Tweet
That Wishful Thinking About Tax Rates
By CHRISTINA D. ROMER
Published: March 17, 2012
By CHRISTINA D. ROMER
Published: March 17, 2012
AT least since Calvin Coolidge, politicians have trumpeted the supply-side benefits of cutting marginal income tax rates. Lower rates will unleash economic growth and the cuts will largely pay for themselves — or so it’s often said. Yet careful studies find little evidence of such effects. Perhaps it’s time to reform tax policy based on facts, not worn-out assumptions.
A family’s marginal tax rate is what its members pay to the government if they earn another dollar. If the government takes a smaller chunk of that dollar, a family has more incentive to earn it. Workers may choose to work additional hours, or a stay-at-home spouse may decide to work outside the home. Likewise, entrepreneurs may invest in a new enterprise or expand an existing one. Lower marginal rates also reduce people’s incentives to shield income from taxes, through legal and illegal means.
The main question is whether these incentive effects are large. If they are, cutting marginal rates could cause a sustained surge of hard work and entrepreneurial activity — and thus reported income. This idea was the essence of President Ronald Reagan’s theory of supply-side economics, and his justification for large, permanent tax cuts in the early 1980s. Mitt Romney, now seeking the Republican nomination for president, cited a similar argument when he proposed cutting all income tax rates 20 percent.
If the incentive effects are small, however, the situation is very different. Cutting taxes would still raise output for a while by putting more money in people’s pockets, and so increasing their spending — a temporary demand-side effect. But lower marginal rates wouldn’t greatly raise output over the long haul through the supply side.
History shows that marginal federal income tax rates have varied widely. Since World War II, the top rate has ranged from less than 30 percent (at the end of the Reagan presidency) to more than 90 percent (throughout the Eisenhower years). The 1964 Kennedy-Johnson tax cut significantly reduced the typical marginal rate paid by American families, but rates rose greatly over the next 15 years as inflation pushed people into higher tax brackets. Rates fell sharply under President Reagan, rose under President Bill Clinton and fell again under President George W. Bush.
If you can find a consistent relationship between these fluctuations and sustained economic performance, you’re more creative than I am. Growth was indeed slower in the 1970s than in the ’60s, and tax rates were higher in the ’70s. But growth was stronger in the 1990s than in the 2000s, despite noticeably higher rates in the ’90s.
Of course, many factors affect the economy, so a lack of correlation doesn’t prove that marginal-rate changes have little impact. That’s why economists have devoted thousands of pages in journals to testing the effects more scientifically.
ONE standard approach is to look for natural experiments in the tax code. Often, a law changes the marginal rate for one group and not others. For example, the 2001 Bush tax cut lowered marginal rates sharply for married couples with taxable incomes of around $50,000, but did little to rates for couples earning slightly less. Using household survey data, economists can compare the behavior of taxpayers whose rates did and didn’t change.
A useful summary measure of such changes’ supply-side effects is the sensitivity of reported income to marginal rates. If people work and invest more in response to tax cuts, their reported income will rise when marginal rates fall. True supply-siders believe that this sensitivity is well over a value of 1, implying that cuts in marginal rates raise reported income enough that government tax revenues nevertheless rise. But a critical review of several natural-experiment studies concluded that the best available estimates of this sensitivity range from 0.12 to 0.40. The midpoint of the range, 0.25, implies that if the marginal tax rate for high earners decreased from its current level of 35 percent to 28 percent (which Mr. Romney proposes), reported income would rise by just 2 1/2 percent.
In a new study, David Romer and I found that changes in marginal rates in the 1920s and ’30s had even smaller effects. (Mr. Romer is my husband and a colleague at Berkeley.) The rate shifts in that era make those after World War II look tame, and varied greatly across income groups. The Revenue Act of 1935, for example, raised marginal rates on the very highest earners to 79 percent from 63 percent, but barely raised rates at all for those below the top one-fiftieth of one percent of households. (Now that was class warfare!)
We found that an increase in marginal rates on an income group leads to a decrease in its reported taxable income relative to other groups. Indeed, because the variation is so large, the effect can be pinned down much more precisely than in most postwar studies. But the estimated impact is very small — almost at the bottom of the postwar studies’ range. One likely reason is that the tax system between the two world wars was very simple — all the instructions and tax forms for the personal income tax fit on just six pages. As a result, there were few legal methods of shielding income.
Where does this leave us? I can’t say marginal rates don’t matter at all. They have some impact on reported income, and it’s possible they have other effects through subtle channels not captured in the studies I’ve described. But the strong conclusion from available evidence is that their effects are small. This means policy makers should spend a lot less time worrying about the incentive effects of marginal rates and a lot more worrying about other tax issues.
Most obviously, the federal budget is on a collision course with reality. Reining in the long-run deficit will have to involve slowing the growth rate of spending. But unless we choose to gut Medicare and Medicaid, additional tax revenue will be needed. This essential truth is the No. 1 factor that should be driving tax policy. And anyone who tells you that the way to raise revenue is to cut marginal tax rates is arguing from ideology, not solid evidence.
Many people have proposed raising revenue by cutting back on deductions and loopholes — so-called tax expenditures. Indeed, all else being equal, such changes are preferable to raising marginal rates. After all, higher rates do have some disincentive effects. Eliminating special tax provisions would also simplify tax preparation, and give people less incentive to try to game the system.
But if moderate increases in marginal rates wouldn’t much affect behavior, a mix of rate increases and cuts in tax expenditures might be a sensible path. Some tax expenditures, like the favorable tax treatment of employer-provided health insurance, may have worrisome effects (by encouraging overly generous plans) and so should be trimmed. But others, like deductibility of charitable giving, may be worth keeping.
Finally, income inequality has surged in recent decades. Raising marginal rates on the wealthy is a straightforward, effective way to counter this trend, while helping to solve our looming deficit problem. Given the strong evidence that the incentive effects of marginal rates are small, opponents of such a move will need a new argument. Invoking the myth of terrible supply-side consequences just won’t cut it.
A family’s marginal tax rate is what its members pay to the government if they earn another dollar. If the government takes a smaller chunk of that dollar, a family has more incentive to earn it. Workers may choose to work additional hours, or a stay-at-home spouse may decide to work outside the home. Likewise, entrepreneurs may invest in a new enterprise or expand an existing one. Lower marginal rates also reduce people’s incentives to shield income from taxes, through legal and illegal means.
The main question is whether these incentive effects are large. If they are, cutting marginal rates could cause a sustained surge of hard work and entrepreneurial activity — and thus reported income. This idea was the essence of President Ronald Reagan’s theory of supply-side economics, and his justification for large, permanent tax cuts in the early 1980s. Mitt Romney, now seeking the Republican nomination for president, cited a similar argument when he proposed cutting all income tax rates 20 percent.
If the incentive effects are small, however, the situation is very different. Cutting taxes would still raise output for a while by putting more money in people’s pockets, and so increasing their spending — a temporary demand-side effect. But lower marginal rates wouldn’t greatly raise output over the long haul through the supply side.
History shows that marginal federal income tax rates have varied widely. Since World War II, the top rate has ranged from less than 30 percent (at the end of the Reagan presidency) to more than 90 percent (throughout the Eisenhower years). The 1964 Kennedy-Johnson tax cut significantly reduced the typical marginal rate paid by American families, but rates rose greatly over the next 15 years as inflation pushed people into higher tax brackets. Rates fell sharply under President Reagan, rose under President Bill Clinton and fell again under President George W. Bush.
If you can find a consistent relationship between these fluctuations and sustained economic performance, you’re more creative than I am. Growth was indeed slower in the 1970s than in the ’60s, and tax rates were higher in the ’70s. But growth was stronger in the 1990s than in the 2000s, despite noticeably higher rates in the ’90s.
Of course, many factors affect the economy, so a lack of correlation doesn’t prove that marginal-rate changes have little impact. That’s why economists have devoted thousands of pages in journals to testing the effects more scientifically.
ONE standard approach is to look for natural experiments in the tax code. Often, a law changes the marginal rate for one group and not others. For example, the 2001 Bush tax cut lowered marginal rates sharply for married couples with taxable incomes of around $50,000, but did little to rates for couples earning slightly less. Using household survey data, economists can compare the behavior of taxpayers whose rates did and didn’t change.
A useful summary measure of such changes’ supply-side effects is the sensitivity of reported income to marginal rates. If people work and invest more in response to tax cuts, their reported income will rise when marginal rates fall. True supply-siders believe that this sensitivity is well over a value of 1, implying that cuts in marginal rates raise reported income enough that government tax revenues nevertheless rise. But a critical review of several natural-experiment studies concluded that the best available estimates of this sensitivity range from 0.12 to 0.40. The midpoint of the range, 0.25, implies that if the marginal tax rate for high earners decreased from its current level of 35 percent to 28 percent (which Mr. Romney proposes), reported income would rise by just 2 1/2 percent.
In a new study, David Romer and I found that changes in marginal rates in the 1920s and ’30s had even smaller effects. (Mr. Romer is my husband and a colleague at Berkeley.) The rate shifts in that era make those after World War II look tame, and varied greatly across income groups. The Revenue Act of 1935, for example, raised marginal rates on the very highest earners to 79 percent from 63 percent, but barely raised rates at all for those below the top one-fiftieth of one percent of households. (Now that was class warfare!)
We found that an increase in marginal rates on an income group leads to a decrease in its reported taxable income relative to other groups. Indeed, because the variation is so large, the effect can be pinned down much more precisely than in most postwar studies. But the estimated impact is very small — almost at the bottom of the postwar studies’ range. One likely reason is that the tax system between the two world wars was very simple — all the instructions and tax forms for the personal income tax fit on just six pages. As a result, there were few legal methods of shielding income.
Where does this leave us? I can’t say marginal rates don’t matter at all. They have some impact on reported income, and it’s possible they have other effects through subtle channels not captured in the studies I’ve described. But the strong conclusion from available evidence is that their effects are small. This means policy makers should spend a lot less time worrying about the incentive effects of marginal rates and a lot more worrying about other tax issues.
Most obviously, the federal budget is on a collision course with reality. Reining in the long-run deficit will have to involve slowing the growth rate of spending. But unless we choose to gut Medicare and Medicaid, additional tax revenue will be needed. This essential truth is the No. 1 factor that should be driving tax policy. And anyone who tells you that the way to raise revenue is to cut marginal tax rates is arguing from ideology, not solid evidence.
Many people have proposed raising revenue by cutting back on deductions and loopholes — so-called tax expenditures. Indeed, all else being equal, such changes are preferable to raising marginal rates. After all, higher rates do have some disincentive effects. Eliminating special tax provisions would also simplify tax preparation, and give people less incentive to try to game the system.
But if moderate increases in marginal rates wouldn’t much affect behavior, a mix of rate increases and cuts in tax expenditures might be a sensible path. Some tax expenditures, like the favorable tax treatment of employer-provided health insurance, may have worrisome effects (by encouraging overly generous plans) and so should be trimmed. But others, like deductibility of charitable giving, may be worth keeping.
Finally, income inequality has surged in recent decades. Raising marginal rates on the wealthy is a straightforward, effective way to counter this trend, while helping to solve our looming deficit problem. Given the strong evidence that the incentive effects of marginal rates are small, opponents of such a move will need a new argument. Invoking the myth of terrible supply-side consequences just won’t cut it.
Comment