Russ Roberts vs. the Great Whatever
By Arnold Kling
So my challenge to Tyler is to tell me what he thinks the stagnation in median income signifies. Has there been a change in the returns to education or creativity? Or is it mostly a statistical artifact? Whichever answer he gives, I would like to see him reconcile it with the panel data-the surveys of economic information that follow the same people over time.
Just about every journalist, and even most economists, talk about cross-sectional slices of the income distribution at different points in time as if they were trends among cohorts. Thus, if you say that “real incomes of workers with only a bachelor’s degree declined over the past ten years,” that sounds as if a particular cohort of college graduates experienced a decline in real income. In fact, what it says is that in a cross-section of workers in 2010, those with a college education have lower average real incomes than those in a cross-section in 2000.
In theory,the following could have happened:
1) The real income of every single college-educated worker went up between 2000 and 2010.
2) However, many high-income college-educated workers aged out of the labor force over that period and many moderate-income college-educated workers joined the labor force.
If the effect of (2) overwhelms the effect of (1), the data would show that “the real incomes of college-educated workers declined.” It could turn out that, when all is said and done, the current cohort of college-educated workers will have lifetime incomes that far exceed those of those who retired over the past ten years.
The longer the time interval that you are comparing, the more serious this issue becomes. I do not think it distorts comparisons of 2010 with 2000 by very much.* But I think comparing 1973 to 2010 using cross-sectional slices is really hazardous.
(*You can argue that 2000 is a cyclical peak and 2010 is close to a cyclical trough, and that is a problem. That is, you can argue that. Not me. The PSST perspective is that the economic restructuring is the story of the decade, not some temporary shortfall of aggregate demand.)
My list of data points for the Great Factor-Price Equalization would include:
–rising incomes in developing countries, even in Africa.
–a dramatic slowdown in the earnings gains of all but the very-highly-educated in the U.S. I believe this data point can survive the not-a-cohort critique given above.
–polarization of incomes occurring in other OECD countries, not just the U.S.
–manufacturing output rising in the U.S. while manufacturing employment is flat
–anecdotal evidence of the pattern of creative destruction: Borders bookstores closing while sales of books on e-readers soar; video streaming displacing physical movie rentals; etc.
Where I disagree with Tyler is on the “stagnation” interpretation. I agree that as a society, we are throwing away a lot of money in our health care, education, and homeland security sectors, and we are probably not getting much return on that. But in spite of all that waste, I think there is still considerable technological progress. I am inclined to agree with Russ that the plight of the not-highly skilled worker in the U.S. these days is a symptom of the rapidity of technological progress, not a slowdown.