In January I made a prediction and kept it a secret. Now I can tell you what it was.
For fun I wanted to see how well I could predict the Economics PhD job market. As a simple test I tried to predict who would be selected for the Review of Economic Studies Tour, a kind of all-star team of new PhDs. Here was my predicted list
- Gharad Bryan (Yale)
- Matt Elliot (Stanford)
- Neale Mahoney (Stanford)
- Alex Torgovitsky (Yale)
- Glen Weyl (Harvard)
- Alex Wolitzky (MIT)
- Alessandra Voena (Stanford)
- Kei Kawai (Northwestern)
And here is the actual list (the RES Tour website is here.)
- Alex Wolitzky (going to Stanford?)
- Daniel Keniston (MIT, going to Yale)
- Mar Reguant (MIT, going to Stanford GSB)
- Kei Kawai (going to Princeton?)
- Alex Torgovitsky (coming to NU)
- Alessandra Voena (going to Chicago?)
- Peter Koudijs (Pompeu Fabra, going to Stanford GSB)
So I got 4 out of 8. (There are usually 7 people, and I predicted 7 originally, but I updated it the next day adding Kawai to the list. I had been so involved in recruiting students from other schools that I had completely forgotten about our own star student Kei Kawai and as soon as I remembered him I added him to the list.)
You can see why I wanted to keep the prediction a secret until the market was over. You can verify my prediction by cutting and pasting the text in this file and generating its unique SHA1 hash (a digital signature) with this web tool and cross check that it is the hash that I originally posted here, and that I tweeted here, and that is reproduced below.
I did this as an experiment to see how easy it is to predict job market outcomes. At the time I made this prediction I had read the files of each of these candidates and interviewed most of them. I didnt know where else they had interviews and I made the prediction before the stage of flyout interviews so I had little information about how their job market was going overall.
Getting half right is about what I expected. To me this is evidence that the market is hard to predict even after having interviewed the candidates. In particular I take it as evidence against the cynical view that the market herds on certain candidates early in the process. Indeed I would not have changed my prediction much even a week or two later when flyout schedules were in place.
Incidentally the Review Tour rosters say something about the strength of PhD programs. Here’s a breakdown of the last 6 years and where the tourists received their PhDs:
Stanford GSB 2
BU, Duke, LSE, Michigan, NYU, Penn, Princeton, Ohio State, Stern, UPF, UCL — 1 each
Northwestern is doing very well! (In addition to producing stars, we also do well hiring them. 3 from the tour in the past 6 years, including Alex Torgovitsky this year, an outstanding hire.)
Also I understand that a team is at work creating the web tool that I suggested here for creating and managing secret predictions. If and when it hits I will announce it here.