Ryan and Gene on RBC

Here and here, respectively.

I think they both make important points, although it looks like the post Gene had about it being a theory of the error term is down... I'd disagree with that a little.

So Gene is right that it doesn't really make sense to call it a cycle theory - it's really just a random downturn theory. But that's true of Keynesians too, and Austrians for that matter. The old accelerator-oscillator is actually a "cycle", but I think most business cycle theories aren't cycle theories. Kind of an interesting wrinkle, I guess.

I do think RBC gets a bad rep in terms of their ideological content. It's precisely because they've been genuinely dedicated to explaining the data that they've (over time) incorporated the frictions and things that Ryan refers to (and as I've mentioned recently - they include more realistic microfoundations, such as including home production in labor supply decisions).

I think one of the biggest weaknesses of RBC is its reliance on calibration. It makes it very hard to assess whether the model is a good way of thinking about the underlying data generating process or whether they have just smashed a round peg into a square hole with a big hammer. Contrast this with even a naive Keynesian model that has some interesting out of sample predictions about some stylized facts, for example, at the zero lower bound - I am much more convinced by that than by repeated ex post calibration. When those naive Keynesian predictions are reaffirmed in New Keynesian models that incorporate all the (in Ryan's words "shitty") microfoundations that an RBC has to offer, I wonder even more what RBC adds.

The point that RBC is just a theory of the error term is both right and wrong. It's wrong no a theoretical level. A shock and a random error are really two different things, I think. But on an empirical level Gene has a point. It's one thing to theorize about how real shocks or technological shocks will impact the economy. It's another thing to not have an actual variable to measure that shock and just assume any perturbaitons are precisely that sort of shock.

This gets into the application of RBC to real life as well. As many have pointed out: exactly what technological shock occurred to throw millions of people out of work in the last five years? It's hard to take that seriously. We have, after all, a massive, obvious, front-page news story demand shock. Why mess with RBC, then?