Mobile Money: Understanding and Predicting its Adoption and Use in a Developing Economy (1812.03289v1)

Abstract: Access to financial institutions is difficult in developing economies and especially for the poor. However, the widespread adoption of mobile phones has enabled the development of mobile money systems that deliver financial services through the mobile phone network. Despite the success of mobile money, there is a lack of quantitative studies that unveil which factors contribute to the adoption and sustained usage of such services. In this paper, we describe the results of a quantitative study that analyzes data from the world's leading mobile money service, M-Pesa. We analyzed millions of anonymized mobile phone communications and M-Pesa transactions in an African country. Our contributions are threefold: (1) we analyze the customers' usage of M-Pesa and report large-scale patterns of behavior; (2) we present the results of applying machine learning models to predict mobile money adoption (AUC=0.691), and mobile money spending (AUC=0.619) using multiple data sources: mobile phone data, M-Pesa agent information, the number of M-Pesa friends in the user's social network, and the characterization of the user's geographic location; (3) we discuss the most predictive features in both models and draw key implications for the design of mobile money services in a developing country. We find that the most predictive features are related to mobile phone activity, to the presence of M-Pesa users in a customer's ego-network and to mobility. We believe that our work will contribute to the understanding of the factors playing a role in the adoption and sustained usage of mobile money services in developing economies.

• The paper uses machine learning and large M-Pesa transaction data to predict mobile money adoption (AUC 0.691) and spending (AUC 0.619), identifying key influencing factors.
• Key predictors of mobile money use include SMS volume, active mobile days, the percentage of M-Pesa users in one's social network, and geographic mobility.
• Understanding these factors allows for targeted marketing, better service design, and enhanced financial inclusion initiatives in developing countries.

Mobile Money: Analyzing and Predicting its Adoption and Use in Emerging Markets

The paper "Mobile Money: Understanding and Predicting its Adoption and Use in a Developing Economy" presents a data-driven investigation of mobile money services, with a focus on predicting the usage patterns and monetary spending using M-Pesa data in an African country. The research provides robust quantitative analysis from a vast dataset consisting of mobile phone communications and M-Pesa transaction records, elucidating factors influencing mobile money adoption and usage.

The research paper delineates three critical contributions. Firstly, it offers a comprehensive analysis of M-Pesa utilization patterns across various services by examining both communication and monetary transaction flows. Secondly, the paper details the construction of machine learning models aimed at predicting future mobile money adoption and spending. The authors employ features from multiple data sources: mobile telephony records, information on M-Pesa agents, the user's social network characterized by M-Pesa activity, and the geographical nature of the user's location. The results of these models show an AUC of 0.691 for adoption and 0.619 for spending prediction models, indicating modest but significant predictive capability. Thirdly, the paper elucidates the most predictive features within these models, offering key insights for the design and deployment of mobile money services in the context of a developing country.

The comprehensive feature engineering process across several domains, including call interactions, social network measures, and mobility metrics, underscores the paper’s methodological depth. Particularly notable is the finding that the number of sent SMS messages and the active days of mobile usage correlate strongly with future M-Pesa usage. Furthermore, the paper reveals the pivotal role of the user's social network, as the percentage of M-Pesa users within a consumer's network strongly correlates with mobile money utilization likelihood. Mobility, exemplified by the radius of gyration, also emerges as a noteworthy predictor.

The implications of this research have broad relevance. The demonstrated ability to predict mobile money adoption can drive more targeted marketing and service design strategies, enhancing financial inclusion efforts in developing economies. Understanding the prevalent role of social networks may inform initiatives aimed at fostering mobile money ecosystems through community and peer influences. Additionally, the insights into transaction predominance in urban versus rural areas and longer money transfer distances compared to call distances reflect the economic dynamics and potential for economic development facilitated by mobile money services.

Moving forward, further investigation into the characteristics of the agent distribution network could refine understanding of its influence on mobile money usage. Cross-regional studies could determine the generalizability of these findings to other emerging markets. Enhanced model deployment in operational settings could offer empirical validation and potential improvements in financial service offerings. Thus, this paper sets a foundational framework for both theoretical exploration and practical application in financial technology within developing regions.