Chapter Two
Inequality
Each year, Parade magazine has a cover story entitled “What People Earn”. In its most recent version, it displays photographs of 124 people together with their earnings, which vary from zero to many millions of dollars.[i] Twenty-five persons, one fifth of the total, are shown with earnings of a million dollars or more. Is this an accurate picture of America today? If 124 people are selected to be representative of our society, how many would be millionaires? The correct answer is: none. Only one of about 500 Americans has earnings that high. It would perhaps make for less titillating reading to show only a handful of people (out of 124) with earnings in excess of $100,000, and no millionaires, but it would be more accurate.
Egalitarians like to produce stories that contrast the lifestyles of the rich and superrich with the middle class and poor, emphasizing the extent and unfairness of inequality. Partly as a result of this, it is no surprise that polls show that Americans greatly overstate the extent of inequality that does characterize our society. Yet if we are to have sensible policies to deal with this issue, it is clearly important to base them on an accurate account of how much inequality does exist. That is the major goal of this chapter.
Lying with Statistics
Most studies that deal with inequality rely on statistics that vastly overstate the extent of true economic inequality in America. As we will see, that is true of the widely used Census data just described. This is not because the Census estimates are inaccurate; they accurately measure what they say they measure. The problem is that what they measure is inadequate to judge the extent of inequality. This is what is measured: One Year’s Before-Tax Money Income of Families (or Households). Why do the resulting data overstate inequality? Let us count the ways.
First, the Census figures are for before-tax incomes.[iii] After-tax incomes are what counts in identifying economic inequalities. The after-tax distribution of income is more equal than the before-tax distribution because the relevant taxes fall more heavily on higher income groups (that is, they are progressive, as I will explain in detail in Chapter Nine)
.
Second, much real income received by people is in nonmonetary form and so is not counted at all. Particularly important are government in-kind (noncash) government transfers that add to people’s incomes but are not counted as income. For example, Food Stamps, Medicaid, and housing assistance are not counted as income because they are in-kind, as distinct from cash, transfers. Even one important cash transfer, the Earned Income Tax Credit, is not counted, even though it contributed about $35 billion in income for low income families in 2005. Counting these uncounted transfers as income produces a more equal income distribution since they are received mainly by low income families
.
Third, there are differences in the number of persons per family/household, as we have seen. Higher income families/households have more persons to support, and not taking this into account leaves the impression that these families are better off than they actually are. One way to adjust for this is to measure not family income, but income per person. On that measure, there is less inequality among the quintiles (fifths) than when just comparing family incomes.
Fourth, there are the great differences in the amount of work done among families, as noted in the last chapter. If one family is working twice as many hours as another, and receives twice the income, we should not conclude that the first family is twice as well off. Indeed, in some cases we might reasonably conclude that the second family is as well-off as the first. Although it is not obvious how best to adjust for the vast differences in work effort among the quintiles, any sensible approach would lead to a lower measure of inequality.
Fifth, it may be inappropriate to focus on income. People’s standard of living is actually better evaluated by how much they consume. As we will see, consumption is more equally distributed than income.
Sixth, people more around in the income distribution from year to year. Looking at any one year’s income can give a greatly distorted position of longer-run economic well-being. Longer-run measures of income are more equally distributed than annual measures.
It is easy to be overwhelmed by all the numbers reporting different measures of inequality, but don’t lose sight of the one incontrovertible finding from this body of research: The Census Bureau’s data on annual before-tax money incomes greatly exaggerate the amount of real economic inequality. Instead of the top income class receiving twelve or fifteen times as much as the lowest class, based on the evidence presented here (and other similar studies), a better measure would be a HiLo ratio of three or four to one. Yet egalitarians continue to use the highly flawed and misleading Census data to argue that inequality is excessive.