Robustness test: Sample selection bias – Heckman’s (1979) Two-stage model
| First stage | Second stage | ||
|---|---|---|---|
| DIS | EIS | EIS | |
| DV= | (1) | (2) | (3) |
| RDI | 42.795***(4.01) | −0.015 (−0.80) | |
| ESGS | 0.644*** (6.83) | ||
| ESGS × RDI | 0.013**(2.18) | ||
| FS | 1.066*** [5.98] | 47.431*** (5.85) | 29.672*** (3.91) |
| LVG | −0.608 [−1.51] | −27.760*** (−3.69) | −27.438*** (−3.99) |
| POV | −0.446 [−1.47] | −21.918*** (−4.25) | −16.817*** (−3.34) |
| CH | 0.763** [2.26] | 25.229*** (3.50) | 23.755*** (3.78) |
| AT | 0.601*** [3.89] | 24.480*** (4.94) | 16.322*** (3.80) |
| AG | −0.103 [−1.38] | −4.554*** (−3.80) | −2.905** (−2.50) |
| IMR | 42.795*** (4.01) | 34.110*** (3.76) | |
| Intercept | −10.581*** [−5.86] | −475.150*** (−5.51) | −313.468*** (−4.00) |
| Year fixed effects | Yes | Yes | Yes |
| Industry fixed effects | Yes | Yes | Yes |
| Pseudo R2/R2 | 0.315 | 0.490 | 0.577 |
| Obs. | 3,035 | 1049 | 1049 |
| First stage | Second stage | ||
|---|---|---|---|
| DV= | (1) | (2) | (3) |
| −0.015 (−0.80) | |||
| 0.644 | |||
| 1.066 | 47.431 | 29.672 | |
| −0.608 [−1.51] | −27.760 | −27.438 | |
| −0.446 [−1.47] | −21.918 | −16.817 | |
| 0.763 | 25.229 | 23.755 | |
| 0.601 | 24.480 | 16.322 | |
| −0.103 [−1.38] | −4.554 | −2.905 | |
| 42.795 | 34.110 | ||
| Intercept | −10.581 | −475.150 | −313.468 |
| Year fixed effects | Yes | Yes | Yes |
| Industry fixed effects | Yes | Yes | Yes |
| Pseudo | 0.315 | 0.490 | 0.577 |
| Obs. | 3,035 | 1049 | 1049 |
Model 1 reports the first stage of Heckman’s (1979) Two-stage model, a probit regression model with the dependent variable (disclosure [DIS]) equal to 1 for firms that have ESG data, and 0 otherwise. Models 2 and 3 present the second-stage regressions, a baseline model regression including the inverse Mills ratio (IMR) which is calculated from the first stage. All regressions are estimated with clustered robust standard errors by firm and include year and industry fixed effects. DV = dependent variable. EIS = environmental innovation score. The t-statistics (z-statistics) are reported in parentheses (brackets). Superscript *, **, and *** indicate significance at 10, 5, and 1% levels, respectively. Table 2 presents the definitions of variables
Sharing content requires targeting cookies to be enabled. Please update your cookie preferences to use this feature.