Assessing alternative indicators for Covid-19 policy evaluation, with a counterfactual for Sweden

We compare different indicators of the spread and consequences of the COVID-19 pandemic, developing a novel method to adjust daily COVID-19 deaths to match weekly excess mortality. Focusing on Sweden, the only country that has good data and did not impose a lockdown, we construct counterfactuals for what would have happened if it had imposed a lockdown, using a synthetic control method. Correcting for data problems and optimizing the synthetic control for each indicator considered, we find stronger effects than previously estimated. Most importantly, studying the ratio of positives to the number of tests, we find that a lockdown would have had sizable effects already after a week. The 3-4 weeks delay highlighted in previous studies appears mainly driven by the large changes in testing frequency that occurred in Sweden during the period considered.