TRUST GAMES: Experimental evidence of how we feel about cheating

Nearly a third of people have a stringent notion of what constitutes cheating, which can reduce trust since they are more likely to expect to be cheated. At the same time, if others perceive these stringent cheating notions, they will be more reluctant to cheat. These are among the findings of research by Professor Paola Giuliano and colleagues, published in the September 2016 issue of the Economic Journal.

Their experimental study sheds light on how we decide whether to trust someone with whom we interact in our daily lives. For example, when taking a cab, we expect the driver to use a reasonably short route even if neither we nor the driver make explicit mention of it. Despite the lack of explicit promise, we may feel cheated if the taxi driver takes an unnecessarily long route.

But according to the experimental results, if taxi drivers anticipate that their customers would feel cheated unless they take the shortest route, they will indeed be more likely to minimise the length of the route they choose – even if, out of self interest, they do not take what is absolutely the shortest route.

The researchers note that many other situations involve some degree of trust and the fear that we might be cheated is a key element in making our decisions. For example, when we ask for financial advice, the advisers do not spell out that they will act solely in our best interest, but we may still judge cheating according to this metric.

Similarly, when we book a vacation through a travel agent, search for the best insurance at a broker or take our car to a mechanic, we may act on implicit notions of how the travel agent, broker or mechanic ought to behave, perhaps feeling cheated or let down when their behaviour fails to live up to these standards.

These notions can actually make us trust in different ways. But they are unobservable.

The new study finds a way to elicit people”s subjective, individual notions of what constitutes cheating in the context of a laboratory experiment known as a trust game. In the game, one person – the sender – sends money over to another person – the receiver. The amount is increased on arriving with the receiver who, in turn, can return all, some or none to the sender.

The research finds that people playing the trust game belong to two categories:

• The majority feel cheated if they don”t get a positive return on their investment (the amount they send to the receiver).

• But a substantial fraction of people – as large as 30% – feel cheated unless they get back half of all the money the counterpart receives. This latter group apparently operates according to a much more stringent notion of what constitutes cheating.

These notions of what constitutes cheating affect our behaviour in several ways. First, the more stringent is the cheating notion, the higher is the expectation of being cheated. Consequently, more stringent cheating notions may reduce the likelihood of trust.

At the same time, receivers” expectations about senders” cheating notions affect their behaviour. If receivers think their counterparts have stringent cheating notions, they will be more reluctant to cheat. Even when they do cheat, they do not cheat fully: rather they typically return something.

”Trust and Cheating” by Jeffrey Butler, Paola Giuliano and Luigi Guiso is published in the September 2016 issue of the Economic Journal. Jeffrey Butler is at the Einaudi Institute for Economics and Finance (EIEF) and the University of Nevada, Las Vegas. Paola Giuliano is at the University of California, Los Angeles. Luigi Guiso is at the EIEF.

Paola Giuliano