May 2010, Robert Cull (DECFP)
On May16-20, IBN co-editor Bob Cull was the guest of the Gates Foundation (Ignacio Mas, Jake Kendall, and Sheila Miller) on a field trip to Kenya along with Dean Karlan (Yale), Jonathan Morduch (NYU), David Roodman (Center for Global Development), Stuart Rutherford (SafeSave), and Dean Yang (Univ. Michigan) to learn about M-PESA, a mobile phone-based system for person-to-person payments and money transfers, and to explore potential research ideas regarding mobile payments.
I. M-PESA Described
M-PESA – the M is for mobile phone and PESA is the Swahili word for money – is a platform for making small-value electronic payments. Although M-PESA customers earn no interest on the balances in their accounts, many also use it to build small amounts of savings. The brainchild of a London-based team within Vodafone led by Nick Hughes and Susie Lonie, the M-PESA concept was pursued by Safaricom, Kenya’s dominant mobile phone operator. Since its inception in March 2007, M-PESA has attracted 9.5 million customers which represent over 40% of Kenya’s adult population. The service is clearly meeting a keen need for secure, low-cost money transfer.
M-PESA was originally conceived as a means of doing microcredit disbursements and repayment, but focus groups revealed much broader demand for moving money around the country. In large part this is because members of many Kenyan families live and work in Nairobi, or other cities, while the rest remain in the village. Tapping into this need, Safaricom’s first advertising campaign to promote M-PESA centered around the simple slogan, “Send money home.”
How does it work? M-PESA is built on the premise that charges are only incurred when a customer “does something” with her money and thus there is no charge for depositing funds. Rather, charges are incurred for sending and receiving funds. For example, after creating our own M-PESA accounts at one of the 16,000+ retail stores throughout Kenya and inserting our new Safaricom SIM cards in our mobile phones, we were ready to deposit funds. We started with a deposit of 300 shillings (about $4). The owner of the store took our cash, entered the transaction on her mobile phone, and then registered the transaction in her log book, which fellow traveler David Roodman signed. Almost immediately, David received a text message on his cell phone confirming the deposit. We then sent the money in David’s account to those of the others in our party, incurring a 30 shilling charge for each transaction. The money in those accounts could then be withdrawn at any M-PESA outlet or at an ATM. We were dealing in small sums, so the withdrawal fee was 25 shillings per transaction. Fees increase to 170 shillings (or 175 at an ATM) for withdrawals of 20,001 to the maximum 35,000 shilling limit. Though the tariff increases with transaction size, the gradient is not steep, which makes small-scale transactions less economical than larger ones.
Another interesting feature is that customers can send money to non-M-PESA clients that have access to a mobile phone. Money is debited from the sender’s account and the recipient receives a code via text message which can then be used to claim the debited amount at any M-PESA store. M-PESA clients pay a commission that is about three times the standard fee when sending to non-customers, though the non-customer pays nothing to withdraw. Why set up the pricing in this way? First, the arrangement acts a sales device to ensure that the non-customer has an enjoyable first experience with M-PESA. Second, and perhaps more subtly, the sender has a strong incentive to persuade the recipient to register with M-PESA to reduce future commission fees. Since the sender is the one with the money, and therefore likely to have the economic power in the relationship, this pricing arrangement induces recipients to sign up. The influx of new registered users, in turn, increases the utility of the M-PESA network for all users.
|The Lake District M-PESA store in Kisumu near Lake Victoria. Its location, directly across the street from the offices of power company, helps make it one of the busiest M-PESA locations in the country.|
Another important aspect of M-PESA’s success is the meticulous attention paid to branding. All M-PESA stores are painted “Safaricom green” (see picture), which ensures a uniform appearance, reinforces the link with the well-respected mobile provider, and makes it easier for customers to locate the outlets. To ensure that customers have almost identical experiences throughout the network, Safaricom hired a third-party vendor named Top Image, which visits each store at least monthly, rating them on visibility of branding and the M-PESA tariff poster (partially visible in the upper left corner of the picture above), in addition to the availability of cash and electronic value to accommodate the flow of customer transactions, and the quality of record-keeping. We visited a lot of M-PESA outlets in Nairobi and in and around Kisumu, and they all looked and functioned a lot like the one pictured above.
II. Payment Applications
M-PESA is now used for all sorts of person-to-person payments including, we were told, settling golf bets on the course. But a new wave of applications is making it possible for customers to make a range of payments to businesses using their M-PESA accounts. The most widely used is for payment of electricity bills, but other businesses are also coming to see M-PESA’s advantages. We visited Bridge International Academy, which is rolling out private schools in some of the poorest areas of Kenya, and is one of the first M-PESA ‘super-users.’ Bridge provides a remarkably low-cost ($4 in monthly tuition per student) alternative to public schools. To do so, all aspects of providing education are, in a sense, commoditized. The curriculum is standardized and details exactly what teachers are to present, down to what they should write on the chalkboard for a given lesson. Construction of the schools – methods, materials, the number of seats in a classroom – is also standardized. Class sizes are large at 55-65 students, but to judge from the burgeoning enrollments, parents are pleased with the quality of education and Bridge’s accountability to them.
|Students at Kingston School in Nairobi. Tuition at the Bridge International Schools can be paid only through M-PESA.|
With a business model so focused on scale and cost efficiencies, M-PESA proved a natural fit for managing tuition payments to Bridge. Indeed, Bridge only accepts payments through M-PESA or directly into its account at Equity Bank. These arrangements mean that Bridge’s employees do not collect cash payments, which would pose a security threat in the neighborhoods where these schools are located. In addition, using M-PESA means parents need not be physically present to make tuition payments and that other members of the extended family (e.g. grandparents) can pay. Electronic payment records also make it easy for Bridge to track families that are delinquent on tuition payments. Our discussions with representatives of Safaricom indicate that the range of person-to-business payment options will continue to grow as new super-users are identified.
III. The Future
By lowering the costs of money transfer, M-PESA has helped to increase market activity, especially outside cities, and that trend will continue. At Jubilee market in Kisumu, shop owners no longer spend two days traveling to Nairobi and back to provide funds for their suppliers. Funds arrive via M-PESA and suppliers draw on them to produce or obtain goods that are then sent to Kisumu. In addition to the time savings, M-PESA provides a secure alternative to traveling with relatively large amounts of cash. In the more remote villages outside of Kisumu, owners of small shops use M-PESA to pay for goods from Jubilee Market. Whereas there used to be little economic activity in these village markets – there wasn’t much cash around and thus shopkeepers had little incentive to keep inventories of goods – M-PESA has made cash less scarce and businesses have responded.
Although the benefits of M-PESA to its customers exceeded my expectations by a long shot, until very recently it remained only a payments service and customers received no interest on their deposits. That changed with the introduction of M-KESHO on May 18, a joint venture of Safaricom and Equity Bank. M-KESHO is a co-branded suite of financial products that will ride on the M-PESA transactional platform. There is no fee to open an M-KESHO account, nor minimum balances or monthly charges, all features shared by the M-PESA account. M-KESHO accounts differ from M-PESA in that they pay interest, do not have an upper limit on account balances, and are linked to credit and insurance facilities provided by Equity Bank. Customers can deposit and withdraw money from their M-KESHO account by transferring value to and from their M-PESA account. The M-KESHO arrangement between Safaricom and Equity is exclusive for one year, and thus the product could be offered by other banks after that. By offering a menu of products and services, M-KESHO promises to bring fuller immersion into the financial system for many ordinary Kenyans.
|Jubilee Market in Kisumu bustles with activity (this is a side street, not the main market). The density of M-PESA outlets is reflected in the three stores at Jubilee.|
IV. The Research Questions: Why Kenya? Why Now?
Recipes to promote financial inclusion in developing countries remain elusive, and so the key factors underlying M-PESA’s success should be of broad interest. Aside from the investments in branding described above, there were a number of features of Kenya that are potentially relevant. First, and as alluded to above, migration patterns within Kenya have split many families and thus generated high demand for domestic remittances services. Second, Kenya was under-banked. In 2007, the ratio of private credit to GDP hovered near twenty percent and there were only 2.5 million bank accounts in a population of 39 million. High levels of financial exclusion provided an open playing field for M-PESA. M-PESA also benefitted from the structure of the mobile telephony industry, since Safaricom enjoyed a market share near eighty percent. A final potential factor is the flexible regulatory treatment afforded M-PESA. Because M-PESA accounts earned no interest and because customer balances were deposited in a trust account in a regulated financial institution (and thus were not used for intermediation), the Central Bank of Kenya allowed Safaricom to operate M-PESA outside the mandates of the banking law under a letter of no objection from CBK that laid down specific conditions for M-PESA operations.
Which of these factors were necessary for M-PESA’s success? This is, of course, not possible to determine from studying only Kenya’s experience. But, as similar services are rolled out in other countries, those cases can be studied to begin deriving a set of lessons. And evidence from randomized controlled trials that vary product offerings and price incentives in a variety of settings could also offer important clues about where take-up rates for mobile payments and banking services are likely to be high.