COVID-19 in 38 countries: a structural breaks approach

In this paper we estimate econometric models of daily data on the number of new cases from COVID-19. Our primary purpose is to test for the presence of structural breaks in each time series and then to incorporate these breaks in the estimation process. Government Response Stringency Index is also included as explanatory variable and interacting with the factor variables resulting from the breaks. We found different number and different dates for the structural breaks across the countries. We also observe that the last structural break is more or less contemporaneous for most of countries (24 in 38). By considering the curvature of the quadratic function we can guess if the peak already occurred. Latin America, Australia, Asia and Africa countries seem to reach already the peak. The same conclusion does not apply to European countries, USA and Canada. Including the GRSI increases the fit significantly.