j.r. leonard

1/Here's a follow-up thread about not trusting macro models. In an earlier thread, I pointed out that permanently low interest rates might have effects on productivity - something that doesn't appear in either standard macro models or macro intuition.https://twitter.com/Noahpinion/status/1100515008453066752 …

2/Here's an Atif Mian thread about other possible outside-the-model effects of low rates:https://twitter.com/AtifRMian/status/1100212663353835520 …

3/I learned a bunch of macro models in graduate school. A whole modeling paradigm. Those models were - and are - very dominant in academia. But in the Great Recession, none of the standard models worked. They all basically failed to predict or describe what was going on.

4/Now, they didn't fail 100%. For example, Ben Bernanke's own models of the "financial accelerator" helped convince him to help bail out the banks, which - though not fair - probably helped avoid a deeper longer recession. But in general, they did fail.

5/But I was not surprised that these models failed. They were massively chock-full of unrealistic assumptions. http://noahpinionblog.blogspot.com/2013/05/what-can-you-do-with-dsge-model.html …

6/The popular models were also linearized, and they generally assumed unique equilibria. That, plus the simplifying assumptions, meant that any really unusual macroeconomic situation - a financial crisis, a deep recession - was outside their purview. http://noahpinionblog.blogspot.com/2013/03/the-swamp-of-dsge-despair.html …

7/This episode taught me that macro models are at best heuristic stories, not ironclad guides to future events. I had thought that everyone else took away the same lesson, but then along came MMT.

8/Suddenly, people who were willing to get in the weeds and criticize the details of mainstream macro models (representative agents! rational expectations! linearity! perfect competition!) are swallowing this far sketchier MMT theory wholesale.