Happiness for All

Review of book by this title by Carol Graham

Carol Graham is at the forefront of the movement to bridge economics, public policy and the growing literature on wellbeing and its determinants. This book is a substantial step forward in that endeavour. It presents innovative work using wellbeing metrics to provide new insights into long-running questions in economics and public policy around inequality, welfare and mobility. It is particularly outstanding in this literature for two reasons. First, it is aware of the limitations inherent in much wellbeing data and utilizes clever techniques to avoid many of them. Second, rather than staying entirely within wellbeing metrics, it instead uses these metrics specifically and deliberately to explore underlying drivers of economic phenomena like discount rates, education decisions and crime rates.  

Before discussing the insights of the book, it’s necessary to discuss data for a moment. The admissibility of wellbeing data has been a controversial issue at least since Easterlin first questioned the “more income = more utility” assumption using wellbeing data in 1974. Early criticisms pertained to the capacity of such data to parse long-term evaluative wellbeing out from shorter-term emotional moods, as well as to the large effects that priming had on responses to subjective wellbeing questions. In particular, people’s responses could be substantially influenced by asking them about negative things like politics before asking them about their subjective wellbeing, or by manipulating them with seemingly inconsequential factors like finding a dollar before the interview (Gilbert 2006). What dimensions of life mattered to people’s evaluative wellbeing could also be influenced by priming. For example, asking about dating in subjective wellbeing surveys has been shown to significantly increase the importance people place on their romantic entanglements when considering their wellbeing relative to control groups who are not asked about dating.

Graham was a member of the National Academy of Sciences Panel on Well-being metrics for Public Policy from 2011–2012 and is also a scientific advisor to the Gallup organisation, which collects the World Values Survey, one of the principle sources of wellbeing data used by researchers. She is very much across the criticisms of wellbeing data, and has been a central part of efforts to address them during data collection. For example, life satisfaction surveys now typically include several questions that allow different dimensions of wellbeing like satisfaction and mood to be teased apart. The questions are also located in surveys in positions where they are unlikely to be affected by questions that come before them. Some noise will still creep into responses. For example, if someone is having a bad day—perhaps they missed the bus before the survey—this will result in most cases in them giving lower responses, even to questions that request a long-term evaluation of life satisfaction. However, the Gallup data sets used in Graham’s book contain thousands of observations. The distribution of these biases throughout survey responses can thus reasonably be expected to match the distribution of the relevant events in the day to day life of respondents. As such, they’re not really a problem.

One issue does remain, which is concerns about scale-norming (Fredrick & Loewenstein 1999, pg. 308). This is where the qualitative meaning of different points on a respondent’s scales changes over time. For example, for a graduate student, completing their PhD might be what they think is needed for them to reach 9/10, while getting a tenure-track position at a top 200 university is what they need to be 10/10 (their ‘best possible life’). However, when they do get that tenure track position, their horizons necessarily push outward, and now their 10/10 is getting to Professor. As such, someone could give a 9/10 in two consecutive waves even though their own assessment is that their life is getting better. They simply can’t communicate that within the strictures of the scale instrument.

Scale norming is a problem for wellbeing research that needs much more investigation, and it does raise concerns about some of the discussion in this book. However, a great deal of Graham’s analysis sidesteps scale-norming worries. For starters, the focus of the book is the badly-off. These people have a huge amount of space in their scale responses to improve into, as it were—they don’t need to rescale. Second, Graham makes ingenious use of data about dimensions of affective wellbeing that are either not measured on a scale, like the number of times someone smiled yesterday, or where it is reasonable to assume that the affective phenomenon in question does exist on a scale, such as someone’s stress burden. These data are used to check the robustness of the scale data, and also provide their own insights. Finally, Graham focuses on questions that can be examined using broad correlations in the data. For these, the suggestive evidence provided by wellbeing metrics is sufficient to provide a rudimentary test of hypotheses, especially as those hypotheses are fairly intuitive; for example, that people who are hopeful about the future invest in it while those who are hopeless have much higher discount rates.      

The central claim of the book is that hope matters. More formally, “the Gatsby curve in opportunities has an additional beliefs and behaviours channel that may perpetuate and deepen inequality over time.” The Gatsby curve shows a correlation between inequality as measured by the gini coefficient and intergenerational economic mobility. As inequality rises, mobility falls. Graham’s book illuminates several new mechanisms that drive this relationship.

The first of these is that stress has differential impacts on the wealthy and the poor, or rather, that the poor and the wealthy experience different kinds of stress and this conditions their behaviour. ‘Good stress’ is associated with goal pursuit, while ‘bad stress’ emerges out of desperation. Individuals who expend a large amount of cognitive energy on contemplating bills and how to secure food have little leftover for long-term planning. They sometimes consequently engage in economically irrational behaviour.

The second, related mechanism is that beliefs about the long-term benefits of hard work affect how much effort people expend in the present. Graham notes that inequality can be a source of individual effort and economic dynamism when it reflects the fact that people can get ahead if they work hard. However, when inequality instead reflects the fact that institutions and incentives favour the already privileged it has a dampening effect on both effort and dynamism. People at the bottom of the distribution develop a sense of hopelessness and consequently disconnect from the economy. Data about people’s attitudes to hard work from the Gallup World Values survey suggest that Americans, especially white former manufacturing workers, are becoming increasingly sceptical that hard work can secure the American dream and are consequently engaging in short-sighted activities like opioid abuse.

A final mechanism that receives substantial attention in the book is that the poor often face higher opportunity costs to attain basic goods than the rich. A notable example is that the bureaucracies that administer universal welfare programs like Medicaid are considerably more efficient and user-friendly than welfare programs that are strictly for the poor. An incisive anecdote in this regard is that the penalties for selling a food stamp for cash are around the same as those for manslaughter and sexual contact with a child under 12.

There are a range of other very interesting themes in the book, including how wellbeing and optimism about the future differ across racial groups in America, how cultural differences in attitudes towards work and fairness can affect behaviour, and how mechanisms for coping with stress and insecurity differ across social classes. These are all insights that bear not only on aspects of economic policy concerning welfare and mobility, but also on the core business of economics, namely predicting human behaviour in a parsimonious manner. The principle merit of Graham’s work is in this demonstration that wellbeing theory and metrics, even in their present infancy, have a lot to contribute to economics and economic policy and deserve to be taken more seriously. Happiness for All not only presents Graham’s original work in this area, but also does an excellent job of succinctly summarising what ideas in the wellbeing literature are most readily adaptable for these purposes.

Overall, Happiness for All is a great read for anyone who demands quantitative methods but also recognises that inequality and social mobility have substantial qualitative and normative dimensions. It is timely in the aftermath of Trump’s election, which many have attributed to feelings of hopelessness and desperation among large segments of America’s population, especially in the rust belt states. It will also be useful to people looking for policy responses to the loss of hope in America, as the final chapter deals specifically with wellbeing-literate ways to rekindle the American dream.   

References

Easterlin, R. (1974), ‘Does economic growth improve the human lot? Some empirical evidence’, in David and Reder eds. Nations and households in economic growth: essays in honour of Moses Abramovitz, New York, NY: Academic Press

Frederick, S. and Loewenstein, G. (1999), ‘Hedonic Adaptation’, in Kahneman, Diener and Schwarz (eds.), Well-being: The Foundations of Hedonic Psychology, New York, NY: Russell Sage Foundation

Gilbert, D. (2006), ‘Stumbling on happiness: ‘, London, UK: Harper Perennial