Asymmetric Information - University of Southern Californiaebayrak/teaching/LECTURES/303/week... · • An asymmetric information problem occurs when a seller knows more about

Asymmetric Information

16

Introduction 16

Chapter Outline

16.1 The Lemons Problem and Adverse Selection

16.2 Moral Hazard

16.3 Asymmetric Information in Principal–Agent Relationships

16.4 Signaling to Solve Asymmetric Information Problems

16.5 Conclusion

16Introduction

Our analysis of markets thus far has assumed that participants have

complete information in market transactions.

• All participants in a transaction know the relevant information.

• All participants face the same level of uncertainty.

In reality, many market transactions occur under conditions of incomplete

information, in which one or more parties to a market cannot determine with

certainty all of the important attributes of a market or a particular good.

This is asymmetric information.

• An imbalance of information across participants in a transaction

The Lemons Problem and

Adverse Selection 16.1

A common manifestation of asymmetric information in markets is the

lemons problem.

• An asymmetric information problem occurs when a seller knows more about

the quality of the good he is selling than does the buyer.

• First formally analyzed by economist George Akerlof, who studied the used-

car market

Consider the market for a used car from a private-party seller.

• Assume there are two types of used cars, good ones (plums) and bad ones

(lemons).

• Half of the cars available are plums, and half are lemons.

• Potential buyers value plums at $10,000 but place no value on lemons.

• Sellers value plums at $8,000 and also have no value for lemons.



Observable Quality

First, think about market outcomes if quality attributes are observable to

both sellers and buyers.

• Since sellers value plums at $8,000 and buyers value plums at $10,000, the

half of used cars that are plums will sell at prices between these two values.

• Both parties are better off following the trade.

‒ Buyers value the cars more than former owners, and former owners are happier

with the money.

• Lemons have no value, so none will be sold.



Unobservable Quality

Now, consider what happens when sellers know whether their offering is

a plum or a lemon but buyers do not.

• Buyers know 50% of cars are lemons.

• Therefore, buyers recognize that the probability that they will purchase a

lemon is 50%.

What is the most a buyer is willing to pay for a used car?

• Since buyers value plums at $10,000 and lemons at $0 and since there is a

50% chance that a given car is a lemon, the most a buyer is willing to pay is

• Any buyer who pays more is worse off (in expectation) from engaging in the

trade.

000,5$50.00$50.0000,10$ EV



Unobservable Quality

Now, think about the owner of a plum who is considering whether to sell.

• The seller values his car at $8,000 but recognizes that because buyers can’t

know for sure whether or not his car is a plum, he could never get more than

$5,000.

• Therefore, he does not offer the car for sale.

The lemons problem is that Pareto-improving market transactions fail to

occur because of asymmetric information.



Adverse selection: market characteristics leading to more low-quality

goods and fewer high-quality goods available

The existence of quality differences is not by itself a market failure;

instead, it is a lack of information.

• This leads to a market price that provides no incentive for plums to be offered

for sale but every incentive for lemons to be offered for sale.

Information asymmetries hurt not only those with little information but

also those with more information.

• Both sides lose because a lack of complete information prevents trades.

• In this example, buyers and owners of plums are both hurt despite the

owners having complete information.



Other Examples of the Lemons Problem

Among the many markets characterized by information asymmetries:

• Used merchandise sold online

• Home improvement

• Vehicle repairs

• Labor

• Insurance



Institutions That Mitigate Lemons Problems

The lemons problem destroys economic value by preventing beneficial

exchanges. In response, institutions to mitigate information asymmetries

have developed.

These can work in three ways:

1. Address the information asymmetry directly by allowing buyers to observe

quality characteristics before a transaction takes place

2. Punish sellers who misrepresent their lemons as plums

3. Use incentives to increase the number of plums brought to market



Reducing Asymmetric Information Directly

• Third-party examinations of quality (e.g., mechanics)

• Offering standardized, unbiased information products (e.g. Carfax)

Incentives for Truthful Quality Reporting

• Reputation (e.g., online feedback)

• Warranties and return policies (offered by seller)

‒ Lemon laws mandate warranties and return policies for new and used vehicles

in many states.

Increasing the Average Quality of Cars Placed on the Market

• Leasing can increase the average quality of used cars by encouraging return,

regardless of quality.



Beyond Used Cars

• Yelp and Angie’s List

• Consumer Reports

• Referrals and references

• Background checks

• Accreditation services



Adverse Selection When the Buyer Has More Information:

Insurance Markets

Thus far we have considered the case of the seller having more information than

the buyer. However, in some markets the buyer holds the information advantage.

Consider insurance.

• Health, auto, life, others

• Requires the seller to pay the buyer compensation in the event of a covered

incident

What information do the buyers have that sellers lack?

• Risk! Sellers are unable to determine quality, or the likelihood that a buyer

will have claims.

• Thus insurance buyers are adversely selected in insurance markets.



Mitigating Adverse Selection in Insurance

A number of mechanisms have emerged to deal with adverse selection

in insurance markets.

1. Group policies

‒ Tying insurance to employment removes the link between the

individual’s riskiness and the decision to purchase insurance.

‒ Pooling individuals reduces the effect of any given poor-risk person.

2. Screening

‒ Detailed questionnaires, health exams, driving records, and so on

3. Denying coverage

Moral Hazard 16.2

Moral Hazard

Moral hazard arises when one party to a transaction cannot observe the

other party’s behavior.

• When quality is difficult to observe, a party to a transaction may have a

financial incentive to engage in fraud.

‒ For example, money managers handling clients’ funds

In insurance markets, adverse selection refers to the problem of

deciding whom to insure and at what price.

Moral hazard refers to the effect of being insured on the behavior of an

individual.

• Knowing you are insured may make you more willing to take risks, since part

of your risk is being borne by a third party.

Moral Hazard 16.2

Moral Hazard in Insurance Markets

Consider comprehensive auto insurance.

• Comprehensive insurance compensates drivers in the event their vehicle is stolen

(among other things).

How might drivers’ behavior change if they know they are

covered by a comprehensive insurance policy?

‒ The coverage may lead them not to try to prevent theft.

Examples:

‒ Parking on the street instead of a garage

‒ Parking in relatively unsafe areas

‒ Not locking the car

Moral Hazard 16.2

Figure 16.1 Moral Hazard in the Insurance Market

Marginalbenefit, MB

Marginalcost, MC

MC

MBMBPI

MBFI

0 Quantity ofactions taken to

improve outcome

*API A*

In a market without insurance, an agent will act to improve the potential for a good outcome up until A*, where the marginal benefit of further action,

MB, equals the marginal cost , MC.

If instead the agent obtains full insurance, there is no marginal benefit to acting to improve the

potential for a good outcome, because he will be fully compensated for any bad outcome.

If the policy offers partial insurance, the agent will have the incentive to take action—but not to the

same degree as in the absence of insurance.

Moral Hazard 16.2

Moral Hazard Outside of Insurance Markets

Another common instance of moral hazard occurs between borrowers

and lenders in financial markets.

• Borrowers’ liability may be limited (e.g., in bankruptcy proceedings), meaning

they may be willing to take unjustified risks with borrowed funds.

The recent financial crisis has highlighted the issue of “too big to fail.”

• Implicit guarantees to large financial institutions may increase risk taking.

Employer–employee relationships often are a source of moral hazard.

• Inability to observe all of employees’ activities provides an opportunity and

incentive for employees to shirk.

Moral Hazard 16.2

Lessening Moral Hazard

Market mechanisms have developed to diminish the effects of moral hazard.

Insurance policies often mandate actions to be taken by the insured.

• Commercial property insurance often requires working smoke detectors and

regularly inspected fire extinguishers.

Policies can be structured to encourage good behavior.

• Some auto insurance deductibles fall after each accident-free year.

• Life insurance premiums usually fall when efforts are made to improve health

(e.g., quit smoking).

Asymmetric Information in

Principal–Agent Relationships 16.3

Principal–agent relationships are a set of economic transactions that

feature information asymmetry between a principal and his hired agent,

whose actions the principal cannot fully observe.

• Involves one party (the principal) hiring a second party (the agent) to perform

a task and being unable to completely observe the effort of the agent

Information asymmetry is insufficient to develop a principal–agent

problem.

• There must also be a misalignment between the incentives and preferences

of the principal and the agent.

• For example, an employer wants her employees to work as hard as possible,

but employees want to work as little as possible without being fired.



Principal–Agent and Moral Hazard: An Example

Consider a mall kiosk selling mobile phones.

• The kiosk is owned by Selena (the principal), who hires a single employee

(Joe, the agent) to staff the kiosk.

• If Joe works hard, the kiosk earns a daily profit of $1,000 with 80% probability

and $500 with 20% probability (expected profit of $900).

• If Joe shirks, the probabilities are reversed (expected profit of $600).

• Joe does not like to work hard and requires at least $150 per day to do so.

• Selena cannot observe Joe’s effort.

Selena would be happy to pay Joe a flat rate of $150 per day to work

hard, but as she cannot observe his effort (and profits are uncertain),

she cannot be sure that he will indeed follow through.

How can Selena encourage Joe to work hard?




Rather than paying Joe a flat rate, Selena can instead pay Joe a wage

that varies with the profits of the kiosk.

• For example, Joe will be paid $255 if daily profits are high ($1,000) and $0

when profits are low ($500).

• If Joe works hard, the 80% probability of high profits yields an expected wage

of $200 and a surplus of $50. (Remember, he is indifferent between $150

and exerting high effort.)

• If Joe shirks, the 20% probability of high profits yields an expected wage of

$50, with no cost to him because he need not exert any effort.

‒ Thus, Joe will be indifferent between working hard and shirking.

• However, if Selena chooses to pay Joe $255 if the profits are high, he will

now clearly prefer to work hard.




Does Selena prefer this arrangement?

• Paying $255 to Joe for high-profit days yields an expected wage of $204.

• This is less than the increase in expected profits associated with Joe working

hard ($300), so she will prefer this compensation plan.

• Expected profits are significantly higher than under the fixed compensation

plan of $150 per day,.



The Principal–Agent Relationship as a Game

This example (and most principal–agent relationships) can be thought of

as a sequential game.

• In the first stage, the principal chooses from a set of contracts.

• In the second stage, the agent chooses a level of effort resulting in payoffs to

the principal and agent.

‒ The goal for the principal is to choose a contract structure that induces

the agent to choose a high level of effort.



Figure 16.2 The Principal–Agent Problem as a Sequential Game

( 750 , 0 )

AgentWork hard

(Worker: Joe)B Shirk

Flat wage

( 450 , 150 )Principal(Owner: Selena)

A( 696 , 54 )

Wage tied to

kiosk’s profit Work hardC

Agent Shirk(Worker: Joe)

( 549 , 51)

PP

P

If Selena pays a flat $150, the interests of Joe and Selena do not align. Joe gains

more by shirking ($150 compared to a payoff of $0 if he works hard, because of his cost of effort). Selena prefers that he

work hard so that she earns $750, compared to $450 if Joe does not work.

If Selena links Joe’s pay to profit, Joe’s and Selena’s interests align. Joe will choose to work hard, earning him $54 and Selena $696. As a result, Selena chooses to link

Joe’s pay to the kiosk’s profit.

Signaling to Solve Asymmetric

Information Problems 16.4

One major result of principal–agent problems is that good agents or

products are not identifiable and therefore cannot command full value.

Signaling is a solution to the problem of asymmetric information in

which the knowledgeable party alerts the other party to an unobservable

characteristic of the good.

Often, economic actors will attempt to communicate their quality via a

signal.

• A costly action taken by an economic actor to indicate something that would

otherwise be difficult to observe

• To signal high quality credibly, a signal must be less costly for high-quality

agents than low-quality agents.



The Classic Signaling Example: Education

Education—specifically the granting of degrees—is a classic case of

signaling.

• Individuals with college degrees earn more in their lifetime than individuals

with only a high school diploma.

Is this because the training that education provides increases an

individual’s value—or something else?




Highly productive workers like to let employers know they are indeed

highly productive so they might receive a job offer and/or a higher wage.

• However, simply telling an employer on your résumé that you are productive

is cheap talk.

• A college degree, however, signals that you are indeed productive.

• College is difficult. It takes time, a lot of money, and an ability to learn and

apply new information.

• These are the same attributes that often make employees productive.

• The fact that a third party (an accredited university) has issued a degree is a

credible signal of productivity.




We can use numbers to make this example clearer.

• Suppose there are two types of workers, high-productivity and low-

productivity.

• Each year of higher education costs high-productivity workers $25,000,

including psychic costs of going to class, finishing assignments, studying for

exams, and so on.

• Each year of higher education costs low-productivity workers $50,000

because of their higher psychic costs.

yC

yC

L

H

/000,50$

/000,25$




• Over a lifetime, high-productivity workers produce $250,000 worth of value to

employers, whereas low-productivity workers produce $125,000 of value

regardless of education.

• Thus, employers are willing to pay high-productivity workers up to $125,000

more in wages. However, employers must be able to tell them from the low-

productivity workers.

• Suppose employers view a four-year degree as a signal of high productivity

and are willing to pay workers with a degree $125,000 more.

Is this an effective strategy?

‒ For this to be effective, it must be incentive-compatible.

‒ High-productivity workers must find it advantageous to use college as a signal,

and low-productivity workers must find a college degree to be not worth the

effort.




• To determine whether the strategy is effective, compute the net benefits for

each type of worker; ignore discounting.

• High-productivity workers earn $25,000 more in surplus with a college

degree, while low-productivity workers lose $75,000.

Therefore, this is an effective strategy, as the low-productivity workers will

choose not to go to college.

000,25$000,100000,125)4( yearsCBenefitNB HH

000,75$000,200000,125)4( yearsCBenefitNB LL



Figure 16.3 Education as a Signal on the Job Market



Other Signals

How might the following be used as signals?

• Buying an engagement ring for your fiancée

• Offering a 10-year warranty (e.g., Hyundai cars)

• Driving an expensive car

• Certifying your products as organic

Since quality is not always observable, signals are a market mechanism

that can overcome some problems associated with information

asymmetry.

Conclusion 16.5

This chapter has examined asymmetric information.

In a world in which quality is not always observable, signals are a market

mechanism that can overcome some problems associated with

information asymmetry.

In the next chapter, we examine other market failures related to

incomplete property rights, including externalities and public goods.

In-text

figure it out

Suppose consumers value a high-quality used laptop computer at $400,

while they value a low-quality used laptop at $100. The supply of high-

quality laptops is QH = PH – 100, while the supply of low-quality laptops

is QL = 2PL – 50. Potential buyers cannot tell the difference between

high-quality and low-quality laptops.

Answer the following questions:

a. Buyers believe there is a 50% probability that a used laptop will be

high-quality. What price would buyers be willing to pay for any used

laptop?

b. If the price you determined is offered, how many high-quality

laptops will be available? How many low-quality? Are buyers

correct that 50% of used laptops for sale are high-quality? Explain.

c. What do you expect to happen over time as information about the

true odds of buying a high-quality used laptop become known?

Explain.

In-text

figure it out

a. If buyers expect that 50% of the used laptops available are high-

quality (meaning that the other 50% are low-quality), the expected

value of a laptop is equal to

Therefore, $250 is the most that buyers would be willing to pay for

a used laptop.

b. If the price of a used laptop is $250, the quantity supplied of high-

quality laptops is QH = 250 – 100 = 150, while the supply of low-

quality laptops is QL = 2(250) – 50 = 450. Therefore, 600 used

laptops, 150 high-quality and 450 low-quality, will be for sale.

The probability of buying a high-quality used laptop is not 50% but

150 / 600 = 0.25, or 25%. Because of asymmetric information, buyers

are not willing to pay much for a used laptop.

250$50$200$100$5.04005.0

In-text

figure it out

c. Over time, buyers will adjust the expected value of used laptops.

This will further reduce the price buyers are willing to pay. Owners

of high-quality used laptops will be even less inclined to sell them,

reducing even more the proportion of high-quality used laptops

available. It is possible for the market to end up with only low-

quality used laptops available.

Additional

figure it out

Suppose consumers value a high-quality used laptop computer at $400

and a low-quality used laptop at $100. The supply of high-quality

laptops is QH = PH – 100, and the supply of low-quality laptops is QL = 2PL – 50. Buyers cannot tell the difference between high-quality and

low-quality laptops before the purchase.


a. Assume that buyers believe there is a 75% probability that a used

laptop will be high-quality. What price would buyers be willing to

pay for any used laptop?

b. If the price you determined is offered, how many high-quality

laptops will be made available? How many low-quality laptops? Are

buyers correct in their assumption that 75% of the used laptops

available for sale are high-quality? Explain.

c. What do you expect to happen over time as information about the

true odds of buying a high-quality used laptop become known?

Explain.

Additional

figure it out

a. If buyers expect 75% of the used laptops available to be high-

quality (meaning that the other 25% are low-quality), the expected

value of a laptop is equal to

Therefore, $325 is the most that buyers would be willing to pay for

a used laptop.

b. If the price of a used laptop is $325, the quantity supplied of high-

quality laptops is QH = 325 – 100 = 225, while the supply of low-

quality laptops is QL = 2(325) – 50 = 600. Therefore, there will be

825 used laptops, 225 high-quality and 600 low-quality, for sale.

The probability of buying a high-quality used laptop is not 75% but

225 / 825 = 0.273, or 27.3%. Because of asymmetric information,

buyers are not willing to pay much for a used laptop.

0.75´400+0.25´$100 = $300+$25= $325

Additional

figure it out

c. Over time, buyers will adjust the expected value of used laptops.

This will further reduce the price buyers are willing to pay. Owners

of high-quality used laptops will be even less inclined to sell them,

reducing even more the proportion of high-quality used laptops

available. It is possible for the market to end up with only low-

quality used laptops available.

In-text

figure it out

Anastasia and Katherine own a café. They run the risk of loss

due to small kitchen fires. This risk can be mitigated by taking

precautions (e.g., purchasing fire extinguishers, training

employees). Assume that the marginal cost of precautions can

be represented by MC = 80 + 8A, where A is equal to the actions

to mitigate the risk of a fire. Likewise, the marginal benefit of

these precautions is MB = 100 – 2A.


a. If the cafe has no insurance, what is the optimal level of

precautions?

b. Suppose the restaurant has insurance that reduces the

marginal benefit of taking precautions to MB = 90 – 4A. What

happens to the optimal level of precautions? Explain.

In-text

figure it out

a. Anastasia and Katherine will take precautions until their

marginal benefits equal their marginal costs.

b. With the insurance policy in place, the benefits of

precautions fall:

The policy reimburses Anastasia and Katherine in case of a fire,

which reduces the owners’ incentive to try to prevent a loss.

AA 8802100

A1020

2A

AA 880490

A1210

83.A

Additional

figure it out

John owns a restaurant, and as with any cooking establishment,

it carries a risk of fire. This risk can be mitigated by precautions

(e.g., purchasing fire extinguishers, training employees). Assume

that the marginal cost of precautions can be represented by MC= 60 + 9A, where A is equal to the actions taken to mitigate the

risk of a fire. Likewise, the marginal benefit of these precautions

is MB = 100 – A.


a. If the restaurant has no insurance, what is the optimal level

of precautions?

b. Suppose the restaurant has insurance that reduces the

marginal benefit of taking precautions to MB = 80 – A. What

happens to the optimal level of precautions? Explain.

Additional

figure it out

a. John will take precautions until their marginal benefits are

equal to their marginal costs.

b. With insurance in place, the benefits of precautions fall.

This outcome is due to the fact that the policy reimburses John

in case of fire, which reduces the owner’s incentives to try to

prevent a loss.

AA 960100

A1040

4A

AA 96080

A1020

2A

In-text

figure it out

Pablo is a struggling artist who wants to rent an apartment from Donald.

Pablo loves to draw, so much so that he often draws on any surface he can

find, including walls.

In fact, Pablo would get $300 in utility from being able to draw on the walls of

the apartment. On the other hand, Donald would like Pablo to leave the

apartment walls clean and free of marks.

If Pablo draws on the walls, it will cost Donald $500 to have the apartment

repainted. Therefore, Donald is considering charging Pablo a damage

deposit of $500.


a. Explain why this situation could be considered a principal–agent

problem. Who is the principal? Who is the agent?

b. Draw the extensive form of this principal–agent problem and use

backward induction to solve for the Nash equilibrium.

In-text

figure it out

a. Donald is the principal and Pablo is the agent. Donald would like

Pablo to treat the apartment the same way he (Donald) would.

However, Donald cannot be at the apartment to monitor Pablo’s

behavior all of the time. Therefore, a principal–agent problem exists.

Donald’s and Pablo’s interests do not coincide.

In-text

figure it out

b. The extensive form of this principal–agent problem:

Given this information, the Nash equilibrium occurs where Donald charges

the damage deposit and Pablo does not draw on the wall.

( 0 , 0)P

If Donald does not charge a damage

deposit, Pablo will want to draw on the

walls because $300 > $0. Donald will

then have to repaint the apartment and

lose $500.

If Donald charges the damage

deposit, Pablo will leave the walls

untouched because $0 > −$200. In

this case, Donald will not lose $500

because he will not have to repaint.

( −500 , 300)

AgentPaint on Walls(Pablo)

B Leave walls untouchedCharge no damage deposit

( 0 , 0 )Principal(Donald)

A( 0 , −200 )

Charge damage deposit of $500

C

Agent

P

P

(Pablo)

Paint on Walls

Leave walls untouched

We can use backward induction to

solve the game.

Additional

figure it out

The Bureau of Land Management (BLM) is considering leasing land for oil

drilling to DBD, Inc. Extracting the oil would yield a $10 million profit for DBD.

After drilling, DBD can choose to plug the well properly and reclaim (clean

up) the land, but this would cost $2 million. If DBD chooses not to reclaim the

land, the BLM will have to take over the responsibility at a cost of $3 million.

(DBD is a small corporation that will go bankrupt if sued for damages.)

The BLM is considering charging DBD a refundable bond in the amount of $3

million to cover the costs of reclamation should DBD not follow through.

Assume the price of the lease is $5 million.


a. Explain why this situation could be considered a principal–agent

problem. Who is the principal? Who is the agent?

b. Draw the extensive form of this principal–agent problem and use

backward induction to solve for the Nash equilibrium.

Additional

figure it out

a. The BLM is the principal and DBD is the agent. The BLM would

like DBD to reclaim the land after drilling. However, the BLM cannot

monitor DBD’s behavior all of the time. Therefore, a principal–agent

problem exists. The incentives facing BLM and DBD do not

coincide.

Additional

figure it out

b. The extensive form of this principal–agent problem:

Given this information, the Nash equilibrium occurs where the BLM

charges the bond and DBD reclaims the land.

( $5M , $2M)

P

We can use backward induction to

solve the game.

If no bond is charged, DBD will not

reclaim the land and the BLM will have

to spend $3 million of the $5 million

revenue from the lease on reclaiming

the land. ($300 > $0).

If the BLM charges a bond, DBD will

be better off reclaiming the land

because $2 million cost is less than

the $3 million bond.

( $5M , $3M)

AgentReclaim(DBD)

B Do not reclaimNo Bond

( $2M , $5M )Principal(BLM)

A( $5M , $3M )

$3M Bond

ReclaimC

Agent Do not reclaim(DBD)

P

P

In-text

figure it out

Last year, Used Cars R Us sold very few cars and ended up with a large

economic loss. The owner, Geoffrey, has developed two strategies to

help the dealership sell more vehicles in the coming year by signaling

that it deals in only high-quality used vehicles:

Geoffrey has decided to pursue one of two strategies:

1. Change the name of the dealership to Quality Used Cars R Us.

2. Offer a 60-day bumper-to-bumper warranty for every car sold.

Answer the following question:

Which is the more effective signal of high quality? Explain.

In-text

figure it out

To be a good indicator of quality, a signal must be relatively cheap for

high-quality producers and expensive for low-quality producers.

Therefore, the better signal is the 60-day warranty.

If the cars sold at Used Cars R Us are truly high-quality, the warranty

will not be very expensive for the dealership. On the other hand, if Used

Cars R Us only sold lemons, the warranty would be very expensive and

negate the benefit of increased sales. Thus, consumers can be more

confident that a dealership offering a warranty has higher-quality

products than those that do not.

The change of the name of the dealership would just be cheap talk. Any

dealership can alter its name, and the cost of doing so will not vary

between high-quality and low-quality sellers.

Additional

figure it out

CheapFix is a big-city mechanic shop that specializes in brake repair.

After a number of poor online reviews of its brake service, management

has decided to hire new mechanics and pursue a new marketing

strategy to change consumer perceptions about the quality of the shop’s

workmanship.

Management has decided to pursue one of two strategies:

1. Change the name of the shop to QualityFix

2. Offer a 3-month warranty on brake service

Answer the following question:

Which of these two strategies is the more effective signal of high

quality? Explain.

Additional

figure it out

Remember, for a signal to credibly indicate high quality, it must be

cheaper for high-quality producers and more expensive for low-quality

producers.

Therefore, the best signal is the 3-month warranty. If the new brake

service is indeed high-quality, this policy should be less expensive than

it would be for a low-quality shop.

On the other hand, the change in the name of the shop is cheap talk,

whose cost does not vary across shops.

Documents

Asymmetric Information - University of Southern Californiaebayrak/teaching/LECTURES/303/week... · • An asymmetric information problem occurs when a seller knows more about