The Cost Disease: Why Computers Get Cheaper and Health Care Doesn't

  • William J. Baumol
Yale University Press, 2012 272pp., hardcover, $30.00 0300179286 | ISBN: 0-300-17928-6

In his latest book, William J. Baumol returns to an insight he (and William G. Bowen) put forth in 1966. They originally observed that prices of goods produced in industries with rising productivity fall relative to prices of goods produced in industries with little or no increase in labor productivity. Because it always and everywhere takes four players to perform a string quartet, they pointed out that the price of musical performances will rise relative to that of most other goods.

Baumol now extends this theory to health care and educational costs, thus explaining why expenditures in these areas have claimed a growing share of total income. However, for health care I believe that claim is wrong. To understand the problem, start with the simple fact that total expenditure (E) on any commodity equals the price (p) times quantity (q): E = p × q. Next, the word 'cost' can mean different things. Sometimes it refers to p (for example, the cost of a television set) and sometimes to E (for example, the total cost of health care). If one slips back and forth between price and total expenditures, confusion is inevitable. For example, with respect to computers, their officially measured price fell 99% between 1980 and 2011, but expenditures for these items increased eightfold between 1980 and 2000 and fell about 30% since then. So in one case the 'cost' of computers fell; in the other it increased.

Indeed, the author's thesis does not explain why, or even whether, expenditure on the product of a technologically stagnant industry should be expected to rise or fall relative to expenditure on other goods. For example, if economy-wide output per worker doubles, one would expect wage rates to roughly double. So if wages of artistic performers keep up with the average, their compensation would also roughly double, and thus so would the price of artistic performances. But whether total expenditures on a given product, be it artistic performances, health care, or anything else, rise or fall depends on the relative effect of rising prices and rising incomes on aggregate demand (or quantity) for that product, with the former leading to reduced demand and the latter to an increase.

To be sure, official indices indicate that current health care prices are 15.7 times what they were in 1960, and total health care spending, at 17.9% of the gross domestic product, is more than triple the 5.2% share in 1960. But has the price of desired health outcomes (reduced pain, extended life and improved functionality) actually risen? Health care is not a stagnant industry, and one can buy vastly more of these desired outcomes with today's menu of health care services (think magnetic resonance imaging and other improvements in diagnostics, modern and powerful drugs, joint and organ replacement and so on). As E = p × q, the question then becomes whether the quantity of desired health outcomes has risen as much as or more than have health expenditures. If so, then the equation tells us that p (the price of those health outcomes) must have fallen, regardless of what official price indices may say.

Like most people, you will not choose to forego currently available services. You will bridle at the amount spent that produces little or no benefit or that is needlessly costly. But, however grudgingly, you will pay the price. If you do, it could mean that, whatever official price indexes may say, the price of achieving desired health care outcomes has fallen, as I outlined at the end of the previous paragraph. Indeed, studies in refereed journals by leading health economists have concluded that if one measures the price of improved outcomes, treating coronaries and mental illness has gotten cheaper. Alternatively, one might say that wholly new services are now available and that these new products are highly valued. A flood of highly valued new goods is not the hallmark of a stagnant industry. To suggest that health care spending is driven by the fact that it costs relatively more now than in the past to have a nurse change the sheets on a bed or provide other technologically unchanged services is absurd.

That the health care industry is rife with inefficiencies is beyond much dispute. But the cost disease isn't about inefficiency; it is about an alleged fundamental structural challenge—that the proportion of income spent on products of technologically stagnant industries necessarily tends to rise. Because of offsetting price and income effects, this outcome for health care is not inevitable (that is, there is no cost disease for this industry) and wouldn't be necessary even if health care were a stagnant industry, which it isn't.

To finish on a positive note, Baumol is powerful and persuasive on one central issue, which is a different premise from the rest of the book that he unfortunately leaves for its end. If economy-wide productivity continues to grow, we can continue to consume more of other goods even if total spending on health care continues to claim ever larger shares of our incomes. Indeed, the stakes in reducing medical waste go up as the size of the health sector increases. The political problems raised by paying for health care for the elderly, disabled and poor through public programs also intensify. Baumol does a major service by pointing out that the continuing growth of health spending will not prevent us from having more of other goods if we want them.