I received the following email the other day:

Given your past criticisms of this issue in your posts, I do not think you will like my co-authored paper, “Microfoundations of the Business Cycle and Monetary Shocks” . . .

Given this lead-in, of course I had to take a look. The paper is by James Holmes, John Holmes, and Patricia Hutton, it’s called “Microfoundations of the Business Cycle and Monetary Shocks,” and they say:

Non-rational expectations can produce larger expected real income for some or all agents than rational expectations. . . . Hence, “rational expectations” are not rational, and “money illusion” can be optimal and satisfy the “Lucas Critique.” . . . Nominal wage rigidity or stickiness can be the rational response of firms or workers and need not be evidence of money illusion.

I asked Holmes why he thought I would not like the paper, as I am not opposed to microfoundations when they are of interest. My opposition is to the attitude that microfoundations are *necessary* for macro understanding. Holmes replied that his point was that his article was an example of how a simple micro model can yield new and unconventional insights which would not have been obtained simply by studying macro patterns in a statistical, model-free way. (Here I’m using the term “model” in the economic sense of a mathematical formulation of individual decisions, rather than in the statistical sense of a joint probability distribution.)

To get to the details of the paper . . . hmmm, I’ll leave that to others. (Paging Rajiv Sethi . . .) I am sympathetic to the sorts of arguments presented in this paper but I don’t have the energy to go through what the authors are actually saying. I have no training in studying this sort of model (my last econ class was in 11th grade!) and I can’t bring myself right now to put in the effort required to follow it all. But I’m setting it out there in case any of you are interested.

It seems odd to me to argue that, because non-rationality can collectively generate Pareto-superior outcomes, that non-rationality is consistent with individual optimizing behavior. I think a more compelling example should demonstrate that a single agent unilaterally deviating to some form of non-rational expectations would benefit from doing so. I’ve only read a page or two of the paper, though, so maybe they actually do something like this!

“My opposition is to the attitude that microfoundations are necessary for macro understanding.” – This is a very interesting sentiment that I agree with and have often wondered whether the same can be said for other disciplines where a bottom-up principled approach exists; namely, Physics. Would you consider Statistical Mechanics an example of this, namely a situation where a Statistical approach yields macro understanding where the micro-approach is futile? (I am unsure, since while Statistical Mechanics makes probabilistic assumptions about the states of particles via the Canonical Partition Function, it still requires “micro”- concepts like energy, so it’s not necessarily a purely statistical approach, from a Statistician’s point of view.)