Testimony before the U.S. House Committee on Ways and Means

Transcript

January 28, 1998

Testimony before the U.S. House Committee on Ways and Means

Bruce R. Bartlett
Senior Fellow
National Center for Policy Analysis

A marriage penalty results when a married couple pay more taxes by filing jointly than they would pay if each spouse could file as a single person. A marriage bonus results when a couple pay less taxes than they would pay as singles. Marriage penalties only result when both spouses have earned income. Single earner couples never pay a penalty and in fact always get a bonus from the Tax Code.

The marriage penalty fundamentally results from progressivity of the Tax Code.¹ Marginal income tax rates rise from 15 percent to 39.6 percent. This causes a marriage penalty because the earnings of the secondary worker (the lower-paid spouse) in effect come on top of the primary earner's. Thus, a secondary worker may find his or her income pushes the total income into a higher tax bracket, resulting in a marginal rate higher than the secondary worker would pay if taxed as a single.

To see how this works, consider a husband with taxable income of $25,000 per year. Under both the single and joint tax schedules he would pay 15 percent tax on that income. If his wife also makes $25,000, however, only the first $17,350 of her income would be taxed at 15 percent. The remaining $7,650 of her income would be taxed at 28 percent, because it puts the couple's total income above the $42,350 ceiling for the 15 percent bracket. Thus she will pay 13 percent more tax on that income (the difference between 15 percent and 28 percent) than she would pay if she were taxed as a single. In this case, that would make the marriage penalty $994 per year.

On the same total income, a couple may either get a tax bonus or pay a tax penalty depending on what the income split is between husband and wife. The couple in the earlier example paid the maximum marriage penalty on their $50,000 joint income because each spouse earned half the income. However, if one spouse earned substantially less than the other, the marriage penalty would have become a marriage bonus. If the husband earned $40,000 per year while the wife earned $10,000, instead of paying a penalty of $994 per year, they would have received a bonus of $910. That is, they would pay $910 less in taxes as a couple filing jointly than they would pay if each were taxed as a single.

The marriage penalty is most likely to strike couples whose incomes are roughly equal. No couple with equal incomes or those within 10 percent of each other receive a marriage bonus and most receive penalties. As noted earlier, no single-earner couples pay a marriage penalty and virtually all, regardless of income, receive a bonus.

To get an idea of how marriage penalties and bonuses affect real people, the Congressional Budget Office (CBO) looked at Internal Revenue Service and Census data. The CBO found that the highest proportion of marriage penalties occurred when the higher-earning spouse made between $20,000 and $75,000 per year. Couples with incomes above and below these levels were more likely to receive a tax bonus for being married.

Thus we see that marriage penalties are most likely to have an impact on couples with middle incomes whose incomes are roughly equal. In an interesting article, Professor Dorothy Brown of the University of Cincinnati College of Law has argued that these two factors mean that blacks are more likely to suffer a marriage penalty, while whites are more likely to receive a marriage bonus from the Tax Code. ² The reason is that among married couples, black women are more likely to work than white women. Furthermore, working black women on average provide a higher percentage of the couple's total income than working white women. According to a 1990 study by the U.S. Commission on Civil Rights, 75 percent of black women work full-time, whereas only 62 percent of white women do. And working black women contribute 40 percent of family earnings, while working white women contribute just 29 percent. ³

Although the marriage penalty is inherent in the nature of progressive tax rates, its magnitude has gone up and down with changes in the tax law. When the income tax was established in 1913, there was no distinction between married and unmarried taxpayers. There was a single rate schedule that applied to both.

The tax problems related to working women were much less in those days because only a small number of married women worked outside the home. In the census of 1900, there were only 769,000 married women in the labor force, out of a total of 27,640,000 workers. Even single women were unlikely to hold a paying job at that time. The female labor force participation rate was just 20 percent in 1900, compared to 86 percent for men. ⁴

In the 1920s, however, a number of couples in community property states began filing separate tax returns, with each spouse claiming half the couple's total income. ⁵ This was justified on the grounds that under community property each spouse is deemed to own half the couple's joint earnings, regardless of who earned them. By contrast, in common law states, the earnings of a spouse generally belonged to that spouse. Among the states with community property laws at that time were Texas, Arizona, Idaho, Louisiana, Nevada, New Mexico, Washington and California.

Initially, the Attorney General of the United States ruled that couples in community property states could split their income for tax purposes. This had the effect of reducing taxes for most couples. For example, if a husband had $20,000 of earnings and his wife had none, they would be taxed as if each earned $10,000. This generally put them in a lower tax bracket and lowered their joint tax liability. Had this state of affairs been allowed to continue, it would have led states to adopt community property laws just to give their citizens a cut in their federal income taxes.

Congress and the Treasury Department attempted to thwart the use of income splitting through legislation and regulations. Eventually, a case reached the Supreme Court on the question of income splitting. In Poe v. Seaborn (1930), the Court ruled that state community property laws did allow couples to split their incomes for federal income tax purposes. And as expected, it did indeed lead several states to change from common law to community property in order to give their citizens a tax cut at no expense to the state. This trend accelerated when tax rates shot up during World War II. By 1948, Oregon, Nebraska, Michigan and Oklahoma had changed their laws to become community property states. ⁶

Obviously, this situation led to a great deal of unfairness, with citizens of some states paying significantly lower federal income taxes than citizens of other states with the same income. The magnitude of the marriage penalty for couples in common law states in 1947 was quite high. Some couples in common law states were paying 40 percent more in federal income taxes than they would have paid in a community property state. A couple with a joint income of $25,000, for example, would have paid $9,082 in federal income taxes in a common law state, but only $6,460 in a community property state. ⁷ As Professor Michael Graetz of Yale recently noted, "this absurd situation did not engender great respect for the integrity of the income tax." ⁸

Congress finally resolved this problem in the Revenue Act of 1948, which extended the principle of income splitting to all married couples. ⁹ This constituted a significant tax cut for most married couples. The bulk of the benefits accrued to couples with middle incomes. ¹⁰

More significantly, almost every married couple saw a sharp reduction in their marginal tax rate - the tax that applies to the last dollar earned. A couple earning $51,000, for example, saw their marginal rate drop from 75 percent to 59 percent between 1947 and 1948. Again, those in the middle brackets, not the rich, were the principal beneficiaries.

In practice, the impact of lower tax rates was mainly on women. Since a married woman's earnings came on top of her husband's, she was in effect taxed at her husband's marginal tax rate on the first dollar of her earnings. With marginal tax rates going as high as 90 percent after World War II, this very strongly discouraged married women from working.

Although the institution of income splitting was highly beneficial to most married couples, it created a problem for single taxpayers. As a result of income splitting, a married couple mow paid significantly less tax than a single earner with the same income. Congress tried to address this inequity in 1951 by creating a new tax rate schedule for single heads of households, which roughly split the difference between the married and single tax schedules.

Singles, however, continued to agitate for tax relief. By 1969, some single taxpayers were paying 42 percent more federal taxes than a married couple with the same income. That year Congress created a new tax schedule for singles that was designed to keep the tax burden on singles and married couples with the same income within 20 percent of each other. This legislation created a significant marriage penalty for the first time. ¹¹ As a result, some married couples now paid more taxes by filing jointly than they would have paid if both filed as individuals. ¹²

Further contributing to the rise of the marriage penalty was the steep rise in the number of women in the labor force. The number of women in the labor force increased by about 50 percent between the late 1940s and the early 1970s. The labor force participation rate for women has continued to rise since and in 1997 was almost double the rate of 1947. This is important because a marriage penalty only occurs when a husband and wife both have earned income. With women working in greater and greater numbers, this means that the likelihood of a couple suffering a marriage penalty rose concomitantly.

As knowledge of the marriage penalty grew, increasing numbers of couples began to take matters into their own hands by getting divorced for tax reasons. One couple, David and Angela Boyter, received national publicity for getting divorced each December, allowing each to file as single for the year, and then getting remarried in January. ¹³ Eventually the IRS cracked down on this charade, but not before moving Congress to action. ¹⁴ By 1981, there was strong political pressure to redress the marriage penalty problem. A variety of proposals were put forward to accomplish this goal. ¹⁵

In the Economic Recovery Tax Act of 1981, Congress attempted to redress the marriage penalty by giving the lower-paid spouse a 10 percent tax deduction on income up to $30,000, for a maximum deduction of $3,000. While this provision did not eliminate the marriage penalty, it did redress the problem substantially for most married taxpayers. ¹⁶

The secondary earner deduction did not live long, however, and was eliminated by the Tax Reform Act of 1986. But because the Tax Reform Act sharply reduced tax rates for most taxpayers, the net effect was to reduce the number of couples suffering a marriage penalty and the magnitude of the penalty. ¹⁷ Nevertheless, some couples were worse off. ¹⁸

The most recent tax legislation with a major impact on the marriage penalty was the 1993 tax bill. ¹⁹ Interestingly, the provision of the legislation that exacerbated the marriage penalty was not the increase in tax rates, but the expansion of the Earned Income Tax Credit (EITC). The EITC is a refundable income tax credit for workers with low earnings. It creates marriage penalties because it is phased out as incomes rise and because it is maximized for workers with two children. ²⁰ No additional credit is available for three or more children in a single qualifying family. Depending on their income, therefore, a two-earner couple might significantly increase their joint EITC benefit by divorcing. And if they have more than two children, the benefits of divorce can be enormous. In 1996, for example, a two-earner couple with four children and each earning $11,000 would have increased their EITC payment from $1,375 to $7,120 by getting divorced, with each spouse claiming two children. ²¹

As noted earlier, the principal effect of the marriage penalty has been on wives, because they generally earn less than their husbands and thus are in effect taxed at their husbands' marginal tax rate. This means that wives generally receive less aftertax income on each dollar they earn than their husbands do. This alone is sufficient to significantly discourage work effort among married women. There is a considerable amount of economic research clearly demonstrating that high marginal tax rates reduce labor supply, especially for married women. ²²

The disincentive effects of high marginal tax rates on married women are aggravated by their looser attachment to the labor force than men and by their child-rearing responsibilities. ²³ Although most married women who work do so because of financial necessity, many do not. Their income is not essential for maintaining a couple's standard of living. Such women may work for a variety of reasons, including the simple joy of doing so. But the consequence is that they are more easily driven from the labor force by tax disincentives than married men are. For this reason, economic theory suggests that married women should be taxed less than married men. ²⁴

Thus it should come as no surprise that tax policies affecting the marriage penalty have had a significant impact on female labor supply. The institution of income splitting in 1948 and the effective reduction in marginal tax rates had a significant effect on women's work decisions. Between 1947 and 1950 the labor force participation rate for married women shot up, raising their share of the female labor force from 46.2 percent to 52.1 percent. Those with a husband present, those most likely to be affected by income splitting, increased their labor force participation most, increasing their share of the female labor force from 40.9 percent to 48 percent. By contrast, single, widowed or divorced women, who gained nothing from income splitting, saw their labor force participation stay flat or decline. The labor force participation rate for men was also unchanged over this period.

A study of the 1981 tax act, which reduced the marriage penalty by instituting a secondary earner deduction, shows that married women's work expanded by almost enough to pay for the deduction's revenue loss. ²⁵ Analysis of the Tax Reform Act of 1986, which lowered the top marginal tax rate from 50 percent to 28 percent, shows that married women responded more strongly to the increased work incentive than men did. ²⁶ Another study estimated that if the marriage penalties remaining after the Tax Reform Act were eliminated, the average married woman would increase her hours worked by 46 hours per year. High-income and low-income women would respond even more strongly, increasing their work hours by 100 hours per year. ²⁷

The latest estimates by economists Martin Feldstein and Daniel Feenberg suggest that the labor supply response of married women to reduction of the marriage penalty could be quite large. Sharply cutting the tax rate on secondary workers could lead to an increase in earnings by such workers of as much as $66 billion per year. ²⁸

In addition to effects on labor supply, the marriage penalty also has an impact on the marriage/divorce decision. There is certainly no question that over time the number of couples living together without marriage has sharply increased. The Census Bureau reports that 523,000 adults of the opposite sex were living together in 1970. By 1996, this figure had risen to 3,958,000. In 1970, just 0.5 percent of couples living together in the United States were unmarried. By 1996, this percentage had risen to 7.2 percent. At least some of this is undoubtedly due to tax considerations.

Several studies have looked at this question. They find that the marriage penalty has a small but significant impact on couples' decision to marry. When the marriage penalty rises aggregate marriage rates fall. There is a much greater impact on the timing of marriage, with couples often delaying marriage late in the year to minimize their marriage penalty. ²⁹ Finally, there is some evidence that taxes encourage divorce, especially on the part of women who are affected most by the marriage penalty. ³⁰

As noted earlier, from 1913 to 1948 Congress adopted an approach to taxation that did not differentiate between married and unmarried persons. There was only one tax schedule and everyone paid the same rates. A single person and a married couple with the same income paid the same tax. Congress did not willingly adopt income splitting in 1948. It was forced to do so out of necessity resulting from the consequences of a Supreme Court case. Nevertheless, the effect was to replace the individual with the family as the fundamental unit for taxation.

It has long been known that a tax system cannot simultaneously do three things: (1) have progressive tax rates, (2) have equal tax treatment of couples with the same income, and (3) be marriage-neutral. ³¹ The last point means that marital status would have no effect on an individual's tax liability. If the first point is accepted, one must choose between the second and third. In 1948, Congress chose the first and second and abandoned the third.

In recent years, a number of tax theorists have questioned Congress's decision. Progressivity is no longer assumed to be a primary criterion of our tax system. Increasingly, tax theorists question whether it is fair to penalize those with higher incomes, while economists produce more and more data on the economic cost of progressivity. At the same time, others question the assumption of family-based taxation. They argue that a system of individual filing would be fairer, simpler and more efficient.

The notion of progressivity has been under attack for many years. Tax experts have long known that exemptions, deductions and exclusions in the Tax Code can easily erode the nominal progressivity of the rate structure. They have also known that progressivity breeds complexity, evasion and imposes a large deadweight cost on the economy. But the idea that "fairness" demanded higher tax rates on those with upper incomes was too widespread to challenge. ³²

By the 1980s, however, opinion had shifted sufficiently that there was now serious support for the idea of a flat tax, one with a single tax rate for all taxpayers regardless of income. So popular was the idea that in 1986 Congress went a long way toward a flat tax by creating a two-rate tax system, with a top rate of just 28 percent. Eventually, even academic tax theorists began to come around to the idea. Now it is common to read criticism of progressivity in leading law journals, where earlier it would have been unthinkable. ³³

At the same time, economists have increasingly come to see the cost of progressivity as extremely high. One study put it this way:

Even a mild degree of progressivity in the income tax system (as measured by the steepness of the marginal rate schedule) imposes a very large efficiency cost. For example, in comparison with an equal revenue proportional income tax, a progressive income tax with average tax rates varying over the life cycle between .23 and .32 and marginal rates ranging from .23 to .43 imposes an efficiency cost greater than 6 percent of full lifetime resources. ³⁴

Since that study appeared, many others have come to similar conclusions about the overall welfare cost of progressivity in the U.S. tax system. ³⁵ As a result, a recent president of the American Economic Association has said, "Today, it is fair to say that many, if not most, economists favor the expenditure tax or flat rate income tax. This group has joined the opponents of progressive taxation in the attack on the income tax." ³⁶

Just as progressivity increasingly has become questioned as a norm of taxation, so too many tax theorists now question whether the family should be the fundamental unit of taxation. They suggest that the individual, rather than the family, is the most appropriate unit of taxation. Such a move would eliminate the marriage penalty completely, but would also eliminate marriage bonuses. Such bonuses, however, may be inappropriate because there is no particular reason why couples should receive special treatment from the Tax Code merely because they are married. To the extent that we wish to aid children, we could target tax deductions or credits directly to the children, rather than families in general. ³⁷

Individual taxation may also be better suited to changing societal mores. In 1948, relatively few women worked, few headed households, and most couples had a single earner. Now women work in almost the same percentages as men, female-headed households are common, and families represent a decreasing share of households. Indeed, growth of the marriage penalty is as much due to demographic changes as changes in the tax law. ³⁸ According to the Census Bureau, nonfamily households have risen from 18.8 percent of all households in 1970 to 30.1 percent in 1996. ³⁹ It is also worth noting that most major industrialized countries use the individual as the basic unit of taxation. ⁴⁰

It is not necessary to completely abandon the family as the basic unit of taxation in order to eliminate the marriage penalty. It would only be necessary to allow couples the choice of filing as singles or jointly. This would preserve marriage bonuses for single-earner couples, but eliminate the marriage penalty for two-earner couples. However, Congress would also have to pass rules about dividing joint income, such as interest and dividends, and allocating itemized deductions, such as for mortgage interest and dependents. ⁴¹

The major objections to the choice approach are complexity, cost and abandonment of the principle that couples with the same income should pay similar taxes. It would be complex because many couples would, in effect, have to do their taxes twice: first jointly and then as singles to see which way they would come out ahead. Also, whatever rules are adopted for allocating joint income and deductions are bound to be complicated.

Allowing couples to choose their filing status would also be costly. According to the CBO, it would have reduced federal revenues by $29 billion in 1996. ⁴² It will also lead to situations in which certain couples will pay less total taxes than others with the same income. This could create pressure in future years for further tax measures to redress this perceived imbalance.

Congress certainly needs to be wary about adding additional complexity to an already overly complicated Tax Code. However, in recent years Congress has enacted a number of very complicated provisions to the tax law involving phase-outs for various tax benefits that also have the effect of worsening the marriage penalty for some couples. For example, the child credit is phased-out for couples with incomes over $110,000 and over $75,000 for singles. This means that a couple making $75,000 each would qualify for the full $500 per child credit if they divorce, but receive nothing if married. ⁴³

Almost any solution to the marriage penalty is likely to increase complexity and raise questions about cost and fairness. ⁴⁴ Short of going all the way to an individual filing system, other options for redressing the marriage penalty include restoration of the second-earner deduction, such as that included in the 1981 tax bill, widening tax brackets and modifying provisions such as the EITC that create marriage penalties. ⁴⁵ Given the cost of full elimination of the marriage penalty and budgetary realities, in the end Congress will probably be forced to choose among these more limited options if it decides to address the issue at all.

A better solution to further tinkering with the Tax Code would be to move toward a flat rate income or consumption tax. By eliminating progressivity, it gets at the root cause of the marriage penalty. ⁴⁶ Although there are many other arguments for a flat tax, this one may prove most persuasive to two-earner couples.

END NOTES