Skip to main content

Verified by Psychology Today

Why We Are All Crooked Bankers

A laboratory game shows we might be generous to some, but ruthless to many.

Key points

  • New research shows that people are simultaneously generous in one-on-one interactions and selfish enough to harm a large group for personal gain.
  • The prospect of a large gain suffices for most people (not just bankers and managers) to steal from a group.
  • We did not need to dangle million-dollar gains in front of people to elicit this behavior. A hundred bucks did the job!
Jaume García Segarra, used with permission
The prospect of a large gain suffices for most people to steal from a group.
Source: Jaume García Segarra, used with permission

Here are two things you probably “know.”

The first is that, since roughly the 1990s, there have been a large number of high-profile corporate scandals demonstrating an appalling degree of selfishness among managers, bankers, and other high-level economic decision-makers.

The second is that social and economic psychologists and behavioral economists have spent the recent decades criticizing traditional economics and showing that the textbook picture of the selfish Homo Oeconomicus is wrong, because people care for others and are basically generous.

If you are raising an eyebrow or two right now, you are not alone. It seems like something might be wrong there. Let us check. On the one hand, the scandals affecting the financial industry “showed that alongside a widespread failure of competence in the banking industry, there was a failure of professionalism and ethics.” These are not the words of some pundit or enraged blogger. It is an official statement of the U.K. government in its 2013 response to the Parliamentary Commission on Banking Standards.

Examples like this abound, and they are a just response to a succession of scandals where bankers and managers were seen to be appropriating resources from a large number of individual investors and citizens. Selfishness!

On the other hand, experimental economists use simple lab paradigms and experimental games to show that people can, in fact, be generous toward one another. For example, in the “Dictator Game,” a participant can split a certain amount of money between herself and a powerless participant with total anonymity and impunity. With no repercussions of any kind, one might reasonably expect her to simply keep all the money. And yet people typically and overwhelmingly give a fraction of the money to the other person. Generosity!

What is going on?

There is an easy way to square these opposite behaviors: The woman in the experiment and the disgraced banker are not the same people. Experimental psychologists and economists typically bring students to the lab. True, the experiments have been replicated with general-population samples, but bankers and managers are not necessarily representative of the general population. The problem, you might think, is that high-level economic decision-makers are a different kind of people. Maybe they were always different and gravitated to certain professions. Or, maybe the economics training that they underwent turned them into egoists. So, we are generous! It’s just them who are selfish.

Sadly, like most simple explanations which make us feel good about ourselves, this one might be wrong.

In our research, we set out to demonstrate that, simply put, the roots of corporate scandals are in all of us. It’s not “them.” It’s us.

In the “Big Robber Game,” we invited hundreds of people to the lab, in groups of 32. These were normal students, as in any typical economics experiment. They interacted anonymously in a number of games along the lines of the Dictator Game, and, as in previous studies, they were basically generous. For instance, when asked how they wanted to anonymously split 10 Euros (of real money) with somebody else, they voluntarily transferred some money, even when the other person was powerless to retaliate.

But here’s the twist. Half of them (the robbers) were given the chance to steal the earnings from the other half (the victims). The victims were a large group: sixteen participants. Altogether, they had around 200 Euros of real money, that they had earned in the Dictator Game and the other games. Each robber was asked whether he or she wanted to steal half the earnings of all the victims, which would be about 100 Euros. In each group, one robber was randomly selected and his or her decision was actually implemented, so the decision was very much real. They could also steal less, say one-third or just one-tenth of the money. Or nothing at all. What would you do?

Most of them just stole as much as they could, like our idea of a corrupt manager would have. The majority stole the maximum (half of the earnings of all victims), over 80 percent of robbers took one-third or more, and almost nobody declined to rob. There were no differences between economics students and those in the humanities or the sciences. There were no gender differences.

Our participants were simultaneously generous when interacting anonymously with one other participant (in the Dictator Game), and selfish to the point of ruthlessness when they could run away with 100 bucks, even though this would severely harm a large group, like a corrupt manager who steals from a pension fund before breakfast and then slips a bill to a homeless person in the street.

How can we be generous with individuals but selfish to the masses?

When considering our own welfare and that of others, our behavior entails a tradeoff between personal gain and concern for others. We obviously like money. But we also dislike inequality, and this is what hundreds of experiments in economic psychology and behavioral economics have (correctly!) shown. However, and this is the key insight, the inequality created by a decision varies wildly if that decision affects a group or just another individual.

The decision to appropriate the entire pie of 10 Euros in the Dictator Game creates a distribution of 100 percent for me vs. 0 percent for the other person, a difference of 100 percentage points. Maximum inequality. With 16 victims, each with 10 Euros, if a Big Robber, who also has 10 Euros, comes around and steals half the earnings of every victim, he or she will end up with a share of 90 Euros out of the group's 170 total, or around 53 percent of the pie. That would leave each person with 5 Euros, around 3 percent of the pie, but half of what they started with. When stealing from a large group, the benefits are far larger, but the inequality you create is smaller. Hence, even if we dislike inequality, the results are completely different and we seem to behave differently.

Our results, published in Nature Human Behaviour, show that, before the Big Robber Game, laboratory experiments rightly revealed a large degree of pro-social behavior simply because they focused on bilateral interactions. But data suggests that the same people who behave pro-socially in such experiments (which is most of us) would behave scandalously when tempted by a high reward obtained by harming a large group. Sadly, this is not about bankers and managers, and we cannot hide behind alleged differences in training or career choice. It is really us. What is more shocking is that the price of our moral principles might be astonishingly low. We did not need to dangle million-dollar gains in front of people to elicit this behavior. A hundred bucks did the job!


Alós-Ferrer, C., García-Segarra, J., & Ritschel, A. (2022). Generous with Individuals and Selfish to the Masses. Nature Human Behaviour, 6, 88-96.Alós-Ferrer, C., García-Segarra, J. & Ritschel, A. Generous with individuals and selfish to the masses. Nat Hum Behav 6, 88–96 (2022).