Maximum Willingness-to-pay (Risk Premium)

1. Downside risk only

Let W = wealth, L = loss if the event occurs, = probability that the event occurs, and
q = maximum willingness-to-pay to avoid the risk = risk premium, then

U (W L) + (1 )U (W ) = U (W q):

The individual, who has wealth W and faces a downside risk of losing L; is willing to
pay at most q to remove the risk. For instance, taking a new vaccine for the prevention of
cervical cancer, paying protection money (protection racket, protection fee, pizzo/ma…a)
to gangsters/racketeers (extortion), ransomware attacks (malware and ransom demand).

Downside Risk and Risk Premium

U (W )
6

`
``
U (W )
` `
` ``
`
U (W L)+
` ``
`
` ``
`
(1 )U (W )
` ``
`
``
= U (W q)
` `
` ``
`
` ``
`
` ``
`
` ``
`
` ``
`
` ``
U (W L) ``

W L W q W L W W

99
 2. Downside risk and Upside potential
Let W = wealth, W1 = wealth if the event occurs, W2 = wealth if the event does not occur,
q = maximum willingness-to-pay to avoid the risk = risk premium, and W1 < W < W2 ;
then

U (W1 ) + (1 )U (W2 ) = U (W q):

The individual, who has wealth W and faces a downside risk of losing W W1 as
well as an upside potential of gaining W2 W; is willing to pay at most q to remove the
risk. For example, the gamble in equation (43), where W W1 = W2 W = h. A plainti¤
(or defendant) in a civil lawsuit contemplates whether to settle it out of the court.

Risk Premium: Special Case

W = W1 + (1 )W2

U (W )
6

U (W2 )
` ``
`
` ``
`
` ``
`
` ``
`
``
U (W1 ) + (1 )U (W2 )
` `
``
= U ( W1 + (1 )W2 q)
` `
` ``
`
` ``
`
` ``
`
` ``
`
` ``
`
` ``
`
U (W1 ) ``

-
W1 + (1 )W2
W1 W1 + (1 )W2 q W2 W

100
 Risk Premium: General Case
W S W1 + (1 )W2

U (W )
6

`
``
U (W2 )
`
` ``
`
` ``
`
``
U (W )
` `
``
U (W1 ) + (1 )U (W2 )
`
= U (W q)
` ``
`
` ``
`
` ``
`
` ``
`
` ``
`
` ``
`
` ``
``
U (W1 ) ``

-
W1 + (1 )W2
W1 W q W W2 W

101
The Demand for Insurance
Let = probability of accident, W = income, L = loss if the accident happened, M =
amount insured, and = premium percentage, then insurance premium = M: There are
two parties in the insurance market: insured (i.e., consumer) and insurer (i.e., insurance
company). Assume that the insured is risk averse, i.e., U 00 ( ) < 0:

Insured
The insured’s objective is to choose M to maximize expected utility:

max U (W M L + M ) + (1 )U (W M)
M

To the insured, ; W; ; and L are exogenous (i.e., given). The …rst-order condition is

(1 )U 0 (W M L + M) (1 ) U 0 (W M ) = 0: (45)

Example: If U (W ) = log W; then the optimal M is given by

(1 ) L+( )W
M = (46)
(1 )

Insurer
To the insurer, revenue = M and cost = M + (1 )0 = M; hence expected pro…ts
= M M: If insurance is actuarially fair (e.g., the insurance market is perfectly com-
petitive), then the expected pro…ts = 0, which implies = : As a result, the premium
percentage = the probability of payout to the insured.

Back to the insured

If the insurance is actuarially fair, then substituting = into (45), one obtains (1
)U 0 (W M L + M ) (1 ) U 0 (W M ) = 0; i.e.,

U 0 (W M L + M ) = U 0 (W M ): (47)

This equation means that the insured will achieve an optimum by choosing M to equalize
the marginal utility of income across the two states (the accident state and the no-accident
state). From mathematics, f (x1 ) = f (x2 ) implies that x1 = x2 if f (x) is a monotonic
function. Since U 00 < 0; U 0 is thus monotonic, hence (47) implies that W M L+M =
W M; which implies that M = L: The insured is fully insured because M = L: As
= and M = L; the insured’s utility = U (W M L + M ) + (1 )U (W M) =
U (W L): Regardless of whether the accident occurs, the insured receives U (W L)
with certainty. Insurance essentially removes uncertainty!

102
Example: If U (W ) = log W and = ; then (46) shows that M = L:

Question: Without any insurance, what is the consumer’s maximum willingness to pay
to avoid the risk?

Let the individual’s maximum willingness-to-pay be q; then q is determined by

U (W L) + (1 )U (W ) = U (W q): (48)

Example: U (W ) = log W; = 0:25; W = 100; 000; L = 20; 000: Then W M L+M =

W M = 95; 000: The individual is willing to pay as much as q to completely avoid the
risk, where 0:25 log(100; 000 20; 000) + 0:75 log 100; 000 = 11:4571 = log(100; 000 q);
hence q = 5; 426: Notice that q = 5; 426 > M = 5; 000: If an actuarially fair insurance is
available, then U (W L) = log(100; 000 5; 000) = log(95; 000) = 11:4616 > 11:4571 :
The consumer is better o¤ with insurance.

Remarks:

1. If the insurance is not actuarially fair, then 6= : For example, if there is admin-
istrative cost C, then the …rm’s expected pro…ts become M M C: Assume
C
that there is perfect competition, M M C = 0; hence = + M > : In this
case, will the optimal M be equal to L?

2. In the real world, even if = , the insurance company can still make money. How?

3. The inequality q > L must hold, i.e., the maximum willingness-to-pay (q) is greater
than the actuarially fair premium ( L).

103
Insurance: Demand and Supply
There are many di¤erent types of insurance policies.

1. Individual:

2. Corporation:

Are the suppliers of insurance policies necessarily risk lovers?

What are some of the oldest forms of insurance policies in the history of mankind?

Are there any weird insurance policies? What are the weirdest ones?

What is the future of the traditional insurance industry?

Will traditional insurance companies be wiped out by FinTech (blockchain, smart

contract)?

Change or fail, sink or swim, insurance versus insurtech

Paul Schulte (December 8, 2016): "Chinese online insurer leaves traditional rivals
in the dust"

Blue, Hong Kong’s …rst purely digital life insurance company, entered into the
market in September 2018. No banking partners and no brokers, therefore no
commission expenses.

104
Problems with Asymmetric Information
Sometimes uncertainty arises because of asymmetric information.

When two parties have asymmetric information, two problems may arise: moral
hazard and adverse selection.

Moral hazard: One party cannot completely observe the other party’s behavior
(cannot exactly tell what the other party has done)

Adverse selection: One party cannot completely identify the other party (cannot
exactly tell who’s who)

1. Moral Hazard
Moral hazard is a pervasive problem in the insurance industry, but it also occurs in other
industries. There are many de…nitions in the literature. Here are some examples.

(i) Snyder and Nicholson’s Microeconomic Theory: "The e¤ect of insurance coverage
on an individual’s precaution, which may change the likelihood or size of losses."

(ii) Pindyck and Rubinfeld’s Microeconomics: "Moral hazard occurs when a party whose
actions are unobserved a¤ects the probability or magnitude of a payment."

"The term “moral hazard”was originally used by insurers to describe the behavior of
individuals who, when they became insured, took less care or incurred greater losses
than if they had remained uninsured. This change in behavior resulted in insurers
incurring greater payo¤s than were expected, expectations being determined by
behavior when uninsured. The hazard was, therefore, to the insurer who, when
setting the premium, based it on the presumption that the insured consumers had
a moral obligation to not change their behavior." John Nyman (2003): The Theory
of Demand for Health Insurance.

Numerical Example: Insurance against car theft

Let W0 = 100; 000; = 0:25; L = 20; 000; U (W ) = ln(W ): Suppose that, if an antitheft
device, which costs $1,950, is installed, then will be reduced from 0.25 to 0.15.

Examples of antitheft devices

Expected utility without installing the device

= 0:25 ln(100; 000 20; 000) + 0:75 ln(100; 000) = 11:4571 :

Expected utility from installing the device

= 0:15 ln(100; 000 20; 000 1950) + 0:85 ln(100; 000 1950) = 11:4590 :

Thus, the device should be installed (as 11:4590 > 11:4571 ).

105
Suppose insurance is available: full insurance for $5,200 ($5,000 for the expected loss
0:25 20; 000; plus $200 for administrative costs), then

Expected utility with the insurance policy

= 0:25 ln(100; 000 5200 20; 000+20; 000)+0:75 ln(100; 000 5200) = ln(94; 800) =
11:4595 :

Expected utility with the insurance policy and the device (assuming that the
insurance company does not know whether the device is installed)
= 0:15 ln(100; 000 5200 1950) + 0:85 ln(100; 000 5200 1950) = ln(92; 850) =
11:4387 :

Expected utility with the insurance policy and the device (assuming that the
insurance company knows for free whether the device is installed, in which case
insurance premium = $0:15 20; 000 + 200 = $3200)
= 0:15 ln(100; 000 3200 1950) + 0:85 ln(100; 000 3200 1950) = ln(94; 850) =
11:4601 :

Thus, the antitheft device will not be installed if the company does not know that it
is installed (as 11:4595 > 11:4387 ). The car owner will just buy the full insurance.
Without installing the antitheft device, the probability of theft is higher (0.25).

This is a moral hazard problem. The availability of insurance increases the likelihood
of car theft. However, if the company knows that it is installed and adjusts the premium
accordingly, then the insured will have the incentive to install the device (as 11:4601 >
11:4595 ):

Case Options Expected Utility

1 No Device, No Insurance 11:4571
2 Device 11:4590
3 Insurance 11:4595
4 Insurance, Device, Premium Unadjusted 11:4387
5 Insurance, Device, Premium Adjusted 11:4601

Without insurance, the device will be installed. With insurance, the device will not be
installed if the premium is unadjusted, hence moral hazard.

Monitoring Cost
It may be costly for the company to check whether the device is installed. Suppose it
costs $10 to …nd out whether the antitheft device has been installed, then the insurance
premium = $0:15 20; 000 + 200 + 10 = $3210; then

106
 Expected utility with monitoring, insurance, and the antitheft device
= 0:15 ln(100; 000 3210 1950) + 0:85 ln(100; 000 3210 1950) = ln(94; 840) =
11:4600 :
Thus, the device will be installed when there is monitoring (even when monitoring
is costly) (as 11:4600 > 11:4595 ).

Suppose the monitoring cost rises to $200, will the device be installed?

Expected utility with monitoring, insurance, and the antitheft device

= 0:15 ln(100; 000 3400 1950) + 0:85 ln(100; 000 3400 1950) = ln(94; 650) =
11:4579 :

In this case, the device will not be installed (as 11:4595 > 11:4579 ).

The above example shows that insurance can change the insured’s behavior, result-
ing in a higher probability that the insured event will occur.
Other examples of moral hazard:

– Deposit Protection Scheme (deposit insurance) in Hong Kong (2004): Who is

for and who is against?
– US Subprime Mortgage Crisis in 2007-09
– EU bailouts: Cyprus, Greece, Ireland, Portugal, Spain, ...?
– Coronavirus pandemic (COVID-19): economic relief policies
Unemployment insurance (contributory), unemployment assistance (means-
tested)
– COVID insurance policy by FWD General Insurance Company Ltd (Hong
Kong)
COVID-19 Care Plan Certi…cate (premium: HK$200)
Daily Cash for Compulsory Quarantine under Hong Kong Government’s
instructions (if sickness is diagnosed): $500/day (14 days), Maximum:
$7,000.
– Insurance frauds
Three Hong Kong students arrested for making bogus robbery claim in
Thailand
Husband: Faked death
Oneself: Suicide
Oneself: Arm-chopping

What is bad about moral hazard?

How to solve the moral hazard problem?

107
Example
The Deadweight Loss of Moral hazard
P 6

B C
M

-
O Q Q0 Q

Let P = price of a doctor’s visit, Q = number of doctor’s visits. Suppose that the
marginal cost of a doctor’s visit is a constant M: Without insurance, P = M and
Q = Q . With insurance, P = 0 and Q = Q0 :

Moral hazard: Q0 > Q

Assume that the demand for doctor’s visits is given by Q = a bP: The deadweight
loss (DW L) is obtained by comparing the following two scenarios:

I. Without insurance: each doctor’s visit costs M

When price is equal to marginal cost (P = M ); a bQ = M; hence Q = a bM:
Consumer Surplus = AM B: Producer Surplus = 0: Total Surplus = Consumer Surplus
bM )2
+ Producer Surplus = AM B = ab M Q2 = (a 2b :

II. With insurance: doctor’s visits are free

Assume that the insurance premium is F: Since doctor’s visits are free, the quantity
consumed at P = 0 is Q0 = a:
Consumer Surplus = OAQ0 F: Producer Surplus = F OM CQ0 : Total Surplus =
bM )2 bM 2
(OAQ0 F ) + (F OM CQ0 ) = AM B BCQ0 = (a 2b 2
:
bM 2
Comparing the total surpluses of I and II, DW L = BCQ0 = 2
:

Robert Pindyck and Daniel Rubinfeld (Microeconomics 8th ed. p.321): Deadweight
loss = Net loss of total (consumer plus producer) surplus.

108
 Moral hazard may be costly to society.

Digression: Should bu¤et be banned?

Is the Welfare Cost of Moral Hazard Excessive?

“The welfare case for insurance policies of all sorts is overwhelming. It follows that
the government should undertake insurance in those cases where this market, for
whatever reason, has failed to emerge. Nevertheless, there are a number of signif-
icant practical limitations on the use of insurance. It is important to understand
them, though I do not believe that they alter the case for the creation of a much
wider class of insurance policies than now exists.”
[Kenneth Arrow (1963): “Uncertainty and the Welfare Economics of Medical Care,”
American Economic Review, p.961]

“Even if all individuals are risk-averters, insurance against some types of uncer-
tain events may be nonoptimal. Hence, the fact that certain kinds of insurance
have failed to emerge in the private market may be no indication of nonoptimality,
and compulsory government insurance against some uncertain events may lead to
ine¢ ciency.”
[Mark Pauly (1968): “The Economics of Moral Hazard,” American Economic Re-
view, p.537]

“The lesson of Mr. Pauly’s paper is that the price system is intrinsically limited
in scope by our inability to make factual distinctions needed for optimal pricing
under uncertainty. Nonmarket controls, whether internalized as moral principles or
externally imposed, are to some extent essential for e¢ ciency.”
[Kenneth Arrow (1968): “The Economics of Moral Hazard: Further Comment,”
American Economic Review, p.538]

“American families are in general overinsured against health expenses. If insurance

coverage were reduced, the utility loss from increased risk would be more than
outweighted by the gain due to lower prices and the reduced purchase of excess
care. ... by raising the average coinsurance rate from 0.33 to 0.50 and 0.67 percent.
The most likely values imply net gains in excess of $4 billion.”
[Martin Feldstein (1973): “The Welfare Loss of Excess Health Insurance,”Journal
of Political Economy, p.251]

“Pauly’s analysis relies on the assumption that consumers’purchases of health care

are not responsive to income. It is clear, however, from empirical studies that the
individual’s demand for medical care generally increases with income. ... Pauly’s
welfare loss is too large because it counts moral hazard due to both price and income
e¤ects, whereas the welfare loss should only be related to moral hazard from price
e¤ects. ...the true welfare loss is about 33 percent of Pauly’s welfare loss.”
[John Nyman (1999): “The Economics of Moral Hazard Revisited”, Journal of
Health Economics, p.812 and p.818]

109
Some market solutions to mitigate moral hazard in the insurance industry:

– deductible
– copayment
– coinsurance
– cap on claims

110
2. Adverse Selection
Asymmetric information: buyers and sellers have di¤erent information about the
quality of the product.

The following are some de…nitions of adverse selection:

(i) Adverse selection arises when products of di¤erent qualities are sold at a single
price because buyers or sellers are not su¢ ciently informed to determine the true
quality at the time of purchase. As a result, too much of the low-quality product
and too little of the high-quality product are sold in the market.

(ii) Adverse selection is said to occur when insurance companies must charge a single
premium because they cannot distinguish between high-risk and low-risk individu-
als; more high-risk individuals will insure, making it unpro…table to sell insurance.

Akerlof’s Lemon Model

Akerlof’s (1970) model is not related to insurance at all.

Lemons: a colloquialism for defective cars

Akerlof’s idea may be illustrated by a simple example. Assume that a good is sold
in indivisible units and is available in two qualities, low and high, in …xed shares
and 1 .

Each buyer is potentially interested in purchasing one unit, but cannot observe the
di¤erence between the two qualities at the time of the purchase. All buyers have
the same valuation of the two qualities: one unit of low quality is worth wL dollars
to the buyer, while one high-quality unit is worth wH > wL dollars.

Each seller knows the quality of the units he sells, and values low-quality units at
v L < wL dollars and high-quality units at v H < wH .

High-quality Unit Low-quality Unit

Buyers’Valuation wH wL
Sellers’Valuation vH vL

If there were separate markets for low and high quality, every price between v L
and wL would induce bene…cial transactions for both parties in the market for low
quality, as would every price between v H and wH in the market for high quality.
This would amount to a socially e¢ cient outcome: all gains from trade would be
realized.

But if the markets are not regulated and buyers cannot observe product quality,
unscrupulous sellers of low-quality products will choose to trade on the market for
high quality. In practice, the markets would merge into a single market with one
and the same price for all units. In other words, buyers are only willing to pay
the average valuation w = wL + (1 )wH ; therefore the market price could not

111
exceed w (assuming that buyers are risk averse or risk neutral). If v H > w; then
sellers with high-quality goods would exit from the market, leaving only an adverse
selection of low-quality goods, the lemons. The market has adversely selected only
the low-quality goods.

However, if v H < w; then there will be no adverse selection. Both high-quality and
low-quality goods will be sold in the market.

The lemon problem: With asymmetric information, low-quality goods can drive
high-quality goods out of the market.

Di¤erence between averse (risk averse) and adverse (adverse selection)

Adverse selection in the labor market: People who believe that they are of high value
may take themselves out of the labor market because they consider the market wage
rate too low.

Adverse selection and dating

– Is there adverse selection in the search for dates/partners?

– Would online dating apps solve (or reduce) the adverse selection problem?

112
Adverse Selection in Car Insurance
Consider the car theft example. Suppose there are two types of car owners: Type A has
already installed the antitheft device but Type B does not have the antitheft device. The
insurance company cannot distinguish between the two types of owners. Assume there
are no administrative costs of underwriting insurance. Let W0 = 100; 000; L = 20; 000;
U (W ) = ln(W ): The antitheft device reduces the probability of theft from 0.25 to 0.15.

If the insurance company o¤ers fair insurance and charges every car owner $5,000
(0.25 20,000) for a full coverage ($20,000), then the expected utility of an insured
car owner = ln(100; 000 5; 000) = ln(95; 000) = 11:4616 : This is because
(a) for Type A owners: expected utility = 0:15 ln(100; 000 5; 000 20; 000 +
20; 000) + 0:85 ln(100; 000 5; 000) = ln(100; 000 5; 000) = ln(95; 000) = 11:4616
(b) for Type B owners: expected utility = 0:25 ln(100; 000 5; 000 20; 000 +
20; 000) + 0:75 ln(100; 000 5; 000) = ln(100; 000 5; 000) = ln(95; 000) = 11:4616

– Type A owners will not purchase this insurance policy because the expected
utility without insurance = 0:15 ln(100; 000 20; 000) + 0:85 ln(100; 000) =
11:4795 > 11:4616
– Type B owners will purchase this insurance policy because the expected utility
without insurance = 0:25 ln(100; 000 20; 000)+0:75 ln(100; 000) = 11:4571 <
11:4616
– There is adverse selection because only the high-risk owners (i.e., Type B) will
purchase the insurance policy.

If the insurance company charges every car owner $3,000 (0.15 20,000) for a full
coverage ($20,000), then the expected utility of an insured owner = ln(100; 000
3; 000) = ln(97; 000) = 11:4825 (which is higher than 11:4795 and 11:4571 ).
Clearly, both Type A and Type B owners will purchase this insurance policy.

Suppose the numbers of Type A and Type B owners are the same, therefore the
proportion of Type B owners is 0.5. If the company o¤ers the $3,000 policy, then
it will lose on average 0:5(3; 000 0:25 20; 000) = 1; 000 for underwriting such an
insurance policy. Therefore, the insurance company will not o¤er this policy.

Suppose the insurance company sets the premium at the average level, i.e., $4; 000
(0:5 5; 000 + 0:5 3; 000 = 4; 000) for a full coverage ($20,000), then Type B
owners will buy the policy because the expected utility = ln(100; 000 4; 000) =
ln(96; 000) = 11:4721 > 11:4571 : However, Type A owners will not buy the
policy because their expected utility without insurance is 11:4795 > 11:4721 :
Therefore, the pooling equilibrium is not viable because the company will lose on
average (4; 000 0:25 20; 000) = 1; 000. There is adverse selection because only
the high-risk owners (i.e., Type B) are insured.

113
The insurance company can separate Type A owners from Type B owners by o¤ering
the following policy: charge $5,000 for a full coverage ($20,000) and 0:15M for a
partial coverge of M dollars (the magnitude of M is to be determined). Under this
policy, a Type B car owner will buy the full insurance but not the partial insurance
if

0:25 ln(100; 000 20; 000 + M 0:15M ) + 0:75 ln(100; 000 0:15M ) < ln(95; 000):

Solving this inequality numerically, M < 3; 000: A Type A owner will buy the
partial insurance M = 3; 000 because

0:15 ln(100; 000 20; 000 + M 0:15M ) + 0:85 ln(100; 000 0:15M )
= 0:15 ln(100; 000 20; 000 + 3; 000 0:15 3; 000)
+0:85 ln(100; 000 0:15 3; 000)
= 11:4803 ;

which is greater than 0:15 ln(100; 000 20; 000) + 0:85 ln(100; 000) = 11:4795 .
A Type B owner will not buy the partial insurance because

0:25 ln(100; 000 20; 000 + 3; 000 0:15 3; 000)

+0:75 ln(100; 000 0:15 3; 000)
= 11:46160134 < ln(95; 000) = 11:46163217

This is called a separating equilibrium. Of course, this coverage will be regarded as

too small by Type A owners as it o¤ers only $3,000 out of a loss of $20,000.

Question: What is the pro…t of the insurance company?

Assume that full insurance is available. If Type A owners could buy a certi…cate
to prove that they have installed the antitheft device, they will be willing to pay at
most YA such that

ln(100; 000 3; 000 YA ) = 11:4795 ;

thus YA = 287: A Type B owner will not buy the certi…cate if

ln(100; 000 3; 000 YB ) < 11:4616 ;

thus YB > 2; 003: Therefore, if it costs a Type A owner $287 to prove that he
has installed the device and a Type B owner $2,004 to fake that he has installed
the device (e.g., forgery is costly), then the Type A owner will pay $287 to do so
but the Type B owner will not pay to do so. This is a signaling equilibrium. It
requires di¤erential costs in obtaining the signal, otherwise the equilibrium will not
be viable. (Application: An MTR Director’s dodgy degree)

114
Reference: 2001 Nobel Prize: Markets with Asymmetric Information
(http://nobelprize.org/nobel_prizes/economics/laureates/2001/press.html)

Pooling equilibrium: An equilibrium in which di¤erent people are treated (paid)

alike or behave alike.

Separating equilibrium: An equilibrium in which one type of people is di¤erentiated

from other types of people.

Signaling equilibrium: An equilibrium in which one type of people takes an action

to send information to a less informed party in order to di¤erentiate themselves
from other types of people.

Screening equilibrium: An equilibrium in which a less informed party takes an

action to di¤erentiate one type of people from other types of people.

Party Action Party

signal
!
informed less informed
screen

Question: adverse selection ! market failure ! government intervention?

115
How to mitigate adverse selection problems?
1. Restrict opportunistic behavior
(a) Universal Coverage
Government provides insurance to everyone or mandates everyone to buy insurance
(e.g., third-party auto insurance)

Firm provides health insurance to all employees rather than paying them a higher
wage and letting them decide whether to buy health insurance on their own

(b) Laws to Prevent Opportunism

Product liability laws

– Hong Kong does not have speci…c liability laws. Main legislations on product
liability are: Sales of Goods Ordinance, Consumer Goods Safety Ordinance,
Control of Exemption Clauses Ordinance, Toys and Children’s Products Safety
Ordiance, Pharmacy and Poisons Ordinance, Dangerous Goods Ordinance,
Trade Descriptions Ordinance
fried …sh grouper (garoupa) …llets in creamy corn sauce ( )
versus fried …sh …llets in creamy corn sauce ( )
Some restaurants had been …ned several thousand Hong Kong dollars for
breaking the Trade Descriptions Ordinance because the …sh …llets in their
"fried …sh grouper …llets in creamy corn sauce" were not grouper …llets.
They used cheaper …sh …llets.
– Singapore has a lemon law
The Lemon Law (e¤ective September 1, 2012) is a law that protects con-
sumers against goods that do not conform to contract or are not of sat-
isfactory quality or performance standards at the time of delivery. Under
the Lemon Law, businesses are obligated to repair, replace, reduce the
price or provide a refund for a defective good.
– US has lemon laws (federal and states)
federal lemon law: Magnuson-Moss Warranty Act (passed in 1975)
To ensure that manufacturers honor their warranties and to reduce
the chance that a consumer will be misled about the nature and scope
of a warranty when making a purchase.
Example: Will adding aftermarket parts void the warranty? Some car
dealerships told customers that using non-OEM parts would void the
factory warranty. The Magnuson-Moss Warranty Act ensures that car
owners have the freedom to choose aftermarket parts and upgrades
without having the warranty stripped or voided.
state lemon laws

116
 An auto manufacturer is obligated to repurchase a vehicle that has
a signi…cant defect that the manufacturer is unable to repair within
a reasonable amount of time. The exact criteria of "unable to repair
within a reasonable amount of time" vary by state.

2. Equalize Information
(a) Screening Action taken by a less informed person (or party) to determine the
information possessed by informed people

Life insurance companies: check health history, lifestyle, and habits of clients

Consumer screening: consumers buy information from objective experts (appraisal),

learn of a company’s reputation

Genetic testing and insurance (The Human Genome Project): Should insurers have
access to genetic test results?

(b) Signaling Action taken by an informed person to send information to a less in-
formed person (or party)

MBA degree?

Establish brand name (What is the world’s most valuable brand?)

Guarantees and warranties

College guarantee: Some US vocational schools guarantee that their students are
…t for the purposes for which the education was intended. If not, send them back
for retraining.

Questions:

Are low-quality or high-quality producers more likely to o¤er warranties for their
products?

Are warranties a sign of weakness or an indicator of con…dence? (The Hong Kong

Monetary Authority rejects the suggestion of legislating the linked exchange rate
HK$7.8 = US$1)

(c) Third-party Comparisons

Consumer groups (for-pro…t or not-for-pro…t …rms), Consumer Council (Consumer

Digest)

117
(d) Standards and Certi…cation

ISO 9000 (International Organization for Standardization) for quality management

standards

Certify doctors, dentists, electricians, real estate agents, car mechanics, beauticians,
plumbers, economists, …nancial analyst, accountant, actuary, ...

Concerns: drive up prices, anticompetitive, barriers to entry

118