Upstarts and Downstarts


EDITOR’S NOTE: The August 31, 1992, issue of National Review, set out to set the record straight about the Reagan administration’s economic record. We reprint the content of the issue here.

The Economic policies presided over by Ronald Reagan were stunningly successful–except to informed opinion, as represented by the academy and the major media. The principal charge against Reagan has become almost a chant: The rich got richer, the poor got poorer, and the middle class was squeezed out of existence.

A key player in the campaign to popularize this view has been Sylvia Nasar of the New York Times, who relied on statistics concocted by Paul Krugman of MIT, who, in turn, garbled some already disreputable estimates from the Congressional Budget Office (CBO).

The purpose of the crusade was obvious. Mr. Krugman has been advocating that we somehow double tax collections from those earning over $200,000, so as to greatly increase federal spending. Miss Nasar openly boasted about “supplying fresh ammunition for those . . . searching for new ways to raise government revenue.” Governor Clinton immediately seized upon the Krugman-Nasar statistics as the rationale for his economic plan to tax us into prosperity.

Since the question is what happened in the 1980s, after the Carter Administration, it makes no sense to begin with 1977, as Mr. Krugman and Miss Nasar do, or with 1973, as the Children’s Defense Fund does. Real incomes fell sharply during the runaway inflations of 1974-75 and 1979-80. Median real income among black families, for example, fell 15 per cent from 1973 to 1980, then rose 16 per cent from 1982 to 1990.

The table shows the actual real income of households by fifths of the income distribution, for the most commonly cited years. There is no question that all income groups experienced significant income gains from 1980 to 1989, despite the 1981-82 recession, and were still well ahead of 1980 even in the 1990 slump. For all U.S. households, the mean average of real income rose b 15.2 per cent from 1980 to 1989 (from $33,409 to $38,493, in 1990 dollars), compared with a 0.8 per cent decline from 1970 to 1980.

This table shows that the “income gap” did not widen merely between the bottom fifth and any “top” group, but also between the bottom fifth and the next highest fifth, the middle fifth, and so on.

A common complaint about these figures is that they exclude capital gains, and therefore understate income at the top. However, the figures also exclude taxes. Average income taxes and payroll taxes among the top fifth of households amounted to $24,322 in 1990, according to the Census Bureau, but capital gains among the top fifth were only $14,972. To add the capital gains and not subtract the taxes, as some CBO figures do, is indefensible. Indeed, all CBO estimates of income gains are useless, because they include an estimate of capital gains based on a sample of tax returns. Since lower tax rates on capital gains after 1977 induced more people to sell assets more often, the CBO wrongly records this as increased income. It also ignores all capital losses above the deductible $3,000, and fails to adjust capital gains for inflation.

The Middle-Class Boom

One thing that we know with 100 per cent certainty is that most Americans – far more than half – did very well during the long and strong economic expansion from 1982 to 1989. In those fat years, real after-tax income per person rose by 15.5 per cent, and real median income of families, before taxes, went up 12.5 per cent. That means half of all families had gains larger than 12.5 per cent, while many below the median also had income gains, though not as large. Many families had to have gained even more than 12.5 per cent, since the more familiar mean average rose 16.8 per cent from 1982 to 1989. Even if we begin with 1980, rather than 1982, median income was up 8 per cent by 1989, and mean income by 14.9 per cent. And even if we end this comparison with the slump of 1990, median family income was still up 5.9 per cent from 1980, and mean income was up 12 per cent.

In U.S. News & World Report (March 23, 1992), Paul Krugman claimed that “the income of a few very well-off families soared. This raised average family income – but most families didn’t share in the good times” (emphasis added). Mr. Krugman apparently does not understand what a rising median income means.

The whole idea of dividing people into arbitrary fifths by income ignores the enormous mobility of people in and out of these categories. What was most unusual about the Eighties, though, was that the number moving up far exceeded the number moving down. A Treasury Department study of 14,351 taxpayers shows that 86 per cent of those in the lowest fifth in 1979, and 60 per cent in the second fifth, had moved up into a higher income category by 1988. Among those in the middle income group, 47 per cent moved up, while fewer than 20 per cent moved down. Indeed, many more families moved up than down in every income group except the top 1 per cent, where 53 per cent fell into a lower category. Similar research by Isabel Sawhill and Mark Condon of the Urban Institute found that real incomes of those who started out in the bottom fifth in 1977 had risen 77 per cent by 1986 – more than 15 times as fast as those who started in the top fifth. Miss Sawhill and Mr. Condon concluded that “the rich got a little richer and the poor got much richer.”

This remarkable upward mobility is the sole cause of “The Incredible Shrinking Middle Class,” featured in the May 1992 issue of American Demographics. Measured in constant 1990 dollars, the percentage of families earning between $15,000 and $50,000 fell by 5 points, from about 58 per cent to 53 per cent. This is what is meant by a “shrinking” middle class. We know they didn’t disappear into poverty, because the percentage of families earning less than $15,000 (in 1990 dollars), dropped a bit, from 17.5 per cent in 1980 to 16.9 per cent in 1990. What instead happened is that the percentage earning more than $50,000, in constant dollars, rose by 5 points – from less than 25 per cent to nearly 31 per cent. Several million families “vanished” from the middle class by earning much more money!

It is not possible to reconcile the increase in median incomes with the often-repeated claim that low-wage service jobs (“McJobs”) expanded at the expense of high-wage manufacturing jobs. Actually, there were millions more jobs in sectors where wages were rising most briskly, which meant competitive export industries but also services. From 1980 to 1991, average hourly earnings rose by 6.8 per cent a year in services, compared with only 4.8 per cent in manufacturing. The percentage of working-age Americans with jobs, which had never before the 1980s been nearly as high as 60 per cent, rose to 63 per cent by 1989.

The Myth of Low-Wage Jobs

An Editorial in Business Week (May 25) claimed that, “according to a just-released Census Bureau study, the number of working poor rose dramatically from 1979 to 1990.” This is completely false. In fact, the report shows that the percentage of low-income workers who are in poverty fell dramatically. Among husbands with such low-income jobs, for example, 35.7 per cent were members of poor families in 1979, but only 21.4 per cent in 1990.

Low incomes, in this report, were defined as “less than the poverty level for a four-person family” ($12,195 a year in 1990). Yet very few people with entry-level or part-time jobs are trying to support a family of four. Husbands now account for only a fifth of such low-income jobs, which are instead increasingly held by young singles and by dependent children living with their parents. Wives had 34 per cent of such jobs in 1979, but fewer than 28 per cent in 1990. That reflects the impressive fact that the median income of women rose by 31 per cent in real terms from 1979 to 1990.

It is true that the absolute number of low-income jobs increased in all categories, but that increase was not nearly as large as the increase in medium- and high-income jobs. All that the rise in low-income jobs really shows is that students living with their parents and young singles found it much easier to find acceptable work. The only reason fewer young people had low-income jobs back in the glorious Seventies is a large percentage of them had no jobs at all! Only 51.4 per cent of single males had full-time jobs in 1974, but 61.8 per cent did by 1989. Young people always start out with low earnings, if they get a chance to start out at all.

In his new book, Head to Head, Lester Thurow writes that “between 1973 and 1990, real hourly wages for non-supervisory workers . . . fell 12 per cent, and real weekly wages fell 18 per cent.” Yet these averages include part-time workers, which his why average wages appeared to be only $355 a week in 1991, even though half of all full-time workers (the median) earned more than $430 a week. Because many more students and young mothers were able to find part-time jobs in the Eighties, that diluted both the weekly and the hourly “average” wage. It most definitely did not mean that the wages of the “average worker” went down, but rather that otherwise unemployed part-time and entry-level workers were able to raise their wages above zero. The increase in part-time jobs also does not mean that families are poorer; rather, they are richer. Out of 19.3 million part-time workers in 1991, only 1.2 million were family heads, and only 10 per cent said they were unable to find full-time work.

The Rich Work Harder

Although the vast majority clearly had large income gains in the Eighties, Mr. Krugman and Miss Nasar nonetheless assert that those at the top had even larger gains, and that this is something that ought to provoke resentment or envy. Yet the figures they offer to make this point are grossly misleading. Moreover, the whole static routine of slicing up income into fifths is bound to show the highest percentage increases in average (mean) incomes among the “top” 20 per cent or 1 per cent. That is because for top groups alone, any and all increases in income are included in the average, rather than in movement to a higher group.

In his U.S. News article, Krugman first claimed that CBO figures show that “Ronald Reagan’s tax cuts” boosted after-tax income of the top 1 per cent “by a whopping 102 per cent.” That figure, though, is based on a “tax simulation model” which estimates “adjusted” incomes as a multiple of the poverty level. The top 1 per cent supposedly earned less than 22 times the poverty level in 1980, but 44 times the poverty level in 1989 – hence the gain of 102 per cent. Yet this is a purely relative measure of affluence, not an absolute gain in real income. As more and more families rose further and further above the unchanged “poverty line” in the Eighties, thus lifting the income needed to be in the “top 1 per cent,” the CBO technique had to show a “widening gap.”

Furthermore, the share of federal income tax paid by the top 1 per cent soared from 18.2 per cent in 1981 to 28 per cent in 1988, though is slipped to 25.4 per cent in 1990. Indeed, this unexpected revenue from the rich was used to double personal exemptions and triple the earned-income tax credit, which was of enormous benefit to the working poor.

By the time Mr. Krugman’s alleged 102 per cent gain at the top had reached the New York Times, it had shrunk to 60 per cent. However, the CBO wrote a memo disowning this estimate too, noting that “of the total rise in aggregate income . . . about one-fourth went to families in the top 1 per cent.” By fiddling with “adjusted” data, the CBO managed to get that share of the top 1 per cent up to one-third. Whether a fourth or a third, these estimates still begin with 1977, not 1980. Between 1977 and 1980, the CBO shows real incomes falling by 6.6 per cent for the poorest fifth. The top 5 per cent fared relatively well before 1980, because everybody else suffered an outright drop in real income.

Even if the Krugman-Nasar figures had been remotely accurate, the whole exercise is conceptually flawed. In every income group except the top, many families can move up from one group to another with little or no effect on the average income of those remaining in the lower group. Above-average increases in income among those in the lower groups simply move them into a higher fifth, rather than raising the average income of the fifth they used to be in. Only the top income groups have no ceiling, as those in such a group cannot possibly move into any higher group. A rap star’s first hit record may lift his income from the lowest fifth to the top 1 per cent, with no perceptible effect on the average income of the lowest fifth. But two hit records in the next year would raise the total amount of income counted in the top 1 per cent, and thus raise the average for that category.

Nobody knows exactly how much income is needed to be counted among the top 1 per cent, because the Census Bureau keeps track only of the top 5 per cent. Census officials argue that apparent changes in the small sample used to estimate a “top 1 per cent” may largely reflect differences in the degree of dishonest reporting. When marginal tax rates fell from 70 per cent to 28 per cent, for example, more people told the truth about what they earned, so “the rich” appeared to earn much more.

One thing we do know, though, is that the minimum amount of income needed to be included among the top 1 per cent has to have risen quite sharply since 1980, because of the huge increase in the percentage of families earning more than $50,000, or $100,000. This increased proportion of families with higher incomes pushed up the income ceilings on all middle and higher income groups, and thus raised the floor defining the highest income groups.

While $200,000 may have been enough to make the top 1 per cent in 1980, a family might need over $300,000 to be in that category a decade later. Clearly, any average of all the income above $300,000 is going to yield a much bigger number than an average of income above $200,000. The CBO thus estimates that average pre-tax income among the top 1 per cent rose from $343,610 in 1980 to $566,674 in 1992. But this 65 per cent increase in the average does not mean that those specific families that were in the top 1 per cent in 1980 typically experienced a 65 per cent increase in real income. It simply means that the standards for belonging to this exclusive club have gone way up. That is because millions more couples are earning higher incomes today than in 1980, not because only a tiny fraction are earning 65 per cent more.

Sylvia Nasar totally misreported the CBO’s complaints with her first article, and audaciously quoted her own discredited assertions in a later New York Times piece (April 21). This front-page editorial changed the subject- from income to wealth. It claimed a “Federal Reserve” study had found that the wealthiest 1 per cent had 37 per cent of all net worth in 1989, up from 31 per cent in 1983. Paul Krugman, writing in the Wall Street Journal, likewise cited this “careful study by the Federal Reserve.” Yet the cited figures are from a mere footnote in a rough “working paper” produced by one of hundreds of Fed economists Arthur Kennickell, along with a statistician from the IRS, Louise Woodburn. It comes with a clear warning that “opinions in this paper . . . in no way reflect the views of . . . the Federal Reserve System.”

At that, all of the gain of the top 1 per cent was supposedly at the expense of others within the top 10 per cent, not the middle class or poor. In any case, the figures are little more than a guess. The authors acknowledge that they “cannot offer a formal statistical test of the significance of the change.”

“The 1983 and 1989 sample designs and the weights developed are quite different,” they write. “The effect of this difference is unknown.” Their estimated range of error does not account for “error attributable to imputation or to other data problems.” Yet it is nonetheless within that range of error for the share of net worth held by the top 1 per cent to have risen imperceptibly, from 34.5 to 34.6 per cent. This is why Kennickell and Woodburn say their estimates merely “suggest that there may have been an increase in the share of wealth held by this top group in 1989.” Or maybe not.

The actual, official Federal Reserve study tells a quite different story. It shows that real net worth rose by 28 per cent among 40 per cent of families earning between $20,000 and $50,000, but by only 6.6 per cent for the top 20 per cent, earning more than $50,000. Since this huge increase in net worth among those with modest incomes means their assets grew much faster than their debts, this also puts to rest the myth that the Eighties was build upon “a mountain of debt.” It was, instead, built upon a mountain of assets, particularly small businesses.

Children without Fathers

What about the poor? There is no question that there has been a stubbornly large increase of people with very low incomes. However, annual “money income” turns out to be a surprisingly bad measure of ability to buy goods and services. In 1988, average consumer spending among the lowest fifth of the population was $10,893 a year – more than double their apparent income of $4,942. That huge gap occurs partly because annual incomes are highly variable in many occupations, and many people have temporary spells of low income, due to illness or job loss. People can and do draw upon savings during periods when their income dips below normal.

Another reason why those in the bottom fifth are able to spend twice their earnings is that many in-kind government transfers (such as food stamps) are not counted as “money income”. Census surveys also acknowledge that a fourth of the case income from welfare and pensions is unreported. And, of course, very little income from illegal activities is reported. In CBO figures, incomes of low-income families are further understated by counting singles as separate families, as though young people stopped getting checks from home the minute they get their first apartment.

Despite such flaws in measured income, nearly all of the income differences between the bottom fifth and the top fifth can nonetheless be explained by the number of people per family with full-time jobs, their age, and their schooling. Among household heads in the lowest fifth, for example, only 21 per cent worked full-time all year in 1990, and half had no job all year. In the top fifth, by contrast, the average number of full-time workers was more than two.

The May 25 Business Week editorial noted that “the percentage of Americans below the poverty line rose from 11.7 per cent in 1979 to 13.5 per cent in 1990.” Yet this poverty rate is exaggerated, because it is based on an obsolete consumer price index that mis-measured housing inflation before 1983. Using the corrected inflation measure, the poverty rate was 11.5 per cent in 1980 and 11.4 in 1989, before rising to 12.1 per cent in 1990. That 12.1 per cent figure, though, is only one of 14 different Census Bureau measures of poverty, and not the most credible. Like income for the “Bottom fifth,” the usual measure of poverty excludes many in-kind transfer payments, as well as cash from the earned-income tax credit. By instead including such benefits, and also subtracting taxes, the Census Bureau brings the actual poverty rate down to 9.5 per cent for 1990, or to 8.5 per cent if homeownership is considered (those who own homes need less cash because they don’t pay rent).

Even by the conventional measure, the poverty rate among married-couple families dropped slightly, from 5.2 per cent in 1980 to 4.9 per cent in 1990, and poverty rates among those above age 65 have fallen quite substantially. On the other hand, among female household heads with children under the age of 18 and “no husband present,” poverty rose from 37.1 per cent in 1979 to 39.9 per cent in 1980, and then to 41.6 per cent by 1990.

The poverty rate among fatherless families, then, is slightly higher now than it was in the previous decade, and is lower if these young women work. (Among female householders with children under the age of 6, the poverty rate among those with jobs dropped from 20.2 per cent in 1979 to 17.9 per cent in 1989, and the percentage of such mothers who worked full-time rose from 24.9 to 30.6 per cent.) But there are so many more female-headed households, and so few of these women work, that the net effect is nonetheless to keep the overall poverty rate from falling. The number of female-headed households with children under age 18 rose from 5.8 million in 1979 to 7.2 million in 1989. In too many cases, these mothers are so young that child-labor laws would not allow them to work in any case.

In March 1991, the average money income of female-headed families with children was only $17,500, and most of that money (plus food stamps, housing allowance, and Medicaid) came from taxpayers. For married couples who both worked full-time, average income was $55,700 before taxes – about enough to put the average two-earner family in the top fifth. Taxing hardworking two-earner families to subsidize broken, no-earner families can only discourage the former, encourage the latter, and thus exacerbate the problems it pretends to solve.

To summarize what actually happened in the 1980s, the “middle class,” and the vast majority by any measure, unquestionably experienced substantial gains in real income and wealth. With millions more families earning much higher incomes, it required much higher incomes to make it into the top 5 per cent or top 1 per cent, which largely accounts for the illusion that such “top” groups experienced disproportionate gains. The rising tide lifted at least 90 per cent of all boats. About 9 to 12 per cent continued to be poor, but this group increasingly consisted of female-headed households with young children. More and better jobs cannot help those who do not work, improved investment opportunities cannot help those who do not save, and increased incomes cannot help families whose fathers refuse to support their own children.

Mr. Reynolds is Director of Economic Research at the Hudson Institute in Indianapolis.