Sticky wages and unemployment

Yesterday, in response to Tyler Cowen's post, I reiterated that sticky wages don't strike me as the major thing we should be worrying about, and that my theory on why people get so caught up on sticky wages is that we've come to equate labor surpluses with unemployment, which is wrong. Since then a lot more people have jumped into the fray. I haven't even had time to digest them all, but I've seen:

- Arnold Klng: "I think it also makes for a good story for how state and local government employment declines in a recession. I have consistently prescribed wage cuts for state and local government workers as a way of maintaining employment there... This is the biggest problem with trying to fit the current situation into the AS-AD paradigm."

- Brad DeLong: "The old Keynesian line was that nominal wage flexibility--and the union-smashing recommended by Hayekians--was a side issue. In an economy with nominal debt contracts downward-flexible nominal wages were likely to produce deeper depressions as the economy was subjected to much stronger downward shocks from the deflation, debt, and bankruptcy cattle prod. Wage inertia was thus a blessing--albeit a poorly-understood blessing--rather than a curse.

I think that everybody open-minded and nuanced is finding themselves moving rapidly toward the old Keynesian position under the pressure of events and data right now." [He also does a good job showing how Cowen's chief empirical example commits the Mulligan fallacy]"

- Robert Waldman

- Scott Sumner

- Jonathan Catalan

- Alex Tabbarok

*****

Two of the more interesting posts were from Karl Smith, who downplays sticky wages and notes real wage pro-cyclicality and then a response post to him by Scott Sumner which says that real wages are pro-cyclical during supply-shock recessions and counter-cyclical during demand-shock recessions.

Smith's point, I think, is very important. The research he is refering to on pro-cyclical wages is summarized by Abraham and Haltiwanger (1995). They note that real wage cyclicality looks very different when you look at microdata than it does when you look at aggregated data because with aggregated data you're not usually looking at comparable labor forces over time. Sumner points to real wage counter-cyclicality in demand-driven recessions and says "I’d add that real wages rose especially sharply in some of the most easily identifiable adverse demand shocks (1920-21, 1929-32, 1937-38.)". Maybe - but the story here is very unclear to me. 1920-1921, for one thing, has been considered a supply shock since Romer's work in the 1980s. The other thing is he has to be looking at aggregate data for these episodes, unless he knows of some microdata that I don't. I doubt it would reverse the finding given the deflationary pressures, but it makes it harder to assess.

*****

This is actually all right up the alley of what I want to do for my dissertation. In my applications I proposed looking at excess worker turnover as a wage adjustment strategy for firms facing sticky nominal wages. I'm not wed to that specific point, but I want to work on something related to worker flows and wage adjustment.