James Schintz

NexThought Monday – 12/8/14: Safaricom, share your M-PESA data!: Why we need data philanthropy in Kenya

The team here at the Microfinance Information Exchange (MIX) have been hard at work producing a new country analysis for Kenya.

One of the more interesting topic threads I encountered while researching mobile money’s reach in Kenya centered around data mining and privacy concerns, specific to Safaricom’s M-Pesa product. I came across a few ordinary netizens who blogged curiously about the nefarious ends towards which all the M-Pesa data could be used. And with risks ranging from corporate greed to Big Brother and monopolies, there is merit for concern. I can think of very few occasions (Microsoft and Standard Oil among them) where such a large portion of a country (15 million users and 2 million transaction per day in a country of 25 million adults) has voluntarily paid for services through a private entity. The scale of the information collected from all these transactions is incredible, supposedly reaching all subsets of the population. But in spite of the risks of data of this magnitude, my mind immediately drifted away from how this information could be abused to how it could aid financial inclusion efforts.

Data philanthropy from private, commercial entities is in vogue. Orange has opened up their data in Senegal and Ivory Coast for research purposes. Digicell helped map population movements in Haiti after the 2010 earthquake. UN Global Pulse has tried to piece together public sentiment towards M-Shwari in Kenya via social media. Also in Kenya, mobile data has been used to map Malaria’s effects. But if M-Pesa followed suit in opening its data coffers, the impact could dwarf that of previous efforts. For regions where obtaining decennial census data can be difficult, the granularity of the M-Pesa data is electrifying. By enabling informed decisions on current market realities, this data has the potential to spark generational shifts for consumers, commercial entities, NGOs and regulators. So Safaricom, share your M-Pesa data, please!

Why am I picking on Safaricom? It’s a rare opportunity to collect comprehensive data from one sole actor. And it would just be so easy. Because of the first mover advantage, M-Pesa is a de facto monopoly and will continue to be so for some time. Yet due to encroaching competition, Safaricom’s vice-grip on the mobile market is no longer an issue of concern; rather, it’s an opportunity to gain deep market insights. Safaricom already releases its data to some Kenyan authorities, so why not make this data available to financial inclusion actors? Intermedia, Gates and FSDK all collect data on mobile money via surveys. Having access to M-Pesa’s transactional level data would be a treasure trove to financial inclusion actors.

Some proposals for using this data include:

  • Identifying at-risk markets for variations in domestic remittances

  • Identifying socio-economic subsets with low up-takes of mobile money

  • Identifying over-indebted communities

  • Identifying communities in need of commercial banks

  • Predicting failure to repay debts at the individual level

MIX is accustomed to the challenge of incentivizing profit-seeking entities to share their data – our website FINclusionLab hosts data on financial services in 18 countries. We’ve found that in some cases, regulation by central banks or telecommunication authorities is a great way to obtain reliable, well-defined data. But often this data is only delivered in large, aggregate chunks in order to protect the competitive advantage of the company. Deriving actionable insights from mobile money data to further financial inclusion requires the availability of granular information outside the scope of normal regulation. In addition, this data needs to be anonymized to protect consumer rights, and made publically available (within reason) to derive a positive cost-benefit relationship from provisioning it.

Absent the prompting of regulators, many companies are reluctant to share any information at all. But luckily, some companies have started to realize that there is not as much benefit as there used to be in keeping all their data secret – and there can be considerable advantages to sharing it. For instance, both Visa and Mastercard have allowed developers to locate ATMs and point of service devices connected to their network, providing a cost-effective way to direct more customers to their services through outside websites or mobile applications. And one of the more innovative API’s I have come across is an experiment by BBVA, which provided a community of 780 programmers with anonymized data on card transactions, challenging them to create applications to benefit their customers.

These examples expose a major benefit to this approach: namely, it can serve both profit-seeking and regulation needs. Financial service providers should be rushing to produce these types of APIs because of the potential for return for data as an asset. And regulators need to sponsor these public data endeavors, encouraging a self-sufficient reporting model among financial service providers, as well as a more equitable distribution of financial services to the general public through information symmetries. Maybe with a little nudge, Safaricom might agree to open their gold mine. If they do, the company, the financial inclusion community – and most importantly, low-income customers – stand to benefit.

James Schintz is an analytics expert focusing on data visualization for the financial inclusion sector, working as analytics lead at Microfinance Information Exchange.

digital payments, financial inclusion, mobile finance, philanthropy, research