NUMBER 104 ~~~~~~~~~M I * ED Precisrn Operations Evaluation Department February 1996 KUPEDES: Indonesia 's Model Small Credit Program By 1983, years of heavy losses and fis- tant as providing them with loans. the years the banks evolved into cal cuts had made the vast network In Indonesia, SIMPEDES, the village a vast network of 3,300 units em- of village banks belonging to the savings program, attracted thousands ploying 14,300 people. But by state-owned Bank Rakyat Indoniesia of depositors. And by mobilizinig rural 1983, heavy losses and fiscal unsustainiable. Shutting the banks savings, it not only provided the banks cuts (the latter in response to down risked serious unlemployment with a stable source of funds, it also Indonesia's structural adjustment and the loss of an important credit de- kept finanicial savings in rural areas, program) had made the banks livery mechanismn. The government thus helping continue development in unsustainable. When the BIMAS responded by taking a novel approach. the countryside. program ended, the government It established a new, nmarket-based ru- decided to transform the bank ral small credit program, the Kredit Finally, the audit highlights the units into self-sustaining full ser- Umumll Pedesan (KUPEDES), aimed main reasons for village banks' success: vice centers to provide financial at transforming village banks into self- simplicity of loan design, which en- services to small borrowers at sustainingfull-servicefinanicial units. abled the banks to keep costs down; ef- market rates. As a first step, the The program's principal goals were to fective management at the unit level banks were restructured. Many of provide credit to small borrowers at backed by close supervisioni and monli- the units were moved to market market rates and to mobilize rural sav- toring by the center; and appropriate centers, and about a third were ings. By the time the World Bank's staff training and performance incen- downgraded to village service $102 millioni loan to KUPEDES was tives. By themselves, however, these posts to reallocate credit resources. approved in 1987, the program was factors would not have been enough for An accounting system was in- already zorking with a profit. By loan the program to succeed. The presence stalled in each unit to ensure ac- completion in 1992, the programn's as- of an existing institutional network of countability and performance. sets had grown 21 times over the 1984 village banks and a thriving economy level. The village banks were reachinlg that had spread to many rural areas The program, Kredit Umum 5 percent of the country's households were also critical. Pedesan (KUPEDES), began in with loans and 15 percent with deposit 1984, on the heels of the financial services. The KUPEDES program A recent OED audit* draws two Background and program design *Perforance audit report, major lessons from the KUPEDES "Indonesa: r eDES experience. First, it is possible to Since the 1970s, Bimbingan "Indonesia: BRI/KUPEDES create a microlending program that Massal, or BIMAS, an agricultural Small Credit Proect," Report serves the poor and is profitable and diversification program with a ru- No. 14511, May 1995. Perfor- self-sustainzing. But to succeed the ral credit component, had been mance audit reports are avail- bank units will need to lend at market providing small loans to farmers and staff from the Internal rates, use their income to finanice their at below market rates. BIMAS documents Unit and from Re- operations, and devise appropriate sav- channeled the loans through unit gional Information Services ings instrumenits to attract depositors. desas (village banks) belonging to Centers. Second, developing a savings instrn- Bank Rakyat Indonesia (BRI), a ment for the poor is at least as impor- state commercial bank. Through percent a month at neighbor- Box: The unit desas' structure and operation hood stores. Private banks of- fered credit at about 3 percent a month. In 1993, the Bank Rakyat structure of each unit was kept Indonesia ran a vast network of simple. Units were staffed by at World Bank loan almost 3,300 unit desas (village least four persons and at most 11. banks) and 620 village service If the business of a unit expanded The World Bank provided the outposts. Together, the units em- beyond the maximum staff limit, The With a provin ployed 16,000 people, represent- the unit was split, thus keeping program with a major boost in ing the biggest single rural bank- the operation small and focused. 1987 through a $102 million loan, ing network in the country. The of which $97 million was for transition from nonmarket to Institutional incentives were onlending and the rest for tech- commercial banking required used to motivate staff and instill nical assistance and training. The a major change in company cul- competition among them. For loan had three objectives: (1) to ture and work ethics. This trans- example, performance bonuses support BRI's efforts to develop formation took place during the allowed 10 percent of the unit the unit desas (village banks) into first four years of KUPEDES, desas' profits (up to one-and-a- a financially viable network of and produced clear operational half months of an employee's full-service subdistrict branches; procedures. salary in a year) to be paid to (2) to encourage BRI to improve staff. To promote competition its resource mobilization to fund Every subdistrict town had among the units, the central of- KUPEDES, thus reducing its reli- at least one unit desa. Staffing fice devised a system to recog- ance on central bank credits; of units was determined by a nize the top 20 performers of a ratio-for example, one credit each region, who then competed and (3) to improve the institu- officer per 400 borrowers, one for becoming the top 20 perform- tional capacity of both BRI and teller per 200 daily transactions ers at the national level. This KUPEDES. The onlending por- in automated units, and one competition was based on several tion of the loan was disbursed teller for 150 daily transactions criteria, including savings mobi- by September 1989, 31 months in nonautomated units. Village lized, outstanding portfolio, ahead of schedule. service outposts were manned short- and long-term loss ratios, by two people. The operational and management. Findings KUPEDES became a model credit program in Indonesia. Be- liberalization reforms introduced cent yearly interest rate, as op- tween 1984 and 1994, it grew six- in mid-1983. The government (in- posed to 12 percent under the fold, much of which occurred cluding the central bank) pro- old subsidized program. In addi- during the first three years. The vided about 210 billion rupees tion, a refundable 0.5 percent increase in the number of new ($20 million in 1984 dollars) in monthly penalty was added as loans and in the maximum loan seed capital and start-up loans incentive for early repayment. In size (to $3,150) led to another and two years' administrative 1984, the program also introduced growth spurt between 1989 subsidy to help cover projected the Village Savings Program, or and 1990. early losses. SIMPEDES, to encourage rural savings and to meet the deposit On two occasions (in 1988 and Simplicity was the critical ele- needs of clients. again in 1991-92), the program ment in the program's design. Its suffered setbacks. Drought, mild only restriction was loan size, By 1986, the bank units had in 1988 and long and severe in which automatically limited the fully used the seed capital and by 1990-92, caused a rural recession, loan's appeal to small borrowers. year's end were turning a profit. pushed up arrears, and reduced (In 1984, the minimum loan was KUPEDES loans had grown three demand for credit. The BRI's ini- equivalent to $24 and the maxi- times and deposits had expanded tial response was to restrict lend- mum to $930.) Otherwise, the fourfold. The initial fear that the ing, which only increased the banks were free to offer loans to high interest rates would make portfolio's deterioration because any creditworthy borrower in any the program unacceptable to it kept many creditworthy clients sector and for whatever purpose. small borrowers never material- from getting loans. But in 1991, Instead of charging a yearly rate, ized. A major reason was the BRI lifted the restrictions and in the banks followed rural lending much higher credit rates demand- 1992, in keeping with the Project traditions, charging a flat lending ed by informal lenders, ranging Agreement with the World Bank, rate of 1.5 percent a month. This in 1989 from 30 percent a month issued a plan of action to reduce method translated into a 33 per- for suppliers of raw material to 6 delinquencies. These steps pro- Februtary 1996 duced immediate results. Arrears Institutional capacity oped by BRI in 1984, SIMPEDES started to decline steadily, from 8.6 offered a safe, convenient, and liq- percent of the portfolio in 1991 to The village banks developed uid savings instrument that paid 5.6 percent in 1994. into an effective channel for bank- savers market-based interest rates ing services, playing an important (positive in real terms, that is, net Who was served role in financial intermediation of inflation) and allowed them un- (see box). Moreover, BRI now has limited withdrawals. The savings KUPEDES made credit avail- a cadre of 16,000 trained employ- plan evolved from market surveys able to small borrowers in villages ees. Banking technology has ex- of villagers and its design was re- across the country. In spite of the panded to village units, enabling fined through pilot studies. Its market cost, small borrowers were the 3,300 units and 622 village ser- purpose was to meet the needs able to obtain reliable credit more vice posts to reach close to two mil- of rural organizations and house- cheaply than from informal chan- lion borrowers or about 5 percent holds. The program allowed de- nels. The majority of borrowers of Indonesia's estimated 36 million posits to be made in the name were from landless and near land- Indonesian households. of an organization under joint less families, who used their house savings held by the group leader plot or house to guarantee their Savings mobilizationi or treasurer. loans. Women participated ac- tively. A 1989 survey found 25 Savings mobilization was To mobilize deposits, the village percent of the sample borrowers to the most successful element of banks used creative approaches. be women. In reality the number of KUPEDES, supplying the bank They conducted strategic mapping women benefiting from KUPEDES units with a stable source of funds. of potential savers, advertised their is probably much higher since By the end of 1993, savings ac- services, visited potential custom- small family enterprises in Indone- counts (excluding current accounts ers, and established links with vil- sia involve both husband and wife. and time deposits) made up almost lage chiefs and community leaders. 28 percent of the accounts and Many organizations responded, The lending program also be- 11 percent of the rupiah value in ranging from village treasuries and came fairly diversified. In 1994 the entire Indonesian banking sys- government offices to religious in- more than half of the loans financ- tem. By 1994, village banks were stitutions, town organizations, and ed microenterprises and small- servicing more than 12 million sav- the like, helping to create a large scale trading activities. Agri- ings accounts. With many house- market for rural deposits. As cultural lending accounted for holds having several savings of 1994, deposits had grown 27- about one-fifth of the lending pro- accounts, the program reached fold from their 1986 levels when gram. The biggest increase was in an estimated 15 percent of the SIMPEDES was adopted as a na- loans for working capital, used country's households. tional program for the unit desas. mostly for home construction and And savers substantially outnum- repairs. Started in 1989, working The unit banks offered a variety bered borrowers, thus helping the capital loans made up 29 percent of of savings schemes, but their major program achieve self-sufficiency the portfolio by 1994. program was SIMPEDES. Devel- (see figure). Moreover, by helping Indicators of self-sufficiency: growth of KUPEDES savings and loans 5000 - 14 . /w ~~~~~12 10 3000 ~ ~ ~ ~ ~ ~ 2000 , 6 1000 1984 1986 1988 1990 1992 1994 1984 1986 1988 1990 1992 1994 -c- Loans outstanding * Number of savers -_ Deposits E Number of borrowers OED Precis to mobilize rural savings within sound banking principles is appli- cal experts who not only under- the banking system, the program cable to most small credit efforts. stood the issues but appreciated provided a stable, credible chan- The use of a variety of institutional the need to adapt solutions to nel for keeping financial savings innovations (including those that Indonesia's unique setting. in rural areas. helped motivate staff), attention to savings instruments, and close su- * A focus on objectives and simplic- The role of the World Bank pervision and monitoring are also ity of design are critical in credit of general applicability. But in the programs geared to self-sufficiency. The World Bank supported end, institutional development KUPEDES's objective of turning KUPEDES at a time when the takes time, and much nurturing. money-losing units into self-sus- program was still fragile and The audit highlights some of the taining profit centers formed the prone to outside pressure, par- important lessons from Indonesia's framework for the program's ap- ticularly because of its contin- experience. proach to the types of financial uing dependency on central bank services offered, the policies and loans. Moreover, Bank staff, * Success of rural credit programs procedures selected, the organiza- aware of the poor results of many is facilitated by sound macroeconomic tional characteristics adopted to run targeted and subsidized credit and sector performance. KUPEDES the program, and the type of sup- programs, strongly supported the was initiated at a time of rapid port and technical assistance re- program's simple, nontargeted growth spurred by Indonesia's quested from the government, design and working principles. structural adjustment and sound donors, and the Bank. The success As a result, the Bank lent credibil- sector development program. The of the BRI unit desas system, and ity to the program's innovative economic expansion fed into the thus the BRI/KUPEDES project, approach and protected it from growth of incomes in rural areas, rested on its simple design, which pressure to move to a targeted and the strong rural infrastructure in turn, was guided by good poli- approach. and agricultural supporting ser- cies: (1) nontargeted credit, except vices programs helped to increase for loan size; (2) nonsubsidized Lessons demand for credit, and as impor- lending; and (3) self-sustainability tant, create the basis for growth of through savings mobilization and KUPEDES evolved in response domestic savings in rural and profitable operations. to the failure of the subsidized small households. BIMAS program and on the * Simplicity in design must be strength of an already existing * Government commitment and backed by careful attention to organ- delivery network. It took a major ownership are at the heart of a izational detail and staff incentives. challenge to the survival of the sound market-based credit program. The program's management insti- delivery system to make the KUPEDES was designed and tuted a system of quality cbntrol KUPEDES's market-based ap- launched by the government, with that emphasized monitoring, proach an acceptable option. As help from long-term foreign con- supervision, and accountability a result, some of the program's sultants from the Ministry of Fi- backed by significant attention to innovations may be unique to nance. Thus, the program was staff training. Management also Indonesia. Nevertheless, the implemented with a high level devised performance incentives, program's reliance on simplicity of government commitment and rewarding employees and units of design and application of guided by carefully chosen techni- for good performance. OED Precis is produced by the Operations Evaluation Department of the World Bank to help disseminate recent evaluation findings to development professionals within and outside the World Bank. The views here are those of the Operations Evaluation staff and should not be attributed to the World Bank or its affiliated organizations. This and other OED publications can be found on the Internet, at http:// www.worldbank.org/html/oed. Please address comments and enquiries to the managing editor, Rachel Weaving, G-7137, World Bank, telephone 473-1719. Internet: rweaving@worldbank.org February 1996