Stationarity Panel data

[Lisa Stein]

hi,

I have a dataset with approximately 160,000 obs. and when I run a unit root test using this code: xtunitroot fisher item_35, dfuller lags(0), this error code appears:
"could not compute test for panel 259", on almost every panel.

How can I fix this problem?


Thank you.

[Wessel de Kroo]

Hi lisa,

it is helpfull to post your dataset using dataex , aswell as using your code see we can see quickly what is wrong with it.

You might want to consider that you might need the a different command to run unit root test on panel data. see for example the command xtfisher

hope this helps

[Phil Bromiley]

You'll increase your chances of a useful answer by following the FAQ on asking questions - provide Stata code in code delimiters, readable Stata output, and sample data using dataex. Note, we don't need your entire dataset - just enough to generate the problem.

My random guess is that your problem is using lags(0). The documentation says "The Fisher test assumes the data are generated by an AR(1) process" - if this is about an ar process, then 0 lags may make no sense. What happens if you use lags(1)? I'd also check that you have enough observations to do this in your panels - 2 observation panels probably won't work with this routine. Alternatively, are your data spaced as you said in xtset? I could see this kind of estimator crashing if you have xtset panel year and then had a lot of missing years.