Monday, March 12, 2012

Peter Diamond On The Slow Recovery Of Employment

I have just returned from the Eastern Economic Association meetings held in Boston over the weekend. Peter Diamond delivered the main plenary lecture on Saturday evening (2/10/12) on "Markets with Search Frictions." While I disagree with some things he said (He thinks the "natural rate" equals "NAIRU" [and that these are both meaningful concepts], and as always wants to "fix" social security, but these were not his main topics), he mostly gave a wise and knowledgeable presentation about the search model of unemployment, going back into its routes and noting many of its limitations and problems, as well as how it is useful, reminding everyone in the audience what fools those in the Senate are who think he is not qualified to serve on the Board of Governors of the Fed.

One general point he made that I had not really thought about, although once pointed out it is obvious, is that labor markets are seriously different from textbook supply and demand models in that there is never a really clearcut equilibrium. There are always vacancies, hence some "excess demand," and some official unemployoment, hence some "excess supply." All the imposed definitions of labor market equilibrium are thus arbitrary.

The most interesting remark to me came in reply to a question from the audience. He had stated in his main talk that "the matching mechanism has broken down during the recent recession." He was asked to elaborate. Drawing on research by Davis, Haltiwanger, and others, he broke it down to the micro sectoral level, although saying that more is going on than just that. But at that level, different sectors have different hiring rates. The one with the most rapid hiring rates is the one with the least hiring, construction. Some with slow hiring rates include education, health, and government. Not all of those latter are growing that much, but health is certainly one of the most rapidly growing. So, quite aside from broader macro issues (including possibly reduced mobility from underwater mortgages, although some studies claim this is not a factor), this sectoral pattern of how the recession has hit has slowed down the hiring/rehiring portion of the matching mechanism.

Outgoing president, Duncan Foley, also gave an excellent talk on physical limits to growth related to climate and energy, but, I shall not comment on that at length here and now (title, "Dilemmas of Economic Growth"), other than to say I largely agreed with it and he showed some very interesting statistics on various things that I had not seen before.


Peter Dorman said...

Based on your account, Barkley, I'd say that Diamond left out an intermediate step in his discussion of labor market S&D.

The traditional Beveridge Curve view, going back many decades, was that the 45-degree line, where unemployment equals vacancies, represented S=D in an S&D diagram, and that prolonged periods away from this line demonstrated that the labor market had no inherent tendency to clear -- good ol' fashioned Keynesianism.

After search theory arrived on the scene (due in significant part to Diamond, of course), this assumption was challenged. Why should u=v have any significance at all, since it was a static condition in a dynamic setting? That is, search is about separating, matching and hiring, which are all rates denominated by time, so why would comparisons of u and v at some moment have any independent importance apart from their effects on the flows of consequence?

The solution that was put forward was, as so often in economics, stationarity: the equivalent in search theory of S=D is the proportion of u to v that remains constant over time given the determinants of the component flows.

I have always thought that this alternative is arbitrary, based on the assumption that the parameters we see at the moment are fixed through all eternity. Maybe Diamond's comment is that he agrees.

To be clear, I think "equilibrium" is largely meaningless in a search model of the labor market, and the more interesting question is whether there is a meaningful notion of full employment. My rough and ready answer would be: pick a period during which your economy produced labor market outcomes that look like full employment to you, and call that u/v full employment. And perhaps Diamond is saying that stationary inflation (NAIRU) is that magic moment. That dissolves the meaning of "full employment" as something separate from price level dynamics, of course.

Barkley Rosser said...


The other Peter only briefly tossed off a line about the natural rate equaling NAIRU, although he definitely said it. He did talk about the Beveridge curve some, with it clearly delineating this relationship, even if it is not clear that its intersection with the 45 degree line is an equilibrium of any meaning or standing.

What he did talk about it is how that curve can shift from time to time and apparently has in a bad way during the past few years. This became sort of his way of stating the appearance of these nastier macro effects going on in the job market, although that shift also presumably partly reflects these micro-based processes that lead to what he called the "breakdown" of the matching mechanism.

There is also an obvious parallel between this discussion and that about the nature of at least real world, short-term Phillips curves, although he completely stayed away from that.

spencer said...

Would love to see those economic series
on various things you had not seen before.