In their paper Subramanian and colleagues show us that both individual and community level health and happiness tend to covary but that the community level health-happiness covariation is quite a bit stronger than that for individuals.¹ Their approach is novel in that they examine the covariation in these two outcomes nested within individuals who are in turn nested within USA communities.

The study design of the Subramanian paper forces the questions: Are people naturally endowed with a degree of happiness that life events, whether positive or negative, are unlikely to influence? Or are there life circumstances, set in motion by social contexts, which have enduring effects on our levels of happiness? If we assume that people have a set point of happiness homoeostasis that only wavers slightly in response to either positive and negative life conditions and events we might expect that happy people will tend to feel healthy, be optimistic, vital, and we would expect health and happiness to covary within individuals across contexts. On the other hand, if it is life circumstances that permanently change our level of happiness then we would be more inclined to suggest that some underlying life conditions are implicated in the codetermination of health and happiness.

Easterlin argues that there indeed are some conditions of life, such as a decline in health status, that permanently change our level of happiness. Pointing to data from the US general social survey, he shows that across the life course, those in poorer health report consistently lower happiness than those in better health and he conjectures that the balance of evidence favours the direction of causality running from health to happiness and not the other way around.² Marriage has a lasting positive effect on happiness and divorce a lasting negative effect. The case of income, however, is quite different. Although cross sectional studies like that of Subramanian and colleagues typically show a significant positive association between income and happiness even in the presence of control variables, he suggests that cohort studies show no trend of increased income being associated with increased happiness. So according to Easterlin’s ideas, health and marriage can permanently affect happiness but, in psychologist’s parlance, there seems to be “complete hedonic adaptation” for income. Income’s inability to nudge us towards happiness, he explains, is attributable to the fact that material desires expand with increased incomes leaving us perpetually at roughly the same distance between our material goals and our ability to attain them.

There also seems to be some kind of hedonic adaptation at higher levels of social organisation. In the USA real personal incomes have grown substantially since the middle of the past century but the proportion of Americans reporting themselves “very happy” has hovered around one third.³ Internationally, happiness is not correlated with GDP per capita beyond a threshold of about $13 000.⁴ Indeed, there is a remarkably similar shape (diminishing returns to both health and happiness beyond a basic threshold) to the relation between income and health and income and happiness at the scale of the nation state—suggesting that income, health, and happiness covary strongly at higher levels of social organisation. This is precisely what Subramanian and colleagues report for US communities.

But why do we not see the same degree of alignment of health and happiness within individuals? Myers states emphatically that “age, gender and income (assuming people have enough to afford life’s necessities) give little clue to someone’s happiness” (page 65), although we know these to be strong predictors of an individual’s health status through years of epidemiological study.³ Health, however measured, tends to decline with age. Women consistently outlive men in most countries around the world. Despite the modest role of health selection effects, the epidemiological evidence supporting the causal relation between income and health status is strong. Thus it would seem that some of the most important common indicators of health and happiness are probably still out there to be measured in a future study. Among other candidates, things like self esteem, extent and quality of social relationships, job satisfaction, and leisure time activities should probably be considered.^3,5 As for the role of community context, we know from the Subramanian piece that healthy communities tend to be happy communities and this finding should force us to consider what forms of social organisation are most supportive of these candidate common indicators of health and happiness.