Thursday, December 15, 2011

Caplan on Wages and Labor

Bryan Caplan writes a post that I have a lot of issues with. First and foremost is this statement that he leads with: "Unemployment is just a labor surplus; since wages are the price of labor, the fundamental cause of unemployment has to be excessive wages."

This is how a lot of people think, and I often feel like I'm in the small minority on this - but I am more confident that I am right on this than I am on almost anything else I write about on here. Unemployment is not "just" a labor surplus (although obviously if there are labor surpluses, that can contribute to unemployment). The labor market could be in equilibrium and we could still have a lot of unemployment.

8.6% of the United States labor force is currently classified as "unemployed". That's 13.3 million Americans. Several million more aren't in this count but matter to us as "discouraged workers" or "underemployed" workers. To be in a labor surplus, you must (1.) have a reservation wage that is equal to or lower than the market wage, and you (2.) have to be jobless. Only the second criterion is involved in your classification as being "unemployed" - the phenomenon we allegedly care about and the data which we look at. We don't know how many people satisfy (1.) and (2.) - we don't know how many people comprise a labor surplus in this country. Nobody seems to have cared enough to even think about how to collect that data. We don't know if that number is cyclical or not. We don't even know if our unemployed uncle or unemployed friend is in that labor surplus or not. And we don't seem to care about that. The social fact we seem to care about is unemployment, not labor surplus. To be unemployed, it's perfectly possible to have a reservation wage above the market wage (and it's not even clear exactly what this means, since one person's labor can be sold in multiple labor markets). Nobody associated with the Current Population Survey will ever issue a survey to you asking you what your reservation wage is. It has nothing whatsoever to do with the phenomenon of unemployment. Unemployment is not the same as labor surplus.

Economists acting as economists care about market efficiency. Economists as human beings care about much more than that, and on the question of unemployment the economists' obsession with market efficiency has been very damaging - we often try to fit a square peg into a round hole. If you want to explain unemployment you can't just copy and paste other concepts like a labor surplus and pretend you've explained it. The best bet we have for explaining unemployment is explaining something that we have a better sense of: employment. We know the determinants of employment. Unemployment seems to increase when employment decreases (this sounds tautological but it's actually not at all). We also need to explain employment in a general equilibrium rather than a partial equilibrium framework. Finally, we need to explain why people might be jobless and unemployed vs. jobless and not unemployed (search theory helps with that). To the extent that we care about discouraged workers and those sorts of people, search theory becomes less essential for answering the questions we care about.

I know that only addresses Caplan's first paragraph - but take my argument seriously, think about it, then read the rest of Caplan's post and see what makes sense and what doesn't.

UPDATE: I also think it's worth clarifying something about "sticky wages". First, the observation of sticky wages doesn't necessarily imply a labor surplus at all (although certainly a wage floor would be great example of sticky wages!). An elastic labor supply curve to the left of equilibrium will give you sticky wages in the data. Why might a labor supply curve get elastic? People with jobs create lives with costs that they try to make predictable. Anyway - it's just important to remember that "sticky wages" just means "wages move less than employment". It doesn't have to mean "wage floor" or any sort of surplus. Second - it's important to know one of hte major reasons why economists talk about sticky wages that has little to do with any of this: sticky wages help explain why the long run aggregate supply curve is vertical but the short run aggregate supply curve is upward sloping. It's probably unfair to say that's the only reason why economists care about sticky wages, but it is a major reason why.


  1. One nitpick I have is that wage rigidity doesn't actually have to happen in the economics of Keynes or Keynesian economics.

    I otherwise would agree with your argument, Daniel Kuehn.

    But I would also add that the high level of unemployment in the United States and other Western countries still reeling from the global financial crisis is simply a low "weight of evidence". What do you make of Keynes's concept of the weight of evidence, Daniel?

  2. This comment has been removed by the author.

  3. Cutting wages would obviously impact AD...if excessive real wages were really the cause of unemployment then why did the period of growing real wages coincide with 1-2% unemployment and the period when they stagnated coincide with much higher unemployment?

    Also, note that Caplan is happy to defend CEO pay as not 'excessive' and therefore causing unemployment.


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