Thread by @Jabaluck

A major difference between econ and epi is that econ is more authoritarian while epi is more democratic. This has positive and negative consequences @epiellie

https://twitter.com/EpiEllie/status/1450454678748254208

Econ is authoritarian in the sense that well-respected senior economists are the gatekeepers to all the top journals. If you want your paper to be widely read, you need to get it past an editor who is a leading authority in a (usually) related field.

This means that for a methodological innovation to diffuse, only a few top people need to be convinced that it is worthwhile and they will ask for it as a condition of publication in the top journals.

Additionally, papers are written to appeal to the very best people rather than the lowest common denominator. My experience talking to epi people is that many decisions are made to avoid confusing readers who are bad at stats.

This is not because most epi readers are worse at stats, but because epi authors might need to get their paper past referees and editors with a weak stats background. This doesn't happen so much in econ for editors at top journals.

An example is balance tests where the repeated feedback I got from epi people was that it's bad to report balance tests for RCTs not because they are not informative, but because they might lead a paper to be wrongly rejected.

https://twitter.com/Jabaluck/status/1428859216580812803

https://twitter.com/Jabaluck/status/1429499856276578322

The downside of authoritarianism is that it is harder to get work published that doesn't fit the established mold. In epi, my experience is that journal editors and referees come from a wide variety of backgrounds with different expectations.

In econ, there is a lot of homogeneity in what top journal editors are looking for. Raj Chetty can get great new data and publish a descriptive paper with some new stylized facts, but a grad student who did the same thing would get hammered by referees and editors.

There are real advantages to the authoritarian model --- good papers are relatively rare, and it's good that we have a few journals with high standards enforced by very knowledgeable gatekeepers to help reduce search costs.

A major unknown is the cost of authoritarianism -- how much truly innovative work disappears entirely because it never receives sufficient exposure.

I haven't seen a systematically analysis quantifying this trade-off and obviously it would be difficult to do, but these kinds of organizational issues are understudied as drivers of scientific progress (see, e.g. @carolyn_sms and @RyanReedHill work)