Okay, going to wade into the MW debates against my better judgment, but hopefully in a way that facilitates productive discussion. 1/N.

Let’s say I take all the empirical MW research at face value (I generally do): this means modest increases in MW have little to no negative impact on employment, do not reduce the number of jobs, and do a number of other good things: reduce racial inequality, poverty, etc. 2/N.

What is the model that rationalizes the findings of MW on employment? A lot of people claim Monopsony, and I was initially persuaded by this (clearly Monopsony would be consistent with these findings) but I’m not clear it is as clean an answer as it first appears 3/N.

Here is a great thread on the sources of monopsony power. The problem is, the source of monopsony power most consistent with the MW story—concentration—appears not to be the main source of wage setting power in the literature (at least, according to Arin’s thread here). 4/N https://twitter.com/arindube/status/1192249126064398336

Where does that leave us? Either monopsony power due to search frictions (dynamic monopsony) is pervasive enough to generate the observed MW effects or its down to monopsony due to heterogenous worker preferences over job characteristics (job differentiation). 5/N.

The former is possible, but seems unsatisfactory (what are the specific frictions in low wage labor markets?). The latter implies minimal welfare effects from monopsony: if mkt power due to workers’ preferences for job characteristics, hard to it argue causes welfare loss. 6/N.

My read is thus that MW has little to no negative empl effect and lots of other positive effects, but that we don’t have a good model for why. Monopsony works, but not perfectly. Not sure what the answer is, but think it involves models with #class conflict and bargaining. 7/7.