Problem in stationarity tests

[Salma Gallas]

I am working with a panel model and I am looking to see if the variables are stationary. With a study of 11 countries and quarterly data for 2000 Q1 -2018Q4. With the Levin-Lin-Chiu tests, I get this output "Levin-Lin-Chiu test requires strongly balanced data" r (498) and "Levin-Lin-Chiu test cannot have gaps in data"
r (498) for other variables. I didn't quite understand what to do. thank you for helping me.

[Øyvind Snilsberg]

The error message says that each country must have the same time points, i.e., one observation for each quarter from 2000Q1-2018Q4.

[Salma Gallas]

Yes that's what I've done for my database. For each country, the observations start 2000 Q1 and end 2018 Q4

[George Ford]

If the data is truly balanced, then you may have an error in the dates. Might create a time variable and see if that works, but that may be tricky if the dates are invalid.

try

unique date

to see if the math works (answer should be periods, or equal to the observation count/countries). This is not determinantive, since you may have replicate dates for a country. Could do unique for each country to see if you get the number of periods for each one.

[George Ford]

This might help as well. Collapse the data by the date var and then look at what dates are there to detect any irregularities.

collapse (mean) pickanyvar , by(datevar)

[Salma Gallas]

I did not quite understand what you mean. I tried all the different stationarity tests but no results. I think my database is well done, the dates are common for all countries, I have missing data for two types of variables. I'm still looking for an explanation of the STATA outputs: Im-Pesaran-Shin test cannot have gaps in data r (498).

[George Ford]

That means your times series has gaps (skipped quarters). The missing values might cause that too. It's a data problem, not a stata problem.