37736 TIMOR-LESTE ACCESS TO FINANCE FOR INVESTMENT AND WORKING CAPITAL Prepared for the Word Bank and the Government of Timor-Leste by John Conroy 1 List of Abbreviations ADB Asian Development Bank AMFITIL Associasao Microfinancas de Timor-Leste BPA Banking and Payments Authority CEP Community Empowerment Project CGD Caixa Geral de Depositos (previously Banco National Ultramarino, BNU) CU Credit Union GDP Gross Domestic Product GoTL Government of Timor-Leste IFC International Finance Corporation IMfTL Instituicao Microfinancas de Timor-Leste MFI Microfinance Institution SEP Small Enterprises Project PEDF Pacific Enterprise Development Facility PSD Private Sector Development SIP Sector Investment Program TFET Trust Fund for East Timor WOCCU World Council of Credit Unions 2 TIMOR-LESTE ­ PRIVATE AND FINANCIAL SECTOR REVIEW REPORT ON ACCESS TO FINANCE FOR INVESTMENT AND WORKING CAPITAL Table of contents Executive summary 3 Recommendations 7 1. Introduction 13 2. Structural issues 15 3. Conceptual and definitional issues 18 The meaning of `private sector development' in Timor-Leste 18 The informal sector and wage-cost disadvantage 20 4. Broad Outline of the Financial Sector 21 5. The demand for financial services 25 6. Supply of financial services: the commercial banks 28 7. Supply of financial services: potential for entry of new commercial banks 32 8. Supply of financial services: The Microfinance Institute of Timor-Leste 35 History and objectives 35 Size and scope of operations 36 Viability and future status of MFITL 37 3 9. Supply of financial services: the microfinance sub-sector 38 10. Supply of financial services: other sources 42 11. Assessment: the adequacy of financial services in Timor-Leste 46 `Large' and medium-scale enterprises 46 Small- and micro-enterprises (including the urban informal sector) 47 Rural households 48 4 Executive Summary This study argues the need for a policy environment supporting both urban informal sector dynamism and rapid transition from subsistence to monetization in agriculture. Such policies must include measures facilitating access to financial services for households, which are the backbone of the informal and subsistence economies. The economy of Timor-Leste is divided between a farm sector in which as many of 80% of workers remain, with most of these still dependent on subsistence production, and a non-farm sector in which micro- and small enterprises are an overwhelming majority. Most urban enterprises operate in an informal environment, while in both the farm and non-farm sectors the household is the basic unit of economic activity. With only about 40,000 workers in the private sector and some 15,000 potential labor market entrants each year, there are severe limits on the capacity of the non-farm sector to offer wage employment to new job-seekers. In these circumstances, it is pertinent to ask: what is the meaning of `private sector development' in Timor-Leste? With such a large rural subsistence element remaining, it is appropriate to commence private sector development there. Diversification and monetization of farm output is crucial to raising rural incomes. Access to financial services is needed for the embryonic rural private sector, to support produce trading and the supply of manufactured goods to a rural economy making the transition from subsistence. Beyond agriculture, `informal' enterprises form a major component of the non-farm sector. The distinction between formal and informal enterprises is problematic in Timor- Leste. While business registrations give some guide to the extent of formal enterprise activity, even many registered entities exhibit `informal' characteristics, in terms of scale, use of family labor, home-based operation, technologies employed and other criteria. Thus while all `formal' enterprises are registered, not all registered entities in Timor- Leste can be regarded as formal. Numbers of workers employed provide some guide to the formal/informal divide. Among formal (i.e., registered) enterprises, while more than 70% have fewer than 10 employees, almost 20% have from 10 to 20 workers. Among informal (i.e., unregistered) enterprises, slightly more than 50% have only a single worker. The distinction is thus between the small and the very small; even in the `formal' sector, more than 90% of entities have fewer than 20 workers. Hence even most formal enterprises conform to the definition of micro-enterprise. In 2004, about half of informal (unregistered) enterprises fell below an annual turnover of $1000. But even in the formal (or registered) sector more than 15% of entities had turnover below this level. Urban enterprises, both formal and informal, were predominantly engaged in trading. Low levels of startup capital also categorize both groups. About a quarter of formal, but fully 90% of informal, enterprises had startup capital of less than a hundred dollars ($100). Few entrepreneurs of either category access non-personal or non-family resources for startup capital, suggesting extremely limited resort to institutional sources of finance and a potentially substantial market for microfinance services. 5 In the urban economy, much informal economic activity is stagnant and derivative in character. A more dynamic informal sector with better access to financial services would bring benefits, including increased alternatives to waged employment, a more equitable distribution of income and increased supplies of goods and services in the consumption baskets of urban dwellers. Greater supply and diversity of informal sector production will support policies of wage restraint designed to reduce Timor-Leste's relative wage-cost disadvantage. The formal financial system is still limited in outreach. It consists essentially of the banking system: three commercial banks, each an overseas branch of a foreign institution, and one local institution operating with a limited banking license. Each of these serves distinct but limited market niches within the broader population. Regulatory and supervisory functions for the banks are performed by the Banking and Payments Authority (the BPA). There are no non-bank financial institutions subject to BPA supervision and there is no insurance company serving Timor-Leste, nor any leasing finance entities. Beyond the formal, regulated, institutions, there are microfinance institutions (MFIs), savings and loan cooperatives, and pawnshops. While MFIs are of limited significance in dollar terms, they have succeeded in reaching between a fifth and a quarter of poor households. There are also informal mechanisms to provide financial services. The financial system of Indonesian East Timor was destroyed in 1999. Since the re- introduction of banking, the level of domestic credit has grown as a proportion of non-oil GDP, suggesting some recovery of financial depth, while rising bank deposits as a proportion of non-oil GDP point to improving levels of financial intermediation. Financed in the aggregate entirely by domestic deposits, bank credit rose very rapidly throughout 2004 to reach almost 22% of non-oil GDP early in 2005. This was only half the proportion for all `low-income' countries, of which Timor-Leste is one. The system was intermediating 87% of deposits to domestic lending. Together with evidence of excess demand for bank credit, this suggests the need for expanded savings mobilization to bring Timor-Leste up to a more appropriate level of financial depth. BPA is a central bank in embryo and is required to promote a market-based banking and financial system. It has explicit responsibility to nurture microfinance and to build savings capacity, including in rural areas. Financial sector development in Timor-Leste is constrained by lack of human resources in the sector, by low levels of `financial literacy' in the population and by a lack of ancillary services in the business community. Crucial deficiencies exist in the legal and regulatory environment for financial services. Most urgent is the establishment of a basic legal framework to establish, recognize and enable enforcement of property rights. There is substantial unsatisfied demand for financial services and the great majority of the population remains untouched by formal financial services. Historical experience from the years of Indonesian rule suggests that aggregate savings in Timor-Leste could increase substantially beyond the 2005 level. With a population of some 200,000 6 households and fewer than 60,000 deposit accounts in 2005, it would be appropriate to plan the extension of services to the great majority of these households, commencing with the 9 district capitals and 65 sub-district towns still largely unserved. A total of 200,000 deposit accounts in all classes of institution by 2010 would be realistic, if the financial infrastructure were made available. Innovative approaches would be necessary. Cellphone-based systems could allow financial institutions to operate away from their branches. For more remote areas, technology exists to allow field staff to provide offline services using handheld electronic devices. In addition, the planned expansion of postal service points would permit cautious trialing of deposit and remittance services using those new facilities. To meet demand for credit, the approach should be incremental, achieving judicious expansion by appropriate and sustainable institutions. Currently-active commercial banks will respond to excess demand for bank credit, while approval for the entry of one or two suitable new banks would increase competition. New entrants will have made their own feasibility studies. This is preferable to target-setting by authorities, who may instead aim to facilitate progressive and market-driven expansion of credit services in response to observed indicators of excess demand at the margin. Appropriate lenders will include MFIs and Credit Unions (CUs) in addition to banks, as well as new non-bank commercial lenders. There is no good case for either a state development bank or a government postal bank to remedy gaps in service provision. The three currently-operating commercial banks each serve distinct market niches. The entry of the third bank seems to have stimulated beneficial competition, evidenced by reduced fees and margins. Recent lending growth has been rapid and a substantial backlog of demand exists, testing the limits of domestic loanable funds. The BPA is ready to consider new applications for entry. Further diversification of banking operations within a competitive framework appears the most promising option for achieving outreach to unserved population groups, while lowering costs and improving service levels overall. Bank BRI, an Indonesian institution specializing in rural- and micro-banking and with particular strength in deposit-mobilization, conducted a study of the feasibility of re-entering the Timorese market in 2003. BRI did not proceed, but if it were to do so it would contribute greatly to diversity and competition within the system. Timor Leste also possesses a microfinance banking institution, the IMfTL. Granting a license for an institution similar to small, locally-based and successful microfinance banks in Indonesia and the Philippines was a positive innovation, offering a distinct and poverty-focused model of operation. Despite its relatively poor performance, some regulatory handicaps, and unresolved issues of ownership and governance, the presence of such a banking model within the Timorese financial system is an asset that should not be surrendered lightly. Amidst deliberations about its future, the option of its being acquired by a private, for-profit, entity for `back-door' entry to the banking system is not appealing. The best option for retaining IMfTL's distinctive focus could involve attracting equity and expertise from an institution such as the IFC and/or a private international, not-for-profit investor active in micro-banking. 7 Microfinance institutions (MFIs) have particular relevance to Timor-Leste's need for financial services, given the micro-scale of most economic units. This is recognized in official policy; the 2002 National Development Plan directed BPA to support donors and NGOs to develop micro savings and credit schemes, especially in rural areas. The BPA does not plan to create a regulatory environment specifically for MFIs. Instead it takes a non-interventionist and broadly supportive stance that tolerates limited deposit-taking by MFIs on the basis that they can be regarded as `membership' institutions. Service providers have formed a national association, AMFITIL, to establish a framework for setting performance and reporting standards, based on international best practice principles. It has eleven member institutions of which four, including IMfTL, are specialist providers. Members report performance data to on a regular basis. At end-2004, reporting institutions had outreach to some 26,600 savers and 19,200 borrowers and MFIs (excluding IMfTL) were dealing with more than 20% of poor households. Sustainability is a major problem for Timorese MFIs and they regard capacity-building as a particular need. AMFITIL provides a suitable platform for coordinating assistance to the sub-sector. The best MFIs have demonstrated comparative advantage in lending to the poor but are constrained in accepting deposits, while IMfTL has had difficulty in lending to the poor but has a legal mandate for deposit-taking. This suggests a working relationship between MFIs and the IMfTL could provide mutual benefits. In terms of the adequacy of financial services, all enterprises are penalized by real interest rates that are relatively high for dollar-denominated loans. Even larger firms with access to external sources of credit suffer from a domestic credit environment in which the absence of supplier- and consumer-credit facilities limits marketing and distribution. Smaller enterprises suffer particularly from the lack of supplier credit. All firms suffer to a greater or lesser extent from a lack of specialized financial services, such as insurance, equipment leasing and venture capital, which lag in the current legal environment, and are disadvantaged by the absence or limited availability of a range of complementary business services. Urban informal enterprises and rural households suffer many of the disabilities mentioned above and also the absence of domestic remittance services. Limited understanding of business and money (`financial illiteracy') is pervasive at this level. In rural households, but also for the urban poor, lack of access to safe, convenient and liquid deposit facilities is a particular deficiency. Recommendations A. General recommendations 1. Among other benefits, the maintenance of low inflation in a stable macroeconomic environment will support the development of an adequate and sustainable supply of financial services. 2. Concerning the role of financial services in stimulating private sector development, the Review recommends that priority be given to measures facilitating access to financial services for households, which are the backbone of the subsistence and informal economies. This will accelerate 8 the transition from subsistence in agriculture and the growth of income- generating activities beyond agriculture. The primary household service need is for deposit services, while credit, remittances and payments services are also required. 3. The Review endorses the emphasis of policies outlined in the GoTL Sectoral Investment Program for Private Sector Development. These call for a policy shift, from direct interventions to assure the availability of financial services, to increasing reliance on intermediation through the financial system and support for the creation of enabling environments. 4. Given the paucity of sources of credit in Timor-Leste, the Review views as appropriate a relatively permissive and non-interventionist approach to the emergence of private non-bank lenders. It would also be important for authorities to be sensitive to the continuing importance of a variety of non- formal and cross-border credit sources, and to appreciate their contribution to financing economic activity in Timor-Leste. 5. GoTL may want to focus on continue to remedy deficiencies in the legislative, regulatory and judicial frameworks affecting financial services, and in the array of relevant professional and business services, with particular emphasis on measures to permit financial service contracts to be secured and enforced. The Review turns to this subject in Section E, below. 6. Concerning the demand for financial services, the Review accepts that unsatisfied demand for credit and unfulfilled capacity for saving are both substantial in Timor-Leste. With regard to credit, the Review makes specific recommendations below concerning the need for a larger number and greater diversity of lending institutions. Also, consistent with its view that the household should be the primary focus of policy, it urges GoTL to set a national target to reach 200,000 deposit accounts, in a variety of institutions, within a five year period. This could commence with extension of services to the 6 district towns and 65 sub-district towns that are largely unserved at present. In addition, extension of services using cell phones and offline electronic devices needs to be trialed and developed. As the national network of postal service points is expanded, cautious trials could be conducted of using these for deposit and remittance services. 7. With regard to expanding the availability of credit, the Review recommends against setting quantitative targets, arguing instead for pragmatic and incremental expansion. Authorities may want to adopt an active stance, facilitating and monitoring the progressive expansion of services by a diversity of sustainable institutions in response to observed indicators of excess demand at the margin. However the Review recommends against the creation of a state banking institution for Timor-Leste. The prior expansion of deposit-taking facilities, to un-reached population strata and into un- served geographic areas, will provide indicators of credit capacity for the information of lenders and the authorities. However the SIP suggestion that postal centers could be employed for lending would add a more complex dimension to their functions, likely to result in failure. In urban areas and the 9 formal sector, reliance on market forces in an enabling environment will support the necessary growth of credit. 8. A major deficiency is a pervasive lack of `financial literacy'. This flows from limited familiarity of Timorese people with the monetary economy. It can also be viewed as an aspect of weak entrepreneurship, traceable to their hitherto limited role in the modern economy. Widespread financial illiteracy prejudices smooth functioning of financial systems and reduces the welfare of retail users of financial services. However, systematic saving by households engenders financial literacy. Financial institutions are among the agencies which can contribute to helping people understand the workings of a monetized economy, beginning with the discipline of saving. While the temptation to overload school curricula should be resisted, there is scope for schools to contribute to the financial literacy of young people. For budding entrepreneurs, Business Development Centers established under the Small Enterprise Project are a valuable resource, meriting further support from the international community and the GoTL. B. For the banking sector 1. A number of considerations support the case for one or more new entrants to the banking sector. These include evidence of excess demand for bank loans, reduced interest margins and improvements in services following the entry of the third bank and the potential for greater system stability with a larger number of lenders. Entry by an additional bank, even if it were specialized in a particular niche, could have a beneficial influence on services across the board in so small a market, as well as contributing to the overall resilience of the sector. Diversification within a framework of competition is likely to be the most promising option for achieving outreach to un-served population groups 2. GoT-L may want to re-examine the possibility of Bank Rakyat Indonesia's resuming commercial microbanking activities, with a view to encouraging the bank to re-work and re-consider its 2003 feasibility study for returning to Timor-Leste. 3. Specialized state banks set up to support rural and agricultural development, including in small island states in the Pacific, have a generally bad record. The case for such a bank in Timor-Leste is unconvincing. Nor would it be wise to see the proposal in the Sector Investment Program for a network of postal service centers to perform banking functions as a prelude to the creation of a state postal bank. 4. IMfTL has value as a distinctive banking model which brings diversity to the banking sector. Small regulated banks catering to a poor and low-income clientele and employing microfinance best practice have done well elsewhere in the region. Timor-Leste has made a regulatory breakthrough by licensing IMFTL as such a bank, even if regulatory constraints currently impede its outreach. The Review recommends that modes of ownership and governance be found that are consistent with its original mandate for serving the poor, as 10 well as being acceptable to the BPA for a class C banking license (to permit regulatory requirements on deposit-taking and liquidity to be relaxed). 5. Concerning options for the future of IMfTL, the least-preferable would be private-for-profit ownership intent on gaining `back-door' entry to the commercial banking sector. The best chance of fulfilling the Institute's original mandate could be to obtain capital and expertise from the IFC and/or a private international not-for-profit equity investor active in micro-banking. 6. While IMfTL should not abandon payroll lending, its micro-lending mandate should be pursued more energetically by a combination of direct retail lending and wholesale lending to the MFI sub-sector. For the former it would need to adopt best practice microfinance principles, and for the latter it could combine lending relationships with arrangements to accept some deposits from MFIs on behalf of their members. 7. A proposal for other commercial banks to engage in such `linkage' relationships with Timorese MFIs is made in the Sector Investment Program, together with a suggestion to provide credit guarantees. Unfortunately, the commercial rationale for seeking to induce bank/MFI linkages in this manner is not yet convincing in Timor-Leste, as credit guarantees raise the possibility of moral hazard. Commercial banks would probably have to be compelled to engage in such linkages on anything more than a `token' basis and such compulsion or suasion would set a bad precedent. 8. While credit guarantees are not advisable, another suggestion, that a bilateral donor might find means to `enhance' the bankability of selected enterprises, should be examined. The proposal involves providing assistance to small- and medium-scale businesses, concurrent with but not tied to their dealings with banks, designed to increase the chances of borrowers' commercial success. Such dealings involve a delicate balance between facilitation and moral hazard. If the right balance could be found this would be a useful initiative. 9. The same bilateral donor is considering whether it might support capacity within the Business Development Centers established under Small Enterprises Project (SEP) II to assist clients in dealing with banks, and specifically to help them prepare loan applications. To the extent such training produces entrepreneurs who have internalized the basic principles of business planning (as distinct from those clutching a pre-cooked loan application) this would be a positive contribution to lowering barriers to access. 10. Monitoring and analysis of developments in the banking sector would be assisted by improvements in the quality and relevance of sectoral lending data reported by the BPA. C. For the microfinance sector (MFIs and CUs) 1. Given the difficulties Timorese MFIs experience in achieving sustainability and their financial vulnerability, priority should be given to institutionalizing their presence in the country. This requires attention to sustainability and less emphasis on expansion except where scale is crucial to sustainability. Such a priority is consistent with a more general strategy of judicious expansion of 11 credit at the margin for all financial service institutions, as recommended by this Review. 2. The policy stance of the BPA in regard to MFIs, that specific regulation of the sub sector is inappropriate at this stage, appears to be sound. It elevates the significance of self-regulation procedures developed by AMFITIL, the industry association, and leaves MFIs more free to innovate and experiment. 3. BPA chooses for now to regard MFI deposit-taking as contributions by `members', and not in breach of the Banking Act. This position is reasonable so long as net balances (the excess of loans over deposits) are substantially in favor of the MFI in each case. Nonetheless, BPA's stance further underlines the need for effective self-regulation by AMFITIL and its members to protect the savings of poor people. 4. Consideration should be given to strengthening AMFITIL's capacity as an effective representative, standard-setter and (increasingly) de facto supervisor of MFIs. The setting and adoption of common performance standards, reported via the AMFITIL network and monitored by it, would be a most positive development for MFIs. While operational sustainability, based on fees levied from members, should be set as an objective to be reached within a determinate timeframe, AMFITIL would need interim assistance to set up and staff a free-standing secretariat. Partnership with external agencies including IFC/PEDF could serve to strengthen AMFITIL's pivotal role within the microfinance sub-sector. 5. MFIs have manifold needs for assistance. Capacity-building, in the broad sense that includes training for management and staff, technical assistance for many operational procedures and exposure to best practice operations in neighboring countries should all be considered. Transport and office equipment, and generators for electricity in many areas, are also necessary. Capacity-building assistance for clients is also needed. This can be seen as an approach to improving the financial literacy of a key group of micro- entrepreneurs. Business Development Centers established under SEP II now have capacity to contribute to this process. The fact that many potential clients are women needs to influence the approach BDCs take to such training. MFIs could benefit from linkages with a rejuvenated IMfTL whose mandate for poverty-lending had been confirmed and whose ownership and governance issues were resolved. Linkages could provide for wholesale lending to, and deposit-collection from, MFIs on behalf of their clients. 6. While at least one credit linkage between a commercial bank and an MFI has been established, the arguments for this occurring on a more general basis are not compelling on commercial grounds. 7. Decree legislation has been passed to regularize the status of the credit union movement, a positive step. CUs are legally entitled to mobilize member savings and serve a market segment between those of MFIs and banks. CUs ought to feature in any concerted campaign to lift national savings capacity, but the danger of loss of deposits by weak CUs gives pause to plans for their expansion. Perhaps the best short-term strategy is for World Council of Credit 12 Unions (WOCCU) or one of its affiliates to maintain low-level assistance relationships with the CU movement. D. The need for new financial products and complementary services Discussion of the adequacy of financial services, in section 10, below, and elsewhere, leads to a checklist of financial products and complementary business services whose availability would support more rapid private sector development. Given the banking sector is relatively better developed, the more immediate financial product deficiencies are found in the non-bank sector. 1. Among bank products, the absence of domestic money transfer, or remittance services, is notable. For the transfer of funds to unbanked areas, the proposed postal transaction centers offer the most likely channel, although electronic service possibilities discussed below in section 5 and elsewhere should also be considered. With many Timorese living overseas and with the commencement of recruitment of Timorese for offshore employment contracts, it should also be possible for the postal network to achieve working relationships with the foreign exchange banks or with Western Union to enable transmission to workers' home villages. This may be an early priority for the network upon its establishment. 2. A number of important non-bank credit products were not available in Timor- Leste at the time of the Mission, although in the case of equipment leasing the legislative and regulatory preconditions were in final draft preparation and have since been passed. Finance companies offering personal loans or facilitating supplier and consumer finance were not active, although there was no legal barrier to their establishment (as seen in the case of the pawn broking operation described in section 10). Given that BPA has a permissive attitude, the principal concern of public policy should be to require transparency in loan contracts, rather than to promote such lending actively. For the largest enterprises, venture capital facilities would be a useful addition to the small range of financing options currently available. However, improvements in the broader legislative environment for company formation, collateral and enforcement of financial contracts would be necessary to promote start-ups in this field. 3. Insurance has been a major gap in the range of financial products available, although regulations are now in place. Applications for insurance licensing however, are still to come. As with banks, the issues of the number of players and the credentials of applicants are policy matters for decision by the BPA. 4. With regard to deficiencies in complementary business services, at least one bank visited by the Mission regarded the absence of notary services as a constraint on its capacity to lend prudently, though this may reflect differences in national banking cultures and practices. It probably also reflects a divergence of national cultures in the legal profession and system in Timor- Leste. Deficiencies in accounting and auditing services are undeniably a barrier to the interpretation of financial statements as normally required for loan proposals. The absence of a national credit register for the banking sector, 13 has been another constraint on willingness to lend. The BPA has taken some preparatory steps towards preparing for the establishment of a credit registry with the support of technical assistance. This is a vital step towards improving information sharing, risk management and willingness to lend. A collateral register is a facility sometimes mentioned as desirable to assist lenders to avoid fraud, but this argument seems less cogent than the case for solving the broader issues of loan enforcement. In short, the solutions to most of these problems is likely to be incremental; action to attract accounting firms to set up shop may be possible but time will be required to sort out problems pervading the legal and judicial systems. E. The legislative and regulatory environment 1. Deficiencies in the legislative and regulatory environment affecting financial service provision are quite well understood. However, the difficulties of putting an appropriate framework in place are due to limited capacity in a Government struggling to secure reform on all fronts and to the need to fashion coherent arrangements in a system with elements drawn from several legal systems. The establishment of a basic legal framework, particularly with regards to property and land, would provide an adequate interim basis for growth in economic and financial activity, pending an incremental process by which specific legislative gaps are filled and more intractable deficiencies are addressed. Legal impediments to the provision of commercial insurance and equipment leasing are among the specific gaps to which early attention needs to be directed. Into this latter category fall systemic weaknesses in commercial dispute resolution and enforcement. These cannot be remedied quickly, given human resource deficiencies in the judicial and bureaucratic systems. More specifically, implementation of the law for insurance is an urgent priority as is the establishment of the insurance industry itself in Timor-Leste. Bankruptcy legislation and regulations are likewise of urgent concern to lenders. Legislation for equipment leasing also needs urgent attention as it offers some relief for the collateral issue. 14 TIMOR-LESTE ACCESS TO FINANCE FOR INVESTMENT AND WORKING CAPITAL 1 Introduction Under the Indonesian regime, the former East Timor had a GDP per capita level around one-third of the Indonesian national average. In consequence, some 30% of population (double the general Indonesian proportion) were below the poverty line. In 1993, 71% of all villages in the province were classed by the Indonesian government as desa tertinggal, or `backward villages' according to criteria of isolation, infrastructure, housing and physical environment. Indicators of human development reflected this situation. Bureaucratic and commercial life was dominated by outsiders, with `modern' sectors such as manufacturing and services (including financial services) largely controlled by non-Timorese. In 1999, East Timor suffered severe loss of infrastructure, displacement of population and disruption of economic activity. While the dramatic events that led to this tragedy are well-known, it is worth recalling some of the consequences. Loss of capital stock was extensive; for example, it is estimated that 70% of housing was wholly or partially destroyed. Dili lost a third of electricity generating capacity and losses elsewhere were higher still. Of particular interest for this Review, all banks were destroyed, their records lost or repatriated to Indonesia, their equipment removed. Not even a single bank safe remained. The financial system in 1999 consisted of 18 branches of commercial banks, comprising 5 branches of state banks, 10 branches of the provincial development bank and three branches of private Indonesian banks. This represented approximately one branch per 47,000 people. In addition a state bank, Bank BRI, operated a dozen village banking units, described below. There were a number of credit unions and a credit union federation (section 9, below) but none of the licensed rural banks (BPRs) or other small community financial institutions common in Indonesia. While some microfinance institutions (MFIs) operated in West Timor, the political environment in the east of the island was not conducive to NGO activity, still less for those NGOs with international connections. The international relief effort under UN auspices, from late in 1999, was concerned initially with humanitarian relief and then increasingly with reconstruction. Among the early initiatives were several projects concerned, directly or indirectly, with elements of financial system-building (section 10, below). After a confused period, during which several other currencies circulated alongside the Indonesian rupiah, the UN transitional administration (UNTAET) addressed the issues of a currency and a monetary authority for an independent Timor-Leste. The US dollar was adopted as sole legal tender early in 2000 and later that year an UNTAET decree set out regulations for bank licensing and supervision. In 2001 another decree established the Banking and Payments Authority 15 (BPA) as an embryonic central bank (section 4). Guidelines for government financial sector policy were set out in a National Development Plan, approved by the newly- elected parliament of independent Timor-Leste in 20021 (section 4). While external assistance continues, the peak of expenditures occurred in 2001. For a time a `bubble economy' flourished, based on construction and the presence of large numbers of international personnel and causing substantial distortions in the wage- and cost-structure. Wage-levels now substantially higher than in Indonesia and other neighboring states will continue to bedevil economic policy and affect economic development. As the foreign presence decreased from 2002, the bubble economy deflated, private sector employment declined and many enterprises experienced difficulty. These conditions were apparent in mid-2005, although recovery from drought and rapid expansion of banking activity were providing some stimulus. Unemployment was high, and 40% of the population lived below the national poverty line. Population growth also posed particular problems for Timor-Leste. Provisional results of a national census conducted in 2004 show a total population of some 925,000, of whom almost 50% were under the age of 162. This population structure, very young by international standards, poses acute problems of educational provision and job creation, as well as suggesting the possibility of continuing high crude birth rates in future decades. Some concern was expressed about urbanization following the events of 1999, but the preliminary data do not permit reliable calculations of urban growth rates. The district of Dili, in which the national capital is located, has 18% of the enumerated population, while the district of Baucau, the second city, holds 11%. Thus Timor-Leste is not highly urbanized although Dili is certainly a primate city in economic terms. Doubts about the validity of earlier population counts make estimates of recent population growth problematic, although the current population structure suggests it has been quite high for the past decade or more, and not simply in response to the trauma of 19993. The implications of current population structure for work force growth are discussed in section 2. In 2005 the IMF described Timor-Leste as `successfully transitioning from post-conflict status.' The Government of Timor-Leste (GoTL) is fortunate to have access to substantial oil and gas resources. Managing tax and royalty revenues and framing economic policy for a `dollarized' economy with a distorted wage structure are tasks which will pose particular challenges. GoTL policy is to eschew debt while adopting a Petroleum Fund model for managing and sustaining the finite stream of oil and gas revenues that will accrue over the next two or three decades. Following such a path successfully will require political will, good governance and fiscal discipline. These challenges occur in an economy with per capita GDP of only $367 in 2004, and a human development (HDI) ranking of 158 (of 177 countries ranked). 1Planning Commission (2002) Timor Leste National Development Plan 2Economic Bulletin of the Banking and Payments Authority, October 2004. 3Thus an ADB estimate put the rate of growth at an annual 2.6% in the period 1990-95. 16 2 Structural issues A distinguishing feature of the economic structure is the sectoral distribution of the labor force. The GoTL's proposed Private Sector Investment Program (PSD SIP)4 operates on a working assumption that perhaps 80% of the labor force is in the farm sector, and around 20% in the non-farm sector, including Government. The share of agriculture in non-oil and gas GDP was 29% in 2003, while private non-farm output accounted for 34% and government provided about 37%. Some 40,000 workers were employed in the private sector, primarily in services, in 2001. The PSD SIP projects an annual labor force increase of around 15,000 in the decade to 20155. It appears that with civil service numbers unlikely to grow significantly, the burden of providing off-farm wage employment will fall on the private sector. Given that sector's small base, even quite rapid growth would fail to provide enough places, should any substantial fraction of the annual increment in the labor force seek wage employment. This unpromising outlook obliges GoTL to consider the prospects for continued absorption of labor within the farm sector, aside from considering how to maximize private sector employment creation. Within the farm sector, a marked structural characteristic is the continuing importance of subsistence agriculture, in which the household is the primary unit of both production and consumption. This means that much output is consumed within households and does not enter into monetary exchange. Similarly, within the private non-farm sector the household is again the dominant unit of production, by number if not by value of output. Most enterprises, whether formal or informal, are very small and most use the services of unpaid (or at least unwaged) family labor. Thus, in agriculture, a 2001 `Suco Survey' produced estimates that, of some 140,000 farm households, about 57,000 produced no saleable surplus and earned no non-farm income. Aside from money transfers, these people operated substantially outside the monetized economy. Some 70,000 households were said to produce some marketable produce, mainly rice and coffee, but were also in varying degrees dependent on subsistence production for food and shelter6. Beyond agriculture, data for the non-farm sector are also limited. It is difficult to quantify numbers of enterprises and the extent of their operations. Because of the generally very small scale and embryonic character of most private sector operations, `informal' enterprises appear to form a major component of the whole. However the distinction between formal and informal enterprises, commonly employed to map private sector operations, is more problematic in Timor-Leste. Most international studies, and ILO 4RDTL, Private Sector Development: Priorities and Proposed Sector Investment Program, April 2005. 5But note higher IMF figure in art IV report of June 2005. This cites a current unemployment rate of about 20%, puts annual population growth of 3%, claims a population structure with 60% aged 16 or under, and, on the assumption of an increase in the participation rate to `a level common in low income countries', estimates that about 20,000 new entrants will soon be added to the work force annually. 6The numbers are cited in RDTL, footnote 4, above. These two categories sum to only 127,000 farm households of the estimated 140,000. While this discrepancy is unexplained, the value of the data lies in their showing the substantial importance of subsistence activities. 17 practice, equate the compliance of an enterprise with regulations, including business registration, with `formality'. But in Timor-Leste, while business registrations give some guide to the extent of formal enterprise activity, even many registered entities exhibit `informal' characteristics, in terms of scale, use of family labor, home-based operation, technologies employed and other criteria. Thus, while it seems reasonable to assume all `formal' enterprises are registered, not all registered entities in Timor-Leste can be regarded as formal. Moreover, data for non-farm enterprises which are both informal and unregistered are fragmentary. Thus the GoTL's private sector investment program was forced to rely largely on anecdotal evidence and observation to conclude that there is `significant' informal business activity in Dili and other towns, but it was unable to quantify it7. According to the PSD SIP, some 10,800 business licenses were issued to the end of 2004, about 30% of these to companies and 70% to individuals. Many or even most of the latter appear to have been household-based enterprises, in many of which use is made of unwaged family labor. This appears to confirm that business registration is not confined to `formal' enterprises but also overlaps the informal; registration cannot be taken as a proxy for formal status. By location, it appears that more than 80% of business registrations have occurred in Dili. By sector, around 75% of registrations appear to have been for service industries, of which 58% were for trading. Construction accounted for 14% overall and manufacturing and processing for less than 10%. The rate of attrition among registered business entities appears to have been high. Only 13.5% of non-company licenses falling due in 2004 were renewed, together with some 40% of company-held licenses. Thus at end-2004 valid registrations were held for only about 1,800 companies and 2,760 other entities. However, non-renewal of registration may not mean the failure of the enterprises concerned. Indeed, given their typically very small scale, it would make sense for some registered enterprises to avoid the relatively high overhead costs of re-registration if follow-up by authorities is not strict. However, while some viable enterprises might decide not to re-register, a high rate of business failure, typical of the micro- and small business sector everywhere, would likely have accounted for much of the attrition. A more recent survey of urban enterprises8, made available in draft, attempted to enumerate all registered business enterprises in the towns of Dili, Baucau and Maliana. Although many registered entities were found to be inactive, the survey was able to enumerate 494 such establishments. The survey also examined a representative sample of 552 informal establishments. Overall, urban enterprises in Timor-Leste, including the registered ones, are confirmed as being very small. However, numbers of workers employed do provide some guide to the formal/informal divide. Among formal (i.e., registered) enterprises, while more than 70% have fewer than 10 employees, almost 20% have from 10 to 20 workers. 7RDTL, Private Sector Development: Priorities and Proposed Sector Investment Program, April 2005. 8World Bank, Timor-Leste Urban Labor Market Survey, 2005 18 Among informal enterprises (i.e., unregistered), slightly more than 50% have only a single worker, while almost all of the remainder employ fewer than 10 workers. The distinction however is between the merely small and the very small; thus more than 90% of `formal' entities have fewer than 20 workers. The survey concluded that `even most formal enterprises would conform to the definition of micro enterprises and some would be considered small enterprises. Only a very small proportion is large or even medium enterprises'. Given this evidence of the typically very small scale of both formal and informal enterprises, it may be useful to assign some scale, however arbitrary, to the terms `large', `medium', small' and `micro' in the context of Timor-Leste. Employment of a hundred or more workers could be taken as the measure for the relatively few enterprises in the `large' category, with 20 to 99 for `medium' and 6-19 for `small'. Micro-enterprises could be defined as those with from 1 to 5 workers. However it is important to stress that most enterprises are likely to be close to the bottom of the range in each case. Returning to the recent urban survey cited above, informal (unregistered) enterprises proved to be almost exclusively household ventures; a majority of workers were family members, and many or most of these were either unpaid or paid in kind. This suggests a possible source of resilience for these firms, that they are not impacted negatively by the comparatively high wages paid by many formal enterprises9. In terms of scale, it appears that about half of informal enterprises fell below an annual turnover of less than $1000. But even in the formal (registered) sector more than 15% of entities had turnover below this very modest level. Urban enterprises, both formal and informal, were found by the survey to be predominantly engaged in trading. According to the survey findings, there was `a surprising degree of homogeneity ...with trading dominating every category of enterprise ­ from the micro to the large'. This generalization is even more true for the informal sector, where 93% of informal entities were found to be in trading and petty maintenance services (auto, bicycle and small machinery repair). The dominance of trading activities in the informal sector opens opportunities for women, with some 43% of entities being female-owned. By comparison, the figure for formal enterprises is only 16%. Low levels of startup capital also categorize both kinds of enterprise, with about a quarter of formal, but fully 90% of informal, enterprises having startup capital of less than a hundred dollars ($100). The survey found that `only about 7% of either category use non-personal or [non-]family resources as startup capital', suggesting extremely limited resort to institutional sources of finance. The importance of such informal sources of finance is discussed in some detail below, in section 10. Other survey data indicate that entrepreneurs in both the formal and informal economies believe that financial access is the most binding constraint on the capacity of their businesses to grow. This suggests a high degree of `financial exclusion' in the sense that enterprises have very limited access to formal financial services. These findings are central to the purpose of this report, and will be discussed at length below. 9Although the corollary of this is that their expansion beyond the family-level of operation might be impacted by the challenge of high wage levels in the formal sector. 19 A particular feature of informal enterprises identified by the survey (and confirmed to some extent by field visits) was their `stagnant' character, said to be due to the weakness of market demand for goods and services produced by them. Another survey, of furniture-making enterprises10, described a situation where micro-manufacturers produce only on demand, calling upon family and other workers as required. In the meantime these workers resort to other activities, a survival mechanism that helps to account for the resilience of this form of industrial organization. The stagnant condition of many micro enterprises in Timor-Leste has implications for the meaning of `private sector development' in Timor-Leste, discussed in section 3(i), below. 3 Conceptual and definitional issues The meaning of `private sector development' in Timor-Leste As noted in section 2, above, the economy of Timor-Leste is divided between a farm sector in which as many of 80% of workers are still found, with most of these still dependent on subsistence production, and a non-farm sector in which micro- and small enterprises are a substantial majority. Most urban enterprises operate in an informal, or unregistered, environment and in both the farm and non-farm sectors the household is the basic unit of economic activity. It is also necessary to remember the extent of destruction and social disruption resulting from the events of 1999. The painful process of rebuilding physical infrastructure, as well as establishing the administrative, legal and policy frameworks for economic activity in an independent state, is continuing. For all these reasons there are severe limits on the capacity of the non-farm sector to offer employment to job-seekers as their numbers grow. Given these circumstances, it is pertinent to ask the question: what is the meaning of `private sector development' in Timor-Leste? With such a large subsistence element remaining in the rural economy, it seems most appropriate to commence private sector development there. This would be done by managing the transition of farm households from subsistence to monetized production, and by encouraging them to produce marketable surpluses. Such surpluses would require increasing levels of agricultural productivity, per unit of both land and labor. Fortunately, a self-contained rural economy producing food merely to the limit of its own subsistence requirements, as in Timor-Leste, will normally have underused resources of labor and land. Given the stimulus of market demand, these can be brought into production to increase output and to diversify into higher value crops. These may be commodities such as coffee, vanilla and high-end coconut products, or they may be foodstuffs for domestic consumption. While the former will support export expansion, the latter can help to improve nutritional standards and consumer welfare more generally, through greater diversity of food supply. 10"Challenges for Business in East Timor, A firm Level Analysis of the Furniture Industry", Timor Institute for Development Studies (TIDS), 2004. 20 This diversification and monetization of farm output is what the Timor-Leste Poverty Policy Study11 described as a `crucial intervention' for raising rural incomes. The study identified crucial interventions for raising rural incomes. Among these, one of the most potent would be action to encourage the heads of rural agricultural households to shift out of full-time agriculture and to become produce traders. Similarly, the PSD SIP pointed to the need for `further commercialization of agriculture and the production of marketable surpluses for domestic and export markets'12. Complementary measures would include the establishment or rehabilitation of periodic markets in rural communities, the improvement of rural road networks, private investment in transport and agricultural equipment and measures to eliminate `financial exclusion' in rural communities. Access to financial services (or `financial inclusion') would bring benefits to the embryonic rural private sector. Principal among these would be the potential of financial services to support petty trading and the supply of formal sector, manufactured, goods to a rural economy making the transition from subsistence. Specialized traders underpin the operations of local markets in villages and sub-district centers. Most other market traders in these centers are members of farm households. They come to sell small amounts of seasonal produce, but also buy consumer goods produced in the formal sector. In Timor- Leste they bring produce to market to gain access to cheap and appropriate Indonesian consumer goods. There is evidence to suggest that in districts where such goods are more readily available the level of informal monetized economic activity is relatively high13. The availability of consumer goods, in marketplaces and village `kiosks', stimulates production and marketing of subsistence foodstuffs and other, higher-value, crops. This in turn improves nutritional standards, especially of wage- and salary-earners, most of whom are in urban areas. As one benefit of a well-managed transition from subsistence, the farm sector would offer higher living standards and better income-generating opportunities to the young, retaining (or at least retarding the departure of) many of those who might otherwise seek income and employment elsewhere. While the transfer of labor from agriculture to industry and services is a necessary structural change, the present imbalance in numbers between the farm and non-farm sectors makes it essential to maintain the capacity of the farm sector to absorb labour. Hence, while it is necessary to strengthen the capacity of the non-farm economy to generate opportunities, the parallel improvement of economic prospects for the young in rural communities is necessary. It is more accurate to describe this as facilitating rural income-generation, rather than as `job'-creation. Policies in the non-farm sector should include those directed to increasing the productivity of household-based economic activities and improving their linkages with 11World Bank (2003), Timor-Leste, Poverty in a New Nation: Analysis for Action. Dili. 12RDTL, Private Sector Development: Priorities and Proposed Sector Investment Program, 2005, p. vi. 13Thus the MFI Moris Rasik reports that its mean loan sizes in a district where the border is `porous' are double those in another where border control has choked off the direct supply of Indonesian consumer goods. Personal communication, Ms Helen Todd. 21 the formal economy. Such an approach recognizes the importance of household economic units as sources of income for urban populations, without neglecting measures to stimulate medium- and larger-scale enterprise. Here again, the early focus of policy should be on income-generation rather than on `job'-creation. The role of financial sector policy would be to increase the financial inclusion of non-farm households, their access to financial services, as a means of increasing incomes, increasing their welfare as consumers and facilitating their more effective linkage with the formal economy. A policy approach that devalues the informal sector and `income-generation' (either explicitly or implicitly) while focusing exclusively on `jobs' fails to acknowledge certain current economic realities of Timor-Leste. These are that, for the many rural households living in a condition close to pure subsistence, the opportunity to commence activities that generate cash income would mark an advance. Another reality is that many urban households accommodate young people for whom formal employment opportunities (or `jobs') can only arise as the result of very high and sustained rates of GDP growth. For these young urban people, participation in a more dynamic informal sector offering income-generating opportunities is a viable, even if second-best, solution. Their participation would contribute to greater equity, supporting social and political stability. Together with an increased marketable surplus from the subsistence economy, it would increase the supply of many goods and services demanded by low-income urban dwellers, thus increasing their welfare as consumers. The informal sector and wage-cost disadvantage Apart from benefits accruing to individuals and consumers, greater aggregate supply and diversity of informal goods and services would support macro-policies of wage restraint, designed to reduce the country's relative wage-cost disadvantage. For reasons of recent history, Timor-Leste finds itself saddled with an uncompetitive wage level. A vigorous informal sector could play a role in diminishing that wage-cost disadvantage over time. Policy should aim to increase the share of informal sector goods and services in the consumption baskets of urban wage earners14. This would complement a policy of wage restraint, aimed at allowing relative wages to settle back to more realistic levels. Since any attempt to introduce formal sector wage-reductions by fiat would be politically- difficult, an alternative would be for Government to aim to maintain relative domestic price stability while nominal wages either remain unchanged or decline under the influence of market forces15. `Dollarization' will help maintain price stability, but fiscal 14It might be more accurate to set the goal as being to restore the share of informal sector output in consumption. The PSD SIP cites an estimate that 86% of employment in 1998 was informal, involving some 316,000 persons, although this included workers in agricultural subsistence (RDTL, 2005, 4). Much urban informal sector activity in Indonesian times was conducted by migrants, leaving scope for `Timorization' of the sector as an element in its revival. 15That such forces can operate is already apparent. The survey of urban enterprises quoted extensively above in section 2 (see footnote 8) suggested that real wages paid by formal enterprises had declined in urban areas over the period 2001-2004, with the deflation of the `bubble economy' due to declining 22 discipline will be necessary to achieve inflation below levels occurring in neighboring countries, Indonesia in particular. Fortunately, Timor-Leste's urban informal economy is not subject to labor-cost disadvantage to the same degree as its formal counterpart16. With a more vigorous urban informal sector, Timor-Leste could secure an increasing aggregate market supply of goods and services appropriate to the consumption baskets of urban wage earners, derived from lower-cost sources. This would tend to reduce pressure for wage increases, and the formal sector's relative wage-cost disadvantage would diminish over time. An increased supply and diversity of foodstuffs and other farm produce from an increasingly monetized farm sector would reinforce this favorable trend. These benefits suggest the need for a policy environment supporting urban informal sector dynamism and the rapid transition from subsistence to monetization in agriculture. Such policies would include measures facilitating access to financial services for households, which are the backbone of the informal and subsistence economies. The next section introduces the discussion of financial services, commencing with a broad overview of the financial system. 4 Broad Outline of the Financial Sector With an economy marked by limited monetization, a narrow productive base and very low per capita incomes, Timor-Leste has a shallow and undiversified financial system. This system has achieved very limited outreach to the general population. The formal system consists essentially of the banking system: three commercial banks, each an overseas branch of a foreign institution, and one local institution operating with a limited banking license. Each of these serves distinct but limited market niches within the broader population. Regulatory and supervisory functions for the banks are performed by the Banking and Payments Authority (the BPA). As `proto'-central bank of Timor-Leste, the BPA has a legislative mandate for its operations17. Its primary objective is macroeconomic, that of achieving and maintaining domestic price stability. In addition it has responsibility for managing the monetary and exchange rate regime. However the second of these responsibilities is redundant so long as the dollarized currency regime is in place, while responsibility for control of inflation rests primarily with fiscal policy. Pending any decision to adopt a national currency, BPA's macroeconomic responsibilities are more prospective than actual. The legislation also gives BPA certain financial system-building responsibilities, and these are of primary importance under present circumstances. They relate to promoting the stability, liquidity and solvency of a market-based banking and financial system, numbers of international personnel from the peak of 2001 and associated reductions in formal wage employment opportunities for Timorese. 16This is true even if informal sector wage levels in Timor-Leste are relatively high, due to the influence of formal sector wages, because of the importance of unwaged family help in informal sector establishments. 17Most importantly, UNTAET Regulation 2001/30 on the Banking and Payments Authority of East Timor, UNTAET Regulation 2000/8 on Bank Licensing and Supervision. 23 together with a safe, sound and efficient system of payments. BPA currently has departments for financial system supervision and for payments and banking operations, and has prepared a corporate plan setting out the road map for its further institutional development over the period to 200718. Beyond the legislation, government aspirations for the BPA are spelt out in a National Development Plan published in 200219. This calls for it to become `a catalyst for significant growth of savings and credit', and to `promote more competition and increasing numbers and types of private banks and financial institutions (including widening rural presence by private banks and the BPA). Specifically, BPA was charged to `support donors and NGOs to develop micro savings and credit schemes especially in rural areas' and to `develop regulations and capacities to promote private superannuation and other savings schemes (including increased saver education and enhanced security for savings)'. Thus the BPA's responsibility to nurture the development of microfinance is explicit, together with its obligation to build national savings capacity, including in rural areas. There are no non-bank financial institutions subject to BPA supervision and there is no insurance company serving Timor-Leste, nor any leasing finance entities. Beyond the very limited set of formal, regulated, institutions, there are a number of microfinance institutions (MFIs), some savings and loan cooperatives, and some pawnshops. While MFIs are of limited significance in dollar terms, they have nevertheless succeeded in reaching a small but significant proportion of poor households. There is also, as might be expected under the circumstances, a range of informal mechanisms in operation to provide financial services. Each of these formal and informal institutions and mechanisms is to be discussed in some detail in later sections of this report. With no restriction on cross-border currency movements it is not possible for the BPA to know the value of currency in circulation, hence not possible for it to calculate standard measures of financial depth. One available ratio, domestic credit provided by the banking sector as a proportion of non-oil GDP, shows some recovery of depth occurring in the financial system since the destruction of the banks in 1999. Another ratio, deposits with the banking system as a proportion of non-oil GDP, has some explanatory power in charting the relationship between loans and deposits. It suggests improving levels of intermediation occurring within the system. Table 1 reports changes in these measures in recent years. These suggest that considerable improvements in financial depth and the extent of financial intermediation have occurred during the recovery since 1999. Table 1 Timor-Leste: some available indicators of financial depth and intermediation Measure 12/2000 12/2001 12/2002 12/2003 12/2004 03/2005 private sector credit as 0.1 0.8 1.5 6.6 20.8 21.6 18BPA (2004) Corporate Plan 2004-2007. 19Planning Commission (2002) Timor Leste National Development Plan, pp65ff. 24 % of non-oil GDP bank deposits as % of non-oil GDP 6.3 13.9 15.9 21.5 24.3 24.8 ratio, credit to deposits 2% 6% 9% 31% 86% 87% Source: calculated from IMF and BPA data. The volume of bank credit has increased substantially since the re-establishment of banking activity in 2000, reaching a level equivalent to 21.6% of non-oil GDP by the end of the first quarter of 2005. The GoTL has not borrowed from the banking system and the data in table 1 relate to credit extended to the private sector and for personal consumption. While international comparisons can mislead, due to differences in the relative importance of non-bank lending institutions, it is worth noting that the proportion for all `low income' countries in 2003 was 45.3%, or roughly double the current level of Timor- Leste20. It is also noteworthy that, in aggregate terms, bank lending is financed entirely by domestic deposits. Indeed, until a spectacular run-up in bank lending from late 2003 the banking system was in a position to deploy substantial liquidity offshore. While the positions of individual banks differ, the system as a whole is now intermediating 87% of deposits to domestic lending. The commercial banks are discussed in some detail in section 6, below, but broadly speaking it appears that the system may be running up against a constraint on the availability of loanable funds. To consider the implications of this it will be useful to consider the situation of the banking sector in the former Indonesian province of East Timor. Since 1999, Timor-Leste has gone from being a peripheral province within a larger national economy to become a national entity in its own right. As one of the poorest provinces of Indonesia up to 1999, East Timor received substantial direct grants from Jakarta. In the mid-1990s these were around $150m annually, or some 60% of provincial GDP. Thus government expenditure, rather than private investment, drove the Timorese economy. 21Reflecting this reality, the provincial banking sector displayed a pattern typical of peripheral regions, with savings mobilized by the banks substantially exceeding bank lending. The imbalance was quite extreme: a factor of about four to one. Thus in March 1999 savings balances totalled IR594bn ($68.4m) while credit outstanding was IR158bn ($18.2m). Private savings mobilization greatly exceeded the opportunities available locally for private investment, leading to an outflow of private resources from the province22. However, in net terms and because of central government grants, external savings amounted to 25% of provincial GDP. In independent Timor-Leste, as shown above in table 1, the performance of the banking sector in intermediating savings to credit has improved rapidly, to the point where, in the 20World Bank, 2005 World Development Indicators. See table 5.5, "Financial depth and efficiency' where ratios for countries with recent experience of conflict include Rwanda (13.9%), Sierra Leone (47.3%), , Sri Lanka (42.1%), Sudan (11.9%). 21Hal Hill and Joao Saldanha, `The key issues' in Hill and Saldanha (eds) East Timor: development challenges for the world's newest nation. Singapore: Institute of Southeast Asian Studies, Asia-Pacific Press, chapter 1. 22Bank BRI was an exception to the overall pattern. Its deposits totalled IR29.8bn, as against loans outstanding of IR25.4bn, at end-1998. 25 March quarter of 2005, bank lending represented 87% of aggregate bank deposits. A decline in the rate of growth of bank lending, evident in that quarter, is reported in section 6, below. This may have reflected an emerging constraint on loanable funds. The principal lending institution reported in mid-2005 that it was holding a substantial backlog of loan applications. Government policy for private sector development is to rely increasingly on organic processes of financial intermediation while eschewing official lending schemes in the future. Under such circumstances, the overall volume of credit may be constrained unless domestic savings mobilization increases or foreign-owned banks can resort to head office transfers. But previous experience in Indonesian East Timor suggests that aggregate savings could increase substantially above the 2005 level, if supported by an adequate infrastructure of deposit-taking institutions available widely to the population. In general, financial sector development in Timor-Leste is subject to a number of constraints, to which reference will be made at appropriate points in this report. For example, financial institutions suffer from the human resource deficiencies which affect all areas of national life, given that few Timorese occupied positions at even middle-level in the Indonesian financial system. System development is also impeded by a generally low level of `financial literacy' among Timorese. This deficiency can be seen as an aspect of a more general weakness in entrepreneurial capacity and experience. It has implications for the demand for financial services, and for private sector development. But it also has implications for the welfare of the mass of Timorese, as consumers. Further, ancillary services necessary for financial institutions are lacking. Finally, crucial deficiencies exist in the legal and regulatory environment for financial services. Turning to the last of these constraints, a facilitating legal and regulatory environment for financial service provision has been slow to emerge. Where legislation exists, its administration is often inefficient, while a weak judiciary creates uncertainty as to the resolution of business disputes and enforcement. An insurance law (which will reduce risk for lenders, among others) has recently been approved but awaits implementing regulations and the issuance of licenses to insurers. Continuing concerns include the enactment of legislation for bankruptcy, to secure creditor rights, which could be regarded as urgent in light of recent rapid increases in bank lending23. This deficiency is a factor behind the pronounced bias in favor of `payroll' lending observable in the loan books of commercial banks and the MFITL. Land legislation and land titling, facilitating the use of land as collateral, are delayed because of technical problems that will require some years to overcome. Many of the records relating to particular parcels of land have been destroyed, and different systems for land titling were in place during Portuguese colonization and Indonesian occupation and competing claims for land and property often exist. There is not yet a law for equipment leasing, which would assist in overcoming collateral problems, though 23 The 2005 IMF Country Report notes that `recent rapid increase in bank credit to the private sector, including small businesses, strengthen[s] the case for faster progress in establishing a legal framework that defines the rights of different categories of creditors in the case of insolvency'. 26 legislation for private leasing of property has been enacted. However, where deficiencies are manifold, it is necessary to prioritize. Most urgent is the establishment of a basic legal framework for property and land, in which rights to property are recognized and enforced. This would provide an adequate interim basis for growth in economic and financial activity, pending an incremental process by which specific legislative gaps are filled and more intractable deficiencies are addressed. In terms of specific gaps, completion of the implementing regulations for insurance and enactment of equipment leasing legislation might also be given priority. 5 The demand for financial services It is almost a truism to say that there is substantial unsatisfied demand for financial services in Timor-Leste, and the perception has stimulated attempts to estimate the size of the market. A difficulty associated with such calculations is whether they are concerned with market demand, or with some notion of the need for financial services. Analysis of market demand takes into account constraints inhibiting borrowers. Many potential borrowers are confronted by obstacles to obtaining credit or to using it effectively. These include a limited range of potential investments, limited market knowledge, or inability to take up opportunities because of `lack of adequate productive skills'24. Some potential borrowers may be inhibited by lack of mechanisms to mitigate risk; the absence of commercial insurance in Timor-Leste is the most obvious example. High real interest rates reduce the profitability of investments, causing some to be ruled out. High transaction costs imposed by the regulatory environment and the requirements of financial institutions will also reduce the profitability of investment. Lack of acceptable collateral or the inability to employ assets as collateral due to administrative shortcomings (such as lack of land titling or other required documentation) are binding constraints on many potential borrowers. With regard to the demand for deposit services, constraints also apply to potential savers. Depositors in Timor-Leste appear to be responsive to changes in the level of real interest rates on savings. Their behaviour is influenced by transaction costs imposed by banks and by travel and time costs incurred in accessing services. Further, inflationary expectations and savers' confidence in financial institutions influence their willingness to hold assets in the form of deposits. Recent memories of destruction and loss of financial assets, followed by inflation in the early post-conflict period, are negative factors. The capacity of Government to rebuild confidence in institutions and to maintain a low-inflation environment will be crucial to the necessary growth of national savings capacity. The population census of 2004 enumerated almost 200,000 households with mean household size of 4.7 persons. On that basis, it would be appropriate to plan for the extension of deposit services to the great majority of these households. Allowing for 24As an ADB evaluation suggested, in explanation of limited lending achieved by MFITL. See ADB Aide Memoire, Grant 8186 TIM (TF) Microfinance Development Project, Project Review Mission (12/2004) 27 some multiple account-holding within households, a total of 200,000 deposit accounts would appear realistic if the financial infrastructure were available and the requirements (cheap, accessible, safe, convenient) were met. Initially it would be sensible to aim at extending services to the 9 remaining district capitals and 65 sub-district towns that remain largely unserved, as well as to their immediate hinterlands. The actual number of deposit accounts in all classes of institution was somewhat less than 60,000 in late 2004 and a target of 200,000 accounts in all classes of institutions should be achievable over a five year period, by 2010. A study of the market for microfinance services was done for AMFITIL, the umbrella body for MFIs25. This aimed to derive estimates of household numbers for the economically active, for those with access to productive resources, and for those likely to benefit from access to financial services. Of a total population of 925,000, some 435,000 (or 47%) were recorded as economically active and constituted the `maximum market' for financial services. After excluding `non-poor' households and waged workers, this exercise produced a number close to 60,000 households, with some 275,000 people. This sub-population was thought likely to need access to micro-financial services, and to be capable of benefiting from them. A private sector review conducted for the IFC26 considered financial sector constraints on private sector development. It adopted the AMFITIL `maximum market' figure of 435,000 persons as an estimate of the `customer base' for financial services. This second study suggested that with a current figure of 27, 000 borrowers only about 6% of that base were being served27. While estimates of demand for deposit services based on population data tend to be credible, estimates of demand for credit are subject to serious reservationsThis is because international experience suggests the great majority of households will make use of deposit services if they are cheap, accessible, safe and convenient. This suggests the need for further investment in banking and other financial infrastructure across the country. The facilities of all financial service providers, including regulated banks and non-bank institutions such as the MFIs and CUs, would need to be employed and expanded to meet a target of 200,000 deposit accounts by 2010. It may be possible for banks to apply international experience with cellphone-based systems that would allow them to open new transaction points away from their branches. These could involve collaboration with retailers and others who already have facilities to handle cash safely. Even beyond the 25Lene Hansen and Novanto Agus (2005), `Financial services assessment in Timor-Leste: Final report'. Dili: Association of Microfinance Institutions in East Timor. 26Unpublished IFC commissioned ECON Analysis, Report No. 2005-022, Assessment of private sector opportunities in Timor-Leste. 27A third study in Timor-Leste, by consulting firm DAI for USAID, uses the Hansen and Agus data to posit the existence of a `second-tier' market for microfinance services. This refers to an intermediate-level group of borrowers, positioned between the formal business market and the microfinance market, who are estimated to number between 5,800 and 8,000 clients. See Development Alternatives Inc. (2005), Report on options for financial service provision to second-tier microfinance clients in East Timor, report prepared for USAID, Dili. 28 range of cellphone coverage, offline transaction systems using handheld recording devices are being trialed for the provision of basic financial services. Dialogue between GoTL and the banks would be necessary to explore necessary incentives and regulatory implications. The pace of international innovation in this field is rapid and authorities in Timor-Leste may want to keep informed of progress. Another possibility may be found in a proposal, in the Sector Investment Program for private sector development, that a planned national network of postal service points should also be equipped to conduct financial sector transactions28. The proposal refers to `postal banking services', to include, inter alia, deposit accounts and payments facilities. This is almost certainly too ambitious given the prevailing low levels of financial literacy in Timor-Leste, including among lower-level public officials. However, if these were simply to be transaction centers, rather than a government postal bank, the proposal would merit some pilot trials, perhaps in conjunction with efforts at introducing IT-based technology, as discussed above. On the other hand the approach to expansion of credit services would best be pragmatic and incremental, rather than target-based. The objective could be to achieve the judicious expansion of credit by appropriate and sustainable institutions. There is evidence of excess demand for credit in the banking sector, to which the currently-active commercial banks will respond, while a decision by authorities to approve the entry of one or two suitably-credentialed new banks would assist by increasing competition. Such new entrants will have made feasibility studies before risking their capital, and this is preferable to target-setting by authorities. Instead, authorities may want to focus on facilitating and monitoring the progressive expansion of credit services in response to observed indicators of excess demand at the margin. Based on world-wide experience it would not be wise for Government to be directly involved in lending, either through a state-owned Development Bank or a Portal Bank. Appropriate lenders will include MFIs and CUs in addition to banks, as well as new non- bank commercial lenders. Market demand for credit will increase as GoTL initiatives designed to promote the transition from subsistence to commercialized agriculture take hold, and as policies to promote economic activity in urban areas bear fruit. The prior expansion of deposit-taking facilities, to un-reached population strata and into un-served geographic areas, will provide indicators of credit capacity for the information of lenders and the authorities. It will also create infrastructure that can be adapted progressively for lending, with the growth of market demand. Active support of MFIs and CUs by the GoTL and donors will assist in this process. 28RDTL, Private Sector Development: Priorities and Proposed Sector Investment Program, April 2005, p. 25. 29 6 Supply of financial services: the commercial banks Three full-service commercial banks operate in Timor-Leste, together with another institution, MFITL (Microfinance Institution of Timor-Leste), which has a limited banking license. This latter institution is discussed separately below, in section 8. The commercial banks are all branches of overseas entities, and operate at the interface of three banking cultures, those of Portugal, Australia and Indonesia. The branches of the Australian and New Zealand Bank (ANZ) and the Caixa Geral de Depositos (CGD)29, of Portugal, were both established in 1999-2000. Bank Mandiri, from Indonesia, opened after Timor-Leste's independence, in 2003. While ANZ is a private sector bank, both CGD and Mandiri are state-owned and it is possible to infer that their decisions to enter the Timor-Leste market involved at least some non-commercial considerations. Each of these three banks serves a fairly distinct market segment but there is some overlap, primarily at the retail level. The arrival of Bank Mandiri as a third player appears to have brought greater competition to the banking sector, both for deposits and loans. Its entry was followed by lower bank charges and the payment of interest on demand deposits, and also provided an alternative to the dominant lender. There is evidence of some reduction in lending margins. BPA data show weighted average lending rates falling by about 1% between 2003 and 2004, from 14.78% to 13.81%. The weighted interest rate on saving deposits rose from 0.18% in 2003 to 0.32% in 2004. A gross spread of rather more than 13% may not be unreasonable for Timor-Leste, under present market conditions of risk and uncertainty, and with annual inflation running at between 2% and 3%. Nonetheless there is obvious scope for further compression of margins as conditions improve and the system matures. An overview of the banking system may provide a useful picture of its early growth, which has been rapid. Table 2 presents aggregate data describing the progress of lending and deposits for the system since 2002. Deposit liabilities increased quite rapidly in the two year period to end-2004, by almost 54%, and rose by a further 3.2% in the first quarter of 2005. Lending increased almost 1300% over the two year period, though from a very small base. While growth appears to have moderated more recently, it still rose by an annualized 30% in the March quarter of 2005. There is demand for more credit, as evidenced by a backlog of loan applications held by the main lending bank. A marked increase in domestic intermediation has occurred in recent years, with the proportion of domestic deposits re-lent within the economy increasing from less than 10% at the end of 2002, to some 87% most recently. In parallel with this, and reflecting the greater capacity of the banking system to deploy its assets locally, net foreign assets of the banking system ran down steadily. The rundown began with the surge in private sector lending from late 2003, and the banks' net foreign assets were almost zero by March 2005. Table 2 29This Portuguese bank opened under the name BNU, familiar to Timorese from the colonial period and by which it is still popularly known. Subsequently the bank rebranded itself as CGD, the larger state banking entity of which BNU is now a part. 30 Domestic intermediation by regulated financial institutions ($m) Transactions 12/02 12/03 12/04 03/05 Loans and advances 5.1 22.1 70.5 75.6 to private sector Deposit liabilities 54.6 72.3 84.0 86.7 Ratio of loans and advances to deposits 9.3% 31% 84% 87% Note: includes the three commercial banks and the MFITL However these aggregate figures disguise important differences in the market positions of individual institutions. Information about particular institutions in a very small banking sector, such as Timor-Leste's, is sensitive. But so far as it is possible to judge, it appears ANZ continues to hold the bulk of its assets offshore. Mandiri appears able to deploy some liquidity outside the country while continuing to expand domestic lending. CGD, on the other hand, appears to have responded to the demand for credit by supplementing domestic resources with offshore liquidity. These contrasts reflect the different operating strategies of the three banks, which are discussed in greater detail later in this section. The growth of deposits and their changing composition are detailed in table 3, which suggests some interesting developments. Growth has been rapid and the relative importance of savings accounts has increased, with savings balances accounting most recently for about a third of total deposits. The growth in demand for term deposits has also been healthy, although their volume is still small. If the savings account is the product of choice for lower-income people, while demand and term deposits are mostly held by enterprises and higher income earners (including resident foreigners), the growth in absolute level and relative importance of savings deposits suggests burgeoning demand for this product. The retail market still consists largely of urban wage earners and the banking system has not so far tested the demand for savings deposit services among people outside formal employment or in rural areas. Table 3 Growth of domestic deposits in regulated financial institutions ($m) 12/00 12/02 12/03 12/04 03/05 Total deposits 20.0 54.6 72.3 84.0 86.7 Demand 19.3 39.1 40.1 49.4 52.4 deposits Savings 0.0 10.2 22.6 27.8 26.7 deposits Time deposits 0.7 5.3 9.6 6.8 7.5 Savings/total -- 18.7 31.3 33.1 30.8 Note: includes the three commercial banks and the MFITL Source, IMF and BPA In terms of strategy and market positioning, ANZ secured the accounts of international agencies operating in Timor-Leste. It has served as the principal provider of banking 31 services to the international community as well as servicing trade with Australia. After some initial and not entirely successful experience of lending, ANZ essentially withdrew from credit operations and has deployed its substantial liquidity offshore. ANZ management cites difficulties with collateral documentation, the absence of notary services and credible financial statements, and other uncertainties in the current legal environment as constraints on lending. The bank appears to have experienced difficulty in applying or adapting its home country lending criteria to the circumstances of Timor- Leste. However, with diminishing numbers of foreign personnel and declining funds transfers by its primary institutional clients, ANZ faces a challenge. While operating from a single office in Dili, it has introduced ATM services for local and international clients and electronic funds transfer services (EFTPOS) with participating merchants in Dili. Bank CGD is the industry leader in terms of assets, customer numbers and branches. It has an office in Dili, further branches in Baucau,Gleno, Maliana, Oecusse and Viqueque and plans to open shortly in Suai. At least some of these might be regarded as longer- term breakeven propositions, but the bank appears to have an aggressive and nationally- focused expansion strategy. This is also reflected in recent rapid lending growth. Although net data for CGD are difficult to obtain, new lending approvals in 2002, 2003 and 2004 appear to have been $4.2m, $20.1m and $67.9m, respectively, and to have amounted cumulatively to some $105.7m up to May 200530. These numbers appear consistent with a figure for loans outstanding (ie, net of repayments) of $59.6m as at September 2004, cited by several studies31. CGD has positioned itself well to benefit from strong links with the GoTL. It is authorized to receive tax and excise payments on behalf of government and has arranged direct salary payment to the accounts of civil servants. This overcomes uncertainties associated with collateral and enforcement, and facilitates `payroll' borrowing, mostly for consumer durables, vehicles and housing rehabilitation. Payroll arrangements are thought to have been a major factor in the rapid expansion of credit. CGD is also active in trade financing with Europe and especially Portugal. CGD's initial lending was done as the executing bank for emergency credits disbursed under the first phase of the World Bank's Small Enterprise Project (SEP I), from late 2000. Repayment of these credits was slow (standing at just over 75% in mid-2005) and very few of the borrowers from SEP I remained as clients of CGD. This experience is discussed at greater length in section 10, below. Bank Mandiri's market niche is based primarily, though not exclusively, on financing trade with Indonesia and servicing Indonesian business interests in Timor-Leste. It is often able to lend against collateral located in Indonesia, on contracts enforceable at law in that country. Operating from a single office in Dili, its lending is almost entirely commercial and mean loan sizes are comparatively large. The bank provides scarcely any consumer credit because of concerns about collateral and contract enforcement. It has similar reservations about micro- and smaller business lending. However it has extended a line of credit to the largest Timorese MFI, and has also in the past provided bridging 30These numbers are approximate, having been taken by the writer from graphs obtained from CGD. 31Hansen and Agus, DAI and ECON Analysis, all cited above. 32 credit to CCT, a coffee producers' cooperative. The bank's competitive approach to savings mobilization has won it market share in retail banking and enabled it to finance strong lending growth in economically important sectors. In terms of bank lending, ANZ is effectively a non-lending institution, while Bank Mandiri extends perhaps 10-12% of bank credit, the IMfTL about 2% and CGD the balance, or more than 85%. Data presented by the BPA for the sectoral allocation of credit suffer from inconsistencies in classification, as reported by the banks. However information obtained from Bank Mandiri provides some additional insights. From inspection of BPA records for late 2004 and early 2005, it appears that overall the banking system provides negligible finance for agriculture, industry or tourism (although Mandiri allocated almost 10% of its lending to industry). In aggregate, the largest loan category appears to be construction, including the rehabilitation of housing, at around 40% of the total (but only 2.3% of Mandiri's lending), followed by `trade and finance' at around 19% (but for Mandiri, 69%). `Individual' borrowing forms some 7% or 8% of the total (Mandiri, 0.1%), while `transport, communications and services' account for around 5% (Mandiri, 15%). There is a large and unexplained residual category in the industry-wide data. This is `other' borrowing, amounting to around 27% of the total. If this could be reclassified it might change our perception of the industry-wide allocation of credit. Overall, it seems reasonable to conclude that Bank Mandiri's lending is more commercial than that of CGD, the majority lender whose numbers dominate the aggregated data. By contrast with CGD, Mandiri offers stronger support for trade, services and industry while providing negligible support for housing or personal consumption. Mandiri's total portfolio at May 2005 amounted to $8.54m, to 91 clients with a mean balance outstanding of around $93,850. By contrast, CGD approved some 6,500 loans in 2004 with a mean size of about $10,500. Rapid growth of lending by the banking system, and the extent to which it is perceived to be financing either consumption or excessive investment in taxis and small buses, has attracted some negative comment32. The IMF's 2005 Article IV consultation process produced an expression of satisfaction with the growth of intermediation, but also a warning concerning the need `to remain vigilant to possible emerging risks, particularly given the rapid growth and sector concentration of bank credit'. However, the Staff Report noted that `the banks' loan portfolios have been assessed as being of appropriate quality and in compliance with supervision regulations, including for loan provisioning'.33 The principal lending institution appears to have the full support of its parent bank and to enjoy good relations with the GoTL; moreover, much of the credit growth is of lending secured against civil service salaries, which is low risk. The authorities are aware of the need for close monitoring and adequate provisioning and appear to regard bank balance sheets as generally sound. 32See, for example the DAI report referenced in footnote 24, above. 33IMF, Democratic Republic of Timor-Leste, Staff Report for the 2005 Article IV Consultation, para 25. 33 7.Supply of financial services: potential for entry of new commercial banks The authorities stand prepared to consider applications for new banking licenses in Timor-Leste. In addition to maintaining systemic stability and safety, BPA's mandate is to oversee the development of a competitive and market-based financial sector. In principle, this suggests the desirability of having more than three full-service commercial banks, especially given the evidence of competitive benefits flowing from the entry of the third bank in 2003. The BPA has since denied several further applications for entry. These denials are said to have occurred for reasons other than the ability of the market to absorb new players. Aside from the merit of any particular application, the viability of new banks depends not only on aggregate market growth but also on whether niches exist in which new entrants could flourish. Diversification within a framework of competition is likely to be the most promising option for achieving outreach to unserved population groups, while lowering costs and improving service levels overall. These issues were canvassed in 2003, when Bank BRI conducted a study of the feasibility of re-entering the market of Timor-Leste from Indonesia. Instead, Bank Mandiri acted to open its Dili branch, while BRI hesitated and the moment for its return appeared to pass. An opportunity was lost for the re- introduction of a `micro-banking' business model radically different from those of the currently operating banks. It is worth studying the rationale advanced in 2003 for Bank BRI's re-entry, and to consider its relevance to present circumstances. At end-1998, BRI's Dili branch supervised 12 BRI Units serving some 12,000 borrowers with a mean loan size about $21034. Arrears as a percentage of loans outstanding (`portfolio status') stood at the creditable level of 3%. BRI's deposit-taking services were at least as important for microfinance as its lending. Deposits exceeded loans by a margin of about 17% at end- 1998, and included many very small accounts. In 2003, USAID financed a feasibility study which recommended the recommencement of BRI Unit operations in Timor-Leste35. This assumed BRI would offer its standard Unit services, micro-loans, deposits and funds transfers. The study presented business plans based on various scenarios and suggested BRI `might not be willing to fund the full cost of the optimal level of investment on its own'. Reasons given were the demands on BRI to finance expansion at home, the `proven returns' on such expansion (thus setting a high level of opportunity cost for BRI's own capital) and the greater costs and risks of operating in Timor-Leste. In hindsight, these reservations appear justified, in that BRI's ambitious domestic expansion program is currently constrained by resources, despite the profit potential of new branches and Units. BRI itself is no longer a fully-owned state bank and is presumably even more focused on shareholder value now than in 2003. 34Calculated at a rate of Rp10,000 per dollar. 35Re-establishing microfinance services in East Timor: Back to BRI? (A study conducted by Bank Rakyat Indonesia, Harvard University Centre for Business and Government and the East Timor Study Group, with support of USAID, May 2003). 34 The 2003 study explored an option under which BRI would fund quite limited expansion from its own resources, under the `basic' or neutral operating cost scenario. The study also explored more expansive scenarios, with BRI assisted by third party capital. Potential sites for additional BRI units were identified in Maliana and Baucau, apart from in and around Dili itself. The optimal outcome was shown to require an upfront external contribution of $0.45 million. This would enable lending totaling about $4.9 million by the end of year 4, to some 2,100 micro-entrepreneurs and around 6,400 wage and salary- earners. The BRI micro-banking business model relies on a high level of net savings mobilization by the Units, achieved by paying positive real rates of interest, with surplus liquidity transferred internally to fund BRI's retail and corporate lending activities. By the end of year 4, the feasibility study showed BRI managing almost 22,000 deposit accounts, with funds mobilized of around $8.4 million. This amount would exceed local lending by some $3.5 million. The surplus liquidity would yield returns for the branch by being deployed internally through the BRI system. Such offshore deployment of liquidity by BRI would mirror the actions of some other banks in the Timor-Leste market. Lending rates for BRI unit loans were to be set at those prevailing in Indonesia: (1.5% per month, flat, equivalent to about 27% per annum, but with rebates for timely repayment). This was considerably higher than charged by the banks in Timor-Leste at the time (up to 18% per annum). Even though BRI expected to offer current deposit rates higher than those available in 2003, from 1.5% to 2.0% per annum, with time deposit rates from 2.0% to 2.25% per annum, it was allowing itself considerably wider interest margins than those enjoyed by banks already in the market. This may be explained by the fact that the BRI Unit banking model incurs higher transaction costs per dollar loaned. BRI could have assisted low-income urban (and with expansion, rural) people by providing a safe, liquid and accessible savings service that is largely lacking at present. The deposit rates it proposed to pay, at a maximum of 2.25% per annum, are close to current inflation rates. By contrast, current commercial bank deposit rates are negative in real terms. And it is likely that BRI, given its familiarity with the Timorese context, would have succeeded in lending profitably to a class of micro- and small entrepreneurs not presently served by commercial banks. In short, BRI's re-entry would have added to the diversity of banking services in Timor-Leste and would very likely have increased the system's outreach to low-income and rural people. The market is now more competitive than at the time of the feasibility study and the scenarios for BRI's re-entry would have to be reworked. From discussions in Jakarta with BRI management, it appeared that BRI might be persuaded to revisit the issue of returning to Timor-Leste, although it is not currently on the bank's agenda. USAID funded the original BRI feasibility study in 2003. Perhaps BRI could be consulted again and third party resources found to enable the study to be reworked in the light of current conditions. 35 Other commercial interests have considered applying for entry to the banking market. As mentioned above, several applications have already been declined. Bank Danamon, an Indonesian private bank that traded in East Timor until 1999, actively canvassed the possibility of returning in 2003. But after foreign interests acquired equity in Danamon it did not proceed. A recent review of private sector opportunities, done for IFC36, reports that CGD might spin-off its Timor-Leste operations. CGD would propose to retain control of a new entity in which GoTL and IFC itself would be minority shareholders. Another possibility is that GoTL itself might seek to remedy perceived deficiencies in the availability of financial services by setting up a special purpose `development' bank. Such proposals reflect a `supply-leading' response to deficiencies of financial services provision. Such an approach would be in danger of ignoring world-wide experiences with poor performance of such banks, which have often proved subject to political interference. When the case for investing resources in a state bank is made nowadays, it is usually in the context of rehabilitating a moribund or failing enterprise to make use of sunk capital, as represented by existing infrastructure and especially by branch networks. It is often argued, in addition, that the human resources associated with such banks have much to contribute to their success, once they are restructured and revitalized. Neither of these arguments can apply in Timor-Leste, where a new state bank would have to be built from the ground up, and no reserve pool of experienced staff exists. The capital and human resources for such a bank would incur very high opportunity costs. Proponents of a state bank may point to the performance of a few distinguished state agricultural banks, including Bank BRI and BAAC of Thailand. But these successes reflect a combination of heavy investment and political will, not easily reproduced elsewhere, together with opportunities for economies of scale in countries with sizeable populations. The performance of state development banks in economies more comparable with Timor-Leste, such as the small island states of the Pacific, does not encourage confidence. Such a bank established by Government in Timor-Leste would pose challenges for governance and could encourage rent-seeking and the politicization of credit. An alternative proposal, that the state might take equity in an existing bank, raises other questions: if the investment was not necessary to establish the bank initially, what social benefits could such investment be expected to yield now? And how might the state's taking equity in one bank affect perceptions of the returns to private investment elsewhere in the sector? BRI's feasibility study implied that the option of a joint venture with Government would be unacceptable to it. No doubt that judgment reflects sober assessment of the constraints a link with Government might entail for successful commercial operations. But if BRI is the very model of a good development bank, why not simply invite it to re-establish itself in the familiar territory of Timor-Leste? 8.Supply of financial services: the Microfinance Institute of Timor-Leste 36ECON Analysis, footnote 23, above. See P. 16. 36 History and objectives The Instituicão de Microfinancas de Timor-Leste (IMfTL, the Microfinance Institute of Timor-Leste, or `the Institute') was established in May 2002 as a regulated quasi-bank. As it is part of the banking system, the Institute's operational data are reported in consolidated statements put out by the regulatory authority, the BPA. For this reason, aggregated banking data cited above, in sections 4 and 6, also include information about the IMfTL. However, discussion of the Institute has been reserved for this section of the review because its operations are both quantitatively and qualitatively distinct from those of the major banks. The IMfTL is a regulated banking institution, registered under an exempting clause of the banking legislation37. This limits its obligation to comply fully with the requirements of the law while imposing certain limitations upon it. Thus it may not accept deposits greater than $1 million in total, nor may it adopt the title of bank. It is subject to a liquidity requirement of not less than 50% of the value of deposits held. Aside from accepting deposits, the Institute is licensed by BPA to extend credit, with a minimum of 65% of portfolio to be devoted to microcredit. This establishes a particular mandate for poverty alleviation for IMfTL, to be fulfilled through the provision of microfinance services to poor and low-income households, and especially to women. This distinguishes it from other regulated financial institutions in Timor-Leste. The IMfTL is also authorized to provide payment and collection services, issue and administer current accounts and checks, and provide safekeeping services for valuables. It has a settlement account with the BPA and is a member of the inter-bank clearing system. IMfTL was established with a TFET grant of $4 million, and it has authorized share capital of $3 million, of which $2 million is paid-up. This is the minimum capital level required under banking legislation. Operating since 2002, the Institute is controlled by a not-for-profit entity, the Foundation for Poverty Reduction of East Timor. This is its interim owner, pending agreement on permanent arrangements. The Board of Trustees of the Foundation includes ex officio members from the donor community and ADB is de facto the controlling trustee. Management of the Institute is in the hands of a Board of Directors. An ADB project under the TFET grant provided assistance to establish the Institute, and while this project concluded at the end of 2004 there is a continuing program of ADB technical assistance. This has terms of reference, inter alia, to advise on a plan for the divestment or dilution of the shareholding in IMfTL with a view to determining the Institute's ultimate ownership. The GoTL is also concerned to see the issue of future responsibility for the Institute clarified and to have its outreach expanded to all districts. Issues associated with the future ownership and direction of the IMfTL are not only technical, however. They require judgments about whether, and if so how best, to maintain the Institute's original mandate for poverty reduction. They also require choice among a number of ownership options, ranging from full state ownership, to some mode of community ownership, to full privatization. 37Section 2.6 of UNTAET Regulation 2000/8 37 Size and scope of operations IMfTL operates from branches in Dili and the district centers of Gleno and Maliana. As an overview of operations, it had deposits of $0.876 million in more than 8000 accounts, with a mean size of about $100, and an outstanding loan portfolio of $1.436 million at end-April 2005. Less than 60% of that amount was disbursed via the main office in Dili, suggesting it is much less Dili-centric than the commercial banks. However, only about 87% of loan accounts, by value, were classified as `on-time' and principal overdue amounted to almost $82,000. Portfolio at risk was estimated at 7.3% overall, although this measure varied widely between branches (for instance, as high as 20% in rural Gleno) and between the several loan products offered. Among loan products the largest category is payroll lending, secured against the salary payments of civil servants. This amounted to some $1.152 million, or 80% of total loans, and was extended to 1,532 borrowers, with mean balance outstanding of $750. These loans are extended for periods of 12 to 24 months, at a monthly rate of 1.5% and with an upfront deduction of 5% from principal as an administration charge. This is equivalent to a 22.5% effective interest rate, per annum. Overdues in this category are very low, with borrowers' salaries credited direct to the Institute. However, although a substantial proportion of this clientele consists of lower-level officials located in rural centers, it is questionable how far such lending goes to satisfy the Institute's original poverty reduction objectives. Other products offered by IMfTLappear more appropriate to its mandate, if of minor importance in its overall lending. These suffer much higher delinquency rates. They are microfinance loans, including group loans with weekly repayments, extended mostly to women (some 1,250, with mean balance outstanding of $46). About 290 market vendors have daily repayment loans at an average outstanding balance of $130. In addition, there is some seasonal crop lending (more than 500 loans with mean balance around $230) and a group of somewhat larger `business' loans (180 at about $400). In general these loans all suffer relatively high delinquency rates, perhaps 25% across the board, with delinquency defined as 30 days or more overdue. Management of overdues is said to be lax. In regard to lending to the poorest, the Institute appears to have adopted the rhetoric of group organization without having internalized the rationale and methods of classic group-based lending, as practiced by the Grameen Bank and its many imitators. Similar problems were noted in connection with group lending conducted in the earlier Timor- Leste Community Empowerment Project, in the emergency period from 200038. Seasonal crop loans are another area of particular difficulty for all agricultural lending institutions, in which the IMfTL has not discovered any comparative advantage. Market vendor loans 38World Bank, Social Development Papers: conflict prevention and reconstruction, Timor-Leste: Independent Review of the Credit Component of the Community Empowerment Project, Paper No. 11, March 2004. 38 on the other hand have been shown elsewhere to perform well with good management. IMfTL needs to review its operating procedures in this area, and to raise the productivity of staff. It has commenced a collections campaign to reduce delinquency with the assistance of village officials. Mobile services operate to collect savings and payments in some areas. Interest is paid at the rate of 0.8% annually on the single deposit product and accounts may be opened with an initial deposit as low as $10 (compared with $50 in the commercial banking system). However the deposit ceiling has not helped to focus management attention on savings, other than to hold them within required limits, and while weaknesses on the lending side are manifest, the Institute's capacity to mobilize deposits to fund growth remains untested. Indeed the Institute is highly liquid, having some $1.8 million on deposit with the BPA. Viability and future status of IMfTL While there is a range of options for the future ownership of IMfTL, the choice between them cannot be regarded as a purely technical issue. The judgment whether the Institute should continue to pursue its mandate for microcredit and poverty alleviation will have implications for ownership and governance arrangements. Such a social objective seems incompatible with private, for-profit, ownership. Nonetheless, in supportive policy and regulatory environments, including the Philippines and Indonesia in the immediate neighborhood, there are numerous instances of small regulated banking institutions thriving with the support of local communities in which they are embedded. Timor-Leste and IMfTL have offered a difficult test case for the proposition that this experience can be generalized. After the departure of Indonesian citizens in 1999, few experienced bank staff were available, and in the case of the Institute it proved particularly difficult to recruit staff with backgrounds in credit management. Problems occurred with the design of credit products. The chosen method of accessing the poor, and women in particular, was the group method of organization. There appears to have been a failure to understand and employ the strengths of the group method. Indeed one IMfTL branch established in a district where an accomplished MFI was already engaged in replicating Grameen banking seems to have decided not to compete with it in group lending. Where IMfTL has tried the method, it appears to have been used as an administrative device rather than an exercise in group dynamics, and results have been disappointing. Despite these shortcomings, ADB technical assistance to the Institute has been renewed and efforts to reduce delinquency and improve operational efficiency are continuing. Payroll lending, while threatened by the growth of commercial bank lending in this market segment, contributes substantially to the Institute's sustainability and must for that reason be maintained. To fulfill its mandate for lending to the poor, IMfTL must increase its microcredit portfolio while simultaneously increasing efficiency. The ADB microfinance project had envisioned IMfTL's acting as a wholesaler to the credit union movement, but (as explained below) it proved too difficult to rehabilitate CUs. However, growth in the Institute's microcredit portfolio might be achieved indirectly by combining 39 wholesale lending to well-performed MFIs with direct retail lending. Freed of the cap on its deposit-taking, or at least allowed some greater leeway than at present, IMfTL might also offer deposit services to MFIs on behalf of their clients. The Institute has value for banking in Timor-Leste in offering the prospect of a distinct and poverty-focused model. If realized, this model would increase the reach and diversity of the system. Aside from improving the internal efficiency of its operations, the Institute would benefit from a mode of ownership and governance both consistent with its original mandate and acceptable to the BPA for a class C banking license (to permit regulatory restrictions on its deposit-taking and liquidity to be relaxed). The least-preferable option would be private, for-profit, ownership intent on gaining `back-door' entry to the commercial banking sector. The best option could involve attracting capital and expertise from the IFC and/or a private international, not-for-profit equity investor active in micro- banking. 9.Supply of financial services: the microfinance sub-sector Timor-Leste's 2002 National Development Plan (cited above in section 4) directed the BPA to `support donors and NGOs to develop micro savings and credit schemes especially in rural areas'. This was to be done as part of a broader strategy of catalyzing `significant growth of savings and credit' in the financial sector. Thus GoTL policy is to support microfinance development and it emphasizes the value of access to savings services as an element in the process. Moreover, while under the banking legislation deposit-taking from the public may only be conducted by licensed banks, the regulatory environment is relatively permissive in regard to the operations of lending entities such as microfinance institutions. The BPA is not disposed to create a regulatory environment specifically for MFIs. Instead it appears to take a non-interventionist, if broadly supportive, stance. This tolerance appears to extend to limited deposit-taking by MFIs on the basis that they can be regarded as `membership' institutions. For the purposes of this survey, the microfinance sub sector is treated as including MFIs and savings and loans cooperatives. Both of these institutional models have received assistance from external agencies and favorable attention from Government since 1999. In regard to MFIs, an important distinction is between the specialist providers of financial services and a larger group offering microfinance as part of a broader portfolio of services. In the absence of an explicit regulatory environment for microfinance, service providers have formed the Association of Microfinance Institutions in Timor-Leste (AMFITIL) to establish a common framework for standard-setting, based on international best practice principles. The Association provides a basis for self-regulation by MFIs, derived from an agreed `code of conduct'. AMFITIL has eleven member institutions of which four, including IMfTL, are specialist providers. Members undertake to report performance data to a secretariat on a regular basis, although capacity constraints prevent full compliance by all members. IMfTL occupies an intermediate position, between the world of regulated banking (where its performance data are recorded by the BPA along with the banks) and the sub-sector of MFIs (where its operations are also incorporated with AMFITIL's data for microfinance outreach). 40 Table 4 Microfinance operations of AMFITIL member institutions at end-2004 Deposits Loans Value Number Mean Value Number Mean Total, including IMfTL $1.28m 26,600 $48 $2.36m 19,200 $123 Total, excluding IMfTL $0.42m 17,100 $25 $1.074m 15,600 $ 69 Note: Data based on reporting by 8 member institutions, including all specialized MF service providers. Data provided by AMFITIL for the microfinance activities of member institutions are summarized in table 4. They indicate that at end-2004, on the basis of reports from 8 member institutions, the Association served some 26,600 savers and 19,200 borrowers (or around 17,100 and 15,600, respectively, if the clients of IMfTL are excluded). The non-regulated MFIs reporting to AMFITIL (that is, excluding IMfTL) had loans outstanding of rather more than a million dollars and held some four hundred thousand dollars of client funds, with a client base of around 17,000 persons. This may correspond to around 17,000 families served by MFIs, since the target clients of most institutions are women. Extrapolating further, an upper estimate of the numbers of family members potentially benefiting from access to microfinance services would be perhaps as many as 85,000. In relation to the estimated population living below the poverty line, perhaps 350,000 to 370,000 in 200439, this number corresponds to less than a quarter of the target group. In terms of geographic coverage, MFIs operate in 10 of the country's 13 districts. Overall, however, access to micro-financial services for the poor in Timor-Leste appears to be both restricted and shallow. MFIs held about $0.4m in savings notwithstanding the prohibition of public deposit- taking under the Banking Act. In some other jurisdictions (for example in the Philippines under the banking legislation of 2000) authorities ignore MFI deposits unless their total exceeds loans advanced (the so-called `compensating balance'). Thus Philippine MFIs have some scope for legal deposit-taking. No such latitude exists under the Banking Act in Timor-Leste, although authorities have not so far chosen to act, apparently preferring to regard MFI clients as `members' rather than as `the public'. One reason may be that a large proportion of savings held by MFIs is contractual and the scope for instability in the system is thereby reduced. Deposits are made as an element in the loan contracts between MFIs and clients; they are not available upon demand but only after loans are discharged. The largest and possibly best-performing MFI is Moris Rasik, which replicates the Grameen bank approach and has secured a line of credit from a commercial bank. With the most pronounced rural orientation of any MFI, it had some 7,000 borrowers with a mean loan balance outstanding around $70, and reported portfolio-at-risk at a creditable 3.6%. Moris Rasik had recently begun to provide larger credits up to $1,000, to persons who had successfully serviced at least four previous loans for progressively higher amounts. Some 20 such loans had been granted by mid-2005, but Moris Rasik 39The AMFITIL Financial Services Sector Assessment (UNDP and USAID, 2005) estimated the 2004 number by extrapolating from the 2001 Timor Living Standards Survey. The estimate is adequate for our purpose since broad magnitudes suffice to show that access is quite limited. 41 management stressed that its core business would continue to lie in provision of smaller loans to a low-income female constituency, with larger loans available for motivation. This MFI also serves about 7,500 savers whose accounts have a mean balance just over $20. As described above, much of this consists of so-called `compulsory' savings, made as part of the loan contract and not available upon demand. Excluding IMfTL, the MFIs reporting to AMFITIL (and including Moris Rasik) all fell well short of sustainability and still required substantial subsidy to continue operating. On an aggregated basis, the four largest operators earned only 57% of revenues required for their operational sustainability, using the standard CGAP measure40. When financing costs were added in, their revenues were reckoned by AMFITIL to be at less than 40% of the level required. All of these MFIs are fortunate to receive support from international NGOs. The Sector Investment Program for private sector development proposes assistance for microfinance. It suggests that the industry needs `institutional support for performance- based capacity building, guarantees for commercial bank loans and support for industry development among providers'. The industry itself sees capacity-building, including IT skills, as its greatest need. `Industry development' is the province of AMFITIL, which has already made progress in setting standards for self-regulation by MFIs and in reporting members' performance. Individual MFIs and their umbrella body would all benefit from such assistance. IFC's Pacific Enterprise Development Facility (PEDF) has an `Access to Finance' program scheduled, for which Timor-Leste will be eligible. Capacity-building assistance for MFIs will be designed to equip them with better credit risk management, internal controls and governance. PEDF will support vocational training programs to improve the skills base and competence of financial institutions, including MFIs, and will seek partnership with AMFITIL for these purposes. Providing guarantees for loans from commercial banks to MFIs would be designed to create linkages between them and thereby to leverage the loanable funds available to MFIs. Leaving aside the potential for moral hazard (which is not negligible, considering the immaturity of the MFI sector) the commercial rationale for seeking to induce bank/MFI linkages in this manner is not convincing. Banks tend to look to such linkages when the regulatory environment requires them to do so (most commonly when they are subject to `priority sector' credit targets) or when the success of well-performed MFIs attracts their attention and they are otherwise looking for expansion opportunities. Credit targets are part of the discredited armory of financial repression, and it would not be wise to apply them in Timor-Leste. And the difficulties Timorese MFIs are experiencing with sustainability do not make them attractive as commercial partners. Sustainability is the Achille's heel of MFIs in Timor-Leste. Given the evidence of their capacity to achieve outreach, priority in assistance should be given to institutionalizing their presence in Timor-Leste. This requires attention to sustainability, and insofar as sustainability requires more scale, then attention must also be given to increasing 40A detailed discussion of the microfinance sub-sector is contained in the AMFITIL report Financial Services Sector Assessment in Timor-Leste, noted above. 42 outreach. Remarks in section 5, above, concerning a policy of judicious expansion of credit at the margin, apply also to the MFIs by reason of their financial vulnerability. One potential source of resources for future expansion, suggested above in section 8, could be the IMfTL. The MFIs and the Institute (after its mandate for micro-lending and poverty alleviation is confirmed and its ownership and governance issues are resolved) could benefit from a linkage arrangement. Under this, IMfTL might wholesale credit to the better MFIs. This might also be extended to permit MFIs to place deposits with IMfTL on behalf of their members, while retaining `contractual' deposits. Before Indonesia's departure in 1999, East Timor had a small system of credit unions (CUs), said to include 27 institutions and about 7,500 members, and with a Credit Union Federation (CUF) at its apex. This was quite separate from the official KUD (Koperasi Unit Desa) system of production cooperatives through which many official Indonesian agricultural programs were initiated, and which no longer exists. After the events of 1999 the CU system also collapsed, due to loss of physical assets and inability to recover outstanding loans. In the period 2001-2003, the ADB project responsible for setting up MFITL also devoted attention to rehabilitating the CU movement. It found few CUs meeting project criteria and with potential for restoration. In the event it was able to assist with the physical reconstruction of four CUs and to provide technical assistance to a number of others. The project's efforts at reviving the CUF also encountered difficulties and the Federation's weakness was perceived as an obstacle to revival of the sector. For the CU movement as a whole, the ADB listed obstacles including low membership, inadequate capital and limited savings mobilization, shortages of trainers and technical staff, and governance and transparency problems. In late 2004, the CUF had 44 CUs, many newly-registered but only 7 of them with more than one hundred members. There appeared to be only three potentially viable CUs, with some 2,000 members between them and deposits of around $0.85m. Around 780 of their members had loans, with balance outstanding about $0.33m, at an average of $420. This loan size is considerably above the mean level of about $70 for loans by members of AMFITIL (excluding IMfTL) and suggests a different social stratum is served. Reported portfolio-at-risk was very high, at 22%. The CU component of the ADB project was discontinued during 2004 and no credit was ever extended to CUs by MFITL, as had been envisioned. GoTL agreed to review the status of the CU movement. A recent decree law, under which responsibility for the CUs rests with the Department of Development and Environment, may have clarified the situation and provided a better basis for future development. However, the Sector Investment Program for private sector development41, while giving prominence to plans for strengthening producer cooperatives, does not devote any attention to credit unions and does not appear to envision further external assistance to the sector. It is important to recall that they are legally entitled to mobilize member savings and that they serve a population niche between those served by MFIs and banks. CUs would ideally feature in any concerted campaign to lift national savings capacity, such as is recommended by this 41RDTL, Private Sector Development: Priorities and Proposed Sector Investment Program, April 2005 43 Review. However the dangers of loss of deposits placed with ill-managed CUs give pause to plans for their expansion as long as fundamental weaknesses remain. Perhaps low-key assistance from the World Council of Credit Unions or one of its national-level affiliates is the most effective solution for the short- to medium-term. Participation in capacity- building programs initiated by IFC/PEDF is another possibility. 10.Supply of financial services: other sources In a market economy there will be demand for financial services. Formal sources of supply will come into existence so far as the regulatory environment permits and to the extent that the level of effective demand and the prevailing cost structure enable services to be supplied profitably. Where market conditions and the regulatory environment militate against the formal supply of services, informal mechanisms will usually emerge to satisfy demand. For example, in the absence of formal financial institutions, informal networks of family and friends will be called upon to meet the needs of households and individuals, while moneylenders will commonly be available to meet their needs on commercial terms. Local moneylenders are pervasive in Indonesia and were certainly present in East Timor up to 1999. Little is known about their operations, but it is possible that some were ethnic Indonesians who left the country in that year. Others may have sustained heavy losses during social upheaval, when debts became uncollectible, and since been unable to recommence operations (as indeed seems to have happened within the credit union movement; see discussion in the previous section of this review). Until recently inquiries revealed little or no knowledge of activity by commercial moneylenders, but recent anecdotal evidence suggest their activities to be rather widespread. The authors of another survey concluded that moneylenders have `all but disappeared' from Timor- Leste42. If that were true it would be a remarkable phenomenon, but appears not to be valid anymore. . Informal transfers between family members and friends are certainly pervasive in Timorese society and are among the means by which households smooth consumption and meet periodic or emergency financial needs, such as school fees or medical expenses. Such transfers also enable a broader cross-section of the population to access formal finance than may be apparent from the data on bank lending. As a prime example, it seems reasonable to conclude that the explosion of payroll lending from late 2003 has financed a variety of micro enterprise activities and created employment for persons other than the immediate borrowers. Concern sometimes expressed about borrowers taking loans excessively large in relation to their salaries may miss the point that such loans are likely to be serviced by households rather than individuals. Ultimately, the capacity to service these loans will be influenced by the profitability of the enterprises financed. The 42AMFITIL, Financial Services Sector Assessment in Timor-Leste, p. 5. The authors state that some 20% of the population used moneylenders pre-1999, and that the 2001 Timor Living Standards Survey had found 5% of respondents then accessing informal commercial credit. 44 profusion of taxis and other small public transport on the streets of Dili does not encourage optimism in this regard. However households will have strong motivation to repay such loans, despite any losses, and it should not be assumed they will fail to do so. Informal mechanisms appear to have played an essential part in restoring trade and commerce in Timor-Leste in the immediate aftermath of the events of 1999. It is important to remember the extent to which Indonesian interests, including the military, dominated the economy up to that time. Yet after the violence and the departure of these interests, and despite total destruction of the domestic financial system, cross-border mechanisms emerged to restore the supply of consumer commodities for survival and of capital goods for reconstruction. Indonesian and ethnic Chinese interests (including returning Sino-Timorese, some long absent from the country) re-established themselves soon after the conflict. New players also stepped in to fill gaps left by departing Indonesian entrepreneurs. Some were local and others were from overseas (including Taiwan and Singapore). They imported basic commodities, foodstuffs and consumer goods, badly needed at the time due to the disruption of normal trade and commerce. In the earliest days the trade was unenumerated and was financed from Indonesia. Goods made their way from Surabaya, Indonesia, via West Timor and across the border (along with new and second-hand vehicles and machinery to revive transport, agriculture and other activities). There was a countervailing flow of physical cash, initially Indonesian rupiah, but increasingly dollars. The arrival of entrepreneurs from other countries diversified sources of commodities and their financing. Thus wholesaling, and associated retail trade, was revived largely by market forces, with limited donor or Government intervention and in the total absence of domestic financing facilities. The purchasing power of the substantial international presence obviously supported revival, while donors provided credit directly to restart economic activity but this came later and was less effective. Offshore financing was still very important in Timor-Leste in 2005, especially in support of wholesaling, although unenumerated trade was apparently much less significant. At a lower level, direct relationships have been forged between Timorese contractors and retailers and Surabayan suppliers, often involving supplier credit arrangements. The role of local financial institutions in such arrangements may be restricted to money transfer. Dollarization also facilitates physical cash flows, of which the BPA has no record. In regard to direct donor involvement in financial service provision, the immediate post- 1999 period saw three attempts to revive lending for the resumption of economic activity, financed by the international community. The first two occurred as components of the Small Enterprise Project, stage 1 (SEP I) and the Community Empowerment Project (CEP), respectively. Both of these were financed by TFET and administered by the World Bank. The third was the ADB's Microfinance Development Project, described above in section 8. Aside from their intended contribution to economic revival, these three projects also had financial `system-building' impacts, both positive and negative. In addition there were a number of private initiatives conducted by international NGOs. 45 These moved from relief to reconstruction and the establishment of MFIs, described in section 9, above. SEP I involved a package of reconstruction activities, including disbursement of a credit line of $4.0m. This was for loans to individuals to finance the resumption of small enterprise, to be channeled via BNU, the predecessor of Bank CGD. SEP was designed to complement the ADB's Microfinance Development Project (one of the outputs of which was the establishment of IMfTL, as discussed above in section 8). Although SEP I became quickly effective (in April 2000), another year passed before the major portion of its credit line was disbursed. In the event, market forces and informal financing sources were much quicker to support the revival of economic activity, as described above. For this reason and because of difficulties encountered while implementing the initial credit line, a further tranche of loans proposed for SEP II was cancelled and the resources were diverted to other project activities. Efforts by Bank CGD to collect the outstanding balance of the original credit line have continued. In mid-2005 the repayment rate stood at a fraction above 75%, or some $3.0m of the original amount loaned, $4.0m. SEP I got into stride a full year ahead of the Microfinance Project, whose relatively slow start saw it play no role in the immediate humanitarian effort and only a delayed part in reconstruction. Both SEP I and the Microfinance project had system-building outcomes. SEP I employed a commercial bank (BNU, now CGD) as intermediary and enabled it to gain lending experience in an unfamiliar market without putting its own capital at risk. This capacity-building appears to have contributed to the bank's more recent high profile as a lending institution43. The Microfinance Project, for its part, led to the establishment of IMfTL, although its related objective of reviving the credit union movement and the CUF had limited success. The third international initiative was the Community Empowerment Project (CEP), directed primarily to creating social capital and establishing local-level governance. But one of its components also provided a credit line to finance the recapitalization of rural micro enterprises disabled by the events of 1999. While funds injected into rural communities by CEP assisted in reviving local economic activity, lending commenced only in the first half of 2001. So, as with SEP I, official finance lagged well behind the arrival of informal and cross-border financing. The CEP experience demonstrated severe limitations of microcredit as a tool for use in immediate post-crisis periods. Evaluations of both SEP I and the CEP credit experience concluded it might have been more appropriate to make cash grants to micro- and small entrepreneurs for the revival of economic activity, rather than attempting to institute formal credit procedures in a population suffering the immediate consequences of crisis. While CEP was not a financial sector project, per se, an official evaluation noted that its credit component `had implications for financial sector actors, including embryonic microfinance institutions and the commercial bank involved in SEP I, as well as for other 43For more information, see World Bank (2003) Report No. 26094, Implementation Completion Report on a grant in the amount of US$4.85 million to the Democratic Republic of Timor-Leste for the Small Enterprises Project (Private Sector Development Unit, East Asia and Pacific Region, June 26, 2003). 46 commercial banks contemplating entry to the market in Timor-Leste'. The evaluation saw CEP as possibly undermining Timorese credit culture, and as being unhelpful to the longer-run sustainability of rural financial services in Timor-Leste44. The GoTL intends a much reduced role for aid-financed lending in the future development of Timor-Leste's financial system. In the Sector Investment Program (SIP) for private sector development, GoTL allocations for micro-credit and SMEs, some $19m in the five years to 2003/04, are programmed to decline to about $5.5m in the five years to 2008/09. The SIP noted recent rapid increases in the capacity of domestic financial institutions to intermediate savings and proposed to shift the focus of external assistance more towards creation of `enabling environments', that is, to developing financial sector framework and policies. Future programs will rely more for private sector financing on organic growth through financial intermediation. To complete the catalogue of developments in Timor-Leste's non-bank, non-MFI financial sector since 1999, Western Union offers money transfer services from branches in Dili and Baucau and is said to serve about 1,500 clients. There do not yet appear to be any finance companies operating to fund purchase of consumer durables or producer goods. There is some evidence of supplier credit from domestic wholesalers to retailers, aside from the cross-border cases mentioned above, but most transactions appear to be conducted on a strictly cash basis. There is little evidence of consumer credit provided directly by retailers. There appear to be at least two pawnshops in operation. One of these visited had opened in April 2005 and in less than two months had accepted pledges from around 650 people, advancing some $110,000, at a mean size of $1,600 per pledge. Pledges redeemed in that period amounted to some $40,000. Standard pledge terms and conditions apply (including interest payable at 9.5% per month, flat) and these had been notarized by the Ministry of Justice, while the enterprise is registered with the Department of Commerce. The firm intended to expand by opening further branches in Dili. Maliana and Baucau were seen as other possible locations, although the owner expressed reservations about security outside Dili. He intended to extend his product range by introducing loans for periods of 12 to 36 months, operating as a finance company and lending against land certificates or to waged employees. He hoped eventually to establish a small regulated bank and to arrange payroll deductions with authorities. For its part, the BPA appeared to have a permissive attitude to the creation of finance companies and other commercial non-bank lenders, so long as their operations do not infringe the Banking Act. 44 See World Bank, Social Development Papers: conflict prevention and reconstruction, Timor-Leste: Independent Review of the Credit Component of the Community Empowerment Project, Paper No. 11, March 2004. 47 11.Assessment: the adequacy of financial services in Timor-Leste `Large' and medium-scale enterprises Few private enterprises in Timor-Leste qualify as `large'. Larger and medium-scale enterprises, mostly in importing and wholesaling, are likely to benefit from cross-border financing arrangements described in section 10, above. For these firms, access to investment- and working capital is not dependent on the adequacy of services provided by the locally-operating banks, though they are affected in various ways by other service deficiencies and by the high-cost operating environment. For larger and medium-scale firms dependent on the local banking system, real interest rates are, as described in section 6, quite high. High rates reflect perceptions of risk, legal uncertainties confronting lenders with respect to enforcement, and high costs of banking operations in Timor-Leste. Interest rates attached to offshore financing arrangements will also reflect some of these disabilities. To the extent such credit lines are secured against offshore assets and are enforceable in more predictable legal jurisdictions, risk premiums may be lower. However, even firms insulated against the disabilities of borrowing onshore may still be handicapped by difficulties associated with the credit environment. For example, importers and wholesalers may be discouraged from using credit as a marketing tool for distribution and retailing. This may help to explain the still-limited development of supplier- and consumer-credit facilities in Timor-Leste. Further, the provision of specialized financial services, such as insurance, equipment leasing and venture capital, lags in the current legislative and legal environment. In addition, all firms operating at this level are affected to a greater or less extent by the absence or limited availability of a range of complementary business services. As previously mentioned, their access to financial services is impeded by the absence of notary services and a centralized credit registry. An important constraint on the adequacy of financial services at the larger and medium- scales of operation is limited competition due to the small number of banks and the smaller number actively lending. Bank Mandiri's lending was restricted to fewer than a hundred enterprises with a mean loan size close to $100,000, and overlapped the medium- and small-scale categories, with the latter described in the next section. While the entry of Mandiri in 2003 improved the situation, the market remains very sensitive to the actions of individual players, as for example with the rapid expansion of lending since late 2003. While systemic failure does not seem to be an issue, a decision by one head office to reduce its exposure to Timor-Leste, or even to withdraw, could prove destabilizing, both to enterprises served by it and to the banking sector as a whole. On the 48 other hand, Bank Mandiri's entry as a specialized business lender was followed by a reduction in lending rates and increased deposit rates, to the benefit of both enterprises and consumers. The case for at least one further new entrant was made above in section 7. Entry by an additional bank, even if it were specialized in one particular niche, could have a beneficial influence on services across the board in so small a market, as well as contributing to the overall resilience of the sector. Small- and micro-enterprises (including the urban informal sector) Small-scale enterprises, and micro-enterprises both formal and informal, experience considerable difficulty in gaining access to commercial bank finance in Timor-Leste. ANZ is essentially a non-lending bank. Bank CGD is the major lender in this category (and indeed, overall) and made some 6,500 loans in 2004 at a mean size of about $10,500. On the face of it, this could constitute a major contribution in support of micro- and small enterprise activity. However, less than 20% of CGD lending appears to be in the `Trade and Finance' category. `Construction' loans, up to 40% of the total, appear to be its major product. Many or most of these are probably made to households, for the reconstruction of family residences. There is another category of `individual' loans, perhaps as much as 10% of CGD lending. This is not to say that such loans, totaling together perhaps half of the bank's lending, are irrelevant to small- and micro-enterprise. The household is the organizational base for most firms in Timor-Leste, as well as being the physical location of business activity in many cases. Some of this construction and individual lending, including payroll loans, may be applied to small-scale business ventures, but there is no way of estimating how much from the reported data. A further difficulty is the large and unexplained residual category in the total lending data reported by the BPA. As mentioned above, in section 6, this amounts to more than a quarter of the total. Future improvements in reporting could assist in clarifying these issues. Overall, it seems likely there is a considerable consumption component in total bank lending, much of it payroll-based. In these circumstances, small- and micro- enterprises have quite limited access to investment and working capital from the banks. They find also that neither supplier credit nor non-bank financing are available to them, due to the lack of such facilities. The USAID Desenvolve Seitor Privadu (USAID-DSP) project45 is considering whether it might be possible to `enhance the credibility' of selected small enterprises in Timor-Leste in the eyes of banks. This would be done by involving them in support programs that increase their creditworthiness and probability of success. Such `enhancements', designed to avoid the moral hazard problems associated with credit guarantees, are an interesting approach to the problem of financing for formal entities at the larger end of the small/micro category. 45DSP, Desenvolve Seitor Privadu, is a three year USAID initiative fielded from 2005. 49 The regulated microfinance institution, MFITL, is mandated to provide financial services to the micro-enterprises of the poor but has not been effective in that task. Instead it focuses largely on payroll lending with a substantial consumption component. While this lending helps to sustain the Institute, it is vulnerable to the expansion of cheaper commercial bank lending to centers outside Dili. MFITL's non-payroll lending is subject to high levels of delinquency and is not sustainable, while its capacity to offer deposit services is constrained severely by regulation. The microfinance sub-sector, including credit unions and MFIs, was discussed above in section 9. The CUs serve a higher income segment than that reached by the MFIs but no data were available concerning the purposes for which they lend. The numbers of clients are in any case very small and the viability of the CU model is an open question. At the lowest level, financial services are provided by the MFIs. These target their lending much more effectively to the support of enterprise than other financial institutions and also attempt to target the poor. At end-2004 they were serving some 15,600 borrowers and 17,000 savers, mainly in rural areas and in most provinces. On a household basis, their outreach might have been as high as 20% to 25% of the population below the poverty line, though to the extent they actually deal with non-poor clients this claim would have to be modified. Remarks made above, concerning disabilities affecting larger enterprises, also apply to some extent at this level of the enterprise hierarchy. The disabilities include high real interest rates, the absence of insurance and leasing services and the limited development of complementary services. These factors impact negatively on the access of small- and micro-enterprises to financial services, especially from regulated financial institutions. The small number of players in the banking sector, and in particular the absence from it of an efficient and commercially-motivated micro-banking institution, is another disability. Limited access to supplier credit and inability to arrange consumer credit for retail sales are other problems for firms at this level. Also, in the urban informal sector, many micro-entrepreneurs are rural-urban migrants who incur travel and other transaction costs to transfer cash for the support of rural households. Given the absence of financial institutions from all but a few rural centers, cheaper and more efficient remittance services are needed. Finally there is another factor that assumes greater importance at lower levels. This is the limited `financial literacy' of micro-entrepreneurs, which diminishes their capacity to make bankable proposals for financing and their capacity to use financial services effectively. Rural households Financial exclusion, a circumstance suffered by the great majority of rural Timorese households, is a key theme in this review. Section 3 above, discussed the meaning of `private sector development' in Timor-Leste and argued that the monetization and diversification of agriculture during the transition from subsistence is the starting point. With some 80% of population still in the farm economy and such a large subsistence element remaining within it, it is appropriate to commence private sector development there. Achieving financial `inclusion' for farm households is crucial in a situation where 50 the great majority continue to lack access to financial services, and especially to convenient, low-cost deposits, loans and remittances. The proposal for rural postal outlets to offer deposit and remittanceservices, discussed above in section 5, has promise. Similarly, possibilities for extension of basic services by means of remote cell phone coverage and the use of handheld offline electronic devices, also discussed above, should be explored and trialled. As the previous discussion has suggested, financial institutions (the commercial banks, IMfTL, the MFIs and Credit Unions) have so far achieved limited coverage. This limited outreach has occurred primarily but not exclusively in urban and peri-urban areas. With the exception of the banks, these institutions have also extended deposit and credit services to farm households in peri-urban and some rural areas, but only to a very limited degree. Apart from a small number of seasonal agricultural loans made by IMfTL, credit to the farm sector has been extended largely to finance off-farm activities of rural households. This is micro-enterprise lending and the discussion in the previous section could be regarded as having dealt with the topic. But the intention here is to focus particularly on rural households whose principal economic activity still lies within the farm sector. For these households, financial inclusion means gaining access to savings, credit and remittance services that will support the monetization and diversification of their sources of livelihood. This is the financial services strategy recommended in this study for `private sector development'. The IMfTL has a mandate to provide financial services to the poor, with the specific requirement that 65% of its lending should be in the form of microcredit. A cap imposed on its deposit-taking has discouraged the Institute from developing capacity for deposit- mobilization. Nonetheless it provides facilities for some 8,000 depositors, about double the number of its loan accounts. Another positive feature of its operations is that the Institute is less focused on Dili than the other banks, with more than 40% of its lending conducted outside the capital. However the great bulk of its loan portfolio is payroll- based. The Institute's other loan products (including seasonal-crop, market-vendor and group loans) have the potential to support the growth and diversification of rural household incomes. But this lending suffers high rates of delinquency and appears unsustainable without considerable reform and reorganization of the Institute's internal processes. The microfinance movement in Timor-Leste has achieved more outreach to rural households than all other financial institutions combined. As calculated above, MFIs may serve a fifth or more of all poor households, the majority of these in rural areas. Even the best MFIs struggle to achieve operational sustainability. However, the movement has active international links and possesses a strong mechanism for information-sharing and self-regulation in AMFITIL, the industry umbrella organization. MFIs also have an explicit role in assisting GoTL policy, as represented by the National Development Plan and the SIP for the private sector, so they may expect continuing support from government and donors. 51 Rural households suffer from the paucity of financial services available and their high cost. This is because MFIs in Timor-Leste labor under considerable disadvantage, resulting in high transaction costs of service provision and slow expansion of outreach. Rural households are also disadvantaged by `financial illiteracy', which is one aspect of the general scarcity of entrepreneurship. This is perhaps a deficit to be expected in rural communities making the transition from subsistence to a monetized economy, but it requires concerted educational effort. Financial institutions are among the agencies which can contribute to helping people understand the workings of a monetized economy, beginning with the discipline of saving. Problems faced by MFIs include low population densities and difficult transport and communications, as well as low degrees of monetization and limited economic activity in the outlying areas where they operate. To these are added the high cost of essential services such as electricity, and human resource limitations among staff. Most MFIs had been in operation for little more than three years by end-2004 and were struggling to achieve economies of scale. Nonetheless their comparative advantage as institutions is in dealing with rural households. By contrast, IMfTL has not so far demonstrated comparative advantage in lending to the poor, despite its mandate to do so. The Institute was also intended to act as a wholesaler of credit to the CU movement, but this did not eventuate due to the difficulty of rehabilitating the credit unions. This suggests the possibility that the Institute could instead act as a wholesaler to the best-performing MFIs, permitting them to expand more rapidly. By doing so it would move closer to fulfilling its mandate for microcredit, even if only indirectly. Whatever their strengths in lending, MFIs find difficulty in demonstrating comparative advantage in deposit-taking, so long as the Banking Act prohibits them from doing so. Prudential principles also dictate that they should not act as deposit-takers in the absence of appropriate supervision. The BPA tolerates their current practice of collecting savings as an element in loan contracts. This seems appropriate so long as net savings remain substantially negative46. However, MFIs could also mobilize voluntary savings (as distinct from contractual deposits) and place these with the Institute on behalf of their clients. Linkages between the Institute and MFIs would have more promise than linkages between other commercial banks and MFIs. Despite the extension of a loan by Bank Mandiri to the MFI Moris Rasik, the commercial logic for such linkages does not appear compelling under present conditions in Timor-Leste. 46Thus in table 4, MFIs are shown as holding deposits equal to about 40% of loans outstanding, and net savings were negative $0.6m. 52