sergiocorreia / reghdfe

Linear, IV and GMM Regressions With Any Number of Fixed Effects
http://scorreia.com/software/reghdfe/
MIT License
219 stars 57 forks source link

[BUG] Backwards compatibility broken for Stata < 14.2 #253

Open mcaceresb opened 2 years ago

mcaceresb commented 2 years ago

[BUG] Backwards compatibility broken for Stata < 14.2

Bug report

Behavior

. reghdfe
invalid syntax
(error occurred while loading reghdfe.ado)
r(198);

Expected output in this case would be

. reghdfe
last estimates not found
r(301);

which is the case in Stata 14.2+. I have only found the above two issues.

sergiocorreia commented 2 years ago

One question is if we should care or not (I haven't made my mind..)

Google says that 14.2 was released back in 2016, almost 6yrs ago, but more importantly, anyone with Stata 14 should have been able to upgrade freely to 14.2 by now, so the relevant date of 14.0 was April 2015 (7yrs ago). So I'm guessing you are running something on a server that uses Stata 13?

As a comparison, in R-land ggplot requires R>=3.3 which was released in 2016.

In terms of alternatives, I can think of three:

  1. Fix it. I guess we could replace the adopath option with two lines (one findfile, one include). Annoying, but not that difficult
  2. Ignore it and bump required version to 14.2
  3. Tell people to use the old version of reghdfe

Thoughts?

mcaceresb commented 2 years ago

@sergiocorreia Yes, sadly I am using Stata in a server that only has 14.1. Assuming these are sensible options to have, if I am being honest I would lean to option 2 and note somewhere what the last reghdfe release that works with 14.1- is.

1 seems annoying for what is admittedly a corner case.