Elliott (1999) Pt test

[mike80]

Dear Tom,

I compared the test statistics for the PT test of Elliott (1999) obtained by the PerronNgMtest procedure with that obtained by the erstest procedure. For my data, they differ slightly. Can you help me to understand where the difference is coming from? The commands are:

@erstest(print,lags=%%autop,det=constant) var start end
@PerronNgMTests(gls,lags=%%autop,det=constant) var start end
disp %%pt

Did I forget something to do?

Thanks in advance!

Mike