Wednesday, August 1, 2012

Wage stagnation is not a simple story

Noah Smith blogged on it the other day, pointing to labor competition from globalization and the Bretton Woods break down, as well as the oil crisis.

Steve Landsburg (HT - David Henderson) also recently pointed out the impact compositional change in the labor force had on median wages. Henderson writes: "between 1980 and 2005, median incomes for non-white women, white women, non-white men, and white men all rose substantially, by 62%, 75%, 16%, and 15% respectively (all in real terms.)", while median real wage growth for all workers was only 3%. The compositional shift towards lower wage workers drove down median wages even thought everyone's wage was growing faster than 3%.

One of the things I pointed out at Henderson's blog is that although Steve makes a good case for the impact of compositional effects, something still looks very wrong. A 15% annual growth in white male wages, for example, translates to a half a percent annual growth. Does anyone think that between 1980 and 2005 white male productivity was growing only half a percent a year? I don't.

So Landsburg and Henderson are right, I think. Compositional effects matter. But something is definitely going on here. Wage stagnation is not a non-story, it's just a complicated story.

Mark Thoma shares further evidence for this view from Lane Kenworthy. Lane points out that the compositional shift has been going on since the early 1950s - when real wages were increasing at a much faster pace (and wage differentials between white males and disadvantaged groups were even higher). It's hard to ascribe the post-1980 or post-1970 stagnation (however you want to date this) to compositional effects if the compositional effects were going on even before the stagnation. We can say compositional change makes median wage growth look lower than it "actually is", as experienced by workers. But I still get the impression it's only a part of the wage stagnation story.


  1. What you and Noah write still isn't convincing until you take into account the other factors. What about non-pecuniary income? BLS statistics don't always take that into account, and when they do it's often in odd ways. What about family size changes?

    Lastly -although I grant this is very difficult to measure and I'd accept analysis without it- what about the behaviour of price indices for the kind of goods low income earners buy?

  2. Another important issue is finding a way to take into account immigration from Mexico. Uneducated immigrants can't be expected to have the same opportunities as others in any economic system.


All anonymous comments will be deleted. Consistent pseudonyms are fine.