Colombia - Manufacturing sector developments and changes in foreign trade and financial policies (Vol. 2) : Detailed analysis and appendices

Report No. 4093-CO F C(olombia tManufacturing Sector Developments and C(hanges in Foreign Trade and Financial Policies (In Two Volumes) Volume II: Detailed Analysis and Appendices January 21, 1983 Projects Department Latin America and the Caribbean Regional Office FOR OFFICIAL USE ONLY Document of the World Bank This docurnent 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.  CURRENCY EQUIVALENTS 1/ Currency Unit = Colombian Peso (Col$) US$1.0 - Col$70.11 Col$1.0 = US$.0143 Col$1.0 million = US$14,164 GLOSSARY OF ABBREVIATIONS ANDI Asociacion Nacional de Industriales (National Association of Manufacturers) ASOBANCARIA Asociacion Bancaria (Banking Association of Colombia) BR Banco de la Republica (The Central Bank) CAF Corporacion Andina de Fomento (Industrial Financing Agency of Andean Common Market) CAT Certificado de Abono Tributario (Tax Rebate Certificate for Non-traditional Reports) DANE Departamento Administrativo Nacional de Estadistica (National Statistical Office) DNP Departamento Nacional de Planeamiento (National Planning Office) CDTs Certificados de Deposito al Termino (Certificates of Deposit) CEPAL Comision Economica para America Latina (United Nations, Economic Commission for Latin America) CFs Corporaciones Financieras (Investment Banks) ECOPETROL Empresa Colombiana de Petroleo (National Petroleum Company) FEDESARROLLO Fundacion para la Educacion Superior y del Desarrollo (Economic Research Institute in Bogota) FFI Fondo Financiero Industrial (Industrial Financing Fund) FIP Fondo de Inversiones Privadas (Private Investment Fund) IFI Instituto de Fomento Industrial (Official Industrial Development Bank) IFS International Financial Statistics, Published by the Inter- national Monetary Fund (IMF) INCOMEX Instituto Colombiano de Comercio Exterior (Colombian Foreign Trade Institute) PROEXPO Fondo de Promocion de Exportaciones (Export Promotion Fund) PV Plan Vallejo (Import Drawback Device for Exporters of Non-traditional Products) UPACs Unidades de Poder Adquisitivo Constante (Indexed Investments Issued by-Savings and Loan Associations) ll December 31, 1982  FOR OFFICIAL USE ONLY VOLUME II DETAILED ANALYSIS AND APPENDICES Page No. I. EXPORT INCENTIVES AND IMPORT PROTECTION FOR INDUSTRY A. The Export Incentive System ................... ........ 2 B. Imporl: Controls and Effective Protection . .... 13 Annex 1: Formulas for Effective Export Incentives ............ 23 ]II. TRENDS AND DETERMINANTS OF MANUFACTURED EXPORTS AND IMPORTS C. Exports of Manufactured Goods in Recent Years . . 25 D. Exports of Major Subsectors .. . 34 E. Does the Exchange Rate Matter? .......................... 40 F. Recent Growth and Change of Imports . . 42 G. Major Determinants of Colombian Imports . . 46 H. Prospects for Increased Manufacturing Exports and ImporLs ........................................................... 47 Annex 2: A Note on Colombian Export Statistics ............. . 56 III. THE OPENING-UP PROCESS AND ITS IMPACT ON I1DUSTRIAL GROWTH AND DEVELOPMENT I. Imports, Exports and Growth .......... .............. 60 J. Investment, Employment and Productivit:y and their Relation to Foreign Trade Policies,.,. I................... 69 K. Involuntary Import Liberalization: The Case of Textiles. 75 L. Barriers to Efficient Import Substitution and Exports: The Case of Capital Goods ............................... 80 Annex 3: The Textile Industry ............................... 89 Annex 4: The Metal Mechanical Industries ................... 139 IV. LIBERALIZING THE FINANCIAL SECTOR: ITS IMPACT ON INDUSTRIAL FINANCE AND INVESTMENT M. Finarncial Repression and Liberalization Attempts ...... 159 N. Sources of Aggregate Investment Finance, Inflation and Interest Rates ............... 163 0. Structural Changes of Industrial Finance .............. - 168 P. Other Issues in Financial Liberalization .............. - 176 V. STATISTICAL APPENDIX This document has a restricted distribution and may be used by recipients only in the performance of their official dluties. Its contents may not otherwise be disclosed without World Bank authorization.  I. EXPORT INCENTIVES AND IMPORT PROTECTION FOR INDUSTRY 1.01 Since the Second World War, industrialization in Colombia has generally been tied to import substitution policies. Although the various incentives provided to industry did not always respond to a coherent indus- trial strategy, they generally operated withir. a development framework which placed priority on the growth of manufacturing activities through import substituion. High effective protection barriers were the principal policy instrument used to promote industrial growth, supplemented by an overvalued exchange rate and ample availability of subsidized credit. While these policies led to significant inefficiency in dcomestic manufacturing activi- ties, they also encouraged domestic and foreign investments. As a result industrial value added expanded rapidly at an average annual rate of 6.2% p.a. during 1950-66, with the fastest growth taking place during the early 1950s, and a substantial slowdown in the early 1960s. 1.02 The early sixties witnessed a series of balance of payments crises and successive devaluations, which led the aut:horities to reconsider the thrust of economic policy. There was increasLng awareness of the strategy's adverse implications for the balance of payments, since import substitution efforts in more complex industries turned out to be import intensive. The recurrent foreign exchange problems eventually led the authorities to seek a more export-oriented policy. Although isolated attempts at stimulating exports had been undertaken before, it was nol: until 1967 that a coherent set of measures was implemented to promote nontraditional exports.1! 1.03 The most important element of the new policies was the introduction of a crawling peg exchange rate system. A package of export incentives was also introduced, including: (a) fiscal incentives (CATS--Certif'icados de Abono Tributario); (b) concessionary credits for export-related activities from the Export Promotion Fund (PROEXPO); and (c) an expanded and more effective import/export regime (Plan Vallejo). The new policies represented an attempt to compensate for the relative price distortions genierated by the import substi'tution effort. Together with a favorable developrient of world trade, this shift in policy emphasis led to impressive results. Between 1968 and 1974 manufactured exports in current prices increased from US$58 million to over US$390 million, and their share in total exports rose fcrom 8% to 28%. Industrial value added increased by 7% p.a. in real terms, and the growth in manufacturing employment of 8.5% p.a. reached unprecedented levels, to a large part as a result of the relatively more labor-intensive characteristics of the leading export subsectors--textiles, apparel, footwear, an(d leather products. This growth of manufactured and other minor exports also succeeded in easing the foreign exchange problems which had plagued the economy in prior years. Net official internationaL reserves rose by more than US$500 million during the 1967-74 period, from a negative position of US$95 million in 1966. 1/ These included all exports other than ccffee and petroleun. The first export promotion measure implemented in 1960 was the import duty drawback system known as Plan Vallejo.  - 2 - A. The Export Incentive System 1.04 Among the most important factors stimulating the growth of Colombia's manufactured exports during 1967-74 were the acceleration of world trade and the increased competitiveness of Colombian industry in inter- national markets. The sharp expansion of world trade of manufactured gocds has been estimated to have accounted for about one-third of Colombia's export growth, 2/ leaving the largest part to improvements of Colombia's competi- tiveness, which in turn was crucially linked to profitability in exporting. The key determinant of that profitability is the real effective exchange rate: the number of pesos that an exporter receives per unit of foreign exchange earned, adjusted for the difference between domestic and inter- national inflation rates. In Colombia, these earnings depend not only on the rate of devaluation of the peso but also on the effective value of the tax export subsidy (CAT), the effective value of subsidized PROEXPO credit, and the conditions of the import duty drawback scheme (Plan Vallejo); these are discussed in some detail below. Certificado de Abono Tributario (CAT) 1.05 The CAT is a tax certificate issued to exporters, valued at a given percentage of the value of exports (FOB). The CAT is freely negotiable and is traded on the stock exchanges, where exporters may sell it at a market- determined discount rate. At maturity, the owner may use the CAT to pay domestic taxes and duties. The cash value of the incentive is thus deter-- mined by (i) the percentage of export value granted as CAT; (ii) the market discount, which depends on the market interest rate and the maturity period( of the CAT; and (iii) the marginal tax rate of the owner (the CAT represents tax-free income). The total volume of CATs in terms of budgetary cost approached the value of about US$90 million in 1974 (nearly 9% of current revenues of the national government). Reforms of export incentives in subse- quent years reduced the value to an average of about US$40 million for the period 1975-77 (2.8% of current revenues). The 1978 value of the CATs was further reduced to US$32 million, or about 1.5% of current revenues. 1.06 The growing fiscal cost of the CATs until 1974 was the main reason for a sharp reduction of their 15% rate which had prevailed for all non- traditional exports during 1968-74.3/ The reduced percentages of CAT effec- tive for 1975 and 1976 were divided into three categories: (i) most agricul- tural and mining products obtained a CAT of 1% of export values; (ii) a large 2/ See J. P. Wogart, Industrialization in Colombia (Tubingen:Mohr, 1978), pp. 121-124. 3/ See J. D. Teigeiro and R.A. Elson, "The Export Promotion System and the Growth of Minor Exports in Colombia" IMF Staff Papers (July 1973), pp. 419-470.  portion of industrial products received 5%; and (iii) the CAT for other non- traditional products was fixed at 7%. FEDESARROLLO has estimated that the weighted average CAT for all nontraditional exports was reduced through these measures from 15% of export value until 1974 to 3.6% in 1975 and 1976.4/ Changes in 1977 provided for an increase of 82% for most industrial products; major exceptions were textile yarn and fabrics, and wood products, which obtained 5%. The CATs for other exports were somewhat reduced on average, and the weighted average CAT declined to 2.5%. In 1978, the CATs were raised in an attempt to compensate for the lagging exchange rate devaluation, and four different rates were established: (i) 12%' for most industrial products; (ii) 9% for a number of agricultural products and several industrial products, the most important being textile yarn and fabrics! (iii) 5% for agricultural products such as meat, fish, vegetables and tobacco; and (iv) 1% for all other nontraditional products. The average weighted CAT thereby rose to 3.8% in 1978. Effective in 1979, a further increase of CAT percentages was implemented, which raised the subsidies to many of the export products previously receiving 5% to 9%, and reduced the number of products at the minimum 1% level. The weighted average CAT for 1979 was at 6%. In 1980, the 1% rate was abolished, leaving CAT rates at 5%, 9% and 12%. Most industrial products continue to receive a CAT of 12%. 1.07 Eligibility for the above CAT perceatages is based on. a number of different factors. Exports which benefit from the duty drawback system (Plan Vallejo) have to have at least 40% of domestic value added in order to obtain the stipulated CAT on the full FOB value. If the domestic value added is below 40%, the percentage of CAT is applied only to the amount of value added. About. half of the approximately US$1 billion of nontracitional exports in 1978 benefited from Plan Vallejo, but no information is readily available on the share of exports which actually obtained CATs, only on the amount of domestic value added. A second group of important exceptions to the above CAT rates is due to the special trade agreements with the U.S. on flowers (since 1974), leather handbags (effective 1978), and textiles and clothing (effective 1979). In order to avoid implementation oi- counter- vailing duties by the U.S., Colombia reduced the CAT for leather handbags from 12% to 6.5%. In addition to the establishment of export quotas, the CAT for garments was reduced from 12% to 9.6%, ar.d that for textile yarn and fabrics was Lowered from an average of 8.4% to 7.6%. These reductions apply only to exports of these product categories to the U.S. A total annual 4/ FEDESARROLLO, Coyuntura Economica (December 1978), Table VIII-6, page 81. The weights are according to the export structure of 1976. It should be noted that this average CAT, as well as the percentages and values calculated in the following text, are not adjusted for marginal tax rates and maturity periods and are t:hus not strictly comparable to the amounts of foregone fiscal revenues mentioned in the previous paragraph.  - 4 - export volume of about US$40 million is affected by the latter twoagree- ments.5/ In addition to domestic fiscal considerations, the threat of countervailing duties has thus become another factor in moderating these incentives. 1.08 A number of reasons for the implementation of the above changes in the level and structure of CATs have been given by the authorities. These have changed over time, and have not always been consistent with each other or with the actual CAT system. The most important reason for the introduc- tion of CATs was the objective of providing exports with an overall compensa- tion for an overvalued exchange rate; consequently, an across-the-board CAT was established for all nontraditional exports. The drastic reduction in CAT rates in 1974, on the other hand, was clearly based on the objective of reducing the budget deficit.6/ After 1974, the strong differentiation of CAT percentages and their frequent changes reflect the desire to use the CATs as a more finely tuned instrument of export promotion. Among others, the following arguments have been mentioned in favor of differentiated CATs:7/ (i) exports with high domestic value added deserve a higher CAT than other products; (ii) exports with insufficient or variable domestic supply (such as agricultural products) should have a low CAT; (iii) exports with good poteln- tial in foreign markets only require a low CAT. More recently, the criteria mentioned for the differentiated increase of CATs in 1978 include the concept of products "critical" for the domestic market and the desire to support industrial products facing strong international competition. Moreover, it would appear that the 1979 rate increase for a substantial number of exports which were previously at the 1% and 5% levels was due partly to the Govern-- ment's announced intentions to use the CAT as an instrument of industrial decentralization and for the selective promotion of labor-intensive export products.8/ 5/ The quotas and the CAT reductions are applied only to exports which enter the U.S. under the Generalized System of Preferences (GSP). About US$20-30 million of garments exports entered the U.S. in 1978 under the assembly provision of U.S. tariff item 807, where quota and subsidy restrictions are not applied. 6/ However, at the time of this policy revision, the authorities did indicate their concern with the impact of these changes on the competitive position of exports, and for a period of time the rate of devaluation was speeded up. 7/ These criteria are listed by INCOMEX in Comercio Exterior, Vol. 10, No. 3. 8/ See the text of President Turbay's speech at the annual meeting of ANDI in 1978.  -5- PROEXPO Credit 1.09 'When the CAT incentive system was cuictailed in 1974, the supply of subsidized export credit from PROEXPO was expanded, and new credit lines were added. In 1974, about US$150 million of short-term credits helped to finance US$680 million of nontraditional exports (22% coverage). The 1978 volume extended by PROEXPO credit was about US$600 miLlion, financing exports of about US$1 billion (60% coverage). Credits to industrial exports during the same period rose even more rapidly from US$101 million (26% coverage) to US$416 million (78% coverage). The expansion of PROEXPO credit was made possible by increased funding. PROEXPO's principal source of funds--an import surcharge on the CIF value of imports--was raised during this period from 1.5% to 5%. This surcharge represents about 20% of total tariff revenues. Export credits are granted through three major lines: (i) a short-term credit line, established in 1972, for the financing of working capital requirements. This credit line is in national currency, with a volume equivalent to about US$500 million in 1977 and US$524 million in 1978; (ii) a medium--term credit line, established in 1974, which finances working capital requirements for export production as well as fixed capiLtal expenditures. Credit is granted for up to five years (three years on average); credlit approvals were US$18 million in 1977 and US$33 million in 1978; (iii) a short-term post-shipment credit line in US dollars established in 1975. Credit approvals in 1977 were US$39 million and about US$50 million in 1978. 1.10 In addition to these three major credit lines, PROEXPO has also established a number of smaller credit lines: (i) a line in US dollars jointly with the Corporacion Andina de Fomentc. (CAF) to finance exports to the Andean Pact; (ii) credit support for the development of frontier areas; (iii) assistance to the Corporacion Financiera Popular for export financing of medium and small-scale enterprises; and (iv) financial assistance to the Instituto de Fomento Industrial (IFI). The total volume of these credit lines was US$20-25 million in 1980. 1.11 The increasing coverage of PROEXPO credit has provided liberal provision of export credit financing. While t:he authorities' monetary policy has aimed at restricting the expansion of domestic credit, since inflation accelerated in 1967-77, the increasing availability of credit from PROEXPO contributed to relieve the tight credit situation faced by the industrial sector. During the five-year period 1974-78, PROEXPO's credit approvals increased at an annual average rate of about 42% in current terms, compared with a growth rate of 31% in credit from commercial banks and financieras. PROEXPO thus constitutes an increasingly important source of credit to the manufacturing sector, with nearly 90% of its credit outstanding being of a short-term nature.9/ 9/ For further discussion of PROEXPO credits and their importance for overall industrial credit policy, see the discussion in Chapter IV.  -6- 1.12 PROEXPO credit provides an interest rate subsidy, which has become more important in recent years as a result of the wider differential between PROEXPO interest rates and the interest rates charged by commercial banks. The value of this incentive increased almost fivefold between 1974 and 1980, as the authorities consciously attempted to compensate for the lagging exchange rate movement with more favorable lending terms. The level of the PROEXPO credit subsidy reached 7.5 cents for every dollar of exports in 1980. Since the subsectoral distribution of PROEXPO credit was quite similar to the actual percentage composition of exports by product categories, the interest rate subsidy operated as an across-the-board incentive. Plan Vallejo (PV) 1.13 The import duty drawback system, known as Plan Vallejo, constitutes the oldest instrument of export promotion for manufactured exports. Under this scheme, export manufacturers establish contracts with the Government for the import of inputs and machinery free of import duties and other import charges. The exporter has three advantages when using Plan Vallejo: (i) aL reduction in production costs, which is especially valuable for export products with a high share of imported inputs; (ii) increased flexibility in the choice of production inputs; and (iii) a reduction of the "red tape" connected with usual import procedures. In addition, import applications under Plan Vallejo face a smaller risk of rejection by INCOMEX, which allows for better production planning. Nearly half of all nontraditional exports in 1977-78 benefited from duty-free imports under Plan Vallejo, compared to a little over one-third in 1974-75. The increased importance of Plan Vallejo in 1977-78 compared to 1974-75 reflects the significant increases in domestic costs relative to international prices between the two periods. Virtually all textile exports benefited from Plan Vallejo. Paper and printing (69%), chemicals (51%), and metal products (53%) also had above-average utilization of the duty drawback scheme. The apparent low utilization for garments (29%) was partly due to the registration of a portion of the exports with their domestic value added only. Also, as in the case of machinery, a large number of garment exports are directed to the Andean Pact countries, and thus do not benefit from Plan Vallejo. On average, during 1977-78, the value of annua:L Plan Vallejo imports was equivalent to 30% of the value of exports benefiting from the program. 1.14 By providing access to inputs at international prices, Plan Vallejo reduces the costs of exporting enterprises that might otherwise be forced to use domestically produced inputs at higher prices. It has been estimated that the cost reduction achieved by Plan Vallejo users during the period 1972-76 averaged about 6.5 cents per dollar of exports.10/ Given the recen,t differences in external and domestic price trends, it is likely that the cost 10/ See M. H. Cardona, "El Crecimiento de las Exportaciones Menores y el Sistema de Fomento de las Exportaciones en Colombia", Revista de Planeacion y Desarrollo (Bogota, 1977).  -7- reductions in later years was somewhat higher. Plan Vallejo has thus provided a substantial benefit to nontraditional exports by reducing the disincentives inherent in the import protection system. Moreover, the benefits provided under Plan Vallejo have been quite stable over time. However, the complicated administrative procedures involved have tended to favor large enterprises, and the delays inhereni: in the scheme have signifi- cantly reduced its value to exporters. Real Exchange Rate 1.15 With domestic and international inflation rates differing signifi- cantly, Colombian policy makers have adjusted the nominal exchange rate on a continuous basis since 1967. As a consequence, it becomes important to examine movements in the real exchange rate which measures the extent to which the rate of devaluation of the peso has offset the difference in the rate of increase of prices in Colombia on the one hand, and in its trading partners on the other. Two different real exchange rates are usually calculated in Colombia: peso-US dollar, and peso-weighted average-basket-of- currencies ("peso/weighted average"). Table 1 presents the development of the nominal and real exchange rates between 1967 and 1980. It reflects the most consistent: of four separate sets of estimates for the peso-dollar real exchange rate. The analysis in all cases yielded essentially the same results, although exchange rates and U.S. and Colombian price indices used were not identical in the different sets of estimates.1!/ The real peso-dollar exchange rate rose consistently from 1967 through 1971, II/ These four estimates were undertaken by Banco de la Republica (BR), Asociacion Bancaria (A), FEDESARROLLO (F), and Morawetz (M). See Statistic.al Appendix Tables 1 and 2. One further set of estimates, by ANDI for 1970-80, gives results very similar to these. (Revista ANDI No. 54, 1981, p. 71).  -8- Table 1: NOMINAL, REAL & REAL EFFECTIVE EXCHANGE RATE, 1967-80 Real Exchange Rate Real Effective Exchange Rate Nominal Exchange Peso/iultiple Peso/'fultiple Year Rate Peso/US$ Currencies a! Peso/US$ Currencies 1967 14.88 83.8 80.4 82.3 79.0 1968 16.48 88.3 85.6 87.8 85.1 1969 17.53 89.4 86.5 89.0 86.1 1970 18.68 94.9 89.5 94.8 89.4 1971 20.26 100.0 100.0 100.0 100.0 1972 22.14 99.3 100.4 99.5 100.6 1973 23.98 95.2 99.5 95.8 100.1 1974 26.66 97.9 107.2 100.7 110.3 1975 31.58 100.4 109.5 93.0 101.5 1976 35.21 97.2 107.0 90.7 99.8 1977 37.20 81.3 93.2 79.6 91.3 1978 39.10 79.9 95.0 81.7 9R7.2 1979 42.55 76.5 90.3 78.7 932.8 1980. 47.28 79.9 94.0 83.2 97.7 1981 b/ 56.26 76.5 87.2 79.6 90.7 Source: IMF, International Financial Statistics, Asociacion Bancaria Colombiana,-and Tables 1 - 4 in Statistical Appendix. a/ Includes 15 of Colombia's most important trading partners (See FN 1, p. 19). b/ January to June 1981.  -9- fluctuated around the 1971 level during 1972-75, and then fell by 20-25% during 1976-81, with the largest single decline (15-16%) occurriRg in 1977.12/ 1.16 Table 1 also presents an estimate for real peso-weighted average exchange rate, based on research at Banco de la Republica (BR), ASOBANCARIA (A), and FEDESARROLLO (F).13/ These series indicate that the real peso- weighted average exchange rate rose consistently during 1967-72, reached a new peak in 1975-76, but then declined continously until June 1981, when it was 13% below what it had been in 1971, and 20% below what it had been in 1975. 12/ Different computations of the real exchange rate have used dlifferent base years; often 1967=100 has been used as the base, but sometimes it has been 1970, 1971, or 1975. In the present report, all series have been converted to 1971=100. Partly, this is for convenience of presentation. But partly, too, there is significance of a sort to the choice of year. If 1967 was set equal to 100, the real peso-dollar exchange rate for June 1981 would have read 91. It might thus seem that the real exchange rate for 1981 is "only 9% below what it w<1s in 1967". But 1967 wvas a year of extreme foreign exchange shortage, i.e., the real exchange rate in mid-1981 was 9% lower than it had been during the worst foreign exchange crisis of the last 20 years. By contrast, setting 1971=100 allows one to see at a glance that by mid 1981, the real exchange wgas at least 25% lower than what it had been a decade earlier, in 1971, when exporting was still profitable in Colombia. 13/ The first two of these sets of estimates yield similar results and are probably more reliable, for reasons of more consistent coverage of countries and sources. During 1967-77, the F series covers 99 countries, but thereafter there are only L1, including U.S., Canada, Japan, W. Germany, France, U.K., Italy, Belgium, Netherlands, Sweden and Switzerland, but excluding one of the most: important markets for Colombia's non-coffee exports, Venezuela. By contrast, the A series consistently includes 15 countries (F's e:Leven plus Colombia's four Andean Group co-members), and the BR series consistently includes 12 countries (F's eleven minus Canada and Sweden plus Spain, Venezuela and Ecuador). The A series uses IMF, International Financial Statistics (IFS) throughout, and corrects the figure, when provisional data are later changed, as often occurs. The BR series also uses IMF, IFS throughout. The F series, by contrast, uses this source during 1967-77, but then switches to The Economist, whose data are available with less delay.  - 10 - Effective Export Incentives 1.17 The computation of the effective value of the export incentives (CAT + PROEXPO credit) is presented in Table 2. The formulas used to calculate these effective values are presented in Annex 1.14/ As indicated earlier, the average value of the CAT as a percent of the value of exports rose very gradually from 14% in 1967 to 18% in 1974, fell abruptly to 6,%' in 1975 (when the nominal CAT was cut from 15% to around 5%), and then increased to 13-14% during 1978-80. The CAT rate used in these calculations is thes one applying to the majority of industrial products. By contrast, a series published by the Banco de la Republica (BR) takes as the basis for the calculation of the effective CAT the average CAT actually paid out to miLnor exports (i.e., total CAT payments as a percentage of the total value of minor exports). This series (Statistical Appendix, Table 4) presumably is heavily influenced by the fact that the CAT for agricultural products (a large proportion of minor exports) has generally been below that for industrial exports. It is thus, arguably, less relevant for the present report than the Morawetz series.15/ 1.18 The effective value of PROEXPO credit as a percentage of the value of exports has shown a marked, fairly steady increase from 1-2% of the value of exports during 1973-76 to 8% of the value of exports in 1980. The reason for this increase is that the rates of interest on PROEXPO credit have been kept at 13-19%, while market interest rates have risen from 20-23% in 19,73-75 to 45% in 1980. Thus, for the industrial products eligible for a CAT of 5%, PROEXPO credit is a more significant incentive than the CAT; for those goods eligible for 9%, the credit incentive is almost as great as that from the CAT; and even for goods eligible for 12%, the credit incentive is worth two-thirds as much as the CAT. Furthermore, as noted earlier, credit has been severely rationed in Colombia during the last couple of years; this adds still more to the real incentive provided by the simple availability of PROEXPO credit. In 1981, several industrialists claimed that they were exporting mainly because they could get PROEXPO credit. The combined effective value of the two export incentives (CAT + PROEXPO credit) is presented in the final row of Table 2. The value of the two incentives rose 141 These are the same as those used by D. Morawetz, Why the Emperor's New Clothes Are Not Made in Colombia, (New York: Oxford University Pres3, 1981), Appendix B. 15/ The BR series shows the nominal CAT falling from 12-15% during 1970-74 to 3-6% during 1975-80; and the effective CAT falling from 14-15% during 1970-74 to 4-6% during 1975-80.  - 11 - TELble 2: NOMINAL ANID EFFECTIVE INCENTIVES TO INDUSTRIAL EXPORTS 1967--80 1967/68 1970/71 1975/76 1979/80 Nominal value of CAT as % of value of exports 15 15 5 12 Discount if CAT is sold in stock market when received, as % of xvalue of CAT. 15.5 7.0 2.8 11.9 Average tax rate on corporate profits (%) 12.5 12.5 20.0 20.0 Percent of value of exports that may be financed by PROEXPO credit - - 80.0 80.0 Rate of interest on PROEXPO credit (: p.a.) - - 18.0 18.0 Mtarket rate of inierest(% p.a.) 18.0 15.0 24.0 41.0 Effective value of CAT, as % of value of exports 14.8 16.2 6.1 13.6 Effective subs:idy implicit in :PROEXPO credit, as % of value of Pexp)orts - - 2.0 6.6 Total effective subsidy to industrial exports (CAT + PROEXPO credit), as percent of value of exports. 14.8 16.2 8.1 20.2 Source: Statistical Appendix Table 5.  - 12 - gradually from 14% to 20% during 1967-74, fell sharply to 8% after the reduction of the nominal CAT in 1975, and rose again to reach 21% in 1980.16/ Real Effective Exchange Rate 1.19 The real effective exchange rate sums up in a single series all of the factors underlying the real exchange rate (rate of devaluation, rates of increase of prices in domestic and partner countries), on the one hand, and the effective value of export incentives (effective CAT, effective PROEXPO credit), on the other. As with the real exchange rate, separate estimates are available for the peso-dollar and peso-weighted average cases, based on different sets of estimates.17/ 1.20 Allowing for the fact that the Banco de la Republica series does not include the effective value of PROEXPO credit, which certainly ought to be factored in, the four series show essentially the same picture. The real effective peso-dollar exchange rate rose consistently during 1967-71, fluctuated around the 1970-71 level during 1972-74, then declined significantly (especially during 1975 with the reduction in the CAT and in 1977 with the slow rate of devaluation), until by mid-1981, it was 20-25% below what it had been a decade earlier. 1.21 Of the three sets of estimates of the real exchange rate of the peso against a weighted average basket of currencies, using the 16/ The Banco de la Republica (BR) document cited above does not include estimates of the effective value of PROEXPO credit. If the BR series for the effective CAT is added to the Morawetz series for the effective value of PROEXPO credit, the effective value of the two incentives are shown to have been relatively constant at 15-16% during 1970-74; they fell sharply to 7% with the reduction of the nominal CAT in 1975, and then, because of the influence of PROEXPO credit, rose gradually during 1967-80, reaching 13% in the later year (see Statistical Appendix, Table 3). 17/ Since ASOBANCARIA and FEDESARROLLO do not themselves present calculations of real effective exchange rates or of the effective value of export incentives, the series for these two institutions have been computed by combining their own real exchange rate series with the Morawetz effective subsidy series. The Morawetz series is the same as the one shown in the original source, except that it has been updated. The Banco de la Republica (BR) series also is presented as it appears in the original BR document.  - 13 - ASOBANCARIA data as the basis is probably the best for the present purposes.18/ That series shows that this real exchange rate rose throughout 1967-1974; but then declined especially in 1975 (with the fall in the nominal CAT) and 1977 (with the fall in the rate of devaluation). By mid 1981 it was about 10% below what it had been in 1971. It is important to nol-e here that, while the tendency toward overvaluation of the Colombian peso can be observed in all series, the difference in the level of cverevaluation is significant. Whereas the peso-dollar relationship has shown a 20% "overvaluation" since 1977, the comprehensive real peso exchange rate, against the baskcet of currencies of Colombia's trading partners, never suffered a revaLuation of more than 10%, but showed rather strong fluctuation between 1977 and 1979.19/ This would lead to the conclusion that it may have beeni less the actual amount of the lag in the exchange rate adjustment than the uncertainty created by the shift from fiscal to monetary irncentives and the renewed price stabilization attempts in 1977, that influenced, the behaviour of exporters. B. Import Controls and Effective Protection 1.22 A stated objective of current industrial sector policy is to reduce levels and dispersion of effective protection rates in order to increase the competitiveness of the sector. Historically, industry has enjoyed signifi- cantly higher rates of effective protection than primary sector activities, but within industrial categories there exists a broad range of effective protection levels.20/ While periods of foreign exchange accumulation could have been propitious for the reduction of import tariffs and the relaxation of controls, the movement toward long-term objectives usually has been hampered by short-run considerations. In partiLcular, the favorable balance-of-payments position enjoyed in recent years has been conducive to some progress in lowering import barriers only with a considerable lag. 18/ As explained earlier, the FEDESARROLLO series contains several inconsistencies, while the Banco de la Republica (BR) series uses a less relevant measure of the effective CAT and omits the effective value of PROEXPO credit. 19/ This is a. reflection of the increasingly stronger fluctuations of exchange rates among the major industrialized countries, with the US dollar depreciating against major other currencies in the late 1970s and appreciating during the early 1980s. 20/ See Thomas Hutcheson, Incentives for Industrialization in (Colombia, Ph.D. dissertation, University of Michigan (1973), and Luis J. Garay, "Analisis de las Estructura Arancelaria de Colombia Vigente en Diciembre de 1974", (Bogota: DNP, 1975).  - 14 - 1.23 Currently, imports are still restricted through a combination of prior deposits, several exchange controls, selective prior licensing, and highly dispersed tariffs. Prior deposits were eliminated in January 1976, reimposed in mid-1976, increased in 1977 and 1979, and then lowered in February 1982. Their implicit cost has been less than 4% of the total value of imports, though they fall more heavily on imports of inputs and consumer goods, which have most of the prior deposit requirements.2 / All requests for imports (registros) are first classified as reimbursable or nonreimbursable and prior license or free (see Statistical Appendix, Table 6 for a breakdown). Nonreimbursable import requests typically originate from direct foreign investors, often petroleum companies importing capital goods; from government and quasi-government importers, especially in cases where foreign loans are used; and from Colombians who are returning from long stays abroad. In the case of reimbursable imports, foreign exchange can be obtained from the Banco de la Republica once the licensing and/or registration requirements at INCOMEX are fulfilled. Overinvoicing of imports to escape exchange controls is regulated by the prior licensing agency, INCOMEX, rather than the Banco de la Republica. 1.24 The main operative restriction on requests for imports relates to the prior licensing system of INCOMEX. All official imports are imported under license. In some cases the licensing requirement also has the dual function of enforcing a government monopoly, e.g., wheat, powdered milk, and petroleum.22/ In other cases the Government simply uses the licensing requirement to regulate imports by direct foreign investors, e.g., petroleum companies. In such cases a simplified procedure--global licenses--can be used to reduce day-to-day bureaucratic problems. In many cases the licensing requirement is still a vestige of the import substituting industrialization philosophy which has not been removed because of political power of various industrial groups, e.g., textiles. There also is pressure by producers to retain the licensing system on goods in which Colombia seems to have a comparative advantage, to prevent dumping or to regulate imports in times of crop failure, e.g., meat and fruits. Finally it should be noted that even nonlicensed imports can be held up when registered by INCOMEX if the price 21/ See S. Clavijo, "Los Depositos Previos de Importacion", Revista del Banco de la Republica, (June 1981). The opportunity cost of these deposits varied between 5% and 10% of the value of imports in the period 1953 to 1978. See J. Garcia, The Effects of Exchange Rates and Commercial Policy on Agricultural Incentives in Colombia 1953-78, (Washington: International Food Policy Research Institute, 1981). 22/ See Statistical Appendix Table 7 for a list of major imports under prior license in 1976-1979.  - 15 - appears unreasonably low,23/ and imports with foreign financing have been held up as part of the Government's drive to cut down on capital inflows. 1.25 Since 1976, over 90% of requests for nonreimbursable imports were under prior license. The percentage of requests for reimbursable imports which required prior license is much lower, only about 50% in 1980. Until 1975 there had been a steady decline in the percentage of registros which required such licensing; thereafter the percentage has remained roughly constant. The percentage of tariff code classifications which are subject to prior licensing has declined steadily, from over 95% in 1971, to 66% in 1975, about 33% in 1979 and 31% in 1980. (See Statistical Appendix, Table 8). Of course such declines are not a priori evidence of lower protection, since they may simply reflect switches in the status of noncompetitive imports. The fact remains that relatively few tariff code classifications were subject to prior licensing in 1980-81. It should also be noted that the request procedure has been streamlined; a decision on imports required only one week in 1981, except for certain special cases cited below. 1.26 Of the 1502 tariff code classifications under prior licensing requirements in 1980, no import requests were made in 606 classifications in 1977, 623 in 1978, and 603 in 1979, an average of about 40%.24/ The lack of requests has been interpreted as an indication of excessive protection which could be liberalized. In fact in some cases :in which Colombia seems to have a comparative advantage and thus is unlikely to receive requests for imports, the prior license requirement represents a desire to control imports if crops should fail or to prevent dumping, e.g., meats (10 classifications) and fruits (12 classifications). In other cases t:he lack of requests probably represents the importers' knowledge that the request will be denied because of desires to protect the industry, e.g., textiles (113 classifications) and wood products (21 classifications). 1.27 Information from INCOMEX indicates l:hat about 89% of requests for reimbursable imports were accepted during 1980 (see Statistical Appendix, Table 9). However, this calculation excludes various special categories of imports discussed below which amounted to about 43% of Registros. Seven percent of the rejections were for various reasons having to do with faulty preparation of the request. Included in these rejections were some requests which quoted prices that in the opinion of the INCOMEX licensing board were overly high or overly low (overinvoicing and underinvoicing). Another 5% of the requests were rejected for substantive reasons; in about 4C% of these cases one of the cited reasons for rejection was the existence of local supplies. Excessive requests relative to historic levels were cited as another reason to deny or delay a request. These rejection rates seem 23/ G. Giraldo, "La Estimacion de la ProteccLon en Colombia, "Revista de Planeacion y Desarrollo" (May-August 1979). 24/ W. Marin, et al., Analisis General de las Importaciones del Regimen de Licencia Previa (Bogota, INCOMEX, 1980).  - 16 - somewhat higher than those cited by Diaz Alejandro for 1971, although the comparability of the coverage is not clear. It also seems probable that the current attitude of INCOMEX has elicited more requests for import licenses than in the mid-seventies. 1.28 A policy authorizing general licenses (licencias globales) for importers of capital goods was established at the end of the 1950s to favor the importation of systems of machinery. Initially, the main advantage of the global license was that the whole project had authorized access to foreign exchange. Decreto Ley 444 of 1967 and subsequent resolutions based on it (#15 of 1967, #20 of 1971, #27 of 1978, and #066 of 1980) required that applications for global licenses contain a study of economic feasibility including documentation supporting the project's contribution to objectives of the national plans and other criteria. As foreign exchange scarcity diminished so did the attractiveness of guaranteed access to foreign exchange. In addition, other concessions were granted, including: (a) longer payment periods permitting importers to use lower cost foreign credit for longer periods; (b) extension from prior deposits after 1974; (c) favorable consideration for the 5% uniform tariff on capital goods imports (since May 1976). However, the granting of this low uniform tariff depends to some degree on lack of competitive domestic production. To these advantages must be added the simplicity of a single import licensing negotiation for the whole project. Given the favorable treatment they receive, requests for imports under global licenses have grown substantially. (See Table 3, the jump between 1979 and 1980 probably reflects a depressed level of applications in 1979 owing in part to the reimposition of prior deposits). In 1980, there was a large increase in demands for imported machinery by civil construction firms involved in the PIN program. Also, demands for imports were stimulated by the lifting of prior deposit requirements and the extension of the permissible period of payment. 1.29 Import requests under Plan Vallejo have also grown steadily, reaching a 1975 peak of 4.1% of total imports. Realized imports under Plan Vallejo peaked in 1976 and amounted to 4.1% of total realized imports (see Table 3 & Statistical Appendix, Table 11). From this peak, import requests under Plan Vallejo declined to only 2.7% of total import requests in 1979, though they rose to 3.6% in 1980. Realized imports under Plan Vallejo also fell, to 2.6% of realized imports in 1980. This decline probably reflects the fall in profitability and thus the quantity of nontraditional exports using imported inputs, rather than any change in the terms of Plan Vallejo. However, it is worth noting that requests for imports under Plan Vallejo took longer to process than those for regular imports under previous license, up to a month as compared to a week. Thus, some liberalization was recently achieved by speeding up the Plan Vallejo process. 1.30 Until 1980 imports of automobiles with a factory price of less than US$2000 were prohibited to avoid competition with the local Renault factory's production. Worldwide inflation of automobile prices substantially eroded the protection provided by that prohibition, and in late 1980, after a sharp rise in auto imports, the floor price was raised to US$5000. A substantial number of requests for auto imports were paralyzed while the decision on the  - 17 - Table 3: IMPORT SPECIFIC AND GLOBAL LICENSES AND IMPORTS UNDEPR PLAN VALLEJO 1970-1980 1970 1975 1978 1979 1980 Prior Licenses l.Total Number 3496 a/ 3128 2218 1592 1502 2.% of Total T.ariff 80 66 46 33 11 Positions II. Global Licenses l.No. of Projects 137 35 124 173 218 2.Value of Imports authorized(:in mill US$) 8.9 163.0 309.8 220.1 382.3 3.% of total imports 9.1 10.8 9.2 4.8 7.0 IL-. Plan Vallejo 1,Registered (in mill. US$) 2L..8 61.7 87.9 124.3 197.1 2-in % of total registered imports 2.7 4.1 2.6 2.7 3.6 3.Actual Imports (in mill. US$) 9.7 47.1 54.7 65.6 92.6 4 in % of total actual imports '.7 3.6 2.6 2.6 2.6 Source- Statistical Appendix Tables 8, 10, 11. a/ 8/71.  - 18 - new minimum price level was reached. Many of these requests were then granted in 1981 once the issue was settled.25/ 1.31 According to the most recent major study on external tariffs.26/ Colombian tariff protection averaged (unweighted) about 26% following the revision of protection in mid-1979. This compares with an (unweighted) average rate of about 28% in the first semester of 1979, 36% in 1975 and 70% in 1970.27/ The aim of the 1979 revision was to reduce protection in order to slow inflation and obtain greater efficiency and competitiveness in the Colombian economy in the long run. This average level is close to the Andean Pact's common minimum tariff, although substantial divergences exist in specific product areas. Thus, the Pact has not presented a barrier to cuts in tariff protection in most areas. As Table 4 shows, the dispersion of tariffs has remained high, with a standard deviation of 18% and skewed to the right, with some items having tariffs of as much as 150% and others having zero tariff protection. In addition to tariffs, imports are subject to a 5% tax which is used to finance PROEXPO loans to exporters (raised from 1.5% in 1975) and a 1.5% tax which is used to finance the Fondo Nacional de Cafe. The 1% consular fee on the requested value of imports was eliminated in 1980. 1.32 No current study is available on effective protection, which uses the ratio of local to world prices of imports.28/ The existing, more recent studies examine effective protection on value added by using an input-output matrix to compare the nominal tariff on the final product with that of the corresponding imported inputs. Thus, these studies neglect non-tariff barriers, the implicit cost of prior deposits, the taxes mentioned in the previous paragraph, and tariff redundancy. They also do not include the reduction in protection associated with Plan Vallejo and global licensing. Finally, the estimates are not adjusted for the changes in the exchange rate which would be necessary to maintain the balance of payments, given a set of macroeconomic policies. Thus the usefulness of these estimates lies in their contribution to understanding the potential, rather than the actual structure and level of protection. 25/ This administrative procedure would have produced an unfounded increase in the number of denied import requests in 1980 and therefore requests for auto imports are omitted from Table 9 of the Statistical Appendix. 26/ Giraldo, op. cit. 27/ Giraldo, op. cit. 28/ The only existing study based on price comparison was undertaken by T. Hutcheson op. cit., comparing world and local prices for 385 products.  Table 4: NOMINAL AND EFFECTIVE PROTECTION AND PERCENTAGE OF ITEMS UNDER PRIOR LICENSE Broad Produict Classes 1975 and 1979 1975 1979 I 1979 It Nominal Items Nominal Effective Items Nominal Effective Items** Protection Previous Protection Protection Previous Protection Protection Previous License License License Av.a/ s.D.b/ x Av. S.D. Av. S.D. Z AV. S.D, Av. S.D. X Primary Products 19 21 - 16 8 26 22 53 15 7 24 20 28 Consumer Goods 47 24 - 43 22 87 50 48 39 22 81 49 32 + Intermediate Products 24 16 - 22 11 32 21 43 20 10 29 18 28 Machinery 28 15*C/ 66**d/ 30 18 42 33 47 27 16 39 30 42 Transpnort Eqiipnment 40 44 89 37 40 82 102 73 34 34 75 91 70 All Imports 36 - 66 28 19 48 43 46 26 18 44 40 33 a. Average b. Standard deviation. c. Estimate d. Esti-,ate ased on. rou-hl1, comparable sections of ta-riff node. SOURCE: Statistical Appendix Table 12.  - 20 - 1.33 According to the most recent investigation by members of the DNP, effective protection averaged (unweighted) about 44% in the second semester of 1979. The standard deviation was 40% and the distribution of effective protection was skewed, with maximum rates of 400% and minimum rates of -46%. No major tariff changes ocurred in 1980 and 1981. Table 4 presents estimated average nominal and effective protection and the number of items on the prior license list in 1975 and 1979, both for imports competitive with production in some major industrial groups. It indicates that by far the highest average nominal rates of protection apply to consumer goods and to transport equipment; the latter reflecting automobile protection, with the other trans- port equipment sectors receiving not more than average rates of protection. Since all these industries use relatively large components of intermediate inputs, often with low protection, their calculated average effective rate of protection is more than double the nominal rate. This compares with calclu- lated average effective rates in the other industrial sectors that are only 30% to 50% higher than nominal rates. 1.34 According to these estimates the machinery sector has the next highest average rate of nominal and effective protection, about equal to the overall average. However, the estimate neglects the Global License facility which allows the import of machinery with only a 5% nominal tariff, if the machinery is not produced locally. Depending on the interpretation of compe- titive local production, this facility substantially erodes the protection given to the Colombian machinery industry, and the rapid growth in global licenses seems to indicate that the interpretation has been fairly libera:L over the past few years. Further, Government imports of capital goods are not subject to duties. Finally, for capital goods imports' financing terms are an important determinant of purchases. High local real interest rates thus are an important form of negative protection, particularly in view of the subsidized foreign interest rates on capital goods sales, but this also is not reflected in the effective tariff rates. In particular, in 1980 Government-sponsored imports of capital goods under the PIN have grown rapidly as domestic producers were limited to 15% of contracts for a variety of reasons including the easy terms of foreign financing. Thus, machinery production probably faces much more international competition than Table 4 would seem to indicate. Specific examination of some products in connection with the investigation of the metal mechanic industry indicated that several domestic capital goods were exposed to negative effective protection of over 20%. 1.35 Primary products have the lowest rates of nominal and effective protection. However, this average is somewhat illusionary, since several products have high individual rates, the maximum being 70%. Many competitive primary imports are also subject to prior license. A comparison of world and local price of some specific products showed that corn, milk, wheat and vegetable oil received nominal protection of between 25% and 50%. 29/ The rates of protection on these commodities seem to have risen after 1974. TIhe 29/ J. Garcia, op. cit., p. 27.  - 21 - aggregate patterns of tariff protection are also reflected in the nominal and effective rates at the three digit industry level. Textiles, garments, and automobiles have the highest protection, followed by leather products. Intermediate products, such as chemicals, paper and metals, have the lowest rates. 1.36 To summarize the discussion on protection, the prior license was the most important non-tariff protection befor-e the reform of 1979. Although average nominal tariff rates have declined by about 5 percentage points between 1975 and 1979, and the consular fee of 1% was eliminated in 1980, much of these declines have been offset by the rise (3-1/2 percentage points) in the PROEXEPO fee and the rise in the implicit cost of prior deposits. The principal reduction in protection in the reform of 1979 thus seems to have been the elimination of licensing requirements from many products. While several key products were still subject to prior license, the reform was certainly in the right direction. Thus the sharp rise in imports in 1980 should probably be attributed to declines in effective protection in a few lines, especilally machinery, petroleum and autos, and to the easing of prior licensing requirements, rather than to a broad based reduction in protection. 1.37 The effect of the exchange rate on imports between 19474 and 1980 is not clear. The depreciation was some 20% less than the differience between the inflationa rates in Colombia and the U.S., as measured by the consumer price indices. This would seem to have encouraged imports. T'he comparison is less favorable to imports, if the Colombian import price index in dollars (which includes petroleum) is used instead of' the U.S. consumer price index. Between 1974 and 1980 the estimated inflation in dollar import prices was 97.8%, and the devaluation of peso/dollar exchange rate amounted to 81.4%, implying a rise in local import prices of 258.8%. In contrast, both the overall wholesale price index and the consumer price index for workers rose by 264%; the wholesale price index for local industry, 254%. Thus by this measure the real exchange rate fell very little over the period, although the rise in petroleum prices distorts the comparison. However, it must also be borne in mind that much of the rise in the nominal value of imports is due to increaed petroleum imports, as discussed in Section 2.45 and 2.46.30/ Finally the wholesale price index for all goods produced and consumed in the country rose 292% between 1974 and 1980, versus 266% for the local wholesale price index of importables; the wholesale price index of manuf'actures produced and consumed in the country rose 213%, versus 258% for local prices of imported manufactures. Comparisons of these figures are limited by the differences in the goods included in the different price indices. However, they seem to indicate that: (a) tariff and nontariff barriers, rather than 30/ Between 1974 and 1979 the wholesale price index rose about 8.5% faster than the product of the exchange rate and the estimated import price index in dollars. The black market exchange rate rose a'bout 10% faster than the legal rate since 1974, owing to the effect of t'he certificado de cambio.  - 22 - exchange rate movements, were the most important means of altering protection of import competing industries in the latter half of the seventies; and (b) local prices of nonmanufactured goods rose much faster than world prices, corrected for the exchange rate depreciation, while local prices of manufactures actually rose less rapidly than those of imported manufactures.  - 23 - ANNEX 1 Page 1 of 2 FORMULAS FOR EFFECTIVE EXPORT INCENTIVES The f'ormulas used to calculate the efEective rates of subsidy that are implicit iTn the CAT and in subsidized PROEX'PO credit (and in the Vallejo Plan drawback scheme) are derived below. In each case, the effective subsidy is expressed as a percent of the value of exports. To simplify the analysis, it is assumed that the firm's costs of production are zero. The results are unchanged if this assumption is relaxed. 1. CAT Question: What is the pretax taxable income (X*) that produces the same net income after taxes as that received by a firm exporting goods valued at 100 pesos which receives an export subs,id,y (CAT) and sells it at a discount (d) in the market? Define: X - income from exports (assumed to be 101) pesos) CAT export subsidy (assumed to be 10%t) d - market discount on sale of CAT (assumed to be 20't) t - rate of company tax (assumed to be 40'%) Before the 1974 tax reform Pesos Income from exports X 100.00 plus Net income from CAT + CAT(1-d) 8.00 = Total pretax income X+CAT(1-d) 108.00 less Exemption of CAT from taixable income - CAT -10.00 = TaLxable income = X-dCAT 98.00 Taxes payable t(X-dCAT) 39.20 Net income after taxes (=108.00 - 39.20) X+CAT(1-d)-t(X-dCAT) 68.80 X 1I-t)+CAT(l-d+td) Now derive the desired pretax taxable income (X*) and the implicit effective CAT (CAT*): X*(L-t) = X(l-t) + CAT(l-d+td) 68.80 X* = X+ CAT(1-d+td) 114.67 (1-t) CAT* = CAT(l-d+td) 14.67 (1-t) After the 1974 tax reform Pesos Income from exports X 100.00 plus Net income from CAT + CA" (1-d) 8.00 = Total pretax income = X -e CAT(1-d) 108.00 less Taxes payable -t[R--CAT(I-d)] -43.20 plus Tax discount of t times nominal CAT + tCAT 4.00 = Net income after taxes = X(L-t) +CAT(l-d+td) 68.80  - 24 - ANNEX 1 Page 2 of 2 Thus, the real subsidy implicit in the CAT is the same as before the 1974 tax reform despite the change after 1974 in the way that the CAT is treated for tax purposes. Timing The above derivations ignore the fact that the CAT is not valid for paying taxes immediately after it is received. The period before it becomes valid for such use varied from 3 to 12 months during 1967-81. The derivations also ignored the fact that the CAT is sometimes sold in the market only after a delay of several rmonths, and not, as is assumed above, immediately after it is received. Defining n as the average number of years delay before the CAT is sold (0.08 n 1.00), and defining r as the market rate of interest, the above formula would need to be altered as follows if such delays are to be taken into account: CAT1 -d + td 7 CAT* = L(1+r)n (1+r)ni 1-t (1+r)n The net effect of taking these delays into account would be to diminish the value of CAT* to some extent. The delays are not taken into account in the present study because the necessary data are not available. 2. PROEXPO CREDIT Question: What is the effective value (P*) of the subsidy that is implicit in the low interest credit that is provided by PROEXPO to exporters for six months? Define: C/X - the percentage of the value of exports for which PROEXPO credit can be received (80% from 1973 onwards, or 40% for users of the Vallejo Plan. For the latter, it is assumed that imported inputs account for 50% of the value of output). r - the market rate of interest (varied between 14% and 45% during 1967-81). rp - the rate of interest on PROEXPO credit (18% from 1973 to mid-1977, 13%, then 17%, then 19%, thereafter). The effective subsidy implicit in PROEXPO credit (P*) is equal to the present value of what is received less the present value of what has to be repaid six months later. That is: (1 + rp)4L1 P* = C L - X (1 + r)2  - 25 - II. TRENDS AND DETERMINANTS OF MANUFACTURED EXPORTS AND :[MPORTS C. Exports of Manufactured Goods in Recent Years 2.01 Colombia's exports of manufactured goods in current US dollars increased almost fourfold during 1970-75, but less than doubled (luring 1975-80, despite the fact that world inflation was much higher duiring the second period (Table 5). World trade in manufactures grew faster during the former period, as witnessed by exports of manufactures from other developing countries such as Korea, Taiwan, Hong Kong. Hcwever, these countries still realized a reaL annual growth rates of 15% during 1975-80 after they had grown by 35% between 1970-75. In contrast to that, Colombia's export rose by real annual rates of 17% between 1970 and 1975 and decelerated to only 4% during the latter part of the decade. What were the major reasons for this decline in export growth? How did the structure of Colombian manufactured exports change during the last five years? Which countries became Colombia's most important customers of manufactured goods? Before discussing these and similar issues, a brief note on the problems of Colombian export statistics is in order. Export Statistics 2.02 The analysis in this report is based mainly on officiaL export statistics. There are several problems involved with these data (see Annex 2 for more details). First, the three sets of export statistics that are available--DANE figures on exports that have allegedly passed the customs barrier (manifiestos), INCOMEX figures on export registrations (registros), and Banco de la Republica's figures on export dLollars converted to pesos (reintegros)--do not always agree. Second, illegal exports, mostly to neighboring countries, do not enter the official statistics. These exports respond to differences in prices in the domesti'c market and neighboring countries, and have been substantial over the ]ast decade or so for a number of industrial goods---sugar, textiles, clothing, cement and tires--as well as for coffee and cattle. Third, overinvoicing and fictitious exporting inflate the official figures. A FEDESARROLLO study has estimated overinvoicing to be at about 6% of official exports. Fictititous exporting seems to have been particularly common in 1974 and 1979. In both years, holders of black market dollars found it feasible and profitable to convert these into pesos at the official exchange rate plus the export subsidy. After each year, the laws and procedures were tightened up to discourage such maneuvers--but some fictitious exporting continued anyway. 2.03 In the clothing industry, Morawetz'/ found that fictitious exports made up about half of all exports that were declared as having passed customs (DANE data) in, 1974, and that five outstandingLy large non-existent firms seemed to account for most of the total (one of these alone "exported" $12 million of clothing in 1974). An analysis of the INCOMEX data con registered clothing exports for 1979 (in which year the registered dollar value of such exports almost, doubled) indicates that this time there were no suspicious 1/ D. Morawetz, Why the Emperor's New Clothes Are Not Made in Colombia (New York: Oxford University Press, 1981).  - 26 - Table 5: EXPORTS OF NANNUFACTURED GOODS, CURRENT AND CONSTANT PRICES 1970 - 1980 (US$ millions) Constant Dollars of 1970 Current Value Index Percentage Dollars Change (M) 1970 117.9 117.9 100 n.a. 1971 152.3 147.0 125 25 1972 188.2 175.6 149 19 1973 236.9 207.4 176 18 1974 a/ 400.8 286.7 243 38 a! 1975 404.7 259.6 220 -9 1976 395.5 238.5 202 -a 1977 416.6 234.8 199 -1 1978 473.3 248.6 211 6 1979 a/ 744.1 346.9 294 39 ! 1930 b/ 772.0 317.0 269 -9 a/ Includes Fictitious exports. b/ Mission estimates. SOURCE: Francisco Piedrahita, "Desarrollo Industrial en la Decada de los Setenta", Revista ANDI, No. 51 (1980).Table 17, based on data from DANE and Banco de la Republica. LCPI2 October 1981  - 27 - large entries but rather literally hundreds of smaller entries, most of them US$200,000 or 'Less, each in the name of an individual rather than an enter- prise. It is possible that there sprang up overnight hundreds of new clothing exporters; but it could also be that some of these exporting enter- prises were fictitious. In the metal manufacturing sector, exports of several millionls of dollars were registered in 1979 in items like metal statues and metal picture frames. The fact that these goods were exported in these voLumes neither before nor since has caused some to believe that these, too, may have been fictitious exports.2/ 2.04 Fictitious exports during 1979 seem to have been "sent" especially to Venezuela. Colombia's industrial exports to Venezuela doubled in nominal dollar terms from 1978 to 1979, and Colombia's exports of all goods to Venezuela jumped from 9.5% to 12.8% of its exports to all destinations--this despite the fact that Venezuela's per capital GDP fell by 2.7% in 1979. In 1980, as fictitious exports were reduced, industrial exports to ilenezuela fell by almost a quarter in nominal dollar terms, and total exports to Venezuela fell back from 12.8% to 7.8% of total exports to all destinations. As a first lower bound estimate at the extent of fictitious exporting in 1979, it would be useful to compare Colombian data on its exports to Venezuela in 1979 with Venezuela's data on imports from Colombia (i.e. data on the same transactions) for the same year. Cn the basis of those compari- sons, but without publishing its methodology or a detailed sector or product level breakdown, INCOMEX has estimated that fictititous exports were 8.9% of exports other than coffee and petroleum in 1979.3/ 2.05 A fourth statistical problem, relating mostly to clothing, is that exports to the United States under the U.S. 807 (offshore assembLy scheme) are registered in Colombia at the value added cnly. Last but nol: least, different institutions and different authors define "industrial exports" differently, piarticularly with respect to products of the petroleum refining sector, preciolas and semi-precious stones, and some processed foods like sugar and bleached rice. Subsectors are also not always consistiently defined: textiles sometimes includes clothing, metal manufactures sometimes includes transport equipment, and so on. 2.06 Previous estimates of total (official. plus unregistered) industrial exports from 1970-77 show total industrial exports exceeding official (DANE) estimates by the following percentages in each of the years 1970--77: 12, 20, 17, 12, 31, 40, 38. The jump after 1974 is explained by the fall in the CAT, which makes illegal exporting relatively more profitable. Since the value of the CAT and PROEXPO export incentives has been raised considerably since 1977, the proportion of total exports that is illegal may have fallen and the proportion of official exports that is overinvoiced or fictitious may have risen since then. 2/ Note, though, that the apparently legitim1ate exports of Colombia's dairy products were believed to be "fictitious" when they first occurred. 3/ (Comercio Exterior, Sept.-Oct. 1980, p.3 n.1).  - 28 - Aggregate Trends of Manufactured Exports 2.07 Exports of manufactured goods in constant dollars increased at an average of 17% p.a. between 1970 and 1975, but decelerated to 4% p.a. between 1975 and 1980, experiencing declines in three of the last five years.4/ During 1971-73, manufactured exports grew in real terms by 25%, 19% and 18% respectively; by contrast, during 1976-78 they grew by -8%, -1% and 6% respectively. The rapid rise and subsequent relative stagnation of Colombia's industrial exports can be seen clearly in the figures on indus- trial exports as a percentage of industrial production and industrial exports as a percentage of GDP (Statistical Appendix, Table 13). Industrial exports rose from around 3% of industrial output in 1970/71 to about 10% in 1974/75, but then gradually declined to about 7% in 1980. As a percentage of GDP, industrial exports rose from about 1.4% during 1970/71 to about 5% in 1974/75, but then declined gradually to about 3.6% in 1980. Similarly, Colombia's share of total LDC manufactured exports rose from 0.69% in 1971 to 0.88% in 1973, but then fell back slightly to 0.85% in 1975-77 and 0.75% in 1978-79. Exports by Sectors and Products 2.08 The distribution of manufactured exports by sectors was relatively stable during 1974-79 (Stastical Appendix, Tables 14 & 15). Food products accounted for 18-21% of the total during most of the period, refined petro- leum products made up 13-17%, and textiles contributed 12-14%. There were some changes, however. The share of metal manufactures rose abruptly from 9% in 1974-76 to 13-14% in 1977-79; the share of chemical, rubber and plastic products fell from 12-13% in 1974-76 to 8-9% in 1977-79; while the share of clothing, in which fictitious and smuggled exports are a particular problem, fluctuated between 4% and 11%. The share of all other manufactured products rose from 11-16% in 1974-76 (products of leather, wood, and non-metallic minerals predominated), to 19-22% in 1977-79 (products of the paper and printing industries were important in the increase, more than offsetting a decline in wood products and furniture). 2.09 The sectors exporting the highest percentages of their output during 1977-80 (Stastical Appendix, Table 16) were non-metallic minerals (16-37%),leather products (18-30%), and oil products (19-28%). These were followed by a group of sectors in which roughly 10% of output was exported on average during 1977-80, including wood products-and-furniture, textiles- clothing-footwear, machinery (the percentage for mechanical was higher, for electrical lower), and paper-and-printing products. Finally, in a further group of sectors, the proportion of total output exported was consistently 5% or less during 1977-80 these were: basic-metals-metal-products (except 1979), transport equipment, chemicals, and, lowest of all, food products (2%). Almost without exception, all sectors showed significant increases in the 4/ Note that since the data for 1974 and 1979 include some fictitious exports, it is difficult to make precise statements about those years. It is also rather misleading to give growth rates of manufactured exports using 1974 or 1979 as the beginning or end year.  - 29 - percentage of output exported between 1967-70 and 1973-75. By ccntrast, only two sectors (oil. products and non-metallic minerals) showed significant increases in the percentage of output exported between 1973-75 arLd 1977-80. This indicates that both, the boom in manufactured exports during 1967-75 was a widespread phLenomenon that touched all sectors and that the subsequent relative stagnation during 1975-80 was almost equally widespread. 2.10 Durin.g the first four months of 1981, there was a 21% increase in manufactured exports in nominal terms compared with the corresponding period of 1980 (Statistical Appendix, Table 17). However, more than a third of the increase was in. tariff cotton and cotton textiles; information obtained from the major textile firms indicates that there was a typical bunching of exports during the first four months of 1981, not to be continued. during the rest of the yea.r. The only other tariff positions in which there was an increase of as much as US$1 million compared with January-April 1980 were plastic product's (US$1.9 million increase), other chemical products (US$1.3 million), other metal products (US$1.3 million) and printing (US$1.2 million). Partial information for January-May 1981 indicates that in the clothing sector, exports fell by 5% in current dollar terms compared to January-May 19E80.5/ 2.11 Between 1970 and 1980, almost all sectors increased their exports in current dollar terms tenfold or better (Statistical Appendix, Table 18). The exceptions are non-metallic minerals (which increased their exports 7 times), leather products (7 times) and wood products and furniture (3 times). At the high end of the spectrum were oil products (560 times), rubber products (20 times) and the metal manufacturing sectors (13-19 times). Food products, oil products, textiles, clothing and metal manufactures had each reached an annual export value of US$100 million or more by 1980, while chemical-plastic-rubber products, non-metallic mineral products, and paper and printing products, had each reached US$50 million or more. By way of comparison, total manufactured exports from all sectors did not reach US$100 million until 1971. According to INCOMEX data, manufactured exports amounted to over a billion dollars in 1980 (US$1,177 mil]ion), and despite the fact that coffee prices remained high in 1980, accounted for 31% of total exports in that year. This may be compared with 1967-70, a year of much lower coffee prices, when manufactured exports made up only 12-16% of total exports, with just three sectors (food products, chemicals and textiles) providing between half and two-thirds of the manufactured export total. 2.12 The manufactured goods (or to be more precise, tariff positions), which registered exports of US$10 million or more in 1980 are shown in Table 6 ("others" ref.ers to all tariff positions with less than US$10 million). The entries are placed in order of value of 1980 exports (See Statistical Appendix, Table 19). It is clear that, although there has been a great increase in the number of goods exported, 15 broadly defined iteris still account for about half of total manufactured exports. These 15 items included processed foods (sugar, beef, sweets, cheeses, shrimps and rice), exterior clothing for both sexes, textile fabrics and yarns (most:ly of cotton), cement:, books and periodicals, leather goods, bedclothe-s and epsilon-caprolactama. 5/ El Tiempo, 8-20-1981, p. 2A.  - 30 - 2.13 The volume of exports of five of these 15 items fell by the follow- ing amounts during 1976-80: textile yarns of cotton (25%), beef (29%), frozen shrimps (17%), bleached rice (81%), and leather travel goods (32%). Exports of a further two items grew in volume terms by less than 5% a year durirLg 1976-80: textile fabrics and leather shoe components. Cement exports grew at 8% a year or better and so did the following products: raw sugar (357%), exterior clothing (15%), books, periodicals, etc. (21%), "melazas" (30%), and cheeses (16%). 2.14 Two categories of comparative advantage have emerged in Colombian industry. One group of export-oriented industries--garments, wood products, furniture, and leather products--is based on domestic raw materials and/or cheap labor, and has a comparatively low degree of technological sophistica- tion. A second group of export products includes printing products, certain ranges of metal products and machinery and the like, which employ relatively more sophisticated but well-established technologies. (Printing also has subsectors which are labor intensive and employ low technology: the produc- tion of "pop-up" children's books for example, where a Colombian manufact:urer has a strong world market position, employs basically unskilled labor and virtualy no machinery.) Colombia's comparative advantage in these sectors is based on such factors as relatively well-trained workers, an established force of mechanics and engineers, and a number of well-managed enterprises with international market ties. These advantages are particularly strong in the printing industry, and recent export growth has been high. Another important reason for Colombian competitiveness in the metal products and printing industries is the possibility of economical small-scale production. Large manufacturers in industrial countries are not able to produce econo- mically short production runs of such items as books, foundry products, and valves. Destination of Exports 2.15 During the late 1970s Venezuela emerged as the single most important market for Colombia's manufactured products. This country, with a comparatively limited but high income market, now takes about 40-50% of Colombia's total industrial exports (legal and illegal). Total official industrial exports to Venezuela increased from US$4 million in 1970 to US$281 million in 1980 (Statistical Appendix, Table 20). Four sectors--textiles and clothing, metal manufactures, transport equipment and cement--have consis- tently made up about half of these exports, with the other half spread quite widely across the industrial spectrum.6/ The main reasons for the rapid growth of manufactured exports to Venezuela seem to be the rapid rise of Venezuelan incomes since the oil price increases of 1973-74 and the emergence of the Margarita Islands, as an important Venezuelan free port since 1975. Andean Group tariff preferences, by contrast, seem to have played a smalL role. Clothing and textiles have consistently made up a quarter to a third of Colombia's industrial exports to Venezuela, with clothing being over two-thirds of this total. Yet clothing is on Venezuela's "list of 6/ Official exports of meat amounted to US$20-40 million p.a. during 1972-79; exports of cheese were US$20 million in 1980.  - 31 - _Table 6: MANUFACTURED EXPORTS 1980 LEVEL AND GROWTH IN VOLUME SINCE ]Q7h Value of Volume of exports exports 1980 1980 ($ millions)