In responding to reviews and criticisms of his new e-book, Tyler Cowen digs in and claims that median income growth (or rather, the relative lack thereof) is the telling indicator of U.S. economic performance in recent decades. From a rhetorical perspective, his insistence is certainly understandable: the divergence between median income trends pre- and post-1973 is indeed dramatic. But from an analytical perspective, does his focus really make sense?
Perhaps, if the idea is merely to substantiate the claim that living standards for ordinary Americans are improving less rapidly than they did in the past. Now while it's possible to debate this claim, I'm not interested in doing so at present. What I care about now is scrutinizing Tyler's argument that this slowdown in the growth of living standards is due to a general slowdown in innovation -- to the arrival, in his words, at a "technological plateau." And on this important issue, I believe Tyler's insistence on looking at median income -- and dismissing the greater relevance of GDP per capita and productivity -- is misguided.
Here is the issue. Median income growth since 1973 is worse than GDP per capita growth over the same period, which in turn is worse than productivity growth (at least since the mid-'90s). The productivity stat's directly contradict the alleged existence of a technological plateau, and Tyler will need to do a much more convincing job of picking apart those productivity numbers if he wants his argument to fly. Meanwhile, the disparity between median income and GDP per capita points to a rival explanation of sagging median income growth. Maybe we're living through a cultural lag, not a technological plateau.
The reason median income hasn't kept up with GDP per capita is that income gains since 1973 have accrued disproportionately to high earners. In other words, we've experienced a big increase in income inequality. And why is that? If we're focused on the growing gap between incomes at, say, the 90th percentile and the 50th percentile (as opposed to the stratospheric gains at the very tippy top of the income scale), a major reason is the increased return to skills due to sluggish human capital development. The continuing increased demand for people with college degrees (or the skills for which a degree serves as a useful proxy) has not been matched by corresponding growth in people with those degrees and skills, and thus the wage premium for college grads opened up after 1980 and has persisted stubbornly since then. As Claudia Goldin and Lawrence Katz put it, in the race between education and technology, technology has been winning handily.
And this really shouldn't be that surprising. Once we exhausted the supply of "smart but uneducated kids" that Tyler writes about, further gains in educational attainment and human capital development would have to overcome deep-seated, class-based differences in values and childraising philosophy. Since the big economic payoff (in the form of the college wage premium) of overcoming this class divide only dates back about 30 years, or a single generation, it's no wonder that these differences are still with us.
Now this cultural lag is a formidable obstacle to future growth prospects, but it's a very different one than Tyler is highlighting. It's not that we're running out of good ideas; it's that we're running out of people who can use good ideas to best advantage.