The Demand and Supply of Health Insurance

1. The Economics of Risk and Insurance Why do we buy insurance? On average we pay more in premiums than we receive in benefits Why do people buy lottery tickets? For every $1 you bet, you receive back $.50 The fact that the outcomes of some events are uncertain, cause us to change our standard utility theory. First we will talk about the problem in general and then will apply it to the health care market. Example 1: Get $900 for sure OR 90% chance to get $1,001 Which do you prefer? Example 2: Lose $900 for sure OR 90% chance to lose $1,001 Which do you prefer?

Expected Utility Rational decision makers will choose the course of action that has the highest expected utility

Think about how you feel about your chances of winning $1million (or your chances of contracting a deadly disease) as your probability improves by 5 percentage points: a. from 0-5% b. from 5-10% c. from 60-65% d. from 95-100% Are these the same? Going from 0-5 is probably different than going from 60-65 possibility effect. Likewise going from 95-100 probably is also different from 60-65 certainty effect. Most people tend to have the following preferences

High Probability Event

Low Probability Event

Gains 95% chance to win $10,000 Fear of Disappointment RISK AVERSE Accept unfavorable settlement 5% chance to win $10,000 Hope of large gain RISK SEEKING Reject favorable settlement

Losses 95% chance to lose $10,000 Hope to avoid loss RISK SEEKING Reject favorable settlement 5% chance to lose $10,000 Fear of a large loss RISK AVERSE Accept unfavorable settlement

Note the bottom left cell would be the possibility effect this would explain lottery ticket purchases

The top right is getting at the idea that when people are faced with very bad options tend to take desperate gambles. The bottom right is where the market for insurance comes from.

Expected Utility Rational decision makers will choose the course of action that has the highest expected utility The vonNeumann - Morgenstern utility function shows the decision makers preferences w.r.t. risk. Indicates how we convert wealth into utility, accounting for different states of the world. Suppose you have the choice between two games: 1) 10,000 for sure 2) A 50% chance at 40,000, and a 50% chance of losing 5,000 When deciding which of the above gambles to take (1 or 2) we compare: EU(2) =.5U(40,000) + .5U(-5,000) EU(1) = U(10,000). This approach allows an unequal weighting of outcomes: a loss of 5,000 may be more undesirable than a gain of 40,000 is desirable. Suppose U(x) = x if x> 0 and -|x| if x < 0 where x is wealth, and wealth from all other sources =0. So now: EU(1) = 10,000 = 100 EU(2) = .540,000 -.55,000 = 100-35.36 = 64.64 Thus this person would prefer the sure thing to the risk even though the expected value of the gamble is greater than the expected value of the sure thing. Now suppose Petes utility function is U(x) = x2 if x>0 and -x2 if x<0 EU(1) = 10,0002 = 100million EU(2) = .5(40,0002) - .5(5,0002) = 800m - 12.5m = 787.5m So Pete would flip the coin

What sets these two individuals apart? Risk lovers vs. Risk averters>

A Risk Averse Utility Function:

Utility

For a Risk Averse person utility increases in wealth but at a decreasing rate : diminishing MU of income. Suppose we are choosing between 2 jobs: 1) 40,000 for sure 2) .5 *20,000 or .5*60,000 The expected value is 40,000 for both, but EU(1)=200 while EU(2)=193. So the risk averse person would choose 1. That is a risk averse person would turn down a fair bet.
20 40 60 Wealth

245 200 193

141

But a risk lover will act very differently:

Utility

For a Risk Loving person utility increases in wealth at an increasing rate : Increasing MU of income Now this person if offered the same opportunity: EU(1) = 1600 EU(2) = 2000 So a risk lover would go for the risky job. That is a risk lover would accept a fair bet
20 40 60 Wealth

3600

2000 1600 400

Risk Neutral:

Utility

For a risk neutral person utility increases in income at a constant rate. So this person would be completely indifferent to a fair bet.

60

40

20

20

40

60

Wealth

2. Why People Buy Insurance or the Demand for Insurance

We can use this theory to explain the market for health insurance Suppose you think there is a .5 probability you will need surgery with a cost of $20,000 and suppose your wealth is 40,000. Suppose your utility function is U(W) = W So that U(20,000) = 141 and U(40,000) = 200 and the expected utility of no insurance is 170.5

Utility

Now suppose you are offered an insurance policy that will cover all of your expenses in the event of illness and this will cost $10,000. Will you buy this? If you do, you get 30,000 for sure and U(30) = 173.2. So yes, you would buy the insurance. You will be better off with the insurance than without (but note your expected income is 30k in both cases).

200 173.2 170.5

141

How much would you be willing to pay for insurance? Or at what income for sure would you be just indifferent to having no insurance? EU(no insurance) = 170.5 So find the W that solves: W = 170.5 or w = 170.52=29,070 So if the insurance policy was 40,000-29,070=10,930, you would be just indifferent to buying the policy. This is known as a Certainty Equivalent The more concave the utility function the larger this amount - or the more the person would be willing to pay to avoid risk. Note that Pete Rose wouldnt buy insurance:

20

29.070 30

40

Wealth

Utility

1600

1000 900 200 20 30 40 Wealth

His utility from no insurance is 1000, but if he gets 30k for sure his utility is only 900 so he wouldnt want it. He would only buy insurance if the expected utility was higher with it than without it. So without it he will get an expected utility of 1000 find the wealth for sure that gives 1000 or X2= 1000 = 31623 so he would only pay 8377 for insurance. But note that the insurance company would not be willing to sell insurance for this price!!

So the more risk averse a person is - or the more concave their utility function is - the more likely they will be to buy insurance Also the lower your assessed probability of getting sick, the less likely you will be to buy insurance Many uninsured are younger with low probability of getting sick, and who may not be very risk averse. So many are uninsured by choice

Note it is the uncertainty that is important here, not necessarily the high cost of the adverse events. There are many high cost events in our lives that we do not have insurance for (my sons needed braces, they will want to go to college, etc.) But the difference is that I know this and can plan for it (I know that in a few years Ill have to pay for college so Id better save now). But with health care, the problem is that many events are not predictable. If I knew that in 5 years I would need heart surgery, I could (in theory) save enough to pay for it by then. But the problem is with uncertainty there is a small chance that I will need it this year. It is this risk that insurance gets us out of.

Conclusions: i. Insurance can be sold only in circumstances where there is diminishing marginal utility of wealth or income risk aversion. ii. Even though people will have less wealth as a result of the purchase of insurance, the increased well-being comes from the elimination of risk. iii. Insurance, by pooling large groups of people with a low rate of incidence allows people to have more access to care than they would otherwise.

Is insurance a mechanism for avoiding individual risk or is it a financing mechanism? An alternative way of looking at the insurance issue is from the standpoint of ability to pay. Note that heart surgery is pretty expensive, and even if I knew it was coming I might not be able to pay for it no matter how well I planned. But since it is a relatively unlikely event, it will only occur to a small fraction of the population. So insurance transfers money from healthy people to the sick and enables them to pay for highly valuable services they would otherwise be unable to afford.

Recall the moral hazard problem The above assumed behavior did not change as a result of insurance. But we know that it will. So that the consumption will increase and so the premium of 1,000 would tend to be too little to cover the insurers costs. So in order to cover costs rates increase. Thus an insurance premium has two components. The first is the premium for protection against risk, assuming that no moral hazard exists. The second is the extra resource cost due to moral hazard. Experience rating vs community rating Experience rating alleviates moral hazard and focuses on individual risk car insurance vs. health insurance. Car insurance does not pay for routine maintenance oil change, new tires, etc. Nor does it pay if you blow your engine. It really only covers the risk of an accident. We dont have a problem if the insurance company totals your car. Yet health insurance tends to pay for routine things we dont get totaled what is going on? Tax treatment tends to encourage generous insurance Easy to distinguish accidents from routine things for autos, more difficult for health Fairness/ethical issues

But note that experience rating reduces the income transfer effect from the healthy to the sick that allows more access to care. Community rating also requires less administrative overhead we just need to get it right on average as opposed to each and every time. Economies of scale in insurance pools.