Association for Public Policy Analysis & Management

Technology adoption decisions in development policy can lead to unintended consequences. This paper explores the effects of the adoption of mobile money of poor households in rural Kenya on their saving behavior, focusing on the two contradicting mechanisms through which the use of the mobile money is expected to affect the savings of the users.  I develop a model of household consumption and saving decisions in a two period setting, where adoption of technology is introduced. I consider several limited dependent variable models with endogenous binary regressors in order to estimate the effect of the mobile money in Kenya, M-PESA, on propensity to save and also on the amount of self-reported monthly savings. Using survey data collected in 2013, the preliminary results show that registered M-PESA users are 17 percent more likely to be savers; moreover the effect of M-PESA use on the self-reported savings per household member is significant and found as 238 Kshs. The decision to be an M-PESA user is taken endogenous. In order to deal with the endogeneity problem, the models are estimated by the maximum likelihood estimation strategy, using instrumental variable approach.