At first glance, there is no obvious institutional explanation for the remarkable manner in which Swedish public policy diverged from the European norm when the novel coronavirus arrived early in 2020. After all, Sweden has all the classic trappings of a typical parliamentary democracy. Its government and prime minister are accountable to a majority of the country’s 349 parliamentarians, who themselves are chosen by the electorate every 4 years. Political conflict is channelled largely through political parties. The judicial branch plays a limited role in politics. Yet as most of its neighbours imposed sweeping behavioural restrictions in trying to forestall the spread of the virus, Sweden tried merely to slow it down and to mitigate its effects through comparatively limited and modest measures (see Baldwin, 2021). Public and private organisations remained, for the most part, open and active through the pandemic. Authorities tended to issue advice and voluntary recommendations rather than legally binding instructions.
As is now well-known, assessments of countries’ relative performance in dealing with covid-19 are complicated and contentious. Much depends on the relative importance ascribed to particular indicators and the benchmarks that are considered appropriate. If, for example, we compare accumulated excess mortality in European countries since the start of the pandemic, Sweden lands in mid-table. A narrower comparison between Sweden and its nearest three Nordic neighbours puts Swedish policy in a less favourable light. On the other hand, there may be less immediately discernible, long-term benefits from, say, keeping open schools for children up to the age of 12 throughout the entire pandemic, as Sweden did.
Our objective in this article is not, however, to evaluate the results of Swedish policy vis-à-vis the pandemic. We focus instead on the reasons for the divergence of this policy from the European norm, which implied comparatively stricter and more comprehensive measures, including “lockdowns”, encompassing testing, and mask mandates. At the centre of our study is the Swedish Public Health Agency, an expert authority. One condition that facilitated the outcome was that the agency strongly preferred Swedish policy to take the direction that it took. However, we look much more closely at a second condition, which we argue was equally necessary for bringing about Swedish policy’s idiosyncratic character. This condition involved the control over policy-making that the agency acquired during the pandemic. It was, we suggest, an extreme example of politicians delegating decision-making authority to experts.
In this article, we describe, classify and, to some extent, explain this deviant case of policy-making in the teeth of an international crisis—what might be described as a modern version of “Swedish exceptionalism,” a phenomenon that, in other aspects of politics, observers had long thought was fading (e.g., Pierre, 2016; Rydgren and van der Meiden, 2019; also Jerneck, 2021, p. 8). Our argument about the delegation to experts is that the structure of Swedish public administration was certainly relevant, but that several other circumstances—an unusually weak and deferential political leadership, a strong and assertive bureaucratic leadership, and the high level of public trust in state institutions—led to this delegation, and thus facilitated the outcome of policy exceptionalism.
The remainder of the article unfolds as follows. First, we review literature about the juxtaposition of technocracy, or the rule of experts, and representative democracy, both generally and in Sweden in particular. After briefly describing our methods and data, we turn to an account of our case. This section starts with the initial Swedish response to the onset of the pandemic, in which the authorities took several positions that continued to shape policy thereafter, while the government took a back seat. The section continues by examining how policy was subsequently modified, though slowly; and how ministers sought, at times, a somewhat more central role. Thereafter, we develop our thesis about the conjunction of public-sector organisation and leadership, political circumstances and public values. Finally, we draw conclusions from our study.
Democracy, Delegation and Expertise
In this section, we discuss debates within political science about the questions of representative democracy, accountability and the influence of experts in decision-making. These are essential for understanding the context of the case.
Representative democracy enables citizens to delegate power to elected representatives. They then hold these representatives accountable for their actions. The responsibility of politicians in this system lies in deliberation; in making laws and taking political decisions; and in recruiting and supervising the public servants, or bureaucrats, who implement these decisions. In practise, however, there are often great information asymmetries between highly skilled and specialised bureaucrats, which in turn may lead to policy drift or agency loss (Strøm, 2006). This is exacerbated when the remit delegated to administration is very broad or when contentious political decisions are delegated to administration—for example, to avoid blame (Hood, 2002). Indeed, blame games—to avoid and allocate blame to others—is common and politically consequential during and after major societal crises, particularly when they appear “uncontrollable” and when the loss of lives is considerable (Kuipers and ‘t Hart, 2014).
Of course, the virological, epidemiological and more general expertise on infectious diseases and public health are especially significant for effective management of pandemics. The question, then, is: what kind of relationships develop between government and experts?
A first body of research focuses on (1) the political uses of expertise. It explores how and when expertise is picked up in policy processes and used by political decision-makers. A second area of research addresses (2) the political influence of experts. It examines how and when experts and expert communities gain authority and occasionally achieve political influence, usually within specific policy domains. It includes the literature on “epistemic communities” and “epistemic authority,” also called networks of “knowledge elites” (Haas, 1992, 2016; Jasanoff et al., 2021; Jerneck, 2021, p. 14).
Less attention has been paid to an additional type of relationship between expertise and policy, which we claim applies in the Swedish response to the covid-19 pandemic: (3) when policy-makers delegate political authority to experts—that is, when non-elected experts are made responsible for shaping and implementing policy (cf. Jerneck, 2021, p. 13; Wahlberg, 2021, p. 346).
At first glance, the “model” of delegation of authority might seem to resemble Plato’s idealised republic, where technocratic knowledge elites are in power, making key decisions on how societal life is to be organised and resources allocated. In reality, however, the delegation of political authority to experts addressed herein does not imply that policy is purely evidence-based, but rather that they are mandated with a wider responsibility to shape policy. This is not only based on science. As a consequence of the sheer size and complexity of the mandate, it also implies responsibility for balancing competing interests, norms and values. The kind of power exercised in this model thus resembles that of “street-level bureaucrats” (Lipsky, 1980), “policy entrepreneurs” (Kingdon, 1995) or what more recently has been called “policy professionals” (Svallfors, 2016). It relates to how non-elected civil servants and consultants exercise political power in that they make decisions on who gets what, when and how (to paraphrase Lasswell, 2018).
The political nature of this model is a result of the mandate given to the experts in question. It can involve an explicit allocation of the power to make authoritative decisions; or an allocation of a broad mandate covering multiple issue-areas which necessitates balancing of conflicting interests and values; or a combination of the two. An example of the former is when a government permits the military to make decisions on when and how to use force, legitimised by national security concerns. An example of the latter is when the government mandates a public agency to control not only contagious disease but also public health in general, which implies balancing the goal of limiting contagion with the negative impact of measures designed to do just that (for example, the psycho-social and economic impact of “lockdowns”).
Given that the third model presented here implies delegation of political authority to experts, the question of categorisation of political expertise comes into play. Experts with decision-making authority might challenge or reject external expertise. When experts battle experts, it boils down to a question of who controls policy formulation and problem definition.
Methods, Concepts and Data
Ours is a study of a deviant case, one in which the outcome deviates from “established generalisations” (Lijphart, 1971, p. 692). The deviation we are concerned with is based on empirical observation—that Sweden’s response to the pandemic deviated considerably from that of other liberal democracies, what we call the European norm. Specifically, the Swedish case is deviant both with regard to the “dependent variable”—the Swedish policy response implied fewer and weaker mitigating measures—and with regard to a crucial “independent variable”—policy-making was delegated to an expert agency, in contrast to most other liberal democracies which maintained central government lead.
We are focused primarily on the Swedish authorities’ decisions in managing the pandemic, the most important of which were taken in early 2020. We will call these decisions a strategy. We demarcate the case temporally, then, by starting it as the novel coronavirus began to spread from China in January 2020, and concluding it a little less than a year later, as vaccines started to become available. (The Swedish vaccination programme deviated from that in other EU countries only at the margins, so we do not take up this part of the strategy.) However, we occasionally refer to events that occurred somewhat later, in order to offer evidence for conditions that may also have applied earlier, when policy was being established and consolidated.
The ensuing empirical analysis first addresses a series of temporal episodes during the pandemic crisis, which turned out to be pivotal “windows of opportunity,” arguably facilitating both particular pandemic policy decisions, and delegation of authority. The concept of crisis is typically understood as implying perception of threat, urgency, and uncertainty (Boin et al., 2005)—three features which certainly applies to the covid-19 pandemic, at least initially. Embedded in the crisis concept as well as convincingly demonstrated in past research is that a crisis not only implies danger, but also opportunity, that it can open the door for policy change and reform (Boin and t’ Hart, 2022). Thus, a crisis implies a series of temporal events or “junctures” involving external shocks as well as decision occasions, which might imply “institutional flux” (Capoccia and Kelemen, 2007, p. 341), demonstrated for example in new forms of delegation of authority. Secondly, the empirical analysis identifies specific conditions which came to the fore during different temporal episodes (cf. Mahoney, 2021, p. 321–322)—mainly with regard to specific features of Swedish public policy and public administration.
The data that we use in this chapter are drawn mostly from secondary sources and from contemporary media reports and statements from public authorities. When media sources involve a text written by someone whose identity is significant in the evaluation of the content, or which involves evidential claims based on the author’s own research, we treat that text as a journal article and include it in our study’s reference list. If, however, the source is a news report in which an event or remark is reported without interpretation, or it comprises a statement by a public agency or the government, we note the source only in parenthesis in the text. All such media and official references are from 2020 unless stated. The full list of these sources, with links to the original publications, is available in an online (Appendix).
We also refer to some of the findings of a public inquiry, known as the Corona Commission, into the management of the pandemic by all public authorities, including the government itself. It was appointed in June 2020. In December the same year, the inquiry published its first installment, on the protection of the elderly (SOU 2020:80, 2020). In autumn 2021, it published its second part (SOU 2021:89, 2021). Its final installment was published in late February 2022 (SOU 2022:10, 2022).
The Case: Sweden and the Pandemic
In this section, we describe the case. We highlight several idiosyncratic components of Swedish strategy: the authorities’ view of asymptomatic/pre-symptomatic contagion; their approach to interventions such as quarantine, testing and masks; their emphasis on recommendations rather than legally binding instructions; and the withdrawn role of government.
Initial Policy and Policy-Making
When the pandemic arrived, Sweden was, in some ways, ready. As recently as in December 2019, the Public Health Agency (Folkhälsomyndigheten, FHM) had issued a policy document on the threat of pandemics, such as new strains of influenza. It defined the goals of its strategy as to “minimise death and sickness in the population” and to “minimise other negative consequences for individuals and society,” through the use of “medicinal […] and non-medicinal measures (such as social distancing)” (Folkhälsomyndigheten, 2019, p. 6).
In April 2020, some weeks after the first wave of covid-19 had taken hold, the Swedish government published a text on its website in which its strategy for dealing with the disease was summarised (Regeringskansliet, 7 April). “The overall goal of the government’s work,” it declared, “is to reduce the rate of the spread of infection, i.e., to flatten the curve so that not many people become ill at the same time.” It seemed clear: the authorities sought to mitigate the effects of the pandemic rather than suppress the virus (also Nature, 21 April). This was despite the dramatic developments in the weeks prior to the publication of the strategy, which had persuaded other countries in Europe to abandon mitigation and take much more comprehensive steps to inhibit contagion. Britain, for example, changed course and imposed what became known as a “lockdown” in mid-March (Freedman, 2020, p. 52–54).
For sure, the Swedish authorities did take measures. Many recommendations, which were formally binding but involved no sanction for non-compliance, were issued, mainly about social distancing. Some more enforceable restrictions were imposed, such as on public gatherings, though they were milder and came appreciably later than in neighbouring countries (Ellingsen and Roine, 2020, p. 8; Karlsten, 2021). Universities and upper-secondary schools switched to digital learning. Otherwise, however, kindergartens and comprehensive schools stayed more or less open. So did shops. So did bars and restaurants, albeit sometimes with restricted opening hours. In Figure 1, the Swedish deviation in early 2020 is captured in quantitative comparative data. Notably, this diagram indicates mainly the varying temporality of policy responses, not the kind or qualitative aspects of decisions made. It shows primarily how Sweden’s initial response was slower than most comparable countries.
Figure 1. Initial coronavirus responses in Northern Europe. Governments’ response to the pandemic are coded into 23 indicators, such as school closures, travel restrictions, vaccination policy, and aggregated on a scale of 0 (no restrictions) to 80 (maximal restrictions). Source: Hale et al. (2021).
As became clear later, during an inquiry by the parliamentary Committee on the Constitution, there was never any formal government decision about the strategy that the Swedish authorities would pursue (Gummesson, 2021). Policy was instead left mostly to public agencies, primarily FHM. It was, in other words, a case of delegation of policy-making authority to an expert-run agency. That agency’s most prominent figure was the head of its Department of Public Health Analysis and Data Management, Anders Tegnell, the country’s chief epidemiologist (statsepidemiolog). He took centre stage in the daily press conferences that the agency began holding in March 2020, and he quickly attained an extraordinary degree of public recognition. Ministers did quite often announce adjustments to policy, even fairly minor ones, at their own press conferences. In practise, however, they were—with a few exceptions—announcing decisions that FHM had made.
FHM was initially relaxed, even dismissive, about the danger to public health. In late January 2020, the agency saw little risk that the virus would reach Sweden (SR, 27 January). A month later, Tegnell was still doubtful (AB, 26 February; Karlsten, 2020a). Even in early March, he remained sceptical about any prospect of community transmission there (SR, 2020). These erroneous forecasts were presumably rooted in its understanding of the novel coronavirus’s spread. At first, FHM saw little risk that infected but pre-symptomatic or asymptomatic people were infectious. Gradually, and discretely, it backed away from that stance. Yet its initial position had big policy consequences, which meant that Sweden diverged markedly from restrictions imposed elsewhere in Europe.
The conviction that infectiousness was almost exclusively associated with symptoms meant, for instance, that there was no need to isolate returning holidaymakers in February and March, even if they had been in Alpine regions badly hit by the virus (SvD, 7 March). It was enough to recommend staying at home if any symptoms were felt. FHM duly criticised companies and public authorities that had asked their employees to isolate on their return from holidays abroad, or that permitted their employees to work from home (SR, 11 March; Expressen TV, 11 March; also Ehrenkrona, 2020). Days later, however, the agency declared that that employers should, after all, consider allowing work from home if possible, “which might have a certain dampening effect” on contagion (FHM, 16 March). Only at the start of April was social distancing advised (FHM, 1 April).
The same understanding of contagion also informed the agency’s recommendation that asymptomatic family members of people with confirmed infections need not stay away from school or work. The logic also implied that only limited measures need be taken in relation to staff at care homes for the elderly. Such measures might have included the requirement that staff wear masks. However, during most of 2020, FHM declined to recommend that anyone wear a mask (see below).
The slow development of a national system of testing for the virus, which subsequently attracted much media inquiry (SR, 2020; SVT, 2020) and the attention of the Corona Commission, was probably at least partly due to scepticism within FHM as to whether mass testing was an efficacious use of resources—a rare view among authorities in Europe (and, indeed, the World Health Organisation). That, in turn, may have been partly because the agency believed that the spread of the virus would soon slow due to the natural post-infection immunity that it would induce in the Swedish population.
Yet the most notable aspect of this initial period of the pandemic, in our view, was the very limited role that politicians played in decision-making (cf. Jerneck, 2021, p. 13–14). The prime minister, Stefan Löfven, declared in early March that the government was prepared to act on any recommendations and requests that the public agencies made (SR, SVT, 10 March). Days later, when asked why the government had not acted sooner to limit public gatherings, Löfven was clear: “It is the Public Health Agency that makes judgments about the spread of the virus” (SvD, 27 March). Consequently, FHM’s monopoly of problem definition and policy formulation was entrenched. The delegation of authority was thus publicly announced at the top political level.
The government did put some pressure on FHM to speed up testing during spring 2020 (SOU 2021:89, 2021, p. 231–300). In early April, it also secured parliamentary approval of a “crisis law,” which enhanced its powers to close certain places, such as shopping centres, in order to hinder the virus. Even then, the law prescribed—reportedly at FHM’s insistence (Kleja, 2020)—that the views of “expert authorities,” obviously including FHM, “be given decisive significance” in the determination of policy (Prop. 2019/20:155: 20). Anyway, the law lapsed at the end of June without being used.
Increasing Tension Between Experts and Government
By the summer of 2020, the power of Sweden’s public epidemiological experts at FHM was arguably at its peak. Levels of reported infection had fallen considerably. Tegnell was interviewed frequently in foreign media; he was usually depicted as the architect of a success storey. He remained confident that the post-infection immunity built up during the spring would protect Sweden from a renewed wave of covid-19; Denmark, which had locked down more harshly then, would now be more exposed, he suggested (FT, 11 September).
Yet the viral tide was turning again in Sweden. FHM played down a renewed bout of infections in September. Still, a recommendation that family members of infected people should self-isolate was adopted (FHM, 1 October). This reduced a stark policy divergence between Sweden and other European countries – although the Swedish period of isolation was shorter and children of all ages were still exempted.
As Sweden’s second wave of coronavirus infections gathered pace in early winter, there were again many deaths in homes for the elderly. Tegnell appeared to abandon the expectation that post-infection immunity after the first wave would protect the Swedish population (SvD, 13 November). As media pressure built, the government began to take some initiatives of its own. They did not always receive wholehearted public support from FHM.
In November, the government stipulated that bars and restaurants should close early (Regeringskansliet, 11 November). A few days later, it initiated a ban on organised public gatherings of more than eight people (Regeringskansliet, 16 November). When interviewed the same day, Tegnell could not explain the proposal, because, he said, it had come from the government, not his agency (Lönegård, 2020). A fortnight later, FHM extended to children its recommendation that family members of infected people should self-isolate (FHM, 1 December), which reduced further Sweden’s deviation from policy elsewhere. The education minister could not explain why the new line had not been adopted sooner; she had to deny that she had pressed the agency to make the change (Exp., 1 December [b]). FHM, however, sounded dubious about the epidemiological benefit.
Also in December, when it became clear that the agency’s earlier prognoses (Folkhälsomyndigheten, 2020) had been much too sanguine, the government hastily instigated a renewal of its “pandemic law” (Regeringskansliet, 4 January). In December, FHM changed its internationally deviant position on the use of masks, albeit only a little. We return to that point later.
The manifest tensions between politicians and experts outside of FHM then appeared to ease, even as Sweden’s second wave intensified. When restrictions on gatherings were tightened further, the measure was said to have been proposed by the agency (FHM, 18 December).
Why Policy Deviance?
In this section, we develop our explanation of why Sweden deviated from the European norm in its management of the pandemic, especially in its early phase. We see three important decision occasions. The first involved the fulfilment of the first necessary condition for the outcome—the initial estimation of the virus—although it is not one that we focus on in any detail. The second two decision occasions were those that led to the fulfilment of the condition of radical delegation to expert decision-makers (see Figure 2).
The first decision occasion involved the initial estimation of the virus that was made within FHM in December 2019 and January-February 2020. We know that there were particular views within the agency on the nature of contagion and on the limited value of non-medicinal interventions. It also became clear, from a radio lecture that Tegnell (2020) gave a few months into the pandemic and from many other public comments, that he was loth to support anything other than voluntary restrictions on people’s behaviour. This was partly because, in his view, voluntarism would be more sustainable in the longer term, and maybe even more effective, than compulsory controls.
Our analysis does not extend to seeking explanations for FHM’s initial understanding of the virus. That the agency took its position, however, was obviously necessary for Sweden’s being subsequently steered towards it in the formulation of national strategy. What our analysis does cover are two specific conditions, both of which concern how policy-making power was organised, which we now discuss in much more detail. These were, respectively, the establishment of FHM’s control over public policy; and the cementation of that control, as potential challenges to it were precluded.
The Experts Take Control
How did a bureaucratic agency assume such a central role in the management of such a huge national crisis? In a way, that it occurred in a Swedish administrative context makes it more understandable. A closer look at the contemporary political context also sheds light on the government’s passivity during the early part of the pandemic.
Swedish Public Policy-Making
Much of what the state does in Sweden is enacted by 341 public agencies. These agencies enjoy considerable autonomy. They are not part of the government ministries, which are small and organised into the Government Offices. Most state employees (263,000) work in the agencies, far more than those who work in the Government Offices (4,600) (Statskontoret, 2020, p. 11, 33; Regeringskansliet, 2021). The agencies are run by the political executive at arm’s length, through general policy instructions and the appointment (and, occasionally, the dismissal) of the agencies’ directors-general (Ahlbäck Öberg and Wockelberg, 2016; Jacobsson and Sundström, 2016; Pierre, 2020). The agencies’ autonomy is underlined by the ban on politicians interfering in the agencies’ handling of individual cases when exercising public authority. This is what is meant by the prohibition of “ministerial rule” (ministerstyre) (Instrument of Government Ch. 12, Art. 2).
Sweden is a unitary state that tends towards decentralisation to local and regional government (Montin, 2016). Regions and municipalities are in charge of healthcare and elderly care, which are regulated by central government through special legislation. As with the public agencies, national government can influence local and regional authorities through legislation and through funding specific activities. Moreover, outside this hierarchical relationship, some policy outcomes are the result of negotiations between the various actors at national and subnational level. The Association of Local Authorities and Regions plays an important role in such negotiations, although it has no constitutional status (Feltenius, 2016, p. 388). This association had a key role in co-ordinating agreements between the state and FHM in organising testing and vaccination programmes – a role that the Corona Commission felt raised questions about public accountability (SOU 2021:89, 2021, p. 25).
FHM was created in 2014 through the merger of the Institute for Communicable Disease Control and the National Institute of Public Health. In 2015 it was given overall co-ordinating responsibility for transmissible-disease control, which had previously been held by the National Board of Health and Welfare (Folkhälsomyndigheten, 2018). The new agency thus received a very wide remit to work for the promotion of public health, of which disease control is one among other important aspects, also including socio-psychological and socio-economic effects of policy measures, such as “lockdowns.” It also has the power to issue regulations and recommendations within its field of responsibility (Ordinance with instruction for the Public Health Agency, SFS 2013:1020, 2013).
Overall, then, Sweden’s decision-making structure is decentralised, even fragmented. With hindsight, co-ordination was always likely to be difficult in a crisis. Plans for the management of a particular type of crisis, involving the prevention and control of disease, were centred on FHM, the Civil Contingencies Agency and National Board of Health and Welfare. However, as the Corona Commission observed, “it was not evident in advance how crisis management would de facto be organised during the pandemic” (SOU 2021:89, 2021, p. 24). Instead, the “responsibility principle” (ansvarsprincipen) applied, which meant that the authorities normally responsible for a certain activity would maintain that responsibility during crises. Moreover, crisis co-ordination was to involve state agencies, municipalities and regions, and also private actors, including voluntary non-profit organisations (SFS 2006:827, 2006; FS 2015:1052, 2015; Krisinformation.se, 2021).
There was, then, a zone of uncertainty around leadership in the Swedish management of the pandemic. There was scope for certain actors to step forward and fill the gap—or to step back.
The Withdrawal of the Politicians
As described above, the political executive has considerable scope to steer public agencies. That it chose not to was, we argue, partly due to prevailing political circumstances.
Coronavirus arrived with the country’s party system in flux (Aylott and Bolin, 2019) and its government commensurately weak. After the election of 2018, government formation had proved extremely difficult. The centre-right party bloc collapsed, which allowed a minority coalition of Social Democrats and Greens to remain in office. Minority governments have been common in Sweden (Bergman, 1993; Lindahl et al., 2019). However, from 2018 to 2019, there were large policy distances between the government parties, and between them and their new parliamentary partners. A strong, stable government was never likely.
In one way, the government’s weakness was not directly exposed by the pandemic, because the policy response was—at least initially—hardly politicised. Perhaps in keeping with a tradition of political ceasefires (borgfred) during national challenges, opposition parties declined to voice anything more than mild criticism of the authorities. For example, even as she was prepared to dispute specific decisions, such as that not to isolate homecoming tourists, one centre-right party leader emphasised that she would “lock arms” with the government in the crisis (Ekots lördagsintervju, 25 April). Only the radical-right Sweden Democrats urged tougher restrictions and, in June, that Tegnell be sacked as chief epidemiologist (Åkesson, 2020). For their part, the Greens, the junior coalition partner, were reportedly disinclined to question what they saw as policy based on scientific evidence (Eriksson et al., 2020, p. 8–10).
In fact, Swedish politics was in flux at another level, too, which may have had more impact on policy. The Social Democrats’ electoral fortunes had declined over time. An election defeat in 2010 induced an acute internal crisis. The party leader resigned. Her replacement lasted just 10 chaotic months before he too was forced out. In this desperate moment, the Social Democrats’ usual elaborate method of leader selection was set aside. The party’s executive committee reportedly offered the job to former ministers who had left politics years before, before turning instead to one of its own members—Löfven, who was then leader of the Metalworkers Union and who had never held any elected political office (Aylott and Bolin, 2021). His foremost merit appeared to be his lack of previous association with any particular ideological or strategic orientation within the party.
Löfven became prime minister in 2014. His government promptly shifted responsibility for co-ordinating crisis management away from the Prime Minister’s Office to the Ministry of Justice and, in particular, the minister of the interior (Regeringskansliet, 2019; Folkhälsomyndigheten, 2021; p. 75). Not surprisingly, given his own credentials and the weakness of his governments, Löfven’s subsequent leadership was largely devoid of significant projects or initiatives. He became instead a broker of compromises between others, in government and in parliament.
It would be in accordance with the pattern during the rest of his time as prime minister, then, if Löfven’s government was inclined to delegate in dealing with the pandemic. It sought no expert advice other than that offered by FHM. That, after all, would have involved the responsibility of choosing between competing recommendations. In a remarkable interview in March 2020, the director-general of FHM, Johan Carlson, confirmed that his agency was “holding the baton, very clearly.” Yet this had not, in his view, been inevitable. “I can well imagine,” he stated, “that the government could take the initiative from its side, but that has not been the case.” Carlson was asked if his agency had been explicitly mandated by the government to steer policy. “Our understanding,” he replied, “is that this is the procedure. That said, we have not tried any other procedure” (Örstadius et al., 2020).
Assertive Bureaucratic Leadership
The reluctance of politicians to take command left a space (cf. SOU 2022:10, 2022). However, some other actor had to fill that space—which, as the quotation above suggests, FHM was prepared to do. Some of that willingness must have been due to personalities. It is not difficult to imagine that Tegnell, in the words of a previous chief epidemiologist, had “an ability to get his way and…good connexions to the Government Offices” (Kleja, 2020).
Here we note two particular examples of assertive bureaucratic leadership. They indicate how experts did not just advise political decision-makers, or even take delegated decisions. They were also prepared, albeit implicitly, to criticise policy decisions with which they did not agree; and argue actively and publicly for the decisions that they had taken.
The Masks Controversy
In some ways, the most notable divergence between politicians and experts concerned the use of masks. As we saw, FHM was sceptical about masks. It argued that there was little evidence of their efficacy and that they might actually increase the risk of infection, due to the need to apply and adjust them and because they might dissuade people from maintaining social distance (FHM, 6 May). By the summer, the Swedish media began to notice how isolated the agency’s view had become internationally (Eriksson, 2020; SVT, 22 July). In September, FHM insisted that, “if and when transmission increases, face masks will be considered—for example, in settings where physical distancing cannot be maintained” (Carlson and Tegnell, 2020, p. 843). In late December, at the height of the second wave, the agency did relent—up to a point. Masks were recommended, but only on public transport, only after the new year and only during two daily rush-hour periods (FHM, 18 December). Adherence thus required of Swedes a high level of awareness, not only of the recommendation but also of the time of day. FHM’s director-general was left “embarrassed” by his own failure to maintain such awareness (SVT, 28 January 2021).
On occasion, the experts’ pronouncements appeared more directly to undermine politicians’ decisions. In autumn 2020, regional governments had been authorised to take their own measures against the virus. In late February 2021, Stockholm was one of several that recommended the use of masks on all public transport—partly, it explained, because compliance with national advice had been poor (Region Stockholm, 23 February 2021). FHM’s support for the measure was lukewarm, at best. Compliance with the regional government’s recommendation by users of public transport in the capital remained limited.
The Comparability Controversy
Tegnell, in his countless media appearances, was always ready to explain robustly why Sweden’s own path had been the right one—and thus why other countries had taken the wrong path (for instance, BBC Radio 23 April). Perhaps surprisingly, his arguments invariably referred not to specific Swedish circumstances, such as its relatively low population density and small households, or its lack of constitutional provision for emergency governance, but rather to general principles. He insisted that his agency had pursued a “classic” strategy of managing a pandemic, as he put it in his radio lecture. The rest of the world, by contrast, seemed to have “gone mad” (Tegnell, 2020).
Yet the issue of comparability, which we mentioned in the introduction to this article, became increasingly contentious as the pandemic developed. For some Swedish (and many foreign) observers, comparison of Sweden with the other Nordic countries, which shared some common geographical, cultural, socio-economic and demographic characteristics, was natural. To them, the correlation between Sweden’s strategy and its relatively high death toll implied an obvious causal relationship.
Tegnell, however, would have none of it. Sweden, he argued, had been hit harder because it was in a different phase of the pandemic than its neighbours (Svenska YLE, 4 April); or because of particular problems in care homes (Exp., 7 April); or because of the virus’s relatively early arrival in Sweden, which made its circumstances more like those of Britain or France (SvD, 13 September); or because Sweden had experienced a relatively light influenza season the previous year, which had left it with many vulnerable old folk (DN, 15 September); or because of Sweden’s high proportion of foreign-born residents (AB, 4 December); or simply because, on the outcome variable, Finland and Norway were the European outliers (SVT, 18 September; Exp., 1 December [b]). In other words, after initially emphasising the difference in Sweden’s strategy compared to those elsewhere in Northern Europe, he later preferred to de-emphasise it, through arguing that policy was, at most, just one factor among several causes of Sweden’s higher death toll.
We take no position on the merits of these competing claims about masks and national comparability per se. Our point relates to the lengths to which Sweden’s chief epidemiologist went to assert that the policies promulgated and implemented by FHM had been correct. If it is unusual that a bureaucratic organisation should take such a prominent role in formulating public policy, it is arguably more unusual that the same organisation should then campaign so energetically to justify that policy and to shape citizens’ interpretation of its results. The agency strove to maintain a monopoly of problem definition and policy formulation, even in the face of critique by external experts (Carlsson et al., 2020; Steineck et al., 2021; also SVT, 19 March; DN 30 November 2021; AB, 24 July) and, more rarely, politicians (SVT, 7 June). In the words of the final report of the Swedish Corona Commission, published in February 2022:
The Government should have assumed leadership of all aspects of crisis management from the outset. It should have been able to overcome the obstacles to clear national leadership that currently exist: government agencies with a degree of autonomy […] The Government had too one-sided a dependence on assessments made by the Public Health Agency of Sweden […] This is not a satisfactory arrangement for decision-making during a serious crisis in society (SOU 2022:10, 2022, Summary in English, p. 4).
Trust in Public Institutions and the Cementation of Strategy
The Swedish tradition of public administration, an unusually passive political leadership and a public agency that was fully prepared to assume control of policy: these conditions combined to establish FHM’s initial analysis of coronavirus as the basis of the Swedish strategy. However, no formal transfer of powers ever took place. There was, then, every chance of a misjudged initial policy then being revised—as in Britain. We suggest, however, that Swedes’ deep-seated faith in their public institutions insulated policy-makers from being influenced in a similar way.
It seems likely that a national approach to the pandemic might be connected to societal attitudes to risk more generally. Cornia et al. (2016) concluded that Sweden, among seven surveyed countries, most resembled a “state-orientated” risk culture, characterised by strong confidence in public authorities’ ability to deal with risk and disaster (p. 294). In such circumstances, it is argued, individuals are expected to take less responsibility than they are in other types of risk culture; the state thus assumes an almost paternalistic relation with the citizen. This also means that the public association of disaster with political accountability tends to be strong. Yet this finding is hard to square with the Swedish experience in the first months of the pandemic, in which the emphasis of public policy was on individual responsibility, while political accountability was unclear.
Trust in public institutions nevertheless explains something of how the Swedish strategy became consolidated. Swedes, like other Nordic citizens, exhibit high levels of interpersonal trust and trust in public institutions (Holmberg and Rothstein, 2020, p. 9–10; Andersson, 2021a, p. 2; OECD, 2021; DN, 30 October). During the early phase of the pandemic, a regular survey of public opinion observed record levels of such trust. This applied to parliament, to government and especially to FHM. In 2020, fully 81 per cent expressed high trust in the agency. The figure dipped in 2021, but was still 65 per cent (Andersson, 2021b, p. 3). Trust in the media, meanwhile, remained at similarly high levels, although, again, public-service television and radio had fallen moderately by 2021 (Andersson, 2021a, p. 2–5).
Not surprisingly, then, Swedes took most of their information about the pandemic from public institutions. FHM was the primary source: nearly nine out of 10 survey respondents reported that, to a fairly or very great extent, they were informed by the agency (see Table 1).
Predictably, we find a strong connexion between reliance on FHM for information and holding a positive view of its competence (the Pearson correlation is−0.406). The correlation is notably weaker, and goes in the opposite direction, among the smaller numbers of people who were reliant to a similar degree on information from foreign media (0.182) or social media (0.063). Such people, we can assume, were more exposed to views on the pandemic that deviated from those propounded by FHM. Similar findings about exposure to foreign media are reported by Johansson et al. (2021). They found that Swedish respondents with higher levels of trust in government were likelier to be careful with hand hygiene, which FHM had emphasised in its information, but less likely to wear a mask—a measure that, as we saw, the agency had set itself against from the start, but which, by autumn 2020, had become widely required in other countries. Conversely, the likelihood of wearing a mask was higher among those accessing foreign media.
This trust in public institutions manifested itself in a “rallying-round-the-flag” phenomenon (Hetherington and Nelson, 2003; Murray, 2017). Indeed, some in the Swedish media were initially hostile to critics, domestic and foreign, of the Swedish strategy and its scientific basis. Interestingly, this defensive reaction was strongest on the left of the political spectrum.
Of course, the authorities’ management of the pandemic was intensely debated in the Swedish media. Commentators were well aware that Sweden’s strategy was different to those chosen by its neighbours. Yet by contrast with countries like America, Britain and Germany, it was the Swedish right that was generally in favour of more behavioural restrictions in order to stem the virus’s spread. Criticism of FHM came mostly from right-of-centre columnists (such as Arpi, 2020). On the left, meanwhile, the authorities’ line attracted enthusiastic support. In Aftonbladet, the main left-leaning newspaper, FHM, and Tegnell in particular, became objects of occasionally rapturous praise (Lindberg, 2020b).
Trust in public institutions helps to explain this political alignment. A revealing indicator is provided by the question of who, in any particular setting, is regarded as an “expert.” For many observers in Sweden in early 2020, that accolade was given especially to those who held positions in the state, including agencies like FHM. Critics of the authorities were sometimes depicted as non-experts, even anti-experts or “populists”—eccentric, unqualified, unscientific, even disloyal (Lindberg, 2020a; Pettersson, 2020; von Schreeb, 2020; further examples are cited in Gustavsson, 2020, 2021; also Vogel, 2020). The dismissal of such dissidents as “amateur epidemiologists” (hobbyepidemiologer) or querulants was despite the fact that many of them had scientific qualifications and voiced opinions that were mainstream in other countries (Ahlström, 2020; Lagercrantz, 2020; Bjurwald, 2021, p. 69–72; Steineck et al., 2021; SVT, 19 March). Away from newspapers’ leader pages, meanwhile, Swedish media reporting of the pandemic often had a rather uncritical character (see Andersson et al., 2021; Bjurwald, 2021). It was easy to infer that many journalists were reluctant to criticise a trusted national institution.
We argue, then, that trust in public institutions and their experts, to the extent of generating a certain hostility to dissenting voices in relation to pandemic strategy, was sufficient to consolidate that strategy and, during the first half of 2020, to insulate Swedish policy-makers from pressure to change course. Because the policy had strong public support, there was little incentive for the media or the political opposition to attack the government or FHM. Only for a brief spell in late spring 2020, and then again with the arrival of the second coronavirus wave in late autumn, did media coverage turn more critical.
In the preceding analysis, we have established that Sweden’s public response to the coronavirus pandemic in 2020 can certainly be classified as a case—arguably an extreme case—of politicians’ delegating public authority to expert bureaucrats. It involved an informal but highly significant transfer of power—a peculiar type of science-policy interface, one in which political authority is explicitly delegated to, and even usurped by, an expert agency (cf. SOU 2022:10, 2022). This, we argue, was necessary for the outcome of Sweden’s policy deviation from the European norm.
For some, this was normatively appropriate. Jonung (2020, p. 5), for example, argues that putting FHM in charge of “framing and designing the policy response to the pandemic” had the advantage of insulating these decisions from political interference—much as monetary policy is the preserve of the central bank, not the government. This, he argues, allowed the agency’s epidemiologists and health economists to take into account “the overall impact of the corona policies on the health of the Swedish people and on the Swedish economy.”
One normative counter-argument might be that the very breadth of a crisis, which necessitates the weighing of values and goals against each other, illustrates precisely the point of having public leadership that is politically accountable for its decisions across the full range of public policy. A second counter-argument might be that extensive delegation, in combination with almost total reliance on information from one particular agent, may lead to an enfeebled principal—in practise, abdication (Kiewiet and McCubbins, 1991, p. 31; McCubbins and Schwartz, 1984, p. 166; also Ahlenius, 2020). The question of political accountability for Sweden’s management of the pandemic is a pertinent one, which is likely to be subject to extended discussion in the country.
While much of the literature on the politics of expertise is about how politicians use expertise (Boswell, 2009), the case at hand shows rather how experts can become political decision-makers themselves. Notably, we found that when science-technocrats moved from the traditional advisory role to become policy-makers, they made use of science in a way that was similar to how elected politicians do it—that is, by mixing science and value-based judgement; by emphasising the need for scientific certainty when arguing against precautionary action; and by rejecting external expert critique as unfounded, thus maintaining a monopoly of problem definition and policy formulation.
The delegation of political authority to science-technocrats is a type of science-policy interface in need of further research. We should not expect that delegation of authority is the same everywhere. Governments might not always be as willing to delegate authority to the extent that it was in the Swedish case. There will probably be few other examples in developed democracies of a single expert agency having such a wide mandate as FHM enjoyed, even before the pandemic.
As for how this degree of delegation came about, we argue that there were three important junctures. The first was FHM’s initial analysis of the novel coronavirus—which it subsequently revised, bringing Sweden more into line with neighbouring countries’ policies, but only slowly and discretely. As we stated, this expert positioning was obviously necessary for Sweden’s adopting the strategy that it did, although it is not in itself something that we described or explained in any depth in our study.
The second and third decision occasions, on which we focused, established and consolidated, respectively, FHM’s control over policy, which, we argue, was far from inevitable, a conclusion made also in the final report of the Corona Commission (cf. SOU 2022:10, 2022). Together, they led to the fulfilment of our second condition for the outcome of Swedish exceptionalism. For sure, Sweden’s tradition of autonomous public agencies, accentuated by the wide mandate that the agency had been afforded, facilitated such radical delegation. Still, it is likely that a Public Health Agency with a less self-confident and assertive leadership, and a political leadership that was keener on taking responsibility in a crisis, would have been sufficient to forestall this radical delegation of policy-making, and thus led to Sweden’s policy being less deviant. The specific case of Sweden’s delegation of policy-making authority during the pandemic is another promising area of future research. Indeed, it is hard to imagine something similar occurring in countries with different administrative traditions, although future comparative research will have to address the issue. It would be particularly interesting to compare Sweden’s handling with that of other deviant liberal democracies, for example New Zealand, which attempted a “zero” Covid-19 strategy and maintained a strong government lead.
Data Availability Statement
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.
All authors listed have made a substantial, direct, and intellectual contribution to the work and approved it for publication.
Conflict of Interest
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fpos.2022.832518/full#supplementary-material
Ahlbäck Öberg, S., and Wockelberg, H. (2016). “The public sector and the courts,” in The Oxford Handbook of Swedish Politics, ed J. Pierre (Oxford: Oxford University Press). doi: 10.1093/oxfordhb/9780199665679.013.8
Aylott, N., and Bolin, N., (eds.). (2021). “The rule of the valberedning? Party leader selection in sweden,” in Managing Leader Selection in European Political Parties (London: Palgrave). doi: 10.1007/978-3-030-55000-4
Capoccia, G., and Kelemen, R. D. (2007). The study of critical junctures: theory, narrative, and counterfactuals in historical institutionalism. World Polit. 59, 341–369. doi: 10.1017/S0043887100020852
Carlsson, M., Einhorn, L., Einhorn, S., Elgh, F., Frisén, J., and Gustafsson, Å. (2020), “Folkhälsomyndigheten har misslyckats – nu måste politikerna gripa in”, Dagens Nyheter. Available online at: https://www.dn.se/debatt/folkhalsomyndigheten-har-misslyckats-nu-maste-politikerna-gripa-in/ (accessed April 14, 2020).
Ellingsen, T., and Roine, J. (2020). Public Policy and the Swedish Model: What was the Swedish Government’s Approach to the Pandemic, and How was it Received in Sweden? Stockholm: Stockholm School of Economics, Institute for Research.
Feltenius, D. (2016). “Subnational government in a multilevel perspective,” in The Oxford Handbook of Swedish Politics, ed J. Pierre (Oxford: Oxford University Press). doi: 10.1093/oxfordhb/9780199665679.013.23
Hale, T., Angrist, N., Goldszmidt, R., Kira, B., Petherick, A., Phillips, T., et al. (2021). A global panel database of pandemic policies (Oxford COVID-19 Government Response Tracker). Nat. Hum. Behav. 5, 529–538. doi: 10.1038/s41562-021-01079-8
Jasanoff, S., Hilgartner, S., Hurlbut, J. B., Özgöde, O., and Rayzberg, M. (2021). Comparative Covid Response. Interim Report. Harvard, IL: Harvard Kennedy School. Available online at: https://ingsa.org/covidtag/covid-19-commentary/jasanoff-schmidt/
Johansson, B., Sohlberg, J., Esaiasson, P., and Ghersetti, M. (2021). Why swedes don’t wear face masks during the pandemic–a consequence of blindly trusting the government. J. Int. Crisis Risk Commun. Res. 4, 335–358. doi: 10.30658/jicrcr.4.2.6
Lindahl, J., Hellström, J., and Bäck, H. (2019). “Sweden: Minority government as the norm,” in Coalition Government as a Reflection of a Nation’s Politics and Society, ed M. Evans (London: Routledge). doi: 10.4324/9780429422379-8
Montin, S. (2016). “Municipalities, regions, and county councils,” in The Oxford Handbook of Swedish Politics, ed J. Pierre (Oxford: Oxford University Press). doi: 10.1093/oxfordhb/9780199665679.013.22
SR (Sveriges Radio) (2020). “Coronapandemin”, P1 Dokumentär: Miniserie, episodes released 27 October (“Den dolda smittan”), 3 November (“Den missade testningen”), 10 November (“Testa, testa, testa”). Stockholm: Swedish Radio. Available online at: https://sverigesradio.se/gruppsida.aspx?programid=5319&grupp=34379
Strøm, K. (2006). “Parliamentary democracy and delegation,” in Delegation and Accountability in Parliamentary Democracies, eds K. Strøm, W. C. Müller, and T. Bergman (Oxford: Oxford University Press).