In November, Joseph Stiglitz, George Akerlof, and Michael Spence shared this year’s award for economics, primarily for their academic contributions to the economics of “asymmetric” information.
Theory Makes Bad Policy
Columbia University’s Stiglitz is particularly noted for his emphasis on the possibility of market failure in competitive health insurance markets because of adverse selection and imperfect information. But his market skepticism has had its own adverse consequences, bolstering support for policies favoring more government regulation and less consumer choice.
In a seminal 1976 article, co‐authored with Princeton economist Michael Rothschild, Stiglitz observed that if insurance buyers know more about their own health than do insurers, then buyers who know their risks are greater than average will purchase more insurance when it’s priced based on average risk. Buyers who know their risks are less than average will purchase less insurance. This sequence of events is termed “adverse selection.”
Over time, insurance premiums will increase as those with risks less than average self‐insure more often and more extensively. In the worst‐case scenario, the overall health insurance market can break down completely.
Twenty‐five years ago, Stiglitz wrote, “Do these theoretical speculations tell us anything about the real world? In the absence of empirical work it is hard to say.”
Unfortunately, health care policymakers have generally acted as if adverse selection was a dominant reality and not just a theoretical curiosity. Such thinking provided legitimacy for all sorts of regulatory interventions and compulsory health insurance schemes that favored equity over efficiency and risk‐spreading over personal health maintenance incentives.
Government Failure, Not Market
By exaggerating the relative significance of adverse selection in health insurance markets, pro‐regulatory forces could blame inefficient markets for leaving parts of the population uninsured.
Market‐oriented reforms like medical savings accounts, defined contribution health plans, and Internet‐based individual insurance options have collided with worries that they would increase adverse selection and make health insurance unaffordable for the high‐risk people “left behind” in traditional employer group plans featuring comprehensive coverage.
To ward off such adverse selection, consumers are denied a level playing field approach to market‐priced health care alternatives. They must maneuver past artificial barriers like mandated benefits, modified community rating, restrictions on risk classification, and regulatory cross‐subsidies, as well as tax policies biased toward employer‐sponsored group health insurance.
Ironically, public policy barriers to competitive choices in health care represent government failure –built not only on imperfect information, but on deeply flawed assumptions.
Adverse Selection Doesn’t Apply
There’s little evidence that individuals and families can identify and anticipate most of their future medical expenses in ways potential insurers cannot.
In a study forthcoming from the Cato Institute, James Cardon and Igal Hendel find little empirical evidence of information asymmetries, market failure, and adverse selection in health insurance markets. They note that differences in health expenditures between the insured and uninsured are mostly due to observable differences in demographics (age, gender) and price sensitivities (higher‐income workers capture more tax subsidies for insurance coverage), rather than unobservable factors related to health status.
Private insurers aren’t helpless or clueless when it comes to adverse selection. They may use such tools as setting limits on plan switching, varying premiums according to the amount of insurance purchased, underwriting and rating based on risk categories, creating more homogeneous risk pools, or relying on the law of large numbers to diversify risks in large employer pools.
Consumer inertia, individual differences in aversion to risk, and the administrative complexities and costs involved in making risk adjustments within group health plans all further limit the applicability of adverse selection theory to the actual health insurance world.
Many of the difficulties we observe in health insurance markets are due to government intervention rather than adverse selection or other market failures.
If insurers are not allowed to charge different premiums to different risks in competitive markets, they may resort to excluding or discouraging coverage of high risks. Cream‐skimming (selecting only the best risks) becomes the insurers’ mirror image of adverse selection by insurance customers.
Political interventions in health insurance markets don’t alleviate underlying differences in risk across customers or eliminate insurers’ knowledge of such differences. They only force insurance companies to cope in inefficient ways and create new problems.
Health status information is most likely to be asymmetric when it is scarce and costly. While government mechanisms prefer to ignore, hide, or shift those information costs, markets create proper incentives to discover efficient ways to signal relevant private information and put it to use.
Need for Reform
Market mechanisms can’t eliminate every unfortunate human experience in health care access, affordability, and quality. Private charity and a backup safety net of transparent, direct subsidies have necessary roles to play.
Beyond a certain point, any reductions in adverse selection will cost more than they are worth. But we do need to open more breathing room for market innovations to fill the real gaps that exist in insurance coverage. Effective reforms include bolstering incentives for new forms of voluntary risk pooling and long‐term insurance contracts, deregulating insurance choices, and providing tax equity for all insurance purchasers.
The growing availability of online health information and insurance products further strengthens the case for empowered consumers. Unlike centralized government “solutions,” markets don’t promise perfect outcomes, just better ones.