Esther Duflo wins the John Bates Clark Award

A great pick, here is one account.  Enter her name, or "Poverty Action Lab," into the Google search function on MR (left hand side of the page), for much more material on her work.  In my view her chance of winning a Nobel Prize is very high.

Comments

Not taking away from Duflo, who I think is a worthy recipient, but she may have gotten it over some other close rivals due to the perception that John List got the shaft in not getting (and no longer eligible). The sign here is that her work involves field experiments, and List was the person who introduced the modern movement to doing those with a paper in JEBO back in the 90s (there have been many field experiments in the past, although not labeled as such, such as the old time and motion studies in work places).

BTW, the usual collection of ignorant and hysterical morons over at econjobrumors are whining about this award, complaining as they did with the Nobel for Elinor Ostrom that it is all "afffirmative action." I guess we should forgive them because the lousy job market this year has a lot of them crankier than usual.

I agree with Arsene...Duflo's work is quite different in nature from List.

In any case, Duflo was the obvious choice to win this year. Theory per se remains important, but quality large-scale empirical work on important questions is so rare that those who did it right (Duflo, Fryer, Mullainathan, etc.) are justly rewarded.

Arsene, Well, I would say that "field experiments" is a broad category that includes
a variety of methods and approaches. What you describe fits the original 1998 paper
by List and Shogren, but a good overview of the broader set of types can be found
in the List and Levitt's JEP paper. Anyway, RCT can be viewed as part of this broader
category, but certainly her approach is different from what List did in 1998 and is
important work. I am not at all knocking her award.

Any references to practical impact of all the trials? For example, I heard that they gave away rice or lentils in return for immunization and that worked well. Has this been standardized?

Overall I am highly skeptical about field experiments as its so easy to fake results. In a vast country like India, who is going to take the trouble to drive for hours over horrible dusty roads in hot conditions to the 10 villages to verify your data?

An excellent choice.

DuFlo won; List did not. Give her credit for amazing work in development. She may save more lives than 90% of economists, combined. Glad to see the trend (w/Olstrom) of awarding people who do applied (=useful) work.

Prize goes to people working in the US, not just Americans. The winner last year was also a Frenchman.

She may well get the Nobel Prize but no cigar from any real economist for work which is either not economics at all or bad economics.

RCT do not adjust well for general equilibrium. I understand that the profession would be enthusiastic about her work at this point, but within ten years it will be relegated to the dustbin of failed empirical strategies.

The GE concern is a real one. Simply put, sometimes GE effects are a valid concern and other times, not so much. A case when it is a genuine concern would be estimating the impact of an additional year of schooling on wages: one person may get a 10% wage boost from an additional year but if the entire population gets more educated, it's not obvious whether everyone's real wage would rise or not.

But when it comes to immunizations, deworming pills or microfinance, it's not clear concerns over general equilibrium are enough to negate the results of relevant RCT research. If you have such an argument or a suggestion for a better empirical strategy, please share.

Barkley is correct. I think that it is clear that Duflo won because List didn't. His non-award caused the committee to give the award every year and that helped Esther get this one. Not only because it is every year, but because field experiments (List) got the shaft a few years ago.

RCT is a subset of what List has pioneered. Randomization was a part of the JEBO paper that Barkley cites of List in the 90s--before Esther was even in grad. school.

I’m only commenting on the development economics angle. Since you can’t eliminate the aid industry overnight, the results of RCTs in development economics might at least nudge the aid industry into more useful field operations, e.g. in health. I’m not suggesting it’s all just reinventing the wheel. However a generation of students might come to view microscopic empirical studies as substituting for political or sociological economics that stand a chance of eliminating poverty across the board. Be honest now -- can poverty action technicians in rural South Asia learn anything significantly more helpful from new randomized control trials and field experiments than European and North American countries did from their *200-year* trials and experiments with the markets, technologies, and institutions that produced the health systems, multi-gendered parliaments, and taxation policies that we enjoy today? It makes me wonder to what extent this is supply-driven within economics (equivalent to iPads generating output and employment). If there were demand for development knowledge, then transfers of the 200-year stock would be productive. That is unless poverty reduction today in a variety of geographies and cultures is totally unlike yesterday in a variety of geographies and cultures?

Maybe I'm confused DG Lesvic, are you saying that unless work can be applied "all people at all times and places" then it is not economics? This is such a bold pronouncement I assume I must be incorrect. Wouldn't that basically eliminate *all* empirical work? And if economics is entirely divorced from empirical work, from the real world then (a) what is the point and (b) how is that not an even larger game played by academics? Not trying to be snarky, trying to figure out what you mean.

In other news, I'm embarrassed that I'd never heard of Duflo until today. Very cool work.

Mike,

According to Ludwig von Mises,

Theories of human action “are, like†¦logic and mathematics, a priori†¦not subject to verification or falsification on the ground of experience and facts†¦both logically and temporally antecedent to any comprehension of historical facts†¦a necessary requirement of any intellectual grasp of historical events. Without them we should not be able to see in the course of events anything else than kaleidoscopic change and chaotic muddle.†

“Action and reason are congeneric and homogenous†¦two different aspects of the same thing. That reason has the power to make clear through pure ratiocination the essential features of action is a consequence of the fact that action is an offshoot of reason†¦Logical thinking and real life are not two separate orbits. Logic is for man the only means to master the problems of reality. What is contradictory in theory is no less contradictory in reality.†

“There is no means of studying the complex phenomena of action other than first to abstract from change altogether, then to introduce an isolated factor provoking change, and ultimately to analyze its effects under the assumption that other things remain equal.†

While, to Milton Friedman, the test of an economic theory was how it worked out in practice, to the man he read out of the science, Ludwig von Mises, such tests could never be conclusive, for there was always the question of whether concurring events were cause and effect or coincidence.

“The question whether there is any connection between them can only be answered by† a theory “established beforehand on the ground of aprioristic reasoning†¦If there were no economic theory†¦economic facts would be nothing more than†¦unconnected data open to any arbitrary interpretation.†

Since they are all complex, “Every historical experience is open to various interpretations and is in fact interpreted in different ways†¦History can neither prove nor disprove any general statement.†

That is not to rule out empiricism altogether. For the basic premises of economic theory, such as the disutility of labor and variety of resources, are derived from observation. But the theory itself is antecedent to all other historical facts.

JT,

Since everything I have given you above is right out of Mises and Rothbard, whatever you're saying about me you're saying about them.

Barkley,

While I realize experiments including randomized ones have been around a long time, what you say might be construed to suggest (a) why their policy impact at government level has not been enormous, and (b) that their policy ambitions are now greater.

I certainly don’t mean to be negative about the prize or recipient, only to note a trend with implications in a field where policy impacts on human futures and to worry a little about it. I restricted myself to development economics because that’s my training and because there’s a history of fashions in development economics often labeled paradigm shifts.

Some now claim such a shift on behalf of RCTs, which is fairly ridiculous even in spite of the potential to ‘scale up’ results. RCT is a marginal revolution in the study of economic impacts and social cost-benefit of marginal interventions like clean water or dams for irrigation. Leaving aside fungibility, it can inter alia show the positive or negative effects of aid financing to public subsidies for deworming of intestinal parasites. (I’m not getting into a host of well known problems, for example replicating results and possible confirmation bias in relation to results unfavorable to aid or in incentives to study ineffective projects.)

But -- big but -- as far as I know the nearest RTCs come to being transformational on a macro-developmental level is in providing evidence that small scale business enterprise works as a force for development. Even though this insight should be obvious, its empirical confirmation apparently helps convince investors that there will be a profitable return.

BTW, Vernon Smith demonstrated more decisively than previously by lab experiment what many believed true by virtue of pure logic and theory, namely that markets approach equilibrium, and that constructivist rationality is far from uniform or universal. Actually that only confirms my bias if favor of big political economy ideas (not necessarily requiring or allowing experimental verification in microeconomics) that are not necessarily new but need re-branding.

Michael,

Astrology, numerology, and phrenology may have a greater impact on gullible minds than reason, science, and economics, but that doesn't make them scientific. None of your experiments, whatever you call them, nor Vernon Smith's either, are economics. They are all distractions from it. Claim whatever you will for them, they are not economics, nor substitutes for it. There are none. Economics is nothing less than the science of civilization, and civilized men ignore it at their peril. As Mises stated, "if they fail to take the best advantage of it and disregard its teachings and warnings, they will not annul economics; they will stamp out society and the human race."

Comments for this post are closed