Realizing the Potential 47000 of Branchless Banking: Challenges Ahead B eing able to make payments conveniently and "payments utility," these options are complicated securely is an essential ingredient in modern for me to handle--and probably too costly for life and commerce. It enables economic livelihoods a bank or other provider to consider.1 Once I am and supports many social relationships (remittances connected to the payments utility, I am more likely between geographically split families and friends), to become an attractive potential client of a broader NOTE communal support actions (e.g., joint buying of range of financial institutions--not just for the lone staples), and public welfare programs (payments to bank or microfinance institution (MFI) that opened an needy families). Yet most people and microenterprises office a long bus ride away from my neighborhood. in developing countries must rely on physical delivery Provision of financial services can then more easily of cash or actual goods to make payments. This be de-coupled from geography (bricks-and-mortar imposes large costs and risks on those beyond the branches), placing more competitive focus on the reach of modern payment networks. range and quality of products and trust in the brand. Access to payment facilities is a major The lucky few have more efficient means of exchange: enabler for achieving universal access to checks, money orders, direct bank transfers, credit finance. Once I have the capability to cards, debit cards, and so forth. All these cut down on easily pay and receive money to and from the need to carry cash, making consumers and their anyone, my range of financial possibilities money more secure. Even handling and exchanging expands. cash is a lot easier: consumers have debit cards with which they can exchange electronic value for cash at In this paper we further develop a broad vision for any number of conveniently located automated teller financial inclusion sketched out in Mas (2008), where machines (ATMs). payments can be easily made through an electronic network. What makes visioning such a payments Affluent people have access to these payment utility possible is the technology we have today, which facilities through the banking system. Should that can be used to bridge distances, close information remain a privilege, or should we start conceiving of gaps, contain settlement risks, and generally reduce retail payment networks as a utility much as we think transaction costs. We are confident that today's FOCUS of access to water, electricity, or phones? What would technology can do the job. Now the challenge is to it take to provide every person with the option of develop attractive services that engage customers participating in modern payment networks--that is, to and workable business models that enable provide them with transactional accounts that enable decentralized, largely private, institutions to build this transactions right from their home, or at least from payments utility. No. 50 specified outlets in the neighborhood, affordably? October 2008 In the next section we first explain the basic financial Access to payment facilities is a major enabler for needs that can be met through a payments utility, and Ignacio Mas achieving universal access to finance. Once I have from that we derive the main characteristics that the the capability to easily pay and receive money to payment utility should have. Then we review the basic and from anyone, my range of financial possibilities logic of branchless banking. Finally, we outline the expands. I can use this capability to put money away main challenges we see for the mass deployment of in a savings account. I can receive disbursement branchless banking channels. In the rest of the paper of a loan and repay it conveniently. I can pay my we go into these challenges in more detail. monthly insurance premiums. Without access to the 1 The term "utility" in this paper is doing double duty, on the supply (provider) side as a description of the standard model of universal service provision for networked industries (akin to the water utility) and on the demand (client) side to indicate the value (utility) of payments and savings for currently unbanked people. 2 A vision of a payments flexible transferring of value now and in the future. In network as a utility principle, the most flexibility would be achieved if at least four basic elements were in place: The need · A payments network that (i) allows value to be Finance can be understood to be a collection of ways routed between all the parties on the "grid" and of transferring value: (ii) provides for ubiquitous points (access nodes) at which cash can be converted to electronic value · Transfers across people (or legal entities) = (or vice versa). payments (for goods bought or sold or services · A range of providers (banks or other financial rendered) or remittances. The customer's objective institutions) that creates and markets a range of is to discharge commercial or social obligations. finance products (e.g., a savings, loan, or insurance · Transfers across space = cash delivery. The product) to meet customers' needs. Customers customer's objective is primarily safety and would need to trust the provider­product package convenience by limiting the amount of cash that enough to accept it in the marketplace. Thus, the he needs to carry around and having the option of provider's brand hinges on relevant products and converting electronic value to cash (or vice versa) trust. on demand, as and where needed. · A basic or transactional account for each individual · Transfers across time = savings (transfer value to that can be used to collect and transfer value the future) or loans (borrow from the future). The electronically. It is the individual's landing point customer's objective is to separate the timing of onto the payment grid. cash inflows and expenditures. · An enforceable legal framework that specifies the · Across future outcomes (or states of the world) = rights and obligations of each player in the system. insurance. The customer's objective is to alter the This would promote trust in the enforcement of nature of risks he is exposed to, by essentially contracts and the protection of consumers. contributing during good times and receiving compensation during lean times. These are the building blocks for a payments network that reaches out to all. But for this payments network The first of these transactions is a spot transaction-- to be useful, people need to be able to transfer taking place here and now--whereas the rest are value through the payment network in a way that a package of a spot transaction (the deposit, the is convenient, reliable and secure, widely available, collection of loan proceeds) against a promise affordable, and useful. of future offsetting transactions. Any financial product or service can be broken down into a set · Convenient--a transactional account, and an easy of transactions between the parties over time. The way of issuing instructions for transactions. fact that all financial products (beyond payments and · Reliable and secure--low probability of loss of transfers) represent bundles of transactions occurring value, low risk of nonacceptance of transactions at different points in time puts trust at the core of because of illiquidity (that is, the cash point has financial service offerings. cash), reasonable certainty on the finality of transactions. Access to finance entails access to a basic legal, · Widely available--ability to transfer money to institutional, and physical infrastructure that permits anyone, and ubiquity of cash-in/out points. 3 Table 1: Comparing retail payments systems against other utilities Utility Key user benefit Network provision Range of services supported Water Health and By state-owned or regulated By the user: drink, cook, shower, sanitation monopolies wash, irrigate Electricity Energy By state-owned or regulated Range of electrical appliances: monopolies heating/cooling, motors Telecoms Communications Licensed operators; oligopoly in Personal communications, and information "last mile" services broadcast radio/TV, Internet content Payments Value transfers By government entities or by Save, pay for insurance, receive associations of licensed or loans, pay for goods, etc. authorized institutions · Affordable--low overall cost relative to people's Yet access to the payments system has never been seen means, and a charging structure that correlates as a basic enabler for individuals or microenterprises. with customers' perceptions of value. Is it time to start conceiving the retail payments · Useful, especially for poor people who have system as a utility? This does not necessarily mean up to now been excluded from formal financial provision by the state, but the creation of a policy services--the services offered on the network framework that adequately incentivizes investment should make it easier for poor people to manage in the infrastructure, promotes universal extension their financial lives, including their ability to (i) of access, and provides minimum basic protections smooth their consumption in the case of seasonal to consumers. or erratic earnings, (ii) handle large life events and absorb unforeseen shocks, and (iii) support their Of course, governments have always been very microentrepreneurial and livelihood activities. involved in payments systems, in at least four ways. Governments need to make many payments (welfare Seen in this light, the retail payments network is akin payments, pensions, teachers' salaries) as well as to a utility (see Table 1). The faucet in the kitchen sink collect fees and taxes in remote areas, so many provides convenient, reliable, and affordable access created a public bank with many branches through to a standard commodity (water), which the user the country (sometimes piggybacking on postal then turns into a range of useful activities (drinking, outlets). Many governments have also used the broad cooking, washing). The power outlet provides postal infrastructure to create a basic postal or giro equally convenient, reliable, and affordable access bank to serve populations in remote communities.2 to a standard commodity (electricity), to which the However, public banks often have been hampered user connects a range of devices (light bulb, toaster, by underinvestment in technology, undertrained vacuum cleaner). In just the same way, access to the and undermotivated staff who are often resistant payments network offers individuals access to a basic to change, weak incentives to deliver services well, service (sending/receiving monetary value) through and lack of competition. Many of these banks have which they can fulfill a range of needs (save, borrow, been increasingly under pressure to become more insure, or send money to anyone). financially self-sustainable or have been privatized, often with the result of curtailing their reach. 2 World Bank (2006) provides a comprehensive survey of postal networks in 55 countries. It reports that in 2004, postal networks supported 390 million accounts, with savings of $94 billion, and made 3.3 billion transactions annually. Interestingly, there were 490,000 post offices in the developing world, compared to 275,000 bank branches. 4 At a more basic level, central banks have traditionally of products--but not financial services. How can enabled basic retail payments by the issuance of cash, financial institutions use these stores, rather than ensuring relative monetary stability and providing build their own retail presence? Finding ways for cash liquidity across the national geography by village stores to offer financial services has three transporting cash to provincial offices of the central main economic advantages for banks: (i) it permits bank. They have also promoted or even incubated increased physical presence in the area for banks or interbank clearing and settlement systems. other providers at drastically reduced set-up costs for banks; (ii) it turns customer service costs into However, these government interventions have left variable costs, insofar as stores are remunerated many segments of the population outside of electronic per transaction; and (iii) it offers the opportunity payments networks. Some private schemes have to create a familiar service environment for poor, emerged to fill the gap between interbank electronic less educated people who may feel intimidated by payments systems and ever-present cash. Western the service style at traditional bank branches. Union is perhaps the main prototype. It used the telegraph network to wire person-to-person transfers 2. Deliver trust through technology. As the bank­ beginning late in the nineteenth century. Payment customer interface shifts to third-party retail card networks, which started in the 1950s, are now environments, ensuring trust--the fundamental going beyond their original purpose of paying for product "sold" by banks--becomes paramount. goods and services to include remittances and cash Technology-assisted processes can be used to deposit/withdrawal services. Their members have substitute for traditional means of ensuring trust, access to a broad range of services, but they mostly so that clients (and banks) have confidence in cater to middle and upper class segments, leaving the transaction carried out by a nonbank agent. the poor behind. With the proliferation of debit cards As described in detail in Mas and Siedek (2008) and ATMs since the 1980s, we can now truly talk of and depicted in the middle column of Figure 1, a retail payments network--for those who can afford the trust infrastructure has four key elements: (i) a bank account. a mechanism to identify and authenticate the customer (and the agent), for example using a The plot so far card; (ii) a point-of-sale (POS) terminal at the store remotely managed by and with a direct The fact that financial institutions have to rely on communications link to the bank, so that the physical branches that are limited in their reach makes bank can authorize and record transactions in it difficult and costly for populations in rural areas real time; (iii) a requirement that all customer or in poorer districts to take advantage of financial transactions are immediately offset against the services. It is also a problem for the state, which store's own transactional bank account and are needs to distribute public sector salaries, pensions, subject to availability of funds by the store, so that and social benefits to a dispersed population. Three no credit risk is involved at the point of transaction; key innovations would permit us to transform the and (iv) a receipt-issuance capability, so that the economics of financial service delivery at the customer customer has recourse if the transaction fails to be end (see Figure 1): recorded appropriately. 1. Use existing retail outlets. Practically every 3. Use existing mobile phones. Set-up costs can be neighborhood, no matter how remote or poor, reduced further by using existing mobile phones has a number of retail stores selling a wide variety as part of the technology platform, for both clients 5 Figure 1: Options for reducing the cost of retail banking infrastructure and merchants, as described in more detail in Mas phones is finding market success in the Philippines, and Kumar (2008). Clients can use their mobile Kenya, and South Africa. phones as the means of authenticating themselves and identifying their account number. In fact, As banks move from offering service through their mobile phones already contain a smartcard (SIM own branches to offering service through branchless card), and it or the phone itself can easily contain a channels (using first ATMs, then retail-based POS "virtual" bank card. Clients would use the mobile systems, and eventually a mobile payment platform), device itself (rather than the store's POS terminal) they can achieve lower unit costs for basic transactional to enter their personal identification number (PIN) services. But they will not be able to support the same and release their account number. Merchants range of services (e.g., account sign up, financial could also use their own mobile phones as a POS advice, cross-selling of financial products) and will not terminal. The mobile phone has similar technical deliver the same customer experience. With ATMs, capabilities as a POS terminal: a screen, a numeric the customer loses human interaction; with store- keypad, and a secure communications link. In based agents, they will not be able to inquire about this case, we have a mobile-to-mobile service. A their accounts with the shopkeeper in the way they solution based on mobile phones can therefore would with a teller; and with mobile payments they will substantially reduce the cost of spreading have to master the user interface on their phone and financial services over many retail environments, interact through a very small screen and keyboard. at least in areas with relatively high mobile phone That is the key trade-off the bank needs to manage penetration. (The receipt would take the form of when deploying an outlet: lower setup cost against a text message; alternatively, a portable printer less satisfying customer experience. In managing this could be attached to the mobile phone of the trade-off, it would be entirely appropriate for a bank merchant.) to adopt a mixed-channel strategy incorporating some if not all of the options in Figure 1. There is no The first two steps above are the essence of the right or wrong; the bank simply needs to match the banking agent model that has had considerable needs of specific customer segments against their success in Brazil. Extension of the model using mobile willingness to pay. 6 We see banks in many developing countries focused is lagging behind some perceptions. There are some on their ATM deployments. ATMs are a lower cost fundamental obstacles we still need to resolve to alternative to rolling out more tellers, by automating make branchless banking an attractive proposition the bank­customer interaction. This creates a for financial institutions. The main challenges are as paradox: is the cost of labor really the main problem follows: for banks in developing countries? No--the problem is (i) the cost of ensuring that tellers are adequately · Understanding what drives customers to use and trained and can be trusted and (ii) their inflexible labor pay for a range of financial and payments services. contracts. In the agent arrangements discussed here, Understanding customer psychology and designing customers retain human interaction at the touchpoint, appropriate services is essential to stimulate supported by a technology-based trust infrastructure. demand. How would making financial transactions So customer-facing staff can be outsourced to almost available electronically transform how people make any store owner. financial choices? · Making the economics work for the cash-in/cash- With such arrangements, we envision a world where out agents. To go to the heart of the problem: people are able to make small deposits into their shops are used to taking high single-digit or low bank account through a variety of cash handling double-digit commissions on selling a bottle of outlets right in their neighborhood. Depositing and cola, a mobile prepaid card, a bar of soap, and so withdrawing money from your account could be just forth. So why would they offer cash-in/out services another product that your local store offers, along for a much lower commission, if that saddles them with toothpaste and mobile prepaid cards. It might with extra security risks and extra trips to the bank? even be possible and appropriate for retail stores to It seems unlikely that in most cases mobilization of work for all banks, and neither depositors nor their microsavings could support commissions at typical banks need to have trust in the store, beyond what retail levels. they would normally expect when buying toothpaste · Providing transactional accounts for all. An or a mobile prepaid card. Moreover, with mobile individual's landing point onto the payment banking you can do cashless transactions right from network is his or her transactional account. It needs your home, coming close to the typical "utility" vision to be easy for people to open such accounts, and of water and electricity in every home. Banks, like providers need to see economic benefit from most service providers, can concentrate on product offering such accounts broadly. In particular, know quality and marketing--that is, branding--and can your customer (KYC) requirements need to be leave retail operations to local shops. practical and proportional to the potential risks involved, and financial providers catering to poor The challenges ahead people need to have a more efficient and lighter back-office infrastructure that works for low- Thus far, we have laid out an admittedly utopian balance accounts. view of retail payments in developing countries that · Identifying shared industry models and would enable universal access to finance--call it a business cases that support the appropriate guiding vision. We are still some way from being able level of cooperation and competition among all to make it a reality. With all the attention, and even the parties involved in the payments network. hype, that branchless banking has been getting in Mobile operators, larger banks, MFIs, and other industry circles and in the media, the actual situation grassroots intermediaries each have a role to play 7 in marketing and delivering a range of services But, to use a phrase that reflects a common concern to the population. But of course reaching the in most technology-led projects, if we build it, will appropriate interoperability standards, aligning they come? Looking at examples from other utilities, private incentives, and building the necessary widespread access to electricity triggered demand shared infrastructure takes time. for lighting and heating, but that outcome may not have been so evident when the thought first occurred It's all a matter of setting up the right incentive to someone to connect every home in the land. structure: the customer needs to see benefit in the service, the agent needs to be adequately rewarded, So what would it take for the vision of universal the financial institutions need to find marketing and connection to a payment utility to be translated into selling to poor people worthwhile, and the various a compelling proposition for poor people whose players in the supply chain need to be comfortable economic activities are largely informal, restricted to in sharing and collaborating to minimize overall a smaller community of local suppliers and customers, infrastructure costs. and more sporadic? There are really two parts to the question: In the next sections, we address each of these four points in turn. But they together raise a broader · To what extent would currently unbanked people question of the role of government in enabling, take up the access offer, by signing up to a promoting, and regulating the development of a new transactional account linked to a card or a mobile ubiquitous payments network. Prudential and market phone enabling electronic transfers right from their conduct regulations will need to be adapted to allow home or from the local store? banking transactions to occur in nonbank locations · And then, how many financial services would they and through nonbank parties. Interesting competition use, how often would they use these services, and policy issues will emerge around converging roles of what would they be willing to pay? The potential for telecommunications and financial intermediaries, and extensive use of the network--getting transactions around the development of interoperability standards through--is what is going to pay for it and motivate and shared infrastructures. Consumer protection all the different players to cooperate in building it. regulations and consumer education programs will need to be bolstered, so that customers with It is important to separate the access component no earlier experience with financial services are not from the services component, at least conceptually, misinformed or harmed by inappropriate choices. All because they have different take-up drivers. The these aspects will need to be worked out in parallel access component is driven by customer comfort with with business and commercial aspects, which we the use of the technology platform (whether card concentrate on below. or mobile phone based) and related aspects, such as security, privacy, reliability, and convenience. The Challenge #1: Understanding services component is driven by the relevance and customer drivers pricing of each service. However, in practice we cannot disentangle the two components if we are talking So far we have talked about the supply side of the about targeting previously unbanked customers. The equation: the infrastructure that would make it only thing we can do is observe customer reaction possible for every citizen to gain access to finance. to a joint offer of access-plus-services. If customers 8 reject a branchless banking offer, is it because they true of the mobile operator-backed projects, because are not drawn by the range of services on offer or is it they want to avoid the perception of offering banking because they are not comfortable with the technology services, which would trigger regulation. But what are through which they need to transact? the customer perceptions driving this behavior? Next we explore why branchless banking channels are used This is why it is difficult to answer these questions more for payments than for savings, by looking at based on what we know now. Early branchless three aspects of customers' financial service purchase banking deployments show encouraging signs of decisions: customer take-up. Widespread use of banking agents in Brazil has motivated banks to deploy · Perceived value proposition. How easy is it agents in every municipality. In the Philippines, Smart for customers to assess and validate the value Communications and Globe Telecom together have proposition of electronic payments and savings over 2.5 million active users of mobile money services relative to the options they are using today? (Smart Money and G-Cash, respectively). Safaricom in · Discipline and commitment. What degree of Kenya has managed to register 3.6 million customers personal commitment is required to make electronic to its M-PESA mobile money service within a payments versus saving? year-and-a-half after launch. These are significant · Pricing structure. How do service provider charges achievements. align with customers' willingness to pay for electronic payment versus savings services? But the picture is less compelling when we look at the breadth of services used and the frequency of Assessing the relative value proposition transactions (Ivatury and Mas 2008 and Porteous 2007). In particular, although poor people seem ready When confronted with a new way of doing something to embrace payment services when they are offered (in this case, from traditional methods for undertaking as a convenient and cheap solution, it does not payments or savings to electronic modes), customers translate into greater financial savings. When money need to understand how they benefit by switching-- hits the account, a vast majority of poor people draw offsetting the positives and the negatives of the new it down immediately and in full. This despite the latent system versus the old. Customers are more likely to demand for better savings products in many countries switch if they can (i) easily assess a priori the extra net (see, e.g., Deshpande 2006). benefits and (ii) readily validate those extra benefits and costs as they experiment with the new approach. Why is use dominated by only a few payment services, How do payment and savings services through rather than broadening to all payments made by branchless banking channels compare on these two individuals and households? And why doesn't this use criteria? of payment services "spill over" into use of a broader range of financial services, particularly savings? To Electronic payments seek to replace payment be sure, this is explained in part on the supply side. methods, such as giving cash to the bus driver or to a Branchless banking providers have really marketed friend who is going to a remote village to carry cash to these services as payments solutions, not so much your relatives, relying on informal trust networks, such as store-of-value solutions--they are modeled more as Hawala dealers, or having to go to another town and on Western Union than on a bank. This is especially queue up to pay your electricity bill. When choosing 9 payments methods, customers are likely to value phone or card) at a network of local stores acting as certainty of delivery, convenience, and cost. These banking agents. Yet customers know that the value of correlate well with electronic value delivery versus cash held by the local stores is unlikely to be sufficient more traditional or informal payment mechanisms. to liquidate their entire savings, which may temper Potential customers can fairly straightforwardly count their confidence in the system. Note that the liquidity the benefit in terms of how much less time you need constraint at local cash-in/cash-out points may be a to spend to send the money, how much faster the problem in the case of one-off payments or transfers money gets there, and how much less of it you lose just as it can be for savings, but customers' evaluation along the way. Moreover, the service concept is of the magnitude of the potential problem will be easily understood so that providers can concentrate quite different in the two cases. To entrust a payment, messaging on the benefits.3 And, once you try it out customers just need to believe that the local agents and send money to your grandmother, you can call have enough cash to meet the value of their transfer; her immediately and verify that she did indeed get in the case of savings, customers may assess the value the money. Each transaction provides immediate of cash in the stores' till against the aggregate savings experiential confirmation of the advantages of the of everyone in the village (e.g., in the event of a new system, which reinforces the choice. natural catastrophe or political unrest). What about savings? Let's consider two key attributes: To sum up, the challenge to induce financial savings preservation of value (probability that value will be by unbanked poor people through electronic offerings maintained or enhanced over time in real terms, net is that there is no quick experiential path that will of costs) and liquidity (reliability, convenience, and make them more comfortable with the safety and cost of realizing the value in cash). Comparisons of soundness of the scheme. The benefits may be less benefits and costs of technology-delivered savings apparent than the risks. schemes versus more traditional brick-and-mortar savings institutions (for people who are already Building discipline and commitment banked) or informal or physical forms of savings (for a majority who are not) are a lot harder to make. The fact that electronic payments flourish under electronic banking schemes but savings seem to lag For one thing, experimentation may be costly if there might also be explained by the relative strength of are account opening or maintenance fees involved. customer commitment to the services. Bill payments Also, there is no defined time horizon over which to generally have an external built-in discipline device-- assess preservation of value. For how long does the your power may be cut off. Remittances to relatives customer need to run an "experiment" to validate often involve social pressures that build commitment whether keeping money in the new form of bank and occur in a regular pattern that helps build accessible through electronic means is "safer" than in discipline through repetition. Savings, on the other the savings institution in the neighboring large town hand, may be harder to "bank" because it can always or under the mattress? Therefore, trust is difficult to be postponed--tomorrow, failing that, next week, build because underlying concerns can linger. and so forth. In terms of liquidity, branchless banking relies on How does one build discipline to be able to acceptance of customers' payment device (mobile save? One effective (though costly) way to do so 3 Safaricom's award-winning advertising for M-PESA shows a stream of dollars physically fly out from the sender's mobile phone, which is then "captured" by the receiver's phone. 10 is to borrow. You "invest" the money upfront (for Realizing the willingness to pay instance to buy an asset) and then face a stream of future payments against it. You need to make these There may be a third explanation for the relative payments; the (contractual) obligation to a third party success of payments relative to savings through builds commitment and establishes a clear timeframe electronic channels, having to do with customer for (future) savings. Asset building through lending acceptance of the providers' service charging offers liquidity today and a promise of discipline in structure (not just the level of fees). the future. The success of microcredit stems in large part from the fact that people's demand for savings is Customers are satisfied when the service provider's so strong that they are ready to do so at substantially charging structure aligns with their own perceptions negative interest rates (the lending rate)--when it of value. This is easier to achieve for payments than comes with a commitment device. The alternative, for savings. Customers understand payments to be to "save up" in the terminology of Rutherford (2001), one-off transactions, each of which is concluded in may be harder to achieve because the implications of a relatively short period; hence, charging the sender deferring payments are less immediate. per payment event makes intuitive sense. How does a world with electronic payments change On the other hand, savings entails a sequence of this? Making savings possible through a ubiquitous related open-ended transactions: there is an a priori network of retail outlets should increase the undefined pattern of deposits and withdrawals. In opportunities people have to save--it can be done at this case, when is it appropriate to charge? We know any time, at any number of convenient locations. You that customers dislike account opening fees (I must might need less discipline if it is convenient to do, if pay just for the privilege of joining?), fixed monthly you have many points in time in your daily life at which fees (even if I do nothing?), deposit fees (I have to pay you can make a deposit. On the other hand, one to give you money?), as well as withdrawal fees (you might speculate whether it actually weakens discipline charge me to get what is rightfully mine?). Because we by making it so much easier to delay action and by cannot pinpoint the benefit of saving to an individual reducing the number of weekly events (e.g., trips customer-triggered event, any charging scheme to the major town nearby) that previously triggered seems arbitrary--and hence hostile--to customers.4 explicit savings decisions. Which way the impact goes probably depends on cultural factors. In fact, door-to-door deposit collectors have succeeded in many markets. For example, susu There is an inherent tension between two customer collectors in Ghana collect a prespecified amount drivers we have discussed: I want to precommit to a from customers each day and return the monthly future path of savings, but I want to retain liquidity collections, less those corresponding to one day, back of my savings--which is tantamount to the freedom to their customers at the end of the month. Thus, to not save or even draw down savings at any time. there is a negative interest of around 7 percent-- One avenue worth exploring might be to combine payable each month.5 more effectively the transactional capability of mobile phone banking with the personal communications One can speculate that customers are willing to pay capability of mobile phones through a system of this extremely high fee because it is actually not reminders and incentives. visible at any point in time: the fee is deducted at 4 In principle, the dislike of bank charges is shared by rich and poor people alike. Even affluent people are often irritated by relatively small ATM fees. But the issue may loom larger for people with low savings balances, for whom charges represent a larger percentage of their savings balance. 5 Here is a back of the envelope calculation. Suppose that daily collections are 1, and a month has 30 days. Thus, the collector's fee is 1 and the average monthly savings balance is 15; 1 in 15 is about 7%. If a customer takes this service for 1 year, the collector's fee is 12 and the average monthly savings balance is still 15; 12 in 15 is about 80%. 11 some undefined point between when customers pay on the agent business, we must ensure that they in their daily contributions and when they get the receive adequate incentives to offer and promote sum back at the end of the month. Is it possible that the cash-in/cash-out service. Banking services have electronic forms of savings have the disadvantage the advantage over other products sold at the store that fees need to be accrued on a specific day and in that they do not require physical handling and do appear as a line on statements and, hence, are more not occupy space on shelves or in the back room, but visible to the customer? they still impose a (often disproportionately) large cash-handling burden on the store (with the attendant In other words, if you wanted to run an online susu costs of more frequent trips to the bank and higher collection scheme in which the customer could check security risk). on daily savings balances, would you deduct your fee a little bit with each deposit (so that the customer In fact, the economics for agents in many branchless daily pays 1 but only sees 0.93 hitting his account), banking environments is still uncertain, for two or would you charge it at the end (so that on the last reasons. First, the volume of transactions is relatively day the customer gets 7 percent less than he saw was low--an agent doing 15­30 transactions per day is the balance just the night before)? How would the not uncommon. This is because of the low number customer feel in either case? of transactions undertaken per customer, coupled with the financial provider's need to create a fairly Challenge #2: Making dense agent network to establish a footprint and offer enough convenience to customers. Second, the economics work retail stores are used to product margins in the for retail agents upper single digits or lower double digits, and that sort of commission is hard to sustain on basic Branchless banking propositions have to satisfy the financial transactions, such as deposits or peer-to- needs of agents as much as customers. Agents are peer payments. Customers may not be willing to pay the bridges between the "old" cash economy and fees in excess of 1 percent,6 and the bank may not the "new" electronic payments utility. Without a find these customers sufficiently profitable to absorb workable set of options for customers to cash in and agent commissions significantly beyond that. But if cash out, they will find limited benefit in storing value that is so, is there enough value in it for the agent? electronically, and they will not find much acceptance of their stored value as a means of exchange. There are three possible ways of making the economics work for the agent, which we explore in In standard branchless banking models, agents are turn below: normal retail outlets: airtime resellers, post offices, lottery houses, pharmacies, down to the smallest · Substantially increasing transaction volumes corner shops. They operate on thin margins on · Shifting the incidence of transactions cost away goods and services sold, and decisions about product from the consumer and cross-subsidizing placement on shelves are strictly a matter of which · Lowering the agents' cost of providing the service products sell enough volume given the margin they produce. If we expect more retail stores to take 6 For reference, 1% is what Smart Money and G-Cash in the Philippines, as well as MTN Banking and WIZZIT in South Africa, charge their customers for depositing cash. Most banks in Brazil, Colombia, and Peru that offer branchless banking go for a fixed payment per cash-in transaction instead of a variable fee. Note that this is the commission charged to the user and is split between the bank or operator and the agent. 12 Substantially increasing perhaps up to one per week per user. Significant transaction volumes transaction volumes will develop only when people use electronic payments for their daily life--in the Ultimately, the success of the payments network communities where they live, to pay for day laborers, will be determined by how many transactions flow buy goods at the store, pay local fees, transfer money through it. Transaction volume will reduce unit between family members, and so forth. transaction costs, incentivize agents to participate as cash-in/cash-out points across the territory, and entice However, there is as yet very limited evidence different types of service providers to join the system. anywhere that branchless banking solutions are The existence of branchless banking transactional being used for intracommunity transactions in any channels may not (at least at first) make customers significant volume.7 Many innovative virtual cash transact more in total, so we must see which of the products have been tried in developed countries, existing transactions can be captured electronically. but they have met very limited success, indicating We offer the following considerations. that cash is very hard to displace for smaller in-person payments. The stronger driver for cash substitution As a starting point, branchless banking providers ought to be security, and hence it may be that mobile need to find ways to automatically fill their customers' money products would be particularly successful as an wallets, through direct deposit of salaries and wages, alternative to cash in particularly crime-ridden areas payments for goods sold or services rendered, or in conflict zones. government welfare or pension payments, or inbound remittances. If these payments are made in cash, it Ultimately, transaction costs still need to come down is less likely that recipients will deposit the cash into significantly if we want people to use electronic their transactional account to fund savings or future payments within their communities and to manage transactions. their own daily financial lives. The system should work for transactions of as little as $2--the daily income More broadly, electronic payments should be for many--on agent commissions of not more than particularly well suited for transactions that low- 2 percent. That means that customers' transaction income people need to do with outside parties (large fees should be in the range of 2­4¢ (1­2 percent of a employers, the state distributing benefits, utility $2 transaction).8 Only then will the size of electronic bill payments, remittances from distant relatives or transactions match the daily cash flows of the poor. friends, loan repayments). Service providers need to take an ecosystem view and attract all those remote A final element that can help drive transaction parties who wish to transact with their customers. volumes at agents is to share the agents across all They are in a two-sided market: the market for end- financial providers--much like ATMs are shared in users (customers) and the market for external parties many countries. This way, they can grab banking wishing to transact with their customers (including business from their entire regular clientele (regardless payers, billers, and other financial service providers). of who they bank with) rather than only from those with the "right" bank. And by eliminating needless Being able to make payments to remote parties is duplication of agents (each serving niche banking particularly valuable to users, but this by itself will customer bases) the same number of transactions not generate a significant volume of transactions-- could be handled by fewer agents. 7 The mobile telephony world is leading in this respect as well, with airtime transfer products that allow airtime gifting between friends or parents buying airtime for their children. 8 Person-to-person transactions through the Smart Money and G-Cash systems in the Philippines are within range: they cost 2­5¢. Compare this with person-to-person charges of M-PESA in Kenya (46¢) and MTN Banking and WIZZIT in South Africa (39¢). 13 Seeking customers other than · Likewise, international remittances remain a highly the end-users willing to pay profitable business, with commissions typically commissions for services in the range of 7­10 percent. The success of remittance networks is based partly on the density Ivatury and Mas (2008) noted that the use of a of in-country termination points. Hence, they are branchless banking system is dominated by one or generally willing to provide a significant commission two "anchor" payment services, which constitute the on termination to reward local outlets' participation main drivers for pulling customers into the branchless in their network. banking platform and account for the bulk of transactions. These are utility bill payment in Brazil and Taking the lead from the very encouraging experience airtime reselling and receipt of remittances in mobile- with these services, we need to find more players operator-centric systems, such as in the Philippines other than low-income end-users willing to pay and Kenya. What these three have in common is that commissions for bottom-of-pyramid transactions. branchless banking providers can take advantage of It is these transactions that should be generating a higher willingness to pay by parties other than the the bulk of commissions, permitting lower marginal beneficiary customer: commissions on user-paid person-to-person payments and microdeposits (Goldfinch 2006). · Utilities traditionally incur high collection costs, and hence they are often willing to absorb a fee We see opportunities in this respect for business- for new forms of electronic bill payment by their to-person payments (e.g., salaries), government-to- customers. By shifting transaction costs from the person payments (e.g., pensions and social welfare payer (the customer) to the payee (the biller), banks entitlements), and person-to-government (e.g., taxes engaging in branchless banking can tap into a and duties). Businesses, whether formal or informal, larger willingness to pay. Banks thus have incentives stand to gain significantly from distributing payments to promote electronic bill payment services, and electronically to laborers or suppliers. This offers them customers find the pricing very attractive relative the opportunity to reduce cash handling costs while to their existing options. they increase their control over spending. Similarly, · Over 1 million of the estimated 2.5 million active governments often pay very high prices to deliver mobile money customers in the Philippines are cash payments to rural remote populations on welfare airtime resellers. In fact, most use the mobile programs or pensions. Branchless banking providers banking platform only to convert electronic money should develop strong payment service propositions into prepaid airtime. This is because mobile for them, at a price that is better than their current operators are able to provide a much higher alternatives but that creates significant commission commission on airtime reselling, typically around for agents. 10 percent, funded by the high product margins available to mobile operators. It is no surprise that There is also an opportunity to tie electronic banking most of these stores advertise the airtime product platforms more closely to electronic commerce (attracting a commission of 10 percent) but not the offerings. Airtime reselling by Smart or Globe is an cash-in/cash-out product (attracting a commission example of that: consumers use the mobile money of less than 1 percent). platform to purchase airtime, and resellers use the same mobile money platform to settle with the 14 operator. By linking the payment with a product sale, Regarding the other component of cost, the use of agent commissions can be higher. store resources (staff time, space): do the agents need to be stores? By taking the banking agent Lowering the cost of providing business out of more established stores, which have the service by the agents higher overheads and are used to higher margins on product sales, it might be possible to reduce the cost There are two main costs for stores offering banking of cashing in and out in poorer communities, while at services: (i) cash handling costs (including working the same time creating new livelihood activity. We capital, trips to the bank, and risk of loss or theft) and could thus turn access to the payments system into (ii) use of store resources, including space and staff. new business opportunities for people and hence go There are potential cost-saving opportunities related beyond convenience and cost as drivers of take-up. to each of these two components. There is a powerful precedent: among the 1 million The cash handling costs depend strictly on the timing airtime resellers on the Smart Money or G-Cash and volume of cash-in versus cash-out transactions platforms in Philippines, in addition to many tiny the store has to meet. If the transactions were "sari-sari" stores, there are many youth acting as perfectly timed (a customer deposit being matched "freelancing" resellers within their social circles. We immediately by an equal and offsetting withdrawal), can even conceive of a new kind of susu collector who there would be practically no cost to the store for uses branchless banking technologies to guarantee offering the service (beyond staff time), and it would a higher level of service and trust to his or her be justified in accepting a much reduced margin users. Another area with promise is bill payments: relative to the other products it sells (which require one can imagine roving bill payers who use their financing and warehousing). mobile phones to make bill payments on behalf of customers--and get a fee in return. This model might Financial service providers could use product portfolio not work everywhere, but it is worth exploring. and marketing levers to balance cash-in and cash-out costs at the community level, so that agents could Challenge #3: Providing manage local liquidity with less (if any) trips to the transactional accounts for all bank. Reducing the burden on stores for fulfilling the cash-in/cash-out function in this fashion should result Universal access requires having a set of financial in significantly lower agent commissions and more institutions that, together, market and sell their reliable liquidity for customers. This would require services to all segments of the population. Financial financial service providers to carefully analyze net institutions have to develop business models that cash trends location by location and to devise ways allow them to find all segments profitable. to stimulate use of cash-in or cash-out services as required. For instance, local cash shortfalls could be Or, to look at it from the customer's point of view: countered through attractive new savings products or so far we have talked about use (broadening the getting more employers to pay wages into accounts; range of services and increasing the number of local cash surpluses could be countered by increasing transactions), but what about access? If we define the range of bills that can be paid at the agent or by access as having a transactional account--a landing offering larger loans into the right segment. point on the electronic payment network--several 15 barriers need to be overcome to be able to ensure fixed accounting cost per customer to a level that all have a transactional account: may render serving poorer customers uneconomic. · Cost base. There needs to be a class of financial MFIs can develop cheaper core banking systems that service providers with a sufficiently low-cost are more tailored to the needs of poorer people. But banking platform (the system that maintains they generally face a problem of scale, because they customer accounts) to be able to profitably offer must amortize (lower, in absolute terms) the fixed accounts to all. system costs over a much smaller asset base. · Revenue model. Recognizing that low-income people have transactional needs but are likely to This makes it difficult to expect rapid progress in maintain relatively low savings balances, financial banking the very poor or destitute, at least on a purely service providers targeting low-income segments commercial basis. There are two basic approaches need to shift from a float-driven model to a that pro-poor market participants could adopt to transactional-driven one. solve this problem. One approach is for MFIs to share · Information issues. All need to be able to provide the back-office system. The shared core banking the necessary information to open an account. system would be developed, managed, and hosted by a third-party application service provider (ASP). Low-cost banking The ASP might adopt a leaner core banking system, reflecting the more basic needs of MFI customers, and To take full advantage of the payment network, and spread the cost over a larger volume of customers. in particular the savings services attached to it, I need This, however, would make the participating MFIs a transactional account. Who will give me one? The very dependent on data connectivity to be able to answer may not be that encouraging when it comes access the banking application remotely. to poor people with no regular sources of income, with low financial savings, and with the need to make The second approach may be to specialize a banking small transactions. For many banks, the revenues and platform to support only one product--transactional costs associated with putting poor people onto their accounts. Through specialization, the core banking banking platforms just don't add up. system could be rid of much of its complexity, and the end result would be something much akin to The complexity and cost of traditional core banking the prepaid platforms used by mobile operators to systems are driven by the high platform requirements maintain the airtime balances of their customers. This for reliability, accuracy, and security; the need to approach can be particularly impactful if nonbank obtain a consolidated picture of the customer across financial institutions (including mobile operators many different products; and regulatory reporting themselves) are allowed to issue prepay cards or requirements. The requirements of core banking other forms of electronic money. By allowing players systems are generally specified against the needs of with more brand affinity for and marketing presence more sophisticated customers, who are presumably with the mass market to offer limited savings and more profitable. Regulations also often mandate transactional services, today's unbanked customers onerous minimum requirements, for example, in would have an alternative path to join the payment terms of maintenance of records. All this raises the network, and it would stimulate competition for financial services at the much neglected bottom of 16 the pyramid. Of course, such players would need to then, can I get on it and use it the way I want to? be licensed with due care for the preservation of the Unfortunately, there are some things about me that public's moneys. service providers will need or want to know, and without that information, they will deny me service. If the market fails to achieve sufficiently broad inclusion There are some information sources that need to be despite these lower cost alternatives, there are two attached to the low-cost payments network described more heavy-handed options left for government. The above to achieve universal access. first one is to mandate universal access for a defined basic account (at a defined price point for customers), The first and fundamental aspect relates to proving in effect requiring banks to cross-subsidize loss-making identity--a legal requirement under Anti-Money customers from the broader pool of customers. This is Laundering and Combating the Financing of Terrorism a common practice in utility sectors, where customer (AML/CFT) rules. It follows that every citizen should fees are geographically averaged (e.g., the customer be able to conveniently demonstrate proof of pays the same for a rural or an urban telephone line, identity. There are certain things that the state can despite the costs being very different). In the case of do to facilitate this: ensuring that there is a single or banking, this is likely to be ineffective: banks will have a combination of cards that are trusted and singly or mandated low-cost products on their portfolio but collectively cover the entire population; providing will neglect to market them effectively. The second electronic access to identity databases so that service option would be for government to offer a subsidy providers can check the validity of identity cards to make poorer or rural customers appear more presented to them; and allowing authorized third profitable for banks. However, we do not yet know parties to physically check identity cards (matching enough about the economic impact from achieving photos and signatures) on behalf of service providers universal financial inclusion to justify such a policy under reasonable risk-related limits. There might also action. be exemptions from proof of identity requirements in low-risk situations where customers can maintain only Nontraditional revenue models small balances and undertake small transactions. For banks to implement a low-cost banking platform is The second informational aspect relates to the not enough to ensure an inclusion or pro-poor focus. ability of customers to borrow. Having affordable Given the low savings balances held by poor people, access to credit may be an important driver for using they will also need to shift their revenue model from branchless banking services. Lending is subject to the traditional float-driven one to a transaction-driven well-known informational asymmetries: I know how one. Also, they will need to develop ways to cross- good a borrower I am; I know what I will do with sell services to customers (from payments to savings the money once I get it; I know the quality of the to insurance, for example) to enhance customer project I want to fund; I know how hard I intend to profitability. work if I get the loan--but the bank cannot be as sure about these things as I am. If these information Knowing me, knowing you asymmetries are not addressed, they are likely to lead to an undersupply of credit (or "credit rationing"). This Let's say there exists a low-cost payments network happens because, if the bank responded to higher in my village (over the air, at retail shops). Even perceived credit risks by raising the interest rate 17 beyond a certain threshold, "safe" borrowers would economies of scale within its scope of activities while be discouraged by the onerous interest rate, leaving at the same time broaden the service offering to its credit demand from only "risky" borrowers who do customers. not care much about the interest rate (because there is a good chance they won't be able to repay anything Interoperability--the basis for sharing--is one of at all anyway). the key considerations of any networked business, and the payments network is no different. To bring There are some pieces of market infrastructure that it back to the customer: once customers have an can help address these information asymmetries. account, they will want to use it for many types of Credit bureaus can report on the past credit history payments. Indeed, the value to customers of joining of borrowers, which can be a proxy for "character" the network depends on how many of their friends, or willingness to repay. Collateral registries can be business counterparties, utilities, financial institutions, used to verify the legal title of property, which is agents, and so forth, are connected to that network. pledged as collateral. By letting service providers If all account-issuing institutions are able to connect know these things about me and my assets, they will to the same national payments network, then it will be in a better position to more accurately "price" the be possible to use any agent for cash-in/cash-out services they are willing to offer me. This can result in and undertake any-to-any payments, and value to all increased use of financial services through branchless customers in principle would be maximized. banking channels. But also branchless banking can help draw people into savings products and build a But the story is more complicated on the supply positive financial profile, which can then become a side. Because for providers, bigger is better in terms stepping stone into credit or other products. customer numbers, larger institutions may in fact opt to maintain their customers on a closed network Challenge #4: Interoperable (i.e., not to interoperate with others) in an attempt to business models turn their size into a comparative advantage. Smaller competitors would have more to gain than they In the above discussion, sharing is a recurring theme: would from sharing customers and platforms; indeed, the agents might be shared (just as the ATMs that interoperability is a great "equalizer." So much so, they functionally replace might be), the mobile phone that if interoperability were mandated at an early platform may be shared, the credit bureau is shared stage of development of a network, it might actually by definition, and even the core banking system may kill off innovation and incentives for growth. Why be shared under an ASP model. Financial service invest to create and grow a network if others will be providers might choose to share rather than build able to latch on at a later point in time and take full their own channels or systems for entirely pragmatic advantage of the network size? reasons: to reduce cost, to accelerate speed to market, and to extend reach beyond what it might Indeed, regulators are rightfully very cautious about be able to achieve on its own. Sharing most bits with mandating outcomes in this area. It is generally other providers allows the provider to specialize better to let the markets work out the optimal timing in what is really important to it, whether that is to and wholesale (interprovider) pricing that drives leverage specific core competencies or to target an interoperable outcome while recognizing the specific markets. In so doing, the provider can harness differential contributions and benefits each player can 18 realize. Interoperability does not mean "free," but of keep track of how much money they owe each course if compensating prices between players are other. For small enough transactions, they are excessive and those are passed on to the customer, recorded and periodically netted off against each then customers may opt out of those services, and other. A clearing function also allows bulk billers the benefits from interoperability will exist only on like utilities to request payment instructions from paper. Regulators need to keep a watchful eye for their customers at multiple banks. competitive dynamics in the payments network, · Interbank settlement. Interbank amounts due but they should avoid intervening too hastily or need to be settled from time to time (i.e., moneys dogmatically. actually need to move from one bank to the other). This is done by transferring money between the Thus the key issue will be how to manage coopetition: accounts that each bank maintains at an agreed to work cooperatively with other providers to build clearing bank--often the central bank. the basis of an interoperable network within which each provider can competitively solicit business from Banks can achieve interoperability at any of these a broader range of people. levels by striking bilateral agreements between them. It is much more common, though, to rely on specific To realize the full shared agent network and any-to- institutions to perform each of these roles: an ATM/ any payment scenario, interoperability would need to electronic funds transfer­POS or card payment switch, happen at several levels: an automated clearinghouse (ACH), and a real-time gross settlement system, respectively. The first two · Terminal equipment. POS terminals through which are often owned by a consortia of banks, while the customers and merchants transact need to accept third is often owned by the central bank. There may payment instruments of any of the participating be multiple competing switches or ACHs, but these banks. Otherwise, merchants would need to deploy also can be interconnected to present a seamless specific infrastructures for each bank, which would interworking from the point of view of customers. make it too burdensome and costly for merchants to work with many banks. Although such an interoperable system should, in · Transaction switching. Customer instructions (for principle, offer a higher level of efficiencies through a cash withdrawal, deposit, or account transfer) aggregation of transactions, such efficiencies may need to be passed on between financial institutions fail to materialize if joint decision making causes involved at each end of the transaction in real suboptimal infrastructure investment decisions. And time, so that the customer can walk away from even if efficiencies are reaped, they may not translate the ATM or banking agent with full and instant into lower cost services for end-users if the benefits gratification (cash in hand, deposit receipt, transfer are appropriated by the operators of the shared confirmation). This provides for transaction finality infrastructure who have a bottleneck control over the from the customer's point of view, even if the entire transaction processing chain. transaction is in the account of another bank and the banks have not yet settled between them. Summing up · Interbank clearing. Transactions for customer accounts that cross over multiple banks need to In the traditional banking environment of high- be accounted for, so that the relevant banks can cost proprietary distribution channels and back- 19 office systems, the corresponding high direct and propagate across the territory? In this endeavor, what indirect costs to customers have resulted in very is the right balance of priorities between fostering limited penetration of current accounts within the technology-enabled innovation and ensuring system population. Many poor and rural people manage stability; between sparking competition and driving their lives without using the services of the formal toward interoperability; between creating financial banking system. For some it is simply not an option, choices for and protecting consumers? whether because of lack of physical bank presence, lack of appropriate documentation, or prohibitive The barriers to customer adoption have less to do cost. For others, the perceived value proposition is with attitudes about technology (fear of change, simply too weak or too fuzzy relative to their current unfamiliar user interfaces, security concerns, etc.) approaches to managing their savings and payments, than with communicating precise customer benefits which may include cash under the mattress, saving in and the financial incentives (or business case) for the physical goods or livestock, and so forth. And in some various providers. Consider the one million tiny "sari- cases, it may be general perceptions, combined with sari" stores and resellers in the Philippines that use insufficient awareness of what banks have to offer, the operators' mobile money platforms to recharge that preclude them from trying something different customers' airtime: they show that an electronic value to what they have always done. Whatever the case, transfer system can go mass market--provided users what formal financial institutions have to offer does have a real incentive to use them. With a population not appear compelling enough for these potential of some 17 million households, that's one out of 17 customers. households in the Philippines taking it up as a side business. Sometimes customers who are new to savings accounts take up the service primarily because it References is a prerequisite to something else: to get a loan from a bank or an MFI, to collect a salary, to receive Deshpande, Rani. 2006. "Safe and Accessible: Bringing remittances from friends and family, or to claim social Poor Savers into the Formal Financial System." Focus payments from the government. The savings product Note 37. Washington, D.C.: CGAP, September. is "pushed" by the provider rather than "pulled" by the customer. There is nothing wrong with that: use Goldfinch, Peter. 2006. "Can Developing Countries of certain financial services requires access to the Afford a Developed World Payments System?" The payments network. But that can be only the beginning Goldfinch Report, Issue 10. Auckland, N.Z.: GFG of the road to financial inclusion. Group, November. Branchless banking offers an opportunity to Ivatury, Gautam, and Ignacio Mas. 2008. "The Early dramatically slash transactions costs and expand Experience with Branchless Banking." Focus Note 46. geographic coverage of formal financial services. Washington, D.C.: CGAP, April. Offering relevant, compelling products will be essential to trigger demand through these new channels. Yet Ivatury, Gautam, and Mark Pickens. 2006. "Mobile realizing this potential also poses several policy Phone Banking and Low-Income Customers: Evidence questions. What is the role of government in causing from South Africa." Washington, D.C.: CGAP and low-cost retail payment and banking infrastructures to United Nations Foundation. No. 50 October 2008 Lyman, Tim , Mark Pickens, and David Porteous. 2008. Porteous, David. 2007. "Just how transformational is Please share this "Regulating Transformational Branchless Banking: m-banking?" FinMark Trust. Available online at www. Focus Note with your Mobile Phones and Other Technology to Increase finmarktrust.org.za. colleagues or request Access to Finance." Focus Note 43. Washington, extra copies of this paper or others in D.C.: CGAP, February. Rutherford, Stuart 2001. The Poor and Their Money. this series. United Kingdom: Oxford University Press. Mas, Ignacio. 2008. "Being Able to Make (Small) CGAP welcomes your comments on Deposits and Payments, Anywhere." Focus Note 45. World Bank. 2006. "The Role of Postal Networks this paper. Washington, D.C.: CGAP, April. in Expanding Access to Financial Services." Global Information and Communication Technologies All CGAP publications Mas, Ignacio, and Hannah Siedek. 2008. "Banking Department Discussion Paper. Washington, D.C.: are available on the CGAP Web site at through Networks of Retail Agents." Focus Note 47. The World Bank. www.cgap.org. Washington, D.C.: CGAP, May. CGAP Mas, Ignacio, and Kabir Kumar. 2008. "Banking on 1818 H Street, NW MSN P3-300 Mobiles: Why? How? For Whom?" Focus Note 48. Washington, DC Washington, D.C.: CGAP, June. 20433 USA Tel: 202-473-9594 Fax: 202-522-3744 Email: cgap@worldbank.org © CGAP, 2008 The author of this Focus Note is Ignacio Mas, an adviser in Hannah Siedek, and Jeanette Thomas for useful comments on an the Technology Program at CGAP. The author wishes to thank earlier draft of this paper. Gautam Ivatury, Kate McKee, David Porteous, Rich Rosenberg, CGAP publications are frequently cited in other works. The suggested citation for this paper is as follows: Mas, Ignacio. 2008. "Realizing the Potential of Branchless Banking: Challenges Ahead." Focus Note 50. Washington, D.C.: CGAP, August.