Covid-19 and life satisfaction across Europe

ABSTRACT

We analyse the importance of the spread of Covid-19 and related policy measures for individual life satisfaction in a sample of 15 European countries. We find that life satisfaction is negatively correlated with the spread of Covid-19 in Northern Europe, whereas this correlation is insignificant in Southern and Western Europe. This result is mitigated by the stringency of policy measures in the North. By contrast, the stringency of policy measures is negatively correlated with life satisfaction in the West. These results indicate that while policy measures can reassure the population and increase life satisfaction, beyond a certain threshold these policy measures may decrease life satisfaction.

KEYWORDS:

• COVID-19
• life satisfaction
• well-being
• cross-country comparison

JEL CLASSIFICATION:

• I31
• I38
• I18

I. Introduction

The ongoing pandemic caused by the coronavirus disease (COVID-19) continues to be a dangerous threat to human health, which by April 2021 had led to 3 million deaths worldwide (WHO 2021). Attempting to contain the spread of the virus and to avoid the breakdown of national health systems, many governments have implemented lockdowns or similar measures of social distancing, including, e.g. workplace closing, school closing, or measures limiting travel activity or social gatherings. While these measures often succeeded in ‘flattening the curve’ of COVID-19 cases and helped to assure capacities in intensive care units in hospitals (Flaxman et al. 2020; Fang, Wang, and Yang 2020), they lead to high costs to economic prosperity. Furthermore, individual welfare has been affected by the economic consequences, including the fear of unemployment, but also by the effects on individuals’ daily activities and their ability to interact with family and friends. The resulting lack of social contacts might have reduced the ability of coping with negative experiences and could have further increased psychological and social suffering such as feelings of loneliness or boredom (Brodeur et al. 2021a, 2021b, Brülhart et al. 2020). Considering these indirect effects, welfare costs related to lockdowns and social distancing might have been larger than expected a priori (Brodeur 2021b). However, the total effect of the policy measures is ambiguous as they might also have provided a comforting feeling of reduced infection risk, decreasing worries about families’, friends’, or own health.

In this study, we combine rich survey data and data from public authorities to assess how COVID-19 cases, policy measures and the economic recession affected individual life satisfaction in Europe. This provides important empirical insights into both direct effects of the pandemic, and unintended side effects of policy measures.

II. Data and methodology

Our analysis is based on three different data sources. First, we use two waves of the ‘Living, Working and COVID-19 Survey’ collected by Eurofound in two survey waves between the beginning of March 2020 and the beginning of May 2020, and the end of June 2020 and the end of July 2020 (Eurofound 2020). The survey was allocated via snowball sampling method and social media advertisements and asked detailed questions on topics ranging from quality of life, work, financial security, and quality of public services during COVID-19. In our analysis, we include the surveyed individuals from 15 EU countries (see footnote 2), covering 48,869 individuals (35,555 wave 1, 13,314 wave 2).¹ Second, we use daily data on COVID-19 cases from the EU Open Data Portal that is based on reports from health authorities worldwide (ECDC 2020). Third, we use daily data on COVID-19 policy measures provided by the University of Oxford and Blavatnik School of Government (Hale et al. 2020; Blavatnik School of Government 2020).

We analyse self-assessed life satisfaction measured on a 1 to 10 Likert scale using linear OLS regression models. Our explanatory variables are the weekly average of daily new COVID-19 cases in a country relative to all other countries, various indicators for government policies, and a large set of control variables, both at the individual and at the country level.² We use relative cases rather than total cases because we expect individuals to weight their own country’s COVID-19 cases relative to those in other countries. The policy measures are summarized by a daily index (from 0–100) that summarizes the stringency of each country’s COVID-19 policy measures (Phillips and Tatlow 2021). The index accounts for school and workplace closings, cancelling of public events, restrictions on gathering, closing public transport, stay-at-home requirements, restrictions on internal movement, and international travel controls.

We interact the explanatory variables of interest with country group indicators to assess how well-being was affected by different policy environments. We construct three country groups which represent different combinations of stringency in policy responses and different time trends in COVID-19 cases (see also Figure 1)³

Figure 1. Relative new Covid-19 cases and stringency index across country groups and time.

Northern European countries followed a more laissez-faire approach with fewer and less stringent measures.

Western European countries had slightly more time to prepare their health system as COVID-19 cases started to rise later than in the southern European countries. On average, measures were less stringent than in the south.

Southern European countries were hit early in the pandemic with high COVID-19 cases, struggled with high death rates, and were among the first European countries to implement stringent measures.

Data sources: See Section 2. – Notes: Relative new cases are defined as average new COVID-19 cases relative to the average new cases in other sample countries at a certain date.

Overall, self-assessed life satisfaction from the first to the second survey wave displayed a slight increase in the North (7.2 to 7.3), and a larger increase in the West (6.6 to 7.1) and in the South, although from a much lower level (5.6 to 6.5).

III. Results

Table 1 shows estimation results for different sets of explanatory variables. Overall, we find that relative new COVID-19 cases are not significantly correlated with individuals’ life satisfaction, controlling for employment status. Yet, the relation is different by country group: the correlation between relative COVID-19 cases and life satisfaction is significantly negativ in the North country group. Individuals in these countries may perceive cases with greater worries as lockdown measures were more laissez-faire and, therefore, less effective in containing the virus. The results indicate that an increase in relative new cases by one unit⁴ in the North decreases life satisfaction by 0.143 on a 0–10 Likert scale (column 4).

Table 1. Regression results

	(1)	(2)	(3)	(4)	(5)	(6)	(7)^b
Relative new cases	-0.023(0.027)		-0.016(0.027)
Stringency Index		-0.009**(0.004)	-0.009*(0.005)
North # Relative new cases				-0.143***(0.032)	-0.065(0.052)	-0.084**(0.030)	-0.081***(0.024)
West # Relative new cases				0.036(0.045)	0.018(0.043)	0.003(0.043)	-0.048(0.039)
South # Relative new cases				-0.015(0.048)	0.014(0.052)	0.007(0.046)	-0.007(0.040)
North # Stringency index					0.019*(0.010)	0.022**(0.007)	0.017**(0.006)
West # Stringency index					-0.011**(0.004)	-0.019***(0.004)	-0.017***(0.003)
South # Stringency index					-0.010(0.006)	-0.010(0.007)	-0.009(0.005)
North	-0.175(0.253)	-0.321(0.291)	-0.331(0.275)	0.108(0.208)	0.522***(0.122)	-0.068(0.203)	-0.286(0.192)
South	0.060(0.121)	0.096(0.107)	0.103(0.115)	0.198(0.269)	0.083(0.273)	-0.232(0.249)	-0.422*(0.212)
Constant	2.963***(0.599)	2.993***(0.598)	3.012***(0.516)	2.553**(1.001)	3.155***(0.975)	2.803**(1.025)	3.621***(0.948)
Socio-demographics	✓	✓	✓	✓	✓	✓	✓
Trend, Time Dummies	✓	✓	✓	✓	✓	✓	✓
Life Satisfaction 2016^a	✓	✓	✓	✓	✓	✓	✓
Country groups # duration policy measures						✓	✓
Country groups # trust in healthcare							✓
N	47,881	47,881	47,881	47,881	47,881	47,881	47,750

Standard errors in parentheses. * p < 0.1, ** p < 0.05, *** p < 0.01 – Additional controls: age groups, gender, education groups, urbanization level, having a partner, living alone/with a partner, health status, financial situation, week dummies, time trend. Reference group: Country groups – West. ^aThis variable controls for differences in live satisfaction across countries before the pandemic and the latest available Eurofound European Quality of Life Survey 2016. ^bThe variable “trust in healthcare” is not available from some respondents which explains the drop in the number of observations to 47,750.

We also find a positive relationship between life satisfaction and the stringency of policy measures in the North. An increase in the stringency index by 10 points in the North is associated with an increase in life satisfaction by 0.019 (column 5). Individuals in the North may welcome more stringent measures to lessen their worries over new COVID-19 cases because they initially face less stringent measures compared to other countries in Europe. Importantly, a robustness test excluding one country at a time shows that the results for Northern Europe are strongly driven by the Swedish experience – the country with the least stringent Covid-19 policy measures during the pandemic.⁵ In contrast to the North, life satisfaction in the West is negatively correlated with the stringency in measures. All these patterns are stable in all model specifications.⁶ The results in Columns 2 and 3 show for overall Europe that the size of the coefficient of the ‘Stringency Index’ stays stable after including relative new cases to the model. Therefore, the average negative effect for Europe as a whole is driven by Western Europe, as a comparison of column 2 with columns 5–7 makes clear. Table A6 in the Appendix features the coefficients for all control variables.⁷

IV. Conclusion

We find that life satisfaction is only negatively correlated with the spread of Covid-19 in the Northern European country group, the correlation is insignificant in the South and in the West. This result is mitigated by the stringency of policy measures in the North. By contrast, the stringency of policy measures is negatively correlated with life satisfaction in the West. This implies that while stringent lockdown measures might be able to offset the worries related to the pandemic, they decrease life satisfaction if they pass a certain threshold in their stringency. This in consistent with Fetzer et al. (2020) who find that in the first wave of the pandemic, there was broad public support for COVID-19 containment measures if governments are seen to take decisive actions. Our results support this view for countries in which measures are initially less stringent, i.e. the North. However, in countries with relatively more stringent measures already in force, individuals seem to assess containment measures negatively.

Our findings highlight that policymaker not only need to consider how their decisions affect the spread of COVID-19, but also how such choices influence overall life satisfaction of their population, as also argued by Layard et al. (2020). Yet, this seems to be a highly complicated quest, as the need for new measures might change over time depending on the overall assessment of COVID-19 cases and the negative consequences of measures already in place.

Disclosure statement

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

Notes

¹ For our analysis, we create a cross-sectional dataset pooling data over both waves. To account for changes over time in our analysis, we include a daily time trend and dummies for calendar weeks in our regression model (see section 3). We weight the observations according to country size in order to make the results representative at the European level.

² In our empirical analysis, we use dummy variables for individual characteristics such as age and gender in order to take into account the oversampling of certain groups (e.g. women and older individuals).

³ Eastern countries are excluded from the estimation sample, as there was no strong increase in Covid cases, and

therefore little stringency measures, during the observation period. Our country groups are defined as follows. South: Cyprus, Greece, Italy, Portugal, Spain. West: Austria, Belgium, France, Germany, Ireland, Luxemburg, Netherlands. North: Denmark, Finland, Sweden.:

⁴ E.g. an increase in relative new cases from 1 to 2, i.e. a doubling of relative new cases.

⁵ Results available from the authors.

⁶ Results are robust to a variety of alternative measures for the spread of Covid-19. Table A3 in the Appendix displays results using relative total cases, Table A4 display results using absolute new cases and Table A5 displays results using absolute total cases as main explanatory variable for Covid-19 spread.

⁷ Results not reported but available from the authors upon request.