Table A2 provides sample characteristics (descriptive statistics) of the good governance indicators and COVID-19 in the 172 countries included in our study. On average, 19,486 COVID-19 cases per million were reported by January 20, 2021, in the study sample, and a maximum of 120,469 of COVID-19 cases per million people was recorded by that time. Also, on average, 355 COVID-19 deaths per million people and a maximum of 1,775 COVID-19 deaths per million people in a country were reported. Additionally, on average, 0.9 COVID-19 vaccinations per hundred people and a maximum of 36.8 COVID-19 vaccinations per hundred people were reported by January 20, 2021 (see Appendix Table A2 for detailed descriptive statistics).
K-means clustering resulted in four clusters of countries based on their good governance data. Figure 3 depicts the map of clustering results of good governance indicators for each country using the K-means clustering algorithm and visualized using PCA. Good governance data was projected into the first two principal components, containing the most variation in the data (see Appendix Figure A1). A value of 87.1 in the x-axis (Principal Component 1) implies that the first principal component accounts for 87.1% of good governance variation. The second principal component accounts for 5.2% of the variation. In total, 51 countries out of 172 countries started COVID-19 vaccination approximately one month after the first public COVID-19 vaccination in the United Kingdom. Cluster 1 consists of 30 countries, mainly African and Middle Eastern countries, with poor governance indicators. None of the Cluster 1 countries started COVID-19 vaccinations. Cluster 2 was the largest cluster and consists of 77 countries with a mixture of relatively good and poor governance indicators. Only 11 of the Cluster 2 countries started vaccinations. However, Bahrain, with 8.47% of its population vaccinated, had one of the highest global vaccination rates. Cluster 3 consists of 41 countries with relatively acceptable good governance indicators, and 23 of those countries started COVID-19 vaccination. Israel, the United Arab Emirates, and Seychelles, which had the highest vaccination rates globally, belonged to this group, with vaccination rates of 36.76%, 21.85%, and 13.39%, respectively. Cluster 4 consists of 24 countries and had the highest good governance indicators. Seventeen of the Cluster 4 countries had started the COVID-19 vaccination in our study period. Also, countries with high rates of COVID-19 vaccination, such as the United Kingdom (8.01%) and the United States (4.94%), were in Cluster 4 (see Appendix Table A3).
Table A4 shows the Confusion Matrix and describes the binary classification prediction results. The classifier evaluation metrics, including accuracy (0.81), sensitivity (0.67), specificity (0.87), precision (0.68), F1 (0.67), and balanced accuracy, confirmed the accuracy of the model and the reliability of the predictions (see Appendix Table A5).
Figure 4 presents each good governance indicator's relative importance in predicting the COVID-19 vaccination status based on the XGBoost binary classification model. The regulatory quality in a country is the most important indicator in predicting COVID-19 vaccination status in a country, followed by voice and accountability, and government effectiveness. The relative contributions of these indicators are 22.7%, 22.2%, and 20.9%, respectively. Also, political stability, rule of law, and control of corruption have relative contributions of 14.8%, 11.4%, and 8%, respectively (see Appendix Table A6).
Results of sensitivity analyses using data up to March 2021 shows that, in total, 131 countries out of 172 countries started COVID-19 vaccinations approximately three months (100 days) after the first public COVID-19 vaccination in the United Kingdom. 11 of the Cluster 1 countries (out of 30 countries) started COVID-19 vaccinations. In Cluster 2, which is the largest cluster (77 countries), 58 countries started vaccinations. 38 countries of Cluster 3 started COVID-19 vaccination. Israel, Seychelles, and the United Arab Emirates (within Cluster 3) still had the highest vaccination rates globally, with vaccination rates of 116%, 102%, and 84%, respectively. All of the 24 countries of Cluster 4 that had the highest good governance indicators had started COVID-19 vaccinations approximately three months after the first public COVID-19 vaccination.
Figure 4 also presents each good governance indicator's relative importance in predicting COVID-19 vaccination status based on the XGBoost binary classification model approximately three months after first COVID-19 vaccine. The voice and accountability in a country is the most important indicator in predicting COVID-19 vaccination status in a country, followed by political stability, and regulatory quality, with relative gains of 21.9%, 19%, and 16.3%, respectively. Additionally, government effectiveness, rule of law, and control of corruption have relative contributions of 16.1%, 13.8%, and 13%, respectively (see Appendix Table A6).