*4.3. Fitting Results for Hypothesis 2*

The results of the empirical calculations around Hypothesis 2 are shown in Table 5. Overall, the Hausman test values for Models 1, 2, and 3 were 21.22, 20.19, and 18.99 at the 1% significance level, respectively, indicating that the FE regression could be chosen. Specifically, in Model 1, the *RE* term has a coefficient of 0.414 in the FE regression and 0.284 in the GMM regression at the 10% significance level, indicating that *RE* has a significant contribution to *SG*. The paper then divides *RE* into SOEs and non-SOEs, which are analyzed in Model 2 and Model 3, respectively. According to Model 2 for SOEs, *RE* has a coefficient of 0.389 in the FE regression and 0.179 in the GMM regression at the 10% significance level. According to the results of Model 3 for non-SOEs, the *RE* term has a coefficient of 0.402 in the FE regression and 0.277 in the GMM regression at the 10% significance level. In summary, the results can be that Hypothesis 2 holds and the contribution of resource endowment to *RE* firms is more significant in non-state-owned firms.

**Table 5.** Measurement results of the relationship between *RE* and *SG*.


Note: \*\*\* *p* < 0.01, \*\* *p* < 0.05, \* *p* < 0.1.

## *4.4. Fitting Results for Hypothesis 3*

The results of the empirical calculations around Hypothesis 3 are shown in Table 6. This paper determines the effect of *RE* on the relationship between *ER* and *SG* by measuring the interaction term between *ER* and *RE*. Overall, the Hausman test values for Models 1, 2, and 3 were 28.08, 21.67, and 20.59, respectively, at the 1% significance level, allowing for the choice of FE regression. Specifically, in Model 1, the *ERA* × *RE* term has a coefficient of −0.016 for the FE regression and −0.31 for the GMM regression at the 5% level of significance. Model 2 shows that the *ERB* × *RE* term has a coefficient of −0.007 for the FE regression and −0.259 for the GMM regression at the 1% level of significance. Model 3 has the *ERC* × *RE* term at the 5% level of significance. In summary, Hypothesis 3 holds that *RE* is able to inhibit the facilitative effect between *ER* and *SG*, and *RE* has the most pronounced inhibitory effect on *ERA* and *SG* compared to *ERB* and *ERC*.

**Table 6.** Measurement results of the relationship between *ER* and *SG* with moderating effect of *RE*.


Note: \*\*\* *p* < 0.01, \*\* *p* < 0.05, \* *p* < 0.1.
