May 18, 2012

Randomized evaluations simultaneously emphasize that context matters and that human preferences are often shockingly similar across cultures and contexts.  For example, as humans, we tend to be present-biased, leading us to procrastinate, over-spend and under-save – as a three-country IPA study on reminders to save highlights. Yet, preferences like present-bias and others manifest in a myriad of ways in different situations. Randomized evaluations look at very specific circumstances to define causality around interventions. IPA seeks to answer some of the “bigger questions” by using the causality randomized evaluations establish to know whether to replicate elsewhere and how to improve, scaling up knowledge and impact. As the relatively new field of behavioral development economics grows, its stakeholders are debating big questions such as the relative importance of context. At the same time, as with any burgeoning field, an infrastructure is constantly evolving to support the field and accelerate the rate at which it makes change.

In Nairobi, Kenya, right outside the Kibera slums, a new actor in this space seeks to both provide infrastructure and gain increased understanding of the ways behavioral preferences don’t just operate differently in different places, but might be different in nature. Busara Center for Behavioral Economics is a new IPA project that sets up an infrastructure in which researchers can conduct behavioral economics studies on respondent populations from Kibera. In its first months, Busara has been building its respondent base through community recruiting. Now that studies are commencing, respondents receive text message invitations to participate in studies based on their demographic information and receive compensation when appropriate through the mobile money service M-PESA. Researcher Johannes Haushofer started the center with the goal of bringing lab-based behavioral economics to developing countries. In an interview with Haushofer, he highlighted three key reasons why he’s excited about what this center could do for randomized evaluation research.

First, Johannes emphasized that he is interested in how behavioral patterns operate differently for different people – for example, he pointed out distinctions between preferences of Kenyan farmers and US undergraduates. On the one hand, he acknowledges that there is no reason to assume people operate fundamentally differently in Kenya than they do at MIT. However, Johannes hypothesizes that the context in which people learn to think and behave is bound to influence their preferences and decisions. For example, perhaps the extent of present-bias differs when one is raised in an unstable socio-political context.

Second, and closely related to the first point, this center can also be used to inform the study design of randomized evaluations. Researchers might ask whether risk preferences or time preferences are the best lever for a randomized evaluation design. Or how much do social norms matter in bargaining? The baseline information from the Busara Center can help researchers get more information on answers to these questions in order to generate initial hypotheses.

Third, the center can help inform randomized evaluations after the fact to help make sense of some of the results. If awareness was improved through a sexual education intervention, but risky behavior itself was not, what behavioral patterns – identified amongst poor respondents in the developing world – lead us to more clarity on why?

Check out the Busara website and Facebook page and let us know what you think: could a center like this help support the growth of this field, enabling randomized evaluations to be more efficient? Will behavioral experiments conducted in the developing world give us a different foundational understanding of some preferences and behavior? Will it really get us closer to understanding why certain policies and programs are effective or ineffective?