Rethinking Macroeconomics, Continued
Two links to discuss in this post. First, David Altig discusses the “output gap,” which I put in scare quotes because it is a concept that I am about to attack. Pointer from the indispensable Mark Thoma, who also provided the link to the Ray Fair paper I will discuss second.
1.The output gap
You might not know it from Altig’s post, but the debate over the output gap can get quite heated. See Greg Mankiw.
I think it is misleading to speak of the output gap. It is part of what I call “hydraulic macro,” which thinks of the economy as producing a single good with a single type of labor and a single type of capital. As I stated here and expanded on here, I think that this is a dangerous simplification.Macro makes a number of simplifying assumptions. There is a longstanding battle over the rationality assumption. Akerlof and Shiller’s book, Animal Spirits, is one huge attack on the rationality assumption. I am sympathetic to that attack. My problem with their book is the implicit assumption that there is no irrationality in government. It is as if one goes from having animal spirits to being perfectly rational simply by switching from the private sector to the public.
But I think that the simplifying assumption of homogeneous output, labor, and capital is equally dangerous. My claim (which is not original with me–it is recognizably Austrian) is that a recession can be thought of as a recalculation. Imagine a central planner who decides to radically change plans. He has a huge recalculation to make in order to figure out where to allocate labor and capital. He says to some people, “Wait a minute. I am thinking. Some of you just have to stand idle while I figure this out.”
The market economy is like that central planner. We are undergoing a Great Recalculation.
In macro, we teach that there are different types of unemployment: structural unemployment, where there is a mismatch between skills and needs; frictional unemployment, where the skills and jobs match up in theory, but in practice people are stuck in the search process having not yet found the right match; and cyclical unemployment, where people are unemployed because of inadequate aggregate demand.
People who are out of a job cannot really tell the difference–they could not tell you whether their unemployment is structural, frictional, or cyclical. As an economist, I would say that if you are laid off from a manufacturing firm that accumulated excess inventory and you will be recalled once the inventory is worked off, then you are cyclically unemployed. Otherwise, I cannot tell why you are unemployed.
Up until about 1980, most of the unemployment in a recession was unemployment that I would recognize as cyclical. However, certainly since 2000, most of the unemployment has not met my definition of cyclical.
The old-fashioned inventory recessions involved one class of miscalculations–firms over-produced relative to demand, and then had to cut back for a while. Those miscalculations are less important today, in part because computerized inventory and integrated supply chains tighten the connection between production plans and reality, and in part because a lot of the actual production of stuff for U.S. consumption is done overseas, so our inventory miscalculations affect foreign workers relatively more and U.S. workers relatively less.
The recalculation that followed the Dotcom bubble was different. So is the current situation, which I call the Great 21st-century Recalculation. (The 1930’s might be called the Great 20th-century Recalculation.)
In conventional, hydraulic macro, we think in terms of this one good called GDP, and the output gap is the difference between how much of this GDP stuff we could produce if everybody were working and how much we are actually producing with all the unemployment over and above “normal.” We assume that “normal” unemployment, which is structural and frictional, is some roughly constant fraction of the labor force.
The way I look at things, we have a huge amount of structural and frictional unemployment these days. There is very little cyclical unemployment–limited to autos and household durable goods. If you measure the output gap using my definition of cyclical unemployment, then the output gap is tiny.
But I think it is better to drop the concept of the output gap altogether. Let us just say that we are in the middle of a Great Recalculation. Before the housing bubble popped and we figured out that our housing and financial sectors were messed up, we had something like full employment. Now, we need a Great Recalculation to figure out a new allocation of skills and jobs that gets us back to full employment. Because we got so messed up before, it will take years for the recalculation to be completed. Meanwhile, if you insist on thinking in terms of a large output gap, you are talking about an imaginary world in which the Recalculation takes place instantaneously. What you are not talking about is a gap that can be closed with government action, unless the government miraculously does the Recalculation faster and more effectively than the market.
2. Macro forecasting
Ray Fair has a new paper. Ray is the last macroeconomic forecaster left in academia. In fact, he was already the last macro forecaster in academia in 1977, when he visited MIT to teach a course in the macro sequence, where I remember John Huizinga sitting in the back of the classroom laughing at the fact that every equation in Fair’s model used the lagged dependent variable. See The Lost History of Macroeconometrics.
In his latest paper, he says that what made it impossible to forecast the current recession was a combination of not knowing the path for: house prices, equity prices, import prices, exports, and random shocks. If you had known all of those, you could have forecast the recession. Not knowing those, from a forecasting point of view it was a “perfect storm.”
The way I would put it is that macroeconometric models built on the assumption of one good called GDP and one type of labor work fine as long as nothing happens in the economy that requires a major recalculation. But once something dramatic occurs, all bets are off. Obviously, that does not leave me highly enamored of macroeconometric models, but you already knew that.