The result of the hausman test is that ein fixed effect regression is perferred and I found the problem of hetroscedasticity, autocorrelation and cross sectional dependency. N>T, a short panel data, I'm not sure which model I should use for the regression.
José-Ignacio Antón
N > T should not pose a problem. You can deal with heteroskedasticity using robust standard errors. Under N near 50, you could just use clustered standard errors (clustered at the unit level). You could still do that with certain caution. In that case, you just carry out xtreg , fe vce(cluster ...).
Alternatively, you have to solve the autocorrelation problem and avoid clustering standard errors at the unit level.
0 votes
0 thanks
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Probability
- Probability Theory
- Reliability
More Yishuang Lei's questions See All
Similar questions and discussions