Political regimes and deaths in the early stages of the COVID-19 pandemic

View author details View Less
  • 1 São Paulo State University, , Brazil
  • | 2 Central European University, , Austria and Democracy Institute, , Hungary
  • | 3 Central European University, , Austria
Full Access
Get eTOC alerts
Rights and permissions Cite this article

This article provides a quantitative examination of the link between political institutions and deaths during the first 100 days of the COVID-19 pandemic. We demonstrate that countries with more democratic political institutions experienced deaths on a larger per capita scale than less democratic countries. The result is robust to the inclusion of many relevant controls, a battery of estimation techniques and estimation with instrumental variables for the institutional measures. Additionally, we examine the extent to which COVID-19 deaths were impacted heterogeneously by policy responses across types of political institutions. Policy responses in democracies were less effective in reducing deaths in the early stages of the crisis. The results imply that democratic political institutions may have a disadvantage in responding quickly to pandemics.

Abstract

This article provides a quantitative examination of the link between political institutions and deaths during the first 100 days of the COVID-19 pandemic. We demonstrate that countries with more democratic political institutions experienced deaths on a larger per capita scale than less democratic countries. The result is robust to the inclusion of many relevant controls, a battery of estimation techniques and estimation with instrumental variables for the institutional measures. Additionally, we examine the extent to which COVID-19 deaths were impacted heterogeneously by policy responses across types of political institutions. Policy responses in democracies were less effective in reducing deaths in the early stages of the crisis. The results imply that democratic political institutions may have a disadvantage in responding quickly to pandemics.

Introduction

The early stages of the COVID-19 pandemic saw remarkable state interventions in social and business life on a scale not seen since the Second World War. Since that disruptive time, many countries have shifted towards democratic governance, experienced unprecedented rates of economic growth and globalisation, and greatly improved the health of their populations. Indeed, there is a large literature on the positive impact that democracy has on public health (Besley and Kudamatsu, 2006; Justesen, 2012; Fujiwara, 2015; Welander et al, 2015; Patterson and Veenstra, 2016; Wigley and Akkoyunlu-Wigley, 2017; Bellinger, 2019).1 However, securing public health during a novel pandemic is quite different. How political regimes have dealt with the COVID-19 public health crisis in its early stages is the question that we address in this article.

On 31 December 2019, China alerted the World Health Organization (WHO) to an outbreak of pneumonia of an unknown cause in the city of Wuhan in Hubei province. The epidemic quickly spread, with cases of COVID-19 confirmed throughout China and elsewhere in the world. The Chinese government’s forceful response had drawn initial praise from global health officials (Kavanagh, 2020). Other autocratic countries in the region were also fast to act. On 18 February 2020 and 10 March 2020, the WHO praised Singapore’s initial efforts to contain COVID-19 infections through aggressive tracing and quarantining of close contacts, as well as comprehensive testing of every case of influenza-like illness and pneumonia. While many international medical experts praised Singapore’s efforts to control the outbreak, others argued that this could well continue the erosion of civil liberties in the city-state (Singer and Sang-Hun, 2020). At the same time, the US and other European democracies (for example, Italy, France and Spain) initially struggled to fight COVID-19 while balancing the defence of their civil liberties and economies. Figure 1 plots the evolution of COVID-19 deaths in some selected countries.

In the graph, the horizontal axis is scaled from 0 to 100 in increments of 20 units and the vertical axis is scaled from 0 to 5.7 in increments of 2 units. The graph shows the following data for different countries. Spain: 5.6 deaths per capita, 33 days; Italy: 5.5 deaths per capita, 40 days since first death; U K: 5 deaths per capita, 30 days since first death; U S A: 4 deaths per capita, 36 days since first death; France: 5.2 deaths per capita, 50 days since first death; Iran: 4 deaths per capita, 48 days since first death; China 1.5 deaths per capita, 90 days since first death. All values are estimated.
Figure 1:

Logarithmic chart of COVID-19 deaths per capita, selected countries

Citation: Journal of Public Finance and Public Choice 37, 1; 10.1332/251569121X16268740317724

Recent research has shown that political factors have influenced policy responses and the public’s adherence to COVID-19 regulations within the US (Adolph et al, 2020; Baccini and Brodeur, 2020; Farvaque et al, 2020) and in a cross-country context (Cheibub et al, 2020; Cronert, 2020; Sebhatu et al, 2020; Cepaluni et al, 2021c; Löblová et al, 2021). COVID-19 raised the question of an authoritarian advantage in disease response (Kavanagh and Singh, 2020), as non-democratic countries with more collectivist cultures solve coordination problems relatively easily (Gorodnichenko and Roland, 2021). The results of only the handful of studies as yet available are mixed. More democratic countries have suffered from higher COVID-19 infection rates, but case fatality rates have been lower compared to less democratic countries (Karabulut et al, 2021). Policy responses of democracies to the pandemic were highly heterogeneous and often similar to measures of autocracies (Cheibub et al, 2010). Nevertheless, Maerz et al (2020) find no relationship between violations of democratic standards for emergency measures and COVID-19 death rates. The analysis that we present here focuses on the number of per capita deaths related to COVID-19 in the early stages of the pandemic and the effectiveness of policy stringency from a comparative political perspective.2 Thus, apart from a rigorous analysis of the link between political regimes, policy stringency and fatality due to COVID-19, we add to the existing literature by evaluating how democratic institutions affect the effectiveness of specific policy responses to the pandemic. Using a variety of empirical techniques, we show that more democratic countries experienced more per capita deaths, were sooner to experience deaths and enacted less stringent and less effective policy responses.

The results are sobering for advocates of democracy. In general, social scientists tend to agree that democratic governance yields better economic, health and social outcomes through more informed, rigorous and accountable policymaking processes (for example, Wittman, 1989; Przeworski and Limongi, 1993; Besley, 2006; Acemoglu and Robinson, 2012; Acemoglu et al, 2019; Bollyky et al, 2019; Dorsch and Maarek, 2019). However, the same features of democracy that are thought to yield better public policies in the long run also work to constrain the speed and incisiveness of democratic decision-making (Weeks, 2008; Malesky and London, 2014). Therein lies the trade-off in democracy that the COVID-19 crisis exposes: policy responses that impinge on personal liberties and privacy that could have contained the spread of the virus were, on average, not pursued in the early stages of the crisis.

The debate on the trade-off between protecting lives and preserving freedom is centuries old. Thomas Hobbes (1970 [1651]) wrote that an absolute sovereign – the Leviathan – is the best solution to protect the lives of its citizens and that civil liberties are of secondary value. Carl Schmitt (2005) argues that a sovereign should maintain the capacity to initiate a ‘state of exception’ in order to speed up the slow processes of democratic politics and its bureaucracy, and that every government capable of decisive action must include a dictatorial element within its constitution. Alternatively, the liberal tradition places greater emphasis on protecting civil liberties. John Stuart Mill (1887) states that it is illegitimate to infringe civil liberties and that power can be exercised rightfully over a citizen only to prevent harm to others. John Dewey (1923) also praises democracy because it provides freedom for the individual to participate in an informed way in the political sphere. In his path-breaking account of democracy, Robert Dahl warns against the appeal of an elite of experts or ‘wise men’ by arguing that a government by guardians undermines people’s autonomy and, thus, their sense of responsibility and ability to learn. Democratic processes are superior because they promote freedom, individual and collective self-determination, and moral autonomy to an extent that no other forms of political regime do (Dahl, 1989). Recently, the debate over how to solve the trade-off between providing security and freedom in democracies gained prominence during the aftermath of the 11 September 2001 (9/11) terrorist attacks against the US. As a response, Bruce Ackerman (2006) proposes a so-called ‘emergency constitution’, which enables the government to take extraordinary actions in the short run to fight against a future attack, while safeguarding human rights and without generating insuperable long-term pathologies.

As COVID-19 spread, countries had to make difficult choices between protecting civil liberties and minimising the risk of deaths. The coercive power of some non-democratic countries may have provided them with an advantage in reducing deaths in the early stages of the pandemic.3 However, over time, the freedom of information in democratic countries might help them to reverse the autocratic advantage that we document in this article (Sen, 2001). Therefore, we hope that our paper reignites the broader debate about the trade-off that democratic societies must grapple with: restricting social and economic interactions to secure public health during emergencies, while maintaining the civil liberties that define liberal democracy. We adopt a broad definition of democracy that considers competitive and transparent elections, the nature of political participation, the extent of checks on executive authority, and the degree of political freedoms and civil liberties (Dahl, 1973; Hadenius and Teorell, 2007).4

The article proceeds as follows. In the next section, we provide a brief overview of the early stages of the COVID-19 crisis. The third section presents the data that we use in the study. In the fourth and fifth sections, we present our main analyses, which use country-level data on the number of confirmed deaths from the Oxford COVID-19 Government Response Tracker (OxCGRT) as our dependent variable. First, looking at cross-sectional data, we regress total per capita deaths on some standard measures of the democratic quality of political institutions in a country. We also pursue an instrumental variable strategy on the cross-sectional data, in which we use a ‘neighbourhood’ instrument, in the spirit of Acemoglu et al (2019). We try to rule out the possibility that the result is capturing the under-reporting of deaths by less democratic countries. Second, using repeated cross-sectional data about the deaths and policies during the first 100 days of the pandemic, we run interaction models to investigate the extent to which democratic institutions impacted the effectiveness of policy interventions to reduce deaths. The final section discusses some rationalisations and theoretical implications of our results.

Background

An outbreak of pneumonia emerged in Wuhan City, Hubei province, in China in December 2019. The cause was identified as a novel coronavirus, which the WHO named ‘COVID-19’ in February 2020. COVID-19, a relative of severe acute respiratory syndrome (SARS), induces such symptoms as a dry cough, sore throat and fever, and in a small fraction of the cases, leads to severe pneumonia requiring intensive care support, especially among the elderly and patients with multiple comorbidities (Sohrabi et al, 2020).

The infection spread via human-to-human transmission, and its reach escalated due to national and international travel (Heymann and Shindo, 2020). The WHO declared that the COVID-19 outbreak constituted a public health emergency of international concern by the end of January and characterised COVID-19 as a pandemic in early March (WHO, 2020b).

On the date that we released the working paper version of this study on 27 April 2020, there were around 2.9 million confirmed COVID-19 cases, and close to 206,000 deaths recorded globally (Coronavirus Resource Center, 2020; Dong et al, 2020). The highest numbers of confirmed cases come from the US, Spain, Italy, France, Germany, the UK, China and Iran, and the incidence of the virus was still rising globally, though at a different pace across countries. While some, such as China, Singapore, Taiwan and Hong Kong, seem to have contained the outbreak, other countries (that is, in Europe and the US) struggled to control COVID-19 in the early stages of the pandemic.

Governments responded to COVID-19 differently, with measures introduced varying greatly, both in their stringency and in their timing. The Chinese experience and the WHO’s recommendations point to the role of quarantine, social distancing and the isolation of infected populations in reducing transmission (Anderson et al, 2020; WHO, 2020a). Accordingly, some of the most common social measures include travel and movement restrictions, bans on public gatherings, school and workplace closings (distance learning and teleworking from home), and closures of non-essential facilities and services.5 Given that there were no vaccines or antiviral drugs in the early days of the pandemic, the widely declared aim of these social-distancing measures was to flatten the epidemiological curve in order to avoid surpassing the capacities of healthcare systems (Anderson et al, 2020). Parallel to social measures, governments tended to invest extra funds in healthcare, and they implemented testing and contact-tracing protocols to stop chains of transmission. The latter two policies became especially important once social measures were lifted to avert a resurgence of the virus (WHO, 2020a).

The scale and nature of these policy responses, and their intrusion into private lives, are unprecedented, especially in democratic regimes, where citizens are less subservient to the ruling power and where individual rights are institutionally protected. Thus, it is interesting to compare how democratic and less democratic regimes fared in terms of death toll and the effectiveness of policy responses in the early stages of the COVID-19 crisis.

Data

Data source and country coverage

Our main data source is the OxCGRT as of 9 April 2020 (after the first 100 days of the pandemic). The OxCGRT is an ongoing data-collection project that includes systematic information on several different common policy responses that governments took as the COVID-19 virus spread, as well as a common aggregate score called the ‘stringency index’ (Hale et al, 2020). We make use of the OxCGRT longitudinal database covering daily updates on policy responses and confirmed COVID-19 cases and deaths from 1 January 2020 until 9 April 2020. We complement the OxCGRT with measures of democratic institutions and a handful of economic and social characteristics of countries. After merging all our data sources, we end up with 106 countries covering all major geographical regions.

Variables

Our main dependent variable is the number of deaths per capita due to COVID-19 in a country, based on the number of confirmed deaths from OxCGRT. The other dependent variable we study is the stringency index, which is a composite measure of the stringency of government responses to COVID-19, running on a scale from 0 to 100, based on the presence and stringency of seven policy measures.6 The main explanatory variable of interest is the level of democracy (an average of the Freedom House and Polity indicator), which ranges from 0–10, where 0 is least democratic and 10 is most democratic (Teorell et al, 2020). In some specifications, we use a binary indicator of democracy defined as 1 if the level of democracy is at least 5 in the country and 0 otherwise. As additional measures of the quality of democratic institutions, we use the Performance of Democratic Institutions indicator (measured on a scale from 1 to 10, where the higher the score, the better the performance) (Bertelsmann Stiftung, 2020).

We make use of several control variables that are potential confounders in our analysis: number of confirmed COVID-19 cases; real GDP per capita; percentage of tropical climate (percentage of the land surface area of each country with tropical climate)7; population density; a proxy for integration in the global economy (total trade as a share of gross domestic product [GDP]); a dummy for experiencing SARS (above 100 confirmed cases of SARS in the country during 2002–03 based on the WHO); and the number of airports in the country.8

We also control for variables indicative of the potential level of misreporting of COVID-19 cases, for example: an accountability and informational transparency index (Williams, 2015); the Transparency Index, measuring the availability of credible aggregate economic data that a country discloses (Hollyer et al, 2014); the coverage of a country’s death registry (UNSTAD, 2021); and the number of total COVID-19 tests per 1,000 people in a country (Ortiz-Ospina et al, 2020).9

Another group of independent variables indicates whether governments introduced certain measures in response to COVID-19 and suggests the stringency of these policies, for example, school closures, workplace closures, cancelling public events, closing public transport, public information campaigns, restrictions on internal movement, international travel controls, fiscal measures, monetary measures, investment in healthcare and vaccines, testing, and contact tracing (Hale et al, 2020). For a detailed definition of each variable, their sources and descriptive statistics please see the Online Appendix (see Cepaluni et al, 2021b).

Political regimes and deaths

Cross-section analysis: ordinary least squares

We begin with a cross-section analysis, where we consider the total deaths per capita at the end date of our data collection. We take the natural log transformation of deaths per capita due to the strong right skew of that variable.10 Technically, we estimate the following regression with ordinary least squares (OLS):
(1)

where Di represents the political institutional measure for country i, Xi is a vector of controls and εi is the error term. In the vector of controls, we include measures for the actual spread of the virus (log of confirmed cases and days since the first confirmed case), vulnerability to the spread of the virus (economic integration in the world economy, percentage of land that is tropical and number of airports), economic development level and recent experience with the SARS pandemic.

Results in Table 1 show that there are highly statistically significant cross-country correlations between political institutional measures and deaths per capita. In Panel A, we present the bivariate correlations, while in Panel B, we include the vector of control variables. The top-line result in column 1 of Panel B indicates that a one-unit increase in the democracy index is associated with a 13 per cent increase in deaths per capita. Figure A.3 in the Online Appendix demonstrates that the result is not being driven by one country in particular.11 Table A.7 in the Online Appendix demonstrates that the result is not being driven by our choice of dependent variable.12 That result is confirmed with the binary measure of democracy (in column 2), where we estimate that countries with democracy scores equal to or above 5 have experienced 71 per cent more deaths per capita than countries with scores below 5. The performance of democratic institutions indicator points in a similar direction (in column 3).

Table 1:

Cross-section OLS regressions – logged COVID-19 deaths per capita

Dependent variable: logged COVID-19 deaths per capita
Panel A: bivariate regressions(1)(2)(3)
Level of democracy (Freedom House/imputed Polity)0.24***
(0.04)
Level of democracy (binary indicator)0.99***
(0.25)
Performance of democratic institutions0.08*
(0.04)
R20.22650.08210.0600
N10410478
Panel B: multiple regressions(1)(2)(3)
Level of democracy (Freedom House/imputed Polity)0.13***
(0.03)
Level of democracy (binary indicator)0.71***
(0.18)
Performance of democratic institutions0.05
(0.03)
Log (confirmed cases)0.53***0.56***0.36***
(0.07)(0.06)(0.07)
Log (real GDP per capita)0.020.080.08
(0.11)(0.12)(0.09)
Percentage tropical climate in 2012–0.00–0.000.00
(0.00)(0.00)(0.00)
Population density–0.00*–0.00*–0.00
(0.00)(0.00)(0.00)
Trade (% of GDP)0.00*0.00**0.00
(0.00)(0.00)(0.00)
SARS–0.10–0.080.21
(0.38)(0.45)(0.37)
Log (airports)–0.21**–0.21**–0.11+
(0.07)(0.07)(0.06)
Days since first case–0.02**–0.02**–0.02***
(0.01)(0.01)(0.01)
R20.71280.69590.5659
N10410478

Notes: All specifications include robust standard errors (in parentheses). + p < .1; * p < .05; ** p < .01; *** p < .001.

As for other controls, naturally, the log of confirmed cases is highly statistically significant and positively correlated with deaths. In some specifications, economic development level is negatively correlated with deaths, as is past experience with the SARS epidemic. Economic globalisation, measured by the openness to international trade, is statistically significant and positive. The number of airports and the days since the first case both correlate negatively with COVID-19 deaths in Table 1.13 Table A.9 in the Online Appendix demonstrates that the main result is robust to additional controls for demographics, healthcare characteristics, cultural variables and regional fixed effects.14

Cross-section analysis: two-stage least squares

The method of instrumental variables provides a general solution to the problem of endogenous explanatory variables, allowing for consistent estimation when explanatory variables are correlated with the error term in a regression model. The endogeneity – or correlation between explanatory variables and the error term – may occur because of reverse causality, omitted variable bias and non-random measurement errors (Wooldridge, 2002: 89–96). We believe that reverse causality is a minor problem in our data, as it is too early to observe the impact that COVID-19 will have on political regimes.15 On the other hand, instrumental variables help us mitigate problems related to omitted variable bias and measurement errors. For instance, dictatorships might under-report the number of COVID-19 cases or economic data because of the lack of state capacity or more constrained public opinion. Consequently, dictatorships might not be better at fighting the virus, just worse at reporting information. Similarly, dictatorships might have a higher number of infectious diseases due to lower levels of public health information and sanitary measures (Jiang et al, 2020), and our findings might underestimate the effect of political regimes on reducing deaths. In our case, as the results of the instrumental variables present the same sign and statistical significance as other methods present in the article, it can also be considered a robustness check of the relationship of interest.

We construct ‘neighbourhood’ instruments for our political institution variables. Following the spirit of the identification strategies of Acemoglu et al (2019) and Dorsch and Maarek (2019), we calculate the regional average values for the political institutional measures, not including the country for which the neighbourhood effect is being calculated. More formally, for a country-specific democracy indicator, Di, and denoting the set of countries in a given region by Ii, we define the instrument for country i as:
(2)

for In other words, the instrument Zi is the jackknifed average of democracy in a region, leaving out the own-country observation.16

We then proceed to estimate the impact of democratic institutions on deaths per capita using two-stage least squares (2SLS). In the first stage, we estimate the instrumented variation in the level of democracy:
(3)
We then use the fitted values from equation (3) to estimate the second-stage relationship:
(4)

where is orthogonal to εi if our instrument Zi is valid. We have constructed the neighbourhood instrument for each of the three different institutional measures that we evaluate. First-stage F-statistics are well above the rule-of-thumb criteria of 10, so we are confident in the strength of the instruments. First-stage regressions in Table A.5 in the Online Appendix show that the instruments are significant at the 0.1 per cent level in all specifications.17

The results in Table 2 suggest that OLS had underestimated the impact of political institutions on per capita deaths. The top-line result from the 2SLS analysis is that a one-point increase in the democracy score is associated with a 30 per cent increase in COVID-19 deaths per capita over the early stages of the crisis. This suggests that estimates about the effect of political institutions from other models that we present in the article are likely to be lower-bound estimations of the actual effect.

Table 2:

Cross-section 2SLS regressions – logged COVID-19 deaths per capita

(1)(2)(3)
Dependent variable: logged COVID-19 deaths per capita
Level of democracy (Freedom House/imputed Polity)0.30*** (0.06)
Level of democracy (binary indicator)2.49*** (0.63)
Performance of democratic institutions0.03 (0.08)
Log (confirmed cases)0.47*** (0.08)0.51*** (0.09)0.37*** (0.07)
Log (real GDP per capita)–0.07 (0.12)0.05 (0.16)0.08 (0.08)
Percentage tropical climate in 2012–0.00 (0.00)–0.00 (0.00)0.00 (0.00)
Population density–0.00 (0.00)–0.00 (0.00)–0.00 (0.00)
Trade (% of GDP)0.00 (0.00)0.00 (0.00)0.00 (0.00)
SARS0.16 (0.40)0.48 (0.47)0.18 (0.39)
Log (airports)–0.27*** (0.07)–0.31*** (0.09)–0.11+ (0.06)
Days since first case–0.01 (0.01)–0.01 (0.01)–0.02** (0.01)
First-stage C-D F-stat46.68824.29110.703
First-stage K-P F-stat41.15524.0788.886
N10410478

Notes: All specifications include robust standard errors (in parentheses). First-stage regressions are reported in the Online Appendix. + p < .1; * p < .05; ** p < .01; *** p < .001.

We have conducted several sensible robustness tests on these 2SLS results. Figure A.3 in the Online Appendix shows that the result is not being driven by any specific country. Table A.6 in the Online Appendix shows reduced-form regressions, where we use OLS to estimate the impact of the instruments directly on the outcome variable. Table A.7 in the Online Appendix demonstrates that the result is not being driven by our choice of dependent variable.18 Table A.8 in the Online Appendix demonstrates that coefficient estimates are quite similar when the two-stage regressions use a Tobit IV estimator. Table A.10 in the Online Appendix demonstrates that the main result is robust to additional controls for demographics, healthcare characteristics, cultural variables and regional fixed effects.

Cross-section analysis: misreporting

A major concern is that the result may be picking up the possibility that less democratic countries are systematically under-reporting deaths. We see two possibilities for how this may occur. The first is that the death data are misreported for political reasons. The second is that governments may be under-reporting simply because they do not have the capacity to maintain a death registry or because they are not testing sufficiently and cannot differentiate between deaths that are related to, for example, heart conditions or COVID-related complications.

To address the first possibility, we include two measures of government transparency. In Table 3, which reports OLS and 2SLS results for the democracy index, we control, first, for information transparency and accountability with a transparency index by Williams (2015) and, second, for the transparency of governmental reporting of economic data using the index of Hollyer, Rosendorff and Vreeland (2014). We suppose that governments who misreport information and economic data for political reasons would also misreport data for COVID-19 deaths.

Table 3:

Cross-section regressions – controlling for under-reporting

Dependent variable: logged COVID-19 deaths per capita
Panel A: OLS regressions(1)(2)(3)(4)(5)(6)
Level of democracy (Freedom House/imputed Polity)0.16** (0.05)0.09** (0.04)0.13*** (0.03)0.17*** (0.03)0.15* (0.07)0.10+ (0.06)
Information and accountability transparency index–0.01 (0.02)
Economic data transparency index0.10 (0.07)
Death registry exists (binary)–0.10 (0.29)
Percentage of deaths registered in registry0.00 (0.01)
Log (tests per 1,000)0.38* (0.15)
Confirmed cases/tests6.39** (2.16)
Full battery of controlsYesYesYesYesYesYes
R20.73250.73190.75960.73250.76650.8083
N10391104844949
Panel B: 2SLS regressions (second stage)(1)(2)(3)(4)(5)(6)
Level of democracy (Freedom House/imputed Polity)0.92** (0.33)0.34*** (0.09)0.30*** (0.06)0.32*** (0.06)0.56** (0.20)0.56* (0.22)
Information and accountability transparency index–0.18* (0.07)
Economic data transparency index–0.09 (0.09)
Death registry exists (binary)–0.40 (0.34)
Percentage of deaths registered in registry–0.00 (0.01)
Log (tests per 1,000)0.61** (0.22)
Confirmed cases/tests5.19** (1.80)
Full battery of controlsYesYesYesYesYesYes
First-stage C-D F-stat7.11030.45546.47039.98211.3457.909
First-stage K-P F-stat5.53327.07041.08937.4675.1474.421
N10391104844949

Notes: All specifications include robust standard errors (in parentheses). + p < .1; * p < .05; ** p < .01; *** p < .001.

To address the second possibility, we include controls for the existence of a death registry, the percentage of deaths registered, the log of COVID-19 tests per 1,000 people and the ratio of total confirmed cases to total tests. If governments under-report due to a lack of information, these controls should pick that up because countries providing information on deaths and testing might also have better reporting on COVID-19 fatalities. The testing data are only available for 49 countries (for which we also have the other controls), but the results are stable to the introduction of these important controls despite the different samples. Particularly with the OLS estimations, the results are quite consistent with our baseline estimations (where β1 = 0.13). The 2SLS estimates remain statistically significantly positive as well with these additional controls, though the strength of the instrument and second-stage coefficient estimates are less stable. As an alternative way to address the issue, we have also considered dropping the least (and most) transparent countries from the sample, as these are the most likely to be under-reporting (over-reporting) in Table A.4 in the Online Appendix.19

Political regimes, policy responses and deaths

We have also investigated the extent to which more democratic countries have responded with more stringent policies and the extent to which policy responses were more effective in reducing deaths in democratic countries. First, we present a result with the cross-section data and then we examine country repeated cross-section observations and (daily) time-varying data.

Political regimes and policy stringency

To begin, we use 2SLS to estimate a cross-sectional relation, where we now consider the index of policy stringency as the dependent variable (see the ‘Data’ section earlier). Consistent with the results concerning COVID-19 deaths, Table 4 demonstrates that more democratic countries pursued less stringent policy responses. Column 2 indicates that the democratic countries had stringency scores more than 18 points lower than non-democracies, on average.

Table 4:

Cross-section 2SLS regressions

(1)(2)(3)
Dependent variable: stringency index
Level of democracy (Freedom House/imputed Polity)–2.12*
(1.01)
Level of democracy (binary)–18.16+
(9.45)
Performance of democratic institutions–0.45
(1.93)
Log (confirmed cases)3.18*2.97+2.63
(1.53)(1.53)(2.00)
Log (real GDP per capita)–1.23–1.93–1.59
(2.28)(2.47)(2.35)
Percentage tropical climate in 2012–0.03–0.02–0.04
(0.05)(0.05)(0.05)
Population density–0.00–0.00–0.00
(0.00)(0.00)(0.00)
Trade (% of GDP)0.030.040.05
(0.05)(0.05)(0.06)
SARS–14.98–17.18–29.12**
(14.02)(14.22)(11.27)
Log (airports)–0.93–0.57–0.72
(1.51)(1.65)(1.78)
Days since first case–0.06–0.060.07
(0.11)(0.12)(0.18)
First-stage C-D F-stat40.59720.4638.758
First-stage K-P F-stat36.99220.9627.587
N10510580

Notes: All specifications include robust standard errors (in parentheses). First-stage regressions are reported in the Online Appendix. + p < .1; * p < .05; ** p < .01; *** p < .001.

Heterogeneous effects of policy responses: linear interaction

We turn now to the daily dimension of the OxCGRT data, investigating the differential effects of policy responses. Our data set consists of pooled cross-sectional and time-varying observations. COVID-19 deaths, COVID-19 confirmed cases and policy responses vary daily. Countries’ sociodemographic characteristics and institutional variables are time invariant.

Our dependent variable of logged COVID-19 deaths is non-stationary, as we failed to reject the Dickey-Fuller test’s null hypothesis that all panels contain a unit root. To reduce the concerns about possible inflated results because of time trends, Figure A.5 in the Online Appendix presents two randomised inference tests stratified by region and with 500 permutations.20 The outcome variable is COVID-19 deaths per capita. The right figure shows that the assignment of democracy to countries is not random.21 However, it produces statistically significant (ρ = 0.001) point estimates similar to our baseline regression first shown in column 1, Panel B, of Table 1. The second figure (left) runs the same test with a random variable, where the point estimate is close to zero. Therefore, we conclude that democracy is endogenous because the regression line is far from the distribution but the effect on death is real. As a second approach, Table A.13 in the Online Appendix presents our standard regressions, employing Driscoll and Kraay (1998) standard errors. These standard errors are heteroscedasticity consistent and robust to spatial and temporal dependence. Our main measure of democracy and its binary indicator are both statistically associated with deaths. The performance of democratic institutions is not statistically significant. Those findings are similar to our cross-section results.22 Unfortunately, the method does not allow factor variables and interaction effects, which are central to our analyses in the following sections. Third, if a time series has a unit-root problem, the first difference (or controlling for fixed effects when T > 2) of such a time series is ‘stationary’. It is also well known that for N > T, the first-difference and fixed-effects models produce similar results (Wooldridge, 2002: 321–5). Hence, instead of taking the first difference of our dependent variable, we choose to present fixed-effects models to compare our results with the previous analysis that employs logged COVID-19 deaths.

Finally, we also estimate survival models in Section A.4 (see Figure A.6 and Table A.19) in the Online Appendix, employing similar model specifications. However, our primary dependent variable is now the length of time (number of days) it takes for a country to experience a COVID-19-related death after recording its first confirmed case of COVID-19. As the pandemic is still unfolding, deaths are ‘censured’, and survival models are handy tools to exploit our data. The results corroborate the article’s findings, as democracies experienced deaths earlier and maintained higher death rates during the first 100 days of the pandemic.

These caveats aside, we turn to our panel data to investigate how individual policy responses correlated with deaths. In particular, we examine the extent to which there were heterogeneous effects of the policy responses on deaths across types of political institution. Here, we exploit the daily variation in the deaths and policy response data in a repeated cross-section OLS regression model with multiplicative interaction terms:
(5)

where: Yi,t is the outcome variable (logged deaths by COVID-19 per capita) in country i on day t; Di is the political regime of country i; Pi,t is the policy action of country i on day t (the treatment); and Di * Pi,t is the interaction effect that we are interested in. Here, the political regime serves as a moderator – a variable that affects the direction and strength of the policy treatment effect. As before, Xi is the vector of control variables.

Table 5 presents the results. We employ (logged) COVID-19 deaths per capita as the dependent variable and our standard vector of controls. All models include day and region fixed effects and robust standard errors.23 To save space, we do not display the control variables, as all results are comparable with other models presented in this study. We consider the following policy response variables: school closing (SC), workplace closing (WPC), cancel public events (CPE), close public transport (CPT), public information campaigns (PIC), restrict internal movement (RIM), international travel controls (ITC), fiscal measures (FM), monetary measures (MM), investment in healthcare (IH), investment in vaccines (IV), testing framework (TF) and contact tracing (CT).

Table 5:

Panel OLS regressions — Policy Responses, Political Regimes and COVID-19 Deaths

(1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)
Dependent Variable: logged COVID-19 Deaths per capitaSCWPCCPECPTPICRIMITCFMMMIHIVTFCT
Level of Democracy (Freedom House/Imputed Polity)-0.03***-0.02***-0.03***-0.01*-0.00-0.02***-0.02***0.02***0.02***0.02***0.03***-0.01*-0.00
(0.004)(0.004)(0.004)(0.003)(0.003)(0.004)(0.004)(0.003)(0.004)(0.003)(0.003)(0.004)(0.004)
School closing-0.38***
(0.020)
democracy x school closing0.04***
(0.003)
Workplace closing-0.36***
(0.025)
democracy × workplace closing0.05***
(0.003)
Cancel public events-0.40***
(0.022)
democracy × cancel public events0.04***
(0.003)
Close public transport-0.40***
(0.023)
democracy × close public transport0.04***
(0.004)
Public information campaigns-0.36***
(0.035)
democracy × public information campaigns0.04***
(0.004)
restrict internal movement-0.37***
(0.025)
democracy × restrict internal movement0.04***
(0.004)
Int travel controls-0.23***
(0.013)
democracy × int travel controls0.03***
(0.002)
Fiscal measures-0.00**
(0.000)
democracy × fiscal measures0.00**
(0.000)
Monetary measures-0.01**
(0.003)
democracy × monetary measures0.00**
(0.000)
Investment in health care-0.00*
(0.000)
democracy × investment in health care0.00
(0.000)
Investment in vaccines-0.00+
(0.000)
democracy x investment in vaccines0.00
(0.000)
Testing framework-0.31***
(0.020)
democracy × testing framework0.03***
(0.003)
Contact tracing-0.31***
(0.022)
democracy × contact tracing0.03***
(0.003)
R20.620.630.620.620.600.620.610.600.600.600.610.600.61
Countries10310310410110199102999894939594
N6046602860125916591658605954578256325432535148564892

Notes: The table employs the standard controls first presented in Table 1. All specifications include day and region fixed effects (not shown) and robust standard errors (in parenthesis). +p < .1; *p < .05; **p < .01; ***p < .001.

As we can see in Table 5, all significant interactions are estimated to have a positive sign, meaning that the policy responses were less effective in reducing deaths in countries with higher democracy scores. To understand the marginal effect of a policy, β2 + β3 × Di, take, for example, the closing public events policy response variable (see column 3 of Table 5). There the marginal effect of cancelling public events is given by –0.03 + 0.004 × Di, so the predicted marginal effect for a full dictatorship (Di = 0) is a 3 per cent reduction of deaths, while for a full democracy (Di = 10), it is a 37 per cent increase in deaths.

Heterogeneous effects of policy responses: non-linear interaction

Here, we relax the assumption of linearity between the multiplicative term , employing a kernel-smoothing estimator of the marginal effect (Hainmueller et al, 2019). The kernel approach allows us to flexibly estimate the functional form of the marginal effect of P on Y across the values of D by estimating a series of local effects with a kernel reweighting scheme.

Formally, the kernel-smoothing method is based on the following semi-parametric model:
(6)

in which f (.), g(.), and γ(.) are smoothing functions of D, and g(.) captures the marginal effect of P on Y. The kernel regression nests the standard interaction model given in equation (5) as a special case when and . In other words, the kernel approach will converge to a linear function when the assumption of linearity is true. However, multiplicative terms can vary freely across the range of D. In addition, if covariates X are included in the model, the coefficients of those covariates are also allowed to vary freely across the range of D, resulting in a flexible estimator that helps to guard against misspecification bias with respect to the covariates.

As all of the significant interaction signs in Table 5 go in the same direction, for simplicity, we present the kernel regression result here with the policy stringency index (as was used in Table 4) interacted with the democracy level. We fit a model with regional and day fixed effects, our standard vector of controls and standard errors clustered at the country level. Figure 2 summarises the main result graphically, presenting the marginal effect of an increase in the policy stringency index as a function of the level of democracy in the country.

In the graph, the horizontal axis is scaled from 0 to 10 in increments of 2 units and the vertical axis is scaled from negative 0.02 to 0.005 unit in increments of 0.005 unit. The graph shows a line that rises from (0, negative 0.008) to (9.5, 0.002). The line is surrounded by a shaded patch, which is wider on the left and narrower on the right. All values are estimated.
Figure 2:

Marginal effect of the stringency index on logged COVID-19 deaths per capita, conditional on political regimes

Citation: Journal of Public Finance and Public Choice 37, 1; 10.1332/251569121X16268740317724

Notes: The model contains log (confirmed cases), log (real GDP per capita), percentage of tropical climate in 2012, population density (people per km2 of land area), trade (percentage of GDP), SARS and log (airports). The specification includes region and day fixed effects and clustered standard errors at the country level. Grey areas represent 95 per cent confidence intervals. The histogram in the bottom of the figure present levels of political regimes.

The relationship in Figure 2 is almost linear, though we are estimating a flexible kernel model. Figure 2 shows that the marginal effect of the stringency index – conditional on different levels of political regime – reduces (logged) COVID-19 deaths per capita. The marginal effect of policy responses on (logged) COVID-19 deaths per capita is negative, conditional on relatively low levels of democracy (below 7.5 points on the Freedom House/imputed Polity scale). Non-democratic countries that adopt stringent policy measures reduce the number of (logged) COVID-19 deaths per capita, whereas highly democratic countries (above 9 points) adopting the same policy responses do not necessarily, as the confidence interval for more democratic countries (above 7.5 points) crosses zero.

Discussion and conclusions

Our analysis demonstrates that in the early stages of the COVID-19 pandemic, more democratic countries experienced deaths on a larger scale, implemented less stringent policy and implemented policies that were less effective at reducing deaths. Despite the findings, we do not neglect the benefits of democracy. Instead, we hope to raise awareness for a discussion about the need to adapt and refine democratic political institutions during periods of crisis responses or emergencies. Therefore, we should not ignore controversial but much-needed debates.

There may be systematic under-reporting of COVID-19 deaths in less democratic countries. In some cases, under-reporting could be a political decision. In other cases, it may reflect a lack of state capacity to perform the testing necessary to determine the real cause of death (The Economist, 2020).24 We have tried to deal with this through our instrumental-variable strategy and by including controls for governmental transparency, testing rates and subsample analyses, but it remains a concern.

Centralised decision-making may be advantageous when it comes to responding to pandemics (Schwartz, 2012). Table 4 shows that more democratic countries imposed less stringent restrictions during the first 100 days of the crisis. With fewer checks built into the policymaking process, public health policy responses can be made quicker and perhaps more incisively in autocratic governments. Especially for policies that impinge on civil liberties and privacy, autocratic governments have a far freer hand in imposing restrictions on their citizens. The trade-off between capacity to protect the public health and personal liberties is a central debate during, and probably well after, the COVID-19 crisis (Harari, 2020). Based on a recent Pew Research Center survey (Silver et al, 2020), the majority of citizens of 14 advanced economies have unfavourable views of China’s responses to the pandemic, suggesting that excessively stringent policies may not be acceptable in democratic regimes.

Autocratic governments may have had an extra advantage to the extent that their citizens are more obedient to governmental decrees, especially those that may disrupt the social and business lives of citizens. We demonstrated that more stringent policies decreased deaths in less democratic countries but not necessarily in fully democratic ones. Whether it comes from higher public support for government initiatives (possibly through the threat of force) or from the government’s ability to stifle debate around their decrees in the media, the greater obedience of citizens in autocracies may have had a role in the lower scale of deaths. Protests against social-distancing restrictions in democracies with different institutional performances, such as Brazil and the US, seem to reinforce this point.

Finally, this article contributes to a substantive topic that will have broad and lasting implications. It may well be many years until we have a clear understanding of how the COVID-19 crisis will impact our societies. As the pandemic started in East Asia, the location of some of the best-managed autocracies, it may be that our sample disproportionately includes autocratic governments with high state capacity. However, replications of our results for later stages of the pandemic also produce similar findings (for example, Cepaluni et al, 2021c).

Moreover, the autocratic advantage may be short-term and only observable in terms of the death toll due to the pandemic. Looking at other performance indicators that COVID-19 will affect in the long run, such as mental health and other well-being measures, as well as the economic recovery, may well melt away the advantage that our estimations have documented in the short term. Under certain circumstances, levels of democracy here and now might matter less than democratic capital acquired through a historical learning process (Gerring et al, 2004; Persson and Tabellini, 2009). For instance, because of freedom of information, large-scale famines that take years to unfold are less frequent in societies than in authoritarian systems (Sen, 2001). However, the long term might be bleaker if we compromise the short term during an emergency or a crisis, for example, terrorist attacks, natural disasters or pandemics. Then, democratic societies and the citizens living in them must develop emergency strategies to respond quickly and more efficiently to future outbreaks of pandemics or similar urgent crises (Boin et al, 2009; Leonard and Howitt, 2010). In our view, failure to deal effectively with pandemics poses a risk to the public’s trust in democratic governance and could contribute to the democratic rollback that is happening in some regions of the world.

Notes

1

The result is, however, disputed by some scholars, notably, Truex (2017) in a global sensitivity analysis. Furthermore, democracy may not improve the health of the poor (Ross, 2006; van der Windt and Vandoros, 2017).

2

Figure A.1 in the Online Appendix (see Cepaluni et al, 2021b) shows the geographic distribution of deaths and policy response stringency.

3

Schwartz (2012) argues that there may be an ‘authoritarian advantage’ in policy responses to pandemics in their comparative analysis of how the severe acute respiratory syndrome (SARS) outbreak was dealt with by China and Taiwan.

4

We also acknowledge that different measures of democracy have their respective strengths and weaknesses, and that the employment of a definition of democracy should be guided by normative and theoretical preferences, even when these measures correlate with each other (Cheibub et al, 2010).

5

Cepaluni, Dorsch and Kovarek (2021a) show that such restrictive policies are less likely in societies with higher levels of geographic mobility.

6

The stringency index is a composite measure that is a simple additive score of seven response indicators: school closures, workplace closures, cancelling public events, closing public transport, public information campaigns, restricting internal movement and international travel controls. The aim of the index is to have a general cross-national measure of policy stringency that allows for systematic comparisons across countries (Hale et al, 2020).

7

Arguably, there is a small association between lower temperatures and the spread of COVID-19 (Sajadi et al, 2020).

8

Tables in the Online Appendix also report results with additional controls that were suggested by the journal’s referees.

9

Additionally, in the Online Appendix, we explore variables that characterise state capacity: an estimate of government effectiveness (Kaufmann et al, 2011); an indicator of the quality of government (PRS, 2019); and a state fragility index (Marshall et al, 2017). Our state capacity measures focus on bureaucratic capacity and the government’s capacity to maintain law and order, as well as its legitimacy.

10

We divide the total deaths by the population (in millions) and then take the log of deaths per capita + 1, so that we do not lose the observations for which there were reported cases but no deaths. Figure A.2 in the Online Appendix shows histograms for the raw data and the log-transformed data. Results are robust to the use of either measure.

11

The figure shows the coefficient point estimate is quite stable at around 0.13 and highly statistically significant as we jackknife the following potentially influential countries: Belgium, China, France, Germany, Iran, Italy, Netherlands, the UK and the US.

12

There, we show that higher democracy index scores are positively and statistically significantly correlated with logged per capita deaths weighted by the fraction of population over age 65, logged total deaths, deaths per capita, deaths per capita weighted by the fraction of population over age 65, the death rate among confirmed cases and the logged death rate.

13

However, we note that both the number of airports and days since the first case have a positive pairwise correlation coefficient with logged per capita deaths.

14

There, we show several specifications where we additionally control for the log of the population, the log of the rate of the population over the age of 65, the probability of an early death due to cardiovascular disease, the prevalence of smoking, the number of hospital beds per 1,000 people, average cultural survey responses to questions about individual civil liberties and acquiescence to rulers.

15

We already know that countries such as Hungary, Poland, Turkey and Russia are adopting anti-democratic legislation. In Hungary, for example, the Parliament passed a law on 31 March 2020 that allows the government to rule by decree, suspend the Parliament and repeal any existing law, indefinitely. However, we cannot observe these changes in our data.

16

We use the regional coding from Hadenius and Teorell (2007), which are Eastern Europe and post-Soviet Union, Latin America, North Africa and Middle East, sub-Saharan Africa, Western Europe and North America, East Asia, South Asia, the Pacific, and the Caribbean.

17

The quality of political institutions in regional countries may influence the death rate through their impact on the regional movement of people and goods, which would violate the exclusion restriction. However, controlling for economic integration and airports should, to some extent, alleviate these concerns. Moreover, including the number of confirmed cases controls for the spread of the virus through such channels.

18

There, we show that higher democracy index scores are positively and statistically significantly correlated with logged per capita deaths weighted by the fraction of population over age 65, logged total deaths, deaths per capita, and deaths per capita weighted by the fraction of the population over age 65.

19

There, we drop countries that are in the lowest decile and highest deciles according to the following: the two transparency indicators and the tests per 1,000 measure.

20

Randomisation inference resamples the treatment variable constructing its statistical distribution and compares it to the estimates of the original regression. The randomisation inference assumes a fixed population sample. Only the ‘treatment’ variable is random. The method tests the null hypothesis that there is no effect of the original random draw on the data by assessing whether the sample realisation of the statistic is consistent with the inferred distribution.

21

This is the reason we employed an instrumental variable in the first place.

22

Controlling for population at risk (see Table A.14 in the Online Appendix) and misreporting (see Table A.15 in the Online Appendix) does not affect the main findings of the panel analysis.

23

The results are robust to several model specifications. For instance, Table A.17 in the Online Appendix controls for day and country fixed effects and produces virtually the same results.

24

In Table A.11 in the Online Appendix, we display repeated cross-section OLS regressions, always controlling for day and region fixed effects and also three measures of state capacity – government effectiveness, quality of government and the state fragility index – in columns 2, 3 and 4, respectively. We still observe more (logged) deaths per capita in more democratic countries, though these results are smaller than in the specifications from the main article.  Table A.12 in the Online Appendix interacts political regime and the three measures of state capacity, controlling for time and region fixed effects. The results for democracy still hold. These results suggest that state capacity can moderate the effect of political regime on the (logged) number of COVID-19 deaths. To illustrate our point, democratic countries with a lot of state capacity, such as Germany and South Korea, did relatively well in fighting the pandemic in its early stages.

Acknowledgements

We thank Fernando Bizarro, Caitlin Brown, László Bruszt, Cristina Corduneanu-Huci, Raphael Cunha, Amanda Driscoll, Mihály Fazekas, Ivan Filipe Fernandes, Danilo Freire, Giampaolo Garzarelli, Julius Horvath, Evelyne Hubscher, Erin Jenne, Martin Kahanec, Miklós Koren, Levente Littvay, Paul Maarek, Alexandru Moise, Anand Murugesan, Kata Orosz, Gabor Simonovits, and Michael Touchton for comments that improved the analysis presented here. We also thank the audiences at São Paulo State University, University of São Paulo’s Center for International Negotiation Studies, Getúlio Vargas Foundation, Central European University, European Consortium for Political Research and at the International Political Science Association workshop ‘Rising Democracies, Interrupted’. Dorsch thanks the Aix-Marseille School of Economics, where he was a fellow in residence during part of this research, for their financial support and hospitality.

Conflict of interest

The authors declare that there is no conflict of interest.

References

  • Acemoglu, D. and Robinson, J.A. (2012) Why Nations Fail: The Origins of Power, Prosperity, and Poverty, New York: Crown Books.

  • Acemoglu, D., Naidu, S., Restrepo, P. and Robinson, J.A. (2019) Democracy does cause growth, Journal of Political Economy, 127(1): 47100. doi: 10.1086/700936

    • Search Google Scholar
    • Export Citation
  • Ackerman, B.A. (2006) Before the Next Attack: Preserving Civil Liberties in an Age of Terrorism, New Haven, CT: Yale University Press.

  • Adolph, C., Amano, K., Bang-Jensen, B., Fullman, N. and Wilkerson, J. (2021) Pandemic politics: timing State-level social distancing responses to COVID-19, Journal of Health Politics, Policy and Law, 46(2): 21133.

    • Search Google Scholar
    • Export Citation
  • Anderson, R.M., Heesterbeek, H., Klinkenberg, D. and Hollingsworth, T.D. (2020) How will Country-based mitigation measures influence the course of the COVID-19 epidemic?, The Lancet, 395(10228): 9314. doi: 10.1016/S0140-6736(20)30567-5

    • Search Google Scholar
    • Export Citation
  • Baccini, L. and Brodeur, A. (2021) Explaining governors’ response to the COVID-19 pandemic in the United States, American Politics Research, 49(2): 21520.

    • Search Google Scholar
    • Export Citation
  • Bellinger, N.M. (2019) Why democracy matters: democratic attributes and human well-being, Journal of International Relations and Development, 22(2): 41340. doi: 10.1057/s41268-017-0105-1

    • Search Google Scholar
    • Export Citation
  • Bertelsmann Stiftung (2020) Transformation Index BTI 2020: Governance in International Comparison,  Verlag Bertelsmann Stiftung, https://www.bti-project.org/en/index/political-transformation.html

    • Search Google Scholar
    • Export Citation
  • Besley, T. (2006) Principled Agents? The Political Economy of Good Government, Oxford: Oxford University Press.

  • Besley, T. and Kudamatsu, M. (2006) Health and democracy, American Economic Review, 96(2): 31318. doi: 10.1257/000282806777212053

  • Boin, A., Hart, P. and McConnell, A. (2009) Crisis exploitation: political and policy impacts of framing contests, Journal of European Public Policy, 16(1): 81106. doi: 10.1080/13501760802453221

    • Search Google Scholar
    • Export Citation
  • Bollyky, T.J., Templin, T., Cohen, M., Schoder, D., Dieleman, J.L. and Wigley, S. (2019) The relationships between democratic experience, adult health, and cause-specific mortality in 170 countries between 1980 and 2016: an observational analysis, The Lancet, 393(10181): 162840. doi: 10.1016/S0140-6736(19)30235-1

    • Search Google Scholar
    • Export Citation
  • Cepaluni, G., Dorsch, M. and Kovarek, D. (2021a) Mobility and policy responses during the COVID-19 pandemic in 2020, SSRN 3817289, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3817289.

    • Search Google Scholar
    • Export Citation
  • Cepaluni, G., Dorsch, M. and Branyiczki, R. (2021b) Political regimes and deaths in the early stages of the COVID-19 pandemic: online appendix https://sites.google.com/view/dorsch/research.

    • Search Google Scholar
    • Export Citation
  • Cepaluni, G., Dorsch, M. and Dzebo, S. (2021c) Populism, political regimes, and COVID-19 deaths,  SSRN 3816398, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3816398.

    • Search Google Scholar
    • Export Citation
  • Cheibub, J.A., Gandhi, J. and Vreeland, J.R. (2010) Democracy and dictatorship revisited, Public Choice, 143(1–2): 67101. doi: 10.1007/s11127-009-9491-2

    • Search Google Scholar
    • Export Citation
  • Cheibub, J.A., Jean Hong, J.Y. and Przeworski, A. (2020) Rights and deaths: government reactions to the pandemic, SSRN 3645410, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3645410.

    • Search Google Scholar
    • Export Citation
  • Coronavirus Resource Center (2020) COVID-19 Case Tracker, Johns Hopkins University & Medicine, https://coronavirus.jhu.edu/map.html.

  • Cronert, A. (2020) Democracy, state capacity, and COVID-19 related school closures, APSA preprint, https://preprints.apsanet.org/engage/apsa/article-details/5ea8501b68bfcc00122e96ac.

    • Search Google Scholar
    • Export Citation
  • Dahl, R.A. (1973) Polyarchy: Participation and Opposition, New Haven, CT:  Yale University Press.

  • Dahl, R.A. (1989) Democracy and its Critics, New Haven, CT: Yale University Press.

  • Dewey, J. (1923) Democracy and Education: An Introduction to the Philosophy of Education, New York: Macmillan.

  • Dong, E., Du, H. and Gardner, L. (2020) An interactive web-based dashboard to track Covid-19 in real time, The Lancet Infectious Diseases, 20(5): 53334.

    • Search Google Scholar
    • Export Citation
  • Dorsch, M.T. and Maarek, P. (2019) Democratization and the conditional dynamics of income distribution, American Political Science Review, 113(2): 385404. doi: 10.1017/S0003055418000825

    • Search Google Scholar
    • Export Citation
  • Driscoll, J.C. and Kraay, A.C. (1998) Consistent covariance matrix estimation with spatially dependent panel data, Review of Economics and Statistics, 80(4): 54960. doi: 10.1162/003465398557825

    • Search Google Scholar
    • Export Citation
  • Farvaque, E., Iqbal, H. and Ooghe, N. (2020) Health politics? Determinants of US states’ reactions to COVID-19, Journal of Public Finance and Public Choice. doi: https://doi.org/10.1332/251569120X16040852770342

    • Search Google Scholar
    • Export Citation
  • Fujiwara, T. (2015) Voting technology, political responsiveness, and infant health: evidence from Brazil, Econometrica, 83(2): 42364. doi: 10.3982/ECTA11520

    • Search Google Scholar
    • Export Citation
  • Gerring, J., Bond, P., Barndt, W.T. and Moreno, C. (2004) Democracy and economic growth: a historical perspective, World Politics, 57(3): 323–64.  doi: 10.1353/wp.2006.0002

    • Search Google Scholar
    • Export Citation
  • Gorodnichenko, Y. and Roland, G. (2021) Culture, institutions and democratization, Public Choice, 187(1): 16595. doi: 10.1007/s11127-020-00811-8

    • Search Google Scholar
    • Export Citation
  • Hadenius, A. and Teorell, J. (2007) Pathways from authoritarianism, Journal of Democracy, 18(1): 14357. doi: 10.1353/jod.2007.0009

  • Hainmueller, J., Mummolo, J. and Xu, Y. (2019) How much should we trust estimates from multiplicative interaction models? Simple tools to improve empirical practice, Political Analysis, 27(2): 16392. doi: 10.1017/pan.2018.46

    • Search Google Scholar
    • Export Citation
  • Hale, T., Petherick, A., Phillips, T. and Webster, S. (2020) Variation in government responses to COVID-19, Blavatnik School of Government Working Paper, 31, 2020–11. Accessed: 7 April, 2020.

    • Search Google Scholar
    • Export Citation
  • Harari, Y.N. (2020) The world after coronavirus,  Financial Times, 19 March, 2020.

  • Heymann, D.L. and Shindo, N. (2020) COVID-19: what is next for public health?, The Lancet, 395(10224): 5425. doi: 10.1016/S0140-6736(20)30374-3

    • Search Google Scholar
    • Export Citation
  • Hobbes, T. (1969 [1651]) Leviathan, Menston, Scolar P., 1969.

  • Hollyer, J.R., Rosendorff, B.P. and Vreeland, J.R. (2014) Measuring transparency, Political Analysis, 22(4): 41334. doi: 10.1093/pan/mpu001

    • Search Google Scholar
    • Export Citation
  • Jiang, F., Deng, L., Zhang, L., Cai, Y., Cheung, C.W. and Xia, Z. (2020) Review of the clinical characteristics of coronavirus disease 2019 (COVID-19), Journal of General Internal Medicine,  35(5):1545-49. doi: 10.1007/s11606-020-05762-w

    • Search Google Scholar
    • Export Citation
  • Justesen, M.K. (2012) Democracy, dictatorship, and disease: political regimes and HIV/AIDS, European Journal of Political Economy, 28(3): 37389. doi: 10.1016/j.ejpoleco.2012.02.001

    • Search Google Scholar
    • Export Citation
  • Karabulut, G., Zimmermann, K.F., Bilgin, M.H. and Doker, A.C. (2021) Democracy and COVID-19 outcomes, Economics Letters, 203: 109840.  doi: 10.1016/j.econlet.2021.109840

    • Search Google Scholar
    • Export Citation
  • Kaufmann, D., Kraay, A. and Mastruzzi, M. (2011) The worldwide governance indicators: methodology and analytical issues, Hague Journal on the Rule of Law, 3(2): 22046. doi: 10.1017/S1876404511200046

    • Search Google Scholar
    • Export Citation
  • Kavanagh, M.M. (2020) Authoritarianism, outbreaks, and information politics, The Lancet Public Health, 5(3): e1356.

  • Kavanagh, M.M. and Singh, R. (2020) Democracy, capacity, and coercion in pandemic response: COVID-19 in comparative political perspective, Journal of Health Politics, Policy and Law, 45(6): 9971012. doi: 10.1215/03616878-8641530

    • Search Google Scholar
    • Export Citation
  • Leonard, H.B. and Howitt, A.M. (2010) Organising response to extreme emergencies: the Victorian bushfires of 2009, Australian Journal of Public Administration, 69(4): 37286. doi: 10.1111/j.1467-8500.2010.00695.x

    • Search Google Scholar
    • Export Citation
  • Löblová, O., Rone, J. and Borbáth, E. (2021) Focus on Czechia, Hungary, and Bulgaria, in Greer, S.L., King, E.J., da Fonseca, E.M. and Peralta-Santos, A. (eds), Coronavirus Politics:  The Comparative Politics and Policy of COVID-19. Ann Arbor, MI: University of Michigan Press.

    • Search Google Scholar
    • Export Citation
  • Maerz, S.F., Lührmann, A., Lachapelle, J. and Edgell, A.B. (2020) Worth the Sacrifice? Illiberal and Authoritarian Practices during Covid-19, V-Dem Working Paper 110, September.

    • Search Google Scholar
    • Export Citation
  • Malesky, E. and London, J. (2014) The political economy of development in China and Vietnam, Annual Review of Political Science, 17: 395419.  doi: 10.1146/annurev-polisci-041811-150032

    • Search Google Scholar
    • Export Citation
  • Marshall, M.G., Gurr, T.R. and Jaggers, K. (2017) Global Report 2017: Conflict, Governance, and State Fragility, Center for Systemic Peace, http://www.systemicpeace.org/vlibrary/GlobalReport2017.pdf

    • Search Google Scholar
    • Export Citation
  • Mill, J.S. (1887) On Liberty, London: Longmans, Green.

  • Ortiz-Ospina, E., Ritchie, H., Beltekian, D., Mathieu, E., Hasell, J., Macdonald, B., Giattino, C., Appel, C., Rodés-Guirao, L. and Roser, M. (2020) Our world in data COVID-19 testing dataset [Online]. https://ourworldindata.org/coronavirus

    • Search Google Scholar
    • Export Citation
  • Patterson, A.C. and Veenstra, G. (2016) Politics and population health: testing the impact of electoral democracy, Health & Place, 40: 6675.

    • Search Google Scholar
    • Export Citation
  • Persson, T. and Tabellini, G. (2009) Democratic capital: the nexus of political and economic change, American Economic Journal: Macroeconomics, 1(2): 88126. doi: 10.1257/mac.1.2.88

    • Search Google Scholar
    • Export Citation
  • PRS (Political Risk Services Group) (2019) International Country Risk Guide,  Political Risk Services [Online]. https://epub.prsgroup.com/products/icrg/international-country-risk-guide-icrg#

    • Search Google Scholar
    • Export Citation
  • Przeworski, A. and Limongi, F. (1993) Political regimes and economic growth, Journal of Economic Perspectives, 7(3): 5169. doi: 10.1257/jep.7.3.51

    • Search Google Scholar
    • Export Citation
  • Ross, M. (2006) Is democracy good for the poor?, American Journal of Political Science, 50(4): 86074. doi: 10.1111/j.1540-5907.2006.00220.x

    • Search Google Scholar
    • Export Citation
  • Sajadi, M.M., Habibzadeh, P., Vintzileos, A., Shokouhi, S., Miralles-Wilhelm, F. and Amoroso, A. (2020) Temperature and latitude analysis to predict potential spread and seasonality for COVID-19, SSRN 3550308.

    • Search Google Scholar
    • Export Citation
  • Schmitt, C. (2005) Political Theology: Four Chapters on the Concept of Sovereignty, Chicago, IL: University of Chicago Press.

  • Schwartz, J. (2012) Compensating for the ‘authoritarian advantage’ in crisis response: a comparative case study of SARS pandemic responses in China and Taiwan, Journal of Chinese Political Science, 17(3): 31331. doi: 10.1007/s11366-012-9204-4

    • Search Google Scholar
    • Export Citation
  • Sebhatu, A., Wennberg, K., Arora-Jonsson, S. and Lindberg, S.I. (2020) Explaining the homogeneous diffusion of COVID-19 nonpharmaceutical interventions across heterogeneous countries,  Proceedings of the National Academy of Sciences, 117(35): 2120108.

    • Search Google Scholar
    • Export Citation
  • Sen, A. (2001) Development as Freedom,  Oxford: Oxford University Press.

  • Silver, L., Devlin, K. and Huang, C. (2020) Unfavorable views of China reach historic highs in many countries, Pew Research Center, https://www.pewresearch.org/global/2020/10/06/unfavorable-views-of-china-reach-historic-highs-in-many-countries/.

    • Search Google Scholar
    • Export Citation
  • Singer, N. and Sang-Hun, C. (2020) As coronavirus surveillance escalates, personal privacy plummets, The New York Times, 23 March, https://www.nytimes.com/2020/03/23/technology/coronavirus-surveillance-tracking-privacy.html.

    • Search Google Scholar
    • Export Citation
  • Sohrabi, C., Alsafi, Z., O’Neill, N., Khan, M., Kerwan, A., Al-Jabir, A., Iosifidis, C. and Agha, R. (2020) World Health Organization declares global emergency: a review of the 2019 novel coronavirus (COVID-19), International Journal of Surgery, 76: 716.

    • Search Google Scholar
    • Export Citation
  • Teorell, J., Dahlberg, S., Holmberg, S., Rothstein, B., Pachon, N.A. and Axelsson, S. (2020) The Quality of Government Standard Dataset, version Jan20, University of Gothenburg, Gothenburg, Sweden. doi: https://doi.org/10.18157/qogbasjan20

    • Search Google Scholar
    • Export Citation
  • The Economist (2020) Tracking COVID-19 excess deaths across countries,  The Economist, 23 July 2021, https://www.economist.com/graphic-detail/coronavirus-excess-deaths-tracker.

    • Search Google Scholar
    • Export Citation
  • Truex, R. (2017) The myth of the democratic advantage, Studies in Comparative International Development, 52(3): 26177. doi: 10.1007/s12116-015-9192-4

    • Search Google Scholar
    • Export Citation
  • UNSTAD (United Nations Statistics Division) (2021) Coverage of birth and death registration – 2021, https://unstats.un.org/unsd/demographic-social/crvs/.

    • Search Google Scholar
    • Export Citation
  • Van der Windt, P. and Vandoros, S. (2017) Democracy and health: evidence from within-country heterogeneity in the Congo, Social Science & Medicine, 194: 1016.

    • Search Google Scholar
    • Export Citation
  • Weeks, J.L. (2008) Autocratic audience costs: regime type and signaling resolve, International Organization, 62(1): 3564. doi: 10.1017/S0020818308080028

    • Search Google Scholar
    • Export Citation
  • Welander, A., Lyttkens, C.H. and Nilsson, T. (2015) Globalization, democracy, and child health in developing countries, Social Science & Medicine, 136: 5263.

    • Search Google Scholar
    • Export Citation
  • WHO (World Health Organization) (2020a) Coronavirus disease 2019 (COVID-19): situation report, 72. https://apps.who.int/iris/handle/10665/331685

    • Search Google Scholar
    • Export Citation
  • WHO (2020b) WHO timeline – COVID-19,  https://www.who.int/emergencies/diseases/novel-coronavirus-2019/interactive-timeline?gclid=CjwKCAjwx8iIBhBwEiwA2quaq-vLip6gKHVCSKcsSmQb_4bm08h6VAINGoPEBP3sbw0Hr9p-yxv7PhoCyT0QAvD_BwE#event-115

    • Search Google Scholar
    • Export Citation
  • Wigley, S. and Akkoyunlu-Wigley, A. (2017) The impact of democracy and media freedom on under-5 mortality, 1961–2011, Social Science & Medicine, 190: 23746.

    • Search Google Scholar
    • Export Citation
  • Williams, A. (2015) A global index of information transparency and accountability, Journal of Comparative Economics, 43(3): 80424. doi: 10.1016/j.jce.2014.10.004

    • Search Google Scholar
    • Export Citation
  • Wittman, D. (1989) Why democracies produce efficient results, Journal of Political Economy, 97(6): 1395424. doi: 10.1086/261660

  • Wooldridge, J.M. (2002) Econometric Analysis of Cross Section and Panel Data, Cambridge, MA: MIT Press.

  • View in gallery

    Logarithmic chart of COVID-19 deaths per capita, selected countries

  • View in gallery

    Marginal effect of the stringency index on logged COVID-19 deaths per capita, conditional on political regimes

  • Acemoglu, D. and Robinson, J.A. (2012) Why Nations Fail: The Origins of Power, Prosperity, and Poverty, New York: Crown Books.

  • Acemoglu, D., Naidu, S., Restrepo, P. and Robinson, J.A. (2019) Democracy does cause growth, Journal of Political Economy, 127(1): 47100. doi: 10.1086/700936

    • Search Google Scholar
    • Export Citation
  • Ackerman, B.A. (2006) Before the Next Attack: Preserving Civil Liberties in an Age of Terrorism, New Haven, CT: Yale University Press.

  • Adolph, C., Amano, K., Bang-Jensen, B., Fullman, N. and Wilkerson, J. (2021) Pandemic politics: timing State-level social distancing responses to COVID-19, Journal of Health Politics, Policy and Law, 46(2): 21133.

    • Search Google Scholar
    • Export Citation
  • Anderson, R.M., Heesterbeek, H., Klinkenberg, D. and Hollingsworth, T.D. (2020) How will Country-based mitigation measures influence the course of the COVID-19 epidemic?, The Lancet, 395(10228): 9314. doi: 10.1016/S0140-6736(20)30567-5

    • Search Google Scholar
    • Export Citation
  • Baccini, L. and Brodeur, A. (2021) Explaining governors’ response to the COVID-19 pandemic in the United States, American Politics Research, 49(2): 21520.

    • Search Google Scholar
    • Export Citation
  • Bellinger, N.M. (2019) Why democracy matters: democratic attributes and human well-being, Journal of International Relations and Development, 22(2): 41340. doi: 10.1057/s41268-017-0105-1

    • Search Google Scholar
    • Export Citation
  • Bertelsmann Stiftung (2020) Transformation Index BTI 2020: Governance in International Comparison,  Verlag Bertelsmann Stiftung, https://www.bti-project.org/en/index/political-transformation.html

    • Search Google Scholar
    • Export Citation
  • Besley, T. (2006) Principled Agents? The Political Economy of Good Government, Oxford: Oxford University Press.

  • Besley, T. and Kudamatsu, M. (2006) Health and democracy, American Economic Review, 96(2): 31318. doi: 10.1257/000282806777212053

  • Boin, A., Hart, P. and McConnell, A. (2009) Crisis exploitation: political and policy impacts of framing contests, Journal of European Public Policy, 16(1): 81106. doi: 10.1080/13501760802453221

    • Search Google Scholar
    • Export Citation
  • Bollyky, T.J., Templin, T., Cohen, M., Schoder, D., Dieleman, J.L. and Wigley, S. (2019) The relationships between democratic experience, adult health, and cause-specific mortality in 170 countries between 1980 and 2016: an observational analysis, The Lancet, 393(10181): 162840. doi: 10.1016/S0140-6736(19)30235-1

    • Search Google Scholar
    • Export Citation
  • Cepaluni, G., Dorsch, M. and Kovarek, D. (2021a) Mobility and policy responses during the COVID-19 pandemic in 2020, SSRN 3817289, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3817289.

    • Search Google Scholar
    • Export Citation
  • Cepaluni, G., Dorsch, M. and Branyiczki, R. (2021b) Political regimes and deaths in the early stages of the COVID-19 pandemic: online appendix https://sites.google.com/view/dorsch/research.

    • Search Google Scholar
    • Export Citation
  • Cepaluni, G., Dorsch, M. and Dzebo, S. (2021c) Populism, political regimes, and COVID-19 deaths,  SSRN 3816398, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3816398.

    • Search Google Scholar
    • Export Citation
  • Cheibub, J.A., Gandhi, J. and Vreeland, J.R. (2010) Democracy and dictatorship revisited, Public Choice, 143(1–2): 67101. doi: 10.1007/s11127-009-9491-2

    • Search Google Scholar
    • Export Citation
  • Cheibub, J.A., Jean Hong, J.Y. and Przeworski, A. (2020) Rights and deaths: government reactions to the pandemic, SSRN 3645410, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3645410.

    • Search Google Scholar
    • Export Citation
  • Coronavirus Resource Center (2020) COVID-19 Case Tracker, Johns Hopkins University & Medicine, https://coronavirus.jhu.edu/map.html.

  • Cronert, A. (2020) Democracy, state capacity, and COVID-19 related school closures, APSA preprint, https://preprints.apsanet.org/engage/apsa/article-details/5ea8501b68bfcc00122e96ac.

    • Search Google Scholar
    • Export Citation
  • Dahl, R.A. (1973) Polyarchy: Participation and Opposition, New Haven, CT:  Yale University Press.

  • Dahl, R.A. (1989) Democracy and its Critics, New Haven, CT: Yale University Press.

  • Dewey, J. (1923) Democracy and Education: An Introduction to the Philosophy of Education, New York: Macmillan.

  • Dong, E., Du, H. and Gardner, L. (2020) An interactive web-based dashboard to track Covid-19 in real time, The Lancet Infectious Diseases, 20(5): 53334.

    • Search Google Scholar
    • Export Citation
  • Dorsch, M.T. and Maarek, P. (2019) Democratization and the conditional dynamics of income distribution, American Political Science Review, 113(2): 385404. doi: 10.1017/S0003055418000825

    • Search Google Scholar
    • Export Citation
  • Driscoll, J.C. and Kraay, A.C. (1998) Consistent covariance matrix estimation with spatially dependent panel data, Review of Economics and Statistics, 80(4): 54960. doi: 10.1162/003465398557825

    • Search Google Scholar
    • Export Citation
  • Farvaque, E., Iqbal, H. and Ooghe, N. (2020) Health politics? Determinants of US states’ reactions to COVID-19, Journal of Public Finance and Public Choice. doi: https://doi.org/10.1332/251569120X16040852770342

    • Search Google Scholar
    • Export Citation
  • Fujiwara, T. (2015) Voting technology, political responsiveness, and infant health: evidence from Brazil, Econometrica, 83(2): 42364. doi: 10.3982/ECTA11520

    • Search Google Scholar
    • Export Citation
  • Gerring, J., Bond, P., Barndt, W.T. and Moreno, C. (2004) Democracy and economic growth: a historical perspective, World Politics, 57(3): 323–64.  doi: 10.1353/wp.2006.0002

    • Search Google Scholar
    • Export Citation
  • Gorodnichenko, Y. and Roland, G. (2021) Culture, institutions and democratization, Public Choice, 187(1): 16595. doi: 10.1007/s11127-020-00811-8

    • Search Google Scholar
    • Export Citation
  • Hadenius, A. and Teorell, J. (2007) Pathways from authoritarianism, Journal of Democracy, 18(1): 14357. doi: 10.1353/jod.2007.0009

  • Hainmueller, J., Mummolo, J. and Xu, Y. (2019) How much should we trust estimates from multiplicative interaction models? Simple tools to improve empirical practice, Political Analysis, 27(2): 16392. doi: 10.1017/pan.2018.46

    • Search Google Scholar
    • Export Citation
  • Hale, T., Petherick, A., Phillips, T. and Webster, S. (2020) Variation in government responses to COVID-19, Blavatnik School of Government Working Paper, 31, 2020–11. Accessed: 7 April, 2020.

    • Search Google Scholar
    • Export Citation
  • Harari, Y.N. (2020) The world after coronavirus,  Financial Times, 19 March, 2020.

  • Heymann, D.L. and Shindo, N. (2020) COVID-19: what is next for public health?, The Lancet, 395(10224): 5425. doi: 10.1016/S0140-6736(20)30374-3

    • Search Google Scholar
    • Export Citation
  • Hobbes, T. (1969 [1651]) Leviathan, Menston, Scolar P., 1969.

  • Hollyer, J.R., Rosendorff, B.P. and Vreeland, J.R. (2014) Measuring transparency, Political Analysis, 22(4): 41334. doi: 10.1093/pan/mpu001

    • Search Google Scholar
    • Export Citation
  • Jiang, F., Deng, L., Zhang, L., Cai, Y., Cheung, C.W. and Xia, Z. (2020) Review of the clinical characteristics of coronavirus disease 2019 (COVID-19), Journal of General Internal Medicine,  35(5):1545-49. doi: 10.1007/s11606-020-05762-w

    • Search Google Scholar
    • Export Citation
  • Justesen, M.K. (2012) Democracy, dictatorship, and disease: political regimes and HIV/AIDS, European Journal of Political Economy, 28(3): 37389. doi: 10.1016/j.ejpoleco.2012.02.001

    • Search Google Scholar
    • Export Citation
  • Karabulut, G., Zimmermann, K.F., Bilgin, M.H. and Doker, A.C. (2021) Democracy and COVID-19 outcomes, Economics Letters, 203: 109840.  doi: 10.1016/j.econlet.2021.109840

    • Search Google Scholar
    • Export Citation
  • Kaufmann, D., Kraay, A. and Mastruzzi, M. (2011) The worldwide governance indicators: methodology and analytical issues, Hague Journal on the Rule of Law, 3(2): 22046. doi: 10.1017/S1876404511200046

    • Search Google Scholar
    • Export Citation
  • Kavanagh, M.M. (2020) Authoritarianism, outbreaks, and information politics, The Lancet Public Health, 5(3): e1356.

  • Kavanagh, M.M. and Singh, R. (2020) Democracy, capacity, and coercion in pandemic response: COVID-19 in comparative political perspective, Journal of Health Politics, Policy and Law, 45(6): 9971012. doi: 10.1215/03616878-8641530

    • Search Google Scholar
    • Export Citation
  • Leonard, H.B. and Howitt, A.M. (2010) Organising response to extreme emergencies: the Victorian bushfires of 2009, Australian Journal of Public Administration, 69(4): 37286. doi: 10.1111/j.1467-8500.2010.00695.x

    • Search Google Scholar
    • Export Citation
  • Löblová, O., Rone, J. and Borbáth, E. (2021) Focus on Czechia, Hungary, and Bulgaria, in Greer, S.L., King, E.J., da Fonseca, E.M. and Peralta-Santos, A. (eds), Coronavirus Politics:  The Comparative Politics and Policy of COVID-19. Ann Arbor, MI: University of Michigan Press.

    • Search Google Scholar
    • Export Citation
  • Maerz, S.F., Lührmann, A., Lachapelle, J. and Edgell, A.B. (2020) Worth the Sacrifice? Illiberal and Authoritarian Practices during Covid-19, V-Dem Working Paper 110, September.

    • Search Google Scholar
    • Export Citation
  • Malesky, E. and London, J. (2014) The political economy of development in China and Vietnam, Annual Review of Political Science, 17: 395419.  doi: 10.1146/annurev-polisci-041811-150032

    • Search Google Scholar
    • Export Citation
  • Marshall, M.G., Gurr, T.R. and Jaggers, K. (2017) Global Report 2017: Conflict, Governance, and State Fragility, Center for Systemic Peace, http://www.systemicpeace.org/vlibrary/GlobalReport2017.pdf

    • Search Google Scholar
    • Export Citation
  • Mill, J.S. (1887) On Liberty, London: Longmans, Green.

  • Ortiz-Ospina, E., Ritchie, H., Beltekian, D., Mathieu, E., Hasell, J., Macdonald, B., Giattino, C., Appel, C., Rodés-Guirao, L. and Roser, M. (2020) Our world in data COVID-19 testing dataset [Online]. https://ourworldindata.org/coronavirus

    • Search Google Scholar
    • Export Citation
  • Patterson, A.C. and Veenstra, G. (2016) Politics and population health: testing the impact of electoral democracy, Health & Place, 40: 6675.

    • Search Google Scholar
    • Export Citation
  • Persson, T. and Tabellini, G. (2009) Democratic capital: the nexus of political and economic change, American Economic Journal: Macroeconomics, 1(2): 88126. doi: 10.1257/mac.1.2.88

    • Search Google Scholar
    • Export Citation
  • PRS (Political Risk Services Group) (2019) International Country Risk Guide,  Political Risk Services [Online]. https://epub.prsgroup.com/products/icrg/international-country-risk-guide-icrg#

    • Search Google Scholar
    • Export Citation
  • Przeworski, A. and Limongi, F. (1993) Political regimes and economic growth, Journal of Economic Perspectives, 7(3): 5169. doi: 10.1257/jep.7.3.51

    • Search Google Scholar
    • Export Citation
  • Ross, M. (2006) Is democracy good for the poor?, American Journal of Political Science, 50(4): 86074. doi: 10.1111/j.1540-5907.2006.00220.x

    • Search Google Scholar
    • Export Citation
  • Sajadi, M.M., Habibzadeh, P., Vintzileos, A., Shokouhi, S., Miralles-Wilhelm, F. and Amoroso, A. (2020) Temperature and latitude analysis to predict potential spread and seasonality for COVID-19, SSRN 3550308.

    • Search Google Scholar
    • Export Citation
  • Schmitt, C. (2005) Political Theology: Four Chapters on the Concept of Sovereignty, Chicago, IL: University of Chicago Press.

  • Schwartz, J. (2012) Compensating for the ‘authoritarian advantage’ in crisis response: a comparative case study of SARS pandemic responses in China and Taiwan, Journal of Chinese Political Science, 17(3): 31331. doi: 10.1007/s11366-012-9204-4

    • Search Google Scholar
    • Export Citation
  • Sebhatu, A., Wennberg, K., Arora-Jonsson, S. and Lindberg, S.I. (2020) Explaining the homogeneous diffusion of COVID-19 nonpharmaceutical interventions across heterogeneous countries,  Proceedings of the National Academy of Sciences, 117(35): 2120108.

    • Search Google Scholar
    • Export Citation
  • Sen, A. (2001) Development as Freedom,  Oxford: Oxford University Press.

  • Silver, L., Devlin, K. and Huang, C. (2020) Unfavorable views of China reach historic highs in many countries, Pew Research Center, https://www.pewresearch.org/global/2020/10/06/unfavorable-views-of-china-reach-historic-highs-in-many-countries/.

    • Search Google Scholar
    • Export Citation
  • Singer, N. and Sang-Hun, C. (2020) As coronavirus surveillance escalates, personal privacy plummets, The New York Times, 23 March, https://www.nytimes.com/2020/03/23/technology/coronavirus-surveillance-tracking-privacy.html.

    • Search Google Scholar
    • Export Citation
  • Sohrabi, C., Alsafi, Z., O’Neill, N., Khan, M., Kerwan, A., Al-Jabir, A., Iosifidis, C. and Agha, R. (2020) World Health Organization declares global emergency: a review of the 2019 novel coronavirus (COVID-19), International Journal of Surgery, 76: 716.

    • Search Google Scholar
    • Export Citation
  • Teorell, J., Dahlberg, S., Holmberg, S., Rothstein, B., Pachon, N.A. and Axelsson, S. (2020) The Quality of Government Standard Dataset, version Jan20, University of Gothenburg, Gothenburg, Sweden. doi: https://doi.org/10.18157/qogbasjan20

    • Search Google Scholar
    • Export Citation
  • The Economist (2020) Tracking COVID-19 excess deaths across countries,  The Economist, 23 July 2021, https://www.economist.com/graphic-detail/coronavirus-excess-deaths-tracker.

    • Search Google Scholar
    • Export Citation
  • Truex, R. (2017) The myth of the democratic advantage, Studies in Comparative International Development, 52(3): 26177. doi: 10.1007/s12116-015-9192-4

    • Search Google Scholar
    • Export Citation
  • UNSTAD (United Nations Statistics Division) (2021) Coverage of birth and death registration – 2021, https://unstats.un.org/unsd/demographic-social/crvs/.

    • Search Google Scholar
    • Export Citation
  • Van der Windt, P. and Vandoros, S. (2017) Democracy and health: evidence from within-country heterogeneity in the Congo, Social Science & Medicine, 194: 1016.

    • Search Google Scholar
    • Export Citation
  • Weeks, J.L. (2008) Autocratic audience costs: regime type and signaling resolve, International Organization, 62(1): 3564. doi: 10.1017/S0020818308080028

    • Search Google Scholar
    • Export Citation
  • Welander, A., Lyttkens, C.H. and Nilsson, T. (2015) Globalization, democracy, and child health in developing countries, Social Science & Medicine, 136: 5263.

    • Search Google Scholar
    • Export Citation
  • WHO (World Health Organization) (2020a) Coronavirus disease 2019 (COVID-19): situation report, 72. https://apps.who.int/iris/handle/10665/331685

    • Search Google Scholar
    • Export Citation
  • WHO (2020b) WHO timeline – COVID-19,  https://www.who.int/emergencies/diseases/novel-coronavirus-2019/interactive-timeline?gclid=CjwKCAjwx8iIBhBwEiwA2quaq-vLip6gKHVCSKcsSmQb_4bm08h6VAINGoPEBP3sbw0Hr9p-yxv7PhoCyT0QAvD_BwE#event-115

    • Search Google Scholar
    • Export Citation
  • Wigley, S. and Akkoyunlu-Wigley, A. (2017) The impact of democracy and media freedom on under-5 mortality, 1961–2011, Social Science & Medicine, 190: 23746.

    • Search Google Scholar
    • Export Citation
  • Williams, A. (2015) A global index of information transparency and accountability, Journal of Comparative Economics, 43(3): 80424. doi: 10.1016/j.jce.2014.10.004

    • Search Google Scholar
    • Export Citation
  • Wittman, D. (1989) Why democracies produce efficient results, Journal of Political Economy, 97(6): 1395424. doi: 10.1086/261660

  • Wooldridge, J.M. (2002) Econometric Analysis of Cross Section and Panel Data, Cambridge, MA: MIT Press.

  • 1 São Paulo State University, , Brazil
  • | 2 Central European University, , Austria and Democracy Institute, , Hungary
  • | 3 Central European University, , Austria

Content Metrics

May 2022 onwards Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 219 219 143
PDF Downloads 149 149 92

Altmetrics

Dimensions