What's So Fair About a Tax on Income? (Part 2 of 2)

Dan R. Mastromarco of the Argus Group, Washington, argues that an honest examination of the income tax would reveal that it is less fair than a consumption tax.

Tax Analysts Document Number: Doc 1999-32575 (25 original pages)

Citations: (October 8, 1999)

VI. The Wrong Camera: The Denominator of the
Tax Incidence Equation.

A. The FairTax Looks Surprisingly Good in a Still Photograph.

[68] Even if we were to go along with the very biased assumption that no one in our cordoned off rooms will ever change their financial position, a sales tax can look pretty good in a still photograph if the right lens is chosen. Including payroll taxes, the current federal tax system imposes as relatively flat (but high) marginal tax rate on wage income. Combined tax rates range from 30.3 percent to 46.3 percent for all but the very lowest income Americans.

[69] The highest marginal tax rate imposed on single persons is for those with taxable incomes between $62,450-$72,600 per year (nonitemizers with earnings for instance between $69,500 [$62,450 plus $4,300 for the standard deduction and $2,750 for the personal exemption] and $79,650. /38/ The highest marginal rates experienced by married couples are for those with taxable income(s) between $104,000 and $145,200. The most affluent taxpayers face a marginal tax rate on their wage income somewhat less than most middle-class taxpayers because investment income is a much larger fraction of affluent taxpayers' income. The chart above shows the marginal tax bracket faced by a single taxpayer under current law without taking into consideration the earned income tax credit (EITC).

[70] On an effective rate basis annually calculated, a sales tax need not be regressive any more than an income tax need be regressive. In fact, the FairTax is one proposal that completely untaxes the poor. Under the FairTax, the marginal tax rate would be a flat 23 percent (tax-inclusive) above the poverty level, but zero on expenditures below the poverty level (as the implicit taxes today are eliminated) because the FairTax plan would provide a rebate equal to the sales tax rates times the poverty threshold (holding taxpayers harmless on purchases up to the poverty level). Those that consume at less than the poverty threshold would actually have a negative tax liability. In fact, married couples have an added advantage. To eliminate the marriage penalty (since poverty statistics reflect that a married couple lives more cheaply than two single persons), the FairTax actually increases the rebate so that a married couple is treated as having a poverty level equal to twice the level of single taxpayers. Thus it totally eliminates marriage penalties -- something that may never occur under current law.

[71] To illustrate the plan's progressive nature we can examine the tax burden that a family of four will have at various annual income levels (or in this case, annual spending levels). If an admittedly frugal family of four spends $11,060 per year (half of the 1999 U.S. Department of Health and Human Services (HHS) poverty level /39/ grossed up to eliminate the marriage penalty), they will pay no federal sales tax at all and will have an effective tax rate of negative 23 percent; through the rebate system they will receive a cash grant of $2,543. A family of four that spends $22,120 per year (the 1999 HHS poverty level grossed up to eliminate the marriage penalty) will have an effective tax rate of 0 percent; they will pay no federal sales tax at all. A family of four that spends $33,400 per year (twice the 1999 HHS poverty level) will have an effective tax inclusive rate of 7.8 percent (taxes over expenditures); they will pay $2,594 in tax. A family of four that spends $66,800 per year (four times the 1999 HHS poverty level) will have an effective tax rate of 15.4 percent; they will pay $10,267 in tax. A family of four that spends $133,600 per year (eight times the 1999 HHS poverty level) will have an effective tax rate of 19.2 percent; they will pay $25,640 in tax. And a family of four that spends $1,670,000 per year (a hundred times the 1999 HHS poverty level) will have an effective tax rate of 22.7 percent; they will pay $379,012 in tax. This is based on 1999 HHS poverty level of $16,700 for a family of four and a 23 percent FairTax rate (tax-inclusive) where the rebate is grossed up to eliminate the marriage penalty.

[72] The EITC changes this picture, sometimes for the better and sometimes for the worse. The EITC is a surprisingly complex provision meant to provide an incentive to low income individuals to work. It is provided to qualified individuals that have wages below a designated level and do not have a disqualifying amount of other income. The credit percentage varies depending on the number of children and is phased out at specified thresholds.

[73] Under the EITC, a qualified individual is someone age 25 to 65 who lives in the U.S. and is not a dependent of another person. Earned income is basically wages and salary. Disqualifying income of $2,300 or more per year results in the taxpayer receiving no credit. Disqualifying income is generally dividends, interest, rents, royalties, or capital gains. /40/

[74] Even with the EITC, a single person with no children enjoys lower effective tax rates and almost always lower marginal tax rates under the FairTax as compared to the current tax system because that person is disqualified from the EITC. The poorest single parents with two children would enjoy lower effective tax rates under the FairTax. The top half of the poor single parents have negative effective tax rates under the FairTax but not as negative as under current law. Those somewhat over the poverty line have low effective tax rates under the FairTax but somewhat higher than current law. Those above about $18,000 have lower effective tax rates than under current law. All of this assumes away embedded costs under current law.

B. The Proper Way to View Distribution Is With a Movie Camera

[75] Looking at the distribution as a moving picture -- taxes paid over the course of a lifetime -- yields a much more accurate view than the still photograph. While we cannot predict each individual's outcome and tax effect, we do know with certainty that distributional analyses predicated on stagnation will yield a distorted image. Each individual in our cordoned-off room is a datum moving in three respects throughout time. The individual's financial condition changes as a function of normal lifetime cycles, their financial condition may change due to luck or hard work, and the individual's behavior may change to the extent that individual can influence tax liability by altering behavior. /41/

1. Distribution models built on annual income are misrepresentative.

[76] Our distributional analysis is also typically based on a knowingly incorrect assumption biased in favor of the income tax: that each individual in the frame of our photograph will remain there in perpetuity. Each individual cordoned off within the income groups will continue to be cordoned off. Their incomes will not change from year to year or over the course of their lives, i.e., each individual will make the same income at 17, as they will at 46 or 76. We have grouped the rich teenager with a part-time job and a trust fund with the family seeking to make ends meet. We are also, in many instances, comparing a 45-year-old father with his 18-year-old daughter to make the case that the father should redistribute more of his income to his daughter (a sort of government mandated minimum allowance). Moreover, in comparison to the consumption tax, no one's savings rate will be altered, and no one's income will either increase or decrease.

[77] Are income changes of the persons within the still photograph so insignificant that we can disregard movement within the income quintiles over time when we estimate distribution? Are they negligible? Mere exceptions to the rule? The answer is clearly "no," based on what we know about mobility and income fluctuations.

[78] Some economists have sought to conduct a longitudinal study of the members of each income quintile. For instance, the Department of Treasury in a study on income mobility used historical panel data. /42/ The panel data came from individual income tax returns filed over a 10-year period (1979-1988) to analyze the extent to which there was significant income mobility for households, or equivalently, whether current annual income was a good proxy for average income over a longer period of time. Their principal conclusions were quite startling:

o The panel data suggest significant household income mobility
over time, especially from lower to higher percentiles of the
income distribution.

o About 86 percent of the tax filers in the bottom quintile had
exited over nine years, moving to a higher-income quintile in
1988.

o In the lowest quintile, a greater percentage of taxpayers
moved to the fifth (highest) quintile over the 10-year period
than remained in the first quintile. In other words, the
poorest quintile was more likely after 10 years to be in the
richest quintile than the poorest.

o The corresponding mobility rates were 71 percent for the
second lowest quintile, 67 percent for the middle quintile,
62.5 percent for the fourth quintile and 35.3 percent for the
top quintile.

[79]Just below is a bar graph that is reproduced from the data in the study. The graph shows what happened to those who were in the first (lowest) income group over the 10-year period. The chart shows the key result outlined above, that most of the people we cordoned off in the poorest room in 1979 were not there by 1988. Assuming that any individual would still be there in 10 years is a bet against the odds.

[80] What was the conclusion of the panel study? It was that current annual income is imperfectly correlated with 10-year average income, suggesting that income observed in a given year is not a very good measure of whether the taxpayer is on average "high income," "middle income," or "low income." /43/

[81] There are several reasons for income fluctuations, and we can view them separately. One is the natural progression of individuals into higher and lower income deciles based on the natural ebbs and flows of income over a lifetime (the life cycle). Older people tend to consume less than the young. They make less than in their peak years. Second, income mobility comes from movements up and down the income scale as a result of different wealth accumulation rates among individuals of the same age and starting points along the income scale. Then there is simply income volatility, such as that experienced by the self-employed who must assume a relatively high risk. The same may be experienced by workers subject to layoffs.

[82] When we ignore mobility or more generally income fluctuations it is well established that we bias our estimates in favor of an income tax and steeply progressive rates and against a consumption tax. /44/ One study, using Canadian data, showed that personal income taxes are less progressive in the lifetime context, whereas sales taxes are less regressive. /45/ Other studies have found that sales and excise taxes may be regressive for annual income, but closer to proportional for consumption.

[83] We can see this most plainly if we consider that the vast majority of Americans spend over the course of their life roughly what they earn over the course of their life. If current trends continue, they may even try to spend more than they earn. According to recent U.S. Department of Commerce information, personal savings and personal savings as a function of income has been negative since February, /46/ and it is now roughly what it was in 1932 and 1933 when the savings rate was negative 1.4 and 2.1 respectfully. /47/

[84] A relatively pure consumption base is 85 percent of GDP. The rest is investment and savings. If we assume that the taxes under a sales tax are revenue neutral with the taxes under an income tax (again ignoring economic growth or compliance costs savings), then the amount of tax paid at the end of one's life should be approximately the same whether or not an income tax or consumption tax was in place. The difference is that, for some reason (policy or otherwise), the income tax system imposes greater tax burdens on those whose income fluctuates more, meaning those who minimize risks or who have greater wealth and can control risk and income fluctuation are advantaged.

[85] The crux of the problem stems from the fact that consumption is far more stable year-by-year than income. As far back as 1957, in his "permanent income theory," Milton Friedman postulated that consumption decisions depend on horizons longer than one year. In determining how much to consume, individuals average their income over long horizons. They borrow or save to achieve a smooth consumption path, although the income can deviate substantially. /48/ Thus, to the extent that individuals are life-cycle consumers, the annual measure of consumption to income is a very poor proxy for the lifetime measure.

[86] In his paper, "The National Sales Tax, Who Bears the Burden?" /49/ Gilbert E. Metcalf, using data from the Bureau of Labor Statistics' Consumer Expenditure Survey (CES), measured the lifetime incidence of a shift from the current income tax to a national sales tax. He found that if annual income is used to rank households (without taking into consideration the items filtered out from that still photograph), tax reform might look very regressive. However, he found that the Schaefer-Tauzin bill (H.R. 2001), the sales tax to which Dr. Pepperell refers, is about as progressive as the current income tax. His principal finding was twofold. First, how we rank people -- by annual or lifetime income -- makes a big difference when we measure the progressivity of a national sales tax. Second, a national sales tax replacement for the income tax is not inherently regressive; it is relatively easy to construct a sales tax that protects the poor from paying any tax and is roughly as progressive as the current income tax. If this were what Dr. Metcalf concluded about the Schaefer-Tauzin national sales tax proposal, he would no doubt reach a better conclusion about a sales tax proposal that repeals the regressive payroll taxes, like the FairTax.

[87] That there is significant income mobility in the United States is a fact known to most economists that model tax distribution. Movement among the lower and upper quintiles is in fact the rule, not the exception. One would think that the significant income mobility would make it somewhat difficult to make comparisons among observations when we do not really know where the taxpayer will wind up in the future. One would think that it might give us pause to conclude that one system regressive unless we had accounted for life cycle fluctuations in income.

[88] Income mobility is one of the major reasons distributional analysis based on annual snapshots are always misleading. All distributional effects are muted in the lifetime context. A still photograph should not be superimposed to portray as a caste system what are in reality moving targets. However, mobility and income fluctuation is all but ignored in the income "fairness" discussions as if it were negligible afterthought. Basing distribution on stagnation within income deciles yields not only a knowingly wrong result, it also yields a result knowingly biased against a consumption tax in every respect.

[89] Economists that base distributional tables on annual estimates argue that data covering the entire lifetime profiles for labor income do not exist, let alone data on the composition of income by lifetime income category. Some panel data do exist, and there are insufficient pleas to obtain the data needed for accurate lifetime analyses.

[90] One final note deserves mention. Although economists believe that lifetime income more accurately reflects the economic status of individuals, it is really not possible to create a tax system that taxes individuals according to lifetime income. It is for this reason that some economists advocate adoption of consumption- based taxes. Since consumption more accurately reflects lifetime income and economic well being, such a system would be more equitable than the current system and subject to fewer vicissitudes. /50/

2. Not all tax plans contribute equally to financial mobility.

[91] It is critical to take into consideration existing financial mobility for truthful modeling of the distribution of current law and alternatives, but the degree to which financial mobility is affected by tax reform is equally important. Increasing the ability of persons to improve their lot in life should be a goal of our tax system.

[92] How does our current system affect the ability of one to raise his standard of living? While we have pointed out the existence of a fair degree of mobility, the truth is that we could hardly have devised a tax system that does a better job of locking an individual within an income group while making it appear that we are helping those who want to excel. If we intentionally set out to make it difficult for a taxpayer to advance to another income level, we might do exactly what we have done.

[93] The analysis goes something like this. To improve one's material position over their "pecuniary peer," to move from one quintile to another, one must earn more money or save a greater percentage in one closed period of time. To improve one's standard of living, to "catch up" to a wealthier pool, it is axiomatic that one must earn greater income in a shorter period of time. However, our income tax system is designed to tax those who seek to increase their financial position faster than others by the application of steeply progressive rates. Our income tax system provides increased resistance when someone seeks to better his or her lot in life through higher wages or earnings. The income tax does not really tax the ability to pay; it generally taxes changes in wealth that occur from other than unrealized appreciation in assets.

[94] To make this point clearer, we can take a simple example with manufactured data. Let us take the case of two imaginary individuals, Mr. Bjorn Silverspoon and Mr. Justin Struggles, who are true to their names. Bjorn is born with a silver spoon in his mouth, so to speak. He spends all that he earns in consumption. Struggles is a successful small-business person, whose income fluctuates and who struggles to save at a considerably higher rate in an effort to approach Silverspoon's wealth.

[95] In this example, Silverspoon would pay $472,374 in income taxes, while Struggles would pay $1,113,309, or 2.4 times as much. Is Struggles in better financial condition than Silverspoon? Should he be penalized because he was seeking to advance his material wealth for himself and his children? For every foot Mr. Silverspoon climbed, he had to take 2.4 steps to acquire the same wealth in 20 years. A sales tax would not impose this disproportionate burden on him. In fact, Silverspoon would pay more in taxes because he is consuming more.

[96] For many, wage income is the only vehicle of transport to a more prosperous life -- apart from perhaps a lucky childbirth or a financially rewarding marriage, which are often themselves by- products of wealth. Yet our income and payroll tax system targets those who try to improve their lot in life by disproportionately penalizing those that have to work the hardest to climb from one financial tier to another. The income tax literally taxes our vehicle of transport from one standard of living to another.

[97] To see the fallacy of the progressive rate structure as a "fair" tax system, look at the EITC. The EITC is supposed to entice families to stay in the working world at low income levels, but if decisions are made at the margin, what is most disturbing is what happens to marginal rates under the income tax with the EITC. Marginal rates for those qualified under the EITC are perhaps the highest marginal rates applicable to any income group. Under current law, marginal tax rates for single parents with incomes above $12,260 are 36 percent and for those over $14,350, 51.4 percent (29 and 44 percent, respectively, if only the employee share of the payroll tax is considered). Since most economists believe that the employer share is borne by employees as lower wages, it is appropriate to consider all payroll taxes. These marginal rates fall to 30.3 percent at $30,095.

[98] Such steeply progressive marginal tax rates punish lower- middle-class workers. Once state taxes are considered, many lower- middle-income single parents keep only 40 cents of each dollar they earn. Once the costs of commuting, child care, and other work-related expenses are considered, choosing to work makes very little economic sense for single-parent families. Under the FairTax, marginal tax rates are lower for all workers in this group earning over $9,000 annually (in many cases the marginal tax rates under the FairTax are about of current law). The chart above depicts the comparison./51/

[99] For a married couple with four children, the effective tax rate is somewhat lower for all families in this group. The chart on the following page depicts the marginal rate for a family with four children.

[100] The marginal tax rate is lower for all families with wages above $9,000. If we assumed that the couple had $2,300 or more in dividends, interest, or capital gains income, then the couple would receive no EITC. Thus, a couple that has saved $46,000 over the years and receives 5 percent interest would be disqualified. A couple that had $10,000 in stock mutual funds and obtained a 23 percent capital gain would also be disqualified. In other words, these couples would not have to be wealthy to be disqualified. Parents who saved for their children's education or for retirement or for other reasons would be disqualified. In this couple's case, the effective tax rate is lower at all wage levels. The marginal tax rate is almost always lower.

[101] We could hardly devise a system that permits greater mobility than a sales tax. Once we have exempted the necessities of life from tax, a sales tax, by design, permits maximum maneuverability. Taxpayers have maximum choice relative to current law in altering their personal liability. Individuals with different income levels will have different spending and saving habits and the maximum capability to affect their future tax liability. By not taxing wages, a sales tax removes the regressive burden of the tax from those who must expend their own sweat equity to improve their lot in life. Moreover, by not taxing income higher on a progressive scale, a properly constructed sales tax would not penalize movement from one level of wealth to another. Equally as important, it enables taxpayers to save tax free, as if there were in effect a universal, unlimited front-loaded and back-loaded IRA. No more restrictions on what to save or on who can save. It removes the tax on the vehicle of upward mobility. /52/ It doesn't penalize those who must invest in their future in favor of those who want to consume for today.

3. A still photograph fails to capture economic growth.

[102]A proper moving picture will also show one other factor: that all the groups will be better off in the long run. The current distortion in favor of consumption over investment does not help our economy or even improve consumption over the long haul: it hurts our economy and depresses consumption. All known economic projections (even by the most skeptical economists) predict a healthier economy with a system that does not punish savings and investment. Typical estimates are that the economy will be 10 to 14 percent larger within 10 years and consumption will grow very substantially. A study prepared by Nathan Associates (March 1996) for the National Retail Institute predicted the economy would grow 5 percent more in 10 years than it would have under the income tax. /53/ This study made every adverse assumption it could. They assumed no compliance cost savings, very low labor responsiveness to lower tax rates on labor, no increased foreign investment (as a result of the U.S. being the only industrialized nation with a zero rate on income), and a very low effect of higher investment on productivity increases from higher levels of investment.

[103] If a national retail sales tax were to increase economic growth from its present low rate to the rate that the U.S. has enjoyed on average for the past 35 years (i.e., by about ths of 1 percent), then the typical American family will see dramatic improvements in their standard of living over time. That seemingly small increase in the rate of growth will increase the disposable income of the median family by $21,000 over the next 10 years, with their annual income being $4,100 higher in the 10th year and more than $10,000 higher in the 20th year.

[104] The attractiveness of investment and the demand for capital will improve as well. After repeal of the income tax, the U.S. will be perhaps the most attractive place to invest on earth. The U.S. will not only have stable political and legal institutions, a relatively well-educated workforce, a good infrastructure, and a large domestic market, but it will be the only jurisdiction in the world with a zero rate of tax on saving and investment. Foreign firms can be expected to build manufacturing plants here, employing U.S. workers. U.S. firms can be expected to stay in the U.S. instead of being forced to move overseas to remain competitive.

[105] If the moving picture shows a benefit from a shift to a consumption tax, what would the same moving picture show about our income groups if we keep the income tax system? The battle to replace the current tax system with a consumption tax is partly a battle to improve the standard of living of the average American family. The poor are disproportionately hurt by economic downcycles. They are the first to be laid off, the last to be rehired, and the least capable of weathering the economic storms. The effect of a consumption tax on alleviating this hardship is not measured in tax distributional tables nor is the unnecessary deleterious effect of today's system.

VII. How Can We Accurately Observe the Image of Distribution?

A. Try a Consumption Lens as an Alternative.

[106] Even if we were to accurately measure whether or not the "concerned taxpayer" pays more or less tax over his lifetime under a sales tax like the FairTax than under the income tax as a function of income, what does this really tell us about equity and fairness? Recall that we began our distributional analyses with a conclusion: that income tax paid was the correct numerator and income groupings were the correct denominator of the ability to pay. However, is income really the best measurement of the ability to pay? If we could base the tax on an illusive concept of the ability to pay, what would we choose as the denominator? Alternatively, if we assume that "fairness" in tax distribution would be a simple expression of:

what would we want "x" to be? Would the best denominator really be income?

[107] There are many possibilities from which to choose. How about how hard we work? How much the market values our work? Income earned during a given unit of time (i.e., our income system today)? How much we earn from our labor (i.e., the payroll tax system today)? Wealth (i.e., the death taxes)? The benefits one gets from the government (i.e., user taxes)? Our level of consumption of generic government services (multinational companies enjoy our global defense commitments)?

[108] If we say it should be income, which is the dominant lens proponents of the income tax choose, we are assuming that the burden of the tax as a function of income equates with fairness. We are saying that the more someone earns in a given unit of time, the more they should pay. However, income within a closed period of time is only one indicator, a surrogate, of the ability to pay. Moreover, it is often a poor surrogate.

[109] What one taxpayer earns in a given year may have nothing to do with what he or she will earn the next year or 10 years, as we have already noted. High income in a one-year period does not mean one has the ability to pay the same proportion of income in the future. We recognize this inequity when we made exceptions for income averaging for farmers in the Taxpayer Relief Act of 1997. /54/ Why not extend the same logic to athletes? Trial lawyers? Miners? Flash- in-the-pan restaurateurs? Stunt men? For people who get sick with a critical illness in their peak? For tax lawyers who invent a new gimmick and gain fleeting notoriety? Also, one-time events like the sale of a business or a farm makes the middle-class taxpayer look "rich" for one year.

[110] Even low income doesn't mean someone lacks the ability to pay any more than high income gives them that ability. The very wealthy don't need taxable income relative to consumption. They can consume from existing wealth. They can also throttle down earnings and throttle up future appreciation, so as to avoid the taxable definition of income. An income tax does not tax based on the ability to pay. Income is merely the means by which individuals try to increase their personal wealth so that they have a better measure of economic security, so they can invest, so they can provide better for their family, provide to charity, save for their education, or pay for things they wish to consume.

[111] Wealth -- which itself may or may not be a fairer determination of the ability to pay -- is not even captured in the income tax. Individuals rich in personal wealth, may have very little taxable income. That is because wealth is defined in assets that they hold -- their homes, properties, securities, collectibles, and other items -- which may or may not have been earned by them and which may or may not generate taxable income. These wealthy individuals can often choose whether or not to create taxable income, since they can restructure their affairs to avoid receiving current taxable income. Far more than the poor or the middle class, the wealthy have the ability to control income flows (as we legally define it) versus appreciation versus consumption. The ability to pay, therefore, is not perfectly defined by how much income someone happens to make in any period, like a year. In individual cases, it is not even roughly defined.

[112] What is wrong with basing a tax on consumption for one's own personal benefit over the course of one's lifetime? Supporters of the consumption tax believe consumption is a better determinant than the ability to pay. Proponents of a consumption tax would posit that there is no fairer tax base than what one individual consumes for his or her own personal well being over the course of their lifetime. Savings and investment builds factories and creates jobs. It is only when this wealth is consumed for the personal benefit of the wealth holder that anything "selfish" is really going on.

[113] Thomas Hobbes stated that it is fairer to tax people on what they extract from the economy, as roughly measured by their consumption, than to tax them on what they produce for the economy, as roughly measured by their income. Irving Fisher postulated in the 1930s that income really isn't income to an individual until it is consumed. If income is not consumed it is only deferred for later consumption or perhaps given to charity or lost to poor investment. If it is saved or successfully invested it is either consumed later by ourselves or the objects of our bounty. If it is saved and lost to bad investment, it at least has been available to the economy. If we tax income and savings, we have simply taxed deferred consumption. And those that are deferring consumption are doing so because they elect not to consume it for themselves immediately but to make the resource available for a future time. /55/

[114] If we were to analyze the distribution of the taxes we pay where "x" is consumption, we would be surprised at how progressive a sales tax system is. Those that choose to consume for their gratification, as opposed to investing, would pay more tax. Those that consume the most for themselves would pay the most tax. Everything and everyone in their place. Profligate consumers will pay more. Savers and investors will pay less. And the world would look right to us.

[115] In the end, if we examine the income tax as if it were a consumption tax, we will find the distribution of the income and sales taxes comparable. The only net difference between the two approaches -- apart from administration -- is that consumption taxes will create higher levels of economic growth, greater efficiency, greater personal choice over taxability, and greater chances for those who strive to improve their economic lot upward. For most taxpayers, unfortunately, consumption is a proxy for income over the course of their lives.

B. Remove the Filters From the Income Tax Lens

[116] If we insist on looking at the distribution of the tax systems through an income tax lens, then it is imperative that we make certain adjustments before reaching any conclusions about the distribution, let alone the "fairness" of our system. To accurately measure distribution of the tax, we could begin by exposing the taxes embedded in the price of goods and services each of us consume. We should not stop there. We should ensure that the income tax distribution tables also treat as taxes paid, the higher cost of consumer interest, as capital owners transfer the tax to capital users. Alternatively, we might estimate the burden of the tax as if it had backwards economic incidence. However, this would require us to make critical assumptions concerning the relationship to taxes on capital and the after-tax return to labor. We should ensure that the hidden transfer of tax burden from evaders to wage earners is accounted for and that the significant compliance costs, which may total as much as one-fifth of all income taxes collected, are accounted for as hidden taxes in the price of goods and services.

C. Use a Motion Picture Camera.

[117] Beyond just ensuring that the image we develop accurately shows all the taxes the poor and middle class pay, we should clearly not make comparisons based on the assumption that taxpayers are frozen in income brackets. Rather, if we want to classify taxpayers based on some measurement of income, then we should follow that taxpayer's progress over the course of a lifetime by using longitudinal data. Panel data does exist and can be enriched. Moreover, we should not forget about the impact of the tax alternatives on the financial mobility of the taxpayer. Finally, if a proposal will improve the standard of living of most Americans, should we not also judge the tax burden by what we have remaining after taxes, rather than what we pay in taxes?

D. Use a Consistent Size Screen to Portray It.

[118] When considering the rate of a national sales tax, or any tax for that matter, one must always decide which of two distinct means of portraying this rate -- the "tax-inclusive rate" or "tax- exclusive rate" -- best expresses the tax burden. Which one we employ changes absolutely nothing in terms of the taxes that are actually raised or paid by the taxpayer under the taxing regime examined, in the same way that measuring a journey in inches or meters does not change the distance. However, how the rate is presented changes how the relative tax burden is perceived by those who wish to compare the merits of competing tax proposals. Confusion results when we compare alternatives under different measuring scales.

[119] The sales tax is particularly susceptible to this confusion because state sales taxes are normally expressed on a tax- exclusive basis, while income, estate, and payroll taxes, as well as the Flat Tax and other VATs, are normally expressed on a tax- inclusive basis. If we were to express a sales tax rate as a percent of the product price as is done in the states, we would be unfairly overstating the burden of the tax when we compare it to what it is meant to replace at the national level. Or conversely, we would be greatly understating the relative burden of the federal income and payroll taxes for those who don't have time to learn the different measuring systems.

[120] Presentation of a rate of tax on a tax-exclusive basis simply means that the rate of the tax is expressed as the tax paid over a base determined after the tax was already imposed (for example, taxable income under our personal income tax system that is net of the tax). In other words, a tax-exclusive rate would be defined as:

[121] The rate therefore reflects the ratio of taxes paid to what is left in the base, such as net of tax income.

[122] On the other hand, defining the rate of tax on a tax- inclusive basis simply means that the rate of the tax is expressed as the tax paid over the base before the tax has been imposed. In other words, a tax-inclusive rate would be defined as:

[123] Since the base of the tax before the tax is imposed is always more than the base after tax (the denominator is greater), expressing the tax in a tax-exclusive way will always yield a higher rate. In other words, it will express the tax as having a higher burden. /56/

[124] Let us take the following example.

Example: An individual earns $1,000 and pays $200 in taxes
(under either a VAT, income tax, or sales tax) but spends the
remaining $800 on a stereo. Although the taxpayer will pay the
same amount of taxes ($200) out of the same amount of pretaxed
income ($1,000) a question arises as to how the rate should best
be expressed? Is the tax rate 20 percent or 25 percent?

[125] Clearly, one might say that the income or Flat Tax rate is the lower rate, 20 percent, since the taxpayer paid $200 on $1,000 of pretaxed income. That is because the income tax and VATs are normally looked at (unquestionably looked on) on a tax-inclusive basis. However, when we view traditional state sales taxes we might say that the state sales tax rate needed to raise $200 of revenues is 25 percent, even though the sales tax rate raises the same amount of revenue as a 20 percent tax-inclusive income or Flat Tax rate. The taxpayer would be considered to have paid the tax at a 25 percent rate since the taxpayer paid $200 of tax on $800 worth of goods exclusive of tax. That is because the state sales taxes are normally looked on on an after-tax or tax-exclusive basis. To use our formula for tax-exclusive representation:

[126] Which is the correct way of expressing this rate? To the casual observer, it is obvious which tax to prefer. All else being equal, one would prefer a 20 percent rate over a 25 percent rate. But that same person may be surprised to find out that they are saying the same thing, and paying the same tax.

[127] The problem with using a tax-exclusive basis for determining the rate of a national sales tax and a tax-inclusive base to portray the income tax is that it can be very misleading. Let us look at a taxpayer who is at the top marginal rate under each taxing scheme. The tax-inclusive and tax-exclusive rates would be compared as shown in the charts just above and just below.

[128] In the tax-inclusive chart, we see comparisons that we are used to seeing. This chart reflects the maximum marginal rate of the current personal income tax system as 43.3 percent. /57/ Here the sales tax rate is 23 percent and the Flat Tax rate is 32.3 percent, reflecting the combined payroll and Flat Tax burdens. /58/ But the tax-exclusive chart indicates that the income tax with the payroll tax bears a maximum marginal rate that is 75.8 percent of the tax- exclusive base. Even the federal individual income tax alone reflects a maximum marginal tax-exclusive base of 43.3 percent. According to the chart above, the Flat Tax bears a maximum marginal rate of 47.7. The FairTax plan bears a maximum marginal rate of 29.9 percent. In this chart, the taxes paid are calculated as a percentage of what remains after tax.

[129] In making comparisons between alternative taxing systems it is important to ensure therefore that these comparisons are consistent, fair in terms of expectations, and are well explained. Fair comparisons eliminate and do not exacerbate confusion over a relatively critical point as the means of expressing the tax rate. The only means to do so is to ensure that a tax-inclusive rate is compared with a tax-inclusive rate.

VIII. Conclusion.

[130] It strikes me that the whole idea of analyzing distribution of an income tax on the basis of income in an annual period, let alone using that method for comparing an income tax to a consumption tax is disingenuous at best and misrepresentative at worst. In the very least, it fans the fires of ignorance. What we do not know about the current distribution of taxes and yet profess to know is shameful.

o we do not know the standards by which distributional equity is
judged, yet we assume a concensus on the standards;

o we do not know what is the right definition of income, yet we
gloss over the details;

o we do not know how much implicit tax consumers pay, so we
ignore it or arbitrarily assign it to capital and labor;

o we do not know how much tax is buried in the user cost of
capital, so we ignore it;

o we do not know how to distribute compliance costs, so we
consider them negligible;

o we know that we have improperly assumed that individual income
is static, when we know mobility is the general rule (and we
consider distribution on an annual basis);

o we do not know the exact effects of economic growth, so we
don't bother distributing the benefits of plans that clearly
increase overall well being.

[131] In short, we do not know the distribution of taxes on individuals, as hard as we try not to admit it. If we don't know the distribution of these taxes, then we are a long way from saying that they are fairly distributed today -- especially when we do haven't defined the standard on which "fairness" is based.

[132] Let me end by suggesting something that those of us who engage in the tax reform debate -- economists, tax lawyers, and advocates -- should always remember. Many of us are in a genuine struggle to improve the system and to make it fairer. The debate is more than a chance to watch an amusing clash between those that believe they have their hearts in the right place and those that are heartlessly helping to advantage the already advantaged. We are all absolutely right to try to evaluate all tax reform proposals -- even the current system -- by the distribution of the tax. However, we must have a goal in mind and sound criteria against which our progress toward that goal can be measured.

[133] What are these goals? One such goal might be upward mobility so that tax policy affords the greatest opportunity for rewards within the control of the taxpayer. Another might be that the tax system allows for the greatest after-tax income of the most number of taxpayers. Another might be that the tax system is based on the ability to pay, at least insofar as it does not tax those who don't have the ability to pay.

[134] The most frustrating part of the arguments about fairness and distribution is that, while the issue is easily politicized because it has the maximum ratio of buzzwords to fuzzwords, it takes patience to understand. It will take a willingness on behalf of the politicians and advocates to define the criteria against which fairness can be judged. It will require the Joint Tax Committee staff and others to develop more accurate measurements of distribution and to expose the current measurements as faulty. Today, every assumption is stacked in favor of the income tax. What is worse, we know that if we base distributional estimates on these assumptions we have a greater chance of being wrong than right. The debate over a consumption tax is a serious intellectual debate: it should begin with ensuring that those charged with defining "fairness" set forth the basis for that standard and that those charged with the responsibility of evaluating distribution do so with science and not bias.

- - - - - - - - - - -

Footnotes (for Part 2 of 2)

/38/ This is based on the 1998 taxable year marginal rates of income and payroll combined.

/39/ The 1999 poverty level is $16,700.

/40/ The credit percentage varies depending on the number of qualifying children. The amount of the credit is the percentage amount times the amount of earned income (up to the earned income amount). Thus, for a person with no children the maximum credit amount is 7.65 percent of $4,460 or $341. For a person with two children, the maximum amount is 40 percent of $9,930, or $3,972.

The credit is reduced (i.e., phased-out) for wages above the phase-out amount by an amount equal to the phase-out percentage times the amount by which wages exceed the phase-out amount. Thus, a person with two children earning between $12,260 and $30,095 has the credit reduced by $2.11 for each $10 they earn over the $12,260 phase-out amount. This is the equivalent of an additional 21 percent tax on earnings in this range.

/41/ We choose to ignore that the law to which the individual is subjected changes, although that is clearly our experience.

/42/ U.S. Department of the Treasury, Office of Tax Analysis, Household Income Mobility During the 1980s: A Statistical Assessment Based on Tax Return Data (June 1, 1992).

/43/ Annual current income tends to underestimate or overestimate the household's average income in a manner dependent on the age of the taxpayer. The Treasury Department found that those taxpayers advancing into higher-income quintiles tended to be younger; whereas those who fall into lower-income quintile over time tend to be older. These findings therefore were not netted for average life-cycle changes, but included both changes from intuitive life-cycle movements as well as from advancements beyond the norm.

/44/ Kotlikoff, "Saving and Consumption Taxation: The Federal Retail Sales Tax Example," in Frontiers of Tax Reform, p. 160 (M. Boskin ed. 1996).

/45/ Davies, "Some Calculations of Lifetime Tax Incidence," 74 American Economic Review 633 (September 1984). Poterba, "Lifetime Incidence and the Distributional Burden of Excise Taxes," 79 American Economic Review 325 (May 1989).

/46/ Savings have been a negative $51.5 billion in February, through 60.7, 74.5, 94.5, 68.4, and 86.2 billion in July. As a function of income, the savings rates have been -0.8, -1.0, -1.2, -1.5, -1.1, and -1.4.

/47/ U.S. Department of Commerce, Bureau of Economic Analysis. See also http://www.asec.org/perssav.htm.

/48/ M. Friedman, A Theory of the Consumption Function (1957).

/49/ Cato Policy Analysis No. 289 (December 8, 1997).

/50/ D. Bradford, Untangling the Income Tax, at ch. 8 (1986).

/51/ This is based on the 1998 taxable year marginal rates of income and payroll combined.

/52/ As far as income fluctuations are concerned, the FairTax effectively reinstates income averaging with its single rate structure. It imposes a tax that is a standard percentage of consumption over the course of one's life; those whose income fluctuates (like those that are improving their standard of living) get disproportionately hurt in an annual income system with progressive rates.

/53/ The National Retail Federation, which wants to oppose a national sales tax, did not like the findings of this study and has funded another study.

/54/ In claiming victory, one member's press release stated that "just as the actual processes of agriculture demands a long-term view, so too does understanding the income and wealth of those engaged in the business. The really good years cancel out the really bad . . . finally resulting in an average."

/55/ When we tax savings, we take away the ability to make those resources available to others. We raise the interest rates for others. When taxpayers save and invest, they contribute to the public welfare. They generate benefits to the community, for their future and their children's future, well beyond personal gratification.

/56/ When calculating the tax-inclusive sales tax base, two algebraically equivalent methods may be used. The tax-exclusive rate may be converted into a tax-inclusive rate by dividing the tax- exclusive rate by one plus the tax-exclusive rate: ti = te / (1+ te). Conversely, a tax-inclusive rate may be converted into a tax- exclusive rate by dividing the tax-inclusive rate by one minus the tax-inclusive rate: te = ti / (1-ti). Alternatively, the tax- inclusive sales tax rate may be calculated by adding the repealed income tax revenue back into the tax base (consumers, after all, would have that money to spend), whereas one would not do so when calculating the tax-exclusive base (consumers would be spending that amount on tax and it would not be appropriation to include it in the calculation of a tax-exclusive base).

/57/ The maximum marginal payroll rate is 15.3 percent, but this rate applies regressively between $0 and $72,600 for 1999. When this rate attaches, it is possible for a tax to apply at a maximum marginal rate of 43.3 percent (28 percent individual income tax rate plus 15.3 percent payroll tax rate).

/58/ While it is beyond the scope of this article, it is important to understand that the Flat Tax rate of 17 percent assumes a substantial reduction in government revenues.

End of Footnotes (for part 2 of 2)