A classic. This is this Nobel Prize winner's most-cited article. Arrow won his Nobel for laying out in mathematical terms the necessary prerequisites for a free market to have optimal outcomes. This is what "welfare economics" in the title means. It has nothing to do with public assistance programs.
This article is heavy going, especially the appendix. (That appendix gives you a taste of what Arrow's book on economic theory is like. If you understand it, please explain it to me!) If a paragraph seems incomprehensible, go on to the next. Also, some of the ideas are anachronistic. Even so, this article repays a careful reading, especially if you have not read much economics before.
The important part of the article begins near the top of page 952. There, Arrow explains why health care doesn't fit the free market ideal. The basic problem is that, unless you know as much about medicine as your doctor, you can't evaluate the quality of advice he or she (mostly "he" in 1963) gives you. He calls this "uncertainty." I see this as the definitive answer to those who claim that success in the market validates a hospital or an insurance plan.
Go back to p. 942 and the First Optimality Theorem.
First off, "utilities" in this sentence means the individual's well-being as he or she sees it. "Utility" is another term that economists use in a peculiar way.
Secondly, be aware that economists use the word "optimal" in a peculiar way. The idea of optimality here, also called Pareto Optimality (named for the economist Wilfredo Pareto), is that an allocation of resources is "optimal" if there is no way to rearrange things that makes somebody better off without making at least one other person worse off. Usually, when you use the word "optimal," you mean "the best." In the economics usage, however, there there are many possible "optimal" allocations.
For example, consider these scenarios:
By the way, another word that economists use in a peculiar way is "efficient." In economics jargon, "optimal" is synonymous with "efficient." In other words, "efficiency" means that no one can be made better off without making somebody else worse off. It's the same Pareto criterion.
The First Optimality Theorem says that if certain assumptions are true, then the free market will produce an optimal allocation of resources. Two of those assumptions are about consumer sovereignty and perfect competition.
If both of these hold, then then sellers make money by pleasing the consumer, offering the combination of quality and price that consumers want. The consumers are in command, subject only to what they can afford to spend. Who runs Exxon-Mobil? You do! (You do, that is, if you have money to spend, and if you argue that consumer sovereignty and perfect competition actually describe our economy.)
If and only if the economy is efficient, all goods and services have an opportunity cost. Having an opportunity cost means that obtaining that good or service requires giving up something else. The value of the next-best something else is the opportunity cost of the good or service.
"There is no such thing as a free lunch." This truism is generally false. It's true only if the economy is efficient. For a big example, consider the U.S. effort in World War II. During the war, Americans' average real consumption of economic goods went up, because the economy had been in a depression, with many unemployed people and unused factories. For the U.S., from an economic standpoint, World War II was a free lunch.
Similarly, as I write this in September 2011, President Obama is proposing to have unemployed construction workers fix roads, bridges, and schools. That fixing would have practically no opportunity cost.
p. 943. The Second Optimality Theorem says that if you don't like what the economy is producing you can fix it -- switch to a different optimal allocation of resources -- solely by redistributing wealth. You don't have to regulate markets. (Of course, as Arrow points out in a footnote, redistributing wealth is politically difficult.)
There are other necessary assumptions to make the First Optimality Theorem true besides consumer sovereignty and perfect competition. Arrow emphasizes those two here because he thinks that both of them are problematic for health care.
One of those other requirements for the free market to produce optimality is "no increasing returns in production." Increasing returns means that bigger is better, that average cost per unit produced goes down as the size of the firm goes up. This assumption, too, is questionable. Beer brewing and other chemical processes are cheaper per unit when bigger. Big insurance companies have an advantage over small ones, too, because the big company is less likely to be forced into bankruptcy by disastrous bad luck. Increasing returns messes up the competitive model. If bigger is better then competition will force firms to grow. If bigger is better, no matter how big you already are, then one firm will eventually dominate each market. That will be the end of competition. (Note, by the way, that increasing returns is the opposite of the law of diminishing returns.)
Another requirement for the free market giving us optimality is the "marketabilty of all goods and services relevant to costs and utilities." (p. 944.) Something is "marketable" in Arrow's sense if you can buy or sell it. Arrow uses the word "commodity" is anything that is marketable. (Business people use the word "commodity" more narrowly to mean goods that are uniform and interchangeable, like bushels of hard red spring wheat.)
What's something relevant to cost and utility that's not marketable in Arrow's sense? Externalities. These are costs and benefits that our decisions to buy or not to buy impose on others who are not a party to the transaction. These costs and benefits are called externalities because they are outside the market. For example, failure to immunize yourself raises the risks for others. Ideally, those other people could pay you to immunize yourself, but there's no way to do that without government or community action to compel people to get immunized or to pay taxes to subsidize others' immunizations.
Back to Arrow's main argument: The first full paragraph on page 946 makes a crucial point about consumer sovereignty in health care. Arrow observes that much of what we buy from physicians is advice. (Some medical care is dexterity, particularly surgery, but much of what we buy from medical professionals is advice.) If we knew enough to know how to evaluate the advice, we would know the advice, and wouldn't have to buy it.
This is a problem with the purchase of any kind of information. Think of detective TV shows or movies where a stoolie says he has information and tries to bargain for payment or a reduced sentence. The detective doesn't know what the information is worth, so he or she does not know for sure what to offer for it. In economics terms, this is a failure of consumer sovereignty.
p. 947 Arrow goes on to speculate that the peculiarities of medicine -- the fact that we have "professional ethics," for example -- stem from attempts, imperfect though they may be, to deal with the basic failure of consumer sovereignty, which Arrow calls the "imperfect marketability of information."
p. 949's last paragraph lists some illustrations of differences between how physicians behave compared with other business persons. Illustration (1) seems quaint today. Illustration (3) shows a problem that was important then and still is today, but managed care has changed the direction of the financial incentive from doing too much to doing not enough. Arrow's ideas on for-profit and non-profit hospitals (p. 950) may (or may not!) need modifying today. This article, by the way, was written before Medicare and Medicaid, so hospitals had to do much more uncompensated care then.
p. 951 Product Uncertainty -- Again, crucial differences between medical care and other commodities. Many products you buy have an information uncertainty component. You don't know, before you try, how well a new laundry detergent will clean your clothes. You may know as little about laundry detergent as you do about medical science, but you can learn from experience with the laundry detergent. You can also gain experience with a particular seller and judge the likely quality of its future advice from how well its past advice worked out. This is a role product brand names play. Health care, for most people, is highly episodic. You may get only one chance to make the correct decision, and the consequences of a mistake can be severe and permanent. In health care, experience generally cannot solve the product uncertainty problem.
p. 953 "Price discrimination" means charging different customers different prices for the same service.
p. 955 "Indivisibility" means that a resource comes in minimum-sized lumps. Workers (people) are indivisible resources. They can only be in one place at one time. This is a major problem for providing health care in rural areas. For many types of facilities, there's a minimum size, and if there are few patients in the area, the average cost per patient becomes prohibitively high.
p. 957 The D. Pricing section is skimmable. It's an argument against the claims of some economists (Reuben Kessel, for example) that doctors are monopolists, pure and simple. The point in here about "elasticity" is this: Elasticity is a measure of the responsiveness of the amount demanded by buyers to changes in the price. (The economics tutorials on elasticity are assigned for next week.) An elasticity of less than one means that the amount demanded does not fall much when price is raised. In such a circumstance, a ruthless monopolist would raise the price, and keep raising until the elasticity became greater than one, meaning that the quantity bought did start to fall significantly. This is called "charging what the market will bear," and it's what monopolies do, in theory. Medical care prices, high as they might seem, are not as high as that. Demand is inelastic at current prices. (One thing managed care has done has been to increase the elasticity of demand that an individual provider faces. If a doctor asks for higher payments from an insurance company, the company may tell him to get lost and direct its "covered lives" to other providers. That is highly elastic demand.)
p. 958 Section IV gets into the theory of insurance, which we will discuss separately. I regard section IV. C. Problems of Insuranceas somewhat distracting from Arrow's main argument. I'd say it's skimmable.
The patient's uncertainty about his treatment is different from his physician's, because of the difference in knowledge and perspective. Arrow suggests as a theoretical idea that physicians (and, by extension, all health care providers) should be paid only for results, not for effort. That would be one way to align the provider's incentive with the patient's well-being. Such arrangements don't exist, however. Results are hard to measure. Instead, we have a system that relies on trust by the patient and social obligation by the physician. The physician is, morally and legally, the patient'sagent. An agent (the provider) acts on behalf of the principal (the patient) as if the agent were the patient.
A legal term for agent is fiduciary. If you are managing someone
else's financial affairs as a fiduciary, you are legally required to act in
the sole interest of the beneficiary, the person whose affairs you
are managing. Physicians are not complete fiduciaries in that sense, because
they also have social obligations, like the reporting of communicable disease,
that may conflict with the patient's interest.