Tornado Alley (PROG)

[fuagf]

Using it to argue lower taxes today, debunked comprehensively.

August 10, 2010, 10:37 am 107 Comments

The Laffer Test (Somewhat Wonkish)

The renewal of claims that tax cuts pay for themselves has led to a revival of discussion about an old question: how high do taxes have to be before further increases actually reduce revenue?

So I thought it might be worth thinking about this question in terms of a simple model of labor supply. Think of an individual facing a marginal tax rate t; and think of the amount this individual produces as depending on effort, which in turn depends on how much of his or her income the individual gets to keep at the margin, i.e., 1-t. Then a little calculus will show that whether a tax hike increase raises or lowers revenue depends on whether the elasticity of effort with respect to earnings — the percentage change in effort from a 1 percent rise in 1-t, the after-tax return to effort — is less or more than (1-t)/t.

An example may make the point clear. Suppose that the top marginal tax rate is 20 percent, so that high earners get to keep 80 percent of what they make. Now raise the rate to 21 percent. This is a 5 (100*1/20) percent increase in the proportion of income collected in taxes; revenue will only fall if effort falls more than 5 percent. Meanwhile, the after tax return to effort falls 1.25 percent (100*1/80). So the elasticity of effort with respect to earnings would have to be more than 4 for revenue to fall.

At a marginal tax rate of 50 percent, the break point is much lower; just 1. And so on.

So what do we know about the elasticity of effort with respect to earnings? Well, history suggests that if anything it’s negative: real wages have trended up over time, but working hours have fallen. That’s not a paradox, because rising wages have an income as well as a substitution effect. When you get richer, you want in general to consume more of everything, and among the things you want to consume is leisure. Against this,you can buy more goodies with an additional hour of work; but the net effect can go either way.

Now, someone might come along and point out that higher taxes aren’t the same thing as lower wages, because those taxes are generally used to finance a more generous welfare state — and this can wash out the income effect. That is, if you impose taxes that bring incomes after tax back to what they were in, say, 1960, but use the revenue to finance generous retirement benefits and free medical care, you should not expect people to work as hard as they did in 1960. And that’s a good point when we’re talking about the effects of high taxes/high benefits for people in the lower part of the income distribution.

But it’s not very relevant to high earners, for whom welfare-state benefits are inevitably small compared with their overall incomes.

So the way I see it, even quite high marginal tax rates on high earners — even rates in, say, the 70 percent range that prevailed pre-Reagan — are unlikely to put us on the wrong side of the Laffer curve by discouraging effort. High earners won’t work much less; they might even work harder, because it takes more effort to make enough to buy that fourth home.

That doesn’t mean, however, that it’s OK to go back to Eisenhower-era 91 percent top marginal rates. The problem with super-high rates isn’t so much that they reduce incentives to work; it’s that they create huge incentives to avoid or evade.

But we’re nowhere near Laffer country now. In terms of taxes and revenue, up is up, down is down.

http://krugman.blogs.nytimes.com/2010/08/10/the-laffer-test-somewhat-wonkish/