A series of recent studies argue that the importance of the role of policy mandated NPIs in shaping the progression of the pandemic might have been overstated. Atkeson et al. (2020) show that in the early phase of the pandemic, transmission rates of Covid-19 declined almost universally worldwide, whereas region specific NPIs varied in their degree of severity. While NPIs typically involved closing non-essential shops and banning gatherings of more than two people, Sweden only forbid gatherings of more than 50 people. Germany only advised to reduce social contacts. Other countries such as France, Italy or Spain strictly ordered citizens to stay at home (Born et al., 2021).
Despite differences in NPIs between Germany and Sweden, a glance at mobility trends indicates similar dynamic behavioural adaptations of individuals as can be seen in the Figure below. The number of visitors to transit stations, workplaces, and places of retail and recreation declined sharply in both countries in March 2020 and rebounded only in the summer months, except for the number of visitors to workplaces in Sweden, which initially lagged behind the increase in Germany but since then roughly follow the same pattern. Despite the different levels of retail visitors in Sweden compared to Germany, the trajectories are strongly correlated as well. In Sweden, therefore, there were similarly strong declines in mobility as in Germany, although no comparable restrictions were imposed in the spring of 2020. Moreover, people seem to adjust their behaviour in anticipation and reduce social contacts even before the introduction of NPIs. Mobility trends during the second Corona wave are very similar: A decline in activity from the end of September and a renewed increase in activity since the beginning of 2021, with another dip in April 2021.
Measurement issues aside this suggests that other common factors across regions might have contributed to transmission rate declines and possibly bias the efficacy of NPIs:
- Distinction between lockdown (elimination of virus transmission domestically) and isolation (border closure to prevent trans-border movements)
- Measurement biases due to omitted variables
- Distinction between effects of natural pandemic-related developments and government induced restrictions
At this point, there is still little evidence to clarify conclusively why some nations like New Zealand, South Korea, or Taiwan are less affected by the pandemic. One hypothesis is that island states can isolate themselves better. However, countries such as Finland, Norway or South Korea achieved similar results by restricting cross-border movement significantly whereas the UK suffered severely from Covid-19 infections during 2020. Recent studies have argued that other country-specific conditions such as population density and age are likely to play an important role in the progression of the pandemic (Allcott et al., 2020). Another omitted variable that potentially positively biases the presumed efficacy of lockdown restrictions is voluntary dynamic behavioural change as indicated by the comparison of mobility trends of Germany and Sweden. Born et al. (2021) also suggest that effects of voluntary social restraint are likely to have been underestimated in the case of Sweden.
Consequently, an endogenization of these omitted variables is advisable in future research to better assess the impact and effectiveness of NPIs. Current research generally lacks the inclusion of these factors.
More issues arise when trying to quantify the economic costs caused by NPIs. Most studies estimate short-term trade-offs between health harms and forgone economic growth. However, there are other relevant costs that – depending on where we are on the time horizon axis – must also be considered. Among the most far-reaching from an economic perspective include loss of education potential, distributional inequality, and (more permanent) structural changes post-Corona. The first and second issue are related to each other because less education particularly affects households who are socio-economically disadvantaged and thus also fuels into inequality issues. The third factor concerns structural changes in the economy, caused by adjustments in consumption and investment behaviour. Due to the still insufficient data basis in Germany, the real and more substantial effects on the economy and society can only be assessed post-corona.
Nevertheless, they already clearly show that the total effect of the trade-offs is not only between health-related costs and lost GDP in the same period. Hence, research that only considers a bivariate relationship between GDP loss and Covid-19 deaths has not yet adequately addressed the important long-run effects of the crisis on society and the economy.
Admittedly, doing a comprehensive cost-benefit analysis is incredibly difficult since there are many subjective values at play and need to be aggregated. Often, cause and effect are conflated, and mono-causal explanations tend to trump more complex approaches. Nevertheless, given the major restrictions of NPIs on social and economic life clarifying these trade-offs, even under uncertainties and with caveats, is crucial.