(It used to be 4 and 5.)
A student in her 5th year who doesn’t have a stellar job market paper is always tempted to stay another year and try to produce something better. This is the ex post incentive of an individual student.
But ex ante the department as a whole would like to enforce a commitment for all students to go on the market in 5, even those whose job market paper at that stage leaves something to be desired. The basic reason is risk aversion. Every year they spend in grad school they produce another signal for the market. Good signals improve their prospects but bad signals make them worse. They would avoid the additional risk by committing to stay only 5 years rather than 6.
Now consider a student whose job market paper in year 5 leaves something to be desired. If she stays another year and produces a good paper, then although she is better off, she raises the bar for her colleagues and thereby strengthens their incentives to stay another year. A department policy that strongly incentivizes students to finish in 5 is needed to prevent the implied unraveling.
But that’s MIT. Then there’s everybody else. Students in other departments have to compete with MIT students for top jobs. At a department like Yale, only the best students will be able to compete for top jobs and this makes them risk loving not risk averse. Instead of wanting to minimize signals, the best Yale students want to produce enough signals in hopes that at least one of them is good enough to give them a shot at a top department job.
So one should expect funding, TAships, and face time with advisors to drop after 5 years at MIT but continue into the 6th year at Yale. (Note: I have no data on this.)
Cheap Talk 
9 comments
Comments feed for this article
February 13, 2012 at 3:15 pm
Ian E
Sorry, what’s a job market paper? I get the idea that it’s like an arms race. I presume it’s a grad student’s research paper that they use as a sample? Or a resume?
February 13, 2012 at 5:27 pm
dan s
yes, a sample – the paper they submit in apps, discuss in interviews and present at job talks – but more than sample in that for many it’s their main or only research output when they go on the mkt
February 13, 2012 at 5:32 pm
brian
caltech has done a good job historically committing to the (4 then 5 year) short phd completion system, but probably has a quality profile more similar to yale than MIT. how? are they in cahoots with their cross country rival? i doubt it. i don’t have a good off-the-cuff answer, but they do seem to view it as a comparative advantage, both for their students on the market and from a recruiting point of view.
February 13, 2012 at 6:25 pm
Anonymous
Isn’t just a matter of selection/quality of the students?
Or, maybe, quality of advising?
February 16, 2012 at 1:24 pm
Drew Fudenberg
This year Harvard has 17 students on the job market; of these 12 finished in 5 years or less, including Azevdeo, Balidiga, Kamada, and Leshno. (theorists). In recent years all our theorists have taken 5 years or less, with Manea at 4 and Kominers at 2.
February 16, 2012 at 1:54 pm
jeff
Harvard = MIT 🙂
February 17, 2012 at 7:58 am
Drew Fudenberg
yeah, I guess looking east from Chicago all those Cambridge schools blur together 🙂
July 12, 2012 at 1:38 pm
k
Heh. There’s other competitors outside the Ivy League. Most students of these departments see their funding vanish after year 4 or 5. The goal changes to finishing a dissertation while holding a temporary job. So now we enter year 7.
Maybe it would be better to have every department report how long students take to finish, and require applicants to read this. This way many students on the margin will not choose to apply and soak up funds which might be better utilized funding other students.
Below the top 10-top 15 departments in the US, it’s a bleeding mess after 4 years with heavy costs – lack of funding, lack of a research environment, rotten social lives. I’m not exaggerating. It needs urgent attention – even at a department ranked say at #50, an average economics graduate student is an expensive investment.
February 23, 2013 at 3:40 am
A few random questions - Page 4
[…] data. Not true, unless if by "tops schools," you literally mean only 2 schools: Why MIT Students Finish in 5 Years and Everyone Else in 6 | Cheap Talk Reply With […]