When will people move away from pure AD theories of unemployment?

Commerce department figures on Friday showed that total sales in 2010 were up 6.8 per cent from 2009, marking the sharpest such increase in more than a decade…

Industrial output is up by 5.9 per cent year-on-year.

Yet the labor market is still "eh."  Here is more, but again note it is wrong to reject the AD factor altogether, though it seems to be becoming less relevant over time.  Arguably AD and AS are interacting in unusual and presumably deleterious ways.

I have read too many blog posts attacking a caricatured version of either RBC theory or a narrowly defined notion of "structural unemployment" which requires excess demand for labor in significant parts of the economy.  As Arnold Kling points out, the labor market shock can be asymmetric in its effects.

From a different direction, here is Scott Sumner criticizing the recalculation argument.  I read Scott as establishing the conclusion that both AD and AS must be at work.


Where would unemployment be if AD (NGDP) was on the prior growth path, or at the location estimated by the prior growth path?

@okin15: I'm rather sceptical of the very idea of "aggreghate demand". I used to work for a petrochemical firm and we never talked about the number of tons of stuff that we sold in a year, and scarcely ever of the total sales revenue we raised. We talked endlessly of the trade in each of the products we sold; in other words, our interest was in analysis, not aggregation. We even talked of how much we might have sold of a product we didn't sell, if only we had manufactured it. But aggregates - nope. Because, I suppose, our customers never phoned to ask for a thousand tons of petrochemicals. They always wanted something specific.

I am an economist and I don't really have a clue what AD stands for either. AD and AS are wonderful ex post explanations for whatever we observe (and usually failed to predict). But specifying them in advance with any confidence or predictive power has always struck me as magic.

How does aggregate demand (or Keynesianism) explain the differences in unemployment rates between ethnic groups?


Your comments appear to be misdirected, but I don't understand what you are saying. Fixed costs do not require prices to be above MC. As long as MC is rising, the fixed costs can be covered by the inframarginal incremental profits.

Anon, I agree. It's expected future output that matters, not current output, which Tyler seems stuck on. Most every economist realizes that the stock market is the best predictor of the economy 6 months out -- however wrong it may be some of the time -- yet Tyler seems to think the fact that the market is still well below its long term trend doesn't mean much.

That said, I think Tyler is creating a strawman when he attacks "pure AD theories of employment". Those arguing for more AD tend to realize it is more complicated than that, but that doesn't mean lack of AD still isn't a big part of the problem.

Mario, OK, but to paraphrase Scott Sumner: why are there so many fewer restaurant workers than a few years ago? Is the restaurant industry being revolutionized? Last time I went to a restaurant they still had a head chef, sous chefs, other cooks, a manager, assistant managers, hosts, bartenders, a lot of waiters and possibly some bus boys. If some new genius is revolutionizing the productivity of the restaurant industry I sure would like to hear about it and get in on it. Seems like Tyler would have mentioned it.

In New York, for example, restaurants are heavily dependent on the financial sector. Complementary industries. There may be fewer restaurants or high-end restaurants which are more service intensive. I am just guessing. (I spend my time on other matters.) But I fear the AD explanation is just the lazy person's (economist's) explanation.

Doesn't it strike you as just a little weird that you can take an idea -- the demand curve -- developed by Marshall for a single small industry and then simply aggregate across the whole economy? (Schumpeter did.) Voila depressions are explained

In any event, since we are not likely to get more stimulus the AD crowd should prepare itself for an never-ending recession --or high unemployment economy. It will just never end. Permanent stagnation and all that post-WWII Keynesian stuff.

"In New York, for example, restaurants are heavily dependent on the financial sector."

In general, restaurants depend for their business on the local economy. Recalculationists might say that the financial sector needs to shrink and New York restauranters are SOL, at least in the short run. But this does not address all other places where restaurants are doing poorly.

"Doesn't it strike you as just a little weird that you can take an idea -- the demand curve -- developed by Marshall for a single small industry and then simply aggregate across the whole economy?"

It should strike you as weird because it's wrong. Modern macroeconomists (both New Keynesians and New Monetarists) use general equilibrium analysis, not supply and demand curves, but the stylized facts about the business cycle are largely unchanged, so we use the term "aggregate demand" for ease of exposition.

Your remarks about the effectiveness of "stimulus" are puzzling. As dirk states the word is ambiguous, but monetary easing has been quite effective, and inflation pressures are modest.

"When will people move away from pure AD theories of unemployment?"

I assume it will be when a theory emerges that is either demonstrably more accurate, more predictive, or gets the same results within a simpler framework.

The "ether theory" was a useful construct for understanding how electromagnetic waves could seemingly travel in a vacuum (by providing the answer "they do not"), but was not more precise than special relativity. We lost some intuition, but gained numerical accuracy and predictability.

In contrast, though introduced early on, the "corpuscular theory" did not become as useful as waves until experiments and classical theories started to deal with quantum systems. So people did not move. Its "correctness" was relative to the data.

Theories that involve AD seem as predictive and accurate given the data as those that involve AD/AS. Most economic theories seem to be at the level of orders of magnitude and proportionality arguments, so there seems little reason to prefer one over another or not relegate each theory to its particular effect: e.g. Boyle's law vs the ideal gas law vs statistical mechanics. In particular, there are not enough graphs of data with error bars and lines going through them or not going through them to warrant adherence to one theory or another.

It looks to me that we are in the Boyle's law (1662) era of macroeconomics. Or maybe the Carnot cycle (1824) era with some of the more sophisticated things like Black-Scholes working away in particular areas of finance. A unifying first principles approach, like statistical mechanics (1870), has yet to be developed nor have we gotten to the fluctuation theorem (1994) level. Not to say there haven't been advances. But someone trying to explain why Watt's steam engine was more efficient than Newcomen's using Boyle's law would be met with skepticism, and any hand-waving about how temperature, pressure and volume are proportional to each other would be roughly as good as any other.

As a side note, I think this may explain why people with no training (e.g. me) believe they can just weigh in on an economics discussion.

Also, no offense. I'm not looking down on economics from a physicist's perch (although we do have the most precise theory-experiment agreement in existence). On the contrary, it seems like an exciting time in economics. I would even wager to think that a new paradigm for economics might arise from these great recession years.

Comments for this post are closed