Thursday, April 26, 2012

Yes, another puff post...

... because I don't feel able to exert the required effort for a more substantive post. But this may be interesting.

It's the last week of classes - I've already had my last econometrics class and U.S. economic history class, and tonight is my last macro class. I was going to share a few things I learned or gained a greater appreciation for in each.

Econometrics
1. Two thirds of the class was math stats (not untypical of graduate econometrics), but one thing that was more atypical (for me) is that we worked a lot more with specific distributions. So I feel much more comfortable demonstrating things with gamma distributions, beta distributions etc. - some of which I had never heard of before much less played around with.

2. This is kind of simple, but the professor also focused more on maximum likelihood than past professors of mine had. I always used to get the intuition behind it, but then we just did stuff on the computer. In this class we actually did a lot of derivations of maximum likelihood estimators by hand, which was an exercise I really appreciated.

3. Throughout the professor emphasized the role of subjectivity and judgement in doing empirical work, which is something that I think a lot of people gloss over - that was good.

U.S. Economic History
1. I got closely acquainted with the organization of work/agency costs take on corporate and labor history. Major players here are Stephen Marglin, Richard Edwards, and Katherine Stone - but going back to Charles Babbage. I wonder if the Kirzner types ever read any of this stuff - it seems very relevant. Back during the Mankiw bruhaha I didn't think all that much of Stephen Marglin. I think more highly of him now (not that my appreciation of Mankiw has declined or anything).

Macroeconomics
1. I always thought highly of Robert Lucas, but I appreciate him even more after this course. That's not even to say I think all of his points were even right. For example, I'm still pretty Keynesian in the way I think about the Phillip's Curve. But this course has only improved my respect for his methodological point.

2. I am more and more convinced that the label "microfounded" is a misnomer. What these models are is grounded in optimizing behavior. For the most part, they really aren't "microfounded" as you would think a word like that would imply.

3. The conservative central banker model still really bothers me... perhaps I'll return to this again at a later point. Assumptions matter a great deal.

3 comments:

  1. Does AU not include micro/price theory in its core year?

    ReplyDelete
  2. Deriving crap by hand with maximum likelihood is the worst thing in the world.

    ReplyDelete
    Replies
    1. No, practising a multitude of ML derivations in preparation for your final econometrics exam only to discover that all those heavy hints dropped by your professor were a red herring... is the worst thing in the world.

      (I do rather like ML though.)

      Delete

All anonymous comments will be deleted. Consistent pseudonyms are fine.