Managerial economics strategy by m perloff and brander chapter 15 asymmetric information

15.1 Adverse Selection• The Problems of Adverse Selection – If consumers lack relevant information, they may not engage in transactions to avoid being exploited by better informed seller

Chapter 15

Asymmetric Information

• Managerial Problem

– During the mortgage market meltdown that started in 2007, record numbers

of mortgage holders defaulted on their loans.

– Why did executives at these banks take risks that resulted in so much lost shareholder value? How can a firm compensate its corporate executives so as

to prevent them from undertaking irresponsible and potentially devastating actions?

• Solution Approach

– We need to examine how a manager’s incentives can lead to excessive risk taking

• Empirical Methods

– A party in a transaction may know less than the other party (asymmetric

information) because of hidden characteristics or hidden actions.

– A more-informed party may exploit the less-informed party, engaging in

opportunistic behavior and creating two problems: adverse selection and

moral hazard.

15.1 Adverse Selection

• The Problems of Adverse Selection

– If consumers lack relevant information, they may not engage in

transactions to avoid being exploited by better informed sellers (adverse selection)

– As a result, not all desirable transactions occur and potential consumer and producer surplus is lost

– In extreme cases, adverse selection may prevent a market from operating

at all

• Adverse Selection Cases

– Two important examples of adverse selection problems are insurance and products of varying quality

15.1 Adverse Selection

• Adverse Selection in Insurance Markets

– How would people react to a fair insurance rate (a rate for insurance equal to the average cost of health care for the entire population)?

– Unhealthy people—people who expect to incur health care costs that are higher than average—would view this insurance as a good deal and many would buy it.

– Healthy people, in contrast, would not buy it because the premiums would exceed their expected health care costs (unless they are extremely risk averse).

– So a disproportionately large share of unhealthy people will buy the insurance

(adverse selection).

– The insurance company’s average cost of medical care for covered people exceeds the population average The company makes a loss.

• Consequences of Adverse Selection in Insurance Markets

– Adverse selection results in an inefficient market outcome: few healthy people are

insured and insurer’s cost is high The sum of CS and PS is not maximized.

– This outcome could be changed with perfect information: healthy people would buy insurance at a lower premium, but the insurer must first verify they are healthy

15.1 Adverse Selection

• Products of Unknown Quality

– Adverse selection often arises because sellers of a product have better information about the product’s quality than the buyer

– In a transaction of a used car, the seller knows whether his car is a

lemon, but the buyer cannot know it (hidden characteristic)

• Consequences of Unknown Quality

– If sellers have more information than buyers, adverse selection may drive high-quality products out of the market (Akerlof, 1970) Why?

– Car buyers worry that a used car might be a lemon They would be willing

to pay only relatively low prices that reflect the possibility of getting a lemon

– However, sellers of excellent used cars do not want to sell their cars at such low prices They do not enter the market

– Adverse selection has driven the high-quality cars out of the market,

leaving only lemons

15.1 Adverse Selection

• Lemons Example

– 1,000 sellers cannot alter the quality of their used cars; 1,000 buyers are

willing to pay $4,000 for a lemon and $8,000 for a good used car

– The demand curve for lemons, D L, is horizontal at $4,000 in panel a of Figure

15.1, and the demand curve for good cars, D G, is horizontal at $8,000 in panel b.

– The reservation price of lemon owners is $3,000, so the supply curve for

lemons, S L in panel a, is horizontal at $3,000 up to 1,000 cars, where it

becomes vertical (no more cars are for sale at any price)

– The reservation price of owners of high-quality used cars is v, which is less than $8,000 Panel b shows two possible values of v If v = $5,000, the supply curve for good cars, S1 , is horizontal at $5,000 up to 1,000 cars and then

becomes vertical If v = $7,000, the supply curve is S2

• Equilibrium with Full & Symmetric Information

– In panel a of Figure 15.1, the equilibrium in the lemons market (D L =S L ) is at e

and 1,000 lemons sell for $4,000 each In panel b, the equilibrium in the

good-car market is at E and 1,000 good cars sell for $8,000 each.

– This market is efficient: the goods go to the people who value them the most

15.1 Adverse Selection

Figure 15.1 Markets for Lemons and Good Cars

15.1 Adverse Selection

• Equilibrium with Incomplete & Symmetric Information

– If information is symmetric and buyers and sellers are equally ignorant

about the quality of cars, EV = 0.5 x 8000 + 0.5 x 4000 = $6,000 A

risk-neutral buyer and a seller would transact at $6,000

– This market is efficient: the goods go to the people who value them the most

• Equilibrium with Asymmetric Information

– When sellers know the quality but buyers do not, there are two possible equilibria

– If sellers value good cars at v = $5,000 and buyers consider EV = $6,000, all cars are sold at $6,000 The equilibrium points are f and F in Figure

15.1 In this case asymmetric information does not cause an efficiency problem, but it does have equity implications

– If sellers value good cars at v = $7,000, they will not sell them at $6,000

Buyers realize good cars cannot be found for less than $7,000 Then, the lemons drive good cars out of the market, only lemons are sold at $4,000 leading to an inefficient equilibrium

15.1 Adverse Selection

• Summary of the Lemons Problem

– If buyers have less information about product quality than sellers do, the result might be a lemons problem in which high-quality cars do not sell even though potential buyers value the cars more than their current owners do

– The lemons problem does not occur if the information is symmetric If

buyers and sellers of used cars know the quality of the cars, each car sells for its true value in a perfectly competitive market If, as with new cars, neither buyers nor sellers can identify lemons, all cars sell at a price equal

to the EV.

• Varying Quality Under Asymmetric Information

– If consumers cannot identify high-quality goods before purchase, they pay the same for all goods regardless of quality

– Firms do not produce top-quality goods if p is the same for High and Low

quality

– The outcome is inefficient, assuming consumers are willing to pay more

for top-quality goods

15.2 Reducing Adverse Selection

• Restricting Opportunistic Behavior

– One method for solving adverse selection problems is to restrict the ability

of the informed party to take advantage of hidden information

– Which type of restriction works best depends on the nature of the adverse selection problem, as we see below

• Mandating Universal Coverage

– Health insurance markets have adverse selection because low-risk

consumers do not buy insurance at prices that reflect the average risk – Such adverse selection can be eliminated by providing insurance to

everyone or by mandating that everyone buy insurance

• Laws to Prevent Opportunism

– Product quality and product safety are known characteristics to sellers but not observed by buyers

– Product liability laws protect consumers from being stuck with

nonfunctional or dangerous products

15.2 Reducing Adverse Selection

• Equalizing Information

– Another method for solving adverse selection problems is to provide

information to all parties

– There are three methods for reducing information asymmetries:

screening, signaling, and third party

• Screening Reduces Adverse Selection

– Insurance companies screen potential customers based on their health records or medical exams They collect information until marginal benefit and marginal cost from the extra information are equal

– Buyers of used cars test or drive the cars, bring a trusted mechanic, or buy only from sellers with good reputation Reputation is not easy to get

in markets where buyers or sellers trade only once, like in tourist areas

15.2 Reducing Adverse Selection

• Signaling Reduces Adverse Selection

– An informed party may signal the uninformed party to eliminate adverse selection However, signals work only when the recipients view them as credible

– Examples: A firm may distribute a favorable report on its product quality

by an independent testing agency; a candidate for life insurance may present a health report signed by a doctor approved by the insurer;

education is also a signal

• Third Party Information Reduces Adverse Selection

– If the information on quality provided by consumer groups, nonprofit

organizations, and government agencies is credible, it can reduce adverse selection

– These groups and organizations also provide standards and certification If these programs inexpensively and completely inform consumers and do not restrict the goods available, the programs are socially desirable

15.3 Moral Hazard

• Moral Hazard and Adverse Selection

– Moral hazard problems come from hidden actions For instance, renters driving rental cars off-road, workers loafing when the boss is not

watching, and lawyers acting in their own interests instead of those of their clients

– We will focus on the insurance market and the principal-agent

relationship

• Moral Hazard in Insurance Markets

– Many types of insurance are highly vulnerable to hidden actions by

insured parties that result in moral hazard problems

– Example: A business insures merchandise in a warehouse against hazards such as fire and theft If merchandise is not selling, the owner faces a

significant financial loss He may burn down the warehouse and make an insurance claim

– Example: If doctor’s visits are free and unlimited with health insurance, the insured may make ‘excessive’ visits

15.3 Moral Hazard

• Moral Hazard in Principal-Agent Relationships

– Principal-Agent problem or agent problem: when responsibilities are delegated, a principal contracts with an agent to take an action that benefits the principal If the agent’s actions are hidden, moral hazard may result.

– Example: A business owner (principal) hires an employee (agent) to work at a

remote site and cannot observe whether the employee is working hard The employee may shirk by not providing all the services they’re paid to provide.

• Reducing Moral Hazard using Efficient Contracts

– The principal and agent can agree to an efficient contract: an agreement in which neither party can be made better off without harming the other party.

– If the parties to the contract are risk neutral, efficiency requires that the combined profit of the principal and agent be maximized

– If one party is more risk averse than the other, efficiency requires that the less averse party bear more of the risk.

risk-– In the previous example, efficiency occurs if the agent works extra hard so total profit is maximized, and if the agent (risk averse party) bears none of the risk.

15.3 Moral Hazard

• Moral Hazard & Efficient Contracts: Ice Cream Shop

– Paul (principal) owns many ice cream parlors across North America He contracts with Amy (agent) to manage his Miami shop Her duties include supervising

workers, purchasing supplies, and performing other necessary actions.

– The shop’s daily earnings depend on the local demand conditions and on how hard Amy works (Table 15.1) Demand can be high or low depending on weather

conditions (50%) and Amy can put in normal or extra effort (valued $40 per day) – Paul is risk neutral because he can pull earnings from the many stores he owns Amy, like most people, is risk averse

– We know an efficient contract requires Amy to bear no risk, but the outcome

depends on symmetric and asymmetric information.

• Ice Cream Shop Efficient Contract & Symmetric Information

– Moral hazard is not a problem if Paul can directly supervise Amy and agree that: Amy earns $200 per day if she works extra hard, but loses her job if she doesn’t – Amy’s zero risk: She gets $200 independently of weather

– Amy’s incentive to work hard: She nets $160 (200-40), better than being fired.

– Paul bears all risk: EV = $200; σ2 = 10,000 (perfect monitoring row in Table 15.2) – Efficient contract: Profit maximized, risk averse agent bears no risk.

15.3 Moral Hazard

Table 15.1 Ice Cream Shop Profits

15.3 Moral Hazard

Table 15.2 Ice Cream Shop Outcomes

15.3 Moral Hazard

• Ice Cream Shop Inefficient Contract & Asymmetric Information

– Moral hazard is a problem if Paul cannot observe Amy’s effort Both agree on a

fixed-fee contract: Amy earns $100 per day.

– Amy’s zero risk: She gets $100 independently of weather

– Amy’s incentive to work normally: If she works normally, she gets $100 But, if she works hard, she only nets $60 (100-40).

– Paul bears all risk: EV = $100; σ2 = 10,000 (fixed wage row in Table 15.2)

– Inefficient contract: Profit is not maximized, although risk averse agent bears no risk.

• Moral Hazard and Selfish Doctors Study Case in China

– Patients (principals) rely on doctors (agents) for good medical advice with respect to drug prescriptions Do doctors act only in their patient’s best interests?

– Lu (2011) found in China that doctors prescribed similarly whether or not a patient had insurance if the doctors received no compensation for prescriptions However, if doctors were compensated, they prescribed drugs that cost 43% more on average for insured patients than for uninsured patients.

– Thus, many of these doctors appeared to be motivated largely by self-interest.

15.4 Using Contracts to Reduce Moral Hazard

• Contracts and Correct Incentives

– A skillfully designed contract that provides strong incentives for the agent

to act so that the outcome is always efficient may solve moral hazard problems

– We will focus on fixed-fee and contingent contracts

• Fixed-Fee Contracts

– Amy could pay Paul a fixed license fee to operate Paul’s shop Paul bears

no risk as he receives a fixed fee, Amy bears all the risk and gets the residual profit

– Paul makes $200 with certainty

– Amy’s incentive to work hard: She earns all the increase in expected profit

from her extra effort EVHARD = $160 > -$100 = EVNORMAL, σ2 = 10,000 – Efficient contract: Licensing fee profit > fixed wage profit in Table 15.2 (360 > 200) However, it does not provide efficient risk bearing

– If Amy is nearly neutral, she picks the license fee If she’s highly averse, she picks the fixed wage

risk-15.4 Using Contracts to Reduce Moral Hazard

• State Contingent Contracts

– In a state-contingent contract, one party’s payoff is contingent on only the state of nature (weather conditions determine low and high demand).

– Contract: Amy pays a license fee of $100 if demand is low and $300 if

demand is high, and keeps all extra earnings (state-contingent row in Table 15.2)

– Amy’s incentive to work hard: Working normal she nets zero She must work hard.

– Amy bears no risk: EVHARD = $160 = EVNORMAL , σ 2 =0.

– Paul bears all risk: EV = $200, σ2 =10,000.

– Efficient outcome: Profit is maximized, risk averse agent bears no risk.