I’ve been commenting on DeLong lately because I’ve been hanging out at his blog pondering this post, which is part of a dispute with Arnold Kling. (The discussion also includes Noah Smith’s post that I previously discussed.) Kling is arguing that we’re in a process of “recalculation”, where the market is trying to re-value certain sectors.* DeLong doesn’t think much of this idea, and argues that the problem is much simpler, inadequate aggregate demand.
Underlying this, of course, are questions of methodology, and DeLong challenges those from the Schumpeterian/Hayekian approach who argue that the economy is too complex to be modeled.
Modeling 140 million workers, 10 million firms, and 20 million commodities is really complex–that’s why we don’t do it, and don’t have a big computer centrally-planning our economy. That is why we use the market system.
But when it comes to business-cycles–to recessions and depressions and downturns–we don’t need to model 140 million workers, 10 million firms, and 20 million commodities: we only need to model two: (OK, four): currently-produced goods and services on the one hand, and (perhaps three types of) financial assets on the other. A business-cycle downturn comes when–for any of a number of possible reasons–there is an excess demand for financial assets and a corresponding deficient demand for currently-produced goods and services…
To this extent DeLong is absolutely right. Of course the economy can be modeled fairly simply. To some extent any complex phenomenon can be modeled simply. That’s an argument I’ve been having with opponents of rational choice theory in the social sciences for years, so I come down solidly on DeLong’s side here. But of course no model is good for everything. Because models, by definition, abstract just portions of reality, every model is good for some purposes and not for others. As powerful as rational choice theory is for general analysis, however, Bryan Caplan has persuasively shown that when it comes to really analyzing voter decision-making, many aren’t really rational at all. So there’s no doubt that the economy can be modeled simply–the question is “what problems is this model well-suited for analyzing?”
And my doubts rise up when he continues with this:
In the background the market system is trying as best as it can to find the best uses and production plans for 140 million workers, 10 million firms, and 20 million commodities given the state of aggregate demand. But that is not of the essence in understanding low capacity utilization and high unemployment. The aggregate demand shortfall is.
My doubts about the utility of the model rise here because Arnold Kling is arguing that in this case the market’s struggle to “find the best uses and production plans” for all those workers and firms is “the essence in understanding” our current problem. He writes:
Most of the time, individuals and firms are adjustment-averse. Adjustment-aversion is rational in a world that is mostly mean-reverting. If you stand pat in the face of a transitory drop in demand, as a firm you avoid paying excess adjustment costs associated with hiring and training, and as an individual you avoid paying the excess adjustment costs associated with making new investments in human capital.
However, the world is not always mean-reverting. There are secular changes taking place as well. In my view, the Internet and globalization have produced a lot of secular change. Because of adjustment-aversion, by 2008 a number of industries, such as the legal industry, had not kept up with these secular changes. Because they had failed to adjust gradually to the underlying technological changes, what we got was a large step adjustment, greatly reducing the size of the legal industry relative to its previous trend. This took place in many other industries as well.
So far, the market has managed to figure out that we do not need so many lawyers. What it has not figured out is where the excess lawyers should go and how to make the transition happen. Again, multiply this out by many sectors, and you get a lot of unemployment.
This sounds right to me, or at least plausibly so. The elegant model of DeLong and Krugman seems to require a simple decline in demand without a substantial change in type of demand. That is, their model is suitable for a “mean-reverting” economic situation, because in their model, demand is demand is demand. The differing components of it are not relevant, so it is treated as an aggregate (hence, aggregate demand). In that case simply increasing demand should suffice to set things right, as DeLong and Krugman, with increasing frustration, keep telling us.
But if the types of demand are changing,then maybe the differing components do matter. Maybe the problem is not a generalized decline in demand but sector-specific declines in demand. That doesn’t mean increasing aggregate demand would be entirely useless–indeed Kling suggests that “[m]aybe expansionary fiscal and monetary policy can speed this process” of readjustment–but it does suggest that the analytical power of the aggregate demand model would be much more limited than normal, and the policy prescriptions derived from it less appropriate or effective than normal.
Critiquing the Aggregate Demand Model
Before just tossing the aggregate demand model aside, however, let’s look at it from DeLong’s perspective. In general the model has served him well. Perhaps in extraordinary cases it might not, but that means those who claim it doesn’t serve well now bear the burden of demonstrating this is an extraordinary case. Just saying it’s so doesn’t cut it. So while I am more sympathetic to Kling’s approach, DeLong’s defense is based on solid intellectual foundations.
Yet ultimately I am not persuaded by DeLong’s defense because I don’t think he, or Krugman, provide a persuasive answer to Kling’s question of what we do with the excess lawyers.
The AS-AD paradigm says that we can go back to something like the 2007 status quo with expansionary fiscal and monetary policy. The P[atterns of] S[ustainable] S[pecialization and] T[rade] paradigm** says that there is no going back. The market needs to figure out new ways to best utilize would-be lawyers and other unemployed workers. Maybe expansionary fiscal and monetary policy can speed this process, but the mechanism by which policy will do this is not as simple and reliable as the AS-AD paradigm suggests.
Instead, we get this kind of argument from DeLong:
When you ask believers in “recalculation” what pattern of production and trade proved to be unsustainable in 2007, they answer: “building so many houses.” When you ask believers why the market economy has been unable to sort out this problem in three years, they answer with nothing–silence. When you say that OK, there were $300 billion of excess houses at the start of 2007 but now construction has been so depressed for so long that there are $1 trillion fewer of houses than trend and why isn’t the 2007 pattern of production and trade sustainable again, they answer once again with nothing–silence.
But part of the dynamic view of the market is that nobody can fully understand the dynamics of the market, so to criticize them for not being able to satisfactorily explain the dynamics of the market is not just beside the point, it’s a little bit dishonest. And to ask why an earlier pattern is not sustainable again is to wholly miss the main claim made by Kling, which is that patterns are shifting, so old patterns are no longer sustainable. DeLong might as well ask why the old pattern of 40 acre farms is not sustainable again, the only difference being that now enough time has gone by that we can understand the answer to that one. Farming still matters, as does construction, but the pattern or structure of it has changed. Now maybe Kling is empirically wrong about that–only time will tell–but that doesn’t change the fact that DeLong is asking the wrong questions.
Is the Aggregate Demand Model Suitable for Unusual Recessions?
Curiously, DeLong’s insistence that the normal model is the appropriate one seems to contradict Krugman’s argument that this is not a normal recession. Krugman is on record arguing that while monetary policy works fine in most recessions, fiscal policy is required in unusual ones such as the current one. But isn’t the unusual recession, the persistent one that’s not responsive to normal monetary policy tools, the one that’s most likely to be more than a mere shortfall in demand, and more of a consequence of real economic restructuring? So which is it? Or is the implicit argument that this is an unusual recession, but not that kind of unusual? If so what kind of unusual is it that it fits a standard model but requires a non-standard policy response? I wouldn’t go so far out on a limb as to argue that no such argument can be made, but I don’t think it has been made, at least not that I’ve seen.
* For reasons that aren’t clear he’s taken to calling this patterns of sustainable specialization and trade, which may do more to confuse the issue than clarify things.
** Paradigm, my ass! It’s a pet peeve, but the watering down of the word paradigm has, I think, stripped it of any useful meaning.