When Is It Democratic to Postpone an Election? Elections During Natural Disasters, COVID-19, and Emergency Situations

Abstract

Election Law Journal: Rules, Politics, and PolicyVol. 19, No. 3 Free AccessWhen Is It Democratic to Postpone an Election? Elections During Natural Disasters, COVID-19, and Emergency SituationsToby S. James and Sead AlihodzicToby S. JamesAddress correspondence to: Toby S. James, University of East Anglia, Norwich Research Park, Norwich, Norfolk NR4 7TJ, United Kingdom E-mail Address: t.s.james@uea.ac.ukToby S. James is a professor of politics and public policy in the School of Politics, Philosophy, Language, and Communication Studies at the University of East Anglia in Norwich, Norfolk, United Kingdom. Sead Alihodzic is senior program manager at International IDEA in Strömsborg, Stockholm, Sweden. Search for more papers by this author and Sead AlihodzicToby S. James is a professor of politics and public policy in the School of Politics, Philosophy, Language, and Communication Studies at the University of East Anglia in Norwich, Norfolk, United Kingdom. Sead Alihodzic is senior program manager at International IDEA in Strömsborg, Stockholm, Sweden. Search for more papers by this authorPublished Online:17 Sep 2020https://doi.org/10.1089/elj.2020.0642AboutSectionsView articleView PDFView PDF Plus ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions Back To Publication ShareShare onFacebookTwitterLinked InRedditEmail View articleAbstractHolding regular elections is an essential feature of democratic practices. The case for postponing elections is often made during emergency situations, however. Despite the critical nature of the issue for democracy, peace, and security, there has been sparse academic literature on election postponement. This article provides a new typology of reasons why elections might be delayed to disentangle the causal factors and normative rationale. It distinguishes the humanitarian case for temporary postponements during natural disasters. It then argues that substantive concepts of democracy and electoral integrity, rather than existing international/national laws and standards, should be used to inform decisions about postponement by relevant stakeholders, be it an electoral management body (EMB), government, parliament, or the judiciary. The possible effects of natural disasters on electoral integrity are traced through a comparative analysis of past cases. The article holds that variations in context and the ability of actors to strategically adapt to situations will make the effects contingent. Nonetheless, holding elections during natural disasters will often lead to severely compromised opportunities for deliberation, contestation, participation, and election management quality. There is therefore a strong, democratic case for time-limited postponement. However, the postponement will break institutional certainty, which could pose threats of democratic breakdown—especially in presidential systems. The best available safeguards for electoral integrity during natural disasters include the introduction or expansion of low-tech solutions such as early voting, strengthened risk management, but also transparency and inclusivity in decision making. Overall, there are important lessons for the broader scholarship and practice of democracy during emergency situations.IntroductionOne of the defining characteristics of a democracy is that it holds regular, periodic elections (Dahl 1971; Przeworski 1999). This requirement was famously enshrined into Article 21(3) of the Universal Declaration of Human Rights (UN General Assembly 1948). It is also embedded into many conceptualizations of electoral integrity (Norris 2014). At the same time, there are occasions where a natural disaster such as an epidemic or tsunami may mean that holding an election will potentially introduce greater threats to human life and security. The problem was laid bare as the coronavirus pandemic spread in 2020. Between February 21 and August 18, 2020, elections in at least 70 countries and territories across the globe had postponed some elections. And yet during the same time peroid, plans for national or subnational elections in at least 54 countries continued (International IDEA 2020a). Court case battles ensued elsewhere, with the Wisconsin Supreme Court blocking an executive order by the state governor to suspend in-person voting for the presidential primaries in April 2020 (Wisconsin Legislature v. Evers 2020).The question of whether elections should be delayed in such circumstances has rarely been given detailed analysis. It has arisen in rare court judgments, such as that in Wisconsin, where arguments have principally focused on constitutionality. The question, however, is clearly critical for practitioners, judicial judgments, and democratic theory as it raises much deeper arguments and is therefore a major gap in the broader democracy and electoral law literature.This article firstly contributes to this literature by providing a new typology of reasons why elections might be delayed to disentangle the causal factors and normative rationale. Seven separate categories are set out which include those which are predominately postponed because of the political interests of the incumbent, but also those which are postponed for technical reasons, peaceful conciliation, or humanitarian postponements that might be necessary because holding an election during a natural disaster would pose a threat to human life and health.Secondly, the article contributes by making the case that the decision of whether to postpone or hold an election should be subject to assessment against broader democratic theory rather than international law and standards. The latter provide useful benchmarks, but there is no guarantee that these international laws and standards are democratic themselves—or best practice in rapidly changing situations. A substantive theory of democracy is used to identify five key principals of electoral integrity which are at risk when natural disasters hit.Thirdly, the article makes the case that the effects of natural disasters on electoral integrity are unpredictable, not just because the nature of natural disasters and contexts will vary, but because actors have agency and can therefore strategically learn and adjust. “Iron laws” about the effects of natural disasters on electoral integrity may therefore not hold up. However, the article holds that there are major threats to opportunities for deliberation, contestation, participation, and election management quality posed by natural disasters. The postponement paradox is that postponing will break institutional certainty, and this may lead to partisan scrobbling which could trigger democratic breakdown and undermine trust in the system—especially presidential systems that are used to fixed terms and states with low levels of political trust. We make specific recommendations about how elections could be conducted during natural disasters, which are that low-tech solutions are used—but also that the process of deciding is as important as what decision.The article will begin with a review of the very limited existing research that has been undertaken on the postponement of elections. It will distinguish the new concept of humanitarian postponements within a new typology of instances where scheduled elections might not take place or have been cancelled at any stage of the electoral cycle to disentangle the causal and normative reasons for this. Next, it will scrutinize the case for the humanitarian postponement of elections or electoral preparations in the face of looming or actual natural disasters which will threaten human life. It will then compare how holding or postponing an electoral event during emergency situations would affect the realization of five key democratic principles: opportunities for deliberation, equality of participation, equality of contestation, electoral management quality, and the institutionalization of rules. Finally, the article outlines the measures that can be taken to protect electoral integrity during a natural disaster. Overall, this article has important lessons for the broader scholarship and practice of democracy during emergency situations.The Postponement of Elections: what do we KNOW?Although elections have been one of the most prominent areas of study in political science (Htun and Powell 2013), the elections that do not take place remain systematically understudied and under-categorized. There is some provision in international law and standards. Article 21(3) of the Universal Declaration of Human Rights notes that states must have regular, periodic elections (UN General Assembly 1948). Article 4 of the International Covenant on Civil and Political Rights, meanwhile, states that “in time of public emergency which threatens the life of the nation and the existence of which is officially proclaimed, the States Parties … may take measures derogating from their obligations under the present Covenant to the extent strictly required by the exigencies of the situation.” International electoral assistance agencies have issued guidance that any emergency provisions must be proportionate, nondiscriminatory, temporary, and limited in scope (Ellena and Shein 2020).Political science research is much more limited, however. Datasets on electoral contests tend to measure the quality of electoral integrity or the opinions of voters in elections that have taken place (Hyde and Marinov 2012; Kelley 2011; Norris, Wynter, and Cameron 2018) and the drivers of electoral integrity within those events (Birch 2011; James 2020; Norris 2015), rather than those elections which did not take place. We therefore know little about “the dog that didn't bark in the night.” Elections that do not take place appear to be common, however. The National Elections Across Democracy and Autocracy (NELDA) dataset on elections lists 144 states that have had some previous experience of a “suspended election” between 1945 and 2015 (Hyde and Marinov 2012).1 Examples cited in that dataset include civil wars such as the fighting between the ruling People's Movement for the Liberation of Angola (MPLA) and the rebel National Union for the Total Independence of Angola (UNITA) in Angola or a military coup such as that under Pervez Musharraf as the Pakistani leader.There has been a wider discussion about the timing of elections. This has considered issues such as the effects of early “snap” elections, where it is constitutionally permissible to hold elections earlier than planned (Smith 2004). The timing of elections are also thought to be important with respect to post-conflict situations and democratic transitions (Alihodžić and Matatu 2019; Reilly 2002). Research on natural disasters and elections has considered the effects on vote share and turnout (Abney and Hill 1966; Flores and Smith 2013). There has been some discussion about the potential steps that election officials can take in the event of a natural disaster in the U.S., with Stein (2015) examining the conduct of the 2012 presidential election in the aftermath of Hurricane Sandy. Morley (2017) similarly considered how U.S. states responded to natural disasters such as the September 11 attacks on New York City, Hurricane Katrina's destruction of New Orleans, and the constitutional cases that then followed. There have also been discussions of specific legal cases where courts have been asked to intervene to postpone an election, such as the 2013 California Gubernatorial Recall Election (Brady 2004). But once an election has been scheduled, should it take place come what may, or are there some circumstances where it should be delayed? There has been little critical examination of the case for postponing elections against democratic theory.Seven Types of Non-ElectionsThe motives and causes of not holding elections are varied and complex. It is important to distinguish these because non-elections are not just power grabs by autocrats—the most intuitive reason for why an election might not take place. Table 1 therefore presents a new ideal-type typology of non-elections. The actors who cancel the election, their rationale, and timing vary in each case. As an ideal-type typology (Collier, Laporte, and Seawright 2008), cases may not always precisely fit into one single category and may have some aspects of each of them. However, the categories provide an important advance in identifying the variety of causes for postponing an election—which can help academics and judicial decision makers determine whether or not they are power grabs and undermine democracy—or are actually necessary steps to reinforce democratic ideals.Table 1. A Typology of Delayed Elections Stopped by whom?Why?When?CancellationsIncumbentMilitaryPower-grabPre-electionCrisis postponementsIncumbentOppositionPolitical stalemateConstitutional crisisViolencePre-electionTransition postponementsTransitional governmentDemocratic reformsPre-electionTechnical delayEMBIncumbentTechnical issuesViolencePre-electionCandidate deathEMBIncumbentEquality of contestation and voter choicePre-electionHumanitarian postponementIncumbentEMBEmergency situation threatening lifePre-electionAnnulledJudiciaryIncumbentMilitaryNot constitutionalPower-grabNot implementablePost-electionEMB, electoral management body.Cancelled elections are those in which incumbents decide not to hold an electoral event as part of a deliberate statecraft power grab. These cancellations can be long term or indefinite where, for example, power is seized through a military coup. There are numerous cases of elections being suspended for prolonged periods of military dictatorships, such as in Lesotho 1986–1993. Pervez Musharraf postponed the 2008 elections in Pakistan. However, they can also be short-term cancellations where leaders might be expecting that they will not be successful in a future contest. Some Venezuelan elections were postponed in 2016–2018 as part of Nicolas Maduro's strategy to undermine opposition parties and their followers, or to buy time to mitigate negative popular opinion (Alarcon and Trak 2019, 126–27, 9–30).Cancelled elections violate democratic norms, but there are a variety of other reasons why elections might not take place. Deep political crises such as those in FYR Macedonia (2012–2016) can cause an institutional breakdown, preventing scheduled elections from taking place (Alihodžić and Matatu 2019). Crisis postponements may therefore occur because there is a political stalemate and constitutional crisis. The causes of a non-election are therefore deeply political, but not a direct power grab, and there is no indefinite cancellation of elections. Parliamentary elections in Egypt were postponed in 2013 when an Administrative Court overrode a decree issued by President Mohamed Morsi calling the election. It also returned the electoral law, the subject of feuding between the opposition and Morsi's ruling Islamists, to the Constitutional Court for review, leaving Egypt “in limbo” (Saleh 2013). Similarly, the Constitutional Court in Macedonia declared the dissolution of parliament unconstitutional in 2016, thereby cancelling the elections (OSCE/ODIHR 2016).Transitional postponements are those where states determine that a broader constitutional framework needs to be established before general elections are held. Both Nepal and Tunisia decided to adopt new democratic constitutions before conducting general elections. To do so, they held Constitutional Assembly elections in Nepal in 2008 and in Tunisia in 2011. The 2007 Constitutional Assembly elections had to be postponed in Nepal due to the absence of a necessary legislative framework (Pokharel and Rana 2019). The motives of actors in transitional postponements are therefore about deepening democratic institutions rather than a power grab, and the constitutional reforms being put in place should involve a clear framework for the holding of elections.Technical delays might also be needed. Elections are huge, complex logistical events which require a vast volume of resources, laws, and staff (James 2020). It might therefore be necessary to delay parts of the electoral process until staff training has been completed, robust information and communications technology (ICT) systems have been established, or electoral registers have been compiled. The Nigerian 2019 presidential and national assembly elections were postponed by the Independent National Electoral Commission, three days before the scheduled elections “following a careful review of the implementation of the logistics and operational plan” (McKenzie, Swails, and Smith-Spark 2019). The decision was announced only five hours before polls were due to open, and the election was put back a week. Delays can be longer. The Bosnian post-war elections in 1996 were postponed for three months due to technical delays. The 1991 census was to be used as the electoral register—but there were mistakes scanning the data. Meanwhile, voter identification was hampered by the problem of issuing documents to internally displaced refugees (Hadziabdic 2019).Elections might be postponed when there is the death of a candidate. This situation has been described as “quite common in local elections” (Rallings and Thrasher 1997, 44). This is especially problematic where the candidate is individually listed on the ballot paper rather than being a candidate through a party-list system. The procedures for postponement or other steps are usually set out in national laws and may depend on the timing of the death with respect to the electoral cycle and the affiliation of the candidate (ACE 2020). In the United Kingdom, if an independent candidate dies ahead of a parliamentary election, then the election continues and is only rerun if the deceased candidate receives the most votes.2 If the candidate is representing a political party, however, then the election is halted immediately, and a new election is called (Electoral Commission 2019, 33). By contrast, in the Republic of Ireland, voting was delayed across the Tipperary constituency in the 2020 general election when an independent candidate died five days ahead of the polls (Holland 2020). The democratic rationale for postponement is a concern that voter choice would be adversely affected if a political party is unrepresented on the ballot paper and parties will be unequally able to contest the election.An outlier is the annulled election, which is a polling event that takes place as scheduled but which subsequently has the result overridden. This might be by a constitutional court who might judge that the election was not constitutional for technical reasons, or because there was evidence of widespread electoral fraud (Vickery, Ennis, and Ellena 2018). For example, the result of the 2019 Malawi presidential election was annulled by the Constitutional Court after it ruled that “anomalies and irregularities have been so widespread, systematic and grave such that the integrity of the result was seriously compromised, and can't be trusted as the will of voters” (Reuters 2020). The 2017 Kenyan presidential elections were annulled after the Supreme Court said that the polls were “neither transparent nor verifiable,” blaming the country's electoral commission for the shortcomings (Burke 2017). Annulled elections are an outlier in so far as the electoral event did take place and the cancellation is post-election. They therefore feature last in the typology. However, we include them because they are like other categories in that they are electoral contests which are scheduled but do not have their results verified.We can also distinguish humanitarian postponements where elections might not take place because of threats to human life, which are considered in more detail next.The Humanitarian Case for Delaying ElectionsThe loss of life during electoral events as a result of electoral violence has seen considerable academic attention. This is usually seen through the lens of deliberate attempts to suppress the opposition vote or wreck the electoral process as part of a campaign to undermine the legitimacy of the government or democratic process (Birch, Daxecker, and Höglund 2020; Höglund 2009). There has been much less focus on the loss of life that can occur where no deliberate perpetrators were involved. There is enormous scope for this to occur, however. The National Electoral Commission (KPU) of Indonesia reported that over 300 poll workers died as a result of fatigue-related illnesses following the 2019 elections (Manafe and Yasmin 2019), for example.There is an obvious humanitarian case for delaying the conduct of elections where it might bring about immediate threats to human life and security. Electoral democracy is a political system which is valued for bringing about democratic ideals—but it is also valued for bringing greater material well-being to citizens (Sen 1999b). Holding elections when they might jeopardize lives would therefore be a counterintuitive use of institutions designed to facilitate individual and collective preservation.There are a variety of scenarios in which the loss of life might occur during the electoral process which were not the result of deliberate action. Natural accidents are major adverse physical events, resulting from natural processes of the Earth. They include earthquakes, tsunamis, wildfires, epidemics, and pandemics (Table 2). There are many examples worldwide of when holding elections during such unexpected events would have threatened human life. The 2019 local elections in Papua New Guinea were forced to be postponed because of an eruption of Mount Ulawun, forcing thousands of people to be displaced (RNZ 2019). Haiti's electoral plans were devasted in 2020 by an earthquake (Zengerle and Guyler Delva 2010). Heavy rains from Cyclone Idai killed more than 750 people and forced the National Electoral Commission to postpone the electoral census in Mozambique (Machel 2019). Climate change forecasts estimate that some of these problems might become more frequent.Table 2. Examples of Emergency SituationsScenarioExamplesNatural hazardsGeophysical (earthquakes, landslides, tsunamis, volcanic activity)Hydrological (flood, landslide, wave action)Climatological (drought, wildfire)Meteorological (cyclones, storms)Biological (epidemic, pandemic, insect/animal plagues)Extraterrestrial (space weather, meteoroids)Technological or man-made hazardsConflictsFaminesDisplaced populationsIndustrial accidentTransport accidentMiscellaneous accident (collapse, explosion)Technological failureSource: authors, based on CRED (2020).One prominent example of a pandemic is the spread of COVID-19 in 2020. A pneumonia with an unknown cause was detected in Wuhan, China, and first reported to the World Health Organization (WHO) Country Office in China on December 31, 2019. The outbreak was declared a Public Health Emergency of International Concern on January 30, 2020. Cases were rapidly reported around the world and it was declared a pandemic by the WHO on March 11, 2020 (WHO 2020b). Between February 21 and August 18, 2020, at least 70 countries and territories across the globe decided to postpone forthcoming elections (International IDEA 2020a). Elections that have previously been postponed for public health reasons include those in light of the Ebola crisis in West Africa (2013–2016), the 2016 presidential election in the Democratic Republic of Congo, and the 2014 senatorial elections in Liberia (Mark 2014). Meanwhile, during the 2009 epidemics of Swine Flu (H1N1) virus in Mexico, campaign restrictions were put in place, such as discouraging political rallies of more than 40 people, urging candidates not to wear neckties (as they were suspected to be potential carriers of viruses), and discouraging kissing (including baby kissing) and handshaking (Lacey and Malkin 2009).There are major risks of casualties throughout the electoral cycle during a natural disaster because it involves the rapid movement and concentration of people within a country and between countries. The compilation of the electoral register often requires the posting of registers outside of polling stations so that citizens and parties can inspect them and door knocking at residential properties to check the accuracy and completeness of the register. The campaign stage of elections involves holding major rallies and political parties reaching out to voters via the canvassing of properties. Election day itself can involve the movement of millions of citizens. Many countries do not provide external voting opportunities and so citizens must travel from overseas in order to vote (Hartmann 2015). Voting involves staffing thousands of polling stations with poll workers—who are often retired and elderly (Clark and James 2017). Counts often take place in crowded areas such as halls or within polling stations where social distancing is difficult with many people present. Touch-screen equipment, or sharing pens to mark a ballot or to sign a voter register, are opportunities for the transmission of infectious diseases.Natural disasters can be distinguished from technological disasters, which are human-made and occur in or close to human settlements (Table 2). Industrial, transport, and other accidents such as rail crashes and explosions which are accidental may justify a humanitarian postponement of part of the electoral process. A technological failure such as a power outage can cause major problems to the electoral process now that technology is often central to the electoral process through e-counting and electronic poll-books, for example (Loeber 2020). Some technological disasters, whose causes are more complex in nature, might not justify humanitarian postponement. Famines can be caused by crop failure owing to environmental conditions—but they are usually also at least partially caused by human failure to prevent famine and weak accountability systems (Devereux 2009). They will therefore not usually justify humanitarian postponements. Famines were famously thought by Amartya Sen (1999a, 178) to be prevented by the holding of elections. But at their peak, unexpected famines could place humanitarian strains as state resources address human health and leave public services undeliverable. External wars, civil wars, and violence can create situations in which human life would be threatened by an election, but they are best diagnosed as crisis postponements.The humanitarian case for postponing elections is also strictly time-limited to that necessary to make elections deliverable, otherwise there are perverse incentives for leaders to exacerbate an emergency to deliberately postpone an election. A protracted emergency would usually provide an opportunity for electoral officials to put in place alternative election arrangements. The U.S. outbreak of COVID-19 in early 2020 gave sufficient opportunity for postal voting mechanisms to be in place for the November 2020 U.S. presidential election, for example (Persily and Stewart III 2020), although there were clear legal and political challenges. Emergencies are sometimes anticipatable. Floods and hurricanes are more likely in some regions than others. Risk assessment and contingency plans can therefore be developed. A late postponement, following a failure to take reasonable action to prepare for an election given a known threat, is therefore closer to a cancelled election if there are also partisan motives at play.Democratic Theory in PracticeAlthough there might be humanitarian reasons for postponing an election, the prospect of not holding an election when one is scheduled is, at first glance, contrary to democratic ideals. To assess whether postponing elections is in line with democracy and electoral integrity requires definitions of these concepts. If a procedural definition of democracy and electoral integrity is used, which defines the concepts in terms of whether a particular set of institutional practices are in place (e.g., Dahl 1971; Przeworski 1999), then it is possible that any postponement will be considered antidemocratic. As David Beetham argues, the weakness of procedural definitions, however, are that they provide no rationale for why these institutions should be considered “democratic,” in the first place, “rather than, say, ‘liberal,’ ‘pluralist,’ ‘polyarchic,’ or whatever other term we choose” (Beetham 1994, 26). International law may or may not allow the postponement of elections—but does that make this just and democratic?Democracy, based on the work of David Beetham (1994), is here considered as a system that achieves political equality and popular control of government (also see Beetham et al. 2008; International IDEA 2019). This is a substantive theory of democracy that considers whether key principles have been achieved, rather than a checklist of procedures. We define electoral integrity as the realization of principles in the conduct of elections that are necessary to support the broader realization of democratic ideals. Building from Garnett and James (2020), these would include, but are not limited to: opportunities for deliberation, equality of participation, equality of contestation, electoral management quality, and the institutionalization of rules.This article now seeks to map some of the underlying causal mechanisms for how natural disasters can shape electoral integrity. This enables the democratic dangers and advantages of election postponement/continuation to be identified. Our argument is that context is important because the prior legal, organizational, economic, and political relationships will vary. Key actors such as incumbents, political parties, and electoral management bodies (EMBs) also have some creative agency and may strategically learn to identify and adopt different strategies. Accurate prediction of the effects of natural disasters on electoral integrity may not entirely be possible, but we can b