Abstract: This paper is divided into four parts. First, it examines the basic economics of the family in modern America. Second, using this simple “model” of family economics, it seeks to explain the basic facts about the distribution of income among families in the United States. Third, it argues that earlier tax reform plans—for example, Steve Forbes 1996 flat tax plan—failed politically because they ignored these economic realities of American families, and suggests that the same fate awaits any plan that repeats the mistake. Finally, this discussion will outline the “LBMC Plan,” which has two parts: The first would replace the federal income tax and the corporate income tax with a flat tax on all spending in the economy (except for a tax credit based on family size which would rebate both the flat tax and Social Security payroll taxes up to the poverty level). This part of the plan amounts to treating all income the way President Bush is proposing to treat dividends: non-deductible when paid by businesses, and non-taxable when received by persons. The second part of the plan would balance the pay-as-you-go Social Security retirement system indefinitely (both getting rid of the current trust fund surplus and preventing future trust fund deficits) by cutting the payroll tax now by 20 percent, in exchange for a reduction in future benefits of up to 20 percent (prorated according to the number of years each worker receives the tax cut).
While serving as one of the five outside economic advisers to the National Commission on Economic Growth and Tax Reform (the “Kemp Commission”1) in 1995-1996, I proposed a fundamental reform of both the tax code and the Social Security system, which was conceived as an approach to taxation that would help the family. I have since simplified that plan for a “family-friendly flat tax.” I now find that my job in explaining it has also been greatly simplified by the extensive recent discussion, of President Bush’s proposal to end the double taxation of dividends. But before considering my proposal, it may be helpful to understand why so many earlier efforts at fundamental reform of the tax code (particularly variations of a “consumption tax” or a national retail sales tax) have failed politically. Therefore, this paper is divided into four parts. First, it examines the basic economics of the family in modern America. Second, using this simple “model” of family economics, it seeks to explain the basic facts about the distribution of income among families in the United States. Third, it argues that earlier tax reform plans—for example, Steve Forbes 1996 flat tax plan—failed politically because they ignored these economic realities of American families, and suggests that the same fate awaits any plan that repeats the mistake. Finally, this discussion will outline the “LBMC Plan2,” which has two parts: The first would replace the federal income tax and the corporate income tax with a flat tax on all spending in the economy (except for a tax credit based on family size which would rebate both the flat tax and Social Security payroll taxes up to the poverty level). This part of the plan amounts to treating all income the way President Bush is proposing to treat dividends: non-deductible when paid by businesses, and non-taxable when received by persons. The second part of the plan would balance the pay-as-you-go Social Security retirement system indefinitely (both getting rid of the current trust fund surplus and preventing future trust fund deficits) by cutting the payroll tax now by 20 percent, in exchange for a reduction in future benefits of up to 20 percent (prorated according to the number of years each worker receives the tax cut). The two parts of the plan were designed to go together; though either could be enacted by itself, it’s logical to expect that, if either happens, both will happen at the same time.
The Economics of the Family Household
The economic theory of the household goes back at least to Aristotle’s Politics. In fact, the Greek word oikonomia literally means “household management.” Many things have changed greatly in the past 2,300 years, but a few things have remained the same. According to Aristotle, the purpose of any human society larger than the family, such as a political society, is to secure a “good life.” But the basic function of the family household, he reminds us, is to “secure life itself.” This means that the household is essentially built around the relationship between a man and a woman, which normally produces children. It that ever stopped being the case, all households and persons would cease to exist within a single human lifetime. Though it is not necessary or even desirable that everyone many and have children, civilization is sustained because most people in fact do.
Is the Family Disappearing?
On the face of it, the statistics would seem to indicate otherwise. In 2000, there were 104.7 million American households.3 Of these, 72.0 million were family households, that is, related persons living together, while most of the 32.7 million non-family households comprised 26.7 million adults living alone. Among family households, 34.6 million were households headed by adults living with their own children under age 18; of these, about 25.2 million were headed by married couples, and the remainder by householders whose spouse was absent. It would seem, therefore, that less than one-quarter of American households match the description of the “traditional family”: a husband and wife living together with their children. Some have hastily concluded from such figures that the traditional family is a rapidly dwindling minority, presumably to be swamped by an ever-increasing majority of unmarried individuals and married couples without children.
But on closer inspection, the apparent decline of the traditional family turns out to be largely attributable to the fact that. Americans are living longer than they used to. Two important facts should be kept in mind in interpreting the figures: About 96 percent, of all Americans who have reached age 65 have been married4; and about 90 percent of women ever married have had at least one child. Since the early 19th century, the share of ever-married women having children has fluctuated between about 79 and 92 percent5; but in the past decade among women aged 18 to 34, the same fraction either has had or expects to have children as did women aged 65 or older—90 percent.6 While the impact of divorce as a social pathology ought not be minimized (the stability of those marriages has declined, partly because of increased divorce7), a snapshot of American households in a single year does not reflect the experience of the persons in the households over their life-times. After all necessary qualifications, it remains true that more than nine often individuals now living alone either have been (or will be) married; and nearly the same proportion of couples now living without children either have had (or will have) children.
Longer Life Changes the Human Life Cycle
The key fact driving these changes is that, in 1900, average life expectancy at. birth in the United States was 48 years for women and 46 for men8; now it’s about 80 for women and 75 for men.9 The result has been to change the nature of the human life-cycle. To oversimplify only a little, for most of human history, most people experienced only two stages of life: dependent childhood and parenthood. These phases were of approximately equal length, and children made the transition to adulthood at a much earlier age. Few experienced either an “empty’ nest” or retirement, simply because most people didn’t live long enough for either. But because of increased longevity, most people can now expect to experience four distinct stages instead of two: dependent childhood, parenthood, the “empty nest,” and retirement. For someone who reaches retirement age, these phases are roughly of equal length: 20 or 22 years for women, somewhat less for men.
Let’s consider how increased longevity has affected each of the four stages. Longer life has made education and training of all kinds more valuable, because the return on this investment (in the form of higher wages and salaries) can be enjoyed for many more years. Hence the period of schooling and dependent childhood has become longer. Because dependent childhood is longer, people are marrying later and the period of active parenthood is longer. In part because more children survive into adulthood, parents are having fewer children than they did when children were far less likely to survive. Most of the growth in the share of households that consist of married couples without children reflects the fact that most parents now live long enough to see their children grow up and start families of their own: the “empty nest.” Finally, most of the increase in the number of householders living alone consists of widows (and widowers). A century ago, though life expectancy at birth was slightly longer for women, there were more elderly men than women—mostly because medical advances reducing the hazards of childbirth were still quite recent. Thanks to such medical advances, most wives are now outliving their husbands by several years.
The Two Kinds of Wealth: People and Property
Despite the many changes, two other economic facts about the family noted by Aristotle also haven’t changed. Nearly all wealth is owned by families living in households—much directly, in the form of houses, land, autos, and so forth, but much indirectly, by owning the shares or bonds issued by business firms. And the wealth owned by households consists of two kinds: what economist Theodore W. Schultz around 1960 termed “human capital”10 (the economically valuable qualities embodied in human persons), and “nonhuman capital” (the useful qualities embodied in property, such as buildings, land, machines, and patents). Both kinds of wealth derive their economic value either from their direct usefulness in meeting human wants or needs (e.g., food’s ability to nourish us, a car’s ability to transport us, a house’s ability to shelter us, the medical skills by which a doctor can heal sickness, a singing voice that other people like to listen to); or else indirectly, by providing the means to produce such directly valuable or “final” goods and services.
As recently as a century ago, both kinds of wealth were still primarily produced and directly owned by the persons in households. Most people lived on farms, and most business firms were simply the farmer’s (or merchant’s or craftsman’s) household—just as, in a still earlier age, the government was essentially the king’s household. But especially over the past century, the two kinds of production have been largely separated. The modern business firm is. historically speaking, an offshoot of the household, which specializes in producing nonhuman wealth, while the modern household itself specializes in “producing” and “maintaining” human persons. Most of the production of nonhuman goods now occurs in business firms, which combine the services of workers and productive property to produce new goods and services. Most productive property is now directly owned by business firms, but families own all business property indirectly, by investing in the shares or bonds issued by such companies.11 One or both of the parents in a family, if not employed by a government or nonprofit organization, are most likely employed by a business to participate in the production of market goods and services. When the firm’s product is sold—for example, when a family purchases a car or a computer—the money from the sale is used to compensate the workers and pay interest and dividends to the bondholders and shareholders, who effectively own the business property that participates in production. (Any profits retained in the firm rather than paid out will increase the value of the shares.)
In some ways, human and nonhuman wealth are similar. Both are usually reproducible; that is, new examples can be produced by a combination of existing human and/or nonhuman resources. Both must be properly maintained in order to remain productive. The economic value of both investments can. depreciate: Machines wear out in use, or their services lose value because of market changes; the same is true of people.
Why Families Put Investment in Their Children First
The two kinds of wealth differ fundamentally, because human wealth is embodied in. human persons, and there is no longer-fortunately-a market for buying and selling human beings, as there are markets for buying and selling all kinds of property. This difference has two side effects, which are not immediately obvious, but are highly significant for the economics of the family: First, for most families, the average rate of return on investment in “human capital,” such as child-rearing and education, is significantly higher than the market rate of return on investment in property Second, the returns on investment in humans diminish much more rapidly than on investment in nonhuman wealth.
When investing in property (say, the stock market), the amount invested does not appreciably reduce the rate of return—at least, not until one is investing many billions of dollars, and even then, not by much. But for most families with dependent children, the real rate of return on investing time and money in child-rearing and education (in terms of higher lifetime earnings of their children) is much higher than the average return that can be received from investing in the stock market. For example, the long-run, inflation-adjusted average rate of return on the stock market is 5 to 7 percent, or about 8 to 10 percent calculated before subtracting the tax on business profits. But estimates of the average rate of return on the costs of child-rearing and education are consistently about 5 percentage points higher than this. For example, a recent survey calculated the gross rate of return on the time and money invested in a college education in the United States at about 19 percent in 1999-2000 for both men and women—that is, about 16 percent beyond inflation.12 This provides a sound economic reason to help explain why most parents pay for their children’s rearing and education first—instead of, say, leaving their children uneducated and investing the tuition money in the stock market. It’s also the main economic reason why most parents, if they cannot afford to do both, pay for raising and educating their children before saving for their own retirement. Families invest in “things” only after they’ve run out of attractive investments in people.
Why ”Empty-Nesters’” Shift to Investing in Property
But now let’s consider the other side of the difference between investing in people and in property: With each extra year or dollar invested in acquiring valuable skills, the rate of return on human capital declines fairly rapidly after a certain point. Even if earnings did not decline with age, as they do for most people, our lifetimes are finite, so the present, value of our human wealth falls as our remaining life diminishes. The same study cited earlier calculated that the rate of return on college education for men falls to about 9 percent at age 40, 7 percent at age 45, and less than 4 percent at age 50. Moreover, after about age 50, human capital depreciates in value. Most of us lose physical and mental energy and economic productivity (Nonhuman capital also depreciates, but. as we will see, unlike human capital, an allowance for this property depreciation is already accounted for in calculating business profits and subtracted from the market rate of return on property.) When the rate of return, on investing further in human capital falls below the market rate of return that can be realized by investing in property, families start to invest, in the latter. Most American families run out of resources before the return on additional investment in their dependent children, falls below the market return on nonhuman capital (though the minority exception is larger than it used to be). By the same token, families with a large amount of wealth after paying for their children’s upbringing have little choice but to invest most of it in property rather than their family’s own “human capital.”
Income Over the Human Life-Cycle
These simple facts about the family suggest the lifetime pattern of labor-market earnings and consumption of market goods shown in the above graph. During our childhood dependency, we have no market earnings. Any expenses for market goods are paid by our parents. This allows us to spend the time (for example, in school) that is necessary to acquire the skills and qualities that will support us through most of life.
By the start of active parenthood, our investment in “human capital” during childhood has begun to earn a return as we supply our labor services to businesses in exchange for labor compensation. The rate of return on our earlier investment in human capital is highest during this stage, and our labor income rises rapidly. Mostly out of their labor income, parents must pay for the consumption of and any investment in market goods by both themselves and their children. For most families this requires some borrowing—for example, in the form of mort-gages or auto loans—from lenders who either have more nonhuman wealth or are already in the next, “empty nest” stage. During the “empty nest” stage, though the rate of return on human capital is lower than during parenthood, absolute labor market earnings rise to reach their peak. Because the children have left home, current expenses go down, so parents are able repay earlier borrowing and begin saving significantly for their own retirement, usually by investing as much as possible in the market for nonhuman capital. Finally, in retirement, parents leave the labor market, which causes labor income to cease, and to pay for current, living expenses they use previous investments in nonhuman capital, along with any government and business pensions they earned during their working years. (For simplicity, the graph shows the level of consumption and investment in market goods as the same in each phase of life. This is not necessarily the case—for example, the costs of “human maintenance” tend to rise with age—but it’s certainly the case that consumption is much more even than income over a human lifetime.)
We can see how well our simple “model” of family economics explains reality by comparing it with Census data on labor market earnings by age and sex (graph above). Generally speaking, the pattern of labor market earnings by age for both men and women is quite similar to the one we have described: Earnings rise rapidly until about age 30, continue rising (but less rapidly) to a peak at about age 50, and then decline to almost zero after retirement age. However, the labor compensation of men is higher at all ages than that of women. Census and Social Security Administration data suggest that the lifetime market labor compensation for men is about twice as high as the lifetime labor compensation for women.
Marriage as an Economic Partnership
What explains the difference? It cannot be claimed, as it might have in the past, that this difference is due mostly to different educational opportunities or discrimination against women. If anything, young women now receive more formal schooling than young men. By focusing on the relationship between parents and children, and on phases of life that are common to everyone, I have left in the background another important economic fact of family life. Husband and wife are, in effect, partners in a small business; and to make the most of their household resources, most married couples work out a specialization of economic roles. One spouse—usually but not necessarily, the husband-concentrates more on earning cash income for the family by working in the labor market. The other spouse—usually but not necessarily, the wife—concentrates on supplying services directly to the household. The exact balance between labor-market and household work depends on the relative education, skills, and aptitudes of the couple. This specialization affects the labor-market earnings of both spouses. The persistent difference in earnings between men and women is partly due to the fact that specialization in the labor market results in higher market labor compensation for the spouse who specializes in this activity, and partly because the value of the services supplied by the spouse working within the household does not show up in government statistics. The difference in lifetime labor market earnings of men and women is especially important in considering reforms, like replacing Social Security retirement pensions with private financial accounts, which would make each person’s retirement income dependent largely on that person’s life-time earnings in the labor market.
The Economic Logic Of Family Income Distribution
Appreciating these basic facts about the family makes it easier to under-stand the distribution of income among different families in the United States. We recall from our earlier discussion of the economic relations of the household and the business firm that the sale of a product by a business firm provides the compensation for its “factors”: the workers and owners of the productive property’ that produced it. And what is true of a single exchange is true if we add up all exchanges. Since factor income and spending on final products are equal, we can view the whole economy either as total spending on final products or as total income to the producers.’13 (That’s why government statistics are called the “national income and product accounts.”)
Looking at the income side, before considering the effect of taxes and government benefits, our economy is comprised entirely of labor compensation and property compensation. About two-thirds of Gross National Income (the counterpart, to Gross National Product) consists of labor compensation (wages, salaries, and fringe benefits), while about one-third is property compensation (dividends, retained profits, interest, rents, and royalties). This ratio has been remarkably constant as far back as the statistics can measure it (at least since 1929). Theory suggests that this is because workers consistently contribute about two-thirds, and property about one-third, of the value of all final products.
How (and Why) Income Is Distributed Among American Families
However, the mix between labor and property compensation among different families varies systematically Our simple “model” of the family suggests, among other things, that the lower the level of family income and wealth, the greater the share of wealth that will be invested in human capital rather than nonhuman capital, and the higher the share of income from labor compensation rather than property compensation. The higher the level of family income and wealth, the higher the share of wealth invested in nonhuman capital and the higher the share of income that comes from property compensation rather than labor income. We also saw that labor compensation and total income are negligible in childhood, rise to a peak in middle age, and then fall back towards nil in retirement. But whether income is high or low, the younger the person, the more of his or her wealth will be invested in human rather than nonhuman capital and the larger the share of income from labor compensation; the older the person, the more wealth will be invested in nonhuman capital and the larger the share of property income.
Whether or not this “model” of the family is accurate for American families in general can be ascertained by considering Treasury Department figures for the distribution of “Family Economic Income” in the United States, a broad measure of both labor and property income.14 (The Congressional Budget Office has a similar measure, known as Comprehensive Household Income.15) For comparison, American families can be divided into percentiles according to family economic income. In 2000, the 20″‘ percentile started at $17,988, the 40th at $34,844, the 60th at $59,019, the 80th at $100,767, the 90th at $140,581, the 95th at $189,835, and the 99th at $462,053.16
How the Figures Must Be Adjusted to Compare Apples with Apples
Family Economic Income is reported after (that is, including) transfer payments received, but before (without subtracting) the taxes that pay for them. To compare apples with apples, we must sometimes place the figures on a consistent after-tax, after-transfer or before-tax, before-transfer basis. Beyond this, we must be aware of some serious inconsistencies in such measures, which, though pointed out by Schultz more than 40 years ago17, have not yet been corrected. Family Economic Income treats labor and property compensation inconsistently by excluding the costs of maintenance and depreciation of nonhuman capital from property income, but including the costs of maintenance and depreciation of human capital in labor compensation. To remove this inconsistency, I will adjust, the Family Economic Income figures when appropriate to measure property’ income inclusive of depreciation, the same as with labor compensation. (I will deal with the difference in treatment of maintenance costs later.)
Whether or not the figure is adjusted in this way, the shares of labor and property compensation, calculated before taxes and transfer payments, agree with our simple “model”: The lower the absolute level of income, the higher the share of labor compensation, while the higher the income, the higher the share of property compensation. For consistency, I will use Adjusted Family Economic Income. The share of labor compensation in pretax, pretransfer Adjusted Family Economic Income is 83 percent for the bottom fifth (quintile) of families, 76 percent for the second fifth, 74 percent for the middle fifth, 77 percent of the next-to-top fifth, and falls to 61 percent for the top fifth of families. In other words, 80 percent of American families receive at least three-quarters of their income in the form of labor compensation, before taxes and transfers. The share of labor compensation drops to 56 percent for the top 10 percent of families, to 50 percent for the top 5 percent of families, and to 42 percent for the top 1 percent of families. This pattern is exactly what our simple “model” of family economics would suggest.
The Effect of Taxes and “Transfer Payments”
Though all income originates as labor or property compensation, we must take into account another kind of income: government “transfer payments” to persons who don’t contribute to producing current out-put and income. This income to the beneficiaries must be taxed or otherwise transferred away from those who originally earned the labor or property compensation. Most households at every income level receive a combination of all three kinds of income. But this combination changes systematically by income level.
For the bottom fifth of all families, transfer payments make up about 50 percent, labor compensation 44 percent, and property compensation 6 percent of after-tax, after-transfer income. The share of after-tax, after-transfer property income is about, the same for the lower four-fifths of families, ranging from 6 to 11 percent, but rises to about 21 percent for the top fifth. For the top 10 percent, the property income share rises to 28 percent, for the top 5 percent, to 35 percent, and for the top 1 per-cent of families, to 45 percent. The share of income from transfer payments falls from 50 percent, for the bottom fifth of families, to less than 1 percent for the top 1 percent, of families. This also is not surprising, since many transfer payments are “means-tested”: restricted to those with low incomes. But unlike pretax, pretransfer income, the distribution of labor compensation after taxes and transfers is “dome-shaped”: After-tax labor compensation contributes 44 percent, of income for the bottom fifth of families, rises to 87 percent of income for families in the 60th to 80th percentiles, then declines steadily to 54 percent, for the top 1 percent of families. (See Graph)
The fact that the distribution of property income is similar when measured before and after taxes and transfers, while the distribution of labor compensation is quite different, strongly suggests that most transfer payments are paid out of the taxes levied on labor compensation. Put another way, progressive tax rates combined width government benefits do not redistribute income from families with higher total incomes to families with lower total incomes. Rather, such redistribution transfers income mostly from families width higher labor compensation to families with lower labor compensation, leaving the distribution of property income relatively unaffected.
The Economic Logic Of Tax Reform
Having grasped the basic economics of the family and the distribution of income among American families, the economic and political logic of tax reform becomes clearer.
There are essentially two substantive choices to be made in considering fundamental tax reform: the tax base and the tax rates. Apart from a couple of taxes levied on wealth (like the estate tax or the local property tax), the two basic approaches to the tax base are an income tax and a consumption tax. However, we face a further choice, which has more to do with appearances than substance: whether the tax is collected on expenditures or on income.
A Prediction Vindicated
The terms “consumption tax” and “flat tax” are often used inter-changeably, because “flat tax” was the term used by Republican presidential candidate Steve Forbes in 1996 for a plan that would impose a flat rate on a tax base that is often described as a “consumption” tax base. An earlier version had been sponsored by Congressman Dick Armey.18 But it’s possible to combine a flat rate with an income tax base, or progressive tax rates on a consumption tax base, and vice versa. The same tax rate will have quite different, results on the distribution of the tax burden, depending on the choice of tax base.
In June 1995 I could not have known that Steve Forbes would start to1 run for president a few months later on the flat tax plan he announced at that time. But just as the commission’s hearings began, I wrote in a memo to Jack Kemp, chairman of the National Commission on Economic Growth and Tax Reform, that “the plain fact is that all the ‘consumption’ taxes are going to self-destruct politically, when it becomes apparent that any revenue-neutral version will represent a tax increase for working families.”19 This proved to be an accurate prediction of what happened to the Forbes plan. Forbes’ candidacy soared in the polls at first as voters were attracted to the elegant simplicity of the plan to replace the current tax code with a much simpler one. But the popularity of the plan sank like a stone once voters learned more about how it would affect their families.
Income v. Consumption Tax
To understand why, we need to1 understand more clearly what is meant by “investment” and “consumption.” Out of what is produced in any year, some of the product is destroyed. This, properly speaking, is “consumption.” Some goods, like the food that we eat, are what we intuitively think of as “consumption.” But the decline in value of a machine as a result of use in production is also (and properly) referred to as “cap-ital consumption.” The same is true also of the “human capital” of workers. If we don’t eat properly, we lose weight; if we don’t maintain our skills, they decline in value. Only the part of a year’s production that remains in existence at the end of the year is, properly speaking, new “investment.” Thus all production (and therefore income) is necessarily devoted entirely to consumption or investment.
In principle, a consumption tax attempts to tax only the current using-up of goods and services. The underlying theory is that exempting new investment will encourage such investment, thus increasing future production and income, and that eventually, all products must be consumed. Obviously, this assumes that the tax on consumption cannot be deferred for very long, or if so, that it doesn’t matter. An income tax taxes the ability to consume or use goods and services, whether or not that ability is currently exercised. By taxing all income when it is earned, the income tax effectively “prepays” the tax on any future consumption out of that income—so consumption should not be taxed separately. There is no economic difference between the two approaches, unless the different lengths of time between production and consumption somehow result in the tax burden being higher or lower under one approach than under the other.
Taxes on Compensation and Spending Are Equivalent
Since economic activity can always be viewed in two ways—either as compensation received by producers, or as expenditures made by the purchasers of the products—whether a tax is levied on compensation or spending is largely a cosmetic difference. In economic terms, taxing half of the income produced by a factor over its lifetime is equivalent to taxing half its total value, and vice versa. Because of this equivalence, it. is possible to devise either a consumption tax, or a tax on all production and income, by taxing either compensation received by producers or expenditures made by purchasers of the products. Thus, a consumption tax base may result either from excluding spending on investment from a tax on all spending; or, equivalently. it may result from excluding the saving that ultimately finances such investment from a tax on labor and property compensation. Similarly, an income tax can result from including savings in taxable income; or equivalently, from including spending on new investment spending (along with spending on consumption) in a tax on purchases. It is also possible to combine the two approaches—to produce either an income tax or a consumption tax—in ways that may be either consistent or inconsistent. A consistent approach taxes all income or spending once. An inconsistent approach may tax some kinds of income or spending more than once, and others not at all.
“Consumption” Equals Labor Income
The Armey/Forbes plan was in fact two different taxes: one tax on labor compensation received by workers, and another “business transfer tax” on business receipts, less part of labor compensation, the cost of raw and intermediate materials, and investment in finished goods.
This mixture might, have been mostly consistent, except for one thing: its peculiar restriction of the term “investment” to mean investment in nonhuman capital, thus excluding investment in human capital. As Theodore Schultz noted more than 40 years ago, “Much of what we call consumption constitutes investment in human capital. Direct expenditures on education, health, and internal migration to take advantage of better job opportunities are obvious examples.”20 By effectively restricting the definition of “investment” to the purchase of new buildings and machines, the Forbes plan was essentially a tax on labor compensation, or equivalently on investment in human capital. New investment in property would be deductible from the business tax base, but interest, dividends, and capital gains derived from such investment were not tax-able when received by individuals.
To understand the implications of the proposed change, compare it with the distribution of the current federal income tax. The current tax code is called (and to a certain extent resembles) an income tax, but is riddled with many anomalies. If an income tax is chosen to fund general purposes, both equity and efficiency suggest that it should fall in equal proportion upon all kinds of income. Yet it’s still true, as Schultz said in 1961, that “Our tax laws everywhere discriminate against human capital” and “in favor of nonhuman capital.”21
Current Law: Disparate Treatment of Investment in People and Property
For example, a bread-making company, before calculating its gross income, deducts the cost of maintaining and operating its bread-making machines. These costs of capital “maintenance” are subtracted before calculating gross pretax income and never appear in national income or product statistics. In calculating taxable income, the business also deducts from its gross income the cost of acquiring the machines, written off over one or more years, as an allowance for the cost of depreciation.
But the worker who operates the same bread-making machine may not write off the cash cost of acquiring his skills (or of investing in his children’s), nor an allowance for their depreciation. Standard deductions and personal exemptions might be considered a deduction for the “maintenance cost” of keeping body and soul together, but unlike the corresponding business deduction, are not equal to the actual outlays for that purpose. Such allowances also do not apply to the payroll tax, which is larger for most American families than the income tax, though the purpose of the Earned Income Tax Credit, instituted by Senator Russell Long in the 1970S, was to exempt income below the poverty level from the payroll tax.
In general, the federal income tax permits deductions from property compensation for the costs of both property maintenance and property depreciation. It also provides a partial deduction from labor compensation for human “maintenance costs,” but none for depreciation.
There are also anomalies in the treatment of property income. There is a federal tax on both personal income and corporate profits, and to some extent, the two codes are integrated to avoid double taxation: For example, a business may deduct interest paid to creditors from its taxable profit, and the interest received by bondholders is taxed as personal income. But the same treatment is not extended to dividends paid, which are not deductible when paid by the corporation, yet are still taxed when received as personal income. Likewise, capital gains on investment property are taxed without reference to whether they result from retained corporate profits which have already been taxed. Moreover, the gains are calculated without reference to inflation. And so on.
More Labor than Property Income in the Tax Base
On balance, however, gross labor compensation is taxed more heavily at the federal level than gross property income. This can be seen from the fact that, while workers consistently produce about two-thirds of all Gross National Income, they pay more than three-quarters of all federal taxes on personal and corporate income. The basic reason is that, even though more discrete taxes are levied on property’ compensation than on labor compensation, a much larger share of gross property compensation than labor compensation is excluded from the tax base.
The next graph shows the effect of these differences on the size and distribution of the tax base. With all its flaws, the current income tax base, by taxing most gross labor compensation and much net property compensation, would result in a progressive distribution of the tax burden, even with a flat tax rate, if all income below the poverty level were excluded by some combination of standard deductions, personal exemptions, or credits.
Now compare the current income tax base with the Armey/Forbes plan. Because, over time, all investment in property or property income would be excluded from the tax base, only labor compensation would be taxed.22
Different Tax Base, Same Tax Rate: Different Distribution of the Tax Burden
Compared with any income tax, a “consumption tax,” at least with “investment” defined as investment in property, includes a significantly larger share of income in the tax base for middle- and upper-middle-income families than it does for the highest-income families. But the total “consumption” tax base is about 23 percent smaller than the tax base including all gross labor and net property income. To raise the same amount of revenue under both tax bases, the “consumption” tax rate must be about 30 percent higher. For example, to raise the same amount of revenue as a flat tax rate of 19 percent on all gross labor and net property income above the poverty’ level, a “consumption” tax would require a tax rate of 26 percent with the same exemptions.
Because of the exclusion of income below the poverty level, the income tax burden on the bottom one-fifth of American families would be the same: zero. (This pertains to the federal income tax burden before tax credits like the Earned Income Tax Credit, which is refundable and therefore can result in a negative tax, and also ignores the payroll tax, which begins on the first dollar of labor income without any exemptions.) For the top fifth of families, the tax burden would fall, and for the top 1 percent of families the tax burden would fall by more than a third, compared with a flat tax on net income. But for families in the 60th to the 80th percentiles, the tax burden would rise, with the tax increase falling entirely on labor compensation—and therefore effectively on families’ investments in the “human capital” of their children. Overall, about twice as many families would receive a significant tax increase as a significant tax cut.
In other words, a flat tax rate on a tax base that taxes only labor compensation, by exempting both property income and investment in property (even if labor income below the poverty level is also tax-exempt) is a “lumpy” tax, not a flat tax: It would shift virtually the entire tax burden onto workers in each income class, and onto the middle class as a group. This remains the case even when the extra growth that could be expected from flatter marginal tax rates is taken into account.
Why the Armey and Forbes Rat-Tax Plans Failed
The Armey version of the flat tax went one step further in raising the tax burden on workers by repealing the Earned Income Tax Credit (EITC). It also would have made workers’ fringe benefits and payroll taxes nondeductible for employers. Under this version, raising the same amount of revenue as current law would have resulted in an increase in the federal tax burden for more than 90 percent of American taxpayers. Though retaining the general outlines of the Armey plan, Forbes reduced the problem at lower incomes by retaining the EITC, but this could not prevent tax increases for middle-income families under a revenue-neutral Forbes plan. (Forbes argued that he would also cut federal spending enough to permit a 17-percent tax rate, but such spending cuts would also make a lower tax rate possible under any other tax plan. Whether we assume that federal tax revenue is higher, lower, or the same as current law in order to compare apples with apples, we must assume that the same revenue is raised under all plans being compared.)
To predict that such a plan would politically “self-destruct,” therefore, did not require a crystal ball: It. was necessary only to1 know the economics of the family and the distribution of income in the United States. To know that American voters would reject such a plan, it was only necessary to know that the tax burden would go up for many more families than it would go down. In political terms, any version of a “consumption” tax that narrowly defines “investment” as investment in property is a nonstarter. It would mean a tax increase on the majority of middle-class households, whose income comes mostly from the return on their investment in “human capital.” Under most such plans, workers’ share of the federal income tax burden would rise from over three-quarters to nearly 100 percent.23
In economic terms, such a tax base would worsen the allocation of resources, and therefore lower rather than raise total output and real income. The assumption underlying such an approach, which I call the “Stork Theory,” is that the number of workers and their skills are “given,” rather than being the result of investment by families, and can therefore be taken for granted.24 Under the Stork assumption, any tax on nonhuman capital or property income will and should be shifted entirely to human capital or labor compensation; but a tax on labor compensation should not be shifted to property’ income.25 As Gary S. Becker has pointed out, this would only make sense if the population were fixed.26 He points out that “even a modest tax on births can have a large negative effect on the number of children.”27 Shifting to a so-called “consumption” tax would lower the birth rate, reduce investment in children, and ultimately reduce the size of the workforce and the economy. Since the return on investment in human capital for most people is higher than on nonhuman capital, even the smaller population would have lower incomes than without the reform. All this has been the case in Japan and Western Europe, wherever the tax burden has been shifted in recent decades heavily towards “consumption” taxes.
Family-Friendly Tax Reform
After outlining these basic facts about the economics of American families, I suggested to the Kemp Commission that the key to a tax reform that would be equitable, economically efficient, and family-friendly is simply to adopt the principle that investment in both human capital and nonhuman capital (or, equivalently, taxation of both property and labor compensation) should as far as possible be treated alike.28
Beyond deducting poverty-level “human maintenance” from taxable income, equal treatment for labor and property income could in principle be done with either a consumption tax or an income-tax approach. The practicability’ of the two, however, is not the same. The consumption-tax approach, for example, would allow for taxation of income after deducting all costs of investment in both human and non-human capital. According to John W. Kendrick’s calculations,29 this would remove as much as half the economy from the tax base, requiring a flat income tax rate as high as 40 percent, on top of the payroll tax—or progressive tax rates rising to an even higher top rate. Such an approach would involve endless complexities of definition: To what extent is a liberal arts education, or a company Mercedes used by an executive, or a computer used by a child for both games and study, to be considered saving or consumption for tax purposes?
I pointed out that many of the undesirable effects of the proposal may be avoided by modifying the Armey/Forbes plan in two ways.30 First, replace the existing standard deductions, personal exemptions, and Earned Income Tax Credit (as well as the more recently enacted child credit) with an Earned Personal Tax Credit, based on the number of persons in each family, against both federal income and payroll taxes up to approximately the poverty level. Second, there would be no separate deduction for the costs of investing in either human or nonhuman cap-ital: no deduction for “expensing” or “depreciating” the cost of an asset. This would have preserved the simplicity’ of the Forbes plan, offers the most neutral treatment of human and nonhuman capital, and. there-fore, is the most efficient allocation of resources. By bringing most of the economy into the tax base, it would have the lowest revenue-neutral marginal income tax rate of any tax-reform plan: about 16 percent. An earlier graph compared the tax burden on such a gross income tax base above the poverty level under a flat tax against both a tax base consisting (like current law) of net income (after subtracting depreciation of nonhuman capital) and a “consumption” tax like the Forbes plan. To raise the same revenue as a flat tax rate of 19 percent on a tax base resembling current law on gross labor and net property income, a “consumption” tax requires a 26-percent tax rate, while the LBMC Plan needs only a 16-percent tax rate. We have also seen that the “consumption” tax would impose a tax increase on at least twice as many American families as would receive a tax cut. But as the graph shows, distribution of the tax burden by income class under the LBMC plan, with its lower tax rate, is slightly more progressive than under the cur-rent net income tax base, which would require a higher tax rate.
The main problem with the Forbes plan was that it increased the tax on a majority of American families. The 1995 version of the LBMC Plan avoided this problem. But the main problem with the 1995 version of that plan, from my point of view, is that the Armey/Forbes plan was not the best starting point for tax reform. As mentioned earlier, that approach is in fact two separate tax systems: one tax on labor compensation (but not property compensation), and another tax collected from businesses on their sales of products, minus materials purchased, some but not all labor compensation, and new investment in plant and equipment. In Europe, the business tax would be called a “consumption VAT,” or value-added tax. (A value-added tax is levied only on the “value added” by production of each firm, not the total cost of the product, like a sales tax. This avoids taxing the same goods more than once during the intermediate processes of production.)
While both the Forbes plan and the 1995 LBMC Plan were much simpler than the current tax code, both were still not easily grasped by the average worker or businessman. Workers found it hard to understand why they had to pay taxes on their wages while owners of property did not seem to pay taxes on interest and dividends.31 As for the LBMC Plan of 1995, businessmen who were used to thinking in terms of the existing, separate tax on corporate profits, on top of the federal income tax on personal income, were terrified of losing their deductions for depreciation of nonhuman capital investment. Within the current tax system, that would indeed be a major change—but not if the corporate income tax itself and its extra layer of taxation were abolished.
Experience suggests that tax reform and the tax system would be much simpler and more “transparent” to voters and taxpayers if there were only one tax to understand instead of two. Therefore, I devised a simpler LBMC Plan. Instead of requiring workers to pay taxes on their labor compensation while allowing a deduction of that compensation when paid by businesses (and using an opposite treatment of property compensation), both taxes would be consolidated into a single “business transfer tax” on all market products and services. Apart, from the cost of raw materials and intermediate goods (to avoid multiple taxation of the same “value added” and income), there would be no deductions for labor or property compensation paid to workers and property owners. The tax would be levied on imports but rebated on exports, bringing U.S. practice into line with most of our trading partners—and incidentally, bringing the $500 billion U.S. trade deficit into the tax base. Virtually all of Gross Domestic Product would be in the tax base. All “value added” would be taxed, but only once. To avoid taxing “human maintenance” costs, while continuing to deduct maintenance costs for nonhuman capital investment, each family would receive a tax rebate equal to the flat tax rate plus the payroll tax rate, up to approximately the poverty level.32 The credit would be worth up to about $1,500 per family member, or $6,000 for a family of four. There would be no “marriage penalty,” nor any penalty (or preference) for mothers working in the home rather than in the labor market. Most people would have no contact with the IRS except to receive their tax rebate. And the IRS would have the task of collecting the tax from only a few hundred thousand businesses, instead of over 100 million households.
Bush Dividend Plan Points the Way
President George W. Bush’s proposal to end the double taxation of dividends is interesting, even though—or perhaps because—it was not the simplest way of eliminating double taxation within the current tax system. The president could have proposed to allow a deduction from tax-able business profits for dividends paid, while leaving the personal income tax on dividends received. This would have made the treatment of interest and dividends the same. Instead, President Bush proposed keeping dividends nondeductible when paid by businesses, but making them non-taxable when received: the opposite of the treatment for interest. The plan has been criticized for its complexity. But the treatment itself is quite simple; the complications stem from the fact that the rest of the tax code treats most other income differently: The complications come from making sure that someone has already paid tax on the income if it is received without tax. The plan is sensitive to the reality that, where taxation is concerned, appearances matter. The sub-stance must be correct, but its rationale and effect must also be clear to the public. The true significance of the dividend proposal is that it points the way toward a radical simplification of the tax code without sacrificing fairness. The simplified LBMC Plan I have described amounts to nothing more nor less than extending to all income the treatment President Bush is proposing for dividends alone: All income should be taxed once, but only once, by making it non-deductible when paid, but non-taxable when received.
Fixing Social Security
With its tax rebate, the LBMC Plan is designed to take into account the payroll tax, which is larger than the income tax for a majority of American families. The second part of the LBMC Plan would also balance the pay-as-you-go Social Security retirement system.
The basic problem facing the Social Security’ system is that the system now has a large surplus, because current benefits to retirees are only 75 percent of current tax receipts; but after a couple of decades, the payroll tax will be bringing in only about 75 percent of currently promised retirement benefits. Meanwhile, the Congress has been plundering the Social Security surplus to fund an equal deficit in spending on the rest of the federal budget. In other words, even when the overall federal budget is balanced, federal tax revenues other than Social Security payroll taxes are less than non-Social Security spending, by an amount equal to the Social Security surplus.
The LBMC Plan would solve the present and future Social Security imbalance by cutting the current OASI (Old Age and Survivors Insurance) payroll tax by 20 percent now, in exchange for a reduction in future retirement benefits of up to 20 percent (prorated according to the number of years the tax cuts are received). Families would be free to invest the tax cut either in their own “human capital” or in the financial markets. After the plan was brought into balance, the payroll tax would be adjusted as necessary to keep the system in balance with a minimum Trust Fund reserve of one year’s benefits. In comparison to various other plans for Social Security’ reform, the LBMC Plan would offer American families a higher rate of return on their retirement savings than is possible from either a privatized retirement system or from reform proposals involving tax increases and/or increases in the retirement age.33
However, balancing Social Security would expose the chronic deficit in the rest of the federal budget. This is why, even though either part of the LBMC Plan could in theory be enacted by itself, practically speaking, both would probably need to be enacted at the same time. In order to finance the a-percentage-point reduction in the payroll tax, the flat income tax rate would have to be about 18 percent, instead of the 16 percent that would equal current-law federal tax revenues. If both reforms were enacted simultaneously, the total annual federal tax bur-den would be roughly unchanged from current law in the near term, while the large tax increases that would otherwise be necessary in coming decades would be avoided. In the process, efficiency and equity would have been restored to both the federal tax system and the Social Security system.
Recent history has shown that would-be reformers of the tax code and of the Social Security system ignore the economic realities of the family at their peril. But if we proceed with an understanding of those realities, it is possible to design fundamental reform of both the tax code and the Social Security system that is fair, economically efficient, and family-friendly.
1. From 1979 to 1988, I had been then-Congressman Jack Kemp’s economic policy staffer, mostly while serving as economic counsel to the Republican caucus in the House of Representatives, of which Kemp was chairman.
2. Named after the four principals of the financial-markets forecasting firm Lehrman Bell Mueller Cannon, Inc., who jointly advocated the plan to the Tax Reform Commission.
3. Statistical Abstract of the United States, 2001, Tables 53, 54, 57, 58, 59.
4. Statistical Abstract, 2001, Table 40.
5. Historical Statistics of the United States: Colonial Times to 1970 (Washington, DC: Bureau of the Census, 1975): Series B40.
6. Statistical Abstract, 1997, Table 107.
7. Statistical Abstract, 2000, Table 145; Statistical Abstract, 1999, Tables 155-161.
8. Historical Statistics, Series B107-109.
9. Statistical Abstract, 2001, Table 98.
10. Theodore W. Schultz, “Investment in Human Capital,” American Economic Review (March 1961): 1-17.
11. The main forms of wealth that represent exceptions to this ownership arrangement are the “common” or “public” goods owned by governments (e.g., defense equipment, government office buildings and equipment, and public parks) and other non-profit institutions like schools.
12. Bloendal, S., S. Fickel, N. Girouard, and A. Wagner, “Investment in Human Capital Through Post-Compulsory Education and Training,” Organization for Economic Cooperation and Development (Paris, 2001): 10. As with business property, when taxes and the social costs of subsidies for investment are taken into account, the “social rate of return” is somewhat lower — but the comparison still shows a significantly higher rate of return on investment in “human capital” than on business property.
13. By saying that income equals product, I am ignoring the balance of international transactions, which would complicate things without affecting the substance of this discussion. National income and product data pertain only to “final” products, since including the value of raw materials and intermediate goods as well as finished goods would result in multiple counting of the same “value added” in production.
14. Cronin, Julie-Anne, “U.S. Treasury Distributional Analysis Methodology,” OTA Paper 85, Office of Tax Analysis, U.S. Department of the Treasury (September 1999).
15. “Effective Federal Tax Rates, 1979-1997,” Congressional Budget Office (October 2001).
16. The absolute figures for income received in cash would be lower than this, but the order of ranking would be similar.
17. “Our tax laws everywhere discriminate against human capital. Although the stock of such capital has become large and even though it is obvious that human capital, like other forms of reproducible capital, depreciates, becomes obsolete, and entails maintenance, our tax laws are all but blind on these matters.” Op. cit., 13.
18. Based on a plan originally devised by economists Robert Hall and Alvin Rabushka in 1981.
19. John D. Mueller, “Hearings with Members,” memo to Jack Kemp, June 27, 1995.
20. Op. cit., 1.
21. Op. cit., 13, 15.
22. The length of time this requires would depend on the transition from the current tax code, which is not spelled out in any of the plans — for example, whether key provisions in current law would be “grandfathered.”
23. I have shown that the revenue-neutral flat tax rate would have to be about 27 percent when such a plan is fully phased in, and would be levied almost entirely on workers’ incomes. John D. Mueller, “The Distribution of a Federal ‘Consumption’ Tax,” paper submitted to the National Commission on Economic Growth and Tax Reform, August 10, 1995.
24. John D. Mueller, “The Stork Theory of Economics,” Family Policy (Jan-Feb. 2001): 15-18.
25. In tax theory, the incidence of a tax and the burden of a tax are not necessarily the same. That is, the person upon whom the tax initially falls may not be the person who ultimately bears the burden. The burden depends on a number of factors, especially on how easy it is to find substitutes for the good that is taxed. The proponents of such tax reform plans are, to a man, adherents of the old neoclassical model — the “Stork Theory” — which, as we saw, assumes that the population is given but the stock of physical capital is not.
26. “The neoclassical conclusion about tax incidence, that in the long run a tax on capital is fully shifted to other factors, no longer holds when fertility is endogenous.” Gary S. Becker, A Treatise on the Fmilay (Boston: Harvard University Press, 1994): 18.
27. Gary S. Becker, Human Capital, 3rd ed. (Chicago: University of Chicago Press, 1994): 23.
28. John D. Mueller, “The Tax Treatment of Human and Nonhuman Capital,” testimony delivered to the National Committee on Economic Growth and Tax Reform, June 21, 1995.
29. John W. Kendrick, The Formation and Stocks of Total Capital (New York: Columbia University Press for the National Bureau of Economic Research, 1976); “Total Capital and Economic Growth,” Atlantic Economic Journal (March 1994): 1-18.
30. As I proposed in 1995. See John D. Mueller, “The LBMC Plan for Tax Reform,” paper submitted as an adviser to the National Committee on Economic Growth and Tax Reform, September 26, 1995. Also, John D. Mueller, “A Flat Tax That’s Good for Families, Not Just for Business,” Family Policy (Nov.-Dec. 1998): 1-16.
31. Because interest and dividends paid to bondholders and shareholders were not deductible by businesses, the tax was already collected at the source. Or at least, it would have been, if teh plan did not also permit “expensing” or deduction of all investment in property.
32. The payroll tax would continue to be dedicated to funding Social Security and Medicare benefits for retired workers.
33. John D. Mueller, “Winners and Losers from ‘Privatizing’ Social Security,” a report undertaken in cooperation with the Employee Benefits Research Institute (EBRI) and Policy Simulation Group, Washington, D.C., March 1999.