64981 OccasiOnal paper Is There a Business Case for Small Savers? T his study examines quantitatively whether or Apparently for these reasons, some leading MFIs not small savers—defined here as the half of have complained loudly that increased competition all savings clients of a microfinance institution in their markets is causing them to rethink whether (MFI) with the smallest deposit account balances— they can continue to provide the level of cross- contribute to or undermine the sustainability of the subsidies that they believe is required to serve MFI. MFI sustainability has long been accorded a small savers. place of central importance in microfinance for reasons that are well known. Perhaps because of As competition continues to mount in the rapidly the necessary focus of microfinance for so many expanding field of microfinance, more and more years on lending, achieving sustainability has often MFIs are likely to ask this question, including been thought of in terms of reducing loan granting those MFIs already taking small deposits as well and recovery costs and charging sufficiently high as greenfields (MFIs that would start up operations loan rates. Arguably, less attention has been paid as regulated entities that can both take deposits to the impact of other products, such as savings and make loans from the outset) and unregulated accounts and particularly small savings accounts, MFIs that are considering formalizing and funding on MFI sustainability. themselves in the future by mobilizing deposits. For these reasons, and also because it is important The Small-Saver Problem to examine the impact of deposit products on MFI sustainability—just as we have for many years for Recent data for five MFIs in Peru are alarming in loan products—a closer look at the true costs and this regard (Portocarrero, Tarazona, and Westley, overall profitability of small savers is long overdue. 2006, Table 29).1 These data indicate that the This is termed an investigation into the “business annual operating costs associated with small- case for small savers� to differentiate it from the balance savings accounts (defined somewhat social case for small savers, in which MFIs justify differently in that study as savings accounts with providing savings services to small depositors even balances under US$100) are generally in the range at very high operating costs because small savers of 200–300 percent of the amount of savings are deemed a poor and underserved group and mobilized. This means that for a savings account therefore deserving of service. with a balance of US$50, the MFI generally incurs US$100–150 per year in operating costs.2 The methodology used to study the cost and profitability of small savers is preeminently This same study (Table 27) also shows that 75 quantitative. Specifically, we attempt to quantify percent of all savings account holders in four the major costs and revenues associated with small Peruvian MFIs with these data have deposit savers and thus estimate the profitability of this balances under US$100 and that these small savers client segment. Although we estimate small-saver provide only 3 percent of all funds mobilized by profitability in only two MFIs, these two MFIs have the MFI through savings accounts.3 In effect, then, been carefully selected and suggest a surprisingly collecting small savings deposits brings large large number of important channels through which No. 18 numbers of savers to MFI branches, costs a great small savers may be a profitable, or even highly September 2010 deal, and provides very little funding to the MFI. profitable, client segment. Glenn D. Westley and Xavier Martín 1 The five MFis consist of one crac (señor de luren) and four cMacs (arequipa, piura, chincha, and pisco). Palomas 2 after a thoroughgoing literature review, we could find only one other study with any estimates of the operating costs of small-balance savings accounts under any definition of “small,� namely, Deshpande and Glisovic-Mezieres (2007), which updated the estimate for cMac arequipa given in the portocarrero, Tarazona, and Westley (2006) study. 3 While we have chosen in this study to define small savers as the smallest half (instead of three-quarters) of all savers, the same general points would apply to our reduced set of clients, though with even higher costs per dollar mobilized than the 200–300 percent range cited in the preceding paragraph and with an even smaller percentage of all savings mobilized than the 3 percent just noted. 2 There are several reasons to think that the business systems. Also very importantly, these technologies case for small savers may be more favorable than is can help the MFI attract and retain clients and thus indicated by the operating costs in the five Peruvian increase the number of savings accounts and loans, studies: marginal cost, cross-sales, technology, the amount of savings mobilized and the amount higher loan rates for smaller and otherwise lent, and the sale of other products. And they may costlier-to-make types of loans, higher fees and/or increase the average amount clients are disposed lower interest rates on savings accounts, and the to save with the MFI since deposited funds can be evolution of small savers to future profitability. more readily retrieved when needed. None of the five Peruvian MFIs whose costs were cited earlier Marginal cost. The operating costs cited for the five used any of this technology at the time they were Peruvian MFIs were estimated by allocating all costs, analyzed. The present study provides rigorously including fixed and quasi-fixed costs, to the array derived estimates of the impact of technology on of loan, deposit, and other products offered by the the cost and profitability of small savers. MFIs.4 In reality, if these five MFIs were to eliminate their small savers, the cost savings would actually be Higher loan rates for smaller and otherwise less than the figures cited earlier since the fixed and costlier-to-make types of loans. If the loans taken much of the quasi-fixed costs would still have to be out by small savers are significantly smaller than paid. We refer to the costs that would actually be average or costlier per dollar lent for other reasons, saved by eliminating small savers as the marginal the MFI may be able to cover all its loan costs and costs (sometimes also called the avoidable costs). even make this lending and small savers profitable by charging higher interest rates and/or fees for Cross-sales. In addition to having a savings these loans. account, small savers may also take out loans, send and receive money transfers, buy insurance, pay Higher fees and/or lower interest rates on bills, and purchase other financial products from savings accounts. If the cost of serving small the MFI. If the MFI earns a profit on the sale of savers is still excessive even when all the points these products, this profit may partially, fully, or made in the preceding paragraphs are taken into more than fully compensate for the high operating account, the MFI always has the option to charge costs of small savings accounts. On the other for the service provided. This is analogous to MFIs hand, small-saver cross-sales may generate losses that want to be sustainable raising their lending for the MFI, and so one cannot say in advance rates, overall or for particularly costly subsets of whether cross-sales will turn out to help or hurt borrowers. In the case of small-balance deposits, MFI sustainability. For example, if the products this would typically involve reducing the interest bought by small savers consist primarily of small rate paid on small savings balances (possibly to loans and if these small loans are unprofitable, zero) and/or increasing monthly or per-transaction then taking into account cross-sales would actually fees on small, or possibly all, accounts. All of these hurt the business case for small savers. measures would be preferable to making small- balance savings accounts completely unavailable to Technology. Automatic teller machines (ATMs), clients at any cost by raising the minimum deposit point-of-sale (POS) devices, and mobile banking size to levels that many small savers could not reach. each have the potential to greatly reduce the Again, the analogy to the lending side is instructive: operating costs of attending to savers, including one of the great achievements of microfinance has small savers, provided the MFI is big enough to been to show that poor people are willing and able overcome the fixed costs of implementing these to pay high interest rates for good loan services, 4 By definition, as the number of clients and their usage of products grow, fixed costs don’t change at all and quasi-fixed costs increase only a little (much less than in proportion to product usage). 3 and that paying such high rates is far preferable to are a very high-cost product for MFIs to offer, with being without the service. Perhaps somewhat less annual operating costs on a marginal cost basis of well known is the fact that poor savers are willing 59–241 percent of the deposit balances of small to accept high negative interest rates on deposits savers in the study year of 2008. (−30 percent and less) for safe and convenient deposit services, such as those provided by the However, these high operating costs are more than Indian deposit collectors described in Rutherford overcome by the profits generated through cross- (2000, pp. 13–17). The fact that poor people with sales of loans and other products to small savers very low balances are willing to pay significant and by the fee income derived from their savings monthly fees to MFIs and banks (including to accounts. On balance, then, small savers are found Centenary Bank, one of the two MFIs we study) to generate large profits, of just over 400 percent of and receive little if any deposit interest in return is the small-saver deposit balances in Centenary and evidence that the use of negative deposit rates is just over 1,000 percent in ADOPEM. Expressing a viable strategy in the formal banking context as this same result in a different way, without the well, at least for good deposit services.5 small savers, these two very profitable MFIs would lose about 30 percent of their total profits. We Evolution of small savers to future profitability. conclude, therefore, that based on our small-saver Even if small-saver loans and savings accounts are profitability analysis, there is a compelling business too small today to make small savers profitable, both case for serving small savers in both Centenary and may increase in size over time by enough to make ADOPEM. This calculation of the profits associated small savers profitable in future years. As a result, with a given client segment (such as the small small savers may be worth serving today even on a savers) is sometimes referred to as the calculation of strictly business basis so that the MFI can reap the total client profitability, reflecting the fact that we rewards of serving them in future years. As we will are looking at the totality of the products consumed discuss in the final section of the paper, there are by the client segment and the overall MFI profits a number of strong and fairly general reasons for that result. thinking that the average size of the savings accounts and loans of small savers may grow rapidly over An important clarification is in order here. This time, which could transform small savers from an paper demonstrates a “business case for small unprofitable client segment into a profitable one. savers�: the high cost of delivering savings accounts is more than offset by profits from the The next two sections of this paper summarize our other products these customers use (and by any results and then describe the methods we used to savings account fees collected from them). But obtain these results. The last three sections discuss this does not automatically constitute a “business the cost, profitability, and future profitability of case for small savings accounts.� Suppose for small savers. a moment that we eliminate all small savings accounts, with their attendant costs, by raising Summary of Findings: the minimum savings account size, but are still Pathways to Profitability able to retain all these clients as customers for loans and other products, which hopefully are In case studies of two MFIs, ADOPEM in the profitable. Under such an assumption (and also Dominican Republic and Centenary Bank in Uganda, under the assumption, which is made here only for we confirm that the savings accounts of small savers expositional simplicity, that the eliminated savings 5 For example, in centenary Bank about half of the depositors had savings balances of Us$16.67 or less in 2008. With a Us$16.67 savings deposit, centenary’s flat monthly fee of Us$0.56 translates into an effective monthly interest rate of 23.4 percent (50.56/16.67), which compounds to 234 percent per year, a cost that goes far beyond the 5 percent per year received by savings depositors in interest. 4 accounts generated little or no fee income), the 1. Loans. Loans are an important source of small- provision of expensive small savings accounts saver profits in both MFIs. In ADOPEM, loans would hurt, not help, MFI profitability. The generate small-saver profits of US$1.22 million, or profitability of providing small savings accounts 91 percent of the total US$1.34 million in profits depends, then, on how many customers and cross- obtained by ADOPEM from its small savers. In sales the MFI would lose if it eliminated the small Centenary, loans yield small-saver profits of US$2.00 savings accounts (or charged higher fees for them, million, or 51 percent of the total US$3.94 million in or otherwise reduced the losses associated with profits derived by Centenary from its small savers. them). Measuring what that loss of clients and Given the importance of loans in total small-saver cross-sale profits would be is a complex task and is profits, it is useful to examine the ingredients that outside the scope of the present study. However, make small-saver loans so profitable in the two MFIs: in the case of the MFIs analyzed, particularly ADOPEM, the profits derived from the small savers’ (1) Lending in general (to all borrowers, not just other product purchases are large in comparison small savers) appears to be profitable in both with the costs of their savings accounts, which MFIs. We deduce this from the following: suggests that these MFIs are better off keeping the a. Both MFIs are quite profitable overall, with small savings accounts if their elimination would after-tax, return-on-asset (ROA) values in cause the departure of an appreciable number of 2008 of 5.53 percent for Centenary and customers, especially loan customers. The rest of 7.20 percent for ADOPEM. this paper concerns itself with the measurement b. Lending is the major revenue generator in and analysis of total client profitability and whether both MFIs, providing 70 percent of total there is a business case for small savers (not small revenue in Centenary and 86 percent in savings accounts), taking the minimum deposit ADOPEM in 2008 (with the remaining size and savings account fees as they actually were revenue generated by savings account during the study year of 2008. As noted in the fees; the sales of other products, such as preceding paragraph, there is a strong business insurance and money transfers; each MFI’s case for small savers in both MFIs. liquidity investments, such as local bank certificates of deposit (CDs) and treasury To generalize from these two MFIs to others, it bills; and foreign exchange and other is useful to examine what makes small savers operating income). profitable in the two case studies. To this end, With both MFIs quite profitable and lending we have identified five sources of small-saver such a major activity, it is very likely that profits in ADOPEM and Centenary, which we call lending is profitable in both MFIs as well. five pathways to profitability: (1) loans; (2) other (2) Small savers in both MFIs are not small products besides loans and savings accounts, such borrowers. For example, at the end of 2008, as life insurance (ADOPEM) and money transfers the average loan balance of small savers who (Centenary); (3) savings account fees; (4) technology; borrowed was 61 percent and 74 percent of and (5) higher loan interest rates for smaller and the average loan balance of all borrowers in otherwise costlier-to-make types of loans. Small- Centenary and ADOPEM, respectively.6 This saver profits in ADOPEM are derived entirely from is important because if the MFI makes a profit the first two pathways, especially the first, while on its overall lending operations, and if small small-saver profits in Centenary come from all five savers have an average loan size that is not pathways. too far below the average for all borrowers, 6 in contrast, the average savings account balance of small savers was 1.3 percent and 17 percent of the average savings account balance of all savers in centenary and aDOpeM, respectively. These data clearly show that small savers are much bigger borrowers than they are savers. This is not terribly surprising since small savers were selected for their small savings accounts. 5 then it is more likely that lending to small These fees consist primarily of a monthly charge of savers will be profitable as well.7 US$0.56 on each savings account. ADOPEM levies (3) The fact that our analysis excludes all fixed no fees on savings accounts. and much quasi-fixed operating costs in computing small-saver profits (as it should— Taken together, the sum of the profits from see the discussion on marginal cost in the pathways 1–3 is 100 percent of small-saver profits previous section) pushes up small-saver in both ADOPEM and Centenary. In other words, profits in ADOPEM and Centenary compared these three pathways are a disaggregation by to studies that include all of the fixed and product of the total profits derived from small quasi-fixed costs.8 savers. The remaining two pathways, technology (4) In ADOPEM only, there is a high rate of cross- and higher loan rates for costlier-to-make loans, selling of loans to small savers, with about are employed only by Centenary. They can be three-quarters of ADOPEM’s small savers viewed as catalysts that increase the profits derived borrowing at any given time during 2008. from loans, other cross-sold products, and savings accounts to the levels noted in points 1–3. As a result of these four factors, lending to small savers in ADOPEM generates large profits. Given 4. Technology. Centenary makes substantial that loans provide ADOPEM with 91 percent use of ATMs, which is illustrated by the fact of the total profits it obtains from small savers, that 51 percent of the deposit and withdrawal ADOPEM’s pathway to small-saver profitability transactions of small savers (and 47 percent for runs primarily through lending. clients overall) were made with ATMs in 2008, and the rest with tellers. ATMs primarily served While Centenary shares with ADOPEM the first three as a magnet that helped Centenary attract and ingredients for loan profitability, it differs on the retain clients and increase lending to small fourth: only about 15 percent of Centenary’s small savers, the amount of other products (money savers are borrowers. Nevertheless, with lending transfers) sold to small savers, and the number in Centenary generally quite profitable for clients of savings accounts held by small savers. As a overall and with small savers taking substantial size result, ATMs significantly boosted profits in all loans (when they borrowed), lending to small savers three product areas—loans, other products, and still provided Centenary with significant profits. savings accounts—by 43 percent, 30 percent, and 34 percent, respectively, increasing Centenary’s 2. Other cross-sold products. The remaining overall small-saver profits by 37 percent to the 9 percent of total small-saver profits in ADOPEM US$3.94 million figure cited earlier. (or US$0.13 million) are generated from the sales of three life insurance products. In Centenary, 5. Higher loan rates for smaller and otherwise 16 percent of total small-saver profits (or US$0.61 costlier-to-make types of loans. Because small million) are derived from the sales of four essentially savers demanded loans that were both smaller money transfer products. and costlier to make for other reasons, Centenary charged small savers 5.8 percentage points 3. Savings account fees. Centenary’s savings more for loans than it charged borrowers overall account fees generate the remaining 33 percent of (34.2 percent instead of 28.4 percent). Were total small-saver profits there (or US$1.32 million). Centenary to abandon its practice of charging 7 The importance of small-saver loans not being too far below the average in size can be explained as follows. Because the cost of making and recovering loans is roughly constant over a broad range of loan sizes (say, Us$100–3,000), profitability tends to fall off sharply as loan size decreases. This occurs because loan revenue falls in proportion to loan size while costs remain roughly constant. 8 The exclusion of these costs also gives more latitude for MFis with weaker overall loan profitability performances and more modest small-saver loan sizes than those of aDOpeM and centenary to still find that lending to small savers is profitable. 6 more for smaller and costlier loans, and thus Methods reduce small-saver loan rates by 5.8 percentage points, all profits from lending to the small savers, The basic purpose of this study is to analyze the and therefore half of Centenary’s total small- cost and profitability of small savers, quantifying saver profits from all sources (of US$3.94 million), the profits associated with the first five pathways would be lost. This illustrates that charging more described in the preceding section while also for smaller and costlier types of loans can be a taking care to measure costs on a marginal basis. powerful tool to cover costs and make small savers The resulting estimates appear to offer substantial a profitable client segment if they make substantial new information given that our literature search use of these types of loans. did not turn up any other studies that made all or even most of these measurements for any MFI. To To these five pathways to small-saver profitability estimate the cost and profitability of small savers, observed in ADOPEM and Centenary, we add a we have had to develop a number of innovative sixth possible pathway: methodologies. These methodologies are generally applicable to measuring the cost and profitability 6. Small-saver profits in future years. Employing of any client segment, not only small savers. The data from ADOPEM for three recent points in time most novel and perhaps important of these new (end of 2006, 2007, and 2008), we find that the methodologies allows one to measure the impact average size of small-saver savings accounts and of technologies such as ATMs, POS, and mobile loans grew very rapidly, a total of 105 percent and banking on the cost and profitability of small savers 83 percent, respectively, for these two products (or other client segment). Measuring the impact of over the two-year period from the end of 2006 to these new delivery technologies with any reasonable the end of 2008. (Available data did not effectively degree of accuracy is a surprisingly formidable task. allow a longer analysis period for ADOPEM or any It appears that previous calculations suffer from analysis at all for Centenary.) Although our results large measurement errors because they did not in this area are based on indicators of profitability, take into account certain key elements involved in rather than on the complete profit calculation making these estimates, in particular, the potentially employed for pathways 1–5, the indicator trends large impacts of the new delivery technologies in are so strong that it appears quite likely that attracting and retaining clients and in increasing ADOPEM’s small savers have indeed become the number of deposit accounts and loans, the much more profitable during the two-year period amounts saved and borrowed, and the number of studied. As discussed in the final section of this transactions carried out by clients. The presentation paper, there are important reasons for believing of these and other methodological points is mainly that this result of strong rising trends in average left to a subsequent paper, though Box 1 (discussed small-saver savings and loan sizes could hold quite in the next section) touches briefly on some of the generally, thus helping to make or further reinforce key points. The present paper primarily focuses on the business case for small savers in many MFIs. summarizing and interpreting the main empirical findings on the cost and profitability of small savers Given all of these possible pathways to profitability— in the two case study MFIs. and taking into account the fact that the revenue derived from loans, other cross-sold products, and savings accounts need cover only marginal Case Study Selection Criteria costs in order for these products to be considered profitable—our educated guess is that many MFIs To illustrate a number of different pathways by are already profitably serving small savers and many which small savers could be profitable, we took more could do so. great care in selecting the two case study MFIs. 7 ADOPEM and Centenary both had multiyear 7 of Table 1) and profitability (line 11). On the track records (up through the study year of 2008) other hand, they are much closer to average in of successfully delivering voluntary savings and the key area of loan delinquency and write-offs microloans to poor clients, as measured by low (lines 12 and 13), though these average levels are loan delinquency and write-off rates, good MFI generally considered quite good. In the area of profitability, rapid growth of savers and borrowers loan officer productivity (line 14), ADOPEM is well and of savings and lending, predominance of above average but Centenary is only a little above voluntary over forced savings, and reasonably low average. And while ADOPEM charges significantly average loan and savings balances. In addition, more than average for loans (line 15), Centenary’s ADOPEM was chosen because (i ) it had a high rate loan charges are quite typical. Finally, the average of cross-selling, particularly of loans, as indicated deposit and loan sizes (lines 9 and 10) suggest that by the fact that 76 percent of small savers in both institutions aim at fairly typical MFI savings ADOPEM also had a loan and (ii ) it made essentially and loan clients, though Centenary does have no use of new delivery technologies. In contrast, an important subset of commercial savers and while Centenary’s rate of cross-selling loans to borrowers that brings up its averages. small savers was only 15 percent, it provides an outstanding example of the use of new delivery Defining the Small Savers technologies to reduce costs, attract and retain clients, and boost client saving and borrowing. In deciding where to draw the line between small Another difference between the two MFIs is that and large savers, we initially thought to use the Centenary charged monthly fees on all savings same US$100 dividing line used by Portocarrero, accounts and imposed higher interest rates on Tarazona, and Westley (2006) in their five Peruvian the smaller loans taken out by small savers, while costing studies cited earlier. The idea would be ADOPEM did neither of these. to make our results more comparable to those. It quickly became apparent that such a dividing line General Characteristics of ADOPEM would be a poor choice, particularly in the case and Centenary of ADOPEM. ADOPEM reaches down to much poorer savers than do the five Peruvian MFIs, To interpret and generalize from the results with the result that 97 percent of ADOPEM savers obtained in the two case studies, it is useful to had savings account balances of under US$100 in compare some of the key characteristics of 2008. Hence, in terms of operating costs, analyzing ADOPEM and Centenary against those of MFIs in ADOPEM’s small savers would be almost the same general (Table 1). The value of each indicator for as analyzing all of their savers.10 MFIs in general (column (3) of Table 1) is computed as the median value over all 1,160 MFIs with 2008 Adjusting the small-saver cutoff value by gross data in the MIX Market worldwide data set.9 national income (GNI) per capita did not seem very helpful either, failing again in the comparison As can be seen in Table 1, ADOPEM and Centenary of the Dominican Republic and Peru. According to are above average MFIs by a number of measures, the World Bank’s World Development Indicators, but not by all. These two MFIs are well above GNI per capita, calculated in U.S. dollars for average in both overall size (lines 1, 4, 6, and 2008, is US$4,390 for the Dominican Republic 9 These data were obtained from the MiX Market Web site on 18 February 2010. The median value for a given indicator is the value corresponding to the 50th percentile, meaning that half of the MFis have values below the median and half have values above the median. For ease of expression, we sometimes use the words “typical� or “average� in referring to the median. 10 Taking a broader view, even if Us$100 were a reasonable small saver cutoff value for many other MFis in latin america besides aDOpeM, it probably would not be so reasonable for large numbers of MFis located in substantially poorer african and asian countries. 8 Table 1: Characteristics of ADOPEM and Centenary Bank, end of 2008 (3) (1) (2) MIX Market Indicator ADOPEM Centenary median value 1. Total number of active savers 88,250 491,757 3,793 a. Number of active small savers 44,125 245,879 b. Number of active large savers 44,125 245,878 2. Total number of borrowers who are not active savers 22,316 0 (“pure borrowers�) 3. Total number of savings and loan clients (5112) 110,566 491,757 4. Total number of loans outstanding 87,060 90,251 8,593 5. Total deposits excluding CDs (US$) a 2,105,689 167,409,939 6. Total deposits including CDs (US$) a 10,996,960 191,413,460 487,209 7. Gross loan portfolio (US$) 39,151,586 151,295,500 4,005,603 8. Average deposit balance per depositor excluding 24 340 CDs (5 5/1, US$) a 9. Average deposit balance per depositor including 125 389 155 CDs (5 6/1, US$) a 10. Average loan balance per borrower (US$) 440 1,633 581 11. After-tax ROA (%) 7.20 5.53 2.12 12. Portfolio at risk . 30 days (%) 3.28 4.60 3.56 13. Write-off ratio (%) 0.83 0.73 0.36 14. Borrowers per loan officer 437 282 245 15. Yield on gross portfolio (nominal) (%) 42.26 31.04 30.72 Sources: All data in this table are from MIX Market except those found in lines 1–6 and 8–9 of columns (1) and (2), which are from the ADOPEM and Centenary management information systems (lines 1–4) and the 2008 audited financial statements found in ADOPEM (2008) and Centenary Bank (2008) (lines 5–6). Line 8 is derived as the ratio of line 5 to line 1, and line 9 is derived as the ratio of line 6 to 1. See www.MIXmarket.org for the definition of indicators taken from MIX Market. MIX Market data for ADOPEM and Centenary may differ somewhat from the apparently comparable data presented above in lines 1–6 and 8–9 of columns (1) and (2) due to, among other things, the use of different exchange rates and the fact that MIX relies heavily on data obtained by means of an email exchange with a designated respondent at each MFI. Notes: Blank cells indicate data not available. Point-in-time data, such as the total number of active savers and gross loan portfolio, are given for 31 December 2008, while flow data, such as income and expenses (which are used in variables 11 and 15), are cumulations over all of calendar 2008. a Because of the large amount of deposits that consist of CDs, particularly in ADOPEM, none of which is held by small savers, we calculate total deposits both ways, excluding CDs in lines 5 and 8 and including CDs in lines 6 and 9. and US$3,990 for Peru.11 This gives a GNI per While this definition means that the cutoff line capita ratio for the Dominican Republic to Peru of measured in U.S. dollars will vary from one MFI 1.10 (54,390/3,990). Consequently, the US$100 to another, it provides in return an important small-saver cutoff line for Peru would be increased uniformity: for all MFIs, we measure the operating to US$110 for MFIs in the Dominican Republic, costs and profits of the bottom (smallest) half of meaning that an even higher percentage of their savers. If a U.S. dollar cutoff line were used ADOPEM’s savers would be considered small. instead, we could end up comparing, for example, the operating costs and profits of the bottom We finally settled on the relative definition of 10 percent of savers in one MFI to the bottom small savers noted earlier: the half of all savings 80 percent in another MFI. If, in fact, savings clients with the smallest deposit balances. account size is linked at all to client profitability, 11 These data are available to nonsubscribers at http://ddp-ext.worldbank.org/ext/DDpQQ/member.do?method=getMembers. Gni is closely related to the more familiar measure of gross domestic product (GDp). 9 then the bottom 10 percent of savers could well account. While standards differ from MFI contain many unprofitable clients and be highly to MFI, nondormant deposit accounts in unprofitable overall, whereas the profitability of ADOPEM and Centenary are those in which the bottom 80 percent of savers is likely to much there has been at least one transaction in the more closely mirror the profitability level of the last two and three years, respectively.12 overall MFI. By examining the bottom 50 percent • Because we are examining whether a certain of clients in both MFIs, we avoid this source of group of clients (small savers) is profitable or noncomparability. not to serve, the focus of this study is savers, not savings accounts. Hence, for each small Because deposit balances can fluctuate considerably saver we must consider the cost of all of over the course of a year, it seems better, if the saver’s deposit accounts and the profits possible, to define small savers according to an earned from the loans, money transfers, and average balance taken over many days rather than other products sold to the saver. In deciding according to the balance on a single day (such as whether a single saver with multiple deposit the last day of 2008). In ADOPEM, the information accounts is small or not, the balances of all system allowed us to compute the average daily of that saver’s accounts are added together. balance for each individual saver over all 366 days Similarly, the balances of joint accounts are of the year 2008. By contrast, in Centenary we had split equally among all account holders. As it to define small savers according to their balance on turns out, multiple and joint accounts are quite a single day. In light of the steadily growing number uncommon in both ADOPEM and Centenary, of savers at Centenary, we chose that day to be the and so the exact way in which they are last day of 2008 in order to analyze as many savers handled is likely to be of little consequence in as possible. The following points summarize other the present study. key aspects of the small-saver definition. • The average daily balance measure used for ADOPEM is calculated only over those days • Centenary offers three kinds of deposit in which the client had one or more deposit accounts: CDs, savings accounts, and accounts open. For example, if a client opens checking accounts. ADOPEM offers only CDs an account in July 2008 and maintains a and savings accounts. Because of the high US$100 balance for the entire second half of minimum account sizes for CDs and checking the year, the average balance is calculated as accounts, no CD holders and few checking US$100, not US$50. To illustrate the reason account holders ended up in the bottom for not counting missing days as having a zero half of all savers arrayed in increasing order balance, consider the following more extreme according to the total amount they had on example: a new depositor who puts US$1,000 deposit with the MFI. Because it simplifies the into a savings account during the last two days analysis considerably, we put all of Centenary’s of 2008. This client seems to us to merit being checking account holders in the large-saver classified as a large saver (who happened group. As a result, all small savers hold only to open an account late in the year) and is savings accounts in both MFIs. counted as such (with a US$1,000 average • Only active savers are considered. Active balance), not as a small saver with a balance savers are those with a nondormant deposit of (US$1,000) 3 (2/366) 5 US$5.46. 12 Typically, inactive savers generate little if any deposit account costs for MFis, including centenary and aDOpeM. This is because inactive savers generally have no teller transactions and hence none of the substantial teller costs and teller-related nonpersonnel costs we find for active savers (rent for the space tellers occupy, paper for the transactions tellers do, depreciation of the equipment and furniture tellers use, etc.). Hence, it would seem much more relevant to know the operating costs associated with the smaller half of the active savers, rather than the smaller half of all savers, and so we carry out our analysis on this former group. 10 Small Savers of ADOPEM and Centenary noted earlier). While small-saver profitability in the most recent year for which we had data (2008) is Applying the above rules, the cutoff point between the main measure we shall use of whether there is small and large savers turns out to be US$7.24 for a business case for small savers, the final section of ADOPEM and US$16.67 for Centenary, meaning this paper discusses how small-saver profits might that 50 percent of ADOPEM’s and Centenary’s evolve in future years and how that evolution might savers have balances below these respective cutoff affect the business case for small savers. lines, and 50 percent have balances above them.13 By any measure, the savings accounts of small Returning to Table 1, we see that the total number savers are expensive (Table 2, lines 1 and 2). Line of active savers in ADOPEM and Centenary at the 1 of the first data column in Table 2 shows the five end of 2008 is 88,250 and 491,757, respectively average cost estimates calculated by Portocarrero, (line 1). The number of active small savers is half Tarazona, and Westley (2006, Table 29), which of these totals: 44,125 and 245,879 (line 1a). In were cited at the beginning of this paper. These addition, ADOPEM has 22,316 pure borrowers (that measure the cost of small-saver savings accounts is, borrowers who are not active savers), giving it as annual operating costs per dollar mobilized and a total of 110,566 savings and loan clients (lines 2 range from 235 percent to 597 percent. These and 3). Since Centenary requires all borrowers to estimates were made using the traditional average maintain a savings or checking account (to which cost approach (also called the full cost approach), loan proceeds are automatically credited and from in which all MFI operating costs, including fixed which loan payments are automatically taken), it has and quasi-fixed costs, are allocated to small- no pure borrower clients. Finally, while Centenary saver savings accounts and the remaining client has far more active savers than ADOPEM, the total segment/product combinations. By comparison, number of loans outstanding at the end of 2008 the corresponding average annual operating costs is roughly the same in the two MFIs: 87,060 in per dollar mobilized for the savings accounts of ADOPEM and 90,251 in Centenary (line 4).14 all clients (not just the small savers) in these same five MFIs range from 10 percent to 21 percent (see Table 29 of Portocarrero, Tarazona, and Westley Introduction to the Business 2006), far more modest expense levels than the Case for Small Savers: Cost of 235–597 percent calculated for small savers. Small-Saver Savings Accounts The final three columns of Table 2 show the We begin our analysis of the business case for operating cost estimates of small-saver savings small savers in this section by discussing the annual accounts that are made in the present study operating costs of small-saver savings accounts. In for Centenary with and without ATMs and for the next section we build on this analysis, moving on ADOPEM (which did not have ATMs). These three to the larger issue of the profitability of small savers columns give the annual operating costs per dollar (a measure also called total client profitability, as mobilized (that is, the annual operating costs in 13 These U.s. dollar values were obtained from their local currency counterparts using the following exchange rates to convert from Dominican pesos (Dr$) and Ugandan shillings (Ugx) to U.s. dollars: Us$1 5 Dr$35 5 Ugx 1,800. These exchange rates are used throughout the study, regardless of whether the values to be converted to U.s. dollars pertain to the end of 2008, all of 2008, or an average of the values at the end of 2007 and of 2008. By adopting a single exchange rate, ratios such as rOa that use a mix of these concepts will be the same in local currency and U.s. dollars. Given the fairly limited variation of the exchange rates during 2008 in both countries, our results will not differ greatly from those obtained by using end of 2008 exchange rates to convert end of 2008 indicators, end of 2007 exchange rates to convert end of 2007 indicators, and daily average 2008 exchange rates to convert 2008 flow data such as profits and costs. 14 note that we have presented data on the number of savers (not deposit accounts) and the number of loans (not borrowers). This is because we are interested in small vs. large savers (as clients) and the cross-sales of other products, especially loans, to these two groups. 11 Table 2: Annual Operating Costs of Savings Accounts Held by Small Savers (in US$ and as a percent of the average savings mobilized from all small savers in 2008) “Baseline�: Centenary Centenary Small-Saver Savings Accounts: Operating Cost 5 Average With ATMs Without ATMs per Dollar Mobilized (%) Cost Studies ADOPEM (Scenario A) (Scenario B) 1. Average cost per dollar mobilized 235%, 238%, 220% 848% nc (traditional full costing) (= line 4 / line 6) 255%, 298%, 597% 2. Marginal cost per dollar mobilized (this 59% 181% 241% study’s methodology) (= line 5 / line 6) 3. Marginal cost / Average cost (= line 2 / line 1 27% 21% nc OR line 5 / line 4) Operating Costs and Savings Account Balances of Small Savers in 2008 (US$) 4. Small-saver savings accounts: total operating 290,374 8,123,480 nc costs in 2008—using the average cost (traditional full costing) approach (US$) 5. Small-saver savings accounts: total operating 77,846 1,737,396a 1,653,567a costs in 2008—using the marginal cost approach (US$) 6. Small-saver savings accounts: 2008 131,910 958,061 684,995 balance—average of the small-saver daily totals in 2008 for ADOPEM; average of the small-saver end 2008 and end 2007 totals for Centenary (US$)b Sources: All calculations in the last three columns are based on data provided by ADOPEM and Centenary. The “Baseline� percentages in the first data column are taken from Portocarrero, Tarazona, and Westley (2006), Table 29. Notes: The exchange rates used for all conversions to U.S. dollars from Dominican pesos (DR$) and Ugandan shillings (Ugx) are US$1 5 DR$35 5 Ugx 1,800. “nc� indicates not calculated. a Some readers may be puzzled by the fact that the operating costs of small-saver savings accounts are higher with ATMs than without. These higher costs stem from the fact, discussed in Box 1, that ATMs attract additional clients and also induce clients to make additional transactions. b Data limitations at Centenary prevented us from using the average of the 366 daily totals of small-saver savings in 2008, as done with ADOPEM. 2008 divided by the average amount of savings savers, or given client segment/product combination, mobilized in 2008) using both the average and such as small-saver savings accounts? As will be marginal cost approaches (except for Centenary discussed further in the forthcoming companion without ATMs, where only marginal cost is methodology paper, marginal cost is the right computed). These values are also high, ranging measure to use whenever assessing viability, whether from 59 percent for marginal cost in ADOPEM to it be the viability of a client segment, product, branch 848 percent for average cost in Centenary. office, or a combination of these, such as small-saver savings accounts. This is an important point since The marginal cost of small-saver savings accounts many (though not all) past costing studies, including includes only the costs that the MFI would save (or many studies employing activity-based costing (also avoid) if the savings accounts of small savers were known as ABC), have used average cost in assessing eliminated, and thus excludes all of the MFI’s fixed viability despite the fact that this can easily lead one to and much of its quasi-fixed costs. As noted, marginal incorrect conclusions. For example, a study using full cost is the appropriate measure to use when trying to costing instead of marginal costing might find that a answer the question: how much would an MFI save certain product offered by an MFI generates negative if it eliminated a given client segment, such as small profits and thus is not viable. If the MFI then drops 12 that product to increase its profitability, it might find marginal cost basis of 59 percent to 241 percent of instead that profits have declined because the MFI has the deposit balances of the small savers. not saved any of the fixed costs or much of the quasi- fixed costs that had been allocated to the product. In this section, we find that these high operating costs are more than overcome by the profits In both ADOPEM and Centenary we find that generated through cross-sales of loans and other marginal costs are substantially below average products to small savers and by the fee income costs. This reflects the large amount of fixed and derived from small-saver savings accounts. On quasi-fixed overhead expenses generally involved balance, then, small savers are found to generate in running financial institutions, often including large profits, of just over 400 percent of their much or all of the following: deposit balances in Centenary and just over 1,000 percent in ADOPEM.17 Expressing the same • The board of directors and management result in a different way, without the small savers, • The staff of such central service departments these two very profitable MFIs would lose about as accounting, administration, audit, finance, 30 percent of their total profits. We conclude, information technology, legal, marketing, therefore, that based on our analysis of the personnel, and risk management profitability (also called total client profitability) of • The nonpersonnel costs associated with all of small savers in 2008, there is a compelling business these personnel (e.g., rent for the space they case for serving this client segment in both MFIs. occupy; depreciation of the desks, furnishings, equipment, and vehicles that they use; and the As discussed in the summary of findings section, electricity needed to provide ambient lighting total small-saver profits can be disaggregated and cooling of the space they occupy and to into five sources of profit, which we have also run their equipment) called five pathways to profitability: (1) loans; (2) other products besides loans and savings However, even with ADOPEM’s marginal cost accounts, such as life insurance (ADOPEM) and per dollar mobilized value of 59 percent, the money transfers (Centenary); (3) savings account operating costs of small-saver savings accounts fees; (4) technology; and (5) higher loan interest are still high, and even higher for Centenary (181 rates for smaller and otherwise costlier types of percent and 241 percent with and without ATMs, loans. Building on that earlier section, this section respectively).15 While the business case for small explains in greater detail how we calculate the savers looks somewhat better from a marginal cost profits derived by the MFIs from each of the five view, small-saver savings accounts still appear to pathways. be a high-cost proposition.16 Our computation of small-saver profits in Centenary Total Client Profitability and the (both with and without ATMs) and ADOPEM is Five Pathways shown in Table 3. Lines 1, 2, and 3 give what may be thought of as the usual three components of In the preceding section, we confirmed the high costs lending costs: operating costs, bad loan costs, of small-saver savings accounts for both ADOPEM and loan funding costs, respectively. These are all and Centenary, with annual operating costs on a calculated on a marginal basis, that is, by computing 15 The methods we used to calculate the cost and profitability of small savers in the case of centenary without aTMs will be explained in the forthcoming companion paper on methodology. Those interested in a brief glimpse of this should see Box 1. 16 The ratio of marginal to average cost for small-saver savings accounts in aDOpeM and centenary is 27 percent and 21 percent, respectively, as shown in line 3 of Table 2. lines 4–6 of Table 2 present the data used to construct the cost per dollar mobilized ratios given in lines 1 and 2. 17 To get these values, we divide line 9 of Table 3 (small-saver profits in 2008) by line 6 of Table 2 (small-saver savings account balances in 2008). 13 Box 1: Simulating Centenary without ATMs (Scenario B) Of Centenary’s 491,757 savings and loan clients at the The meaning of the first parameter is that one-third end of 2008, 256,221 (or 52 percent) were ATM users. of all ATM users would completely leave Centenary— We define an ATM user as a client who held an ATM taking all of their deposit, loan, and other business with card and made at least one ATM withdrawal during them—if Centenary did not have ATMs. In the ATM 2008. While we would like to have defined an ATM user survey, the primary reason respondents gave for user as a client who held an ATM card and made at this is the great convenience of ATMs, which has at least least one ATM withdrawal or deposit during 2008, we two important dimensions: (i) the availability of ATMs are forced to use our definition because we can trace 24 hours a day, seven days a week and (ii) the fact that only ATM withdrawals to individual clients (and thus transactions could generally be made more quickly with know whether the client is a small or large saver), not ATMs than tellers since ATM waiting lines were usually ATM deposits. Fortunately, two pieces of evidence shorter and ATM locations were more numerous (ATMs indicate there is likely to be very little difference were in all 32 branches plus 10 additional locations). between these two definitions. The two-thirds of ATM users who would stay with The first piece of evidence comes from a nationally Centenary even without ATMs would reduce their representative sample survey we conducted of 252 savings balances by an average of 25 percent Centenary clients (151 small savers and 101 large because of the reduced convenience of accessing savers), who were interviewed after making an ATM their deposited funds. This means that if there were deposit or withdrawal transaction. This survey found no ATMs, the overall savings balances of ATM users that every one of the 252 interviewees who made an would fall by half, as indicated by parameter 4. (This ATM deposit in the last month also made an ATM value of 1/2 is calculated as follows: Because 2/3 of all withdrawal. The second piece of evidence is the fact ATM users would remain Centenary clients if ATMs that there were 5,101,810 ATM withdrawals but only were eliminated, each with a deposit balance that is, 524,460 ATM deposits made by all Centenary clients on average, 3/4 of what it was with ATMs, then overall in 2008, so that withdrawals are by far the dominant deposit balances would be half of what they were with ATM transaction. ATMs since (2/3) 3 (3/4) 5 1/2.) To calculate the amount by which ATMs impacted As can be seen from the remaining parameter values, small-saver operating costs, average savings account the number and value of loans to ATM users would balances, and profitability, we estimated the following both fall by 1/3 if ATMs were eliminated (parameters six key parameters for all ATM users using the 2 and 3). Finally, parameters 5 and 6 tell us that 1/2 of above survey as well as an analytical (case-control/ ATM deposit transactions and 2/3 of ATM withdrawal time series) method. We find that about the same transactions would not be replaced by teller deposit parameter values hold for both the small- and large- and withdrawal transactions if ATMs were eliminated, saver subgroups of all ATM users: again because of the greater inconvenience of using tellers instead of ATMs. 1. Share of ATM users who would leave Centenary if there were no ATMs: 1/3 A forthcoming companion paper on methodology will 2. Share of loans that would not have been made if explain how these parameter estimates were made and there were no ATMs: 1/3 then used to calculate what the cost and profitability of 3. Reduction in loan portfolio if there were no small savers would have been if Centenary did not have ATMs: 1/3 ATMs. These without-ATM cost and profitability values 4. Reduction in deposit balances if there were no (shown in the last column of Tables 2 and 3, respectively) ATMs: 1/2 permit us to compute the impact that ATMs have had 5. Share of ATM deposit transactions that would be on small-saver cost and profitability. The companion eliminated (not done with a teller) if no ATMs: 1/2 paper will also discuss how these impact estimates 6. Share of ATM withdrawal transactions that would be might be made for MFIs employing branchless banking eliminated (not done with a teller) if no ATMs: 2/3 technologies, such as mobile phones and POS devices. 14 how much total operating costs, total bad loan By means of their savings accounts, small savers costs, and total loan funding costs would change if provide a small share (approximately 2 percent) small savers were eliminated. of their own loan funding needs in all three cases (Centenary with and without ATMs and ADOPEM). Line 1 measures the operating costs associated Hence, the loan funding costs shown in line 3 have with both the savings accounts and loans of two components: the relatively small amount of small savers, in contrast to the operating costs interest paid on the small savers’ own savings given earlier in Table 2 (line 5), which measure accounts plus the much larger amount of interest the operating costs of only the savings accounts paid on the remaining loan funding sources, which of small savers. To compute the profits from must be tapped because small savers borrow far lending to small savers we must count both more than they save.18 Summing up lines 1–3, we loan and savings account operating costs, the get the total costs of small-saver savings accounts latter because the savings accounts mobilized and loans, shown in line 4. from small savers provide a part of the funding for the loans extended to small savers. As can Total small-saver profits (Table 3, line 9) are obtained be seen in the subcategories of line 1, the by summing the following three components: personnel costs that would be saved if the profits from lending, fees on savings accounts, small-saver loans and savings accounts were and profits from other products. The profits from eliminated come overwhelmingly from branch lending to small savers (line 6) are obtained by personnel, mainly loan officers who make subtracting the line 4 costs from the loan interest loans to small savers and tellers who attend and fee revenue given on line 5. As shown on line their savings account and loan transactions. 7, Centenary, but not ADOPEM, collects significant Most of the remaining personnel costs come fee income from small-saver savings accounts, from (i ) branch- and headquarters-based back particularly from a monthly savings account fee office staff working on the loans and savings (called a ledger fee there). In addition, Centenary accounts of small savers and (ii ) loan officer and levies a fee of approximately US$0.11 for each teller support and supervisory personnel, also withdrawal clients make using an ATM. Centenary in the branches and headquarters. The major also charges clients for new and replacement ATM nonpersonnel costs associated with small savers cards, but loses money on these sales because the (which are mainly driven by the loans and savings cost of producing and delivering the cards is greater accounts of small savers) fall into six categories than the fees charged (Table 3, footnotes b and c). in both ADOPEM and Centenary: transportation Finally, line 8 reports the profits earned on other (especially for loan officers), rent, depreciation, products besides loans and savings accounts: three paper and printing, electricity (purchased and life insurance products in the case of ADOPEM self-generated), and telecommunications. and four essentially money transfer products in the case of Centenary (electronic crediting of salaries The bad loan costs of small savers (line 2) are to client deposit accounts, electronic payment of small compared to their operating and loan school fees for the children of clients, and two all- funding costs (lines 1 and 3). This is not very purpose money transfer services, RGTS/EFT and surprising given the low loan delinquency and Western Union). Summing lines 6–8 gives line 9, write-off rates noted earlier (Table 1, lines 12 the total profits derived from small savers from all and 13). sources, before income taxes. 18 as noted earlier, the loan funding costs in line 3 measure the marginal cost of funds (as opposed to the more commonly encountered concept of the average cost of funds). Just as we measure the operating costs of serving small savers as the amount by which total operating costs would be reduced if small savers were eliminated, we measure small saver loan funding costs as the amount by which total loan funding costs would be reduced if small savers were eliminated. The forthcoming companion methodology paper delves further into how to calculate the marginal cost of funds. 15 Table 3: Profitability of Small Savers in ADOPEM and Centenary, 2008 (US$) (2) (3) Centenary Centenary (1) with ATMs without ATMs Small Savers Only ADOPEM (Scenario A) (Scenario B) 1. Operating costs of savings accounts and loans (5a1b) 1,045,879 4,755,262 4,174,098 a. Personnel costs (5a11a2) 598,601 3,363,061 3,088,691 a1. Branches 543,292 3,191,986 2,996,281 a2. Headquarters 55,309 171,075 92,411 b. Nonpersonnel costs 447,278 1,392,201 1,085,407 2. Bad loan costs (provision expense net of recoveries of 173,393 1,029,203 840,842 written-off loans) 3. Loan funding costs (5a1b) 1,110,986 4,279,120 3,505,724 a. Interest on small-saver savings accounts 6,398 194,193 129,881 b. Interest on remaining loan funding 1,104,588 4,084,927 3,375,843 4. Total costs of lending, including costs of own savings 2,330,258 10,063,584 8,520,665 accounts, a loan funding source (511213) 5. Revenue from lending: loan interest and commission 3,547,056 12,064,666 9,915,274 revenue a 6. Profits from lending (5524) 1,216,798 2,001,081 1,394,609 7. Fees on savings accounts (5a1b1c) b 0 1,322,599 1,015,914 a. Monthly fee on savings accounts (ledger fees) 0 1,249,177 1,015,914 b. ATM withdrawal fee 0 205,431 0 c. Sale of ATM cards (net of card production and distribution 0 −132,008 0 costs)c 8. Profits from other products d 127,439 613,315 459,405 9. Total profits from small savers, before income 1,344,237 3,936,996 2,869,928 tax (561718) Sources: All table entries are based on data provided by ADOPEM and Centenary and refer only to small savers. a For Centenary, in addition to loan interest, this line includes an upfront fee of 2 percent of loan amount and the proceeds from selling loan application forms for US$2.78 per loan. ADOPEM also charges upfront legal and processing fees totaling approximately 2–4 percent of loan amount. b Centenary Bank’s fees on savings accounts are as follows: a. Monthly savings account fee 5 US$0.56 per month b. ATM withdrawal fee 5 US$0.11 per withdrawal c. Sale of ATM cards 5 US$1.11 new, US$5.56 replacement c The entry for Centenary Bank is computed as US$92,899 in sales revenue minus US$224,908 in expenses, where expenses consist primarily of the costs of the blank ATM cards, card embossing and printing, and distribution to clients through Centenary tellers. d The breakdowns of the line 8 totals are as follows. Profits from ADOPEM’s three life insurance products: Transition funds 5 US$314 Burial insurance 5 US$1,460 Credit insurance 5 US$125,665 Profits from Centenary’s four other products are given in each line below, first for Scenario A (with ATMs) and then for Scenario B (without ATMs): Salary entries (electronic crediting of salary payments) US$335,602 US$238,507 School fee payments (to client-designated schools) US$183,248 US$154,235 RTGS/EFT money transfers US$86,001 US$60,031 Western Union money transfers US$8,464 US$6,632 16 Table 4: Small Savers: Sources of Profitability Centenary with Centenary without (4) ADOPEM ATMs ATMs Percent Increase (1a) (2a) (3a) in Centenary (1) Percent (2) Percent (3) Percent Small-Saver Profits Profits of Total Profits of Total Profits of Total Due to ATMs (US$) Profits (US$) Profits (US$) Profits 5 [(2) 2 (3)] / (3) 6. Profits from 1,216,798 91 2,001,081 51 1,394,609 49 43 lending 7. Fees on savings 0 0 1,322,599 33 1,015,914 35 30 accounts 8. Profits from 127,439 9 613,315 16 459,405 16 34 other products 9. Total profits 1,344,237 100 3,936,996 100 2,869,928 100 37 from small savers Note: The row numbers and labels here are the same as those used in Table 3. The profits data in columns (1), (2), and (3) here are also the same as in columns (1), (2), and (3) of Table 3. We now consider the breakdown of total profits Pathway 4: ATMs. ATMs make a substantial into the five pathways noted earlier. contribution to Centenary’s profits, as can be seen in column (4) of Table 4. This column gives Pathways 1–3: Loans, Other Products, and the percentage increase in small-saver profits Savings Accounts (respectively). As shown in that Centenary obtains from ATMs and shows Table 4, the percentage breakdown of total small- that ATMs significantly increase all three profit saver profits into the three product pathways components, lifting overall profits by 37 percent. is quite different for ADOPEM and Centenary. ATMs do this primarily by helping Centenary to Columns (1), (2), and (3) of Table 4 reproduce the attract and retain clients, which in turn increases profits data for ADOPEM, Centenary with ATMs, lending volume, the number of savings accounts, and Centenary without ATMs, respectively, that and the sale of other products, and hence the are shown in lines 6–9 of Table 3. Columns (1a), profits earned in all three of these areas. (Box 1 (2a), and (3a) of Table 4 show the percentage of discusses the effects of ATMs in more detail.) total profits that are due to lending (line 6), savings account fees (line 7), and other products (line 8). As Pathway 5: Higher loan interest rates. Finally, shown in these three columns, ADOPEM’s profits because of the higher cost types of loans and are derived mostly from lending (91 percent), with smaller loans that small savers disproportionately 9 percent of total profits coming from the three demanded, Centenary charged small savers 5.8 insurance products and no profits derived from percentage points more for loans than it charged savings account fees (since ADOPEM had no such borrowers overall.19 Had Centenary abandoned its fees). Centenary’s profits (both with and without practice of charging more for types of loans that ATMs) are more diversified: approximately 1/2 were costlier to make (namely, microenterprise, from lending, 1/3 from fees on savings accounts, housing, and agricultural loans) and for smaller and the remaining 1/6 from the four money loans, and thus reduced small-saver loan rates by transfer products. 5.8 percentage points, all profits from lending to 19 in 2008, centenary’s portfolio yield, a measure of the effective interest rate charged to borrowers, was 28.4 percent for all borrowers and 34.2 percent for borrowers who were small savers, giving this 5.8 percentage point differential. portfolio yield is calculated as the following ratio (for example in 2008 in Us$): (loan interest plus commission revenue in 2008 in Us$) / (average loan portfolio in 2008 in Us$). 17 small savers, and therefore half of Centenary’s a. The far larger size of commercial loans total small-saver profits from all sources, would compared to the other four loan types. The have been lost.20 This illustrates that charging average outstanding balance of commercial more for smaller and costlier types of loans can be loans for all borrowers is about US$15,000, a powerful tool for making small savers a profitable versus US$1,000–2,000 for the other four client segment if they make substantial use of loan types. these kinds of loans. b. The much more limited analysis and collection efforts needed for consumer A Further Exploration of the Loan Rate loans, especially since all consumer loan Differential clients are salaried employees whose paychecks are deposited directly into To gain further insight into the 5.8 percentage their Centenary accounts, from which loan point differential between what small savers and repayments are automatically deducted by all clients pay for loans, it is useful to examine the Centenary’s computer system. Reflective effects of loan type and size on Centenary’s costs of this situation, consumer loan officers and loan rates. We start with the effect of loan disbursed 834 loans per loan officer in type and then turn to loan size. 2008, versus 331 for loan officers overall.21 Centenary’s five main types of loans fall into two Turning to the effect of loan size, small savers groups: paid loan rates that were 2–3 percentage points higher than rates for borrowers overall for each • Those loans that are more costly per dollar one of the five types of loans. (This can be seen lent and have higher interest rates, namely, by the data just given in the two bullets above; microenterprise, housing, and agricultural loans. for example, all borrowers paid 34 percent for These loans have interest rates of 34 percent, microenterprise loans and small savers paid 33 percent, and 31 percent, respectively, in 37 percent.) In practice, this occurred because 2008 on a portfolio yield basis for all borrowers the generally smaller size loans taken by small (and 37 percent, 35 percent, and 34 percent, savers gave these borrowers less bargaining respectively, for small savers). power with which to negotiate interest rate • Those loans that are less costly per dollar reductions. From Centenary’s point of view, this lent and have lower interest rates, namely, is a reasonable outcome since smaller loans of commercial and consumer loans. Both of these a given type are more expensive per dollar lent loan types have interest rates of 24 percent in than larger loans. 2008 on a portfolio yield basis for all borrowers (and 26–27 percent for small savers). The We conclude that Centenary’s loan pricing lower cost per dollar lent of commercial and structure broadly reflects differences in cost and consumer loans reflects: leads to small savers paying more for loans than 20 reducing the overall loan rate charged to small savers by 5.8 percentage points, starting from the current rate of 34.2 percent, implies a reduction in small-saver loan rates of 16.8 percent (5 5.8/34.2). For ease of illustration, we assume that loan demand would not have been stimulated by this rate cut, so that the amount of loan interest and commission revenue paid by small savers would have been reduced by 16.8 percent. Taking 16.8 percent of the Us$12.06 million that centenary collected from small savers in loan interest and commissions (Table 3, line 5), gives a reduction in loan revenue and therefore profits of Us$2.02 million (516.8 percent 3 Us$12.06 million). This represents all of centenary’s profits from lending of Us$2.00 million (Table 3, line 6) and a tiny bit more, or 51 percent of centenary’s total profits from all sources of Us$3.94 million (Table 3, line 9). all calculations here and in the remainder of this section’s text refer only to the case of centenary with aTMs (scenario a), not to centenary without aTMs (scenario B). 21 loans disbursed per loan officer is a better measure of loan officer productivity in centenary than the usual loans carried per loan officer. This is because most loan officer time is spent on loan analysis rather than on loan monitoring and collection. 18 Figure 1: Should We Serve Small Savers Now? (based on what we know now and on a strictly business basis) Are small savers profitable in future years? Yes No 2. Yes, serve small savers now Are small Yes → 1. Yes, serve small savers now (but perhaps drop them in savers future years) profitable 3. Yes, serve small savers now? 4. No, don’t serve small savers No → now if future profits are now enough to offset losses now borrowers overall because small savers’ loans grow rapidly over time, as we find has happened in are more expensive to make. This higher cost recent years at ADOPEM.23 This growth in average of small-saver loans occurs because (i ) small savings account and loan size could convert small savers disproportionately demand more of the savers from an unprofitable client segment into a three higher cost types of loans (microenterprise, profitable one.24 housing, and agriculture) and less of the two lower cost types of loans (commercial and consumer),22 Figure 1 shows graphically and in a more complete and also because (ii ) the average size of the and rigorous fashion why the consideration of small savers’ loans is less than the average for future profits opens up an additional pathway borrowers overall for each one of the five types to profitability for the small savers. For ease of of loans. exposition we refer to the year in which we do the small-saver profitability study (2008 in our case) as The Evolution of Small Savers: “now� or “the current year,� and the years after A Final Pathway to Profitability that (2009 and afterwards in our case) as “the future� or “future years.� Figure 1 shows the four To the five pathways for achieving small-saver possible cases of present and future profitability in profitability, we add one more: profits in future years. the form of a 2 3 2 grid. The two rows tell whether The idea is that in cases in which small savers are or not small savers are profitable now (the first found to be unprofitable in the year studied (2008 row indicating that they are and the second row in the case of ADOPEM and Centenary), they still indicating that they are not). The two columns tell may be worthwhile serving in that year in order to whether or not small savers are profitable in the reap the profits from serving them in later years. future (the first column indicating that they are and As we will discuss, there are a number of strong and the second column indicating that they are not). fairly general reasons for thinking that the average The four cases are numbered 1–4. We defer for size of small-saver savings accounts and loans may now the question of how we might estimate future 22 at the end of 2008, for example, small savers held 25 percent of centenary’s overall loan portfolio, a figure that is surpassed by the 36 percent, 32 percent, and 48 percent small savers held of centenary’s portfolio of microenterprise, housing, and agriculture loans, respectively. small savers held only 6 percent and 23 percent of centenary’s portfolio of commercial and consumer loans, respectively, which are below the 25 percent overall average. similar results are obtained when we consider the portfolio at the beginning of 2008 (instead of at the end of 2008) or the amount of loans disbursed during all of 2008. 23 The data required to do this analysis were not available for centenary. 24 in the summary of findings section we explained why increases in average loan size generally lead to increased profits from lending, at least if all other factors—such as loan interest rates, loan operating costs, delinquency rates, and the cost of funds—are held constant. The same is true of increases in average savings account size, again subject to the proviso that all other factors are held constant. This proviso would mean, for example, that the average number of transactions carried out per savings account should not rise as average savings account size increases, so that operating costs per savings account remain constant. With constant operating costs per savings account, the only additional cost to the MFi of larger savings accounts is the extra interest it must pay on them. Because savings account interest rates are typically quite low (often 5 percent a year or less), growing savings accounts will generally provide an MFi with one of the cheapest sources of loan funding available to it, almost certainly cheaper than issuing certificates of deposit or borrowing from all but the most subsidized of sources. as a result, the MFi’s profits from lending should rise as additional funds become available to it from the growth in the average size of small saver savings accounts. 19 profitability, but suggest a possible approach later only current profits, but (ii ) we can also serve in this section. small savers now even if they are found to be unprofitable now, provided that future profits are The four numbered boxes in Figure 1 answer the large enough to compensate for current losses. question: should we serve small savers now, based on the information we have now and on a strictly In the remainder of this section we explore how the business (profitability) basis? In three of the four future profitability of small savers might be analyzed cases the answer is either an unconditional yes using multiyear data. We then discuss how our (cases 1 and 2) or a conditional yes (case 3); in the findings might or might not generalize to other MFIs. last case (case 4) the answer is a clear no. The ADOPEM Cohort Analysis and To see the logic behind these answers, let us Its Interpretation begin by examining the first row (cases 1 and 2), in which a current year profitability study (like At the time we did this study, the data needed to that presented in Table 3) finds small savers to calculate small-saver profitability were available up be profitable. In case 1 the answer is an easy and through 2008, but not for 2009 and afterwards. To unconditional yes since small savers are profitable attempt to understand how small-saver profitability now and in the future, so of course we would serve might evolve in those future years, we analyzed them now. Interestingly, in case 2—in which small the recent past, looking for trends in key indicators savers are profitable to serve now but (based on of small-saver profitability in the years leading up the information we have now) are estimated to be to and including 2008. Because the available data unprofitable in the future—we would still serve for Centenary did not allow us to do a satisfactory small savers now but perhaps drop them in future multiyear analysis, we present only the case of years if our forecasts of their unprofitability are ADOPEM. borne out by future data. So, again, we would unconditionally serve small savers now. ADOPEM’s history of mobilizing deposits from any great number of clients is relatively short. As Turning to the second row, in which small savers a result, we begin our multiyear analysis there are unprofitable now, case 4 is an easy decision at the end of 2006, dividing all clients present since small savers are unprofitable now and in the at that time into three groups.26 The 31,192 future; hence, there is no business case for serving “pure borrowers� have a loan but no savings them now. Case 3 is the one that provides the account on 31 December 2006. The remaining additional pathway for finding that small savers 23,980 clients all have a savings account on 31 are worthwhile to serve now, provided that future December 2006 (and may or may not have a profits are large enough to offset current losses.25 loan). We divide the savers into two equal size groups of 11,990 each, with small savers having We conclude that a multiyear analysis truly does the lowest average daily balances during 2006 provide an additional pathway to profitably and large savers having the highest average serving small savers now because (i ) we still serve daily balances during 2006. That is, we use the small savers if they are found to be profitable same procedure employed in our main 2008 now (regardless of their future profitability), just profitability analysis for dividing all savers into as we do in a single year analysis that considers small and large savers. We then follow all three 25 The forthcoming companion paper on methodology will discuss criteria for determining whether likely future profits are “large enough� to offset losses now. 26 Had we begun the analysis at the end of 2005, we would have been able to analyze the behavior of only 8,800 total savers instead of the nearly 24,000 available at the end of 2006. aDOpeM began mobilizing deposits in late 2004. 20 groups of clients for two years—from the end of attrition in the number small-saver savings 2006 to the end of 2007 and finally to the end of accounts over the two-year analysis period, the 2008—leaving each and every client in the group average size of the remaining accounts increases to which they were originally assigned regardless by 105 percent. This pushes up total small-saver of what they did after 31 December 2006 (started savings account balances by 68 percent. The borrowing, stopped borrowing, opened a savings combination of having to attend to 19 percent account, or closed a savings account). This type fewer accounts, but receiving 68 percent more in of study is called a “cohort analysis,� indicating total savings, almost certainly increases the profits that it examines fixed groups of people (cohorts) derived from the small-saver cohort.27 In fact, over time. even if total small-saver savings account balances had remained unchanged, small-saver profits The ADOPEM cohort data are presented in would most likely have risen. This is because the Table 5. The main conclusion we draw from the 19 percent reduction in the number of small- cohort analysis is that because the average size saver savings accounts would most likely reduce of small-saver savings accounts and loans in the number of savings account transactions and ADOPEM grew rapidly over the study period, therefore savings account operating costs. At the small savers appear to have become much more same time, small-saver savings accounts would profitable during this time. Moreover, as we argue have continued to provide the same amount of at the end of this section, there are strong reasons loanable funds as before, allowing ADOPEM to to believe that similar (though perhaps more earn the same profits from making loans with moderate) rising average size and profitability these funds as before. Putting that together with trends could hold for small savers in many MFIs. the reduced savings account operating costs Hence, this final pathway to profitability may be (from the 19 percent attrition), we know that of real practical significance. On the other hand, profits from small savers would increase. Profits our cohort analysis covers only three annual data would almost certainly be even higher with total points (spanning two years), an unfortunately small-saver savings account balances rising 68 short time series. Also, our analysis is based percent (instead of remaining unchanged). This on indicators of profitability, rather than on a is because, apart from operating costs (which we complete profit calculation, although the indicator have already taken account of), the only other trends are so strong that it seems quite likely that cost ADOPEM must pay for these extra funds ADOPEM’s small-saver cohort has indeed become is deposit interest. As discussed earlier, savings significantly more profitable over time. account interest rates are typically very low, and so this would provide ADOPEM with very low- To see how these conclusions were reached, we cost loan funding. As a result, ADOPEM’s profits turn to the data in Table 5, looking first at the from lending should rise as these additional savings accounts of small savers and then at their funds become available to it from larger savings loans. accounts. Just as we found earlier that small savers are not The next three data lines of Table 5 show necessarily small borrowers, here we find that that small-saver loans most likely become small savers today are not necessarily small savers substantially more profitable over time as well. tomorrow. Looking at the first three data lines of This is because overall loan volume increases Table 5, we see that while there is 19 percent slightly (4 percent), which suggests a small 27 at least this is true as long as we hold all other variables besides those in Table 5 constant. This proviso would rule out, for example, the existence of such contrary trends as a rapid rise in the number of transactions per account, which could push operating costs up over time despite there being fewer accounts. 21 Table 5: ADOPEM Cohort Analysis, end of 2006 to end of 2008 11,990 CLIENTS WHO WERE SMALL SAVERS ON 12/31/06 Total percent change from 12/31/06 12/31/07 12/31/08 12/31/06 to 12/31/08 Savings accounts Number of savings accounts 11,752 11,246 9,511 219.1% Average balance (US$) 5.81 8.88 11.91 105.0% Total balance (US$) 68,294 100,825 114,535 67.7% Loans Number of loans 9,606 7,087 5,468 243.1% Average balance (US$) 444 631 814 83.4% Total balance (US$ millions) 4.26 4.47 4.45 4.4% 11,990 CLIENTS WHO WERE LARGE SAVERS ON 12/31/06 Total percent change from 12/31/06 12/31/07 12/31/08 12/31/06 to 12/31/08 Savings accounts Number of savings accounts 12,064 11,942 11,401 25.5% Average balance (US$) 45.09 42.47 44.74 20.8% Total balance (US$) 543,992 507,218 510,092 26.2% Loans Number of loans 9,597 7,597 6,193 235.5% Average balance (US$) 655 890 1,102 68.2% Total balance (US$ millions) 6.29 6.76 6.82 8.5% 31,192 CLIENTS WHO WERE PURE BORROWERS ON 12/31/06 Total percent change from 12/31/06 12/31/07 12/31/08 12/31/06 to 12/31/08 Savings accounts Number of savings accounts 0 10,022 13,250 - Average balance (US$) 0 20.90 20.86 - Total balance (US$) 0 209,434 276,334 - Loans Number of loans 31,200 19,617 13,698 256.1% Average balance (US$) 521 743 937 79.8% Total balance (US$ millions) 16.25 14.58 12.83 221.1% Source: ADOPEM. increase in interest income, while the number We conclude that as long as the average balances of loans falls considerably (43 percent), greatly of small-saver savings accounts and loans increase reducing the operating costs of attending small- by enough to offset the effects of attrition, and saver loans. Again, the rapid rise in average size thus prevent the total balances of savings accounts (this time of loans, instead of savings accounts) and loans from falling, small-saver profits will most is behind the favorable profit trend. likely increase.28 In the case of ADOPEM, average 28 in fact, even if the rise in average balances falls short of this, small-saver profits may still increase. For example, with aDOpeM’s 43 percent loan attrition, the operating cost savings will most likely be substantial. even if total small-saver loan balances had dropped moderately, reducing interest income somewhat, profits could still have risen because the reduction in costs could be much greater than the loss in revenue. 22 savings and loan balances increased by more than since that time, there were undoubtedly many this, especially savings balances, which should new savers in both the small- and large-saver push up small-saver profits even further. cohorts at the end of 2006, when these cohorts were established. (Although it is arbitrary, new It will help us to understand the rapid rise in the savers might be defined as savers who opened average small-saver savings and loan balances their savings account in the last six or 12 months.) in ADOPEM—and thus assist us in generalizing As new clients leave their funds on deposit longer from these results to other MFIs—if we note two and longer with a good MFI, such as ADOPEM, characteristics of ADOPEM’s large savers and pure many are likely to become increasingly reassured borrowers, the other two groups in Table 5. First, that their funds are safe and convenient to access, in contrast to the sharp rise in the average savings and thus become increasingly willing to deposit account balances of small savers, average savings additional funds. Consequently, in a cohort account balances of large savers are essentially analysis, the average deposit balance would have stagnant (falling 1 percent over the two-year period). a tendency to rise over time in both the small- and Second, average loan balance of all three cohorts large-saver cohorts.30 rises rapidly during the 2006–2008 analysis period. On the second point, regression to the mean is What might lie behind the rapid increase in the a phenomenon that is widely observed for many average loan balance of all three groups? At least economic variables, including household savings. part of the answer is likely to be ADOPEM’s use The idea is that a variable that is extreme on its of the progressive lending scheme. Widely used first measurement will tend to draw closer to the by MFIs, progressive lending starts new borrowers average (or mean) in later measurements. The with small loans and works progressively up to reason for this tendency is that extreme values larger and larger loans, provided the preceding are often due, at least in part, to transitory effects. loans were repaid satisfactorily. For example, very low household savings levels may be due to unusually low profits from a family On the savings side, we believe that there are at business, job layoffs, unusually large medical or least two factors at play in explaining the rising other expenses, or other negative shocks. Very average balances of small savers and the stagnant high household savings levels may be due to such average balances of large savers: the presence of factors as having recently had unusually good many new savers and the phenomenon known as success in business or in controlling household regression to the mean. expenditures. Returning to ADOPEM, as a result of this phenomenon, we expect there to be a To understand the importance of new savers, tendency for the average savings balance to rise we begin by recalling that there is widespread over time in the small-saver cohort since this group agreement that the most important issues to is likely to contain many depositors whose savings savers in deciding where to save are usually level is unusually and transitorily low. Similarly, the the safety of their deposits from loss first and average savings balance of the large-saver cohort foremost, followed by convenience (being able will have a tendency to fall over time. to easily deposit and withdraw their money).29 With ADOPEM mobilizing deposits only since late Now let’s look at the combined effect of these 2004 and the number of savers growing rapidly two factors. In the small-saver cohort, both factors 29 For example, see Branch (2002) and Deshpande (2006). 30 This may be viewed as a form of progressive lending only in reverse, with clients depositing more as the MFi earns their trust. 23 act in the same direction, pushing up the average References savings balance of this group. This may explain, at least in part, the rapid rise in the average savings ADOPEM. 2008. Memoria Anual (Annual Report). balance of ADOPEM’s small-saver cohort. In the Santo Domingo: ADOPEM. www.BancoAdopem case of the large-saver cohort, the two factors act .com.do in opposite directions, tending to cancel each other out. This may at least partly explain the stability of Branch, Brian. 2002. “Savings Product Management: the average savings balance of ADOPEM’s large- Establishing the Framework.� Chapter 3 in Brian saver cohort. Branch and Janette Klaehn, eds. Striking the Balance in Microfinance: A Practical Guide to Mobilizing These three arguments (one for lending and two Savings. Washington, D.C.: Pact Publications for for savings) are likely to carry over from ADOPEM World Council of Credit Unions. to many other MFIs, since Centenary Bank. 2008. Annual Report and Financial • Progressive lending is widespread among MFIs. Statements. Kampala: Centenary Bank. www. • Microfinance is generally a rapid growth CentenaryBank.co.ug industry. Moreover, the relatively recent introduction of savings products in many Deshpande, Rani. 2006. “Safe and Accessible: Bringing MFIs means that deposit-taking will often Poor Savers into the Formal Financial System.� Focus be an especially fast growing area within Note 37. Washington, D.C.: CGAP. microfinance as new markets are penetrated and new savings clients are rapidly acquired. Deshpande, Rani, and Jasmina Glisovic-Mezieres. Consequently, there may be particularly high 2007. “The True Cost of Deposit Mobilization.� percentages of savings clients who are new to Washington, D.C.: CGAP. http://www.cgap.org/gm/ an MFI at any point in time. document-1.9.2117/45789_file_THE_TRUE_COST_ • Regression to the mean is a well-accepted, OF_DEPOSIT_MOBILIZATION.pdf general phenomenon, with household savings a classic illustration of the principle. MIX. 2009. Microbanking Bulletin, No. 18 (Spring). Washington, D.C.: MIX. www.TheMIX.org We conclude that there are important general forces tending to push up the average savings Portocarrero, Felipe, Alvaro Tarazona, and and loan balances of small-saver cohorts over Glenn Westley. 2006. “How Should Microfinance time. This means that even if small savers are not Institutions Best Fund Themselves?� Washington, profitable to serve now, they could be in future D.C.: Inter-American Development Bank. years, providing one final reason why MFIs may find it beneficial to serve small savers even from a Rutherford, Stuart. 2000. The Poor and Their business point of view. Money. New Delhi: Oxford University Press. No. 18 September 2010 Please share this Occasional Paper 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, 2010 The authors of this Occasional Paper are Glenn D. Westley and Xavier Noel Rutebemberwa, Robinah Isakwa, James Matovu, Deo Ilungole, Martín Palomas of CGAP. The authors would like to thank the many Charles Onen, Betty Matovu, Nelson Ogua, Bob Mayeku, and Kellen people at ADOPEM and Centenary Bank who welcomed us to their Ssali at Centenary. We would also like to thank Henry Sempangi and institutions and generously provided the data and information we Samson Odele of Microsave Uganda for their excellent work running needed to carry out this study. While there are too many individuals to randomly selected locations all over Uganda to carry out the ATM to name them all here, we owe a special debt to Mercedes Canalda user survey. Many other people made important contributions to this de Beras-Goico, Mercedes Pimental de Canalda, Eva Carvajal, Sonia study with their ideas and assistance, including Julie Abrams, Anita Reyes, Fernando Pérez, Juan Francisco Terrero, Gloria Román, Eddy Campion, David Cracknell, Jennifer Isern, Gautam Ivatury, Ignacio Mas, Santana, María Estela Terrero, Ivan Moquete, Hector Almanzar, Olga Kate McKee, Mark Pickens, Beth Rhyne, Marguerite Robinson, Evelyn Araujo, Cecilia Ramón, Quisqueya Domínguez, Blanca Español, Wilson Stark, Blaine Stephens, and Victoria White. Finally, Alexia Latortue, Peña, and Rosa de León at ADOPEM and Simon Kagugube, Joseph Kate McKee, Rich Rosenberg, Mark Schreiner, and Jeanette Thomas Kimbowa, Joseph Lutwama, George Thogo, Philip Irama, Jennifer all contributed helpful comments on an earlier draft of this paper and Kaggwa, Katimbo Mugwanya, Peninah Kasule, Andrew Ssemaganda, Anna Nunan very ably edited the final version. A final thanks goes to Francis Ogwang, Fred Mukwanga, Catherine Jamwa, Charles Kabanda, the Bill & Melinda Gates Foundation, which provided partial financial Patrick Woyaga, Expedito Kalyango, Dennis Echeru, Dominic Lagom, support for this work. The suggested citation for this Occasional Paper is as follows: Westley, Glenn D., and Xavier Martín Palomas. 2010. “Is There a Business Case for Small Savers?� Occasional Paper 18. Washington, D.C.: CGAP, September.