Tax Subsidies and the Incentive to Work

Uwe E. Reinhardt is an economics professor at Princeton. He has some financial interests in the health care field.

Last week a brouhaha erupted over a passage in Appendix C of aCongressional Budget Office report, Budget and Economic Outlook 2014-24.

TODAY’S ECONOMIST

Perspectives from expert contributors.

In that appendix, “Labor Market Effects of the Affordable Care Act: Updated Estimates,” the agency reported its estimate that the Affordable Care Act “will reduce the total number of hours worked, on net, by about 1.5 percent to 2.0 percent during the period from 2017 to 2024, almost entirely because workers will choose to supply less labor – given the new taxes and other incentives they will face and the financial benefits some will receive.”

The agency estimated this reduction in hours worked as the “full-time-equivalent workers of about 2 million in 2017, rising to about 2.5 million in 2024.” The agency hastens to point out that this number does not represent jobs no longer offered by employers but, for the most part, the decision of employees not to work.

Opponents of the Affordable Care Act and many news reports quickly seized upon this estimate, characterizing it as “dropping a bomb” or having “nuked” Obamacare. Joseph Rago of The Wall Street Journal attributed this interpretation of the data to an exposé by my fellow Economix blogger Casey B. Mulligan.

Commentators supporting the Affordable Care Act pointed out thatthe pro-growth effect of the law’s lower health costs would swamp any antigrowth effects from a lower labor supply and that if some Americans decided to work less, given the incentives they face, they would yield available jobs to others willing to work but unable to find a job, which on balance would be a good thing.

It is worth reading Appendix C of the C.B.O. report to get a feel for what is really at stake here. In that appendix the agency explains, for example, that its “estimate that the A.C.A. will reduce employment reflects some of the inherent trade-offs involved in designing such legislation.” Further clarification was offered in a “Frequently Asked Questions” statement by the C.B.O.

The agency alludes there to the fact that any program offering means-tested public assistance to citizens will do much good but implicitly confront citizens with higher marginal tax rates that may induce these beneficiaries to work fewer hours or even retire. As Professor Mulligan properly notes, it is not the economist’s job to come to a judgment on this trade-off.

The clearest and easiest vehicle for explaining this trade-off is thenegative income tax. It was most famously proposed in this country by Milton Friedman, the late conservative economist and Nobel laureate, in his classic “Capitalism and Freedom” (1962). What ultimately happened to this idea is well told in a history of the concept by Jody T. Allen. It need not detain us here.

Friedman explained his rationale for and the modus operandi of his idea clearly in a 1968 video clip, during a conversation with the late William F. Buckley Jr.

In this scheme a, say, family of four with an earned income (call it Y) below a certain threshold income level (call it X) receives public financial assistance calculated as flat “tax rate” of “t” on the amount (Y – X). In other words, the tax – actually a negative tax or a subsidy – is

TAX = t(Y – X), when Y is less than X.

Because the TAX in this expression is negative when, as we assume, Y is less than X, economists would say that the family “pays” a negative income tax. In plainer in English it means it is receiving a public subsidy.

In the video clip, by the way, Friedman assumes a flat marginal income tax rate of t = 50 percent, a rate even higher than the marginal tax calculated by Professor Mulligan for the Affordable Care Act.

The graph below depicts Friedman’s idea, although I use different numbers, because the Consumer Price Index has increased by a factor of about 6.5 since 1968. In that graph, I assume that the minimum guaranteed income (the maximum cash assistance) for the family in question is M = $20,000 if it has no earned income or any other income at all, and I assume two different tax rates: t = 25 percent and t = 50 percent.

The red line in the graph assumes a tax rate of t = 50 percent and shows how quickly federal cash assistance melts away as the family’s earned income rises, which means that the family faces a strong disincentive for work to earn income. The threshold income below which a subsidy is received here is X = M/t = $20,000/0.5 = $40,000 only.

The blue line assumes a tax rate of only t = 25 percent, driving the threshold income to X = 20,000/0.25 = $80,000. That scheme confronts the family with less of an incentive not to work and earn income; but the price for that is a much higher federal assistance budget.

The sizes of the total federal assistance budget implied by these two alternative tax rates cannot be read off the graph, because households are not evenly distributed along the income range on the horizontal axis; more households will be in one income range than another. But whatever the distribution of households along the income scale may be, it is clear that, for a given minimum guaranteed income (M), the size of the federal assistance budget rises when the tax rate decreases. At the 50 percent tax rate, for example, no household with an income above $40,000 and below $80,000 gets any assistance. At the 25 percent rate they do.

In designing such a scheme, lawmakers thus must trade off among (a) the generosity toward the poorest members of society (M), (b) the disincentive to work baked into the scheme and (c) the total federal budget for cash assistance in the form of negative taxes. Whatever is put in place will naturally be a compromise.

Although a negative income tax scheme is the clearest way to illustrate the trade-off to which the Congressional Budget Office alluded, that trade-off is inherent in any means-tested federal assistance program intended for the poor.

Details on the nature of that means-testing built into the Affordable Care Act, for example, can be found on the Kaiser Family Foundation’s convenient website. If readers wish to get a feel for how quickly the amount of federal assistance under the program melts away with higher earned income in their own situation, I recommend playing around with the foundation’s handy premium-subsidy calculator.

Finally, to return to the estimated reduction in work effort reported in Appendix C of the Congressional Budget Office report, the following should be kept in mind: it is merely an estimate.

Calculating the marginal tax rate inherent in the Affordable Care Act, which can be done with reasonable accuracy, is only half the story.

The rest depends on what assumption one makes concerning the likely behavioral response of employed or unemployed workers to changes in marginal income-tax rates.

There is not just one response coefficient that every economist would use. Usually the literature offers a fairly wide range of previously estimated response coefficients. For that reason, different economists can easily come up with different estimates of reduced worker hours.

Taking into account both estimated changes in demand and supply, Professor Mulligan, for example, projects a reduction in the total number of hours worked under the Affordable Care Act as equivalent to 3 percent relative to the number of hours likely to have been worked in the absence of the law. The C.B.O. has that number at 1.5 to 2 percent over the coming decade. Other economists might arrive at an even lower estimate.

That, alas, is the nature of economic analysis.

via:http://economix.blogs.nytimes.com/2014/02/14/tax-subsidies-and-the-incentive-to-work/