key: cord-0890338-xc17hhwj authors: Chae, Seung Hoon; Kim, Wukki; Park, Hyung Jun title: At odds? How European governments decided on public health restrictions during COVID-19 date: 2022-02-15 journal: Public Health DOI: 10.1016/j.puhe.2022.02.001 sha: f23909c5213983dccfc0346eec8c02eb29cf6b76 doc_id: 890338 cord_uid: xc17hhwj Objectives To understand how politics, economics, and public health restrictions affected each other during the COVID-19 pandemic. Methods We employ seemingly unrelated regressions (SURs) upon a monthly dataset of government approval ratings, the stringency index, the time-dependent reproduction number (R), and unemployment, allowing the residuals in each regression to be correlated with each other. We also conduct sensitivity tests using weekly data and the growth in polls. Results The study covers 27 European countries from April 2020 to April 2021. A unit increase in the R and COVID-19 cases per million increment the stringency index by 23·742 and 4·207 respectively; a unit increase in stringency boosts the incumbent’s popularity by 0·384; the poll positively affects the stringency index; stringency has negative effects on the R; the poll and stringency index have opposite effects on unemployment. Results The study covers 27 European countries from April 2020 to April 2021. A unit increase in the R and COVID-19 cases per million increment the stringency index by 23.742 and 4.207 respectively; a unit increase in stringency boosts the incumbent’s popularity by 0.384; the poll positively affects the stringency index; stringency has negative effects on the R; the poll and stringency index have opposite effects on unemployment. Conclusion Political and economic pressures did not hinder the government from introducing stronger measures. 1 Background 1 COVID-19 has had an unprecedented impact on public health, causing a mass outbreak 2 accompanied by a considerable number of deaths. Where the number of COVID-19 patients 3 soared past the capacity of a nation's health care system, COVID-19 patients could not be 4 treated with proper management, and the death rate increased steeply. 1 Moreover, the medical 5 repercussions of the pandemic reached beyond the patients who were directly affected by the 6 disease: by limiting patient access to hospitals, the pandemic has led to delays in treatment and 7 diagnoses for diseases other than COVID-19 as well. 2 8 Containing the spread of the virus induced disease was, therefore, a central task for 9 many governments across the world in 2020 and 2021. To contain the infectivity of COVID-10 19, national governments imposed a range of restrictions from the mandatory wearing of masks 11 to more stringent restrictions such as lockdowns. Public health decisions, however, were also 12 inevitably interlocked with economic as well as political considerations. One the one hand, 13 politicians have been penalized in the polls for steep rises in infection. 3 On the other hand, the 14 very restrictions that were seen as effective for curbing the infection of COVID-19 were 15 arguably afflicting the economy 4-6 and politically agitating citizens 7 . In March 2021, Prime 16 Minister Modi announced a sudden national lockdown on India, for instance, severely 17 damaging the economy. Citizens did not comply with lockdown restrictions, leading Modi to 18 eventually ease the lockdown, and India's COVID-19 situation spiraled out of control. 9 As vaccinations rolled out, some scholars had prematurely predicted that social 20 distancing would be over by the fall of 2021. 7 However, while vaccines may have reduced both 21 the number of new infections and the severity of the illness 4 , the rise of the much more 22 infectious Omicron variant and the possibility that yet another highly infectious virus may 23 arrive at our doorsteps in the coming years call for an enhanced readiness against global 24 pandemics. We contend, therefore, that a thorough investigation of the determinants of 25 COVID-19 health restrictions is still wanting. Indeed, echoing previous fears about the 26 politicization of public health policies 10 , some medical doctors have cautiously raised concerns 27 that the field of medicine is becoming "co-opted into a political programme of population 28 control" 11 , even urging public health scientists to "play politics" 12 . If we are unable to resolve 29 this state of disquietude, the next pandemic might just have to be fought with an army of 30 disillusioned and politicized medical professionals. Therefore, we also employ seemingly unrelated regression (SUR) models allowing the 48 residuals in each regression to be correlated with each other. As sensitivity tests, we apply the 49 same set of analyses using the weekly averages of the variables. First of all, the relationship between the independent and dependent variables varies 81 considerably depending on country-specific contexts. Depending on whether a polity is a 82 multiparty system or a two-party system, for instance, the baseline level of support for the 83 incumbent (or largest incumbent) party would vary considerably. To account for this variation, 84 we first use a set of fixed-effects models. However, the variables we seek to analyse are also 85 causally intertwined. Unemployment can affect the approval ratings, the stringency index, and 86 4 log scale. To accommodate this causal complexity, we use seemingly unrelated regressions 94 (SURs) to allow the error terms of these regressions to be correlated with one another. 95 96 Sensitivity tests 97 As sensitivity tests, we run the same set of models using weekly averages instead of monthly 98 (Table 2) . These tests allow us to compare long-term versus short-term considerations. Because 99 unemployment data was only available on a monthly basis, unemployment is excluded in 100 weekly model. Another set of sensitivity tests run the same models using monthly averages of 101 the growth rate in polls rather than the absolute values of polls. (Appendix Table A2 ) Because 102 of country-specific political contexts, some countries have inherently higher or lower polls than 103 others; by using the growth in polls, we can discard such country-specific differences. The included nations and their baseline characteristics are described in the Appendix (Table 107 A1). Table 1 reports the results of each regression derived with the fixed effects (top) and SUR 108 (bottom) models. As expected, the error terms of dependent variables are significantly 109 correlated (Appendix , Table A2 ), affecting both the size and significance of the variables. Consequently, we will be using the SUR models in our analysis. In Model 1, the SUR indicates estimates for the fixed-effects models (Appendix , Table A3 ). Accordingly, our analysis will 135 again be based on the SUR models. The results closely follow those of the main analysis. As There is no funding to declare. under siege from COVID-19 as hospitals overwhelmed before holidays | Reuters Impact of COVID-19 on Cancer Care: How the 269 Pandemic Is Delaying Cancer Diagnosis and Treatment for American Seniors 271 3. The political consequences of the Covid pandemic | VOX Disentangling policy effects using proxy data: Which shutdown 275 policies affected unemployment during the COVID-19 pandemic Inequality, fiscal policy and COVID19 278 restrictions in a demand-determined economy How Will COVID-19 Affect the Health Care Economy? COVID-19 Pandemic, Unemployment, and Civil Unrest 283 Underlying Deep Racial and Socioeconomic Divides. The Growing Gap in Life 284 Expectancy by Income: Implications for Federal Programs and Policy Responses How Did India's COVID-19 Crisis Become a Catastrophe? | Time Beyond policy analysis: the raw politics behind opposition to healthy public 289 policy Is science ever enough? Dare to play politics. The Lancet Tracking R of COVID-296 19: A new real-time estimation using the Kalman filter How POLITICO Poll of Polls tracks polling trends across Europe A global panel database of pandemic policies 304 (Oxford COVID-19 Government Response Tracker) Unemployment by sex and age Shepherd K. Tensions over restrictions spark violence and defiance among protesters as 309 Trump pushes states to reopen. The Washington Post Scuffles and Arrests as Anti-Lockdown Protesters March Through London agent-based simulation of an influenza epidemic in a city Association of Stay-at-Home Orders With 324 COVID-19 Hospitalizations in 4 States Effectiveness of Penalties for lockdown violations during the 327 COVID-19 pandemic in Germany Government responses and COVID-19 deaths: Global 329 evidence across multiple pandemic waves COVID-19 Helicopter Money, Monetary Policy and Central Bank 332 Independence: Economics and Politics Comparisons of Covid-19: Policy, Politics and the Price of Life. Environmental and 336 Resource Economics A literature review of the economics of 338 The comparative politics of 341 COVID-19: The need to understand government responses Democracy in times of the pandemic: explaining 344 the variation of COVID-19 policies across European democracies Dying in a Leadership Vacuum