“Economic analysis of law”

Higher School of Economics

Moscow, Russia

June 2005

Essay

“Adverse Selection and Moral Hazard
in Contract Law”

by

Nicole Petrick

 

 

 

Table of Contents

1  Introduction. 3

2  Legal and economical interpretations. 4

a)  Contract, contract law and contract theory. 4

b)  Asymmetric information. 5

c)  Adverse selection. 5

d)  Moral Hazard. 6

3  Negative effects of adverse selection and moral hazard. 8

a)  Transaction costs. 8

b)  Incomplete contracts. 9

4  Incentives to avoid adverse selection and moral hazard. 11

a)  Signaling. 11

b)  Deductible provision. 12

c)  Indemnity contracts versus valued contracts. 14

5  Summary. 15

References. 16

 

 

1  Introduction

To an economist a contract is an agreement under which two parties make reciprocal commitments in terms of their behavior, thus being a “bilateral coordination arrangement”. This formulation touches on the legal concept of the contract but also transcends it. Over the last thirty years, the “contract” has become a central notion in economic analysis, giving rise to three principal fields of study: “incentives”, “incomplete contracts” and “transaction costs”.

This paper will focus on the topic of adverse selection and moral hazard in contract law. While adverse selection, a problem involving hidden information, anticipates ex ante, moral hazard, a problem involving hidden action, reveals an ex post phenomenon. Both imply negative effects on contract efficiency and prices and touch either field of study: incentives, incomplete contracts and transaction costs.

The paper will approach the topic by defining the objective of each both adverse selection and moral hazard first, starting from a definition on both and including definitions on contract, contract law and contract theory.

Negative effects, such as the effect on transaction costs as well as the influence of incomplete contracts will be talked about in the third part.

Contractual and economic solutions such as signaling, deductibles as well as indemnity and valued contracts in order to minimize or avoid both adverse selection and moral hazard is been looked upon thereafter.

 

2  Legal and economical interpretations

a)  Contract, contract law and contract theory

A contract is any promise or set of promises made by one party to another. In case of a breach of contract, the law provides a remedy. The promise or promises may be expressed either written or oral or may be implied from circumstances.

Where a contract is a single connection between parties, contract law is the portion of civil law that interprets written agreements between parties and resolves disputes between them in general.

Contract theory in its place is the body of legal thought that investigates normative and conceptual problems in contract law. The central problem of contract theory is the question, “Why are contracts enforced”. A prominent answer to this question focuses on the economic benefits of enforcing bargains. Contract theory bases itself on the principle of pareto-efficiency, where the pareto-optimum shall be achieved.

Contract law itself has several purposes to achieve efficiency also such as to enable people to cooperate by converting games with a non-cooperative solution into a game with a cooperative solution and thus to achieve efficient results. Optimal commitment to performance and the securing of optimal reliance are other purposes of contract law that shall ensure an efficient outcome. Contract law also supplies default terms that shall be efficient in order to minimize transaction costs of negotiating contracts and the law shall correct market failures by regulating the terms of a contract.

If a contract would be complete and efficient, without gaps and failures contract law would not be necessary, at least not in the sense of settling arguments in court. But a complete contract implements that every possible contingency is anticipated, associated risks are efficiently allocated and all relevant information has been communicated. However, this would imply tremendous transaction costs and thus collide with the 4^th purpose of contract law, which is to minimize those costs for the goal of efficiency.

In Economics, the theory of contracts is a part of the information economics and describes how economic actors use particular contractual arrangements to deal with information asymmetries. Contracts itself or better to say the cooperation and commitment towards a contract can be and are basically modeled in games (bargain theory/ game theory) using the principal-agent approach.

b)  Asymmetric information

Asymmetric information or “information asymmetry” describes the situation where one party in a contract has information which is superior to that of the other party – or in common words: one party knows more than the other party does. It basically describes the failure of the parties to communicate to each other all the relevant information, needed for both parties within the contract. But information asymmetries also occur when information is intentionally hidden (hidden information). Those asymmetries consequently lead to inefficient behavior, due to being the cause of “adverse selection”.

c)  Adverse selection

Adverse selection in its place is a term from the new institution economics and designates a condition, in which on a market unwanted results are systematically obtained. The basic model to describe adverse selection was first derived by George A. Akerlof in 1970, where he used the now well-known “market of lemons” example to demonstrate adverse selection in practice:

Imagine a market for used cars. Staying close to reality it is quite likely that there will be both good and bad quality cars available. For simplicity we assume a distribution of 50% good cars – the “cherries” – and 50% bad cars – the “lemons”. Only people selling cars know the exact quality of the car, while the potential buyers/ the clients are not able to observe the quality of a car to 100% certainty. Once again for simplicity, we even assume that they do not know the quality at all.

Now we suppose that buyers and sellers equal in number and that sellers would like to receive 200 currency units for a good and 100 currency units for a bad car, while buyers are willing to pay 240 currency units for a good and 120 currency units for a bad car. In this case, the market would clear with all cars being sold, giving a pareto-efficient solution.

However, since buyers do not know the quality of a car offered, they calculate the expected value of a car and are not willing to pay more than that expected value. In this example the expected value equals:

P_B=E(V)=0.5 x 240 + 0.5 x 120=180.

The expected value of a car, equaling the maximal price a buyer will be willing to pay is now 180 currency units. This price lies below the required price for a seller to sell his good car (the cherry) which was at 240 currency units. Consequently, sellers will not offer their cherries but only their lemons for a price of 120 each. As a result only 50% of the market volume will clear, leading to an inefficient solution.

The above example from Akerlof perfectly describes and explains adverse selection, where it derives from (information asymmetry) and what a tremendous negative effects it has on the goal of achieving an efficient solution.

As mentioned before, adverse selection is a problem that we face ex ante – before a contract is made. Actually, adverse selection is the hinder to contracts, as seen in the example. Thus it can evade that a contract will not be made although it would have been pareto-efficient and – in a game without information asymmetries – would have led to a cooperative solution.

d)  Moral Hazard

Moral hazard is the name given to the risk that one party to a contract can change their behavior to the detriment of the other party once a contract has been concluded – ex post.

The most well known examples of moral hazard come from insurance. Fire insurance, for instance, gives people an incentive to commit arson, especially if they are operating a failing business and decide that they would rather have the cash from the insurance proceeds on the buildings than the buildings themselves. Many, perhaps most, police investigations of arson are the result of leads from suspicious insurance adjusters. More generally, insurance may encourage riskier behavior, such as sloppy fire prevention.

Thus, moral hazard is a problem of hidden action, supported by information asymmetries analogous to adverse selection. In the context of the principal/agent theory moral hazard derives from the fact that the incentives of both principal and agent may not be perfectly aligned. As an example for an imperfect alignment the costs for the principal may lie below the costs for the agent to fulfill the contract. However, the risk of moral hazard only occurs whenever it is difficult for the principal to monitor the agent.

Similarly to the effect of adverse selection moral hazard hinders the commitment towards a contract. Figure 1 demonstrates this negative effect with the help of the example on insurance coverage.

Figure 1. Moral Hazard and Optimal Hedging

As it can be easily seen, the price of coverage with insurance rises exponentially rather than linear under moral hazard. As a result the game ends in coverage below the coverage without moral hazard and does thus not result in the efficient level. The same holds for contracting in general.

 

3  Negative effects of adverse selection and moral hazard

Those negative effects talked about do not only arise due to the appearance of adverse selection or moral hazard in general but also due to the fact that the former and the latter create spillovers if they want to be avoided.

a)  Transaction costs

Adverse selection, deriving or being caused by asymmetric information can be avoided by the help of collecting additional information, if the other party of the contract is not willing to contribute this information itself.

In Akerlof’s model the buyer then would have to source information on used cars in order to have the ability to know or at least to abstract information of a car’s quality to a certain extent when looking at it.

However, sourcing information comes together with additional costs that are known as transaction costs in contract law and contract theory.

In general it is said that ignorance (not acquiring more information) is rational when the costs of acquiring additional information exceed the expected benefit from gaining it, thus the buyer, referring to the principal-agent theory here the principal, would have to and will source more information as long as the costs of retrieving it are below the costs of allocating the expected loss:

C(I) < C(L) x P(L)

where C… cost function, L… loss, P… probability.

For Akerlof’s example we receive the following:

The value and thus price of a car, determined by the price the seller wants to have equals P_c=200 currency units (cu) for a “cherry” (a good car) and P_l=100 cu for a “lemon” (a bad car). The price a buyer would be willing to pay, his personal valuation of the car, equals V_c=240 cu for a cherry and V_l=120 for a lemon. The loss a buyer would have to allocate if he buys a car that he expected to be a cherry but that turns out to be a lemon:

L = V_c – P_l = 240 – 100 = 140.

We assumed that there are 50% of lemons (p_l) and 50% of cherries (p_c) on the market. Consequently the expected loss of the buyer equals:

P(L) = p_l x L = 0.5 x 140 = 70.

As a result, the buyer will source more information as long as the costs of this additional information do not exceed 70 currency units. In any case, these costs, referred to as transaction costs in contract law and contract theory, are to the detriment of efficiency and might hinder the enforcement of contracts. Ways to avoid adverse selection without rising transaction costs to a very high extent will be talked upon later in this paper.

b)  Incomplete contracts

Contracts or their fulfillment have risks. While some risks are allocated explicitly others may remain silent, meaning that they are not explicitly mentioned.

If a contract remains silent about a risk, the contract is said to have a “gap”. Gaps might be in a contract, because the possibility of a certain risk is not foreseen, those gaps are called “inadvertent gaps”. If the possibility of a risk is rather remote, than the gap not covering this risk is called “deliberate gap”. In either case the contract is said to be “incomplete”.

If parties negotiate explicit terms to allocate risks, they bear a certain transaction cost compared to transaction costs with a positive probability if they leave a gap.

The case of moral hazard challenges contracts in many ways. Firstly, moral hazard is an ex post problem – thus occurs after a contract was already made. Secondly, moral hazard derives from hidden action and is thus not foreseeable for the other party of the contract. Thirdly, hidden action in the case of moral hazard is mostly intentionally so that the party committing hidden action will surely not disclose this information to the other party on any occasion.

The problem occurring is that gap filling would imply to source more information. Thus, referring to the principal-agent theory where the principal is not able to monitor the agent, the principal would have to source more information on the agent.

This might be in most cases not possible, since the agent counteracts in disclosing his intentions. Additionally, sourcing more information is a weight factor on the transaction costs. However, the principal will collect more information as long as the cost of additional information is below the cost of allocating the expected loss:

C(I) < C(L) x P(L)

with C…cost function, I… Information, L…Loss, P…Probability.

If this additional information then leads him to a result that would suppose to further allocate risks in the contract, he will equivalently do so as long as the costs of allocating a risk are below the costs of allocating the expected loss:

C(A_R) < C(A_L) x P(L)

with A_R….allocation of risk, A_L… allocation of loss.

Although efficient default rules enable the parties to minimize transaction costs of negotiating contracts by leaving gaps it might still not hinder negative effects from moral hazard or prevent them at all. Especially in the insurance sector it can be seen that moral hazard does not mean a breach of contract and thus would not lead to actions like “gap-filling by courts” or the hypothetical bargain case.

As for the example of the insurance sector moral hazard appears in the way that an individual, after buying an insurance might become less attentive, knowing that he is well insured or in the case of health insurance makes now use of medical consultation more than before, since his insurance now pays the fees. In any of those cases no contract is breached, although the costs of performing for the principal (the insurance company) are higher ex post as assumed ex ante due to the ex post change in behavior of the agent (the client). Due to this effect of moral hazard with insurance on the insured’s incentive to reduce losses, fewer or at least no full insurance coverage will be available. The intuition behind this is that if insurance causes moral hazard then less insurance limits moral hazard. As a result, less then the pareto-efficient insurance coverage will be offered and the market does not clear.

But how can these effects be limited without reducing the range of coverage?

 

4  Incentives to avoid adverse selection and moral hazard

There are economic as well as pure legal incentives in order to avoid or at least to reduce adverse selection and moral hazard. This chapter chooses signaling, deductibles as well as indemnity and valued contracts to show four of the possibilities to handle the problem.

a)  Signaling

Adverse selection, being a problem of asymmetric information can be overcome by simple communication. However, communication between two parties of a contract is not always as simple. In Akerlof’s model for example the sellers of lemons could communicate just as the sellers of cherries that their car is of high quality. This situation would leave the buyer in the same situation as before when he had no idea of the quality at all. The game would end in the same inefficient result, where only lemons would be traded. In this situation not only 50% of buyers are dissatisfied but also 50% of sellers – the ones that wanted to sell their cherry. Now we see that the sellers (our agents in this case) would have an incentive to signal the buyers (the principal) the quality of their car. In praxis this could be easily done by giving a warranty on the car. Cherry sellers for example could offer a 6 months warranty for any problems that would occur. Lemon sellers will not do so, because they fear that they truly would have to fulfill the warranty.

Speaking in the context of contracts, the agent who sells a cherry would be willing to allocate risks differently compared to agents, selling lemons. The principal could then abstract information about the car’s quality from the signal the agent sends concerning his willingness to allocate risks.

Signaling, in an economic sense of view, is an action that has economic consequences and that must have different costs for different types. A signal helps the principal to determine true types. The most commonly used example to describe the economics of signaling is the classic example of the job market, the so-called “Spence signaling model”:

On the job market there are two different types of individuals, those with low productivity (L) and those with high productivity (H). While L has a marginal product of g, H has a marginal product of b. There are q fractions of H and (1-q) fractions of L on the market. If firms could distinguish high productivity workers from low productivity workers, workers would be paid according to their productivity, their marginal product. However, in this model firms can not observe the quality of a worker and thus workers are paid the equilibrium wage which is calculated by computing the average marginal product: 

MP = [q x g + (1-q) x b] / [q+(1-q)] = qg+(1-q)b.

Now we assume that workers can acquire some education that does not improve their quality. The costs for an education level e are c_ge and c_be, where c_g< c_b.

If we choose an e such that (g-b)/c_b < e < (g-b)/c_g there will be a separating equilibrium where high productivity workers acquire e, low productivity workers do not acquire education at all. Consequently, education is used as a signal and workers with education are paid g while those without education will be paid b. In this example the equilibrium is still inefficient, since education does not increase productivity and thus output and average wages still remain the same, additionally c_ge is lost. However, the example perfectly describes the principle behind signaling from an economic point of view.

b)  Deductible provision

Another possibility to both avoid or at least to minimize adverse selection and moral hazard comes again from the insurance sector. It is either called co-insurance, insurance-to-value in property insurance or deductible provision. In any variation it is still a re-allocation of risks from the principal to the agent in the case of a loss to the principal, caused by the agent. If for instance the agent hides information (adverse selection case) or hides action (moral hazard case) that will impose ex post a higher cost of performance to the principal than ex ante this deductible provision would allocate a percentage of the costs to the agent so that the incentive of hiding information or action will be reduced.

In an example we assume that e.g. the insurance company (principal) pays 80% of any potential damage, hence the client (agent) would have to pay 20% of the potential damage. Consequently all clients are willing to decrease risky behavior in order to avoid higher costs themselves.

This is analogously adaptable to the contract world in general. Suppose a promisee (the principal) allocates the risk of non-performance or bad performance of the promisor (the agent) to a certain percentage also to the promisor. Then the promisor will have an incentive to commit to performance instead of deviating from the ensured behavior. Thus, deductibles reduce moral hazard.

Additionally, deductibles can also reduce adverse selection. Recalling, adverse selection occurs when the principal can not monitor the agent. As for the case of insurance, at a given price, high risk people will buy more coverage than low risk people will do. In the insurance industry the principal would offer multiple policies with different deductibles and different prices per currency unit of coverage. Thus, lower deductible policies or higher coverage policies will have a higher price per currency unit of coverage. Consequently, higher risk people (the agents) will choose the lower deductible or higher coverage policy where lower risk people will choose the higher deductible or lower coverage but thus lower price policy.

Speaking for contracts in general the principal would offer different contracts that differ in their allocation of risks. For example the agent will be rewarded a higher wage for performance but also a higher penalty for non-performance or bad performance in contrast to a lower wage but less penalty for bad or non-performance. The agent who most likely will deviate from its expressed behavior will choose a contract that allocates fewer risks to him, even if the wage will be lower. Trustworthy agents will choose the contract with higher wage but also with a higher risk allocation towards them. Thus, deductibles also reduce adverse selection.

c)  Indemnity contracts versus valued contracts

Another possibility to reduce moral hazard is the use of indemnity or valued contracts. In both cases the agent pays a certain amount himself.

Within an indemnity contract the agent pays based on the amount of loss that occurred – thus ex post. Controversially, the agent pays a pre-determined (ex ante) amount in the case of a loss within a valued contract. The payment in both cases occurs only in the case of and after a loss; the two contract types differ only over the time at which the amount of payment is determined.

Consequently, the type of contract to choose is largely explained by the costs of assessing the value in case of a loss and the possibility of the occurrence of moral hazard. Thus, if the amount of loss can be assessed at low costs following the loss, an indemnity contract would be the better option. However, if moral hazard is less likely to be a problem, fixing the payment ex ante will avoid costly haggling following a loss. In both cases low transaction costs and thus efficiency lie still in the field of attention.

Indemnity as well as valued contracts have the goal of avoiding moral hazard and also re-allocate not only the risk but even more the loss towards the agent. Both contracts differ only, as said before, over the time when the amount will be set – the re-allocation is done ex post.

 

5  Summary

The paper shows that both adverse selection and moral hazard are not only a phenomenon but a result of economic decision making. Where adverse selection hinders parties from even efficient contracting, caused by rational decision under asymmetric information, moral hazard is the rational but immoral intention of a party to a contract to ex post re-adjust its position to the own advantage but to the detriment of the other party.

In contract law both adverse selection and moral hazard hinder contracting and pose challenges to contract law in order to minimize or avoid the negative extends of the former and the latter.

There are several approaches in either economic or legal field of study in order to opt-out moral hazard and to avoid adverse selection. All imply effects on the transaction costs but aim to achieve a pareto-efficient result and to achieve efficiency in general, even if it is not always the case, as the example of signaling shows.

On the whole it can be said that a re-allocation of risks or losses towards the agent (the promisor) always imposes a positive effect on the reduction to the incentive of moral hazard but also on adverse selection.

 

References

Robert Cutter, Thomas Ulen. (2000): Law and Economics. Third Edition. Addison-Wesley. pp 225-328.

C.F. Zinnes. (2004): Moral Hazard and Adverse Selection. Institutions, Governance and Economic Development. University of Maryland-College Park.

Journal of Insurance Issues. (2005): Vol.28. No. 1. pp.1-13.

McGraw-Hill, Irwin. (2004): Insurability of Risk, Contractual Provisions, and Legal Doctrines. Chapter 10.

G. Dionne, P.-C. Michaud, M. Dahchour. (2004).: Separating Moral Hazard from Adverse Selection in Automobile Insurance: Longitudinal Evidence from France. Working paper 04-05. Canada Research Chair in Risk Management, HEC Montréal.

Prof. Dr. Helmut Gründl.(2005): Organisations- und Entscheidungstheorie. Dr. Wolfgang Schieren-Lehrstuhl für Versicherungs- und Risikomanagement. Humboldt-University of Berlin.