CRS Report Compares Marriage Penalty Bills

                    MARRIAGE PENALTY LEGISLATION:

[1] President Bush's tax proposal and H.R. 6, (as passed by the House), have different approaches to the marriage penalty. The Administration proposal, in addition to rate changes, has a second- earner deduction for 10% of income (up to $30,000) earned by the lower earning spouse. H.R. 6 would increase the standard deduction and width of the 15% rate bracket for joint returns to twice the size of singles, eliminating the penalties for taxpayers in the 15% and 28% brackets but adding to any marriage bonuses. This report compares these alternative proposals.

[2] Taxes can go up or down as a result of marriage, depending on the income of the two spouses. These penalties and bonuses arise from the progressive tax structure and the decision to impose taxes on a household basis.

[3] For much of the middle class, marriage penalties are low. For couples without children, the maximum marriage penalty at 2001 income levels for the 60% of taxable returns subject to the 15% marginal rate in 1997 is $225; most couples that itemize have no penalty at all. The second-earner deduction in the administration proposal virtually eliminates the marriage penalty for these couples. Even in the 28% rate, which covers another 26% of taxable returns, the second earner deduction along with the flatter rates results in no or negligible penalties. Overall, these taxpayers account for the middle 75% of joint returns. Taxpayers with children could have small remaining penalties, depending on how penalties are defined. Most of those in the 15% bracket who might have still have penalties that are significant relative to income are removed from the tax roles entirely through the additional child credits.

[4] Lower income couples may incur penalties through the Earned Income Credit (EIC) under either proposal, although it is difficult to address EIC marriage penalties and bonuses. The 12% of taxpayers in higher brackets may have remaining penalties. A large fraction of these returns do not have large penalties because their marginal tax rate is the flat capital gains tax rate. Any penalties are substantially reduced or eliminated, however, due to the lower rate structure in the Administration proposal.

[5] The important difference between H.R. 6 and the Administration proposal is that H.R. 6 expands marriage bonuses in most cases while the Administration proposal generally reduces them due to flatter tax rates. H.R. 6 combined with rate cuts would increase bonuses in comparison to the Administration proposal and, in many cases, in comparison to current law.

[6] Measurement of the marriage penalty for couples with children is highly dependent on assumptions about the allocation of children for tax purposes. When married couples are compared with cohabitating singles, where issues of incentives and fairness suggest attention be focused, cohabitating singles are much less likely to have children and when they do, tend to have low incomes that are unlikely to be subject to regular income tax. Thus, the issues of measuring the marriage penalty for families with children are less important than they may initially appear.

                              CONTENTS

[8] Taxes can go up or down as a result of marriage, depending on the income of the two spouses. These bonuses and penalties arise from the progressive tax structure and the decision to impose taxes on a household basis, a decision that may have historical roots but is also amenable to adjusting the tax burden to reflect relative ability to pay (or relative standards of living). ¹ In particular, the marriage penalty or bonus is affected by the allocation of income between the two spouses, by the relative size of standard deductions, the relative bracket widths of the tax structure, the earned income tax credit, and a variety of other features of the tax system (e.g. the importance of capital gains income which is taxed at a flat rate, and the points at which various phase outs occur).

[9] There are several different approaches to reducing the marriage penalty. One proposal uses the approach of increasing standard deductions and the first rate brackets for joint returns to twice that of singles, an approach that reduces penalties but magnifies bonuses. This proposal was passed by the House (H.R. 6) in April, 2001 and also by the full Congress in 2000. Historically, the tax system employed a second-earner deduction, which tended to be more focused on (although not precisely targeted to) reducing penalties. A third option, which was adopted by the Senate in 1999, was optional joint filing, which precisely targets marriage penalties and does not change bonuses (assuming that these can be defined), but adds complexity.

[10] President Bush's proposal includes a mix of provisions: a second earner deduction, a general lowering and flattening of the tax rate structure, and a new lower bracket that has a width for joint returns twice that of singles. The second earner deduction tends to reduce penalties without expanding most bonuses; the rate cuts reduce penalties and bonuses (which, after all, arise from progressive rates). The relationship between the bracket widths for the new lower 10% rate increases bonuses without affecting penalties.

[11] The House has separately passed the rate cuts in the President's proposal as H.R.3, but has not combined the two bills, and it is not clear at this point what form an omnibus bill would take. If the original rate cuts are retained, either a second earner deduction or a general expansion of rate brackets might be considered. This paper examines the rate cuts with the second earner deduction and the consequences for marriage penalties and bonuses as contrasted with the marriage penalty legislation contained in H.R. 6, with commentary on optional separate filing adopted by the Senate in 1999.

ASSESSING MARRIAGE PENALTIES OR BONUSES

[12] It is surprisingly difficult to measure marriage penalties and bonuses. The measure of the marriage penalty requires a counterfactual: what would the combined taxes of the two spouses be if they were not married? Unfortunately, there is no simple answer to this question because marriage itself tends to be a life-changing event. Indeed, the answer may be different if the question is: "What would the tax be if the spouses had never married?" rather than "What would the tax be if the couple divorced?" These issues become much more important for certain types of married couples. Most couples would not have had children without marriage, and having children can dramatically reduce tax liability. In divorce, who gets custody of children can alter tax liability. Never marrying would probably also have altered career paths (and thus the amount of earned income of each partner), savings behavior (and thus the amount of unearned income), the allocation of savings and consumption into different types (including tax exempt forms such as owner occupied housing that also tend to affect the choices of itemizing deductions). Getting divorced would probably also alter these effects, unless the divorce were solely for purposes of tax minimization.

[13] The decision about how to measure the marriage penalty does not alter the effects of the two proposals (which are still determined with reference to the married couple's joint circumstances), but rather the original magnitude and remaining magnitude.

[14] In the next section we calculate the marriage penalty in general for couples without considering the effects of children, where there is much less uncertainty about how to measure the penalty. The following section discusses the issues surrounding marriage penalties for couples.

MARRIAGE PENALTIES AND BONUSES: COUPLES WITHOUT CHILDREN

[15] Despite all of the attention focused on marriage penalties, it may be surprising to know that for the middle class, marriage penalties are small or nonexistent, particularly in cases where the penalty is well-defined. According to data from the Internal Revenue Service Statistics of Income, about 60% of joint returns that have tax liability before credits paid taxes at the 15% rate in 1997; for these taxpayers, the marriage penalty can be no larger than $225 (measured at 2001 income levels). For taxpayers who itemize and whose itemized deductions are at least as large as the combined standard deduction for two singles, no penalty exists at all.

MARRIAGE PENALTIES FOR THE "MIDDLE CLASS"

[16] The "middle class" is typically defined very broadly and in this section we discuss effects that cover taxpayers with marginal tax rates of 15% (covering 60% of taxable returns and 51-52% of all returns) and with marginal tax rates of 28% (covering 26.2% of taxable returns and 23% of all returns). Overall this broad range covers all but the bottom 12-14% of returns that have a zero marginal tax rate and the top 12% or so. (For the bottom, the 12% refers to no liability before credits and the 14% with no liability after credits; since the earned income credit, but not other credits is refundable, the number with a zero marginal tax rate falls in between.) Thus, roughly speaking, these taxpayers represent the broad range of the middle three quarters of the distribution of joint returns.

[17] The following examples, based on 2001 income levels, assume that all income is taxable, and assume a constant rule for the choice of itemizing deductions, so there is a switch at the point of itemizing. Note that these assumptions overstate the marriage penalty as a fraction of economic income, particularly at the top and bottom of the distribution. At these ends of the distribution, untaxed or preferentially taxed income (transfer payments at the lower end and capital gains and untaxed income, such as tax exempt interest at the higher end) is common. The middle class also has higher incomes than adjusted gross income indicates, primarily due to untaxed fringe benefits. Note also that the marriage penalties assume evenly divided income where marriage penalties are usually at their maximum, especially at higher levels of income. ²

[18] Table 1 shows the marriage penalties from the regular income tax for returns falling in the 15% bracket, representing slightly more than the bottom two thirds of our "middle class." (It excludes the earned income credit (EIC).) These returns assume income is evenly split between the couples. The income points are chosen to show peaks and valleys of the penalty as a percentage of income. The highest level of the marriage penalty as a share of income is reached at the exempt level for singles, $14,900 in 2001. But for most joint returns, the marriage penalty is relatively small. Both the second earner credit, and lower rates, in the Administration proposal would reduce the marriage penalty (and in cases where taxpayers itemize, turn it into a bonus). H.R. 6 would eliminate the penalty, as would optional filing.

[19] Table 2 shows the maximum marriage penalties for returns falling in the current 28% bracket. These returns can be affected by the width of the 15% bracket as well as the standard deduction, and their marriage penalties peak at about 1.6% of income (again, assuming that incomes are evenly divided).

    TABLE 1: MARRIAGE PENALTY AS A PERCENTAGE OF INCOME, RETURNS IN

[21] Because the second earner deduction is actually larger the more evenly divided incomes are, there could be a smaller effect, as noted in the footnote to Table 1. Regardless of how it is calculated, however, it is clear that the Administration proposal virtually eliminates the small marriage penalty that currently exists.

[22] The same is true of the 28% bracket shown in Table 2. The maximum marriage penalty as a percentage of income is reached at the top of the 28% bracket for two singles (and income of $73,768). This penalty is reduced by the second earner deduction until it is only 2/10 of a percent of income. Again, H.R. 6 in most cases eliminates the marriage penalty, although it does not do so entirely because of the phase outs of itemized deductions and personal exemptions that begin at $132.750.

[23] When looking at the marriage penalty, therefore, it is clear that both the Administration proposal and H.R. 6 are approaches that eliminate or virtually eliminate these middle class marriage penalties, with H.R. 6 being more precise, but with the Administration proposal being quite successful.

[24] The more important differences arise from the effects of the two proposals on marriage bonuses. Table 3 illustrates the marriage bonus (at its maximum) for the relatively common case of the one-earner family. As this table illustrates, marriage bonuses tend to be larger than penalties, and are a significant fraction of income. Bonuses begin at the exempt level for a single individual. Bonuses reflect both the higher standard deduction for joint returns compared to a single individual, as well as the additional personal exemption associated with the non-earning spouse.

[25] The Administration proposal and its rate cuts lower the marriage bonus at very low income levels, because it lowers the rates from 15% to 10%. However, the proposal creates a new type of bonus because the new 10% rate bracket is twice as wide for joint returns as for singles, a feature that may have been chosen to avoid increasing the marriage penalty. This new bracket, in fact, follows the general approach used in H.R. 6, but limited to the new 10% bracket. In this income category, however, H.R. 6 increases the marriage bonus at lower and moderate income levels only because it increases the standard deduction. Hence at the points where deductions are itemized, H.R. 6 has smaller bonuses than the Administration proposal. Optional filing, of course, would not affect bonuses.

[26] In addition to the comparison of H.R. 6 and the Administration proposal, a fourth column has been added calculating the effects of combining H.R. 3 and H.R. 6, that is, keeping the rate structure of the Administration proposal but altering the standard deduction and 15% rate bracket to make them double the width of the single bracket. These results indicate that choosing this approach, given the rate structure, would lead to higher bonuses than those with the second earner deduction and higher bonuses than in present law in many cases.

[27] Table 4 shows the one-earner bonus for returns in the 28% bracket. In this case the bonuses tend to be smaller for the Administration proposal than either for current law, or for H.R. 6. That effect occurs because of the lower rate, 25%, which offsets the effect of the 10% bracket H.R. 6, however, expands the bonus because it increases the 15% rate bracket for married couples to twice that of singles.

TABLE 3: ONE EARNER MARRIAGE BONUS AS A PERCENTAGE OF INCOME, RETURNS IN THE 15% BRACKET, NO CHILDREN

 Income  Current Law      Administration     H.R.6      H.R. 6 &

LOWER INCOME COUPLES AND THE EARNED INCOME CREDIT

[28] For lower income individuals, the tax system still has an impact through the earned 'income credit, and marriage penalties can arise from this system as two individuals who marry can be phased out of the EIC. For example, a married couple without children with equal earnings and total income at the filing threshold would have a marriage penalty of 4.6% of income due to the phase-out of the EIC. Of course, an equal split at these low combined income levels would imply very low annual earnings, probably reflecting part time work. These measures would, however, be much larger in the case of couples with children where the EIC is much larger.

[29] Marriage bonuses can also occur with the EIC, if an individual without children and earnings marries someone with children and little or no earnings.

[30] None of the proposals includes significant changes in the earned income tax credit. H.R. 6 has a percentage increase in the earned income tax credit for joint returns as compared to single and head of household, while the bill that passed last year increased the phase-out threshold by $2,000. (A proposal being considered in the Senate Finance Committee would increase the phase-out threshold by $3,000). These proposals would not have very large effects, and both would increase bonuses slightly while reducing penalties slightly.

[31] This decision may reflect in part how difficult it is to address the marriage penalty in the case of the EIC where targeting poor families via their total income is paramount. If we simply allowed even the most targeted of the proposals, optional filing, and even ignoring the consequences of children, we could find couples with very large incomes becoming eligible for the EIC. (A similar point could be made for an additional "second earners" credit.) For example, a family with a combined income of $100,000 where one spouse earned a low salary (say $8,000) would be eligible for the EIC which does not phase out until slightly over $10,000. This issue would become even more complicated given the larger EICs with children, where the phase out ends at slightly over $28,000 for one child and $32,000 for two children. Thus, relatively wealthy families could become eligible for the EIC.

[32] Nor do the alternatives become more palatable. To double the phase out points of the EIC, using the same approach as H.R. 6 for marriage penalty relief in the regular income tax system, would mean that families with children and with $60,000 or more would be eligible for the credit. Nor is the approach of increasing the credit itself very effective, since the problem is the phase out and not the level of the credit.

HIGHER INCOME COUPLES

[33] Higher income couples, representing the top 12-13% of the income distribution, present their own problems in assessing the marriage penalty, even in the absence of children. First, as incomes increase, the possibility of excluded income and of capital gains income taxed at flat rates increases. For couples in the 15% bracket about 3% of income is from capital gains; for couples in the 28% bracket perhaps 6% to 10% of income is from capital gains (based on the closest matches to income categories). At higher income categories, the share that was in capital gains rises to 20%, 30% and 60% of income. In 1997, half of the individuals in this upper part of the distribution had a top marginal tax rate that was due to capital gains. Any penalties that arise from existing law would, therefore, be lower as a percentage of income than the penalties listed in table 2, and hence effectively eliminated by either approach for half of these taxpayers.

[34] Secondly, problems about the division of non-wage and unearned income become more important, particularly at the very high end. Adjusted gross income ranges from $141,000 to $171,000 for the 31 % bracket, and up to $368,000 for the 36% bracket. While about 80% of income in the $50,000 to $100,000 class is from wages, only 70% of the $100,000 to $200,000 class is from wages, and the share declines until it reaches less than 30% for income over $1,000,000.

[35] In addition, an equal distribution, or even close to equal, distribution of income between the spouses becomes much less likely in the upper end of the distribution, so that illustrations based on even incomes would be much less representative of typical two earner couples. Thus, while calculations of bonuses might be more realistic, calculations of penalties would be misleading.

[36] Despite these uncertainties that make illustrative examples problematic, it is clear that the penalties in most of these higher income categories are not very large, and will decline. For the 31% bracket, which includes about 3.3% of joint returns,the maximum penalties assuming even division of income and full taxability begin at about 0.9% of income, rise to a peak of 1.3% and then fall to 1.2% at the top of the bracket; the administration proposal will reduce those penalties to 0.3%, largely because the former 28% and 31% brackets are now combined into a single 25% bracket. Most of the penalty that would now apply in this group is due to the itemized deduction phase out and not rate brackets. (A proposal by Senate Finance Committee Chairman Grassley and Ranking Member Baucus would make the former 31% bracket subject to a 28% rate, however, which would restore the marriage penalty effect because it reflects the differences in the two rates. The same proposal, however, would increase the itemized deduction phase-out start point, reducing the penalty and also includes the H.R.6 proposals). H.R.6 would actually leave slightly larger penalties than the Administration proposal in many cases, beginning at 0.1% but rising fairly quickly, with a peak of 0.7%, then declining to 0.6%. These penalties occur because the other brackets are not corrected; penalties would be quite similar to those in the Administration proposal, however, if H.R. 6 were combined with H.R. 3. Bonuses would also be reduced by the Administration proposal in general, and would be increased by H.R. 6.

[37] At higher levels, penalties would remain, although they will be reduced by both bills, but these remaining taxpayers account for only 3% of taxpayers.

COUPLES WITH CHILDREN

[38] As noted in the previous section, couples with children present serious problems in modeling marriage penalties and bonuses. The presence of children in a household can dramatically alter tax liabilities. Children justify a head of household rate schedule which has wider brackets than the single return, implying a larger penalty and smaller bonus as compared to joint return than would be the case in comparing single and joint returns. Children increase personal exemptions and, if under 17, make the taxpayer eligible for a child credit. Earned income credits, which reach much further up the income brackets and are much larger, accrue with children. Thus, how children are divided between the spouses can have profound consequences for tax liability.

METHODS OF ALLOCATION

[39] At least five potential ways of allocating children's tax consequences might be considered and several have been specifically used in studies. Some of them are most plausible when thinking of a married couple divorcing rather than never marrying; others are most plausible when thinking of a couple never marrying.

(1) The Tax Minimization Model. This model assumes that the children are allocated in a way that minimizes tax liability. If a couple has two children and has similar incomes, for example, each would take custody of one of the children and move to a separate household so that both could obtain head-of-household status. This was the basic approach used initially in the economics literature and by the Congressional Budget Office in its study of the marriage penalty. ³ This choice tends to lead to the largest measures of penalties and the smallest measure of bonuses. Obviously, as an optimization strategy, this approach has some shortcomings because it requires not only the emotional and social costs of a sham divorce but also increased economic costs (maintenance of two households) to achieve tax minimization. Except for very high income taxpayers, such a strategy would be unlikely to be followed.

(2) Resource Pooling Tax Minimization Model. This model, developed by Treasury, assumes that tax minimization is modified to keep economic costs unchanged, so that divorced individuals continue to share the same household and otherwise maintain their economic relationships; i.e. they would be unmarried cohabitators. ⁴ Only one parent (the one with the highest income) would be able to claim head of household status.

(3) Divorce Model. This model, used as an alternative by both Treasury and CBO, would allocate children in a way typical of most divorces, where the children are allocated to the lesser earning spouse (who is typically the mother). Treasury finds that this model produces an aggregate $28.9 billion net bonus for all couples, compared to the $1.6 billion net penalty they find in their tax minimization model.

(4) Ignore Head of Household Status. In this view, head of household status may be seen as a special dispensation to help single parents and should not be viewed as the standard against which to measure the marriage penalty. The allocation of children (and their personal exemptions and credits) would still matter for measuring bonuses and also for measuring penalties where incomes tend to be disparate. One might say this view is implicit in H.R. 6 and other congressional proposals where the changes are made to match the single return, not the head of household return. A similar argument might be made for child credits.

(5) Ignore Children. This view would argue that the counterfactual for most couples would be that they would not have had children had they not married. In this case the marriage penalties and bonuses would be the same as in the previous section, except there might be recognition that married couples with children will not have tax liabilities at higher incomes.

[40] Because it is impossible to assess these bonuses we do not present tables of penalties and bonuses as we did for single individuals, and it is t impossible to make general observations. Nor is it easy to choose a model, in large part because the rationale for the concern about the marriage penalty. In the next section, however, we suggest that common rationales for concern about the marriage penalty would suggest a focus on lower income individuals and then discuss the parameters of these bonuses and penalties under various alternatives.

RATIONALES FOR CONCERN ABOUT THE MARRIAGE PENALTY

[41] As noted at the end of the previous section, part of the uncertainty surrounding the measurement of a marriage penalty is that the reason for being concerned about the penalty has not been clearly articulated. There are two potential reasons to be concerned about the marriage penalty: fairness and incentives.

[42] If the marriage penalty is a case of concern about fairness, then there is a trade-off. There is considerable evidence to suggest that married couples are already being taxed somewhat more lightly than singles with the same ability to pay, but that in general the system is relatively even handed. ⁵ (Families with children are treated more generously whether headed by married couples or singles.) Thus, any unfairness in the marriage penalty must arise from the comparison of married couples with cohabitating couples. However, the numbers of these latter groups are very small. According to March 1998 Census data, there were 197 million people over 18, and 110 million were married. However, of the 87 million unmarried individuals (some never married, some widowed, and some divorced) only about 8.5 million were living with unmarried partners of the opposite sex. Thus, at most one would be increasing inequities between the 110 million married couples and the slightly under 80 million single individuals, in order to more evenly treat the 110 million relative to the 8.5 million. Such a fairness argument does not hold up well, especially since some married couples, in fact a significant fraction, receive a bonus from being married. Using a "divorce" model (discussed above), CBO found that 37% of married couples have penalties and 60% have bonuses. Thus, simply lowering tax rates for joint returns would probably magnify the existing average favorable treatment of married couples relative to cohabitating individuals. It would stress the pursuit of "fairness" between a minority of married couples and a tiny minority a' singles, at the cost in increasing the discrepancies between much larger groups.

[43] Another reason to be concerned about the marriage penalty is the incentive structure which, again, may cause individuals to cohabitate rather than marry. This concern may be more legitimate, but it is probably of much greater importance for certain types of couples. Of course, the relatively small number of cohabitating couples, and the fact that most of them are probably cohabitating for reasons other than tax consequences suggest that not very many people are discouraged from marriage or encouraged to engage in a divorce, because of the tax consequences. In particular, it is likely that most middle income individuals would not see the financial incentives as large enough to divorce, and that the presence of children, in particular, would make that outcome less likely.

[44] The characteristics of cohabitating couples also differ from those of married couples in general. About half of married couples have children, but only about a third of cohabitating couples have children; cohabitating couples tend to be relatively young. While no data directly report the poverty rates among this group, it is likely that a larger fraction of families with children is not subject to regular income tax liability. These families would file tax returns as heads of households or singles, and the fraction of these groups, particularly heads of household, that do not pay taxes, is higher than in the case of singles. In 1997, about 35% of head of household returns had no tax liability before credits, while only 12% of joint returns had no tax liability; 21% of single returns had no tax liability.

[45] Data on child poverty are also instructive. In a study that focused on cohabitating couples based on the 1990 census, the poverty rate even after taking into account earnings of both partners was 26.9%, compared to a poverty rate of 5.8% for children whose parents are married. ⁶ If even for those not in poverty, a large fraction of these families tended not to be much above the poverty line. Another 19% of children would be classified as poor by considering only one of the partners, and when income was combined, the average income was only 1.75 times the poverty line. Since the current tax system excludes families with children well above the poverty line from tax and since public assistance received by these families and counted as income would not be taxable, it is likely that a significant fraction of these families do not incur a regular income tax liability.

[46] These data do suggest that if our concern is about a comparison to cohabitating couples, there are two important groups: couples, mostly young, who do not have children and, in the case of couples with children, mostly lower income couples.

[47] The Administration proposal will further remove these families with children from concerns about the tax system, because of the lower rate of 10% and the increased child credit. Under current law, a family with one child would have no tax liability until income reached $19,633 (assuming that all income is taxable); a family with two children would not have tax liability until income reached $25,866. Under the Administration proposal, with the larger child credit, these amounts will rise to $27,684 and $38,456, amounts close to the poverty level, and closer to the income levels where itemized deductions tend to displace standard deductions. Thus, a marriage penalty or bonus will be feasible for only a small slice of families in the 15% bracket whose incomes are high enough to be taxable but who do not itemize. And even in this case, considering the head-of- household schedule for one of the spouses, the marriage penalty can be no larger than $540. Looking at maximum values, for a $30,000 couple, the penalty would be 1.8%, which would be reduced to $220 (0.7% of income) by the Administration proposal and to $315 (1.1%) under H.R. 6. These amounts will decrease as incomes rise and, again, become subsidies for itemizers. These taxpayers, along with those with no liability, account for approximately the bottom two-thirds of the joint returns.

[48] These couples may experience a penalty due to the earned income credit, but, as noted earlier, none of the proposals are really focused on the EIC and it is difficult to devise a way of correcting for marriage penalties without undermining much of the purpose of the EIC. And higher income individuals are not particularly comparable to most cohabitating couples with children.

CONCLUSION

[49] This analysis of the effects of legislative proposals on the marriage penalty indicates that for the middle class, the marriage penalty, which is already relatively small, will be essentially removed by the rate reductions and second earner exemptions in the administration proposal, capturing essentially the same penalty elimination objective of the original marriage penalty legislative proposals in the previous Congress. The proposal will also reduce marriage penalties in the higher brackets by a larger amount than H.R. 6 standing alone. This result is a natural outcome of the rate reduction and flattening in the President's proposal, and explains why marriage penalty relief in the form of a second earner deduction was eliminated when rate[s] were flattened in 1986.

[50] H.R. 6, the congressional marriage penalty proposal, would, however, increase existing marriage bonuses, while the administration proposal would reduce them. H.R. 6 combined with the rate cuts and credits portion of the proposal passed by the House as H.R. 3 would increase marriage bonuses compared to the Administration proposal with the second earner deduction, and in many cases would increase bonuses compared with current law.

² Because the rate brackets are discrete spans of income, the marriage penalty can be the same over a span of aggregate income even with somewhat different divisions of income. For example all families with two earners in the 15% bracket pay the same $225 penalty as long as they do not itemize deductions and as long as the income of the lower earning spouse is as large as the singles standard deduction plus personal exemption. For taxpayers in the 28% bracket, the maximum penalties are encountered only if the second earner's income is large enough to absorb the 15% singles rate bracket, etc. The surest way, however, to be sure that one is constructing a maximum penalty is to assume equal division of incomes.

³ U.S. Congressional Budget Office, For Better or for Worse: Marriage and the Federal Income Tax. Washington, DC, June 1997.

⁴ Office of Tax Analysis Paper 82, Defining and Measuring Marriage Penalties and Bonuses, November 1999. See http://www.ustreas.gov/ota/ota82_revised.pdf.

⁵ See tables in CRS Report RL30419, The Marriage Tax Penalty: An Overview of the Issues. Note also that the argument can be made that one-earner couples are already favorably treated because of the extra value of household production due to the spouse not formally employed in the labor market being implicitly exempt from taxation.

⁶ Marcia Carlson and Sheldon Danziger. Cohabitation and the Measurement of Child Poverty. Review of Income and Wealth. Vo. 45, June 1999, pp. 179-191.