Covering the Uninsured

Featured in print Bulletin on Aging & Health

The U.S. is the only wealthy industrialized nation that does not guarantee universal health care for its citizens. Currently 47 million Americans, or 18 percent of the non-elderly population, lack health insurance. The problem of the uninsured has been a major concern for policy makers for years and is back in the spotlight this year due to the presidential election and a recent major health care reform in Massachusetts.

In "Covering the Uninsured in the U.S." (NBER Working Paper 13758) researcher Jonathan Gruber reviews the lessons of economic research regarding the uninsured and conducts simulations to show the effects of several popular reform proposals to increase access to care.

Gruber begins by noting that most non-elderly Americans receive health insurance through their employer. The reasons for this include risk pooling (insurers like groups of employees because they offer a predictable distribution of medical risk, unlike people buying individual insurance policies who may be doing so because they are already sick), administrative costs (these are lower in a group because fixed costs are spread over more members), and the tax subsidy (insurance premiums paid by the firm are exempt from individual income taxation).

In answer to the question "who are the uninsured?" Gruber notes that they tend to be lower income, yet not all the uninsured are poor. Seventy percent are in families with a head who is a full-time, full-year worker and either is not offered insurance at work or does not take it up.

Standard economic models predict that under certain conditions, people should want to fully insure themselves against medical risk. So how can we explain the widespread failure to insure? One explanation is that insurance is actuarially unfair, meaning that expected benefits are less than premiums. This can occur because of adverse selection - buyers in the individual insurance market are more likely to be sick, so insurers set premiums at the level necessary to cover their costs, making insurance a bad deal for healthy people - or because of administrative costs, which average 12 percent of premiums in the U.S.

A second possible explanation for the failure to insure is that hospitals essentially provide free catastrophic care, since they are required by law to treat individuals who arrive in an emergent state regardless of ability to pay and often fail to collect payment for these services. Hospitals spend $30 billion per year on such uncompensated care, but evidence is mixed as to whether the provision of uncompensated care deters insurance purchase.

Perhaps the most frequently given explanation is the high cost of health insurance. Evidence has linked higher insurance premiums with a rise in the number of uninsured, yet the mechanism by which this occurs is not clear. If premiums are rising because the cost of treating illness is rising, people should be even more interested in insuring themselves against the risk of medical expenditures. It may be that individuals value insurance but lack the ability to borrow against future income to pay for it, that individuals have time inconsistency problems and overvalue premium costs today relative to future medical expenditure risk, or that government regulations force insurers to cover services that individuals value at less than cost.

From the puzzle of the uninsured, Gruber turns to another puzzle - most insured people have more generous insurance than what would be optimal. Moral hazard - the idea that having insurance increases the prob-ability people use medical care and the amount they use - means that the optimal insurance plan includes significant cost-sharing in order to balance the gains from insurance with the costs of moral hazard. Gruber lists some possible reasons for overinsurance, including the tax subsidy to employer-provided insurance, regulatory requirements, psychological motives (people prefer not to associate financial transactions with medical care even if they must pay more up front), and the offset hypothesis (skipping primary care visits will lead to higher expenditures later).

Gruber concludes "there are a variety of hypotheses for why so many individuals are uninsured, but no clear sense that this set of explanations can account for 47 million individuals. Moreover, an equally important and related mystery is why insurance is so much more generous for most individuals than would be suggested by optimal insurance theory." Gruber calls for the development of a unified theory that can explain both the large number of uninsured and overinsurance among the insured.

Gruber also asks why we should care about the uninsured. One reason is externalities, costs that insured people bear as a result of uninsurance. These include physical externalities from communicable diseases (uninsured people may spread disease to others) and financial externalities from uncompensated care. However, these costs are too small to provide a major rationale for universal coverage. A second reason is "job lock" - workers may be afraid to change jobs for fear of losing their health insurance, which can lead to a mismatch between workers and jobs and lower U.S. productivity. Other reasons include paternalism (people may not realize the value of insurance) and redistribution (many of the uninsured are low-income).

Finally, Gruber develops a microsimulation model to estimate the effects of various reforms to increase cover-age. The model uses data from the Current Population Survey and data on health insurance premiums. The model first calculates how the reforms would change the price of different kinds of insurance, then runs these through a detailed set of behavioral assumptions (based on the best available evidence) to predict their effect on the behavior of individuals and firms.

Gruber considers two modest reforms to increase access - expanding Medicaid to cover low-income adults and introducing tax credits for the purchase of non-group insurance (also limited to low-to-moderate income families). To make meaningful comparisons, the reforms are designed have similar coverage impacts, reducing the number of uninsured by either 5 or 10 million persons.

The Medicaid simulations show that the cost per newly insured person would be about $5,000, that most of the benefit would go to very low-income individuals, and that there would be relatively little crowd-out (dropping of private insurance to enroll in Medicaid), particularly with the smaller reform. The tax credit simulations show that a higher fraction of those using the credit would be previously insured and that many workers would lose their insurance due to firms dropping their coverage. This policy is more expensive, costing about $8,000 per newly insured person, and more of the benefits go to people higher up the income distribution.

Gruber also simulates the effect of more fundamental reforms. These reforms include features like premium subsidies for low-income individuals, new pooling mechanisms, and individual mandates, but maintain private insurance as a centerpiece of the health care system. He finds that absent a mandate, only about half the uninsured would obtain coverage, although this reform compares favorably to tax credits in that it covers many more people at only slightly higher cost. With a mandate, nearly everyone is covered at relatively low cost to the government per newly insured person, but this is achieved by forcing individuals to spend money on insurance that they would prefer to spend elsewhere. Finally, he considers funding health insurance expansions by removing the tax subsidy to health insurance. He finds that this would generate more than enough revenue, but would be a "net loser" for more than half of families.

Gruber closes by noting that his study largely ignores the other major problem plaguing the U.S. health care system, rapidly rising costs. He notes that most cost control measures currently being discussed, like electronic medical records, will lower costs very little if at all. Effective control of health care costs will involve denying coverage for care that does little for health but that consumers now want. He concludes "until we are understand more fully which health care spending is justified and which is not, we are not prepared to take on the American public on cost control. The fundamental insight of this round of reform is therefore not to hold the attainable goal (universal coverage) hostage to the (currently) unattainable goal, fundamental health care cost control."

Gruber gratefully acknowledges funding from the Kaiser Family Foundation.