Some Thoughts on the Health Care Debate

It’s been painful watching the ridiculous arguments and accusations in the current U.S. debate over health care. This issue in particular seems to have brought out the worst of democracy, with a rationally irrational public shouting about death panels and calling names. Health Economics is not one of my fields, but some things can be said without being a specialist in the field.

First of all, many people are comparing health systems around the world and making unjustifiably strong claims, sometimes using questionable data or results. For example, the oft-cited WHO study ranking countries’ health care systems focuses almost entirely on the distribution of health care, rather than the level, and has error bars so large that the results are almost worthless. Rankings of life expectancy are also of questionable use when comparing countries’ health care systems, as life expectancy can vary for reasons unrelated to health care systems. For example, Americans tend to be more obese, get in more care accidents, and experience violent crime more often than people in other countries. These may be serious problems with American society, but they are not problems with the health care system. Another measure that has been used recently is a comparison of preventable deaths for people under 75 years of age. This suffers from a problem similar to life expectancy—Americans may be more prone toward particular kinds of health problems, and treatment of them may be less successful, for reasons unrelated to the health care system. A real study of the subject would need to control for more variables.

Similarly, one hears that around 50 million Americans do not have health insurance. While technically correct, this number is of dubious value. What fraction of those 50 million people qualify for government aid but have not used it? What fraction are wealthy but have opted not to buy health insurance? How many actually went for an extended period without health insurance, and how many went without for a short period only?

There are also questionable statements about health care in other countries. They are sometimes presented as utopian, and at other times as hellish, depending on the position taken by the discussant. Conservatives like to claim that wait times in Canada for medical tests and procedures are outrageously long; others claim that waits lists in Canada are reasonable.

Do other countries do a better job controlling costs? David Goldhill argues that if they do, it’s not by much:

From 2000 to 2005, per capita health-care spending in Canada grew by 33 percent, in France by 37 percent, in the U.K. by 47 percent—all comparable to the 40 percent growth experienced by the U.S. in that period. Cost control by way of bureaucratic price controls has its limits.

Perhaps other systems get the level of costs down, but cannot control growth, meaning they will end up where the U.S. is eventually.

None of this is to say that the U.S. health care system could not use some dramatic changes. To explain what kinds of changes might work—and by “work”, I mean provide the same or better care at a lower cost to more people—I need to talk about some basic economics, and then some particular economics of health care.

Students in any economics class learn that resources are limited, or scarce. We cannot all have as much of anything as we want. This is true of cars, gasoline, televisions, computers, clothing, leisure, and everything else people enjoy consuming. One can have more of something, but only at the cost of giving up something else. Over time, economic growth allows us to produce more goods and services with fewer resources going into each of them, but this is a slow process. Even with growth, getting more of something is not free.

This applies to health care just as much as it applies to everything else. From society’s point of view, providing more health care means providing less of something else. When people become doctors, nurses, medical device engineers, or other medical specialists, they cannot also be computer programmers, chefs, or produce other non-medical goods and services. No health care system—government run, purely free-market, or anywhere in between—can produce as much free health care as everyone wants. They can be more or less efficient, depending on the incentives of the system (which I will discuss later), but unlimited free health care is not an option, and anyone that tells you it is possible is lying to you.

Economists call the process by which things get allocated to people “rationing”. In a purely market-based system, economists say that the price acts as a “rationing mechanism”—people who are willing and able to pay enough for something get it, and people who are not willing or able to pay enough for something do not get it. In a purely government-run system (say, Soviet-style central planning) bureaucrats decide who gets what, and base it on a calculation of needs, political influence, or some other measure. There are possibilities in between the two extremes.

In any health care system (as in every part of an economic system), someone ultimately says “no”. There are entities that decide who will get care, and who will not. In the U.S. those entities are insurance companies (who decide what to cover and what not to cover, as well as to whom policies are sold), governments (state and federal, which administer programs for government employees, the poor, and the elderly), and individual consumers, whose incomes limit the amount of health care they can afford. All these people—insurance companies, governments, consumers—can say “no” to getting a particular kind of care. Obviously a consumer would prefer never to hear “no” from the other two groups of decision makers.

One might object that doctors get to decide who gets what kind of care, if any, and to an extent that’s true—they (hopefully) will not treat a torn ACL with chemotherapy. Yet doctors are ultimately paid by someone, and if that someone (consumers, the government, insurance companies, charities, whoever) stops paying, they will eventually reduce the amount of care they provide.

In other countries the decision makers may be government officials. The NHS in the U.K. decides which treatments are acceptable for what health problems, and which ones are not acceptable (and therefore will not be funded). When controversies arise in such systems, they are often over whether the government should pay for one treatment or another (one would expect such controversies to arise most often when new treatments become available). Governments can also ration care using queues—care is given out first-come, first-serve. The waiting time for a service depends on what kind of resources a government allocates to it, and how efficiently those resources are used.

Whether government decision makers do a better or worse job than private decision makers is an interesting question, and you might have gathered from the discussion of international comparisons above, it is a difficult question to answer. Nonetheless there is a consensus that the U.S. system is not working as well as it ought to. Health care is getting more expensive rather than less (contrast health care costs with, say, the cost of DVD players). Why is this so?

There are two economic primary theories to explain why this would be the case: adverse selection and moral hazard (both having to do with the economics of insurance). Both can be expected to drive up the costs of health care, pricing many consumers out of the market.

The adverse selection argument goes like this: Suppose that there are two types of consumers, those who will get sick (the unhealthy) and those who won’t (the healthy). Suppose consumers know which type they are, but insurers do not. Then the healthy types, knowing that they do not get sick, will not buy health insurance. The unhealthy types, knowing that they will get sick, will try to buy health insurance. Should insurers sell to them?

They should not. If only the unhealthy types buy, then selling health insurance is a sure way to lose money. Each consumer is going to incur, say, $10,000 in average medical costs, which the insurer will have to pay, and then pay his workers and stockholders. But the consumer isn’t willing to pay more than $10,000 for the insurance policy, so there’s no way for the insurer to make a profit.

This is an extreme version of the adverse selection story, in which the insurance market completely disappears. A more realistic version would involve probabilities, rather than certainties, and when one creates such a model, an interesting result can emerge: If enough healthy consumers drop out of the market, health care can get sufficiently expensive that only the most risk-averse, wealthy consumers (of either healthy or unhealthy type) can afford to buy. That is, even if I am a healthy consumer, I might still want to buy insurance because I am particularly risk-averse—but I can’t, because it’s too expensive, because of adverse selection. If everyone could be forced to buy insurance, it would be cheaper for all. (Here is a clever related argument regarding car insurance, suggesting that cities that require all drivers have car insurance should experience lower rates.) Adverse selection is why health insurance companies try so hard to screen for preexisting conditions, smoking, and other factors that can increase the risk that a consumer will become ill; they are trying to sort out the good risks from the bad.

Does this argument help explain why health care is so expensive in the U.S.? I don’t think so. Recent reforms in Massachusetts have succeeded in getting almost everyone in the state (97% of the population) insured, yet health care costs have risen dramatically, and the state is having trouble with its budget. Adverse selection is really an argument about health insurance costs, not necessarily health care costs.

Moral hazard is a simple response to incentives: Once insured, people have fewer incentives to take precautions to prevent future illness, because the medical costs will be paid for by a third party (the insurer or the government). This makes small medical problems into big medical problems, driving up the cost of treatment (and the cost of insurance). In fact, in the U.S., around eighty to ninety percent of all health care expenditures are paid for by third parties (again, insurers or governments, rather than the patients). This has additional incentive effects beyond simply failing to take precaution. People stop shopping around when someone else is paying. They also request additional procedures or tests, or go to the emergency room instead of a less-costly physician. Doctors, who are paid for procedures and tests performed, have little incentive to resist, leaving it up to insurance companies or the government to say “no”—which, naturally, makes consumers angry, and they exert pressure to make “no” a difficult thing to say.

When a large portion of consumers are no longer paying attention to costs, and no longer shopping around, competition stops functioning. Competition—the process that gave us the $20 DVD player, instead of the $900 DVD player we started with—is an important mechanism for disciplining producers, whether they are producing cars, newspapers, computers, or cancer treatments. Competition keeps producers on their toes, forcing them to find better, more efficient ways to do things. I find the moral hazard/lack of competition argument compelling. In health care, competition seems broken.

Robin Hanson has argued that overconsumption of medical services means we could cut health care spending in the U.S. in half (by simply denying care or refusing to undertake procedures or tests) with almost no effect on health care outcomes. If correct, is a pretty stunning result—if we can get people to consume more carefully, we can dramatically reduce costs.

What reforms should be enacted? It is possible that moving to something like Canada’s system might be an improvement; I don’t know. I am skeptical because our government has budgetary problems with the current large medical program, Medicare. Taking on additional large expenses will be problematic. Furthermore, I do not think our federal government is good at running technocratic programs. Remember, there are more than 300 million people in the U.S. (compared to 33 million in Canada); I am not convinced that the federal government can run a health care program that serves them all. That is, I do not think the U.S. government would run Canada’s health care system as well as Canada has. The more government involvement in the system, the less competently I think the U.S. government performs. I shudder to think how badly the U.S. government would run a system like the U.K.’s NHS.

This brings me to a reform that I think would work. Bryan Caplan, Tim Harford, and others have praised Singapore’s health care system. It seems designed to address all the problems people have with the U.S system, without requiring a large amount of government management or intervention in health care markets. It is certainly not laissez-faire, but it is not single-payer or nationalization, either.

You can read Caplan’s account of Singapore’s system here. I think the most important features of Singapore’s health care system are:

1) Mandatory medical savings accounts: Every person pays into a savings account that accumulates over time, and can even be passed on to one’s children. The money in the account can only be used for medical care. This gives consumers an incentive to shop around—they are spending their own money, after all. It may also give them an incentive to undertake preventative care, as doing so increases the amount in their account in the long run. This policy puts competition back in health care markets.

2) Mandatory published price lists: All hospitals and medical facilities must say what their procedures and tests cost. This allows consumers to shop around, and even establish reputations for low prices. Try calling a local hospital in the U.S. and ask them what a procedure costs; they usually won’t be able to tell you. Your insurer might be able to tell you what they pay a particular hospital, but they might not. This policy, like the first, encourages competition.

3) Optional low-cost catastrophic coverage: When health care expenses go over a certain threshold, the government steps in and pays. Because catastrophic illness is uncommon, this coverage is relatively cheap to provide. Private insurance can also be purchased. This is sort of like the “public option” of the current health care reform proposal in the U.S., but it is far less generous, covering extreme situations rather than routine medical expenses.

There is one aspect of Singapore’s policy that I’m not so sure about: supposedly the government pays for“80% of basic health care services”. I’m not sure what that means. I’m guessing checkups and other services that catch disease early. It surely can’t mean a much broader swath of health care services, or the “buffet mentality” (as they call it) would emerge, with some people overconsuming care, driving costs up. So far, Singapore has managed to provide a high quality of care with low costs (that rise slowly!).

I think such a system might be manageable in the U.S. because the government’s role in it is simple; competitive markets do most of the hard work. There are other reforms that should also be enacted in the U.S., such as removing (or at least equalizing) the tax advantage of getting insurance through an employer, and making interstate health insurance sales legal. There may be other ways to get similar results to that of Singapore, perhaps with health care vouchers (much like education vouchers).

Realistically, however, I do not think we will get reforms that will make a significant difference, for better or worse.

Thinking Out Loud: An Idea for Stimulus

I've already talked about why I don't like the Cash for Clunkers program. In short, it is not likely to be environmentally friendly, since the process of producing new cars creates pollution, and because old cars must be disposed of. It also has a bit of a broken window fallacy aspect to it, destroying something old just so we can have the "pleasure" of replacing it.

Another reason to disapprove of this program is that the government has chosen to pick winners and losers. Car purchases get stimulated, but other purchases do not. The government is altering relative prices, encouraging them to buy cars when, absent the reduced relative price of cars, people might benefit more from buying new clothes, or a computer, or bus passes, or whatever. Resources are not being put to their highest-valued use; rather, they are going to the use the government favors.

On the other hand, the program has been hailed as one of the few pieces of clearly successful stimulus spending. Tax cuts handed out last summer were mostly saved rather than spent, for example, yet people are scrambling to take advantage of Cash for Clunkers. My in-laws were considering buying new cars at some point, but they have moved their purchasing decision up to right now, rather than later. In terms of economic stimulus, one could argue that this is a good thing--sure, people are just transferring their car purchases through time, from the future to right now, but on the other hand, we presumably need stimulus now, rather than at some point in the future when the economy has recovered.

Why is Cash for Clunkers different? Why are my in-laws running out to buy cars now, whereas everyone just saved the tax cuts? Because this program is temporary. There has been a total of $3 billion allocated to this program, and once it runs out, the program is over. Apparently the temporary nature is enough to get people to run out and buy now.

So why not apply this more generally? Instead of having the government pick this or that place to spend money--on roads here, on a solar power plant there, and on some representative's pork barrel project some other place, why not just give people time-limited cash, which must be spent at retail?

That is, the government could take, say, $200 billion of the stimulus package, and say "Every American may receive up to $1,000 in federal spending for any purchase of their choice. However, once the money is gone, the program is over. If you do not use up your $1,000 before the money is gone, it disappears."

There is no change in relative prices (except for that between consumption and saving, which is part of the point). Retailers would have a form to fill out, similar to that of Cash for Clunkers, which would allow them to receive payment from the government.

One problem is that people might have a hard time tracking their spending. For example, if I buy $800 worth of stuff at Wal-Mart, I have $200 left over for other uses. Costco doesn't know this, so when I go to Costco and try to charge $300 to the government, there will be paperwork and bureaucracy as the government tries to figure out what to do with the $100 overcharge. Cash for Clunkers doesn't run into this problem, since every new car far exceeds the $4,500 maximum subsidy. One solution may be for the government to issue everyone a debit card with up to $1,000 on it, and then when a total of $200 billion is spent, all the cards get cancelled.

There are potential public choice problems, too. This sounds like something politicians would love to abuse. They would probably love to hand these out during booms, too, defeating the counter-cyclical purpose for which I proposed them. I suppose there could also be problems with fraud.

Still, it could be an improvement over the mess that is current stimulus spending, couldn't it? I'd love to hear additional criticism in the comments.

An Answer to the Exam Question I Should Have Asked

Former student Michael Duch provides the following answer to the question I posed:

The demand for water is inelastic! The fact that total revenue rose after the price increase implies that the percent change in price > percent change in demand, which gives us an elasticity ratio less than 1.

And the price increases should have been enacted first because they would have reduced the amount of water used while allowing people to maximize their utility. If this failed, which, it probably would have (assuming I'm right about the inelasticity of water) restrictions should have been imposed. I think the conservation of water in a drought takes precedence over allowing someone to maximize utility. But that's just me.

Some comments: Michael is correct in his conclusion that the own-price elasticity of demand for water must be less than one, and his reasoning is correct, too. Own-price elasticity of demand is:

Absolute value ( (percent change in quantity demanded)/(percent change in price) )

...and total revenue is simply Price * Quantity. If price goes up, quantity must go down (that's the Law of Demand), but what if revenue still rises? Then the increase in the price outweighs the decrease in quantity. In terms of elasticity, the denominator must be larger than the numerator, meaning own-price elasticity of demand is between zero and one--which we call inelastic. Good job, Michael!

Regarding the second paragraph, I agree that it would be better to raise the price first, but I would like to elaborate on why that is a good idea. Raising the price allows people to adjust in a variety of ways. For example, someone who has fruit trees in their yard might be willing to pay extra in order to make sure that the trees bear fruit, whereas a simple watering ban doesn't allow that option. Other people might decide it's better to just forgo watering altogether, or to collect rainwater, or take other actions. A price increase allows this variety of responses, while still achieving a given reduction in use.

Watering bans are usually riddled with exceptions for car washes and golf courses; there is no economic justification for this. If the water to be used at a car wash is too valuable for that use, then it shouldn't be used. The car wash, faced with an increase in water costs, will try to pass that cost increase to customers in the form of higher prices. If customers are unwilling to pay it, they won't, and the car wash will shut down. This is exactly what we should want to happen. We should want low-valued uses of water to stop, and allow only high-value uses of water to persist. Simple restrictions on use can't achieve this, but a price increase can.

More on Cash for Clunkers

Here's some more on the environmental effectiveness of Cash for Clunkers. Nina Shen Rastogi argues that it was too costly. This AP piece makes the same point, claiming that the program will save the equivalent of shutting down all cars and power plants in the country for one hour--a miniscule fraction of our annual pollution production. Jeffrey Miron, one of my favorite economists (if you haven't read his 1995 Journal of Economic Perspectives article on the economics of drug prohibition, you should do so immediately), argues against the program here.

It has surely stimulated auto sales, although with a $3,500 or $4,500 subsidy, you can stimulate lots of things. It would make more sense to give people a payroll tax cut or even a direct payment, so they can spend the money on what is important to them, rather than trying to decide what specific products people should be buying.