levpet --- (1)r2000 and (2)Insufficient variation to identify...

[Shane Wu]

Hi All. I was running levpet for computing TFP of firms in my data. Two errors came up:

(1) levpet lny,free(lnl) proxy(lnm) capital(lnk) revenue reps(250) level(99)
...no observations
error #2000 occurred in program nllp_bs_gmm

i guess it's because firms in my data rarely have more than 6-year observation?
For a check, i use a data with firms all have 6-year observation run the order again. i got the second error:

(2)
Insufficient variation to identify the capital and intermediate input coefficients separately. Bootstrapping for first-stage parameters only.

Anyone knows how to fix it?
I've search the forum and got no answer. Call for help...Thank youuuuuuuuuu!!!

[Mina Wu]

Have you by any chance figured out the answer to any of the questions, in particular question #2? Ive got the same problem, and noticed few unanswered posts reporting it too.

[Erick Gonzales]

I read in a similar post that you should check if the nllp_bs_gmm.ado file is located in the right folder (folder n). Check if this is the case (C:\ado\plus\n).
I experienced that after checking that variables are numeric, the behavior of missing values and declaring the panel data, among others, the message "r(2000) no observations" kept appearing. What worked for me is to have consecutive years in the variable that sets the time for the panel data (xtset panelid year). My "year" variable was 2003 and 2009. When I changed this to consecutive values, for example, 1 and 2, the program worked. I wanted to share this just in case. If you can, please let us know if it works or how you solved the problem. Cheers!

[Julia Boltz]

Hi @Shane and @Mina,

you able to find the solution to problem #2 ("Insufficient variation to identify the capital and intermediate input coefficients separately. Bootstrapping for first-stage parameters only.") I have the same problem and have not found a solution to it.

I have a panel data from 5 different countries and if I ran levpet for each country individually, it works. I only get the problem for the combined dataset.

I would appreciate your help.

Best,
Julia