Between lockdown and calm down. Comparing the COVID-19 responses of Norway and Sweden

ABSTRACT

This article studies the difference in the government response to COVID-19 in Norway and Sweden drawing upon theories of agenda setting, crisis management and multi-level governance. Despite having virtually identical systems of government and, initially, facing similar infection threats, Sweden opted for far less strict countermeasures than Norway. While Norway’s government response was similar to that of many European countries, Sweden received international attention for choosing befuddlingly soft measures, mostly recommendations and guidelines. This divergence is discussed vis-à-vis the multiple streams lens of agenda setting theory, highlighting differences in institutional and organisational legacies (e.g., the autonomy of government agencies) and intergovernmental relations (e.g., whether infectious disease expertise and authority is located at the local level).

KEYWORDS:

• COVID-19
• sweden
• norway
• autonomy
• expertise
• multiple streams
• intergovernmental relations
• local governments

Introduction

The question of why governments responded differently to COVID-19 has occupied social scientists since the spring of 2020. Surprisingly, COVID-19 caused extensive damage in countries considered very well-prepared to handle a health crisis, like the United States (Cameron et al. 2019). Toshkov, Yesilkagit, and Carroll (2020) found general governance capacity to be negatively associated with the speed of countries’ policy response to COVID-19. They speculated that leaders in high-capability contexts were overly confident that a pre-programmed government response would contain the virus. Generally, ‘mysterious’ heterogeneity in the COVID-19 response across similar countries tested the imagination of scholars (Cheibub, Hong, and Przeworski 2020, 12).

This paper describes and aims to explain why COVID-19 was met by very different government responses in the two ‘fraternal twin’ countries of Norway and Sweden. Beyond banning large assemblies of people, Sweden avoided strict regulation. The authorities provided guidelines – nudges – on how to prevent the spread of the infection, they isolated infected individuals and traced their contacts. Sweden’s response was thus ‘a remarkable exception’ in Europe (Bouckaert, Galli, and Kuhlmann 2020, 769), whereas Norway’s was similar to that of other European countries with closed schools and businesses, strict travel restrictions, etc. (Christensen and Lægreid 2020).

That such different responses to the crisis could occur in neighbouring countries with nearly identical state–society relations and politico–administrative systems is paradoxical (Plümper and Neumayer 2020). The purpose of the article is thus to answer the question: What explains the difference in the response to COVID-19 between Sweden and Norway? We focus primarily on the crisis experience during the spring of 2020, with a special emphasis on agenda setting, the role of autonomous government agencies, and how the government response was co-shaped and coordinated between national and subnational governments. On selected issues, the empirical study covers events through the first quarter of 2021.

We begin by presenting the analytical framework. Then, we give a background by presenting Sweden’s and Norway’s systems of governance and a timeline of important events in the two countries. Next, we describe government responses to COVID-19, focusing on the balance between liberalism and paternalism, politics and administration at the central level, and centralisation/decentralisation between national and subnational levels of government. Finally, we explain the differences between the countries and relate the findings to the theories in our framework.

Analytical framework

According to Kingdon (1984), three streams intersect in a policy window when an issue rises to the fore and public policies are enacted: a ‘problem stream’ of issues warranting public action, a ‘policy stream’ of actors offering policy options and a ‘political stream’ of factors, such as changes in national mood, advocacy campaigns or a crisis. Policymakers are bombarded with information originating from different sources and they must, according to Jones and Baumgartner, ‘focus on some of this information and ignore most of it. This selective attention process has critical consequences for policymaking, and especially how the political system prioritizes problems for policy action’ (Jones and Baumgartner 2012, 7). In some situations, policymakers are part of a ‘policy community’ with ‘shared ideas about the policy at stake’ (Walgrave and Varone 2008, 367). In other situations, actors in and around the political–administrative establishment engage in ‘framing contests’ (‘T Hart and Tindall 2009, 23), trying to influence the way a policy problem is looked upon and, ultimately, to shape public opinion (Mintrom and Norman 2009, 652).

The way we think of a problem can be decisive for policy making (Bacchi and Goodwin 2016). The importance of the conceptualisation of the problem was addressed in the early implementation literature: ‘what is a problem is, after all, both subjective and socially defined’ (Wolman 1981, 437). We, thus, need to probe the assumptions that underpin how problems are represented.

When a crisis occurs in the political stream, a ‘random’ policy window opens, one characterised by urgency and uncertainty (Boin et al. 2005, 3). Often, governments must respond to a crisis before its magnitude, causes and effects are well understood (Wetzelburger et al. 2019, 100). To avoid uncertainty paralysing the response, governments engage in crisis planning. When planning and when putting the plan into action, policymakers make important decisions about where to strike the balance between conflicting demands. Here, we focus on three such balancing acts or dimensions. The first dimension, liberalism–paternalism, concerns how much responsibility citizens must shoulder themselves for managing a crisis and how much responsibility the state takes.

Secondly, the politics–administration dimension concerns where to place executive responsibility when a crisis occurs. In a serious health crisis like COVID-19, the expectation would be that decision authority is concentrated (Boin and Hart 2012); for example, that authority is shifted from an executive agency to a government ministry under tighter political control (Kirlin 2020). A decentred regime with authority located in an agency at arm’s length from its political masters might be conducive to evidence-based policymaking. However, due to uncertainty, a crisis is not a situation in which politicians can rely on one set of experts to dominate the policy stream with well-established ‘ideas about the policy at stake’ (Walgrave and Varone 2008, 367). Moreover, experts cannot ‘answer the question: which risk is acceptable and which is not’; they ‘can only supply more or less uncertain factual information about probabilities’ (Beck 2003, 59). Due to urgency, politicians must emphasise precaution and meet a crisis with countermeasures, even if the evidence about a hazard and cause–effect relationships is uncertain (Bourguignon 2015).

Thirdly, the centralised–decentralised dimension concerns striking a balance between meeting a crisis with a coherent, nationwide response and a flexible, localised response (Kettl 2003). This relates to multi-level governance – a lens through which to describe and analyse political systems with several and interdependent tiers of authority. As seen through this lens, policymaking is multi-centric and policies are a result of coordination between tiers (Cairney 2020; Hooghe and Marks 2001). When a crisis hits a system in which authority overlaps between levels of government, the general expectation would be a shift towards a nationwide response. In a unitary state, this would be the national government in a dominant position vis-à-vis subnational levels of government (Chisholm 1989, 42). The expectation would be, therefore, that the national government centralises authority to respond to the crisis and intensifies vertical coordination of residual decentralised crisis management responsibility. To overcome resistance to the suspension of local autonomy, the government can play on ‘rally around the flag’ rhetoric and symbols (Chowanietz 2011).

Research design

The research design of this study is a comparative case study using a most-similar-systems design. We compare the context and features of two instances of a specific phenomenon: government responses to COVID-19. The comparison aims to discover contrasts, similarities or patterns across the cases. The case study method allows for interpretations rather than statistical generalisations. The advantages of the method are counterpoised by problems like deference to case idiosyncrasy (Zartman 2005). Even so, the case study is an appropriate method for many important research tasks in the social sciences (Flyvbjerg 2006). The many and strong similarities between Sweden and Norway mean we ‘hold constant’ many relevant factors and make the comparison a viable strategy for identifying differences in the political–administrative structure and culture that are likely explanations for cross-country differences in the government responses to COVID-19.

Empirical material includes government documents, parliamentary proceedings, legal acts and secondary sources like official statistics, media reports and scientific research. Uniform data is published by government-appointed expert evaluation committees in both countries (Kvinnsland et al. 2021, Swedish Government 2020) but much information necessary for a thorough investigation remains to be produced.

Research context

This section describes similarities and differences between the two countries on factors that existing research suggests can explain a country’s response to the COVID-19 crisis (Frey, Chen, and Presidente 2020; Cheibub, Hong, and Przeworski 2020; Helsingen, Refsum, and Gjøstein et al. 2020; Plümper and Neumayer 2020). Sweden has a larger population than Norway (10.3 million compared to 5.4 million), but the demographic profiles of the populations are similar. Both countries are sparsely populated and relatively urbanised; about 87% of Swedes and 82% of Norwegians live in urban areas (Ritchie 2018). Immigrants make up about 18% of Sweden’s population and 13% of Norway’s (Eriksen 2019). Both countries have a high level of generalised trust and social capital (Rothstein and Stolle 2003), and thus a good basis for voluntary commitment to government guidelines (Toshkov, Yesilkagit, and Carroll 2020).

Moreover, both countries had minority coalition governments in 2020; Sweden had a centre–left government and Norway had a centre–right coalition government. The political culture is consensual and rationalistic; it typically takes time to reach a decision, but then implementation runs smoothly (Dahlström 2015). Both countries have three levels of government. Sweden has 21 regions and 290 local governments, while Norway has 10 regions and 356 local governments. Local governments have relatively high autonomy from the national government and are responsible for a wide array of services, including primary education, primary health services and elderly care (Ladner, Keuffer, and Baldersheim 2016). Local governments predominantly operate elderly care services themselves; only about 20% (Sweden) and 10% (Norway) of these services are contracted out to voluntary and for-profit organisations in the private sector (Vårdföretagarna, 1 September 2020; Statistics Norway 2020). Both countries have a strong welfare state with universal access to high-quality healthcare (Esping-Andersen 1990). In Sweden, hospitals are a regional government responsibility; in Norway, specialist healthcare is provided by state health authorities. The 2019 Global Health Security Index ranked Sweden seventh and Norway sixteenth of 195 countries based on their level of preparedness for handling an infectious disease outbreak (Cameron et al. 2019).

National executive government agencies play a major role in policymaking and policy implementation in both countries. Sweden has about 340 and Norway about 150 such agencies. The position of agencies is different in the two countries, though. By law and custom, Swedish agencies have comparatively higher autonomy from ministerial interference. The government influences its agencies through legislation budgets, not through ad hoc instructions in singular issues, and the government will rarely change legislation without an initiative or support from the agency that administers the legislation. In Norway, the principle of ministerial governance is strong, meaning that government ministers by law and culture have political responsibility for their whole-policy portfolio. Agencies are more integrated into the ministries, and government ministers can, generally, instruct agencies what to do (Askim, Bjurstrøm, and Kjærvik 2019).

COVID-19 responses

We start this section by presenting a timeline of events (Table 1) . As detailed accounts are readily available from exiting research, reports and data sets, we keep this overture brief.

Table 1. COVID-19 response timeline 2020–2021

Sweden	Norway
	31 Jan: Municipalities and the Directorate of Health temporarily authorised to make binding decisions using the Infection Control Act
	6 March: Visits to elderly care institutions (b). First case of a municipality deciding to close schools and kindergartens (b).
9 March: Tests for symptomatic travellers from risk areas (g)	9 March: Directorate of Health temporarily authorised to make binding decisions using the Health Preparedness Act
12–14 March: 500+ assemblies (b); restriction on international travel (g)	11–15 March: Lockdown: 500+ assemblies, kindergartens and educational institutions, restaurants, swimming pools, sports and cultural events, quarantine for travellers from non-Nordic countries, self-isolation for infected travellers, etc. (b) Many municipalities decide on supplementary, stricter measures
	16 March: Border closure and control (b).
17 March: Distance teaching at secondary and higher education (g); travel restrictions outside EU (b)
	19 March: No visits to secondary homes (b)
27–30 March: 50+ assemblies, visits to elderly care institutions (b)	25 March: 5+ assemblies (g).
27 March: COVID-19 Act temporarily expands government powers. The COVID-19 bylaw enacted.
7 April: Overall government strategy published
16 April: Communicable Diseases Act changed to expand government powers
	24 April: Municipalities instructed to build capacity to test 5% of inhabitants per week (b)
	27 May: COVID-19 Act expires
October 21: Stricter rules first introduced in one region (g)
November 4: Limitation on the number of visitors in bars and restaurants (b)
	December 2020: State coordination of local government responses at regional level (b)
8 January 2021: New pandemic law
By the end of March 2021, both countries had administered
about 18 vaccine doses per 100 residents (Roser et al. 2021).

Note: g = guidelines and recommended actions; b = bans, closures and mandatory actions

Sources: Christensen et al. (2021), Kvinnsland et al. (2021), Ludvigsson (2020), Roser et al. (2021).

The first cases of COVID-19 infection were registered 31 January in Sweden and four weeks later, 26 February in Norway. Both countries registered the first cases of COVID-19-associated deaths in mid-March. Updated and detailed information about the number of people infected and hospitalised with COVID-19 and the number of COVID-19-associated deaths in the two countries is available from the national health authorities and from the European Centre for Disease Prevention and Control (2020). Table 2 shows selected key figures.

Table 2. COVID-19 Infections, deaths and hospitalisations

Data per 100,000	Time (end of)	Sweden	Norway
Infections, accumulated	Q2 2020	669	162
	Q4 2020	4,476	918
	Q1 2021	8,003	1,782
Deaths, accumulated	Q2_2020	54	5
	Q4 2020	93	8
	Q1 2021	134	13
Hospitalisations, at-the-time	Q2 2020	9	0.4
	Q4_2020	28	3
	Q1 2021	18	5

Source: European Centre for Disease Prevention and Control (2020)

Next, we describe the government response to COVID-19 in the two countries using the basic conceptual triad introduced above. For overview, key comparative observations are summarised in Table 3.

Table 3. Summary of key comparative observations

Conceptual triad	Sweden	Norway
Liberalism-paternalism	Legislation and culture strongly emphasise individual liberties and the responsibility of individuals to protect themselves from communicable diseases	Comparatively stronger emphasis on the government’s responsibility to take measures to protect society from communicable diseases
Politics-administration:	Decision authority in the central government largely decentred to the bureaucracy. The government took a more active and visible role from end of 2020.	The government took active lead mid-March. Considerable decision authority decentred to the bureaucracy throughout but under heavier shadow of hierarchy than in Sweden.
Centralised-decentralised	No communicable disease decision authority delegated to municipalities, some delegated to regions. Predominantly coherent, nationwide government response throughout.	Considerable decision authority delegated to municipalities, who used this authority actively. Overall pattern of varied, localised response interrupted by shorter periods of nationwide coherent government response.

Liberalism vs. Paternalism

The liberalism–paternalism dimension concerns how much responsibility citizens must shoulder themselves for managing a crisis and how much responsibility the state takes. On infection control, both countries have a history of successful voluntary-based infection preventive measures, such as vaccinations. Both countries have communicable diseases legislation emphasising that ‘volunteer preventive measures should be applied first, whenever possible’ (Helsingen, Refsum, and Gjøstein et al. 2020, 7). However, Norway’s Infection Control Act authorises the government to make binding decisions and implement quarantine and travel bans and other tough measures in the case of a health crisis. By contrast, Sweden’s Communicable Diseases Act places particularly high emphasis on the responsibility of each individual, reflecting a deep-rooted mindset in Swedish infection control that says the individual should be allowed to choose how to protect themselves on the basis of their own ethical considerations (DN Debatt, 9 June 2020). Sweden’s Communicable Diseases Act makes only very limited provisions for lockdowns. It is generally assumed that the Swedish Constitution places constraints on the government’s use of ‘illiberal’ measures like banning visiting elderly institutions and restricting travel. We mention, though, that some commentators, for example, political science professors Dahlström and Lindvall, have argued that the constitution does not block the use of such legislation in a situation like the COVID-19 crisis (see Christensen et al. 2021). Some have suggested that Sweden’s political culture has made its political leaders assume that the constitution restricts their choices more than it actually does (Christensen et al. 2021).

In general, it seems that Norway was more prepared for a crisis, or at least, for a crisis to be handled by the authorities. Sweden has plans for war situations, but was surprisingly ill prepared for a pandemic. As an illustration, the brochure ‘If Crisis or War Comes’ (Swedish Civil Contingencies Agency 2018) – sent to all households in Sweden – even does not mention the possibility of a pandemic. Relatedly, Norway’s response was relatively more influenced by gerontology-related science; Sweden’s response was more influenced by a public health perspective.

Decision authority in the central government: Politics vs. administration

The politics-administration dimension concerns where in the central government executive responsibility is located in a crisis. In a pandemic crisis, important executive authorities at the national level in Sweden are the Ministry of Health, the Public Health Agency (issues regulations and takes action), the National Board of Health and Welfare (crisis preparedness in the health system and in the social services) and the Civil Contingencies Agency (emergency management). In Norway, the equivalent authorities are the Ministry of Health and Care Services (national responsibility for health preparedness; declares a health crisis, thus activating roles of other authorities), the Directorate of Health (coordinates mitigation efforts and implements infection control measures), and the Institute of Public Health (monitors the epidemic situation and supervises and advises state and local authorities on infection control).

Based on the analytical framework, the expectation was that decision authority would be concentrated during COVID-19. In both countries, infection control legislation and -planning place heavy responsibility on executive agencies for handling a crisis. Both countries also follow the responsibility principle, which says that the authorities that have responsibility for a function in normal circumstances, have responsibility for the function also in a crisis situation (Kvinnsland et al. 2021, 56). This division of responsibility was maintained in the spring of 2020 in the handling of COVID-19, with the Public Health Agency as the centre of action in Sweden and the Directorate of Health in Norway.

Norway’s Directorate of Health was delegated responsibility for coordinating the health-sector response starting on 31 January 2020 and has kept this role into 2021. When COVID-19 turned into a full-blown crisis in March, the Directorate temporarily had extraordinary extensive decision-making responsibilities due to emergency procedures in the Infection Control Act. Out of convenience, the 12 March lockdown decisions were formally enacted by the Directorate, even though the Prime Minister, other government ministers and the Ministry of Health were actively involved in the decision-making process. A government-appointed evaluation committee later argued that the government’s allowing such important decisions to be made by anyone but the King in Council constituted a breach of the Constitution (Kvinnsland et al. 2021). In late March, the government enacted a COVID-19 bylaw that placed more responsibility for day-to-day use of the Infection Control Act on the Ministry of Health. The Directorate maintained its coordination role, continued to be the government’s closest advisory body throughout the crisis, and implemented government policies. Decision authority and executive responsibility in Norway was thus not centralised, for example, by the government reassigning authority to government ministries.

The similarity in executive responsibility of the central government agency in the two countries should not distract us from an important difference. As described above, agencies in the two countries have different autonomy from political influence, and the COVID-19 experience did not deviate from this pattern. In the spring of 2020, Sweden’s COVID-19 policy agenda was dominated by the Agency for Public Health, particularly State Epidemiologist Dr. Anders Tegnell. The director of the Public Health Agency was relatively invisible in the public discourse, as were government ministers; Tegnell presented and explained the government response to COVID-19. According to political scientist Jon Pierre, ‘Cabinet ministers [took] a backseat while urging citizens to follow the experts’ advice’ (Pierre 2020, 483). Consequently, Swedish public discourse about how to contain the virus – the policy image (Baumgartner and Jones 1991) – was scientific in content and tone. Whether the relative inactivity of the Swedish government was a political choice or a consequence of legislation is a matter of debate. Two leading experts, Johan Hirschfeldt and Olof Petersson (2021), claim the executive agencies have less autonomy from political control than Swedes generally think. At the end of 2020, politicians also seemed to be more active in public appearances.

Norway’s COVID-19 policy agenda was more dominated by the Minister of Health, in tandem with the Directorate of Health and with local governments in a key cameo role (Christensen and Lægreid 2020). Norway’s Directorate of Health was a key actor but, unlike its Swedish sister agency, the directorate is not exempt from instructions from the government. We do not interpret the government’s delegation of coordination and decision-making to the directorate as a means for shielding COVID-19 policymaking from political interference but rather as a means for ensuring capacity for swift responses to a developing crisis. The Infection Control Act does not oblige the directorate to inform the government of its decisions in advance. During the COVID-19 crisis, the government was often informed prior to the directorate making decisions, but decision authority lay with the directorate (Kvinnsland et al. 2021).

Norway’s public discourse was far less scientific in content and tone than Sweden’s. The Norwegian public was left with the impression that decisions about how to respond to COVID-19 were made by elected politicians who balanced many different considerations, including scientific advice, proportionality, legality, precaution, actions in neighbouring countries and the public mood. In Sweden, authorities took on a balancing role instead. That public appearances were dominated by the government’s chief epidemiologist does not mean there was an absence of debate. Disagreements between scientific camps played out in the mass media. Many critics were part of the medical profession. The health authorities were criticised, for example, for not prescribing a lockdown and for not recommending the use of protective masks.

Intergovernmental relations: centralised vs. decentralised

The third dimension of crisis management we focus on concerns the balance between meeting a crisis with a coherent, nationwide response and with a flexible, localised response (Kettl 2003). The location of infectious disease expertise at subnational level is notably different between the two countries. In Sweden, the regional authorities appoint communicable disease doctors, who are responsible for planning, organising and leading regional responses to a pandemic based on the mandate provided by the Communicable Diseases Act. By locating infectious disease expertise at the regional level, Sweden has organised it alongside specialist healthcare. Norway’s Infection Control Act places infectious disease expertise at local government level, thus alongside primary healthcare and elderly care services. Local governments can respond to a health crisis by implementing a range of local measures, including local quarantines, travel restrictions, assembly restrictions and closure of kindergartens, schools, businesses and events. These competencies are placed with the elected municipal council and, in urgent cases, with the chief medical officer in the municipality.

Based on crisis management scholarship, we expected the national government to centralise authority to respond the crisis and to intensify vertical coordination of any residual decentralised executive responsibility. Overall, the first expectation did not hold while the second did to some extent.

Norway’s decentralised regime produced a heterogeneous government response to COVID-19. Several municipalities, including Oslo and Bergen, the two biggest, closed schools and kindergartens before the government did so on a national scale on 12 March. Moreover, several municipalities banned non-residents from entering and imposed a fourteen-day quarantine for residents entering from outside. According to Kvinnsland (2021: 277–285), 95 of Norway’s 356 local governments enacted local bylaws using the Infection Control Act in 2020. These enactments were most frequent in March and April 2020, thus effectively topping up the national lockdown with quarantines for people coming to the municipality, isolation for those infected, bans on private assemblies and other local restrictions. In the majority of the cases, the local bylaws were enacted by the chief medical officers using the urgency procedure, not by the municipal council – a practice that was later criticised for a lack of democratic accountability by a government-appointed evaluation committee (Kvinnsland et al. 2021, 284). The willingness of Norway’s local governments to take action provoked reactions from national health authorities. For example, local governments in the northern half of the country decided to quarantine travellers from the southern part, despite the Confederation of Norwegian Enterprise urging local governments against restricting travel and imposing quarantines and the government issuing a circular saying the national regulations were sufficient (Kvinnsland et al. 2021). The government lacked authority to overturn local bylaws, but it did have tools to make them ineffective. The Minister of Justice and the Director of Public Prosecutions instructed prosecutors and the police not to prioritise following up on breaches of local bylaws.

Another case of intergovernmental friction in Norway concerned residence in second homes. When the COVID-19 infection numbers in urban areas rose, local governments in rural areas where holiday homes are located urged their urban holiday homeowners to stay away, fearing their healthcare capacity would be overwhelmed. These local governments also successfully lobbied the government to ban travel to holiday homes. On 19 March, the government banned owners of second homes from travelling outside of their registered home municipality. Some local governments urged (to no avail) the police and military to enforce the ban and, if necessary, evict people from their holiday homes. The ban was lifted after 1 month.

State control of local governments intensified somewhat in Norway when the Directorate of Health supervised local governments’ local pandemic plans in January and February 2020 and required improvements in many instances. Moreover, new municipal reporting requirements were implemented. For example, local governments were required to report to the Directorate of Health daily or weekly, with information about stocks of medical equipment and numbers of infections, etc. Moreover, in December 2020, the government introduced a regime with a stronger role for the state in deciding the response at subnational level. Local governments were required to participate in weekly coordination meetings with the understanding that the state could impose its will by enacting a set of policies for a whole region. The first instance of the state making this type of regional decision occurred in January 2021 (Kvinnsland et al. 2021, 283). Overall, though, the state allowed the local governments to act within the responsibilities given to them by the Infection Control Act. The heterogeneity of the government response to COVID-19 was thus a result of conscious design, or stated differently, an outcome that accorded with the guiding principles of responsibility and subsidiarity (Kvinnsland et al. 2021, 56).

In Sweden, subnational levels of government played a far less prominent role in responding to COVID-19 than in Norway. Regions are responsible for stocking testing equipment and other medical necessities for healthcare services, and local governments are responsible for stocking medical equipment for elderly care services. Many subnational governments were criticised during spring 2020 for not being prepared for a pandemic and relying too heavily on a ‘just in time’ philosophy that proved inadequate for a pandemic (Pierre 2020, 488). For example, there were considerable differences among regions in their provision of antibody tests (Läkartidningen, 15 July 2020). A public debate about who was to blame ensued, with the national health authorities stating that the number of tested people should be much higher and regions dismissing testing for the sake of testing. Many regions prioritised the testing of vulnerable groups and medical staff. A national strategy for testing was produced in mid-May, and in June an agreement was reached between the government and the Swedish Association of Local Authorities and Regions (SALAR) on funding for testing.

The Swedish government tasked the National Board of Health and Welfare with assisting regions and municipalities in the purchasing and distribution of personal protective equipment and medical supplies (Socialstyrelsen, 29 April 2020). The board was also tasked with coordinating (including, if necessary, redirecting patients between) intensive care units across regions. This centralisation of authority caused friction. In their comments about the proposed new rules, SALAR was critical of the government authorising a national agency to overrule regions and municipalities (as reported in Swedish Government, Proposition 2019/20:155, 16).

SALAR also has a role both as a partner in mitigating the pandemic and as a special interest organisation beyond any roles defined in legislation. The organisation has a strong position as a coordinating agency between the regions and the national level. Numerous government decisions during the COVID-19 crisis instructed government organisations to collaborate with each other and with civil-society organisations (‘samverkanskrav’) and no organisation was mentioned more often than SALAR (The Swedish Agency for Public Management 2020, 34). It is convenient for the government to strike deals at the central level, and not with 21 regions. An example is the agreement about vaccination in December 2020 where regions do the work but SALAR reports the results to the government.

Echoing what occurred in Norway, a number of Swedish local governments wanted to refuse to offer elderly care to non-residents. Many elderly ‘fled’ big cities to stay in their second homes during the spring. A number of local governments unsuccessfully urged the government to propose a temporary change in the Social Services Act. The Swedish Parliamentary Ombudsman criticised the municipality of Orust for restricting service to non-residents, stating that local governments ‘cannot choose whether to follow the law or not’ (Justitieombudsmannen (The Parliamentary Ombudsmen) 2020). A number of municipalities lost court cases concerning unlawful refusal to provide elderly care to non-residents.

Other instances of intergovernmental friction also occurred. The City of Stockholm successfully appealed in court against a decision by the Work Environment Authority to close a nursing home for insufficient use of protective equipment, such as masks (SvT Nyheter, 8 August 2020). Vertical coordination intensified, in that the Health and Social Care Inspectorate investigated whether necessary health care was given to residents of nursing homes and other elderly care facilities (IVO, 7 July 2020). Regions were strongly criticised for not taking full responsibility for individual medical assessments at homes for the elderly during the pandemic (IVO, 24 November 2020).

Thirty Swedish local governments banned visitors to their elderly care institutions weeks before the government legalised such visiting bans (Svenska Dagbladet, 16 June 2020). The City of Stockholm introduced a ban on visits with the interesting motivation from the leading politician that the city was forced to break the law to save lives because the Public Health Agency was slow to act (Aftonbladet, 12 March 2020).

Generally local and regional authorities acted within their normal responsibilities but in some cases, extensions were made. For example, from July 2020 a temporary law gave local governments authority over closures and preventive measures at bars, cafés and restaurants.

Conclusion and Discussion

This article does not pass judgment on the adequacy and proportionality of the COVID-19 government response in either country. Although Norway’s response appears the most effective at the end of 2020 (OECD 2020; Plümper and Neumayer 2020), Sweden’s response might well be viewed more positively in the long run. At the time of writing, government-commissioned evaluation committees have published their first in-depth reports (Kvinnsland et al. 2021; Melin et al. 2020) and more are forthcoming in 2021 and 2022. That other governments across Europe (including Austria, Denmark, and France) have commissioned in-depth and self-critical evaluations will allow comparative understanding of responses to COVID-19 in the coming years.

The aim of this article is to understand why Sweden’s response to COVID-19 was softer than Norway’s, despite Sweden being affected more by the pandemic. The two countries are similar in almost all relevant explanatory factors. Following the most-similar-systems logic, the explanation for the difference we are interested in can be found among factors that vary. To discuss this, we return to Kingdon’s streams.

The problem stream

Sweden’s agencies enjoy extraordinary autonomy from political interference. The way this difference is rooted as much in tradition as in formalities supports Toshkov, Yesilkagit, and Carroll (2020, 23) notion that ‘institutional and organizational legacies’ play a part in explaining country differences in government responses to COVID-19. They set the stage for pandemic policies. Sweden’s infection control legislation and the philosophy behind it place comparatively high emphasis on the responsibility of individual citizens in preventing the spread of infections. The threshold for selecting strict measures is therefore lower in Norway than in Sweden. The Swedish government trusted authorities and individuals to handle the problem without extensive political involvement which, in comparison with other countries, is remarkable (Andersson and Aylott 2020, 6). They might be seen as having abdicated their responsibility, but this might be an effect of underlying assumptions about the role of the government in crisis situations. The pandemic was not primarily seen as a political issue.

Another difference in the conceptualisation of the problem is that the Norwegian government emphasised scientifically supported healthcare to elderly more than the Swedish government did. In our interpretation, this reflected a key difference between the two countries in what kind of problem COVID-19 was perceived to be – a healthcare crisis or a public health crisis. The government response in Norway included swifter and more decisive moves to restrict outsider access to elderly care institutions and a far higher presence of doctors and trained nurses in elderly care (Kvinnsland et al. 2021, 345). This reflected a relatively stronger emphasis on the healthcare-aspects of the crisis. The Swedish strategy had more of a broader public health perspective by considering repercussions for people and everyday life throughout society at large. Elderly care in local governments was, at first, not a priority. It was overshadowed by anxiety about intensive care capacity of regional healthcare institutions. A Swedish Government Commission (Melin et al. 2020) in December 2020 criticised structural shortcomings for having ‘led to residential care being unprepared and ill-equipped to handle a pandemic’ (p. 2). ‘(I)t took far too long before attention was paid to the specific problems and shortcomings in municipal elderly care’ (p. 7).

In our interpretation, one explanation for this difference in problem perception was the more proactive role played by local politicians and municipal chief medical officers in Norway, with almost a third of all Norwegian local governments making local bylaws to combat the pandemic. Had they wanted to, decision makers at the national level could hardly have acted on any ‘shared ideas’ they might have had ‘about the policy at stake’ (Walgrave and Varone 2008, 367). Norway’s local governments were simply too active in the ‘framing contests’ (‘T Hart and Tindall 2009, 23), predominantly pushing the government towards a lockdown. In March and April 2020, local responses were constantly referenced in ‘race to the top’ debates in national news media, with ‘top’ being the most heavy-handed measures to prevent the virus from spreading.

An additional peculiarity of Sweden’s response was that its policymakers seemed disinterested in the way other countries handled the situation. The framing of a policy issue can be influenced by the choices of others experiencing a crisis at the same time (Toshkov, Yesilkagit, and Carroll 2020). In a financial crisis, a refugee crisis or a pandemic, decision makers look to other countries for solutions (Starke, Kaasch, and Hooren 2013). As COVID-19 hit all countries at the same time and media coverage was intense and internationally oriented, the expectation would be that politicians look to other countries for clues on how to respond and that responses converge across countries. Norway’s government did so explicitly, but Sweden’s apparently did not. Intense contact with health authorities in neighbouring countries did not lead to any mainstreaming of policies. Moreover, Swedish decision makers appeared unconcerned about international critique, an observation that echoes Sweden historically seeing itself as a ‘moral superpower’ (Ruth 1984).

The policy stream

Based on crisis management research, we expected both countries to respond to COVID-19 by centralising decision-making authority from agencies to government ministries. Surprisingly, neither did. In contrast to other countries (Kirlin 2020), Norway and Sweden largely stuck to the plan when hit by COVID-19 and sustained a decentred response regime. This is noteworthy in both countries, for different reasons. In the Swedish case, the decentred regime produced a response so soft that it befuddled outside observers. The government could have taken the wheel and introduced harder regulations, but chose not to. The decentred regime and its response remained undisrupted; even when evidence suggested Sweden was hit harder than many other countries by the pandemic, it still responded more softly than almost everybody else (OECD 2020). In the Norwegian case, a regime with decision-making authority delegated to an executive agency produced what the Prime Minister in a press release called ‘the strongest and most intrusive measures we have had in Norway in peacetime’, measures that ‘interfered directly with people’s everyday lives and how our society works’ (Office of the Prime Minister, 12 March 2020). This relatively decentred regime remained intact even when very strict measures were continued and despite evidence suggesting the virus was more contained in Norway than in most comparable countries. Critics in the political opposition and the legal establishment argued that the government response burdened the bureaucracy with a responsibility that popularly elected leaders are best equipped to shoulder (VG, 27 October 2020).

Differences in the distance between organisational functions affects which considerations are balanced against each other in organisational decision-making (Egeberg 2003). Having local epidemiological expertise inside its own organisations meant that Norwegian local governments could quickly respond to COVID-19. In many instances, local leaders did not consider the national response sufficient to solve what they considered to be an urgent local problem. Debates occurred throughout the country about how to use local powers to combat the pandemic. Since Norwegian local governments have considerable autonomy, local variance in exposure to the infection and local risk factors meant that a patchwork of local school closures and travel bans were layered on top of the national government response. Local COVID-19 responses affected the national response in Norway instead of local governments only implementing national policies. We therefore consider the Norwegian response to COVID-19 a case of multi-layered policy co-formation (Hill and Hupe 2014, 147). Channels for exerting influence bottom-up included bilateral contacts between local mayors and the Minister of Health, lobbying by the Association of Local Governments, and, as just mentioned, local responses fuelling ‘race to the top’ news debates.

The political stream

There are no clear signs in either country that the COVID-19 crisis has harmed citizen trust in government. Remarkably, appreciation of the government response is greatest in Sweden, where damage, so far, is greatest. A study of popular perceptions of the measures in the two countries shows that Swedes placed more trust in their government and health authorities than Norwegians did, and Swedes had a more positive perception of the information they received from their government during the pandemic (Helsingen, Refsum, and Gjøstein et al. 2020). The study also shows that Norwegians thought closing schools was a good idea, while Swedes thought it was good to keep them open. However, Swedes expressed far lower trust in the government’s provision of care for the elderly during the crisis than for the general population (Kantar Sifo, 4 May 2020).

In many countries, including Sweden and Norway, traditional disharmony between the centre and the periphery has been fuelled by the differences in the spread of the pandemic. Capital cities are prone to being more affected and, most importantly, at an earlier stage in a pandemic. In Norway, this disharmony was very evident during COVID-19, with local governments outside urban areas using their powers to restrict travel, impose quarantines and attempt to restrict elderly care to residents in elderly homes. Differences in restrictions between territories caused debate and antagonism.

One discussion initiated by the pandemic in Sweden is the need for centralisation, shifting balance to central government based on the hope of speeding up decision-making processes. Any freedom to act locally could be followed by demands for control of that freedom. The national level could take over responsibilities for (parts of) healthcare and elderly care. The chairman of the Swedish Association of Local Authorities and Regions stated that ‘the risk of nationalisation is evident’ (Dagens Samhälle, 23 June 2020) and media reports indicate that all Swedish political parties favour stricter control of health care and elderly care (SVT, 13 February 2021).

A blame game was played to a certain extent, most evidently in Swedish discussions about testing capacity. Central authorities give ‘advice’ and cannot be blamed for local or regional authorities not following it. Local or regional authorities claim they follow advice from central authorities and cannot be blamed if that advice is vague or delivered too late. In November 2020, the Director General of the Public Health Agency stated that testing would have been easier to achieve with one actor instead of 21 regions. This view was supported by the Minister of Social Affairs and the National Testing Coordinator, Professor Harriet Wallberg (Sydsvenska Dagbladet, 15 November 2020). There have also been some signs of decentralisation requests, however. The lack of medical expertise at nursing homes has led to demands to allow local governments to employ doctors (Läkartidningen, 7 July 2020). Likewise, local governments have indicated a need for more discretion, such as local bans on elderly care visits.

In Norway, the viability of decentralised governance in infection control could be questioned. The smallest local governments preside over only a few hundred residents. Local health preparedness plans were, in many cases, outdated. At the same time, there were also clear benefits of having infection control, elderly care and primary health at the same level of government. Even so, there was a lack of clarity on how far the Infection Control Act allows local governments to enact their own local travel bans. Accountability issues arise when national government officials, like police and county governors, undermine local government infection control measures.

Moreover, the effects of privatisation are criticised. Elderly care lacks personnel, training and equipment, which critics claim is due to cost reductions after privatisation. Stocks of medical equipment that used to be the responsibility of the state-owned pharmacy company simply have not been replaced after pharmacy privatisation in Sweden. This had not been a priority, leading to ad hoc solutions under COVID-19.

In Sweden the development after the first half of 2020 included a more active government and audits showing structural problems in elderly care in general. Though not everything developed according to plan, a window of opportunity might now exist to change existing laws relating to the division of responsibilities among governmental levels and to raise general preparedness for handling this and future pandemics. Responding to a second wave, the Swedish Government in November introduced stricter regulations, similar to those of other countries. These restrictions were not initiated by the Health Agency, however. The government communicated them without the presence of officials from the agency. This has been interpreted as a clear break from the previous policy of letting the agency decide on all measures (Svenska Dagbladet, 22 November 2020). The Swedish government did not oppose any of the policies suggested by the agency during the spring but seems to have changed course. A need to show leadership and perhaps reduced faith in the agency’s ability to predict developments might explain this shift. The diminishing of the rally-around-the-flag effect also has led to open criticism of the government. They are accused of doing too little and too late. At the time of writing (June 2021), crisis management is indeed up for a critical debate in Sweden, with several government initiatives having been launched to address the problems the pandemic has highlighted.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Jostein Askim

Jostein Askim is a professor in the Department of Political Science at the University of Oslo, Norway. His research focuses on political leadership, political-administrative relations, and administrative reform. Journals in which he has published include the Journal of Public Administration Research and Theory, Public Administration, and Government and Opposition.

Tomas Bergström

Tomas Bergström is a senior lecturer in the Department of Political Science at Lund University, Sweden. His research focuses on leadership, organisational reform and conflict resolution in the public sector, especially in cities and municipalities. He is the co-editor, with J. Franzke, S. Kuhlmann and E. Wayenberg, of The Future of Local Self-Government. European Trends in Autonomy, Innovations and Central-Local Relations, Palgrave Macmillan, 2021.