PUBLIC ECONOMICS AND FINANCE

INTRODUCTION TO SOCIAL INSURANCE

1
ROLE OF GOVERNMENT IN INSURANCE

• New topic: Insurance

• Two kinds of insurance: social vs. self

• Self insurance: various private means of insurance against adverse events
• Social insurance: government intervention in the provision of insurance

• In the US, social insurance programs include:

• Unemployment insurance
• Disability insurance
• Workers’ compensation
• Social Security
• Medicaid/Medicare 2
INSURANCE MODEL

• Two “states of the world”: good and bad

• Payoffs: different payoffs (e.g., income) in different states: πGood and πBad
• πGood > πBad

• Probabilities: don't know which state of the world you will be in, but you do
know the probabilities
• Probability of bad state: p
• Probability of good state: 1-p
3
ROLE OF GOVERNMENT IN INSURANCE

• What is my expected payoff?

• Answer: (1-p)*πGood + p*πBad
• i.e., (prob good state)x(payout in good state)+(prob bad state)x(payout in bad state)

• Example: I earn $30,000 per year. There is a 1% chance I get hit by a car and
need to pay a medical bill of $30,000 and 99% chance I’m healthy.

• Expected payoff = .99*30,000 + .01*0 = $29,700

4
STRUCTURE OF INSURANCE

• Now suppose you have the opportunity to insure your payoffs

• i.e., the insurance company pays you a certain amount if you end up in the bad state of
the world

• In return, you pay an insurance premium

• Note: you pay the premium regardless of which state of the world you end up in

• How much should the premium be?

5
ACTUARIALLY FAIR INSURANCE

• An actuarially fair insurance premium is equal to the expected payout by

the insurance company:
• Premium = (probability of bad state)x(payout in bad state)
• Example: premium = .01*30,000 = $300 if insurance company pays full medical bill

• What is the expected payoff for a consumer who buys actuarially fair insurance?
• Example: expected payoff = .99*(30,000-300) + .01*(0+30,000-300)=29,700

• Result: same expected payoff if buy actuarially fair insurance or not!

6
ACTUARIALLY FAIR INSURANCE

• So why would anyone ever buy actuarially fair insurance?

• Both options (actuarially fair insurance vs. no insurance) give the same payoff in
expectation: $29,700. BUT...
• Without insurance: receive $29,700 on average
• With insurance: receive $29,700 with certainty

• Which would you prefer? Why?

7
EXPECTED PAYOFF VS. EXPECTED UTILITY

• Important difference between expected payoff and expected utility

• Expected payoff= p*πBad +(1-p)* πGood
• Expected utility= p*U(πBad)+(1-p)*U(πGood)

• If expected payoff=expected utility, then I get the same utility from my first dollar
of income as my 1000th dollar of income etc.

• Why might expected payoff ≠ expected utility?

• Answer: Diminishing marginal utility
• Diminishing marginal utility  risk aversion
• Constant marginal utility  risk neutral 8
EXPECTED PAYOFF VS. EXPECTED UTILITY

• Example: Suppose U(x)=√x

State of the Payoff Expected Expected

World Payoff Utility
No Insurance Accident (1%) 0 29,700 171.5
No Accident (99%) 30,000
Actuarially Accident (1%) 29,700 29,700 172.34
Fair Insurance
No Accident (99%) 29,700

9
INSURANCE AND RISK AVERSION

• Risk averse consumers: always fully insure when offered actuarially fair
insurance (i.e., a risk averse person would like to consumption smooth)

• Risk neutral consumers: indifferent between no insurance and actuarially fair

insurance

• Will consumers ever pay more for insurance than the actuarially fair premium?
• Risk neutral: no
• Risk averse: yes (willing to pay an additional risk premium)

10
RISK POOLING

• Typically, we think of insurance companies as being risk neutral. Why?

• Suppose there is a 1 in 10 chance of any individual getting sick. If individual buys

insurance, this risk is transferred to the insurance company
• Only one customer: a lot of uncertainty about insurance company's costs
• If 10 customers: more certainty about costs
• If 1000 customers: even more certainty about costs

• Insurance companies risk pool across individuals, lowering their total risk

11
ASYMMETRIC INFORMATION
• We established the benefits of insurance
• But why get the government involved? Why can't individuals insure themselves?

• Suppose there are two types of customers: careful and careless

• Each has different probability of bad state  different actuarially fair premiums

• Ideally, insurance companies could charge each customer a different premium

• But if there is asymmetric information (e.g., customers know their “type”, but
insurers don't), then the problem gets more complicated...

12
ASYMMETRIC INFORMATION: WHAT CAN
INSURANCE COMPANIES DO?

• Example: Insurance company has 200 customers with different likelihood of

getting into an accident
1
• 100 careful customers (p= %)
2

• 100 careless customers (p=5%)

• Insurer can’t observe who is careful and who is careless

• What can insurance companies do?

13
ASYMMETRIC INFORMATION: WHAT CAN
INSURANCE COMPANIES DO?

• Solution 1: Charge two different actuarially fair premiums and ask customers if
they are careful or careless
• Careful premium = .005*30,000=$150
• Careless premium = .05*30,000=$1,500

• Result: careless people will have an incentive to lie about being careless
• All customers say they are careful and pay the low premium  insurance company
loses money and goes out of business

14
ASYMMETRIC INFORMATION: WHAT CAN
INSURANCE COMPANIES DO?

• Solution 2: Set premium = average expected payout across all customers

100∗.005∗30,000+100∗.05∗30,000
• Premium = = $825
200

• Will careless people want to buy this insurance?

• YES. Expected costs = $1,500 > premium

• Will careful people want to buy this insurance?

• NO (unless really risk averse). Expected costs = $150 < premium

• Result: careful people won't want to pay, insurance company loses money since only
careless join and insurance company charges them less than actuarially fair premiums
• Adverse Selection
15
ASYMMETRIC INFORMATION: WHAT CAN
INSURANCE COMPANIES DO?

• Solution 3: Set premium = expected payout of careless customers only

• Premium = .05*30,000 = $1,500

• Will careless people want to buy this insurance?

• YES. Expected costs = premium

• Will careful people want to buy this insurance?

• NO (unless really, REALLY risk averse). Expected costs = $150 < $1,500

• Result: Careful won't buy insurance, insurance company doesn't lose money
• However, only careless get insured
16
INSURANCE AND MARKET FAILURES

• Adverse selection causes insurance companies to not insure all customers

• “Death spiral”

• But any customer who is even a little risk averse will want to fully insure and
insurance companies would want to provide it for them
• Under-provision  Market failure!

• Will there always be adverse selection?

• NO: if careful customers are very risk averse, they might be willing to pay the
actuarially fair premium of the careless
17
ASYMMETRIC INFORMATION: WHAT CAN
INSURANCE COMPANIES DO?

• Solution 4: Create two different insurance products with two different prices
• High coverage: pays more for an accident, but higher premium
• Low coverage: only pays some of the expenses, but charge lower premium
• e.g., the healthcare market has a low-cost, minimal coverage catastrophic insurance
plan and a high-cost traditional insurance plan

• Suppose the high coverage plan is set to the actuarially fair price of the careless
• Will careless people want to buy this insurance? Yes
• Will careful people want to buy this insurance? Maybe

18
POOLING VS. SEPARATING EQUILIBRIUM

• If both buy the high coverage…

• Called a pooling equilibrium

• If careless buy the high coverage and careful buy the low coverage…
• Called a separating equilibrium

• How do we know which type of equilibrium we end up in?

• Answer: depends on customers' utility function (i.e., level of risk aversion)
• If careful customers are really risk averse, they'll pay the high premium for the high
coverage
• Numerical example in this week’s recitation
19
INSURANCE AND MARKET FAILURES

• We showed before that any risk averse customer will want to fully insure

• Efficiency = everyone fully insures (i.e., no market failure)

• So when do we have a market failure in the private insurance market?

• Adverse selection?
• Pooling equilibrium with full insurance?
• Separating equilibrium?

20
INSURANCE AND GOVERNMENT INTERVENTION

• Reasons for government intervention

• Adverse selection
• Experience rating
• Redistribution
• Externalities
• Job lock
• Individuals underweight probability of bad events (paternalism)

21
NEXT CLASS

• Examples of social insurance programs

• Health insurance (next class)
• Income insurance: Social Security & Unemployment insurance (following class)

22