The value of face-to-face: Search and contracting frictions in Nigerian trade 
(Job market paper)

Distance between buyers and sellers can create search and contracting problems: how to find out what goods are available in far away places, and ensure they are actually delivered? Traveling to do business in person is one way of dealing with both, transforming a remote transaction into one that is face-to-face. I estimate the magnitude of search and contracting frictions in a developing country context by exploiting the fact that travel is a common, costly, and easily observable strategy for coping with them. I collect transaction-level panel data from Nigerian importers of consumer goods that combines the “what” of trade (e.g. products, quantities) with variables describing “how” trade is conducted (e.g. travel, payment terms). To account for patterns inconsistent with a full information environment, I build and estimate a model that embeds a search problem and a repeated game with moral hazard into a monopolistically competitive trade framework. Welfare from imported consumer goods would be 29% higher in the absence of both frictions. I decompose the total barrier into parts attributable to search and to contracting, and show why the effects will be larger in markets with low consumer spending, high firm entry/exit rates, and frequently changing products. The results suggest that greater attention to market integration policies beyond transportation and tariffs could have large welfare effects, particularly in developing countries. In counterfactual scenarios, I show that regulation of air travel between Nigeria and China would yield gains in Nigeria on the order of $650 million per year through consumer goods trade alone, while existing financial services do little to mitigate frictions because they do not offer a better contract enforcement technology than travel or repeated interaction.

Reputation and relational contracting: Experimental evidence on contracting frictions in West African trade (with Shelby Grossman)
[Baseline survey currently underway]
In theory, freely flowing information about past behavior can substitute for formal contract enforcement. However, little is known about whether or how reputation alleviates contracting frictions in practice.  In this project, we experimentally vary the availability of public information about Nigerian used clothing traders' transaction history, via a new mobile money escrow service. Currently, these traders travel to neighboring Benin to make purchases in person as often as 2-3 times per week. We investigate whether increased access to information about potential trade partners' track records substitutes for costly monitoring and changes market outcomes. Two treatment arms identify to what extent any effect operates through a selection channel (traders are able to identify partners who are of a reliable type ahead of time) versus a moral hazard channel (partners behave better knowing that their actions will become public). 

Direct and spillover effects of a business training program: Experimental evidence from urban Uganda (with Greg Fischer and Dean Karlan)
Is lack of knowledge about how to manage a business a major impediment to the growth of small and medium enterprises (SMEs) in developing countries? If so, will the increasing number of public and non-profit technical assistance programs aimed at training entrepreneurs raise overall productivity in the economy, or simply redistribute surplus across firms? We conducted a randomized evaluation of the Women Mean Business (WMB) program, which offers classroom training and long-run mentoring to female small business owners in urban Uganda.  Tracking participants for three years after the end of the program, we find that it had substantial and lasting, but not transformational, effects on attitudes, business practices, and financial outcomes, including a 17 percent increase in monthly profits.   Surprisingly, we also find a positive ‘rising tide’ general equilibrium effect at the neighborhood level:  in areas in which a larger fraction of local businesses were (randomly) invited to join the program, control group firms who did not participate enjoy higher revenues and profits even though they do not appear to adopt the new business practices themselves.