Hi all,

I am working on a probability model by utilising a panel probit regression. One of the independent variables of this model is actually generated by another regression. Specifically, it is the residuals of another regression.

Now I am wondering the outcome of the first regression and the probit regression could be endogenously determined. In this situation, I need to perform propensity score matching, right? Please see the syntax I used:

Code:
. teffects psmatch (MarketLeveraget) (AcquirerStatus LeverageDeficit Size Markettobook Profitability stockre
> turn i.Year, probit)
To run the probit regression I used -xtprobit- command as the data I have is panel. But I can't use xtprobit even though I can add the time dummy (-i.Year-). Is this a problem?

Here is what Stata gave me back:

Code:
Treatment-effects estimation                   Number of obs      =      2,539
Estimator      : propensity-score matching     Matches: requested =          1
Outcome model  : matching                                     min =          1
Treatment model: probit                                       max =          1
--------------------------------------------------------------------------------
               |              AI Robust
MarketLevera~t |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
---------------+----------------------------------------------------------------
ATE            |
AcquirerStatus |
     (1 vs 0)  |  -.0270506   .0098036    -2.76   0.006    -.0462653   -.0078358
--------------------------------------------------------------------------------
What exactly shodul I infer from this table? Is the number of matches requested problematic? I can clearly see that there is a relationship between the AcquirerStatus and MarketLeverage. Is this a problem?

Many thanks!