Tuesday, September 22, 2009

The origins of behavioral economics: Kahneman interview

Below is a great interview with Daniel Kahneman: Intuition and Rationality (full transcript).

Some years ago I wrote about Choices, Values, and Frames. See here for a nice discussion of behavioral economics.

Obvious to psychologists, difficult for economists. (See here for discussion by Robert Shiller, whose wife is a psychologist.)

Then when we worked on decision making, which was the second thing we did together, we published the theory paper in Econometrica, which is the prestige theory journal in economics. We were not intending to influence economics at all. It was just the best journal for that kind of article. If we had published exactly the same article in Psychological Review it would not have got the Nobel Prize, it [would have been seen as] a trivial detail. If we'd had a hostile referee in Econometrica we would have gone to Psych Review, but the big event that caused the prize is that some economists were interested, particularly an economist who's now my closest friend, Richard Thaler, who became interested in our work. He is the one who both taught me economics and brought these ideas into economics for their behavioral [value.]

So, it's because of the work of economists that I got the prize. I didn't get the prize for work that Amos and I did just because it was very good work. We got the prize (or I got the prize, because he couldn't share in it) because it had an impact on economics; but the impact was because people in economics saw it was relevant. They were a minority of people, and they did brilliant work.

The complexities of life:

It must have been the fall of '41, when there was a curfew for Jews. We were also supposed to be wearing a yellow star, and there was a curfew which I think was 6:00 PM. I was in first or second grade and I'd gone to play with a friend and I was going home and I missed the curfew, I was late. And so, I turned my sweater inside out and walked home, and as I was coming close I remember the street was deserted and there was this German soldier walking towards me. He was wearing the black uniform and I knew that was not good. That was the uniform of the SS. We were walking towards each other and as we were coming close he sort of beckoned me, and of course I went there, and he picked me up and hugged me. I remember being terrified that he would see the Star of David inside my sweater. Then he put me down and took out his wallet and showed me a picture of a boy and gave me some money. That's a formative memory because of what it meant about the complexity of things. I remember being very fascinated at the time by this and by stories of Hitler liking flowers and kissing babies. The complexity of evil was much on my mind as a seven- or eight-year-old.

[While traveling, missing my kids, I find myself smiling goofily at every 3-5 year old I see!]

Psychology research for the Israeli army:

One of the things I did do that I'm still quite proud of is I set up an interviewing system for the Israeli army to interview recruits for combat units. I learned in 2002 [that] that interview is still in place, fifty years later. ... In 1954 Paul Miel published a very important book in which he showed that clinicians are much less good at what they're doing than they think they are, and that their ability to forecast events and psychological behavior in the future is really quite limited. ... I set up an interview that didn't leave a lot of room for clinical intuition on the part of the interviewer. The interviewer went through a script and found out what the person had done as a civilian, basically. It turns out that when you use that information to produce ratings of various strengths, you get ratings that are predictive and are useful on whether somebody is responsible, or appreciates manliness, or is sociable, things like that. [Personality factors are stable and useful! Expert prediction is weak; algorithms beat experts.] In addition, it turned out (and this was important to me years later) that when people had acquired that information, then their clinical judgment was, in fact, good, whereas if they were trying to form a clinical impression [without a script] they couldn't do it.

These observations that I made at that time became quite essential to my research about fifteen years later and they were central to the work that ultimately led to the Nobel Prize. So, it started quite early, by accident.

[Note added: (September 2010) I met Kahneman at Scifoo this year and asked him briefly about his personnel selection experience. He seemed quite proud that the Israeli military is still using his methodology today.]

Advice on research:

My main advice would be to look at the natural traps that await anybody who does research. One of the real traps is to get trapped by a research program that is fundamentally uninteresting, and also, the reluctance to cut your losses so that when you've done something that is not very interesting, wasting time trying to publish it is almost always a big mistake, at least in my judgment. Look for things that are worth doing and discard things that don't work, and know that if you're reasonably good at what you're doing, and otherwise you shouldn't be in that profession, but if you're reasonably good at it, then ideas are a dime a dozen. You'll get lots of ideas. You need some ideas that work and that keep you excited and that you can do something with. And so, not getting stuck on an idea that doesn't work, but looking for other ideas -- that would be my general advice. That certainly has been how I've operated.


CW said...

In the title: "inverview"?

(Feel free to delete this.)

CW said...

More: It appears that "Amos Tversky" should be replaced by "Daniel Kahneman" in this post. (They were collaborators, of course.) Am I missing something?

CW said...

In the third excerpt I believe "Paul Miel" should be "Paul Meehl". I realize the error was in the Berkeley transcript.

Steve Hsu said...

Chris -- Oops! Thanks -- I must still be jetlagged :-)

Tim said...

It's an interesting interview, but it mostly left me wishing it actually was with Tversky.

Amusingly, this post is now the first result for a Google search on "interview Tversky".

Ian Smith said...

"Choices, Values and Frames" should be "Choices, Values, and Frames". Another non-arbitrary grammatical rule violated.

Steve Hsu said...

Re: "Choices, Values, and Frames", thanks -- fixed now!

Kathkenne said...


Tks very much for post:

I like it and hope that you continue posting.

Let me show other source that may be good for community.

Source: Behavioral interview

Best rgs

Blog Archive