Chapter 17
Markets with Asymmetric Information
Chapter 17 2©2005 Pearson Education, Inc.
Topics to be Discussed
Quality Uncertainty and the Market for Lemons
Market SignalingMoral HazardThe Principal-Agent ProblemManagerial Incentives in an Integrated
FirmAsymmetric Information in Labor
Markets: Efficiency Wage Theory
Chapter 17 3©2005 Pearson Education, Inc.
Introduction
When one party has more information about relevant variables than the other, we have asymmetric information
Asymmetric information is all around us Sellers usually know more about the quality
of their product than buyers Workers know more about their own skills
and abilities than employers Managers know more about investment
opportunities than shareholders
Chapter 17 4©2005 Pearson Education, Inc.
The Market for Lemons
Ex: The market for used cars Two types of used cars – high quality and
low quality First, assume that buyers and sellers can
distinguish between the carsThere will be two markets – one for high quality
cars and one for low quality cars
Chapter 17 5©2005 Pearson Education, Inc.
The Market for Lemons
SH is higher than SL because owners of high quality cars generally require a high price to sell them
Similarly, DH is higher than DL because consumers are generally willing to pay more for higher quality cars
Chapter 17 6©2005 Pearson Education, Inc.
The Market for Lemons
PH PL
QH QL
SH
SL
DH
DL
$5,000
50,00050,000
$10,000
Price for low quality cars is $5000.
Price for high quality cars is $10,000.
50,000 of each type are sold.
Chapter 17 7©2005 Pearson Education, Inc.
The Market for Lemons
Now assume sellers can distinguish between the cars but buyers cannot This will allow us to see the role of asymmetric
informationSince, under complete information, the same
number of high- and low-quality cars are sold, under asymmetric information buyers may think the odds that a car is high quality are 50/50 Since buyers can’t distinguish, they will view all cars
as medium quality (with demand DM)Keep in mind that no car is actually of medium
quality
8©2005 Pearson Education, Inc.
The Market for Lemons
PH PL
QH QL
SH
SL
DH
DL
5,000
50,00050,000
10,000
DL
HQ cars are perceived as MQ and sell for $7500.
DM
25,000
7,500
DM
75,000
7,500
DH
LQ cars are perceived as MQ and sell for $7500.
25,000 HQ cars end up being sold.
75,000 LQ cars end up being sold.
Chapter 17 9©2005 Pearson Education, Inc.
The Market for Lemons
Since only 25,000 HQ cars are sold while 75,000 LQ cars are sold, a buyer will change her subjective probability that the car is high quality to 25/75 from 50/50 Buyers will now view all cars as “low-medium”
quality (with demand DLM)
10©2005 Pearson Education, Inc.
The Market for Lemons
PH PL
QH QL
SH
SL
DH
DL
5,000
50,00050,000
10,000
DL
Price will drop below $7500.
DM
25,000
7,500
75,000
7,500
DM
DLM
DLM
Subjective probability of buying a HQ car will drop below 25/75.
Chapter 17 11©2005 Pearson Education, Inc.
The Market for Lemons
You get the picture: as price drops, a buyer’s subjective probability of the proportion of HQ cars decreases, and this in turn causes price to drop even further
This vicious cycle will continue until the price drops to PL and all cars offered for sale are LQ. In well-functioning markets, this cycle takes a
very short time to complete
Chapter 17 12©2005 Pearson Education, Inc.
The Market for Lemons
With asymmetric information: Low quality goods can drive high quality
goods out of the market - the lemons problem
The market can fail (in the above example, mutually beneficial trade for HQ cars does not occur)
Adverse selection occurs; in equilibrium, trade takes place mainly for lower quality goods
Chapter 17 13©2005 Pearson Education, Inc.
Market for Health Insurance
Ex: Health Insurance Older individuals have difficulty purchasing health
insurance at almost any price – Why?They know more about their own health than the
insurance company doesUnhealthy people are more likely to want insurance
than healthy people; therefore the proportion of unhealthy people in the pool of those insured rises
As the pool tilts toward unhealthy, prices rise; in turn the pool tilts even more toward unhealthy
Eventually, the only people left are the most unhealthy, and the price of insurance becomes extremely high
Chapter 17 14©2005 Pearson Education, Inc.
Market for Auto Insurance
Ex: Auto insurance Insurance companies know that some drivers
have a low probability of being in an accident and some have a high probability
If they can’t distinguish among drivers, they will base premiums on the average experience
Safe drivers will... (you complete the rest)
Chapter 17 15©2005 Pearson Education, Inc.
Market for Insurance
How do we solve the problem of asymmetric information?
A possible solution is to pool risks The government can take on this role, as
with MedicareProblem of adverse selection is eliminated
Chapter 17 16©2005 Pearson Education, Inc.
Market for Credit
Ex: Credit cards Asymmetric information regarding the ability
to repay debt Potential for lemons problem To solve this, banks and credit agencies use
credit histories, and even share credit histories, to gauge credit-worthyness
Chapter 17 17©2005 Pearson Education, Inc.
Importance of Reputation and Standardization
Asymmetric information is everywhere Antiques, art, rare coins – real or counterfeit? Restaurants – check the kitchen? Home repair – climb up to see job?
Chapter 17 18©2005 Pearson Education, Inc.
Implications of Asymmetric Information
How can these producers provide high-quality goods when asymmetric information will drive out high-quality goods through adverse selection? Reputation
You hear about restaurants or stores that have good or bad service and quality
StandardizationChains that keep production the same everywhere
You look forward to a Big Mac when traveling, even if you would not typically buy one at home, because you know what to expect
Chapter 17 19©2005 Pearson Education, Inc.
Market Signaling
Another way to alleviate the problem of asymmetric information is through market signaling Market signaling is the process by which
sellers send information to buyers about product quality
For example, a potential worker (selling labor) will usually submit a resume to the potential employer (buying labor), thereby signaling his/her ability and skills
Chapter 17 20©2005 Pearson Education, Inc.
Market Signaling
Not all signals are the same Dressing well for the interview is a weak
signalWeak because even unproductive employees
can dress well Having an advanced degree from a reputable
university is a strong signalTo be effective, a signal must be difficult for
sellers of LQ products to imitate Degree in Financial Engineering from Polytechnic
Chapter 17 21©2005 Pearson Education, Inc.
Model of Job Market Signaling
Simple model of signaling Two groups of workers Group I: Low productivity
Marginal (and average) product of labor = 1 unit per year
Group II: High productivityMarginal (and average) product of labor = 2 units per
year There is an equal number of Group I and Group II
workersAverage product of labor for all workers = 1.5 units
per year
Chapter 17 22©2005 Pearson Education, Inc.
Model of Job Market Signaling
Assume competitive markets (no profit) Price = $10,000/unit Employees average 10 years of employment Revenue from Group I worker = $100,000
(1 x $10,000 x 10 years) Revenue from Group II worker = $200,000
(2 x $10,000 X 10 years)
Chapter 17 23©2005 Pearson Education, Inc.
Model of Job Market Signaling
With Complete Information Wage = revenue from worker
Group I wage = $10,000/yr.Group II wage = $20,000/yr.
With Asymmetric Information Wage = average revenue from all workers
Group I & II wage = $15,000/yr.Group I workers essentially gain at the expense
of group II workers
Chapter 17 24©2005 Pearson Education, Inc.
Model of Job Market Signaling
Consider a potential signal: y = education index (e.g., years of education,
reputation of university, GPA, etc.) C(y) = cost of attaining educational level y
Direct costs tuition textbooks
Indirect costs opportunity cost of doing something else suffering through FE 601
Chapter 17 25©2005 Pearson Education, Inc.
Model of Job Market Signaling
Assume education is more costly for Group I for any level of y Why?
Low productivity workers may simply be less studious
Low productivity workers progress slower through degree programs
Group I CI(y) = $40,000y
Group II CII(y) = $20,000y
Chapter 17 26©2005 Pearson Education, Inc.
Model of Job Market Signaling
Assume education is only valuable as a potential signal Does not improve your skills
Can we find an equilibrium where Group I workers choose less education and Group II workers choose more education?
We have an equilibrium if we can identify a level y* such that Group I optimally chooses < y* and Group II chooses ≥ y* Employers will be able to tell workers apart based on
level of education
Chapter 17 27©2005 Pearson Education, Inc.
Model of Job Market Signaling
Some observations: Since y* is the only threshold and education is costly,
no one will choose y > y* and no one will choose y* > y > 0
y > y* sends the same signal as y = y*y* > y > 0 sends the same signal as y = 0
Ed level of y* signals GII and wage = $20,000/yr
Ed level of 0 signals GI and wage = $10,000/yr
Note that there is no 1 right solution to the level of y*, but for education to be an effective signal, firms must identify a correct solution
Chapter 17 28©2005 Pearson Education, Inc.
Model of Job Market Signaling
How much education will individuals obtain given that firms use this decision rule?
Benefit of education B(y) is the increase in wage associated with each level of education
B(y) is initially 0, which corresponds to the $100,000 base 10-year earnings B(y) continues to be zero until ed reaches y*
Chapter 17 29©2005 Pearson Education, Inc.
Model of Job Market Signaling
How much education to choose is a cost-benefit comparison Obtain education level y* if the benefit, B(y), is at least as
large as the cost, C(y). Group I:
No education if $100,000 < $40,000y* no education if y* > 2.5
Group II: No education if $100,000 < $20,000y* no education if y* > 5
Therefore, any level of y* between 2.5 and 5 will work in distinguishing Group I workers from Group II workers Make sure you understand this
Chapter 17 30©2005 Pearson Education, Inc.
Model of Job Market Signaling
Suppose y* = 4 People in Group I will find education is not
worthwhile, since $100,000 < $40,000×4Benefit lower than cost
People in Group II will find education is worthwhile, since $100,000 > $20,000×4
Benefit higher than cost
Therefore, y* = 4 is an effective separating equilibrium
31©2005 Pearson Education, Inc.
Model of Job Market Signaling
Value ofCollege
Educ.
$100K
Value ofCollege
Educ.
Educationlevel
Educationlevel
0 1 2 3 4 5 6 0 1 2 3 4 5 6
$200K
$100K
$200KCI(y) = $40,000y
B(y)B(y)
y* y*
CII(y) = $20,000y
Group IIGroup I
Chapter 17 32©2005 Pearson Education, Inc.
Model of Job Market Signaling
Conclusion Education is valuable as a signal, even in the
absence of other benefitsIn reality, of course, education makes you more
productive
Chapter 17 33©2005 Pearson Education, Inc.
Market Signaling
Another example of market signaling: Guarantees and warranties Sellers of HQ products signal quality and the
cost to LQ producers of offering guarantees and warranties is too high
Chapter 17 34©2005 Pearson Education, Inc.
Moral Hazard
Moral hazard occurs when a party whose actions are unobserved can affect the probability or magnitude of an adverse event If your home is insured, you might be less likely to
lock doors or install a security system If you have a flat salary, you might decide to work
less hard Here we have hidden action rather than hidden
information
Chapter 17 35©2005 Pearson Education, Inc.
Moral Hazard
Ex: Determining the Premium for Fire Insurance Warehouse worth $100,000 Probability of a fire:
.005 if conduct a $50 fire-prevention program.01 if don’t conduct the program
If the insurance company cannot monitor the company’s decision whether to conduct the fire-prevention program, how does it set the premium?
Chapter 17 36©2005 Pearson Education, Inc.
Moral Hazard
With the program the premium is: 0.005 x $100,000 = $500
However, once insurance is purchased, the owners no longer have an incentive to run the program Probability of fire increases to 0.01 Premium increases to (0.01 x $100,000) =
$1000
Chapter 17 37©2005 Pearson Education, Inc.
The Principal-Agent Problem
This leads us to a discussion of agencyTypically, owners (principals) do not
continuously monitor managers (agents), so the managers can potentially take actions which are detrimental to the owners
This creates a principal-agent problem which arises when agents pursue their own goals, rather than the goals of the principal
Chapter 17 38©2005 Pearson Education, Inc.
The Principal-Agent Problem
Some examples of principals and agents Company owner (principal) wants to
maximize profits, managers and employees (agents) want to take it easy
U.S. residents (principals), politicians (agents)
Parents of a baby (principals), babysitter (agent)
Chapter 17 39©2005 Pearson Education, Inc.
The Principal-Agent Problem
A Principal-Agent Problem in the Private Sector: Management vs. Shareholders Ownership of most U.S. firms is not highly
concentrated It is costly to monitor managers, and when
ownership is diffuse the benefits don’t justify the costs for any 1 shareholder
Chapter 17 40©2005 Pearson Education, Inc.
The Principal-Agent Problem
Managers may pursue their own objectives Increase company size and market share,
which ultimately will lead to more respect and power
Increase own salary and bonuses Increase own job security
Chapter 17 41©2005 Pearson Education, Inc.
The Principal-Agent Problem
Limitations to managers’ ability to deviate from owners’ objectives Stockholders can oust managers with help
from board of directors Another firm may attempt a takeover Compensation can be structured to better
align managers’ and owners’ incentives
Chapter 17 42©2005 Pearson Education, Inc.
The Principal-Agent Problem
The problem of limited stockholder control shows up in executive compensation In 2002, Business Week reported that, on
average, CEOs of large corporations earn $13.1 million per year, and that the annual growth rate in executive compensation has been in double-digits
Chapter 17 43©2005 Pearson Education, Inc.
CEOs vs. All Workers: Salaries
All Workers Top 100 CEOs
1970 $32,522 $1.3 Mil.
1999 $35,864 $37.5 Mil.
CEO compensation has gone from 40 times the pay of average worker to over 1000 times*
*Based on 1/18/2003 NYT article
Chapter 17 44©2005 Pearson Education, Inc.
Incentives in the Principal-Agent Framework
When a principal-agent problem arises, the best solution is to design a reward system for the agent such that his/her incentives are aligned with the principal’s goals This usually involves tying the reward to
some observable variable
Chapter 17 45©2005 Pearson Education, Inc.
Incentives in the Principal-Agent Framework
Ex: Watch manufacturer Uses labor and machinery Manufacturer’s goal is to maximize profit Machine repairperson can work with low effort
or high effortLow effort increases the probability that machines
will break down, thereby reducing profitsHigh effort reduces the probability that machines
will break down, thereby increasing profitsEffort is not observed (monitoring too costly)
Chapter 17 46©2005 Pearson Education, Inc.
Incentives in the Principal-Agent Framework
Some other assumptions: Revenue also depends on other factors
besides repairperson’s effort Effort cannot be deduced from revenue level
This means that effort is not ex-post verifiable
Chapter 17 47©2005 Pearson Education, Inc.
The Revenue from Making Watches
Revenue
Poor Luck Good Luck
Low Effort
(a = 0)$10,000 $20,000
High Effort
(a = 1)$20,000 $40,000
effort not
ex-post
verifiable
Chapter 17 48©2005 Pearson Education, Inc.
Incentives in the Principal-Agent Framework
Repairperson’s (agent’s) goal Maximize wage less the cost of effort
Assume cost = 0 for low effort, cost = $10,000 for high effort
Note that this goal implies that the repairperson is risk-neutral
Owner’s (principal’s) goal Maximize revenue less repairperson’s wage
Choose a payment scheme that will maximize profits Since effort is not observable, the wage can only be
set as a function of revenue, not effortw(R) = repairperson’s wage as a function of revenue
Chapter 17 49©2005 Pearson Education, Inc.
Incentives in the Principal-Agent Framework
Lets compare two payment schemes: Scheme1: Fixed wage
w = 0 (why zero? How much does the wage need to be if cost is 0?)
Then, effort = low, wL – cost(effortL) = 0
E[profit] = E[revenue] – E[wL] = $15,000 - $0 = $15,000
Scheme2: Bonus arrangement w = 0 if r = $10,000 or $20,000 w = $24,000 if r = $40,000 Then, repairperson will choose effort = high, since
E[wH] – cost(effortH) is $12,000 - $10,000 = $2000 but E[wL] – cost(effortL) is $0 - $0 = $0
Incentivize the repairperson to choose effort = high
E[profit] = E[revenue] – E[wH] = $30,000 - $12,000 = $18,000
Thus, both are better off under the bonus arrangement
Chapter 17 50©2005 Pearson Education, Inc.
Incentives in the Principal-Agent Framework
Suppose there are many repairpersons, so there is a lot of competition for the job
If you are the watch manufacturer what payment scheme would you choose? Clearly, a fixed wage is not a good idea The bonus from previous slide is a better idea But even better:
w = 0 if revenue = $10,000 or $20,000w = $20,001 if revenue = $40,000 This scheme allows you to extract maximum
rents; E[profit] = $19,999.50
Chapter 17 51©2005 Pearson Education, Inc.
Incentives in the Principal-Agent Framework
Conclusion A clever incentive structure can induce the
agent to aim for the goals set by the principal
Chapter 17 52©2005 Pearson Education, Inc.
Managerial Incentives in an Integrated Firm
In integrated firms, division managers have better (asymmetric) information about production than central management
Two Issues How can central management elicit accurate
information? How can central management achieve
efficient divisional production?
Stop
here
Chapter 17 53©2005 Pearson Education, Inc.
Managerial Incentives in an Integrated Firm
We will focus on firms that are integratedHorizontally integrated
Several plants produce the same or related products
Vertically integrated Firm contains several divisions, with some
producing parts and components that others use to produce finished products
Chapter 17 54©2005 Pearson Education, Inc.
Managerial Incentives in an Integrated Firm
Possible Incentive Plans
1. Give plant managers bonuses based on either total output or operating profit
Would encourage managers to maximize output
Would penalize managers whose plants have higher costs and lower capacity
No incentive to obtain and reveal accurate cost and capacity information
Chapter 17 55©2005 Pearson Education, Inc.
Managerial Incentives in an Integrated Firm
2. Ask managers about their costs and capacities and then base bonuses on how 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
Chapter 17 56©2005 Pearson Education, Inc.
Managerial Incentives in an Integrated Firm
If manager estimates capacity to be 18,000 rather than 20,000, and if the plant only produces 16,000, her bonus increases from $8000 to $9000 Don’t get accurate information about capacity
and don’t insure efficiency
Bonus still tied to accuracy of forecast
Chapter 17 57©2005 Pearson Education, Inc.
Managerial Incentives in an Integrated Firm
Modify scheme by asking managers how much their plants can feasibly produce and tie bonuses to it
Bonuses based on more complicated formula to give incentive to reveal true feasible production and actual output If Q > Qf ,B = .3Qf + .2(Q - Qf)
If Q Qf ,B = .3Qf - .5(Qf - Q)
Chapter 17 58©2005 Pearson Education, Inc.
Managerial Incentives in an Integrated Firm
Assume true production limit is Q* = 20,000 Line for 20,000 is continued for outputs
beyond 20,000 to illustrate the bonus scheme but dashed to signify the infeasibility of such production
Bonus is maximized when firm produces at its limit of 20,000; the bonus is then $6000
Chapter 17 59©2005 Pearson Education, Inc.
Incentive Design in an Integrated Firm
Output(units per year)
2,000
4,000
6,000
10,000
0 10,000 20,000 30,000 40,000
Bonus($ peryear)
8,000
If Qf = 30,000,bonus is $4,000,
the maximumamount possible.
Qf = 30,000
Qf = 10,000
If Qf = 10,000,bonus is $5,000.
Qf = 20,000
If Qf = Q* = 20,000,bonus is $6,000.
Chapter 17 60©2005 Pearson Education, Inc.
Efficiency Wage Theory
In a competitive labor market, all who wish to work will find jobs for a wage equal to their marginal product However, most countries’ economies
experience unemployment
Chapter 17 61©2005 Pearson Education, Inc.
Efficiency Wage Theory
The efficiency wage theory can explain the presence of unemployment and wage discrimination In developing countries, productivity depends
on the wage rate for nutritional reasons
Chapter 17 62©2005 Pearson Education, Inc.
Efficiency Wage Theory
The shirking model can be better used to explain unemployment and wage discrimination in the United States Assumes perfectly competitive markets However, workers can work or shirk Since performance information is limited,
workers may not get fired
Chapter 17 63©2005 Pearson Education, Inc.
Efficiency Wage Theory
If workers are paid market clearing wage w*, they have incentive to shirk
If they get caught and fired, they can immediately get a job elsewhere for same wage
Firms have to pay a higher wage to make loss higher from shirking
Wage at which no shirking occurs is the efficiency wage
Chapter 17 64©2005 Pearson Education, Inc.
Efficiency Wage Theory
All firms will offer more than market clearing wage, w*, say we (efficiency wage)
In this case, workers fired for shirking face unemployment because demand for labor is less than market clearing quantity
Chapter 17 65©2005 Pearson Education, Inc.
Without shirking, the market wageis w*, and full-employment exists at L*
Demand forLabor
w*
L*
Unemployment in a Shirking Model
Quantity of Labor
Wage
SL
No-ShirkingConstraint
The no-shirkingconstraint gives
the wage necessaryto keep workers
from shirking.
we
Le
At the equilibrium wage, We the firm hires Le workerscreating unemployment of L* - Le.
Chapter 17 66©2005 Pearson Education, Inc.
Efficiency Wages at Ford Motor Company
Labor turnover at Ford 1913: 380% 1914: 1000%
Average pay = $2 - $3Ford increased pay to $5
Chapter 17 67©2005 Pearson Education, Inc.
Efficiency Wages at Ford Motor Company
Results Productivity increased 51% Absenteeism was halved Profitability rose from $30 million in 1914 to
$60 million in 1916
Chapter 17 68©2005 Pearson Education, Inc.
Important concepts
Asymmetric informationLemons problemAdverse selectionPooling risksMarket signalingSeparating equilibriumMoral hazardPrincipalAgentPrincipal-agent problem
Chapter 17 69©2005 Pearson Education, Inc.
HomeworkQuestions for review: 4 Exercises: 5,8,9,10
Due in 2 weeks
Next week: final exam