Guest post by Jeff Mosenkis of Innovations for Poverty Action.
There have been a few related things over the past couple months that all get at this tension between how applied/practical vs. theoretical policy-related research should be:
- At the AEA Ely Lecture in January, Esther Duflo suggested economists should be more like plumbers, tinkering and adjusting, concerned with the details. For example, whether a voucher system works can really come down to the nitty gritty. The theoretician focuses on whether a voucher is a good idea, but in practice, lots of tiny decisions, like how the voucher is distributed, and what kinds of information the designer chooses to put on the voucher can make the difference between whether the program works or not. Papers rarely even discuss those kinds of day-to-day details of a program, but she argues if the field wants to make a difference, details matter. Video and print versions available here.
- Beatrice Cherrier puts it into historical context, including the “physics envy” some use to describe the fields’ march towards more complex mathematical models in recent decades.
- Side note: In Cherrier’s interview on the Economics Rockstar podcast she talks about tracking how the MIT model of highly quantitative economics came to be so popular in the U.S. along with using The Wire as a teaching tool.
- For another historical alternative model of how to think about economics, see The Economist’s article on the history of the Cambridge school of economics. The thinking there was less concerned with mathematical models and more with training economists who’d understand the social and political contexts in which their work would be used.
- In Nature, Duncan Watts asked “Should social science be more solution-oriented?” He cites an organizational scholar’s likening of that field to the Winchester mansion in California, based on a dream the rifle company heiress had:
Because the dream didn’t specify any particular plan for the house, however, she embarked on an open-ended construction project in which hundreds of rooms, stairwells and other elements of a normal house were added over nearly 40 years of continuous construction with no overall objective other than to keep building. The result was an agglomeration of components, each of which was individually well-constructed, but that did not cohere into any sort of functional whole: stairways ran directly into walls, doors did not open, stained glass windows were installed in interior rooms with no light exposure, and so on. In Davis’s view, organizational science has the same problem: although each individual contribution must comply with strict disciplinary standards, no attention is paid to how all the contributions fit together; as a consequence, they do not.
He suggests a solution in which research contributions are judged not on their theoretical contributions but on how well they actually solve a social problem, the way prize-oriented contests do (such as the Netflix or SpaceX challenges). This would also incentivize cross-disciplinary collaborations.
- What might this look like? Take a look at the New Yorker profile: “Can Behavioral Science Save Flint?” It’s a very engaging ride-along with cognitive scientist Maya Shankar, of the Obama White House Behavioral Sciences Unit. In the waning days of the Obama administration she got on a plane to Flint, Michigan, and worked tirelessly to try to resuscitate the relationship between health officials and residents who’d been hurt by their governments’ poisoning of their water. (A literal plumbing crisis ruining thousands of lives.)
- She started with listening to the problem and to the affected people, then brainstorming on what tools from the social scientists’ toolkit might be able to help. It’s a very inspiring read about what social scientists can do.