Bias-corrected State-by-state Forecasts of COVID-19 Deaths

Since testing for COVID-19 infections is not at all randomized over the general population, most epidemiological model forecasts of deaths are subject to `selection bias.' This paper updates and supplements Vinod and Theiss (2020), where the bias correction using generalized linear models (GLM) and inverse mills ratio (IMR) are described in detail. We include state-by-state forecasts using Poisson regression to predict one-week-ahead cumulative deaths from logarithms of current cumulative infections. We hope that the details provided here will help local governors and mayors in their opening up decisions