Perennial Sowell Fallacies

By Tom Lowe

February 21, 2000

I catch much flak from friends and acquaintances because of my harsh comments directed towards Thomas Sowell and his writings, to which my reply is simply that Sowell, being in a position of prestige and power, is hardly in danger of being thrown into the street as a result of anything I may say or write, but, on the other hand, is in dire need of refutation for the distortions and untruths which he regularly purveys in his writings. His recent column "Perennial economic fallacies," http://www.jewishworldreview.com/cols/sowell020700.asp, is a good case in point.

As economists have been pointing out for some time, inequality in America has been increasing. The difference in incomes between the top and bottom have been increasing since Congress, under the influence of Ronald Reagan, enacted the large tax cuts benefiting large corporations and wealthy individuals and relaxed antitrust and securities regulation. The top 1% of wealth-holders, for instance, now owns more than 40% of the wealth of the entire country. The people in the bottom quintile (20%) of income have lost ground during one of the longest economic booms in our history. There are numerous reasons advanced for this increasing gap, including corporate downsizing, the flight of manufacturing overseas, union-busting by the Reagan and Bush administrations, and a bull market that increased the holdings and income of stockholders while leaving the earnings of non-stockholders behind. It is likely that all of the contributed to the increasing inequality.

The conservative response to this has been denial, then obfuscation. Sowell's column is a textbook example of that process. Let's look at his points. Here's what he claims:

  1. The poor do not remain poor. Mobility is so high that an "absolute majority of the people who were in the bottom 20 percent in 1975 have also been in the top 20 percent at some time since then."
  2. Real income per capita has risen 50 percent over the same span of time when household income has remained virtually unchanged because households are getting smaller.
  3. Actual examinations of flesh and blood people show no nutritional differences between people in different income brackets. Americans in the lower income brackets today are slightly more likely to be overweight than is the rest of the population.

 

The mobility argument

Sowell's argument rests mainly upon the mobility argument, that people do not stay poor long, but move up the wealth ladder over time. Furthermore, the wealthy often move down the wealth ladder. He bases this argument on the Annual Report of the Federal Reserve Bank of Dallas in 1995, which also found that only 1% of the American population remain permanently in the bottom 20% of income earners. Needless to say, if this is true, it is good news, indeed.

Unfortunately, that is not the case.

One of the blessings of the Internet is the ability to check sources, and Sowell usually provides enough clues in his columns to trace the origin of his figures. In this case, the document is "By Our Own Bootstraps: Economic Opportunity & the Dynamics of Income Distribution," http://www.dallasfed.org/fed/annual/1999p/ar95.html, a beautifully-produced, lavishly illustrated, paper by W. Michael Cox and Richard Alm, purporting to show that the economic mobility of the American population completely mitigates the growing economic inequality. There is some merit in this position. Persons usually earn less when they enter the job market than they do later. A person's retirement income will usually be lower than what they earned prior to retirement. It therefore stands to reason that earnings over a lifetime probably would be a better indicator of economic well-being than a statistical slice of the population at any given time. In many ways, it's like the difference between an earnings statement, which covers a period of time, and a balance sheet, which is a snapshot of the assets and liabilities of a firm at a point in time.

Cox and Alm, having made this point, then run into a problem: there isn't much data available on how people fare over time. They found two sources: The first was the University of Michigan's Panel Survey on Income Dynamics, http://psidonline.isr.umich.edu/, a tracking study from 1968 to the time of the study (1995). Because of difficulties in the collection of data, Cox and Alm dwell upon the period from 1975 to 1991, covering 3,725 individuals. The second source is a 10-year study by the U. S. Treasury of persons who had filed tax returns for 10 consecutive years.

The use and misuse of statistics is often a matter of common sense. In studies that track people over a period of time, called "longitudinal" studies, it is essential that the sample of persons studied be representative of the population to which one intends to generalize one's findings. In many statistical studies, 3,725 persons is more than adequate to represent millions of persons, but only if the members of the sample are chosen properly. A sample consisting of only New Yorkers, or professional football players, or prisoners, to take several obvious examples, would clearly not be representative of the American population. More to the point, by selecting a less-representative starting sample with a particularly low income, percentage increases in income will be overstated at the end of the period.

The Cox and Alm paper flunks the statistics test with respect to the University of Michigan study. Exhibit 5 on page 8 of their report tells the tale; it shows the average annual income of the lowest quintile to be only $1,153 in 1993 dollars, which is indicative of the kinds of workers who comprise this group, undoubtedly among them 16-year-olds with paper routes, otherwise successful self-employed business people having a bad year and married women working just a few weeks in a year. This "low-income" category consisted of anyone aged 16 or older whose earnings ranked in the lowest quintile of the distribution of earned income in 1975. " And the fact that their average earnings increase 22-fold over the next 16 years has no obvious implications for most of America's working poor today." See Duncan, Boisjoly and Smeeding, "How Long Does It Take a Young Worker to Support a Family?" http://www.northwestern.edu/ipr/publications/nupr/nuprv01n1/duncan.html. Further, Duncan, Boisjoly and Smeeding add;

One way of reframing the Cox-Alm study is to estimate where the individuals whose 1975 family incomes placed them in the bottom 20% of the family-income distribution ended up after 16 years. The answer is a half-full/half-empty one, common in mobility research: nearly half (47%) of the poor were still at the bottom in 1991, but 6% did make it into the top quintile and one in five made it into the top half of the income distribution.

Cox and Alm's paper was highly misleading because of they way they selected their baseline income.

The 1992 U. S. Treasury study, cited on page 12 of the report is equally suspect. The study found that 86 percent of those in the lowest income bracket moved to a higher grouping. Further two-thirds of them reached the middle strata or above with almost 15 percent making it to the top fifth of income earners.

Paul Krugman, fully cited below, http://www.prospect.org/web/page.ww?section=root&name=ViewPrint&articleId=5232, easily disposes of this:

But this report was based on what we may charitably call a strange procedure. Here's what Hubbard's report did: it tracked a group of individuals who paid income taxes in all ten years from 1979 to 1988, and compared their incomes not with each other but with those of the population at large. The restriction to individuals who paid taxes in all years immediately introduced a strong bias toward including only the economically successful; only about half of families paid income taxes in all ten years. This bias toward the successful was apparent in the fact that by the end of the sample period the group contained very few poor people and a lot of affluent ones: indeed, only 7 percent of the sample were in the bottom quintile by the sample's end, while 28 percent were in the top quintile. More important, by comparing the sample with the population at large rather than with each other, the report essentially treated the normal tendency of earnings to rise with age as representing social mobility. The median age of those whom the study classified as being in the bottom quintile in 1979 was only twenty-two.

Thus, we can dismiss Cox and Alm's paper as virtually worthless as a source to prove the high level of mobility Sowell claims to exist in the U.S.

Comments

Mobility has not changed significantly over the last 25 years and is no higher in the U.S. than in other industrial countries. See Sawhill and McMurrer, "How Much Do Americans Move Up and Down the Economic Ladder?", http://www.urban.org/oppor/opp_031.htm. When increasing inequality is unaccompanied by a compensatory increase in mobility, it follows that society is pulling apart between the haves and have-nots.

The relationship between inequality and mobility involves societal choices as well as individual ones. We Americans have historically been willing to tolerate considerable inequality as long as there is a fair opportunity for individuals to move up the economic ladder The real question is how to get a satisfactory balance of the two.

I want a chance to grab the brass ring, but if I fail, I do not want to fall into quicksand.

The class structure of our society also limits mobility, although it is one of those things that are not talked about. Nevertheless, the society in which our parents move, with its manners, customs, neighborhoods, social organizations and even watering holes, is highly determinative of our future position in life. The higher up the social ladder, the more social practices and customs tend to exclude the uninitiated. Persons at the bottom of the economic and social ladder want as much mobility as possible, for that is their only hope for a better life. The wealthy understandably do not desire mobility, either for themselves or for others further down the ladder, for they have nowhere to go but down.

 

Real Per Capita Income and the Size of Households

Sowell's next argument is equally vaporous. He states "If you are serious about considering the well-being of flesh and blood human beings, then you can talk about their real income per capita. But alarmists avoid that like the plague, because it would expose their little game for the fraud that it is."

Why is Sowell always demonizing categories of people - liberals, alarmists, plaintiffs, activists, anyone with whom he disagrees? Hardly a column goes by without vitriol being cast at someone. Surely one can advocate without vilifying one's opponents. Persons concerned about the increasing gap in income and wealth are hardly alarmists; they are simply concerned about a nation becoming less and less just, with a wealthy upper-crust and no middle class. And to call a legitimate concern "fraud" is deceitful in itself.

Per capita income equals the sum total of all incomes divided by the population. In other words, it is an average. Sowell correctly states that, since the bottom 20 percent of households contains 39 million people, while the top 20 percent contains 64 million, that comparing households is like comparing apples and oranges. But one cannot logically conclude from this that per capita income is a reliable indicator of economic well-being. A rising per capita income can conceal an absolute decline in the income of one segment of the population if it is counterbalanced by disproportionate increases in another segment.

Consider the following highly simplified hypothetical example of three people, A, B, and C, on an island:

Assume that A, B and C each start out earning $100. Per capita income can easily be calculated thus:

($100+100+100)/3 = $300/3 = $100

which intuitively makes sense; if everybody makes the same amount, then each person makes the average.

Now assume that A's income drops to $50 and C's income increases by $50. The reader can verify that adding $50, 100 and 150 and dividing by three results in the per capita income remaining the same, even though A has seen his income drop by 50% and C has seen his income rise by 50%.

Suppose that, starting at the original position A's income drops to $50, B's goes up to $110 and C's to $160. Per capita income will then change to

($50+110+160)/3 = $320/3 = $106.67,

an increase of 6.67%. Even though per capita income rises, A loses ground.

Thus per capita income cannot reflect the distribution of incomes or changes in individual incomes but only the overall change. A's life may indeed go from comfortable to destitute as his income is cut in half, while B modestly improves his position by 20% and C can live in luxury with a 60% more income.

Sowell's rhetoric is designed to conceal the real truth about incomes in America: that the bottom fifth of the nation has seen a drop in constant-dollar income (constant-dollar income is income corrected for the decrease in purchasing power of the dollar over time) since 1980, the middle class's income has remained stagnant and the wealthy have greatly increased their wealth and income. In the last couple of years, the incomes of the lower and middle sectors of the population have begun to see improvement, but nothing like the gains of the top 10 percent and especially the top 1%. In spite of gains in productivity approximating 1%/year, workers whose productivity has increased have not seen a proportionate increase in their income. It has been appropriated by the upper classes and mostly the top 1%. See Paul Krugman, "The Rich, the Right and the Facts: Deconstructing the Income Distribution Debate," The American Prospect No. 11, Fall 1992, http://www.prospect.org/web/page.ww?section=root&name=ViewPrint&articleId=5232.

The Results of Poverty: The Poor are Fatter so they must be OK

Finally, we come to the low point of Sowell's column:

Despite desperate effort of activists to keep "hunger in America" alive as an issue by manipulating numbers, actual examinations of flesh and blood people show no nutritional differences between people in different income brackets. In contrast to the gaunt and undernourished poor of other times and places, Americans in the lower income brackets today are slightly more likely to be overweight than is the rest of the population.

There's no doubt that if you eat junk food, lard, and other staples of the poor, you are likely to weigh more. But Sowell is ignoring the overwhelming statistical evidence that high poverty levels in a state are highly correlated with significant social ills, such as infant mortality, teen pregnancy, teen violent deaths and low birth weight babies. See 1999 Kids Count Data Online http://www.aecf.org/kidscount, published by the Annie E. Casey Foundation. Using a spreadsheet, it is easy to calculate the correlation between the percent of children in poverty and the social ills set out above. The results are sobering and conclusive. Poverty, as it is measured today, causes bad things to happen. It is not a mere abstract construct, invented by "bleeding-heart" liberals to justify increased welfare expenditures. More babies die, more teens get killed and more teenagers get pregnant in states with high poverty levels. These results are consistent with experience and common sense. And these kids, contrary to the Panglossian assertions of Sowell and the Federal Reserve Bank of Dallas, don't move up, at least not in significant numbers.

Perhaps Sowell has some other explanation for these high correlations, but I doubt it. Otherwise, he would not have selected weight alone as a factor to prove that the poor in this country fare well.


Copyright 2000, Thomas Lowe. All rights reserved. Published in The Jackson Progressive, http://www.jacksonprogressive.com. Noncommercial reproduction of this article in its entirety is authorized, provided that this notice accompanies any reproduction.