Paternalism attitudes and the happiness value of fundamental freedoms

Konrad, Kai A.; Simon, Sven A. Article — Published Version Paternalism attitudes and the happiness value of fundamental freedoms Economics of Governance Provided in Cooperation with: Springer Nature Suggested Citation: Konrad, Kai A.; Simon, Sven A. (2025) : Paternalism attitudes and the happiness value of fundamental freedoms, Economics of Governance, ISSN 1435-8131, Springer, Berlin, Heidelberg, Vol. 26, Iss. 1, pp. 23-56, https://doi.org/10.1007/s10101-024-00322-y This Version is available at: https://hdl.handle.net/10419/319164 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. http://creativecommons.org/licenses/by/4.0/ Vol.:(0123456789) Economics of Governance (2025) 26:23–56 https://doi.org/10.1007/s10101-024-00322-y ORIGINAL PAPER Paternalism attitudes andthehappiness value offundamental freedoms KaiA.Konrad1 · SvenA.Simon1 Received: 31 July 2024 / Accepted: 12 December 2024 / Published online: 4 February 2025 © The Author(s) 2025 Abstract Governmental regulations and prohibitions may effectively reduce self-harming behavior, but might also restrict personal freedom rights. This paper investigates into individuals’ preferences for the trade-off between being protected from taking self-harming actions by the state, and the desire to be free to choose. We introduce the Paternalism Preference Index to measure this preference. Our results indicate considerable heterogeneity in how individuals support, or object to, such governmental prescriptions. More importantly, we show the importance of the paternalism preference in a survey on life satisfaction during the Covid-19 pandemic. We find a reversed relationship between the preference for paternalistic state interventions and the valuation of personal freedoms. Individuals who prefer to have a larger choice set at the expense of committing potentially self-harming actions derive a higher value from personal freedoms in terms of life satisfaction. Keywords Paternalism preferences· Paternalism Preference Index· Governmental regulation· Self-harming behavior· Life satisfaction· Civil liberties JEL Classification H11· H12· C90· K38 We thank Bruno S. Frey, Clemens Hetschko, Ronnie Schöb, Raisa Sherif, Alois Stutzer, Lisa Windsteiger, two anonymous reviewers, participants of the 2021 annual congress of the international institute of public finance, of the 2022 ZEW Public Finance Conference, the ECONtribute Research Workshop 2022, of the first and second edition of the workshop ‘Paternalism: Origins and Consequences’ in Berlin, of workshops in Lutherstadt Wittenberg and Munich, and our colleagues at the Max Planck Institute for helpful comments and suggestions, and Stefan Bruckmeyer, Magnus Haselier, and Agnes Schelling for excellent research assistance. * Sven A. Simon sv[email protected] Kai A. Konrad [email protected] 1 Max Planck Institute forTax Law andPublic Finance, Marstallplatz 1, 80539Munich, Germany 24 K.A.Konrad, S.A.Simon 1 Introduction In most societies some choices on individual conduct are made by the citizens themselves and some other choices are made by the state and on behalf of the individual. Prominent examples are legal restrictions such as the obligation to wear seatbelts in cars or helmets on motorcycles, mandatory health insurance or forced retirement savings. But how much should the state interfere with individuals’ choices in their presumably best interest? Vice versa, how much self-harming behaviors should the state tolerate before choices are restricted? In this paper, we study citizens’ attitudes to the trade-off between individual choice and the protection from welfare-reducing and self-harming behavior. We call this the preference for governmental paternalism. We develop an innovative survey instrument to measure these preferences, the Paternalism Preference Index. In a large-scale representative survey experiment, we document the heterogeneity of paternalism preferences and demonstrate the validity and relevance of our index by examining its relationship with individuals’ valuations of personal freedoms. Whether or not individual preferences for paternalism matter is of basic epistemological interest. Similar to personality indices common in psychology, preferences on paternalism might be an imprint or character trait of the individual. This is also of high policy relevance: Such a trait could determine individual attitudes to public governance and governmental behavior. Policymakers must weigh freedom rights against other basic rights such as life and security of person when interfering with individuals’ private interests. Many modern democracies attach similar importance to a number of rights. The United Nations firmly establishes this in the Universal Declaration of Human Rights, which mentions liberty together with the right to live and security of person in Article 3. The ideal-typical version of these fundamental norms and values is also found in the US Declaration of Independence:‘We hold these truths to be self-evident, that all men are created equal, that they are endowed by their Creator with certain unalienable Rights, that among these are Life, Liberty and the pursuit of Happiness.’ The quote suggests a connection between life, liberty and happiness in general. We design an indicator that captures individual preferences toward governmental paternalism along different dimensions of life. Individuals’ attitudes to such strict interventions might be based on a number of motivations, and might differ based on the specific area and direction in which the government intervenes. However, all interventions have one fundamental trade-off in common. Some individuals may wish to be free to make decisions about their lives, even if those decisions involve situations in which they might end up choosing self-harming actions and making mistakes. Others may wish to be protected by the state from such dangers through appropriate (behavioral) regulations. Related to the definition by Dworkin (1972), we refer to strict paternalistic interventions as policies that (i) restrict an individual’s 25 Paternalism attitudes andthehappiness value offundamental… freedom of choice (ii) without (explicit) consent of the respective individual (iii) primarily in order to make the respective individual better off.1 Using large-scale survey evidence from Germany, we show that the heterogeneity of preferences toward paternalistic governmental interventions is large and ranges from high approval to high disapproval. Then, we use the events of the Covid-19 pandemic: In an attempt to contain the spread of the virus, Germany imposed strong constraints on personal freedoms. Our survey elicits how individuals perceived the constraints on personal freedoms in spring 2020, and how these affected their articulated well-being. To establish the relationship of the paternalism preference and the valuation of personal freedom rights, we test whether the individual paternalism index score is correlated to the reported effects of freedom-reducing infection control measures. Importantly, in the context of our study, it is not crucial for the identification that the motivation for the restrictions has been (exclusively) paternalistic in nature. The hypothesis we test is that paternalism-skeptic individuals value personal freedoms more than paternalism-friendly individuals. A restriction to personal freedoms should lead to larger reduction of life satisfaction for paternalism-skeptic individuals, independent of the motivation of the restrictions. Moreover, paternalism-skeptic individuals should attribute a larger part of their life satisfaction losses tochanges in personal freedom. This may be due to a larger valuation of personal freedoms for instrumental and for intrinsic reasons. Our analysis is related to sets of several areas of scientific interest. First, the growing literature on paternalism. Two papers explore paternalism preferences empirically but from a different research angle. Pedersen etal. (2014) study responses to a set of attitudinal questions that partially overlap with our set of questions. Their research focuses on whether individuals suffer from self-control problems and are sophisticated enough to know about it. Individuals may like governmental choice restrictions and prohibitions to compensate and correct for their own inability to control themselves. Treger (2023) explores the preference for coercive and noncoercive paternalistic interventions in different need domains. She hypothesizes that individuals have a preference for coercive paternalistic policies in the basic needs domain, and a preference for non-coercive paternalistic policies in the higher needs domain. In contrast to us, both studies put a stronger emphasis on libertarian paternalism, which suggests the governmental use of nudges.2 More importantly, we take a consequentialist perspective and show that the paternalism preference is closely connected to an individual’s valuation of personal freedoms. 1 In the empirical analysis, we somewhat relax the third element and require that a policy is at least partially motivated to make the individual better off. Moreover, our measure is not strictly bound to this definition and able to accommodate different elements. For example, Shiffrin (2000) replaces the welfare consideration in condition (iii) by the interference of the paternalist with the individual’s own interests or matters that legitimately lie within the individual’s control. For an overview of conceptual matters, see Dworkin (2013). 2 For a seminal outline of the concept see Thaler and Sunstein (2003). Hertwig and Ryall (2020) provide a critical discussion and analyze the quality of nudging outcomes if the rule maker follows an own, nonbenevolent agenda. 26 K.A.Konrad, S.A.Simon A natural next step is the question what determines the preference for paternalism.3 The literature discusses various reasons for governmentally imposed choice restrictions, consequences of such restrictions, and reasons for why individuals might like or dislike such restrictions. Mill (1859/2014) highlighted that the tradeoff between individual choice and governmental prescriptions can be seen as an issue of private information. He suggested that a government that assumes decision power on individuals’ genuinely private matters is likely to make wrong decisions. The trade-off between individuals’ superior information as regards its idiosyncratic tastes and the government’s economies of scale and scope as regards systematic information aspects is the starting point in Konrad (2024). He constructs an information-asymmetries-based micro foundation of attitudes toward paternalism and highlights possible feedback loops between political decision making, individual consumption tastes and endogenously formed capabilities to make good decisions. The preference for paternalism is determined by the instrumental value of personal freedom here. Complementary is the literature on the intrinsic/non-instrumental valuation of freedoms. Buchanan (2005) postulates that some individuals might like to delegate private decisions, if this comes with a shift of the burden of responsibility for the consequences of these decisions. Related is also Bartling etal. (2014). They suggest that there is an intrinsic value of decision rights and explore a method for the measurement of such decision rights preferences. Also addressing the preference for decisions on private matters, Benz and Frey (2008) and Benz etal. (2004) take a more straightforward behaviorist approach: individuals might have procedural preferences. They may attribute intrinsic value to own decision rights and might value the way in which choice outcomes are determined. Sen (1988) addressed freedom of choice as a genuine desire. He argues that, in an appropriately broadened decision framework, a procedure can be interpreted as a consumption item such that procedural preferences become part of standard economic decision theory. These views and explanations for individuals’ attitudes on the allocation of decision rights are not mutually exclusive and might add or compound. They suggest that there might be heterogeneity of preferences for paternalistic interventions within a given population. Our research addresses this heterogeneity issue. Indeed, individuals differ in their views on specific decision delegation issues, and in how they assess the tradeoff between protection by the government and the freedom of making own choices. These differences are not uniform across issues, but we find a common underlying factor that aggregates and describes individuals’ preferences on this trade-off. Moreover, results of our robustness analysis in different socio-economic subgroups provide some tentative evidence on the instrumental and intrinsic component of the paternalism preference. 3 Complementary, Ambuehl, Bernheim and Ockenfels (2021) suggest, individuals might care about how and why the government should paternalize not themselves, but others. Redistribution goals, normative values, ethical, cultural and religious judgements, altruistic motives based on own perceptions of other peoples capabilities and true desires, and externalities of individual behavior can be drivers for what an individual thinks is appropriate for the government to prescribe to others. 27 Paternalism attitudes andthehappiness value offundamental… Second, there is a major literature that relates freedom of choice to subjective well-being. By using mainly panel data, exploiting variation between countries and variation within countries over time, several studies diagnose a positive wellbeing impact of freedom. Foundational contributions are Inglehart and Klingemann (2000), and Veenhoven (2000). Frey and Stutzer (2000, 2002) use individual survey data from Switzerland, and focus on the role of democracy variables using variation between cantons. More recent studies are Graafland and Compen (2015), Jackson (2017), Nakazato etal. (2017), and Spruk and Kešeljević (2016). Closest to our paper is Hadsell and Jones (2020) that points to individual approval or disapproval of governmental policy as a possible determinant of life satisfaction.4 For an account of the large empirical literature on the relationship between economic freedom and economic performance, see the survey by Hall and Lawson (2014). Our results show that individuals attribute a large share of life satisfaction losses to restrictions in personal freedoms during the first wave of the pandemic in 2020. However, quantitative effects should be taken with a grain of salt because we rely on remembered life satisfaction and self-assessed effects, and restrictions to personal freedoms have been salient. Third, our analysis contributes to the large literature that identifies and measures personality traits. A meta-study by DeNeve and Cooper (1998), for instance, lists 137 personality traits that have been distinguished and analyzed. Some of the analyses study the correlation between such traits and well-being (see, e.g., Steel etal. 2008). Beside the paternalism index, we elicit two well-studied personality traits. The first one is psychological ‘reactance’ (see the foundational work by Brehm 1966 and a review by Miron and Brehm 2006). The reactance trait measures the motivational arousal of the threat of the elimination of one’s freedom and disapproval of being lectured and patronized by others. The second trait is the ‘locus of control’ (Rotter 1990). It can be described as the extent to which a person believes that the course of her life depends largely on her own actions rather than on the actions of others or on chance and luck. We test whether paternalism preferences can be seen as independent of these existing personality concepts, and whether paternalism preferences have more explanatory power for the valuation of personal freedoms than these two related personality traits. The remainder of this paper is structured as follows. Section2 describes our survey experiment. Section3 presents descriptive results on the Paternalism Preference Index and analyses its relationship to the valuation of personal freedoms. Section4 provides a discussion of the survey design and section5 concludes. 4 Their questionnaire generates a composite ‘pro-capitalism’ index for each subject and compares this preference index with an established measure of Economic Freedom in North America. They examine the deviation between these measures as a possible explanatory variable for life satisfaction. Economic freedom may have important instrumental significance, and this might provide a theory explanation for the correlations found. 28 K.A.Konrad, S.A.Simon 2 The survey experiment 2.1 The survey design 2.1.1 Overview The overarching research goal is to measure an individual’s preference on the question of how far governmental regulation should intervene in pre-dominantly private matters with the aim of improving an individual’s welfare, and to what extend it should respect an individual’s decision autonomy even for potentially bad choices instead. To this end, we develop a questionnaire of 11-items, the Paternalism Preference Index. In order to assess the validity of the indicator, we rely on subjective well-being during the Covid-19 pandemic. We use the restrictions of personal freedoms to test the hypothesis that individuals who are skeptical of paternalistic interventions value personal freedoms more in terms of life satisfaction than individuals who approve such interventions. 2.1.2 Implementation The first wave of the survey tests our main hypothesis and focuses on the imposition of restrictions. It was conducted end of June/beginning of July 2020. The sample size was 4,411. A second exploratory wave focuses on the phasing-out of restrictions and was conducted in July 2022. It mirrors the 2020 timing for 2022. The sample size was 1,100. The composition of survey participants was representative of the German population with respect to gender, age (18–69), income, and state of residence. The survey company respondi administered the panel, the distribution of the surveys, and the participant payments. The project received ethics approval by the Ethics Committee of the Department of Economics of LMU Munich (‘Attribution of Happiness Losses,’ date of approval 06/09/2020). Fig. 1 Survey procedure 29 Paternalism attitudes andthehappiness value offundamental… 2.1.3 Procedure Figure1 provides an overview of the survey structure. On the first page, participants received the information that the survey was independent research by Max Planck scientists. After giving consent to participate and elicitation of basic socio-economic characteristics, subjects state their remembered life satisfaction in early 2020 on a 1 to 10 Likert scale.5 ; 6 Participants then answered the Paternalism Preference Index as well as the two established personality traits Psychological Reactance and Locus of Control. Subsequently, subjects were randomly allocated to different information treatments in the first wave of the survey,7 and were then asked for their remembered life satisfaction in April/May 2020 on the same scale as before. In the following, we asked subjects to attribute their change in remembered life satisfaction between April/May and January to changes in six different domains of life. After eliciting additional questions on the personal affectedness by the Covid-19 pandemic, it concluded with an individual’s current life satisfaction at the time of the survey and further socio-economic controls. 2.1.4 Paternalism Preference Index The key explanatory variable is subjects’ views on paternalistic governmental interventions. We measure their preference with a questionnaire that contains 11 items on such policies. Subjects indicate their opinion on each item on a 5-point Likert scale, ranging from complete approval (1) to complete disapproval (5). The full list of items is: 1. Everybody should have the right to do things that can harm himself/herself. 2. Behaviors that endanger people themselves should be banned. (reversed score) 3. Nobody knows better than myself what is good for me. 4. The most important task of the state is to protect me comprehensively from dangers of any kind. (reversed score) 5. I would like to see the state relieve us citizens of more tasks and responsibilities. (reversed score) 6. The state should ban unhealthy food. (reversed score) 7. The state should consistently protect us from drug abuse. (reversed score) 5 Such scales are commonly used, and a prominent example is the World Value Survey. Our question and scale is adapted from Q49 in the 2017–2021 World Values Survey Wave 7 (Haerpfer etal. 2020) that is “All things considered, how satisfied are you with your life as a whole these days? Using this card on which 1 means you are ‘completely dissatisfied’ and 10 means you are ‘completely satisfied’ where would you put your satisfaction with your life as a whole?” with answers on a Likert scale from 1 (‘completely dissatisfied’) to 10 (‘completely satisfied’). 6 Diener etal. (2018) provide an overview on subjective well-being measures. They discuss robustness and reliability of self-reported measures and ascribe good reliability to them. They also elaborate on the effect of potential confounding factors. See also the discussion in the surveys by Frey and Stutzer (2002), Helliwell and Putnam (2004), Clark etal. (2008), Diener (2012), and Clark (2018). 7 The information treatments focused on Covid-19 policies in different European countries and are orthogonal to the research question investigated here. These treatments did not have any significant effects and we abstain from discussing them in what follows. 30 K.A.Konrad, S.A.Simon 8. When and how much I save for retirement I would like to decide entirely on my own. 9. Everyone should decide for himself/herself whether he/she will wear a bicycle helmet. 10. The state is overdoing it with the large number of building regulations. 11. Old vehicles with low energy efficiency should have their operating license revoked in the medium term. (reversed score) The list includes items aiming at the general preference for paternalistically motivated interventions, and items aiming at specific policies in the health, safety, and pension savings domain. Some items focus exclusively on the individual herself. Other items also include the regulation of externalities to some extent, or have more of a normative character and refer to the society as a whole. They have in common that policies restrict the freedom of choice without explicit consent of the respective individual. However, we do not require the strict assumption that the interventions have been exclusively motivated to make the respective individual better off, but rather that it is at least one of the motivations for such policies. Technically, we rely on factor analysis to condense the variance in the response pattern that is due to the paternalism preference into an individual index score. Appendix A.1 provides the technical details on the construction of the index. We also test the robustness of our indicator by using only a subset of items.8 2.1.5 Outcome variables The first outcome variable is the difference in remembered life satisfaction between January and April/May of the respective year. Conceptually, this difference measures participants’ retrospective assessment of how the events of the pandemic have affected their life satisfaction. For simplicity, we will refer to the difference as their change in life satisfaction. In the first wave of the survey (2020), Germany was still largely unaffected by the pandemic in January (Covid-19 stringency index of 6 points by end of January as measured on a scale from 0 (no restrictions) to 100 (strictest response), Mathieu etal. 2020), but in April/May, the stock of infections was considerable, the number of new infections was substantial, and restrictions to civil liberties peaked (Covid-19 stringency index between 60 and 77 points). Hence, the period captures drastic restrictions to personal freedoms. We discuss our life satisfaction measures in detail further in section4.9 Formally, the change in life satisfaction is defined as the difference of remembered life satisfaction in April/May and remembered life satisfaction in January: Δ= Life Satisfaction April/May 2020 − Life Satisfaction January 2020 8 In particular, an estimation of the paternalism preference based on the subset of general items (Q7, Q8, Q10, and Q11) leads to qualitatively similar results compared to the full version with 11-items. 9 In the second wave of the survey (2022), restrictions in January were still widespread (Covid-19 stringency index between 41 and 64 points), but most of them terminated in the beginning of April by law (Covid-19 stringency index between 33 and 18 points). Hence, this period captures the phasing-out of restrictions. Due to habituation effects and previous anticipation of changes in 2022, it is natural to expect stronger effects in the first wave in 2020. 37 Paternalism attitudes andthehappiness value offundamental… the paternalism preference. We refer to the following subsection for an analysis of different socio-economic subgroups. Table 1 Multivariate analysis of the contribution of the domain ‘Personal Freedom’ to the change in life satisfaction Notes: The table presents results of an ordinary least squares (OLS) regression. The dependent variable are the points attributed to the category ‘Personal Freedom’ (min: -5; max: + 5) for explaining the total change in life satisfaction between January 2020 and April/May 2020. ‘Paternalism Preference’ refers to the individual preference for paternalistic governmental interventions, ‘Reactance’ is the individual score based on the Hong Psychological Reactance Scale, and ‘Locus of Control’ is the individual score based on Kovaleva etal. (2012). Control variables include a dummy for female participants (‘Female’), participants‘ age (‘Age’), life satisfaction in January 2020 (‘Life Satisfaction January’), the sum of absolute points distributed to categories (‘Sum of Absolute Points’), and a dummy for participants with low trust in government (‘No Trust Government’). The set ‘Basic Control Variables’ contains the family status, categorized household income, categorized education, and a dummy for households with children. The set ‘Further Control Variables’ contains a dummy for self-employed participants, a dummy for individuals in the health sector, a dummy for individuals in the education sector, a dummy for (university) students, a dummy for unemployed individuals, a dummy for pensioners, a dummy for mothers, a dummy for individuals living in Eastern Germany, and categorized city size. The set ‘Self-Reported Affectedness’ contains a dummy for a perceived high own health threat, a dummy for perceived high health threat of close relatives, a dummy for financially concerned individuals, a dummy for individuals with concern of a job loss, and a dummy for subjects reporting challenges with the care of children and elderly relatives. Robust standard errors clustered at the state level in parentheses, *** p< 0.01, ** p< 0.05, * p< 0.1 Variables (1) (2) (3) (4) (5) (6) Paternalism Preference 0.451*** 0.450*** 0.439*** 0.460*** 0.389*** 0.411*** (0.066) (0.065) (0.065) (0.066) (0.058) (0.056) Female − 0.541*** − 0.468*** − 0.466*** − 0.455*** − 0.492*** − 0.452*** (0.075) (0.059) (0.058) (0.062) (0.055) (0.051) Age − 0.004 − 0.008* − 0.008* − 0.009** − 0.009** − 0.008** (0.003) (0.004) (0.004) (0.004) (0.004) (0.004) Locus of Control − 0.063 − 0.052 − 0.069 − 0.096 (0.044) (0.054) (0.056) (0.061) Reactance − 0.055 − 0.022 0.034 0.048 (0.046) (0.037) (0.039) (0.043) Life Satisfaction January − 0.022 − 0.029 − 0.032 (0.023) (0.023) (0.023) Sum of Absolute Points − 0.027** − 0.019 − 0.010 (0.012) (0.011) (0.012) No Trust Government − 0.799*** − 0.829*** (0.104) (0.102) Constant − 0.653** − 0.532* − 0.543* − 0.144 0.054 0.308 (0.249) (0.297) (0.280) (0.209) (0.211) (0.208) Basic Control Variables YES YES YES YES YES YES Further Control Variables NO YES YES YES YES YES Self-Reported Affectedness NO NO NO NO NO YES Observations 4,411 4,411 4,411 4,411 4,411 4,411 R2 0.043 0.046 0.047 0.052 0.067 0.078 38 K.A.Konrad, S.A.Simon The two baseline specifications in Table1 include a set of basic socio-economic control variables (1) and a set of further socio-economic control variables (2). In both specifications, the preference for paternalistic interventions is significantly and positively correlated to the contribution of the domain ‘Personal Freedom.’ Individuals who look favorable (skeptical) at paternalistic governmental interventions, and who have consequently a positive (negative) value for the ‘Paternalism Preference Index,’ attribute a less (more) negative contribution to the domain ‘Personal Freedom.’ Hence, restrictions in the personal freedoms are related to a smaller loss of life satisfaction for individuals who favor paternalism than for individuals who dislike paternalism. We can rephrase the effect in terms of differences between participants in different deciles of the paternalism index. The marginal effect of moving from the very paternalism friendly ninth decile to the very paternalism skeptic first decile of the Paternalism Preference Index is a difference of 1.03 points allocated to the domain ‘Personal Freedom,’ and from moving from the medium paternalism friendly third quartile to the medium paternalism skeptic first quartile is a difference of 0.51 points. This correlation is robust to a number of robustness specifications. Specification (3) employs the two personality traits ‘Psychological Reactance’ and ‘Locus of Control’ as further control variables. The psychological resistance of (a threat to) the elimination of freedoms and the belief that the course of one’s life largely depends on one’s own actions might be related to the perceived loss in life satisfaction due to restrictions in personal freedoms. However, we find no evidence for a significant correlation of these traits to the contribution of personal freedoms. More importantly, the coefficient on the paternalism preference remains qualitatively and quantitatively unaffected by the inclusion of these personality traits. This is further evidence that the preference for governmental paternalism is a different concept as compared to the two personality traits developed in psychology. Conceptually it measures a ‘social trait’ rather than a ‘psychological trait:’ the relationship between the individual and a paternalistic State, and this measure turns out to be highly relevant. The final set of robustness specifications controls for the baseline level of life satisfaction in January 2020 and the sum of absolute points attributed to different domains of life (specification (4) to (6)). The former variable takes into account that those with greater levels of remembered pre-pandemic life satisfaction may have a larger scope for pandemic effects on different domains of life in general, while the latter variable controls for a potential correlation between paternalism preferences and an individual’s expressiveness in allocating points to domains. This protects against the explanation that paternalism-skeptic individuals generally allocate more points to different domains without actually valuing personal freedoms more than their paternalism-friendly counterparts. Moreover, specification (5) includes a dummy variable for individuals with low trust in government, which may capture a general scepticism to governmental interventions of any sort (both of paternalistic and non-paternalistic nature). Specification (6) adds a set of control variables that capture self-reported concern and subjective affectedness by the Covid-pandemic along different dimensions. These control variables might all be correlated to the perceived loss in life satisfaction due to restrictions in personal freedoms. However, 39 Paternalism attitudes andthehappiness value offundamental… the reason for including these variables is solely to test the robustness of the predictive power of the paternalism preference. Regression results indicate that the results are robust to the additional checks. The coefficient on the paternalism preference is qualitatively and quantitatively unaffected. Result 2 Paternalism preferences are correlated to an individual’s valuation of personal freedoms: The smaller the preference for paternalism, the larger the valuation of personal freedoms. 3.2.3 Robustness results Finally, we presents three additional robustness checks. First, paternalism-skeptic subjects may not value personal freedoms more than their paternalism-friendly counterparts, but may have experienced a larger loss in remembered life satisfaction in general. In turn, they may have attributed more points to the different domains of life. Figure5 (left) shows points attributed to personal freedoms conditional on the change in remembered life satisfaction. The solid black line refers to subjects with a paternalism preference below the participant with the median preference, the gray dashed lines to those subjects with a preference above. For both groups, the larger the loss in life satisfaction, the more negative points are attributed to personal freedoms. Importantly, holding the level of the change in remembered life satisfaction constant, the more paternalism-skeptic subjects attribute more negative points Fig. 5 Robustness results for the valuation of personal freedoms. The left figure shows points attributed to personal freedoms (min: -5 points, max: +5 points) conditional on the reported change in life satisfaction (min: -10 points, max: +10 points). The right figure shows points attributed to personal freedoms conditional on self-reported affectedness during the first wave of the pandemic (min: -10 points, max: +10 points). Individuals to the left feel more affected by the restrictions, individuals to the right feel more affected by the health threat. In both figures, the dashed gray line refers to individuals with a paternalism preference above the median participant, and solid black line refers to individuals with a paternalism preference below the median participant. Error bars indicate mean of respective group ± standard errors 40 K.A.Konrad, S.A.Simon to personal freedoms. Hence, given the change in remembered life satisfaction, a larger part of the change in life satisfaction is due to restrictions in personal freedoms as compared to more paternalism-friendly subjects with the same change in life satisfaction. Second, we control for the subjective binding of restrictions. Life satisfaction losses along personal freedoms might be particularly salient to subjects who felt more constrained by restrictions than by the health threat. In turn, they might attribute more negative points to changes in personal freedoms. At the same time, they may have the perception that the government has taken away consumption possibilities from them. In turn, they may indicate a generally more skeptical attitude toward any governmental intervention. Figure 5 (right) displays points attributed to personal freedoms conditional on subjectively assessed binding of restrictions.12 Those individuals who felt more constrained by restrictions (left side) attribute on average more negative points to personal freedoms than subjects who felt more constrained by the health risks (right side)(Wilcoxon-Mann–Whitney test: p<0.01 ). However, even conditional on the perceived binding of restrictions, paternalism-skeptic individuals (solid black line) attribute more negative points to personal freedoms than the paternalism-friendly half (gray dashed line) does. This holds both for individuals who felt more constrained by the restrictions and for those who felt more constrained by the health threat (Wilcoxon-Mann–Whitney test: p<0.01 , respectively). Hence, it seems unlikely that the relation between the valuation of personal freedoms and the preference for paternalism is determined by the individual saliency of restrictions alone. Third, we investigate the valuation of personal freedoms in different subgroups (parental status, age cohort, income level, employment status) and test the exploratory power of paternalism preferences in these groups. Based on socio-economic characteristics, the pre-pandemic instrumental value of personal freedoms may have differed. For example, consumption opportunities of personal freedoms might be lower for parents. If paternalism preferences and the pre-pandemic consumption value of personal freedoms are correlated, paternalism preferences would mostly reflect differences in the instrumental value of personal freedoms. In this case, differences within subgroups along the paternalism preference should be small and aggregate findings might be driven by specific subgroups. In contrast, if the paternalism preference is not only co-determined by the forgone consumption opportunities of personal freedoms, the paternalism preference should be predictive of the valuation of personal freedoms also within subgroups. The latter is what we observe in the data. While there are pronounced differences between subgroups in the loss in life satisfaction and points attributed to personal freedoms,13 the paternalism 12 After the attribution of the change in life satisfaction to different domains of life, subjects reported the perceived personal affectedness/confinement (i) by the health threat of the Covid-19 (abstracting from restrictions) and (ii) by measures/restrictions implemented against the spread of the virus (abstracting from the health threat). Both variables ranged from 1 (‘not constrained at all’) to 10 (‘very constrained’). Subjects are ordered according to the difference in both scales. 13 The greatest loss in life satisfaction is reported by mothers ( − 1.39 points), and the least loss is reported by unemployed ( − 0.59 points). The most negative points are attributed to personal freedoms by students ( − 1.44 points), the least negative points by unemployed ( − 1.10 points). 41 Paternalism attitudes andthehappiness value offundamental… preference (as determined in the complete population) is predictive of the valuation of personal freedoms within all subgroups considered. By illustration, mothers with a paternalism preference below the median participant attribute significantly more negative points to personal freedoms than mothers with a paternalism preference above the median participant (Wilcoxon-Mann–Whitney test: − 1.79 points vs. − 0.89 points, N=302 vs. N=372 , p<0.01 ). The same holds for fathers ( − 1.29 points vs − 0.23 points, N=207 vs. N=288 , p<0.01 ), young people ( − 1.69 points vs. − 0.96 points, N=453 vs. N=477 , p<0.01 ), retired people ( − 1.43 points vs. − 0.92 points, N=311 vs. N=333 , p=0.01 ), people with low income ( − 1.61 points vs. − 0.84 points, N=479 vs. N=590 , p<0.01 ), people with high income ( − 1.64 points vs. − 0.82 points, N=700 vs. N=599 , p<0.01 ), and unemployed individuals ( − 1.54 points vs. − 0.81 points, N=68 vs. N=106 , p=0.05 ). These consistent within-subgroup differences might be tentative evidence for an intrinsic component of paternalism preferences. Specifically, paternalism-skeptic individuals may have a valuation of personal freedoms beyond its mere consumption value. However, the finding may also come from remaining differences in the instrumental value of personal freedoms in the respective subgroups. 3.3 Second wave: exploratory analysis oflifted restrictions The second wave of the survey in July 2022 addressed the relationship between the phasing-out of restrictions to personal freedoms and the paternalism preference. The timing of events in the second wave is a mirror image of the events of the first wave. The beginning of 2022 was still characterized by substantial restrictions while restrictions were phased out in April/May 2022. The analysis of the second wave is exploratory in nature and we did not postulate a hypothesis similar to the first wave. Between January 2022 and April/May 2022, remembered life satisfaction changed only by 0.04 points, on average. This might reflect that changes in the late stage of the pandemic were predictable, more gradual, and may have taken more time to affect life satisfaction, whereas the first wave of the Covid-19 pandemic changed life circumstances drastically, immediately, and unexpectedly. The domain ‘Personal Freedom’ still contributed most to the increase in life satisfaction, but accounted only for 34 percent of total positive points attributed to different domains of life. Moreover, there is an asymmetry in the second wave along the paternalism dimension. Remembered life satisfaction for paternalism-skeptic and paternalismfriendly subjects did not differ in January 2020 (7.26 vs. 7.21 points, WilcoxonMann–Whitney test: p=0.57 ). In contrast, remembered life satisfaction in January 2022 was different in both groups (6.45 vs. 6.82 points, Wilcoxon-Mann–Whitney test: p=0.01 ). Figure6 illustrates the effect on target measures for the second wave of the survey in 2022. Similar to Fig.4 (left),Fig.6 (left) displays results for a median division of outcome variables along the preference for paternalism. The change in life satisfaction is small and not significantly different in both groups ( − 0.03 versus 0.11, Wilcoxon-Mann–Whitney test: p=0.42 ). Opposite to the results for the imposition of restrictions, the paternalism-friendly half reacts stronger to the phasing-out 42 K.A.Konrad, S.A.Simon of restrictions along the dimension of personal freedoms. Paternalism-friendly individuals attribute a larger part of their increase in life satisfaction to personal freedoms than the paternalism-skeptic individual do (0.62 versus 1.18 points, Wilcoxon-Mann–Whitney test: p<0.01 ). Note, however, that a comparison of points attributed to personal freedoms between both waves is not meaningful due to the small change of life satisfaction in the second wave. Points attributed in the second wave correspond to a much smaller effect in terms of life satisfaction as compared to points attributed in the first wave. Figure6 (right) presents the effect on target variables along deciles of the preference for paternalism. Opposite to the first wave, there is no monotonic relationship between the paternalism preference and the target variables. The change in life satisfaction fluctuates around 0 when moving from the most paternalism-skeptic first decile to the least paternalism-skeptic tenth decile. Moreover, the effect of personal freedoms for the change in life satisfaction is positive but stable for paternalismskeptic individuals and those close to the median preference (first decile to seventh decile of paternalism preference). In contrast, the paternalism-friendly individuals (eighth decile to tenth decile) attribute more points to personal freedoms. The asymmetry along the paternalism preference compared to the first wave may allow for two interpretations. First, effects of lifted restrictions might take longer to materialize for paternalism-skeptic subjects. Second, paternalism-skeptic individuals might attribute not only an instrumental value to personal freedoms but may rather value personal freedoms intrinsically. In turn, the knowledge and experience that personal freedoms can be removed might lead to a persistent effect even when freedoms are restored. Fig. 6 The effect of paternalism preferences on the two key outcome variables in 2022. The left figure shows the change in life satisfaction (light gray) and the points attributed to the domain ‘Personal Freedom’ (dark gray) by division along the median preference for paternalism. For each comparison, the left (right) bar refers to the subgroup that is more skeptical (welcoming) of paternalistic interventions than the median participant. Error bars indicate mean ± standard errors. The right figure shows the change in life satisfaction (upper figure) and the points attributed to personal freedoms (lower figure) along the deciles of the paternalism preference 43 Paternalism attitudes andthehappiness value offundamental… 4 Discussion ofthesurvey design Two points are noteworthy on our survey design and require a discussion. First, we rely on remembered instead of actual life satisfaction. A considerable literature discusses remembered utility versus experienced/actual utility (see, e.g., Kahneman and Riis, 2005; Kahnemann, 2011; Mogilner and Norton, 2019). Acknowledging the differences between the two concepts, we compare studies that measured actual life satisfaction during the first wave of the Covid-19 pandemic to remembered values in our study. Closest to us, Windsteiger etal. (2022) used sample characteristics and elicitation methods similar to ours. They assessed actual values of life satisfaction for April/May 2020 and remembered values for January 2020. We do not find any major difference between their actual and our remembered value of life satisfaction for April/May 2020 (6.15 vs. 6.14, Wilcoxon-Mann–Whitney test: p=0.87 ). We also find only a minor difference between the remembered value for January 2020 in both studies (7.31 vs. 7.23, Wilcoxon-Mann–Whitney test: p=0.09 ). Further evidence comes from Bittmann (2022), who uses actual life satisfaction data from the German National Educational Panel Study (Blossfeld etal. 2011). The pre-pandemic-to-pandemic drop in life satisfaction between 2018/2019 and May/June 2020 by an average of 1.00 points (on a scale 0-10) is qualitatively and quantitatively comparable to the drop by 1.10 points (on a scale 1-10) in our study. Schmidtke etal. (2024) confirm the qualitative reduction in life satisfaction in an event study with somewhat smaller quantitative effects. Finally, the drop in life satisfaction from 7.14 points in 2019 to 6.74 points in 2020 as elicited in the Glücksatlas (SKL Glücksatlas 2022) is also qualitatively in line with our data, and we find similar results regarding most affected subgroups (parents and particular mothers). In summary, studies using actual/current life satisfaction find qualitatively similar results, while differences in the absolute size of the effects are not of major interest for our study. We only focus on whether a respondent’s paternalism preferences cause a qualitative difference in the size of the drop in life satisfaction. Second, the Paternalism Preference Index does not include questions directly related to the Covid-19 pandemic. The timing of the questionnaire at the beginning of the survey makes it also unlikely that subjects were primed toward the pandemic restrictions before. This should limit potential reverse effects of the developments during the pandemic on the preference for paternalism. Moreover, it should also limit the concern that participants report a larger loss in life satisfaction/attribute more negative points to personal freedoms and state a skeptical attitude toward paternalistic governmental intervention in order to express a consistent disapproval of the Covid-19 restrictions. Even though restrictions have been more salient in 2020, we do not find a difference in the paternalism preference between the first wave of the survey in 2020 and the second wave of the survey in 2022 (section3.1). This is in line with Schatz and Petersen (2023), who show that attitudes toward governmental prohibitions in the private domain remain stable for the period 2014 – 2022, including the period before and the period during the Covid-19 pandemic. Finally, controlling for the perceived subjective binding of restrictions in our robustness analysis (section3.2.3) and controlling for individuals’ attitude toward the Covid-19 44 K.A.Konrad, S.A.Simon restrictions with one of our control questions does not change our key result on the relationship of the paternalism preference to the valuation of personal freedoms. Hence, it seems unlikely that the preference for paternalistic intervention changed as a reaction to the pandemic and related governmental actions or that results are driven by subjects wishing to express a consistent disapproval of restrictions. 5 Conclusion In this paper, we develop a new index that measures the preference for governmental laws and regulations to ban potentially adverse or self-harming behavior in the private domain. The index addresses the degree to which an individual is willing to delegate personal decisions to the government that she could also very well take for herself. Examples of past interventions may include regulations to reduce traffic casualties, such as the obligation to wear seatbelts. We find that the index has considerable explanatory power for how individuals assessed the freedom-restricting measures that the government took to contain the pandemic. We show that the costs of restrictions to personal freedoms during the early months of the pandemic depend on individuals’ general preferences for governmental paternalism. This speaks to the larger question whether paternalism preferences are useful as an explanatory variable for the impact of governmental behavior. Appendix A.1 Construction ofindices Appendix A.1 provides additional results on the construction of the Paternalism Preference Index, psychological reactance and the locus of control. As discussed in section2, both psychological reactance and the locus of control have some conceptually related elements to the Paternalism Preference Index. Hence, we elicted both concepts to test their relation to the Paternalism Preference Index. Moroever, we test their correlation to an individual’s valuation of personal freedoms. Construction ofthepaternalism preference index For the Paternalism Preference Index, we compose a set of 11 statements (see section2). We use the respondents’ agreement or disagreement with these statements to measure the position of the individual participant with regard to the trade-off between the desire to choose and decide what risks to take, and the desire to be protected by the state from dangers arising from one’s own behavior. For each statement, individuals have to indicate whether they fully agree (score = 1), rather agree (score = 2), neither agree nor disagree (score = 3), rather disagree (score = 4), or fully disagree (score = 5). The score for reversed items is converted by six minus the original score. Individual answers are condensed into a single index value by factor analysis. For robustness, we also compute an index value based on the aggregate score of all items. Both index values are highly correlated ( corr =0.93 ). The result 45 Paternalism attitudes andthehappiness value offundamental… regarding the construction of the Paternalism Preference Index refer to pooled data from the first wave and second wave, but are robust to the anaylsis of both waves independently. Table 2 shows the results of the factor analysis for the Paternalism Preference Index. We find one factor with an eigenvalue greater than 1 and retain this factor according to the eigenvalue criterion. While factorloadings on all items have the expected sign, we find six items with a loading of 0.5 or higher. The common factor explains a somewhat lower share of the variance of item 3 (knowledge of things that are good for oneself), item 8 (choice of pension savings), and item 10 (building regulations). Cronbach’s alpha is at 0.70, and iterative elimination of items with factor loadings below 0.5 would only lead to a marginal increase of the scale reliability to 0.76. We then use the estimated weights of the statements for this factor to calculate each individual index scores. Construction ofthereactance index In our survey, we used the 14-item version by Hong and Page (1989) to elicit psychological reactance. We include this measure for two reasons. First, individuals who react strongly on (the threat of) the elimination of their freedom in general might also be particular skeptical toward freedom-restricting paternalistic interventions. Secondly, individuals scoring high on psychological reactance may have suffered particularly from the restrictions to personal freedoms in the course of the Covid-19 pandemic. Table3 shows results of the factor analysis for the reactance index. We find one factor with an eigenvalue considerably greater than 1, and one factor with an eigenvalue close to 1. We retain only one factor according to the previous Table 2 Eigenvalues (upper part) and factorloadings (lower part) for the Paternalism Preference Index Factor1 Factor2 Factor3 Factor4 Factor5 Eigenvalues 2.32578 0.95673 0.27883 0.07647 0.05746 Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 Item 1 0.5674 0.2362 − 0.2224 − 0.0928 − 0.0250 Item 2 0.7499 − 0.0433 − 0.0278 − 0.0808 − 0.0441 Item 3 0.1395 0.5082 0.1061 − 0.0714 − 0.0343 Item 4 0.6014 − 0.2533 0.0102 − 0.0108 0.1289 Item 5 0.5060 − 0.1827 0.2448 − 0.0665 0.0838 Item 6 0.5167 − 0.1019 0.0752 0.0846 − 0.1282 Item 7 0.5567 − 0.1501 − 0.1523 0.0643 − 0.0362 Item 8 0.0220 0.4599 0.1553 − 0.0693 − 0.0047 Item 9 0.2599 0.3589 − 0.1899 0.0837 0.0600 Item 10 0.1727 0.4001 0.0002 0.1199 0.0871 Item 11 0.3599 0.1001 0.2613 0.1192 − 0.0335 46 K.A.Konrad, S.A.Simon literature (e.g. Windsteiger et al. 2022). Factorloadings on all items have the expected sign and range between 0.45 and 0.65, and Cronbach’s alpha is 0.87. Construction ofthelocus ofcontrol To measure the locus of control, we rely on the four-item version by Kovaleva etal. (2012). The intuition to include this measure is that the degree to which participants perceive the course of their life to be an outcome of their own actions and choices might be correlated to the preference for paternalistic governmental interventions. Morever, participants with a high locus of control might have had a larger reduction in life satisfaction due to the restrictions of their personal freedoms. Table4 shows the results of the factor analysis for the locus of control. We find one factor with an eigenvalue greater than 1 and retain this factor according to the eigenvalue criterion. Factorloadings on all items have the expected sign and range between 0.45 and 0.60, and Cronbach’s alpha is at 0.64. Effect ofpsychological reactance andlocus ofcontrol onkey variables Finally, we test the predictive outcome of reactance and the locus of control on the two key outcome variables, the change in remembered life satisfaction and the points attributed to personal freedoms. Table 3 Eigenvalues (upper part) and factorloadings (lower part) for Reactance Index Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 Factor 6 Eigenvalues 4.5784 1.0484 0.1882 0.1512 0.0595 0.0210 Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 Factor 6 Item 1 0.6557 − 0.2176 0.1674 0.0134 − 0.0023 0.0116 Item 2 0.4990 − 0.2367 0.1110 0.0666 − 0.0088 0.0738 Item 3 0.6336 − 0.4342 0.0984 − 0.0160 − 0.0152 0.0119 Item 4 0.5114 0.3246 0.1142 0.0493 − 0.0292 0.0093 Item 5 0.5696 − 0.0849 − 0.1836 0.1782 0.0690 − 0.0003 Item 6 0.5667 0.3727 0.1274 0.1228 0.0140 − 0.0116 Item 7 0.4815 0.2502 − 0.1183 0.1801 − 0.0396 0.0094 Item 8 0.4928 0.2890 0.0868 − 0.0413 − 0.0274 − 0.0590 Item 9 0.6589 − 0.3848 − 0.1320 0.0064 0.0581 − 0.0307 Item 10 0.5378 0.1968 0.0431 − 0.1140 0.1450 − 0.0274 Item 11 0.5590 0.2221 − 0.0671 − 0.1482 0.0799 0.0562 Item 12 0.5680 0.2511 − 0.1595 − 0.0928 − 0.0977 0.0198 Item 13 0.6445 − 0.0554 − 0.0665 − 0.1372 − 0.0825 0.0183 Item 14 0.5851 − 0.2380 − 0.0075 − 0.0230 − 0.0623 − 0.0768 53 Paternalism attitudes andthehappiness value offundamental… life satisfaction. Figure10 (right) refers to the second wave of the survey in 2022. A comparison to Fig. 9 shows that the structured estimation approach identifies changes to personal freedoms (the phasing-out of restrictions) as the only domain of life with a relevant influence on life satisfaction. However, in comparison to the imposition of restrictions, the effect of the phasing-out of restrictions has a much smaller effect. Effect ofpaternalism preference onkey outcome variables Finally, Fig.11 displays results for a median division of outcome variables along the preference for paternalism. For the first wave of the survey in 2020, results are line with the main analysis in section3.2.1. Participants with a paternalism preference below the median, i.e., those who tend to be negative toward the governmental interventions, have a more negative contribution of restrictions to personal freedoms to their total change in life satisfaction than the paternalism-friendly half has ( − 0.53 versus − 0.35 points, Wilcoxon-Mann–Whitney test: p<0.01 ). For the second wave of the survey in 2022, results of the structured estimation approach provide a somewhat different result along the paternalism preference as compared to unstructured approach with points attributed to personal freedom (see section3.3). There is no significant difference in the contribution of personal freedoms for the change in life satisfaction between paternalism-skeptic individuals and paternalism-friendly individuals (0.15 versus − 0.09 points, Wilcoxon-Mann–Whitney test: p=0.88 ). This is not surprising because the structured estimation approach accounts for the small change in life satisfaction in both groups and corrects for an individual’s responsiveness when allocating points to the different domains of life. Fig. 10 Contribution of changes to different domains of life to the total change in remembered life satisfaction in units of life satisfaction. Figure shows the contribution of different domains of life to the total change in life satisfaction in units of life satisfaction. Left figure refers to the first wave of the survey in 2020 (compare to Fig.8) and right figure refers to the second wave of the survey in 2022 (compare to Fig.9). Error bars indicate mean ± standard errors 54 K.A.Konrad, S.A.Simon Acknowledgements We thank Bruno S. Frey, Clemens Hetschko, Ronnie Schöb, Raisa Sherif, Alois Stutzer, Lisa Windsteiger, two anonymous reviewers, participants of the 2021 annual congress of the international institute of public finance, of the 2022 ZEW Public Finance Conference, the ECONtribute Research Workshop 2022, of the first and second edition of the workshop ‘Paternalism: Origins and Consequences’ in Berlin, of workshops in Lutherstadt Wittenberg and Munich, and our colleagues at the Max Planck Institute for helpful comments and suggestions, and Stefan Bruckmeyer, Magnus Haselier, and Agnes Schelling for excellent research assistance. Author contributions Conceptualization: KAK and SAS, Survey design: KAK and SAS, Survey implementation: SAS, Analysis and graphic work: SAS, Writing: KAK and SAS, Resources: KAK. Funding Open Access funding enabled and organized by Projekt DEAL. Funding for this project came from the general budget of the Max Planck Institute of Tax Law and Public Finance. Data availability Data can be made available in the data repository “keeper” with a DOI upon publication of the manuscript. Declarations Conflict of interest The authors are not aware of any conflict of interest that are relevant to the content of this article. Ethical approval The project received ethics approval by the Ethics Committee of the Department of Economics of LMU Munich (‘Attribution of Happiness Losses,’ date of approval 06/09/2020). Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission Fig. 11 The effect of paternalism preferences on the two key outcome variables. Figure shows the change in life satisfaction (light gray) and the contribution of the domain ‘Personal Freedom’ in units of life satisfaction (dark gray) by a division along the median preference for paternalism. For each comparison, the left (right) bar refers to the subgroup that is more skeptical (welcoming) of paternalistic interventions than the median participant. Error bars indicate mean ± standard errors 55 Paternalism attitudes andthehappiness value offundamental… directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/ licenses/by/4.0/. References Ambuehl S, Bernheim BD, Ockenfels A (2021) What motivates paternalism? An experimental study, American Economic Review 111(3):787–830 Bartling B, Fehr E, Herz H (2014) The intrinsic value of decision rights. Econometrica 82(6):2005–2039 Benz M, Frey BS (2008) The value of doing what you like: Evidence from the self-employed in 23 countries. J Economic Behav Org 68(3–4):445–455 Benz M, Frey BS, Stutzer A (2004) Introducing procedural utility: Not only what, but also how matters. J Inst Theor Econ 160(3):377–401 Brehm JW (1966) A theory of psychological reactance. Academic Press, New York Bittmann F (2022) How trust makes a difference: The impact of the first wave of the Covid-19 pandemic on life satisfaction in Germany. Appl Res Qual Life 17:1389–1405 Blossfeld HP, von Maurice J, Schneider T (2011) The National Educational Panel Study: need, main features, and research potential. Z Erzieh 14:5–17 Buchanan JM (2005) Afraid to be free: Dependency as desideratum. Public Choice 124:19–31 Clark AE (2018) Four decades of the economics of happiness: Where next? Rev Income Wealth 64(2):245–269 Clark AE, Frijters P, Shields MA (2008) Relative income, happiness, and utility: An explanation for the Easterlin Paradox and other puzzles. J Econ Literature 46(1):95–144 DeNeve KM, Cooper H (1998) The happy personality: A meta analysis of 137 personality traits and subjective well-being. Psychol Bull 124(2):197–229 Diener E, Oishi S, Tay L (2018) Advances in subjective well-being research. Nat Human Behav 2(4):253–260 Diener E (2012) New findings and future directions for subjective well-being research. Am Psychol 67(8):590–597 Dworkin G (1972) Paternalism. Monist 56(1):64–84 Dworkin G (2013) Defining Paternalism. In: Coons C, Weber M (eds) Paternalism: Theory and practice. Cambridge University Press, Cambridge, pp 25–38 Frey BS, Stutzer A (2000) Happiness prospers in democracy. J Happiness Stud 1:79–102 Frey BS, Stutzer A (2002) What can economists learn from happiness research? J Economic Literature 40(2):402–435 Graafland J, Compen B (2015) Economic freedom and life satisfaction: Mediation by income per capita and generalized trust. J Happiness Stud 16(3):789–810 Hadsell L, Jones AT (2020) The company you keep: Satisfaction with life, economic freedom, and preference-policy mismatch. J Comp Econ 48(3):642–657 Hall JC, Lawson RA (2014) Economic freedom of the world: An accounting of the literature. Contemp Econ Policy 32(1):1–19 Haerpfer C, Inglehart R, Moreno A, Welzel C, Kizilova K, Diez-Medrano J, Lagos M, Norris P, Ponarin E, Puranen B etal (eds.), (2020) World Values Survey: Round Seven - Country-Pooled Datafile, Madrid, Spain & Vienna, Austria, JD Systems Institute & WVSA Secretariat. https:// doi. org/ 10. 14281/ 18241.1, Master Survey Questionnaire, http:// www. world value ssurv ey. org/ WVSDo cumen tatio nWV7. jsp (accessed: January 26, 2021) Helliwell JF, Putnam RD (2004) The social context of well-being, Philosophical. Trans Royal Soc London, Series B: Biol Sci 359(1449):1435–1446 Hertwig R, Ryall MD (2020) Nudge versus boost: Agency dynamics under libertarian paternalism. Economic J 130(629):1384–1415 Hetschko C, Knabe A, Schöb R (2019) Looking back in anger? Retirement and unemployment scarring, Demography 56:1105–1129 Hong S, Page S (1989) A psychological reactance: Development, factor structure and reliability. Psychol Rep 64(3):1323–1326 56 K.A.Konrad, S.A.Simon Inglehart R, Klingemann H-D (2000) Genes, culture, democracy, and happiness. In: Diener E, Suh EM (eds) Culture and Subjective Wellbeing. MIT press, Cambridge, MA USA, pp 165–183 Jackson J (2017) Free to be happy: Economic freedom and happiness in US States. J Happiness Stud 18(4):1207–1229 Konrad KA (2024) The political economy of paternalism. Public Choice 201:61–81 Kovaleva A, Beierlein C, Kemper CJ, Rammstedt B (2012) Eine Kurzskala zur Messung von Kontrollüberzeugung: Die Skala internale-externale-Kontrollüberzeugung-4 (IE-4), GESIS-working papers 2012_19 GESIS Köln Mathieu E, Ritchie H, Rodés-Guirao L, Appel C, Giattino C, Hasell J, Macdonald B, Dattani S, Beltekian D, Ortiz-Ospina E, Roser M (2020) Coronavirus pandemic (COVID-19). Our World Data 5:947–953 Mill JS (1859/2014) On Liberty, with an introduction by W.L. Courtney, Enhanced Ebooks Miron AM, Brehm JW (2006) Reactance theory - 40 years later. Zeitschrift für Sozialpsychologie 37(1):9–18 Nakazato N, Nakashima K, Morinaga Y (2017) The importance of freedom in the East and the West over time: A meta-analytic study of predictors of well-being. Soc Indic Res 130(1):371–388 De Neve J-E, Ward G, De Keulenaer F, Van Landeghem B, Kavetsos G, Norton MI (2018) The asymmetric experience of positive and negative economic growth: Global evidence using subjective wellbeing data. Rev Econ Stat 100(2):362–375 Pedersen SK, Koch AK, Nafziger J (2014) Who wants paternalism? Bull Econ Res 66:147–166 Rotter JB (1990) Internal versus external control of reinforcement: A case history of a variable. Am Psychol 45(4):489–493 Schatz R, Petersen T (2023) Freiheit: Die Mehrzahl der Deutschen fühlt sich eingeschränkt. InnoVatio Verlag, Zürich Schmidtke J, Hetschko C, Schöb R, Stephan G, Eid M, Lawes M (2024) Does worker well-being adapt to a pandemic? An event study based on high-frequency panel data. Rev Income Wealth 70(3):840–861 Sen A (1988) Freedom of choice: Concept and content. Eur Econ Rev 32:269–294 SKL Glücksatlas (2022) Retrieved 27.02.2023 from: https:// www. sklgluec ksatl as. de/ artik el/ deuts chlan dsgluec ksniv eauerholtsichnurleicht. html Shiffrin SV (2000) Paternalism, unconscionability doctrin, and accomodation. Philosophy Public Affairs 29(3):205–250 Spruk R, Kešeljević A (2016) Institutional origins of subjective well-being: Estimating the effects of economic freedom on national happiness. J Happiness Stud 17(2):659–712 Steel P, Schmidt J, Shultz J (2008) Refining the relationship between personality and subjective wellbeing. Psychol Bull 134(1):138–161 United Nations (2020) Covid-19 and Human Rights: We are all in this together, April 2020, Retrieved 15.01.2021 from: https:// www. un. org/ en/ uncoron aviruscommu nicat ionsteam/ wearealltoget herhumanrightsandcovid19respo nseand Thaler RH, Sunstein CR (2003) Libertarian paternalism. American Economic Rev 93(2):175–179 Treger C (2023) When do people accept government paternalism? Theory Exp Evidence, Regulation Governance 17:195–214 Veenhoven R (2000) Freedom and happiness. A comparative study in 44 nations in the early 1990’s’. In: Diener E, Suh EM (eds) Culture and Subjective Wellbeing. MIT press, Cambridge, MA USA, pp 165–183 Windsteiger L, Konrad KA, Ahlheim M (2022) Curtailment of civil liberties and subjective life satisfaction. J Happiness Stud 23:2157–2170 Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.