Transcript Asymmetric Information

Asymmetric
Information
Chapter 23
Slides by Pamela L. Hall
Western Washington University
©2005, Southwestern
Introduction


Managers (insiders) of firms can increase their profit by taking actions
based on insider information (information that is not available to public)
If insiders obtain positive (negative) information about a company, they
buy (sell) the company’s stock with expectation that stock’s price will rise
(fall) when positive (negative) information is publicly announced
 Such insider trading of securities is illegal
• However, incorporating information not known by all agents in market pricing is
generally legal


When selling a commodity, agents are not legally required to provide full
disclosure of information on a commodity
Previously, we generally implicitly assumed or explicitly stated symmetry
in market information as a characteristic of market structure
 Assumed all agents had costless access to this information
• Symmetry existed with both buyers and sellers having the same market
information

For example, one of the explicit characteristics of perfect competition is agents’ perfect
knowledge
2
Introduction

In general, market information is costly, and this cost may vary between
buyers and sellers
 Resulting in asymmetric information held by agents
• One set of agents may be more knowledgeable about a commodity than another
set
 Information cost may vary among agents as a result of differences in
education and experience about commodity
• Examples include




A firm possessing limited information about a potential worker’s abilities
A used car buyer not having complete repair and maintenance history on an auto
An insurance company not knowing risky behavior of a potential insurer
When commodities are homogeneous and their characteristics are
readily available
 Cost of determining these characteristics is small
 Assumption of market symmetry would generally hold
• For example, symmetric information generally holds for commodity futures market

Where, except for delivery dates, all futures contracts on same commodity have
identical characteristics
3
Introduction

In contrast, asymmetric information will generally
exist for heterogeneous commodities with
characteristics that are costly to determine
 An example is the vehicle market
• Condition of a vehicle is difficult to determine without costly
testing


Heterogeneous nature of used vehicles prevents a general
determination of a vehicle’s condition based on examination of other
like vehicles
A major consequence of asymmetric information is
possible disappearance of markets
 Which result in an inefficient allocation of resources
4
Introduction

Aim in this chapter is to demonstrate how missing markets
and associated efficiency losses result in presence of
asymmetric information
 Asymmetry in information generates two types of outcomes
• Adverse selection


Where one agent’s decision depends on unobservable characteristics that
adversely affect other agents
 Use used-automobile market to illustrate missing market and resulting
market inefficiencies
We discuss mechanisms of signaling and screening as second-best Paretoefficient mechanisms for addressing these inefficiencies
• Moral hazard


A contract is signed among agents with one agent being dependent on
unobservable actions of other agents
Using a principal-agent model, we derive inefficient level of precaution taken
by agents
 Evaluate mechanisms (such as coinsurance) designed to address
inefficiencies
5
Introduction

Asymmetric information is relatively new area for applied economic analysis
 In 1970 George A. Akerlof was first to address problems and solutions associated with adverse
selection



Knowledge of moral hazard has been around since advent of insurance in 18th century
However, only recently have applied economists investigated ramifications of moral hazard
on economic efficiency
Asymmetric information in a market can result in market inefficiencies
 If information concerning characteristics of a commodity is not freely available, inefficient
allocations may result

One type of asymmetric information is called adverse selection (also called hidden
information)
 An informed agent’s decision depends on unobservable characteristics that adversely affect
uninformed agents
• Classic example is market for used cars

Assume used cars can be grouped by quality into two groups
 Within each group, used cars are homogeneous and are associated with a single price
• Comparing vehicles between groups, they are heterogeneous in quality and thus are not valued by identical
•
prices
In a free (symmetric) information case, two heterogeneous groups of commodities (used cars) would have
separate markets with associated prices p1 and p2
6
Lemons Market

Generally, sellers of a used car know vehicle’s history
 Can determine its market price at zero or very minimal cost

In contrast, buyers do not have this knowledge
 Cost of determining information for each group of used cars is prohibitive
 Without information, buyers may base their market price determination on
average quality of used cars available


Asymmetry in information results in buyers only willing to pay up to
average price of used cars available, p
At average price sellers would be willing to supply only
 QS = qj(pj)
 pj ≤ p
• qj(pj) is supply of used cars in group j
 Above-average used cars, associated with pj > p, will not be offered for sale
• Sellers would be unwilling to supply their cars for less than vehicle’s market value
7
Lemons Market

Buyers realize above-average used cars will not enter market at average price
 So average quality of used cars offered in market is less than average quality of used
cars available

They will adjust downward their willingness-to-pay for used cars offered in the
market
 Only sellers who value their cars below this new lower price will supply vehicles
• Average quality of used cars offered in market will once again decline

Tâtonnement process will continue until only lowest-quality group of used cars
are offered and sold in market
 When only lowest-quality group is offered for sale, any asymmetry in information
vanishes


With symmetric information, buyers’ and sellers’ expected prices match and only
a market for lowest-quality group exists
Missing markets for other groups of used cars represent market failure
 These lowest-quality cars are popularly referred to as lemons
• Market failure associated with adverse selection is called the lemons problem
8
Lemons Market


In Figure 23.1, lemons problem is illustrated for two quality groups of
used vehicles, reliable cars, , and lemons, 
Curves S and S are supply curves for lemons and reliable cars,
respectively
 Supply curve for reliable cars is above lemons curve
• Indicates sellers of reliable cars are only willing to supply these higher-valued
vehicles at prices above the lemon
 Demand curves for lemons and reliable cars are represented by D and D,
respectively
• Buyers are willing to pay a higher price for reliable compared with lemons

So reliables demand curve is above lemons demand curve
 Given free information, market is able to discriminate between these two
types of cars
• So market-clearing prices exist for both reliables and lemons markets

Equilibrium price and quality for reliables are p and * and for lemons p and *
9
Figure 23.1 Lemons problem
10
Lemons Market

Asymmetric information in form of adverse selection
prevents buyers from freely distinguishing reliable cars from
lemons
 Buyers may know proportion of automobiles that are reliable and
lemons
• But are unable to distinguish quality of a given automobile


Overall market demand, QD, facing sellers will be horizontal
sum of lemons and reliables demand curves
Total supply of cars, QS, is horizontal sum of lemons and
reliables supply curves
 Resulting equilibrium price and quantity are p' and Q'
• Loss in ability of market to distinguish between reliables and lemons

Results in number of lemons offered on market increasing from * to S' and
number of reliables declining from * to S'
11
Lemons Market


Demand for lemons decreases from * to D' and demand for reliables
increases from * to D'
Imbalance within markets will result in some buyers who expect to
receive a reliable car instead receiving a lemon
 As buyers realize ratio of reliables to lemons is declining in market, they will

adjust their expected quality downward
Participation in number of buyers wanting a reliable car will decline as
expectation of obtaining a reliable car in market decreases
• Resulting downward shift in demand curve further drives reliable cars out of
market


Further erodes demand for reliable cars
Tâtonnement process will continue until buyers only expect lemons to be
supplied, so their market demand curve is D
 Such a market will then supply * automobiles at a price of p *, and a
missing market will exist for reliable cars
• Market is unable to allocate both supply of reliables and lemons efficiently to
buyers

It is unable to price discriminate across quality differences
12
Lemons Market


An efficient allocation would result in quality discrimination
 Buyers would have market choice of purchasing either reliable cars or lemons
Without market ability to quality discriminate, some buyers may by chance purchase a reliable
car
 But these may not be buyers with highest willingness-to-pay
• Failure of market to allocate commodities based on willingness-to-pay results in an inefficient allocation


Cause of this missing market and inefficient allocation of resources is an externality between
sellers of reliable cars and lemons
As illustrated in Figure 23.1, as number of sellers offering lemons increases
 Buyers’ expectations regarding quality of vehicles in market is affected
• Price buyers are willing to pay declines

Adversely affects sellers of reliable cars by preventing them from selling their vehicles and improving efficiency

Externality between sellers for reliable cars and lemons has distributional implications

 Owners of lemons may receive more than their automobile is worth and owners of reliable cars less
Buyers possessing limited information generally benefit sellers of lemon products
13
Lemons Market

Problem of adverse selection exists in other markets
 For example, in insurance market buyers of insurance know more about
their general health than any insurance company
• Unhealthy consumers are more likely to buy insurance


Because healthy consumers will find cost of insurance too high
Tâtonnement process will continue until only unhealthy consumers purchase insurance
 Will make selling insurance unprofitable
 Another example is labor market
• Workers’ potential productivity is unobservable by a hiring firm
• But workers themselves know their productive capabilities


Tâtonnement process will result in only less-productive workers being hired
Market failure resulting from adverse selection explains
 Why a new automobile declines in value once it is driven off lot
 Why insurance is so high for a previously uninsured driver or a person with

no medical history
Why salaries start low with a potential for frequent raises once a person is
hired
14
Second-Best Mechanism Designs

U.S. health care costs are nearly double that of other nations and
outpace inflation
 Firms and workers are faced with rising premiums and cutbacks in coverage

A national health insurance program can avoid inefficiencies of adverse
selection in health care
 By making purchase of insurance compulsory
• Unhealthy citizens benefit from insurance premiums below their expected health
•
costs
Healthy citizens can purchase insurance at lower rates


Such a government policy is called cross-subsidization
 Healthy consumers pay a portion of health care for unhealthy consumers
One justification in favor of Medicare for elderly
 By providing insurance for all elderly, adverse selection is eliminated
• However, without knowing agents’ private information, obtaining Pareto-optimal
•
allocation is not possible
Acquiring such information is costly

So only a constrained or second-best Pareto optimum can be obtained
15
Second-Best Mechanism Designs

In general, insurance companies can avoid adverse selection by offering
group health insurance plans at places of employment
 Called pooling—both healthy and unhealthy consumers are pooled together
• Insurance premiums are based on average cost of health care


Adverse selection is eliminated by requiring all employees to participate
Government agencies can improve functioning of markets by providing
free information or requiring product information prior to sale
 Many government agencies currently provide information useful for making
market decisions
• Examples include U.S. State Department cautioning tourists about visiting a
particular region, USDA publishing situations and outlooks for agricultural
commodities

An example of requiring product information is FDA’s requirement for
food labeling on processed foods
16
Signaling



Both buyers and sellers can potentially benefit from creating markets
that were missing due to adverse selection
Provides incentives for developing market mechanisms to mitigate
market failure associated with adverse selection
Signaling
 Mechanisms that transfer information from informed agent to uninformed
agent

Naive signal on part of a buyer
 Asking sellers quality of a commodity—for example, asking a used car
dealer condition of a car
• Cost of such a signal could be high if signal is inaccurate and commodity is
purchased
 An example of a particularly weak signal
• Where cost of providing a signal is the same for all sellers regardless of quality of
their product
 Appearance can be another weak signal
17
Signaling

For a strong signal, a signal must have an associated lower cost for
sellers offering relatively high-quality commodities
 Compared with cost for sellers offering poor-quality commodities
• Examples of strong signals used by firms are reputation and standardization


Firms offering higher-quality commodities have an advantage over other firms in
establishing a reputation for quality
 For example, construction subcontractors can provide a signal for quality
construction by developing a list of satisfied customers
One problem with reputation as a signaling device
 Delay associated with establishing a reputation
• Problem may be partially avoided

By supplementing reputation with guarantees and warranties as explicit signals of
product quality
 For example, in 1980s, as a counter to Japanese auto manufacturers’ reputation
for producing quality cars
 U.S. manufacturers offered extended 100,000-mile warranties as a signal of
improved quality
18
Signaling

Such signals are useful in cases where buyers lack information on
quality of some commodity that they do not purchase on a regular basis
 For regularly purchased commodities that vary in quality
• Firms will attempt to standardize commodities they are offering to signal quality


For example, a fruit and vegetable wholesaler will attempt to always offer same quality
of produce
Through standardization, sellers send a strong signal that buyers can
expect a quality product from them
 Some firms advertise such standardization as a market signal

In general, a signaling mechanism will be employed by informed agents
 Agents are not always the seller
• Agent could be an antique dealer purchasing items for his shop

Through experience, dealer will have a greater knowledge about market than sellers
 A reputable dealer could employ signaling mechanisms to separate him from
unreputable dealers
19
Signaling

Concept of signals was first developed by Michael Spence
in a labor market context
 A strong signal of a person’s labor productivity is education
• Education generally improves a person’s productivity

However, even if it did not, it is still a strong signal of productivity
• Any admission requirements to a university or college will only result in
higher-quality individuals entering the institution

“Quality in quality out” is signal sent to employers
• Consumers, firms, and government agencies have also used gender,
race, color, religion, and national origin as signals for labor productivity


But these signals, besides being illegal in U.S., are generally weak
Exceptions are when insurance companies target insurance rates by such
characteristics as age and gender
 Some segments of society may feel use of such discriminatory signals
is morally wrong and thus should always be illegal
20
Signaling

Economic theory does not pass judgment on morality of
signals
 But it does provide a framework for determining economic
consequences of restricting such signals
• Theory would indicate that a government restriction on one signal would
result in firms adopting related signaling mechanisms to maintain profits

For used car market, reliable car dealers will be able to offer
signals
 For instance, in form of warranties
• At a lower cost than lemon dealers, as illustrated in Figure 23.2

Lower warranty cost will result in lemon dealers being unable to compete in
offering warranties
 Thus, only reliable dealers offer warranties
21
Figure 23.2 Signaling
22
Signaling

Through these signals market for used vehicles can
now be separated into two markets
 Lemons and reliables
• Market equilibrium for lemons is Pareto efficient by corresponding
to free-information equilibrium (p*, *)
• Market supply and demand curves for lemons did not shift

Introduction of a signal for reliable cars established a separate
market for lemons
• Market supply curve for reliable cars shifts up from S to S'

Represents increased cost associated with offering warranties
• As a result of supply shift, equilibrium quantity of reliable cars is
below free-information quantity of *

Equilibrium price of p' is above free-information price of p*
 Results in a deadweight loss area of CAB
23
Signaling


Such a market equilibrium is called a separating equilibrium
 It segments pooled market for lemons and reliables into two markets
However, this is only a second-best Pareto-efficient
outcome
 Because in markets with free information, sellers do not incur extra
expense of signals
• Deadweight loss of removing inefficiency



Cost of removing externality
Both producer and consumer surplus loss
Proportion of costs paid by buyers and sellers depends on
relative elasticities of supply and demand for reliable cars
 In long run, as elasticity of supply becomes more elastic
• Larger proportion of signal cost is passed on to buyers
24
Screening

Top three techniques to prevent used car scams
 Have a mechanic inspect vehicle
 Run a Vehicle History Report
• Will reveal if vehicle was flooded, rebuilt, salvaged, stolen, or totaled
 Never sign anything stating “as is, no warranty”

• Obtain at least a 30-day warranty
Symmetric information associated with free information results in a
Pareto-efficient allocation
 Pareto preferred to an allocation with signals
 However, signals can be a second-best Pareto-efficient outcome if they
result in a separating equilibrium, which improves efficiency

• Not all signals do this
Weak signals resulting in a pooling equilibrium
 Signals of different quality sellers cannot be differentiated
• Do not separate markets so market efficiency is not improved
 Buyers may attempt to distinguish or screen various commodities offered
25
Screening

Screening exists when a buyer employs a mechanism for sorting
commodities offered by sellers
 Examples of screening are

• Buyer having a used car inspected prior to purchase
• Employer offering internships prior to employment
In general, screening is employed by uninformed agent
 Can be either buyer or seller

For example, price discrimination, discussed in Chapter 13, is a form of
screening
 Seller does not have information on buyers’ willingness-to-pay for

commodity
By screening buyers based on their characteristics, sellers can create
separate markets and practice price discrimination
26
Screening

In some cases, buyers rely on another firm or
consumer (third parties) for screening
 For example
• A consumer may acquire her dentist, house painter, doctor, or
maid through a recommendation from another consumer
• Or a firm may screen commodities and sell information to
potential buyers
• Magazine Consumer Reports is in the business of screening
commodities
 Major third parties for screening are government
agencies providing market information
27
Screening

Consider used car market
 Potential buyers may screen vehicles by having them inspected

As illustrated in Figure 23.3, cost of screening will shift demand for both
lemons and reliable cars downward
 From D to D' for lemons market and from D to D' for reliable market
• Resulting separating-equilibrium prices, p' and p', are lower than free-
•
information equilibrium prices, p* and p*
Separating-equilibrium quantities, ' and ', are lower than free-information
equilibrium quantities, * and *


Cost of screening is sum of deadweight losses in lemons market (shaded area CAB)
and reliable market (shaded area DEF)
Both signaling and screening have potential for reducing asymmetric
information and yielding a second-best Pareto-efficient outcome
 Cost of reducing asymmetric-information externality
• Cost of signals or screening may offset any market efficiency gains

They may or may not improve social welfare
28
Figure 23.3 Screening
29
Principal-Agent Models

Concept of moral hazard
 Developed from study of insurance market
 An insurer has no control over policyholder not taking precautions toward reducing
probability of an insured event from occurring

Term moral hazard (also called hidden actions) is derived from condition that a
policyholder may take wrong (immoral) action by not taking proper precautions
 For example, an auto insurance firm has no control over hidden action of a
policyholder leaving car keys in an unlocked car



Moral hazard lasts over life of some established agreement
Moral hazard may result if purchase of a commodity establishes future returns or
utility of an agent being dependent on actions of another agent
Moral hazard is not restricted to issuance of insurance
 It generally exists whenever one agent (principal) depends on another agent (agent)
to undertake some actions
• If agent’s actions are hidden from principal, asymmetric information is present

Market inefficiencies may result
30
Principal-Agent Models

In general, contracts establishing such dependence
are designed to mitigate potential moral hazard
problems
 Problems in designing contracts result from principalagent problem
• Examples include


Owners of a firm who are unable to observe a manager’s work ethic
Instructor’s inability to observe how hard a student is actually
studying
• In these examples, agents have ability to hide actions

Uninformed principal wants to provide informed agent with efficient
incentives for fulfilling contract
31
Pareto Efficiency with No Moral
Hazard

As an illustration of no moral hazard, assume agents face an expected loss
associated with some event
 Examples are losses from fire damage to their business or an auto accident

Without any insurance, consumers face full cost of some negative event, which
reduces their welfare
 Can mitigate negative impact by taking precaution
• An increase in level of precaution can both reduce likelihood of event occurring and
magnitude of loss when event does occur


An objective of a consumer is to determine optimal level of precaution, P
Assume total cost of precaution at first increases at a decreasing rate and then
increases at an increasing rate with level of precaution (Figure 23.4)
 A basic level of precaution offers a great deal of protection with little increases in cost
• Examples are driving with traffic instead of against it, locking your car when shopping
 At basic level of precaution, precaution costs are increasing but at a decreasing rate
• At some point an additional level of precaution will result in costs increasing at an increasing
rate
32
Figure 23.4 Total and marginal
cost curves for precaution
33
Pareto Efficiency with No Moral
Hazard

If TC(P) is total cost function for precaution
 Then TC'(P) > 0 and at first TC"(P) < 0 and at some precaution level TC"(P) > 0

For example, at first a great deal of fire protection can be purchased with a small
investment in a smoke detector
 For additional protection, fire extinguishers can be purchased at a higher cost per unit
• Followed by a sprinkler system at an even higher cost per unit

Associated with a given level of precaution is an expected loss
 Expected loss is probability of event occurring times total loss

Objective of a consumer is to determine efficient level of precaution that
minimizes overall cost (sum of expected losses and cost of precaution)
 F.O.C. is
• TC'(P) = -EL'(P)



TC'(P) is marginal cost, MC(P)
-EL'(P) is marginal benefit of precaution, MB(P)
Marginal benefit is reduction in expected losses associated with an increase in precaution
34
Pareto Efficiency with No Moral
Hazard


EL'(P) < 0, so marginal benefit, -EL'(P), is positive (Figure 23.5)
Optimal level of precaution, P*
 Where marginal cost equals marginal benefit
• If marginal benefit is greater than marginal cost

An increase in precaution would reduce EL(P) more than increase in TC(P)
 So overall costs fall
• If marginal benefit is less than marginal cost


A decrease in precaution would reduce TC(P) by more than increase in EL
 So overall cost will fall
Optimal level of precaution is illustrated in Figure 23.6
 Positively sloping marginal precaution-cost curve represents assumption of increasing


per-unit precaution cost
Negatively sloping marginal precaution-benefit curve represents assumption of
decreasing reduction in expected loss as precaution increases
At P*, where marginal benefit equals marginal cost
• Overall costs are minimized
 To left (right) of P*, marginal benefit is greater (less) than marginal cost
• Consumer has an incentive to increase (decrease) precaution
35
Figure 23.5 Expected losses
36
Figure 23.6 Pareto-efficient precaution
level with no moral hazard
37
Insurance Market with No Moral
Hazard

As discussed in Chapter 18, a risk-averse agent will not voluntarily take
on additional risk
 Will seek out opportunities for avoiding risk

Insurance allows an agent to shift risk of a negative event onto another
agent (an insurance company)
 In event of a loss, such as a flood, an insurance company compensates
agent for loss

Assume contract (policy) between principal (insurance company) and
agent (consumer) is actuarially fair insurance
 If consumer can purchase insurance covering full expected loss for a given
level of precaution
• Consumer no longer suffers a loss from a negative event, EL = 0
 However, consumer must pay premium, which, for actuarially fair insurance,
is equal to EL
• Assuming no moral hazard, insurance company will want to design a policy where
expected payout varies by level of precaution a consumer takes
38
Insurance Market with No Moral
Hazard



Premiums would be higher for a low level of precaution by a consumer
 Decline as level of precaution increases
Let A(P) represent insurance premium, so A'(P) < 0
 A"(P) > 0
Consumer is still faced with problem of determining optimal level of precaution that
minimizes overall cost of taking precaution and now paying insurance premium

F.O.C. is

 TC'(P) = -A'(P)
Consumer equates marginal precaution cost, TC'(P), to marginal precaution benefit, -A'(P)
 Marginal precaution benefit is additional savings in premium costs from an additional increase in
precaution


As illustrated in Figure 23.6, with no moral hazard and actuarially fair insurance
 Results in same level of precaution as in no-insurance case
In general, assuming agents gain some utility from having an insurance company assume
risk (assuming agents are risk averse)
 Then P*, with insurance, is a Pareto-efficient level of precaution
39
Insurance and Moral Hazard



Unfortunately, Pareto-efficient level of precaution is generally not
possible
Hidden level of precaution by consumers makes cost of designing an
insurance policy where premiums are based on every level of
precaution prohibitive
In extreme case of moral hazard, where insurance company cannot at
all determine level of precaution
 Insurance premium would not be a function of consumer’s precaution level

Assuming insurance company sets premium at Pareto-efficient level of
precaution, P*, consumer’s objective is
 Optimal solution is for consumer to not take any precaution, P = 0
• Zero level of precaution increases risk of negative event occurring

Results in insurance company having to pay higher-than-expected claims
 This is root of terminology “moral hazard” for insurance company (principal)
40
Insurance and Moral Hazard

Unless insurance company can design policies that provide
incentives for consumers to take precaution
 Tâtonnement process will result in no insurance company able to
pay all its claims from revenue generated by premiums
• Inefficiency of zero precaution associated with moral hazard is
represented by deadweight loss, area ABC, in Figure 23.6
 Tâtonnement process toward an equilibrium can also result in
instances where agents are overinsured
• For example, due to falling property values or a failing business, an
agent may realize that level of insurance is more than property is worth

If this information is hidden from insurance company (adverse selection)
 Hidden action of not taking any precautions to prevent fire or even
causing business to burn down can increase returns
 For this reason, in fire investigations owners are always possible
suspects
41
Coinsurance


Deadweight loss associated with moral hazard can be
reduced by inducing consumers to take some precaution
One type of inducement, employed by many health
insurance companies, is coinsurance
 Require consumer to pay some percentage of cost, so insurance
company pays less than 100% of loss
• Actual percentage paid varies, but a common rate is for an insurance
•
company to pay 80% and consumers to pay remaining 20%
As percentage of loss a consumer pays increases, less risk is shifted to
insurance company and consumer is more willing to take precaution

Consumers will tend to seek lower-cost treatments rather than alternative
higher-cost treatments
42
Coinsurance

If, for example, consumer pays 20% of cost along with a fixed
premium A, then consumer’s objective is

F.O.C. is
 =0.2EL'(P) = TC'(P)
As illustrated in Figure 23.7, marginal benefit curve tilts
downward and intersects marginal cost curve at second-best
Pareto-efficient equilibrium level of precaution PO > 0
 Deadweight loss is reduced from area ABC to DEC
Only when moral hazard can be eliminated will a Paretoefficient solution P* exist


43
Figure 23.7 Coinsurance
44
Deductibles


Writing insurance policies with deductibles is another option insurance
companies employ for increasing agents’ precaution level
Require agents to incur all loss up to some dollar limit
 For example, if an auto insurance policy has a $500 deductible provision
• First $500 in damages is paid by car owner, and insurance company pays any
remaining damages
Generally, the higher the deductible, the lower will be the insurance premiums

•
Insurance companies will incorporate deductibles into their policies
when agents’ basic level of precaution is so low that insurance
companies cannot earn normal profits
 For example, without some deductible for auto insurance, our roadways
could take on a bumper-car atmosphere
• Resulting in dramatic insurance premium increases with few if any consumers
willing to be insured
45
Deductibles

With deductible provision, optimal level of precaution for a consumer is
determined by
• DA is level of deductible
 Maximum cost a consumer will incur is deductible DA
• However, if overall cost of precaution plus expected losses is less than DA


Consumer can lower his cost below DA
F.O.C. for minimizing cost is
 TC'(P) = -EL'(P)
• If DA > min[TC(P) + EL(P)], yielding optimal level of precaution P* (Figure 23.6)
• If DA < min[TC(P) + EL(P)], a zero level of precaution, P* = 0, is optimal level

With DA as lowest possible level of cost
 Expenditures on precaution will not result in any additional benefits
46
Deductibles

Deductibles allow consumers to insure against large losses
 But be responsible for any relatively small expected losses below deductible
 Reduces deadweight loss associated with moral hazard

As DA increases, deadweight loss is reduced, as consumers will likely
choose no-insurance level of precaution, P*
 Consumers who are more willing to take risk will self-insure by seeking
higher insurance deductibles
• However, with increases in DA, risk-averse consumers are worse off since they
are less able to shift this risk to another agent (insurer)

Other options available to insurance companies for increasing agents’
precaution level are
 Combinations of coinsurance and deductibles
 Subsidizing preventive care
• Health insurance policies will generally


Have both deductibles and coinsurance provisions
May also offer preventive care such as annual physical examinations and routine blood
tests at reduced cost
47
Employer and Employee
Relations

Moral hazard exists whenever asymmetric information in the
form of hidden actions is prevalent in a principal-agent
agreement
 For example, moral hazard can exist between an employer
(principal) and an employee (agent)
• Unless an employer can constantly monitor productivity of employees



Employees can engage in leisure while working (shirking) by reducing their
level of effort
 For example, employees’ surfing the Net has become a major form of
shirking
Asymmetric information on level of employees’ productivity
creates inefficiencies
An objective of employers is to design contracts that provide
employee incentives directed at improving productivity and
reducing shirking
48
Pareto Efficiency with No Moral
Hazard


Major incentive for employees’ effort is
compensation they receive for supplying their labor,
in form of wage income
Assuming symmetric information (no moral hazard)
 Can determine Pareto-efficient level of employee effort,
E*
• By considering employer’s objective function and an employee’s
participation constraint


No moral hazard implies that an employer can
observe an employee’s level of effort
Assume employer determines labor contract and
employee can then either accept or reject contract
49
Pareto Efficiency with No Moral
Hazard

An employer is concerned with productivity of an employee
 Denoted by production function q = f(E)
• q is some output level
 Given a per-unit output price of p and wage rate based on an
employee’s effort w(E)
• Employer’s objective is maximizing profit from this employee
• Employee has a cost of increasing effort in form of total opportunity cost
from lost shirking, TCE(E)


Let MCE(E) represent marginal cost of effort
 So MCE(E) = TCE(E)/E
In general, as illustrated in Figure 23.8, this marginal
opportunity cost is U-shaped
50
Figure 23.8 Employee’s marginal
opportunity cost of effort
51
Pareto Efficiency with No Moral
Hazard

Marginal cost of effort may at first decline
 For very low levels of effort (to left of EM), additional effort results in
marginal cost of effort declining
• Spending so much time shirking, a little additional effort results in lower
marginal opportunity cost
 At relatively higher effort levels (to right of EM), any additional effort
raises this marginal opportunity cost

Employee’s payoff for E level of effort is
 Difference in wage income, w(E)E, and total opportunity cost, TCE(E)
• w(E)E – TCE(E)
 Instead of working for this particular employer, employee could be
engaged in other activities
• Being employed by another employer, being self-employed, or being
immersed in total leisure
52
Pareto Efficiency with No Moral
Hazard

Assume next-highest payoff from these alternatives is U°
 Employee will be willing to work for an employer if payoff is at least as
great as U°

Specifically, if
 w(E)E – TCE(E) ≥ U°
 U° is reservation-utility level, and equation is participation constraint
• Employer must pay at least level U° if he expects to hire employee

Employer’s objective is then
 Subject to w(E)E – TCE(E) = U°

Constraint is an equality because if w(E)E - TCE(E) > U°
 Employer could lower wages and still hire employee
53
Pareto Efficiency with No Moral
Hazard

Substituting constraint into objective function yields
• F.O.C. is



MRPE = MCE
For profit maximization, employer will equate marginal
revenue product of an employee’s effort, MRPE
 To employee’s marginal opportunity cost of effort, MCE
Solving this F.O.C. for E results in Pareto-efficient level of
employee effort, E*
 Illustrated in Figure 23.9, where MRPE is equated to MCE
54
Figure 23.9 Pareto-efficient level of
effort for a risk-averse employee
55
Pareto Efficiency with No Moral
Hazard

Compensation scheme necessary for obtaining employee
effort level E*
 Where level of compensation just equals reservation-utility level plus
employee’s cost of effort
• w(E)E = U° + TCE(E)
 If employer is risk neutral and employee risk averse, employer will
fully insure employee against any wage risk
• Employer will offer a fixed wage rate, w* = w(E*)
• Optimal contract when effort is observable


Specifies Pareto-efficient effort level E*
Fully insures a risk-averse employee against income losses
 When employee is also risk neutral, insurance is not necessary
• Any compensation scheme where wages are a function of profits, with
w(π)E = U° + TCE(π) will be efficient
56
Inefficiency with Moral Hazard

In many cases, cost of monitoring effort prohibits constantly
observing an employee’s level of effort
 For example, an employer is generally unable to observe a night
clerk at a convenience store or a truck driver for a furniture company

When effort is not observable, Pareto-efficient effort level
comes in conflict with result of full insurance
 Only method for increasing employee effort is relating wages to
firm’s profit
• Random nature of profit results in employee assuming some uninsured
risk


Such conflicts create inefficiencies unless employee is risk neutral
A risk-neutral employee is only concerned with expected profit
 Would not be concerned with any random nature of profit
 Indifferent with taking uncertain profit in place of a certain wage
57
Inefficiency with Moral Hazard

In contrast, when an employee is risk-averse
 Increased share of profit relative to a certain
wage does affect employee
 Incentives for increased effort are directly
associated with an employee’s cost of increased
risk
• Results in an additional constraint on employer
Employer not only maximizes profit subject to participation
constraint
 w(E)E – TCE(E) ≥ U°
 But also is subject to an incentive-compatibility constraint
 Must offer a compensation scheme that gives an
employee an incentive to choose required effort level

58
Inefficiency with Moral Hazard

When employer can directly observe employees’
efforts
 Employees will put forth required level of effort regardless
of their desire

In contrast, when effort is not directly observable
 Employees can shirk by not putting forth required level of
effort
• To avoid such shirking, employers must offer a compensation
scheme to induce employee to offer E* units of effort


Determined by setting employee’s payoff associated with E* at least
as great as payoff for any other level of effort
 w(E*)E* – TCE(E*) ≥ w(E)(E) – TCE(E)
 For all levels of effort E
At any wage below this constraint, employee will shirk
59
Inefficiency with Moral Hazard

In the extreme case, not considering this incentive-compatibility constraint
 Results in an employee seeking an effort level independent of his wages
• Employee would minimize his total effort

Optimal solution is for employee to totally shirk and not exert any effort
 Analogous to zero level of precaution associated with insurance
 Illustrated in Figure 23.6
• Unless employer can design contracts that provide incentives for employees to choose effort



Tâtonnement process will result in a zero level of effort
Inefficiency of E = 0 associated with moral hazard is represented by deadweight
loss, area ABC in Figure 23.9
Designing employment contracts with compensation mechanisms that take into
consideration this incentive-compatibility constraint will provide incentives for
employees to increase their work efforts
 Will reduce inefficiency associated with wages not directly linked with level of effort
• However, such contracts will be second-best Pareto-efficient allocations


Still result in risk-averse employee not being fully insured
Only with symmetric information associated with no hidden action on part of
agent (employee) will a Pareto-efficient allocation exist
60
Residual Claimant

Large poultry enterprises use residual claimant production contracts with
independent farmers to raise chickens
 By having farmers assume risk of raising chickens
• Farmers will have incentives to take necessary effort to prevent disease and other
possible adverse effects on chickens

USDA Economic Research Service estimates 52% of approximately
50,000 farms with poultry or egg production in 1995 reported use of
production contracts
 Value of poultry and eggs produced under such contracts accounted for 85%
of total value of all poultry and egg production


Farmers without contracts tended to be either large owner-integrated
operations or independents providing poultry and poultry products to
local markets
In poultry contracting, employees are residual claimant to output
 An example of a second-best Pareto-efficient compensation scheme
incorporating incentive compatibility
61
Residual Claimant

A residual claimant is an agent (farmer) who
receives payoff from output after any fees are paid
to a principal (poultry enterprise)
 An employee will maximize payoff by
• Equating marginal revenue product of employee’s effort to
employee’s marginal opportunity cost of effort

Examples of residual claimant contracts are franchises and
employee buyouts
 Fast-food enterprises are a typical example of franchising
 Owner of a fast-food establishment pays parent company a
fixed fee for right to operate (franchise)
 Employees become residual claimant
 Compensation is now dependent on profits of firm minus lump
sum payment to owners
62
Residual Claimant

Specifically, employees’ objective is to maximize consumer
surplus plus economic rent minus franchise fee

F.O.C. is
 MRPE = MCE
Although marginal benefit equals marginal cost, risk-averse
employees are not able to fully insure against losses
 Results in a second-best Pareto-efficient allocation

• Such residual claimant contracts are very popular when employees are
able to take precautions and reduce possible losses in profits at a lower
cost than owners
63