What is the ideal length of a job market paper?

Question

I am a 4th year PhD student, and I am substantially done with a paper (area: Political Economy, Theory) that my advisor thinks is very interesting and would be worthy of publication in a good journal if appropriately extended and revised. However, realistically, the paper would be about 25-30 pages including proofs. My advisor's input is that a really long JMP is not really "journal ready", and my advisor personally prefers shorter, cleaner papers.

On the other hand , it seems like virtually every job market paper is in the 45+ page category. I think a short paper might be good because it might be read more carefully. On the other, I'm worried that the page length itself is a signal, but I'm not really sure of what.

Answer 1

Short papers are great to read: the ideal length is as short as possible to convey what you've done.  (Published papers are often long because of add-ons requested by referees.)
That being said, job market papers are a bit special--you want to show what you can do.  People will examine your job-market paper not just for the new economics it contains, but also for evidence of your skills.
A _short paper with a long appendix _might be what's called for...it will make it easier for everyone to see what you have discovered, while allowing the theorists  to check your detailed proofs, the applied econometricians to appreciate your robustness checks,  etc. And when your paper is published, some of the appendices can be online...(although the magic of modern refereeing will make your published paper look somewhat different from your job-market paper no matter what you do.)

Answer 2

Would you have given a job to
   (1) John Nash for the equilibrium existence paper (just over 1 page),
   (2) Francis Crick or James Watson for the DNA structure paper (3 handwritten pages, 1 printed)?
Quality is (at least should be) the primary criterion. And for a given quality, quantity should have a negative marginal product.

Comments

If no one looks at your application or JMP, what is the effective quality of your application? Zero.

---

At least at the top levels, it won't be that people "look at" your application or JMP; they will hear and talk about it: "OMG, have you seen such and such a paper?"