Introduction

The year 2020 was marked by the implementation of lockdowns on a global scale to contain the COVID-19 pandemic by drastically minimizing human contact. This public health measure was instantly described as “medieval”Footnote 1 by several political actors but, facing the surge of the pandemic, many governments in Europe opted for this choice. The social science literature has already begun to analyse the attitudes of citizens towards lockdowns (Mariot et al., 2021), the impact of health measures on trust in governments (Glaurdić et al., 2021; Oude Groeniger et al., 2021), the responses of so-called “populist” parties to this restriction of freedom of movements (Katsambekis & Stavrakakis, 2020) and the political aspects of lockdowns that were presented as health measures during the first wave of COVID-19 (Bristielle, 2021). However, few studies have examined the different realities that were provoked by lockdown measures in different countries and the evolution of these measures over time. For instance, during the first lockdown, French citizens could not leave their home without a self-signed “derogation to travel” certificate proving a “compelling reason” to go out, while in Germany, the ban on leaving one’s home was strongly recommended but not imposed.

This chapter has a dual objective. On the one hand, it aims to highlight differences in the stringency of lockdowns across countries. On the other hand, it seeks to identify some explanatory factors behind the observed differences. The comparative analysis builds on the EXCEPTIUS database and covers 20 countries in the first wave, and 15 in the second wave. The study period is from January 2020 to December 2020 and covers the first wave (January–June 2020) and the second wave (July–December 2020) of COVID-19. First, we distinguish between strict and flexible lockdowns, using an indicator of coerciveness of the lockdown by compiling different variables, such as the order to “stay home”, the number of kilometres of travel allowed outside home, or the presence or self-completed forms for leaving home. Secondly, we look at whether the level of coerciveness of lockdown measures taken by governments was linked to the health situation in the country (circulation of the virus, number of deaths), and/or to political and economic criteria (wealth of the country, confidence of citizens in the institutions, party family of the ruling party) and if they had significantly impacted on the number of deaths.

State of Art and Hypotheses

Previous studies have already shown that several factors come into play when analysing the differences in the measures that governments adopted to deal with the COVID-19 pandemic in the first wave. For example, Engler et al. (2021) show that the quality of democracy in countries had an important impact on the restriction of individual freedom during the health crisis (see also Kuebler in this volume): the more governments protected freedom outside of crisis periods, the less restrictive measures they implemented during the COVID-19 crisis. Bristielle (2021) and Engler et al. (2021) show that, during the first wave, the measures depended largely on trust in institutions and political actors: the higher was the level of trust in political institutions, the less stringent the measures adopted. These works have highlighted the deeply political aspects of the measures put in place to curb the epidemic. Considering this political dimension, we formulate three hypotheses:

  • H1: During both waves, the coercive nature of the quarantine measures implemented in each country did not depend on the intensity of COVID-19 circulation.

  • H2: During both waves, the effectiveness of health policies was not related to the severity of lockdown measures.

  • H3: During both waves, countries with the highest levels of public trust in institutions implemented less restrictive quarantine measures than countries with low levels of trust in political institutions.

Lesschaeve et al. (2021) examined the economic dimension of the lockdown measures implemented to contain the COVID-19 epidemic. Their study showed that during the first wave in southern and eastern Europe, people were in favour of implementing these measures despite their “high economic costs”. However, in the long run, countries with a low GDP would be less inclined to implement strong restrictive measures that affect the country’s economy, which leads us to our fourth hypothesis:

  • H4: During the second wave, countries with the lowest GDP per capita implemented fewer lockdown measures than others.

Finally, this chapter looks at the effects of political party ideology on the response to the COVID-19 epidemic. Diverse ideologies— for example, liberalism, socialism, conservatism (Mudde, 2004)—can be associated with different positions that advocate both strong restrictions to protect the weakest and hospital care, and more flexible responses in order to guarantee individual freedom and collective rights. Katsambekis and Stavrakakis (2020) have shown that populist parties in Europe responded differently depending on national particularities, and that the right-left axis did not seem to have played a major role to explain their positions. For example, in France, the radical left, represented by La France insoumise (Unbowed France, LFI), and the radical right, represented by the Rassemblement National (National Rally, RN), assessed the situation differently, explaining that the COVID-19 crisis validated their respective ideologies. For LFI, the COVID-19 epidemic was the result of the ecological crisis, while for the RN it was linked to open borders and mass immigration (Chazel, 2020). However, during the first wave both parties defended the implementation of strong lockdown measures. This brings us to the fifth hypothesis:

  • H5: During both waves, the political ideology of the ruling party had no impact on the intensity of the lockdown measures.

To initially test these hypotheses, the chapter uses data coded in 20 European countries for the first wave and 15 countries for the second wave by the EXCEPTIUS project. We compiled an indicator of the stringency of lockdown measures put in place to contain the COVID-19 pandemic during the first and second waves, based on seven ordinal variables, each graduating from low to high the coercive intensity of single COVID-19 measures.

The aggregation of the indicator is done as follows: for each wave, and for each variable, we consider the most coercive measure taken by a country whether at the national or subnational level, considering that, depending on the state form of the countries, major decisions may have been taken mainly at one of the two levels (see Chap. 6).

We then divide each variable by the number of categories it contains, and then we divide the indicator by the number of variables (7) so that its value lies between zero and one. Thus, 0 corresponds to no coercive measure at all and one corresponds to the maximum possible coercive measures. Each value is the maximum of what a country has implemented during each wave (no matter if this score describes 1 day or 3 months); therefore, the indicator does not aim at capturing in which country the measures have been implemented the longest, but rather at capturing which are the maximum measures (in terms of severity) that have been implemented by the countries over a wave.

The indicator is then used to examine the correlation between the stringency of COVID-19 containment measures and contextual factors regarding health, political and economic conditions. Results provide insights into why such measures have been put in place in certain countries and enable to confront the working hypotheses. Respective variables (see respective distribution in appendixes 1 and 2) include:

  • The number of deaths relative to population per million people: we rely on the dataset available on COVID Tracker.Footnote 2 On the one hand, we seek to determine whether there is a correlation between the intensity of the lockdown measures put in place during each wave and the number of deaths at the end of each wave (using as a basis the number of deaths relative to population per million people on 30 June 2020 for the first wave and on 31 December 2020 for the second wave) (H2). On the other hand, we aim to find out whether the measures implemented during each wave were implemented according to the circulation of the epidemic (H1). Considering that the detection of COVID-19 cases varied considerably from one country to another (e.g., depending on whether the tests were reimbursed or not), we look at the number of deaths relative to population per million people at the beginning of each wave to evaluate the circulation of the virus. For the first wave, as COVID-19 only appeared in Europe in March 2020, we use as a basis the number of deaths relative to the population on 31 March 2020,Footnote 3 and for the second wave on 1 July 2020.

  • Confidence in national governments: we rely on the November 2019 Eurobarometer data to account for H3.

  • Level of GDP per capita in 2019: we rely on the European Commission data to test H4.

  • The ideology of the political party of the government in office in March 2020: we distinguish six types of parties in power to examine H5: far left (1), social democratic left and green parties (2), centre (3), liberal right and conservative right (4), far right (5). Regarding coalition governments, we have selected the party to which the president or prime minister belongs.

Analysis and Results

Different Types of Lockdowns

Three groups of countries can be distinguished according to the intensity of the measures implemented: light lockdown (from 0 to 0.33), medium lockdown (from 0.33 to 0.66) and strict lockdown (from 0.66 to 1).

The data show that two countries (out of 20) had light lockdowns in place in the first wave (Finland and Sweden), and five countries (out of 15) in the second wave (Denmark, Estonia, Finland, Poland, Sweden). The first group of countries is characterized by the fact that the governments in place relied primarily on the responsibility of citizens rather than on coercion: the injunctions to stay at home were primarily recommendations, and many shops remained open during the period. For example, in Finland, in the first and second waves (with a respective lockdown score of 0.24 and 0.17), the government recommended that cafés and restaurants be closed, and banned gatherings of more than ten people.

The Swedish case received particular attention. The indicator of coercive lockdown measures shows that there were no measures during the first wave (0) and a very light closure measure of non-essential shops during the second wave (0.07). The Swedish centre-left government led by Stefan Löfven (Sveriges Socialdemokratiska Arbetareparti [Swedish Social Democratic Party]) had announced during the first wave of COVID-19 that the fight against the virus was a “marathon, not a sprint”. For instance, the Swedish government encouraged citizens to work from home, when possible, discouraged travel and going out for those who felt ill and for citizens aged 70 or older. During the first two COVID-19 waves, the Swedish government stood out by the trust it placed in its citizens to limit their contacts. When, in most European countries, masks started to become compulsory after the first wave (see de Saint Phalle’s chapter in this volume), the government stood out again, as wearing masks never became compulsory. All over Europe, the Swedish case was taken as an example by the supporters of the lifting of sanitary restrictions (e.g., showing that another model was then possible), and it was mobilized by other political and scientific actors as an example not to be followed (e.g., showing the impasse of the search for herd immunity).

Regarding medium lockdowns, 14 countries out of 20 implemented them during the first wave (Austria, Belgium, Czech Republic, Denmark, Estonia, Germany, Greece, Hungary, Ireland, Luxembourg, Netherlands, Poland, Portugal, United Kingdom) and 9 countries (out of 15) did so during the second wave (Austria, Belgium, Czech Republic, Hungary, Italy, Ireland, Portugal, Spain, United Kingdom). In most of these countries, governments have also called on citizens to take responsibility. This was the case in the Netherlands, for example, where the government declared relying on the “intelligence” and “self-discipline” of citizens. There was no form to fill in to leave home, no time limit and no space limit.

Finally, 4 countries (out of 20) implemented strict lockdown during the first wave (Spain, France, Italy, Romania) and only 1 (out of 15) during the second wave (France). During the first wave, countries such as Spain, Italy and France relied on coercion. In these three countries, citizens were forbidden to leave their homes except for essential reasons such as buying food or walking their pets. Very heavy fines could be imposed, and prison sentences could be incurred for non-compliance with lockdown (e.g., in France, up to 3750 euros and 6 months in prison). Citizens were required to carry a self-signed sworn statement/travel certificate every time they left their home to justify a compelling reason. In France, during the first lockdown, citizens were allowed to go out for 1 h/day, within a maximum radius of 1 km. France is a special case as during the second wave it implemented both strict lockdown measures and a strict curfew, while most other countries chose one or the other.

Initial Explanation of the Stringency of Lockdowns

When we look at the impact of the circulation of the virus on the lockdown measures, we can observe differences between the two waves (Figs. 12.1 and 12.2). In the first wave, a positive correlation between the circulation of the virus and the severity of the lockdown measures implemented is observed (r = 0.53): the higher the level of virus circulation (proxied by the number of deaths at the beginning of each phase), the stricter the lockdown measures—thus invalidating H1. However, for the second wave, if the correlation remains positive it appears much lighter and not significant (r = 0.27). As such, the posited relationship between the intensity of circulation of COVID-19 and the coerciveness of lockdown measures decorrelates and decreases over time.

Fig. 12.1
A line graph plots the indicator of coercive measures plotted against deaths per million. The fitted line displays a linear increasing trend. The 95% confidence interval area also increases over time, widening throughout the year.

Indicator of coercive measures and deaths per million habitants, beginning of the first wave

Full size image
Fig. 12.2
A line graph plots the indicator of coercive measures against deaths per million. The fitted line exhibits a linear increasing trend. The 95% confidence interval widens throughout the year, particularly less at 400 on the x-axis.

Indicator of coercive measures and death per million habitants, beginning of the second wave

Full size image

Regarding the impacts of economic and political factors, the analysis of the first wave data reveals that only the level of institutional trust had a significant and negative impact on the types of lockdown measures implemented (Fig. 12.3). The lower the level of institutional trust, the higher the intensity of coercive measures (r = −0.61). Similar results are observed in the second wave. The link between institutional trust and coercive measures is still significantly negative with a similar intensity (r = −0.59), thereby confirming H3 for the two waves. 

Fig. 12.3
2 line graph plots the indicator of coercive measures against deaths per million. The fitted line exhibits a linear decreasing trend. The 95% confidence interval widens throughout the year, particularly less at 40 on the x-axis indicating a decreasing trend.

(a) First (on the left) and (b) second (on the right) waves. Indicator of coercive measures and trust in political institutions

Full size image

Meanwhile, it seems that the GDP level, standing for the capacity of the countries’ economy to cope with strong lockdown measures and the economic slowdown it implies is not linked with the coerciveness of implemented lockdowns (r = −0.24 with no significance for the first wave and r = 0.01 for the second wave). This result invalidates H4. In addition, the type of political party in power does not appear as a significant factor either to explain the decision to implement strong or weak lockdowns, neither during the first wave (r = 0.01) or the second wave (−0.16), here confirming H5 (Figs. 12.4 and 12.5).

Fig. 12.4
2 line graph plots the indicator of coercive measures against G D P per capita. The fitted line exhibits a linear decreasing trend. The 95% confidence interval widens throughout the year, particularly less at 40000 on the x-axis indicating a decreasing trend.

(a) First (on the left) and (b) second (on the right) waves. Indicator of coercive measures and GDP per capita

Full size image
Fig. 12.5
2 line graphs plot the indicator of coercive measures against the positioning of the ruling party on a left-right axis. In the first graph, the fitted line remains consistent. In the second graph, the line has a linear decreasing trend. The 95% confidence interval widens throughout the year, particularly less at the center on the x-axis, indicating a decreasing trend.

(a) First (on the left) and (b) second (on the right) waves. Indicator of coercive measures and party in power

Full size image

Finally, regarding the impact of lockdown measures on mortality, during the first wave and the second wave, only a very low correlation exists with weak significance (r = 0.19 for the first wave, r = 0.34 for the second wave), partially validating the low impact of lockdown measures on mortality (H2) (Fig. 12.6).

Fig. 12.6
2 line graph plots the indicator of coercive measures against deaths per million. The fitted line exhibits a linear increasing trend. The 95% confidence interval widens throughout the year, particularly less at 400 and in the 2nd graph is less at 1000 on the x-axis indicating an increasing trend.

(a) First (on the left) and (b) second (on the right) waves. Indicator of coercive measures and death per million (30 June/31 December 2020)

Full size image

Conclusion

This chapter has provided two main explanations for decision-making in times of crisis. On the one hand, while almost all European countries drastically restricted the movement of citizens and urged them to stay at home, there were also large differences between these countries. The lockdowns implemented to contain the circulation of COVID-19 have shown very different realities. On the other hand, only two of the four variables we selected explain the level of severity of the lockdowns put in place. The circulation of the virus had a significant impact (the higher the circulation, the stricter the lockdowns) but only during the first wave, while the level of trust of citizens in political institutions was a much more robust explanatory factor during both waves (the lower the level of trust, the stricter the measures).

The chapter also showed that neither the level of GDP nor the incumbent political party explained the severity of the lockdown measures during the first two waves. Regarding the level of GDP, no correlation can be made between the latter and the indicator of coercive measures.

Regarding the ideological position of the governing party, it is interesting to note that the left-right axis does not seem to have had any impact on the position taken by the different political parties on lockdown measures. For instance, the left was largely divided on the policy responses to the COVID-19 epidemic. In France, the main left-wing party, LFI opposed lockdown measures from the second wave onwards. The party proposed the reopening of essential and non-essential shops and the introduction of a “rolling society” (société par roulement). In Spain, Podemos (We Can), the Spanish counterpart of LFI, widely criticized the “laxity” of right-wing parties in some autonomous communities led by the conservative Partido Popular (Popular Party, PP), as in the Community of Madrid, and defended lockdown as a measure to protect the most vulnerable citizens.

To go further, it could be interesting to study the impact of complementary context variables such as democratic level of the country, hospital capacity or trust in the police. Furthermore, a double synchronic and diachronic analysis could be performed on the first and second wave (and also include the third wave) to categorize the countries according to the strictness of the lockdowns for each wave but also in an evolutionary way to perform an average over the study period.