In lieu of an abstract, here is a brief excerpt of the content:

Journal of Health Politics, Policy and Law 26.5 (2001) 829-834

[Access article in PDF]


Mark V. Pauly
University of Pennsylvania

Kenneth Arrow's article made research in health economics respectable, but it did more than that. It also made it interesting. His essay was the gracious response of an already distinguished economic theorist to an invitation to write something for the Ford Foundation on the economic properties of the medical services industry. Before this article appeared in 1963, economists, even those interested in industry applications, had either steered clear of trying to understand (much less analyze) a product and an industry that appeared to depart so greatly from the competitive model, or they had written policy papers alternately justifying or condemning the differences from that model. Either perspective was largely irrelevant because medical care was "special" or because economics could be applied only to show that the guild and regulatory features of this industry were the result of a long-term conspiracy by medical providers to gain monopoly rents at the expense of efficiency and consumer welfare.

Arrow's article was and still is exciting, I believe, for two reasons:

1. It showed how some behaviors in medical markets could be brought within the purview of standard economic models of competing, maximizing agents. The primary examples of this type are those having to do with insurance.

2. It offered an explanation that atypical institutional arrangements in medical care markets are a reaction to special features of this market. The primary source of the problem here is imperfect or asymmetric [End Page 829] consumer information, and the hypothesized solution was nonmarket organizations, public or private. In doing so, it discussed concepts that made (and make) economists attentive but uncomfortable, like trust and morals. (These concepts do not tend to disquiet lawyers or policy makers.)

The essays in this remarkable collection reflect on these two different sets of ideas. They analyze and dissect them, they contrast their implications for what medical markets were like then and now, and how they have changed, and they offer suggestions about where to go from here. I will not summarize the individual essays, but in what follows I will offer my views on the reactions to the two sets by these scholars, by researchers in general, by the market, and by the policy arena.

From the viewpoint of applied microeconomic theory, a major contribution of Arrow's article was the development of some typically elegant and provocative models of optimal insurance. There was a policy question here, of course. In 1963, why was it that many medical care services were not covered by insurance and that many people had no health insurance coverage at all? (The proportion of uninsured then was about twice what it is now.) Arrow began the discussion that actually generated the most-published commentary on his article by restating the well-known proposition that if insurance is available at actuarially fair premiums, all risky events should be insured. But he then pointed out in the text, and developed further in an often-quoted appendix, that in real markets with insurer administrative costs, the optimal pattern of insurance is full coverage above a deductible. He also hinted at (though not forthrightly enough for my tastes) the idea that moral hazard--any effect of insurance coverage on expected losses--could also explain incomplete coverage and that one remedy here would be proportional co-insurance.

The introduction to health economics of the topic of moral hazard ignited a firestorm of interest in the impact of insurance on the process of care, with both theoretical and empirical dimensions. The major empirical topic was the measurement of the extent of moral hazard, and the major theoretical question was how that magnitude would affect the ideal design of an insurance policy that relied on patient cost-sharing to limit medical spending. The Rand Health Insurance Experiment defined the measurement of moral hazard as a key empirical question and provided bulletproof evidence on the question, but both the theoretical and empirical investigations of this topic continue the extension of Arrow's initial modest...


Additional Information

Print ISSN
pp. 829-834
Launched on MUSE
Open Access
Archive Status
Archived 2005
Back To Top

This website uses cookies to ensure you get the best experience on our website. Without cookies your experience may not be seamless.