Do Tax Cuts for the Wealthy Create Jobs?

Over at The Economist's Democracy in America blog, Matt Steinglass is pondering the differential stimulus effect of tax cuts for the wealthy, and tax cuts for everyone else:

But what about the stimulative effect of the upper-income, as opposed to lower-income, tax cuts? How big will it be? Mike Konczal points us to research on this subjectby the Congressional Budget Office. In September, the CBO found that those $100 billion in tax cuts on income above $250,000 would reduce unemployment in 2011 and 2012 by...somewhere between 0.1% and nothing at all.

The CBO's Doug Elmendorf explains: "(T)he economic impact per dollar of revenue reduction from the full extension would be smaller than that from partial extension because a greater proportion of the tax savings from the full extension would go to relatively high income households, which tend to spend less of an increase in income than lower-income households do."

Meanwhile, today the Center for American Progress's Michael Linden and Michael Ettlinger take the principle a step further and ask how many jobs would be created by letting the Bush tax cuts for income above $250,000 expire, and then using that revenue to fund a bigger payroll-tax cut for lower earners. Their answer, based on a combination of multipliers from the CBO and from economist Mark Zandi: an extra 500,000 jobs.

The implication is that tax cuts for the rich aren't very stimulative, while tax cuts for the middle class are. But this is not quite how I would interpret that chart.

According to the CBO, the low estimate is that temporarily extending the Bush tax cuts for the middle class aren't very stimulative, while extending the tax cuts for those who make over $250,000 doesn't show up at all; in the high estimate, extending the tax cuts for incomes above $250,000 reduces employment by about a tenth of one percent, while extending the tax cuts for incomes below $250,000 lowers it by about half a percent.

It's true that there is some differential effect. But most of that figure is not driven by the fact that wealthy save their tax cuts, while the middle class spend them; it's driven by the fact that the tax cuts for incomes below $250,000 are much larger, in terms of the federal budget, than tax cuts for higher incomes.

According to CNN, the two-year cost of the tax cuts for high earners will be about $75 billion, while the estimated cost of the cuts for incomes below $250,000 is about $310 billion, or four times larger. In other words, in the "high estimate" world, a $300 billion stimulus composed only of tax cuts on income above $250,000 would reduce unemployment by 40 basis points, while one of a similar size composed entirely of tax cuts on income below that level would lower it by about 50 basis points. That isn't nothing, but it isn't a particularly large difference in effectiveness, either.

Presumably this effect is also operating in the "low estimate" column. If tax cuts for high earners were exactly as effective as tax cuts for lower earners, the added unemployment reduction from the tax cuts on high earners would only be .05%. If they're even very marginally less effective--say it lowers unemployment by .044%--it will disappear from the chart. That's not because there's "no effect at all"; rather, it's a rounding error.

It's important to note that in the "low estimate" world used by the CBO, all sorts of stimulus--not just tax breaks for the wealthy--is not very effective. So while you can use this scenario to make arguments against the tax breaks for high earners, you're also weakening the argument for other forms of stimulus you make like better. Meanwhile, in a world where the middle class tax cuts created 500,000 jobs, the tax cuts on the wealthy probably create something in the neighborhood of 100,000 jobs--not a huge drop in the unemployment rate, but probably rather important to those 100,000 people who are currently out of work.

Myself, I don't think we can afford to extend any of it. We've got a massive fiscal problem ahead of us, and extending all these tax cuts until 2012 will simply mean that both parties will be vying with each other to bust the budget ahead of the election. But it seems ludicrous to me to claim that tax cuts for the middle class are affordable, excellent stimulus, while the much smaller tax cuts for the wealthy cost too much relative to the stimulus they deliver. The differences are really pretty marginal, and not particularly well measured.