Panel Data Estimation of Driscoll–Kraay standard errors with plug-in estimators (Optimum number of lags)

[Sorphasith jumong]

Dear Experts and Researchers, Would you please suggest on how to estimate Driscoll–Kraay standard errors with plug-in estimators (automatic optimum number of lags) or what are the command to estimate this procedure with automatic lag selection that proposed by Andrews (1991) and Newey and West (1994). Many thanks in advance

[Fei Wang]

xtscc (from SSC) estimates DK standard errors with the number of lags, by default, being selected according to Newey and West (1994). But note that the rule of lag selection is simpler than the paper and tends to give shorter lags than the optimum.

[Sorphasith jumong]

@Fei Wang could you please provide full command and example ? Many thanks

[Fei Wang]

Please run the code below to install xtscc

ssc install xtscc

and see instructions and examples in the help file and the Stata Journal article.

[Sorphasith jumong]

@Fei Wang Many thanks, I am still confused. So, could you please give an example of maximum lag selection from this procedure: m(T) = floor[4(T/100)^2/9]. Thanks in advance

[Fei Wang]

For example, if the length of periods in your panel data is 40 (T = 40), and then,

m(T) = floor[4(T/100)^2/9] = floor[4(40/100)^2/9] = floor[3.263] = 3

[Sorphasith jumong]

@Fei Wang Got it, Many thanks