Informing policy—or at least the potential to do so—is a critical criterion for deciding whether it is worth the time and cost of measuring experienced well-being (ExWB) in national flagship population surveys (for example, the American Community Survey and Current Population Survey in the United States or the UK’s Annual Population Survey) or in more focused domain-specific surveys, such as the Health and Retirement Study, the English Longitudinal Survey of Ageing, and various crime, health, or housing and neighborhood surveys. If the relevance or appropriateness of ExWB as an instrument for decision making, policy evaluation, or monitoring purposes cannot be established, then the case for government-supported data collection becomes difficult. An overarching question is whether self-reported ExWB metrics add analytic content above and beyond the existing dashboard of statistics—e.g., those based on income and health data—more traditionally used to measure well-being. In other words, to what extent do results of subjective well-being (SWB) research go beyond the realm of the interesting and thought provoking (which has already been established) to a point that they might refocus policy or even directly inform it? This chapter identifies a number of policy areas for which ExWB measures show promise.
A number of recent studies support the validity of SWB concepts and data when applied to policy-relevant social science research. Kahneman and Deaton (2010) and Stevenson and Wolfers (2013) used data collected in the Gallup-Healthways Well-Being Index to estimate the impact of income and income-normalized effects on evaluative well-being (life evaluation) and ExWB; understanding this relationship could be a consideration
in tax and social program policies. Oswald and Wu (2009) used data from the Behavioral Risk Factor Surveillance System to rank U.S. states based on hedonic analyses of regional variation in such factors as precipitation, temperature, sunshine, environmental greenness, commuting time, air quality, and local taxes. Diener and Chan (2010) argue that people’s emotional states causally affect their health and longevity, concluding that the data are compelling, though “not beyond a reasonable doubt.”
From longitudinal prospective studies to experimental mood inductions where physiological outcomes are assessed, the data have shown strong associations indicating that high positive and low negative emotions are likely beneficial to health and longevity. Recent work on well-being and cardiovascular disease finds a comparatively strong relationship between people’s emotional states and the behaviors that affect the risk for cardiovascular disease; research on the relationship between evaluative well-being measures and health is more mixed (Boehm and Kubzansky, 2012).1 A line of research (e.g., Steptoe et al., 2005) has established that SWB measures relate in a predictable manner to physiological measures, such as cortisol levels and resistance to infection.
Developing more robust and comparable measures of people’s SWB can also play an important role for decisions aimed at improving the living and working conditions of different population groups, including children or older adults. These measures hold the promise of predicting later outcomes and well-being for children associated with different custodial arrangements or of providing evidence about the relative impact of different factors (such as health status, employment status, transportation and mobility, and social isolation) that prevent older people from living in conditions of greater autonomy. Such measures—many of which should be based on longitudinal data—may shed light on the importance of people’s appreciation of their own health (beyond objective measures of their physical functioning) for the quality of various dimensions of their lives. Given the emphasis in the United States, as elsewhere, on enhancing people’s physical and mental health—beyond disease prevention—information on SWB, including ExWB, can play an important role in guiding policies and delivering higher-quality services. Another issue is how to measure the effects of painful but needed policies, such as austerity, that produce short-term pain but long-term gain.
The unique policy value of ExWB measures may lie not in assessing how income does or does not relate to an aggregate-level tracking of emo-
1 Similarly, Cohen et al. (2003) examined how Positive Emotional Style predicts resistance to illness. The authors controlled for other social and cognitive factors associated with Positive Emotional Style and compared resistance to rhinovirus or influenza virus of a group characterized by being happy, lively, and calm with a group characterized as anxious, hostile, and depressed. They found significantly different rates of symptom reporting and concluded that Positive Emotional Style may play a more important role in health than previously thought.
tional states but in discovering many actionable relationships that otherwise escape attention: commuting patterns, accessibility of child care, exercise, interaction and connectedness with neighbors or friends, understanding impacts of corruption, presence of neighborhood amenities and other city planning issues, divorce and child custody2 laws, and the like. Many potential applications rely on analyses of ExWB measures that are tied to time-use and activity data. These targeted areas can be (and have been) improved at many levels, from company policies that improve well-being—and possibly, in turn, improve productivity and lower absenteeism—to community or regional planning policies. ExWB measures seem most relevant and useful for policies that involve weighing costs and benefits when there are nonmarket or not easily quantifiable elements involved—for instance, government consideration of spending to redirect an airport flight path to reduce noise pollution, funding alternative medical care treatments when more is at stake than maximizing life expectancy, or selecting between alternative recreational and other uses of environmental resources.
The possibility of using aggregate-level SWB statistics for broad population monitoring purposes has also been raised. The UK’s Office for National Statistics (ONS) has expressed the view that multidimensional measures of the progress of society are needed, focusing on a “triple bottom line”: economy, social, and environment and sustainability. ONS states that “overall monitoring of progress” is one possible goal of SWB data. To the extent that this becomes useful, there is certainly consensus (see OECD, 2013; or Office for National Statistics, 2011) that SWB measures should be viewed as one set in the much broader array of indicators through which populations are monitored and policies informed.3 Statistics capturing trends in a population’s health, poverty and income distribution, home production, and environmental degradation are all crucial, as are SWB measures,
2 Child custody and child care discussions raise the issue of whether the SWB of children should be tracked, an issue not addressed in this report. Pediatric SWB measures are being developed as part of the Patient Reported Outcome Measurement Information System, which is designed to produce numeric values indicating patients’ state of well-being or suffering and their ability or lack of ability to function. See http://www.nihpromis.org/default [October 2013].
3 No one is seriously discussing replacing other monitoring statistics with an SWB catch-all. The National Income and Product Accounts (NIPA), for example, have proven to be extraordinarily valuable historically, and the core concept is powerful and useful to preserve in something close to its current form. The same can be said of other health, economic, and social “headline” statistics. That said, all measures have limitations and appropriate use constraints. The gross domestic product measure derived from the NIPA is only one piece of evidence among many used for evaluating economic progress and performance. The shortcomings of focusing only on market transactions and measuring their impact in terms of market prices have been well documented (e.g., National Research Council, 2005; Stiglitz et al., 2009). Effective social and economic policies require much more.
which may serve to connect the patchwork of social science statistics that, together, create a portrayal of where a people are as a society and where that society is heading. Thus, an SWB “account” or set of indicators would supplement other key social and economic statistics. Sir Gus O’Donnell, chairing a commission on how well-being data can be used by the UK central government, has communicated an urgency in moving toward greater use of SWB for use in policy making—“if you treasure it, measure it.” He has outlined a number of policy areas—encouraging altruism and volunteering, community spending decisions, and carbon reduction, to name a few—where he believes SWB data could be used to effect positive changes.
These important goals notwithstanding, the panel does not expect SWB (experienced or evaluative) to produce a single number on the state of the nation or to replace established statistics, such as gross national product (GDP),4 unemployment rate, or vital statistics. SWB is multidimensional—perhaps more so than measures of market output, unemployment, or mortality rates; there is no comprehensive single measure of happiness or of suffering. ExWB, in particular, does not establish any sort of overall measure of social well-being, but its measurement is proving useful when applied to specific questions, such as evaluating end-of-life care or child custody options.
CONCLUSION 5.1: ExWB data are most relevant and valuable for informing specific, targeted policy questions, as opposed to general monitoring purposes. At this time, the panel is skeptical about the usefulness of an aggregate measure intended to track some average of an entire population.
At this point, evidence about interactions between ExWB and other indicators is inconclusive. For example, on the relationship between income and ExWB, Deaton and Stone (2013b) note that, at least cross-nationally, the relationship between aggregate positive emotions (here, meaning day-to-day ExWB) and per capita GDP is unclear:
The countries of the former Soviet Union are among the unhappiest in the world, unhappier than the Congo, Benin, or Chad, for example, and Italy and Denmark are unhappier than Mozambique, Sudan, and Rwanda.
They conclude that such revelations cast doubt on using measures of ExWB to provide an overall assessment of human well-being: “While it makes sense for SWB measures to paint a different picture than GDP, it is
4 In reality, the national income accounts and the labor market statistics are also multidimensional. Neither yields a single measure that fully summarizes their rich detail.
hard to credit a measure that says that Denmark is worse off than Rwanda; being happy is a good thing, but other things surely outweigh it in any credible overall assessment of life.” It is important to note that Deaton and Stone’s point pertains only to data from the Gallup “happiness yesterday” question. As pointed out in the World Happiness Report (Helliwell et al., 2012), for life-evaluation questions—namely, the Cantril ladder of life, life satisfaction, and happiness with life as a whole—rankings of countries consistently show Denmark near the top and Rwanda near the bottom.
An important, but poorly understood aspect of SWB, is its causal associations—both between factors and reported SWB and between SWB and various outcomes. This is, of course, a difficult problem in many areas of social science. Heckman (2000, p. 91) described the difficulty in establishing causal relationships: “Some of the disagreement that arises in interpreting a given body of data is intrinsic to the field of economics because of the conditional nature of causal knowledge. The information in any body of data is usually too weak to eliminate competing causal explanations of the same phenomenon. There is no mechanical algorithm for producing a set of ‘assumption free’ facts or causal estimates based on those facts.”
This critique seems especially pertinent for analyses based on SWB data, given their inherent nature. For pure program evaluation, a full understanding of causality is not always necessary, but in general, we would like to know how self-perceptions of well-being influence behavior, as well as what conditions and factors influence perceptions of well-being. In most analyses, it is not obvious whether positive and negative emotions are the dependent or the independent variables. The link between positive emotions and health appears stronger than the link between negative emotions and health, but we do not know the extent to which high positive ExWB creates better health or the extent to which better health creates conditions for high positive ExWB. Clearly, both can be taking place. The relationships between income and various SWB dimensions could also embody this kind of circular uncertainty.
As described in Chapter 2, experienced and evaluative types of wellbeing may have very different causal properties, and certain policies may only address one or the other of these dimensions of SWB. Those which aim to enhance longer-term opportunities may even impart negative short-term effects on daily experience. A policy designed to enhance living quality at the end of life, for example, focuses on the hedonic dimension (which is at least one of the objectives of palliative care, that is, relieving suffering), while a policy aimed at enhancing the education and opportunities of youth focuses on life evaluation (and the anticipation of the impact of
education). Thinking in terms of process versus outcomes, one can imagine that acquiring the skills and agency to lead the lives associated with high levels of life satisfaction is, at least at times, associated with stress and other experiences that could undermine happiness or even health. One can also imagine respondents with low expectations, agency, or capabilities finding contentment in particular daily experiences, such as socializing and eating, at the expense of longer-term objectives, such as investments in education and health. The example comes to mind of people who are obese and unhappy—but less unhappy than high-obesity cohorts that have even worse health and lower income mobility (Graham, 2008). It is likely that there are comparable effects for smokers, among other examples. In such contexts, considering only one dimension of well-being, such as ExWB in this instance, could lead to bad policy outcomes, and vice versa.
If daily experiences are negative enough, they might overturn the longer-run objectives of policies. A good example comes from George Akerlof’s work on identity (Akerlof and Kranton, 2010). He cites work by Robert Foot Whyte on youth in gangs in New York City who receive scholarships to go to top boarding schools. Often they do not fit in at the new schools and find the experience so unpleasant that they drop out. When they return home, they no longer fit into their home environments. The bottom line of the story is that the daily experience eventually determined the long-run outcomes (Akerlof and Kranton, 2010). Krueger and Mueller’s (2011) work on the hedonic well-being of the unemployed shows that the longer the sadness associated with failed job searches is prolonged, the more likely they are to quit searching for jobs, ultimately affecting their global life satisfaction evaluations as well.
Momentary feelings and experience drive some health behaviors—eating and smoking habits, for instance—while global memories drive other kinds of behavior, such as economic decision making. For example, a person does not think about his or her car most of the time, even while driving. But, when choosing a car to purchase, the global memory is of the car because the person has been prompted. Thus, ExWB measures can reveal the well-being differences between daily activities better than long-term measures such as life satisfaction (evaluative well-being). This makes the ExWB measures ideal for assessing factors that vary across people’s days. In contrast, life satisfaction is more likely to reflect general, long-lasting factors such as unemployment, income, or a happy marriage, although it is easy to see how these circumstances could directly impact ExWB.
ExWB measures may also be capable of uncovering the impact of objective conditions that are themselves not known by individuals. For example, air quality is known to influence mood and behavior, and even life satisfaction (see Luechinger, 2009), but it is difficult for people to recognize such associations and report them, while subjective reports of feelings (which
may fluctuate as a function of air quality) may provide more accurate—or at least more useful—information. It is exactly these sorts of associations that the combined use of granular time-use and emotions approaches, such as Ecological Momentary Assessment or the Day Reconstruction Method (DRM), are capable of identifying.
Friendships and socializing (connectedness) stand out as additional factors extremely important to ExWB. Connectedness is also important to life satisfaction and other evaluative well-being metrics, but it may be more important in relative terms to the evaluative well-being of those respondents with less means and opportunity than of those who have greater capabilities and other overarching life objectives.5 In this instance, agency may be an important mediating factor (Graham, 2011; see also the findings by Diener et al., 2010, on religion and friendships around the world).
Thus, the various types of SWB measures reveal distinctly different things. While people with children tend to evaluate that aspect of their lives as highly important and meaningful, time spent with small children is often reported as the least enjoyable time of the day in time-use surveys (as any busy parent who has had to drop all else at work to take a sick child to the doctor can attest, although the experience hardly results in less love for the child).6 Understanding this difference—for example, that child rearing can cause quite intensive stress, even in the context of deep affection and it being a desirable aspect of life—could help policy makers better understand the constraints faced by those individuals or cohorts without the means to cope with that stress, among other things.
CONCLUSION 5.2: To make well-informed policy decisions, data are needed on both ExWB and evaluative well-being. Considering only one or the other could lead to a distorted conception of the relationship between SWB and the issues it is capable of informing, a truncated basis for predicting peoples’ behavior and choices, and ultimately compromised policy prescriptions.
5 Robert Sampson’s Chicago neighborhoods study (Sampson and Graif, 2009) reveals the importance of connectedness to the well-being of neighborhoods. One of many examples is the variation, even among relatively poor areas, in the resilience of different neighborhoods to the 1994 heat wave in the city. Sampson’s findings were used to support the creation of a new (for 2013) Neighborhood Social Capital module of the U.S. Department of Housing and Urban Development’s American Housing Survey. The survey asks about trust in neighbors, friends in one’s neighborhood, interactions, connectedness, and so on. SWB questions might add an insightful dimension to this module, in that they could reveal nonmonetary elements of people’s surroundings that influence their well-being.
6 See, for example, the findings on women in Texas by Kahneman and Krueger (2006). New work by Deaton and Stone (2013a) finds that parents have more positive affect but also more negative affect.
One example where these downsides could occur is considering only ExWB in the case where obese individuals are less unhappy in high-obesity cohorts than in lower-obesity cohorts. A second is considering evaluative well-being when looking at the effects of acquiring skills and agency without considering the stress, function of emotions, and other health effects more closely linked to ExWB.
The evidence suggests that life satisfaction correlates more strongly with external factors such as income and economic region, whereas ExWB measures correlate more strongly with personality. This difference raises questions about how people adapt (discussed in Chapter 4) and about the feasibility of improving ExWB in the long term. A possible implication of adaptation is that, if people are accustomed to living in deplorable economic conditions (so the conditions are no longer reflected in their ratings of pain, stress, and discomfort), their chronic suffering is no less real or in need of policy attention just because they have become used to it. Describing this issue, Sen (1985, p. 14) wrote:
A person who is ill-fed, undernourished, unsheltered, and ill can still be high up in the scale of happiness or desire fulfillment if he or she has learned to have “realistic” desires and to take pleasures in small mercies … the metric of happiness may, therefore, distort the extent of deprivation in a specific, and biased way … [and] it would be ethically deeply mistaken to attach a correspondingly small value to the loss of well-being because of this survival strategy.
Deaton, reiterating Sen’s point, concluded that:
we should not base policy on a measure that is subject to hedonic adaptation. Yet the extent to which any particular measure of SWB is actually subject to the adaptation critique is a question that can be investigated empirically, so that it is possible that Sen’s concern is hypothetical, or is hypothetical for some measures but real for others. Note also that Sen does not deny the goodness of happiness in and of itself, only that it is an unreliable indicator of overall well-being.7
Much of the adaptation question has to do with the distinction between overall life satisfaction and day-to-day experience and with the time horizon of interest. Optimization of short-term versus long-term well-being (both at individual and aggregated levels) may imply different policy actions. A program to reduce fat intake or smoking may reduce ExWB in the short run but increase it (via the health covariate) over the long run. Life-
7 Presentation by Angus Deaton to the Panel on Measuring Subjective Well-Being in a Policy-Relevant Framework, December 2012.
cycle modeling and interplay of ExWB measures with evaluative well-being measures will play a role in advancing the assessment of SWB for specific policies.
CONCLUSION 5.3: The type of ExWB measurement employed for policy use will depend on the specific questions to be addressed. In some cases, global-yesterday measures may suffice, but in other cases a DRM-type measure may be more valuable, because it captures time-use and allows associations between affect and specific activities (which may be selected with the research question in mind). In general, ExWB measures are likely to be most valuable to policy when they (1) capture time-use and (2) associate affect with specific activities, as these kinds of data are amenable to being applied to answer specific questions (as opposed to all-purpose, tracking-type questions).
SWB data and statistics are helpful for identifying areas of need and informing policies targeted at subgroups of the population. As emphasized throughout this report, the panel believes the most compelling case for SWB data is its potential to identify populations that are suffering and to help in the study of the sources of that suffering. On the positive-emotion side, there is promising research indicating that health benefits are associated with certain emotional states—but the policy application is less obvious. Here the panel examines several specific possibilities for using SWB data in policy decisions.
The health domain seems a good starting point for thinking about ExWB and policy. Quality-adjusted life years (QALYs) are usually used to assess health interventions, as they combine the quantity and quality of life into a single metric. Positive ExWB is not normally used in the quality assessments, although negative feelings such as anxiety and depression have been. Because QALYs are usually derived from general population perceptions of how health affects quality of life, they do not capture the actual experience of poor health. While QALYs give a metric for the quantity and quality of life, they have been criticized in several key respects (Garau et al., 2011; Loomes and McKenzie, 1989). As Graham (2008) and others have noted, their adoption to health problems could make policy decisions based on perceptions unreliable and inaccurate. Health policy decision making that utilizes ExWB measures, either in addition to QALYs or incorporated
into a revised QALY metric, could be a significant advance. Of relevance to this discussion, it is worth considering the demographic shift in life expectancy and the concurrent increase in the number of older people living with a chronic health condition. A metric based in part on adding years of life may be less useful than a measure of expected ExWB while living with a chronic condition. Interventions could aim to improve ExWB without targeting intractable underlying health problems.
In the United Kingdom, where the burden of social care is increasingly placed on family members, the emotional burden of chronic illness on the family is not adequately captured by the current QALY metric. Given this shortcoming, ExWB could be an appropriate metric for capturing the experience of ill health among patients and their care givers. Similarly, many people with disabilities receive informal family-based care, rather than institutional care. The public and private monetary costs of these two modes differ greatly. A policy-relevant question is, for a subpopulation of persons receiving informal care, are they, net of a range of covariates, experiencing greater SWB than a sample of those receiving formal care? And what is the potential burden, captured in terms of SWB, for the care givers in the informal sector? Might the latter burden be offset to some extent by a higher level of eudaimonia or purpose? These are important but unanswered questions. For evaluating these kinds of policies, simple end-of-day or global-yesterday measures may be sufficient in some cases. For others, the DRM or time-use methods may be more appropriate, although the burden on patients of collecting these data may be too great to merit recommending such methods in all instances.
Richard Frank gave the panel a number of examples from the medical realm for which ExWB metrics are particularly well suited and provide added value.8 Self-reports of SWB are likely to add useful information in instances where medical interventions have a desired outcome that is something other than merely an increase in life expectancy, where reflections of successful treatment and support extend beyond signs and symptoms and into domains such as functioning and social integration, and where parties other than the patients are affected by treatment and symptoms (care givers, family members, and others).
Valuing end-of-life treatment options is another area that calls out for more nuanced measurement than what simple life-expectancy numbers can provide. Considerable health care costs accrue at the end of life; in many cases, considerable benefits are derived from that care as well. Some agencies, such as the National Institute for Health Care and Excellence (NICE) in the United Kingdom, have explicitly raised the cost-effectiveness
8 Presentation by Richard Frank, Harvard University, to the Panel on Measuring Subjective Well-Being in a Policy-Relevant Framework, March 2012.
thresholds for coverage of end-of-life treatments (by 50 percent in the case of NICE). But is this the “right” policy, and how would one know? To help answer such questions, better data are needed on the impact that end-of-life treatments have on patients and on families and care givers. ExWB is a central part of that impact. Dolan et al. (2013), for example, have assessed the impact of health and life satisfaction on tradeoffs between quality- and length-of-life scenarios. These concepts may be especially important for the end of life, where the balance between predominantly purposeful and pleasurable activities might change (conceivably in either direction).
Terminally ill people often report high levels of purpose, which may translate into a higher reported life satisfaction than many would predict. Cancer patients’ will to live has been shown to vary by large amounts over the course of a month, and only somewhat less so over 12-hour periods. These differences can be explained by how the patients felt at the time they were asked about their will to live. Dolan (2008) looked at data on the life satisfaction (evaluative well-being) of cancer patients and found that levels worsen when the cancer is in remission. One possible interpretation of the data is that the imminence of death allows people to “get their house in order” and to solidify a sense of purpose in their lives, whereas remission casts uncertainty in a way that unsettles these thought processes. As with other areas of direct policy applications, more research is needed, including research on the interplay of evaluative well-being and ExWB.
To understand the full costs and benefits of treatment, all of the SWB ripple effects that flow from these circumstances—the immediate effects on patients and their families and the longer-term effects on families after the patient dies—need to be measured and valued. To date, there have been no serious attempts to consider the spillover effects on others over time in such cases. These kinds of results will be of interest to patients deciding upon treatments; clinicians concerned with establishing patient preferences; policy makers deciding on the cost-effectiveness of different interventions; and academic audiences in medical decision making, psychology, and economics.
At key decision nodes or key stages in disease progression, “standard” information could be elicited from patients and close family members on the health-related quality of life according to validated condition-specific and generic measures. Such questions would allow for comparison of the results from that assessment with the results of other studies that have used these measures (including all the recent submissions to NICE). Addition of ExWB measures to such assessments would allow for investigation of the degree to which different people adapt in different ways to their changed circumstances and would enable service providers to reflect more accurately the “epidemiology” of the treatment experience.
Many of the policies that may be informed by SWB require data capable of revealing contrasts at the local, or at least subregional, level. ONS is formally looking into possible applications; its case studies include
• Civil Service People Survey—insights into staff well-being to help steer engagement and human resource policies;
• Well-being of job seekers—joining up of mental health and job seeker services;
• Cabinet Office evaluation of the impact of National Citizen Service on the well-being of participants;
• Local government initiatives and policies; and
• Impact of sport and culture on well-being.
Another policy domain where ExWB measures may be useful is in the delivery of benefits. For example, beginning in 2014, the UK government is replacing statements of special educational needs with a simpler assessment process. Parents with a care plan will have the right to a personal budget for their child’s education and health support. This policy will enable parents to choose the support and services that they believe are right for their child, instead of local authorities being the sole decision makers. ExWB seems a suitable element to include among the measures used to assess the impact of this change, given the link between autonomy and SWB (although evaluation of a program that aims to affect a child’s education and health must surely be centered on the education and health outcomes of those children).
Disability and attendance allowances in the United Kingdom are currently paid to individuals to spend on whatever they wish, to support their independent living. Plans to remove these benefits and place the funds in local authority social-care budgets were shelved after campaigns stressed the importance of personal allowances in people’s well-being. But if policy changes such as these were to be implemented, then ExWB could be a suitable complementary measure, along with measures of objective well-being, to assess potential impacts. For evaluating policy changes in the delivery of benefits, simple end-of-day or global-yesterday measures may be adequate, although DRM and time-use assessments may be able to capture specific changes in ExWB while interacting with a child with special needs. It might also be possible to ascertain which activities that disability and attendance allowances support have particularly positive consequences for ExWB. If improvement in ExWB is afforded by access to the social activities and networks that higher disability and attendance allowances would make possible, with consequent improvements in health and reductions in the need
for services, then there might be both moral and cost imperatives to giving priority to funding allowances for the disabled and sick.
It is usually assumed that measures of evaluative well-being (life satisfaction) are appropriate for considering work-related policies, but ExWB would add a useful dimension untapped by evaluative well-being measures—such as in the case of policies addressing statutory retirement, unemployment, and working conditions. This extension is consistent with the aforementioned theme that measures of both evaluative well-being and ExWB are needed to provide a comprehensive picture of SWB. For example, in 2011, the policy on retirement in the United Kingdom was changed so that employers are no longer able to force employees to retire at age 65. Being able to continue working if inclined, even if unlikely to change people’s overall evaluations of their life, may well increase the positive-emotion aspects of their ExWB. Simple ExWB measures might be adequate for assessing this domain.
The UK policy to increase the statutory retirement age disproportionately affects women in their 50s. For example, a woman currently 55 years of age who thought she would be able to retire at 60 now finds she is not able to receive a state pension until age 66. Women in this age group typically exit the labor force to care for grandchildren, elderly parents, or both, but, without their state pension, they may not be able to afford to leave work to take up these family-care responsibilities. ExWB measures could capture the total burden of paid and unpaid work in late middle age, which other measures of well-being do not capture. Because the United Kingdom and the United States increasingly rely on informal care-giving to support an aging population, it might be important to know more about the decision-making processes involved in balancing paid and unpaid work.
Policies concerned with working conditions, rights, and practices are another domain in which ExWB could play a part. Insecure work has been shown to have almost as great an effect on SWB and the risk for anxiety and depression as does unemployment (Burchell et al., 1999; Ferrie et al., 2002). An interesting and important research question is the extent to which good work conditions and practices improve positive emotions, or at least remove a source of stress. One policy issue such research would obviously inform is whether or not flexible labor market policies are associated with a lower level of positive ExWB in the population. Along these lines, there is a large literature on job satisfaction and the quality of working life, although much of this research has been done in conjunction with overall life satisfaction metrics. Clark et al. (2008) examined the relationships between job satisfaction, wage changes, and future quitting behavior using data from the German Socio-Economic Panel. They found, as did Bertrand and Mullainathan (2001), that job satisfaction was as strong a predictor of the probability of quitting or changing jobs as was wage change. Taylor (2006) investigated day-of-week effects on job satisfaction and SWB. Com-
muting time and its relationship with ExWB is another often-cited policy application of SWB information (Kahneman and Krueger, 2006). For example, in deciding whether or not to create high-occupancy toll lanes in metropolitan areas, the well-being of people of different incomes who travel the highway has to be examined, along with the network effects—the consequences for those whose travel choices are affected even if they themselves do not use the highway—when estimating full aggregate costs and benefits of a new policy.
Other aspects of planning laws and the built environment could also be evaluated using ExWB measures. Cities that provide easy access to convenient public transportation and to cultural and leisure amenities, that are affordable, and that serve as good places to raise children or to keep older residents better connected have happier residents (Leyden et al., 2011). Data generated by surveys of neighborhood social capital, such as the American Housing Survey (conducted by the U.S. Department of Housing and Urban Development) or Robert Sampson’s survey of Chicago neighborhoods (Sampson and Graif, 2009) are useful to researchers investigating whether and how changes to the built environment can promote SWB (the alternative hypothesis being that happier people tend to have more autonomy over where they choose to live). Creating spaces and buildings that encourage and promote SWB, as a worthwhile investment for public health, is an idea that has gained some currency with policy makers. An architectural think-tank book, Building Happiness: Architecture to Make You Smile, attests to the acceptance of measuring ExWB to inform policy in this area (Wernick, 2008). Moreover, the relationship between the environment and ExWB can affect policy in areas other than just public health. For example, social and economic benefits go hand in hand with the experiential benefits. Still, more data and more research are needed to better understand what ExWB measures add to assessments of the benefits of green space, transport, or clean and safe urban areas.
Another policy-relevant domain involves the SWB of the unemployed and their experiences as they undergo prolonged job searches, as work by Krueger and Mueller (2011) has shown. Not only did they find that the SWB of the unemployed declines with the duration of unemployment spells; they also found that the time spent involved in job search is particularly unhappy and the unhappiness increases with the time spent in job search (measured both with life-satisfaction and sadness variables). These effects on the unemployed provide an example of how low ExWB related to the process could in the end undermine individuals’ incentives to persist, ultimately reducing their capacity to achieve higher levels of evaluative wellbeing in the future.
Yet another area with possible policy implications is the SWB of refugees or immigrants. SWB metrics can be used to help assess how well they
are adapting and assimilating to their new environments, which in turn has repercussions for social stability, investments in children’s education, and so on. Does scoring better on either dimension lead to better adaptation or assimilation skills? Do low levels of ExWB, as these people experience the process of adapting to a new environment, lead to lower levels of success in the job market and other areas where greater success would contribute to higher levels of evaluative well-being in the future? Some initial evidence suggests that migrants are more likely to be unhappy prior to migrating, as well as post-migration (Graham and Markowitz, 2011).
Other examples from the literature where ExWB measures have been proposed for informing cost-benefit policy analyses include the following:
• Evaluating trade-offs between inflation and unemployment (Gandelman and Hernández-Murillo, 2009);
• Environmental policies (Ferreira and Moro, 2009);
• Full valuations of cash transfers, earned income tax credit, food stamps, back-to-work programs, and other social policies (Blattman et al., 2013);
• Connectedness (or loneliness/isolation) and health among the elderly; given demographic trends, what will more isolation mean? (Helliwell, 2002);
• Quality dimension of child care arrangements; experiences of parents at work with or without subsidies and/or child care, and with or without health insurance (Brodeur and Connolly, 2012); and
• The effects of different custody arrangements on the SWB of adolescents in divorce situations (e.g., Amato, 1999).
Beyond and apart from informing policies, there is an important role for ExWB measures in advancing research in behavioral sciences, epidemiology, medicine, and even law. Further, such data perform a general information role of interest and value to the public and media. This informing function of basic science often ultimately leads to policy relevance and innovation of science generally. For these reasons, it is worthwhile for governments (and others) to continue to learn about the SWB of the population, especially given that people’s well-being, both subjective and objective, is often the ultimate objective of public and private policy. Media and the general public have shown great interest, for example, in information about why some groups—defined by various characteristics or by place—seem happier than others. ONS has explicitly expressed, as part of its Measuring National Well-being Program, the goal of “an accepted and trusted set of National Statistics to help people understand
and monitor national well-being.”9 The underlying belief here is that the need for basic descriptive information is enough justification to warrant data collection, even if causal links between SWB measures and social and economic outcomes (or vice versa) have not yet been established. A broader resonance with the public is driving the recent movements, such as those by ONS and the statistics offices of other countries to implement SWB measurement. At the moment, this informational role is dominated by measures of evaluative well-being. Much of the value of the U.S. Decennial Census—which, granted, is required by the U.S. Constitution for the purpose of drawing political districts and also provides data used for all manner of federal programs—is in its by-product of descriptive information about who Americans are as a society. In general, support of many of the U.S. federal surveys is validated by this extremely important role in producing information regarding the public good.