Work on the other aspects of subjective well-being needs to be balanced with the need for clarity in communication with those outside the research community. While a synthetic indicator that combines different dimensions of subjective well-being is probably not a strategy that is well grounded in current understanding of the different constructs, there is often some demand for combining information in summary form. Effort should be made to ensure that these measures, if they are to gain traction in a policy setting, are reasonably easy to understand and compare.
Related to this question is an issue that studies on subjective well-being share with other topics in this volume: how to think about, and communicate, what constitutes a meaningful change or difference in the measure considered. Especially in large samples, a very small difference may be statistically significant but not very important for policy-making. This can be particularly problematic when considering differences on a scale that has no inherent meaning. One option is to phrase differences in terms of some calibration (i.e., a difference of magnitude x corresponds to the increase in subjective well-being associated with an increase in personal income of y, although much more research is needed to establish a universal unit of measurement). It should also be noted that the original work on the day reconstruction method presented differences in experiential well-being across groups by aligning those differences with the affect associated with typical daily activities, providing an informative ruler for interpreting affective differences.
Applications to Policy
Experimentation and Innovation
The application of subjective well-being to policy is at an early stage, and there is still much to learn. At this point, there is sufficient understanding of methodological issues and consensus on the best way to address most of those issues to cautiously move forward. Experimental initiatives will, in turn, generate new questions and more progress on the methodological issues. However, many policy applications will have to wait until a sufficiently large and long cross-country data set has been built up, which will take time. In the meantime, there are a variety of ways that progress can be made.
Subjective well-being can complement existing policy analysis methods, and we recommend taking steps to consider all three dimensions of subjective well-being in order to obtain a complete picture. However, we realize that this recommendation is much less specific than we would like it to be, and the reason for this follows. Evaluative measures of subjective well-being (life satisfaction, Cantril ladder, Diener scale) seem to us to have a conceptual advantage over the eudaemonic and experiential subjective well-being measures in that they target people’s summary evaluations of their current lives, which at face value appears more consistent with their choices and therefore with the economic concept of “utility.”
Experiential measures may be less ideal because of concerns about true adaptation to changing environmental situations (see Sen’s “happy peasant” arguments), though experiential measures do capture an essential aspect of well-being—how one feels. Eudaemonia addresses broader meaning and aspirations, which are undoubtedly an important aspect of life, yet it appears less directly applicable as a measure of utility. An example of this approach tracked evaluative and experiential well-being for people who migrated from Tonga to New Zealand, and showed how these measures illuminated their complex psychological transition (Stillman et al., 2015).
There are drawbacks, though, with exclusively adopting evaluative well-being as the primary measure for policy. First, as mentioned above, we believe that experience is an essential part of subjective well-being that should not be omitted. Second, as reviewed in the methodological sections of this chapter, evaluative measures are prone to being perturbed and, possibly, biased by a number of irrelevant factors, which raises questions about recommending their use for policy-making—particularly if the data are to be used as a proxy for utility, rather than as one of a variety of subjective well-being measures that might inform policy decisions.
Additional work should be done on cost-benefit analysis, to understand whether and how results from analyses based on subjective well-being valuation can complement existing methods. Early efforts have produced some extreme results, potentially due to income measurement issues (i.e., if the estimated partial effect of income on subjective well-being is small, and does not represent the “true” value of income, scaling other effects by this coefficient will lead to implausibly large monetary effects); and because models and theories were not sufficiently developed to allow sound interpretation. In addition, some other measures are particularly amenable to links with subjective well-being research—for example, the Quality-Adjusted Life Year (QALY) and the Well-Being-Adjusted Life Year (WELBY).
Some fairly low-cost initiatives would be to routinely collect, and report on, subjective well-being indicators in program evaluation questionnaires, and to routinely add subjective well-being questions to questionnaires such as labor force surveys or surveys carried out in schools. In the European context this will be done using the EU-SILC as a vehicle, as two to three indicators (including life satisfaction) will be collected on a yearly basis. As shown in the sidebar on page 213 listing articles released in 2015, data on subjective well-being is an important outcome not only in itself, but also as an input or driver of other outcomes of interest, and can help researchers provide a richer data analysis.
Researchers, in turn, need to spur applications to policy, in part to demonstrate to NSOs that investment in subjective well-being data is worthwhile. Much of the current research on subjective well-being is difficult to apply to policy, even experimentally. A greater focus on policy applications in the literature (e.g., on policy-amenable drivers of subjective well-being) would be helpful. Coordination among researchers, policy-makers, and NSOs may be very valuable, a role that the OECD may be well placed to play.
Cautions in Using Subjective Well-Being Data for Policy
Well-meaning but naive policy changes may make people worse off due to the complex inter-relationships between choices, prices, and heterogeneity in subjective well-being, as well as general equilibrium effects. Airport noise is one example: people live near airports for a reason, typically because of lower housing prices, or because they do not care too much about noise, and well-meaning policy (e.g., to reduce traffic at night) could actually make them worse off because of the effect of lower noise in raising house prices. (This example is taken from a personal communication with Angus Deaton.) We need to understand why people live where they do, and build models forecasting how location could change in response to policy changes. Hedonic models of sorting are well established in labor economics (e.g., Rosen, 1986) and urban economics (e.g., Roback, 1982), and this type of work can be extended to subjective well-being.
A less ambitious but still important goal for policy is simply to provide information on subjective well-being and let people and businesses use it as they see fit. To use a term from David Halpern, “de-shrouding” subjective well-being means giving the public information on the correlates of subjective well-being—informing them, for example, that priests, or people without children, or people who live in Denmark are happier than others (clearly, some of this has already occurred in the news media). To some extent, this is simply providing information, which might in principle be useful to someone considering a move to Copenhagen.
However, such information may also be misleading, partly because of the considerable challenges in identifying causality in these studies. The types of people who become priests are very special, and the standard battery of variables used to control for differences between groups (e.g., sex, age, educational background, income) is unlikely to fully control for the difference between people who become priests and those who select other occupations. In this regard, subjective well-being data are no different from data on average income: while it is informative to know that doctors earn more than the average worker, it is the case that many individuals do not have the training, aptitude, or temperament t
o work as doctors. Therefore, to continue with the subjective well-being example, entering the clergy may not yield the expected gains in subjective well-being that were anticipated on the basis of the de-shrouded well-being averages.
In addition, results from observational studies are averages, even if they are averages within groups, and as such may not apply to a given individual. So the application of subjective well-being data must be done in a considered way, given that the potential for unintended consequences are far from academic.
Conclusions
Up to this point, we have not made recommendations for how subjective well-being measures could be used in policy applications, i.e., recommendations beyond the generic suggestion echoed from the 2015 NAS report for the use of both evaluative and experiential subjective well-being measures. While we continue to agree that this is a reasonable approach, here we describe our concerns about this position and lay out recommendations taking those concerns into account.
The choice of the subjective well-being measure that will be used to inform policy should be directed by a theory or model of whatever phenomenon is under consideration, which should direct the subjective well-being construct that best serves the model and policy objectives. As discussed throughout this chapter, explicit depiction of the potential pathways by which subjective well-being influences or is influenced by other variables is paramount for properly specifying the measures selected, for study design, for structuring analyses, and for allowing appropriate interpretation of the results. Without such considerations, investigators are prone to arrive at incorrect, and possibly counterproductive, conclusions about how subjective well-being is impacted by a specific policy.
In the light of the considerations made in the preceding sections, we would like to conclude by providing a limited number of recommendations that could guide research, measurement efforts, and policy application of subjective well-being data in the future.
1. Continue regular, frequent, and standardized collection of subjective well-being data by NSOs. Use the OECD Guidelines to create a standardized evidence base, and aim for re-evaluation of guidelines in the future, once a sufficient evidence base is established.
2. Ensure that these data are collected in a way that allows estimation of the joint distribution of subjective well-being with other variables, and that the other variables (in particular, income) are well measured.
3. Focus on subjective well-being measures beyond life evaluation, and examine the relationship between different aspects of subjective well-being.
4. Continue to collect information on time use and experiential well-being, and intensify efforts to collect such data in low-income countries.
5. Focus efforts to resolve methodological issues on systematic inter-personal differences in response styles, which are not amenable to solution through standardized questionnaires.
6. Develop theories and build models of how different types of subjective well-being function as predictors and outcomes, and how they relate to the other variables one is considering; and develop models of people’s sorting based on preferences and policy changes.
7. Add subjective well-being measures as outcomes in studies of randomized experiments and natural experiments to help identify causal mechanisms.
Notes
1. Albeit often looking at longer-term aspects of affective experience, such as feelings and emotions in the last two weeks, which can confound evaluative and experiential well-being.
2. http://worldhappiness.report/.
3. A remaining concern is that, even though context effects can be reduced or eliminated through good survey design, the notion that subjective well-being measures are particularly vulnerable to them would imply that the underlying construct is not stable. This is, however, difficult to test empirically and is controversial. That said, substantial evidence on the validity of life evaluation measures, and their consistent relationship to objective factors, suggests that people can and do provide meaningful responses to these evaluative questions.
References
Aassve, A., B. Arpino, and N. Balbo (2016), “It takes two to tango: Couples’ happiness and childbearing,” European Journal of Population, Vol. 32(3), pp. 339–354, https://pdfs.semanticscholar.org/5f7f/cf43dcfcbca04e4d8cc90ab078f3190c4661.pdf.
Aghababaei, N. and A. Błachnio (2015), “Well-being and the dark triad,” Personality and Individual Differences, Vol. 86, pp. 365–368, https://doi.org/10.1016/j.paid.2015.06.043.
Beegle, K. et al. (2011), “Methods of household consumption measurement through surveys: Experimental results from Tanzania,” Journal of Development Economics, Vol. 98(1), pp. 3–18.
Benjamin, D.J. et al. (2014), “Beyond happiness and satisfaction: Toward well-being indices based on stated preference,” American Economic Review, Vol. 104, No. 9, pp. 2698–2735.
Berlin, M. and N. Kaunitz (2015), “Beyond income: The importance for life satisfaction of having access to a cash margin,” Journal of Happiness Studies, Vol. 16(6), pp. 1557–1573, https://ideas.repec.org/a/spr/jhappi/v16y2015i6p1557-1573.html.
Blanchflower, D.G. and A.J. Oswald (2008), “Is well-being U-shaped over the life cycle?,” Social Science & Medicine, Vol. 66, pp. 1733–1749.
Boylan, J.M. and C.D. Ryff (2015), “Psychological well-being and metabolic syndrome: Findings from the midlife in the United States national sample,” Psychosomatic Medicine, Vol. 77(5), pp. 548–558.
Brewster, C.E. (2015), “Churchmanship and personal happiness: A study among rural Anglican clergy,” Rural Theology: International, Ecumenical and Interdisciplinary Perspectives, Vol. 13(2), pp. 124–134, www.tandfonline.com/doi/abs/10.1179/Z.00000000050.
Broderick, J.E. et al. (2016), “A qualitative study of themes considered in life satisfaction ratings in the U.S.,” unpublished manuscript, University of Southern California.
Brown, T.T. (2015), “The subjective well-being method of valuation: An application to general health status,” Health Services Research Journal, Vol. 50(6), pp. 1996–2018, http://www.ncbi.nlm.nih.gov/pubmed/25762183.
Christodoulou, C., S. Schneider, and A.A. Stone (2014), “Validation of a brief yesterday measure of hedonic well-being and daily activities: Comparison with the day reconstruction method,” Social Indicators Research, Vol. 115, pp. 907–917.
Clark, A.E., C. d’Ambrosio, and S. Ghislandi (2016), “Adaptation to poverty in long-run panel data,” Review of Economics and Statistics, Vol. 98(3), pp. 591–600.
Clark, A.E., S. Flèche, and C. Senik (2015), “Economic growth evens out happiness: Evidence from six surveys,” Review of Income and Wealth, Vol. 62(3), pp. 405–419.
Clark, A.E. and Y. Georgellis (2013), “Back to baseline in Britain: Adaptation in the BHPS,” Economica, Vol. 80, pp. 496–512.
Clark, A.E. et al. (2008), “Lags and leads in life satisfaction: A test of the baseline hypothesis,” Economic Journal, Vol. 118, pp. F222–F243.
Coffey, J.K., M.T. Warren, and A.W. Gottfried (2015), “Does infant happiness forecast adult life satisfaction? Examining subjective well-being in the first quarter century of life,” Journal of Happiness Studies, Vol. 16(6), pp. 1401–1421, https://link.springer.com/article/10.1007%2Fs10902-014-9556-x.
Cramm, J.M. and A.P. Nieboer (2015), “Social cohesion and belonging predict the well-being of community-dwelling older people,” BMC Geriatrics, Vol. 15(1), p. 30, http://www.researchgate.net/publication/275051964_Social_cohesion_and_belonging_predict_the_well-being_of_community-dwelling_older_people.
Crane, M. et al. (2016), “Correcting bias in self-rated quality of life: An application of anchoring vignettes and ordinal regression models to better understand QoL differences across commuting modes,” Quality of Life Research, Vol. 25(2), pp. 257–266.
Cuellar, I., E. Bastida, and S.M. Braccio (2004), “Residency in the United States, subjective well-being, and depression in an older Mexican-origin sample,” Journal of Aging and Health, Vol. 16(4), pp. 447–466, https://doi.org/10.1177/0898264304265764.
> Cummins, R.A. (2001), “The subjective well-being of people caring for a family member with a severe disability at home: A review,” Journal of Intellectual and Developmental Disability, Vol. 26(1), pp. 83–100.
Cummins, R.A. et al. (2015), “A demonstration of set-points for subjective wellbeing,” Journal of Happiness Studies, Vol. 15(1), pp. 183–206, http://dx.doi.org/10.1007/-013-9444-9.
D’Addio, A. et al. (2014), “Using a quasi-natural experiment to identify the effects of birth-related leave policies on subjective well-being in Europe,” OECD Journal: Economic Studies, Vol. 2013/1, http://dx.doi.org/10.1787/_studies-2013-5k3tvtg6fvmq.
Deaton, A. (2012), “The financial crisis and the well-being of Americans,” Oxford Economic Papers, Vol. 64(1), pp. 1–26.
Deaton, A. (2008), “Income, health and wellbeing around the world: Evidence from the Gallup World Poll,” Journal of Economic Perspectives, Vol. 22(2), pp. 53–72, https://dx.doi.org/10.1257%2Fjep.22.2.53.
Deaton, A. and A.A. Stone (2016), “Understanding context effects for a measure of life evaluation: How responses matter,” Oxford Economic Papers, Vol. 68(4), pp. 861–870, https://academic.oup.com/oep/article-lookup/doi/10.1093/oep/gpw022.
De Neve, J-E. et al. (2015), “The asymmetric experience of positive and negative economic growth: Global evidence using subjective well-being data,” Review of Economics and Statistics, Vol. 100(2), pp. 362–375.
Devins, G.M. et al. (2015), “Distancing, self-esteem, and subjective well-being in head and neck cancer,” Psycho-Oncology, Vol. 24(11), pp. 1506–1513, https://doi.org/10.1002/pon.3760.
Devoto, F. et al. (2012), “Happiness on tap: Piped water adoption in urban Morocco,” American Economic Journal: Economic Policy, pp. 68–99.
Diener, E. (1984), “Subjective well-being,” Psychological Bulletin, Vol. 95(3), https://ssrn.com/abstract=2162125.
Di Tella, R. et al. (2010), “Happiness adaptation to income and to status in an individual panel,” Journal of Economic Behavior & Organization, Vol. 76(3), pp. 834–852.
For Good Measure Page 31