Estimating the impact of COVID-19 on mortality using granular data

Abstract

We present an extension of the Li and Lee model to quantify mortality in five European countries during the COVID-19 pandemic. The first two layers specify pre-COVID mortality, with the first one modeling the common trend and the second one the country-specific deviation from the common trend. We calibrate this part of the model using annual data from 1970 to 2019 and then add a third layer to capture the country-specific impact of COVID-19 in 2020 and 2021. The calibration of the added layer is based on data with a higher granularity in time, since we analyze weekly instead of annual data. We also investigate whether estimates improve if we increase the granularity over the ages, utilizing data we obtained for single ages instead of the usual aggregated age groups. We complement our analysis by presenting mortality forecasts based on different possible scenarios for the future course of the pandemic and a backtest in which we compare predictions of Dutch mortality improvements from 2021 to 2022 against their realizations. The results from this backtest can be used to update mortality forecasts as new observations become available.