The Interplay of Policy and Institutions during COVID-19

ABSTRACT

Are COVID-19 spread and mortality related to different countries’ government mobility restriction policies, democratic institutions, and cultural norms? Leveraging data from 140 countries, we find that policy, institutions, and vulnerabilities interact to determine pandemic spread and mortality. A delay in restricting international mobility increases pandemic mortality. Combining vulnerabilities with a delay in domestic mobility restrictions increases mortality. Democratic countries have faster policy responses and lower pandemic mortality, but they also face more adverse effects from a delay in restricting domestic mobility. More individualistic countries have a higher pandemic spread, and their delay in domestic mobility restrictions is associated with higher pandemic mortality.

KEYWORDS:

• Coronavirus
• pandemic spread
• mortality
• lockdown
• institutions
• culture

Acknowledgments

We thank George Clarke (editor), Bob Cull, Sebastian Galiani, and two anonymous referees for comments. Fang, Xu, and Yi have not received funding for this study. Peng thanks the Jindal Chair at UT Dallas. The views are the authors’ own and do not implicate the World Bank or its member countries.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

¹ See Alfaro et al. (2020), Barrios and Hochberg (2020), Borgonovi and Andrieu (2020), Brzezinski et al. (2020), Ding et al. (2020), Frey, Chen, and Presidente (2020), Galasso et al. (2020), Glaeser et al. (2021), Papageorge et al. (2021), and Simonov et al. (2020).

² In this article, we focus on the role of policy and institutions. Although health sector development, including health expenditure, is related to the outcomes of pandemic spread and mortality, it could be a function of some of our key variables, such as democracy, log GDP per capita, and the urban share. Moreover, given our focus, the relationship between the development of the health sector and the pandemic is beyond the scope of this study. We thank one of our anonymous referees for pointing this out.

³ Data available at https://github.com/OxCGRT/COVID-policy-tracker; see Hale et al. (2020).

⁴ Data available at World Development Indicators | DataBank (worldbank.org).

⁵ In particular, we measure the distance to Beijing in kilometers. Other functional forms do not make a significant difference.

⁶ That is, each country is classified as “rich in this continent” and “poor in this continent,” with the threshold being the mean GDP per capita of the latest year, and if the two cultural variables are missing, they are replaced with the mean values of the continent-income-level cell.

⁷ Data available at https://www.who.int/data/gho/indicator-metadata-registry/imr-details/2389.

⁸ Data available at https://sentinel.esa.int/web/sentinel/user-guides/sentinel-5p-tropomi.

⁹ We also control for the indicators for missing Trust and Individualistic Culture.

¹⁰ That is, East Asia and Pacific, Europe and Central Asia, Latin America and Caribbean, Middle East and North Africa, South Asia, and Sub-Saharan Africa. Since the U.S. and Canada are just two countries, we combine them with Europe and Central Asia.

¹¹ This is implemented in Stata following Belotti, Hughes, and Mortari (2017).

¹² We find little interaction effects between DelayI and the variables Democracy, Trust, and Individualistic Culture, and we, thus, again opt to focus on their interactions with DelayD.

¹³ To simplify, we base our selection on the specification using the full sample of 140 countries.

¹⁴ The coefficient of democracy has a similar magnitude but is insignificant in the non-imputing sample. We should focus on Column (1) here since the democracy variable does not need imputation, and the full sample result is more relevant.

¹⁵ The discussion here is based on the full sample. The qualitative results are similar except for Democracy$\times$SO2, in which case the full sample is more relevant anyway.