Please, I'd like to know whether Driscoll-Kraay standard error model approach (Stata command XTSCC) able to handle non-stationary I(1) panel data sets? In other words, is this a requirement for panel to be stationary in order to get consistent results by applying this method?
For instance, the panel data is DV(1); key IV(0) and the rest control variables are IV(1)
Would be glad to hear and learn from your expertise.
Thanks beforehand.
Samy Azer , Greetings! Thank you for your response, with due respect, but I have to say that I watched the video you shared and quite sure it does not give me any answer on the issue I raised in the question. Well, it does talk about robust standard errors, however it neither talks about stationarity of the variables, not it talks about the particular XTSCC Driscoll-Kraay approach for the panel data sets. But I'm still wondering why you recommended it in this situation and I stand to be corrected if I missed something you wanted to say based on that video. Thanks anyways.
Elina Kim Thank you for your message. I am sorry if I have sent a material not precisely matching your question. I am sending you a paper by Hoechle D titled, Robust standard errors for panel regressions with cross-sectional dependence. I hope this helps with your question.
I have come across an interesting study by Pazienza (2015) where he applied Driscoll-Kraay standard errors to his model. Apparently, in addressing non-stationarity of variables, he controlled it by simply using first difference of the variables that are stationary at 1st level differencing.
Reference:
Pazienza, P. (2015). The relationship between CO2 and Foreign Direct Investment in the agriculture and fishing sector of OECD countries: Evidence and policy considerations. Intellectual Economics, 9(1), 55–66. https://doi.org/10.1016/j.intele.2015.08.001
- Badges
- Science topic
- Similar topics
- Psychology
- Mental Processes
- Learning