The economic reaction to non-pharmaceutical interventions during Covid-19
Segarra, A., Teruel, M.: "The economic reaction to non-pharmaceutical interventions during Covid-19", Economic Analysis and Policy
Countries have applied different economic policy support as a response to the application of non-pharmaceutical interventions during Covid-19 pandemic
There appear differences between rich and poor countries and more and less touristic countries.
Our results show how each non-pharmaceutical intervention (i.e. the cancelation of public events or the closure of public transports) implied a different level of economic support and how structural characteristics, such as GDP or dependence on the tourism industry, were relevant to the decision-making process
These findings constitute an insightful ex-post analysis for policy-maker in the context of so-called "lock-down policies". First, our results show that governments responded quickly to the economic shock. Second, their economic stimulus was more intense in richer countries and those which have a sectoral structure more vulnerable to the pandemic, this being particularly among countries highly dependent on international tourism. Furthermore, it is crucial that countries continue with the coordination of different policies to ensure a balanced recovery. Third, we observe the strong interlinkages between those interventions to control the pandemic and the economic answer given by governments. These results reveal an initial "do whatever it takes" policy to contain the pandemic and avoid damaging the economy in the short term. However, once the recovery phase arrives, policymakers adjust their policies by trying to compensate for the unequal impact on different social groups and economic activities.
This research has been carried out by Agustí Segarra Blasco, Mecedes Teruel and Sebastiano Cattaruzzo of the Department of Economics of Universitat Rovira i Virigili. The authors belong to the Industry and Territorial Research Group and to ECO-SOS.
We have made three relevant contributions in this empirical study. Firstly, at an empirical level, although many studies have used daily time economic support data, in this work, the time dimension has been fully exploited in the analysis of economic responses. On the contrary, previous contributions, such as Khalid et al. (2021) and Elgin et al. (2020), analysed the same phenomenon aggregating the time dimension into an average level. Economic support is crucial in avoiding social breaks and ensuring aggregated demand, so data at a more precise time scale facilitates better analysis of the policies deployed during the current pandemic (Graph 1).
Graph 1. Share of countries that deployed economic support index according to the degree of intensity
Second, we highlight how responses have been country-dependent. The pandemic has dramatically affected all countries, but the reactions are asymmetric depending on their individual characteristics. Dosi et al.'s (2020) hypothesis is confirmed here, as our results show that heterogeneous initial conditions affect individual governmental responses to the pandemic. For instance, although Covid-19 affected the tourism industry at a global scale, the policy response in terms of speed and depth differed according to the economic framework. Finally, we contribute to the analysis of the differences in the NPIs applied to reduce social interactions during the outbreak and in public aid to households. NPIs are necessary to control the pandemic, but they constrain the economy by affecting the purchasing power of households. Our results highlight that the imposition of NPIs is quite heterogeneous and its economic response also.
The discrete non-negative nature of the dependent variable generates non-linearities that make the usual linear regression models inappropriate and forces to one opt for counting models. Consequently, we opted for the model proposed by Wooldridge (1999, 2000) that implies the estimation of robust standard errors clustered at the country level. This specification is fully robust to violation of the strict exogeneity assumption (Wooldridge, 2000) as well as unmodeled serial correlation in the dependent variable within a country. Moreover, in panel data, the presence of country-specific "unobserved heterogeneity" (Wooldridge, 2005) such as the internal government dynamics and the extent of regional differences are undeniable, and these unobservable factors influence the way in which countries react to pandemics and make decisions to support the economy. Finally, this modelling choice is also helpful in minimizing potential biases that derive from cultural and social normative differences among those geographical areas under consideration.