Mobile Money in Tanzania

In developing countries with sparse retail banking branches, mobile telecom net- works have emerged as major providers of financial services bypassing traditional banks. Using individual-level mobile money transaction data in Tanzania, we find that the vast majority of these transactions can be classified as either (i) transferring money to others, (ii) transporting money for short distances, or (iii) storing money for short to medium periods of time. We find that the demand for long-distance transfers is less elastic than for short-distance transfers suggesting that the mobile networks compete with traditional cash transportation by bus drivers, in addition to competing with each other. Using the revealed preferences for transportation and storage transactions, we monetize the economic damage caused by a high level of crime. We estimate the willingness to pay to avoid walking with cash an extra kilometer and to avoid storing money at home for an extra day to be 1.1% and 1% of an average transaction, respectively. We propose a Pareto superior price discrimination scheme where cash-out fees that follow a transfer are set to zero, while otherwise cash-out fees are set a bit below transfer fees.