FMOLS - positive coefficient and negative t-stat, is it possible?

[johny kos]

Dear Statalists,

I would like to ask you how it is possible after running a FMOLS model using the xtcointreg command to have a positive coefficient and a negative t-stat.
I have attached you an image of my model.
I am not sure if my command has any problem or not.
Thank you in advance.

[Jeff Wooldridge]

johny: Please show what you typed and the Stata output in between code delimiters. I'm making a habit now of not opening attachments because it has resulted in a computer virus recently. I haven't used the xtcointreg command but I might be able to help.

[johny kos]

Dear Professor,

I would like to thank you for your response. It honours me.
Of course, I totally understand. Below, I have written the 3 codes I used for my models:
For the FMOLS:

xtcointreg lnefpc  lngdppc lngdppc2 lnrenpc eci kofgi trade , est(fmols)  eqtrend(0) xtrend(0)  stage(1) kernel(bartlett) bmeth(neweywest)

For the CCR:

xtcointreg lnefpc  lngdppc lngdppc2 lnrenpc eci kofgi trade , est(ccr)  eqtrend(0) xtrend(0)  stage(1) kernel(bartlett) bmeth(neweywest)

For the DOLS:

xtcointreg lnefpc  lngdppc lngdppc2 lnrenpc eci kofgi trade , est(dols)  eqtrend(0) xtrend(0)  dlead(1) dlag(1) vlag(2)  stage(1) kernel(bartlett) bmeth(neweywest)

Following the same commands, I run several different models and this case (of opossitve signs for beta and t-stats) are observed

[Lina Nova]

johny: Please show what you typed and the Stata output in between code delimiters. I'm making a habit now of not opening attachments because it has resulted in a computer virus recently. I haven't used the xtcointreg command but I might be able to help.

Dear Professor,
Thanks for your previous response, I got similar results as the ones exposed before. Is any way to interpret this?

[Jeff Wooldridge]

As I said previously, I need to see the commands you typed and what Stata reported.