Mozambique - Small and Medium Scale Enterprise Development Project

Document of The World Bank FOR OFFICLAL USE ONLY Report No: 18015 IMPLEMENTAT.ION COMPLETION REPORT REPUBLIC OF MOZAMBIQUE SMALL AND MEDIUM SCALE ENTERPRISE DEVELOPMENT PROJECT (Credit 2082-MZ) June 16, 1998 Private Sector and Finance Africa Region This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Jts contents may not otherwise be disclosed without World Bank authorization.  CURRENCY EQUIVALENTS Currency Unit = Metical (MT) US$1.0 = 11,850 (April 1998) ABBREVIATIONS AND ACRONYMS BM Bank of Mozambique BCM Banco Comercial de Mo,ambique/Commercial Bank of Mozambique BIM Banco Internacional de Mo,ambique BPD Banco Popular de Desenvolvimento/Popular Development Bank BSTM Banco Standard Totta de Mo,ambique/Standard Totta Bank CFD Caisse Fran:aise de Developpement ERP Economic Rehabilitation Program PFI Participating Financial Intermediary SMEDP Small and Medium Scale Enterprise Development Project ULC United Leasing Company FISCAL YEAR January 1 - December 31 Vice President Callisto Madavo Director Phyllis Pomerantz Sector Manager Thomas Allen Team Leader Arnold Sowa Team Member Gabrielle Rooz  FOR OFFICIAL USE ONLY IMPLEMENTATION COMPLETION REPORT REPUBLIC OF MOZAMBIQUE SMALL AND MEDIUM SCALE ENTERPRISE DEVELOPMENT PROJECT (Credit 2082-MZ) TAE,LE OF CONTENTS Page Preface Evaluation Summary ..................................................................... i-v Part I: Project Implementation Assessment ....................................................................1I A. Introduction ................1 B. Background ............................................................I C. Statement of project objectives and evaluation ......................................... . 4 D. Major factors affecting the project and implementation record .......................... . 5 E. Project documentation and data ............................ ................................5 F. Achievement of objectives ...........................................................S5 1. Relieve scarcity of term finance to viable enterprises to develop the industrial sector ............. ..............................................S5 2. Rehabilitating the banking system's intermediary role for termn lending 13 G. Project sustainability ............................................................ 14 H. Bank performance . . . 15 E. Borrower performance . . . 16 J. Assessment of outcome ............................................................ 16 K. Future operations . . . 17 L. Lessons learned ................................................................... 17 PartII: StatisticalAnnexes ..................................................................... 19 Appendixes A. Mission's aide-memoire ................................................................... 25 B. Borrower's contribution to the ICR ................................................................... 28 C. Cofinancier contribution to the ICR ................................................................... 34 D. Borrower's comments on draft ICR ................................................................... 35 This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  IMPLEMENTAT][ON COMPLETION REPORT REPUBL][C OF MOZAMBIQUE SMALL AND MEDIUM SCALE ENTERPRISE DEVELOPMENT PROJECT (Credit 2082-MZ) Preface This is the Implementation Completion Report (ICR) for the Small and Medium Scale Enterprise Development Project in Mozambique, for which Credit 2082-MZ in the amount of SDR 25.1 million (US$32.0 million equivalent) was approved in December 1989, and became effective in June 1990. Parallel financing was provided by the European Investment Bank and Caisse Franqaise de Developpement. The Credit was closed in December 1997, one year after the original closing date of December 1996. Final disbursements were made on April 17, 1998. Total Credit disbursement was estimated to be SDR 22.4 million (US$32.0 million equivalent), and an estimated SDR 2.7 million (US$3.6 million equivalent) will be canceled. The ICR was prepared by Gabrielle Rooz, and the Team Leader was Arnold Sowa, AFTP1. It was reviewed by Ms. Phyllis Pomerantz, Country Director, CD2, Mr. Paul Murgatroyd, Financial Sector Lead Specialist, AFTP1, Mr. Brian Falconer, Principal Operations Officer, OS2, and the previous task managers. The Sector Manager is Mr. Thomas Allen. The borrower and Caisse Franqaise de Developpement provided their own evaluation that are included as appendixes to the ICR. Preparation of this ICR was begun during IDA's final supervision mission in September 1997. The ICR mission took place in December 1997 during which the mission discussed the effectiveness of this operation with staff of the apex unit, the participating financial intermediaries, and a sample of eight firn beneficiaries selected at random. The mission members would like to thank them for their kind cooperation and assistance in the preparation of this report. The report is based to a large extent on their inputs. The report is also based on material in the project file. The report was sent to the borrower and cofinanciers for comment. The borrower contributed to preparation of the ICR by commenting on the draft ICR.  IMPLEMENTATION COMPLETION REPORT REPUBL][C OF MOZAMBIQUE SMALL AND MEDIUM SCALE ENTERPRISE DEVELOPMENT PROJECT (Credit 2082-MZ) Evaluation Summary Introduction 1. The Small and Medium Scale Enterprise Development Project (SMEDP) was prepared in 1988 to support and complement the Government's Economic Rehabilitation Program (ERP) launched in 1987. The project was designed to rehabilitate the industrial andfinancial sectors. The project built on the short-term import support assistance provided to the industrial sector by external aid and started to address the longer-term issue of industrial investment and rehabilitation. To complement this objective, the project sought to rehabilitate the banking sector and to revive prudent investment lending to enterprises. The SMEDP focused on the small-scale enterprises. An ongoing Industrial Enterprise Restructuring Project was approved in parallel in 1989 and focused on larger scale enterprises. 2. The ERP was the first comprehensive reform program implemented by IDA to assist the Government transform a command economy into a market-oriented one. The measures were implemented through two Rehabilitation Credits (1987 and 1989) to first address macroeconomic distortions, sectoral allocations of expenditures, and credit policy. Prior to the ERP, IDA provided assistance to the country through an Economic Action Program initiated in 1984 which was essentially viewed as a "quick-fix"-type lending operation and had a limited content policy. Project Objectives 3. The project's main objectives were to: * Relieve scarcity of term finance to financially viable enterprises in the industrial, mining, agriculture, transport, and construction sectors. * Rehabilitate the banking sector's capability to provide term lending. * Assist the Government in formulating an industrial policy. 4. The project included two related components: * An Enterprise Financing Component (LTS$28.6 million equivalent) for a line of credit to be on-lent by the central bank, through an apex unit, to the participating financial intermediaries (PFIs) * An Institutional Strengthening Component (US$3.4 million) for technical assistance and training to the PFIs and staff of the apex unit in the areas of project appraisal and processing. 5. Cofinancing or parallel financing included Caisse Fran aise de Developpement and the European Investment Bank for a total amolnt of US$4.5 million equivalent for the credit line. The cofinanciers financed eight sub-loans in total, of which one was cofinanced with IDA. Prior to effectiveness, UNDP financed a pre-operational training program which was continued under the project for staff from commercial banks, the apex unit, and IDIL, the public business services agency. 6. The objectives of the SMEDP were relevant but untimely. The project supported the Government's first attempts to introduce fundamental changes to transform the economy, revitalize and  develop the industrial sector, and rebuild the banking sector. In this context, the objectives were justified. 7. Their achievement was, however, largely impaired by two main factors: (a) the difficult political and economic environment in which the Government had to manage the transition from a centrally planned to a market-driven economy; and (b) weak financial institutions and lack of key reforms in the banking,sector which were initiated only in 1992. The likelihood of a reversal in macro and interest rate policies was then high. 8. Also, the potentially high cost to the financial institutions, which were to channel termn financing to high-risk sub-borrowers under highly unstable political and economic conditions and lacked the required skills, was not sufficiently recognized. Finally, the adverse effects of the economic liberalization policy on domestic producers during the transition years were not sufficiently taken into consideration. Nor were the infrastructure deficiencies caused by an ongoing civil war during the first years of implementation with enduring effects throughout implementation. 9. lhe design of the SMEDP reflected the same lack of sequencing found in the different elements of the enterprise reform policy of the ERP, as well as inadequate attention to their linkages. Two of these elements were: (a) the commercial and legal framework, in particular lack of enforcement of commercial contracts, as reflected in the accumulation of bad debt in the productive sector and between that sector and the financial sector, because of past commercial practice and state credit provisions; and (b) delayed reform of the banking sector. The legacy of the former contributed to the high default rate on sub-loans. Delayed reform of the banking sector led to the macroeconomic instability which dominated during much of implementation. Monetary leakages through the two former state-owned banks generated monetary instability and the consequent macroeconomic instability. The combined effect of a weak banking sector and macroeconomic instability was one of the major factors contributing to the project's poor results. Implementation Experience and Results 10. Overall, the project did not achieve its objectives. Although most of the funds from the credit line were disbursed, IDA's transferred resources did not finance viable investments, based on a currently high default rate of 73% on the sub-loans. Training to upgrade commercial bank staff skills in project appraisal and processing techniques did not yield the expected results, also as evidenced by the high default rate. 11. The credit line financed a total of 134 sub-projects, of which 63% were new projects. The largest share of total investment and number of sub-loans were in the industry and manufacturing sectors (27% each), with a 55% concentration in Maputo. Investmnents were expected to generate a total employment of 5,029 jobs, at a cost of US$5,627/job (ex-ante). Based on actual data from one commercial bank and an 8-firm sample visited as part of the ICR mission, a combined total number of 664 jobs were created, for a total sub-loan amount of US$3,044,419, at an average investment cost/job of US$4,585. Thus, 10% of IDA funds created 664 jobs. If all loans were good loans, the actual job creation would be close to the appraisal estimate of 6,000. However, although firms' performance could improve if the current economic trends continue, the present uncertain status of many firms tends to make the overall employment creation questionable. 12. Initially, three commercial banks participated in the on-lending program--two state banks which were privatized towards the end of the project, and a private bank. Two additional PFIs, one private bank ancl one leasing company, participated in the program toward the end of the project. The two former state banks on-lent 87% of IDA funds and together accounted for almost all defaulting sub-loans.  iii 13. The project was implemented in an unstable political and economic environment and during periods of high inflation. Although interest rates were liberalized in 1994, they remained at fixed and negative levels until end of 1996, as the necessary pre-conditions for a market-driven financial system were absent. One of the major causes was the predominance in the banking system of state-owned banks which had operated for years with relatively poor commercial cultures in a non-competitive environment and under a system of state-directed credit allocation. 14. Lending rates, both on commercial loans and sub-loans, were negative in real terms during most of project execution. Real negative rates resulted in attractive subsidies for the sub-borrowers. If the majority of problem sub-projects does not prove financially viable overtime, these implicit subsidies could add to the country's already excessive debt burden, further worsening its debt service payment capacity. Finally, negative rates discouragedl domestic resource mobilization, which in turn perpetuated a scarcity situation and adversely affected the project's sustainability. 15. Achievement of the institutional strengthening component, which accounted for only 11% of total Credit, was unsatisfactory. Skills of commercial bank staff in accounting, project appraisal and processing were deficient and did not improve significantly. The pre-operational training program was insufficient, both in scope and duration. Of a total of 50 staff who received training, only three were left in the banks' credit department, two in one bank, and one in the other. Financial assistance provided to the firms for assistance in developing business plans, but without the adjunct ongoing technical assistance, resulted in poor appraisal of the financial viability of the sub-projects contributing to the high default rate. Local private and public entities did not have the required skills in this area to provide meaningful support services to the firms. The objective of establishing an apex unit to create a long-term structure for future IDA or other donor credit lines was not achieved. 16. Project Sustainability. The project was implemented under highly uncertain conditions. The project, as designed and carried out under thlese conditions, is not sustainable. Changes which would pave the way for macroeconomic stabilizatiion started to take place in 1995, with 1996-1997 as the transition years. It remains to be seen how the firm beneficiaries will perform over a given time horizon, or the survival rate, under the current much improved economic conditions. Only then can true rates of return, both financial and economic, be measured, hence the project's impact on the firm beneficiaries. The project's sustainability in terms of financing productive capacity is thus uncertain. However, while not quantified, a large number of sub-projects may have had negative returns, and those sub-projects which will survive are likely to have low returns because of the long delays in project benefits. The project's sustainability in terms of building the PFIs' capacity to provide ongoing term lending is unlikely. 17. Bank performance. The project was prepared and implemented during a period of fundamental economic and institutional reforms. The Bank's assistance to the country was hindered by a highly uncertain political situation following the aftermath of 16 years of a civil war which ended in 1992. However, in the rush to provide term lending to help the Government relieve a scarcity in term financing to small firms, and in the absence of a developed private sector, the Bank over-relied on state-owned or controlled public entities as participating agencies. While the Staff Appraisal Report (SAR) correctly highlighted the limitations of the Business Environment Study which provided the basis for project design, the project was nonetheless prepared on the Bank's optimistic assumptions as to the pace of the reform program. Neither the study nor the SAR were complemented by sufficient financial sector work. One fundamental flaw in the project design, which was also found in the overall reform program, was to have failed to recognize the crucial linkages between the financial and real sectors. Because Mozambique was considered a strategic country at the time, recommendations of the Levy report were not taken fully into account. Neither were OD 8.30 during implementation. With the benefit of hindsight, the project should have been substantially delayed and/or designed as a small pilot operation, as opposed to a full-fledged investment pn)ject, given the uncertainty as to the pace of the reform  iv program, lack of knowledge about the SME sub-sector, and the country's very low institutional capacity. Supervision tended to focus on improving the slow disbursement rates, at the expense of project impact. While corrective actions were taken following a change in task management at the end of the project, by suspending on-lending of funds through the two state banks, these actions failed to have an effective remedial impact, as most funds were already disbursed, and the project was about to close. Borrower performance. Until their privatization, the Government's inadequate control over the two state banks was largely responsible for its failed monetary policy, resulting in high inflation. Project performance would have benefited from the apex unit's greater involvement in the monitoring and supervision of sub- projects, especially given the poor recovery rate on the sub-loans. Assessment of Outcome 18. Based on project results, project outcome is unsatisfactory. Future Olperations 19. Measures to build on reforms already achieved under prior adjustment operations are part of the proposed Economic Management Reform Credit (FY99). The EMRC will particularly focus on fiscal reforms and on improving the business environment to foster private-sector-led growth. Fiscal reforms will include financial deepening and better coordination between monetary and fiscal policies, and private sector measures will aim to establish a legal and administrative framework in line with market- driven economic structures. A follow-up investment operation is also slated for FY99. 20. Notwithstanding substantial achievements in macroeconomic policy and privatization programs over the last five years, private-sector growth response has been modest, especially with respect to local Mozambican-owned firms. The proposed Mozambique Enterprise Development Project (the PoDE) will aim to improve the business environment and accelerate broad-based economic growth through capacity building of private firms and institutions. Particular emphasis will be placed on developing Mozambican-owned firms' export potential and linking domestic suppliers with foreign investors. Specifically, the project will: (a) assist enterprises in raising the technical skills of their management and workers; (b) provide financial resources for export development and enterprise technology upgrading; and (c) improve private and Government institutions' capability to deliver support services to the business community. Key Lessons Learned 21. The key lessons learned are: * Implementation of this project confirms the validity of OD 8.30 which were not rigorously followed during project implementation, particularly with respect to term lending under high inflationary periods and by weak institutions, and distortions created by directed credit and targeted credit subsidies, in particular interest rate subsidies. Lines of credit are not useful tools when implemented under unstable economic and business conditions because risks associated with new investments are too large. * Lines of credit are not the appropriate instruments to address sectoral policies. Lines of credit are most likely to be adversely affected by macro and sectoral distortions which are best addressed in an overall macroeconomic policy framework. Continuing commitments under FILs should be linked to the effective implementation and maintenance of a satisfactory policy framework, a justifiable pre- condition for an IDA line of credit as well as for its successful implementation.  v + It is very difficult to stop commitments under a project which has no clearly established ongoing conditionalities once the project has been launched. Ongoing sub-project commitments should be explicitly linked to project performance. Project implementation design should include explicit monitoring performance indicators to assess the continued eligibility of financial institutions and other entities, public or private, sub-project performance, and technical assistance impact. * The financial sector should not be viewed essentially as a "non-real" delivery mechanism to support the real sector, with little inherent importance in itself. Prospects for success are poor when PFIs are unsound and lack the skills and commercial orientation necessary to make good credit decisions. + It is very difficult to take a financially poorly performing bank with non-commercial objectives and weak management--characteristics most likely to be found in state-owned than private banks--and turn it into an effective financial interme,diary. * Interest rates should be variable to allow the pricing of capital to be market-determined and to reduce interest rate risk for lenders and borrowers. During high inflationary periods, an indexing mechanism might be best, as opposed to Government-adjusted rates, although experience such as in Brazil has shown that indexing tends to perpetuate inflation. An indexing mechanism might also be applied to Government's fee for currency risk coverage, if the Government is to bear the cost of foreign exchange. Proceeds from the fee should be deposited in a fund in a commercial bank to earn interest. * Implementation of the project raises the usefulness of "quick-fix"-type lending operations in highly uncertain conditions. When local capacity is low, and/or fundamental economic and sectoral reforms are underway, a small pilot project and more prolonged training programs are more suitable than a full-scale operation to fully take into account the time factor during the transition years, and adapt and build knowledge about the (sub)sectors involved. The pilot project should include a simple, straightforward, and focused approach to reduce the final costs to the banks and firms. The approach should also be sufficiently flexible to allow for changes as needed in project execution. * Technical assistance to build SMEs' capacity is as important as their access to capital. This technical assistance need not be a supplement to sub-loans, but could be a stand-alone component. It should be demand-driven. * The apex unit should be managed under a management contract arrangement. This is especially crucial if the objective of setting up an apex unit is to transfer the monitoring and supervision aspects closer to the financial intermediaries ancl final beneficiaries. Monitoring indicators should be set to assess the unit's operating performance. Finally, in addition to monitoring a database on sub-loans and projects, the apex unit should possess or develop an analytical capability to assess the project's impact on an ongoing basis. This would allow introducing needed changes in the project scope and/or design throughout implementation. Under this project, the apex unit is meant to be a "sunset organization".  PART 1: PROJECT IMPLEMENTATION ASSESSMENT A. Introduction 1. The Small and Medium Scale Enterprise Development Project (SMEDP) was prepared in 1988 to support and complement the Government's Economic Rehabilitation Program (ERP) launched in 1987. The project was designed to rehabilitate the industrial andfinancial sectors. The project built on the short-term import support assistance provided to the industrial sector by external aid and started to address the longer-term issue of industrial investment and rehabilitation. To complement this objective, the project sought to rehabilitate the banking sector and to revive prudent investment lending to enterprises. The SMEDP focused on the small-scale enterprises. An ongoing Industrial Enterprise Restructuring Project was approved in parallel in 1989 and focused on larger scale enterprises. B. Background 2. In the years preceding preparation of the SMEDP, the effects of years of civil war and a centrally planned economy since the country's independence from Portugal in 1975 were devastating. Social indicators were among the worst in the world. An estimated 60-70 percent of the population lived in absolute poverty, the agriculture sector was below subsistence levels, and internal fighting and sabotage had destroyed much of the rural infrastructure, Real GDP declined by an estimated 12.8 percent p.a. on average in 1980-87, and share of industry, as a % of GDP, by 22 percent on average over the same period. As a result of a deterioration in the country's export and import capacity, Mozambique became overly dependent on emergency assistance and external aid, with an external debt stock equal to 3.5 times GDP and 45 times exports. Finally, the country's extremely weak institutional capacity could be improved only by training an entire new generation, "a task that would take more than a decade".' 3. The adjustment process to overhaul the economy from centrally planned to market-oriented was tackled by successive adjustment programs and credits, supported by the IMF and other donors. The first IDA Credit to Mozambique was the Rehabililation Program in 1984. The Credit had a limited policy content and was viewed essentially as a "quic]k-fix"-type lending operation. It supported some specific producer price reforms and changes in Mozambique's import procurement procedures. The Credit was successful in establishing a Bank/Government dialogue on, and commitment to, structural adjustment and the Bank's central role in the process. It was followed by the more comprehensive Economic Recovery Program (ERP), 1987-89. The program's adjustment measures were included in the Second Rehabilitation Credit (1987) (SRC) and the Third Rehabilitation Credit (1989) (TRC). 4. The ERP's broad objective was to create greater reliance on market mechanism, reduce state intervention in economic activity, and integrate the national economy within the international economic environment. The reform program's primary strategy was to re-establish a market-based incentive structure in the productive and distributive sectors. Initial reforms of price and import controls were supported by the SRC and TRC, as well as the first Credit. All three Credits were largely effective in implementing these reforms and maintained the impetus for change, despite the adverse effects of the insecurity situation until the signing of the FPeace treaty in 1992, and the severe drought of 1991-92 which devastated the agricultural sector. Anothler external shock was the country's loss of key markets in Eastern Europe when aid from, and trade with, the former Soviet Union stopped in 1991, resulting in loss of foreign exchange of over 8 percent of GDP. 5. In 1987, at the start of the ERP, as in 1989 when the TRC and the SMEDP were prepared, the primary problem faced by the Bank and the Government was to design a set of reforms in the context of ' World Bank. Mozambique, Financial Sector Study, Report No. 10269-MOZ, September 1992.  2 an ongoing war situation. It was recognized that the uncertain situation would markedly impact on the responsiveness of changes to the reforms. Consequently, the reform program was designed based on a gradualist approach to take into account the war situation, the time required to allow the population to adjust to a fundamentally different economic and social environment, and the severe constraints on the supply side, including infrastructural deficiencies. 6. Throughout the 1989-95 adjustment period, major progress was made towards relaxation of pricing controls and adoption of prices more closely linked to international prices. However, this process led to domestic industry facing increasingly intense international competition for which it was not well prepared. ]First, the obsolete production processes and grossly distorted financial structure placed it in a weak positiion. Also, although domestic prices had been realigned with the international price structure to a considerable extent, the effectiveness of the reform program was seriously hampered by the failure to unite the parallel and official product markets. This failure was caused by the influx of emergency assistance, of which a large percentage was illegally finding its way to parallel markets; the increased flow of illegal imports from neighboring countries; the tax evasion in the parallel markets, both on international trade and domestic transactions, rendering domestic producers uncompetitive (and reducing the revenue flow to the government); the continuing, although not as large, overvaluation of the local currency. Substantial imports into the domestic economy through parallel market channels were facilitated by the ineffectiveness of the national customs system and the fiscal system's deficiencies. 7. Effectiveness of the reform program in the areas of enterprise and financial sector reforms was also limited. State ownership of the productive enterprises and financial institutions was still dominant. Reforms in the financial sector had been viewed as less of a priority as reforms in the enterprise sector, and were delayed as a result. Failure to recognize the crucial linkages between the two sectors was to severely imipact on successfully reforming the enterprise sector. The banking sector was supporting loss- making state enterprises. The design of the SMEDP was based on the same oversight of the linkages between the financial and real sectors as in the reform program. Delayed reforms of the banking sector led to the imacroeconomic instability which dominated during much of the execution of the SMEDP. Monetary leakages through the two state-owned banks generated monetary instability and the consequent macroeconomic instability. The combined effect of a weak banking sector and macroeconomic instability was one of the major factors contributing to the project's poor results. 8. The next adjustment operation was the Economic Recovery Credit (FY92) which supported the early phases of financial sector reforms, and further integration and liberalization of the exchange rate. This Credit was followed by the Second Economic Recovery Credit (SERC) (FY94) and the Third Economic Recovery Credit (TERC) (FY97). The SERC focused on strengthening fiscal and monetary policy and an interlinked program of enterprise and financial sector reforms. One of the major achievements of the SERC was to establish macroeconomic stability and low inflation through successful reforms in the financial sector--privatization of state banks, opening up of the sector to an increasing number of private banks, bringing in more competition and resulting in more and better financial services. Macroeconomic stability was also achieved through the privatization of state enterprises which represented the large, non-performing borrowers from the banking system. The ongoing Industrial Enterprise Restructuring Project was restructured in 1994 to provide more focused technical assistance to privatized firms which also contributed significantly to the acceleration of the privatization program. The privatization program in turn contributed to a faster eco-iomic growth through increased industrial output.  3 9. Macroeconomic Indicators, 1987-1997' . A ------_ ---------Actua --------------------- 997 Estimates 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 DNP* INE* Real GDP Growth Rate (%) 4.6 8.2 6.5 0.9 4.9 -0.8 18.8 4.5 1.4 6.4 7.9 7.9 Inflation (end-period) (%) - 49.4 35.0 47.1 35.2 54.5 43.6 70.2 54.1 16.6 5.8 5.8 Private Investment/GDP 18.8 21.1 18.9 23.7 26.0 31.1 38.0 30.1 32.2 31.5 25.8 13.6 Public Investment/GDP 17.3 22.2 21.46 20.9 19.7 17.6 16.4 20.9 19.0 16.8 19.3 14.3 GrossDomestic SavingslGDP -2.4 -2.1 -6.7 1.5 4.5 5.5 10.4 5.7 18.0 19.9 22.5 11.1 Gross National Savings/GDP 2.2 12.4 3.2 15.2 12.5 15.8 19.6 15.7 17.7 17.5 21.7 11.3 Public Sector Surplus -22.9 -27.0 -24.8 -29.2 -24.9 -26.3 -22.2 -29.7 -20.8 -17.0 -20.1 -14.9 (Deficit)/GDP /1 Real Effective Exch. Rate/2 -61.0 -36.0 3.0 0 -16.0 -24.0 -1.0 -1.0 -4.0 4.0 10.0 10.0 Exports (GNFS)/GDP 13.0 15.7 16.0 16.0 21.8 23.7 21.3 24.3 27.4 27.6 26.4 19.5 Imports (GNFS)/GDP 51.4 61.1 63.:3 59. 63.0 66.9 65.3 69.6 60.6 56.0 49.0 36.2 Cuffent Account /GDP /3 -25.4 -23.3 -28.2 22.0 -16.5 -18.6 -21.9 -20.5 -22.7 -22.4 -13.4 -9.9 Total Debt Service/GDP /4 40.5 43.2 37. 34.6 34.1 40.7 34.1 32.0 28.4 21.8 18.1 13.4 Source: CAS 1997. /1 = excl. grants; /2 percent change; (-) = devaluation; /3 = after grants. /4 = total debt service due before debt relief, including IMF. DNP = National Planning Commission; INE = National Statistics Institute. 10. Economic growth. Real GDP growth hides important annual and sectoral variations. In 1993, the year following the peace accord, GDP growth is estimated at 19 percent p.a. as the influx of refugees were returning to the rural areas and resumed agricultural production. By contrast, in the same year, industrial production fell sharply. The declining trend was reversed in 1995 when manufacturing output increased by 18.5 percent following the privatization of large, loss-making state enterprises. The transport and commercial sectors also experienced strong growth over the same period, estimated to be 6.4 percent in 1996 and 7.9 percent in 1997. 11. The financial sector. At the time the SMEDP was prepared, the financial sector included the Bank of Mozambique (BM); Banco Popular de Desenvolvimento (BPD), a second state-owned bank which was largely directed to the agriculture sector; Banco Standard Totta (BSTM), a private bank; and a state insurance company. The Bank of Mozambique was the central bank as vvell as the largest commercial bank. Its primary function was to meet the needs of the centrally planned economy. Banking services were limited to accepting deposits, to be lent out to Government and state enterprises. The Bank undertook the first sector study in 1992 which served as a basis for outlining a reform program. The study stated that as far as the financial sector of Mozambique was concerned, it was more a matter of forming, rather than reforming one. 12. One of the first priorities was to separate BM's central and commercial banking functions. The separation was completed in 1992 by establishing the Bank of Mozambique as the central bank and the separate Commercial Bank of Mozambique (BCM). While this measure resulted in producing a clean balance sheet for the central bank, it did riot address structural problems, and the accumulated non- performing loans from state enterprises for the period 1987-1992, representing 4 percent of GDP, were transferred to BCM. Measures to rehabilitale, restructure, and privatize BCM and BPD were supported under the Second Economic Recovery Program in 1994. BCM was first restructured owing to its extremely poor financial condition: the bank's losses were equivalent to 12 percent of GDP. It was eventually privatized in July 1996 when it became difficult to keep its lending operations within agreed credit ceilings. BPD was privatized in September 1997. Prior to their privatization, the Government's inadequate control over the state banks was largely responsible for its failed monetary policy. In 1996, immediately before their privatization, a substantial financial deterioration in the two banks caused them to run excessive overdrafts. 2 Statistical data (level of GDP and ratios to GDP) from the National Planning Commission (DNP) tend to be somewhat misleading due to inadequacies in the national accounts. The Govemment has, recently adopted statistics from the National Statistics Institute (INE) as official data, starting with the year 1991. Statistics from DNP are used in this report for time series purposes, but [NE 1997 estimates are also shown.  4 13. Over the life of the project, Mozambique's financial sector gradually evolved from a predominantly state-controlled system to one which is now fully private. These changes were progressively reflected in the type of financial intermediaries participating in IDA on-lending program. By project completion, two additional private intermediaries were channeling funds from the line of credit (Banco International de Mozambique (BIM) and United Leasing Company (ULC)). However, the impact of an emerging competitive financial market had relatively little effect on the project's outcome as these changes took place towards the end of the project. Hence, with respect to the participating financial intermediaries, the focus of the report is on the three commercial banks which participated in the on-leiading program from the outset: BCM, BPD, and BSTM. C. Statement of Project Objectives and Evaluation Overall objective * Support and complement the ERP reforms by rehabilitating the financial and industrial sectors. Main objectives + Relieve scarcity of term finance to viable enterprises, building on the short-term import financing assistance provided so far by the donor community * Rehabilitate the banks' capability to provide term lending + Assist the Government in formulating an industrial policy Line-of-caedit * US$28.6 million equivalent (Enterprise Financing Component) * Funds on-lent by the apex unit to PFIs. Technical assistance * US$3.4 million equivalent (Institutional Strengthening Component). and training Cofinanciers * Caisse Franaaise, European Investment Bank, for a total of US$4.5 million equivalent. 14. The objectives of the SMEDP were relevant but untimely. The project supported the Government's first attempts to introduce fundamental changes to transform the economy, revitalize and develop the industrial sector, and rebuild the banking sector. In this context, the objectives were justified. 15. Their achievement was, however, largely impaired by two main factors: (a) the difficult political and economic environment in which the Government had to manage the transition from a centrally planned to a market-driven economy; (b) weak financial institutions and lack of key reforms in the banking sector which were initiated only in 1992. The likelihood of a reversal in macro and interest rate policies was then high. 16. Also, the potentially high cost to the financial institutions, which were to channel term financing to high-risk sub-borrowers under highly unstable political and economic conditions and lacked the required skills, was not sufficiently recognized. Finally, in addition to infrastructure deficiencies, the adverse effects of the economic liberalization policy on domestic producers during the transition years were not sufficiently taken into consideration. 17. As stated in para. 7, the design of the SMEDP reflected the same lack of sequencing found in the different elements of the enterprise reform policy of the ERP, as well as inadequate attention to their linkages. Two of these elements were: (a) the commercial and legal framework, in particular lack of enforcennent of commercial contracts, as reflected in the accumulation of bad debt in the productive sector and between that sector and the financial sector, because of past commercial practice and state credit provisions; and (b) delayed reform of the banking sector.  5 D. Major Factors Affecting the Project and Implementation Record 18. The major factors undermining the project's success were then, to a large extent, the combined effects of the above factors: * An unstable political and economic environment. * Failure to take into account in the project design the gradualist approach of the ERP. * Failure to recognize the linkages between the financial and real sectors in the preliminary sector work3 which provided the basis for project design. This resulted in lack of financial sector work to complement the preliminary study * Because Mozambique was viewed at the tirme as a strategic country, recommendations from the 1989 Report of the Task Force on Financial Sector Operations4, and subsequently OD 8.30, were not taken fully into account, especially with respect to term lending in high inflationary periods and distortions created by directed subsidized credit. + The Bank's over-reliance on state-owned or controlled public entities as participating agencies: state- owned banks, as financial intermediaries vhich did not have the incentives to make proper decisions in the allocation of capital; and IDIL, a state-run business advisory services, in the rush to provide term lending to help the Government relieve a scarcity in term funds to small firms, and in the absence of a developed private sector. + The Bank's emphasis on disbursement rates to measure project performance, rather than actual impact. 19. The project was appraised in April 1989 and negotiated in June 1989. The Credit was approved by the Board in December 1989, signed in February 1990, and became effective in June 1990. The closing date was extended by one year, from December 31, 1996, to December 31, 1997 to allow for disbursements on committed funds. E. Project Documentation and Data 20. During implementation, the apex uniit maintained a large database of basic raw data on the lending program: sources and uses of funds, sectoral and regional distribution, and expected job creation. Unfortunately, data on file were not updated after the sub-loans were approved and disbursed, and no data were kept on the status of sub-loans or on the actual performance of the firms. Statistical information presented in this report is based on these ex-ante data, as well as available data collected from the PFIs during the ICR mission, and interviews with the PFIs and a sample of eight firms selected at random. The PFIs (mainly the former two state banks whose combined shares represented 87 percent of total lending from IDA funds) did not monitor their internal costs associated with the sub-loans. The ICR mission was not able to obtain financial statements from the PFIs, especially the former state banks, to calculate basic financial ratios, such as earnings performance and exposure, or age of, and trend in, arrears. Visited firms did not produce financial statements or records to permit a re-assessment of rates of return on their sub-projects. F. Achievement of Objectives (a) Relieve scarcity of term foreign exchange financing to viable enterprises to develop the industrial sector 3World Bank. Mozambique - The Development of Industrial F'olicy and Reform of the Business Environment, Report No. 7795, May 1990. 4World Bank. Report of the Task Force on Financial Sector Operations, No. R89-163, August 1989.  6 21. The credit line under the Enterprise Financing Component was to provide foreign exchange for fixed asset investment; associated incremental permanent working capital; and consulting services for project preparation. The following table shows the financing arrangements under the project for the line of credit: Financing Arrangements Disbursement Mechanism On-lending Mechanism Repayment Mechanism Terms: 40 years, standard 100 percent FE - 70 percent LC IDA Grace period: 10 years Interest rate: 0.75 percent _ ~~~bears cuffency nsk refinance 90-80 percent * Central Ban fee to PFIs: 73/65 % PFI current account debit mechanismn 0 Apex Unit on final on-lending rate same due dates as those on individual sub-loans l ~~~changed to 50% in 1993 30 percent of LC bears credit risk on full 1 ~~~sub-loan value Sub-loan denominated and repaid in local currency 10-20 percent equity Firms Maturity: 3-12 years On-lending Interest Rate: payable every 3 months Grace period: : 1-4 years FE= foreign exchange; LC = local costs. 4 90 percent for expansion/rehabilitation projects. 80 percent for new projects. 22. Resource Transfer. About US$28.8 million, out of an initial allocation of US$28.6 million equivalent, IDACreditAllocationMbPFI were disbursed under this component. 23. Participating Financial Intermediaries (PFIs). IDA Credil; allocation among the PFIs closely reflected BCM 20% the banks' relative share in the banking sector: BCM 67 67% percent, B]PD 20 percent, and BSTM 5 percent. By contrast, lILC's 6 percent share was relatively high, especially given the leasing company's late participation in the on-lending program, in the last year of the project. On the other hand, BIM's 2 percent participation was low relative to the bank's size in the sector. Source: Apex unit. 24. The enterprise financing component of US$28.8 million equivalent supported a total investment of about M4ts. 431.8 billion, or US$81.4 million equivalent. Of this amount the EIB and Caisse Franqaise de Developpement financed in parallel about US$4.5 million equivalent, or 6 percent, the PFIs about Mts 76.7 billion from their own funds, or 18 percent, and the firms about Mts. 115.1 billion, or 27 percent. Thus, total local contribution to the credit line represented 44 percent of total investment. Firms' participation was their equity contribution to total investment costs, of 10-20 percent. PFIs' contribution, as recorded by the apex unit at the time of sub-loan approval, represented the local costs component of the sub-loan: the working capital portion of the sub-loan, and/or IDA 30 percent local costs requirement when not covered by the firms. Under the project, PFIs were not required to have direct exposure, although their participation was estimated at appraisal to be about 7 percent. 25. In reality, the banks' contribution was greater. Additional loans from the PFIs to the firms, at commercial rates higher than the subsidized on-lending rates, resulted from frequent cost overruns experienced by the sub-projects, compounded by unpredictable inflation. The cost overruns were in large part due to lengthy (one year on average) procedures in the complete approval/disbursement cycle. They were also due to additional financing needs for working capital requirements resulting from  7 underestimated initial investment costs. BCM reported an average 30 percent increase beyond its initial contribution. Requiring a greater equity participation from the firms was seldom a workable alternative solution as they did not have the resources. Most of them had already found the 20 percent equity requirement difficult to meet. (Normally, new investment projects should not have leverage higher than 2:1, except in very unusual low-risk situations. New, very risky sub-projects, such as those financed by IDA line of credit, had very high debt/equity ratio, which was likely to undermine the financial conditions of the PFIs.) In no case was the sub-borrower's equity investment re-assessed after the sub- loan was approved, nor was the debt/equily ratio recalculated. Finally, because of exceptions to the global credit ceilings in effect under the IMF program granted to "special" programs, including the IDA line of credit, the 1992 sectoral study raised the issue of leakages through the banking system contributed by the line of credit (and other donors) adding to an already difficult monetary control situation. 26. The effect of this increase on the PF'Is' portfolio was mitigated by the small proportion the credit line represented as a percentage of their overall lending portfolio: BPD 9 percent and BSTM 6 percent. BCM's figure was not available for preparaltion of the report, but was also assumed to be low. While the banks indicated that IDA funds relieved a scarcity in term credit to credit-constrained enterprises, they showed a mixed response to the on-lending program for the reasons explained below. 27. First, as stated above, since the PFIs were allowed to circumvent credit ceilings by lending from this credit line, they used the IDA line of credit only to complement the ceilings--although not always adhered to by the former state banks, nor reinforced by the Government for state banks. The banks were inclined to sidestep this credit to offer more attractive short-term commercial loans to their customers because of higher spreads on commercial loans and lending restrictions associated with the directed funding element of the credit line. In doing so, they were able to take advantage of the substitutability, albeit limited and distorted, in the local financial markets. This was certainly the case for BSTM, the private bank and the only bank to abide by credit ceilings. 28. Second, PFI use of their own resources resulted in a term mismatch of their balance sheet, given the short-term nature of deposits in Mozambique. 29. - --Third, the banks indicated that, without the IDA line of credit, they would not have channeled resources to small sub-borrowers because of- inadequate collateral, high risks, and high transaction costs relative to the size of the loans. Added incentives resulted from the Government assuming the currency risk and sub-loan repayments in local currency.5 BSTM circumvented high transaction costs associated with high-risk sub-borrowers by lending only to long-time customers with an established track record. This practice largely eliminated the need for' project evaluation for which the bank did not have staffing resources. 30. Fourth, poor loan collection under the line of credit reduced the banks' profitability. Reduced profitability could only reinforce the banks' bias against small firms and will make it difficult for them to include these firms into their normal lending portfolio. This in turn affected the program's sustainability. 31. The fifth and last factor was the relatively high fee charged by the central bank on the PFIs' final on-lending rates to the firms to cover the apex unit's operating costs and the Government's currency risk, resulting in unfavorable spreads to the PFIs. This spread was initially set at 73 percent of the final on- lending rate for sub-loans up to seven years, and 65 percent above seven years. It was later changed to 5Since the banks did not monitor their internal costs associated with the sub-loans, it was not possible to deternine the magnitude of costs related to the processing and supervision of sub-loans and risks, if any.  8 50 percent for all new sub-loans following the project mid-term review in November 1993, when the high spread was recognized as a major deterrent to PFI participation. As no market-determined interest rates existed at the time, this method was used as opposed to the central bank charging the PFIs their marginal (or average) cost of capital. The banks indicated that this fixed cost, even amended, contributed to their negative returns. 32. In 1996, the central bank established, as its minimum on-lending rate, a reference base rate which closely followed the discount rate, but which was always set below the discount rate. Should the final on-lending rate charged by the PFIs to the firms be lower than the reference rate, the central bank's fees were charged on the reference rate in order to ensure an adequate cost coverage for the central bank. At Credit closing, the reference rate was 12 percent, compared to the discount rate of 12.95 percent. 33. On-lending Rates. Under the terms of the project, on-lending rates on sub-loans were to be variable and market-determined. On-lending terms and conditions, and progress in achieving positive on-lending rates in real terms were to be reviewed annually by IDA and the Government if review of the general interest rate structure was no longer undertaken under an ongoing IMF program. 34. UJntil 1994, when interest rates were liberalized, the Government followed a policy of administratively fixed rates, which were adjusted in 1991 to positive real levels and fluctuated between positive and negative in real terms in 1992-1993. However, even when fully liberalized, interest rates did not respond to market forces because the pre-conditions for a market-driven system to work were not in place. A major cause was the predominance in the banking system of state-owned banks which had operated for years in a non-competitive environment and under a system of state-directed credit allocation. 100- 90 CAmual Inflation so * * w~ ;;;; Ct ~ ~ ~ Subloan Rate 70 60 - Commercial Rate 50 40 20 ~~~~~~~~~~~~~~~~~~~~~~~~- - - Subloan Rate 10 -10 -20 - - - Subloan real rate -30 -40 -so -60 -- 2Commercial real 91 A J 0 92 A 3 0 93 A J 0 94 A 3 0 95 A J 0 96 A 3 0 97 A J 0 rate Month Sources: Commercial banks, Bank ofMozambique, National Statistics Institute. 35. Thus, during the 1994-96 transition period from a state-controlled to market-oriented system, commercial rates remained fixed and negative in real terms. IDA sub-loans (and other donors' lines of credit) benefited from a preferential rate, fixed by the Government below the commercial rate, to induce borrowing by the targeted small firms. The result was an attractive subsidy for small firms as their borrowing cost did not reflect the true opportunity cost of capital. At the end of 1996, with a substantial fall in inflation, real rates turned positive. Small firms who contracted their debt when rates were heavily subsidized saw an unexpected real increase in their debt-servicing obligations contributing to collection difficulties. 36. Given rising inflation and real negative rates on average during project implementation, the banks would have found it difficult to earn a positive rate of return on their assets and maintain a real  9 return on their capital (not only on the credit line but on their overall portfolio as well), resulting in an erosion of their capital base, even if high bad debt had not caused losses. Losses resulting from a high percentage of bad loans (on their overall portfolio, including the line of credit) contributed to a deterioration of their portfolio. 37. At the end of 1996, BSTM which had 100 percent good sub-loans indicated it was able to earn a nominal profit margin of 5 percent, which it found adequate. On the other hand, BCM and BPD reported that they could not post positive margins on the sub-loans which were transferred to them when they became private banks as the two banks on-lent the majority of bad sub-loans. At Credit closing, on- lending rates charged by the commercial banks still followed closely the commercial rates of 19-25 percent. BCM's gross spread on on-lending rate was 10 percent, BPD 12.5 percent, and BSTM 9 percent. BCM was negotiating with the central bank a reduction in the 50 percent servicing fee. 38. Currently, the 19-25 percent commercial rates in Mozanbique are high relative to the estimated low inflation rate of 5.8 percent. This is because, although a number of reforms have been put in place, the banking sector is still underdeveloped and cash-based. Banks, which have weak treasury management, tend to hold a high percentage of non-earning assets (equal to about 30-40 percent of deposits) as excess reserves to respond to the cash needs of their clients. This increases the spread they require between deposits and lending rates. Another factor is the relative lack of competition among banks. 39. Recovery Performance. The arrears situation remained serious, as shown in the table below. Total no. of No. Paying No. Rescheduled/ No. in No. in PFIs active sub-projects Regularly Paying Irregularly No. Defaulting Grace Period Collection Process BCM 89 6 37 41 3 2 BPD 23 2 4 9 4 4 BIM 1 - - - 1 _ BSTM 8 6 - 2 ULC 13 8 1 - 4 Total 134 22 42 50 14 6 40. The overall default rate of 73 percent (excluding sub-loans in grace period) was mainly the result of poor lending by the former state banks, with a higher concentration in BCM, given the bank's higher lending share relative to the other banks. No data were available from BCM and BPD regarding the age and accumulation of arrears, or their proportion as a percentage of their total outstanding portfolio. The negative impact of the large number of problem sub-loans on the banks' overall portfolio was assumed to be relatively small because they constituted a small proportion of their total portfolio. The two banks estimated about 67 percent of sub-loans in arrears were collectable. In contrast, BSTM had a recovery rate of 100 percent, and ULC 99 percent. 41. The repayment situation resulted from a number of factors. The firms' inability to repay reflected the difficult economic environment prevailing during implementation. A large proportion produced for the local markets. The low recovery rate was also due to poor appraisal of sub-projects, lack of ex-post supervision by the PFIs leading to an inability to trace some sub-borrowers, a culture of non-repayment of loans, reinforced by poor loan collection from the banks and a virtually non- functioning legal enforcement system. 42. The recovery performance also reflected a flaw in project design in that IDA did not have the explicit power to suspend new commitments by PFIs who experienced poor loan collection performance. Although the Project Agreement required independent annual audits of the banks, this requirement, which was a legal covenant, was only complied with twice, once in 1992 and again in 1994 when IDA  10 first attempted to stop lending through the forner state banks. The first unqualified audits were produced in 1996. 43. Because of poor loan recovery, the rediscount mechanism set up under the project to recycle repaid funds was never used, thereby increasing the small firms' dependency on the original credit line. 44. The PFIs were responsible for assessing the sub-borrowers' creditworthiness and bore the credit full risk. This was meant to ensure a thorough credit analysis and ex-post supervision of sub-projects. Staff did not perform this normal commercial lending task satisfactorily because of insufficient staffing resources and serious lack of technical skills, a legacy from years of state-directed credit allocation. Staff compensated for this lack of skills by relying on IDA's or the apex unit's prior review of credit applications. 45. Neither assumption of full credit risk nor potential losses from bad loans were incentives strong enough to ensure adequate credit risk assessment and commercially oriented lending from the banks. The two largest lenders of the credit line were state entities which, for most of project execution, the Government kept bailing out until their privatization. Total assumption of their losses by the Government amounted to the equivalent of US$93 million in 1993 under the IMF's program. The Government injected an additional US$80 million equivalent to cover BCM's losses when the bank was privatized. The amount of BPD's non-performing loans and other bad assets assumed by the Government is not yet fully known, as it will be fully determined once the one-year period in which the new private owners can turn over to the Government any losses discovered during that year has been completed. 46. Characteristics of Sub-projects and Sub-loan Realization. Appraisal Estimates Actual Estimates * Number of Sub-projects 110-120 134 * Type of investment: - new 85 (63 percent) - expansion 10 (7 percent) - rehabilitation 39 (29 percent) * Job creation: 6,000 5,029 (ex-ante) - maintained 1,046 - new 3,983 * Average investment cost/job(US$) 4,800 5,627 (ex-ante) * Average sub-loan size (US$) 250,000 211,194 Sub-loan size (US$'000) No. of sub-loans 1-49 29 50-99 37 100-249 36 250-499 17 500-999 12 >1,000 3 * Average maturity 7 years * Average grace period 2 years 47. Sub-Project Achievements. Based on a sample of 30 sub-projects, ex-ante financial rates of return were high in nominal terms, averaging 42 percent, compared to an appraisal estimate of 12 percent in real terms. A minimum economic rate of return of 12 percent (in constant terms) was to be achieved for sub-projects above US$250,000. Neither the PFIs nor the apex unit calculated realized returns and were not required to do so. Given the high number of problem sub-projects, the majority may not be financially viable and expected returns may not be achieved.  11 48. Firm Size. To qualify as a small and medium enterprise, firms were to % f i Total no. of sub-loans and value by firm size employ no more than 200 workers (before 35% expansion). About 76 percent of sub- 30% r-nOota.lOvasue -F borrowers employed no more than 50 25% |total -L workers (after expansion), with 40 percent 20% mm of their sub-loans adding to less than o 1 fn 100,000 each. Total value of their share 5% I_ was 68 percent. Very small-scale 0v l A borrowers (with 10-19 employees) - g 0 0 Oo A accounted for a total of 41 sub-loans (31 irm size percent), but only 18 percent share of the funds. It is interesting to note that this Source: Apex Unit. group's repayment performance tended to be better. Amott Credit Denmnd b Year (in US$ illion) 20 - . . 8G 15 7.0- 10 6.0- 5 5.0 _ 0 4.0- F] 5.0- 20 -n; 1.0 l2 0.0. +C-m2sieR 1991 1992 1993 1994 1995 1996 -25 Year lo93 2994 295 19 Source: Apex Unit. 49. There seemed to be a positive relationship (correlation coefficient = .8) between credit demand and the level of real interest rates, with a time lag. That is, higher subsidized rates resulted in higher demand for investment in the following year. The time lag could be due to a delayed response to changes in interest rates by the firms, the time required to prepare business plans, credit rationing, and delays by the banks in processing firms' sub-loan requests. However, due to so few observations, this is not statistically significant. Also, lack of a 1:1 relationship may be due to (a) firms' uncertainty with respect to future changes in inflation and on-lent rates, and (b) firms not expected to repay. In this case, firms made business decisions irrespective of the level of interest rates. 50. Sectoral Impact. The manufacturing and industry sectors each accounted for 27 percent of both total investment and the number of sub-loans. Agriculture and fishing accounted for 16 percent, transport 14 percent, services 6 percent, tourism 9 percent, and mining 1 percent. Inh.a b..e Z.m bezi, 6% a ~~~~~~~~~~~~~~~za~~2 /-Tanspor l2Y / 9 14% AS&FWh Mining . p 1%~~~~~~~~~~~~~1 nd us 0M a f 1%N11 Sectoral Distribution ofSubloans Regional Distribution of Subloans Source: Apex unit.  12 51. Regional Impact. About 55 percent of total sub-loans, both in terms of value and number, were in Maputo. This seemingly high concentration ratio is in fact lower than might have been expected, given that, traditionally, economic activities in Mozambique have tended to be highly centralized in or around the capital city. To some extent, the project's decentralization objective was achieved. 52. Employment Generation. By sector, expected job creation was the highest in the %ofdor industrial and manufacturing sectors, za o accounting for a total of 1,733 (of which 160 1,357 new jobs) and 1,430 jobs (of which i,xo 1,231 new jobs), respectively. Expected job wo creation, by type of sub-project, was 4* consistent with uses of funds. New investments accounted for 63 percent of total sew Tomn T an. investment and were expected to generate 50 percent of total jobs. However, the high credit demand for new investments was not ExpectedJobCreation byLoan Size consistent with the state of the economy for no. of jobs (in US$ '000) which a higher investment demand for 1,60 business expansion or rehabilitation should -4.200 - have been expected. With respect to borrower soo size, expected job creation was the highest in 400- [ H the 100-249 employee range, with a total o 0l 1-49 50-99 100-249 250-499 500-999 >1,000 estimate of 1,404 jobs. Loan size Source: Apex Unit. 53. UJnfortunately, no ex-post figures were available, except from BSTM which confirmed that the estimatedl number of jobs created was realized. The combined total of the bank's figures and an 8-firm sample visited by the ICR mission gives an actual number of jobs created of 664, for a total sub-loan amount of US$3,044,419, making an average investmnent cost/job of US$4,585. Thus, about 10% of IDA funds created 664 jobs. If all loans were good loans, the actual total number of jobs created would be close to the appraisal estimate of 6,000. While it remains to be seen how the total number of firm beneficiaries will perform over a longer period of time under the current much improved economy, the present uncertain status of many firms tend to make the overall employment creation questionable. 54. Environmental Issues. There were no major environmental issues related to the implementation of sub-projects. In the few sub-projects where this could have been the case, the issues were dealt with by IDA either during the prior review process or the supervision missions to ensure compliance with project requirements. 55. Current Status of sub-Projects. Of a total of 134 firm beneficiaries, 22 (16 percent) are considered profitable, 92 (69 percent) are operating with difficulty, and six (4 percent) went out of business and are in judicial collection. 18 firms have negotiated an extension of the grace period because of cash flow problems. 56. VVhile economic and infrastructure conditions were important contributing factors, the major problems faced on the firm level were lack of technical know-how, lack of production management and managerial skills, and poor market assessments. Technical appraisal of sub-project applications tended to emphasize the mechanics of pro-forma balance sheets and cash flow analysis and did not give sufficient weight to more qualitative factors, such as adequate management skills and market assessments, and the  13 applicant's ability to operate a small business. Firms' problems in these areas are confirmed by the banks in their latest supervision report. 57. In some cases, however, there appears to be little correlation between firms' profitability and their loan repayment performance. Non-repayment of sub-loans tended to be encouraged by lax loan collection by the former state banks and irnplicitly encouraged by a virtually non-functioning legal enforcement system. In other cases, overleveraged firms, unable to obtain additional loans, included as part of their retained earnings for expansion or working capital what should have been used to service their debt. 58. Of the eight firms visited as part of the ICR mission, four were considered profitable, of which three were larger, well-established enterprises. Of these firms, two were making payment on their sub- loan--one refinanced its sub-loan with a short-term commercial loan at more advantageous terms. The other two were financing their business expansion or working capital with their loan payments. The other four firms were new, smaller businesses operating with difficulty and unable to service their debt obligations on a regular basis. All four were experiencing market-related and/or management problems. Furthermore, the firms complained of the lack of transparency in the sub-loan approval process, suggesting possible corruption. (b) Rehabilitating the banking system's intermediary rolefor term lending 59. The Institutional Strengthening Component (US$3.43 million equivalent) was designed to rebuild the banking system's capability for term financing, and help establish (a) a local capacity for business advisory services to SMEs, and (b) a local permanent structure to monitor IDA or other donor future lines of credit. Total disbursement for this component was US$3.23 million equivalent. It is interesting to note the imbalance between the credit line component, which represented 89 percent of total credit, and this component, on which the project's success critically depended, which represented only 11 percent. 60. Provisions for the training of banking staff were clearly inadequate, in light of their low level of skills identified at preparation. The training program, while useful, was not sufficient in duration, as the PFIs received, prior to effectiveness, a total of only six-week training and three workshops of two weeks each, in the areas of project appraisal and processing. The training program was funded by UNDP. US$0.5 million was provided for additional training under the project for a total of 50 staff from commercial banks, the apex unit, and IDIL. Commercial bank staff indicated they found the training useful, but those who received training were subsequently transferred to other positions within the banks. Of the 46 staff trained, only two are left in B3PD credit operations department, and one at BCM. Of the two people trained in the apex unit, only one is left, the other staff was transferred to another division of the central bank. The project provided for early, limited pre-operational training programs, but nothing to correct the actual problems that emerged during implementation. Furthermore, to provide training only to credit department staff when the two state banks had many serious weaknesses outside the credit department could only impair their ability to iimplement the line of credit. 61. A banking institute was established, in 1994 under the ongoing Financial Sector Capacity Building Project. The institute includes on-going training support to commercial bank staff as part of its activities, with mixed success. 62. The project also provided financial cassistance to enterprises to prepare business plans. Firms could use the services of private local consultants or IDIL, the Government's already established business services bureau. Fees charged for services could be financed by sub-loans. Provisions for this fundamental aspect were also clearly inadecluate. Lack of local capacity in this area reinforced the  14 banks' dependency on IDA or the apex unit's prior review. Low quality of IDIL services was the major factor for low demand for its services. 63. In 1995, IDA suspended disbursement in support of IDIL, pending an independent audit of its operations. In line with project requirements, IDIL was to charge an up front fee for its services. These fees were to gradually cover its operating costs for the bureau to become self-sufficient. After five years in operation, IDIL could not account for the cash revenues received, hence IDA request for the audit, to which IDIL never agreed. IDIL is currently operating on a government-funded budget. 64. An apex unit was established in the central bank to manage the credit line with the objective to create a long-term structure for future IDA lines of credit, or those of other donors. The unit was initially independent, operating with its own budget and staff, and was later absorbed by the central bank in 1994. Throughout project execution, the apex unit confined its role to that of an administrator of funds. In line with project requirements, the apex unit devoted most of its time to evaluating applications and ensuring conformity with Bank procurement guidelines. These responsibilities which duplicated normal day-to- day functions of commercial banks represented an additional step in the approval process, causing further delays and increasing the sub-borrowers' final costs. Because the PFIs bore the credit risks, the apex unit did not see sub-project monitoring and supervision as part of its responsibilities. Another reason was the automatic debit mechanism built into the project which allowed the central bank to debit the PFIs' current account on the same due dates as those on individual sub-loans, irrespective of the firm's capacity to repay. Currently, the apex unit is managing sub-loans of the ongoing Industrial Enterprise Restructuring Project and other donors' lines of credit. 65. Procurement. In line with project requirements, firms were to obtain three independent quotations for goods to be procured under the IDA Credit. Given the small size and nature of their requirements, and the limited local market, small firms found it difficult to meet IDA's procurement requirement. This requirement proved neither economic nor efficient, as firms sought the required quotations in neighboring countries or abroad, resulting in delayed processing. The requirement also tended to add to costs, as suppliers charged for price quotations. G. Project Sustainability 66. The project was implemented under chaotic conditions, both at the macroeconomic level and in terms of disruptions of key infrastructure (roads, utilities, market distribution channels, etc.) The project, as designed and carried out under these conditions, is not sustainable. Changes which would pave the way for macroeconomic stabilization started to take place in 1995, with 1996-1997 as the transition years. It remains to be seen how the firm beneficiaries will perform over a given time horizon, or the survival rate, under the current much improved economic conditions. Only then can true rates of return, both financial and economic, be measured, hence the project's impact on the firm beneficiaries. The project's sustainability in terms of financing productive capacity is thus uncertain. The project's sustainability in terms of building the PFIs' capacity to provide ongoing term lending is unlikely. 67. Sustainability of future operations rests on three main areas. (a) sustained macroeconomic stability to remove the need for Government subsidies to guarantee interest and exchange rate stability and promote longer term growth; (b) sustained progress in financial sector development to further competition, to improve banking supervision, and to increase efficiency in domestic resource mobilization and allocation. SMEs' increased reliance on internal resources (their own equity and retail bank contributions), as opposed to external resources (IDA, Government subsidies), will ultimately attest to sustainabiility; and (c) further actions in private sector development in the legal and administrative framework to change outdated pre-independence laws and regulations.  15 H. Bank Performance 68. Preparation and Appraisal. The project was prepared and implemented during a period of fundamental economic and institutional reforms, as well as a highly uncertain political situation following the aftermath of 16 years of a civil war which ended in 1992. Results of the Business Environment Study, which provided the basis for project design, were thus highly tentative. While the Staff Appraisal Report (SAR) correctly higlhlighted the study's limitations, the project was nevertheless prepared on the Bank's optimistic assumptions as to the pace of the reform program, and neither the study nor the SAR were complemented by sufficient financial sector work. With the benefit of hindsight, the project should have been substantially delayed and/or designed as a small pilot operation, as opposed to a full-fledged investment project, given the uncertainty as to the pace of the reform program, lack of knowledge about the SME sub-sector, and the country's very low institutional capacity. 69. During appraisal, emphasis was placed on setting up the credit line, settling BM's lending arrears, recapitalizing the banking system as a whole, and establishing the apex unit. The fact that it was necessary to recapitalize the banks to be eligible to participate in the line of credit was symptomatic of their uncertain status. Project preparation and appraisal was unsatisfactory. 70. Major weaknesses in project design were (a) failure to take into account the country's unsuitable macroeconomic and business environment; (b) lack of explicit links between project performance and sectoral policies and IDA ongoing commitments; (c) lack of monitoring performance indicators for the commercial banks and for sub-projects; and (d) training and technical assistance inputs which were not commensurate with the country's institutional capacity requirements. In addition, project design would have benefited from experiences and lessonis drawn from other Bank lines of credit elsewhere, to the extent these can be transferred. 71. Supervision. Project supervision tended to focus on improving the slow rate of disbursement. The two main causes identified during the November 1993 project mid-term review were unduly long approval processes and unfavorable spreads for the PFIs. At that time, only 27 percent of total project funds, compared with projected 48 percent, had been disbursed. Amendments were introduced to correct the slow rate by: (a) increasing the free-limit ceiling, from US$250,000 to US$500,000 to decentralize more approvals of sub-projects to the apex unit, and (b) reducing the central bank's servicing fee to correct earlier disincentives to the banks. Steps were also taken to address an accumulation of non- performing loans. However, little emphasis was placed on the commercial banks' continued eligibility to participate in the IDA on-lending program. 72. IDA's decision to increase the free limit on sub-projects to improve slow disbursements on the sub-loans at a time when the quality of the sub-projects, as evidenced by poor repayment records, was relatively poor reflected a blatant disregard for the impact of the line of credit on the PFIs. The banking sector's role was then essentially viewed as a "non-real" delivery mechanism to support the real sector, with little inherent importance in itself. For the first five years of project implementation, IDA supervision was deficient overall. 73. In the last years of implementation, following a change in task management, focus of supervision started to shift from disbursement rate to PF][s' portfolio performance. Measures were introduced with a view to improving that performance. IDA requested, with little success, that the two former state banks develop action plans to improve and monitor collections of non-paying loans. In 1996, IDA suspended their participation pending achievement of a 10 percent recovery rate. Delays in sub-loan approvals were also reduced, thereby reducing the final costs of funds to firms. These measures, although sound and needed, came too late as, by then, 75 percent of total funds had been disbursed, and the project was in its last year. Closing the project then was< discussed but did not materialize because the Government  16 argued of an adverse impact on the small firrns, given the scarcity of term credit. As more private banks and two leasing companies were entering the banking sector and expressed interest or participated in the on-lending program, the project was kept open. Other ways of providing term finance to SMEs, other than through the formal banking sector, were also explored. For the last two years of project implementation, IDA supervision was satisfactory. E. Borrower Performance 74. ThrougL3ut implementation, the Government's inadequate control over the former two state banks was largely responsible for the macroeconomic instability, as well as their poor loan recovery performance, which prevailed until project closing. This instability in turn contributed to the wide fluctuations in inflation, leading to the need to subsidize interest rates on the sub-loans, and to recapitalize the banks. Project performance would have benefited from the apex unit assuming a stronger monitoring and supervision of sub-projects and, to this end, establishing better coordination with the PFIs. Finally, frequent changes in management adversely affected the unit's proper functioning, as each new director had to learn about its operations. 75. Comp?liance with legal covenants and audits. Compliance with legal covenants was generally satisfactory, except with respect to interest rate policy and submission of annual audits. However, although most covenants sought to address major issues in the banking system (settlement of banks' lending arrears, action program to rehabilitate the banking system and upgrade the banks' accounting capabilities), these issues were addressed only once and after Credit effectiveness, or no concrete actions were implemented beyond preparing the action plan. The end result was that, despite satisfactory one- time compliance with the covenants, the problems kept re-surfacing since generic issues and causes of the banking sector's poor performance were not addressed. The commercial banks' lending arrears were a recurrent problem throughout project execution, including under the IDA line of credit, although somewhat masked by grace periods on loans. The Government recapitalized the banks three times. Bank staff accounting skills were deficient. Compliance with financial statement audits was not satisfactory in terns of timely submission of audit reports. Qualified audit reports were not dealt with by the Government, despite IDA's expressed concerns. J. Assessment of Outcome 76. Overall project outcome is unsatisfactory. * Relevance. The project was relevant but untimely as the objectives were unsuited to the country's conditions. * Efficacy. 1. The project's transferred resources to relieve a scarcity in term funds did not support viable investment projects, as evidenced by the currently high default rate on the sub-loans. 2. T}e project's institution-building objective was not achieved. Impact of technical assistance and training provided to: (a) PFIs, to develop skills in credit risk assessment and project appraisal, was limited; (b) apex unit, to build a permanent local structure, although more successful, was nonetheless not sustainable; and (c) IDIL, to build a local capacity for business advisory services, was limited. 3. The project did not assist the Government in achieving a more rational interest rate structure. Throughout most of project execution, in the absence of the necessary pre-conditions in the banking system for interest rates to be market-determined, rates remained administered, and negative in real terms, discouraging domestic resource mobilization, perpetuating a scarcity situation., and adversely affecting the project's sustainability. Subsidized rates may also have had the  17 effect of leading to larger loans than necessary, attracting marginal borrowers, and eliminating necessary spreads for making lending viable to the PFIs. 4. Because of poor project results, the project did not assist the Government in formulating an industrial policy. Although progress of policy reform was to be monitored as part of the annual project implementation review, funds earmarked for studies were never used for that purpose. * Cost-Effectiveness. Implementation of the project came at a significant net cost to the economy. Negative real interest rates on sub-loans resulted in a net transfer of resources to firms. If the majority of problem sub-projects does not prove viable investments, the resulting subsidies could add to the country's already excessive debt burden, further worsening its debt service payment capacity. While not quantified, a high number of sub-projects may have had negative rates of return, and even those which will survive are likely to have low, if not negative, rates of return because of the long delays in project benefits. The former state banks would have incurred substantial losses on the sub-loans had it not been for Government's recapitalizations before their privatization. If the banks' estimated recovery rate of 670/o on the sub-loans does not rmaterialize, and if the Government no longer covers their losses, they stand to sustain substantial losses on the sub-loans. + Ratings. Financing of viable projects: unsatisfactory Institution-building: unsawisfactory Interest rate policy: unsatisfactory Industrial policy: unsatisj2ctory K. Future Operations 77. Measures to build on reforms already achieved under prior adjustment operations are part of the proposed Economic Management Reform Credit (FY99). The EMRC will particularly focus on fiscal reforms and on improving the business environment to foster private-sector-led growth. Fiscal reforms will include financial deepening and better coordination between monetary and fiscal policies, and private sector measures will aim to establish a legal and administrative framework in line with market- driven economic structures. A follow-up i:nvestment operation is also slated for FY99. 78. Notwithstanding substantial achievements in macroeconomic policy and priivatization programs over the last five years, private-sector growth response has been modest, especially with respect to local Mozambican-owned firms. The proposed Mozambique Enterprise Development Project (the PoDE) will aim to improve the business environment and accelerate broad-based economic growth through capacity building of private firms and institutions. Particular emphasis will be placed on developing Mozambican-owned firms' export potential and linking domestic suppliers with foreign investors. Specifically, the project will: (a) assist enterprises in raising the technical skills of their management and workers; (b) provide financial resources for export development and enterprise technology upgrading; and (c) improve private and Governmerit institutions' capability to deliver support services to the business community. L. Lessons Learned * Implementation of this project confirms the validity of OD 8.30 which were not rigorously followed during project implementation, particularly with respect to term lending under high inflationary periods and by weak institutions, and distortions created by directed credit and targeted credit subsidies, in particular interest rate subsidies. Lines of credit are not useful tools when implemented under unstable economic and business conditions because risks associated with new investment are too large.  18 * Lines of credit are not the appropriate instruments to address sectoral policies. Lines of credit are most likely to be adversely affected by macro and sectoral distortions which are best addressed in an overall macroeconomic policy framework. Continuing commitments under FILs should be linked to the effective implementation and maintenance of a satisfactory policy framework, a justifiable precondition for a IDA line of credit as well as for its successful implementation. * It is very difficult to stop commitments under a project which has no clearly established ongoing conditionalities once the project has been launched. Ongoing sub-project commitments should be expliciltly linked to project performance. Project implementation design should include explicit monitoring performance indicators to assess the continued eligibility of financial institutions and other entities, public or private, sub-project performance, and technical assistance impact. * It is very difficult to take a financially poorly performing bank with non-commercial objectives and weak management--characteristics most likely to be found in state-owned than private banks--and turn it into an effective financial intermediary. * The financial sector should not be viewed essentially as a "non-real" delivery mechanism to support the real sector. Prospects for success are poor when PFIs are unsound and lack the skills and commercial orientation necessary to make good credit decisions. * Implementation of the project raises the usefulness of "quick-fix"-type lending operations in highly uncertain conditions. When local capacity is low, and/or fundamental economic and sectoral reforms are underway, a small pilot project and more prolonged training programs are more suitable than a full-scale operation to fully take into account the time factor during the transition years, and adapt and build knowledge about the (sub)sectors involved. The pilot project should include a simple, straightforward, and focused approach to reduce the final costs to the banks and firms. The approach should also be sufficiently flexible to allow for changes as needed in project execution. * Technical assistance to build SMEs' capacity is as important as their access to capital. This technical assistance need not be a supplement to sub-loans, but could be a stand-alone component. It should be demand-driven. * Interest rates should be variable to allow the pricing of capital to be market-determined and to reduce interest rate risk for lenders and borrowers. During high inflationary periods, an indexing mechanism might be best, as opposed to Government-adjusted rates, although experience such as in Brazil has shown that indexing tends to perpetuate inflation. An indexing mechanism might also be applied to Government's fee for currency risk coverage, if the Government is to bear the cost of foreign exchange. Proceeds from the fee should be deposited in a fund in a commercial bank to earn interest. * The apex unit should be managed under a management contract arrangement. This is especially crucial if the objective of setting up an apex unit is to transfer the monitoring and supervision aspects closer ito the financial internediaries and final beneficiaries. Monitoring indicators should be set to assess the unit's performance. Finally, in addition to monitoring a database on sub-loans and projects, the apex unit should possess or develop an analytical capability to assess the project's impact on an ongoing basis. This would allow introducing needed changes in the project scope and/or design throughout implementation. Under this project, the apex unit is meant to be a "sunset organization".  19 Table 1: Summary of Assessments A. Achievement of Objectives Substantial Partial Negligible Not Applicable Macro Policies Sectoral Policies Financial Objectives i Institutional Development Physical Objectives Poverty Reduction i Gender Issues Other Social Objectives i Environmental Objectives Public Sector Management Private Sector Development Other (specify) B. Project Sustainability Likely Unlikely Uncertain C. Bank Performance Highly Satisfactory Deficient Satisfactory Identification Preparation Assistance Appraisal Supervision D. Borrower Performance Hlighly Satisfactory Deficient Satisfactory Preparation Implementation Covenant Compliance Operation (if applicable) E. Assessment of Outcome Ilighly Satisfactory Deficient Satiisfactory  20 Table 2: Related Bank Programs, Credits, and Sectoral Studies Credits Purpose Year Status Preceding Operations * Economic Action Program * Restore investment and economic 1984 Completed services for immediate production results. * Economic Rehabilitation Program 1987 Completed (ERP) (1987-89) * Second Rehabilitation Credit * Reduce macroeconomic imbalances and 1987 Completed price, and exchange rate distortions. * Thi rd Rehabilitation Credit * Improve foreign exchange allocation, 1989 Completed budgetary policy, price and distribution policy, reform trade tariff structure, and restructure state enterprises. Following Operations * Industrial Enterprise Restructuring * Restore production and efficiency in a 1989 Ongoing Project selected group of major industrial and agro-industrial enterprises. * Economiic Recovery Credit * Improve foreign exchange allocation and 1992 Completed management, establish a legal and regulatory framework to restructure banking sector and initiate privatization of state enterprises. * Second Economic Recovery * Strengthen fiscal and monetary policy, 1994 Completed Credit support interlinked program of enterprise and financial sector reforms. * Financial Sector Capacity Building * Support the overall enterprise and financial 1994 Ongoing Project reform program. * Third Economic Recovery Credit * Support the privatization of state-owned 1997 Ongoing banks, rationalization of the tariff and indirect tax regime, budget management reform, liberalization of the cashew market, and private concessioning of CFM. Sector Studies * Development of Industrial Policy * Study of broad industrial policy within the 1990 Completed and Reform of the Business context of the changed business environment Environment following the ERP reforms. * Financial Sector Study * Study on financial sector to recommend 1992 Completed improvements and outlines sequence of reforms. * Impediments to Industrial Sector * Study to identify impediments to industrial 1995 Completed Recovery recovery, assess growth potential, and propose strategy.  21 T'able 3: Project Timetable Steps in Project Cycle Planned Actual Identification - Initial IEPS 4/26/88 4/26/88 Final EPS 1/18/88 1/18/88 Preparation Preappraisal Appraisal 4/20/89 4/20/89 Negotiations 10/5/89 10/5/89 Letter of Development Policy (if applicalble) n.a. n.a. Board Presentation 12/20/89 12/20/89 Signing 2/5/90 2/5/90 Effectiveness 6/8/90 6/8/90 Mid-term Review 8/94 8/94 Project Completion 6/30/96 12/31/96 Credit Closing 12/31/96 12/31/97 Table 4: IDA Credit Disbursements - Cumulative Estimated and Actual (in US$ million) FY90 FY91 FY92 FY93 FY94 FY95 FY96 FY97 FY98 Total Appraisal 0.25 4.68 4.5 6.0 6.0 7.0 3.57 - 32.0 Estimate Actual 2.5 0.4 0.7 3.0 6.1 6.9 4.8 7.2 0.4 32.0 Actual as a % of 1,040% 9% 16% 50% 102% 100% 134% - 100% Estimated . _ Date of Final Disbursement April 17,1998 Table 5: Key Indicators for Project Implementation Ky ImplementtinIndict r in >SA R A.-.-l Es0timate Aictua Estimates A. Credit Component 2. No. of Sub-projects 110-120 134 3. Employment generation 6,000 5,029 (ex-ante) 4. Repayment rates on sub-loans 16% B. TA Component 1. Training of PFI Participants 50 50 Table 6: Key Indicators for Project Operation (see above table) Table 7: Studies Included in Project none carried out.  22 Table 8A: Project Costs : . , -Itemns - - - Appraisal Estimates /I Actuial'2 -___.-;___-__.__.__--.____-__.__. (US$ million) ( -lUS$milion) Local Foreign Total Local Foreign Total costs costs costs costs costs costs A. Enterprise Financing Component 10.5 35.57 46.07 56.7 20.2 76.9 IDA - 28.57 28.57 8.6 20.2 28.8 EIB, UTNDP /2 - 7.0 7.0 - - - Enterprises 5.5 - 5.5 22.0 - 22.0 PFIs 3.0 - 3.0 26.1 - 26.1 Government /3 2.0 - 2.0 n.a. - - B. Institutional Strengthening Component 0.2 3.93 4.13 - 3.7 3.7 IDA 3.43 3.43 - 3.2 3.2 UNDF' 0.5 0.5 - 0.5 0.5 Government 0.2 - 0.2 - - - TOTAL 10.70 39.50 50.20 56.7 23.9 80.6 /1 Includes donors' cofmancing. /2 Includes only UNDP. EIB and Caisse Fran,aise provided parallel financing. /3 n.a.: not available. Government's contribution to project costs (salaries of apex personnel, and the operating costs of the unit when it became part of the central bank) was not available for preparation of the ICR. Table 8B: Project Financing | -7-:S,our,ce -- Appraisal Estimates Actual Estimates ;:__ '_E__ -_'_'__ -_.__ ._._'__'_-_' ______(USS million) ( -US$ million) Local Foreign Total Local Foreign Total costs costs costs costs costs costs IDA 32.0 32.0 8.6 23.4 32.0 European Investrnent Bank, UNDP - 7.5 7.5 - 0.5 /1 0.5 Enterprises, PFIs 8.5 - 8.5 48.1 - 48.1 Government 2.2 - 2.2 n.a. - - TOTAL 10.7 39.5 50.2 56.7 23.9 80.6 /1 Includes UNDP only. Table 9: Economic Costs and Benefits Indicator Appraisal Estimate Latest Estimate 1. Line of Credit: ERR - All sub-projects with ERR above 12% not available FRR - minimum 12% not available  Table 10: Status of Legal Covenants Original Revised Description of Covenant Covenant Present Fulfillment Fulfillment Section Type Status Date Date l Comments Credit Agreenmnt 3.04 10 C 06/30/91 - Bank of Mozambique to prepare an action program for separation of central and commercial banking functions and implement it thereafter. 3.05 9 C 12/31/91 - Govemment to complete an action program to settle lending arrears of commercial banks, and implement it thereafter. 3.05 9 CD 06/30/92 06/30/93 Government to complete an action program for the rehabilitation 12/31/94 of the banking system, including its recapitalization, performance benchmarks, and extension of competition, and implement thereafter 3.06a 2 C 3/31/91 - Government and BM to review jointly with IDA progress on Interest rates were brought to positive achieving positive weighted average real lending interest rates. levels in 1991, but fluctuated between positive and negative levels in real terms, until their liberalization in 1996. 4.01 1 C March of - Government to furnish IDA external audit report of statement of 1996 audit received. I I | each year _ [ expenditures and special account transactions. Project Agreement 2.08a I C 03/31/91 - Review of PFIs' accounting capabilities. _ 2.08 b I C 03/31/91 - BM to prepare action program to strengthen PFIs' accounting W capabilities, including work required, resources needed, design of training program, and list of equipment and materials. 2.08c l C 03/31/91 BM to review jointly with IDA progress in implementation of above action program. 3.01b 1 C 09/30/92 BM to furnish IDA certified copies and audit report of PFIs' Not consistently complied with by the I_________ I________ I________ _________ _________ financial statements. PFIs, nor reinforced by BMAIDA. Covenant types: Present status: I . Accounts/audits C = covenant complied with. 2. Financial performance/revenue generation for beneficiaries CD = complied with after delay 3. Flow and utilization of project funds. CP = complied with partially 4. Counterpart funding. NC = not complied with. 5. Management aspect of the project or executing agency. 9. Monitoring, review and reporting. 10. Project implementation not covered by categories 1-9. 11. Sectoral or cross-sectoral budgetauy or other resource allocation. 12. Sectoral or cross-sectoral policy/regulatory/institutional action.  24 Table 11: Compliance with Operational Manual Statements Statement number and title Describe and comment on lack of compliance OD 8.30: Financial Sector Operations Lack of compliance with: - macroeconomic environment - fihancial sector policies, in particular interest rates and targeted subsidized credit - legal and regulatory framework Table 12: Bank Resources: Staff Inputs Stage of Project Cycle Planned (weeks) Actual (weeks) Weeks US$ '000 Weeks US$ '000 Through Appraisal n.a. n.a. 38.3 109.0 Through Board n.a. n.a. 6.4 17.0 Supervision n.a. n.a. 202.3 515.8 Completion 15.0 35.2 15.0 35.2 n.a.: notavailable. Table 13: Bank Resources: Missions /1 Project Activity Month/Year Days in fields No. of Persons Specialized Staff Skills Ratings Pre-apprafisal n.a Appraisal - - - n.a. Supervision 1 11/90 5 2 S Supervision 2 3/91 10 I E S Supervision 3 8/91 5 2 E S Supervision 4 12/91 5 2 E S Supervision 5 11/92 15 3 E S Supervision 6 3/93 10 1 E S Supervision 7 6/93 25 3 E,F S Supervision 8 11/93 25 4 E,F S Supervision 9 7/94 n.a. 2 E S Supervision 10 3/95 15 2 PSD, F S Supervision 11 7/95 5 1 PSD S Supervision 12 12/95 15 3 PSD S Supervision 13 3/96 15 2 PSD U Supervision 14 8/96 10 2 PSD U Supervision 15 3/97 20 1 PSD U Supervision 16 5/97 18 1 PSD U Supervision 17 9/97 19 2 PSD, 0 U Completion 12/97 14 2 PSD, 0 U E = Economist N.A. = not applicable F = Financial Economist U= Unsatisfactory 0 = Operations Officer/Analyst S = Satisfactory CO = Country Officer /I Best estimates based on Form 590s.  25 APPENDIX A AIDE-MEMOIRE SMALL AND MEDIUM SCALE ENTERPRISE DEVELOPMENT PROJECT ,(Credit 2082-MZ) IDA Mission - Implemeintation Completion Report - December 1L997 This is a draft evaluation summary of the Implementation Completion Report on the Small and Medium Scale Project (Cr. 2082-M[Z). [Note: figures are preliminary estimates and subject to revisions.] Obje,ctives The Small and Medium Scale Project was designed to support the Government of Mozambique's policy to develop the industrial sector by stimulating long-term investment, starting with small and medium sized enterprises. The project was an integral part of the ERP and sought to fill a credit supply gap for term financing (mostly for imported capital investments) while reforms of the financial sector were underway. This credit was to be directed to specific sectors through a two-tier mechanism and subsidized by the Government. To complement that objective, the project also sought to rehabilitate the capacity of the banking system to channel resources to productive sectors. Achievement of Objectives Participation and collection performance of the Financial Institutions (PFIs). At the time the Credit was approved in 1989, Mozambique's banking sector was dominated by state- owned banks, mainly BCM and BPD. Since then, the banking sector has been privatized. To the three original PFIs--BCM, BPD and BSTM, two others have recently been added, BIM and ULC. The PFIs' lending share is: BCM (70%), BPD (16%), ULC (8%), BSTM (5%), and BIM (1%). BCM's predominant lending share may be explained by the fact that it was the bank which traditionally serviced the selected sectors. The banks' collection performance is poor for the former state-owed banks and satisfactory for the privately owned PFIs. BCM has the highest concentration of bad loans, since it absorbed the majority of the funds and given its low loan recovery (52% are defaulting). BCM's new management is currently reviewing the situation to decide on a course of action on a case-by-case basis. BPD's performance seems to be improving, and the bank's new management has started to take legal action against defaulting firms. As of now, approximately US$32.5 million of the total IDA Credit has been disbursed, leaving of a balance of about US$3.3 million. The Credit is expected to be fully disbursed at the time of closing.  26 The causes of BCM's and BPD's arrears situation vary, ranging from long average processing time for subprojects (it takes about 6-9 months from the time a client presents a subproject proposal to a bank for financing until the subproject actually starts implementation, resulting in rescheduling of grace periods which increases the client's interest payments), inadequate technical appraisal of subprojects, irregular supervision, and no efforts in loan collection from non-paying customers when the banks were still state-owned. In the case of BPD, this was reinforced by the bank's former management's decision not to invest any of the bank's own funds in a subproject. The processing time situation applies not only to BCM and BPD, but to the other PFIs as well. Here too, the causes are many, ranging from the country's own customs procedures to lengthy import process from abroad, and also the Bank's own procurement and disbursement requirements. The DCA should have included monitoring and performance indicators, a clearly stipulated maintenance of loan collection ratio and conditions under which disbursements from the line of credit would be suspended, and action plans for improved performance. The fact that the Central Bank debited the banks' account automatically was not an effective measure, in and of itself, to ensure that the banks adopt a more commercially oriented approach to credit-risk management, as BCM and BPD were state-owned banks. Firm-level and Subproject Achievements. An estimated 61% of the sub-loans were made for new projects, 31% for rehabilitation, and 8% for expansion. It is difficult to assess if the subprojects have achieved their projected economic and financial rates of return because none of the PFIs calculates ex-post rates of return. The same applies for job creation, so that any assessment to estimate the social impact of the project in terms of employment generation would be based on estimates at the time of subproject appraisal. There is nonetheless some consensus among the PFIs and the PMU that the social impact objective has been met to some extent, and that the bad loans situation of BCM and BPD, by far the problem banks, does not necessarily mean that defaulting firms are not operating or even profitable. Sectoral and Regional Impact. The sectoral distribution is as follows: 46% of total lending went to industry, 14% to transport, 13% to services, 7% to tourism, 9% to fishing, 5% to agro-industry, 2% to mining, and 4% other. In terms of regional distribution, Maputo benefited from about 70% of total lending, Sofala 12%, and the remaining 18% went to the other regions. An attempt was made during project preparation to reach the Beira region with the establishment of one of the two small business advisory services units. This did not materialize due in part to IDIL's poor performance and the resulting low demand for its services. On-lending Rates, Credit Ceilings, and Government Subsidies. Under the terms of the project, on-lending rates were to be variable market rates, the Government was to bear the foreign exchange risk, and the commercial banks the credit risk on on-lent funds. During project negotiations, the Government agreed to review annually with the Bank progress on achieving positive real on-lending interest rates, terms and conditions. A controlled lending rates and foreign exchange rates regime prevailed in Mozambique until 1996. From 1990-94, commercial lending rates could not exceed 48%, and 46% from 1994 until the liberalization of interest rates in 1996. Those rates were below inflation rates. At the same time, the Government imposed credit ceilings on the banks. The Government gave the IDA line of credit, as well as other donors' lines of credit, a preferential rate below the market lending rates to induce firms to borrow from those lines. The banks used the line of credit to either complement their credit restrictions, or more recently to have access to local currency in short supply. However, the IDA  27 line of credit did not prove attractive at first to the commercial banks because the spread was not high enough to cover their administrative costs for the sub-loans, after the PMU had retained its percentage for management fee and to cover the Government's "insurance" on foreign exchange fluctuations. The split was later changed to 50:50, when the Bank complained of slow disbursements on the credit line. Any benefits derived from this project should be weighted against the implicit costs to the Government associated with the subsidies on interest rates and foreign exchange. The need to subsidize those two rates to spur investment may have been warranted at the time the Small and Medium Enterprises project was designed, but should be reassessed in the current economic context. Since the Government may abscrb all or part of the losses of the previously state-owned banks as part of the negotiations of the privatization process, that cost should also be added to the subsidies costs. PMU/commercial banks relationship, and PMU's performance. The PMU maintains a good relationship with the commercial banks and assumes its responsibility of record checking and keeping satisfactorily. However, the PMU should have been more active in ensuring that it receives regularly the supervision reports from the commercial banks, especially when the banks stopped sending them altogether. World Bank's Special Account (S'A) and procurement procedures. The SA was opened with Citibank-New York for the letters of credit for imports. As already mentioned above, this arrangement was one of the factors which added to the lengthy process. Another factor was the Bank's three quotations requirement, since in the majority of cases when this requirement applied, there was only one local supplier, and in order to comply, the firm had to inquire outside of Mozambique, often in neighboring countries, which led to additional delays and costs. The Bank's internal procedures should be flexible enough so that they easily adapt to the needs of the private sector. As they stand now, those internal procedures are too slow and cumbersome for the fast pace characteristic of the private sector, and the client is the only one to ultimately bear the cost. Maputo, December 1997  28 APPENDIX B SMALL AND MEDIUM SCALE ENTERPRISE DEVELOPMENT PROJECT (Credit 2082-MZ) Borrower's contribution to the ICR (translated version) IMPLEMENTATION REPORT ON THE SMALL AND MEDIUM SCALE ENTERPRISE DEVELOPMENT PROJECT (PDPME)' Introduction The Small and Medium-Scale Enterprise Development Project was approved in 1990 as part of ithe series of current economic and social policy measures designed to rehabilitate the national economy. The specific objective of the project was to ensure the active participation of economic operators throughout the country, which was to be promoted via the government's plan to offer a system of incentives to small and medium-sized businessmen and small and medium- scale enterprises to make them more dynamic. The project was intended to finance the mining, agro-food, and manufacturing industries, and the tourism, transportation, equipment repair, canning, and services sectors. Since this project ended on December 31, 1996, negotiations were undertaken with the World Bank to extend the period for using this line of credit for another year, to enable the project disbursements to be completed. This report aims to provide a comprehensive survey of the impact of the project during its implementation period. 1.1. Project Objectives The overall project objectives consisted essentially in the following: * Increase domestic production and reduce dependency on the outside world, by import substitution and the growth and diversification of exports; * Increase the number of jobs, and thereby reduce delinquency, especially on the periphery of urban areas; * Renovate, modernize, and expand the installed productive capacity and facilitate the emergence of new industries; * Mobilize private savings, and promote their use in productive sectors; * Inform and educate the country's business community, the only way to guarantee that they will play an active part in the national reconstruction effort; * Improve the performance of institutions capable of providing support for small and medium scale enterprises, and especially public institutions and the banking system; * Identify the sectoral economic policy instruments best suited to attain the goals indicated earlier, among others. 1 Includes all donors' lines of credit.  29 1.2. Sources of Financing To finance the Project in question, the government, through the Bank of Mozambique, received assistance from international financial institutions, and especially the World Bank and IDA, the European Investment Bank (EIB), and the Caisse francaise de developpement (CFD). * On February 7, 1990, the World Bank made approximately US$36 million (SDR 25.1 million) available for this purpose; * As of July 18, 1990, about US$7 million (ECU 6 million) were provided by the EIB as cofinancing. Use of these funds was canceled on July 18, 1994, in view of the fact that the line of credit had barely been used. The primary reasons why these funds were not used to a greater extent are as follows: (i) The length of the process of evalualting and approving subprojects; it took nearly a year and a half to approve a subproject, and this created serious problems for the Bank of Mozambique in its dealings with commercial banks and their customers; (ii) the EIB funds covered. only 50% of the overall needs of each subproject, which meant that alternative sources always had to be found to cover the remainder; - The CFD made about US$9.5 million (FF 50.0 million) available to finance the foreign currency component for support for private sector investment. The use of those funds was suspended on March 11, 1993. This action was taken on the heels of the decision made by the French government to suspend credit operations to poor countries, as a way of halting the continuous growth in the external debt service of those countries. 1.3. Operational Instruments To achieve these objectives, the Bank of Mozambique, in its capacity as the central bank, was designated as the institution responsible for implementing and coordinating the project. An appropriate structure (the former UGP), abolished in February 1994, was created for the purpose, and was later incorporated into the DOC, forming the Investment Projects Core Group [Nuicleo de Projectos de Investimento] (NUPI). NUPI spared no efforts, whelther in organizing courses to train credit analysts for commercial banking, or promoting senninars to circulate information on the available lines of credit in the business community in Maputo as well as in the rest of the provinces. FACIM and the mass media also served as a vehicle for disseminating information. Commercial banks were invited to play a major role. They proved to be increasingly active participants in the economic recovery process. To this end, the following banks signed a contract of participation in the line of credit referred to: the Banco Comercial de Mocambique (BCM); the Banco Popular de Des envolvimento (BPD); the Banco Standard Totta de Mocambique (BSTM); the Banco de Fomento e Exterior (BFE); the Banco Portugues do Atlantico, which is now part of the BIM; the Banco Internacional de Mocambique (BIM); the ULC Mocambique (Leasing Company); and Credicoop, which is responsible for selecting the best customers and the most deserving subprojects to receive financing. The Institute for the Development of Local Industry (IDIL) has been receiving specific assistance under the Project. This has enabled it to provide these and other services to small and  30 medium-scale enterprises. This does not mean, however, that other firms providing technical and financial assistance cannot operate in this area at the same time. 1.4. Project Evaluation Seminar On July 4, 1994, a Seminar was held in Maputo in the form of an informal brainstorming session, to accomplish the following objectives: (i) to assess the effectiveness of the line of credit and to determine how its accessibility and use could be improved; and (ii) to identify ways to improve the business climate, especially for small and medium-scale enterprises. The seminar was organized at the request of the World Bank, following a study of the national business community in the form of a survey prepared and conducted by Eduardo Mondlane University in October and November 1993. The purpose of the survey was to identify the priimary problems encountered by SME's and how those problems affected their ability to set up in business and operate. This would make it possible to determine systematically the princilpal constraints they encounter. Of the 125 businesses interviewed that (i) had submitted bids for the line of credit for support to SME's and (ii) were listed in the survey provided by the National Planning Commission (CNP), 39 were chosen at random from the survey provided by the CNP, 58 were selected on the basis of the information on bidders for the line of credit, and 28 were businesses in the informal sector. In the conclusions of the document, the obstacles that were found to have a serious adverse effect on small and medium scale enterprise development were divided into three categories of descending importance: (i) In the first group, the obstacles regarded as most serious were financial problems and problems related to inflation and price volatility. As far as financial problems are concerned, difficulties in gaining access to working capital were the primary constraint for these businesses, followed by access to investment financing and problems encountered in loan negotiations with banks. As for problems related to price instability, the businesses interviewed indicated that the greatest obstacles were the high costs of raw materials and capital replacement. (ii) The businesses put the following problems in the second category: * Expenses incurred as a result of taxes and mortgage recording fees; * The uncertainty surrounding economic and political strategies; * Access to foreign currency. Heavy emphasis was placed on macroeconomic problems. The businesses interviewed indicated that economic stability and the functioning of the financial system were of critical importance to development of the private sector in Mozambique. It is difficult for SMEs affected by these issues to get beyond them in order to focus on matters related to their own operations. The importance attached to access to foreign currency, in the context of the current economic liberalization policy, could be interpreted as a difficulty in obtaining financial resources to pay for imports of raw materials; (iii) The problems ranked in the third group had to do with procurement, infrastructure, and technology. Problems related to regulations and laws governing the operations of SME's were also placed in this category.  31 Despite the seminars organized by NUPI (the former UGP) to disseminate information in certain provinces, they seem to have had little impact in business circles. The poor performance of commercial banks in providing information on this facility to the public is also largely to blame for this situation. The economic sector which has made the most active use of the proceeds of the Credit has been industry, since 154 of the applications filed, or 53.55% of total demand, was from that sector. It is followed by the transportation sector with 46, services with 41, fisheries with 29, tourism with 22, agro-industry with 17, mining with 8, and other sectors accounted for 13 applications. In terms of amounts of financing, industry absorbed US$65.4 million, followed by the transportation sector at US$17.9 million, services at US$12.1 million, tourism at US$10.3 million, fishing at US$5.8 million, agro-industry at US$2.9 million, mining at US$2.4 million, and others at US$5.0 million. Applications to finance new businesses or ventures numbered 187, which put this category in first place with 40.26% of total requests. One hundred and three (103) of the applications, or 36.34%, were to renovate already existing enterprises, and 40 requests were submitted to expand operations, equivalent to 10.69% of the total. 1.6. Disbursements The following table shows the actual behavior of yearly disbursements up to December 1997. Table 1.3. Comparative Chart of Proiected and Actual Disbursements (in millions of US$) Years I1990 1991 1992 1993 1994 1995 1996 1997 Total Actual 2.5 0.6 1.0 5.9 5.6 5.2 8.7 4.0 33.5 The relatively low level of disbursements shown in Table 1.3 is attributed to the following factors: (i) The inefficiency of the specialists involved in all levels of the import process, as a result of their lack of experience during the first two years of the project; (ii) the complexity of the procedures to be performed, especially with regard to procurement for operators located outside the city of Maputo; and (iii) various bureaucratic procedures responsible for delays both in approval of letters of credit by local and foreign banking institutions and in delivery of the goods by suppliers outside the country. 1.7. Repayments Of the 147 subprojects approved, 137 are already in the collection phase. Of the remaining 10, 4 are in the grace period and 6 have been canceled. Generally speaking, the loan recovery rate has not been satisfactory. The following table presents a general picture of the debt service, by bank:  32 Table 1.4. Current Debt Service Status of Commercial Banks (Dec. 1996) Projects/Banks BCM % BP % BSTM % ULC % BIM % Total approved 102 100 24 100 8 100 I 100 I 100 Paying regularly 12 11.8 5 20.8 2 25 0 0 0 0 Paying irregularly 11 10.8 3 12.5 0 0 0 0 0 0 Defaulting 59 57.9 12 50 0 0 0 0 0 0 In grace period 4 3.9 1 4.2 6 7.5 1 100 1 100 Rescheduling 10 9.8 2 8.3 0 0 0 0 0 0 requested Cancelled ?6 5.8 1 4.2 0 0 0 0 0 0 The reasons for the poor recovery rate vary, but the following are some of the primary ones: (i) Poor performance on the part of certain commercial banks, especially in monitoring during the period following approval of the subprojects; (ii) Inability of customers to pay, as a result of the change in the business climate (lack of a market for finished products); (iii) Deliberate default by customers as a result of diversion of their earnings to other purposes (in some cases, they used their revenue for re-investment); (iv) Deliberate refusal to pay (bad faith), on the basis of allegations that the funds were from the World Bank and so no payments were owed to the bank in question (this is related to the situation indicated in (i)). As a result of these factors, banks, including the BCM and the BPD, were already prevented from using all the lines of credit financed by the World Bank, unless they achieved at least (i) a 10% recovery rate, in order to conclude any applications they may have pending; and (ii) 90%, in order to once again be eligible for the lines of credit. With a view to minimizing the portfolio of non-performing loans associated with this line of credit, commercial banks decided to pursue one of the following courses of action: (i) go to the courts and file suits: (ii) invite their customers to reschedule their debt payments according to the payment capacity of each customer; and (iii) promote activities to make customers aware of the need to honor their commitments with banks. However, in accordance with the terms of the participation contracts, regardless of the performance of the customers or final beneficiaries in making timely payments of the amounts due, the commercial banks in question must reimburse the Bank of Mozambique the capital owed in addition to the interest payable on the due date. Thus, as of December 31, 1997, the Bank of Mozambique was charging the following amounts for annual interest (in millions of meticais): Table 1.5. Trend in Interest Rates Charged by the Bank of Mozambigue (in millions of meticais) 1991 1992 1993 1994 1995 1996 1997 Total BCM 0 0 22 2 259 2 584 22 051 22 484 49 380 BPD 0 0 55 244 388 4 382 10 230 15 209 BSTM 0 0 0 174 230 118 611 1 131 BIM 0 0 0 0 0 0 0 0 ULC 0 0 0 0 0 0 719 719 Total 0 O 77 2 677 3 202 26 548 34 024 66529  33 1.8. Special Account The special account was closed on March 31 of this year. The unused balance of US$853,592.43 was returned to the World Bank on March 26, 1998. The amounts incorrectly credited and debited to this account, which balance out to US$32,118.90, will be transferred to the account of the exchange fund at Bankers Trust in New York. 1.9. Accounting of the Operations Generally speaking, all operations are entered in the accounts and reconciled regularly, on a monthly basis, with the DCO. The accounts were audited by external auditors, Ern'st &Young, up to December 31, 1996. 1.10. Social Impact According to statistics for investment applications submitted during the period under analysis, at least 4,340 new jobs were created, and 12,102 workers were guaranteed work as a result of the rehabilitation of the companies where they were already employed. The average amount in foreign currency requested for each subproject is US$250,000. With regard to the demand for this line of credit and how it was distributed, the most frequent applicants were businesses with 50 or less employees, as shown in Table 1.6. below: Table 1.6. Distribution of Demand by Beneficiary Businesses (By size or number of employees) i Size of Staff No. of Businesses % of Funds Allotted From 0 to 50 employees 107 73 From 51 to 100 employees 23 16 From 101 to 200 employees 12 8 Over 200 employees 5 3 Total 147 100  34 APPENDIX C SMALL AND MEDIUM SCALE ENTERPRISE DEVELOPMENT PROJECT (Credit 2082-MZ) Cofinanciers' contribution to the ICR Assessment of sub-projects financed by the Caisse Francaise de Developpement * Transports Cassamo: Provided parallel financing in the amount of FF 3,610,200 to add six Renault trucks to the truck fleet of a freight company to meet an increasing demand. The company also received financing from the World Bank project in 1992 to buy four Volvo vehicles. Assessment: The objectives of the project have been met, and the project is being implemented on schedule, despite the closing of the Volvo auto shop in Maputo. * Linga Turist Lta.: Provided parallel financing in the amount of FF1,500,000 to build an ocean-front resort on the peninsula of Linga Linga. Assessment: The project did not achieve its objectives and implementation would be completed only if the investor can find 1,300,000 rands to complete works. Major factors contributing to the unfinished project can be attributed to legal weakenesses in the establishment of the two shareholding companies and in the equity contribution, as well as the total lack of supervision of the works. The investors are for a large part responsible for the failed project because of their inabilily to meet the financing requirements which were too ambitious. This project, which was financed through an inefficient banking system, is testimony to the limitations of the exercise. Despite the availability of virtually risk-free foreign exchange which was then in very short supply, the project did not achieve its objectives. * Hotel Moqambique: Provided parallel financing in the amount of FF 3,800,000 to renovate the Hotel Macambique. Assessment: An assessment of the financial viability of the project cannot be made at this stage because of lack of accounting and financial statements. Also, the project is still in the grace period phase. * Ginwala: Provided parallel financing in the amount of FF 2,275,622 to help this oil- producing company establish a unit to manufacture plastic bottles for the cooking oil to a more marketable size. Assessment: The project objectives were achieved.  35 APPENDIX D SMALL AND MEDIUM SCALE ]ENTERPRISE DEVELOPMENT PROJECT (Credit 2082-MZ) Borrower's comments to draft ICR (translated version) Maputo, June 8, 1998 TO: Private Sector Finance (Fax No.: (202) 522-1198) Africa Regional Office World Bank Washington, D.C. Attn: Gabrielle M. Rooz FROM: Bank of Mozambique, Maputo Credit Operations Department Investment Projects Bureau (Fax No.: (258) 1 42 34 65) RE: COMMENTS: THE WORLD BANK REPORT ON THE PDPME We have received the Implementation Report on the Small and Medium-Scale Enterprise Development Project (PDPME), in response to which we wish to make the following comments. Overall, the report is very thorough, and presents the main problems encountered subsequent to the signing of the Credit Agreement, approval of the subprojects, and the implementation of the latter up to recovery of the subloans. In particular, it identifies some key lessons learned from the project, and these will be of great help in improving our operations in the future. We have identified some discrepancies, such as the following: The report quotes the Credit numiber as 1794-MZ, instead of 2082-MOZ. There are also a number of mistakes, such as the following: The World Bank states that the closing date was December 1998, even though the year 1998 is not yet over. Furthermore, closure was in December 1997, a year later than the date originally scheduled (December 1996). The European Investment Bank (EIB) and the Caisse francaise de developpement (CFD) financed eight subprojects, not seven as the report states, even though one of these was financed simultaneously by EIB and IDA (i.e. that relating to UGC).  36 There are also some discrepancies in amounts, but these are not very important, because, whereas your report bases its analysis on PDPME subprojects financed exclusively by the World Bank, the NUPI report reflects the overall position, involving as it does three international financial institutions: the World Bank (IDA), the Agence frangaise de developpement (AFD, the former Caisse fran9aise de developpement, or CFD), and the European Investment Bank (EIB). This accounts for the differences in the numbers of subprojects approved, workers employed and jobs created, the regional distribution of subprojects, their breakdown by type of economic activity, their classification into "new"v, "rehabilitation", or "expansion", etc. In our view, the NUPI version is accurate, and this is reflected in the title of the report: Relat6rio de Execuqdo do PDPME ("Report on the Implementation of the PDPME"); i.e. the project in general, rather than "Report on the Implementation of the PDPME (IDA/World Bank)". However, if the World Bank does decide that NUPI should prepare a report refening exclusively to the IDA-financed aspects of PDPME, it should call for it and we will produce it, although at the time that Ms. Gabrielle M. Rooz was preparing the World Bank report, she asked us for the addresses of EIB and CFD, and we inferred from this that the approach she was adopting was general and not restricted, because otherwise it would have made no sense to forward the report to the cofinancing agencies. We find it difficult to comment on the percentage (82%) that the World Bank quotes for arrears unrecovered by BPD (Banco Popular de Desenvolvimento) and BCM (Banco Comercial de Mocambique), because since December 1996 we have received no information from these banks regarding their lending from this and other lines of credit. As regards the relationship between the Bank of Mozambique and the commercial banlcs, the repayment rate for the Credit proceeds has been 100%, since this is safeguarded under the Participation Contract, which authorizes BM to debit the accounts of these banks on the days that repayments fall due. These are all the comments we wish to make on the report. Sincerely, /s/ [Illegible] Mrs. Irene C. Mauricio