Green Microfinance in Europe

Austria, Belgium, Bulgaria, Cyprus, Croatia, the Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, the Netherlands, Poland, Portugal, Romania, Slovakia, Slovenia, Spain, Sweden, the United Kingdom; Iceland, Montenegro, Republic of Macedonia, Serbia and Turkey; Albania, Bosnia and Herzegovina, Kosovo.

To test the normality, we plot the distributions for the square root of the global MEPI and of the MEPIs for the five environmental dimensions, for the two populations for all the nine categories previously stated, against a normal distribution, we employ a Q–Q plot that compare the different quantile of the distributions and we perform a Shapiro–Wilk normality test. All these tests turned out to be satisfied at an adequate level of accuracy, for almost all the subsamples.

We observed two potential outsiders and as robustness check we performed four regressions: one with both potential outsiders, two without one of the potential outsiders; and one without both outsiders; and we verify that the results for the level of significance of the various coefficients are robust. We then proceed to check the normality of the residuals. We observed that the residuals for the regression using the global MEPI as independent variable do not have normal distribution. We then decided to use the square root of the MEPI as independent variable. We then verified that these residues follow a normal distribution by drawing a Q–Q plot and doing a Shapiro–Wilk test that turned out significant. We then checked the absence of heteroscedasticity using the White's test and the Breusch–Pagan test. The Variance Inflation Test checked the absence problems related to multicollinearity. We then performed a couple of simple tests to check the absence of problems related to omitted variables. We also checked that the residues has zero expectation value.