« You're a bastard | Main | Gerry Gunderson and the battle at Trinity College »

Handicappling the Clark medal

Justin Lahart reports:

Friday, the American Economic Association will present the John Bates Clark medal, awarded to the nation’s most promising economist under the age of 40.

The Clark is often a harbinger of things to come. Of the 30 economists who have won it, 12 have gone on to win the Nobel, including last year’s Nobel winner, Paul Krugman. Other past winners include White House National Economic Council director Lawrence Summers and Steve Levitt, of Freakonomics fame. Since it was first awarded in 1947, the Clark has been given out every two years, but beginning next year it will be given out annually.

With a deep pool of young talent to draw from, there’s no sure winner. But among economists, the clear favorite is Esther Duflo, 36, who leads the Massachusetts Institute of Technology’s Jameel Poverty Action Lab with MIT colleague Abhijit Banerjee.

Ms. Duflo has been at the forefront of the use of randomized experiments to analyze the effectiveness of development programs. If teacher attendance is a problem in rural India, for example, what happens if teachers are given cameras with date and time stamps and told to take a picture of themselves and their students each morning and afternoon? Ms. Duflo and economist Rema Hanna tried it out and found that in the “camera schools,” teacher absences fell sharply and student test scores improved. Does giving poor mothers 60 cents worth of dried beans as an incentive to immunize their children work? It works astoundingly well. By answering these kinds of problems, Ms. Duflo, her colleagues, and the many economists around the world she has helped inspire, are uncovering ways to make sure that money spent on helping poor people in developing countries is used effectively.

Harvard University’s Sendhil Mullainathan, who founded the Poverty Action Lab with Ms. Duflo and Mr. Banerjee, is also likely on the Clark short list. He’s a leading light in the fast-growing field of behavioral economics, studying ways that psychology influences economic decisions. For one paper, he and frequent co-author Marianne Bertrand sent out fictitious resumes in response to want ads, randomly assigning each resume with very African American sounding or very white sounding names. The resumes with the very white names got far more call backs. Mr. Mullainathan, 36, is also applying behavioral economics insights to development problems. One insight: The behavioral weaknesses of the very poor are no different than the weaknesses of people in all walks of life, but because the poor have less margin for error, their behavioral weaknesses can be much more costly.

Emanuel Saez at the University of Calif.-Berkeley, another Clark candidate, has been tenaciously researching the causes of wealth and income inequality around the world, with a focus on the what’s happening at the very tip of the wealth pyramid. But because there is very little data on the very rich, Mr. Saez, 36, and his frequent co-author Thomas Piketty have combed through income tax figures to come up with historic estimates. Among their findings: That before the onset of the financial crisis, the income share of the top 1% of families by income accounted for nearly a quarter of U.S. income — the largest share since the late 1920s.

Posted by Tyler Cowen on April 23, 2009 at 12:08 PM in Economics | Permalink

Comments

I would like to handicap the odds that any of these people, who I am sure are fine human beings, will make any contribution whatsoever to reducing poverty.

Posted by: josh at Apr 23, 2009 12:27:32 PM

josh - what's your opinion of Duflo's work then? That is the focus of everything Duflo does - focusing on things that work at the micro level.

Posted by: jason voorhees at Apr 23, 2009 12:33:35 PM

Is the test/control poverty work economics?

Posted by: bastiat at Apr 23, 2009 12:41:34 PM

I've covered some of Mullainathan's work in my undergrad econometrics class, and I'm a fan of his. I wouldn't mind seeing him win.

Posted by: Colin at Apr 23, 2009 12:49:06 PM

I am confused. Duflo and Saez are both French. Isn't the medal given only to Americans?

Posted by: John at Apr 23, 2009 1:10:30 PM

That's an interesting thought of Mullainathan's, that the weaknesses of the poor are no different from those of others, but they have less margin for error. It needs a little more definition, though: do we mean that their weaknesses are no different in kind, or no different in degree? Unfortunately, the linked paper is very thin--really no empirical data at all, just vapid rhetoric. It sounds like a well-meaning social worker speaking to her friends at lunch.

Posted by: y81 at Apr 23, 2009 1:10:42 PM

Those are the three candidates for the award? I dont want to denigrate their work in any manner what so ever but one would hope that a medal with such prestige would be awarded for breakthroughs of a bit larger magnitude. The bottom two are especially blah.

Posted by: John Pertz at Apr 23, 2009 1:12:10 PM

Duflo's citizenship may be in France, but her employment is at MIT, and she is a member of the AEA. I think that's all it takes.

Posted by: jason voorhees at Apr 23, 2009 1:15:34 PM

Is the test/control poverty work economics?

==

It's probably easier to answer this question if you first define "economics."

Posted by: jason voorhees at Apr 23, 2009 1:16:01 PM

Re: randomized trials. I once heard someone quip that if the profession moves more towards original data collection and randomized trials, then the gap between the haves and the havenots will grow even more. A professor at Harvard or MIT can afford to send 10 graduate students into Ugana, traipsing around interviewing and running experiments. A professor at any other university likely cannot.

That is of course independent of the scientific value, but just wanted to make a note.

Posted by: jason voorhees at Apr 23, 2009 1:17:39 PM

Looks like a strong bias towards MIT PhDs in the recent history of the bates medal...3 of the 5 in the past 10 years, with Duflo and Saez also alums

Posted by: md at Apr 23, 2009 1:19:02 PM

I think Duflo and Saez are no less American than Acemoglu was when he won in 2005...while neither Saez or Duflo have US citizenship (according to their CV's), they both have at least been in the U.S. since they started graduate school (which is not true for Acemoglu).

Posted by: Nick at Apr 23, 2009 1:40:14 PM

To bad we dont' have a $10 million dollar prize for young economists who are helping us to understand how economics can possibly become a conceptually sound and causally powerful explanatory enterprise -- because economics so far isn't that, and it needs all the help it can get.

And we need a $1 billion dollar prize for economic departments which are making an effort to develop graduate programs which give young economists the competence to fix and improve economics as a global causal explanatory enterprise, something most of them now don't do.

Posted by: Greg Ransom at Apr 23, 2009 2:22:14 PM

Mullainathan deserves to win it, but I don´t like the change to giving the prize anually.
The Nobel prize is running out of people who deserve to be awarded, it will happen the same thing to the Bates Clark ( Nobel is even worse because it´s usually shared among 2/3 economists)

Posted by: Bruno at Apr 23, 2009 2:28:26 PM

“Ms. Duflo and economist Rema Hanna tried it out and found that in the “camera schools,” teacher absences fell sharply and student test scores improved.”

So it takes PhD and years of research to prove that introducing efficient controls leads to better results?

Is the true reason for nomination a feel-good topic of the research?

Say, they'd found that in businesses with time-stamping machines installed number of workers coming late fell sharply and profits improved. Would they even be considered for award?

Posted by: Ozornik at Apr 23, 2009 3:13:16 PM

These picks are completely biased towards reduced form empirical work without much theoretical motivation. These economists are great at what they do but there are many other (less glamarous) fields which deserve as much if not more recognition. The author of the article would certainly not have written about susan athey winning in 2007 because her work is complex and harder to explain to a naive WSJ reader.

Posted by: abcdefg at Apr 23, 2009 3:40:56 PM

Say, they'd found that in businesses with time-stamping machines installed number of workers coming late fell sharply and profits improved. Would they even be considered for award?

==

If the study had internal validity, it would almost certainly get published.

Posted by: jason voorhees at Apr 23, 2009 3:58:27 PM

A couple things:

~Non-US citizens are eligible for the Clark medal. David Card, for example, is Canadian, and Acemoglu is Turkish, to name just two.

~Duflo is extremely worthy, despite the "over focus on empirics by the AEA"-type argument. There are lots of young economists doing theoretical research and being applauded by the AEA (and will probably win the Clark sometime soon). Raj Chetty is just one example. Duflo does empirics, yes, but it's really made development economics what it is with randomized control trials.

~Sadly, Rob Shimer should win (because Duflo is 36 and can win again, and because he's the agenda setter in macro/labor) but won't (because I think he just turned 40 in the past few months).

Posted by: Andrew at Apr 23, 2009 5:59:42 PM

~Sadly, Rob Shimer should win (because Duflo is 36 and can win again, and because he's the agenda setter in macro/labor) but won't (because I think he just turned 40 in the past few months).

==

John List also fits this "sadly" bill. He is too old to win it, too. And to circulate something I've heard a million times by now, it's that he lost to Athey that led the committee to move to an annual prize instead of bi-annual.

Which I think is good, but what's the optimal number of prizes anyway? The profession is much bigger than it was when they started this, so presumably once every two years does not mean what it meant years ago. But what is the optimal prize? Are these really incentive mechanisms anyway?

Posted by: jason voorhees at Apr 23, 2009 6:43:12 PM

he and frequent co-author Marianne Bertrand sent out fictitious resumes in response to want ads, randomly assigning each resume with very African American sounding or very white sounding names. The resumes with the very white names got far more call backs.

What's this?? I thought market competition had abolished racial discrimination in employment - that all those nasty racists had gone bankrupt. Hmm. Need to think again.

Posted by: Bernard Yomtov at Apr 23, 2009 7:41:59 PM

he and frequent co-author Marianne Bertrand sent out fictitious resumes in response to want ads, randomly assigning each resume with very African American sounding or very white sounding names. The resumes with the very white names got far more call backs.

I don't know what this study has to do with racism. Why is "Greg Baker" or "Emily Walsh" a "very White sounding name"? Here is a black man named "Gregory" and here is a

Before I believed that the discrimination that Mullainathan and Bertrand found had anything to do with racism, I would need evidence that people named Lakisha were equally represented among the rich and the poor, or at least, among rich blacks and poor blacks. I suspect the effect may be a discrimination against the poor and against those who deviate from the preferred naming practices of the elite. Both of things seem bad to me. They both seem like social problems, but they don't seem like racism. Mislabeling them as such can only hinder work to eliminate discrimination.

Posted by: Curt Fischer at Apr 23, 2009 8:00:54 PM

he and frequent co-author Marianne Bertrand sent out fictitious resumes in response to want ads, randomly assigning each resume with very African American sounding or very white sounding names. The resumes with the very white names got far more call backs.

I don't know what this study has to do with racism. Why is "Greg Baker" or "Emily Walsh" a "very White sounding name"? Here is a black man named "Gregory" and here is a black woman named "Baker". I wonder how often people asked them, "Hey, what's up with your White sounding name? That's crazy!" Or maybe people tell them, "Oh, Gregory, it's a good name your last name is Royal instead of Baker, otherwise I'd've thought you were White!"

Before I believed that the discrimination that Mullainathan and Bertrand found had anything to do with racism, I would need evidence that people named Lakisha were equally represented among the rich and the poor, or at least, among rich blacks and poor blacks. I suspect the effect may be a discrimination against the poor and against those who deviate from the preferred naming practices of the elite. Both of things seem bad to me. They both seem like social problems, but they don't seem like racism. Mislabeling them as such can only hinder work to eliminate discrimination.

Posted by: Curt Fischer at Apr 23, 2009 8:01:37 PM

A Pareto optimal improvement would be to bomb Berkeley's economics department into the Stone Age.
Is there anyone there who isn't a left-wing tax-consuming thief?

Posted by: AADL at Apr 23, 2009 9:19:57 PM

"So it takes PhD and years of research to prove that introducing efficient controls leads to better results?

Is the true reason for nomination a feel-good topic of the research?

Say, they'd found that in businesses with time-stamping machines installed number of workers coming late fell sharply and profits improved. Would they even be considered for award?"

(a) It isn't always true that introducing efficient controls leads to better results. Whether or not a given control that appears to be efficient, and may have proved efficacious in other contexts, actually improves attendance and productivity depends on other cultural and cognitive factors at work in the particular system.

(b) Whether or not a given control is efficient is problematic even when other variables are controlled for. Many controls that common sense suggests are efficient (for example time-stamping arrival at work) can result in camouflaging behaviour that reduces productivity, retention and attendance more than if no such control existed.

Whether or not the time stamped photograph method was effective in Indian schools is a legitimate research topic. Without an empirical project it would be difficult to argue theoretically one way or the other. There are many examples of enormously expensive IT project failures that foundered on such simple false assumptions.

Posted by: Phillip Capper at Apr 23, 2009 10:17:53 PM

Before I believed that the discrimination that Mullainathan and Bertrand found had anything to do with racism, I would need evidence that people named Lakisha were equally represented among the rich and the poor, or at least, among rich blacks and poor blacks. I suspect the effect may be a discrimination against the poor and against those who deviate from the preferred naming practices of the elite. Both of things seem bad to me. They both seem like social problems, but they don't seem like racism. Mislabeling them as such can only hinder work to eliminate discrimination.

==

They tried to control for socio-economic status by varying the address of the resumes.

Posted by: anony at Apr 23, 2009 10:50:08 PM

Post a comment