Development, Technology Adoption, and Social Networks

Abstract

Agriculture remains a key component of economic development, but the methodology for how development policies are determined has changed for developing countries. In the last decade, the focus of economic growth in developing countries has shifted from country-wide prescriptions to testable micro-development programs at the local level. As international development focuses in on local programs, social networks have been identified as a key component for their effective deployment. This dissertation analyzes the effects of a social network-based intervention. It contributes to the economics literature on identifying social network effects by implementing a randomized encouragement design to develop social capital, while simultaneously introducing a new method of development training. The program implemented here is comprised of two parts, and was conducted with female-headed households in rural Uganda, that were growing a relatively new cash crop, cotton. The first part conducted social network-based information games in 20 sample villages, in which each participant was trained in one aspect of cultivating cotton, and encouraged to attain a full set of knowledge on growing cotton through her assigned learning networks. They were presented with two different incentives schemes for accumulating information: competitive and team incentives. The second portion of the program paired the surveyed individuals at random with other game participants. These pairs were encouraged to develop team goals across the growing season and a time schedule for networking as well as update and share their learned information from the games on a regular basis. The estimated effects of the SNI, which comprise this dissertation, include both the effects from the information games and the effects of the mentored pairing; that is, the impact of acquiring one information point and one new link. I compare the effects of this program to a standard agricultural training program that was concurrently conducted during this research, in which extension agents taught the same information that was presented in the information games but with a traditional classroom-based teaching method. My games analysis shows that females learn more when presented with competitive incentives. The total number of learning points learned during competitive incentives first order stochastically dominates the total number of learning points learned during team incentives. However, for the dissemination of one specific information point, team incentives are better at ensuring that a unique information point reaches the entire group. Difference in difference estimates, controlling for the training program, show that the overall SNI program had significant effects on the average farmer, with diminishing returns for higher yielding farmers. I find that these average effects are comparable to the effects of the conventional training program, but at a fifth of the implementation cost. A closer examination shows that the SNI program has its most significant effect for farmers growing around the average output when the program was started in 2009 (100-200 kgs/acre), while the Training program has its greatest and most significant impact for those yielding above the average output in 2009. Therefore, the two programs are not necessarily substitutes in how they effect change. My research shows that a competitive incentive structure coupled with social network-based learning serves as an effective paradigm for improving outcomes for the poorest producers.