Guest post written by Uwe Reinhardt Mr. Reinhardt is the James Madison Professor of Political Economy at the Woodrow Wilson School at Princeton University.
In his comment on my post responding to Sally Pipes' original post on the relative costs of employment- and government-sponsored health insurance, fellow economist Chris Conover alerts us to the idea that raising taxes causes a so-called "deadweight loss." It is a proposition well rooted in standard economic theory and has ample empirical support. I found it helpful that Chris brought up this point and thank him for it. But it is not relevant in this case.
The economic theory of the deadweight loss is straightforward. If the basis of taxation is a particular transaction—e.g., a sale, or income earned—then taxing it triggers incentives to do fewer of these transactions. The potential value that these lost transactions could have added to society is the deadweight loss of taxation. A glaring example of the deadweight loss of a tax was a 10% tax imposed on luxury boats costing more than $100,000 imposed in 1990. It is said to have deeply hurt the U.S. boat building industry, and therefore was subsequently repealed.
The actual magnitude of the deadweight loss triggered by a tax depends, of course, on the particular context. Some transactions are less sensitive to the disincentive engendered by taxation than are others. For example, income earned in 9 to 5 jobs is less sensitive to the disincentive inherent in taxes in comparison to hourly workers, who can change the number of hours they work in response to a tax on hourly income. Not surprising, then, estimates of the magnitude of the deadweight loss as a percentage of the tax raised have varied enormously, ranging from 10% to as much or more than 100%.
Furthermore, one must be mindful of what kind of tax is involved. Economists talk of Pigovian taxes—named after the British economist Arthur Cecil Pigou (1977-1959) who wrote extensively on it—specifically designed to reduce inefficiencies in the economy. Taxing the output of a producer who pollutes the environment is an example. In that case, taxes trigger a welfare gain, not deadweight loss, because we explicitly want the polluting activity reduced.
Similarly, it behooves one to inquire into the use to which a particular tax is put. For example, suppose a sales tax is imposed on alcohol or tobacco products and the proceeds are used to immunize children. What is the net, overall deadweight loss of this tax-and-transfer policy here?
Leaving aside these fine points, Chris Conover is entirely correct in asserting that in some context the deadweight loss from taxation should be added to the cost of tax-financed activity. That would be proper, for example, if the choice was between 100% private financing and 100% tax financing of an activity, and both options were feasible on economic, ethical and political grounds. Chris suggests a surcharge of 25% for this purpose. Many economists would use an even higher figure.
My question, though, is whether the choice we have been discussing in these posts is such a context.
Consider again the Medicaid population, described in the first two charts of my earlier post. Would it be economically and ethically feasible to have employers sponsor health insurance for this mixed, high-risk population? Surely not.
By definition, the pauperized elderly population eligible for dual Medicare-Medicaid coverage—often in nursing homes—would not be candidates for employment. Nor would the blind and disabled, except in very rare circumstances. Nor would be the children covered by Medicaid. Nor would be their moms. If they were employed by an employer sponsoring health insurance, they would not be in Medicaid in the first place. If they are working and are in Medicaid, it must mean that their employer pays them very low wages and does not sponsor insurance, as many employers with predominantly low-wage workers do not.
So employment sponsored coverage for the Medicaid population is not a realistic option, unless employers—heaven forefend—were mandated somehow to take care of that population. How could that be done, and should it be done?
But what about moving the entire Medicaid population out of government-run insurance into private insurance, without employers as intermediaries? Would that not get rid of the dreaded deadweight loss?
How could it? The very reason Medicaid beneficiaries are in Medicaid is that their household budgets are too small to be able to cover actuarially fair or even community-rated private health insurance premiums.
So if we wanted to move this clientele into private insurance—as well over half of Medicaid beneficiaries actually already are—then someone other than the beneficiaries themselves would have to pay private insurers the premiums for this clientele. That "someone else" is, of course, the government. And to be able to make these payments, government would have to raise taxes, which, of course, would trigger—you guessed it—the deadweight loss of taxation.
A maxim in economic analysis is that when a particular cost is incurred under all options being considered, then that cost item can be ignored in comparative analysis. Which explains why I ignored the deadweight loss, even though as an economist I was aware of it. Roughly the same amount of deadweight loss would be incurred whoever performs the purchasing function on behalf of Medicaid beneficiaries: government itself or private managed care companies.
So, in effect, Sally Pipes was right in ignoring the deadweight loss, although, judging by the content of her column that I criticized, I very much doubt that she was even aware of this subtle point of economic analysis.
0 comments:
Post a Comment