The decisions of small firms to offer health insurance are fairly sensitive to the tax subsidization of insurance prices. Moreover, among firms that offer health insurance, the level of insurance spending is very sensitive to tax subsidies.
In 1987, 14.8 percent of non-elderly Americans were without health insurance. By 1997, 18 percent of the non-elderly population did not have health insurance, a 25 percent increase. Despite a recent decrease in that percentage, 40 million non-elderly Americans still lack health insurance. A common prescription for reducing the number of uninsured is to introduce new tax subsidies for health insurance costs. Yet federal and state governments already spend over $100 billion annually in subsidies to health insurance, by excluding employer spending on health insurance from taxable employee wages.
In Taxes and Health Insurance (NBER Working Paper No. 8657), NBER Research Associate Jonathan Gruber investigates what we have learned about the effects of the existing tax subsidy to health insurance, and what it implies for future tax policy towards the uninsured. He documents that the primary reason employees are uninsured is that they are not offered insurance in the workplace. Small and low wage firms in particular are unlikely to offer insurance. Take-up of insurance when offered, in contrast, is quite high in all types of firms.
Gruber notes that there are a variety of avenues through which tax policies can affect insurance status, and he reviews the evidence on responsiveness along these dimensions. He concludes that the decisions of small firms to offer health insurance are fairly sensitive to the tax subsidization of insurance prices. Moreover, among firms that offer health insurance, the level of insurance spending is very sensitive to tax subsidies; this partly reflects a shifting of costs to employees as employer spending is less subsidized.
On the other hand, worker decisions to take-up insurance do not appear to be price sensitive. There does seem to be some scope for substitution between private and public coverage as the relative subsidies of these forms of insurance change, as well as scope for switching coverage across spouses. But there remain important unanswered questions, particularly about the sensitivity of the uninsured to subsidies to the price of insurance, and how the insurance market would react to widespread subsidies.
Gruber draws on this evidence to discuss its policy implications. His findings suggest that removing or reducing the existing tax subsidy to employer-provided health insurance could lead to significant increases in the number of uninsured. For example, completely removing the tax subsidy to health insurance would lead to 22 million workers losing their employer-provided insurance coverage. Likewise, subsidies to small and low wage employers to offer insurance could significantly increase insurance coverage.
But subsidies to employee spending on insurance are unlikely to be an effective route, since take-up among eligible employees is already so high and their take-up decisions do not appear to be price sensitive. Subsidies to non-group insurance purchase would allow significant numbers of uninsured to purchase insurance, but would also subsidize the existing spending by those holding non-group insurance, and would potentially erode the group market. Given the high costs of insurance in the non-group market, subsidizing insurance in that market may be a very costly means of reducing the number of uninsured.