This column by ACRU General Counsel and Senior Fellow for the Carleson Center for Public Policy (CCPP) Peter Ferrara was published August 31, 2011 on Forbes.com.
Last week we discussed the broad prosperity throughout society produced by Reaganomics and the resulting 25 year economic boom that started in 1982. We showed how that has been obscured by changing demographics and cultural factors, and how the resulting pattern of incomes closely reflected productivity and the productive performance of different workers.
For many people, this is essentially a religious issue. They are devoted to what they want to believe based on faith and emotion, and don’t want to be confused by facts or logic. For these readers, I am going to continue to challenge your faith.
Again, the best work on this issue is the brilliant 2006 book by Alan Reynolds, Income and Wealth. I reprise and add to this analysis in my own recent book, America’s Ticking Bankruptcy Bomb.
Further fundamental errors in understanding the equality of Reaganomics arise from a failure to distinguish between more Americans getting richer, and only the rich getting richer. For example, major media outlets decried during the Reagan boom era a supposed declining percentage of Americans in some fixed, moderate income range, such as between $35,000 and $50,000 per year, which they took as indicating a shrinking or vanishing middle class.
But during periods of sustained economic growth, such as during the 25-year Reagan economic boom, growing incomes would naturally take a higher percentage of Americans past any fixed, moderate income range. Reynolds notes that the data published in a high profile Washington Post feature raising such an alarm refuted the significance of the claim. It showed, in fact, that the percentage of households with real incomes higher than $50,000 increased from 24.9% in 1967 to 44.1% in 2003, while the percentage with real incomes lower than $35,000 fell from 52.8% in 1967 to 40.9% in 2003. This represents the middle class getting richer, with incomes growing beyond the fixed income range, rather than a shrinking middle class.
Wall Street Journal senior economics writer Steve Moore made the same point, demonstrating the soaring prosperity of middle America during this period in writing, “in 1967 only one in 25 families earned an income of $100,000 or more in real terms, whereas now [in 2004], one in six do. The percentage of families that have an income of more than $75,000 a year has tripled from 9% to 27%.” An exasperated Reynolds rightly laments the statistical confusion reflected by major media institutions and even researchers who should know better, writing:
Such complaints — deploring the rapidly rising percentage of families earning more than $50,000-$100,000 per year — are persistent yet inexplicable. A larger percentage of Americans earning more income is apparently being confused with a quite different concept of ‘inequality’ — namely, the same number or percentage of people earning more money….Yet there is no sense in which a larger percentage of workers earning higher incomes can possibly have a harmful effect on other people with ‘less-skilled’ jobs.
If class warriors cannot get this simple point, they will never understand more complex statistical errors, which Reynolds has been an intellectual leader in exposing. Any top income group, whether you are measuring the top 1%, 5%, 10% or 20%, by definition has no income ceiling like the lower income groups. For example, the middle 20% will always be between some $X thousand per year and some $Y thousand. But the top 20% will include all incomes above the 80th percentile, up to the highest earning individual billionaires in the country.
The implication of the above is that any average or mean of incomes in the top 20% will always be much higher than the median of incomes in the top 20%, which is defined as the income level with half of those in this fifth above, and half below. The median will consequently always provide a much more accurate reflection of the typical income earner in any top income group than any average or mean. Discussions focusing on the change in “average” incomes of the top 1%, or top 20%, or top anything in between, are, therefore, always misleading. The reality in any top income group is always better reflected by focusing on changes in the median income of the group. Reynolds provides this example, “Mean income for the top 10 percent is about two-thirds larger than median income, showing how mean averages greatly exaggerate the level of typical incomes of top income groups.”
This becomes especially misleading when comparing the top income group to any lower income group, or to changes in their relative income shares over time, when those income shares are measured by average incomes rather than median incomes. That is because in any lower income group, with both an upper and lower bound to the incomes included, the average or mean income will always be much closer to the median. So in focusing on averages within each income group, the realities of lower income groups will be compared to a misconception of the top income group.
Reynolds illustrates this by examining Federal Reserve data regarding incomes of different subgroups of American workers. In that data, the average or mean income of the top 10% of households seems to increase much more from 1989 to 2004 than the average or mean income of the next highest 10%, or of any lower income group. That would seem to imply growing inequality, with the rich getting richer faster than any other group is improving. But when the more accurate median income is considered, the income of the top 10% grew at virtually the same rate from 1989 to 2004 as the bottom 20%, and as the second lowest 20%. That indicates widespread and relatively even progress among all income groups over the period.
A similar statistical problem arises from income changes affecting the bottom limit, or threshold, of any top income group. The statistical illusion here is that rapidly rising incomes among households with incomes below the top income group can appear to increase the incomes of the top group. For example, as the incomes of those in the second 10% of income earners grow into what was formerly the top 10%, the bottom limit, or threshold, defining the top 10% must increase, because only 10% of income earners can fit into the top 10%. As the threshold increases, the average of incomes above the higher threshold will inevitably be higher. That will appear to be an increase in incomes of the top 10%. But it is an effect of the increase in incomes of those below the threshold. Any increase in incomes of the original top 10% would be in addition to this statistical effect.
Reynolds illustrates the problem with this specific example. The top fifth of household incomes began at $68,352 in 1980 (in 2004 dollars). By 2004, the incomes of so many in the second 20% had increased above the former $68,352 threshold that the top 20% of income earners now started at $88,029. As Reynolds explains, “If you calculate a mean average of all the income above $88,029 in 2004, you are bound to come up with a larger figure than if you averaged all the incomes above $68,352 (as we did in 1980). The average in 2004 excluded incomes between $68,352 and $88,029–incomes that were included in the average in 1980.”
In this case, the average of the top 10% is being “pushed up from below by rising numbers of people moving up–leaving what used to be considered a ‘middle class’ income and ‘joining the ranks of the rich.'” Or as Reynolds has more famously said, “so many people have been getting rich there isn’t room for them all in the top fifth.” The essential point, to paraphrase
Reynolds, is this statistical effect does not mean the rich are getting richer, it means more people are getting rich, which is a reflection of rising general prosperity, rather than rising incomes at the top.
Confusion arising from this statistical illusion led to media reports insisting that the great majority of income growth, if not all, over substantial periods, such as 1970 to 2000, had occurred only among the top income group. These reports claimed that as a result incomes of the bottom 80% had not increased at all over that period, or even had fallen. Reynolds cites New York Times writers David Cay Johnston and Paul Krugman as making such claims. But as Reynolds notes, Census Bureau data showed not only real income gains among every income group over the period, but accelerating gains, with real incomes among the lower income groups increasing even faster in the later years. CBO data showed as well real income gains for every income group over the period.
The resulting misconceptions among the public are affecting public attitudes and public policy to this day. The result is bad policy that is badly harming current economic growth and prosperity, and their future prospects.