Published on Development Impact

Wealth and the endogeneity of behavior

This page in:

Allow me to take the occasion of the 236th “birthday” of my native-born country (celebrated on July 4th here in the U.S.) to go far afield and discuss a topic that, while grounded in empirical social science, doesn’t touch directly on impact evaluation. The topic is how the personality traits of an individual may be related to his or her relative wealth.

A recent spate of research on this topic is summarized in an enjoyable article in New York Magazine. The article centerpiece is a new paper by Piff and co-authors that begins with a review of previous studies linking the greater resources and independence found among the “upper classes” and what they term “self-focused cognitive tendencies”. Self-focused tendencies in turn may facilitate unethical behavior – apparently in observational data there is a clear correlation between professed tolerance for unethical behavior such as cheating and whether the respondent is of high “social class”. (The authors define social class as an individual’s rank with respect to others across the three dimensions of wealth, occupational prestige, and education.)

Piff and co-authors then relay the results of several naturalistic field studies they devised, including one that involves observation of whether drivers of “upper-class” cars are more likely to cut-off pedestrians at a particular cross-walk in Berkeley California. It turns out drivers behind the wheel of cars in the top two quintiles of car status cut-off the unlucky student serving as the study pedestrian about 45% of the time. Drivers in less valuable cars (i.e. older, less expensive models) cut-off the pedestrian only about 30% of the time. (One obvious difficulty with this measure of class is that a car is often an aspirational good, at least for certain folks, and thus may be a particularly noisy measure of overall social status.)

If we accept this link between social class and self-prioritizing behavior, it is natural to ask whether individuals with these behavioral tendencies are more motivated by wealth and hence select into an “upper class” or, rather, the experience of relative wealth itself induces a generalized lack of empathy and reduced cognizance of others. So Piff and co-authors move on to additional experiments that attempt to experimentally prime a feeling of belonging to either a higher or lower social class among randomized study subjects. They do this by asking the subjects to compare themselves to people with either the most or least money in society (or the most or least education and respected jobs). Apparently earlier studies that use this approach have noted a reduced ability to identify with others’ emotions after a subject was experimentally manipulated to identify as “upper class”. In this experiment, after subjects were manipulated they were then presented with a jar of candy ostensibly meant for children in a nearby laboratory and told they could take some if they so wanted as they delivered the jar to the other lab. As you may probably guess, participants induced to feel “upper class” took significantly more candy (that would otherwise go to children).

The authors conclude from this and additional experiments that individuals from upper class backgrounds behave more unethically in both naturalistic and laboratory settings and that, to some-extent, this unethical behavior is induced by cognizance of relative privilege.

While this strikes me as incredibly interesting, the conclusions are far from iron-clad as yet (I imagine Piff and co-authors would agree). I mentioned problems with the social class measure used in the car experiment. In addition, in the candy jar experiment, the population is certainly WEIRD since it is comprised of UC Berkeley undergrads. In fact a recent working paper by Trautmann and co-authors analyzes a large scale panel data set from the Netherlands with various objective measures of class as well as elicited statements of ethical judgments. Here the story is not so clear. There is no apparent systematic relation between social class (variously defined as wealthy, high income, prestigious job, and high education) and ethical views, although wealthy Dutch respondents are significantly more likely to claim that it can be justified to cheat on taxes or commit adultery than the less wealthy. Furthermore upper-class subjects are not anymore likely to choose to betray when assigned to the second mover in an anonymous two-player trust game.

The contrasting results in the studies can be due to numerous factors (and their interactions). We have different populations, different measures of “social class” – subjectively defined (i.e. induced in the subject) or assessed on the basis of observable characteristics – and different measures of the expression of ethical views. Another difference is the prevailing cultural norm and institutional framework that governs how comfortable a subject feels in either expressing self-focused tendencies or exhibiting related behavior. Even if self-focused tendencies exist to an equal degree in the two populations, the extent to which privileged individuals feel comfortable enough to “let it out” may be a function of the prevailing environment.

The biologist Edward O Wilson recently wrote a fascinating piece arguing that the capacities for both ethical pro-social behavior as well as selfish behavior exist in each of us. This dual nature is a direct consequence of two evolutionary forces: between-group and between-individual selection. Personally, I have little doubt that we contain both ethical (pro-social) and greedy tendencies to various degrees. Perhaps, under certain institutional frameworks and cultural norms, the privileged among us are more likely to express the greedy side if they face comparatively little sanction?

So back to the July 4th holiday. The theme of group vs. individual rights, including the topic of insurance from risk vs. individual responsibility, has been an ongoing political discussion since the founding of the U.S. all the way through to the present day (to wit, recent discussions around health care reform). This has resulted in a rather unique institutional setting that today yields some of the highest rates of wealth inequality among developed countries. The psychological studies discussed here serve as a reminder that not only do we shape our political and economic institutions but are likely shaped by them in turn.



Jed Friedman

Lead Economist, Development Research Group, World Bank

Join the Conversation

The content of this field is kept private and will not be shown publicly
Remaining characters: 1000