2 - Pondicherry University DSpace Portal

AN EMPIRICAL VALIDITY OF MARKET INTEGRATIONHYPOTHESIS WITH SPECIAL REFERENCE TOTHE PEPPER MARKET IN KERALATHESIS SUBMITTED TO PONDICHERRY UNIVERSITYFOR THE AWARD OF THE DEGREE OFDOCTOR OF PHILOSOPHYIN ECONOMICSM.V. KANAKARAJDEPARTMENT OF ECONOMICSSCHOOL OF MANAGEMENTPONDICHERRY UNIVERSITYPONDICHERRY 605 014INDIAAPRIL 2005

The minor corrections as suggested by the Examiners areincorporated in the following waySuggestions of Dr. Tan Hui BoonPara 9 0P Problem statement given in page no.3.P Need of the sturly given in page no. 13.> "List of Tables and Figures" has deleted from coiltelits page.Para 10P Results of Akailce Inforn~ation Criterioi~ are given in Table 5.3a (pageno. 158). Disc~ission is added in page 149 and 150.k Table 5.4 amended as Table 5.4a (page no. 159) and Table 5.4b (pageno. 1 60). Its discussion is incorporated in page 110.1 5 1.Suggestions of Dr. L.P, SwamhathanP IIistory of pepper, cultivatioi~ and narration of the crop were given in pagejl2f A

M.V. KANAKARAJPh.D. ScholarDepartment of EconolnicsPondicherry UniversityPondicherry-6050 14DECLARATIONI hereby declare that the Thesis entitled, "AN EMPIRICAL VALIDITYOF MARKET INTEGRATION HYPOTHESIS WITH SPECIALREFERENCE TO THE PEPPER MARKET IN KERALA", submitted by mefor the award of degree of Doctor of Philosophy in Econolnics is a record ofresearch work done by me and that the Thesis bas not previously formed thebasis for the award of any Degree, Diploma, Associateship or any othersimilar title.Place : PondicherryDate : 4 K ~ 2 0 0 5 M.V. Kanakaraj

ACKNOWLEDGEMENTFirst and forelnost let me express my profound gratitude toDr.K.Sham Bhat, Professor and Head, Department of Economics,Pondicherry University for his sagacious guidance, pungent criticism andprobity in approach. His presuinptuousness facilitated me to opt for the presentarea of research and also to complete the thesis in the stipulated time.I am indebted to my revered teacher Prof. P.P. Pillai, forinely Head,Department of Econoinics, Dr.Johi1 Mathai Centre, Uiliversity of Calicut, forbeing a constant source of inspiration which helped me to land in PondicherryUniversity.Inspite of his official preoccupation, I got invaluable statisticalassistance froin Dr.S.Amanulla, Reader, School of Management, PondicherryUniversity. I aln deeply obliged to him for his expertise, humility andcoininitinent towards the subject.I take this opportunity to express my gratitude to the faculty inelnbersof the Department of Economics, Pondicherry University for their timelyassistance and eilcouragen~eilthroughout my stay at the University.My sincere thanks are due to the Principal of Government M.G.M.P.G.College, Itarsi, Madhya Pradesh, Head of the Department of EconomicsPr0f.P.K. Mishra and all the Colleagues who smoothly facilitated my sojournill Pondicherry .It is my duty to convey my gratefulness to the Ministry of HigherEducation, Government of Madhya Pradesh for sanctioniilg study leave andextending all official assistance for the completion of the thesis.

As a matter of fact, no word can convey my heartfelt gratitude to mymother. my wife Nimya and my daughter Vismaya for their affection andperseverance showered on me inspite of several odds.The family members of Pr0f.K. Shain Bhat owe much respect for theirforbearance during my visit to his house.My heartfelt thanks are due to Mr.J. Manoj, Senior Assistant,Establishment, Pondicherry University, Mr. P.R. Madhusoodan and Mr.K.Durairaj, P1i.D. Scholars of Department of Economics, PondicherryUniversity for their timely assistance.Special thanks are to Mr. Vinod Krishna, Assistant Professor ofZoology and Mr. K.L. Chourey, off'ice staff, botli of GovernmentM.G.M.P.G. College, Itarsi, who pleasantly sorted out all the administrativetechnicalities during my study period.I will be failing in my duty if I sun not acknowledging the hostelinmates of Pondicherry University, who have nurtured a family atmospliereand extended all sorts of help.It is a matter of pleasure to express my deep sense of gratitude to theLibrarian and assistants of Pondiclierry University Library, Centre forDevelopment Studies, Trivandruin, Institute for Social and Ecoi~omic Change,Bangalore, Directorate of Arecanut and Spices Development, Calicut, IndianInstitute of Spices Research, Calicut, Spices Board, Kochi, Directorate ofMarketing and Inspection, Kochi and India Pepper and Spice TradeAssociation, Kochi for the library assistance offered to me.Mr. D. Venkata Ramana Moorthy, Office Manager, Department ofEconomics, Pondicherry University has extended all assistance without anyhesitation throughout my stay. Let me acklzowledge sincere thanks to him.

The painstaking task of reading the manuscript was undertaken by Dr.P. Bhaskaran Nair, Reader, Academic Staff College, Pondicherry University.Let me convey my gratefulness to him for hearing me in the midst of his busyschedule.Last but not the least, my sincere thanks are to Mr.R.Saravaoan, NewLink Coinputers, Opposite to Pondicherry University-I1 Gate, for his dexterityand proinptness in executing all the coinp~rter work.

CONTENTSCHAPTERPAGENO.I. INTRODUCTION11. REVIEW OF LITERATURE111. MARKETING EFFICIENCY AND MARKETINTEGRATION HYPOTHESES -A THEORETICAL FRAMEWORKIV.TRENDS IN AREA, PRODUCTION,PRODUCTIVITY AND PRICES OF PEPPERV. AN EMPIRICAL VALIDITY OF MARKETINTEGRATION HYPOTHESIS WITHREFERENCE TO THE PEPPER MARKET INKERALAVI.SUMMARY AND CONCLUSIONSBIBLIOGRAPHY

No.4.1.LIST OF TABLES AND FIGURESTitlePage No,TRENDS IN SHARE OF PEPPER IN SPICES EXPORTFROM INDIA 122TRENDS IN COUNTRY WISE AREA, PRODUCTIONAND YIELD OF PEPPER 124INDIA'S PERCENTAGE SHARE TO WORLD ACREAGEAND PRODUCTION OF PEPPER 127TRENDS IN COUNTRYWISE EXPORT OF PEPPER 128TRENDS IN STATE WISE AREA, PRODUCTION ANDPRODUCTIVITY OF PEPPER 130AREA UNDER MAJOR CROPS IN KERALA 132STATES / U.T. SHARE OF AREA AND PRODUCTION OFPEPPER TO INDIA 133TRENDS IN DISTRICT WISE AREA, PRODUCTION ANDYIELD OF PEPPER IN KEFL4LA 13 5COMPOUND GROWTH RATE OF VARIOUSINDICATORS OF PEPPERRESULTS OF DICKEY-FULLER (DF) AND PHILLIPS-PERRON (PP) TESTS (1 974.4 TO 2003.3) 156RESULTS OF DICKEY-FULLER (DF) AND PHILLIPS-PERRON (PP) TESTS (1974.4 TO 1991.6 AND 1991.7TO 2003.3) 157RESULTS OF AKAIICE INFORMATION CRITERION 15 8RESULTS OF JOHANSEN'S MULTIPLECOINTEGRAI'ION TEST 158RESULTS OF ERROR CORRECTION MODEL DURINGPOOLED PERIOD 159RESULTS OF ERROR CORRECTION MODEL DURINGPOOLED PERIOD 160RESULTS OF ERROR CORRECTION MODEL DURINGPRE-EFORM PERIOD 161RESULTS OF ERROR CORRECTION MODEL DURINGPOST-REFORM PERIOD 1625.7 IMPACT OF PRICE OF PEPPER ACROSS THE SELECTEDMARKETS (AT ONE LAG) 163Figure 4.1. TRENDS IN PRICE OF PEPPER ASSEMBLINGMARKETS OF KERALA (POQLED PERIOD) 138

CHAPTER IINTRODUCTIONPhysiocrats were the earliest school of thinkers who have highlightedthe significance of agriculture. A sound agricultural base has been a matter ofself-esteem for any country. Agriculture has inultifarious interlink with severalmacroeconomic aggregates. Food security via agriculture is the hallmark ofnational security. 'There is an age-old causal relation between agriculture andeconoinic development. Agriculture sector is the lifeblood of any economy.Firstly, it is the source of food supplies to the country. An uninterrupted,qualitative food supply ensures healthy people and hence, a healthy economy.Secondly, it supplies basic raw inaterials to industries. Rapidindustrialisation and hence econoinic development is possible only throughsustained agricultural developmeilt. As a corollary, by providing raw inaterialsto industries, it transfers inanpower to non-agricultural sector and hencefacilitates overall developlnent of the economy. Agricultural sector also acts asa market for industrial products right froin fertilisers to agriculturalimplements. Therefore, the growth of the industrial sector depends on thestrength of agricultural sector and hence, econoinic growth and developinent.Thirdly, it is a good foreign exchange earner throug1-1 enlarged exports.Nowadays in the era of globalisation, strength and stability of a country ismeasured on the basis of foreign exchange reserves. Foreign exchange is a

pre-requisite for development, via import of necessary machinery andtechnology.Fourthly, agricultural sector provides a base for capital formation.Capitalforination is supposed to be a pre-condition for econolnicdevelopment. The main source of capital formation is rural savings. Higherinarltetable surplus in the agricultural sector paves the way for rural savingsand hence, economic developinent via capital formation.Fifthly, it directly contributes to the domestic product of a nation.Contribution of agriculture sector to Indian econoiny is quite significant.Nature has blessed this nation with 50 per cent of its land suitable forcultivation against a global ratio of 10 per cent(Sharma and Ainbastha,1995). Indian agriculture nearly contributes 25 per ccnt towards GrossDomestic Product and about 70 per cent depends on it for their livelihood(Govt of India, 2003). It provides exnployinent to 56.7 per cent of country'sworltforce and accounts for 14.7 per cent of the country's total export earnings(Panchal, et. al, 2003).An efficient inarlteting system in agricultural sector is essential to keepthe pace of agricultural growth. In ordinary parlance an agricultural marketcan be said to be efficient if there is a mif for in price for an identical productprevailing in the entire market area. Agricultural ~narketing is a specialisedactivity which endeavour to establish an equilibrium between production andconsumption (Ghatage, 1958). However, it is the type of competition prevalent

etween the contracting parties that deterlni~ic whether there will be a rise inrevenue to the fariilers, fall in consumer's price or rise i11 prolit of iniddleinenor a coinbination of all these possibilities (Jasdanwalla, 1966).Problem StatementMarketing efficiency is of great significance as it is the pivot tofanner's response to agricultural production and inarlteted surplus. In ancfficieilt inarketing system, producers are able to get reimunerative prices totheir products and coilsuiners to get the product at affordable prices.Attainment of remunerative prices will teinpt thein to produce more; and thiswill lead to greater inarltetable surplus. Greater lnarltetable surplus will helpto generate capital forinatioil and foreign exchange. Crop specialisationthrough tlic principle of coinparative cost advantage is possible only in acompetitive and efficient inarltet.In other words an efficient ~narltetingsystem is the pre-condition to achieve food security.The eff'xcient marketing systein depends on marlcet mechanism. Andthe e~~iciency of inarlcet mechanism in turn depends on the regional priceiiitegration. Uniformity of prices can be attained through the integration ofnational economy. The difference between prices at various locational pointsshould not be higher than warranted by reasonable calculation of transport andother costs (Gadgil, 196 1).For examining the efficiency of a marketing systein, one mustinevitably look at the degree to which village primary, secondary and terminal

marlets are related to each other (Lele, 197 1).Thus the concept of marketintegration emerges in the picture. Whenever the actions of agents of onemarltet affect the actions of agents of other inarltets, it is said to be a situationof integrated market. Markets are ltnown to be interlinked when transaction inonc influences tlie terms of exchange in other inarlcets. According to Faminowand Henson (1990) integrated inarltets are those wherein prices are determinedinterdependently; which is assulncd to nlean that pricc change in one will befully passed on the others. Monke and Petzel (1984) defined integratedmarkets as marltets in which prices of differentiated products do not behaveindependently. Ravallion (1986) observes that an equilibrium will have theproperty that, if trade taltes place at all between any two regions, then price inthe importing region equals price in the exporting region plus the unittransport cost incurred by moving between the two. Goodwin and Schroeder(1991) cautions that inarltets that are not integrated inay convey inaccurateprice inforination that might distort producer's marketing decision andcontribute to inefficient product movelnents.Thus, an interrelated or interdependent movement of prices betweenspatially separated inarkets can be said to be a situation of market integration.An integrated marltet system is synonymous with an efficient marketingsystem. Therefore, the conccpt of market integration percolates the basic

principle of attaininent of maxilnuin utility with the most eff.icicnt utilisationof resources available in the marketing system.*Policy inalcers have argued that liberalisation is required to attaininarl

Petzel ( 1984). Heytens ( 1986), Ravallion (1 986), Delgado ( 1986), Dahlgrailand Blank (1 992). Sorensen ( 1993), Kallfass (1 993), Zhao ( 1995), Zanias( 1993), Gardner and Brooks (1 994), Silvapulle and Jayasuriya (1994),Baharuinshah and Habibullah (1 994), Cawalho, et. a1 (1 994), Alexander andWyeth (19941, A~igulo and Gil (1996), Bijinan (1996), Rozelle, et. a1 (19971,Baulch (1997), Munir, et. at (1997), Khedhiri (1999), Isniet, et al, (1999),Asche, et. a1 (1 999) etc*.At ~~atioilal level also, attempts were made to investigate the lnarltetii~tegration hypothesis. The significant studies are done by Lele ( 1967, 197 I),Blyn (1973), Thalcur (1974, 1998), Rudra (1980), Naik and Arora (1986),Patiiaik (1988), Palaskas and White (1993), Padinanablian (1993),Narasiml~ain (1994), Sinharoy and Nair (1994), Nasurudeen and Subramanian(1995), Mathew, et. a1 (1997), Behura and Pradhan (1998), Ghosh (2000),Basu and Dinda (2003), Prainod Kuinar and Sharma (2003) etc.*The above studies at National and International level have attempted toverify the niarl

approach, coefficient of variation, parity bound model, Engle-Granger'scointegration and Johansen's lnultiple cointegration technique were ernployedto test the validity of market integration hypothesis. Most of the studies haveuscd monthly wholesale price to examine the inarket integration hypothesis.Few studies were relied either on daily or weekly prices. A big chunk ofstudies were able to reveal an existence of strong market integration in theiranalysis. Rejection of inarlcet integration hypothesis is colnparatively a raresit~iation.Thus, it can be cited that alinost all these studies investigated theinarket integration hypotliesis of food crops. Majority of these studies foundout the existence of strong forin of inarltet integration arnoizg agriculturalinarltets. However, the existing literature reveals the following lacunae.Most of the studies at national and international level has verified thenlarltet integration hypothesis of food crops. In some of the regionaleconoinies non-food crops are doininant over food crops. According toNational family health survey (2001), only 28 per cent of Indian agriculturalarea is covered under non-food crops; while the same is 76 per cent to the stateof Kcrala. Still inarlcet integration analysis related to non-food or cash cropswere alinost neglected.At the national level, studies related to the inarltet integration ofagricultural products are mainly related to states such as Maharashtra, TamilNadu, West Bengal, Punjab, Gujarat, Andhra Pradesh, Haryana, Orissa and

ICerala. Here also it can be noticed that food crop is the major iten1 of analysis.However, an integrated study pertaining to Indian economy as such and rest ofthe states are lacking.Ainong non-food crops, spices is a major item. India is known as thelallcl of spices and has a dominant position in the production of spices andcondiinents and accouiits for 35 per cent of global trade. Among Indian states,Kerala has inore or less dominance over production of solme major non-foodcrops. Pepper accounts 97 per cent, rubber 92 per cent, cashew 85 per cent,cardainoin 70 per cent, ginger 60 per cent and coconut 43 per cent of Indianproduction (Kerala State Land Uses Board, 1997). Tliougl~ Kerala had a nearrnonopoly over pepper production and also fetches valuable foreign exchange;no serious attempt has so far been made to know about the validity of inarltetintegration of pepper market.On inethodological ground also, the earlier literature shows certaindrawbacks. Majority of studies have used price series correlation to measureinarltet integration. But correlation coefficient can not be taken as a reliableindicator of market integration since comnlon trends in tiiiie series analysismay make an upward bias of the results. It can also be maintained that aperfect monopoly or price fixing by a central authority can produce acoefficient of one as a perfectly competitive market. Hence, though correlationindicates the nature of price movements among markets, it can not be taken asa true indicator of market integration.

To get rid of the problem froln correlation, economists began to applyregression techniques. Regression has the benefit of adding tiine trend to therelationship. But in practice, studies ignore the time series properties of pricedata, and the results obtained inay be biased and inconsistent. Hence,regressio~~ results are affected by econometric shortcoinings as spuriousregression, non-stationarity in data series and inappropriate use of firstdifference etc.Now the long run relationship between non-stationary series can betested by the cointegration approach. Cointegration gives a way to reconcilekindings ol' non-statiunarity with the possibility of testing relationship ainonglevels of econoinic variables. The modern technique that can be applied to testinarltet integration is Sohai~sen's rnaxiinuin liltelihood method. This method isused to know the cointegration properties in multiple time series analysis.In a coiltrolled or regulated economy; prices will never reflect the truemarket situation. Liberalisation is intended to release control over inarltets andhence, to allow the forces of market to play its own role. It is maintained thatthe actual transmission illechanisln of inarltet can be realised only through thefree play of demand and supply. With easing of controls, farmers will be ableto move their products to that inarltet froin where they can fetch remunerativeprices. Then price difference will reflect only the transportation costs. Thus,liberalisation will help in removing bottleneclcs in transporting products in

etween markets. This eliminates the problem of excess supply or demand inany of tlie markets; and will lead to uniform price throughout the system.Econoinic reforms and liberalisation initiated by the Government ofIndia in one way or other have also transmitted into the Kerala economy.Certain drastic policy measures with private participation in infrastructuresuch as transportation sector, power sector, inforination technology etc. hasinitiated by Kerala government. Administrative pricing and distribution ofcommodities at controlled prices havc eliininated to a greater extent. Froin thevery old days, market restriction on pepper was negligible. In pace witheconoinic reforins, however, export and licensing procedure become simple.Thus, with opening of inarkets to global competitors, a cominodity likepepper, which has high export exposure, will benefit through a unified market.Therefore, one can readily suspect that liberalisation will have a directimpact on market integration and hence marketing efficiency. On this countsome international studies (Fafchainps, 1972, Cawalho, et. al, 1994, Rozelle,et. al, 1997) have shown that liberalisation had a positive effect by reducedinarlceting margin and hence liberalisation and coinmercialisation madelnarltet efficient. At national level no serious attempt is made to know theeffect of liberalisation on marketing efficiency. It is natural to believe thatwith easing of regulations and restrictions there is greater chance fordissemination of market information which will ultimately lead to an efficient:competent market.

On policy perspective, inarket integration studies has severalimplications. An integrated market is required to iinplement the followingpolicies. A famine policy intended to smoothen food availability and securityis possible in a unified market. Agricultural policies of any kind can besuccessfully implemented only wit11 an integrated marltet. Pricing policiesineant for remunerative prices to producers and affordable prices to consuinersbecome effective only in a unified marltet system. Fruits of liberalisation canbe attained to the producers and consulners through a well integrated marketstructure. Successful iinpleinentation of foreign trade policies has a directbearing on the efficiency of inarket. Long term planning on the basis ofcomparative cost advantage by the fariners is possible only through anintegrated inarket.Therefore, it can be maintained that studies on inarlcet integration arepertillelit in several respects. But no serious attempt to know the marketefficiency of a doininant crop like pepper has not taken place either at nationalor rcgional level. Pepper is produced througliout the length and breadth ofKerala either as a lnonocrop or as a mixed crop. There is a large number ofsmall and inarginal fariners producing pepper. Usually these producers used tosell their products to small traders which is spread throughout the state. Thesesinall traders after collecting the produce resell it to the wholesale dealers. Thewholesale dealers then transport the article to the major pepper assemblingcentres of the state viz., Tellicherry, Kozhilcode, Kochi and Alleppey. These

assembling centres in turn transport pepper to the terlninal inarket - Kochi;froill where the products are exported to various parts of the world. Thus, insuch a trading network the following issues emerges.(i) To what extent pepper lnarkets of Kerala economy is integrated? Ifit is integrated, is there any transmission mechanism of prices between themarkets?(ii) To what extent, recent ecoi~oinic liberalisation had an impact oninarket integration and transmission inechanisin of prices between the marketsof pepper product of Icerala economy?Objectives of the StudyOn the basis of the issues mentioned above, the proposed studyexamine the following objectives,(i)to exainine the earlier literature pertaining to market integrationhypothesis with special reference to agricultural commodities and toidentify the gap of the studies.(ii)to evaluate critically the inarket integration hypothesis and theinethodological ilnproveineilt to test the hypothesis over the years andto identify ail appropriate inethodological tool for verifying thehypothesis.(iii)to examine the trend and pattern of area, production, productivity andprices of pepper with special reference to Kerala economy. Besides, we

attempt to investigate the iinpact of econolnic liberalisation of 1991 onarea and production of Kerala's pepper economy.(iv)to investigate the empirical validity of inarket integration hypothesisamong major pepper assembling centres of Kerala and to identify thetransmission inechai~isin between the markets during pre and postreforinperiods.Need of the StudyThe present study is justifled on tlze following grounds.(I)Since pepper is a doininant cash crop of Kerala, inarket integrationstudy of pepper will be worthwhile to know the efficiency of peppermarketing.(2) Study on the effect of liberalisation on pepper econoiny of Kerala willbe fruitful to evaluate the efficacy of econo~nic reforms.(3) Comparative analysis of pepper acreage and productivity of coulitrieswill be useful to adopt necessary policy changes.(4) Knowledge of price transinission mechanism will help to forinulatenecessary marketing strategies to root out inflexibility in the market.MethodologyRatio and percentage methods are employed to examine the trend andpattern of area, production and productivity of pepper economy of Kerala.Coinpound growth rates are also computed to examine the trend of area,production and productivity. Dummy variables were introduced to the trend

equation in order to identity the impact of liberalisation. Dickey-Fuller andPhillips-Perron test were einployed to verify the existence of marketintegration hypothesis during pre and post-reform periods. Besides, Johansen'slnultiple cointegration technique and error correction inodel were einployed toidei~tify the transmission mechanism of prices between the selected inarltetsduring pre and post-reforin periods of pepper markets of eral la.*'l'ellicherry, Kozhikode, Kochi and Alleppey are the inajor pepperassembling markets of Kerala. To get a clear picture of inarket integration ofpepper, all these inarlcets were talten into consideration. Due to non-availability of continuous daily or weekly pepper prices, monthly wholesaleprice of pepper for all the inajor assembling centres of Kerala for the periodApril 1974 to March 2003 were selected to verify the inarltet integrationhypothesis. To know the impact of ecoiiomic liberalisation on pepper inarketintegration of Kerala, the period of study has been divided into two. They are:(i) April 1974 to June 199 1 and (ii) July 199 1 to March 2003.Data SourceThe study is exclusively relied on secondary data. Many governmentalpublications were extensively used to obtain data under several heads. Majorsources are various issues of Statistics for Planning, Econoinic Review,Agricultural Abstract of Kerala, Statistical Abstract, Spices statistics, Cocoa,Arecanut and Spices statistics, Arecanut and spices data base, Area and* See Chapter five for detailed discussion.14

production of spices in India and the world. Pepper statistical year book,International pepper news bulletln and the unpublished data of Directorate ofArecanut and spices developinent for the period 1960-6 1 to 2002-2003.Limitations of the Study1 The present study is 111nlted to cash crop of pepper Its results can notbe generalised to other cash crops or for the non-food sector of Keralaeconomy as such, because crops are heterogeneous in nature and theextent of market also varies2 Val~dity of inarltet tntegration hypothesis is limited to monthly database for the year 1974- 20033 Our study is restricted to four ilialor pepper assembling inarkets ofICerala It need not necessarily reflect the actual behaviour of all thereinaining local inarkets of the state which are not so doininant withreference to pepper i~larket4 Following liinitations of Johansen's inultiple cointegration techniquealso applies(a) Likelihood ratio statistics are obtained under the normalityassuinption(b) Assuines that the underlying data generating process has a finiteorder autoregressive representation with known lag structure.(c) Johansen's cointegration tests are sensitive to underparameterization in the lag length, though not to overparameterization.

Scheme of the Study:The present study consists of six chapters and it is presented below.Tl~e first chapter deals the issues, objectives, methodology, data source,liin~tations and scheme of the study.Earlier literature pertaining to the study area and the gap of the studiesare examined in second chapter.Theoretical analysis of marketing efficiency and market integrationhypotheses were exainined in the third chapter. Besides, it evaluates thevarious statistical and ecoiloinetric techniques employed by earlier studies toverilji the lnarltet integration hypothesis and transmission mechanisin of pricesbetween the inarltets.Chapter four deals with the trend and pattern of area, production,productivity and price of pepper in Kerala economy. It also analyses theiinpact of liberalisation on trend and pattern of area, production, productivityand price of pepper.An e~npirical validity of market integration of the selected pepperinarltets of Kerala during pre and post-reform periods were examined in thefifth chapter. Besides, it also verify the transmission mechanism of pricesbetween the markets of pepper in Kerala during pre and post-reform periods.Summary and conclusions are presented in the sixth chapter. Further, itdeals with the policy implications and scope of future work in this area.

CHAPTER I1REVIEW OF LITERATUREIn this chapter it has been attempted to review the earlier work done inthe area of market integration. Studies on inarltet integration of agriculturalcommodities and allied activities were taken into consideration. For simplicityand clarity, the order of presentation is as follows:(a)(b)Studies related to International level.National level studies.International Level StudiesFafcliamps (1972) examined reasons behind the wealthier farmers ingrowing cash crops. The first part of the paper presents a simple theoreticalinodel of crop portfolio choice, the second part on effects of consuinptioilpreference on output choices and the third part the possible effect of marketintegration on optimal crop choice is simulated for various types of producers.Sim~~lation are based on Taylor approximation. Parameters used for simulationare chosen to represent a typical third world fanning household. The authorargues that the correlation between individual and aggregate output is alsolikely to decrease with market integration, thereby reducing the price andrevenue correlation. A glimpse of the strong pleading for market integrationcan be obtained through - 'food' market integration via reduced trade

estriction. better roads and transportation, and 1 or government food shops canbe a powerful tool to boost cash crop production and to increaseresponsiveness of small fanners to price incentives.Hays and Mc Coy (1978) examined spatial and temporal aspects ofmarketing efficiency for the traditional marketing system for millet andsorghum in Nortliern part of Nigeria. Analysis of pricing efficiency of theinarlteting system was accomplished by examin~ng movement between pricesat fifteen selected locations in Nigeria's four Northern States during 1969-71.Spatial price relationship were analysed by examining inter-market pricedifferentials in relation to transport and other transfer costs. Temporal pricerelationship were analysed by examining significance of storage costs as afactor In explaining seasonal price rises. The analysis revealed that positiveprice spreads is due to an erratic nature of supply, an inadequate disseininationof information on prices and supply in the various markets and lack ofspecialisation in trade by traders. Storage operations were considered to be thereason for interteinporal price increases. The author points out that lack ofinarlcet integration ainong the fifteen urban locations studied resulted in spatialprice differentials that in some cases exceeded transfer costs.Harris (1979) made a detailed survey of studies on inarket performanceand inarket integration. The survey encompasses studies undertaken in India,United Kingdom, Afiica, Nigeria etc. Until then almost all studies have usedcorrelation coefficient as a measure of market integration and competitiveness.

Citing several studies Harris maintains that high correlation coefficient inaycharacterize a situation of physical disconnection and low coefficientscharacterise regions with coi-nplex trading pattern. The author indicated thatcorrelatioil coefficient analysis fails to explain market integration due tosecularly rising prices and it is caused by increased population growth andeffective demand, widening range of crop varieties, trading relationship of ajoint destination market, monopoly procureinent at fixed prices, inflationarytrend etc.Lundahl and Petersson (1982) gave some more evidence of problemsregarding the use of price series correlatioii as a tool of il~arket integration.Besides, an attempt is made to calculate correlation coefficient in the sameinanner as carried out by Blyn with Cumining's data for Haitian inarket for theperiod 1969- 1974. Rice, grain, millet, grain corn, ground corn and red beansare the products chosen for the analysis. The average correlation coefficientsfor the raw series, after grouping and detrending were obtained. The averagecoefficients for each of the products is approximately in the same range as theone calculated by Blyn (1973). The lower correlation coefficient is citedprobably for trade which is not uni-directional. The scholar argues that thespecific nature of Haitian marketing system also nullifies the use of priceseries correlation for inarket integration.Analysing market integration of international trade in cotton was thesubject matter to Monke and Petzel (1984). The data for the estimation of the

pair wise price relationship included twenty series of monthly prices. Monthlyprices were averaged to produce annual series for the period 1962-79. Theprices are deflated by the world bank index of CIF prices to generate stabletime series and reduce the possibility of spurious correlation. Thirteeninternational price series are included and the data are organised by staplelength. Bivariate price regression and hedonic index estimation are themethods used to identi@ whether differentiated products are amenable totreatment as homogenous commodity. The analysis reveals that theinternational cotton inarlcet is integrated across the shorter staple lengths; andsliort, medium and long staple cotton inay be treated as a hoinogeneouscommodity. The authors were of the opinion that coi~sumption rather thanproduction adjustments are the constraints for price lnoveinents of exports ofalternative countries.Heytens (1986) employed Ravallion inodel to examine the validity of~narket integration l~ypothesis with reference to Gari (processed cassava) priceand Yam prices. The results indicate that gari market coinprised a fairly wellintegrated system after the first five years. Yam prices froin a subset ofEastern Nigerian cities result showed a dismal integration. Local seasonalitywas identified as the source of poor intcgration. Besides, the author points outthat as a matter of fact the Ravallion model gives a much broader range ofresults than earlier bivariate correlations.

Ravallion (1986) throws light on the inferential dangers in usingbivariate correlation or regression coefficients as a measure of spatial marketintegration in agriculture. He suggests that the main dangers of the simplebivariate mnodel can be avoided if the static bivariate inodel is extended into adynamic model of spatial price differentials with the same data. By acceptingshort run dynamic adjustment process. Ravallion offers an approach to testlong run inarket integration. The analysis was done by using data on theinterregional price differentials for rice in Bangladesh during the turbulentpost-independence period (1972-75). Ravallion posits an autoregressivedistributed lag relationship among each local price of a coininodity and anappropriate reference price level. The analysis revealed that marketsegmentation performs poorly as a restricted forin of the general model for alldistricts and short run integration continues to be weak when long runintegration is imposed.Delgado (1986) developed a variance coinponent method to test foodgrain market integration in Nigeria. The approach is to decompose thevariance of food grain prices into components. The model was applied toeighteen inonths of weekly grain prices for twenty two villages in NorthernNigeria for the period August 1976 to March 1978. Empirical result reveal thatinarltets are not well integrated in the six inonths covering the harvest period.Dahlgran and Blank (1992) investigated the integration of a system ofdiscontinuous and continuous markets. The discontinuous markets are those in

wh~cl~ transact~ons do not occur during same time period. They observe thatwhen a d~scontinuous market is part of a spatial system, the degree ofintegration of the continuous markets may depend on operation ofdiscontinuous markets. For emp~rical test of integration. data from six westernU.S. alfalfa markets from April 1, 1984 to March 29, 1987 were used. Theanalys~s revealed that continuous markets are less integrated duringdiscontinuous market operation and long run market integration differ byseasonSorensen (1993) attempted to study the impact of product marketintegration on welfare of an economy. Usually there are welfare gains fromintegration either due to increasing returns to scale or to firms' or unions' lossof market power. The author's main purpose was to show that in an econoinywith centralised wage setting, integration of product market inay give rise tolower welfare. The experiment was made with a general equilibrium model. Itis shown that highly centralised as well as coinpletely decentralised econoiniesin general, have better employment performance than econoinies with a degreeof centralisation in between. Utility function is applied to illustrate howintegration of product market inay give rise to a decline in welfare. Theanalysis concluded that the real wage is higher and employment, real incomeand profits are lower when product markets are integrated.

Kallfass ( 1993) examined the impact of long terin contracts and verticalintegration between farm and the food industry would reduce costs andimprove the competitive situation of German agriculture. The author'sanalysis revealed that the choice among spot market sales, long-term contractsand vertical integration depends on key factors such as specific physicalassets, specific location and difficulty in lnonitoring quality as suggested intheoretical analysis. Kallfass' findings do not confirin the hypothesis thatgreater vertical integration is necessarily cost saving. Hence, he argues thatgovernment policy should not distort coinpetition between differentdistribution system by favouring a particular type of co-ordination.Zanias (1993) investigated the degree of spatial inarket integration inEuropean com~nunity of agricultural product markets. Failure to observe asingle price tllrouglrlout the conlinunity could be attributed to one or more ofthe following: (i) lack of linkage by arbitrage between agricultural rnarlcetsbetween member states (ii) impediments to efficient arbitrage and (iii)imperfect competition in one or more of the markets. Zanias with the help ofco-integration analysis investigated the impact of the above reasons to nullifythe force of the law of one price in the European coininunity agriculturalproduct market. Law of one price is tested for four European communityagricultural products which differ both in terms of product characteristics andpolicy framework namely, soft wheat, cow's milk, potatoes and pig carcasses.Test result reveals an existence of a single price in the soft wheat market; and

it may be due to the operation of miniinurn intervention prices rather than thedifferent markets being truly integrated in a spatial sense. Zanias observes thatlnarlcet integration fails in the European corn~nunity milk ~narlcets due to non -tariff barriers to intra - coinmunity trade or imperfect competition. In the caseof pig carcasses and potatoes, cointegration is established in three out of sixcases.The intention of Gardner and Brooks (1994) was to examine the extentto which econoiilic reforms in Russia constituted genuine price liberalisation,i.e., food prices that respond to supply and demand conditions. Both linearand non-linear equations were worked out to test inarlcet integration. Testresult showed that for every commodity, the hypothesis that= 1 in all citescan be rejected at one per cent level of significance and hence, there is littleconsistency across coi~lmodities in which city is most closely integrated withthe Moscow price. The results of city market prices also indicated a similartrend. From OLS estimates strong evidences for lack of consistent relationshipbetween the distance and income variable and the city price difference werealso obtained. They have observed that the bchaviour of these price series isdominated by oblast - level (forinal or inforinal) regulation of enterprises andmarl

on the monthly average price of rice of five selected markets for the periodJanuary 1975 to December 1989. Johansen's multiple cointegration techniquewas applied to test inarket integration. The result of the Johansen's multipleco-integration technique indicates that Phillippines rice markets are generallywell integrated in the long run with Manila as the central market.Baharurnushah and Habibullah (1994) made an attempt to determinewhether prices of black and white pepper in a market are in parity with pricesin a reference marltet. The cointegration method developed by Engle andGranger (1987) was employed to analyse the long run relationsl~ip betweenprices in different markets. The period of observation spans from the firstweek of January 1986 to the last week of December 1991. The test resultsshows that regional pepper markets in Malaysia are spatially linked. Theauthors concluded that due to low transportation cost and risk, the degree ofcointegration is unaffected by distance and hence, price changes are fully andiininediately passed on to the other markets.Carvalho, et. al. (1994) examined the agro-industrialverticalintegration process in Brazil's sugar cane and alcohol sector in the period1970-92. The analysis revealed that the Brazilian sugar sector already hasvertical integration and it was due to well established relaxation of restriction,seasonality of the raw material, emphasis on profit, and administrative pricepractices.

Alexander and Wyeth (1994) employed Granger method ofcointegration and Johansen's maximum likelihood procedure to test marketintegration and employed data on monthly prices from January 1979 toDecember 1990. The method was illustrated with data on prices in differentparts of the Indonesian market. The authors observed that the consumer priceIndex (CPI) is consistently cointegrated with all the rice price series, whichmeans an existence of apparent market integration. Besides, this analysis alsorevealed that supply sources are more important than demand sources indriving prices.Zhao (1995) in an article seeks to draw general trends for thedevelopinent of the integration of agricultural production, processing andinarlteting in China. He concluded that there is a high degree of integration ofagricultural production, processing and marketing. Besides, he pointed out thatintegration of agricultural production, processing and inarketing and the closelinks between agricultural production supply and sales result from thedevelopinent of market econo~ny .Angulo and Gil (1996) analyscd the impact of vertical integration onprice transmission in the Spanish poultry sector by employing error correctionmodel. Feed price, producer's price and consumer prices are considered forthe analysis. Monthly data fioin January 1981 to December 1992 wereconsidered for an empirical analysis. Angulo and Gil's coinputation ofimpulse response functions and decomposition of the forecast errors variance

shows that producer and consumer prices immediately rise after shock. Theyfurther pointed out that these response endure is for approximately two years,indicating that poultry fin tend to price according to long term goals, therebyshowing the price adjustment process.Fafchamps and Gavian ( 1996) studied spatial integration of livestocklnarltet in Niger by using co-integration approach. The study shows thatlivesloclc markets are poorly integrated. Prices are seldonl co-integrated,suggesting that large price differentials occasionally persist between adjacentareas for long periods of time. Parity bound approach indicates that one has toassunie high transportation costs and large quality variation to reconcile thedata with efficient spatial arbitrage. Besides, the analysis confirins descriptivestudies that have einphasised regional segmentation in West-African livestocktrade.Bijinan (1996) examined the link of biotechnology and verticalintegration in the Dutch potato chain. In general studies on the impact ofbiotechnology have stated that development and introduction of thistechnology may lead to vertical integration on the agrofood chain, makingfarmers morc dependent on the input supply industries and the food industry.It is pointed out that even without biotechnology the agrofood sectorexperiences structural changes leading towards more horizontal and verticalintegration. Further, the analysis revealed that biotechnology will reinforce thetrend only if consumers accept products made with the new technology.

Rozelle. et. al. ( 1997) examined the impact of liberalisation on rurallnarket integration in China. The impact of inarlcet integration on pushingproducers to inore effectively utilise their comparative advantage was alsoanalysed. Price and market liberalisation were taken as a way of raising theefficiency of China's food econoiny and to increase rural income. A uniqueand colnprehensive set of data on provincial prices of major food coininoditiesbetween 1988 and 1995 were considered for the analysis. The analysisrevealed a falling coefficients of variation for provincial rice and maize and itimplies a sign of increasing integration. The number of pair of province thatbecame integrated went up by more than four times for rice markets and moretl~an doubled for inaize markets during 1988-89 and 1991-93. This is anindication of an expanding geographic range of spatial market integration.Illcrease in rank correlation coefficient results reveals that liberalisationpolicies appear to have been successful in encouraging farmers to move intocrops in which they have a comparative advantage.Baulch (1997) developed an alternative inethodology known as ParityBound Model (PBM) to test Philippine rice inarket integration. The authorargued that all the conventional tests (price correlation, causality, errorcorrection and co-integration) rely on price data alone and fail to recognize thepivotal role played by transfer costs. Transfer costs(comprisingtrailsportation, loading and unloading costs and traders' normal profit)deternine the parity bounds within which the prices of a homogeneous

comlnodity in two geographically distinct inarkets can vary independently.Violations of the spatial arbitrage conditions indicates that there areiinpediinents to trade between inarltets and to be viewed as evidence of lack of~narlzet integration. The parity bound lnodel developed in this paper assess theextent of inarket integration by distinguishing among three possible traderegimes. They are: (i) the parity bound (spatial price differential equalstransfer cost) (ii) inside the parity bound (price differentials are less thantransfer costs) and (iii) price differentials exceed transfer costs. To assess thestatistical reliability of the parity bound model, a series of Monte Carloexperiments were performed. Three alternative trading scenarios for the spatialprice equilibrium lnodel integrated, partially integrated and independentinarltets are considered in the Monte Carlo simulations. The analysis revealedthat the sum of the probabilities of trade (i) and (ii) are interpreted as theprobability of inarket integration of Philippine rice inarkets.Munir, et. al. (1997) analysed market integration of Indonesianvegetable market. Four selected vegetables (chilli, shallot, potato and cabbage)in nine consumer and three producer markets in Indonesia are considered forthe analysis. The results revealed that none of the markets are segmented.Furtlner, short run and long run inarltet integration tests revealed thattransportation and product perishability are the important factors in explainingthe speed of price transmission.

Khedhiri ( 1999) empirically analysed agricultural market integration inTunisia. Cointegration technique was e~nployed to examine the objective. Theempirical result shows that the degree of inarket integration is low forwholesale market, particularly for the storage products. Besides, the analysisrevealed that the distance between markets and the volume of transaction cannot explain the lack of linkage between the markets.Tsmet, et. al. (1999) evaluated the long run spatial price relationship inIndonesian rice markets and factors affecting the degree of market integration.By relying on the weekly pricc data for thc period 1982- 1993, they employedmultivariate co-integration test for verifying ~narltet integration. Besides, theyclassified their evaluation into pre-self sufficiency and post-self sufficiencyperiod. The co-integration tests revealed a smaller degree of marketintegration in Indonesian rice market. Further, the analysis revealed thatgoverninelit intervention in terns of rice procureinent significantly influencedmarlcet integration during the period of post- self sufficiency (1985-93) andthe pooled period (1 982-93).Asche, et. a1 (1999) by using the Johansen procedure analysed worldsalmon inarket integration to test the law of one price and to evaluate thepossibility of product aggregation. Their elnpirical investigation also include(a) a co-integration analysis of world Salinon export prices during 1986-1996,(b) an analysis of the dynamic relationship between the price series and (c) anerror correction model which assess short run responsiveness of the prices to

one another. For empirical analysis five species of salmon were considered.Since inultivariate cointegration test indicates four cointegration vectors(hence one cominon stochastic trend in the system) the scholars conclude thatthere is one market for all salinons. Parameter stability test indicates that thesalmon inarket is well integrated during the study period.National Level Studies1,ele (1967) examined market integration of Sorghum prices in WesternIndia. Five priinary inarltets in Sholapur district and two terminal inarltets aresclected for the analysis. The analysis is based on weekly wholesale prices forthe period 1958 to 1963. The two hypotheses tested in this article are: (i)inarlcets of agricultural coininodities in underdeveloped countries are closelyinterrelated (ii) Price differences between markets do not tend to be greaterthan transport costs because of the competitive nature of wholesale trade.Correlation coefficient is used to test the degree of market integration. Leleobtained high correlation coefficicnt between prices and maintains that itsupport the hypothesis that agricultural markets are fairly competitive and thatprice inoveinents in a single market are influenced by prices in other markets.Lele (1971) made an extensive study of market integration of Indiangrain markets. Comparable varieties price data for the year 1954-1965 of rice,wheat and Jowar in the four major states of West Bengal, Tainil Nadu, Punjaband Maharashtra were considered for the analysis. Correlation coefficient isused to test market integration. The analysis revealed the following31

conclusions: ( i ) H). and large collusion. either tacit or overt is uncomlnon inthc Indian grain trade. High profits earned by t'ew traders are not inonopolisticrcturns but can bc attributed to the large volume of operations resulting fromtheir command of capital(ii) Examination of regional price disparitiessusgests that grain markets are closely related to each otl~er (iii) The study oflnarltet integration suggests that a reaso~lably well organised competitivesystem of private trade exihls in India and (iv) Existence ul' price differencebetnee11 regions, are rnainly due to lack of adequate transportation facilitiesand hindrance to perfect mobility i~nposed froin outside the trade sector suchas transport bottlenecks and official restrictions.Blyn (1973) questions the validity of using correlation coefficients totest the presence of market integration. He maintains that even if markets arewell integrated, correlation measures of tlieir price series will not necessarilybe high. Besides, he observes that time series correlation should be restrictedto residuals remaining after the trend and seasonal components have beenremoved. An increase in population may affect all prices in a region, even ifeach market within the region was independent of others. Blyn reworkedCuinming's eight year collection of monthly wheat prices in eight Punjabinarkets and Delhi by eliminating trend and seasonal influences. The analysisrevealed that even if inarkets are well integrated, correlation coefficients maynot be high because these marltets are not simply supply centres but alsocentres of iinportance for local consumption.

I h:ll\ur ( 1074) e?\a~r~ined priciilg efficie~~cy of the marketing system bynnalq'sing price trends. marliet integration and price spread in the inarketingchannel of Gyj arat foodgrains. Foodgrains Lvere pertained to baj ra. j owar,paddq and wheat. Weekly ~vholesale prices from 1965- 197 1 were used for theslatistical analysis. Carrelatioil coefficient is used to test the degree of marketintegration. Test result shows correlation coefficient to be higher in certainmai-kets and low and negative in the case of' bajra. And for paddy and jowarcoel'ficient seems to be high. I11 the case of wheat; correlation is relatively highsii~ce it is relatively a scarce cornlnodity in Gujarat. Thus, the analysisrevealed that [he existing foodgrain inarketing system on the whole is notefficient.liudra (1980) made a critical analysis of' the concept of inarltetingeficiency as defined in several studies of Indian agriculture especially of UinaJ. Lele and Z.Y. Jasdanwalla. He argued that without any scientific basis theseauthors are propagating the idea that markets for agricultural com~noditieswork successfully in India. Uina J. Lele's contention that product prices areequalised in foodgrains marltet is criticised by the author by saying that thereis very little theoretical analysis of the concept of competition in itsapplication to foodgrains. Competitiveness of the market is questioned on thebasis of correlation coefficient that prices can frequently be unifos111 undermonopoly or oligopoly, not even reflecting any difference due to transportcosts or storage costs. The claim of single price for different parties entering

thc grain marl\et i\ also cl~iestioned. I he concept or cficienc! nas cleared by/l

intcrc\l

3'lad1-:1\ tit! . Salc~n. C'oi~nhatorc. Matiurizi.l iriincl\ eli, l~rnakula~n andE\;oihap~~r. 1 hc ~il~olcsale ncckl! pricc 1;)s thc pcriod 1982-83 ncre consideredti)r thc anallsis. High corrclritic~n coefticient mas obtair~ed in all the markcts.I lwcc, the anal)-sis concluded that thc ~narket ofjaggery ;ire \$ell integrated.Natasirnhalla ( 1994) examined the integration of groundnut markets ofIQiiq alscerna regioi~ of Andhra Pradesl~. The study covered a period of sixyears fso~l~ 1973 to 1979 with daily price data. Narnsimham have considered13ombay as tenninal markct and Hyderabaci and Madras as regional markets.Koyc1i.s distributed lag ~nvdel was e1lzp1oyt.d to examine tho ob.jective. Theauthor observes that the results from the statistical analysis justify thel~ypothesis that the groundnut oil price in a given market is being influencedby the groundnut oil price in the itninediately higher level marftct. Thus, thea~~alysis proves that the oil price integrates the groundnut ~narkets vertically.Sinharoy and Nair (1994) examined the ~noveinents in internationalprices of Indian pepper reflect the variations in such prices of othereconomies. Dickey - Fuller, Augmented Dickey - Fuller and co-integrationtechnique was einployed to examine the objective. Monthly spot prices ofIndia, Indonesia and Brazil for the eighties are used far the analysis. Resultssilow thd the international prices of pepper for Indonesia and India havemoved synchronously in the long run despite short run drifts and it is due tooligopolistic nature of the world market of pepper.

Na\ur~~dccn iincl Suhr:i~niit.zi,in ( 1905 Ic.\:l~ni~icci tilt. 1 tilidit!of' (i)\ct-tical integration of' sccd prim ra price of its oil anti cahc and (ii) tiori~ont;tIintegration ol~prices oi'diff'ercnt oils. '1 cn oil products nerc considered hr theanal? sis ol' bori~ontul and \ crtical pricc integration. Kc)) ck's distributed laginodcl Ifas emplo~ ed to test the integri~tion of'oil prices. I Iorizontal integrationtest rcsults relealed that the price of'groundnut oil intluenced thc prices of allotl~cr oils c~ccpt castor oil. C'ustor oil price mas intluenccd onl>1 by linseed oilprice since they are substitutes. Vertical integration res~~lt:, rcirealed that thereexist some imperfection in seed price formation. It is also inferred that tl~eprice of industrial oil ir~fluenced the price of edible oils but not \. lice 'versa.The researchers concluded that vertical integration i17oilseed price was ln~ichquiclcer as coinpared to horizontal integration in oil priccs.An attempt is made to study the long run behaviour of the farm pricesof coconut in various rnarlcets of Kerala by Mathew, et. al. (1997). Averageyearly farm price considered for the analysis is arrived by taking a simpleaverage of monthly farm prices of 23 years data from 1970 to Dec. 1992 for25centers in Kerala. The co-integration method developed by Eilgle andGranger is employed in the study to test the coconut market integration. Thetest result indicated that the farm price of coconut in various markets of Keralawas integrated of order one. After establishing the order of integration of eachvariable, pairwise co-integration were carried out with the farm price ofTrichur market as the independent variable and the respective farm price in

other m;lrl\ct.; as clcpenllc~~t \ :tsiitl~lcs. I hc 4tN re\ calccl that all the lnarkctcsccpt C'alicut rnitrLct \+ csc intcgsatcd \-i.ith 1 richur 1narht.t.I'hnhur ( 1OL)X) madc a dctailcd study of' the concept of marketingufficienuj . (:orrc.l;ititw cc>efficicnt is emplo~cd as a stntjsric;~l tool lo ~ncasurethc ~ieg-L.L. of'tllaskct integration. 'lilcckl ii holcsale prices ot'it. heat market of(ii~.jarat a11d apple iri dil'ferent ter~rii~~al markets of' India fbr thc year 1985-86to 1995-96 Lr crc empirically tested to examine the objective of the studq7. Theanalysis revealed that wheat and apple markets are integrated. Further, theallal ysis cautioned that high degree of' integration may cainc siinplq. as a resultot'collusion on the part ol'traciers.Behiira and Pradhan (1998) made an attempt to identify marine fishmarkets in Orissa are integrated and efficient. 'The analysis relied on datapertaining to the ~veekend marine fish prices for the last week of twelvemonths for the period January 1984 to December 1992 from among 30 oddfish markets. Bivariate price correlation as well as the inetl~odology developedby Engle and Granger (1987) has been einployed to show whether marine fishil~arkets are cointegrated or not. To test the univariate price series forstationarity, the Augmented Dickey Fuller test was also employed. Theanalysis revealed that the bivariatc correlation coefficieilt ranged between 0.60to 0.85. Augmented Dickey Fuller test revealed that the price series formarine fish in the selected markets in the state are stationary after firstdifference. The test statistic of cointegration test obtained for all the painvise

mashet~ ;I~C!i)iltld to 1~less than thc aslmptotic critical \.;tlut t.1-cn at It.) pcrcent lei cl csccpt that ot'C'~ltta1;- f'aradip pairs. I Icncc. ~IIC analysis c011~1udet.ithat marine tish lnarliets in tile state arc not integrated and it is mainlyattributed to poor intinstructure facilities at landing centrcs as well as atterniinal sccondarqv rnarkets.Cihosh (2000) examined spatial integration ol'ricc 111arkels in India. Thecn~pirical analysis was carried out on the basis of data ot' ~rlonthly tvholesaleprices of rice for the period from March 1984 to April 15397. l'rice data relatesto state-specitic varicties of rice quoted in different ~narket centres of fourselected states viz. Hihar, Orissa, Ilttar Pradesh and West Berlgal. To exatninewhether intra-state and inter-state regional rice markets are integrated andlinked together into a single econo~nic ~narliet. the Maximum Likelihood (ML)method of co-integration developed by Johansen and extended by Johansenand Juselius was used. To examine the univariate time-series properties of thedata and for non-stationarity, Augmented Diclcey Fuller test was conducted.The finding of one coinmon stochastic trend for Uttar Pradesh iiriplies that allthe prices are pair-wise co-integrated. On the other hand the presence ofinultiple coinmon stochastic trends in Bihar, Orissa and West Bengal signifiesthat the prices are not pair wise co-integrated.Basu and Dinda (2003) attempted to evaluate ei~lpirically spatialintegration of potato inarket in Hooghly district of West Bengal. Bivariateprice correlation as well as co-integration test and error-correction method

dci clopcd I-r?lfngle-Granger has heen used to shot$ \vhtlther potato marketsarc intcgratecl or not. The stud) is based on time scrics data un wholesale andrclilil prices of potato in the selected three importilnt inurhet centres namely.C'hu~npadanga. l'arnkc~vsh~i.ur and Shcariipl~ull~ in the district of Hooghly fortlw period ol'.ianuary 1998 to Dcccmber 2000. On the basis of'l~igh values ofcorrelation cocflicicnt, the anal~sis rcvealed that marlicts are stronglycorrelated ancl they are highly inter dependent in price fonnntion.I'ramod Iturnas and Shas111a (2003) tried to evaluate price integrationand pricing cfficicncy to the state of Haryana. Johansen'sinultiplecointcgration method was e~nployed to test price integration. lhe integrationtests were carried out with the monthly wholesale price of coarse paddy forfour ~narltets of Haryana. To linow the inipact of liberalisation the period wasdivided into pre-libcralisation (October 1978 to September 1989) and post -liberalisation period (October 1989 to September 2001). The multivariatecointegration tests results indicates the presence of three cointegrating vectorsat one per cent level of significance for both pre and post - liberalisationperiods. It implies that all thc four paddy markets are cointegrated and henceexhibit a long run relationsl~ip.I-Iowever, results of error correction inodel reveals a very weakassociation ainolig these markets. The authors observed that this weakassociation is because of paucity of availability of information and lack ofquicker dissemination of available information. But the adjustment process

nas found to be quicker in post-liberalisation period in coinparison with pre-li beralisation period.Concluding RemarksFroin the earlier studies it can be observed that for the last four decades,a series of studies have undertaken to verify the validity of market integrationhypothesis of various agricultural crops and products of allied activities.Immense studies were conducted at international level in colnparison withnational studies. Most of the studies have talten food crops for their analysis.Some experts have also attempted to evaluate inarlceting efficiency of certainnon-food crops. Different statistical and econometric tools like Correlationcoefficient, Coefficient of variation, Regression analysis, Ravallion model,Autoregressive model, Koyck's distributed lag model, Variance coinponentapproach, Engle-Granger's cointegration, Johansen's multiple cointegrationand Parity Bound Model were employed to test the validity of marketintegration hypothesis.Most of the studies have used monthly wliolesale price to examinemarket integration hypothesis. Some of the studies have relied either on dailyor weekly prices. Majority of the studies were able to identify the existence ofstrong form of market integration. Rejection of market integration is a rarephenomenon. However, the existing literature reveals the following lacunae:

(I)Most of the studies at national and International level have given muchemphasis to food crops. Market integration analysis related to non-foodcrops or cash crops were alinost neglected.(ii)At the national level, studies are related to states such as Maharashtra,Tamil Nadu, West Bengal, Punjab, Gujarat, Andhra Pradesh, Orissa andHaryana. There is only a single study of coconut market pertaining toKerala economy. But studies on pepper, which is a dominant crop ofKerala is lacking.(iil)Some international studies have shown that economic liberalization hada positive effect on inarketing efficiency. At the regional level, noserious attempt is made to know the effect of econoinic liberalization onn~arlteting efficiency.(iv)On inethodological front also there are some drawbacks. Much of theearlier studies have relied on correlation, regression and Engle-Grangercointegration techniques. But these techniques have several limitations.Correlation simply shows the associatioil between two variables.Regression technique gives an idea of the effect of one variable on theother. Engle-Granger method is used to know the nature of relationbetween bivariate marltets. However, the studies employing Johansen'smultiple co-integration test to identify the existence of marketintegration across multiple marltets are rare in the literature.

The above lilnitations related to market integration studies calls forinore studies on regional cash crops. Evcn though only one-fourth of thelndian agricultural area is devoted for non-food crops, Kerala has earmarkedinore than three-fourth of its area for non-food crop cultivation. Among nonfoodcrops of Kerala, pepper contributes to 97 per cent of Indian production.But no serious attention has been paid to the study of market integration of thiscrop. Hence, the present study is an attempt to fill this gap.

CHAPTER I11MARKETING EFFICIENCY AND MARKET INTEGRATIONHYPOTHESES - A THEORETICAL FRAMEWORKIntroductionThe concepts of 'integrated market' and 'efficient market' are usedinterchangeably. To have a better understanding of the concepts of marketingefficiency and inarket integration, it is necessary to know the concepts of'marliet', 'integration' and 'efficiency' and the way in which these areinterrelated. Generally market is any region in which the buyers and sellersinteract each other, in which the price of a good tends to uniformity. Usually,price prevailing in the market depends on the extent of market; which in turndepends on the nature of competition or efficiency prevailing in the market.The extent of competition or marketing efficiency in its turn dependson the niarltet structure, marllet conduct and inarltet performance. Marketingefficiency is determined by two factors - economic efficiency and technicalefficiency. Econo~nic efficiency deals with matters related to trading or pricingto enhance the degree of competition. Technical efficiency on the other, triesto apply the least cost input combination.There are two criteria to measure inarlceting efficiency. One is pricespread and the other is market integration. A brief account of the methods tomeasure price spread will be dealt in this chapter. Besides, a detailed analysis

of various statistical and econo~netric tools einployed by earlier authors tomeasure the degree of market integration will also be made. Hence, the presentchapter is intended to examine these concepts in its theoretical and empiricalperspective. Besides, it also attempt to identify the limitations of earlierinethodology and the course of future study.Concept of MarketFor conceptual clarity of the term 'marlcet', let us briefly discuss someof the iinportant definitions of market. These definitions in general, can becategorized into three on the basis of the emphasis they were given. The firstone emphasizes the existence of a public place for transaction. The termmarket is a derivative of a Latin word 'inercatus' to denote a market place -thereby meaning merchandise, trade or a place where business is conducted(Graviii, 1929). According to Jevons, "the central point of market is a publicexchange inart or auction rooms, where the traders agree to meet to transactbusiness.. . the traders may be spread over a whole town or region of country,and yet make a market, if they are by ineans of fairs, meetings, published pricelists, the post office or other wise in close communication with each other"(Quoted from Marshall, 1961). Cochrane (1957) observes that market is somesphere or space, where the forces of demand and supply were at work, todeterinine or modify price since the ownership of some quantity of a good, orservice was transferred and certain physical and institutional arrangementsinight be in evidence.

The second category undermines the need for any specific location orspace. According to Cournot (197 1) "not any particular inarltet place in whichthings are bought and sold. but the whole of any region in which buyers andsellers are in such free intercourse with one another that the prices of the samegoods tend to equality easily and quickly". Bliss and Stern (1982) opined thatinarltet refers to exchange of the services of factors take place and thearrangements in force for organizing that exchange. There is no implicationthal the market is in any sense a formal one with a specified location; still isthere any suggestion that the market is perfect or competitive. According toStonier and Hague (1982) market is "any organization whereby the buyers andsellers of a good are lcept in close touch with each other, whenever the inarketis open, either because they are in the same building or because they are ableto talk by telephone at a moment's time".Other than the existence of location or space; the third set of definitionsgive emphasis on the prevailing price out of the interactions of agents.Hotelling (1929) in analyzing the relationship between prices in competingmarkets; focuses on market for identical goods separated by distance. Stigler(1969) defined market as "the area within which the price of a good tends touniformity, allowance being made for transportation costs". An observation onsimilar line was made by Cournot (1971), shortrun deviations of prices areallowed in this definition, but arbitrages or substitutability insure that they arerelated in the long term. "A market is a group of people and firms who are in

contact with one another for the purpose of buying and selling somecorntnodity. It is not that every members of the lnarket inay be in contact withevery other one; the contact may be indirect" (Dorfman, 1979).Though there are some differences in defining the term market amongeconoinists, one can observe that the basic requirement for a market is thattrading or exchange should take place between buyers and sellers. It inay bedirect or indirect, rnay be in a small region or the entire globe; may be ofvis~ial contact or invisible contact. In our analysis, we use the term market asdefined by Stigler. In this definition by lnalcing due allowance fortransportation cost, the existence of a unified price got prominence. Hence,price is tlze villain of market; and it depends on the extent of market.Extent of the MarketOne of the proiainent roles of inarltet is to facilitate exchange betweenbuyers and sellers. Stigler maintains that market area embraces the buyerswho are willing to deal with any seller, or the seller who are willing to dealwith any buyer or both. It can be inaintained that the actual test of lnarket isthe uniformity of price inovements within the market. This criterionencompasses the crucial role of cornpetition in dominating the priceinovenlents within each section of the market.The idea of exchange and price formation will be clear by observing theview of Stigler and Sherwin (1985). "It is inherent in any exchange, whetherof one good for another good or for money, that there be a rate of exchange47

etween the quid and the quo; a quantity of something is exchanged for aquantity of something else. Therefore to say that a market facilitates themalting of exchange is equivalent to saying that markets are where prices areestablished. One may quote a price for a colninodity on the moon if one isvisiting that celestial body, but one can only establish a price by making atrade. The ~narltet is the area within which price is determined: the market isthat set of suppliers and demanders whose trading establishes the price of agood".The organizational structure of ~narltet strongly determines the processby which prices and output are determined in the real world. Koutsoyiannis(1979) has suggested three basic criteria for market classification. They areproduct, substitutability and interdependence criterion. Bain has suggestedanother criterion for market classification, namely the condition of entry,which measures the ease of entry in various inarltets.Harris, (1984) opined that the analysis of the structure of coinmodityinarltets norinally proceeds down a list of characteristics of their organisation:size, distribution, location, entry condition, agent and product differentiation,information and so on. She observes that the numerical size of the sector andits concei~tration are the two structural aspects most important for the analysisof mercantile power. Salvalore (1998) identifies four different types of inarltetorganizations.

(a)(b)(c)(d)Perfect competition at one extremeMonopoly at the opposite extremeMonopolistic coinpetition andOligopoly in betweenHowever, it can be maintained that the actual market power depends onthe coinpetition or inonopoly power. The tilt of this power determines thebenefits either to the buyer or to the seller. Competitive power is one of thebasic criteria to distinguish various forms of market. To understand theextent of coinpetition or efficiency, it is necessary to lulow the structure ofmarltet.Market StructureThe extent of market depends on several factors. According to Bain,(1968) three distinct approaches can be followed to understand the extent ofcoinpetition or inarketiiig efficiency in the inarlteting system. They are:(i)(ii)(iii)inarket structure,inarlcet conduct, andinarlcet performance.Seller concentration, firm's size, buyer concentration and entryconditions are the basic elements of market structure. These elements in oneway or the other influence inarket integration. Seller concentration or buyerconcentration inhibits the free flow of goods and services among marltets, Thisin turn distorts the spirit of a unified or integrated market. Similarly if the49 .

entry condition is restricted, the biggest firm may control the entire lnarket andthis lead to wealcly integrated marltets. Thus, these ele~nents of marketstructure affects the degree of competition in the marlcet and that in turninfluence the magnitude of lnarket integration. Therefore the degree of marketintegration is determined by the structure of market.Bain refers by the term market structure to "those characteristics of theorganization of a inarket which seems to influence strategically the nature ofco~npetition and pricing within the market".The characteristics of marketorganization eii~phasized were the degree of seller concentration, size of thedistributing firms, degree of buyer concentration, degree of productdifferentiation and the condition of entry in the market. The views of Georgeand Singh (1970), Garoian (1971), Purcell (1973), Caves (1977), Dahl andHaln~xond (1977) and Bhide, et. a1 (1981) were the same as that of Bain(1 968).However, the above views are not able to highlight the significance ofvarious marketing channels and intermediaries in analyzing the marketstructure. In the view of Schultz (1946) inarket structure, includes all thestrategic variables, which control or influence the behaviour of differentagencies involved in the market.An all-encompassing version was given by Cundiff and Still (1972). Tothem, market structure was the whole net work of ~narlceting institutions thatserviced society's needs. At one end of the network, producers initiated the

flow of goods and services and various interinediaries such as wholesalers andretailers n~aintained the flow. finally discharging the goods and services forconsumer's use. 'To Lele (1973) market structure included various marketchannels. interinediaries, and traders involved in moving the produce froinproducers to the consumers.According to George (1984) market structure could be defined as allthe agencies involved either vertically or horizontally in the selling and buyingof the produce. It includes different marketing channels, their form and marltetshares and the marlcet environment.Thus, the market structure through various marketing channelsinfluences the nature of competition and pricing within the market through theintermediaries. However, it is in the inarltet structure that the inarltet agentshas to function. And it is the structure that deterinine th~;.ks@.a:'~r,i~u~~pf the" - ' ,: -.. iJ*I/--"^ ..+*, ,".; -'hb." d , . ".*; '1,.&" ;; . Y"ij*inarltet i.e.; conduct of the marltet.*

actions of governinent also determines the market conduct and thus marketintegration. Government restriction and regulation hampers dissemination ofmarket information and it will lead to distorted price determination by theeconomic agents. This ultitnately caters to an inefficient and non-integratedmarket. On the other hand, the behaviour of econoinic agents in an economywhich is liberated from controls will be conducive for an efficient and wellintegrated inarltet.In the opinion of Bain (1968), inarlcet conduct refers to the pattern ofbehaviour followed by the enterprise in adopting or adjusting to the ~narkets inwliicli they sell or buy, in particular methods employed to determine prices,sales proinotion and co-ordination policies and the extent of predatory orexclusionary tactics directed against established rivals or potential entrants.According to Moore, et. al. (1973) ~narltet conduct colnprises several methodspractised by traders to attract the custoiners in their fold. It includes severalprice cornpetition methods and non - price inducements. According to Purcell(1973) inarltet conduct refers to the actions and behaviour of firins within thegiven structure. Pricing policies, selling cost, non-price coinpetition are allsome of the activities of inarket conduct.Hence, ~narltet conduct resembles the behavioural pattern of enterprise.It comprises of various decisioi~ making techniques of Grin in determiningprice, output, sales promotion policies and other tactics to achieve theireconomic goals. Thus, given the structure of the market, market conduct

determines the outcome. The result of market behaviour of market agents infact resembles market performance.Market PerformanceThe economic result of inarket structure and i-narkct conduct representsrnarltet perfor~nance. Market performance resembles price level, profit margin,level of investment, reinvestment of profit etc. In an economy, if the pricefixed by the fir111 is just equal to average cost (the condition in perfectcompetition)? the market is said to be performing well or efficient or is called awell integrated one. Similarly, a less profit margin, norrnally indicates anefficient inarket performance. In other words, through the level of prices, thelevel of profit margin etc., one can determine the degree of market integration.Therefore, market perforinance has a direct bearing on market integration.In the view of Bain (1968), market performance deals with theeconomic results that flow-from the system in terins of its pricing efficiency,its flexibility to adopt to new changing situation etc. It represents theeconomic results of the structure and conduct.According to Narver and Savitt (197 l), marketing performance was thenet result of the conduct and was measured in terms of net profits, rate ofreturn on owner's equity, efficiency with which plant, equipment and otherresources were used and so on. Stifel's (1976) analysis of inarket performanceis in relation to its structural conditions and conduct with regard to pricing andproduct policies.

From the above observations it can be maintained that inaritetperformance is the combined result of market structure and rnarlcet conduct.Marketing performance has several connotations. As pointed out byNarasilnhain (1994), to study the extent of competition, in marketing acommodity, marlcet performance approach seems to be more appropriate. Inother words, one can say that marketing performance actually percolatesmarketing efficiency. To Shrivastava (1996), if the structure, conduct andpcrtbrinance of the inarketing system bears a proof about the efficiency, it willpercolate in the form of greater income, saving, capital formation andinvestment. In this context it is pertinent to understand the concept ofinarlceting efficiency.Marketing EfficiencyMarlceting efficiency is considered to be a pre-requisite for promptdelivery of goods. Proinpt delivery of good at a reasonable price is possibleonly if the market works in a competitive way. Competitive mechanism ispossible only when the lnarlcet agents are free to exercise their actions. Anefficient marketing system implies that price spread or inarlceting margin isfairly less. In market integration terminology, prices in spatially separatedmarkets will get differed only by transaction costs among markets. Lowerprice spread also implies that both consuiners and producers are gaining fromaffordable price and reasonable profit. Hence, an efficient marketing systemiinplies the existence of market integration.54

Experts have viewed the concept of marketing efficiency in differentttajs. A brief look at the views can be presented under three heads. They are(i) Maxiinization of input output ratio as a resemblance of marketingefficiency. (ii) Competition or effective market structure as an indicator ofrnarketing efficiency and (iii) Lower price spread or marketing margin as acondition of marltetiiig efficiency. The examination of these approaches arepresented below:(i) Maximization of input output ratio as a resemblance of marketingefficiencyKohls' (1967) analysis was on the basis of optimizing behaviour ofeco~io~nic agents. It is the inaxiinisation of input-output ratio, output beingconsumer's satisfaction and input as labour, capital and management thatinarlceting fir~ns employed in the productive process.(ii) Competition or effective market structure as an indicator ofmarketing efficiencyAccording to CIarlc (1954) the three components of effectiveness, costand their effect on performance on marketing functions and services which inturn affect production and consuinption constitute inarketing efficiency.Jasdanwalla (1966) opined that marketing efficiency signifies the effectivenessor competence with which market structure performs its designated functions.

(iii) Lower price spread or marketing margin as a condition of marketingefficiency.The higher the price spread. the greater the inefficiency in theinarlceting system and a minilnuin price spread denotes an efficient marketingsystem. One can consider a market~ng system efficient if it performs thefollowing functions - observes Singh, et. a1 (198'7)An adequate marketable surplus to be ensured.0 Prevalence of lower price spread.0 Accessibil~ty of agricultural inputs to be ensured to farmers at areasonable price.On the whole, there is no unanimity of opinion on the concept ofinarlceting efficiency. Some are giving emphasis to raise output by loweringinput. Here no specific analysis of price structure is made. In the second view,importance is given to elimination of wasteful marketing costs or coinpetenceof market structure. As per the third view, price spread is considered as anindicator of marketing efficiency and it is more realistic one. A regulatedinarket with low marketing costs and marketing inargin is said to be anefficient one. Marketing efficiency or the integrated marketing systein alsodepends on inarket structure, the nature of colnxnodity and the socio-politicalsystem. Price stability can also be considered as an indicator of efficientinarket system. Hence, it can be cited that there are several factors thatdetermine marlceting efficiency.

Determinants of Marketing EfficiencyEconomic efficiency and technical efficiency are the two determinantsof marketing efficiency. They are explained below:(a)Pricing, Trading or Economic Efficiency:Usually economic efficiency is a matter to be considered to enhance theconditions for competition and pricing of commodity in a market. Chahal andGill (1991) observes that pricing or econoinic efficiency either relates tofunctional deficiencies or to the degree of coinpetitioil or inonopoly and toeconoinic structure existing within the tnarlteting system. To them in anefficiently operating market, prices will be related in the following manner.(i)Prices should only differ (due to transportation costs) betweengeographic areas of a country,(ii)The price of storable coininodity at one point in time should notexceed price in a previous period of time by more than the cost ofstorage plus normal profit, and(iii)The price of the processed products, should only exceed the price ofunprocessed product by processing costs plus normal profit.According to Lipsey and Harbury (1992) economic efficiency has twocomponents. They are: (i) Productive efficiency, and (ii) Allocative efficiency.Productive efficiency is a situation when it is not possible to produce more ofany one good without producing less of any other good. Allocative efficiencyinvolves choosing between productively efficient bundles. Resources are said

to be allocatively efficient &hen it 1s not possible to produce a combination ofgoods different froin that currently being produced. which will allow any oneperson to be made better off without making at least one other person worseoff.Thus, as the term denotes it concerns matters related to trad~ng orpricing so as to enrich the degree of competition. When there is enrichment inthe degree of competition, the possibility of price spread will be lower. Lowerprice spread ensures remunerative and affordable prices to various economxagents. Hence, effective nieasures of pricing efficiency ensures an efficientinarltet system.{b)Operational, Technical or Organisational Efficiency:The emphasis of operational efficiency is on the cost of marketinginputs by keeping the cost of physical operations to the least possible. Brunk(1950) held that one of the primary purposes of marketing research is to findways of increasing efficiency in the physical handling and processing of good.Lau and Yotopoulos (1971) defined technical efficiency as "a firm isconsidered more technically efficient than another if, given the same quantityof measurable inputs, it consistently produces a larger output". To quoteHenderson and Quandt (1971), the production function differs from thetechnology in that it presupposes technical efficiency and states the maximumoutput attainable from every possible input combination. The best utilizationof every particular input combination is a technical, not an econoinic problem.

All these definitions are unanimous in pointing out that a technicallyefficient systeln should ensure least cost combination. And an ideal marketings~steln emanates from optimuln marketing efficiency resulting fromoperational and economic efficiency. Hence. a market through econolnical andorganizational efficiency tries to function effectively. If the organsiational andpricing structure sinoothens free flow of market information it will lead to anintegrated market. Hence, marketing efficiency is concerned withenhancement of utility with the most efficient utilization of scarce resourcesavailable in the rnarlteting system; which is the basic principle of economics.Measurement of Marketing Efficiency - CriteriaUsually in the literature there are two criteria that can be used tomeasure marketing efficiency. One is price spread and the other is marketintegration.Price Spread: A product has to pass through several distribution channels soas to reach to the consumer. Therefore, it is natural that every distributionchannel require a fair share. Longer the channel greater will be the share ofthese intermediaries in the consumer's price. Price spread is denoted as thedifference between the price received by the producer and the price paid bythe consuiners for a coinmodity at a point of time. Lesser the difference; moreefficient is the market system. If the intermediaries charge just the normaltransaction costs, consumers in the central and peripheral markets can get the

article allnost at the same prtce. If this is realizable. ~t IS a situation of efficientmarketing system or it characterizes an integrated market.According to Dhondyal (1989) price spread simply compares the totalvalue of the product that comes in the bacltdoor of the business with the totalvalue of that which goes out of the front door. Bq distinguishing price spreadfroin ~narlteting inargin, Dhondyal (1989) states that price spread can bewithin the same city but the marketing margin is a wider term which is usedfor various levels of outstation market also. The concept of price spread wasconceptualized by George (1972) as the difference between the retail price ofproduct and its value in production. The cost incurred and the profit gained byintermediaries are generally included i.e; charges for assembling, processing,storing, transporting, wholesaling and retailing. These definitions tries to giveemphasis on the difference between what producers are able to get and whatconsumers are bound to pay. Hence, they are unaniinous in portraying pricespread as the charges spread among intermediaries.Soine of the statistical techniques used to measure the magnitude ofprice spread can be discussed below: The inetliod followed by Hays and McCoy ( 1 978) can be explained asPP,] = P, -(KC,, +TC,, +AS,,)wherePP,, - parity price of one unit in the i th market in relation to j thmarket,

P, - the actual retail price of one unit of the article at the i th market,E3C ,, -handling costs involved in moving one nit from the j th to the iTC,, -transport cost for moving one unit froin j th to the i th market,andAS,, - assembler's charge in moving one unit from the j th to the i thmarltet.Now the actual price spread between any two markets would bePSI, = PP,, - l',wherePS,,- price spread for one unit between i th and j th marltet, andP, - the actual retail price of one unit in j th maricet.I11 a perfectly competitive market, where the product is moving fromthe j th to the i th market, PP, would always be equal to P, and therefore, pricespread would be zero. A positive price spread would provide an opportunityfor traders to make abnormal profits.The method followed by Hays and McCoy (1978) is simple incalculation. Almost all the intermediary charges are included in thecalculatioi~. Without deviating much froin the above method, another way ofcalculation was used by Nailc and Arora (1986). Concurrent method was usedby Naik and Arora to compute the price spread. Before proceeding to computethe price spread, the following percentage share has to be obtained.61

PRPSRP, = 2 x I00R"PSRP, = Percentage share in retail price retained by the i th intermediary.PR,= Price retained by the i th intermediary, andK, = Retail price per unitLPSCO, = ---I- x 100R,PSCO, =percentage share in retail price incurred as cost by the i thintermediary, andc, - the cost incurred by the i th intermediary per unit.Now the percentage share in retail price retained as net margin(PSNM,) or price spreadPSNM, = PSRP, - PSCO,Lower values of PSNM, indicates higher inarketing efficiency and viceversa.The method suggested by Hays and Mc Coy (1978) is an allellcoinpassingone than by Naik and Arora (1986). In Hays' method whilecomputing price spread all sorts of transaction cost has taken intoconsideration. It indicates the actual share retained by the intermediaries afterproviding necessary allowances. However, both techniques assert that lowerprice spread indicates greater marketing efficiency. A zero price spread is the

optirnurn level in attaining highest inarlieting efficiency. But this is only atheoretical possibility which can be attained in a perfectly competitive market.Market IntegrationBefore analyzing the concept of market integration; let us know whatthe notion of integration is? To integrate means unify into a whole. Theeconomic ilnplication of integration is that an element of efficiency isattainable in the unified operation than in the independent actions.According to Mc Donald (1953) the integrated econoiny is one inwhich separated econolnic process is so functionally related to every otherprocess that the totality of separate operation form a single unit of productionwith characteristics of its own. Mc Donald (1953) puts some of themanifestations of integration as(a)Many diverse, specialized and independent econornic processes oroperations, none of which is complete or self sufficient.(b)A system of relationship between the various processes whichserves to register this interdependence upon the conduct of eachprocess so that all are caused, in some manner to fall under theoverall plan.(c)A concatenation of processes in unified pursuance of the aims andpurposes of the larger scheme of things.(d)A mutual replenishment to spent resources to the end that thecontinuity of each and all processes shall not be jeopardized.63

lie-allocation of productive resources is the integral part of integration.'The idea behind integration is that an efficient management of the overallindustry or to say the economy for the well-being or betterment of society.Having dealt the concept of marliet and integration we can proceed toltnow the concept of market integration and its relevance in economics.Marltet integration is considered to be a useful parameter to measuremarketing efficiency for temporal and spatial analysis.I3orowitz (1981) maintains that it is cominon in econolnics to definemarket integration on the basis of price determination. Relevance of theconcept of market integration will be clear if one loolts at the view of Dercon(1995). "Marltet integration analysis can assess the transmission speed of pricechanges in the main inarlcet to the peripheral markets. A reduction in the timelag of transinitting price signals suggests better arbitrage and therefore animprovement in the functioning of markets".Marltet integration is the process by which price interdependenceoccurs. To Faminow and Benson (1990) the usual definition in the literature isthat integrated markets are those where prices are determinedinterdependently; which is assumed to mean that price change in one rnarlcetwill be fblly passed on the others.Goodwin and Schroeder (1991) ca~itions that inarltets that are notintegrated inay convey inaccurate price inforination that inight distortproducer marketing decision and contribute to inefficient product movements.

~ctually what inarket integration delii ers to the econon~y: 1% ill be explicitfrom the following views. Information on market integration provides specificevidence as to the competitiveness of the market. the effectiveness of arbitrage(Carter and Halnilton. 1989) and the efficiency of pricing (Buccola, 1983).Delgado (1986) opined that a well integrated market system is essential tohousehold food security in both food deficit rural areas and those witnessing arise in the relative importance of non-hod cash cropping. To know theworlting of market, an understanding of inarket integration measurement willbe useful.Monke and Petzel (1984) defined integrated market as markets inwhich prices of differentiated products do not behave independently. Spatialmarket integration refers to a situation in which prices of a commodity inspatially separated markets move together and price signals and informationare transmitted smoothly across the markets. Spatial market performance maybe evaluated in terms of the relationship between the prices of spatiallyseparated inarltets and spatial price behaviour in regional inarltets may be usedas a measure of overall lnarltet performance (Ghosh, 2000).Behura and Pradhan (1998) defined inarket integration as a situation inwhich arbitrage causes prices in different markets to move together. Morespecifically two markets inay be said to be spatially integrated; when eventrade takes place between them, if the price differential for a homogeneous

commodity equals the transfer costs involved in moving that co~nmoditybetwecn them.An equilibriuln will have the property that. if trade takes place at allbetween any two regions, then price in the importing region equals price in theexporting region plus the unit transport cost incurred by moving between thetwo. If this holds then the markets can be said to be spatially integrated -observes Ravallion (1986).According to Slade (1986) two trading regions are integrated if pricechanges in one region cause price changes in the other. The transmissionrnechanisrn could be that price increases in one region result the productmoving into that region from the other, hence reducing the supply of productin the exporting region and causing price to increase.Hence, an interrelated or interdependent movement of prices betweenspatially separated ~narlcet can be said to be a situation of marlcet integration.Several statistical techniques were einployed to test the nature of marketintegration. Since we are concerned with the testing of marltet integrationhypothesis, it is obvious to review all these available techniques.Techniques to test Market Integration HypothesisMany empirical techniques have been developed and employed toinvestigate the relationship that exists across space and time. It is from theseresults drives the conclusion about the magnitude of competition or integrationor marketing efficiency that exists in a marketing network. Let us review the66

techniques einployed over the years in the area of market integration researchof agricultural products. Some of the important techniques are:(i)(ii)(iii)Price Series Correlation.Variance Component Approach.Ordinary Least Square Framework.(a) Ordinary Least Square method.(b) Autoregressive Model.(c) Koyck's Distributed Lag Model.(d) Ravallion Model.(iv)Cointegration Technique:(a)Stationarity and unit root tests- Dicltey - Fuller Test- Augmented Dicltey - Fuller Test- Phillips - Perron Test(b)(c)Engle-Granger Model of CointegrationError Correction Model.(v)Parity Bound Model.(i) Price Series CorrelationThe degree of association of price formation in one market with theother can be shown through a zero order correlation matrix of prices in thesemarkets. The system assumes that with random price behaviour expected of anon-integrated market, i.e., bivariate correlation coefficient will tend to zero.On the other, in a perfectly integrated market, correlation coefficient is

expected to be unity. Correlation coefficient can be estimated by the followingforinula:wherer = correlation coefficient,PI, = price of the cornrnodity in the first tnarltet at i th point of time,P2, = price of the coininodity in the second market at i th point oftime,-.P, = mean price in the first marltet, and-P, = iliean price in the second market.Correlation coefficient is considered to be a convenient measure ofmarket integration on two counts - price data is the only required data and iseasily accessible and calculation is simple. This technique is based on theassumption that if markets are perfectly competitive and spatially wellintegrated price differences among markets will reflect only processing andtransportation costs; and correlation coefficient will be equal to one.Accordingly, higher correlation coefficient implies that the inarkets are well orstrongly integrated; and a lower coefficient specifies a weak form of marketintegration signifying lack of market information, transport bottlenecks, lackof product homogeneity or an element of monopoly power.

I-lo~%~ever. an arraj of cr~ticism has started in using correlationcoefficient as a measure of market integration. Blyn (1973) pointed out thatbecause of corninon trends there may be an upward bias to the results. Blynfurther states that the trend may be due to rising demand occasioned bypopulation increase that inay affect all parts of the region or due to commonclimatic condition. Here all price series in a region would be affected by suchinfluences even if each market within the region was independent of others.Blyn therefore cautions that time series correlation need to be restricted toresiduals remaining after the trend and seasonal coinponents have removed.Price series correlation method has also been criticised by Harris(1979) on the ground that a high correlation between the markets does notnecessarily mean that these two markets are well integrated in tlze sense that acompetitive network of traders exists which ensures that agricultural goodsmove between market places in swift response to price difference that exceedtransport cost. Lundahl and Petersson (1 982) have also cited their criticisinalmost on the same line with Blyn and Harris.Problems in using correlation coefficient were also earinarked byHeytens (1986). Heytens observes that though prices in an efficient marketsystein tend to move together, they may do so for other reasons (generalinflation, coininon seasonality) or other coininon factors may producesympathetic but unrelated price changes. It is further maintained that a perfectmonopoly or price fixing by a central authority can just easily produce a

coefficient of one as a perfectly competitive market. Therefore correlationcoefficients are not unequal indicator of inarket conditions and applicationsbecome inore indiscriminate. Petzel and Monke (1979-80) also assures theabove observation. Harris (1979) and Tim~ner (1974) too pointed out thatmarlcets inay be spatially integrated, but demonstrate low price correlationbecause of changes in the geographical direction of price formation.(ii)Variance Component ApproachThe technique developed by Delgado (1986) is to test time series ofprices for seasonal differences in the price integration of markets. Theapproach is to decoinpose the variance of prices into compoiients. The modelof price for a crop can be written asWherePIIS' = 111'~) + v,'~' + u IS' + z j;)Super subscripts S = 1 . . . . Nuinber of seasons,1tPi,- 1 .... Nuinber of markets,= 1 , . . . Number of weeks in season,= price of the article concerned in market 'i' in week 't',in = the meail price of each season,Vi = a constant location (village) effect,U, = a constant weekly time effect, andZit = a stochastic interaction term.

T~2.o important assu~nptions of a season are :(i)(ii)Variance of prices for a given crop is constant over the season.transport and transaction costs for marketing a given crop between twomarkets are constant subject to a random disturbance over the season.Analysis is done for each crop and season. Variations around meanprice is divided into two constant deviations and a stochastic tenn. Theyardstick of integration is that the price spread between markets staysapproxi~nately constant, subject to randoin variations either way. Equation (1)shows that after removing a common seasonal trend (U,) and seasonal meanprice for each village (m+V,) interaction between the residual price term (Z,,)across villages are independent.More formallyE (Z,t 21,) = 0 - (2) where i ;t 1If equation (2) can be shown to hold jointly with a reasonable degree ofstatistical confidence for all pairs of markets, the system of markets is judgedintegrated.Equation (2) can be tested by estimating the Z elements from the pricedata. For it one has to remove the long term trends and constant effectspeculiar to a particular location by subtracting mean seasonal price for eachvillage and crop from weekly price data.-Thus P,: = P,, - Pi - (3)

Where the notation is consistent with equation ( 1) and the mean pricefor village 'i' is calculated separately for each crop and seasonHence P,; = Lrt - ZIt - (4)The next step in retrieving the stochastic Z,, is to eliminate the weeklytime effect Ut, which is conceptually constant across villages but different foreach week. Netting this effect out removes the coinmon seasonal trend andthis eliminates spurious correlation of price lnoveinents arising from seasonalinfluences. The variance component method perinits statistical inference froma sample of time series of market prices concerning seasonal and regionaldifferences in the variance of prices.The major limitations of this inodel are (i) it assumes constant varianceof price over the season (ii) transaction cost between two inarltets are alsoassumed to be constant. Once these restrictions are relaxed, the inodel may notbe able to measure the exact degree of rnarlcet integration.(iii) Ordinary Least Square Framework(a)Ordinary Least Square MethodSeveral researchers have tested integration of agricultural coinmoditymarkets with the OLS method and it is presented below:wherePlt - price at location i at time t,Pjt = price at different location at time t, and

T,,, = indicator of transportation and trarlsaction costs betweenlocation i and j at time t.In order to obtain a linear equation. log of equation (1) is taken forestimation. The estimation is focusing on PI as the "elasticity of pricetransmission". If pl = I the market is said to be integrated.The above method has a few limitations. No serious attention hashowever, been given to the properties of the error term. Unbiasedness requiresthat the error term has no discernible structure, otherwise the price of centralmarket can not be said to possess all inarltet information and the past historyof peripheral ~liarket price. Fusther, as a matter of fact, the notion of non-stationary pi, and pi, raises doubt about the consistency of the estilnatioil of PI.(b) Autoregressive ModelThe autoregressive model which was employed by Heytens (1986) totest market integration will be explained below.a,(L)P,, =P, (L)F,+Y (L)X + U,,- 1- lt(1)where+ 1 ........ kandt=l ......... n.P,, = price in market i at time t,-P, = reference price at t,X = vector of seasonal and other relevant variables in market i attime t with the same collection of variables used in all vectors,X it, overall markets and all time period,

1 J,, = an error t erin andu, (L), PI (L) and y (L) denote the polynomials.- 1a,(L)=I- cxl,(Ia)- ................ - qnLnPI (L)= p,,, +PI, L+ ........ ep,,l, L"'y, (L) = y,, +yL+ ................. + Y I,, Ll'1 IFor the empirical analysis equation (1) will be rewritten with firstdifference of local price on the dependent variabie.-where A PI, =I",, -P ,,-, and A' = P,, - P,Where ai, = 1. For siinplicity equation (2) can be written for one lageach for local and reference market.removing A, equation (3) becomesPi, -PI,-, =(a, -l)(Pit -?,-I)+Pio (Ft -Pt-,) +(ai +pio +Pi\ -1) FL-1 +?' X+Uit-1(4)

Equation (4) specifies the changes in local price as a function of thechange in the reference price for the same period, last period's spatial pricemargin, last period's reference market price and local market characteristics.p,,, - measures the extent to which local market participants know theinarlcet conditions of reference marltet.a,-[ - rneasures the extent to which last period's spatial pricedifferential is reflected in this period's local market price.Here inarltet 'i' could be called segmented ifPI* = Pi, = 0 (5)Which can be determined by testing equation (4) against the followingrestricted inodel with an F testAcceptance of equation (6) indicates that the price in inarltet i dependsonly on its own lagged values and local market characteristics.Now if P, =I$, (L)=l (3 Pi, = 0) (7)And a1 = 0 - (8) , then inarltet 'i' is integrated with the referencemarket in one time period.When n=l, market integration as indicated by equation (7) and equation(8) implies the absence of local price autocorrelation.Heytens maintains that some problems are obvious in the model.Detennination of appropriate reference prices and variable specification will

e a mattes of concern where a broad understanding of the market is limited.There will be the existence of simultaneous equation bias. The model'sparameters are likely to be sensitive to the time length of data. Though themodel can handle problem raised by common time trend, it cannot deal thesituation when direction of comlnodity flow between rural and urban areasreverses with the season.(e) Koyck's Distributed Lag ModelWhen the regression model includes not only the current but the laggedvalues of the explanatory variables, it is called a distributed lag model. Moyckhas proposed an ingenious method of estimating distributed lag models(Madnani, 1986).Pit = a+poP,t +p, P ,t-, + .......,,.....*...., + P, p,,-k +u, (1)wherePi, - the price of the i th product in period t,Pi, - the price of j th product i11 period t, anda and p are parameters. Assuming that the P 's are all of thesame sign, Koyck assumes that they decline geometrically as followsP, = P, ,IK (2) k= 0, I....Wliereh, such that 0 < 1 < 1 is known as the rate of decline or decay of thedistributed lag and (1 - 1) is the speed of ad-justment.

Equation (2) explains that each successive P. is numerically less thaneach preceding P implying that as one goes back into distant past the effect oflag on Pit becomes progressively smaller. By assuming non-negative values forA, Koyck rules out the p's from changing sign and by assuming A < 1 he giveslesser weight to the distant P's than the current one and ensures that the sum ofp's gives the long run inultiplier in a finite amount nainelyAs a result of equation (2); equation (1) can be written asPi, =@+POPjt +J3, hP ,,-, + p, h 'P,,-~ + .................... +lJt (4)As still the model is not amenable to easy estiination due to large number ofparameters, Koyck lags equation (4) by one period.Pi,-, =a+P,P,,-, +p, hPi,-? + 0, A'P,~-~ + .................... +Ute, (5)Multiplying equation (5) by 3LhP,+, =ha+p, hP,,-, +PI h' P,,-, + P, h' Pi,-, + ..................... hut-, (6)Subtracting equation (6) from equation (4)Pit -hPi,,= a(1-h)+p,P,, + (U, -hU,-,)RearrangingPi, =a(l-h)+p,P, + hPi,-, +V,whereV, = (U, -hut,)

Positive signs are expected forand h for market integration inequation (7).Multicollinearity is resolved by replacing Pi,l. Pit-2 by a single variablePit- I -Here we have started wit11 a distributed lag model but ended up with anautoregressive model. The presence of lagged explanatory variable violatesDurbin-watson 'd' test. Therefore, one have to test the serial correlation byDurbin-watson 'h' test.The P gives the short-run price adjustment corresponding to a unitchange in j th price.Long run adjustment is measured through equation (3).That is P, =Po -) . The error term V, possess OLS properties.[ lcIn Koyck's transformed model the presence of lagged dependentvariable raises some problems.In the new formulation the error term V, = ( U, - h U,-, ) is found to beautocorrelated.E(Vt V,I)= E(U,- hut-I) (Ut-I - hut-2)= E (Ut UbI - Ut h Utm2 - 1 u2t-l + h2 Ut-I Ut-2)

The lagged variable Pi,., is also not independent of the error term V, i.e.E (V, V ) 0 This is because Pit directly depends on V,. Similarly Pi,.i onVtTI. Rut since V, and V,., are not independent, Pi,., will obviously be relatedto V,.Due to these two problems, ICoyck's distributed lag inodel give rise tobiased and inconsistent estimates. Again it assuines that the impact of pastperiods decline successively in a specific way. But in reality this may not bethe case.(d) Ravallion ModelRavallion (1986) developed an econometric inodel of spatial pricedifferentials. It is assumed that there are a number of local markets and acentral markct. The pattern of price formation ainong N markets, whereinarlcet 1 is the central ~narlcet is summarized by the modelPI = F1 (P2, P3, ........... PN, XI) (1)and Pi = Fi (Pi, X,) (2)where i = 2 ............... NXi (i = 1 ..... N) is a vector of other influences on local markets.By incorporating a dynamic structure to equation (1) and (2),econometric model of T period series of prices for N regions is assumed

where Pit = the price of central tnarket andPi, = the price of peripheral market.Ravallion used equation (3) to test several hypothesis.Market is segmented if bii = 0 (4)Short run inarlcet integration is possible If bi,, = 1 (5)For lagged effects aii = bij = 0 {,6)If (5) and (6) are accepted, then one can say that market 'i' is integrated withthe central xnarltet with one time period.A weak form of inarket integration will also be tested in which thelagged effects vanish on an average.takes whereFor long run market integration consider the forin that equation (3)Pit = Pi* , P,, = P; and e,, = 0 for all t; thenMarket integration now requires thatFor long run integration equation (3) was reestimated in the following fonn

pi+ =(a,, -l)(P ,,-, -PI,- )+x a,, (P,: -P )[-, 1I-!+z+b,,Ap,, (bto-lit a,, +b,, )AP,?- +X,, C, -e,,I=Ih=!The Ravallion model extracts more inforination on the nature of spatialprice differentials. This model avoids the inferential dangers in using spatialprice correlation. It permits price series for each local market to have its ownautoregressive structure and a dynamic relationship with market prices in atrading region. Ravallion's dynamic approach permits a clear distinctionbetween short run market integration and integration as a long run tendency inthe short run adjustment process.However, the Ravallion model is beset with several problems. Palaskasand White (1993) observes that even if the correct estimation procedure isadopted, the coefficient estimates of tlie stochastic equation can be impreciseif the dynamics are of a relatively high order, the reason being~nulticollinearity between lagged values of the explanatory variables. Againspecification in levels raises the problem of spurious correlation associatedwith the regression of trending variables in levels. Baulch (1997) maintainsthat Ravallion model is based on assessing the co-movement of price dataalone and fail to recognize the pivotal role played by transfer cost.(iv) Cointegration Technique:Cointegration can be regarded as the empirical counterpart of thetheoretical notion of long run equilibrium relationship. The development of

cointegration technique form a fbrlnidable achieve~nerlt of time serieseconometrics in the 1980s. Cointegration analysis has been necessitated by theearlier approach which generally ignored or misrepresented the time seriesproperties of the price series and hence. serious flaws in the estimationprocedure. As a matter of fact, several macro economic time series exhibitstrend like behaviour. Granger (1966) expressed this as the series having muchof their spectral power at low frequencies and Nelson and PIosser (1982)argued that this persistence was captured by modelling the series as having aunit auto regression root (being integrated of order one). Stock (1999)maintains that "the achievement of cointegration analysis, as developed byGranger (1986), Ganger and Weiss (1983) and Engle and Granger (1987) wasto provide a unified fraineworl< in which to understand and to reconcile theapparent conflict between spurious regressions and economically meaningfullong term relations".Cointegration technique is a three-stage procedure. Firstly, variableshave to be pre-tested for stationarity. A series is said to be integrated of order'd', I(d), if it has to be differenced 'd' times to produce a stationary series.Once stationarity is obtained, variables are to be tested for cointegration orlong run relationship. Two series are cointegrated of order (1,1), if theindividual series are I(1) and a linear combination of thein called thecointegrating regression is I(o). After getting cointegrated relationship, theresiduals from the equilibrium regression can be used to estimate the error

correction model. Thus, it can be shown that in the case when tIva series areI(1) and are cointegrated. the inodel can be given an error correctionrepresentation.(a) Stationarity and Unit root tests:To develop models for time series, it is important to know whether ornot the underlying stochastic process that generated the series can be assumedto be invariant with respect to time. If the characteristics change over tiine.that is; if the process is non stationary, it will be difficult to represent the timeseries over past and future intervals of time by a siinple algebraic model. Onthe other, if the stochastic process is fixed in time, that is; if it is stationary,then one can model the process via an equation with fixed coefficients that canbe estimated from the past data.The presence of unit roots in time series points toward non-stationarityof the series. Regression will be spurious if both independent and dependentvariables show the presence of unit root. To have a compatible model,variables should be of same order sf integration. Unit root test starts with thelevel series, takes the difference and tests for the presence of unit roots byregressing in the first difference on lagged variable of the series. If oneobserves the presence of unit root, the series is said to be non-stationary. Nowthe exercise is to be repeated by taking the second difference and so on untilthe series become stationary. Some of the important tests used to check

stationarity are Dicltey-Fuller test, Augmented Dickey-Fuller test and Phillips- Perron test.Dickey-Fuller Test (DF)To test whether the series Y, is stationary, the test have been providedby Dickey and Fuller (1979, 198 1) and it is presented below:AYt = a, + p Y,,, + e,r 71 est result reveals that Yt is stationary if p < 1, non-stationary if p = 1,and non-stationary and explosive if p > 1.If 'p' is negative and statistically significant, the alternative hypothesisthat Yt is integrated of order I(1) is accepted. Dickey and Fuller derivedcritical values for the test from Monte Carlo experiments and is given as 't'statistics.One of the major flaws of DF test is that the problem of serialcorrelation is endemic. It is also cited that autoregressive or moving averageerrors have a big effect on the power of DF test.Augmented Dickey Fuller Test (ADF)It includes additional lags to inop up serial correlation. It alsoincorporates additional nuisance parameters. Further, data based selection oflag length can be used with little adverse effect. ADF is augmenting theregression equation of DF by adding sufficient terms in AYt.I and it ispresented below:

Whcre k is selected to be large enough to ensure the error e, as a white noise.Interpretation of ADF results are same as that of DF test.One of the serious defects of ADF test is that too rnany lags reduce thepower of the test to reject the null of a unit root since the increased nuinber oflags necessitates the estimation of additional parameters and a loss of degreesof freedom. The degrees of freedoin decreases since the nuinber of parametersestimated has ~ncreased and because tlie number of observation has decreaseddue to additional lags.It is important to note that the Dickey-Fuller test assumes that the errorsare independent and have a constant variance. This raises some problems.Firstly, the true data generating process may contain both autoregressive andinoving average components. Secondly, one call not properly estimate 'p' andits standard error ullless all the autoregressive terins are included in theestimating equation. The third problem steins from the fact that Dickey-Fullertest considers oi~ly a single unit root.Now, by relaxing the assulnptions of Dickey-Fuller test, a newinethodology was developed by Phillips and Perron for testing stationarity ofdata series.

Phillips-Perron TestThe distribution theory supporting the Dickey-Fuller test assumes thatthe errors are statistically independent and have a constant variance. Whileusing this methodology utmost care is to be talten to ensure that error terms areuncorrelated and have a constant variance. Phillips and Perron (1988)developed a generalization of the Dickey-Fuller procedure that allows forfairly mild assu~nption concerning the distribution of errors.Consider the following regression equation.Yt = ao* + a,* Y,., + pt andY, =a, +a, y, -1 +a2 (t -T/2)+p,where 'I' = number of observation and the disturbance term p, is such thatEp, = 0, but there is no requirement that the disturbance term is seriallyuncorrelated or hornogeneous. Instead of Dicltey-Fuller assu~nption ofindependence and homogeneity, the Phillips-Perron test allows the disturbanceterm to be wealtly dependent and heterogeneously distributed.Phillips and Perron characterize the distribution and derive teststatistics that can be used to test hypothesis about the coefficients a,* and ii,under the null hypothesis that the data are generated byyt =y*4 + ptIf the coefficients are negative and statistically significant, tlie series is said tobe stationary. The critical values of the Phillips-Perron statistics are preciselythose given for the Dickey-Fuller test.86

Monte Carlo studies find that the Phillips-Perron test has greater powerto reject a false null hypothesis of a unit root. Monte Carlo studies have alsoshown that in the presence of negative moving average terms, Phillips-Perrontest tend to reject the null of a unit root whether or not the actual datagenerating process contains a negative unit root. But in practice, the choice ofthe most appropriate test can be difficult since one never know the true datagenerating process. Enders (1995) observes that a safe choice is to use bothtypes of unit root tests; and if they reinforce each other, one can haveconfidence in the results.(b) Engle-Granger Model of CointegrationLet there exists a constant h such that X, is h Y, is I(o). When thisoccurs, X, and Y, are said to be cointegrated but the variable 2, = X, - h Y, isstationary I(o), h is the cointegrated parameter (Granger 1986).If X, and Yt are I (I), it is necessary that 2, be I(o).When the series are integrated of order one, to estimate the long runrelationship between X, and Y, one has to run the 0I.S regression given belowX, = a + h Y, + 2, and test whether the residuals Z, are stationary. Afterrecovering the residuals, cointegration test can be done in the following way.First the estimated residuals fromX, = a + h Y, t 2, are used to construct a Durbin-watson statistic(OW) and is compared with the critical value given in Engle and Yoo

( 1987). If the estimated CRDW is above the critical value. the null hypothesisof non-cointegration is rejected. Then CRDW test is reinforced byconstructing Dickey-Fuller and Augmented Dickey-Fuller statistics.Dickey-Fuller Test (DF)DF is computed by running the following regression model.Z, =a+bZ,-, +El,where Zt is the residual fioin the cointegrating regression. If the 't'statistic of 'b' coefficient is less than one and statistically significant, theexistence of cointegration between series is accepted.Augmented Dickey-Fuller Test (ADF)ADF test is based on the following regression model.It t-statistic of 'b' coefficient is negative and statistically significant,then the variables are said to be cointegrated.(c) Error Correction ModelGranger (1986) and Engle and Granger (1987) have demonstrated thatif Y and X are both 1(1) variables and cointegrated, an error correctioii inodelexists. The principle behind this model is "there often exists a long-runequilibrium relationship between two economic variables. In the short run,however, there may be disequilibrium. With the error correction mechanism, a

proportional disequilibrium in one period is corrected in the next period"(Ramanathan, 1 995).Error Correction Model i~lcludes last periods' equilibrium error as wellas lagged values of the first difference of each variable. The degree ofdisequilibriuin can be assessed by examining the relative magnitude andstatistical sigr~ificance of the error correction coefficient. Error correction~nvclel coinbines the long term inodel with the short-term dynainics when Y,and X, are cointegrated of order (1,l) the variables have the error correctionThe coefficients a,, a,, a,!, aj2, a2, and a2~ shows the short rundynainics of the system. If both ay, a, are zero, it assumed that there is noerror correction.Limitations of Engle-Granger Cointegration(i)Engle-Granger procedure is a bivariate model which ignore thelinkage that may operate through a third inarltet.(ii)The existence of inore than one long run relatioilship cannot becaptured by co-integration technique.

(iii)Tlie tests conducted for identiQing the driving forces in the marketignore the probability of existing multiple corn~non trends; whichwould imply ~nultiple dominant markets.(v)Parity Bound Model (PBM)Baulch (1997) developed a PBM to test rnarltet integration. Baulchargued that time series techniques involving Granger causality, errorcorrection and cointegration rely on price data alone and fail to recognize therole of transfer costs. These approaches were unable to distinguish integratedfrom independent markets when both were subject to a common, exogenousinflationary process.PBM extends earlier work on stochastic frontier and switchingregression models. Transfer costs (comprising transportation, loading andunloading costs and trader's normal profit) determine the parity bound withinwhich the prices of a homogeneous coininodity in two geographically distinctniarltet can vary independently.PBM assesses the extent of market integration by distinguishing amongthree possible trade regimes: Regime 1, at the parity bound (spatial pricedifferentials equals transfer costs) Regiine 2, inside parity bound (pricedifferential < transfer costs) Regiine 3, outside parity bound (price differentialtransfer costs).Deviations of the inter-market price spread fromextrapolated transfer costs in any period inay be composed into threeco~nponents. The first error term (et) allows transfer costs to vary between9 0

periods. The second error term (u,) captures the extent to which pricedifferentials fall short of the parity bound and the third error term (V,)measures by how inuch price differentials exceed transfer costs.The PBM is specified aswhere Regime 1 isand Regime 3 ishl, hZ = probabilities of regime 1 and 2, .Y, = the absolute value of natural logarithm of the price spread betweeninarkets i and j in period t,Kt = logarithm of nominal transfer cost in period t,

0 e. cr ,,. CT , = standard deviation of three error terms e,, u, and v, , and4 and cp - denotes standard normal density and distribution functions.To obtain probability estimates for the three regimes of the PBM, thelogarithm of this function may be maximized nuinerically with respect to hi,12, 0 c, 0 u, and o , using the David-Fletcher-Powell algorithm.Statistical hypothesis tests for the purpose of market integration can beconducted by testing the null hypothesis that h, + h2 = 1.The PBM allows for inarltet to be integrated in some periods but not inothers. Statistical reliability of the PBM can be assessed with Monte Carloexperilneilts.Baulch also gives the limitations of PBM.(i)Since only contemporaneous spreads are used in its estimation, it ishard for the PBM to take into account the type of lagged priceadjustment postulated by causality and Ravallion models.(ii)Precise estimation of transfer cost is essential. Inaccuracies inestimation of transfer costs will lead to high cr ,, and problelns with theconvergence of the maximum likelihood procedure.(iii)Violatioils of the spatial arbitrage condition indicate lack of marketintegration but they do not pinpoint its causes.

Concluding Remarks:The present chapter tried to examine the theoretical background of thepresent study and various related concepts of inarket integration and inarketingefficiency. Besides. the present chapter also reviewed the various statisticaland econometric tools employed in the earlier literature to verify the validityof market integration hypothesis.A series of techniques right from correlation coefficient, variancecomponent approach, autoregressive model, distributed lag model, Ravallionmodel, Engle-Granger cointegration technique to Parity Bound Model wasexplained in detail. All these techniques in general can be employed to test thevalidity of the inarket integration hypothesis in a bivariate frainework. In ourstudy, pepper is a product dominated in multiple markets. Hence, Johansen'sinultiple cointegration method' is an appropriate method to examine thevalidity of market integration hypothesis pertaining to the pepper inarket ofKerala.For detailed discussio~l on Jolianse~i's Metliodology See Chapter five.

CHAPTER IVTRENDS IN AREA, PRODUCTION,PRODUCTIVITY AND PRICES OF PEPPERSpices Economy - An OutlineSpices have played a crucial role in the history of human civilization.Christopher Morley defined spice as the plural of spouse (George, 1989).According to Webster, spices are specifically any of various aromaticvegetable production (Khan, 1990). In the view of International spices group,"spices are any of the flavoured or aromatic substance of vegetable originobtained from tropical or other plants, coinmonly used as condiinents oremployed for other purposes on account of their fragrance, prcservative ormedicinal qualities". In brief, one can say that spices are agricultural productssignificant for their taste, aroma, flavour and colours in food, beverages, aspreservative, as medicine, as a substance in perfume industry and what not.Since pre-historic times Inan has been using spices to flavour his food(Padinai~abhan, 1976). The significance of spice was put by Ridley (1912) as,"the history of cultivation and use of spices is perhaps the most romantic storyof any vegetable product".The number of items enlisted as spices differs from country to country.Forty-one items were included by the American Spice Trade Association.

Nearly 107 spice varieties are reported by John (1969). Spices Board hasincluded 52 iteins in the list. Bureau of Indian standards notifies 63 spices asgrown in the country.From time immemorial, India is known as 'the land of spices' to theworld. Let us take a historical record of spices of India. The earliest literaryrecord in India can be found in Rig Veda (6000 B.C). Manu, the proponent oflaw (around 4000 F3.C) was awarc of the origin of spices. The epic Rainayanaalso mentions about spices. References about Indian spices were made by theBabylonians and Assyrians (around 3000 BC) and in the Old Testainent (1000B.C) of the Bible (Sivarainan and Peter, 1999). Susruta (500 B.C.) alsomentions about spices in the writings of Ayurvedic texts.Alexandria (Alexander the Great, 332 B.C) was a iilajor trading centrein spices between the East and the Mediterraneans (Purseglove, et. al, 1981).Travellers like Hiuen Tsang (629 AD), Masudi of Baghadad (890-956 AD),Abd~~l Feda (1273-1331 AD), Rashid-Ud-Dina (1300 AD) etc. had madevaluable citatioils regarding the treasure house of Indian spices (George,1989). By fifteenth century, European countries, especially Portugal, Spainand U.K. showed keen interest with the East in spice trade. Thus, froin severalhistorical incidents, it can be observed that spices are the spring of inspirationfor waging war, expeditions, voyages and even for romance.This treasure of spice had lured explorers to Indian shores for centurieswhich became a spectator for the rise and fall of several empires. India got a

wide-range of agro clilnatic conditions suited Sor its varied cultivations. Allkinds of spices can be produced from different parts of Indian soil. That ishow India got a unique position in the world as the largest producer of spice.Some of the important spices grown are black pepper. cardamom, tunneric,ginger, chillies, coriander, cumin, fennel, fenugreek, celery, aniseed, saffron,clove, nutmeg, cinnamon etc. According to the importance in foreign trade andinternal marketing spices such as black pepper, cadamom, ginger, chillies andturmeric are grouped under major spices and the remaining as ininor spices.Major spices and seed spices are export oriented whereas tree spices and otherminor spices are consumed internally (Ravindran and Manoj Kuinar, 200 1).Spices contributed to 1.24 per cent of India's total export earning, andits share in export earnings from agricultural and allied products works out tobe 8.5 per cent during 1999-2000 (Spices Board, 1999-2000). This sectoralone has the potential of earning an annual foreign exchange of over Rs.1000crore by way of exports to the Indian economy (Behera and Indira, 2002). Thissignifies the iinportance one has to attach to the spices economy of India.Since pepper is one of the important spices, it is better to know its origin,cultivation, propagation, harvesting, value added products and marlteting.A Brief History of PepperPepper from piper nigrum is one of the important spices of the world. Awide genus, with over thousand species piper is from the Botanical family ofpiperaceae (Spices Board, 1993). The Greek name Pepperi, the Latin piper and96

the English pepper all derived from the Sansltrit 'pippali', which was the namefor the long pepper (Purseglove, et. al. 198 1). Piper nigruln is a native of theinonsoon forests along Malabar coast of south western India. ChristopherMorley has called pepper as the king of spices and cardamom as the queen ofspices (George, 1989).'Two kinds of pepper were mentioned by Theophrastus in the fourthcentury B.C. Discorides in the first century A.D. mentions black and longpepper. Mention was also made by Roinan writers in the fifth century A.D.(Majeed and Prakash, 2000). Pioneering sea voyages of the 1 51h century wereintended to find sea routes to the sources of pepper and other aromatic objects.Landing of Vasco da Gaina in 1498 in the Malabar coast of India is alandinarlc in the history of spices (Mathew, 2002). This followed intense fightbetween empires to control for Eastern spice producing regions. Referencesabout the use of medicinal properties of pepper were seen in Materia Medicaof Ay~weda which dates back: to 6000 B.C. Pepper thus has historical andgeographical significance besides trade importance.CultivationPepper is mostly herbaceous or woody climbers or shrubs in the tropicsof both hemispheres. It is a perennial climber to 10 rn or more in height. Thepepper plants has 10-20 main adventitious roots from the base of the maturestein which penetrates to a depth of 1-2 in and there is an extensive mass ofsurface feeding roots (Purseglove, 1968). The climbing branch becomes stout97

4-6 c.ln. in diameters at the base with a thick flake - like bark; the internodesare 5-12 c.m. long. Spikes are borne opposite to leaves on the plagiotropicbranches and are 3-15 c.m. long; bearing 50- 150 minute flowers borne in theaxils of ovate fleshy bracts. The fruit is a sessible, globose drupe 4-6 c.m. indiameter, with a pulpy pericarp, borne in spikes 5-15 c.m. long. Each spilteproduce 50-60 single seeded fruits. The unripe fruit is green with the exocarpturning red when ripe and drying black. The seed is 3-4 lnin in diameters witha minute embryo, little endosperm and copious perisperm (Purseglove, 1968).A well-distributed rainfall and high temperature is required for this cropof wet tropics. Generally pepper has been cultivated as far as 20' north andsouth of the equator. A well drained alluvium soil with humus content is idealfor its cultivation. Though a humid climate crop, it can be grown in placeswhere rainfall ranges from 125-200 cm. It can thrive well at a minimuintemperature of 10°C and a maxiinurn of 40°C temperature and upto 1500 mabove sea level. Pepper can be grown in a wide range of soil which is acidic innature with pH 4.5 - 6.0 (Sit, et. al, 2002).Pepper has been cultivated either as a monocrop or mixed crop on livestandards (i.e. live and non live support). In India the land under blaclc pepperis classified as (Sivaraman, et. al, 2002):(i)Coastal areas where pepper is grown in almost every homestead orplot of land.(ii)Slopes and valleys.

(iii)Hills at an elevation of 800 in - 1500 in with shade trees in coffee/tree plantation.(iv)Valleys as a mixed crop in arecanut gardens.Pepper is generally propagated by stern cutting. Over 75 varieties /cultivars are popular ainong farmers. Varieties of pepper are cultivatedaccording to the soil and climatic conditions and be naturally resistant to pestsand disease of the region. Sorne of the varieties that are popular in India arePanniyur, Kariinunda, Icottanadan, Balanltotta, Neelamundi, Narayakodi,Araltulamunda, Kalluvally, Ailnpiriyan etc. (Ravindran and KalIupuracltal,2000).Pepper plantation is beset with several diseases. Phytophthora foot rootdisease is a serious problem in India. Two inajor diseases spreading in Iceralaare stunted disease caused by piper nigruin strain of cucuinber mosaic virus(CMV - P,) and yellow inottle disease caused by piper yellow inottle virus(PY M.V) (Govindan, et. al, 2003).Prime harvesting of pepper is usually done in the third year. Thisperennial plant lasts for about 20 to 25 years. With good yield and underfavourable conditioils it may long for more years. The maturation period ofpepper varies froin 5 to 6 months in Indonesia and 7 to 8 months in India fromflowering to harvest.

Value Added ProductsThe two primary products of piper nigruin that are internationallytraded are black pepper and white pepper. Black pepper is the whole driedfruit; white pepper is the fruit from which the mesocarp has been removed.Some of the other value added products of black pepper produced andexported froin India are pepper powder, pepper oil, pepper oleoresin, piprine,dehydrated green pepper, pepper in brine, pink pepper, frozen green pepperand encapsulated spices. Being a native of western ghats, pepper constitutes animportant ingredieiit of several indigenous inedicines of India.Since time iinineinorial India has been in the international trade ofpepper. Indian pepper called "Malabar black pepper" is popular ininternational markets. The particular grade "Tellicherry Garbled Extra Bold"fetches tlie highest price due to its size and supreme quality. Majority ofIndian exports is to the U.S.A, foriner USSR and East and West Europeancountries. India contributes about 25 per cent to 30 per cent of the worldproduction and Kerala accounts for over 96 per cent of the production in India(George, 1989).Tl~us, it can be observed that ainong spices, pepper is a doininantproduct. It is an important contributor to foreign exchange. On an average, thecontribution of pepper in spices export earning is found to be 41.2 per centduring 1960-61 to 1999-2000 (Table 4.1). On the domestic front, pepper isconsumed throughout the length and breadth of the country. Other than a main

spice in food items; it has several medicinal uses also. Hence, it can beotxerved that pcpper is a universal com~nodity consumed throughout theglobe. Now as part of integrating world trade, there is unanimity abouteconomic liberalisation and globalisation among nations. On this global effortGovernment of India too has initiated several reforrn measures.The process of liberalisation has started in July 1991. The initial focusof structural adjustment was on industry.Later on agricultural tradeliberalisation also became part of liberalisation agenda. Some of the importantmeasures are reinoval of import control on several agricultural products,relaxation of quantitative restriction on imports and exports, bringing a varietyof articles under Open General Licence Policy; delicensing certain imports,slashing of basic in~port duty etc. (Gulati and Kelley, 1999). These relaxationsopened new avenues for international co~npetition for exports. Hence, thesetrade policy reforms exposed Indian agriculture to world markets. Nowbecause of the export exposure of pepper, these trade policy reforins definitelywill have an effect on pepper econoiny.On this background, the present chapter investigates the trend ofdifferent iiidicators of pepper with spccial reference to Kerala state. Besides, italso investigates the ilnpact of recent economic reform on different indicatorsof pepper product of Kerala.

MethodologyRatio and percentage methods were employed to examine the trend inarea, production, productivity and prices of pepper in India and Kerala.Co~npound growth rate of various indicators of pepper are calculated by usingthe following semi-logarithmic regression equation:whereLogyi= a+bT+uy; = indicators of pepper,t = time trend, andu = randoin error termC.G.R. = [ Antilog b - 1 ] x 100.Multiple linear regression analysis were carried out to examine theimpact of economic reforins on area, production and prices of pepper and it ispresented below:whereYi=a0+pi t+PzD+u,aiidYi=a,+ PI t+P2D+P3Dt+uYi = different indicators of pepper,t = time trend,D = duinmy variable, i.e. '0' for pre-reform and ' 1' for post-reform period,Dt = dummy x trend, andu = random error terin.

Empirical Results and DiscussionFor the sake of si~nplicity and clarity the empirical results are classilledunder the following heads:(i)(ii)(iii)(iv)(v)Trends in Countrywise area, production and yield of pepperTrends in Countrywise export of pepperTrends in Statewise area, production and productivity of pepperTrends in Districtwise area, production and yield of pepperTrends in Price of pepper assembling markets of Kerala.(i) Trends in Country wise Area, Production and Yield of Pepper:Some of the major producers of pepper in the world are India,Indonesia, Brazil, Malaysia, Vietnam and China. Minor contribution are alsomade by Sri Lanka, Thailand, Madagascar and Mexico. In this section specialemphasis will be given to India's acreage and production of pepper. Table 4.2accounts for a country wise analysis of acreage, production and productivityfor the year 1973-2003. India's percentage share to world acreage and worldproduction of pepper for the same period is presented in Table 4.3.India has earmarked the inaxilnuln area of land for pepper cultivation inthe world during the period of our study. In 1973-74 the area of peppercultivation in India was 121 thousand hectares which comes to 61.5 per cent ofthe world (Table 4.2 & 4.3) and it went up to 173 thousands hectares in 1990-91. Further, it can be observed that there is only a marginal increase in the areaof pepper cultivation after 1990-91. It implies that after economic reform,

there isn't any significant change in Indian acreage of pepper. This may bebecause of the mixed cropping pattern and the perennial feature of peppercrop.The compound growth rate of area of pepper cultivation during prereforin(2.5 per cent) is greater than the post reform (1.7 per cent) period(Table 4.9).The estimated lnultiple linear regression equation to identify the impactof econoinic reform on area of pepper cultivation in India is presented below.Froin the estimated equation it can 'be observed that there is an increaseof 0.0 10642 per cent of area of cultivation of pepper per year during the periodof our study. Besides, estimated coefficients reveals that econoinic reformdoesn't have ally iinpact on area of pepper cultivation in India. This nonresponseon area of pepper cultivation may be because of the followingreasons. Firstly, pepper is inainly cultivated as a mixed crop. So the area ofpepper will change only if there is a change in the area of the main crop.Change in the area of main crop depends on how far econoinic refonn will

exert an influence on it. Secondly, all of a sudden shift in the area of peppercultivation is not feasible due to its perennial character.In terms of quantum of production, India ranks first in 1973-74, with aquantity of 28.7 tliousand tonnes out of 106.8 thousand tonnes. In the lateryear India iost its unique position either to Malaysia or Indonesia. In 1979-80Malaysia got the highest production of 37.4 thousand tonnes out of 126.2thousand tonnes. Except for few years, Indonesia ranked first and India was inthe second position in production throughout the eighties (Table 4.2).However, India has allnost regained its position during 90's. Indonesiaand Brazil were in neck to neck colnpetition in terms of quantum ofproduction of pepper. The year 1993- 1994 witnessed an increased productionby India with 50 thousand tonnes out of 182 thousand tonnes which comes to27.5 per cent. Indonesian production was nearly half of the Indian productionduring 1995 to 1999. Besides, the trend of pepper production was more or lessinixed during 2000-2003.The high level of pepper production in Indonesia is because of thefollowing facts. In Indonesia pepper is cultivated on a coinmercial basis.Highly productive vines are used for cultivation. Once the vines becameunproductive and senile, new vines are transplanted. The Indonesian fannersare considering pepper as a major crop than their Indian counter part (mixed orsubsidiary crop). Fanners producing pepper on a commercial base use deadstead as climbers.

Indian production of pepper has increased by inore than two fold during1973-74 to 2002-2003. During the saine period coinpound growth rate ofpepper production is worked out to be 4.7 per cent. Further, the coinpoundgrowth rate of pepper production during pre-reform is slightly higher than thepost reforin period in India (Table 4.9).Multiple linear regression analysis was carried out to identify theiinpact of econolnic reforin on pepper production in India. The result of theestimated equation is presented below.The value of iishows that 75.6 per cent change in production ofpepper is explained by time trend and dummy variable. Pepper production ofIndia has increased 0.0169 per cent per annuin during the period of our study.The duininy coefficients are found to be statistically insignificant, whichimplies that economic reform does not have any significant iinpact on pepperproduction in India. This inay be because of the earlier experience fanners hadwith other plantation. For instance, during the nineties there was a suddenincrease in the international price of cocoa. By seeing the bullish behaviour ofmarket prices farmers have shifted their plantation to cocoa cultivation. This

aised the quantum of cocoa production, which ultimately caused the prices tofall even below the cost of production. This in turn, forced the farmers to keepaway froin cocoa cultivation. Such experiences may be the reason for theimmediate non-response of farmers on economic liberalisation and hence onpepper production.On the productivity front India is showing a dismal picture (Table 4.2).In 1973-74 Malaysia ranlts the list with a productivity of 3,23 1 kg Per hectare,Brazil got the second position with 2,202 kg and the Indian figure was just 236kg. In simple terms, Malaysian productivity is 14 times inore than Indianyield. During 70's Malaysian productivity is far ahead of other nations. Brazilbecanze a leader in productivity front in the eighties.The nineties has witnessed another interesting picture.Thailandbecame a stiff challenger to all other countries by having inore than double theproductivity of Malaysia. Throughout the entire period of our study, Indianproductivity has remained more or less in the range of 200-400 kg per hectare.Colnpound growth rate of yield of pepper in India are found to be 1.6per cent, 2.0 per cent and 3.7 per cent during pooled, pre-reform and postreforin periods respectively (Table 4.9). It can also be observed that postreforincompound growth rate of pepper export (both in quantity and valueterms) from India is far greater than the pre-reform compound growth rate(Table 4.9).

India can claim a better position both in acreage and production but onproductivity front its position is bleak. From the results of Table 4.3, it can beobserved that on an average India cultivated pepper about 5 1.25 per cent areaof world acreage. But Indian share to world production on an average is 24.57per cent. This perccntage share of lndian pepper in world production clearlyindicates the predominance of India in world pepper market. But, productivityof pepper in India is relatively low. The reasons for low Indian productivitycan be cited as follows. Majority of pepper farmers are considering pepperonly as a secondary crop. It means that it is cultivated as a mixed crop inarecanut or coconut gardens. Here attention is inaiiily given to the primarycrop. This mixed cropping syste~n leads to another serious problem. Peppervines used to grow with the support of these live steads (climbers). Thereforewhatever manure is applied to pepper plant will be squeezed by these livesteads. Soine other reasons are the continuous cultivation of poor yieldingvines, existence of senile and unproductive vines, loss due to pests, diseasesand drought. All these factors kept Indian productivity to its low level.However, concerted effort is required to increase productivity. Farmersshould be encouraged to cultivate pepper as a rnonocrop and dead steadsshould be used instead of live steads as climbers. Highly productive vinesshould be planted. All the senile and unproductive vines should be replaced.Tilnely application of insecticide is required. Implementation of the above

measures will help the country to increase its productivity and hence, canbenefit froin its dominance in acreage and production.(ii) Trends in Countrywise Export of Pepper'l'rends in countrywise export of pepper for the period 1973-2000 ispresented in Table 4.4. India's contribution in the year 1973 was 32 per cent;the highest ainong pepper exporting countries of the world. Tl~roughoutseventies Indian share ranged between 12 to 32 per cent. Indonesian shareranged between 16 to 30 per cent, whereas Malaysian share is in between 24-35 per cent. In the seventies Malaysia topped the countries. On an averageIndia contributed 30 per cent to world export during the period of our study.Besides, Indonesia, Brazil and Malaysia remained stiff competitors to India inpepper export.Hence, one can maintain that on an average wit11 5 1 per cent share ofworld acreage, 25 per cent of world production and 30 per cent of worldexport, India's dominance in pepper is undoubtful. In this context it will beinteresting to kizow the statewise contribution of pepper cultivation andproduction.(iii) Trends in State wise Area, Production and Productivity of PepperKerala, Karnatalta, Tainil Nadu, Pondicherry, Andaiz~an and NicobarIslands are the states / Uiiion Territories where pepper is produced. Treiids ofstatewise distribution of area, production and productivity of pepper for the

period 1970- 197 1 to 1999-2000 is presented in Table 4.5. In 1970-7 1, Iceralacultivated pepper in an area of 117.54 thousand hectares and accounted for25.03 thousand tonnes of production. Kerala's acreage of pepper remainedmore or less the same throughout the seventies. ICarnataka's share both incultivation and production was stable throughout the eighties. The unionterritory of Pondiclieny contributed marginally to the Indian kitty of pepper.One of the noteworthy point of nineties is that Andainan and Nicobar islandshas also emerged as a pepper producing region. It can also be mentioned thatsome of the North Eastern states have also started pepper cultivation on anexperimental basis, but its contribution is not to an amount that call beaccoul~ted for. However, it can be observed that the highest acreagc andproduction of pepper is accounted to the state of Icerala during the period ofour study.The area under major crops of Kerala from 1960-6 1 to 1998-1999 ispresented in Table 4.6. In 1960-61, Kerala had a gross cropped area to thetune of 2,348.86 thousand hectares, out of which paddy occ~lpied the dominantposition with 33.16 per cent. Ainong spices, pepper got the highest acreage of99.75 thousand hectares with 4.2 per cent. Tremendous increase can be citedin almost all the non-food crops froin 1970-71 onwards. However, it can beobserved that other than coconut and rubber, pepper recorded the highestpercentage increase in acreage during the period of our study.

Coinpoiu~d growth rate of area of pepper cultivatioil in I

the second rank with a good margin from Karnataka. It can be concluded thatPondicherry maintained a triple fold productivity of India and Kerala hasrecorded as the second highest productive region during the period of ourstudy.Cornpound growth rate of pepper production in Kerala is 4.0 per centfor the pooled period (Table 4.9). Besides, the post reforin compound growthrate of pepper production (2.2 per cent) is less than the pre-reforn period (3.0per cent). Coinpound growth rate of pepper yield in Icerala is 1.4 per cent, 1.2per cent and 1.4 per cent during pooled, pre-reform and post-reform periodrespectively (Table 4.9).The estiinated multiple regression equation to identify the iinpact ofecolio~nic reform on pepper production ill Icerala is presented below.The results of estiinated regression equation reveals that there is 0.0125per cent increase in area of pepper production in Kerala per year. Besides, theestimated coefficients reveals that ecolloinic reforms has failed to lnalte anysignificant impact on pepper production in Kerala. The reason can be cited asfollows. During the last decade there was an increase in the international priceof cocoa. Lured by this high price, Kerala farmers have shifted to cocoa

cultivation at the cost of some perennial crops. This lead to over production ofcocoa. It ultimately resulted in reducing the price below the cost of production.This forced the Kerala fanners to revert to the earlier crops. Such a bitterexperience may be the reason for non-response of Kerala fanners to pepperproduction.On the whole, it can be observed that Icerala contributed 95 to 98 percent of India's cultivation and 85 to 98 per cent of India's pepper production(Table 4.7). Thus, Kerala has a near inonopoly over pepper cultivation andproduction over other Indian states.(iv) Trends in District wise Area, Production and Yield of PepperPepper cultivation can be seen throughout the length and breadth of thestate. The reason is that pepper can be cultivated as a mixed crop, either incocoiiut garden, in cardainoin plantation, in arecanut garden or can be grownwith jackfruit or mango tree. Pepper plant is susceptible to the various climaticconditions of the statc. Hence, it can be considered as a more general crop thanany other spices of the state. Table 4.8 gives a district wise profile of area,production and yield of pepper for the period 1981-82 to 1998- 1999.ICannur district has accounted for one-fourth acreage in 1981-82. Thedistricts of Kottayam, Kozhiliode and Idukki are also dominant in peppercultivation.On the production front, ICannur topped the list followed byKozhikode, Wynad, Quilon and Idukki districts. The lowest productivity wasearmarked by Palghat district during 198 1 - 1982.113

Out of a total plantation of 182.38 thousand hectares of pepper crop inthe state Idukki registered 49.75 thousand hectares followed by Wynad with40.21 thousand hectares in 1998-99. Other than these two districts, Kannurand Kozhiltode districts also play a significant role in the area of cultivation ofpepper. The district of Idukki has registered a remarkable yield of 590kglhectare, followed by Wynad with 452 lcglhectare during 1998- 1999. Idulclcidistrict has registered a four fold increase in acreage of pepper cultivation:while its production has increased 15 fold and productivity by inore than 3fold during the study period. With inore than 5 fold increase in acreage, 6 foldincrease in production and one fold increase in productivity, Wynad accountedthe next highest increase. Thus, during the study period it can be observed thatIdullti, Wynad, Kozhiltode and Kannur are the four districts where peppercultivation is dominant.The reasons for dominance of pepper production in the districts ofIduklci and Wynad can be because of the following facts. Both these districtsare situated in high ranges which are surrounded by hilly regions. Thisgeographical positioil coupled with the climate and the suitable soilculminated in higher level of production in these two districts. Alleppeybecame the district with lowest area, production and productivity. Thisposition of Alleppey can be cited to its geographical position. The district isinaroolled with lakes and back waters and therefore may not be conducive for

pepper plantation. However, all districts of Kerala have its ow11 share to becontributed in the state kitty of pepper production.We have already observed that Idukki, Wynad, 1.ozhiltode and I

(v) Trends in Price of Pepper Assembling Markets of KeralaTo analyse the price trend, inonthly wholesale price of pepper for theperiod April 1974 to March 2003 were considered. Graphical representation ofprice trend of Alleppey, Kochi, Kozhikode and Tellicherry assemblingmarkets were shown in Figure 4.1.It can be observed that prices of all these four assembling lnarltetsinoved in a steady direction till April 1985. Between April 1986 and April1996 there were some upward movements. During 1996-1997, price of someinarlcet has drifted froin other inarlcets. A steep rise in the montlily wholesaleprice can be observed in April 1997 which has sustained till April 2000. Thisincrease can be cited due to the following reasons. Dwring this period therewas a fall in production of some of the major pepper producing countries suchas Brazil and Indonesia. Indonesian vines were severely affected by pests anddiseases. Hence, the reduction in quantum of production benefited the Indianfariners through increased prices. 011 the whole, it can be deduced thatinonthly wholesale price of all these four iliarkets were moved synchronouslyduring the period of our study.The results of multiple linear regression analysis to identify the impactof econoinic reforin on inonthly wholesale price of selected pepper marltets ofKerala are presented below.

Note : Parentheses shows 't' value.*shows significant at 1 per cent level.*%shows significant at 5 per cent level.* * *shows significant at 10 per cent level.AL - Alleppey Marltet, KO = Kochi Marlcet, KZ = Kozhilcode Marketand TL = 'l'ellichewy MarlcetIt can be observed that the value ofranges in between 0.75 and0.81. All the F-statistics are found to be statistically significant. The trendcoefficients ranges in between 0.0032 and 0.0037. There isn't n~uch variationsin the trend coefficients of all the marltets, which implies that prices of theseinarkets have inoved in the same direction and tlze level of increase in prices inall the selected niarlcets are more or less the same. Even though the interceptcoefficients are negative, all the slope coefficients are found to be positive andstatistically significant. It implies that econoinic reform is able to make apositive influence on the monthly wholesale price of pepper in all the majorpepper assembling inarkets of Kerala. This positive influence of econoinicreforin on selected inarlcet prices of ICerala is due to the fact that there isgreater transparency in dissemination of inarlcet information andcoinpetitiveness ainong the sellers and buyers.

Concluding RemarksFrom early settleinelit onwards. spices were known to hulnan beings.Out of the total Indian exports 1.24 per cent was contributed by the spicessector. Anlong spices, pepper got a predoininant role in India. Pepper earns aforeign exchange to about 40 per cent in spices export. In the world area andproduction of pepper, fifty per cent of the area cultivated beloilgs to India andits contribution in world production of pepper is above 25 per cent.Other major pepper producers in the world are Indonesia, Brazil,Malaysia, Vietnaln and China. Minor contributions are also shared by SriLanka, Thailand, Madagascar and Mexico. On productivity front Brazil,Malaysia and Indonesia are far ahead of India. This is because of thecoininercial cropping pattern followed by these countries. Indian productivityis low because of the use of low yielding vines, existence of senile andunproductive vines, mixed cropping pattern, loss due to diseases and naturalcalamities.However, on an average, India contributed 30 per cent towards worldexport of pepper. Similarly Indonesia, Brazil and Malaysia also havecontributed significantly towards world pepper export.Ainong the Indian states, Kerala has a near monopoly over pepper with95 to 98 per cent of the area and 85 to 98 per cent of production. Karnataka,Tamil Nadu, Pondicherry and Andaman and Nicobar Islands are some otherregions from where pepper is produced. On productivity front Pondicheny

had a triple fold productivity of the country and can be accounted for its smallacreage. Second highest productivity is registered by Kerala during the periodof our study.All the geographical regions of the state are blessed with pepperproduction. The state is cultivating pepper mainly as a mixed crop. Though allthe districts contributed towards state's production Idukki, Wynad. Kozhikodeand Kannur are the dominant districts. The geographical position of Idukkiand Wynad districts (high range area) are so conducive for pepper cultivation .Kozhikode and Kannur are producing world cIass pepper. The Iowestcontribution of Alleppey is because of the fact that, the district is maroonedwith lakes and backwaters.It can be concluded that post-reform colnpound growth rate of pepperyield and export froin India are greater than the pre-reform growth rate.Whereas, the pre-reforin compound growth rate of area and production ofpepper in India and area, production and yield of pepper in Kerala are greaterthan the post-reforin growth rate. Multiple linear regression result shows thateconoinic reforin is not able to inake any significant impact on area andproduction of pepper in India and pepper production in Kerala. This nonresponseof economic reforins on pepper production may be due to thefollowing factors. Pepper is considered only as a secondary crop. The area ofpepper directly depends on the area of the main crop which is used as thestead. If the area of the main crop is influenced by economic reform, the area

of pepper will also be influenced indirectly. Similarly, farmers used to makechanges in perennial crop acreage only if they are able to observe a steadyiinprove~nent in the price of the product.However, pepper cultivation in ICerala has been slightly affected due toeconolnic reform. Silnilarly economic reforin has positively affected themonthly wholesale price of all the selected pepper assembling markets ofKerala. With econoinic reforms and trade liberalisation there is greaterdissemination of inarltet information which ultimately lead to co~npetitivenessamong econoinic agents.

Table 4.1. Trends in Share of Pepper in Spices Export from IndiaYearSpices Export(Rs. Crores)Pepper Export(Rs. Crores'Pepper Export as% of Spices~xport (value)1960-61 16.39 8.49 51.81961-62 17.52 8.07-- 46.11962-63 13.37 6.57 49.1

Sources: Government of India, Arecanut and Spices Database 2002,Directorate of Arecanut and Spices Development, Department ofAgriculture and Co-operation, Ministry of Agriculture, Calicut,Icerala.Government of India, Spice Statistics, Fourth Edition, Spices Board,Ministry of Commerce, Cochin, Icerala.Government of India, Statistical Abstract, Central StatisticalOrganisation, New Delhi.

Country121.40 469 289.32 141.48 501Y A P Y PA P Y A P - Y A P --YIndia 107.35 22.71 212 109.40 18.22 167 125.12 34.00 272 132.81 31.34 236 149.93 48.09 327Indonesia 79.00 39.56 501 80.00141.24 515 80.00 32.00 400 80.00 37.00 462 80.00 36.00 450Malaysia 11.36 23.40 2060 10.51 1 16.56 1575 5.04 16.00 3173 --- 5.30 15.50 2925 7.70 14.00 1818-.Brazil 26.42 29.26 1107 20.45 35.38 1730 19.00 30.50 1605 16.00 25.30 1581 19.00 27.00 1421ISri Lanka 5.31 1.51 / 284 5.51 2.59 '470 6.38 2.68 420 16.90 1.88 272 17.60 2.51 1142Madgascar 6.10 2.39 1 392 6.12 2.61 426 6.20 2.80 452 6.20 2.80 452 6.20 3.00 484Total(inc1uding 237.10 others)124.67 526 233.66 123.66 529 244.25 125.99 516 258.98 1991-92CountryA P3.38234.581988-891989-901990-91A P Y A P Y A P Y India168.26 43.42 258 171.4955.19 490 173.431 47.95 276 184.20 52.01---89.87 47.00 523 100.00 50.00 500 118.00 1 53.00 449 ( 95.05 ; 61.00Malaysia 10.00 22.80 2280 11.30 27.50 2434 11.25 31.00 2756 10.30 31.00Brazil 30.00 33.00 1100 30.00 30.00 1000 30.00 31.50 1050 50.00 50.00Sri Lanka 7.63 1 3.44 451 8.35 3.50 ( 194 9.13 1.99 218 8.60 2.85Madgascar 6.30 3.50 538 6.47 3.38 522 6.50 3.38 520 6.50 Total(inc1uding 322.19 176.57 548 340.57 190.45 520 362.77 194.74 537 374.12 1992-93Y A P Y J282 189.39 50.76 2681642 98.003010 10.00 26.00 , 260010001 35.00 - 27.50 786Indonesia 62.00 1 633 '--l____3135206278.806.50367.1 13.26 3703.38 1 520204.55 557I__&-- .

Table 4.3. India's percentage share to world acreage andproduction of pepperYear1973-741974-751975-761976-771977-781978-791979-801980-811981-82India's percentageshare to worldacreage61.560.158.457.153.247.753.952.151.5India's percentage shareto world production26.924.021.621.118.416.821.921.320.42001-022002-03Note : Calculated on the basis of Table 4.2.23.419.9

Table 4.4. Trends in Countrywise Export of Pepper1979Qty20.825.037.425.20.92.6111.9(%share)18.622.333.422.50.82.31001980Qty26.329.731.532.00.63.1123.2(Ohshare)21.324.125.625.90.482.51001981Qty20.634.028.646.92.13.0135.2(%share)15.225.121.234.71.52.2100(Quantity in '000 tonnes)CountryQty.India 31.6lildonesia 125.5Malaysia 24.1Brazil 13.8Sri Lanka 0.5Madagascar 3.2Total 198.719731974 19751976(% share) Qty. (% share) 'Qty. (% share)Qty (%share)32.0 26.3 29.3 24.2 26.2 20.5 19.225.8 24.4 15.71 17.5 30.21 33.6 15.2 31.4 16.5 34.1 24.2 37.9 22.735.614.0 0.5 3.2 15.2 0.3 2.2 16.9 0.3 2.4 17.8 0.1 3.5 19.3 0.1 3.8 20.0 0.01 3.9 18.80.0093.7100 189.9 100 92.2 100 106.5 100Qty.24.630.928.917.10.63.7105.81977(%share)23.329.227.316.20.63.5100Qty.15.737.036.630.41.22.2122.71978(%share)12.830.229.824.41.01.8100

CountryIndiaIndonesia1987 1988 1989Q~Y.Q~Y. (% Qwshare)43.01 36.7 36.19 28.730.0025.641.51 1 32.934.4842.14(%)share25.431.11990 (%share)15.842.91991 1992Qty. (% Qty.(Oh Qty. share)share)28.89 20.6 19.66 13.0 22.68 47.68 34.1 49.67 32.8 61.44 Sources: Government of India, Cocoa Areacanut and spyG Statistics, 1970-1 983, 1989-1 993International Pepper Community, Pepper Statistical Year Book 1986, 1995/1996, and Various Issues.

Table 4.5. Trends in Statewise Area, Production and Productivity of PepperStatesKamatakaA2.16Kemla , 117.54TamiINadu 1 0.231970-71P Y1.07 49525.0 , 2130.05 1 217Pondicheny ] 0.03 1 0.01 1 333 1 0.01Total 1 119.96 1 26.16 1 21s I 118.63I1971-72A I P2.05 1 1.00116.34 1 25 10 , 216 , 116.340.23 1 0.05 1 217 1 0.250.0126.16Y488A3.201972-73P0.9825.150.05I 11000 1 0.01 ] 0.01221 1 119.80 1 26.19Y3062162.00(Area in '000 hectares; Production in '000 tomes, yield kg/ha)1000 / 0.01219 1 121.721973-74A I P3.28 1 0.91Y A277 118.25 1 235 372 27.75 3.32118.410 1 % 1 0 04 0 - -- 18I28.70 1 2360.01121.921974-75A1975-76P Y0.91 1 274 3.40 / 0.94 27627.23 230 1 108.25 24.58 2270 n4 1 237 1 n 77 I n ni 1 125".-,- 0.0128.18 ( 231 1 111.93P-.--25.57Y. "4228

Table 4.6. Area under major crops in Kerala ('000 hectares)Note: Figures in parentheses shows percentage to gross cropped area.Sources: Government of Kerala, Statistics for planning, Directorate of Economics and Statistics, Tl~iruvanmthapuram, Aug. 2001Government of Kerala, Agricultural Statistics of Kerala 1992-93; Ecollornic Review and Agricultural Abstract of Kerala.Pillai. P.P, Kerala Economy - Four Decades of Development.

Table 4.7. States / U.T. Share of area and production of pepper to IndiaPercentage Area of Pepper tes / U.7'. to Total1984-851985-862.3 1 96.72.1 97.20.90.70.010.011333.62.095.297.41.20.6 0.02

Table 4.8. Trends in District wise Area, Production and Yieid of Pepper in Keraia(A: Area '000 ha., P: Production '000 tonnes, Y: Yield Kg per ha.)PY2333103541081981-82 1985-861988-891991-921992-93DistrictsA P Y A P Y A P I Y A P Y A PThiruvananthapurm, 5.38 - 1.73 323 5.06 1.57 309 3.92 0.80 205 4.43 0.95 214 4.16 0.97Quilon 9.80 2.78 285 7.89 2.98 1 378---8.22 1.94 235 7.87 2.39 304 8.16 2.53P athmamthitta- 4.68 1.52 326 6.32 2.16 342 4.88 1.71 350 5.12 1,81Alleppey4.82 0.97 201 3.64 0.65 178 1.58 0.41 259 2.02 0.23 114 1.95 0.21t

----ITable 4.9. Compound Growth Rate of Various Indicators of PepperIlldicatorsExport of pepper (quantity)from IndiaExport of pepper (value)from IndiaArea of pepper in IndiaProduction of pepper inliidiaYield of pepper in IndiaArea of pepper in KeralaProduction of pepper inKeralaYield of pepper in KeralaPooled Period1.9(6.362)*11.9(1 8.918)*3.2(9.185)*4.7(10.990)'k1.6(3.667)*2.5(8.066)*4.0(8.649)"1.4(5.43 1)*Note : * Shows significant at one per cent** Shows significant at five per cent.Pre-reformPeriod1.9(4.880)" -11.8(1 3.973)*2.5(3.875)*3.9(3.483)"2.0(2.126)""1.6(2.934)*3 .O(3.403)*1.4(2.692)"Post-reformPeriod6.4(1.761)36.4(1 1.447)"1.7(1.284)3.7(4.195)*3.7(3.202)'"1.1(0.935)2.2(2.390)**1.2(1 -368)

CHAPTER VAN EMPIRICAL VALIDITY OF MARKET INTEGRATIONHYPOTHESIS WITH REFERENCE TOTHE PEPPER MARKET IN KERALAIntroductionPerformance of any sector or an economy depends on the degree ofmarket efficiency. Generally, efficiency of inarket can be judged on the basisof prevailing price in the whole market system. If there exists a uniform pricethroughout the lnarltet system, it can be considered as a situation of efficientmarket. In a way one can observe that the efficient marltet and unified priceare synonymous. The very existence of unified price throws light on the longrun equilibrium. Long run equilibrium in general, is achievable only byphasing out short run divergences. Drift of short run divergences ensures auniform price or in other words long run equilibrium.In a situation of long run equilibrium, all market agents are said to be inan economically viable position. In other words, producers at the village level,traders, middleinen and consumers of the product are able to attain theireconomic goal. That is to say the motive of all these agents are linked by a keyfactor - price. A unified price resembles an integrated market. Excess ordeficiency of demand or supply will be wiped off or compensated by theintegrated inarket through price mechanism. Thus, the concept of market

integration becornes an integral part of efficient market systeln or long runequili briuin.The attainment of long run equilibriulil with a unified price will besustainable only if the market forces are free to play its own role. Any policyor activity that hinders ~narket forces to act freely will directly favour someeconoinic agents at the cost of others. In other words the optimum efficiencyof the inarket in toto will be hindered. To attain opti~nuln efficiency the bestway is to liberate the economy from all sorts of controls. In this context byfooting on classical laissez - faire policy, there is a global consensus on theconcept of liberalization and econornic reforms.Earlier thinkers were of the opinion that agriculture is the onlyproductive sector of the economy. This statement can not be taken to meanthat all the remaining sectors are non-productive. It just reminds one to attachthe importance to be given to the agriculture sector. A more efficientagriculture sector will be able to make a strong foundation for both thesecondary and tertiary sector.In this context, attempts were made to know the efficiency ofagricultural sector by relying on studies of market integration* of severalagricultural commodities. But studies on validity of market integrationhypothesis pertaining to cash crops are limited. It can be observed that studieson pepper; which is a dominant cash crop of Kerala; is not able to get the- -* For detailed discussion see chapter two.140

equired attention. More than 96 per cent (Kerala State Land Use Board, 1997)of Pepper cultivation and production of the country is froin the state of Kerala.Besides, it fetches a sizeable foreign exchange also.Pepper has a remarkable export exposure. From time in~mnemorial,pepper is a doininant spice exported to various countries froln Kerala. It isconsumed throughout the length and breadth of the coulltry because of itsexport exposure and medicinal properties. Thus, pepper being a globallyconsumed product, the recent globalisation and econoinic reforms might havein some way or other affected the pepper market.In line with the global view, Government of India too has initiated aseries of econoinic reforms since 1991. The motive behind all these policychanges is to inalce the market efficient and hence to attain econoinicdevelopment at an earlier spell of time. These policy changes at national level,directly or indirectly influence all the constituents of the country. In thisperspective, it is imperative to believe that economic reforms have made somesweeping changes in the pepper markets of Kerala. With the abolition oflicence raj and controlled pricing, producers are able to transport theirproducts to the inarket from where economic gains seems to be high. In otherwords, liberalisation can be instrumental in an unhindered disselnination ofmarket information throughout the system.On the above background, the present chapter attempts to examine thevalidity of market integration hypothesis with special reference to the selected

pepper markets of Kerala and if it is so, then it also identify how fartransmission mechanism of prices of pepper product across the selectedinarkets of Kerala exist? Besides, the present chapter investigates the influenceof recent economic reform on the validity market integration hypothesis andtranslnission lnechanism of prices across the selected pepper markets ofIcerala.MethodologyFour inajor pepper assembling inarltets of Kerala, namely Tellicherry,Kozhiltode, Kochi and Alleppey were selected for the analysis. Monthly dataof market prices of pepper product for April 1974 to March 2003 werecollected for examining the objectives of our study. To examine the impact ofeconoinic reforin on the validity of marlcet integration hypotl~esis andtransmissioii mechanism of prices between the selected inarkets, price serieswere divided into two. They are: (i) April 1974 to June 1991, and (ii) July1991 to March 2003.Validity of market integration hypothesis and transmission mechanismof prices will be examined with the help of Dickey-Fuller test, Phillips-Perrontest, Johansen's multiple cointegration technique and Error Correction Models.It consists of three stages. They are (i) Data series has to be tested forstationarity, (ii) Verification of cointegration between the markets, and (iii)Once the inarket is cointegrated, error correction model should be estimated toidentify transmission mechanism of prices among the markets.142

In our study, Dickey-Fuller and Phillips-Perron tests' were conductedto know the stationarity of the variables and examination of market integrationhypothesis. Johansen's multiple cointegration technique was employed toverify the validity of cointegration between markets. After obtainingcointegration, transmission mechanism of excess prices between the marketswere tested with the help of error correction model.Johansen's multiplecointegration and the error correctioil models are explained below:Multiple Cointegration TestCointegration analysis became a widely used technique for examiningthe behaviour of two or inore than two data series. When variables have a unitroot, Engle and Granger showed that cointegration can be an einpiricallyuseful method to examine such relationship (Elliott, 1998).The Engle-Granger technique is a single equation regression residual-based test. Thoughit is a simple and attractive test for bivariate model, it doesn't perform well ina lnultivariate situation.A recently popular approach to analyse cointegrated system wasdeveloped by Johansen (1988, 1991). It extends the Engle-Granger procedureto a multivariate context where there may exist Illore than a singlecointegrating relationship among a set of 'n' variables (Sarker, 1993).Johansen proposed a maximum likelihood method for estimating long run"or detailed discussion on Dickey-Fuller and Phillips-Perron Tests see Chapterthree.

equilibrium relationship or cointegrating vector and derives likelihood ratiotest for cointegration in a Gaussian Vector Error Correction Model.Following Johansen and Johansen and Juselius (1990), the model isexplained on the following way. Given a vector of Y of 'n' potentiallyendogenous variables, the inodel of Y as an unrestricted vector autoregressionwith k lags of Y can be specified asYt=aY,_[+ ............... +ai,Yt-k+~t......................Where u,- IN (0, C) (1)and Y is (n x 1) and ai is an (n x n) matrix of parameters.In its reduced forin with each variable in Yt regressed only on laggedvalues of both itself and all other variables can be specified asAY,= CIAYt-I + ...... + Ck-, Yt-k+l f 'itYt-k + U, ...........(2)where Ci = - (i - a, - .........-ai);(iZ 1, .... k- 1)7t can be represented asn = a p', where a is the speed of adjustment to disequilibrium, whilep' is the matrix of long run coefficients in such a way that p' Ybk in equation 2represents upto (n-1) cointegration relationship in a multivariate modelensuring that Y, converge in their long run steady state solutions.Rewriting equation 2 asAY, + a J3'Yt.lc = CI AY,, + ...... i- AYt-k+l -+ ut ........... (3

enables correcting short run dynamics by regressing AY,.k and Ytmk separatelyon the right hand side of equation 3.Thus the vector V,, and Vkt are obtained from,AYt = Xl AYt., + ... .. + Xt.{ AYt..kil + Vat ,..,.....,.(4)Yt-k = Z1 AYtml + ...... + Zk-, AYtk+l 3- Vlcr ........... (5)Equation 5 is used to form the residual (product moment) matrices asRij =Z' C Vit Vfjt (1, J=0, k)The null hypothesis that there are 'r' cointegrating vectors is tested byusing two liltelihood ratio tests called the trace test and the inaxiinuln eigenvalues. If HI: is a special case of Hz: for r = p, then the likelihood ratio tracestatistic is defined asSimilarly, tlie lnaximum eigen value statistic for testing H2(r) inHz(r+ 1) can be defined as- 2 ~n(~;r/r+l) = -~1n(l-i,+,)The asyinptotic distribution of these likelihood ratio tests representmulti-variate version of the Dickey-Fuller distribution.Thus, when there are 'n' price series and (n-1) potential cointegratingrelationship, it first tests the null of zero cointegrating relationship; if that isrejected, it then test the null that there is only one cointegrating relationshipand proceeds in stepwise way to test the null of higher number of such

elationships upto (n-1). The test statistic is compared with the critical valueswhich are generated through simulations and are reported in Johansen andJuselius (1990').Johansen's method has several superiority over Engle-Granger'scointegration method. Engle-Granger does not provide any procedure fortesting multiple cointegrating vectors when there are three or Inore variables.Again for conducting the test of integration in Engle-Granger method, it isnecessary to identify the central (exogenous) and the peripheral (endogenous)markets. Rut Johansen's multiple cointegration method treats all the variablesas explicitly endogenous and takes care of the endogeneity problem byproviding an estimation that does not require arbitrary choice of variable fornormalization. Further, since Johansen's test is carried out in a vectorautoregressive inodel of reduced fonn, the problem of simultaneity can beavoided.By examining the distributional properties of maximuin likelihoodestimates of cointegrating vector, it is observed that the maximum likelihoodestimator is super-consistent, syininetrically distributed and median unbiased -asyrnptotically and that an optimal theory of inference applies (Cheung andLai, 1993). Johansen's approach also provides with a very flexible format forinvestigating the properties of the estimates under various assumptions aboutthe underlying data generating process (Sarker, 1993).

Error Correction ModelWhen variables are cointegrated, the residuals from the equilibriumregression can be used to estimate the error correction model. With errorcorrection model, price transmission in time series can be estimated. Errorcorrection model einployed to know price transrrlission among selectedmarkets of the present study are presented below.AAL, = a, +Pi EC+ ALt-, + a,, ALt-, + ICOt-1 + KOt-? + a;, KZt-1+ + TLt-l + TLt-i + e,,AKZ, = a, +pi EC+ a,, KZ,-, + a,? KZt-, + a,, AL,-, + a,, AL,-, + a,, KO,-,+ a,, KO,-, i- a,, Kzt-, + a42 KZ,-, + e,,where,EC - Error correction term,AL - Alleppey inarket price,KO - Kochi market price,KZ - Kozhikode inarket price, andTL - Tellicherry market price.

If 'pi'(i=l.. . . . .4) has a negative sign and statistically significant, itcan be inferred that there is an indirect strong long term price transmissionmechanism across the selected markets of pepper product in Kerala. Thevalues of 'aiYs depict the degree of association between respective individualmarkets.Empirical Results and DiscussionBefore examining the existence of cointegration, it is necessary todetermine the order of integration of the selected data series. Stationarity testswere einployed to identify whether the series have finite variance and atendency to return to the mean. To check stationarity of data series, unit roottest has been employed. Both Dickey-Fuller and Phillips-Perron tests wereconducted for the pooled (April 1974 to March 2003), pre-reforin (April 1974to June 1991) and post-reform period (July 1991 to March 2003). Further, thenecessary tests were carried out on the log of monthly wholesale prices of theselected pepper markets of Kerala.Dickey-Fuller and Phillips-Perron test (with and without trend) resultsfor the monthly price series of the selected markets are presented in Table 5.1and 5.2 respectively. Initially, tests has been conducted on the price series inlevels, then on the first difference. Test results of price in levels of the selectedmarkets with and without trend revealed that Dicltey-Fuller test does not rejectthe null hypothesis at least at five per cent level during pooled, pre-reform and

post-reform periods. However, Phillips-Perron test has rejected unit roothypothesis in certain markets at least at 5 per cent level of significance. Thenull hypothesis was rejected (both Dickey-Fuller and Phillips-Perron tests)during pre-reform, post-reforin and pooled periods at one per cent level in thecase of first difference of price series of the selected pepper ~narlcets of Kerala.Hence, it can be concluded that price series of the selected pepper markets ofKerala were integrated of order one during pooled, pre-reforin and post-reforlnperiods.It implies that market integration hypothesis pertaining to theselected market prices of pepper product of Kerala were accepted during theperiod of our study.The reason for the integrated nature of individual markets can be citedas follows. Alleppey inarlcet is the nearest assembling centre for all thesouthern districts of Kerala. As far as Kozliikode market is concerned, it is amajor pepper producing district and the district is surrounded by highlyproductive pepper producing areas. Kochi market is an assembling centre andit is also a terminal market in which pepper products are exported to variouscountries. Finally, Tellicherry market has its own unique features that qualitywise the products of Tellicherry market is internationally acclaimed and islabelled as "Tellicherry Garbled Extra Bold".After confirming that the price series of the selected pepper markets ofKerala are integrated of order of one, the next procedure is to examine therank of 'n' or the cointegrating vectors. Optimum lag was decided on the basis

of Altaike Information Criterion (Table 5.3a). The necessary co-integrationvector was estimated on the basis of optimum lag selected. Generally,integrated inarltets must share a cominon trend and will have commonintegrating vector. For examining this issue Johansen-Juselius method ofmultiple cointegration test were conducted and its results were presented inTable 5.3 for the period of our study. The likelihood ratio test results indicatethe presence of three cointegrating vectors at one per cent level of significancefor the pooled period. It implies that all the four markets were in factcointegrated and shares market information on changes in prices. Thus theeillpirical evidence suggests that all the selected pepper markets of Kerala doexhibit a long-nm relationship during the pooled period of our analysis.But once data are subdivided into pre and post-reform periods,cointegration results are not in parity with the pooled data results. Here thenull hypothesis of zero cointegrating vector has rejected. Proceeding further,the null hypothesis of one cointegrating vector has accepted; which means thatthere is only one factor which integrates the inarltets. The number ofcointegrating vectors differs during pre and post-reforin periods compared topooled period. However, it implies that there is sharing of inarltet informationamong the four markets during pooled, pre-reforin and post-reform periods inthe long run.After examining the long-run association between the markets, the nextstage is to identify the direct and indirect transmission mechanism of prices

etween the markets and it will be captured with the help of error correctionmodels. To get an autocorrelation free error correction model, two months' lagwas considered. Besides, we restricted ourselves to the order of one toestimate the required Error Correction Model for simplicity. convenience andcomparison between the periods of our study.The results of the Vector Error Correction Model (VECM) for pooled,pre-refor~n and post-reform periods are presented in Tables 5.4a, 5.4b. 5.5 and5.6 respectively. In comparing tables 5.4a and 5.4b it can be observed thatECM results with one error correction term is more significant than the ECMresults with three error correction terms. Therefore our further analysis isconcentrated on table 5.4a. Since error correction equations are estimated indifference form, R~ values are comparatively low ranging from 0.09 to 0.44during the period of our study. It can be observed from the results of Table5.4a that all the error correction coefficients except one are significant wit11 thedesired negative sign. It implies that there is a strong indirect pricetransmission of pepper product froin one market to the other during the pooledperiod. There is a direct transinission of pepper prices from Alleppey marketto all the other markets. Kozhikode market pepper price is directly influencedby Kochi market pepper price. Kozhikode market price of pepper also has adirect bearing on Alleppey, Kochi and Tellicherry market pepper prices.Again, Tellicherry market price of pepper is directly able to influence the

Alleppey and Kozhikode market prices of pepper. Further, it can be concludedthat Alleppey market is the dominant among the selected markets of Kerala.The results of Table 5.5 reveals that error correction coefficients of allthe markets are highly significant with a positive sign; which means that thereis only a weak association of indirect transmission mechanism of prices fromone market to the other during pre-reform period. Here, Alleppey inarketpepper price directly affects the pepper price of Kozhikode and Tellicherrymarkets. Pepper price of Kochi market is also able to influence pepper priceof Alleppey and Tellicherry markets. In the pre-reforin period, Kozhikodemarket pepper price has influenced only Alleppey market pepper price.Si~i~ilarly there is a direct pepper price transmission froin Tellicherry market toAlleppey and Kozhikode markets.Table 5.6 results reveal that half the number of error correctioncoefficients are significant with negative sign and the remaining are positiveand insignificant, which implies that indirect price transmission is strongerduring post-reform period than pre-reform period. The influences of Alleppeyand Kochi markets are similar to those in pooled period. In the case ofKozhikode market pepper price, there is a direct transmission of pepper pricesto Alleppey and Tellicherry markets. But Tellicherry market pepper price isnot able to influence directly the pepper price of any of the other markets. Asin pooled period, Alleppey market turned out to be the dominant inarketduring the post-reform period also.

Thus, the results indicates that price transmission between the selectedmarkets of Kerala is comparatively stronger during post-reform period thanthe pre-reforin period. In other words, through economic reforms and tradeliberalization, there is greater dissemination of market information amongmarkets; which ultimately lead to well integrated markets and thereby excessprice in one pepper market has transmitted into the other. It can also beobserved that (Table 5.7) Alleppey lnarltet is the doininant market followed byKozhikode, Icochi and Tellicherry markets during our period of study.Alleppey market is not a predoininantly pepper produciiig market, but itis able to decide prices of other markets to a greater extent. The main reason isthat Alleppey is the nearest assembling centre to all the southern districts ofIcerala and it resulted to an edge over other markets. As far as Kozhikodeinarlcet is concerned, it is one of the dominant pepper producing districts ofKerala. Besides, Wynad the second largest pepper producing district of Keralais nearest to the Kozhikode market. These factors are responsible for peppermarket of Alleppey to be dominant followed by Kozhikode market.Lesser dominance of Tellicherry in comparison with Alleppey andKozhiltode markets is due to its unique features. Pepper products ofTellicherry market is superior in quality and is ranked as one of the best ininternational markets (George, 1989). Hence, it is not able to influence othervariety of adjacent markets due to peculiar variety of pepper product.

Kochi market is the terminal market and it is not able to make a desiredimpact like the other selected markets of Kerala. This may be due to the factthat surrounding areas of Kochi inarket is not dominant in pepper production.Further, Kochi market is a place from where the pepper products of otherassembling centres are procured and exported. So, Kochi market price ofpepper product will have an influence on international inarket prices than thedomestic market prices.Concluding RemarksDickey-Fuller, Phillips-Perron, Johansen's multiple cointegration andError Correctioil Models were employed to examine the objectives of thestudy. Dickey-Fuller and Phillips-Perron test results revealed that all theselected markets of Kerala are well integrated during pooled, pre-reform andpost-reform periods. Johansen's multiple cointegration results exhibit a long-run relation between the four assembling markets during the period of ourstudy.The error correction model results reveal that there is strongtransmission inechanisin of prices from one market to the other during pooledperiod. Transmission mechanism of prices across the inarkets is strongerduring post-reforin period than the pre-reform period. This inay be due to theinfluence of recent econo~nic reforms and liberalisation in dissemination ofmarket information.Further, it can be observed that Alleppey is the dominant marketfollowed by Kozhikode, Kochi and Tellicheny markets. Alleppey market is

doininant because it is the most accessible assembling centre to the southerndistricts of Kerala. Do~ninance of Kozhilcode market is due to the fact that it isa highly pepper producing district surrounded by similarly productive areas.Kochi market is primarily exposed to international trade than domestic trade.Hence. it is not as dominant as Alleppey and Kozhikode markets. Tellicherrymarket is also not able to make much influence like other inarkets due to itsunique quality of pepper.

Table 5.1Results of Dickey-Fuller (DF) and Phillips-Perron (PP) TestsTest Statistics at levelsMarket without trend with trendPrice DF PP DF PP(1) (2) (3 (4) (5AIIeppey -1.410 -4.543* -2.285 -17.188*Icochi -1.237 -2.477 - 1.652 -8.582*Kozhikode -1.256 -2.981** - 1.748 -12.943*Test Statistics at First differenceWithout trendWith trendAlleppey -9.599" -270.652" -9.594* -270.143*Kochi -8.765* -362.782* -8.771 * -362.187*Kozhiltode -10.1 11 * -323.106:k -10.1 18"322.517*Tellicherrv -10.169"292.86* -10.173"292.322*Note : * shows significant at one per cent**shows significant at five per cent.Critical values are -3.98 (1 per cent) and -3.42 (5 per cent) for constant andtrend; -3.44 (1 per cent) and -2.87 (5 per cent) for constant only.

Table 5.2Results of Dickey-Fuller (DF) and Phillips-Perron (PP) TestsTest Statistics at levelsMarlcet without trend with trendPrice DF PP DF PP(1) (2)-(3) (4) (5)Alleppey -1.186 -3.619* - 1.596 -10.390*Kochi - 1.086 -2.457 - 1.492 -6.442%I(lozhikode -1.156 -3.414** -1.710 - 11.789*Tellicherry -1.203 -3.902* - 1.663 -1 1.386*Test Statistics at First differenceWithout trendWith trendAlleppey -8.026* -1 70.871 * -8.0 14* -170.350*Kochi -9.648* -225.77 1 * -9.630* -225.33*I

Table 5.3aResults of Akaike Information CriterionSample : 1974.4 to 2003.3No. of lags 0 1 2 3 4Sample : 1974.4 to 1991.6-12.095 -12.056 -12.030 -12.046 -1 1.94Sample : 1991.7 to 2003.3-7.570 -7.541 -7.480 -7.468 -7.344Table 5.3 bResults of Johansen's Multiple Cointegration TestNull hypothesis Eigen Liltelihood Critical Criticalvalue ratio value value(1 per cent) (5 per cent)Sample: 1974.4' to 2003.30.218 178.879* 70.1 63 .O0.162 98.197* 48.5 42.40.106 40.475* 30.5 25.30.115 3.792 16.3 12.3Sample: 1974.4 to 1991.6r=O 0.871 422.480* 70.1 63 .Ordl 0.468 11.494 48.5 42.4r12 0.009 1.856 30.5 25.3r

Table 5.4aECMResults of Error Correction Model during Pooled Period(1)A Alleppey(2)-0.05A Kochi(3)-0.003A Kozhikode(4)-0.02A Tellicherry(5)-0.09A (Kochi (-1))A (Kochi (-2))(1.67) ***-0.09(0.89)0.06(3.14) *-0.21(1.93) *-0.05(1.8 1) ***0.42(4.05) *0.08(3.25) * --0.006(0.06)0.041 (3.24) * (l.22) (4.37) * 1 (1.75) ***A (Tellicherry (- 1 )) 0.44 0.17 0.58 0.20A (Tellicherry (-2))ConstantR'0.04(0.42)14.07(0.3 1)0.14-0.07(0.6 1)17.90(0.39)0.090.27(2.43) * *17.52(0.40)0.410.07(0.79)15.06(0.39)0.26Log likelil~oodAICSC-2798.9116.2816.39-2806.8516.3216.44-2789.4216.2216.33-2751.5516.0016.12Note: Parentheses represent t-values*significant at one per cent,**significant at five per cent, and"* significant at ten per cent.

Table 5.4bResults of Error Correction Model during Pooled PeriodECMl(1 1A Alleppey(2)-0.02A Kochi A Kozhikode A Tellicherry(3) (4)(5)-0.002 -0.03 -0.07(1.42) (2.56)"; (1-58) (2.89)*A (Kochi (-1)) -0.06 -0.22 0.32 -0.005(0.62) (1-63) (3.12)* (0.08)A (Kochi (-2)) 0.04 -0.03 0.06 0.03(0.57) (0.44) (0.82) (0.41)A (Kozhiltode (- 1)) -0.42 -0.10 -0.89 -0.56(3.6)4c (2.42)** (6.8) 4c (5.8)*A (ICozhikode (-2)) -0.05 -0.02 -0.25 -0.18(0.89) (0.64) (1.56) (2.59)*A (Tellicherry (- 1 )) 0.34 0.18 0.64 0.19(2.42)** (1.15) (1.52) (1.72)***A (Tellicherry (-2)) 0.02 -0.06 0.18 0.06(0.26) (0.89) (1.96)*qc (0.93)Constant16.23 18.54 19.23 14.72Log lilcelihoodAIC-2542.3 118.6218.72SC-Note: Parentheses represent t-values-2636.5818.5318.64-2842.3218.4218.53-253 1.5218.2018.32*significant at one per cent,**significant at five per cent, and*** significant at ten per cent.

ECMTable 5.5Results of Error Correction Model during Pre-reform Period(1)I/ A Alleppey A Kochi / A Kozhikode I A Tellicherry(2)0.19(3)0.10(4)0.20(5)0.2 1A (Kochi (-2))ConstantR'~2Log likelihood0.1 1(1.13)*significant at one per cent,**significant at five per cent, and**%significant at ten per cent.-0.15(1.50)A (Kozhiltode (-1)) -0.39 -0.06 -0.63 -0.34(1.65) *** (0.26) (2.51) *+ (1.42)1 1 1 1A (Kozhiltode (-2)) -0.78 0.20 -1.10 -0.82(1.30)' (0.86) (4.39)* (3.35)*1A (Tellicherry (- 1)) 1.61 0.34 1.14 0.85(5.34)#' (1.12) (3.56)* (2*7i)*A (Tellicherry (-2)) 1.40 0.03 1.31 0.98(4.63)* (0.10) (4.07)* (3.13) *AICSC1 1.73(0.55)0.390.36-Note: Parentheses represent t-values16.44(0.75)0.140.100.17(1.65) *4c*13.69(0.60)0.370.34-1450.26 -1454.76 -1462.68---14.31 14.4314.47 14.52 14.600.14(1.37)11.50(0.52)0.370.34-1456.8314.3814.54

Table 5.6Results of Error Correction Model during Post-reform PeriodECM(1)A Alleppey(2)-0.04A Kochif 3)0.01A Kozhikode 1(4)0.03A Tellicherry(5)-0.09(2.23)" (2.33) ** (2.06)HcA ( Alle~~e~ (-2)) 0.13 0.26 0.10(1.01) (1.90) (0.78)A (Kochi (-1)) -0.03 -0.24 0.63(0.19) (1.44) (4.12) 4cA (Kochi (-2))A (Kozhikode (- 1))0.05(0.32)-0.44(2.67)*-0.09(0.58)-0.12(0.73)0.15(1.04)-0.89(5.66)"0.08(0.58)-0.57(3.92) *A (Tellicherry (-1))(1.60)A (Tellicherry (-2))(0.19)Constant 1 21.01 1 26.12 1 24.59 1 22.80Log Iikelihood-1199.01AIC17.14SC17.35Note: Parentheses represent t-values-1201.89 1 -1191.98 1 -1179.8617.1917.39I17.0417.25I16.8717.08"significant at one per cent,**significant at five per cent, and* * * significant at ten per cent.

Table 5.7lmpact of Price of Pepper across the Selected Markets (at one lag)---+ Alle- Icochi Kozhi- TelliliodecherryAlleppeyPre-reform periodICocliiKozhi-ItodeTclllchelr)~Alleppey-- -- -Post reform periodI

CHAPTER VISUMMARY AND CONCLUSIONS~~undness of any econolny depends on its agricultural base. A welldeveloped agriculture sector is a pre-requisite for the develop~nent ofsecol~dary and tertiary sectors and hence the economy. Similarly, the natureand magnitude of macroeconomic aggregates also depends on the agriculturalsector. Indian agriculture, by contributing a quarter to its Gross DomesticProduct, by supportiiig nearly three-fourth of its population and by providinge~nployinento inore than half of its workforce; has a significant status in theeconomy.An efficient inarlcet systein in agricultural sector is required to keep thepace of agricultural growth. And an efficient market systein depends on~narlcet inechanisin and the efficiency of market mechanism in turn depends onregional price integration. Uniformity in prices can be attained through theintegration of an economy. Therefore, market integration is the process bywhich price interdependence occurs. Integrated markets are those where pricesare determined interdependently; which is assumed to mean that price changesin one will be fully transmitted on to others.Market integration is a useful parameter to measure marketingefficiency for temporal and spatial analysis. An integrated inarket is requiredfor an effective iinplementation of econo~nic reforms, liberalization policies,

agricultural or famine policies. With market integration, all economic agentsare able to attain an econoinically viable gain. Hence, the principle ofattainment of inaxirnuin utility with the most efficient utilization of resourcescan be obtained through market integration.Now, if the economy is ruled by controlled pricing, the existing pricewill never reflect the true market situation. Free play of market forces ispossible only in a liberalized economy. Only the forces of demand and supplycan inalce price transmission effective. Liberalisation of economy fromcontrols will enable farmers and traders to get remunerative gain. In such anatmosphere, the process of liberalization got pro~ninence throughout theworld. On this line, Government of India too has initiated several policies aspart of economic reforms and liberalization. The aim of all these policyinitiatives is to attain efficiency in the market.Against this background, attempts were made to test the validity ofmarket integration hypothesis of various agricultural crops at national andinternational level. Some efforts were aIso made to know the effect ofliberalization on inarlcet integration. The majority of these studies hasconcentrated on food crops. Even though some of the Indian states aredominant in cultivation of non-food crops; market integration studies of thesenon-food crops are rare in literature.The state of Kerala has devoted three-fourth of its agricultural area tonon-food crops. Among non-food crops, the state is having a monopoly over

pepper which is considered to be a major item of spice. But no effort is madeto explore inarket efficiency of this doininant crop either at regional ornational level. The entire state cultivates this crop. All the small and bigfarmers sell their products to traders of their region. These traders in turntransport the products to the wholesalers. Then the wholesalers of the statetrade the products to the pepper assembling markets namely, Tellicherry,Kozhikode, Kochi and Alleppey, whichever is accessible. All the assemblingcentres then transport it to the terminal market namely Kochi, which isconsidered to be the exporting centre.The reform process in India has started in July 1991. The initial focuswas on industrial sector. Trade liberalization in agriculture later became partof econoinic refonns. Removal of iinport controls on certain agriculturalproducts, relaxation of quantitative restriction on imports, delicensing someimports and slashing of basic import duty are some of the important policyinitiatives. This necessitated an impetus for international competition forexports. Hence, these policy initiatives exposed Indian agriculture to globalmarkets. Since pepper is exposed to international markets, economic reformand liberalization will definitely affect the pepper economy of Kerala. In sucha trading scenario following issues emerges.(i)How far pepper markets of Kerala are integrated? If they are integrated,how far price transmission mechanism works out among the markets?

(ii)What will be the impact of econoinic reforms on market integration andtransmission mechanism of prices across the pepper markets of Kerala?Against this background, we attempted to investigate the following objectives:(i)examination of earlier literature pertaining to inarket integration studiesof agricultural commodities and to identify the gap in the studies.(ii)critical evaluation of inarket integration hypothesis and its earliermethodology employed to verify market integration hypothesis whichwill enable us to derive suitable inethodology to verify the marketintegration hypothesis.(iii)examination of trend and pattern of area, production, productivity andprice of pepper with special reference to Kerala economy and also toidentify the impact of liberalization on these indicators of pepper.(iv)investigating the validity of inarlcet integration hypothesis amongpepper assembling markets of Kerala and if it is, then to identify thetransmission inechanism of prices of pepper during pre and post-reformperiods.SummaryThe present study consists of six chapters. By introducing the conceptof inarltet integration, the first chapter briefly explains the objectives of thestudy. Uiider inethodology of the study ratio, percentages and compoundgrowth rates were considered to examine the trend of area, production,productivity and prices of pepper. To identifp the impact of liberalization,167

evity, the earlier literature have reviewed under two sections nanlejy;international level studies and national level studies. A series of studies haveundertaken at international level. Studies at national level are less incomparison with international studies. All these studies have attempted toverify the market integration hypothesis of various crops of agricultural sectorand products of allied activities. Most of the studies were mainly attempted toverify the market integration hypothesis of food crops. However, someattempts were made to identify the validity of market integration hypothesis ofnon-food crops. At international level, some studies were conducted to knowthe effect of liberalization on market integration.Most of the studies have used monthly wholesale price to examine themarket integration hypothesis. Few studies have relied either on daily orweekly prices. A major number of studies were able to reveal the existence ofstrong market integration in their analysis.The third chapter analysed the theoretical and methodological issuesrelated to marketing efficiency and market integration hypotheses. Theconcepts of market integration, marketing efficiency and its variouscomponents were dealt in detail. Price spread and market integration are thetwo criteria to measure marketing efficiency. The techniques employed tomeasure price spread and market integration were detailed in this chapter.Some of the important techniques employed to test the market integrationhypothesis were price series correlation, variance component approach,

ordinary least square method, autoregressive model, Koyck's distributed lagmodel, Ravallion model, cointegration technique and parity bound model. Thepresent chapter concludes with a critical evaluation of all these statistical andeconometric tools.Common trends and seasonal colnponents may make an upward bias inthe results of correlation coefficient. Variance component approach fail tomeasure the degree of market integration once the assumption of constantvariance of price over the season and transaction cost relaxes. Attention is notpaid to the properties of error term and stationarity of data series in ordinaryleast square framework. Regression techniques have generally ignored ormisrepresented the tiine series properties of price series and hence, there areserious flaws in the estimation procedure. The Engle-Granger cointegrationtechnique is a bivariate inodel that ignore the linkage that inay operate througha third-market.With the presentation of a brief history of pepper, the fourth chapteranalysed countrywise and statewise area, production and productivity ofpepper. Countrywise export of pepper was also analysed. Similarly adistrictwise analysis of area, production and yield of pepper pertaining to theeconoiny of Kerala were also examined. Besides, price trend of major pepperassembling markets of Kerala were also investigated. Simple ratio,percentages and compound growth rate were used to examine the trend andpattern of pepper. By introducing duminy variables, multiple linear regression

models were estimated to identify the impact of liberalization on variousindicators of pepper.The validity of market integration hypothesis of pepper assemblinglnarkets of Kerala were empirically tested in the fifth chapter. Four majorpepper asselnbling inarkets of Kerala, viz; Tellicherry, Kozhikode, Kochi andAlleppey were selected for the analysis. Monthly data of market prices ofpepper product for April 1974 to March 2003 were collected to verify theinarket integration hypothesis. Price series were divided into two to examinethe impact of econo~nic reform on the validity of market integrationhypothesis and transmission mechanism of prices between the selectedmarkets, They are: (1) April 1974 to June 1991, and (ii) July 1991 to March2003,Dickey-Fuller and Phillips-Perron tests were conducted to know thestationarity of the variables and examination of market integration hypothesis.Johansen's inultiple cointegration technique was employed to verify theexistence of cointegration between the markets. After obtaining multiplecointegration between the selected markets, transmission mechanism of excessprices across the inarkets were tested with the help of error correction model.The last chapter summarises arguments of the thesis and explorespolicy implications and future agenda of research.

FindingsThe main findings of our study are:(1) Most of the studies at national and international level have given muchemphasis on food crops. Market integration studies related to non-foodcrops or cash crops were allnost neglected.(2) At national level, there is only a single study of coconut marketpertaining to Kerala economy. But studies on pepper, which is adominant cash crop of Kerala are lacking.(3) Some international studies have sliown that economic liberalization hasa positive effect on marketing efficiency. At regioiial level, no seriousattempt is made to know the effect of econoinic liberalization onmarketing efficiency.(4) On inethodological front, studies employing modern techniques such as~~iultiple co-integration test and error correction models to identify theexistence of market integration across multiple markets are rare in theliterature.(5) Critical evaluation of all the statistical and econoinetric techniquesreveals that Johansen's multiple cointegration technique is superiorthan other techniques to test the validity of market integrationhypothesis in a multivariate framework.(6) It is observed that nearly fifty per cent of the area of cultivation ofpepper in the world belongs to India and its contribution to the world

production of pepper is about 25 per cent. This lower share in pepperproduction can be attributed to lower productivity. The reason for lowerproductivity can be because of the mixed cropping panem. use of lowyielding vines, existence of senile and unproductive vines, loss due todisease and natural calamities.(7) Other than India, Indonesia, Brazil, Malaysia, Vietnam and China arethe other pepper producers. Sri Lanlta, Thailand, Madagascar andMexico are the ininor contributors. Productivity of Malaysia and Brazilare high because of their colninercial cropping pattern.(8) Ainong the Indian states, Kerala has a near monopoly over pepper with95 to 98 percentage of the area and 85 to 98 percentage of production.Karnatalta, Tamil Nadu, Pondicherry and Andainan and Nicobar islandsare the other regions froin where pepper is produced. The highestproductivity has recorded by Pondicherry due to its small acreage.(9) All the regions of Kerala state are blessed with pepper production; eventhough Idukli, Wynad, Kozliikode and Kannur are the dominant pepperproducing districts. The geographical position of the districts of Idukkiand Wynad coupled with its suitable soil resulted in higher level ofproduction. Similarly, the districts of Kozl~ikode and IGinnur arerenowned for their international quality of pepper.

(10) Prices of all the assembling markets of Kerala have movedsynchronously. This may be because of the improved transportationsystem and information technology prevalent in the state.(1 1) Further, it is revealed that post-reform compound growth rate of pepperyield and export from India are greater than the pre-reforn growth rate.But the pre-reform compound growth rate of area and production ofpepper in India and area, production and yield of pepper in Kerala aregreater than the post-reform growth rate. Farmers' non-response toeconomic reforins on pepper production may be due to the followingfactors. Pepper is considered only as a secondary crop. The acreage ofpepper also depends on the acreage of the main crop which is used asthe stead. Similarly acreage shift of a perennial crop like pepper alsodepends on the steady nature of market price.(12) Multiple linear regression result shows that economic reform is not ableto make any significant impact on area and production of pepper inIndia and pepper production in Kerala. However, pepper cultivation inKerala has positively affected by economic reforins. With economicreform, there is greater dissemination of market information which leadto competitiveness ainong econoinic agents.(13) The economic reform is able to influence the monthly wholesale priceof Alleppey, Kozhikode, Kochi and Tellicherry assembling markets of

Kerala. This positive influence is because of the fact that with refomthere is greater transparency in dissemination of inarket information.(1 4) Dickey-Fuller and Phillips-Perron test results shows that all the selectedmarkets of Kerala were well-integrated during pooled, pre-reform andpost-reform periods. The well-integrated nature of the selected marlsetsmay be because of the following reasons. Alleppey rnarltet is thenearest asseinbling centre to the southern districts of Kerala. Kozhikodei~~arlcet is a highly producing district and is surrounded by equallyproductive regions. Kochi inarltet is acting both as an assembling centreand as a terminal market. Tellicherry inarket has the superior quality ofpepper in the world.(1 5) Johansen's multiple cointegration results exhibits a long run relationbetween the four pepper assembling rnarltets during the period of ourstudy. It implies that all the four pepper markets used to share inarketinformation in a transparent way.(16) The error correction model results revealed that there is strongtransinission mechanisn~ of prices from one inarket to the other duringpooled period. Again, transmission mechanism of prices across themarkets is stronger during post-reform period than the pre-reformperiod. This may be due to the influence of recent economic reformsand liberalization in dissemination of inarltet inforination.

(17) It has also been observed that Alleppey is the dominant marketfollowed by Kozhikode, Kochi and Tellicherry markets. Alleppeyinarltet is do~ninant because it is the nearest pepper assembling centreto the southern districts of Kerala. Dominance of Kozhikode rnarket isdue to the fact that it is a highly pepper producing district surroundedby similarly productive areas. Since Kochi market is primarilyconcerned of exports, it is not dominant as other markets. Tellicherrymarket is not able to exert much influence like other markets due to itspeculiar quality of pepper.Policy llnnplicationsThe main policy iinplications of the study are:(1) We are able to observe that economic reform is not able to exert thedesired impact on several indicators of pepper. This may be because ofthe fact that the policy initiatives of economic reform has notimplemented in toto. Therefore, policy initiatives should beiinple~nented in toto.(2) It is explored that Indian pepper productivity is one of the lowest in theworld. Farmers should be promoted to cultivate pepper as a major cropand only dead steads are to be used as climbers.

Future Agenda of ResearchThe main future agenda of research in the area are:(1) Marltet integration study of pepper by incorporating more wliolesaleinarltets of different pepper producing Indian states can be undertakenand can also examine the transmission inechanism of prices. Besides,state specific validity of the hypothesis can be explored.(2) Daily or weekly prices of various market information are better thann~onthly market price behaviour. Hence, the entire experiments can beworked out on the basis of daily or weekly price behaviour of peppermarltet .(3) The present experiment can be extended to other cash crops of Keralaand to the rest of the states in India.

BIBLIOGRAPHYBooksBain, Joe. S. (1968), Industrial Organisation, Second Edition, John Wiley andSons, Inc, New Yorlc.Bliss, C.J. and N.H. Stern (1982), The Economy of the indian Village, OxfordUniversity Press, Oxford.Caves, Richard ( 1977), American Industry, Structure, Conduct, Perjbrmance,Englewood Cliffs, Prentice Hall.Clark, F.E. (1954), Principles of Marketing, In Thakur, D.S. (1998),"Agricultural Marketing Efficiency: Agenda of 21'' Century", TheBihar Journal of Agricultural Marketing, Vol.VI, No.2, PP. 133- 149.Cournot, A.A. (1971), Researches into the Mathematical Principles of theTheory of Wealth, A.M. Kelly, New York.Cundiff, Edward W. and Richard R. Still (1972), Basic Marketing ConceptsDecisions and Strategies, Prentice Hall, New Delhi.Dahl, Dale C and Jeroine W. Hamlnond (1977), Market and Price Analysis.The Agricultural Industries, McGraw Hill, New Delhi.Dhondyal, Dr.S.P. (1989), Problems ' of Indian Agriculture, FriendsPublications, Meerut.Dorfinan, Robert (1979), The Price System, Prentice Hall of India PrivateLimited, New Delhi.Enders, Walter (1995), Applied Econometric Time Series, John Wiley & Sons,Inc.Granger, C.W.J. and A. Weiss (1983), Time Series Analysis of Error-Correction Models, In S. Karlin, T. Amerniya and L.A. Goodman(Eds), Studies in Econometrics, Time Series and Multivariate Statistics,In honour of T.W. Anderson, Academic Press.Gravin, J.L. (Ed) (1929), Encyclopaedia Britannica, The EncyclopaediaBritannica Company Limited, London, Vol.XIV.Gulati, Ashok and Tim Kelley (1999), Trade Liberalisation and IndianAgriculture, Oxford University Press, New Delhi.

Harris, Barbara (1984), State and Market, Concept Publishing Company, NewDelhi.Henderson, J.M. and R.E.Quandt (19711, Micro Economic Theory, AMathematical Approach, Second Edition, McGraw Hill BookCompany, Singapore.Jasdanwalla, Z.Y (1966), Marketing Eficiency in Indian Agriculture, AlliedPublishers Private Limited, New Delhi.John, W. Parry (1969), Spices, Chemical Publishing Company Incorporation,New Yorlc.Khan, M.T. (19901, Spices in Indian Economy, Academic Foundation, Delhi.Kohls, Richard. L. (19671, Marketing of Agricultural Products, Macinillan,New Yorli.Koutsoyiannis, A. (1979), Modern Microeconomics, Second Edition,Macinillan Press Limited, London.Lele, Ulna. J (19711, Food Grain Marketing in India- Private Performanceand Public Policy, Cornell University Press, Ithaca and 1,ondon.Lele, Ulna. J. (1973), Foodgrain Marketing in India, Popular Prakasham Ltd,Bombay.Lipsey, Richard G. and Colin Harbury (1 992), First Principles of Economics,Second Edition, Oxford University Press, Oxford.Madnani, G.M.K. (1986), Introduction to Econometrics - Principles andApplications, Third Edition, Oxford and IBH Publishing Co., NewDellii.Marshall, A. (196 l), Principles of Economies, 8" Edition, Macrnillan, London.Moore, John R, S.S. Johl and A.M. IUlusro (1973), Indian Food grainMarketing, Prentice Hall, New Delhi.Narasiihain, N.V. (1994), A Model for Commodity Price System Analysis,Himalaya Publishing House, Bombay.Narver, John and Ronal Savitt (1971), The Marketing Economy: An AnalyticalApproach, Holt Rine Hart and Winston Inc., New York.Padmanabhan, D. (1976), The Spices, National Council of EducationalResearch and Training, New Delhi.

Padrnanabhan. K. (19931, Economic Analysis of Marketing, Printwell, Jaipur,India.Pillai, P.P. (1 994). Kerala Economy - Four Decades of Development, Instituteof Planning and Applied Economic Research, John Mathai Foundation,Dr. John Mathai Centre, University of Calicut, Thrissur, Kerala.Purseglove, et. al. (1 98 11, Spices, VoL I, Longinan Scientific & TechnicalGroup Limited, New York.Purseglove, J.W. (19681, Tropical Crops - Dicotyledous, Vo1.2, Longman,London.Rainai~athan, R. (L995), Introductory Econometrics, Fort Worth: DrydenPress.Ridley, H.N. (1912), Spices, Macmillan, London.Salvatore, Doininick (1998), Managerial Economics in u Global Economy,Third Edition, McGraw Hill International Edition, Economic Series.Shrivastava, O.S. (1 996). Agricultural Economics, Rawat Publications, Jaipurand New Delhi.Singh, G.N, D.S. Singh, Rain Iqbal Singh (1987), Agricultural Marketing inIndia, Analysis, Planning and Development, Chugh Publication,Allahabad, India.Stigler, G. J. (1 969), The Theory of Price, Macmillan, London.Stock, James. H. (1999), A Class of Tests for Integration and Cointegration.In "Cointegration, Causality and Forecasting, Robert F. Engle andHalbert White (Eds), Oxford U~iiversity Press.Stonier, Alfred and Douglas C. Hague (1982), A Text book of EconomicTheory, FiAh Edition, The English Language Book Society andLongman Group Limited.

Alexander, Carol and John Wyeth (1994), "Co-integration and MarketIntegration: An application to the Indonesian Rice Market", Journal ofDevelopment Studies, Vol. 30, No. 2, PP. 303-328.Angulo, A.M and J.M. Gil. (1996), "Vertical Integration and PriceTranslnission in the Spanish Poultry Sector", World AgriculturalEconomics and Rural Socioloa Abstracts, Vol. 38, No. 9, P.709.Asche, Frank, Helge Breinnes and Cathy R. Wessells (1999), "ProductAggregations, Market Integration and Relationship between Prices: AnApplication to World Salmon Markets", American Journal ofAg~icultural Economics, Vol. 81, No.3, PP. 568-58 1.Baharuinshah, Ahmad Zubaidi and Muzafar Shah Habibullah (1994), "PriceEfficiency in Pepper Markets in Malaysia: A Co-integration Analysis",Indian Journal of Agvicultural Economics, Vo1.49, No.2, PP.205-2 15.Rasu, Jyotish Prakash and Sou~nyananda Dinda (2003), 'Market Integration;An Application of Error Correction Model (ECM) to Potato Market inHooghly District, West Bengal", Indian Journal of AgriculturalEconomics, Vo1.58, No.4, PP. 742-751.Baulch, Bob (1997), "Transfer Costs, Spatial Arbitrage and Testing for FoodMarket Integration", American Journal of Agricultural Economics,V01.79, N0.2, PP.477-487.Behera, Souinya and M. Indira (2002), www.comodityindia.coin. Vo1.2, Issue.I, PP.4-12.Behura, Debdutt and Durga Charan Pradhan (1998), "Co-integration andMarket Integration; An Application to the Marine Fish markets inOrissa", Indian Journal of Agricultural Economics, Vol.53, No.3,PP.344-350.Bhide, Shashanka, Ashole Chowdhury, Earl 0 Heady and M.A. Muralidharan(1981), "Structural Changes in an Agricultural Assembling Market:. ACase Study of Arecanut (Betelnwt) Market in Mangalore, KamatakaState", Indian Journal of Agricultural Economics, Vol.XXXVI,No.2,PP.25-34.Bijinan, J. W. (1 996), 'LBiotechnology and Vertical Integration in the DutchPotato Chain in Economics of Innovation - The Case of Food Industry",World Agricultural Economics and Rural Sociology Abstracts, Vo1.38,No.9, P.709.

Blyn, George (1973), "Price Series Correlation as a Measure of MarketIntegration", Indian Journal of Agricultural Economics. Vo1.28, No.2,PP.56-59.Brunk, M. (1950), "Marketing Research in Operational Efficiency -Proceedings, National Marketing Workshop", In R.L. Kohls ( 1956)Towards a more meaningful concept of Marketing Efficiency, Journalof Farm Economics, Vo1.38, PP.68-73.Buccola, S .T. (1 983), "Risk Preferences and Short-run Pricing Efficiency",American Journal of Agricultural Economics, Vo1.65, PP.587-9 1.Carter, C.A., and N.A. Hamilton (1989), "Wheat Imports and the Law of OnePrice", Agribus, Vo1.5, PP.489-96.Carvalho, F.C. De; Marques S.A; Maia M.L and Yoshil R.J. (1994), "Studyof Market Integration in Sugar-alcohol Agri-business in the State ofSao Paulo 1970-92. Brazil", World Agricultural Economics and RuralSociology Abstracts, Vo1.3 6, No.2, P. 129.Cl~ahal, S.S and K.S. Gill (199 l), "Measurement of Marketing Efficiency inFarm Sector: A Review", Indian Journal of Agricultural Marketing,Vo1.5, N0.2, PP. 138-143.Cheung, Yin-Wong and Kon S. Lai (1993), "Finite Sample Sizes ofJohansen's likelihood ratio tests for cointegration", Oxford Bulletin ofEconomics and Statistics, Vo1.55, No.3, PP.3 13-328.Cochrane, Willard W. (1957), "The Market as a Unit of Inquiry inAgricultural Economics Research", Journal of Farm Economics,V01.39, N0.2, PP.21-40.Dahlgran, Roger. A and Steven C. Blank. (1992), " Evaluating the Integrationof Contiguous Discontinuous Markets", American Journal ofAgricultural Economics, Vo1.74, No.2, PP.469-475.Delgado, Christopher. L (1986), "A Variance Components Approach to FoodGrain Market Integration in Northern Nigeria", American Journal ofAgricultural Economics, November, PP.970-979.Dercon, Stefan (1995), "On Market Integration and Liberalisation: Methodand Application to Ethiopia", The Journal of Development Studies,Vo1.32, No.1, PP.112-143.

Dickey, David. A and Wayne. A. Fuller (1979), "Distribution of theEstimators for Autoregressive Time Series with a Unit Root", Journalof American Statistical Association, Vo1.74, No.366, PP.427-43 1.Dickey, David. A and Wayne. A. Fuller (1981), "Likelihood Ratio Statisticsfor Autoregressive Time Series with a Unit Root", Econornetrica,V01.49, No.4, PP. 1057-107 1.Elliott, Graham (1998), "On the Robustness of Cointegration Methods WhenRegressors Allnost have Unit Roots", Econometrica, V0l.66, No.1,PP.149-158.Engle, R.F. and B.S. Yoo (1987). "Forecasting and Testing in CointegratedSystem", Journal of Econometrics, Vo1.35, PP. 143- 159.Engle, R.F. and C. W .J. Granger (1 987), "Cointegration and Error Correction,Representation, Estimation and Testing", Econornetrica, Vol.55, No.2,PP.25 1-276.Fafchamps, M and S. Gavian (1996), "The Spatial Integration of LivestockMarkets in Niger", Journal of African Economies, Vo1.5, No.3, PP.3 66-405.Fafchamps, Marcel (1972), "Cash Crop Production, Food Price Volatility andRural Market Integration in the Third World", American Jownal ofAgricultural Economics, Vo1.74, No. 1, PP.90-99.Faminow, Merle D and Bruce L. Benson (1990), "Integration of SpatialMarkets", American Journal of Agricultural Economics, Vo1.72, No. 1,PP. 49-62.Gadgil, D.R. (1961), "Agricultural Price Policy - A Suitable Basis forAchieving Targets", Yojana, September.Gardner, Bruce L. and Karen M. Broolcs (1 994), "Food Prices and MarketIntegration in Russia. 1992-93", American Jownal of AgriculturalEconomics, Vo1.76, No.3, PP.64 1-646.Garoian, Leon (1 97 l), "Implications of Changes in Market Structure onExtension Marketing Programmes and Adininistration", Journal ofFarm Economics, Vol.XLII1, No.3, PP.674-683.George, M.V and A.J. Singh (1970), "Structures, Conduct and Performance ofWholesale Vegetable Markets in Punjab", Agricultural Marketing,Vol.13, No.1, PP.1-9.

George, M.V. (1 984), "Changing Market StructuresiMarketing Policies':Indian .Journal of Agricultural Economics, Vo1.39, No.4, PP.586-5 89.George, P.S (1972), "Role of Price Spread in Determining Agricultural PricePolicy", Agricultural Situation in India, Vol.XXVI1, No.9, PP.617- 19.Gliatage, M.B. (1958), iDevelopment Work of Agricultural Marketing - ItsScope and Method". Agricultwal Marketing, Vol. 1, No. 1, P.3.Ghosh, Madhusudan (20001, "Co-integration Tests and Spatial Integration ofRice Markets in India", Indian Journal of Agricultural Economics,V01.55, No.4, PP.6 16-626.Goodwin, Barry. K and Ted C. Schroeder (1991), "Co-integration Tests andSpatial Price Linkages in Regional Cattle Markets", American Journalof Agricultural Economics, Vo1.73, No.2, PP.452-464.Govindan, M. Joy and K.M.Sreekuinar (2003), "Viral Diseases of BlackPepper and Their Management", Indian Journal of Arecanut, Spicesand Medicinal Plants, Vo1.5, No.3, PP.96- 102.Granger, C.W.J. (1966), "The Typical Spectral Shape of an EconomicVariable", Econometrica, Vo1.34, PP. 150- 16 1.Granger, C.W.J. (1986), "Development in the Study of CointegratedVariables", O~ford Bulletin of Economics and Statistics, Vo1.48, No.3,PP.213-228.Harris, Barbara (1979), "There is Method in My Madness or Is It Vice Versa?Measuring Agricultural Market Perforinance", Food Research InstituteStudies, Vol.XVI1, No.2, PP. 1 97-2 1 8.Hays, H.M and I.H Mc Coy (1978), "Food Grain Marketing in NorthernNigeria; Spatial and Temporal Performance", The .lournal ofDevelopment Studies, Vo1.14, No.2, PP. 182- 192.Heytens, Paul J. (1986), "Testing market integration", Food ResearchInstitute Studies, Vol.XX, No. 1, PP.25-40.Horowitz, X (1 98 l), 'Warket Definition in Antitrust Analysis - A RegressionBased Approach", Southern Economics .hurnal, Vo1.48, PP.291-304.Hotelling, H, (1929), "Stability in Competition". Economic Journal, Vo1.39,PP.4 1-57.

Ismet, M. Barkley A.P and Llewelyn R.V (1999), "Government Interventionand Marlcet Integration in Indonesian Rice Market", World AgriculruralEconomics and Rural Sociology Abstracts, Vo1.4 1, No. 3, P. 196.Johansen, S. (1988), "Statistical Analysis of Cointegrating Vectors", Journalof Economic Dynamics and Control, Vol. 12, PP.23 1-254.]ohansen, S. ( 199 1), "Estimation and Hypothesis Testing of CointegratingVectors in Gaussian Vector Autoregression Models", Econometrics,V01.59, PP.1551-1580.Johansen, S. and K. Juselius (1990), "Maxiniuin Lilteliliood Estimation andInference on Cointegration with Application to the demand forMoney", Oxford Bulletin of Economics and Statistics, Vo1.52, PP.169-210.Kallfass, H.H (1993 ), "Cost Advantages Through Vertical Integration in theAgricultural Sector-Germany", Wovld Agricultural Economics andRural Sociology Abstracts, Vol. 3 5, No.9, P.75 1.Khedhiri, S. (1 999), "Empirical Analysis of Agricultural Market Integration inTunisia", World Agricultural Economics and Rural SociologyAbstracts, Vol. 4 1, No. 1 1, P. 869.Lau, L.J. and P.A. Yotopoulos (1971), 'fA Test for ReIative Efficiency andApplication to Indian Agriculture", American Economic Review,Vol.LX1, No. I, PP.94- 109.Lele, Ulna. J (1967), "Market Integration: A Study of Sorgliuin Prices inWestern India", Journal of Farm Economics, Vo1.49, No. 1, PP. 147-159.Lundahl, Mats and Erling Petersson (1982), "Price Series Correlation andMarket Integration: Some Further Evidence", Indian Journal ofAgricultural Economics, Vo1.37, No.2, PP. 1 84- 190.Majeed, Muhai~iined Ph.D. & Lakshini Prakash Ph.D. (2000), "The MedicinalUse of Pepper". International Pepper News Bulletin, Vol.XXV, No.1,PP.23-29.Mathew, A.G. (2002), "Pepper Oil and Oleoresin", International Pepper NewsBullettin, Jan-June Issue,Mathew, M. Thomas, V. Venkitachalain and S ebastian (1 997), "Long RunRelationship of Coconut Market in Kerala; A Co-integration",Agricultural Situation in India, PP ,777-7 80,

Mc Donald, Stephen. L (1953), "The Process of Integration and AgriculturalProblems", JournaZ offirm Economics, Vo1.35, No.3, PP.37 1-3 80.Monke, Eric and Todd Petzel (1984), "Market Integration: An Application toInternational Trade in Cotton", American Journal of AgricuituralEconomics, PP. 481-487.Munir, A, Suresh Waran. S, H. Selassie. and J.C. Nyankore (1997), "MarketIntegration in Developing Countries - A Case Study of SelectedVegetables in Indonesian Market", World Agricultural Economics andRural Sociology Abstracts, Vo1.39, No. 12, P.946.Naik, Gopal and V.P.S Arora (1986), "Marketing Pattern and PricingEfficiency in Indian Arecanut Market", Indian Journal ofAgriculturalEconomics, Vol. 4 1, No.2, PP. 172- 182.Nasurudeen, P and S.R Subrainanian (1995), "Price Integration of Oil andOilseeds", Indian Journal of Agricultural Economics, Vo1.50, No.4, PP.625-633.Nelson, C. and C. Plosser (1982), "Trends and Random Walks inMacroeconoinic Time Series", Journal of Monetary Economics,V01.10, PP.139-162.Palaskas, Theodosios B and Barbara Harriss White (1993). "Testing MarketIntegration: New Approaches with Case Material from the West BengalFood Economy", The Jownal of Development Studies, Vo1.3 0, No. 1,PP. 1-57.Patnaik, I

Purcell, Wayne. D. (1973), "An Approach to Research on Vertical Coordination.The Beef System in Oklahaina", American Journal ofAgricultural Economics, Vol.LV, No.2, P.64.Ravallion, Martin (1986), "Testing Market Integration", American Journal ofAgricultural Economics, V0l.68, No. 1, PP. 102- 109.Ravindran, P.N. and Johny A. Kallupurackal (2000), "Black Pepper ResearchUnder the All India Coordinated Research Project on Spices", IndianJournal of Arecanut, Spices and Medicinal Plants? Vo1.2, No.3, PP.71-78.Ravindran, T.C. and K. Manoj Kumar (2001), "Production, Export andMarketing of spices in India", Indian Journal of Arecanut, Spices andMedicinal Plants, Vo1.3, No.4, PP.246-257.Rozelle, Scott, Albert Park, Jikum Huang and Hehui Jin (1997),"Liberalisation and Rural Market Integration in Cl~ina", AmericanJournal ofAgricultural Economics, Vo1.79, No.2, PP.635-642.Rudra, Ashok (1980), "On the So Called Marketing Efficiency", IndianEconomic Review, Vol. XV, No. 4, PP.23 1-242.Sarlcer, Ralchal (1993), "A Maximum Liltelihood Cointegration Analysis ofCanadian Lumber Exports", Canadian Journal of AgriculturalEconomics, Vo1.4 1, No. 1, PP.97- 110.Schultz, T. W. (1 946), "Changes in ~conbrnic Structure Affecting AmericanAgriculture", Journal of Farm Economics, Vo1.28, No. 1, PP. 15-27.Sharxna, M.L. and S.K. Ainbastha (1995), "Globalisation of Indian Econoinywith Special Reference to Indian Agriculture", The Bihar Journal ofAgricultural Marketing, Vol.111, No. 1, PP.3 5-39.Silvapulle, Parain and Sisira Jayasuriya (1 994), "Testing for Philippines RiceMarket Integration: A Multiple Co-integration", Journal of AgriculturalEconomics, Vo1.45, No.3, PP.369-3 80.Sinharoy, Saikat and Sanjith R. Nair (1994), "International Trade and PepperPrice Variations: A Co-integration Approach", Indian Journal ofAgricultural ~conomics, Vol. 49, No.3, PP. 417-425.Sit, Arun Kumar, K. China Chenchaiah. and G.C. Acharya (2002), "BlackPepper - A Profitable mixed crop in Areca based Cropping Systemunder Terai region of West Bengal", Indian Journal of Arecanut, Spicesand Medicinal Plants, Vo1.4, No.4, PP. 155- 157.

Sivarainan, I< and K.V. Peter (1999), "An Overview of Spices Developinent inIndia", Indian Journal of Arecanut, Spices and Medicinal Plants, No. 1,PP.l-11.Slade. M. (1986), "Exogeneity Test of Market Boundaries Applied toPetroleum Products", Journal of Industrial Economics. Vol. 3, PP.29 1 -304.Sorensen, Jan Rose. (1993), "Integration of Product Markets and LowerWelfare in An Economy with Centralized Wage Setting", ScottishJournal ofPolitica1 Economy, Vol. 40, No.3, PP.323-329.Spices Week (2004), www.cornmodityindia.com.- (Various Issues).Stigler, George. J and Robert A. Sherwin. (1985), "The Extent of the Market",Journal of Law and Economics, Val-XXVIII, PP.55 5-5 85.Thaliur, D.S. (1974), "Food Grail1 Marltetiilg Efficiency: A Case Study ofGujarat", Indian Journal of Agricultural Economics, Vo1.29, No.4,PP.61-74.Thakur, D.S. (1998), "Agricultural Marketing Efficiency: Agenda of 21''Century", The Bihar Journal of Agricultural Marketing, Vol.VI, No.2,PP. 133- 149.Panchal, et. a1 (2003), The New Indian Express, Chennai, Oct. 27, P. 12.Timiner, C. Peter (1974), "A Model of Rice Marketing Margin in Indonesia",Food Research Institute Studies, Vol. XIII, No.2, PP. 145- 167.Zanias, G.P (1993), "Testing for Integration in European CommunityAgricultural Product Markets", .Journal of Agricultural Economics,V01.44, NO. 3, PP.4 18-427.Zhao, J.W. (1995), "The Development Trends of Integration of AgriculturalProduction, Processing and Marketing - Chinese Rural Econon~y",World Agricultural Economics and Rural Sociology Abstracts, Vo1.37,No 2, P. 122.

ReportsGeorge, Dr.C.K. (1989), Spices of India - An Overview, Proceedings of theWorkshop on Strategies for Export Development of Spices (Ed),C.K.George, et. al., Spices Board, Cochin, PP.1-14.Governinent of India (1 970- 1993), Cocoa, Arecanut and Spices Statistics,Directorate of Arecanut and Spices Development, Ministry ofAgriculture, Department of Agriculture and Co-operation, Calicut,Kerala (Various Issues).Government of India (1973-2003), Area and Production of Spices in lndiaand the World, Spices Board, Ministry of Commerce, Kochi, Kerala.(Various Issues).Government of India (1 990-2000), Statistical Abstract, Central StatisticalOrganisation, Ministry of Statistics and Prograinine Implementation,New Delhi. (Various Issues).Government of India (1991), Spices Statistics, Spices Board, Ministry ofCommerce, Kochi, Kerala. (Various Issues).Government of India (2002), Arecanut and Spices Database, Directorate ofArecanut and Spices Development, Department of Agriculture andCooperation, Ministry of Agriculture, Calicut, Kerala. (Various Issues).Government of India (2003), lndia - 2003, A Reference Annual - Compiledand Edited by Research, Reference and Training Division - PublicationDivision, Ministry of Information and Broadcasting.Government of Kerala (1970- 1980), Agricultural Abstract of Kerala,Department of Economics and Statistics, Thiruvananthapuram, Kerala.(Various Issues).Government of Kerala (1970-2002), Economic Review, State Planning Board,Thiruvananthapurain, Kerala (Various Issues).Government of Kerala (1970-2002), Statistics for Planning, Directorate ofEconoinics and Statistics, Thiruvananthapuram, Kerala. (VariousIssues).Government of Kerala (1 992- 1993), Agricultural Statistics of Kerala, (VariousIssues).International Pepper Corninunity (1 980- 1999, International Pepper NewsBulletin, Jakarta, Indonesia. (Various Issues).

International Pepper Coininunity (1986- 1996), Pepper Statistical Year Book,Jakarta, Indonesia. (Various Issues).Kerala State Land Use Board, (1997), Land Resource Based Perspective Planfor 2020 AD, Viltas Bhavan, Thiruvananthapuram,National Family Health Survey (2001), (NFHS-2), India 1998-99 - Kerala -International Institute for Population Sciences - Mumbai, India,December.Sivaraman, K, M.S. Madan and M.Tarni1 Selvan (2002), Black Pepper Guide,Directorate of Arecanut and Spices Developmel~t, Ministry ofAgriculture, Government of India, Calicut, Icerala.Spices Board (1993), "Know your Spices", Indian Spices, Vol.30, No.2&3,Kochi.Spices Board (1999-2000), Spices Export Review, Ministry of Commerce,Government of India, Cochin, Kerala.Stifel, Laurence D. (1976), "Imperfect Competition in a Vertical MarketNetwork: The Case of Rubber in Thailand", Singapore: Teaching andResearch Forum - A Reprint, No.3.