Hello all,

I have recently come across to xtdcce2 and would like to ask some questions.

- If I use xtdcce2 d.y d.x d.z, lr(L.y x z) p(L.y x z) nocross , does this mean it will be a regular pmg estimation with (1 1 1) lag structure? When I estimate like this, there is a huge difference between the coefficients and their statistical significances compared to what I get from xtpmg.

-Does xtdcce2 allow an estimation with a different lag structure? For example, can I use d.y d.x, lr(L.y x z) p(L.y x z) if I would like to estimate it (1 1 0)?

-As I understand from regular pmg estimation, the speed of adjustment parameter is heterogenous; therefore I took out L.y from the pooled bracket - although I am not sure if I can do that - , as a result the coefficients and p values changed, but they are very still different from what I get from xtpmg. As I read the 'help xtdcce2' I see the long run is estimated through ols rather than ml, but I'm not sure if this estimation difference would create such a big difference. Therefore I couldn't really understand what could be going wrong.

-Another related question to the previous one, when I take out the speed of adjustment parameter from the pooled bracket and use 'showindividual" I see the individual coefficients and their probs for x and z, but not for the adjustment parameter. Is there another way to see the individual adjustment parameters?

-Related to the individual results, I see almost all of them highly insignificant while the means are all significant no matter if I use nocross or crosssectional(_all). Is it really possible to have all insignificant individual x and z's while being significant in the mean?

Many thanks in advance if you have any guidance for any of my questions.