Principal-Agent Problems in Fisheries - lah@sam.sdu.dk

42Principal-Agent Problems in FisheriesNIELS VESTERGAARD42.1. INTRODUCTIONA “principal-agent problem” arises whenever anindividual or public agency or regulator (the principal)has another person, office, or firm (the agent)perform a service on its behalf and cannot fullyobserve the agent’s actions, inducing informationasymmetry. In economics, the traditional example isthe potential conflict of interest between ownershipand management, but any delegation of authoritymay give rise to this problem. The classic examplein fisheries is the relationship between vessel ownerand crew members. Principal-agent theory focuseson mechanisms to reduce the “problem” of asymmetricinformation, such as defining and selectingthe “right” types of agents, implementing incentivecontracts, including instituting forms of monitoringand various amounts of positive (“carrots”)and negative sanctions (“sticks”). The underlyingassumption is that the agent’s interests may differfrom those of the principal. Strategies to mitigatethe problem produce a type of transaction cost,reflecting the fact that without cost, it is impossiblefor principal to be sure that agents will act in theprincipal’s best interest. In other words, Pareto efficientallocation cannot be obtained.There have been many areas of applicationsof the principal-agent paradigm in economics. Inan employer-labor setting, an employer serves therole of principal, while workers act as agents. Theclassic example of the principal-agent relationshiphas a landlord monitoring the activities of a tenantfarmer. In regulated industries, the regulator acts asprincipal, designing an incentive scheme or contractfor the firms (agents) whose activities are being regulated.From the insurance sector, the policyholdermight change behavior after becoming insured.The applications of principal-agent theory torenewable ocean resources, including fisheries,have been diverse and in some instance superficial.To understand this, let us look at the areaof regulation. The economics of regulation can bedivided into two main approaches: public interesttheory and interest group theory. In public interesttheory, the regulations promote overall publicinterest, while interest group theory views the purposeof regulation as promoting narrow actionsand interests of certain groups in the society. Thegeneral setup is that the government wants firmsand/or households to adjust their behavior in sucha way that certain things are prevented or somespecific goals reached. In reality, it is very difficultfor the government to control the actions takenby firms and/or households. Figure 42.1 shows avery simplistic interaction between different actorsin the regulation of fisheries. The fisheries policy isformulated by the politicians and implemented bythe ministry of fisheries. An important part of theministry of fisheries is the section of control andenforcement because this section interacts with thecourts to determine the final interpretation of thelaws. So, it is not sure that the intentions of the563

564Policy Instruments and Perspectivespoliticians are put into practice. The fishing firmsmay consist of the owners, managers (e.g., the skipper),and a range of employees or crew members.The skipper and the crew members are actuallydoing the fishing, not necessarily the owner. Eventhough the figure is missing the consumer side, it issufficient to show that the links between the politiciansat the one hand and the skipper and crew onthe other hand are imperfect. When the government(principal) cannot precisely control the fishermen(agent) has been termed a principal-agent problem.In the political economy literature, the situationin which a state agency (agent) pursues its owninterest, which may differ from the interest of thegovernment (principal), is called a principal-agentproblem. Here the problem may be connected tothe rent-seeking issue that arises when agents fromthe regulated industry try to influence the regulationin their favor (e.g., lobbying using resourcesto acquire subsidies). Both of these problems arisewith conflicting objectives and strategies betweenagents and the regulators, and they both contributeto government failure (for an overview, seeEdwards 1994).The problem with this approach to regulation isthat although it might emphasize that actors whoare going to be regulated by the government cannotbe expected to adjust passively to the regulation, itdoes not explicitly state the asymmetric informationproblem, which is a market failure (see Hanleyet al. 1997). In fact, it is the opinion of the authorthat studies of the interplay between a principal andan agent or the regulator and the regulated shouldhave some kind of asymmetric information problemattached before it can be called a principal-agentstudy. Otherwise, it is “just” a study where the regulatortakes the reaction by the agent into accountPoliticiansMinistry/regulator(control and enforcement)CourtLobbyingOwnerManager/skipperCrewFIGURE 42.1 Interactions among the politicians, thegovernment, and the fishing firmswhen the regulation is designed, which every soundregulation model has to do. In fact, several of theapplications of the principal-agent model in fisheriesdo not include the information problem, and hencecannot be called principal-agent analyses (see section42.3). 1 Jensen (2008) provides a recent reviewon uncertainty and asymmetric information in thefisheries regulation economics literature, where thepurpose is to distinguish when uncertainty can beinterpreted as asymmetric information.This chapter presents principal-agent theory,with an emphasis on identifying the instances whenit is suitable to apply the principal-agent approachand what are the gains in knowledge. It then surveysdifferent applications of the principal-agentapproach in fisheries.42.2. PRINCIPAL-AGENTTHEORYKwerel (1977) wrote:In a world of perfect information, optimal regulationof an isolated economic variable would berelatively straightforward. Unfortunately, we donot live in such a world. Regulatory authoritiestypically find that the information which theyneed during the planning phase is known only bythose who are to be regulated. In this situationa serious incentive problem may arise. Unless asystem can be designed which makes the objectivesof the individual agents coincide with theregulator’s objectives, self-interested agents willsystematically deceive the regulatory authoritywhen asked to reveal their information.The economics of information and incentives hasdeveloped greatly in recent years. In the economictheory of contracts, agents are characterized byprivate information that determines their ultimateactions. The two key features of principal-agentproblems are that (1) the principals know less thanthe agents about something important, and that (2)their interests conflict in some way (see Sappington1991; two more recent introductions are Macho-Stadler and Pérez-Castillo 2001; Laffont and Martimort2002).There seem to be two sources of asymmetricinformation problems:• Problems where agents can do some costlyaction to improve outcomes for the principal,

Principal-Agent Problems in Fisheries 565but the principal cannot observe the action.These are known as moral hazard or hiddenaction problems. For example, this caninvolve effort in a production process or protectionagainst risk.• Problems where there are different types ofagents, and principals cannot distinguishbetween them. 2 These are known as adverseselection or hidden information problemswhen the types are fixed and the question iswhich agents will participate. For example,this can include efficiency in terms of cost orwillingness to pay for a given good.While the type of the agent is unknown to the principal,the latter nevertheless is assumed to haveprior information before the contract is negotiated,in terms of the statistical distribution of the typeand other relevant characteristics of the agent. Thechallenge for the principal is to set up a contractscheme enforcing truthful revelation of the privateinformation, thus allowing a second-best optimumto be attained for the economic variable of interest(production level, environmental externality, etc.).This is called the revelation principle. 3 In practice,contracts are largely used in domains such as insurance,public regulation, industrial relationships,agricultural production, and employment procedures.This suggests that asymmetric informationcan be present in many situations and must betaken into account in almost every regulatory andempirical analysis.The timing of the interactions between principaland agent is shown in figure 42.2. A contractis offered to the agent, which the agent can eitheraccept or refuse. If the agent refuses, the interactionstops. If the agent accepts, the contract is executed.The sequence of decisions is important because ateach point in time the agent and principal have tomake decisions based on the available information.Therefore, a correct classification of the asymmetricinformation situation is important (see Macho-Stadler and Pérez-Castillo 2001).The advantage of the principal-agent approachis that the incentives at stake are highlighted. Thefocus is on the regulation problem, where the issueis to align the incentives of the private firms so thesocial objectives can be met. Put in another way, infisheries, if the reactions of the fishermen are nottaken into account when formulating the fisherypolicy of keeping, for example, the stock size withinoptimal size (e.g., assume that the fishermen passivelyadjust to quota settings), then the policy canlead to suboptimal and unexpected results. Anotheradvantage of the principal-agent approach is thatthe information problems are explicitly taken intoaccount when setting up the incentive scheme. Theprincipal submits to the agents a menu of contractsthat are conditioned on what both the principaland the agent can observe and on what can be verifiedin the legal system. In other words, asymmetricinformation is included explicitly, because thecontract cannot be conditioned on what only theagents know.Adverse SelectionAgentknowshis typePrincipaloffers acontractAgentacceptorrefusesThecontractisexecutedMoral HazardPrincipaloffers acontractAgentaccept orrefusesAgentexerts aneffort ornotTheoutcomeis realizedand thecontract isexecutedFIGURE 42.2 Timing of contract offers and interactions between theprincipal and the agent

566Policy Instruments and Perspectives42.3. APPLICATIONS OFPRINCIPAL-AGENT THEORY TOFISHERIESThe principal-agent model can be applied to differentareas in fisheries such as share contracts, internationalfishing agreements, illegal landings anddiscards, safety, and invasive species.42.3.1. Share ContractsShare contracts are used as the main remunerationsystem in commercial fisheries all over the world.The fishermen or crew are paid not a fixed wage perhour but as a share of the revenue obtained by thecatch of the vessel. These systems vary across countries,vessel types, and fisheries. In some systems,some of the common variable costs are deductedbefore the revenue is divided by the crew and thevessel owner, and in some systems the crew receivessome minimum wage combined with a share of therevenue obtained.Sutinen (1979) analyzed the social desirabilityof the share system and concluded that “the sharesystem of remuneration is viewed as making a significantpositive contribution to the developmentof a fishing industry.” This result is driven by riskaversevessel owners and competitive labor markets.Under the share system, the vessel owners canspread some of the risk to the crew, reducing therisk cost. Sutinen does not explicitly mention themoral hazard problem due to private informationabout the effort level of the crew, and he does notcall the approach a principal-agent analysis. However,he writes that the share system “provides awork incentive that makes it less costly to extractthe desired level of labour services from the crew.”The analysis is done in a static framework.In two studies, Hämäläinen et al. (1986,1990)analyzes the issue of share-fishing in a principalagentand a dynamic fishery model. They use theterm “Stackelberg game” as interchangeable withthe principal-agent model because the principal(called leader in the Stackelberg game) defines thegame and takes into account the reaction of theagent (called follower in the Stackelberg game).Because the principal defines the game, the reactionsof the agents are passive in the sense thatthey, given the choice of the principal, optimizetheir behavior. In this model, the agents define theirlabor supply as a function of the share contract andthe stock level. In both studies, Hämäläinen et al.use a dynamic fishery model in a principal-agentsetup. Again, like other applications, there is noinformation problem in the model, because thereis no asymmetric and private information. Theyargue that the reason for setting up the model inthis way is that a contract based on labor time isnot enforceable because of the high transactionscost in, for example, supervision of labor. The issueof risk sharing is not included in the model. In their1986 paper, Hämäläinen et al. assume myopicbehavior, because the owners and the fishermen arenot coordinating their actions. In the case of a pureshare contracts, they found, with myopic behaviorunder open access, that the share is equal to theelasticity of harvest with respect to labor input.They analyzed—given the share system—the possibilitiesto induce social optimal fishing and showedthat a combination of licenses and taxation of harvestmay lead to the optimal solution. In their 1990paper, Hämäläinen et al. studied a fishery wherecooperative vessel owners (the principal) hire unorganizedfishermen (agents) to operate vessels. Oneach vessel, the fishermen’s salary is a share of thevalue of the catch. The results show that harvestshares of myopic fishermen will be reduced whencartels are established. The results suggest thatshare-fishing is a self-adaptive and time-consistentremuneration system as it automatically incorporatesdifferences in individual crews’ labor suppliesdue to, among other things, fishermen’s skills. Theyfound that optimal regulation is accomplished by aconstant subsidy on the price of fish.Stanley (2007) studies relative payment contractsin a principal-agent framework between larvagatheringagents and their boat-owning employerprincipals who supply seed to shrimp farms. Theeffort level by the agents is not observable by theboat owners, and the output of the agents is relatedto the effort level and the unknown and stochasticenvironmental factors. A relative payment contractis designed such that the payment depends on therelative performance of the agents. That is to say, ifthe production of an agent is higher than the averageproduction of all agents, the agent receives abonus payment depending on production level andon whether the vessel owner is earning profit. Thestudy incorporates the sources of production riskcreating income shocks to gatherers. Two data setsfrom a Honduran coastal fishery case were usedto test the hypotheses concerning contractual performanceacross environments. Which contractprovides the highest mean income (and variation)

568Policy Instruments and Perspectivesstock, and the probability of fishermen being eitherlow cost or high cost.42.3.3. SafetyThe issue of safety in fisheries is analyzed by Berglandand Pedersen in 1997. The background is thatit is necessary for the public authorities to engage insafety issues to secure a supply of communicationand navigation systems used by all vessels, becausethese are public goods. In addition, to reduce accidentrisks, the supply of these public services willincrease the production possibilities in the fishingindustry. However, they found that moral hazardeffects might occur because the supply of public servicescould induce the individual rational fishermento behave in a way that increases risks because theymay insert less of the safety-reducing and more ofthe safety-increasing private inputs as long as thetotal losses experienced when accidents happen arehigher than the losses each of the fishermen are facedwith. Their analysis has no information problem;the market failure in the model is the public goodcharacteristic of safety. As moral hazard is definedin this survey, the adjustment of the fishermen tothe provision the public good is not hidden.42.3.4. Discards and IllegalLandings: Moral HazardJensen and Vestergaard (2002b) treat output regulationin fisheries as a moral hazard problem, becausethe fishermen have private information about theircatches. What the regulator normally can observeunder reasonable cost is the landings. Observingcatches is in many cases too costly. 5 With catchesas private information, there is an incentive problem,because information about the real catches isimportant of several reasons. First, if the total realcatch is higher than the level set by the regulator,this might represent an unsustainable harvest level,leading to direct long-term economic loss. Second,because the real catch is unknown, the assessmentof the stock level will—all things equal—be moreuncertain, and again, this can lead to situationswhere the assessed the stock level is different thanthe real stock size—a problem that the InternationalCouncil for the Exploration of the Sea (ICES) hasbeen dealing with for several years (see, e.g., ICES2007). Third, all this leads to implementation ofthe comprehensive control and enforcement systemto ensure that the total catches are held within thetotal allowable catches. In other words, the informationabout the variable—catches—is typicallyknown only by the fishermen, who are to be regulated.Self-interested fishermen will systematicallydeceive the regulator, when asked to reveal theirinformation about catches, unless a system can bedesigned that aligns the motives of the fishermenwith the social objectives. Using results from thenonpoint pollution literature (Segerson 1998), anincentive scheme is formulated based on the state ofthe stock. The basic assumption is that in the caseof many fishermen, asymmetric information anduncertainty the regulation of fisheries are a complexproblem, where the assessment of each fisherman’sreal catches is prohibitively costly. This leadsto defining the problem as moral hazard in groups.Solving this problem in the fishery case results informulating the policy instrument as a function ofthe state of stock biomass. The tax rate (the policyinstrument in this case) is equal to the expectedmarginal net social cost from exceeding the optimalcatch divided by the fisherman’s biological response.The tax structure eliminates “free-riding,” becausethe fishermen pay on the basis of the full marginalcosts that illegal landings generate (the difference inuser costs). In this way, compliance with the totalquota is ensured—the incentive for illegal landingsis avoided.Hansen et al. (2006) address the informationproblems in the mechanism proposed by Jensenand Vestergaard (2002b). Instead of grounding thetax rate on stock size and individual fishermen’scost function, Hansen et al. (2006) suggest a taxthat depends on knowledge of the principal of theaggregate cost function and total catches of allfishermen. The tax is a function of the aggregatecost function and the fishermen’s ex ante reportsof planned catch. Hansen et al. show that this taxsystem will secure nearly optimal catches, leastcostproduction, and nearly optimal entry-exitincentives.Jensen and Vestergaard (2007) analyze the issueof discards and illegal landings and moral hazardwhen there is uncertainty in stock size. The purposeof the study is to analyze incentive contracts as asolution to the stock externality problem, the problemof imperfect and private information aboutcatches and the stock uncertainty problem. So, theincentive problem due to private information aboutcatches, where landings or self-reported catches(e.g., in logbooks) are observable, is addressed in arealistic situation where there is uncertainty about

Principal-Agent Problems in Fisheries 569the stock size. In other words, there are multiplemarket failures, and it is well known that with severalmarket failures, multiple policy instrumentsmust be used to secure a first-best optimum. Theincentive contract offered to the risk-adverse fishermenconsists of the two taxes: (1) a stock tax basedon the difference between the target year-end stocksize and the expected stock size at the end of theyear multiplied by an unit stock tax rate that is adeclining function of the self-reported catches; and(2) voluntary self-reported catches in logbooks aretaxed using a tax-rate in terms of per unit of individual,voluntary self-reported of catches in the fishingperiod. Given stock uncertainty and risk aversion,the stock tax rate and the self-reported tax-rateboth depend on the variance of the uncertain stocksize and the risk-aversion function. Further, there isan interaction between the two tax rates. The reasonis that tax rates have to be balanced againsteach other on the margin, because the stock taxrate is a declining function of self-reported catches.The uncertainty and risk aversion make it favorablefor the fishermen to have a positive level of theself-reported tax rate. The use of a stock tax alonein a situation with uncertain fish stocks and riskaversefishermen allows for a second-best optimum.In this situation, it is not possible to reach a fulloptimum. However, it is shown in the analysis, ifin addition to the stock tax a tax on voluntary selfreportedcatches is imposed, a first-best optimumis reached.42.3.5. Invasive SpeciesMacPherson et al. (2006) presents a dynamicprincipal-agent model of aquatic species invasionsin which a manager, who is concerned about thespread of invasive species across lakes by boaters,sets management controls on a lake-by-lake basis,and boaters make a series of trip decisions duringthe course of the season based on the controlsimposed by the manager. The results of a simulatedinvasion of Eurasian watermilfoil (Myriophyllumspicatum) highlight interesting aspects of theoptimal management policies under two differentmanagement objectives: maximizing boater welfareand minimizing milfoil spread. As such, this studyendogenizes resource user behavior in the managementdecisions related to a plant species invasion byallowing the lake manager, the principal, to anticipateboater reaction to management activity andthe invasive species. There are two fundamentalreasons to endogenize boater movements in a modelof the spread of aquatic invasions by boaters: (1) toprovide a better forecast of the rate and direction ofspread of the invader and (2) to accurately estimatewelfare effects. The welfare effect of closing a lakehas a direct welfare effect because now access tothe lake requires keeping a boat on the lake, but italso has an indirect welfare effect because it shiftsthe spread of the invasive, which may leave societyworse off overall. In the analysis there are no asymmetricor private information problems. Therefore,the approach taken does not belong to the principal-agentapproach under asymmetric informationand incentive contracts, but the approach is connectedto the more general principal-agent principleunder the economics of regulation mentioned in theintroduction.42.4. CONCLUSIONThis overview shows that applications of the principal-agenttheory to fisheries have been diverseand not very systematic. The overview also demonstratesthat the use of term “principal-agent modelapproach” has not been consistent in the fisheriesliterature. The traditional principal-agent approachconsiders asymmetric information as a marketfailure problem in the general relationship wherethe principal wants to induce the agent or agentsto do certain things. Several applications have notincluded the information problem, which changesthe research approach toward game theory. Froma policy management point of view, regulatingassuming full information will create an inefficientallocation of resources, and therefore the regulatorhas to create an incentive compatible regulationwhere the agents will reveal their private information.This has a cost, but the allocation is betterthan just going ahead with regulation assuming fullinformation.The state of the research so far indicates thatmore work needs to be done with respect to theinformation requirements for regulation and henceto be much more explicit about what a contractbetween the regulator and the fishermen may bebased on. There is a lot of potential in this area,because regulation of fisheries in many countriesis based on licenses that easily could be extendedto include more items, for example, fishing behavior.Another area is signaling, where fishermenmight be able to signal their type cost-free and in

570Policy Instruments and Perspectivesthis way reduce the information problem. Further,the issue of designing a fishing contract between acoastal state and a distant state has not been fullyexplored.Two other areas where the principal-agentapproach could be applied to in fisheries are invasivespecies and marine reserves. Invasive speciesmay in some marine areas threaten the marineenvironment and biodiversity. As a preventive measure,agents that are assumed to have a high riskof transferring species from one marine area toanother area could be offered an incentive contractwhere they are compensated for reducing the riskof moving species to the foreign areas. With respectto marine reserves, the regulator might compensatethe fishermen for their loss due to establishment ofthe reserve, but at the same time they might haveless information about the cost and/or actions ofthe fishermen, raising the problem of setting up aproper incentive scheme (see for a very recent application,Quach 2008).Acknowledgments I thank Knud Sinding andAnders Skonhoft for helpful comments and suggestionsto an earlier draft of this chapter.Notes1. One could speculate whether asymmetricinformation between different units in the governmentis market failure or government failure. It isnot important for our purpose, because we wantto look at cases where objectives and asymmetricinformation conflict.2. In some literature, there is a distinctionbetween whether the agent knows something apriori or will know something (see Mas-Colellet al. 1995). In fact, moral hazard issues and thesituation when the agent is coming to know morethan the principal is the original sense of the principal-agentproblem. Adverse selection then refersto the case where information about some characteristicsof commodities is not observable by all theparticipants. However, in recent years the way weuse the term seems to be the most used.3. The revelation principle says (in words)that whatever outcome you can achieve, you canachieve by giving the agent an incentive to tell thetruth, and you do not lose anything by making theagent tell the truth.4. In Macho-Stadler and Pérez-Castillo (2001,section 3B.2), the same moral hazard modelbetween a coastal and distant state is formulatedwith asymmetric information based on Gallasteguiet al. (1993). However, since the model is formulatedon observed catches, it leads to paradoxicalresults, where higher catches than allowed arethrown back into the sea.5. However, in some fisheries a 100 percentobserver program is in place to ensure that thecatches are reported correctly.ReferencesBergland, H., and P.A. Pedersen (1997). Catch regulationand accident risk: The moral hazard offisheries. Marine Resource Economics 12(4):281–292.Clarke F.H., and G.R. Munro (1987). Coastal states,distant water fishing nations and extendedjurisdiction: A principal-agent analysis. NaturalResource Modeling 2: 81–107.Edwards, S.F. (1994). Ownership of renewableocean resources. Marine Resource Economics.9(3): 253–273.Gallastegui, M.C., E. Iñarra, and I. Macho-Stadler(1993). Contratos de pesca desde la perspectivade la teoría de la agencia. Estudios deEconomía 20(2): 329–354.Hämäläinen, R.P., J. Ruusunen, and V. Kaitala(1986). Myopic Stackelberg equilibria andsocial coordination in a share contract fishery.Marine Resource Economics 19: 175–192.Hämäläinen, R.P., J. Ruusunen, and V. Kaitala(1990). Cartels and dynamic contracts insharefishing. Journal of Environmental Economicsand Management 19: 175–192.Hanley, N., J.F. Shogren, and B. White (1997). EnvironmentalEconomics in Theory and Practice.Houndmills, U.K.: Palgrave Macmillan.Hansen, L.G., F. Jensen, U.S. Brandt, and N. Vestergaard(2006). Illegal landings: An aggregatecatch self-reporting mechanism. American Journalof Agriculture Economics 88(4): 974–985.ICES (2007). Cod in Skagerak. ACFM Advice.International Council for the Exploration ofthe Sea. www.ices.dk/committe/acom/comwork/report/2007/may/cod-kat.pdfJensen, F. (2008). Uncertainty and asymmetricinformation: An overview. Marine Policy 32:89–103.Jensen, F., and N. Vestergaard (2001). Managementof fisheries in the EU: A principal-agentanalysis. Marine Resource Economics 16(4):277–291.Jensen, F., and N. Vestergaard (2002a). A principal-agentanalysis of fisheries. Journal of Institutionaland Theoretical Economics 158(2):276–285.Jensen, F., and N. Vestergaard (2002b). Moral hazardproblems in fisheries regulation: The case

Principal-Agent Problems in Fisheries 571of illegal landings and discard. Resource andEnergy Economics 24(4): 281–299.Jensen, F., and N. Vestergaard (2007). Asymmetricinformation and uncertainty: The usefulness oflogbooks as a regulation measure. EcologicalEconomics 63: 815–827.Kwerel, E. (1977). To tell the truth: Imperfect informationand optimal pollution control. Reviewof Economic Studies 44: 595–601.Laffont, J.-J., and D. Martimort (2002). The Theoryof Incentives. Princeton, N.J.: PrincetonUniversity Press.Macho-Stadler, I., and J.D. Pérez-Castillo (2001).An Introduction of the Economics of Information.New York: Oxford University Press.MacPherson, A.J., R. Moore, and B. Provencher(2006). A dynamic principal-agent model ofhuman-mediated aquatic species invasions.Agricultural and Resource Economics Review35(1): 114–154.Mas-Colell, A., M.D. Whinton, and J.R. Green(1995). Microeconomic Theory. New York:Oxford University Press.Quach, N.T.K. (2008). Creation of Marine Reserveand Incentives for Biodiversity Conservation.Unpublished manuscript, Norwegian Collegeof Fishery Science, University of Tromso.Sappington, D.E.M. (1991). Incentives in principalagentrelationships. Journal of Economic Perspectives5: 45–66.Segerson, K. (1988). Uncertainty and incentives fornonpoint pollution control. Journal of EnvironmentalEconomics and Management 15: 87–98.Stanley, D.L. (2007). Risk management in gatheringeconomies. Journal of Development Studies43(6): 1009–1036.Sutinen, J. (1979). Fishermen’s remuneration systemsand implications for fisheries development.Scottish Journal of Political Economy26(2):147–162.