I’m cracking away on my macro paper and thought I’d post about the paper on which it is based.
Basically, as I wrote about in an earlier post, recessions at graduation cause college grads to opt for graduate school at higher rates. This obviously makes the application cycle more competitive, and the cohort that comes out as a result is thought to be stronger.
This paper creates the coolest dataset ever – from scratch! Funnier still, all of my professors are actually in this dataset! I hope to be in this dataset, someday. (For my other paper, I’d say the exact opposite – it’s a dataset of New Hampshire criminals…)
The authors used a JSTOR tool to aggregate the publication record of every single person who received an Economics PhD between 1955 and 2004. They then computed things like the number of citations, number of publications, and Hirsch indexes for these subjects, and used these measures of productivity to quantify the “productivity premiums” caused by macroeconomic conditions.
I am blown away by the amount of work that was put into creating this dataset, and the productivity premiums are pretty impressive as well. The data itself is fascinating, though – there’s an alarmingly long left tail in the distribution for “number of publications” – most people who come out with a PhD never publish a single thing. The same goes for the lower tiers, with the highest tier of PhD programs having the lion’s share of publication and citation. The paper also used tons of different economic indicators to measure conditions at graduation – basic unemployment rates, as well as the DOW industrial average and the S&P 500 to measure economic fluctuations. Their estimates of the productivity premium were consistent across these different measures.
I’m delving into the Roy model for my paper, trying to see if there’s a variable switching cost that affects graduates’ decisions. Sure, everyone will allocate their talent according to economic conditions – but can everyone do it that well? Is talent uniformly transferable? I hope to find the answer in my macro paper!