The difficulties involved in splitting a restaurant bill are the stuff of legend or, at least, the stuff of Life, the Universe and Everything by Douglas Adams. Adams introduced the concept of Bistromathics and the Bistromathic Drive: the mathematics generated by people’s odd behaviour in restaurants is apparently so powerful and strange, he declared, that it can fling a starship across a galaxy in seconds.
Bistromathics emerge from the interaction of two powerful forces in social science: freeriding and social pressure. Eating out is a social occasion and your behaviour will affect how others see you. Yet the opportunity for freeriding is obvious: if six diners split the bill equally, the additional cost to any one diner of ordering that £12 starter will be just £2, with his companions picking up the balance. The split-the-bill rule offers an excellent opportunity to try the fancier options at little extra cost. Of course everyone else is likely to reach the same conclusion – but then they’d be a fool to do otherwise.
Do people really freeride in this way? A large number of laboratory experiments, typically involving students sitting at computers and interacting anonymously, have found that we freeride less than textbook economic theory would predict. We care about other people, it seems – either about their wellbeing or what they think of us.
But there’s an alternative explanation for this curious outbreak of human decency: perhaps people are just confused by the laboratory setting and the laboratory games. There’s some evidence, for instance, that people in these free-rider games start to take advantage of each other after a few practice runs.
In 2004, three researchers, Uri Gneezy, Ernan Haruvy and Hadas Yafe published an intriguing study in The Economic Journal. Gneezy and his colleagues had conducted an experiment in a real restaurant, in Haifa, Israel, inviting groups of six diners to order food – but giving them different systems for settling the bill.
The idea wasn’t to simulate the real restaurant experience, when one dines with friends and can expect social consequences for good or bad behaviour: the researchers had invited complete strangers to eat together, ostensibly for the purposes of researching the effect of food on emotions.
What was the point of creating such an odd dining experience? Instead of trying to understand how people behave in restaurants, the researchers were trying to use the restaurant as a better kind of laboratory. Laboratories can be confusing, but everybody understands that if you’re sitting in a restaurant and you’ll be splitting the bill, you have an incentive to order the lobster.
Diners were, at random, offered three different billing rules: split-the-bill, pay-your-share, or on-the-house. They were also asked to order food by writing their choices down, without discussion. This odd request was made less odd by the fact that they were all filling in questionnaires at the time.
Homo economicus immediately emerged: diners ordered, on average, 37 shekels worth of food when paying their own way, 51 shekels when splitting the bill, and 82 shekels when the experimenter picked up the tab for everyone. (A small follow-up experiment hinted that people splitting the bill six ways behave similarly to those paying one-sixth of their own bill.)
We freeride in restaurants, then – at least, we freeride in the odd circumstance of having lunch with total strangers and filling in questionnaires. But there’s a twist to this: the experimenters asked, out of curiosity, whether people would prefer to split the bill or pay their share. Most people would have paid their share, given the choice. And yet in social settings, splitting the bill remains the default option. Is there a deeper or more fascinating subject than Bistromathics?
Tim Harford’s new book, “The Undercover Economist Strikes Back”, is published by Little, Brown