Six Questions with Chris Udry

This is the first in a potential new series of posts of short interviews with development economists. Chris Udry was one of the pioneers of doing detailed fieldwork in development as a grad student and has continued to be one of the most respected leaders in the profession. While at Yale he taught David, and advised both David and Markus, and is famous for the amount of time he puts into his grad students. Most recently he has moved from Yale to Northwestern. We thought this might be a good time for him to reflect on his approach to teaching and advising, and to share his thoughts on some of the emerging issues/trends in development economics.

1. Let’s start with your approach to teaching development economics at the graduate level. The class when you taught David in 1999 was heavy on the agricultural household model and understanding micro development through different types of market failures. Most classes would involve in-depth discussion of one or at most two papers, with a student assigned most weeks to lead this discussion. There was a lot of discussion of the empirical methods in different papers, but no replication tasks and the only empirical work was as part of a term paper. How has your approach to teaching development changed (or not) since this time?

Try as I might, I have made little progress on changing my basic approach to teaching. The papers and topics have changed, but the essence of my graduate teaching remains the in-depth discussion of a paper or two each class. I’ve tried to expand the use of problem sets, and had a number of years of replication assignments. The first was hindered by my own inadequate energy (it’s hard making up decent questions!). I found that replication exercises required too much time and effort in data cleaning by students relative to their learning gain. Students were spending too much time cleaning, merging and recreating variables and too little time thinking about the ideas in the paper. I’ll reassess assigning replication this year, because there may now be enough well-documented replication datasets and programs available. With these as a starting point, it would be possible to get quickly into substantive issues in the context of a replication.

2. As someone who has successfully advised a lot of students, many of whom are now faculty themselves, what advice to you have to offer on how to be a good advisor, especially for people just starting out as junior faculty, perhaps in places without a lot of other development economists? David and Markus remember finding it tough to transition from being a PhD student to suddenly becoming an advisor – in terms of thinking about the responsibility involved, learning how to advise people on something on which you are not necessarily an expert yourself, and in helping students identify and focus on a topic that is both interesting and tractable.  A related question about advising is how/whether you advise differently for people who might be considering policy jobs rather than academic jobs?
I love advising Ph.D. students. I’ve been lucky enough to have a constant stream of astonishingly good people to work with. That makes it pretty easy. I’ve seen advisors succeed with very different models. For example, your advisor Peter Phillips had a fantastic, “lab” system in which new PhD students joined existing research projects, there was a lot of co-authoring, and dissertations often grew directly out of that work. This model obviously works in development as well. I’m much more hands off: all my students have generated their own independent projects, with lots of comments and some suggestions from me, but I think little direct guidance.  I don’t consciously treat students going into policy jobs rather than academic jobs any differently.
 
3. Looking back on your own research, are there papers you now think very differently about and perhaps interpret differently than you did at the time because of subsequent work you or others have done? Or work that you find is often mis-interpreted by others and you would like to clarify?
Probably all of them. A good example is “Gender, Agricultural Productivity and the Theory of the Household.” At the time, I was mostly satisfied making the argument that we could reject the null hypothesis of a Pareto efficient allocation of resources with the household. Now, though, it is obvious to me that with sufficiently good data and a well-designed empirical strategy, one could reject household Pareto efficiency anywhere. The interesting questions are more difficult: when is it most informative to do analysis at the level of the individual, the household, the family, networks of various types? What are the frictions and constraints that hinder efficient reallocations at these different levels of analysis? These are the questions that matter for our understanding of social organization.

4. Many more students now end up collecting their own data, and often also running their own field experiments as part of their PhD. Having observed this over the years, are their common mistakes you see occurring in how people do this, or general advice you would offer to someone considering this as part of their PhD?
A common error is a misperception that one must collect one’s own data or run one’s own experiment to be credible as a development economist. But our comparative advantage need not be field research, and indeed most of us are not especially strong in the management tasks that are needed to run large surveys or experiments. Students should resist feeling some sort of obligation do field work if their project does not require it.
For those who should do field research. I’ll mention four traps that I and students have fallen into. The first three are general; I think that the last is most common among students working in cultures other than their own.

1. Too narrow a focus. That’s about the last thing you expect to hear from an advisor – we typically are pushing our students to keep narrowing their vague ideas down to a researchable topic. But if it is worthwhile doing field work at all, it is because we don’t have a full idea of the range of possibilities of what we might find. So we should keep resources in reserve to ask additional questions or to do a follow up.
2. The related problem of getting so involved in the day-to-day administration of field work that we lose track of the bigger questions we are asking. Even if the first paper to come out of the field research is already planned, the main reason we are doing field research ourselves is to generate new ideas. We should reserve time for writing, modelling, testing out these ideas.
3. Our data collection methods can be amateurish and unscientific. We don’t take sufficient advantage of the protocols developed over decades of survey research around the world. We don’t learn enough from each other’s successes and failures. The World Bank LSMS unit has done some work to improve this, the JDE has had a special issue, and IPA, JPAL and now the GPRL at Northwestern(!) are working on this as well.
4. An overreliance on a limited range of local experts. Field work often starts with a few contacts in an academic institution, NGO or government. A partner in a local organization, an RA or a translator is one’s constant companion. If we are new to the environment, there are multiple layers of incomprehension and confusion, and there is a natural tendency to permit our views to be shaped by these few people we are connecting with. We should put more effort into expanding the network of people with whom we connect. 

5. The field work you did in Nigeria and then later in Ghana included some long stretches in each country.    This let you spend a lot of time talking with people in the field, often not even in the context of an interview.   With the move to experiments, it seems like field work involves much less time in the field, especially for faculty.   How do you think this has changed (or not) the questions people ask and the answers they come up with?  
It’s not the time in the field that is a concern to me, it’s how the time is used. If the research is very well-defined ex ante – as must be the case in many experiments and reaches its extreme form with pre-registered studies, then the field work is mechanical. This is necessary and important work, but constrains the range of one’s vision. Researcher interactions with people engaged in whatever they are studying are limited and overly structured. They are missing the most exciting (and fun) aspect of field research, which is the generation of new hypotheses sparked by unexpected information that emerges in conversation and discussion. Who knows what insights we lose when that kind of time is replaced with more time in the project office?

6. In a recent blog post, David discussed the issue of other researchers wanting to go back to re-interview participants in a study originally collected by someone else. Has this been something you have ever done, or discussed? Any thoughts on the feasibility and desirability of this?
 The ethical quandaries are important and difficult, but well laid out in David’s blog post, so I won’t discuss these. I have never re-interviewed participants in someone else’s studies, but other researchers have re-interviewed respondents of one of my field projects.  Tavneet Suri spent time in the communities in Akuapim in Ghana that Markus Goldstein and I had surveyed five years previously. She documented remarkable changes in the external environment facing pineapple exporters, driven by the introduction of sea freight routes that substantially reduced transportation costs. This, in turn, changed the relationship between farmers and exporters in important and subtle ways that Rahul Deb and Suri model in their JDE paper. This is a great example of a light touch “re-interview”; it did not depend on access to individual level data from the earlier research, but by building on the original survey was able to observe institutional changes that would otherwise have been obscured.
 
Let us know if you have ideas for other people you would like interviewed in this series, or questions you would like advice on from other development economists.