GAB is delighted to welcome Nils Köbis, a researcher at the Max Planck Institute for Human Development, to contribute today’s guest post:

Suppose that you need some sort of official license, such as a fishing permit. Would you consider obtaining that license—or obtaining it more quickly—by paying a bribe? Now suppose that you are traveling in a foreign country and you need a similar sort of license. Would you consider paying a bribe to get that license in that foreign setting—if we assume that the probability of getting caught and the possible penalties are the same as in your home country, but that bribery is much more common by citizens of that country?

Are your answers to the two questions the same? Do you think other people’s answers to those questions—or, more importantly, their actual behavior—will be the same or different, depending on whether they are at home or abroad?

This question implicates a more general issue in moral and behavioral psychology. Some believe that the moral constraints on our behavior are relatively stable: In the example above, some people believe paying bribes is wrong and won’t do it, no matter where they are, while others are willing to pay bribes, at least if they the advantage of doing so is big enough and they think they will probably be able to get away with it—again, without reference to other aspects of the surrounding context. But some research has suggested that the (perceived) behavior of others can exert a strong pull on our moral compass (see, for example, here, here, and here).

To further explore this question, my collaborators and I conducted a study that involved online experiments with 6,472 participants from 18 nations, in which the participants played a bribery game based on the our opening example. Our findings were both surprising and intriguing, and suggest that our inclination to engage in corrupt behavior is influenced by our stereotypes (not always accurate) of people from diifferent countries.

Our experimental set-up was as follows: In each round of the game, one player is assigned the role of the “citizen” who wants to obtain a public license, and another player is assigned the role of the “public official” who can grant or withhold that license. The citizen decides whether to offer a bribe, and if so, the public official decides whether to accept it. The participants received a real monetary reward that was higher if a bribe was offered and accepted, to simulate the private benefits that accrue to the parties to a bribe transaction. A citizen who offered a bribe that was rejected suffers a monetary cost, a feature that is intended to capture the risks associated with offering bribes to officials who turn out to be honest. We also told the participants (truthfully) that at the end of the experiment we would make a monetary donation to a global NGO, and that the amount of this donation would be reduced by a fixed about every time a citizen-official pair in the game agreed on a bribe. This feature of the experiment was designed to simulate the cost of bribery to society. Players played multiple sessions of the game, sometimes as a citizen and sometimes as an official. Importantly for our purposes, although the game was played online (rather than in face-to-face interactions), players were told the nationality of their partner in each session of the game. So, for example, if an Italian were matched with a German, the Italian would know her partner was German, and the German would know her partner was Italian, even though they would not know much else about one another.

We found that the players’ willingness to pay bribes was substantially affected by the nationality of their partner. When in the role of the citizen, participants were more likely to offer bribes to public officials from nations that are perceived as having a high level of corruption. That may be because the structure of the game—like the structure of many real bribery transactions—is such that one only wants to offer a bribe that one believes will be accepted. But we also found that even when players were assigned the role of public official, they were much more likely to accept bribes from the citizen if the player in the role of citizen was from a country perceived as more corrupt. This finding is more striking, because the player in the public official role always earns more by accepting bribes. Indeed, the nationality of each player’s partner was a more important factor in determining that player’s behavior in the game than the player’s own nationality.

Of course one should be careful not to exaggerate the conclusions that can be drawn from this sort of experiment for real-world behavior. Nonetheless, we think these findings are consistent with the view that people’s sensitivity to ethical constraints is influenced by their perceptions of what the people around them view as morally acceptable. When playing the game with a partner from a country stereotypically perceived as more corrupt, players seemed to be less concerned about engaging in corrupt behavior—even though there were no external sanctions for being perceived as corrupt by the other player (again, these were anonymous online interactions) and even when the player, in the public official role, did not have to make any guess about what the other player was likely to do.

Speaking of stereotypes, we were also able to use our experimental set-up to assess the accuracy of participants’ stereotypes about the likelihood of corrupt behavior in different countries. In each game session, we asked the player in the citizen role to predict in advance whether their partner would accept the bribe, and we gave a bonus payment if the prediction was correct. We found that there are indeed consistent stereotypes about different countries’ tendency toward corruption: participants tended to have rather consistent views about which other countries’ players would be most likely to accept bribes. But perhaps more importantly, we found that these stereotypes were not very accurate. In particular, participants vastly overestimated the willingness of players from countries with a reputation for corruption to accept bribes.

In short, we find that, at least in our study, most people appear to be conditionally corrupt: They bribe willingly when matched with a partner from a country with a reputation for corruption—even though this national-level reputation wildly overstates the actual likelihood that players from those countries will behave corruptly in our game. More research is of course required to assess how robust these results turn out to be. But if this is right, it suggests that one of the most important ways to reduce actual corruption is to combat exaggerated stereotypes about the corrupt tendencies of particular countries or nationalities, and to emphasize that ethical opposition to bribery and other forms of corruption is in fact widely shared, even if actual rates of corrupt conduct do vary substantially across countries.