78234 focus note Advancing Financial Inclusion through Use of Market Archetypes T he evolution of money from physical cash to digital form is redefining financial services as an information business. This, in turn, is generating • In a Convergence Battle market archetype, branchless banks and retailers (and perhaps MNOs in the future) fiercely compete for the same optimism around the long-term prospect of cashless customer in urban areas, while the countryside or “cash-lite� societies, where most people have remains underserved. access to low-cost, convenient, and broadly available • In the Pervasive Social Banking market archetype, financial services. Research indicates that these digital historical financial inclusion success achieved cash models (often called branchless banking, mobile through social banking leads to regulations that banking, or mobile money) can increase financial heavily favor future social banking, leaving less access for unbanked segments by reducing the room for innovation. cost-to-serve for providers and making service more convenient for customers. Branchless innovators who For each of these market archetypes, this paper “get it right� can help accelerate the pace at which suggests a distinct agenda that can help lead the financial inclusion happens. market toward financial inclusion. This evolution will create confusion before it creates The Legacy Economics clarity. It will shake the competitive game board of Bank Branches by shifting which industry players create economic value and what role they play—a process that is Only a few short years ago the physical nature dynamic and often difficult to predict. As branchless of currency dictated that banking operate within financial ecosystems develop and markets begin the nine-to-five, brick-and-mortar, labor-heavy to shift, industry players often cite diverse (and constraints of a traditional retail business. If there conflicting) views about their roles. In parallel, is one factor that drives success of any traditional policy makers struggle to promote regulation that retail activity, it is the high frequency of the right can move the market forward since the “forward� type of customer passing by a retail store, in this step is not clear. case, a bank branch. This explains the underlying basis for the old retail adage “location, location, This Focus Note provides a framework that location.� regulators, policy makers, financial service providers, donors, and investors can use to identify the most Research suggests that we can extrapolate this productive next steps in their respective markets. It retail principle from the street corner to the country argues that countries can be broadly grouped into level. An analysis of 148 countries covering 5.6 No. 86 three market archetypes—distinguished by broad billion people suggests that population density April 2013 economic, demographic, and policy environment and per capita income influence financial inclusion, characteristics—that represent three different as illustrated in Figure 1. This chart computes the Xavier Faz and Ted Moser starting points in the journey to financially inclusive population-weighted adoption of formal bank ecosystems. Branchless financial ecosystems, accounts in countries clustered along dimensions therefore, develop differently in these markets. of gross domestic product (GDP) per capita and population density (Demirguc-Kunt and Klapper • In the Mobile Leapfrog market archetype, 2012). These data show both income and density mobile network operators (MNOs) fill a banking matter (income being the most powerful by roughly infrastructure gap to increase the percentage of a factor of two). Combined income and density is the population that has access to services. associated with still-higher financial inclusion. 2 Figure 1. Percentage of adults (15+ yrs and older) with a formal bank account Weighted average per group of countries 60 40 33 20 1,000 0 64 60 60 Population Density 40 40 Pop/Km2 23 20 20 125 0 0 60 60 60 49 40 40 34 40 18 20 20 20 0 0 0 $4,000 $8,000 $15,000 GDP/Capita Current USD Source: Demirguc-Kunt and Klapper (2012). Income and density seem to be significant factors, but how these two dimensions impact financial are probably not the only factors involved. Several ecosystems. By “ecosystems� we mean the type other factors may lead a market to have a different of actors involved and the role they play in driving financial inclusion level than its income/density financial infrastructure. Comparing ecosystems profile would suggest. One example is a country’s by country (banking infrastructure, cell phone cultural attitude toward banking, which can lead to infrastructure, and presence of retail chains) and a different level of inclusion than expected. Other the breadth of financial services offered across examples are countries that have had rapid recent socioeconomic segments in urban and rural economic growth. Because the banking industry localities, we found distinct links between income/ often takes time to catch up with such rapid growth, density market groupings and financial ecosystems: retail banking shows less penetration than would be predicted. Yet other examples are countries that only • Within common income/density country groups, recently privatized their economies after decades of financial service ecosystems seemed relatively similar. state control; in this case, financial inclusion may be • Across different income/density country groups, higher or lower than expected. ecosystems had meaningful differences. Financial Ecosystems and In other words, countries with similar income/ Income/Density Archetypes density characteristics travel the road to branchless banking from similar starting points, while countries Having seen the variation in financial inclusion with different income/density characteristics make across income and density, we wanted to explore their journey from different starting points. 3 Box 1. The Positive Disruption of Branchless Banking The power of income/density combinations to explain the combined cost of establishing a service point current financial inclusion levels (Figure 1) implies that and carrying out a transaction. By using digital and cost reduction and elimination of the distance effect mobile technologies and leveraging existing retailer remain two key factors in expanding financial access. footprints, these new models bridge gaps in branch • Cost reduction. Innovations that reduce customer infrastructure to make customer convenience a reality acquisition, customer transaction, and customer and offer more affordable services. service costs empower financial services providers to serve lower-income clients without compromising Figure B1.A illustrates this disruptive power. Banking profitability. agent transaction costs can beat branch costs by 50 • Elimination of the distance effect. The time and cost percent; automatic teller machine (ATM) transaction that an individual must spend travelling to access costs in high-traffic locations can beat branch costs by a financial service is a major inhibitor of adoption. as much as 90 percent. Even more encouraging, the If this distance effect can be eliminated, pent-up fast pace at which technology is evolving suggests that demand (and new service adoption) should follow. today’s digital money models are still in their infancy: high-speed wireless networks are still being rolled These factors help explain why one of the most out, mobile handset prices (including smartphones) potent enablers of new financial inclusion is the are falling toward mass affordability, self-service retail deployment of technology-enabled branchless kiosks carry out increasingly sophisticated tasks, and banking models. By escaping the economics of stand- business model innovators keep entering the market. alone, labor-intensive branches, these models reduce Figure B1.A. Transaction cost through branches, agents, and ATMs Example from Latin American financial institutions Normalized transaction cost USD/transaction 1.2 1.1 1.0 0.9 Cost per transaction at branches 0.8 USD 0.7 – 1.0 Cost per transaction 0.7 USD 0.6 0.5 Cost per transaction at agents 0.4 USD 0.3 – 0.6 0.3 0.2 0.1 0.0 0 2,000 4,000 6,000 8,000 10,000 12,000 Number of transactions Source: World Bank (2010); Wireless Intelligence (2011); CGAP Country Notes (2012). Figure 2 identifies the countries that fall into the of the population living in each cell of the matrix is income and density matrix from Figure 1 using the noted in the small rectangles. same breakpoints. Countries are color coded based on their region, and only countries with more than The extremes of this data set show the clearest 3 million people are included. The aggregated size market groupings, and as such, we have 4 Figure 2. Financial inclusion environments Countries with population > 3 million and GDP/Capita < USD 15,000 Total Population in group: Regions: Million Sub-Saharan Africa Latin America & Caribbean East Asia, South Asia and Pacific Europe and Central Asia Middle East and North Africa Pervasive Social Banking Bangladesh India* Indonesia* 1,000* 1,634 Burundi Nepal El Salvador China Rwanda Pakistan Guatemala Thailand Malawi Philippines Haiti Dominican Republic Nigeria Sri Lanka West Bank & Gaza Uganda Vietnam 681 1,426 125 Mobile Leapfrog Convergence Battle Population Benin Mauritania Bolivia Angola Albania South Africa Russian F. Density Burkina Faso Mozambique Honduras Colombia Azerbaijan Malaysia Turkey Pop/Km2 Cameroon Niger Nicaragua Ecuador Belarus Argentina Central Afr. Rep Senegal Paraguay Peru Bosnia & Herz. Brazil Chad Sierra Leone Egypt, Arab Rep. Algeria Bulgaria Chile Congo, Dem. Rep. Somalia Iraq Iran Serbia Costa Rica Congo, Rep. Sudan Morocco Jordan Turkmenistan Mexico Cote d'Ivoire Tanzania Syria Tunisia Panama Eritrea Togo Yemen Uruguay Ethiopia Zambia Armenia Venezuela, Ghana Zimbabwe Georgia Croatia Guinea Afghanistan Kyrgyz Republic Hungary Kenya Cambodia Moldova Kazakhstan Liberia Laos Tajikistan Lithuania Madagascar Papua N.G. Ukraine Poland Mali Uzbekistan Romania 831 286 810 $4,000 $8,000 $15,000 GDP/Capita Current USD * Minimum 100 million contiguous population at >1,000 pp/km2 Source: World Bank (2011). highlighted and labeled three corners of the that exist are not widespread. The customer bases matrix as distinct market archetypes. While any of commercial banks or retail chains are typically effort to define a world’s worth of markets in only only 15–25 percent of the population. three categories will be full of exceptions, we believe that these three do a robust job of making In contrast to banking and retail chain development, strategic distinctions. Middle cell environments are MNOs have developed significant networks for more likely to be hybrids of these three archetypes distribution of prepaid airtime in convenient top-up than distinctively different environments. A brief locations that help them gain and retain customers. summary of the characteristics of each market While these airtime networks are loosely structured, archetype follows. when combined with high mobile phone penetration rates they put MNOs in touch with 65–80 percent of 1. Mobile Leapfrog markets the population. In these markets, MNOs often take the (low income, low density) lead in launching mobile banking initiatives, particularly money transfers. Some also create partnerships with In these markets, individuals live too far apart and microfinance banks or small commercial banks to offer account balances are too small for widespread more holistic branchless banking solutions. bank branch economics to work well. As a result, consumer retail banking systems are generally 2. Convergence Battle markets underdeveloped, with little banking infrastructure (higher income, low density) and few branch-based access points. By the same token, the retail sector remains largely As the name implies, in these markets all major types fragmented—there are few retail chains, and those 1 of branchless banking providers (banks, retailers, and 1 Throughout this paper, the term “retail chains� refers in particular to formal retail stores with a relatively large number of outlets (such as convenience stores, supermarkets, pharmacy chains, among others). These chains coexist alongside informal merchants. 5 MNOs) have a strong enough market presence that they often plays a key role in the formation of these banks, converge on and fight for each other’s customers. Higher there is a “public good� nature in these businesses. income per capita and higher urbanization enable major The result is scaling of microfinance for the poor, and commercial banks to have a relatively strong financial some socially oriented commercial banking for lower- inclusion footprint (often 60 percent or more of the income segments. Financial service penetration figures population). Retail chains are also strong and possess are high relative to GDP per capita. At the same time, sufficient operational expertise to carry out financial there are flaws in the system—many accounts end up services partnerships or directly provide financial services. inactive, because commercial banks seeking to meet MNOs have networks that also reach 60 percent or more government social banking targets provide accounts of the population, but they do not usually offer direct to individuals who don’t use them, or some customers financial services. Instead they distribute the services of are excessively offered credit, creating a microlending major commercial banks, typically because regulation bubble and over-indebtedness. prevents them from offering their own services, but sometimes because banks’ brand equity in financial Because government and bank (commercial/ services is so strong. Customers’ expectations of nongovernmental organization) collaboration has convenience (e.g., ability to bank through their mobile advanced financial inclusion in the past few decades, phone, conduct mobile payments, or bank where they policy makers continue to favor schemes where banking shop) often force collaboration across actors, even when entities take the lead in branchless banking partnerships they are fierce competitors. and MNOs and retail agents adopt supporting roles. Competition for value between commercial banks and Rural areas represent the main challenge for financial MNOs in the context of this uneven playing field has led inclusion in these markets. Commercial banking to slow-to-form and slow-to-act partnerships. penetration in these areas remains low because traditional financial institutions lack the required Financial Inclusion across business case for branch-based operations to be Market Archetypes profitable, and while agent banking is beginning to develop, it often mostly focuses on urban branch To provide a more in-depth profile of countries in decongestion. In addition, microfinance institutions each market archetype, we have profiled financial (MFIs) in this environment struggle to generate inclusion in nine countries (three per archetype).2 sustainable financial results at reasonable interest Figure 3 illustrates some key characteristics. The rates because labor costs are too high in relation to height of the bars represents the percentage of the the size of the loans needed by poor households. population that has a bank account. The bottom part (darker color) represents commercial bank adoption; 3. Pervasive Social Banking markets the top part (lighter color) represents penetration (low income, high density) of microfinance. Retailers often host the services of banks, so they don’t appear here as separate figures. The key characteristic of this market archetype is the pervasiveness of social banking, implemented through On average, the effect of having a significantly high noncommercial banks and through government population density has resulted in 2 to 2.5 times more mandates for commercial banks to partially address banking penetration3 than for countries with roughly social equality. High population density near branch the same income per capita (but low or average sites creates attractive economics for social banking. population density). Some of the difference is driven Client incomes and account sizes are low, but so are by commercial banking (via social mandates); the rest labor costs, creating the potential to profitably bank is driven by microfinance. Current initiatives in Mobile low-income customers. Because the poor are the Leapfrog countries are likely to change this picture, largest segment they serve, and because government but for the time being, those initiatives have not yet 2 These countries were selected considering the progress of adoption of branchless banking, the breadth of business models observed, and geographic coverage. 3 Penetration of formal financial services by commercial banks, MFIs, or any other formal financial service provider. 6 Figure 3. Social and commercial banking penetration levels (by market archetype and country) 100% 90% 79% = Micro�nance Banking 80% 1% = Commercial Banking 70% 64% 63% 1% Population 60% 7% with Bank 55% Accounts 50% Percent 15% 41% 40% 40% 78% 30% 19% 25% 63% 23% 56% 28% 2% 12% 20% 40% 5% 10% 10% 22% 22% 18% 12% 10% 0% India Indonesia Bangladesh Pakistan Ghana Senegal South Africa Brazil Mexico GDP per capita 1,489 3,495 735 1,194 1,570 1,119 8,070 12.594 10,064 USD Pervasive Mobile Convergence Social Banking Leapfrog Battle Source: World Bank (2011); CGAP Country Notes; Reserve Bank of India; MicroSave (2011); USAID (2011); The MIX (2011); Bangladesh Bank. led to high bank account penetration—early wins getting there—after all, if branchless economics have been in providing remittances or person-to- make sense, why wouldn’t branchless banking just person money transfers, which are not counted here. happen quickly and naturally? The reason is that there are often several sources of friction. The Countries in Convergence Battle markets have opportunity to grow down-market might be unclear; a markedly higher financial inclusion rate, but in players may be concerned that they are investing contrast to Pervasive Social Banking markets, this in a marginally profitable segment of the market. progress has been led by commercial banks in In other cases, regulations may be outdated but tandem with retail store partners. Microfinance in supported by industry groups who profit from the these environments has had a lesser role in driving way the market has worked in the past. In still other banking penetration, possibly due to a higher cases, players who lead a partnership might design income per capita (which impacts cost structure) the partnership’s revenue and profit splits in a way relative to client account or loan size. that gives them disproportionate benefits, thereby creating a disincentive for supporting partners to Different Starting Points, invest, or to participate enthusiastically. And even Different Journeys if partner economics are equitable, the long-term question of “who owns the customer?� may make The three market archetypes represent three some critical partners hesitate. different starting points in the journey to financially inclusive ecosystems. The technology and business Because the issues that cause this friction vary rationale that enable branchless banking models across market archetypes, it is useful to think of a may be universal, but they will have a different “financial inclusion agenda� as one that identifies impact on the financial ecosystems found in each the main barriers in the journey to a financially of the three market archetypes. inclusive ecosystem, as well as the biggest opportunities that need to be pursued for a greater It may seem unnecessary to focus on the “journey� percentage of a country’s population to be served toward branchless banking or on an “agenda� for through formal financial services. Such an agenda 7 would necessarily focus on low-income populations, markets’ low-income/low-density profiles make but might not be restricted to that group; in some commercial, branch-based banking unattractive. countries the emerging middle class may be Banks in these environments tend to focus their unbanked as well. An effective financial inclusion business on providing financing to governments, agenda would help prioritize efforts and align key larger businesses, and the wealthier part of the market players along a consistent roadmap. As population (about 20 percent). They lack a strong expected, each of the three market archetypes has retail branch footprint. a distinct financial inclusion agenda. The same income/density profile hinders rapid development of large, organized retail chains. The remainder of this paper describes three distinct Poorer consumers won’t typically shop at branded financial inclusion agendas, one for each of the destination stores—the time and cost of travel is too three market archetypes. We hope that this type high, and lower prices can be found in the informal of thinking will strengthen efforts by enterprises, retail sector. As a result, global retailers allocate their investors, and policy makers to accelerate financial investment capital to other geographies first. Not inclusion in their respective markets. only does this preclude retail chains from playing a major financial services role, it also makes agent Mobile Leapfrog Agenda network development more difficult and expensive for banks since they do not have the option of Mobile Leapfrog markets are characterized by low conveniently partnering with large retail chains. income per capita and low population density. In This is where the MNOs’ advantages come into this context, three market dynamics play out. play. They already have a commercial relationship with millions of customers—mobile phone 1. MNOs are best positioned to drive financial penetration in Mobile Leapfrog markets ranges inclusion because they have better economics from 60 percent to 90 percent (see Figure 4). for branchless banking than banks. Because For all practical matters, MNOs already have bank branch profitability is driven largely by the their own version of a retail network—the airtime number of individuals with sufficient income living distribution network that they’ve developed as within a convenient radius of the branch, these part of their voice business. Furthermore, MNOs Figure 4. Mobile Leapfrog market archetype Sample Country Profiles Ghana Senegal Pakistan % Penetration Access % Penetration Access % Penetration Access (1) (1) (1) Points(2) Points(2) Points(2) MNOs 87 - 67 - 59 - Commercial Banks 18 1,317 10 400 22 5,546 Retailers 20-25 20-25 100 Size of largest chain - - - [all formal retail outlets](3) [ 550 ] [ NA ] [ 6,000 ] Notes: (1) Penetration of MNOs based on estimates of unique users as % of total population. Penetration of commercial banks based on % of adults with access to at least one financial service from commercial banks. (2) Number of access points for commercial banks estimated as # of bank branches plus # of ATMs; banking agents not considered since most existing agents do not serve bank customers. (3) Retail chain access points is a lower-bound estimate based on formal retail chain store outlets (including among others: food stores, pharmacies, textile, hardware, construction material, convenience stores, gas stations). Source: World Bank (2011); Wireless Intelligence (2011); Bank of Ghana Annual Report; InterMedia market research (Ghana); USDA Foreign Agricultural Reports (Pakistan, Ghana, Senegal) (2007-11); CGAP Financial Access Survey (2010); State Bank of Pakistan (2010); CGAP Country Notes for Ghana, Pakistan and WAEMU (2012). 8 Box 2. The Power of Partnerships Figure B2-A measures the cost of delivering financial  ot surprisingly, banks retain the lowest cost of storing • N services via the three different actors commonly funds because the cost of complying with prudential involved in branchless banking services. It incorporates regulation is marginal to their core business. the cost of acquiring and retaining a customer, the MNOs have the lowest transaction costs because their •  cost of safely storing a customer’s savings, and the core business already bears the cost of the infrastructure cost of carrying out related financial transactions in required to initiate and process transactions. a scenario where more than 1 million customers are being served through a branchless banking model. Note that a hypothetical three-way partnership using The chart shows the following: the retailer’s ease of acquiring customers, the bank’s cost of storage, and the MNO’s transaction efficiency • The cost of acquiring a customer is an important yields a total cost position 60 percent below that of part of the equation. MNOs and retailers can have a traditional bank branch approach. It is no surprise lower costs than banks because they can leverage that partnerships are often a goal if not a reality in existing infrastructure and customer bases. branchless banking ecosystems. Figure B2-A. Estimated unit cost of �nancial service provision 14 12.7 12 Banks, MNOs, retailers combining to leverage their 4.0 respective comparative 10 advantages 0.5 8 USD 6.6 6.9 6.9 6 2.1 3.2 4.8 4 8.2 1.4 2.1 1.4 0.5 2 3.1 2.3 2.3 0 Bank MNO Retailer Best-in class actor Acquisition Safe store-of-value Cost to serve USD/Cust USD/Cust/Mo USD/Cust/Mo Source: CGAP analysis of retailer, bank, MNO economics (2011) can find a suitable return on investment delivering given tacit permission, but not explicit, written financial services that yield moderate or low stand- approval to MNOs. This hinders full investment by alone profitability, because they have already MNOs out of concern that their rights to operate paid to acquire their customers and because may be removed or that future reforms might financial services bring additional core business reduce the attractiveness of the business. Clear synergies, such as lower customer churn, added and unequivocal regulatory permission for MNOs average revenue per user, and savings on airtime is a foundational step toward financial inclusion in distribution (see Figure 5). Mobile Leapfrog countries. Given that MNOs represent the Mobile Leapfrog Given this context, MNOs are assuming team market’s best chance for financial inclusion, it is leader roles in Mobile Leapfrog geographies, not surprising that central bankers and financial recruiting banks as junior team members whose supervisors in these markets are among the most role is to provide licensed deposit-taking globally progressive in granting MNOs a role in accounts. In some cases, MNOs choose to financial services. That said, some regulators have partner with a regulated microfinance bank that 9 Figure 5. MNO business case — Share of direct and indirect revenue for two major mobile money implementations % of Annual Revenue INDIRECT REVENUE Churn reduction 1% Airtime distribution savings 29% 33% 11% 12% Voice use uplift 3% 59% 52% DIRECT REVENUE Major implementation Major implementation West Africa East Africa Source: CGAP analysis of mobile money business case (2011); GSMA MMU (2011). has a large client base among the poor or with financial products and segments) that MNOs don’t a second-tier commercial bank (Safaricom did have. both in Kenya). To maximize control and return Payments have generated a significant amount of on investment, an MNO might even buy a bank social benefit where they have achieved scale. The (as Telenor Pakistan did when it bought Tameer Kenya market has been the benchmark for payments- in Pakistan, a regulated microfinance bank). To led inclusion progress; growth is now picking up add points of presence, MNOs may choose to in Uganda and Tanzania as well. But adoption of supplement their own airtime networks with third- payments and stored value alone does not add up to party agents. In some Mobile Leapfrog countries, a robust set of financial inclusion services. “virtual� chains have been created by aggregators One promising MNO development might be of small merchants. These aggregators, which may fast-growing, MNO-marketed and distributed be entrepreneurs (e.g., INOVA in Burkina Faso), life insurance, promoted either as a stand-alone payment systems vendors (Visa, MasterCard), or product or as a free reward for customer loyalty firms that distribute goods across merchants (e.g., (as MNO Tigo does in Ghana, in partnership with fast-moving consumer goods distributors such as MicroEnsure.) Tigo’s parent, Telenor—a major Coca-Cola), can partner with MNOs to expand MNO with a focus in Asia—has now invested in the MNOs’ access points. In these cases, MNOs MicroEnsure and plans to roll out MNO-based bring additional transactions and revenues to the microinsurance in other countries). MNOs who network. partner with insurance companies offer insurers MNOs focus on transfers with limited interest in 2. access to a stable risk pool as well as the ability a full-service product line; however, long-term to collect small premiums frequently at a low financial inclusion progress requires full-service cost, which enables a new low-income-oriented products. MNOs usually launch their financial product that has significant cost synergies with the services with payment services: remittances, payments business. person-to-person payments, person-to-business payments—and noninterest-bearing, stored-value Beyond insurance, a more challenging product “float� accounts. Broader financial services require line question is whether MNOs will promote real abilities (such as assessing an individual’s likelihood savings accounts or simply promote stored value as of repayment, or pooling risk across multiple a substitute for cash. Strategically, a float account 10 or stored-value e-money wallet4 is more attractive and another three friends who use a second for MNOs (higher MNO profitability plus MNO MNO, there are nine points of usage opportunity customer relationship control) compared to a bank- (rather than six) if the two MNOs became based savings account that the MNO facilitates. interoperable. If another three friends who Savings products for low-balance customers are use a third MNO became interoperable, points not highly profitable, and they create a tighter of usage among them increase to 27. If a high bond between bank and customer. However, proportion of an individual’s network transacts research indicates that if the poor have longer- on an interconnected system, this exponential term surplus funds, they prefer the security of a effect could be sufficiently powerful to convince bank account, and they value the “out of sight, out customers to change behavior and adopt mobile of mind� personal discipline created by a harder- payments. to-withdraw savings program. Poor families who The general MNO bias toward proprietary payment have savings are better positioned to weather systems is driven in part by easier engineering and unexpected downturns in income or costly health in part by the intense rivalry among competing care bills. MNOs will better serve true financial mobile carriers. It may also stem from efforts to inclusion if they add this customer option to their replicate the success of M-PESA and its owner product line. Safaricom in Kenya. M-PESA is a Safaricom-owned Finally, MNO-led financial services teams will need system that became a proprietary standard (rather to decide whether to provide credit in cases where than a shared standard) in Kenya. For many Kenyan weak client information makes risk management users, M-PESA’s market share has been so high difficult. Microfinance addressed this problem in that for all practical purposes the limitations of the past through group risk pooling, but more noninteroperability have been insignificant. Many broadly, financial service providers are now branchless banking articles have held up Kenya as aggressively searching for ways to create individual a model for the future. credit risk profiles or their proxies. MNOs are in the But can Safaricom’s creation of a de facto standard process of creating individual credit score proxies and de facto M-PESA interoperability through for low-income customers, enabling MNO-led dominant market share be replicated in other teams to extend credit one client at a time. This countries? Figure 6 implies no. During the key highlights the value to an MNO of picking a high- years of M-PESA’s take off (2007–2009), Safaricom quality banking or microfinance partner who has had an 80 percent to 20 percent market share experience in credit product development for low- lead over its only major Kenyan competitor, and income clients. Safaricom was the only player on the market with 3. MNO interoperability is critical for high adoption. 5 a mobile payments service. Safaricom’s unusual One obstacle has slowed the take-off of mobile level of dominance created its own “network payments in many countries. MNOs have built effect.� In 2011, MNO country-leader market closed-loop, proprietary payment systems that shares were 20–40 points lower than Safaricom’s work only within their own networks. As a result, were in 2007 (even in Kenya). There were three to person-to-person transfers often require that both four major players per market rather than two. In sender and recipient use the same MNO. addition, multiple MNOs within the same country This hurdle to usage prevents customers from offer mobile payments. Even MNOs who hold the making mobile payments part of their normal second or third position in the market (in terms of routine. Without that hurdle, customers could share) can afford to offer payment solutions, due adopt the service more quickly. The swing to the availability of technology platforms and is dramatic because it is exponential. If an solutions that are fully operated and maintained individual has three friends who use one MNO by technology providers. 4 Nonbank, stored-value accounts are not considered savings accounts. Regulation often defines caps in balances, and products are not marketed for savings. 5 “Interoperability� as used here is defined as the ability of a user to send money to or receive money from a customer of a different mobile financial service provider. 11 Figure 6. MNO concentration 2011 Market shares in selected countries (2011), and Kenya (2007) 100% 8% 6% 6% 19% 13% 6% 14% 18% 19% 10% 26% 26% 21% 80% 17% 16% 26% 13% 17% 33% Market Share (SIMs) 19% 60% 16% 21% 27% 35% 31% 81% 25% 40% 65% 60% 55% 52% 49% 45% 20% 39% 34% 30% 0% Kenya Kenya Senegal Paraguay Uganda Ghana Tanzania Bolivia Cote Pakistan 2007 2011 d’Ivoire Leader #2 player #3 player Rest of market Source: Wireless Intelligence (2011). The adoption rate that M-PESA triggered in MNO network. Both Visa and MasterCard Kenya through near-interoperable dominance will can point to the impact that their respective be difficult to replicate in other Mobile Leapfrog switches have made in U.S. bank payment cards countries without truly interoperable platforms. years ago: after years of achieving less than 10 Figure 7 compares the adoption rate of mobile percent market penetration using proprietary payments in several similar Mobile Leapfrog approaches, the interoperability of third-party environments two years after the launch of each cards enabled market penetration to rise to over service. Adoption by end-users in a country with 70 percent in ensuing decades (see Figure 8). de facto interoperability (Kenya 2007–2009) was A government, social investor, and/or software •  six to seven times greater than adoption in more provider might create an interoperable platform, competitively fragmented African markets (Uganda, then stimulate end-user demand for MNO- Tanzania) and 17 times greater than in Pakistan. independent e-money wallets (e.g., Rêv in Rather than focusing on the low-odds vision of Mexico). The decreasing costs of smart phones creating a dominant proprietary standard, leading and the emergence of ultra-low-cost mobile MNOs in Mobile Leapfrog environments would financial apps suggest the long-term potential be well-advised to consider the impact that of this approach. collaborative interoperability could make in market A mobile payments provider might become a •  take-off. There are multiple ways that interoperable standard in a particular country if it were made standards could come about: available as an open system to MNOs in that environment. MNOs might create a joint standard by adopting •  How does competition respond to the evolution a common switch or protocol. from proprietary to interoperable standards? In the A third-party payments switch or agent network •  first scenario, where all players agree to a common might convince MNOs to unite around an standard through negotiation, there is an overnight interoperable technology layer that it provides. “big bang� effect. In the other scenarios, MNOs Visa recently launched a mobile platform in second in size in terms of market share and below Rwanda that can enable connections across adopt interoperability, either individually or as a mobile accounts from different banks on any group, to gain market share against the market 12 Figure 7. Mobile payment adoption velocity Mobile Leapfrog market archetype 40% 34.3% = Four years after launch 35% = Two years after launch 30% Kenya’s 2-year adoption multiple 25% 18% Percent adoption of 20% mobile payments 15% 6X 7X 17X 10% 16% 5% 3% 2% 1% 0% Kenya Uganda Tanzania Pakistan Sources: The Bill & Melinda Gates Foundation (2010); World Bank (2010); assumes 70% of stated Safaricom customers are active users. Figure 8. Impact of credit card interoperability (U.S. example) Mobile Leapfrog market archetype 80 70 National introduction of 60 BankAmericard (Visa) in 1966 and Master Charge (Mastercard) in 1969 50 Percent of U.S. families 40 with bank payment 30 cards 20 10 0 1960 1970 1980 1990 2000 2010 Sources: Federal Reserve Survey of Consumer Finances leader. If this group’s initiative proves successful, Financial inclusion agenda for Mobile Leapfrog the leading MNO—who has the most to lose environments through interoperability—feels compelled to Key opportunities to drive financial inclusion in follow over time. Mobile Leapfrog markets include the following: 13 • Regulators and policy makers: − Ensure that regulation explicitly and clearly Box 3. Technology and Microfinance allows MNOs to operate and take leadership One clear trend in microfinance is the use of roles in partnerships, including nonbank issuance technology to lower costs and support innovation. The cost-and-access benefits of branchless banking of e-money.6 have as much potential to help MFIs lend to even • Banks, MNOs, and retailers: poorer communities and reach rural clients as they − MNOs should seek to build partnerships that do to help commercial banks reach the lower- middle class. Technology is also enabling the include financial institutions (commercial banks delivery of individual (vs. group) loan products and or MFIs) and merchant aggregator networks to is helping to make microsavings, insurance, and maximize coverage and broaden the spectrum of remittance products financially sustainable. Finally, potential financial products that may be offered to by placing electronic tablets and mobile money technology in the hands of their loan officers, MFIs payments customers. They should also rethink the can usher in multiproduct client solutions and trade-off among interoperable systems (with higher create client-centric relationship management, adoption) and proprietary systems (with stronger while using cash-in agents to reduce the challenge customer retention incentives) across MNO-bank of officer cash transport. platforms. In general, they should aim to maximize What might hamper smaller MFIs’ opportunities customer adoption and create network effects. in mobile financial services is their ability to drive − Banks and MFIs should seek to partner with scale deployment of technology-enabled business models. By contrast, commercial banks often have MNOs and implement growth strategies the advantage of having a larger customer base leveraging potential to expand product lines (to and typically offer a broader set of services, which include insurance, savings, and credit). in combination can drive the economies of scale • Funders and social investors: that make alternative channels economically viable. MFIs that have achieved larger scale (i.e., millions − Encourage experimentation to expand MNO- of customers) can benefit in a similar manner as linked products beyond payments. commercial banks. − Social venture capitalists could consider investments in disruptive players that help (2,000–10,000 stores in size) are several orders of introduce interoperability. magnitude bigger than in Mobile Leapfrog markets. The level of cell phone penetration is roughly the Convergence Battle Agenda same, with 60–85 percent penetration. If the primary dynamic in Mobile Leapfrog markets Because a certain level of inclusion progress is one of MNOs filling a coverage vacuum, the has already been made in Convergence Battle primary dynamic in Convergence Battle markets countries, the inclusion agenda should take into is one of banks, retailers, and MNOs each using account secondary levels of inclusion, not just branchless banking models to improve convenience primary levels. For instance, a household that has levels for already banked, partially banked, and received a loan may qualify as “included,� but it newly banked customers—and in doing so trying to would be “more included� in the financial system achieve an edge in “owning� what is really a shared if it also had a savings account and insurance. customer relationship. As research in a higher-income market (Mexico) indicates in Figure 10, on a product-by-product In these markets, banks, retailers, and MNOs all basis there remain many unserved and underserved have solid though not pervasive penetration levels, customers not yet benefiting from a full range of which makes sense given higher income per capita services—even among those who are not poor. (see Figure 9). Bank accounts number around 60 percent of the adult population, three to four Given that there are three strong types of contenders times the bank account intensity found in Mobile for financial services leadership, and each has a robust Leapfrog environments. The largest retail chains customer relationship in its core business, it is no 6 For more details refer to Tarazi and Breloff (2010). 14 Figure 9. Convergence Battle market archetype Sample Country Profiles Mexico Brazil South Africa % Penetration Access % Penetration Access % Penetration Access (1) (1) (1) Points (2) Points (2) Points (2) MNOs 65 - 71 - 83 - 55,200 360,000 38,600 Commercial Banks (3) 56 78 63 (+450K POS) (+ >1Mn POS) (+150K POS) Retailers 10,000 3,500 2,000-5,000 Size of largest chain - - - [all formal retail outlets](4) [ >100,000 ] [ >320,000 ] [ >12,000 ] Notes: (1) Penetration of MNOs based on estimates of unique users as % of total population. Penetration of commercial banks based % of adults with access to at least one financial service from commercial banks. (2) Number of access points for banks is estimated as # of branches plus # of ATMs outside branches (assuming there is on average one ATM/branch) plus total number of banking agents. (3) An estimation of total point-of-sale (POS) devices is provided since these are access points for payments through debit/credit cards. Some fraction of these points are usually used for “cash-back.� (4) Organized retail refers to formal outlets of retail stores typically with more the one branch. Estimate is lower bound amount of stores in food, pharmacies, textile, hardware, construction material, convenience stores, gas stations. Source: World Bank (2011); Wireless Intelligence (2011); CNBV (2011); México ENIGH (2010); México SHCP Financial Inclusion Survey (2009); FinScope South Africa (2011); FEBRABAN (2011); ABRAS, ABRAFARMA and ANAMACO (2011) Figure 10. Mexican banking customer pro�les SAVINGS CONSUMER CREDIT Segments of population in income brackets “C-� to “E� Segments of population in income brackets “C-� to “E� Served Underserved Not served Source: CGAP analysis based on the study “Segmenting the Base of the Pyramid in Mexico� (2011) wonder that a convergence battle for the customer among the emerging lower-middle income class. is unfolding. The key market development dynamics However, regulation often creates friction that hinders in this environment are as follows: the full deployment of agent-based banking. 1. Banks invest in agent banking to grow revenue in a Many Convergence Battle economies are growing capital-efficient manner, to bring greater convenience at a reasonably strong pace. These economies to current customers, and to add new customers from often have a consolidated banking industry, 15 with four to six well-branded national banks. process that is particularly valuable in countries This dynamic creates a race for growth among with high income disparities between urban and the banks. Rather than simply growing through rural locations. A bank’s current customer service aggressive new branch construction, banks are locations are represented by the inner circle. It growing through the addition of bank agent is here where branch-and-spoke structures are networks and self-service technologies such as replacing the traditional branch-only system. ATMs and online banking. The middle circle represents lower-middle class areas, where those rising economically might The banks’ approach is rooted in a current join the formal financial system. It is here that a customer service problem as well as in smart mostly spokes system is used, with fewer branches asset management. Today’s bank branches are deployed to reflect the need for a lower-cost often congested, which degrades the customer’s approach. Because the same branchless banking experience. Two types of activities occur within technology is used to serve both current and the branch: transactions—simple services—that new customers, a bank that invests in branchless could be carried out by less skilled staff or by convenience for current customers today also self-service technology, and solutions—more lowers its cost of reaching new, lower-income complex financial products and services that unbanked individuals tomorrow. require more personal time and skilled support. Branch congestion is worsening as urban areas Banks that invest in branch-and-spoke models experience demographic, economic, and density often encounter regulatory requirements that (e.g., residential high rise) growth. Banks have seem effective for the inner circle but might be responded to this problem with a simple principle: costly to meet as banks expand toward the middle transactions should take place outside of the and outmost circles. A single agent operating branch, and more engaged customized customer model may not be effective or efficient in all service should take place inside. To enable localities; banks may need flexible approaches for branchless transactions, banks have been building recruiting and phasing out agents and adopting out networks of agents and self-service ATMs and different schemes for managing liquidity. In kiosks in current customer neighborhoods, as other cases, while regulations permit the use of well as deploying Internet banking websites. The agents, requirements around the account-opening result is a “branch and spoke� network designed process (e.g., the need to handle paper-based to provide customers with more convenience and a better experience. These networks on average also provide lower-cost access points for banks as Box 4. Legal Challenges to Agent Banking shown in Figure 1. In Brazil, a Convergence Battle country where banks have achieved significant growth through A parallel build-out of branchless service points agent networks, banks are facing two types of legal in new areas enables banks to acquire customers challenges. One challenge is from unions that claim with improving incomes as they rise into the lower- that captive bank agents are really employees by labor law and, therefore, should be paid the middle class. Banks can target geographic hot same wages as bank workers (they are usually paid spots where these economic climbers are located less.) A second challenge is that various groups and use agents and ATMs to provide affordable advocate for legislated caps on bank agent fees. but close-to-client service levels. A branch may If agent labor costs were adjusted up and prices were capped, it would cast doubt on the viability not even be part of this infrastructure—it may of agent economics. Lack of clarity on these two be just a management node for agents. This issues puts at risk the significant investment made approach brings the bank more agility as well as by banks in adopting financially inclusive models. Legislative clarity would help further financial more customers. inclusion. While this is an issue in Brazil today, the dynamics of agent banking create the potential Figure 11 presents this growth strategy. The for this issue to arise in any Convergence Battle diagram applies income/density principles market. within a country (rather than across countries)—a 16 Figure 11. Single country bank coverage build-out Convergence Battle market archetype Unclear model ? Mostly spoke ? Branch and Extension spoke Extension of bank- of bank- led led model? model? Mobile Mobile leapfrog? leapfrog? Highest income/density ? Moderate income/density ? Lowest income/density = branch income/density = spoke (ATM/agent) = coverage areas in = managing spoke a country market documentation) and limitations on fees that can be This kind of in-country segmentation may already be appropriately passed on to the customer, hinder happening. In Mexico, Telcel’s launch of a person- the business case for incorporating new customers. to-person funds transfer service in partnership with Banamex may support a longer-term strategy to 2. MNO-led approaches may be needed in rural areas. move into rural remittances (though clearly the What is the bank’s game plan for the outer initial focus is urban). In Brazil, two of the country’s circle? Economically speaking, it is not clear that largest banks—Banco do Brasil and Bradesco bank-agent-and-ATM economics work below a (through card acquirer Cielo)—recently bought an certain level of income or density. There may equity interest in the mobile payments subsidiary be significant poverty-stricken or rural areas (Oi Paggo) of Oi, an MNO. These examples speak that it just can’t reach. Can a further iteration to the different approaches that may be called for of the bank’s model work? Or does the bank- in the outer circle, lowest income/density areas led model need to flip to an MNO-led model, where traditional banks cannot reach. with nonbank issued e-money and banks playing These types of initiatives will require greater the junior team member role? This would create support from regulators in Convergence Battle a “pocket� of MNO-led financial inclusion, markets. Ideally, MNOs would be allowed to a Mobile Leapfrog story, in the midst of an compete with banks nationwide (e.g., through the otherwise bank-led market. issuance of e-money7), and free market choices 7 In this Focus Note, “e-money� refers to “electronically recorded value issued against the receipt of equivalent value� as described in Tarazi and Breloff (2010). Nonbank issuers of e-money may offer services to transfer value between customers, make payments to merchants or utility companies, or redeem the value in cash. 17 would improve coverage of gaps in rural areas. accounts for low-balance savers, as might be An imperfect transitional compromise may be found in Pervasive Social Banking markets such one that creates “exception zones� to bank-led as India. As a result, the time cost of travel to national regulation that explicitly allow MNOs to a bank and the financial cost of paying small- lead financial service initiatives in tough-to-reach balance fees can make the use of savings accounts rural areas. prohibitive to small savers, even though they are technically available. E-wallets, whether prepaid Solving this issue is even more important because card or especially mobile-phone based, require traditional microfinance models may not be able less travel by the small saver and do not charge to reach many of the 20–30 percent impoverished fees on savings. and unbanked in higher-income Convergence Battle markets. The reason is that the ratio of 4. Strong retailers distribute bank services, MFI staff costs to client loan sizes in Convergence positioning themselves to negotiate a maximum Battle markets is much higher than in markets split of value and selectively trying to own the that produce more sustainable microfinance customer. organizations. The higher the GINI inequality The spokes in Figure 12 are often retail stores. coefficient, the more intense this problem Paradoxically, in Convergence Battle environments, tends to be. This helps explain why for-profit retailers represent both the strongest partners to microfinance has been accompanied by extremely banks and the strongest competitors for value that high real interest rates in Mexico, even at scale banks traditionally capture. (Compartamos), and why more socially oriented microfinance has struggled to achieve market From a cost perspective, retailers have an penetration and economic viability in Brazil and advantage compared to banks because they have South Africa. access to an already-paid-for store footprint. From a customer experience perspective, a well-designed 3. Beyond rural inclusion, increased e-money retail format can compete favorably with both bank flexibility could improve urban payments services branches and MNOs. On the one hand, the retailer levels and provide affordable substitutes for low- can provide a wider range of services and customer balance savings accounts. support than the MNO. On the other, retailers can E-money is most closely associated with making provide these services in the same store that the financial services available for the first time to rural customer must visit anyway for food and basic populations. However, e-money also improves the goods, providing greater convenience. quality of certain financial services in urban areas. Several types of retailers are relevant to branchless It is particularly helpful in two ways. banking model development: First e-money enables less costly and more • Mass merchandise chains. These massive convenient money transfer/payment services. stores can provide financial services similar While low-income urban populations in to bank branches. The large store format Convergence Battleground markets can already provides space for a full financial services access transfer/payment services through banks counter. Although mass merchandiser chains and remittance companies, these services require have a relatively small number of stores, they travel on the part of the payer (and the recipient boast high traffic per store, and the ability to in person-to-person scenarios). Minimum charges provide customers with “everything you need are typically higher than with e-money, and under one roof� convenience. In most cases, recipients have to convert their payments to cash, mass merchandisers have chosen a major unlike using a mobile wallet. bank as a service delivery partner. Even then, Second, e-money provides a substitute to a low- large merchandisers push the boundaries of balance savings account through the e-wallet what financial services they can offer through function. In many Convergence Battleground store loyalty cards. In a few cases (Wal-Mart markets there are no legislated no-frills/no-fees in Mexico; Falabella and Ripley in Chile), the 18 Figure 12. Pervasive Social Banking market archetype Country Profiles India Indonesia Bangladesh Access Access Access % Penetration (1) % Penetration (1) % Penetration (1) Points (2) Points (2) Points (2) MNOs 76 1.0-1.5M 66 - 62 - Commercial Banks 40(3) 286,000 22 19,100 12(4) 8,365(4) (+635K POS) (+182K POS) MFIs 15(5) 168,254(6) 19 44,100 28 18,022 Retailers Size of largest chain 1,000 4,800 70 - - - [all formal retail outlets] [ 3,000 ] [ 13,650 ] [ 600 ] (7) Notes: (1) Penetration of MNOs based on estimates of unique users as % of total population. Penetration of commercial banks based % of adults with access to at least one financial service from commercial banks. (2) Number of bank access points estimated as # of branches plus # of ATMs outside branches (assuming on average one ATM/branch) plus total number of banking agents; for India, 50% of ATMs is taken to be “offsite� according to RBI. (3) India commercial bank penetration is based on institutions under “scheduled commercial banks,� public and private, including rural regional banks. (4) Bangladesh commercial bank penetration based on “scheduled commercial banks� and “nonbank financial institutions.� (5) India microfinance penetration based on MFI customers and SHG members. (6) India microfinance access points based on number of MFI branches and SHG village organizations. (7) Retail chain access points is a lower-bound estimate based on formal retail chain store outlets (including among others: food stores, pharmacies, textile, hardware, construction material, convenience stores, gas stations). Source: World Bank (2011); Wireless Intelligence (2011); GSMA (2011); CGAP Country Notes (2012); Reserve Bank of India; NABARD; Bank Indonesia; Bangladesh Bank; MicroSave (2011); USAID (2011); MIX (2011). retailer ended bank distribution and acquired would find difficult or inconvenient to serve its own banking license. directly. From a financial inclusion perspective, • Convenience store chains. The store “footprint� chains focused on poor clients can provide of these chains is the best for financial valuable financing of items that improve quality inclusion, because they often locate close to of life, such as a refrigerator or oven. These lower-income neighborhoods. Convenience chains should consider enriching the financial stores are best suited for certain basic financial role they play by promoting item-specific transactions: cash-in/cash-out, stored-value savings programs (i.e., lay-away) that encourage card or phone top-up, remittances, and clients to save in advance of their purchases payments. Some convenience retailers have rather than incurring debt as the only financing organized themselves as independent agents option. While credit may be more profitable, and, in a reversal of roles, have signed up savings options may improve sales and market multiple banks as partners. Oxxo in Mexico share of stores that offer it. is a strong multibank agent example. Oxxo 5. Convergence battles will extend to government also illustrates how a strong distribution social payments. partner can also become a direct provider of financial services. While Oxxo proudly partners Governments in Convergence Battle countries with several major banks for certain financial often provide significant levels of social welfare services, Oxxo is also preparing to launch payments to their poorest citizens (called its own stored-value card to provide a more government-to-person [G2P] payments). The convenient client solution. inefficiency and nontransparency associated with • Home furniture and appliance retail chains. physical distribution of government benefits mean These retailers serve as an intermediary that that governments can achieve high returns on their aggregates credit opportunities that banks investment by implementing digital social payments. 19 To date, government-owned banks have been payments through mobile accounts (in combination the vendor of choice for digital social payment with other channels) to conduct payments. distribution. They establish bank accounts for In Mexico and Brazil, banks are carrying out each recipient and certify a network of cash-out market research to find out what kind of product agents. Studies of early-stage cost savings per innovations might convince social payment recipient in Brazil and South Africa have been recipients to become true banking customers. very promising, indicating that the government The intensity of bank and MNO competition for may save up to 30–40 percent of its distribution G2P contracts may help crack the code of financial costs through use of mainstream savings accounts inclusion for some of society’s poorest members. and digital infrastructure (Bold, Porteous, and Rotman 2012). (In countries without pre-existing Key opportunities to drive financial inclusion in infrastructure, governments must make an initial Convergence Battle markets include the following: build-out investment.) • Regulators and policy makers: The most disappointing aspect of G2P payments − Develop secondary inclusion metrics (adoption has been that payment recipients have not become of a broad range of services vs. single-product financial services clients. Social payment recipients adoption) to capture a more refined picture of typically empty their accounts immediately or financial inclusion progress. within days of receipt, leaving balances at zero, − Ensure that the cost of complying with agent and the account unused until the next payment banking regulations does not surpass the arrives. Because of inactive recipient accounts, benefits of implementing them to enable the distributing bank’s role has been closer to that banks to use growth/outreach strategies (e.g., of a digital post office rather than a branchless simplified account opening, simple bank agent financial service provider. recruitment, less restrictions on fees). This creates an opportunity for MNOs. If people − Enable MNOs to lead financial service initiatives could receive their payment more conveniently in the market (e.g., by allowing nonbank issuance on a cell phone account or stored-value phone, of e-money) to both better serve and improve would they store their surplus instead of quality of service to urban populations. immediately converting it to cash? Would they − If enabling MNOs to compete nationally is not begin making digital payments? Would odds possible in the near term, financial regulators increase of successfully marketing additional and policy makers may consider enabling or financial products to payment recipients? even incentivizing MNOs to lead the provision Early evidence from mobile G2P pilots in Colombia8 of financial services in “exception zones� seems to indicate that answers to these questions where income/density economics limit the are yes. Mobile payment recipients retained a effectiveness of agent-based banking models or fraction of their social payment as stored value, microfinance. Government agencies in charge implying that it was convenient to receive and of cash transfer programs should aggressively use their social payment through the phone. explore mobile G2P distribution pilots; aim After an initial learning curve to understand the to deliver low-cost payments; maximize client full capabilities and features of the product, they convenience; and convert clients to financial made use of other functionality (balance inquiries, services users. cash-outs at ATMs, airtime purchases). Moreover, • Banks, MNOs, and retailers: they indicated the desire to use their mobile − Banks should create a branchless banking phone as a safe storage for longer-term funds (i.e., investment portfolio that is balanced between saving). The pilots were short, and formal results short-term returns (decongesting branches have yet to be published, but the government is in the inner circle) and medium-term growth now implementing broader strategies that involve (acquiring lower-middle-class customers in the 8 Results of the study have not been made public yet 20 middle circle); consider developing an “outer all of Bangladesh, the island of Java in Indonesia, circle� game plan that might involve playing a and significant swathes of India). While not all of secondary role in partnerships with MNOs. India or Indonesia are as dense, these high-density − MNOs should consider developing a mobile zones have played a significant role in shaping financial services strategy taking the lead national government financial inclusion policy and for outer circle subregions of the country. regulation. In parallel, they should consider bank partnerships for mobile banking in wealthier/ There are structural reasons why high population denser areas. density, low income per capita, and successful − Convenience store retail chains/aggregators: social banking are linked. Despite low client Develop independent multibank agent income, it is viable to run a bank branch if there models (via convenience chains or aggregated is a high volume of clients nearby (branches in convenience merchants) with particular focus on this environment typically serve 15,000 to 20,000 stores near low-income communities. customers/branch), and the product mix is skewed − Home furniture and appliance retail chains: toward loans9 (as opposed to savings). Low income Consider adding layaway options to complement per capita also has a positive effect on the viability current credit-based financing options. of social banking because staff wages are relatively • Funders and social investors: low. As a result, what is distinctive about this − Improve understanding of low-income market archetype (see Figure 12) is less the cell households that are underserved or unbanked, phone, commercial banking, or retail infrastructure, by supporting public-good, demand-side and more the additional relevance of the social research to help providers develop products that banking infrastructure. are better tailored to customers’ needs. − Conduct research to raise awareness and Governments in these environments have advocate solutions to the rural financial coverage played a pivotal role in the way social banking gap (e.g., income/density measurement within has developed, and continue to influence the country, evaluation of bank-led and MNO-led financial sector through political clout, allocation options). of resources, and/or direct bank ownership. Social − Promote conversion of large convenience store banking involves more than government funding chains into multibank agent networks. of a state-owned bank (after all, many developing countries nationalized one or more of their banks in the 1950s or 1960s, and many still have specialized Pervasive Social Banking Agenda development banks today). What distinguishes social banking is the extent to which government Countries where average income per capita is very policy has driven meaningful financial inclusion low, but population density is significantly high, among the poor while achieving sustainable have distinctively developed large, successful social economics, thus enabling the government to create banking systems. These countries have an annual significant social impact with modest taxpayer income per capita of less than US$4,000, similar to subsidies and in many cases making profits. This Mobile Leapfrog environments. But while Mobile success has made social banking the cornerstone Leapfrog environments have population densities of financial inclusion policy in these environments. of 50–150 persons per square kilometer, Pervasive Social Banking environments have densities 10 Can the pervasive social banking model be times as high (more than 1,000 persons per square improved? Are there geographic areas or product kilometer). The three countries in this archetype all lines within these countries where existing have at least 100 million persons living contiguously social banking models have not been able to fill at 1,000 persons/square kilometer density (nearly the financial inclusion gaps? Despite the high 9 This refers to institutions that, even if funded primarily by client/member deposits, promote microcredit at the center of their product offering. 21 penetration of social banking in these environments, agents for banks. More recently, these restrictions some financial services still remain undelivered or were lifted, allowing banks to more freely choose unsustainably structured. Services such as low- the kind of third parties that can serve as agents. balance savings accounts, transfers, and remote Yet most bank agents still reflect past efforts to area financial services are good examples. More meet quotas, and insufficient attention is given than half of the adult population remains unserved. by providers to design quality service delivery (Chen and Thoumoung 2012). New initiatives by The common thread emerging from the three commercial banks to tap into new agent networks countries analyzed for this market archetype— could yield additional financial inclusion. India, Indonesia, and Bangladesh—is that their • Regulations are highly restrictive on MNOs (or impressive historical achievement of financial nonbanks) to provide electronic payments11 or inclusion success through social banking led to issue e-money. As much as these services have the regulations that may slow the future emergence of potential to reduce the cost of conducting business next-generation financial services business models. in the everyday lives of the poor, they also bring While this pattern is broadly accurate, each of large, foreign, and profit-driven MNOs into the social these three countries is unique. What follows is banking system. By closely restricting how MNOs each market’s individual story, illustrating both play, the Indian market misses out on new models that commonalities and differences. might be able to cover less penetrated areas of the country, and on the integration of banking services with payments models via mobile. Experiments with India MNOs and banks in partnership are being tested, but India is the most complex market in the Pervasive have been slow to develop so far. Social Banking archetype due to its massive • Giving MNOs an innovator’s role is all the more scale, multiple cultures, global investor interest, important given that government historical and pioneering government. The country is now restrictions on international retailers have kept the entering a new era of financial inclusion initiatives retail sector fragmented, with a corresponding (see Box 5). Branchless banking opportunities are reduction in retailer-based financial services emerging at the same time that policy makers are innovation. setting limits on privately led microfinance. The result is a reform wave that simultaneously aims to more strictly monitor MFIs, to cap profit-seeking Bangladesh (by way of a margin cap on nonbank financial While Bangladesh has four major state-owned companies [NBFC] MFIs10), and to mandate the banks, financial inclusion has been driven by roll-out of agent-based banking while keeping nonprofit NGOs who have been successful in commercial banks responsible for constructing and building a profitable microfinance model to leading these deployments. scale. Most of these organizations received early government and donor agency support from the The challenge that India’s pioneering regulations 1970s through the 1990s although growth over now face is to balance a top-down social banking the past two decades has been driven without approach with bottom-up business model significant additional subsidies. Technically a for- innovation by nonbanking actors. The following profit bank, Grameen Bank serves more than 8 are some examples: million poor on social business principles. Together with large NGOs, such as BRAC, ASA, and Buro • Regulations initially allowed only bank agents Bangladesh, microfinance reaches nearly every with a “proven� social mission (NGOs, retired corner of Bangladesh, operating on a sustainable government/bank employees, servicemen) to act as basis. These institutions make profits but are all 10 MFIs that operate under the legal form of the NBFC Act, which allows them to lend but not to take deposits from the public. 11 Closed loop would be possible, but not broader open-loop systems with a stronger value proposition to customers. 22 Box 5. History of Pervasive Social Banking in India India began a series of efforts to formalize finance, from a government crackdown in the state of Andhra especially in rural areas following independence. This Pradesh, MFIs remain active in the rest of India. included nationalization of the banking system and Overall, India’s microcredit industry has established the creation of a new class of regional rural bank. The a sustained presence and remains a significant Reserve Bank of India also implemented regulations contributor to financial inclusion. designed to replace the countryside money lender with commercial bank branches. For every new commercial On the savings side, regulations made commercial branch opened in an already-banked geography, four and state-owned banks responsible for providing new branch openings were required in territories no-frills savings accounts to low-balance holders. designated as “unbanked.� India’s commercial banks The provision of these accounts is more a necessary added 30,000 branches in unbanked areas during this obligation of doing business in India than a self- period, and cut money lender share by half. However, sustaining economic activity. Many accounts are progress came at an unsustainable cost to the banks: dormant. It is yet to be seen whether the viability of loan repayments rate were only 42 percent, and costs this approach will improve with branchless models. per client helped were in the thousands of dollars (Burgess and Pande 2003). Today, the Government of India has embarked on a shift to use technology models to improve the Rather than continuing to insist that commercial effectiveness of social banking. India’s first nationwide banks play a “last mile� lending role through new identity system is under development; this will serve branch openings, in 1991 regulators instead asked as a key link in the delivery of social payments and can commercial banks to support emerging microcredit help monitor the use of microcredit. A rapid expansion models (Mahajan and Navin 2012). This began with of agent banking is planned: commercial banks self-help groups (SHGs) in the 1990s; and by the have been asked to develop targets to reach nearly 2000s, regulators had also added NGOs and NBFC 75,000 unbanked villages. The National Payments MFIs. This led to a marked increase in the availability Corporation of India, a bank-owned entity working of small loans in rural areas across India, carried out by in close consultation with government, is promoting private organizations but under the guidance of state a national switch that is interoperable among all bank directed policy. agents and banks. This supported the rapid growth of India’s microcredit India’s technology initiatives are notably aimed at industry (which today serves about 140 million Indian improving the effectiveness of social banking without clients—about 15 percent of the adult population— expanding the types of actors that could provide new including SHGs, MFIs, and regional rural banks).a levels of financial inclusion (MNOs, retailers). While private MFIs have recently faced a set-back a Based on information from MIX (2010) and the National Bank for Agriculture and Rural Development of India (2010). grounded in a social mission to serve Bangladesh partnership with MNOs has led to fast growth, now and, therefore, retain their social banking character. reaching more than 3 million customers through more than 35,000 agents. Another early mobile banking As in India, this success has raised expectations that platform is Dutch Bangla Mobile, offered by a socially future financial inclusion happens primarily through a responsible commercial bank. The platform now has social banking model. Unlike India, the government almost 900,000 customers and 15,000 agents. In has not prohibited MFIs from providing savings each case, although the banks provide the account accounts. This has empowered MFIs to provide a infrastructure, the MNOs are able to own the mobile banking platform to MNOs who are more customer relationship. willing to work in partnership because of a lower perceived competitive threat. For example, BRAC Bangladesh is a Pervasive Social Banking market to the NGO has offered the “bKash� platform to MNOs watch for alternative models. This said, the present as a mobile financial service linked to its affiliated approach does not allow MNOs to issue individual deposit-taking BRAC Bank.12 BRAC’s nonthreatening e-money wallets, which may leave some business 12 bKash’s press release for its launch in June 2011 described the platform as “a full- scale mobile phone-based payments switch� in addition to being “an extension of BBL (BRAC Bank Limited).� 23 model options for financial inclusion untested. However, the main benefit of this e-money system In addition, nearly 50 commercial banks are not is narrowly focused—provincial citizens use e-money involved in experimentation. to make payments that otherwise require travel to make in person (IFC 2010). This is in itself a real Indonesia financial inclusion achievement. But limitations in cash- Indonesia’s history of financial inclusion has out preclude real banking services from flourishing revolved primarily around government-owned through this model. Bank Rakyat Indonesia (BRI), the world’s largest microfinance and small and medium enterprise There are several signs that more profound (SME) lending bank boasting 28 million active branchless change is on the way. Branchless clients (13 percent of the adult population), with banking regulations that allow cash-in/cash-out and roughly 6,800 branches and 7,000 ATMs. Its 28 MNO-bank partnerships look to be finally coming percent return on investment rivals the financial to fruition. Bank Mandiri has also just finalized a performance of any Indonesian bank (BRI 2010). The deal with the PT Pos Indonesia post to open bank bank has evolved over time from lending to savings branches at post office outlets. 13 A BRI-Telkom and from microfinance/SME services to corporate partnership for a hybrid mobile bank account is financial services. At the same time, BRI has been meant to be announced soon. partially privatized via minority investment. As with other leading social banking markets, Indonesia BRI has a strong presence in Java, which at 1,024 will be challenged to balance top-down rules that persons per square kilometer is the nation’s densest create order and protect incumbent social banking province, housing nearly 60 percent of the population. providers with new model innovation. The benefits of BRI has significant presence as well across the other supporting innovation include the following: 9,000 named islands through a “tiered� scheme of branches (going all the way down to small kiosks) that • Commercial banks that are trying to target low- allow it to reach much lower density areas efficiently. income segments can provide healthy competition Without formally adopting agent banking, BRI has in if they are truly allowed to adopt agent banking. effect achieved a similar outcome in reach and cost • Allowing full cash-in/cash-out agents for e-money efficiency. In addition to BRI, 1,700 provincially focused would broaden the financial services available rural banks (BPRs) and several hundred thousand to Indonesia’s remote poor (beyond the current cooperatives help serve the islands outside of Java. benefit of bill payment services provided today). Despite all of this activity, research indicates that the inclusion gap across the country remains significant, A summary of opportunities to drive financial particularly in regions outside the island of Java. inclusion in Pervasive Social Banking markets include the following: Due to the government’s historical success at social banking, plus the power of incumbency, branchless • Regulators and policy makers: banking has been slow to take off as a new tool for − Regulators should strike the right balance financial inclusion. Banks cannot use agents to conduct between depending on the known actors regular banking transactions. Rules developed five responsible for the past successes of social years ago allow banks and nonbanks to issue e-money, banking, and the potential for healthy but e-money agents (which cannot act as bank agents) competition and innovation that can be can provide only cash-in services (in practice, primarily introduced through new nonbanking actors. In used for bill paying). Several commercial banks and the most cases this implies allowing social banks largest MNOs—including Telkomsel, and Indosat— to develop e-money platforms that MNOs have issued e-money wallets, and today along with can aggressively use to bring mobile financial competitors claim more than 11 million subscribers. services to unserved customers. 13 Through this deal, the postal company PT Pos Indonesia also invests in Mandiri Bank through one of its subsidiaries, the Bank Sinar Harapan Bali (The Jakarta Post 2013). 24 − Ensure that social banking targets for commercial Battleground levels of GDP per capita, although players don’t lead to a false sense of progress with significantly higher population density than (e.g., inactive bank accounts, inactive agents, Latin American markets, supporting even higher other examples of fulfilling quotas without levels of financial inclusion. achieving the underlying financial inclusion • Over 300 million live in population densities similar goals). Objective evaluation of impact is critical. to Mobile Leapfrog markets, albeit with higher • Banks, MNOs, and retailers: income per capita than those environments. − Social banks that can take deposits should • The country’s 20th century history of state- aggressively provide attractive platforms for controlled economy makes Pervasive Social MNOs to incentivize them to invest heavily in Banking issues relevant to China, even though promoting financial services. its income/density profile does not fit that − Commercial banks should be on the lookout for environment. new agent banking opportunities, both through nonretail networks and through retail chains that While this paper does not directly address China’s may emerge in the future. issues, we believe the combined issues raised in − MNOs who have access to a mobile money all three market archetypes provide the building platform in partnership with a social bank should blocks for a Chinese financial inclusion agenda. A invest aggressively in promoting mobile financial China note that applies these market archetypes services. MNOs who do not yet have access to would be valuable. such a platform, should clearly articulate the value proposition to society for greater MNO Implementing a Financial leadership in certain financial inclusion products Inclusion Strategy Based and in areas of the country that are underserved on Market Archetypes by a bank-led model. • Funders and social investors: The ideas expressed in this Focus Note have − Invest in research that demonstrates what is implications for the planning, investment, and working best in branchless banking across organizational approaches of those involved in Pervasive Social Banking markets: promoting financial inclusion. • Document patterns of success across the three Pervasive Social Banking markets to identify 1. National teams of business and government conditions through which successful mobile leaders who are trying to develop effective banking is emerging in financially underserved financial inclusion strategies may gain new insights areas. by choosing comparators from countries within • Document patterns of the most successful similar market archetypes, including those outside agent banking networks established by of their own geographic region. commercial banks especially in fragmented 2. Investor and donor strategies for financial retail environments. inclusion are likely to be more focused and to generate better results if they are thought about A Brief Word on China by market archetypes. Specialized funding pools or staff organization by common income/density There is one giant data point in the center of the market archetypes is one possible approach. This three-by-three income/density matrix that warrants same principle holds true for industry knowledge its own discussion—China. The country is located in management players and broader knowledge the middle, touching the corner of all three market dissemination strategies. archetypes, making it aptly placed. Within China’s 3. In large and highly varied countries, financial 1.3 billion population: inclusion leaders should debate whether a single archetype approach best fits their • Over 600 million people—a population larger country, or whether a multi-archetype approach than all of Latin America—live in Convergence is needed. This could apply to Convergence 25 Battle countries that have a much less wealthy ———. 2012. “Branchless Banking Country Notes� and dense rural population or to island-based for Brazil, Mexico, Ghana, WAEMU, South Africa, geographies whose income and density varies India, and Pakistan. Washington, D.C.: CGAP. dramatically. 4. It may be appropriate to measure financial inclusion CGAP and World Bank. 2010. Financial Access progress taking into account the market archetype 2010: The State of Financial Inclusion through the to which a country belongs rather than using Crisis. Washington, D.C.: CGAP and World Bank. nominal metrics. For instance, financial inclusion of 40 percent in a Mobile Leapfrog environment might Chen, Gregory, and Aimthy Thoumoung. 2012. be viewed as a more impressive accomplishment “National Survey of Branchless Banking Agents in than financial inclusion of 70 percent in a India: Towards High Quality BC Customer Service Convergence Battle country. Points.� Washington, D.C.: CGAP and the College of Agricultural Banking. References CNBV (Comisión Nacional Bancaria y de Valores) México. 2011. Financial Sector Reports and Statistics. ANTAD (Asociación de Tiendas de Autoservicio y Departamentales, A.C.) 2012. “Desempeño en Das, V. S. 2008. “Social Responsibilities of the ventas en el mes de enero 2012.� México: ANTAD. Indian Banking Sector—How to Promote Access to Finance for the Unbanked Masses and Underserved Australia Indonesia Partnership and IFC Advisory Sectors.� Address at the Indian Banking Services Indonesia. 2010. “Mobile Banking in Conference, Centre for Analytical Finance, Indian Indonesia: Assessing the Market Potential for Mobile School of Business, Hyderabad. India. Technology to Extend Banking to the Unbanked and Underbanked.� Indonesia: Australia Indonesia Demirguc-Kunt, Asli, and Leora Klapper. 2012. Partnership and IFC Advisory Services Indonesia. “Measuring Financial Inclusion: The Global Findex Database.� World Bank Policy Research Paper Bangladesh Bank. 2011. Financial Sector Reports 6025. Washington, D.C.: World Bank. and Statistics. Bangladesh: Bangladesh Bank. Faz, Xavier, and Paul Breloff. 2011. “Segmenting ———. 2012. “Mobile Financial Services in the Bottom of the Pyramid in Mexico.� Washington, Bangladesh: An Overview of Market Development.� D.C.: CGAP. Bangladesh: Bangladesh Bank, July. FEBRABAN (Federacao Brasileira de Bancos). 2011. Bank of Ghana. 2011. Annual Report. Ghana: Bank “Bancarizacao e Inclusao Financeira no Brasil.� of Ghana. Presentation at conference of the FELABAN, July. Bold, Chris, David Porteous, and Sarah Rotman. FinMark Trust. 2011. “FinScope South Africa 2011 2012. “G2P Payments and Financial Inclusion: Report.� South Africa: FinMark Trust. Evidence from Four Countries.� Focus Note 77. GSM Association Mobile Money for the Unbanked Washington, D.C.: CGAP. Blog. 2011. Burgess, Robin, and Rohini Pande. 2003. “Do Rural GSM Association Wireless Intelligence. 2007–2012. Banks Matter? Evidence from the Indian Social Banking wirelessintelligence.com Experiment.� Discussion Paper No. DEDPS/40. London: London School of Economics and Political Science. Huérfano, Edgar. 2012. “De la población 60% no accede al banco.� 2012. México: El Economista. CGAP. 2011. “Building the Business Case for Banks in Branchless Banking—Brazil Study.� Washington, Hulme, David, and Karen Moore. 2005. “Why Has D.C.: CGAP. Microfinance Been a Policy Success? Bangladesh 26 and Beyond.� U.K.: Institute for Development Philip, Darrel. 2012. “Death of the Mom & Policy and Management. School of Environment Pop Store—Is It Inevitable?� Blog post. www. and Development. University of Manchester. Chillibreeze.com. INEGI (Instituto Nacional de Estadística y PT Bank. 2010. “Annual Report.� Rakyat, Indonesia. Geografía). 2010. “Encuesta Nacional de Ingreso PT Bank. y Gasto de los Hogares 2010.� Aguascalientes, México: Instituto INEGI. RBI (Reserve Bank of India). 2009. “Branch Banking Statistics.� Mumbai: RBI, March ———. 2011. “Indicadores Oportunos de Ocupación y Empleo.� Boletín de prensa núm. 011/1219 de ———. 2011 “Financial Sector Reports and enero de 2012. Aguascalientes. México: INEGI. Statistics.� Mumbai: RBI. Intermedia Knowledge Center. 2010. “Case Rochadi, S. Budi. 2010. “The Role of Central Bank Study: Extending Financial Services Access to in Financial Education Process in Indonesia.� More Ghanaians.� Washington, D.C.: Intermedia Presentation at the OECD-Banque du Liban Knowledge Center. International Conference on Financial Education, Beirut, Lebanon. Kale, Vishal. 2012. “Category Analysis of Organised Retail in India.� Blog post. reflectionsvvk.blogspot. SHCP (Secretaría de Hacienda y Crédito Público). mx, India. 2009. “México Financial Inclusion Survey.� México: SHCP. Kumar, Kabir, and Toru Mino. 2011. “Five Business Case Insights on Mobile Money.� Washington, Sipahutar, Tassia. 2013. “Bank Mandiri Builds D.C.: CGAP. Partnership with PT Pos, Taspen.� The Jakarta Post, 1 February. Mahajan, Vijay, and T. Navin. 2012. “Microfinance in India: Growth, Crisis and the Future.� India: Basix. State Bank of Pakistan. 2008–2009. “Pakistan 10 Year Strategy Paper for the Banking Sector Martowijoyo, Sumantoro. 2007. “Indonesia Reforms.� Pakistan: State Bank of Pakistan. Microfinance at the Crossroads: Caught Between Popular and Populist Policies.� Essays on Regulation and ———. 2011. Financial Sector Reports and Supervision, No. 23. College Park, Md.: Microfinance Statistics. Pakistan: State Bank of Pakistan. Regulation and Supervision Resource Center. Tarazi, Michael, and Paul Breloff. 2010. “Nonbank MIFA (Microfinance Initiative for Asia). 2009. E-Money Issuers: Regulatory Approaches to “Bangladesh: Microfinance and Financial Sector Protecting Customer Funds.� Focus Note 63. Diagnostic Study.� Washington, D.C.: IFC and KfW Washington, D.C.: CGAP. Bankengruppe. Thomas White International. 2012. “Retail Sector in MIX (Microfinance Information Exchange). www. Brazil: Riding the Wave of Middle Class Growth and themix.org Consumer Credit Boom.� BRIC Spotlight Report. Chicago: Thomas White International. NABARD (National Bank for Agriculture and Rural Development). 2010. “The Status of Microfinance USDA (U.S. Department of Agriculture). 2005– in India 2009–10.� India: NABARD. 2011. USDA Foreign Agricultural Service, Global Agricultural Information Network Reports on the Oxford Business Group, The. 2011. “The Report: Retail Food Sector for Pakistan, Ghana, Senegal, Ghana 2011.� Ghana: Oxford Business Group. Brazil, Mexico, India, Bangladesh and Indonesia. 27 World Bank. 2010. “Improving Access to Financial Growth through Acquisition.� Jakarta, Indonesia: Services in Indonesia.� 2010. Jakarta, Indonesia: World Finance, February. www.worldfinance.com World Bank. Wright, Graham, Mukesh Sadana, Puneet Chopra, World Bank. World Development Indicators by the and Manoj Sharma. 2011. “Why E/M-Banking Will World Bank DataBank, at databank.worldbank.org Soon Reach Scale in India (An Optimistic View).� India: MicroSave. World Finance. 2011. “BRI Maintains Momentum: Organic Growth through New Outlets; Inorganic No. 86 April 2013 Please share this Focus Note with your colleagues or request extra copies of this paper or others in this series. CGAP welcomes your comments on this paper. All CGAP publications are available on the CGAP Web site at www.cgap.org. CGAP 1818 H Street, NW MSN P3-300 Washington, DC 20433 USA Tel: 202-473-9594 Fax: 202-522-3744 Email: cgap@worldbank.org © CGAP, 2013 The authors of this Focus Note are Xavier Faz, Sr. Financial Sector Gates Foundation), David Ferrand (the Kenya Financial Sector Specialist of CGAP, and Ted Moser, Senior Partner at Prophet consulting Deepening Trust) as well as Gregory Chen (CGAP) and Katharine and a 25-year board member of Opportunity International, a global McKee (CGAP) for their invaluable input to this paper. The Technology microfinance network. Moser also serves on the Technical Committee of and Business Model Innovation Program at CGAP is co-funded by the CGAP’s Technology and Business Model Innovation program. Bill & Melinda Gates Foundation, CGAP, The MasterCard Foundation, The authors would like to thank Rodger Voorhies (the Bill & Melinda and the UK Department for International Development (DFID). The suggested citation for this Focus Note is as follows: Faz, Xavier, and Ted Moser. 2013. “Advancing Financial Inclusion through Use of Market Archetypes.� Focus Note 86. Washington, D.C.: CGAP, April. Print: ISBN 978-1-62696-012-1 epub: ISBN 978-1-62696-014-5 pdf: ISBN 978-1-62696-013-8 mobi: ISBN 978-1-62696-015-2 UKa from the British people