Table 3.
Benchmark regression results.
| GTI |
Multicollinearity test |
|||
|---|---|---|---|---|
| RE | FE | VIF | Tolerance | |
| EGC | 0.0919*** (9.56) | 0.0528*** (4.39) | 1.080 | 0.925 |
| SIZE | 0.1665*** | 0.2454** | ||
| (4.78) | (2.31) | 1.670 | 0.600 | |
| ALR | −0.0785 | −0.1964 | ||
| (-0.44) | (-0.49) | 1.550 | 0.647 | |
| NPM | 0.3427*** | 0.4473** | ||
| (2.95) | (1.89) | 1.090 | 0.918 | |
| FAR | −3.7870*** | −2.5001*** | ||
| (-15.55) | (-3.65) | 1.090 | 0.921 | |
| TQ | 0.0801*** | 0.0205 | ||
| (3.48) | (0.56) | 1.210 | 0.830 | |
| MS | 0.0059*** | −0.0031 | ||
| (7.04) | (0.73) | 1.200 | 0.831 | |
| Cons |
−5.3230*** (−7.04) |
– |
– |
– |
| Individual fixed effect | No | Yes | (Mean) 1.27 |
|
| Time fixed effect | No | Yes | ||
| Industry fixed effect |
No |
Yes |
||
| Number of samples | 12246 | 4708 | – | – |
| Hausman | 38.79 (0.00) | – | – | |
| Wald-P | 0.00 | 0.00 | – | – |
Note:(1)When using Stata software to estimate panel Poisson's counting models, there is no constant term estimation value for fixed effects. (2) The difference in samples number is automatically adjusted by Stata software. (3) Robust standard error is used. (4) The values in parentheses are z-statistics. (5) * * *, * *, and * represent significance at the levels of 1 %, 5 %, and 10 %, respectively.