Social networks, phone money transfers, and rainfall shocks: Evidence from Rwanda

A large literature describes how local risk sharing networks can help individuals smooth consumption in the face of idiosyncratic economic shocks. However, when an entire community faces a large covariate shock, and when the transaction costs of transfers are high, these risk sharing networks are likely to be less effective. In this paper, we document how a new technology – mobile phones – reduces transaction costs and enables Rwandans to share risk quickly over long distances. We examine a comprehensive database of person-to-person transfers of mobile airtime and find that individuals send this rudimentary form of “mobile money” to friends and family affected by natural disasters. Using the Lake Kivu earthquake of 2008 to identify the effect of a large covariate shock on interpersonal transfers, we estimate that a current-day earthquake would result in the transfer of between $22,000 and $30,000 to individuals living near the epicenter. We further show that the pattern of transfers is most consistent with a model of reciprocal risk sharing, where transfers are determined by past reciprocity and geographical proximity, rather than one of pure charity or altruism, in which transfers would be expected to be increasing in the wealth of the sender and decreasing in the wealth of the recipient.

The data sits under “Appendix C. Supplementary data”. The data is included in Stata.dta and do file format. Also included is PHP code and documentation.

The data file includes an anonymized version of the phone survey data, which contains asset information and household weights. Mobile phone operator data and Rwandan Household Survey data used in the paper are not available in the download. A ReadMe file provides details for requesting this data.