for each firm in the multistage race of Harris and Vickers to reach the 

finishing line of the race first, we leave it open for firms to decide which network they 

want to build. It is this feature that can lend profit maximisation by firms to a complex 

network evolution pattern. Also, the amount of effort put into R&D stochastically moves 

firms to a new stage in the race, whereas in our model prices – the variables of strategic 

interaction – deterministically move firms to a new market share. 

While a lot of work has been done in the area of R&D, very little of it specifically 

addresses markets with network externalities. Katz and Shapiro look at the choice of 

product compatibility in a market with technological progress. They look at both the 

cases of deterministic and uncertain technological progress. However, technological 

progress in their model is exogenous and they look at a two-period game, which does not 

provide a very suitable framework for looking at the dynamic pricing strategy of firms 

and the resulting evolution of market shares. Choi (1994) looks at a two-period model of 

a monopoly in a market with network externalities. He studies the incentive of the 

monopolist to introduce an incompatible improved product in the presence of network 

externalities. Kristiansen (1998) studies the consequences of network externalities on the 

riskiness of R&D projects chosen by an entrant and an incumbent. He shows that the 

incumbent chooses a risky project that very often lets a new firm with an incompatible 

technology enter. In addition, the entrant has an incentive to choose more certain projects 

than are socially optimal and these strengthen the possibility of adoption of an 

incompatible technology. 

This paper develops and analyses a model of technological uncertainty where firms 

compete strategically in prices in the presence of network externalities. Technological 

uncertainty is modelled as an improvement in the quality of the firm’s product. 

Whether the firm gets the innovation, however, is uncertain. In the simplest model, 

once a firm gets the innovation, the race ends and there is no further innovation. 

Once the innovation occurs, when there is no uncertainty, the model looks like that of 

Katz and Shapiro – the firm with the quality plus network advantage sells to all the 

consumers subsequently. We attempt to analyse how firms share the market by 

strategically choosing their prices before they succeed in getting an innovation. 

The advantage of a firm is measured by the sum of its network size and the quality of 

its product. If the firms are in a technological race but the probability of innovation is 

small, then again, in the absence of innovation, the firm with the bigger advantage sells to 

all the consumers. If, however, those firms expect innovation with a high probability then 

so long as neither firm has a very big advantage both firms take turns in selling to the 

consumers in the absence of innovation. Thus, if the firms are on an equal base, neither 

firm gets pushed out unless one of the firms gets an insurmountable advantage in the 

process. That is because with firms that are sufficiently close in terms of their advantage, 

neither firm finds it worthwhile to push the other firm out of the market since the ‘losing’ 

firm prices more aggressively than the ‘leader’ as it has more to gain by staying in the 

market. 

72




4 Asymmetric R&D races 

We provide some anecdotal observations across various industries that show a strong 

rivalry between a leader in the field and the nearest rival(s) on the technological front and 

its efforts to gain market share. 

We identify those interactive patterns as technological races. We can even extend 

those observations to megamergers and alliances as products of network economies that 

did not protect those networks from relentless technological attacks by upstarts, new 

entrants or Schumpeterian entrepreneurs hailing ‘creative destruction’. This asymmetric 

view reflects the match between the ‘Old Guard vs the Vanguard’ (Hamel and Switzer, 

2004) in the sense that an incumbent’s position endows her with a natural network 

advantage which could only be overcome by a breakthrough innovation of a new entrant 

depending on some probability of innovation success. We have manifestations of intense 

rivalry and ‘head-to-head’ competition across industries, be it advanced microprocessors 

between Intel and American Micro Devices (AMD), (Markoff and Lohr 2003), specialty 

pharmaceuticals against chronic diseases between Merck, Glaxo and Pfizer, and between 

biotech companies Amgen and Biogen. In the expanding market of software related web 

services it is Microsoft against IBM, Sun Microsystems and Oracle (Lohr, 2003), in 

servers it is IBM vs Sun and H-P (Clark, 2003) in consumer electronics and design it is 

Sony against Matsushita or Sanyo (Belsen, 2003). 

5 Modelling a race in a network 

We start out with a simple model in which two firms, Y and Z, are manufacturing a 

product using incompatible technologies: the product is infinitely durable. The firms have 

products of identical qualities that we normalise to zero. For simplicity the marginal cost 

of production is assumed to be zero for both firms. 

We define: The network base of firm i at time t, b it , is the total number of consumers 

it has sold to until time t, i ∈ {Y, Z}. 

The network advantage of firm i at time t, n it ≡ b it – b jt , i, j∈ {Y, Z}. 

A new consumer arrives in the market periodically. Each consumer has an inelastic 

unit demand for the product and must decide the firm from which to buy the product as 

soon as he/she enters the market. Consumers care about how many other consumers buy 

the same product because of network externalities. That is, they are likely to choose 

products of the larger network over those of the smaller. They also care about the quality 

of the product and its price. For simplicity, all ‘representative’ consumers have identical 

preferences. We may state their utility functions as u (n t , q t, p t ) where q t is the quality of 

the product as a perfect substitute for the network advantage and p t is its price. 

Throughout the phase of adding uncertainty to the innovation process the consumers are 

expected to be risk-neutral. We assume that there is a threshold difference, D, in network 

sizes of the two firms such that if neither firm’s network is D (critically bigger than the 

other’s) then consumers get the same network benefits from the two firms. More 

realistically, one could think of D not just as a number but also as an ‘order of 

magnitude’. Therefore, if neither firm has a network advantage of D then consumers buy 

from the firm which charges a lower price if the quality is the same. Consumers are 

indifferent about the two firms if they charge the same price. Once a firm establishes a 

73




network advantage of D over the other firm, consumers get a much bigger benefit from it. 

Thus, if both firms charge the same price then consumers prefer the firm that has the 

network advantage of D (or more). It follows that once a firm gets a network advantage 

of D it becomes a (temporary) monopolist of the market. 

If both firms did not have identical qualities to start out with, then we could think of 

the network advantage as comprising the network difference plus the quality difference, 

that is, treat network and quality as perfect substitutes in the consumer’s utility. 

The proposed rule for consumer decisions implies that consumers are myopic, that is, 

the current base of consumers with the firms is what matters to them. While in practice 

we would expect consumers to form expectations about the future networks of firms, and 

thus introduce the possibility of multiple fulfilled-expectations equilibria, our assumption 

of myopic consumers would help select the most efficient equilibrium outcome among 

them. Both assumptions – of myopic consumers and consumers with fully rational 

expectations – are extreme ones, and what is more realistic is something in between the 

two. Having myopic consumers has the added advantage of making the analysis more 

tractable. 

Let n Yt represent the state of the system (we could equivalently choose n Zt to be the 

state variable). Thus state 1 denotes the situation where firm Y has one (order of 

magnitude) more consumer than firm Z, state – 1 denotes the situation where firm Z has 

one (order of magnitude) more consumer than firm Y. We define 3 to be the absorbing 

state for firm Y and – 3 the absorbing state for firm Z, that is we let D equal 3. The net 

discounted payoff of a firm in its absorbing state is M, for the other firm likewise. The 

absorbing state can be thought of as the threshold advantage in quality or network, which 

is big enough to deter the other firm from ever trying to catch up. Thus the firm that has 

this advantage becomes a monopolist in the market. It is the sole seller in the market 

making monopoly profits of M whereas the other firm leaves the market and thus gets 

zero profits. 

The product is subject to technological innovation. We assume that an innovation 

pushes a firm to its absorbing state. Thus, we assume that the innovation is equivalent to 

a big enough network advantage that gives a firm an insurmountable lead over the other 

firm, in many given cases of asymmetrical network advantage the innovations need to be 

‘radical’. The innovation, however, is uncertain, either a firm gets the required quality 

improvement or there is no improvement at all. The probability that a firm gets the 

innovation in any time period is α ∈ [0,1/2] that is the same for both firms. We could 

think of α > 0 as representing the technological environment in which the firms operate. 

We assume that once one firm succeeds, there are no further innovations in the market. 

We also assume for notational simplicity that at most one firm gets the innovation. Thus, 

the probability that neither firm gets the innovation in any one time is 1 – 2α. In this 

situation, we look at the innovation process as ‘exogenous’ (though we could make it 

‘endogenous’ by letting the probability of innovation success depend on the network size 

as a proxy for the degree of accumulated knowledge (learning) in determining innovation 

success). 

We could think of the technological innovation as a new feature that the firms try to 

introduce in their respective products. Once one firm gets its innovation, it can get a 

patent even though the firms produce incompatible products. We assume that the new 

improved product that a firm might introduce is compatible with its older product. 

74




Compatibility, in this model, translates into network sharing – compatible products share 

a network. 

Firms compete in prices and thus must decide what price to charge at each of the 

states in S = {–2, –1,0, 1,2}, in order to maximise their respective profits given by 

π Y (s 0 ,p) = ∑ t δ t (1–2α) t–1 [(1–2α)p Yt x Yt + αM] + δ τ+1 I(s τ+1 = 3)M, t = 0,1,2, …. , τ 

π Z (s 0 ,p) = ∑ t δ t (1–2α) t–1 [(1–2α)p Zt x Zt + αM] + δ τ+1 I(s τ+1 = 3)M, t = 0,1,2, …. , τ 

x 

it 

s 

t+ 

1 

⎧1 

⎪ 

= ⎨0 

⎪ 

⎩ 

0 or 1 

= s 

t 

+ x 

if p 

if p 

if p 

Yt 

it 

it 

it 

, 

< p 

> p 

= p 

jt 

jt 

jt 

where p it is the price charged by firm i in period t, x it ∈ {0, 1} is the number of period t 

consumers that the firm i sells to: τ + 1 is the time period in which a firm reaches its 

absorbing state even without getting the innovation: s 0 ∈ S is the initial state; and 

p = (p 1 , p 2 ), where 

p i = { p it }, t = 0,1,2,...,τ. 

We thus have a game where the firms are identical except in the sizes of their networks. 

There is uncertain technological innovation but the size of the innovation and the 

probability of innovation are again identical for both firms. Therefore, by comparing the 

solution of a game with no uncertainty with the solution of our game, we should be able 

to gain valuable insights into the effects uncertainty has on the competition between two 

firms in the presence of network externalities. 

We look for Markov equilibria (Fudenberg and Tirole 1993), which are 

non-collusive and symmetric. In such an equilibrium we consider a firm that is indifferent 

about selling or not selling. Thus if firm Y is the firm not selling at state s, then 

p Y (s) + δ ∏ Y (s+1) = δ ∏ Y (s–1) 

where ∏(s) represents the continuation payoff starting from state s. If Z is the firm not 

selling at state s then 

p Z (s) + δ ∏ Z (s–1) = δ ∏ Z (s+1). 

In such an equilibrium the non-selling firm ensures that even if against all expectations 

the consumer buys its product then it will not make negative profits in equilibrium. 

Definition 1 An equilibrium is non-collusive if neither firm charges a positive price 

at a non-absorbing state. Alternatively, if either firm charges a positive price at a 

non-absorbing state, equilibrium is collusive. 

Definition 2 An equilibrium is symmetric if p Y (s) = p Z (–s) for all permissible s and 

x Y (s) = x Z (–s) for all permissible s ≠ 0. 

If instead of having an infinite horizon game, we look at a game with a long but finite 

horizon, the non-collusive equilibria are the ones we expect would most likely be 

sustainable. 

We first consider the case in absence of uncertainty, that is, α = 0. 

75




Proposition 1 For α = 0, the symmetric non-collusive Markov equilibrium prices are 

unique: the firm with a positive network advantage makes all the sales in equilibrium. 

Proof: Suppose firm Z sells at s = 2. If firm Y were to sell to the consumer at s = 2, 

then in the next period it would get a profit of M. Therefore, if in a non-collusive 

equilibrium firm Y does not sell at s = 2 as we have assumed, it must be getting a 

continuation payoff at least as big as M next period. There are two possibilities for s = 1: 

either firm Y sells or firm Z does. 

Suppose firm Y sells at s = 1. Then once the state reaches 1, it must alternate between 

1 and 2 forever more from then on. The continuation payoff to firm Y at state 1 then is 

just the discounted sum of the price it charges to the consumer at s = 1 every other period. 

For this to be at least M, the price at s = 1 must be positive and equilibrium is then 

collusive. Suppose firm Z sells at s = 1 also. Then once s = 1 is reached, it must remain 

within the states –1, 0, 1 forever more. But then again for firm Y’s continuation payoff at 

s = 1 to be at least M, it must get a positive payoff at s = 0 or –1, a contradiction. Thus, Y 

must sell at s = 2. Suppose firm Z sells at s = 1 and Y sells at s = 2. Then once state 1 is 

reached, the state must remain within the states –1, 0, 1. In a non-collusive equilibrium 

where neither firm charges a positive price, Y’s continuation payoff at s = 0 cannot be 

positive. Z’s continuation payoff at s = 2 is zero. Similarly, the smallest price Z is willing 

to charge at s = 1 cannot be less than zero. However, firm Y’s continuation payoff at s = 

2 is then positive and at most zero at s = 0, it is therefore willing to charge a negative 

price at s = 1 in order to lure the consumer into buying its product. But that means the 

consumer should buy from firm Y at s = 1 since its price is lower than Z’s, a 

contradiction. Thus firm Y must sell at s = 1 also. 

We have so far shown that if a symmetric, non-collusive Markov equilibrium exists, 

then it must have the firm with the network advantage selling. We now show that such an 

equilibrium does indeed exist and that it is (almost) unique. 

If in an equilibrium the firm with the network advantage always wins then the 

continuation payoffs of both firms at s = 0 must be zero in a symmetric equilibrium since 

both firms are willing to give up exactly the same surplus in order to win the consumer. 

This surplus is equal to the continuation payoff from winning the current, which has to be 

at least zero. 

At s = 1, the lowest firm Z is willing to charge is zero since whether it sells to the 

current consumer or not, it is going to get a payoff of zero from then on. But that means it 

must charge a price of zero at s = 2 also. The continuation payoff at s = 0 from winning is 

then δ 2 M. 

Thus, both firms charge –δ 3 M at s = 0 and get zero discounted payoffs. It can be 

checked that these are equilibrium prices. 

Thus, the unique equilibrium prices are as follows: 

p(2) = p(–2) = 0 

p(1) = p(–1) = 0 

p(0) = – δ 3 M. 

Both firms charge the same prices. Firm Y sells to the consumer at s = 1, 2,... firm Z sells 

to the consumer at s = –1, –2, either firm could sell at s = 0. 

Thus, we see that in the absence of uncertainty, the firm that is already ahead in terms 

of network size is the ‘winning’ firm. The other firm that has a network disadvantage is 

76




unable to overtake the firm that is ahead. This result is similar to that obtained by Katz 

and Shapiro (1992). Once a firm acquires a bigger network, the other firm is unable to 

overcome its disadvantage. Since in our model, quality and network are perfect 

substitutes in the consumer’s utility, we could replace network by quality to get the same 

result. The firm with the higher quality sells to all consumers, the lower quality firm is 

unable to push out the better quality firm from the market, given that both firms start out 

with the same network size. 

We now introduce uncertainty into the game and see how firms behave. As in the 

previous case, the only difference between the firms is in their networks. The uncertainty 

faced by firms is symmetric: both firms have the same probability of getting an 

innovation and the effect of an innovation is symmetric for both firms. 

Proposition 2 There exist δ * and α * such that for δ * < δ < 1 and 0 < α < α * the 

non-collusive symmetric equilibrium prices are unique. The firm with the bigger network 

sells to the consumer, either firm could sell when s = 0. 

Proof: 

The proof proceeds along a sequence of Lemmas. 

Since we are looking for symmetric equilibria, at s = 0 the lowest price both firms are 

willing to charge must be the same. Thus, in equilibrium both firms must be indifferent 

between selling or not at s = 0. In other words, at s = 0 both firms have the expected 

payoffs. Again, since we are looking at symmetric equilibria we need to only focus on 

behaviour at s = 0, 1, 2. The different possibilities at s = 1 and s = 2 determine the 

different candidates for equilibrium outcomes. There are four possible candidates for 

equilibrium network advantage evolution, given that no innovation occurs in the 

meantime. 

(i) x Y (1) = 1, x Y (2) = 1, 

(ii) x Z (1) = 1, x Y (2) = 1, 

(iii) x Z (1) = 1, x Z (2) = 1, 

(iv) x Y (1) = 1, x Z (2) = 1 

where x i (s) = 1 means that at state s firm i sells to one consumer (and thus j ≠ i sells to no 

one since there is only one consumer in the market at any time period) in the absence of 

innovation having taken place. 

Define p i * (s) to be the lowest price firm i is willing to charge and p(s) to be the 

equilibrium price at s. The lowest a firm is willing to charge at any state is just the 

difference between the profits if it lost the current consumer and if it won it. In other 

words, the maximum a firm is willing to give up in order to attract a consumer is exactly 

the amount of extra profit the firm can expect to make by selling to that consumer. Define 

π i W (s), π i L (s) to be the expected profits of firm i next period onwards from respectively 

winning and losing the current consumer at state s, let π i (s) be the ex ante expected 

payoff of i at state s. 

We prove the proposition using the following Lemmas: 

Lemma 1 

Proof: 

Case (iii) cannot occur in equilibrium. 

Suppose (iii) described the equilibrium network evolution. Then, 

p Y * (1) = p(1) = δ(1 – 2α)p(0) (1) 

77




This is just the difference in payoffs to firm Y from losing versus winning the current 

consumer. 

Similarly, 

π Y W (1) = 0 + δαM + δ (1 – 2α) π Y (1) 

π Y L (1) = (1 – 2α)p(0) + δαM + δ (1 – 2α) π Y (1) 

p Y * (1) = π Y L (1) – π Y W (1). 

p Z * (0) = p(0) = δ(1 – 2α)p(1) (2) 

Equations 1 and 2 together imply that 

p(1) = δ 2 (1 – 2α) 2 p(1). 

But that, in turn, implies that either p(1) = 0 or δ = 1 and α = 0. Since we are looking at 

the case where α ≠ 0 , p(1) must be zero, which means from (2), p(0) = 0. Thus 

π 

Y 

αM 

(0) = π 

Z 

(0) 

≡ π(0) 

1 − δ( 

1 − 2α) 

. 

We can calculate p Z * (2) and p Y * (2) (= p(2) since Y is the losing firm at s = 2) 

p Z * (2) = – δαM – δ 2 (1 – 2α)π(0) 

p Y * (2) = δαM + δ 2 (1 – 2α)π(0) – δM. 

But for δ < 1, p Y * (2) one selling to the 

consumer at s = 2 , a contradiction. 

Lemma 2 Case (iv) describes the network evolution in equilibrium only when 

δ 2 (1 – 2α) 2 > 1/2 . 

Proof: π Z W (1) = 0 + δαM + δ (1 – 2α) π Z (1) 

π Z L (1) = (1 – 2α)p(2) + δαM + δ (1 – 2α) π Z (1) 

p Z * (1) = p(1) = π Z L (1) – π Z W (1) 

⇒ p(1) = δ(1 – 2α)p(2) (3) 

We can calculate p(2) = p Y * (2) using the same method: 

δαM δ(1−2 α) 

p(2) = 

1−δ( 1−2α ) 

p(1) 

−δM 

− 

1 −δ 2 

( 1 − 2 α ) 2 

) 

* 

z 

2 

2 

δαM δ (1− 

2α) 

(2) = 

− 

2 

1− 

δ( 

1− 

2α) 

1− 

δ ( 1− 

2α) 

p 

2 

p(2 

(4) 

Substituting the expression for p(2) in (3), we get, 

2 

2 

2 

2 

δ (1− 

2α) 

αM 

δ (1− 

2α) 

2 

p(1) = + 

p(1) − δ (1− 

2α)M 

(5) 

2 

2 

1− 

δ(1− 

2α) 

1− 

δ (1− 

2α) 

Similarly, we calculate p(0) using the relation of (3) to get p(0) = 0 since 

78




2 2 

δ (1−2 α) δ(1−2 α) 

p(0) = p(2) − 

1 − (1−2 ) 1 − (1−2 ) 

2 

2 2 2 

δ α δ α 

p(1) . 

This in turn implies that p Y * (1) = 0. 

For case (iv) to occur in equilibrium, p(1) must be non-negative. That leads to p Y * (1) 

≤ p(1) ; it also implies that p(2) is non-negative, therefore p Z * (2) ≤ 0 ≤ p(2) from Equation 

(4). From (5) we see that p(1) is non-negative only if δ 2 (1 – 2α) 2 ≥ 1/2. 

Lemma 3 As δ → 1 and α → 0, the pattern of network evolution described in (iv) 

can only occur in a collusive equilibrium. 

Proof: As δ → 1 , α → 0 , δ 2 (1 – 2α) 2 ≥ 1/2. Thus from Lemma 2 case (iv) can occur 

in an equilibrium. But when δ 2 (1 – 2α) 2 ≥ 1/2 , we see from (5) that p(1) > 0, that is, the 

equilibrium is collusive. 

Lemma 4 Case (i) describes the pre-innovation network evolution in a non-collusive 

equilibrium only when 3α [1 + δ (1 – 2α) ] ≤ 1. 

Proof: 

We calculate the lowest prices firms are willing to charge as described earlier on 

p Y * (2) = δαM + δ 2 (1 – 2α) αM + δ 3 (1 – 2α) 2 M – δM + δ (1 – 2α) p(1) + 

δ 2 (1 – 2α) 2 p(2) 

p Z * (2) = p(2) = − δ α M − δ 2 (1 – 2α) α M 

p Y * (1) = δ 2 (1 – 2α) αM + δ 3 (1 – 2α) 2 αM – δ 2 (1 – 2α) M – δ (1 – 2() p(2) 

p Z *(1) = p(1) = δ 2 (1 – 2α) αM – δ 3 (1 – 2α) 2 αM 

p(0) = – δ (1 – 2α) p(1) – δ 2 (1 – 2α) 2 p(2) – δ 3 (1 – 2α) 2 M 

= δ 3 (1 – 2α) 2 M [2α + 2δα (1 – 2α) – 1]. 

We can use the above equations to solve for p(0), p (1), and p(2), simultaneously. It can 

be checked that p Y *(2) ≤ p(2) always holds, but that p Y *(1) ≤ p(1) holds only if 

3α [1 + δ (1 – 2α) ] ≤ 1 (6) 

It can be checked that when inequality (6) is satisfied, then p(0), p(1), and p(2) are all 

negative. Note that (6) holds when δ → 1 and α → 0. Thus, when inequality (6) is 

satisfied the network evolution described in case (i) occurs as a result of the non-collusive 

equilibrium described by the prices given above. Both firms charge p(s), s ∈ {−2, −1, 0, 

1, 2}, where p(0), p(1) and p(2) are as given above and p(−2) = p(2), p(−1) = p(1). At s = 

1, 2 firm Y sells to the consumer in the market, at s = − 1, − 2 firm Z sells to the 

consumer, and at s = 0 either firm could sell. 

We see that the equilibrium when there is no uncertainty is just the limit of the above 

equilibrium as α → 0. 

Lemma 5 Case (ii) describes the pre-innovation network evolution in a noncollusive 

equilibrium only when 1 − 3δ 2 (1 – 2α) 2 ≤ 0 or when 

δ 4 (1 – 2α) 4 + 4 δ 2 (1 – 2α) 2 – 1 ≤ 0 and α − 3δ 2 (1 – 2α) 2 − (1 – 2α) (1 – δ) ≥ 0. 

Proof: Suppose (ii) describes the pattern of equilibrium network evolution in the 

absence of innovation. Then, p Z * (2) = p(2) and p Y * (1) = p(1): 

79




M (1 2 ) 

* (2) = δα δ − α 

− 

p(1) 

2 

1 − δ(1 

− 2α) 

1 − δ (1 − 2α) 

(7) 

p 

Z 

2 

2 

2 

M (1 2 ) 

*(2) = δα δ − α 

+ 

p(0) − δM 

2 

1− 

δ(1− 

2α) 

1− 

δ (1− 

2α) 

(8) 

pY 

2 

p 

Y 

2 

M (1 2 ) 

*(1) = δα δ − α 

2 

+ 

p(0) − δαM 

− δ(1− 

2α)p(2) 

δ (1− 

2α)M 

2 

2 

1− 

δ(1− 

2α) 

1− 

δ (1− 

2α) 

(9) 

2 

2 

M (1 2 ) 

* (1) = δα δ − α 

− 

p(1) δαM 

2 

1 − δ(1 

− 2α) 

1 − δ (1 − 2α) 

(10) 

p 

Z 

+ 

2 

p(0) = δ (1 – 2α) p(1) (11) 

Using Equations (7) and (11) to substitute in the values of p(2)and p(0) in Equation (9), 

we get 

M 

p (1) =− 

(12) 

⎡ 

2 2⎤ 

2 2 

⎢ ⎛ ⎞ 

1 12 

⎥ ⎛ ⎞ ⎛ ⎞ 

⎢ −δ 12 1 

⎜ 

− α⎟ ⎥δ ⎜ 

− α⎟ ⎜ 

−δ⎟ 

⎢ ⎝ ⎠ ⎥ ⎝ ⎠ ⎝ ⎠ 

⎣ 

⎦ 

⎡ 

⎤ 

⎡ 

2⎤ 

⎛ ⎞ 

1 12 

⎢13 2⎛ ⎞ 

⎢ −δ−α ⎥ 12 

⎥ 

⎢ ⎜ ⎟ ⎥ ⎢ −δ ⎜ 

−α ⎟ 

⎣ ⎝ ⎠⎦⎢ ⎝ ⎠ 

⎣ 

⎥ 

⎦ 

For (ii) to occur in equilibrium, it must be true that p Z * (1) ≤ p(1) and p Y * (2) ≤ p(2). It can 

be checked that p Y * (2) ≤ p(2) iff 1 − 3δ 2 (1 – 2α) 2 ≤ 0, or δ 4 (1 – 2α) 4 + 4 δ 2 (1– 2α) 2 – 

1 ≤ 0. It can also be checked that p Z * (1) ≤ p(1) iff 1 − 3δ 2 (1 – 2α) 2 ≤ 0 or α − 3δ 2 α 

(1 – 2α) 2 − (1 – 2α) ( 1 – δ) ≥ 0, thus proving the lemma. 

Lemma 6 

For δ → 1 and α → 0 the equilibrium described in Lemma 5 is collusive. 

Proof: For δ → 1 and α → 0, 1 − 3δ 2 (1 – 2α) 2 < 0, thus case (ii) can occur in equilibrium 

from Lemma 5. But it can be checked from (12) that for 1 − 3δ 2 (1 – 2α) 2 ≤ 0 and δ < 1, 

p(1) is positive. 

Thus from Lemmas 1 to 6 we can infer that as δ → 1 and α → 0, the (almost) unique 

non-collusive equilibrium shows the following properties. In the absence of innovation 

both firms charge the prices as below: As δ → 1 and α → 0, the equilibrium prices in the 

absence of innovation are: 

p(2) = p(−2) = − δαM − δ 2 (1 − 2α) αM 

p(1) = p(−1) = − δ 2 (1 – 2α) αM − δ 3 (1 – 2α) 2 αM 

p(0) = δ 3 (1 – 2α) 2 M [2α + 2δα (1 – 2α) – 1]. 

As α → 0, the equilibrium prices in the absence of innovation are: 

p(2) = p(−2) → 0 

p(1) = p(−1) → 0 

p(0) → −δ 3 M. 

Firm Y sells to the consumer in the market at s = 1, 2 and firm Z sells when s = − 1, − 2. 

When s = 0, either firm could sell but both earn the same expected profits when s = 0, 

that is, x Y (1) = x Y (2) = x Z (−1) = x Z (−2) = 1 and x Y (0) = 1 or x Z (0) = 1. 

80




Thus, in the limit as the probability of innovation approaches zero, the equilibrium 

approaches the non-collusive equilibrium without uncertainty. 

Proposition 3 There exist δ * and α * such that for δ * < δ < 1 and α * < α < 1/2, the 

non-collusive symmetric equilibrium prices are unique. The firm with network advantage 

of 2 and the firm with network advantage of 1 sell. 

Proof: For Lemmas 1, 2, 4 and 5, we infer that the only possible outcome in 

equilibrium has the firm with the advantage of 2 and the firm with the disadvantage of 1 

sell prior to an innovation occurring. From Lemma 5 we see that for δ close to 1 and α 

close to 1/2, the equilibrium prices are negative. The equilibrium is, therefore, noncollusive. 

As δ → 1 and α → 1/2, the equilibrium prices are as follows: 

p(2) = p(−2) → − M/2 

p(1) = p(−1) → 0 

p(0) → 0. 

Propositions 2 and 3 illustrate the effect of uncertainty on equilibrium behaviour in a 

market with network externalities. In Proposition 1 we saw that when the only difference 

between two firms is in the sizes of their networks of consumers, the firm with the bigger 

network sells to all consumers that subsequently enter the market. 

However, once we introduce uncertainty about the quality of a firm’s product this no 

longer is true. Even if the uncertainty faced by the firms is fully symmetric and the only 

difference between the firms is in the size of their respective networks, the firm with the 

bigger network does not always sell in equilibrium. For a sufficiently high success rate in 

innovation, it is the firm with the smaller network that sells provided the network 

distance is not too large. 

The reason for this strikingly different outcome compared to the situation without 

uncertainty is as follows: When there is no uncertainty, the only way either firm can 

realise positive profits is by reaching its absorbing state. Outside of their respective 

absorbing states both firms make the same profits that is zero. Thus the way for either 

firm to make positive profits is to sell to all the consumers and reach its absorbing state. 

However, since the firm with the larger network needs a smaller addition to its network to 

reach its absorbing state, it needs a shorter time span to reach its absorbing state, hence its 

discounted profits from selling to all consumers is larger. Not selling to any consumers, 

on the other hand, gives both firms zero profits. Thus, since the firm with the bigger 

network has more to gain from selling to all subsequent consumers while not selling 

gives it the same profits as the firm that is behind, it is willing to price more aggressively 

in order to lure the consumers. Thus, in equilibrium the firm that is ahead always sells. 

The introduction of uncertainty, even when it is uniform across the firms, alters the 

incentives of the firms to fight for the consumers. When the difference in network (size) 

between the two firms is sufficiently small and the probability of success is high the 

presence of uncertainty reduces the incentive of the firm that is ahead to fight for the 

current consumer relative to that of the firm with the network disadvantage. That is 

because with a high rate of innovation the leading firm has a high probability of getting 

its monopoly profits whether it sells to the current consumer or not. The firm with the 

network disadvantage, on the other hand, has a bigger stake in winning the current 

consumer since by not letting the other firm get close to its absorbing state, it keeps its 

own chances of getting the innovation and thus reaping the high monopoly profits alive. 

81




However, when the leading firm has a big network advantage, that is, is close to its 

absorbing state, its incentive to sell is much higher since it can then guarantee itself the 

monopoly profits by selling to the current customer. Thus as δ → 1 and α → 1/2, we see 

the firm with a network advantage of 2 selling and also the firm with a network 

disadvantage of 1. 

Until now all the discussion has been in terms of network advantage or disadvantage. 

We could equally well have had the discussion in terms of quality advantage or 

disadvantage since network and quality are perfect substitutes in a consumer’s utility. 

Thus without uncertainty, in the presence of network externalities, if the only difference 

between two firms was in the quality of their products then the firm with the better 

quality product would sell to all the consumers. With uncertain quality improvement 

however, if the probability of quality improvement were high enough then the firm with 

the lower quality would sell to the consumer. Both firms would remain in the market until 

one of the firms got an innovation that gave it an insurmountable lead over the other. 

6 Extensions and conclusions 

We have looked at a simple model of a market with network externalities where there is 

technological innovation. It would be interesting what kind of modifications to our basic 

model might change some of the results. The model is based on the assumption of 

symmetry in terms of the innovation that the firms expect. 

If we introduce asymmetric innovation this would permit us to endogenise innovation 

to the extent that we could look at the decisions of firms to improve quality. If we look at 

a more dynamic model of having sequential innovation and multiple races 

(<strong>Gottinger</strong> [2002b]) we might be able to see whether there is a clustering of innovations, 

and by examing the size of innovations, whether firms would rather tend toward one big 

risky innovation or rather for a series of small ones which are less risky. Under welfare 

theoretic considerations there may be related questions on whether there is too much or 

too little innovation in view of a social optimum (<strong>Gottinger</strong>, 2003). 

Our simple model shows how competing firms behave in a market with network 

externalities before the resolution of technological uncertainty. If our firms are in waiting 

(for the breakthrough innovation to come true) but the probability of innovation is small, 

then the firm that has a network advantage makes all the sales. However, if the 

probability of innovation is large, then it is not always the firm with the bigger network 

that sells. While a firm with a network advantage of 2 would make the sale, it is the firm 

with a network disadvantage of 1 that makes the sale if it is lucky to catch this 

innovation. 

Alternatively, it could also serve as an explanation for why a firm, that is in the end, 

driven out of the market may just hang on there for a while, with a network disadvantage, 

before it finally gives up and exits (or is exited from) the market. 

Furthermore, as some sort of by-product of this analysis, we could pursue the 

interaction between network effects and innovation for market structures. If network 

effects are ‘competitive barriers’ that could only be overcome by radical innovation then, 

by implication, the synergy of network effects and radical innovation through the 

incumbent would constitute a formidable powerhouse to consolidate monopolistic 

positions in dynamic product markets where ‘winner-takes-all markets’ are created. 

82




Acknowledgments 

I want to say thank to A. Aykac, Sophia Antipolis, J. McGee, Warwick, I. Png, 

Singapore, for their helpful comments and suggestions. 

References 

Belsen, K. (2003) ‘Sony again turns to designs to lift electronics’, New York Times 

(Online Edition), Feb. 02. 

Choi, J.P. (1994) ‘Irreversible choice of uncertain technologies with network externalities’, Rand 

Journal of Economics, Vol. 25, pp.382–401. 

Clark, D. (2003) ‘Sun unveils blade servers in bid to counter IBM, H-P’, Wall Street Journal 

(Online Edition), Feb.22. 

Evans, D.S. and Schmalensee, R. (2001) ‘Some economic aspects of antitrust analysis in 

dynamically competitive industries’, National Bureau of Economic Research, Cambridge, 

MA. Working Paper 8268, http://www.nber.org/papers/w8268. 

Fudenberg, D. and Tirole, J. (1993) Game Theory, MIT Press, Cambridge, MA. 

Futia, C. (1980) ‘Schumpeterian competition’, Quarterly Journal of Economics, Vol. 94, 

pp.675–695. 

<strong>Gottinger</strong>, H.W (2002b) Sequential Innovations and R&D Racing, http://www.csef.it/ publications 

2001/. 

<strong>Gottinger</strong>, H.W (2003) Economies of Network Industries, Routledge, London. 

<strong>Gottinger</strong>, H.W. (2002a) ‘Modeling stochastic innovation races’, Technological Forecasting and 

Social Change, Vol. 69, pp.607–624. 

Hamel, G. and Switzer, L. (2004) ‘The old guard vs. the vanguard’, The Wall Street Journal 

(Online Edition), Feb. 23. 

Harris, C. and Vickers, J. (1987) ‘Racing with uncertainty’, Review of Economic Studies, Vol. 54, 

pp.1–21. 

Katz, M. and Shapiro, C. (1992) ‘Product introduction with network externalities’, The Journal of 

Industrial Economics, Vol. 40, pp.55–83. 

Kristiansen, E.G. (1998) ‘R&D in the presence of network externalities: timing and compatibility’, 

Rand Journal of Economics, Vol. 29, pp.531–547. 

Lohr, S. (2003) ‘Competitors shape strategy to gain edge in web services’, New York Times (Online 

Edition), Feb. 3. 

Markoff, J. and Lohr, S. (2003) ‘Advanced micro issues challenge to rivals’, New York Times 

(Online Edition), April 19. 

Uttenback, J.M. (1994) ‘Mastering the dynamics of innovation’, Harvard Business School Press, 

Boston. 

83



84

1.6 HIGH SPEED TECHNOLOGY COMPETITION 


Int. J. Business and Systems Research, Vol. X, No. Y, XXXX 1 

High speed technology competition 

<strong>Hans</strong>-Werner <strong>Gottinger</strong> 

IMS, University of Maastricht, 

Maastricht, The Netherlands 

IIR, Hitotsubashi University, 

Tokyo, Japan 

E-mail: hg528@bingo-ev.de 

Abstract: We consider a technological race in which competitive firms engage 

in a class of differential games and some firms have priority of moves over 

others. The firm that has the right to move first is called the technological 

leader and the other competing firm is called the follower. An example of 

this type of sequential move game is the Stackelberg model of duopoly. 

The open-loop Nash equilibrium conditions in a sequential move game is 

derived and the results for technological race problems are derived as is a 

comparison of the strategies of leader and the follower. 

Keywords: industrial organisation; dynamic competition; differential games; 

technological leadership. 

Reference to this paper should be made as follows: <strong>Gottinger</strong>, H-W. (XXXX) 

‘High speed technology competition’, Int. J. Business and Systems Research, 

Vol. X, No. Y, pp.XXX–XXX. 

Biographical notes: <strong>Hans</strong>-Werner <strong>Gottinger</strong> has been the Director of the 

Institute of Management Science, Maastricht, NL and the Professor of 

Economics in the Institute of Innovation Research., Tokyo, Japan. He has 

internationally taught and widely published in the areas of industrial, energy 

and environmental economics and management. In these areas he has also been 

an advisor for international organisations. He is the author for the recently 

published book titled Economies of Network Industries (Routledge, 2003) and 

Innovation, Technology and Hypercompetition (Routledge, 2006). 

1 Introduction 

In highly competitive technological industries new challenges and opportunities are 

arising in the new product development arena. Driven by global markets, global 

competition, the global dispersion of scientific/engineering talent and the advent of new 

Information and Communication Technologies (ICT) a new vision of product 

development is that of a highly disaggregated, distributive process with people and 

organisations spread throughout the world. At the same time, products are becoming 

increasingly complex requiring numerous engineering decisions to bring them to market. 

Competitive pressures mean that ‘time to market’ has become a key to new product 

success. However, at the same time, it is important to keep the innovation and quality 

dimensions of the new product at their optimal level. 

Copyright © XXXX Inderscience Enterprises Ltd. 

85




A central question to address is, how should firms invest in innovation and what are 

the implications of such investments for competitive advantage? Understanding why a 

firm benefits from investments in innovation and quality illuminates issues of 

competitive strategy and industrial organisation. In the field of competitive strategy, 

much attention has been devoted to the concept of core capabilities (Teece et al., 1997). 

Understanding how firms make optimal investments in the face of competition reveals 

the nature of competition and provides theoretical and managerial implications for 

developing core competence and dynamic capabilities. 

In major parts of competitive analysis involving R&D decisions the focus is on 

breakthrough innovations which could create entirely new markets, for example, in 

studies featuring patent racing between competing firms. In more common competitive 

situations we observe firms, however, competing by investing in incremental 

improvements of products. It is an important aspect when innovation is considered to be 

manifested in product quality, process improvements and in the overall quality culture of 

an organisation (Kanter, 2006). For example, after product launch, incremental 

improvement of different aspects of product quality, improvements in various business 

processes and an incremental adoption of a quality culture are quite real-world 

phenomena. Some firms operate in a simultaneous product launch situation while others 

compete sequentially by adopting the role of leader or follower. The strategic 

implications in these diverse circumstances can be treated within a unified framework of 

dynamic stochastic differential games. 

In recent years, with the emergence of e-business and a supply chain view for 

product development processes, multiple firms with varying and at times conflicting 

objectives enter into collaborative arrangements. In such situations, the competitive 

strategy based on quality and innovation could potentially permeate in those 

collaborative setups. A recent example is the collaborative venture of Sony and Samsung 

to build a cutting edge plant for LCD flat screen TV though displaying ongoing fierce 

rivalry in new product launches in exactly the same product categories. When innovation 

and quality levels form the core of a firm’s capabilities, each member in the supply chain 

would have an incentive to invest and improve their dynamic capabilities. This leads to 

tacit competition among collaborative product development partners by means of active 

investment in innovation and quality. 

Although the forces of innovation are central to competition to emerging technically 

dynamic industries, they also affect mature industries where life cycles, historically, were 

relatively strong, technologies were mature and demands stable. 

A strategy for technology must confront primarily what the focus of technical 

development will be. The question is what technologies are critical to the firm’s 

competitive advantage. In this context, technology must include the ‘know-how’ the firm 

needs to create, produce, market its products and deliver them to customers. As a major 

step in creating a technology strategy it has to define those capabilities where the firm 

seeks to achieve a distinctive advantage relative to competitors. For most firms, there are 

a large number of important areas of technological ‘know-how’ but only a handful where 

the firm will seek to create truly superior capability. 

Having determined the focus of technical development and the source of capability, 

the firm must establish the timing and frequency for innovation efforts. Part of the timing 

issue involves developing technical capabilities, and the rest involves introducing 

technology into the market. The frequency of implementation and associated risks will 

depend in part on the nature of the technology and the markets involved (e.g. disk drive 

86



High speed technology competition 3 

versus automotive technology), but in part on strategic choice. At the extreme, a firm 

may adopt a rapid incremental strategy, that is, frequent, small changes in technology 

that cumulatively lead to continuous performance improvement. The polar opposite 

might be termed the great leap forward strategy. In this approach, a firm chooses to make 

infrequent but large-scale changes in technology that substantially advance the state of 

the art (<strong>Gottinger</strong>, 2006, Chap. 6). 

As an example of the importance of innovation strategy in product development, we 

notice that IBM created and continues to dominate the mainframe segment, but it missed 

by many years, the emergence of the minicomputer architecture and market. 

The minicomputer was developed and its market applications exploited by firms such as 

Digital Equipment Corporation (DEC) and Data General (DG) back in the 1970s. 

Summarising, in a mainstream scenario, high speed technological competition is 

likely to be driven by 

1 companies competing in fast changing, diverse networks and 

2 companies confronted with ever shortening product/technology life cycles 

because of multiple interactions between current and emerging technologies and 

product diversity or what may be circumscribed as ‘combinatorial innovation’ 

(Varian et al., 2004). 

A clear example provided by Arthur (2000, p.3): 

“If you look at genomics it is certainly heavily based on biology, but it is also 

highly dependent on digital computing! 

Genomics, then, is as much a manifestation of digital computing as it is a 

manifestation of molecular biology. What happens is that some of the new 

technologies become base technologies, and these give rise to manifestations, 

some of which themselves become base technologies. Manfestations then 

appear at an even faster rate.” 

To represent the features of leader-follower type in a sequence of technological racing 

we consider a class of differential games in which some firms have priority of moves 

over others. 

The firm that has the right to move first is called the leader and the other competing 

firm is called the follower. A well-known example of this type of sequential move game 

is the Stackelberg model of duopoly. In this type of interaction the open-loop Nash 

equilibrium conditions in a sequential move game can be derived. It would lead to a 

comparison of the strategies of leader and the follower. 

Section 2 presents the essence of leader-follower type interactions through the 

format of a differential game which is accommodated to a competitive market 

situation in Section 3. The properties of their interactions are derived in Section 4. 

Section 5 handles various market asymmetries in the leader follower type situation. 

Finally, Section 6 summarises and draws up conclusions and directions for future 

research. 

2 A differential game formulation 

For our modelling effort we follow the notations and symbols as explained below. 

N: number of firms, T: finite time horizon for the strategies, i, j: superscripts to 

denote competing firms in a duopoly, t: an instant of time in the dynamic game setup, 

87




u(t): investments in innovation effort (expenditure per unit time), R(t): net revenue rate 

for the firm at t, R o 

: product category net revenue rate for the existing product, 

R 1 

: product category net revenue rate for the new product, x o 

: quality level of the existing 

product. 

Several examples illustrate the importance of innovation and quality as 

competitive weapons. In the computer mainframe arena IBM replaced Remington 

Rand and Sperry Univac as market leader in the 1950s, and its subsequent growth 

outperformed its competitors so rapidly that in 1963 its data process revenues were 

four times larger than the combined revenues of its eight main rivals in the US market 

(Hoffman, 1976). 

In the model, we assume that enhancements of quality are achieved by climbing a 

performance ladder that may squarely embrace technology but could expand to other 

criteria uniquely identified with quality. Let the quality of the product at time t be x(t). 

In the context of total quality management, as quality levels increase it becomes even 

more difficult to climb the performance ladder. The hypothesis behind the formulation is 

that a firm needs to make higher innovation investments targeted towards the 

improvement of quality. To capture this dynamics a negative feedback effect of the 

present state quality on the rate of change of product quality (x " (t)) is considered. 

The state dynamics is 

" 

# 

x ! t" $ K# u! t "$ % %x! t" 

(1) 

where K is proportional to the level of capital investment in development technology and 

L is the proportionality constant for the influence of present quality on the speed of 

further quality improvements, # is the innovation resource productivity parameter. 

Based on (1) the quality of the product at time t is 

t 

# 

x! t" $ xo 

* 

& 

( K# u! s "$ % %x! s" ' 

) & s 

(2) 

% 

The product quality provides a means for evaluating the product’s attractiveness in the 

market in the presence of other competing products. The firm’s market share is a 

function of both its own product quality and the product quality of rivals (the initial 

product quality x o 

may reflect the reputational capital of the firm, either low or high, and 

may induce her to different innovation investment efforts). 

The difference lies in the planning horizon. In this competitive setup, the 

planning horizon extends beyond the date of launch and the competing firms continue 

investing in product-process innovation until the end of the growth phase of a product. 

Further, 

+ x% 

R% 

( % , t - , 

i * 

p 

. x% / x% 

R! t" 

$ 0 

i 

. x !, 

" 

R 

' 

( , 

i * 

p 

, t - , 

. 

1 x !, " / x !, 

" 

The cumulative development cost of the new product at time t is given as: 

T*! t" 

$ * , 

+ u! s",& 

s (4) 

% 

(3) 

88




the firm’s cumulative profit at time t is determined as follows, 

, 2 ! t" $ TR! t" % T*! t" 

(5) 

where TR(t) and TC (T) are total revenues and costs at time t, respectively. The total 

revenue function is given by: 

TR! t" 

$ * , 

R! s"& 

s (6) 

% 

where R(3) is given in (3). The firm’s decision set is 4 $ {u(t)}. Notice that the firm 

precommits on the date of product launch T p 

. The cumulative profit function, T 2(5), is 

defined as the total profit by end of the window of opportunity with decision 564. 

The firm’s decision problem can be stated as, 

7 9 8 7 9 8 9 7 9 

8 

./x 

5 64 

, 2 ! 5 " $ TR 5 % T* 5 $ , 2 5 

(7) 

The combination of Equations (1)–(6) generates an explicit representation of firm i’s 

cumulative profit by the end of time horizon. This substitution yields: 

& x% 

, 2 9 75 

9 

8 $ ./x5 

64 : R 

% 

, 

i * 

( x% / x% 

, 

# 

xo 

/ * K 

% 

'# u! s "$ % %x! s"& 

s 

/ 

' 

% 

, 

# 

* 

xo 

/ * K# u! s "$ % %x! s"& s / x !, 

" 

% 

% 

* 

, 

% 

+ u! s",& 

s' 

;) 

p 

7 p 8 

R , , 

(8) 

* 

where x 

% 

and x j 

(T), respectively are the quality of existing and new products of the 

competitor. Considering a duopoly, the optimisation problem written above can be 

reformulated as a differential game problem with state variable for the firm i 

(the competing firm is represented by superscript j) given as x i (t); the control variable 

u i (t). In the terminology used in optimal control and differential games, the salvage term 

for firm i, 

i 

i 

x% 

< !, ( x !, ""141R% 

, 

i * 

x / x 

/ R 

' 

i 

p 

% % 

i 

, 

i i 

# 

i 

= x% / * K & 

' 

! "' 

% 

' 

! "& >7 % 8 

% ( u s ) % x s s , , 

p 

i 

, 

i i 

# 

i * 

& ' 

% 

/ *% 

' ( ) % 

' 

/ 

x K u ! s " % x ! s"& s x !, 

" 

(9) 

where 

i 

, 

* * 

! " 

& * 

# 

* 

x , $ x & 

% 

/ 

% 

2 

! "' 

% 

2 

! " 

' 

* : 

K ( u s ) % x s 

; 

& s (10) 

( ) 

89




3 A sequential differential game between two competing firms 

We extend the conceptualisation of a hyper-competitive scenario by considering a 

sequential differential game as being representative of a leader-follower or 

incumbent-entrant competitive situation. The game is characterised by information 

asymmetry where the follower is aware of the innovation and quality levels of the 

leader’s products. The motivation for considering this scenario is its close 

correspondence with many real life competitive cases. Knowing the investment strategy 

of the leader, the rival firms can formulate their own strategies. Therefore, the firm acting 

as a leader chooses a decision path that maximises the objective for all conceivable 

responses that can be taken by the follower(s). In the case of sequential games a 

hierarchical play differential game approach is used to model the competitive situation 

and to obtain the open-loop Stackelberg Nash equilibrium. The issue of subgame 

perfectness and commitment is extremely important in these solutions. Adding to the 

previous notation we let T p 

, date of product launch by leader, T p 

/ ?, date of product 

launch by follower. The leader is represented by the superscript i and the follower is 

represented by the superscript j. Since it is a sequential game the time of product launch 

for the two players is such that the leader launches the product at time T p 

. Later, the 

follower launches the product after time ? at T p 

/ ?. A continuous improvement in the 

product is considered and therefore the entire time interval t 6 @0, TA for the leader and 

t 6 @T p 

B TA for the follower needs to be optimised. The state dynamics of the leader is 

" i i 

# 

i 

$ & ' 

' 

% 

' 

x ! t" K ( u ! t ") % x ! t" 

(11) 

The rate of quality improvement x "i (t) increases with investments u i (t). The factor L 1 

x i (t) 

suggests a natural decay in quality in the absence of any investments. The differential 

game formulation for leader is: 

i i i 

7 p 8 = 7 8 7 8> 

* , 

% 

./x , C u ! t "(, $ % u ! s"& s / < ,( x !, " 

(12) 

where 

s. t. 

i i i 

i 

R% x , 

% p 

/ R' 

x !, 

"! " ' 

i * i * i i 

% 

/ 

% 

/ 

% 

/ 

7 p 8 

? R x !, " , %, 

%? 

i i 

< 7, ( x !, 

" 8 D / 

x x x !, " x x !, " x !, 

" 

3 

i 

# 

i i i i 

x t $ K & 

' 

u t ' % %' x t x $ x 

% 

, x , 

(13) 

! " ( ! ") ! "( !%" ( 4ixe&( ! "47ee 

(14) 

Next, the follower’s problem formulation is discussed. Owing to the sequential nature of 

the game, the information about the leader’s quality and about the leader’s investments is 

known to the follower. Specifically, it is assumed that the knowledge gained by the 

leader’s investments ‘spills over’ to the follower’s quality improvement dynamics. It is 

plausible to assume that in the context of a ‘leader-follower’ competition, the level of 

quality improvements of the follower indeed depends not only on its own innovation 

efforts but also on the knowledge pool available because of the leader’s investments. 

The fact that the leader often cannot wholly conceal his efforts or can he credibly 

announce the commitment he has made, make the situation quite complicated. To address 

these issues fully in this context it would require a subtle and rich analysis of games with 

incomplete informations (Robson, 1990). 

90




A formal treatment of ‘spillover effects’ would enrich this model. The spillover 

effect is modelled by considering a linear additive term M 2 

u i (t). The follower’s state 

dynamics is 

3 * * 

# 

* i 

$ & ' 

2 

% 

2 

/ 

2 

x ! t" K ( u ! t ") % x ! t" - u ! t" 

(15) 

It is assumed that the spillover is a function of the investments made by the leader u i (t). 

In this term M 2 

is a very small number which quantifies the amount of spillover from the 

leader to the follower. The term M 2 

u i (t) would restrict the analysis to the case where the 

follower emulates the innovation and product quality of the leader, as opposed to setting 

his own technology and quality standards. Now with all considerations wrapped up, a 

leader-follower dynamics can be stated. 

The follower’s differential game formulation is given as: 

, 

7 p 8 * = > 7 8 

, p 

, %, 

p * * * 

$ %* 

= u 7s / , 

p 8> & s / < 7, ( x !, 

" 8 

* * * * * 

./x , 2 u ! t "(, $ % u ! s" & s / < , ( x !, 

" 

where < j (T, x (T)) is defined as, 

s.t. 

* 

x 

< !, ( x !, "" 4R % 

, 

x 

* 

% 

i * 

% 

/ x% 

p 

% 

* 

= x !, " 7, %, 

p 

%? 

8> 

x 

/ R 

/ 

x !, " x x !, " x !, 

" 

* 

% i 

% 

? R 

* ' 

/ 

% 

i 

/ 

* 

* * 

# 

* i 

$ & ' 

2 

% 

2 

/ 

2 

x ! t" K ( u ! t ") % x ! t" - u ! t" 

(18) 

where M 2 

u i 

(t) represent the spillover of knowledge, assumed to be a function of 

investments by the leader, M 2 

is a small constant such that 0 - M 2 

-- 1. 

where 

Furthermore, x 

* !%" $ x 

* 

% 

( T fixed, x j (T) free 

i 

t 

i i 

! " 

& i 

# 

i 

x , $ x & 

% 

/ 

% 

' 

! "' 

% 

' 

! " 

' 

* : 

K ( u s ) % x s 

; 

& s 

( ) 

* 

t 

* * 

! " 

& * 

# 

* i 

x , $ x & 

% 

/ 

2 

! "' 

% 

2 

! " / 

2 

! " 

' 

* : 

K ( u s ) % x s - u s 

; 

& s 

, p ( ) 

(16) 

(17) 

4 Analysis of the model and discussion 

A conventional tool for solving the problem is the application of Pontryagin’s maximum 

principle for open-loop Stackelberg equilibrium conditions. Interested readers can refer 

Dockner et al. (2000) for details regarding Pontryagin’s approach to solving sequential 

91




differential game problems. The equilibrium results are expressed in the form of 

properties as given below. 

4.1 Equilibrium results 

Property 1: The maximised costate variables for the follower are function of time and are 

given by, 

E 

9 $ E e 

(19) 

' 

E 

% 2t 

%' 

E e 

9 ' 

$ (20) 

2 %2 

% t 

where 

E 9 ' 

and E 9 

2 

are the costate variables reflecting the marginal price for a unit 

increase in follower firm’s own state and the state of the leader i, E 1 

(0) $ E 01 

, E 2 

(0) $ E 02 

are known constants. 

Property 2: The Stackelberg equilibrium investment by the follower in product 

development is given as: 

* 

* 9 * % 

'8!' " 

2t 

%# 

! " & 

2 

# E ' 

%' 

u t $ ( K e ) (21) 

Next the leader’s problem is investigated. The leader knows the follower’s best response 

to each control path u i (3). 

Property 3: The maximised costate variables for the leader are given by, 

F 

F 

F e 

9 %' 

t 

' 

$ 

%' 

(22) 

9 % 2t 

2 

$ F 

%2e 

(23) 

F F e 

% 

9 $ % 2t 

(24) 

9 %9 

where F 9 ( ' 

F 9 2 

and F 9 

9 


increase in the leader’s own state, the state of the follower and the costate of the follower. 

F 01 

, F 02 


are constants. 

Property 4: The Stackelberg equilibrium investment by the leader in product 


i 

u ! t" 

t 

i % 

'8!' " 

' t 

%# 

9 

& K' 

# F 

%' 

e ' 

$ : % 2t 

; 

'% 

- 

2 

F 

%2e 

( ) 

Property 5: The equilibrium state trajectory of performance improvement of the follower 

is given as: 

(25) 

& 

x t % x e - 

i 

'8!'%# 

" 

%% 2t 

i %' 

t 

* e 

& 

2 

'# F ' 

%' 

! " : 

* 

% t K e 

$ 

2 % 

/ 7% '/ 8 2 : % ; 

2t 

% 

2 

: ('% 

- 

2 

F 

%2e 

) 

( 

% 2t * % 2t 

7 ' e 8 K2 7K2 # E%' 

e 8 

/ % / 

* * 2 

8!' % " 

# # 

' 

;) 

(26) 

92




Property 6: The equilibrium state trajectory of performance improvement of the leader is 

given as: 

i i 

' 

i % 

8!' " 

' t 

# %# 

%% t 

i e 

& 

i 

%' 

i 

& K' 

# F 

%' 

e ' 

' 

x ! t " $ :%' x% / 7% '/ e 8 K' 

: % 2t 

; 

; 

%' : ('% 

- 

2 

F 

%2e 

) ; 

( ) 

For the analysis, the leader and the follower are assumed to be symmetric and therefore 

the corresponding values for the follower j are also assumed to be the same. 

The first observation is that the costate variable of both the leader and the follower 

increases more rapidly with time 

The marginal utility of a unit increase in quality, exhibits a convex increasing trajectory. 

With comparable values of parameters the plot of the costate trajectory of the follower 

has an upward exponential drift with time. The plot begins at time t $ T r 

, since the 

follower initiates product development activities only after the leader has already 

launched the product. Regarding the costate trajectory of the leader we see that F 1 

represents marginal utility from a unit increase in quality of the leader’s product. 

The costate variable for both the leader and the follower firm increase for the 

entire planning horizon. This follows since the leader and follower are competing on the 

basis of their quality levels. Firms increase their market shares based on their 

relative quality to that of the competitor’s. After launch the product faces its introduction 

and growth stage. 

In these phases, maintaining higher quality becomes even more important since the 

product sales are directly influenced by product quality. This leads to a convex 

increasing trajectory for the costate variable. Thus, as a second observation. 

The investment in innovation made by the follower increases more rapidly with time 

The investment strategies of the follower are purely a function of its own costate 

variable. The follower’s costate variable increases in a convex fashion. This in turn 

results in a convex increase in follower’s control trajectory. With identical parameter 

values we can derive that the follower compensates for a delayed entry into the market 

by increasing its investment intensity. Such an increase in investment results in increased 

quality and therefore high total revenue for the follower. Thus we observe. 

The investment in innovation made by the leader initially increases rapidly with time but 

later increases at a decreasing rate 

The leader’s investment trajectory is sigmoidal or S-shaped. Because of the nature of the 

game, the leader enters the market before the follower. While formulating his 

equilibrium investment strategy, the leader takes into account the evolution of his own 

costate and also the costate of the follower. 

Being early in the market, the leader takes into account all possible courses of action 

that a follower may choose. 

We note that the leader’s rate of investment increases in the initial phase and then 

starts to decrease. That is, the leader capitalises on the advantage of early market entry by 

increasing the intensity of investments and gaining a high market share. Subsequently, 

after the follower’s entry, the leader reduces the intensity of investments and thereby 

reduces the amount of spillover that is potentially possible. 

(27) 

93




When the parameter values of the leader and follower are identical, the follower invests 

higher than the leader 

In the course of time, initially the follower maps its investment strategy to that of the 

leader. However, subsequently the leader starts reducing the rate of investments, while 

the follower continues with the high investment rate based strategy. 

As a further observation, the rate of increase of the follower’s product quality increases 

with time 

The quality of follower’s product increases for the entire time-horizon. Moreover, 

this increase is a convex function. As can be inferred from the problem formulation, the 

quality level of the follower is a function of investments made by the leader and the 

follower. Specifically, the quality trajectory is influenced by both the convex profile of 

investments made by the follower and also . 

by the spillover effect of the investments 

made by the leader. This results in a convex increase in quality improvements. The state 

trajectory is therefore upward sloping in convex form. 

The follower starts accruing revenue only after T p 

/ ?, where T p 

is the date of launch 

of the leader. In the game context of quality-based competition, the follower 

compensates for the delayed entry by increasing the quality levels at a fast rate. 

The rate of increase of the leader’s product quality decreases with time 

The leader’s investment trajectory follows an S-shaped trajectory. In the problem 

formulation, the leader's quality improvement is only a function of its own investments. 

Moreover, it is assumed that the leader doesn’t obtain the advantages of spillover of 

knowledge gained by the follower’s investments. Furthermore, the leader needs to 

deliberately avoid maintaining very high investments to ensure that the follower 

doesn’t achieve huge gains from spillovers. Under such a setup it must be noted 

that the leader chooses an S-shaped trajectory for investments. In such an investment 

profile the investment increases rapidly in the initial phase but then eventually 

tapers off. Corresponding to this investment profile, the rate of quality improvement 

is high in the initial phase but then eventually the rate of improvement starts 

decreasing. Thus, the quality trajectory of the leader is concave. This leads to another 

observation. 

When the parameter values of the leader and follower are identical, the quality of the 

follower’s product is higher than that of the leader’s product 

With identical parameter values the follower and leader are perfectly symmetric in their 

capabilities. Moreover, the leader has the advantage of earlier market entry, whereas the 

follower obtains the gains of information asymmetry and the associated knowledge 

spillovers from the leader. It is reasonable to assume that the follower employs his 

inherent competence and the benefit of spillover of knowledge from the leader’s 

investments to increase quality at a much faster rate. Under such circumstances the 

overall gain and loss in the competitive game is dictated by the relative revenues 

achieved and the costs incurred by the two players. With higher investments than the 

leader the follower incurs higher costs. At the same time these investments also lead to 

higher quality levels and therefore revenues for the follower. In the event of the value 

of R 1 

being very high the follower wins the game while the results favour the leader if 

R 1 

is low. 

94




5 Firm asymmetries 

We now turn to firm asymmetries and their implications. It is likely that firms take up the 

role of leader and follower based on inherent strengths and weaknesses. Therefore, an 

explicit consideration of firm asymmetries is very important and could potentially lend 

more insights in a sequential game setup. Asymmetries may be addressed in at least four 

ways: Firstly, it is possible that organisational approaches and techniques (such as total 

quality management) can be used to make the product development process more 

cost-efficient and effective. This would influence the value of the resource productivity 

parameter #. Secondly, it is possible to have an advantage in product development by 

making capital investment in technology development. This is equivalent to considering 

asymmetries in the value of K. Thirdly, it is possible that the obsolescence parameter 

of firms denoted by L is different. Finally, different values of M suggest different 

levels of spillover of knowledge from the leader’s investments to the follower. The 

firm with higher value of L has a higher obsolescence or decay in quality. We 

restrict assessment of asymmetries by considering different values for the parameters: 

(# i and # j ), (K 1 

and K 2 

), (L 1 

and L 2 

) and (M l 

and M 2 

). 

5.1 Innovation resource productivity 

Parameters # i and # j denote the innovation resource productivity parameter of the two 

firms. An asymmetry could result if the competing firms have differing capabilities in 

making a productive use of investments. The skill set of employees, training and 

development activities, the quality culture are some of the reasons for such an 

asymmetry. For example, (# i $ 0.2) G (# j $ 0.1) suggest that firm i has a higher 

innovation productivity over j. An investigation of different values of # suggest that the 

firm with higher innovation productivity also invests a higher amount in product 

development. From the results we observe that the follower has a relatively higher level 

of investments in innovation as compared to the leader. 

Therefore, an increase in the value of # of the follower will only result in making 

these investments still higher. Instead, an increase in leader’s # would provide some 

interesting insights in that it will show that the leader sails ahead with the follower 

having no chance in catching up. It clearly shows that the leader has a higher investment 

than that of the follower for almost the entire planning horizon. As can also be noted, the 

quality of the leader’s product is always higher than that of the follower. Hence, an 

asymmetry in terms of resource productivity parameters provides a totally different result 

than what was observed when the competing firms were symmetric. It suggests that the 

leader with an advantage in terms of early market entry as well as a higher level of 

resource productivity indeed continues having higher revenues. Under this situation the 

leader doesn’t worry much about the spillover to the follower since its own resource 

productivity enables attaining higher revenues by increasing quality levels relative to that 

of the follower. With a high value of R 1 

it can be conjectured that the result favours 

the leader. 

5.2 Capital investment parameters 

Firm asymmetries can be analysed by evaluating different values for K and the 

implications on control and state trajectories. K l 

and K 2 

are proportional to the level of 

95




capital investment in development technology by the two competing firms. If 

(K 1 

$ 20) G (K 2 

$ 10), it suggests that the leader i has higher levels of capital 

investment in development technology as compared to the follower j. The implications of 

a higher value of K for the leader can easily be derived. The parameter K exerts a 

positive effect on innovation investments. Similar to the conjecture regarding different 

values of #, the intuition behind this effect is that with a higher level of capital 

investments in development technology, a firm targets the investments towards 

increasing the product quality. Thus with a higher value of K the investment by the 

leader is relatively higher than that of the follower for most part of the planning horizon. 

However, in the latter part of the planning horizon, the follower’s investment in fact 

shoots up while that of the leader tapers off. Such an investment profile would have an 

impact on the state trajectory. Furthermore, the product quality of the leader is higher 

than that of the follower for most part of the planning horizon. Eventually, at the end 

of the planning horizon, the high growth in the lower’s investment results in higher 

quality than that of the leader. In the model, the relative difference in resources due to 

strategic investments is captured in asymmetries in the value of K. Given adequately high 

values of R 1 

, the leader i would have relatively higher profits than the follower j. 

5.3 The obsolescence parameter 

L 1 


characterise the decay or in other words the obsolescence effect of quality. 

Different values of L 1 


help evaluate the difference among the two firms regarding 

this effect. (L 1 

$ 1) G (L 2 

$ 0.7) suggests that the obsolescence effect for leader i is higher 

than that for the follower j. An investigation into the dynamics of innovation investments 

reveals that with higher value of the parameter L, a firm would make higher level of 

investments. Interestingly, with a difference in L 1 


the shape of investment 

trajectory of the leader is now convex. That is when the leader faces a high obsolescence 

effect the investments are increased at a faster rate to ensure an increasing state 

trajectory. However, as opposed to parameter a and K such an increased control 

trajectory doesn’t translate into higher quality. In fact, in this situation, the leader has 

lower product quality inspite of higher investments. A high decay effect puts the leader 

into a disadvantageous position. In such a situation, under equilibrium control the leader 

always loses the competitive game. 

5.4 Spillover factor M 2 

M 2 

characterises the amount of spillover from the leader to the follower. In this section 

the impact of reduction of the spillover factor M 2 

is considered. Evidently in this case the 

leader would be less worried about the amount of advantage the follower obtains from 

the leader’s own investments. If the spillover effect is low, the leader pursues 

investments more aggressively, consistent with the incentive enhancement of spillover 

effects in the context of network externalities (Saaskilahti, 2006). Under such 

circumstances, the control trajectory of the leader is also convexly increasing and is only 

marginally lower than that of the follower. 

Regarding the impact of such investments on the state trajectory, the quality of the 

follower is higher than that of the leader and both obtain a convex increasing state 

trajectory. But unlike #, K and L, a different value of M 2 

doesn’t correspond to any 

advantage related to resources, core competence or dynamic capabilities. This is purely 

96




an exogenously defined variable based on technology and industry types. Under changed 

circumstances with respect to M 2 

the leader doesn’t really gain much in the competitive 

game. The results in fact point to the follower winning the game due to much higher 

levels of quality and only marginally higher levels of investments. 

6 Conclusions and extensions 

There is apparently a paradox in trying to assess, both empirically and theoretically, the 

impact of competitive pressure on innovation and growth. On one hand, according to the 

tradition originating in Schumpeter (1942) the prospective reward provided by monopoly 

rent to a successful innovator is required to stimulate sufficient R&D investment and 

technological progress. 

On the other hand, the incentives to innovate are weaker for an incumbent 

monopolist than for a firm in a competitive industry (Arrow, 1962). When competition is 

intense in the product market, innovation may even be seen as the only way for a firm to 

survive. In neo-Schumpeterian models of endogenous growth (Aghion and Howitt, 1992; 

Grossman and Helpman, 1991), innovation allows a firm in an industry to take the lead 

and gain profit. But the monopoly rent enjoyed by the winner is only temporary, and a 

new innovator, capitalising on accumulated knowledge, is always able to ‘leapfrog’ the 

leader unless the leader is endowed with advantages of firm asymmetries. In the recent 

research literature, Aghion et al. (1997) and Aghion et al. (2001), supposing a duopoly in 

each sector, both at the research and production levels, have introduced what they call 

‘step-by-step innovation’, according to which technological progress allows a firm to 

take the lead, but with the lagging firm remaining active and eventually capable of 

catching up. This model has been extended by Encaoua and Ulph (2000), allowing for 

the possibility that the lagging firm leapfrogs the leader, without driving it out of the 

market. Here we adopt a simplified approach to consider spillover effects. Griliches 

(1979) lays out the conceptual framework and provides an early discussion of the 

importance of spillover effects of R&D. Later in Griliches (1992) reviewed the recent 

empirical evidence on spillovers, and tentatively concluded that spillover effects may be 

substantial. The model can be extended by evaluating the improvement in product quality 

with learning effect. This effect is very well documented in the literature and thus the 

extension to incorporate learning effects is straightforward. Specifically, learning can be 

used to characterise the dynamics of evolution of product quality and the characterisation 

of costs associated with innovation investments. 

For analytical simplicity the revenue function is treated as a salvage value. As an 

extension the incorporation of the dynamics of revenue in the analysis of the model can 

be explored. In the sequential play game, the state dynamics of the follower can be 

considered to be dependent on the state of the leader at previous time instant. 

This modification would allow for analysis of a model in which the leader uses 

open-loop Nash equilibrium strategy for innovation investments whereas the follower 

would adopt a Markovian Nash equilibrium strategy. 

Additional insights can be gained by this modification for leader-follower 

competitive dynamics in new product development. Firm asymmetries can be explored 

by considering multiple parameters at the same time. This would enable a richer 

understanding of the strategies that a leader and a follower should adopt based on their 

strengths and weaknesses. 

97




The conventional approach used in innovation and R&D research viewed the process 

as one with constant returns, competitive output and factor markets and no externalities. 

However, such a framework doesn’t offer a full explanation of productivity growth. 

For a better understanding it therefore becomes very important to consider increasing 

returns to scale, R&D spillovers and other externalities and disequilibria. 

References 

Aghion, P. and Howitt, P. (1992) ‘A model of growth through creative destruction’, Econometrica, 

Vol. 60, pp.323–352. 

Aghion, P., Harris, C. and Vickers, J. (1997) ‘Competition and growth with step-by-step 

innovation: an example’, Review of Economic Studies, Vol. 41, pp.771–782. 

Aghion, P., Harris, C., Howitt, P. and Vickers, J. (2001) ‘Competition, imitation and growth with 

step-by-step innovation’, Review of Economic Studies, Vol. 68, pp.467–492. 

Arrow, K. (1962) ‘The economic implications of learning by doing’, Review of Economic Studies, 

Vol. 29, pp.155–173. 

Arthur, W.B. (2000) ‘Myths and realities of the high-tech economy’, Talk given at Credit Suisse 

First Boston (CSFB) Thought Leader Forum, 10 September. 

Dockner, E., Jorgensen, S., VanLong, N. and Sorger, G. (2000) Differential Games in Economics 

and Management Science, Cambridge: Cambridge University Press. 

Encaoua, D. and Ulph, D. (2000) ‘Catching-up or leapfrogging’, Cahiers de la MSE, Vol. 5, 

pp.10–22. 

<strong>Gottinger</strong>, H.W. (2006) Innovation, Technology and Hypercompetition, London: Routledge. 

Griliches, Z. (1979) ‘Issues in assessing the contribution of Research and Development to 

productivity growth’, The Bell Journal of Economics, Vol. 10, No. 1, pp.92–116. 

Griliches, Z. (1992) ‘The search of R&D spillovers’, The Scandinavian Journal of Economics, 

Vol. 94, pp.S29–S47. 

Grossman, G. and Helpman, E. (1991) Innovation and Growth in the Global Economy, Cambridge, 

MA: MIT Press. 

Hoffman, W.D. (1976) ‘Market structure and strategies of R&D behavior in the data processing 

market – theoretical thoughts and empirical findings’, Research Policy, Vol. 5, pp.334–353. 

Kanter, R.M. (2006) ‘Innovation: the classical traps’, Harvard Business Review, November, 

pp.72–83. 

Robson, A.J. (1990) ‘Duopoly with endogeneous strategic timing: stackelberg regained’, 

International Economic Review, Vol. 31, pp.25–35. 

Saaskilahti, P. (2006) ‘Strategic R&D and network compatibility’, Economics of Innovation and 

New Technology, Vol. 15, No. 8, pp.711–733. 

Schumpeter, J.A. (1942) Capitalism, Socialism and Democracy, New York: Harper. 

Teece, D.J., Pisano, G. and Shuen, A. (1997) ‘Dynamic capabilities and strategic management’, 

Strategic Management Journal, Vol. 18, No. 7, pp.509–533. 

Varian, H.R., Farrell, J. and Shapiro, C. (2004) The Economics of Information Technology, 

Cambridge: Cambridge University Press. 

98



Appendix 


Mathematical Preliminaries, Definitions and Theorems 

This section covers some of the salient analytical aspects specific to a sequential play 

game. 

For a finite horizon T let L and F denote the leader and follower respectively. Let 

x denote the vector of state variables, u L the vector of control variables of the leader and 

u F the vector of control variables of the follower. Assume x 6 R n , u L 6 R mL and u F 6 R mF . 

Definition 1: The initial value of the follower’s costate variable E i 

is said to be 

non-controllable if E i 

(0) is independent of the leader’s control path u L (t). Otherwise, it is 

said to be controllable. 

The definition suggests that if the costate variable is controllable, the follower’s control 

variable u F (t) at time t depends also on future values of u L (3); that is, on values u L (s) with 

s G t. 

Theorems and Proofs 

Theorem 1: The maximised costate variables for the follower are a function of time and 

is given by 

where 

E 

9 $ E e 

(A1) 

' 

E 

% 2t 

%' 

E e 

9 ' 

$ (A2) 

2 %2 

% t 

E 9 ' 


2 


increase in the follower firm's own state and the state of the leader i; E 1 

(0) $ E 01 

, E 2 

(0) $ 

E 02 


Proof: Stackelberg equilibrium conditions are derived by constructing the Hamiltonians. 

The analytical solution is derived for the follower firm j by writing the Hamiltonian as: 

* 

* * # 

! " E & 

* * i 

/ $ % & ' 

' 2 

! " 

2 

! " 

2 

! "' 

( u t ) / K u t % % x t / - u t 

( ) 

i 

# 

/ E & 

i 

i 

2 

K' u ! t " % %' 

x ! t" 

' 

( ) 

(A3) 

The necessary conditions for optimality are: 

* 

/ $ % 

(A4) 

* 

u 

From (A5) 

E $ % / : E $ % / 

(A5) 

9 * 9 * 

' * 

x 2 

i 

x 

x !%" $ x 

(A6) 

* * 

% 

E E E E 

9 9 % 2t 

' 

$ % 

2 ': 

' 

$ 

%' 

e 

(A7) 

E E E E 

9 9 %' 

t 

2 

$ %' 2: 

2 

$ 

%2e 

(A8) 

99




where E 1 

(0) $ E 01 

is the known positive constant denoting the initial value of the 

costate E 1 


(0) $ E 02 

is a known negative constant denoting the initial value of the 

costate E 2 

. 

Theorem 2: The Stackelberg equilibrium investment by the follower in product 

development is given as 

* 

* 9 * % 

'8!' " 

2t 

%# 

! " & 

2 

# E ' 

%' 

u t $ ( 

K e ) 

(A9) 

Proof: Differentiating (A3) with respect to u j (t), we obtain: 

* 

* * 

# %' 

u* 

$ %'/ E 

'# 

& 

2 ( ! ") 

' 

/ K u t 

(A10) 

By Equation (A10) to zero and some algebraic manipulations, the following expression is 

obtained for optimal effort in product development: 

* 9 * 9 

2 ' 

* 

' 8!' %# 

" 

u ! t" 

$ & ( K # E ' ) 

(A11) 

Substituting the expression for E i 

in (A11): 

* 

* 9 * % 

'8!' " 

2t 

%# 

! " & 

2 

# E ' 

%' 

u t $ ( K e ) 

(A12) 

Theorem 3: The maximised costate variables for the leader are give n by, 

F 

9 $ F e 

(A13) 

' 

%' 

t 

%' 

F 

F e 

9 % 2t 

2 

$ 

%2 

(A14) 

F F % 

9 $ % % 2t 

9 %9e 

(A15) 

where F 9 ( ' 

F 9 2 


9 


increase in the leader’s own state, the state of the follower and the costate of the 

follower; F 01 

, F 02 



Proof: The Stackelberg equilibrium conditions are derived by constructing the 

Hamiltonian. An analytical solution is derived for the leader firm I by writing the 

Hamiltonian as: 

/ & u t ' & K u t % x t ' 

i 

i i i # i 

$ % ( ! ") / F 

' 

! " % 

' 

! " 

( ) 

/ F & % / ' / F E 

( ) 

* 

* # 

* i 

2 

K2u ! t " % 

2x ! t" - 

2u ! t " 

9 '% 

2 

@ 

A 

(A16) 

The necessary conditions for optimality are 

i 

/ $ % 

(A17) 

i 

u 

F $ %/ : F $ %/ : F $ %/ E 

(A18) 

9 i 9 i 9 i 

' i 2 * 

x 

x 9 

' 

i 

x !%" $ x 

(A19) 

i 

% 

100



From (A18) 


F F F F 

9 9 %' 

t 

' 

$ %' ': 

' 

$ 

%' 

e 

(A20) 

F F F F 

9 9 

% 2t 

2 

$ % 

2 2: 

2 

$ 

%2e 

(A21) 

9 % 2t 

9 

$ % 

9% 2: 

9 

9 

$ % 

%9e 

(A22) 

F F F F % 

where F 9 ( ' 

F 9 2 


9 


increase in the leader’s own state, the state of the follower and the costate of the 

follower; F 01 

, F 02 



Theorem 4: The Stackelberg equilibrium investment by the leader in product 


i 

u ! t" 

i 

i % 

'8!' " 

' t 

%# 

9 

& K' 

# F 

%' 

e ' 

$ : % 2t 

; 

'% 

- 

2 

F 

%2e 

( ) 

Proof: Differentiating (A3) with respect to u i (t), we obtain: 

(A23) 

i 

i i 

# %' 

& ' 

' ' 

F 

2 2 

/ i $ %'/ F # K u ! t" 

/ - 

u 

( ) 

(A24) 

By Equation (A24) to zero and some algebraic manipulations, the following expression is 

obtained for optimal effort in product development: 

i 

u ! t" 

i 

i 9 

'8!'%# 

" 

9 

& K' 

# F ' 

' 

$ : 9 ; 

'% 

- 

2F 

2 

( ) 

(A25) 

Substituting the expression for 

i 

u ! t" 

i 

i % 

'8!' " 

' t 

%# 

9 

& K' 

# F 

%' 

e ' 

$ : % 2t 

; 

'% 

- 

2 

F 

%2e 

( ) 

F 9 ' 

and F 9 2 

in (A25) yields 

(A26) 

Theorem 5: The equilibrium state trajectory of performance improvement of the follower 

is given as 

& 

* * 

x ! t " $ % x / %'/ e - : ; 

i 

'8!'% 

# " 

%% 2t 

i %' 

t 

e 

& 

2 

'# F ' 

: 

% t K 

%' 

e 

2 % 7 8 2 

% 2t 

% : 

2 (: '%-2 F%2e 

; 

( 

) 

* * 2 

# 8!' % # " 

% 2t * % 2t 

' 

/ %'/ e K K # E e ; ) 

7 8 2 7 2 %' 8 

(A27) 

Proof: The expression for the optimal state trajectory can be obtained by considering the 

state dynamics given in Equation (11). Substituting u i (t) $ u i (t) 9 and u j (t) $ u i (t) 9 in 

(11) the following first-order differential equation can be obtained 

* i i 9 

* * 9 

# 8!' %# " 

* 

& K' 

# F ' 

' 

x ! t" $ K & 

2 ( K 

2 

# E ' 

' ) % % 

2x ! t " 7, p 

/ H % t8 

/ - 

2 : 9 ; 

('% 

- 

2F 

2 ) 

i 

'8!'%# 

" 

(A28) 

101




The first order differential equation can be solved with the initial condition x 

* !%" $ x 

* 

% 

; 

The resulting expression for x j (t) 9 is: 

& 

x t % x e - 

i 

'8!'%# 

" 

%% 2t 

i %' 

t 

* e 

& 

2 

'# F ' 

%' 

! " : 

* 

% t K e 

$ 

2 % 

/ 7% '/ 8 2 : % ; 

2t 

% 

2 

: ('% 

- 

2 

F 

%2e 

) 

( 

% 2t * % 2t 

7 ' e 8 K2 7K2 # E 

%' 

e 8 

/ % / 

* * 2 

# 8!' %# 

" 

' 

;) 

(A29) 

Theorem 6: The equilibrium state trajectory of performance improvement of the leader is 

given as: 

i i 

' 

i % 

8!' " 

' t 

# %# 

%% t 

i e 

& 

i 

%' 

t 

& K' 

# F 

%' 

e ' 

' 

x ! t " $ :%' x% / 7% '/ e 8 K' 

: % 2t 

; 

; 

%' : ('% 

- 

2 

F 

%2e 

) ; 

( ) 

(A30) 

Proof: The expression for the optimal state trajectory can be obtained by considering the 

state dynamics given in Equation (14). Substituting u i (t) $ u i (t) 9 in (14) the following 

first-order differential equation can be obtained: 

i i 

i 9 

# 8!' %# 

" 

* 

& K' 

# F ' 

' 

i 

! " $ 

' : ' 

! " 

9 ; % 

'% 

- 

2F 

2 

x t K % x t 

( ) 

(A31) 

i 

i 

The first order differential equation can be solved with the initial condition x !%" $ x% 

; 

The resulting expression for x i (t) 9 is: 

i i 

' 

i % 

8!' " 

' t 

# %# 

% t 

i 

%' 

t 

& K' 

# F 

%' 

e ' 

: 

' % 7 8 

; 

' 

% 2t 

' 

: '% 

2 

F 

%2 

; 

% 

i e 

& ' 

x ! t " $ % x / %'/ e K : ; 

% ( - e 

( 

) 

) 

(A32) 

102

1.7 SEQUENTIAL TECHNOLOGY CHOICE AND R&D RACING 


103


1.7 SEQUENTIAL TECHNOLOGY CHOICE AND R&D RACINGN 

104



105



106



107



108



109



110



111



112



113



114



115



116



117



118



119



120



121



122

1.8 INTERNET OF THINGS ECONOMIC AND INDUSTRIAL DIMENSIONS 


!"#$%"&#'("&)*+(,%"&)*(-*.,/'"$//*&"0*12("(3'2/*4$/$&%25* 

"#$%&!'()*+!$$),$"-! 

.//0+112223456785709:;64.68?1@16@:7=! 

A>6+!$#3$$'BC1@36@:7=3"#$%#'#)3$)! 

DEEF+!"-"C,%)B-!(G=68/*&!DEEF+!"-"C,%)'H!(I8;687*! ! 

! 

65$*!"#$%"$#*(-*65'"7/8*12("(3'2*&"0*!"0,/#%'&)* 

9'3$"/'("/* 

!"#$%&'(#'(%)*++,#-'(% 

E/=J/78>?654L!EMNOMKP!P>849;/683!)L!"#$%L!003!$$),$"-3!A>6+!$#3$$'BC1@36@:7=3"#$%#'#)3$)! 

6'4',7'1+!O9,2JS! 5>??9865J/6>8! :75J?7! 

5>??>8L!2.65.!78J:;7A!947=!0J=/6560J/6>8L!5>;;J:>=J/6>8!J8A! 

68/7=J5/6>83! W7:! "3#! /75.8>;>56J;! 87/2>=]68;>?7!74478/6J;!/>!?>A7=8! 

4>56J;!68/7=J5/6>8!J4!27;;!J4!X>=!:J;!:94687443!W.6;7!W7:! 

"3#!59==78/;S!A>?68J/74!/.7!D8/7=87/L!87/2>=]!78=]682J=A4!J8>/.7=!J;L!5>??>8;S!=7X7==7A!/>!J4! 

/.7!E7?J8/65!W7:!J8A!4>?7/6?74!=7X7==7A!/>!J4!W7:!-3#3!M.7! 

J;! >X! /.7! E7?J8/65! W7:! (7?:=J5689=]4*!64!/>!?J=]!90!27:!5>8/78/! 

68!J!2JS!/.J/!?J]74!6/!98A7=4/J8AJ:;7!:S!?J5.6874L!J;;>268! :7.J[7! ?>=7! 68/7;;6=?J/4!2>9;A!J;;>2! 

?J5.6874!/>!0=>5744!J8A!4.J=7!AJ/J!>8!/.76=!>28L!26/.>9/!/.7! 

877A! X>=! .9?J8! ?7A6J/6>83! O;>80?78/4! 68! 

D8/7=87/! /75.8>;>;>=! F7/2>=]4! J8A! 

F7J=! d67;A! P>??9865J/6>8! (FdP*! 9468! d=7c9785S! 

DA78/6/S! (NdDY*! /J! :778! 7[>;[688[7=X! /.747! /2>! /75.8>;>=! 

F7/2>=]4L! 64! ;7JA68! 872! 0>446:6;6/674! J8A! [646>843! M.7! 

0>446:6;6/S!>X!J!X=J?72>=]!/.J/!2>9;A!J;;>2!A6=75/!?J5.687, 

/>,! ?J5.687! 5>??9865J/6>8! >[7=! /.7! D8/7=87/! .J4! ;7A! 

=747J=5.7=4! />! 78[646>8! /.7! :787X6/4! >X! :=688;687!J8A!J;;>268!0J=/6560J/7!68!/.7!27:! 

J4! J! [J4/! 87/2>=]! >X! J9/>8>?>94L! 47;X,>=A957A!J!0J=JA6



! D8/7=8J/6>8J;!e>9=8J;!>X!a9468744!J8A!K5>8>?654!N747J=5.!"#$%&!'()*+!$$),$"-! $$'! 

! 

D8/7=87/,;6]7!87/2>=]3!Y7[6574!J/!.>?7!8>!?J//7=!/.76=!46`7L! 

J=7!=!0J446[7!NdDY!/J8A6/6>87=! J=7! J//J5.7A! />! J! /=J84576[7=! />! 

5>??9865J/7! 26/.! J! ;>5J;! 47=[7=3! M.64! ;>5J;! 47=[7=! 90AJ/74! 

/.7! 4784>=! >9/09/! AJ/J! 68/>! /.7! D8/7=87/! /.=>95J;! 

47=[7=3! M.7! AJ/J! X=>?! D8/7=87/! 64! J557447A! X=>?! J! A74]/>0L! 

?>:6;7!0.>87!>=!J8S!A7[657!/.J/!64!5>8875/7A!/>!/.7!D8/7=87/3! 

I87! >X! /.7! ?J68! JA[J8/JX! /.64! ]68A! >X! 87/2>=]68;[68?7! J00;6J8574! J8A! A7[6574! 64! /.J/! 

27! 5J8! .J[7! 5>8/=>;! >[7=! >9=! A7[6574! /.=>99/! J! :J;! 

=7J5.!2.7=7!5>??9865J/6>8!64!0>446:;73!M.7!0>/78/6J;!J=6474! 

>8! 87J=;S! 7[7=S! 45J;7! J4! 27! 5>9;A! 6?J8>?>94! 47;X,A=6[68874*L! 

4?J=/! .>?74! J8A! 4?J=/! 56/6743! D>M! 5J8! :7! 6?0;7?78/7A! 68! 

:9468744!>=!68A94/=6J;!78[6=>8?78/4!:J47A!>8!/.7!=7c96=7?78/! 

78J:;7A! /75.865J;;S! :S! ?J5.687,/>,?J5.687! (Q"Q*! 

5>??9865J/6>8!(O;0JSA68!["]*3! 

M.7=7! J=7! ?J8S! A7[6574! 5>8875/7A! />! /.7! D8/7=87/3! d>=! 

7\J?0;7L! />! :! 27! .JA! />! 84!?J89J;;SL!:9/!8>2!27!.J[7!0=68/7=4!/.J/!5J8!:7! 

5>8875/7A! 26=7;744;S! /.=>9?7! 87/2>=]3! M.7! 

0=68/68:!5J8!:7!A>87!X=>?!J8S2.7=7!J8A!>8!J8S!A7[657! 

495.!J4!J!5>?09/7=!>=!J!4?J=/0.>873!D8!/.7!4J?7!2JSL!J;;!/.7! 

J00;6J8574!J8A!A7[6574!J/!.>?7!5J8!:7!5>8/=>;;7A!=7?>/7;S3!D/! 

64! J5.67[7A! :S! J==J8=]!J8A!5>8875/68!J!A75646>8, 

?J]68=]! :75J947! /.7! 

A75646>8,?J]688!>8!/.7!4/J/94!>X!/.7! 

A7[6574!68!/.7!87/2>=]3! 

D8!J8!D>M!78[6=>8?78/L!/.7!5>8/=>;;68??9865J/6>8! 

>X! A7[6574! 68[>;[74! ?95.! 4784>=! 87/2>=]4! 26/.! 6809/! J8A! 

>9/09/!0>=/43! 

W.J/! A64/68M! X=>?! /=JA6/6>8J;! 87/2>=]! 

/75.8>;>,X>;A+! (6*! /.7! ^/.68X! A6[7=47! 

8J/9=7! 495.! J4! ?J5.687! 0J=/4L! 5>8/=>;! 986/4L! 5>??9865J/6>8! 

A7[6574LfL! (66*! /.7S! 78=7! 5>?0;7\! 

5>??9865J/6>8! 46?6;J=! J8A! 90! />! /.7! ;7[7;! >X! .9?J8! 

5>??9865J/6>83! a3! O=/.9=! [-]L! /.7! E/J8X>=A! 75>8>?64/! J8A! 

XJ/.7=! >X! ^685=7J4688A!K5>8>?S!J8A!7\7?0;6X674!6/!;6]7!/.64+! 

gf!6X!D_?!A=6[684!O82!>X!/=JXX65!2.7=7!?S!5J=!64!68! 

J! 5>8[7=4J/6>8! 26/.! /.7! 5J=4! J=>98A! 6/! /.J/! J=7! 68! J! 

5>8[7=4J/6>8! 26/.! 8A! 75>8>?S! 64! 5=7J/68=! 94! ,,,! 4;>2;SL! c967/;SL! J8A! 

4/7JA6;S!,,,!J!A6XX7=78/!2>=;Ah3! 

CA%D"4E-(*;#1%6'$'"(45%*#%B#+'(#'+%*F% 

35,#-$% 

D>M! ?7J84! 5>8875/68! /.7! D8/7=87/L! 6/! 5J8!:7! J8! 

>:@75/! >=! J! A7[6573! M.7! 5>??9865J/6>8! >X! A7[6574! >8! /.7! 

D8/7=87/,;6]7!87/2>=]!5J8!:7!>87!46A7A!>=!/2>!46A7A!68!/7=?4! 

>X! 87/2>=]68??9865J/6>8! 5J8! :7! J! 

46?0;7\! ]68A! >X! 5>??9865J/6>8! >=! A90;7\! 5>??9865J/6>83! 

M.7! D8/7=87/! .J4! :778! 9468X! /75.8>;>?! (5>??7=56J;*&! /.7! 5>??9865J/6>8! 

:7/2778! /.7! A7[6574! .J4! 7[>;[7A! J4! D8/7=87/! /75.8>;>07A3!M.7!872! 27:!A7[7;>0?78/!>8!/.7!27:!J=7! 

W7:! "! J8A! W7:! -! /75.8>;>;>;>=!A7[6574! 

/>! 5>??9865J/7! 26/.! /.7! 27:! A6=75/;SL! ?J]688,?J]6807=! 

J;=6/.?!/>!?J]7!J!A75646>8!J8A!J5/!J55>=A68=!684/J857L!5>846A7=!J8!J5/6[7!NdDY!5.60!7c96007A!26/.! 

J! =7;7[J8/! 4784>=&! /.64! NdDY! ;J:7;! 949J;;S! ?! /.7! 0>27=7A! 4784>=! 68! /.7! 5.60! J8A! 478A4! /.J/! 

AJ/J! />! /.7! 47=[7=! :9/! A>74! 8>/! /J]7! J8S! J5/6>83! V>27[7=L! 

2.78!27!??J8A4!68!5>?09/7=!;J8>47! J8! 

J00=>0=6J/7!J5/6>8!X=>?!/.7!X!J5/6>843!D8!/.64!5J47L! 

27! J=7! /J;]689/! ?J5.687,/>,?J5.687! (Q"Q*! 

5>??9865J/6>8! 2.7=7! /.7! ?J5.687! 5>??9865J/74! 26/.68! 

/.7?47;[74!J8A!07=X>=?4!J8!J00=>0=6J/7!J5/6>83!% 

D8A94/=6J;!A=6[68=574!68!/.7!D>M!5>8/7\/!! 

O55>=A68! D>M! A7[7;>0?78/L! D>M! :9468744! 7[>;[74! 

/.=>9! /7;7, 

>07=J/6>8! J8A! /7;7,0=747857! J4! J! =749;/! >X! /.7! 68/7=J5/6>8! 

X>=574! >X! :>/.! ?J=]7/! 09;;! J8A! /75.8>;>//680?78/!=>JA?J0!68A65J/74!/.J/!/.7!0=>;>8/689>94;S! 5>8/=6:9/7! />! /.7! 

A7[7;>0?78/! >X! D>ML! 2.6;7! /.7! 5>??7=56J;6`J/6>8! 68! /.7! 

?J=]7/!0;J57!64!J8>/.7=!]7S!64497!=7;J/68!/.7!0=>X! 

D>M3! M.7=7X>=7L! D>M! :9468744! 64! 0=>?>/7A! :S! 68A94/=6J;! 

A=6[68=574!>X!:>/.!/75.8>;>;>8!:76890?78/! >X! =7;J/7A! /75.8>;>0?78/!>X!D>M!/75.8>;>94!0>/78/6J;!/>! 

i=?4! 68! D>M! :94687443! d>=! 7\J?0;7L! ?J8S! 6?0>=/J8/! 

/75.8>;>9A!5>?09/68;>=! 87/2>=]! /75.8>;>?>/7A! /.7! A7[7;>0?78/! 

>X!D>M!:9468744!/>!J!872!;7[7;!(W.6/?>=7!7/!J;3![)]*3! 

E75>8AL!?J=]7/!09;;!64!A7i87A!J4!J8!688>[J/6[7!X>=57!:76807A! :S! i=?4! 68! =740>847! />! J8! 6A78/6i7A! J8A1! >=! 

0>/78/6J;! ?J=]7/! 877A! 5>8A6/6>8J;! >8! J! =?>94! ?J=]7/! A7?J8A! 68! D>M! :9468744! 

0=>[6A74! 980=757A78/7A! ?J=]7/! >00>=/986/674! />! i=?43! d>=! 

7\J?0;7L! D>M! J00;65J/6>84! 68! .>?7! J00;6J8574L! .7J;/.5J=7L! 

J8A! J9/>?>:6;74! J=7! 09;;7A! :S! 872! ?J=]7/! A7?J8A4! ;6]7! 

.>947.>;A!A7?J8A!X>=!.>?7!J9/>?J/6>8L!0J/678/!A7?J8A!X>=! 

594/>?6`7A! 47=[657L! J8A! 594/>?7=! A7?J8A! X>=! 68/7;;68J;! M7;75>??9865J/6>8! j86>8! (DMj*! X>=! 

684/J857! A7i874! /.7! D8/7=87/! >X! M.68=!/.7!D8X>=?J/6>8!E>567/SL!78J:;688875/688L! 7\64/68;[6807=J:;7! 68X>=?J/6>8! 

J8A! 5>??9865J/6>8! /75.8>;>



$$%! VJ84!W7=87=!R>//68X!M.688>?65!J8A!D8A94/=6J;!Y6?7846>84! ! 

! 

/.7! /.68?7! 5>8875/7A! 68! /.7! D>M3! I/.7=! 

A7i86/6>84!X>594!>8!D8/7=87/,=7;J/7A!J4075/4!>X!/.7!D>ML!495.! 

J4! D8/7=87/! 0=>/>5>;4! J8A! 87/2>=]! /75.8>;>8!47?J8/65!5.J;;78M!=7;J/68!/.7! 

4/>=J=8! >X! ;J=;9?74! >X! 

68X>=?J/6>8!(O/`>=6!7/!J;3![%]*3!! 

W.6;7!/.7=7!64!8>!986[7=4J;!A7X686/6>8!X>=!/.7!D>ML!/.7!5>=7! 

5>8570/! 64! /.J/! 7[7=SAJS! >:@75/4! 5J8! :7! 7c96007A! 26/.! 

6A78/6XS682! /.7?! />! 5>??9865J/7! 26/.! >87! J8>/.7=! J8A! 

26/.!>/.7=!A7[6574!J8A!47=[6574!>[7=!/.7!D8/7=87/!/>!J5.67[7! 

4>?7! 947X9;! >:@75/6[73! M.7! 5>=7! 4688570/4! 98A7=;S68M!J=7!8>/!8723!d>=!S7J=4L!/75.8>;>=! 87/2>=]4! .J[7! :778! 947A! 68! 68A94/=6J;! 

J8A! ?J89XJ5/9=688/7\/4! X>=! /=J5]685]3!M.7!6A7J!>X!A6=75/!?J5.687,/>,! 

?J5.687!5>??9865J/6>8!64!J;4>!8>/!/.J/!872L!J4!6/!64!:J465!/>! 

/.7!6A7J!>X!/.7!D8/7=87/!68!2.65.!5;678/4L!47=[7=4!J8A!=>9/7=4! 

5>??9865J/7!26/.!7J5.!>/.7=3!W.J/!/.7!D>M!=70=7478/4!64!J8! 

7[>;9/6>8!>X!/.7!947!>X!/.747!7\64/68;>X! 

/.7! 89?:7=! J8A! ]68A4! >X! A7[6574! J4! 27;;! J4! /.7! 

68/7=5>8875/6>8! >X! 87/2>=]4! >X! /.747! A7[6574! J5=>44! /.7! 

D8/7=87/3!d>=!7\J?0;7L!?>4/!A7[6574!59==78/;S!>8!/.7!D8/7=87/! 

27=7!>=6X!/.7!D8/7=87/!J8A!.J[7! 

68/7574468=J=]! 5J0J:6;6/6743! 

M.747!A7[6574!685;9A7!47=[7=4L!A74]/>04! ;J0/>04L!/J:;7/4!J8A! 

4?J=/!0.>8743!W.J/!/.7!D>M!0=>0>474!64!/>!J//J5.!/75.8>;>! 7[7=SAJS! A7[6574L! 495.! J4! J9A6>1[6A7>! =7576[7=4L! 4?>]7! 

A7/75/>=4L! .>?7! J00;6J8574L! 7/53! J8A! ?J]688;687L! 

7[78!6X!/.7S!27=7!8>/!686/6J;;S!A746/.7=!?J@>=!7[>;9/6>8J=S!5.J8?647A!:S!/.7! 

D>ML!64!/.7!68/78!>X!87/2>=]4!/.J/!5>8/J68!/.747!A7[6574L! 

?J]689=!7\J?0;7L!NdDY!.J4!:778!947A!X>=!S7J=4!/>!/=J5]!0=>A95/4! 

/.=>9X!/.7!4900;S!5.J683!V>27[7=L!>857!/.7! 

0=>A95/! ;7X/! /.7! 4.7;X! >X! J! =7/J6;! >9/;7/L! /.7! ?J89XJ5/9=7=! _4! 

J:6;6/S!/>!/=J5]!/.7!>:@75/!2J4!;>4/3!! 

b6]72647L! 5>849?7=4! 27=7! 98J:;7! />! ! /.7! 

;6X75S5;7!68X>=?J/6>8!>X!0=>A95/4! /.7S!09=5.J47A3!aS! 9M! 0=>?6474! />! 78J:;7! 0=>A95/! 

/=J57J:6;6/S!/.=>99/!/.7!78/6=7!0=>A95/!;6X75S5;73! 

M.7! 7?7=;>X!A7[6574!J8A! 

/.7=7X>=7! /.7! >=68570/! >X! D>ML! .7857! 64! 7\/78A68! 

J?:678/! 68/7;;68>?>94! 5>8/=>;3! M>! AJ/7L! J! 

89?:7=!>X!/75.8>;>;[7A!68!D>ML!495.!J4!26=7;744! 

4784>=! 87/2>=]4! (WEF4*L! :J=5>A74L! 68/7;;684L! 5;>9A! 

5>?09/68=]!A7078A688!/.7!4/J8AJ=A6`J/6>83! 

Y7078A688! [J=6>94! /75.8>;>=! /.7! 

6?0;7?78/J/6>8L! /.7! A7X686/6>8! >X! /.7! D>M! [J=6743! V>27[7=L! 

/.7!X98AJ?78/J;!>X!D>M!6?0;674!/.J/!>:@75/4!68!J8!D>M!5J8!:7! 

6A78/6X67A! 986c97;S! 68! /.7! [6=/9J;! =70=7478/J/6>843! W6/.68! J8! 

D>ML! J;;! /.68! 7\5.J85744!AJ/J!J55>=A68!0=7A7X687A!45.7?743! 

GA%H9'4,F,4%3'45#*0*-,'$%"#1%I$"-'% 

O/!/.7!5>=7!>X!/.7!6A7J!>X!/.7!D8/7=87/!>X!M.68/6>8!/.J/!7[7=SAJS! ^/.68A4! 26;;! :7! 

7c96007A! 26/.! /=J5]688! 64! X9;;S! J5/9J;6`7AL! ^/.68! 5>8/J68! ?>=7! 

4>0.64/65J/7A!0=>574468=]68:@75/4!/>!98A7=4/J8A!/.76=!78[6=>8?78/4L! 

68/7=J5/!26/.!07>0;7!J8A!:75>?7!4?J=/3!b6]7!J8S!68X>=?J/6>8! 

4S4/7?L! /.7! D>M! 26;;! =7;S! >8! J! 5>?:68J/6>8! >X! .J=A2J=7L! 

4>X/2J=7! J8A! J=5.6/75/9=74L! J8A! /.=>9X/2J=7! ?>=7! J8A! 

?>=7!O=/6X656J;!D8/7;;68>?>94! J8A! 47=[7! J4! 68/7;;68,?J]68



! D8/7=8J/6>8J;!e>9=8J;!>X!a9468744!J8A!K5>8>?654!N747J=5.!"#$%&!'()*+!$$),$"-! $$C! 

! 

E784>=!87/2>=]4!J=7!A7[6574!/.J/!?>86/>=!5.J=J5/7=64/654!>X! 

/.7! 78[6=>8?78/! >=! >/.7=! >:@75/4! 495.! J4! /7?07=J/9=7L! 

.9?6A6/SL! ?>[7?78/L! J8A! c9J8/6/S3! W.78! ?9;/60;7! 4784>=4! 

J=7! 947A! />! J4! J! 

26=7;744! 4784>=! 87/2>=]! (WEF*3! W6=7;744! 4784>=! 87/2>=]4! 

5>8/J68! /.7! 4784>=4! /.7?47;[74! J8A! ?JS! J;4>! 5>8/J68! 

;;75/!AJ/J!X=>?!/.7!4784>=4!J8A!0J44!6/!>8!/>!J! 

47=[7=3!! 

W.6;7! 4784>=4! ^47847_! /.7! 4/J/7! >X! J8! 78[6=>8?78/! >=! 

>:@75/L!J5/9J/>=4!07=X>=?!J5/6>84!/>!JXX75/!/.7!78[6=>8?78/!>=! 

>:@75/!68!4>?7!2JS3!O5/9J/>=4!5J8!JXX75/!/.7!78[6=>8?78/!:S! 

7?6//6898AL! ;6! 2J[74! >=! 7[78! 4?7;;43! M.747! 

5J0J:6;6/674! J=7! >87! 2JS! /.J/! D>M! >:@75/4! 5J8! 5>??9865J/7! 

26/.! 07>0;73! O5/9J/>=4! J=7! X=7c978/;S! 947A! 68! 5>?:68J/6>8! 

26/.! 4784>=4! />! 0=>A957! 4784>=,J5/9J/>=! 87/2>=]43! I87! 

7\J?0;7!>X!/.7!947!>X!J5/9J/>=4!68!495.!J!87/2>=]!2>9;A!:7! 

/.7!947!>X!J!4784>=!/>!A7/75/!/.7!0=747857!>X!5J=:>8!?>8>\6A7! 

68!J!=>>?!J8A!/.7!947!>X!J8!J5/9J/>=!/>!0=>A957!J!;>9A!8>647! 

J;7=/680;7!/>!/.7!A7/75/6>8!>X!/.7!.J=?X9;!?:68J/6>8! >X! 4784>=4! J8A! J5/9J/>=4! 5J8! 78J:;7! >:@75/4! />! 

46?9;/J87>94;S! :7! J2J=7! >X! /.76=! 78[6=>8?78/! J8A! 68/7=J5/! 

26/.!07>0;7L!:>/.!J;4!>X!/.7!D>M3! 

D8!0=J5/65J;!68A94/=6J;!94JM!0=7478/;S!7?:=J574! 

NdDY,78J:;7A! 6A78/6X65J/6>8! J8A! /=J5]68;>[6A7! 49XX65678/! =7J;,/6?7! 

68X>=?J/6>8! J:>9/! /.68ML! 2.65.! J=7! [7=S! 947X9;! />! 

?J89XJ5/9=7=4L!A64/=6:9/>=4L!J8A!=7/J6;7=43!d>=!7\J?0;7L!NdDY! 

J00;65J/6>8! 68! 4900;S! 5.J68! ?J8J[7! 

68[78/>=S! ?J8J?7! 6A78/6X67A! JA[J8/J=! 5>4/L! 46?0;6X67A! :9468744! 0=>574474L! J8A! 

6?0=>[7A!7XX656785S3! 

d6[7!S7J=4!JL!6/!2J4!=70>=/7A!/.J/!-!07=578/!>X!Kj!:J47A! 

5>?0J8674! J=7! 9468![$#]*! 

:9/! 26/.! J! 4/=>82/.! 0>/78/6J;3! D8! NdDY,:J47A! 

J00;65J/6>84L! )'! 07=578/! >X! X6=?4! ! X>=! J55744! 5>8/=>;L! "k! 

07=578/! X>=! 4900;S! 5.J68L! ")! 07=578/! X>=! X=772JS! />;;4L! "B! 

07=578/! X>=! 4759=6/S! 5>8/=>;L! "$! 07=578/! X>=! 0=>A95/! 5>8/=>;L! 

J8A!$)!07=578/!X>=!J447/!?J8J;[7! [7=S! A6[7=46X67A! 40756X65J/6>84! 

68!/7=?4!>X!5>??9865J/6>8L!5>?09/J/6>8L! ?7?>=SL!J8A!AJ/J! 

4/>=J=! /=J84?6446>8! 5J0J56/6743! O8! D>M! 

J00;65J/6>8! 5>8464/4! >X! ?J8S! /S074! >X! A7[65743! O;;! /S074! >X! 

.J=A2J=7! A7[6574! 4.>9;A! :7! 27;;! >=9=]! J8A! :7! J557446:;7! [6J! J[J6;J:;7! 5>??9865J/6>83! 

MS065J;;SL! A7[6574! 5J8! :7! >=??9865J/6>8!09=0>47!>[7=!/.7!D8/7=87/3! 

D>M! 5J8! :7! J8! J94! 87/2>=]4L! 

495.! J4! WEF4L! 26=7;744! ?74.! 87/2>=]4L! ?>:6;7! 87/2>=]4L! 

J8A! W6=7;744! b>5J;;S! O=7J! F7/2>=]! (WbOF*3! M.747! 

87/2>=]4!.7;0!/.7!/.688,?J]68?09/J/6>8L! J8A! AJ/J! 7\5.J88L! /.7! =7;6J:;7! 5>??9865J/6>8! :7/2778! 8!26/.!=74075/! 

/>! D>M3! M.7! X! =98868?0;65J/7! 

>0/6?6`J/6>8! J;=6/.?! ;>5J;;S! :S! 7\0;>6/68=]! 

]8>2;7A?09/J/6>8J;! 5>?0;7\6/S! 64! 4.6X/7A! X=>?! 

/.68!/.7!0/6?J;!=>9/7!J8A!0J=J?7/7=! 

[J;974!X>=!/.7!:/J687A3!M.64!64!X7J46:;7!46857! 

/.7!46`7!>X!/.7!?J68!64!68!/.7!>=A7=!>X!J!X72!>X!/784! 

68!5>?0J=64>8!26/.!/.7!46`74!>X!/.68??9865J/6>8! 

=7c96=7?78/4! [J=S! X=>?! >87! A7[657! /S07! />! J8>/.7=3! M.7! 

/.68M! 5J8! .J[7! [7=S! A6XX7=78/! 5J0J:6;6/674! X>=! 

5>?09/J/6>8L! ?7?>=SL! 0>27=L! >=! 5>??9865J/6>83! d>=! 

684/J857L! J! 57;;9;J=! 0.>87! >=! J! /J:;7/! .J4! ?95.! :7//7=! 

5>??9865J/6>8!J8A!5>?09/J/6>8!5J0J:6;6/674!/.J8!J!40756J;, 

09=0>47! 7;75/=>865! 0=>A95/! 495.! J4! J! .7J=/! =J/7! ?>86/>=! 

2J/5.3!E6?6;J=;SL!/.68X! m9J;6/S! >X! E7=[657! (m>E*L! 68! 0J=/659;J=L! 68! /.7! J4075/4! >X! 

A7;JSL! 787=849?0/6>8L! J8A! =7;6J:6;6/S3! d>=! 7\J?0;7L! 

?686?6`68=! 5>??9865J/6>815>?09/J/6>8! 

09=0>474! 64! J! ?J@>=! 5>84/=J68/! X>=! /.7! :J//7=S,0>27=7A! 

A7[6574! 26/.>9/! 7XX65678/! 787=/! 5=6/65J;! X>=! /.7! A7[6574! 26/.! 0>27=! 

4900;S!5>8875/6>83! 

JA%D('"1+5%*F%8990,4"+,*#%8('"$% 

D>M! 64! J8! 68/7=5>8875/7A! 87/2>=]! >X! 4?J=/! >:@75/4! J8A! 

A7[65743! I:@75/4! 68! /.7! 2>=;A! >X! /.7! D8/7=87/! >X! M.684/;S! =7X7==7A! />! J4! 4?J=/! >:@75/4L! J=7! J9



$$k! VJ84!W7=87=!R>//68X!M.688>?65!J8A!D8A94/=6J;!Y6?7846>84! ! 

! 

/.>47!0=>?0/7A!:S!/.7!87/2>=]68X!7[7=SAJS!>:@75/4h!(DMj! 

[']*3! 

M.7! D>M! 64! J! 5>8[7=X! [J=6>94! [646>84L! 7[>;[688! /.7! 4/J]7.>;A7=4_! X>594! >X! A7[7;>0?78/3! 

O/`>=6!7/!J;3![%]!6A78/6i74!/.=77!5;J4474!>X![646>84!26/.68!/.7! 

=747J=5.! 5>??986/S! >X! /.7! D>M+! M.68=678/7AL! D8/7=87/, 

>=678/7A! J8A! E7?J8/65,>=678/7A3! M.7! M.68=678/7A! [646>8! 

X>59474!>8!/.7!M.68X!M.68??9865J/6>84!(FdP*L!W6=7;744!E784>=! 

J8A! O5/9J/>=! F7/2>=]4! (WEOF*L! ?65=>0=>5744>=4L! 

7?:7AA7A! 4S4/7?4L! -Y! J8A! :J=5>A74! />=678/7A![646>8!64!5>857=87A!26/.!/.7!0=>/>5>;4! 

X>=!5>??9865J/6>8!87/2>=]43!E>?7!0=>/>5>;4!J=7!/.7!D8/7=87/! 

G=>/>5>;L! DG['! J8A! /.7! W7:! >X! M.68



! D8/7=8J/6>8J;!e>9=8J;!>X!a9468744!J8A!K5>8>?654!N747J=5.!"#$%&!'()*+!$$),$"-! $"#! 

! 

78>9!4900>=/!=>:94/!785=S0/6>83!M>!78J:;7!785=S0/6>8!>8! 

/.7!D>ML!J;=6/.?4!877A!/>!:7!?JA7!?>=7!7XX65678/!J8A!;744! 

787=849?688! 45.7?74! 

J=7!877A7A!(pJ8!J8A!W78![""]*3!! 

D8! JAA6/6>8! />! 785=S0/6>8L! 6A78/6/S! ?J8J=/J8/! 5>?0>878/! >X! J8S! 4759=6/S! ?>A7;! J8A! 986c97! 

6A78/6X67=4!J=7!74478/6J;!/>!D>M!A7[65743!M.747!6A78/6X67=4!?JS! 

:7! 947A! />! 74/J:;64.! 07=4>8J;! 6A78/6/674! J/! X68J856J;! 

684/6/9/6>84L!6A78/6XS!6;;7/.7=!X985/6>843!! 

M.94L!7849=68:@75/4!J=7!2.>!/.7S!4JS!/.7S!J=7! 

64!74478/6J;!/>!D>M!4955744!(QJ.J;;7!7/!J;3!["-]&!N>?J8!7/!J;3! 

["B]*3! 

G=6[J5S!! 

O4!?>=7!J8A!?>=7!>:@75/4!:75>?7!/=J57J:;7!/.=>9ML! 

/.=7J/4!/>!07=4>8J;!0=6[J5S!:75>?7!?>=7!47=6>943!D8!JAA6/6>8! 

/>! 4759=68! ?J]7! 49=7! /.J/! 6/! A>748_/! XJ;;! 68/>! /.7! 

2=>8X!AJ/J!>287=4.60!877A!/>!:7!JAA=7447A! 

68!>=A7=!/>!7849=7!/.J/!947=4!X77;!5>?X>=/J:;7!0J=/6560J/68M3!! 

M.94L!/.7!>287=4.60!>X!AJ/J!5>;;75/!7A!X=>?!4?J=/!>:@75/4! 

?94/!:7!5;7J=;S!74/J:;64.7A3!M.7!AJ/J!>287=!?94/!:7!J449=7A! 

/.J/! /.7! AJ/J! 26;;! 8>/! :7! 947A! 26/.>9/! .641.7=! 5>8478/L! 

0J=/659;J=;S! 2.78! /.7! AJ/J! 26;;! :7! 4.J=7A3! G=6[J5S! 0>;65674! 

5J8!:7!>87!J00=>J5.!/>!7849=68X!68X>=?J/6>83! 

E?J=/! >:@75/4! J8A! =7JA68M! 5J8! 7J5.! :7! 

7c96007A!26/.!0=6[J5S!0>;656743!W.78!/.7!>:@75/!J8A!=7JA7=! 

5>?7!68/>!5>8/J5/L!/.7S!5J8!7J5.!5.75]!/.7!>/.7=!_4!0=6[J5S! 

0>;65S!X>=!5>?0J/6:6;6/S!:7X>=7!5>??9865J/68?J8!7/!J;3! 

["B]*3!! 

LA%.4*#*/,4%D'#'F,+$% 

D8!J!49=[7S!>8!D>M!J8A!6/4!^75>8>?65!787=



$"$! VJ84!W7=87=!R>//68X!M.688>?65!J8A!D8A94/=6J;!Y6?7846>84! ! 

! 

5>9;A!27;;!.J[7!0>46/6[7!75>8>?65!6?0J5/4!:9/!5>9;A!27;;!:7! 

87X!4759=6/S!:=7J5.74L!5S:7=!06=J5S!J8A! 

;>44! >X! 0=6[J5S! 0>/78/6J;;S! 68A95688>?65! 

AJ?J9;A! :7! 7[78! >9/! >X! 0=>0>=/6>8! 26/.! J! 

986[7=4J;! 45J;J:;7! J8A! 45J;68X! D>M3! O4! 68! /.7! 

75>8>?654!/=7J/?78/!>X!J556A78/!;J2!/.64!=7c96=74!J44744?78/! 

>X!J!A7;65J/7!:J;J857!>X!/=JA7>XX4!J8A!5>?0784J/6>83!! 

E9??J=6`68M! 0=>?6474! ?9;/60;7! 

75>8>?65!:787X6/4!495.!J4!! 

($*! ;>27=688! 5>4/4! X>=! [J=6>94! 68A94/=674! J8A! 

/.7! 78/6=7! 75>8>?SL! 6?0=>[68X! 0=>A95/4L! 47=[6574! 

X>=!5>849?7=4L!J8A!c9J;6/S!>X!;6X7!68!78[6=>8?78/J;!0=>/75/6>8! 

J8A!787=[646>8L! 

("*!78J:;68=?J857!68!J!:=>JA! 

47/! >X! [J;97! 8>?65! 2/.! J8A! 27;;! :76899/!5=7J/68:4!68!4>?7!68A94/=6743! 

MA%N;+;('%O,('4+,*#$% 

E6857! /.7! D>M! .J4! 8>/! S7/! :778! X9;;S! =7J;6`7A! s! 

5>8570/9J;;S!J8A!68!0=J5/65J;!6?0;7?78/J/6>84L!6/!?656>94! />! X>=75J4/! /.7! X9/9=7! A6=75/6>84L! /.7! 7\/78/! >X! 

68A94/=6J;! :=>JA7868X! D>M3! p7/L! /.7! 

?J89XJ5/9=68=! />8! /.7! W>=;A! K5>8>?65! d>=9?_4! (WKd*! "#$'! O889J;! 

Q77/684L! J;4>! 27! .J[7! .JA! J! 5>?0=7.7846[7! 

4.>25J47! J/! /.7! PKaDM! J8A! /.7! VJ88>[7=! D8A94/=6J;! dJ6=! 

("#$'L! "#$%*! 26/.! 5>?0J8674! 495.! J4! E67?784L! OaaL! D8/7;L! 

DaQL! V9J276L! tMKL! a>45.L! l9]JL! dJ895L! lJ2J4J]6! 68! 

0J=/87=4.60! 26/.! 4>X/2J=7! X6=?4! 4.>268@75/4L! J8A! 

7\0>468[7=! $##! 5J474! >X! =7J;,2>=;A! 6?0;7?78/J/6>84! >X! 

D8A94/=S!B3#3!d9/9=7![646>84!>X!/.7!D>M!26;;!JXX75/!6/4!59==78/! 

A7[7;>0?78/!J8A!?94/!/.7=7X>=7!:7!5>846A7=7A3!! 

I87! X9/9=7! [646>8! X>=! /.7! D>M! 64! /.7! W7:! >X! M.68X! M.680>474! /.7! 947! >X! 27:! 4/J8AJ=A4! />! X9;;S! 

68/7:@75/4! 68/>! /.7! W>=;A! W6A7! W7:! (WWW*3! 

j468;>=!A7[7;>07=4!/>! 

:96;A! J00;65J/6>84! 9468:@75/4! J8A! 7\64/68/>5>;4! 5J8! ?>=7! 7J46;S! 78J:;7! /.7! 68/7=>07=J:6;6/S! J8A! 

5>??9865J/6>8!>X!A6XX7=78/!A7[65743!O!?J4.90!64!J!W7:!"3#! 

5>8570/! 2.7=7! J8! J00;65J/6>8! 9474! AJ/J! J8A! X985/6>8J;6/S! 

X=>?!J![J=67/S!>X!27:!=74>9=5743!E>?7!=747J=5.7=4!0=>0>468X! M.68A7;! 49X!J00;65J/6>84!(R968J=A!7/!J;3![-"]*3! 

W6/.!/.7!X!/.7!W7:L!/.7!W>M!J8A!D>M!J=7!>8! 

/.7! 7AX! 7\07=678568AJSL! 27! J=7! >87! 

4/70! 5;>47=! />! /.64! [646>8! A97! />! ;J/74/! JA[J8574! 68! 27:! 

47=[6574L! 6A78/6i5J/6>8! /75.8>;>8[7==]4L! 2.65.! ?J]7! 5>??9865J/6>8! 5J0J:6;6/674! 

J8A! 0=>57446827=! J[J6;J:;7! 68! 685=7J468;[68! /.7! 4>, 

5J;;7A! ^W7:! >X! M.68M*L! J8! 78[6=>8?78/! 2.7=7! 

7[7=SAJS!A7[6574!495.!J4!/=JXi5!;6A6/674!J=7!=75>8/=>;;J:;7! [6J! /.7! D8/7=87/3! P7=/J68;SL! 4/J=/68?! J8! D8/7=87/! >X! 87J=;S! >87! /.>94J8A! ?6;;6>8! 5>?09/7=4L! 

/.7!W7:!8>2!/9=84!/>!:7!J8!D8/7=87/!>X!87J=;S!$##!:6;;6>8!>X! 

/.68=!A7[6574!0=74J?!J8!D>M!/>!J!W>M3! 

M.94L!68!J!W>M!0;J/X>=?!/.7!2>=];>JA!64!09/!J/!/.7!7\/=7?7! 

;7[7;!J8A!45J;J:6;6/S!64!J!5>?09;4>=S!=7c96=7?78/3!! 

M.64! A>74! 8>/! =7X7=! />! J8S! /75.8>;>=! J8S! 87/2>=]! 

4/=95/9=7L!:9/! >8;S! />! /.7! 6A7J! >X! 68/7=5>8875/688875/68?09/7=4! 26/.! /.7! D8/7=87/3! j47! >X! 

/.7! W7:! J4! /.7! 0;J/X>=?! .>4/684688875/7A! 

A7[6574! 5J8! :7! 7\0;J687A! :S! ?9;/60;7! /75.8>;>8>?65!:787i/4L!J!X72!>X!2.65.!685;9A7!A70;>S?78/L!.6X! 4/J8AJ=A6`7A! 

5>??9865J/6>8! 0=>/>5>;4! J8A! /.7! 75>4S4/7?! 5=7J/7A! /.J8]4! 

/>! W7:! "3#! 0J=JA6



! D8/7=8J/6>8J;!e>9=8J;!>X!a9468744!J8A!K5>8>?654!N747J=5.!"#$%&!'()*+!$$),$"-! $""! 

! 

899'#1,R2%H/"(+%!*/'%HE'0'+*#%O'$,-#% 

S%8#%B00;$+("+,7'%.R"/90'%% 

O! .>?7! :75>?74! 4?J=/! 6X! /.7! J55744! J8A! 5>8/=>;! >X! /.7! 

A7[6574!J8A!9/6;6/674!68!/.7!.>?7!4/J=/!/J]688!(27! 

J=7! 8>/! /J;]689/! O=/6X656J;! D8/7;;607=! 47/! >X! 

684/=95/6>84!64!!/.7?L!J8A!>8;S!6X!QJ5.687!/>!?J5.687! 

5>??9865J/6>8!64!.J007868!?J5.687!;7J=868446:;7! 6X! 4784>=! 87/2>=]4! 5>8/=>;! /.747! A7[65743! 

M75.8>;>/>5>;4L! 87/2>=]4! ;6]7! NdDYL! FdPL! t6>/.! J8A! [J=6>94! 4784>=4! 

(/7?07=J/9=7L! GDNL! 4?>]7L! J8A! .9?6A6/S*! 5J8! ?J]7! J! 

A75646>8,?J]682! 6;;94/=J/6[7;S! 

7\0;J68!/.7!5>??9865J/6>8!0=>57443! 

E?J=/!.>?7!87/2>=]3! 

W7!5J8!477!4?J;;!56=596/4!(4784>=4!J8A!4?J=/!56=596/4*!X6\7A! 

>8!J;;!/.7!A7[6574!J8A!J00;6J8574L!/.747!4784>=4!J8A!A75646>8, 

?J]68??9865J/7! 26/.! J55744! ! J! ;>5J;! =>9/7=! /.J/! 0=>[6A74! 

D8/7=87/! 5>8875/6>8! /.=>9! 

/.7!D8/7=87/3! 

E?J=/!.>?74!.J[7!/.7!X>;;>2688!?J]6827=!KXX65678/!O00;6J8574! 

V6!9/09/! 

J4! /.7! =7849?7! ?686?J;! 0>27=3! 

O00;6J8574! ;6]7! .627=! 4J[68! ?J]7! 4?J=/! 9/6;6`J/6>8! >X! 787=9/!5>?0=>?64688!5>?X>=/!;7[7;43!! 

-3!N7?>/7!O55744! 

M.7! ?J68! 09=0>47! >X! 4?J=/! .>?74! 64! />! J5.67[7! 5>8/=>;! 

>[7=!J00;6J8574!=7?>/7;S3! 

Q>4/!>X!/.7!/6?74!27!477L!07>0;7!X>=!426/5.!>XX!/.7! 

J00;6J8574!2.6;7!=94.68!2>=]!A97!/>!2.65.!/.7S!.J[7!/>! 

0JS! .628! >X! 

J00;6J857_4! A97! />! 7\57446[7! 94J! J! 986[7=4J;! 5>8/=>;J=7! J557446:;7! X=>?! J8S2.7=7! 

=7?>/7;S!26/.!>9=!4?J=/0.>874!J8A!5>?09/7=4L!/.7!947=!5J8! 

426/5.!/.7?!J4!.714.7!2J8/4!J8A!?>86/>=!/.76=!.>?73!! 

B3!E759=6/S!! 

! 

W7!:96;A!>9=!.>?7!/.7!2JS!27!2J8/L!J8A!27!J;4>!2J8/!6/! 

/>!4/JS!/.7!4J?7!2JSL!4>!4JX7/S!64!J;4>!/.7!0=6?J=S!5>857=8!68! 

J! 4?J=/! .>?73! d>=! 4JX7/S! X=>?! X6=7! 2J/7=! J8A! .60=6J/7! 4784>=4! 4.>9;A! :7! 684/J;;7A! />! J;7=/! 

94!68!5J474!>X!X6=7!:=7J]>9/!>=!2J/7=!;7J]J?7! 

5J8!J;4>!:7![9;87=J:;7!/>!=>::7=S3!M>!0=7[78/!>9=!.>947!X=>?! 

=A!/.7!X>>/J9A3! O;4>! />! =?7A! >8! J8S! 98J9/.>=6`7A! 

78/=674! J8A! 4940656>94! :7.J[6>=L! 68X=J=7A! J;J=?! 4S4/7?! (6/! 

5>9;A! :7! J! GDN! 4784>=! >=! :9=X! 

/.7! ;J/74/! /75.8>;>0;7*L! 4>! /.7! >07=J/6>8! >X! 

/.747!J00;6J8574! 4.>9;A!:7!7J46;S!98A7=4/J8AJ:;7!X>=!07>0;7! 

>X!J;;!J9:;74.>>/!>8!6/4!47;X3! M.7!68/7=XJ57!>X!5>8/=>;!4.>9;A! 

J;4>!:7!J4!46?0;7!J8A!X=678A;S!J4!0>446:;73! 

d>=! 6;;94/=J/6[7! 09=0>474L! 27! .J[7! A746A7;! />! 

A7?>84/=J/7! D8/7=87/! >X! M.68=! =7?>/7! ?>86/>=688/=>;;68X! J00;6J8574! =7?>/7;S! /.=>9X! /.64! ?>A7;! 64! />! 

0=>[6A7!8>80.S465J;!5>8/=>;!>[7=!J00;6J8574!J8A!/>!8! /.7! /7?07=J/9=7! J8A! .9?6A6/S! ;7[7;4! J/! .>?7! 

/.=>9?!J8S2.7=7!68!/.7!2>=;A3!M.64! 

]68A! >X! J! ?>A7;! .7;04! /.7! 0.S465J;;S! 5.J;;78! 5>8/=>;! J00;6J8574! J4! 6/! ?6=! /.7?! />! ?>[7! />! /.7! 426/5.! 7[7=S! /6?7! /.7S! 2J8/! />! 

5>8/=>;! 6/3! Q>=7>[7=L! /.7! 0=>0>47A! ?>A7;! 64! 5>4/! 7XX75/6[7! 

J8A!0>27=!7XX65678/3!! 

M.7! 5>8570/! >X! D8/7=87/! >X! M.68[7A! :S! 5=7J/6807=! 68/7=XJ57! J8A! A746??9865J/7! 26/.! /.7! 

D8/7=87/3!O[J6;J:;7!>A4!J8A!c9J8/6/S!5J8!:7!]8>28!68!J! 

468:@75/4! J8A! >A4! J=7! 98A! 

>:@75/4!5J8!:7!/=J5]7A!94688570/!:S!/=J5]680=6J/7! 4784>=4! X6\7A! />! /.7?3! O! ?>:6;7! 

J00;65J/6>8!5J8!:7!5=7J/7A!/>!5>8/=>;!/.747!J00;6J8574!X=>?! 

>9=! 4?J=/! 0.>8743! aS! J==J886/>=7A!X=>?! 

J8S2.7=7! :S! ;>;>94!J00;65J/6>84!X>=! 

7J4S!J557446:6;6/S3! 

6'F'('#4'$% 

u$v! 

u"v! 

u-v! 

uBv! 

K[J84L! G3! P3! J8A! O8898`6J/JL! Q3! ("#$"*3! D8A94/=6J;! D8/7=87/L! 

R787=J;!K;75/=65!(RK*!N70>=/L!F72!p>=]3! 

O;0JSA68L! K3! ("#$'*3! QJ5.687! b7J=868



$"-! VJ84!W7=87=!R>//68X!M.688>?65!J8A!D8A94/=6J;!Y6?7846>84! ! 

! 

u)v! 

W.6/?>=7L! O3L! O=6L!b3L!D7=JL!O3!J8A!Q>=J:6/>L!R3!("#$#*3!hM.7!D8/7=87/!>X! 

M.68?09/7=!F7/2>=]4L!)B($)*L!"%C%s"C#)3! 

d68?JL!D3L!E6?07=;L!K3!J8A!V785.L!R3!("##k*3!gO!@>68/!=>JA?J0! 

X>=! E7?J8/65! /75.8>;>X! M.68;>=J8A6L! Y3L! E65J=6L! E3L! Y7! G7;;7X! /.688L! J00;65J/6>84! J8A! =747J=5.! 

5.J;;785!F7/2>=]4L!$#!(%*L!$Bk%s$)$'3! 

u$Bv! b6L!b3L!J8A!b69L!e3!("#$"*3!gO8!7XX65678/!J8A!X;7\6:;7!27:! 

47=[6574,:J47A! ?9;/6A64560;68J=S! A7460/6?64J/6>8! 

X=J?72>=]! X>=! 5>?0;7\! 788!ES4/7?4L!'(-*L!-B)s-%$3! 

u$)v! E5.JXX7=4L!V3L!l>?868>4L!F3L!GJ;;>/L!Q3L!M=>9447L!a3L!F6;44>8L! 

Q3L! J8A! I;6[76=JL! O3! ("#$$*3h! E?J=/! 56/674! J8A! /.7! X9/9=7! 

68/7=87/+! M>2J=A4! 5>>07=J/6>8! X=J?72>=]4! X>=! >078! 

688>[J/6>8hL! 68! Y>?68X! M.68=! J?:678/! 

J4464/7A! ;6[68577A68X! /.7! E7[78/.! D8/7=8J/6>8J;! 

P>8X7=7857! >8! D8X>=?J/6>8! M75.8>;>84! 

(DMFR*3! 

u$Cv! d;n84/=95/68??986,! 5J/6>84! 

68!5>?09/7=!J8A!68X>=?J/6>8!4567857!(Z>;3!-"*L!$'s"'3!a7=;68+! 

E0=68874L! e3! Y3! ("#$"*3! g7, 

a9468744! 4S4/7?4! 68/78+! J! 4S4/7?4! 07=4075/6[7hL! 

D8X>=?J/6>8!M75.8>;>84! =747J=5.! (IN*! 68! 47=[657! 

68A94/=674+! J! 5>?0=7.7846[7! =7[672hL! ES4/7?4! N747J=5.! J8A! 

a7.J[6>=J;!E567857L!-#!(-*L!-##s-)-3!! 

u""v! pJ8L!M3!J8A!W78L!m3!("#$$*3!ga96;A68X!M.68/>5>;!:J47A!>8!68X>=?J/6>8! 

/75.8>;,! >?09/3! E563! D8/7;;3! ES4/3! K8[6=>83! $#BL! 

$B-s$Bk3!! 

u"-v! QJ.J;;7L! G3L! aJ:J=L! E3L! G=J4JAL! F3! N3! J8A! G=J4JAL! N3! 

("#$#*3hDA78/6/S! ?J8J=]! />2J=A4! D8/7=87/! >X! 

M.68M*+!N>JA?J0!J8A!]7S!5.J;;78?J8L! N3L! FJ@7=JL! G3! J8A! b>07`L! e3! ("#$$*3hE759=68X!M.68?09/7=4L!BB(k*L!)$s)C3! 

u")v! d;7645.L! K3! ("#$-*3! gW.J/! 64! /.7! D8/7=87/! >X! /.68?!.//0+!1!1!2223!6?3!K/.`3!5.!1! 

7A95J/6>8!1!VE$#!1!OjMIDYbOaE,WG,aDtOGG,)-30AX3!! 

u"'v! R>//68X! R>[7=8J857L! 

F72!p>=]+!I\X>=A!j86[3!G=7443! 

u"kv! R>//68X! P6/6743! 

M.7! 68/7=87/! >X! K[7=S/.682! 27! ;6[7hL! 68! 

E0756J;!D4497!>X!d>=76X!M.68!/.7!W7:!>X!M.689=57, 

>=678/7A! J=5.6/75/9=7! J8A! :74/! G=J5/6574hL! 68! j5]7;?J88L! Y3L! 

VJ==64>8L! Q3! J8A! Q65.J.7;;74L! d3L! KA43! ("#$#*3! O=5.6/75/68X!M.68



132

2. 

MARKETS 

AND 

COMPETITION

2.1 INNOVATION, DYNAMICS OF COMPETITION AND MARKET DYNAMICS 


!"#$%&'&()*+),-(%.&(()/&(&0"#$)1)2*3456)7*48) 

9-:3%#0;%*.)



!"##$%&'()*+,*-,*./0123,*4%%"56#$"%)*78%69$:;*"8!4='!B4,C8#!:#,A@#A,8)!I@"AB$8#8,!E4:!+B$,8::8'!5;!#"8!JA4D+#4#+98!'+4#+*=!*8! 

+=!,8:$8@#+98!+='A:#,+8:!SI@"8,8,?!.X30?.XX6T)! 

! 

G*,! "+-"U#8@"! +='A:#,+8:! E+#"! 4! -D*54D!



7+%8&9':!$*!;0:&,'::!A!! 

! 

M80:! *$+! *&+5:! &,! 8&48G('%8! &,.0:(+&':! %$,(&,0$0:! &,,$9-(&$,! &:! (8'! ,-5'! $*! (8'! 4-5'! -:! 

%$5K'(&(&$,!&,!&,,$9-(&$,!.+&9':!(8'!$0(%$5':!&,!4+$S(8!1!S'@*-+'!%8-,4':!-,.!:0+9&9-@A! 

! 

V&(8$0(! &(! :(+0%(0+-@! 4+$S(8! &,! 5-+W'(:! S$0@.! ,$(! F'! K$::&F@'! S8'(8'+! &,! K+$.0%(! 

('%8,$@$4&':1!:'+9&%':!$+!$K'+-(&$,-@!K'+*$+5-,%'A!C,,$9-(&$,!F-:'.!*&+5!4+$S(8!S$0@.!F'!-! 

W')!&,.&%-($+!*$+!-!*&+5T:!:0:(-&,-F@'!*0(0+'A! 

! 

7*('+!+':$@9&,4!(8'!&::0'!8$S!-,.!S8)!@-+4'!*&+5:!-+'!-%(&9'@)!',4-4'.!&,!&,,$9-(&9'!-%(&9&()!-,! 

&5K@&%-('.! ,'L(! :('K! %$0@.! F'! 8$S! (8')! %$,(&,0'! ($! F'! &,,$9-(&9'! $+4-,&:-(&$,:A! 2,.'+! (8'! 

K+':05K(&$,!(8-(!@-+4'!*&+5:!F'%-5'!T@-+4'T!F'%-0:'!$*!T4+$S&,4!(8+$048!&,,$9-(&$,T!(8'+'!&:! 

(8'!X0':(&$,!8$S!(8')!W''K!$,!F'&,4!&,,$9-(&9'!*$+!(8'&+!@&*'(&5'A! 

! 

M8+$048!:(0.&':!$*!&,,$9-(&$,!&,!&,.0:(+&-@!$+4-,&Y-(&$,:!S'!@'-+,!(8-(!+$048@)!(S$G(8&+.!$*! 

&,,$9-(&$,:!-+'!$*!(8'!K+$.0%(!&5K+$9'5',(!()K'!S8&@'!$,'G(8&+.!&:!$*!(8'!K+$%'::!&,,$9-(&$,! 

()K'!=%$:(!+'.0%&,4!$+!K+$.0%(&9&()!',8-,%&,4>A!V'!5-W'!0:'!$*!(8&:!$F:'+9-(&$,!F)!-::&4,&,4! 

(8'!*&+:(!W&,.!($!K+&9-('!$+4-,&Y-(&$,:!=&,.0:(+&-@!*&+5:>!S8&@'!(8'!:'%$,.!($!-!K0F@&%!:'+9&%'! 

%',('+'.!$+4-,&Y-(&$,1!.':&4,'.!-,.!+'40@-('.!-:!:0%8A! 

! 

!*+,-././*012+345-.26.47!.74-23082/00*93./*02 

V'! *$%0:! 8'+'! $,! :$5'! W')! +':0@(:! *+$5! F'8-9&$+-@! '%$,$5&%! 5$.'@:! &,! &,.0:(+&-@! 

$+4-,&Y-(&$,=! CZ>! (8-(! @'-9'! *$$(K+&,(:! *$+! 'LK@$+&,4! (8'! &,,$9-(&$,! .+&9'! -,.! %-K-F&@&()! $*! 

*&+5:!&,!K-+(&%0@-+!&,.0:(+)!:'((&,4:A!! 

! 

C,!%8-+-%('+&Y&,4!-,.!K$+(+-)&,4!(8'!&,('+,-@!'%$,$5)!$*!@-+4'!*&+5:!



!"##$%&'()*+,*-,*./0123,*4%%"56#$"%)*78%69$:;*"!#"6!:*7'+#+*79!-*A6,7+7-!896>!47'!9*;6!:"47:6)! 

! 

B"6! :?499+:4?! $*+7#! *! 

*76! C*8?'! *596,A6! ;*,6! +77*A4#+*7! +7! #"*96! +7'89#,+69! #"4#! 4,6! :"4,4:#6,+O6'! 5@! 4! "+-"6,! 

'6-,66!*!4?#"*8-"!7*!:4894?+#@!9"*8?'!56!+;$8#6'!#*!#"+9!,6?4#+*79"+$)! 

! 

N*;6!;*'6?9!:*79+'6,!#"6!,6?4#+*79"+$!56#C667!$,*'8:#!;4,=6#!9#,8:#8,6!D+)6)!:*7:67#,4#+*7E! 

47'! +77*A4#+*7)! ! S7! #"6! :*7#6(#! *! #"6! $,*545+?+#@! *!C"67!+;+#4#+*7!+9! 

7*#!$*99+5?6>!47!+7:,6496!+7!#"6!78;56,!*



7+%8&9':!$*!;0:&,'::!! .$5&,-(':! (8'! *&+5:R! .'%&:&$,A! Y8',! :0%%'::! &:! :($%8-:(&%1! 8$N'9'+1! (8'! (8+'-(! 

*+$5!(8'!+&9-@!&,,$9-(&,4!&:!@'::!-%0('A!C,!(8'!%-:'!$*!.+-:(&%!&,,$9-(&$,:1!(8'!%$5O'(&(&9'!(8+'-(! 

&:!(8'!:-5'!*$+!F$(8!*&+5:!-,.!&(!&:!(8'!O+$*&(!&,%',(&9'!=N8&%8!&:!@-+4'+!*$+!%8-@@',4'+:>1!,'Z(!($! 

.&9'+:'!,$,G5$,'(-+)!&,%',(&9':1!(8-(!.'('+5&,':!(8'!@'9'@!$*!



!"##$%&'()*+,*-,*./0123,*4%%"56#$"%)*78%69$:;*"?!4!,:?@9#A!4!D+':!;4,+:#8!*99!#":?:!#":*,+:?!?"4,:!#":!=*BB*7! 

4+B! *



7+%8&9':!$*!;0:&,'::!



!"##$%&'()*+,*-,*./0123,*4%%"56#$"%)*78%69$:;*"! 94,! -,74#7,! 

7=*! #"7! 7:$"48+8! *!4!.P31>!N"4$)!2E)! 

! 

H,4'+#+*!*



7+%8&9':!$*!;0:&,'::!A! 

! 

S8':'! %$,:&.'+-(&$,:! :044':(! (8-(! &(! &:! 0:'*0@! ($! .&:(&,40&:8! F'(L'',! .&**'+',(! '%$,$5&%! 

',9&+$,5',(:1! -%%$+.&,4! ($! (8'! +'@-(&9'! &5N$+(-,%'! $*! %$5N'(&(&$,! &,! &,,$9-(&$,! -,.! 

%$5N'(&(&$,!-(!(8'!N+$.0%(!5-+M'(!@'9'@!(8-(!%-,!F'!'TN'%('.!&,!(8'!5-+M'(A!S8'!@&('+-(0+'!$,! 

,'(L$+M!&,.0:(+&':!&:!N-+(&%0@-+@)!0:'*0@!&,!'TN@-&,&,4!L8)!-!5-+M'(!5-)!'T8&F&(!L&,,'+G(-M':G 

-@@!N+$N'+(&':A! 

! 

4,)5"%/2677,8)02 

U'(L$+M! '**'%(:! 5-)! F'! -,! &5N$+(-,(! *-%($+! &,! .'('+5&,&,4! (8'! %$5N'(&(&9'! ',9&+$,5',(A! 7! 

=N$:&(&9'>! .&+'%(! ,'(L$+M! 'T('+,-@&()! 'T&:(:! L8',! (8'! .'5-,.! *$+! -! :'+9&%'! &,%+'-:':! F)! -,! 

'T(+-! 0,&(A! C,! $+.'+! *$+! (8&:! '**'%(! ($! F'! O.&+'%(P! (8'! +'-:$,! *$+! (8'! &,%+'-:'! &,! .'5-,.! 50:(! 

%$5'!.&+'%(@)!*+$5!(8'!-..&(&$,-@!%$,:05N(&$,!$*!(8'!:'+9&%'!&,!V0':(&$,1!L&(8$0(!,''.!*$+!-! 

:(+-('4&%! +':N$,:'! F)!:0NN@&'+:!&,!(8'!:-5'!$+! +'@-('.! 5-+M'(A! E$+! 'T-5N@'1! (8'! .'5-,.! *$+! 

N0F@&%! ('@'N8$,)! :'+9&%':! =&A'A! (8'! .'5-,.! *$+! :0F:%+&N(&$,:! ($! ,'(L$+M:>! 5-)! &,%+'-:'! -:! 

5$+'!0:'+:!%$,:05'!(8&:!N+$.0%(!='A4A!',('+!&,($!:0F:%+&N(&$,!-4+''5',(:!L&(8!N0F@&%!,'(L$+M! 

$N'+-($+:>!:&5N@)!F'%-0:'!5$+'!0:'+:!-@@$L!-!4+'-('+!,05F'+!$*!N$(',(&-@!%$,,'%(&$,:!($!F'! 

5-.'!$*!(8'!,'(L$+MA! 

! 

;)! %$,(+-:(1! -! =N$:&(&9'>! &,.&+'%(! ,'(L$+M! 'T('+,-@&()! 'T&:(:! L8'+'! (8'! .'5-,.! *$+! -! :'+9&%'! 

:0NN@&'.!&,!$,'!+'@'9-,(!5-+M'(!8-:!-!:&4,&*&%-,(!'**'%(!$,!:0NN@)!$*!-,$(8'+!N+$.0%(!=()N&%-@@)! 

&,! -! .&**'+',(! +'@'9-,(! 5-+M'(>! :0%8! (8-(! (8&:! &,! (0+,! &,%+'-:':! .'5-,.! &,! (8'! *&+:(! +'@'9-,(! 

5-+M'(A! 

! 

7,!'T-5N@'!$*!&,.&+'%(!,'(L$+M!'**'%(:!&:!(8-(!$*!%$5N0('+!:$*(L-+'!*$+!-!N-+(&%0@-+!$N'+-(&,4! 

:):('5!N@-(*$+5A!7,!&,%+'-:'!&,!%$,:05N(&$,!&,!(8'!5-+M'(!*$+!:0NN@)!$*!$N'+-(&,4!:):('5:!&:! 

@&M'@)! ($! F+&,4! F','*&(:! ($! (8'! 5-+M'(:! *$+! :0NN@)! $*! %$5N-(&F@'! :$*(L-+'1! F)! 'TN-,.&,4! (8'! 

5-+M'(!:0%8!(8-(!-!4+'-('+!9-+&'()!$*!:$*(L-+'!&:!$**'+'.!$+!(8'!0,&(!N+&%'!.'%+'-:':!=:&,%'!*&T'.! 

%$:(:!$*!:$*(L-+'!.'9'@$N5',(!%-,!F'!:N+'-.!$9'+!-!4+'-('+!,05F'+!$*!N$(',(&-@!F0)'+:>A!C,! 

(0+,1!(8'!F','*&(:!($!(8'!:$*(L-+'!5-+M'(:!+',.'+!(8'!:-5'!$N'+-(&,4!:):('5!5$+'!9-@0-F@'1! 

-,.! (80:! &,%+'-:'! .'5-,.! *$+! &(A! ;0(! (8'! '**'%(! &:! &,.&+'%(! F'%-0:'! &(! +'@&':! $,! -! :(+-('4&%! 

+':N$,:'!$*!:$*(L-+'!:0NN@&'+:!($!(8'!&,%+'-:'!&,!(8'!%$,:05N(&$,!$*!$N'+-(&,4!:):('5:A! 

! 

9"%&:"+).12.+;2



!"##$%&'()*+,*-,*./0123,*4%%"56#$"%)*78%69$:;*""6,65!47'! 

56##6,! $,*'?:#9! '+9$84:6! *8'6,! *59*86#6! $,*'?:#9! =,*A! #"6! A4,F6#)! O7! A4,F6#9! 9?5R6:#! #*! 

;6,#+:48!'+==6,67#+4#+*7@!4!=+,A!B*,!A*,6!4::?,4#68!#"6!$4:6!*=! 

#6:"7*8*-+:48!4';47:6!A4I!6(+9#9@!*?#9+'6!>"+:"! 

=+,A9!:477*#!9?,;+;6!+7!#"6!A4,F6#C)!N"+9!6:*7*A+:!$,*:699!:47!-?+'6!*7!#"6!74#?,6!*=!A4,F6#! 

'+96! 

+77*;4#+*7!A4*,96V!!N"+9!:477*#!56! 

479>6,6'!4#!#"6!-676,+:!86;68@!-+;67!#"6!:?,,67#!86;68!*=!?7'6,9#47'+7-!*=!#"696!6==6:#9)!W47"+:"! A4



7+%8&9':!$*!;0:&,'::!! ($! -%%'::! 

'R('+,-@! :$0+%':! $*! M,$N@'.4'! ='A4A! :%&',(&*&%! M,$N@'.4'>S! 8-9&,4! -%%'::! ($! :$5'! (-%&(! 

M,$N@'.4'!N8&%8!&:!,$(!(+-,:*'+-E@'!=L'+8-L:!E'%-0:'!'5E$.&'.!&,!$+4-,&O-(&$,-@!+$0(&,':>S! 

-,.!8-9&,4!-%%'::!($!:$5'!:L'%&*&%!%$5L@'5',(-+)!+':$0+%':1!'A4A!:L'%&*&%!&,L0(!*-%($+:A! 

! 

*+,&)-./"0123'2- 

J8'!%-L-E&@&()!$*!-!*&+5!($!&,,$9-('!&,!L-+(&%0@-+!-+'-:!N&@@!E'!-**'%('.!E)!(8'!'R(',(!$*!(-%&(! 

M,$N@'.4'! :0LL$+(&,4! &(:! &,,$9-(&9'! -%(&9&()1! -,.! N8'(8'+! (8'! *&+5! 8-:! -%%'::! ($! (8&:! 

M,$N@'.4'!='A4A!-%T0&+'.!(8+$048!-!L+$%'::!$*!@'-+,&,4FE)F.$&,4>A! 

! 

J-%&(!M,$N@'.4'!%-,,$(!E'!%$.&*&'.1!(+-,:*'++'.!$+!&5&(-('.!E)!'R('+,-@!L-+(&':!E0(!&:!+-(8'+! 

'5E$.&'.!&,!+$0(&,':!-,.!:M&@@:!$*!(8'!*&+5A!J8'!('+5!U(-%&(!M,$N@'.4'K!&:!0:'.!($!',%-L:0@-('! 

(8'! &.'-! $*! M,$N@'.4'! (8-(! &:! *$0,.! ($! E'! 0:'*0@! ='A4A! &,! -,! $+4-,&O-(&$,-@! +$0(&,'! $+! -! :M&@@>! 

N&(8$0(! E'&,4! -+(&%0@-('1! -,.! (8'+'*$+'! N&(8$0(! E'&,4! .&+'%(@)! (+-,:*'+-E@'! ='A4A! :$@.! -:! 

&,*$+5-(&$,>A! 

! 

455678*465-859-:;8A! 

! 

Y'!:8$0@.!-%%'L(!(8-(!(8&:!.&:(&,%(&$,!&:!,$(!-E:$@0('A!/-%8!&,,$9-(&$,1!&,!'9'+)!5-+M'(1!N$0@.! 

4','+-@@)!&,9$@9'!E$(8!'@'5',(:!$*!&,9',(&$,!-,.!-.$L(&$,A!D'N!&,,$9-(&9'!&5L+$9'5',(:!N&@@! 

E'!4',0&,'@)!0,&T0'!($!(8'!&,,$9-(&,4!*&+5!:&,%'!(8')!N&@@!-@N-):!&,9$@9'!-!.'4+''!$*!&5&(-(&$,! 

-,.! -.$L(&$,! *+$5! $E:'+9-(&$,! $*! (8'! N$+@.! $0(:&.'! $*! (8'! *&+5A! 7,)! &,9',(&$,! N&@@! .+-N! ($! 

:$5'!'R(',(!$,!L-:(!&,,$9-(&$,:!-,.!M,$N@'.4'A!P$,9'+:'@)1!'9',!N8'+'!-!*&+5!:''5:!($!E'! 

.&+'%(@)! -.$L(&,4! &,,$9-(&$,:! -::$%&-('.! N&(8! &,9',(&9'! '**$+(! E)! $(8'+! $+4-,&O-(&$,:1! &(! N&@@! 

,''.!($!-.-L(!(8':'!($!&(:!$N,!%8-+-%('+&:(&%:!-,.!%&+%05:(-,%':1!N8&%8!+'T0&+':!-,!'@'5',(!$*! 

&,9',(&$,A!/9',!N8'+'!&,,$9-(&$,!&:!@-+4'@)!.+&9',!E)!&,9',(&9'!&5L+$9'5',(:!&,!(8'!:'+9&%':! 

L+$9&.'.!E)!:0LL@&'+:1!*&+5:!0:&,4!:0%8!:'+9&%':!5-)!,''.!:$5'!%+'-(&9'!'**$+(!($!&,%$+L$+-('! 

(8':'!&,($!(8'&+!L+$.0%(&$,!L+$%'::':!-,.!L+$.0%(:A!J8&:!9&'N!&:!0,.'+@&,'.!E)!%$55',(:!*+$5! 

Z'@:$,! =6[[I1! LAGBB>! *$@@$N&,4! -,! 'R(',:&9'! %+$::F%$0,(+)! '5L&+&%-@! :(0.)! $,! &,,$9-(&$,Q! 

\](8'!E0@M!$*!(8'!'**$+(!&,,$9-(&$,!,''.:!($!E'!.$,'!E)!(8'!*&+5:!(8'5:'@9':A!Y8&@'!(8')!5-)! 

.+-N! $,! $0(:&.'! .'9'@$L5',(:1! :&4,&*&%-,(! &,('+,-@! '**$+(! -,.! :M&@@! &:! ,''.'.! ($! %$5L@'5',(! 

-,.!&5L@'5',(!(8':'A^! 

Z$,'(8'@'::1!N'!%-,!%$,%'&9'!$*!.&:(&,40&:8&,4!(8'!'R(',(!($!N8&%8!(8'!&,,$9-(&$,!-::$%&-('.! 

N&(8!:0LL@)!$*!-!*0(0+'!5-+M'(!&:!-.$L(&9'!+-(8'+!(8-,!&,,$9-(&9'1!'9',!&*!E$(8!'@'5',(:!50:(! 

E'!L+':',(!($!:$5'!.'4+''A! 

! 

! 

!"#$%&'()!"!#$%&'()!*$+!#%&',%'!-,.!/.0%-(&$,1!2,&('.!3&,4.$5! 66! 

145



!"##$%&'()*+,*-,*./0123,*4%%"56#$"%)*78%69$:;*"!4!?@,#";,!'+!



8+%9&:';!$*!(9'!M+&5-+)!5-+K'(?B! 

! 

8! %$55$,A)! %&('.! 'O-5MA'! &;! (9-(! $*! M+&,('+;! -,.! %$5M-(&FA'! &,KG+'*&AA! %-+(+&.4';1! L9'+'! 

.'5-,.! *$+! (9'! A-(('+! &;! .'+&:'.! ',(&+'A)! *+$5! %$,;05M(&$,! $*! (9'! *$+5'+B! D,! (9&;! %-;'1! (9'! 

;%$M'! *$+! %$5M'(&(&$,! &,! &,,$:-(&$,! .&+'%('.! ($! +'*&AA! %-+(+&.4';! L&AA! F'! -**'%('.! F)! L9'(9'+! 

%-+(+&.4'!5-,0*-%(0+'+;!%-,!&,('+-%(!;0%%';;*0AA)!L&(9!(9'!M+&,('+;!5-+K'(B! 

! 

8! A';;! $F:&$0;! 'O-5MA'! &;! 5-+K'(;! %$,%'+,'.! L&(9! -.:'+(&;&,4! >(9'! ;0MMA)! $*! -.:'+(&;&,4! 

;M-%'?B!D,!5$;(!%-;';1!(9'!'O&;(',%'!$*!(9'!-.:'+(&;&,4!5-+K'(!&;!.'M',.',(!$,!%$,;05M(&$,!&,! 

-!+'A-('.!%$,(',(!5-+K'(!>F'!(9-(!(9'!;0MMA)!$*!('A':&;&$,!;'+:&%';1!(9'!;0MMA)!$*!,'L;M-M'+;1! 

5-4-P&,';1!$+!(9'!;0MMA)!$*!%$,(',(!$,!(9'!D,('+,'(?B!8,-A);&;!$*!%$5M'(&(&$,!&,!(9'!&,,$:-(&:'! 

-%(&:&()!+'A-('.!($!',(+)!&,!(9'!-.:'+(&;&,4!5-+K'(;!50;(!M-)!-((',(&$,!($!(9'!,''.!*$+!(9'!*&+5! 

($! ;&50A(-,'$0;A)! $M'+-('! ;0%%';;*0AA)! &,! (9'! %$,(',(! 5-+K'(;B! #$! (9'! +'A-(&$,;9&M! F'(L'',! 

-.:'+(&;&,4!+'A':-,(!5-+K'(;!-,.!%$,(',(!+'A':-,(!5-+K'(;!*-AA;!0,.'+!(9'!%-('4$+)!$*!-.N0,%(! 

5-+K'(;B! 

! 

Q*!M-+(&%0A-+!&5M$+(-,%'!&;!L9'+'!-!%$5M'(&(&:'!M+&5-+)!5-+K'(!&;!-;;$%&-('.!L&(9!-,!-.N0,%(! 

5-+K'(!&,!L9&%9!(9'+'!&;!5$,$M$A)B!R90;!*&+5;!%$5M'('!&,!-!F+$-.!M+&5-+)!5-+K'(!-,.!'-%9! 

-%9&':';!5$,$M$A)!;0MMA)!&,!-!,-++$L!-.N0,%(!5-+K'(!>M+&,('+;!-,.!+'*&AA!%-+(+&.4';?B!!D,!;0%9! 

%-;';!L'!5-)!'OM'%(!*&+5;!($!0;'!-,)!5-+4&,-A!&,%+'-;'!&,!M+$*&(;!+'-A&P'.!&,!(9'!;'%$,.-+)! 

5-+K'(! ($! *0,.! -! .'%+'-;'! &,! M+&%'! &,! (9'! %$5M'(&(&:'! M+&5-+)! 5-+K'(! -,.! (9'+'F)! &5M+$:'! 

(9'&+! +'A-(&:'! M$;&(&$,! -4-&,;(! %$5M'(&($+;B! S$+'$:'+1! L'! L$0A.! 'OM'%(! (9'+'! ($! F'! %9'%K;! 

-4-&,;(!%'+(-&,!()M';!$*!-,(&G%$5M'(&(&:'!F'9-:&$+!&,!(9'!5$,$M$A)!A&,K'.!5-+K'(;!.0'!($!(9'! 

(+-,;*'+!$*!%$5M'(&(&:'!'**'%(;!*+$5!(9'!M+&5-+)!5-+K'(B! 

! 

R90;! &,! -.N0,%(! 5-+K'(! %-;';1! (9'! '**'%(! 5-)! F'! ;0%9! (9-(! &,,$:-(&:'! -%(&:&()! +'A-('.! ($! (9'! 

;'%$,.-+)! 5-+K'(;! L&AA! .'M',.! $,! (9'! -F&A&()! $*! (9'! *&+5! ($! ',('+! (9'! M+&5-+)! 5-+K'(! 

;0%%';;*0AA)T! $+! ! (9'! -F&A&()! $*! (9'! *&+5! ($! U4-&,! -%%';;V! ($! (9'! ,'%';;-+)! A&,K;! F'(L'',! (9'! 

M+&5-+)!-,.!!A&,K'.!5-+K'(;B! 

! 

.6)7"%804996:);0 

R9'+'! ;''5! ($! F'! %A'-+! ;&5&A-+&(&';! F'(L'',! -.N0,%(! 5-+K'(;! -,.! ,'(L$+K! '**'%(;1! F0(! (9'! 

*$+5'+!&;!.'*&,'.!-F$:'!-;!L9'+'!%$,;05M(&$,!$*!-!M+$.0%(!&,!$,'!5-+K'(!%+'-(';!.'5-,.!&,! 

-,$(9'+1! L9&A'! ,'(L$+K! '**'%(;! -+'! 5$+'! +'A-('.! ($! 9$L! %$,;05M(&$,! $*! -! M+$.0%(! &,! $,'! 

5-+K'(!&,%+'-;';!.'5-,.!*$+!'&(9'+!>&?!(9-(!M+$.0%(!$+!>&&?!-!M+$.0%(!&,!-!.&**'+',(!5-+K'(!(9-(! 

&,! (0+,! 5-)! A'-.! ($! &,%+'-;';! &,! .'5-,.! *$+! (9'! $+&4&,-A! M+$.0%(! >;''! W$((&,4'+! >HII7?! *$+! 

.&;%0;;&$,!$*!,'(L$+K!'**'%(;?B! 

84-&,!,'(L$+K!'**'%(;!9-:'!&5M$+(-,(!&5MA&%-(&$,;!*$+!(9'!%-M-F&A&()!$*!*&+5;!($!;0MMA)!*0(0+'! 

5-+K'(;! -;;$%&-('.! L&(9! &,,$:-(&$,B! #0MMA)! ($! (9'! *0(0+'! 5-+K'(! L&AA! +'X0&+'! -! ;(+-('4)! *$+! 

&,('+-%(&,4!L&(9!(9'!,'(L$+K!'**'%(;!&,!;0%9!-!L-)!-;!($!5-K'!(9'!',(+)!;0%%';;*0A!-,.!;0;(-&,!-! 

M$;&(&$,! &,! (9'! 5-+K'(B! D,! M-+(&%0A-+1! ,'(L$+K! '**'%(;! %-,! 4','+-('! &5M$+(-,(! F','*&(;! *$+! 

&,%05F',(! ('%9,$A$4&';! -,.! (90;! %+'-('! &,('+G('5M$+-A! A&,K;! F'(L'',! %0++',(! -,.! *0(0+'! 

5-+K'(;1!-;!.&;%0;;'.!&,!(9'!*$AA$L&,4!;'%(&$,B! 

! 

! 

!"#$%&'()!"!#$%&'()!*$+!#%&',%'!-,.!/.0%-(&$,1!2,&('.!3&,4.$5! 67! 

147



!"##$%&'()*+,*-,*./0123,*4%%"56#$"%)*78%69$:;*"



8+%9&:';!$*!$+!&(;!.';%',.-,(?!&,!(9'!*0(0+'!5-+O'(B! 

! 

D,!(9'!%-;'!$*!&,.&+'%(!,'(L$+O!'**'%(;1!(9'!+'A':-,(!A&,O!F'(L'',!5-+O'(;!&;!&,.&+'%(!F'(L'',! 

(9'! %0++',(! +'A':-,(! 5-+O'(! -,.! (9'! *0(0+'! 5-+O'(1! F'%-0;'! &(! 'P&;(;! (9+$049! -! 9$+&S$,(-AA)! 

+'A-('.!5-+O'(B! 

! 

#0%9!+'A-(&$,;9&M;!-+'!(9'+'*$+'!&5M$+(-,(!&,!-!A-('+!;(-4'1!-,.!.&;%0;;'.!*0+(9'+!F'A$LB!K9'! 

*$AA$L&,4! (-FA'! ;044';(;! M$;;&F&A&(&';! *$+! &,('+G('5M$+-A! A&,O;! F'(L'',! 5-+O'(;! (9-(! 5-)! M0(! 

*&+5;!L9&%9!%0++',(A)!;0MMA)!-!+'A-('.!%0++',(!5-+O'(!-(!-!;&4,&*&%-,(!%$5M'(&(&:'!-.:-,(-4'!&,! 

(9'!%-M-F&A&()!($!;0MMA)!(9'!*0(0+'!5-+O'(!(9-,!*&+5;!(9-(!.$!,$(B! 

! 

*+,+%+-.+/0 

849&$,1!TB!-,.!=BU&**&(9!>HII7?!R$5M'(&(&$,!-,.!U+$L(91!R-5F+&.4'1!Q-B!QDK!T+';;! 

8++$L1!3BVB!>6WJH?1X/%$,$5&%!N'A*-+'!-,.!(9'!8AA$%-(&$,!$*!=';$0+%';!*$+!D,:',(&$,X1!&,!(9'!=-('!-,.!Y&+'%(&$,!$*! 

D,:',(&:'!8%(&:&()Z!/%$,$5&%!-,.!#$%&-A!E-%($+;1!T+&,%'($,Z!T+&,%'($,!2,&:B!T+';;1!JIWGJH7! 

>6W[\?1!=&:-A+)!&,!6W[W?1_/5M&+&%-A!#(0.&';!$*!D,,$:-(&$,!-,.!Q-+O'(!#(+0%(0+'_1!&,!`-,.F$$O!$*! 

D,.0;(+&-A!a+4-,&S-(&$,1!@$AB!DD1!=B!#%95-A',;''!-,.!=BYBN&AA&41!'.;1!85;('+.-5Z]$+(9!`$AA-,.! 

Y$;&1!RB!>6W[[?1!_#$0+%';1!T+$%'.0+';!-,.!Q-%+$'%$,$5&%!/**'%(;!$*!D,,$:-(&$,X1! 

V$0+,-A!$*!/%$,$5&%!^&('+-(0+'!HJ>b?1!66HIG66\6! 

/-;A')1!YB!-,.!VB3A'&,F'+4!>HI6I?1!]'(L$+O;1!R+$L.;!-,.!Q-+O'(;1!R-5F+&.4'Z!R-5F+&.4'!2,&:B!T+';;! 

/:-,;1!YB1!`-4&01!8B!-,.!=B!#%95-A',;''!>HIIJ?1!D,:&;&FA'!/,4&,';1!`$L!#$*(L-+'!TA-(*$+5;!Y+&:'!D,,$:-(&$,!-,.! 

K+-,;*$+5!D,.0;(+&';1!R-5F+&.4'1!Q-B!QDK!T+';;! 

! 

!"#$%&'()!"!#$%&'()!*$+!#%&',%'!-,.!/.0%-(&$,1!2,&('.!3&,4.$5! 67! 

149



!"##$%&'()*+,*-,*./0123,*4%%"56#$"%)*78%69$:;*"**'!B.C32D:!7*E'9':!F$+?'E9':!4?'!;G#+E4#9':!A#"!+?!#"9!=E*54E!A!M?'G8#,+4E!^,-4?+_4#+*?:!;4,W9#!7*,

2.2 NETWORK ECONOMICS FOR OPEN SOURCE TECHNOLOGIES 


!"#$%&'()*%)+,&*$-)+&*'%&'().'('$/&0)1,2/%$-)3)4,-567)8,59) 

:2;-*&$%9K*!L:$E3>) M#:>! 78A$9:! G:9K>8&8D%:4*! N>>8O3;%8>*! P:;Q8$R! @98>8E%94*! S%9:>4%>D*! NG! 

N>?A4;$?!T:O:&8#E:>;!UHVTW!%4!3!9$A9%3&!#K:>8E:>8>!X8;K!=$8E!;K:!#8%>;!8=!O%:Q!8=!;K:! 

%>?%O%?A3&!=%$E!3>?!;K:!:98>8E:=%;4!;8!3!=%$E!=$8E! 

%>>8O3;%>D!3$:!&%R:&;!=$8E!;K:!489%3&!X:>:=%;4'!N>!;K:!3X4:>9:!8=!3>%4E! 

#$:O:>;%>D!%;*!;K:!X:>:=%;4!;8!3>!%>>8O3;%>D!=%$E!3$:!&%R:&;$D!=%$E4'!N>!4A9K!3!49:>3$%8*!=%$E4!3$:!A>&%R:&>8O3;:'!GKA4!3998$?%>D!;8! 

98>O:>;%8>3&! ;K%>R%>D*! =%$E4! >::?! ;8! K3O:! 48E:! 48$;! 8=! $:Q3$?! =8$! %>>8O3;%>D'! N>;:&&:9;A3&! 

#$8#:$;;4!3>?!98#43;%8>'!"!X%D!#8$;%8>!8=!;K:! 

HVT!&%;:$3;A$:!K34!=89A4:?!8>!;K:!8#;%E3&!#3;:>;4[!?A$3;%8>!3>?!X$:3?;K!3>?!;K:!%>9:>;%O:!8=! 

=%$E4!;8!%>>8O3;:'! 

! 

Y8Q:O:$*!3!?%==:$:>;!;$:>?!K34!:E:$D:?!;K:4:!?3D!#$8&%=:$3;%8>! 

8=! K%\;:9K! U>:;Q8$RW! %>?A4;$%:4'! N>4;:3?! 8=! ;$D! ;8! D:;! :]9&A4%O:! 8Q>:$4K%#! $%DK;4*! 3>! 

%>9$:34%>D!>AEX:$!8=!=%$E4!3$:!E3R%>D!;K:%$!;:9K>8&8D8&8D8!&8>D:$!#$8#$%:;3$!48A$9:_!UI3>:*!@98>8E%4;*!(++/W'!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 

! 

N>! ;K:! 48\93&&:?! X$8Q4:$! Q3$*! %>! ;K:! ,``+4*! Q:! K3O:! Q%;>:44:?! %>;:>4:! 98E#:;%;%8>! %>! ;K:! 

E3$R:;! =8$! %>;:$>:;! X$8Q4:$4! X:;Q::>! J%9$848=;! 3>?! P:;493#:'! P:;493#:! K3?! 3! E358$! K:3?! 

4;3$;!8>!J%9$848=;*!98>;$8&&%>D!`+a!8=!;K:!X$8Q4:$!E3$R:;!XD!%>!;K:!E3$R:;'!b%;K!;K:!:>;$D3D:?!%>!3!$39:!;8! 

K3O:!;K:!X:4;!3O3%&3X&:!#$8?A9;'!! 

! 

L%O:>!;K:!%>;:>4:!98E#:;%;%8>!X:;Q::>!;K:!;Q8!=%$E4*!X?!8=!,``-!J%9$848=;!Q34!#$%9%>D! 

;K:!N>;:$>:;!@]#&8$:$!=$::'!N>!98>;$34;*!P:;493#:!Q34!9K3$D%>D!98$#8$3;%8>4!&%9:>4%>D!=::4!=8$! 

! 

! 

! 

@,AB/*C09:!3>?!@?A93;%8>*!B>%;:?!C%>D?8E! 

! 

,F/! 

151



!"#$%&'()*%)+,&*$-)+&*'%&'().'('$/&0)1,2/%$-)3!++.14) 

5,-678)9((2'):)!;/*-?:) 

) 

2(*%G) A0'*/) K/,L('/6) MB) A0') '%") ,C) ?NN:8) O*&/,(,CA) 0$") (A,-'%) $) -$/G') &02%P! ,C) Q'A(&$;'R() 

H$/P'A)(0$/'6)Q'A(&$;')'#'%A2$--B)C,--,L'")(2*A)$%")(A$/A'")G*#*%G)$L$B)*A()K/,L('/)C/''6) 

) 

S0') 'TA'%"'") K$AA-') K'AL''%) O*&/,(,CA) $%") Q'A(&$;') 0$") *A() A,--) ,%) A0') ;/,C*A() ,C) K,A0) 

&,H;$%*'(6)9%)?NNU)Q'A(&$;')&$H')2;)L*A0)$)%'L)(A/$A'GB)$%")"'&*"'")A,)/'-'$(')*A()(,2/&') 

&,"'8)A0')$&A2$-)-*%'),C);/,G/$HH*%G)-$%G2$G'8)C,/)A0')Q'A(&$;')V,HH2%*&$A,/6)S0*()$--,L'") 

2('/()$%")"'#'-,;'/()A,)-,,P)*%(*"')A0')L,/P*%G(),C)A0')K/,L('/8)A,)H,"*CB)A0')(,CAL$/')$%") 

'#'%)A,)/'"*(A/*K2A')A0')%'L)#'/(*,%)2%"'/)A0'*/),L%)K/$%")%$H'8);/,#*"'")A0$A)A0')H,"*C*'") 

(,2/&')&,"')L$()$-(,)C/''-B)$#$*-$K-'6)S0')L0,-')*"'$)*()A,)A2/%)A0')'%A*/')*%A'/%'A)&,HH2%*AB) 

*%A,)$)#$(A)/'('$/&0)"*#*(*,%)C,/)Q'A(&$;')K/,L('/6)) 

) 

S0') A'/H) W,;'%) (,2/&') (,CAL$/'X) 0$() K''%) L*"'-B) 2('") *%) A0') ;,;2-$/) $%") ;/,C'((*,%$-) 

-*A'/$A2/') 35$/*$%) $%") +0$;*/,8) =>>YZ) ['/%'/) $%") S*/,-'8) =>>746) 9%(A'$") ,C) P'';*%G) A0'*/) 

A'&0%,-,GB) ;/,;/*'A$/B8) A0') C*/H() L*--) "*(A/*K2A') *A) C/''-B6) 9A) *() A0*() ;0'%,H'%,%) A0$A) A0*() 

&0$;A'/) $AA'H;A() A,) 'T;-,/'6) \') L*(0) A,) (A2"B) A0') "'&*(*,%() ,C) C*/H() L0'A0'/) A,) P'';) A0'*/) 

A'&0%,-,G*'();/,;/*'A$/B),/)%,A6) 

) 

D#'%)A0,2G0)A0')L0,-')2%,/A0,",T),;'%)$;;/,$&0)H$B)(''H)&,2%A'/*%A2*A*#'8)Q'A(&$;')L$() 

%,A)A0'),%-B),%')L0,)'H;-,B'")*A6)!;$&0'8)$);/,G/$H)C,/)('/#*%G)L,/-")L*"')L'K)(*A'(8)$%") 

+'%"H$*-8) $) ;/,G/$H) A0$A) /,2A'() $%") "'-*#'/() *%A'/%'A) '-'&A/,%*&) H$*-8) $/') 'T$H;-'() ,C) C/'') 

,;'%)(,2/&');/,G/$H()A0$A)",H*%$A')A0')H$/P'A6)];'%)(,2/&')$;;/,$&0'()0$#')K''%)'T;$%"'") 

A,) A0') K*,A'&0%,-,GB) $%") 0'$-A0) &$/') *%"2(A/*'(8) D&,%,H*(A) 3=>>7$46) [*%2T8) $%) *%&/'$(*%G-B) 

;,;2-$/),;'/$A*%G)(B(A'H)&/'$A'")*%)?NN?)*()$%,A0'/)&-$((*&)'T$H;-'),C)(2&&'((C2-),;'%)(,2/&') 

(,CAL$/'6) O$%B) ,C) A0') ;/,G/$HH'/() $%") (,CAL$/') "'(*G%'/() $"#,&$A*%G) ]++) H$B) (0$/') $) 

2A,;*$%)#*(*,%),C)(,CAL$/')"'#'-,;H'%A8),/)A0'B)H$B)(*H;-B)L$%A)A,);/,#')A0'H('-#'()A,)K') 

K'AA'/) A0$%) (,CAL$/') G*$%A) O*&/,(,CA6) ^,L'#'/8) A0') L0,-') *"'$) ,C) ]++) H$B) %,A) K') (,) $%A*< 

&$;*A$-*(A*&) $() *A) (''H(6) 9A) *() 0$/") A,) K'-*'#') A0$A) ;/,C*A



!"##$%&'()*+,*-./012,*3'#4"(5*67"%"8$'9*:"(*;



!"#$%&'()*%)+,&*$-)+&*'%&'().'('$/&0)1,2/%$-)3!++.14) 

5,-678)9((2'):)!;/*-?:) 

) 

&,%(2H'/) J$('6) K0$A) *() J'&$2(') A0') ",L%(A/'$H) (,CAL$/') &,H;$%*'() A02() &$%) A$;) *%A,) A0*() 

-$/G'/)%'AL,/M),C)&2(A,H'/(6)K0*()*H;/,#'()A0')N2$-*AB),C)A0')O+6)K0*()*()A0')*%"*/'&A)'CC'&A),C) 

%'AL,/M),%)&,%(2H'/(P)2A*-*A*'(6) 

) 

O;'%)(,2/&')(B(A'H)$;;/,$&0'()'%A$*-)(,H')A/$%(*A*,%$-);/$&A*&$J-')"*($"#$%A$G'()$G$*%(A)A0') 

;/'#$*-*%G)(A$%"$/"6) 

) 

Q'&$2('),C)J2*-"*%G)2;)$)%'AL,/M)A0'/')L*--)J')C'L'/)$;;-*&$A*,%()L0*&0)$/*(')$()/,$"J-,&M() 

A,L$/"() C2/A0'/) "*(('H*%$A*,%8) C,/) 'R$H;-'8) $() *%) A0') &$(') ,C) S*%2R8) ,%-B) $) (*%G-') "*G*A) 

;'/&'%A$G') L,/M() C,/) &,H;2A'/) G$H'() 3D&,%,H*(A8) =>>7J46) 9%&,H;$A*J*-*AB) *((2'() L*A0) A0') 

",H*%$%A),;'/$A*%G)(B(A'H)$/')$-(,),%'),C)A0')J*GG'(A),J(A$&-'()C,/)C2/A0'/)"*(('H*%$A*,%8)*%) 

;$/A*&2-$/8) C,/) -$/G') &,/;,/$A*,%() L0,) L,2-") C$&') (*G%*C*&$%A) (L*A&0*%G) &,(A() L*A0) ;,((*J-') 

,;'/$A*,%$-)"*(/2;A*,%(6)O%)A0')J'%'C*A)(*"'8)0,L'#'/8),;'%)(,2/&')(B(A'H()H$B)*%"2&').TU)*%) 

&,H;-'H'%A$/B);/,"2&A()$%")('&A,/()J'&$2('),C)-,L'/)-*&'%(')C''(8)A02()-,L'/)&,(A(),C).TU8) 

A0'')$/')H,/')*%&'%A*#'()A,)'%-$/G')$%")'R;$%");/,"2&A)"'#'-,;H'%A)$%")$;;-*&$A*,%()J'B,%") 

&,/') *%%,#$A*,%(6) Q'&$2(') ,C) *A() &,%A*%2,2(-B) ,%G,*%G) &0$/$&A'/) *A) *() H,/') -*M'-B) A,) J') 

*%&/'H'%A$-) A0$%) ('H*%$-) L0*&0) ;,('() N2'(A*,%() ,C) A0'*/) ,L%) $() A,) L0'A0'/) ,;'%) (,2/&') 

(B(A'H()&$%)G'%2*%'-B)C,(A'/)*%%,#$A*,%6)V,/)"'"*&$A'")$;;-*&$A*,%()$()*%*A*$--B)J2*-A)2;)C/,H) 

(&/$A&0)(2&0)$()('$/&0)A0/,2G0)W,,G-')*A)H$B)B*'-")(,H')J*G)$"#$%A$G'()8)%,A)A0')-'$(A)J'&$2(') 

,C)&,H;$/$A*#'-B)-,L)*H;-'H'%A$A*,%)$%"),;'/$A*,%$-)&,(A(6) 

) 

K0*();$;'/)*(),/G$%*X'")$()C,--,L(Y)K0')%'RA)('&A*,%)-,,M()$A)A0')-*A'/$A2/')/'-$A'")A,)A0*()L,/M6) 

9%)+'&A*,%)Z6)L');/'('%A)A0')H,"'-6)9%)+'&A*,%)78)L')-,,M)$A)A0')'N2*-*J/*2H)/'(2-A()$%")G*#') 

*%A'/;/'A$A*,%(6)+'&A*,%)[)*%"*&$A'()'RA'%(*,%()A,),2/)H,"'-)$%")"/$L()&,%&-2(*,%(),%)C2/A0'/) 

2('6) 

* 

+,)-.)&.'/*0&1-.2&.'*,.3*4#-.*5"6%1-*7-1(."8"'&-2* 

K0') *((2'() &,#'/'") *%) A0*() ;$;'/) /'-$A') A,) L,/M) ,%) %'AL,/M) 'RA'/%$-*A*'(8) .TU8) '%A/B) 

"'A'//'%&')$%")-*&'%(*%G6) 

) 

.TU) *() $%) 'RA'%(*#'-B) /'('$/&0'") $/'$) *%) *%"2(A/*$-) ,/G$%*X$A*,%) (A$/A*%G) L*A0) !//,LP() 

;*,%''/*%G)$/A*&-')L0'/')0')$(M'"Y)\]0$A)*()A0')G$*%)C/,H)*%%,#$A*,%)A,)$)C*/H)A0$A)*()A0'),%-B) 

,%')A,)2%"'/A$M').TU8)G*#'%)A0$A)*A()*%%,#$A*,%)*();/,A'&A'")JB)$);$A'%A),C)2%-*H*A'")"2/$A*,%^_) 

3!//,L8)?`I=46))+*%&')A0'%)A0'/')0$()J''%)$)(;$A'),C)/'('$/&0),%).TU)&,#'/*%G)*((2'()(2&0)$()A0') 

*%&'%A*#'() A,) *%%,#$A'8) ;$A'%A) /$&'(8) L'-C$/') *H;-*&$A*,%() ,C) .TU8) &0,*&') $%") $",;A*,%) ,C) 

A'&0%,-,G*'(6)a,L'#'/8)H,(A),C)A0*()L,/M)0$()C,&2('"),%)&,%#'%A*,%$-)H$/M'A()/$A0'/)A0$%),%) 

H$/M'A()L*A0)%'AL,/M)'RA'/%$-*A*'()*%)L0*&0)"B%$H*&),/)b+&02H;'A'/*$%c)&,H;'A*A*,%)'#,-#'() 

3D#$%()$%")+&0H$-'%(''8)=>>?46)]')'R;-*&*A-B)-,,M)$A)A0')'CC'&A),C)%'AL,/M)'RA'/%$-*A*'(),%).TU) 

&,H;'A*A*,%)$%")*%A/,"2&')A0');,((*J*-*AB),C)C*/H()%,A)L$%A*%G)A0'*/)A'&0%,-,GB)A,)J');/,;/*'A$/B) 

L0*&0)*()%,A)/'&,G%*X'")JB)A0')A/$"*A*,%$-).TU)-*A'/$A2/'6) 

) 

]*A0*%)A0')A,;*&),C).TU8)A0'/')0$()J''%)(,H')L,/M)"'#,A'")A,)-*&'%(*%G6)K0*()-*A'/$A2/')A$M'()$() 

A0')(A$/A*%G);,*%A),%'),/)H,/')C*/H()0$#*%G)$);$A'%A6)S*&'%(*%G)*()A0'%)$)H'$%(),C)"*(('H*%$A*%G) 

$%) *%%,#$A*,%6) !H,%G) A0') *%&'%A*#'() C,/) -*&'%(*%G) $/') ;/,"2&A) H$/M'A) &,H;'A*A*,%) L0*&0) 

&/'$A'() *%&'%A*#'() C,/) H$%$G'/() L0,) L,2-") ,A0'/L*(') 'R;-,*A)A0'*/) H,%,;,-B) ;,(*A*,%(8) &,(A) 

($#*%G() A,) A0') -*&'%(''() L0*&0) &,2-") J') $;;/,;/*$A'") JB) A0')-*&'%(*%G) C*/H8) $%") -,L'/*%G) 

/*#$-(P)*%&'%A*#'()A,)*%#'%A)$/,2%")A0')*%%,#$A*,%6)F$AX)$%")+0$;*/,)3?`d[4)-,,M)$A)A0') 

*%&'%A*#'()A,)'%G$G')*%)-*&'%(*%G),%&')$%)*%%,#$A*,%)*()"'#'-,;'")*%)$)L,/-"),C);'/C'&A);$A'%A(6)) 

) 

154 

!"#$%&'())@)+,&*'AB)C,/)+&*'%&')$%")D"2&$A*,%8)E%*A'")F*%G",H) 

) 

?I:)



!"##$%&'()*+,*-./012,*3'#4"(5*67"%"8$'9*:"(*;-! 5#,4#:-;! *B! 4! ,:5:4,?"! ?:>#+@:5!#*!'+55:J+>4#:!+>B*,J4#+*>!J4;!C:!#**!':!G.HH3I! 

:($!#"+5!B45"+*>!?4>!+>?,:45:!$,*B+#5!C;! 

-+@+>-!?*>5LJ:,5!4>!4'':'!!"#$"%!&$'#*!+>?L,!#":!5:#L$!?*5#!*B!$L,?"45+>-!#":!J4+>!?*J$*>:>#)!! 

9":!,:5L!*$:>!5;5#:J!+5!!+#! 

+5!J*,:!:!#":!5"4,:! *B!! 

#":!J4+>!?*J$*>:>#!+>!#":!#*#4! ,:@:,5:'! ?*J$4,:'! #*! #":! #,4'+#+*>45+>-! 5#:4'! *B! 5+>-! C:+>-! L5:'! 45! 4! J:4>5! *B! '+BBL5+>-! 4>! +>>*@4#+*>! ()%$*' +#! "45! 

*??L,,:'A!+>!*L,!J*':5+>-!4?#L4:,! G.H3TA! .H38I! 4>'! D4#E! 4>'! F"4$+,*! G.H3TA.H38I! 4,:! #":! $+*>::,+>-!K*,=5! 

:($-! #":! +J$5! *B! >:#K*,=! :(#:,>4! +>'L5#,+4+E4#+*>)! S4,,:'! 5+':! 4>'! 5"*K! #"4#! #":! :(+5#:>?:! *B! >:#K*,=! 

:(#:,>4!$,*C'!#"L5!#*!4!JL! *B! 4! >:K! 5#4>'4,'! J+-"#! C:! #**! 5'!F4:,M5!K*,=5!C;!+>?-! C#)!9":;!L5:!#":+,!J*':B*,?:!#":!+J$*,#4>?:!*B!?*>5LJ:,! 

:($:?#4#+*>5!+>!J4,=:#5!K+#"!>:#K*,=!:(#:,>4-!?*J$4#+C+'!K":#":,!5+':!$4;J:>#5!4,:!B:45+C'!F"4$+,*!G.H38I!5"*K! 

#"4#! +>! #":! $,:5:>?:! *B! >:#K*,=! :(#:,>45!L>':,!K"+?"!B+,J5!J4;!L5:!?*J$4#+C+-! 

?*J$:#+#+*>)! 

! 

W"+:#K*,=! :(#:,>4:K! $,*'L?#! K"+?"! :JC*'+:5! #:?">*!:UL+:K! $,*'L?#! +5! +>#,*'L?:'! #**! 5**>! ?*J$4,:'! #*! #":!5*?+45! 

C:?4L5:! #":! 5**>:,! #":! $,*'L?#! +5! +>#,*'L?:'A! #":! 5**>:,! #":! B+,J! ?4>! 5#4,#! CL+-!+#5!C:>:B+#5)!["*+!G.HH0I!!4! 

J4,=:#!K+#"!>:#K*,=:(#:,>4



!"#$%&'()*%)+,&*$-)+&*'%&'().'('$/&0)1,2/%$-)3!++.14) 

5,-678)9((2'):)!;/*-?:) 

) 

K')(A2"*'()A0')*%&'%A*#'),C)A0')H,%,;,-*(A)A,)*%A/,"2&')$%)*%&,H;$A*L-')*H;/,#'");/,"2&A)*%) 

A0') ;/'('%&') ,C) %'AM,/N) 'OA'/%$-*A*'(6) F/*(A*$%('%) 3?JJI4) (A2"*'() A0') &,%('P2'%&'() ,C) %'AM,/N) 

'OA'/%$-*A*'() ,%) A0') /*(N*%'(() ,C) .QR) ;/,S'&A() &0,('%) LB) $%) '%A/$%A) $%") $%) *%&2HL'%A6) K') 

(0,M() A0$A) A0') *%&2HL'%A) &0,,('() $) A,,) /*(NB) ;/,S'&A) A0$A) A,,) ,CA'%) -'A() $) %'M) C*/H)M*A0) $%) 

*%&,H;$A*L-')A'&0%,-,GB)'%A'/)$()&,H;$/'")A,)A0')(,&*$-),;A*H2H6)9%)$""*A*,%8)A0')'%A/$%A) 

0$() $%) *%&'%A*#') A,) &0,,(') H,/') &'/A$*%) ;/,S'&A() A0$%) $/') (,&*$--B) ,;A*H$-) $%") A0'(') 

(A/'%GA0'%) A0') ;,((*L*-*AB) ,C) $",;A*,%) ,C) $%) *%&,H;$A*L-') A'&0%,-,GB6) .'G*L'$2) $%") .,&N'AA) 

3?JJT4)$%$-BU')&,H;$A*L*-*AB)&0,*&'(),C)AM,)C*/H()M0*&0)H2(A)$-(,)"'&*"')M0'%)A,)*%A/,"2&') 

A0'*/) G,,"() *%) $) H$/N'A) &0$/$&A'/*U'") LB) %'AM,/N) 'OA'/%$-*A*'(6) V0'B) (0,M) A0$A) A0') C*/H(W) 

*%&'%A*#'() A,) $&0*'#') &,H;$A*L*-*AB) "';'%") &/2&*$--B) ,%) A0') A*H') $A) M0*&0) A0') "'G/'') ,C) 

&,H;$A*L*-*AB) H2(A) L') &0,('%6) V0') &2//'%A) ;$;'/) *() $-(,) /'-$A'") A,) A0') -*A'/$A2/') ,%) 

*%#'(AH'%A)$()$)H'$%(),C)'%A/B)"'A'//'%&'6)X*/H()&,H;'A')%,A),%-B)M*A0)'O*(A*%G)C*/H(8)$()*() 

A/$%(;$/'%A8)L2A)$-(,)M*A0);,A'%A*$-)'%A/$%A(6)9%)$%)'OA'%(*,%),C)A0')+;'%&')H,"'-)R*O*A)3?JT>4) 

(0,M() A0$A) $%) *%&2HL'%A) C*/H) H$B) H$N') *//'#,&$L-') &,HH*AH'%A) ,C) *%#'(AH'%A) *%) ,/"'/) A,) 

$-A'/)A0')*%*A*$-)&,%"*A*,%(),C)A0');,(A)2%-*N')A0')'%A/$%A)3OD()(>4*([')&,%(*"'/)$)AM,)(A$G')G$H')*%)$)H$/N'A) 

M*A0)%'AM,/N)'OA'/%$-*A*'(6)V0');/,"2&A(),C)A0')C*/H()$/')*%&,H;$A*L-')3*6'68)A0'B)0$#')(';$/$A') 

%'AM,/N(46) 

) 

9%) A0') C*/(A) (A$G'8) C*/H() &0$/G') $) -*&'%(') C'') 3+4) A,) ,A0'/) ",M%(A/'$H) C*/H() C,/) 2(') ,C) A0'*/) 

;/,"2&A)3C,/)'O$H;-'8)\*&/,(,CA)-*&'%('()A0'*/)[*%",M(),;'/$A*%G)(B(A'H)A,),A0'/)&,H;$%*'() 

A,)"'#'-,;)$;;-*&$A*,%()(,CAM$/'46)V0')%2HL'/),C)",M%(A/'$H)C*/H()M*--*%G)A,)M,/N),%)$)C*/HW() 

;/,"2&A)3H*4)"';'%"(),%)A0')-*&'%(')C'')&0$/G'")$%")$-(,)A0')*%*A*$-)%'AM,/N),C)A0')C*/H)3A0') 

&,%(2H'/)L$('),%)C,/)C*/H)*46) 

( 

-(%)(.(/(,%(0(+%8)(.(1(2( 3?4) 

) 

.(%3($)H'$(2/'),C)A0');,A'%A*$-)H$/N'A)(*U'),C)",M%(A/'$H)C*/H()*%"';'%"'%A),C)A0')-*&'%(')C'') 

$%")*%*A*$-)%'AM,/N6) 

) 

V0')(H$--'/)A0')-*&'%(')C'')&0$/G'"8)A0')H,/')C/''-B)A0')C*/H)"*(A/*L2A'()*A()A'&0%,-,GB)$%")*%),2/) 

A'/H*%,-,GB8) A0') H,/') ],;'%^) A0') C*/H) *(6) [') $((2H') %,) &,H;'A*A*,%) *%) G'AA*%G) 

",M%(A/'$H) C*/H() A,) L2B) A0') C*/H(W) -*&'%('(6) -%) "'&/'$('() $() +*) *%&/'$('() L'&$2(') $() A0') 

-*&'%('() L'&,H') H,/') 'O;'%(*#') C'M'/) ",M%(A/'$H) C*/H() L2B) A0'H6) V0*() *() S2(A) A0') 

(A$%"$/") $/G2H'%A) C,/) $) ",M%M$/") (-,;*%G) "'H$%") *%) ;/*&'6) !() C,/) A0') /'-$A*,%(0*;) 

L'AM''%( -%) $%") ,%(4) (*%&') A0') ",M%(A/'$H) C*/H() $/') "'#'-,;*%G) ;/,"2&A() A0$A) &,2-") L') 

2('")*%)&,%S2%&A*,%)M*A0)A0')C*/HW();/,"2&A)3-*N')&,H;-'H'%A(48)A0')-$/G'/)A0')*%(A$--'") 

L$('),C)&2(A,H'/()A0$A)A0')C*/H)0$(8)A0')-$/G'/)*()A0');,A'%A*$-)"'H$%")C,/)A0')",M%(A/'$H) 

C*/H(W);/,"2&A6))V02(8)A0')H,/')A0')",M%(A/'$H)C*/H()$/')M*--*%G)A,);$B)C,/)A0')-*&'%('6) 

) 

_'A)2()"'C*%'`))9C)))+% ( !a)+5 ( !A0'%)C*/H)*)&0,,('()A,)L')H,/'),;'%)A0$%)C*/H)S6)3!-A'/%$A*#'-B(C*/H)S) 

&0,,('()A,)L')-'((),;'%)A0$%)C*/H)*4*( 

) 

) 

156 

!"#$%&'())@)+,&*'AB)C,/)+&*'%&')$%")D"2&$A*,%8)E%*A'")F*%G",H) 

) 

?IJ)



!"##$%&'()*+,*-./012,*3'#4"(5*67"%"8$'9*:"(*;##*! 5:@@! *?#$?#)! N:9;:! D:! ,:5#,+;#! #":! 

:9#,49#!#*!;"**5+9-!4!@+;:95:!*?,9*#!-4=:)!B"?5H!#":!X'*D95#,:4=!



!"#$%&'()*%)+,&*$-)+&*'%&'().'('$/&0)1,2/%$-)3!++.14) 

5,-678)9((2'):)!;/*-?:) 

) 

!()*%),A0'/)AI,



!"##$%&'()*+,*-./012,*3'#4"(5*67"%"8$'9*:"(*;!'9:+5+*;!*?!?+,@!A!4;'!#"9!9BC+D+E,+C@!D+:9;59! 

?995!?*,!#"9!#!+;!#"9!H*C,;*#!5CE-4@9G!#*!#"9!'+??9,9;#!D+:9;59! 

?995!#"4#!#"9!+;:C@E9;#!:4;!:"4,-9)!I*,!$,**?5!$D9459!,9?9,!#*!#"9!4$$9;'+()!89!:4;!5"*!*C#!*?!#"9!@4,J9#!4;'!?*,!"!##>##" ! = #+#!+5!#*!9;#9,)!K"9,9!9(+5#5!"! ? ##>##" ! = ! 

5C:"!#"4#!+;!#"9!+;#9,L4D!M"!# = #6#"! ? N##"9!E95#! ,95$*;59! *?! #"9! 9;#,4;#! +5! #*! :"4,-9! #"9! D4,-95#! 

4::9$#4ED9! D+:9;59! ?99! #"4#! @4J95!*$#+@4D! (A! 9(4:#D>! /G! 4;'! ?*,! ƒO! P! ƒ! ? ! +#! +5! *$#+@4D! ?*,! #"9! 

9;#,4;#!#*!9;#9,!4;'!:"4,-9!+#5!C;:*;5#,4+;9'!*$#+@4D!D+:9;59!?99!,95CD#+;-!+;!4!$*5+#+L9# 

@A# 

% 

&'()*(+%,%&'(%-(.(-/%)0%12 ) 3%4%567%8%7(9(*+:6(%9'(%(69*;%7(+%5/%0)--)@/$% 

.) A;#,>!+5!ED*:J9'!+?!Q.!R!7!S!(O! * !"!T!R!7J)! 

7) I*,!T!R!7J!"!Q.!R7!S!(O! * !"!TRJ&7G!9;#,>!+5!,95#,+:#9'!)!#"9!9;#,4;#! 

9;#9,5!#"9!@4,J9#)!Z*!?995!4;'!@4J9!+#5!@*;*$*D>!$,*?+#5!! "+-"! #"9;! #"9! +;:C@E9;#!:4;! ,95#,+:#! 

9;#,>!4;'!@4J9!@*;*$*D>!$,*?+#5)!O?!+#5!4'L4;#4-9!+5!"*!$,*?+#5)!K"C5! 

+;!#"+5!:459!#"9!+;:C@E9;#!+5!4ED9!#*!4:#!45!4!@*;*$*D+5#!+;!#"9!*C#$C#!@4,J9#!EC#!;*#!+;!D+:9;59! 

?995)! 

! 

O?!#"9!+;+#+4D!;9#



!"#$%&'()*%)+,&*$-)+&*'%&'().'('$/&0)1,2/%$-)3!++.14) 

5,-678)9((2'):)!;/*-?:) 

) 

!()&,/,--$/*'()A,)J/,;,(*A*,%)=)K')G'A)A0')/'(2-A)A0$A)A0'/')*()%,)'%A/B)"'A'//'%&')$%")K')$-(,)G'A)$) 

&,H;$/*(,%) L'AK''%) A0') H,%,;,-B) -*&'%(') C''() ,C) A0') *%&2HL'%A) $%") *A() -*&'%(')C''() K0'%) $) 

;,A'%A*$-)'%A/$%A)'M*(A(6) 

) 

!"#"$$%#&'(')*+#+',-'."'+./#&'0+/+##+.1+2' 

!A)'!"# $# 'A0')'%A/$%A);/,"2&'()%,A0*%G)$%")A0')*%&2HL'%A)N2(A)H$O'()A0');/,C*A()*A)K,2-")$()$) 

H,%,;,-*(A) *%) A0') ('&,%") (A$G') ,2A;2A) G$H'6) +*%&') !"# $# ' *() A0') -$/G'(A) -*&'%(') C'') C,/) A0') 

*%&2HL'%A)A0$A)"'A'/()'%A/B8)*C)#!"# %# !#!"# $# 'A0'%)A0'/')*()L-,&O'")'%A/B6)9C)!" % #&#!"# $# '0,K'#'/8)A0'%) 

'%A/B)*()%,A)L-,&O'"6)P2A)*A)$-(,)H'$%()A0$A)A0')*%&2HL'%AQ();/,C*A()H2(A)L')*%&/'$(*%G)$A)!"# $# ' 

3(*%&')A0')H,%,;,-B);/,C*A()$/')&,%&$#')*%)!" # '$%")C/,H)R'HH$)!6=)K')O%,K)A0$A)*%)A0')*%A'/#$-)S!"# 

$ ''!"# (# ]#'#A0')*%&2HL'%A)H$O'()H,%,;,-B);/,C*A()46)T'%&')A0')*%&2HL'%A)K*--)%,A)C*%")*A);/,C*A$L-') 

A,)&0$/G')!"# $# ',/)-'((6)U02()'%A/B)*()%,A)"'A'//'"6)) 

) 

!()$)&,/,--$/B)A,)U0',/'H)=)K')G'A)$)&,H;$/*(,%)L'AK''%)A0')H,%,;,-B)-*&'%(')C''(),C)A0') 

*%&2HL'%A)$%")*A()-*&'%(')C''()K0'%)$);,A'%A*$-)'%A/$%A)'M*(A(6) 

) 

!"#"$$%#&'3'4/'/*+'+56,$,7#,68'06"9"$&'$,1+.-+':++':;



!"##$%&'()*+,*-./012,*3'#4"(5*67"%"8$'9*:"(*;9:!#":!;:#A*,C!#:,=H!A:!5::!#"4#!#":!+;9>=D:;#!9"**5:5!#*!D:!:L:;!=*,:!*$:;!A"+8:! 

#":!:;#,4;#!9"**5:5!#*!D:!8:55!*$:;)!! 

! 

O,*=!#":!P$$:;'+(!QP!)./@!A:!5::!#"4#!5#4,#+;-!9#+*;!+;!!%"# +;! 

*,':,!#*!D,+;-!#":!=4,-+;48!:8#!+#!+5! 

,:!#":!9:;#,48!+':4!*#>,:! 

,:5:4,9"!:':'! #":! #,4'+#+*;48! M#:9";*8*-?! 5#:48+;-M! :=:,5H!;4=:8?!#"4#!9*;5>=:,5!A+88!$:,9:+L:!#":+,! 

$,*'>9#! #*! D:! ;*#! 45! 4##,49#+L:! 45! #":! +;9>=D:;#J5H! :L:;! +48+#?)! E#! 

A*>8'!5#+88!D:!+;#:,:5#+;-!#*!5::%#":!,:5>8#!*-"!#*!4!8:55:,!:(#:;#)!I":! 

:R>+8+D,+>=!;*A!;::'5!#*!5#,+C:!4!D484;9:!D:#A::;!#":!#A*!*$$*5+;-!:8'! :($:9#! #":! ,:5>8#! #"4#! #":! +;9>=D:;#! -*:5! =*,:! 

*$:;!#*!D:!,:+;9:!9*=$:#+#+*;!+;!#":!



!"#$%&'()*%)+,&*$-)+&*'%&'().'('$/&0)1,2/%$-)3!++.14) 

5,-678)9((2'):)!;/*-?:) 

) 

$%) *"'%A*&$-) $%") *%"';'%"'%A) "'H$%") C,/) *A() -*&'%('(6) J0*() &,2-") K') L2(A*C*'") *C) A0'/') $/') 

%2H'/,2() ",M%(A/'$H) C*/H() -,,N*%G) C,/) ;/,(;'&A*#') A'&0%,-,G*'() A,) "'#'-,;6) !) H,/') 

/'$-*(A*&)('AA*%G)H$B)K'),%')M0'/')K,A0)A0')*%&2HK'%A)$%")A0')'%A/$%A)$/')&,H;'A*%G)*%) 

-*&'%(') C''() C,/) ",M%(A/'$H) C*/H() A,) K2B) A0'*/) -*&'%('(6) D$&0) ",M%(A/'$H) C*/H) K2B() $A) 

H,(A),%')2%*A),C)A0')-*&'%('6)) 

) 

O,M'#'/8)A0*()(0,2-")%,A)&0$%G'),2/)/'(2-A()"/$(A*&$--B)$%")M')(0,2-")(A*--)('')P,;'%P)$()$%) 

'Q2*-*K/*2H)(A/$A'GB6)) 

) 

9%)C$&A8)M*A0)&,H;'A*A*,%8)K,A0)A0')*%&2HK'%A)$%")A0')'%A/$%A)$/')-*N'-B)A,)G,)H,/'),;'%) 

K'&$2(')%,M)A0'B)%''")A,)&,H;'A')A,)('--)A0'*/)-*&'%('(6) 

) 

R')&,2-")*%A/,"2&')&,(A()*%A,)A0')A'&0%,-,GB),C)A0')C*/H(6)9%),2/)&2//'%A)H,"'-8)M')0$#') 

%,/H$-*S'")K,A0)C*/H(T)&,(A()A,)S'/,6)U*CC'/'%&'()*%)A0')A'&0%,-,G*'(),C)A0')C*/H()H$B)(M$B) 

,2/)/'(2-A()'*A0'/)M$B)"';'%"*%G),%)M0,)0$()A0')K'AA'/)3-'(()&,(A-B4)A'&0%,-,GB6)9A)M,2-") 

$-(,)K')*%A'/'(A*%G)A,)('')0,M)A0')&,2;-*%G),C)A0')PA'&0%,-,GB)(A'$-*%GP)'CC'&A)$%");,(*A*#') 

&,(A()H$B)M,/N6)V,/)'W$H;-'8)*C)A0')*%&2HK'%A)C/''-B)"*(A/*K2A'()*A()A'&0%,-,GB)A0'%)A0') 

&,(A() C,/) A0') '%A/$%A) H$B) C$--) K'&$2(') *A) &$%) %,M) P(A'$-T) A0') A'&0%,-,GB) ,C) A0') 

*%&2HK'%A6) 

) 

9%),2/)H,"'-8)M')",%TA)0$#')'%A/B)"'A'//'%&')K2A)M')0$#')/'(A/*&A'")'%A/B)*%)M0*&0)A0') 

'%A/$%A)('--()-*&'%('()*%)A0')C*/(A)(A$G'8)K2A);/,"2&'()%,A0*%G)*%)A0')('&,%")(A$G'6)J0*() 

H$B) (''H) $) K*A) 2%/'$-*(A*&6) R*A0) ;,(*A*#') C*W'") &,(A(8) M') (0,2-") K') $K-') A,) 0$#') A0') 

(A$%"$/")'%A/B)"'A'//'%&')/'(2-A6)R')M,2-")(A*--)'W;'&A)/'(A/*&A'")'%A/B)K2A),%')M0'/')A0') 

'%A/$%A)*()T/'(A/*&A'"P)A,)(,H');,(*A*#'),2A;2A)-'#'-6) 

) 

!%,A0'/);,((*K-')'WA'%(*,%)*()H$N*%G)*%%,#$A*,%)(A,&0$(A*&8)*6'6)/';-$&*%G)A0')-*%N)!"#$%#&#$#'(%)# 

M*A0) $%#) X)$'(%*+%)# *#M0'/') + % #*() (A,&0$(A*&6) R') $/G2') A0$A) A0*() $G$*%) (0,2-") %,A) &0$%G') ,2/) 

/'(2-A() Q2$-*A$A*#'-B) (,) -,%G) $() A0') /'(2-A*%G) "*(A/*K2A*,%() C,/) Q2$-*A*'() 'W0*K*A) ('&,%") ,/"'/) 

(A,&0$(A*&) ",H*%$%&')



!"##$%&'()*+,*-./012,*3'#4"(5*67"%"8$'9*:"(*;/:N0/ONN2! 

EK*F*I+'B5)!A)!=.>>8?!@S"B!EK*F*I+K5!*U!AB#C*,D5M:!-)0#()*01&)*+$%&'()*+$&,$-).'/0(1*+$5(6*)17*01&)$.0:!133O3//! 

S"B!EK*F*I+5#!=7//048$9:)$5;#)3N?!@EK*F*I+K!V55YB5!+F!Z#4F'4,'+`4#+*FM:!+F!;4IB5!\+GGB,:!B'):!!#+#3&44')13*01&)/$*).$ 

2G'10HI$J&+13H$K#/#*(3"$-//'#/:!A*,#"!a*GG4F':!ABC!b*,D:!.8NO.13$ 

^4,,BGG:!;!4F'!Z4G*FB,:!_)!=.>38?!@VF5#4GGB'![45B!4F'!9*I$4#+J+G+#H%!VFF*W4#+*F:!3N?!@LF!#"B!c+KBF5+F-!*U!VFF*W4#+*F5M)!K*).$%&'()*+$&,$23&)&413/$.8:!N/0ON7/! 

R4#`:!\)!4F'!Z"4$+,*:!9)=.>38?!@>7?!!d8?!@TeP!+F!\4,DB#5!C+#"!AB#C*,D!E(#B,F4G+#+B5M:!-)0#()*01&)*+$%&'()*+$&,$-).'/0(1*+$5(6*)17*01&)$ 

.0:!N2.ON01! 

cB,FB,:;)!4F'!;)!S+,*GB!=7//0?:!S"B!EK*F*I+K5!*U!SBK"F*G*-H!Z"4,+F-%!L$BF!Z*Y,KB!4F'![BH*F'M:!a4,W4,'![Y5+FB55!ZK"**G!]*,D+F-! 



!"#$%&'()*%)+,&*$-)+&*'%&'().'('$/&0)1,2/%$-)3!++.14) 

5,-678)9((2'):)!;/*-?:) 

) 

*++,-./01 

+%""21"213(4"%45161 

I0'%)A0')H,%,;,-*(A)0$()$);/,"2&A),C)J2$-*AB)!"#$%")&0$/G'()$);/*&'),C);9)A0'),%-B)&,%(2H'/() 

K0,) L2B) C/,H) *A) $/') A0') ,%'() C,/) K0,H) $# %# !"# %# !# &" # ' # (# )"# # !) >6) M0'/'C,/') "'H$%") C,/) A0') 

H,%,;,-*(AN();/,"2&A)'J2$-()*#%#!"#%#!#&" #' #(#)"#+##9AN()/'#'%2'3.4)'J2$-())"#,*#%#!"#%#!#&" #' #(#)"-+6)M0') 

,;A*H$-);/*&')C,/)('--*%G)*A(),2A;2A)*()A02()G*#'%)LBO) 

) 

" .P" ; 9 ) Q )!)R)J9)R)#)S9 ), )) 

) 

,/))))))))))); 9 T QS9 T Q3)!)R)J9)R)#)S9 ), 4P)=))))))))))))))))))))))))))))3!?4) 

) 

M0')H,%,;,-*(AN();/,C*A)*()A0'/'C,/') 

$9 T) Q)[)3)!)R)J9)R)#)S9 ), 4P)=] )= )R)C9)3U9)R)S9 ), )



!"##$%&'()*+,*-./012,*3'#4"(5*67"%"8$'9*:"(*;B,>45>5!#">! 

$,*=+#!=,*:!#">!=+,5#!#>,:!D5+;B>!=,*:!#">!E*9,;*#!-4:>A!"#!+5!'>B,>45+;-!+;!ƒ+$%!8#!#">!*$#+:4?!ƒ+! 

#"+5! :95#! ! >(4B#?C! *==5>#! !>F9+?+,>45!#">!=+,:!+H5!,>@>;9>! 

=,*:!?+B>;5+;-!+5!9;4==>B#>'!;!4!5:4??!+;B,>45>!+;!ƒ+!=,*:!+#5! 

$,>@+*95!*$#+:4?!?>@>?!+;B,>45>5!#">!,>@>;9>!=,*:!#">!?+B>;5>!=>>5!!'>B,>45>!+;!,>@>;9>!=,*:!'+,>B#!54?>5!+5!?4,->,!5+;B>!"#!+5!?4,->,)!I"95!G+#"! 

#">!?4,->,!ƒ6!$,*=+#5!4,>!'>B,>45+;-!+;!+H5!?+B>;5>!=>>5!4#!#">!$,>@+*95?C!*$#+:4?!ƒ+%%!K+@>;!#">! 

B*;B4@+#C!*=!#">!$,*=+#!=9;B#+*;!+;!4!=+,:H5!*G;!?+B>;5>!=>>!#"+5!,>F9+,>5!4!?*G>,+;-!*=!ƒ+!+;!*,'>,!#*! 

,>4B"!>F9+?+;B>!#">!,>4B#+*;!=9;B#+*;!+5!'*G;G4,'!5?*$+;-)! 

! 

L>::4! 8.)! I">,>! >(+5#5! =M N ! =*,! #">! +;B9:;#! 59B"! #"4#! =*,! 4??! =M! O! =M N !!#">! *$#+:4?! 

,>5$*;5>!*=!#">!>;#,4;#!+5!#*!5#4C!*9#!*=!#">!:4,P>#!4;'!=*,!=M!Q!=M N! +#!+5!#*!>;#>,)!=M N !+5!#"95! 

#">!?4,->5#!?+B>;5>!=>>!=*,!#">!+;B9:;#!#"4#!'>#>,5!>;#,C)! 

! 

R,**=&%L>#!!'( ) !!#">!?+B>;5>!=>>!=*,!G"+B"!'*!+!/!+5!#">!5:4??>5#!?+B>;5>!=>>!#"4#!:4P>5!"*!,'*&!'( ) $!S!/)! 

T">;!'*!+!-%!"*!S!/A!4;'!'(!+!'( ) !#">;!!*!S!/!4;'!"!*!/!"ƒ*!+!/!U!P!Q!/)!I"95!,>'9B+;-!'*!G+??! 

*;?C!:4P>!!*!!;>-4#+@>!G"+?>!4!,!'*!+5!;*#!4;!4BB>$#4!?+B>;5>!=>>)!V*!>;#,C!+5!'>#>,,>'!4#! 

'( ) %! 

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 

I*!B4?B9?4#>!'( ) !G>!5>#!!"*!>F94?!#*!W>,*!4;'!'*!+!-!G+#"!+#)! 

T>!#">;!->#! 

! 

8U7DP!!!ƒXY!!!DP!U!(M! 0 !!!!=M N Y!!!"(M! * !S!/!!!!!!!!!!!!!!D8ZY! 

! 

I">,>=*,>A!=M N !S!D.!U!"Y!(M! !!!8!!!P!!!!!!!!!!!!!!!!!!!!!!!!!!!D81Y! 

0 

! 

X;#,C!+5!'>#>,,>'!4#!4;C!ƒ(!.!'( ) !!#**A!5+;B>!#">;!#">!5:4??>5#!'*!!#"4#!:4P>5!"*!,'*&!'($!S!/)!:95#!! 

;>-4#+@>!A5>>!D8ZY)![*,!'(!/!'( ) !#">!5:4??>5#!!' * !#"4#!:4P>5!"*!,'*&!'($!S!/)!+5!$*5+#+@>A!#"95!#">,>!'*! 

>(+5#! 4BB>$#4! ?+B>;5>! =>>5! #"4#! -+@>! #">! >;#,4;#! $*5+#+@>! $,*=+#5)! J>;B>! #">! >;#,4;#! G+??! 

>;#>,!4;'!5*!#">,>!+5!;*!>;#,C!'>#>,,>;B>)!I"95!'( ) !#0!+;'>>'!#">!?4,->5#!?+B>;5>!=>>!#"4#!#">! 

+;B9:;#!B4;!B"4,->!4;'!5#+??!'>#>,!>;#,C)! 

! 

L>::4!87&%%I">,>!>(+5#5!=M \ !Q!=M N !, 59B"!#"4#!+;!#">!+;#>,@4?![!=M N A!=M \ ]!!+#!+5!*$#+:4?!=*,!=+,:!X!#*! 

B"4,->! #">! ?4,->5#! 4BB>$#4! ?+B>;5>! =>>! =X!!#"4#! :4P>5! (X! S! /! 4;'! =*,! =M! Q! =M \ ! #">! *$#+:4?! 

,>5$*;5>!=*,!=+,:!X!+;'9B>5!+#!#*!$,*'9B>!4!$*5+#+@>!4:*9;#)! 

! 

R,**=&%[,*:!>F94#+*;!D8]Y!G>!P;*G!#"4#!#">!5#!,>5$*;5>!=*,!#">!>;#,4;#!#*!=M \ +5!-+@>;!;!=M!S!=M \ A!!= \ X!!:95#!!45!-+@>;!!(H5)!#">;!G">;!=+,:!M!B"*5>!'( 1 !4;'!=+,:!X!B"*5>!+#5!?+B>;5>! 

=>>!*$#+:4??CA!! +#5! ?+B>;5>! =>>! *$#+:4??CA! #">;! +;! #">! 

59F9>;#!*9#$9#!-4:>!=+,:!X!G*9?'!>;'!9$!5>??+;-!4!$*5+#+@>!4:*9;#)!M=!!"*G>@>,!=+,:!M! 

B"*5>!'(!;!+;!#">! 

59F9>;#!*9#$9#!-4:>!=+,:!X!G*9?'!!,>F9+,>'!#*!5>??!4!;>-4#+@>!4:*9;#)!V+;B>!#"4#!+5! 

! 

!"#$%"##$%&&'()'*+)*,-&./).0123&455,6)01)72.7)% .13! 

! 

165



!"#$%&'()*%)+,&*$-)+&*'%&'().'('$/&0)1,2/%$-)3!++.14) 

166 

5,-678)9((2'):)!;/*-?:) 

) 

%,A);,((*J-')D)K2(A)('A()*A()-*&'%(')C'')$A)A0')-'#'-)A0$A)K2(A)H$L'()*A(),2A;2A)*%)A0')%'MA)(A$G') 

N'/,6)O'%&') P0'%) A0'/') *() %,) '%A/B) J-,&L8) *%) A0') *%A'/#$-)[!"# $ %!"# & ]!!C*/H) D) &0$/G'() A0') -$/G'(A) 

$&&';A$J-')-*&'%(')C'')"'!A0$A)H$L'()('!)!*+!Q*/H)9)H$L'();/,C*A()$&&,/"*%G)A,)A0')H,%,;,-*(AR() 

;/,C*A)C2%&A*,%6)S0*()*()P0'/')A0')*%'T2$-*AB)M*)!)>)C,/)A0')'%A/$%A)*()J*%"*%G6) 

) 

Q,/)"#!,!"# & !A0')'%A/$%A)K2(A);-$B()*A()2%&,%(A/$*%'")J'(A)/'(;,%(')*%)A0')+A$&L'-J'/G)G$H'*+!"# & )*() 

A0') (H$--'(A) -*&'%(') C'') C,/) A0') *%&2HJ'%A) $J,#') P0*&0) A0') '%A/$%A) ;-$B() *A() 2%&,%(A/$*%'") 

,;A*H$-)C''()$%"),2A;2A)*%)'T2*-*J/*2H*+ 

+ 

,%""-+"-+.(/"%/0+1+ 

U');/,#')A0')A0',/'H)*%)$)C'P)(A';(6) 

) 

-.//0!12+!'3456!78!9:;!7"!?@!A!B!C!(# ; "!1!A!B=+! 

) 

U')C*%")A0')(H$--'(A)"'!A0$A)G*#'()2()('!?"'%!"# D C!V)>! 

Q/,H)3!W4)P')G'AX) 

) 

!)Y)=3L")ƒD4)")3L)Y)M9 > )")C9 Z 4)")#M9 , )V)>) 

) 

+2J(A*A2A*%G)A0')'M;/'((*,%)C,/)"# D !f C/,H)'T2$A*,%)3!=4!*%A,)A0')$J,#')G*#'()2(X) 

))))))))))))))))))))))))))))))))))))))))))))))))))) 

ƒD $) V)!)Y=L)")3?)Y)=#4M9 , [\) 

)))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))) 

ƒD $ !78!A0')-$/G'(A)$&&';A$J-')-*&'%(')C'')C,/)A0')'%A/$%A)P0'%)A0')*%&2HJ'%A)&0$/G'()"# D ) 

U0'%)ƒD $ %)>)8)#'!)$A)ƒD $ )*()%,%)>6! 

) 

]*#'%)A0')&,%&$#*AB),C)#'!)*%)ƒD)))H'$%()A0$A)*C)!ƒD %)>)8) 

$ 

A0'/') *() %,) $&&';A$J-') -*&'%(') C'') C,/) A0') '%A/$%A) A0$A) G*#'() *A) $) ;,(*A*#') ;/,C*A6)O'%&') '%A/B) *() 

J-,&L'")*C)!ƒD %)>)8)*6'6)*C) 

$ 

) 

!Y=L^)3?Y=#4M9 ,) 

) 

ƒD V) >) *() A0') -$/G'(A) $ ƒD ) C,/) P0*&0) '%A/B) *() J-,&L'") (*%&') *C) $ ƒD ) >) >) A0'%) A0'/') ",'() 'M*(A) $%) 

$ 

$&&';A$J-')-*&'%(')C'')C,/)A0')'%A/$%A)A0$A)G*#'()*A);,(*A*#');/,C*A6) 

! 

-.//0!1EFG;5!1AB="!?@!A!B!C!(# ; !"!1!A!=HB%!.3456!78!5.8457!I74J!4J.!73



!"##$%&'()*+,*-./012,*3'#4"(5*67"%"8$'9*:"(*;.=H#?!(9! * ! 

! 

IE##+F-!#*-D#"D,!#"+5!B*F'+#+*F!A+#"!#"4#!8*,!F*!JK*B.!=#?!(9 * "!@T! 

&!" . /"!!"#/#567#[8#9#:+# #7&5#9:$-#1"# ; #]# 

! 

! 

!" . !+5!'DB,D45+F-!+F!!!"##*4#!"#/#ƒ" ; !+8! 

.!=7#?!(9 * ! 

! 

["E5!8*,!7@&P!=!3# ?;@# 7@&P=



!"#$%&'()*%)+,&*$-)+&*'%&'().'('$/&0)1,2/%$-)3!++.14) 

) 

5,-678)9((2'):)!;/*-?:) 

ƒ! . "#+*%&')!! $ )*()&,%&$#'8)A0'),;A*H$-)-*&'%(')C'')C,/)A0')*%&2HJ'%A)H2(A)J')G*#'%)JB)A0'),%') 

A0$A)H$K*H*L'()!! $ "#)M$--)A0*()-*&'%(')ƒ! + "#+'AA*%G)!!! $ /")%!# N)>)C/,H)3!?>4)O')G'A) 

))))))))))))) 

ƒ9 + )N)[PQ)")I!)")3:)R)S=#4)K9 , ])T?I))))))))))))))))))))3!??4) 

) 

U*#'%),2/)"'C*%*A*,%),C#ƒ! . &#%!# #ƒ! . )*(#,;A*H$-)C,/)'()V)>6)) 

) 

W02()*C))#*#+,)VP-.#*+$/#'! 0 8)A0'%)A0')*%&2HJ'%A)&0$/G'(#ƒ! $ &#A0')'%A/$%A)&0$/G'()%(#G*#'%)JB)*A() 

/'$&A*,%)C2%&A*,%8)$%")J,A0)C*/H();/,"2&');,(*A*#')$H,2%A(6) 

) 

*%""+,"+,!"%"--.%$,/, 

X/,H) Y'HH$() ,+#$%") !S) O') *%C'/) A0$A) C,/) ,#*#)1+) $-.#*+$/# '!# 0 8) A0') *%&2HJ'%A) &0$/G'()*A() 

,;A*H$-)H,%,;,-B)-*&'%(')C''8)ƒ! 2 "##Z'%&')A0')(A$A'H'%A),C)A0')&,/,--$/B)*()A/2')*%)A0*()&$('6)X/,H) 

Y'HH$)!78)O0'%)=!TP)R))13#%#-.#*+$/#'!#%4,#*#)1+"#A0'),;A*H$-)%!#56#ƒ! . # "#[2A)*%)A0*()(&'%$/*,) 

H,%,;,-B);/,C*A()$/')*%&/'$(*%G)*%)A0')*%A'/#$-)[#%! 7 &#%! 8 ]#$()O')($O)*%)Y'HH$),9#6)W02()*A)H2(A) 

J')A0')&$(')A0$A)%!# 2 >ƒ! . )N)%! & ##"#X*%$--B8)O0'%)))#*#+,)VP-.#*+$/#'! 0 )&#A0'),;A*H$-)-*&'%(')C'')C,/) 

A0')*%&2HJ'%A)*()G*#'%)JB)ƒ! + )*%)'\2$A*,%)3!??46) 

) 

ƒ9 +) ")ƒ9 ] )N))")[=!)R)Q)R)?S)3?)R)=#4)K9 , ])T?I))6) 

W02()O0'%))#*#+,#:3-.#*+$/#'! , 8)ƒ9) ] >ƒ9 + )N)ƒ9 % ) 

, 

*%""+,"+,0(1"%12,3, 

9A) &$%) J') &0'&Q'") A0$A) *%) A0') &$('() O0'%) A0'/') *() 2%/'(A/*&A'") '%A/B) $%") O0'%) A0'/') *() 

/'(A/*&A'") '%A/B) O*A0,2A) H,%,;,-B) ,2A;2A8) %!#

2.3 VERTICAL COMPETITION AND OUTSOURCING 


!"#$%"&#'("&)*+(,%"&)*(-*.,/'"$//*&"0*12("(3'2/*4$/$&%25* 

"#$%&!'()*+!,$%-,""! 

./0123456!781285!95:5;05*&!GHHI+!","E-F%)J!(K81285*! ! 

6$%#'2&)*7(38$#'#'("*&"0*9,#/(,%2'":*'"*&*;,88)5CM=!N5:48717CM=!O:787;2:3=!HNPQNOR!R783/1>28C=!S/82:4=!T55:U:VML477B:7;! 

4)%5+*-%*6+$%".*+51-3% 

WL83!XB!T7>>28C52>278!L86!K/>37/L8:53! 28! ;L8/[L:>/2783! 7[! 

17:]-283!:56!0M!2887\L>2\5!3/??125551! L86! S2:! L3! 3/??1257! 

653]>7?! ;L8/[L:>/278-0L356! 17:]-283B! G86556=! >453! L8! 5\71\28C! 

?7A5/3!;5;052>278! L3! 

L665C2:L11M!0M!.755Z>! 

23! 58\23LC56! 28! A42:4=! L>! L8M! C2\58! ?728>! 28! >2;5! 7[! >45! 

278342?=! 07>4! >45! 0/M545! 3/??125278! 32;/1>L857/31MB! X5! 4L>! >45! 

?278! [7;58>3! 28! 2887\L>278! 0M! 

;5;0545!3/??1M!:4L28!23!>7!28:4522L1! 

72\5! :4L8851! ?7A545! 3/??1M! :4L28B! Q137! 28!L! 

32>/L>278!A4545!0/M545!0/M52\51M! ?/45! :278! 7[! L! 3/03>2>/>5! >5:48717CM! 

0M! 28\53>28C! 28! 2887\L>278B! N45! ?278! [745! 

0/M5423!:L35!A7/16!05!>7!512;28L>5!>45!>5:48717CM!17:]- 

28!L86!05:7;5!2865?58658>B! 

X5! :L8! 177]! L>! 47A! >42:L1!28>5C278+! 

$*!H>L?153!28!>45!7[[2:5!?3!286/3>23>5?=! 

42C4-:73>! \L1/5! :4L28B! N45M! 58\2327856! L! [/>/7;556! 0M! 3475828C! >45! :4L28! >7! :L?>/423! \L1/5! ?278! A2>4! L! ;7651! >4L>! 

3/??756! 2>3! L21! 3>74! 2>3! 7A8! :58>278! :58>557! 512;28L>5! 

;26615;58!L86!28:



! G8>5278L1!f7/3!>4L>!>45!:L!L3!L8!5ZL;?15! 

7[! i>45! 65:7;?732>278! 7[! >45! \52:L11M! 28>5C56! 0/328533! 

;7651j![45!785!>4L>!/356!>7!05!\55C56!L3!?L!7[! 

L!342[>!28!L!\L1/5!85>B!G8!3/:4!:2L8:53=![2545! 

:5CM! 7/C4>! >7! 05!28! ?1L:5! >7! 

65[586! >45! L0212>M! >7! L??5! \L1/5! 0M! 

583/4L>! >45! 3/??12545! ! >45! [2! >7! 7A8! L! L015! >7! ?/>! >45;351\53! 28! L! 

?732>278!>7!15\57!>45![2=! ?4L>! 2>! 37/45! ;2:3>L8628C!3/::533[/1!?!6532C8!0/>!L!623L3>51!28!>45! 

3/??1M! :4L28! [756! 0/>! :4L1158C56! 0M! 

Q6\L8:56!S2:2783!>4L>! 

7::/423! 286/3>58>! 65[2856! 0M! >423! 

/?3>7! 

05AL45!bG8>51!G83265c!3M8658628C! >423! L! >7! >45! /?3>2783B!9/45! $d)#3=! >45! ?2:5! 7[! 35:786! 37/! :L8! L:>! L3! L! 35:786! 37/4L>!37;5!3715! 

3/??125756c!>4527;545M! 4L6! 17:]56! 28! >452! 6532C83! >7! >4L>!7[!>45! 

3/??125/45! 286/3>45!5\71/>278!7[!;L!3>/>2\5! 3>5CM! [757! ?1M! A4215! 7[[52:L1! >7! >4L>! 7[! >45! 2887\L>28C! [2423! 

4/!:7/16!/35!2>3!C7! 

3/??7!L!1L4!L!\25A!>7!65\517?28C! 2>3! 

7A8!85Z>!C585278!?3B!^756! 

L3!L!35:786!37/54L8!4L1[!L!67e58!:7;?L8253!A578=!"##$*B! 

Q8! 5ZL;?15! :L8! 05! [7/86! 28! >45! 3/??1M! :4L28! 7[! 

L/>7;702153! (R1L7=! $dd$=! R4L?3B%=! )*B! Q/- 

>7;70215! ?278! 05C283! A2>4! 6532C8=! A42:4! :78323>3! 7[! 

>43B! N45![2!515;58>!23! >45!:78:5?>!2>351[=! 

>45545!6532C8!7[!>45! \542:15!:L8!05! /35[/11M! 62\2656! 

28>7!>45!;L:45!65\517?;58>!7[!>45!0L32:!:4L3323=! 

3/0-L335;01M! L86! :7;?7858>! 3?5:2[2:L>278*! L86! >45! ;2:45! 65\517?;58>=! >7! LC2783=! 7[! >45! 

\542:15l3!:783>2>/58>!:7;?7858>3*B!N45![2!>A7!515;58>3!7[! 

6532C8! >586! >7! 05! /865L]58! 0M! >45! :L2:/1L45!L335;015L]53!:4L!7278B! 

G8! >45! [L:5! 7[! 28>5835! :7;?5>2>278=! 47A5\545! :73>3! 

L337:2L>56! A2>4! 65\517?28C! 85A! \542:15! ?7>M?53! 4L\5! 

28:3! (>45! ;2:7! 

C170L1! 3>L86LL]5! [/11! M! [745! 6532C8! 7[! 3/0- 

L335;01253!L86![745!:77278!7[!>45!35:786!L86!>4225!;L8/[L:>/!:78>5!>7!>45!?B! 

N45! 3>278! 7[! >4535! 3/0-L335;015278B! I7! 328C15! 

3/??125M?53! 7[! 3/0-L335;01MB! ^745! A77;7>2\5! 5_/2?;58>! 

;L8/[L:>/3! 5[[73! 78! 3>L55;3=! 3?L28C!L86!A2863:278!23!>4L>!>45!3/??1M!7[! 

?L2:/1L5;3! 4L3! 05:7;5! :78:58>56! 

L;78C3>!D/3>!L!4L86[/1!7[!;L8/[L:>/5:48717C2:L1!:7;?5>2>278!23!/356!>7!5ZL;285!>45!:7;?5>2>278! 

05>A558!>45!0/M545!3/??1252CL>278!7[!>423!>M?5! 

7[! 0/M5278! A7/16! 15L6! >7! L! 05>>5L8628C! 7[! >45! 6M8L;2:3! 7[! :711L07278! L;78C! 

3/??1M!:4L28!?L85528C! L3?5:>! 7[! >45! ?45! [L:>! >4L>! >45! 

3/??125!>L]5!28>7!L::7/8>!>45!2845!28:58>2\53![745! 

0/M57! 65\517?! L! b0L:]3>7?c! >5:48717CM! A42:4! :L8! 05! 

3/03>2>/>56![745!3/??1256!:7;?7858>B!N45!3/??1254!>45! 

]87A156C5! 7[! >423! 28>58>! 7[! >45! 0/M53! 3/:4! >4L>! >45! 

?3! L278! 7[! 3/03>2>/>5! 

>5:48717CM! 0M! >45! 0/M52;5! A458! 3/:4! 2887\L>278! 

;L>5!:L8!05!L[[5:>56!0M!Pk9!5[[73B! 

N45! ;7651! [7278! 23! 56! >7! L! :L35! 28! >45! 

5:787;2:3! 7[! 5Z4L/3>2015!



,$F! WL83!XB!T7>>28C52>278!L86!K/>37/4L>!23!355]28C!>7! 

28\58>! L! 3/03>2>/>5! >5:48717CMB! N45M! L66278! 28! >45! 5:787;2:3! 7[! 5Z4L/3>2015! 5:48717C253! >7! 3/03>2>/>5! [75! 8L>/56! >7! 623:7\57?! >5:48717C253! 87>! 781M! >7! 

7\545!?4L>!!L137! 

>7! 5:48717C2:L1! 

:7;?5>2>278! 23! ?56! 28! H5:>278! "B! G8! H5:>278! ,! >45! 

62[[52L1! CL;5! 23! L8L1Me56! L86! >45! SL1"*+)#% 

N45! 87>L>2783! L86! 3M;0713! /356! 28! >45! ;7651! L![745!0/M55!>2;5!475C253B! 

>+!Q8!283>L8>!7[!>2;5!28!>45!6M8L;2:!CL;5!35>/?B! 

/ a (>*+!a/M5278!5[[7!(Pk9!5Z?5862>/2;5*B! 

/ H (>*+! H/??125/!>7!05!28:/!>7!05!28:/!28:/4L>! A7/16! 05! L! ?5! 

3/03>2>/>5![745!3/??125278B! 

!+!L8>!7[!>2;5!L>!A42:4!>45!15\51!7[!2887\L>278! 

(04278*!23!*+!>45!?M!>4L>!>45!0/M28C![245!2887\L>278!?!0M!>2;5!>=!>4L>!23!.*!p!^(>*B! 

] #$ +![/8:>278!45!3A2>:428C!5B! 

Z(>*+!3>L>5!6M8L;2:3!5Z>!7[!0/M54L>!23!0L356!78! 

2887\L>278! L86! _/L12>M! ?L5B! Q! 785-! >7-! 785! 

:745!0/M57!>45![28L1!:783/;5!15L3>!785!:7;?7858>!23! 

! :7;?5>2>278! 28! >45! 

[28L1!:783/;5B! 

G>! 23! L33556! >4L>! L! 785! q! >7! -! 785! :745!L8L1M323!28!>A7!623>28:>! ALM3B!^2=! 

>423!L117A3!5Z?12:2>!5ZL;28L>278!7[!65?58658:5!05>A558!>45! 

>A7! :711L0728C! ?L8545! 5\71/>278! 7[! ?7A5/278B! H5:786=! >423! 15863! L8L1M>2:L1! 

32;?12[2:L>278! [745! ?! [745! 

:73>!7[!?278!23!: H (/ H *!m!#!?5B!H/??1255!7[! 

?278! 23! L! [/8:>278! 7[! ?278! 5[[7! / H !L86!23! 

L33/;56!>7!05!"/ H B!^72:L1!32;?12:2>M!"!23!L33/;56!>7! 

05!5_/L1!>7!$B!N45!5!7[!?278!23!65[2856!L3=!/ H +!(#=#*! 

$! [#=! / H ]! N45! 45! 3>L>5! \L45! :78>5Z>! 7[! >45! ;7651! 3?5:2[2:L>278! L86! >45! 

/8652783=!>45!2:L1! 

[745!0/M545!3/??125278!7[!>45!3/??12545!5!7[!:4L8C5!7[!*!p!-!"!/ H !(Z(>**!p!%!/ H !(Z(>**&!Z(#*!p!Z 7B! 

Q3! >45! ?278! 5! 28:45! 278! 5! 7[! >45! 

3/??1257!:5!L!3/03>2>/>5!>5:48717CM! 

A42:4! 23! L33/;56! >7! 58L015! ?278! 7[! >45! ?! 

:7;?7858>!L>!L!/82>!:73>!: a !r!: H B!G8!>45!5\58>!7[!L!3/::533[/1! 

:278!7[!3/:4!L!>5:48717CM!>45B! N45! 28\53>;58>! ?712:M! 28! 2887\L>278! 

:47358! 0M! a! 23! 6587>56! 0M! / a +! (#=! #*! $! s#=! #*B! N423! 

28\53>;58>! 45! 0/M5278! 7[! >45! *B! aM! 28\53>28C! 28! 

2887\L>278=!>45!0/M545!>2;5!?L>4![745!0/M5*!23!C2\58!0M=! 

] • (>*!p!/ a !(Z(>**=!](#*!p!#B! 

N45!7/>:7;5!7[!2887\L>278!28\53>;58>3!23!:7832657!05! 

/8:5L28B! N45! >45! Pk9! 28\53>;58>3! 

678l>!?545!6L>5!7[!3/::533!7[!2887\L>278B!Q! 

8L>/!ALM!>7!278! 

23!>7!3/??735!>4L>!>45!283>L8>L857/3!?M!7[!>45!0/M528C=! :7862>278L1! 78! 87>! 4L\28C! 37! [L278! 7[! 2>3! :/!23!L33/;56!>7! 

>L]5! L8! 5Z?7858>2L1! [7! 7[! >45! 

L::/;/1L>56!]87A156C5!15\51=!](>*B! 

^(>*!@![$!q!^(>*]!p!/ a !(Z(>**!!!!!!!!!!!!!!!!!!!!!!!($*! 

A45*!23!>45!?M!>4L>!>45!0/M545!2887\L>278!?!0M!>2;5!>&!>4L>!23!.*!p!^(>*B! 

`5>! u(/ H *! p! s?(/ H *! q: H (/ H *v! w/ H ! 05! >45! 3/??125L8>L857/3! ?! [/8:>278B! N423! 283>L8>L857/3! ?! 

[/8:>278! L:>3! L3! >45! 70D5:>2\5! [/8:>278L1! [745! 3/??12545!?5Z>=!>45!0/M5!:7;?5>2>278B!N45! 

0/M53! 28! 2887\L>278! >7! 05:7;5! 2865?58658>! 7[! >45! 

171



! G8>5278L1!f7/! 7[! ?278! 5[[7! : H (/ H *! 23! :78>28/7/31M! 

62[[52L015!78!s#=( / H *+!: H (#*!p!#!78!s#=( / H *!L86!:l H (M*!23!L! 

:783>L8>B! N45! :73>! 7[! 2887\L>278! 5[[7! : n (/ a *! 23! >A2:5! 

:78>28/7/31M!62[[52L015!78!s#=!#*&!: n !(#*!p!:) n !(#*!!p!#&! 

:+ n !m!#!78!s#=!#*&!L86!:) n ! (M*! $! #! L3!M!$! #B! N423! 

L33/;?>278! 3/CC53>3! >4L>! :73>3! 28:!L137!3/CC53>3!>4L>!e52;L1!L86![745!3>5CM!/ a! A5!:L8!:78[285!7/5! 

>7!28>5423!;L!35>>28C=!2>!23!L33/;56!>4L>!>45!?!58D7M3! 

L! \5278! 

5! 7[! >45! 3/??12545! _/L8>2>M! 65;L8656! ?5! 

>2;5!28!>45!;LB!N45!28\5278!74545!?278!7[!/82>3!?2;5B! 

?(/ H (Z(>***pL%0"/ H (>*pLq0/ H (>*&!Lm: H m#=!0m#!LL8>3!("*! 

N45!\L!7[!?278!23!L33/;56!>7!05!1285L278!7[!?278!5[[7!/ H B! 

: H !=!: $ ![/ H]=!: $ !m!#!23!L!?M!:783>L8>!!!!!!!!!!!(,*! 

!"#$%&'()*7("(8()9*:%(0,2#'("*&"0*:%'2$6* 

N45! 3>5C253! / H (B*! L86! / a (B*! :783>2>/>5! L8! 5_/212045! 

?278!5[[7!/ H (B*!;LZ2;2e53!>45!3/??12545! 2887\L>278! 5[[7! / a (B*! ;LZ2;2e53! >45! 0/M55C253B!N45![7117A28C!L33/;?>2783!L45! ?! 

[/8:>278!L86!>45!:73>!7[!5[[7![/8:>278B! 

(2*! N45! 28\5278! ?(/ H *! 23! >A2:5! 

:78>28/7/31M!62[[52L015!78!s#=( / H ]+! 

A45!7[!5[[7![/8:>278![745!0/M57!05! 

3>1M!:78\5Z!L86!28:278!5[[7!/ a B! 

: n !=!: " ![/ a] " = ! : " !m!#!23!L!?M!:783>L8>!!!!!!!!!!('*! 

I7>5!>4L>!>423!?45!3L;5!3>L>5!6M8L;2:3![745! 

0/M545!3/??1252\5![/8:>278L1!7[!3/??12545!5Z?5:>56!?!/8>21!>45!L8>!7[!>2;5!!=! 

f H (Z*!=!O /a(B*! # ! !5 - !s?(/ H (Z(>***!q: H (/ H *vw/ H !Z(>**]6>! 

p!O /a(B*! # ! !5 - ![L!-!0/ H (Z(>**!%!: $ (/ H *(Z(>**/ H (!Z(>**]!6>!!!(%* ! 

I7>5!>4L>!>45!3/??125![7!m!!!(328:5! 

65;L86!\L823453*B!N45!70D5:>2\5![/8:>278L1!7[![245! 

5Z?5:>56!CL28![278!5[[73=!C2\58!0M! 

f a (Z=!/ a *!=!O /a(B* !! # ! !5 - s&(/ H (Z(>***!q!: n (/ a *(Z(>***]6>!+!! ! # !5 - !' a !(/ a *6>! 

p!O /a(B*! # ! !5 - !s&(/ H (Z(>***!q: " (/ a *(Z(>***] " 6>!+!! ! # !5 - !' a !(/ a *6>!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!()*! 

&(/ H (Z(>***!=!! # /H !s?(M*6M!q/ H! (Z(>**?(/ H! *!=!! # /H ![L!-!0M]6M!%/ H (Z(>**![L!-!0/ H (Z(>**]!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!(F*! 

23!>45!3/



,$d! WL83!XB!T7>>28C52>278!L86!K/>37/***!q: H (/ H *v/ H Z(>**]6>! 

=!O /a(B*! # ! !5 - ![L!-!0/ H (Z(>**!%!: $ (/ H *(Z(>**]/ H (!Z(>**6>! 

3/0D5:>!>7!Z • (>*!p!%!/ H !(Z(>**&!Z(#*!p!Z 7 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!(E*! 

f a (Z=/ a *!=!O /a(B*! # ! !5 - !s&(/ H (Z(>***q: n (/ a *(Z(>***]6>!+!! ! # !5 - !' a !(/ a *6>! 

N45!62[[52L1!CL;5!65[2856!0M!(E*!L86!(d*!05178C3!>7!>45! 

:1L33! 7[! ?25:5A235! 65>52:! CL;53! (97:]85! L1B=! 

"###*+! N45! CL;5! 4L3! >A7! ;7653! 781M+! ;765! #! 23! L:>2\5! 

05[7278!L86!;765!$!05:7;53! 

L:>2\5!2[![245!3/03>2>/>5!>5:48717CMB! 

N4/3=! >45! ;73>! 785! 3A2>:4! 7[! ;765B! N45! 

3A2>:428C!>2;5!23!>45!278!7[!!!23!^=!L3!C2\58!28!5_/L>278!($*B! 

C'%7#"1:$+$%)A%*6-%D)2-1% 

N45!;7651!23!L8!5Z>583278!7[!L!?>28C5***] " 6>!+!! ! # !5 - !' a !(/ a *6>! 

3/0D5:>!>7!Z • (>*!p!%!/ H !(Z(>**&!Z(#*!p!Z 7 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!(d*! 

3>L>278L278L! n H! (4=!Z*!23!L! 

;L??28C!n H +!S!,(#=!#*!$!s#=!( / H v=!A4545! 

35>! 7[! ;7653B! a/M5L>278L278! L??2783! 7[! >45! 3/??1252\51M![745!65[282>278!L86!>4577! ?25:5A235! 65>52:! 62[[52L1! CL;53! 

?45!Q??5862Z=!>4535!\L1/5![/8:>2783!L7!05!07/8656!L86!:78>28/7/31M!62[[52L015!L86![74L>! 3L>23[M! >45! WL;21>78-fL:702-a511;L8! (Wfa*! 

5_/L>2783B! 

3! 7[! / H-! s#=! ( / H v!L86!/ a-! s#=#*! >4L>! 

;LZ2;2e5!>45!2\5!\L1/5![/8:>2783!7[!>45!3/??12545!0/M5278!($#*!L86!($"*!L56!L3!0 H (Z*!L86! 



! G8>5278L1!f7/45! :73>! 7[! 5[[7! [/8:>278! [2783!("*!L86!5_/L>278!('*=!2\51MB! 

T H (e*!=!;LZ!{L!%!0/ H !%!: $ !/ H !]!/ H !%!/ H !e.#!1!/ H!1!( / H}!!($d*! 

T a (e*!=!;LZ!{e/ a !%!: " (/ a * " .#!1!/ a !}!!!!!!!!!!!!!("#*! 

n H (#=Z**!=!L5!>45!?732>2\5!5_/212045!3/??125458!! 

! "# $ 

%&"'& 

"&#()*+$' , - . *%/ 0 $)' . -1&1% . +' !!!!!!!!!!!!!!!!!!!("'*! 

! 2# $ 134 & 51$1%6 7 111&1 "&#()*+$' ,- . *%/ 0 $)' . -1&1% . + 

!!!!!!!!("%*! 

"&*+$' , - 

;5$(%$3*=6%G[!n H 2 !L86!n a 2 !6587>5!>45!?732>2\5!5_/21202783!7[!>45!3/??125458! 

8 9 # $ :;! < 9 *=>?-1 

3*@$4 A -1 

8 B # *=> ?- $ 6C 1&16 C *=>?- 

$ 1: 

3*@$4 A - !!!!!!!!!!!!!!!!!!!!!!!!!!!(")*! 

34 3 

$ "D5*@E4 A -3 F6 C E54 3 G;: 3@ 

3"3*@$4 A -"4 3 

!!!!!!!!!!("F*! 

;5$(%$3*>6!Q>!5_/212045!!>2;5!>! 

:L8!05!5Z?$$+)#% 

?@/$%A&#'("* ! 5_/212045! 3/??125278!5B! 

N45! 7035278! :L8! 05! 5Z?1L2856! 0M! 5Z?1745! 

;787?71M!?278!5!7[!>45!3/??12545!C2\58!35>>28CB! 

N45!;787?71M!?!7[!>45!3/??12545![2!77!>45!;787?71M!?278!5!C2\58!0M!L!/ H; !=!L/!"(0+: $ *B! 

N45!3/??1257!05!L!;787?71MB!G8!>45!C2\58! 35>/?! >45! 3/??125278!5!>7!CL28!L3! 

;/:4!7[!;787?71M!?3!L3!?7332015!05[7!7[!>45!;LB!Q8!211/3>278!7[!3/:4!L!;787?71M!?278!23! 

?



,"$! WL83!XB!T7>>28C52>278!L86!K/>37/45!3/??1253!L!85CL>2\5!5[[5:>!78!>45! 

0/M5278!5B! 

N45! 7035278! 3>L>56! 28! "(2*! 23! 28>/2>2\5! L86! :L8! 05! 

703545! 5Z?5! 0M! >45! 0/M556B! X2>4! 37;5! L1C502783! >45! 

5_/2120278!5!7[!>45!0/M545! L07\5! 5Z?! :L8! 05! 87>56! >4L>! >45! 

3/??1252\5!5[[5:>! 

78! >45! 0/M5278! 5[[73B! K035278! $! L86! "! 

:4L545! 8L>/45!C2\58!35>>28C!>453!87! 

;7>2\L>278![745!>A7!?1LM57!:711L075!L86!5L:4!L:>3!28! 

2>3!053>!28>5B!N45!3/??12545!;787?71M! 

?7A545!0/M528/53!28\53>;58>3!28!2887\L>278!>7! 

:5!3/03>2>/>5!>5:48717CMB! 

W7A5\52783B! I7>5=! >4L>! 28! 5_/L>278! ("d*! A5! 4L>!A2>4!L!67A8AL278B!N45!28\5



! G8>5278L1!f7/7! 

Q86! 

( 63 % 

7 8 "9 8 

":#; $ 8 +%? @ $>28C52>278j=! 

!"#$%"&#'("&)* +(,%"&)* (-* .,/'"$//* &"0! 12("(3'2/* 4$/$&%25! 

'("*=!)F-F$B! 

sdv! T7>>28C5

2.4 SUPPLY-CHAIN COOPETITION 


!"#$%"&#'("&)*+(,%"&)*(-*.,/'"$//*&"0*12("(3'2/*4$/$&%25* 

"#$%&!'(")*!+,-,$! 

./0123456!781285!9:;!"#$%!(4??@*AABBBC3



+G! V:83!WC!S7??28D5;*! ! I/@@1M-R4:28!R77@5?2?278! ! 

! 

:8:1M_53! ?45! /`!:=!28Y!{?!>!#*!g? /I>/`!=!M!}CN45!Y2;3?!42??28D!?2E5!?7!?45! 

@728?!M!/865;!?45!3@5/`!C)! 

@(_)! *b@;2



! H8?5;8:?278:1!e7/;8:1!7Y!`/328533!:86!O@CG#]! :86! /?212_28D! ?45! ;53/1?3! 

Y;7E!(')!:86!(%)!D2Z53 ! 

6!g? /I>/`!=!g? /I>/`[E(/ I (?)>!(/`(?))6?!+!/ I (?)# I !6W I (?)!!!/`(?)#`! 

6W`(?)]!!!!!!!!!!!!(+)! 

B45;5!?45!Y/8



,#! V:83!WC!S7??28D5;*! ! I/@@1M-R4:28!R77@5?2?278! ! 

! 

c / (_)!=!3/@ /I !28Y /`!O _ !(5 !,' 0 /I>/`)!=!28Y /`!3/@ /I !O _ !(5 !,' 0 /I>/`)!>!Y7;! 

_!! !f![%`"!($!!!- " )!±!.p!]!/!"![%`"!($!o!- " !!!"!&!-)!o!"!,!($!!!& " )]!($")! 

:86!?45!:337! 

/`/(_)!=![%`/#`]![($!!&-!)0 + !!!$]!/!(!($!!!& " )(0 + ) " !!!$)!]!($=)! 

(")!R:35!`! 

N457;5E!"C!HY!?45!65D;55!7Y!E:;[5?!:6Z:8?:D5!7Y!:!3/@@125;! 

7Z5;!:!0/M5;!-!3:?23Y253!?45!!-!>!- + !B45;5!- ! !>- + ! 

:;5!:3!65Y2856!28!(a)>!?458!?45!Z:1/5!7Y!?45!623$$+)#% 

W5!



! H8?5;8:?278:1!e7/;8:1!7Y!`/328533!:86!O!:!3/@@125;!B7/16!D:28!0M!E:[28D! 

28Z53?E58?3! 28! 2887Z:?278! ?4:?! :;5! 2865@58658?! 7Y! ?45! 

28Z53?E58?3!E:65!0M!?45!0/M5;C!c/;?45;E7;5>!:!3/@@125;!B7/16! 

:137! D:28! 0M! E:[28D! 28Z53?E58?3! 28! 2887Z:?2783! ?4:?! :;5! 

@7332015!3/03?2?/?53!7Y!?45!0/M5;r3!@;76/



182

2.5 A TALE ON CHINESE INNOVATION AND COMPETITION 


A Tale on Chinese Innovation and Competition: A Case Study 

of the Chinese Telecommunications Equipment Industry 


STRATEC Munich Germany 

www. stratec-con.net 

Extended Abstract. 

We examine the four largest Chinese suppliers of stored program control 

(SPC) switches: Datang Telecom Technology (DTT), the Great Dragon 

Information Technology (GDT), Huawei Technologies (Huawei), and 

Zhonxing Telecommunication Equipment (ZTE). The tale shows how 

these four Chinese suppliers competed against the two largest SPC switch 

manufacturers that had foreign joint venture partners :Shanghai Bell an 

Alcatel joint venture, and Beijing International Switching System (BISC), 

a Siemens joint venture , and further how they competed globally against 

well established suppliers of telecommunications equipment (TE)such as 

Alcatel, Cisco, Ericsson, Fuijitsu, NEC, Nokia, Nortel, and Siemens. 

First, the fast growth in the period after 1985 of two suppliers, Huawei 

and ZTE, requires appropriate attention. Their success threatens well 

established incumbents’ plans to dominate the global and China markets 

for TE. ZTE started operations in 1985, and Huawei in 1988. In slightly 

over ten years, Huawei became the number one supplier of TE in China. 

In 1998. Huawei’s annual revenues exceeded those of the top two TE 

suppliers that had foreign joint venture partners: Shanghai Bell and BISC, 

and since then had the gap growing. Still, in the mid eighties, the 

possibility that any of the four Chinese suppliers could pose a serious 

threat to established global suppliers seemed very improbable. Today, 

Chinese suppliers compete aggressively against these well established 

global suppliers and their joint ventures in China. 

For example, reputable market share tracking firms ranked Huawei the 

number one supplier in the global market for new extended switching 

equipment in 2003, the number one supplier in the global market for new 

generation networks in 2004, the number two supplier in the global 

market for digital subscriber line (DSL)access multiplexers in 2003, the 

number three supplier in the global market for long distance wavelength 

division multiplexers, and the number four supplier in the global market 

for optical transmission. Huawei and ZTE have been ranked the number 

183



three and eight suppliers in the global market for integrated access 

networks. As the Wall Street Journal reported repeatedly in 2003 to 2005, 

executives of North American and European vendors of TE have become 

increasingly concerned about the head-to-head competition from Chinese 

suppliers. 

On the basis of previous models of competitive racing (<strong>Gottinger</strong>, 2006, 

2009) we trace empirically the various stages of competitive strength in 

this strategic industry. 

In the first stage we observe that the four startups targeted the basic need 

for infrastructural development in telecommunications in Western and 

rural China in supplying low cost telecommunications gear to those areas 

which were less lucrative for foreign vendors and joint ventures and were 

heavily encouraged by the Chinese national government in a sort of 

infant industry protection. This kind of asymmetic competition separated 

the startups from the established players in the Chinese market, call it the 

separation stage. 

In the course of this stage the’four horsemen’ underwent technological 

learning either through indigenous innovation or imitation of some sort, 

even industrial espionage in China,therefore gaining competitive strength 

and competing against foreigners on large scale projects in the Chinese 

market. This is termed the convergence stage. When asymmetric 

competition turns symmetric we observe competitive convergence , in 

which each technology’s development is directed at expanding its appeal 

not only in its own home market but in its rival’s as well. 

While the Chinese companies with the implicit support of the Chinese 

government continued to gain market share against foreign competitors 

and as their technological learning adavanced product quality at lower 

cost they expanded in actively seeking to bid successfully for 

telecommunications projects in developing and emerging economies 

where they gained further strength by competing on given product quality 

and lower prices. This is the globalization stage. 

However, whether Chinese companies keep on growing outside their 

home market will largely depend on whether they turn into genuine 

sustainable innovation leaders rather than followers. So far they have not 

gained a notable footage in advanced markets for smart networks. 

184



Disruptive Innovation 

If we consider the competitive positioning of high technology firms as a 

technology race in which falling behind, getting ahead and catching-up in 

industry leadership is the name of the game. We may come across 

specific situations that would be connected with ’disruptive innovations’ 

that could lead to a dramatic paradigm shift of that race. The management 

of disruptive technologies originated with the substantial works by C. 

Christensen et al.(2004) and C. Christensen (2005). 

Disruptive innovations introduce a new kind of product or service that is 

actually worse initially, as judged by the performance metrics that main 

stream customers value. In terms of the Chinese market where firms 

came up with simpler appropriate performance level of network gear at a 

much lower price for use in rural China those could be identified as 

disruptive. They also contributed to a market separation that 

distinguished two different tiers of market segmentation applying to rural 

and urbanized areas. In an industrial organization context incumbents 

have a high probability of beating entrant attackers when the competition 

is about incremental innovations, but almost always lose to attackers 

armed with disruptive innovations (Christensen et al, 2004).. At the 

bottom line a cumulation of incremental innovations on a particular 

design may end up in a disruptive innovation as will on the upper end 

technologically radical innovations . This may lead to market situations 

and revenue streams of ’winner-take-all’ or ’winner-take-most’. 

We may categorize two aspects of disruptive innovations that relate to 

(1) new markets and (2) quality characteristics. 

(1) Products offered are too expensive, too complicated, too difficult too 

maintain. 

(2) Product quality characteristics are appropriate,’ good enough’ with 

significantly lower pricing. New entrants compete profitably , learn and 

invest in R&D while pricing at deep discounts. 

Competitive Phases 

In adapting Christensen’s theory of innovation to the Chinese telecom 

industry we may distinguish clearly competitive phases such as 

separation( ’isolation’ as Christensen’s term), convergence and 

disruption. These phases seem to be more technologically motivated 

through innovation but may be supported by strategic direction or other 

factors. 

In a competitive separation (isolation) regime technologies do not interact 

(in terms of being replaceable) in the course of their development. 

185



SUMMARIZING REMARKS 

This case study is of interest as a general attempt in identifying Chinese 

efforts on innovation at the interface of culture, politics, technology and 

industry. The subject is so complex that it is unlikely to find a consistent 

interrelated interpretation of innovation with 'Chinese Characteristics'. 

(The well known Japanese growth economist Michio Morishima in the 

1980s did this for Japan (CUP) and ended up with a taxonomy which yet 

failed to provide a consistent interpretation of Japanese innovation) In my 

brief comments I limit myself on a case study based Chinese high 

technology industry 

To substantiate some of the claims on industrial innovation one needs a 

set of industry specific case studies such as the one submitted (A Tale of 

Innovation and Competition in the Chinese Telecom Industry) . A few 

remarks are in order. 

Chinese advanced technology and science based industries tend to be 

biased toward" product innovation through commercialization" --- quick 

market access and then incrementally improving product performance 

(quality) in subsequent releases.Indigeneous innovation could also 

involve 'rapid prototyping' in intentionally or inadvertently bypassing 

IPRs in favour of quick market introduction. Indigeneous innovation 

could also be a camouflage word for nurturing Chinese startups (where 

state directed subsidization acts as the equivalence of private sector 

venture capital or angel investment), and shielding them from foreign 

competition (at least in the domestic market for a while) in support of 

'infant industry' protection and expansion. This is coupled with state 

sponsored private entrepreneurship with encouragement of 

Schumpeterian entrepreneurship as examples in the Chinese network 

equipment industry clearly show. Competition in the 

Chinese domestic market among indigeneous firms could be fierce and 

adopts characteristics of a 'technology race' but since the domestic 

market is so large and with rising incomes 

demand growing in leaps there sometimes is a lack of incentives to go for 

global markets (with some notable exceptions outlined by industrial and 

development policies). 

Chinese innovation resembles a 'learning-by-doing' (Arrow) effect as the 

Japanese have done over several decades whereas the Chinese have 

accelerated this process by choice and through competition. Indigeneous 

innovation through domestic competition may limit market risk and 

therefore protect the company against total financial failure (bankruptcy). 

By industrial policy targeted companies can experiment on product 

development without severe downside risk. You don't bet the company if 

the product fails in the marketplace, therefore you can come up with 

186



incremental changes, whether true innovation or not, and let the market 

decide in the Chinese seller's market without fearing drastic financial 

penalties. No Western companies can take these (financial) risks in their 

domestic markets besides being blocked by Chinese entry 

barriers(reinforced by Chinese regulation) 

Christensen,C., Anthony,S.D. and Roth,E.A. (2004),’Seeing what’s next: 

using the theories of innovation to predict industry change’, 

Boston:Harvard Business School Press 

Christensen,C. (2005), The Innovator’s Dilemma, New York, 

HarperCollins 

<strong>Gottinger</strong>,H.W.(2006), Innovation, Technology and Hypercompetition, 

Routledge, London 

---- (2009), Strategic Economics of Network Industries, NovaScience, 

New York 

--------------------------------------------------- 

187



188

2.6 CHINESE INNOVATION AND COMPETITION 


! 

!"#$%"&#'("&)*+,-'"$--*&".*/0("(1'0-*2$-$&%03* 

!"#$%&!'!()&!!*$!& & 

+,-./0123&45./52&6,52&#"7&!"#$&'1889)::;;;>?4,9O!DP-/5>4*2Q



& T582?5M8/45M.&^,0/5200&M53&I=454@/=0&G202M?=1&!"#$%&!'!()&!!*$!& !$& 

& 

@,.8/9.2\2?0& /5& !""$7& 812& X,@-2?& F1?22& 0,99./2?& /5& 812& U4?2/>5& =4@928/84?0& M53& M0& 812/?& 82=154.4>/=M.& .2M?5/5>& 

>.4-M.& @M?Z28& U4?& .45>& 3/08M5=2& ;MQ2.25>81& 3/Q/0/45& M3QM5=23& 9?43,=8& V,M./8N& M8& .4;2?& =408& 812N& 2\9M5323& /5& 

@,.8/9.2\2?07& M53& 812& X,@-2?& b4,?& 0,99./2?& /5& 812& >.4-M.& M=8/Q2.N&022Z/5>&84&-/3&0,==200U,..N&U4?&82.2=4@@,5/=M8/450& 

@M?Z28&U4?&498/=M.&8?M50@/00/45&'I=454@/087&E298/5>&2=454@/20&;12?2&812N& 

`FI[0& 32Q2.49@258& 4Q2?M..& ;M0& ?2.M8/Q2.N& @4?2& @43208& >M/523& U,?812?& 08?25>81& -N& =4@928/5>& 45& >/Q25& 9?43,=8& 

'I=454@/087&d=8& 

I/>18& 0,99./2?0& /5& 812& >.4-M.& @M?Z28& U4?& /582>?M823& M==200& 0,08M/5M-.N&4,80/32&812/?& 14@2& @M?Z28& ;/..&.M?>2.N&329253& 

528;4?Z025,/52&/554QM8/45&.2M32?0&?M812?& 

!"##&R,M;2/&;M0&./0823&X,@-2?&d52&M53&`FI&M@45>&812& 81M5&U4..4;2?0M/523&M&548M-.2&U448M>2& 

849&F25& 812/?& @M?Z28& 

9M?8/=,.M?7& 1M0& -225& 0/5>.23& 4,8& M0& 812& 3M?Z& 14?02& /5& 812& 940/8/45&/5&M&09M8/M.&i-4M?3&>M@2[N[&7&/8&=4,.3&-2& 

'W2?N& ?2=258.N7& -2=M,02& 4U& M..2>23& 02=,?/8N& =45=2?50& 7& @M8=123& M53& 8?M=23& 84& E=488& ^44?@M5& '#YBY(& 45& @/./8M?N& 

R,M;2/& 1M0& -225& -M??23& U?4@& -/33/5>& 45& 9,-./=& 08?M82>N& M53& /5Q/820& M& @M8=1/5>& /5& 82=154.4>N& ?M=20& M53& 

/5U?M08?,=8,?M.& 9?4A2=80& /5& H,08?M./M& M53& U?4@& =4@9M5N& -,0/5200&;M?0&'L488/5>2?7&!""B7&!""Y(N[&'XMZM/&M53&FM5MZM7& 

#YY!%&L488/5>2?7&!""B(& & 

!"#"(& -,8& @4?2& 092=/U/=M..N& 812N& -481& M?2& 2@-23323& /5& 

T5&Q/2;&4U&32Q2.49@258&4U&3N5M@/=&=4@928/8/45&;2&8?M=2& S2/[=1/20&4U&=4@928/8/Q2&08?25>81&/5&81/0& 

08?M82>/=&/53,08?N2& ;2& 4-02?Q2& 81M8& 812& 81?22& 08M?8,90& 

?'%9-:%@)#5-A*$%)B%@)0A-*+*+)#%B).% 

8M?>2823& 812& -M0/=& 5223& U4?& /5U?M08?,=8,?M.& 32Q2.49@258& /5& C-*;).D%=#2>$*.+-$% 

82.2=4@@,5/=M8/450&/5&S2082?5&M53&?,?M.&J1/5M&/5&0,99.N/5>& 

.4;& =408& 82.2=4@@,5/=M8/450& >2M?& 84& 81402& M?2M0& ;1/=1& 

d52&4U&4U&812&@408&/@94?8M58&?20,.80&U?4@&812&/53,08?/M.& 

?2V,/?23& .200& 0491/08/=M823& 2V,/9@258& M53& ;2?2& .200& 

4?>M5/]M8/45& ./82?M8,?2& /0& 812& 0,>>208/45& 4U& M5& M5M.N8/=M.& 

.,=?M8/Q2&U4?&U4?2/>5&Q2534?0&M53&A4/58&Q258,?20&& 

3/08/5=8/45& -28;225& 9?43,=8& @M?Z28& =4@928/8/45& M53& 

?2>/45M.& 3/09M?/8N7& 258?N& /5& 81402& @M?Z280& ;2?2& 12MQ/.N& 

=4@928/8/45&/5&/554QM8/45M?3& ;2& U4..4;& @MA4?& 

25=4,?M>23& -N& 812& J1/5202& 5M8/45M.& >4Q2?5@258& M.04& /5& 

=45=2980&/5&WM?/M5&M53&E1M9/?4&'#YYD(&M53&6&9?43,=8/45&9?4=20020&M53&=4080(254,0& /554QM8/45& 81?4,>1& Gg_7812?2U4?2& >M/5/5>& 

M?2& =450/32?23& ?2.M8/Q2.N& [08M8/=[7& 9?43,=8& @M?Z28& 

=4@928/8/Q2& 08?25>81& M53& /5=?2M0/5>.N& =4@928/5>& M>M/508& 

=4@928/8/45& /0& 812& @M/5& =1M552.& 81?4,>1& ;1/=1& 812& 

U4?2/>52?0& 45& .M?>2& 0=M.2& 9?4A2=80& /5& 812& 258/?2& J1/5202& 

=4@928/8/Q2&9?4=200&8MZ20&9.M=2225=27& /5& ;1/=1& 2M=1& 82=154.4>N[0& 

9?43,=8/45& 9?4=20020%& 81/0& =4@928/8/45& /0& 4U825& M004=/M823& 

32Q2.49@258&/0&3/?2=823&M8&2\9M53/5>&/80&M992M.&548&45.N&/5& 

;/81&812&/32M0&4U&M&=4@928/8/Q2&81?2M8&U?4@&/554QM8/45& 

S1/.2&812&J1/5202&=4@9M5/20&;/81&812&8M=/8&0,994?8&4U&812& 

',5=44?3/5M823(& /554QM8/Q2& M=8/Q/8N& 81M8& =M5& -2& /3258/U/23& 

J1/5202&>4Q2?5@258&=458/5,23&84&>M/5&@M?Z28&01M?2&M>M/508& 

;/81&812& 

! "#! 

190



!C& 

RM50&S2?)&J1/5202&T554QM8/45&M53&J4@928/8/45)H5&T53,08?N&JM02&4U&812&L.4-M.& & 

F2.2=4@@,5/=M8/450&IV,/9@258&T53,08?N& 

9?4092=8&4U&/58?43,=8/45&4U&9?43,=80&4?&02?Q/=20&81M8&;/..& 

=4@9282& /5& 812& U,8,?21& M& 9?4=200& 4U& 

.2M?5/5>*-N*34/5>& /5& 

/554QM8/457&;2&2\92=8&812/?&32=/0/450&?2>M?3/5>&/554QM8/Q2& 

08?M82>N&84&-2&/5U.,25=23& 

0/>5/U/=M58.N& -N& 812& /554QM8/Q2& 08?M82>/20& 2@9.4N23& -N& 

?/QM.&U/?@0&j&UM/.,?2&84&/554QM82&0,==200U,..N&;/..&.2M3&548& 

A,08& 84& .400& 4U& 948258/M.& 9?4U/8& -,8& M.04& ?/0Z& UM../5>& -21/53& 

/554QM8/Q2&?/QM.0N& 

4U&[08M8/=[&M53&[3N5M@/=[&=4@928/8/45&812&0M@2&./520& 

M0& M-4Q27& 812& 08M8/=& 3/@250/45& 4U& =4@928/8/45& /0& ,0,M..N& 

?2.M823&84&9?/=2&=4@928/8/457&;1/=1&8MZ20&9.M=2&>/Q25&812&028& 

4U&9?43,=80&4?&02?Q/=20&81M8&M?2&@M?Z2823&-N&U/?@0,/3M5=2&U4?&=4@928/8/45&94./=N&9?M=8/=2&/554QM8/45&M53&@M?Z28&3N5M@/=0&/584& 

M==4,58& U4?& 812& 9,?94020& 4U& =4@928/8/45& 94./=N1& M5M.N0/0& 4U& =4@928/8/45& ;/81/5& M& ?2.2QM58& @M?Z282&/5&812&=45=2980&4U&=4@928/8/45& 

,023&U4?&M00200@258,/3M5=2&32Q2.4923& 

/5& 0,-02V,258& 02=8/450& /0& 84& ?2.M82& 81202& 3/@250/450& 4U& 

=4@928/8/45&84&9?M=8/=M.&=45=2980&,023&/5&=4@928/8/45&94./=N& 

M00200@2587&/5&9M?8/=,.M?&84&@M?Z28&32U/5/8/45&M5M.N0/0M5/]M8/45&@432.07&M53&812&?2Q/2;&4U&08,3/20& 

4U&528;4?Z&/53,08?/207&0,>>208&81M8&=4@928/8/Q2&/582?M=8/45&/5& 

/554QM8/Q2&M=8/Q/8/20&@MN&8MZ2&3/UU2?258&U4?@0&M==4?3/5>&84& 

812&08?,=8,?2&4U&9MN4UU0&4U&/554QM8/Q2&M=8/Q/8/20&84&[;/552?0[& 

M53&[.402?0[M5/]M8/45&@432.07&M53&812&?2Q/2;&4U&08,3/20&4U& 

528;4?Z& /53,08?/207& 0,>>208& 81M8& =4@928/8/Q2& /582?M=8/45& /5& 

/554QM8/Q2&M=8/Q/8/20&@MN&8MZ2&3/UU2?258&U4?@0&M==4?3/5>&84& 

812&08?,=8,?2&4U&9MN4UU0&4U&/554QM8/Q2&M=8/Q/8/20&84&[;/552?0[& 

M53&[.402?0[&-252U/80&84&812&@M?Z280& 

U4?&0,99.N&4U&=4@9M8/-.2&04U8;M?27&-N&2\9M53/5>&812&@M?Z28& 

0,=1&81M8&M&>?2M82?&QM?/28N&4U&04U8;M?2&/0&4UU2?23&4?&812&,5/8& 

9?/=2&32=?2M020&'0/5=2&U/\23&=4080&4U&04U8;M?2&32Q2.49@258& 

=M5&-2&09?2M3&4Q2?&M&>?2M82?&5,@-2?&4U&948258/M.&-,N2?0(&0N082@&@4?2&QM.,M-.27&M53&81,0&/5=?2M02&32@M53& 

U4?&/8/=& 

?2094502& 4U& 04U8;M?2& 0,99./2?0& 84& 812& /5=?2M02& /5& 812& 

=450,@98/45&4U&492?M8/5>&0N082@0



& T582?5M8/45M.&^,0/5200&M53&I=454@/=0&G202M?=1&!"#$%&!'!()&!!*$!& !O& 

& 

-2882?&81M5&2\/08/5>&9?43,=80&45.N&-N&04@2&,02?0&'4?&U4?&04@2& 

,020(&M0/32&9?/=2&=450/32?M8/4507&/0& 

=450/32?23&-2882?&81M5&2\/08/5>&9?43,=80&-N&M..&,02?0&'U4?&M..& 

,020(%&125=2&812&/32M&81M8&9?43,=80&=M5&-2&?M5Z23&M==4?3/5>& 

84&M&lV,M./8N&.M332?m?4,90&81M8& 

=M5&=42\/08&;/81&4.32?&9?43,=8&>?4,90252?M..N& 

M004=/M823&;/81&M&9?4=200&;12?2-N&52;&M53&-2882?&9?43,=80& 

3/09.M=2&4.32?&4-04.282&9?43,=80&U?4@&812&@M?Z28& 

lV,M./8N&;/534;m&2\/0807&4,80/32&;1/=1&U/?@0&=M5548&0,?Q/Q2& 

/5& 812& @M?Z28(,/32& 45& 812& 

5M8,?2& 4U& @M?Z28& 3N5M@/=0& 81M8& @MN& MUU2=8& M& 9M?8/=,.M?& 

/53,08?N>208& 81M8& /5=,@-258& U/?@0& @MN& 

1MQ2&3/UU2?258&/5=258/Q20&84&/554QM82&81M5&258?M580/Q25&812&3/UU/=,.8/20&81M8&/5=,@-258&.2M3/5>&U/?@0&@MN&1MQ2& 

84&32M.&;/81&3?M08/=&=1M5>20&o&M&U4?@&4U&4U&/58?M*U/?@& 

0N082@/=&/52?8/M&o&812&4994?8,5/8/20&U4?&0829*;/02&/554QM8/45& 

@MN&?2.M8/Q2.N&UMQ4?&/5=,@-258&U/?@0252?M.&8124?N252?/=& .2Q2.7& 

>/Q25& 812& =,??258& .2Q2.& 4U& ,532?08M53/5>& 4U& 81202& 2UU2=80&812&./5Z&-28;225&9?43,=8&@M?Z28& 

=4@928/8/45&M53&/554QM8/45,/3M5=2&U?4@&812&./82?M8,?2&45&81/0&/00,27&;1/=1& 

/0&92?1M90&-208&M33?20023&45&M&=M02*-N*=M02&-M0/01& eKFE& ;2?2& 0Z/@@23& 4UU& U?4@& 

M58/=/9M8/5>&@M?Z28&9M?8/=/9M5807&45.N&84&Z54;&.M82?&81M8&?/QM.& 

08M53M?30& -M023& 45& 52;& 82=154.4>/20& 'S/*b/(& -N9M00/5>& 

eKFE7& ;4,.3& =?2M82& =4@9.282.N& 52;& @M?Z280& 548& -2/5>& 

M58/=/9M823



!B& 


F2.2=4@@,5/=M8/450&IV,/9@258&T53,08?N& 

@M?Z280&812&?M5>2& 

4U& 9400/-.2& 0,-08/8,8202& @MN& 

MUU2=8&812&08?,=8,?2&4U&812&@M?Z28&M0&32U/5237&U4?&2\M@9.27&-N& 

@MZ/5>& 2=454@/20& 4U& 0=M.2& @4?2& 4?& .200& /@94?8M5820&@MN&-2&3,2&84&/554QM8/45&,532?8MZ25&-N&812&U/?@0&/5& 

812& @M?Z287& 4?& 3,2& 84& UM=84?0& 4,80/32& 812& @M?Z28& 0,=1& M0& 

2\82?5M.&82=154.4>/=M.&32Q2.49@25807&4?&01/U80&/5&=450,@2?& 

8M0820&M==4,58&4U&81202&UM=84?0&/0&548&0/@9.N&M&V,208/45&4U& 

2\M@/5/5>& 258?N& 9400/-/./8/2020& 

4Q2?&8/@22?7&812&8;4&@2?>/5>&U/?@0&@MN& 

=,??258.N& -2& 492?M8/5>& /5& =4@9.282.N& 029M?M82& ?2.2QM58& 

@M?Z2807& 81,0& ?M/0/5>& 54& =,??258& =4@928/8/45& =45=2?52&/5&812&U,8,?27&0,=1&81M8&81202&U/?@0&M?2&2\92=823&84& 

-2&=4@928/84?0&/5&812&U,8,?22?& 

014,.3& -2& -.4=Z23%& 0/@9.N& 81M8& =4@928/8/45& M5M.N0/0& @MN& 

5223&84&-2&U4?;M?3*.44Z/5>&M0&84&14;&=,??258& @M?Z280&M?2& 

./Z2.N&84&2Q4.Q2&/5&U,8,?21./>1823&M-4Q27&M&U/?@&@MN&548&-2&M=8/Q2&45&M5N&=,??258& 

@M?Z280&-,8&@MN&-2&32Q2.49/5>&9?43,=80&81M8&;4,.3&=4@9282& 

/5&812&U,8,?2M8/45& 45& 812& -M0/0& 4U& 04@2& 9?2*32U/523& /53/=/M>208&812&-252U/80&4U&812& 

U,?812?&M5M.N0/0&9?494023&/5&81/0&08,3N20& /5& ?2.M8/Q2& 

9?/=20&%&81202&U2M8,?20&M?2&3/0=,0023&-?/2U.N&-2.4;&4U&9M82580&/0&329253258&45&04@2&U4?@&4U& 

54Q2.8N7& /554QM8/45& /0& 52=200M?N& 'M.814,>1& 812& ?2.M8/Q2& 

/@94?8M5=2&4U&/554QM8/45&/5&812&=4@928/8/Q2&9?4=200&/0&548& 

U,..N&208M-./0123(/207&=1M?M=82?/08/=0&M53&9?/=20



& T582?5M8/45M.&^,0/5200&M53&I=454@/=0&G202M?=1&!"#$%&!'!()&!!*$!& !q& 

& 

7"C"$HA+63B(.$35$)& & 

!! /58?43,=8/45& 4U& 9?43,=80& 81M8& 2@-43N& M& 52;& 

82=154.4>N%& & 

!! 014?8&9?43,=8&./U2*=N=.20%& & 

!! 1/>1&9?494?8/45&4U&@M?Z28&8,?54Q2?&M==4,5823&U4?&-N& 

52;& 9?43,=80& '/$*.:%"#2% 

G".D-*%F:#"0+5$% % 

M53&R,M;2/&81402&7&/5&812&=4,?02&4U&32Q2.49@2587& 

?2M..N& 45.N& 81?22& _FF7& `FI& M53& R,M;2/& 1MQ2& -2=4@2& 

/582?5M8/45M..N&4U&=45=2?57&M53&812N&M?2&812&34@/5M58&U4=,0& 

4U&81/0&08,3N& M53& T>2.& '!""#(& 08,3/23& 81202& U4,?& J1/5202& 

0,99./2?0& ;/81& 8;4& A4/58& Q258,?207& ^2/A/5>& T582?5M8/45M.& 

E;/8=1/5>& EN082@& '^TEJ(7& M& E/2@250& A4/58& Q258,?2& M53& 

E1M5>1M/&^2..7&M53&H.=M82.*a,=258&A4/58&Q258,?24Q2?5@258& M00/08M5=2& ;2?2& 

/@94?8M58&84&J1/5202&0,99./2?0[&08?M82>/20&;1/.2&U4?2/>5&A4/58& 

Q258,?20&?2./23&45&U4?2/>5&=4@9M5/204Q2?5@258(4Q2?5@258&1M0&8?/23&84&32=?2M02&812&/@94?8& 

4U&E+J&0;/8=120&23&82.291452&492?M84?0&84&,02& 

9?43,=80&U?4@&5M8/45M.&@M5,UM=8,?2?0&84&`1M5>&M53&T>2.&'!""#(&7&/5&812&2M?.N&#YY"0& 

J1/5202*4;523& U/?@0& 12.3& 45.N& M-4,8& #& 9=& 4U& 812& @M?Z28& 

01M?2&4U&=258?M.&4UU/=2&2\=1M5>20M/523& 

@4?2& 81M5& 1M.U& 4U& 812& 5M8/45M.& @M?Z28& 0/5=2& #YYq& 

08?45>& =1M..25>20& U?4@& 34@208/=& U/?@0& M53& .4;2?& 9?4U/80& 

04@2&.2M3/5>&6M9M5202&U/?@07&0,=1&M0&b,A/80,&M53&XIJ7&1MQ2& 

;/813?M;5& U?4@& J1/5M[0& E+J& @M?Z287& M53& 04@2& .2M3/5>& 

S2082?5& E+J& 9?43,=2?0& 1MQ2& -225& =450/32?/5>& .2MQ/5>& M0& 

;2..M/508&A4/58&Q258,?20&M53&U4?2/>5&=4@928/8/45&/5& 

?,?M.& M53& ?2@482& J1/5202& @M?Z280& ;1/=1& ;2& =M..& 812& 

029M?M8/45&91M02&4U&=4@928/8/45&812N&1MQ2&-225&.2M?5/5>&45[& 

=4,?027& =45Q/=8/45& M53& =M9M-/./8N[& 'T^K& ^,0/5200& T508/8,82& 

45&J4@9M5N&L?4;81(&81?4,>1&=M8=1,9&3?/Q25&/554QM8/45&84& 

812& 2\8258& 81M8& 812N& =4,.3& =4@928282& 12M3& 45& ;/81& @MA4?& 

U4?2/>5&0,99./2?0&4?&812/?&A4/58&Q258,?20&/5&@MA4?&,?-M5&M?2M0& 

M.45>& 812& =4M08M.& =?20=258& 4U& J1/5M7& 08?28=1/5>& U?4@& Rf7& 

E125]257& E1M5>1M/& M53& ^2/A/5>25=2&91M02& 4U& 1/>1& 

82=154.4>N& U/?@0& M0& M& 82=154.4>N& ?M=2& /5& ;1/=1& UM../5>& 

-21/537&>288/5>&M12M3&M53&=M8=1/5>*,9&84&/53,08?N&.2M32?01/9& 

/0&812&5M@2&4U&812&>M@2&'L488/5>2?7!""B7J1M9.2& U/\23& ./52& =M??/23& M& O7"""& G2@/@-/& 

'GK^(&=408&7&M53&4U825&M54812?&O7"""&GK^&;M0&522323&M0& 

M5&2\923/8/5>&U22&'_2S40Z/57!""#(&/5U?M08?,=8,?M.&?2U4?@&M53&>?4;81&/5& 

J1/5M&/8&;M0&32=/323&84&,02&82.2=4@@,5/=M8/450&2V,/9@258& 

0,99./2?0&84&9?4@482&812&32Q2.49@258&4U&545*A4/58&Q258,?2& 

=M9M-/./8N& /5& J1/5M& M0& 2M?.N& M0& @M?Z28& 8?M50/8/45/5>& ;M0& 

8M?>2823&'X4,>18457!""q7&J1M9&528;4?Z&2V,/9@258&@M?Z28&4?/>/5M..N& 

U4,?& =4@9M5/20& ;2?2& /3258/U/23& M0& Z2N& 2\9452580& 4U& 81/0& 

08?M82>N)& 6,.45>& 'L_F(7& _M8M5>& '_FF(7& `145>\/5>& '`FI(& 

52;& 9?43,=8& 32Q2.49@258& /5[M99?49/M82[& 82=154.4>N& 81M8& 

! "'! 

194



!D& 


F2.2=4@@,5/=M8/450&IV,/9@258&T53,08?N& 

=4,.3&.2M3&84&M&9M?M3/>@&01/U8&4U&81M8&?M=22524,0&U/?@0&=M@2&,9& 

;/81&0/@9.2?7&0,UU/=/258&92?U4?@M5=2&.2Q2.&4U&528;4?Z&>2M?&M8& 

M&@,=1&.4;2?&9?/=2&U4?&,02&/5&?,?M.&J1/5M& ;1/=1&=4,.3&-2& 

/3258/U/23& M0& M99?49?/M82& 82=154.4>N@258M8/45&M99.N/5>&84&?,?M.&M53&,?-M5:@28?494./8M5&M?2M0.N7& 812& 0M@2& 844.0& 4U& @M?Z28& 92528?M8/45& ;2?2& 

M99./23&-N&J1/5202&U/?@0&.M82?&45&45&/5&4812?&2@2?>/5>&M53& 

8?M50/8/45M.&2=454@/20& & 

Q/?8,M..N& M..& M?4,53& 812& ;4?.32823& M8& ?,?M.& M53& ?2@482& M?2M0& 4U& J1/5M7& =.2M?.N&/5&M& 

32Q2.49/5>&2=454@N&7&=4,.3&-2&82?@23&[?2Q2?02&/554QM8/45[& 

/5&812& @2M5/5>&4U&L4Q/53M?MAM5&M53&F?/@-.2&'!"#!(&0/5=2& 

812N& 4?/>/5M823& /5& 812& 52230& M53& ?2V,/?2@2580& 4U& 81/0& 

/5U?M08?,=8,?M.& 02=84?& 81M8& ;M0& 029M?M82& 84& 812& M3QM5=23& 

,?-M5&M?2M0&/5&J1/5M4?/]2&8;4&M092=80&4U&81202&/554QM8/450&81M8& 

?2.M82&84&812&3/=1484@N&4U&812&82.2=4@&@M?Z28&/5&J1/5M&/5&812& 

2M?.N&#YY"0)& 

'#(&52;&@M?Z280&M53&'!(&V,M./8N&=1M?M=82?/08/=0443& 

254,>1[&;/81& 

0/>5/U/=M58.N& .4;2?& 9?/=/5>& M8& 3229& 

3/0=4,580"#"$E3B@(+&+&=($GI,.(.$ $ 

@M?Z280&0/>5M..23&812&08M?8&4U&812&=4@928/8/Q2&=45Q2?>25=2& 

9M815&0,99./2?0&=4@92823&12M3*84*12M3& 

U4?8&12&,?-M5&M53&?/?M.&J1/5202&@M?Z28&M53&8125&/5&812&>.4-M.& 

@M?Z28?2M82?& 

5,@-2?0& 4U& 3/>/8M.& 0,-0=?/-2?& ./520& '_Ea(& /5& J1/5M& 81M5& 

U4?2/>5& 0,99./2?0& /5=.,3/5>& 812/?& A4/58& Q258,?20& 

3/3'SE67!""C%_/88-2?52?7!""C(&=45=2?523&M-4,8&812& 

U/2?=2& =4@928/8/45& U?4@& J1/5202& 0,99./2?0& /5& 812& >.4-M.& 

@M?Z2805M..23&M.?2M3N&M&8?M50/8/45&U?4@&812&=4@928/8/Q2& 

=45Q2?>25=2& 9M81& 84& 812& =4@928/8Q2& >.4-M./]M8/45& 9M81& U4?& 

J1/5202& 0,99./2?025=2& 9M81& 

82=154.4>/20& 2Q4.Q2& 84& =4@9282& 12M3*45& U4?& 812& 0M@2& 

=,084@2?05& 

0,99./2?0& 04.3& 1/>1& =M9M=/8N& E+J& 0;/8=120& 84& 0M8/0UN& 812& 

52230&4U&.M?>2&7&UM08&>?4;/5>&=/8/202823& 812& /5U?M08?,=8,?M.7& 

545*=450,@2?&@M?*Z28&M53&M3323&QM.,2&84&UM=/./8M82&>?4;81& 

9?4092=80&/5&,532?32Q2.4923&J1/5202&?2>/450&=2582?0'.2Q2.O(1M/& 

^2..[0& 9?43,=80& ;2?2& 04.3& 84& .M?>2& 3/08?/=8& =2582?& '.2Q2.& 

#(79?4Q/5=/M.&0;/8=1/5>&=2582?0&'.2Q2.&!(&M53&=/8N&0;/8=1/5>& 

=2582?0&'.2Q2.&$(4&U?4@&.M?>2&84&0@M..& 

>24>?M91/=&,5/805& 0,99./2?0& 

M53&812/?&A4/58&Q25,?20&M==4,5823&U4?&qD9=&'KTT7&!""C(& & & 

H54812?& ?2M045& U4?& 812& 029M?M8/45& -28;225& J1/5202& 

0,99./2?0& M53& 208M-./0123& 0,99./2?0& ;M0& 81M8& ?2>,.M84?N& 

?2V,/?2@2580&-N&812&+408&M53&F2.2=4@@,5/=M8/45&^,?2M,0& 

'+F^0(& /5& ?,?M.& M?2M0& ;2?2& V,/82& 3/UU2?258& 81M5& 812& 

?2V,/?2@2580& 4U& 812& 9F^0& /5& 812& =/8/20& 

F1,07&/5&2\9.M5/5>&812&91M020&4U&@M?Z28&32Q2.49@258&U4?& 

J1/5202&U/?@07& 

/5&82?@0&4U&1/084?N&;2&@MN&/3258/UN&812&029M?M8/45&91M02& 

?4,>1.N& ;/81& 812& 92?/43& 4U& @/3& #YD"0& 84& @/3& #YY"07& 812& 

=45Q2?>25=2& 91M02& U?4@& @/3& #YY"0& 84& 2M?.N& !""0& M53&812& 

>.4-M./]M8/45&91M02&U4?&812&?208&,58/.&54;/20&'./Z2&E+J& 0;/8=120(& 

-,8& M.04& 84& -?4M3& 9?43,=8& >?4,90& M0& 528;4?Z& ?4,82?0& M53& 

02?Q2?0& M0& ;2..& M0& 84& =450,@2?& 253& 9?43,=80& 0,=1& M0& 

0@M?8914520&M53&8M-.28&32Q/=20&814,>1&81402&@M?Z280&1MQ2& 

2@2?>23& 45.N& ?2=258.N/=M..N& @48/QM823& 81?4,>1& /554QM8/45& -,8& @MN& -2& 

0,994?823& -N& 08?M82>/=& 3/?2=8/45& 4?& 4812?& UM=84?0& 0,=1& M0& 

>4Q2?5@258&3/?2=823&/53,08?/M.&94./=N/20&34& 

548&/582?M=8&'/5&82?@0&4U&-2/5>&?29.M=2M-.2(&/5&812&=4,?02&4U& 

812/?&32Q2.49@25825=2&91M02&7&82=154.4>/20& 

2Q4.Q2& 84& =4@9282& 12M3*45& U4?& 812& 0M@2& =,084@2?0.4-M./]M8/45&9M81&=4@928/8/45&2Q4.Q20&81?4,>1& 

82=154.4>N& 92?U4?@M5=2& M53& .4=M.& @M?Z28& =1M?M=82?/08/=0& 

0,=1& M0& ?2>,.M84?N& ?2V,/?2@2580& 4U& 32Q2.49/5>& 2=454@/207& 

U4?2/>5& >4Q2?5@258& 0,994?8& M53& 0,-0/3/]M8/45& /=.,3/5>& 

948258/M.&=4??,98/45&0=12@20&/5&82532?/5>1&=M9M=/8N&E+J& 

0;/8=120&/5&#YYO&U4?&0M.2&/584&812&J1/5202&,?-M5&M53&?,?M.& 

9.M8U4?@0'I3V,/087&!""$(



& T582?5M8/45M.&^,0/5200&M53&I=454@/=0&G202M?=1&!"#$%&!'!()&!!*$!& !Y& 

& 

>"!"$G)3:A6+$F(=(/3@B('+$J$0,'A5,6+A)&'K$E,@,L&/&+D$,':$ 

7&!""!(&M53&a/,7&!""$(2& 9M?80& 4U& 812/?& 

?2Q25,2&/5&Gg_&?2V,/?23&84&32Q2.49&1/>1&=M9M=/8N&0;/8=120252?M823& 4Q2?& 812& 

9?2Q/4,0& U4,?& N2M?0&84&32Q2.49&/80&RJ& 0;/8=10&84&32Q2.49&/80&RJ&0;/8=11& =M9M=/8N& 0;/8=120& ;M0& =450/32?23& M5& /@94?8M58& 

-?2MZ*81?4,>1& 84& 8?M50/8& 84& 812& =45Q2?>25=2& 91M02>2?&528;4?Z&M3QM58M>2&/0&./Z2.N&84&M88?M=8&@4?2&=,084@2?0& 

/5&812&M-025=2&4U&/554QM8/451&9?4-M-/./8N7&8125&MU82?& 

?2M./]M8/45& 81/0& U/?@& ;/..& >28& M12M3& 4U& /80& ?/QM.0& M53& >M/5& 

@M?Z28&01M?2&45&/80&-?M53&/5329253258&4U&9?/=/5>



$"& 


F2.2=4@@,5/=M8/450&IV,/9@258&T53,08?N& 

.2M32?01/9&940/8/45& & 

812N&1MQ2&548&45.N&84&=M8=1&,9&-,8&M.04&84&.2M9U?4>&812/?& 

?/QM.0N&U?458/2?&/8&/0&M&@,=1&@4?2&3/UU/=,.8&81/5>&84&342M?7& 

J/0=4& EN082@07& =450/32?0& R,M;2/& M0& /80& 84,>1208& ?/QM.& 

'SE67!"#!3(/=& 940/8/45/5>& 4U& 812& 

/53/>2524,0& J1/5202& U/?@01& 812/?& =4@928/8/Q2& 91M020& 

U4..4;23& 812& ?,.20& 4U& L4& 4?& S2/[=1/7& M& 82??/84?/M.& >M@2& 

320/>523& 84& 0,-3,2& N4,?& ?/QM.& -N& 09M8/M..N& 25=/?=.2& M53& 

=4?52?&1/@&M53&=M98,?2&1/0&940/8/45&-N&34@/5M8/5>&1/@&4Q2?& 

812&-4M?3M?37&812&J1/5202&@M?Z28&.44Z0&./Z2&M&L4& 

-4M?3& ;1/=1& ;125& 2\9M5323& 84& 2@2?>/5>& M53& 8?M50/8/45M.& 

2=454@/20& 258M/.0& 812& >.4-M.& @M?Z28M@2&/0&548&4Q2?&N28N&/0&=4@92../5>&-,8&52230& 

U,?812?&?202M?=1&812&849&#""&J1/5202&2.2=8?45/=0:TF&=4@9M5/20/5>& 

M53&8?M50/8/45M.&2=454@/20&-,8&M.04&02Q2?M.&dIJ_&=4,58?/202&=M9M=/8N&E+J&2\=1M5>2&M53&M==200& 

528;4?Z&2V,/9@258&/5&#YY$&M53&#YYC7&?2092=8/Q2.N/=&9?/4?/8/20)&'#(&_2Q2.49&U?4@&-2/5>&M& 

0/5>.2& 0;/8=1/5>& 9?43,=8& =4@9M5N& 84& M& @,.8/*9?43,=8& 

=4@9M5N7&'!(&M33?200&812&52230&4U&812&,?-M5&@M?Z28&M0&;2..& 

M0& 81402& 4U& 812& ?,?M.& @M?Z28& M53& '$(& M33?200& 812& 52230& 4U& 

/582?5M8/45M.& & 

KM?Z280&M53&M==200&9?43,=80M>23& /5& ?202M?=1& 9M?852?01/90& 'G6W0(& 

;/81&.M?>27&;2..*208M-./0123&U/?@0&0,=1&M0&F2\M0&T508?,@25807& 

K484?4.M&M53&H>2?2&EN082@0&'`FI[0&M55,M.&?294?8&4U&!""$(N[7&X2;&h4?Z)& 

d\U4?3&e5/Q*^,0/5200& S22Z'!"#!(7[`FI& M/@0& 84& 9M00& ?/QM.& 

R,M;2/&/5&0@M?891452&0M.20&/5&J1/5M[7&6M57_



& T582?5M8/45M.&^,0/5200&M53&I=454@/=0&G202M?=1&!"#$%&!'!()&!!*$!& $#& 

& 

^2/A/5>)J4582@94?M?N&J1/5M&+?200& & 

sBt! J1/5M& F49& #""& I.2=8?45/=0& J4@9M5N& a/08'!""$(7& 

;;;M?& & 

s##t! _2S40Z/576/8M.& 0;/8=1/5>& 08M8,0& M53& 

U4?2=M08[7&;;;N& M53& 

RN92?=4@928/8/457&a45345)&G4,8.23>2& & 

s#Ot! L488/5>2?7R2)&KTF&+?200& & 

s!qt! +4?82?7K1& F2=154.4>N& 

J4@928/8/457&JM@-?/3>2)&RM?QM?3&e5/Q?200& 

=4,.3&0V,22]2&-/>&9.MN2?0[7&KM?=1&q& & 

s$Ct! F12& SM..& E8?228& 64,?5M.& '!"#!-(7[R,M;2/& M/@0& U4?& 1/>12?& 

9?4U/.2&/5&0@M?8914520[7&KM?=1&D& & 

s$Ot! F12& SM..& E8?228& 64,?5M.& '!"#!=(7[R,M;2/& M/@0& 84& -2=4@2& 

@4?2&@M8,?2[7&H0/M&F2=154.4>N7KM?=1&#D& & 

s$Bt! F12& SM..& E8?228& 64,?5M.'!"#!3(7[J/0=4& JId& 0220& J1/5M[0& 

R,M;2/&M0&/80&84,>1208&?/QM.[7&H9?/.&B& & 

s$qt! F12& SM..& E8?228& 64,?5M.& '!"#!2(7[J1/5M& SM?50& R,M;2/& 

G294?8&=4,.3&1M?@&e

2.7 COMPETITIVE POSITIONING THROUGH STRATEGIC ALLIANCES 


!"#$%&$%'()("*(%+,",-(.("#/%012$%3/%41$%3/%3333% 1 

!"#$%&'&'(%)$"*'&'"+'+,)&-."/,-)*&.0&%,'1)022'0+1%) 

3".#0&'"+4).%('%5)0+6)*7+&-%*'*) 

!"#$%&'(#'()*+,,-#.'() 

Institute of Economic Analysis, 

University of Osaka (KGU), 

Osaka, Japan 

E-mail: hansgottinger@aol.com 

!"#$%&'$( This paper explores competitive positioning through network 

competition on the basis of alliance formation (strategic alliances, joint 

ventures). From a strategic perspective, technological competition will be 

refined and expanded into new markets, or new markets will be created through 

alliance formation. Alliance formation could speed up competitive positioning 

and technological leadership in strategically important, though geographically 

diverse, markets. Revenue management tools can help in linking alliance 

formation to better competitive outcomes, thereby improving strategic 

directions. 

)*+,-%.#( competition; corporate governance; managerial economics; 

networks; revenue management; strategic alliance. 

/*0*%*1'* to this paper should be made as follows: <strong>Gottinger</strong>, H-W. (xxxx) 

‘Competitive positioning through strategic alliance formation: review and 

synthesis’, !"#$%&$%'()("*(%+,",-(.("#, Vol. x, No. x, pp.xx–xx. 

23"43-5%&673'8 1-$*#( Dr. <strong>Gottinger</strong> is a Professor of Economics at the 

University of Osaka (KGU), and the Director of the Institute of Management 

Science, Maastricht, NL. He has internationally taught and widely published in 

the areas of industrial, energy and environmental economics. In these areas he 

has also advised international organisations. His most recent forthcoming book 

titles, !""1),#51"/%6(78"121-9%,":%;9#=*7#*=(, illuminated the impact of the rise of the 

railway and related industries during the 19th century on the development of the modem 

capitalist corporation. During the Industrial Revolution, fixed assets produced value 

through the creation and distribution of hard goods, often across long distances. The wide 

geographic, temporal and financial requirements of managing and operating a railway 

compelled the invention and evolution of substantially more sophisticated and structured 

corporate organisations. Earlier organisational forms and business practices were 

incapable of effectively managing in this new environment. As such, the railway 

compelled the evolution of the modern global corporation. Later, the breadth and depth of 

modern corporations encouraged the evolution of ever more complex functional 

Copyright © 200x Inderscience Enterprises Ltd. 

199



2% ;?@$%A1##5"-(=% 

hierarchies to coordinate disparate resources within expansive, multi-divisional firms 

(Chandler, 1962, 1990). 

We witness a similar phenomenon as we enter the 21st century. In contrast to the 

Industrial Age, the emerging economy primarily generates value through the creation, 

dissemination and application of knowledge. Since the 1980s, networking technologies 

have created a dynamic similar to that of the railway in the 19th century, influencing the 

options for corporate structures, relationships and competition. The most important and 

far-reaching impacts occur as a result of how people and firms use these tools to create 

value. The conflict between proprietary ownership as a necessary means for profit and the 

social nature of knowledge comprises the fundamental dynamic compelling the 

transformation of organisational forms during the present period. 

In partial response to this challenge, building and maintaining corporate alliances has 

become an increasingly important capability for the pursuit of both operational 

efficiencies and competitive advantage. Alliances have been transforming competition 

and, as such, corporate strategy. The primary objective of this paper is to develop and 

apply a new conceptual framework providing insight into the impact of alliances on 

strategy (Section 5). 

Before we will deal with some pertinent issues of strategic environments that 

facilitate or constrain alliance formation, 

Section 2 shows how managerial strategy drives competition but in turn is driven by 

competition. Section 3 shows how Transaction Cost Economics (TCE) on one side limits 

the strategic manoeuvrability but, on the other hand, provides new opportunities to get 

ahead of rivals. 

The role of corporate governance in view of achieving superior competitive 

performance of the firm is discussed in Section 4. Section 5 builds the case for alliance 

formation to strengthen competitive performance. Section 6 pinpoints the dimensions of 

network economics in achieving leading market positions. Section 7 adds the time 

dimension as the facilitator of change management. Section 8 outlines the scope of 

applications of network formation to various high-tech industries. 

>&8-08#$%&$*5+(8'->>3$>*1$?8;1'*%$&31$+8&1.8'7&15*8 

Constructing and executing strategy, proactively or passively, requires predictions, yet 

the future always presents uncertainty. Herein lies the dilemma of the strategist – 

committing resources to an uncertain future. As market and technology change 

accelerates, uncertainty becomes of increasing concern. In a most basic challenge to 

strategic planning, in the view of Porter (1980), how can firms effectively position 

themselves and their offerings if marketplace positions keep changing? Alternatively, 

executing on strategy in the real world requires firms to commit to directions they believe 

will provide sustainable profits. Commitment theory, as advanced by P. Ghemawat, 

argues that strategy requires resource investments that might be difficult or even 

impossible to recover from in the future, if too many strategic decisions turn out to have 

been wrong (Ghemawat, 1991). Strategically phrased, you cannot arrive anywhere in 

particular if you do not commit to a direction – and you cannot know in advance if you 

have chosen a good destination. Developing advantageous competency and resource 

combinations require substantial time and effort, so re-direction can be costly. 

200



% B1.



4% ;?@$%A1##5"-(=% 

substantially decreased the uncertainty regarding the marketplace availability of inputs or 

capabilities to accomplish economic objectives. If one needs something made or 

accomplished, one can probably (out) source it. 

Certainly, the control over resources provided by vertical integration still provides a 

compelling rationale for the formation and boundaries of firms; however, the uncertainty 

of another sort increasingly dominates as we move to the future. Rapid technology and 

marketplace change threatens firms by challenging the relevance of every product, 

service and business model. This represents the uncertainty regarding what to execute or, 

by extension, what, exactly, to integrate. Which competencies and other resources will 

provide a competitive advantage over the long term, most effectively integrated into the 

firm? While this issue presents the most significant challenge in dynamic markets, even 

mature industries can transform as a result of many factors, such as changing regulatory 

regimes, new entrants, technologies or business models, or a significant re-direction by 

an established player. Moreover, smaller firms typically do not have the capability or 

resources necessary to achieve a broadly integrated firm, even if they have the 

philosophical wherewithal. 

The greater the uncertainty regarding the future, the more firms must construct a 

portfolio of strategic options, without compromising the focus necessary to successfully 

compete. So effective strategy requires commitments towards an uncertain future, while 

marketplace change threatens to undermine or even destroy a firm’s value creating 

capabilities. 

Markets characterised by rapid technology change exhibit high levels of alliance 

formation compared with mature industries, which tend to exhibit consolidation and even 

decline (Hagedoorn, 1993). The high frequency of alliance formation in emerging 

industries will become evident when we look at the biotechnology and pharmaceutical 

industries, which evidenced a ninefold increase in operative alliances between 1993 and 

2000 (<strong>Gottinger</strong>, 2004). More mature industries encountering re-organisation or 

transformation also experience increases in alliances, as occurred when the global 

chemical industry experienced severe overcapacity in the 1980s. Alliances and mergers 

allowed the industry to rationalise production in a manner that competing firms could not 

have accomplished individually. These alliances represented an industry-wide alternative 

to consolidation through Mergers and Acquisitions (M&As), which occurred as well 

during this period. 

So marketplace actors apply network strategies in the periods of significant change 

and uncertainty. This seems to contradict the notion that uncertainty engenders vertical 

integration. We are dealing with differing types of uncertainty, but the responses are 

more similar than might appear. Vertical integration and alliances both integrate activities 

more closely than in the open market. To an extent, alliances represent integration by 

other means. Nonetheless, alliances present significant strategic and managerial contrasts 

to integrated firms, not the least of which is the opportunity for firms to interact with a 

broader, more diverse and potentially more productive universe of knowledge creating 

opportunities than is possible within a single firm. Let us first examine its precedents in 

the established organisational economics and strategy literature. 

202



% B1.64*D In the early 1990s, Eastman Kodak Corporation became a pioneer in 

outsourcing Information Technology (IT) services. While many corporations outsourced 

various aspects of their IT operations, Kodak was one of the first large, multinational 

corporations to outsource the majority of its corporate IT function. After a long period of 

planning, Kodak outsourced to three firms: Digital Equipment Corporation (DEC), 

BusinessLand and IBM. Although a Harvard Business School case on the outsourcing 

strategy (Applegate and Montealegre, 1995) reports overall success of the outsourcing 

venture, the process of building the integration and management processes between these 

four firms proved more challenging than Kodak’s CIO originally imagined. Kodak did 

achieve cost savings as a result; however, the overall cost savings were lower than 

anticipated. Kodak found that by outsourcing they saved the costs from shifting IT 

personnel and infrastructure to the firms whose core competencies were IT services. 

Nonetheless, outsourcing introduced new costs involving developing staff and processes 

203



6% ;?@$%A1##5"-(=% 

to manage, or rather, coordinate the inter-firm efforts. Control costs were swapped for 

coordination costs. 

The TCE perspective biases the discussion of this example towards the costs of 

achieving business results. Other motivations for outsourcing should include higher 

quality and improved corporate focus on the part of all firms in outsourcing relationships. 

Effective outsourcing should provide equal or higher-quality service than would be 

possible in-house, or at least a satisfactory level of service, at a similar or lower cost. 

Thus, outsourcing should represent the net value creation between the contracting 

entities, allowing each to focus on its core competencies. While the TCE perspective 

approaches a successful outsourcing arrangement as a more efficient governance 

configuration, focusing on costs can neglect or de-emphasise the potential benefits of 

greater value creation. 

The transaction cost TOF presents transaction costs reduction as the organising principle 

of firms; as such, TCE should prescribe any organisational form that provides the lowest 

transaction costs for the economic value created. It is partly a result of the efficiency and 

effectiveness of the modem corporate form that a transaction cost approach has provided 

theoretical justification for the existence of firms. It is also partly historical. Firms, in the 

traditional definition, exist, so they therefore must provide transaction cost advantages 

over market governance, given TCE’s foundation in bounded rationality (Simon, 1947; 

Williamson, 1985) and the path-dependent nature of organisational and institutional 

development (Nelson and Winter, 1982; Hannan and Freeman, 1989). Moreover, the 

original context within which Coase developed his transaction cost insights to 

understanding firm existence and boundaries likely biased the foundation of the 

approach. Coase extrapolated from his experience working at the Ford Motor Company 

in the 1930s, recognising the company’s vertical integration strategy. 

He postulated that vertical integration was justified to decrease the transaction costs 

that the firm would have otherwise incurred purchasing inputs in the open market. In the 

1930s, the industrial production model defined by Ford strove almost exclusively for 

operational efficiency. TCE is fundamentally an efficiency-based approach to 

understanding firms. Industries and strategies defined by operational efficiencies seem 

best suited to TCE. A number of researchers have asserted over the past two decades that 

inter-organisational relationships are often not well explained by TCE, particularly when 

efficiency issues do not represent the primary decision factors (Mariti and Smiley, 1983; 

Zajac and Olsen, 1993; Hamel and Prahalad, 1994). Decisions to pursue new firm 

competencies, or to form long-term R&D alliances are unlikely to be atomisable into 

discrete transactions (Argyres and Silverman, 2004). 

E8 

F-%6-%&$*85-G*%1&1'*8 

Alliances can be a critical strategic mechanism for the long-term success of ventures; 

however, alliances include a broad range of types, characterised by a number of factors 

such as integration, co-ownership, trust and longevity (Inkpen, 2001). Varied forms of 

alliance constellations also create varied constraints for participants, from the long-term 

commitments into which they enter to other potential partner opportunities which 

participants forego to build and maintain membership. For instance, membership in 

certain strategic blocks can preclude involvement with competing blocks, the clearest 

204



% B1.



8% ;?@$%A1##5"-(=% 

So when do network arrangements foster a competitive advantage, relative to firm or 

market organisation? When are the assets or resources worth more as part of an alliance 

than within an individual firm or the open market? How do networks assist in the 

creation, appropriation and sustainability of value? 

H8 

=%3G31580-%'*#8"*731.8&443&1'*80-%>&$3-18 

After all, why do firms form alliances? 

The implications of network strategies vary substantially, depending on the purposes 

for which a particular network of firms forms, as well as the purposes for which the 

network actually operates (which are not always identical). The creation and management 

of alliance constellations must be understood in light of motivations for their creation. 

While TCE presents a compelling approach to understanding firm boundaries 

(Milgrom and Roberts, 1992, Ch. 2), managers clearly engage in M&A, firm growth and 

divestiture, re-organisation and other boundary-shifting initiatives for motivations quite 

distinct from transaction cost minimisation. An extensive survey of decision makers 

involved in alliances in the early 1980s found that none of them cited decreasing 

transaction costs as a primary rationale for their decisions (Mariti and Smiley, 1983). 

Marketplace competition encourages firms to minimise costs, but it also compels firms in 

other directions: to acquire firms to pre-empt or respond to competitors, to adjust 

strategic vision in the face of disruptive technologies, to innovate to create value over the 

long term, to name a few. Innovation by its nature requires investment in the creation of 

new knowledge and capabilities, whether these capabilities are new to a particular firm or 

new to the marketplace. In the case of knowledge capital creation, such as in R&D 

partnerships, ‘transaction’ costs in the form of inter-firm coordination can in many cases 

be higher than an internally controlled effort; however, the overall value created by the 

combination of capabilities between firms can outweigh the increased costs of 

coordination. Zajac and Olsen’s notion of ‘transactional value’ attempts to reflect the 

importance of value calculations in understanding inter-organisational arrangements 

(Zajac and Olsen, 1993). 

While value creation and cost reduction reflect the complementary notions for the 

factors that encourage firms to change their boundaries through M&A or interorganisational 

arrangements, these approaches do not assist much in elucidating the 

complexity of motivations influencing organisational decisions. In broad terms of the 

field of industrial organisation, cost and value creation provide appropriate 

generalisations. For the purposes of applied corporate strategy, cost and value paradigms 

are by themselves woefully limiting, particularly given an ever-uncertain future. 

Economists attempt to reflect the behaviour of market actors based on preferences, 

available data and other decision factors, as well as provide prescriptions for the most 

effective economic actions as a result of the insights. The study of corporate strategy 

attempts to decipher why firms organise and act the way they do, how perhaps they 

should act regarding the development and execution of effective strategies (but often do 

not) and the outcomes associated with various scenarios (Grant, 2001). The fact that 

much of the strategy literature seeks grounding in economics reflects their 

complementary nature. Implicit in both of these disciplines is the notion of motivation 

(Milgrom and Roberts, 1992, Ch. 16). Economic behaviour, while observable in the 

marketplace, can only be modelled in the classical sense of optimising functions (even 

206



% B1.&$3-18 

What are the key building essentials the firms might need for network formation? There 

are potentially limitless reasons, but all rational, optimising motivations for forming 

alliances from a firm’s perspective’ can be categorised into three types, aside from purely 

financial motivations. 

1 4(#F1=G%(71"1.57C. A firm is attempting to compete in some manner under the 

conditions influenced by network economies. Network economies refer to the 

situations in which network economic phenomena significantly influence the 

dynamics of a particular industry or market (<strong>Gottinger</strong>, 2003). 

2 B1.



10% ;?@$%A1##5"-(=% 

environmental, institutional and competitive factors present in a particular industry, 

market and/or economy. The following list presents a summary of the examples 

collected: 

Factors favouring alliances 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Knowledge sharing and creation 

Distribution 

Risk sharing 

R&D outsourcing 

Network effects 

Leveraging complementary competencies 

Manufacturing outsourcing 

Increasing focus on specific competencies 

Valuation 

News 

Entering a new product line or line of business 

Entering a new market, geographic or offering brand leverage 

Capital access investment 

Pre-emptive move 

Market spanning 

While such a list might never be complete, most factors for alliances can be dominantly 

and usefully categorised into one, two or three of the types presented. The taxonomy need 

not be exhaustive to be effective. The point is that these three categories – Network 

Economics, Competencies and Market Structure (NCM) – provide the dimensions that 

most powerfully present implications for the creation and prosecution of network 

strategy. Competencies refer to the internal decisions a firm makes regarding the 

competencies it develops in-house, which competencies it chooses to source through 

partners, and which competencies it might leave for the market to provide. Market 

structure refers to the environmental, institutional and competitive factors of a particular 

industry, market and/or economy. Network economics refers to whether network 

economic phenomena significantly influence the dynamics of a particular industry or 

market under consideration. Network economics includes all of the issues traditionally 

associated with this field of economics, such as demand-side economies of scale, network 

effects and positive and negative feedback loops (Shapiro and Varian, 1999; Varian, 

Farrell and Shapiro, 2004). In such environments where fixed costs are high, variable 

costs low and decreasing on scale, demand is volatile and competition is keen, profitable 

growth can be achieved only with sophisticated price discriminating strategies designed 

to help you provide the right product, at the right time, for the right customer, at the right 

price, that is on ‘mass customisation’, as a major focus of revenue management. 

208



% B1.



12% ;?@$%A1##5"-(=% 

the firm’s effective monopoly. Any government mandated monopoly falls into this 

category as well, but can also overlap with other categories, as in the case of national 

telecommunications monopolies (network economics) or national control of raw 

materials mining and export. In the case of an oligopoly, existing players might, officially 

or unofficially, ally to maintain the status quo, as in the case of a cartel. The opposite can 

also be true. In highly disaggregated markets, firms might decide to ally in order to create 

economies of scale, lowering costs and developing bargaining power against suppliers. 

Economists use metrics such as the Herfindahl Index or a top four-firm concentration 

ratio to represent market concentration. 

A literature exists on the implications of the level of market concentration in the 

industrial organisation literature. The purpose here is not to develop and test many 

possible hypotheses related to market concentration and alliance formation, but rather to 

underscore that market structure plays a role in motivating the creation of networks of 

firms, which holds the implications for network strategy. 

Market structure dimensions are not limited to monopolistic or oligopolistic 

conditions. As evident from the examples presented earlier, the ‘market structure’ 

category includes institutional influences such as government regulation and anti-trust 

issues that might motivate inter-organisational arrangements over acquisition or market 

relationships. Here we define ‘institutional’ in the terms developed by North in the 

economic field of institutional analysis (North, 1990). These institutional issues relate to 

the market structure category of the taxonomy, given the fact that they influence the 

organisational decisions between firms (i.e. to integrate or ally) by impacting competitive 

market conditions, providing rules of the game. In this sense, it is not necessary to 

separate market structure and institutional issues into separate categories. The insights 

turn out on close inspection to be similar. 

H$M% J=5)(=C%D1=%,225,"7(CK%71.



% B1.6-%&48.3>*1#3-1(8%*6%*#*1$3158'7&15*8-G*%8$3>*8 

Firm constellations change over time. Sometimes, this occurs as a natural process of 

growth, or a deepening of contact between firms. Other times, participant firm objectives 

diverge. By definition, the NCM taxonomy underscores the influences that motivate 

firms to form alliances. As those influences change on a given constellation, so too will 

their nature and construction. The economics literature does address firm life cycles when 

211



14% ;?@$%A1##5"-(=% 

considering long-term decisions. In these cases, economists consider ‘entry and exit’" of 

firms over the course of the term in question. 

Nonetheless, the assumptions of firm perpetuity, both by researchers and 

practitioners, can be misleading. Rarely do firms last more than a few generations. Those 

that do, usually undergo significant change over time. While firms like Ford Motor 

Company continue to create value based on the original purpose for the company 

(automobile production), other venerable firms compete in completely different markets 

to their predecessors. Nokia’s current dominance of the wireless handset and 

telecommunications industries bears very little relationship with its past, which has 

included long periods in forest products, paper and pulp, and as a conglomerate with 

many unrelated lines of business. Can the Nokia of 1865 , the year of its founding, be 

considered the same firm as today? So, despite any legal continuity (and a company’s 

legal status can change), firms change markedly over time. Given that firms are by no 

means perpetual, researchers should treat them as entities finite in time and seek to 

understand firm lifespans and growth cycles (Jovanovic, 1982; Greiner, 1972). 

Similarly, hybrid forms should be treated as arrangements in flux, developed in 

response to, and as a result of, participant firms’ motivations and changing marketplace 

conditions. The taxonomy provides varied insights into the life cycles of firm 

constellations. Consortia created to motivate standards development in an emerging 

industry, at the interface of network economics and competencies dimensions, would be 

likely to migrate further to the network economics after the standard has been developed 

and disseminated. Once the standards race has been won (not necessarily by the 

consortia), the motivation for the alliance recedes, and the consortia is likely to be 

disbanded unless alternative benefits have been developed between member firms. By 

contrast, alliances in the competencies dimension often take one of the two predominant 

directions, assuming the competencies that motivated the alliances continue to be 

perceived as useful by the participants. Either firms in a competency-based alliance build 

integrated, mutually dependent competencies, which may encourage alliance longevity, 

or one or more of the participant firms decides to disengage from the alliance, perhaps 

having changed strategic priorities or having acquired the competency in house. Often, 

firms form alliances to develop a competency in-house, with the assistance of a partner. 

Sometimes firms do so avowedly, while in other cases knowledge appropriation occurs in 

a more surreptitious manner. Firms can participate in an alliance until they have acquired 

necessary knowledge from their partner, and then defect from cooperation. Microsoft has 

become notorious for this, particularly with small firms. Andy Groves’ famous 

‘paranoia’” (Groves, 1996) has been part of the reason Intel’s partnership with Microsoft 

has endured successfully for so long. Intel has maintained a healthy suspicion regarding 

its alliance with Microsoft. If one or more of the alliance members’ strategic interest, in 

particular competencies changes, competency alliances are most likely to be affected. 

Market structure provides the most obvious example of how the factors in each 

motivation space can impact alliance evolution. Government regulations can very quickly 

upset the competitive balances between constellations of firms, as has occurred in most 

instances of de-regulation or privatisation. The massive, global airline alliances (such as 

One World and Star Alliance), fit in the dimensions Network Economics and Market 

Structure. Airline de-regulation in the USA and elsewhere exerted a significant impact 

during the 1980s and 1990s, leading to a proliferation of airlines. Moreover, network 

economics heavily impacts the airline industry, characterised by communicating 

networks of routes, hub-and-spoke methods though anti-trust regulations continue to exist 

212



% B1.*1#3-18$-8G&%3-;#831.;#$%3*#8 

The network formation dimension can be applied to high-tech industries to assist in 

understanding the evolution of constellations of firms and network strategy under varied 

conditions. As an example, the dimension applies to understand the evolution of alliances 

within the biotechnology and pharmaceutical industries. The analysis will provide a proof 

of principle of the efficacy of the dimension, as well as an example of its application. 

A cursory examination of industry network strategy might suggest that specific 

dimension types dominate in given industries. Most clearly, constellations of firms in 

industries dominated by network economics, such as telecommunications equipment or 

the airlines industries, would likely coalesce within the network economics space. 

Industries dominated by the creation of new knowledge, such as those substantially 

involved in commercialising emerging technologies (e.g. biotechnology, optical 

computing), would favour the competencies dimension space. 

While these distinctions make intuitive sense, and further validate the application of 

the taxonomy, any purely static application of the taxonomy neglects a good deal of its 

explanatory power. Networks of firms, and network strategy in particular, must be 

understood dynamically. Strategy aims to change or maintain market relationships to the 

advantage of a particular firm or group of firms, relative to competitors (and in many 

cases relative to partners as well). Transforming market relationships for competitive 

advantage, as well as attempting to sustain an advantage, requires an understanding of 

how and why the relationships between firms change over time. While a particular 

category of the taxonomy might dominate an industry during a given period, examination 

over time will undoubtedly uncover shifting influences. A network strategist must 

become adept at monitoring industry developments critical to alliance formation, 

recognising these changing conditions early on, and quickly developing responses to 

maintain as advantageous a position as possible. An effective network strategy should 

maintain a vigilant eye for events substantially impacting any of the dimension spaces 

within or between industries, or throughout a sector or economy-wide. Such events can 

transform the relationships between firms, and can provide alert firms with the 

opportunity to drive transformations to their own advantage, or at least to successfully 

respond. 

213



16% ;?@$%A1##5"-(=% 

Both biotechnology and large pharmaceutical firms compete in an industry 

characterised by rapid technology change; in particular, these firms depend on the 

creation of new knowledge. Alliance competencies should be prevalent in any market 

characterised by fast-changing, intangible assets, given the difficulties inherent in trading 

intangibles; moreover, in industries with very high rates of technology change, 

technologies can be introduced that create new market segments, obsolesce existing 

product lines and create substantial competitors from previously little-known firms. 

Under such conditions, few firms can afford to conduct research in enough directions to 

build sufficient R&D options. Alliances offer opportunities for firms, in essence, to 

outsource R&D efforts, creating options on knowledge development, without requiring 

merger or acquisition. Biotechnology provides a critical example of such intangible asses 

that can be difficult to trade, and are, as well, not necessarily appropriate candidates for 

acquisition. Additionally, under conditions of fast change and high uncertainty, the 

network forms of governance provide preferred access to information, decreasing 

information asymmetries and allowing firms involved in a network to scan a broader 

environment. 

/*0*%*1'*#8 

Applegate, L.M. and Montealegre, R. (1995) N,C#.,"%O1:,G%B1$K%+,",-5"-%!"D1=.,#51"%>9C#(.C% 

#8=1*-8%>#=,#(-57%P225,"7(C. Boston: Harvard Business School Press. 

Argyres, N.S. and Silverman, B.S. (2004) ‘R&D, organizational structure and the development of 

corporate technological knowledge’, >#=,#(-57%+,",-(.("#%&1*=",2, Vol. 25, pp.929–958. 

Besanko, D., Dranove, D. and Shanley, M. (2000) N71"1.57C% 1D% >#=,#(-9 (2nd ed.). New York: 

Wiley. 

Chandler, A.D. (1962) >#=,#(-9%,":%>#=*7#*=(. Cambridge, MA: MIT Press. 

Chandler, A.D. (1990) >7,2(% ,":% >71#=,#(-9. New York: Free Press. 

<strong>Gottinger</strong>, H.W. (2003) N71"1.5(C%1D%4(#F1=G%!":*C#=5(C. London: Routledge. 

<strong>Gottinger</strong>, H.W. (2004) ‘Biotech and pharma industries: application of network economies’, 

R8,=.,S(*#5C78(%!":*C#=5(, Vol. 66, pp.1475–1478. 

Grant, R.M. (2001) B1"#(.#=,#(-9%P",29C5C. Oxford: Blackwell. 

Greiner, L. (1972) ‘Evolution and revolution as organizations grow’, ;,=),=:% T*C5"(CC% '()5(F, 

Vol. 50, pp.37–46. 

Groves, A. (1996) U"29%#8(%R,=,"15:%>*=)5)(. New York: Norton. 

Gulati, R. (1999) ‘Network location and learning. The influence of network resources and firm 

capabilities on alliance formation’, >#=,#(-57%+,",-(.("#%&1*=",2, Vol. 20, pp.397–420. 

214



% B1.#=,#(-57% 

+,",-(.("#%&1*=",2, Vol. 12, pp.105–124. 

North, D.C. (1990) ‘!"C#5#*#51"C/% !"C#5#*#51",2% B8,"-(% ,":% N71"1.57% R(=D1=."7(. Cambridge: 

Cambridge Univ. Press. 

Oxley, J.E. and Sampson, R.C. (2004) ‘The scope of governance of international R&D alliances’, 

>#=,#(-57%+,",-(.("#%&1*=",2, Vol. 25, pp.723–749. 

Porter, M.E. (1980) B1.#=,#(-9. New York: Free Press. 

Shapiro, C. and Varian, H.R. (1999) !"D1=.,#51"% '*2(C. Boston, MA: Harvard Business School 

Press. 

Simon, H. (1947) P:.5"C#=,#5)(%T(8,)51=. New York: Macmillan. 

Stinchcombe, A.L. (1990) !"D1=.,#51"%,":%U=-,"5S,#51"C$ Berkeley, CA: University of California 

Press. 

Varian, H., Farrell, J. and Shapiro, C. (2004) 68(%N71"1.57C%1D%!"D1=.,#51" 6(78"121-9. Raffaele 

Mattioli Lectures, Cambridge: Cambridge University Press. 

Williamson, O.E. (1975) +,=G(#C%,":%;5(=,785(C/%P",29C5C%1D%P"#5#=*C#%!.



216

3. 

GROWTH 

AND 

DEVELOPMENT

3.1 MICROECONOMIC FOUNDATIONS OF ECONOMIC GROWTH (REVIEW) 


Supplementing: Strategies of Economic Growth and Catch-Up . H.W. <strong>Gottinger</strong> and M. Goosen, 

eds., NovaScience, New York, 2012 

<strong>Hans</strong> W. <strong>Gottinger</strong>, STRATEC Germany 

www.stratec-con.net 

A Comprehensive Review of the Microeconomic Foundations of Economic Growth 

This review provides a summary of the main findings of the literature review (which, as 

supporting document for the text “Strategies of Economic Growth and Catchup”, is attached 

separately ). The section then provides a brief discussion of key theories and concepts that we drew 

from the literature 

Summary Findings from the Literature 

Technological innovation has an important role to play both as part of the competitive process and 

as a driver of economic growth and development. However, perhaps surprisingly, for a long time 

innovation was considered a field where economists, as opposed to other scientists and 

engineers, only viewed this as a ‘black box’.. The forces underlying the process of technological 

change were believed to be substantially independent of economic incentives and mostly affected by 

the exogenous evolution of scientific knowledge and its technological applications. 

It has been only recently, mainly over the last two decades, that the economic forces behind 

technological innovation have started to be investigated in more detail within mainstream 

economics. As a result, a wide variety of theories and models, sometimes very diverse in spirit, 

describing the economics of innovation are now available. All these theories share the common 

aim of providing a conceptual foundation for understanding how innovation affects the economy, 

how economic forces affect the emergence of technological changes, and the decision-making 

processes through which technological innovation occurs. 

We have made no attempt to provide the most representative and comprehensive summaries of 

each strand of literature; this is not our goal. Rather it is to give a sufficient summary to enable 

subsequent identification of how the study can be taken forward to develop analytical tools for 

assessing market dynamics in competition policy investigations. 

This section provides a brief summary of some key elements of the different strands of literature 

overviewed in much of this work. 

Industrial organization 

Studies of dynamic competition within economics have developed from once 

inherently static neoclassical economic theories to modern analyses of innovation and dynamic 

competition. 

Nonetheless, a strong relationship with neoclassical economics remains in the general 

methodological approach adopted in these theories, which are characterized by the analysis of 

equilibrium models where fully rational economic agents make optimizing choices. 

Generally, traditional models of innovation focus on the study of firms’ incentives to invest 

resources in Research and Development (R&D) activities. Game-theoretical models developed 

in the industrial organization literature have investigated firms’ R&D decisions in strategic 

1 

219


3.1 MICROECONOMIC FOUNDATIONS OF ECONOMIC GROWTH (REVIEW)S 

environments. (Endogenous growth models, discussed in the next sub-section, have developed 

the study of market dynamics in models that explain the relationship between firms’ investments, 

innovation and economic growth.) 

Game-theoretical models suggest that there are two main forces that underlie firms’ investment in 

R&D: the search for higher profits and the threat posed by falling behind potential innovating 

rivals. Game theoretical models study these forces in a variety of market situations and address 

issues such as the interplay between innovation and market structure, the dynamics of 

competition and the nature of the relationship between intensity of competition and innovation. 

These models provide a rich picture of what the possible strategies and industry equilibria in 

dynamic markets can be. However, general predictions, that can be considered appropriate 

across all situations and industries, are scarce. On one hand, this is a result that seems to stress 

the lack of predictive power of these models, i.e. (almost) everything can be rationalized; on the 

other hand, however, the variety of results seems well to fit with the variety of observed 

behaviours and “equilibria”. There is no general model that can uncritically be applied to any 

case: the understanding of the specific characteristics of the single situation needs to underpin 

any appropriate choice of a modelling framework. 

Despite the absence of general results, these models are certainly useful tools to understand 

firms’ incentives to invest in R&D activities in strategic environments and to suggest what main 

factors may be central in shaping the nature of dynamic competition. For instance, these models 

suggest that: 

• In order to understand R&D investments in strategic environments it is necessary to 

understand how innovation may affect profits both of successful and non-successful 

innovators. The first perspective captures the idea that firms want to innovate to increase 

their profits; the second captures the idea that firms want to innovate to maintain 

competitiveness. 

• 

• 

The relationship between concentration of an industry and its rate of technological 

innovation is certainly complex and in general not a causal one: both should be thought of 

as the outcomes of the operation of market forces and exogenous factors such as the 

nature of demand, technological opportunity and the conditions governing appropriability. 

Dynamic competition may be characterized by persistent dominance of the incumbent 

leader or by action-reaction whereby incumbents are overtaken by a rival whose 

incumbency is itself then short-lived. The nature of market dynamics depends on a 

number of factors, such as the type of innovation, i.e. drastic or non-drastic, the 

uncertainties involved in R&D activities, the nature of patent protection and of knowledge 

spill-overs, the intensity of product market competition, etc. 

• 

When the relationship between competition and innovation is investigated, it is necessary 

to be clear what the notion of intensity of competition describes and how this relates (or 

does not relate) to market structure. Indeed a market where competition is tougher may 

be more concentrated simply because inefficient firms cannot survive. There may be a 

2 

220



trade-off between the intensity of static competition and innovation. In general, the 

relationship between intensity of competition and innovation need not be monotonic at all. 

Endogenous growth models have recently developed along the earlier game theoretical literature on 

innovation in the context of studies that seek to explain the relationship between innovation and 

economic growth. These models suggest that innovation, resulting from intentional R&D 

investments by profit-maximizing firms or simply by unintentional learning-by-doing, is a 

fundamental driver of economic growth in the long run. 

Early Schumpeterian endogenous growth models stressed the importance of ex-post rents for 

innovation: competition would have a detrimental effect on innovation by decreasing the rents that 

an innovator would be able to appropriate. More recent models have emphasized another 

mechanism by which competition affects innovation: tougher competition may increase the 

incentives of firms to innovate in order to escape from fierce competition. These recent studies 

suggest that the relationship between competition and innovation may not well be monotonic and 

that instead, one should expect an inverse-U shaped relationship: when competition is low, an 

increase in competition would foster innovation; the reverse would happen when competition is 

fierce. 

The result that competition may be conducive to innovation is also obtained in studies where the 

traditional behavioral assumption of profit-maximizing firms is relaxed. When principal-agent 

considerations are introduced to explain managers’ behaviors, another mechanism by which 

competition may favor innovation is suggested: the speed of innovation may be retarded by the 

slack of managers who tend to avoid private costs associated with innovation. When competition 

intensifies, the higher threat of bankruptcy may force managers to speed up the process at which 

new ideas are adopted. Hence, competition may be conducive to faster rates of innovation. 

New economic geography 

New economic geography is a branch of economics which is mainly concerned with spatial 

aspects of economics. In particular, it seeks to explain why and how given economic activities 

concentrate geographically, either within individual countries (agglomerations) or between 

different countries (industrial clustering). Furthermore, it considers the inter-relationships between 

geographical concentration of industry, international trade and economic development. 

At a broad level, several themes emerge from the literature. 

• 

• 

Geographical concentration of economic activity will occur where transport costs are low 

enough and the labor force is mobile. 

Key characteristics of a particular industry have also an important role to play. The higher 

the scope for vertical integration, the higher the share of intermediate goods in output and 

the lower the labor intensity are, the higher will be the drive for concentration. 

3 

221



Industrial concentration implies the possibility of geographical proximity of firms that lays 

the foundation for technological cooperation and thus lead to increased innovative 

activity. 

• 

A high degree of concentration leads to very large markets — and the larger the size of 

the market, the higher the potential reward of successful innovations. 

The evolutionary economics of innovation 

The evolutionary approach to the study of innovation has been developed on very different 

methodological basis than those underlying traditional economic models of innovation. In 

particular, we observe the rejection of the modelling assumptions of rationality and equilibrium 

that are fundamental to the traditional approach. 

Evolutionary economics looks from the outset at dynamic processes. In particular, it is associated 

with the use of analogies from evolutionary biology to explain economic growth and the process 

of competition. Thus the cornerstones of an evolutionary analysis of competition and innovation 

are variety, selection and imitation. 

At a basic level, using Darwinian analogies, we can begin to appreciate the role of the market in 

selecting the more fit firms (efficient and profitable), products and techniques at the expense of 

less fit firms, products or techniques. In addition to this effect, we would expect to see imitation of 

winning ideas by those whose survival is otherwise threatened (although this is limited by the tacit 

nature of knowledge). 

Inherent in this model of competition is the association between competition and experimentation 

and variety. A variety of experiments allows, through the process of selection, for greater 

economic progress than would be available through uniform optimization. 

Thus economic development and innovation can be seen as a combined effect of selection (via 

competition) from a variety of competing routines and practices as well as the more endogenous 

process of agents seeking improved routines and practices. While the latter is certainly 

incorporated, if treated somewhat differently, under the mainstream neoclassical approach, the 

emphasis on selection from variety seems an important addition to this approach. 

This suggests that models found in other branches of economics might miss something important 

when they analyze dynamics with reference to homogenous profit-maximizing firms: namely the 

benefits of selection from heterogeneity in capabilities and innovative experimentation. 

Systems of innovation 

At an aggregate level, the evolutionary approach to the study of economic growth draws attention 

to the importance of institutions in the process of economic growth. The key findings of the 

literature are as follows: 

• Innovation and its diffusion take place within systems of interconnected organizations and 

institutions. Important constituent elements of such systems are organizations such as 

4 

222



firms, governments and universities. Institutions, with which these organizations interact, 

reflect laws and statutes (e.g. the institution of patent protection) as well as more abstract 

elements (e.g. cultural aspects of the economy such as the spirit of entrepreneurial 

activity). 

• 

• 

Innovation within the system will depend not only on single institutions but also on the 

nature and intensity of interactions between the various elements of the system. 

When considering the interaction between competition and innovation, we must remain 

aware that the effects of such interaction will depend on the evolving institutional 

background against which agents in the economy operate. 

The overview of the cliometric literature has shown that cliometrics is a methodological framework 

for the study of economic history. The existing body of cliometrics draws attention to the following 

factors relevant to “systems of innovation”: 

• Economic growth is a result of a complex interaction between institutions and markets. 

Examples of institutions that are frequently important for innovation are the protection of 

property rights (for example, against piracy) and patent policy. 

Innovation plays a critical role in economic progress and development. This involves not 

only technological innovations but also changes in institutions. Thus, institutions 

themselves evolve according to changing circumstances. These two last points 

complement the concepts of systems of innovation found in the macroeconomic field of 

the evolutionary approach. 

5 

223



INTRODUCTION 

1 

1.1 

1.2 

1.3 

1.4 

2 

2.1 

2.2 

2.3 

3 

3.1 

3.2 

3.3 

4 

4.1 

4.2 

4.3 

4.4 

4.5 

4.6 

ECONOMICS OF INNOVATION 

Introduction 

Industrial Organization 

Endogenous Growth Models 

New Economic Geography 

THE EVOLUTIONARY ECONOMICS OF INNOVATION 


Microeconomics of Innovation 

Systems of Innovation and the Role of Institutions 

STRATEGIES OF INNOVATIVE FIRMS 

Innovation and Competitive Advantage 

Timing of Innovation and Life Cycles 

Issues Relating to Network Industries 

BIBLIOGRAPHY 


Endogenous Growth 

New Economic Geography 

Evolutionary Economics 

Cliometrics 

Strategic Management and Networks 

224

7 



INTRODUCTION 

This is an overview, with references to the seminal literature on the spectrum of dynamic 

approaches in micro and meso economics covering at least the last two decades – the 

overview should avoid a pure focus on mainstream approaches and draw benefits from 

overlaps with neighboring disciplines such as systems theory, cliometrics and 

institutionalism. 

This is a brief overview and classification based on the relevant literature of the most 

important economic strategies applied by innovative companies to enter new markets or 

introduce new products. 

Section 1 presents the survey of mainstream theories and models of innovation and Section 2 

discusses evolutionary economics, systems theory, cliometrics and institutionalism. 

Section 3 focuses instead on firms’ strategies related to innovative activities. 

225



1 

1.1 

ECONOMICS OF INNOVATION 


The neo-classical approach to innovation 

In neo-classical theories, production opportunities are described by specifying a production 

function, which describes the technological relationship between inputs and output. Technological 

progress is captured by the ability of a firm, or the economy, to produce, for each combination of 

inputs used, more output. However, such progress was for a long time essentially considered 

exogenous, as if it were knowledge that falls, like manna, from heaven. 

Nevertheless, neo-classical economists were conscious that the consideration of innovation as an 

economic phenomenon posed some challenges to established theories of perfect competition 

and that a trade-off may exist between static and dynamic efficiency. 

Innovation is essentially the production and commercial application of new knowledge. Neoclassical 

economists, following the pioneering work of Arrow (1962), have usually treated 

knowledge as information, that is, as a pure non-rivalrous and non-excludable good. The 

concepts of non-rivalry and non-excludability are crucial in understanding the neo-classical 

economic analysis of innovation and can be explained as follows: 

(a) Non-rivalry implies that the same piece of knowledge can be consumed by different 

individuals, the marginal cost involved being essentially zero. Knowledge is 

considered to be essentially disembodied (as opposed, for instance, to knowledge 

embodied in organizational routines — a central feature of the evolutionary approach 

overviewed in section 2) and available in a common stock or pool. In other words 

knowledge is essentially assimilated to information. 

(b) Non-excludability is the inability of the innovator to prevent others from accessing the 

knowledge produced. Non-excludability arises largely because it may not be feasible, 

or it is extremely costly, to monitor and charge for the use and enjoyment of a piece of 

knowledge. In traditional models of innovation, it is often related to the notion of 

knowledge spillovers, which refers to the positive externality by which the knowledge 

generated by an agent enhances the technological capabilities of other agents or their 

ability to discover new knowledge. Non-excludability is not entirely an intrinsic quality 

of knowledge and depends largely on the legal regime in place and, in particular, on 

the system of property rights (e.g. patents). Indeed, a patent might allow the innovator 

to appropriate the value of an (otherwise non-excludable) innovation by granting the 

exclusive right of its commercial use for a certain period of time. 

The neo-classical analysis of the economic forces underlying the production of new technological 

knowledge follows from the general theoretical framework described above and is developed on 

the assumption that economic agents are rational and act to maximize their utility or profit. 

When non-rivalry is coupled with non-excludability, and both are considered to be essential 

features of technological knowledge, innovation acquires the characteristics of a pure public good 

and, accordingly, profit maximizing agents are considered not to have incentives to invest in its 

production. 

8 

226



In fact, if the process of imitation of an innovation were immediate, a firm would not have the 

incentive to invest resources in R&D activities since ex-post competition would not allow the 

innovator to earn profits to pay back its investment. However, although a single firm would not 

have incentives to discover new knowledge, innovation may well be beneficial for society so that 

a role for public intervention in the form of direct R&D investment and/or appropriate public policy 

is usually called for. 

These considerations suggest that a trade-off may exist between static and dynamic efficiency 

and provide the economic rationale for some important public policies such as public funding of 

research and the definition of a patent system that protects the innovator from imitation. The 

basic trade-off of a patent system is that between the static costs that derive from the innovator’s 

potential monopoly power and the long-term benefits associated with the discovery of the new 

knowledge. 

Core models of dynamic competition 

In the last two decades, the analysis of innovation and technological change has attracted much 

attention within mainstream economics, due to the development of two closely related disciplines 

— game-theoretical industrial organization and endogenous growth theories. As a consequence, 

many of the themes that were once a feature of the less conventional theories of technological 

advance have now found their way into economics. 

Studies of dynamic competition within mainstream economics have developed once inherently 

static neoclassical economic theories into modern analyses of innovation and dynamic 

competition. Nonetheless, a strong relationship with neoclassical economics remains in the 

general methodological approach adopted in these theories, which are characterized by the 

analysis of equilibrium models where fully rational economic agents make optimizing choices. 

Core models of dynamic competition are highly stylized and this is also reflected in the 

definition of innovation adopted. In these models, innovation is usually defined as the reduction of 

production costs or as the commercialization of new and/or better products. Apparently this 

definition identifies innovation with technological change and downsizes the importance of other 

forms of innovation, such as managerial and organizational changes. 

However, notwithstanding the emphasis placed by these models on innovation as the commercial 

use of new or improved products and processes, this stylized definition of innovation can also be 

considered to encompass non-technological changes inasmuch these ultimately matter because 

result in a reduction of production costs or new and better products. 

What mainstream models suggest is that the particular form of innovation (e.g. new technological 

knowledge rather than new organizational knowledge) is not central to the understanding of firms’ 

incentives to exert effort to improve their ability to produce efficiently and to market new and or 

better products, which is the main question addressed in these theories. 

9 

227



Similar to the definition of innovation , also the treatment of the process by which new knowledge is 

generated is very stylized in these models. Knowledge is usually considered as deriving from 

learning-by-doing or from intentional R&D investments, optimally chosen by profit-maximizing 

agents. The R&D process is captured by the definition of a deterministic or stochastic knowledge 

production function which related the input used in R&D — notably capital, labor, available 

knowledge – to the new knowledge produced and/or the timing of innovation. 

Game-theoretical models study firms’ incentives to invest resources in R&D activities in strategic 

settings and address issues such as the interplay between innovation and market structure, the 

dynamics of competition and the nature of the relationship between intensity of competition and 


In general, these models are structured on the fundamental distinction between static and 

dynamic competition. Static or product-market competition refers to the strategic interaction that 

arises between firms, taking as given the capabilities of these firms in terms of the variety, quality 

and production costs of the products in their portfolios. Dynamic competition refers instead to the 

strategic interaction among rival firms to change these capabilities. 

This distinction is usually captured by the structure of game-theoretical models: firms are 

considered to play a game where they first invest in innovative activities and then compete at the 

product-market level. The distinction is a fictitious one to a certain extent, since in reality both 

dimensions of competition are continuously overlapping, but it is a convincing and powerful 

conceptual categorization since it allows distinguishing activities that occur over different timehorizons, 

i.e. short-term price competition versus. long-term dynamic competition. When firms 

compete in the short-term they have to take as given a number of constraints in terms of their and 

their rivals’ “capabilities” that derive from previous innovative activities; when competition is 

considered over a longer-time horizon, however, these “capabilities” become endogenous to the 

competitive process as well. 

Despite the absence of general results, game-theoretical models are certainly useful to 

understand what fundamental strategic forces and incentives drive firms’ investments in R&D and 

what factors may lead to different types of dynamic competition (i.e. persistence of monopoly or 

action-reaction). 

Endogenous (or new) growth models, pioneered by Romer, Lucas, Aghion and Howitt have tried 

to explain the economic forces driving the technological progress underlying economic growth. In 

these models economic growth, in general, derives from the growth of the stock of knowledge 

available to society, which in turn is the outcome of economic choices of profit-maximizing agents. 

Earlier models developed in the 1960s were instead mainly focused on capital accumulation and 

did not treat technological change as an economic phenomenon amenable to economic analysis. 

Although these models are based on the equilibrium methodology typical of the neo-classical 

approach they are inherently dynamic. The evolution of the stock of knowledge is explained as 

either being the product of unintentional learning-by-doing activities or the outcome of R&D 

investments undertaken by profit-maximizing agents and rewarded by the monopolistic rents that 

the innovator is able to appropriate. Externalities involved in the production of new knowledge 

10 

228



usually play an important role in these models, and in particular knowledge spillovers by which 

the knowledge generated by an agent enhances the technological capabilities of other agents or 

their ability to discover new knowledge. 

This section considers first game-theoretical microeconomic models of innovation developed in 

the industrial organization literature. It then considers endogenous growth theories and the new 

economic geography theories. 

1.2 Industrial Organization 

The game-theoretical literature on innovation, which was developed mainly in the 1980s and 

1990s, addresses the strategic interaction that characterizes R&D investment decisions in 

oligopolistic industries. These models provide the fundamental building blocks of more recent 

endogenous growth models, which cannot be properly assessed without reference to this earlier 

literature. 

More recently, game-theoretical models have focused on the analysis of competition and 

innovation in markets characterized by network effects, on the analysis of co-operation in 

research and development activities and on the study of the relationship between intensity of 

competition and innovation. 

The starting point of this section is a categorization of game-theoretical models that should act as 

a guiding map to this literature. Understanding the nature of the models, and not only their 

results, is unavoidable to assess the adequacy of a particular model as theoretical reference in a 

practical case. The results of game theoretical models are generally very sensitive to the specific 

assumptions adopted and only by having clear the link between the former and the latter can 

these studies offer useful guidance for practical analyses of competition and market dynamics. 

The following sections focus on some of the issues that are addressed in the game-theoretical 

literature and that are most relevant to the analysis of competition and innovation for antitrust 

purposes: 

• 

• 

• 

• 

the relationship between market structure and innovation; 

the issue of persistence of monopoly in dynamic markets; 

the relationship between toughness of product market competition and innovation; and 

R&D co-operation. 

An overview of the game-theoretical literature on R&D 

Game-theoretical models of innovation explain the strategic interaction that underlies firms’ 

investments in research and development activities. The focus on strategic behavior 

distinguishes this literature from earlier decision theoretic models, e.g. Kamien and Schwartz 

(1980). 

11 

229



Game-theoretical models of innovation can be categorized as belonging to one of three broad 

paradigms: 

• Non-tournament models. These models capture the strategic interaction that can arise in 

those cases where firms pursue different paths of technological advance so that a 

discovery made along one path can be used irrespectively on whether a competitor 

innovates along another technological path. These models have been investigated by 

Dasgupta and Stiglitz (1980a), Spence (1984) and Judd (1985). 

• 

• 

Deterministic auction models. Auction models of innovation describe R&D competition in 

terms of rival firms bidding for a given process or product innovation. Important 

contributions include Gilbert and Newbery (1982), Katz and Shapiro (1987) and Vickers 

(1986). 

Stochastic tournament models (or “patent races”). These models capture the strategic 

interaction that arises when firms compete to be the first to introduce an innovation and 

the issue of timing is crucial to firms’ strategic interaction. Papers in this category include 

Loury (1979), Lee and Wilde (1980), Dasgupta and Stiglitz (1980b), Grossman and 

Shapiro (1987), Harris and Vickers (1987), Reinganum (1981,1982). 

Each of these approaches describes R&D investments in a particular setting, which may be an 

appropriate depiction of reality in some circumstances and a less appropriate theoretical 

reference in relation to some other industries or to particular stages of a technology life cycle. 

Despite the differences, however, game-theoretical models of innovation illustrate that there are 

two central economic incentives at the basis of firms’ investments in R&D activities, whose 

relative importance varies according to the specific environment considered: the profit incentive 

and the competitive threat (or replacement effect). 

The profit incentive, or stand-alone incentive, is related to the notion that allocating resources 

to innovative research and development would, if successful, increase a firm’s profits. This 

incentive can be usually thought of as the incentive to invest in R&D of a firm that takes the 

innovation decision in isolation. In fact, it is equal to the difference between the profits that the 

firm would get if it innovates and those that it would get if it does not innovate, all else equal. 

The competitive threat, or pre-emption incentive, arises when strategic interaction in innovation 

activities is considered: a firm not only has to take into account the benefits connected to 

innovation but also the possible loss of competitiveness where it does not innovate and a 

competitor does. This incentive reflects the threat posed by the existence of an active rival and 

can be thought of as the difference between the profits if the firm innovates and the profits if it is 

left to a rival firm to innovate. 

12 

230



In general settings, both the profit incentive and the competitive threat may play a role in shaping 

the nature of dynamic competition. This would depend crucially on the specific characteristics of 

the market environment considered and, in the literature, on the specific assumptions adopted in 

each model. 

Non-tournament models 

Non-tournament models of innovation investigate R&D investments in a setting where there can 

be multiple discoverers and firms do not compete to be the first to innovate. In these models 

symmetric competitors cannot prevent each other from getting equivalent improvements from 

spending equivalent amounts of resources in R&D. 

For this reason, such models may explain R&D investments in those environments where firms 

pursue different research paths so that an advance made on a single path can be appropriated 

irrespective of rivals’ advances on other research trajectories. 

A typical non-tournament model, such as the one by Dasgupta and Stiglitz (1980a), is structured 

in 3 stages: an entry stage, an R&D stage and a market competition stage. Firms first decide 

whether or not to enter an industry; active firms would then invest resources in R&D activities to 

decrease their marginal cost of production or increase the quality of their products and would 

finally compete in the marketplace. 

R&D activities are described in terms of a deterministic production function that specifies the size 

of innovation according to the amount of cost reduction or quality improvement per unit of R&D 

investment. Each firm can reduce its costs or improve the qualities of its products independently 

of parallel innovation by rival firms, although it is usually assumed that a firm can indirectly benefit 

from a rival’s R&D activity for the existence of knowledge spillovers. 

Knowledge spillovers capture the notion that innovative knowledge produced by R&D activities 

may, to some extent, be non-excludable so that firms can, at least partially, appropriate the results 

of R&D activities of a rival. Cohen and Levinthal (1989) suggest that such learning may not be 

costless and that firms may have to invest in own R&D programs in order to access external 

knowledge and benefit from spillovers from innovating rivals. 

Non-tournament models have especially been employed to assess the extent of innovation 

achieved by the operation of market forces and how it compares to the socially optimal one, to 

discuss technological policy, and to investigate the relationship between market structure (i.e. 

concentration) and innovation. 

However, non-tournament models assume that there are a large number of research paths and 

the relationship between R&D investments and rewards from innovation is essentially continuous. 

For this reason they are not well suited to describe market environments where innovation takes 

the form of a race and competition is for the market rather than in the market. 

Although this may be an appropriate description for technological change in some markets, in 

many others innovation is a more disruptive process that leads to winners and losers in very 

different competitive positions. In these circumstances, the rewards to R&D investments are 

13 

231



discontinuous, performance is rewarded on the basis of the rank within the set of realized 

performances, and competition takes the form of rival firms racing to be first to innovate 

(Dasgupta, 1986). An example of this type of race could be that characterizing competition 

between Intel and AMD to introduce better PC processors. 

Auction models 

Auction models of innovation describe R&D competition in terms of rival firms bidding for a given 

process or product innovation. The innovation is allocated to the firm which makes the highest bid 

and the issue is to investigate the identity of the firm which would have the larger willingness to 

pay for the innovation. 

In general, the maximum bid that a firm would make is equal to the difference between the profits 

it would get if it were successful in the auction and those it would get were it not. For this reason, 

auction models of innovation emphasize the competitive threat as the incentive underlying 

competition in innovation. 

Auction models have been widely used to investigate dynamic competition and whether market 

dynamics are characterized by persistent dominance by a technological leader or by “creative 

destruction” whereby the incumbent is overtaken by some rival whose incumbency is itself shortlived. 

In a single auction model, the issue of dynamic competition is investigated by considering a set of 

firms with different technological levels, i.e. marginal costs in the case of process innovation, and 

studying their incentives to purchase an innovation. Market persistence is the result of such 

competition if the winner of the auction is the technological leader, otherwise dynamic competition 

is characterized by action-reaction. 

Assessing market dynamics in a single-innovation auction model, however, may not help to 

understand dynamic competition because in these models firms do not take into account how the 

outcome of the current bid may affect the outcome of future auctions. These inter-temporal links, 

however, may be an important element to explain firms’ incentives to invest in R&D in some 

contexts where a sequence of innovative opportunities is in prospect. For this reason, the 

question of dynamic competition may be better posed in the context of sequences of innovation 

rather than of a single innovation. 

Auction models are intrinsically deterministic and this leads to the somewhat disappointing result 

that losers do not commit resources in R&D and that the process of technological advance is not 

really explained. 

14 

232



Stochastic tournament models 

Tournament models of innovation capture the particular interaction that arises in those markets 

where dynamic competition takes the form of a race to be the first to make a discovery. As 

opposed to deterministic auction models, tournament models consider the case where there is 

technological uncertainty. This can be captured, for instance, by considering a stochastic 

relationship between R&D effort and success. 

As such, these models emphasize timing in the context of R&D competition and are well suited to 

describe those environments where competition is for the market rather than in the market. This 

nature of market dynamics can arise because technological change proceeds along a small 

number of research trajectories and there is competition to be the first to innovate because 

success by a firm would somehow prevent appropriability by subsequent innovators. 

The advantage of being the first to innovate is usually explained on the basis of the ability of the 

winner to appropriate the value of the innovation by means of patent protection. More generally, 

however, appropriability may depend on factors other than patent protection, such as lead-time, 

access to complementary resources and secrecy. The relative importance of various 

appropriability factors is likely to be different across industries. (Levin et al., 1987) 

In auction models it is the competitive threat that characterizes firms’ incentives to invest in R&D 

activities. When randomness is introduced, however, both the profit incentive and the competitive 

threat are relevant in firms’ strategic interaction. 

In general, these forces need not be symmetric and have to be considered for each firm involved: 

one firm may face the greater competitive threat and another may have the greater profit 

incentive. Clearly, if a firm faces both the greater profit incentive and competitive threat it would 

undertake more R&D and hence be more likely to win the race. However, where there are 

asymmetric incentives, the nature of the outcome is not as straightforward. 

The literature usually focuses on the case where the competitive threat exceeds the profit 

incentive for all firms involved in the race. However, it could be the case that the profit incentive 

for both firms exceeds the competitive threat and that competition in R&D takes the form of a 

waiting game where firms tend to behave as free-riders and invest little resources in R&D. This 

type of situation has been considered, for instance, by Katz and Shapiro (1987) who explicitly 

account for the possibility of imitation and licensing of innovation. 

An important distinction among tournament models is related to the assumption on the nature of 

R&D expenditures and whether these are contractual (e.g. Dasgupta and Stiglitz, 1980, Loury, 

1979) or non-contractual (e.g. Lee and Wilde, 1980; and Reinganum, 1981,1982). In the first 

case R&D outlays take the form of a lump-sum cost incurred at the outset; in the second case 

R&D can be considered as a flow — a cost that it is incurred until the successful innovation is 

achieved. 

Beath, Katsoulacos and Ulph (1994) discuss the implications of these different assumptions. 

When R&D is of the contractual kind, equilibrium investments can be explained on a relatively 

15 

233



simple marginal benefit/marginal cost condition. On the other hand, when R&D is of the noncontractual 

kind, it is the competitive threat that is the main determinant of a firm’s R&D. 

As with auction models, tournament models have also been developed to consider a sequence of 

tournaments in which firms complete a number of R&D stages before competing in the market. 

Examples of models that fall in this category are Fudenberg et al. (1983), Harris and Vickers 

(1987) and Grossman and Shapiro (1987). 

In these models innovation is obtained only as the result of the completion of a number of stages. 

A crucial issue that arises in such a setting is the nature of the innovative process and in particular 

whether each firm has to discover each technological step itself before it can move on to discover 

the next or whether instead the follower firm can compete directly with the leader for the new 

state-of-the-art technology. 

The latter case may arise if knowledge spillovers are relevant so that a laggard firm has access to 

(but may not have right to use) the same technological knowledge of the technological leader. On 

the other hand, knowledge may be excludable to some extent or in-house R&D is very important 

to assimilate external knowledge so that the technological frontier has to be achieved 

independently by each firm. 

The Grossman and Shapiro, Park and Harris and Vickers step-by-step models assume that each 

firm has to discover each technological step itself before it can move on to discover the next. 

Reinganum (1985) and Beath et al. (1987) are examples of leapfrog models where a follower is 

allowed at any time to compete directly with the leader for the new best-practice technology. 

In a step-by-step model firms can compete neck-to-neck or have different technological levels. In 

a leapfrog model firms are always asymmetric in terms of their technological ability. This 

distinction turns out to be crucial in the analysis of the relationship between competition and 

innovation, as recent contributions in the endogenous growth literature surveyed below have 

shown. 

Endogenous growth theories have further developed these non-tournament models by 

considering sequences of tournaments where firms compete to innovate and compete in the 

product market before starting the following race to innovate. 

Market structure and innovation 

A long-standing question in the economics of technological change has been the nature of the 

relationship between market structure and innovation. 

The relationship between concentration of an industry and its rate of technological innovation is 

complex. Market structure may have an impact on the rate of innovation, but innovation is also 

an important factor that shapes market structure. In fact, it is necessary to recognize that the 

16 

234



relationship between concentration and innovation is not a causal one: both are the endogenous 

outcomes of the operation of market forces and exogenous factors such as the nature of 

demand, technological opportunity, the conditions governing appropriability, and pure chance. 

The classical point of departure in the economic analysis of the relationship between static market 

structure (i.e. concentration) and innovation is the work by Arrow (1962). Arrow considers the 

case where a cost-reducing innovation is exogenously available and investigates firms’ 

willingness to pay for the innovation under different market structures. For a drastic innovation, 

Arrow’s analysis shows that a firm that is already a monopolist would have lower incentives to 

innovate than a firm that is currently in a perfectly competitive environment, essentially because it 

would have the lower profit incentive. 

On the other hand, innovation is certainly an important factor that affects market structure: 

innovation is a means by which a firm tries to escape the constraints imposed by competition. 

Studies in the Schumpeterian tradition have emphasized the importance of ex-post market power 

for firms’ incentives to innovate. Some degree of market power is necessary for a firm to cover its 

R&D outlays: dynamic and static efficiency are somehow conflicting. This is a theme that has 

been well developed in the recent literature on endogenous growth. 

Dasgupta and Stiglitz (1980a) discuss the relationship between concentration and innovation in a 

non-tournament model. If there are exogenous entry barriers, an increase in the number of firms 

causes each firm to spend less on R&D in equilibrium, however total R&D expenditure increases 

with the number of firms. When entry is considered to be endogenous, one would observe more 

innovation in those industries that are characterized by a higher degree of monopoly power, 

although no causality should be imputed to this relationship. 

Sutton (1998) adopts an innovative approach to the study of technology and market structure and 

achieves a very general, albeit somehow loose, result on the relationship between concentration 

and R&D. He is able to characterize, under very general assumptions, a lower bound to market 

concentration (in large markets) that holds in those industries where the effectiveness of R&D in 

raising consumers’ willingness to pay is high and the product groups within the industry are 

sufficiently close substitutes. By contrast, in those industries where R&D is not effective or where 

product groups are not close substitutes, very fragmented market structures can also be 

equilibrium configurations no matter how large the market size is. 

The models discussed above consider the relationship between product market structure (i.e. 

concentration) and innovation. In the context of models where firms race to innovate another 

interesting question concerns the relationship between the number of firms that are part of the 

race and the pace of technological advance. 

17 

235



Loury (1979) considers a tournament model of innovation where R&D expenditures are 

committed upfront (that is, the probability of success depends on the scale of the R&D lab) and 

shows that increasing the number of firms reduces the expected date of invention. Lee and Wilde 

(1980) consider the case where the probability of innovation is related to the research intensity 

and show that as the number of firms in the industry increases, the equilibrium rate of investment 

per firm increases and success by any one firm is hastened by an increase in the number of 

firms. However, the first success date is hastened by an increase in the number of firms in both 

models. 

Reinganum (1982) considers a setting where firms can vary the research intensity as in Lee and 

Wilde (1980) but does not assume that the rate of expenditure is constant over time. Instead, 

firms may adjust R&D intensity in response to elapsed time and rival progress. Reinganum 

shows that, in this setting, when imitation is not possible, an increase in the number of firms 

increases the equilibrium rate of investment for each firm and decreases the expected time of 

innovation. When there is no full patent protection, the relationship is ambiguous and depends on 

the relative payoffs to the innovator and the imitators. 

The persistence of dominance 

In many industries characterized by fundamental long-term market dynamics competition may 

take the form of competition for the market rather than competition in the market. 

In these markets the issue is not whether more or less concentration is associated with faster 

technological progress but whether market dynamics would be characterised by persistent 

dominance of the incumbent leader or by action-reaction whereby incumbents are overtaken by 

some rivals whose incumbency is itself short-lived. 

The dynamic evolution of market structure depends on both abilities and incentives of the 

incumbent and the rivals to innovate. Game theoretical models are well suited to analyse the 

incentives underlying R&D investments and the resulting evolution of market structure. If we 

focus on economic incentives, and set aside differences in R&D abilities, persistence of monopoly 

or action-reaction can be related to the different profit incentive and competitive threat faced by 

the leader and the follower(s). 

Gilbert and Newbery (1982) use the auction model to examine potential competition at the R&D 

stage and reverse Arrow’s (1962) result: potential competition matters and a monopolist may 

have more incentives to innovate than a potential entrant because the incumbent monopolist may 

face a larger competitive threat. 

The profit of a successful incumbent who innovates is that of a monopolist firm, whereas if it were 

the entrant to innovate each firm would get the profit of a duopolist. Hence, the competitive threat 

of the incumbent can be measured as the difference between the profit of a monopolist firm and 

the profit of a duopolist firm. The incumbent’s competitive threat, instead, is simply equal to the 

profit of a duopolist. This implies that the incumbent would have more incentives to innovate (i.e. 

the larger competitive threat) if, as it is normally the case, the profit of a monopolist is greater than 

the combined profits of two duopolists. 

18 

236



Katz and Shapiro (1987) study a general auction model and show that for minor innovations, the 

industry leader will typically be the innovator, whether or not imitation and licensing are feasible. 

For markets where patent protection is strong, they predict that the major innovations will be 

made by industry leaders. But if imitation is easy, industry followers or entrants will make the 

major discoveries. 

Reinganum (1983) in a tournament model shows that when the first innovator captures a 

sufficiently high share of the post innovation market, then the incumbent firm invests less on a 

given project that does the potential entrant. This is because the incumbent has less incentive 

than the challenger to shorten the period of its incumbency. 

Beath, Katsoulacos and Ulph (1995) show that in a tournament model under Cournot 

competition, a large innovation or a large initial gap results in persistent dominance while a small 

initial gap and a small innovation results in action-reaction. However, if market competition is 

Bertrand there is persistent dominance irrespective of the size of the innovation or the initial gap. 

This is a result that points towards the possible trade-off between static and dynamic efficiency 

suggested also by Vickers (1986). 

Market dynamics have also been investigated in models that consider a sequence of innovations. 

Reinganum (1985) considers a sequence of drastic innovations and shows that market dynamics 

are characterised by a process that resembles Schumpeter’s process of creative destruction: the 

incumbent invests less than each challenger in each stage. 

Vickers (1986) considers a sequence of non-drastic process innovation in the context of the 

auction model. He compares market dynamics under Bertrand and Cournot competition and 

finds that when the product market is very competitive then there is increasing dominance, but 

when it is not very competitive there is action-reaction. When innovation is drastic, then market 

dynamics are characterised by increasing dominance. 

Reinganum (1989) observes that the differences in the results obtained in different models can be 

ascribed to the different roles that the profit incentive and the competitive threat play in stochastic 

tournament and deterministic auction models. In a deterministic model the incentive to pre-empt 

(larger for the incumbent) dominates the firms’ decision. When success is stochastic, however, 

the threat from the rival innovating is less acute. In the case of drastic innovations, the 

competitive threat is the same for both firms and it is the profit incentive (which is larger for 

challengers) that determines R&D investments. The relevance of the profit incentive extends to 

the case for some non-drastic innovation. 

Competition intensity and innovation 

Market structure is often associated with the concept of competitiveness: usually, a high level of 

concentration in an industry is interpreted as weak competition. This view is based on a 

symmetric Cournot model, where price-cost margins are higher as the number of firms increases. 

However, it is preferable to disentangle the notions of market structure and toughness of price 

competition as, among others, Sutton (1998) and Boone (2000, 2001) have done. 

19 

237



In a simple world with homogeneous firms, toughness of price competition can be considered as 

being related to the level of price-cost margins given any level of concentration. Hence, a 

differentiated Bertrand market would be considered to be more competitive than a differentiated 

Cournot market because price-cost margins would be lower in the former, for any level of 

concentration. As a result, more competitive markets may allow fewer firms to profitably survive. 

Recently, Boone (2000,2001) has offered an interesting contribution to the study of the 

relationship between toughness of competition and innovation. Boone defines intensity of 

competition on the basis of an axiomatic approach, which allows the result to be general to a 

large class of specific game-theoretical models. Boone considers four axioms: 

• 

• 

• 

• 

the lower limit of competition is such that firms are not affected by opponents’ actions; 

the least efficient firm in the market loses as competition becomes more intense; 

if the leader is far enough ahead, he gains as competition becomes more intense; and 

if all active firms have similar costs, they all lose as competition becomes more intense. 

Boone (2001) shows that this definition encompasses, for instance, a switch from Cournot to 

Bertrand competition and a reduction in travel costs in typical horizontal product differentiation 

models. 

When the notion of intensity of competition is defined in such a way, it is normally inversely related 

to equilibrium concentration. For instance, in Boone’s setting, if intensity of competition is low, 

then a large number of firms can be active in the market because the less efficient firms can 

survive as well. On the other hand, fierce competition is associated with high concentration 

because only the most efficient firms can survive. 

On the basis of this definition of product market competition, Boone (2001) investigates the 

relationship between intensity of competition and firms’ incentives to invest in R&D. The central 

finding of this study is that the relationship between market competition and innovation may not 

be monotonic; this is so because varying intensity of competition is associated with different 

identity of the innovator. 

Boone (2001) explains this interesting result as follow. When intensity of competition is low it is 

the follower that would be the next innovator, whereas when intensity of competition is large, it is 

the current technological leader that is likely to innovate. When it is the follower that innovates, 

tougher competition implies lower profits and hence lower incentives to innovate. However, when 

it is the current leader to innovate, an increase of toughness of price competition would further 

increase the profits related to his technological leadership and hence would increase the value of 

innovation for the firm. Hence, the relationship between toughness of price competition and 

innovation would be U-shaped. 

Denicolò (2002) observes that Boone’s paper generalises the findings of a number of other 

studies. Delbono and Denicolò (1990) considered a homogeneous product market and found 

that Bertrand duopolists have greater incentive to innovate than Cournot duopolists. Bester and 

20 

238



Petrakis (1993) and Bonanno and Haworth (1998) showed respectively that horizontal and 

vertical product differentiation could reverse this result. Qiu (1997) compared the two regimes in 

a non-tournament model and find that is such a framework the incentive to innovate is greater 

with Cournot competition even in the case of homogenous products. 

Other studies that investigate the relationship between toughness of competition and innovation 

fall within the endogenous growth literature ,Aghion and Griffith (2005) and are surveyed in the 

relevant section of this 

review. 

R&D Co-operation 

Increasingly, research joint ventures (RJVs) have become a widespread form of industrial cooperation. 

In parallel, the study of the effects of R&D co-operation has also emerged as an 

important research topic when considering industry policy. 

Co-operative R&D ventures can take many forms, ranging from R&D joint ventures, to crosslicensing 

agreements and various informal types of technology trading or information-sharing 

agreements. 

Research Joint Ventures themselves can cover a variety of arrangements. One type of RJV is 

the traditional joint venture, in which two or more parties create a separate entity in which they all 

have equity interests to conduct well-defined R&D projects for their benefits. Another type is the 

research consortium. A third form of RJV is the venture capital investment by firms in a standalone 

start-up company. 

Economists have investigated the extent to which RJVs might allow firms to internalize spillovers, 

co-ordinate their research activities and achieve higher R&D efficiency. Co-operative R&D is 

thought to be socially beneficial for several reasons: 

• 

• 

• 

RJVs can alleviate the under-provision of R&D effort that results from technological 

spillovers and other sub-optimal R&D decisions. 

Co-operation can lead to greater dissemination of R&D results. It can improve R&D 

efficiency through good research design and information sharing, avoiding needless 

duplication of resources. 

Co-operative R&D enables the firms to share risks, to exploit synergies, pool different 

complementary assets, and exploit increasing returns to scale in R&D. It can enable firms 

to overcome a cost-of-development barrier impenetrable to any one of them alone. 

Adding to these effects, an important issue that is not very developed in the industrial organization 

literature but is emphasized in the evolutionary economics and the management literature on 

21 

239



inter-firm co-operation is that cooperative ventures may be an important means by which firms 

exchange specialised know-how and tacit knowledge. 

Countering these social benefits is the fear that firms participating in co-operative R&D might use 

the venture to engage in anti-competitive practices or that they might free ride on each other. 

RJVs and R&D investments 

Since the early contributions to the literature on RJVs, the effects of R&D co-operation on the 

amount of resources invested in innovative activities have been extensively investigated. 

This attention is immediately related to the study of the potential of RJVs to correct the suboptimal 

R&D decisions that may characterise independent decision-making. 

The results of these studies accord support to the presumption that RJVs, by internalising 

externalities which affect the efficiency of non-cooperative decision-making, can help to correct 

the non-optimality of independent investments in R&D. In explaining what impact RJVs may 

have on R&D investments, knowledge spillovers are widely acknowledged as a crucial factor, and 

their consideration plays an important role in most economic models of RJVs. 

In their seminal paper, d’Aspremont and Jacquemin (1988) pioneered an influential approach to 

study the effects of co-operation on R&D investments. They consider a two-stage model in which 

duopolists first conduct partially inappropriable research leading to a reduction in unit cost, and 

then compete à la Cournot in the product market. The focus of their analysis is on the 

comparison of the magnitude of cost-reducing technical advance achieved when firms conduct 

R&D competitively versus co-operatively, in the presence of spillover effects. 

The difference between the independent decision-making equilibrium and the co-operative one 

results from the balance of the spillover internalisation effect and the reduced strategic incentive 

effect. Joint decision on the levels of R&D expenditures internalises two externalities disregarded 

under independent decision-making: 

• Research knowledge is partially inappropriable, and it can benefit other firms without the 

innovator being able to fully appropriate the value of the knowledge created. This 

externality causes sub-optimality of independent R&D investments. Its internalisation 

under joint-decision making (spillover internalisation effect) will tend to increase cooperative 

R&D investment relative to the non-cooperative level. 

• 

The innovative activity of a firm comprises also a negative pecuniary externality on the 

profits of the rival through a clear strategic market interaction. Joint decision-making 

internalises this externality (reduced strategic incentive effect) and tends to decrease the 

co-operative level of R&D investment relative to the non-cooperative equilibrium. 

The balance of these opposite forces is crucial to understand the effects of co-operation on the 

level of resources invested in research and, in turn, on the appropriateness of co-operation to 

reduce the gap between the private and social incentives for doing R&D. 

22 

240



In case of “high” spillover rates, the spillover internalization effect would dominate the reduced 

strategic incentive effect, the net effect being an increase in the R&D expenditures. By contrast, if 

spillovers are “low”, R&D co-operation would cause R&D investments to decrease relative to the 

non-cooperative equilibrium. 

The contributions of Kamien, Muller and Zang (1992), Suzumura (1992) Simpson and Vonortas 

(1994), have proposed extensions of the basic modelling framework to cover an arbitrary number 

of firm, general demand and cost assumptions. 

RJVs and R&D efficiency 

As we have discussed, considerable attention has been devoted to the study of the effects of 

RJVs on the amount of resources allocated to R&D, and of the advantages of technological cooperation 

in terms of correcting inefficient R&D efforts that independent decision-making involves. 

Another commonly acknowledged advantage of R&D co-operation is identified in the increased 

efficiency of R&D activities: RJVs not only can affect the absolute levels of R&D spending, but 

also the amount of R&D spending per unit of cost reduction/quality improvement achieved. 

Indeed, to the extent that co-operative R&D is more widely disseminated than individually 

conducted R&D, ex ante co-operation increases the efficiency of R&D efforts because a single 

investment benefits a greater number of firms. 

The efficiency gain can have three types of positive effects. First, sharing lowers the cost of 

investment for each firm, which may induce them to conduct more R&D. Second, for a given 

level of R&D investment it might increase the effective amount of R&D. Third, co-operative R&D 

can eliminate the wasteful duplication that would occur if several firms separately undertook the 

same projects. Even if several firms continue to conduct separate R&D under an ex ante 

agreement, they still can improve the efficiency of their efforts by co-ordinating them. 

Some studies have formalized these intuitions, by investigating the effects of co-operation on the 

level of information-sharing and the co-ordination of research activities. 

Katz (1986) presented an early, albeit isolated, study in which firms choose the level of spillovers 

in the RJV, by determining the amount of information-sharing achieved. However, the 

endogeneity of spillovers is restricted to the co-operative equilibrium only, and this limits the 

significance of the comparison of the (exogenous) non-cooperative and (endogenous) cooperative 

information-sharing levels. 

To fully exploit the endogenization of knowledge spillovers, and explain the differences of 

information-sharing levels in the co-operative and non-cooperative equilibrium it is necessary to 

23 

241



model both as the result of firms’ choices. Katsoulacos and Ulph (1998) have presented a 

stochastic model of innovation with this feature. 

The authors adopt a structured approach to model knowledge spillovers by distinguishing two 

relevant dimensions: the adaptability of the research to the other firm and the amount of 

information-sharing that actually takes place. 

They show that the non-cooperative equilibrium generally produces minimal spillovers; in this 

case, RJVs might lead to a higher information-sharing. However, if firms operate in different but 

complementary industries, it is possible to achieve maximum information-sharing also in the noncooperative 

equilibrium. 

Poyago-Theotoky (1999) analyses endogenous spillovers in the context of a simple nontournament 

model of R&D where firms are engaged in cost-reducing innovation. It is shown that 

when spillovers of information are treated as endogenous firms never disclose any of their 

information when choosing their R&D non-cooperatively. Under co-operative R&D, firms will 

always choose to fully share their information, i.e. a research joint venture will operate with a 

maximum spillover value. 

The relationship between the research paths followed by firms can be manifold, spanning a 

continuum from perfect complementarity to perfect substitutability. Co-operation in R&D can take 

advantage of joint research design, by choosing the number of labs to operate for optimally 

exploiting such substitutability or complementarity. 

The issue of organisation design in an RJV has been further considered by Katsoulacos and Ulph 

(1998) who relate this choice to the nature of research paths followed and the endogenous level 

of information-sharing chosen by firms. An important result of this study is that cost 

considerations, dominant when research outputs are close substitutes and firms are willing to 

share a lot of information, are only one of the factors behind the RJVs decision as to whether to 

operate one or two labs. 

Indeed, the RJV may close a lab for anti-competitive reasons, to prevent having to face a very 

competitive situation when both firms discover. Or, when there are very strong complementarities 

between the research outputs of the two firms, the RJV might prefer to keep both labs open in 

order to exploit such complementarities. 

1.3 Endogenous Growth Models 

Technological change is a fundamental driver of economic growth. However, only recently have 

the economic forces that underlie technological progress in a dynamic economy been 

investigated. Advances in this field have been associated with the development of endogenous 

growth theories, one of the most fertile grounds for economic research in the last 15 years. The 

relevant research agenda has been set on the study of economic growth as the result of 

knowledge production — we shall focus on technological knowledge in this survey — which is 

explained as the outcome of economic decisions. 

24 

242



Before the development of endogenous growth theories economic models of growth were 

essentially models of capital accumulation developed on the basis of Solow’s (1956) seminal 

contribution. These models suggested that the simple accumulation of physical capital cannot 

sustain long-run economic growth as long as its marginal productivity is decreasing and not 

bounded from zero. Long run economic growth had to be attributed to the effect of some other 

factor, exogenous to the economic choices analyzed in these models, such as technological 

change. 

In fact, even in older exogenous growth theories that studied capital accumulation, technological 

change was acknowledged as one of the major factors that explain the ability of an economy to 

grow in the long run. Certainly, advances in technology are a desirable element in any 

explanation of growth: “a story of growth that neglects technological progress is both ahistorical 

and implausible”(Grossman and Helpman,1994,p.26). Cliometric studies have also examined the 

relationship between knowledge 

and economic growth as shown in the historical background below. 

Nevertheless, for a long time the forces underlying advances were not explained on the basis of 

economic incentives but were essentially related to the exogenous evolution of scientific 

knowledge and its technological applications. 

Endogenous growth theories or new growth theory have certainly contributed to fill this gap. 

These theories are based on product differentiation and imperfect competition, economies of 

scale and increasing returns instead of perfect competition and diminishing returns put forward by 

neo-classical theory. Freeman (1994) argues that these theories have been considerably 

influenced by the neo-Schumpeterian school of innovation. 

New growth theory not only emphasizes knowledge accumulation and diffusion, in the form 

mainly of human capital (Lucas, 1988, Rebelo, 1991) or technological innovation (Romer, 1990; 

Aghion and Howitt 1992), as fundamental drivers of economic growth but also the economic 

decisions that underlie both. 

The model developed by Rebelo (1991) represents the first generation of endogenous growth 

models that underscore the importance of human capital accumulation. In this model, there is no 

exogenous productivity growth given that the entirety of output growth is explained by human 

capital and capital stock accumulation. 

As such, this review focuses on the second generation of endogenous growth models that 

consider technological innovation as well as human capital accumulation. In general, these 

models relate economic growth to innovations that improve firms’ productive efficiency and/or 

lead to the production of new or better intermediate and consumer goods. 

8 

9 

25 

243



Innovations result from new knowledge produced as the outcome of firms’ learning-by-doing 

and/or intentional R&D activities. In general, R&D models of endogenous growth have drawn 

heavily from the earlier game theoretical literature on innovation and extended those analyses to 

more dynamic settings.As such, they have developed the analysis of long-standing questions 

investigating, for instance, the relationship between the intensity of competition and innovation 

and the nature of market dynamics. 

Recent economic analysis in this field has offered interesting new insights on the relationship 

between competition and innovation. Despite some suggestive conclusions, however, 

endogenous growth models suffer from the same limitations as their constituent micro 

foundations: results often depend crucially from the assumptions adopted and the specific 

structure of the model studied. Unfortunately, in many cases some of these important 

assumptions relate to features of market environments that are hard to observe and evaluate. 

Economic theory provides no general results that can be considered to hold in any market situation, 

but rather specific considerations that suggest which features of a market environment are worth 

considering to assess the nature of competition, market dynamics ,their relationship and growth. 

(Aghion and Griffith, 2005). 

Learning-by-doing 

Knowledge generated by learning-by-doing can be a significant engine of economic growth and 

models of endogenous growth with technological progress driven by learning-by-doing have been 

studied by Romer (1986), Young (1991,1993), Stokey (1988) and Aghion and Howitt (1998). 

Learning-by-doing, since the seminal work of Arrow (1962), Solow (1997) has been considered an 

important source of technological knowledge. Firms can produce new knowledge without investing 

in institutionalized research and development activities but rather, as a by-product of their usual 

production activity. Indeed, experience has been shown to be an important source of 

technological improvements in many industries such as artificial fibres, semiconductors and 

memory chips. 

26 

244



Generally, in these models learning-by-doing is the unintentional by-product of production or 

investment activities and it is usually considered to result in technological knowledge that is only 

partially appropriable. The existence of knowledge spillovers is in fact generally acknowledged: a 

firm can, in part at least, benefit of the knowledge generated by the learning-by-doing of other 

firms. 

Young (1991, 1993) studies a setting where new technologies are discovered through R&D 

activities but are initially inferior to the older technologies they seek to replace. Experience, 

however, generates incremental improvements over time, which allow the new technology to 

supplant older technologies. These incremental improvements are only achievable to some 

extent, in the sense that there is a bound to the technological advance that can be generated by 

learning-by-doing. This hybrid model emphasises that inventive activity and production 

experience are complementary forces in driving technological change. 

R&D models 

Endogenous growth models have also investigated economic growth as driven by the 

accumulation and diffusion of knowledge generated by Research and Development activities, 

undertaken by profit-maximising firms. 

These models investigate the process of discovery of new or better products. Accordingly, a 

distinction can be drawn between those models that treat innovation in terms of the development 

of new varieties of products and those that consider sequential improvements of the quality of 

existing products. 

The major difference between these two types of models is that the latter capture an important 

characteristic of the innovative process, namely that new inventions entail an element of 

destruction because they make old technologies or products obsolete. For this reason, early 

endogenous growth models with sequential improvements in the quality of products are called 

Schumpeterian. 

Despite these differences however, the two types of models share the same fundamental 

structure. In both models, firms invest resources to acquire the exclusive right to manufacture a 

new product and it is usually assumed that R&D activities generate knowledge spillovers that 

benefit other research firms. The production of new knowledge and its diffusion drive economic 

growth in the long run. 

Variety of products 

Romer (1990) provides the seminal model where technological progress is captured in terms of 

an increasing variety of products. The increase of the variety of intermediate or final goods is 

reflected in higher productivity or utility and hence economic growth. 

27 

245



New varieties of products are discovered as outcomes of R&D investments and remunerated by 

the rents that accrue to the successful innovator who can benefit from patent protection (or, more 

generally, of other appropriability factors such as lead time, secrecy, access to complementary 

resources). 

In these models, it is generally assumed that R&D activities lead both to new knowledge that can 

be appropriated by the innovator and new general knowledge that cannot be appropriated but 

instead contributes to a stock of publicly available knowledge that facilitates further discoveries. 

Not only is the production of new knowledge necessary for economic growth, but also its 

diffusion. 

A distinguishing feature of Romer’s (1990) growth model is that it studies horizontal product 

innovations that involve no obsolescence: new goods are never close substitutes for existing 

goods (i.e. each new product finds its own horizontal niche). 

Although this feature of the model may detract from its realism is some cases, it is not necessarily 

so. There are some industries where innovation is fundamentally directed at the introduction of 

new varieties of products rather than at improving products’ qualities. An example discussed in 

the literature is the flowmeter industry where technological advance takes the form of the 

development of new types of flowmeters. In this industry there is no escalation of R&D on a 

single technological trajectory and related product group but proliferation of product groups. 

Quality of products 

Aghion and Howitt (1992) examine growth as driven by industrial innovations that improve the 

quality of products. 

Vertical innovation models of this kind introduce, relative to models with expanding varieties of 

products, the consideration of the process of creative destruction: a new innovation makes, to 

some extent, old products obsolete and negatively affects the economic rents of previous innovators. 

For this reason, models with qualitative improvements are also called Schumpeterian because they 

embody Schumpeter’s idea of creative destruction by emphasizing the process by which new 

products (and innovators) displace old products (and innovators). 

Endogenous growth models with vertical innovations have very elaborated micro-foundations 

describing the nature of market dynamics, which are similar to and develop further earlier gametheoretical 

models of innovation. As such these models have developed the study of dynamic 

competition. 

28 

246



Aghion and Howitt (1992) study a model of economic growth based on creative destruction. 

Innovation consists of a higher quality intermediate good that can be used to produce output 

more efficiently than before. Research firms invest resources in R&D activities that stochastically 

determine the time of the new innovation; the arrival rate of new innovations depends only upon 

the current flow of input to research. Firms’ R&D investments are motivated by the prospect of 

monopoly rents that can be captured when a successful innovation is patented. However, unlike 

the horizontal innovations considered in Romer (1990), these rents would last only until the next 

innovation occurs, when the current innovation would become obsolete. 

Aghion and Howitt consider the cases of both drastic and non-drastic innovation. In the first case 

the new innovator is unconstrained by potential competition from the previous patent; in the latter 

case, such a constraint is binding. 

In the case of drastic innovations market dynamics take the form of a continuous process of 

action and reaction whereby at each point in time the market is dominated by a monopolist whose 

incumbency is short lived. In theory, a new innovation could be introduced by either the current 

monopolist or an outside research firm. The value of innovation is the expected incremental 

present value of the flow of monopoly profits generated by the innovation over an interval whose 

length depends on the occurrence of the next innovation. Because this incremental profit is lower 

for an incumbent monopolist, only outside research firms would innovate and the monopolist 

chooses to do no research. 

The reason for this result lies in the fact that, given the assumption and the structure of the model, 

firms’ incentives are driven only by the profit incentive, which in the specific case is always larger 

for an external firm. This is the replacement effect emphasized by Arrow (1962) whereby a 

monopolist would have lower incremental profits from innovation than a perfectly competitive firm. 

In the case of drastic innovation the competitive threat does not play a role because the flow of 

profit is independent of the identity of the innovator. 

Innovations are non-drastic if the previous incumbent could make a positive profit when the 

current one is charging the unconstrained optimal monopolistic price. In the case of non-drastic 

innovation the competitive threat may potentially play a role in shaping equilibrium investments in 

R&D activities. However, Aghion and Howitt assume that the monopolist chooses to do no 

research also in the case of non-drastic innovations, i.e. assume that the efficiency effect is 

smaller than the replacement effect. As a result market dynamics take the form of continuous 

action/reaction. 

Finally, Aghion and Howitt consider the case in which firms can affect not only the frequency but 

also the size of innovations. Their finding is that innovations would be too small if they were 

drastic; in the non-drastic case, the tendency to make innovations too small is in part mitigated by 

the incentive for innovators to move away from their competitive fringe. 

Grossman and Helpman (1991) have built on the model by Aghion and Howitt to consider an 

economy where a continuum of, and not only one, intermediate goods. Each product has its own 

quality ladder and entrepreneurs target individual products and race to bring about the next 

generation. In each industry, success occurs with a probability per unit of time that is proportional 

29 

247



to the total R&D resources invested to improving that product. The authors allow free entry in the 

race for the next generation of products and assume that potential entrants can learn enough 

about the state of knowledge to compete to produce the new state-of-the-art quality. These 

assumptions imply that, without a cost advantage, industry leaders do not invest resources to 

improve the quality of their own state-of-the-art products. This is because of the replacement 

effect whereby the leader would obtain an incremental flow of profits that is less than that of an 

external firm. 

Aghion and Howitt (1998) introduce an element of heterogeneity in innovative activity that 

captures the distinction between fundamental and secondary research. They first investigate this 

heterogeneity by considering fundamental research as deriving from R&D activities and 

secondary innovation from learning-by-doing. Each innovation resulting from research consists of 

a potential new product, and learning-by-doing leads to improvements of the quality of goods that 

have been invented. They also consider the possibility that learning-by-doing contributed to a 

general stock of knowledge that creates new opportunities for research as well. Moreover, 

Aghion and Howitt (1998) consider both the case where learning is not appropriated by the firm 

but is shared by all firms and where the quality enhancement of learning is fully internalized. 

Finally, they consider the case where the distinction between fundamental and secondary 

innovations is captured in terms of research versus development. 

A feature of Aghion and Howitt’s (1992) influential model, shared by most early models of 

endogenous growth with quality ladders and leapfrogging, is that more competition (i.e. higher 

elasticity of demand for each firm) is associated with less innovation and growth, a very 

Schumpeterian result. The reason for this conclusion is that these models capture only an 

increase of ex post product market competition and not of ex ante competition as well. Some 

recent studies modify this feature of earlier models and reach a different conclusion on the 

relationship between competition and innovation: tougher competition may increase the 

incentives of firms to innovate in order to escape from a more competitive market state. 

In the model developed by Aghion, Harris and Vickers (1997) and subsequently analysed by 

Aghion et al. (2001) and Aghion et al. (2002), the study of the relationship between product 

market competition and innovation is carried out in the context of a model of endogenous growth 

with vertical differentiation in a duopolistic industry where the incumbent firm can innovate. 

The model considers a sequence of tournaments with step-by-step innovations: the technological 

follower has first to catch up with the leading-edge technology before being able to become the 

new technological leader. This assumption can be justified on the grounds that innovative 

knowledge is to some extent tacit so that a rival has to invest in its own R&D to catch up and 

cannot simply rely on knowledge spillovers. 

Aghion, Harris and Vickers (1997) focus on the comparison between Cournot and Bertrand 

competition to capture different levels of market competition. Aghion et al. (2001) and Aghion et 

al. (2002) use the same model to investigate the relationship between innovation and competition 

where the latter is captured in terms of a substitutability parameter at the product market 

competition level. 

30 

248



The substantial difference of these models compared to earlier Schumpeterian models lies in the 

fact that, by considering the possibility that the incumbent innovates, it is the incremental profits of 

each firm that drive R&D investments and the impact of product market competition has to be 

evaluated both ex ante and ex post. Tougher product market competition might reduce a firm’s 

pre-innovation rents by more than it reduces post-innovation rents, hence it may increase R&D 

firms’ investments. 

In these models, technological progress does not involve leapfrogging. Hence, a typical industry 

can be in either of two states: it can be a leveled industry where firms compete neck-to-neck or 

an unleveled industry where one firm is technologically ahead of the follower. 

The effect of an increase in product market competition would have to be evaluated considering 

the impact in both leveled and unleveled industries. The impact is actually twofold: on the one 

hand the degree of product market competition affects equilibrium R&D investments in each of 

the possible state of the industry (leveled and unleveled); additionally, product market 

competition has an impact on how often the industry will be in one of either states. 

In a leveled industry increased competition would spur innovation because it would increase the 

incentives of firms’ to get ahead of the rival (the new “escape competition effect”); in an unleveled 

industry, increased competition would reduce the incentives to innovate for the classic 

Schumpeterian effect. 

The reasoning in Aghion et al. (2002), which leads to the conclusion that the relationship between 

intensity of competition and innovation has an inverted-U shape, is discussed below. 

Consider the case of very large innovations, in which case the leader never innovates and the 

largest gap between the leader and the laggard is one technological step. 

When there is not much product market competition, there is hardly any incentive for firms to 

innovate when they compete neck-to-neck, and the overall innovation rate will be highest when 

the industry is in an asymmetric state. Thus the industry will spend relatively more time in the 

levelled state where the escape-competition effect dominates. An increase in product market 

competition would then result in larger incentives for the firms to innovate. Hence, if the degree of 

competition is initially low, an increase will result in a faster innovation rate. 

When initial product market competition is high, there is relatively little incentive for the laggard 

firm in an unleveled state to innovate and a relatively large incentive for a neck-to-neck firm to 

leave the leveled state. The consequence is that the industry will spend most of the time in the 

unleveled state where it is the Schumpeterian effect that dominates. Tougher product market 

competition reduces the Schumpeterian effect so that when the degree of competition is initially 

high, an increase may result in a slower average innovation rate. 

These results extend to the general model with no upper bound to the technological gap. 

Aghion et al. (2002) also provide empirical evidence that support their claim of an inverted-U 

relationship between intensity of competition and innovation. 

31 

249



Aghion et al. (2001) use the same framework to investigate the relationship between imitation and 

innovation and conclude that this is also characterized by an inverted-U shape. Again this may 

be the case because easier imitation may decrease a firm’s pre-innovation rents by more than it 

reduces post-innovation rent. 

Denicolò and Zanchettin (2002) consider the relationship between competition and growth in the 

case of non-drastic innovations by comparing innovation and growth under Bertrand and Cournot 

competition in a standard leapfrogging model. The main finding of this study is that: 

“[…] when the size of innovations is sufficiently large, the equilibrium rate of growth is 

unambiguously greater with Bertrand than with Cournot competition. For smaller 

innovation, Cournot competition may (but need not) create greater incentive to innovate. 

The intuition is that more competition entails lower prices but at the same time leads to greater 

productive efficiency (hence greater rates of growth) in that it lowers the market shares of less 

productive firms (Aghion and Shankerman, 2000). When firms are symmetric only the first effect is 

at work. When technical progress creates asymmetries between firms the productive efficiency effect 

becomes important and the effect of more competition on innovators’ profits is generally ambiguous. 

However, the productive efficiency effect must dominate if innovations are close to being drastic. 

D’Aspremont et al. (2002) investigate the relationship between competition and innovation in a 

tournament model under uncertainty where, unlike Aghion et al. (2001), the number of firms in 

each industry is endogenously determined and the race entails the possibility of several 

simultaneous winners and spillover effects. The main conclusion of the paper is that the 

relationship between toughness of market competition and the incentives to innovate is nonmonotone. 

The key for this result lies in the possible multiplicity of successful innovators in the 

same period. 

Introduction of Agency Considerations 

The adoption of a step-by-step technological assumption is not the only means by which the 

relationship between competition and growth takes a different nature that in the traditional 

Schumpeterian model. Aghion and Howitt (1998) consider also an alternative approach, based 

on a study published later by Aghion, Dewatripont and Rey (1999) that leads to conclusions on 

the relationship between competition and growth that are different from those of the classic 

Schumpeterian model. 

This approach is based on relaxing the traditional behavioural assumption that firms are profit 

maximizing. Non-profit maximizing behavior is a common assumption in the literature that 

emphasises the existence of agency problems between producers and their financiers. Aghion et 

al (1997) consider a principal-agent model of the firms where managers are interested in 

maximizing some private benefits of control while minimizing private costs that innovation would 

entail such as the costs of reorganizing the firm or training costs. 

32 

250



Within this new setting, the impact of intensity of competition on growth can be positive because 

tougher competition may lead to managers to speed up the adoption of new technologies to avoid 

bankruptcy and the loss of control rights and hence to faster economic growth. 

The literature on the importance of competition for incentives is extensive and it suggests that 

there are other important mechanisms though which competition may have a positive impact on 

incentives is through the improvement of the ability of assessing managers’ efforts and hence of 

monitoring them. At the same time, similarly to other areas of research, theoretical predictions of 

these studies are not as immediate as these simple considerations may suggest. 

Nickell (1996) provides empirical evidence on the relationship between intensity of competition 

and innovation by studying the impact of competition on both the level and the growth of total 

factor productivity of a large number of UK manufacturing companies. The main findings of this 

paper are that 

• 

• 

market power, which is captured in terms of market share, generates lower levels of 

productivity; and that 

competition, which is captured by the number of competitors or by lower levels or rents, 

leads to faster total factor productivity growth. 

Blundell, Griffith and van Reenen (1999) also study the relationship between competition and 

innovation. Their study provides evidence that 

“[…] within industries in was the high market share firms who tended to 

commercializemore innovation although increased product market competition in the 

industry tended to 

stimulate innovative activity” (p. 550) 

For a very comprehensive survey of empirical studies on competition and innovation see the 

recent OECD study by Ahn (2002). 

1.4 New Economic Geography 

New economic geography was effectively initiated by Krugman (1991). The main goal of this 

strand of research is to provide a theoretical explanation of why economic activity and population 

tend to concentrate, that is to say, why cities and industrial belts come into existence and keep on 

expanding at the expense of the periphery. This approach is similar to new growth theory in that 

the models assume imperfect competition and product differentiation, economies of scale and 

increasing returns. 

33 

251



The underlying story is that manufactured goods will be produced in regions where nearby 

demand is higher, as this allows firms to minimize transport costs. Demand will come either from 

agriculture or from the manufacturing sector itself (manufacturing workers). Because of a 

possible positive feedback between production of and demand for manufactured goods, 

manufacturing production will tend to relocate into these regions. 

This phenomenon is formalized by the canonical modelling of core and periphery (Krugman 

(1991)) that assumes that the economy is split into two sectors - namely manufacturing and 

agriculture – and into two regions (the core and periphery). In manufacturing, monopolistic 

competition à la Dixit-Stiglitz (1977) prevails: manufacturing produces differentiated products in 

the presence of increasing returns. In addition, labor, specific to manufacturing, is mobile 

between the two regions. Wages, both in nominal and real terms, can differ between the two 

regions. The driving force behind labor mobility is indeed differences in real wages. 

Transportation costs when shipping from one region to the other occur in the form of Samuelson’s 

iceberg cost (Samuelson (1954)). This means that transport costs are not explicitly considered 

but rather it is assumed that only a fraction of goods arrives, with the difference “smelting down” 

during transportation. 

By contrast, agriculture is characterized by perfect competition with homogenous goods and 

constant returns to scale. Wages equal prices and labor, specific to agriculture, is immobile 

between the two regions. Furthermore, no transportation costs are assumed to occur when 

goods are moved from one region to the other. Assuming constant returns to scale and the 

absence of transport costs secures wage equalization in agriculture. It should be also noted that 

agriculture is evenly divided between the two regions. 

For concentration in manufacturing to take place in one of the regions, several conditions have to 

be fulfilled. First, a large share of population should be allocated to the manufacturing sector so 

as it generates large demand (that triggers the virtuous circle). Second, economies of scale 

should be high enough and transportation costs low. 

Three cases can be considered. 

1 First, in the event that transport costs are large, there is an equal division of 

manufacturing between the two regions. 

2 

3 

If transportation costs are low, manufacturing still can remain equally divided between 

regions. However, this situation is not a stable one (unstable equilibrium) for the following 

reason. If manufacturing were only slightly more important in one of the regions, this 

would lead to an increasing share of manufacturing in this sector accompanied by a 

dwindling of manufacturing in the other region. At the end of this process, manufacturing 

ends up being concentrated in one of the regions, and this can be referred to as the coreperiphery 

pattern. 

The third case, characterized by intermediate transportation costs, can lead to either 

complete specialization or equal division between regions. 

34 

252



Extensions 

The three-region model 

A possible extension of the basic model is to simultaneously consider three regions. Keeping the 

assumptions of the two-region model would yield basically the same results: No concentration of 

manufacturing activity in the presence of high transport costs, extreme concentration in one of the 

three regions if transport costs are low, and the combination of the two previous situations in the 

case of intermediate transport costs. 

Transport costs in agriculture 

The second extension of the basic two-region model consists in releasing the assumption related 

to the absence of transport costs in agriculture. Empirical evidence clearly indicates that transport 

costs exist not only for differentiated goods produced in the manufacturing sector, but also in the 

other sectors, and that these costs are likely to be at least as high as the cost for manufacturing 

products. For this reason, iceberg-type transport costs are allowed for in agriculture. As a 

corollary, wages and consequently prices of agricultural products do not equalise between the two 

sectors. 

The introduction of transportation costs for agricultural products when shipped from one region to 

the other results in higher agricultural prices in the destination region (with highly concentrated 

manufacturing at the expenses of the other region). Hence, higher prices for agricultural 

products, translated into higher costs of living would actually discourage labour mobility between 

the two regions, and thus hinder the concentration of manufacturing activity. 

It seems very unrealistic to assume that the same homogenous agricultural good is produced in 

the two regions. The next step is therefore to consider differentiated agricultural products. The 

results are similar to the previous ones. In general, it is possible to reach the conclusion that 

while high transport costs in agriculture tend to prohibit concentration in manufacturing, a 

decrease in costs leads to concentration. 

Urban agglomerations 

A second strand of the new economic geography is concerned with urban agglomerations and 

attempts to answer questions as follows: 

• 

• 

• 

• 

How do cities come into existence and why do they exist in the longer run? 

How do new cities form and which are the mechanisms underlying the growth of cities? 

What is the reason that cities turn out to be linked one to another through a hierarchy? 

What is the role of natural advantages, such as harbors, in the development of cities? 

35 

253



International trade 

New trade theory, as developed by Krugman (1980), has much in common with new economic 

geography. The former aims at investigating international specialization phenomena in the case 

of two or more countries, which can be viewed as the equivalent of agglomeration and 

concentration within an individual country. It should be noted that the logic and the mechanisms 

at work are very similar and what makes the difference are different restrictions to the basic model 

as discussed below. 

Fujita et al. (2001) develops a model with intermediate goods. In this two-country framework, 

each individual country is composed of two sectors, namely manufacturing and agriculture with 

labor being assumed to be perfectly mobile between agriculture and the manufacturing sector. 

By contrast, labor is assumed to be immobile between the two countries. It is worth drawing 

attention to the fact that this model corresponds exactly to the canonical model with the difference 

that two countries instead of two regions are considered. There is no labor mobility between the 

two countries (regions) and intermediate goods are added to the model. 

The introduction of intermediary goods is normally associated with input-output matrices. Put 

differently, firms are assumed to be linked with each other through input and output linkages. The 

model avoids the inclusion of new sectors by assuming that the range of goods produced in 

manufacturing serves not only for final consumption but also as input for manufacturing itself. 

As a result, manufacturing uses labor and intermediary goods as inputs. In accordance with the 

model, the greater the gamut of intermediary goods, the lower the costs of production. In addition 

to that, if those goods are available locally, prices do not include transport costs. As far as sales 

are concerned, what firms produce is sold either to consumer as final goods or to manufacturing 

firms as intermediate goods. 

On the one hand, high trade costs leave manufacturing equally divided between the two countries 

because firms have to supply consumers. On the other hand, low trade costs translate into the 

full concentration of manufacturing activity in one of the countries on the grounds of backward 

and forward linkages. In the presence of intermediate trade costs, concentration is possible but 

not inevitable 

This model also manages to give an explanation for differing nominal and real wage levels in 

different countries. According to the model, the real wage in the country where manufacturing 

concentrates rises relative to those in the other country. One reason for this is that demand for 

labor increases and so does the wage level as a consequence of concentration in the first 

country. Another reason is that the country without manufacturing production has to pay the bill of 

transport costs of the imported manufactured goods, which in turn decreases the real wage. 

36 

254



Economic development: a sequential process 

In this stage of the analysis, an exogenous growth process is added to the model and it is 

assumed that demand for manufactured goods rises faster than potential supply. A model 

consisting of two countries is considered. The model seeks to explain under which conditions 

industrial concentration starts spilling out from one country to the other. 

The point of departure is to consider one country with highly concentrated manufacturing. It is 

shown that concentration will continue until the gains of backward and forward linkages in the 

industry exceed the potential loss that could be associated with the fact that wages are 

considerable higher in the industrialized country. If the losses are higher than the gains (that is 

the wage gap exceeds a given threshold) firms start moving into the other country. These firms 

then create their inter-firm linkages in the other country, which, through a positive feedback 

process will attract other firms and thus accelerate the concentration process. 

However, there are some interesting questions to answer. For instance, which industry is the first 

to move, and how does the industrial structure of countries at different stages of industrialization 

change? This is studied using a five-country model in which each country is composed of seven 

different industries. According to the model, the most labor intensive industry will be the first to 

move into the other country, followed progressively by less-labor intensive industries. However, 

the latter will move faster. 

The authors suggest that the inter-industry structure of the economies changes depending on 

several factors, including: 

• 

• 

the higher the share of final consumption goods in sales, the higher the probability that 

firms move (forward linkage); and 

the lower the intermediate input requirement, the quicker firm moves (backward linkage). 

Industrial clustering 

Fujita et al. (2001) construct a two-country, two-industry model with a single factor of production. 

Agriculture is dropped in the analysis and two manufacturing industries are used instead, both 

monopolistically competitive. The question is: under which circumstances do concentration or 

dispersion occur? 

The model provides us with two conclusions. First, in the event that one industry is strongly 

connected to the other industry through input-output linkages, concentration will not happen. In 

contrast, in the case of strong intra-industry links, the model predicts a high concentration, that is, 

each country specializes in one of the industrial activities. When considering trade costs, the 

results are similar to the previous ones: the two industries remain present in both countries if 

trade costs are high, but a dwindling in trade costs allows for and subsequently leads to 

specialization of the countries. 

37 

255



The authors suggest that with deeper economic integration bringing about lower trade costs, 

European economies are expected to undergo a concentration process and will converge 

towards the economic structure of the US. 

The two-country, two-industry framework is then replaced by a model including more than two 

industrial sectors. Although the model cannot predict why one industry concentrates in one 

country and not another, it can provide a theoretical underpinning in favor of the argument that in 

case one country loses a particular industry as a result of some sort of exogenous shock, there 

exists no mechanism that could ensure that, after the shock, the industry in question regains 

ground in the country under consideration. 

Empirical works related to industrial clustering show that the US experienced a high concentration 

of economic activities and a high degree of industrial clustering chiefly during the period prior to 

WW-I. (see e.g. Kim (1995)). Patterns of concentration and industrial clustering remained 

relatively unchanged afterwards. Studies focusing on Europe indicate a strong concentration and 

clustering which has occurred during the last 20 years or so. Even so, the level of concentration 

and sectoral clustering in Europe is not as high as in the US (e.g. Midelfart-Knarvik et al. (2000)). 

In a recent study with a strong cliometric influence (see section 2), Davis and Weinstein (2002) 

set out to investigate agglomeration patterns in Japan based on a unique data set stretching over 

8000 years. Data drawn from archaeological findings and population census indicate that 

dispersion in population has been historically high in Japan. However, variation in population 

density considerably widened from the outset of the industrial revolution. The authors argue that 

although increasing returns can hardly explain the origins of geographical agglomeration, they 

play a prominent role in explaining temporal dynamics, i.e. the acceleration of geographical 

concentration throughout the last 200 years. New light is also shed on the temporary nature of 

even very large shocks. For instance, the massive bombardment of Japanese cities at the end of 

WW-II does not appear to have permanently altered the spatial structure of the Japanese industrial 

structure. 

38 

256



2 THE EVOLUTIONARY ECONOMICS OF INNOVATION 

This section discusses the main findings from the evolutionary approach to the study of 

innovation and technical change. It then discusses the relationship between innovation and 

institutions (known as “systems of innovation”). 

2.1 Introduction 

The evolutionary approach to the study of innovation and technological change arose largely 

because the mainstream theory was considered inadequate to tackle the inherent disequilibrium 

dynamics that characterise technological innovation. Nelson and Winter (1984) point out that 

both the economy as a whole and the economic actors taken individually can be considered to 

evolve continuously, similar to the evolutionary process described in biology. Hence the name of 

evolutionary economics. 

Dissatisfaction with the traditional approach is almost pervasive and embraces the definition of 

innovation and knowledge, the description of the R&D process and more generally of the 

innovation process, the modelling approach based on the notion of full rationality and equilibrium, 

and the lack of consideration of institutional factors. 

Though these themes trace back to earlier researchers — in particular, the Schumpeterian 

perspective, which places emphasis on dynamics and competition through innovation — dynamic 

theorizing in evolutionary economics is often seen to start with the models of Nelson and Winter 

(1982). 

Although evolutionary economics is perhaps not a coherent theory with a common methodology, 

it may be categorised as a group of various approaches that share the same criticisms of the neoclassical 

approach. For example: 

• First, evolutionary economics does not treat economic agents as rational, in the sense 

they are in neo-classical economics. They do not maximize returns through calculating 

the optimal strategy from a range of probabilistic outcomes. 

• 

Second, evolutionary economics does not seek to analyse economic phenomena from 

equilibrium outcomes. In particular, dynamic change within the economy is considered 

best analyzed in terms of the processes that govern it, which are inherently disequilibrium 

dynamics. 

Evolutionary models draw their theoretical foundations on different sources and particularly on 

biological models, evolutionary game theory and institutionalism in order to capture essentially 

dynamic and irreversible processes. 

The literature tends to place emphasis on innovation and selection from variety as the force of 

economic development. Of particular importance is the need to see innovation as more than 

simple technological development (and much more than the production of knowledge as a public 

good) and to analyze economic development through technological and organizational 

improvements (which lead to the introduction of new products, inputs and techniques). 

39 

257



In evolutionary theories innovation is in fact defined more broadly than in mainstream models and 

as a multidimensional phenomenon. Innovation is not only a reduction in production costs or the 

commercialization of new products but also a more efficient reorganization of a firm’s activities, 

the adoption of more sophisticated managerial techniques, access to new markets, etc. 

This richer definition emphasizes that innovation extends outside the realm of technology in two 

ways: many innovations are not technological but involve other aspects of a firm’s activities such 

as its organizational structure; moreover, even technological innovations often affect nontechnological 

activities of a firm: technological and organizational changes are often importantly 

intertwined. 

Another important difference with mainstream models of innovation is that evolutionary models 

emphasize the tacit and specific nature of knowledge, as opposed to the idea of knowledge as 

information that somehow underlies mainstream models of innovation. 

In terms of the process that leads to the discovery of new products, evolutionary models describe 

the importance of learning, and of organizational learning often embodied in organizational 

routines as a source of knowledge. The process through which innovations emerge is by no 

means linear, but it is characterized by complex feedback mechanisms and interactive relations 

involving science, technology, learning, production, institutions, organization, policy makers and 

demand (Edquist 1999). 

Evolutionary economics can be looked at both from a microeconomic and macroeconomic 

perspective. The microeconomic aspects of evolutionary economics are mainly concerned with 

how innovation arises at the firm and market level. In contrast, the macroeconomic strand of 

evolutionary economics places innovation into the social and economic context and examines the 

dynamic process through which the innovation systems generate innovations through a complex 

and reciprocal relationship network connecting different components of the system. 

2.2 Microeconomics of Innovation 

Knowledge and rationality 

At the heart of evolutionary economics is a definition of knowledge and of its characteristics that is 

substantially different for the neo-classical assimilation of innovation to information. In each 

technology there are elements of tacit and specific knowledge that cannot be formalised and 

cannot, therefore, be diffused in the form of information. These firm-specific, local and cumulative 

features of knowledge are, to various extents, essential characteristics of innovation. 

The term tacit knowledge is used to encapsulate the idea of knowledge that is found to be useful 

(e.g. in an organizational routine or a skill) without being directly accessible to consciousness 

(Pavitt, 2002) or articulable (Winter, 1987). Thus individuals can be considered not to understand 

all that they know — tacit knowledge cannot be fully explained, written down or otherwise 

diffused; yet it remains valuable. By contrast the term information can be taken to mean ‘explicit’ 

knowledge, which can be codified and transmitted. 

40 

258



Information, as explicit knowledge, can be diffused around the economy (and fits most closely 

with the neo-classical concept of knowledge as a public good). Tacit knowledge cannot be 

passed from agent to agent in this way, but it is transferred and transmitted implicitly. Firms, and 

other economic organizations, can be seen as vehicles through which tacit knowledge is 

replicated through routines and techniques (Foster, 2001). 

Furthermore, as mentioned in section 1.2.5, drawing mutual benefits from tacit knowledge that 

cannot be explicitly transferred between firms may bring benefits to inter-firm co-operation in 

innovative activity that complement those benefits identified in the industrial organization 

literature. 

Support is found for this concept of tacit knowledge from psychological experiments. Nelson and 

Winter (2002) cite studies that examine the difference between processes acquired through 

rationality (understanding the underlying nature of the problem and calculating strategies for 

success) and learning-by-doing (learning successful strategies through trial and error without a 

complete theory on why such strategies are successful). They note that “a close logical 

connection between a learned task and a newly presented task does not necessarily indicate a 

potential for easy transfer”. Thus there is an important difference between knowledge from 

understanding and knowledge from practice (tacit knowledge). 

Such results indicate that the more knowledge that is embedded in firms, through working 

practices, research programs and ongoing developments, is tacit, the less transferable this 

knowledge is to outsiders. To some extent, if a firm cannot rationalize why a particular routine has 

proved beneficial, but continues with it on the basis of past performance, then it will be less able 

to adapt to a slight change in conditions and less able to apply successful strategies to slightly 

different problems. Thus, it is perhaps unrealistic to expect firms to re-optimize their strategies 

immediately when circumstances change — by its very nature a process of trial and error (or 

imitation) takes time. 

In a similar manner, the idea of tacit knowledge helps explain the large international differences in 

living standards and economic development that seem inconsistent with knowledge being a 

public good transferred and adopted at minimal cost. At a more micro level, it may be expected 

that different organizations may obtain different levels of output from ostensibly the same inputs. 

Evolutionary economists not only disagree with the neoclassical tradition on the definition of 

technological knowledge but also on fundamental assumptions on the behavior of economic 

agents. 

Evolutionary economics does not treat economic agents as rational, in the sense they are in neoclassical 

economics. This is not solely because economic agents do not have full information, but 

more importantly because they have limited cognitive powers; they do not profit-maximize (given 

their information set) but rather aim at sufficiently high levels of success to enable survival in a 

dynamic environment. 

41 

259



Related to this, Dosi (1988) emphasizes that innovative search is characterized by strong 

uncertainty whereby the list of possible events is unknown and one does not know either the 

consequence of particular actions for any given event. 

Instead of being rational and thriving for maximizing profit, economic actors are assumed to 

develop routine behaviors, which they follow as long as they turn out to work well. In case the 

routine behaviour fails to function, they will be revised and changed so that new routines may be 

developed. As a result of the evolutionary process, viable routines prevail whereas the less 

successful ones are eliminated. It must be mentioned, however, that these changes are not 

taking place continuously but are rather unevenly distributed over time and tend to occur 

sporadically. 

Innovation and competition 

Variety is an important part of evolutionary theories of innovation and competition and can be 

traced back to the assumptions on limits to economic agents’ rationality. 

A classic issue raised by the concept of Schumpeterian competition as a process of creative 

destruction is that this, on its own, would imply that economic growth leads to suppression of 

variety and ultimately to monopoly (Nelson and Winter, 2002, p 35). Thus, to counter this 

prediction, there is need to consider how new variety originates. 

Metcalfe (1998, p98) draws the link between the rejection of the neo-classical assumption of 

rational maximizing agents, and the generation (and re-generation) of variety: 

“…the positive side of bounded rationality is that it frees the imagination from the limits of 

calculation. When problems become too complex to be well-defined, let alone solved 

analytically, one is inevitably dependent and judgment, conjecture and the guiding hand 

of experience.” 

The point here is that while rational agents, profit-maximizing under identical resource and 

information constraints should, by definition, reach the same strategy, more realistic agents, 

guided by judgment and experience, will reach a range of preferred strategies. This process 

can be expected to lead to significant variety in innovative solutions, within which there are 

elements of path dependence as agents bring past experience to bear on current problems. 

From these factors leading to the generation of heterogeneity, competition proceeds to bring 

economic development. As discussed below, this is because competition leads to improvement 

through selection, adaptation and imitation. 

Foster (2001) emphasizes how the traditional (mainstream) discussion of competition in 

economics has much to do with static outcomes and little to do with dynamics. By modelling 

competition in terms of equilibrium, competition is seen as timeless, rather than a process of 

development. A key feature of evolutionary economics is its emphasis on competition as a 

process rather than an outcome. 

42 

260



Adopting a more dynamic approach, evolutionary economics is in part associated with the use of 

analogies to evolutionary biology to explain economic growth and the process of competition. At 

a very basic level, the notion of competition through natural selection associated with Darwin 

seems similar to the process by which economic competition selects more fit (efficient and 

profitable) firms at the expense of less fit firms. 

In this context, Mokyr (1999a) defines the Darwinian model “as a system of self-reproducing units 

(techniques) that changes over time.” Such a system appears to be best characterized by the 

following three properties: 

• The relationship through which the underlying structure determines the manifested 

entities. From an economic viewpoint, an underlying structure is the “useful knowledge” 

that provides the framework for the “feasible” techniques (manifested entity) i.e. what 

society potentially is able to do. 

• 

• 

A Darwinian system is likely to undergo dramatic changes over time. Techniques 

reproduce themselves between two time periods through either repetition or learning and 

imitation. This process then entails constant changes part of which cannot be foreseen 

and which can turn out to destabilize the current system. An important question that this 

raises is the pace of innovation. Is the innovation process a gradual one or does it occur 

in a stepwise manner? This is indeed a crucial issue (for example, in cliometrics as 

discussed below, the debate on whether the first industrial revolution as it occurred in Britain 

created discontinuity or whether it can be regarded as a smooth transition). 

Excessive variety in a Darwinian system implies that the actual number of techniques 

exceeds the number sustainable in the system. This results in a selection process. 

Mokyr (1999a) argues that selection can operate through three channels. The first one is 

the standard neo-classical mechanism where techniques are chosen in the event they 

maximize an objective function incorporating supply and demand side consideration as 

well as externalities. Second, there is inertia in the system that eliminates useless 

techniques. Finally, and more importantly, selection not only happens in the market but 

also at the social level. 

Inherent in this model of competition is the association between competition and experimentation. 

A variety of experiments allows for economic progress beyond the scope that would be 

considered achievable through comprehensive calculation and strategy (i.e. if each agent had to 

specify all permutations of outcomes, the likelihood of each, and the interrelationships involved). 

Nelson and Winter (2002, p 28) identify that the economic development ‘puzzle’ is how the vast 

technological progress and efficient forms of organization seen in the modern age could have 

arisen given the cognitive limits of humans and organizations. To this end, they propose that 

competition is the means through which the deficiency of human rationality and ability are 

mitigated to achieve favorable economic outcomes: 

“Neo-classical economics discovers social value in human selfishness, but not virtue 

that 

is robust against the human limitation of incompetence — and the possible role of the 

market process in achieving that robustness is not featured.” 

43 

261



Adopting this position, limited rationality, rather than being an impediment to well-functioning 

markets, is actually seen as part of the process of competition, and through it, innovation. If 

agents possessed far greater cognitive powers and rationality, the importance of competition in 

delivering economic development would be reduced. 

Arguably there is an interesting correspondence between this concept of competition and the role 

of competitive markets in conveying information, most famously associated with Hayek (1945): 

“The marvel is that in a case like that of a scarcity of one raw material, without an order 

being issued, without more than perhaps a handful of people knowing the cause, tens of 

thousands of people whose identity could not be ascertained by months of investigation, 

are made to use the material or its products more sparingly; that is, they move in the right 

direction.” 

In this example, the focus is on competition as a force for distilling information in an efficient 

manner. 

Thus the process of competition can be seen to bring benefits exactly because agents in the 

economy do not have full information — for example, through “price signals” competition tackles 

the deficiencies and asymmetries of information across the economy. Similarly, as discussed 

above, the process of competition can be seen to bring benefits exactly because agents in the 

economy are not fully rational — for example, selection from experimentation tackles the inability 

of economic agents to make optimal use of all the resources and information they have at their 

disposal at a point in time. 

Yet the Darwinian reference to competitive selection does not quite mean the same in 

evolutionary economics as in biology. Foster (2001, p120) argues that competition in economics 

should allow for “inheritance of behavioral characteristics acquired from experience in particular 

environments.” In other words, competition is not just about selection between firms (raising 

industry standards through elimination of weaker firms) but also the benefits of learning through 

creative endeavors, experience and also imitation that allows individual firms to make progress. 

(In fact Foster sees this process as more reflective of the evolutionary theories of Lamarck than 

Darwin — a significant difference being that Lamarck considered that organisms evolve because 

they are willing or need to evolve, rather than on the chance-orientated basis of Darwin). 

In particular, any competitive advantage one firm has over another depends on the extent to 

which the factors underlying that advantage are imitable. While species may not be able to 

imitate other species (imitation is fundamentally different from the process through species may 

adapt in a similar way to similar circumstances), firms are often able to imitate their competitors at 

low cost. Therefore an evolutionary approach to economic should not be so Darwinian as to miss 

the importance of imitation in economic development. Furthermore, the role of imitation in 

economic development will be governed by the ability of firms to maintain their competitive 

advantage through patents, and secrets, as well as relative speeds of innovation where timing 

matters to success. (How firms seek competitive advantage is discussed further in section 3 

within the context of the management and strategy literature.) 

44 

262



Thus economic development and innovation can be seen as a combined effect of selection (via 

competition) from a variety of competing routines and practices as well as the more endogenous 

process of agents seeking improved routines and practices. The role of ‘dynamic’ competition in 

this regard is quite different from its “static” force in constraining prices. 

This emphasis on competition as a selection process is supported by a quite separate strand of 

empirical literature that seeks to decompose industry-level productivity growth to examine the 

‘microstructure’ of this growth. Such studies use panel data of firms (or plants within multi-plant 

firms) to examine the extent to which economic development at the industry level is driven by 

productivity improvements within firms (‘internal restructuring’) versus expansion (and entry) of 

high-productivity firms and contraction (and exit) of low-productivity firms (‘external 

restructuring’). 

For example, Disney et al. (2000) analyze a dataset which covers the period from 1980 to 1992 

and contains data for 140,000 UK manufacturing establishments a year. They examine the 

relationship between industry-level and firm-level dynamics. Of particular significance they find 

that external restructuring explains approximately 50 per cent of changes in labor productivity, 

and 90 per cent of changes in total factor productivity, over the period. These results indicate that 

in the industries studied, competition in terms of selection from variety was a very important driver 

of productivity growth. 

Note that such studies focus more on technological and organization improvements than drastic 

innovations (major new lines of products) that would not be fully captured in the productivity 

studies. But it is clear to see how evidence of internal versus external restructuring, in the 

context of industry-level economic development fits in with the evolutionary concepts of 

competition as a selection process. 

Industry structure and innovation 

Schumpeter is often associated with the hypothesis that large firms with market power are more 

innovative than small ones, although Fagerberg (2002) argues that Schumpeter was more 

interested in the difference between new firms and old firms. Evolutionary economics offers an 

interesting perspective on the relationship between firm size and innovation. 

Fagerberg (2002) cites models by Nelson and Winter from the early 1980s that seek to explain 

why larger firms may be more innovative than smaller firms. On the one hand, if firms use 

retained profits to finance R&D, large successful firms will have advantages over smaller firms. 

On the other hand, a large firm is able to derive greater benefit from finding a better “routine” 

because it can put it to use across a larger number of units of production. Such analysis would 

suggest that large firms are at a competitive advantage, although Nelson and Winter are reported 

to see this effect counteracted by large firms (with more market power) having a higher profit 

target (price/cost ration) and this may provide some restraint in these dominating smaller firms. 

These models also suggest that if large firms are aggressive in pursuing their advantage over 

small firms, and there is in turn a tendency towards concentration, large “imitating” firms may do 

well at the expense of smaller innovators. Thus where technological progress is endogenous, 

productivity may be hindered compared to a situation of a more varied industry structure. 

45 

263



An alternative way of considering whether large firms are more innovative than small firms is to 

focus more on heterogeneity in ability. To the extent that competition leads to the growth of 

innovative firms, we might simply expect causality to run the other way: it is not that large firms 

are inherently better innovators than smaller firms, just that the better innovators will be 

successful and hence grow in size. Theories based along these lines would therefore be treating 

innovation as exogenous to firm size, with causality running from the former to the latter. 

This exposition between exogenous and endogenous innovation is valid more generally than firm 

size. Nelson and Winter (2002) draw a distinction between industries that are characterized by 

“science-based” innovation and industries that are characterized by “cumulative” innovation. 

Science-based industries are considered to be those where the thrust of innovation comes from 

R&D activities outside that industry. For example, firms in such an industry might benefit 

principally from external scientific developments and innovative activities by their suppliers. By 

contrast, in cumulative industries, innovation stems from R&D activity within each of the firms in 

that industry. 

An important difference may be expected between science-based industries and cumulative 

industries. In the former, where firms reap the benefits of innovative activities from outside their 

own industry, new entrants and previous laggards may be expected to have similar scope for 

development than industry leaders. Just because a firm has performed poorly in the past, or is a 

newcomer to an industry, should not totally undermine its ability to take advantage of external 

innovations if these are exogenous. 

However, if the innovative activity within an industry comes predominantly from the efforts of firms 

within that industry, we may expect a significant degree of path dependency. Successful 

innovators in one period will gain advantage over other firms that allows them greater potential for 

innovation the next period. The cumulative effect of this is that less innovative firms fall further 

and further behind, and incumbents have some degree of advantage over new entrants. 

High-level policy implications 

In discussing policy implications, Fagerberg (2002) draws distinction between evolutionary 

approaches and neo-classical approaches. Neo-classical economics places emphasis on the 

failure of markets to deliver “public goods”, and makes the case for government support to 

provide the public good of knowledge. By contrast, evolutionary economics does not treat 

knowledge as a public good in this sense; for example tacit knowledge exists in routines and 

practices rather than the consciousness and cannot be transferred from one party to another in 

the way that a formula or design can. Adopting a perspective more in line with the evolutionary 

approach would therefore provide les support for government funding of knowledge as a public 

good. 

Furthermore, evolutionary economics argues that variety is an important part of the scope for 

innovation and development, while the role of heterogeneity per se is underplayed in neoclassical 

economics. Policy proposals stem from this stance. For example: 

46 

264



!Rather than subsidizing R&D in well-established firms in traditional sectors, one 

might put the resources into new types of activities or actors, not necessarily with the 

expectation that these would do extremely well, but because the entire system 

(including thetraditional sectors) might benefit from such increased diversity” 

(Fagerberg, p 41). 

2.3 Systems of Innovation and the Role of Institutions 

The emphasis on the notion of “systems” of innovation suggests that innovation is not a 

phenomenon which results from the activities of isolated innovative firms, but the result of 

complex important interactions between firms and other organizations and institutions. Each 

innovative activity by a certain organization involves an element of reliance on external sources 

and innovation in the economy results from the complementary contributions of different 

organizations in a particular institutional environment. 

This strand of the literature draws together many common themes between two related 

approaches — evolutionary economics and cliometrics. What is cliometrics? 

Cliometrics was initiated by Alfred Conrad and John Meyer (Conrad and Meyer 

1957,1958) and has been popularized by Robert Fogel and Douglass North. 

Cliometrics is also often labelled as new economic history or quantitative economic 

history. It attempts to provide new insights into economic history, mainly by 

considering issues related to economic development and growth. 

A more precise definition for cliometrics is given by McCloskey (1978): “a 

cliometrician is an economist applying economic theory (usually simple) to historical 

facts (not always quantitative) in the interest of history (not economics)”. Despite 

substantial initial criticism, cliometrics has become a well-established branch of 

economics and in 1993, the NobelPrize in Economics was awarded to Robert Fogel 

and Douglass North. 

According to Claude Diebolt , cliometrics can be broken down into three major 

categories. First, descriptive cliometrics, which seeks to disentangle the economic 

and social influence that might have exerted an influence on historical events and 

processes. 

Second, positive cliometrics, which is concerned with organizing data based on 

which then different theoretical conjectures can be tested. Third, normative 

cliometrics, which cannot be tested empirically and which deal with ethical questions 

such as efficiency,distributive and social justice and with how to ameliorate social 

welfare. 

Cliometrics is largely based on reconstructed long (secular) time series or large panel 

data related to historic events. The theoretical background underlying cliometrics was 

initially neo-classical theory, but it has been evolving continuously with 

developments in 

general economics over the past 40 years or so. Recent works in this field are 

essentially related to evolutionary economics as a theoretical background. 

47 

265



Defining ‘Innovation Systems’ 

Different definitions of the concept of innovation systems have been offered in the literature, and, 

despite the inevitable differences, all point towards the importance of the notion that invention, 

innovation and its diffusion result from interactions between different complementary 

organisations and institutions. 

Freeman (1987) defines a system of innovation as: 

“[…] a network of institutions in the public and private sectors whose activities and 

interactions initiate, import, modify and diffuse new technology”. 

Lundvall (1992) defines a system of innovation as a system that includes all parts and aspects of 

the economic structure and the institutional set-up affecting learning as well searching and 

exploring. Alternatively, Gregerson and Johnson (1998, pp. 5) provides the following description 

for system of innovation: 

“…overall innovation performance of an economy depends not only on how specific 

organizations like firms and research institutes perform, but also on how they interact 

with each other and with the government sector in knowledge production and 

distribution. Innovating firms operate within a common institutional set-up and they 

jointly depend on, contribute to and utilize a common knowledge infrastructure. It can 

be thought of as a system which creates and distributes knowledge, utilizes this 

knowledge by introducing itinto the economy in the form of innovations, diffuses it 

and transforms it into something valuable, for example, international competitiveness 

and economic growth.” 

In general a system of innovation is defined in terms of its components and the relations through 

which the components are linked with other another. 

The elements of the system of innovation are organizations and institutions. Institutions exert a 

considerable influence on organizations, and simultaneously, organizations also impact on the 

institutional environment. Institutions may give rise to new organizations as well as being the 

origin of new institutions. 

Edquist (2001) argues that organizations are formal structures that are created with a well defined 

goal in mind on the one hand, and institutions such common habits, routines, established 

practices, rules or laws on the other. Organizations within the system can be broken down into 

two categories - primary and secondary. The activities listed above are associated with primary 

organizations whereas secondary organizations impact on how primary organizations behave 

when performing fundamental activities. 

Schoser (1999) argues that the institutional setting can be broken down into formal and informal 

institutions and that the factors identified by Gregerson and Johnson (1998) – institutional set-up, 

knowledge infrastructure, specialization patterns, public and private demand structure and 

government policy – may impact on innovation either directly or indirectly. 

Typical organizations in an innovation system are private firms, universities, government 

laboratories and industrial research associations, among which a division of labor is realized, 

48 

266



mainly coordinated by non-market means. These various organizations undertake different 

activities within the system, depending on their particular objectives, research mechanisms and 

incentive structures. 

The division of labor within a system of innovation explains why the systemic view holds that not 

only the components of a system of innovation are important but crucially also their links and 

interactions. Metcalfe (1995) observes that connectivity within a system is achieved by a variety 

of mechanisms as for instance: 

• 

• 

• 

mobility of employers in the labor market; 

grants and contracts for research; and 

informal networks, e.g. the links between user firms and their suppliers 

How ‘Innovation Systems’ Work 

There are various activities undertaken within a system. Liu and White (2001), identify five 

activities: research; implementation, i.e. manufacturing; end-use; education and linkage. Bergek 

and Jacobsson (2002) put forward the following functions of a system of innovation: creation of 

new knowledge; guidance of the research process; supply of resources; creation of positive 

externalities and creation of markets. 

Pearce (2002) offer a different approach to defining the activities of a system of innovation. 

Pearce identifies basic research as the first element on which applied research is based. Product 

development including market research is then built on both basic and applied research. The 

final stage of the system of innovation is adaptation and marketing of the products. The first two 

stages depend heavily on research institutions, government support as well as firms. By contrast, 

the last two stages mainly concern private firms. 

Nonetheless, Nelson (1996) notes that the notion of systems of innovation does not imply that the 

system was consciously designed or that interactions between the components of the system 

work are smoothly and that the system is coherent as a whole. 

Systems of innovation are often considered to be national, although it is debated whether national 

boundaries can be deemed relevant. 

On the one hand it is argued that the element of nationality may derive from different factors such 

as the national focus of technological and other policies, laws and regulations that have an impact 

on the innovative environment (Metcalfe, 1995). Nelson (1996) argues that because the 

education and research system, public infrastructure, laws and financial institutions keep some of 

their national characteristics, differences across different nation innovation systems are likely to 

persist even in the longer run. 

On the other hand, the boundaries of systems of innovation in different countries are getting 

blurred with the rise of international firms present in several countries and increasingly intensifying 

cross-country inter-firm connections to share knowledge and innovate. 

49 

267



It may also be argued that systems of innovation may be local and regional rather than national 

and that one may think of different systems of innovations at various levels of aggregation. 

It is widely recognized in the literature that there is a dilemma in national systems of innovation. 

This dilemma is whether to create incentives for innovation or whether to foster diffusion of 

innovation. At the heart of this problem is the patent system that encourages innovation by 

allowing for reaping the benefits and which at the same time hinders diffusion. 

The role patents play in innovation is controversial. For instance, Levin et al (1987) report a fairly 

limited role of patents when it comes to protecting new technology. A brief discussion of some of 

the findings of the literature review regarding patents is presented in the box below. 

Mokyr (1992, 1999a) also suggests that some institutional arrangements may hinder innovation. 

He points out that self-organizing systems become the cornerstone of evolutionary economics 

and notes that: 

“…the most interesting property of these systems is that they resist change. Resistance 

to change is essential for any system if it is to function and not degenerate into chaos.” 

However, he also emphases that resistance in society might have not only a stabilizing effect but 

can also act as a break on innovation. Indeed, high resistance can slow down or even eliminate 

the introduction of innovations or even deter innovation activity. 

Opposition against new innovations might come from both owners and employees. Mokyr 

(1999a) enumerates the following sources of resistance: 

• 

• 

• 

• 

economically motivated resistance; 

ideologically motivated resistance; 

systemic resistance; and 

frequency resistance. 

However, the success of an innovation hinges on two factors. On the one hand, the greater the 

value of the equipment and the more specific the skill that are threatened by a new innovation, 

the higher the incentive to resist. On the other hand, in the event that the innovation generates 

high social gains, the incentive to establish the innovation will be equally high. 

Second, the introduction of the new technology requires time for people to become familiar with it 

and to make best use of it. Atkenson and Kehoe (2002) build a quantitative model of technology 

diffusion in order to simulate this slow transition from a low towards a high-productivity regime. 

The model seems to fit actual data rather well for the late-19th century US manufacturing sector. 

One important feature of the model is that the new technology can be adopted only by building 

new plants. This assumption is consistent with empirical literature on the microstructure of 

productivity growth that finds that multi-plant firms often achieve productivity gains by closing 

50 

268



down inefficient plants rather than improving efficiency within those plants (e.g. Disney et al, 

2000). 

For the model to produce the slow transition, another assumptions are needed. Manufacturers 

using the old technology dispose of a slowly accumulated stock of knowledge. At the outset of 

the transition (from old to new technology), they are hesitant to relinquish this stock of knowledge 

in favour of the new technology. The higher the stock of knowledge compared to the one related 

to the new technology, the slower they are to go for the new technology. 

51 

269



The role of patents 

A number of papers attempt to analyze the history of patents and the patent system in 

theUS. For instance, Sokoloff (1988) examines 4,500 patents accepted in the US 

during the period 1790-1846 and sheds light on the fact that the number of patents 

changed at an uneven pace over time. Apart from an important rise of roughly 300 

per cent in patents that can be observed during 1805-10 and 1820-36, the number of 

patents recorded stayed fairly constant during the rest of the period. 

Using a different set of US patents, Phillips (1992) is concerned with the period 

spanning from 1831 to 1879 and confirms the finding of Sokoloff (1988). Accepted 

patents rose sharply over the periods 1950-60 and 1962-68, but remained relatively 

stable in other years. If the number of patents is viewed as a good proxy for 

innovative activity, it could be argued that innovation follows a stochastic path. 

Sokoloff (1988) suggested that the surge in the number of patents occurred between 

1820 and 1936 was due to major changes in the patent system that enforced 

incentives for innovative activities. In fact, patent theory tells us that the higher the 

expected return to an innovation, the higher the incentive to innovate. 

Khan (1995) puts forward that the US patent system, which was introduced in 1790 

,and especially the pro-innovation attitude of judges when accepting patent demands 

fostered substantially innovative activity in 19 century US. 

Phillips (1992) emphasizes the demand-driven nature of patent activity. He argues 

that not only accumulation of human capital, but also the increasing demand for 

patentable ideas and new technology in the context of increasing industrialization, 

which lead to an increase in accepted patents. 

However, some others have argued that patents do not truly reflect innovations. 

MacLeod (1986) points out that patentees are likely to react sensitively to the 

economic and financial conjuncture. She sets out to show that the increase in 

registered patents between 1691-1693 was influenced by the increased availability of 

risk capital. It remains an unanswered question, however, whether innovations were 

also that dependent on the economic environment. MacLeod also notes that the 

ongoing war gave a significant impetus to military innovation. Another important 

question here is the degree of development of the patent system. It is argued that 

weaknesses in the system probably exerted a deterrent or at least not conductive for 

patenting new innovation. 

MacLeod (1992) investigates the 19 century British mechanical engineering industry 

and argues that the existence of a patent system was one key to innovation. The 

system provided certainty for innovators so that they could reap the benefits of their 

innovation. Therefore, the uncertainty prevailing before 1830 due to the unreformed 

and inappropriate patent system encouraged secrecy, which in turn worked against 

the diffusion of new technology. 

52 

270



Do Different “Systems” Matter? 

There are differences of opinion as to whether differences between “Innovation Systems “ matter 

in terms of promoting innovation and technical change. 

One factor that can promote technology diffusion is the presence of multinational companies 

worldwide. However, Saviotti (2000) documents substantial differences that persist between 

national systems of innovation of different countries. He observes that different institutional and 

organisational configurations result in different specialization patterns across countries (see also 

section 1.4 on new economic geography.) 

Nelson (1996) also sets out to analyze the cause of differences in different national systems of 

innovation. Innovation systems are analyzed in three groups of countries, namely large highincome 

countries such as the US, Germany, Japan, small high-income countries such as Sweden 

and developing, low-income countries, e.g. South Korea and Argentina. 

Based on this categorization of countries, it is argued that the size and the affluence of a 

particular country mainly determine its capability to innovate. In addition, economic and political 

circumstances and priorities also have a significant role to play in shaping the innovation system. 

Further to this, Nelson (1996) emphasis the importance of competition. Strong competition is 

believed to promote innovations. However, he argues that domestic markets especially in small 

countries fail to play this role and therefore international competition encountered in world 

markets, that is on export and import markets, is crucial in fostering innovative activity. 

A strong and competitive export sector is thought to be the engine of economic growth in many 

economies. However, for the export sector to withhold international competition, the national 

system of innovation should be shaped to increase innovation and thus support exports. 

Nonetheless, competitiveness can be viewed and defined differently in high-income and lowincome 

countries. High wage level in developed countries implies that only new, improved 

products and the constant amelioration of the underlying technological process offer possibilities 

to boost competitiveness. By contrast, in less developed countries with lower wage levels, 

competitiveness is to be improved by diffusing and learning foreign technology. 

Security policy considerations also seem to forge the shape of the innovation system. Nelson 

(1996) argues that the bulk of government backed research and development is concentrating on 

the defense industry in the US and Britain. Military inventions are then applied to civil problems 

giving birth to new products. 

Furthermore, the current German and Japanese systems of innovation deeply root in the past 

when the system was designed to back the production of better weaponry. However, after World 

War II, those innovative capacities were turned toward civilian production, and this led to 

flourishing exports. 

This is a clear sign for the strong continuity in national innovation systems over time. Nelson 

(1996) points out that the German and French innovation systems in 1990 were very similar to 

53 

271



those in 1890. This phenomenon seems to hold true for the majority of developed countries 

except for the US where a major change occurred after 1945. 

Grupp et al. (2001) set out to investigate the system of innovation in Germany throughout the past 

150 years and come to the conclusion that the innovation system has been very stable over the 

period under investigation. According to the authors the industrial research system can be 

associated with an increase in innovative activity. When analyzing how R&D expenditures and 

innovation activity have changed from 1850 to 1999, they found that the German industrial 

research system has been very stable. In particular, they provide some more in depth 

econometric analysis, which reveals that during 1850-1913, patent activity can be predicted by 

growth in human capital. Furthermore, there is causal link from patent activity to standard of 

livings whereas standard of living influenced human capital. 

These results clearly indicate the presence of a virtuous circle that results in substantial economic 

growth. Results for the second period (1951-99) are less straightforward: the standard of living 

variable appears exogenous and turns out to explain patent activity as well as on public and 

private expenditures in R&D. 

The education system and especially universities have a significant role to play in the formation 

and the successful functioning of the innovation system. One important question to be addressed 

is how and to what extent the research and teaching orientation of universities satisfies the need 

for technological innovation. 

In fact, the design of a good interface between public research organizations and private firms 

has been central to innovation policy in most advanced Countries. This interface is considered 

important both because it may allow research conducted in public organizations to be more easily 

translated into economic growth and competitiveness and because it may result in fertile 

combination of complementary knowledge-basis and skills. Countering these benefits, the 

concern that closer linkages between public and private research organizations, and an increased 

privatization of university research, may diminish the variety of reward systems in the economy 

and in the long-term the potential of public research organizations to contribute to the innovation 

system. 

Evidence from Studies using Cliometric techniques 

Cliometrics also offers some observations on the role of the relationship between innovation and 

institutions. 

Neo-classical theory posits that economic agents maximise aggregate income in the absence of 

transaction costs. However, since Coase (1960), it is clear that transaction costs do matter. For 

instance, Wallis and North (1986) show that in 1970, 45 per cent of US GNP was linked to the 

transaction sector. As a corollary, the institution setting, which, to a large extent, determines the 

magnitude of transaction costs, has a considerable impact on the final outcome. 

What matters in the long- run is the continuous interaction between institutions and organizations 

that exerts an influence on developments in the institutional environment. Organizations are born 

and are operating within the given framework of the institutions and their activities reflect the 

54 

272



opportunities provided by institutions. If, for example, security is lax, piracy will come into 

existence, whereas if, say, property rights are properly enforced, actors may engage in more 

innovative activities. (This approach is indeed very close to the systems of innovation) 

Olson (1996) puts forward that institutions are essential for efficiently functioning markets. 

Nonetheless, it does not necessarily hold true that institutions provide per se the conditions 

required for efficiency in markets and therefore they are subject to changes. If the rules of the 

game are perceived by the organizations as too restrictive, they may attempt to modify the 

institutional setting. The driving force behind changes is, according to North, learning by 

individuals. And it is argued that competition contributes significantly to the learning process 

given that learning is indispensable for survival. That is, the greater the degree of competition, 

the bigger the incentive for learning. 

In a series of seminal studies, Douglass North (1990,1991) underscores the importance of social 

and economic institutions, which are likely to affect innovation and hence long-term economic 

development. According to North, institutions form the incentive structure of a society, and as a 

consequence, the political and economic institutions are the underlying determinants of economic 

performance. 

The institutional environment provides a general framework for social and economic interactions. 

Institutions fall into two categories. The first category can be referred to as formal constraints and 

is composed of rules, laws, constitutions, property rights, whereas the second one is labeled as 

informal constraints and is made up of customs, traditions, behavior, conventions, codes of 

conducts etc. 

In North (1994), the evolution of institutions is analyzed in a historical perspective. It is argued that 

there are some primitive institutional settings that are most unlikely to experience changes 

triggered by pressures coming from the inside. In three types of exchanges under consideration, 

namely tribal society, regional economy and long-distance caravan trade, learning, and hence the 

accumulation of knowledge and skill will not lead to changes, simply because innovation is 

perceived as something endangering the system’s survival. This is also borne out by Posner 

(1980). 

North (1968) also explores the impacts that changes in the institutional environment have on 

technological progress. He analyses the case of ocean transportation and finds that from 1600 to 

1860, the costs of ocean shipping dropped by 50 per cent due to huge increases in total factor 

productivity. The question is whether productivity gains are driven by technological advances or 

by other factors. 

North argues that in the period 1600-1784 productivity advances were slow, and virtually all the 

productivity growth came from decreased crew size and the better use of shipping time, that is the 

reduction of time spent in harbors. This productivity advance may be understood within the 

evolutionary economics framework as an improvement in organizational routines and working 

practices. 

55 

273



Between 1814 and 1860, North finds that total factor productivity experienced a considerable 

acceleration compared with the previous period, and grew roughly 10 times faster than before. 

North argues that the bulk of these productivity gains can be attributed to larger ships and to an 

increased load factor. 

Essentially, an increase in the size of the market and a sharp reduction in piracy explain these 

changes in the characteristics of ships. On one hand, shipping capacity was significantly better 

used because of immigrants pouring in America: Ships that used to return to America without 

carrying anything but the ballast were transporting men. On the other hand, security became 

better because of the elimination of piracy. This led to a reduction in the need for military personal 

and equipment on board. All in all, North emphasizes that productivity gains were not due to 

technological progress but rather the consequence of organizational innovations and changes in 

the institutional environment. 

With regards to the first industrial revolution in 18th and 19th century Britain, Landes (1994) argues 

that it was not pure accident that the first industrial revolution occurred in Britain and not 

elsewhere, e.g. in France. This is because of the structure of the industry and the institutional 

setting, which made it more likely that major inventions would be produced in Britain. However, 

Crafts (1995a,b) disagrees with this view and suggests that major inventions follow a stochastic 

process and thus those inventions could have occurred with roughly the same probability in 

France than in Britain given that the things in common between the two countries outweighs the 

differences between them. 

According to Mokyr (1993), two types of inventions can be distinguished. The first is 

“macro invention” and describes inventions of paramount importance for the economy as a whole. 

The second is “micro invention” and consists of small steps by which already extant techniques 

are ameliorated. Mokyr (1993) points out that macro inventions “… do not seem to obey obvious 

laws, do not necessarily respond to incentives, and defy most attempts to relate them to 

exogenous economic variables. Many of them resulted from strokes of genius, luck or 

serendipity.” Mokyr notes that Britain was by no means in a better position in the realm of 

macro inventions compared with France but the institutional and economic environment were 

more conducive as regards micro inventions. 

51 

56 

274



3 STRATEGIES OF INNOVATIVE FIRMS 

This section sets out some of the strategies applied by companies to enter new markets and 

introduce new products. This is addressed by looking at the strategies employed to gain and 

protect such an advantage. 

The section then looks how the optimal timing of innovation can be used as a strategy and 

includes a discussion of life cycles more generally. Finally, the section considers competition in 

network industries. 

Innovation and Competitive Advantage 

In considering the strategies employed by innovative firms, it is useful to draw a distinction 

between two broad categories: 

• 

• 

how a firm gains a competitive advantage over rivals; and 

how a firm can protect and maintain that competitive advantage. 

In the following sections, we discuss further these strategies and their relationship with 

innovation,although it is necessary to be aware that the distinction between these components is not 

clear-cut. 

Value creation: how firms gain a competitive advantage 

The first strategy focuses on how firms can create more value, which can be achieved in principle 

through: 

• 

• 

• 

cost reduction underpinned by efficiency improvements; 

increasing quality and value of existing products and services via differentiation relative to 

competitors; and 

developing new products and services. 

At the centre of value creation is the notion of competitive advantage popularized by Porter 

(1985). Porter’s view is this: 

“Competitive advantage is at the heart of the firm’s performance in competitive 

markets.After several decades of vigorous expansion and prosperity, however, many 

firms lost sight of competitive advantage in their scramble for growth and pursuit of 

diversification. Today the importance of competitive advantage could hardly be 

greater. Firms throughout the world face slower growth as well as domestic and global 

competitors that are no longer acting as if the expanding pie were big enough for all.” 

In accordance with Porter (1985), competition is a continuous quest for competitive advantage. 

57 

275



External and internal sources of competitive advantage can be identified. External sources such 

as changing consumer patterns or changes in technology are exogenous to firms. Since firms 

are heterogeneous in their resources and capabilities, they are not equally able to adapt to such 

changes, and thus some firms are able to realize competitive advantage through more timely and 

efficient responses to exogenous factors. 

The internal source of competitive advantage is essentially the capability to innovate, by which we 

mean inventive effort, rather than adaptive or imitative behavior. 

Cost leadership 

As highlighted above, value to customers can be created by means of decreasing costs, through 

which a firm is able to reduce prices relative to competitors. 

At the extreme, focus on this strategy can be described as thriving for cost leadership, which 

refers to the situation when a firm has the capability to produce similar products at significantly 

lower cost than its competitors. There are several routes through which a firm can achieve cost 

advantage, for example: 

• 

• 

Economies of scale and scope yield benefits to the extent that an increase in production 

volume will entail a reduction in unit cost. 

Economies of learning describe experience-based learning. The more complex the 

process technology or the product, especially in terms of tacit knowledge (i.e. that which 

cannot be codified and transferred) the larger the scope for benefits from learning-bydoing. 

This provides means for established companies to experience cost advantage 

over new firms. 

Enhancing process technology and process design allows for greater productive 

efficiency and lower production costs for the same output. 

Similarly, enhancing the organizational efficiency and routines (for a given technology) will 

increase productive efficiency. 

Reducing input costs allows for lower unit costs for the same level of productivity. 

Because of the use of different suppliers, geographical differences in input prices and 

different bargaining power, different firms may pay a different price for the same set of 

inputs. 

Augmenting capacity utilization: In the presence of high fixed costs, unit costs are to be 

decreased by increasing capacity utilization. On the other hand, overcapacity resulting in 

overtime pay, premiums for night shifts and rising maintenance costs also increase unit 

costs. 

58 

276



Product quality and differentiation 

Changes to product quality (in particular quality-price trade-offs) and product differentiation 

compared to rival products represent the second facet of value creation. 

A firm can set out to differentiate its product according to the tangible and intangible aspects of 

the products it manufactures. The former concerns physical characteristics and performance of 

the product, while the latter relates to perceived, mainly social and psychological faculties. 

The nature of competition in a market will be associated with whether firms strive to gain 

competitive advantage through innovation in physical attributes of the products they supply, or the 

intangible characteristics. For example, where an industry is mature and there appears little 

scope for changes in product innovation and cost leadership, competitive advantage may be 

achieved through a successful branding exercise that provides differentiation relative to rivals. 

Product innovation 

While changes to the quality of a product represent some degree of product innovation, 

innovative behavior is likely to be greater with regard to the introduction of new products and 

services that have features to mark them as distinct from current products. 

A strategy of product innovation may have the effect of creating new markets (in competition 

policy terms, the introduction or products for which there are no good substitutes on the demandside) 

but may also, though not necessarily, destroy old markets (as consumers cease 

consumption of previous generation of products in favor of the new products). 

When it comes to assessing how to exploit an innovation to the maximum, that is to maximize 

profits related to the innovation, firms can choose from several options depending on 

• 

• 

How much risk can or are they willing to take, and thus how high a return do they expect; 

How many resources can or are they willing to put at disposal when exploiting the 


Risks are high in emerging, innovative industries. First, firms have to face technological 

uncertainties (such as the direction technology develops in and which technical standards are 

established in the industry). Second, there are substantial market uncertainties. In other words, it 

is difficult to predict the potential size and expansion of the market. Nonetheless, these risks are 

manageable to some extent via: 

• 

• 

• 

Co-operation with lead users 

Limiting financial risks related to financing innovation 

Maintaining a high degree of flexibility and reactivity. 

59 

277



Grant (2002) enumerates five different strategies through which innovative firms can reap the 

benefits of their innovations. These strategies, shown in Table 3.2, can be associated with 

differing involvement of he firms, different risk-taking, resource requirements and firm size. 

Strategies in Innovative Markets 

Strategy 

Internal 

commercialization 

Joint venture 

Strategic alliance 

Outsourcing 

Licensing 

Risk 

Necessary firm size Resource requirement 

High risk, high returns and 

Medium-large High 

complete control 

Sharing investment and risk but 

also risk of conflict with the partner Medium-large Resource and 

capability synergies 

company 

Medium among several firms 

Limits risk but increases 

dependence on the outside world 

Small risk and small returns 

Small 

Small 

Access to outside 

resources 

Low 

Source: Adapted from Grant (2002 p342). 

Each of these strategies is described below: 

• The strategy of internal commercialization means that the firm commercializes the new 

product — that is, the output of the innovation — without any external productive help. 

This is the highest level of involvement that goes hand in hand with the heaviest 

investments and therefore the highest risk. Nevertheless, the firms can have the whole 

process under its control. In doing so, the resource requirements are substantial, which 

implies that the firm is expected to be fairly large to be able to bear the burdens. 

• 

• 

• 

• 

The decision to start a joint venture still implies a great deal of involvement. By contrast, it 

also helps to share the cost and the risk of the investment. In addition to that, the 

participating firms can pool their resources together and thus benefit from synergies. On 

the other hand, however, they run the risk of potential disagreement and possible 

differences in corporate culture. 

Strategic alliance represents a middle way between total and very small involvement. 

Risk and resources can be shared. 

Outsourcing is suited to both large and small companies so as to enable access to 

outside resources without too much commitment. 

Licensing is typically — but not exclusively — employed by small firms, which do not 

possess the necessary resources to exploit the innovation on their own. 

Protecting competitive advantage 

The second strategy can be understood as efforts to protect and maintain a competitive 

advantage that has been achieved through value creation. In particular, it concerns efforts firms 

60 

278



may take to prevent imitation by rivals, since in competitive environments, imitation is a crucial 

force that undermines the competitive advantage of a firm over time. Such behavior may be 

both legal (e.g. patenting) and illegal (e.g. certain types of anticompetitive behavior that inhibit 

rivals access to a market). 

Once a firm has been able to gain competitive advantage in some respect, it will have incentives 

to protect and maintain this. In particular, where a firm has benefited relative to competitors from 

an important innovation it will be keen that rival firms do not undermine this advantage through 

imitation. 

One way in which firms achieve such protection is through various legal institutions put in place 

for this purpose (e.g. patent policy). These are highlighted below before we turn to other ways in 

which a source of competitive advantage can be maintained. 

The use of legal institutions against imitation 

Intellectual property rights are legal institutions that may protect the competitive advantage a firm 

has realized through innovation by restricting the degree of imitation of the innovation that can 

take place. 

Patent policy provides a firm with protection against direct imitation of certain types of inventive 

innovation, for a period of time, subject to disclosure of information relating to the innovation. 

Similarly, copyright protection prevents the direct copying of new content. In both cases the aim 

of the legal protection is to provide incentives for the innovation to take place, exactly because in 

the absence of this protection such incentives for such innovation would be mitigated by the 

threat of direct imitation and risk of immediate erosion of any competitive advantage accruing to 

the innovation. 

Nonetheless, there are ways in which patent policy may be used by firms against the interests of 

consumers. Pre-emptive patenting describes the case where firms gather a series of patent 

around an initial innovation, not for the purposes of using those patents in the introduction of new 

products, but rather to stop other firms form inventing around the original innovation. Thus the 

new patented products or technologies are neither used by the patentee nor licensed to other 

firms, but their value derives from the fact that rivals cannot use them rather than the use the 

holder makes of them. 

Alternatively, where a particular product or service protected by patent represents a bottleneck to 

a range of related markets, the patent holder may be able to use its protection in the supply of the 

patented product to enjoy competitive advantage in the supply of products in these related 

markets. 

In some sense, there is overlap between creation of competitive advantage through product 

differentiation and maintenance of this advantage through branding. Where firms can link an 

innovative product to a successful brand image, and protect that brand from imitation (e.g. via 

trademark law), the advantage from the initial innovation can be maintained in the face of 

imitation as to the products physical characteristics. For example, in the pharmaceutical industry, 

61 

279



successful creation of goodwill, reputation and brand recognition allow firms to enjoy some (albeit 

less) competitive advantage over a drug that they have created once the patent has expired. 

Other protection against innovation 

Besides the use of intellectual property rights, the competitive advantage associated with an 

innovation can be obtained through other mechanisms. 

The first strategy is secrecy. Where an innovation is hard to observe or hard to codify, the 

innovative firm may able to maintain competitive advantage simply by not facilitating the flow of 

information about the innovation to competitors. 

A second strategy to protect competitive advantage is to merge with rivals. This could provide 

protection where a particular rival is capable of reducing the competitive advantage a firm has 

gained through innovation. Of course, the merger could bring other benefits that are associated 

with creation rather than maintenance of competitive advantage. For example, the combination 

of tacit knowledge in the merged parties may bring greater capability to innovate. 

Finally, a firm may be able to with sufficient market power may be able to protect its competitive 

advantage by using this power to restrict rivals’ ability to compete with it. Besides predation, a 

firm may be able to use competitive advantage with regard to one good or service, and spread 

this advantage to related services, for example: 

• denying the rival access to some important input service, ranging from necessary raw 

materials and components to services that allow supply of a service to particular 

customers, or even IP rights; or 

• 

bundling or tying products such that rivals are forced to compete against the bundle of 

services offered by a firm and the competitive advantage the firm holds is transferred 

across the range of products supplied together. 

Not only could such a strategy help maintain the competitive advantage from the original line of 

innovation, but it may also serve to expand the economic area of which returns from the 

competitive advantage can be enjoyed. However, it is essential to recognize that such strategies 

may also be conducive to value generation. For instance, with respect to bundling, this pricing 

option may be the only way in which sufficient revenue could be generated to cover the initial 

outlay required for an innovation. 

3.2 Timing of Innovation and Life Cycles 

Having set out firm strategies according to two broad categories, we discuss below issues of 

timing. The first sub-section considers how the success of innovation, and thus the optimal 

innovation strategy, may be dependant on timing. The second sub-section considers product 

‘life-cycles’ more generally. 

62 

280



First-mover advantage 

An essential question in innovative markets is to choose a strategy whether to enter early in the 

market and invest heavily in research and development in the hope of future profits and market 

leadership or to adopt a wait-and-see policy and to follow the technology leader by copying the 

innovation. The benefits of a first-mover strategy relies on three conditions: 

• 

• 

• 

Length of the lead time (i.e. the time over which the innovation is protected from imitators 

and followers) either through IP rights or less formally. 

The extent to which complementary resources are required while exploiting the 

innovation. If the need for complementary resources is relatively low, the initial investment 

and therefore the risk is lower. This may stimulate firms to try to be the first-mover. 

The possibility to set a standard is also likely to give an impetus for first-movers. 

The reason why standards are set is closely connected with positive network externalities 

(discussed further below in the context of the economics of networks). The higher the number of 

users of the standardized product, the more valuable the product to individual user. Positive 

network externalities have three major sources: 

• 

• 

• 

users of the product are connected with each other through networks (such as 

transportation and the internet); 

the ease with which complementary products are available (for example, software 

applications for Windows); and 

switching cost between different networks: the larger the network the customer uses, the 

lower the costs to switch to another network (for example, mobile telephony networks). 

Generally, technological standards tend to result in ever increasing positive network externalities. 

If a technological standard appears to be dominant in the industry, more and more users are likely 

to choose this one that may easily lead to winner-takes-all situation. This is one reason why 

innovative firms try hard to establish their technology as the industry standard. 

Shapiro and Varian (1999) identify three possible and highly complementary strategies how firms 

could establish their own technology as the standard. The first strategy builds on obtaining the 

support of other, rival companies and firms producing complementary products, possibly through 

tacit agreements. The second approach consists of market pre-emption. Finally, and importantly, 

the firm aiming at winning the standard war has to make the impression from the very outset that 

its standard will be accepted at the end of the standard war. Put it another way, marketing and 

other communication techniques can be employed so as to create self-fulfilling expectations 

among customers and rivals. 

However, different firms may have different strategic windows. These are defined as the time 

period during which they are able to seize market opportunities in line with their resources and 

capabilities. Smaller firms tend to have shorter strategic windows because they cannot wait for a 

63 

281



long time to introduce new innovations and procure competitive advantage as opposed to large 

corporation that do not tend to rush in introducing new technologies and take unnecessary risks 

given their solid financial background. 

The role of the life-cycle of an industry 

The position in the life cycle of a particular product, market or industry is likely to have a role to 

play as regards the strategy adopted by any firm towards the creation and protection of 

competitive advantage. Life cycle theory sets out a useful analytical framework for analyzing this. 

In this context, there is some ambiguity in the relevant literature regarding terminology between 

product life cycles and industry life cycles. In principle these are different concepts since industry 

usually refers to a wider set of entities that a product, but these terms refer broadly to the same 

approach and are sometimes used interchangeably. In some sense the concepts of an industry 

life cycle is associated with the industrial organization literature (Bonaccorsi and Giuri (2000), 

Horvath et al. (2000) and Klepper and Kenneth (2001)), while marketing literature may be more 

likely to use term product life-cycle. In this overview we refer to product life cycles. 

Based on a series of case studies and empirical papers, Klepper (1996, pp. 564-565) comes up 

with five major regularities that best characterize the product life cycle: 

• 

• 

• 

• 

• 

The number of new entrants rises very rapidly at the beginning and then starts declining 

and eventually stabilizes at a very low level. 

The number of incumbents increases in the initial phase and then starts falling 

accompanied by a steady increase in output. 

Initially, market shares change quickly, i.e. some firms grow fast and others loose ground. 

However, after a while, market shares stabilize. 

Simultaneously, there is an increase in the diversity of products competing in the market 

because of a rise in product innovation. 

However, as the market gets more mature and with the consolidation of the firms, 

incumbents spill more energy on process innovation. 

Werker (2003) provides two more stylized facts associated with the product life cycle complement 

the analysis above. First that prices fall sharply at the outset of the life cycle with standardization 

and a growing number of customers. However, later on, prices decrease at a lower pace and the 

decrease finally comes to a halt. Second, that the beginning of the life cycle favors entry 

whereas established firms are favored to new entrants at later stages of the life cycle. 

More generally, the life cycle of a product may typically have four phases: (i) introduction; (ii) 

shakeout; (iii) maturity; and (iv) decline. 

During the first phase, i.e. the introduction, a few companies produce a wide range of goods with 

substantial differences in product features partly on the grounds of differences in the underlying 

64 

282



technology. The competition among technologies yields rapid product innovation. Werker (2003) 

notes that at this stage of the product cycle, consumers’ preferences are not well established. 

Therefore, firms exert effort to discover the unclear preferences through high rates of innovation. 

The second stage of the life cycle usually witnesses a shakeout and the emergence of a 

dominant technology, accompanied by standardization of around this technology. This lays the 

ground for further rapid innovation. The competitive environment forces unsuccessful firms to exit 

the market and leads to mergers and acquisitions. Product differentiation in this phase can be 

achieved mainly through new product design and by continually improving quality. 

For instance, Klepper and Simons (2001) provide empirical evidence for dramatic shakeouts that 

occurred at some point in the automobiles, tires, televisions and penicillin industries in the US. 

They suggest that earlier entrants are more likely to survive shakeouts. Shakeouts do not seem 

to be triggered by drastic events but rather the consolidation in these industries was a result of the 

process of competition during which early entrants gain high market shares through innovation. A 

similar conclusion is drawn by Horvath et al. (2000), who find that the shakeout in the US beer 

brewing industry occurred from 1880 to 1890 was a consequence of earlier mass entry and the 

subsequent competition. 

When an industry reaches maturity, products tend to become increasingly homogeneous and 

could be viewed as commodities. In this context, and to counter this tendency, branding and the 

introduction of complementary services represent an effective way to differentiate products. Price 

competition also tends to become more intense in this phase of the life cycle. Hence, cost 

advantage is a crucial aspect of competitive advantage. Simultaneously, the technology 

developed in earlier stages becomes diffused and firms seek to improve this technology 

incrementally. 

The final phase can be referred to as the decline of the industry. During this phase fierce and 

possibly destructive price wars become the norm given that product differentiation is extremely 

difficult. The lack or at least the very low pace of product and process innovation is one reason 

for this. As a result, some firms decide to exit from the market and this further contributes to the 

decline. 

Different industries may be very different as regards the intensity of competition and the pace of 

innovation in the same stage of the life cycle. Accordingly, three types of industries can be 

identified: 

• In local monopoly markets, products are very specialized and target narrow market 

segments. Products are of high quality with high unit values, which implies that volume of 

production is low. Because of substantial differentiation, competition is very limited. 

Examples for this kind of market are found amongst the following areas: highway and 

residential construction, surgical appliances, defense industry. 

• 

Traditional industrial markets are characterized by homogenous products where 

competitive advantage can be usually achieved via cost leadership, branding or product 

65 

283



variety. Passenger cars and household equipment are examples of products belonging to 

this category 

• Schumpeterian markets are dominated by product innovation, where incumbents can be 

completely displaced by entrants. These markets might be seen within the computer 

software industry and within consumer electronics. 

3.3 Issues Relating to Network Industries 

Network industries give rise to a number of issues that are likely to affect — or at least influence 

— the strategies used by firms to innovate. The characteristic features of these industries 

originate from network externalities. A positive network externality arises when “a good is more 

valuable to a user the more users adopt the same good or compatible ones”. A common 

distinction is that between direct and indirect network externalities. 

Direct externalities typically occur in a two-way (physical) network. This externality reflects the 

fact, for instance, that a telephone user benefits from others being connected to the same 

network: an additional telephone increases the number of potential communication within the 

system, and thus the value of membership. 

Indirect externalities arise from the fact that a network of users increases the incentives to 

produce compatible products that are complementary with the network good. These externalities 

arise in systems, which can be defined as “[…] collections of two or more components together 

with an interface that allows the components to work together” (Katz and Shapiro, 1994, p93). 

The typical system comprises of a “platform” (or hardware) and of “applications” (or software) that 

can only be used with a platform. Compatibility refers to the possibility that software can be used 

with a particular platform. The network externality arises when users make their purchase over 

time, because in the presence of economies of scale a greater number of complementary 

products can be supplied at a lower price when the network grows. 

The existence of network externalities has a substantial impact on the way technologies are 

chosen and promoted. In this section we follow closely Katz and Shapiro (1994) overview of 

systems competitions and their identification of three main issues that have been addressed in 

the literature: 

• 

• 

technology adoption decisions; 

product selection decisions; and 

. 

66 

284



• compatibility decisions. 

Technology adoption decisions 

A number of issues that arise in markets characterized by network externalities can be addressed 

by considering a system only, setting aside competition between different systems. 

The existence of direct network externalities drives a wedge between private and social 

incentives to join the network. A single user does not consider the benefits that would accrue to 

others by his joining the network and as a result the market may lead to network of inefficient 

small size. In fact, the benefits that an individual would take into consideration when choosing 

whether to join a network depends not only on the current size of the network, but also on its 

expected future size. 

Consumers’ expectations play a central role in driving market outcomes where there are 

technology decisions in markets with important network effects. An important consequence is that 

multiple equilibria can easily arise. For instance, if all consumers expect no one else to join the 

network, then its size would be zero, even if the network may be valuable for consumers; if all 

expect everyone to join, the network may achieve a large size. 

Indirect externalities may have a different impact on the size of the network. In a system market, 

one consumer’s adoption decision would does not affect other consumers, given the prices and 

varieties of products available. The externalities arise indirectly through the impact of one 

consumers’ adoption decision of the future variety or prices of applications. Katz and Shapiro 

(1994) note that in such an environment, if cost conditions are consistent with the existence of a 

competitive equilibrium, a perfectly competitive equilibrium arises in system markets. When, at 

the other extreme, a monopolist is the sole supplier of the components of a system, there are 

issues regarding multi-product pricing and inter-temporal pricing. 

We follow Katz and Shapiro (1994) closely to discuss these issues. Consider the case where the 

monopolist can commit to the prices of both components of the system: 

• 

• 

with fixed proportions technology, the monopolist simply sets the price that maximizes 

profits in the usual way. 

with a variable proportions technology, issues of bundling and price discrimination arise 

but again the inefficiencies are attributable to monopoly power rather than externalities. 

If the supplier cannot commit to prices in advance, buyer expectations are again crucial: a 

monopolist would like to convince consumers applications will be available at low price in the 

future. After consumers are locked in, however, the monopolist may raise price to the monopoly 

level. 

In the literature, it is usually assumed that consumers’ expectations on the future prices of 

applications are affected by the quantity of hardware currently sold which may act as a signal: a 

larger base may lead to cheaper and/or more varied applications. Hence, the monopolist may 

want to lower the price of hardware to create a larger network and software aftermarket. 

67 

285



Product selection decisions 

Competition in network industries may take the form of competition between incompatible 

systems. 

Network externalities may lead to a de facto standardization whereby everyone uses the same 

system. Due to positive feedback elements, a system may become dominant once it has gained 

an initial edge, a phenomenon which is usually referred to as “tipping”. 

Two main factors, however, may limit tipping and sustain multiple networks: 

• 

• 

network externalities being exhausted at a smaller network size; 

consumers’ heterogeneity and product differentiation. 

In the latter case multiple networks would reflect consumers’ love for variety and there is a typical 

trade-off between variety and standardization. 

Competition between systems may be very intense, at least before a dominant system, if any, 

would emerge as dominant. For instance, firms may be engaged in very intense price 

competition at an early stage for seeking to establish an installed base and achieve leadership. 

This competition can be interpreted as firms bidding for future monopoly profits. 

Competition may also be played in trying to affect consumers’ expectations about the dominant 

system that would emerge. 

An important implication of the network externalities literature is that the market may settle in an 

equilibrium where the dominant system is that with a lower social evaluation. Markets can exhibit 

“excess inertia” and remain locked into an obsolete standard, even though a better one is 

available. 

However, markets can also exhibit “excess momentum” whereby the market tips inefficiently to 

new technologies. This could, for instance, arise when competition is between an older 

technology which is competitively supplied and a new technology which is sponsored. The new 

technology that is sponsored may have an advantage over an older technology that is more 

competitively supplied because the sponsor of the technology may engage in pricing below costs 

or other types of investments (with the hope of recouping these later once the technology is 

established). Nonetheless, this perceived risk is not straightforward — pricing very low from the 

outset may be the only way of introducing an efficient new technology, and tipping may be 

inevitable. 

Compatibility choices 

Compatibility between different networks/systems is often a choice variable of firms. Two types of 

compatibility can be considered: 

• horizontal compatibility: between two comparable rival systems; and 

68 

286



vertical compatibility: between successive generations of a technology. 

Compatibility entails both social benefits and costs: 

• 

• 

• 

• 

• 

compatibility expands the size of both networks thereby avoiding the cost of participating 

to two different networks (e.g. duplicate equipment); 

compatibility in systems may lead to lower production costs for economics of scale, 

learning effects, etc; 

compatibility enhances variety by allowing consumers to combine components from 

various systems; 

the risks of adopting a particular technology are lower; and 

costs derive from the mechanism by which compatibility is achieved. Standardization 

may lead to a loss of variety and may prevent the development of new incompatible 

systems; adapters that allow interfacing have a cost themselves and may work 

imperfectly. 

The nature of competition in the market is affected by compatibility decisions. For systems that 

are compatible, competition is essentially at the level of each component. For incompatible 

systems, competition is at the network or system level. 

Katz and Shapiro (1986) consider the impact of compatibility on pricing competition over time. 

Price competition is relaxed at earlier stages of the product life cycle because the market loses its 

winner-take-all feature. For the same reason, however, competition may be intensified in later 

stages of the product life cycle. 

Compatibility decisions may be affected by a number of factors: 

• 

• 

• 

asymmetries in the probability of being the winner, e.g. reputation; 

product differentiation; and 

installed base 

Side payments may help firms to reach an agreement on compatibility. When they are not 

feasible, it is useful to distinguish markets where a firm can unilaterally impose compatibility and 

those where a firm can unilaterally impose incompatibility. 

In general, there may be different mechanisms and institutions that govern compatibility choices: 

standards committees, unilateral action and hybrid mechanisms. (Farrell and Saloner, 1988) 

69 

287



4 

4.1 

BIBLIOGRAPHY 


Arrow, K J (1962) “Economic Welfare and the Allocation of Resources for Invention”, in 

Universities National Bureau Committee for Economic Research. The rate and direction of 

inventive activity Princeton, Princeton University Press. 

Baumol,William (2002) The Free Market Innovation Machine,Princeton,N.J.,Princeton Univ.Press 

Beath, J, Poyago-Theotoki, J, Ulph, D, (1998), “Organization design and information-sharing in a 

research joint venture with spillovers”, Bulletin of Economic Research, 50:1, pp.47-59. 

Beath, John, Katsoulacos, Yannis and Ulph, David (1987) “Sequential product innovation and 

industry evolution”, Economic Journal (Supplement), Vol 99, p32-43. 

Beath, John, Katsoulacos, Yannis and Ulph, David (1994) “Strategic R&D innovation” in Cable, 

John (ed) Current Issues in Industrial Economics, Basingstoke: Macmillan. 

Beath, John, Katsoulacos, Yannis and Ulph, David (1995) “Game-theoretical approaches to the 

modelling of technological change” in Stoneman, Paul (ed) Handbook of the Economics of 

Innovation and Technological Change, Oxford UK and Cambridge USA: Blackwell. 

Bester, H. and Petrakis, E. (1993) “The incentives for cost reduction in a differentiated industry”, 

International Journal of Industrial Organization, No16, p519-534. 

Boone, Jan (2000) “Competitive pressure: the effects on investments in product and process 

innovation”, RAND Journal of Economics, Vol 31, No 3, p549-569. 

Boone, Jan (2001) “Intensity of competition and the incentive to innovate”, International Journal of 

Industrial Organization, Vol 19, p705-726. 

Cohen, W M and Levinthal, D A (1989a) “The Implications of Spillovers for R&D Investment and 

Welfare: A New Perspective” in a Link and K. Smith eds, Advances in applied micro-economics 

vol. 5: The factors affecting technological change. Greenwich, Conn: JAI Press. 

Cohen, W M and Levinthal, D A (1989b) “Innovation and learning: the two faces of R&D – 

implications for the analysis of R&D investment”, Economic Journal, Vol 99,p569-96. 

d’Aspremont, C, Jacquemin, A, (1988), “Cooperative and non-cooperative R&D in a duopoly with 

spillovers”, American Economic Review, 78, pp.1133-1137. 

Dasgupta, P and Stiglitz, J E (1980a) “Industrial structure and the nature of innovative activity”, 

Economic Journal, 90:266-293. 

Dasgupta, P and Stiglitz, J E (1980b) “Uncertainty, industrial structure and the speed of R&D”, 

Bell Journal of Economics, 11:1 28. 

Dasgupta, P. (1986) “The theory of technological competition”, in J.E. Stiglitz and F. Mathewson 

eds., New Developments in the Analysis of Market Structure, London: Macmillan. 

70 

288



Delbono, F. and Denicolò, Vincenzo (1990) “R&D investment in a symmetric and homogeneous 

oligopoly: Bertrand vs Cournot”, International Journal of Industrial Organization, p297-313 

Fudenberg, D & Tirole, J (1983) “Learning-by-doing and market performance”, Bell Journal, 

14:522-530. 

Fudenberg, D., Gilbert, R, Stiglitz, J. and Tirole, Jean (1983) “Preemption, leapfrogging and 

competition in patent races”, European Economic Review, Vol 22, p3-31 

Gilbert, R J and Newbery, D M G (1982) “Preemptive patenting and the persistence of monopoly”, 

American Economic Review, 72:514-526. 

Greenlee, P, Cassiman, B, (1999), “Product market objectives and the formation of research joint 

ventures”, Managerial and Decision Economics, 20(3), pp.115-130 

Grossman, G M and Shapiro, C (1987) “Dynamic R&D competition”, Economic Journal, 

97:372-387. 

Hagedoorn, J, Schakenraad, J, (1992), “Leading companies and networks of strategic alliances 

in information technologies”, Research Policy, 21, pp. 163-190. 

Harris. Christopher and Vickers, John (1987) “Racing with uncertainty”, Review of Economic 

Studies, Vol 54, No 1, p1-21 

Jacquemin, A, Soete, L, (1994), “Cooperation in R&D, efficiency and European antitrust policy”, 

European Review, 2, pp.65-72. 

Judd, K L (1985b) “On The Performance of Patents”, Econometrica, 53:567-585. 

Kamien, M, Muller, E, and Zang, I, (1992), “Research joint ventures and R&D cartels”, American 

Economic Review, 82, pp.1293-1306. 

Katsoulacos, Y, Ulph, D, (1998), “Endogenous spillovers and the performance of research joint 

ventures”, The Journal of Industrial Economics, 3, pp.333-357. 

Katz, L M and Shapiro, C (1986) “Technology Adoption in the Presence of Network Externalities”, 

Journal of Political Economy, 94: 822-841. 

Katz, M L and Shapiro, C (1986) “How to License Intangible Property”, Quarterly Journal of 

Economics, 101:567-590. 

Katz, M L and Shapiro, C (1986) “Technology Adoption in the Presence of Network Externalities”, 

Journal of Political Economy, 94: 822-841. 

Katz, M L and Shapiro, C (1987) “R&D Rivalry with Licensing or Imitation”, American Economic 

Review, 77:402-420. 

71 

289



Katz, M L, (1986), “An analysis of cooperative research and development”, RAND Journal of 

Economics, 17 (Winter), pp.527-543. 

Katz, M L, Ordover, J A, (1990), “R&D cooperation and competition”, Brookings Papers on 

Economic Activity, Special Issue, pp. 137-91. 

Kleinknecht, A, Reijnen, J, (1992), “Why do firms cooperate on R&D? an empirical study”, 

Research Policy, 21, 347-360. 

Lee, T and Wilde, I (1980) “Market Structure and Innovation: A Reformulation”, Quarterly Journal 

of Economics, 94:429, 436. 

Levin R, Klevorick, A K, Nelson, RR and Winter, S G (1987) “Appropriating the returns from 

industrial research and development”, Brookings Papers on Economic Activity, vol. 3:783-820. 

Loury, G C (1979) “Market Structure and Innovation”. Quarterly Journal of Economics. 93(3): 

395-410. 

Perloff, Jeffrey , Karp, Larry and Golan, Amos (2007) Estimating Market Power and Strategies, 

Cambridge,Ma.: Cambridge Univ. Press 

Poyago-Theotoki, J, (1999), “A note on endogenous spillovers in a non-tournament R&D 

duopoly”, Review of Industrial Organization, 15(3), November, pp.253-262. 

Qiu, L. (1997) “On the dynamic efficiency of Bertrand and Cournot equilibria”, Journal of 

Economic Theory, NO 75, p213-229 

Reinganum, J F (1981a) “Dynamic Games of Innovation”, Journal of Economic Theory, 25:21-41; 

1985 “Corrigendum” 35:196-197. 

Reinganum, J F (1981b) “On the Diffusion of New Technology: A Game Theoretic Approach”, 

Review of Economic Studies, 48:395-405. 

Reinganum, J F (1981c) “Market Structure and the Diffusion of New Technology”, Bell Journal, 

12:618-624. 

Reinganum, J F (1982a) “A Dynamic Game of R and D: Patent Protection and Competitive 

Behaviour”, Econometrica, 50:671-688. 

Reinganum, J F (1982b) “Strategic Search Theory”, International Economic Review, 23:1-15. 

Reinganum, Jennifer F. (1983) “Uncertain innovation and the persistence of monopoly”, American 

Economic Review, Vol 73, p741-48 

Reinganum, Jennifer F. (1985) “Innovation and industry evolution”, Quarterly Journal of 

Economics, p81-99 

Segerstrom, P., 2007. Intel economics. International Economic Review 48 (1), 247–280 

Simpson, R D, Vonortas, N S, (1994), “Cournot equilibrium with imperfectly appropriable R&D”, 

The Journal of Industrial Economics, 1, pp.79-92 

72 

290



Spence, A M (1984) “Cost Reduction, Competition, and Industry Performance”, Econometrica 

52:101-121. 

Sutton, John (1998) Technology and Market Structure, Cambridge, Massachusetts and London, 

England: The MIT Press. 

Suzumura, K, (1992), “Cooperative and non-cooperative R&D in an oligopoly with spillovers”, 

American Economic Review, 82, pp.1307-1320. 

Vickers, John (1986) “The evolution of industry structure when there is a sequence of 

innovations”, Journal of Industrial Economics, Vol 35,p1-12 

4.2 Endogenous Growth 

Aghion,Philippe and Giffith,Rachel (2005) Competition and Growth, Cambridge,Ma., The MIT 

Press 

Aghion, Philippe and Howitt, Peter (1988) “Growth and Cycles through Creative Destruction”, MIT 

Mimeo. 

Aghion, Philippe and Howitt, Peter (1992) “A model of growth through creative destruction”, 

Econometrica, Vol 60,p323-51. 

Aghion, Philippe and Howitt, Peter (1998) Endogenous Growth Theory, Cambridge, 

Massachusetts and London, England: The MIT Press. 

Aghion, Philippe and Shankerman, Mark (2000), “A model of market enhancing infrastructure” 

CEPR Discussion Paper, NO.2461. 

Aghion, Philippe, Bloom, Nicholas, Blundell, Richard, Griffith, Rachel and Howitt, Peter (2002) 

“Competition and innovation: an inverted U relationship” NBER Working Paper, no.9269. 

Aghion, Philippe, Dewatripont, M. and Rey, P. (1997) “Competition, financial discipline and 

growth”, Review of Economic Studies, Vol 66, No 4, p825-852. 

Aghion, Philippe, Harris, Christopher and Vickers, John (1995) “Competition and Growth with 

Step-by-Step Technological Progress”, mimeo. 

Aghion, Philippe, Harris, Christopher and Vickers, John (1997) “Competition and growth with 

step-by-step innovation: an example”, European Economic Review, Vol 41, p771-782. 

Aghion, Philippe, Harris, Christopher, Howitt, Peter and Vickers, John (2001) “Competition, 

imitation and growth with step-by-step innovation”, Review of Economic Studies, Vol 68, p467- 

492. 

Aghion,Philippe and Griffith,Rachel (2005) Competition and Growth, Cambridge,Ma., The MIT Pr 

Ahn, Sanghoon, “Competition, innovation and productivity growth: a review of theory and 

evidence”, OECD Economic Department Working Paper, no 317. 

Arrow, K J (1962) “Economic Welfare and the Allocation of Resources for Invention”, in 

Universities National Bureau Committee for Economic Research. The rate and direction of 

inventive activity Princeton, Princeton University Press. 

73 

291



Blundell, Richard, Griffith, Rachel and Van Reenen, John (1999) “Market share, market value and 

innovation in a panel of British manufacturing firms”, Review of Economic Studies, Vol 66, No 3, p 

529-554. 

Chen,Derek H.C. and Kee,H.L. (2005),”A Model on Knowledge and Endogeneous Growth”, World 

Bank Policy Research Working Paper 3539, March 

D’Aspremont, Claude, Dos Santos Ferreira, Rodolphe and Gérard-Varet, Louis-André (2002) 

“Strategic R&D investment, competition toughness and growth”, mimeo. 

Gilbert, Richard J. and Newbery, David M.G. (1982) “Preemptive Patenting and the Persistence 

of Monopoly”, American Economic Review, 72:514-526. 

Grossman, Gene M. and Helpman, Elhanan. (1991) “Quality ladders and product cycles”, 

Quarterly Journal of Economics, Vol 106, p557-586. 

Lucas, R E, (1988). “On the Mechanics of Economic Development”, Journal of Monetary 

Economics 22(1): 3-42. 

Nickell, Stephen, J. (1996) “Competition and corporate performance”, The Journal of Political 

Economy, Vol 104, No 4, p724-746. 

Romer, Paul M. (1986) “Increasing Returns and Long Run Growth”, The Journal of Political 

Economy 94(5): 1002-1037. 

Romer, Paul M. (1990) “Endogenous Technological Change”, The Journal of Political Economy 

98(5) part 2: 71-102. 

Solow, R M (1956) “A Contribution to the Theory of Economic Growth”, Quarterly Journal of 

Economics 70(1): 65-94. 

Solow,Robert (1997) Learning from ‘Learning by Doing’, Stanford: Stanford Univ. Press 

Stokey, Nancy L. (1988) “Learning by doing and the introduction of new goods”, Journal of 

Political Economy, Vol 96, No 4 (August), p701-171. 

Young, Alwyn (1991) “Learning by Ding and the Dynamic Effects of International Trade”, Quarterly 

Journal of Economics 106(2): 369-406. 

Young, Alwyn (1993) “Invention and Bounded Learning by Doing”, Journal of Political Economy 

101(3): 443-472. 

4.3 New Economic Geography 

Davis, Ronald R. and David E.Weinstein (2002) Bones, Bombs, and Break Points: The 

Geography of Economic Activity, 92(5), pp. 1269-1289. 

Fujita, Masahisa, Paul Krugman and Anthony J. Venables (2001): The Spatial Economy: Cities, 

Regions and International Trade, MIT Press. 

74 

292



Fujita, Masahisa and Jacques-François Thisse, (2002) Agglomeration and growth with migration 

and knowledge externalities, mimeo, April. 

Kim, S. (1995) “Expansion of markets and the geographic distribution of economic activities. The 

trends in US regional manufacturing structure: 1860-1987” Quarterly Journal of Economics, 

110(4), pp. 881-908. 

Krugman, Paul Anthony J. Venables (1995) “Globalization and the Inequality of Nations” 

Quarterly Journal of Economics, 110(4), pp. 857-880. 

Krugman, Paul (1980) “Scale Economies, Product Differentiation, and the Pattern of Trade” 

American Economic Review, 70(5), pp. 950-959. 

Krugman, Paul (1981) “Intraindustrial Specialisation and the Gains from Trade” Journal of Political 

Economy, 89(5), pp. 959-973. 

Krugman, Paul (1991) “Increasing Returns and Economic Geography” Journal of Political 

Economy, 99(3), pp. 483-499. 

Martin, Ron and Peter Sunley (1996) “Paul Krugman’s Geographical Economics and Its 

Implications for Regional Development Theory. A Critical Assessment” Economic Geography, 

72(3), pp. 259-292. 

Midelfart-Knarvik K.H., H.G. Overman, S.J. Redding and A.J. Venables, (2000) The Location of 

European Industry, European Commission, DG ECFIN, Economic Papers No. 142, April. 

Neary, Peter (2001) “Of Hypes and Hyperbolas: Introducing the New Economic Geography” 

Journal of Economic Literature, 39, June, pp. 536-561. 

Overman, Henry G., Stephen Redding and Anthony J. Venables (2001) “The Economic 

Geography of Trade, Production and Income. A Survey of Empirics” London School of 

Economics, Centre for Economic Performance, September. 

Robert-Nicoud, Frédéric (2002) “The structure of the simple ‘New Economic Geography’ models” 

London School of Economic, Centre for Economic Performance, August. 

Samuelson, Paul A. (1954) “The Transfer Problem and Transport Costs II: Analysis of Effects of 

Trade Impediments” The Economic Journal, 64, pp. 264-289. 

4.4 Evolutionary Economics 

Baumol,William (1993) Entrepreneurship, Management and the Structure of Payoff, Cambridge,Ma. 

The MIT Press Chapts. 1-7 

Bergek, Anna and Staffan Jacobsson (2002) The emergence of a growth industry: A comparative 

analysis of the German, Dutch and Swedish wind turbine industry, in Stan Metcalfe and Uwe 

Cantner: Change, transformation and development, Springer pp. 197-228. 

Denicolò, Vincenzo and Zanchettin, Piercarlo (2002) “Competition and growth in a neo- 

Schumpeterian model”, mimeo Università di Bologna. 

75 

293



Dosi, G (1988) “Sources, procedures, and microeconomic effects of innovation”, Journal of 

Economic Literature, Vol 26, No 3, p 1120-1171. 

Edquist, Charles (2001) The Systems of Innovation Approach and Innovation Policy: An account 

of the state of the art, lead paper presented at the DRUID Conference, Aalborg, June 12-15, 

2001. 

Etro, Federico (2008), “Growth Leaders”, Journal of Macroeconomics 30, 1148-1172 

Fagerberg, J (2002) “A layman’s guide to evolutionary economics”, mimeo, September 2002. 

Foster, J (2001) “Competition, competitive selection and economic evolution”, in Garrouste, P and 

Ioannides, S (eds) Evolution and Path Dependence in Economic Ideas, London: Routledge. 

Freeman (1987) Technology, policy and economic performance: lessons from Japan, 

London: Pinter. 

Freeman (1994) Critical survey: The economics of technical change, Cambridge Journal of 

Economic, 18, pp. 463-514. 

Hayek, F. A. "The Use of Knowledge in Society" American Economic Review, 35, 

(September 1945), p 519-530. 

Hauknes, Johan (1999) “Innovation systems and capabilities” STEP Working Paper, Oslo. 

Levin, R.C. – Klevorick, A.K. – Nelson, R.R. – Winter, S.G. (1987) Appropriating the returns from 

industrial research and development, Brookings Papers on Economic Activity, 3, pp. 783-820. 

Liu, Xieling – White, Steven (2001) Comparing innovation systems: A framework and application 

to China`s transitional context, Research Policy, 30, pp. 1091-1114. 

Lundvall, Bengt-Åke (1992) National systems of innovation: Towards a theory of innovation and 

interactive learning, London: Pinter. 

Metcalfe, J S (1995), “The Economic Foundations of Technology Policy: Equilibrium and 

Evolutionary Perspectives”, in Paul Stoneman ed., Handbook of the Economics of Innovation and 

Technological Change, Oxford UK and Cambridge USA: Blackwell. 

Metcalfe, J S (1998) Evolutionary Economics and Creative Destruction, London: Routledge. 

Nelson, Richard R. (1996) The sources of economic growth, Harvard University Press, Chapter 

10: National innovation systems: A retrospective on a study, pp. 274-301. 

Nelson, R and Winter, G (2002) “Evolutionary theorizing in economics”, Journal of Economic 

Perspectives, Vol 16, No 2, p 23-46. 

Nelson, R and Winter, S (1982) An Evolutionary Theory of Economic Change, London: Belknap 

Press. 

76 

294



Pavitt, K (2002) “Knowledge about knowledge since Nelson & Winter: a mixed record”, SPRU 

Electronic Working Paper Series, No. 83. 

Saviotti, Pier Paolo (2000) Networks, national innovation systems and self-organisation, 

University of Pierre Mendes-France, mimeo. 

Segerstrom, P., 2007. Intel economics. International Economic Review 48 (1), 247–280 

Stoneman, Paul (ed) (1998) Handbook of the Economics of Innovation and Technological 

Change, Oxford: Blackwell Publisher. 

Winter, S (1987) “Knowledge and competence as strategic assets” in Teece, D (ed) The 

Competitive Challenge, New York: Harper and Row. 

Zeira, J.(2003) Innovations, Patent Races and Endogenous Growth, mimeo. Harvard University. 

4.5 Cliometrics 

Atkenson, Andrew and Patrick Kehoe (2001) “The Transition to a New Economy After the 

Second Industrial Revolution”, NBER Working Paper, No. 8676, December 

Coase, Ronald (1960): The Problem of Social Cost, Journal of Law and Economics, 3(1), pp. 1- 

44 

Coleman, D.C. and Christine MacLeod (1986) “Attitude to New Techniques: British Businessmen” 

1800-1950, Economic History Review, 39(4), pp. 588-611 

Conrad, Alfred H. and John R. Meyer (1957) “Economic Theory, Statistical Inference, and 

Economic History” Journal of Economic History, 17(4), pp. 524-44 

Conrad, Alfred H. and John R. Meyer (1958) “The Economics of Slavery in the Ante Bellum 

South” Journal of Political Economy, 66(2), pp. 95-130. 

Crafts, N.F.R (1987) “Cliometrics, 1971-1986: a survey” Journal of Applied Econometrics, 2(3), 

pp. 171-192 

Crafts, Nicholas (2002) “Productivity growth in the industrial revolution: A new growth accounting 

perspective”, mimeo, London School of Economics, January. 

Crafts, Nicholas F.R and C.K. Harley (1992) “Output growth and British industrial revolution: a 

restatement of the Crafts-Harley view” Economic History Review, 45(4), pp. 703-730 

Crafts, Nicholas F.R and Terence C. Mills (1997) “Endogenous innovation, trend growth, and the 

British industrial revolution: reply to Greasley and Oxley” Journal of Economic History, 57(4), pp. 

950-956. 

Crafts, Nicholas F.R. (1995a) “Exogenous or Endogenous Growth? The Industrial Revolution 

Reconsidered” Journal of Economic History, 55(4), pp. 745-772 

Crafts, Nicholas F.R. (1995b). “Macroinventions, Economic Growth, and `Industrial Revolution` in 

Britain and France” Economic History Review, 48(3), pp. 591-598. 

77 

295



Crafts, Nicholas F.R. (1996) “The First Industrial Revolution: A Guided Tour for Growth 

Economists” American Economic Review, 86(2), pp. 197-201. 

Fischlow, Albert (1965) American Railroads and the Transformation of the Ante-Bellum Economy, 

Harvard University Press, Chapter 2. 

Fogel, Robert W. (1962): “A quantitative approach to the study of railroads in American economic 

growth: a report of some preliminary findings” Journal of Economic History, 22(2), pp. 163-197. 

Fogel, Robert W. (1964): Railroads and the American Economic Growth: <strong>Essays</strong> in Econometric 

History, Baltimore: Johns Hopkins University Press. 

Fogel, Robert W. (1965) “The reunification of economic history with economic theory” American 

Economic Review, 55 (1/2), pp. 92-98. 

Fogel, Robert W. (1966a) “The new economic history: its findings and methods” The Economic 

History Review, 19 (3), pp. 642-656. 

Fogel, Robert W. (1966b) “Railroads as an analogy to the space effort: some economic aspects” 

Economic Journal, 76(301), pp. 16-43. 

Goldin, Claudia (1995) “Cliometrics and the Nobel” Journal of Economic Perspectives, 9 (2), pp. 

191-208. 

Greasley, David and Les Oxley (1994) “Rehabilitation sustained: the industrial revolution as a 

macroeconomic epoch” Economic History Review, 47(4), pp. 760-768. 

Greasley, David and Les Oxley (1997a) “Endogenous growth or “Big Bang”: two views of the first 

industrial revolution” Journal of Economic History, 57(4), pp. 935-949. 

Greasley, David and Les Oxley (1997b) “Endogenous growth, trend output, and the industrial 

revolution: reply to Crafts and Mills” Journal of Economic History, 57(4), pp. 957-960. 

Gregerson, Brigitte and Björn Johnson, (1998) National Systems of Innovation as a Framework 

for Innovation Policy, The United Nations University, April. 

Heckman, James J. (1997) “The value of quantitative evidence on the effect of the past and the 

present” American Economic Review, 87(2), pp. 404-408. 

Jones, Charles I. (1999) “Was an industrial revolution inevitable? Economic growth over the very 

long run” NBER Working Paper, No. 7375, October. 

Khan, Zorina B. (1995) “Property rights and patent litigation in early nineteenth-century America” 

Journal of Economic History, 55(1), pp. 58-97. 

Landes, David (1994) “What room for accident in history? Explaining big changes by small 

events” Economic History Review, 47(4), pp. 637-656. 

Landes,David (1998), The Wealth and Poverty of Nations, W.W.Norton, New York 

78 

296



Lorenz, Edward H. (1991) “An evolutionary explanation for competitive decline: the British 

shipbuilding industry, 1890-1970” Journal of Economic History, 51(4), pp. 911-935. 

MacLeod, Christine (1992) “Strategies for innovation: the diffusion of new technology in 

nineteenth-century British industry” Economic History Review, 45(2), pp. 285-307. 

Maddison, Angus (2001) The World Economy: A Millennial Perspective, OECD, Development 

Centre Studies. 

McCloskey, Donald N. (1978) “The achievements of the cliometric school” Journal of Economic 

History, 38(1), pp. 13-28 

Mokyr (1992) “Technological inertia in economic history” Journal of Economic History, 52(2), pp. 

325-338. 

Mokyr, Joel (1993) Introduction: The New Economic History and the Industrial Revolution, in Joel 

Mokyr (ed.): The British Industrial Revolution: An Economic perspective, Oxford. 

Mokyr, Joel (1997) Invention and Rebellion: Why Do innovation Occur at All? An Evolutionary 

Approach, paper presented at the Conference on Minorities and Economic growth, Bar Ilan 

University, June 2-5 1997. 

Mokyr, Joel (1999a) Innovation and Rebellion: Why do Innovations Occur at all? An Evolutionary 

Approach, in Elise Brezis and Peter Temin, eds., Minorities and Economic Growth. Amsterdam: 

Elsevier Publishers, pp. 179-203. 

Mokyr, Joel (1999b) Knowledge, Technology, and Industrial growth During the Industrial 

Revolution, Northwestern University, mimeo. 

North, Douglass C. (1968) “Sources of productivity change on ocean shipping, 1600-1850” 

Journal of Political Economy, 76 (5), pp. 953-970. 

North, Douglass C. (1990) Institutions, institutional change, and economic performance, 

Cambridge University Press. 

North, Douglass C. (1991) “Institutions” Journal of Economic Perspectives, 5(1), pp. 97-112. 

North, Douglass C. (1994) “Economic performance through time” American Economic Review, 

84(3), pp. 359-368. 

North, Douglass C. (1997) “Cliometrics – 40 years later” American Economic Review, 87(2), pp. 

412-414. 

North, Douglass (2005), Understanding the Process of Economic Change, Princeton Univ. Press 

O’Rourke, Kevin H. and Jeffrey G. Williamson, (2002) “From Malthus to Ohlin, trade, growth and 

distribution since 1500” NBER Working Paper, No. 8955, May 

Olson, Mancur (1996) “Big bills left on the sidewalk: why some nations are rich, and other poor” 

Journal of Economic Perspectives, 10(2), pp. 3-24. 

Olson,Mancur (1982),The Rise and Decline of Nations, Yale Univ. Press 

79 

297



Pearce, Robert (2002): National systems of innovation and the international technology strategies 

of multinationals, University of Reading. 

Phillips, William H. (1992) “Patent growth in the old dominion: the impact of railroad integration 

before 1880” Journal of Economic History, 52(2), pp. 389-400. 

Posner, Richard (1980) “A theory of primitive society, with special reference to the law” Journal of 

Law and Economics, 23, pp. 1-54. 

Schoser, Christof (1999) “The institutions defining national systems of innovation: a new 

taxonomy to analyse the Impact of globalisation”, University of Freiburg, Discussion Paper. 

Shiue, Carol H (2002) “Transport costs and the geography of arbitrage in eighteenth-century 

China” American Economic Review, 92 (5), pp. 1406-1419. 

Smith, Chauncey (1890) “A century of patent law” Quarterly Journal of Economics, 5(1), pp. 44- 

69. 

Sokoloff, Kenneth L. (1988) “Incentive activity in early industrial America: evidence from patent 

records, 1790-1846” Journal of Economic History, 48(4), pp. 813-850. 

Sullivan, Richard J. (1994) “Estimates of the value of patent rights in Great Britain and Ireland, 

1852-1876” Economica, 61(241), pp. 37-58. 

Triplett, Jack E. (1999) “The Solow productivity paradox: what do computers do to productivity?” 

Canadian Journal of Economics, 32(2), pp. 309-334. 

Wallis, John J. – North, Douglass C. (1986): Measuring the Transaction Sector in the American 

Economy, in Engerman, S.L. – Gallman, R.E. (eds), Long-term factors in American economic 

growth, University of Chicago Press, pp. 95-148. 

4.6 Strategic Management and Networks 

Audretsch, D, Baumol, W and Burke, A (2001) “Competition policy in dynamic markets”, 

International Journal of Industrial Organization, Vol 19, p 665-693. 

Bonaccorsi, Andrea and Paola Giuri (2000) “Industry life cycle and the evolution of an industry 

network” LEM Working Paper Series, No. 4, December 

Ellig, Jerry and Daniel Lin (2001): A taxonomy of dynamic competition theories, in Jerry Ellig (ed): 

Dynamic Competition and Public Policy: Technology, Innovation and Antitrust Issues, Cambridge 

University Press. 

Evans, D and R. Schmalensee (2001) “Some economic aspects of antitrust analysis in 

dynamically competitive industries“ NBER working paper, No.8268. 

Farrel, Joseph and Garth Saloner, “Coordination through committees and markets”, RAND 

Journal of Economics, Vol 16 (Spring), p70-83 

<strong>Gottinger</strong>,H.W. (2003), Economies of Network Industries, London: Routledge, 2003 

<strong>Gottinger</strong>,H.W. (2006), Innovation, Technology and Hypercompetition, London: Routledge, 

2006 

80 

298



Grant, Robert M. (2002): Contemporary Strategy Analysis, 4th edition, Blackwell Publishing 

Katz, Michael L. and Shapiro, Carl (1986) “Product compatibility choice in a market with 

technological progress”, Oxford Economic Papers, Special Issue on the New Industrial 

Economics, November. 

Katz, Michael L. and Shapiro, Carl (1994) “Systems competition and network effects”, Journal of 

Economic Perspectives, Vol 8, No 2 (Spring), p93-115 

Klein, Jeremy (2002): Beyond Competitive Advantage, Strategic Change, 11(6), pp. 317-326. 

Klepper, Steven (1996) “Entry, exit, growth and innovation over the product life cycle” American 

Economic Review, Vol. 86, No 3, pp. 562-583. 

Klepper, Steven and Kenneth Simons (2001), “Industry shakeouts and technological change” 

mimeo. 

Jovanovic, Boyan (1982) “Selection and the evolution of industry” Econometrica, Vol. 50 No 3, pp. 

649-670. 

Leonard-Barton, D (1992) “Core capabilities and core rigidities: a paradox in managing new 

product development” Strategic Management Journal, Summer, pp. 111-125. 

Liebowitz, S.J. and Stephen E. Margolis (1994) “Network externality: an uncommon tragedy” 

Journal of Economic Perspectives, p133-150. 

Liebowitz, S.J. and Stephen E. Margolis, (1995) “Are network externalities a new source of 

market failure?” Research in Law and Economics¸ Vol 17, p.41-46 

March, J.G. (1991) “Exploration and exploitation in organizational learning” Organizational 

Science, 2, pp. 71-87. 

Porter, Michael (1990): The Competitive Advantage of Nations, The Free Press, New York 

Porter, Michael E (1985): Competitive Advantage, The Free Press, New York 

Posner, R (2000) “Antitrust in the new economy“ John M. Olin Law and Economics Working 

Paper, No 106, University of Chicago. 

Shapiro, C and H.R. Varian (1999) Information Rules: A Strategic Guide to the Network 

Economy, Harvard Business School Press, Boston. 

Simons, Kenneth (2001) “Product market characteristics and the industry life cycle”, mimeo. 

Tushman, M L and E. Romanelli (1994) “Organizational transformation as punctuated equilibrium: 

an empirical test” Academy of Management Journal, 37, pp. 1141-1166. 

81 

299



Wecker, Claudia (2003) “Market performance and competition: a product life cycle model”, 

Technovation, Vol 23, No 4, pp. 281-290. 

Comprehensive treatments of Reviewer(<strong>Gottinger</strong>) 

Economies of Network Industries, Routledge : London, New York 2003 

Innovation, Technology and Hyper-Competition Routledge, London, New York 2006 

Strategic Economics of Network Industries, NovaScience, New York 2009 

Strategic Alliances in Biotech and Pharmaceuticals, NovaScience, New York 2010 (with F.Floether 

and C.Umali) 

Strategies of Economic Growth and Catch-Up, NovaScience, New York 2012 (with M. Goosen, and 

M.Takashima) 

82 

300

3.2 ECONOMIC GROWTH, CATCHING-UP, FALLING BEHIND AND GETTING AHEAD 


301



302



303



304



305



306



307



308



309



310



311



312



313



314

3.3 CATCH-UP AND CONVERGENCE 


Journal of Applied Economics and Business 

CATCH‐UP AND CONVERGENCE: 

MECHANISM DESIGN FOR 

ECONOMIC DEVELOPMENT* 

*Paper presented at the 1st International Congress Jean Baptiste Say, Research Network on Innovation, 

Boulogne‐sur‐Mer, France, August 27‐30, 2014 

<strong>Hans</strong> <strong>Gottinger</strong> 


gottingerhans@gmail.com 

Abstract 

This paper identifies factors that promote long‐run sustainable economic growth and industrial catchup 

in achieving superior performance. An informationally efficient mechanism design of the Hayek‐ 

Hurwicz type would contain as essential elements allocative efficiency with a limited state role, 

fostering of private entrepreneurship for high intensity dynamic competition through industry 

specific ‘technological racing’ within a supporting market structure, free trade and intellectual 

property rights (IPRs), flexible labour markets, low taxes on labour, capital income and profits. 

Emphasis on high value‐added network industries directed toward increasing returns. If resource 

endowment among countries are similar these factors would eventually lead to convergence through 

catch‐up performance, otherwise they would lead to divergence. 

Key words 

Catch‐Up; Convergence; Mechanism Design; Economic Development 

Taking a philosophical point of view, this may be seen as the mechanics for the 

implementation of Adam Smith’s invisible hand: despite private information and pure selfish 

behavior, social welfare is achieved. All the field of Mechanism Design is just a generalization 

of this possibility. – Noam Nisan (2007) 

INTRODUCTION 

What are the major factors for economies to succeed on catching‐up in terms of GDP 

or GDP per capita (of purchasing power parity PPP) as an indicator of prosperity? 

As has been broadly covered in the economic growth and development literature 

(<strong>Gottinger</strong> & Goosen, 2012), there have been convergence theories on advanced 

(developed) economies that have been partially verified for some OECD economies, 

67 

315




Catch‐up and Convergence: Mechanism Design for Economic Development 

but equally there have been observations of divergence among some developed and 

developing (transitional) economies that appear to show a growing gap. Technology 

adoption within regions of a national economy (say , China) were a major factor of 

less developed regions to catch up. Catch‐up metrics could also be identified in 

disaggregated form next to aggregates such as GDP (total), GDP per head, Human 

Development Index (HDI), for example, industry/technology index, infrastructure 

indicator (transportation, communication networks), R&D expenditure 

(government, private industry) and competitiveness (global market share of key 

industries). A process of catching up induced by industrial races may tend to 

converge over time within a bloc of similar countries if technological and 

educational endowment is similar as covered by Abramovitz’ (1986) famous ‘social 

capabilities’, as key industries engage in more incremental and complementary 

innovation that through international trade and foreign direct investment (FDI) 

spread to emerging industries in likewise developing economies. In the post World 

War II history what was the mechanism that induced Japan to be on a path of catchup 

growth in the 1960s and 1970s in terms of per capita income growth? As 

suggested in the Asian miracle (Krugman, 1994), one factor was input growth in key 

value‐added industries through capital expansion, the other cumulative 

technological advancement through largely incremental innovation leading to 

superiority in some industries. But as soon as the technological frontier was close to 

be reached it became increasingly difficult to dominate the market. Also in a wide 

array of high‐tech markets it became increasingly decisive to have integrative 

technologies to catch complementary and increasing returns markets. 

The paper shows that catch‐up processes should be primarily understood as 

technological races, establishing new industries that allow free flow of information 

through entrepreneurial activity and innovation. Under these circumstances, it is 

more likely that an economic mechanism design generating more information, 

choices, economic freedom and market transparency supported by democratic 

institutions, will have a better economic performance record, in a sustainably long 

run, than any other that fails on the information distribution and strategic incentive 

side, which would coincide with various types of socialist systems. 

The content of the paper is as follows. Section 2 gives a brief historical outline on the 

interrelation between economic development and catching‐up paths. Section 3 

sketches the basic structure of algorithmic mechanism design (AMD) as it pertains to 

computational economic systems, in particular to superior performance of 

decentralized (distributed) dynamic systems. Section 4 shows the connection 

between industrial competition and ist aggregation to macro competition on a 

national or regional level – in view of technological racing. Section 5 identifies 

technological frontiers on the firm level (FTF) as embedded on the industry level 

(ITF), amenable to a statistical profiling of technological evolution and innovation. It 

shows some rules of technological race behaviour resulting in statistical indicators 

68 JOURNAL OF APPLIED ECONOMICS AND BUSINESS, VOL.2, ISSUE 3 – SEPTEMBER , 2014, PP. 67‐95 

316




on an industry level. Sec. 6 looks at industrial (in)efficiencies in catch‐up countries as 

benchmarked against the industrial leader. Conclusions follow and open problems 

are discussed in the last section. 

A BRIEF HISTORICAL REVIEW 

In a historical context, a catch‐up process becomes multidimensional, emerging with 

England, moving to the US and a few European economies, Germany, France, later 

to Japan and the newly industrialized economies (NIEs) in Asia overlapping the 

BRICs. The argument is that due to expansion of trade through globalization the 

catch‐up process is multidimensional and multispeed (Wan, 2004). Even for a few 

increasing in overall speed the benchmark of catch‐up keeps shifting from one or a 

few to many more. For example, in the past WW II period, industrial and economic 

growth of Japan was driven toward the US economy less so toward Europe; the 

pattern of growth and catch‐up of China hits many more emerging economies with 

significant potentials. We have modelled this by a more complex differential game as 

in <strong>Gottinger</strong> and Goosen (2012). 

As a precursor of economic mechanism design, from an economic history 

perspective, we first identify Gershenkron (1962) who emphasized state action on 

industrialization setting free guided entrepreneurial activities through targeted 

industrial policy (along the line of Japan and later South Korea, Taiwan) (Lee, 2013). 

This compares to Abramovitz’ (1986) and Abramovitz and David’s (1992) catching 

up process where self‐reinforcing, industry specific, competitive market forces, 

rather than state action, would initiate and sustain a technological race resulting in 

leadership positions across a range of industries. Here the roots of growth rest in the 

microeconomic industry structure and new industry creating high technology 

entrepreneurship as conceptionally and empirically explored by Scherer (1992, 1999). 

The World Bank in the past takes a middle ground, propagating the state’s role to 

develop social capital which then by itself creates social capability to induce an 

upward potential through market forces. A more activistic role has recently been 

favoured by Justus Lin (2013), the former Chinese Chief Economist of the World 

Bank. Amsden’s development statism is a further extension of Gershenkron, 

catching‐up as a process of learning how to compete through ’Late Industrialization’ 

(Amsden, 1989). Other than the main catchup processes listed here have been 

reviewed by Burkett and Hart‐Landsberg (2003), and Fagerberg and Godinho (2004). 

In ’Achieving Rapid Growth’ Sachs and Warner (1996), in comparing growing 

middle income countries identify major factors such as allocative efficiency (with 

government interaction low, markedly less regulation), high degree of competition, 

free trade. flexible labour markets, low taxes on labour, capital income and profits. 

While convergence among developing economies may be facilitated by becoming 

69 

317





more alike in terms of sources of growth such as technological progress, innovation, 

levels of physical capital, labour productivity, quality of human capital, extent of 

trade openess, industrial structure and institutional framework, one clear 

distinguishing factor may involve increasing returns mechanisms (IRMs). IRMs 

eliminate any kind of convergence and allow for ongoing quasi‐linear growth. 

Anecdotal observations as put forward by Easterly and Levine (2002) can be 

summarized as stylized facts on growth mechanisms. 

(1) Factor accumulation does not account for growth differences but total factor 

productivity does for a substantial amount; 

(2) There are hugely growing differences in GDP per capita. On a global scale 

divergence and not conditional convergence is the major concern in 

development policy; and 

(3) Among developing economies growth is not persistent over time. 

If movable through market forces all factors of production flow to the same places 

suggesting important externalities. 

A SIMPLE MECHANISM DESIGN FOR CATCHUP AND 

DEVELOPMENT 

A simple mechanism design emanates from the following paradigmatic situation: 

Let there be k agents in an (Internet) economy that collectively generate demand 

competing for resources from a supplier. The supplier herself announces prices 

entering a bulletin board accessible to all agents (as a sort of transparent market 

institution). In a simple form of a trading process we could exhibit a “tatonnement 

process” on a graph where the agents set up a demand to the suppliers who 

advertise prices on a bulletin board which are converted to new prices in interaction 

with the agents. 

The tatonnement process in economics is a simple form of an algorithmic 

mechanism design (AMD) (Nisan & Ronen, 2001), which in modern computer 

science (CS) emerges as an offspring to algorithmic game theory (Nisan et al, 2007). 

Algorithmic ingredients apply to rational and selfish agents having well defined 

utility functions repesenting preferences over possible outputs of the algorithm. A 

payment ingredient motivates the agents. Mechanism Design Theory (MDT) aims to 

show how privately known preferences of the entire population can be aggregated 

towards a ’social choice’ that drives the mechanism of the whole economy. 

(i) Each agent or group of agents have some some private input represented 

by its type. Its type is embedded in public knowledge or resource 

endowment as the ’social environment’; 

(ii) There is a production function (or output specification) that associates 

each type t = t 1 , ..., t n with a set of socially allowable outputs (productions) 

oO; and 


318




(iii) 

Each agent’s preferences are codified as real‐valued utility functions noted 

by u i (t i ,o). 

The universal function is specified in linear terms as u i = r i + v i (t i ,o) where r i 0 is the 

initial endowment (or resource), or the agent’s wealth, which the agent attempts to 

optimize. 

If we assume that the mechanism is truthful to the extent that the agents report their 

real type then truth‐telling is the only dominant strategy. This is considered a 

truthful implementation. The societal objective function is simply the aggregation of 

all agents valuations. A ’maximizing mechanism design process’ (MDP) is called 

utilitarian or ’Hayekian’ if its objective function is the sum of all agents’ utilities. 

In the context of the internet economy an MDP would enable users of network 

applications to present their ’quality of service’ demands via utility functions 

defining the system performance requirements (<strong>Gottinger</strong>, 2013). The resource 

allocation process involves economic actors to perform economic optimization given 

scheduling policies, load balancing and service provisioning. 

Distributed algorithmic mechanism design (DMD) for internet resource allocation in 

distributed systems is akin to an equilibrium converging market based economy 

where selfish agents maximize utility and firms seek to maximize profits and the 

state keeps an economic order providing basic public goods and public safety 

(Feigenbaum et al, 2007). A distributed algorithmic mechanism design thus consists 

of three components: a feasible strategy space at the network nodes for each agent 

(or autonomous system), an aggregated outcome function computed by the 

mechanism and a set of multi‐agent prescribed strategies induced by the 

mechanism. 

For such DMDP in place an internet economy can be shown computationally and 

informationally efficient in the sense of Hurwicz and Reiter (2006), corroborated 

from a CS view by Conitzer and Sandholm (2002). Furthermore, for efficient macromanagement 

it would satisfy ’moral hazard’ and incentive based concerns, and in 

view of ’informational constraints’, e.g. adverse selection, it may also be superior in 

performance since in a market based decentralized capitalistic system due to keen 

competition among operating managers ‐ there will be more of higher type 

performing managers keeping truthful payoff‐relevant information than in a 

comparative socialist planning system with plenty of lower type performing 

managers providing lower quality or less payoff relevant information. 

A distributed algorithmic mechanism design (Feigenbaum et al, 2007) being 

computationally efficient in a large decentralized internet economy is a powerful 

paradigm to substantiate claims by Hayek (1945) that an industrialized economy 

71 

319





based on market principles has an overall better performance than socialist type 

economies of a similar nature and scale. It is a paradigm that even contemporary 

theorists of MDT seem to have partially missed (Myerson, 2008) and that puts 

historically the socialist planning debate in a new light which ironically, by some 

proposals, has been conducted on the basis of computational feasibility and 

superiority. 

Such a Hayekian MDP also extends to a dynamic real economy which also invokes 

highly desirable properties of incentive structures (Myerson, 2008) and knowledge 

creation through hi‐tech entrepreneurship. This suggests that an MDP of the 

Hayekian‐Hurwicz type should be more likely to generate a long‐run sustainable 

growth and development process with comparative greater welfare benefits than 

what any socialist type planning could achieve. This is in compliance with Hayekian 

development ideas as put forward recently by Easterly (2013). 

The focus of this paper would be to explore growth and development generating 

structures and factors that are compatible with a Hayek‐Hurwicz design scheme for 

the developing world. 

INDUSTRIAL AND MACRO COMPETITION 

The striking pattern that emerges in the innovative activities of firms is their rivalries 

for a technological leadership position in situations that are best described as races 

or hypercompetition (Harris & Vickers, 1987; <strong>Gottinger</strong>, 2006b). A race is an 

interactive pattern characterized by firms or nations constantly trying to get ahead of 

their rivals, or trying not to fall too far behind. In high‐technology industries, where 

customers are willing to pay a premium for advanced technology, leadership 

translates into increasing returns in the market through positive network 

externalities. Abramovitz (1986), in explaining the catch‐up hypothesis, lays stress 

on a country’s social capability in terms of years of education as a proxy of technical 

competence and its institutions. Competing behaviour is also a dynamic story of 

how technology unfolds in an industry. In contrast to any existing way of looking at 

the evolution of technology, racing behaviour, though in character more ’a 

productivity race than a runner’s race (Abramovitz & David, 1997), recognizes the 

fundamental importance of strategic interactions between competing firms. Thus 

firms take their rivals’ actions into account when formulating their own decisions. 

The importance of this characterization is at least twofold. At one level, racing 

behaviour has implications for appreciating technology strategy at the level of the 

individual firm; at the other level, for understanding the impact of policies that aim 

to spur technological innovation in an industry or country. 

On a national scale, simple catch‐up hypotheses have put emphasis on the great 

potential of adopting unexploited technology in the early stage and the increase of 

self‐limiting power in the later stage. However, the actual growth path of the 


320




technological trajectory of a specific economy may be overwhelmingly constrained 

by social capability. The capability endogenously changes as states of the economy 

and technology evolve. The success of economic growth due to diffusion of 

advanced technology or the possibility of leapfrogging is mainly attributable to how 

the social capability evolves. In other words, which effects become more influential: 

growing responsiveness to competition or growing obstacles to it on account of 

vested interests and established positions? 

Observations on industrial patterns in Europe, the US or Asia point to which type of 

racing behaviour is prevalent in global high‐ technology industries. The pattern 

evolving from such conduct could be benchmarked against the frontier pursuit type 

of the global technological leaders. Another observation relates to policy inferences 

on market structure, entrepreneurship, innovation activity, industrial policy and 

regulatory frameworks in promoting and hindering industry frontier races in a 

global industrial context. Does lagging behind one’s closest technological rivals’ 

cause a firm to increase its innovative effort? The term ‘race’ suggests that no single 

firm would want to fall too far behind, and that everyone would like to get ahead. If 

a company tries to innovate more when it is behind than when it is ahead, then 

‘catch‐up’ behaviour will be the dominant effect. Once a firm gets far enough ahead 

of its rivals, then the latter will step up their efforts to get closer. The leading 

company will slow down its innovative efforts until its competitors have drawn 

uncomfortably close or have surpassed it. Of course, the process of getting closer to 

may be much easier than surpassing the rival This process repeats itself every time a 

firm gets far enough ahead of its rivals. Of course, catch‐up may only consistently 

apply to the next rivals but will not impact the leader. This is called ’persistent 

leadership’. On a national level catchup processes like this may not lead to 

convergence. 

An alternative behaviour pattern would correspond to a business increasing its 

innovative effort if it gets far enough ahead, thus making catch‐up by the lagging 

companies increasingly difficult. For any of these businesses there appears to be a 

clear link to market and industry structure, as termed ‘intensity of rivalry’. 

We investigated two different kinds of races: one that is a frontier race between itself 

and the technological leader at any point in time (‘frontier‐ sticking’ behaviour), or it 

might try to actually usurp the position of the leader by ‘leapfrogging’ it. When there 

are disproportionately large payoffs to being in the technical lead (relative to the 

payoffs that a firm can realize if it is simply close enough to the technical frontier), 

then one would expect that leapfrogging behaviour would occur more frequently 

than frontier‐sticking behaviour. Alternatively, racing toward the frontier creates the 

reputation of being an innovation leader hoping to maintain and increase market 

73 

321





share in the future. All attempts to leapfrog the current technological leader might 

not be successful since many lagging firms might be attempting to leapfrog the 

leader simultaneously and the leader might be trying to get further ahead 

simultaneously. Correspondingly, one could distinguish between attempted 

leapfrogging and realized leapfrogging. 

Among the key issues to be addressed is the apparent inability of technologyoriented 

corporations to maintain leadership in fields that they pioneered. There is a 

presumption that firms fail to remain competitive because of agency problems or 

other suboptimal managerial behaviour within these organizations. An alternative 

explanation is that technologically trailing firms, in symmetric competitive 

situations, will devote greater effort to innovation, so that a failure of technological 

leaders to maintain their position is an appropriate response to the competitive 

environment. In asymmetric situations, with entrants challenging incumbents, 

research does demonstrate that startup firms show a stronger endeavor to close up 

to or leapfrog the competitors. Such issues highlight the dynamics of the race within 

the given market structure in any of the areas concerned. 

We observe two different kinds of market asymmetries with bearing on racing 

behaviour: risk‐driven and resource‐based. When the incumbents’ profits are large 

enough and do not vary much with the product characteristics, the entrant is likely 

to choose the faster option in each stage as long as he has not fallen behind in the 

contest. In view of resource‐based asymmetries, as a firm’s stage resource 

endowment increases, it could use the additional resources to either choose more 

aggressive targets or to attempt to finish the stage sooner, or both. Previous work 

has suggested that a firm that surges ahead of its rival increases its investment in 

R&D and speeds up, while a lagging firm reduces its investment and slows down. 

Consequently, preceding effort suggests that the lead continues to increase. 

However, based on related work for the US and Japanese telecommunications 

industry when duopolistic and monopolistic competition and product system 

complexity for new products are accounted for, the speeding up of a leading firm 

occurs only under rare circumstances. For example, a company getting far enough 

ahead such that the (temporary) monopoly term dominates its payoff expression, 

will always choose the fast strategy, while a company that gets far enough behind 

will always choose the aggressive approach. Under these conditions, the lead is 

likely to continue to increase. If, on the other hand, both monopoly and duopoly 

profits increase substantially with increased aggressiveness then even large leads 

can vanish with significant probability. 

Overall, this characterization highlights two forces that influence a firm’s choices in 

the various stages: proximity to the finish line and distance between the firms. This 

probability of reaping monopoly profits is higher the farther ahead a firm is of its 

rival and even more so the closer the firm is to the finish line. If the lead company is 


322




far from the finish line, even a sizeable lead may not translate into the dominance of 

the monopoly profit term, since there is plenty of time for the lead situation to be 

reversed, and failure to finish first remains a probable outcome. In contrast, the 

probability that the lagging company will get to be a monopolist becomes smaller as 

it falls behind the leader. This raises the following question: what kind of actions 

cause a firm to get ahead? Intuitively, one would expect that a firm that is ahead of 

its rival at any time t, in the sense of having completed more stages by time t, is 

likely to have chosen the faster strategy more often. We will construct numerical 

estimates of the probability that a leading firm is more likely to have chosen a 

strategy faster to verify this intuition. 

Moving away from the firm‐led race patterns revolving in a particular industry to a 

clustering of racing on an industry level is putting industry in different geoeconomics 

zones against each other and becoming dominant in strategic 

product/process technologies. Here racing patterns among industries in a relatively 

free‐trade environment could lead to competitive advantages and more wealth 

creating and accumulating dominance in key product/process technologies in one 

region at the expense of others. There appears to be a link that individual races on 

the firm level induce similar races on the industry level and will be a contributing 

factor to the globalization of network industries. 

Thus similar catch‐up processes are taking place between leaders and followers 

within a group of industrialized countries in pursuit of higher levels of productivity. 

Supposing that the level of labour productivity were governed entirely by the level 

of technology embodied in capital stock, one may consider that the differentials in 

productivities among countries are caused by the ‘technological age’ of the stock 

relative to its ‘chronological age’. The technological age of capital is the age of 

expertise at the time of investment plus years elapsing from that time. Since a 

leading state may be supposed to be furnished with the capital stock embodying, in 

each vintage, technology which was ‘at the very frontier’ at the time of investment, 

the technological age of the stock is, so to speak, the same as its chronological age. 

While a leader is restricted in increasing its productivity by the advance of new 

technology, trailing countries have the potential to make a larger leap as they are 

provided with the privilege of exploiting the backlog in addition of the newly 

developed technology. Hence, followers being behind with a larger gap in 

technology will have a stronger potential for growth in productivity. The potential, 

however, will be reduced as the catch‐up process goes on because the unexploited 

stock of technology becomes smaller and smaller. However, as new technologies 

arise and are rapidly adopted in a Schumpeterian process of ‘creative destruction’, 

their network effects induce rapid accelerating and cumulative growth potentials 

75 

323





which are catalyzed through industry competition. In the absence of such a process 

we can explain the tendency to convergence of productivity levels of follower 

countries. Historically, however, it fails to answer alleged puzzles as to why a 

country, such as the United States, has preserved the standing of the technological 

leader for a long time since taking over leadership from Britain in around the end of 

the nineteenth century and why the shifts have taken place in the ranks of follower 

states in their relative levels of productivity (i.e. technological gaps between them 

and the leader). Abramovitz (1986) poses some extensions and qualifications on this 

simple catch‐up hypothesis in an attempt to explain these facts. Among other factors 

than technological backwardness, he lays stress on a country’s social capability in 

terms of years of education as a proxy of technical competence and its political, 

commercial, industrial, and financial institutions. To become effective social 

capability may also include or expand to ’deep craft’, a ’set of knowings’ on 

technological performance and industrial techniques (Arthur, 2009). The social 

capacity of a state may become stronger or weaker as technological gaps close or 

grow and thus Abramovitz argues that the actual catch‐up process does not provide 

itself to simple formulation. This view has a common understanding to what another 

economist, Olson (1996), expresses to be ‘public policies and institutions’ as his 

explanation of the great differences in per capita income across countries, stating 

that any poorer states that implement relatively good economic policies and 

institutions enjoy rapid catch‐up growth. 

The suggestion should be taken seriously when we wish to understand the 

technological catching‐up to American leadership by Japan, in particular during the 

post‐war period, and explore the possibility of a shift in standing between these two 

countries. This consideration will directly bear on the future trend of the state of the 

art which exerts a crucial influence on the development of the world economy (Juma 

& Clark, 2002; Fagerberg & Godinho, 2004). These explanations notwithstanding, we 

venture as a major factor for divergent growth processes the level of intensity of the 

racing process within the most prevalent value‐added industries with cross‐sectional 

spillovers. These are the communications and information industries which have 

been shaped and led by leading American firms and where the rewards benefited 

their industries and country. Although European and Japanese companies were part 

of the race they were left behind in core markets reaping lesser benefits. (Since ICT 

investment relative to GDP is only less than half in states such as Japan, Germany 

and France compared to the US, 2% vs. more than 4% in 1999, this does not bode 

well for a rapid catch‐up in those countries and a fortiori, for the EU as a whole). 

Steering or guiding the process of racing through the pursuit of industrial policies 

aiming to increase competitive advantage of respective industries, as having been 

practiced in Japan, would stimulate catch‐up races but appears to be less effective in 

promoting frontier racing. Another profound reason lies in the phenomenon of 

network externalities affecting ICT industries. That is, racing ahead of rivals in 


324




respective industries may create external economies to the effect that such economies 

within dominant industries tend to improve their international market position and 

therefore pull ahead in competitiveness vis‐à‐vis their (trading) partners. As 

Krugman (1991) observed: ‘It is probably true that external economies are a more 

important determinant of international trade in high technology sectors than 

elsewhere’. The point is that racing behaviour in leading high‐growth network 

industries by generating frontier positions, create critical cluster and network 

externalities pipelining through other sectors of the economy and create competitive 

advantages elsewhere, as supported by the increasing returns debate (Arthur, 1996). 

In this sense we can speak of positive externalities endogenizing growth of these 

economies and contributing to competitive advantage. All these characteristics lay 

the foundations of the ‘Network Economy’. 

The Network Economy is formed through an ever‐emerging and interacting set of 

increasing returns industries; it is about high‐intensity, technology driven‐racing, 

dynamic entrepreneurship, and focused risk‐taking through (free) venture capital 

markets endogenized by societal and institutional support. With the exception of 

pockets of activity in some parts of Europe (the UK and Scandinavia), and in specific 

areas such as mobile communications, these ingredients for the Network Economy 

are only in the early stage of emerging in Continental Europe, and the political 

mindset in support of the Network Economy is anything but prevalent. As long as 

we do not see a significant shift toward movements in this direction, Europe will not 

see the full benefits of the Network Economy within a Global Economy. 

Racing behaviour on technological positions among firms in high‐ technology 

industries, as exemplified by the globally operating telecommunications and 

computer industries, produce spillover benefits in terms of increasing returns and 

widespread productivity gains. Due to relentless competition among technological 

leaders the network effects result in significant advantages in the value added to this 

industry contributing to faster growth of GDP, and through a flexible labour market, 

also to employment growth. This constitutes a new paradigm in economic thinking 

through network economies and is a major gauge to compare the wealth‐creating 

power of the US economy over the past decade against the European and advanced 

Asian economies. It is interesting to speculate on the implications of the way 

companies in major high‐technology markets, such as telecommunications, split 

clearly into the two major technology races, with one group of firms clearly lagging 

the other. 

The trajectories of technological evolution certainly seem to suggest that firms from 

one frontier cannot simply jump to another trajectory. Witness, in this regard, the 

gradual process necessary for a firm in the catch‐up race to approach those in the 

77 

325





frontier race. There appears to be a frontier ‘lock‐in’, in that once a company is part 

of a race, the group of rivals within that same race are the ones whose actions 

influence that company’s strategy the most. Advancing technological capability is a 

cumulative process. The ability to advance to a given level of technical capability 

appears to be a function of existing technical capability. Given this path dependence, 

the question remains: why do some firms apparently choose a path of technological 

evolution that is less rapid than others? Two sets of possible explanations could be 

derived from our case analysis, which need not be mutually exclusive. The first 

explanation lingers primarily on the expensive nature of R&D in industries like 

telecommunications and computers which rely on novel discovery for their 

advancement. Firms choosing the catch‐up race will gain access to a particular 

technical level later than those choosing the frontier, but will do so at a lower cost. 

TECHNOLOGICAL FRONTIERS 

The evolution of a cross section of high technology industries reflects repetitive 

strategic interactions between companies in a continuous quest to dominate the 

industry or at least to improve its competitive position through company level and 

industry level technological evolution. We can observe several racing patterns across 

industries, each of which is the result of a subset of firms jockeying for a position 

either as a race leader or for a position near the leader constituting a leadership club. 

The identification and interpretation of the races relies on the fact that different firms 

take very different technological paths to target a superior performance level with 

the reward of increasing market shares, maintaining higher productivity and 

profitability. In a Schumpeterian framework such races cannot be interpreted in a 

free‐riding situation where one firm expands resources in advancing the state of 

technology and the others follow closely behind. Such spillover interpretations are 

suspect when products are in the domain of high complexity, of high risk in 

succeeding, and different firms typically adopt different procedural and 

architectural approaches. 

The logic underlying this evolution holds in any industry in which two broad sets of 

conditions are satisfied. First, it pays for a firm to have a technological lead over its 

rival; it also boosts its market image and enhances its reputational capital. Second, 

for various levels of technological complexity among the products introduced by 

various firms, technological complexity can be represented by a multi‐criteria 

performance measure, that is, by a vector‐valued distance measure. The collection of 

performance indicators, parameters, being connected with each other for individual 

companies form an envelope that shapes a ‘technological frontier’. The technological 

frontier is in fact a reasonable indicator of the evolving state of knowledge (technical 

expertise) in the industry. At any point in time the industry technology frontier (ITF) 

indicates the degree of technical sophistication of the most advanced products 

carried by companies in that industry in view of comparable performance standards. 


326




Firm level technology frontiers (FTF) are constructed analogously and indicate, at 

any point in time, the extent of technical sophistication achieved by the firm until 

that point in time. The evolution of company and industry level frontiers is highly 

interactive. Groups of company frontiers are seen to co‐evolve in a manner that 

suggests that the respective firms are racing to catch up with, and get ahead of each 

other. 

A data set could focus on a given set of products (systems) by major European, 

American or Asian enterprises in those industries for a sufficiently representative 

period of market evolution. In principle, we can identify at least two races in 

progress in the industries throughout a given period of duration. One comprises the 

world frontier race in each of those industries, the other, for example, the European 

frontier race which technically would constitute a subfrontier to the worldwide race. 

The aggregate technology frontier of the firms in a particular race (that is, ITF) is 

constructed in a manner similar to the individual FTFs. Essentially, the maximal 

envelope of the FTFs in a particular race constitutes the ITF for that race. The ITF 

indicates, as a function of calendar time, the best achievable performance by any 

firm in the race at a given date. 

A statistical profiling of technological evolution and innovation relates to 

competitive racing among rival companies. Among the (non‐exclusive) performance 

criteria to be assessed are: (1) frequency of frontier pushing; (2) technological 

domination period; (3) innovations vs. imitations in the race; (4) innovation 

frequency when behind or ahead; (5) nature of jumps, leapfrogging or frontiersticking; 

(6) inter‐jump times and jump sizes; (7) race closeness measures; (8) interfrontier 

distance; (9) market leading through ‘market making’ innovations; and (10) 

leadership in ‘innovation markets’. 

A race may or may not have different firms in the leadership position at different 

times. It may be a tighter race at some times than at others, and in general, may 

exhibit a variety of forms of interesting behaviour. While analysis of racing 

behaviour is left to various interpretations, it is appropriate to ask why the firms are 

motivated to keep on racing at all. As access to superior technology expands the 

scope of opportunities available to the firms, the technology can be applied in a 

range of markets. However, leading edge technology is acquired at a cost. It seems 

unlikely that all the companies would find it profitable to compete to be at the 

leading edge all the time. Also not every firm has access to equal capabilities in 

leveraging a given level of technological resources. Firms may, for example, be 

expected to differ in their access to complementary assets that allows them to 

appropriately reap the benefits from their innovation. It is reasonable to assume that 

whatever the level of competence of a company in exploiting its resources it will be 

79 

327





better off the more advanced the technology. Based on this procedure an analysis 

will show how dynamic competition evolved in the past. 

Unlike other (statistical) indicators (such as patent statistics) referring to the degree 

of competitiveness among industries, regions and countries concerned, the proposed 

measures cover behavioral dynamic movements in respective industries, and 

therefore are able to lend intrinsic predictive value to crucial economic variables 

relating to economic growth and wealth creation. The results are likely to provide 

strategic support for industrial and technology policy in a regional or national 

context and will enable policy makers to identify strengths and weaknesses of 

relevant players and their environments in those markets. While this process looks 

like a micro representation of dynamic technological evolution driving companies 

and industries into leadership positions, we may construe an analogous process that 

drives a region or a nation into advancement on a macro scale in order to achieve a 

higher level pecking order among its peers. This may allow using the micro 

foundations of racing as a basis for identifying clubs of nations or regions among 

them to achieve higher levels and rates of growth. 

Catch‐up or Leapfrogging 

It was Schumpeter (1947) who observed that it is the expectation of supernormal 

profits from a temporary monopoly position following an innovation that is the chief 

driver of R & D investment. Along this line, the simplest technology race model can 

be explained as follows: A number of firms invest in R&D. Their investment results 

in an innovation with the time spent in R&D subject to some varying level of 

uncertainty. However, a greater investment reduces the expected time to completion 

of R&D. The model investigates how many firms will choose to enter such a contest, 

and how much they will invest. However, despite some extensive theoretical 

examination of technological races there have been very few empirical studies on 

this subject (Lerner, 1997) and virtually none in the context of major global 

industries, and on a comparative basis. 

Technological frontiers at the firm and industry race levels offer a powerful tool 

through which to view evolving technologies within an industry. By providing a 

benchmarking roadmap that shows where an individual firm is relative to the other 

firms in the industry, they highlight the importance of strategic interactions in the 

firm’s technology decisions. From the interactive process of racing could emerge 

various behavioural patterns. Does lagging behind one’s closest technological rivals 

cause a firm to increase its innovative effort? The term ‘race’ suggests that no single 

company would want to fall too far behind, and that everyone would like to get 

ahead. If a firm tries to innovate more when it is behind than when it is ahead, then 

‘catch‐up’ behaviour will be the dominant effect. Once a firm gets ahead of its rivals 

noticeably, then rivals will step up their efforts to catch up. The leader will slow 

down its innovative efforts until its rivals have drawn uncomfortably close or have 


328




surpassed it. This process repeats itself every time a company gets far enough ahead 

of its rivals. An alternative behaviour pattern would correspond to a firm increasing 

its innovative effort if it gets far enough ahead, thus making catch‐up by the lagging 

firms increasingly difficult. This looks like the ‘Intel Model’ where only the paranoid 

survives (Grove, 1992). For any of these forms there appears to be a clear link to 

market and industry structure, as termed ‘intensity of rivalry’ by Kamien and 

Schwarz (1982). 

We group two different kinds of races: one that is a frontier race among leaders and 

would‐be leaders (first league) and another that is a catch‐up race among laggards 

and imitators (second league). Though both leagues may play their own game, in a 

free market contest, it would be possible that a member of the second league may 

penetrate into the first, as one in the first league may fall back into the second. 

Another aspect of innovation speed has been addressed by Kessler and Bierly (2002). 

As a general rule they found that the speed to racing ahead may be less significant 

the more “radical” (drastic) the innovation appears to be and the more likely it leads 

to a dominant design. These two forms have been applied empirically to the 

development of the early Japanese computer industry (<strong>Gottinger</strong>, 2006a), that is, a 

frontier racing model regarding the struggle for technological leadership in the 

global industry between IBM and ‘Japan Inc.’ guided by MITI (now METI), and a 

catch‐up racing model relating to competition among the leading Japanese 

mainframe manufacturers as laggards. 

It is also interesting to distinguish between two sub‐categories of catch‐up 

behaviour. A lagging firm might simply try to close the gap between itself and the 

technological leader at any point in time (‘frontier‐sticking’ behaviour), or it might 

try to actually usurp the position of the leader by ‘leapfrogging’ it. When there are 

disproportionately large payoffs to being in the technical lead (relative to the payoffs 

that a firm can realize if it is simply close enough to the technical frontier), then one 

would expect that leapfrogging behaviour would occur more frequently than 

frontier‐sticking behaviour (Owen & Ulph, 1994). Alternatively, racing toward the 

frontier creates the ‘reputation’ of being an innovation leader facilitating to maintain 

and increase market share in the future (Albach, 1997). All attempts to leapfrog the 

current technological leader might not be successful since many lagging firms might 

be attempting to leapfrog the leader simultaneously and the leader might be trying 

to get further ahead simultaneously. Correspondingly, one should distinguish 

between attempted leapfrogging and realized leapfrogging. This phenomenon 

(though dependent on industry structure) appears as the predominant behaviour 

pattern in the US and Japan frontier races (Brezis et al, 1991). Albach (1993) cites 

studies for Germany that show otherwise. 

81 

329





Leapfrogging behaviour influenced by the expected size of payoffs as suggested by 

Owen and Ulph (1994) might be revised in compliance with the characteristics of 

industrial structure of the local (regional) markets, the amount of R&D efforts for 

leapfrogging and the extent of globalization of the industry. Even in the case where 

the payoffs of being in the technological lead are expected to be disproportionately 

large, the lagging companies might be satisfied to remain close enough to the leader 

so as to gain or maintain a share in the local market. This could occur when the 

amount of R&D efforts (expenditures) required for leapfrogging would be too large 

for a lagging firm to be viable in the industry and when the local market has not 

been open enough for global competition: the local market might be protected for 

the lagging local companies under the auspices of measures of regulation by the 

government (e.g. government purchasing, controls on foreign capital) and the 

conditions preferable for these firms (e.g. language, marketing practices). 

When the industrial structure is composed of multi‐product companies, as for 

example it used to be in the Japanese computer industry, sub‐frontier firms may 

derive spill over benefits in developing new products in other technologically 

related fields (e.g. communications equipment, consumer electronic products). These 

companies may prefer an R&D strategy just to keep up with the technological 

frontier level (catch‐up) through realizing a greater profit stream over a whole range 

of products. 

What are the implications of the way firms split cleanly into the two technology 

races, with one group clearly lagging the other technologically? The trajectories of 

technological evolution certainly seem to suggest that firms from one frontier cannot 

simply jump to another trajectory. Witness, in this regards the gradual process 

necessary for the companies in the Japanese frontier to catch up with those at the 

global frontier. There appears to be a front line ‘lock‐in’ in that once a firm is part of 

a race, the group of rivals within that same race are the ones whose actions influence 

the firm’s strategy the most. 

Advancing technological capability is a cumulative process. The ability to advance to 

a given level of technical capability appears to be a function of existing technical 

potential. Given this ‘path dependence’, the question remains: why do some firms 

apparently choose a path of technological evolution that is less rapid than others 

are? We propose two sets of possible explanations, which need not to be mutually 

exclusive. The first explanation hinges primarily on the expensive nature of R & D in 

industries like the computer industry, which rely on novel scientific discovery for 

their advancement. Firms choosing the subfrontier will gain access to a particular 

technical level later than those choosing the frontier, but will do so at a lower cost. 

Expending fewer resources on R & D ensures a slower rate of technical evolution. 

The second explanation relates mainly to technological spillovers. Following the 

success of the frontier firms in achieving a certain performance level, these become 


330




known to the subfrontier firms. In fact, leading edge research in the computer 

industry is usually reported in patent applications and scientific journals and is 

widely disseminated throughout the industry. The hypothesis is that partial 

spillover of knowledge occurs to the subfrontier firms, whose task is then simplified 

to some extent. Notice that the subfrontier firms still need to race to be technological 

leaders, as evidenced by the analysis above. This implies that the spillovers are 

nowhere near perfect. Company specific learning is still the norm. However, it is 

possible that knowing something about what research avenues have proved 

successful (for the frontier firms) could greatly ease the task for the firms that follow 

and try to match the technical level of the frontier company. 

Statistical Metrics of Industrial Racing Patterns 

Statistically descriptive measures of racing behaviour can be established that reflect 

the richness of the dynamics of economic growth among competing nations. The 

point of departure for a statistical analysis of industrial racing patterns is the 

aggregate technological frontier represented by the national production function as a 

reasonable indicator of the evolving state of knowledge (technical expertise) in a 

nation or region which is the weighted aggregate of all industries or activities that 

themselves are represented by industry technology frontier (ITF). Firm level 

technology frontiers (FTF) are constructed analogously and indicate, at any point in 

time, the weighted contribution of that firm to the industry on standard industry 

classification. 

In this context we define ‘race’ as a continual contest for technological superiority 

among nations or regions with key industries. Under this conceptualisation a race is 

characterised by a number of countries whose ITF’s remain ‘close’ together over a 

period (T) of, say, 25 to 50 years. The distinctive element is that countries engaging 

in a competition have ITF’s substantially closer together than those of any company 

not in the race. A statistical analysis should reflect that a race, as defined, may or 

may not have different countries in the leadership position at different times. It may 

be a tighter contest at some times than at others, and in general, may exhibit a 

variety of forms of industrial behaviour. We look for clusters of firms who’s ITFs 

remain close enough throughout the duration (formal measures of closeness are 

defined and measured). We identify races to take place at any level of industrial 

performance between the very top and the very bottom throughout 50 years 

duration that is racing from the bottom to racing to the top. 

One comprises the world frontier race in each of those industries, the other a 

subfrontier race (say, North America, Europe, East Asia, China, India, Latin 

America, Africa) which technically would constitute a subfrontier to the world, 

allowing under the best of circumstances for the subfrontier to be the frontier. Since 

83 

331





the level and breadth of industrial activity is reflected as an indicator for economic 

welfare, racing to the top would go parallel with economic growth and welfare 

enhancing, whereas racing from the bottom would correspond to poverty reduction 

and avoiding stationary (under)development traps. 

Characterization of Statistical Indicators of Industrial Racing 

While a variety of situations are possible, the extremes are the following: (a) one 

country may push the frontier at all times, with the others following closely behind, 

(b) some countries share more or less equally in the task of advancing the most value 

generating industry technology frontiers (ITFs). Depending on the situation the most 

value generating industries may be high technology based increasing returns or 

network industries that are able to induce complementary emerging industries with 

high potentials. Extreme situation (a) corresponds to the existence of a unique 

technological leader for a particular race, and a number of quick followers. Situation 

(b), on the other hand, corresponds to the existence of multiple technological leaders. 

Assessment of Frontier Pushing: .The relevant statistics for the races relate to counting 

the times the ITFs are pushed forward by countries or regions at large within a 

global or regional frontier. Frontier pushing can be triggered through industrial 

policy by governments or well fostered entrepreneurship in an advanced capitalistic 

system 

Domination Period Statistics: Accepting the view that a country/region has greater 

potential to earn income and build wealth from its technological position if it is 

ahead of its race suggests that it would be interesting to examine the duration of 

time for which a country can expect to remain ahead once it finds itself pushing its 

ITF. We statistically define the ‘domination period’ to be the duration of time for 

which a country leads its particular race. It is interesting to note that the mean 

domination period is virtually indistinguishable for the three races, and lies between 

three and four years. A difference of means test cannot reject the hypothesis that the 

mean years of domination tend to cluster but hardly converge. So countries in each 

of the races can expect to remain ahead approximately in proportion to their 

technological capability and more than the amount of time after they have propelled 

themselves to the front of their respective races. However, the domination period 

tends to be a more uncertain quantity in the world frontier race, to a lesser degree in 

the EU frontier race than in any smaller regional races (as evidenced by the lower 

domination period standard deviation). 

Catch‐up Statistics: If key industries of a country push to innovate more when they 

are behind than when they are ahead, then ‘catch‐ up’ behaviour will be the 

dominant effect. For each country/region, these statistics compare the fraction of the 

total innovations carried out by industries in that country (i.e. the fraction of the total 

number of times that their ITFs advance) when it was engaging in its race when 

lagging, with the fraction of times that the country actually led its race. In the 


332




absence of catch‐up behaviour, or behaviour leading to a country increasingly 

dominating its rivals, we would expect to see no difference in these fractions. Then 

the fraction of time that a country is ahead of its race could be an unbiased estimator 

of the fraction of innovations in its key industries that it engages in when it is ahead. 

Relevant data, however, suggest that this is usually not the case. They appear to 

show that the fraction of times a state leads its race at any development level in a 

group or club is larger than the fraction of innovations that occur when the country 

is ahead, i.e. more innovations occur when the country is lagging than would be 

expected in the absence of catch‐up or increasing dominance behaviour. A major 

exception would arise if the country would act like an ‘Intel Economy’, where 

unchallenged leadership in key industries creates incentives to increase the lead to 

its rivals. Catch‐up behaviour is supported by additional observations, as derivable 

from convergence and conditional convergence in the economic growth process that 

countries make larger jumps (i.e. the ITFs advance more) when they are behind than 

when they are leading the race 

Leapfrogging Statistics: From this, the distinction emerges between two kinds of catchup. 

A lagging country might simply try to close the gap between itself and the 

technological leader at any point in time (frontier‐sticking behaviour), or it might try 

to actually usurp the position of the leader by ‘leapfrogging’ it. When there are 

disproportional larger incomes per head when being in the technical lead (relative to 

a situation that a country can realize if it is simply close enough to the technological 

frontier), then one would expect that leapfrogging behaviour would make it a more 

attractive incentive than frontier‐sticking behaviour. 

All attempts to leapfrog the current technological leader might not be successful 

since many lagging firms/industries might be attempting to leapfrog the leader 

simultaneously. Correspondingly, we observe both the attempted leapfroggings and 

the realized leapfroggings. It appears likely that the leapfrogging phenomenon 

would be more predominant in the premier league than in following up leagues. 

Interfrontier Distance: How long does ‘knowledge’ take to spillover from frontier to 

subfrontier industries? This requires investigating “interfrontier distance”. One 

measure of how much subfrontier industries’ technology lags the frontier industries’ 

technology could be graphed as “subfrontier lag” in terms of calendar time. At each 

point in time, this is simply the absolute difference in the subfrontier performance 

and the frontier performance time. The graph would clearly indicate that this 

measure has been declining or increasing more or less monotonically over the past 

50 years to the extent that the subfrontier industries have been able/unable to catch 

up with the frontier industries. A complementary measure would be to assess the 

difficulty of bridging the lag. That is, how much longer does it take the subfrontier to 

85 

333





reach a certain level of technical achievement after the frontier has reached that 

level? Thus it might very well turn out that the interfrontier distance may be 

decreasing though the difficulty in bridging the gap is increasing. 

Race Closeness Measure (RCM): None of the previous analyses tell us how close any of 

the overall races are over a period of time. The races are all distant/close by 

construction, however, some might be closer than others, We define ‘a measure of 

closeness’ of a race (RCM) at a particular time as follows: 

RCM (t) = 0 N Fi (t) – Fj (t) 2 /N (t) (1) 

where Fi (t) is country’s i ITF at time t, Fj(t) is country’s j comparable ITF at time 

t = max ITF(t) for each i, j and N(t) is the number of active key value‐generating 

industries at time t. 

The measure (Equation 1) thus constructed has a lowest value of 0, which 

corresponds to a ‘dead heat’ race. Higher values of the measure correspond to races 

that are less close. Unlike the earlier characteristics (domination period length, 

innovation when ahead versus when behind, leapfrogging versus frontier‐sticking) 

which investigate the behaviour of a particular feature of the race and of a particular 

industry in relation to the race frontier, the RCM is more of an aggregate statistic of 

how close the various racing parties are at a point in time. The closeness measure is 

simply an indication of parity, and not one that says anything per se about the 

evolution of the technological frontier. To see this, note that if none of the frontiers 

were evolving, the closeness measure would be 0, as it would be if all the frontiers 

were advancing in perfect lock‐step with one another. 

TABLE 1. ITF SHIFTS ACROSS AGGREGATED INDUSTRIES 

Aggregate Industries 1980‐ ITF (max = 100) 2010 GDP (%) 

US 80 85 70 

EU 60 75 60 

China 15 60 50 

USSR 30 35 40 

India 25 40 30 

Brazil 20 30 25 

Japan 70 70 65 

We talk about value‐added increasing returns industries over a period of 30 years. 

The industries comprise ICT, Consumer Electronics, Chemicals and Materials, 

Automobiles, Pharma/Biotech, Machine Tools, Medical Instruments, 

Aerospace/Defense, Energy Technologies, and HT Transportation Systems. Industry 

sectors can be assigned to various countries/regions such as US, EU, China, Russia, 

India, Brazil, Japan (Table 1). We benchmark the industry technology frontiers (ITFs) 

accordingly, that is, highest ‘state of knowledge’ at time t is 100 pc. The countries’ 

rank to the max ITFs is assessed as the share of the max ITF. The assessment 

intervals are spaced in five year intervals starting in 1980 until 2010. 


334




Efficiency 

Let’s explore the inefficiency of the follower nations; i.e., the negative effect on the 

potential technology gap stemming from inefficient social and institutional factors. A 

good example of cross industrial inefficiencies over a historically representative 

period (1810‐2000) is Russia that was hardly advancing economically against 

underdeveloped benchmark countries and falling behind leading economies, 

reinforced through the bolshevik revolution and its underperforming economic 

mechanism design (Gaidar, 2012). Increasing efficiencies deblock catch‐up in lagging 

countries (Juma & Clark, 2002). Efficiency is found by dividing a nation’s estimated 

fixed effect by the regional adoption rate. As defined here, it is quite robust to 

different estimations and samples. The relative efficiencies of the nations within 

regions appear to conform to common beliefs. For example, in Europe, the 

Netherlands, Belgium and Switzerland are the most efficient while Turkey, Portugal 

and Greece are the least efficient. In East Asia, Hong Kong is the most efficient while 

Indonesia and Thailand are the least efficient. Finally, in Latin America, Mexico and 

Argentina are at the top and Honduras and Bolivia at the bottom. Another way to 

discuss the findings is to consider the time required to catch‐up. Previously, Parente 

and Prescott (2004) showed that countries with lower levels of relative efficiency will 

adopt modern technologies at much later dates. Conversely, one could argue that if 

those countries adopt modern technologies concurrently with their low level of 

relative efficiency then their rates of growth would stay at a subpar level of their 

potential. 

One major source of efficiency generation for a country, according to Parente and 

Prescott (2004), is belonging to a ‘free trade club’ that improves efficiency through 

greater industrial competition. We calculate the required time period until the 

nations reach their frontier when only the catch‐up term and inefficiency are allowed 

to vary across regions and countries. Two frontiers are considered: nations’ 

inefficiency frontier and the leader nation’s frontier. The latter requires that the 

inefficiency levels fade away in time which we assume occurs at the rate of . The 

European countries, with the exception of Turkey, all seem to have reached their 

inefficiency reduced frontier. The same is true for most of the East Asian countries. 

Thus, these nations will not catch‐up with the US without higher accumulation rates 

or improved efficiency. For Latin America, most countries are still catching‐up with 

their inefficiency frontier, so that if accumulation rates were the same catch‐up 

would still take place through diffusion of disembodied technology. Of course, if 

inefficiency levels remain then a follower could never completely catch‐up with the 

leader by taking advantage of the technology gap alone. As an illustrative example, 

for the required time to catch‐up with the leader if inefficiency levels were 

improving at the rate much of Europe and Latin America could then approach the 

87 

335





frontier faster than East Asia on account of East Asia’s lower rate of technology 

adoption. This begs the question of what determines these. (in)efficiencies? 

It is reasonable to expect a tradeoff between a general technology level (GTL) of a 

nation’s leading industries and its institutional efficiencies (IE). Thus, using an 

aggregate score, (GTL, IE) , say, a country may be in the top rank of GTL but weak 

on IE which may be surpassed in growth by one which is lower in GTL rank but 

strong on IE. 

CONCLUSION AND FURTHER DISCUSSION 

Economic growth over the long‐run can only be achieved in the course of a real, 

sustainable value‐creating process through industrial performance and open 

markets in which technology and innovation are the key facilitators. Nations with 

their industries engage in rival contests in what we term industrial races within a 

given international trade regime. This reflects a micro‐economic based behavioral 

focus on economic growth (positive or negative). It builds a deeper foundation to 

explanations of economic growth than conventional macro‐economic texts . It also 

uncovers the true sources of growth as a tool for growth diagnostics (Rodrik, 2007) 

allowing to embrace other observations on urban growth and non primarily 

economic factors. In an influential paper in Foreign Affairs entitled ‘Can India 

overtake China’ Huang and Khanna (2003) first looked at macro‐economic factors, 

which favor China. They then considered micro‐economic structures and behaviors 

such as competent indigenous entrepreneurship, a sound capital market, an 

independent legal system, property rights and a grass roots approach to 

development. The latter all favor India in the long run, say over the next fifty years. 

In a widely covered empirical investigation on global growth patterns we concur 

with Easterly and Levins’ (2002) finding that it is not factor accumulation, per se, but 

total factor productivity that explains cross‐country differences in the level of GDP 

growth rates. This total productivity in turn is derived from technology (innovation) 

transfer and diffusion, its’ supporting institutional characteristics and cultural 

dependence. Of course, on a deeper level, considerations of merely formal 

institutions may not suffice for explanations but instead forms of economic 

mechanism design may be called for that effectively deal with (enforce rules on) 

‘moral hazard’ and ‘adverse selection’ issues (Myerson, 2006). Economic growth in a 

decentralized system would be fully supported by a a Hayek‐Hurwicz mechanism 

design. 

Observations on firm‐led racing patterns emerging in oligopolistic market structures 

of particular high tech industries, and the clustering of racing on an industry level 

are putting companies in different geo‐economic zones against each other, becoming 

dominant in strategic product/process technologies. Here racing patterns among 

industries in a relatively free trade environment could lead to competitive 


336




advantages and more wealth creating and accumulating dominance in key product / 

process technologies in one region at the expense of others. The question is whether 

individual contests on a firm level induce similar effects on an industry level and if 

so, what controlling effects may be rendered by regional or multilateral policies on 

regulatory, trade and investment matters? The point is that racing behaviour in 

leading high technology industries by generating frontier positions create cluster 

and network externalities pipelining through other sectors of the economy and 

creating competitive advantages elsewhere, as supported by the ‘increasing returns’ 

debate. In this sense we can speak of positive externalities endogenizing growth of 

these economies and contributing to competitive advantage. 

We are about to show in the upcoming chapters how technological racing, rivalry 

and competition instigates a process of innovation, industrial and market evolution 

and how it extends to larger entities than firms and industries to regions and 

national economies or economy networks. It will show what drives economic 

growth and globalization, which industries are most significantly affected and how 

technological racing results in value generation in increasing returns and network 

industries. Furthermore, we consider how the emergence of selective managerial 

strategies is most likely to carry success in the pursuit of corporate and industrial 

policies. 

Welfare enhancing technology racing as a constituent element of the capitalist 

process reinforced by globalization provides social benefits far exceeding the costs. 

Even more important, any alternative path, other than the competitive, would likely 

be inferior given the costs in that it would generate a less valued and less welfare 

producing technology portfolio. That is, even if the competitive process is wasteful, 

(for example, in parallel or correlated technology development) its unique high 

value innovation outcome far exceeds the benefits of any alternative path. There is 

historical, observational and analytical evidence given in <strong>Gottinger</strong> and Goosen 

(2012). 

On a national scale simple catch‐up hypotheses have put emphasis on the great 

potential of adopting unexploited technology in the early stage and the increase of 

self‐limiting power in the later stage. However, an actual growth path of 

technological trajectory of a specific economy may overwhelmingly be constrained 

by social capability. The capability also endogenously changes as states of the 

economy and technology evolve. The success of economic growth due to diffusion of 

advanced technology or the possibility of leapfrogging is mainly attributable to how 

the social capability evolves (i.e., which effects become more influential: growing 

responsiveness to competition or growing obstacles to it on account of vested 

interests and established positions). Another observation relates to policy inferences 

89 

337





on market structure, entrepreneurship, innovation activity, industrial policy and 

regulatory frameworks in promoting and hindering industry frontier races in a 

global industrial context. Does lagging behind one’s closest technological rivals 

cause an industry to increase its innovative effort? 

On an industry level, among the key issues to be addressed is the apparent inability 

of technology oriented corporations to maintain leadership in fields that they 

pioneered. There is a presumption that firms fail to remain competitive because of 

agency problems or other suboptimal managerial behaviour within these 

organizations. An alternative explanation is that technologically trailing firms, in 

symmetric competitive situations, will devote greater effort to innovation, so that a 

failure of technological leaders to maintain their position is an appropriate response 

to the competitive environment. In asymmetric situations, with entrants challenging 

incumbents, research does demonstrate that start‐up firms show a stronger 

endeavour to close up to or leapfrog the competitors. Such issues highlight the 

dynamics of the race within the given market structure in any of the areas 

concerned. 

Catch‐up processes are taking place between leaders and followers within a group of 

industrialized countries in pursuit of higher levels of productivity and economic 

growth. Supposing that the level of labour productivity were governed entirely by 

the level of technology embodied in capital stock, one may consider that the 

differentials in productivities among countries are caused by the ‘technological age’ 

of the stock used by a country relative to its ‘chronological age’. The technological 

age of capital is a period of technology at the time of investment plus years elapsing 

from that time. Since a leading country may be supposed to be furnished with the 

capital stock embodying, in each vintage, technology which was ‘at the very frontier’ 

at the time of investment, ‘the technological age of the stock is, so to speak, the same 

as its chronological age’. While a leader is restricted in increasing its productivity by 

the advance of new technology, trailing countries ‘have the potential to make a 

larger leap’ as they are provided with the privilege of exploiting the backlog in 

addition of the newly developed technology. Hence, followers being behind with a 

larger gap in technology will have a stronger potential for growth in productivity. 

The potential, however, will be reduced as the catch‐up process goes on because the 

unexploited stock of technology becomes smaller and smaller. However, as new 

technologies arise and are rapidly adopted in a Schumpeterian process of ‘creative 

destruction’, their network effects induce rapid accelerating and cumulative growth 

potentials being catalyzed through industry racing. 

In the absence of such a process, we can explain the tendency to convergence of 

productivity levels of follower countries. Historically, it fails to answer alleged 

puzzles of why a country, such as the United States, has preserved the standing of 

the technological leader for a long time since taking over leadership from Britain in 


338




around the end of the nineteenth century and why the shifts have taken place in the 

ranks of follower countries in their relative levels of productivity (i.e. technological 

gaps between them and the leader). Abramovitz (1986) poses some extensions and 

qualifications on this simple catch‐up hypothesis in the attempt to explain these 

facts. Among other factors than technological backwardness, he lays stress on a 

country’s ‘social capability’ (i.e. years of education as a proxy of technical 

competence and its political, commercial, industrial, and financial institutions). The 

social capability of a country may become stronger or weaker as technological gaps 

close and thus, he states, the actual catch‐up process does not lend itself to simple 

formulation. This view has a common understanding to what another economist, 

Olson (1996), expresses to be public policies and institutions as his explanation of the 

great differences in per capita income across countries, stating that any poorer 

countries that adopt relatively good economic policies and institutions enjoy rapid 

catch‐up growth. The suggestion should be taken seriously when we wish to 

understand the technological catching‐up to American leadership by Japan, in 

particular, during the post‐war period and explore the possibility of a shift in 

standing between these two countries. This consideration will directly bear on the 

future trend of the state of the art, which exerts a crucial influence on the 

development of the world economy. 

These explanations notwithstanding, we venture as a major factor for divergent 

growth processes the level of intensity of the racing process within the most 

prevalent value‐added industries with cross sectional spillovers. These are the 

communications and information industries, which have been shaped and led by 

leading American firms and where the rewards benefited their industries and 

country. Though European and Japanese companies were part of the race they were 

left behind in core markets reaping lesser benefits. The IT investment relative to 

GDP, for example, used to be only less than half in countries such as Japan, 

Germany and France compared to the US. This does not bode well for a rapid catchup 

in those countries. Steering or guiding the process of racing through the pursuit 

of industrial policies aiming to increase competitive advantage of respective 

industries, as having been practised in Japan, would stimulate catch‐up races but 

appears to be less effective in promoting frontier racing. Another profound reason 

lies in the phenomenon of network externalities affecting IT industries. That is, 

racing ahead of rivals in respective industries may create external economies to the 

effect that such economies within dominant industries tend to improve their 

international market position and therefore pull ahead in competitiveness vis‐a‐vis 

their (trading) partners. 

The point is that racing behaviour in leading high growth network industries by 

generating frontier positions create critical cluster and network externalities 

91 

339





pipelining through other sectors of the economy and creating competitive 

advantages elsewhere, as supported by the increasing returns debate (Arthur, 1996). 

In this sense we can speak of positive externalities endogenizing growth of these 

economies and contributing to competitive advantage. 

All these characteristics lay the foundations of the ‘Network Economy’. The latter is 

formed through an ever emerging and interacting set of increasing returns 

industries, it is about high‐intensity, technology driven racing, dynamic 

entrepreneurship, focussed risk‐taking through (free) venture capital markets 

endogenized by societal and institutional support. 

Racing behaviour on technological positions among firms in high technology 

industries, as exemplified by the globally operating telecommunications, and 

computer industries, produce spillover benefits in terms of increasing returns and 

widespread productivity gains. Due to relentless competition among technological 

leaders the network effects lead to significant advantages in the value added to this 

industry, contributing to faster growth of GDP, and through a flexible labour 

market, also to employment growth. This constitutes a new paradigm in economic 

thinking through network economies and is a major gauge to compare the wealth 

creating power of the US economy against the European and advanced Asian 

economies. 

REFERENCES 

Abramovitz, M. (1986). Catching up, forging ahead, and falling behind, Journal of 

Economic History, 66, 385‐406. 

Abramovitz, M. & David, P. (1995). Convergence and Deferred Catchup’, Chapter 1 

in Growth and Development: The Economics of the 21 st Century, Ralph Landau, 

Timothy Taylor, and Gavin Wright, eds., Stanford CA: Stanford University Press. 

Albach, H. (1994). Information, Zeit und Wettbewerb (Information, Time and 

Competition), J. H. von Thuenen Lecture, in Jahrestagung des Vereins fuer 

Socialpolitik, Muenster 1993, Berlin: Duncker und Humblot, 113‐154. 

Albach, H. (1997). Global competition among the few, The Ehrenrooth Lectures, 

Swedish School of Economics and Business Administration, Res. Reports 40, 

Helsingfors. 

Amsden, A. H. (1989). Asia’s Next Giant, New York: Oxford University Press. 

Arthur,W. B. (1996). Increasing returns and the new world of business, Harvard 

Business Review, July–August, 100‐109. 

Arthur, W.B. (2009). The Nature of Technology, New York: Free Press. 


340




Brezis, E., Krugman, P. and Tsiddon, D. (1991). Leapfrogging: a theory of cycles in 

national technological leadership, National Bureau of Economic Research (NBER), 

Working Paper # 3886. 

Burkett, P. & Hart‐Landsberg, M. (2003). A Critique of “Catch‐Up“ Theories of 

Development, Journal of Contemporary Asia 33(2), 147‐171. 

Conitzer, V. & Sandholm, T. (2002). Complexity of Mechanism Design, Computer 

Science Dept., Carnegie Mellon Univ., Pittburg,Pa., Disc.Paper, May 28, 1‐8. 

Easterly, W. & Levine, R. (2002). It’s not Factor Accumulation: Stylized Facts and 

Growth Models, Central Bank of Chile , Working Papers No. 164. 

Easterly, W. (2013). The Tyranny of Experts, Economists, Dictators, and the 

Forgotten Rights of the Poor, New York: Basic Books. 

Fagerberg, J. & Godinho, M. M. (2004). Innovation and Catching‐Up’, Chap. 20 on 

Handbook of Innovation, Oxford: Blackwell. 

Feigenbaum, J., Schapiro, M. & Shenker, S. (2007). Distributed Algorithmic 

Mechanism Design, In: Nisan et al, Algorithmic Game Theory, Cambridge: 

Cambridge Univ. Press, 363‐384. 

Gaidar, Y. (2004). Modern Economic Growth in Russia and Catch‐Up Development, 

Herald of Europe, 1‐11. 

Gaidar, Y. (2012), Russia, A Long View. Cambridge, MA: MIT Press. 

Gerschenkron, A. (1962). Economic Backwardness in Historical Perspective, 

Cambridge, MA: Belknap Press. 

<strong>Gottinger</strong>, H. W. (2006a). Global Technological Races, Japan and the World Economy 

18, 181‐193. 

<strong>Gottinger</strong>, H. W. (2006b). Innovation, Technology and Hypercompetition, London: 

Routledge. 

<strong>Gottinger</strong>, H. W. & Goosen, M. (2012), Strategies of Economic Growth and Catch‐Up, 

New York: Nova Science. 

<strong>Gottinger</strong>, H. W. (2013). Quality of Service on Queueing Networks for the Internet, 

iBusiness 5, 1‐12. 

Grossman, G. & Shapiro, C. (1987). Dynamic R&D Competition, Economic Journal, 

97, 372‐387. 

Grove, A. S. (1999). Only the Paranoid Survive, New York: Doubleday. 

Harris, C. & Vickers, J. (1987). Racing with uncertainty, Review of Economic Studies, 

54, 305‐321. 

93 

341





Hayek, F. A. (1945). The Use of Knowledge in Society, American Economic Review 

35, 519‐530. 

Huang, Y. & Khanna, T. (2003). Can India overtake China?, Foreign Policy July‐Aug. 

Hurwicz & Reiter, S. (2006). Designing Economic Mechanisms, Cambridge Univ. 

Press, Cambridge 

Juma, C. & Clark, N. (2002). Technology Catch‐Up: Opportunities and Challenges for 

Developing Countries, Supra Occ. Papers, Research Centre for the Social Sciences, 

Univ. of Edinburgh, UK. 

Kessler, E. H. & Bierly, P. E. (2002). Is faster really better? An empirical test of the 

implications of innovation speed, IEEE Transactions on Engineering Management, 

49, 2‐12. 

Kamien, M. I. & Schwarz, N. L. (1982). Market Structure and Innovation, Cambridge: 


Krugman, P. (1991). Increasing Returns and Economic Geography, Journal of 

Political Economy, 99(3), 483‐499. 

Krugman, P. (1994). The Myth of Asia’s Miracle, Foreign Affairs 73 ,63‐78. 

Lee, K. (2013). Schumpeterian Analysis of 


Economic Catchup. Cambridge: 

Lerner, J. (1997). An empirical exploration of a technology race, Rand Journal of 

Economics, 28(2), 228‐224. 

Lin, J. Y. (2013). Against the Consensus, Cambridge: Cambridge University Press. 

Myerson, R. B. (2008). Perspectives of Mechanism Design in Economic Theory, 

American Economic Review, 98(3), 586‐603. 

Nisan, N. & Ronen, A. (2001). Algorithmic Mechanism Design, Games and Economic 

Behavior 35, 166‐196. 

Nisan, N., Roughgarden, T., Tardos, E. & Vazirani, V. V. eds. (2007). Algorithmic 

Game Theory, Cambridge: Cambridge University Press. 

Olsen, M. (1996). Big Bills left on the Sidewalk: Why some nations are rich and others 

poor, Journal of Economic Perspectives 10, 3‐24. 

Owen, R. & Ulph, D. (1994). Racing in two dimensions, Journal of Evolutionary 

Economics, 4, 185‐206. 

Rodrik,D. (2007). One Economics, Many Recipes: Globalization, Institutions and 

Economic Growth, Princeton: Princeton University Press. 


342




Sachs, J. D. & Warner, A. M. (1996). Achieving Rapid Growth in the Transition 

Economies of Central Europe, Harvard Institute for International Development, 

Discussion Paper 544, July. 

Scherer, F. M. (1992). International High‐Technology Competition, Cambridge: 

Harvard University Press. 

Scherer, F. M. (1999). New Perspectives on Economic Growth and Technological 

Innovation, Washington D. C.: Brookings Institution Press. 

Schumpeter, J. (1947). Capitalism, Socialism and Democracy, New York: Harper, 

Rev. Edition. 

Wan, H. Y. (2004). Economic Development in a Globalized Environment, East Asian 

Evidence, Dordrecht, Kluwer. 

95 

343



344

3.4 ESCAPING FROM DEVELOPMENT TRAPS 



ESCAPING FROM DEVELOPMENT 

TRAPS: INDUSTRIALIZATION AND 

RACING FROM THE BOTTOM 

<strong>Hans</strong> W. <strong>Gottinger</strong> 1* , Celia Umali 2 

1 


2 

Faculty of Economics, Nagasaki University, Nagasaki, Japan 

*umari@nagasaki‐u.ac.jp 

Abstract 

For development economies, escaping costly development traps an industrialization policy is likely to 

be gradual rather than take a “big push” form and becoming more balanced over time. Under certain 

conditions the optimal industrialization policy should be more unbalanced the weaker are the 

sectorial linkages, the stronger are increasing returns, entrepreneurial resources, and the smaller are 

the domestic market size and the lesser the degree of dynamic competition. We show how to make 

tradeoffs at different levels of development and from the perspective of the industrialization debate in 

a historical context of modern development policies. 

Key words 

Development traps; Tradeoffs; Industrialization. 

INTRODUCTION 

The cumulative literature on industrialization has formalized the long‐standing idea 

that development traps are the result of a failure of economic organization rather 

than a lack of resources or other technological constraints. The so‐called “big push” 

models of industrialization have shown how, in the presence of increasing returns, 

there can exist preferable states to advance the economic states of countries in 

contest with other countries. Such a view not only provides an explanation for the 

co‐existence of industrialized and less industrialized economies, but also a rationale 

for government intervention to coordinate investment in a “big‐push” toward 

industrialization. Moreover, unlike competing theories, these models emphasize the 

temporary nature of any policy. Thus, industrialization policy involves facilitating an 

adjustment from one equilibrium to another rather than any change in the nature of 

the set of equilibria per se. 

5 

345



<strong>Hans</strong> W. <strong>Gottinger</strong>, Celia Umali 

Escaping from Development Traps: Industrialization and Racing from the Bottom 

While recent formalization makes clear the possible role for the government in 

coordinating economic activity, little has been said about the form such policy 

should take. Is there a conceptual model to analyze the question: what precise form 

should the “big push” take? It should be part of mechanism design for economic 

development (<strong>Gottinger</strong>, 2014). It is argued that while many different 

industrialization policies can be successful in generating escapes from development 

traps, the form of the policy that minimizes the costs of this transition depends on 

the characteristics of the economic situation at hand. Factors such as the strength of 

the complementarities, externalities and increasing returns, among others, all play a 

role in influencing the nature of a “getting‐ahead” industrialization policy. Such 

ideas were already present in the debates in development economics in the 1940s 

and 1950s regarding the form of industrialization policy. The models underlying 

these less formal debates inspired the recent more formal research but the policy 

elements of these have not been addressed, to date, in any substantive way. 

The paper proceeds as follows. We first give a brief history to recall different 

development strategies proliferating in the literature, in Section 2. Then we show 

how the increasing returns debate on industries impact structural change and 

development paths, in Section 3. Section 4 gives the industrialization policydevelopment 

context in an optimization framework. Conclusions follow in Section 5. 

A BRIEF HISTORY OF MAJOR DEVELOPMENT STRATEGIES 

Principal among the earlier policy debates was that surrounding the efficacy and 

costs involved in the alternative strategies of “balanced” versus “unbalanced 

growth.” Rosenstein‐Rodan (1943, 1961) and Nurkse (1952, 1953) provided the 

rationale for the notion that the adoption of modern technologies must proceed 

across a wide range of industries more or less simultaneously. It was argued that the 

neglect of investment in a sector(s) could undermine any industrialization strategy. 

Reacting to this policy prescription was the “unbalanced growth” school led by 

Hirschman (1958) and Streeten (1956). They saw the balanced strategy as far too 

costly. The advantages of multiple development may make interesting reading for 

economists, but they are gloomy news indeed for underdeveloped countries. The 

initial resources for simultaneous developments on many fronts are generally 

lacking. By targeting many sectors, it was argued that scarce resources would be 

spread too thin‐ so thin, that industrialization would be thwarted. It seemed more 

fruitful to target a small number of “leading sectors” (Rostow, 1960). Then those 

investments would “….call forth complementary investments in the next period with 

a will and logic of their own: they block out a part of the road that lies ahead and 

virtually compel certain additional investment decisions” (Hirschman, 1958: 42). 

Thus, the existence of complementarity between investments (in particular those 

involving human capital) and increasing returns motivated an unbalanced approach 

(Easterly, 2002). Curiously, at the same time, “complementarity of industries 

6 JOURNAL OF APPLIED ECONOMICS AND BUSINESS, VOL. 3, ISSUE 1 ‐ MARCH , 2015, PP. 5‐13 

346




provides the most important set of arguments in favor of a large‐scale planned 

industrialization”(Rosenstein‐Rodan, 1943: 205). Further, one of the first to preview 

the connection between Big Push, Poverty Traps and Takeoffs was the essay by W. 

Easterly (2005) who integrated historical sources with present day modern 

development strategies. Both sides appeared to have agreed that a “big push” was 

warranted, but they disagreed as to its composition. Our purpose here is to use the 

guidelines provided by the more recent formalization of the “big push” theory of 

industrialization to clarify the earlier debate of the appropriate degree of focus for 

industrialization policy. After all, the more recent literature has stressed the roles of 

complementarities and increasing returns that both schools saw lying at the heart of 

their policy prescriptions. 

The seminal article formalizing the “big push” theory of industrialization is that of 

Murphy et al, (1989). In their model, firms choose between constant returns and an 

increasing returns technology based on their expectations of demand. However, 

these choices spill over into aggregate demand creating a strategic interaction among 

sectors in their technology adoption decisions. Thus, under certain conditions, there 

exist two equilibria: with all firms choosing the constant returns or all choosing the 

increasing returns technology. Clearly, in the latter equilibrium, all households are 

better off. 

While the Murphy et al, (1989) model shows how increasing returns (and a wage 

effect) aggregate to strategic complementarity among sectors, it does not lend itself 

readily to the debate concerning the degree of balance in industrialization policy. 

First, the static content leaves open the question of whether the intervention should 

take the form of anything more than indicative planning. Second, the most 

commonly discussed policy instrument in the industrialization debate is the 

subsidization of investments. However, in the Murphy et al, (1989) example, use of 

this instrument biases one toward a more unbalanced policy. To see this, observe that 

it is the role of the government to facilitate a move to the industrializing equilibrium. 

This means that the government must subsidize a sufficient amount of investment to 

make it profitable for all sectors to adopt the modern technology. 

Given the binary choice set, there then exists some minimum critical mass of sectors 

that must be targeted to achieve a successful transition. A greater range of successful 

industrialization policies might be more plausible, however, if firms had the choice 

of a wider variety of technology to choose from (<strong>Gottinger</strong>, 2006; <strong>Gottinger</strong> & 

Goosen, 2012). One might suppose that targeting a large number of sectors to 

modernize a little and targeting a small number of sectors for more radical 

modernization might both generate a big push. Thus, to consider the balanced 

approach properly, a greater technological choice space is required. 

7 

347





INCREASING RETURNS, STRUCTURAL CHANGE AND 

DEVELOPMENT PATHS 

What would be the choice variables available to the government provided it would 

be able to pick up what is likely to be increasing returns industries in the future? 

First, in each period, the government can choose the set of firms that it targets for 

structural change. Second, for each targeted firm, the government can choose a 

target level for ‘increasing returns industry’ modernization in the period. Along this 

vein, the government could choose to target the same number of firms in each period 

but induce those firms to modernize gradually over time. Or in contrast, the 

government chooses a single level of modernization to occur across all firms and all 

periods. It then targets a mass of firms each period for entry and modernization. This 

means that industrialization policy is solely characterized by the critical mass of 

sectors targeted, and the target level of modernization. The level of modernization 

could be sequentially expanded by infrastructural upgrading across the board to 

benefit all major sectors as suggested by the Chinese economist Justin Yifu Lin 

(2013). 

Given a parameterized development path, the most significant parameter represents 

the strength of increasing returns in the technology adopted by industrial sectors, 

which generates a rationale for “big push” intervention. A “big push” can be 

activated if the economy is stuck in a “development trap” from which an escape 

could be made through sufficient coordination of decisions by input producers. For a 

developing economy in its early phase a “poverty trap” is a special case of a 

“development trap” defined by Barro and Sala‐i‐Martin (1995: 49) as a stable steadystate 

with low levels of per capita output and capital stock. This is a trap because, if 

agents attempt to break out of it, the economy has a tendency to return to the lowlevel 

steady‐state. Only by a very large change in their behavior, can the economy 

break out of the poverty trap and move to the high‐income steady state. To evaluate 

the economic characteristics, i.e., the strengths of complementarities and increasing 

returns, would affect the government’s policy choices and industrial policies (Gans, 

1994). 

Big Push theories of industrialization could lead to ‘development traps’ if 

sequential industrialization would add more diminishing returns than increasing 

returns industries which could be a result of government’s coordination failure. 

This would point to deficiencies in institutional quality as outlined by North (1990) 

impacting economic performance. They could give explanations for decade long 

lackluster performance of Latin American economies (Fukuyama, 2008). When a 

development trap is purely the result of coordination failure, to escape from the trap, 

would technically require the government to synchronize the expectations of 

individual agents (entrepreneurs) with targeting investment in industrialization 

activities. If a government were to announce that firms should modernize to a 


348




certain degree, even if this were believed perfectly by individuals and firms, each 

firm might still have an incentive to wait before investing. In that case, the optimistic 

expectations by the government would not be realized and the policy would be 

ineffective. Irreversibility and the time lag of production mean that history rather 

than expectations matter (Krugman, 1991). The previous level of industrialization 

determines what path the economy will take in the future. This is why it is difficult 

to characterize the industrializing paths of the economy. There is econometric 

evidence that a contributing factor toward the emergence of development traps is the 

lack of surpassing some threshold of technological integration in the industrializing 

(manufacturing) sector (Ortiz et al, 2009). 

INDUSTRIALIZATION POLICIES AND DEVELOPMENT 

In the context of a big push development strategy the government faces a tradeoff 

between the number of sectors it targets and the degree to which it wishes them to 

modernize, that is, it chooses the critical mass of sectors that must be targeted at any 

point in time in order to generate an escape from a development trap and to achieve 

increasing returns. Let’s take a simple case where the industrialization policy takes 

the form of a “big bang”, that is intervention occurs for one period only granting that 

the resources exist in that period to allow for such a policy. This means that the 

industrialization policy is solely characterized by the critical mass of sectors targeted 

s * and the target level of modernization ƒ. 

Suppose naturally that individual transition costs are non‐decreasing in ƒ, the 

optimal critical mass in terms of ƒ can be described by the path 

s * (ƒ¸) (ƒ + 1) ((1‐)(‐1)/ L) 

varieties of the industrial economy. 

 

 

sI ) + sI with sI as the basic input 

Substituting this into the objective function with cost c (ƒ,1; sI, ), the ‘big bang’ 

industrialization policy problem becomes 

min (ƒ + 1) ((1‐)(‐1)/ L) sI ) c (ƒ,1; sI, ) 

where use is made of the symmetry of the cost functions and the fact that s sI 

firms are targeted. could represent any given exogenous parameter, i.e. or 

L a given parameter linked to L 

In designing an optimal industrialization policy it shows that a cost minimizing 

policy in the industry transition entails setting certain development model 

(exogeneous) parameters such as labour productivity improvements (), upstream 

firms discount future earnings () the fixed size of the labour force (L ), the number 

of basic industrial sector varieties (sI) , the product linkages between intermediate 

input producers (), and the use of the intermediate input composite (), the latter 

two showing a certain degree of interaction referring to as the returns to 

9 

349





specialization (Romer, 1986). Discussing these parameters qualitatively in terms of 

comparative statics would indicate industrial change. Raising any of these 

parameters , , L , and sI increases the marginal returns to upstream firms in both 

their entry and modernization decisions. 

Raising means that sunk costs are translated into labor improvements more 

effectively. Similarly, since the costs of modernization and entry are carried today 

and most of the returns occur in the future, the more likely they are to undertake 

those actions. A large market, a higher L , also raises the marginal return to entry 

and modernization. Finally, more industrial varieties mean that the past level of 

industrialization is greater, thereby, reducing the marginal costs of inducing firms to 

adopt more modern technologies. Given this, the responsiveness of firms to 

inducements by the government is enhanced when any of these parameters is raised. 

Therefore, the higher are these parameters, the fewer firms need to be targeted to 

facilitate an escape (from a development trap). Of these parameters has probably 

received the most discussion. In many ways, this parameter represents the strength 

of increasing returns in the technology adopted by upstream producers. This is 

because higher levels of imply that, when they choose to modernize, upstream 

firms will choose technologies involving greater sunk (or fixed) costs. Therefore, 

while one requires some degree of increasing returns or economies of scale in 

production to generate a rationale for a “big bang” intervention, the stronger are 

those increasing returns to support a more unbalanced industrialization policy. This 

relates back to arguments made on balanced vs. unbalanced growth. Of the three 

other parameters, only the discount rate seems to have been given a potential role 

in the past debate on industrialization policy. Matsuyama (1992) interprets the 

discount rate as measure of effectiveness of entrepreneurship in coordinating 

investment, with a low discount rate indicating existence of greater entrepreneurial 

resources. If so, then the above result seems to imply that with a relative scarcity of 

entrepreneurial talent a more balanced approach should be followed. 

The comparative statics results for and require more restrictions because each of 

these has two effects. On the one hand, lowering and increasing raises the 

strength of strategic complementarities among upstream sectors. This tends to favor 

a more balanced growth approach. On the other hand, and each affect the 

marginal returns to entry and modernization of firms. The second effect reinforces 

the first and leads to more balanced strategy that is, lowering and lifting 

increase the marginal returns to entry and modernization. A lower also implies 

stronger technical complementarities. This effect is sometimes referred to as the 

returns to specialization (Romer, 1987). The consequence is that a lower raises the 

marginal returns to employing greater variety of inputs in production. The higher is 

the weaker linkages among intermediate input sectors. Conversely, stronger 

linkages between sectors raise the marginal return to targeting an additional sector 

for change supporting the arguments of the balanced growth strategy. 


350




Looking at , it is a measure of the appropriability of the returns from supply as an 

additional intermediate input. As Romer (1994) discusses, the larger is , the greater 

is the surplus gained by intermediate input producers from the employment of their 

product in final goods production. Therefore, producers of inputs targeted in an 

industrialization policy are more likely to react positively (in terms of adopting 

better technology) when the appropriable returns from the introduction of their 

variety is larger. This effect would tend to favor a more unbalanced approach as 

increases. 

Summarizing, we have outlined the role of several parameters in influencing the 

kind and degree of balance in industrialization policy. Factors addressed in the 

earlier literature such as strength of linkages, increasing returns and entrepreneurial 

resources all influence the composition of the ‘big push’. By considering a ‘big bang’ 

policy, some results are possible. For instance, strong increasing returns in 

conjunction with weak sector linkages tend to favor a more unbalanced approach in 

order to minimize costs. 

CONCLUSIONS 

A major problem of the industrialization debate is the timing of the industrialization 

policy and its degree of focus is complex and dependent on the characteristics of the 

case specific economy. A ‘big push’ perspective on industrialization does not imply 

that transition can be a simple matter of coordinating expectations via some kind of 

indicative planning. Nor does it mean that policy must be balanced and take a ‘big 

bang’ form in order to be successful. A wide variety of industrialization policies can 

generate a ‘big push’ and the choice between them is therefore a matter of costs. 

In a dynamic model, however, this wide variety of industrialization policies makes a 

characterization of the optimal policy quite difficult. To take advantage of full 

marginal modernization and entry costs, a gradual policy is always optimal. 

Moreover, in a policy of gradual entry, the number of sectors targeted in each period 

is rising over time. However, pairwise interactions between choice variables and 

exogenous parameters tend to be qualitatively ambiguous in a dynamic setting. For 

instance, strong increasing returns accompanied by weak sector linkages tend to 

favor a more unbalanced approach in order to minimize costs. The former effect 

favors the arguments of the balanced growth school, while the latter was part of the 

intuition of the unbalanced growth school. 

11 

351





REFERENCES 

Barro, R. A. & Sala‐i‐Martin, X. (1995). Economic Growth, Cambridge, Ma.: MIT 

Press. 

Easterly, W. (2002). The Elusive Quest for Growth, Cambridge, Ma.: MIT Press. 

Easterly, W. (2005). Reliving the ‘50s’: the Big Push, Poverty Traps and Takeoffs in 

Economic Development, Center for Global Development, Washington, D.C., WP No. 

65, Aug. 2005, 1‐37. 

Fukuyama, F. ed. (2008). Falling Behind, Explaining the Development Gap between 

Latin America and the United States, Oxford: Oxford University Press. 

Gans, J. (1994). Industrialization with a Menu of Technologies: Complementarities, 

Market Structure and Multiple Equilibria, Economics Department, Stanford 

University. 

<strong>Gottinger</strong>, H. (2006). Innovation, Technology and Hypercompetition, London: 

Routledge. 

<strong>Gottinger</strong>, H. & Goosen, M. (2012). Strategies of Economic Growth and Catch‐Up, 

New York: Nova Science. 

<strong>Gottinger</strong>, H. (2014). Catch‐Up and Convergence: Mechanism Design for Economic 

Development, Journal of Applied Economics and Business, 2(3), 67‐95. 

Hirschman, A. O. (1958). The Strategy of Economic Development, New Haven: Yale 


Krugman, P. R. (1991). History versus Expectations, Quarterly Journal of Economics 

106(2), 651‐667. 

Lin, J. Y. (2013). Against the Consensus, Cambridge: Cambridge University Press. 

Matsuyama, K. (1992). The Market Size, Entrepreneurship, and the Big Push, 

Journal of the Japanese and International Economies, 6, 347‐364. 

Murphy, K. M., Shleifer, A. & Vishny, R. W. (1989). Industrialization and the Big 

Push, Journal of Political Economy, 97, 1003‐1102. 

North, D. C. (1990). Institutions, Institutional Change and Economic Performance, 

Cambridge: Cambridge University Press. 

Nurkse, R. (1952). Some International Aspects of the Problem of Economic 

Development, American Economic Review, 42, 571‐583. 

Nurkse, R. (1953). Problems of Capital Formation in Underdeveloped Countries, 

Oxford: Basil Blackwell. 

Ortiz, C. H., Castro, J. A. & Badillo, E. R. (2009). Industrialization of Growth: 

Threshold Effects of Technology Integration , Documentos de Trabajo, 


352




Departamento de Economia, A.A. 25360, Cali, Columbia. 

Romer, P. M. (1986). Increasing Returns and Long‐Run Growth, Journal of Political 

Economy, 94, 1002‐1037. 

Romer, P. M. (1987). Growth Based on Increasing Returns Due to Specialization, 

American Economic Review, 77(2), 56‐62. 

Romer, P. M. (1994). New Goods, Old Theory, and the Welfare Costs of Trade 

Restrictions, Journal of Development Economics, 43, 5‐38. 

Rosenstein‐Rodan, P. N. (1943). Problems of Industrialization of Eastern and 

Southeastern Europe, Economic Journal, 53, 202‐211. 

Rosenstein‐Rodan, P. N. (1961). Notes on the Theory of the “Big Push”, In: Economic 

Development for Latin America, H. S. Ellis and H. C. Wallich (Eds), New York: St. 

Martin’s. 

Rostow, W. (1960). The Stages of Economic Growth, Cambridge: Cambridge 


Streeten, P. (1956). Unbalanced Growth, Oxford Economic Papers, 167‐190. 

13 

353



354

3.5 SOME APPLICATIONS OF A RESULT IN CONTROL THEORY TO ECONOMIC PLANNING MODELS 


355



356



357



358



359



360



361



362



363



364



365



366



367



368



369



370



371



372



373



374



375



376

3.6 OPTIMAL ECONOMIC GROWTH WHEN CO2 CONSTRAINTS ARE CRITICAL 


377



378



379



380



381



382



383



384

3.7 ECONOMIC CHOICE AND TECHNOLOGY DIFFUSION IN NEW PRODUCT MARKETS 


385



386



387



388



389



390



391



392



393



394



395



396



397



398



399



400



401



402



403



404



405



406



407



408



409



410



411



412



Technology Diffusion 119 

Mansfield, Edwin, "Technical Change and the Rate of Imitation". Econometrica, Val. 29, 

1961, pp. 741-766. 

-, Industrial Research and Technological Innovation. New York 1968. 

Manski, Charles, Daniel McFadden, Structural Analysis of Discrete Data with Econometric 

Applications. Cambridge, Mass., 1981. 

McFadden, Daniel, "Conditional Logit Analysis of Qualitative Choice Behavior". In: Paul 

Zarembka (Ed.), Frontiers in Econometrics. New York 1973, pp. 105-142. 

Metcalfe, J. Stanley, "Diffusion of Innovation in the Lancashire Textile". Manchester School 

of Economics and Social Studies, Val. 10, 1970, pp. 35-50. 

Nabseth, Lars, George F. Ray, The Diffusion of Industrial Processes: An International Study. 

Cambridge 1974. 

National Science Foundation, Technology in Retrospect and Critical Events in Science 

(TRACES). Washington, D.C., 1968. 

Nelson, Richard (Ed.), Government and Technical Progress: A Cross lndustry Analysis. New 

York 1982. 

-, Sidney G. Winter, An Evolutionary Theory of Economic Change. Cambridge, Mass., 1982. 

Oster, Sharon, ''The Diffusion of Innovation among Steel Firms". The Bell Journal of Economics, 

Val. 13, 1982, pp. 45-56. 

Peterka, V., Macrodynamics of Technological Change: Market Penetration by New Technologies. 

1.1.A.S.A. RR-77-22. Laxenburg, Austria, November 1977. 

Reinganum, Jennifer F., "Technology Adoption under Imperfect Information". The Bell Journal 

of Economics, Val. 14, 1983, pp. 57-69. 

Rosenberg, Nathan, Inside the Black Box: Technology and Economics. New York 1982. 

-, Ralph Landau (Eds.), The Positive Sum Strategy. Washington, D.C., 1986. 

Scherer, Frederic M., Innovation and Growth: Schumpeterian Perspectives. Cambridge, 

t\1ass., 1984. 

Spinrad, Bernard 1., Market Substitution Models and Economic Parameter. 1.1.A.S.A. 

RR-80-28. Laxenburg, Austria, July 1980. 

Telser, Lester G., "A Theory of Innovation and Its Effects". The Bell Journal of Economics, 

Val. 13, 1982, pp. 60-92. 

* * * 

Zusammenfassung: ökonomische Wahl und Verbreitung der Technologie auf den 

Märkten für neue Produkte. - Dieser Aufsatz versucht modellhaft zu entwickeln und zu 

schätzen, wie sich Innovationen unter den Produzenten im Falle von Spitzentechnologien 

ausbreiten. Nach der Einführ'ung einer Innovation soll es annahmegemäß Firmen geben, die 

die Innovation übernehmen, und Firmen, die sie imitieren; außer diesen beiden Firmengruppen 

gibt es keine anderen Marktteilnehmer. Die grundlegende Annahme ist, daß es einen 

inneren Zusammenhang zwischen dem Muster der Diffusion und der Dynamik des Marktes 

gibt. Es wird unterstellt, daß eine "repräsentantive" Firma in einem gegebenen Markt vor die 

Wahl gestellt ist, ob sie sich auf dem Markt betätigen will oder nicht bzw. ob sie aus ihm 

ausscheiden will oder nicht, und daß die Wahlmöglichkeiten von wirtschaftlichen Parametern 

413



414

Hans W. Gottinger, Essays on Technology, Markets and Development

Create successful ePaper yourself

Delete template?

Save as template?