COVID-19 and democracy: a scoping review

Abstract

Background

The resilience of democracy is tested under exogenous shocks such as crises. The COVID-19 pandemic has recently tested the resilience of democratic institutions and practices around the world.

Aim

The purpose of this article is to scope the early research literature that discusses democracy and the COVID-19 pandemic. We review scientific journal articles published during the first two years of the pandemic. We ask three research questions in scoping this body of literature: (1) what are the key topic areas of all published research that associates itself with both democracy and COVID-19, (2) what kinds of conceptual and theoretical contributions has research literature that more specifically discusses democracy under the pandemic produced, and (3) what are the impacts of democracy to the pandemic and vice versa according to empirical research?

Methods

The scoping review methodology draws on systematic literature search strategies, computational methods, and manual coding. The systematic Web of Science search produced 586 articles for which we conducted a Correlated Topic Model. After technical and manual screening, we identified 94 journal articles that were manually coded.

Results

The early research on democracy and the COVID-19 pandemic offers a versatile body of scholarship. The topic modeling shows that the scholarship discusses issues of crises, governance, rights, society, epidemiology, politics, electorate, technology, and media. The body of papers with conceptual and theoretical contributions has offered new insights on the difficulties, possibilities, and means to maintain democracy under a pandemic. Empirical research on democracy’s impact on the COVID-19 pandemic and vice versa varies in terms of methodology, geographical scope, and scientific contributions according to the direction of influence studied. Democracy appears to have a significant impact on some aspects of policy responses and epidemiological characteristics of the pandemic. In most parts of the world, the scope, franchise, and authenticity of democracy narrowed down due to the pandemic, albeit in most cases only temporarily.

Conclusions

A significant number of papers show that the pandemic has accentuated democratic backsliding but is unlikely to have undermined established democracies that have proved resilient in face of the pandemic. But empirical research has also made visible some weak signals of antidemocratic tendencies that may become more accentuated in the longer run.

Background

Democracy is inextricably linked to crisis [1]. Democracies are often perceived to be in crisis due to the absence of some features which we consider as definitional of democracy [2]. While different theories of democracy may focus on the absence of different features, most would agree that exogenous shocks, especially large-scale crises, such as financial crises or pandemics, are the key factors that challenge and test the durability of democratic institutions and practices. Democracy rarely flourishes under large-scale crises and crises tend to have negative impacts on democracy; but democracy may as well recover, revive, and sometimes even strengthen after crises [2].

The resilience of democracy in face of external shocks has recently gained much research attention [3]. Democratic resilience can be defined as the capacity of democratic institutions and practices to absorb and recover, adapt, innovate, or transform in response to shock or crisis. Democracy is a contested concept, both an ideal and a system of government, and inclusive of procedural and substantive elements, and ultimately about collective decision-making, which means that democratic resilience inherently lacks the conceptual specificity of resilience as it is found in some other disciplines [3]. As a political form, democracy is the perpetual absence of something more, an always pending agenda that calls for the redress of social ills and further advances in the manifold matters [4]. This means that there are neither external safeguards for democratic politics nor any kinds of internal guarantees that democracy will be maintained under crises: democracy survives crises only if citizens continue to engage with democratic politics under them. Hence, the impacts of crises to democracy depends on how much of the society is kept under democratic control, to what degree people rely on democratic institutions to solve conflicts and problems, and to what degree people participate in democratic politics in the crisis conditions [2].

No crisis has recently tested democracy as much as the COVID-19 pandemic [5, 6]. The pandemic has had enormous impacts to the society, economy, health systems, and everyday lives globally. The cause of the pandemic, the new severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), was first identified from an outbreak in Wuhan, China, in December 2019. The World Health Organization (WHO) declared a Public Health Emergency of International Concern on 30 January 2020 and a pandemic on 11 March 2020. The COVID-19 pandemic has also given rise to an exceptionally broad and rapidly growing body of research in various scientific disciplines and numerous fields of research. The research has also been followed by an exceptionally large number of published literature reviews. The high number of reviews can be in part explained by the necessity to synthesize findings for policymaking purposes [7].

Our scoping review is the first attempt to scope early research on democracy and the COVID-19 pandemic. We combine semi-systematic and integrative approaches to the scoping task. Our review combines systematic search strategies, computational methods, and manual coding. The scoping review covers all scientific articles included in the Web of Science database by the end of March 2022. The period studied covers roughly the first two years of the pandemic, which saw the rise and, towards the end of the period, the decline of wide policy responses around the world. The body of literature discussed here represents only early research on the topic, and much more research is likely to be published on the topic in the near future. Scoping the research published during this period is important, as this literature has, at least in principle, been available for policy experts, policymakers, and public officials during the acute phases of the pandemic. We are thus scoping the peer-reviewed research that has potentially had an impact on the policy and politics of the pandemic response and, possibly, the maintenance of democratic politics in crisis conditions.

We focus on three issues in our review. First, using computational methods, we scope the topics of published scholarship that associates itself explicitly with democracy and the COVID-19 pandemic. Second, we scope the conceptual and theoretical contributions of research literature that more specifically discusses democracy under the pandemic. Third, we scope the empirical research on the impacts of democracy to the pandemic and vice versa.

Methods

Scoping reviews and democracy research

The purpose of a scoping review is to map a complex research field [8]. Unlike systematic reviews that seek to synthesize evidence on clearly defined topic or phenomenon, scoping reviews seek to scope the topics of a body of literature, clarify concepts, identify knowledge gaps, or/and to investigate research conduct [9]. Scoping review is viewed as helpful for understanding complex research fields that are in nature highly heterogeneous and fast growing [10]. Scoping reviews that define a field in this way have been considered especially useful in political sciences [11]. Scoping reviews that deal with highly complex fields can deploy various review methods [9]. Typical to scoping reviews is to combine elements of semi-systematic and integrative review methods [12].

Democracy qualifies as a highly complex and heterogeneous research field. The concept of democracy is essentially contested and deployed differently by schools of thought differentiated by various normative, philosophical, and theoretical commitments [13]. Research on the COVID-19 pandemic qualifies as a fast-growing field. The evolution of the COVID-19 pandemic over time and variation and change in government responses to the pandemic suggest that the complexity of the research field may also have increased over time.

It is important to note here that two ‘academic worlds’ of democracy research exist [14]; one of theoretically grounded (positive) empirical research into real-world democratic, democratizing, or non-democratic regimes, institutions, and practices; and another of political philosophy that critically and normatively assesses different conceptions and (possible) practices of democracy. In the former world, the configurations of political organization that can be called ‘democratic’ and the research objects addressed by research depend on the theory or model of democracy onto which the research is founded. This research can be approached with systematic and semi-systematic review strategies. In the latter world, the concept of democracy is deployed to discuss a broad variety of theoretical and normative issues in various topic areas that range far beyond political systems, institutions, and practices [15]. An unclearly bounded, conceptually moving, and often incommensurable field easily escapes systematic review methods, and requires more interpretive and integrative methods [11].

Like semi-systematic scoping reviews, we ask broad research questions, use systematic search strategies, and focus only on scientific articles to recognize key themes and assess the state of knowledge in early research on the topic. But, like integrative scoping reviews, we are not focused on the details of research conduct but seek to scope and, where possible, synthesize the key approaches, insights, and arguments presented in research. Identifying the key topics of the ‘second world’ research would be laborious with solely manual coding. Therefore, we use computational methods that can deal with a large body of research to assist our work. This review strategy allows us to scope the key topics of both the ‘first world’ and the ‘second world’ of democracy research as well as to provide more nuanced scoping of the findings and contributions of research in the former ‘world’. Our review is conducted in three stages: collection and screening of the sample, computational topic modeling, and manual coding (see below).

Research questions and limitations

Our review is guided by three research questions:

• (1) What are the key topics of research that deploys the concepts of democracy and the COVID-19?

• (2) What kinds of conceptual and theoretical contributions has been produced by research that discusses democracy under the pandemic?

• (3) What is the mutual relation between democracy and the COVID-19 pandemic according to empirical research: how has democracy affected the pandemic and how has the pandemic affected democracy?

The main limitation of our review is that we scope only the topics under which democracy and the pandemic is discussed, and the key insights and arguments (e.g., conceptualizations, theories, empirical findings, normative assessments) that have been presented. Due to the high proportion of commentaries, essays, and conceptual papers it would add limited value to discuss empirical research settings in detail. We discuss issues of research conduct only by identifying the broad research strategies and key indicators used and geographical areas addressed in empirical research. We leave the issue of more detailed research design and methods for later systematic reviews. Another limitation is that we exclude from our review the monographs and edited volumes that have been published on the topic during this period (e.g., [16,17,18]). These publications include complex arguments that rely on highly varied conceptualizations of democracy and scopes of analysis. Their summarization requires thematically more focused research questions and integrative review methods.

Sample and technical screening

We conducted a search in the Web of Science database with a simple search string covid-19 AND democracy addressed to all search fields to identify relevant terms for more focused searches. We used the term ’covid-19’ to focus only on the specific pandemic and exclude previous pandemics from the query. We addressed all search fields to allow for the possibility that another term would be used in the title or abstract. The search included all items indexed in the databases by 31 March 2022. The search produced a body of 617 items. We excluded from this body all other items than published journal articles (such as unpublished conference presentations, posters and working papers), one retracted article, and all articles without an abstract. With these measures, we were left with 586 articles, which we use as a sample for answering our first research question through topic modeling (see below).

A quick reading of the article titles suggested that most items were unlikely to address democracy or/and the pandemic in a way that provides answers to our second or third research question. The reading also suggested that some referrals to the pandemic used different terms for COVID-19 in the title and main text. We changed the search string on the COVID-19 to ‘covid’ OR ‘pandemic’ OR ‘current public-health emergency’ for our further screening. We then deployed various technical measures to narrow down the corpus. We included only articles that explicitly used search terms democra* and covid-19/pandemic/current public-health emergency in the title or/and abstract. We thus use the inclusion of a key concept in the title or abstract as a proxy for an explicit discussion of the concept or some phenomena described by it, and the inclusion of two concepts as a proxy for explicit discussion on some relation between the two concepts or some phenomena described with both concepts. We also included in the review only articles whose main text was in English. With this screening, the items narrowed down to 383 articles. This body of literature comprises of all research that can potentially answer our second and third research questions.

The works excluded by this measure include some publications that discussed closely related topics to the publications included in our review but did not associate the topic explicitly with democracy. These include legal studies on the constitutional aspects of government responses to the COVID-19 pandemic, political science scholarship addressing changes in power relations during the pandemic, and multidisciplinary studies on human rights in connection with the pandemic. Even though these publications are excluded from our review, they may provide some additional insights on topics that other scholars have explicitly associated with democracy. Addressing these studies in further reviews requires more focused systematic reviews.

Topic modeling

We conducted a topic modeling for the 586 items to generate an initial understanding of the contexts in which democracy and the COVID-19 pandemic have been simultaneously mentioned by a wide body of research. We implemented Correlated Topic Model (CTM) [19] using Structural Topic Model (STM) package in R to study the topics in the abstracts of the 586 articles. In case of no covariates used, the STM model reduces to implementation of CTM [20]. Common pre-processing steps were taken, including removal of punctuation, stopwords, and numbers, followed by the removal of infrequent terms which only appeared in maximum of one document. We did not use stemming, as it is shown that stemmers produce no meaningful improvements for the process [21].

To select the number of topics, we ran the model first with k = 5, 10, 20, 30, 40 and 50. On the second run we narrowed the scope to k = 5–20. From the second run we chose the best number of topics to be 9. To select the best model, we estimated a set of 20 separate 9-topic CTMs with different initializations. The model selected involves a trade-off between maximized semantic coherence and exclusivity of the topic-word vectors, which is typical to model selection [20]. We explored the estimated topics using the words associated with each topic, and then validated the topics by selecting one document with the highest document-topic loading per each topic and manually checked the abstract and title of those documents. All steps in the topic modeling process were performed using R. We report our findings in the next section of the paper.

Manual screening and coding

Next, our analysis shifted to the 383 articles. We screened the articles manually in two phases: first through abstracts and then through full texts. The purpose of the screening was to identify the articles that provide conceptual or theoretical contributions to democracy research (second research question), or/and discuss the impacts of democracy to the pandemic or vice versa (third research question). In both phases we worked with Rayyan software [22]. Due to the complexity of the topic at hand, we did not use blinded decision-making regarding inclusion/exclusion. This allowed us to reflect upon the decisions as a team. Conducting an unblinded review has not been found to increase the risk of bias in systematic reviews if review decisions are documented [23]. We documented our discussions as notes in the Rayyan software.

In the first phase, screening required very little interpretation. For example, several publications were omitted on the basis that they studied something that occurred in a ‘democratic’ or ‘non-democratic’ country during the pandemic but did not otherwise address democracy or associate their more specific research object with democracy. More interpretation was needed in two specific types of cases. First, we excluded articles in which the publication presented democracy or the pandemic as a contextual factor but did not address it directly. Excluded articles typically studied something that took place during but not due to the pandemic or/and presented their findings as a potential but not (at least yet) actual concern for democracy. We used labels ‘covid-19 not the research topic’ and ‘democracy not the research topic’ in Rayyan to mark these publications. Second, we excluded papers that addressed issues explicitly associated with democracy and the pandemic but did not discuss either of the two directly. Articles that treated both democracy and the pandemic only as covariates for something else were also excluded here. We used labels ‘indirect link to the pandemic’ and ‘indirect link to democracy’ in Rayyan to document these exclusion decisions.

124 articles remained after manual screening of the abstracts. In the second stage, our attention turned to the full texts to screen and code the remaining articles. The measure excluded 30 articles. The main reason for excluding an article at this stage was that the association with democracy or the COVID-19 pandemic in the abstract proved indirect in the full text. For example, the full text of one article [24] shows that the article explicitly discusses liberalism and not democracy under the pandemic, even though the term ‘liberal democracy’ is used in the abstract. Another reason for exclusion was that the use of the term ‘democracy’ (or some of its variation) in the abstract proved as a synonym or label for something else in the full text without addressing its democratic qualities or “democraticness”. For example, one paper used the label ‘democratization’ explicitly as a synonym for widening user/patient involvement without other references to democracy [25].

Our final sample for answering the second and third research question includes in total 94 articles, which are all listed in the references. We used the topics produced by topic modeling to provide preliminary labels to each article in Rayyan. We added labels to divide the articles to three broad categories: (1) all articles with conceptual, normative, and theoretical contributions (second research question), (2) empirical articles discussing democracy as a determinant of some aspects of the COVID-19 pandemic (first part of the third research question), and (3) empirical articles discussing the pandemic’s impacts on democracy (second part of the third research question).

For each of the three broad categories we gathered key insights from the articles into a separate Excel table. We first collected the conceptual and theoretical contributions regarding the COVID-19 pandemic to one table. In case of the two other broad categories, we split tables by thematic labels. We used two initial labels under the second broad category – epidemiology and policy responses – which ultimately covered all articles in the broad category. The challenge with the third broad category was that most articles deployed different models or conceptions of democracy; hence, we needed labels that address change but can be used independent of the model or conception of democracy. We used Dryzek’s [26] notions of franchise, scope, and authenticity of democracy (see below) as our initial labels. We then collected the key aspects of democracy addressed or/and democracy indicators used, geographical focus areas, and findings and/or conclusions of each paper to the tables.

Results

Topic modeling

We conducted a CTM on the abstracts of the 586 articles to scope the topics of research that mentions the COVID-19 pandemic and democracy. Nine topics offered the most consistent account in terms of semantic coherence and word-topic exclusivity trade-off. The nine topics were relatively equal in proportion. The most common topic was expected to appear in 14.2 per cent and the least common in 8.5 per cent of the abstracts. We interpreted the most probable and frequent and exclusive (FREX) terms to label the topics. We used R to randomly select articles whose abstract involves the topic to validate the labels.

Table 1 CTM results and topic labels

The labels of the topics are listed in Table 1. Some clarification is needed. We labelled the most common topic as ‘crisis’. Common to the abstracts of this topic is to represent something as (or as being in) a crisis. While the topic includes explicit discussions on the crises of democracy caused the pandemic, it also includes many other discussions such as the legacies of previous (political, economic or health) crises and their implications to the functioning of democracy under or policy responses to COVID-19. We labelled the second most common topic as ‘governance’, as the term was among the most frequent and expected in this topic. However, as the topic includes numerous articles that deal with politics, policies and leadership related specifically to national policy responses to the pandemic, it could be also titled ‘policy response’. We used this latter term later as a label in manual coding.

Three topics are relatively homogenous. The topic ‘rights’ largely revolves around issues of human rights during the pandemic, the topic ‘epidemiology’ around statistics regarding the pandemic, and the topic ‘media’ around social media and news. Two of the topics are much more heterogeneous. The topic labeled ‘society’ discusses a variety of civic and economic issues raised by the pandemic. Common to these is the highlighting the importance of non-state actors; hence, ‘society’. The topic labeled ‘technology’ discusses highly varied issues such as population surveillance, vaccine development, and remote work in equally varied contexts. Common to these issues is the concerns caused by the rapid adoption of new technologies under the pandemic; hence, ‘technology’.

Conceptual and theoretical contributions

The breadth of conceptual contributions varies in the scholarship reviewed. Broader conceptual contributions to democracy scholarship can be found in individual conceptual papers (see Table 2) and special issue introductions [27, 28], while critical assessments and essay articles typically address some more specific and limited aspects of democracy. Some have asked how the pandemic may have changed the conceptions of democracy [29]. As expected from early research on any topic, the body of literature involves few syntheses of research on the state of democracy under COVID-19. The notable exceptions here are Hellmeier et al. [30], who seek to synthesize the impacts of the COVID-19 pandemic to democracy through a comprehensive analysis based on the Liberal Democracy Index of the Varieties of Democracy (V-Dem) dataset, and Afsahi et al. [27], who build a synthesis of 32 democracy scholars’ early insights on the impacts of the pandemic to democracy.

Table 2 Articles with conceptual and theoretical contributions

Some articles conceptualize and specify the social and political conditions of a large-scale pandemic. One paper introduces an index to capture these conditions. Edgell et al. [31] have constructed The Pandemic Violations of Democratic Standards Index (PanDem) to assess the extent to which states have violated different types of human and political rights during the pandemic. The index is used widely in empirical research (see below). Another paper introduces a new conceptual approach to democracy to capture these conditions. Parry et al. [32] describe a ‘systemic view of democracy’ that conceptualizes how participation and deliberation has occurred across private, public, and empowered spaces of communication and contestation under the pandemic. While some have conceptualized different aspects of the COVID-19 pandemic from the perspective of democratic theory [33, 34], others have discussed the qualities of some aspects of democratic politics in the conditions of a large-scale pandemic [35,36,37,38]. Some have also discussed whether democracies are intrinsically inferior or superior to autocracies in dealing with pandemics [39].

The rest of the papers in this category conceptualize and critique some specific challenges or dangers that the COVID-19 poses to democracy. Nikolova [40] focuses on the legitimization of imbalances of powers and normalization of social distance as having negative effects on democracy, while Peng and Berry [41] focus on the negative impacts of the pandemic to freedom of movement and privacy. Otherwise, the focus ranges from macro-level issues such as government-civil society relations [42] to micro-level case studies such as particular surveillance technologies [43]. Some, but not all papers, provide prescriptions for dealing with the outlined challenge [44]. While most papers in this category are essays or conceptual papers with empirical examples, one review article also exists on democratic accountability [45].

The relations between democracy and the COVID-19 pandemic

Next, our attention turns to empirical research on the relations between democracy and the pandemic. Three key differences exist in this research depending on the direction of influence studied. First, the research strategies. Research on democracy’s impact on the pandemic is almost exclusively statistical, and typically based on different democracy indices (see Table 3) and either pandemic response indices or epidemiological statistics. Few articles that address the other direction of influence uses democracy indices. The articles addressing this direction are mostly descriptive and methodologically varied, typically deploying interpretive methods to assess the quality of impacts.

Second, the geographical scope of research varies between the two directions. Most research on democracy’s impact on the pandemic is based on global comparisons with more than 100 countries included in the analysis. The findings of the few articles that focus on a narrower group of countries offer somewhat different findings from these. Very few articles focus on the sub-national level and more in-depth comparisons of individual countries. In contrast, the research addressing the pandemic’s impact on democracy largely addresses individual countries or comparisons of relatively few countries.

Third, the scientific contribution of the published scholarship varies according to the direction. Few papers addressing democracy’s impact on the pandemic aim at theory-building. Few papers address or specify the mechanisms and processes through which democracy has tangibly influenced the pandemic in the cases studied. Due to the extensive reliance of this research on the existing indices, much of the future theory-building work to which this research may contribute is thus limited to factors that can be accommodated to the most popular indices. In contrast, most papers that address the pandemic’s impact on democracy engage with in-depth cases and pursue wider theoretical contributions with them.

Table 3 Democracy indices used in the reviewed research

The impacts of democracy to the COVID-19 pandemic

Policy responses

Most studies addressing the impacts of democracy to pandemic responses draw their indicators from various democracy indices, The Oxford Covid-19 Government Response Tracker (OxCGRT), or/and the PanDem index. Most analyses discuss a wide group of countries. Some single country analyses [46] and sub-national analyses [47, 48] also exist (see Table 4).

Table 4 Articles discussing the impacts of democracy to policy responses

One key topic in this research concerns the type and timing of policy responses. Research on the type of policy responses deals largely with the issue of stringency and violations of democratic principles in connection with specific measures. Dempere [49] argues that countries with the highest democracy indexes (using various indicators from different indices) applied the softest social constraints measured by the daily average stringency index. These countries exhibited the shortest outbreak response time and the most extensive daily average tests per thousand. Chiplunkar and Das [50] show that non-democracies (Polity IV) impose more stringent policies (OxCGRT) prior to their first COVID-19 case, but democracies close the gap in containment policies and surpass non-democracies in health policies within a week of registering their first case. Democracies with greater media freedom respond more slowly in containment policies, but more aggressively in health policies. Engler et al. [51] find that in countries where the quality of democracy (Democracy Barometer) is higher in normal times, governments were also more reluctant to adopt policy measures (OxCGRT and PanDem) that are potentially in conflict with democratic principles. Lundgren et al. [52] study the declaration of a state of emergency. They find that weak democracies (V-Dem) with poor preparedness (GHS) have been considerably more likely to opt for a state of emergency than dictatorships and robust democracies with higher preparedness. Research on timing offers more mixed findings. Chen et al. [53] do not find significant predictive power of democracy (EIU) on the speed of government responses. However, Sebhatu et al. [54] shows that governments in countries with a stronger democratic structure (V-Dem) were slower to react in the face of the pandemic but were more sensitive to the influence of other countries.

The remaining articles in this category focus on more specific policy areas. The econometric analysis of Erić et al. [55] shows that democracy contributes to the economic policy response to pandemic, while Lins et al. [56] study the impact of the political regime type (V-Dem) to the adoption of stay-at-home requirements (OxCGRT) and find no major influence.

Epidemiological characteristics

Like in previous category, most research on the impact of democracy to the epidemiological characteristics of the pandemic are focused on the country level and address broad country groups (see Table 5). In fact, only Palguta et al. [57] discuss sub-national issues: they show that COVID-19 infections grew significantly faster in voting compared to non-voting constituencies in the Czech Republic.

Table 5 Articles discussing the impacts of democracy to the epidemiological characteristics of COVID-19.

The key issues addressed here are the impacts of democracy to COVID-19 cases, deaths, and case fatality rates (CFR; i.e., proportion of people diagnosed with a certain disease and end up dying of it over time). Most studies associate democracy with higher levels of COVID-19 incidence globally. Using various indices, Dempere [49] and Karabulut et al. [58] show that countries with the highest democracy index scores suffered a more severe pandemic impact. Higher levels of incidence are especially found among countries being classified as having “full democracy” by the EIU Democracy Index [59, 60]. Similar findings can be found from narrower country groupings. For example, Jardine et al. [61] find that non-democratic regimes had much shorter doubling time of cases compared to functional democratic Muslim-majority countries.

Others suggest that the relation between the extent of democracy and COVID-19 incidence is not linear but is shaped by numerous moderating factors. For example, Achim et al. [62] find that in high-income countries higher levels of democracy (as measured by EIU and various V-Dem indices) reduce the spread of COVID-19 while in the low-income countries its influence is exactly the opposite. Chen et al. [63] show that democracy levels (EIU) moderate the effects of policies on infection and death rates (OxCGRT).

Research on the relation between democracy and COVID-19 deaths offers varied findings. Lago-Peñas et al. [64] find that the coefficient between the extent of political rights and COVID-19 deaths is negative and statistically significant but only for estimates using accumulated data up to September 2020. Annaka [65] shows that authoritarian countries do not necessarily tend to have fewer COVID-19 deaths than their democratic counterparts (as defined by Polity and V-Dem indices). Vadlamannati et al. [66] suggest that more equitable access to health care increases testing rates and lowers the mortality rate from COVID-19, but egalitarian democracy (V-Dem) shows the opposite effect.

Research on CFR has provided different findings over time. Research on the early stages of the pandemic associate democracy with higher CFR. Using the Polity IV index, Sorci et al. [67] found moderate evidence suggesting that countries with a democratic regime were those with the highest CFR. Norrlöf [68], using the FH index, finds that liberal democracies have a higher CFR than other regime types (although liberal democracies do not have higher cases per capita than other regime types). Serikbayeva et al. [69] find that the level of democracy (FH) has a statistically significant positive impact on CFR in non-free countries, and that the likelihood of a higher death rate is lower in non-free countries compared to free countries. Yao et al. [70], using the EIU index, suggest that a higher Democracy Index is associated with (and moderated by increased hospital beds and healthcare workforce per capita) more deaths from COVID-19 at the early stage of the pandemic in all countries. However, later research offers somewhat different results. Karabulut et al. [58], using various indices (FH, Polity, and V-Dem), show that the observed CFR are in fact lower for democratic countries in a longer time period.

The impacts of the pandemic to democracy

Research on the pandemic’s impacts on democracy requires some further tools for interpretation. We use Dryzek’s [26] three dimensions of democracy – scope, franchise, and authenticity – to map out different types of impacts to democracy. The three dimensions are not dependent on a particular theory or model of democracy but can be applied across different conceptions of democracy. Scope refers to the extent to which different areas of life are under democratic control. Franchise refers to the effective number of participants who exercise influence over a democratic decision. Authenticity denotes the degree to which democratic control is substantive (rather than symbolic) and engaged by competent (rather than incompetent) and reflective (rather than inconsiderate) actors.

Most research insights presented in the sample deal with only one of the three dimensions. Some (albeit few) papers discuss more than one dimension, and thus appear more than once in the following sections. The two exceptions that escaped our attempts to categorize their insights are the paper that introduces the PanDem index [31] and another paper that discusses the state of democracy in the world in 2020 [30]. These papers do not attribute the violations of democracy or the state of democracy to any one specific conception of democracy but to more general principles that are relevant to various conceptions of (liberal) democracy. The problem here is that the violation or enactment of principles can be interpreted differently depending on how exactly democratic politics is understood. For example, where one theory of democracy that focuses on citizen rights might regard the curtailing of freedom of movement as a curtailment of the scope of democracy, another theory that focuses on participation might regard it as a curtailment of democratic franchise. To avoid such conflations, we have excluded the papers from our analysis in this section. Their more general conceptual and theoretical contributions have been presented above.

Scope

Most papers belonging to this category focus on democratic institutions of decision-making (see Table 6). There are only two exceptions here, one focused on freedoms and another on political rights. Cassani [71], who studies the impacts of the policy responses to COVID-19 to citizen freedoms, finds a widening freedom divide between autocratic and democratic regimes. Kinowska-Mazaraki [72] shows that Poland curtailed the right of assembly and protest, hence limiting the scope of democratic action.

Table 6 Articles discussing the impacts of the COVID-19 pandemic to the scope of democracy

The remaining twelve papers of this category deal with the state of exception or related aspects of the expansion of executive powers and limiting of democratic accountability and deliberation. A few papers discuss ‘executive aggrandizement’ [73] during the pandemic. In some countries like Australia [74], the democratic accountability of the executive was (due to popular protests only temporarily) abandoned to provide leeway for the making of pandemic response policies. Some observe that technocratic policymaking by public health officials [73, 75] replaced democratic procedures in pandemic responses. Others observe a similar tendency in the case of the military [76].

A significant number of papers in this category argue that the pandemic aggravated the already ongoing and more general expansion of executive powers to replace previously democratic politics in ‘democratically backsliding’ countries. The argument has been made in the cases of El Salvador [77], Georgia [78], Hungary [79], India [80] and Indonesia [81]. Others have found that the pandemic has not deepened existing democratic deficiencies. This case has been made the European Union [82] and some individual countries like the Czech Republic and Slovakia [73, 79]. Lewkowicz et al. [83], utilizing the V-Dem indices and the PanDem index, show that the stronger the rule of law and the higher levels of electoral democracy, the lower the risk of democratic backsliding has been in the face of the pandemic. Previous strengthening of democratic accountability mechanisms has also been found to decrease the likelihood of democratic backsliding [84].

Franchise

All papers in this category discuss the impacts of the pandemic to elections (see Table 7). The difficulty of holding elections under a pandemic have been widely noted. Only a case study on Israel shows that election turnout can be maintained through effective containment procedures, logistics, and communications [85]. Otherwise, the articles of this category observe a decreasing voter turnout during the pandemic. The countries in which this has been observed include Chile [86], Ghana [87], Ethiopia and Mali [88], and India, Pakistan and Afghanistan [89]. Some also argue that the pandemic has halted the efforts to instill democratic elections [90].

Table 7 Articles discussing the impacts of the COVID-19 pandemic to the franchise of democracy

Authenticity

Three common themes can be found among the papers included in this category (see Table 8). First and the most common theme addressed here is related to democratic legitimation and justification of policymaking under the pandemic. Research suggests that some countries legitimized the expansion of executive powers democratically (e.g., Portugal [35]), whereas some others did not (e.g., India [91]). Mixed interpretations have been made regarding the Italian case [35, 92]. In some countries like Israel, the lack of democratic justification for executive aggrandizement led to wide popular backlashes, hence demonstrating democratic resilience [93]. But in Germany, a similar backlash did not occur, which raises questions about the degree of authenticity and resilience in the country [94]. Some have also addressed the preconditions for democratic emergency politics. Truchlewski et al. [95] argue that, by ‘buying time’ through effective emergency politics the EU enabled its member states time to democratically deliberate upon and justify their policy responses.

Table 8 Articles discussing the impacts of the COVID-19 pandemic to the authenticity of democracy

Another theme concerns the democratic virtues and vices of new communication technologies that popularized during the pandemic. Some positive impacts to authenticity are observed here. New online forms of election work are observed to have activated the youth to participate in election work in Singapore [96] whereas online scientific surveys offered South Africans a way to express popular views under lockdowns [97]. Anecdotal evidence suggests that even though online platforms may maintain or even enhance the quality of deliberation [98], they may also be unrepresentative of the broader communities [99]. Another discussion concerns the role of traditional information sources for democratic actorhood. Casero-Ripollés [100] observes that legacy media consumption surged in the United States during the pandemic. However, Baekkeskov et al. [101] also note that media discourses became much less deliberative and more monotonous during the pandemic.

The third common theme concerns the support for democratic politics. Here, the evidence is highly varied. Bol et al. [102] find that lockdowns increased satisfaction with democracy in Western Europe. But in the case of Italy, Pedrazzani et al. [103] report that evaluations of democracy became more negative with social proximity to the disease and with individual perceived vulnerability. Despite observing the rally effects documented in contexts of interstate conflict, no evidence of a broader shift in democratic attitudes due to the pandemic can be observed in Brazil [104] or Haiti [105].

Discussion

The early research on democracy and the COVID-19 pandemic offers a diverse body of literature. Our topic modeling suggests that the scholarship that mentions the two concepts deals with various issues: the nature of crises brought by the pandemic, epidemiological characteristics, political behavior, the governance of responses to the pandemic, the (temporary or longer-term) narrowing down of citizen rights amidst the pandemic, the virtues and vices of new technologies, and societal challenges and change. 94 articles discussed the relation between democracy and the pandemic more systematically. The body of papers with conceptual and theoretical contributions in this sample has offered new insights on the possibilities, difficulties, and means to maintain democracy under severe health crises such as pandemics. This research has given rise to new indices to track violations of democratic principles in crises, new criteria for governing rapid policy responses democratically, new ideas on how to organize elections under health crises, and warnings about the longer-term impacts of new policies, technologies, and discourses with anti-democratic qualities.

Empirical research on democracy’s impact on the COVID-19 pandemic and vice versa also offers a versatile body of research. We find that the methodologies used, the geographical scope of research, and the scholarly contributions vary according to the direction of influence studied. Research on democracy’s impacts on the epidemiological characteristics of and policy responses to the pandemic are largely based on democracy indices, country-group-level analysis, and varying timeframes. Democracy appears to have a significant impact on some aspects policy responses and epidemiological characteristics of pandemics. Be it about timing of policy measures, preferred types of measures, or preferences over the stringency of measures, democratic countries are likely to produce responses that somewhat differ from non-democratic countries as well as from each other. Democratic and non-democratic countries do not necessarily perform in a vastly different degree in dealing with pandemics in the short run. Beyond these observations, the results are somewhat mixed depending on the democracy indices, epidemiological and policy indicators, and time periods studied. Hence, further empirical research and meta-analyses are needed to say anything conclusive about democracy’s impacts on the pandemic. In-depth case studies and qualitative research is needed for theory-building.

Research on the pandemic’s impacts on democracy are largely based on qualitative research and discuss relatively few countries at a time. Many gaps still exist. For example, the impacts of the pandemic to democratic participation in the civil society and the quality of deliberation and representativeness in policymaking contexts were not explored systematically in the body of literature scoped here. Longer-term time series are needed to study the pandemic’s impacts on democracy globally.

Thus far, most findings concerning the impacts of COVID-19 to democracy raise some concern. In most parts of the world, the scope of democracy narrowed down due to the pandemic, albeit in most cases only temporarily. But in the already democratically backsliding countries, the pandemic offered new conditions for broadening executive powers time- and scope-wise beyond what may have been necessary to tackle the pandemic. The evidence concerning the franchise of democracy is very much limited to elections and election turnout, but it suggests that policy responses to the pandemic will have a major impact on the conduct of elections. Much is needed to maintain high degrees of democratic franchise. The authenticity of democratic politics has been compromised in various ways in the COVID-19 pandemic. Most importantly, the pandemic revealed alternatives to democratic politics. While the popular support for democratic politics decreased during the pandemic in some democratic countries, others broadly mobilized against policies without democratic justification. While new communication technologies are no panacea for maintaining authenticity, online channels may offer some opportunities to renew it.

Conclusions

If the conceptual and theory papers reviewed above have offered important insights and hypotheses for further research, then the empirical research reviewed gives equally important reasons to keep a close eye on future events and test the hypotheses. Many papers have argued that the pandemic has accentuated different forms of democratic backsliding but is unlikely to have undermined democracy as such thanks to various mechanisms from constitutional checks and balances to popular backlashes that have proven the resilience of established democracies. Yet, empirical research shows some weak signals of antidemocratic tendencies that may become more accentuated in the longer run, ranging from the emergence of anti-democratic discourses to positive popular reactions to authoritarian forms of governance. Thus, it remains to be seen whether the longer-term impacts of the COVID-19 pandemic will eventually prove detrimental to democracy, and whether democracy will remain as resilient in the next large-scale health crisis as it did under COVID-19.