Happiness Inequality #3: Putting It All Together
Two days back I noted that happiness inequality today is at much lower levels than in earlier decades, despite rising income inequality. What lies behind these trends?
Economists typically think about inequality as reflecting three influences.
First, differences in average levels of happiness between groups may have changed. And indeed, in yesterday’s post, I showed that differences in happiness by gender and race have narrowed, even as differences by education widened.
A second influence is that inequality may have changed within even narrowly defined groups. To get at this, we broke the sample up into 24 sub-samples, reflecting a division by gender, three separate age groups, and four education levels. Within 20 of these 24 sub-samples, we found that happiness inequality has declined, and in no case did we find a group for whom there is a statistically significant increase in happiness inequality.
We also ran a more complex analysis, tracing the evolution of the inequality of happiness as a function of each of six demographic variables. The results are shown in the following graph:
While some of these lines are hard to distinguish, the main insight is that the trend decline in the variance of happiness within each group is surprisingly similar across groups.
The third contributor to happiness inequality is changes in the composition of the population. For instance, falling numbers of high school dropouts yield lower happiness inequality.
Putting these three pieces together, we tried to figure out the extent to which the changing distribution of happiness reflects changes between groups, changes within groups, and changing composition.
Our key finding is that most of the movements in happiness inequality reflect changes in happiness inequality within even narrowly-defined demographic groups, and these changes are quite pervasive.
While this is a statistical explanation, it simply begs the question: What changes could have narrowed happiness inequality so pervasively? And juxtaposing our observed trends in happiness inequality with measures of income inequality — which have pretty much risen for the past four decades in a row — presents a real puzzle. How might we reconcile these trends?
Our sense is that there may be important trends in the non-pecuniary domain that have had a major equalizing effect. But once you start thinking about the possibilities, the list of suspects is pretty long: what has happened to leisure, to family life, to religiosity, to communities, to notions of procedural fairness, belief in the American Dream, or to personal freedoms? Could these changes yield a more equal distribution of happiness. And how could we test these theories?
At this point, our research has simply documented the facts about the evolution of U.S. happiness, and Betsey and I are yet to dig into the bigger question of why. So let me ask: What do you think explains these trends in happiness inequality?