**4. Results and discussion**

The descriptive and correlation statistics for all variables for 120 countries over the period of 1985 to 2019 are reported in **Tables 1** and **2**. Real GDP (*gdp*) and government environment protection expenditure (*gov\_env*) are expressed in the logarithm. The correlation statistics in **Table 2** shows no high correlation amongst the variables.

**Table 3** displays the overall results for the whole sample and high-income countries. The lagged values of CO2 emission are consistently significant across all models in the overall sample and the high-income countries. The proxy for economic growth, *gdp* is significant except for eq. 4 but *gdp*<sup>2</sup> is negative but insignificant. The results suggest that economic growth continues to increase CO2 emission in the overall sample and high-income countries. The presence of an inverted *U*-shaped EKC could not be established based on these results. The rate of economic growth is positive and significant for all regression but the same could be generalized in the case of highincome countries. The positive and significant impact of financial development as proxied by domestic credit (*dom*\_*credit*) indicates how improvement in credit access, for example, promotes more production and hence, higher CO2 emission. Results are


#### **Table 1.**

*Descriptive statistics.*


#### **Table 2.**

*Correlation.*

consistent for the overall sample and high-income countries; results are significant for eqs. 1–3. International trade (*trade*Þ is not significant except for eq. 3 in the overall sample.

**Table 4** illustrates the results for upper-middle and lower-middle- and low-income countries. In the case of upper-middle-income countries, lagged term of CO2 emission is positively related to CO2 emission which is in line with the use of S-GMM. Evidence of an inverted *U*-shaped EKC is present implying the existence of a certain threshold point where reduction of CO2 emission takes place with better use of technology and more sustainable production techniques. These results are consistent with Le and Nguyen [13], Shahbaz et al. [14], and Le et al. [12]. More financial development implies higher emission of CO2. For lower-middle- and low-income countries, *gdp* is positive and significant and *gdp2* is negative and significant which suggests the existence of an inverted U-shaped relationship between economic growth and CO2 emission. Domestic credit and trade openness have the expected sign which both positive and significant, indicating the growth of trade and financial development leads to increased CO2 emission, hence, environmental pollution. Results are fairly consistent across all specifications.

