In Virus Veritas Lockdown and Happiness Under COVID-19

Since antiquity, philosophers have reflected on what contributes to a better life and makes people happy. In recent decades, empirical studies on this issue have flourished in a variety of disciplines (for a review, see among others Dolan et al., 2008, and Powdthavee 2007, or the monograph Frey 2018). However, while understanding which factors impinge on subjective well-being is of utmost importance in general, it becomes even more intriguing in situations where some major determinants of happiness are jeopardised by widespread contingent factors. This surely happened in 2020 due to the COVID-pandemic.

The whole World Happiness Report 2021 focused on the effects of COVID-19 on the structure and quality of people’s lives (Helliwell et al., 2021). According to it, during this time emotions changed more than did life satisfaction. The aim of the present paper was to exploit the COVID-19 pandemic as a “natural experiment” to investigate individual happiness and the effects of some period-specific determinants. Moreover, unlike the WHR 2021, we carried out the research during the lock-down period and, instead of analysing satisfaction with aspects of individual life, we investigated on the factors people believe contribute most to individual well-being. This was done on the assumption that under an extraordinary and dramatic event reference values may change or, at least, what people really care about may come to the fore more spontaneously.

A Latin saying, in vino veritas (in wine lies the truth), suggests that persons under the influence of alcohol are more likely to express their true views and desires. Alcohol loosens moral restraints, enabling people to express what they really think and feel. We rephrased the saying as in virus veritas, under the assumption that COVID may have a similar effect on people’s feelings, being an extraordinary event that has dramatically changed the way our lives are organized and conducted. What people really care about may emerge more spontaneously than at other times, or at least, COVID may change the relative importance of things (scale of values) and the way they relate to individual happiness. Of course, we are not suggesting that studies conducted in “normal” times are deceiving and misleading. Our main point is that there are special moments in life when people see their attitudes and preferences in a different and extraordinary way. By observing agents in such situations, we may capture information that is otherwise not available. The pandemic gave us this opportunity. The evidence we tried to obtain during this unique natural experiment is also unique, as well as complementary (not necessarily substitutive) of evidence captured in normal and ordinary times.

The COVID-19 pandemic is a unique event, unlike anything we have seen in the recent past. Its uniqueness lies not only in the dimensions of the viral attack but also in the response to the threat by policy makers in various countries. The lockdown measures implemented in some countries forced millions of people to remain at home for weeks and in some cases, as in Italy, for about two mouths. It goes without saying that the Italian experience was also unique in the violence of the epidemic, the human, social and economic costs and the time spent in the lockdown. It had and continues to have many dramatic effects in terms of deaths and economic costs, all of which have yet to emerge. However, it certainly (and even more) modified people’s habits and threatened their lives and relationships, especially in the lockdown period (Helliwell et al., 2021).

This is different from what Kahneman (2011) defines as a focusing illusion. In his view, contextual factors may cause judgment bias. Instead of assessing their lives as a whole (target question), people focus on their contingent affective state (heuristic question). Rather, it is more in line with what Thaler and Sunstein (2008) sustain regarding arousal state, which varies in time, giving rise to dynamically inconsistent appraisals. In a cold state (in our case “COVID-off”), individuals do not fully appraise the importance of certain aspects of their lives. In a hot (aroused) state (in our case “COVID-19”), they appraise them differently, due to the effects of “arousal” on their mood or value system. In the language of economics, one can say that actual experience modifies agents’ perceptions, changing their indifference curves and thus their appraisal (Bimonte & Punzo, 2016). Whether or not this modification endures is another question.

Empirical studies on life-evaluation have identified a number of variables explaining changes in individual happiness. The variables can be grouped in three main categories: economic (such as income, employment status), social (education, marital status, social capital) and health-related (physical and mental) (DeLeire and Kalil 2010; Dolan, Peasgood, and White 2008; Easterlin 2003; Frey, 2018; Frey, Stutzer, and Easterlin 2002; Helliwell, Layard, and Sachs 2017; Rojas 2011). There is now consolidated empirical evidence showing that once basic needs are met, material possessions do not contribute significantly to individual happiness (DeLeire and Kalil 2010; Helliwell 2003; Kasser, 2002). It has been demonstrated that relational goods and social capital also matter (Bruni and Stanca 2008; Diener & Seligman 2009), and people whose values are centred more on material possessions or income aspiration have a greater probability of being less happy (Kasser, 2002; Killingsworth and Gilbert 2010).

The COVID-19 pandemic and the policy response (lock-down) provide a kind of counterfactual situation. The latter concerns how things would have been under a particular/hypothetical circumstance. In our specific case, the (rhetorical) question is: what would happen to individual happiness if people were faced with a threat to their economic, social and relational situation? Would relational aspects still prevail over material possessions in determining individual happiness?

Q1. In the dramatic lockdown period, income, health and interpersonal relationships were threatened. How did this threat affect reported happiness?

Q2. Under these threats, were social capital and relational goods as opposed to more material aspects still considered important determinants of happiness?

Q3. Did the effect of lock-down on happiness depend on individual characteristics?

To address these questions, we carried out a web-survey during the COVID-19 lockdown period. Ordinal Logit Regression (OLR) was used to detect whether (and how) perceived or actual COVID impacts, respondents’ profiles and factors considered important for a happy life related to individual happiness. As for the latter aspect, unlike other papers that investigate people’s satisfaction with aspects of their lives, we investigated their opinion of what matters for a happy life, while assessing their happiness and self-reported impact of the pandemic on income, personal relations and health. Our outcomes suggest that under lockdown, the main direct effect on happiness was through the effect on interpersonal relationships. Subjects who reported that COVID and lockdown had threatened their interpersonal relationships were significantly less likely to report higher levels of happiness. This result is always true, irrespective of the regression model used. Its strength and significance increase when controlling for other important covariates. Moreover, social capital and social interpersonal relationships also emerged as important determinants of happiness. An important gender effect emerged as well. However, considering the sampling procedure, imposed by time constraint and the difficulty of administering questionnaires during lockdown, results has to be taken with caution.

Table 4 summarizes the main estimates. The four columns represent different specifications of the model. Specification I is the model of Eq. (1) while the others progressively exclude variables that turned out to be non-significant in the first specification. Diagnostic tests confirmed the robustness of the results. The first three rows show the effect of what we considered the main direct drivers of happiness in our model. To make things simpler, we collapsed the original four-level survey responses (see Table 1) of these three variables into two levels (dummy variables) aggregating “a lot” with “somewhat” and “little” with “almost nothing”. The results show that of these three direct drivers of happiness during the pandemic and lockdown, only relational-covid appeared to have a significant (negative) effect on the probability of a higher level of happiness. This is so, irrespective of the regression model used, and its significance improves when controlling for other covariates.⁸ Although they had the expected sign, income-covid and health-covid turned out not to be significant. It is worth recalling that, unlike what normally done in other studies, these variables were evaluated in term of impact, i.e. variation induced by COVID.

The effect of relational-covid was remarkably great. Persons who reported that COVID affected their relational life a lot or to some extent had a 45%⁹ lower probability to report higher levels of happiness. Income-covid had the expected negative sign and had less impact than relational-covid (18%), but was not statistically significant. These results were also confirmed by the other models. This may be partly due to the fact that the social lockdown and consequent difficulty in spending money made money temporarily less useful than usual and therefore less desirable. If it is not money itself that affects happiness but the goods and the services that money buys, the obvious difficulty of purchasing goods during lockdown may have reduced the influence of income on happiness. The same was true of health-covid. Fear of COVID was quite common in our population but did not influence the probability of being very happy.

With regard to the role of personal aspects, age and education turned out to be not statistically significant. This is not surprising. The happiness literature is unclear about the link between age and subjective well-being, nor does it show any clear-cut conclusions on the link between education and subjective well-being. Some empirical studies have found a positive effect of education on happiness (Di Tella et al., 2001, Stevenson and Wolfers, 2008, Cuñado and de Gracia, 2012), while others have shown a non-significant (Inglehart & Klingemann, 2000) or even negative effect (Clark and Oswald 1996). While a direct positive effect of education on SWB is understandable, being evidence of “self-confidence” or “self-esteem”, the negative effect has been explained by the fact that a higher level of education may increase the misalignment between expectations and fulfilment. This was suggested by a paper on highly educated Italians (Ruiu and Ruiu 2019).

In our case, results may depend on the characteristics of the sample, where about 60% of respondents had at least a bachelor’s degree. The variable had a low variability and was skewed on the right. Education may be used as a proxy for employment status and hence income. Both are usually associated positively with happiness. Due to the percentages mentioned just now, this effect was not caught by our sample.

Interestingly, the effect of religious faith appeared to be positive and significant, i.e. believing increases the probability of being happy. Those who defined themselves as religious had more than twice the probability (108%) of reporting a higher level of happiness than those who defined themselves as not religious. The question about religious faith was very broad (Would you define yourself as a believer?) for two main reasons: first to lighten the questionnaire and obtain more reliable answers; second, we were not interested in differentiating faiths or observation. Our idea was to capture an internal sense of religious faith and to determine whether it can affect happiness, especially in unusual times like the pandemic. The variable scored one if the answer was positive and zero otherwise (i.e. no or do not know).

This result is in line with other studies that found positive relationships between subjective well-being (SWB) and religious faith in various populations, irrespective of religion (see Minkov et al., 2020 and references therein). The role of participation in groups and networks with religious inspiration was stressed by Lim and Putnam (2010). Since our questionnaire was administered during lockdown, participation was prohibited and this effect should be absent or very small. In our case, we can guess that feeling part of a believer community, in most cases the Catholic Church, may have a positive impact on happiness, especially in anxious and uncertain times, such as the pandemic. Moreover, faith is somehow a form of certainty and self-assurance which can improve people’s self-confidence. In a period such as lockdown, this can have a positive effect on happiness.

Interestingly enough, hometown population also had a significant coefficient. In the regression we constructed a dummy variable equal to zero if the population was less than 10,000, and one otherwise. The coefficient was negative and significant: living in a small village contributed to individual well-being. Those who lived in larger towns had 35% less probability of reporting a higher level of happiness.

The relationship between happiness and town size, or happiness in rural versus urban areas, has often been studied, without reaching clear-cut conclusions (Itaba, 2016). Some studies found higher levels of SWB in rural areas (Shucksmith et al., 2009, Knight and Gunatilaka, 2010, Davern and Chen, 2010); others found the reverse relationship (Millward and Spinney, 2013), while in some cases no sharp disparity emerged (Mookherjee, 1992; Best et al., 2000). This relation may depend on the stage of a country’s development: in less-developed countries happiness is considerably greater in urban settings, but this urban–rural differential tends to disappear in developed countries (Easterlin et al., 2011).

Using a very large dataset for community-level life satisfaction in Canada, Helliwell and co-authors recently concluded that “life is significantly less happy in urban areas” (Helliwell et al., 2018).¹⁰ Our results seem in line with Helliwell’s. On theoretical grounds, living in a small village during a pandemic can have negative and positive side effects. First of all, living further from large hospitals, where the quality of healthcare can be perceived as higher, can increase anxiety. Some on-line services common in big cities are not available in small towns (e.g. home delivery of food). When people are forced to stay at home, the size of the house may be more important than the size of the town. Houses are generally larger in small towns due to lower housing prices, and more likely have gardens and surrounding countryside where people can potter or walk without breaking the lock-down rules or infecting/being infected by others. Finally, remaining at home in a small town engenders less feeling of isolation than in a city (Mouratidis and Yiannakou 2022). There is more feeling of community and people help each other. Neighbourhood relationships are real and tangible. “Mr. Smith” is not just a name on a door seen from the elevator, but a person in flesh and blood, known to the townspeople. Our result suggests that the positive elements are more important than the negative ones, and is consistent with the results of other studies, since many positive effects of living in a small village are relational. As evidenced in the last World Happiness Report, trust and the possibility to count on others are very important aspects for life evaluations, especially in the face of crises (Helliwell et al., 2021).

Another interesting finding was the significant negative coefficient for gender. In Italy, women seem to be less happy than men (ISTAT, 2020). Our results suggest that this effect is strong. Indeed, according to our computation, being female reduced the probability of being very happier by about 40%. The reasons invoked to explain the lower level of happiness of women in Italy may also explain the lower probability of happiness in our particular sample, namely the double burden of work of wives who work at home and outside (see Zoch et al., 2021). Although we do not have data on this aspect, it is likely that most of the burden of the lockdown was carried by wives. During lockdown schools were closed and children stayed home, dramatically increasing the time dedicated to childcare. If the division of housework is unfair and the bulk of it falls on wives, it is not surprising that lockdown reduced the probability of female happiness.

In this last section we concentrate on the relationship, if any, between happiness and individual beliefs about happiness determinants. We try to establish whether beliefs about what is really important for being happy are somehow related to reported levels of happiness. Again, we did not ask respondents about their satisfaction with different domains of their lives, but how important they believed these domains to be for individual happiness. As stated before, the questionnaire contained two groups of questions aimed at appraising respondents’ attitudes regarding the main drivers of subjective well-being. We distinguished personal and institutional/social aspects (Table 2). Then, we carried out a Principal Component Analysis to reduce the dimensionality of our dataset. It found five components that we called: social capital, health & relationships, public goods, materialism and populism.

The first point worth noting is that the regression results preserved the distinction between the social and individual dimensions that we postulated. As Table 4 shows, only social capital and health & relationships were statistically significant. In both cases the sign was positive. Social capital turned out to be an important determinant of happiness. This may appear partially counterintuitive. Italy cannot be defined as a country with robust social capital. To the contrary, corruption is high, tax evasion is dramatic and widespread, and organized crime is stronger and more pervasive than in other countries. Individuals who believe that these elements are important may therefore be expected to display a lower level of happiness, given the Italian situation. However, once again, it is worth recalling that we investigated what people believed to be important to pursue happiness rather than how satisfied they were with aspects of their own lives and environment. Our results suggest (and somehow confirm) that pro-social and less materialistic beliefs foster happiness. Believing that social capital is an important driver of subjective well-being also means believing in a cooperative idea of society and feeling that a society which faces an external enemy, such as the coronavirus, in a united and collaborative way has a greater chance of success.

This result was consistent with and reinforced by our findings with respect to health & relationships. Relational goods are local public goods which are simultaneously consumed and produced (Gui and Sugden 2005). Examples of relational goods are affective relationships such as love, family relationships, friendship, neighbourhood connections, and many other kinds of social interactions. Our findings suggest that people who say that sound health and relationships are an important determinant of happiness are more likely to report a higher level of happiness, although this applies with different strength to women and men. In fact, the interaction term between “Gender” and “Health and relationships” turned out to be negative and significant.

In a period when health and relations are widely threatened, this may appear counterintuitive, especially if we consider that subjects who reported a worsening of their interpersonal relationships due to lockdown also showed a lower level of happiness. We would expect persons who give relatively more importance to relational goods to have a lower probability of reporting a higher level of happiness. Deeper analysis shows that this result is consistent, if we distinguish between “attitudes” and “COVID impacts”.

Being pro-social and pro-relational is good for individual well-being. According to our assumption “in virus veritas”, this is particularly true in dramatic times, when people focus more on what makes life worth living. Moreover, those who give relatively more importance to health and relational goods are likely to organize their lives coherently. They are more likely to have a healthy life-style and to observe the rules of hygiene. They feel equal to the challenge of the virus and may therefore be happier. Similar considerations apply to relational goods. Those who believe that health and relational goods (friendships, sentimental relationships) are important drivers of well-being are likely to organize their lives accordingly. They build a rich and varied network of friendships and invest in cheerful sentimental relationships with their partners. It is therefore likely that persons who value relational goods, consume more of them than those who believe that such goods are not important. Since a positive correlation between relational goods and happiness has been demonstrated in different empirical studies (Becchetti et al., 2008, Gui and Stanca, 2010), it is legitimate to expect that those who value relational goods end up reporting higher subjective well-being. On the other hand, in line with our “counterfactual” hypothesis, because they value relational and social aspects, those who more experienced a worsening of these aspects turned out to be less happy. To say it in a sentence: the more you appreciate a good the more you suffer when is gone.

The other three principal components, public goods (quality of health services, quality of other public services and pollution), materialism (job stability, high income, social status) and populism (immigration and loss of national identity) were not significant.

As far as materialism and populism are concerned, the happiness literature suggests some interesting results. In a major meta-analysis, Dittmar et al., (2014) showed a negative association between materialism and well-being with small negative correlations between various measures of materialism and various aspects of well-being. Our definition of materialism was less rigorous than that of most other scholars, but in our sample we did not find any statistically significant relationship between the degree of materialism (as we defined it: aggregation of income, social prestige and a stable job) and the probability of reporting a higher level of happiness. Again, we can interpret this outcome as further confirmation that during a pandemic, materialist aspects of life do not have any effect on happiness or people reconsider own scales of values.

Interestingly, a recent study suggested that unhappier people were more likely to vote for populist parties (Algan et al., 2018). We did not find any statistically significant relationship between the measure of populism that we adopted and reported levels of happiness. Subjects who believed that immigration and loss of national identity were important determinants of happiness did not appear to report a different level of happiness from others. Under lockdown, issues such as immigration and loss of national identity are presumably perceived as less of a threat and therefore did not have a major effect on happiness. This may also be due to the fact that during the pandemic, the media and news were monopolised by pandemic-related issues.

This study was based on a research assumption: “in virus veritas”. Paraphrasing a Latin saying, it assumed that in a hot or aroused state (here due to the threat of coronavirus infection), people are more likely to express their true views and desires and may modify their scale of values. Lockdown was an extraordinary time in Italian life. Faced with an invisible and dangerous enemy, the way people organized their lives changed dramatically. As a consequence, what they really cared about (what makes life worth living) emerged more spontaneously.

Building on this assumption, we exploited the lockdown, imposed in Italy in March 2020 in response to COVID-19, to conduct a web survey into people’s beliefs about the social and personal aspects most important in the pursuit of happiness, and to appraise individual happiness and the impact of the pandemic on some aspects of personal life (income, relationships, health). The main aim was to test whether any relationship existed between beliefs, COVID impacts and reported happiness. Our aim was to determine whether the picture was so different from those recorded in “ordinary” periods in Italy and elsewhere. In other words, we set out to discover whether the unique situation in which Italians were forced to live could give us new information on the relationship between happiness and its determinants.

Our heuristic hypothesis was that the pandemic and the policy response (lockdown) represented a counterfactual situation that allowed us to address the following rhetorical questions: what would happen to individual happiness if people were faced with a threat to their economic, social and relational situation? Would relational aspects still prevail over material possessions in determining individual happiness? Is there any relationship between people’s rescaled beliefs and reported happiness? Since such a situation actually arose in 2020, undermining people’s feeling of security, we exploited the lockdown period to address these questions.

The result is consistent with other aspects unveiled by our study. Religious faith, living in a small village and male gender increased the probability of reporting the highest level of happiness. The first two aspects are directly related to relational/community issues that may have mitigated the impact of the pandemic on relational goods. If this interpretation is correct, the gender effect is also coherent. Women suffered the burden of COVID more than men and their relational ties were probably more affected.

Lastly, our study also showed that people who believe that social capital and relational goods are important determinants of individual happiness tended to have a higher probability of reporting the highest level of happiness. Our results therefore tell us that the pandemic led people to appraise immaterial aspects more than material possessions, and this was a reward in terms of individual happiness. Because social capital is important for happiness, those who most experienced a deterioration in personal relations reported a lower level of happiness. The income effect, although it had the right sign, proved to be not statistically significant, as did being a materialistic type of person.

In conclusion, our study exploited the natural experiment of lockdown to obtain interesting insights into happiness and its determinants in a unique situation. The results are intriguing and add a new perspective to the study of happiness. They somehow confirm what the World Happiness Report (2021) evidenced for the pre-lock-down period. It seems that emotions changed more than did life satisfaction.

The main shortcoming of the study was its small, non-representative statistical sample due to the time constraint and the difficulty of administering questionnaires during lockdown. This prevents any generalizability of findings. Lockdown was a (hopefully) not reproducible event that makes our analysis unique. However, considering the above mentioned shortcomings, it should be taken as such, i.e. an opportunity to address/discuss about some intriguing issues in a special environment/context, aware that, to use Corbin & Strauss (1990: 191), our results and theoretical formulation applies to these situation or circumstances but to no others. The alternative was to renounce and waste the “opportunity” to carry out such “natural experiment”.