Colombia - Special study - developmental impact of financial assisted projects

Report No. 1037-CO Colombia: Special Study Developmental Impact of Financiera-Assisted Projects January 23, 1976 Latin America and Caribbean Projects Department Development Finance Companies Department FOR OFFICIAL USE ONLY Document of the World Bank 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. L E C O P y  CURRENCY EQUIVALENTS (Average Exchange Rate-Selling) 1967 1968 1969 1970 1971 1972 1973 1974 US$1.0 = Col$ 14.54 16.33 17.36 18.49 19.98 21.92 23.67 26.14 Col$1.0 = US$ .0688 .0612 .0576 .0541 .0501 .0456 .0422 .0382 ABBREVIATIONS BR Banco de la Republica CEDE Centro de Estudios Sobre Desarrollo Economico CFP Corporacion Financiera Popular DFC Development Finance Company GDP Gross Domestic Product IER Internal Economic Rate of Return IFF Industrial Financing Fund IFR Internal Financial Rate of Return OED Operations Evaluations Department PIF Private Investment Fund This report was prepared by Messrs. Ernst Loeschner, David Cook and Javier Nogales based on a study carried out by consultants from CEDE (Centro de Estudios Sobre Desarrollo Economico, Andes University, Bogota). Messrs. Paul Knotter, Luis Liberman and Ernst Loeschner collaborated with CEDE in designing the field study.  FOR OFFICIAL USE ONLY TABLE OF CONTENTS Page I. INTRODUCTION 1 II. OVERALL SOCIO-ECONOMIC IMPACT OF PROJECTS IN SAMPLE 4 ( i) Financial and Economic Efficiency 4 ( ii) Employment Generation Impact 9 (iii) Income Distribution Impact 12 III. COMPARATIVE RESULTS 16 ( i) Comparisons between different sizes of projects 16 ( ii) Comparisons between projects sponsored by firms 17 of different sizes (iii) Comparisons between sectors 20 ( iv) Regional comparisons 23 IV. SUNMARY, CONCLUSIONS-AND RECOMENDATIONS 24 ANNEXES: I Annex 1 Characteristics of Financiera Lending Annex 2 Sizes of Sponsoring Firms Annex 3 Project Information Annex 4 Partial Indicators of Economic Benefits Annex 5 Direct Employment Impact of Projects Annex 6 Indirect Employment Impact on Unskilled Labor Annex 7 Distribution of Net Economic Benefits by Income Group Annex 8 Relative Project Size and Profitability Annex 9 Ranking Orders of Internal Rates of Return Annex 10 Cumulative Distribution of Project Sizes in Sample Annex 11 Cumulative Distribution of Lending by Cmpany ?ize Annex 12 Efficiency Comparisons by Size of Projects Annex 13 Resource Allocation by Sector and Region Annex 14 Correlation in Sectoral Ranking Annex 15 Labor Productivity and Wages Annex 16 Methodology to Evaluate the Impacts of Financiera- Assisted Projects on Foreign Exchange, Employment and Income Distribution 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.   COLOMBIA SPECIAL STUDY DEVELOPMENTAL IMPACT OF FINANCIERA-ASSISTED PROJECTS I. INTRODUCTION 1.01 This study is sixth in the series of Bank Special Studies on the development impact of DFC (development finance company) operations. The previous studies of DFCs in Korea, Turkey, India, Tunisia and Iran concen- trated on two major elements. The first entailed an assessment of the socio- economic benefits of projects financed by the DFCs. The second was an evaluation of the developmental impact of the DFCs as institutions, including their contribution to the process of mobilizing and allocating funds, their effect on the economic and institutional environments, their relationships with governments and clients, and the evolution of their policies in the face of changing circumstances. 1.02 The Colombian Special Study differs from the five preceding ones in some important respects. Because there were originally as many as five, and currently seven, private institutions (financieras) through which the Bank has been channeling credit to Colombian private industry, it was decided to concentrate primarily on the developmental impact of DFC-financed sub-projects, rather than on the wider development functions of the institutions. In addition, this study investigated the socio-economic impact of DFC lending in considerably greater depth than the previous ones. Thus it attempted to assess not only the economic and financial efficiency of sub-projects but also their effect on direct and indirect employment generation and on income distribution, and to examine the variations of all these factors according to industrial sector, location and size of sponsoring companies. The primary objectives were both to increase the Bank's understanding of the socio-economic impact of DFC lending and to suggest ways in which to improve future DFC operations in Colombia and elsewhere. Background to the Financieras 1.03 The Colombian financieras were established to provide long- term financing, both debt and equity, to industry and other productive enter- prises, since commercial banks were providing mainly short-term credit. The oldest and largest of the financieras, Corporacion Financiera (CF) Colombiana, came into operation in 1959 in Bogota. Between 1959 and 1964, four other financieras j were set up in Colombia's main industrial-commercial regions. Since then, ten more financieras have been established; however, they are mostly much smaller than the initial five. 1.04 Bank Group assistance to the financieras started with IFC equity investments in CF Colombiana and CF Nacional in 1961; subsequently, TIFC invested in each of the other three older financieras. The first Bank DFC loan of US$25 million to Colombia (451-CO) was signed in 1966. The loan was made to Banco de la Republica (BR, Colombia's Central Bank) for onlending via the five older J Corporaciones Financieras Nacional, del Valle, de Caldas and del Noarte  -2- financieras to cover the foreign exchange costs of projects of productive enterprises in the private sector. Together with four later loans, Bank lending via these financieras now totals US$163 million, excluding a US$10 million portion of the Development Program and Export Expansion Loan (842-CO of June 28, 1972) to which the financieras also had access. 1.05 Since 1966 lending by the financieras has grown rapidly (at about 25% annually) and they now represent the most important source of institutional credit for manufacturing industry. World Bank credit lines are the single most important source of funds for the financieras, accounting for about 36% of their loan portfolios and 27% of their total assets. However, they also have access to other major credit lines administered by BR j/, in addition to their own equity and bond issue resources. The financieras have extended loans to all but the smallest industrial companies, but comparisons between the projects financed via IBRD loans and other BR credit lines indicates that IBRD funds have usually been used for the larger loans required to finance the projects of the larger companies (Annex 1, Exhibit 2). 1.06 On average, financiera loans have covered about 25% of the costs of large projects and a higher proportion for small ones. Loans have been made for projects in almost all regions of Colombia but the volume of lending reflects the industrial-conmercial concentration in the five regions where the financieras are located (accounting for 90% of lending). Similarly, lending has been spread widely over the different industrial sectors, but six sectors-- textiles, food and beverages, chemicals, paper, non-metallic mineral products and metal mechanics industries--account for the bulk of financiera lending (Annex 1, Exhibit 4). 1.07 The operations of the financieras were examined as part of a study of past Bank lending to Colombia carried out in 1971 (Operations Evaluation Report: Colombia, Volume V--The Corporaciones Financieras). This study pointed out that a high proportion of the proceeds of Bank loans had been concentrated on a relatively small number of large manufacturing enterprises, and raised some doubts, on the basis of aggregate data and the use of partial indicators,as to the economic and financial efficiency of large manufacturing enterprises relative to those of medium and small scale. 1.08 As a result of the OED suggestions the Fifth Financiera Loan (903-CO; signed in June 1973) incorporated some new features designed to secure a wider distribution of Bank funds. First, three additional financieras were allowed to participate (although only two--Occidente and Santander--eventually qualified).j/ O of which the Private Investment Fund (PIF) and Industrial Financing Fund (IFF) are the most important (see Annex 1). j In addition, a separate loan was prepared for small scale industry to be channeled through the Corporacion Financiera Popular (CFP), a government- owned DFC. The CFP loan was signed in January 1975. Only companies with total assets of less than Col$20 million are eligible to participate in the loan--a group that received less than 10% of previous Bank lending via the financieras.  - 3 - Second, a limit of US$4 million was placed on the amount of IBRD funds that could be outstanding (from all financieras) to any one industrial concern, including its subsidiaries. Exceptions to this limit would only be granted for projects with high economic priority and demonstrated lack of alternative sources of finance at reasonable terms. Other refinements introduced included the requirement that Economic Rate of Return calculations would be performed on any project for which the Effective Rate of Protection exceeded 25%. 1.09 In the light of the above features, a comparison between projects and firms of different sizes has been a subsidiary objective of this study. This comparison touches upon the socio-economic characteristics of projects/ firms of various sizes and on the access to credit by firms in different size groups. Method of Investigation 1.10 The results of this study are based on a sample comprising 29 projects which had been in operation for at least one year. Results were computed from actual figures for historical periods and latest forecasts for the remaining years of projects' lives. In choosing the projects a stratified sampling technique was employed. Since an important focus of the study was to gauge the develop- mental contribution of projects which had received Bank funds, only those projects where the Bank had financed part of the investment cost were chosen. The sample was also selected to reflect the relative financing shares of the five financieras and the relative importance of various industrial sub-sectors in their portfolios. Moreover, it was important that an adequate number of relatively small and large projects were represented as well as various sizes of sponsoring firms. 1.11 Although the list of projects cannot be described as a completely random sample, it is representative of the distribution of past financiera lending (see Annexes 1 and 11). There are six industrial subsectors represented in the sample as follows: textiles with seven projects, chemicals with six projects, and non- metallic minerals, food processing, paper and metal industries with four projects each. These six sub-industries account for about 75% of historic financiera lending and 69% of Colombian industry's contribution to GDP. The five financieras are represented as forlows: Colombiana with eight projects, Caldas with three, and Na6ional, Norte and Valle with six projects each. 1.12 The project results are based on field work and analysis undertaken by consultants from CEDE (Centro de Estudios Sobre Desarrollo Economico, Andes University, Bogota). Terms of reference for the study were drawn up jointly. CEDE obtained the project data through field visits to all 29 companies and the five financieras. Financiera staff was involved in most field visits and Bank staff members participated in some. CEDE developed a comprehensive computer ,program to analyze the project data and subsequently prepared a detailed report describing their methodology and explaining their findings and conclusions. 1.13 Unlike the previous five studies, most of which had followed a derivative of the OECD method, the Colombian study used the UNIDO approach j to economic project analysis. However, the difference in the methodologies employed is more in degree than in kind. For instancp, the method used For a complete description of the methodology, see Annex 2 of CEDE's report Estudio de Proyectos Financiados por Corporaciones Financieras, Bogota, november 1974, and UNIDO, Guidelines for Project Evaluation UN, New York, 1972.  -4 - in this study to gauge economic efficiency was complex in that an attempt was made to estimate the effect of economies of scale in production and price controls on consumer or producer surplus. Furthermore, material inputs and outputs were valued differently depending on whether their effect was to influence domestic consumption or production, or to change the level of imports or exports. The employment and income distribution impacts of projects were analyzed by following a methodology which is described in summary form in Annex 16. The methodological differences are, however, of minor importance for purposes of comnaring the Colombian results with those of the Drevious DFC studies. 1.14 In order to interpret the results of the study, it is important to understand the concept used to define a project. Only two of the 29 projects included in the sample were sponsored by new companies. The rest involved the addition of new product lines or the expansion of production by existing companies. Since a particular financiera-assisted project was often preceded or followed by investments related to it, it was in many cases necessary, for the purpose of this study, to redefine the "project' as comprising all invest- ments related to the product of the financiera-assisted project which were undertaken by the sponsoring company after a specific date. The date chosen varies from company to company but in most cases the project consists of all investments made after 1969. II. OVERALL SOCIO-ECONOMIC IMPACT OF PROJECTS IN SAHPLE 2.01 This chapter describes the results obtained from all 29 projects in the sample. It attempts to analyze the effects which these projects have had in terms of financial and economic efficiency and their impacts on employment generation and income distribution. Annexes 2 through 7 contain relevant data, project by project. (i) Financial and Economic Efficiency 2.02 Overall, the 29 financiera-assisted projects studied exhibited a high degree of efficiency in the utilization of capital, both financially and in terms of net aggregate consumption. The average financial and economic rates of return (weighted according to their investment costs) were 18% and 32%, respectively. These rates compare favorably with the estimated marginal product- ivity of capital of 10% j and the social marginal preference for consumption over time in Colombia (12-18%).j/ 2.03 The individual project results that yielded the above averages showed a wide spread between projects, particularly for economic rates of return. There were also unexpected extremes in terms of absolute magnitude of financial and economic returns. Nevertheless, the averages are not particularly / Harberger, "The Marginal Capital Productivity in Colombia", 1968. For the methodology of calculation, see UNIDO Guidelines for Project Evalua- tion, 1972, Chapter 10. In this study, a 12% discount rate has been used to calculate the net present values of aggregate consumption.  sensitive to the deletion of extreme values: 21 projects (i.e. 70% of the sample) are either financially or economically attractive, with IFRs (Internal Financial Rates of Return) above 10% and IERs (Internal Economic Rates of Return) above 18% respectively. Two-thirds of the projects were attractive in both financial and economic terms. Annex 4 shows the returns of each project and the table below sunmarizes their distribution. Table 1 Distribution of Project Returns Number of Projects with Returns Between Neg-5% 5-10% 10-20% 20-30% 30-50% 50-75% 75+ % Financial Returns (IFR) 4 4 6 7 4 4 - Economic Returns (IER) 3 2 4 6 5 4 5 2.04 Some Extreme Cases. There were five projects with an IER below 10% and four of them also showed poor financial returns. These four projects were all affected by marketing problems that led to low margins and lower than expected sales. Two synthetic textile projects were affected by rapidly rising raw material costs and price-sensitive domestic demand. A project to produce paper bags and laminated paper operated on a very low margin arising from heavy domestic competition. The fourth project, involving the production of laundry soap, has not yet broken even because price controls were imposed on substitute products (detergents). Although the events which adversely affected these projects were not easily forseeable, the quality of the marketing analysis carried out on these projects (their total investment cost of Col$l9 million was considerable), and on some of the successful projects, did not appear to be fully adequate. In fact, few appraisals analyzed the sensitivity of return projections to price and volume changes or contained systematic demand surveys. 2.05 The fifth project with an economic rate of return below 10% was perhaps the most striking. It showed a significant divergence between financial and economic benefits (IFR 19%, IER 0.5%). This project, to produce chewing gum, involved the substitution of labor by imported machinery. The project was small (Col$1.5 million) but enabled significant operating cost savings resulting from the reduction in the number of workers from 112 to 91. Thus the project paid off for the sponsors (40% after-tax return on equity), but the opportunity costs of job elimination and machinery imports produce the low IER. Two other projects in the sample also resulted in a net loss of direct jobs, but in these cases the economic rates of return proved satisfactory. These examples reinforce the need for including IER calculations in project appraisals. IER analysis would have shown, at the time when the chewing gum project was considered, the undesirability of the investment from a socio-economic point of view. 2.00 On the brighter side, there were five projects with economic rates of return in excess of 75%. Their financial returns were also healthy, in three cases ranging from 40% to 60%. Three of the projects were relatively small expansions equivalent in size to less than 10% of the sponsoring companiestassets. In these cases the high returns are in part due to the high leverage effects of expansions eliminating bottlenecks. Nevertheless, common to all these projects  - 6 - were low capital/output ratios which enabled them to export a substantial proportion of their production i/ and to command major shares of the domestic markets for their products (plastic laminates, metal nuts and springs, toilet paper and polyethylene bags). 2.07 The most interesting project had a wide divergence of financial and economic returns (IFR = 15%, IER = 140% which is highest among all projects in the sample). The project was comparatively large (Col$27 million) and represented an expansion equivalent to 30% of the company's assets. The subject of production is toil,et paper. A combination of three factors contribute to the unusually high IER: a low-cost input (old newspapers); significant employment of unskilled labor in the collection of waste paper in various parts of the country; and price controls on the final project (CEDE estimated that consumers would be willing to pay about 20