Article body

Over the last decades, happiness has received increased attention. Psychologists have been investigating understandings[1] (Biswas-Diener et al., 2009; Kashdan et al., 2008; Keyes and Annas, 2009; Waterman, 2008) and determinants of happiness (Diener and Seligman, 2004). Positive psychology (a branch of eudaimonia) has elaborated on “authentic happiness” (Seligman, 2002) while researchers like Daniel Kahneman (1999) coined “objective happiness,” a modern version of Benthamian hedonia. The Easterlin Paradox[2] has been intensely studied and debated (e.g. Clark et al., 2008; Easterlin, 1974, 1995; Easterlin et al., 2010; Graham, 2009; Stevenson and Wolfers, 2008). Finally, philosophers have been discussing the foundations of various understandings of happiness and the relationship between happiness and well-being (e.g., Annas, 1993; Haybron, 2010; Kraut, 1979; Sumner, 1996).

Public decision-making is affected too. Researchers advocate for national and international measures of happiness (e.g. Diener, 2000; J. F. Helliwell and Barrington-Leigh, 2010; Kahneman and Krueger, 2006; Kahneman et al., 2004a; Krueger, 2009). Political decision-makers have increasingly become receptive to the idea of creating happiness indicators. International organisations lobby for reforming existing indicators or elaborating new ones that include happiness (e.g., United Nations). Happiness (often understood as subjective well-being) is on the agenda of local and central governments too (Saamah et al., 2012, p. 3). Some already have or will soon have their own happiness indexes—including Bhutan, southern Denmark, Canada (unofficially), the city Somerville in United States, United Kingdom, and Japan—while other governments consider developing such tools (Stiglitz et al., 2009).

This increased attention appeals to a normative evaluation because social indicators have deep moral and political underpinnings and implications, in particular when implemented by public institutions.[3] There is a need for an analytical overview of the philosophical and political challenges. Political philosophy and public ethics have not so far proposed an encompassing overview of these challenges.

This article represents such an attempt. It reviews some of the main questions that happiness indicators raise, organized in a taxonomy (see the table below); the criticisms that can be lodged against happiness indicators; and replies that can be opposed to these criticisms. The taxonomy is structured along four categories: The first includes questions about which understanding of happiness should be measured? The second relates to the population that should be monitored: whose happiness to measure? The third gathers methodological questions about the construction of these indicators: how to elaborate such an index? The last category is about the use of indicators by public institutions: how to apply happiness indicators?

By evaluating the construction and use of diverse happiness indicators, this article neither endorses happiness as the ultimate goal of life, the definitive metric of justice, nor assumes that happiness research and some of its political outcomes cannot be challenged at a foundational level (e.g., as being a controversial conception of welfare). In this article I review some philosophical and political issues raised by happiness indicators. The methodology is to take seriously the diversity of indicators that are or could be proposed by researchers and institutions. The point is not to discuss happiness as a philosophical concept or to defend the politics of happiness. It is simply to provide a normative evaluation of happiness measurement when carried on by public institutions.

The originality of this work in public ethics is that it covers the main issues about building happiness indicators, the debates they generate, and the possible replies that could be framed in good faith and with close attention to happiness research. Furthermore, it highlights the importance of the strategy of diversification-hybridization. Indeed, it is shown that most worries about happiness indicators (e.g., moral/epistemological partiality) can be deflected by building composite indicators (i.e., indicators including several understandings of happiness) or by favouring the elaboration of various indicators for capturing the complexity of human happiness. The conclusion recaps challenges that are of interest for normative thinkers, in particular on democratic grounds. It also presses the point that attention from normative thinkers is required even more now that the process of diversification-hybridization is already being discussed in social sciences and is a reality of public decision-making.

Table

Structure of the Argument

Structure of the Argument

-> See the list of tables

1. What to Measure?

When building happiness indicators, the first step is to identify the basis of the indicators, i.e. the type of happiness that is being analyzed. This step is important because it may express underlying political choices and commitments. Thus the first section introduces us to (1.1) the three understandings of happiness available for building indicators, (1.2) the proper characterization of happiness, (1.3) the broad philosophical issues such indicators raise, (1.3) the two-pronged criticism of political or epistemological partiality that could be brought against either hedonia or eudaimonia, and (1.4) the issues proper to indicators based on subjective well-being.

1.1 Which Happiness?

In the literature, there are three main understandings of happiness: hedonia, eudaimonia, and subjective well-being (SWB).[4] Each articulates two interrelated dimensions: a descriptive one (Happiness is X or Y) and a normative one (Happiness as X or Y is good because of A or B). They are interrelated in the sense that the normative value is rooted in the descriptive dimension (e.g., hedonia is good because pleasures or positive emotions are good things for human welfare, eudaimonia is good because flourishing or functioning well is a good thing for human welfare, etc.).

Hedonia is an affair of affects (or sometimes emotions (Diener, 1984; Diener et al., 1999), even if the hedonic nature of emotions might be disputed). An individual is happy to the extent that he or she is “feeling” happy, i.e., that individual is experiencing positive affects. And, overall, a life may be called qualified as “happy” in the hedonic sense if the balance between positive and negative affects is positive. The World Happiness Report defines it as “affective happiness” (Sachs, 2012, p. 6). Authors such as Jeremy Bentham, Francis Edgeworth, and Daniel Kahneman are representative of this understanding.

The second understanding, subjective well-being (SWB hereafter), has three components: positive affects, negative affects, and life satisfaction (Pavot, 2008). Positive and negative affects are hedonic whereas life satisfaction is of a different nature, implying self-assessment (poll respondents judge their whole life or life domains such as work, family, social relations, etc.). The World Happiness Report characterizes this form as “evaluative” (Sachs, 2012, p. 6) because respondents are required “to make a retrospective judgment about how their lives are going overall” (Tiberius, 2006). SWB, especially the life satisfaction component, is widely used for national and international surveys and popular among psychologists (e.g., Ed Diener), heterodox economists (e.g., Richard Easterlin, Bruno Frey) and philosophers (e.g., Wayne Sumner).

The third understanding, eudaimonia, identifies happiness with the development of personal faculties. An individual is happy in the eudaimonic sense to the extent that he or she is virtuous (Aristotle), functions well—rather than simply “feels” good—(Keyes and Annas, 2009), is self-determined (Ryan et al., 2008), or is personally expressive (Waterman, 1993). So “happiness is something like flourishing human living, a kind of living that is active, inclusive of all that has intrinsic value, and complete, meaning lacking in nothing that would make it richer or better” (Nussbaum, 2005, p. 171). Despite notable differences, eudaimonic views share a common idea: an “happy” life consists in the actualization of individuals’ potential that could be related to morality (e.g., virtues in Aristotle), agency (e.g., capabilities), intellectual capacities (like in John Stuart Mill), flourishing, etc. Psychologists (e.g., Carol Ryff, Martin Seligman, Alan Waterman), philosophers (e.g., Julia Annas, Martha Nussbaum) and economists (e.g., Amartya Sen) have discussed this understanding.

A last, important, point: one may deny that eudaimonia is about happiness by claiming, for instance, that eudaimonia is about well-being, assuming that happiness and well-being are two different things. There are several ways of making sense of the distinction. A common way is to postulate that happiness is about psychological states whereas eudaimonia is not. The problem is that there are eudaimonic conceptions that are subjective (Kraut, 1979), or contain a strong psychological dimension (e.g., Waterman’s personal expressiveness, Singer and Ryff’s psychological well-being, Nussbaum’s view on eudaimonia). One may acknowledge this, but maintain the point that eudaimonia is something else than “happiness” (in a sense that often is unclear).

To this, it can be replied that happiness is a continuum ranging from feeling to functioning well, a continuum that expresses the deeper idea that the life is going well according to the individuals themselves. To some extent, this is the complexity that SWB tries to capture by combining affective and evaluative dimensions. Hedonists may contest the label of “true happiness” to eudaimonia and eudaimonists may do the same with hedonia. These controversies express something common in philosophy and beyond: interpretative conflicts about the true nature of values and principles. This article adopts a neutral posture on the true nature of happiness in order to present a structured overview of the challenges faced by all types of indicators that monitor happiness as understood in its diversity defined by many philosophers, psychologists, sociologists, and economists.

1.2 Concepts, Conceptions, and Understandings

But before going any further it is worth to ask a simple but fundamental question: what are hedonia, SWB, and eudaimonia? Are they “forms,” “types,” “kinds,” “dimensions,” “conceptions,” “concepts,” of happiness or something else?

A possibility is that hedonia, eudaimonia and SWB are conceptions of happiness: there would be a concept (happiness) that would receive different interpretations (conceptions). This concept would exist independently of any particular interpretation. A problem with this view is that if hedonia, SWB and eudaimonia expresses a common idea, it might not be a single, unified, concept.

A second possibility is that hedonia, SWB, and eudaimonia are different concepts (Haybron, 2010, p. 31). Hedonia would be a subjective feeling (i.e., a psychological state) while eudaimonia would be an objective manner of being or functioning and SWB the combination of affective and evaluative conditions. A variant is to posit that hedonia is about happiness and eudaimonia about well-being. This view has the merit of simplicity, but this simplicity is misleading. As stated, some eudaimonic conceptions are subjective or rooted in specific psychological states. Also, SWB is not a self-standing concept; it is a construct made of hedonic components (affects) and life satisfaction (individuals’ judgment on their life).

The purpose of this article is not to reach the truth about happiness, but to review public policy issues raised by happiness indicators. We therefore need to find a characterization of hedonia, SWB, and eudaimonia that serves this purpose. In this article I chose the term of understandings. The word has the merit of accommodating the opposite view of hedonia, eudaimonia, and SWB as conceptions and concepts. Its encompassing nature leaves room for a general overview of public policy issues conveyed by happiness indicators. This choice might be tackled as accepting the terminology framed by happiness researchers and any ensuing confusion.

One might also deny any relevance to happiness research as such or consider that social scientists are so confused about the “essence” or true nature of happiness that they end up talking about different things. These criticisms call for two comments.

Firstly, the confusion that would reign in happiness research is sometimes exaggerated. The fact that different authors have different understandings of happiness does not mean that the whole research is crippled by confusion. An analogy is useful: the fact that philosophers mobilize different understandings of equality (e.g., in regard to the metric, to the rule of distribution, etc.) does not imply that equality as an object of philosophical investigation is crippled by confusion.

Secondly, the criticism would be fair if researchers never defined their object or if they disregarded the diverse uses of “happiness”. But, on the one hand, most researchers are clear on their understanding, how it is constructed, how it differs from other understandings. On the other hand, they are also clear on the diversity that exists within their field as illustrated by various debates (Biswas-Diener et al., 2009; Kahneman, 1999; Kahneman et al., 1997; Kashdan et al., 2008; Keyes and Annas, 2009; Pavot, 2008; Waterman, 1993, 2008).

For a moral thinker interested in public policies, the challenge is precisely to offer a structured view of the philosophical and political issues of building indicators of happiness in its diversity. In other words, it is to adopt a pluralistic and neutral view on happiness when dealing with public ethics issues.

1.3 Broad Philosophical Issues

The choice of an understanding by researchers and public institutions is fundamental. Depending on this choice, a given index will emphasize dimensions of well-being at the expense of others (with the limit that no understanding of happiness is exhaustive of welfare), which will affect the feed-back that institutions receive. The choice raises two main challenges: the respect for pluralism and the nature of “true” happiness.

a) Respect for pluralism. Despite significant overlaps, citizens do not share identical moral, religious, or cultural views. Individual opinions diverge on what constitutes the good life, even if this divergence is often overstated. Therefore, institutions need to pick an understanding of happiness that can accommodate different conceptions of the good life. Because hedonia, SWB, and eudaimonia are never purely descriptive when used in the public realm (they are measured, compiled, monitored for political purposes ranging from monitoring to decision-making), the normative charge that indicators are carrying should be compatible with most of the reasonable comprehensive doctrines endorsed by individuals.

Axiological diversity is an issue. But diversity also has a cultural dimension. Understandings of happiness may vary depending on cultural membership (Vazquez and Hervas, 2013, pp. 33-34). People may also weight components (e.g. positive or negative affects) or understandings (e.g., hedonia, eudaimonia) differently (Diener, 2009). Recognizing pluralism when building happiness indicators is an acute issue for any society as well as for international comparisons.

b) Nature of “true” happiness. There is also a question as to which understanding most closely matches “true,” “authentic,” “real,” or “veritable” happiness (if such a thing exists, which is controversial in itself), and not something illusory or delusive.

If we start with hedonia, this understanding is often criticized for giving an unrefined or partial and, as a result, unappealing view of human well-being (Alexandrova, 2005; Haybron, 2010). This is either because a life structured around hedonistic goals is intrinsically undesirable or because hedonia leads to undesirable outcomes such as the pursuit of short-term pleasures at the expense of lasting satisfaction or self-flourishing. There is a further criticism. Hedonia is accused of being an indiscriminate theory that does not (cannot) discriminate pleasures according to their moral content (Nussbaum, 2010). A pleasure will remain a pleasure, no matter if it results from sadistic or antisocial behaviour.

If hedonia is endorsed by the state, through an indicator, as (part of) what has value in life (i.e., as (part of) the good life), lifestyles based on temporary, immediate, or even immoral pleasures might be promoted at the expense of more ambitious and fulfilling life paths. This leads to two broader criticisms, which indiscriminately apply to hedonia and eudaimonia.

According to the neutralist critique, it is objectionable for the public to endorse any comprehensive doctrine, independently of the fact that this doctrine is hedonic or eudaimonic. A partisan critic argues that when institutions adopt hedonic indicators, they endorse an inferior moral view (hedonia) at the expense of a preferable alternative (eudaimonia or SWB).

However, in cases where hedonia is adopted for some indicators without any further public endorsement, it is possible to argue that a moral ideal is not being promoted as such. But the choice of hedonia as the baseline of an index may still distort the feedback received since other important components of well-being are not taken into account.

1.4 Political and Epistemological Partiality

This criticism of partiality, either political—the public commitment to a specific interpretation of the good life—or epistemological—the construction of biased indices—, can also be addressed to eudaimonia.

In regard to political partiality, the objection is that eudaimonia expresses a perfectionist view. If happiness lies in the development of human capacities, which could be understood as the highest intellectual abilities (as expressed by Aristotle and John Stuart Mill), eudaimonia could also be criticized for being biased in favour of a particular conception of the good life (Landes 2013).

Not all eudaimonic understandings are equally vulnerable to this criticism. Objective understandings based on a predefined list of precise items classified in a moral hierarchy are particularly vulnerable, contrary to understandings focusing on general human traits or abilities that are not overly specific. For example, the capability approach emphasizes general human functioning (e.g., a self-sufficient, socially integrated, individual life) rather than specific functioning (Nussbaum, 1992; Sen, 1999). Another example of a not-so-vulnerable understanding is Self-Determination Theory, which grounds eudaimonia as psychological well-being in individual autonomy and intrinsic motivation (Ryan et al., 2008).

As regards epistemological partiality, eudaimonic indexes can be criticized for disregarding the affective aspect of happiness (as well-being): if they are based on eudaimonia alone, the indexes will provide a partial view of the actual well-being of a given population. They will leave out of the picture presumably important elements for living a happy life (e.g., positive feelings such as pleasure or joy).

Both hedonic and eudaimonic supporters may reply by asserting the epistemic and/or moral superiority of either hedonia or eudaimonia. The difficulty with this response is for hedonia supporters to dismiss eudaimonia (and vice-versa) they must prove that the life aspects that the other theory encompasses do not contribute to happiness in any reasonable sense. If it is shown that pleasures/pains or flourishing/functioning are part of widespread understandings of happiness, it becomes difficult to exclude them from the construction of an indicator that attempts to measure the happiness of a population. The reason is that individuals themselves assess their condition by appealing to these understandings.

1.5 Subjective Well-Being and Diversification-Hybridization

One strategy for avoiding this difficulty is to combine elements of hedonia and eudaimonia into a single index. Even if SWB does not include eudaimonic components per se (Ryan and Deci, 2001), it partially adopts this strategy of diversification (by referring to the different understandings that might be separately accounted for in an indicator or a set of indicators) or hybridization (by referring to the creation of a composite indicator drawing on different understandings) (Diener, 2000).[5] Because SWB is made of positive affects, negative affects, and life satisfaction,it contains affective and evaluative aspects. Therefore, an indicator based on SWB is more resistant to the partiality objection (without completely deflecting the criticism). This versatility has contributed to the popularity of SWB among researchers who work on indexes (Diener and Seligman, 2004).

Since SWB tracks both affective and evaluative dimensions, it is less vulnerable to the partiality objection. Yet reservations have been raised about SWB’s capacity to capture adequately culturally diverse understandings of happiness (e.g., the valorization of negative affects and de-valorization of positive affects, the lesser emphasis on the importance of self-evaluation in certain cultures (Diener, 2000)). The problem is that SWB components are measured by requiring individuals to reply to questions (“are you satisfied with your life?,” “are you satisfied with your work/family/social relations?”) or to provide a cardinal evaluation (“how depressed / elated / etc. do you feel?” on a 0 to 3 scale for instance) that can be interpreted differently according to respondents’ cultural backgrounds. This might be a problem for a national index since it leads to aggregate data that does not have the same content (and meaning) in different places and populations.

For international comparisons, the problem is twofold: the same indicator applied to different countries may not measure the same thing because people may interpret the questions differently depending on their culture and, therefore, the answers cannot be compared. Individuals may also value the diverse domains or emotions under evaluation differently. Some data may appear, at first sight, negative or positive, whereas they are not intended as such by respondents, which undermines international comparisons and can render interpretations difficult. For instance, certain negative affects are positively valued in Eastern cultures while certain positive affects are despised (Diener, 2000, p. 39).

Partiality is the main criticism made against happiness indexes based only on either hedonia or eudaimonia. Researchers usually address this objection by using or invoking the potential diversification-hybridization of indicators, by employing SWB measurements or other composite indicators (e.g., the Pemberton Happiness Index, which tracks positive affects, negative affects, general well-being, eudaimonic well-being, hedonic well-being, and social well-being (Vazquez and Hervas, 2013)). In other terms, the combination of different elements, which capture different understandings of happiness, make the partiality criticism less convincing by rendering indicators more encompassing.[6]

Like other composite indicators, SWB may deflect most of the criticisms in relation to partiality. However, they still face a shortcoming. They do not account for functioning, flourishing, or individual capabilities. But it is far from being a fatal defect since other, eudaimonic, components may be added up to a given indicator or used to supplement the information given by SWB. Thus happiness research has internal resources for turning happiness indicators into less partial indicators.

A final criticism regards the supposedly monstrous nature of hybrid-diversified indicators. One may claim that they compile unrelated elements (e.g., affects, self-assessment, flourishing) for creating a chimer. On theoretical grounds, it might be the case: hybrid-diversified indicators might be judged monstrous since they blur “some” boundaries between understandings of happiness. Practically, it is another matter. On political grounds, it is difficult to grasp the monstrous nature of hybrid-diversified indicators and see in it a problem for public policy. Public monitoring and decision-making continuously rely on composite indicators (e.g., UN Human Development Index, OECD Better Life Index). If the goal of indicators is not to say the truth of happiness, but to inform public policies on some, complex and multilayered, aspects of human welfare, then theoretical monstrosity is no defect. To the contrary, it might constitute an advantage. Furthermore, if designers and public users of these indicators are aware of their composite structure and the nature of the various components, it is difficult to see what kind of monstrosity it could be.

2. Whom to Include?

In this section we deal only briefly with the scope of happiness indexes: whom to include? The topic is not unimportant, but it raises issues that go beyond the scope of this paper—i.e., the inclusion of three specific groups—residents/citizens, non-human animals, and future generations. Although this is only a quick overview of this issue, it should be stressed that researchers and public institutions need to identify the population(s) they wish to survey (which implies providing justifications) when they construct and implement an index. For national indicators, the choice has practical and moral implications because it expresses an underlying view about the limits of the moral community, understood as the community whose interests are taken into account even if happiness indicators do not offer a complete and definitive view on welfare and, so, might be completed by other indicators.

2.1 Membership Scope

Should national indicators measure the happiness of citizens, legal residents, or anyone present on the territory regardless of status (Durand and Smith, 2013, p. 120)? This question might seem rhetorical since, in practice, most existing indicators and surveys do not discriminate among people on the basis of their immigration and citizenship situation. But the question is still relevant since institutions may have an interest in monitoring different populations separately. In addition, when international studies claim, for instance, that Swiss are the happiest people in the world (J. Helliwell et al., 2015, p. 26), is it really about Swiss citizens or about residents in Switzerland? The question is both political and epistemological.

Furthermore, public institutions may have difficulties collecting data from various population categories. For instance, illegal refugees usually do not appear on national statistical database (and so are not picked by large-scale surveys that use such databases) or prefer to remain invisible for obvious reasons. Another example is homeless or migrant people who may not appear in happiness indicators because they do not have a constant residence. It is a shortcoming because happiness indicators ought to monitor the well-being of these groups (except if decision-makers consider these groups as socially unimportant and their well-being as morally irrelevant).

On moral grounds, if the well-being of these groups is not monitored, there is a risk of accentuating some of the vulnerabilities they face by reducing their social visibility and, as a consequence, their political weight. In short, is a national index only about the happiness of the nationals, the legal residents, all the residents whoever they are, or only the residents with a fixed address (excluding for instance homeless people)?

Also, should different populations be aggregated in a single index or monitored through different indexes? This question implicitly raises the diversity issue discussed above. If members of cultural minorities understand hedonia or eudaimonia differently or value SWB components differently (e.g., the value of negative vs. positive affects), it will affect the happiness index. In short, national indexes will aggregate material that is not understood in the same way or data that does not refer to the same psychological or objective conditions. Consequently, the salience of happiness indicators will be diminished.

2.2 Generation Scope

This question could be summarized as follows: should future generations be included in a happiness index (Brülde, 2010, p. 572)? The inclusion of non-existing individuals echoes a series of discussions in the field of intergenerational justice that are not addressed in these pages. But, generally speaking, if we agree on the very general principle that happiness of future generations is morally relevant, the issue is how to calculate such happiness and integrate it into an indicator that also tracks the situation of existing generations.

This question contains three aspects. The first aspect is moral. It is about combining the happiness of different generations in a single index—i.e., what relative weight to give to existing and future people. Grouping the well-being of several generations in a single index (or combining several indexes) raises technical issues too, the most important being the appropriate discount rate to be imposed on the (potential) happiness of future generations. The third aspect is epistemological: is happiness the same when we consider actual or future individuals? Are we comparing the same thing? This question will presumably grow in importance as sustainability becomes a primary goal of institutions and as we increasingly rely on happiness and well-being indexes.

2.3 Species Scope

Should non-human animals be included in happiness indexes (Brülde, 2010, p. 572)? It may be argued that, in order to be exhaustive, happiness indexes should include non-human animals. Or the argument can be more moral in the sense that excluding non-human animals cannot be justified on moral grounds.

A quick glance over the literature may suggest that some understandings of happiness are better suited than others for applying to animals. By claiming that happiness is reducible to the presence of pleasure and the absence of pain, classical hedonism, for example, offers an analytical framework that seems readily applicable to non-human animals (especially if no distinction is made in regard to the nature or quality of pleasures (Crisp, 1998, pp. 9-13)).

Another aspect of the issue is that researchers and institutions may favour some understandings of happiness over alternatives because of their ability to include non-human animals in indicators. In that case, it is very likely that qualitative hedonism, SWB (through the life satisfaction component), and “intellectual” eudaimonism (e.g., Aristotelian rationalism or virtue ethics) constitute inferior options in comparison to classical hedonism, subjective emotional state theory, or “naturalistic” eudaimonism.[7] In sum, taking the species issue seriously impacts on the choice of understanding for building happiness indexes.

The scope question encompasses issues that are central to the measurement of happiness. Most of the research does not identify the membership scope: indicators are often presented as recording the happiness of everyone on a given territory without any further analysis of the people included and how they are represented in a given index. Moreover, future generations and non-human animals are excluded from current initiatives for developing happiness indexes. Such exclusions may be seen as morally problematic, which leaves room for debate in which philosophers have a direct stake.

3. How to Collect the Data?

As with any indicator, the construction of happiness indicators raises two separate questions: how to collect the data? (Durand and Smith, 2013, pp. 119-130) and how to build the indicator? The latter covers issues such as the choice between monetary and non-monetary indexes, the rules of aggregation, the weighing of the various components in the case of composite indicators, cardinal vs. ordinal measurement, and so forth. Despite its importance, the issue of how to build indicators is not addressed in these pages. It raises questions of engineering that that are beyond the scope of this article, which focuses on philosophical and political challenges of happiness indicators, not technical ones. So this section focuses on the first set of challenges created by the collection of data.

3.1 Methods of Data Collection

The two methods for collecting data are objective and subjective reporting. The distinction can be formulated as being between directly observable and unobservable material (Frey, 2008, p. 162). Subjective reporting records data that are not directly observable from outside like affects, feelings, emotions, sensations, and so forth. Investigators have no direct access to the raw material (e.g., individuals’ feelings, emotional states, or satisfaction with life). Objective reporting tracks the “hard data” such as functioning, capabilities, or flourishing that can be observed from outside.

The split operates at two levels: the nature of the object under evaluation (pleasure, pain, life satisfaction, flourishing, etc.) and the nature of the assessment (biochemistry, brain waves, hormones, questionnaire, self-report) (Veenhoven, 2002). On the first level, objective assessment focuses on things that exist independently of the subject’s mental states, while the subjective is about psychological states. On the second level, the objective implies external criteria and observers, whereas the subjective is mostly about self-reporting. There is an overlap between the two, which should be acknowledged.

Despite these methodological issues, this section outlines the common understanding and practices of data collection: subjective reporting refers to an indirect access to happiness, whether that happiness is hedonic, eudaimonic, or subjective well-being. Information is obtained through self-reports or evaluations. Objective reporting refers to direct measurement of individuals’ happiness through external observation with minimal mediation (which excludes self-reports and assessments).

3.2 Objective Reporting

Favoured by economists for indicators such as Gross Domestic Product, objective reporting seems a natural fit for eudaimonic understandings of happiness based on an objective list. For example, it is possible to calculate a capability index, inspired by the works of Martha Nussbaum (2010, pp. 110-111), through the external observation of the performance of individuals (e.g., Human Development Index). A possible criticism is that it is not fine-grained enough for capturing the complexity of part of human well-being that relates to happiness, especially the subjective dimension. It does not collect all the information that matters for evaluating happiness—namely, feelings, emotions, affects, moods, and satisfaction.

Against this criticism, two counter-arguments can be made in defence of objective reporting, one negative and one positive. The negative argument is to stress the difficulties of subjective reporting—e.g., the risk of manipulation or inaccuracy.[8] But this reply is convincing only insofar as it could be proven that the lack of subjective elements is less of a serious defect than the lack of reliability of subjective reporting.

The positive argument is that objective reporting can directly access individual subjective elements. Possible methods include brain waves (Layard, 2005) or physiological screening (e.g., cortisol, oxytocin). But one may argue that objective reporting offers only unmediated data, and is therefore only distantly related to what is supposed to be measured (e.g., affects) because hormonal reactions or brain waves are different from the subjective states they are correlated with. As a result, subjective reporting would still be valuable for individual feelings, affects, satisfaction, or moods. In any case, happiness researchers use only marginally objective indicators, so there is no need to elaborate.

3.3 Subjective Reporting

Subjective reporting is the most common method for collecting happiness data. The methodology is to ask individuals about their lives as a whole (life-satisfaction); specific domains like work, family, social relationships, and so on (domain-satisfaction); or particular moments, feelings, and emotions (affects). This method is used by major happiness surveys and indicators, such as the World Values Survey, the Eurobarometer, and Gallup’s Global Barometer of Hope and Happiness. Subjective reporting is the method par excellence, especially for SWB (i.e. positive and negative affects plus individual satisfaction) (Kashdan et al., 2008, pp. 221-222; Ryan and Deci 2001, p. 144).

There are several ways of collecting individuals’ self-reports depending on the sort of happiness that is monitored, each of which has specific issues.

a) Affects. Central in hedonia and present in SWB, positive and negative affects may be measured as experienced or remembered states (Kahneman et al., 1997; Kahneman and Krueger, 2006). Experienced states are measured “on the spot,” at the very moment individuals experience specific activities or situations. Measuring experienced happiness (or utility in Kahneman language, which is derived from the Benthamian view that equates utility with the total aggregate of pleasure and pain) is complex and demanding since individuals must be prompted while they carry on their daily activities.

The Experience Sampling Method (ESM) is a method for measuring experienced mental states (Csikszentmihalyi, 1990). Participants are equipped with electronic devices that regularly require them to enter information on their affective and emotional states (elated, bored, stressed, angry, etc.) and on the kind of activity they are currently undertaking, and other information (e.g., the time when the activity started). The Day Reconstruction Method (DRM) is an alternative method for tracking remembered happiness (utility). Participants have to keep a diary of their daily activities (Kahneman et al., 2004b), report information such as the kind of activities they undertook, the length of each activity, the affects, feelings, or emotions experienced during these activities, and so on.

The choice between experienced and remembered utility is a trade-off between practicability and accuracy. Studies in psychology have shown that individuals make mistakes when they recall past events, especially when they try to recollect mental states and emotions (Gilbert, 2006). As shown by Kahneman et al. (1997), remembered utility substantially diverges from experienced utility (understood as instant utility).[9] When individuals are asked to recall specific events and what their affects were at those moments, they average the peak affect (i.e., the most intense moment) with the affect at the end of the event. The so-called peak-end rule produces hedonic accounts that diverge from experienced utility, especially when understood as the integral of the intensity and length of affects.[10] This alteration is due to the combined effect of judgment and memory.

A psychometric method such as ESM reduces the noise inherent to self-reporting, but it suffers from a practical shortcoming: it is very demanding for the respondents and costly to implement. Proponents willingly acknowledge that this method is not fit for large-scale indicators (Kahneman et al., 2004a, p. 431). The alternative, DRM, does not interrupt individuals in their activities and reduces implementation costs, but re-introduces significant noise due to the time gap between the experience and the individual report. In addition to the effect of the delay, individuals, consciously or not, filter out their experiences, producing distorted accounts.

b) Self-evaluations and judgments. Self-evaluative and judgmental components are usually collected through questionnaires where individuals are asked to situate their happiness (or satisfaction) on a scale. A popular method is the Cantril self-anchoring scale (e.g., Gallup) where individuals are asked to imagine the best possible world and to weigh their actual life against it. Other methods include life satisfaction questions (“On the whole, are you very satisfied, fairly satisfied, not very satisfied or not at all satisfied with the life you lead?”) and happiness questions (“Taking all things together, would you say you are very happy, rather happy, not very happy or not at all happy?”) that are part of numerous indicators like the Eurobarometer, the Gallup World Poll, and the World Values Survey.

While the mediation of individual self-assessment, i.e. a person’s evaluation of his or her own situation, is a problem for monitoring affective states, it is precisely what is tracked by evaluative components: how individuals judge their life as a whole or by domains (e.g. work, family, personal relations). Self-assessments therefore offer a weighted picture of life: when individuals express how satisfied they are with their life, work, family, etc., they elaborate on their objective situation by taking into account subjective elements such as life goals, moral or religious values, etc. By doing so, they produce a weighted judgment, a re-construction of how their life or parts of it is going from their own point of view.

However, self-reports of life satisfaction may be more prone to biases than affects because they are constructed judgments. But techniques are available for making self-reports more robust (Diener, 1994). Also, circumstances under which questionnaires are administered influence responses (Diener, 2000, p. 35; Kahneman et al., 2004a, p. 430). Respondents may have distorted views on how their life is going (Gilbert, 2006). They might also distort their self-reports for various reasons: they may be under the sway of social conventions (e.g., in regard to the expression of personal satisfaction) or have various psychological motives (e.g., to appear in a better light in the eyes of the investigator).

c) Functioning and flourishing. The last category is eudaimonic happiness. Even if objective reporting is important for monitoring eudaimonia, especially for objective list theories, subjective methods are used too. For instance, the Personally Expressive Activities Questionnaire is made up of a series of questions about a given activity that concentrate on factors that contribute to feelings of flourishing, accomplishment, or completeness (Waterman, 2008, p. 236). Another example is Psychological Well-Being: respondents are asked to rate their level of disagreement/agreement in regard to statements covering six eudaimonic domains (self-acceptance, environmental mastery, positive relations, purpose in life, personal growth, autonomy) (Ryff and Keyes, 1995).

In both surveys (as for other eudaimonic monitoring tools), respondents are asked to provide a self-evaluation. As phrased by Julia Annas and Corey Keyes (2009, p. 198) “our measures of well-being reflect individual’s judgments of their functioning in life.” For the Psychological Well-Being, they evaluate their own functioning across several areas generally. For the Personally Expressive Activities Questionnaire, they evaluate their more limited eudaimonic experience during a given activity. The methodology is so close to the one used for self-evaluation and judgments that we will not go any further and will now consider criticisms against subjective reporting.

3.4 Objections

Whether subjective reporting is used for monitoring affects, life satisfaction, or eudaimonia, it needs to reply to two objections.

The first is the adequacy of the measured object to capture happiness: to what extent is the measured thing really happiness? It can be asked whether experienced utility, remembered utility, life-satisfaction, or personal expressiveness reflects true happiness. Eudaimonists may argue that experienced and remembered utilities offer a measure of emotions that has nothing to do with “true” happiness understood as human flourishing or functioning. Hedonists may reply that personal expressiveness, life-satisfaction, and other evaluative components track individual qualities such as a person’s self-reflective capacities or judgment on his or her situation that are not related to happiness.

To be fully convincing, these criticisms ought to prove that happiness indicators based on hedonia, eudaimonia, or SWB (depending on the point of view of the critique) do not overlap with some shared understanding of happiness or folk conception. Since it is reasonable to consider that affective and evaluative components capture different dimensions that a significant number of individuals identify with happiness, it is difficult to argue that to measure experienced utility, remembered utility, life satisfaction, flourishing, or any other hedonic/eudaimonic/SWB concept is to measure something other than happiness from people’s own point of view.

The second objection tackles the reliability of the measurement. Affective states as well as evaluative judgments may (or not) be components of happiness. The quarrel is not about this. The criticism is that researchers use tools that do not offer reliable accounts of these components. At first sight, the objection is supported by psychological findings from happiness research itself. As Kahneman et al. (2004a, p. 430) put it, “the life satisfaction and happiness questions that are used in well-being research request the type of global assessment that people perform poorly on in the psychological laboratory.” Academics such as Daniel Kahneman (2011) and Daniel Gilbert (2006) have built part of their careers on this idea of individual insufficiency. Thus if we consider that individual rationality is bounded, that individuals suffer from cognitive biases, that they adopt (flawed) heuristics, and that they have trouble correctly remembering what made them happy (Gilbert and Wilson, 2000), how can subjective reporting be trusted (Landes 2013)?

It may be argued that two reasons make it difficult to disregard subjective reporting. First, these measures have something to do with happiness even in a loose sense. They record emotions, affects, evaluations, and judgments that matter for individuals who experience them (the folk conception of happiness). When an individual rates a given experience as painful or pleasurable, that individual positively or negatively values an episode of his or her life on affective grounds. When that individual reports being satisfied or dissatisfied, he or she positively or negatively values how his or her life is going. Self-assessment of satisfaction, feelings, etc., even partly mistaken, still expresses something of value: individual experience. In other words, one’s judgments about one’s own affects or satisfaction have some prudential value.

Secondly, personal authority in regard to one’s own welfare forces us to take seriously individual affective or evaluative self-assessments, even if they do not accurately track happiness as a complex of feelings/emotions or an informed judgment on one’s own situation. This second reason is not epistemic. It doesn’t imply or require that individual statements perfectly reflect a person’s well-being with a perfect, or even high, degree of accuracy. The bottom line is that individual statements cannot be overridden qua expressions of an authoritative agent without further qualification (e.g., impaired judgment).[11] It is a question of principle.

The reliability objection is not a knockout argument. First of all, it suggests that more accurate monitoring tools are needed, not that we should give up on measuring happiness. An analogy with inequality is illustrative: because the Gini coefficient is an imperfect snapshot of income inequalities, it does not imply that it is worthless. In that respect, such imperfections argue in favour of improving or supplementing the Gini coefficient, not dropping it. The same goes for happiness indicators. Instruments like ESM or DRM represent improvements in the measurement of hedonic states when compared to retrospective evaluations. In addition to forging better tools, one way of addressing reliability is to diversify the components of the indicators (the diversification/hybridization strategy) and/or to assess happiness measurements through the correlation with other indicators such as brainwaves, blood pressure, physiological monitoring, etc. (Frank, 1988; Kahneman and Krueger, 2006; Kahneman et al., 1997; Layard, 2005).

4. What to Do?

Happiness indexes are assumed to have two qualities, especially those based on subjective reporting. They promote more encompassing views on welfare and they provide a picture of human welfare that is closer to what actually matters for individuals than most common alternatives like GDP. In addition, they are tools for monitoring the evolution of happiness and well-being in a given country, as well as differences across various groups within a country or across countries (as long as the questions are understood in the same manner in the countries compared).

From the perspective of political philosophy, the use of happiness indicators by public institutions raises numerous challenges. So far this article has addressed those related to the identification of a conception of happiness, the choice of a population to be monitored, and the collection of data. There are further issues that are about the institutional use of indicators. Happiness indicators are subject to four criticisms: uselessness, manipulation, capture, and non-neutrality. These four objections derive from the previous methodological points. The purpose of this last section is to present an encompassing view of the objections to attempts to use happiness indicators for public decision-making. This section also gives an overview of the possible replies to these objections.

4.1 The Uselessness Objection

Happiness indicators are worthless for public policy for different reasons.

a) Happiness indicators are useless because the results cannot be compared, especially at the international level. The inability to compare the results may stem from the fact that individuals understand questions about happiness differently [12] or from the fact that they hold different understandings of happiness. As a result, individual responses will be poorly informative:

Many of measures still leave a lot to be desired. Probably the least useful are the surveys that simply ask people to say how ‘happy’ they are. Since people interpret the word ‘happy’ differently, different respondents will effectively be answering different questions.

Haybron, 2013, p. 44

Three replies are possible. The first one is to concede the lack of comparability only for international comparisons. National indicators are still useful for longitudinal evaluations since they monitor a population that, except under dramatic circumstances such as war, massive emigration, or genocide, keeps a relatively stable composition. This continuity gives an opportunity for identifying significant variations in happiness because the biases presumably remain constant through time if the population remains the same.

A second reply is to underline that evaluative components of happiness and SWB are more likely to depend on cultural norms than affective/emotional ones (Vazquez and Hervas, 2013, p. 32). There is still room for international comparisons, at least by using hedonic components such as affects and reducing the weight of evaluative dimensions.

The last counter-argument is to engage with the premise of the argument by pointing out that it is based on too radical a view of pluralism according to which moral or cultural differences are so profound that it will generate insurmountable incommensurability. If it is right that people may understand questions differently or give different weight to different components, it is still possible to compare groups of people from different cultural backgrounds by adjusting or correcting the data or by analyzing these differences. Finally, it should be noted that cultural differences are often exaggerated.

b) Happiness indicators are useless because individuals quickly adapt to changes in life circumstances (Frey, 2008, pp. 164-165). This is the most controversial finding of the literature stemming from the Easterlin Paradox (Easterlin, 1995).[13] Also, a study (now contested) suggested that lottery winners and newly tetraplegics rapidly adapt to their new situations, almost returning to their level of happiness before the lottery gain or the accident that caused the handicap (Brickman et al., 1978). Thus, aspiration and hedonic treadmills will render happiness indicators poor in useful information. Moreover, public policies grounded on such information would be ‘Sisyphus policies’ that pursue impossible social improvements since happiness would be fated to return to its long-term level.

The criticism is based on the premise that individuals fully adapt to changes in their life conditions and concludes that no public policy can have a lasting effect on collective happiness. But the amplitude of individual adaptation has been exaggerated by initial studies (e.g., on tetraplegics and lottery winners): individuals do not completely adjust to life changes, especially when changes are non-pecuniary (Easterlin, 2003). In addition, indicators may still record short-term variations in happiness levels—i.e., before adaptation. Moreover, the different dimensions of happiness (affects, self-assessment, etc.) or understandings (affective, evaluative, etc.) adapt differently depending on the kind of changes in a person’s life circumstances (Kahneman and Krueger, 2006, pp. 9-14). As a result, happiness indexes can still be useful for identifying potential social reforms.

c) Happiness indicators are useless because the pursuit of happiness makes sense only at the individual level. The criticism is that happiness does not pose any particular problem as a personal goal, but does as a political one. There is a practical or/and normative gap that cannot be overcome, between the value of happiness in personal life and its political pursuit (Duncan, 2010, pp. 172-173; Landes 2013).

From a practical point of view, individuals hold so many different understandings of happiness that there cannot be such a thing as public happiness. Consequently, the usefulness of happiness indicators is limited. From a normative point of view, individuals hold so many different understandings of happiness that the state should not enact politics of happiness. Echoing Rawls’s objections to the separateness and sacrifice of the minority (Rawls, 1971, pp. 26-27), the risk is that public happiness will impinge on individuals who hold original, marginal, minority views on happiness. Consequently, it undermines more than the usefulness of happiness indicators: it undermines its moral desirability.

Such criticisms are relevant only insofar as institutions intend to use happiness indexes for maximization purposes (such as Bentham’s felicific calculus). But this is not necessarily the case. As a matter of fact, happiness indicators are often viewed by researchers and decision makers as not having the final say on human welfare. Indicators may be used only as part of more general socioeconomic feedback on the state of the population. They can serve to spot vulnerable groups within the society that could not be identified otherwise (i.e., through traditional socioeconomic indicators) (Stiglitz et al., 2009, pp. 10-11). They can also serve to monitor the evolution of a given population or to compare different groups within a given population. They may be used by entities that are not directly involved in public decision making—e.g., advocacy groups or non-profit organizations. Even if individuals do not share the same conception of happiness, happiness differentials among groups or the evolution through time of the happiness of specific groups may still be a valuable source of information for institutions and various organisations.

To summarize, the pursuit of public happiness is not the only reason institutions want to measure it. They may use indicators as complements of other decision-making tools (as recommended by the Stiglitz-Sen Commission) that assess other dimensions of welfare (some unrelated to happiness) and for the sole sake of having a more finely tuned image of vulnerability. Thus, the objection is a non sequitur: it is not because the pursuit of happiness could only make sense at the individual level that happiness indicators are necessarily useless.

d) The final criticism is that happiness indicators are useless because they try to quantify something that should not be. The criticism is sometimes made as part of a more general point against the contemporary obsession with accounting for any dimension of human life (Jany-Catrice, 2012). The problem with happiness indicators is that they try to quantify segments of human experience that should not be quantified. There is a risk that the happiness determinants (e.g., family, friends, personality traits, etc.) and understandings (hedonia, eudaimonia, SWB) will be reduced to things, objects, and, as a result, happiness will lose what makes it valuable in individual lives.

Against this criticism, it can be said that, first, it is difficult to determine in what sense evaluating dimensions of life and compiling data to form indicators leads to reification. That criticism could be made against all public policy tools that rely on evaluation and aggregation (e.g., unemployment, violence, birth rates). Due to its radical premise and implications, the last criticism may not be addressed here in a way that could be judged satisfactory by those who argue for it. But, there is also not so much to discuss since the objection tackles the justification for having any index at first place.

4.2 The Manipulation Objection

Happiness indicators may be manipulated to serve certain interests (Frey, 2008, pp. 166-167). Subjective reporting, at the core of most initiatives for building indexes, facilitates such manipulation.

Different agents may manipulate happiness indicators: individuals, public institutions, and political parties. Especially when happiness is the focus of public policies (Frey and Gallus, 2013, p. 4206), citizens may alter their self-reports in order to punish the government. Subjective reporting facilitates manipulation because there is no “fundamental” (i.e., easily accessible background facts, the “raw material”) against which to check the collected information. Also, institutions may manipulate indicators by diverse means: legal, legal but morally problematic, and illegal (like increased and/or targeted public spending, propaganda, creative accounting, and so forth) (Frey and Gallus, 2013, pp. 4207-4209). Happiness indicators will be more easily manipulated than other indexes based on objective data such as GDP or HDI (Frey and Gallus, 2012, pp. 103-104). Finally political parties might “cook the facts” by misinterpreting or perverting the data provided by indicators (De Prycker, 2010, p. 595). If this is an issue (and as such problematic, whether they are about happiness or other data), it says nothing about the indicators per se since the manipulation happens ex post, on already gathered and analyzed data.

One counter-argument is to point out that so-called objective indicators can be manipulated too. GDP, inflation, and public debt for instance are, qua objective measures, assumed to be more verifiable and therefore less easy to “cook,” but as a matter of fact they have been repeatedly falsified (e.g., the 2010 Greek public finances scandal). As expressed by Derek Bok (2010, p. 40), “in light of these weaknesses, the results of happiness studies seem, if anything, more reliable than many familiar statistics and other types of evidence that legislators and administration officials routinely use in making policy.” Of course this counter-argument does not make the case for happiness indicators based on subjective reporting. Rather it suggests that the distinction between objective and subjective indicators is not one between indicators that can be manipulated and those that cannot be manipulated. Both types can be bent to specific ends.

A reply to this counter-argument is to emphasize the relative advantage of indicators based on objective reporting in comparison to those based on subjective reporting. The reply agrees that both types of indicators can be manipulated, but affirms that it is easier to twist the latter. Moreover, agents have incentives to do so. Governments have incentives to give more weight to particular happiness indicators and to try to talk down others; respondents can easily alter their self-assessments depending on their intention.

It might be possible to agree with the spirit of this reply but still believe that manipulation can be significantly reduced by various tactics. (While all address governmental manipulation, only the first is effective for preventing the alteration of self-reports by respondents.) First, happiness indicators might incorporate more objective reporting such as brain waves, biochemical activity, and so forth. This data can be used for building indicators that will be teamed up with subjective indicators for producing more encompassing measurements. Or this data can be directly incorporated into existing indicators for building composite indicators.

Secondly, the responsibility for collecting data and building indicators may be delegated to independent organizations, as it is already the case for central banks for inflation (Frey and Gallus, 2012, p. 108) or national statistics institutes. There are several ways for preserving the independence of these institutions. One is to set up constitutional and legal provisions that guarantee independence, such as for the European Central Bank.

Finally, the number of happiness indicators may be multiplied in order to diversify the sources that inform public policies and render manipulation more difficult to undertake or more visible when undertaken. This argument may be supported by a consequentialist argument that emphasizes the virtue of a diversity of opinions for resisting propaganda. In other words, a diversity of happiness indicators will reduce the risk of manipulation, but also will enhance public information and democratic debates. To conclude, the three tactics illustrate the potential of diversification/hybridization measures for meeting the manipulation objection.

4.3 The Capture Objection

Happiness indicators give too much power to experts in the political decision-making process.

If public policies are based on happiness indicators, governance will be captured by experts and will ultimately lead to a “government of the experts.” Such capture will be accentuated by the degree of complexity of happiness indicators.

However, it is difficult to see how happiness indicators are or could be much more complex than other indicators routinely used for political decision-making (e.g., GDP, Human Development Index, Gini coefficient, balance of trade). If the capture of the decision-making process is a legitimate source of worry and if complexity is what grounds this worry, many dimensions of decision making in modern societies must raise identical concerns. Thus this objection is not so much against happiness indicators per se than against indicators as governance tools.

A counter-argument is to underscore that public institutions could provide extensive information on the construction of such indexes. By being completely transparent about the data and methodology used, they may reduce the role of experts as the hermeneutists of happiness indicators. Transparency could then serve as a safeguard against attempts to capture the decision-making process (and incidentally against the manipulation objection). In conclusion, the worry about the capture of democratic processes is legitimate, but it does not justify a rejection of happiness indicators (if it did, opinion polls would also be problematic). To the contrary, it advocates for finding instruments and processes that reduce this risk.

4.4 The Non-Neutrality Objection

Happiness indicators promote specific moral, political, or cultural norms.

This objection is the immediate consequence of the methodological choices imposed by the construction of happiness indexes. Since different understandings of happiness do not account for the same aspects of human life, happiness indicators are bound to offer partial views on human well-being depending on the understanding chosen. Furthermore, public policies based on them express this moral and political partiality. As a result, moral, cultural, religious, or specialist doctrines are favoured at the expense of others, which undermines the ideal of state neutrality in regard to axiological diversity.

a) The first counter-argument is to claim that happiness indicators are neutral, so there is no risk of violating the principle of neutrality. This response is articulated by Ed Diener and Martin Seligman when they write that they “believe that measures of well-being are—and must be—exactly as neutral politically as are economic indicators” before adding that such measures are “descriptive, not prescriptive, and should remain so” (Diener and Seligman, 2004, p. 24).

However, it is unclear in what sense happiness measures are politically neutral “as economic indicators are.” First of all, if Diener and Seligman are right in the sense that SWB measures, for example, do not contain an ought, it is still a far cry from guaranteeing political neutrality, especially politically neutral use. They are other ways to be non-neutral than by explicitly endorsing or promoting political (or moral, religious, cultural) views. As seen above, the choice of a conception of happiness for building an indicator is non-neutral in a double sense: it indicates that the architects of the indicator have chosen an understanding of happiness over alternatives (but without explicitly endorsing it) and it provides a measuring tool focused on specific dimensions or understandings of happiness at the expense of alternatives.

b) A second counter-argument is to challenge the interpretation of the ideal of neutrality as requiring fully neutral political decisions. It may be argued that no one ever seriously defends the principle of completely neutral public policies. Neutrality matters as a question of degree—i.e., that it is preferable to have policies that are as neutral as possible (Dworkin, 1985, p. 191). Therefore, any rebuttal of happiness indicators rooted on their lack of neutrality may rely on too radical and, therefore, too indefensible a view. The crux of the objection is to ask what could be wrong about the fact that happiness indicators may favour specific conceptions of the good life. In other words, non-neutrality in itself is insufficient for proving that there is a moral issue.

As a reply to the second counter-argument, it could be contended that the neutrality issue is still pertinent. The utopian nature of a full-blooded ideal of neutrality may be acknowledged, but there is still the question of how much non-neutrality to accept. Since the idea of public institutions oppressively promoting a particular conception of the good life will (presumably) appear problematic to most citizens, there is a need to define the conditions for acceptable non-neutral decisions. It could be argued that happiness indicators cross this line because they influence public policies (through feedback) in a direction that advantages some individuals to the detriment of others, without a morally valid reason.

c) A third counter-argument is to recognize the lack of neutrality of happiness indicators and to accept the fact that it may be problematic, but to limit the divergence from the ideal through a “strategy of diversification/hybridization,” either internal or external.

Internal diversification is about diversifying the constituents of happiness indicators in order to cover the broadest possible spectrum of happiness understandings through, for instance, the inclusion of positive and negative affects, moods, life-satisfaction, domains-satisfaction, functioning, and so forth (Diener, 2000, pp. 35, 40; Vazquez and Hervas, 2013, p. 39). A further step could be to allow respondents to rate the importance of the diverse constituents of the indicator when their responses are collected.

External diversification can be enabled by, as already mentioned, allowing independent actors to develop their own happiness indexes and taking them into account for political decision-making. As such, the diversity of views on happiness will be accounted for from various perspectives and through various indicators. Consequently, public decision making will be made more pluralistic (Frey and Gallus, 2012, p. 108).

Conclusion

The previous pages have shown that happiness indicators raise issues related to the conception of happiness on which they operate, their internal construction, and the ways in which they are used. In addition, the specific issues raised by happiness indicators have been presented systematically and discussed, and tackled one by one. It should be noted that this article was neither an endorsement of happiness indicators nor a rebuttal of the possibility of measuring happiness. Instead, the purpose was to identify philosophical and political challenges of happiness indicators and to try answering them in good faith—i.e., by taking happiness research seriously.

From a bird’s eye view, three encompassing issues present an immediate interest for political philosophers and normative thinkers in public ethics. I would like to conclude this article by succinctly presenting them.

Firstly, the adequate balance among understandings of happiness within hybridized-diversified indicators opens up research avenues for political philosophers. In short, how to weight hedonic, eudaimonic, or SWB components within a single indicator? Or, how to balance different indicators (tracking hedonic, eudaimonic, or SWB components) within a given decision-making process? An easy answer is to consider that the appropriate balance is different depending on the indicator, context, institution in charge, and political aim. The answer might appear too simplistic, but the devil is in the detail. Arguably one of the tasks of applied political philosophy could be to look into such details to depart from the over-generality and lack of specificity that often characterizes discussions on happiness measurement.

Secondly, there are institutional and democratic issues: how should these indicators be handled in liberal democracies? Various sub-questions emerge. For instance, there is the question of the distribution of the responsibility for administrating these indicators: who should be in charge? In a democracy the choice of the institutions, public and private, that provides the information and necessary feedback for political decision making is never trivial for reasons evoked in this article (e.g., risks of manipulation, capture, involvement of citizens).

A related question is: how should such responsibility be exercised by the institutions (private or public) in charge? In other words, the mandate of these institutions, when they build happiness indicators, needs to be defined. Behind these issues, there is one that is much more fundamental and that relates to the public status of happiness. Public institutions need to be clear about the role played by happiness in the more general framework of quality of life. They ought to articulate the new data that will be provided by these emerging indicators with more classical measurements. They also need to determine the importance they place on happiness as a political goal.

This means that there are further reflections to carry out on happiness as a political value (among other values). This also means that an underlying institutional theory of happiness and well-being indicators is required. Political theorists and normative thinkers may have a stake in contributing (more than they currently do) to the debates about the public value happiness of course, but also about the institutional implementation of such value. For instance, one of the tasks of this institutional twist is probably to determine the adequate degree of transparency of happiness indicators or the value of happiness arguments in specific debates (e.g., on unemployment, climatic change, work environment, social policies).

Last but not least, there are issues of inclusion: how should these indicators be handled in pluralistic liberal democracies? If we consider that democratic struggles have often been about ending the exclusion of particular individuals or groups from the decision-making process, as well as taking into account their interests and welfare, happiness indicators raise issues in this dimension too. The second section (whom to include?) echoes this of course, but not only that. Transparency issues illustrate this challenge of better inclusion of citizens in public decision-making processes. Moreover, it could be argued that due to their subjective nature and their multidimensional aspects (feelings, functioning, life satisfaction, etc.), happiness indicators require the inclusion of all stakeholders (citizens, non-citizens, and animals) much more than traditional indicators like GDP.

In conclusion, if it is accepted that institutions have a responsibility to design public policies that enhance the quality of life of their citizens (partly captured by happiness indicators), and if the pluralistic nature of happiness is acknowledged as a legitimate constraint, political philosophers and normative thinkers have to discuss the diversification-hybridization strategy. If the idea is taken seriously that individuals hold divergent views on the determinants and the “true” nature of happiness, the creation of indexes capable of accounting for such plurality is either morally or strategically necessary. Furthermore, the inclusion of stakeholders appears to be necessary too.

Happiness indicators should be at the core of both public debates and academic research. The need for such a debate is reinforced by three trends. The first is the rising concern about traditional indicators of welfare such as GDP and the harsh criticisms that they are attracting in a situation of environmental and economic crisis. The second is the necessary reforms the world is facing for addressing global issues such as climate change, depletion of natural resources, and anemic economic growth, which implies, if not completely changing, then at least amending political goals, economic structures, and so forth. The last trend is the current momentum enjoyed by happiness measurement (J. Helliwell et al., 2013). As specialists of political design, political philosophers have a “natural” competence for being more present in these debates that they currently are.

For these reasons, as well as the increased interest from psychologists, sociologists, and political theorists, normative thinkers should pay more attention to the issues related to the construction and political use of these new indicators, which includes happiness ones. As stated, there is a “natural” fit for political philosophers and normative thinkers because these issues intersect with themes that have always been utterly important for philosophy like pluralism, democracy, and inclusion, as well as themes that benefit from a growing interest in the profession, like institutional design and social architecture.