Table 4.
Threshold and linear model estimation.
Variable | Threshold Estimation | FE | SYS-GMM |
---|---|---|---|
lnagg*I(q≦γ1) | −2.2145*** | ||
(−8.6486) | |||
lnagg*I(γ1 < q≦γ2) | −3.4455*** | ||
(−6.8035) | |||
lnagg*I(q>γ2) | −1.9292*** | ||
(−5.7364) | |||
lnagg | −0.5807779** | −1.136857*** | |
(−2.15) | (−5.48) | ||
lnfdi | 0.1697648*** | 0.0605472*** | |
(11.05) | (9.59) | ||
lntec | −0.8821*** | −0.4590924*** | −0.2184677*** |
(−12.7169) | (−6.39) | (−5.55) | |
lnstr | −1.5079*** | −0.1471305 | −1.182222*** |
(−7.7661) | (-0.64) | (−13.76) | |
lnreg | 0.0168 | 0.0385123* | 0.0009014 |
(0.6836) | (1.76) | (0.09) | |
cons | 0.119965 | 0.3791804 | −0.0006207 |
(−0.2) | (0.90) | (−0.00) |
The t or Z statistics are in parentheses in the table. “***”, “**”, “*” indicates significance levels of 1, 5, and 10% respectively. SYS-GMM is two-step.