Using the SARIMA Model to Forecast the Fourth Global Wave of Cumulative Deaths from COVID-19: Evidence from 12 Hard-Hit Big Countries

The COVID-19 pandemic is a serious threat to all of us. It has caused an unprecedented shock the world’s economy, and it interrupted lives livelihood millions people. In last two years, large body literature attempted forecast main dimensions outbreak using wide set models. this paper, I short- mid-term cumulative deaths from in 12 hard-hit big countries around world as 20 August 2021. data used analysis were extracted Our World Data dataset. Both non-seasonal seasonal autoregressive integrated moving averages (ARIMA SARIMA) estimated. showed that: (i) ARIMA/SARIMA forecasts sufficiently accurate both training test by always outperforming simple alternative forecasting techniques chosen benchmarks (Mean, Naïve, Seasonal Naïve); (ii) SARIMA models outperformed ARIMA 46 out 48 metrics (in future values), i.e., on 95.8% considered accuracy measures (mean absolute error [MAE], mean percentage [MAPE], scaled [MASE], root squared [RMSE]), suggesting clear pattern data; (iii) forecasted values fitted very well observed (real-time) for period 21 2021–19 September 2021 almost analyzed. This article shows that can be safely medium-term predictions deaths. Thus, approach help government authorities monitor manage huge pressure exerting national healthcare systems.