Tunisia - SOFOMECA Foundry Modernization and Expansion (Fourth Industry) Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 9740 PROJECT PERFORMANCE AUDIT REPORT TUNISIA SOFOMECA FOUNDRY MODERNIZATION AND EXPANSION PROJECT (LOAN 2301-TUN) JUNE 26, 1991 Operations Evaluation Department 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. ACRONYMS AND ABBREVIATIONS AIIC - Arab Industrial Investment Company ARMICO - Arab Mining Company BDET - Banque de Ddveloppement Economique de Tunisie BTEI - Banque de Tunisie et des Emirats d'Investissements BTKD - Banque Tuniso-Koweitienne de D6veloppement BTQI - Banque de Tunisie et de Qatar d'Investissements CKD - Completely Knocked Down CMT - Complexe M6canique de Tunisie EMI - Electro-Mechanical Industries GF - George Fischer Aktiengesellschaft GM - General Motors ICM - Industries Chimiques Maghr6bines IDB - Islamic Development Bank KFW - Kredit-Anstalt fur Wiederaufbau KHD - Kloeckner-Humboldt-Deutz OED - Operations Evaluation epartment PCR - Project Completion Report PMG - Project Management Group PPAR - Project Performance Audit Report RVI - Renault Vhicules Industries SAR - Staff Appraisal Report SOFOMECA - Soci6td de Fonderies et de M6canique STIA - Soci6td Tunisienne d'Industries Automobiles STUSID - Soci6td Tuniso-S6oudienne d'Investissement et de D6veloppement tpy - Tons Per Year CUPENCY EQUIVALENTS At Project Appraisal (1982) - 1 Tunisian Dinar (TD) - 3$1.685 November 1990 - 1 Tunisian Dinar (TD) - U;$l.220 WEIGHTS AND MEASURES 1 metric ton (t) = 1,000 kilograms (kg) 1 metric ton (t) 2,204 pounds (lb) FISCAL YEAR Government and SOFOMECA: January 1 to December 31 FOR OMFCL4AL UE OMLY THE WORLD BANK Washington. D.C. 20433 U.S.A. Oce of DIectOCnWal Ope"atinn Evalimun June 26, 1991 MEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECT: Project Performance Audit Report on Tunisia- SOFOMECA Foundry Modernization and Expansion Project (Loan 23o1-WN Attached, for information, is a copy of a report entitled "Project Performance Audit Report on Tunisia - SOFOMECA Foundry Modernization and Expansion Project (Loan 2301-TUN)" prepared by the Operations Evaluation Department. Attachment This document has a restricted distribution and may be used by recipients only in the performance or their official duties. Its contents may not otherwise be disclosed without World Bank authorization. FOR OFFICIAL USE ONLY PROJECT PERFORMANCE AUDIT REPORT TUNISIA SOFOMECA FOUNDRY MODERNIZATION AND EXPANSION PROJECT (LOAN 2301-TUN) TABLE OF CONTENTS Preface . . . . . . . . . . . . . . . . . . . . . . . . . Basic Data Sheet . . . . . . . . . . . . . . . . . . . . . . . . . . in Evaluation Summary . . . . . . . . . . . . . . . . . . . . . . . . . vii I. INTRODUCTION................... ..... . . 1 II. OBJECTIVES.................. ..... . . . . 4 III. IMPLEMENTATION EXPERIENCE......... ..... . . . . . 5 A. Implementation Problems......... ..... . . . . 5 a. Implementation Delays........ ..... . . . . 5 b. Marketing . . . . . . . . . . . . . . . . . . . . . . . 5 c. Technical Problems . . . . . . . . . . . . . . . . . . 7 d. Organization and Manning.......... .... . . 7 e. Control Systems and Services........ .... . . 8 f. Financial Problems............. .... . . 8 B. Reasons for Implementation Problems. ......... . . . 9 a. Industrial "Integration" Strategy and the EMI Sector 9 b. Excessive Reliance on a few Key Customers (STIA and CMT) 12 c. Role of the Banks........... ..... . . . 14 d. Management Weaknesses...... ..... . . . . . . 16 e. Consultants Problems........ .... . . . . . . 17 IV. RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 A. Production Capacity . . . . . . . . . . . . . . . . . . . . 18 B. Productivity . . . . . . . . . . . . . . . . . . . . . . . 19 C. Quality . . . . . . . . . . . . . . . . . . . . . . . . . . 21 D. Output and Sales . . . . . . . . . . . . . . . . . . . . . 21 E. Exports . . . . . . . . . . . . . . . . . . . . . . . . . . 22 F. Project Costs . . . . . . . . . . . . . . . . . . . . . . . 23 G. Financial Results . . . . . . . . . . . . . . . . . . . . 23 V. SUSTAINABILITY . . . . . . . . . . . . . . . . . . . . . . . . 23 A. Closure . . . . . . . . . . . . . . . . . . . . . . . . . . 25 B. Sale . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 C. Continuation as a Company . . . ... . . . . . . . . . . . . 26 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. TABLE OF CONTENTS (cont'd) Page. No VI. FINDINGS AND LESSONS . .27 A. Findings ........................ . . . . ........27 a. Market Issues . . . . . . . . . . . . . . . . . . . . 27 b. Management Issues ................. . 28 B. Lessons . . . . . . . ................ . . . 28 Aw Borrower's Comments . . . . . . . . . . . . . . . . . . . . . . . . . 31 PROJECT PERFORMANCE AUDIT REPORT TUNISIA SOFOMECA FOUNDRY MODERNIZATION AND EXPANSION PROJECT AN 2301-TUN) 1. This is a Project Performance Audit Report (PPAR) on SOFOMECA Foundry Modernization and Expansion Project, Loan 2301-TUN, in the amount of US$16.8 million to Socidt6 de Fonderies at de Mdcanique (SOFOMECA), with the guarantee of the Government of Tunisia. The Loan was approved on June 2, 1983 and became effective on April 16, 1984. The original closing date of December 31, 1986 was postponed to June 30, 1988. The Loan was fully disbursed on January 31, 1989. 2. The PPAR was prepared by the Operations Evaluation Department (OED) and is based on the Project Completion Report (PCR) prepared by the Europe, Middle East and North Africa Regio i (EMENA) of the Bank and issued in 1990,1; the Staff Appraisal Report (SAR), the President's Report, the loan documents, the summary of the Executive Director's meeting at which the Project was considered, a study of the project files and discussions with Bank staff. An OED mission visited Tunisia in November 1990. That mission reviewed, inter_alia, the effectiveness of the Bank's assistance with SOFOMECA's management, Government's officials and development banks' representatives. Their kind cooperation and valuable assistance in the preparation of this report is gratefully acknowledged. 3. The PCR provides a good account and assessment of the project experience, and discusses the performance of the Bank and the borrower. The PPAR elaborates on particular aspects such as the critical issues of concern that led to the genesis of this loan in conjunction with other Bank loans financing electrical and mechanical industries in Tunisia; evaluates the problems encountered in implementing the various components of the project; and ascertains the key factors that determined the loan's outcome and sustainability of the project. The PPAR, then, draws further lessons from the project experience. 4. The draft PPAR was sent to the Borrower for comments. The comments reveived from SOFOMECA were taken into account in finalizing the report and are reproduced as an Annex. 11 Project Completion Report, Tunisia - SOFOMECA Foundry Modernization and Expansion Project (Loan 2301-TUN), Report No. 8456, March 21, 1990. - Lii - PROJECT PERFORMANCE AUDIT REPORT TUNISIA SOFOMECA FOUNDRY MODERNIZATION AND EPANSION PROJECT (LOAN 2301-TUN) BASIC DATA SHEET LOAN POSITION (Amounts in US$ Million) As of April 30. 1991 Credit Original Disbursed Cancelled RRa Outstanding 2301 16.80 16.75 0.05 5.83 10.93 gUUATIVE ESTIMATED AND ACTUAL-DISBURSEMENTS EM& EIR9 Appraisal Estimate (US$M) 1.60 16.80 Actual (US$M) 0.58 16.751' Actual as % of Appraisal (%) 96.1% 99.7% Date of Final Disbursement: January 31, 1989 PROJECT DATES Oginal Actual Project Brief 11/24/81 11/24/81 Yellow Cover 03/25/82 10/25/82 Negotiations 04/18/83 04/18/83 Board Approval 06/02/83 06/02/83 Signing 06/16/83 06/16/83 Effectiveness 12/31/83 04/16/84 Loan Closing 12/31/86 06/30/88 The rest of the committed funds ($50,000) was cancelled. * iv - STAFF INFUTS (staffweeks) E = EI E8A EDI EISA EDI EHi EX82 WjQ TD.TAL Preappraisal 2.1 11.0 - . - - - - 13.1 Appraisal - 49.0 - - - - - - 49.0 Negotiations - 4.4 - - - - - - 4.4 Supervision - 0.2 14.0 27.3 5.4 10.7 5.8 16.9 0.5 80.7 Total 2.1 64.6 14.0 27.3 5.4 10.7 5.8 16.9 0.5 147.2 MISSION.DAT No. of No. of Staff Date of Month/Year Weeks Persons wnso Report Preparation Appraisal 12/79 1.0 1 1.0 11/80 0.8 1 0.8 04/81 1.4 3 4.2 06/81 2.0 1 2.0 11/81 2.0 2 4.0 12/81 0.4 1 0.4 04/82 1.8 3 5.4 07/82 0.8 2 1.6 Supervision I 10/83 1.0 2 2.0 11/83 Superrision II 05/84 1.0 2 2.0 05/84 Supervision III 04/85 1.0 1 1.0 04/85 Supervision IV 03/86 1.6 2 3.2 04/86 Supervision V 07/86 2.0 2 4.0 08/86 Supervision VI 07/87 1.4 1 1.4 08/87 Supervision VII 06/88 2.6 2 5.2 06/88 . v * OTHER PROJECT DATA Related Projects: Borrower/Executing Agency: BDET Project: Electrical and Mechanical Industries (EMI) Loan No.: 2113-TUN Amount: US$30.5 million Approval Date: 12/02/82 Borrower: Republic of Tunisia Project: S-cond Electrical and Mechanical Industries (EMI) Loan No.: 2554-TUN Amount: US$54.0 million Approval Date: 06/04/86 Project: Industrial and Trade Policy Loan Loan No.: 2781-TUN Amount: US$150.0 million Approval Date: 08/05/87 - vil - PROJECT PERFORMANCE AUDIT REPORT TUNISIA SOFOMECA FOUNDRY MODERNIZATION AND EXPANSION PROJECT (LOAN 2301-TUN) EVALUATION SUMMARY Introduction 1. In 1983, a Bank loan was made 3. The SOFOMECA Project was to Socidtd de Fonderies et de implemented against a deteriorating M6canique (SOFOMECA), a foundry economic background, particularly established in 1964 as part of a after 1985. The agricultural and maintenance workshop attached to the industrial sectors performed poorly, Tunisian railway company. SOFOMECA and the automotive and tractors became a separate Government-owned subsectors, which were to absorb the company later on and sold iron and largest part of SOFOMECA's increased steel castings a variety of production of castings, experienced a industrial sectors, mainly in the number of problems and plant Tunisian market (para. 1). closures, with a direct negative impact on SOFOMECA. 2. In the late 1970s, the Tunisian authorities launched a program to Ojetv foster higher labor productivity and increased local value added content 4. In the framework of the general through a more integrated development policy to support the electrical and of the industrial sector. The mechanical industries sector, the Electrical and Mechanical industries basic objective of the SOFOMECA (EMI) (including foundries) were Project was to contribute towards the thought to be well placed to achieve development of the Tunisian foundry these objectives because they were industries by: M meeting labor-intensive and, although small efficiently the growing domestic and insufficiently integrated, demand for foundry products; (ii) offered prospects of increasing improving the quality, productivity, technical competence and quality. It and cost competitiveness of castings; was felt that the foundry industry and (iii) creating a sound basis for needed particular support if it were exports (para. 17). to become the "backbone" for the development of the Tunisian 5. Basically, the Project had engineering industries. SOFOMECA tbi elements: was, by far, the largest foundry in Tunisia, and was selected as the i A modernization/expansion project which would serve this program for the existing iron and purpose. In support of this effort, steel foundry to increase its the Bank approved a US$16.8 million capacity from 5,200 to 7,500 tons per loan to SOFOMECA in June 1983 (paras. year. 2-16). - viii - * Construction of a new iron the root of the considerable foundry with a single-shift capacity difficulties which arose later on and of 7,800 tons per year, or a two- led to Implementation problems sh'4t capacity of 12,000 tons per (paras. 83-84). yea. 9. Co-financiers and shareholders * Installation of new sand played a rather passive role during treatment plant with a 20,000 tons of implementation, at least until 1987; sand/year capacity (para. 18). but even when they realized the depth of the project difficulties were not Implementation Expriencc able to influence SOFOMECAs management and the course of events. 6. Implementation issues have They intervened too la- (paras. 70- arisen in six areas: (a) 78). implementation delays; (b) marketing; (c) imbalances in the output of the different sections of the new foundry and other serious technical problems; 10. Most of the project's (d) poor project management objectives have not been achieved. performance and severe disputes with expatriate consultants; (e) lack of a LiLtdPoUction GaRacity: preventive maintenance program; and Imbalance in the new foundry limits (f) catastrophic financial situation the two shift production to 12,400 due to major marketing problems, tons per year, while the steel heavy indebtedness, and cost overruns foundry effective capacity is only (paras. 22-45). 1,800-2,000 tons of acceptable quality castings. 7. The project originally assumed that SOFOMECA' s market was to be Low Productivity: Productivity largely connected to the demand for remains at a very low level. castings from a new diesel engines SOFDMECA's new foundry has a and tractor company (CMT) and from an productivity (in man-hours/ton) which automotive assembly plant (STIA) is less than half of European which would use mechanical castings. foundries, while the old semi- Due to the economic crisis and to a mechanized foundry has a productivity drastic change in Government policy almost four times lower than similar which led to import liberalization of European foundries. transport equipment, these projects never materialized, with an extremely L uali&: Quality (judged negative impact on SOFOMECA's new by the percentage of scrapped foundry project (paras. 58-69). castings in total output) remains abnormally low, az least in the old 8. A Project Management Group was foundry. However, quality has been set up but remained weak because of improving in the new foundry. lack of experience, particularly in planning and scheduling (para. 79). L Outpu: Output has been Project management and Bank staff significantly below original could not decide on the need to projections, and capacity utilization remove a principal expatriate reached 57% only in 1990. consulting firm which did not perform as expected. Major disputes with JJDnrofitable EXRorts: SOFOMECA this consulting firm may have been at has been able to increase exports, - ix - mainly to Europe, from 500 tons in export-oriented new foundry. The 1983 to some 3,900 tons in 1989, 4,068 tons figure fits with the because of the pressing need to find current estimated 5,000 capacity of export outlets to improve capacity the old iron foundry (para. 115). utilization--but at unprofitable export prices. At current export 13. By normal financial criteria, prices, SOFOMECA cannot break even. SOFOHECA should close down operations because of its high level of * Catastrophic Financial indebtedness and its inability to Sltuation: SOFOMECA's current regain profitability for, at least, financial position is extremely 3-4 years because of high production serious due to low sales and high cost, low export prices, and only debts. Net losses have occurred slowly increasing sales. If it is to every year since 1984, and totalled continue operations, a major TD 16.9 million in 1983-1990. refinancing exercise will be SO1,OMECA must repay TD 41.6 million necessary as part of an overall before the end of 1994. It is unable rescue plan (para. 116). to repay its loans, and the Bank now receives - 'ments from the Tunisian 14. Factors which argue in favor of State who -.ranteed the Bank's loan. SOFOMECA continuing operations are: In the interim period, it will be necessary to finance severe operating * The new foundry is generally losses, and it is also likely that well-designed and contains some of there will be some essential minimum the best equipment of its type. capital expenditures if the plant performance is to improve (paras. 85- * It has the ability to produce 110). high grade engineering castings and could be an important supplier not Sustainability only to Tunisian industry, but also to the whole Maghreb region, 11. To be sustainable, the project particularly Algeria, where demand is has to increase both production and forecast to outstrip supply. unit prices and reduce costs. A break-even analysis for the new * SOFOMECA is not only an foundry indicates that a tonnage of important employer in the area, but 29,472 tons/year would be required to also acts as valuable training ground break even. This is far beyond the of Tunisian technicians and managers. present capacity of the new foundry (12,000 to 14,000 tons). The * Significant progress has financing charges are so high that it already been made by the SOFOMECA is impossible to break even at management in commissioning and current low export prices (para. developing the new iron foundry, and 111). the operation still has considerable scope for improvement. 12. As regards the old foundry, the break-even volume is calculated to * SOFOMECA may move into 4,068 tons with an average selling reasonable profitability in a 4-5 price of TD 1,168/ton, which reflects year time period, if it successfully the much higher proportion of local completes programs of market sales with better prices compared to developments (e.g., in North Africa) exports than in the case of the and operational improvements. There - x - seem to be reasonable chances that at an early stage of the project SOFOMECA could increase castings (para. 134). exports to Algeria (para. 117). * Weak Project Management and 15. The Bank has requested an Lack of Resolve in Critical independent survey of SOFOMECA's Circumstances. performance, capabilities and market potential in order to determine, 19. Weak management (and hesitant along with the other investors, what Bank staff) could not come to a actions should be taken to improve decision to remove the main the situation. Consultants have expatriate consultan as soon as undertaken a study, jointly financed problems started. Also, they could by the Bank and SOFOMECA, which has not decide, when the question came up recommended a course of action aimed in 1985, whether to go ahead with the at improving the present financial construction of the new foundry or situation. This study has concluded simply cancel the project. that the best option would be to keep Insufficient supervision efforts and the company in operation, providing waning Bank staff attention later on that major efforts in finding export contributed to the failure of the markets and improving productivity project (para. 136). are made (para. 129). B. Leasons Findings and Lessons 20. The lessons to be derived from A. Findings this experience are as follows: 16. The major findings are related * A large foundry project should 4o market and management issues: be considered on its own merits and not just as a "piece" of a globally * Overreliance on too few designed industrial strategy. A Domestic Customers and Insufficiently large industrial project should not Profitable Export Sales. be implemented just because the Bank and the Government have singled out a 17. Assumptions made in the late particular subsector (EMI), or even 1970s regarding the market were because a Bank financed line of overoptimistic. The economic crisis credit has already been granted to forced the Government to change its such a subsector. protective industrial policies, liberalize imports, and abandon a * An industrial project should number of Government-financed or - avoid relying on just one or two supported projects such as a diesel customers, particularly when full, engine and a.i automotive plant which iron clad commitment cannot be were to be SOFOMECA's main clients. obtained from the start from these customers. If such commitment is not 18. Only lip service was paid to available, the project should be export sales. SOFOMECA is now trying reduced in size or postponed. Also, to expand exports; but experience has if alternative (export) markets are shown that building export markets thought to provide a realistic requires considerable time and possibility, they should be expertise, and it is much more immediately explored and not prudent to integrate export planning considered as an alternative to be * xi - studied only when domestic markets this proves unavoidable, closer fail to materialize. supervision during project implementation is needed to monitor * If, at appraisal, project both management and consultants' management is thought to be weak, the performance. Other investors and co- project should not be financed unless financiers with a stake in the steps are taken during project project must also be involved, preparation to strengthen it. particularly when they have local Excessive reliance on consultants may experience and contacts. prove a risky approach. However, if PROJECT PERFORMANCE AUDIT REPORT TUNISIA SOFOMECA FOUNDRY MODERNIZATION AND EXPANSION PROJECT (=OA 2301-MU) I. INTRODUTION 1. In 1983, a Bank loan was made to Soci6t6 de Fonderies et de Mdcanique (SOFOMECA), a foundry originally established in 1964 as part of a maintenance workshop attached to the Tunisian railway company. It became a separate Government-owned company later in 1964 and sold iron and steel castings to a variety of industrial sectors, mainly in the Tunisian market. 2. Until the early 1980's, the company consi-ted of the following departments: * an iron foundry with a capacity of 3,400 tons/year in two shifts, * a steel foundry with a capacity of 1,800 tons/year with two-shift molding and three-shift melting, * a machine shop/railway bogies assembly shop with a resource capacity of 70,000 man-hours per annum. 3. The equipment installed generally dated back to the mid-1960s when the foundries were established. Few improvements had been made. The relatively old equipment and production deficiencies and bottlenecks led to low productivity and poor quality of the finished castings. The major reasons for low productivity were: poor quality of the locally supplied casting sand, use of lower than desirable temperatures of the molten metal in iron castings, worn out condition of equipment in molding, lack of auxiliary equipment for material flow, non- rational operations in the finishing shop and poor environmental conditions. These contributed to high rates of rejects, poor surface finish and, as a result, to excessive grinding and finishing requirements to bring the castings to acceptable standards. 4. Despite these problems, SOFOMECA's production had gradually increased over the years, growing, for example, from 3,636 tons of iron and steel castings in 1975 to 4,695 tons in 1982. Sales had grown from 3,193 tons to 4,448 tons during the same period- -exports, mainly to North Africa and France, averaging 20% only of the total. In 1980-82, SOFOMECA production levels averaged more than 80% of rated capacity and the company could less and less keep up with demand. 5. The company had, however, a low profitability performance because of a weak price-cost relationship and poor productivity of the existing facilities. SOFOMECA prices were relatively competitive with imported prices, but the costs of inputs (pig iron, coke, silicate, electricity) were higher than those paid by foundries in European countries. On productivity, SOFOMECA required double the man-hours needed to produce one ton of castings compared to similar foundries in Europe, primarily due to constraints in its facilities. The cost of labor, - 2 - despite low wage levels, was almost equal to that in a European foundry due to low productivity (98 man-hours per ton against about 50 in Europe). 6. However, overall, financial performance of the Company was acceptable. Nevertheless, there were problems of high level of receivables and inventories and a deteriorating liquidity position. 7. SOFOMECA was, by far, the largest foundry in Tunisia. Its output of 4,695 tons in 1912 compared to 2,740 tons for Fonderies Rdunies, the second largest plant while the combined output of several small foundries was 400 tons only. Local production thus amounted to 7,835 tons, while imports reached 14,172 tons, i.e., about 2/3 of domestic consumption. Exports were less than 1,000 tons. The market share of local foundries had remained the same during several years and it was, therefore, thought that there was ample scope for import substitution, provided quality, productivity and cost competitiveness of castings were improved. 8. SOFOMECA was led to embark in an import substitution program because, in the late seventies and during the preparation of the Sixth Development Plan (1982-1986), the Tunisian authorities felt the need to emphasize higher labor productivity and increased local value added content, in particular through more integrated development of the manufacturing sector. The engineering industries (foundries, basic metals, electrical and non-electrical machinery, transport equipment, etc.) were thought to be well placed to help achieving these objectives because they were labor intensive and, although small and insufficiently integrated, offered prospects of increasing competence and quality leading to a higher local value added content. 9. A number of engineering products were identified such as foundry castings which could in principle be efficiently import-substituted, using simple or intermediate labor-intensive technologies with a domestic market large enough to sustain economic operations. However, the foundry industry was underdeveloped and weakly integrated within the engineering subsector and needed particular development if it was to become a "backbone" for the development of the Tunisian engineering industries. Due to the small domestic market and the diseconomies of scale, there was a need for rationalization of future production among Tunisian foundries. 10. Five specific measures were recommended to address foundry-industry- related constraints: (i) raise productivity through modernization, expansion and specialization; (ii) raise the quality of foundry products through quality control, training and availability of good local sand; (iii) expand product lines where Tunisia had a comparative advantage in substituting imports, such as castings for automotive and mechanical parts; (iv) compensate for the small domestic market by increasing domestic integration of castings into the production program of vehicle assembly plants; and (v) establish a sound basis for exports through the above measures, encouraging quality and cost competitiveness. 11. A key factor was the importance given to the vehicle assembly subsector. The Government strategy was to increase the production of trucks and buses through, in particular, the expansion of Soci6td Tunisienne d'Industries - 3 - Automobiles (STIA) which has licensing arrangements with Fiat and Renault/Berliet. Local production of diesel engines by a new company, Complexe M6canique de Tunisie (CMT) was to be purchased in part by STIA, thus reducing its dependence on imported components. CMT was also to produce tractors and agricultural implements. CMT was to buy 4,000 tons of mechanical castings from SOFOMECA in 1990 while STIA was to consume directly 1,200 tons of SOFOMECA's castings. Since SOFOMECA's output was to grow from 4,700 tons in 1982 to 14,000 tons in 1990, purchases by CMT and STIA of 5,200 tons of castings represented as much as 58% of SOFOMECA's projected increase in production. 12. The SOFOMECA project was developed in the framework of the preparation of the Sixth Plan (1982-86). As indicated, the prevailing idea at that time was to "deepen" the industrial sector, in particular through integration of engineering industries. Among the latter, prinrity was given to the vehicle assembly subsector. In that context, SOFOMECA was seen as just one element in the global "integration" strategy and in charge of supplying castings directly or indirectly to the expanding transport equipment industry. 13. The Bank endorsed the Tunisian industrial policy to increase the local value added content and to expand exports. It also supported the view that engineering industries were a high priority sector,' and approved a US$30.5 million loan for an Electro-Mechanical Industries (EMI) project in March 1982. The EMI project was supposed to finance, through a line of credit to BDET, new and modern plants in engineering subsectors, including transport equipment. Interestingly, the Bank agreed that Tunisian manufacturers be allowed to charge ex-factory prices up to 18% above foreign prices for comparable products. 14. Since the Bank had decided to support the EMI sector, it was also inclined to assist the foundry industry. Following a "Study of Foundry Sector in Tunisia" by a British consultant, the Bank prepared a "Development Strategy Paper on Foundry Industry" in 1982, focussing on measures to raise productivity and quality, on import substitution and increased integration of castings and on the need to establish a sound basis for exports. The conclusions of the strategy paper were discussed with the Tunisian authorities and its major aspects were reflected in the Tunisian Sixth Development Plan. Based on the strategy, the Government decided to implement two foundry modernization and expansion projects, SOFOMECA and Fonderies Rdunies, which were to remain the main two foundries supplying primarily the domestic market. The Fonderies Rduaies Project (then in the study stage but later given up) was to consist of the modernization and expansion of its existing facilities from 3,500 tons per year to 7,000 tons per year, mainly for the production of "traditional" castings, i.e., not for the automotive sector. 15. The two foundry projects were to address the major identified constraints in the foundry industry. Productivity in the foundries was to be raised through (i) rehabilitation, modernization and expansion of existing facilities, (ii) rationalization of production among the two foundries, and (iii) technical it Review of the Electrical and Mechanical Industries, Report No. 2666-TUN, June 4, 1980. The Sixth Plan ranked these industries third among priority sectors, immediately after agriculture and energy. assistance and training to improve management organization and the quality of the labor force. Quality of foundry products was to be raised through the identification of local sand quarries and the establishment of a sand treatment plan for supply of quality casting sand to SOFOMECA and other foundries, and through the modernization of existing facilities, as well as training of the labor force. 16. The SOFOMECA Project was also supposed to meet the Government's objective to promote private sector participation in industrial development as the ownership structure of SOFOMECA was to change from a publicly owned to a semi- public and privately owned company. II. OBJECTIVES 17. The objective of the Project was to contribute towards the development of the Tunisian foundry industry by: (i) meeting efficiently the growing domestic demand for foundry products; (ii) improving the quality, productivity and cost competitiveness of castings; and (iii) creating a sound basis for exports (SAR, para. 5.01) 18. Basically, the Project had three elements: a. A modernization program for the existing iron and steel old foundries with the objective of increasing their capacities as follows: - iron castings to 4,800 tons per year on two shifts, - steel castings to 2,200 tons per year on two shifts for molding operations and three shifts for melting operations b. Construction of a new iron foundry to produce engineering quality castings, with a single shift capacity of 7,800 tons/year and a two- shift capacity of 12,000 tons/year. c. Installation of a new sand treatment plant with a two-shift capacity of 20,000 tons of sand/year for external sale to other foundries, as well as internal use. 19. The Project also included a program for improvements in SOFOMECA's organization and management, and transfer of technical know-how and assistance for training by a German company (KHD), the technical partner in the CMT project in order to produce castings efficiently and economically (SAR, para. 5.02). 20. To help financing the total Project cost of US$54.06 million, the Bank approved Loan 2301-TUN in the amount of US$16.8 million in June 1983. III. IMPLZU=TAION EXPERIENCE 21. The Project Completion Report (Parts I and II) deals in great detail with the implementation experience of the Project, particularly as regards technical aspects. A Bank-financed Consultant Report prepared in June 1990 (i.e., after the PCR was drafted) provides additional data on implementation problems. Finally, information obtained in the field has shed some more light on what effectively happened during the period 1983-1990. After first reviewing what the main implementation problems were, this section mainly aims at analyzing and commenting on the reasons for such problems. A. Implementation Problems 22. Major implementation issues have arisen in six areas: (a) implementation delays; (b) marketing; (c) technical problems; (d) organization and manning; (e) control systems and services; and (f) financial. a. Implementation Delays 23. The modernization program for the old iron foundry commenced during 1983 and was largely completed in time, i.e., by the end of 1984. As regards the old steel foundry, the actual modernization work was deferred to January 1986 in order not to have the two foundries out of operation at the same time. More importantly, the new iron foundry project suffered from many delays and eventually did not start production until late 1988/early 1989, i.e., 32 months later than anticipated. Reasons were: late start of engineering studies for the new foundry, execution of piling work not included in the project, inadequate performance of the main consultant, import formalities, project management and blocking of loan funds by frustrated local banks. The sand treatment plant started operations in December 1987, with a 26-month slippage (PCR, page 1.7). During this time, plans for the tractor (CMT) and the vehicle (STIA) plants, which the new foundry had been planned to service with castings, had not gone ahead, thus leaving the foundry with no real load from the Tunisian market and he:ee with a considerable amount of spare capacity. b. Marketing 24. Mainly as a result of the collapse of the CMT and STIA companies, project implementation &-ffered and objectives could not be met as planned after the new iron foundry start.d production in 1989/1990. 25. Production and sales have evolved as follows (in tons): - 6 - 1980 1985 1986 1987 1988 1989 1990 1990 (Fore cast) (SAR) Pyroducttoa Iron 3,396 5,560 5,800 3,429 5,046 8,330 9,978 11,800 Steel 1,391 1,417 1,120 727 1,220 1,130 1,018 2,200 Total 4,787 6,977 6,920 4,156 6,266 9,460 10,996 14,000 Iron 3,249 5,154 4,730 3,408 4,650 8,078 10,106 11,800 Steel 1,243 1,158 929 819 1,164 1,118 781 2,200 Total 4,492 6,312 5,659 4,227 5,814 9,196 10,887 14,000 26. It may be seen, then, that the production of iron castings increased from around 3,396 tons/year in 1980 to 5,560 tons/year in 1985, following the first modernization program. This increased further to 8,330 tons in 1989 and 9,978 tons in 1990 with the new foundry coming on stream. 27. Production of steel castings, by contrast, has not shown any appreciable increase as a result of the modernization program. Production in 1980-85 averaged 1,490 tons/year compared to an average of 1,135 tons/year in 1988-90. 28. Capacity utilization has been as follows (in tons): % Capacity Capacity (2 shifts) Output Utilization Projected Atual Proiected Actual Projected ActuAal (SAR) (SAR) (SAR) (1990) (1990) (1990) (1990) (1990) (1990) Old Fovndry 7,500 6,8001' 7,000 4,843 93 71 New Foundry 12,000 12,500 7,000 6,153 58 49 Total 19,500 19,300 14,000 10,996 72 57 11 Including the steel foundry, which was supposed to reach a capacity of 2,200 tons but because of the age and condition of the molding equipment and the low level of coremaking capacity, is considered as having a real maximum capacity of only 1,800-2,000 tons of good steel castings per year. - 7- c. Technical Problems 29. In general terms, a spacious modern facility using a high standard of equipment was implemented in the new foundry. However, problems have occurred due to the incomplete understanding by SOFOMECA personnel of the operation of the equipment. Moreover, some imbalance existed in the potential output of the different sections of the foundry, and there were certain engineering design faults and inadequate maintenance and organization standards. 30. In the old foundry, the recording of metallic inputs and outputs has been poor, so losses are difficult to identify. Also, the cupolas have been lined down to restrict their hourly output to adjust to the reduced production level (the design output was 4,9 tons/hour and the current output 2,9 tons/hour). This is likely to be causing extensive problems with slag preparation, temperature and coke utilization. 31. As regards the sand-treatment plant, the civil engineering, structural steelwork, electricity, fluids and office packages were not well defined in terms of quantities at the time the initial project cost estimate was made. This resulted in numerous surprises during execution (PCR, page 20). 32. Although the sand mills have a nominal capacity of 150 tons per hour, the system does not appear to be able to produce above 114 tons of mixed sand per hour; there would, therefore, be a shortfall at full capacity. Also, sand is being lost in many parts of the system and this leads to large quantities of sand escaping from the system at all levels of the plant. d. Organization and Manning 33. SOFOMECA productivity remains very poor and organization is a weak point as shown, for example, by the fact that 12 managers currently report to the managing director which is too broad a span of control. 34. Although the project was supposed to bring the plant's productivity to the level of similar European plants, a study by a British consulting firm in 1990 shows that the SOFOMECA new foundry had a productivity of 35,6 man- hours/ton, while an average of six European foundries showed a productivity of 17,4 man-hours/ton, i.e., twice less. Similarly, SOFOMECA old semi-mechanized foundry needed 76,8 man-hours/ton against 20,6 man-hours/ton in a similar French foundry. 35. Productivity problems persisted during implementation becau!e SOFOMECA' s management was unrealistic (and so was Bank staff in appraising productivity's future progress) and slow to respond to adverse changes (PCR, para. 11-iii). 36. For the execution of the Project, SOFOMECA established a Project Management Group (PMG). Assurances with respect to maintaining an adequate staffing had been obtained at appraisal. The number of staff assigned to the PMG grew with the progress of the engineering studies and as of 1986 with the start of the assembly work in the new foundry, the group was finally staffed. However, in 1987 the numbers were appreciably cut back owing to the slowdown in the works - 8 - due to the blocking of fin&ncing by local banks. In the course of the second half of 1988, the numbers were again strengthened, especially at technician and labor levels, prior to start up which took place in October 1988. Throughout the period, the PMG suffered from a lack of experience in this type of work, especially in the planning and scheduling side. Moreover, serious disagreements between the industrial engineering firm, George Fischer (GF), who was hired to assist the PMG in managing the project and SOFOMECA, had a very bad impact on the implementation of the project. e. Control Systems and Services 37. Technical assistance, provided under the Project through a contract with a German company (KHD), was limited to training of executives in KHD's plants. There has been a serious lack of operational preventive maintenance programs and it is only during the fourth quarter of 1989 that a Maintenance Division was set up. The production staff in the new iron foundry provides downtime records for the facility, but there is not enough operating history to allow any evaluation of the state of repair of the equipment. The state of repair of the equipment in the old iron foundry seems reasonable, if its age is taken into account. 38. Reasons for the poor maintenance includes the lack of written prevention maintenance procedures and plans and the lack of widespread transfer of information from the equipment manufacturers to SOFOMECA personnel. Moreover, lack of training and qualifications --particularly among older, uneducated staff- - has slowed down the systematic implementation of normal prevention maintenance programs. Staff resistance has prevented preventive maintenance systems to be put in operation. f. Financial Problems 39. Between 1980 and 1986, SOFOMECA performed in a relatively modest way, varying from a net profit of TD 142,000 in 1980 to a net loss of 99,000 in 1986. 40. 1987 was a poor year, with the increased financing charges associated with the modernization of the old iron foundry coinciding with production problems due to the plant movements and generally difficult selling conditions. This resulted in a net loss of over TD 1.2 million for the years, i.e., about 20% of sales which amounted to TD 5.6 million in that year. 41. After a partial recovery in 1988, with increased throughput from the old iron foundry and an improved performance from the steel foundry, SOFOMECA then slumped back into heavy losses in 1989, i.e., TD 4.6 million compared to TD 9.7 million sales. 42. Most of these losses originated in the new iron foundry which had major marketing problems. The old iron foundry is also a loss maker while the steel foundry is just able to break even. 43. The PCR gives a detailed account of the difficult financial problems experienced by SOFOMECA during implementation. In particular, following the SOFOMEGA Board's decision in 1986 to have an external audit conducted of SOFOMECA and of the project because of the implementation long delays and marketing - 9 - problems, the local development banks (STUSID, BTKD and BTEI) blocked the remainder of the loan funds granted for financing the project. When the audit showed a financing deficit of TD 5.1 million in February. 1987, SOFOMECA was asked to cut back capital expenditures. Moreover, BTEI made continuation of the financing conditional on resolution of the financial problem. The funds began to be released in November 1987 only after the Ministry of industry's intervention. 44. The base cost of the project was 5.7% above estimate (in Tunisian Dinars) and that despite the cuts made in the investment in March 1987 at the request of the banks. Interest during construction, however, was 307% higher than originally estimated, as a result of the late start, which meant that the interest payable had to be placed under this head since the grace periods in the financing expired for the most part as of the end of 1986 or in 1987. Including interest during construction, total project cost was 21% above appraisal estimate. 45. In dollar terms, the estimated actual project cost was about US$58.4 million (after cancellation and reduction of certain equipment lots), about 8% higher than the appraisal estimate of US$54 million. B. Reasons for Implementation Problems a. Industiial "Integration" Strategy and the EMI Sector 46. This Audit agrees with the PCR view that the underlying assumptions of the SAR were seriously flawed in retrospect. The SAR failed to foresee problems of SOFOMECA's major clients (CMT and STIA), Tunisia's economic slowdown, and the increasing competitiveness of the iron and steel castings market. (PCR, para. 8). 47. In fac-, the entire approach to industrial development (need to "deepen" the local value added content which led to giving the highest priority to further integration of the engineering industries, including transport vehicle assembly, diesel engines, foundry products, etc.) proved much too ambitious and unrealistic, in particular given the limited time frame available (a few years) to achieve stated objectives. SOFOMECA difficulties during and after implementation of the project also basically stemmed from the adoption of a "new industrial policy" for EMIs in the late 1970s which proved unsustainable in the following decade due to the economic crisis and the dramatic turnabout in macroeconomic policies. Unfortunately, this industrial policy was fully endorsed by the Bank with damaging consequences for both the EMIs and the SOFOMECA project. 48. Although private investment had increased outstandingly during the 1970s, the public sector still played an important role in the manufacturing sector and accounted for 61% of investment in 1977-1981. Most of the larger enterprises in Tunisia were publicly owned, especially in the chemicals, construction materials and engineering subsectors. The Ministry of Industry was responsible for identifying projects which were then included in the Development Plan and financed by the state. Such projects in 1978 included CMT, STIA II - 10 - (automobile assembly), SOFOMECA extension, ICM (fertilizers), El Foulad Steel Mill extension, oil refinery extension, etc. 49. In 1979/80, the Government decided to favor an increased participation of the private sector in order to accelerate exports and employment. New measures were introduced in 1981 to expand exports and increase their local value added content. It was assumed that the State's role would be mainly limited to investments in energy, water, railways and eventually mining and phosphate fertilizers and few large Government-financed industrial projects were carried out after 1980/81. The policy was instead to finance projects through newly created development banks supported with Arab capital (STUSID, BTKD, BTEI, etc.) 50. In 1979/80, value added in the Tunisian manufacturing sector was still rather low. The main "traditional" subsectors were textiles, footwear, and construction materials. Engineering industries remained small representing only 2% of CDP and 14% of value added in manufacturing. As a consequence of the focus of assembly activities, the value added of the subsector's output had remained low at an average of 25%. 51. In order to increase forward and backward linkages in the industrial sector, it was decided to give priority to engineering industries which were thought not only as having a good labor potential, but also as being able to increase domestic value added with the production of local components. A key subsector was to be the locally assembled vehicle industry. New projects included the expansion of STIA for the production of trucks, buses and pickups, the implementation of CMT to produce tractors, agricultural implements and diesel engines, and the establishment of a General Motors/Isuzu plant for the production of pick-ups. Diesel engines produced by CMT were to be incorporated in STIA's trucks and buses and in CMT's own tractors. 52. The vehicle assembly strategy was based on the idea that compensation agreements would be signed with foreign partners (Fiat, Renault, GM/Isuzu, Volkswagen. Peugeot, Mercedes, etc.). Imports of components for assembly (CKD) were to be compensated by exports of components up to 50% of CKD value.- The 50% ratio was to be reached over a five year period. 53. Compensation agreements were effectively signed in 1982/83 but could not be implemented since, by end 1985 and in 1986, the Government was forced to abandon its "integration" policy due to a dramatic shortage of foreign exchange linked to an economic crisis. The Tunisian Dinar devaluation of August 1986 resulted in a substantial increase of the cost of CKDs and other parts. Imports of CKDs were largely stopped or reduced while assembly plants were asked to implement compensation agreements not over a five year period but immediately, which proved technically unfeasible. STIA II and GM/Opel plants closed down while construction work under way for other plants stopped. STIA was suffering from financial losses, higher imported components costs, poor sales, understaffing and costly investment made necessary for the car assembly on top of the trucks and buses line. STIA I (trucks and buses) was kept open but the 1 And, by exports of complete vehicles and diesel engines. - 11 - integration program (brake drums from SOFOMECA and diesel engines from CMT using castings from SOFOMECA) was brought to a halt. 54. The CMT project was to be implemented in 3 main stages ' and produce in 1988 at full capacity tractors and engines to be inco-:porated in tractors or sold to STIA. Stage 1 was implemented only in the first half of 1985 (1872 tractors were sold with a small profit in 1985). However, the 1986 crisis resulted in consumers credit rationing, in an increase in customs tariffs and a steep rise in imports costs as a result of the Dinar devaluation. Tractors sales fell rapidly while Tunisian farmers still preferred to buy Imported tractors with well known brand names. CMT's financial situation deteriorated not only because of higher import costs and lower Fales, but also because the investment program for stage 2 (engines) proved also much more expensive than anticipated. Efforts were made to accelerate integration, to cut down on investments by concentrating on the production of high unit value components, to sign some compensation agreements, and to increase purchases from local sub-contractors. However, farmers were not forthcoming to buy CMT tractors and sales remained dramatically low. Government measures to subsidize CMT tractors were not implemented and CMT finally closed in August 1989. Foreign exchange losses and high stocks due to lack of sales resulted in a DT 6.3 million deficit in 1989. 55. As indicated above, in 1980 the Bank formally endorsed the Government policy to promote EMI development. BDET pipeline of projects specifically included the CMT project (diesel engines) and the STIA extension. These projects as well as export "compensation" and other auto components projects were also included in the list of projects eligible for BDET financing. Moreover, a nominal 18% to 21% protection accompanied by quantitative restrictions was permitted for these industries on the ground of "infant" protection. At one time, SOFOMECA rehabilitation was also included in the BDET pipeline provided that the company was to be privatized.' 56. The 1980 Bank Report neglected, however, essential, major aspects of the Tunisian industrial policy particularly regarding, (i) excessive quantitative restrictions and high tariffs and (ii) the foreseeable public sector deficit which inevitably would present further subsidizing of public enterprises, many of which were purchasing goods produced by EMIs. There was thus a substantial risk that with a falling demand such EMIs might have to close. Assembly plants such as STIA and CMT were uneconomical in the medium term and to produce components locally for these companies was thus dangerous (SOFOMECA was to supply castings to STIA and CMT). What should have been done at that time should have been to insist on lowering protection rather than to accept an 18 to 21% protection level supported with im7ort quotas and thus make sure that assembly plants would have been able to withstand competition and export at competitive prices. 11 Stage 1, assembly of tractors in 1984; stage 2, introduction of assembly of diesel engines and machining of engine parts (including castings from SOFOMECA) in 1985; and stage 3, assembly and production of all parts and components in 1987. 1' The EMI Project. SAR - Report No. 3689-TUN. March 2, 1982. - 12 - 57. However, a major problem was that few reliable data were available on costs and effective protection in EMIs. The Tunisian authorities were extremely enthusiastic about the promotion of EMI, but had not yet been reconciled with the idea that such industries had to be in,ernationally competitive and hence unprotected or little protected, in order to succeed in the long term. The Tunisian industry had historically been highly protected and the "compensation" export strategy combined with Bank and various Arab development banks financing, was thought sufficient to help promoting EMIs behind a solid protection wall. The Bank probably felt that supporting EMIs (and SOFOMECA) would help developing a labor intensive sector and technoloyical progress. Insisting on greater, immediate liberalization of the economy (prices, tariffs and quotas) was judged, at that time, unrealistic in the framework of an EMIs sectoral loan and in any case premature. The 1986 crisis and subsequent developments proved that such a purely sectoral approach was dangerous because it did not provide enough leverage to deal with basic macroeconomic and overall industrial strategy issues bearing on EMI competitiveness. Moreover, the underlying assumption that "compensation" agreements with foreign partners could be fully worked out in 5 years only proved equally wrong since the 1986 crisis intervened before such agreements could be fully implemented. b. Excessive Reliance on a Few Key Customers (STIA and CMT) 58. The SOFOMECA project was just an element of the strategy giving high priority to EMIs and in particular to the vehicle assembly subsector. The Ministry of Industry felt that foundry products should be included in specified "integration" programs and to that effect identified the relatively old SOFOMECA plant, built in the mid-1960s, together with Fonderies Reunies, with the objective of increasing production of castings fcr the automotive and tractor industry. At the same time, quantitative restrictions and relatively high tariffs were to enable EMIs and the SOFOMECA project to progressively improve efficiency and quality until they would, one day, become internationally competitive. 59. In 1982, the company's production of around 4,700 tons accounted for about 60% of Tunisian output of castings. In the domestic market, SOFOMEGA represented a share of 18%. Its major domestic clients were primarily public enterprises such as the railway company, water distribution authority, housing authority and some machine building enterprises. Its main clients abroad were located in France and Algeria but exports averaged only about 20% of sales in the 1970s due to relatively poor quality of the production and other constraints. 60. The project assumed that SOFOMECA's market position would change significantly, raising its s'Aare in the domestic market to 40% in 1988. SOFOMECA's market for the mechanical castings of the new foundry was to be largely connected to the demand of CMT and STIA. CMT was to be SOFOMECA's major direct client of mechanical castings with annual requirements of about 4,000 tpy in 1990 while STIA was to consume 1,200 tpy of SOFOMECA's castings. Consequently, SOFOMECA was to sell 5,200 tons of castings to CMT and STIA, i.e., 58% of the production increase from 4,700 tons in 1982 to 14,000 tons in 1988. 61. Such dependence on two potential customers proved extremely dangerous. At the time of appraisal in 1982, the integration of mechanical castings from the - 13 - project in CMT's and STIA's production was said to have been finalized by SOFOMECA in collaboration with these enterprises (SAR, para. 9.14). Further, the Government had indicated that it would cause CMT to carry out CMT's investment program for the period 1983-88 according to a schedule agreed with the Bank (SAR, para. 4.08). 62. The Bank decided to finance SOFOMECA as a separate project because its cost (US $54 million) was too high to be included under the EMI line of credit to BDET. But Bank staff was convinced at that time that the rationale to support EMIs (endorsed by the 1980 Industrial Mission), could also be applied to SOFOMECA provided STIA and CMT would provide "guaranteed" market for SOFOMECA castings. 63. Bank staff and SOFOMECA's management believed strongly that these two projects were to be effectively implemented in the framework of the Government industrial strategy to promote EMIs. Discussions took place with CMT and STIA and "letters of agreement" were received from the two firms, the aim being to dovetail SOFOMECA expansion with CMT and STIA investment programs. 64. One disquieting factor was, however, that CMT was to have been partly financed by the German KFW, thus strengthening the argument that CMT was a reliable partner. However, as early as April 1982, KFW informed the Bank that it was taking a "skeptical position" towards the CMT project and that it wanted to approach it "in a prudent way". Also, EMENA's Agriculture Division 2 advised the Industrial Projects Department that "the CMT project may have been based on too optimistic demand projectiens for tractors." The projection of the share of the tractor market of 75% expected be captured in 1988 by CMT seems high, unless resulting from a high tariff ptotection U. Farmers, when given a choice, are oiten conservative and prefer to stick to known markets with satisfactory after sales services" (Memorandum dated March 18, 1983). This proved entirely right afterwards, but was igaored at the time since the "mood" was very upbeat in favor of supporting the overall EMI strategy in Tunisia. 65. The Bank was not able tj obtain submission by the Government by December 1983, of "proposals" for the inregration of diesel engines and castings into products produced by STIA unde future technical assistance and licensing r.greements with foreign partnerd. The existing agreements were to expire on December 31, 1983. According to M-Anutes of Negotiations (Memorandum dated April 26, 1983), the Tunisian delegation stated that they expected the present partners (Berliet and Fiat) to continue to work with STIA beyond the end of 1984. However, a reference to a date in the Guarantee Agreement would have weakened the Tunisian negotiating position. The reference to a date was therefore exclude from the Guarantee Agreement and included in the agreed Minutes of Negotiations. Furthermore, the reference date was changed to March 21, 1984 by which time the draft proposals will have been negotiated with the new, or confirmed, present partners...." 66. In May 1984, the Bank had not yet received information on plans related to the integration of CMT motors and SOFOMECA castings with STIA's own program. 1' Protection was to be limited to 18% but complemented by Ruantitative restrictions which resulted de facto in a high level of protection. - 14 - This had been promised during negotiations for March 31, 1984. The Bank expressed concern about the adverse impact of the integration of the CMT motors in STIA of: (i) the dela3 in the selection of a new STIA partner, and (ii) apparent new vehicle projects in Tunisia including engines comparable to CMT engines which could affect STIA market positions and therefore CMT's integration progress. 67. In 1985, the Bank again complained that it had not yet received the proposals for the integration of CMT engines and SOFOMECA castings in the future STIA production program under licensing with new foreign partners. Furthermore, when a Bank mission inquired about other vehicle assembly projects which may have affected the future market position of STIA in the CMT engine range, such as the assembly of pick-ups by GM/Isuzu, Mercedes and Volkswagen, the Director General of Industry responded that such projects were encouraged in order to increase competition. The proliferation of such small assembly operations, however, was economically very doubtful and was to jeopardize existing projects, including SOFOMECA. The Director General of the Plan indicated that a strategy study for the vehicle assembly industry was to be carried out shortly with German assistance. The Bank simply recommended that this strategy be reviewed by another mission (Supervision Report, April 29, 1985). 68. One year later, i.e., in April 1986, STIA had not yet finalized the selection of its new partner, although this was expected by June 1986 while the integration was to be started by January 1986. However, definite plans on the marketing arrangements with STIA were essential to back up the viability of the SOFOMECA project. But at that time came the economic crisis, the devaluation of the Dinar and the stoppage of the STIA expansion project, as well as the severe cuts in the CMT investment program which finally never went much beyond the mere assembly stage, and this eliminated much of the use of SOFOMECA castings from the new foundry. 69. There seems to be little doubt that the Bank was much too confident that CMT and STIA projects would be implemented. With hindsight, given warnings from KFW, Bank agricultural experts negative opinion on future tractors demand and the unwillingness of STIA representatives to commit themselves in 1983, it would have been wiser to restrict the Project to the financing of SOFOMECA's old foundry modernization and to postpone the new foundry project, mostly geared to the sale of castings to CMT and STIA. Once it had been agreed, during negotiations, not to insist on a full agreement from STIA to buy engines and castings, the Bank had lost any leverage and successive missions could only complain about STIA's delays in finding a new partner but without any real impact. c. Role of the Banks 70. The Project was cofinanced by three locally established joint-venture development banks, Tunisian-Saudi Investment and Development Bank (STUSID), Tunisian-Kuwaiti Development Bank (BTKD), and Tunisian-Emirates Investment Bank (BTEI)U which were to provide US$5.56 million as equity and US$14.99 million U BTEI's participation was subsequently divided evenly with BTQI (Qatar), i.e., US$0.5 million each. - 15 - as long-term loans. Additional share capital amounting to US$12.06 million was to be raised from Arab Mining Company (ARMICO), from Jordan; Arab Industrial Investment Company (AIIC), from 1raq; and Islamic Development Bank (IDB), from Saudi Arabia.- As a result, non-resident foreign shareholders (ARMICO, AIIC and IDB) had a much larger share of the share capital that the local development banks. 71. The Bank loan of US$16.8 million was provided directly to SOFOMECA and guaranteed by the Government. As indicated above, long-term loans were also made by th3 Tunisian shareholders for US$14.99 million and commercial banks were to finance US$0.95 million. Total project cost was US$54.06 million. 72. From the start, the banks faithfully followed the guidance of the Tunisian authorities and of the Bank. They believed that Bank experts had rightly identified a sound project equally supporLed by the Government in accordance with a "new" industrial policy aiming at further "integration" of the industrial sector, in particular through the promotion of the EMI subsector. 73. These banks had understood that firm contracts had been signed vith CMT specifying, in great detail, castings sales and prices as well as timing of deliveries to CMT. They also believed that the Government was fully committed to the implementation of the CMT project and to STIA expansion. 74. When the Project's implementation was delayed, some banks, such as BTEI and BTKD, realized that difficulties were mounting. They decided to step in and requested that RVI (Renault) be made responsible for the management of assembly and training programs since major problems had arisen earlier with George Fischer, the main consultant. But RVI was apparently not fully put in charge by SOFOMECA, which may have insisted in doing too many things on its own. 75. In general, Tunisian minority shareholders felt they did not get much support from non-resident Arab majority shareholders, which gave SOFOMECA a great degree of autonomy.Y Also, Tunisian banks were disappointed by the fact that, despite mounting problems, Bank' s supervision missions remained rather infrequent (an average of 1.7 mission per year in the four-year period 1983-1986). As a result, Tunisian banks felt that the faith that they had put into Bank staff to solve all implementation problems was not fully warranted. At the same time, they believed they were not supported by non-resident Arab banks while SOFOMECA's management would not efficiently take into account their suggestions. 76. As explained above (para. 43), the banks finally succeeded in having an audit completed on SOFOMECA, which had shown a large deficit in 1985. Loan funds ' IDB participation was subsequently made in the form of a leasing agreement (US$1.4 million) and another bank, BEST, invested US$0.9 million while providing leasing funds for US$0.7 million. It Non-resident banks sought more active Government intervention in order to protect their equity participation, while local banks were frustrated to see their credits unused due to implementation delays, and tried to intervene directly in the Project. - 16 - were blocked by local banks from February to November 1987, thus delaying the project for nine months. 77. It must be noted, however, that loans from some local development banks (BTKD, BTQI, and particularly BTEI) had terms and conditions more unfavorable than foreseen, with shorter grace period and higher interest rates, which helped jeopardizing SOFOMECA's financial position, particularly in the start up period of the new foundry. In any case, the deteriorating financial situation of the company during implementation soon resulted in delays in payment of interests. At end 1989, interest unpaid to local development banks amounted to TD 2.1 million and to other banks (leasing, commercial) amounted to TD 0.22 million. Unpaid principal was TD 0.34 million. The local development banks were not really able to supervise and influence the project's implementation. Full reliance on Government policies and on the World Bank was the major reason for them to invest in the project (as minority shareholders and major lenders) but this proved insufficient and they were not repaid when difficulties arose. A more active attitude on the banks' part from the start may have helped to solve some of the major implementation problems (poor company's management and marketing issues), thus helping to prevent the catastrophic financial situation which SOFOMECA is now facing. 78. The Government's role during implementation was also relatively passive. SOFOMECA's general manager was an engineer and a senior civil servant who had been involved earlier in a rather successful Bank project, and believed that the SOFOMECA project would succeed with Bank assistance, good market prospects and sound Government's industrial policies (i.e., EMI's "integration" in a protected market). However, the drastic change in such policies in 1986 helped sealing SOFOMECA's fate at that time, resulting in the disappearance of major markets for castings. Thus, although Government rarely directly intervened in SOFOMECA's affairs during implementation, it indirectly played a determinant role in 1986 and following years. At present, the Treasury is now repaying the Bank's loan which it guaranteed and that SOFOMECA is now totally unable to repay. d. Management Weaknesses 79. Delays in implementation can also be explained by the inefficient management of the Project. A Project Management Group was set up but remained weak because of a lack of experience, particularly in the planning and scheduling side (PCR, page 23). 80. The appointed general manager was a good negotiator able to deal with Banks and ministries, but with little practical industrial experience. Although assisted by a capable deputy general manager experienced in foundry work, it proved difficult for him to fully managed a project of that size. It must be remembered that there are few independent foundries as large as SOFOMECA (most are captive plants such as the large Peugeot's foundry in France) and the challenge was then particularly strong. When relations between SOFOMECA and the main consultant (GF) deteriorated (see para. 89), the PMG was almost alone responsible for the implementation of the project, which proved too much of a burden. - 19 - 89. The new foundry is, in general terms, a spacious modern facility using a high standard of equipment. There are however, problems in the following areas: * the incomplete understanding of SOFOMECA personnel of the operation of the equipment; * certain engineering design faults; * currently inadequate standards of maintenance, cleanliness and safety. 90. To meet these problems, substantial technical assistance would have to be requested from the suppliers of the major items of equipment and a major program should be established to train personnel to improve maintenance, cleanliness and safety. It is to be recalled that substantial technical assistance and training programs were included in the project but, obviously, were not sufficient since serious problems remain. 91. The design capacity of ner iron foundry is understood to be about 14,000 tons per year on two shifts, although it is currently operating at substantially below this level. However, the existing smelting facilities limit the two shift production of molding units to 12,400 tons per year, way above the nominal molding capacity of 16,2-0 tons. The molding units have to start and stop because of shortage of liquid metal supply. B. ProductivJy 92. SOFOMECA's labor force grew from 542 in 1982 to 852 in 1989, but declined slightly to 840 in 1990. These results are somewhat higher than employment of 787 in 1988/90 projected at appraisal. 93. On the basis of optimistic production of castings (based on an earlier 1990 forecast of 17,685 tons, while the actual output was about 11,396 tons) and employment data of 852, productivity calculations are as follows: - 20 - 014 Fondry pew Foundry Toal iron aw Production (tone) 4,524 1,345 11,816 17,685 (1990 forecast) Foundry - Manning 174 123 184 481 - Man-houre/ton 64.5 153.4 26.1 45.6 Foundries and Services - Manning 268 192 318 778 - Man-houre/ton 99.4 239.5 45.1 73.8 Ot-*r Departments - Manning TOTAL NANNING 852 Sources Preliminary evaluation of SOOECA Foundry operations. Knight Vendling Ltd., England. June 1990. 94. These productivity calculations indicate a very low level of productivity for the three foundries together of 45.6 man-hours/ton. As anticipated, the productivity calculated for the new iron foundry at 26.1 man-hours/tons is significantly better than that calculated for the old iron foundry at 64.5 man- hours/ton. These figures rise to 45.1 man-hours/ton and 99.4 man-hours/ton respectively if the other service departments are taken into account. 95. In order to provide some measures of international performance, productivity figures for a group of mechanical foundries in Europe and one semi- mechanical jobbing foundry in France have been compared with the new iron foundry and the old foundry, respectively. 96. Results are as follows: Mechanical Foundryn' Man-hours/ton - Average of 6 European Foundries 17.4 - SOFOMECA New Iron Foundry 35.6 Semi-mechanized FoundryU/ - French Jobbing Foundry 20.6 - SOFOMECA Old Iron Foundry 76.4 97. The productivity of the steel foundry is 153.4 man-hours/ton, which is twice as poor as the average European standard. A Excluding charge preparations but including maintenance. - 21 - 98. The cost of labor for SOFOMECA is estimated to be 17% only (SAR projection for 1990 was even lower, i.e., 11%). Consequently, even should productivity improvement be pursued actively, the cost savings would not contribute as significantly to the financial performance of the business as in one equivalent European foundry. Nevertheless, it is important that productivity be improved in light of the necessity to develop export markets for the new iron foundry and also of its general financial situation. C. Quality 99. Quality may be judged by the percentage of scrapped castings for each of the foundries: m m122 New Iron Foundry 14.3 5.7 Old Iron Foundry 12.3 12.9 Steel Foundry 11.4 12.8 100. It may be seen that control has been gained over the scrap in the new foundry, but the old iron foundry continues at relatively high levels. Many of the problems with the old iron foundry are considered to be associated with the operation of the melting plant. Overall, quality control systems are being developed and the quality of scrap reporting is impressive (daily reports). D. Output and Sales 101. In 1990, output has been significantly below original projections (in tons): Forecast (SAR) Actual DUtput Iron Castings 11,800 9,978 Steel Castings 2.200 1.018 14,000 10,996 Capacity utilization Iron Castings 90 57L" Steel Castings 100 56.6 Based on a 12,400 tons/year real capacity for the new iron foundry and 5,500 tons for the old iron foundry. Capacity utilization was projected at 90% in the SAR, assuming capacity of 4,800 tons of iron castings in the old foundry (two shifts) and 7,800 tons in the new foundry (one shift). I Based on a 1,800 tons/year real capacity. - 22 - 102. Reasons for low capacity utilization include learning curve difficulties during the run-up of the new facilities to ftll production (the steel foundry was closed for 3 months in 1990 due to an electrical breakdown of the arc furnace) but mainly the lack of markets. As regards iron castings, local competition from Fonderies Rdunies and 18 other small foundries is relatively strong. SOFOMECA's iron castings production for the domestic market rose from 3,150 tons in 1983 to 4,267 tons in 1989, while total iron castings capacity is now estimated at 17,500 tons. Iron castings exports were only 3,811 tons in 1989 and at particularly low prices (TD 0.749/Kg against TD 1,110/Kg for domestic sales). 103. Exports of steel castings remain low (107 tons in 19891') because of the relatively strong domestic market (1,011 tons sold in 1989). But, it must be remembered that this market remained practically stagnant in the 1980s, i.e., only around 1,100 to 1,200 tons per year. E. Eorts 104. One of the project's objectives was to provide a "sound basis for exports of iron and steel castings." This meant that although in a first phase the new foundry would sell mostly in the domestic market to CMT and STIA it could later on expand exports since capacity utilization could be increased using a second shift. Iron castings were to reach 500 tons in 1990 (compared to 200 tons in 1983), i.e., 3.5% of projected sales in 1990. 105. Exports amounted to about 42% of sales (i.e., 3,918 tons out of a total of 9,196 tons) in 1989. This ratio is obviously much higher than originally anticipated, but only because domestic sales have increased much less than projected. It is true, however, that SOFOMECA has been able to increase exports from 500 tons in 1983 (including 300 tons for steel castings) to 3,918 tons in 1989 because of the pressing need to find export outlets. Results obtained in 1990 indicate that exports are growing further. 106. However, a major problem (which was not entirely foreseen at appraisal) was the low export prices. In 1989 the average export price for iron castings was TD 0.749/Kg subdivided between TD 0.706/Kg to European customers and TD 1.277 to North African customers. Unfortunately, export sals to North Africa (mainly Algeria) only represented 7.5% of the total tonnage of iron castings exported in 1989. The low European prices are explained by the fact that exports are for low-priced brake drums and flywheels and not for higher-priced products required by the tractor, truck, pumps and valve manufactures. The need to increase unit export price is essential if SOFOMECA is to be able to improve its financial viability and to breakeven.g/ At current selling prices, SOFOMECA cannot breakeven. MI No steel castings exports were assumed in the SAR. 0-i The breakeven point in the new iron foundry would have to be TD 0.900/Kg (including depreciation and financing) for sales of 16,194 tons. - 23 - F. Proiect Costs 107. The estimated actual project costs were about US$58.4 million, about 8% higher than the appraisal estimate of US$54 million. There was a 21% overrun in Tunisian dinars due to the devaluation of the dinar and high interest payments during construction (+307%) because of prolonged delays in project implementation. G. Financial Results 108. SOFOMECA showed a net loss of TD 6.0 million in 1989. The steel foundry lost TD 0.3 million while the old iron foundry showed a TD 0.5 million loss. But mainly, the new foundry posted a TD 5.2 million loss due to low volume of sales. 109. SOFOMECA's current financial position is catastrophic. The 1989 balance sheet showed the company to be under pressure on the ratio between current assets (TD 8.8 million) and short-term liabilities (TD 10.9 million). The major debt repayment schedule indicates that payments of TD 51.3 million will be required between 1990 and 1998, with TD 41.6 million being payable before the end of 1994. Out of total repayments of TD 51.3 million in 1990-1998, sums owed to the Bank total TD 18.3 million. Since SOFOMECA is unable to repay its loans, the Bank receives payments from the Tunisian State who guaranteed the Bank's loan. 110. In the interim period, it will be necessary to finance severe operating losses, and it is also likely that there will be some essential minimum capital expenditures if the plant performance is to improve. V. SUFTAINABILITY 111. To be sustainable, the project has to increase both production and unit prices. A breakeven analysis for the new foundry (based on figures extracted from SOFOMECA's 1990 budget) indicates that a tonnage of 29,472 tons/year would be required to breakeven on an annual cost structure of TD 11.7 million divided between variable costs (46%) and fixed costs (54%) and assuming an average selling price of TD 669 per ton. This is far beyond the present capacity of the new foundry (12,000 to 14,000 tons). Excluding the financing and depreciation costs, the breakeven point is 9,476 tons/year, which is within the current capacity of the unit. The financing charges are, however, so high that it is impossible to breakeven at current low price levels. 112. The effect of different selling price assumptions on the breakeven points both including and excluding depreciation and financing charges is as follows: - 24 - Selling price Breakeven Point (in tons) TD/t Incl. Depr Excl. Depr. + Financina + Financing. 669 29,472 9,476 U 800 18,301 5,884 900 14,594 4,563 1,000 11,592 3,727 113. SOFOMECA' s costs are much higher than comparable costs and selling prices in Europe. Low core iron castings (brake drums, flywheels) sell in Europe for as low as TD 680/ton/Kg, while high core products (cylinder blocks, heads) sell for TD 1,440/Kg. If the new foundry is to move into economic viability it cannot remain on low priced European work (SOFOMECA exports to Europe mostly low core products). Because of the lack of experience of SOFOMECA personnel in manufacturing cored work on the new foundry lines, higher price levels should be sought by moving to higher-priced North African and Middle Eastern work in the short term. It should then be possible to build up experience in providing cored work so that SOFOMECA will have more market options available to it in the medium term. 114. As regards the old iron foundry, the breakeven volume is calculated to be 4,068 tons with an average selling price of TD 1,168/ton, which reflects the much higher proportion of local sales with better prices compared to exports than in the case of the export-oriented new foundry. The 4,068 tons figure fits with the current estimated 5,000 capacity of the old iron foundry. 115. A breakeven analysis for the steel foundry shows that, at a price of TD 885 per ton, the breakeven tonnage is 1,637 tons/year which is substantially in excess of the performance of recent years when production averaged about 1,100 tons because of the plant's lack of profitability and the age and condition of its equipment. The steel foundry appears to be a loss maker despite its strong captive local market position. However, its performance may be improved by the new smelting capacity presented by a small furnace in the new iron foundry. Some minor investments might be able to provide significant improvement in its results and enable production to eventually reach its annual 1,800-ton capacity. 116. By normal financial criteria, SOFOMECA should close operations because of its high level of indebtedness and its inability to regain profitability for at least 3-4 years. If it is to continue operating, a major refinancing exercise will be necessary as part of an overall rescue plan. 117. Factors which argue in favor of SOFOMECA continuing operations are as follows: - 25 - a. The new foundry is generally well-designed and contains some of the best equipment of its type. b. It has the ability to produce high grade engineering castings and could be an important supplier not only to Tunisian industry, but also to the whole Maghreb region. c. This could be particularly significant in the context of Algeria, where demand is forecast to outstrip supply unless now iron foundry facilities are built. d. SOFOMECA is not only an important employer in the area, but also acts as valuable training ground of Tunisian technicians and managers. e. The company's problems have largely arisen from its attempt to support two projected factories (CMT and STIA) in Tunisia which have eventually not proceeded. f. Significant progress has already been made by the SOFOMECA management in commissioning and developing the new iron foundry, and the operation still has considerable scope for improvement due to low productivity and quality. g. SOFOMECA may move into reasonable profitability in a 4-5 year time period, if it successfully completes programs of market developments (e.g., in North Africa) and operational improvements. 118. There are three main options available to SOFOMECA: closure, sale, or continuation as a company.1' A. Closuare 119. The closure of the new foundry would allow some small returns to be made to the company's creditors if the equipment could be resold. However, it seems likely that the majority of the debts would have to be written off, because of the low value of second-hand equipment and the extent of the investment in the foundry buildings. Tunisian and other North African countries would also lose a potential supplier of major strategic significance. 120. The old iron foundry and the steel foundry seem capable of operating profitably, without significant capital investment. There seems little point closing the old foundries and keeping the new foundry open. They should only be closed in the event of the total closure of the site. 121. Total closure may be necessary if the company's creditors are not prepared to accept any rescue plan. It would, however, seem advisable to try to maintain the older facilities, because of their usefulness. This might be done AD6 Cf. Preliminary Evaluation of the SOFOMECA Foundry Operations. Knight Wendling LTD, Stockport, England - June 1990. - 26 - by reopening them under a separate name, while SOFOMECA itself goes into liquidation. B. Aig 122. In light of the current financial conditions, it seems unlikely that SOFOMECA could be sold as "going concern" with its current level of indebtedness. There might be some chance of selling the new foundry, if it could be set up as r a separate company with a relatively sound financial structure. It would, however, be of more interest to a prospective buyer, if it were to be fully operational and had more experienced production staff, rather than being at the current build-up stage. Some European investors have been contacted , but indicated that although they were searching for cheap sources of castings supply, they considered SOFOMECA as a relatively short term supplier and were rather looking at possibilities in Eastern Europe. The sale of the company does not seem a reliable option in te short term. C. Continuation as a Company 123. This would seem to be the most viable option in order to sustain the project. This could also be the best hope for the creditors to obtain some return on their investment. 124. Obviously, actions would be required to achieve minimum production levels over the next four to five years in order to first reach the breakeven point and then some profitability level. Such actions include: (i) Marketing 125. Maghreb and Middle East markets should be investigated and a major sales drive undertaken. At the same time, selected long-run work should be obtained from the European market. (ii) Production 126. A preventive maintenance system should be implemented while maiming levels should be reviewed with the objective of making significant improvements in productivity in the next 2-3 years. Also, scrap reduction should be obtained and castings yields improved in each foundry. (iii) Investments 127. Any continuation should be done initially in minimizing capital expenditures. The only investments that should be sanctioned in the short term would be financial and management assistance programs. All the other items should be made conditional on SOFOMECA management achieving production targets.lU I In the medium term, physical investments may include a new melting unit in the new iron foundry, electric furnace in the old iron foundry and other auxiliary equipment to remove bottlenecks and to improve performance. 27 - (iv) Financial 128. It is inevitable that SOFOMECA will continue to make losses for several years due to its severe financing costs and the slow buildup of the new foundry to full production. The lending banks must be prepared to re-schedule their debts or turn some into equity, if the company is to survive. 129. The Bank has requested an independent survey of SOFOMECA's performance, capabilities and market potential in order to know, with the other investors, what actions should be taken to improve the situation. Consultants have undertaken a study, jointly financed by the Bank and SOFOMECA, which has recommended a course of action aimed at improving the present financial situation. This study has concluded that the best option was to keep the company in operation, providing that major efforts in finding export markets and in improving productivity are being made. VI. FINDINGS AND LESSONS A. Findings 130. The major findings are related to market and management issues. a. Market Issues 131. Assumptions made in the late 1970s regarding the market were over- optimistic and even proved radizally wrong. It was assumed that the Tunisian industrial sector would continue to grow and become more integrated while pursuing highly protectionist policies and financing large government-owned industrial projects. Such approach proved definitely wrong in 1986 when a deep economic crisis forced the Government to change its economic and industrial policies. This resulted in increased liberalization of imports, decreased protection levels and abandonment of a number of government-financed or supported projects such as CMT and STIA. 132. The assumption was also that CMT and STIA would be ready in time to purchase SOFOMECA's output. However, no real commitment could be obtained at negotiations from the Government regarding STIA production programs. The willingness of the Bank to go ahead despite such lack of complete guarantee proved a major mistake since STIA's closure had, later on, a dramatically negative on SOFOMECA's operations. 133. The assumption that only two customers (CHT and STIA) would alone be in a position to absorb almost 60% of SOFOMECA's increased output, proved equally wrong. Believing that this could effectively happen was taking a major, excessive risk as was demonstrated subsequently. 134. Only lip service was paid to export sales. So much reliance was put on sales to CMT and STIA, and on other domestic sales that 3.5% only of the 1990 output was supposed to be exported. It was only assumed that should difficulties - 28 - arise with CMT and STIA, SOFEMECA would turn to export markets, but experience has shown that building such markets takes considerable time and expertise, and it is much better to integrate export planning at an early stage of the project. b. Management Issues 135. The assumption was made that because SOFOMECA's management did not appear to be very strong, an experienced consultant would be put in charge of implementing and supervising the project. This proved an unwise decision since, as it occurred, major disputes developed between the company and the consultant due to: (i) the fact that the consultant had assumed unrealistically low levels of actual work to be performed, and (ii) severe personality clashes and adaptation problems on the consultant's part. The Bank and SOFOMECA moved also too fast in accepting the consultant's feasibility study recommendations (some of these made too hastily). This did not give them enough time in carefully determining the following steps: planning, engineering and supervision, with subsequent delays and increased costs. 136. Weak management and hesitant Bank staff (who had perhaps excessively relied on one consultant) could not resolve themselves in removing the consultant as soon as problems started. Also management and Bank staff could not decide (when the question came up in 1985) whether to go ahead with the construction of the new foundry or cancel the project. Strong initial interest showed by the Bank for the project led Tunisian authorities, local development banks, non resident Arab banks and SOFOMECA's management itself to believe that the project was a "World Bank project" which had inevitably to succeed and they consequently did not pay themselves at the project all the attention which would have been necessary. Insufficient supervision efforts (compared to SOFOMECA' s considerable problems) and waning Bank staff attention later on contributed to the failure of the project. B. Lessons 137. The lessons to be derived from this experience are as follows: a. large foundry project has difficult technical and management issues to face. It should thus be considered on its own merits and not just as a "piece" of a global industrial strategy which unrealistically led to the rapid promotion of engineering industries (including castings). A relatively large individual project should not be implemented just because the Bank and the Government have singled out a particular subsector (EMI) or even because a Bank financed line of credit has already been granted to such subsector, as this was the case. Excessively "logical" links between policy support and individual project financing may prove very costly. b. An industrial project should avoid relying on just one or two customers, particularly when full, iron clad commitment cannot be obtained from the start. If such commitment is not available, the project should be reduced in size or postponed. Also, if alternative (export) markets are thought to provide a realistic possibility, they should be immediately explored and not considered - 29 - as an alternative to be studied only when domestic markets fail to materialize. c. If. at appraisal, the project management is thought to be weak, the project should not be financed unless extraordinary measures are taken during project preparation to strengthen it. In view of the key importance of management factors excessive reliance on consultants may prove a risky approach. However, if this proves unavoidable, extra supervision resources have to be made available during project implementation to monitor both management and consultant's performance. Other investors and co-lenders with a stake in the project must also be involved, particularly when they have local experience and contacts. - 31 - ANN BORROWER'S COMMENTS Transcription of the telex sent by Mr. Negib Lahouar from SOFOMECA on May 20. S(All coments were taken into account and incorporated in the PPAR). TLX. 460/91-NL/KM 20.05.91 ATT MR: GEORGE C. MANIATIS / BANOUE MONDIALE CHEF DIVISION PAR INTERIN EVALUATION DE PRETS A L'AJUSTEMENT MACRO-ECONOMIQUE ET INDUSTRIE - DEPARTEMENT DE L'EVALUATION RETROSPECTIVE DES OPERATIONS BANQUE MONDIALE. NS ACCUSONS BONNE RECEPTION DU RAPPORT D'EVALUATION RETROSPECTIVE DU PROJET SOFOMECA, QUE VS AVEZ BIEN VOULU NS TRANSMETTRE ET VS INVITONS A PRENDRE NOTE DES MODIFICATIONS QUE NS AVONS APPORTEES A CERTAINS PASSAGES DU RAPPORT: PAGE -VII- PARAGRAPHE 10: LIGNE ..... 5 REMPLACER "BUREAU D'INGENIERIE ALLEMAND" PAR "BUREAU D'INGENIERIE GF" PARAGRAPHE 12: - LIGNE .... 2 REMPLACER 'UNE FAIBLE CAPACITE DE PRODUCTION' PAR 'UNE CAPACITE DE PRODUCTION LIMITEE" - LIGNE... 1-5 REMPLACER "CAPACITE DE 1500-1800 TONNES" PAR "CAPACITE 1800-2000 TONNES" - LINE. .... 18 REMPLACER "59%" PAR 57% PAGE - 4 - PARAGRAPHE 15: - LIGNE.....8 REMPLACER "IL ETAIT PREVU DE CREER DANS LE CADRE DU PROJET SOFOMECA DES CARRIERES DE SABLE ET UNE USINE DE TRAITEMENT DE SABLE" PAR 'IL ETAIT PREVU DE RECHERCHER DES CARRIERES DE SABLE LOCAL ET D'Y ERIGER UNE USINE DE TRAITEMENT DE SABLE" 32 - PAGE - 6- PARAGRAPHE 24: RAJOUTER AU TABLEAU DES REALISATIONS DE L'ANNEE 1990 .... PRODUCTIQ FONTE . . . 9978 ACIER . . . .=11l TOTAL . . . 10996 FONTE . . . 10106 ACIER . . . _Z1 TOTAL . . . 10887 PAGE - 7- PARAGRAPHE 25: - LIGNE .... 4 REMPIACER "10340 TONNES EN 1990" PAR "9978 TONNES EN 1990" PARAGRAPHE 27: REMPLACER LES COLONNES REALISATIONS DU TABLEAU PAR LES CHIFFRES REELS 1990 CAPACITE PRODUCTION UTILISATION (2EQUIPES.) CAPACITE (%) - VIEILLE FONDERIE 6800 T. 4843 71 - NOUVELLE " 12500 T. 6153 49 TOTAL 19300 T. 10996 57 PARAGRAPHE 29: - LIGNE ....5 REMPLACER "ET PARCE QU'IL EXISTAIT" PAR "PAR AILLEURS IL EXISTAIT" - LIGNE .... 6 REMPLACER "SERVICES" PAR "ATELIERS" - LIGNE ....7 REMPLACER "C'EST EGALEMENT A CAUSE DES FAUTES" PAR "ET D'AUTRES FAUTES" - LIGNE ..8-9 REMPLACER "DE lA PROPRETE ET DE LA SECURITE" PAR "DE L'ORGANISATION DES CHANTIERS" AVANT DERNIERE LIGNE PAGE 7 REMPLACER "1500-1800 TONNES" PAR "1800-2000 TL:NES" PAGE - 8 - PARAGRAPHE 30: - LIGNE .... 3 REMPLACER LA PHRASE COMMENCANT PAR "DE PLUS, LA DIMENSION DES CREUSETS DES CUBILOTS, A ETE REDUITE, DE FACON A REDUIRE LE DEBIT HORAIRE POUR S'ADAPTER AU NIVEAU DE PRODUCTION...." ......... LE NUMERO DE CE PARAGRAPHE DEVIENT -31- - 33 - PARAGRAPHE 31: • LIGNE ... .1 RAJOUTER DEVANT TRAITEMENT DE SABLE LE MOT "USINE DE" ..... LE NUMERO DE CE PARAGRAPHE DEVIENT -32- PARAGRAPHE 32: - LIGNE .... 1 REMPLACEÀ "L'USINE DE TRAITEMENT DE SABLE" PAR "LA SABLERIE DE LA NOUVELLE USINE" - LIGNE .... 2 REMPLACER "MINIMA" PAR "NOMINALE" - LIGNE ....2 REMPLACER "L'USINE DE LA SOFOMECA" PAR "CET ATELIER" PARAGRAPHE 33: - LIGNE .... 1 REMPLACER "SE POSE" PAR "EXISTE" - LIGNE .... 3 REMPLACER "NOTAMMENT" PAR "ENTRE AUTRES" PARAGRAPHE 36: - LIGNE .... 5 REMPLACER "D'USINAGE" PAR "DE MONTAGE" PAGE - 9 - PARAGRAPHE 36 - LIGNE .... 4 REMPLACER "ALLEMAND" PAR "GF" PARAGRAPHE 37 - LIGNE .... . ENLEVER LE MOT "MALCRE" - LIGNE .... 2 RAJOUTER APRES LE MOT (KHD), "S'EST LIMITEE A LA FORMATION DE CADRES DANS LES USINES DE CETTE ENTREPRISE" - LIGNE .... 3 REMPLACER "JUSQU'A RECEMMENT, LE SERVICE CHARGE DE LA MAINTENANCE N'A PAS ETE MIS EN PLACE" PAR "A PARTIR DU 4EME TRIMESTRE 1989 UNE DIVISION CHARGEE DE LA MAINTENANCE A ETE MISE EN PLACE" PARAGRAPHE 38: - LIGNE .... 6 REMPLACER LE MOT "D'EDUCATION" PAR "DE FORMATION" PARAGRAPHE 41: - LIGNE ... .4 REMPLACER "4,1" PAR "6,0" REMPLACER "9.6" PAR "9,7" PAGE - 10 - PARAGRAPHE 45: -LIGNE ..... 2 RAJOUTER APRES 58,4 MILLIONS DE DOLLARS "(APRES SUPPRESSION ET REDUCTION DE CERTAINS LOTS D'EQUIPEMENTS.) - 34 - PAGE - 15- PARAGRAPHE 64: - LIGNE .... 2 REMPLACER LE MOT "KWF" PAR "KFW" - LIGNE .... 4 REMPLACER LE MOT "KWF" PAR "KFW" PAE-20- PARAGRAPHE 83: - LIGNE .... 3 REMPLACER "UN BUREAU D' INGENIERIE INDUSTRIELLE ALLEMAND, TANDIS QU'UN AUTRE BUREAU D'ETUDES ALLEMAND" PAR "LE BUREAU D'INGENIERIE INDUSTRIELLE GF, TANDIS QUE KHD FOURNISSAIT...." - LIGNE .... 5 REMPLACER LE MOT "ALLEMAND" PAR "GF" PARAGRAPHE 84: - LIGNE .... 2 REMPLACER LE MOT "ALLEMAND" PAR - LIGNE ....8 - LIGNE ... 11 "GF" PAGE - 21 - PARAGRAPHE 25: - LIGNE .... 3 REMPLACER "D'ACIER" PAR "FONTE" - LIGNE ... .4 RAJOUTER DEVANT "LA FONDERIE D'ACIER" LE MOT "CELUI DE ...." - LIGNE .... 5 REMPLACER TOUTE LA LIGNE PAR "A COMMENCE FIN 1985 ET A ETE ACHEVE EN DECEMBRE 1987" PAGE - 23 - TABLEAU PRODUCTIVITE - LIGNE .... 2 CORRIGER CHIFFRES PRODUCTION (TONNES) VIEILLE FONDERIE 4524 AU LIEU DE 6524 1345 AU LIEU DE 1245 PAGE - 25 - PARAGRAPHE 99: REMPLACER CHIFFRES REBUT ANNEE 1990 PAR IES CHIFFRES REELS: - NOUVELLE FONDERIE..................... 5,7 - VIEILLE FONDERIE FONTE ...............12,9 - FONDERIE ACIER .......................12,8 PARAGRAPHE 101: CORRIGER CHIFFRES REALISATIONS PRODUCTION 1990 COMME SUIT: - PIECES DE FONTE ......................9978 - PIECES D'ACIER ...................... 1018 10996 - 35 - CORRIGER CHIFFRES UTILISATION CAPACITE COMME SUIT .... PREVISION (REPI REALISATION PIECES FONTE 68.2% 57% PIECES ACIER 100.x 56.6% PARAGRAPHE 102: - LIGNE .... 10 REMPLACER 17900 TONNES PAR 17500 TONNES PAGE - 27 - PARAGRAPHE 108: - LIGNE ...... 1 REMPLACER CE PARAGRAPHE PAR: "LES PERTES NETTES DE LA SOFOMECA SE CHIFFRAIENT A 6,0 MILLIONS DE DT EN 1989 - LES COMPTES DE LA FONDERIE D'ACIER ETAIENT DEFICITAIRES DE 0,3 MILLION DE DINARS - DANS LA VIEILLE FONDERIE DE FONTE, LES PERTES SE MONTAIENT A 0,5 MILLION DE DINARS TANDIS QU'ELLES ATTEIGNAIENT UN MONTANT DE 5,2 MILLIONS DE DT DANS LA NOUVELLE FONDERIE ET CELA A CAUSE DU FAIBLE VOLUME DES VENTES" ENFIN, VS DEMANDONS DE CHANGER LE MOT "GOUVERNEMENT" PAR "ETAT". MEILLEURES SALUTATIONS NEGIB LAHOUAR / PDG SOFOMECA