Number of Companies = 10, 

Time per Company =9, 

Number of Independent Variables = 6

I used First Difference GMM with results as follows:

2. The coefficient of the lag of the DV which is used as an independent variable is less than 1 and statistically significant (p<.001).

3. AR(1) is significant (p<.05)

4. AR(2) is nonsignificant (p>.05)

6. Wald Test is significant (p<.01).

Since #1 and #5 are not OK, then First Difference GMM is inappropriate.

I tried the Two-Step Difference and the Systems GMM and all are not OK. Can you suggest an alternative?