A Comprehensive Approach To Removing Marriage Penalties


Prepared By
C. Eugene Steuerle
Senior Fellow
The Urban Institute

Sitting in the jury box those few days, I couldn't help but think about how our legal system -- taxes, transfers, Census measures -- would treat this particular household. The man was suing a company which he claimed entered his premises illegally and seized some goods that belonged to him, his wife, and his daughter. Only it wasn't his premises, it was the apartment of a friend. And the goods were not all his, some belonged to the constant stream of temporary occupants of this household. And he was not formally married, or so claimed his sister who occasionally lived in the same household. And his daughter was living in another city.

This household is not all that unusual. Each of us today knows many households whose members have permanent or temporary relationships with other adults, whether in dormitories, nursing homes, rental apartments with friends, convents, or the homes of parents or children. The Census Bureau tells us that male and female "cohabitating" partners alone have risen to eight percent of all households, but this is clearly an understatement. A significant share of these households would never report this information to Census, IRS, or any other agency.

The definition of a household is important for more than counting purposes. In particular, our government programs tend to base taxes and transfers on how the "household" or tax unit is defined. Small changes in definition, however, can have large consequences for individuals. In many cases, these tax and transfer systems have been built around stereotypes of the family that ignore the existence of many modern household arrangements and unintentionally provide strong disincentives to marriage and work(1). In my own research I have found:

    • Someone earning annual wages between one-quarter of the minimum wage and three times
    the minimum wage often faces a combination of explicit and implicit tax rates of 70 percent
    (for example, see Figure 1)(2). Until very recently, these high rates applied mainly to
    traditional, nonworking, welfare recipients who would both lose welfare benefits (through
    an implicit tax rate on income) and pay direct taxes on their wages if they worked. Today
    they apply to a significant portion of those working at moderate wages, although the full
    social implications probably will not be felt until the system has matured.(3)

    • This same set of tax rates impose very high penalties are also imposed on marriage -- or not
    divorcing -- especially in low- and moderate-income families. What happens is that the tax
    rate that normally applies to single low- and moderate-income workers -- usually ranging
    from 15 percent to 30 percent for Social Security and income taxes -- jumps to 70 percent
    or so when they marry low-income heads of household. Essentially the new new married
    partner's earnings are now added to the measure of household resources when government
    phases out benefits. Cohabitation or serial relationships, of course, avoid this large bump up
    in effective tax rate.

These large marriage penalties reflect the crazy quilt of tax and expenditure policies which directly affect the family.(4) It is a quilt weaved with a piece of thread here, and a string there; most importantly, it has no overall pattern or shape. If the reasons for these developments were pernicious, they would probably be easier to deal with, but they are primarily the unintended consequences of what at first appear to be reasonable goals. For example, military and foreign service pensions grant an extra benefit of protection for widows and widowers, a worthwhile goal. To limit costs, however, they often take away the benefit if the widow or widower remarries someone with sufficient earnings. The consequence is a marriage penalty on those who remarry rather than cohabitate with a new partner.

The Congressional Budget Office estimates that 26.8 million couples pay additional income tax simply because of marriage (while 21.5 million receive marriage bonuses)(5). But this is only the tip of the iceberg. An extraordinary array of marriage penalties also arise from the high rate at which welfare and other expenditure programs reduce benefits in the presence of a spouse with earnings. Because penalties are built into so many tax and transfer programs, almost every couple is faced with potential rewards from divorce or nonmarriage at some point or another, and many partners with marriage bonuses will one day pay tax penalties if their incomes become more equal.

Warnings are in order: this is not a simple tour, and it requires branching off at times to explain some related policies affecting the family. For example, giving benefits to married couples is sometimes favored as a means of compensating for the costs of raising children, but providing adjustments for children is a much more direct way of achieving that goal(6). The area is fraught with technical complexity, which is one reason few elected officials understand when or how they are creating marriage penalties. It's not enough simply to be against marriage penalties if one doesn't understand what causes them in the first place. Finally, because of conflict among legitimate principles, this "family" issue breaks down conventional wisdom of what it means to be a liberal or conservative, if such labels have any meaning in this context. Both Democrats and Republicans, right and left, share responsibility for the large historical increase in the size of these penalties.

My own view is that we should make substantial efforts to reduce marriage penalties because they violate traditional notions of fairness, and they weaken the natural tendency to rely upon marriage as an institution for mutual support. Further, they tend to subvert citizens' allegiance to a government of laws: those who manipulate taxes and transfers through marriage decisions come to view the law as expressing few values or norms; those who value the law but pay penalties as a consequence come to view it as unfair. I consider this as a special problem in poorer communities where government subsidies for not marrying have become a major source of financial support for the community itself.

Somewhat reluctantly, I also raise the issue of whether marriage for religious purposes should automatically imply marriage for purposes of the state. I say "reluctantly" because the legal support of marriage as a social institution was designed partly to protect spouses. But this was in a world where the state generally encouraged, or was neutral toward, marriage for almost everyone -- not in a world where it penalized marriage for substantial portions of the population. I also say "reluctantly" because conflicting principles imply no clear-cut righteous answer, even though a large tax on moral commitment cannot help but have moral implications.

To remove most marriage penalties, as I will demonstrate, requires reconsideration of practically the entire shape of our transfer and tax systems. This tail literally would wag the dog, or perhaps more precisely, one can't take the tail somewhere without bringing along the whole dog.

Independently from my own views, most of this essay will pull together, for what I believe the first time, a fairly comprehensive analysis of what causes marriage penalties throughout the expenditure and tax systems and the pros and cons of alternative methods to remove or reduce them.

Required Conditions for Creating Marriage Penalties

Now let us turn more precisely to the conditions under which marriage penalties arise. Except for the income tax, which is a special case that will be discussed later, I do not believe that these conditions have been specified elsewhere in the literature. For those considering the reduction or elimination of marriage penalties, a technical understanding of these conditions is crucial if they wish to do more than merely tinker with the transfer and tax system.

    Marriage penalties or bonuses will arise in almost any tax or expenditure system meeting
    the following two conditions:

    (1) A tax (or subsidy), explicit or implicit, assessed on the basis of household or family
    resources,(7) and

    (2) Different marginal or incremental tax (or subsidy) rates at different levels of resources.

Explicit taxes are those that derive from quite visible tax rate schedules as in the income tax. Implicit taxes are those that derive from the "phase-out," claw back, or elimination of some benefit as resources increases. Resources may include income, wealth, or consumption (or even the addition of a new "breadwinner" in the case of some widow's and widower's benefits), but in most programs household income is the base on which the phase out of benefits is assessed. The marginal or incremental tax rate refers to the tax imposed on an additional dollar of resources, not the average rate on all resources. The increased cost in taxes of an rise in income, say, from $10,000 to $10,001 is the marginal tax rate.(8)

Economies of Scale in Larger Households: The Taxation of Committed Sharing. Some believe that taxing or subsidizing on the basis of the household (condition 1) should involve marriage penalties -- imposing higher taxes or fewer benefits on a married couple than on two single individuals with the same combined income as the couple. The argument is that there are economies of scale in households because of shared facilities and goods. One TV may be enough for two people, one person may be able to prepare a meal for two just as easily as for one, and so on. Thus, sharing is a process that adds economic benefit over and above the money income that individuals receive. Through the process of sharing, the household effectively achieves economies of scale in consumption.

One way this plays out is through adherence to providing benefits or taxes according to a "poverty" scale. Because of the way that they were constructed, most poverty scales imply that a couple needs only about 1 1/2 times the income of a single person to avoid poverty. Suppose a program would be designed to cover exactly the cost of poverty. Then if two people married they would get only 3/4 of the benefits they would receive if they remained as single individuals and cohabitated.

If taxes or benefits had no effect on behavior, and we were trying to treat all households equally on their ability, then it is correct, technically, that the household sharing goods and services would have greater ability than an equal size, equal income grouping of people not sharing. The problem with using this argument to justify marriage penalties is not that there are no economies of scale from sharing. There are, and, indeed, these gains reinforce other natural instincts to engage in mutual support.

Economies of scale, however, apply to almost all sharing arrangements -- dormitories, old age homes, movie houses, and cohabitation. Yet, in practice today, marital vows of allegiance are the only type of arrangement that is taxed. At one point welfare programs tried to determine if there were cohabitating partners but has largely abandoned this effort because of the intrusion involved and the difficulty of providing adequate proof.

In those communities where marriage is no longer the norm -- and these communities are growing -- this natural social incentive to achieve economies of scale in living arrangements does not disappear, but merely is converted into forms that avoid the marriage contract. (Recall the household involved in the jury trial noted at the beginning of this paper.) For example, adult males in marriage-discouraged communities often still live with someone, only now they are more likely than before to stay with their mothers, with other relatives or friends, or in serial relationships rather than with a spouse or with their own children. The tax and transfer systems say that these males deserve significantly lower levels of taxation and higher levels of support than males with equal incomes who marry. If they are fathers, it tells them that they can support their children better by remaining unmarried.

The economies of scale argument can be turned on its head. Since sharing can increase the effective well being of individuals even with no nominal increase in their nominal income, society might want to subsidize it rather than tax it. A similar efficiency argument lays behind some of the subsidies offered for education and other income-improving activities.

Marriage Penalties: A Classic Liberal-Conservative Compromise

Marriage penalties are a classic example of the type of liberal-conservative compromise that has dominated policy-making for several decades. The primary (although not only) problem arises in a modern welfare state that attempts to provide a variety of benefits to households, but then wants to limit costs by confining eligibility (thereby creating conditions 1 and 2). Note, however, that marriage penalties do not derive from more traditional government programs that provide benefits to the community on a more universal basis, such as defense, primary and secondary education, highways, or even Medicare. These programs may be good or bad, efficient or inefficient, but they do not involve the separating out and labelling of groups by economic status. None of these programs create marriage penalties because the individual gets the same benefits whether married or not.

It is usually the way in which specific expenditures and taxes are targeted that penalties arise. Liberals, wanting social programs to be as progressive as possible, often try to concentrate whatever benefits are available at the bottom of the income distribution. Conservatives, wanting to limit the cost to government, also want to limit the benefits. Both motives, progressivity and budget containment, are honorable. The compromise usually used to achieve these goals, however, is to phase out benefits quickly as income (or wealth) increases in the household. This effort attempts to achieve "target efficiency" by maximizing help to the poor for the lowest stated expenditure cost.

To tax experts, phase-outs are implicit taxes with almost identical economic effects as explicit taxes.(9) Their effects, however, are often hidden (see (a)). The main justification for putting these income phase outs in expenditure programs rather than a direct tax like the income tax is that with less universal programs, the additional tax is assessed only against the beneficiaries of the program itself. For the most part, however, most government policy makers have little idea of how all these systems combine together.

Since each new expenditure and tax subsidy program tends to have its own unique, built-in, phase out, households in America (and in most developed countries) literally face dozens of tax systems, both big and small, most of which arise out of subsidy programs. The Joint Committee on Taxation recently identified twenty-two provisions in the income tax alone that resulted in a taxpayer's marginal tax rate differing from the statutory tax rate.(10) For example, if I lose 50 cents of a benefit when my income goes up by $1, the effective tax rate from that benefit program alone is 50 percent. Now start to think about all of the multiple program in the expenditure and direct tax systems that are phased out -- welfare, food stamps, housing allowances, earned income tax credits, Medicaid, child credits, educational assistance, personal exemptions, eligibility for participation in individual retirement accounts, exclusion from the minimum tax, and so on. The 1997 tax legislation, small by historic standards, added several new phase-outs all by itself. It is not hard to see why so many households in all income ranges, but especially those with low incomes in welfare programs, face tax rates of 50 percent, 70 percent, or even 100 percent for some of their income.(11)

These high tax rates affect not only extra income earned through work. They affect any income introduced into a benefit-receiving household through marriage. Some examples are presented in Table 1 and Figure 2.(12) One of the worst cases involves a low-income single head of household earning about $10,000 a year -- exactly the situation to which recent welfare reform encourages heads of households to move. Such an individual does succeed in moving out of traditional welfare (defined as Aid to Families With Dependent Children or its replacement, Temporary Assistance to Needy Families), but still receives a variety of other supplements, such as food stamps and earned income tax credits. The problem is that if this benefit recipient now marries a single person earning $20,000 a year, their combined income would fall by almost 30 percent, or close to $9,000, because of the marriage alone! What happens is that the income of the new partner, normally taxed at a rate of about 15 to 30 percent through federal and state income taxes and Social Security tax is suddenly taxed at a rate of 70 percent or more simply because it is combined with the income of the benefit recipient to force further phase out of income assistance.

Confirming this analysis with a sample of representative taxpayers, Dickert-Conlin and Houser find that the net gain to separating for the median married poor family was 16 percent of income. Similarly, they find that the median loss in income from marriage is equal to 12 percent of unmarried income for all poor single women with children. However, while their estimates include most taxes and transfers, the marriage penalties would be even larger if programs like Medicaid and housing assistance were included.(13)

These programs may explain in part why the literature on fatherhood reports that in many low-income families and neighborhoods fathers feel little sense of accomplishment in staying around to marry and raise children in a traditional family structure. Government in effect has declared that working fathers in low-income, two-parent families are a liability.(14) Whatever the changes in cultural standards or mores that may have led to this situation, the government has created enormous barriers to responsible fatherhood (whether of one's own children or stepchildren) in the context of marriage. The total income of those communities and housing complexes dominated by income assistance, children born out of wedlock, and the absence of married couples would fall significantly if individuals in these communities would marry in patterns closer to national averages. In these communities, government has effectively pronounced that marriage is a foolish exercise -- even though marriage is a principal route out of poverty.(15)

These penalties in expenditure programs are not all confined to programs to provide income assistance. Take the case of some widows' and widowers' benefits. Policy makers treat the widow or widower as a welfare-type recipient who no longer "qualifies" as needy if she marries someone who has his or her own source of income. Examples here include benefits for widows and widowers of career military and foreign service personnel. The effect is the same -- a large tax on marriage.

The Traditional Marriage Penalty Issue: The Direct Income Tax

While here we emphasize the marriage penalties arising out of many subsidy and tax programs, Congress in recent years has turned its attention mainly to a moderate subset of all the marriage penalties it has created. Members of Congress and many students of the individual income tax often associate marriage penalties only with the progressive rate schedule in that tax and with its requirement for joint filing in the case of married couples. These features of the American income tax meets the two conditions that create marriage penalties or bonuses. In effect, the progressive rate schedule often requires an individual's income to be taxed at a different rate depending on whether or not that individual is taxed as a single person or as part of a joint return.(16)

While this paper addresses marriage penalties arising from multiple tax and subsidy systems, the history of the income tax is relevant for understanding the fixes that are often proposed today.(18) Before 1948 taxpayers filed as individuals, so in most states there were no marriage penalties or bonuses arising from the rate schedule (no condition 1)(19). Between 1948 and 1969, there were no penalties arising from the marriage of two single individuals who married, but for a different reason.(20) Although there was joint filing as a household by this time, the 1948-1969 period also entailed what was called "income splitting." Essentially individuals were allowed as married couples to "split" their income equally and then be taxed as if they were each single individuals. This effectively allowed them to file with an assumed share of income that would minimize total tax burden, although it simultaneously created marriage bonuses in almost all married households with two earners.(21)

After 1969 Congress turned its concern toward single individuals who paid more tax than married individuals who had equal earnings but split their income with a spouse. It abandoned complete income splitting, and gave married couples a tax treatment that effectively divided their income roughly as if it were about 70-30 (70 percent earned by one spouse, 30 percent by the other) or 80-20. What this meant was that couples with more equal income splits such as 50-50 paid marriage penalties even while those with less equal splits approaching 100-0 still got marriage bonuses.

With some exceptions, these are the conditions that still apply today. There is no real formula or principle defining where an income split should lead to a marriage penalty or bonus, and the crossover point tends to vary by the total income of the household. As noted above, Congress has also added an earned income tax credit that phases in and phases out, and it has phased out many income tax benefits ranging from eligibility for IRA deductions to personal exemptions. Like income-conditioned expenditures, these types of provisions do create marriage penalties, but they cannot be analyzed in the traditional income tax framework because they meet condition 2 through a different type of tax rate structure -- one where the marginal rate on additional income falls as income rises.

Options for Removing or Reducing Marriage Penalties

To reduce or eliminate marriage penalties essentially involves addressing either of the two conditions -- household filing or multiple rates -- that create the penalties. There are several approaches or options, and sometimes they can be combined.(22) These options and some of their related difficulties are outlined below.

Flattening of the combined tax rate schedule. No, this isn't a pitch for or against supply-side economics. But the consequences of a variable rate schedule must be recognized. Tax reform in 1986 attempted to get at the problem of marriage penalties partly by flattening the income tax rate schedule in a way that did not reduce overall progressivity. By evenly taxing all sources of income and removing deductions and tax shelter opportunities -- that is, by offsetting the benefits of rate reduction for each income group with fewer tax preferences -- this reform was able to maintain progressivity while reducing and flattening statutory rates. Congress has since been moving in the opposite direction with the income tax by reducing the tax base and raising tax rates in budget agreements in 1990, 1993, and 1997. Still, as noted above, the income tax is not the main problem when it comes to marriage penalties.

Complete elimination of variable rates would mean that income would face the same tax rate whether the income was combined in marriage or not. The principal application would be in transfer programs, not the income tax. As a practical step, one could try to limit the combined marginal tax rate for low and moderate income individuals moving out of welfare programs to around 30 percent or 40 percent rather than the 70 percent (sometimes 100 percent or more) rate they now often face. Then, when they moved into the income or Social Security tax systems, they would again face a similar combined rate (in general, the 15.3 percent Social Security tax plus a 15 percent bottom rate bracket in the income tax). The marriage would be penalized much less and it would matter less how the income was split as long as each person faced the same tax rate inside or outside of marriage. As explained below, it is possible to use this method to flatten the rates only in the lower and middle classes, and use some other method to deal with higher income individuals in the income tax. This would allow the tax system to retain a fair amount of progressivity.

Note that a true flattening of tax rates at low-income levels requires a wide variety of legislative and administrative action. It requires reconsideration of almost every income-conditioned tax and expenditure program on the books. Today, however, these programs often do not even share administrative records and their administrators have little idea how they overlap. It requires abandonment of the liberal-conservative compromise placing so much stress on progressivity and measured (although not always real) budget saving within every income-related program, taken one at a time. Given changes that could simultaneously be made in other features, such as the level of minimum benefit and the direct income tax rates at higher income levels, it is unclear whether overall progressivity would be reduced by this type of effort. Nonetheless, phase-outs cannot be sought by Congress every time it deals with an individual expenditure or tax subsidy program without playing havoc with marriage penalties.

Income Splitting. Rather than trying to reform the system by flattening tax rates, one could try to work on the conditions under which family or household filing is required. One of the older methods of doing this in the income tax was to allow income splitting. To repeat, income splitting effectively taxes married couples as if each was an individual filing a single return and having exactly the right share (one-half) of the couple's total income so as to minimize tax liability. Returning to the days of income splitting would deal mainly with marriage penalties arising from the direct income tax rate schedule that applies to the middle- and upper-income classes.

Unfortunately, income splitting eliminates only those marriage penalties arising from a rate schedule where rates are always successively as high or higher at higher levels of income. But our two conditions demonstrate that marriage penalties can also occur when rates fall as income increases. Because of all the phase-outs and implicit tax schemes Congress has adopted, the real tax system now imposes such a rate structure on a large portion of the population. Thus, when one moves through the phase out ranges of the earned income tax credit, food stamps, Medicaid, the itemized deduction limitation, the IRA contribution limit, and so forth, marginal tax rates fall rather than rise.

Take the simple case of the welfare recipient who considers marriage. Even with split income and individual filing, marriage still can introduce income that reduces or removes eligibility for welfare payments. Assume a welfare/tax structure that provides to unmarried adults $5,000 of benefits at zero income, no tax and no benefits at $10,000 of income, and a tax of $1,000 at $20,000 of income. Before marriage, a couple with $0 income for one partner and $20,000 for the other would get benefits of $5,000 and taxes of $1,000 for a net benefit of $4,000. If they marry and split their income, then each is treated as having $10,000 of income. They then get no benefits and pay no taxes for a net benefit of $0. True, income splitting applied to the progressive rate schedule part of the income tax did reduce the couple's income tax slightly, but that only slightly offset their loss of other benefits. The net income of the couple still fell dramatically because of the marriage itself.

Still another issue is that income splitting in the income tax would increase marriage bonuses for households where income is not split very evenly. This raises the cost of action. A compromise here is to limit income splitting, say, to the first $50,000 of income and, of course, use that technique only with the income tax.

Income splitting is not a way of dealing with efforts of parents to raise children, since the marriage bonuses go to both those with and without children. Where adjustments are desired because of the cost of raising children, they can be achieved through child credits and dependent exemptions without giving additional bonuses to all married couples. Put another way, if the goal is to give parents some reduction in tax or increase in expenditures because of the presence of children, spousal adjustments are a poorly targeted device.

Mandatory Individual Filing. For income tax purposes Congress could return to the days of mandatory individual filing. While it still might be easy to move around property income to achieve income splitting through the back door, that is not true for earnings from work, which, with the increase in the number of two-earner couples, is the dominant concern today -- unlike a few decades ago. Community property state rules are also no longer considered to be a major issue because of a change in legal interpretation; individuals are no longer allowed to split their income and file as single individuals to avoid marriage penalties under the current income tax.

With respect to phase-outs and welfare programs, however, pure individual filing could mean that the nonworking spouse of a millionaire might be entitled also to welfare benefits. An intriguing possibility here, and one that I increasingly favor, is that wage subsidies like the earned income tax credit accompany the worker and not the family, and that child credits or subsidies accompany the child. Canada is currently experimenting with such an approach with respect to wage subsidies, although eligible recipients are limited to those already on welfare.(23) I recognize that some high income families would then get a credit or subsidy. But high income individuals and households could still more than pay for any small subsidy through the overall progressivity of the tax and transfer system.

Worrying about whether someone paying millions of dollars in tax gets a small subsidy here or there is simply not worth the trouble. On average, higher income families can be made to pay for these changes through an explicit tax rate structure. The standards applied to different expenditure programs are simply not consistent. Thus, we now have programs that allow high-income individuals to benefit from the much larger and more expensive Social Security or Medicare programs or from public school education, yet progressivity is not removed in these systems that do not have benefit phase outs. The rich simply pay more than their share of taxes to support these systems. Put another way, progressivity is determined by the net benefit and tax structure, not by the benefit structure in one or even several programs.

My main concern here is with trying to address issues related to parenthood, marriage, and work among low and moderate income individuals. If subsidies were applied at low wage levels on an individual basis, they would not create the current strong incentives for absent parents (mainly fathers) and lack of marriage in low-income communities. The trick again is that once a program is aimed at the individual, marriage can have no effect on benefits or taxes paid. This type of reform could significantly change the environment of those communities. Take an example: if a single mother earning $10,000 a year received a wage rate subsidy like the Earned Income Tax Credit on an individual basis, she would not lose it if she married someone with income of his own. Similarly, if a low-earning male married into a family, he would still be eligible for any wage rate subsidy that was available for low-income workers living with dependent children. With individually based programs, his earnings would not affect her credit, and her earnings would not affect his.

Optional Individual Filing. Mandatory individual filing would partially "pay" for the removal of income tax marriage penalties by removing marriage bonuses as well. An alternative is optional individual filing, a procedure that allows married couples either to garner marriage bonuses, or when there are penalties for joint filing, to avoid them by filing as separate individuals. Optional filing increases administrative costs in deciding how to file, as the taxpayer often must calculate taxes two different ways to see which is cheaper. When options are put into place in several areas -- e.g., choosing between regular individual retirement accounts (IRAs) or Roth IRAs -- the combination of options that have to be considered rise in an exponential fashion. Revenues also fall and costs rise when individuals can optionally choose the best system to apply to themselves.

Unfortunately, many taxpayers are already in a world of optional individual filing. The main difference is that the benefit now is granted only to those who are able or willing to treat the act of marriage as the option by which to trigger filing status. Putting optional individual filing into the statutes would simply extend those benefits from unmarried couples to married couples. Then the discretionary act triggering the lower tax burden would switch from avoiding marriage to filing an alternative type of return. Many states already allow optional individual filing for state income tax purposes. They often avoid the increased complexity of two calculations, however, because they are really systems based on the taxation of the individual, not the household, and, hence, joint filing is almost always an inferior option.

Special Subsidies for Second Earners. In 1981, Congress adopted a special subsidy for second earners by exempting from tax a small percentage of the income of the lesser-earning spouse of a married couple. In 1986 this subsidy was removed because of its complexity and because the flattening of the rate schedules had removed many marriage penalties in the income tax. One of the difficulties of this method was that it was not very efficient, it gave subsidies to some families who already had marriage subsidies, and it did not target benefits to cases where marriage penalties were the worst.

A Comprehensive Approach

In my view, taxing a large share of marital commitments makes little sense in any society, much less one searching for ways to revive or foster community (or "communitarian") spirit among its members. After all, the primary feature of community is to share, and the most basic form of sharing is between two people or within a family. Admittedly, the research in this field does not prove that removal of marriage penalties would have a significant effect on behavior.(24) It would not, for instance, by itself reverse the sexual revolution. But the empirical research is not good at detecting the influences of policy on long-term social norms, as opposed to whether slight variations in tax rates result in behavioral variations among individuals at a point in time. These social effects take place over very long periods and often do not show up in the statistical studies that are made across individuals at a point in time or that follow some individuals for a few years.

Although the marriage penalties within the income tax have been around since 1969, moreover, we have only recently moved to a society where the very large marriage penalties from income assistance and wage subsidy programs have been extended well into the middle class and beyond the stereotypical poor, nonworking, welfare recipient. It is doubtful that the long-run influences of any of these conditions have yet to be fully experienced by society. As an aside, it is interesting that policy makers seem willing to accept evidence that high marginal rates influence behavior negatively at the top but not at the bottom of the income distribution.

Independently from whether marriage penalties will significantly affect behavior in a narrow sense, I believe they have a corrosive effect on society and especially on those low-income communities where the penalties are so high that average societal rates of marriage would significantly reduce the total income of the community. Marriage penalties violate almost everyone's sense of fairness because they penalize only that type of sharing done through moral and legal promises. These penalties further discourage responsible fatherhood and motherhood. Finally, where there are economies of scale, one can turn the argument on its head. That is, if economies of scale are a way of improving economic well-being by stretching the value of each dollar further, then one shouldn't want to discourage gains from economies of scale any more than one should want to discourage other income-improving behaviors, such as education.

What all this implies in practical terms is that to deal with marriage penalties in a thorough manner, Congress almost inevitably has to reconsider the entire range of explicit and implicit taxes it has imposed on income. Most of the hidden taxes it has adopted over the years would need to be reconsidered, pulled into an integrated whole, and, where appropriate, replaced by direct, explicit taxes.

Gradually over time, a combination policy that I increasingly have come to favor would not only replace implicit with explicit taxes, but would also contain some of each of the following elements:

     • Reduce combined marginal tax rates on low- and moderate-income individuals so that
      they did not rise much above the rate that applies to middle- and higher-income
      individuals (in this latter case, ranging today from about 30 percent to 45 percent). Politically,
      I would justify this last reform on the basis that we should consider tax disincentives as important
      or more important for low and middle income taxpayers as for upper income taxpayers.(25)
     
      Movement to this type of rate, however, would entail more universal benefits than under
      current law and less concentration of benefits at the very bottom. One could achieve this end
      without reducing real benefits in existing programs for low-income individuals if a portion
      of the revenues that derive from economic growth would be used to reduce the tax rates
      applying mainly to these low- and moderate-income families.
     
      • Use the direct tax rate schedule as the primary means to establish overall progressivity
      and abandon the complicated effort to put "progressivity" into everything government
      does. This requires recognition that each new phase out introduced is simply another tax rate
      schedule added onto the existing system.
     
      • Within the income tax, apply income splitting to lower- and some middle-income ranges.
      For example, if a single individual got a standard deduction of $3,000, a couple would receive
      a standard deduction of $6,000. If the single individual faced a tax rate of 15 percent for the
      first $30,000 of income(26), a couple would face a tax rate of 15 percent for the first $60,000
      of income. Recalling our discussion above, this means that the couple would never face a
      higher tax rate if married than if single, at least for income up to that level.(27)
     
      • For income over and above the levels to which income splitting would apply, use
      mandatory or optional individual filing. For a given rate structure, this would be both less
      expensive and retain a higher degree of progressivity than income splitting at all levels. Although
      optional filing is more complex, the complexity would be limited if it applied only at higher
      income levels.
     
      • If income splitting is unacceptable at any income level, enact mandatory or optional
      individual filing at lower and middle-income levels. While more complex than needed at
      those income levels, it would still be preferable to current law.

Let me again be clear that the issue of how to make adjustments for children for the most part is a separable issue. Marriage bonuses, for instance, do nothing for single heads of households with children, while providing extra benefits for those who no longer are raising children. Therefore, family adjustments for children can be targeted more directly through changes in such devices as the child credit that was first made available in 1997 tax legislation for 1998 and beyond. Nonetheless, we must note that this credit, along with a related child and dependent exemption, are currently more complex than required and are among the multitude of programs that have phase-outs that need to be addressed because of the marriage penalties they create. While a very important family issue, adjustments for children are not addressed in this paper but elsewhere.(28)

As I have indicated, it is not clear to me that policy makers fully comprehend how to achieve changes of this scale and magnitude. Incremental changes here and there may reduce marriage penalties little, and they may be more than offset by new marriage penalties introduced every time there is some new phase out or implicit income tax introduced. The momentum for change may have to come from acceptance of a broader principle. For example, a law might limit the combined marginal tax rate facing a low- or moderate-income worker to no more than the tax rate applying to the highest income individuals. To implement that goal, however, would also require years of effort to coordinate administrative structures in all the government's many programs, converting the crazy quilt of family policy into a more consistently designed overall pattern.

As long as policy makers are not able to deal with the issue in any comprehensive way, perhaps some private consideration ought to be given to examining the consequences of allowing marriage for religious or moral purposes without legal marriage being an automatic consequence. This is a matter that needs to be considered carefully and wisely, for there are a variety of issues -- especially those dealing with spousal rights -- that must be given substantial consideration. Nor, given the conflict of legitimate principles, is it a matter on which anyone can be righteous. Still, consideration of an option on this level might highlight the nature of the problem and the dilemmas which current law has established. A primary goal in all the alternatives I have suggested is to find a pragmatic way to remove from families -- particularly, low-income families -- the moral quandary of having to reduce the family support they can give each other and their children if they marry.

1. See Eugene Steuerle, Edward M. Gramlich, Hugh Heclo and Demetra Smith Nightingale, The Government We Deserve, Washington, DC: Urban Institute Press, 1998.

2. Explicit tax rates are those that can be seen directly in the tax rate schedule of income and Social Security taxes; implicit tax rates, as discussed in more detail below, are those that derive from a loss of benefits as income increases.

3. See Eugene Steuerle and Linda Giannarelli, “The True Tax Rates Faced by Welfare Recipients,” in National Tax Association, Proceedings of the Eighty-Seventh Annual Conference, 1995, 1996, p. 123-129.

4. Still another source of marriage issues arise with spousal benefits in Social Security. As it turns our, these provide a “marriage bonus” that is much more generous for the spouse of a rich person than the spouse of a poor person and is independent of any child-raising performed by the family. This particular design of spousal benefits contributed significantly to two lesser known facts about Social Security: it provided greater net transfers (benefits over and above taxes paid) to the rich than the poor over most of its history; and it continues to provide significantly higher benefits to families with one earner than to families with two earners when they earn the same total family income and pay the same total Social Security tax. See Eugene Steuerle and Jon Bakija, Retooling Social Security for the 21st Century: Right and Wrong Approaches to Reform, Washington, DC: Urban Institute Press, 1994, p. 207-215.

5. See, Congressional Budget Office, “For Better or Worse: Marriage and the Federal Income Tax,” Washington, DC, June 1997, p. 13.

6. For decades after World War II families with children were made to bear higher and higher portions of the total tax burden, as family adjustments such as the dependent exemption did not keep pace with economic growth. While my research here was partially instrumental in the creation of two offsets -- child tax credits in 1997 and increases in the dependent exemption in 1986 -- a typical middle-income working family with children still bears a much larger share of the tax burden than it did a few decades ago. See, Eugene Steuerle, “The Tax Treatment of Households of Different Size,” in Rudolph G. Penner (ed), Taxing the Family, Washington DC: American Enterprise Institute, 1993. See also, Eugene Steuerle, “Taxation of the Family,” Statement before the Committee on Ways and Means, United States House of Representatives, April 15, 1997.

7. A rare exception is when the government removes potential penalties by assuming conditions at least as good as, if not better, than individual filing. This was the case with income splitting in the income tax, discussed below.

8. Also, note that while we will confine our examples below to those provisions that tax or take away something from households as their resources increase, it is possible to create marriage penalties in less common, pure subsidy systems that never phase out, as long as they meet our two basic conditions.

9. The purest example of hiding a tax system in an expenditure program comes from an expenditure that is universally available, except at higher income levels. If a phase out rate applies to everyone’s income, it operates just as if it were part of the income tax. One proposal that came close to this condition was President Bush’s proposal to create an almost universal health credit that would phase out quickly once a family moved above poverty-level income. If the phase out were explicitly put as a tax rate in the income tax instead, it could have achieved almost the same end, but if done that way, the proposal would have appeared much more expensive and explicit tax rates of near to 100 percent would have formally shown up in the income tax rate schedule. With a phase out rather than a direct tax rate, the cost of the proposal appears smaller and no tax rates are revealed.

10. These differences included some phase-ins and floors, as well as the phase-outs on which most our attention is directed here. See Joint Committee On Taxation, “Present Law and Analysis Relating to Individual Effective Marginal Tax Rates,” Washington, D.C.: U.S. Government Printing Office, 1998.

11. As an aside, one reason that meaningful income tax reform is so hard to achieve is that most proposals only deal with the direct income tax, while ignoring the dozens of other income tax systems we now have. What does it mean, for instance, to have a “flat” tax or a “consumption” tax if only one of the many “income taxes” are being flattened or converted to a consumption base?

12. Examples are taken from C. Eugene Steuerle “The Effects of Tax and Welfare Policies on Family Formation,” in “Strategies to Strengthen Marriage - What Do We Know? What Do We Need to Know? Papers presented at a Family Impact Seminar Roundtable Meeting, June 23-24, 1997, Washington, DC.

13. See Stacey Dickert-Conlin and Scott Houser, “Taxes and Transfers: A New Look at the Marriage Penalty,” National Tax Journal 51(2), June, 1998, pp.175-217.

14. See Eugene Steuerle, Edward Gramlich, Hugh Heclo and Demetra Smith Nightingale, The Government We Deserve, Washington, DC: Urban Institute Press, 1998, p. 40-43.

15. See Greg Duncan, Richard Coe and Martha Hill, “The Dynamics of Poverty,” in Greg Duncan (ed.), Years of Poverty, Years of Plenty. Institute for Social Research, Ann Arbor, MI: 1984.

16. There was a brief period of time, before the days when transfer systems were large, when most marriage penalties did arise out of the progressive income tax rate schedule and its standard deduction (which may be considered for this purpose as nothing more than part of the rate schedule -- that part which taxes some initial levels of income at a zero rate).

17. See, Congressional Budget Office, “For Better or Worse: Marriage and the Federal Income Tax,” Washington, DC, June 1997.

18. For a longer discussion of the history behind income tax marriage penalties see, Congressional Budget Office, “For Better or Worse: Marriage and the Federal Income Tax,” Washington, DC, June, 1997. See also, Joint Committee on Taxation, “Present Law and Background Relating to Proposals to Reduce the Marriage Penalty Tax,” Highlights and Documents, January 28, 1998.

19. An exception was in community property states where a married individual could lower liability effectively by treating income as if earned by a spouse. In those states, there was a marriage bonus to many couples.

20. A new head of household rate schedule for single individuals with dependents would create some marriage penalties.

21. The change-over from individual filing was forced on Congress partly by complications created by community property states in the absence of universal income splitting. In those states, income was essentially divided equally between the partners in marriage -- so that individual filing by married couples in those states effectively created the same situation as income splitting. Unfortunately, in the days of individual filing this meant that couples in community property states had a distinct advantage relative to couples in other states. By adopting joint filing and income splitting, Congress extended to all couples the benefits previously enjoyed only by those in community property states See Michael J. McIntyre and Eugene Steuerle, Federal Tax Reform: A Family Perspective, Washington, DC: The Finance Project, 1996.

22. For another discussion of some proposals to reduce marriage penalties in the income tax see, “Reducing Marriage Taxes: Issues and Proposals,” Joint Committee on Taxation, May 1998. This report addresses only those penalties arising in the income tax and does not get at penalties in expenditure programs.

23. The experiment goes under the name of the “Self-Sufficiency” Project.

24. See for example, James Alms and Leslie Whittington, “Til Death or Taxes Do Us Part: The Effect of income Taxation on Divorce,” Journal of Human Resources, Spring 1997, James Alms and Leslie Whittington, “Income Taxes and the Timing of Marital Decisions,” Journal of Public Economics, 64(2), May 1997, James Alms and Leslie Whittington, “The Rise and Fall and Rise...of the Marriage Tax,” National Tax Journal, 49(4), Dec 1996. See also, Stacy Dickert-Conlin, “Taxes, Transfers and Family Structure: Are They Related?,” Department of Economics and the Martin School of Public Policy, University of Kentucky, April 1996. See also, Robert Moffitt, “Incentive Effects of the US Welfare System: A Review,” Journal of Economic Literature, 30(1), March 1992.

25. Advocates of purer forms of supply-side economics will note that a more even set of marginal rates goes against the movement toward a “head tax,” where the highest tax rate applies to the first dollars of income so that marginal rates can be moved toward zero as income increases. However, even most advocates of a flat rate tax do not argue for movement toward a head tax.

26. For example, one could double, for couples, the income level at which the 15 percent rate bracket ends.

27. One would need to eliminate the separate schedule for heads of household if the goal were to remove marriage penalties in the presence of children. This paper has not dealt in depth with this other source of complication. As indicated in the text, however, adjustments for children generally should be made through child credits and exemptions, not through attempts at income splitting (which is what the head of household schedule does). Since heads of household tend to be poorer on average than other households, these adjustments for children may need to be raised to avoid any tax increase if the head of household schedule is eliminated.

28. See, Eugene Steuerle, “The Tax Treatment of Households of Different Size,” in Rudolph G. Penner (ed), Taxing the Family, Washington DC: American Enterprise Institute, 1993. See also Eugene Steuerle, “Taxation of the Family,” statement before the Committee on Ways and Means, U.S. House of Representatives, April 15, 1997, and Eugene Steuerle, “Child Credits,” statement before the Finance Committee, U.S. Senate, August 7, 1995.

The Communitarian Network
2130 H Street, NW, Suite 703
Washington, DC 20052
202.994.6118
comnet@gwu.edu