Now Updated with proofreading!
Political debates about taxes and tax rates in the U.S. often focus on the rich and claims about the incentive effects of different tax rates. Rarely mentioned these days are the poor. Indeed, the Republican demands in the last few years that tax rates should be cut for the high-income rich are primarily about claims of incentive effects. And, no, high-income rich isn’t redundant; it’s precise. There are at least two types of “rich”: High-income rich, which pay income taxes, and the high-asset, low income rich which pay much less. (I suppose there’s another type, the spiritually-rich, but that’s the domain of some other blog.) The claim is made that if tax rates are raised or raised too high, then that provides a disincentive to work and the rich will not work as much. It is often asserted that this is simple micro-economics–that people respond to incentives–and should be obvious.
There’s a problem with the claim, though. Actually there are two problems. First, there’s very little empirical evidence of higher tax rates on the highest end, the rich, actually reducing their efforts to earn income. Indeed, numerous studies (I don’t have time at the moment to look for citations) have found in “natural experiments” that the rich really don’t respond to higher tax rates by working less and earning less. Several studies have found that in situations where a large metropolis straddles two or more states, such as NYC, and different neighboring states changed their tax rates on the rich, the rich did not in fact do what they threatened or what would appear “rational”: move to the lower tax state in the same metro area. There’s also substantial longitudinal evidence in the U.S. and other countries that shows when tax rates on the rich were a lot higher, such as in the 60’s and 70’s, effort and incomes were no less than in more recent times.
The other problem is the whole idea that the “rational” response to higher tax rates is to reduce one’s effort and income actually doesn’t hold microeconomic water. It’s actually irrational to respond that way unless the marginal tax rate is truly so high that it approaches or exceeds 100%. The average tax rate, the percents you normally hear on TV, isn’t what affects incentives. Instead, it’s the marginal rate, or how much an extra dollar earned is taxed, that changes how we behave. Even then, a raising a marginal tax rate might reduce the incentive or attractiveness of additional effort and gross income, but won’t become a true dis-incentive until it becomes very, very high. An example: Let’s suppose someone makes $1,000,000 a year and is taxed $400,000. Such a person is said to pay a 40% average tax rate or effective tax rate. But averages and effective rates tell us nothing about incentives. Incentives deal with changes in behavior at the margins – the incremental changes. If micro is clear about one thing and has been since the 1870’s, it’s that decisions and changes in behavior depend on changes in marginal costs and marginal benefits. What matters is the taxes on the marginal, the incremental, change in income. What matters is the marginal tax rate. The only reliable way to figure the marginal tax rate is to compare two different amounts of income, preferably with only a small difference between them, the taxes paid and the after-tax-income that results. What people work for is to get after-tax, spendable income.
So let’s continue the example. Suppose the existing tax code, with all of its exemptions, deductions, rates, credits, etc, says that $1,000,000 income pays $400,000, but that $1,010,000 income pays $405,000 in taxes, then we have an increase in income of $10,000 of which $5,000 is used to pay the additional taxes. After-tax income rises from $600,000 to $605,000, leaving a net increase in after-tax income of $5,000. This means we have a marginal tax rate of 50%. There be a disincentive effect only if opportunity cost (usually leisure) of the additional time/effort needed to generate the higher income is judged to be greater than the $5,000 increase in after-tax income. Empirical evidence indicates that is not likely. On the other hand, if the marginal tax rate were 100%, it would mean that $1,010,000 in income requires $410,000 in taxes. At a 100% marginal tax rate none of the additional effort results in more after-tax spendable income, so obviously it doesn’t make sense to exert the extra effort.
So what are the marginal tax rates for the highest brackets in the U.S.? Even if all income comes from wages, the highest marginal rate is now around 38%. Even if you include state or city income taxes, the marginal rates faced by the rich aren’t greater than 50% even in the most onerous tax-happy states. For the really rich, most income comes from capital gains and not wages. Capital gains have a much lower marginal tax rate of close to 23-24% (including the 2013 Medicare tax on capital gains). Evidence is pretty clear that such marginal rates do not provide a disincentive to additional work.
But, now I want to return to the poor. We often assume that the poor don’t pay much in taxes. That’s true in total since they’re poor–there’s not much there to tax. But, marginal tax rates still exist. And they affect incentives. In fact, it’s the working poor that face the most serious disincentives to work and earn income. Our tax code is actually set up to make it rational for the poor to not try to earn more income! As University of Southern Cal Professor Edward McCaffery notes on CNN.com,
…some of the working poor face marginal tax rates “approaching 90% as they lose benefits attempting to better themselves.”
Readers were incredulous, asking how it could be that in a nation with a top federal income tax rate of 39.6% on individuals making more than $400,000 a year, anyone could face a 90% rate.
It is true. Marginal tax rates, especially for those below the top rate brackets, are chaotic, confusing, and all over the map.
As a result, some of the working poor face extremely high rates on their next dollar earned. Tax scholars and economists have long known this. Dan Shaviro of NYU published a study in 1999 showing marginal tax rates above 100% on the working poor; specifically, he illustrated that a single parent earning $10,000 would lose over $2,500, after taxes, by earning another $15,000, pushing her income to $25,000.
Obviously, this is a policy failure. We want to support the working poor, but we want them to be able to increase their incomes, join the middle class, and leave dependency behind. Yet the way most welfare and aid to working poor programs are structured, a working poor person can find themselves in a situation where working additional hours or getting a modest raise in wage will actually result in less after-tax spendable money.
The problem is even worse, as Professor McCaffery points out. The tax code exerts a genuine disincentive to getting married or to staying married if you are among the working poor. Yet, we know that stable marriages and two-income households are often the key to escaping poverty for both the present and next generations .
It’s appropriate to talk about the incentive effects of tax rates. Incentive effects should be part of the thinking when writing the tax code, just as reasons for government revenue should be a part. But when we talk about incentive effects of tax rates, we must focus on the marginal rates and we really should be talking about the poor. Not the rich.