Paul Zak reports on measuring the level of the hormone oxytocin in experimental subjects after they have played a game involving trust and co-operation.
People were recruited and paid $10 for showing up. Then they took seats in a large computer lab and were matched up in pairs, but this done completely anonymously so that no one knew (or would know) the other person in his or her pair. One-half of the participants (decision-maker 1s) then had the opportunity to send none, some or all of their $10 show-up fee to the other person in their pair. Whatever is sent is tripled. So, if $4 was sent, the other person would have $22 ($4 tripled, plus the $10 show-up fee the second person receives). The second decision-maker could then send some amount of this money back to decision-maker 1, but need not. This is how we produce a social signal of trust: decision-maker 1’s only reason to transfer money to the other person is because he or she trusts that that person will understand why the money is being sent to them, and in turn will return some to them (be trustworthy)…
The Nash equilibrium says that if you are decision-maker 2, you should prefer more money than less so you should not be trustworthy (that is, return any money to decision-maker 1). Decision-maker 1 should realize this and therefore never send anything to the second person. Yet we see abundant trust in the lab and in daily life.
I am skeptical of these laboratory studies. I think that human behavior in situations with a strong personal component is different from human behavior in an impersonal market. I see economics as studying behavior in the context of markets.
For Discussion. If there is a hormone that increases trust and co-operation, would raising the level of that hormone be good for people?
READER COMMENTS
David Thomson
Aug 18 2003 at 12:28pm
I doubt that we will ever be able to prove this one way or another. At best, we will come up with tentative theories—that may later seem ridiculous.
Arnold Kling might also want to reread Aldous Huxley’s “Brave New World.” I think we best think twice before swallowing pills to allegedly enhance our lives.
Eric Krieg
Aug 18 2003 at 12:59pm
Yeah, only a pot smoking hippie would think that the answer to the world’s problems is more “cooperation”.
Give a test subject that hormone and put them out on the streets of downtown Brooklyn for five minutes. They would be a victim of a con-artist or some other criminal in that time.
There are times when skepticism, unfriendliness, and downright misanthropy are called for. It is bred into New Yorkers as a defense mechanism.
Paul Zak
Aug 19 2003 at 6:14pm
The reason we run laboratory experiments, and indeed neuroeconomics experiments, is that theory does not predict observed behavior, and subjects themselves cannot tell us why they do what they do. Thus we need some other type of measurement, in this case, hormone measurements. This is instructive as it gets to the deep mechanisms generating behavior, and show us, e.g. that trustworthy behavior is driven by largely unconscious physiologic reactions to social signals from others. Economics is about the efficient allocation of resources, both in personal and impersonal markets. Indeed, the bulk of transactions we engage in have another human on the other side of the counter. The work of my lab seeks to bring humans back into social science rather than assume “economic” agents. By the way, we find no evidence that subjects are guillible when their oxytocin levels are high; indeed our subjects who were trusting on average made more money than did those who followed the “rational” economic prediction. My work does not advocate that human ingest drugs to make them cooperative, but identifies the deep physiologic factors that produce trustworthiness.
Bernard Yomtov
Aug 20 2003 at 10:06pm
Paul,
OK, that makes sense, but there are still pretty serious questions about these sorts of experiments.
The most obvious is whether they are “scaleable,” to use a buzzword. That is, why should I believe that people making decisions involving thousands of dollars, or more, behave the same way as undergraduates playing games involving tens of dollars?
And, again assuming that the subjects were college students, aren’t there significant social influences on behavior to suggest that maybe the results don’t carry over?
Comments are closed.