When it comes to taxes and GDP, there are no simple conclusions. Here are three reasons why it's really hard to say how tax changes affect the economy

615 tax 1040.jpg

White House

Mitt Romney claims that cutting taxes will supercharge the economy. Barack Obama says that raising taxes on the richest earners won't dampen the recovery. Who's right? Maybe both. Maybe neither. And the unfortunate truth is that we might never know.

Last week, the Congressional Research Service published a new study that looked at whether cutting taxes for high-earners slowed down the economy. Their answer: Not really. Cutting the top marginal rates on earnings and investment had little effect on economic growth.

But the literature on taxes and growth in the U.S. is excruciatingly complex. While many studies have found that recent tax cuts did little to promote growth, others concluded that higher taxes discourage work and investment and slow the economy down. Rather than pick a winner, let's eat our humble pie and remind ourselves that black-and-white answers are hard to come by. Here are three reasons why it's so difficult to say for sure what tax cuts and tax increases really do to economic growth.

(1) It depends on why Congress is changing taxes in the first place.

A recent paper by Christina Romer and David Romer reached the stark conclusion that an "exogenous" tax increase of 1 percent of GDP would lower real GDP considerably -- by 2 to 3 percent. But the word "exogenous" is playing a big role in the sentence.

Romer and Romer posit that significant tax changes since World War II were for one of four purposes: (1) to counteract a recession or inflationary period; (2) to match changes in government spending; (3) to pay down a deficit; and (4) to promote long-run growth. It's hard to measure the effect of tax changes in the first two scenarios because the effect of recessions, inflation, or big changes in government spending are often greater than the effect of the tax change itself, the authors write. It is only "exogenous" tax increases designed to reduce the deficit or promote long-term growth that the Romer's predict will slow down the economy.

But even that conclusion gets an important caveat. Late in the paper, Romer and Romer acknowledge that deficit-driven tax increases might actually grow GDP over time by reducing long-term interest rates or giving businesses confidence in the future of the economy.

(2) It depends on what the Federal Reserve does.

One reason why it's so hard to say how Congress's tax laws affect the economy is that there's another Washington body that might have an even stronger impact on short-term economic growth. The Federal Reserve. If the Fed thinks the economy is getting too hot, it can raise interest rates. If it thinks the economy is too chill, it can lower interest rates. That makes it awfully difficult to determine how taxes  "The short-run macroeconomic consequences of tax changes depend on how the Federal Reserve changes monetary policy in response to the tax change," Harvard economist Martin Feldstein wrote in a 2008 review of taxes and behavior. Even if some tax changes have longer-term effects on businesses and overall industries, the short-term is awfully noisy.

(3) It depends on what's going on in the rest of the world.

You might think of taxes as a thermostat for the economy -- that raising or lowering rates will have a clear and foreseeable effect on growth. In reality, fixing taxes is more like setting a thermostat on an outdoor patio and pretending you can control the weather. You can have a target temperature, but unpredictable conditions tend to mess with your plans.

There is a lot more to an economy than taxes. We've acknowledged that some tax changes are hard to measure because they coincide with big swings in government spending or important action taken by the Federal Reserve. But there's also ... everything else. China's growth, Brazil's currency, Europe's bailout fund, Pennsylvania's natural gas production. These things don't have much to do with the top marginal tax rate being 35% or 40%. As you look back at the huge swings in GDP over the last 15 years, it's fairly clear that little of it can be pinned to the smaller changes in our tax code. The Clinton tax increases didn't create the '90s Internet-fueled productivity boom (although there's an argument that the investment tax changes encouraged the Nasdaq bubble). The Bush tax cuts didn't create the housing bubble. Obama's stimulus tax credits have probably had very little to do with the pace and nature of our recovery.

This is not to say that taxes don't matter. They do matter. But when a campaign tells it can predict the future under its tax plan, there are good reasons to doubt their magical numbers.