When taxes are high, raising them hurts. When they’re low, not so much.

October 25, 2012

Do tax cuts help the economy grow?


Most economists would say yes. Certainly during a recession, tax cuts — especially those targeting low-income people who are likely to spend the extra money — have a strong stimulative effect. Mark Zandi has argued that payroll tax cuts generate $1.24 of economic activity for every $1 they cost. The effects of cuts to consumption taxes are a bit smaller, and the IMF estimates that you get anywhere from $0.90 to $1.70 in growth per dollar in tax cuts or spending increases.

Even outside of recessions, UC Berkeley's Christina Romer and David Romer have looked at post-World War II tax changes in the United States and found that increasing taxes by 1 percent of GDP hurts growth by about 3 percent. That implies that tax cuts would, conversely, improve growth considerably. But the Romers also found no such relationship between the World Wars.

So which is it? How could tax cuts at one point have a big effect and at another period have no effect? A new study by Stanford's Nir Jaimovich and Northwestern's Sergio Rebelo takes a stab at answering that. They note that usually economists assume that the effects of tax cuts or hikes are linear -- that is, a 1 percent of GDP increase or decrease in taxes has the same effect, no matter where taxes were before that. Raising taxes from 19 percent to 20 percent of GDP would have the same effect as raising them from 49 to 50 percent, in other words.

But Jaimovich and Rebelo argue that taxes don't really work like that. A tax hike when taxes are already high, they figure, ought to do more damage than one when they're low. When taxes are higher, people are likelier to be taking them into consideration when weighing whether to work more or less, and so people will be more responsive to changes than when they're low and no one is really factoring them into their decision making.

You're likelier to care about a change in the price of a house than a change in the price of a soda, just because you're likelier to be paying attention to the price of a house, period. Same logic for taxes. This would explain why tax increases and cuts in the 1920s and 1930s, when taxes were very low, didn't have much effect while increases and cuts after 1945, when the top marginal rate reached 95 percent and stayed at or above 70 percent until 1981, had a bigger effect.

Jaimovich and Rebelo built a model to see how taxes affect growth if they're right about the effects of taxes not being linear. They evaluated a simplified version of the current tax code, with two tax brackets: 15 and 35 percent. They then saw what happens to growth if you increase or decrease the top rate.

The main effect of higher rates, they argue, is on the willingness of people to start companies. All else being equal, you're less likely to want to start a company if you're going to be facing really high rates if successful. So Jaimovich and Rebelo's results differ depending on whether everyone in an economy has the same level of entrepreneurial skill or if their abilities in that area vary. The green line shows changes if everyone is equally able to successfully start a company, and the blue line shows changes if (as seems likely) some people are better at it than others:

Source: Jaimovich and Rebelo

The x axis is how much higher the top bracket is from the bottom bracket; the y axis is the growth rate. So if everyone is equally able to start a company, a hike from, say, 35 percent to 45 percent has some effect, though not a big one. But if some people are better entrepreneurs than others, then the effect is basically zilch, and until you get into the 90 percent range you don't really see any effect of tax increases at all.

The results are even more dramatic if you assume that people don't know how good they'd be at starting companies. In that case, there's literally no effect until rates hit 75 percent, since everyone who could start a company views it as in his or her best interest to do so:

Source: Jaimovich and Rebelo

So the more realistic Jaimovich and Rebelo's model gets, the less likely it is to see any effects on growth of increasing taxes for high earners, and certainly if the increases are as small as the ones Obama has been proposing. That's in line with a recent Congressional Research Service study that found zero effect of changes in top tax rates on growth.

There's definitely a point at which high tax rates for the rich hurt growth. But we're nowhere near there yet.

Show Comments
Most Read Business
Next Story
Sarah Kliff · October 25, 2012