RECENT DEVELOPMENTS IN THE MARRIAGE TAX: A COMMENT AND DECOMPOSITION A. J. CATALDO, II

Transcription

1 RECENT DEVELOPMENTS IN THE MARRIAGE TAX RECENT DEVELOPMENTS IN THE MARRIAGE TAX: A COMMENT AND DECOMPOSITION A. J. CATALDO, II * Abstract - This study uses the 1989 Internal Revenue Service Statistics of Income public use file to produce measures of frequency and magnitude for several marriage taxes. These measures are generated, separately, for itemizer and nonitemizer taxpayers, while holding taxpayer-specific marginal tax rates constant. Separate measures are also provided by taxpayer adjusted gross income class. Distinguishing between marriage tax penalties (MTPs) and those taxes/credits associated with qualifying dependency exemptions, findings suggest that studies failing to make this distinction may overstate the magnitude and frequency of MTPs. INTRODUCTION In a recent issue of this journal, Feenberg and Rosen (1995) provided the results of their study of marriage tax penalties (MTPs) and marriage tax * Department of Accounting, R. B. Pamplin College of Business, Virginia Polytechnic Institute and State University, Blacksburg, VA bonuses (MTBs) for different classes of taxpayers. Their study provided measures relevant to periods both before and after President Clinton s Omnibus Budget Reconciliation Act of 1993 (OBRA93). Using the 1989 Statistics of Income (SOI) data produced by the Internal Revenue Service (IRS), which they aged to project income growth from the 1989 to 1994 tax years, the methodology and analysis were comparable to those used by Rosen (1987). However, their study did not include a decomposition of the MTP elements, as suggested by McIntyre (1988). 1 This paper provides for the separation of major MTP elements, using the same database used by Feenberg and Rosen (1995). The impact of earned income tax credits (EITCs) is excluded from this decomposition. 2 The development of EITC measures would have required the use of a divorce -based methodology comparable to that employed by Rosen (1987) and Feenberg and Rosen (1995). The measures produced in this paper, therefore, excluded the possibility of generating increased MTPs via the 609

2 NATIONAL TAX JOURNAL VOL. XLIX NO. 4 creation of additional, preferentially treated head of household (HH) taxpayers. 3 The results indicate that the separate tax rate tables (or rate effects), required to be used by married (as opposed to single (SGL)) taxpayers, remain the most punitive of the penalties associated with marriage. Nonitemizers, faced with different standard deduction amounts (or base effects), and Social Security (SS) recipients also face substantial penalties, with other MTPs averaging only relatively small amounts. DATA AND METHODOLOGY This decomposition of separate MTPs uses the same 1989 IRS SOI data used by Feenberg and Rosen (1995). These measures were developed from 31,983 tax returns for married taxpayers filing jointly (MFJ) and differs from their study in three respects. First, the EITC has been excluded from this decomposition. As Feenberg and Rosen (1995) suggest, the post-1993 expansion of the EITC is a major contributor to the generation of MTPs. Furthermore, the EITC, prior to OBRA93, did not provide EITC benefits to taxpayers without dependents (Cataldo, 1995a). Therefore, inclusion of the EITC for any pre-1994 tax year dependent analysis of MTPs confuses two separable decisions: (1) the decision to marry and (2) the decision to have children. 4 Second, also excluded are MFJ taxpayers in adjusted gross income (AGI) classes below $1 and above $100,000. Approximately 94 percent of MFJ taxpayers used by Feenberg and Rosen (1995) remain in the sample for the present study. 5 These exclusions were made for several reasons. The 1989 IRS SOI provides evidence that taxpayers with AGIs below $1 are frequently highincome taxpayers and generate disproportionately large mean EITCs. These taxpayers may be engaged in aggressive, EITC- (and, therefore, MTB-) maximizing tax-planning strategies (Cataldo, 1995b). Taxpayers with AGIs above $100,000 are excluded to avoid itemized deduction and personal exemption phaseouts and the related distortions which might otherwise occur were this relatively small sample of highincome taxpayers included in the analysis. Finally, returns are analyzed separately for itemizer and nonitemizer taxpayers. Feenberg and Rosen computed... the joint tax liability of each couple... under the assumption that a divorce occurs (1995, p. 95). The divorcebased methodology employed by Rosen (1987) and Feenberg and Rosen does not distinguish between itemizer and nonitemizer taxpayers. This may mask the impacts of MTP/MTB overstatements/understatements via the net results of arbitrary (though systematic) allocations of deductions to the spouse with the higher earnings (Rosen, 1987, and Feenberg and Rosen, 1995, respectively). It also results in the conversion of marginal itemizers (MFJ) to nonitemizers (SGL and/or HH). 6 TYPES OF MTPs Measures are presented for five MTPs. They represent the most easily identified and quantified marriage taxes and are summarized separately for itemizer and nonitemizer taxpayers in Table 1. Table 1 measures were computed prior to any adjustment for sample weights. 7 With the exception of net capital losses, the penalties contained in Table 1 are directly or indirectly AGI based. These MTPs were developed by recalculating 610

3 RECENT DEVELOPMENTS IN THE MARRIAGE TAX TABLE 1 MEAN MTPS (UNWEIGHTED) FOR AFFECTED MFJ TAXPAYERS AND BY MTP TYPE (NONITEMIZER AND ITEMIZER TAXPAYERS ANALYZED SEPARATELY) Number of Taxpayers Affected, Percentage Affected [percent], and Mean MTP ($) for Affected Taxpayers (n = varies) Mean Marriage Tax Penalties: Affected Taxpayers Nonitemizers the tax, after doubling the exemption or AGI-based amount for two single taxpayers, as follows: [ SINGLE 2 ] MARRIED = 1 MARRIAGE TAX PENALTY Itemizers Total sample size 15,598 16,385 Base 15,598 1,367 [100%] [8%] ($149) ($18) Rate 2,977 8,975 [19%] [55%] ($654) ($687) Social Security (SS) [6%] [3%] ($1,366) ($1,145) Net Capital Loss (NCL) [2%] [4%] ($401) ($527) Alternative Minimum Tax (AMT) neg [1%] ($1,526) ($2,867) 1 Table 2 provides a summary hierarchy of these 1 MTPs, before and after IRS SOI sample weights have been applied (as indicated). Also provided in Table 2 is a breakdown of their frequency of occurrence, by MTP type and AGI class. The algorithms used for the computation of the measures presented in Tables 1 and 2 are contained in the Appendix. The marginal tax rate (MTX) was held constant for all computations other than those specifically targeted at quantifying tax rate schedule based MTPs. The MTPs presented in Tables 1 and 2 are summarized below. Base Effects (Base). The most frequent penalty associated with marriage has been referred to as the base effect. It evolves from the fact that the (1989) standard deduction amount available to married taxpayers ($5,200) is less than that available to two single taxpayers ($3,100 2 = $6,200). This MTP was calculated by increasing the standard deduction amount available to all relevant taxpayers by $1 thousand and holding the marginal tax rate constant. As 2 Table 1 illustrates, this MTP affects all nonitemizer taxpayers. The 2 Rosen (1987) and Feenberg and Rosen (1995) divorce-based studies resulted in the generation of HH taxpayers. The measures provided in Tables 1 and 2 do not. The standard deduction amount available to nonitemizer HH taxpayers was $4,550 for the 1989 tax year. This amount exceeded that available to SGL taxpayers by $1,450 ($4,550 less $3,100), constituting a singles or no dependent tax penalty. The HH filing status is not a function of marriage, but a function of the availability of qualifying dependents. Rate Effects (Rate). The largest penalty for married taxpayers remains that resulting from the different marginal tax rates, imposed and beginning with the 1971 tax year (Brozovsky and Cataldo, 1994). It was this change, the move to separate rate tables based solely on the taxpayer(s) filing status, which first raised concerns regarding the potential for material inequities resulting (solely) from the decision to marry. 611

4 NATIONAL TAX JOURNAL VOL. XLIX NO. 4 TABLE 2 MEAN MTP BY TYPE FOR ALL MFJ TAXPAYERS (N = 31,983) Unweighted Mean MTP Weighted Mean MTP Rate $254 $197 Base $74 $73 Social Security (SS) $58 $58 Net Capital Loss (NCL) $13 $15 Alternative Minimum Tax (AMT) $10 3 $1 Percentage of MFJ Taxpayers Affected by each MTP (by MTP Type and AGI Class) AGI Class Rate Base SS NCL AMT $1K $10K 92.7% 0% 0% 2.1% 0.2% $11K $20K 86.9% 0% 0.1% 1.7% 0.1% $21K $30K 75.8% 0% 4.6% 2.1% 0.1% $31K $40K 60.8% 0% 9.9% 2.0% 0.2% $41K $50K 42.4% 48.5% 9.8% 2.2% 0.2% $51K $75K 22.6% 91.9% 3.2% 3.4% 0.4% $76K $100K 10.8% 97.7% 0% 6.2% 1.8% Overall 53.0% 37.4% 4.5% 2.7% 0.4% The SGL, HH, and MFJ 15/28 (28/33) percent breakpoints were $18,550, $24,850, and $30,950 ($44,900, $64,200, and $74,850), respectively. Again, the preferential treatment afforded HH taxpayers represents a singles or no dependent tax penalty. The use of a divorce-based methodology (and the generation of HH filing status taxpayers) would result in something other than the inclusion of penalties and bonuses resulting only from marriage. Because the generation of HH filing status taxpayers is not represented in the measures provided in Tables 1 and 2, the confounding of marriage- and dependency-based penalties and bonuses is, again, avoided. Social Security (SS). The next largest penalty is that associated with the taxation of SS benefits. SS benefits were taxed at up to 50 percent (through 1993) for single/married taxpayers with provisional income levels above $25/$32 thousand. The calculation of SS-based MTPs was performed by recalculating the tax, using twice the (1989) single taxpayer exclusion amount ($25,000 2 = $50,000) and holding the marginal tax rate constant. Taxable SS, as originally reported, was then subtracted from this recalculated amount to restate that amount of taxable SS benefits to be included in the taxpayer s AGI. Approximately 35 percent of the taxpayers reporting gross SS benefits experienced some level of MTP. The unadjusted incidence of some taxable amount of SS benefits for itemizer taxpayers (944 or 79.4 percent) was nearly double the amount observed for nonitemizer taxpayers (1,176 or 40 percent). Net Capital Loss (NCL). NCLs are subject to statutory limitations of $3,000 per year, per return, constructively making twice this amount available to single taxpayers. 8 This MTP was developed by recalculating the tax after (potentially) doubling this $3,000 limitation, while holding the marginal tax rate constant. 612

5 RECENT DEVELOPMENTS IN THE MARRIAGE TAX Alternative Minimum Tax (AMT). The amount of the 1989 AMT was determined to be any excess of the tentative minimum tax over the regular tax. The exempt portion of AMT income for the 1989 tax year was $40 ($30) thousand, less 25 percent of AMT income in excess of $150,000 ($112,500) for married (single) taxpayers. The calculation of the MTP for taxpayers with AMT was made through recalculations at twice the exempt amounts available to single taxpayers. Again, to avoid the confounding of rate-based MTPs, the marginal tax rate was held constant. THE INCIDENCE OF OTHER PENALTIES ASSOCIATED WITH MARRIAGE Table 3 summarizes those MTPs developed and not developed in this paper. Discussion of selected, but not quantified, MTPs is, however, warranted. Their relevance, given the assumptions that must be made to develop reasonable algorithms, is discussed below. Child Care Credit. The child care tax credit, like the EITC, is a function of the presence or absence of dependents. It is not an MTP, per se. This tax credit is AGI-based and ranged from 30 to 20 percent for the 1989 tax year. Approximately 9.5 percent (3,053) of the 1989 IRS SOI based MFJ taxpayers included in this sample claimed the child care tax credit. Deferred Compensation Limitations. Active participants in certain employersponsored retirement plans were subject to reductions (or the complete elimination) of their IRA deductions. Reductions occurred when AGIs exceeded $25 TABLE 3 MTP COMPONENTS BY INTERNAL REVENUE CODE SECTION (IRC ) Marriage Tax Penalty IRC Notes Not Developed: Earned Income Credit 32 3 Summarized in Tables 1 and 2: Different tax rates tables (Rate) 1 6 Different standard deduction amounts (Base) 63(d) 2, 5, and 6 Social Security benefits (SS) 86 1, 4, and 6 Net capital loss limitation (NCL) and 5 Alternative minimum tax (AMT) 55 6 Not Developed: Child and dependent care credit 21 3 and 6 Deferred compensation limitation 219 1, 5, and 6 Moving expenses and 5 Gain on sale of personal residence and 4 Credit for the elderly 22 3 and 6 Expense election limitations 179 Unemployment insurance income 85 Accounting period elections 6013 Passive activity loss limitations 469 Notes: (1) = Above the line item, included in AGI. (2) = Below the line item, excluded from AGI. (3) = Tax credit. (4) = Income recognition item. (5) = Deduction item. (6) = Directly or indirectly AGI based. 613

6 NATIONAL TAX JOURNAL VOL. XLIX NO. 4 thousand ($40,000 for single (married) taxpayers and complete elimination when AGI levels reached $35,000 ($50,000)). The thresholds for married taxpayers were less than double those available to single taxpayers, resulting in the potential for MTPs. However, only 255 (368) of the nonitemizer (itemizer) MFJ returns, representing a total of 1.9 percent of the returns in the sample, fell within the (affected) $40,000 to $50,000 AGI class. 9 Moving Expenses. The moving expense deduction was available only to itemizer taxpayers for the 1989 tax year. 10 Like NCLs, this statutory limitation created a $3,000 ceiling, per return, for moving expenses. This limitation was effectively doubled for two single taxpayers, creating the potential for an MTP. Only 2.9 percent (477) of the MFJ itemizers claimed moving expense deductions. The IRS SOI does not provide data regarding those amounts in excess of the $3,000 ceiling, so it is not possible to double this per return ceiling for a related MTP calculation. Gain on Sale of Personal Residence. The exclusion of up to $125,000 of gain from the sale of a personal residence by taxpayer(s) aged 55 or over does not apply separately to each spouse. Taxpayers taking this once in a lifetime exclusion (and divorcing later) are not only prevented from retaking this exclusion, but (in the event of remarriage) their new spouses are as well. Unmarried taxpayers could constructively shield double this potential exclusion amount as each party separately takes this exclusion. The incidence of taxpayers (over 65) 11 with gains on sales of their personal residences was very small. The 1989 IRS SOI contained only 23 (43) nonitemizer (itemizer) taxpayers with mean gains of $15,944 ($15,054). 12 Credit for the Elderly. The 1989 credit for the elderly or disabled was 15 percent of $5,000 ($7,500), reduced by one-half of any AGI in excess of $7,500 ($10,000) for single (married) taxpayers. Again, the potential for an MTP exists. However, the incidence of taxpayers claiming the credit for the elderly was very small. The 1989 IRS SOI contained only 6 (59) nonitemizer (itemizer) taxpayers with mean tax credits of $217 ($230). Conclusions Feenberg and Rosen (1995) did not include the decomposition of MTP components suggested by McIntyre (1988) in his response to Rosen s (1987) earlier version of this study of marriage tax penalties. This paper provides this decomposition for five of these penalties and provides some discussion of the relative importance of five additional MTPs, which were more difficult to disaggregate. These measures were produced by holding all other variables (e.g., marginal tax rates) constant, to the extent practicable, and were conducted while retaining 94 percent of the data used by Feenberg and Rosen (1995). Furthermore, this study did not use the divorce-based methodology employed by Feenberg and Rosen (1995), or its predecessor study by Rosen (1987), which generated HH taxpayers, artificially increasing penalties to which single taxpayers are also subject and confounded penalties for marriage with those related to the presence/absence of 614

7 RECENT DEVELOPMENTS IN THE MARRIAGE TAX qualifying dependency exemptions. Any future studies of the MTP, which fail to decompose (McIntyre, 1988) and/or include the ever-increasing EITC impacts in their MTP calculations, will generate EITC-driven increases in MTPs and may provide a misleading foundation for subsequent extensions. Finally, Feenberg and Rosen (1995) have suggested that future research efforts might be directed toward married/cohabitating taxpayer decisions to accurately reveal their marital status for income tax purposes. This proposal (and other possible studies), targeting the varying taxpayer groups likely to benefit (or be penalized) by specific MTP components, is facilitated by understanding which AGI classes are affected most severely (and likely to generate, for example, the maximum amount of regression model-based systematic variance) by each MTP component. This paper provides the necessary guidance, in the form of an AGI-related MTP decomposition, toward such efforts. ENDNOTES The author would like to thank Professor Fritz Scheuren, Director of the Statistics of Income Division Individuals of the Internal Revenue Service (retired) for providing the 1989 sample data used for this study. Also, the insights provided by two anonymous reviewers are appreciated. 1 McIntyre took exception to Rosen s (1987) treatment of the EITC as a tax (versus a spending) policy measure and suggested that these measures be disaggregated. 2 Because the Rosen (1987) and Feenberg and Rosen (1995) studies did not disaggregate the EITC, a uniformly distributed, nonitemizer-based simulation was conducted to approximate the impact of the EITC. AGIs of $1,000, $2,000,..., $100,000 were allocated to single (0 percent, 10 percent,..., 100 percent) and HH (100 percent, 90 percent,..., 0 percent) taxpayers. The tax for each of these 11 allocation scenarios (and for each of the 100 AGI levels) was then compared to the tax resulting for married taxpayers, filing jointly. Repeating this procedure with and without the EITC resulted in an (unweighted) mean EITC impact of $ Alternatively, consider the fact that single taxpayers are also penalized by the preferential rate and base effects provided to HH taxpayers. Therefore, these preferential treatments granted to HH taxpayers might just as easily be referred to as a singles tax penalty, a position inconsistent with the underlying rationale for the calculation of the MTP. To the extent that additional HH taxpayers were generated, a divorce-based methodology overstates the MTP. This is clearly a function of the decision to have children and not a function of the decision to marry as their article would suggest. 4 This distinction is provided for in Rosen (1987) and Feenberg and Rosen (1995). The results of Table 2 in both studies (generally) suggests narrower ranges of MTPs/MTBs for childless taxpayers. 5 Of the 48.1 million 1989 MFJ returns filed, 0.4 million (0.8 percent) and 2.5 million (5.2 percent) were represented by AGI classes below $1 and in excess of $100,000, respectively (SOI, 1992). 6 It is intuitively unappealing to suggest that the economies of scale, as they relate both directly and indirectly to tax deductible living expenses, would remain constant for these couples after divorcing. The relevance of this point can be associated with the ability to pay concept of taxation and is also ignored by proponents of marriage neutrality (McCaffrey, 1993). 7 The IRS SOI is a stratified random sample. Failure to adjust raw (unweighted) observations for the sample weights provided by the IRS can result in significant distortions. 8 It is interesting to note an inconsistency, in that this MTP does not extend to IRC Section 1244 stock treatment, which effectively limits annual losses to $50 thousand per person. 9 If one were willing to assume that primary and secondary taxpayer IRA contributions would remain stable under a policy which effectively doubled the thresholds for this affected group, (unweighted) mean MTPs for the affected nonitemizer (itemizer) taxpayers would approximate $595 ($497) dollars. 10 For post-1993 tax years, the deduction for moving expenses was restored to its pre-1987 status as an above the line deduction. 11 The IRS SOI did not provide information with regard to once-in-a-lifetime elections or taxpayer s age. The additional over 65 personal exemption was used as a proxy, admittedly understating the significance of this potential MTP. 12 I will note that the distributions of these gains were remarkably similar for itemizer and nonitemizer MFJ returns. Skewness and kurtosis were (1.492) and (4.315), respectively, for nonitemizer (itemizer) taxpayers. 615

8 NATIONAL TAX JOURNAL VOL. XLIX NO. 4 REFERENCES Brozovsky, John and Anthony J. Cataldo. A Historical Analysis of the Marriage Tax Penalty. Accounting Historians Journal 21 No. 1 (June, 1994): Cataldo, Anthony J. The Earned Income Credit: Historical Predecessors and Contemporary Evolution. Accounting Historians Journal 22 No. 1 (June, 1995a): Cataldo, Anthony J. The EITC: Investigation of a Wealth-Based Phase-Out. Tax Notes 67 No. 4 (April 24, 1995b): 565. Feenberg, Daniel R., and Harvey S. Rosen. Recent Developments in the Marriage Tax. National Tax Journal 48 No. 1 (March, 1995): McCaffrey, Edward J. Taxation and the Family: A Fresh Look at Behavioral Gender Biases in the Code. UCLA Law Review 40 No. 1 (April, 1993): McIntyre, Michael J. Rosen s Marriage Tax Computations: What Do They Mean? National Tax Journal 41 No. 2 (June, 1988): Rosen, Harvey S. The Marriage Tax Is Down but Not Out. National Tax Journal 40 No. 4 (December, 1987): U.S. Department of the Treasury. Internal Revenue Service. Statistics of Income Division. Individual Income Tax Returns Publication 1304 (Rev. 9-92). Washington, D.C.: GPO, APPENDIX SUMMARY OF THE ALGORITHMS USED TO CALCULATE THE MTPS PRESENTED IN TABLES 1 AND 2 Rate: If TI $30,950, then MTP = $0 If TI > $30,950 and $37,100, then MTP = [ TI $30,950 ] [ 28% 15% ] If TI > $37,100 and $74,850, then MTP = $ If TI > $74,850 and $89,800, then MTP = $ [[ TI $74,850 ] [ 33% 28% ]] If TI > $89,800, then MTP = $ 1,547 Base: SS: NCL: If nonitemizer, then [[ SGLSTD 2 ] MFJSTD ] MTX = MTP If itemizer, then [[ SGLSTD 2 ] ITEM ] MTX = MTP [[ SGLSS 2 ] MFJSS ] MTX = MTP If NCL $ 3,000, then MTP = $ 3,000 MTX If NCL < $ 3,000, then MTP = NCL MTX AMT: If AMT $ 4,200, then MTP = AMT If AMT > $ 4,200 and $23,100, then MTP = $4,200 If AMT > $23,100 and $42,787.50, then MTP = $ 4,200 + [[ AMT $23,100 ] 25% ] If AMT > $42, and $65,100, then MTP = $8, If AMT > $65,100 and $97,650, then MTP = $ 8, [[ AMT = $65,100 ] 25% ] If AMT > $97,650, then MTP = $0 SGLSTD = The standard deduction available to nonitemizer, SGL taxpayers for the (1989) tax year ($3,100). MFJSTD = The standard deduction available to nonitemizer, MFJ taxpayers for the (1989) tax year ($5,200). MTX = The MTX for the MFJ taxpayers prior to any adjustment (i.e., held constant). TI = Taxable income. SGLSS = The SS exclusion for SGL taxpayers. MFJSS = The SS exclusion for MFJ taxpayers. 616