42 Principal-Agent Problems in Fisheries NIELS VESTERGAARD 42.1. INTRODUCTION has a landlord monitoring the activities of a tenant farmer. In regulated industries, the regulator acts as A “principal-agent problem” arises whenever an principal, designing an incentive scheme or contract individual or public agency or regulator (the prin- for the fi rms (agents) whose activities are being reg- cipal) has another person, offi ce, or fi rm (the agent) ulated. From the insurance sector, the policyholder perform a service on its behalf and cannot fully might change behavior after becoming insured. observe the agent’s actions, inducing information The applications of principal-agent theory to asymmetry. In economics, the traditional example is renewable ocean resources, including fi sheries, the potential confl ict of interest between ownership have been diverse and in some instance super- and management, but any delegation of authority fi cial. To understand this, let us look at the area may give rise to this problem. The classic example of regulation. The economics of regulation can be in fi sheries is the relationship between vessel owner divided into two main approaches: public interest and crew members. Principal-agent theory focuses theory and interest group theory. In public inter- on mechanisms to reduce the “problem” of asym- est theory, the regulations promote overall public metric information, such as defi ning and selecting interest, while interest group theory views the pur- the “right” types of agents, implementing incentive pose of regulation as promoting narrow actions contracts, including instituting forms of monitor- and interests of certain groups in the society. The ing and various amounts of positive (“carrots”) general setup is that the government wants fi rms and negative sanctions (“sticks”). The underlying and/or households to adjust their behavior in such assumption is that the agent’s interests may differ a way that certain things are prevented or some from those of the principal. Strategies to mitigate specifi c goals reached. In reality, it is very diffi cult the problem produce a type of transaction cost, for the government to control the actions taken refl ecting the fact that without cost, it is impossible by fi rms and/or households. Figure 42.1 shows a for principal to be sure that agents will act in the very simplistic interaction between different actors principal’s best interest. In other words, Pareto effi - in the regulation of fi sheries. The fi sheries policy is cient allocation cannot be obtained. formulated by the politicians and implemented by There have been many areas of applications the ministry of fi sheries. An important part of the of the principal-agent paradigm in economics. In ministry of fi sheries is the section of control and an employer-labor setting, an employer serves the enforcement because this section interacts with the role of principal, while workers act as agents. The courts to determine the fi nal interpretation of the classic example of the principal-agent relationship laws. So, it is not sure that the intentions of the 563 564 Policy Instruments and Perspectives politicians are put into practice. The fi shing fi rms when the regulation is designed, which every sound may consist of the owners, managers (e.g., the skip- regulation model has to do. In fact, several of the per), and a range of employees or crew members. applications of the principal-agent model in fi sheries The skipper and the crew members are actually do not include the information problem, and hence doing the fi shing, not necessarily the owner. Even cannot be called principal-agent analyses (see sec- though the fi gure is missing the consumer side, it is tion 42.3).1 Jensen (2008) provides a recent review suffi cient to show that the links between the politi- on uncertainty and asymmetric information in the cians at the one hand and the skipper and crew on fi sheries regulation economics literature, where the the other hand are imperfect. When the government purpose is to distinguish when uncertainty can be (principal) cannot precisely control the fi shermen interpreted as asymmetric information. (agent) has been termed a principal-agent problem. This chapter presents principal-agent theory, In the political economy literature, the situation with an emphasis on identifying the instances when in which a state agency (agent) pursues its own it is suitable to apply the principal-agent approach interest, which may differ from the interest of the and what are the gains in knowledge. It then sur- government (principal), is called a principal-agent veys different applications of the principal-agent problem. Here the problem may be connected to approach in fi sheries. the rent-seeking issue that arises when agents from the regulated industry try to infl uence the regula- 42.2. PRINCIPAL-AGENT tion in their favor (e.g., lobbying using resources to acquire subsidies). Both of these problems arise THEORY with confl icting objectives and strategies between agents and the regulators, and they both contrib- Kwerel (1977) wrote: ute to government failure (for an overview, see Edwards 1994). In a world of perfect information, optimal regu- lation of an isolated economic variable would be The problem with this approach to regulation is relatively straightforward. Unfortunately, we do that although it might emphasize that actors who not live in such a world. Regulatory authorities are going to be regulated by the government cannot typically fi nd that the information which they be expected to adjust passively to the regulation, it need during the planning phase is known only by does not explicitly state the asymmetric informa- those who are to be regulated. In this situation tion problem, which is a market failure (see Hanley a serious incentive problem may arise. Unless a et al. 1997). In fact, it is the opinion of the author system can be designed which makes the objec- that studies of the interplay between a principal and tives of the individual agents coincide with the an agent or the regulator and the regulated should regulator’s objectives, self-interested agents will have some kind of asymmetric information problem systematically deceive the regulatory authority attached before it can be called a principal-agent when asked to reveal their information. study. Otherwise, it is “just” a study where the reg- ulator takes the reaction by the agent into account The economics of information and incentives has developed greatly in recent years. In the economic theory of contracts, agents are characterized by Politicians private information that determines their ultimate Lobbying actions. The two key features of principal-agent Owner problems are that (1) the principals know less than Ministry/regulator the agents about something important, and that (2) their interests confl ict in some way (see Sappington (control and enforcement) Manager/skipper 1991; two more recent introductions are Macho- Stadler and Pérez-Castillo 2001; Laffont and Mar- timort 2002). Court There seem to be two sources of asymmetric Crew information problems: FIGURE 42.1 Interactions among the politicians, the • Problems where agents can do some costly government, and the fi shing fi rms action to improve outcomes for the principal, Principal-Agent Problems in Fisheries 565 but the principal cannot observe the action. The timing of the interactions between princi- These are known as moral hazard or hid- pal and agent is shown in fi gure 42.2. A contract den action problems. For example, this can is offered to the agent, which the agent can either involve effort in a production process or pro- accept or refuse. If the agent refuses, the interaction tection against risk. stops. If the agent accepts, the contract is executed. • Problems where there are different types of The sequence of decisions is important because at agents, and principals cannot distinguish each point in time the agent and principal have to 2 between them. These are known as adverse make decisions based on the available information. selection or hidden information problems Therefore, a correct classifi cation of the asymmet- when the types are fi xed and the question is ric information situation is important (see Macho- which agents will participate. For example, Stadler and Pérez-Castillo 2001). this can include effi ciency in terms of cost or The advantage of the principal-agent approach willingness to pay for a given good. is that the incentives at stake are highlighted. The focus is on the regulation problem, where the issue While the type of the agent is unknown to the prin- is to align the incentives of the private fi rms so the cipal, the latter nevertheless is assumed to have social objectives can be met. Put in another way, in prior information before the contract is negotiated, fi sheries, if the reactions of the fi shermen are not in terms of the statistical distribution of the type taken into account when formulating the fi shery and other relevant characteristics of the agent. The policy of keeping, for example, the stock size within challenge for the principal is to set up a contract optimal size (e.g., assume that the fi shermen pas- scheme enforcing truthful revelation of the private sively adjust to quota settings), then the policy can information, thus allowing a second-best optimum lead to suboptimal and unexpected results. Another to be attained for the economic variable of interest advantage of the principal-agent approach is that (production level, environmental externality, etc.). the information problems are explicitly taken into This is called the revelation principle.3 In practice, account when setting up the incentive scheme. The contracts are largely used in domains such as insur- principal submits to the agents a menu of contracts ance, public regulation, industrial relationships, that are conditioned on what both the principal agricultural production, and employment proce- and the agent can observe and on what can be veri- dures.