Greg Mankiw Forgets to Offer Data for his Biggest Claim

Greg Mankiw’s paper in the Journal of Economic Perspectives’ symposium on the top one percent is generating plenty of commentary. Josh Bivens and I have a contribution in that symposium and some new evidence that casts doubt on one of Mankiw’s key claims: that the doubling of the income share of the top one percent reflects the increased economic contributions, or productivity, of those in the top one percent. Specifically, Mankiw claims “that changes in technology have allowed a small number of highly educated and exceptionally talented individuals to command superstar incomes in ways that were not possible a generation ago.” Mankiw’s evidence for this is pretty thin, and we offer contrary evidence. This will take two blog posts, this one addressing the correspondence of growing educational wage disparities and the rise of the top one percent and the next one focused on executive pay.

Before getting down to business, let’s dispose of the distracting discussion by Mankiw and others (e.g. Chrystia Freeland) that we are discussing the incomes of superstar ‘innovators’ like Steve Jobs or J.K. Rowling. The majority of those in the top one percent are financial sector professionals and executives, not ‘innovators.’ Moreover, it is the growth of financial sector and executive incomes, as Jon Bakija and co-authors document, that explains roughly two-thirds of the income growth of the top 1.0 or top 0.1 percent (an analysis we argue understates the role of executives and finance). Besides, as Dean Baker points out, even superstar innovators benefit from a government set system that skews rewards upwards.

So, what evidence does Mankiw offer? First, he rests on the work of Claudia Goldin and Lawrence Katz (2008) book, The Race between Education and Technology, that skill biased technological change continually increases the demand for skilled labor” and concludes that “the story of rising inequality, therefore, is not primarily about politics and rent-seeking but rather about supply and demand.” Mankiw acknowledges, however, that “Goldin and Katz focus their work on the broad changes in inequality, not on the incomes of the top 1 percent in particular. But it is natural to suspect that similar forces are at work.” Mankiw then argues that the growth of top one percent income shares and the premium earned by skilled relative to unskilled workers that Goldin and Katz study “follow a similar U-shaped pattern.”

That’s not much evidence and, in any case, it’s not even true: the growth of top one percent incomes has not followed the same pattern as the wage premium of ‘skilled’ workers (meaning the wage premium earned by college graduates). Look at the figure below, which shows the regression-adjusted (log) wage premium of college graduates (including those with advanced degrees) relative to non-college educated workers and the ratio of top one percent incomes (including capital gains) to the incomes of the bottom ninety percent. These are shown as an index, with 1979 set to be zero, because the data are on far different scales. The first thing to stand out is that the top one percent income advantage moves like the stock market, much like the pay of executives and financial professionals, rising rapidly in the late 1990s, crashing after the tech bubble burst and again in 2008 and recovering as the stock market does so. The graph doesn’t show the 1970s, but as Goldin and Katz (and, obviously, Richard Freeman) have shown, the college wage premium fell in the 1970s. The top one percent share did not fall in the 1970s. These are different phenomena. I have had many conversations about this with Larry Katz (who, as I mentioned, Mankiw references and who also happens to be Mankiw’s colleague at Harvard). Katz agrees that the top one percent and the bottom 99 percent dynamics are driven by different factors. Greg Mankiw should walk down the hall and talk to him about this!

It is also the case that labor economists (including Goldin and Katz) have long noted that the college wage premium flattened out in the mid-1990s: it grew nineteen points from 1979 to 1995, grew another four points by 2007 and grew an additional half a percentage point by 2011. That is, the college wage premium grew roughly a fourth as much in the last fifteen years as it had in the prior fifteen years. Did the top one percent income advantage follow the same pattern? Absolutely not! The top one percent income advantage grew as much between 1995 and 2011 (despite the financial crash) as it did between 1979 and 1995. The growth of the top one percent income advantage grew far more between 1995 and 2007, before the crash, than it did in the earlier period. That is probably a stronger suggestion of a different pattern. CEO compensation grew a lot in 2012, not surprisingly, along with a growing stock market, so I suspect the top one percent is on its way to reestablishing its 2007 income advantage. There is no reason to think the income gains of the top one percent reflect the same forces as those driving education wage differentials.

Mankiw, by the way, has at best a partial story since he does not address some of the key dynamics of the growth of the top one percent income shares: the growth is not just due to the greater concentration of wage income at the top but also reflects the greater concentration of capital income (capital gains, business income, interest, dividends and so on) at the top (receiving 56 percent of all capital income in 2007, up from a 32 percent share in 1979). Moreover, about a third of the growth of the top one percent income share is due to the increase of capital income and the corresponding erosion of wage income. There is a capital versus labor dimension to all this. Again, this is all presented in Josh Bivens’ and my paper, The Pay of Corporate Executives and Financial Professionals as Evidence of Rents in Top 1 Percent Incomes.

Mankiw later notes that “The key issue is the extent to which the high incomes of the top 1 percent reflect high productivity rather than some market imperfection” and turns to a discussion of CEO pay. That is the topic of the forthcoming blog post, where I will draw on our paper to show that CEO pay has grown far faster than that of other high wage earners (the top 1/1000th) and that the CEO advantage relative to other very high wage earners has grown more than that of the college wage premium. That’s pretty substantial, and runs counter to the claim that CEO pay simply reflects the market for talent. More later.