Forecasting the Effectiveness of COVID-19 Vaccination Using Vector Autoregressive with an Exogenous Variable: On the Cases of COVID-19 in Indonesia

Abstract

This study aims to forecast the COVID-19 spread in Indonesia involving vaccination factors using vector autoregressive with exogenous variables (VARX). The COVID-19 spread represented by active, recovered, and death case rate indicators acts as endogenous variables, while the COVID-19 vaccination represented by second-dose vaccination rates acts as exogenous variables. Because the sum of three COVID-19 spread indicators in one day is one, only two indicators with the highest correlation rates are involved in VARX modelling. The other indicator is practically projected by subtracting one from the sum of two indicator projection results. Based on the analysis results, the active and recovered case rates are two indicators chosen in VARX modelling. Using Akaike information criterion, the most suitable VARX model to project the case and recovered case rates are VARX (7, 1). This model is expected to help the Indonesian government project the COVID-19 spread in Indonesia.