A Tweetstorm on Imperfect Information Processing
Because inflation today is caused by firms changing price today & inflation tomorrow is caused by a different set of firms changing price tomorrow, sticky inflation in the face of a big macro shock means firms are changing prices without fully incorporating the new information.
— Miles Kimball (@mileskimball) February 22, 2024
Mankiw and Reis interpret their model as a model of sticky information: firms periodically wake up and get all the new macroeconomic information, then set a new price path accordingly.
— Miles Kimball (@mileskimball) February 22, 2024
Macroeconomic information is actually very easy and cheap to get ahold of. Just google "FRED" and you have a cornucopia of macroeconomics information. Getting macro information is not hard. What is hard is appropriately *processing* macroeconomic information.
— Miles Kimball (@mileskimball) February 22, 2024
When information processing is very costly, it makes sense to economize on it. Hence, most firms only process macroeconomic information into their prices at long intervals. This gives rise to sticky inflation.
— Miles Kimball (@mileskimball) February 22, 2024
In the last half century we have seen modeling *imperfect information* become a solved problem. By contrast, we remain miserable at trying to model *imperfect information processing.*
— Miles Kimball (@mileskimball) February 22, 2024
And avoiding modeling agents who are too smart by modeling agents who are too stupid (e.g., adaptive expectations, finite-state automata, agent-based models) is not that helpful.
— Miles Kimball (@mileskimball) February 22, 2024
Anyone who can come up who can come up with a model of imperfect information processing that is elegant enough that it gets taught around the world in core graduate micro will be a hero.
— Miles Kimball (@mileskimball) February 22, 2024
For more on these issue of finite and scarce cognition (a synonymous phrase for imperfect information processing), see my paper "Cognitive Economics." It isn't easy to publish a methodology paper. This is mine. https://t.co/ZtCbpFdfzz
— Miles Kimball (@mileskimball) February 22, 2024