DURHAM — In the abstract, most people view taxes as being somewhere between a necessary evil and downright evil. However, taxes are not abstract.
There are two primary public policy motivations to levy a tax. The first is to collect money and redistribute it for purposes that markets won't achieve. For example, we would have a smaller military if it were funded by people's donations. Using taxes to fund the military (or anything else) is premised on a claim that the market won't provide the optimal spending level.
The second motivation for a tax is to decrease a particular behavior. Cigarette taxes are designed to reduce the amount that an individual smokes by raising the price of cigarettes. Though some continue to smoke, the amount smoked is less than it would be if cigarettes were cheaper.
The proposed tax on high-cost health-insurance policies is the second type of tax: The behavior this tax seeks to reduce is being overinsured. The tax is designed to encourage persons to have less insurance and therefore avoid the tax. Less insurance would increase out-of-pocket costs, reducing health-care use, slowing premium growth.
For a family policy, the total amount of premiums paid by employers and employees above $21,000 ($8,000 for individuals) would be taxed at 40 percent. If a family had a policy worth $25,000, then a tax bill of $1,600 (40 percent of $4,000) would be levied on the health insurer. Insurers and employers will pass this cost on to employees.
This tax is a central part of the Senate reform (Baucus bill) approach, but it is not in favor in the House.
The CBO estimates that levying the insurance tax would reduce federal spending by $201 billion over 10 years. Most of this -- $142 billion -- is estimated to occur as individuals choose lower-cost plans to avoid the insurance tax. As employers reduce the premiums they pay for employees, wages will rise because premiums are simply a part of total compensation. Wages are taxable whereas premiums are not, resulting in an increase in income taxes paid. The trigger level of the tax would be indexed, but would grow much slower than health-care costs have for the past 25 years.
The tax on high-cost insurance policies is a de facto capping of the amount of tax-free income that may be received via employer-paid health insurance. I would prefer a more straightforward capping of the tax exclusion of employer-paid premiums because it would initiate a cultural conversation about health-care costs that more directly focused on the end user of care: patients. The high-cost tax has found favor because it would be politically easier to sell a tax on insurance companies than it would be to end a popular tax loophole that benefits those with good workplace health insurance. But they both work in more or less the same way.
The high-cost insurance tax is the part of the Baucus bill that would most predictably slow health-care cost inflation. Further, it has a built-in self-correcting mechanism: If costs continued to rise much faster than inflation, more policies would be subject to the tax, making it a more responsible financing mechanism down the road. It is true that all high-cost plans are not of the "Cadillac" variety. Some are simply in high-cost areas, but this tax would differentially affect the most expensive outlier areas, which is needed.
The income tax increase that is favored in House reform bills has three primary problems. First, it does nothing to slow cost inflation. Second, it adds a new source of financing to a system whose main problem is unsustainability due to cost. Third, because wages are growing much slower than health-care inflation, income taxes would be unable to keep pace with the cost of subsidies down the road. On balance, the high-cost insurance tax is a more responsible way to finance reform than an income tax increase.
Donald H. Taylor Jr. is an assistant professor of public policy at Duke's Sanford School of Public Policy. His blog http://www.donaldhtaylorjr.blogspot.com is available for discussion of this article, one in a weekly series exploring the issue.