Ch17_Pindyck

8/22/2019 Ch17_Pindyck

1/81

Chapter 17

Markets with AsymmetricInformation


2/81

2005 Pearson Education, Inc. Chapter 17 2

Topics to be Discussed

Quality Uncertainty and the Market forLemons

Market Signaling

Moral Hazard

The Principal-Agent Problem

Managerial Incentives in an Integrated

FirmAsymmetric Information in Labor

Markets: Efficiency Wage Theory


3/81


Introduction

We can see what happens when someparties know more than othersasymmetric information

Frequently a seller or producer knowsmore about the quality of the productthan the buyer does

Managers know more about costs,competitive position and investmentopportunities than firm owners


4/81


Quality Uncertainty and theMarket for Lemons

Asymmetric information is a situation inwhich a buyer and a seller possessdifferent information about a transaction

The lack of complete information whenpurchasing a used car increases the risk ofthe purchase and lowers the value of the car

Markets for insurance, financial credit andemployment are also characterized byasymmetric information about product quality


5/81


The Market for Used Cars

Assume

Two kinds of cars high quality and lowquality

Buyers and sellers can distinguish betweenthe cars

There will be two markets one for high

quality and one for low quality


6/81



High quality market

SH is supply and DH is demand for high quality

Low quality market

SL is supply and DL is demand for low quality

SH is higher than SL because owners of highquality cars need more money to sell them

DH

is higher than DL

because people are willingto pay more for higher quality


7/81


The Lemons Problem

PH PL

QH QL

SH

SL

DH

DL

5,000

50,00050,000

10,000

DL

Market price for high quality carsis $10,000.

Market price for low quality cars is$5000.

50,000 of each type are sold.


8/81



Sellers know more about the quality of the usedcar than the buyer

Initially buyers may think the odds are 50/50

that the car is high qualityBuyers will view all cars as medium quality with

demand DM

However, fewer high quality cars (25,000) and

more low quality cars (75,000) will now be sold Perceived demand will now shift


9/81

2005 Pearson Education, Inc. 9

The Lemons Problem

PH PL

QH QL

DH

DL

5,000

50,00050,000

10,000

DL

Medium quality cars sell for $7500,selling 25,000 high quality and 75,000

low quality.

The increase in QL reduces expectationsand demand to DLM. The adjustment

process continues until demand = DL.

DM

25,000

7,500

75,000

7,500

DM

DLM

DLM


10/81



With asymmetric information:

Low quality goods drive high quality goodsout of the market - the lemons problem

The market has failed to produce mutuallybeneficial trade

Too many low and too few high quality carsare on the market

Adverse selection occurs; the only cars onthe market will be low quality cars


11/81


Market for Insurance

Older individuals have difficulty purchasinghealth insurance at almost any price

They know more about their health than the

insurance company Because unhealthy people are more likely to

want insurance, the proportion of unhealthypeople in the pool of insured people rises

Price of insurance rises so healthy people withlow risk drop out proportion of unhealthypeople rises, increasing price more


12/81



Ex: Auto insurance companies are targeting acertain population males under 25

They know some of the males have low

probability of getting in an accident and somehave a high probability

If they cant distinguish among insured, they will

base premium on the average experience

Some with low risk will choose not to insure,which raises the accident probability and rates


13/81



A possible solution to this problem is topool risks

Health insurance government takes on roleas with Medicare program

Problem of adverse selection is eliminated

Insurance companies will try to avoid risk byoffering group health insurance policies atplaces of employment and thereby spreadingrisk over a large pool


14/81



The Market for Credit

Asymmetric information creates the potentialthat only high risk borrowers will seek loans

Can end up with a lemons problem again

However, banks and credit agencies usecredit histories to gauge risk of borrowers


15/81


Importance of Reputation andStandardization

Asymmetric Information and Daily MarketDecisions

Retail sales return policies

Antiques, art, rare coins real or counterfeit

Home repairs unique information

Restaurants kitchen status


16/81


Implications of AsymmetricInformation

How can these producers provide high-quality goods when asymmetricinformation will drive out high-quality

goods through adverse selection?Reputation

You hear about restaurants or stores that havegood or bad service and quality

Standardization Chains that keep production the same

everywhereMcDonalds, Olive Garden


17/81


Implications of AsymmetricInformation

You look forward to a Big Mac whentraveling, even if you would not typicallybuy one at home, because you knowwhat to expect

Holiday Inn once advertised No

Surprises to address the issue of

adverse selection


18/81


Lemons in Major LeagueBaseball

Rules in baseball changed so that after 6years a player could either re-sign withtheir team or become a free agent and tryto sign with another team

Free agents create a secondhand marketin baseball players

If a lemons market exists, free agents shouldbe less reliable (disabled) than renewedcontracts


19/81


Player Disability

Days on Disabled List per Season

PreContract

PostContract

PercentChange

All Players4.73 12.55 165.4

RenewedPlayers

4.76 9.68 103.4

FreeAgents

4.67 17.23 268.9


20/81


Lemons in Major LeagueBaseball

Findings

Days on the disabled list increase for bothfree agents and renewed players

Free agents have a significantly higherdisability rate than renewed players

This indicates a lemons market


21/81


Market Signaling

The process of sellers using signals toconvey information to buyers about theproducts quality

For example, how do workers letemployers know they are productive sothey will be hired?


22/81


Market Signaling

Weak signal could be dressing wellIs weak because even unproductive

employees can dress well

Strong SignalTo be effective, a signal must be easier for

high quality sellers to give than low qualitysellers

Example Highly productive workers signal with

educational attainment level


23/81


Model of Job Market Signaling

Assume two groups of workers

Group I: Low productivity

Average Product & Marginal Product = 1

Group II: High productivity

Average Product & Marginal Product = 2

The workers are equally divided between

Group I and Group IIAverage Product for all workers = 1.5


24/81



Competitive Product Market

P = $10,000

Employees average 10 years of employment

Group I Revenue = $100,000

(10,000/yr. x 10 years)

Group II Revenue = $200,000

(20,000/yr. X 10 years)


25/81



With Complete Information

w = MRP

Group I wage = $10,000/yr.

Group II wage = $20,000/yr.

With Asymmetric Information

w = average productivity

Group I & II wage = $15,000/yr.


26/81



If use signaling with education

y = education index (years of highereducation)

Assume all benefits encompassed in years ofeducation

C = cost of attaining educational level y

Tuition, books, opportunity cost, etc.

Group I CI(y) = $40,000y

Group II CII(y) = $20,000y


27/81



Cost of education is greater for the lowproductivity group than for highproductivity group

Low productivity workers may simply be lessstudious

Low productivity workers progress more

slowly through degree program


28/81



Assume education does not increaseproductivity with only value as a signal

Find equilibrium where people obtaindifferent levels of education and firmslook at education as a signal

Decision Rule:

y* signals GII and wage = $20,000

Below y* signals GI and wage = $10,000


29/81



Decision Rule:

Anyone with y* years of education or more isa Group II person offered $20,000

Below y* signals Group I and offered a wageof $10,000

y* is arbitrary, but firms must identify

people correctly


30/81



How much education will individualsobtain given that firms use this decisionrule?

Benefit of education B(y) is increase inwage associated with each level ofeducation

B(y) is initially 0, which is the $100,000base 10 year earningsContinues to be zero until reach y*


31/81



There is no reason to obtain aneducation level between 0 and y*because earnings are the same

Similarly, there is no incentive to obtainmore than y* level of education becauseonce hit the y* level of pay, there are no

more increases in wages


32/81



How much education to choose is abenefit cost analysis

Goal: obtain the education level y* if the

benefit (increase in earnings) is at leastas large as the cost of the education

Group I:$100,000 < $40,000y*, y* >2.5

Group II:$100,000 < $20,000y*, y* < 5


33/81



This is an equilibrium as long as y* is between2.5 and 5

If y* = 4

People in Group I will find education does not pay andwill not obtain any

People in Group II will find education DOES pay andwill obtain y* = 4

Here, firms will read the signal of education andpay each group accordingly


34/81

2005 Pearson Education, Inc. 34

Signaling

Value ofCollege

Educ.

$100K

Value ofCollege

Educ.

Years ofCollege

Years ofCollege

0 1 2 3 4 5 6 0 1 2 3 4 5 6

$200K

$100K

$200KC

I

(y) = $40,000y

B(y) B(y)

y* y*

CII

(y) = $20,000y

Optimal choice ofyfor Group II

Group IIGroup I

Optimal choice ofyfor Group I


35/81


Signaling

Education provides a useful signal aboutindividual work habits and productivityeven if that education does not changeproductivity.


36/81


Market Signaling

Guarantees and Warranties

Signaling to identify high quality anddependability

Effective decision tool because the cost ofwarranties to low-quality producers is toohigh


37/81


Moral Hazard

Moral hazard occurs when the insuredparty whose actions are unobserved canaffect the probability or magnitude of apayment associated with an event

If my home is insured, I might be less likelyto lock my doors or install a security system

Individual may change behavior because ofinsurancemoral hazard


38/81


Moral Hazard

Determining the Premium for FireInsurance

Warehouse worth $100,000

Probability of a fire:

.005 with a $50 fire prevention program

.01 without the program

If the insurance company cannot monitor tosee if the program was run, how do theydetermine premiums?


39/81


Moral Hazard

With the program the premium is:0.005 x $100,000 = $500

Once insured owners purchase the

insurance, the owners no longer have anincentive to run the program, thereforethe probability of loss is 0.01

$500 premium will lead to a loss becausethe expected loss is now $1,000 (0.01 x$100,000)


40/81


Moral Hazard

Moral hazard is not only a problem forinsurance companies, but it alters theability of markets to allocate resources

efficiently

Consider the demand (MB) of driving

If there is no moral hazard, marginal cost of

driving is MCIncreasing miles will increase insurance

premium and the total cost of driving


41/81


The Effects of Moral Hazard

Miles per Week

$0.50

50

Costper

Mile

$1.00

$1.50

$2.00

D = MB

MC (w/moral hazard)

With moral hazardinsurance, companies cannot

measure mileage. MCgoes to $1.00 andmiles driven increases to 140

miles/week inefficient allocation.

140

MC (no moral hazard)

100


42/81


Reducing Moral Hazard Warranties of Animal Health

Scenario

Livestock buyers want disease-free animals

Asymmetric information exists

Many states require warranties

Buyers and sellers no longer have anincentive to reduce disease (moral hazard)


43/81


The Principal Agent Problem

Owners cannot completely monitor theiremployees employees are betterinformed than owners

This creates a principal-agent problemwhich arises when agents pursue theirown goals, rather than the goals of the

principal


44/81



Company owners are principals

Workers and managers are agents

Owners do not have complete knowledge

Employees may pursue their own goalseven at a cost of reduced profits


45/81



The Principal Agent Problem in PrivateEnterprises

Only 16 of 100 largest corporations have

individual family or financial institutionownership exceeding 10%

Most large firms are controlled bymanagement

Monitoring management is costly(asymmetric information)


46/81


The Principal Agent Problem Private Enterprises

Managers may pursue their ownobjectives

Growth and larger market share to increase

cash flow and therefore perks to themanager

Utility from job, from profit, and from respectof peers, power to control corporation, fringe

benefits, long job tenure, etc.


47/81



Limitations to managers ability to deviate

from objective of owners

Stockholders can oust managers

Takeover attempts if firm is poorly managed

Market for managers who maximize profitsthose that perform get paid more so incentiveto act for the firm


48/81



The problem of limited stockholdercontrol shows up in executivecompensation

Business Weekshowed that average CEOearned $13.1 million and has continued toincrease at a double-digit rate

For the 10 public companies led by the

highest paid CEOs, there was negativecorrelation between CEO pay and companyperformance


49/81


CEO Salaries

Workers CEOs

1970 $32,522 $1.3 Mil.

1999 $35,864 $37.5 Mil.

CEO compensation has gone from 40times the pay of average worker to over1000 times


50/81


CEO Salaries

Although originally thought that executivecompensation reflected reward for talent,recent evidence suggests managers

have been able to manipulate boards toextract compensation out of line witheconomic contribution


51/81


CEO Salaries

How have they been able to do this?

1. Boards dont typically have necessaryinformation and independence to negotiate

effectively2. Managers have introduced forms of

compensation that camouflage theextraction of rents from shareholders

Stock options (not counted as expenses)


52/81


CEO Salaries

Rent extraction has increased asconsultants are hired to determineappropriate pay for CEO

Firm usually wants to provide at least theaverage of other companies, so salarieshave been rising rapidly

With publicity increasing, CEO salariesseem to be rising less rapidly

Th P i i l A t P bl


53/81


The Principal Agent Problem Public Enterprises

Observations

Managers goals may deviate from the

agencies goals (size)

Oversight is difficult (asymmetric information)

Market forces are lacking


54/81



Limitations to Management Power

Managers choose a public service position

Managerial job market

Legislative and agency oversight (GAO &OMB)

Competition among agencies

Th M f N P fit


55/81


The Managers of Non-ProfitHospitals as Agents

Are non-profit organizations more or lessefficient than for-profit firms?

725 hospitals from 14 hospital chains

Return on investment (ROI) and averagecost (AC) measured

Th M f N P fit


56/81



Return on Investment

1977 1981

For-Profit 11.6% 12.7%

Non-Profit 8.8% 7.4%

Th M f N P fit


57/81



After adjusting for differences in services:

AC/patient day in non-profits is 8% greaterthan profits

Conclusion Profit incentive impacts performance

Costs and benefits of subsidizing non-profitsmust be considered

I ti i th P i i l A t


58/81


Incentives in the Principal-AgentFramework

Designing a reward system to align theprincipals and agents goals--an example

Watch manufacturer

Uses labor and machinery

Owners goal is to maximize profit

Machine repairperson can influence reliability

of machines and profits



59/81




Revenue also depends, in part, on the quality

of parts and the reliability of labor

High monitoring costs make it difficult toassess the repairpersons work



60/81



Small manufacturer uses labor andmachinery to produce watches

Goal is to maximize profits

High monitoring costs keep owners frommeasuring the effort of the repairpersondirectly

The Re en e from Making


61/81


The Revenue from MakingWatches

Poor Luck Good Luck

Low Effort(a = 0)

$10,000 $20,000

High Effort

(a = 1) $20,000 $40,000

Incentives in the Principal Agent


62/81




Repairperson can work with either high or

low effort

Revenues depend on effort relative to theother events (poor or good luck)

Owners cannot determine a high or low effortwhen revenue = $20,000



63/81




Repairpersons goal is to maximize wage net

of cost

Cost = 0 for low effort

Cost = $10,000 for high effort

w(R) = repairpersons wage based only onoutput



64/81



Choosing a wage:

w = 0; a = 0; R = $15,000

R = $10,000 or $20,000, w = 0

R = $40,000; w = $24,000

R = $30,000; Profit = $18,000

Net wage = $2,000

w = R - $18,000 Net wage = $2,000

High effort



65/81



Conclusion

Incentive structure that rewards the outcomeof high levels of effort can induce agents to

aim for the goals set by the principals

Managerial Incentives in an


66/81


Managerial Incentives in anIntegrated Firm

In integrated firms, division managershave better (asymmetric) informationabout production than central

management

Two Issues

How can central management elicit accurate

information?How can central management achieve

efficient divisional production?



67/81



We will focus on firms that are integrated

Horizontally integrated

Several plants produce the same or relatedproducts

Vertically integrated

Firm contains several divisions, with some

producing parts and components that othersuse to produce finished products



68/81



Possible Incentive Plans

1. Give plant managers bonuses based oneither total output or operating profit

Would encourage managers to maximizeoutput

Would penalize managers whose plants

have higher costs and lower capacity No incentive to obtain and reveal accurate

cost and capacity information



69/81



2. Ask managers about their costs andcapacities and then base bonuses onhow well they do relative to their

answers Qf= estimate of feasible production level

B = bonus in dollars

Q = actual output

B = 10,000 - .5(Qf - Q)

Incentive to underestimate Qf



70/81



If manager estimates capacity to be18,000 rather than 20,000, and if theplant only produces 16,000, her bonus

increases from $8000 to $9000Dont get accurate information about

capacity and dont insure efficiency

Bonus still tied to accuracy of forecast



71/81



Modify scheme by asking managers howmuch their plants can feasibly produceand tie bonuses to it

Bonuses based on more complicatedformula to give incentive to reveal truefeasible production and actual output

If Q > Qf ,B = .3Qf+ .2(Q - Qf)If Q Qf ,B = .3Qf- .5(Qf - Q)



72/81



Assume true production limit is Q* =20,000

Line for 20,000 is continued for outputs

beyond 20,000 to illustrate the bonusscheme but dashed to signify the infeasibilityof such production

Bonus is maximized when firm produces at

its limit of 20,000; the bonus is then $6000

Incentive Design in an


73/81


Incentive Design in anIntegrated Firm

Output(units per year)

2,000

4,000

6,000

10,000

0 10,000 20,000 30,000 40,000

Bonus($ peryear)

8,000

IfQf=30,000,

bonus is $4,000,the maximumamount possible.

Qf =30,000

Qf =10,000

IfQf=10,000,bonus is $5,000.

Qf =20,000

IfQf = Q* =20,000,bonus is $6,000.


74/81


Efficiency Wage Theory

In a competitive labor market, all whowish to work will find jobs for a wageequal to their marginal product

However, most countries economies

experience unemployment


75/81



The efficiency wage theory can explainthe presence of unemployment and wagediscrimination

In developing countries, productivity dependson the wage rate for nutritional reasons


76/81



The shirking model can be better used toexplain unemployment and wagediscrimination in the United States

Assumes perfectly competitive markets

However, workers can work or shirk

Since performance information is limited,workers may not get fired


77/81



If workers are paid market clearing wagew*, they have incentive to shirk

If they get caught and fired, they can

immediately get a job elsewhere forsame wage

Firms have to pay a higher wage to makeloss higher from shirking

Wage at which no shirking occurs is theefficiency wage


78/81



All firms will offer more than marketclearing wage, w*, say we(efficiencywage)

In this case, workers fired for shirkingface unemployment because demand forlabor is less than market clearing quantity

Unemployment in a Shirking


79/81


Without shirking, the market wageis w*, and full-employment exists at L *

Demand forLabor

w*

L*

Unemployment in a ShirkingModel

Quantity ofLabor

Wage

SL

No-ShirkingConstraint

The no-shirkingconstraint gives

the wage necessaryto keep workers

from shirking.

we

L e

At the equilibrium wage, We thefirm hires L eworkers

creating unemployment ofL* - L e.

Efficiency Wages at Ford Motor


80/81


Efficiency Wages at Ford MotorCompany

Labor turnover at Ford

1913: 380%

1914: 1000%

Average pay = $2 - $3

Ford increased pay to $5

Efficiency Wages at Ford Motor


81/81

Efficiency Wages at Ford MotorCompany

Results

Productivity increased 51%

Absenteeism was halved

Profitability rose from $30 million in 1914 to$60 million in 1916

Documents

Ch17_Pindyck