Researchers on the economics of payment and managed care are well aware of the issue. Ellis (1998) labels underprovision of care to avoid bad risks as “skimping.” Newhouse et al. (1997) call it “stinting.” Cutler and Zeckhauser (1997) call it “plan manipulation.” As Miller and Luft (1997:20) put it:

“Under the simple capitation payments that now exist, providers and plans face strong disincentives to excel in care for the sickest and most expensive patients. Plans that develop a strong reputation for excellence in quality of care for the sickest will attract new high-cost enrollees….” The flip side, of course, is that in response to selection incentives the plan might provide too much of the services used to treat the less seriously ill, in order to attract good risks.

“Too much” is meant in an economic sense. A plan, motivated by selection, might provide so much of certain services that the enrollees may not benefit in accord with what it costs the plan to provide them (Newhouse et al., 1997:28). An important implication of this observation is capitation and managed care can be expected to generate too little care in some areas and too much care in others.2 This leads, then, to the questions: How does a regulator know which services a managed care plan is skimping on/over-providing to affect risk selection? Even if the regulator did know, what could it do about it?

Motivated by these questions, public regulatory bodies and private payers have recently become very interested in monitoring the quality of care in managed care plans. Monitoring health care quality has become an industry, virtually overnight, though no single approach has gained wide acceptance. Monitoring consists of identification of measureable standards (consumer satisfaction, health outcomes, quality of inputs) against which a plan’s performance is compared. There are many drawbacks to this approach, from a policy and an economic standpoint. At a recent conference, observers noted that standards have proliferated, and it is difficult to find standards that are sensitive to system characteristics (Mitchell et al., 1997).

The standards are at best imperfect indicators of value to enrollees. Ranking the importance of different standards is largely arbitrary. Quality can be too high as well as too low, and existing approaches are all oriented to a minimum not a maximum standard. Gathering information on many standards for many plans in a timely fashion is very expensive. Plans do not all have adequate administrative capability (Gold, 1995). Enrollees move in and out of plans, making measures based on performance at the person level difficult to implement. Rewarding a subset of quality indicators may distort performance by health plans.