issue from five perspectives: the costs of health care, access to care, quality of care, efficient use of human resources, and the future of academic health centers.

The essential questions are the following: Do we know now what effect very high numbers of physicians might have? Will we be able to project the effects with a satisfactory level of precision, especially as the system itself is evolving unpredictably? Do the answers differ depending on what assumptions are made about the degree of penetration of managed care and the size of the residual fee-for-service (FFS) sector?

As noted in Chapter 2, forecasts of supply and, especially, requirements are fraught with many uncertainties and difficulties. A small mismatch of supply and requirements in one direction or the other probably should not be taken as evidence of a true or meaningful difference. In this discussion, therefore, the committee considers the above-mentioned issues on the assumption that substantial numbers of physicians are currently available, large numbers are continuing to be produced, and even higher numbers—indeed, a surfeit in most experts' views—can be forecast for the near future.

Using standard models of microeconomic theory, one might expect that a significant increase in the number of physicians would increase competition, produce lower incomes for physicians, and reduce health care costs. Theory, however, is not entirely or consistently borne out in practice.

Few data document whether ''competition" in the classic sense—among physicians per se—has increased, although the fact of discounts, sometimes sharp, in physician fees and hospital charges suggests that competitive forces have been at work in recent years. Clearly, the nation's health care system is moving in a rapid and unstructured way to far greater levels of competition than it has ever known. This phenomenon appears mainly to involve competition among health care plans, integrated delivery systems, and hospitals.

As the country moves to larger, but fewer, networks and systems and downsizes in certain areas such as hospital beds, physicians will increasingly find themselves attached to or employed by such plans; in principle, competition among them—even if not overt—might increase. Nonetheless, that scenario is somewhat afield of traditional ideas of direct competition among individuals or groups (in the form, for instance, of price wars or limitation of staff privileges). In sum, no clear evidence exists about the nature of the relationship between a