Small-scale enterprises in Colombia : a case study

SMALL-SCALE ENTEPRISES IN COLOMIBIA: A CASE STUDY by Albert Berry and Armando Pinell-Siles Series: Studies in Employment and Rural Development No. 56 Division: Employment and Rural Development Department: Development Economics Development Policy Staff International Bank for Reconstruction and Development This paper was prepared as a background piece for the Bank's study on Small-Scale Enterprise Development (RPO 671-59). It is being issued for comment and discussion purposes only and is not to be quoted without the author's and the Division's permission. The views it reflects are those of the authoi and not necessarily those of the Bank. The authors are indebted to Mr. Carlos Singer and Ms,. Judy Lu for research assistance and computations respectively in the preparation of this work. Washington, D.C., July 1979 SMALL-SCALE ENTERPRISES (SSEs) in COLOMBIA: A CASE STUDY Table of Contents Page No. I. INTRODUCTION . * * * * * * * . * . . N . . . . . . . . . . 1 II. OVERALL ENVIRONMENT FOR SSEs AND ITS EVOLUTION OVER TIME . . 4 Foreign Trade Regimes . . . . . . . . . . . . . . . . . . . . 10 Locational and Regional Factors . . . . . . . . . . . . . . . 11 III. MANUFACTURING, COMMERCE AND SERVICES . . . . . . . . . . . . 1. MANUFACTURING . . . . . . . . . . . . . . . . . . . . . . 14 A. Factory Sector . . . . . . . . . . . . . . . . . . . 14 B. Household Industries . . . . . . . . . . . . . . . . 16 C. Very Small Firms (Shops) . . . . . . . . . . . . . . 25 2. COMMERCE . . . . . . . . . . . . . . . . . . . . . . . . 38 3. SERVICES . . . . . . . . . . V C S . . . . . . . . . . . 40 Appendix: Data Sources for Manufacturing . . . . . . . . . . 43 IV. DYNAMIC CHANGES IN THE FACTORY MANUFACTURING SECTOR . . . . . 46 1. Aggregate and Sectoral Trends . . . . . . . . . . . . . . 46 A. Structural Changes in Manufacturing: 1956-1975 . . . 48 B. Distribution of the Net Increase in Employment and Establishments . . . . . . . . . . . . . . . . 51 2. Age-Distribution of Factory Establishments . . . . . . . 54 3. Mbbility Patterns in Factory Establishments . . . . . . . 56 A. Aggregate Patterns . . . . . . . . . . . . . . . . . 56 B. Mobility Patterns of Different Industries . . . . . . 66 4. Survival and Death Rates in Factory Establishments . . . 69 V. RELATIVE EFFICIENCY AND CAPITAL INTENSITY BY PLANT SIZE . . . 75 1. Introduction . . . . . . . . . . . . . . . . . . . . . . 75 2. Distinctive Features of SSEs in the Factory Sector: 1970 93 Capital Intensity . . . . . . . . . . . . . . . . . . . . 94 Capital Productivity . . . . . . . . . . . . . . . . . . 101 3. Scale Economies and Diseconomies . . . . . . . . . . . . 103 Page No. VI. THE DETERMINANTS OF THE SIZE DISTRIBUTION OF ENTERPRISES . . . 110 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . 110 2. Economic Policies . . . . . . . . . . . . . . . . . . . . 113 A. Credit and Technical Assistance . . . . . . . . . . . 113 B. Other Policies . . . . . . . . . . . . . . . . 117 APPENDIX B: MEASURES OF CAPITAL USED IN THE FACTOR SECTOR . . . . 184 List of Tables Table No. Title Page No. II-1 Colombia: Changes in Employment Structure, 1951, 1964, 1973 . . . . . . . . . . . . . . . . . . . . . 6 11-2 Colombia: Occupational Distribution of Heads of House- hold in the Poorest 20% of Households (). . . . . . 9 III-1 Employment and Labor Force in Manufacturing, by Size of Establishment and by Rural/Urban Census Years 18 111-2 Colombia: Employment in Manufacturing Sector 19 111-3 Employment in Household Industries, 1973 . . . . . . . 20 111-4 Employment in Household Industries: "Rest", 1973 . . . 21 111-5 Employment in "Cabecera" Household Industries, 1973 22 111-6 Manufacturing Employment by Two Digit Industries and by Factory/Cottage-Shop . . . . . . . . . . . . . . 24 III-7a Distribution of Establishments, Employment, Wages, Production and Value Added by Manufacturing Sector for Very Small Firm . . . . . . . . . . . . . . . . 26 III-7b Percentage Distribution of Establishments, Employment, Wages, Production and Value Added by Manufacturing Sector for Very Small Firms . . . . . . . . . . . . 27 III-8a Distribution of Establishments, Employment, Wages, Raw Materials and Production by Manufacturing Sector for Very Small Firms, 1970 . . . . . . . . . . . . . 28 III-8b Percentage Distribution of Establishments, Employment, Wages, Raw Materials and Production by Manufacturing Sector for Very Small Firms, 1970 . . . . . . . . 29 Table of Contents Table No. Titles Page No. 111-9 Characteristics of Very Small Establishments in Non-Metropolitan Areas, 1970 . . . . . . . . . . . 35 IV-1 Colombia: Sectoral Increases in Establishments, Employment and Average Size of Plant in Manufac- turing 1956 to 1975 . . . . . . . . . . . . . . . 50 IV-2 Changes in Employment Structure by Plant Size, 1953 and 1956 to 1975 . . . . . . . . . . . . . . . . . 31a IV-3 Colombia: Age Distribution of Factory Establishments by Size, 1970 . . . . . . . . . . . . . . . . . . 55 IV-4 Plant Size in 1966 and in 1062, for Plants Identi- fied as Existing in Both Years . ... . . . . . . . 59 IV-5 Change in Average Size of Plants Identified in 1962 and 1966 (4496 Plants) . . . . . . . . . . . . . 60 IV-6a Changes in Plant Size Between 1970 and 1975 for a Sample of Factory Establishments Identified in Both Years . . . . . . . . . . . . . . . . . . . . 61 IV-7 Evidence of Growth Rates of Sets of Plants or Firms Identified over Periods of Time . . . . . . . . . 65 IV-9 Colombia: Estimated Death Rates of Factory Establish- ments According to Size, 1956-71 . . . . . . 70 IV-10 Colombia: Death Rates of Factory Establishments According to Size: 1965-71 . . . . . . . . . . . . 72 IV-11 Colombia: Estimated Death Rates of Factory Establish- ments According to Size: 1967-71 73 V-1 Estimates of Net Social Benefit per Unit of Capital, by Plant Size, Averages for 1956-1967 . . . . . . 80 V-2 Horsepower/Total Employment, 1966 . . . . . . . . . 85 V-3 Value Added to Horsepower Ratios for Small and Medium Plants Compared to Large Plants, 1965 . . . . . . 88 V-4 Characteristics of 5-Digit Sectors by Number of Estab- lishments in each 5-Digit Sector . . . . . . . 91 V-5 Colombia: Characteristics of Small and Large Firms in Selected Factory Sectors, 1970 . . . . . . . . . . 95 Table of Contents Table No. Titles Page No. A-1 Occupied Population by Economic Activity and Urban- Rural Breakdown, 1973 . . . . . . . . . . . . . . 119 A-2 Per Capita Income, Population, GDP and Sectoral Value Added by Departments, 1960 and 1975 . . 120 A-3 Estimates of the Job Position Structure in Manu- facturing, 1953, by Size of Plant . . . . . . . . 122 A-4 A Comparison of the 1953 Small Scale Survey and Total Cottage Shop in 1953, by Industry . . . . . 123 A-5 Composition of Small Scale Production Employment, 1951 and 1973 . . . . . . . . . . . . . . . . . . 124 A-6 Average Earnings by Plant Size: 1953 and 1970 . . . 125 A-7 Relative Productivity of Very Small Establishments 126 A-8 Job Position Structure of Employed Persons in Manufacturing . . . . . . . . . . . . . . 127 A-9 Commerce: Size Distribution of Establishments of at least Five Workers, 1970 . . . . . . . . . . . . 129 A-10 Job Position Composition of Salespersons, 1964 and 1973. . . . . . . . . . . . . . . . . . . . . . . 130 A-11 Size of Establishment, 1967: Wholesale and Retail . 131 A-12 Commerce: Very Small Establishments by Departments, 1970. . . . . . . . . . . . . . . . . . . . . . . 132 A-13 Colombia: Services: Size Distribution of "Large" Establishments 1970, by Type of Activity . . . . 133 A-14 Services: Employment in Very Small Establishments, 1970 . . . . . . . . . . . . . . . . . . . . . . 137 A-15 Services: Establishments of 1 to 4 Employees, by Departments, 1970 . . . . . . . . . . . . . . . . 138 A-16 Colombia: 1953 Industrial Census: Factory Sector . 139 A-16a Factory Employment by Departments, Selected Years . 140 Table of Contents Table No. Title Page No. A-17a Size Distribution of Establishments in Manufacturing, by Sectors, 1956 . . . . . . . . . . . . . . . . . . . . . . 141 A-17b Size Distribution of Employment in Manufacturing, by Sectors, 1956 . . . . . . . . . . . . . . . . . . . . . 142 A-17c Manufacturing: Va'ue Added in the Factory Sector Accord- ing to Size Qf Establishments, 1956 . . . . . . . . . . . 143 A-17c Employment in the Factory Sector, 1966 . . . . . . . . . . 144 A-17c Manufacturing Value Added in the Factory Sector According to Size of Establishments, 1966 . . . . . . . . . . . . . 145 A-18 Number of Establishments in the Factory Sector, 1966 . . . 146 A-18a Colombia: Size Distribution of Establishments and Employ- ment (June 1975) . . . . . . . . . . . . . . . . . . . . 147 A-18b Percentage Distribution of Establishments and Employment (June 1975) . . . . . . . . . . . . . . . . . . . . . . 148 A-19 Absolute Increase in Establishments and Employment between 1956 and 1975 by Sectors and Size Categories . . . . . . . 149 A-20 Percentage Distribution of Increase in Establishments and Employment between 1956 and 1975 by Sectors and Size Categories . . . . . . . . . . . . . . . . . . . . . . 150 A-21 Manufacturing: Size Distribution of a Sample of Factory Establishments According to their Position in 1970 and 1975. . . . . . . . . . . . . . . . . . . . . . . . . . . 151 A-22 Employment Growth of Plants Identified as Existing in both 1970 and 1975, by Number of Workers . . . . . . . . . . . 165 A-23 Composition of Factory Output Growth, 1962-66 and 1970-75 . 166 A-24 Implicit Underreporting in DANE and ICSS Employment Data by ?lant Size, 1975 . . . . . . . . . . . . . . . . . . . 167 A-25 Reported Average Wage (including Fringe Benefits), by Plant Size, 1956, 1963, 1970 and 1975 . . . . . . . . . . . . . 168 A-26 Description of 30 Manufacturing Subsectors . . . . . . . . 169 A-27 Colombia: Number nf Establishments and Employment in 30 Selected Factory Sectors . . . . . . . . . . . . . . . . 171 Table No. Table of Contents Page No. A-28 Estimates of Labor and Capital Productivity, and of Capital Intensity, by Plant Size, 1970 . . . . . . . . 174 A-29 Capital Labor Ratios by Output Level of Plant, 1970 . . 175 A-30 Colombia: Factory Sector Regressions of Capital Intensity by Firms in Specific Industry . . . . . . . 176 A-31 Colombia: Factory Sector Recessions on Labor Producti- vity (Log V) by Firms in Specific Industries . . 177 N A-32 Colombia: Returns to Scale and Productivity Differentials Between 'Small' and 'Large' Factories, as Derived from Regressions Estimates . . . . . . . . . . . . . . . . 178 A-33 Colombia: Sectors Classified According to Labor Pro- ductivity Differentials between Firms of 100 and 25 Workers. . . . . . . . . . . . . . . . . . . . . . . . 179 A-34 Relation of Horsepower and Book Value of Fixed Assets, by Two-Digit Sector, Cerca 1970 . . . . . . . . . . . 180 A-35 Estimates of Factory Employment, 1964, 1966, 1975 . . . 181 A-36 Labor Productivity by Industry and Size of Plant, 1956 . 183 Note on the Small Enterprise Research The following paper is one of a series of background and working papers on the research project on small enterprises. In Colombia, other papers in preparation include studies based on in-depth interviews with a sample of small and medium firms to obtain first-hand information on enterprise origins and growth, entre- preneurial backgrounds, product types and qualities, product markets, production technologies, use of capital and labor, sources of finance and other questions. A study of the markets for and use of second hand equipment is also afoot. The present paper is intended to complement these interviews by providing a broad background "mapping" of patterns of small enterprise development in Colombia; and it is intended to integrate the various studies once they are complete. Parallel studies, though with somewhat different emphases, are being undertaken in India and the Philippines. I. INTRODUCTION The subject of this paper is small-scale enterprises (SSEs) in Colombia, broadly defined to include not only established firms in manu- f'acturing, commerce and services, but also more "informal" activities . such as household industries, artisan production, etc. The idea of fenterprise" normally involves a definite physical location, production of goods or services for the market and some form of permanence and regularity which casual and other forms of employment may lack. Why are SSEs a special subject of study? Basically, because the following properties are attributed to them: First, SSEs are important in terms of employment and the demand for labor ; it is believed that promotion of SSEs should, by increasing the demand in labor, alleviate the employment problem in an efficient way and exert upward pressure on wage 1/ 2/ rates. Second, at least with regard to the informal sector, SSEs are believed to have a positive effect on the income distribution because of their labor intensity and their capacity to provide an alternative to those that could otherwise be unemployed or underemployed. Third, SSEs are believed to be a breeding ground for entrepreneurial talent. Less frequently, they are viewed as a place where otherwise wasted entrepre- neurial talents can be put to use, as with persons unable to exercise 1/ A critical discussion of the arguments for assisting SSEs is contained in P.I. Dhar and H.F. Lydall, The Role of Small Enterprises in Indian Economic Development, New York, 1961, Ch. 2. 2/ This is related to the presumptions that SSEs.are less capital-intensive and have higher capital productivity than larger firms. -2- their talents in larger firms. Finally, SSEs are believed to face substantially different conditions from those of large firms, in terms of access to credit, technical assistance and technology, ability to influence government policies, and other disadvantages whose cumulative effect is to.weaken their competitive position. Data and other limitations preclude the analysis of some of these issues in this study. The main questions that will be addressed are the following: 1. What are the main features of SSEs in manufacturing, commerce and services? What is the employment absorption of SSEs? In which sectors do SSEs predominate? 2. What are the trends in the size distribution of firms, and what factors explain these trends. Iow do structural changes in the economy affect the development of SSEs? What is the impact of regional or locational factors on SSEs? What is the role of economies of scale? In the aggregate or within well-defined industrial sectors: (a) are SSEs more, or less, "efficient" than larger firms?; (b) are SSEs more, or less, capital-intensive than larger firms? -3- 3. What are the dynamics of growth in SSEs? What are the characteristics of the age distribution of firms according to size? Are large firms predominantly 'old'? Are there significant differences in the survival rates of SSEs as compared to larger firms? What factors account for the larger rotation of SSEs? Is there upward mobility of firms in the industrial sector? To what extent are SSEs able to outgrow their initial condition and become medium or large firms? 4. What has been the impact of government policies on SSEs development? Do SSEs face significantly different conditions from those of larger firms with regard to credit, technical assistance, supply of inputs (imports) and technology? Finally, what are the prospects of SSEs in Colombia? -4- II. OVERALL ENVIRONEENT FOR SSEs AND ITS EVOLUTION OVER TIE While the Colombian economy continues to be substantially agricultural and small scale in orientation, these features are less characteristic with the passage of time. This section describes some structural changes that have taken place since 1950 and which may have had bearing on the development of SSEs. The share of agriculture in GDP (at factor cost in constant prices of 1970) has declined considerably, from 38.8% in 1950 to 26.9% in 1975. During the same period the share of manufacturing value added increased from 13.5% to 18,7%. Within manufacturing it is possible to distinguish between production in households and very small firms on the one hand, and 'factory' production on the other. (Factory production is defined, following the Colombian convention, as that occurring in plants of 5 or more workers, with some other plants included if their output level is high enough.) The available evidence indicates an absolute increase in both types of production since 1950 but a considerably larger expansion in the factory sector. Manufacturing value added (at 1970 constant factor cost) increased at an annual rate of approximately 6.6% between 1950 and 1975. The share of factory value added in total manufacturing value added, measuring in current prices, has risen from about 70-75% in 1950 to somewhere in the neighborhood of 80-85% in 1975, so that of establishments of less than five workers has fallen from 25-30% to 15-20%. l The national accounts report percentages of 26.2 and 12 esectively in 1958 pesos, and of 24.5 and 12.6 in current prices. These ratCS are likely to be understated, possible substantially. Urrutia and 'illaba made independent estimates in 1953 and 196A, resulting in (continued) -5- The structure of employment has been affected by the changing production patterns. The share of the agricultural sector in the total occupied population has fallen steadily since 1950, and even absolute agricultural employment has probably declined since 1964. From 1951 to 1964 it increased significantly, absorbing about 25% of the total increase 1/ in employment in the economy as a whole. This trend was reversed later: from 1964 to 1973 agricultural employment probably fell by about 5%, possibly even as much as 10%. The significant changes in patterns of agricultural employment also had repercussions on other areas of economic activity. Most important were the changes in demographic patterns: whereas from 1951 to 1964, 'cabecAras' absorbed about 78.3% of the net increase in population, from 2/ 1964 to 1973, this percentage increased to 85.2%. 1/ However, since the initial share of agricultural employment was 53% absorption of 25% of the increase'was not sufficient to maintain the share over time. 2/ DANE, Boletin Mensual de Estadistica #314, September 1977. (cont. footnote 1 of previous page) output levels ranging from 20 to 100% above those of the Banco de la Republica (which elaborates the national accounts). (Miguel Urrutia and Clara Elsa Villalba, "El Sector Artesenal en el Desarrollo Colombiano," Revista de Planeacion v Desarrollo, Vol. 1, No. 3, Octubre 1969.) Their range probably captures the true value of output, but since there is some underreporting of output in the factory sector as well, it is unlikely that the share of the cottage-shop or artisan sector in total manufacturing output would be more than 30% above that estimated by the Banco de la Republica, e.g., it would be quite unlikely to exceed 20% of total output in 1975. /l Ta].- 11-1: COL~M1lA: CIANGES IN EMIPLOYMENT STRUCTURE 1951, 1964, 1973 Percentage Distribution of Net Emiploymient Increase /2 /3 Employiment Tnereases _ 1951 1964 /4 onoice toDr ANE ANE- S `NA 1973 1951-73 1951-64 1964-73 1951-73 1951-61 1964-73 Agriculure 2,023,281 2,308,048 2,071,599 48,318 284,7é7 -236,449 2.45 25.27 -28.20 Mining 61,223 77,293 47,566 -13,657 16,070 - 29,727 -.70 1.42 -3.54 1anfacturing 460,907 623,796 879,334 418,427 162,889 255,538 21.29 14.45 30.48 Ellecricity, Gas, Water 10,472 12 625 33,192 22,720 2,153 20,567 1.15 .19 2.45 Construction 132,922 209,883 288,768 155,846 76,961 78,885 7.93 6.83 9.48 Commllerce 203,774 418,919 870,683 666,909 215>145 451,764 33.93 19.09 53.89 Trasportation 130,083 182,411 265,938 135,855 52,328 83,527 6.91 4.64 9.96 w inancial Services 598,093 880,542 110,647 647,619 282,449 365,169 32.95 25.06 43.56 Social Services 1,135,065 Not Specified 134,854 168,855 17,787 -117,067 34,001 -151,068 -5.95 3.01 -18.02 Total 3,755,609 4,882,372 5,720,579 1,964,970 1,126,763 838,207 100.00 100.00 100.00 /1 Tie concepts used are as follows: for 1951, econoimically active population (ineludes thde nnepiloyed and considers people of at least 12 years of age). For 1964 and 1973: occupied population (refers to people of at least 12 years of age in 1964 and of at leasl 10 years of age in 1973). /2 DANE, Censo de Poblacion de Colombia 1951 RTsumen, Table 33, p. 154. / ' SENA, Plan Quingavl Capitulo 1, Diagnostico 1977. Based on DANE, XIII Censo Nacional de Poblacion, 1964, Table 36 /4 Author's est imates based on the latest revisioiis of population totals for DANE's XIV Censo de Poblacion, 1973 (INef314 ; and National lousehold Sur-:z; for 1971. See Table A-i below). Sources: See afotnot es above. Note: 'Th 1964 figures ire (sLiimtes S4ICe Llie census did not present data on employIent by seCtor. The 1973 figures arL LStllInaLes since theu fiiil eensus results have not been published. Tle total eIploymenl figurte for 1973 is probably too. -7- Between 1951 and 1973, employment in manufacturing about doubled. (See Table III-1 below). However, the growth waw nct uniformly distributed over this period. From 1951 to 1964 the annual percentage growth in manufacturing employment was 2.2% whereas from 1964 to 1973 it was 4.5% or higher. In other words, whereas on average 12,100 new jobs were created in manufacturing between 1951 and 1964, from 1964 till 1973 there were over 33,000 new jobs created every year, or nearly three times as many as in the previous period. Other sectors that have absorbed a large share of the employment increase after 1964 are commerce and services. Outside agriculture, the additional employment generated between 1964 and 1973 was distributed as 1/ follows: Manufacturing (20.8%); Construction (6.5%); Commerce (36.8%); Transportation (6.8%); Services (29.8%) and other sectors (-0.7%). In other words, about one in every three additional jobs outside agriculture came about in commerce, and almost 30% in services. It is of interest to consider whether a significant proportion of the employment increase in these sectors involved underemployment, either in the form of part-time engagement or very low productivity occupations; this subject which will be expanded on below. Average labor productivity and average incomes of persons engaged in the manufacturing sector are above the average for the economy as a whole. In 1973, When about 15.5% of the employed labor force was engaged in 1/ Distributing the changes in employment is unspecified sectors among the remaining sectors in proportion to their unadjusted shares. These figures are tentative since the 1973 census has not been published in final form. -8- manufacturing, that sector's share of value added at factor cost was 21.3%. Correspondingly, the share of poor families whose head is engaged in manufacturing is a little below the share of all employment in that sector. Indirect evidence on the income structure associated with diverse branches of economic activity can be obtained from a 1974 house- 1/ hold survey, in which the households were ranked according to their household per capita income. Table 11-2 presents the distribution of the bottom twenty percent of the households (i.e., the 'poorest' quintile of households, according to the main economic activity of the head of the 2/ household. The bottom 20% of households receive only 5.2% of income, indicating that their per capita income is about 25% of the overall Colombian per capita income. For the country as a whole, 58.2% of the poorest quintile of 3/ households are in rural areas, indicating that the income structure still contains the seeds that will motivate future migrations to urban areas. The distribution of the poorest quintile of households presents the follow- ing features: the urban self-employed and wage labor account for 16.0% and 18.6%, respectively, of the bottom 20% of all households. The urban 1/ See Selowsky, Marcelo, The Distribution of Public Services Across Income Groups: A Case Study of Colombia, World Bank, 1977, mimeo. 2/ This information does not reflect exclusively the earning capacity of the head of household since it also depends on size of the household and number of earners in addition to the head of family. However, if it is assumed that these qualities are randomly distributed among households independently of the specific sectors in which the heads of household are occupied, then the data will approximately reflect the earning capacity of the main wage earner in poor households. 3/ This figure is probably upward biased due to greater relative underreporting of income in rural as compared with urban areas. -9- Table 11-2: COLOMBIA: OCCUPATIONAL DISTRIBUTION OF HEADS OF HOUSEHOLD IN THE POOREST 20% OF HOUSEHOLDS (%) I. Urban 1. Self-employed 16.0 a. Manufacturing 5.2 b. Services 10.8 2. Wage labor 18.6 a. Manufacturing 4.9 b. Construction 5.5 c. Services 8.2 II. Rural 58.2 III. Ill-Defined Categories.L 7.2 100.0 1/ Composed mainly of self-employed in household activities. Source: Marcelo Selowsky, Op.cit., p. . - 10 - manufacturing sector comprises 10.1% of the poorest quintile of households, 5.2% are self-employed and 4.9% are wage earners. About 10.8% of the poorest quintile of households are self-employed in urban commerce and services, and 8.2% are wage-earners in those sectors. Persons in "ill defined" activities account for about 7% of the heads of household in the poorest quintile. Some of these persons are likely to be self-employed and engaged in household industries; as the data suggest, such activities are characterized by low earnings and productivity. In many instances these informal SSEs arise as a response to the lack of more lucrative employment opportunities in the economy at large. Since some rural families are engaged in manufacturing, and some household activities also fall in this category, perhaps more like 15% of household heads in these families are engaged in manufacturing of one sort or another. Independent evidence from a sample of very small firms (1 to 4 workers) in 1970, reveals that average earnings in these firms are consi- 1/ derably below the levels paid in the factory sector. Consequently, it can be ex.pected that among the poorest heads of households, those that are wage-earners are engaged primarily in small firms, since minimum wage legislation and other legal constraints (e.g., social security) are generally easier to avoid there than in larger firms. Foreign Trade Regimes Foreign trade regimes have had an indirect influence on the development of small enterprises. The trade regime that prevailed from 1/ See section III.1 below. Note that wages for persons of comparable skills differ much less than do average earnings by size of plant. But persons with relatively low skills appear to be found dispropor- tionately in smLller establishments. - 11 - the 1950s up to 1967 can be characterized as one of heavy emphasis on import-substituting industrialization and frequent foreign exchange crises due to the maintenance of a fixed exchange rate in the presence of domestic inflation. The mechanism that was established for dealing with these crises 1/ was a system of import licenses biased in favor of large firms. As a result, small enterprises were frequently denied access to imports of machinery and raw materials, thus diminishing their ability to compete with larger firms. In 1967 the trade regime was liberalized at the same time that a floating exchange rate was adopted. The import system was liberalized and exports were encouraged. Since then the discriminatory import policies against SSEs appear to have been largely eliminated. On the export side, however, it appears that mostly large firms have been in a position to exploit the 2/ available fiscal incentives for exports. Locational and Regional Factors Locational and regional factors have exerted a strong influence on the development of small-scale enterprises in particular, and on producion and valued added in manufacturing, commerce and services in general. From 1960 to 1975 the largest three departments in terms of population and GDP 1/ Diaz-Alejandro, C., Foreign Trade Regimes and Economic Development in Colombia, "ew York, National Bureau of Economic Research, 1976, Chapter . 2/ See Albert Berry and Carlos Diaz-Alejandro, "The New Colombian Exuorts: Possible Effects on the Distribution of Income," in A. Berry and R.. Soligo (editors), Economic Policy and Income Distribution in Colombia, Boulder, Colorado, Westview Press, 1979. - 12 - (Bogota DE, Antioquia and Valle) absorbed the largest share of the net 1/ 2/ increases in population (46%) and GDP (50%) (See Table A-2). The initial level in 1960 and the changes which took place in per capita income, population and GDP in each department between 1960 and 1975 provide concise information on the level and evolution of demand in each region. Those departments where demand has expanded fastest have been able to generate larger net production increases in manufacturing, commerce and services. However, the expansion in these departments is probably not entirely attributable to demand factors. The presence of a strong industrial base in 1960 has meant that technological progress plus the benefits of increasing returns to scale could accrue over time as the scale of operations increased either as a result of an internal expansion of departmental demand or due to easier access to other markets (for instance, through lower transportation costs). On the other hand, the less developed departments have been able to capture only a minor expansion of productive activities in manufacturing, commerce and services. 3/ With respect to cottage shop prouction the firt feature to be noticed is the relatively low- level and snaller increase vis a vis the 1/ Lach of sirilar data for previous years inhibits the analysis but it is thought that the trends were similar to those taking place after 1960. 2/ These increases are substantially larger than the corresponding shares of these departments in 1960. In other words, the largest departments expanded population and GDP at higher rates than the country as a whole. 3/ Comprises production in household industries pl,s very small firms of less than 5 workers and annual production of less than Col.$24,000. - 13 - factory manufacturing sector. The share of these very small productive units in manufacturing value added was probably about 20-25% in 1960 and 1/ 15-20.7% in 1975. Only about 10% of the total expansion in manufacturing 2/ value added between 1960 and 1975 was in such enterprises. This expansion was distributed mainly in response to departmental population changes, as can be seen from comparing the departmental shares in the respective net increases. (See Table A-2). The most industrialized departments have absorbed a lower share of the increases in cottage-shop value added than of total manufacturing. This reflects on the one hand the local character of demand for cottage-shop output, and on the other the disadvantages it faces when confronted in larger markets or alternatively, the weaker linkages that exist between cottage-shop and the factory sector. This description applies mainly to the departments of Antioquia and Atlantico and is less adequate in the cases of Bogota DE and Valle, where a non- negligible expansion in cottage-shop has taken place. / See the discussion at the beginning of this chapter. 2/ The extent to which SSI may have expanded to become medium or even large firms is the subject of Chapter IV Section 3. Such an expansion is unlikely to have taken place in the smaller or stagnant departments, however. 3/ Note that the data in this table show a smaller share of value added in cottage shop than indicated in the text above. They apparently are based on unadjusted output estimates such as those included in the national accounts. - 14 - !II. MANUFACTURING, COMME RCE AND SERVICES 1. HANUFACTURING A. Factory Sector Overall employment in manufacturing increased at an average annual rate of about 3.2% or a little higher over the period 1951-73. (See Table III-I). The factory sector--defined as including firms of five or more workers--expanded its employment at a higher rate of 4.5% so that its share in total employment rose from 4.8 to 7.7%; the non-factory sector grew at 2.0 to 2.3% per year, probably less than aggregate employment. Consequently its share of manufacturing employment declined from about 60% in 1951 to 47-50% in 1973, while its share of total employment fell from about 7.3% to somewhere in the range 6.9 to 7.3%. Within the factory sector, the share of employment fund in smaller firms has fallen slightly over the same period. Although the informal sector accounts for a substantial share of manufacturing employment, its impact on overall value added is considerably less significant, i.e., it has a substantially lower labor productivity than does the factory subsector. Probably 25-30% of manufacturing value 1/ added in 1953 and 15-20% in 1975 were generated in this sector. The 1/ Banco-de-la-Republica estimates for value added in manufacturing distinguish between factory and non-factory, including in the latter household industries and firms of less than 5 workers with annual production of less than Col$24,000. These definitions are thus com- parable with the ones used for reporting employment except -or 1973-75 where the data only cover firms of at least 5 workers, independently of the firm's output value. The Banco's figures indicated that 26.4% of manufacturing value added came from the non-factory sector in 1953 and 12.7% in 1975. There is evidence, however, that these figures understate the contribution of the informal sector. A study by Rliguel Urrutia and C. 7lsa de Sandoval "El Sector Artasenal , .ci) concluded that the Banco de la Republica data are li-ely to underestimate value added in the informal sector by at least 28%. Banco output data imply that the informal sector share of value added has fallen more sharply than its employment share, indicating larger productiviLy increases in the factory sector. But no solid data underlie these estimates. -15 - informal sector's shares of employment and value added in the manufacturing sector have thus declined over time. Employment in the informal sector is still a substantial percentage of total manufacturing employment, but its productivity (value added per worker) is lower and has increased at a lower rate than in the factory sector. Within the factory sector the share of factory employment and value added in small firms (5-49 workers) has also decreased over time. In other words, currently large firms (50+) are the ones that have expanded employment and labor productivity at the highest rates, followed by smaller firms (5-49), and by informal enterprises. Some informal sector workers are engaged only part-time in manufacturing activities. This may explain part of the difference in productivity per 1/ worker between the informal and formal sectors; the rest is attributable 1/ Statistics bearing directly on this issue are not available. it is clear that some small operations operate on a less continuous base than do larger ones due to difficulties in acquiring raw materials, inability to operate without the manager, etc. Whether this affects the productivity calculations depends on how the figure for persons involved in the informal sector is derived. In this study the objective has been to estimate a full time eouivalent but, especially for the earlier census years, where tne distinction between being employed and participating in the labor force may have been imprecise, this may not have been achieved. Note that, when all activities are taken into account, persons engaged in informal sector activities may work as many or more hours per week than do those working in the modern sector. In a 1975 survey, Kugler, et al., found that in the moden sector men worked an average of 52.9 hours per week and waomea 3 .8, hile in tie informal sector the corresponding figures were 57.9 and 60.2. (Bernardo Kugler, Alvaro Reyes and Martha I. de Gutierrez, Education y Mercado de Trabajo Urbano en Colombia; Una Comparacion Entre Sectores Modernos y no Hodernos, Bogota, Corporacion Centro Regional de Poblacion, Julio 1978, p. 36A. Bourguinon, using data from a DANE survey on employment/unemployment found persons in the tradi- tional sector reported 6-7% less hours per week than in the modern sector. (Francois Bourguignon,"Poverty and Dualism in the Urban Sector of Developing Economics: The Case of Colombia," mimeo, February 1973.) Probably the data used by Kugler, et al., is the more reliable. The two definitions of modern were slightly different, so this too could have explained the discrepancy. In neither case did the author oresent a comparison of hours worked in the traditional and modern subsectors of manufacturing. DAE's industrial statistics show data on hours worked by sector and there is no obvious general tendency in industries where most workers are in small plants to report lower hours per worker than do industries where most workers are in large plants. (e.g., DATE, I Censo T_dustrial 1970, pp. 22-26. -16 - to differences in endowment of capital per worker, technology, education, market position (larger firms tend to have oligopoly or monopoly power), etc. Within the factory sector labor productivity differs substantially by the size of establishment. In 1953, the share of recorded factory value added generated by small plants (5 to 49 employees) was 28.9%, although employment in these firms was 46.7% of factory employment. In 1975, plants in the same size group accounted for 30% of factory employment, but only l/ 13-15% of factory value added. Consequently, over this period the increase in productivity per worker in small plants (5 to 49 workers) was lower than in the factory sector as a whole; the average productiviy per worker in these plants in 1975 was only about 40% of the average level corresponding to the factory sector as a whole. The non-factory sector includes household industries on the one hand and small shops of less than five workers on the other. These are the subject of the next two subsections. Household industries accounted for approximately a third of all employment in manufacturing in 1973, and very small firms represented about 12%. Thus the great majority of workers not in establishments of 5 or more workers were found in household industry. B. Household Industries Household production is especially important in consumer goods; about 74% of household manufacturing employment (as recorded in the 1973 1/ Using unadjusted DANE data, the latter figure was 9.5%, but 13.8% if value added is expanded to correct for undercoverage in small-size categories (in the same proportion in which ICSS data exceeded DAE employment figures for each size category). - 17 - population census) is concentrated in food, beverages, tobacco, textiles, clothing, footwear and leather. (See Table 111-3). Household employment accounts for over 50% of total employment in food beverages and tobacco, for 40% in wood and furniture, and for 35% in textiles, clothing, footwear and leather. It is also important in some non-traditional sectors although to a smaller extent, e.g., household industries provide one out of four jobs in nop-metallic mineral products. Household employment, like overall manufacturing employment, is predominantly urban: 69.7% of household employment was located in 1/ increasingly 'cabeceras' in 1973. These ratios reflect Colombia's/urbanized character. In 1973, 49.5% of the population lived in cities of more than 10,000 2/ 3/ inhabitants and 25% in four cities of more than 500,000 inhabitants. Household employment in rural areas is concentrated mainly in food, beverages and tobacco (61.2% in 1973) and textiles, clothing and 4 / leather (29.0%). (Table 111-4). In contrast, only 30% of employment in the urban household sector in 1973 was in food, beverages, and tobacco and 1/ 'Cabeceras' are administrative centers which closely overlap 'urban' areas, the latter being defined in Colombia as centers of at least 1,500 persons. I.e., whereas 'urban population in 1973 was 12,550,441 that of 'cabeceras' was 12,453,339. (Unadjusted DANE data from the 1973 Population Census). Urban population accounted for 59.5% of the population in 1973. 2/ DANE, Boletin Mensual de Estadistica #314, September 1977, p. 30. 3/ There may be a problem of classification in that essentially rural centers are listed as urban. However, this is not likely to be significant' in 1973 only 10% of the population was registered in 'cabeceras' of less than 10,000 inhabitants. 4/ Percentages are calculated with respect to the total number that provided information on industrial activity, -oble II-l: '9MLOY.Itri AD LABOR FORCE 1 N ANUTACUTIING, ST SIZE OF EStABLISM1ET AD BY haLU AAN CENSUS YEARS thousands) 1951 1964 1973 1. Tocal Labor Force 433 675.6 1025-1053 Id Is or higher" zol.oy-enc 464 621.3 925-550 or higher ?Lants of 5 workers 184.5 310.0 489 Ocher 279.5 311.5 436-461 :ndependeat Workers & =mily 172.3 188.3 Helpers k Household 319,04" Uneployed 19 54.1 100-103 2. Urban Labor F.orce Cabeacera 383 370.3 Emp Loyment .lanrs of 5 %orkers Ocher Independent 'orkers & 9anity Relpers Sousehold 222.4 /n /1 3. Rural Labor Force kResco de Los 100 104.8 1Hunicipios) Employmenc ?Lan s of 1 5 Workers Other Zndependen WOrkers & Fmily Helpers Lk Household 96.6 , Census figure plus 51. b Census figure plus 3%. It An esriace of 1023-1039 was reached on the basis of a cotal populatio of 22.7 million and the assumpcion that 16 to 18% of persons for whom sector of activity was ac given in the census were found In manufacturing, excepc for persons In the ared forces for vhm the rato was assumed to be zero. A higher rario was plausible than for persons whose sector was reported, since a disproportionace share of chose en2erared but for whom sector of activity was nor repored were In occupations usually associated wich nanufaccuring. This range appeared a litcle lay given the asimatr of amploymenc and was adjusted upward slightly. I. Arbitrarily assumes 1% unemploymenr. L- Arbicarily assmes 8% unsploymenc- f he lower eactnae assumes the sare share of coral smployemec in amaiacuring as is reported Io Table A- (15.21.) , though wich a higher scinace of tcal employment (6014.3 chousand). Th, *pper ascimace Is chose co be modecly higher, ::c an upper Ui-nic ascimac. it is well below che figure which would merge if che share of uran employment in manufacturing were esciiaced as the average of the figures reported in the household surveys of !are 1972 and Lace Z974 cspectiveLy '-56 and E28). Accepting the share of rural emlaoyment found .n manufacturing as char shown to 7able A-L, and che Just cited rechodology for the arban figure would imply a cocal of 985 thousand. 3ne of :he 171 nacional surveys presenced o figure of over 1 million an upper li:iic. ... Or with oucput above 24,000 pesos. ib Based on an esti-ace of 200,000 in 1953, which 10 tn turn based on the induscrial nsua of that year. It Itsced 199,116 employed persons, bur about 10.5 thousand were in plants of Less chan 5 workers. Although this census sems to have had better coverage than any succeeding one, a comparison of Its reporting of job position scruccure wich the L951 population census implies an Ltplausibly high zu:noer of paid employees to be found in plants of less than five workers. We have therefore assumed a faccory I.e. planc of workers or more) labor farce of 200,000 In 1953. Even this fIgure leaves he ratio of paid workers/cocal employment -a the coctage shop sector iMlausibly high, ar 28.7, compared with ur ascrimaca of 18-19 in 1964. .be implicit growh rate of 4.17. per year over 1957-64 was e:traoolaced back to .951 to arrive ac the figure of 134.5. . 'ansu3 fiure plus 4. Our ecinace of manufacturing labor force is the census .igure adjuscad upward by 4.5%. Caasus figure plus 2.5%. Zur esci=ace of the manufaccuring labor force is adjusted up 3. frem roar of the census, and it seems likely chac few independent warkers would ce "inemployed," ased .n a ilignti laowet escimace of the coral Labor force than used tn calculacing :Ms resc of the figures rsenred here, rqm Urrucia and Villalba, p,cit., pp.0-4?, adjusted upward. Based on faca from the 1951 population census which recorded number of persons living n 'resco e los municiptas and lapenent on the varIous sactors. Table Ill-2. COLOMBIA: EMPLOYMENT IN MANUFACTURING SECTOR (Absolute Figures in Thousands) Annual Growth Rate () 1953 1964 1973 1975 19A3-75 1953-64 1964-73 1973-75 1. FACTORY SECTOR (5 or more 200.00 310.0 489.0 518.3 4.4 4.1 5.2 3.0 workers) Small Factories (5 to 49 employees) 93.8 125.7 153.0 156.4 2.4 2.7 2.2 1.1 Large Factories (50 or more workers) 106.2 184.3 336.0 361.9 5.7 5.1 6.9 3,8 11. NON-FACTORY SECTOR 285.1 311.5 436-461 n.a. 2.3-2.6 0.8 3.8-4.4 n.a. Very small enterprises (1 to 4 employees, not household industries) 116-141 Household Industries 320 III. TOTAL 485.1 621.5 925-950 n.a. 3.5-3.6 2.3 4.5-4.8 (or higher) Non-Factory as % of Total Employment 58.8 50.1 47,1-48,5 (or higher) Source: DAN, Population Census, Industrial Census and Annual Survey of Manufactures. MeLhodology: For all years, data for factories of 50 workers and up are from DANE; Annual Survey of manufacturing estimates of Total Factory Employment (Plants of five workers or more) are from A. Berry, "A Descriptive History....." op.cit., Statistical Appendix, Table A-183. That source estimates underreporting by a variety of cross checks with other Sources. Employment in plants of 5-49 workers is calculated as a residual. Total manufacturing employment is based on the population censuses of 1951, 1964 and 1973. The 1953 estimate used here is an interpolation between 1951 and 1964. Since definitive results of the 1973 census have not yet been published the estimate for that year is based on the partial data available (see also Table A-1). The non tactory sector is calculated as a residual. Table i1 -2: COLNIl A: EMPLOYMENT IN MANUFACTURINc SECTOR (Ab:olute Figures ijt: Thousands) AnulI Growth Rate ( 1953 19614 1973 1975 19:,3-75 1953-64 1964-73 1973-75 vAT'oURY SECTOR (5 or mort, 200.00 310.0 489.0 518.3 4.4 4.1 5.2 3.0 woi, ke rs) SmAlA Factories (5 Lo 49 employees) 93.8 125.7 153.0 156.4 2.4 2.7 2.2 1.1 Large Factories (50 or more workers) 106.2 184.3 336.0 361.9 5.7 5.1 6.9 3.8 11. NloN-FACTORY SECTOR 285.1 311.5 436-461 n.a. 2.3-2.6 0.8 3.8-4.4 n.a. Very sma11l enterprises (I Lo 4 employees, not hotusuh id industries) 116-141 loutsehold Industries 320 111. TOTAL. 485.1 621.5 925-950 n.a. 3.5-3.6 2.3 4.5-4.8 (or higher) Non-Factor1 as 7 of ToLni IlployielL 58.8 50.1 47.1-485 (or higher) Source: DANE, PopulatiOn Cunsus, industrial Census and Annual Survey of Maulfaeturus. "Or all years, data for f,ctuies of 50 workers and up are from DANE; Annual Survey of Lianufacturing estimates of oIal acLory Employment (Plants of five workers or more) are from A. Berry, "A DeSCriptive History.. " op.cit., Stiatist lu] Appendix, T,ab l A-183. That source estimates undecreporting by a variety of cross checks with other Sources. Employment in plants of 5-49 workers is calculated as a residual. Toual manufacturing eploymeut is based on the population censuses of 1951, 1964 ai. 1973. The 1953 estimate used here is an interpolation between 1951 and 1964. Since definitive results of the 1973 census have not yet 1eeu1 published the estimate tr thaL year is based on the partial data available (see also Table A-1). The uil lactury sector is calcuoled as a residual. Table 111.3: EMPLOYMENT IN HOUSMHOLD INDUSTRIES, 1973 Number of Total Households /2 Employ- Employment According to Number of Workers in the Household Manufacturing Sector With Industry ment 1 2 3 4 5 6+ Food, Beverages Tobacco 35,697 104,271 l,300 18,250 16,787 16,374 11,773 29,787 Textiles, Clothing Leather 43,105 93,928 21,953 22,925 12,777 10,168 5,980 20,125 Wood, Furniture 11,068 26,752 5,424 4,937 3,913 2,973 1,626 7,879 Paper, Printing 1,459 4,066 411 946 1,038 239 314 1,118 Chemicals, Oil and Coal Products, Rubber and Plastics 1,895 6.452 501 1,106 478 699 1,248 2,420 O Non-Metallic Minerals 3,920 12,549 1,280 2,311 1,137 1,823 1,011 4,987 Basic Metals 243 403 130 132 141 - - - Metal Products, Machinery 2,914 10,640 1,105 1,227 1,505 1,082 761 4,960 Other 3 133 7,974 1,245 1,832 1,217 784 737 2,159 Without Information 21,763 52,009 9,145 13,351 7,224 6,004 3,209 13,076 Total 125,197 319,044 52,494 67,017 46,217 40,146 26,659 86,511 /1 A household industry is one which produces manufacturing goods in the household for sale. This category does not include shops, stores, beauty parlors, etc. (DANE XIV National Population Census and III of Hlousing, 1973, p. 37). /2 Total Number of households in 1973 was 3,471,834. Source: DANE, XIV National Population Census and 111 of Housing, 1973 Advanced sample and departmental summaries Table 13, p. 52-3. The original data have been expanded according to the under- enumeration refJected in the 1973 Census. The latest population totals for 'cabecera' and Iresto' are taken from DANE, BME 314, September 1977, p. 30. The number of household industries and employmunt hiavc been expanded separately in 'cabecera' and 'resto' in the same proportions in which the JuteLSt revisions exceed the original population estimaes in 'cabecera' and to' r~'ciuv Table 111.4: EMPLOYMENT IN HOUSEHOLD INDUSTRIES: "REST", 1973 Number of Total Households /2 Employ- Employment According to Number of Workers in the Household Manufacturing Sector With Industry ment 1 2 3 4 5 6 Food, Beverages, Tobacco 14,205 47,426 3,604 6,137 7,087 8,463 6,347 15,788 Textiles, Clothing Leather 10,926 22,469 4,825 6,635 3,272 4,715 1,209 1,813 Wood, Furniture 1,809 2,832 1,209 837 181 242 151 212 Paper, Printing - - - - - Chemicals, Oil and Coal Products, Rubber and Plastics 213 470 76 152 91 - 151 - Non-Metallic Minerals 1,251 3,838 378 786 453 643 79 1,499 Basic Metals--- Metal Products, Machinery - - Other 181 392 60 181 - - 151 - Without Information 7,145 19,338 2,397 4,739 3,039 2,374 1,590 5,199 Total 35,730 96,765 12,549 19,467 14,123 16,437 9,678 24,511 /1 "Rest" refers roughly to rural. /2 The total number of rural households in 1973 was 1,222,492. Source: See Table 11-3. 一 23 37.7% in textiles, clot'-fting and leather. (Table 111-5). The remaining urban household employment was in wood and furniture (12.6%), metal products and machinery (5.6%), non-metallic mineral products (4.6%), and other minor activities. Over the period 1951-64 there was a significant increase in the share of manufacturing employment in factories (establishments of at least five workers), from 40% in 1951 to about 50% in 1964. (Table III-1) During 1964-73 it is not yet entirely clear how the artisan or non-factory sector evolved,'but a best guess (see Tables III-1 and 111-2) is that its employu-., growth was quite rapid, not too much below that of the factury- subsector, so that the factory subsector's share of total manufacturing employment rose only to 51-53% by 1973. During this latter intercensal period, the employment absorption of factories does appear to have been faster in intermediate and capital goods industries. Textiles, clothing, footwear, leather, wood and furniture also register increases in share of factory employment. The only group of industries where there appears to have been a significant decline in the share of employment in factories is food, beverages and tobacco. Although factory employment increased in absolute terms in these sectors, aon-factory employment seems to have increased even faster. (These results are tentative pending availability of data from the 1973 population census). During 1951-64 the factory sector grew significantly faster than the small scale or cottacle shop sector, especially in textiles (where cottage shop employment fell in absolute terms), rubber and products, metal and products (though growth was fast in cottage shop as well). In tobacco, factory employment fell sharply while cottage Manufacturing Yzploymcnt by -wo Digit Irdustriea and by Fnctory/Cottage-Sbop 1938-1964 Industry 1938 1944/5 1951 1964 Food-Total 29.387 47,311 73,889 Factory >6,153 32,172 39,355 /a /c 41,561. Cott;gc-Shop C,23,234 7956 -:- 32,328 Baveragcs-Total 6,358 11,772 19,216 Factory >:1,975 9,671 10,699 16,420 CottagL-Shop j.4 , 383 1073 2,796 Tobacco-Total 10,167 10,654 8,574 Factory >5,059 - 8 - - 2,099 Cottago-Shop 4 5,108 3,555 Textiles-ToLal 114,68 b 60,691 62,362 Factory },1,551 29,961 35,127 44,099 Cotte.e-Shop 10,133 25,564 18,263 Clothing & Footear Total 160,783 143,933 153,265 Factory >R,770 18,347 7j,05. 31,510 Cottage-Shop e157,998 117,1137 121.755 Wood & Products Total 44,356 57,202 78,245 factory n.a. 10,154 8,804 10,989 Cottage-3hop 44,356 48,398 67,256 Paper & Products-Total 1,798 5,613 Factory 610* 1,636 5,485 Cottage-Shop 162 128 Pxinting-Total 7,132 10,767 16,935 Factory n.a. 5,552 7,263 11,812 Cottage-Shop 47,132 3,504 5,123 Leather & Productz-Total Tutal 7,024 6,006 9,488 Factory 3,409 3,867 * 4,482 Cotta,,-Shup <7,024- . 4,139 5,006 Rubber & Products Total 2,329 5,425 Factory 576* 2,280 6,900 Coctage-Shop 49 -1,475 Chemicals-Total 4,722 25,752 Factory ,27111,146 19,819 ac 2761 5,E90 8,958 Cottage-Shop (11961 2,188 5,933 Petroleu-a & Coal Products Total 1,607 4,880 Factory 1,142 1,477 2,026 CottU&-Shop 132 2,854 Non-Ifetallic Kincral Producto-Total J',5.7 25,592 36,907 Factory ,' ,307 13,291 16,840 25,493 Cottage:-Shop 113,710 8,752 11,414 Metal 6 Ietal Products Total 25,22f 55,754 139,426 Factory ' 1,417 9,414 : 79 51,549 Cottage-Shcp 413,00) 87,877 Total =..435.863 655,961 Factory - - - -. 6528,9571 151,7 283,571 CottaE-Shop 272.4' 372,390 a) Greater-r than - i of thc fi=r,.n in the cc1-nt a:t hedes t11,tories (non-deputenoatc) for vh.cn no he to brpiti r le-:t dcwn at too two dir,1 lc-el wcj ,,do, an d a w .auvtries clain.tcd ac iellaceus b) Innlndrk '0,,5 pcriia an.d `a "rndustrios ofAn . '.d Te-stoolo nibre" . a w-b:t81LTvro rxn. c) Dubicius figures. May be due to misreporting. paz-in pe n~eJ-.j, ot CP,rn, 1 ~~ :72H3Jttcer were itat L'-,! c.,I - 25 - shop grew modestly. As Table 111-6 highlights for this period, employment in both factories and cottage shop tended to grow fairly fast in the dynamic sectors like metal and products, chemicals, and printing; in industries where overall growth was slow, as in clothing/footwear, employment often grew fast in neither subsector. In textiles factory employment grew at the expense of cottage shop. C. Very Small Firms (Shops) Under this category are included those establishments of 1 to 4 1/ workers which are not household enterprises. It appears that in 1973 2/ about 12-15% of total manufacturing employment was in these plants. They have several common characteristics with household industries. The sectoral distribution of employment is fairly similar except that the share of food, beverages and tobacco is larger for household industries and that of textiles, clothing, footwear, and leather is smaller than in very small establishments, at least judging from the employment composition reflected in the 1970 DANE survey (Table 111-3). Probably about 30% of the employment ir, wood and furniture in 1973 originated in these very small establishments; their employment share was also important in textiles, clothing, footwear and leather (14.9%);food, beverages and tobacco (10.6%); metal products and machinery (10.7%);and paper and printing (11.7%). 1/ It is a question of definition whether household activities should in principle be considered establishments, but it is clear that a census of establishments would not normally cover them simply because of the problem of identifying such activities. 2/ In this year it is possible to make an estimate of the employment in very small, non-household enterprises by subtracting factory and house- hold employment from total manufacturing employment. Table III.7t: DISTRIBUTION OF ESTABLISHMENTS, EMPLOYMENT, WAGES, PRODUCTION AND VALUE ADDED BY MANUFACTURING SECTOR FOR VERY SMALL FIRMS /1 (SECOND INDUSTRIAL CENSUS, 1953) Wages, Social _ Epl0yment Security Number of Payments Production Value Added Manufacturing Sector Establishments Total Paid Non Paid (000's $ 1953) (000's $ 1953) (000's $ 1953) Food (2,470) (5,890) (1,790) (4,100) (1,069) (21,995) (7,129) Beverages (260) (520) (210) (310) (222) (1,742) (919) Tobacco (800) (1,910) (770) (1,140) (463) (3,993) (1,971) Food, Beverages, Tobacco 3,530 8,320 2,770 5,550 1,754 27,730 10,019 Textiles (1,770) (4,070) . (510) (3,560) (313) (9,530) (3,601) Clothing /Footwear (18,260) (28,380) (6,520) (21,860) (6,551) (116,399) (47,482) Leather (840) (1,420) (380) (1,040) (368) (5,591) (2,338) Textiles, Cloth./Ft., Leather 20,870 33,870 7,410 26,460 7,232 131,520 53,421 1 Wood (2,040) (3,520) (1,050) (2,470) (856) (10,351) (5,793) Furniture (2,980) (5,050) (1,620) (3,430) (2,182) (16,432) (9,575) Wood, Furniture 5,020 8,570 2,670 5,900 3,038 26,783 15,368 Paper (10) (10) (-) (10) (-) (48) (31) Printing (240) (660) (430) (230) (590) (2,789) (1,994) Paper, Printing 250 670 430 240 590 2,837 2,025 Rubber (60) (100) (10) (90) (7) (128) (112) Chemicals (470) (1,040) (350) (690) (247) (3,887) (1,569) oil & Coal (H H- H H Rubber, Chemicals, Oil & Coal 530 1,140 360 780 254 4,015 1,681 Non Metallic Minerals 1,460 3,930 1,750 2,180 1,489 6,462 4,306 Basic Metals 50 110 50 60 107 511 371 Metal Products (1,330) (2,330) (850) (1,480) (1,055) (7,608) (4,663) Non Electrical Machinery (80) (160) (50) (110) (59) (430) (280) Electrical Machinery (650) (1,030) (300) (730) (309) (4,222) (2,490) Transport Equipment (920) (2,300) (1,090) (1,210) (1,086) (7,507) (4,401) Met. Prod., N.E.M., E.M., T.E. 2,980 5,820 2,290 3,530 2,509 19,767 11,834 Other 1,420 2,130 370 1,760 571 8,986 5,576 Total (36,110) (64,560) (18,100) (46,460) (17,544) (228,611) (104,601) /1 Defined as firms having less than five workers and less than $24,000 of output in 1953. Source: DANE, BMF 72, March 1957, p. 15-17. Table III.m. PERCENTAGE DISTRIBUTION OF ESTABLISIMENTS, EMPLOYMENT, WAGES PRODUCTION AND VALUE ADDED BY MANUFACTURING SECTOR FOR VERY SMALL FIRMS /. (SECOND INDUSTRIAL CENSUS, 1953) Percentage Percentage of Employment % of Wages % % of and S.S. Produc- Value Manufac turing Sector toabi shmen Total '0R ai d N on Paid payments i n Ad ded Food (6.8) (9.1) (9.9) (8.8) (6.1) (9.6) (6.8) Beverages (.7) (.8) (1.2) (.7) (1.5) (.8) (.9) Tobacco (2.2) (3.0) (4.2) (2.4) (2.6) (1.7) (1.9) Food, Beverages, Tobacco 9.7 12.9 15.3 11.9 10.0 12.1 9.6 Textiles (4.9) (6.3) (2.8) (7.7) (1.8) (4.2) (3.4) Clothing / Footwear (50.6) (44.0) (36.0) (47.0) (37.3) (50.9) (45.4) Leather (2.3) (2.2) (2.1) (2.2) (2.1) (2.4) (2.2) Textiles, Cloth./Ft., Leather 57.8 52.5 40.9 56.9 41.2 57.5 51.0 wood (5.6) (5.5) (5.8) (5.3) (4.9) (4.5) (5.5) Furniture (8.3) (7.8) (9.0) (7.4) (12.4) (7.2) (9.2) Wood, Furniture 13.9 13.3 14.8 12.7 17.3 11.7 14.7 Paper (- (-) (-) (-) (-) (-) () Printing (.7) (1.0) (2.4) (.5) (3.4) (1.2 (1.9) Paper, Printing .7 1.0 2.4 .5 3.4 1.2 1.9 Rubber (.2) (.2) (.1) (.2) (-) (.1) (.1) Chemicals (1.3) (1.6) (1.9) (1.5) (1.0) (1.7) (1.5) Oil & Coal()()(-(-()()() Rubber, Chemicals, Oil & Coal 1.5 1.8 2.0 1.7 1.4 1.8 1.6 Non Metallic Minerals 4.0 6.1 9.7 4.7 8.5 2.8 4.1 Basic Metals .1 .2 .3 .1 .6 .2 .4 Metal Products (3.7) (3.6) (4.7) (3.2) (6.0) (3.3) (4.5) Non Electrical Machinery (.2) (.2) (.3) (.2) (.3) (.2) (.3) Electrical Machinery (1.8) (1.6) (1.7) (1.6) (1.8) (1.8) (2.4) Transport Equipment (2.5) (3.6) (6.0) (2.6) (6.2) (3.3) (4.2) Metal Prod., N.E.M., E.M T.E 8.2 9.0 12.7 7.6 14.3 8.6 11.4 Other 3.9 3.3 2.0 3.8 . 3.3 3.9 5.3 Total (100.0) (100.0) (100.0) (100.0) (100.0) (100.0) (100.0) Number 36,110 64,560 18,100 46,460 17,544 228,611 104,601 /1 Defined as-5 workers and '_$24 000 in 1953. Source: DANE, BME 72, March 1957, p.15-17. 'fable III.84: DISTRIBUTION OF ESTABLISHMENTS, EMPLOYMENT, WAGES, RAW MATERIALS AND PRODUCTION BY MANUFACTURING SECTOR FOR VERY SMALL FIRMS, 1970 Employment Wages Raw Number of Wages S.S.Payments Materials Production Manufacturing Sector Establishments Total Paid Non Paid (( 1970) ($ 1970) ($ 1970) ($ 1970) Food, Beverages, Tobacco 3,901 10,951 5,279 5,672 30,512 1,188 374,356 557,319 Textiles, Clothing and Footwear, Leather 10,415 20,095 6,483 13,612 47,214 1,384 294,811 563,431 Wood , Furniture 4,876 10,309 4,006 6,303 35,722 876 141,147 313,112 Paper. Printing 733 1,969 1,048 921 9,321 403 23,494 59,453 Rubber, Chemicals, Oil & Coal, Plastics 326 877 478 399 3,761 162 27,985 50,440 Non Metallic Minerals 848 2,318 1,335 983 11,673 303 28,905 70,405 1 Basic Metals 126 375 208 167 1,805 110 6,016 13,000 Metal Products, Non Electrical Machinery, Electrical Ilachinery, Trarsport Equipment 2,423 5,888 2,868 3,020 26,432 779 120,759 241,705 Other 481 1,191 505 686 3,586 110 18,352 36,266 Total 24,129 53,973 22,210 31,583 170,055 5,316 1,035,826 1,905,132 /1 From sample of 3,925 establishments having less than five workers. Source: Unpublished DANE data. Table III.8b: PERCENTAGE DISTRIBUTION OF ESTABLTSIMENTS, EMPLOYMENT, WAGES, RAW MATERIALS AND PRODUCTION BY MANUFACTURING SECTOR FOR VERY SMALL FIRMS, 1970 /1 Percentage Percentage of Employment Percentage of Wages of % Raw % Manufacturing Sector Establishments TotalPaid Non Paid Wages S.S.Payments Materials Production Food, Beverages, Tobacco 16.2 20.3 23.8 18.0 17.9 22.3 36.1 29.3 Textiles, Clothing and Footwear, Leather 43.2 37.2 29.2 43,1 27.8 26.0 28.5 29.6 1ood, Furniture 20.2 19.1 18.0 20.0 21.0 16.5 13.6 16.4 Paper, Printing 3.0 3.6 4.7 2.9 5.5 7.6 2.3 3.1 Rubber, Chemicals, Oil & Coal, Plastics 1.4 1.6 2.2 1.3 2.2 3.0 2.7 2.6 Non Metallic Minerals 3.5 4.3 6.0 3.1 6.9 5.7 2.8 3.7 Basic Metals .5 .7 .9 .5 1.1 2.1 .6 .7 Metal Products, Non Electrical Machinery, Electrical Machinery, Transport Equipment 10.0 10.9 12.9 9.6 15.5 14.7 11.7 12.7 Other 2.0 2.2 2.3 2.2 2.1 2.1 1.8 1.9 Total 100.0 100.0 100,0 100.0 100.0 100.0 100.0 100.0 /2 Number 24,129 53,973 22,210 31,583 170,055 5,316 1,035,826 1,905,132 /1 From sample of 3925 establishments of less than 5 workers in 1970. /Z2 Wages, Materials and Production in thousands of $ 1970. Source: Unpublished DANE data. - 30 - Some detailed information on very small establishments is 1/ available in samples taken in 1953 and 1970. (See Tables 111-7 and I-I-8). Three changes in the sectoral distribution of establishments and employment between 1953 and 1970 are suggested by comparison of the two samples: First, food, beverages and tobacco increased their share of establishments (9.8 to 16.2%) and employment (12.9 to 20.3%); second, textiles, clothing, footwear and leather decreased their share of estab- lishments (57.8 to 43.2%) and employment (52.5 to 37.2%); and finally, wood and furniture augmented their participation in establishments (13.9 2/ to 20.2%) and employment (13.3 to 19.1%). 1/ According to the 1973 figures, the blown-up results from the sample of 1970 would have to be more than twice as large as documented in unpub- lished DANE data. 2/ Since each sample misses most of cottage shop employment, these com- parisons are only suggestive with respect to how the composition of cottage shop may have changed between these two years. For 1953 there is a fairly close relationship between our estimates of total cottage shop employment composition by industry (Table 111-6) and the composition of the DANE sample (Table 111-7), though they differ sharply on the importance of food, possibly due to an error in our estimate of total cottage shop employment. In 1953, the DANE sample covered only about one quarter of all cottage-shop production. One would assume that it would tend to catch the larger of these establishments and miss mainly household enterprise, though the share of paid employment in total employment in the sample was only about 20%. The 1970 sample covered less than 15% of total cottage shop for that year, as nearly as can be guessed. It cannot be assumed that it gives a good picture of employment composition of the sector, but the comparison with the 1953 sample may at least be valid in a certain subset of all cottage-shop, presumably the more visible part of it; whether these samples excluded household industry is not clear. It is worth noting that the trends just cited appear to charac- terize cottage shop employment as a whole in the cases of food, beverages, and tobacco (a sharp increase) and textiles, clothing, footwear, and leather (a marked decrease) but not in the case of wood products and wooden furniture, where the overall figures suggest a decreasing share (Table A-20). - 31 - The organization of production in very small firms may have shifted towards a lower participation of unpaid family members. Whereas 72% of employment in the firms captured by the 1953 sample were unpaid family workers, only 58.5% of those reported in the 1970 sample were in l/ a similar condition. Textiles, clothing, footwear and leather together still account for the largest share of unpaid workers, both with respect to the total employment in those sectors and with respect to the total number of unpaid workers, although these percentages have declined from 1953 to 1970. Due to substantial changes in monetary and relative prices over this period, it is not possible to compare trends in productivity between these very small firms and the factory sector. At the sectoral level, however, the following patterns can be established: whereas in 1953 productivity per worker (gross output per worker) in food, beverages and tobacco was slightly below the average for the total of small firms, by 1970 this pattern had changed and productivity in this sector was 44% higher than average. On the other hand, gro,s output per worker in 1/ Since the 1970 survey included a smaller share of total cottage shop employment (at most 15%) than did the 1953 survey (about 25%), this comparison may be misleading. If the 1970 sample penetrated less into the smaller establishments, its unpaid/total employment ratio would tend to be lower. This is suggested also by the relatively high average wage paid in the 1970 sample establishments. In the 1953 sample average wage was 63% of that in plants of less than 10 workers but falling in DANE's "factory" category (i,e., at least 5 workers or 24,000 pesos gross production). In 1970 the comparable ratio was 80% even though coverage was less complete (and therefore probably higher wage than otherwise) among these small "factory" establishments (Table A-6). -32 - textiles, clothing, footwear and leather was 9.5% above average in 1953 but by 1970 it was about 20% below the average for all very small establishments. Presumably these trends in productivity are responsible at least in part for the sectoral shifts in employment and establishments observed earlier. 1/ Recorded productivity per man in these very small firms is only about 22% of that prevailing in the factory sector as a whole. This differential is about the same as that observed in 1953, although it is interesting to note that as compared with a specific size of plant (e.g., those of 25-49 workers), the gap has narrowed; in 1953 the very small establishments reporting in the sample had a labor productivity 23% of that found in plants of 25-49 workers; in 1970 the corresponding ratio 2/ was 40%. Part of this increase may be illusory (due to possible non- comparability of the two samples (see footnote 1, p. 30), but it could not be fully explained in this way. Together with the increase in the relative wage of those very small establishments over 1953-70, it suggests that the pressure of rising real wages and better job opportunities may have helped induce higher labor productivity by these establishments. Part of the increases in average wages and labor productivity for this group is the result of the changing composition by industry, which was alluded to 1/ Gross output divided by total number of employees. 2/ This is consistent with an unchanged ratio of labor productivity of these very small establishments to that o'f factory manufacturing as a whole given that average plant size of factories rose substantially over the period so that the size 25-49 workers had nearly average labor productivity in 1953 but fell below average labor productivity in 1970. - 33 - earlier; clothing and footwear has become substantially less important, and higher wage industries more so. As just noted, average labor productivity of the very small establishments in current pesos was about 22% of the average for factory manufacturing in both 1953 and 1970. Between those two years, labor productivity of the latter rose by about 75% (based on the annual average increase calculated for 1949-51 to 1967-70, and presented in A. Berry, '"A Descriptive History ..." op. cit., Table 1). So it is clear that real labor productivity must have risen substantially for these sub-factory establishments as well, even if some of the estimated increase for them is a statistical illusion (due to possible non comparability of the two surveys, as noted above) and some due to relative price increases for the products of such establishments. Part is undoubtedly due to changing output composition among these establishments (e.g., away from clothing/ footwear/textiles and toward food processing and other industries). How much is due to productivity improvements within specific industries is difficult to judge; it is clearly a matter of great interest. Earnings per paiu worker in very small firms appear to be substantially lower than in the factory sector. In 1953 they were about 41% o average for the factory sector, whereas in 1970 they were down to 30%. The increase in average real earnings (deflating by the national blue collar cost of living series) was 58.5% for these very small establish- ments as contrasted with 216% for the factory subsector. Differences in skills, education, technology and complementary factors of production account for part of these differences in earnings (and in labor productivity). -34 - Much of the wage differentials are due to different labor force composition, by educational level, age, experience, etc. (More extensive reference to this issue is made in Chapter V below). Very small firms satisfy primarily local demands and are less concentrated in metropolitan areas than factories. Whereas 44.7% of the establishments and 42.7% of the employment in very small firms were located outside the main metropolitan areas, only 22.2% of establishments and 18.3% of employment in the factory sector were similarly located in 1970. Very small plants in non-metropolitan areas were especially concentrated in the traditional consumer goods fields of food, beverages, tobacco, textiles, clothing, leather and wood/furniture; these industries accounted for 83.5% of the employment (Table 111-9) whereas in metropolitan areas the corres- ponding share was 76.6%. Plants in non-metropolitan areas make slightly less use of paid labor also--39% as contrasted with 42.7% in metropolitan areas. Considering the lower earnings and other presumed disadvantages of household industry and very small plants as contrasted with factory manufacturing, one may ask whether the maintenance of high employment in the former sectors has been mainly a safety valve phenomenon, the result of poor employment prospectives elsewhere and consistent with stagnant or declining real earnings, or mainly a reflection of rising productivity and ability to compete? Over the recent period 1953-64 Urrutia-Villalba have estimated that average cottage-shop (this includes both household industries and very small firms, i.e., all operations involving less than 5 workers) incomes rose by 24%, while real wa2es in factory industry rose - 35 - Table 111.9: CHARACTERISTICS OF VERY SMALL ESTABLISHMENTS IN NON-METROPOLITAN AREAS, 1970 /1 Number of Employment Production Establish- Economic Activity ments /2 Total Paid ($ 000's) Food, Beverages, Tobacco 1,830 5,424 2,839 273,289 Textiles, Clothing, Leather 5,195 9,285 2,520 268,558 Wood, Furniture 2,221 4,533 1,658 130,580 Paper, Printing 148 448 233 15,741 Chemicals, Oil & Coal Pro- ducts, Rubber, Plastics 53 128 53 22,573 Non Metallic Mineral Products 419 1,200 730 37,194 Basic Metals 17 50 26 736 Metal Products, Machinery Equipment 795 1,705 807 61,709 Other 116 265 120 8,622 Total 10,794 23,038 8,986 819,002 /1 Non-Metropolitan refers to areas outside the cities which had at least 200,000 inhabitants in 1973, i.e., the metropolitan areas of Bogota, D.E. and Soacha, Medellin and Valle de Aburra, Barranquilla and Soledad, Cartagena, Manizales, Pereira and Santa Rosa, Bucaramanga, Giron and Floridablanca, Cali and Yumbo. /2 Defined as having between 1 and 4 workers. Source: Unpublished DANE data. -36- by 66%. In some departments where the factory subsector grew least the figures indicate income decreases for cottage-shop workers, as in Caldas, Cauca, Tolima, and Norte de Santander. On a cross-departmental basis - there appeared, however, to be little relationship between the percent increase in real factory wage and real artisan income. Over the longer period 1953-70, the comparison between DANE's surveys of very small establishments in those years implies a real wage increase of nearly 60%, as indicated above. Though the sets of workers involved are not the same as those included in the Urrutia-Villaba calculations (but rather a subset of the latter), the two wage increase estimates are consistent with each 2/ other. Urrutia-Villalba found that both income levels varied widely by sector; the lowest incomes were those of women working in confecciones (clothing) and the highest those of mechanics in automobile repair shops. Clothing and furniture used primitive techniques while food and auto- mechanic were more advanced. In DANE's 1970 survey, the food/beverage/ tobacco workers earned the lowest wages, those in textiles/clothing/footwear the next lowest, while the rest achieved wages near the level of the firms of 5-9 workers reported by DANE in that year. The evidence just cited is consistent with the plausible, but at this point unprovable hypothesis that average cottage shop earnings have trended upwards over much of the century. Changing composition of employment from lower to higher productivity activities and from rural to 1/ Urrutia and Villalba, "El Sectoral Artesenel ..." op. cit. 2/ They might suggest a somewhat faster increase during 1964-70 than during 1953-64 though this would be speculative. - 37 - urban would be expected to exert an upward push on average earnings. Evidence on labor mobility, especially in recent decades, would also suggest that when other laboring groups are gaining, it is unlikely that this one would be losing. Small scale industry (defined for the moment as plants of 5-24 workers) and ends to be important in the same industries as cottage-shop production, though with some interesting differences. In 1964 the major employers in this plant size range were food (19 thousand), clothing and footwear (14 thousand), nonmetallic minerals (7.7), metal products, excluding machinery (5.4). Wooden furniture, where cottage-shop production is so important, has relatively.few small factories (only 3.0 thousand workers in the size range cited), and transport equipment relatively few 1/ (5.1 thousand). In these subsectors the growth process has involved increasing numbers of cottage shop producers rather than the growth of cottage-shop establishments into small factories; economies of scale are apparently unimportant. In the cases of the other two major cottage-shop establishments into small factories; economies of scale are apparently unimportant. In the cases of the other two major cottage-shop subsectors, clothing/footwear and food, small scale industry is fairly and very important, respectively. The gradual decrease of cottage-shop employment in textiles may have been associated either with a shift of workers to larger plants or with the growth of cottage-shop firms into small factories, though the former seems more likely, if the lack of contemporary small plants is at all indicative. 1/ See A. Bery, "Relevance and Prospects..." p. 18. These figures are adjusted upward from those of DANE to allow for underreporting in this size range. - 38 - 2. COMMERCE The available statistical data on commerce and services is considerably more restricted than that on manufacturing. Substantial differences in the concepts and coverage of census and surveys over time make it difficult to carry out intertemporal comparisons, or analysis of trends and changes in the size distribution of firms. For this reason, the analysis will be mainly limited to the 1970 census and surveys of Commerce and Service Establishments. In 1970 there were about 500-510 thousand persons occupied in commerce. Probably about 165-175 thousand were in establishments of five or more workers, judging from the estimate that 161 thousand were so 2/ located in 1967. (Table A-11) A sample of very small establishments (1-4 employees) accounted for 146,762 establishments and 249,288 employees. (Table III-10). In other words, about 83% of the total occupied population are accounted for in one of these groups. The rest is probably 'informal' or casual employment. In contrast to manufacturing, and parallel to services, the bulk of the employment is in small firms: two-thirds in estaolishment of less than five employees, and only about 15% in establishments of at least 20 1/ See A. Berry, "Urban Labour Surplus and the Commerce Sector: Colombia," Yale Economic Growth Center Discussion Paper #178, June 1973, p.10. 2/ A 1970 commerce census of firms of at least 5 employees registered 6,792 establishments and 125,370 employees. (See Table A-12). This figure is implausibly low unless the 1967 census' figures were severely upward biased. It is possible that the structure of commerce moved in the direction of smaller establishments during these years (see Table A-F) but it could not have done so abruptly enough to make the 1970 figure plausible. - 39 - employees. Moreover, a larger fraction of the employment in these two sectors is in the self-employed category. For instance, according to a national household survey undertaken in 1971, 37.5% of the occupied population in manufacturing was self-employed, whereas the corresponding l/ percentage for commerce, restaurants and hotels was 60.3%. Data were not available for private sector services, but the ratio would be high there too. Consequently, registered enterprises in these sectors will cover a smaller fraction of employment than in manufacturing industries. Most of the very small firms are family-operations: owners and unpaid family workers represented 78.5% of all employment in the establishments of less than five employees reported in the 1970 survey, and presumably an even higher share in the universe of all such small establishments. Wholesale firms account for only about 40% of establishments (based on 1967 estimates--see Table A-11) and for less than 15% of the employment in commercial firms. Most of the employment in this type of commercial activity is in large establishments; in 1967, 47% was in establishments of more than 20 employees, which contrasts markedly with retail trade. Commercial firms of less than 5 employees are dominated by retail trade, especially by groceries stores which accounted for 64.6% of the small establishments reported in the 1970 survey and 57.7% of the employ- ment in those firms. These small grocery stores are mostly family-owned (only 10% of the employees are wage-earners); they satisfy local needs and 1/ DANE, Encuesta Nacional de Hogares, Etapas 3, 4 y 5, Fuerza de Trabajo, Bogota, DANE, 1976, p. 313. -40- are more widely scattered across Colombia than are manufacturing establishments. Whereas the main four departments (Bogota DE, Antioquia, Valle and Atlantico) had 56.1% of the small manufacturing establishments (1-4 employees) reported in the 1970 plants survey and 57% of the asso- ciated employment, only 50.2% of the small commercial establishments and 1/ 51% of their employment were concentrated on these departments. (See Table A-12). Commercial establishments of at least 5 employees are more concentrated in these departments than their smaller counterparts. Establishments engaged in selling clothing and footwear articles accounted for 9.2% of all small commercial establishments reported in 1971 and for 10.7% of employment. Other important trading activities in small firms are general merchandise, and pharmaceutical products. Small commercial establishments tend to be of recent origin: 35% of establishments operating in 1971 were created within the previous two years, and 65% had been in operation for less than five years. It can be presumed that the survival rate for these enterprises is fairly low, since the net increase in establishments every year is considerably smaller 2/ than the newly created establishments. 3. SERVICES Whereas employment in services amounted to well over a million persons in 1973, most of it was outside the realm of registered establish- ments; ie., in domestic services, professional services (e.g., doctors), public administration and defense, and other categories. A 1970 Census 1/ In 1973 these departments represented 41.6% of the population. 2/ Although once again it must be noted that the figures overstate creation of new establishments. See the discussion on page - 41 - of 'large' service establishments (of 5 or more employees) registered 57,754 employees in 4,609 firms (Table A-13) and a 1970 sample of 'very small' firms (of less than 5 workers) estimated 114,058 employees and 55,639 establishments. (Table A-14). Whereas there was presumably some underenumeration of establishments, especially the small ones, it is clear that the employment in registered firms accounts for only a fraction of the total employment in the sector. Further, a high 31% of the total employment in establishments of at least five workers takes place in those that have between 5 and 10 employees and 57.6% in those with less than 20 employees. The largest establishments (100 or more workers), account for only 13.1% of the employment in this group, whereas in manufacturing, firms of similar size account for over 50% of the total employment in the 2/ 'factory' sector. Restaurants and coffee shops account for substantial shares in services provided by firms of at least five workers, according to the 1970 survey: 54.6% of establishments listed in the 1970 survey, 46.9% of 3/ 4/ employment and 40.7% of sales. Automobile repairs represent 19% of 1/ Some service establishments were excluded altogether from the 1970 Census and Sample; for instance, barbershops, beauty parlors, parking lots and others. 2/ Includes plants that have at least 5 workers. 3/ The corresponding percentages with respect to total services will be lower on account of the exclusion of some sectors from the Census and Surveys (see footnote 1). This also applies to other percentages described below. 4/ Within this sector 63.6% of the enterprises have less than 10 employees and 92% have less than 20 employees. - 42 - the establishments, 14.4% of employment and 13.8% of sales, in services firms of at least 5 workers. Another important subsector is hotels and motels, with 7.2% of establishments, 14.4% of employment, and 14.3% of sales in 'large' services' establishments. A large number of services which are notoriously unimportant in firms of more than 5 employees are carried out in smaller firms. For instance, shoe repairs account for an important percentage of the very small establishments in services. A large part of the service establish- ments of 1-4 employees are restaurants and coffee shops, and account for 1/ 52% of thc employment in service enterprises of this size. Automobile and motorcycle repair shops account for 14.8% of service employment in very small establishments. Other activities whose employment is well represented in service establishments of 1-4 workers are the following: hotels and lodging (7.9%), electrical and mechanical repair shops (6.9%), shoe repair shops (4.8%) and laundry services (4.75). The pattern of concentration in the.main four departments (Bogota, Antioquia, Valle and Atlantico) is also evident in services: about 60% of the employment in service establishments of 1-4 workers and about 72% of the employment in enterprises of at least five employees is located in these departments (Table A-15). The age distribution of enterprises of 1-4 workers presents important similarities with those of manufacturing and commerce. About 2/ 38.5% were started between 1970 and 1971. 1/ This share is actually overestimated due to the exclusion of certain types of service establishments from the 1970 sample. See footnote 1 at the beginning of section 111.3. 2/ The survey was carried out in 1971. - 43 - Appendix: Data Sources on the Manufacturing Sector Estimates of total employment in manufacturing are obtained from population census (1951, 1964 and 1973) and household surveys (e.g. 1971 and more recent years). These estimates comprise both formal 1/ and informal employment. Formal employment is intended to refer to tfactorv" activities whereas informal activities cover mainly household industries and artisan-type production in very small establishments (less than five workers). The distinction between formal and informal, or factor-y and non-factory, is sometimes difficult to make, especially in the case of small firms which may be 'modern , in the sense that they employ relatively advanced production techziQuas. Mhenever possible, different definitions have been used in order to see whether the results were sub- stantially affected or not. The data sources for factory employment are: (1) DANE, Industial Censuses (1953, 1970); (2) DAKE, Annual Survev of Maufactures (up to 1975); and (3) ICSS, Social Security records 1971-1975. A cautionazry ante on the data is necessar; the Annual Survey of -Hanufactures undertakan by DANE has changed its ccvarage substantially over the yaars. For instance. ul to 1/ Subject to data railability, emlo2*int 'u:s refer to occupied persons and aot to theacooricll-activ pop:u1:ion' .ic inclu U nea=mplio V C. The reaso S or e d h n l is nha1 their classi,-ic-tion aogsectors mybeariry and mr-- re"'ziat th~ asi_rcAions of -> t,bor fc a an ze ac opor%tuite avaiLb,e in a ven sector. -44 1969 the cut-off size for exclusion of firms was 5 workers, whereas from 1971 onwards it was intended that only firms of ten or more employees 1/ would receive adequate coverage. Thus a comparison of survey data per- taining to years before and after 1970 would indicate a substantial drop in the number of small establishments of 5-9 employees, whereas in fact no real change may have taken place. Consequently, proper comparison can be made only if data for different years are equally inclusive. Moreover, the coverage of firms of less than fifty employees is less extensive than that given to larger firms, so that even considering only firms of more than ten employees, the DANE surveys would underestimate the number of smaller firms. A confirmation of this is given by comparing social security data with DANE records. Consequently, at least for the year 1975 it was concluded that social security data represent more accurately the formal or factory sector. Since most of the uncertainty as to inclusion refers to very low size categories, their exclusion may be a prerequisite for carrying out meaningful comparisons across years. The data collected by DANE refer to establishments and not to firms. In principle several establishments could be part of the same firm. In practice, however, the differences seem to be very small. Social security records, which register firms, show a considerable similarity with DANE data, except for small firms in which coverage, rather than discrepancy between firms and establishments, varies. 1/ Coincidentally, the coverage of social security records has increased substantially since 1971 and can supplement, and in some years substi- tute DANE employment data. - 45 - Non-factory or non-formal employment,is derived as the difference between total and 'factory' employment. Additional information on the non-factory sector is available from several sources. The 1973 Population Census provides data on household industries, including the distribution of employment according to manufacturing sectors (2 digits), and urban-rural breakdown. The industrial census of 1953 attempted and partially succeeding in registering artisan and household production (i.e., firms of 1-4 employees l/ and whose value of production did not exceed Col$24,000 in 1953-/). The 1970 Industrial Census provides information on a sample of firms of less than five employees, but without a constraint on the value of production. To a certain extent the last two sources are comparable and will be used to detect trends in the characteristics of very small firms. Due to the very nature of the informal sector, the statistical information available is limited and subject to larger margins of error than are factory data. For these reasons, a dis- proportionately large part of this paper will be devoted to the factory sector, even though it is recognized in advance that the employment absorption of the informal sector in Colombia is almost as important as that of the factory sector. A final note on the data is necessary. As could have been expected, the available statistical information was incomplete for many of our purposes. To the extent that it was considered justifiable to adjust the data for known biases, or interpolate values for missing information, such steps have been taken; explanations have been included so that independent researchers could reexamine these modifications. A corollary of this is that the data so adjusted are likely to entail larger measurement errors than the rest, and should be treated accordingly. 1/ Those firms that exceeded the indicated level or had more than five employees were classified as part of the 'factory'sector. -46 - IV. DYNAMIC CHANGES IN THE FACTORY MANUFACTURING SECTOR 1. Aggregate and Sectoral Trends The organization of this chapter is as follows. In the first section the focus is on trends in the factory sector since the 1950s, considering both aggregate trends for the factory sector taken as a whole and changes in the composition of the factory sector, by industry and by size of establishment. Section 2 analyzes the age-distribution of factory enterprises according to size. The third section, based partly on longitudinal data for a large sample of factories that operated between 1970 and 1975, studies the dynamic patterns of mobility of firms according to size, trying to assess the extent to which small firms are able to increase in size and become medium or large firms. These mobility patterns are studied both at the aggregate level and for specific industries. The fourth and final section studies survival and death rates for factories classified according to size. According to t.he industrial census of 1953 the smallest size category (5-9 employees) accounted for a substantial number of establishments (70.8%) in the factory setctor, but only employed about 21% of all factory employment, produced 10.7% of gross output and 9.7% of value added. On the other hand, firms of at least fifty employees represented only 5.4% of all factory establishments, but employed 53.3% and produced 61% of gross output and 71.1% of value added in the factory sector in 1953 (Table A-16). Reported figures are, unfortunately, likely to be substantially underestimated, more so for variables related to value of output and for smaller plants for whom underreporting is easier (they may have no organized accounts at all). - 47 - For plants of under 20 or 25 workers, it is probable that underreporting of 50% for value of output, value added, etc., is fairly frequent, and the ratio of reported to true values might be in the range of 50 - 70% at 1/ the bottom end of this size range. It is unclear how much underreporting characterizes the large plants, e.g., the sociedades anonimas. They frequently maintain separate accounts for internal and for tax purposes, so there is little doubt that output figures are understated here too, but probably not by more than 10 - 30%. It thus appears very likely that reported figures understate the output and labor productivity levels of small plants relative to large ones, with the differential being perhaps 20-30% from the smallest to the largest. Obviously, conclusions about the relative efficiency of smaller plants or their share in total output can be substantially affected by such misreporting. Employment in firms of 10 to 49 workers doubled or nearly so 2/ over the period 1953-75. However, the share of these plants in employ- ment of all plants with 10 or more workers has declined from about a third in 1953 to not much over 20% in 1975. The relative decline is even more pronounced in the case of the smaller firms of 10 to 19 workers, from about 19% in 1953 to perhaps 7% in 1975. The overall size of establishments of 5 or more workers increased from an average of about 21 in 1953 to around 40 in 1975. 1/ Based on discussions with small producers and knowledgeable observers. 2/ Comparisons including firms of 5 to 9 workers are less reliable due to differences in coverage in 1953 and 1975. -48- Similarly, although the value added generated in small factories has increased over time (in real terms), it has grown at a slower pace than in larger factories. With reference to factories of 10 or more employees, the share of value added in firms of less than fifty workers l/ declined from 21.3% in 1953 to approximately 10-11% in 1975. In summary, small factories have traditionally accounted for a high share of the number of establishments but a fairly low percentage of factory employment and value added. In relation to factories of at least 10 workers, small plants of less than 50 workers accounted for only a little over 20% of factory employment and 10-11% of factory value added in 1975. Furthermore, the shares of factory employment and value added in small plants have been declining significantly over time, even though both have increased in absolute terms. The trends in small factories may reflect on the one hand, a greater likelihood of extinction than for larger firms or the process of growth of firms: small firms being trans- formed into larger units. Patterns of survival of firms and upward mobility over time are the subject of Sections 3 and 4 below. A. Struc:ural Changes in Manufacturing: 1956-1975 Between 1956 and 1975 the structure of production in the factory sector has changes in several ways. The most important transformation 1/ For 1975 there is considerable underestimation of small firms in DANE data. For this reason, social security (ICSS) data have been used for adjusting upwards the relevant numbers: DANE's value added data have been expanded in the same proportion in which ICSS data exceeded DANE employment records. The adjusted share of value added for 1975 is presented above. -49 - concerns the changing composition of output: import-substitution has led to the expansion of intermediate and capital goods sectors, which were in an incipient stage in 1956. The number of establishments and the employment engaged in these productions (e.g., basic metals, metal products, machinery, transport equipment) have considerably expanded over this period. In general, between 1956 and 1975, establishments producing non- durable consumer goods have expanded at a much slower percentage rate than those engaged in consumer durable goods, and both have increased less rapidly than intermediate and capital goods industries. The percentage increase of employment by sectors has followed a similar pattern. (See Table IV-1). In non-durable consumer goods industries, employment has generally increased much faster than the number of establishments, leading to substan- tial increases in the average size of plants between 1956 and 1975. A partial exception is the textile industry, where a considerable increase in 2 '/ employment (perhaps 110%) has been accompanied by an equally significant increase in the number of establishments (84.5%), resulting in 'only a slight 3/ increase in the average size of plant (10-15%). Electrical machinery, basic 1/ The figures on numbers of establishments by industry which underly these statements and the figures presented in subsequent paragraphs are unadjusted DANE statistics. Since the 1975 data are quite incomplete with respect to smaller plants, the growth of number of plants is understated in relation to that of employment, so the growth of average plant size is overstated. 2/ The level of employment in textiles as of 1975 is open to some question. See the note with Table A. 3/ It should be noticed, however, that the size of plant in this sector is substantially above the average for the factory sector as a whole. - 50 - Table IV.1 COLOMBIA: SECTORAL INCREASES IN ESTABLISHME:NTS. EMPLOYMENT AND AVERAGE SIZE OF PLANT IN MANUFACTURING 1956 TO 1975 (Factory Establishments of at least five workers) 1956 Percentage increase between 1956-75 in Number of Average Number of Average Establish- Employ- Size of Establish- Employ- Size of mants ment Plant ments ment Planq (1) (2) (3) (4) (5) (6)'1 Food 1,631 32,030 19.6 11.3 101.8 81.6 Beverages 190 11,829 62.3 -28.4 74.3 143.3 Tobacco 221 5,362 24.3 -85.5 - 40.6 309.5 Textiles 318 35,989 113.2 84.5 95.4 5.8 Clothing/Footwear 1,345 26,616 19.8 35.2 131.4 71.2 good 275 5,431 19.7 32.4 40.4 5.1 Furniture 333 4,667 14.0 103.3 185.6 40.7 Paper 57 2,963 52.0 89.5 140.5 26.9 Printing 320 8,349 26.1 127.8 144.4 7.3 Leather 137 3,981 29.1 83.2 157.3 40.2 Rubber 45 4,773 106.1 137.8 74.0 -26.9 Chemicals 308 10,588 34.4 88.0 233.8 67.4 Oil and Coal 14 2,027 144.8 233.3 113.1 49.2 Non-Metallic Minerals 603 18,217 30.2 8.6 67.1 54.0 Basic Metals 53 4,902 92.5 437.7 230.2 -38.6 Metal Products 327 8,045 24.6 303.1 481.7 44.3 Non-Electrical Machinery 125 2,123 17.0 164.8 310.1 54.7 Electrical Machinery 119 2,960 24.9 468.1 620.4- 26.5 Transport Equipment 293 8,571 29.3 165.5 145.4 - 7.8 Other 144 3,544 24.6 322.9 546.8 52.8 Total 6,858 202,967 29.6 73.2 143.6 40.5 /1 The relationship among columns (4), (5) and (6) is the following: g6 = (100) ["')+g5 - 1] where g4' g5 and g6 represent respectively the data for a given sector 100+g4 in-columns (4), (5), and (6). Source: DANE, Annual Survey of Manufactures, 1956 and ICSS, Social Security Records, 1975. - 51 - metals and metal products are the sectors which expanded fastest in terms of aiumber of establishments and also in terms of employment, leading to not too-large increases in the average size of plant over the period 1956-75. B. Distribution of the Net Increase in Employment and Establishments Most of the net increase in factory employment between 1956 and 1975 has been located in large firms: 54% in firms of more than 200 1/ employees and 15.5% in firms of between 100 and 199 workers. (Table IV-2). Although most of the net increase in establishments between 1956 and 1975 has been in the small size range, these plants have not been -ble to absorb an important fraction of the overall employment increase. However, only 5-7% of the net employment increase in the factory sector 2/ during this period was located in firms of 5 to 19 workers and only 13.6% 3/ in enterprises of 20 to 49 workers. The process of import substitution in some industries has been highlighted by considerable employment expansion with respect to the 1956 1/ This ignores the changes in the classification of firms: e.g., small firms become large. See Section 3 below. 2/ The figures of Table IV-2 indicate 6%, but they are, of course, less precise than those referring to the larger plants. 3/ There has also been considerable concentration of factory manufacturing employment in Bogota during this period, its share i(with the rest of Cundinamarca) rising from 24% in 1953 to 34.4% in 1975 (probably a slight exaggeration due to changes in coverage of the data over this period). Departments like Antioquia and Valle maintained their shares while nearly all others lost. (Table A-16.5) Table IV-2: CHANGES IN EMPLOYMENT STRUCTURE BY PLANT SIZE, 1953 and 1956 to 1975 Plant Size Best Estimate ICSS (Number of Workers) 1953 1956 1964 1975 Data 5- 9 40,932 40,200 40,000-50,000 39,700 (29,667) 10- 14 15,069 16,114 25,000 24,500 (22,458) 15- 19 10,224 8,100 19,700 (18,456) 16,187 20- 24 7,507 7,606 18,500 (74,752) 25- 49 21,624 25,838 29,602 54,000 50- 74 15,297 18,908 35,360 106,188 (62,552) 75- 99 10,511 14,213 24,405 n 100-199 26,091 35,957 71,715 (73,142) 200 73,218 115,214 229,901 (213,440) Total 200.0 225.0 294.6-304.6 318.3 Sources and Methodology: The estimates for 1953 are based on the industrial census of that year (See Table A-16). An upward adjustment of 2%, 10%, and 30% was effected to the census figures for the categories 15-24 workers, 10-24 workers, and 5-9 workers respectively. (The last group's employment was estimated by the author, since the census lumped a number of establishments of 1-4 workers into this -category). - 52 - level. The most important sector in terms of its share in the net employment increase between 1956 and 1975 is metal products: out of every (net) additional one hundred factory jobs created between 1956 and 1975, 13.3 jobs were in this sector. Similarly in the case of establishments: 19.7% of the net increase in enterprises during this period was in metal- products firms. Of the net increase in factory establishments and employ- ment, respectively, between 1956 and 1975, metal products (19.7% and 13.3%),.together with electrical machinery (11.1% and 6.3%), transport equip- ment (9.7% and 4.3%), "Other' (9.3% and 6.6%), basic metals (4.6% and 3.9%) and non-electrical machinery (4.1% and 2.3%) account for totals of 58.5% and 36.7%. The employment increase in these sectors has not been concentrated primarily in the largest establishments: plants of 200 or more employees accounted for just 12.1% of the total net increase in factory employment, equivalent to about a third of the total increase in employment in these sectors. The remaining shares with respect to the net 1/ employment increase in this metal and metal products industries are distributed as follows: firms of 5-19 workers, 6.1%; those of 20-49 workers, 6.7%; enterprises of 50-99 employees, 5.6%; and finally, firms of 100-199 2/ workers, 6.1%. Some industries have followed quite different patterns. For instance, some non-durable consumer-goods sectors, such as beverages and tobacco have been characterized by a decline in the number of establishments, 1/ All involve metal working except part of "Other." 2/ Percent figures presented in this and the next paragraphs are based on Table IV-1 and are not quite comparable with those of Table 111-2. - 53 - and employment has fallen, at least in small plants. Textiles have contributed importantly to the net increase in employment over this period (11.8%, but most of this increase has been concentrated in the largest firms (200 or more workers), which accounted for 8.7% of the factory. Clothing and footwear are perhaps the only non-durable consumer goods sectors in which a considerable expansion of employment has been widely distributed among firms of different sizes. Out of every 100 additional factory jobs created between 1956 and 1975 12 were in these industries, distributed as follows: size 5-19 workers (.8), size 20-49 (2.0), size 50-99 (1.4), size 100-199 (1.8) and finally size 200+ (6.0). The structural changes that took place between 1956 and 1975 can be decomposed into two periods; the year 1967 signaling the benchmark between a period of heavy concentration on import-substitution industriali- zation (1956-1967), and a subsequent period of broader export-promotion 1/ efforts (1967 till present). During the latter period the annual growth rate of factory employment (5.9%) and value added (7.1%) were higher than in the previous period (3.3% and 5.6%). This acceleration of growth may be due to the changing industrial policy, though it has been influenced also by a faster overall rate of economic growth (which in turn may have been fostered in part by the export oriented strategy). 1/ Exports of manufacturers increased significantly over the period; nevertheless, they accounted for less than 10% of industrial production in 1973. - 54 - The previous results present static comparisons between firms in 1956 and 1975. Dynamic issues, for instance, the extent to which presently 'large' firms have their origin in small firms, are the subject of Section 3 below. 2. Age-distribution of factory establishments In general, larger firms tend to be older than smaller ones, at 1/ least in the legal sense of the term, 44% of firms (plants?)' that had 200 or more employees in 1970 started operations in or before 1950, and another 32% between 1951 and 1960. Thus 76% of these firms were at least 10 years old in 1970, whereas less than 10% were younger than 5 years. At least 50% of.the firms that in 1970 employed 50 or more employees had been started in or before 1960, the percentage being higher for the categories representing larger firms. In the case of smaller firms (less than fifty employees), the percentage of 'old' firms in 1970 also decreased with size; only about 23% of the firms in the 1-4 category were listed as having been founded on or before 1960, whereas more than 53% reported being founded in the three years just prior to 1970. (Table IV-3). In 1970 the reported median age for all factory establishments was approximately seven years, indicating that about half of them had been created after 1963. The median age of establishments of 200 or more employees was 17-18 years, i.e., approximately half of these large firms 1/ Unfortunately, economic age may differ from legal age. When firms are sold, moved from one place to another, or change juridical form they may be treated as having disappeared and been replaced by a new firm. It will be necessary to study in depth how age was defined in these data before conclusions can be firmed up. Table IV. 3, COLOMBIA: AGE DISTRIBUTION OF FACTORY ESTABLISHMENTS BY SIZE, 1970 Tniriali Ypar nf Operatfço1 Establishuent Not $ize (1Nunber 1951- 1957- 1961- 1968- Recorded Row of Workers) (1950 1956 1960 1965 1966 1967 1969 1970 or 1971 /1 Total 1-- 4 2] 20 19 28 11 18 63 75 3 258 5- 9 246 213 326 484 142 173 564 356 40 2,544 10-14 126 107 103 243 66 77 233 141 15 1,111 15-19 93 63 64 137 39 50 116 67 18 647 20-24 49 55 61 92 32 33 86 35 8 451 25-49 155 123 160 264 69 68 172 87 27 1,125 50-74 78 61 69 92 25 21 55 28 10 439 75-99 41 38 46 51 8 7 20 8 5 224 100-199 89 55 63 76 12 9 27 7 14 352 200+ 139 60 41 44 7 4 i 4 6 316 Column11 Total 1,037 795 952 1,511 411 460 1,347 808 146 7,467 /1 The census refers to 1970 but was carried out in 1971; some firms have declared 1971 as the year they started operations. Source: Unpublished DAUE data, 1970 Census of "large" manufacturing establishments, - 56 - had been started before 1953. In sharp contrast, small establishments are of more recent origin: firms of 25-49 workers had a median age of 7 years, and those of 5-9 employees only 4-5 years. Age of establishment data are particularly suspect for the smallest firms, however, and it is virtually certain that average age is 1/ substantially underreported. Still, it appears probable that average age is not high and that turnover is substantial. For a variety of reasons, including a fairly acute process of competition, it is possible that only a limited number of small firms 2/ remain in operation for long periods of time. 3. Mobility patterns in factory establishments A. Aggregate Patterns An interesting aspect of industrial.development is the evolution patterns [(growth, decline, change of output composition, etc.)] of individual firms over time and their relationship to the dynamics of the manufacturing sector as a whole. Does most employment and output growth 1/ In most developing countries, small scale establishments rent their production site, partly to leave themselves less vulnerable to taxation. Being renters increases their likelihood of moving in the first place, and the continuing preference to avoid taxes and bureaucratic constraints provides an additional incentive to be mobile. 2/ If DANE data on creation of firms were accurate and related to " economic" creation, then their median life would be quite short, as can be seen from a comparison of the pattern of (reported) creation of firms in 1970 with that of surviving firms which were born in the immediately preceding two years: whereas 53.4% of firms created in 1970 were very small (1-9 workers), only 46.5% of the surviving firms of 1968-69 were in that category; it would not be possible to impute such a difference to firms that outgrew their initial size category. - 57 - occur in existing firms or new firms, in large ones or small ones? How do such patterns depend on the ratio of overall expansion of the sector and other possible determinants? Some evidence is available with respect to the 1960s and the 1970s. For the period 1962-66, Todd compared employment size category of plants in the two years, finding that of those identified as existing in both years probably about the same number decreased number of employees l/ as increased it. (See Table IV-4) For the about 4,500 plants making up this group, total employment rose by 4,530, or 3.4% over the four years, i.e., less than 1% per year. If these plants were typical of all plants existing in both years, then less than half of the recorded growth in employment of 8.5% over the period could be attributed to expansion of plants existing in both years. But these plants may not have been typical 2/ in this regard; in any case it may be more interesting to know what share of employment expansion is due to new firms than to new plants. It is of interest to note that for plants below 25 workers there was virtually no tendency toward employment growth at all, whereas in plants of 25 workers or more in 1962, an average growth over the four years of 4.5% 1/ See John Todd, Efficiency and Plant Size in Colombian Manufacturing, Yale Ph.D. Dissertation, 1972, pp. 41-45. If a plant changed location, name, or any other item in its set of identifying variables it did not enter this set, which thus included less than half of all plants. While more of these plants moved to a lower size category than to a higher one, tne distribution of firms by size within most categories would suggest that similar numbers moved up as down when intra size category movements are allowed for. 2/ For example, fast growing ones might be forced to change locale more frequently than stable ones. - 58 - occurred. (Table IV-5). Output growth would in general be faster than employment growth, but the relationship between the rate for firms above 25 workers in 1962 and those below 25 would presumably be similar. A Superintendencia de Sociedades Anonimas study of manufacturing corporations operating over the full 1963-67 period suggests a rate of output growth of 1/ about 3% per year (12.9% over the period) if the appropriate price deflator for these corporations was that used in the national accounts for the sector as a whole. Overall manufacturing output rose at about 5% per year over these years, and that of the factory sector a little faster. Between 1970 and 1975 the total net increase in employment in 2/ establishments registered by DANE was 109,737 employees. Of these it has been possible to account for 60,133 new jobs created in 3,721 matched or paired firms (i.e., those that operated between 1970 and 1975 and were individually traced in both years). (Table IV-6a). The difference is not entirely attributable to new firms that came into being between 1970 and 1975. Many small plants that operated in 1970 could not be properly identified in 1975 even though they continued to operate, since DANE moved 1/ Suoerintendencia de Sociedades Anonimas, La Industria Manufacturera, 1969. The study presented data for 315 corporations, a of the 471 under the vigilance of the Superintendencia in 1963 and accounting for about 90% of the value added of all (national) corpo- rations in that year. (See Superintendencia de Sociedades Anonimas, Revista de la Superintendencia de Sociedades Anonimas, 1966, Bogota, 1966.) 2/ The employment increase would have been larger if DANE had covered smaller establishments in 1975 as extensively as in 1970. See below. A plausible guess would be that factory employment (i.e., plants of five or more workers) increased by 120-125 thousand woLkers, or about 5.5% per year. Table IV-4: PLANT SIZE IM 1966 AND IN 1962, FOR PLANTS IDENTIFIED AS EXISTING IN BOTIl YEARS Size in Size class in 1966 (Number of workers) Percent moving to a: 1962 (No. lower higher of workers) 5 5-9 10-14 15-19 20-24 25-49 50-74 75-99 100-199 >200 Total size class size class <5 923 32 2 0 1 0 0 0 0 0 957 0 3.6 5-9 S1 1301 14 3 '4 1 0 0 0 0 1374 3.7 1.6 10-14 9 25 633 11 10 6 0 1 1 0 696 4.9 4.2 15-19 2 32 151 69 50 24 4 1 1 0 334 55.4 24.0 20-24 1 11 55 36 52 45 7 1 0 0 208 49.5 25.5 25-49 2 10 22 29 56 248 60 12 6 1 446 27.7 17.7 50-74 0 0 2 1 1 48 76 29 16 0 173 30.1 26.0 ul 75-99 0 1 0 0 0 6 17 36 22 1 83 28.9 27.7 100-199 0 1 0 1 0 1 6 9 88 22 128 14.1 17.2 >200 1 0 0 0 0 0 0 0 8 88 97 9.3 0 Total 989 1412 879 150 174 379 170 89 142 112 4498 Source: John Todd, Efficiency ..., op. cit, p. 42, based on unpublisied DANE data. -60 - Table IV-5: CHANGE IN AVERAGE SIZE OF PLANTS IDENTIFIED IN 1962 AND 1966 (4496 PLANTS) Number of Percentage Change workers Average Plant Size (Relative to average of in 1962 1962 1966 1962 and 1966 values) <5 3.0 3.1 '1+3.3 5- 9 7.3 6.7 -8.0 10-14 11.6 12.1 +4.3 15-19 17.2 15.7 -2.9 20-24 21.6 21.7 + .5 25-49 35.1 36.7 +4.6 50-74 61.5 62.5 +3.3 75-99 84.8 89.2 +5.2 100-199 136.3 145.5 +6.7 >200 532.4 555.0 +4.4 Source: Todd, op. cit., p. 43. - 61 - Table IV-6a: CHANGES IN PLANT SIZE BETWEEN 1970 AND 1975 FOR A SAMPLE OF FACTORY ESTABLISHMENTS IDENTIFIED IN BOTH YEARS Plant size in 1970 (No. of Plant Size in 1975 (Number of Workers) workers) 5-24 25-49 50-99 100 Total 216 23 8 2 249 1656 178 62 18 1914 5-9 2573 781 566 306 4226 917 603 504 288 2312 1.52 1.00 .83 .47 3.84 1086 283 36 7 1412 16412 5202 664 130 22408 10-24 15837 9354 2327 783 28301 -575 4152 1663 653 5893 -.95 6.90 2.76 1.08 9.79 179 474 211 35 899 5405 15885 7997 1338 30625 25-49 3233 17066 13844 4713 38856 -2172 1181 5847 3375 8231 -3.61 1.96 9.72 5.61 13.68 21 88 288 158 555 1292 5514 19519 12005 38330 50-99 350 3443 20845 22638 47276 -942 -2071 1326 10633 8946 -1.56 -3.44 2.20 17.68 14.87 1 6 34 565 606 112 727 4105 194033 198977 >100 10 247 2707 230764 233728 -102 -480 -1398 36731 34751 -.16 -.79 -2.32 61.08 57.79 1503 874 577 767 3721 24877 27506 32347 207524 292254 Total 22003 30891 40289 259204 352387 -2874 3385 7942 51680 60133 -4.77 5.62 13.20 85.94 100.00% Note: The five items in each cell are as follows: 1. Number of establishments. 2. Initial employment (1970) 3. Final employment (1975) 4. -Increase in employment 5. Percent contribution to net employment increase. Source: Unpublished DANE data. For explanations, see notes to Table A-21. - 62 - toward excluding plants of less than 10 workers during this interval. In other words, the net increase in reported employment of 49,604 workers is the net result of the appearance of new plants, the demise of old ones, and the growth of firms which operated in 1970 but which it was not possible to identify at a later year (1975). Something over half of net employment growth evidently occurred in existing firms, possibly considerably more. While total factory employment was probably growing at about 5.5%, the plants which were identified as existing in both 1970 and 1975 were characterized by employ- ment expansion at an annual average rate of 3.8%. If these plants were typical of all plants which existed in 1970, and if none had disappeared over the five year interval, then such growth would have accounted for about 70% of all employment growth in manufacturing. All one can conclude with any confidence, though, is that a majority of employment growth occurred in existing plants. Among the plants identified in'both 1970 and 1975, percent growth of employment was on average faster the smaller the size in 1970 (Table IV-6b), being over 25% in those in the 10-49 range in 1970 and 17.5 for 2/ those already over 100 workers in 1970. In spite of higher growth rates, 1/ The firms that are purposefully excluded from these data are: (1) firms that existed in 1970 but ceased operating later and (2) new firms that came into being after 1970. Differences in the coverage of establish- ments also account for a substantial portion of those firms that were not 'matched' or paired in 1970 and 1975. E.g., in 1970, according to DANE, there were 2,542 establishments of 5-9 employees, but in 1975 DANE registered only 536, and only 249 were matched. Although a number of the 2,542 establishments of 5-9 workers in 1970 may have ceased operating by 1975, the underenumeration of firms of less than 10 employees in 1975 accounted for a large share of all unmatched firms between the two years. 2/ See next page. - 63 - the small firms identified in both years did not contribute significantly to the total net increase in employment; mainly on account of their relatively low share of employment in the factory sector. Out of every one hundred additional jobs created in firms which operated between 1970 and 1975, 27.33 were in firms of less than fifty workers in 1970, and only 13.64% in firms of less than 25 persons. Firms of 50-99 workers in 1970 accounted for 14.9% of the net employment increase. Most of the employment increase in established plants took place in firms which had at least one hundred employees in 1970, 57.8%. The percent of total employment growth over the period accounted for by these plants was, however, under 30%, and since it seems unlikely that many large plants which were in existence in both years would have moved or otherwise become unidentifiable, it would appear that large plants, so defined, accounted for considerably less than half of employment growth. Large firms might, however, accounted for a much higher share. One of the reasons for a substantial concentration of the net employment increase in large firms is the apparently smaller tendency to 1/ decline in the case of larger plants. 1/ Only 6.8% of the establishments which in 1970 had over a hundred workers declined to a lower size category. This might, however, be due to the open ended character of this top category. In firms of 50-99 workers in 1970, 19.6% had declined to lower size categories by 1975. About 19.9% of plants of 25-49 workers declined to a lower category. Footnote 2/ of previous page 2/ In enterprises of 5-9 workers, employment more than doubled over this five-year period. The rate of growth of employment in firms of 5-9 employees is clearly overestimated since DANE survey data in 1975 would have captured mainly those that had been successful. Those that remained in that size category are likely to have been omitted from the survey. Similarly, firms of 10-49 that declined their size to less than 10 workers would not be reflected in the data; their exclusion would thus inflate the net employment increase of small firms. -64 - Upward mobility exists mainly in plants of adjacent size categories. About 20.6% of the plants of at least one hundred workers in 1975 came from plants of 50-99 workers in 1970, but only 5.7% from establishments of less than fifty workers. Among firms of 50-99 employees in 1975, 36.6% had between 25-49 workers in 1970, but only 7.2% had less than 25 workers. Finally, 32.4% firms of 25-49 workers in 1975 had between 10-24 employees in 1970, and only 2.6% had between 5-9 employees. The size category with the highest mobility is that of 50-99 workers: about 43.8% of its plants in 1975 are newcomers from lower sizes. Unfortunately the relationship between growth and movement to a hig!1er size category depends on the width of a given category and the distribution of plants within it, so the statistics just discussed must be interpreted with caution. Further, firm size movements are more relevant for analizers of mobility than are plant size movements, and the former are not readily available at this time. It is worth noting that data in large Sociedades Anonimas (firms rather than plants) suggested output growth rates a little less than 60% of aggregate manufacturing growth rate over 1963-67. The output growth rate of 3% for these corporations would seem consistent with an employment growth rate of just over 1% in plants of above 100 workers. If the behavior of plant employment is indeed revealing (i.e., if it gives clues as to how firm employment is changing) then the contrast between the periods 1962-66 and 1970-75 is interesting. During a period of relatively slow total growth in manufacturing (1962-66 or 1963-67), relative employment growth of already established plants was faster for large ones than for small ones. (Table IV-7), Table IV-7: EVIDENCE OF GROWTH RATES OF SETS OF PIANTS OR FIUS IDENTIFIED OVER PERIODS OF TIME Annual Annual Employment Growth Annual Employment Growth Ratg Qf Sample Units Rate of Factory Sector Output Growth Period Sample 100 workers >100 Workers in DANE DANE Annual Output Rate of in initial initial year Adjusted Growth Rate Manufacturing All Year of Sample Sector Units Total Factory 1962-66 4498 0.85% 1.2% 0.45% 20% 3.1% .a, 5.5% 5.9% Plants 1963-67 315 n.a. 3.1% ~5.0% 5.6% Corporations La 1970-75 3721 3.8. 3.3% 4.5% 5.6% 5.6% 7.0% 7.5% Plants / Plants of 25 workers in 1970. Small ones were exclu4ed on the grounds that too many whose employment declined to below 15 or so would have been missed in 1975, thus upward biasing the growth rate. Sources: Tables IV-5, IV-6 and Superintendencia dF Socie4ades Anoutnios, La Industria Manufacturera, Bogota, 1969. - 66 - The growth of average labor productivity was almost 3% per year in the factory subsector. During a rapid growth period (1970-75) the smaller established plants had faster employment growth than the larger ones, overall factory employment grew faster and labor productivity in that subsector grew less rapidly (about 2% per year). One might guess that these differences could be related to a different industrial composition of growth, but the difference between the patterns of 1962-66 and 1970-75 are not dramatic, except perhaps in the greater role of metal products in the latter period (see Table A-23). B. Mobility Patterns of Different Industries It is of interest to consider whether the pattern of relative growth of larger and smaller plants differs among manufacturing industries. Three industries, food products (312), textiles (321) and clothing (322) were responsible for 40.2% of the total employment increase in the above cited sample of firms that operated both in 1970 and 1975. In the case of food products, the ability of small plants (less than 50 workers in 1970) to grow has exceeded that for plants of similar size in' the factory sector as a whole. The textile industry (321), on the other hand, expanded mostly through its large enterprises; only 12% of the net employment increase between 1970 and 1975 occurred in plants of less than fifty workers in 1970 as compared to 76.6% in plants with at least one hundred employees in 1970. During this period exports of textile products expanded substantially and it appears that small plants did not participate extensively in that expansion. Clothing industries (322) accounted for 9.5% of the net employ- ment increase between 1970 and 1975 in the sample of plants operating in - 67 - both years, an expansion attributable at least in part to the growth of clothing exports. The pattern of expansion was similar to that of the entire factory sector, except that firms which in 1970 had between 25 and 49 workers have expanded atypically fast. Even relatively small plants benefitted from export incentives granted under the Vallejo Plan, 1/ which permitted liberal imports of raw materials for export production. The beverages industry (313) shows signs of increasing concentra- tion: 72.4% of the net employment increase in this industry was concentrated in plants of more than a hundred workers. Similarly, in leather products, only plants of more than 50 workers were able to expand significantly. The inadequate provision of raw materials for leather has been .mentioned by small producers as one of the main constraints they face. Footwear industries (324) and wood and furniture (331-332) provide clear examples of product where small plants have been able to expand faster than larger ones, although their overall growth was not very large. plants Of the net employment increase in the footwear/included in the sample under discussion, only 15% took place in firms of at least one hundred workers, 32.1% in firms of 50 to 99 workers, and the rest (53.9%) in firms of less than 50 workers. In wood and furniture industries (331-332), 87% of the net employment increase took place in plants which in 1970 had less than fifty workers. Most of this increase was in firms of between 25 and 49 employees. To a large extent, the expansion in this industry was a result of the emphasis on construction activity which characterized the period * 1972-74. Another construction-related industry is "other non-maletallic mineral products" (369), which includes cement and bricks. Surprisingly, 1/ Some of these incentives were curtailed in 1974. -68- employment expansion in these plants was fairly low, and most of it occurred in plants of more than fifty workers. Plants in 'other chemical products' (352) increased employment substantially between 1970 and 1975, but mostly among already large plants. Similarly with clay, porcelain and glass products (361-362), where employment in plants of more than one hundred workers (in 1970) absorbed 72% of the net employment increase between 1970 and 1975 in the entire industry. The employment increase between 1970 and 1975 in paper and printing industries (341-342) has been distributed more broadly among plants of different sizes. Significant upward mobility is observed in plants which in 1970 had between 25 and 49 workers. Even more significantly, the employment expansion in metal products (381) has reached the lower-sized plants. Plants of 10 to 24 workers in 1970 increased their labor force by 30% in five years (equivalent to 6.3% annual growth rate) and accounted for 27% of the net employment increase in the industry. Employment in electrical and non-electrical machinery, transport equipment and professional and scientific equipment industries expanded significantly over the period 1970-75, accounting for almost 14% of the net employment increase in the sample of plants that operated in both years. Plants that had between 10 and 24 workers in 1970 increased their employment by about 55%, while those starting with 25 to 49 workers had an average increase of 36% over 1970-75. The larger plants have not grown as fast. - 69 - In summary, growth patterns vary across industries. The above results identify industries in which the growth performance (defined in terms of employment) of initially small or medium plants has been strong, and those in which it has been weak. Noteworthy inclusions in the former category are clothing, footwear, food, and metal products (both machinery and other items). The patterns of growth and decay of individual SSEs provide one test of economic viability, although it must be remembered that outcomes may be substantially affected by often adverse circumstances faced by SSEs (e.g., substantially higher interest rates than those of large firms). The impact of these will be assessed in Section 6. The next section studies survival and death patterns among small factories. 4. Survival and death rates in factory establishments The 1970 industrial census registers the year when a firm was established. The number of firms that declared having been born in year t and were registered by the census are those that survived from year t till 1/ 1971. (Table IV-9) This number can be compared with the actual stock of firms operating in year t, and a crude death rate (or alternatively a survival rate) can be computed. This procedure can be applied to firms of different size categories--if one can allow for the fact that firms which were initially 'small' may have outgrown their initial classification--and corresponding gross death rates can be obtained, bearing in mind, as discussed earlier, that the statistics tend to overstate plant mortality, as changes of name, location, etc., make it impossible to identify the same economic entity at a later date. Further, the coverage of the 1970 1/ Year in which the 1970 Census was actually carried out. Table IV. 9 ; COLOMBIA: ESTIMATED DEATH RATES OF FACTORY ESTABLISHMENTS ACCORDING TO SIZE,4956-71 Establishments that existed Estimated Establishments it 1971 and were Annual Death Rate (%) Establishment that existed created on or /1 of initial stock of Size in 1956 before 1956 Difference firms (1956-71) 10- 24 2,039 493 -10546 -9.0 25- 49 758 278 - 480 -6.5 50- 74 259 139 - 120 -4.1 75- 99 123 79 - 44 0 . 100-199 184 144 - 40 200+ 134 199 65 Total 3,497 1,332 -2,165 -6.2 /1 A negative sign may mean that those firms ceased operations or that they jumped to a higher size category. The latter applies only to firms of more than 200 workers which show a larger number of firms in 1970 than in 1956. It is assumed that large firms have not gone out of business, and that would mean a larger flow of previously smaller firms taking up their places. Source: DANE, Encuesta Anu4l de Manufacturaq 1956 4nd 1970, and unpubliAhed datA grQm the Jnduatr'ial Cenpup of 1970. 71 1 Manufacturing Survey was incomplete compared to all the earlier ones used in these comparisons. With reference to enterprises that operated in 1956, the calculated (upward biased--see footnote 1 ) death rate of smaller establishments is higher than for medium sized firms and both are higher than for larger firms. Firms of 10-24 employees have been disappearing at an annual rate of 9% whereas in firms of 25-49 and 50-74 workers the rates were 6.5% and 4.1%, respectively. 1/ There is no evidence of disappearance of large firms. Similar mortality patterns are found among smaller establishments that operated in 1965. (Table IV-10). For comparative purposes, the appropriate annual death rates for the period 1965-71 are reported. Firms of 10-24 and 25-49 workers declined at an annual rate of 12.2% and 4.0%, respectively. On the other hand, establishments of at least 50 workers have 1/ Only for those size categories for which there was evidence of an actual decline, have death rates been calculated. Based on the evidence that upward mobility of plants takes place mainly among adjacent size categories, it was concluded that with respect to 1956 (Table IV-9), the net increase in plants of at least 200 workers (+65) was the result of expansion of firms of 75 to.199 workers (rather than -eath) and consequently the death rate for these latter plants was not computed. For the smaller plants, the estimated -mortality may be too high for one group and too low for another because of movement among size classes,but this sector does not significantly affect the overall rate; it is upward biased, however, because of (a) failure to identify firms in the 1970 ,sample which really did exist since 1956, and (b) the relative undercoverage of the 1970 survey relative to the earlier ones. Some feel for this latter problem can be deduced from Table A-24, which indicates undercoverage in 1975. For the category 10-24 workers in 1970, it was less serious, perhaps a little under 40%. This degree of underreporting could obviously bias the results seriously. It would seem, however, that DANE's underreporting of small plants of 10-24 workers (an undercoverage which increased notably when plants of 5-9 workers were dropped from the planned coverage in the late 1960s) might arise mainly from failure to catch new plants as they appeared rather than from dropping of plants already on their list. To the extent that this was the case, the differential underreporting for this size group in 1970 would not upward bias the plant mortality rate estimates under discussion. Table IV. 10 COLOMBIA: DEATII RATES OF FACTORY ESTABLISJIMENTS ACCORDING TO SIZE: 1965-71 Establishnents Annual Death created before end Rate (%) Establishments of 1966 and operäting 1965- Estab.lislent operating in in 1971 .ifference 1971 Size 1965 10- 24 2,600 1,193 -1,407 -12.2 25- 49 897 702 - 195 - 4.0 50- 74 314i 300 - 14 75- 99 1731 176 3 100-199 276 283 7 200+ 265 284 19 Total 4,525 2,938 -1,587 - 6.9 /1 Sea footnote in Table Source: DANE, Encuesta Anual de Manufacturas, 1965 and DANE, unpublished data from the 1970 Industrial Census. Table IV-.ll: -COLOMBIA: ESTIMATED DEATH RATES OF FACTORY ESTABLISHMENTS ACCORDING TO SIZE: 1967-71 Establisments created Annual Death Establi-sli-lens before end of 1967 and Rate (%) operatling in 1965- Establishmient that existed 1971 ic 1971 1971 Difference 17 Size In 1967 10- 24 2,191 1,490 701 -9,2 25- 49 910 839 - 71 50- 74 341 346 5 75- 99 184 191 7 100-199 283 304 21 200+ 244 295 51 Tcital 4,153 3,465 - 688 -4.4 /1 Only firms of at least 10 workers. See footnote In previous table, Source: DANE, IndustriäNManufactuia Nacional 1967 and unpublishedANE data fm the 1970 Industrial Census. - 74 - tended to outgrow their initial size categories. Although the annual percentage decline in firms of 10-24 workers has been larger than for older firms (1956), this has been compensated by a better performance in firms of 25-74 workers, especially in the group of firms of 50-74 workers which did not decline at all from 1965 to 1971. Firms that operated. in 1967 show a considerably better (annual percentage) survival rate up to 1971 than firms that were in operation a two years earlier. (Table IV-4). The only size category that shows/large decline is that composed of establishments of 10-24 workers, the annual percentage death rate being about 9%. The number of plants in the range 25-49 workers fell but not dramatically. All other plants seem to have tended to move upwards to larger size categories. It can be asserted that the improved performance of small plants after 1967 is due to two main factors: first, the better overall economic environment that resulted from the trade liberalization measures of 1967 and the accompanying faster growth of G.D.P., and second, the improvement in credit conditions to small and medium firms, which is partly associated with the creation in 1968 of a governmental lending institution (the 1/ Corporacion Financiera Popular) designed to attend the needs of those firms. It is reasonable to assume that firm mortality would be more severe over 1965-67 than over 1967-70 since the former two years saw a less stable and expanding economy. While the data underlying the discussion of this section have serious weaknesses, it seems reasonably safe to conclude from them that there is a high rate of demise among small firms; just how high and for what reasons is not yet clear. 1/ These issues will be pursued in greater detail in section VI below. Totice, however, that substantial improvements have taken place only in the case of firms of more than 25 workers. - 75 - V. RELATIVE EFFICIENCY AND CAPITAL INTENSITY BY PLANT SIZE 1. Introd: uction A firm's "efficiency" may be defined in various ways. A common short hand definition is the ratio of social value of output to social cost of inputs, where the term social is used to denote that the prices applied to both outputs and inputs may not be equal to market pricee- if there are market imperfections prices may not be equal to social opportunity costs (in the case of factors) or marginal social benefits (in the case of goods). Thus if the labor used by a firm would otherwise have been unemployed, it would be invalid to assume that labor costs reflect use of a scarce resource; the social opportunity cost of the labor could be as low as zero. The definition just cited relates closely to the contribution a firm's existence matices to a country's national income. Unless the value of its output exceeds the social cost of resources used, that contribution is zero or negative. Such a definition disregards other goals of economic policy than maximization of,current output. Accordingly, it may be desirable to take a broader perspective, within which the effects of a firm's activities on employment, income distribution, future output levels (via the savings it generates and technological improvements it undertakes, for example) and other variables of interest are taken into account. In short, the firm characteristics one looks at depend on one's judgment as to which of its possible effects on the economy are of interest. When one asks the question "How much would be produced by the factors currently employed by a firm if it (the firm) did not exist?" it is evicnt that for the answer to be "More than the firm's current output," - 76 - (i.e., for the firm to be socially inefficient) there must be one or more market imperfections preventing the factors used by the firm from going to where their productivity would be higher. A firm's inefficiency cannot be demonstrated persuasively until that imperfection or imperfections is located. ThE need for analysis of market imperfections is highlighted when one takes note of the sort of methodology usually employed in comparing efficiency across firms, industries, or whatever. On the input side, to take an example, it is customary to define a few skill categories and assume all workers within a given category have the same productive capacity. Those workers are then assumed to have the same social cost in any efficiency calculations. If in fact the productive capacity of one firm's workers is below average, such a calculation would imply that it was inefficient when this was in fact not the case since the lower factor productivity was simply due to inferior factors. The alternative methodology, whereby it is assumed that a firm is efficient unless a market imperfection which leads to its being able to employ factors of given quality more cheaply than do other firms, would in this case lead to the correct conclusion that the firm was efficient, since no imperfection would be identified. In general it is desirable to approach the issue of relative efficiencies of different firms from both these perspectives, since neither by itself makes use of all available relevant information and each is quite vulnerable to error. Among firms which are inefficient in the sense that the inputs they use could be more productive elsewhere, it is important to make distinctions, according to the difficuilty of effecting the resource transfer which would be necessary if the firms did not exist. Where the imperfection - 77 - is subsidized access to public credit, the presumption is that there would be no difficulty in effecting such a transfer, so such a firm may be assumed to be inefficient in the most basic sense that its demise would lead more or less automatically to an increase in the total output of the economy. When a firm with a low output/input ratio uses factors which would not be easily transferable to another use, it cannot be viewed as inefficient given existing factor markets, though it would be inefficient and would presumably cease to exist were those imperfections abolished. In principle, it is not possible to judge the absolute efficiency (we define a firm as efficient in absolute terms if its demise would not raise national income) or the relative efficiency (one firm is defined as more efficient than another if its demise would lead to a greater decline in national income relative to its current value added than would be true for the other firm) of firms without a complete specification of the economy and all the complexities which could be reflected in a general equilibrium model and an assumption as to whether any market imperfections are expected to appear or disappear vis a vis the present situation. In practice, ideal eype calculations cannot be effected due to inadequate understanding of the structure of the economy, so general equilibrium models are not likely to help greatly in such analyses for the time being. Comparisons of relative efficiency by firm or plant size in Colombia cannot hope to be "refined" in the sense of taking due account of the different general equilibrium implications of the transfer of resources to or from firms of a given size. Nor can the measurements of factor inputs be very precise, especially for smaller firms. Calculations must thus be viewed as tentative and suggestive rather than conclusive. - 78 - One final point should be clarified. The finding that a particular type of firm is inefficient has no specific policy implication pending the discovery of the cause or nature of its inefficiency. There must be some difference in production function or factor prices underlying inefficiency and if the inefficiency shows itself in the form of lower factor incomes 1/ than could be earned in alternative uses, then there must be some barrier to the mobility of one or more factors to those alternative uses. Sometimes the most appropriate policy involves raising the efficiency of currently inefficient units, sometimes it involves terminating the existence of at least some of them, either directly or by encouraging factors to move elsewhere. Sometimes efficiency as defined abbve is a particularly misleading indicator, as when .onopoly power is used to maintain high prices; since the high monopoly price does measure marginal value in use, the high benefits recorded as a result of the monopolist's production are real, but they are high partly just because he is a monopolist. Individual firms in a competitive industry, whose social contribution was greater, would not score as high as such amonopolist. Since on average large firms have more monopoly power than do small ones, this aspect is likely to create a bias in favor of the former, in the sort of measures which can be used. A final measurement complexity involves the level of aggregation of product for which efficiency estimates are made. The greater the level of aggregation, the greater the likelihood that "efficiency" comparisons among groups of firms will be hard to interpret. If the product is not identical or nearly so, (both in physical characteristics and 4n location 1/ This would not necessarily be the case if the producing unit is a monopoly or is subsidized by the state. - 79 - and nature of the market) differences in the ratio of (social) value of output to (social) value of inputs may be the result of differences in market structure rather than of efficiency. Even for the same item, however, the problem of different firms selling in different locations can arise (e.g., if some sell in larger cities where prices may be higher or lower depending on the product, and others in smaller centers), so whether measured differences in value of output/value of inputs do reflect differences in efficiency is not related in any simple way to the heterogeneity of the product category. This problem constitutes one more reason that any com- parisons of efficiency be viewed as approximations and be made in the light of as full a knowledge of the context as possible. With these issues and caveats in mind, we may turn to the question of what can be said about systematic differences in factor productivities and in overall efficiency of plants of different sizes. The easiest factor productivity to measure and to compare across plant sizes is that of labor. Even here, however, the errors may be substantial, especially at the lower end of the size scale. As noted above, underreporting of output may be as high as 25-50% in small plants, while it is unlikely to reach these magnitudes 1/ in large ones. Employment is probably reported accurately in large plants; there may be some underreporting in small ones if the respondents feel obliged to maintain input-output ratios not too far from the true ones, but it is 2/ presumably less serious than in output. Lumping all these considerations 1/ In one way, DANE data overestimate value added, snecifically the failure to subtract some input qosts from gross value of output. This upward bias is probably greater in larger plants. 2/ There might even be overreporting if there are a lot of part time workers who are all mentioned by the respondent in small plants. Whether this occurs is not clear. - 80 - Table V-1: ESTIMATES OF NET SOCIAL BENEFIT PER UNIT OF CAPITAL, BY PLANT SIZE, AVERAGES FOR 1956-1967 Number of Workers 1-14 15-49 50-199 > 200 Very Small Small Medium Large 1. VA/HP 2.63 2.97 4.33 2.69 (3.06) 2. VA-((WB)*TO/TR))/HP 1.56 1.87 3.09 1.8 (1.79) 3. VA-(TG,,Craft Wage)/HP 1.56 2.24 3.75 2.35 (1.79) 4. VA/Kf 1.01 1.06 .83 .48 (1.17) 5. VA-((WB)-TO/TR))/Kf .60 .66 .59 .32 ( .68) Index based on value for Size Class #9 ( .200 workers) 1. VA/HP 98 110 161 100 (114) 2. VA-((WB)*TO/TR))/HP 86 103 171 100 (99) 3. VA-(TO*Craft Wage)/HP 66 95 159 100 (76) 4. VA/Kf 210 221 173 100 (243) 5. (VA-((WB)*TO/TR))/Kf 187 206 184 100 (.212). Symbols: VA - Value added. WB - Wage bill (including fringe benefits) TO/TR - The ratio of total employment to paid workers Kf - Fixed capital, estimated by Todd using DANE investment data in conjunction with data on horsepower. HP - Horse power Craft Wage - Average wage in plants of <15 workers. Source: Adapted from John Todd, Efficiency ..., op. cit., Table 37, p. 98. Underlying data were from DANE's annual manufacturing survey. Note: Figures in parentheses reflect Todd's adjustments to value added, as explained in Part 2 of Section B of his thesis. - 81 - together, one might guess that relative labor productivity would be under- stated by 10-30% in the smallest plants compared to the large ones of 200 workers and up. This possible bias in the data is small compared to the differences in labor productivity by plant size, however. At the aggregate level, the reported differential between plants of 5-9 workers and those of 200 and over is typically 3-4 fold. (In 1956 it was 3.3 fold, and in. 1975 it was 3.8 fold.) When due account is taken of differences in labor "quality," these labor productivity differentials are substantially reduced. Recent comparisons of wage rates across plant sizes suggest (a) that after allowance for differences in personal characteristics of the workers there are rather 1/ small true wage differentials between larger and smaller establishments and (b) that those differences in personal characteristics would lead one to predict average wages twice as high in the largest plants than in the smallest. If both misreporting (relative underreporting by the small plants) and differences in labor quality across size categories are taken into account, a reported difference in labor productivity of say 250% might reduce 2/ to one of perhaps 25-50%. Since it is not clear whether the two problems 1/ See Francois Bourguignon, "Poverty and Dualism in the Urban Sector of Developing Economies: The Case of Colombia," mimeo, 1978, and Bernardo Kugler, Alvaro Reyes and Martha I. de Gutierrez, "Educacion y Mercado de Trabajo Urbano en Colombia: Una Comparacion Entre Sectores Modernos y no Modernos," Bogota, Corporacion Centro Regional de Poblacion, Julio 1978. In his analysis of DANE's seven city 1974 household survey, Bourguignon defined the modern sector as including all large production units of five or more employees, plus all individuals with a university education and all government employees, (op. cit., p. 708). (cont.) - 82 - While the gross income differential between the modern and traditional sectors was 1.86, most of it was related to educational and other charac- teristics likely to affect personal productivity. Bourguignon performed a wide range of tests providing some confidence in his conclusion that the typical modern-traditional differential (holding personal characteristics constant) is probably fairly low, e.g., under 20 percent. If, as is plausible to assume, the modern sector selects workers of higher capacity, given education, experience, etc., then the true differential--holding ability constant--would be even less than Bourguignon's figures show. On the other hand, fringe benefits which are higher in the modern sector, are seriously underreported. It is possible that, if the modern-traditional earnings gap was indeed as low as Bourguignon estimates it to have been in 1974, that it had fallen from earlier periods when it was more marked. This hypothesis remains to be tested, however. 2/ The differential measured by Bourguignon was between the "modern" sector broadly defined to include establishments of five or more workers plus some other workers, and the rest of workers. "Modern" thus included the full range of plant sizes involved in our discussion of the factory manufacturing sector. He did not analyse differentials within the modern sector. His analysis referred to the whole urban economy, not just manufacturing. In light of the very large reported average earnings differentials by plant size in the factory sector, it seems reasonablE to guess that between the largest and smallest plants the differential is larger than the 20% or :30 reported by Bourguignon in the modern/traditional break. - 83 - just mentioned have changed in importance over the years, it is not possible to judge whether true wage and labor productivity differentials have been increasing or decreasing. The unadjusted DANE data indicate a l/ gradual widening of reported wage differentials between 1956 and 1970. (See Table A-25). Between 1970 and 1975, it appears that the opposite has happened. For the plants of less than 25 workers this result cannot be relied on since coverage continued to decline during those years, but the phenomenon characterized plants of 50-99 workers too, where coverage presumably did not change significantly. Measurement of capital poses much more serious problems than labor, rendering estimates of both capital productivity and total factor productivity open to considerable error. Todd's analysis was the first in depth attempt at such measurement in Coloubia. All raw data were from DANE. Like other analysis, he was not able to include non fixed capital in the estimates, despite its evident importance in both small and large establishments. 1/ Decreases in coverage of small plants in the late 60s create problems of comparability with the earlier years. In some years the use of data from earlier years in some small plant biased their figures downward. But in 1969, whose wage ratios by size are not far from those reported in Table A-25. For 1970, this latter problem should not have been serious, according to a recent DANE document. (DANE, Memoria Censos Economicos 1970, Bogota DANE, 1975, p. 42.) The decreasing coverage of small plants would be expected to lead to an upward bias in such variables as wage ratios, labor productivity, etc. - 84 - At an aggregate level Todd estimated that the average output/ fixed capital ratio over 1956-67 size categories (up to 14 workers) fell by about 30% between the smallest and the medium ones (50-200 workers), then fell sharply at the largest one (200 workers and up), for which the ratio was less than half that for the rest of the sector (Table V-1). He notes that the ratio of fixed capital installed horsepower appears to increase substantially with plant size so that horsepower would not be a good proxy for fixed capital. Where prices may be higher or lower depending on the product, and others in smaller centers) so whether measured differences in value of output/value of inputs do reflect differences in efficiency is not related in any simple way to the hetero- genity of the product category. This problem constitutes one more reason that any comparisons of efficiency be viewed as approximations and be made in the light of as full a knowledge of the context as possible. With these issues and caveats in mind, we may turn to the question of what can be said about systematic differences in factor productivities and in overall efficiency of plants of different sizes. Todd's measure of efficiency was value added minus the cost of some factors other than capital divided by the fixed capital stock (this letter being based on installed horsepower data adjusted by investment data). Todd's analysis was the first indepth attempt of such measurement in Colombia. All raw data were from DANE. Like other analysts, he was not able to include non fixed capital in the estimates, despite its evident importance in both small and large establishments. Table V-2: HORSEPOWER/TOTAL EMPLOYMENT, 1966 Legal Form of Number of Workers Organization <5 5-9 10-14 15-19 20-24 25-49 50-74 75-99 100-199 > 200 Total Proprietorship 1,40 1.10 1.21 1.66 1.21 1.45 1.16 1.25 2.34 2.11 1.30 Limited partnership 4.87 3.41 2.36 2.88 2.39 2.54 1.75 1.89 2.39 3.43 2.57 Corporation 14.15 7.97 6.67 8.77 5.50 13.78 4.36 4.93 5.44 5.23 5.47 Value Added/Horsepower 1966 ('000's of 1952 pesos) Proprietorship 1.8 2.3 2.4 2.8 3.3 3.4 3.6 4.7 2.5 3.1 2.6 Limited partnership 1.6 1.7 2.2 1.9 3.2 2.8 4.2 3.6 3.6 2.8 3.0 00 Corporation 4.0 2.8 1.5 1.9 2.6 1.8 4.1 4.0 3.6 2.8 2.9 1 Source: DANE, unpublished data. Note: Figures are not corrected for methodological error affecting small plants after 1962. It affects primarily monetary values like value added, leading to a downward bias in value added/horsepower estimates for plants of less than 15 workers. - 86 - Todd's measure of efficiency was value added minus the cost of scarce factors other than capital divided by the fixed capital stock (this latter being based on installed horsepower data adjusted by investment data His estimate of fixed capital productivity is less than half as high in the largest size category (200 workers and up) as in plants below 15 workers, and the measure of efficiency, when all inputs were assumed to be priced at their scarcity values, showed only a slightly smaller advantage in the smaller plants. Only when a given ("craft") wage was assumed to measure the opportunity cost of all workers did the smallest plants fail to achieve the highest ranking (row 3 of Table V-1), although efficiency remained-a decreasing function of size across the other three categories distinguished. Todd classified some input and output data by juridical form as well as size category, observing that in the aggregate there was a clear relationship between horsepower/employment and juridical form, fcr a given size category; the same was true for the ratio of value added to horsepower. For a given size category horsepower/employment was greatest for corporations, next highest in limited partnerships and lowest in proprietorships (Table V-2). For corporations this ratio actually tended to decline with size, for limited partnerships it was about constant and for proprietorships an l/ increasing function of size. This suggests that those_factors which affect capital intensity (factor prices, available information, etc.) are more 1/ These relationships would be somewhat different if fixed capital rather than horsepower was being considered, but it seems safe to presume that the difference by juridical form would still exist. - 87 - closely related to juridical form than to size of plant. Juridical form, may of course, be simply a good proxy for size of firm, since many small plants whose juridical form is the corporation are presumably small branches of a large firm. The high capital intensity of small plants which are part of corporations suggests that the size/capital intensity link is less technologically based than might have been surmised. Alter- natively, of course, it might be that corporations concentrate in the industries whose range is atypically capital intensive and that such capital intensity is equally possible at small scales. It is not possible with the available data to distinguish the two competing hypotheses. A final interesting observation from Todd's work is that there may be additional induced improvements in efficiency when sectors become more size diversified. Value added/horsepower ratios are highest in those sectors whose output is more evenly spread 'across size classes, and they are lowest in those sectors dominated by large plants. This superiority of the more size diversified sectors also holds.within the individual size classes. In the largest size class, for example, (that with 200 or more workers), VA/UP was uniformly higher by a wide margin in sectors where output is relatively evenly spread across size classes than in sectors which were 1/ dominated by large plants for both 1960 and 1965. Todd did not estimate fixed capital at the two digit level, but did note that across three digit industries the ratio of value added to horsepower is often higher for the largest plants and often lower (see Table V-3), i.e., at more disaggregated levels there is considerable 1/ Todd, Efficiency ..., oD. cit., P. 131. - 88 - Table V-3: VALUE ADDED TO HORSEPOWER RATIOS FOR SMALL AND MEDIUM PLANTS COMPARED TO LARGE PLANTS, 1965 R = VA/HP In plants of less than 200 workers VA/HP In plants of more than 200 workers Sector No. Name R >, 1.5 2030 Fruits and vegetables 2060 Bread 2320 Knitted fabrics 2360 Silk 2440 Women's clothing 2460 Hats 2710 Wood pulp and paper 3110 Chemicals 3140 Soap 3180 Waterproofing and Adhesives 3320 Glass and glass products 3330 Ceramics 3340 Cement 3410 Iron 3530 Cutlery 3540 Non-electric stoves 3550 Aluminum productp 3630 Industrial machinery 3710 Electrical machinery 3810 Boats and boat repair 3830 Auto assembly R < .67 2020 Milk products 2050 Milled products 2080 Candy and chocolate 2090 Other foods 2130 Beer 2140 Soft drinks 2510 Wood 2810 Printing 2910 Leather 3020 Rubber products for home use 3210 Oil refining 3350 Cement products 344-0 Non-ferrous metals 3520 Hand-tools and hardware 3730 Electro-mechanical equipment 3740 Wire and cable 3760 Electrical installation equipment 3850 Automobile repairs 3860 Airplane repairs Source: Todd, op. cit.,p. 135. - 89 - variability in this ratio, and possibly in fixed capital productivity as well. Todd's analysis does leave unresolved the question of whether in the 1950s - 1960s differences in fixed capital productivity and in social efficiency across plant size in manufacturing as a whole- reflected similar differences at the two or three digit levels or were mainly the result of the aggregation process. We now turn to an analysis of data in 1970, in hich we attempt to compare firms which can be thought of as producing a similar product. Any relationship between capital intensity and size which is not due to technology can best be analysed when the type of production is fairly homogenous. The measure of capital used in the discussion which follows is "book value of fixed capital," as recorded in DANE data. The interpretation of this statistic is of course difficult, and in many countries book values seriously understate the current real.value of capital by any reasonable economic definition. This appears to be the case in Colombia, with the understatement,probably falling in the range of one half to two thirds. This leaves considerable uncertainty in any comparisons of capital productivity, etc., and calls for consideration 1/ Otherwise, as noted above, both technological issues and market differences may be involved, e.g., differences in value added per worker across sectors may result not only from differences in capital intensity--which respond to technological and market conditions, including risk--but also from differences in the degree of competitiveness in each sector. Capital intensity may depend more on the type of economic activity than on size. Some activities (e.g., oil refineries) are constrained by the technology to be capital intensive, while others (e.g., furniture production) are not,so that relative factor prices are more likely to be important in determining factor proportions. 2/ See next page. - 90 - of which groups of firms are likely to have lower book value/true value ratios. Bearing these considerations in mind, the industrial census of 1970 has been fragmented into 5-digit Industrial Classification Sectors. Table B-1 in Appendix B presents the distribution of these establishments according to size (measured by number of workers per establishment). Colombia's industrial production, although diversified, is still largely for the domestic market; only 6.75% of manufacturing output was exported in 1973. Since in several senses the internal market is rather limited, many types of production are carried out by only a handful of enterprises. About 42% of the 435 five-digit sectors in which Colombia has established production are supplied by less than four firms in each, 1/ and otly 27 sectors have at least 50 firms in them. Table V-4 presents 1/ If it is assumed that not more than one large firm (of 100 or more workers) operates in those sectors which have less than five firms of any size, then it follows that about one third of these sectors of less than five firms are dominated by a large firm. 2/ (Of previous page)' An estimate of factory fixed capital by Berry in 1969 is 45.3 billion pesos. (See A. Berry "A Descriptive History of Colombian Industrial Development in the Twentieth Century," Statistical Appendix Table A-32 and A-184) This could imply a 1970 figure of 47-48 billions. The reported figure for book value at end of 1970 is 18.4 billions. Since coverage was incomplete, this figure would have been a little higher had the whole sector been reported. But it seems improbable that more than 50% of true capital value was reported by responding firms. There is no single "correct" way to measure value of fixed capital but even allowing for this ambiguity, the book value figure is evidently far below economic value. This is somewhat surprising in light of the fact that incentives to under- report would seem to be smaller for capital than for output (for example) and since inflation was not so high just before 1970 as subsequently. Note that a ratio of .5 in the book value/true value of fixed capital would seem consistent with the reported output capital ratio in 1970 of 1.37. With a reported labor share of 35.4%, this-would imply a gross profit to fixed capital ratio of 89%. If the true ratio (net of depre- ciation) were .40%, the degree of understatement of fixed capital would be over 50%. (For a further discussion, see Appendix A.) - 91 - the distribution of these sectors according to the number of firms in each, and the importance of "large" plants (100 or more employees) in each group. Table V-4: CHARACTERISTICS OF 5-DIGIT SECTORS BY NUMBER OF ESTABLISHMENTS IN EACH 5-DIGIT SECTOR Number of Plants in Each 5-Digit Sector 1-4 5-9 10-19 20-29 30-49 50+ Total 1. No. of 5-digit sectors 183 89 77 36 23 27 435 2. Percent distrib- ution of 5-digit sectors 42.1 20.5 17.7 8.3 5.3 6.2 100 3. No. of Establish- ments of 100 or more employees 59 117 184 94 50 164 668 4. Total no. of enterprises 416 584 1,030 852 895 3,690 7,467 5. Percent of all firms which are large (Row 3/ Row 4), 14.2 20 17.9 11 5.6 4.4 8.9 Source': Table B-1, DANE Industrial Census 1970. Lest the reader be misled by the reliance placed on the 5-digit industrial classification, a warning should be given as to the somewhat arbitrary nature of any industrial subdivisions and the implicit danger that "similar" productions may not be listed under the same heading whereas "dissimilar" items may be classified together. Adopting a five-digit classification reduces but does not eliminate the danger of comparing 1/ activities which may have only superficial resemblance. 1/ See next page. - 92 - Of the 435 five-digit sectors, those consisting of the largest number of establishments were selected for individual analysis at the firm level. This procedure presents several advantages: (1) First, the extent of "monopolistic" factors that could distort productivity measures is minimized; (2) second, atypical firms can not so easily distort the inferences about performance; (3) third, since the market for industrial products is rather limited, t-hse sectors which have a large number of firms are bound to include small firms in their midst; and (4) finally, given the pattern of specialization of Colombia's industrial sector, it can be seen that more than 50% of all enterprises in the factory sector in 1970 were included in only 30 industrial sectors at the 5-digit level! Table A-27 presents the size distribution of establishments and employment in the selected 30 sectors. These 30 five-digit sectors are mainly those in which there is a large number of SSEs. These sectors account for 54.74% of the number of establishments and 27.86% of the l/ employment of the entire factory sector. But the smaller plants (less than 25 workers) included in these sectors have a much higher representation: they account for 59.76% of the establishments and for 56.0% of the employ- ment in all factory plants in that size category. 1/ Composed of 435 five-digit sectors. 1/ (Of previous page) For instance, "pots and pans" and production of electrical machinery would be classified under the same heading of "metal products" at the 2-digit level. Production of automobiles, motorcycles and bicycles would be linked at the 3-digit level; the last two items would be under the same heading even at the 4-digit level Cables and metal cans could be distinguished only at the 5-digit level. -93- There are substantial differences among the selected sectors. Perhaps the most important is the relative presence of large plants: In 12 of the 31 sectors, more than 60% of the employment is in firms of at least fifty employees. Seven of the selected sectors are in food production (311-312). Clothing (322), footwear (324), sawmills (331), furniture (332) and metal products (381) have three industrial subsectors represented in the sample of 30 sectors. The remaining industries are soft drinks (313), tobacco (314), textiles (321), paper products (341), printing (342), pharmaceutical products (352) and non-metallic mineral products (369). In order to have a broader representation, basic metal industries of iron and steel (371) have been included among the selected sectors, bringing the total to 31 specific industries. 2. Distinctive Features of SSEs in the Factory Sector: 1970 In those cases in which the number of plants in a given 5-digit sector was large enough, the data have been divided into two groups: factories of less than 25 workers (henceforth referred to as SSEs) and those of 25 or more workers ('large' plants). This procedure was applied to 11 of the 31 selected sectors (Table V-5), specifically: bakery products, wheat and corn bread (31171); textiles, knitted synthetic undergarments, shirts, sweaters, etc. (32134); clothing, men's and boy's jackets, suits, and trousers (32201); clothing, women's and girl's blouses, dresses, etc. (32202); clothing, men's and boy's plain cotton and synthetic shirts (32206); footwear, men's leather shoes and boots (32402); sawmills, sawnwood (33111); furniture, non-metallic tables, chairs, etc. (33202); printing, checks, -94- stamps and other commercial printing (34203); non-metallic mineral products clay construction products, bricks and tiles (36912); non-metallic mineral products, concrete and cement construction products (36992). Capital Intensity Large firms tend to have a higher capital-labor ratio in 9 of the 11 sectors; the geometric mean in the group of large firms is higher than the corresponding mean for SSEs in all sectors except 32202 and 33111. In 6 of the 11 cases, including the two just cited, the difference between the (geometric) means of capital-labor ratio in large firms and SSEs does not exceed 15%. In 2 of the 9 sectors in which large firms are on average more 1/ capital-intensive, the difference in geometric means is between 27 and 30%, whereas in the remaining three cases it exceeds 80%. (Table V-5) Even within well specified 5-digit sectors, the differences between small and large firms can thus be large, though on average they are much less 2/ than in the manufacturing sector as a whole. In 2 sectors, the capital intensity of large firms is twice as large as that of the corresponding small firms (36912) and (36992), and in a third one (32402) the difference is close to 80%. These differences may be due in part to aggregation problems that still exist at the 5-digit level, but more importantly, they are likely to 3/ reflect the dualistic structure of production in some sectors. For instance, 1/ Sectors 32201 and 34203, respectively. 2/ For factory manufacturing as a whole, plants of 25 workers or more had a fixed capital/labor ratio about 170% above that of small plants, when the ratio is defined as the total book value of fixed capital in a group of firms divided by the total number of workers (see Table A-28). 3/ See next page. Table V.5: COLOMBIA: CIIARACTERISTICS OF SMALL AND LARGE FIRMS IN SETECTED FAITORY SECTORS, 1970 Means and Coefficients of Variation in 11 Five-Digit Sectore Sectors 31171 32134 32201 32202 32206 32402 Coeff. of Coeff. of Coeff. of Coeff. of Coeff, .f Coeff. of Mean Variation Mean Variation ilean Variation Mean Variation Mean Variation Mean Variation L 2 Large Firms: VK 1.79 (.81) 1.67 (.88) 2.31 (,68) 2.35 (.71) 2.43 (.70) 2.36 (.65) L 41.84 (.47) 68.50 (.80) 75.84 (.87) 46.40 (.47) 60.11 (.69) 78.05 (.68) Y 58.50 (10.15) 63.20 (6.72) 59.01 (8.50) 62.47 (8.23) 60.91 (8.98) 58.09 (13.09) K/L 11,830.10 (.78) 14,217.10 (,81) 9,365.61 (.84) 7,191.10 (.92) 8,347.36 (.58) 10,126.39 (.79) L2 small Firms: VK 1.55 (.99) 1.38 (.80) 2.01 (1.11) 2.09 (.70) 1.98 (.85) 2.81 (.80) L 7.67 (.46) 9.90 (.59) 9.34 (.55) 9.16 (.47) 12.66 (.45) 8.90 (.53) Y L1 62.41 (8.05) 66.60 (3.00) 62.94 (8.28) 65.63 (6.16) 64.88 (7.26) 64.06 (6.46) KIL 10,578.72 (.96) 12,686.58 (.90) 7,355.47 (.97) 8,133,12 (.77) 7,447.99 (.88) 5,691.33 (.95) 33111 33202 34203 36912 36992 Sectors Coeff. of Coeff. of Coeff. of Coeff. of Coeff. of Mean Variation Mean Variation Mean Variation Mean Variation Mean Variation 12 Large Firms: VK .94 (.90) 3.29 (.84) 1.52 (.72) .93 (.98) 1.33 (.76) L 46.62 (.69) 44.17 (.64) 52.45 (.81) 48.19 (.50) 51.25 (.70) Y 64.42 (7.21) 60.21 (10.20) 55.22 (14.90) 55.98 (11.17) 60.53 (11.07) K/L 18,462.56 (.90) 6,493.78 (.68) 22,379.49 (.83) 20,768.65 (1.18) 19,087.71 (1.05) t2 Small Firms: VK 1.02 (.97) 2.03 (.85) 1.04 (.92) 1.10 (.99) 1.75 (1.07) L 11.11 (.52) 8.05 (.46) 9.29 (.48) 7.73 (.63) 9.35 (.45) Y1_ 61.51 (8.06) 62.97 (8.34) 60.06 (10.41) 58.83 (12.21) 61.66 (9.93) K/L 19,187.23 (.90) 7,313.67 .(.89) 17,311.04 (.92) 10,319.60 (.91) 8,282.50 (.97) LI Means are geometric means except in the case of year, where y standards for arithmetic average (e.g., 58 represent 1958) and the accompanying number is the standard deviat:i. In the other cases the number in parenthesis approximates the coefficient of variation, provided the frequency distribution of the variable is log-normal, VK Value added - capital ratio. L Number of -workers Y Year when firm started operations. K/I-- Capital - labor ratio. "Small" refers to firms of less than 25 workers, all others are "large". Source: Unpublished data from DANE Industrial Census, 1970. - 96 sector 36912 is described as 'Non-metallic mineral products: clay construction products (bricks and tiles). It is well known in Colombia that small brick 1/ firms in this sector adopt different technologies from those of larger firms. The second sector (36992) also involves non-metallic mineral products for construction (concrete and cement) and the differences in capital intensity can be explained similarly. The third sector which involves non-negligible differences in capital-intensity between small and large firms is 'footwear: men's leather shoes and boots' (32402). Although the term 'dualistic' may be too narrow for capturing the variety of production techniques in a given sector, it nevertheless illustrates the gulf that separates the large modern factories from the more traditional shoe shops still widely scattered across Colombia. Nevertheless, comparisons of means are harder to interpret if there are substantial variations within the groups that are being compared. In 7 of the 9 cases in which large plants are, on the average, more capital- intensive than SSEs, the coefficient of variation of the capital-labor ratio 2/ 3/ is higher for smaller firms. In other words, smaller firms are generally less homogeneous and present wider differences in capital intensity than do larger firms. This conclusion may be due to the smaller firms facing a greater range of capital prices (or degrees of access to capital) whereas nearly all large firms have good access. 1/ Consult the study of the brick industry by M. A. Bailey. 2/ The coefficient of variation as measured by the standard devi'ation of the logaritb. of the capital-labor ratio. This measure is adequate only if the capital-labor ratios are log-normally distributed. 3/ The two exceptions are sectors 36912 and 36992. In these sectors the coefficient of variation of the capital-labor ratio among large firms exceeds 105%. -97- Summing up, in general book values of fixed assets suggest that larger plants tend to be more capital-intensive than smaller ones, but the differences are not very significant in the majority of industries studied. Furthermore, smaller plants present higher variation in their capital intensity than do large plants. At the same time, there are some sectors in which the differences in capital intensity across sizes are substantial and indicate dualistic production schemes. It should be noted that the relationship between capital intensity and size is likely to depend to some degree on the measure of size. This analysis uses the customary dolombian measure, size of labor force. If capital input were used, a stronger positive relationship would be expected. In a basic sense the most appropriate measure would be a measure of all factor inputs or of output. For recent years, firms are classified by output level on an aggregate basis, which permits a comparison between the size-capital intensity relationship using that measure of size as compared with number of workers. (See Tables A-28 and A-29 for further details.) As shown below, when ranking of plants is by number of workers the top 12% have a capital intensity 3.6 times that of the bottom 37% BOOK VALUE OF FIXED CAPITAL/WORKER: 1970 Plants ranked by number of workers Plants ranked by output Bottom 37% of plants 18.30 13.07 Top 12% of plants 65.53 71.16 - 98 - (groups of firms chosen to minimize interpolation), but when ranking is by output the corresponding ratio is 5.4. Were capital the measure of size it would probably be above 6. It is not clear whether the size- capital intensity relationship would be as sensitive to the measure of size for individual industries as it is in the aggregate, but there is little doubt that it would emerge more positive were any other measure of size used. Capital intensity differs markedly across industries. Small firms engaged in the production of men's leather shoes and boots show the lowest average (5,700 $Col. of 1970 per worker) while the highest ratio corresponds to large printing firms (almost four times higher). In four of the eleven sectors the (geometric) average capital-labor ratio did not 2/ exceed 9,500 $Col. per worker, in either large or small firms. Textile firms (sector 32134) have (geometric) average capital-intensities of 12,700 $ol. and 14,200 $Col. per man for 'small' and 'large' firms, respectively. Bakeries producing wheat and cornbread show capital-labor ratios of 10,600 and 11,800 $Col. per man for 'small' and 'large' firms respectively. In addition to large printing firms, the highest average capital-labor.ratios are to be found among large firms producing non-metallic 1/ All money values in this section are in $Col. of 1970. 2/ The sectors are: clothing (32201, 32202, 32206) and household furniture (33202). - 99 - 1/ mineral construction products (sectors 36912 and 36992), and in sawmills 2/ (sawnwood). In order to explore further the relationship between (fixed) capital-intensity and size, regression methods have been utilized. In this case it was not necessary to dichotomize the range between 'small' and 'large' plants since the specific number of employees in each plant could be used as its measure of size. Table A-30 presents the 31 regressions, one for each of the industries under discussion. In order to capture non- linearities, the log of the capital-labor ratio (log K/N) is regressed against log of employment (log N) and its square (log N)2. The year when the plant started operations is also included among the independent variables in order to correct for possible bias in the book value of fixed assets due to inflation. The results show that in only seven of the 31 industries, is there 3/ a statistically significant association between capital intensity and size, The estimated capital-labor ratios for firms of 100 workers are not systematically higher or lower than those of firms of 25 workers. Diagrams of the capital-labor ratio as it relates to the size of the firms, for enterprises belonging to the same industry, are shown in Appendix C. These diagrams confirm on the one hand the lack of clear statistical association between size and capital intensity, and on the other the higher dispersion 1/ With capital-labor ratios of 20,800 and 19,100 $Col. per man, respectively. 2/ Where capital per worker is 18,500 $Col. and 19,200 $Col. for large and small firms, respectively. 3/ Sectors 31403, 32134, 32402, 32404, 32202, 34203 and 36912. - 100 - in the capital-labor ratios of smaller enterprises. While the use of output (or total inputs) to measure size would presumably have generated a more positive size-capital intensity relationship, the above results seem to make it clear that many other factors than size play a role in the determination of capital intensity, even within fairly narrowly defined industries. There are two main explanations for these results. Firstly, firms face different factor prices and these have' repercussions on the capital-intensity of production. Ceteris paribus, those firms that have access to credit under more favorable terms, will tend to adopt more capital-intensive techniques than others which do not have those facilities. There is evidence that larger firms face lower interest rates and pay higher wages than smaller firms, factors that encourage the mechanization of production. At the same time, it appears that even among small firms (as measured by the number of employees), there are considerable differences in the factor prices faced, leading to the coexistence of a variety of techniques. To the extent that interest rates reflect a risk premium, those firms that are able to provide credit guarantees to lenders are likely to face lower effective interest rates. And some small firms may not need much outside capital, and may not have lucrative alternative uses for their own capital; such firms face a low operative price of capital. Secondly, the lack of association between capital intensity and size may be due to differences in the utilization of the stock of capital. To the extent that large firms are capable of working for two or three shifts per day--which SSEs may be less able to do because of the combined managerial - 101 - and supervision tasks that fall on the owner of the firm--they may be able to have lower capital-intensity even if they utilize more mechanized 1/ methods of production. In other words, a highly mechanized process of production may involve less capital investment per worker than a less mechanized technique that is only partly utilized. Thus the dispersion of capital-labor among both small firms and large ones may be partly due to differences in the utilization of fixed assets. Capital Productivity In seven of the eleven selected sectors with relatively large ..-numbers of plants, value added per peso of fixed assets was higher in large firms. The exceptions were sectors 33111, 32402, 36912 and 36992. The last three sectors were described above as having a dualistic production structure (because of the considerably higher capital intensity of larger firms). In 2/ these sectors the net productivity of capital is higher in the smaller firms. The various statistical and methodological caveats cited above should be reiterated here, in light of the possibly surprising finding that fixed capital productivity appears to increase with plant size. First, it is likely that DANE statistics involve greater underreporting of output for small plants than for large ones. Less obviously, it may be that book value of fixed 1/ The distinction between capital-intensity and mechanization of production is explored in A.K. Sen, Oxford, 1975. 2/ The net productivity of capital (V/K) can be- shown to be equivalent to V/L, where V is value added, L is number of workers, V/L is labor K/L productivity and K/L is the capital-labor ratio. If the percentage difference in the capital-labor ratio of large and small firms were the same as the percentage difference in the productivity of labor, the productivity of capital would be the same for both groups of firms. - 102 - capital understates total capital more for large plants. either because it understates the economic value of fixed capital more or because fixed capital is a smaller share of total'capital'(of which there is some evidence, cited above). Finally, larger plants are more likely to have some degree of monopoly power, and thus to have reported output inflated by monopoly prices. Lower capital productivity for smaller plants cannot be ruled out as implausible, however, since the opportunity cost of their capital may sometimes be fairly low, or the rate of return may be below the rate expected by the entrepreneur. -The finding of lower capital productivity in smaller plants is of course less surprising in light of the small reported differences in capital intensity by size of plant. SSEs show more variability not only with respect to capital- labor ratios, but also with regard to the productivity of capital. The coefficient of variation of the value added-capital ratio is higher in SSEs in nine of the eleven selected sectors. This suggests that, whatever the average performance of SSEs, at their best they can be quite impressive in terms of capital productivity, and raises the possibility that much could be gained by public policies having the effect of bringing the productivity levels of the less productive SSEs up to those of the more productive ones. 1/ The exceptions are sectors 32134 and 32202. In the last one the relevant difference between small and large firms is very small. One poscible illusionary reason for this greater variance may be a greater variance in accuracy of reporting by small firms, especially of output, than by large ones. This seems very likely, given the greater range of accounting practices (or non-practices), fear of being taxed, etc., which is likely to characterize the small plants. - 103 - In order to probe deeper into the relationship between capital productivity and size, it is desirable to isolate other variables which, in addition to size, may have an independent influence on capital producti- vity, namely, capital intensity and year when the firm started operations. The average age of small firms is lower than that of large firms in ten of 1/ the eleven sectors, and, as is noted below, productivity is positively associated with age. 3. Scale Economies and Diseconomies Overall static efficiency depends both on capital productivity and on labor productivity, and is expressed as "total factor productivity." In order to assess the extent of economies of scale on the one hand, and the efficiency in small-scale enterprises vis-a-vis larger firms on the other, regression methods have been applied to observations at the establishment level corresponding to specific 5-digit sectors, utilizing once again the. data of the 1970 Industrial Census. One of the limitations of the regression approach described is that it constructs the production function on the basis of all firms in a given industry, i.e., without excluding those that can be considered technically 'inefficient.' In other words, the production function in this case represents the average performance of firms rather than the actual production frontier defined on the basis of only 'efficient' firms. 1/ The exception is sector 33111. In addition to being older,large firms show less uniformity with respect to the year when they started operations. This is perhaps the only feature in which the coefficient of variation is lower in the case of SSEs. Note, though, that age statistics may be rather misleading, especially those of the smaller firms. 2/ 'Inefficient' in the sense that using the same level of inputs they arrive at a lower level of output than other firms in the data. - 104 - In the case of SSEs which have a larger dispersion than large firms regarding productivity this implies that the performance of the best SSEs may be blurred by that of the less efficient firms. The other major problems in the analysis are the possible differential underreporting of output, the inadequate measure of capital, and the assumption of homogeneous labor across firm sizes. The latter, especially, is likely to bias the results towards finding economies of scale, since average labor quality appears to be a positive function of firm size. The analysis starts by postulating the existence of a modified 2/ Cobb-Douglas production function. ht (b+c log N) d u Vi = Ae Ni Ki ei where V = value added in ith firm A = a constant t = year when the firm started operations Ni = number of workers in ith firm Ki = capital stock (of book value of fixed assets) 1/ This relationship holds at the aggregate (all manufacturing) level, if one may judge by wage and labor force composition data. Whether it holds as strongly, or at all, within such narrow industries as are the basis for analysis here remains to be seen. / A similar function is used in Z. Griliches and V. Ringstad, Economies of Scale and the Form of the Production Function, North Holland, 1971, p. 88. The main difference lies in the variable "year when the firm started operations." Its inclusion here is intended to correct for undervaluation in the book-value of the capital stock (K) due to inflation. -105- u 3= (multiplicative) random error and h, b, c and d are parameters to be estimated After some transformations it can be shown that: V. 2 K log - = a + ht + g log N. + c (log N.) + d log- + ui N. N. 1. 1 where 0 = log A andg= b + d - 2/ This function has been estimated for different 5-digit sectors. It reduces to the simpler Cobb-Douglas function if the parameters h, g and c are assumed to be zero. The parameter d stands for the elasticity of value added with respect to capital. The elasticity of net output (value added) with respect to employment (EV/EN) is variable and depends on the level of employment, i.e., EV EN= b + 2 c log N where a hat ' stands for the estimate 3/ of the corresponding parameter. The implied scale elasticity is E b + 2 c(lr N) + d = 4/ 1 + g + 2 c(ln N). There are increasing, constant or decreasing returns to scale depending upon whether the value of the elasticity (E) exceeds, equals or is smaller than unity, respectively. 1/ The ui are independent random variables with zero mean and constant varance. 2/ In order to avoid spurious results, the dependent variable is the logarithm of the ratio of two variables. 2 3/ The variable (log N) and its corresponding parameter C are intended to capture non-linearities in the production function as would arise, for instance, if small and large firms use different technologies or have differences in productive efficiency. 4/ See Griliches and Ringstad, op.cit., p. 88. - 106 - The estimated equations (See Table A-31) allow for the existence of variable returns to scale. The same industry may have increasing returns to scale up to a certain size, then constant returns and finally decreasing returns to scale in the large-size ranges. Of course, some industries may be characterized by only one type of returns to scale throughout the range of operations of firms, e.g., constant returns to scale. Table A-32 presents two estimates of the scale elasticity for firms of 25 and 100 workers in each of the 31 selected industries. The impact of the scale of the firm on productivity is captured by 2 coefficients of two variables: log L and (log L) . At least one of the corresponding coefficients is statistically significant in eight of the selected 31 sectors. Small-scale enterprises perform better than large firms in two of these eight sectors: mill products: wheat flour (31161) and footwear: women's leather shoes (32403). In five of the remaining six sectors there are strong indications of economies of scale: soft drinks (31343); sawmills: doors and windows (.33116); non-metallic household furniture: tables:and chairs (33202); pharmaceutical products: lecithin and enzyme (35221); and structural metal products: doors windows and frames (33131). In the case of men's clothing industries (32201) there are only slight indications of economies of scale. In industries where economies of scale prevail, SSEs would be expected to be less able to withstand competition from larger firms than would be the case under constant or decreasing returns to scale. In the remaining 23 sectors it is less adequate to characterize the returns to scale since the size variables do not present statistically - 107 - significant coefficients and consequently it is not possible to reject the existence of constant returns to scale. Nevertheless., a considerable majority of the industrial sectors being studied present some evidence of increasing returns to scale. The returns to scale evidence may -be comple- mented by comparisons of labor productivity. Given that the capital labor ratio tends to rise with size in most of the industries, some increase in labor productivity would be consistent with approximately constant returns to scale, but large increases in labor productivity would probably signal increasing returns. Table A-33 has classified the sectors according to the relative difference in labor productivity between firms of 25 and 100 1/ workers in the same industry. If it were assumed that nearly constant- returns to scale prevail if labor productivity in firms of 100 workers exceeds that of firms of 25 workers by less than 25%, then 9 of the 31 selected sectors would be characterized by constant returns. These, together with the three sectors where decreasing returns prevail, increase to 12 the number of sectors where SSEs would appear to have, if not a competitive advantage, at least not too large a disadvantage (Table A-32). In summary, the production function estimates provide an approxi- mate idea of the extent of economies of scale in 31 selected manufacturing industries in which SSEs have considerable participation. There are few sectors where diseconomies of scale seem to prevail. In the majority of industries, the scale variables are not statistically significant. Although statistical tests do not rule out the existence of constant returns to scale in these sectors, there are other indications of economies of scale in most of the selected industries. 1/ It is assumed that small and large firms have the same capital-labor ratio; as will be explained below, this may not be an important characterization. - 108 - Among the variables that explain labor productivity differences among plants, the capital-labor ratio is considerably more important than the size of the plant. Whereas the size of the plant has statistically significant coefficients in only eight of the 31 selected sectors, the coefficient of the capital-labor ratio is highly significant in most of these sectors. In principle this coefficient measures the elasticity of value added with respect to fixed capital, and in all sectors its value is bounded between 0.07 and 0.638. The lower bound is undoubtedly an underestimate, since the share of profits in value added can be expected to be higher. Nonetheless, in most cases where this coefficient is ifmplausibly low, its value is not statistically significant. In general, the results demonstrate that however imperfect, the measures of fixed capital assets are indicative of installed capacity in each establishment. A possible bias in the measure of fixed assets due to inflation has been partly corrected by including in the regressions the year when the firm started operations. Although no exact association can be expected between the age of a firm and that of its capital stock (due to additional investments), the fact that a large number of small firms are very young 1/ contributes to this relation and corrects for the relative undervaluation of older f4,xed assets due to inflation. Although not significant, the coefficieiLc of the variable 'year when the firm started operations' is negative in 20 of the 31 selected sectors. This means that ceteris paribus the younger the firm the lower its labor productivity. 1/ The age distribution of firms according to size is explored in Section - 109 - Technical efficiency is only one dimension of the enterprise, one of various determinants of private profitability and of social efficiency. Different firms may face different factor markets (e.g., pay lower wages or higher interest rates) and have different markets for their products. For instance, a small firm supplying a local market may not be exploiting the economies of scale that are associated with large production levels, but may nevertheless be able to compete adequately with large firms if its competitors face high transportation costs for reaching that market. These and other factors affecting the size distribution of firms will be analyzed in Section IV below. The above analysis does not address directly the question of relative efficiency of plants of differing sizes. Economies of scale may exist but if factor prices induce larger plants to use socially costly factor combinations, the large plants may be less efficient from an economic point of view even if more so in a certain technical sense. In general one would expect firms to fare less well when judged by the economic criterion than when judged by the technical one. -110- VI. THE DETERMINANTS OF THE SIZE DISTRIBUTION OF ENTERPRISES 1. INTRODUCTION In previous chapters the analysis has highlighted the importance of several factors which have had bearing on the development of SSEs in particular and the size distribution of manufacturing enterprises in general. In this chapter these several strands are brought together and the impact of economic policies of SSEs is briefly assessed. The con- cluding section analyses the scope of SSEs in Colombia. It is not possible to disentangle the impact of each of the forces that have contributed to the present size distribution of plants. Some of the main determinants may, however, be mentioned. At the aggregate level of the manufacturing sector, part of the dispersion of firm and plant size is due to different levels of economies of scale across industries. Within a given industry, the presence of absence of economies of scale plays a role. Plants are thus larger in petrochemicals than in clothing. But there is also a substantial range of sizes within most industries, a result of several factors. One is differential factor prices; firms facing a high price of capital and a low price of labor tend not only to have higher L/K ratios but also to be of smaller size, since in many industries it appears that the optimal L/K with a given factor price ratio decreases with output level, i.e., the production function is not homogeneous. A third factor contributing to a range of sizes, especially when economies or diseconomies of scale are_ not too prominent, is the range of ages of firms. Firms tend to grow - 111 - over time, the growth being facilitated by improvement of their market potential through greater contacts with and recognition by clients, by internal funds for growth generated out of past profits and by greater access to capital markets. The cited factors tend to produce a substantial diversity of firm and plant sizes. Whether average plant or firm size is high or low depends mainly on a country's level of development. When capital is scarce in general, and especially so for firms without gGud access to the credit market, it is both natural and (probably) desirable that many small firms exist, desirable at least to the extent that the hypothesized greater potential of small firms to employ labor intensive techniques holds true. As capital becomes less scarce and capital intensive techniques increasingly have a comparative advantage over labor intensive ones, larger firms are likely to be increasingly competitive (in the social sense) with small ones and the whole size structure of firms and plants is likely to move up. Such a shift of the size structure also tends to result from the evolution of firms to larger size,'not simply because their optimum size increases over time but because under imperfect capital markets their privately optimal size depends a great deal on the resources available for investment. In one sense it can be said that a large number of SSEs, especially in the informal sector where the barriers to entry are lower, arise as a response to the lack of employment opportunities in larger firms. (Indirect evidence on this can be obtained by the relatively low earnings that the self-employed are able to derive from these - 112 - activities). The lack of well paid jobs in the larger scale sector is, however, a natural result of the scarcity of capital in the system as a whole, so it is no accident that many workers be employed in SSEs with relatively low earnings. It is inevitable (the low earnings) and presumably desirable, as such SSEs appear to be relatively efficient users of.capital, the scarce resource. As noted above, productive efficiency and economies of scale have had an impact on the size distribution of plants in Colombia. Analysis of detailed industries has suggested that the extent of economies of scale may be fairly broad. 1/ This implies that in an industry characterized by economies of scale, the larger firms till be in a better position to undercut the smaller firms within the industry: their supply price will be lower than that of rivals. This process may be one cause of the higher mortality rates observed for SSEs. There is a limit to the advantages that economies of scale can confer to larger firms. If markets are fractioned by distances and by natural constraints imposed by mountainous terrain, transportation costs restrict the ability of large firms to undercut local producers. It can be claimed that in the case of Colombia a large number of SSEs owe their survival to the natural protection granted by the complex topographical 1/ Though, as discussed above, the data available to analyze this question is quite defective, and there are reasons to believe that our regressions have overstated true economies of scale (See p. ). -113- conditions of the country. A considerably larger percentage of the very small firms (1-4 workers) are located in non-metropolitan areas: 44.7% as compared with 22.2% of factory establishments. On the other hand, the large population shifts towards the maif urban centers that have taken place over several decades, can also be interpreted as a tendency to exploit the potential economies of scale that result from less frag- mented markets. In summary, it should be emphasized that any efficient manufac- turing sector is likely to have a wide range of firm sizes. Whether it is too wide in Colombia, or whether greater efficiency could be attained were more resources put to work in large firms or on small ones, is difficult to judge. 2. ECONOMIC POLICIES Colombia, like most developing countries, has lacked a unified policy towards small-scale enterprises. The implicit criterion has been to let market forces determine the extent to which SSEs would participate in a given line of production,. rather than precluding or promoting such a participation through direct intervention. Nevertheless, the environ- ment in which SSEs have developed has not been unaffected by government policies of a less general character. A. Credit and Technical Assistance Credit is one of the most important limiting factors for SSEs, and the terms under which it becomes available vary substantially among firms. 1/ Those firms that do not have access to institutional sources 1/ However, in many instances lack of credit is a symptom of poor performance and not its cause. -114- of lending, may have to borrow in extra-bank markets. The current monthly interest rate in the extra-bank markets varies from 3 to 5% - approximately twice as high as the nominal interest rates charged by official lending institutions - and is substantially positive in real terms. It is believed that as much as 20% of all credit may pass through the extra-bank markets, although it is not possible to verify this figure. The extra-bank market can fulfill a useful role for SSEs in supplying short-term credit require- ments. But any extensive use of this market for any but quite short term purposes is bound to represent a real drain on the resources generated by the firm and thus affect its viability. Frequently the need for funds is a short run one, e.g., to finance the purchase of raw materials. How often the unavailability of lower cost long run credit constrains the small firm is not clear. In order to increase the employment and output contributions of the SSE sector, it may be necessary to increase the life expectancy of small firms. A substantial number of SSEs are created every year, but it appears that many of these cease operating after a few years. 1/ Longer. life expectancy sometimes requires better productivity and improved ability to compete with other firms. Often it requires the availability of good information prior to the setting up of the plant, i.e., many firms which go under after a short period of time may simply have been badly thought through, such that success was never a very real possibility. 1/ As discussed below, the data probably exaggerate this phenomenon somewhat, but it seems nevertheless to be a real one. - 115 - Technical advice may be necessary to avoid such errors. Credit may be needed at various points of time. Similar to the 'infant industry argnment' for protecting some domestic industries against foreign compe- tition in their early years, an argument can be presented in favor of the 'infant firm'. Special technical assistance in their formative years - undertaken, for instance, by lending institutions - might significantly affect the survival of these firms. This measure, as well as others, could be used to diminish the gulf separating new and established firms and thus contribute to 'protecting' incipient SSEs. One of the principal results of the analysis of individual establishments has been the large variability in performance that exists among small firms. Furthermore, it has been established that SSEs have higher death rates than larger firms. These factors, together with the higher transaction costs involved in lending to SSEs, imply that even under perfectly-working capital markets, competitive interest rates would be higher for SSEs. Nevertheless, the differences in interest rates between bank and extra-bank markets appear larger than those which on the average would be required to compensate lenders for additional risk and transaction costs. In this sense, the differentials imply the existence of 'imperfections' in the capital market. These imper- fections have efficiency implications: in order to be financially viable, small firms may require a substantially higher gross 1/ rate of return than larger firms. 1/ Gross of interest payments. -116 - Nevertheless, the considerable dispersion in the capital-labor ratio among SSEs indicates that the incidence of capital costs may vary substantially. Those firms that are able to provide adequate guarantees can appeal to instituional sources of lending and avoid higher interest rates. Some are essentially self financed. In addition to favorable credit conditions, large firms have also had easier access to imported capital goods and raw materials until 1967. The quantitative restrictions that prevailed on imports discri- minated against smaller firms. 1/ In 1964 the share of purchased inputs which were imported rose from less than 10% for plants of under 20 workers to 26% for plants of 200 workers and up. 2/ While these figures probably exaggerate the difference in that small plants presumably buy a higher share of their inputs, including imported ones, from intermediaries (in which case imported inputs are not so.classified) it seems likely that a considerable difference has existed. How large it may be intra-industry is not clear. Credit conditions for SSEs have markedly improved after 1968, with the creation of the Corporacion Financiera Popular (CFP). In fact, it has been noted above there has been a decline in the mortality 1/ See C. Diaz-Alajandro, Foreign Trade Regimes in Colombia. 2/ Todd, Efficiency...... op. cit., p. 71. - 117 - rate of SSEs from some time in the late 1960's, especially among firms of less than 50 employees. The creation of CFP, the better economic environment that followed the trade liberalization policies of 1967 and the generally buoyant state of the economy may all have contributed to this. The smaller firms among SSEs have been assisted primarily by CFP and by Caja Agraria. The efforts of Caja Agraria have fluctuated over time and seem to have declined in recent years. The unexpected rise in inflation since 1974 has affected the resource position of major lending institutions, which have been forced to weigh their broad development assistance to SSEs with the necessity of generating adequate internal resources for future operations. 1/ B. Other Policies Generally, SSEs pay lower wages than larger firms. This feature appears to be more prevalent the smaller the firm. In addition, employment taxes and contributions (e.g., for social security) are more likely to be avoided in the case of SSEs. Thus, the cost of labor to SSEs is presumed to be lower than that of larger firms. Recent studies have suggested however, that wage differentials by size of firm or plant may be less than had earlier been thought. (See the discussion in Section above). 1/ In some instances, this has meant expanding the relative share of lending to the larger firms among SSEs, since lower transaction costs are involved. -118- Compliance with minimum-wage legislation is less strict among SSEs. Until 1971, there was in effect a lower minimum wage for smaller firms. Since then there has been a uniform minimum-wage, but real wages have declined over time, a fact which may have had a bearing on the significant factory-employment expansion that has taken place during the 1970s. Trade unions are permitted only in firms of at least 25 workers. This is another reason why wages among smaller firms have shown a tendency to be below those of larger factories. In Colombia there is no legislation on monopolistic practices or collusive arrangements. In this regard, large firms in industries with high concentration ratios can exercise their monopolistic power in the expense of smaller firms. 1/ Sometimes this power is exercised at the product market; more frequently, perhaps, the small producer's position is prejudiced by the exercise of monopoly power over the sale of intermediate inputs. A case in point is the control of intermediate inputs to small producers of leather products and packing boxes by a few large firms. In the case of some industries the policy of import substitution and the relatively small domestic market have fostered the development of a few large firms that account for a substantial percentage of the output generated in each of these industries. In fact, in a large number of well- defined industries (5-digit level), there is only a handful of firms. The barriers to entry in these industries are likely to be formidable, especially for SSEs. 1/ Several recent studies have discussed the degree of concentration, e.g., Superintendencia de Sociedades Anonimas, Conglomerados de Sociedades en Colombia, Bogota, Editorial Presencia, 1978; Gabriel Misas, Estudio de las Relaciones Entre Pequena y Gran Industria: El Caso Colombiano, BKogota, Colciencias, 19/8. Table A-1 OCCUPIED POPULATION BY ECONOMIC ACTIVITY AND URBAN-RURAL BREAKDOWN, 1973 Urban /1 Rural /2 Percentage Distribution Economic Sector ("Cabecera") ("Resto") Total "Cabecera" "Resto" Total Agriculture 344,949 1,726,650 2,071,599 16.7 83.3 100.0 Mining 21,419 26,147 47,566 45.0 55.0 100.0 Manufacturing 694,850 184,484 879,334 79.0 21.0 100.0 Electricity, Gas, Water 29,968 3,224 33,192 90.3 9.7 100.0 Construction 241,383 47,385 288,768 83.6 16.4 100.0 Commerce 765,303 105,380 870,683 87.9 12.1 100.0 Transportation 229,263 36,675 265,938 86.2 13.8 100.0 Financial Services 106,897 3,750 110,647 96.6 3.4 100.0 Social Services 1,011,945 123,120 1,135,065 89.2 10.8 100.0 Not Specified 12,737 5,050 17,787 71.6 28.4 100.0 Total 3,458,714 2,261,865 5,720,579 60.5 39.5 100.0 /1 /2 For definitions of "Cabecera" and "Resto" see Table Source: Estimates based on the latest revisions of population totals of DANE's XIV Censo de Poblacion, 1973 and National Household Survey for 1971. Table A-2 PER CAPITA INCOME POPULATION, GDP AND SECTORAL VALUE ADDED BY DEPARTfENTS, 1960 AND 1975 Per Capita Income Population Gross Domestic Product Dept'l Share Dept'l Share ($ of 1970) in Net /2 (Millions of 1970$) in Net 1960 1975 % Increase 1960 1975 Increase(%) 1960 1975 Increase(%j Bogota, D.E. 7,106 9,245 30.1 1,301.0 3,197.9 21.73 11,996.3 37,671.2 26.00 Meta 5,597 8,085 44.4 124.5 290.6 1.90 785.9 2,592.8 1.83 Territorios Nacionales 2,215 7,896 256.5 193.0 402.0 2.39 516.4 3,503.7 3.02 Atlantico 5,088 7,522 47.8 599.4 1,081.9 5.53 3,768.2 9,865.3 6.17 Cesar 5,525 7,325 32.6 200.4 509.7 3.54 1,285.4 4,048.3 2.80 Valle 5,271 7,080 34.3 1,503.2 2,486.9 11.27 9,677.9 20,804.3 11.27 Bolivar 4,602 6,720 46.0 600.8 954,1 4.05 3,249.3 7,321.9 4.12 La Guajira 3,536 6,368 80.1 131.5 254.3 1.41 500.1 1,676.3 1.19 Santander 3,988 6,207 55.6 912.6 1,230.6 3.64 4,407.4 9,000.5 4.65 Risaralda 3,951 6,187 56.6 391.3 509.7 1.36 1,805.6 3,574.6 1.79 Antioquia 4,064 5,945 46.3 2,143.5 3,310.3 13.37 10,637.1 23,772.1 13.30 1 Cundinamarca 3,531 5,933 68.0 1,051.6 1,284.5 2.67 4,832.3 9,020.2 4.24 Tolima 3,760 5,892 56.7 798.2 1,057.8 2.97 3,532.6 7,103.0 3.62 C Sucre 3,072 5,457 77.6 281.4 438.7 1.80 995.8 2,555,2 1.58 Magdalena 3,628 5,281 45.6 438.8 624.2 2.12 1,A12.0 3,765.5 1.98 Cordoba 3,885 5,178 33.3 486.4 818.5 3.80 2,128.6 4,664.1 2.57 Caldas 4,182 5,085 21.6 648.8 781.4 1.52 3,121.0 4,568.9 1.47 Quindio 3,802 4,458 17,3 279.4 375.8 1.10 1,244.7 1,880.1 0.64 IIuila 3,442 4,295 24.8 372.7 533.0 1.84 1,464.8 2,908.1 1.46 B,,-yaca 3,332 4,275 28.3 968.8 1,240.3 3.11 3,868.3 6,158.2 2.32 Norte de Santander 4,036 3,987 -1.2 482.6 794.2 3.57 2,251.7 3,593.4 1.36 Cauca 2,531 3,502 39.5 549.6 672.3 1.41 1,624.0 2,705.9 1.10 Narino 2,725 3,310 21.5 650.9 920.0 3.08 1,961.7 3,314.1 1.37 Choco 1,203 1,417 17.8 163.9 234.8 0.81 247.3 409.9 0.16 Colombia 4,205 6,207 47.6 15,274.2 24,003.0 100.00 77,714.4 176,477.6 100.00 /l The departments are listed according to their per capita income in 1975, which differ from GDP per capita because of government taxes and transfers, and ownership of productive factors. /2 The departmental share in the net population increase is given by the actual population increase between 1960 and 1975 in a given department divided by the total population increase for Colombia. Similarly for other variables in the table except for per capita income in which only the actual percentage increases with respect to 1960 are presented. Sou,ce: Departmento Nacional de Planeacion,Diueantas Regionales de Colombia 1960-1975, 1977. Table A-2 continued: PER CAPITA INCOME, POPULATION, CDP AND SECTRAL VALU' ADDED BY DEPARTMENTS, 1960 AND 1975 Value Added in Manufacturtn[ Small industry/1 Counnerce Personal Services Dept'l Share Dept'l Share Dept'l Share Dept'l Share (millions of 1970$) in Net in Net in Net In Net 1960 1975 Increase() 1960 1974 Increase(%) 1960 1975 Increase (.) 1960 1975 Increase(%) Bogøta, D.E. 2,892.4 8,223.4 26.06 308.7 591.2 17.85 2,167.5 6,531.6 26.20 1,296.4 4,112,3 39.28 Meta 47.3 169.0 0.60 23.6 84.4 3.84 16.4 191.4 1.05 60.7 127.8 .93 Territurios Nacionales 47.2 478.0 2.11 47.2 432.2 24.33 26.8 989.9 5.78 53.4 136.7 1.16 Atlantico 1,(80.9 2,614.4 7.50 147.8 118.4 -1.85 1,017.8 2,355.8 8.03 314.8 826.7 7.14 Cesar 102.1 183.6 0.40 86.3 96.6 .65 117.1 1,067.3 5.70 34.5 91.5 .79 valle 2>405.0 6,120.5 18.17 326.3 501.4 11.06 2,107.2 3,433.2 7.96 657.0 1,386.4 10.17 Bolivar 433.1 1,340.9 4.44 116.6 153.3 2.31 577.2 1,267.5 4.14 184.8 367.9 2.55 La Cuajira 17.5 86.8 0.34 17.1 80.2 3.98 187.4 970.1 4.70 16.5 39.7 .32 Santander 648.8 1,671.6 5.00 112.7 271.7 10.04 863.1 2,016.0 6.92 300.4 647.5 4.84 Risaralda 332.8 855.7 2.56 126.7 166.2 2.49 343.2 982.5 3.83 126.7 217.0 1.25 Antioquia 2,753.8 7,078.2 21.14 239.9 260,7 1.31 1>765.6 3,171.6 8.44 858.8 1,798.2 13.10 Cundluiarca 637.2 1,306.2 3.27 72.7 42.9 -1.88 144.5 415.2 1.62 591.0 959.4 5.14 Tolima 247.6 343.7 0.47 113.8 141.2 1.73 566.8 1,079.9 3.08 225.5 320.5 1.32 Suere 58.3 72,9 0.07 55.1 72.4 1.09 104.1 262.8 .95 50.8 82.6 .44 Magdalena 116.6 203.7 0.43 80.3 130.1 3.14 254.1 349.8 .57 110.7 168.1 .80 Ctirdoba 65.2 151.4 0.42 38.4 121.9 5.27 387.7 882.5 2.97 93.8 264.6 2.38 Caldas 291.9 726.1 2.12 99.9 131.2 1.97 864.2 834.7 -.17 249.5 323.8 1.03 QUi[idlu 144.8 236.4 0.45 75.2 98.8 1.49 179.1 268.5 .53 90.8 155.3 .89 lulla 144,0 243.0 0.48 88,0 135,3 2.98 207.8 318.3 .66 107.2 179.0 1.00 boyaca 432.0 832.6 1.96 69.1 67.8 -.08 188.9 619.3 2.58 233.8 305.7 1.00 Nurte de Santander 155.5 290.9 0.66 55.4 111.0 3.51 547.1 734.4 1.12 146.4 311.7 2.30 Cauca 166.8 327.9 0.79 62.6 81.9 1.21 64.1 272.8 1.25 81.8 109.3 .38 230.2 348.9 0.58 164.0 226.1 3.92 197.1 521.5 1.94 91.7 197.2 1.46 Choco 27.3 ?6.2 - 26.4 19.1 -.46 11.7 21.9 .06 17.1 32.5 .21 Coloinbia 13,478.3 33,-1,9i.U 100.00 2,553.8 4,136.0 100,00 12,906.5 29,558.5 100.00 5,994.1 13,161.2 100.00 /1 iludes household production and establishnmnts of less than 5 workers with annual production of less than ý24,000. Source: DLepartmento flacional de Planeacion, Cuentas Regionales de Colombia 1960-1975. - 122 - Table A-3: ESTIMATES OF THE JOB POSITION STRUCTURE IN MANUFACTURING, 1953, BY SIZE OF PLANT Plants of Reporting Residual Total /a 5 or more Plants of workers less than 5 workers (1) (2) (3) (4) Panel A Self Employed 13,652 50,634 146,254 210.54 Paid 175,030 24,350 75,180 274.57 /b Total 188,682 74,994 221,434 485.11 Panel B Self Employed 17,047 166 495 210.57 Paid 188,953 67,257 274.57 Total 200,000 232 752 485.11 Sources: For Panel A Col(s). 1 and 2 are based on DANE's 1953 industrial census, as reported in Boletin Mensual de Estadistica #72, Marzo 1957, pp. 16-17 and 23-24 Panel B are the authorts adjustment to the figures of Panel A. /a Interpolated between 1951 and 1964 figures. lb Assuming the percent paid was 56.6. - 123 - Table A-4: A COMPARISON OF THE 1953 SMALL SCALE SURVEY AND TOTAL COTTAGE SHOP IN 1953, BY INDUSTRY Estimated Total Sample Cottage Shop Food 5,890 9.1 9,831 /a 3.62 Beverages 520 0.8 1,206 0.44 Tobacco 1,910 3.0 3,649 1.34 Textiles 4,070 6.3 25,002 9.20 Clothing/Footwear 28,300 44.0 118,231 43.51 Wood 3,520 5.5 49,849 18.34 Wooden Furniture 5,050 7.8 Paper products 10 - 159 0.06 Printing 660 1.0 3,629 1.34 Leather 1,420 2.2 4,206 1.55 Rubber & products 100 0.2 - - Chemicals 1,040 1.6 2,476 0.91 Petroleum & Coal - - 341 0.13 products Non-Metallic Mineral 3,930 6.1 8,957 3.30 products Basic Metals 110 0.2 metal products, ex- cept machinery & 2,330 3,6 transportation equip. 116.27 Non-Electric machinery 160 0.2 9.2 44,214 Electrical machinery 1,030 1.6j Transportation 2,300 3.6 Equipment Various 2,130 3.3 Total 64,560 271,175 lb 100.0 /a The 1951 value was implausibly below those of both 1938 and 1964. /b Less than our estimate of total cottage shop since territories are not included due to lack of data. Sources: Data for the DANE sample is from BME #72, March 1957, pp. 15-17. The estimates of total cottage shop employment are an interpola- tion of figurps presented in Table 111-6. - 124 - Table A-5: COMPOSITION OF SMALL SCALE PRODUCTION EMPLOYMENT, 1951 AND 1973 Percent exclu- 1951 1973 ding Miscel- laneous & no Number Percent Number Percent Information Food 7,956 3.01 4.77 115,222 30.89 36.95 Beverages 1,073 0.41 Tobacco 3,555 1.35 J Textiles 25,564 9.68 Clothing & Footwear 117,937 44.68 55.93 114,023 30.57 36.56 Leather products 4,139 1.57 Wood products 48,398 18.33 37,061 9.94 11.88 Wooden furniture Paper products 162 0.06 1.39 6,035 1.62 1-1.94 Printing 3,504 1.33 J _J Rubber products 49 0.02) Chemicals 2,188 0.83 .90 7,329 1.96 2.35 Products of Petro- 132 0.05 J leum & Coal Non-Metallic Minerals 8,752 3,32 14,867 3.99 4.77 Base Metals 778 0.21 0.24 Metal Products, except Machinery & Equipment 40,575 15.37 16,528 4.43 5.30 Non-electric machinery Electric machinery Transportation equipment Miscellaneous & no information _ 61,174 16.40 - Total 263,984 100.0 373,017 100.0 100.0 Sources and Methodology: The data for 1951 are from TAble 111-6. For 1973 the data are a rough approximation, being the sum of household employment reported in the 1973 population census preliminary figures (adjusted as in Table A-1) plus the 1970 figures for small scale enterprises; these latter were not adjusted to estimated 1973 levels. - 125 - Table A-6: AVERAGE EARNINGS BY PLANT SIZE: 1953 AND 1970 1953 1970 Plant Size Index: Annual Wage Index: (Number of Annual Plants of Plants of Workers) Wage 25-49 Workers 25-49 Workers = 100 = 100 Very small (not 969 40.9 7,896 50.7 fulfilling definition of factory) Other < 10 (ful- 1,527 64.5 9,811 63.0 filling factory definition) 10-14 1,873 79.1 11,551 74.1 15-24 2,061 87.0 12,864 82.6 25-49 2,368 100.0 15,578 100.0 > 50 2,704 114.2 30,468 195.6 100-199 25,700 165.0 > 200 37,429 240.3 All > 5 workers 2,375 100.3 26,305 168.9 or > 24,000 gross produc- tion Sources: Data for 1953 are from A. Berry "Real Wage Trends...", Table A-13 and from B.M.E. #2. Data for 1970 are from Table A-25. - 126 - Table A-7: RELATIVE PRODUCTIVITY /a OF VERY SMALL ESTABLISHMENTS 1953 Plant Size Index: 1970 (Number of Thousands. of Plants 25-49 Thousands of Index: Plants of Workers) pesos per Workers = 100 pesos per 25-49 workers year year = 100 < 5, (and 1.60 23.3 16.11 39.9 output < 24,000 pesos) 5-9 and < 5 if 3.47 50.6 23.50 58.2 output >24,000 pesos 10-14 4.40 64.1 30.45 75.5 15-24 5.01 73.0 31.08 77.0 25-49 &.86 100.0 40.35 100.0 > 50 10.04 146.4 86.18 213.6 All Factory 7.53 109.8 72.49 179.7 /a Value Added per Worker. Sources: For 1953, Boletin Mensual de Estadistica #72, pp. 15-17, pp. 23-25. Table A-8: JOB POSITION STRUCTURE OF EMPLOYED PERSONS IN MANUFACTURING White Blue Independent Dates Collar Collar Employers Workers Family Workers Total 1971 National 190,338 297,048 37,655 211,690 37,104 774.837 (March-April) 172,842 243,349 34,340 136,354 25,332 613,219 Cabeceras Bogota* 78,155 44,660 9,625 24,640 6,545 163,625 (47.76) 1971 National 183,334 317,352 39,586 229,081 31,971 801,931 July-August) 163,615 258,535 33,865 150,691 18,341 625,654 Cabeceras Bogota* 54,080 48,672 8,788 21,632 3,380 136,552 1971 National 217,206 362,143 33,967 292,889 61,236 967,748 November-December 196,388 277,004 30,307 179,729 24,143 707,878 Cabeceras Bogota* 57,281 55,223 9,947 25,382 7,203 155,036 H (36.95) 1977 National 204,490 384,768 42,034 293,581 87,799 1,012,672 September-October (20.19) (28.99) Cabeceras 188,167 319,780 30,394 192,702 40,493 771,536 (24.39) (3.94) (24.98) (5.25) Bogota* 82,134 40,560 11,492 35,828 6,084 176,098 (46.64) (20.34) 1974 - June 437,163 Four Cities+- 159,208 (36.41) Bogota 96,104 65,928 13,120 35,424 2,952 213,528 (45-00) (30.87) (6.14) (1.38) 1975 174,623 205,934 19,673 75,194 6,757 482,181 Seven Cities A (36.22) (42.71) (4,08) (15,59) (1.40) Bogota 95,255 78,127 8,937 32,720 3,540 218,579 (43.6) (14,97) Table A-8 (continued) White Blue Independent Dates Collar Collar Employers Workers Family Workers Total 1974 - November 240,090 343,228 30,068 198,217 28,288 839,891 Cabeceras (3.58) (23,60) (3.37) Bogota 106,097 58,843 8,434 34,549 4,696 212,669 (16.3) Independent White Collar/Blue Collar Employers Workers Family Workers Total 1976 March/April 418,784 20,484 92,849 6,512 538,629 (77.74) (17.24) 1977 March Seven Cities 461,095 29,360 92,715 8,702 591,872 (77.90) (4.96) (15.67) (1.47) OO 1977 September 510,903 23,797 98,535 13,900 647,135 (78.95) (3.68) (15.23) (2.15) Bogota 236,540 15,358 39,766 5,493 297,157 (13.38) ) Percentages of total manufacturing employment. Cabecera. + Bogota, Medellin, Cali, Barranquilla. A Bogota, Medellin, Cali, Barranquilla, Bucaramanga, Manizales, Pasto - 129 - Table A-9: COMMERCE: SIZE DISTRIBUTION OF ESTABLISHMENTS OF AT LEAST FIVE WORKERS, 1970 No. of Workers Wholesale Retail per Establishment Trade Trade Total 1 -4 Establishments 64 418 482 Employment 218 1,407 1,625 5 - 9 Establishments 610 2,303 2,913 Employment 4,305 15,719 20,024 10 - 19 Establishments 637 1,410 2,047 Employment 8,656 18,644 27,300 20 - 49 Establishments 358 621 979 Employment 10,467 18,102 28,569 50 - 74 Establishments 58 110 168 Employment 3,521 6,731 10,252 75 - 99 Establishments 17 47 64 Employment 1,453 4,139 5,592 100+ Establishments 42 97 139 Employment 9,384 22,624 32,008 Total Establishments 1,786 5,006 6,792 Employment 38,004 87,366 125,370 Source: DANE, Census of "Large" Commercial Establishments, 1970, includes some firms of less than five workers. - 130 - Table A-10: JOB POSITION COMPOSITION OF SALESPERSONS, 1964 AND 1973 Independent Family White Blue Employer Workers Help Collar Collar Other Total 1964 19,113 147,158 9,798 110,750 - 2,035 288,854 (6.66) (51.31) (3.42) (38.61) 1973 74,417 171,604 7,285 145,774 43,970 443,050 (18.64) (43.00) (1.82) (36.53) - - ( ) Percentages of persons classified by job position. Sources: The population census of 1964 and the advance estimates for 1973, presented in BME, #289, Agosto, 1975, p. 28. Table A-11: SIZE OF ESTABLISHMENT, 1967: NHOLESALE AND RETAIL 1967, Wholesale Retail. . Total Number Percent Number Percent Number Percent Size of Establishment Number of of of Number of of of Number of of of (Number of Workers) Establishments Workers Workero Establishments Workers Workers Establishments Workers Workers (1) (2) (3) (4) = (7) -(1) (5) == .(6) 7), Q (8) (9) (8)-(2) < 5 5732 11,616 20.10 152,537 194,013 62.81 158, 269 205,629 56.08 5 -9 1404 9,534 16.50 6,501 41,121 13.31 7,905 50,655 13.81 10 - 19 727 9,587 16.59 2,009 26,013 8.4z 2,736 35,600 9.71 20 -49 385 11,384 19.70 773 21,333 6.91 11,158 32,717 8.92 50 -74 61 3,530 6.11 107 6,078 1.97 168 9,608 2.62 75 -99 29 2,3/12 4.05 62 5,238 1.70 91 7,580 2.06 100 40 9,792 16.95 72 15,117 4.89 112 24,909 6.80 Total 8378 57,785 100.00 162,061 . 308,913 100.01 170,439 366,698 100.00 Source: Berry, "Urban Labor Surplus..., op. cit. p. 30. Table A-12 COMMERCE: VERY SMALL ESTABLISHMENTS BY DEPARTMENTS, 1970 Employment Wages Sales Number of Fringe (in 000's of Department Establishments Total Paid Wages Benefits Pesos) Antioquia 15,003 26,756 7,593 70,592 3,795 4,400,817 Atlantico 8,762 15,151 2,733 21,904 311 1,710,764 Bogota 32,726 56,638 14,910 123,225 7,129 5,075,419 Bolivar 5,110 8,623 1,358 10,107 172 813,444 Boyaca 7,585 12,565 1,644 9,238 484 717,824 Caldas 4,006 6,864 1,762 11,639 1,124 1,100,411 Cauca 2,179 2,855 323 3,563 52 433,668 Cesar 1,566 2,309 279 2,224 50 174,816 Cordoba 2,419 3,992 1,072 6,707 474 527,327 Cundinamarca -9,309 16,027 2,672 15,300 1,412 1,398,624 Choco 827 1,228 190 1,373 48 283,344 Huila 1,994 3,272 546 3,686 132 321,581 La Guajira 752 1,244 277 2,180 19 192,990 Magdalena 2,533 4,454 665 4,746 56 426,877 Meta 2,558 3,929 608 4,595 168 319,228 Narino 3,225 5,843 845 5,974 153 406,122 Norte de Santander 5,084 8,439 1,380 10,058 354 803,932 Quindio 2,464 3,993 1,085 7,048 702 1,006,206 Risaralda 2,607 4,834 1,604 12,216 1,304 1,180,488 Santander 10,35,6 17,163 2,576 18,299 1,003 1,573,896 Sucre 1,614 2,906 421 2,410 45 313,451 Tolima 6,063 9,612 1,481 12,246 1,738 1,268,419 Territorios Nacionales 822 1,692 729 14,618 231 369,380 Valle 17,148 28,899 6,760 56,595 1,213 5,072,002 Total 146,762 249,288 53,513 430,545 22,170 29,891,029 /1 Defined as having less than five employees. Source: Unpublished DANE data. - 133 - Table A-13:. COLOMBIA: SERVICES: SIZE DISTRIBUTION OF "LARGE"' ESTABLISEMENTS 1970, BY TYPE OF ACTIVITY /1 Number of Employment Wages Establish- Paid Fringe Type of Activity ments Total Workers Wages Benefits Restaurants, Coffee Shops 1 - 4 165 578 366 2,086 161 5 - 9 1,436 9,808 7,650 42,628 3,868 10 - 19 713 9,171 7,982 51,323 5,143 20 - 49 171 4,623 4,342 32,702 4,049 50 - 74 17 1,011 972 10,552 1,505 75 - 99 5 424 418 3,554 388 100+ 9 1,451 1,435 18,987 4,344 Total 2,516 27,066 23,165 161,832 19,458 Hotels and Lodging 1 - 4 32 118 73 449 -28 5 - 9 227 1,514 1,149 6,854 697 10 - 19 97 1,256 1,011 6,412 588 20 - 49 51 1,513 1,432 10,794 1,492 50 - 74 13 779 751 6,013 1,159 75 - 9 8 686 682 8,329 1,470 100+ 12 2,447 2,426 43,114 11,095 Total 440 8,313 7,524 81,965 16,529 Transportation Services 1 - 4 3 10 9 88 6 5 - 9 34 249 216 2,462 309 10 - 19 19 269 239 2,668 478 20 - 49 13 349 329 5,897 711 50 - 74 1 62 60 379 48 100+ 2 667 564 13,966 8,662 Total 72 1,606 1,417 25,46.0 10,214 Storage Services 5 - 9 3 26 25 357 82 10 - 19 1 12 12 121 23 20 - 49 6 233 233 4,608 1,196 50 - 74 1 64 64 740 210 Total 11 335 334 5,826 1,511 Real Estate Services 1 - 5 11 33 21 140 12 5 - 9 50 337 287 3,468 502 10 - 19 32 405 373 4,691 793 20 - 49 13 368 355 8,331 1,406 50 - 74 1 69 69 784 185 75 - 99 1 79 79 3,420 826 100+ 2 297 153 5,815 1,558 Total 110 1,588 1,337 26,649 5,282 - 134 - Table A-13 (Continued) Number of Employment Wages Establish- Paid Fringe Type of Activity ments Total Workers Wages Benefits Publicity Services 1 - 4 3 10 6 63 8 5 - 9 24 163 137 1,950 264 10 - 19 20 264 248 3,729 615 20 - 49 9 274 256 8,637 1,503 50 - 74 4 239 237 13,089 2,151 Total 60 950 884 27,468 4,541 Services To Firms 5 - 9 3 19 14 96 9 10 - 19 4 49 44 253 21 20 - 49 4 125 121 1,933 369 50 - 74 1 70 70 137 23 Total 12 263 249 2,419 422 Machinery Rental 1 - 4 1 3 2 16 2 5 - 9 14 98 80 873 67 10 - 19 7 92 82 768 95 20 - 49 4 113 108 1,300 247 50 - 74 5 292 291 5,781 1,073 100+ 3 619 619 25,007 8,167 Total 34 1,217 1,182 33,745 9,651 Other Machinery Rental 1 - 4 3 9 6 118 6 5 - 9 1 7 6 74 2 10 - 19 6 77 70 1,102 183 20 - 49 3 68 68 830 173 Total 13 161 150 2,124 364 Social & Community Services 1 - 4 1 4 4 19 2 5 - 9 4 25 25 215 16 10 - 19 3 37 37 348 30 50 - 74 2 127 116 2,429 390 100+ 1 116 116 1,822 272 Total 11 309 298 4,833 710 Shoe Repair Shops 1 - 4 2 5 3 27 3 5 - 9 13 76 56 658 73 10 - 19 4 44 38 299 44 20 - 49 1 21 20 98 8 Total 20 146 117 1,082 128 Electrical Equipment Repair 1 - 4 24 75 49 393 24 5 - 9 61 381 301 3,247 276 10 - 19 27 336 291 2,853 411 20 - 49 12 371 361 4,061 800 100+ 3 357 357 4,966 1,063 Total 127 1,520 1,359 15,520 2,574 - 135 - Table A-13 (Continued) Number of Employment Wages Establish- Paid Fringe Type of Activity ments Total Workers Wages Benefits Automobile Repair 1 - 4 143 475 297 2,839 217 5 - 9 472 3,003 2,341 22,354 1,706 10 - 19 183 2,325 2,062 23,713 2,514 20 - 49 75 1,925 1,847 25,909 3,034 50 - 74 7 426 426 7,438 2,244 75 - 99 2 181 179 3,401 556 Total 882 8,335 7,152 85,654 10,271 Watch Repair 1- 4 1 2 1 8 5- 9 2 13 11 195 11 10 -19 1 14 14 120 8 Total 4 29 26 323 19 Other Repair Services 1- 4 21 75 42 366 35 5- 9 54 353 276 2,973 458 10 - 19 23 270 234 2,230 245 20 - 49 8 219 209 2,498 304 50 - 74 2 121 120 1,638 261 Total 108 1,038 881 9,705 1,303 Laundry Services 1 - 5 4 11 5 51 11 5 - 9 39 263 200 1,351 122 10 - 19 60 813 731 6,218 916 20 - 49 49 1,424 1,362 12,162 1,947 50 - 74 6 369 353 4,128 652 75 - 99 3 276 274 3,567 1,005 100+ 7 1,030 1,024 13,955 3,984 Total 168 4,186 3,949 41,432 8,637 Domestic Services 1 - 4 3 11 9 72 11 5 - 9 4 28 23 108 7 20 - 49 2 57 56 525 101 100+ 2 518 518 2,533 503 Total 11 614 606 3,238 622 Photography Studios 5 - 9 3 23 19 135 18 10 - 19 1 16 14 345 - Total 4 39 33 480 18 Other Personal 1 - 4 1 4 4 37 6 5 - 9 4 23 19 167 11 10 - 19 1 12 9 112 12 Total 6 39 32 316 29 - 136 - Table A-13 (Continued) Number of Employment Wages Establish- Paid Fringe Type of Activity ments Total Workers Wages Benefits Total 4,609 57,754 50,695 530,071 92,283 1 - 4 418 1,423 897 6,772 532 5 - 9 2,448 16,409 12,835 90,165 8,498 10 - 19 1,202 15,462 13,491 107,305 12,119 20 - 49 421 11,683 11,099 120,285 17,340 50 - 74 60 3,629 3,529 53,108 9,901 75 - 99 19 1,646 1,632 22,271 4,245 100+ 41 7,502 7,212 130,165 39,648 /1 "Large" establishments according to DANE's definition, are those that employ at least five workers, although the data include some which are smaller. Establishments of less than five workers are covered more fully in Table Source: Unpublished DANE data. Table A-141 SERVICES: EMPLOYMENT IN VERY SMALL ESTABLISHMENTS, 1970 Number of S.S. Income from Establish- Employ- Wages Payments Services Sector ments ment (000's of 1970 $) Costs Rendered 6310 Restaurants, Coffee Shops 28,023 59,858 6320 Hotels, Lodging 3,418 8,969 63 Total 31,441 68,827 133,292 3,601 1,334,215 4,056,611 7191 Transport Services (including Travel Agencies) 81 243 7192 Warehouses, Storage 167 416 71 Total 248 659 2,673 132 13,109 262547 8310 Real Estate 342 950 8325 Publicity Services 322 781 8329 Other Services Rendered to Industries 34 78 8330 Machinery Rental (Heavy) 31 73 8331 Other Machinery Rental (Light) 91 165 83 Total 820 2,047 11,348 751 126,775 213,356 9511 Shoe Repair Shops 3,865 5,423 9512 Electrical and Mechanical Repair Shops 4,570 7,878 9513 Automobile & Motorcycle Repair 6,808 16,851 9514 Watch Repair Shops 1,855 2,580 9519 Other Repair Services 2,519 4,385 9520 Laundry Services 3,513 5,408 95 Total 23,130 42,525 10,261 2,966 365,306 913,872 Total 55,639 114,058 257,934 7,450 1,839,404 5,210,385 Source; Unpublished DANE data, /1 Table A-15 SERVICES: ESTABLISHMENTS OF 1 TO 4 EMPLOYEES, BY DEPARTMENTS, 1970 Employment Wages Sales Number of Fringe (in 000's Departments Establishments Total Paid Wages Benefits of Pesos) Antioquia 8,236 16,943 8,352 48,196 1,268 2,449,271 Atlantico 2,013 4,780 2,407 13,609 215 169,865 Bogota, D.E. 15,416 31,354 16,818 72,037 3,301 937,511 Bolivar 867 1,962 894 5,682 64 61,518 Boyaca 1,235 2,455 1,289 4,843 172 49,777 Caldas 2,157 4,213 2,304 7,721 337 96,223 Cauca 626 1,088 359 11640 8 23 944 Cesar 366. 767 456 13217 38 15 850 Cordoba 390 878 568 2,436 22 38,237 Cundinamarca 2,318 4,198 1,412 5,614 97 87,735 Choco 315 539 272 934 9 19,499 Huica 914 2,026 891 3,255 109 47,185 La Guajira 83 227 183 585 - 73651 Magdalena 752 1,539 804 2,205 17 49,023 Meta 916 1,714 666 2,412 46 46,550 Narino 1,523 2,647 997 3,377 3 60,394 Norte de Santander 1,465 3,401 2,102 93877 179 100,973 quiuid.9 1,309 21522 1,401 53019 312 76,296 Riearada ¾480 3,169 1,564 6,636 278 98,514 Santander 2,786 6,948 4,255 17,923 434 176,781 Secre 145 348 150 826 18 11,708 Tolia 2,368 4,471 1,872 6,635 283 103,319 Territo4Qs N4conales 166 331 186 2,359 30 16,588 Ylle 7,662 14,962 7,505 32,895 210 465,941 Total 55,508 113482 57,707 257,933 7,450 5,210,384 Source: Unpublished DANE data. - 139 - Table A-16: COLOMBIA: 1953 INDUSTRIAL CENSUS: FACTORY SECTOR 5-9Ll 10-14 15-24 25-49 50+ Total Number of Establishments 7,959 1,189 849 634 612 11,243 % 70.8 10.6 7.5 .5.6 5.4 100 Employment 41,735 13,699 15,870 21,624 106,188 199,116 % 21.0 6.9 8.0 10.9 53.3 100 Output 410,134 180,106 266,126 641,315 2,342,526 3,840,207 % 10.7 4.7 6.9 16.7 61.0 100 Value Added 144,774 60,248 79,581 148,371 1,065,723 1,498,697 % 9.7 4.0 5.3 9.9 71.1 100 /1 Includes those firms of less than five employees that had a total production of over $Col. 24,000 in 1953. Source: DANE BME No. 72, 1957, pp. 16 and 23. The figure for employment in plants of 5 or more workers presented in Table 111-2 is higher than that shown here since this table includes over 10,000 workers in plants of less than 5 workers. Table A- 16.a: FACTORY EMPLOYMENT BY DEPARTMENTS, SELECTED YEARS /2 Percentage Distribution Years Years DeparIII"i:s 1953 1960 1966 1971 1975 1953 1960 1966 1971 1975 Antioquia 47,278 64,328 76,121 98,399 119,639 23.7 25.4 25.4 25.1 24.3 Atlantico 18,344 23,083 27,308 31,015 33,863 9.2 9.1 9.1 7.9 6.9 Cuudinamarca (incl. Bogota) 47,859 76,535 90,787 131,522 169,172 24.0 30.2 30.3 33.5 34.4 Valle 34,729 42,220 52,439 69,471 83,869 17.4 16.7 17.5 17.7 17.0 Bolivar 6,036 5,528 5,103 6,383 8,850 3.0 2.2 1.7 1.6 1.8 Boyaca 3,340 4,941 5,059 7,295 9,248 1.7 1.9 1.7 1.9 1.9 Caldas 5,717 8,201 10,764 1.9 2.1 2.2 /3 12,134 12,294 6.1 4.92.3332 Risaralda 126,444 12,790 15,605 2.2 3.3 3.2 Cauca 1,550 1,434 2,292 1,174 4,393 .8 .6 .8 .3 .9 1 Huila 1,010 885 1,078 1,465 2,082 .5 .3 .4 .4 .4 Magdalena 1,3N0 1,330 1,539 1,339 1,977 .7 .5 .5 .3 .4 C Narino 2,772 3,093 3,432 2,870 3,335 1.4 1.2 1.1 .7 .7 a Norte de Santander 3,451 2,800 2,678 2,320 2,635 1.7 1.1 .9 .6 .5 Santander 12,461 10,586 12,365 10,606 13,756 6.3 4.2 4.1 2.7 2.8 Tolima 5,181 3,012 2,918 3,389 5,613 2.6 1.2 1.0 .9 1.1 Rest 1,581 1,318 3,935 4,447 7,595 .8 .5 1.3 1.1 1.5 Total 199,116 253,387 299,215 392,686 492,396 100.0 100.0 100.0 100.0 100.0 /1 The data have not been adjusted to exclude firms of less than five workers. However, since the employment in these very small firms is a small fraction of total employment.the overall results are not expected to be biased significantly. 1 /2 1953, 1960 and 1966 based on DANE data. 1971 and 1975 based on ICSS data. /3 Department creaied in 1966. Sources: 1953 - DANE, II Censo Industrial, 1953. 1960 - DANE, BME 141 (December 1962), Pages 28-9. 1966 - DANE, Industria Manufacturera in BME 211, Pages 52-61 and BME 212, Pages 32-59. 1971, 1975 - ICSS, reprinted in DANE, UME 301 (August 1976), Pages 177-185, Tables 4 and 5. Table A-1.7. -Si 1STIBUT0N SIF ESTABLISIMENTS IN MANUFACTURING, BY SECTORS, 1956 Number of Workers per Establishment 1 - 4 5 - 9 10-14 15-19 20-24 25-49 50-74 75-99 100-199 200- Total Manufctluring Se-ctor Establishmuents 1oud 856 906 274 112 67 151 58 24 24 15 2,487 Bverages 39 59 29 13 6 29 6 18 22 8 229 Tubacco 35 86 52 29 12 22 7 4 5 4 256 TcxLIles 162 107 32 29 14 53 22 6 21 34 480 Footwear/Clothing 845 702 217 115 69 139 46 24 24 9 2,190 Wood 140 178 44 - 15 4 16 3 5 5 5 415 Furniture 77 227 47 21 4 21 7 4 1 1 410 Paper 3 13 6 4 3 17 8 - 5 1 60 PrinLing 77 136 56 36 15 49 10 2 10 6 397 L.eather 97 61 25 9 7 16 7 3 5 4 234 - Rubber 2 11 13 8 1 5 1 - 1 5 47 hiemicials 134 101 52 28 27 49 20 11 13 7 442 Oil and Coal Products - 2 2 2. 4 2 - - - 2 14 Non-MLtallie Minerals 253 292 90 66 27 60 25 5 24 14 856 Basie ietals 2 21 7 8 4 6 1 1 1 4 55 Metal Products 56 141 53 30 25 47 12 6 9 4 383 Non-Electrical Machinery 25 54 38 10 6 10 5 - l 1 150 Electrical Machinuery 33 44 11 18 15 19 8 - 2 2 152 Transport Equipment 103 158 52 19 13 26 7 5 6 7 396 otlier 38 62 24 10 10 21 6 5 5 1 182 Total 2,977 3,361 1,124 .582 333 758 259 123 184 134 9,835 Source: Ulnpublishd DAN[ data from Annual Survey of Hanufactures, 1956 and DANE Statistical Yearbook 1957, Table 376, Page 541. Table A-17b: SIZE DISTR3IBUTION OF EMPLOYNT IN MANUACTURING, BY SECTORS, 1956 Number of Workers per Establishment 1 - 4 5 - 9 10-14 15-19 20-24 25-49 50-74 75-99 100-199 200+ Total Manufacturing Sector Employment hood 2,369 5,957 3,013 1,856 1,527 5,093 3,288 2,044 3,304 5;948 34,399 1everages 107 370 345 209 136 989 360 1,539 2,983 4,898 11,936 Tubacco 88 638 618 460 237 739 409 336 670 1,255 5,450 Tex1tles 509 696 380 437 310 1,817 1,365 522 3,049 27,413 36,498 Footwear/Clothing 2,500 4,498 2,416 1,886 1,524 4,784 2,729 2,021 3,505 3,253 29,116 Wood 437 1,111 509 240 86 581 168 420 691 1,625 5,868 Furniture 189 1,523 560 360 87 684 417 366 137 533 4,856 Piaper 6 88 74 73 64 554 489 - 779 842 2,969 Printing 362 901 661 605 335 1,603 590 171 1,401 2,082 8,711 Lcalher 314 387 300 155 146 560 422 263 671 1,077 4,295 Rubber 8 74 160 133 22 198 62 - 178 3,946 4,781 Chemcals 560 631 604 471 575 1,726 1,203 945 1,889 2,544 11,148 Oil and 'oal Pr,duteLs - 10 22 33 80 70 - - - 1,812 2,027 Nun-Metalllc Minerals 751 1,943 1,085 1,094 651 1,938 1,551 415 3,420 6,120 18,968 Basic Mt±tals 5 142 87 140 87 208 61 93 175 3,909 4,907 iMet41 Products 171 868 634 513 549 1,728 736 536 1,284 1,197 8,216 N4un-EleteLrical N-ichlinery 79 370 441 174 130 309 255 - 143 301 2,202 Electric,) Mahelinery 102 278 134 293 338 620 445 - 280 572 3,062 Transport Equipient 344 1,180 608 314 282 853 374 407 885 3,668 8,915 Utr 101 419 287 168 220 784 363 433 647 223 3,645 To tal 9,00, 22,084 12,938 9,614 7,386 25,838 15,287 10,511 26,091 73,218 211,969 Sourcu: Unpublished DANE data from Annual Survey of Manufacturers, 1956 and DANE Statistical Yearbook, 1957, Table 376, Page 541 /1 Table A-17: MANUFACTURING: VALUE ADDED IN THE FACTORY SECTOR ACCORDING TO SIZE OF ESTABLISIMENTS, 1956 Number of Workers pýer Estabishnnr Manufacturing Sector 5-9 10-14 15-19 20-24 25-49 50-74 75-99 100-199 200 Total Food 33,761 20,613 16,971 10,721 55,638 49,277 33,177 49,299 89,392 358,579 Bevcagces 1,971 2,859 1,851 764 23,313 11,137 79>862 99,684 136,071 357,512 TobaCco 1,896 1,461 1>161 1>767 5,127 8,529 11,509 18,556 130,966 180,972 Text les 2,675 1,774 3,055 2,155 11,472 7,066 2,337 24,500 324,209 379,243 Clothing/Footwear 15,692 8,267 6,696 5,285 17,886 11>036 8,174 20,718 28,439 122,193 Wood 5,642 1>560 981 617 2,638 1,270 1,827 3,421 2>967 20,923 Furniture 6,128 2,125 1,504 410 3,148 2,598 1,968 193 3,396 21,470 paper 331 580 201 237 4,326 3,493 - 4,341 12,190 25,699 Printing 3,962 2,747 2,904 1,341 10,043 4,940 2,199 14>208 29,280 71,624 LeJtier 1,482 1-,361 762 786 2,912 4,036 1,274 5,686 16,227 34,526 Rubber 734 1,366 1>100 270 640 734 - 1,233 52,190 58,267 Chlmilcals 4,137 5,071 4,151 5>227 19,182 17,874 11,242 30,704 45>643 143,231 Oil & Coal Products 74 1,569 201 1,702 2,411 - - - 54,183 60,140 Nun-MeLallie Mineral Products 6,738 3,570 3,637 2,500 10,346 7>576 2,017 23,665 80,863 140,912 Basie Mecals 1,242 503 1,580 484 1,890 573 568 907 39,907 47>654 Mekýtal Produets 3,215 3,136 3,301 2,905 10,402 4,684 3,270 10,604 10>284 51,801 Non-Electrical Machinery 2,697 1,746 750 759 2,01 1,363 - 918 5,065 15,309 Electrical Machiinery 1,657 1,157 1,687 2,719 5,058 4,503 - 3,038 4,209 24,028 Transport Equipment 4,215 2,728 1,646 1,291 4,218 2,120 2>000 5,701 19,387 43,306 Other 1,957 1,538 933 2,237 7,239 6,131 3,556 3,749 1>909 29,249 Tktail 100,206 65,731 55,072 -55,072 199,630 148,940 164,980 321,125 1,086,777 2,186,638 /I DANE d.1La excludlng tirns of less Lhan f:ive workerp. Souuee; U :, A\pu riu GeniirA de EstadiLi ca, 1957, Table 377, page 541. Table A-17c: EMPLOYMENT IN THE FACTORY SECTOR, 1966 Nanufac turing Sector 5 9 10-14 15-19 20-24 25-49 50-74 75-99 100-199 2001- Total Food 6,613 4,094 1,335 1,448 5,277 2,679 2,492 5,745 11,086 40,769 Beverages 331 227 190 133 628 857 704 3,729 9,453 16,252 Tobacco 397 322 117 133 64 223 162 415 1,556 3,389 Textiles 705 823 183 237 2,464 1,621 613 2,598 36,252 45,496 CIoLhing/Footwear 4,286 3,090 920 1,003 4,238 2,819 2,639 3,741 7,440 30,176 Wood 1,035 827 250 288 866 269 95 1,045 1,502 6,177 6"rniture 1,083 627 273 233 986 225 166 405 63 4,691 PapeLr 133 369 194 156 623 595 484 1,082 2,685 6,321 Printing 1,103 1,476 326 588 1,451 482 565 1,381 4,440 11,812 Leather 416 466 121 151 307 191 181 *854 1,528 4,215 Rubber 104 118 85 193 460 297 180 - 5,482 6,919 Chemicals 846 898 277 371 2,420 1,661 1,634 4,822 9,188 22,117 Oil and Coal Products 41 57 - 45 64 51 91 .186 1,535 2,070 Non-Metallic Minerals 2,397 1,735 525 817 2,912 1,328 1,265 2,722 11,624 25,343 Basic Metals 19 155 32 46 63 65 - 790 2,775 3,945 Metal Products 1,424 1,670 644 889 3,106 2,093 1,445 3,716 5,300 20,287 Non-Electrical Machinery 606 711 253 216 518 612 345 522 1,747 5,530 Electrical Machinery 528 532 229 305 1?157 630 752 2,028 4,176 10,337 Transport Equipment 1,498 1,586 358 437 1,234 791 1,178 1,030 6,101 14,213 Other 566 673 184 284 1,353 1,361 1,069 1,881 1l222 8,593 Total 24,131 20,474 6,496 7,973 20,191 18,850 16,060 38,692 125,785 288,652 /I DANE Data excluding firms of less than five workers Source: DANE: Industria Manufacturera, 1966. Table A-17c: MANUFACTURING VALUE ADDED IN THE FACTORY SECTQR ACCORDING TO SIZE OF ESTABLISHMENTS, 1966 /1/2 Manufacturing Sector 5-9 1014 15-19 20-24 25-49 50-74 75-99 100-199 200+ Total Food 136,121 129,497 47,905 73,319 311,753 131,418 131,210 389,148 845,667 2,196,039 Beverages 5,189 4,840 7792 5,402 25,797 60,474 39,615 710,014 937,262 1,796,385 Tobacco 2,654 2,368 1,181 996 468 13,740 6,422 87,838 442,540 558,207 Textiles 6,751 17,541 2,795 6,437 54,594 41,403 18,593 74,685 1,529,209 1,752,108 Clothing / Footwear 39,678 32,713 14,342 15,917 78,743 47,224 65,938 80,485 253,071 628,110 Rood 11,500 10,332 4,072 4,656 18,645 4,664 5,310 22,811 61,133 143,124 Furniture 10,329 7,447 4,424 4,194 17,467 4,855 2,558 9,252 34,110 94,638 Paper 2,601 5,231 4,483 6,270 20,048 49,377 33,814 63,071 230,924 415,820 Printing 11,696 19,111 6,040 10,404 35,207 18,212 17,347 42,304 232,307 392,628 Leather 4,987 ',210 3,402 3,630 5,246 6,603 7,605 39,937 89,442 166,063 Rubber 1,160 2,267 3,217 5,002 14,540 10,202 7,849 - 316,916 361,154 Chemicals 27,394 19,808 9,643 17,560 225,826 118,978 116,665 430,800 748,700 1,715,375 Oil & Coal Products 2,744 3,020 12,757 5,936 6,362 9,945 5,014 20,503 335,826 389,351 Non Metallic Minerals 20,989 20,021 - 13,536 59,406 31,335 48,640 106,883 568,592 882,158 Basic Metals 243 2,918 2,206 2,312 2,032 3,527 - 75,237 135,240 223,714 Metal Products 19,644 23,185 10,305 22,150 82,198 67,529 62,879 155,476 217,879 661,234 Non Electrical Machinery 9,931 11,316 4,983 6,200 10,703 15,300 9,658 22,376 65,370 155,837 Electrical Machinery 8,355 7,991 4,559 9,882 44,448 31,266 39,747 76,408 295,788 518,443 Transport Equipment 14,757 20,742 8,052 7,896 27,358 21,146 31,205 27,076 113,633 271,870 Other 7,766 11,469 3,470 6 071 78,735 44,571 36,501 104,578 53,631 346,791 Total 344,492 357,034 155,630 227,759 1,119,575 731,767 686,569 2,538,884 7,507,340 13,129,049 /1 Iin thousands of current pesos. /2 DANE duta excluding firms of less than five workers. Source: DANE, Industria Manufacturera, 1966 /1 Table A-18 NUNBER OF ESTABLISIIMIENTS IN TifE FACTORY SECTOR, 1966 Manufacturing Sector 5 - 9 10-14 15-19 20-24 25-49 50-74 75-99 100-199 2001- Total Food 1,026 344 80 66 155 44 29 40 29 1,813 Beverages 51 20 11 6 17 15 9 28 20 177 Tobacco 60 28 7 6 2 4 2 3 5 117 Texciles 106 72 11 11 66 27 7 20 41 361 Clotliig/Footwear 658 266 56 45 124 46 31 28 18 1,272 Wood 164 70 15' 13 25 4 1 7 5 304 Furniture 163 53 16 11 29 4 2 3 2 288 Paper 19 30 12 7 16 10 6 8 4 112 Printing 169 123 19 27 42 9 7 10 9 415 Leather 64 38 7 7 8 3 2 6 4 139 Rubber 13 10 5 9 14 5 2 - 6 64 Chemiceals 130 77 16 17 70 27 19 37 26 419 Oil and Coal Products 6 5 - 2 2 1 1 1 2 20 Noti-etallie Minerals 361 147 31 38 83 23 15 20 28 746 Basic Metals 3 13 2 2 2 1 - 6 6 35 Metal Products 212 142 39 41 91 36 17 29 15 622 Non-Electrical Machinery 87 59 15 10 16 10 4 4 5 210 Electrical fachinery 75 45 14 14 32 10 9 16 9 224 Transport Equipment 231 136 21 20 36 13 14 8 10 489 Other 84 56 11 13 39 23 12 14 4 256 Total 3,687 1,734 388 365 869 315 189 288 248 8,083 /1 DANE Data excluding firns of less than five workers. Source: DANE, Industria Manufacturera, 1966. _ Т г t i k � "Га<,1е ;4-1tl�� С0lонысл: S1ce итs�lиlвиrlод ое escл.вllstu+rмcs ли0 t1и�lnУиега• (зиыk 1975) [цииLсс оЕ Еирlиуиея pcr Ечсыыllчtипиnс � _ /L 1- 4 5- Ч 1J-1L 1$-19 20-45 50-'l') i00-19Ч 2UUг 1'ucal ечкгь- �,р1ау- е5саъ- �,�р1оу- Еысаь- трlоу- ев[аВ- >:агрlоу- евсыь- ешрlоу- еч[ыь- /еп�рlоу- еяеад- /егырlоу- еысаь- lernptuy- 'еасиь- /1:шрlоу- t:�,uuiнtcuгl� Sг.,wr�- 11ч;ипепС( шспс 11s11meпt/ тепЕ 11s1идигс) шепс 11ч1ияеn[/ юопс 1/ahment/ шипС 11s1гшепсl wenc llчЬшепд weru 11_lwrenc шепс 1tsluw:nc/ шсис Fцид 1,7б4 3,760 705 4,Ь7б 2Ч7 Э,479 171 2,832 350 10,5ы8 1А 1u,.122 77 10,'!44 64 2'2,4ы2 Э,57Ч бы,7ы7 Hevurцgus 2Э 58 15 Чб 8 96 5 87 26 д5б 20 1,41Э 32 � 4,406 30 1Э,б$ы I59 20,670 'fцLauco 17 97 4 2Ч � 4 4д 2 ЭЬ 9 2ыЭ Э 181 5 655 $ 1,ЧSD 4Ч 9,221 ' Тсх и]иы 257 55б 152 1,и10 • 82 Ч97 45 7б4 129 4,0}9 ЬЬ 4,502 46 б,Ов7 б1 52,ы?) В44 70,ыы4 C1uchingjfuatиcar 1,двб 1,926 722 4,744 299 3,499 � 16tl 2,829 Э8в 12,166 129 в,757 62 ы,Т15 5l 20,ы1б 1,705 б5,51'1 ЧоиJ 409 94В 20Э 1,Э79 63 Т11 Э8 Ы5 42 1,272 6 412 6 71д 6 2,51ы 77Э д,5>9 Furnlcuec 907 1,9ыЭ 371 2,15] 120 1,472 57 �Ч46 11д 7,761 э4 2,Э09 14 1,956 1 ),l72 1,584 15,772 рлреr 9б в1 90 19б 16 154 3 50 29 98i 13 ду7 9 1,147 tl Э,ь57 144 7,'tU7 есlисlпь 55в 1,27о 29] 1,92д 127 1,А0 ь4 1,оа9 1Ь9 з,59s Э5 2,41s 2й 3,1зs 1Э Б,729 1,2ы7 21,ы 1 г lсцыкг 2и9 424 вб sn 42 489 28 47о 47 1,502 2э 1,640 ц 2,2ы7 ы Э,2ы5 kбo 10,669 ,� g tw6бur Ь7 140 '?б У71 11 132 11 190 37 1,048 11 Л9 2 3.2Э 9 5,6Ы 174 ы,444 iI , CheпdcaLs 269 612 15б 1,U29 73 84D 54 914 1э$ 4,213 59 4,9b5 60 д,ы5l 42 15,327 ы4д зS,9$5 Ог1 .,пд wu1 2$ б0 s 3о з эз 1 1в 5 119 з 189 1 105 2 J,ы2s 4s 4,эТ9 tlец-мсс.л171е !ц пехлls 609 1,эб7 24l 1,59в д1 947 59 1,000 159 4,95б 59 4,ОЧ5 24 3,э1д Э2 14,526 l,2Ь4 Эl,tl00 ииslс миtцlч 171 390 ыб 58ы _ 47 548 Э1 526 62 1,94Ч J3 2,2вЭ 17 2,011 11 в,2ЕЭ 45б 16,57ы . Nи[ц1 ргцдиисч 4J8 2,Э00 465 7,tl50 22$ 2,6)2 1э0 2,18Ч 295 9,124 110 7,эЧ! 58 7,940 Э5 14,474 2,25б 49,100 Nuu-Htuc[r1ca1 Nлufцnery 1ы$ 405 1l5 76ц 58 684 44 7:f9 8t 2,Э77 1Ч 1,'167 10 1,229 4 t,65L 5!6 'J,l'26 - eluccrlerl k1лeLlnury 750 1,582 275 1,в23 110 1,'2)9 Ь$ 1,112 132 Э,Ч4д $1 7,42Э 26 Э,46д l7 6,27'2 2,426 22,50G 'lгаиslюг[ Г:циlртсп[ 961 2,U72 7Б9 2,462 150 1,755 66 1,104 137 ],SЧб 7� 2,1i4 14 1,ы22 l4 7,в42 1,7.3Ч 23,107 uclmr 6t0 7,Э47 211 1,404 93 1,ОЧ0 5б 946 140 4,Эб5 35 Э,7ыБ 7'! 4,Ь7Ь 19 G,Gyй 1,2l9 'L4,2Бы 7:лса1 10,6$1 29,31tl й,�4Чb 29,об7 1,90Ч 22,45д 1,098 1ы,456 2,4ыб 74,702 9!2 G2,5в2 5ЭЭ 7Э,142 442 2lЭ,440 '2'2,527 57Т,765 1! "lLe 20U/ чlzе ee[egary 1ncludea ап црныril edjusewunc Со ыссципс For ету7аушеп[ [п р��Б11с encerpclse5, чоше оЕ иh1еЬ и е 1iчСид 1п 114t1L UnJnc "J1C1cLe1 Еисlсiсч" Lu[ пгс пос tгмlигlсд tu 1CSS J�to, '('Ьи adJusCmunt rиргсыепс5 ипlу 12 flnun апд 9,673 ешр7оуеев, дlвсгlЬиСед ач fо17оЧ¢: ВеVьгаgеч 4 апд 4,0$5; 011 <пд Coal�praduccs 1 anJ 7,ЭО1; СlгтLсаlч 4 апд А,27б) TrnnsporC Equlryпепс 3 unJ 1,041. :й.игсес AJjuчc�J т�д unpuBllshud 1C5S �а[и. Тде бихlе ддсы Еог June 1975 ецсер[. Еиг Cht Uоригсшепtч аЕ Аи[1oqyla апд 4�11k- ипдс6 +п•е For lюcciaLer 1975 tlис+г: 1иг ч и lыдичсхlе5 tLiч cnL1e шау unJec5eacc cuwl иирlцрпеп[, us fuilged ду ептрасlвоn н([ir OaNk flgurcч fur 197$- ',71е шлlп ехлтрlи 1ч tuxeilus. нУiиrе the 1C5S даси чhои lонег еарlыутипС 1еvеlч tиг р).�т[ aiпs. Т1ги r1Lf1c[ence сбгаlд Le explatncd 1Р [qere uas,u rapid етрlиутепд fncreasc Gцсчсеи Junu (1C5S Jata) и:,д NovewLcr (11АыС длгп), - � � . . . , ч �' ь � � t - �t:�Llu;1-1КЬс pdNCB17l'AGE G15•Гд1дUТ[ОН DF tiSCAW.1511м8У1•S ANU E2IPWYT1ifR• (.1UNE 1ЧТ5) - ^ Nuwбer оЕ Ешр1о �ы Earrбllчшnunr � _ „�,,,, ._._"'---_..'"-_��.._.�._�- 1�t'R`.'.".��'__--.,_._.._"_.-_.�_---�_'--� ._.'__-- _ __._�-.�_ 1 4 _ 5- Ч 10-14 _ 15-19 20-49 _ 50•9г! l6U-194 7UOr _ __. •_-'Си_r.�t _ _ г.ц[дU'�-%kmploy- 8S[aR- /l'�пр7оу- Е9[�Ь-�Т�дрlау- Еч[5h- IEi.ploy- kчСдп- Е,яу1)у- ёыШб- 7Ешрlоу- ЕыСи6- ГF:+прlоу- lьспб- /Ёшр3оУ' E,r.ab- /:yq.laY- tLuгцh,ctuxtьg 5vc Wr. 3lahncncs/� went 11ь1иыап[s( шеп[ 1lstгwencч/ wun[ 1lагипепсs/ шлл[ ltьhиеп[ч/ menr 31sЬшап[ч/ шаnс 1lrluuanCчl ment 11sЬш,ч,гц1 тдпt llыьигпглl шь.и . •'_'_ "' "' _""-"-"",..._.._._-..�..,,-..._ lочд 49.Э 5.5 1у J v,8 д.Э 5.1 4.8 4.1 9.8 15_5 4.2 1S. 2.2 15.п L Я �z.ч и)1.!1 tг1n.0 yuvctлgcs 1й.5 ,з у,4 ,5 5.0 .5 3г1 .4 36,4 4.1 12.о G,8 20,1 21.Э 18.9 бб,1 10U,0 10U,0 'гилдиг.а Э4.7 1.1 e.z .9 д_2 1.5 4.1 1.1 1е,4 д,д ь,1 s,7 10,2 20.з 1о,2 ь0,5 1о0,0 1cw,0 - Сск•lle� J6-4 ,8 18,U 1.4 Ч.7 l.4 5.7 1.1 15.3 5.8 7.8 Ь,3 i.4 8.6 7.9 74,6 100,0 160,0 t С1осмlпg/F'ии[ысаr 5э,9 6.0 1Ч_5 %.2 д,3 5.Э 4.5 4,7 10.5 18.6 Э.5 lЭ.Ч iJ 13.4 l.4 Э1.8 10U,0 l00.6 ялаа 52_9 11.1 2Ь.Э ц.б 8.2 8,5 4.9 7.'t 5.4 14,8 .8 4.8 .В 8.4 .8 2Ч,а 106,0 1UU,0 Funьtrur.: 57.� 13.0 26_9 14.1 ).б 9.4 Э,б • о,2 7.4 22.0 2.1 15.2 .9 12.д .2 ).7 100.0 100,0 Рдреr 25,0 1.1 20.8 2_7 11,1 2.Т 2.1 .7 ?0,1 1Э.7 9.6 12.4 6.Э 15.9 5.б SOJ lUU.O 10U,6 егlпсlиь 4з,4 5.9 23.1 ы.9 9.9 7.6 5,0 5.0 1Э,1 16.ь 2,7 и.1 1.9 1а.5 1.0 Э1,1 106,6 100.6 1,nл-.1,аг 45.4 4.0 18.7 5.4 9.1 4.6 6,1 й.4 10,2 14,1 5.0 15.4 Э,] 2L 4 1.7 Э0,8 1UO,U 100,0 лиБLес 7д,5 1,7 1й,9 2.0 Ь,Э 1.6 6,3 2.3 21.Э 12.4 6,3 Ч,'[ 1.1 Э.д �.'2 67,0 100,U 100.0 ь Гlшшtсдlч 31,7 1.7 1d.q 2,9 fl,6 2.4 6.4 2.5 15,9 11.7 7,0 12.1 ).1 24.6 5.0 42.1 106,0 100.6 °' . г 011 илд с.,и1 55.G 1.4 11,1 .7 б.7 ,tl 2.2 ,4 11,1 2.7 6.7 й.3 2.2 2,4 4.4 tl7.3 100,0 �100,0 Т+ип-м t.ьlltc мlпагиlь 4д.2 4.3 19.1 5.6 6,4 3.0 4.5 3.l 12.6 15.6 4.7 12.9 1.9 1U.4 2.5 45.% 100,0 100,0 Чs>1с Ме[л1s 77.5 2,4 18,Ч %•S 10.Э 3.3 Ь,д Э,2 11,6 11.В 7,2 1Э.8 2•Ч 12.1 2•9 56.0 100.U 10U.0 меtаl Fcnduг,[ч 41,Ь 4.7 2U,6 6,2 10,0 5,4 5,8 4.5 1Э_1 18.6 � 4.9 11,6 2,6 16,2 1.6 - 29.4 10U,6 100.0 Uwrk[а..гrlслl Млс1ьlпсеу 15.9 4.4 22,3 8,4 13.'[ 7.S 8,5 8.1 15.7 2G,0 Э J 1Э.9 1.9 1Э,5 .В 18,t 7UU.U LOO,U Еlгссгt�нl hWг.Llпcry 52_G 6.5 19,3 0.0 7 J 5.6 4.6 4.9 9.Э I].2 Э.Ь 14.') 1.tl 15.1 1.2 2].4 100.0 tOU.U 1x,msp�r[ kyulpuьiut 55.1 9.0 21_2 10.7 8.6 7_6 7.8 4.д 7,6 16.Ч 1.В у.'Э .3 7,9 ,е ЭЭ,Ч 100.и 100.U и[lь r 5U,0 5,Ь 17,5 5.д 7,6 й.5 4.о Э,9 1.1.5 18.U 4.5 15.6 2.7 19,3 1•b 27,4 IUU.U 1UU.U Ги[.ь1 47.Э 4.5 20.6 5,Т д,5 4.Э 4.9 ).б 11,0 14.4 4.0 12,1 2.4 1й,3 'LA 41,2 10U,U 100.0 ;;ииг.а; t�e Fuucnu[еу �ьn� sо�т�и se� Tдbie 1. TabLe A-19ý INGREASE IN ESýrABLISMLENTS ANU DErijLEN 1956 AND 1975 BY SEM'ORS AND SIZE CKVECORIES (FIMS ot 5 ur morý "ýPIOYL(-S) 5 - 9 10-14 15-19 20-49 50-99 100-199 200k Total Estab- 1EU,ploy- Estab- luploy- Estab- JEwplay- Estab- /Employ- E5c2b- IFJ-PIOY- Escab- JEmploy- Estab- /Effiploy- Estab- IE",plOY- plýnt lifflun.ntsl ýøt 1"lýa,L.iltsl WLIlt Wýi"enes/ ment llsli,.Cncsl -ent liý1w1c11c81 wone 115JU.C.IL91 DiCftt Uýlimýnts/ ~nt -2øj -1281 23 466 50 976 132 3968 69 1990 53 6940 49 16534 W 32593 44 - 274 -21 -249 - 8 -122 4 10 1423 22 8760 - 54 8783 - 82 - 609 -48 -570 -27 -424 -25 693 8 - 562 - Is 1. 695 -189 - 2178 45 314 sti 613 16 327 62 1972 38 2615 25 3034 33 25464 269 34339 20 246 82 1083 53 9ý3 180 5858 59 ø07 38 5270 42 175h3 474 34970 25 228 19 222 23 375 22 605 2 176 1 27 913 89 2L')4 104 630 73 872 36 5B6 93 2590 23 1526 13 1819 2 639 344 8662 17 108 10 120 1 23 9 367 5 4013 4 368 7 2BIS 51 4163 161 1027 71 849 28 484 105 1657 23 1654 14 1234 7 4"7 409 12052 25 185 17 189 19 315 24 796 13 955 12 1616 4 220B 114 6264 Kabb.r is 97 - 2 - 28 3 s? 31 828 10 717 l 145 4 1715 62 3531 55 398 21 21ý4 26 443 59 1912 2217 47 6962 35 12579 271 24755 011 C-I prýdUCL, 20 Is - 1 31 3 169 1 105 - 2013 6 2292 r MinL-t ptdueLs - SI - 345 - 9 -138 7 94 72 2367 29 212ý 109 18 8406 52 12216 blýi. H.L.IS 65 446 40 461 23 386 52 1654 31 2129 12 1836 9 4374 232 11266 ýl,i.i I-rudoeLs 324 2182 172 2038 100 1676 223 6847 92 6119 49 6656 31 132ll 991 38155 Mm-EJ-Leical flaýhl,wý:Y 61 399 20 243 34 565 65 1938 14 1012 9 1086 3 1355 206 6596 231 151,5 99 1144 47 819 ga 2990 43 2978 24 3188 15 5700 557 18364 T-,,,p.CL 211 1282 98 1147 47 790 94 2761 20 1373 8 937 7 4174 485 12464 7 151 985 69 8d3 46 778 109 3361 44 2990 28 4029 18 6431 465 1937.1 1135 7583 785 9520 516 a842 L195 4t478 530 36784 349 47051 308 140222 50111 2914B(l 195L, - d3tý fc"ý Annu.l S.",y øf and DÅNE, AA,,arl. V..Lad15LiC., 195ý, Tdble 376, 541. lq;ý Adjuýled and unpublished ICSS data. S- TilbIL . _ � - I � > � + г 3 •1.�1�1е А-?0: 1�Е1iСЕиглL:Е D[sCluuO�i'10N OF TNCцEnSE 1N £5'1'ЛO1,1511MEttI'5 Ali� EHPLOYNEN'[ Bk'гЧЕЕN 1956 aNU 1975 иУ sLC1'UкS ANO �Т2Е 1:А•гGе0ы1ЕS (L`•iiws of 5 ог more аирlоуес�) tluuober of kmployces рег ЕеtаЫlslгигеnг 5 Ч 10-1rг 15-I9 20-49 50-99 100-199 _ 200и 'Со[д1 �i:.[дЬ- Ewploy- р;[аЪ- Br,ploy- EseaU- Eniploy- Escaб- Етрlоу- EstaЬ- /Ешрlоу- иЕsсггL- Gmploy- EstaU- Fшрlоу- Ез1а6- Fauploy- МяииГасштiиg 5агшг 17ь7w,ruCS/ гuent 1хsЬтеп�ч/ m�nC 11ч1шгецч/ теnг цзtгтерсs/ огеие 1lsbmencs/ тепС 1tslua.:ncs/ ment 11s1mгcntв/ теп[ 11.ыЬгиеnсs/ weuC Е'и.,д -4.0 -.4 .5 .2 1,2 _3 2.б 1.4 1_4 1_7 1,1 2,4 1.0 S.7 't.7 11,2 и�.vсхаьич - ,9 -.1 -.4 -.1 - ,2 - - .2 - ,1 - .1 - .2 ,2 ,5 .4 3,0 -l_1 3.0 '1'�i,..,;�o -1,6 -.2 -1.,0 -.2� - .5 -.1 - _5 - .2 - .2 - .? - - - .2 -9_В - .7 'с..:ii.ь _9 .1 ].0 .2 .Э .1 1.2 .7 .Е< .9 .5 1.0 J 8_7 5.4 ц.0 ��1.�Чtпt:;Fог,си���� ,4 _1 1.6 .4 1.1 .Э 3.6 2.0 1.2 1.4 .В 1.6 .!i Ь.О 9.4 1'L.0 � Чиид .5 .1 .4 .1 ,5 ,1 .4 .2 - - ,1 - - - .Э t,д ,8 ! мггсггll�rо r 2.1 .2 1.5 .Э J .2 1.9 ,9 .5 .5 .Э .6 - .2 6.9 3.0 рар��с .Э - .2 - - - .2 .1 .1 .1 .l .1 .1 1_0 1.0 l.4 lriuг,IUg �.2 .4 l.4 .3 .б ,2 2 д ,б ,5 .6 ,Э ,6 ,1 1,b 8.2 4.1 I� г �.,гц�с .5 .l .3 .1 .4 .1 .5 .3 .9 _Э .'L .Ь .1 .8 2.Э 'г.1 h. о П�ЬL�г .Э - - - .1 - .b, .3 .2 _2 _ - ,1 .6 1.2 1.2 г СlгстСс.гlч 1.1 .1 _4 ,1 ,5 ,2 1,2 ,7 ,б _В ,9 2_4 .7 4.Э 5.4 8.5 ОхI & С�.г7 Ртдгггсs ,1 - - - - - - - ,1 ,1 - - - ,7 ,1 .П t3�n-MucvllIC М(иесаl р[oгlucCS -1.0 --1 - ,_ - - ,1 - 1_4 _8 ,G ,7 - - ,4 2.9 l А 4.2 O.,slc г1и�дls 1,Э .1 ,8 _2 _5 ,l 1.0 ,Ь ,б ,7 .'1 .6 .2 1,5 4.6 3,9 t(г1�1 рrидыисц 6,5 .7 Э.4 .7 2_U ,б 4.4 2.Э l.U 2.1 I_U 2,3 .б 4,S 19.7 13.Э Noгi-Eleпriwl HecLln.:ry 1.2 ,1 .4 .1 ,7 _2 1.Э .7 _Э ,Э .2 .4 .1 .S 4_1 2,J t:3вutricdl hlucSiii7eкy 4.b _5 2,0 � .4 .9 .Э 2.0 l_0 ,'l 1,0 ,5 1,l ,3 2,U 1!_1 б.Э -tr.гnsparc Еуи�риюггс 4.2 .4 2.U .4 .') .Э !_9 .9 .4 .5 .2 .Э _l !.4 9.7 4.Э ucLer �.U .Э 1,4 .Э .9 .Э 2.2 1.2 .9 1.U .G 1,4 .4 L.2 9.7 6.G 'l'.а aI -12.b 2.6 15.6 Э,7 lU.3 '].0 27,гi 14.2 10,6 l2_Ь 7.0 16.1 b.1 48.1 100.0 IUU,U 5и�гиа.,: 1J5❑ 1lnpubllshud tk1N1; даси frouг Апишгl 5игv�у ut М•+nufaa[urur; апд l1ANL Auuдriu Eьrпdis[7co, 1'157, '1'аЬ1с Э76, р. 5й1_ 1975; лдjацс.:д ьпд unpWзllsb�d {С85 д.гL,г, ^ lab _lek- 21.ý SIZE. DISTRI3UTIMI OP A SAMPLE OP FACTORY ESTABLISý=ITS ACCORDI-NG TO TREIR FQSZTION LI 1970 AND 1973 311-312: FOOD-PRODUCTS 7inal Year 1975 5 TITAL 3 T.5 1 ;7 3,17 7 3 9.35 n ).,,T til 5121 ,,1, 5 ýQS7 t 7ýA 137 1 1?,; -ý 7 jý, 73 7;ý 7 5 10.15 zi i P ýji 7 5 1 7112 2b3c >ýkQ9 <9 I ýNA 13 33,3 5 5 i 11Jýe T ý,I * T ýs 3 C: ;2 13 c 10.64 7 RS -lu ký T 7 ý2 >2q50 2ý;k34 .5ý1,1 ' T3 527 2qr'? i s-jý . ;ý 5 8 7 5 -14, C $ 34.33 5 * i u 7 1 ti ý4 )h 7g2 /1 Es tab Lishmenrs 7 k U i n Iý7n 5 A 2;ý. il i I Iý:i 2 inirial Employment (1970) 7 7 LL tj I il uq710 /3 Final Emplovment-(1975) 791 I7AU ýéý(si qý28 _ý -4 * Tncrease in EmpLoyment k 100.007 7-s 't contribution to Nec Employmenc Increase -q -1cE CF!-LI, ;,,Tý<Y 1,ý y Sources-. See Table IV-6a. 川 153 ,LA1M7AC'11MIING. SIZE DISTU.BUTZON OF A SAM-PLE OF GTORY ESTABLISI=N'TS ACCORDING TO THEI'ý 20SITION LN 197Q AND 1975 321: TEMLZS Final Year 1975 (t II n. T-l-r AL q3 2 U 5 127 74 2 7 k q A, 1.82 11 c -1u11 T 15 iý, 3A t4 1 5 S 5 ..D 8Q k. 37 2.37 1 (2 ý2 CI;Jm T ;ý3 7v. ý lim 15 2 A5n li 43 S"Ih, 1: s 3A7 sol 773 S t l C 3 7.85 C -11i'. T 4 177a U m IS i n 252q us69 s il 1 3 1 35 t A i C 5 2 81 A 3 c I. UM T 3 119407 4 <J s 3 2 S:Im rs 1; ý 57n2P 571 7ý 3-jn ;ý 3 -75 7 1 7 ý.; 41 n slim, c 3 76.64 p7p /l Es tab Lishments 1 3,2 i giý 1; ý1 54q87 F2- Initial Employment (1970) 1, 1 li h ti ;' u -ý-3 Final Emplovment (1975) u c :;r k7 !4 Increase in EmpLoynenc #1 100.007 ý5 % Concribucian ýo Nec Finplayment ýncriýase jq;r -IF C 4 1 rvr F>JT;;y t$ ýQF-PLACI"' ý4Y Source: See Tablp- ZV.16. -154- ?Stц1UFACTURI:FG: SIZE JISTRI9UTION OF е: SAi`1PI:E oF r'АСТОАУ EST�BLISIfii1E.цTS �CCORDING SO THEZR Р052?ION IN 197� а,ЧD 1975 322: сг,откlггG F1na1 Year 1975 Х ♦ м t ,Н м л ь * (l�П, +г • * г5;ац) мta5,uct ь��n,э9i . oa4a9i * t'1Тд4 i*#trмiььskr�y#�г*1г*вммtyrвstrktr�вtr#у��ьв,нtуаttв�мt* * 1 а � п а � ь 1: ?р С�,�Чt , ��А r . r . ь Q ь 1и7 �г1н, IS {�,�) ь �д5 м . * . , �So у 3Qa S<.1н, Fg * oS . .. �ь . ь 15п # г°5 g�1ю, С5 4.3о * _ м о у. . f i * л п t* Х!* у k* м у+ k! у# т Х*+ м м t* у к i#�м у i м f у+ t k х 4 Х�м ## f 1 i# i�'+Г Х * 7:: . �� . З« о� tu3 С7�Rгт * i 1"" г �51 + ц5 Х ... ь ! Ь�Ь ai!`+, j5 г1�,гиi • 1л5; м ои� у ?n5 * - . 7гп1 SJ`+, F5 , _19 м иЯп + 1Ьы м . .ь 5°5 9Jм, CS м м - ь . * . �. 10.45 +пХ*Хrу..вгьуа*arkьt*hlytA}lмyifl.тiff4ta'+eвiiyt#М**м4! у 1 � ! �i t 19 м з# 92 �`�о~р " . 4,1 ► �?р4 м уяц . 11а # зг�я gг��+, is °(гiга�i ь �1� Х Ja�a + 1?.3t м 7Ql; + ц5�1 Stfч, FS �°,~' Х .73� . 15Q tг 5U7 г А27 ь 1г1Q3 $Un, С5 * м - + • * ® в 19.20 j t х м f i в t г#* f Ь* у ь а* ь у ь*# м м у в# Х i* м f л# 1г t# М• # t} f�*# Х**#'м R 2 4 � , з м г� � г5 + ir, м 5е С�иИг * 149 м 74? г 17г� * 1го4 � 391гi ,у�1�, тз '� (j;з,��i д 5? ■ 9ац а 1��Z в г11q ь цуb4 SU`+, �3 . .1 й7 : .�г1° ь 9г r� о07' у 55ц gLI1r CS * • + ь . ь . у 9.73 *Хм*Х�tв*tkyXi4мitftY'*МК'k'FRfttiy*1�1t*Rtпt•1Г*мiдум}*в�ХN*�, г t м 1 ь а t ab i 5l С?U�T * . у 1,� . 53о м 11�зг t iг�ьи su��� rs (ltг�, * _ у ��с м 3ц7 * 1ио7я ь• 153F� 911�, F5 . . ��qoc, . _ * .5о е -1А3 Х зии~ . j2(1и SUx, CS . < r • iг • м . + 5о.29 ' . ) 1*� У i 6 й в 1 м i* у в R у k! в## h у Х i t i* i i+ t 1 М#* i^ } А* i#} R# 1�' у'� # f Х'м } . . 1 t; * -�n м �р * ,у, i 3?_2 /1 EscablisEunents �о74 * ;1Ьц . 7v79 t 12а�.� , �tгуи5 �? Initi:aL Етрlоутепt (1970) .`з{,,� . i sдм с дг13� * «о5 + 1 илио * гг,13?, /3 Fiaal Employment (1975) . .;�� . а1% + �1b + St25 а 5aQ1 /4 Increase in Етрlоутепс . ,, у • * _ * - у 100.оо% /5 % CoпtrlbuCion Со УеС °трlоутепС Increase �+лч. а:^� �S� ". vFCвitvc .5.1�г ТЧ _iVF п� �7дЕ` CFLL.i, 7л1Е ГЧТ"v iS '�EPL+�Г.F� дУ ,�.�SЧЕ9 iN dn��vF с�_А,'•F. Source: See ТаЫе LV.16. mAIWACTURING. SIZE DISTRIBUTION OF A SAMPLE OF 7ACTORY ESTABLISHmENTS ACCORDING TO THELR POSITION IN 197Q AND 1975 323.- 12-ATEER PRODUCTS FinaL Year 1975 Ik,(25,44) (su 'C?Q) QQqQ9ý T'1TAL 2 51 1ý- x 5 15 SI", C 3 1U N- T 21 1 27o 5,.111, 15 3.13 173 5W1, 15 (25, 'lý; 7 jm, ø ý4 -8.26 ?j 1 29 79 3,j-, Ts t 5 3 8 ý4 785 Sii F3 34.75 C 273A 2R39 T 5 39n3 511,1 68.52 f, Ssrablishmencs 9 > 7 7;; /7 Inirial Em-ploymenc (1970) 9 7 ?t7 qjý 31ý9i tl Ni' I ý -3 FLnal Emrýdoymenc (1975) W cl !k?s /4 Increase in Employ-ment 100.00% T5- " Concribution Co Nec Emplovment Increase k- - ýFr. il;;- -,Fr-,Årf-ir qij- rt, 3NiE iQ 10F rri -Y 1) 3Y ')AS-FQ A.;ýjv- Source: See Table IV,16. . ...... .... . - 156 - M4ANUFACTURING: SIZE DISTRIBUTION OF A SAMPLE OF FACTORY ESTABLISHMENTS ACCORDING TO THEIR POSITION IN 1970 AND 1975 324: FOOTWEAR Final Year 1975 * * * 4*p.I .r25,49)l *(5C,9) 4 oa9q99 TSTAL * * ~4444* 44 *4 ** 4* ** * *** * *** * ** * ** ***** * 7 * In * 1 C10MT 5 l a * 2 aUM tS (5 84 * S3 4 • * 28 SUM, 'S * ' * b? * aU' * - 1'06 SUM, CS 16.62 * * 1" * * C j! 3 CUT S 373 * 135 * - * * q63 3'N TS - ,, * 1U1 ' 305 * b * - ; 5 S 'S .67 * tto * - 52 sUm, CS -* * - 5.92 * I * * 4 ¶ aib CIUNT n 301 * 120 * 2L S8 Slim, ts- t2,:4 * 13 35n * 3c;7 * 103 823 Su, 75 -25 4 177 * 7 275 StUM, C * - * .- * 31.32 3 7* 5 15 CIUNT 17* * 5n2 * 33 trl13 SUM, TS 54,1 - 125 * 58 * Ao? * 1295 SUm, F5 - * .5< * '86 4* ?l70* 282 SIIM, CS - 4 - * - 32.11 4 ~ * 0 * 2 * 8 C')UNT - 2* 21 * 523 1569 l m S (Ion, - • 132 * 16 1h92 SUl, S o. * - * t11u 237 313 S C3 S- - * -14.00 *1* ~ 444 0 *«* *-****4++************** ****444 ** * 32 * 24 - 15 * i 85 /1 Establishments a71 b * 937 - 1 A8a 3775 12 Initial Employment (1970) T iT L 413 q64 * 1 110 226c; 4653 /3 Final Fmployment (1975) ?54 ? * 173 * A8i a78 /4 Increase in Employment . 4 . 4 - 4 - ~ 100.00%5 % Contriburion to Net Employmenc Increase * F** FlA rF VF Å T V 9F NU4 fx] -3MlE 'I: MJRE CFLLS# ONE ENTRY IS REDLCEn Ry DÅSHES IN Ad()VF Source: See Table IV.16. - 157 - wANUFACTURING: SIZE DISTRIBUTION OF A SAMLE OF FACTORY ESTABLISHMENTS ACCORDING TO THEIR POSITION'IN 1970 A.ND 1975 331-332: WOOD AND FURNITURE Final Year 1975 *rf * *r * (ton, 5.2L) *(25, (50,99) * qqqq T'"7It. fr r rf ~.f f r~~ rf f.r*r~ff * * * * * * * * ** **** * * rffr* * frrf rf * ** ** a*f 1* 13 f .3 0 16 CMM~T fr f fr fr - ~ 1t än I3 53,9) * I? fr * 7 * - fr223 3U'M, PS 2 15 * -• - 107 S.1,j CS - -14.92 fr* R fr 2 4d CCIUNT 553 * 1ih * 34 .74 SuM, 1S 1,?4 46,1 * 273 * 125 * - ASA Sim, FS * 4-3r * 117 * - 110 SU1M, CS -- .r15.34 fff*frr*frrf*frfffrrff*fr**fr**+***********frf** fr*******f 7 i< * 7 2 C1>T * ?48 * 320 * 251 * 56 835 311,4 1 2(,49S 1 11 * 351 * 495 * 277 1242 Suim, F3 kr -47 * 3c * 2a4 * ?21 u07 311M, C3 * - * • 56.76 0 * 3 t a f 3 12 CnuNT ( - * 197 * 373 * 2c46 816 StM, o S frt-Q * 39t * 35d A5u SU"O F3 - -7R * 10 38 SU), CS -f - 5.29 * 0 0 f * 0* 4 4 COUNT r - fr • *f - * 1303 1303 SUM, TS , - * - * 1359 1358 S, S 9 f - f -r 55 55 3U, C3 - 7.67 1A P ?] 5 * ~* 1O6 /1 Establishmencs A - 55 * t 8 * hOS 3818 7- Initial Employment (1970) T L7* 4 * 101 h R * o 535 /3 Final Employment (1975) f .j * 153 * 303 * 134 717 74 Increase in Employment -r - * - f 100.00%75 7 Contribution to Nec Employment Increase *xE 9 i1P IP MFG4TTvF 5Up P, 3NE tR9 4OPE 1ELLS, ONE FNTRY 1S REPLICE) SY DASHES (1 AÅHOVE 01.4A1E, Source: See Table IV.16. - 138 - MANUFACIURING: SIZE DISTRIBUTION OF A SAMPLE OF FACTORY ESTABLISHDENTS ACCORDING TO THEIR POSITION IN 1970 AND 1975 341-342: PAPER AND PRINTING Final Year 1975 * * * b(100,,. t524) *(25, U94 (50,Q-) * :9999 7TALOÅ 5 * I b 0 b 0 17 CIUNT 125 * 1 - * . 133 3ti , TS (3,q) 1Qs * 31 ? 19 3i-, AiS 63 * 2 * - - 6 SUi, CS - 1 2.72 110 * 2 * 2 e * CIUNT * 1633 b £65 * k5 k - 21a3 SUm, 13 ( 3,i73 * i7h * i1 * * s678 SUM, ;S Q5 4 L * 76 * 5 * 35 SUM, CS * - •. * - * - ' 16.96 I Qi 1 '9 *3 78 clUkT a17 * 13<1 4 727 * 15' 2bt SuM T$ o (2 5,49 4* 1b * 1?02 * 2n * 3110 sum, Fs S 33 * 75 * £75 * 35p 7 1.1,4, Cs e * b -24,38 *fb*DW* **bl******b******* ************b*b****f*b* 5 * i1 * In 3u CJUNT *1 * 2n A 10A3 * A 2294 <'Im, IS Q (51,99 b 3 * 152 * t146b 1263 23 SU&l F -5 b * 63 * t!5 z50 SU' CS - - * - • 11.10 b * * 2 * 30 * 33 CUN - 117 * 215 * 10590 9 1016 slim. iS ( , - on * 156 12133 1 ?232q S) , F3 99999 * * -77 * -i9 * 1939 * 1413 S11t4 CS * - b - • • b * 44.81 **k *.,* b******* ***b******b***** ******b***b***** 2 7* 70 b a 'd 300 /1 Establishments 276 é 2213 b 2060 * I15bA 1117 ~2j Initial Employment (1970) T T i193 25S * 26?5 b 13904 21270 72 Final Emp1oyment (1975) .A3 * 335 b 5h5 * 2336 * 3153 /4 Increase in Employment - 100.007 15 % Contribution to Nec Employzent Increase bg*Aw ic;l: «iF tE$IATTVF RUM T 2ME ri9 JPE CELL, øCNE ENTRY S REPt ACFn 8Y )ÅASES IN AROVE pi, ÅNIE Source: See Table IV.16. l Э -1�9- 1 !�L�YUfACTURING: SIZE DISTRIBUTION OF � 5д1`1PLE OF ..iCT6RY cST�,дL25гйiEiд'L5 �CCORDIYG ТО TЪIEIгZ POSITZON IY 1970 :цУD 1975 352: OTHER C9E7dICAL PRODцCTS . iinal Уеаг 1975 - • в к i # * ь ь в # (tOAr ь к ГSоац� к(�5,иа) yc5n,a41 . ао449) � т�тi� кrxиrrrrsrrrf*#виьвьвя#ьхгвьь#ьь*rrri��r�ввiy#rs#fв*s . ь ? и А+ (1 к• j ь З СПUА'�Т ь 1 i . . . . : о t 2и ���Mr ts (�,а) . t°• + • • • * 1ц7 « tb5 ЗUн, Fg ь 3< .. r , ь 1 3А # 1 � 1 9U'+, CS ь ь + к . х . t 3.69 м#пкгьвьььiьsь,гtи#wк#rrvitвw+r*вьхп+вах*trвв,rr,rtr,r,г#*,ь ь 5г ь 11 # 1 к 1 t r�5 сзичт + 7h3 к 20п ь ?-3 * 15 ',t 1 р21 gUм, IS lt"1,7ц? • 71ё м 37b + ЬЗ х !21 # 1Р72 3UH, F9 ь .71 к I7h # ип м 1аl, � �51 8ц�, Cs г а • t . ь • е 6..]7 ьп.#кьt*rьквьtь+км:#f-t##ьт,гв+е+�в#+*ew�xt��*вfв:lotiw в 7 а Z7 # 7: п �. и1 �pur�т в 1а9 к д57 г ?79 ь . * 13�4 911н, I3 °(2i,19i ь 137 х а5о # а43 ь • r t5uq Н F � 9U , S °; к ..b� ь D2 к 1•Т7 ь . ь 2�: 9U�, CS и +r : . . . к . _ ,ь 5.37 `� лrьвхв ,rsьr ¢ьь+в#t,rtts,ь+#,гвtr.ьrrst# ьу # а х � + х#,г ,г ,t vte в#rь мпi �" . [ х з* 1 т х 13 �r зи �;7Uит � ь 5! ь t�гi ы ] I32 # 1n2t в Z3°!� 3Uн, I3 L[5�,��i ь 22 к f15 # t251• ь d08i ,ь 5aбq gU�r F8 � ь .?� к +7о s 119 # 1nb0 + ib71 SUyь С5 ь к . ♦ . ► . ь ?8.06 гхггь•э,ььььхr*ь,rг#f#згw,ьiг+iлв�ньiгх#�гаrгв#iг!!г*r+ftв#вsмчгi х П к ^, х 5 s 5п � 55 C�UNT - ь - ь • ь b1Q t 1ПАtЗ r 11и3г gцн, �9 (1пл, ь . и - + 353 * 13�25 * 13580 gU", �S qaqvyi ь . + • : •.2%,ц к 2ц1? . 21�8 gUM, С5 х и ► к + # . s 5о.28 ь,гьькл*ьпыммаtьье#+гьs#+ьаь#+гмь#rхгяьs##а#в,r.гг*ьiгх*ав ь hг_ ► а•1 # 3п + Ь5 : ( �9 /1 Escablishmeacs • ь tп��д к (�51 ь ?п52 * 11Q5A r 1h�19 �� Iпicial Еърlоутепt (1970) • 1?iл� . дЧ4 и 1�5п ь rl�� в 1557u ♦ 2Пnд5 /3 riпal ЕтрlоутепС (1975) ь .1SЭ + 13о + 7п * 371� ь iдlb I4 Increasa ia rmnloyment + - • + * + � - а 100.00% /5 °L Contгibution Со ;Iet Employment Increase •,•.. яЕгл��SF �Р rF;дriv� 5�Ч ?v �мF г�а ��)дЕ CF�L3� пАгЕ Ент!•цУ Ts aEa�.sCE� Hr ЗаSнЕS ru дЯ(1vF о1 л•r6, Sоитсе: See ТаЬ1е IV.16. 160 =ACTURING: SIZE DISTRIBUTION OF A SkWLF OF FACTORY ESTABLISHKENTS ACCORDING TO THEIR POSITION IN 1970 AND 1975 361-362: CLAY AND PORCEUIN PRODUCTS, GT-,kSS PRODUCTS Final Year 1975 tio, 'hý(sn,QQ) qQQqqý TITAL 2 CIUýj T 17 SUMP TS 26 sut. --5 9 sum, C 5 .43 0 1ý COUNIT jt,' * 59 * 185 3um; 15 119 * 135 * 253 sum, Fs 7 88 3 ,J ý, i C S 3.25 is CI-UNT ? 8 * 30? * 171 * ugg 3 U m ý f $ 2 5 a ý 3 * 31 1 * 372 * 706 3UM, F 5 5 * * 2n1 207 sum, c 5 9.90 r) 1) N T 54 * 272 * 277 603 3UM, 1's 22 * 263 * 611 A96 3uý1, Fs .32 * 3 3 il ?q3 311t4 p C 3 14.01 I * Ih 17 CrIUMT 1 15 * S738 5953 sum, T 3 (100, 814 728P 73hh su", ýýs ýj tsud 1513 sumt CS 72Zý9 1 2 1 1 0 * 20 55 /1 Escablishments PI 17m * 55ý * hil i s 7157 12 initial Employment (1970) T -IT ÅL 1R9 ý1,1 6 * 7 1 ý3 * 7119-ý 4 ý 14 7 ý3 Final Emilayment (1975) 18 * JAJ * I 7;k 2 0 <? 0 ý4 Inerease in Emplovment 100.00%ý5 7 Conribur-ian co Net EMDloymene Increase F 5 Å r T v 91 ýJ'4 TI, ',NIE nq w'jPE CELLSP 'INE E>JTPY t5 REPIL.ACErN SY 2A9MER, 1'4 A8rIVF- Source: See Table, IV.16. - 161 - tIANUFACTURING: SIZE DISTRIBUTION OF A SAMPLE OF FACTORY ESTABLIS=MNTS ACCORDING T0 THEIR POSITION IN 1970 AND 1975 369: OTHER NON-METALLIC MINERAL PRODUCTS Final Year 1975 ** * 0 , 14 21 2,9 *(5n,QQ) QQQq T949 n7AL * *** * *tn ***** å 1t 1 0 0*. * n 12 .CrUNT 4 * 9 SUM, iS ( ) 117 * 27 QU * uSUm PS .31 * 14 * • * 50 SU CS S - * • + 6.33 o 12 * t * o * 79 ClUNT * 5n * 237 * 24I * • t211 SU1s, 13 (1),24 * 52 * 19* * * 39 U-", 1S -98~ * tl ø 35 * Sum, CS 12.42 ************************ 1 * 3b * 9 * * 63 C-UNr 529 * 1262 * 3h3 * - 2s5 s, 1S 2,9 353 * 2S93 * - * 241 SUt, PS * .176 * 31 ø 130 * -.13 3U49 CS * - -1.64 ' * 20 * a 28 COUNT 229 * I 4n4 * 315 * 19L8 SUm, IS 50 , - 1 7* * I2n * 652 * 2250 SU', FS -51 1& * 337 * 302 SU, CS 38.27 I * t * 3 * 26 29 C1UNT .. 4 * 351 * 842 780 sum, IS 0 n 221 * 890m * 9¶32 sU,i FS * - * -12-7 * 7 352 'S U CS - ** •• 44.61 S 5 * 53 * 33 * 3n 21 1 /1 Establishments csh * 1 73-h6 * ?1L 2 7 Q 1Q4187 , /2 Initial Employment (1970) T 1r4. 32 * 1994 * 216 * 956n 14 q716 73 Fiñal Employment (1975) .3 * 152 * 4* il/ * 7AQ /4 Increase in Employment 100.00%/ % Contribution to Nec Employment Increase **#*** 1g- ty 5 - TVE ' U" T IJNE nQ 'OPE CELLSø ýNE EmTqY IS REPla4CE1l R, 36SES T', Anvý pLANE. Source: See Table IV.16. - 162 - kANUFACTURING: SIZE DISTRIBUTION OF A SAMPLE OF FACTORY ESTABLIS1ýNTS ACCORDING TO THEIR POSITION IN 1970 AJD 1975 381: METAL PRODUCTS Final Year 1975 i* i * (100, *5ral *e25,'49 +(50,94) * 999991 i*A * * * * * * ii n* i4* * * * i* * *** ** ******* * * * ** ** * i* * * **** 27 3 * 0 * 0 30 COUNT I k 22 * * • ?36 %UM, tg (SIo1 * 324 4 9 * * -93 30% FS * 110 k 67 - ¶77 3, C3 * * * - * 5.03 110 * l? * 7 * i * 160 COlUNT i732 764 * 135 * 23 2671 4 S, 1S (1 * ¶69f * 1t07 * a31 * 10a 3s32 Ss 3 S 2 623 * 296 q 158 SUtim CS ** . 27.23 * *i******** * **** * * i* * *i* * i* i* **** * ** **** * ** * i* * ** * * 16 * 51 * 17 * i 85 C1UNT 435 * 1637 * 439 i 43 . 9 30M, IS 2 2 * 130 1 075 15 * 3326 som, F3 14 * 14 i 4136 t 107 517 SU>l, CS * • • •14.69 I ? o 0à i 64 CIIUN T 65 * 7 a4 * P767 * 78Q a 1365 T (50,99 i * 427 * >952 * 2 * a913 St FS -Su * -317 * IRS * 734 58 SIU`l, C3 S* * * - * • * 15.37 4 3 * 2 i I * 4 b CrUNT 320 i 129 i 95CIA 916 3149, 13 (tnn, 341 S i 75 w 11075 11234 sul, FS 9 - i -156 * -53 - 1527 1318 511 CS * . * • 37.46 t54 a 110 * 65 54* 395 /1 Establishmencs b p£496 3413 i 3669 * 10403 2000 Initial Employmenc (1970) T 296 k 3837 * 4533 * l2852 *2351 73 Final Employment (1975) .213 * '05 * 61 * 2 411 3518 14 Increase in Employment 4 • * - i • • - * 100.00% - Concriburion to Net Emplovment Increase é**** 2FCå 1S9 4,-eGåTTVF 9Um yý 3NE n9 M0E CELLs, rnNE EjT4Y 13 RFPLACE0 RY DA94ES IN 4AOVF ALANE, Source: See Table IV.16. |_+Ç - y - 163 - MANUFACTURING: SIZE DISTRIBUTION OF A SAM2LE OF FACTORY ESTABLISNTS ACCORDING TO THEIR FOSITION IN 1970 AND 1975 382-335: NoN-ELECTRICAL MACHINERY, ELECTRICAL MACHINERY, TRANSPORT EOUTPMENT, PROFESSIONAL AND SCIENTIFIC E2UTPENT Final Year 1975 b b ne * * * * * ( .<b **-* é eS.2<fl k *(5fl, coQqQ9ý rITåL 3 * 0 * * 4l C1UNT S t 3- * SHM, 1S (5,a b W2 * 91 - b - 213 SUM, FS * * 6 * - * •121 S , CS - * - - b - * é 1.43 ** * 9 k b * * ~* * * **.* * * * * * * * *** * * * * 4** b* * 9* * 93 * 29 * 12 * 1 ¶314 C11UN,T * i3R7 * ')1 é ?14 * 17 ' 1"7- SU . TS il, 2, 1 2 3 -3 7A8 * 0A 32b2 S'', F5 ø '1 57u * 9, 5 i155 U , CS é - * * 13.84 ] 9 ý,7 * 31 1 112 C1UNT k 563 * )4 5 1 %56 * 17b 38 0 S111 TS o (3 * 2n;n¶ 198 * 717 5 16 SU, S G-N03 * 1?5 1 2 * 93A 1360 su?, C3 -** - * - 16.30 **** g***àk********* **********b*** j15 s7 37 * 7'r C1UNT :57 * '9 t'2d /aa S11 IS (5q5, - * 357 * P774 * 2ø1 * 3QuS S3 FS - 0 . n0 * 265 * 118 1 1 SI , CS - b 12.59 9kbA**+r**bbe+*4**q9**********b*********tb**** 1 * 3 Sc &L CIUNT * 112 b 125 l 4it 1SLIi, 160s3 $0, IS ( m 37 2s5 037 n709 su , Fs 4 .2 b -175 * s?2 Q1656 SUm, Cs - -* * b* 55.81 k******* k***+*****kbb**b*********,*****+*b*b****. b P1 5 U104 * 3 * S 314 /1 Establishments 1 5 3 9 "280 * 17034 21967 2 Initial Employment (1970) ø l u i 44' b 37'¶9 b 3795 > 3971 * 3510Q 3 Final Employment (1975) -2 i 211 9 1505 * 134 43<2 2- Increase in Employment * - + - b - * 100.00%15 % Contribution to Net Employment Increase Sok-.k* u cr cr \ ;';TTVu Se" 1N JNE nQ DEe. ONE ENTRY S3 rFPLACE SY ^gsES ab l vI PLA4t., Source: See Table IV.16, - 164 - FOOTNOTES TO tABLE A-21 /1 Establishments.according to initial and final year, classified by size (measured by number of employees). /2 Employment in these establishments in 1970. /3 Corresponding employment in these same firms in 1975. /4 Absolute increase in employment in these same firms between 1970 and 1975. /5 Percentage contribution to employment of firms of a given size in 1970 and 1975 to the total employment increase in operating firms between 1970 and 1975. Source: The following tables are derived from the matching of 3721 establishments which operated both in 1970 and 1975. The original sources were the 1970 and 1975 industrial directories. It was possible to trace 3721 establishments from 1970 to 1975 and classify them in terms of their employment in both years. Thus the cell in row (10,24) and column (25,49) refers to plants that in 1970 had between 10 and 24 employees but which by 1975 had between 25 and 49 workers, and similarly for other cells. Each of these cells contains five numbers, the first one indicates the absolute number of establishments so classified. the second and third numbers present the initial (1970)and find (1975) level respectively, employment in those plants: the fourth datum is the absolute increase in employment for those firms and the fifth one is simply the percentage increase in these firms employment with respect to the total employment increase in the 3721 firms between 1970 and 1975. The bottom right of the table succintly presents the measuring to be attached to each of the five numbers for each cell in the Table. Table A-22: EMPLOYMENT GROWTH OF PLANTS IDENTIFIED AS EXISTING IN BOTH 1970 AND 1975, BY NUMBER OF WORKERS 1970 All Plants Reporting to Plants Identified Percentage Plant Size DANE, 1970 also in 1975 Increase in (Number of Number of Employment Employment Employment, Absolute Workers) Plants Employment Est. in 1970 in 1975 1970-75 Increase 5-9 2,542 17,216 249 1,914 4,226 120.8 10-24 2,210 33,681 1,412 22,408 28,301 26.3 25-49 1,121 38,196 899 30,625 38,856 26.9 50-99 660 45,659 555 38,330 47,276 23.3 > 100 668 211,497 606 198,977 233,728 17.5 34,751 Total DANE 7,459 347,159 3,721 292,254 352,387 20.6 60,130 Adjusted Total 395, 400 518,300 31.1 122,900 F Source: Unpublished DANE data. - 166 - Table A-23: COMPOSITION OF FACTORY OUTPUT GROWTH, 1962-66 AND 1970-75 Industry 1962-66 1970-75 Percent of Percent Percent of Percent Total Growth Growth Total Growth Growth Food 16.46 44.9 15.84 56.6 Beverages 14.38 21.1 12.34 36.1 Tobacco 4.74 13.4 3.90 55.9 Textiles 13.15 10.6 13.18 23.0 Clothing/Footwear 5.88 39.4 3,60 34.1 Wood & Products ) (9 1.35 90.6 1.02 p35.2 Wooden Furniture) ) Paper & Products 2.98 52.6 2.91 43.9 Printing 2.98 21.3 2.85 39.4 Leather & Products 1.10 26.2 0.87 33.2 Rubber & Products 2.64 6.3 2.57 58.0 Chemicals 10.75 28.6 11.55 49.9 Petroleum & Coal Products 3.38 18.3 2.17 17.2 Non-metallic minerals 4.56 10.5 5.51 37.1 Basic Metals 1.97 25.4 2.80 13.4 Metal Products 11.40 40.9 15.24 67.5 Miscellaneous 2.28 - 3.65 - Total 25.9 43.3 Source: Weights for 1962 were based on figures for gross value added in 1962 at 1958 prices, as presented in DANE, BME #224, Mayo 1970, p. 138. Growth rates are from the national accounts. Weights for 1970 are from DANE, III Censo Industrial 1970. - 167 - Table A-24: IMPLICIT UNDERREPORTING IN DANE AND ICSS ENIPLOYMENT DATA, BY PLANT SIZE, 1975 /a Plant Size (Number of DANE Statistic 1 ICSS Statistic 1 Workers) Estimate of Table IV-2 Estimate of Table IV-2 5-9 .073 .747 10-14 .533 .917 15-19 .665 .937 20-24 .689 ) ) 1.030 25-49 .932 ) /a Assuming the estimates shown in Table IV-2 are accurate. Table A-25: REPORTED AVERAGE WAGE (INCLUDING FRINGE BENEFITS), BY PLANT SIZE, 1956, 1963, 1970 AND 1975 Plant Size (Number of 1956 1962 1970 1975 Workers) Wage Index Wage Index Wage Index Wage Index 5-9 1,945 55.7 3,993 46.9 9,837 38.3 25,296 50.3 10-14 2,224 63.7 4,761 55.9 11,551 44.9 27,078 53.9 15-19 2,412 69.1 4,948 58.1 12,633 49.2 26,984 53.7 20-24 2,528 72.4 5,234 61.5 13,117 51.0 29,143 58.0 25-49 2,803 80.3 6,096 71.6 15,578 60.6 32.606 64.9 50-74 3,099 88.7 6,942 81.5 18,552 72.2 38,858 77.3 75-99 3,476 99.5 7,808 91.7 20,658 80.4 42,666 84.9 100-199 3,492 100.0 8,514 100.0 25,700 100.0 50,258 100.0 > 200 4,246 121.6 10,113 118.8 37,429 145.6 74,526 148.3 Total 3,311 94.8 7,864 92.4 26,305 102.4 57,420 114.3 Source: For 1956 and 1962, data from DANE, Anuario General de Estadistica. For 1970, DANE III Censo Industrial, 1970 and for 1975, Industria Manufacturera 1975, Avance. - 169 - Table A-26: DESCRIPTION OF 30 MANUFACTURING SUBSECTORS 5-digit ISIC 1/ Description 31161 Mill products: wheat flour 31163 Mill products: rice flour 31168 Mill products: coffee 31171 Bakery products: wheat and corn bread 31172 Bakery products: cakes and pastries 31182 Sugar refining: raw brown sugar 31212 Beverages: ground and instant coffee 31343 Soft drinks 31403 Cigars and semi-finished cigarettes 32134 Textiles: knitted synthetic undergarments, shirts, sweaters and other outer garments 32201 Clothing: men's and boy's jackets, suits and trousers 32202 Clothing: women's and girl's blouses, dresses, coats, pants and skirts 32206 Clothing: men's and boy's plain cotton and synthetic shirts. 32402 Footwear: men's leather shoes and boo s 32403 Footwear: women's leather shoes 32404 Footwear: boy's leather shoes 33111 Sawmills: sawnwood 33116 Sawmills: doors and windows 33117 Sawmills: floors, partitions and other construction products 33202 Furniture: tables, chairs and other household furniture (non-metallic) - 170 - Table A-26: (Continued) 33203 Furniture: office, commercial and institutional furniture (non-metallic) 33206 Furniture: mattresses and bedsprings 34122 Paper products: cardboard packing containers 34203 Printing: checks, stamps and other commercial printing 35221 Pharmaceutical products: manufacturing of lecithin and enzyme 36912 Non-Metallic mineral products: clay construction products (bricks and tiles) 36992 Non-Metallic mineral products: concrete and cement construction products 38122 Metal products: manufacturing of household furniture and accessories 38131 Structural metal products: doors, windows and frames 38197 Metal products: wire, cable, nails and accessories 1/ This is a slightly revised version of the International Standard Industrial Classification (ISIC), Revision 2, carried out by DANE, - 171 - Table A-27: COLOM1BIA: NUMBER OF ESTABLISEMENTS AND EMLOYMENT IN 30 SELEGTED FACTORY SECTORS /1 /2 5 Digit Size of Establishne.t (Measured by Sectors Number of Employees) *5* * * * * * * * * (1 245 *25t49 (499 * * T7TAL * * * * 31161 * 43 17 * 4 * 0 * 64 * 584 557 * 261 * _* t1402Ž ** **** * ** ***** * * <* ** ** * ** ** ** ******* ** * * **** ***** * 5 * * * * 31163 * 116 * 17* 2 * 0 * 135 32591R. 531 * 147 * * 1937 ****** * *2 * **** ***** ********* * * 313168 * 11 * 2 * 22 * 7 * 61 * 107 * 738 * 1424 5525 * 3522 ******* * **,****** ***f* **************~****'* * *. ** 3W71 *. 662 * 37 * 21 * 3 * 713 515 1210 739 * *** * 31172 * 0 * 6 * 2 * 6 49 * 437 * 193 * 1.06 * 19i3 * 934 *********** ********************** ** ** * * *** * * * ** 311s2 * 107 * 6 * S'* 10 * 119 * 897 * 227 * 352 * 1250 * 2 **************** ** * *** * ** * ** ** ** ** **** * *** ** ** * *** ** ** * * * *, * 31212 * 37 * 2 * 2 * 47* 48 * .360 * 171 * 156 * 05 * 1492 ***** **** ****************************~ **** * ** 313143 * 15 * 10 * 8 22 * SS * 161 * 338 * 625 * 5586 * 6710 ** * * * 34(403 * 714 * 3 * 2 * 3 * 8 * 701 * 99 * 1.29 * 1.155 20* * * * **** * * * * * ** * * *** ******** ** <4* * * *ø<*<* * * ** ** 3213~4 * 31 * 16 * 9 * 6 * 6 * 357 * 3814 * 66* 15833 * 3100 * * * *k* ~ * ** '****<44 **** * ** ** ** ** ******* **ø* * * * * 320i* i50 * 31 * 26 * 28* 235 * 572 * 10711 *< 179 763~ * ~2027 * 4** * * * ***********« **~******* ** * 4 4* *** * * **** * * * *4 * 32202 * 76 * 30 * 11 * 4 * 121 - 172 - Table A-27: (Continued) 818 * 1087 * 753 * d85 * 3143 * ** * * 32206 * 33 * 36 * 21 * 13 * 103 452 . 1324 * 1425 * 2660 * 5870 ** ** * * * ****** * ~* * * * fr* * ** ** *** * ***** * ** **** * ** * * * * *r * 32402 * 82 7 * 12 * 7 * 108 * 778 * 26 * 761 * 1438 3233 ** *** * ** ********* ** r* *r* * * * * * * * ** ** *** ** * *** * *** * * * * * 32403 * 1U5 * 7 * 4 0 * 126 * 1053 * 211 * 267 * * 1531 ** ******** ****************************************** * * * * *r 32404 * 38 * 8 * 3 * 1 * 50 337 * 243 *· 219 * 102 * 901 * ** * ?* * ** ** * Nfr* * *** ** ** **** ** * * * r** **** ** ** * ** ** ** * * * * * 33111 * 73 * 14 * 4 93 919 * 475 * 127 * 808 * 2329 * *. * *r * 33116 * 38* 6 * 2 * 0 * 46 383 . 194 * 104 * 686 ** * * *** * **** * * *$* * ** r* ****** **** ** *** *** * ** ** * ** *** * 33117 * 57'* 6 * * 0 * 66 599 * 285 * 63 * * * 947 33202 * 28 *I 25.* 8 *2 * 283 * 210 * 807* 56 * 649(* 4162 ** *** ********* f********r**********r******************** * * * * * 33203 * 39 * 4 4 * * 48 ** ** 33203 * 32. * 3 * ( * j * 46 333 * 113* 6 * 508 * * *r* * *** ** * * * ** **.* * ** ** * * ** ***** * * r* r* *r* * *** **** * * * * * *r * 34122 * t * 1 * 49 932 * 43 * 40 * 2 2 * 5146 *** * ** ** * ** * fr****** * ** ** ** ** * ** * * ****** * r*r* * * * ** * * ** * 3123 * 291 * 35 * 5t * 9 * 346 * 30t3 *. 1158 * .790 * 366 * 8621 173 - Table A-27: (Continued) * * * * * 35221 2 2 * 14 * 11 * a * 67 * 30 * 465 * 755 * 2657 * 4257 * * * * * 35912 * 158 * 33 * 15 * 3 * 209 2 3 0 *, 1134 * 1037 * 48; 3934 36,992 * * 29 * 9 * 6 * 212 y1 6 099 *L 603 -!A 4 4i1 * I*.* * 331,22 * 29 * 14 * 10 * 55 * * * 329 *. 4 * 600 * . 553 * * 9 3131 * 1513 * 9 * 0 * 157 * 1270 . 317 * 90 * . * 2247 * k* ** * *** * * * ***** ******* * * ** ** ** ** * ** r* ** * * *** 38197 * 0 * 4 10 * 7 * 51 * ..359 l 42 * 6 * 1090 * 2266 * * * **** ***** * ****************** * * * * * T3TAE * 2994 * 471 * 2 4 4 150 * 3859 * 29015 * 16009 * 16460 * 35236 * 96720 Footnotes /1 For each 5 digit sector, the first row indicates the number of establishments while. the second one shows employment, cach classif ied according to size. /2 The sectors inciuded are the 30 five-digit industrial sectors which had the largest number of establishments (at least 46 each), These sectors account for 51.74% of the establishments and 27.86% of the employment in the entire factory sector. The firms of less than 25 workers in these 30 sectors employ 56.0% of all factory employment in firma of the same size category. In other words, the criterion of selection is - as expected - biased in favor of inclusion of small firms. - 174 - Table A-28: ESTIMATES OF LABOR AND CAPITAL PRODUCTIVITY, AND OF CAPITAL INTENSITY, BY PLANT SIZE, 1970 Size of Plant Value Added Value Added Book Value of Fixed (Number of Worker Book Value of Capital/Worker Workers) ('000 pesos) Fixed Capital ('000 pesos) 1-4 20.35) 1.33) 15.30) 5-9 23.66) 1.28) 18.48) 10-14 30.45) 28.27 1.36) 1.29 22.39) 21.86 15-19 31.37) 1.15) 27.28) 20-24 30.76) 1.42) 21.66) 25-49 40.35) 1.26) 32.02) 50-74 48.80) 1.34) 36.42) 75-99 58.11) 80.25 1.57) 1.37 37.01) 58.56 100-199 84.50) 1.73) 48.84) > 200 96.61) 1.30) 74.32) Total 72.49 1.37 53.09 Source: DANE, III Censo Industrial, 1970, op. cit. -175- Table A-29: CAPITAL LABOR RATIOS BY OUTPUT LEVEL OF PLANT, 1970 Output of Plant Book Value of Fixed Assets ('000 pesos) Worker < 200 10.02 200-400 14.86 400-600 15.21 600-900 17.33 900-1,500 18.83 1,500-2,000 21.94 2,000-4,000 23.65 4,000-14,000 36.03 10,000-50,000 50.76 > 50,000 86.34 Source: DANE, III Censo Industrial, 1970, p. 18. 1 У .1 в. '�дьt� А-з0: гхзLоегыlл: ЕвеlцRц s�г:еоК ItzcкLSStuNS о� слгlТаl. 1N'r�NS11'У 8У FT�S 1N seECT1'1а lиииsг8х Sи�счс 311b1 311оЭ 3 U ьд Э1171 з1112 31182 J1212 Э1347 Э1403 3'L134 32201, 32202 Э2206 1.Г�;1. - 1_ЗаЬ03 .1495в - .212ь2 .Э5555 - .85827 1.19а75 54991 .114а4 - 1.ЗьGВ9 .в6094 .25202 .02Ь95 - .27451 � 1.5?690 .8б296 .7Э538 ..Э1692 .7599а 2.i0696 .62655 .49142 .В4в57 .455дд .271?5 .5Э008 .4bdD4 г.сн; ь sq .22sоз - .оч9ое - .иа942 - .о50од 2оз1й - .1494в .с485� .аз925 .23в80 - .z421e - .01]ы - .оо455 .о5ЭЧ4 ` ,2б70д . .15747 ,10429 .06097 .12384 .Э3603 .09308 .06432 .113д4 .аб478 .0Э72Э .08692 ,06481 г 1ЕАК ,0255д А1129 .дU86? - .016Э0 - .0163tl - _04564 - .00331 .0155В - .015ы3 - .056д7 - ,ц0863 - .02180 - .00{93 � .01035 .009В6 .01261 ,00495 .0167Э .03388 .009ЭЭ .00905 .0142д .02464 .ООВ10 ,01059 .00980 СОи�САиС 11.50242 10.37275 10.19дЬы 1.76782 11.0101б ,1U.81045 9.2b12б 8.Э4556 10,G3637 12..09177 д.90350 10,ЭЧЭ)6 9.38054 К2 .120 .025 .094 .028 .199 .117 .096 .ЗЗЬ .Э57 .152 .05В .04В .ОЭВ l' 2.359 .В28 1 Э91 5.ЗВ7 Э.З2'2 1.057 1.4Э7 б.564 5.36Э 2.86В 3.423 1.й84 1.117 _ ИигroЕК UE' S6 99 44 567 . 44 28 4Э 4Э 3Э 52 171 92 88 E5tABLTSl4ЧE17t5 • i �._�_.,.�- - - Stг:cor Э2402 3240З 32404 3Э111 3311Ь 3Э117 ЭЭ202 33203 3Э206 3412_ Э420Э 35221 36912 ? <_цL'L 1.14db3 .1882д Э.27143 -.595Э9 - .70990 - 1.40520 - .576ВВ .21Вб7 1.l39В7 1,�L`106 - .41902 - ,31060 - 1.3578Э .4962б .82р5б 1.1411b .54954 1.69160 1.42173 .3Э4В0 1. U057 1,9148ь 1.00292 .21095 .Н1142 .75702 шG L 8Q - .11495 �3'�ОЗ - .SU118 ,11а40 .1594? .301В5 .07991 - .04117 - .D738S - ,1ЭВ04 .07991 .074ВЭ .26669 ,07590 .14936 ,19135 .OBGG6 .Э1913 .25897 .05695 ,20у08 .35_д^_ .1546] .03184 .L0503 .12161 � YL•'AR - .01916 - .0Э505 - .01297 -.д16д5 .01349 - .01051 - .0180Э - .0:'.986 - .J4009 .02375 - ,00418 .,00908 - .0041Э .01264 .01£92 .011.74 .01Э13 .02067 .01Э93 ,00650 .01678 ,01844 .01123 .00458 .D1064 ,01457 , с0ИSглиr 7.99ыыь 1о.1гбs7 4.ь1ь2ь 11.ч0297 9.д31ы 1l,349ыВ 1о.^,а2оВ 1о.ы 9зz 9.41га4 б.2ьз5о 1o,569s3 я.бо2lо и.зч416 � К2 ,206 .177 .,31ы ,д37 .0Э5 .U83 .052 .128 .29Э .156 ,U40 .138 .151 , F 5.777 5.169 . 4.204 .927 - .367 1.Э35 3.94В 1.271 3.')97 2.3Э9 3.922 2.245 4.22д NОНыЕК UЬ` 71 7b 31 77 З4 48 221 30 Э3 42 ,2В8 46 75 ESfABLISfцIENCS Sес�ог ЭЬ992 3110 Эы1'L2 Зд131 ЭВ197 ; 1АС L .14465 .55877 - ,Чд22Ч .40113 - .OS82U .40687 .42022 .7684D .517Э4 .90i12 1.:,а У. sц .Озив5 - .а177о .16531 - .ОьвВ1 .о7�40 ` .оь25о .�7+597 _12ва1 .о9ьЭ1 .13409 %ЕлК - .О1Ь06 ,U0413 - .051Э8 - .029Э0 .OUB6b � _00737 ,01512 .01390 .0492б . .01894 c0us1'Аиf 9.51023 8.13971 13.В765'L 10.63761 9.07ь06 К2 .144 .326 .291 .Оы8 .190 К F 9.57Ь 5.646 4.'782 4.2ЭЭ 2.976 NUh1llEК OF 175 ` 39 3S 1Э6 42 Е5'СА В7.15(ц4Еиf S ,.,,.,._... UnpuLlished UANE двса fcan 1970 Iг,дизСдаl Cenaus. i г ' 1 I 'L:-1_л�]l. COl11Htl1A: FйCtt1KY SЕССОК 10:C1:5S10N5 ON IЛ80К 1'KOOOL"f1VI'1У (ШС V) 8У FC1Ut5 IИ 5РlCIp1C 1NUU51'K1ES • � N 5еасиг 311b1 31165 з11б8 91U1 31172 31182 31I12 9IЭ4з 3140з 32134 3220I Э2202 32206 1.лК L 3.7бу67 .Э9абы .ЧЬдб8 .1Ь946 -.240з0 .15955 .57373 .tl3939 -.з9233 .26024 .31172 -.45ыЬЧ -.29136 1.04347 .86420 .54а17 .19753 .47157 1.111а7 .42504 ,29120 .56ы47 ,3113Э .U83b .34)4Э .76975 Lлg L St) -.,б0912 .U8412 .06787 -.00240 .0Ы82 .0 р 25 -.02274 -.057Э6 .10Э02 -.Оо4ы5 .0590Э .ОtlыЭ1 .06722 .1824В .15776 ,07766 .03760 .07484 .19279 .06275 .09ы27 ,0779Э .044д5 .021б9 .05697 .051Э0 Усдг -.OU633 -.013Ь4 -,0126tl -.ООЭЧО -.005зб .Wb6l .OO6U8 ,ООо55 .ОООыб -.UUOуB .0Ы84 -,OI097 -,00137 .ОО7й2 .00994 _00444 .003D8 .00990 .02008 .00628 .U0556 .00975 .01714 .00477 .00711 .00Т73 1.лы (к!L) .о9zгь .ьэвsы .sгz91 .2ойз8 .хч910 .4ьs15 .о7989 .07гоs .7в0ЭS .3xzзs .2яsы7 .ЗЭо51 .20570 - .09411 -10273 .11774 .02597 .09250 .116Ы .10757 .094ы2 .11918 ,09527 .04506 ,069а7 .Оа602 Силисr»[ 4.842Ы 5.01100 й.ыы721 7.77ы71 8.09172 3.В8355 7.дЧ779 7.2125Ч 6.22464 6.14024 7.Э7671 ы,Uу04I. ы.21з54 2 - К .25400 294о2 .41318 .1426ы 31894 .SOOtlB .44717 .64044 .64-fl8 .36151 .70895 .29526 ,21Э51 - F 4.34122 9.78Т18 Ь.86502 2Э.Здзб2 4.58494 5.7Т034 7.б8423 16.957Ttl 12.711)9 6.652д0 18.55Э97 Ч.i122G 5.Ы29Э Ил. of Eвta611ы1uucncr 5б 99 44 5b7 44 2д 43 43 Э3 52 П1 92 В8 э # _� ' - �- Э2402 92403 З2404 33111 33116 Э3117 - 3Э202 ЭЭ20Э 1Э20б 74122 Э420Э 35221 36912 Log L -.23854 1.Э9505 .У19з1 -.41i)00 .98аЧ5 -.44590 -.41]ЬВ .20924 .26062 -.1tl758 .18д47 L 1203] -.01051 .z7so3 .ь4712 .7 Ыз9 .Зчs2о .ы 26ь .7ьбs7 .х14ье .786az 1.4784г .аеб4а .1з47з .ьзвб9 .з90Э4 3 Log L 5Q • АЧ093 -.2SЛ2 -.02253 .07Э19 .22178 .097Э0 .04740 -.O1004 .02444 .04287 .01104 -.09445 .ОЗд61 ,04142 .11778 .12518 .05444 .12706 .14022 .0Эб4Э .11712 .27305 АТ4Э9 .02042 .08302 .0Ы37 Усвг -�.012Э0 �-.00970 -.ООд22 -.00690 .UD374 -.OU149 .00001 -.U03U3 -.O15B0 .Q0768 .00229 ,UОу62 -.00324 , -� .ООбыS .0137Э .00701 ,ООы25 .00825 .0074д .00421 .0119Э .01526 .00565 .00291 .00843 .001Э5 " 1 - Log (К/L) .ЭЭ716 .29447 .1320Э .181Э4 .09116 J7974 .24961 .Э'l5д1 .2060Э .29ы92 .18В00 .1854В .20627 .04515 .09241 .11243 .07273 .07239 .09040 .04Э2Э .1Э167 .14250 .07724 .03764 .121°5 .059д5 Слпsгв»t 7.ы)0]2 6,06300 tl.b6071 9.30051 i.70187 tl.75893 7,tl8727 6.16128 8.0874Э 7.8з144 7.48146 5.933Ь[ 7,61125 К2 ' .4Э205 .20915 .21532 .12Э46 .25д44 .1613б .19Э90 .7Э4Ь2 .29922 .38577 .224з4 .44'245 .97446 У 12_55 и9 Ч.ь9411 1,78з67 2.5Э51в 2.5266ы 2.ОЬы30 12.9вв7t 9.14919 2.9ыыви 5.ао951 zo.i625o 8.150з] ю.47sв5 Nл. л1 Еsсдб11в1иыепсв Л 76 31 Т] 34 48 22l ЭО з3 42 28д 46 75 Зь992 3Т1о зs12г зы и1 зв197 1лВ 4 -.16696 .251а8� .Э0222 -_5599Э .18514 .22155 .23952 .й689з � .3767b .44йВ9 1.оы i. 5Q .05095 -.01947 -.00Э09 .15502 -.Оо5д5 .03406 .0256Т .ОТВ15 .06081 ,0)390 Усдг � -,00Э04 -.00728 .00011 .00106 -,00244 ,00407 00&42 .0047В .00605 .01043 WS iК�1J .27515 .J2703 .57281 _21291 .457ST .0{16Э ,09400 .100ы3 .05485 .08909 Сллдwпс 7.50277 6.952ы5 4.05910 8.47 U 2 5.426В7 К2� .3Э260 .47241 .Ы500 .2029Э .55767 F 21.18U32 7.Ь1д9Л 1з.57765 8.Э3ы01 1L бб214 ил. ui е=сдбииьл,ввсы t7s з9 эч tзь ч2 5илгсе; Uupu621yhed UANE [Ю�д F[лщ InJ�sertыl Genвun, ]970. • 178 - Table A-32: COLOMBTA. RETMS TO SCALE AND PRODUCTIVITY DIFFERENTIALS BETWEEN 'SMALL' AND 'LARGE' FACTORIES, AS DERIVED FROM REGRESSION ESTIMATES Estimated Scale Elasticities Ratio of labor producti- in firms of. vities in plants of 100 25 workers 100 workers and 25 workers /1 ( -; , 1) ( 2) 33-161 0.848 /2 31163 1.143 1.376 i743 31168 1.031 0.843 0.92 3X171 1.154 1.147 1.23 31172 1.170 1.347 1.43 31182 1.271 1.318 1.50 31212 1.427 1.364 1.73 31343 1.470 1.311 1.72 31403 1.560 1.846 1.77 32134 1.229 1.215 1.36 32201 1.068 1.232 1.23 32202 1.110 1.355 1.38 32206 1.141 1.328 1.38 32402 1.025 1.138 1.12 32403 0.740 /2 /2 32404 1.074 1.0i_2 l.b-6 331-11 1.001 1.204 1.15 33116 1.439 2.054 2.81 33117 1.180 1.450 1.55 33202 1.213 1.483 1.62 33203 1.143 1.117 1.20 33206 1.418 1.486 1.87 34122 1.088 1.207 1.23 34203 1.260 1.291 1.46 35221 1.512 1.250 1.69 36912, 1.240 1.343 1.49 36992 1.161 1.30Z 1,38 3710 1.126 1.072 1.15 38122 1.282 1.274 1.47 38131 1.338 1.768 2.15 38197 1.147 1.131 1.21 Source: Regressions of log (V/N) against log K/N, log Nj(logN) and (Y) year when firm started operations (Table A-31). Original data from 1970 industrial Census. In order to isolate the impact of size on product- ivity from other possible effects, it has been assumed that both (K/N) and (Y) are the same in both sizes of firms. /I (Value added per worker in firms of 100 employees)/(Value added per worker in firms of 25 workers). In order to assess the independent influence of size on productivity, it is being assumed that the capital-labor ratios are equal in both sizes of plants. /2 !a sectors 31161 and 32403 there are decreasing returns to scale in firms of 25 workers. Larger firms show even lower returns. In effec ,, estimates for firms of 100 workers indicate that their productivity of labor would ba I Iss chan half of thac in s=11 firms. Inspcccion of :hc or ,:inal daca reveals, however, that chcrc arc only four "large' firns of az loasc 1fifty workers in each of these two sectors, having an average 65 and 67 workers per firm, respectively. Consequently the estimates for firms of 100 workers are not presented since it seems clear that these firms are not viable. -179- Table A-33: COLOMBIA: SECTORS CLASSIFIED ACCORDING TO LABOR PRODUCTIVITY DIFFERENTIALS BETWEEN FIRMS OF 100 AND 25 WORKERS /1 (1) (2) (3) (4) (5) 31161** 31171 31163 31212 31403 31168 32201** 31172 31343** 33116* /2 32403** 32402 31182 33117 33206 32404 32134 33202** 38131** 33111 32202 35221* 33202 32206 34122 34203 3710 36912 38197 36992 38122 /1 X = (Value added per worker in firms of 100 workers)/(Value added per worker in firms of 25 workers). /2 One and two asterisks indicate that in the regressions explaining labor productivity (log V/N), at least one scale variable is statistically significant at the 5% and 1% level, respectively. Source: Table A-31. Table A-34: RELATION OF HORSEPOWER AND BOOK VALUE OF FIXED ASSETS, BY TWO-DIGIT SECTOR, CERCA 1970 Total Energy Book Value Percent of Output Energy Consumed/ Capacity,1968 of Fixed in Plants of 200 Energy Capacity, 1966 (1000 horsepower) Assets: 1970 VFA 70/HP68 or more Workers (KWH/HR) (1) (2) (3) (4) (5) Food 185.4 3,361.6 1.813 54.6 1,437 Beverages 65.7 1,268.1 1.930 72.6 1,492 Tobacco 5.09 198.5 3.900 84.9 1,127 Textiles 161.3 2,667.6 1.654 89.8 2,739 Clothing/Footwear 12.53 408.9 3.263 56.5 1,914 Wood & Products 31.04 378.5 1.219 55.4 756 Wooden Furniture 7,94 76.8 .967 27.9 749 Paper & Products 74.89 759.2 1.014 59.2 1,641 Printing 11.94 533.0 4.464 63.6 1,942 Leather & Products 16.66 113.1 .679 63.2 952 Rubber & Products 31.52 270.4 .858 83.6 1,826 1 Chemicals 108.2 2,189.5 2.023 54.8 3,192 H Petroleum & Coal Products 58.51 992.8 1.697 98.2 1,340 1 Non-Metallic minerals 166.5 1,351.7 .811 71.0 1,995 Base Metals 48.19 1,414.5 2.935 78.4 2,592 Metal Products, Except Machinery & Trans- portation Equipment 81.00 808.6 .998 40.8 1,161 Non-electric Machinery 17.53 443.1 2.528 44.0 1,034 Electric Machinery 19.86 393.2 1.980 55.4 1,997 Transportation Equipment 19.69 296.7 1.507 73.7 1,341 Miscellaneous 19.65 502.7 2.558 1,813 Total 1,143.2 18,428.8 1,612 66.6 1,932 Sources: DANE, Industria Manufacturera Macional, 1968 and III Censo Industrial, 1970. Table A-35: ESTIMATES OF FACTORY EMPLOYMENT, 1964, 1966, 1975 Average Annual Large Plants Average Annual Growth Rate (> 100 workers) Growth Rate (Percent) (Percent) 1964 1966 1975 /a 1966-75 1966 1975 1966-75 Food 47,000 48,840 67,670 3.7 16,831 36,108 8.8 Beverages 18,100 18,362 21,046 1.5 13,182 18,992 4.1 Tobacco 5,400 5,350 3,713 /c -4.2 1,971 3,143 5.3 Textiles 44,100 47,083 77,606 5.7 38,850 63,931 5.7 Clothing & Footwear 33,000 35,850 65,512 /b 6.9 11,181 26,113 (29,590) 9.9 Wood/Products 6,500 6,733 9,113 3.4 2,547 2,724 0.7 Wooden Furniture 6,000 6,817 15,312 /b 9.4 1,098 2,134 ( 3,128) 7.7 Paper 5,500 5,994 11,045 7.0 3,767 6,764 6.7 Printing 13,000 13,644 20,232 4.5 5,821 11,124 7.5 Leather 4,600 4,945 8,474 6.2 2,382 5,331 9.4 Rubber 6,900 7,025 8,308 1.9 5,482 5,962 0.9 Chemicals 20,000 21,405 35,936 lb 5.9 14,010 25,353 (24,979) 6.8 Oil & Coal 2,100 2,316 4,526 /c 7.7 1,721 3,935 9.6 Non-Metallic Minerals 27,500 27,941 32,462 1.7 14,346 20,294 3.9 Basic Metals 5,000 6,065 16,961 12.1 3,565 14,390 16.8 Metal Products 20,000 21,186 33,326 5.2 9,016 14,048 5.0 Non-Electric Machinery 6,500 7,400 16,617 9.4 2,269 7,663 14.5 Electric Machinery 9,200 9,791 15,838 5.5 6,204 7,911 2.7 Transportation Equip. 18,000 18,383 22,304 /d 2.2 /d 7,131 13,581 7.4 Miscellaneous 9,000 10,337 24,014 9.8 3,103 12,115 16.3 Total 307,400 326,097 510,015 5.1 164,477 301,616 7.0 Ia Where not otherwise noted, the figure is based on DANE statistics, but replacing DANE's figures for the size categories 5-9 and 10-14 workers with those of ICSS. /b ICSS figure. /c DANE figure. /d Figure for 1975 appears to be downward biased relative .to 1964 and 1966 due to change of definition of category. Sources and Methodology: Figures for 1964 are estimated on the basis of DANE statistics and figures from the population census of 1964. DANE statistics were adjusted upward where the job position composition of persons recorded in the census but not in DANE's survey suggested undercoverage, i.e., where a small share of such persons were independent workers and a high share were paid employees. The total figure is a little below our best estimate of 310 thousand. Figures for some industries in 1975 are those of Table A-18a, including employment reported in plants of less than five workers (for industries where DANE's figures exceeded that of ICSS (Table A-18a) the former was used, adjusted upward slightly in most cases, this adjustment brings the total almost exactly to our best estimate, and it seems a reasonable assumption that the industries with large employment in plants of 1-4 workers would be the ones where undercoverage of plants of five workers and up would be greatest. The 1966 figure is an interpolation between the other two, taking account of the slow growth of total factory employment over 1964-66 and the growth over 1964-75 for each industry. 0O Table A-36: LABOR PRODUCTIVITY BY INDUSTRY AND SIZE OF PLANT, 1956 Labor Size of Plant (Workers) Industry Productivity: Food Firms >200 5-9 10-14 15-19 20-24 25-49 50-74 75-99 100-199 > 200 Total Firms 5-9 Food 2.65 5.67 6.84 9.14 7.02 10.92 15.22 16.23 14.92 15.03 10.43 ,Beverages 5.43 5.33 8.29 8.86 5.62 23.57 30.94 51.89 33.42 28.96 29.95 Tobacco 35.14 2.97 2.37 2.52 7.45 6.93 20.85 34.25 27.70 104.36 33.21 Textiles 3.08 3.84 4.67 6.99 6.95 6.31 5.18 4.48 8.04 11.83 10.39 'Clothing/Footwear 2.50 3.49 3.42 3.55 3.47 3.74 4.04 4.04 5.91 8.74 4.20 Wood/Products 0.36 5.08 3.06 4.09 7.17 4.54 7.56 4.33 4.95 1.83 3.57 Wooden Furniture 1.58 4.02 3.79 4.18 4.71 4.60 6.23 5.38 1.41 6.37 4.86 Paper 3.85 3.76 7.84 2.75 3.70 7.81 7.14 - 5.57 14.48 8.74 Printing 3.20 4.40 4.16 4.80 4.00 6.27 8.37 12.86 10.14 14.06 8.22 Leather 3.93 3.83 4.54 4.92 5.38 5.20 9.56 4.84 8.47 15.07 8.04 Rubber 1.33 9.92 8.53 8.27 12.27 3.23 11.83 - 6.93 13.23 12.19 Chemicals 2.73 6.56 8.40 8.81 9.09 11.11 14.86 11.90 16.25 17.94 12.85 Oil & Coal Products 4.04 7.40 71.32 6.09 21.28 34.44 - - - 29.90 29.67 Non-Metallic Minerals 3.81 3.47 3.29 3.32 3.84 5.34 4.88 4.86 6.92 13.21 7.43 Basic Metals 1.17 8.75 5.78 11.29 5.56 9.09 9.39 6.10 5.18 10.21 9.71 Metal Products 2.32 3.70 4.95 6.43 5.29 6.02 6.36 6.10 8.26 8.59 6.30 Non-Electric Machinery 2.31 7.29 3.96 4.31 5.83 6.51 5.35 - 6.42 16.83 6.95 "Electric Machinery 1.23 5.96 8.63 5.76 8.04 8.16 10.12 - 10.85 7.36 7.85 Transportation Equipment 1.48 3.57 4.49 5.24 4.58 4.94 5.67 4.91 6.44 5.29 4.86 Other 1.83 4.67 5.36 5.55 10.17 9.23 16.89 8.21 5.79 8.56 8.03 Total 3.27 4.54 5.08 5.72 7.46 7.73 9.74 15.70 12.31 14.84 10.31 - 184 - APPENDIX B MEASURES OF CAPITAL USED IN THE FACTOR SECTOR The concept of book value as reported in DANE's 1970 industrial survey appears to coincide with that used by the Superintendencia de Sociedades Anonimas in reporting value of fixed assets for the Sociedades Anonimas. In 1965 and 1966, for which years we have data on value added and fixed assets for the same sets of sociedades, the ratio of the former to the latter was 1.47 and 1.45 respectively. 1/ In 1970 DANE reports a ratio of value added to book value of fixed assets for sociedades anonimas of 1.45. The Superintendencia figure includes non depreciating fixed assets and depreciating ones; the depreciated value of the latter is simply historical cost minus depreciation (based presumably on tax legis- lation); while in the calculation of total assets a category entitled "1valorization and devalorization" is presented, one which does appear to include an item for writing up the value of some fixed assets to allow for price increases, any such write-ups do not enter the recorded value of fixed assets. Since sociedades seem to have used essentially the same figures in recording book value of fixed assets to DANE it must therefore be presumed that these data are not adjusted for inflation, as well as reflecting non-economic depreciation rules. 1/ See Superintendencia de Sociedades Anonimas, Revista de la Superin- tendencia de Sociedades Anonimas, 1966, Bogota, D.E., Imprenta Nacional, 1968, p. 45, 55. -185- The Superintendencia data show information on other assets of sociedades anonimas. As of 1964 the structure of assets which would appear to be directly related to productive activities was approximately as shown in Table B-1. Table B-1 Millions of Pesos Cash 478.4 Short-term credit extended 3,668.7 (a) Apparently related to productive activities 2,634.1 (b) Other 1,034.6 Inventories 3,264.9 Investments (stocks, bonds) /b 766.9 Non-depreciable fixed assets: /d historical cost minus depreciation plus valorization 1,390.0 (645.7) Depreciable fixed assets: 4,115.7 /c(3,446.2) Total Assets Probably Related to Production If 11,887.1 Excluding land (Z500) /e 11,387.1 ( ) Historical cost minus depreciation. /a Includes loans to stockholders, loans to employees, "various" loans, etc. /b Includes several other items of insignificant magnitude. /c Histcrical cost was 5,180.6. /d Includes some assets which are reproducable but not yet completed, as well as land. /e A little less than twice the reported (historical?) value. If Intangibles and the like are excluded. -186- The figures may overestimate the need for cash in production (since it is all included) or be too high or too low on short term credit related to production. And fixed assets are probably understated, even after allowance for valorization. Taking these figures literally, however, the ratio of fixed assets to total capital (excluding land) was .44. If the valorization of fixed assets is not taken account of, their share of the corresponding total would be .36. Their share of total reported assets, including intangibles is .28. The above discussion relates to what might be termed gross productive capital, but the representative firm is of course a debtor as well as a creditor, and the net capital tied up in its production process is the more relevant quantity. It could be argued that credit from clients and suppliers should be subtracted (but not credit from banks) to arrive at a valid figure. For 1964, this would decrease capital by nearly one billion pesos (975.8 millions), to 10.4 billion. In that case the distribution would be fixed capital .482 inventories .314 cash, net credit, etc. .204 It seems probable that the fixed capital share may be still a little understated in these figures, though perhaps not too seriously. It seems likely that the inventory/value added ratio is higher for sociedades anonimas than for other producers. In 1964 it appears to have been about 0.59. 1/ For manufacturing as a whole in that year 1/ Accepting the DANE figure for value added and the Superintendencia figure for inventories (including national and "foreign" sociedades), see Superintendencia, ibid, p.92, 141. - 187il it was 0.412, based on DANE data. 1/ This would imply a ratio of only .288 for other plants. In 1970, the ratio for all manufacturing, according to DANE data, was .457. Overall, it would seem plausible to assume that reported book value of fixed capital is between one-quarter and one-half of total capital involved in production; a best guess might be one-third. And its relation to total capital would appear to be a decreasing function of plant or firm size. In 1964 gross value added (DANE statistics) of the sociedades anonimas was 5,821.4 millions, implying an output/capital ratio of .560 of the last cited figure for capital was accepted. DANE value added figures are upward biased and are gross of depreciation so the net value added/capital ratio would be below .5. 1/ The inventory estimate is from Berry, "A Description" Statistical Appendix, Table A-53.