Strange dictators

An old game is turned upside down, and the results are surprising

An old favourite in behavioural science, often wheeled out to show that we are not ‘rational’, is the Dictator Game. According to standard economic theory, people should be acting to maximize their own economic gain. This is not the behaviour observed, and so the game can be seen as an indicator of our ‘irrational’ altruism or generosity.

The game, first introduced in 1994 by Robert Forsythe and colleagues at the University of Iowa, goes as follows. There are two participants, one of whom (the dictator) is given a sum of money. Initially, the second person (the receiver) gets nothing, but the dictator can decide to give away any proportion of the sum received, literally between 0% and 100%. The receiver must accept this amount and cannot intervene in any way.

It was a development of the Ultimatum game, first described in 1982, in which the receiver can reject the amount offered by the other player, the proposer. If that happens, the whole sum is forfeited, and no one gains anything. The big difference is that here the receiver has control over the outcome, but of course at her own expense. If the amount on offer is deemed too low, both parties lose.

A rational, self-interested, utility-maximizing person would act predictably in all roles in a one-off game. (If the game is repeated, things get more complicated.) In the Dictator game, the dictator would simply keep all the money. In the Ultimatum game, the receiver is better off with even the smallest amount, so that is what a rational proposer would offer.

This is what dictators do — source: Dictator Games, A Meta Study

Yet in practice, people rarely act in that way. Dictators turn out to be a lot more generous than expected. A metastudy by Christoph Engel of the Max Planck Institute analysed the data from more than 20,000 observations in 83 papers on the subject. While in six of the reported 616 treatments Dictators actually do give an average of nothing, over the whole range of observations the average is no less than 28.35%.

Another metastudy by Jean-Christian Tisserand and colleagues at the Université de Franche-Comté looked at both the Dictator and the Ultimatum game. They found that on average the dictators offered 25.6%. The proposers in the Ultimatum game gave on average 41.04% away.

The effect of fear of retaliation — source: A meta-analysis on the ultimatum and dictator games

For firm believers in homo economicus, the Dictator game’s results are the most baffling. While having nothing to lose, dictators give up more than a quarter of their endowment. The significant difference with the Ultimatum game can be explained by the threat of punishment by the receiver. But that, in turn, is no less mystifying on the part of the receiver, since rejecting any non-zero offer the receiver makes is, of course, an ‘irrational’ sacrifice.

One explanation for this behaviour is that people value other things than simply economic gain, in return for which we are prepared to make economic sacrifices. We appear to have an innate sense of fairness and an aversion for (excessive) inequality, which inspire us both as a dictator or a proposer, and as a receiver. In other words, we are prepared to pay for fairness.

At least, that is the case when a monetary gain is to be split between two people. But what if it concerned a non-monetary harm instead? This is the question Alexander Davis and colleagues at Carnegie Mellon University asked themselves in a recent paper.

And a rather intriguing question it is. The debate around altruism does indeed seem to focus mostly on the sacrifices (generally of money) made for the benefit of others. They very rarely look at a form of altruism that involves taking over a burden from another person, or suffering some form of harm to reduce the harm to someone else. Yet that is, as Davis and co describe, a very common form of altruism. In most families, members share relatively unpleasant chores like washing up or taking out the rubbish. Many people donate blood, and some of us put their lives at risk for the benefit of others in the emergency services or in the armed forces.

Now it is true that most instances of such altruism involve direct or indirect reciprocity. If we do a chore to reduce the burden on someone else, they might return the favour. One day we may need a blood transfusion, and then we’ll be glad someone else donated blood, so that might inspire us to do so ourselves.

What makes the research by Davis et al particularly interesting is the absence of such a reciprocity context to motivate altruistic behaviour. For their experiment they used tokens, each of which represented an obligation to submerge a hand into a container with ice water for a given number of seconds. The dictators were asked to divide a number of tokens between themselves and an anonymous receiver. Like in a conventional Dictator game, where they were able to keep all the money, they were at liberty to allocate all the tokens to the receiver and thus escape the unpleasant ice water experience.

The findings are remarkable. On average, the dictators kept more than 40% of the tokens, and thus took more than 40% of the harm. Overall, the dictators allocated more of the harm to the receiver in 47% of the cases. But they shared the harm equally in 44% of the cases, and in 9% they took more of the harm than they allocated to the receiver.

The experiment suggests that we are considerably more generous in the context of non-monetary harm than regarding monetary gain — even when there is no external motivation to do so. (Remember that in a conventional game the dictators kept 75–80% of the money.)

This is not easily explained. Is the discrepancy due to that between the gain and harm frame, or between the monetary and non-monetary frame? Do we have a different set of moral rules when we are concerned with the harm we choose to inflict on others, and when we are concerned with sharing our wealth? Perhaps further research will shed more light on exactly what is behind this asymmetry.

But that shouldn’t stop us from reflecting on the actual behaviour itself, irrespective of its motivation. We humans seem to have a deep desire to reduce the harm to others, even if this is to our own detriment. That makes us strange, even incompetent, dictators.

And that is a hopeful and uplifting thought, no matter why or how.

Originally published at on April 7, 2017.

Thank you for reading all the way to the end — if you enjoyed the story, please consider recommending it by clicking/tapping the ❤️ icon to the left. That brings a smile to my face, and more importantly, it signals it as worth reading to others. And of course do share it with your network via Twitter, Facebook or LinkedIn using the handy icons. Thanks!

Accidental behavioural economist in search of wisdom. Uses insights from (behavioural) economics in organization development. On Twitter as @koenfucius

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store