Printer Friendly

Integrating the tax burdens of the federal income and payroll taxes on labor income.

"The [Social Security] tax is the neglected stepchild of tax policy analysis." (1)


When politicians, or even the average guy on the street, debate whether the allocation of the federal tax burden is "fair" and "efficient," they typically do so by talking only about the federal income tax burden. For example, columnist Jerry Heaster, correctly reporting that the top one percent of income earners in 1999 paid about thirty-six percent of "all federal personal income taxes," writes, "How much tax do high income Americans need to pay before the class warriors will be satisfied?" (2)

But the fact of the matter is that about as much federal revenue is collected from the payroll taxes (Social Security and Medicare) as from the personal income tax, (3) and the payroll taxes are extremely regressive, significantly diluting the percentage of the total federal tax burden paid by the wealthy. For example, in a year 2000 study, the staff of the Joint Committee on Taxation estimated that, under the law at that time, those earning more than $200,000 would pay 42.7% of the total federal personal income tax collected, while the top 1% would pay 33.6%. (4) If, however, all federal taxes are measured, including the regressive Social Security and Medicare taxes, the shares plummet to 27.5% for those earning more than $200,000 and 18.6% for the top 1%. (5) These reductions will become even more dramatic after the full phase-in of the 2001 tax cuts for the super wealthy. (6)

While the tax debate centers on the federal income tax burden, provocative empirical studies published by economists Andrew Mitrusi and James Poterba in 2000 show that nearly two-thirds of American households now pay more in federal payroll taxes than income taxes, (7) chiefly because of significant increases in the payroll tax burden over the last twenty years. This results in much higher effective tax rates, as well as a higher total tax burden, for the lower and middle classes than is probably widely appreciated. Further, it results in a higher portion of the federal tax burden being borne by labor income (as opposed to capital income, which is concentrated in the wealthier households) than is probably widely appreciated. In other words, there is a huge elephant in the room that few politicians talk about and even fewer citizens recognize.

This article explores the historical reasons why we have three separate taxes on labor income at the federal level--the income tax, the Social Security tax, and the Medicare tax--and considers how and why the government might integrate the tax burdens on labor income in some way today. (8) For ease of discussion, this article will refer to the "double tax" on labor income, which 'consists of the income tax on labor income and the two payroll taxes on labor income lumped together. (9)

The case for integration rests more on equitable grounds than on grounds of economic efficiency. In one sense, the efficiency argument is a function of the extent of the disincentive effect that a double tax on labor income has on work effort. Because of the $84,900 wage ceiling that caps the Social Security tax base, (10) and the low rate for the Medicare tax that continues to apply above that ceiling, there is likely no disincentive effect at the margin for wage earners or self-employed individuals earning more than $84,900. Indeed, the income effect (11) might predominate for workers earning slightly below the wage ceiling, who might work longer and harder to exceed the margin and begin taking home more after-tax income for dollars earned above the ceiling.

The case is certainly more complicated for those earning significantly less than the wage ceiling, but there might not be as large a disincentive effect as might be imagined for several reasons. The obvious reason is that, to the extent the worker has control over hours worked, we do not know whether the income effect predominates over the substitution effect or vice-versa. Moreover, lower- and middle-class taxpayers might have far less control than the upper-class taxpayers over the extent of work effort. For example, the worker might not have the freedom to choose to work only thirty-five hours per week instead of forty hours if the employer rigidly defines the worker's job to require forty hours. In short, it might not be possible to disentangle the myriad incentives that affect work effort and isolate the incentive effects of the payroll taxes in particular.

Another factor might be the worker's perception, whether accurate or not, regarding the "rate of return" that she receives in retirement on her "insurance contribution." The disincentive effect on work effort should be weakened if the worker believes that the return is reasonable. While low-wage workers have historically obtained a high "return" on their "insurance contributions," (12) Kotlikoff and Sachs argue that low-wage workers tend not to appreciate this, because some low-wage workers, such as domestic workers, have striven to avoid imposition of the tax (and the subsequent enjoyment of the return) through a failure to report self-employment income. (13) This tendency, however, might be due simply to a pattern of high "discount rates" among low-wage workers. That is to say, low-wage workers might require an unreasonably high rate of return before they are willing to defer consumption to the future. (14)

In any event, to the extent that payroll taxes do discourage work at the margins for low- and middle-class taxpayers and thus raise efficiency concerns, or encourage the underreporting of wages, integrating the payroll and income tax burdens could help to ameliorate these effects.

The more persuasive argument for integrating the payroll tax burden on labor income with the income tax burden on labor income is simple equity. The combined tax burden on labor income from the payroll and income taxes on the poor and middle classes has increased dramatically in the last two decades, and the increase in tax burden is clearly due to substantial increases in the Social Security and Medicare tax rates as well as the Social Security wage base during that time. (15) In other words, there has been a dramatic increase at the federal level on the taxation of labor income, particularly for the middle class. Moreover, this period also witnessed a decreasing reliance on the corporate tax, one of the biggest sources of capital taxation. (16) Michael Graetz argued:

If the basic principle of tax justice requiring taxes to be distributed in accordance with ability to pay is to be satisfied, a significant portion of the tax burden must be borne either by wealth (or wealth transfers) or income from capital. Taxes on labor income alone--even if, unlike [Social Security] taxes, they were progressively structured--do not produce taxation based upon ability to pay, for those with the greatest ability to pay often have channeled their monies into capital. (17)

One can defensibly argue that this state of affairs is simply unjust, particularly at a time when the tax burdens of the super wealthy are being significantly reduced through repeal of the estate tax (18) and reduction in the top income tax (19)--all in a period of unprecedented wealth concentration in this country. In considering whether the distribution of the tax burden is "fair" by examining the extent to which it is "progressive," the right inquiry might not be whether the percentage of tax collected across income strata is progressive but rather how taxes affect the distribution of after-tax discretionary income and wealth. (20) If this is right, information on income and wealth concentration is crucial to the inquiry, and "wealth in America is more highly concentrated today than at any time since 1929...." (21) The wealthiest 1% of households has more than doubled its ownership of private assets in this country since 1976. Whereas the top 1% owned about 19% of our country's assets in 1976, it now owns nearly 40%. (22)

Examining Federal Reserve data, Edward N. Wolff found that between 1983 and 1998--a period that coincides with substantial increases in the payroll tax burdens on the labor income of the poor and middle classes--the trend in wealth concentration accelerated. As reported by Tom Redburn:

The number of households with a net worth of more than $1 million nearly doubled, to 4.78 million from 2.41 million. At the same time, the ranks of the truly rich exploded even more, with those worth more than $10 million nearly quadrupling, to 239,000 from about 67,000.

That's great. A rich society like the United States should have plenty of rich people--and the more the merrier. But what is disturbing is what happened to just about everybody else.

"The richest 1 percent accumulated 53 percent of the total gain in marketable wealth over the 1983-1998 period," Mr. Wolff wrote. "The next 19 percent received another 39 percent, so that the top quintile accounted for 91 percent of the total growth in wealth, while the bottom 80 percent accounted for a mere 9 percent."

"The results indicate rather dramatically," he added, "that the fruits from economic growth in the last few decades were enjoyed by a surprisingly small part of the population--the top 20 percent, and particularly the richest 1 percent." (23)

Though most agree that extreme wealth concentration can lead to plutocracy and can damage democratic values, there can never be universal agreement on the proper criterion for arriving at what constitutes the "most" equitable distribution of the tax burden across the members of the population. Nor will there ever be universal agreement regarding the proper balance between equity, however defined, and economic efficiency and growth, or even the "best" tax policy approach to support growth.

For example, a tax system premised on the classic utilitarian theory of minimal aggregate sacrifice would result in taxing only Bill Gates until his income was brought down to the next highest income earner, taxing those two until their incomes were brought level with the third, etc., until the requisite amount of revenue was raised. Yet, such an approach to "equity" concerns--involving exclusive taxation of the rich at confiscatory rates--completely ignores other criteria of fairness, such as just dessert, and also completely ignores the effects of such a tax system on economic efficiency and investment. (24) At the other extreme, an equal sacrifice or head tax might free up huge amounts of capital for investment by slashing taxes on the wealthy but completely ignores the intuitive notion of the decreasing marginal utility of money (sticking with classic utilitarian theory for the sake of comparison).

There is also no simple agreement within the realm of economics itself regarding how best to promote growth, as exemplified by the disagreement between supply-side and demand-side economists regarding the role of savings in economic growth. As summarized by John F. Witte:

According to Keyne's analysis of the depression, excess savings do not necessarily lead to growth; rather, they stifle consumption, which is the engine that stimulates economic activity. The policy argument that follows from this analysis is that one should reduce taxes, but reduce them for the poor and middle class, who consume the most. Supply-side theorists stress the need for investment and an increase in savings to fuel corporate expansion. The corresponding policy recommendation is again to cut taxes, but in this case the taxes of upper-income groups that have a higher propensity to save and greater capital to invest. (25)

In short, those who believe that the tax burden ought to favor capital substantially may not see any need to explore how the income tax and payroll tax burden on labor income might be integrated. But for those troubled by the increasing percentage of the tax burden borne by labor income (as opposed to capital income) and by the increasing magnitude of wealth and income concentration, this article shall press on with its exploration.

The most straightforward way to integrate would be to repeal the payroll taxes (the Social Security tax and Medicare tax) and raise the revenue formerly raised via those taxes through the income tax. After all, "Social Security is just a part of a larger tax-transfer system that matters solely insofar as it affects people." (26) A combined system could make these allocative and distributional choices much more transparent, even if the decisions are to keep the relative allocations and distributions the same. Having several, separate taxes at the federal level on labor income masks the combined effective tax rate and tax burden imposed on labor income.

Nevertheless, there are sound reasons for maintaining the payroll taxes as separate taxes. Just as when they were adopted, the separate taxes might be perceived by the average taxpayer as the equivalent of "insurance premiums," with the benefits received in retirement as merely the receipt of insurance benefits previously purchased. While, as later described, this is an absolutely inaccurate description of the Social Security and Medicare systems, the common perception can lead to political support for the old-age and survivor's benefit and Medicare benefits, which are generally redistributive in a way that many consider good as a matter of public policy. (27) In other words, some might argue that there are sound public policy reasons to nurture what amounts to an inaccurate view of the Social Security system in order to protect it--a "benign propaganda" of sorts. The insurance analogy that leads to widespread political support could be lost if the separate payroll taxes were repealed and the programs funded with general revenue from the individual and corporate income taxes. (28)

Short of repealing the payroll taxes entirely, full or partial integration might consist of an income tax deduction for the employee portion of the Social Security and Medicare taxes. Alternatively, payroll taxes could be made creditable against income tax liability, with a gross-up in the compensation inclusion if the employer-paid portion were also made creditable by the employee. Since I assume, however, that revenue constraints would not likely allow crediting even the entire employee portion of such taxes, I also assume that any plausible proposal would be limited to either deducting or crediting a portion of the employee portion. That issue is what this article addresses.

It makes little sense to allow a deduction (or a credit) for a tax within the same tax. That is to say, it makes little sense for the federal income tax to be deducted for purposes of computing how much federal income tax is owed. As Senator Henry F. Hollis said when a deduction for federal taxes was repealed in 1917, "It is a pure matter of expediency. If you so arrange the income tax this year that you allow those who pay it to take back a third of it next year, you have simply got to put on a bigger tax ...." (29) It would make the tax system more complicated with no fundamental shift in the tax burden.

But the matter is not so clear-cut when considering whether a flat wage tax of approximately fifteen percent on wages up to, but not exceeding, approximately $85,000 with no deductions or credits (30) ought to be deductible or creditable under an income tax on both wages and capital with no ceiling and progressive rates up to approximately thirty-five percent. Absent massive changes in the structure of the income tax itself, either a deduction or a refundable credit for a portion of the payroll tax should make the tax burden more progressive and less oppressive on the labor income of lower- and middle-income taxpayers. If the argument for integration of the payroll and income tax burdens is to rest on equity grounds, however, the spending side must also be considered. That is to say, welfare might not actually be increased for low- and middle-wage workers if they suffer disproportionately any spending cuts that may accompany the integration.

For this reason, this article might be more than a year too late. Substantial budget surpluses were forecasted prior to the enactment of the Economic Growth and Tax Relief Reconciliation Act of 2001 (2001 Act). (31) Once our elected representatives decided to decrease tax collections significantly to decrease this surplus, they had to decide which form the tax reductions should take. Integrating the payroll and income taxes in a manner considered in this article would have resulted in the allocation of much more of the tax cut to the lower and middle classes than occurred with the decision to repeal the estate tax and decrease the marginal income tax rates on the highest income earners without a concomitant reduction in social spending. Absent a rollback of those tax decreases, which seems unlikely, the revenue lost with any integration proposal would have to be matched with spending cuts today, since no budget surpluses are foreseeable in the near future. (32) If the spending cuts of $X that were needed to p ay for the lost revenue of $X that integration would produce were aimed disproportionately at the poor and middle classes, then no net change in utility for those classes would result. Indeed, they might be worse off if they were deprived of necessary services that they could not obtain in the marketplace for a price equal to (or less than) their tax savings.

For that reason, this article explores one other possibility that does not involve explicitly integrating the tax burdens of the income and payroll taxes on labor income but might nevertheless address the equity concerns inherent in the high combined tax rates on such income for the lower and middle classes. That remedy is to repeal the wage ceiling on the imposition of the Social Security tax, which would allow drastically slashing the relatively high marginal rates that we now have under that tax, while maintaining such a ceiling in the payment formula (33) in order to protect the redistributive function. The rate reduction could be calibrated to be revenue neutral and thus would not require spending cuts. Eliminating the wage ceiling on the Social Security tax, while maintaining a ceiling under the payment formula, might be perceived as unfair if the tax is thought to be the equivalent of a private pension contribution. The arguments developed in this article, however, which build on seminal work done by P atricia Dilley, show why such a perception of the Social Security system is fundamentally misguided. Viewed as a general tax that supports the infrastructure of a capitalist economy by maintaining consumption spending by the retired, etc., there is no reason why the tax should not be structured in a progressive fashion, as is the income tax itself. With a flat-rate tax, the only way to make it progressive (and to lower the single, flat marginal rate) is to expand the wage base upward.

Part II briefly sets out the double tax on labor income that arises under the income, Social Security, and Medicare taxes and even more briefly describes the current income tax treatments of the latter two taxes as well as the benefits received under the Social Security and Medicare programs. Part III provides information on the history of relative income tax and payroll tax burdens in order to place the issue in historical perspective. Part IV explores the history of the separate payroll tax itself. Part V confronts arguments that the combined Social Security and Medicare payroll tax is not a double tax (in addition to the income tax) at all because it purchases individual future benefits that, on average, exceed the tax payments made. The issue of whether we ought to link the payroll tax paid to future benefits received on an individual basis is a critical aspect of analyzing how they ought to be treated under an income tax. I reject that linkage. Even though I reject that linkage, Part VI nevertheless explores how payroll taxes as well as cash benefits and medical care received under Social Security and Medicare might be treated under the income tax if we accept that linkage as a legitimate one. Part VII discusses these issues from the more persuasive view that severs the tax from possible futu re benefits. At bottom, it argues that a portion of the employee payroll taxes, both Social Security and Medicare, ought to be creditable, dollar for dollar, against any income tax liability, with a refundability feature capped by reference to a reasonable "personal exemption" amount. Cash Social Security payments received, if any, ought then to be fully includable in gross income under the income tax when received, though the value of medical care received under the Medicare program ought to remain excludable. (34)


American taxpayers are generally subject to three federal taxes imposed on wages and self-employment income, as opposed to capital income. The earliest of these taxes was the so-called income tax, which is today actually a hybrid income/consumption tax, first enacted in 1913 after adoption of the Sixteenth Amendment to the Constitution. (35) Though commentators bemoan the difficulties of fully taxing income from capital under our current income tax, (36) the tax is (at least nominally) imposed on income both from labor and from capital. (37) Indeed, one of the fundamental impetuses underlying the move toward income taxation was a desire to tax income from the capital of the wealthy. This income essentially escaped taxation under the various consumption taxes, such as imposts and tariffs, which funded the operations of the federal government prior to 1913. (38) The debates surrounding the concept of income taxation clearly evidenced the policymakers' intention to achieve greater fairness in the allocation of t he tax burden. They show a common understanding that prior modes of taxation were consumption taxes that were regressive in nature and that an income tax would more fairly apportion the tax burden to those with greater wealth. (39)

In view of this early desire to use income taxation to effectuate a more progressive allocation of the tax burden, the income tax, though of earliest application, did not apply to a great many taxpayers (40) until the advent of World War II, when it was forever "transformed from class tax to mass tax." (41) The early desire to use the tax to reach chiefly the capital income of the wealthy was attenuated by the need for larger revenues to fund the war effort and post-war government policies, though the rate structure of the income tax remained highly progressive in this era. While middle-class taxpayers were subject to the tax on a massive scale for the first time, the marginal rates applicable to these taxpayers were significantly lower than the ninety-percent plus rates applicable to the highest income earners.

The two so-called payroll taxes discussed here are imposed on labor income only, and they were introduced at different points in time. The first was the Social Security tax, enacted in 1935 (42) and substantially broadened in 1939 (43) (and many times thereafter) as part of President Roosevelt's New Deal. For employees, the tax base for this tax is "wages," and an equivalent self-employment tax is imposed on self-employed individuals. For 2002, the Social Security tax is imposed on the first $84,900 of wages or self-employment income, (44) with no personal exemptions, deductions, or zero-bracket amount of any sort, unlike in the income tax. The ceiling increases each year according to the increase in average wages in the U.S. economy. Wages or self-employment income earned above the ceiling amount are free from these taxes. Therefore, Bill Gates pays the same in Social Security tax as a college professor earning $84,900 per year in wages. The employee must pay tax at a rate of 6.2% on this tax base, (45) while the employer pays an additional 6.2% (46) (summing up to 12.4%), though economists generally agree that the incidence of the employer portion of the tax falls mostly on the employee in the long run through wages that are depressed by an equivalent amount. (47) The self-employed pay the entire 12.4% directly. (48)

The Medicare tax, also imposed on a tax base of wages (or self-employment income), was enacted in 1966. (49) The Medicare tax, unlike the Social Security tax, is now imposed on all wages and self-employment income, without limit, and is imposed at a combined employer and employee rate of 2.9% (1.45% on each). (50) Once again, the self-employed pay the entire 2.9% directly. (51) Thus, wages and self-employment income up to $84,900 (as of 2002) are taxed at a total combined flat rate of 15.3%, with no exemptions or deductions, while wages and self-employment income above that ceiling continue to be taxed at a flat rate of 2.9%.

None of the payroll and self-employment taxes are creditable against the tax due under the income tax, and none of them are formally deductible by employees under the income tax. Because the incidence of the employer portion of these payroll taxes likely falls on the wage-earner, however, the employee gets an effective deduction for these taxes, as though payment of the taxes were included in the employee's gross income as additional wages, and then the employer portion of the tax were deducted by the employee. Thus, it is more accurate to say that, in substance, the employee is effectively allowed to deduct one-half of the 15.3% aggregate payroll tax economically borne by him but not the other half. To achieve parity, the self-employed are expressly allowed to deduct one-half of their self-employment taxes under section 164 of the Internal Revenue Code (Code). (52)

Although money is fungible, and thus tracing government outlays from any particular tax stream is conceptually problematic, the Social Security and Medicare taxes are commonly perceived to be the source of funds for several kinds of government benefits. (53) The biggest and most well-known is the old-age and survivors benefit, payable to those in retirement who have worked at least forty quarters (three-month periods) in the paid labor force (unless exempted from the system (54)), as well as their surviving spouses. Joint federal-state programs dealing with disability payments are also funded by these payroll taxes.

Under the complex rules in section 86, a certain portion of old-age and survivors benefits is taxed under the income tax when received. Daniel Shaviro explains the rules as follows:

Under a maddeningly complex set of rules, Social Security benefits are excluded for income tax purposes by individuals with modified adjusted gross incomes below $25,000 ($32,000 for married couples). Above that point, the benefits are 50 percent includable, and the resulting income tax revenues officially attributed to the Social Security Trust Fund. The inclusion percentage rises to 85 percent for individuals with modified adjusted gross incomes above $34,000 ($44,000 for married couples), with the extra revenues from the increase in inclusion percentage being officially attributed to Medicare. (55)

In 1999, about 45.7 million people received old-age and survivors benefits, while 9.6 million individual income tax returns reported taxable benefits, a record high. (56) Disability payments are treated the same as old-age and survivors benefits. (57) The value of free medical care received in kind pursuant to the Medicare program is not taxed under the income tax.


As of 1999, nearly two-thirds of families pay more in payroll taxes than they do in income taxes, while fewer than one-quarter of families pay more in income taxes than they do in payroll taxes. (58) By comparison, in 1979, only forty-four percent of families paid more in payroll taxes than income taxes. (59) Moreover, the families that pay more in payroll taxes than in income taxes are overwhelmingly low- and middle-income families. Mitrusi and Poterba showed that an overwhelming majority of families with adjusted gross incomes of $100,000 or less in 1999 paid more in payroll taxes than in income taxes; however, very few families with adjusted gross incomes of more than $100,000 paid more in payroll taxes than income taxes. Mean income taxes do not reach approximate parity with mean payroll taxes until one reaches $75,000 to $100,000 of adjusted gross income (AGI), increased by certain items of untaxed income (adjusted AGI). (60) "At income levels below $50,000, more than three-quarters of families have payr oll tax bills that exceed their income taxes. At adjusted AGI levels above $200,000, income taxes exceed payroll taxes for virtually all families." (61)

These conclusions are only marginally less dramatic if solely the employee share of payroll taxes is taken into account. As noted earlier, most economists assume that the incidence of the employer share of Social Security taxes falls on the employee. Nevertheless, the last column in the table above breaks out the percentage of families for whom the employee share of payroll taxes exceeds the personal income tax.

Not surprisingly, excluding half of each family's payroll tax burden substantially reduces the likelihood that payroll taxes exceed personal income taxes. However, employee payroll taxes still exceed personal income taxes for 35 percent of all families, and for more than 40 percent of all families that pay taxes. Families in the lowest income ranges, with adjusted annual AGI of less than $20,000, have a greater than 50-percent chance of paying more in employee payroll taxes than in personal income taxes. Conditional on paying any taxes at all, these families' chances of paying more in employee payroll taxes than in personal income taxes are greater than four in five. (63)

Mitrusi and Poterba also illustrated the increasing burden of payroll taxes relative to income taxes between 1979 and 1999, and this increase in burden was most dramatic for middle-income taxpayers with adjusted AGIs of between $30,000 and $100,000. For example, the table below shows that, in 1979, 45.8% of families with adjusted AGIs of between $30,000 and $40,000 pald more in payroll taxes than in income taxes, but that percentage nearly doubled to 85.7% by 1999. Even more dramatic was the increase from 4.6% to 55% for families with adjusted AGIs of between $75,000 and $100,000. Moreover, the increases are even more pronounced if the sample is limited to only families that actually paid taxes. The table shows that for families with 1999 adjusted AGIs between $10,000 and $20,000, for example, 71.2% of all families but 87.5% percent of families with positive tax liabilities had 1999 payroll taxes in excess of personal income taxes. Mitrusi and Poterba also showed that these effects were particularly dramatic for two-earner married couples, where the cap on taxable earnings can result in a greater aggregate payroll tax liability for a given level of family earnings. (64)

What to make of these observations? If the allocation of total revenue receipts between income and payroll taxes had remained relatively constant in this period, these numbers actually might be quite positive. Under such an assumption, they could possible demonstrate that the income tax burdens of low- and middle-income families have been substantially shifted to the upper classes during this period, which would mean that relatively constant payroll tax payments could then exceed income tax payments for more members of this class. It is true that the income tax reaches fewer low-income families today than in the past, due mainly to the enactment and expansion of the earned income tax credit. (66) When the relative contributions of payroll taxes and income taxes to total national revenue are considered, however, it becomes apparent that these numbers are due in large part to substantial increases in the payroll tax rates as well as substantial increases in the Social Security wage base in the last two decades. Payroll taxes have funded an increasingly larger share of total federal tax revenue since the early 1980s. Even though the taxes are imposed at a flat rate, the wage ceiling feature of the tax base means that it is both regressive (with the average tax rate falling as wages increase) and concentrated in the poor and middle classes. The significant increases in the payroll tax rates and the wage base since the early 1980s explain why these taxes have become an increasingly larger share of the total tax burden for those families with less than $100,000 of income. "Until 1963, federal receipts from the individual income tax were more than twice as great as federal payroll tax revenues. In 1995, payroll tax receipts were approximately equal to income tax receipts." (67) This observation is confirmed by data showing that, when income and payroll taxes are combined, most upper-income taxpayers realized an aggregate tax decrease between 1979 and 1999, while low-income taxpayers generally experienced a tax increase.

If we restrict the analysis to families that paid either payroll or income taxes..., the data show that 70.2 percent of families faced lower income taxes in 1999, but that only 37.5 percent faced lower income and payroll taxes. At low family income levels, most families experienced a tax increase between 1979 and 1999. We need to focus on families in income strata above $50,000 before the probability of paying lower combined taxes in 1999 than in 1979 rises above 50 percent. [These results] provide a stark demonstration of the importance of combining income and payroll taxes when considering recent changes in tax burdens. (68)

These numbers will become even more pungent when phase-in of the 2001 Act is complete. (69)

I have been unable to find data comparable to that of Mitrusi and Poterba for the early years of the payroll taxes. Nevertheless, it is fair to say that the increasing tax burden on labor income of the lower and middle classes (from the combination of payroll and income taxes on labor income) evident from 1979 to 1999 is a relatively new phenomenon. Consider two other emblematic tax years: 1939 and 1960. (70) In 1939, the payroll tax was established but the income tax had not yet become a mass tax. In 1960, the income tax had itself become a mass tax, imposed on the lower classes in addition to payroll taxes. Moreover, that year is only seventeen years before the beginning of the twenty-year period examined by Mitrusi and Poterba.

In 1939 (and, indeed, through 1949), the combined employer and employee Social Security tax was two percent (one percent each), which was imposed on wages up to $3,000. (71) Because of the generous personal exemptions under the income tax--$2,500 for married couples and $1,000 for single taxpayers, along with $400 for each dependent, (72) at a time when few households earned as much each year--the two-percent payroll tax was the only tax paid by the vast majority of lower- and middle-class workers. Even during the "high tax" years of World War I, "at most 13 percent of the labor force [paid] income taxes." (73) When tax collections decreased in the 1920s after the end of the war, Yale economist T.S. Adams wrote that the income tax "touches directly perhaps only 5 or 6 percent of the population." (74) By 1939, only about five percent of the population paid income taxes. (75)

By 1960, the income tax was entrenched as a mass tax, with approximately seventy-three percent of the population covered by taxable returns. (76) The personal exemptions had been reduced to $600 for a single individual and $1,200 for a married couple in 1948, and they remained unchanged by 1960 in spite of the reduced economic value of those amounts with inflation. But the income tax rates applicable to the lower classes were still relatively low, so that those with adjusted gross incomes of less than $5,000--when the average annual salary in the early 1960s was $4,700 (77)--contributed just 8.8% of aggregate individual income tax revenue in 1965. (78) Moreover, while the payroll tax wage base increased to $4,800--almost precisely the average annual salary--and the combined Social Security tax rate increased to six percent (three percent each for the employer and employee), (79) the payroll tax, which did not yet include the Medicare tax, was still low compared to modem standards. Therefore, it is fair to say that the combined tax rate on the labor income of the poor and middle classes was still low by comparison to today. The bulk of federal revenue was collected from the upper classes, and the balance between taxation of labor income and taxation of capital income was not nearly as skewed toward labor income as today.

This brief review suggests that one means by which we could decrease the tax burden on the labor income of the lower and middle classes to restore a balance closer to historical standards (without explicit integration of the two taxes) would be to repeal the wage ceiling under the Social Security tax. The Social Security tax would then apply to an unlimited amount of wages, like the Medicare tax. Elimination of the wage ceiling would allow a slashing of the Social Security tax marginal rate. In other words, eliminating the wage ceiling could pay for substantial rate cuts (perhaps closer to the three-percent rates of 1960) in a revenue-neutral fashion. Such lower marginal rates should also reduce any behavioral distortions, and thus sacrifices in economic efficiency, that accompany the relatively high Social Security tax rates of today. (80)


It was not inevitable that Social Security benefits would be funded from a separately collected and ear-marked payroll tax on wages. The Social Security program created in 1935 could have been funded with general tax revenues collected from the individual and corporate income taxes (as well as tariffs and other federal excise taxes). Indeed, some influential members of President Roosevelt's Commission on Economic Security, which drafted the Social Security proposal, advocated such an approach. Wilbur J. Cohen accompanied University of Wisconsin Professor Edwin E. Witte to Washington when Professor Witte, who "virtually single-handedly wrote the entire wide-ranging committee report," (81) was appointed to head the effort. Cohen later recalled that "[d]uring the formation of the committee's proposal, Emergency Relief Administrator [Harry] Hopkins favored some kind of broad, comprehensive, unified program that would involve, exclusively or substantially, federal general revenues. Other proposals along this line were pending in Congress and were supported by some social workers and members of Congress." (82)

Unless the Congress expanded the income tax to reach the middle and lower classes, however, using general income tax revenues to fund Social Security benefits would have meant that the wealthy would essentially pay for the program. In 1935, the wages of most workers were not taxed under the income tax because of the high personal exemptions. Conservatives, who "charged that the social security conception violated the traditional American assumptions of self-help, self-denial, and individual responsibility," (83) would not likely have supported such a blatantly redistributive program. Indeed, such a system could even aggravate racial animosities. "'The average Mississippian,' wrote the Jackson Daily News, 'cant imagine himself chipping in to pay pensions for able-bodied Negroes to sit around in idleness on front galleries, supporting all their kinfolks on pensions, while cotton and corn crops are crying for workers to get them out of the grass.'" (84)

To avoid this, one option could have been to extend the income tax from the wealthy to the lower classes, with the result that their wages would then come within the income tax system. If that had been done, their wage income would have been taxed only once at the federal level--even as the income tax itself expanded to raise more revenue with World War II--and the increased revenues obtained under the general income tax from this expansion could have funded the payments made under the program. In other words, Social Security spending would have been simply one more government program supported by general tax revenues. The government, however, would collect those general tax revenues from the lower and middle classes (which would benefit from the new program) as well as the wealthy.

But there were important political reasons for the separate tax. If the "tax" could be sold as an "insurance contribution" rather than a "tax," and Social Security benefits perceived as simply the return for which prior premiums were paid, President Roosevelt believed that "by virtue of a statutory 'compact' between the contributors and Congress,... a future President and Congress could not, morally or politically, repeal or mutilate the 'entitlement' character of the program." (85) His observation has held true. (86)

Moreover, there was simply no need at the time to expand the income tax downward, as opposed to enacting a separate tax on wages, in order to avoid a future "double tax" problem, since the income tax was itself thought at the time to be a tax that would never reach the lower and middle classes. It was, therefore, not likely foreseen that the labor income of these lower- and middle-class workers would soon be taxed twice--once under the new payroll tax and once under the income tax. As Professor Carolyn Jones related:

In testimony before the Senate Finance Committee in 1932, Herbert Hoover's Treasury Secretary, Ogden Mills, aptly described the very limited scope of the individual income tax up to that time. "We have become accustomed," he said, "to high exemptions and very low rates on the smaller taxable incomes. That is our fixed conception of an income tax and it is very difficult as a practical matter to change fixed conceptions of this character." (87)

She went on to note that this "fixed conception" remained prominent through the middle and late 1930s in Roosevelt's New Deal, when the payroll tax was enacted. "In public pronouncements, Roosevelt and prominent Congressmen linked income taxpaying to plutocracy and rejected imposition of income taxes upon 'average' citizens. (88) Moreover, Congress agreed.

Congress was... quite clear as to who should not be paying income taxes. When Senator LaFollette proposed reducing exemptions to $2,000 for couples and $800 for singles [in 1935], he was soundly defeated. At a time when three-fourths of American families were at or below the $2,000 level at which they could live decently, Sen. Alben Barkley argued that LaFollette's measure would hurt the "average citizens" and "average families" "whether we consider the average man as one who receives less than $5,000 a year or the one who receives less than $10,000 a year--we can make up our own average to suit our own view of what an average ought to be." (89)

Instead, Congress "increased surtaxes on those with incomes over $50,000, making a top bracket of 79 percent for income over $5 million. For three years thereafter only John D. Rockefeller qualified for this most stratospheric of tax brackets." (90) This attitude toward income taxation continued through the 1930s.

This perception of the income tax as a weapon to be used only against the wealthy persisted in 1936. In his acceptance speech at the Democratic convention in Philadelphia, Roosevelt claimed to have enlisted in a war against the "economic royalists" who had crowded out "many thousands of small business men and merchants who sought to make a worthy use of the American system of initiative and profit." In an October 1936 campaign address at Worcester, Massachusetts, FDR spoke of the "struggle to maintain democracy in America"--a struggle between "the vast majority of our citizens" and "a small, but powerful group which has fought the extension of [the] benefits of democracy, because it did not want to pay a fair share of their cost." The politics of class division were clearly at work. Tax relief should be given to the average working American, while the wealthy should be more heavily taxed.

The Revenue Act of 1937 again played upon the. theme of unfairly low level of taxes paid by the wealthy. The Administration recommended loophole closing, and Congress responded with hearings detailing tax avoidance and evasion by sixty-seven wealthy families. Foreign and domestic personal holding companies, hobby losses, incorporated yachts and country estates, and personal service corporations were just a few of the devices resorted to by the well-to-do. For Roosevelt, plutocratic tax evasion and avoidance focused income tax reform upon the very small group of Americans already subject to income taxation instead of upon measures that would have broadened the tax base and the role of the income tax as a source of federal revenue.

Polls from 1938-39 suggest lack of public support for reductions in taxes "on people with high incomes", and a plurality of support for publicity of the income tax returns "of rich men." Proposals for base broadening were not well received. A poll in January 1938 asked if a married man earning less than $2,500 a year should be required to pay a federal income tax. Eighty percent said no, although that figure fell to 64 percent in a February 1939 poll that asked a similar question. This view supported exemption of over three-fourths of the American population from income taxation. (91)

When the income tax itself was, however, finally expanded to reach the labor income of the lower classes as well beginning in the early 1940s with World War II and the subsequent Cold War, the wage income of the lower and middle classes was, for the first time, doubly taxed at the federal level. But, as noted above, the rates were relatively low for the lower classes under both tax systems at this time. (92) Also as noted above, however, this is increasingly no longer true for lower- and middle-class taxpayers, bringing the double taxation problem for them to a head only now--in the twenty-first century. (93)


To the extent that payroll taxes are not really "taxes," the "double tax" (or high tax on labor income) fairness critique with respect to middle- and lower-class labor income disappears. As more fully described below, some economists have implicitly argued that the payroll taxes are not actually "taxes" to the extent that they purchase an equivalent amount of future benefits personal to that particular taxpayer. Further, politicians have been quick to jump on such implications. Senator Phil Gramm, for one, said that a payroll tax cut is tantamount to "giving a tax cut to people who do not pay taxes." (94) One of the key issues that must be resolved in determining the propriety of integrating the two taxes on labor income, therefore, is the question of whether payroll taxes paid with respect to a particular worker are properly viewed as purchasing future retirement and health benefits for that worker and are thus not a "tax." I ultimately reject this approach. While studies examining the tax paid and benefits received by various demographic cohorts can provide helpful information in designing the tax-and-transfer system and in predicting funding needs, their approach (if taken literally) fundamentally mischaracterizes the Social Security system.

In measuring tax burdens, some economists assert that the marginal payroll tax rate should be calculated by reducing the legislated rate to take account of the present value of the future old-age and survivors benefits expected to be received under the Social Security system upon retirement. As Mitrusi and Poterba phrase it, "From a life-cycle perspective, the marginal burden of the social security payroll tax on an individual's current earnings is not the current marginal rate, but rather this marginal rate net of the marginal increase in the present value of future program benefits that flows from higher current earnings." (95) Or, as Burkhauser and Turner stated, "When a marginal dollar of wage earnings yields an increase in future social security benefits, this offsets the stated payroll tax on that dollar." (96)

This analytical approach reveals that those retiring in the present (as well as in the past) will--as a group--receive far more in benefits than they paid in payroll taxes, and thus their marginal payroll tax rate during their working lives was low or negative. Those retiring in the not-distant future, however, will increasingly face higher and higher marginal payroll tax rates. For example, Burkhauser & Turner state:

In our calculations, we show that for retired workers aged 65 in 1982, the increase in the present value of social security benefits associated with an additional dollar of covered earnings at younger ages exceeded the additional social security taxes paid on those earnings. Thus, for workers who anticipated future benefits, social security acted at the margin as a wage subsidy rather than as a tax. This situation will not continue for younger workers now in the system. It is likely that for many workers born after World War II, the true payroll tax will operate as a tax throughout life, although at less than the legislated rate. (97)

As the quotation above indicates, Burkhauser and Turner thus characterize as a true "tax" only those payments that do not bring a directly related individual benefit, as opposed to the more diffuse benefits realized by society as a whole when general tax revenues are spent for the common good (i.e., for such goods as defense, our federal court system, etc.).

Others reach similar conclusions. The Congressional Research Service recently released a report that examined "the value Social Security provides each generation of workers in relationship to the Social Security taxes they pay." (98) It concludes:

Under the economic assumptions most commonly used by the Social Security Trustees and congressional policymakers, a worker who always earned the MINIMUM WAGE, who retires at age 65 in 2001, and has a dependent spouse will recover the value of the RETIRMENT PORTION of his or her [Social Security] taxes plus interest in 3.4 years; the payback times for a similar worker retiring in 2010, 2020, and 2030 are projected to be 3.9, 4.1, and 3.9 years, respectively.

In 1980, it took 4.4 years for a worker who always earned the Social Security MAXIMUM WAGE, and who retired at age 65 with no dependent spouse, to recover the value of his AND HIS EMPLOYER'S [Social Security] taxes plus interest. The payback period for a similar worker retiring in 2001 is 37.7 years, and in 2030 will be 102.6 years.

For workers with no dependent spouse retiring at age 65 in 2030 who always earned an AVERAGE WAGE, at a 1% real interest rate they would recover their and their employer's [Social Security] taxes in 15.1 years; at 2% in 22.1 years; and at 3% in 33.6 years. However, at 4% and higher real interest rates, the payback times become virtually infinite, i.e., the taxes would never be recovered. (99)

Stated a different way, Steuerle and Bakija conclude that "a high-/averagewage two-earner couple turning 65 in 2030 faces net taxes or negative transfers of approximately $173,500 . . .; their benefits will be worth about 74 percent of their contributions." (100) And stated in yet another way, Caldwell et al. conclude that people born in 1995 and later would enjoy a rate of return approaching zero, meaning that they "would barely get back what they put in, making the program akin to a mandatory zero-interest long-term savings account or to a program that transferred about 80 percent of their tax contributions to members of older generations and then invested the remaining 20 percent at a market interest rate." (101)

Mitrusi and Poterba reach similar conclusions on the numbers, and they also report how many taxpayers in each income spectrum have a higher marginal payroll tax rate than marginal income tax rate if the stated payroll tax rate is adjusted for net future benefits to be received under the Social Security system. (102) They conclude:

While the fraction of families in the lowest income category (less than $10,000 per year) for whom the marginal payroll tax rate exceeds the marginal income tax rate is still substantial (44.6 percent), for most other income groups this probability drops sharply. For all families, there is only a 9.2 percent chance that the net-of-benefits payroll tax rate exceeds the marginal personal income tax rate. For families with adjusted AGI above $20,000, the chance that the net-of-benefits payroll tax rate exceeds the income tax rate is very low. This very substantial change in part reflects the fact that small reductions in marginal payroll tax rates, beginning at a level of 15.3 percent, bring many families across a key threshold--the 15-percent income tax bracket. Small changes therefore change their payroll tax rate from greater than to less than their income tax rate. (103)

Like many generally stated statistics, these results can, however, take on an aura of solidity in general discourse that masks their inherent and deep uncertainties.

Not only do analysts disagree on the proper techniques to use in making calculations, there are often fundamental disagreements involving subjective factors: what work patterns to use; what part of the Social Security tax to count; whether or not to include dependents' benefits; whether or not to include the employer's share of the tax; what rate of interest to use; and whether or not to include the effect of income taxation of benefits. (104)

As one example, variations in the interest rate used to calculate the "return" on payroll taxes have huge effects on the results.

For example, workers retiring at age 65 in 2030 who always earned the average wage would recover the retirement portion of their and their employer's [Social Security] taxes in 11.9 years at a 1% real [i.e., above inflation] interest rate, in 16.1 years at a 2% real interest rate, in 23.5 years at a 3% real interest rate, and 44.4 years at a 4% real interest rate. At a 5% real interest rate, however, the point is reached where the annual interest earned on the accumulation of taxes plus prior interest income exceeds the annual benefit payments, so that the principal is never depleted--in fact, it plus interest grows indefinitely. (105)

I do not mean, however, to quibble with the interest-rate or other demographic assumptions made in these studies or even to quibble with their results. Rather, what I challenge is their implicit message that it is appropriate to view the Social Security system as nothing more than a government-sponsored pension plan when analyzing the distribution of "tax" burdens. If Social Security "taxes" are viewed as nothing more than the equivalent of government-mandated pension plan contributions, which will generate an individual return personal to the taxpayer, then they are arguably not "taxes" at all (and thus not a "double" tax on labor income).

One of the most vociferous opponents of the analogy of Social Security taxes to pension plan contributions is Patricia Dilley. (106) She views the analogy as nothing more than misleading rhetoric, (107) no matter how common and widespread the impression.

Payroll taxes are merely a method of financing the system, not the basis for benefits earned and paid out. Benefit calculations are made based on earnings recorded in the Social Security system .... [B]enefit calculations do not take into account the amount of taxes paid, and benefits cannot be reduced in the event of a failure to pay such taxes by the employer who is responsible for withholding [Social Security] taxes from workers' paychecks. The [benefit] system could as easily be financed through income tax revenues, like other government expenditures, without any impact on the earnings-based benefit structure. (108)

In short, Professor Dilley views the payment of Social Security taxes as unrelated to the receipt of future benefits. Rather, the retiree's economic right to a claim on the country's resources to support consumption in retirement is attributable to the claimant's prior demonstration of "worthiness" through his attachment to the workforce for the requisite number of quarters. It is the "lifetime of work," (109) not the payment of Social Security taxes, that raises the Social Security entitlement. This is the antithesis to both the retirement annuity model and the relief-of-poverty model, which is based on need rather than worthiness.

The public retirement entitlement is a public obligation designed to protect the public interest in social stability and orderly labor force exit by the elderly. The public entitlement is backed by the public taxing power and meets the public need for assurance of old age income security for all workers through redistribution of tax revenues. The notion of public advance "funding," and indeed the emphasis of the original designers of Social Security on payroll tax financing as the equivalent of private pension contributions, is an example of a useful analogy taking over the analysis and distorting the comparison beyond meaningful limits. (110)

Important design features of the system also highlight this disconnection between taxes and benefits. (111) For example, married workers who are paid less than their spouses will receive benefits based on their spouses' higher earnings. Thus, they receive no "benefit" for Social Security taxes paid by themselves and their employers. This is also true for individuals who work fewer than ten years.

Moreover, a portion of the Social Security tax base goes toward paying dependent benefits. Nevertheless, workers must bear the full payroll tax even if they have no dependents. In addition, only the thirty-five highest years of earnings are counted toward computing Social Security benefits. Thus, "a person who works for 15 years at $30,000 gets as much in benefits as someone who works for 30 years at $15,000; a person who works 50 years at $10,000 gets much less than someone who works 25 years at $20,000 and so forth.. " (112) Finally, the disconnection between payroll tax payments and benefits received, on an individual basis, is particularly evident in the case of Medicare, since there is no correlation between the amount of taxes paid and the amount of Medicare benefits that one can receive.

Probably the most fundamental disconnection between taxes and benefits, however, is that Social Security taxes paid today by Taxpayer X are not linked to Taxpayer X's future benefits (as would be an annuity contribution by Taxpayer X). Instead, these taxes fund payments to current retirees under our pay-as-you-go system. (113) Indeed, because annual Social Security tax receipts currently exceed by a substantial margin the amount currently paid in benefits each year, the surplus receipts have been routinely spent to support general federal government activities and programs.

Because the presently retired generation will receive more in aggregate benefits than they paid in aggregate taxes (even with an interest-like return factored in), the system is effectively redistributive between generations (unlike a private annuity system).

Tax rates, in particular, were relatively low for those who are currently retired or retired in the past, but have risen considerably for those who are relatively young today. Since the formula for determining benefits in Social Security gives no credit for paying at a higher rate of tax, individuals gain nothing out of paying a higher rate. One result has been a large redistribution among generations, with more going to those around when the system was younger and rates were lower. Some parts of this redistribution were intended: in general, the goal was to help the old through the contributions of the young, especially when poverty rates among the old were high. (114)

Moreover, unlike a private annuity, the "return" (benefits) on "principal" (essentially, average lifetime wages) is deliberately calibrated to be redistributive within a single generational cohort as well. That is to say, a greater percentage of low-income workers' earnings is replaced under the Social Security formula than is replaced for high-income workers.

The [Social Security benefit] formula divides earnings into three brackets. As in the income tax system, each bracket has a percentage rate associated with it. The first dollars of earnings are replaced at the highest percentage rate, the next level of bracketed earnings at a lower rate, and the highest earnings at the lowest rate. Workers with the very lowest levels of earnings have all their earnings contained within the first bracket and consequently replaced at the highest percentage rate. (115)

Under current law, those three percentage replacement rates are 90%, 32%, and 15%. (116) A worker with low earnings, in other words, may see 90% of those earnings replaced. A worker with somewhat higher earnings will see the first chunk replaced at a 90% rate but the next higher chunk replaced at only a 32% rate, and so on. Adding to the redistributive nature of the benefit formula is the fact that earnings of a worker above the taxable wage base (currently $84,900) are ignored in calculating the benefit to be received. (117)

All of these system-design elements, which are inconsistent with the private-pension analogy, point more generally to the public purposes underlying the Social Security system articulated by Professor Dilley. The taxes paid can be conceptualized as supporting society-wide goals and benefits rather than merely individualized cash benefits. More specifically, Social Security can be conceptualized as supporting society-wide goals and benefits that happen to require cash payments and the provision of health care during retirement. These goals and benefits are rooted in nurturing a stable economic environment through subsidizing consumption expenditures by the retired and by encouraging retirement itself, thus permitting progress in the workplace by succeeding age cohorts. (118) As Wilbur Cohen phrased it, Social Security was a "built-in stabilizer of consumer buying power, permitting the continued functioning of the market-price-profit system in a work-oriented free enterprise economy." (119) President Roosevelt, when signing the bill, also referred to this aspect of the new program.

It is a structure intended to lessen the force of possible future depressions. It will act as a protection to future Administrations against the necessity of going deeply into debt to furnish relief to the needy. The law will flatten out the peaks and valleys of deflation and of inflation. It is, in short, a law that will take care of human needs and at the same time provide the United States an economic structure of vastly greater soundness. (120)

Commissions appointed in the early years to study and make recommendations with respect to the Social Security system echoed this theme when they recommended using general tax revenues to help fund Social Security. They did so because of the general, more widespread societal benefits, including a more stable economic infrastructure, achieved through the system. For example, the 1937-1938 Advisory Council on Social Security Final Report stated:

Since the nation as a whole, independent of the beneficiaries of the system, will derive a benefit from the old-age security program, it is appropriate that there be Federal financial participation in the old-age insurance system by means of revenues derived from sources other than pay-roll taxes. ....

Governmental participation in financing of a social insurance program has long been accepted as sound public policy in other countries. Definite limits exist in the proper use of pay-roll taxes. An analysis of the incidence of such taxes leads to the conviction that they should be supplemented by the general tax program. The prevention of dependency is a community gain in more than social terms. (121)

The Advisory Council on Social Security, Old-Age and Survivors Insurance in 1948 similarly reported to the Senate Finance Committee as follows:

The Council believes that the Federal Government should participate in financing the old-age and survivors insurance system. A Government contribution would be a recognition of the interest of the Nation as a whole in the welfare of the aged and of widows and children. Such a contribution is particularly appropriate, in view of the relief to the general taxpayer which results from the substitution of social insurance for part of public assistance.

... In our opinion, the cost of financing the accrued liability should not be met solely from the pay-roll contributions of employers and employees. We believe that this burden would more properly be borne, at least in part, by the general revenues of the Government. (122)

And, writing in 1946, economist Eliot J. Swan wrote:

It would probably not be desirable to abandon the contributory principle entirely, but in a comprehensive system, benefits could be met in large part from general revenues. The wider the insurance coverage and the more general the understanding and acceptance of social insurance as a necessary public responsibility that, like public education, conveys indirect benefits to all members of the community, the more appropriate it becomes to draw upon general revenues. The right of individuals to benefits and the guarantee that future obligations will be met do not rest on the fact that claimants have a previous contribution record nor on reserves built up from such contributions, but on the harmony of social insurance with the economic and social desires of the nation. (123)

In short, the repeated recommendations that the government use general revenues to supplement payroll tax revenues are premised on the broader societal goals and purposes sought to be accomplished through Social Security. Thus, the Social Security tax can be viewed as an undifferentiated federal tax that supports widespread societal goals, just as the presence of a well-developed judicial system supports our economy and benefits all of us, not simply those who actually appear in court.

The temptation to link Social Security taxes paid by a particular worker to future benefits to be received by that particular worker is understandable. After all, the benefits (unlike most government benefits) are defined in cash and can be calculated, if imperfectly, on an individual basis. Nevertheless, the attempt fundamentally mischaracterizes the system for the preceding reasons. Taxes simply are not investments. Moreover, we do not commonly engage in the same inquiry with respect to the "income" tax paid. That is to say, we do not perceive the income tax burden nominally shouldered by an individual to be reduced by the value of benefits purchased with those taxes, such as a functioning court system, national defense, regulated capitalism, etc. Indeed, if we did so, those in the tax brackets that are nominally the highest might be seen as subject to much lower rates of taxation, since the wealthiest in society arguably benefit most from the expenditures necessary to maintain our regulated capitalist system. The illogical conclusion of such an exercise is that the only real "tax" paid by the population as a whole is the amount of revenue collected that i s wasted instead of spent to provide benefits to the population.

Return to the proposal briefly discussed earlier regarding the repeal of the Social Security tax wage ceiling--and the concomitant slashing of the Social Security tax marginal rates--as an alternative to explicitly integrating the tax burden of the two taxes on labor income. Significantly lower marginal rates under the Social Security tax, closer to historical standards, could deal with the equity problem while not creating a huge budget shortfall. This would be true, however, only if the progressive payment formula, under which wages above the $84,900 wage ceiling are not factored into the replacement formula, remained intact. What about the argument that repealing the ceiling on the tax side while maintaining it on the benefit side is unfair?

That argument is, once again, premised on viewing the Social Security tax paid by a particular individual as "purchasing" his or her future correlative retirement benefit. Once the Social Security tax is seen as a true tax that helps to support the infrastructure of our capitalist system by maintaining consumption spending in the retired, decreasing the severity of the hills and valleys of economic growth and contraction, etc., then that criticism is weakened. If the tax is separated from the benefit system, there is little reason to see why it should not also be progressive through broadening the tax base (which would allow the lowering of rates), just as the income tax is progressive.

But suppose that the reader is not convinced and rejects my discontent with linking payroll taxes with the future benefits to be received on an individual basis. Such a rejection does not, in fact, necessarily mean that the payroll tax burden on labor income and the income tax burden on labor income ought not to be integrated, though the equity arguments for doing so might be lessened. The reader who accepts the personal and individual link between the old-age and survivors benefit received and Social Security taxes paid explicitly conceptualizes Social Security taxes as forced individual savings, like a pension plan. Perhaps such forced savings ought to be deductible under the income tax (a form of integration) under the analogy to those pension plan contributions that are tax-favored under our 124 hybrid income/consumption tax. (124) The mandatory nature of the Social Security system, however, weakens any analogy to the tax-preferred retirement savings regimes, which seek to provide an incentive for increas ing voluntary retirement savings behavior by providing (more favorable) consumption tax treatment to such savings than would otherwise occur under a pure income tax. The more convincing conceptual approach involves how integration analysis might proceed if the link between the payroll taxes and benefits received is rejected. (125)


A. The Old-Age and Survivors Tax as Forced Individual Savings

As indicated in contemporary writings, early notions of "social security" were grandiose, involving funding for almost anything that could contribute to overcoming defective social organization in the modern industrialist state, such as education and vocational training, improved nutrition, planned utilization of the labor supply to adjust for changes in supply and demand, health care, housing-even the creation of enriching (126) forms of entertainment, such as opera. Today, however, many people view the Social Security system, particularly the old-age and survivors portion, as nothing more than a pension fund similar to those provided by private employers. This popular perception is acknowledged in the Interim Report submitted by the commission appointed by President Bush to study and make recommendations with respect to the Social Security system. "Many people believe that Social Security is a national pension fund in which workers make 'contributions' to an investment account called the 'Trust Fund.' When a worker retires, dies or becomes disabled, they believe that his contributions, plus interest, are taken out of an account to pay benefits." (127) And, as indicated earlier, there might actually be instrumental reasons for perpetuating this conception. What might the income tax consequences of this portion of the payroll tax be if it is conceptualized as forced individual savings?

Under a pure income tax, as conventionally described, the tax base consists of amounts spent on (or, in certain instances, the market value of) personal consumption plus net wealth increases (or less net wealth decreases). (128) While contributing to a savings account or purchasing another type of investment asset would be immediately deductible under a consumption tax, since it is an outlay that does not purchase immediate personal consumption, savings under a pure income tax are nondeductible capital expenditures. (129) The contribution to the savings account or the purchase of the investment asset changes the form in which wealth is held but does not diminish wealth and thus is nondeductible under an income tax.

Thus, if the payroll tax is considered to be an instance of forced individual savings, such as a deposit to a passbook savings account, it would be neither deductible under nor creditable against a pure income tax. Moreover, the inside build-up of the account would be taxable as it accrued. We do not, however, have a pure income tax but rather a hybrid income/consumption tax, with a realization requirement to boot, under which much savings goes untaxed, as it would under a consumption tax.

Chief among these favored savings are most contributions to pension plans and similar vehicles that comply with rigorous nondiscrimination, funding, vesting, and other standards. Both amounts contributed by employers and amounts contributed by employees (or eligible self-employed individuals) to these plans are excludable (or deductible), and the inside build-up attributed to the contributions are not taxed to either the employer, the employee, or the trust fund holding the assets. Amounts distributed from the trust are then fully includable when received. That is to say, employees enjoy consumption-tax treatment for this class of savings. Other classes of savings, such as the passbook account or pension-plan contributions to nonqualified pension plans, do not enjoy consumption-tax treatment.

The employer portion of the Social Security tax, the economic incidence of which is most likely borne by the employee, is essentially accorded consumption-tax treatment. The employee excludes the portion of "his" wages that his employer pays in tax, which is economically equivalent to inclusion of those taxes by the employee as wages followed by an offsetting deduction. Moreover, the inside build-up on the investment account is not taxed to either the employer, the employee, or the trust fund itself. Upon eventual payment, the rules in section 86, (130) which require inclusion of a portion of Social Security benefits by taxpayers exceeding certain thresholds of income, can be seen as roughly requiring inclusion of the benefits attributable to the employer portion of the tax. (131)

What about the employee portion, however? From the beginning of the Social Security system, payments of the employee portion of the Social Security tax have never been deductible or excludable from the employee's gross wages for income tax purposes. The inside build-up on the account is, like the employer portion, not taxed to anyone. Why is not the employee portion of Social Security taxes deductible under the income tax, just as are some employee contributions to qualified pension plans and regular individual retirement accounts (IRAs) today? If it were made deductible under this pension analogy, then section 86 should also be amended to require full inclusion of benefits received, just as with other pension distributions. This revision would greatly simplify the law. (133)

One reason, of course, could be that the pension analogy simply is 134 inapposite for the reasons described above. For present purposes, however, if we indulge in the notion that the employee portion of the Social Security tax is analogous to a pension-plan contribution, why is it not deductible?

Such an approach would likely have significant after-tax consequences for most taxpayers. If a taxpayer's tax bracket is constant over time, Daniel Halperin has shown that the

deferral of compensation [under the consumption-tax treatment provided to qualified plans] does not affect the tax burden on the compensation itself. While the tax is deferred, the tax base is increased by the after-tax rate of return on investment. In present value terms, the tax bases and thus the resulting tax are equal. (135)

But many, if not most, taxpayers are likely in higher tax brackets in their earning years than in their retirement years. If this is true, then deduction of the payments in the high-earning years of middle age and inclusion in the low-earning years of retirement can mean a higher total after-tax return than would occur if Social Security taxes were nondeductible during the earning years and benefits were excludable. A slightly different way of looking at it is that, if the taxpayer's marginal rate is lower in retirement, inclusion at that time is the effective equivalent of taxing a portion of the taxpayer's lifetime earnings at this lower rate. Perhaps such a policy is worthwhile as a rough form of lifetime averaging. In short, marginal rates are not typically constant through the lifecycle, and most taxpayers are more likely to be in a higher tax bracket in the years of payment than in the years of receipt.

Several reasons converge to explain why Social Security taxes were not made deductible by analogy to deductible employee contributions to qualified plans. Deductible employee contributions to pension plans and IRAs are a relatively recent phenomenon, beginning in earnest only in the 1980s. There were not many pension plans in the early twentieth century, and employers funded the few that were there. (136) Economic incidence analysis aside, the employer was seen as funding the plan, not the employee. Moreover, most employees did not even owe income taxes, since, as briefly described earlier, income taxes were imposed on only a small minority of the wealthiest citizens. Therefore, an income tax deduction for the few employee contributions that were made would have been a moot point. Finally, the income tax consequences of pension plans were themselves just beginning to be threshed out in the era when the Social Security system was adopted. (137) This combination of facts, explored in more detail below, seems th e most likely explanation of why the employee portion of the Social Security tax was made nondeductible under the income tax.

When Congress adopted the Social Security system, there were few pension plans and few participants.

When the Social Security program was enacted, only about six million persons, 15 percent of those employed, held jobs covered by any sort of retirement system; only a tiny handful--perhaps 100,000 to 200,000--actually were receiving a pension. The poorhouse toward the end of life--with all its horrors--was a very real part of America. (138)

The most widely accepted estimate of pension plans and participation is that between 1875 and 1929 around 400 industrial pension plans were established, with about that number still in operation on the eve of the Depression.

The companies establishing these plans employed about 10% of the industrial labor force--a force substantially less than the entire work force including agricultural and casual labor. It is not clear, however, how many of those employees were eligible to participate in the plan or would eventually qualify for benefits. According to a recent estimate, although industrial pensions were not uncommon by 1935, no more than 4% of male workers and 3% of female workers met the requirements for receipt of pensions at that time. (139)

Moreover, almost none of the plans that did exist were funded in advance. Rather, they took the form of promises by employers that employees would be paid in retirement if the employee were still employed with the same employer at that time.

At the time of the stock market crash in 1929... , only a small minority of American workers were covered by private pension plans and an even smaller percentage could ever expect to draw benefits because of long vesting periods and eligibility terms that often required the worker to be employed by the firm at the time of retirement in order to receive benefits. Moreover, most pension plans were not funded in advance, or were inadequately funded, as no legal funding requirement was imposed on employers until much later. (140)

Industrial pension plans of large businesses, such as railroads, evolved as an aid to personnel management, to ease the older worker out of the work force

in a manner that economized on labor costs while producing a smooth flow of employees through the firm.... The introduction of the mandatory retirement provision arrangement promised to harness expenses. . . by allowing the firm to replace older, high-wage workers with smaller numbers of younger, lower-paid employees. Managers believed these savings would substantially offset the additional cost of paying pensions to retiring employees. (141)

These early plans were pay-as-you-go plans, with employers paying benefits out of current earnings, and these payments were deductible by employers as ordinary and necessary business expenses under a 1911 ruling interpreting the 1909 corporation excise tax that pre-dated the income tax. Problems with funding crunches soon arose, however, and reformers argued that employers should treat pension costs as accruing during the entire working life of the employee instead of expensing them when paid after retirement. (142) The reformers also encouraged advanced funding of these pension promises. In 1919, the Internal Revenue Service (Service) allowed employers to deduct advanced-funding payments to a trust so long as the trust was "organized entirely separate and distinct from the corporation, having its own set of books, making its own investments, and paying its own expenses, legal title of which does not remain in the corporation . . ." (143) The employees on whose behalf these contributions were made did not hav e to include the amounts in their incomes until actually received, since (with the typically extremely weak rights the workers had in these pension funds) it was not at all certain that the employees would ever, in fact, receive them. (144) Thus, the big income tax issue in the early days of funded plans was whether the separate trust should be taxed on the income earned on contributions in the interim between the contribution and distribution to retired employees. In 1921, Congress provided that stock bonus and profit-sharing trusts were not taxable so long as the trusts were made irrevocable by the companies funding them. While unclear whether irrevocable pension trusts were similarly treated under the 1921 law, even though not explicitly mentioned, the Code was formally amended in 1926 to make pension trusts tax-exempt. (145)

Thus, when the Social Security system was adopted, employer contributions to pension trusts were deductible (if the pension trust qualified), employees did not include these contributions in their gross income, employees did not include the earnings on these contributions in their gross income while they accrued, and they included benefits only when actually received. The employer portion of Social Security tax payments mirrored this treatment to a great extent. Employers were allowed to deduct these payments immediately, the employee did not include as wages that portion paid by the employer, and no earnings on the employer's contribution were includable by the employee prior to retirement. One big difference is that none of the Social Security benefit received in retirement, even that portion attributable to the employer contribution that was never taxed to the employee, was includable by the employee at the time the Social Security program was adopted. As noted earlier, a portion of Social Security benefit s may today be includable for income tax purposes. (146)

As for the employee portion of the tax, it was nondeductible, as were other employee contributions (if any) to pension plans, but most employees did not owe income tax under the "class tax" of the time. Therefore, a deduction by the employee under the income tax for the Social Security tax that he or she paid would not have been a ripe issue for the vast majority of employees.

When the class tax became a mass tax with legislation in the early 1940s, however, the tax treatment of employee contributions to pension plans did become a ripe issue. Writing in 1943 in the Harvard Law Review, Harvard Dean Erwin N. Griswold argued that Congress ought to change the law so that employee contributions to pension funds, including the employee portion of the Social Security tax, would be deductible under the income tax when made, placing them on the same footing as contributions by an employer. Thus, there are two interesting features to his article: the substantive change that he argued with respect to employee contributions to pension plans, and the fact that he assumed, without discussion, that the employee portion of the Social Security tax was properly conceptualized as such a contribution. He described the law at that time as follows:

Where the cost of providing retirement funds is paid entirely by the employer under an approved pension plan, the employee pays no tax until the retirement benefit is actually paid to him. But this is not the case with respect to amounts which the employee himself pays to provide his future pension benefits. The money which he earns is taxable to him when he earns it, and the amounts which he pays to provide retirement benefits are not deductible by him, whether he pays them directly himself or they are withheld from his pay by his employer pursuant to the contract of employment. Thus, the law specifically provides that the amount of social security tax withheld from an employee's wages is not deductible in computing his income tax. There is a similar provision as to amounts deducted from wages under the Railroad Retirement Act. The Treasury ruled some time ago that the amount of employees' contributions under private pension plans is not deductible in computing their tax, and this ruling was applied to deduc tions from the salaries of municipal employees. The same position has long been taken by the Treasury with respect to deductions from the salaries of federal employees under the Civil Service Retirement Act, and this conclusion has recently been sustained by the Tax Court. This rule is in effect legislated into the Internal Revenue Code by the recent pension trust amendments. (147)

He then went on to criticize that law.

From the point of view of the employee, a true pension or retirement allowance is income in the year in which he receives the money.... What the employee earns during his productive years must, for all practical purposes, be spread over the period of his life. What he receives after his retirement is in reality his income then, for then is when it comes in to him. To tax him on it at the top bracket of the graduated rates of his earning years is an unfair failure to recognize the economic facts.

Arguments of the sort just outlined have been adopted in the statute to make pension contributions of the employer taxable to the employees only when the pension payments are actually received. But no such consideration is given to the pension contributions of the employee, and the question may fairly be asked whether there is any sound basis for this distinction.... [I]t is hard to find any substantial reason for making a distinction between amounts paid by the employer to provide future pensions and those withheld from the employee for the same purpose. In both cases, the employee's current productive capacity is being utilized to make provision for his retirement. Neither amount is received by the employee any more than the other, for he does not have any more right to obtain presently the amount withheld from his pay than to obtain the additional amount paid by the employer. Indeed, the distinction between the amounts paid by the employer and those withheld from the employee is almost completely formal. A plan may be set up under which the employee's salary is stated to be $100, and the employer withholds $5 for pension purposes and pays an additional $5 to the pension fund. Under such a plan, the employee actually receives payment of $95 of his salary and $10 is paid to the pension fund. Precisely the same economic result would be reached if the employee's salary were stated to be $105, with $10 withheld by the employer as the employee's contribution to the pension fund, or if the employee's salary were fixed at $95, with $10 paid to the pension fund by the employer. Yet the tax consequences will vary sharply according to whichever one of these formal plans is used.

As long as the plan is really a pension plan, the reasons which have already led to the conclusion that the employer's payment in such a case should not be taxable to the employee until the employee actually receives it, should lead to the same conclusion with respect to the similar payments which are withheld from the employee's wages, either under state or federal law or under the terms of the employment contract. To achieve this result, the tax statutes should be expressly amended so as to provide that amounts paid by an employee to provide bona fide pension benefits after his retirement should be deductible from his current income. (148)

In effect, Dean Griswold argued that the reality of economic incidence ought to be recognized. The distinction between whether the "employer" makes the payment or the "employee" makes the payment is solely formal. The economic incidence will fail on the employee in the form of depressed current wages in an equivalent amount. Whether the payment is nominally said to fall on the employer or employee should not make a difference in tax treatment. If the "employer' s" contribution should be excluded from the employee's wage income for income tax purposes, then so should the "employee's" contribution. (149)

While the Service did approve as "qualified plans" some salary reduction plans in 1956, (150) his argument did not carry the day legislatively until enactment of the IRA provision in 1974 and enactment of section 40 1(k) in 1978. Widespread adoption of section 401(k) plans utilizing employee contributions did not occur until the 1980s (151) So, at least as a practical matter, it has only been since the 19 80s that there has been a noticeable difference in treatment between some employee contributions to qualified pension plans (deductible) and the payment by the employee of the employee portion of the Social Security tax (nondeductible).

One response to the argument that Social Security taxes ought to be deductible because they constitute forced savings analogous to qualified pension plan contributions might be that the Social Security system is technically not a "qualified plan." This argument, however, is not persuasive. To be sure, the precise features regarding vesting, participation, etc., required of qualified plans are not present in the Social Security system, but the "big idea" picture is absolutely present. Under federal law, Social Security participants are vested after a certain number of paid quarters in the work force, there is no discrimination in favor of highly paid participants, withdrawal cannot be made until age sixty-five, etc. In other words, the concerns that led to the strictures imposed on qualified plans are concerns that are dealt with in an equally effective way under the Social Security system.

A more decisive argument for rejecting the qualified-plan analogy, however, is that the decision whether or not to make a contribution to a section 401(k) plan or an individual IRA is typically voluntary on the part of the employee, while Social Security tax payments are mandatory. The argument would be that only voluntary savings need to be made deductible (in deviation from a pure "income" tax) in order to encourage the taxpayer to make the savings decision. Since Social Security tax payments are mandatory, there is no similar incentive effect at play, and tax-favored treatment for certain types of savings (as opposed to all savings) is justified only because of is incentive effect on behavior.

Scholars disagree about whether the origins of the favorable treatment accorded qualified pension plans and similar arrangements were premised on this incentive effect. James Wooten argues that policymakers at the time appreciated that these earnings ought to be attributed to workers, not to the employers, and that most workers owed no income tax at the time. Therefore, the tax exemption of the pension fund was adopted, he argues, in order to avoid a bias against deferred compensation, i.e., to achieve tax neutrality with current compensation (a rate of zero for most workers). (152)

Norman Stein, on the other hand, sees little evidence for such prescience on the part of policymakers. He finds other evidence indicating that the exemption was adopted simply to encourage employers to adopt funded pension plans. He believes that the record suggests that both Congress and the Executive shared an understanding that funded pension and profit-sharing plans served a public virtue and that a tax-deferral regime should be extended to such plans in order to encourage employers to sponsor them. The perspective from which the government viewed the question of tax subsidization, however, would not have been through the modem lens of plans as tax-advantaged deferred wages (and thus as tax savings to the participants), but from the then predominant understanding of plans as corporate commitments to make future gifts. Congress might have seen allowing a current deduction of employer contributions, and exemption of plan income, as a direct tax advantage to the employer rather than to the employee. (153)

[I]t is not at all clear that tax advisors, or legislators, of the day, thought that individuals not subject to tax were subjected to a tax penalty when they invested in a taxable financial intermediary. . . . Most significant though, in a profit-sharing plan with delayed vesting, the employer, rather than the employee, might have been understood as benefiting from the deferral until the employee obtained a nonforfeitable interest. Thus, the exemption of the profit-sharing trust might have been seen as a subsidy to the plan sponsor (tax-exempt advance funding of compensation that did not have to be paid until the future), which might have had the effect of encouraging employers to adopt profit-sharing plans. (154)

But it is probably not necessary to resolve this dispute for current purposes because, regardless of the origins of the tax-favored treatment, the commonly cited reason why tax-favored treatment is continued today is precisely the reason described by Stein: to encourage employers to sponsor plans. Even though we are concentrating on the employee contribution, the voluntariness of the payment remains critical to assessing whether it ought to be deductible under a "forced savings" analytical construct. For example, in 1981, Congress allowed deductions for voluntary employee contributions to certain plans in lieu of a contribution to an IRA but denied deductions "for so-called mandatory employee contributions, i.e., those required as a condition of employment or in order to participate in the plan or to obtain the benefit of employer contributions." (155) The voluntariness of the contribution is thus a crucial ingredient to the current premise underlying tax-favored treatment for only certain kinds of savings, a nd not others, under our income/consumption tax. (156) It therefore can be argued--if this "forced savings" analysis is to be used at all--that the tax consequences of the employee payment of mandatory employee Social Security taxes should mirror the tax consequences of disfavored savings, such as an employee contribution to a nonqualified pension arrangement, which is typically not deductible by the employee. (157)

B. The Medicare Tax as Insurance Premium

As stated earlier, I believe the case for linking the Medicare tax to future medical care to be received is even weaker than is the case for linking the Social Security tax with the receipt of old-age and survivors benefits, since there is no correlation between the amount of tax paid and the benefits received. A reader who once again disagrees with this contention explicitly conceptualizes the payment of Medicare taxes as the payment of an insurance premium to cover future medical care. How would the tax payment and the receipt of medical care be treated under the income tax under such a conceptualization?

If the payment of the Medicare tax is explicitly viewed as a payment for medical insurance that will provide care later in life, then the payment should be treated like any other health-insurance-premium payment for income tax purposes. That is to say, it would generally be treated as a personal-consumption expense but would nevertheless be deductible to the extent that it, combined with other healthcare expenditures for the year, exceeded 7.5% of the taxpayer's adjusted gross income. (158) The receipt of medical care in kind later in life would then not be includable, since it would simply constitute the actual receipt of what was purchased earlier. The fact that a particular taxpayer may have paid total "premiums" that do not approach the fair value of the medical care actually received should not be troublesome, as this happenstance commonly occurs in all fields of insurance, and the government does not tax such gains under the income tax.


Above, I rejected the link between the old-age and survivors portion of the Social Security tax and the. Medicare tax and any future individual benefits that the taxpayer may receive. (159) I view the payroll taxes as undifferentiated federal taxes used to support federal spending (for the currently retired generation). Under that view, the taxes paid and benefits received by any particular taxpayer ought to be analyzed independently of each other for income tax purposes. How might the integration analysis proceed under such a view?

One possibility to consider is whether the payment ought to be deductible under an income tax for reasons that are quite different from the earlier discussion regarding the qualified plan analogy. Another possibility is whether it ought to be creditable against that tax. In both cases, the benefits received, severed from any connection with the tax, should also be analyzed separately under income tax theory. Parts A and B consider the possibility of a deduction or credit, respectively, for the old-age and survivor's portion of the Social Security tax, as well as the tax consequences of the receipt of cash benefits on retirement. Part C considers the potential tax consequences of the receipt of Medicare services.

A. An Income Tax Deduction for the Severed Social Security Tax

As noted earlier, a pure income tax base is conventionally assumed to include both personal consumption expenditures and savings outlays. Thus, expenses (as opposed to capital expenditures) incurred for business or investment purposes are generally deductible. A mechanical approach to this analysis might conclude that, since the payment of a tax is not in pursuance of an income-producing activity, it must be nondeductible personal consumption. Going back to first principles, however, one could argue that such an approach is too simplistic.

The principle underlying the choice of income taxation as a tax base is generally agreed to be the ability-to-pay principle. (160) It is because a tax base of income" is thought to best represent one's ability to contribute to the fisc that it is adopted as the tax base. If that is true, then even certain personal expenditures ought to be deductible if they sufficiently compromise ability to pay. Indeed, this view underlies the personal exemption and standard deduction, which protect a bare subsistence amount from taxation, though clearly personal consumption. It also explains the deduction for extraordinary medical expenses (i.e., those in excess of 7.5% of AGI) (161) and for extraordinary personal casualty losses (i.e., those in excess of 10% of AGI). (162)

In one sense, then, mandatory government extractions of many sorts ought to qualify for deduction since they represent amounts not available for contribution to the fisc--even if one chooses to call them "personal consumption" outlays because they are not directly connected to an income-producing activity. As noted earlier, it makes no sense to allow deduction of the income tax itself under the income tax, even though it is a mandatory extraction, since it would result in a greater administrative burden with no fundamental change in the distribution of the tax burden. That is not true, however, of other taxes (so long as they are sufficiently nondiscretionary in nature), such as certain state and local taxes and, notably, Social Security taxes. While not directly mentioning Social Security taxes, Joseph Dodge stated:

Taxes ... fulfill the requirements of prima facie deductibility. The fact that a given taxpayer might reap personal benefits from government having a value equal to the taxes does not negate the fact that the taxes are nondiscretionary expenses, since the taxpayer had no control over how the taxes are spent. The benefits are fortuitous. (163)

For this reason, Social Security taxes could defensibly be made deductible under income tax theory.

Under the view that the income tax consequences of the Social Security tax ought to be analyzed separately from the receipt of benefits, the taxation of the benefits themselves should also be analyzed separately. As a cash payment that is not a "welfare" payment (164) and therefore which generally represents ability to pay, it should be fully includable when received under this analysis. The operation of the personal exemption, standard deduction, and lower marginal rates on low earnings under the income tax would combine to protect the ability-to-pay value on receipt.

Notice that this combination--deduction of the taxes in the earning years and inclusion of the payment in the benefit years--is precisely the same result obtained under a "forced savings" analysis which further accepts the characterization of this "forced savings" as comparable to tax-preferred contributions to certain pension plans. As argued earlier, I believe that both the "forced savings" analysis is itself weak and (even if the forced savings characterization is accepted) that the analogy of this forced savings to tax-preferred pension plan contributions is weak. (165) Nevertheless, we end up at the same place!

Ultimately, however, I believe that the deduction approach is the second-best alternative. The main reason why state and local taxes are deductible, rather than creditable, under the income tax is to prevent states from raiding the Treasury and completely shifting all tax receipts to state coffers. (166) That worry is not present with the federal payroll taxes, and I believe a credit for such taxes is conceptually the most defensible approach.

B. An Income Tax Credit for the Severed Social Security Tax

The idea of a credit under the income tax for Social Security taxes is not new. One of the original purposes of the earned income tax credit (167) was "to provide a wage bonus to low-wage workers in order to offset the burden of the social security payroll tax," (168) though it has long ago exceeded those modest beginnings. Moreover, in 2001, Representative Thomas M. Barrett of Wisconsin introduced a bill that would provide an individual refundable credit against the income tax of up to $300 ($600 for joint filers) of payroll taxes. In their proposed cash-flow consumption tax replacement for the Code, Senators Nunn and Domenici would "allow both businesses and individuals to offset their consumption tax liability with the Social Security payroll taxes they pay .... " (170) Further, Jonathan Forman and others recommended that Social Security taxes be made partially creditable against the income tax. (171)

The argument for a credit is that the two taxes ought to be fully and explicitly integrated simply because they are both taxes on wage income to support federal expenditures. Collecting federal taxes on labor income under two separate systems masks the high effective combined tax rate imposed on the middle and lower classes. The Social Security tax is not viewed, in this analysis, as simply a forced expenditure of any old sort that ought to be deductible under income tax theory because it is nondiscretionary and thus reduces ability to pay, but rather as the payment of an actual tax on wages, equivalent to the income tax on wages, both of which raise revenue for federal spending.

Allowing an income tax credit for Social Security taxes paid raises the additional issue of the extent to which payroll taxes paid in excess of income tax owed ought to be refundable under this conceptual approach. The resolution of this issue depends not only on the obvious revenue constraints of unlimited refundability but also, perhaps, on whether the lack of any exemption in the payroll tax system is troublesome. Assume, for example, that the employee portion of the old-age and survivors tax is made creditable against the income tax up to $1,500, that a waitress earns $25,000 and pays $200 in income taxes and $1,500 in the employee portion of the old-age and survivors tax, and that a lawyer earns $200,000 and pays $40,000 in income taxes and $5,000 in the employee portion of the old-age and survivors tax. Allowing a dollar-for-dollar credit of these payroll taxes against income taxes with no refundability feature would eliminate the waitress's meager income-tax bill but would not truly recognize her decre ased ability to pay taxes, since the payroll tax attaches to the first dollar of earned income with no exemption. Indeed, all the credit would do is significantly reduce the lawyer's income tax bill.

Perhaps payroll taxes ought to be refundable in an amount not to exceed the equivalent of a reasonable personal exemption. For example, in my above hypothetical, suppose that payroll taxes were creditable against the income tax up to $1,500 and that such credit was fully refundable since it is determined that $1,500 is a reasonable "personal exemption" equivalent. The waitress would now owe no income tax and would obtain a $1,300 tax refund. The lawyer's income tax would be reduced by $1,500. The effect of such a ceiling would mean that the lawyer receives no payroll tax "personal exemption," since his payroll tax is far less than his income tax, and thus refundability--where the personal exemption "equivalent" would kick in-would never come into play.

Whether or not refundable, a credit approach severs the link between the tax and benefit and thus would also require independent analysis of the receipt of cash benefits in retirement. As a cash payment that is not a "welfare" payment and therefore which generally represents ability to pay, it should be fully includable under this analysis. (172) The operation of the personal exemption, standard deduction, and lower marginal rates on low earnings under the income tax would combine to protect the ability-to-pay value on receipt.

C. The Medicare Portion of the Payroll Tax

If the Medicare tax is viewed as an undifferentiated federal tax to support federal spending, and thus separate from the receipt of any medical care under the program later in life, then the payment and receipt should each be analyzed separately. For the same reasons that it would be defensible to make the Social Security tax either deductible or creditable in the income tax under such an approach, the Medicare tax might also be made deductible or creditable. That analysis need not be restated here. The interesting question here in either case would be how to treat the receipt of medical care under the program later in life, analyzed separately as an independent receipt.

Whereas the receipt of cash benefits under the old-age and survivors program represents resources over which the taxpayer has control and which can be spent in any fashion--and thus represents taxable ability to pay--the receipt of personal consumption in kind in this instance arguably does not and thus should not be taxable.

Under a rigid and mechanical application of the Shanz-Haig-Simons concept of income, some might argue that personal consumption received in kind (i.e., for free) ought, as a theoretical matter, to be valued and included in the tax base in all events, unless a specific exclusion provision can be found in the Code.173 The fact that everyone enjoys free personal consumption literally every day that goes untaxed, such as the free concert while waiting for the subway in the morning, is more a matter of administrative convenience than grand theory.

Another view, however, is that the personal-consumption component of the Shanz-Haig-Simons concept of income should be interpreted with the underlying ability-to-pay norm that informs it clearly in mind. The reason why resources freely spent on personal consumption (above a bare subsistence amount) are taxed is because they represent resources under the control of the taxpayer that could fairly be called upon for contribution to the fisc. Consumption received in kind may or may not be, depending on the circumstances.

The expression "net increases in wealth plus consumption" is susceptible to two readings because consumption can either be purchased or simply received in kind without charge. (An example of in-kind consumption without charge would be baby-care products sent to the parents of septuplets by publicity-seeking manufacturers.) Under one reading, "net increases in wealth plus consumption" means gross increases in wealth (inflows or receipts other than consumption received in kind without charge) less gross decreases in wealth (outflows or outlays including consumption expenditures) plus consumption (whether purchased through expenditures or received in kind without charge). This reading, which probably has the greater number of adherents, normatively mandates taxing the value of consumption received in kind without charge. Under the other reading, "net increases in wealth plus consumption" means gross increases in wealth (other than consumption received in kind without charge) less decreases in wealth-other than c onsumption spending. This second reading treats "consumption" as only a principle of nonsubtraction (i.e., nondeductibility) for purposes of calculating decreases in wealth. Consumption expenditures are taxed because they are not deductible (and thus remain in the tax base) but consumption received in-kind is not considered an increase in wealth. In other words, the second interpretation of "net increases in wealth plus consumption" treats taxable consumption as equal to the amount spent by the taxpayer on consumption and ignores consumption received in kind. (174)

If the taxation of personal consumption is seen chiefly as a rule of nondeductibility, then whether the taxation of consumption received in kind ought to be valued and taxed would turn on whether the receipt can fairly be viewed as the equivalent to the receipt of cash followed by a free-spending choice by the taxpayer--a nondeductible outlay. If it can, as should typically be the case in the employment context and with the provision of in-kind consumption by closely held corporations to their owners, then the consumption should be valued and included. The receipt in kind can fairly be seen in that case as freeing up cash or assets that we feel comfortable the taxpayer would have otherwise spent himself absent the free receipt, (175) and the freed-up cash or assets represent ability to pay. If it cannot be--if the receipt does not fairly represent the receipt of cash followed by a free-spending choice by the taxpayer--then the receipt should not be seen as truly representing ability to pay in the sense that c onsumption expenditures chosen by the taxpayer represent ability to pay. (176) If we can not be sure that the taxpayer would have purchased the consumption himself, we can not be sure that other cash or assets were freed (and thus available for contribution to the fisc) from having to be spent on the same consumption. And the receipt--services or quickly consumable assets with little or no utility to anyone else in the marketplace--does not itself represent ability to pay.

Under this view, the free receipt of medical care under the Medicare program would not be valued and included in the income tax base. It is fair to generalize that it is unclear whether the taxpayer would be able, absent the provision of the free care, to purchase the care himself. We can not be sure that the receipt of free care freed up other assets (thus representing ability to pay) that otherwise would have been spent on the medical care, and the medical care itself does not represent ability to pay.


The policy of nondeductibility and noncreditability of payroll taxes under the income tax, though deliberate in 1935, has continued without reexamination as the income tax world around the payroll tax has changed dramatically. Few payroll taxpayers in the 1930s paid income tax, which was a tax paid only by the wealthy minority of the population. Moreover, few thought the income tax would ever be paid by the middle and lower classes.

When the income tax was expanded to reach the middle and lower classes, however, suddenly the wages of the lower and middle classes were subject to double taxation at the federal level--once under the payroll tax and once under the income tax. For many years, this situation was ameliorated by the fact that both the payroll tax and the income tax rates imposed on the middle and lower classes were relatively low. With the dramatic expansion of the payroll taxes in the last several decades, however, this is no longer true today.

Payroll tax revenue now nearly equals personal income tax revenue, and this payroll tax burden is borne predominantly by the middle class. The combined employer and employee payroll tax rate is now more than fifteen percent, and nearly two-thirds of households pay more in payroll taxes than they do in income taxes. This is true particularly in low- and middle-class households--those with incomes of less than $100,000 per year. At the same time, the wealthiest households are experiencing significant estate and income tax decreases while also enjoying an ever-growing share of after-tax national wealth. Even before enactment of the 2001 Act, Mitrusi and Poterba concluded that the poor and middle classes experienced a tax increase between 1979 and 1999 because of the increasing take of the payroll taxes, while the wealthy enjoyed a tax decrease. The magnitude of these trends was exacerbated with enactment of the 2001 Act. For reasons of equity, the time has come to begin considering whether the payroll tax burden ought to be integrated with the income tax burden through either a deduction or a credit. I believe that the most conceptually defensible means to achieve integration would be to allow workers a refundable credit for a portion of payroll taxes paid against income tax owed in an amount equal to a reasonable "personal exemption" equivalent. Social Security benefits would be fully includable under the income tax when received in retirement, while medical care received in kind would continue to be excludable.

Having said this, my cynical side thinks it unlikely that any such proposal could get far in the near future. One reason is that such a plan would be expensive, and there is no more budget surplus to pay for it. For this reason, an alternative to integration might be to repeal the Social Security wage ceiling of $84,900 (just as under the Medicare tax) and slash the marginal rates as low as possible to retain revenue neutrality, while maintaining the payment formula (with the wage ceiling) as it is today. With low marginal rates under both the Social Security and Medicare taxes, the multiple tax burden on the labor income of the poor and middle classes should not be as objectionable as it is today. Once the Social Security and Medicare taxes are seen as true taxes that fund government spending that helps to support the infrastructure of our regulated capitalist economy, and not as equivalents to private pension plans or insurance contributions, then objections to repealing the wage ceiling should be muted. If progressivity in the tax burden is generally justified, then there is no reason why the Social Security tax should be predominantly borne by the middle and lower classes.

1999 Family                       Families      Mean      payroll
Adjusted                         (millions)   Personal      Tax,
AGI (000s)                                   Income Tax   Employer
                                             (dollars)   & Employee

All Families

<10                                28.83         -262 *       400
10-20                              23.65         -239 *     1,475
20-30                              18.11        1,062       2,833
30-40                              13.26        2,519       4,199
40-50                              10.57        3,942       5,405
50-75                              18.92        6,248       7,170
75-100                              9.93       10,629       9,950
100-200                             9.30       20,755      12,570
200-500                             2.03       64,481      14,394
500-1000                            0.36      182,264      17,873
>1000                               0.18      799,100      28,848
All                               135.15          841         603

All Families With Tax Liability

<10                                13.70         -551 *       842
10-20                              19.16         -285 *     1,821
20-30                              17.25        1,118       2,975
30-40                              13.05        2,562       4,267
40-50                              10.52        3,962       5,432
50-75                              18.85        6,272       7,197
75-100                              9.92       10,636       9,957
100-200                             9.29       20,775      12,583
200-500                             2.03       64,503      14,399
500-1000                            0.36      182,345      17,880
>1000                               0.18      798,446      28,853
All                               114.31        7,350       5,270

                                    % of         % of
1999 Family                       Families     Families
Adjusted                            with         With
AGI (000s)                        Combined     Employee
                                 Payroll Tax  Payroll Tax
                                 > Personal   > Personal
                                 Income Tax   Income Tax

All Families

<10                                 46.4         46.2
10-20                               71.6         58.3
20-30                               81.7         43.2
30-40                               85.7         34.9
40-50                               74.5         27.9
50-75                               66.3         19.1
75-100                              55.0          6.3
100-200                             16.2          1.5
200-500                              0.1          0.1
500-1000                             0.0          0.0
>1000                                0.0          0.0
All                                 62.1         34.6

All Families with Tax Liability

<10                                 97.6         97.5
10-20                               87.9         71.4
20-30                               85.5         45.1
30-40                               86.9         35.2
40-50                               74.9         28.0
50-75                               66.5         19.2
75-100                              55.1          6.3
100-200                             16.2          1.5
200-500                              0.1          0.1
500-1000                             0.0          0.0
>1000                                0.0          0.0
All                                 73.2         40.8

* The negative income tax is due to the impact of the earned income tax


(000s)                           1979  1984  1989  1994   1999

All Families

<10                              46.4  46.3  46.4  46.4   46.4
10-20                            71.2  70.8  71.3  71.5   71.6
20-30                            66.7  78.8  81.1  81.5   81.7
30-40                            45.8  58.8  84.5  85.3   85.7
40-50                            37.5  48.0  68.9  73.5   74.5
50-75                            19.1  28.8  63.4  65.5   66.3
75-100                            4.6   7.3  41.3  49.1   55.0
100-200                           1.0   1.5   9.7  13.4   16.2
200-500                           0.4   0.4   0.1   0.1    0.1
500-1000                          0.2   0.3     0     0      0
>1000                             0.4   0.5     0     0      0
All                              41.8  47.0  59.5  61.1   62.1

Families with Positive Personal
Income Tax or Payroll Tax

<10                              97.7  97.4  97.6  97.6   97.6
10-20                            87.5  87.2  87.6  87.8   87.9
20-30                            70.0  82.7  85.0  85.4   85.5
30-40                            46.4  59.7  85.8  86.5   86.9
40-50                            37.7  48.2  69.2  73.8   74.9
50-75                            19.1  28.9  63.7  65.8   66.5
75-100                            4.6   7.3  41.3  49.2   55.1
100-200                           1.0   1.5   9.7  13.5   16.2
200-500                           0.4   0.4   0.1   0.1    0.1
500-1000                          0.2   0.3     0     0      0
>1000                             0.4   0.5     0     0      0
All                              49.4  55.6  70.2  72.2   73.2

(1.) Patricia E. Dilley, Breaking the Glass Slipper-Reflections on the Self-Employment Tax, 54 TAX LAW. 65, 65 (2000) [hereinafter Dilley, Glass Slipper].

(2.) Jerry Heaster, Tax Burden Weighs on Wealthy, CLEV. PLAIN DEALER, Feb. 17, 2002, at G3.

(3.) See Andrew Mitrusi & James Poterba, The Distribution of Payroll and Income Tax Burdens, 1979-1999, 53 NAT'L TAX J. 765, 765 (2000) [hereinafter Mitrusi & Poterba, Distribution].

(4.) STAFF OF THE JOINT COMM. ON TAXATION, DISTRIBUTION OF CERTAIN FEDERAL TAX LIABILITIES BY INCOME CLASS FOR CALENDAR YEAR 2000 2 (2000) [hereinafter JOINT TAX COMMITTEE]. As noted in the text, it turned out to be about 36% for the top 1%, presumably because of the exuberant stock market and unusually high capital gain realizations, most of which were realized by the top 1%. "Almost three-quarters of American capital-gains income comes from just 1.7 percent of taxpayers." Gene R. Nichol, Law's Disengaged Left, 50 J. LEGAL EDUC. 547, 551 (2000).

(5.) JOINT TAX COMMITTEE, supra note 4, at 3.

(6.) See infra notes 31-32 and accompanying text. The story does not end at the federal level, though. The Center on Budget and Policy Priorities released a study that shows that during the 1990s many states significantly decreased the percentage of revenue collected through the progressive income taxes and significantly increased the percentage collected through regressive sales taxes. See Nicholas Johnson & Daniel Tenny, The Rising Regressivity of State Taxes, Center on Budget and Policy Priorities, at (Jan. 15, 2002). As a result, the current percentage share of state tax liability relative to the late 1980s has decreased for the wealthy while it has increased for the poor and middle classes.

(7.) See infra note 58 and accompanying text.

(8.) The issue of whether the Social Security system ought to be abolished and replaced entirely with private savings is beyond the scope of this article. Suffice it to say that if the Social Security system is abolished and replaced with private retirement accounts individual to each worker, this article becomes (obviously) moot. This article attempts to explore the tax burden on labor income if the Social Security and Medicare taxes remain in place.

The short version of my take on privatization, however, can perhaps be gleaned from what I have to say in the rest of this article. The privatizers generally dislike the redistributive nature of the payment system under Social Security. As described later, see infra notes 115-17 and accompanying text, a much higher percentage of wages is replaced for low-wage workers than for high-wage workers. A strict system of privatized accounts, with no matching payments for low-wage workers or other system modifications, would end the redistribution from higher-wage workers to lower-wage workers that is embedded in the current system. It would also shift market risk to the participant. Many middle- and upper-income workers have private retirement savings subject to market risk as a balance to the "guaranteed" nature of the social security benefit, while lower-wage workers have only social security. At this point, it seems to me that both shifts would be unwise policy choices. While there might be creative ways to craft a privatized system that maintains both security and redistribution in some fashion, we also have to be concerned about the administrative costs of shifting to private accounts as well as the price effects of vastly increased capital infusions into the capital markets that would accompany private accounts invested in the private stock and bond markets. Nevertheless, I think we have only begun to earnestly explore alternatives. In the meantime, however, I think no one can dispute the success of Social Security at dramatically reducing poverty in the elderly population. See MICHAEL J. GRAETZ & JERRY L. MASHAW, TRUE SECURITY 115-19 (1999). Moreover, its success as a mechanism that helps to stabilize the economy by stabilizing consumption in retirement also should not be undervalued.

(9.) Some people might consider nonsensical the characterization of the payroll taxes as imposing a "double tax" on labor income (in addition to the income tax), since the taxes are all going to the federal Treasury. For these people, there is only a "single" federal tax on labor income, albeit collected through several different mechanisms that result in combined effective and marginal rates that are higher than commonly understood. Focusing on whether it is a "double tax" masks the essential issue of whether these rates are "too high" under either a fairness or efficiency inquiry. I do not believe that the substance of the critique changes much, if at all, under these two different ways of framing the issue.

(10.) See Tom Herman, Tax Report: A Special Summary and Forecast of Federal and State Tax Developments, WALL ST. J., Jan. 2, 2002, at A1; see also discussion infra Part II.

(11.) Some workers, facing a high marginal tax rate on the last dollar, might choose not to earn that dollar and substitute untaxed leisure (the substitution effect), while others will be encouraged to work even harder to reach certain after-tax income goals (the income effect).

(12.) See infra notes 95-103 and accompanying text.

(13.) See DANIEL SHAVIRO, MAKING SENSE OF SOCIAL SECURITY 31-32 (2000) (discussing research by Kotlikoff and Sachs to this effect).

(14.) See, e.g., Jonathan Clements, "I Plan to Save Like Crazy--Someday," WALL ST. J., June 26, 2001, at C1; Mary Kane, Behavioral Economists Can't Say Why We Spend, CLEV. PLAIN DEALER, May 21, 2001, at 4-C. Both articles discuss how people generally tend to overvalue the present compared to the future, what Harvard's David Laibson termed "hyperbolic discounting." Id. I would expect that discount rates may be even higher among the poor and lower-middle classes, as compared to the upper classes, if we assume the decreasing marginal utility of money as income and wealth rise.

(15.) See discussion infra Part III.

(16.) "Simultaneously with this great tax increase on labor income, we have experienced an almost equally striking decline in taxes on capital, most dramatically in the portion of federal revenues generated by the corporate income tax." Michael J. Graetz, The Troubled Marriage of Retirement Security and Tax Policies, 135 U. PA. L. REV. 851, 862-63 (1987) [hereinafter Graetz, Troubled Marriage].

In contrast to the enormous growth in taxes on wages, the percentage of total revenues generated by the corporate income tax has declined sharply in the past forty years. The corporate income tax produced more than 30 percent of federal revenues in 1953, but now produces only about 12.5 percent.


Writing nearly two decades ago, Richard Kaplan noted that "the increasing significance of social security taxes exacerbates the federal tax system's bias against earnings from labor--wages, salaries, and self-employment profits...." Richard L. Kaplan, The Shifting Burden of Federal Taxes, 19 TAX NOTES 3, 4 (1983). The material in Part III, infra, describes more fully recent research documenting this increase.

(17.) Graetz, Troubled Marriage, supra note 16, at 863-64.

(18.) The estate tax is extremely progressive. See generally Deborah A. Geier, The Death of the "Death Tax"?: An Introduction, 48 CLEV. ST. L. REV. 653, 654 (2001).

(19.) See infra notes 31-32 and accompanying text.

(20.) Cf. Martin J. McMahon, Jr. & Alice G. Abreu, Winner-Take-All Markets: Easing the Case for Progressive Taxation, 4 FLA. TAX REV. 1 (1998) (arguing both that the progressivity of the tax burden ought to be measured and distributed by reference to growth in after-tax incomes and that economic efficiency is not necessarily harmed thereby).

(21.) David Cay Johnston, More Get Rich and Pay Less in Taxes, N.Y. TIMES, Feb. 7, 2002, at A22.

(22.) Nichol, supra note 4, at 550-51.

(23.) Tom Redburn, Honoring, and Paying All Those Who Serve, N.Y. TIMES, Oct. 28, 2001, [section] 3, at 4.


(25.) Id. at 47.

(26.) SHAVIRO, supra note 13, at 147.

(27.) "Social Security is meant to be redistributive...." Lee Cohen et al., Social Security Redistribution by Education, Race, and Income: How Much and Why, at euerle-Carasso_Paper.pdf at 1 (May 17-18, 2001) [hereinafter Lee Cohen et al.]; see also infra notes 113-17 and accompanying text.

(28.) Moreover, earmarked taxes for social welfare and healthcare spending can reduce the risk of fiscal bankruptcy.

With general fund financing, each budgetary item is supported ultimately by the revenues of the whole budget; however, this creates an incentive for overexpansion and waste. With the earmarking arrangement, in contrast, the deep pocket of the whole budget is not available; therefore, the incentive for overexpansion and waste is effectively curtailed. Consequently, the risk of fiscal bankruptcy is greatly reduced. For measures such as national health insurance and social welfare policies that involve substantial fiscal outlays, the potential risk of having runaway growth is acute. In short, the compartmentalization consideration makes the earmarking arrangement not only theoretically straightforward but practically relevant.

Bingyuang Hsiung, A Note on Earmarked Taxes, 29 PUB. FIN. REV. 223,227-28(2001).

(29.) 65 CONG. REC. 6324 (1917).

(30.) See discussion infra Part II.

(31.) Economic Growth and Tax Relief Reconciliation Act of 2001, Pub. L. No. 107-16, 115 Stat. 38 (2001).

(32.) See, e.g., Deficits are Foreseen For at Least 3 Years, N.Y. TIMES, Nov. 29, 2001, at A1.

(33.) See discussion infra Parts II-III.

(34.) One final observation before proceeding is worthwhile. As might be gleaned from the brief description above, this article will analyze the possibilities for integration by reference--in part at least--to income tax theory in the old-fashioned sense of the term. It considers, for example, whether the payment of a tax ought to be deductible or creditable under traditional notions of what the tax base ought to be if what we seek to tax is "income," as conventionally defined or as actually implemented in our hybrid income/consumption tax system. I am fully aware that this kind of analysis has fallen on hard times in legal academia. This methodology has been labeled the "internal coherency approach" and uses such criteria as coherency, consistency, and clarity of the tax law. See Joseph Bankman, The Business Purpose Doctrine and the Sociology of Tax, 54 SMU L. REV. 149, 155 (2001). Under the "efficiency/welfarist" approach, in contrast, "the value of coherency, consistency, and clarity is entirely instrument al." See id. (quoting Mark P. Gergen, The Common Knowledge of Tax Abuse, 54 SMU L. REV. 131, 144 (2000)). The inquiry under the efficiency/welfarist approach is only whether the deduction or credit makes sense on efficiency grounds (in the sense of neoclassical economics) or advances social welfare, "understood as some aggregate of well-being of individuals in society" (in the sense of welfare economics). Louis Kaplow, A Framework for Assessing Estate and Gift Taxation, in RETHINKING ESTATE AND GIFT TAXATION 203 (William G. Gale et al eds., 2001). While there may be overlap, tax policy analysis is nevertheless distinctly different under these two approaches, no matter what the tax policy topic.

Bankman notes that those who champion the internal consistency approach "are skeptical ... about the normative assumption that underlies the welfarist/efficiency approach: that efficiency is all that matters," while "[t]he efficiency/welfarist approach regards internal coherency-based scholarship as without normative force. Internal coherency scholarship can tell the welfarist something about the consequences of a particular provision, but until those consequences are tethered to an efficiency/welfare analysis, internal coherency scholarship cannot say anything about the desirability of that provision." Bankman, sup ra, at 156. Bankman goes on to observe that, in his opinion, the efficiency/welfarist group is somewhat in the lead and that "[t]hose who cannot at least participate in the discourse of efficiency and welfare run the risk of being marginalized. Id. at 157.

I think that Bankman is certainly correct in his appraisal of which kind of scholarship seems to be more valued in todays law-and-economics world. Moreover, this fact is relevant, I believe, to the topic of this article, as there is arguably a relationship between this fact and the shift toward the kinds of more regressive taxation that is exemplified by the shift toward heavier taxation of labor income and lighter taxation of capital income. Fairness arguments have steadily lost ground to the more scientifically based arguments grounded in economics, where economic growth (which is often argued arises from freeing capital from taxation), rather than a fair allocation of the costs of government across the population, is the paramount concern. See generally RONALD F. KING, MONEY, TIME, & POLITICS: INVESTMENT TAX SUBSIDIES & AMERICAN DEMOCRACY (1993) (detailing the shift in acceptable tax policy rhetoric from fairness to economic growth after World War II); M. Susan Murnane, Selling Scientific Taxation (2001) (unpublished paper, on file with author). In commenting on the rise in empirical economic research in tax policy, John F. Witte stated, "[N]umbers tend to drive out non-numbers in the policy-making process, JOHN F. WITTE, supra note 23, at 47-48. This approach, therefore, can result in two dangers:

What is alarming is that the uncertainty of the estimates and the qualifications that are a part of scholarly discourse may be lost as unqualified users zero in on simple aggregate conclusions to build and support political arguments. The other danger of this research is that the emphasis in the empirical studies is on efficiency losses, and the equity side of the equation can be easily lost. This is bound to weight the outcome in favor of less progressivity. ...

Id. at 48 (footnote omitted).

Those firmly in the efficiency/welfarist camp will no doubt see this article as lacking any normative force, since it chiefly elaborates the ways in which one can view the payment of payroll taxes, and the receipt of government benefits, from within the paradigm of a tax on "income" or consumption. In this sense, it is descriptive, seeking the internal consistency that is the hallmark of the unfashionable internal coherency approach, but this description is rooted in what I see as fairness, in the sense of the fair distribution of the aggregate tax burden. One means by which we have sought to honor that virtue in the past is by allocating the tax burden across the members of the population by reference to individuals ability to pay, which I believe is a powerful (if imprecise) norm. The abilityto-pay value has traditionally been thought to justify use of income as the tax base (as opposed to, say, consumption). See generally J. Clifton fleming, Jr. et al., Fairness in International Taxation: The Abili ty-to-Pay Case for Taxing Worldwide Income, 5 FLA. TAX REv. 299 (2001) (situating the ability-to-pay norm within income taxation theory in the context of taxing foreign-source income and citing the relevant authority discussing the ability-to-pay value at the margins). Analyzing how a payroll tax ought to be treated under income tax theory can be seen, therefore, not merely as an empty tautological exercise but rather as a proxy in the pursuit of equity, ultimately grounded in the ability-to-pay fairness norm.

Of course, the welfare economists would respond that social welfare, not fairness, ought to govern this analysis, but I leave that analysis to them. But, from what I can see, the outcome under a "fairness" approach might well mirror the outcome under a "welfarist" approach in most cases.

Moreover, I believe that an analysis that seeks internal consistency within the current tax system can have persuasiveness in the political system. I am, after all, ultimately seeking a response in that system. Indeed, perhaps I ought, for that reason, to abandon the internal consistency approach and consider a public choice analysis of payroll and income tax integration, but I also leave that analysis to others.

(35.) Actually, there was a short-lived income tax during the Civil War, and the income tax enacted in 1894 was ruled unconstitutional in 1895 as an unapportioned "direct tax." The adoption of the Sixteenth Amendment permitted an income tax without regard to apportionment. See generally JOHN F. WITTE, supra note 24, at 67-75.

(36.) See, e.g., Daniel Shaviro, Colloquium on Financial Instruments: Risk-Based Rules and the Taxation of Capital Income, 50 TAX L. REV. 643, 645 (1995) (positing that "a zero tax rate [as opposed to a negative tax rate] on capital income is often the best that can be done...")

(37.) At the individual level, income from capital is more lightly taxed than labor income due to the ability to defer taxation of most capital income by choosing when to realize it, by taxing capital gains at lower rates than ordinary income, and by completely forgiving built-in gain at death because of the step-up in property basis at death under current section 1014. Because the incidence of the separate corporate tax might fall on all individual holders of capital, however, the corporate tax might, in a rough way, equalize the tax burden on labor and capital in the income tax. In other words, integration of the corporate and individual tax might exacerbate the overtaxation of labor income relative to capital income in the income tax, while integrating the payroll tax burden with the income tax might tend toward better equalizing the tax burden across labor and capital. Cf. David Cay Johnston, 2 Courts Reject I.R.S. Efforts to Limit Tax Shelters, N.Y. TIMES, June 22, 2001, at Al (noting that "Treasury Secretary, Paul H. O'Neill, called last month for the elimination of corporate income taxes, an idea that he said intrigued President Bush").

(38.) See generally JOHN F. WITTE, supra note 24, at 77 (noting that Congressman Hull "forcefully defended the [proposed income tax] as based not on consumption, as the tariff and excise taxes were, but on ability to pay").

(39.) See generally Erik M. Jensen, The Taxing Power, the Sixteenth Amendment, and the Meaning of "Incomes", 33 ARIZ, ST. L.J. 1057 (2001) (recounting the debates surrounding the enactment of the Sixteenth Amendment, focusing in particular on the move from consumption taxation to income taxation to shift more of the tax burden to the wealthy).

(40.) Fewer than two percent of workers filed income tax returns between 1913 and 1915. See JOHN F. WITTE, supra note 24, at 78.

(41.) Carolyn C. Jones, Class Tax to Mass Tax: The Role of Propaganda in the Expansion of the Income Tax During World War II, 37 BUFF. L. REV. 685, 687 (1989).

(42.) See Social Security Act of 1935, Pub. L. No. 74-271, 49 Stat. 620 (1935).

(43.) See Social Security Act Amendments of 1939, Pub. L. No. 76-379, 53 Stat. 1360 (1939). Robert M. Ball, who began working in the Social Security administration in 1939 and was its Commissioner from 1962-1973, has written, "Even before monthly annuities were payable to a wage earner under the original Act, the program was completely revamped. In a very real sense, the 1939 Act more than the old-age benefit provisions of 1935 formed the Social Security structure that we know today." Robert M. Ball, The 1939 Amendments to the Social Security Act and What Followed, in THE REPORT OF THE COMMITTEE ON ECONOMIC SECURITY OF 1935, at 161, 162 (50th anniversary ed. 1985) [hereinafter 1935 REPORT]. Among other changes, the 1939 amendments broadened the beneficiary class from the worker to the worker's family, including spouse, children, and even parents. See John J. Corson, The Old-Age and Survivors Insurance Program, in WAR AND POST-WAR SOCIAL SECURITY 58, 59-61 (Wilbur J. Cohen ed., 1942). But the precipitating iss ue that caused Congress to revisit the program so soon was a controversy over the financing mechanism. The recession of 1937 provoked fears that "the original method of financing the old-age benefit system--by accumulating a large reserve account--might serve as a drag on future economic growth." Alan Pifer & Forrest Chisman, Foreword to 1935 REPORT, supra, at xi. Ball described the controversy, which has resonance for today's discussions about the funding of Social Security, as follows:

What seemed most important at the time was a continuing battle between those who favored a substantial build-up of reserves that would earn interest and later reduce the level of needed contributions, and those who favored raising only the amount of money needed to pay current benefits--the so-called "pay-as-you-go" approach.

The controversy over financing came to a head when Senator Vandenberg of Michigan questioned Social Security Board Chairman Arthur J. Altmeyer at a Senate Finance Committee hearing on February 22, 1937. Senator Vandenburg felt that the accumulation of funds, estimated to reach $47 billion, had serious consequences for sound fiscal policy. Investing them in government securities as provided by the law Would require a permanent national debt which he and many others considered undesirable. Yet if the Social Security funds were invested in non-federal securities there would have been new federal involvement in private economic activity. Perhaps the most important criticism was that the reserves would encourage extravagant federal spending, because Social Security would have created a ready source for federal borrowing. From the opposite viewpoint, an earnings reserve was considered a way to avoid an eventual general revenue contribution, and on this ground had the personal support of President Roosevelt.

Ball, 1935 REPORT, supra, at 163-64.

Congress did modify the law so that, "in place of a large reserve account for old-age benefits, the adoption of a 'pay as you go system' with a partial reserve meant each generation would finance its parents' retirement out of current earnings." Pifer & Chisman, supra, at xi.

When looked at together, the 1935 and 1939 Social Security Acts essentially put in place all the basic features of the Social Security retirement and survivors' insurance programs as they exist today: entitlement based on earnings reflecting a long term attachment to the paid work force; pay-as-you-go financing; and a benefit formula weighted in favor of low income workers.

Patricia E. Dilley, Taking Public Rights Private: The Rhetoric and Reality of Social Security Privatization, 41 B.C. L. Rev. 975, 1031 (2000) [hereinafter Dilley, Rhetoric].

(44.) See Herman, supra note 10.

(45.) I.R.C. [section] 3101(a).

(46.) I.R.C. [section] 3111(a).

(47.) See, e.g., JOSEPH PECHMAN, FEDERAL TAX POLICY 232-33 (5th ed. 1987) ("In the long run, most of the payroll taxes are probably paid by the worker; it makes no difference whether the law imposes the tax on the employee or on the employer."); Andrew Mitrusi & James Poterba, The Changing Importance of Income and Payroll Taxes on U.S. Families, 15 TAX POL'Y & ECON. 95, 98 (2001) ("We assume that workers ultimately bear the employer share of the payroll tax in the form of lower wages.") [hereinafter Mitrusi & Poterba, Changing Importance]; SHAVIRO, supra note 13, at 10 ("Economists generally agree that the distinction in who pays which portion of the payroll tax has no effect on economic incidence, and that at least in the short run the entire tax is largely borne by workers ...."); C. Eugene Steuerle & Jon M. Bakija, Basic Features of the Social Security System, 62 TAX NOTES 1457, 1457 (1994) (noting that "[m]ost economists agree that practically all of the burden of the 'employer' portion of the Social Secu rity tax falls on the worker, who effectively pays through reduced wages").

(48.) I.R.C. [section] 1401(a). See generally Dilley, Glass Slipper, supra note 1 (describing both the conceptual and administrative difficulties pertaining to the self-employment tax base, which seeks to replicate the wage tax base within the income tax system).

Also introduced in the Social Security Act of 1935 was the Unemployment Tax, which today is imposed on every employer at the rate of 6.2% of the first $7,000 of wages of each employee and which funds the joint Federal-State unemployment compensation program. I.R.C. [section][section] 3301, 3306(b). There is no explicit employee tax. Employers can credit up to 90% of eligible contributions made to state unemployment funds against the federal tax. I.R.C. [section] 3302. Unemployment payments are includable in gross income. I.R.C. [section] 85. The unemployment tax and payments are not further discussed in this article.

(49.) Tax Adjustment Act of 1966, Pub. L. No. 89-368, 80 Stat. 38 (1966).

(50.) I.R.C. [section][section] 3101(b), 3111(b).

(51.) I.R.C. [section] 1401(b).

(52.) See I.R.C. [section] 164(t).

(53.) See Dilley, Glass Slipper, supra note 1, at 65 n.2 (describing these benefits and providing the various pinpoint citations in the Social Security Act).

(54.) For example, as an employee of a state university with its own retirement system, I am ineligible to participate in Social Security and do not have the 12.4% payroll tax withheld from my paycheck, though Medicare tax is still withheld. See generally Ball, in 1935 REPORT, supra note 43, at 169 (stating that some state and local government employees remain outside the Social Security system).

(55.) SHAVIRO, supra note 13, at 17.

(56.) See Tom Herman, Tax Report: A Special Summary and Forecast of Federal and State Tax Developments, WALL ST. J., July 18, 2001, at A1.

(57.) See Thomas v. Comm'r, T.C.M. (CCH) 1653 (2001). Disability payments are not further discussed in this article.

(58.) "In aggregate, 23.7% of families have personal income tax liabilities in excess of their payroll taxes, 62% have payroll taxes in excess of personal income taxes, and 14.3% have equal liabilities (usually zero)." Mitrusi & Poterba, Distribution, supra note 3, at 771 tbl.2 notes; see also Mitrusi & Poterba, Changing Importance, supra note 47, at 101 ("Payroll taxes exceed income taxes for 62 percent of all families, and for 73 percent of those with either income or payroll tax liability.").

(59.) Mitrusi & Poterba, Distribution, supra note 3, passim.

(60.) "Adjusted AGI" is adjusted gross income plus the following items: employer-paid Social Security and Medicare taxes, the portion of Social Security benefits received that is excluded from gross income under section 86, interest received on state and local bonds that is excluded under section 103, and deductible contributions made to individual retirement accounts and so-called Keogh plans. See id. at 770.

(61.) Id. at 771.

(62.) This table is taken from Mitrusi & Poterba, Changing Importance, supra note 47, at 100.

(63.) Id. at 101.

(64.) See id. at 108-09.

(65.) This tale is taken from id. at 105.

(66.) See infra notes 167-68 and accompanying text.

(67.) Mitrusi & Poterba, Distribution, supra note 3, at 765.

(68.) Id. at778-79.

(69.) It also should be noted that this reduction in progressivity at the federal level has occurred at the same time that reductions in progressivity have occurred at the state level. See supra note 6.

(70.) The Statistics of Income Service at the Internal Revenue Service does have a small statistical sample on computer tape for 1960, but it does not contain any data on Social Security taxes. And, of course, there are no computer files for 1939. E-mail from Peter Sailer, Statistics of Income Service, Internal Revenue Service, to Deborah A. Geier, Professor of Law, Cleveland-Marshall College of Law, Cleveland State University (July 10, 2001) (on file with author).

(71.) See PECHMAN, supra note 47, at 332 tbl.A-11.


(73.) JOHN F. WITTE, supra note 24, at 125.

(74.) Thomas S. Adams, How Federal Taxes Are Made, 15 NAT'L TAX ASS'N PROC. 331, 337 (1923) (quoted in James A. Wooten, The "Original Intent" of the Federal Tax Treatment of Private Pension Plans, 85 TAX NOTES 1305, 1315 (1999)).

(75.) See SELTZER, supra note 72, at 62 tbl.9.

(76.) See id.

(77.) See Bob Herbert, Get Well, George, N.Y. TIMES, July 12, 2001, at A23.

(78.) See SELTZER, supra note 72, at 74 tbl.15 (the sum of .7%, 1.6%, and 6.5% for AGI of $1,000 to $2,000; $2,000 to $3,000; and $3,000 to $5,000; respectively), 73 chart 6. I do not have this precise information for 1960, but I have no reason to believe that it differed significantly from 1965.

(79.) PECHMAN, supra note 47, at 332 tbl.A-11.

(80.) See supra notes 9-14 and accompanying text.

(81.) Wilbur J. Cohen, The Social Security Act of 1935: Reflections Fifty Years Later, in 1935 REPORT, supra note 43, at 5, 7 [hereinafter Wilbur J. Cohen].

(82.) Id.


(84.) Id. Apparently, there was also a fear that national legislation on pensions would serve as a precedent for federal intrusion in the area of race relations at the state law level. Edwin E. Witte kept a contemporaneous diary during the time that he served as Executive Director of the Committee on Economic Security though this contemporaneous diary was not published until his death in 1962. In it, he wrote that the old-age assistance portion of the proposed legislation

was very bitterly attacked, particularly by Senator Byrd, on the score that it vested in a federal department the power to dictate to the states to whom pensions should be paid and how much. In this position, Senator Byrd was supported by nearly all of the southern members of both committees, it being very evident that at least some southern senators feared that this measure might serve as an entering wedge for federal interference with the handling of the Negro question in the South. The southern members did not want to give authority to anyone in Washington to deny aid to any state because it discriminated against Negroes in the administration of old age assistance.

EDWIN E. WITTE, THE DEVELOPMENT OF THE SOCIAL SECURITY ACT 143-44 (1962) [hereinafter EDWIN E. WITTE]. He remarked that "[t]he fact is that it had never occurred to any person connected with the Committee on Economic Security that the Negro question would come up in this connection." Id. at 144. Nevertheless, he describes changes that were made to the bill to satisfy these concerns. See id. at 144-45.

(85.) Wilbur J. Cohen, supra note 81, at 7-8.

"I guess you're right on the economics," Roosevelt conceded when told that the employee contributions were a mistake, "but those taxes were never a problem of economics. They are politics all the way through. We put those payroll contributions there so as to give the contributors a legal, moral, and political right to collect their pensions and their unemployment benefits. With those taxes in there, no damn politician can ever scrap my social security program."

LEUCHTENBURG, supra note 83, at 133.

(86.) Another reason that extending the income tax downward would have been a problematic approach is that the Social Security program had many exemptions in its early history. It was thought unfair to impose taxes on all lower- and middle-class families when only a portion of them would be eligible to obtain future benefits under the system. Excluded from the system were the self-employed as well as those employed in agriculture, domestic service, by the Federal government, and by state and local governments, as well as all charitable, educational, religious, scientific, etc., organizations and their employees. See EDWIN E. WITTE, supra note 84, at 152-57. Writing in 1946, Eliot J. Swan noted that "the fact that only certain groups were included made it difficult to justify the use of general revenues." Eliot J. Swan, Economic Aspects of Social Security (1946), reprinted in READINGS IN SOCIAL SECURITY 559, 560-61 (William Haber & Wilbur J. Cohen eds., 1948). While this formally created a separate tax on the wages of much of the lower and middle classes, it is not too far afield to characterize the 1935 legislation as expanding the income tax downward, albeit under a separate fiat-rate schedule and with no deductions, to the lower and middle classes to fund increased federal spending on a new social program. The vast majority of laborers owed no income tax on their labor income, and the vast majority of laborers had no significant capital income, so taxing only their wages under the payroll tax was tantamount to expanding the income tax to them, albeit under a different rate schedule.

(87.) Jones, supra note 41, at 688 (citing Revenue Act of 1932: Hearings on H.R. 10236 Before the Senate Comm. on Finance, 72d Cong., 1st Sess. 3 (1932)).

(88.) Id. at 689.

(89.) Id. at 691 (footnotes omitted) (citing 79 CONG. REC. 13,207-13,208 (1935)).

(90.) Jones, supra note 41, at 690-91 (footnotes omitted).

(91.) Id. at 691-93 (footnotes omitted).

(92.) See discussion supra Part III.

(93.) See discussion supra Part III.

(94.) Heidi Glenn & Warren Rojas, Column A, Column B: Washington Orders Up $50-$75 Billion Stimulus, 93 TAX, NOTES 167, 169 (2001).

(95.) Mitrusi & Poterba, Changing Importance, supra note 47, at 114-15.

(96.) Richard V. Burkhauser & John A. Turner, Is the Social Security Payroll Tax a Tax?, 13 PUB FIN. Q. 253, 254 (1985); see also C. Eugene Steuerle & Jon M. Bakija, How Social Security Redistributes Income, 62 TAX NOTES 1763, 1766 n.13 (1994) (collecting other similar studies).

(97.) Burkhauser & Turner, supra note 96, at 263. Steuerle and Bakija agree. They conclude that "almost all individuals who have retired in any year between 1940 and today--no matter what their income level or family type--have received large positive transfers from Social Security beyond the sum of their contributions to the system and a reasonable rate of return on those contributions." Steuerle & Bakija, supra note 96, at 1767. They also conclude that "[p]ositive net transfers are eliminated for high-/average-wage two-earner couples retiring after the turn of the century [and that] high-wage single workers and two-earner couples retiring in the 2020s and later will face very large negative transfers (or positive net lifetime taxes) from the system." Id. at 1770. The one exception they see is the one-earner couple.

High-wage one-earner couples retiring in the near future receive very large transfers, often exceeding $100,000 . ... Indeed, high-income one-earner couples retiring before the turn of the century continue to receive larger net transfers than anyone else. Under our projections, positive subsidies continue to flow to high-wage one-earner couples retiring as late as 2050. As we move into the next century, moreover, average-wage one-earner couples will continue to receive larger transfers than low-wage couples.


(98.) GEOFFREY KOLLMANN, CONGRESSIONAL RESEARCH SERVICE, SOCIAL SECURITY: THE RELATIONSHIP BETWEEN TAXES AND BENEFITS FOR PAST, PRESENT, AND FUTURE RETIREES (June 22, 2001), reprinted in CRS Examines Social Security Benefits for Past, Present, and Future Retirees, TAX NOTES TODAY (July 6, 2001) (LEXIS, FEDTAX lib., TNT file, elec. cit., 2001 TNT 129-15) [hereinafter CRS REPORT].

(99.) Id.

(100.) Steuerle & Bakija, supra note 96, at 1770.

(101.) SHAVIRO, supra note 13, at 33 (citing Steven Caldwell et al., Social Security's Treatment of Postwar Americans, 13 TAX POL'Y & ECON. 109 (1999)).

(102.) See Mistrusi & Poterba, Changing Importance, supra note 47, at 113 tb1.8.

(103.) Id. at 117.

(104.) CRS REPORT, supra note 98.

(105.) Id.; see also SHAVIRO, supra note 13, at 34 (stating that "[w]hat rate to use. . . has proven highly controversial" and noting that rates used in recent studies ranged from two percent to seven percent).

(106.) See generally Dilley, Rhetoric, supra note 43; Patricia E. Dilley, The Evolution of Entitlement: Retirement Income and the Problems of Integrating Private Pensions and Social Security, 30 LOY. L.A. L. REV. 1063 (1997) [hereinafter Dilley, Entitlement]. In criticizing the integration of the employer portion of Social Security taxes with the nondiscrimination rules for qualified plan contributions, Nancy Altman has also rejected the analogy.

Employers are paying a percentage of their worker's salaries into Social Security. However, the payment is a tax, not a pension contribution. The benefits for which the employer is contributing are not for its current workers, but for workers now retired, from an earlier generation, perhaps before this employer was even in business. To construe the employer's contribution as something other than a general tax and the employee's benefit as somehow purchased in part by the employee's employer is carrying a useful political fiction to an illogical extreme.

Nancy J. Altman, Rethinking Retirement Income Policies: Nondiscrimination, Integration, and the Quest for Worker Security, 42 TAX L. REV. 433, 485 (1987) (footnote omitted).

(107.) "[T]he misleading private pension analogy used to sell the Social Security program from the New Deal forward [is] directly attributable to the distorting effect of the insistence on individual property rights as the basis for economic rights." Dilley, Rhetoric, supra note 43, at 984.

(108.) Id. at 1000 (footnotes omitted).

(109.) Id. at 1003.

(110.) Id. at 1035.

(111.) A technical level at which the link between taxes and benefits is severed is in the legislation itself, which formally treats them separately, though this separation was done purely for concerns about the legislation's constitutionality and did not reflect a substantive decision to break the link between taxes and benefits. For that reason, the technical separation is best viewed as just that.

In the original version of the bill drafted by the Committee on Economic Security, the new tax and the benefit structure provisions were closely intertwined, clearly indicating a link between this new federal tax and this new benefit system. But

[i]t was widely feared that so great an expansion of federal activities would be ruled unconstitutional by the then conservative Supreme Court. . . . For these and other reasons, . . . the old-age provisions were divided into separate taxing and spending clauses--to make the program seem less like an insurance system operated by the federal government and more like an exercise of the government's expressed constitutional powers to tax and spend in new ways.

Pifer & Chisman, supra note 43, at xi.

Edwin E. Witte, in his contemporaneous diary of events, noted that the Supreme Court ruled the Railroad Retirement Act unconstitutional "a few days after the Senate Finance Committee began executive sessions on the social security bill" and that "[l]anguage used in this decision seemed to apply also to the old age insurance provisions of the social security bill." EDWIN E. WITTE, supra note 84, at 100. He recounts what happened as follows:

The entire plan was very objectionable to [House draftsman] Mr. Beaman and seemed unconstitutional to the leading members of the Ways and Means Committee. These members agreed that if the benefits and the tax were really separated, the plan might be constitutional, but that there was no real separation between these two in the original bill. Beaman was then instructed to redraft the bill to make such a separation. The major change was to drop all reference to an old age insurance fund. Instead of such a fund, an old age reserve account was set up in the Treasury; but the proceeds of the tax were not allocated to the old age reserve account or the payment of old age benefits. All provisions which related the benefits to the tax paid by the employers and their employees were eliminated, as well as all references to any contractual rights to benefits. What the Ways and Means Committee and Beaman, acting under its instructions, sought to do was to get away from an insurance plan altogether, but to establish the e quivalent of such a plan without resorting to a definite insurance scheme.

Today's beneficiaries are not living off financial assets accumulated in the past. Today's workers are not accumulating financial assets for the future. Workers "invest" their payroll taxes not in financial assets but in the willingness of future politicians to tax future workers to pay future benefits.

Id. at 146-47. In short, the technical separation between the taxing provisions and the spending provisions was done solely out of concerns for the constitutionality of the provisions. The legislation was held to be constitutional in Stewart Machine Co. v. Davis, 301 U.S. 548 (1937) and Helvering v. Davis, 301 U.S. 619 (1937).

(112.) Lee Cohen et al., supra note 27, at 4.

(113.) As the President's Commission to Strengthen Social Security recognized:

PRESIDENT'S COMMISSION TO STRENGTHEN SOCIAL SECURITY, INTERIM REPORT (July 24, 2001), reprinted in Bush's Social Security Commission Issues Text of Interim Report, TAX NOTES TODAY (July 25, 2001) (LEXIS, FEDTAX lib., TNT file, elec. cit., 2001 TNT 143-43) [hereinafter INTERIM REPORT].

(114.) Lee Cohen et al., supra note 27, at 3.

(115.) Nancy J. Altman, The Reconciliation of Retirement Security and Tax Policies: A Response to Professor Graetz, 136 U. PA. L. REV. 1419, 1431 (1988).

(116.) Lee Cohen et al., supra note 27, at 3.

(117.) See Altman, supra note 106, at 476-78. It must also be noted that these redistributive effects are reduced if one takes into account the reduced life expectancies that are typically associated with low income. "[L]ower income and shorter life expectancies are related. . . . As a result, [the] likelihood of [low-income people] receiving Social Security retirement benefits is probably lower than for high-income people." Steuerle & Bakija, supra note 96, at 1772; see also Study Shows Poor People Die Sooner Than Affluent, CLEV. PLAIN DEALER, Nov. 12, 2001, at E4.

This redistributive function would be difficult to replicate with private accounts, which, depending on one's point of view, is either a strength or a weakness of privatization proposals. But see SHAVIRO, supra note 13, at 146-57 (proposing that private accounts be coupled with redistributive lump-sum payments at retirement from the accounts of high earners to the accounts of low earners).

(118.) See Dilley, Rhetoric, supra note 43, at 1032 (noting that Social Security and related programs were seen as labor force management programs).

(119.) Wilbur J. Cohen, supra note 81, at 10.

(120.) Presidential Statement Signing the Social Security Act (Aug. 14, 1935), reprinted in 1935 REPORT, supra note 43, at 145.

(121.) 1937-1938 ADVISORY COUNCIL ON SOCIAL SECURITY FINAL REPORT, reprinted in 1935 REPORT, supra note 43, at 198. Later, Wilbur Cohen recalled that "[t]he Vandenburg-Murray amendment which was in the law during 1943-1950 authorized a government study, but this provision was never utilized." WILBUR J. COHEN, RETIREMENT POLICIES UNDER SOCIAL SECURITY 5 n.15 (1957).


(123.) Swan, supra note 86, at 561.

(124.) See discussion infra Part VI, which reviews this argument. I ultimately conclude that it is weak.

(125.) See discussion infra Part VH.

(126.) See, e.g., Smith Simpson, Fundamental Principles, in WAR AND POST-WAR SOCIAL SECURITY 1 (Wilbur J. Cohen ed., 1942); Osvald Stein, Fundamental Programs, in WAR AND POST-WAR SOCIAL SECURITY, supra, at 6, 6-9.

(127.) INTERIM REPORT, supra note 113.


(129.) See generally id. at 467-83 (contrasting income and consumption taxation and describing how the nondeductibility of capital expenditures under an income tax is the defining difference between it and a consumption tax).

(130.) See supra note 55 and accompanying text.

(131.) See Daniel Halperin, Description of Current Tax Treatment of Compensation iv (2001) (unpublished manuscript, on file with author) [hereinafter Halperin, Description].

(132.) The initial legislation made clear that employees could not deduct their mandatory Social Security tax payments under the income tax. Section 803 of the original act provided: "For the purposes of the income tax imposed by Title I of the Revenue Act of 1934 or by any Act of Congress in substitution therefor, the tax imposed by section 801 shall not be allowed as a deduction to the taxpayer in computing his net income for the year in which such tax is deducted from his wages." 1935 REPORT, supra note 43, at 93.

(133.) Alternatively, the nondeductibility of the employee portion of the Social Security tax could be analogized to the Roth IRA, under which contributions are nondeductible but distributions are excludable. If this approach were taken, fifty percent of Social Security benefits, representing the employer contribution that was effectively excluded by the employee, would be includable, while the remaining fifty percent, representing the undeducted employee contribution, would be excludable. For the reasons stated in the text regarding the variability of marginal rates over a lifetime, however, I think that the deduction-inclusion approach would be preferable.

(134.) See discussion supra Part V.

(135.) Halperin, Description, supra note 131, at ii. Halperin compares a $100 contribution to a qualified plan with a 10% (tax-free) rate of return on behalf of a 40% bracket taxpayer with a taxable contribution of $60 (paying $40 to the Treasury at the time of contribution) with the same rate of return and demonstrates that the tax on the distribution from the qualified plan is precisely equal to the future value of the $40 tax paid at the time of contribution to a taxable plan. See id. at ii n.7. This equivalency holds true, however, only if the taxpayer is in the same tax bracket at both points in time.

(136.) "Most early pension plans were noncontributory. That is, the plans were financed completely by the employer who generally interpreted the obligation to pay benefits as voluntary." Dilley, Entitlement, supra note 106, at 1114.

(137.) "Only five pages of Murray Latimer's huge study [published in 1932], Industrial Pension Systems, discuss federal tax laws." Wooten, supra note 74, at 1307.

(138.) Robert M. Ball, The 1939 Amendments to the Social Security Act and What Followed, in 1935 REPORT, supra note 43, at 161, 161.

(139.) Dilley, Entitlement, supra note 106, at 1113-14.

(140.) Dilley, Rhetoric, supra note 43, at 1028-29.

(141.) Wooten, supra note 74, at 1307-08.

(142.) See id. at.1308-10.

(143.) Id. at 1313 (quoting O.D. 110, 1 C.B. 224 (1919)).

(144.) In 1921, Congress provided that contributions to "stock bonus and profit-sharing" plans were not includable by the employee "until...distributed or made available to the extent that it exceeds the amounts paid in by him." See Wooten, supra note 74, at 1313 (quoting Revenue Act of 1921, ch. 136, [section] 219(f), 42 Stat. 227, 247 (1921)). The Service informally extended the same treatment to pension plan contributions, and this treatment was codified in 1926. See id. at 1314.

(145.) See id. at 1317.

(146.) See supra notes 55-56 and accompanying text.

(147.) Erwin N. Griswold, The Tax Treatment of Employees' Contributions to Pension Plans, 57 HARV. L. REV. 247, 247-48 (footnotes omitted) (emphasis added).

(148.) Id. at 248-50 (footnotes omitted) (emphasis added).

(149.) Dean Griswold recognized that perhaps neither should be rendered nontaxable to the employee.

A literal argument can be made in support of [immediate taxation to the employee] even where the employee's contribution is deducted from his pay and he has no right to receive it in cash, since the employee has acquired something of value in his retirement contract. The question remains, however, whether it is wise to tax the employee currently on what is actually so remote, though important, a benefit.

Id. at 248. He then went on to note that employer contributions were not taxable to the employee when made and that there is no logical reason to treat employee contributions differently.

(150.) See Daniel Halperin, Cash or Deferred Profit-Sharing Plans and Cafeteria Plans, 41 N.Y.U. INST. ON FED. TAX'N 39-1, 39-6 to 39-11(1983) [hereinafter Halperin, Cash or Deferred].

(151.) IRA eligibility was severely restricted, and section 40 1(k) received little attention when first enacted. It was only after an accountant, Theodore Benna, saw the potential in section 401(k) that it became popular as a pre-tax investment vehicle for employees. The New Yorker magazine told the story this way:

In 1974, the first individual retirement accounts . . . were introduced, but the standards for qualification were strict, and they didn't really catch on. In the Tax Reform Act of 1978, legislators loosened things up a bit by allowing workers to contribute their cash bonuses to retirement savings accounts on a tax-deferred basis. The wording of this clause, No. 401(k), was vague, and it attracted the attention of R. Theordore Benna, an employee-benefits consultant in Langhorne, Pennsylvania.

One Saturday afternoon in 1980, Benna, who was then thirty-nine years old, was helping one of his clients, a local bank, to redesign its employee pension plan when he had a thought. If cash bonuses could be sheltered from tax under clause 401(k), why couldn't regular income be sheltered in the same way? There didn't appear to be anything in the statute that specifically ruled it out. "My approach was that if the code doesn't say, 'Thou shalt not,' then thou should be able to".... He designed a retirement plan that would allow employees to contribute a portion of their paychecks to a savings account on a pretax basis. A few months later, Benna's own firm, the Johnson Companies, launched the first 401(k) plan. In November, 1981, the Internal Revenue Service gave Benna's creation its official blessing. With legal approval, the new savings plans spread rapidly, and by 1985 more than ten million employees had one.

John Cassidy, Striking it Rich, THE NEW YORKER, Jan. 14, 2002, at 63, 63-64.

(152.) See Wooten, supra note 74, at 1311.

(153.) See Norman P. Stein, Some Lessons From History: The Origins of Pension and Profit-Sharing Taxation, 1914-1942, 58 N.Y.U. INST. ON FED. TAX'N--EMPLOYEE BENEFITS AND EXECUTIVE COMPENSATION 12-1, 12-7 to 12-8 (2000).

(154.) Id. at 12-27.

(155.) Halperin, Cash or Deferred, supra note 150, at 39-21.

(156.) "[T]he favorable treatment of qualified plans is said to be an incentive for savings. If so, it would seem to be unnecessary when savings are mandatory." Halperin, Description, supra note 131, at ix.

(157.) One contrary response to this might be that there do exist some mandatory qualified pension plan payments that are deemed to come from the employee, rather than the employer, that the employee may exclude from his or her income. My own situation provides a personal example. As an employee of the state of Ohio, a fixed percentage of my negotiated salary is mandatorily withheld from my paychecks and contributed to the State Teacher's Retirement Association, a defined benefit retirement plan. The state of Ohio makes an "employer" contribution as well. Neither the employer contribution nor the portion withheld from my negotiated salary, i.e., my "employee" portion, is includable in my gross income for federal income tax purposes. Yet, I have no say in whether or not I wish to make such a "contribution." One response is that the portion deemed to come from the "employee" under this scheme is arbitrary. Under reasoning similar to that made by Dean Griswold, see supra notes 147-49 and accompanying text, the d istinction is purely formal. My formal "salary" could just as easily be designated as the amount net of my "mandatory contribution," with the entire contribution paid by my employer. For example, assume that my negotiated salary is $100,000 per year, with $10,000 mandatorily withheld and paid to the state retirement fund and an additional $10,000 paid to that fund by the state. It is just as reasonable to characterize this arrangement as one in which my salary is $90,000 with a $20,000 pension contribution paid by the state. Moreover, "[S]tate and local governments are allowed, through a mere declaration to that effect, to treat mandatory employee contributions to their retirement plans as if they were employer contributions." Halperin, Description, supra note 131, at ix.

(158.) See I.R.C. [section] 213.

(159.) See discussion supra Part V.

(160.) See generally DODGE ET AL., FEDERAL INCOME TAX, supra note 128, at 19-37 (generally describing tax policy tools, including the ability-to-pay fairness norm).

(161.) See I.R.C. [section] 213(a).

(162.) See I.R.C. [section] 165(c)(3), (h)(2)(A).

(163.) JOSEPH M. DODGE, THE LOGIC OF TAXES 123 (1989). Most state and local taxes are deductible under I.R.C. [section] 164(a). The deduction for sales taxes was repealed in 1986. While clearly the major impetus for repeal was to raise revenue in the base-broadening effort that accompanied significant rate reductions (and to reduce the administrative problems associated with substantiation of sales taxes paid), perhaps sales taxes can be seen as less "nondiscretionary" than state income or property taxes, since the amount incurred depends on one's decisions regarding how much of one's income to consume, as opposed to save.

(164.) Governmental expenditures that constitute general welfare payments are excludable from gross income. See Rev. Rul. 63-136, 1963-2 C.B. 19.

(165.) See discussion supra Part VI.A.

(166.) See DODGE ET AL., FEDERAL INCOME TAX, supra note 128, at 323.

Although [section] 164 allows a deduction for certain state and local taxes. . . , the U.S. government does not allow any tax credit for state income taxes paid. Otherwise, the states would be able to destroy the federal government by fully appropriating its main revenue source. For example, if a taxpayer's federal tax bill is $100, a state could enact a tax schedule that would also create a tax bill of $100. If the $100 state tax were creditable against the $100 federal tax, not a dollar would go into the Federal Treasury.


(167.) See I.R.C. [section] 32.

(168.) Jeffrey B. Liebman, The Impact of the Earned Income Tax Credit on Incentives and Income Distribution, 12 TAX POL'Y & ECON. 83, 85 (1998); see also S. REP. No. 94-36, at 11 (1975) (mentioning the same purpose for the earned income tax credit); H.R. REP. No. 103-111, at 609 (1993) (noting that the expansion of the credit to workers without children was intended, in part, to reduce the impact of payroll taxes on those with a lower ability to pay). See generally Dennis I. Ventry, The Collision of Tax and Welfare Politics: The Political History of the Earned Income Tax Credit, 1969-99, 53 NAT'L TAX J. 983 (generally describing the origin and purpose of the credit).

(169.) See H.R. 493 Would Allow Credit For Payroll Taxes, TAX NOTES TODAY (Mar. 5, 2001) (LEXIS, FEDTAX lib., TNT file, elec. cit., 2001 TNT 43-101). Feldstein and Samwick, who support a completely privatized account system, would provide a credit against the income tax for the portion of wages mandatorily contributed to these accounts "[b]oth to make the proposal more politically palatable and to discourage other, possibly dubious enactments by Congress." SHAVIRO, supra note 13, at 131.

(170.) GRAETZ, DECLINE, supra note 16, at 215.

(171.) See Jonathan Barry Forman, Promoting Fairness in the Social Security Retirement Program: Partial Integration and a Credit for Dual-Earner Couples, 45 TAX LAW. 915, 955-64 (1992). Implicit in the notion that a credit ought to be allowed only to the dual-earner couple is the notion that it is legitimate to link the taxes paid to future benefits received, because the reason he argues for the credit in this instance is the failure to earn any "return" on the lower-earning spouse's tax payment. Since I see any attempted linkage between the tax and any personal future benefit as illegitimate, I see no reason to limit the credit to dual-earner couples. Indeed, so limiting the credit further feeds the notion that these taxes "earn" a personal return in retirement for the taxpayer.

(172.) See supra Part VII.A.

(173.) E.g., I.R.C. [section][section] 105, 119, 132.

(174.) DODGE ET AL., FEDERAL INCOME TAX, supra note 128, at 33.

(175.) Moreover, inclusion of consumption received in kind as wages or dividends (absent an explicit exclusion provision) is absolutely necessary to protect the survival of the tax. If wages and dividends were tax-free simply if they were cast in the form of in-kind consumption, the tax base would evaporate. It would also be economically inefficient to encourage casting wages and dividends in this form. The class of free in-kind consumption outside these contexts is likely not large, "which makes exclusion of these benefits less troublesome than it would be in the compensation or dividend contexts." Id. at 103.

As a matter of textual exegesis, the argument that most in-kind consumption received outside of the wage and dividend contexts might not be "gross income would be as follows. Section 61 specifically lists compensation for services rendered and dividends as items that are includable, and there is no limitation that such items must be received in cash or in the form of tangible property to be includable. With respect to consumption received in kind that cannot be said to be wages or a dividend, the only language under which it could be taxed would be the catch-all language, "all income from whatever source derived," in the introductory sentence of section 61. Perhaps there is more room in this ambiguous language to conclude that in-kind consumption that cannot fairly be said to reflect the equivalent of the receipt of cash followed by a free-spending choice is not "income" representing resources under the control of the taxpayer and thus available for contribution to the fisc. Cf. United States v. Gotcher, 401 F.2d 118 (5th Cir. 1968) (though using different reasoning, concluding that a free trip that was provided neither as compensation for services rendered nor as a dividend, and perhaps could not be fairly characterized as the equivalent of the receipt of cash followed by a free spending choice, was excludable by Mr. Gotcher, although includable by Mrs. Gotcher).

(176.) See DODGE ET AL., FEDERAL INCOME TAX, supra note 128, at 103 (discussing noncompensatory and nondividend in-kind consumption).

Deborah A. Geier *

* Professor of Law, Cleveland-Marshall College of Law, Cleveland State University.

[c] Deborah A. Geier 2002. I would like to thank the following people for their comments on earlier drafts, though of course they should be held harmless regarding the final result: Anne Alstott, Joseph Dodge, Cliff Fleming, Dan Halperin (times two), Louis Kaplow, Leandra Lederman, Julie Roin, and Eric Zolt.
COPYRIGHT 2002 Virginia Tax Review
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2002 Gale, Cengage Learning. All rights reserved.

 Reader Opinion




Article Details
Printer friendly Cite/link Email Feedback
Author:Geier, Deborah A.
Publication:Virginia Tax Review
Date:Jun 22, 2002
Next Article:S corporation loss limitations: the Tax Court provides potential hope for related party debt restructurings.

Related Articles
The worst tax: how payroll taxes have hurt America's working class.
The Economic Effects of Taxation.
Who pays federal taxes?

Terms of use | Copyright © 2014 Farlex, Inc. | Feedback | For webmasters