Editor’s Note: Kay-Yut Chen, a leading experimental economist at HP Labs, was the subject of a March 2006 Scientific American profile written by Marina Krakovsky. Chen and Krakovsky recently collaborated on a just-released book, Secrets of the Moneylab: How Behavioral Economics Can Improve Your Business, about the practical lessons from research by Chen and fellow behavioral economists. In the following edited excerpt, Chen and Krakovsky explain why we sometimes reject deals that we might be better off accepting.
Have you ever rejected an unfair deal? Doing so might seem perfectly sensible. Unfortunately, our fairness radar often activates not because we’re getting a bad deal, but only because we compare ourselves with some reference point and find ourselves short.
Consider this incident from Kay-Yut’s lab. He and his colleagues have run hundreds of experiments over the years, and most of the time things go smoothly: participants show up, play for a few hours, and get paid for their efforts. But sometimes things don’t go as planned. One day, a local student was one of several people who came in to participate in an experiment.
The researchers made clear to all the participants that they’d get $25 for showing up, but that how much they took home depended on how well they played the game. As a result, some people would end up with more, others with less. And no matter how badly they played, they wouldn’t have to pay the lab out of their own pockets. Their losses would come out of the $25 they started with. Sounds fair, right? But this particular student did so badly that at the end of three hours of play she came away with only $1.35. To her, that didn’t seem fair. Even though she didn’t gamble away her own money, she felt that her time was worth more than $1.35. She was upset enough to complain to Kay-Yut and, eventually, to his boss. The two stood by the rules. Doing otherwise wouldn’t be fair and would change the dynamics of future games. But the researchers learned a lesson: people compare against a benchmark (whether it is the minimum wage, the $25 baseline, or what most players were earning in the same game). So when people ask how much others earn in the same game, the researchers now give a stock answer that doesn’t reveal much except an unwillingness to invite comparisons.
When in 2005 Washington Ballet dancers rejected a deal with the ballet company’s owners over a planned Italian tour, it was because the dancers—knowing that the State Department per diem is $150—felt shortchanged by a per diem offer of $50. Would their reaction be different if they’d learned that other touring groups were getting a $50 per diem? There’s good reason to believe they’d be happier, and here’s why. It’s hard to know how much it will cost to eat in Rome or what quality of meals to expect an employer to subsidize; it’s much easier to look around and see what somebody else is getting. When we get more than that, we feel we’re getting a decent deal; when we get less, we feel cheated.
But there’s another issue that makes it hard to reach agreement: when people in a negotiation can choose which group to compare themselves with, they’ll make a self-serving choice—and, of course, that choice is likely to be different for the two parties in a negotiation. For example, Carnegie Mellon economists Linda Babcock and George Loewenstein surveyed presidents of school boards and teachers’ unions all over Pennsylvania to see which districts the two groups considered comparable to their own for purposes of salary negotiation. This is the kind of information both sides use to set their targets for a fair negotiation outcome, much as the Washington Ballet dancers used data from the State Department. The teachers tended to list districts whose teachers earned salaries higher than theirs, while the school boards listed districts with lower-paid teachers. The average salary difference between the two groups’ numbers—$711 per year, or 2.4 percent of the average teaching salary—is huge when you consider that the average raise at the time was less than 5 percent per year.
For example, suppose teachers in one district earn $35,000 on average. They compare themselves with teachers in a district where the average salary is $36,700—a number that would require nearly a 4.9 percent raise to reach. Instead, they ask for a “modest” 4 percent raise, which would bring their salaries up to $36,400—not quite as high as they’d like, but in their view a decent compromise. The school board, on the other hand, sees this as an unreasonable demand because the board’s reference point is a school district with an average salary of $36,000 (or $700 less than the teachers’ reference point). So the school board offers $36,200, which it sees as “more than generous.” The result: although both sides made concessions, they still couldn’t reach agreement—and each thought the other side was making unfair demands. Indeed, when the researchers analyzed the data, they found that these biased judgments of what’s fair were leading to teachers’ strikes. For example, strikes were more likely to occur in districts with more varied districts around them, where the gap between teachers’ and school boards’ perceptions of a fair comparison was wider.
Excerpted from Secrets of the Moneylab: How Behavioral Economics Can Improve Your Business by Kay-Yut Chen and Marina Krakovsky by arrangement with Portfolio, a member of Penguin Group (USA), Inc., Copyright (c) Kay-Yut Chen and Marina Krakovsky, 2010.