0s40 0 a-, LA; Urrutia Winners and Losers in Colombia's Economic Growth of the 1970s - Winners and Losers in Colombia's Economic Growth of the 1970s A WORLD BANK PUBLICATION Winners and Losers in Colombia's Economic Growth of the 1 970s Miguel Urrutia Published for The World Bank OXFORD UNIVERSITY PRESS Oxford University Press NEW YORK OXFORD LONDON GLASGOW TORONTO MELBOURNE WELLINGTON HONG KONG TOKYO KUALA LUMPUR SINGAPORE JAKARTA DELHI BOMBAY CALCUTTA MADRAS KARACHI NAIROBI DAR ES SALAAM CAPE TOWN Copyright © 1985 by the Intemational Bank for Reconstruction and Development / The World Bank 1818 H Street, N.W, Washington, D.C. 20433, U.S.A. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying recording or otherwise, without the prior permission of Oxford University Press. Manufactured in the United States of America The World Bank does not accept responsibility for the views expressed herein, which are those of the author and should not be attributed to the World Bank or to its affiliated organizations. The findings, interpretations, and conclusions are the results of research supported by the Bank, they do not necessarily represent official policy of the Bank. The designations employed, the presentation of material, and any maps used in this document are solely for the convenience of the reader and do not imply the expression of any opinion whatsoever on the part of the World Bank or its affiliates conceming the legal status of any country, territory, city, area, or of its authorities, or concerning the delimitation of its boundaries or national affiliation. First printing January 1985 EDITOR Emmanuel D'Silva PRODUCTION Yamile M. Kahn BOOK DESIGN Yamile M. Kahn FIGURES S. A. D. Subasinghe COVER DESIGN Joyce C. Eisen Library of Congress Cataloging in Publication Data Urrutia, Miguel. Winners and losers in Colombia's economic growth of the 1970s. Bibliography: p. Includes index. 1. Colombia-Economic policy. 2. Income distribution- Colombia. 1. Title. HC197.U789 1984 339.2'2'09861 84-12539 ISBN 0-19-520468-9 Contents Foreword vii Preface ix Chapter 1. Introduction 3 Pessimistic Opinions 3 Methods and Conclusions of This Study 4 Chapter 2. Changes in Incomes of Different Occupations 9 The Poor in the Rural Sector 9 Occupations of the Urban Poor 18 The Urban Middle Class 29 Conclusion 52 Chapter 3. Changes in Real Income of a Set of Rich and Poor Families in Cali 55 Characteristics of the Sample 56 Trends in the Real Income of Two Social Classes 59 Changes in Spending for Food 61 Other Indicators of Welfare 62 Single-Parent Families 65 Occupational Mobility 66 Comparison of Income Data from the Survey and Wage Series 67 Conclusion 69 Chapter 4. The Evidence from Household Surveys 73 Changes in Income Distribution between 1964 and 1972 74 Alternative Comparison of Changes in Income Distribution in the 1960s 85 Changes in Income Distribution between 1971 and 1978 85 Conclusion 90 Chapter 5. Changes in Colombia's Urban Poverty 93 Earlier Studies on Poverty Levels 93 Urban Poverty 99 Conclusion 113 v vi WINNERS AND LOSERS IN COLOMBIA'S GROWTH Chapter 6. Some Hypotheses on the Determinants of Changes in Income Distribution in Colombia 117 Labor Force Trends in the Countryside 117 Labor Force Trends in the Cities 119 Sectoral Growth in the 1970s 120 The Impact of the Supply of Education 122 The Effect of Growing Inflation Rates on Income Distribution 123 The Disappearance of Dualism in the Labor Market 124 Import Substitution and Distribution 126 The Impact of Monetary and Credit Policy 127 The Impact of Fiscal Policy 128 Distribution of Income from Capital 129 Conclusion 130 References 135 Index 139 Foreword SOCIAL SCIENTISTS HAVE HAD A DIFFICULT TIME fitting Colombian develop- ment experience of the past thirty years into fashionable categories such as monetarism, structuralism, bureaucratic-authoritarianism, and such. Here is another Latin American country without price stability, but where annual inflation has seldom gone above 35 percent. This is a nation with import and exchange controls, but where the extremes of South American protectionism have been avoided. Colombia does not allow its peso to float freely, but its crawling peg keeps it not too far from reasonable levels. Direct foreign investment is welcome, but selectively. Colombia has not launched grand social experiments, but its health and education indicators show steady improvements. Faced by this disconcerting eclecticism, too many international social scientists have ignored Colombia. That the eclectic Colombian eystem has, for nearly thirty years, generated significant and steady growth while maintaining a reasonably democratic society seems to have decreased its appeal both to those fascinated by socialist revolutions and to those impatient for a restoration of pure laissez-faire in the economic sphere. Neither righteous European socialists nor eminent North American professors feel compelled to study (for two weeks) the Colombian case firsthand. Mercifully, the international press carries no articles about a Colombian miracle. It is a safe bet that either the Allende or the Pinochet administration in Chile has generated more pages in international journals than fifty years of Colombian growth. When the Colombian experience is brought up in learned symposia, it is more likely to be dismissed with a silly-clever remark about drugs than discussed seriously. Colombians, quite wisely, enjoy their solitude. vii viii FOREWORD Miguel Urrutia is one of the few scholars who has offered the international academic community systematic and balanced analyses of Colombian development With this book he does it again, this time challenging the notion (prejudice?) that Colombian growth during the 1970s served only to make the rich richer. He plunges into the tricky area of income distribution, unfazed by data problems, and sorts out with aplomb the bits and pieces of available evidence. His superb knowledge of Colombia and his analytical skills add weight to his cautious conclusions that in fact overall income distribution may have improved at the end of the 1970s, and that the 1970s appear to have witnessed a reduction in Colombian absolute poverty. His plausible conjecture that the Colombian democratic system may have had something to do with those two trends will not be popular among those who sneer at "bourgeois reformism" nor among those whose dogma is that politics can only worsen market outcomes. Students of long-term development will find Urrutia's book most useful. Those anguished by the melancholy picture offered by Latin American economic performance and prospects during the early 1980s may take comfort in the long view offered by this volume. While avoiding the spectacular crises found elsewhere in Latin America, thanks largely to fairly prudent management, Colombian development has been negatively influenced by the bad news coming from abroad, especially from its immediate neighbors. At a time when the intemational economic system seems bent on rewarding political submission more than prudent economic management, Colombia faces difficult tradeoffs between austerity and autonomy. A serene reading of Urrutia's volume will offer an authoritative reminder that achieving both growth and the reduction of poverty, while maintaining national autonomy, is something Colombian policymakers have been able to do in the past and can continue to do in the future. CARLOS F. DIAZ ALEJANDRO Columbia University September 1984 Preface THIS BOOK IS THE RESULT of a collaborative effort among various researchers connected with Fedesarrollo, a private nonprofit research institute in Bogota, Colombia. A good part of the research was financed by the World Bank, but the last phase of preparation was possible thanks to the time the United Nations University allowed me to take away from my duties as vice rector for development studies at its headquarters in Tokyo. Clara Elsa de Sandoval, an old friend who was my research assistant in my previous book on income distribution in Colombia, helped me to calculate the rural income distribution trends. Mauricio Carrizosa analyzed poverty levels, based on data from household surveys, and Martha de Higuera and Maria del Rosario Sintes helped put together the data on wages by occupation. Hernando Gomez Duque managed the data and the field work for the Cali survey of rich and poor families. Finally, Rakesh Mohan and Alvaro Pachon kindly made available much valuable data from their own research projects. That the book is better than the first draft is largely due to the thoughtful comments of the anonymous peer reviewers assigned by the World Bank. They made me work hard to produce the revised version, but it was worth it. Finally, I am grateful for the comments and support of Fred Jaspersen, Nicholas Carter, and Guy Pfeffermann. Because I am old-fashioned, I still write in longhand. My secretaries, Arcelia Ramirez, Yoshie Sawada, and Akiko Hara, have been very patient and helpful. My only regret is that the deep economic crisis of the early 1980s will make it necessary for me to reexamine the income distribution trends in Colombia to see whether the improvements in income distribution ix x PREFACE registered in the late 1970s have continued into the new decade. For the first time in Colombian postwar history, the nation has had three years in a row of low economic growth, and the impact of this disaster on income distribution is not clear. The very rich have not done well since 1980, and many of the owners of the new financial groups are now in jail or in exile, having been able to take little money with them. Profits are down and many industries have gone broke. Unfortunately, the situation for the poor is not good either. Rural real wages have stopped increasing and unemployment is up. But since the inflation rate has come down, some real wages have increased. Consequently, it is difficult to tell what has happened to income distribution as a whole. In this book I analyze the trends in income distribution during a period of rapid economic growth. Toward 1986,1 will have to analyze what happened to Colombia when economic stagnation set in. MIGUEL URRUTIA September 1984 Winners and Losers in Colombia's Economic Growth of the 1970s 1 Introduction COLOMBIA EXPERIENCED SUBSTANTIAL ECONOMIC GROWTH during the decade of the 1970s, but the question of who has benefited from this growth is the subject of some discussion. The weight of opinion is that the growth in the 1970s did not help the poor or improve the distribution of income. This study examines a wide array of data and arrives at a different conclusion-namely, that the income distribution did not deteriorate and that the real incomes of the poorest workers did, in fact, improve significantly during the decade. An examination of who wins and who loses when a developing country experiences economic growth is always important. It is also appropriate for Colombia for two reasons. First, two governments in the 1970s-the administrations of Misael Pastrana (1970-74) and of Alfonso Lopez Michelsen (1974-78)-followed policies specifically designed to improve the distribution of income. Second, the democratic nature of Colombia's political institutions should help eliminate biases toward income con- centration. Pessimistic Opinions Many informed people in Colombia believe, however, that the recent experience of economic growth has not clearly benefited the lower-income families. A few sample statements will illustrate this. In September 1980, the Colombian weekly Consigna characterized the country's development process as follows: For the Colombian people-except the elite-the decade from 1970 to 1980 was a bad experience: the quality of life did not improve; disparities among regions and between urban and rural areas widened; child labor was an appalling phenomenon; and the middle class was proletarianized 3 4 WINNERS AND LOSERS IN COLOMBIA'S GROWTH because of the worsening situation of the white-collar workers. These are the conclusions reached by one of the most respected research institutes in the country, under the direction of the National Association of Financial Institutions (ANIF).' In May 1980, the liberal newspaper El Tiempo summarized the introduction of a new book by Senator Hernando Agudelo Villa, an ideologist of the left wing of the Liberal party, in which he states: The social situation in Colombia is more distressing and unfortunate every day because economic growth has not meant a growing and equitable improvement for the low-income sectors of the population, because inflation has deteriorated real wages, and because the phenom- enon of a bad distribution of income has become more clear.2 Even the major employer federations seem to be concerned about the problem of a deteriorating income distribution. In April 1981, five major employer pressure groups drafted a joint document in which they criticized the govemment's economic policy. The document stated that "in relative terms, the growth of the economy in the 1970s produced results unfavorable to the low and middle classes."3 In general, academic opinion concurs. In 1981, J. A. Bejarano argued that the deterioration in income distribution that began in 1970 continued until the end of the decade.4 Foreign economists agree. Albert Berry and Ronald Soligo, in their 1980 study, assert: "The tentative judgement must be that inequality has increased during the 1970s, and possibly during the late 1970s as well."5 This finding has profound policy implications, especially in view of the fact that the authors begin their book by stating: Ironically, inequality increased in Colombian society precisely when those responsible for economic policy began to take a more direct interest in the problem of income distribution. Both the Pastrana and Lopez administrations formulated and carried out at least some policies whose purpose was to improve the distribution of income. Yet most of the empirical evidence suggests that income distribution worsened in the 1970s.6 This statement would suggest that government efforts to improve distribution have not been effective. Methods and Conclusions of This Study Fortunately, the Colombian data do not support such a pessimistic opinion. First of all, distribution policies were given varying degrees of emphasis in the two administrations mentioned earlier. These policies INTROD UCTION 5 clearly were more important in the Lopez Michelsen administration than under Pastrana. Furthermore, redistribution policies in the fields of education, health, nutrition, foreign trade policy, financial policy, and fiscal policy cannot yield results in the short run. Therefore, assessing the impact of policies put into effect in 1974 on the basis of 1975 data-the most recent statistics available to Berry and Soligo when they published their study-may not be methodologically correct. Second, a complete analysis of all the existing statistical data shows that the income distribution did not worsen in the 1970s and that the real incomes of the poor improved significantly, especially in the latter half of the decade. The reason the results of this study differ so much from the opinion of many observers is that the most commonly available statistics suggest a process of income concentration. National income data on salaries and statistics on real industrial wages show little improvement in an economy with rapidly growing income per capita. This seems to suggest a worsening of the relative income position of labor. A simple comparison of income distribution derived from labor force surveys suggests the same thing. Gustav Ranis, for example, in a 1980 study analyzes income survey data and, on the basis of the distributions derived from these surveys for various years, concludes that there has been some income concentration through- out the Colombian development process. Although he warns about the quality of the data, he uses the survey information without adjustments to arrive at conclusions on income distribution trends. He believes that "some deterioration can be seen prior to the early 1960s, followed by a slight improvement in the late 1960s and early 1970s, with a new worsening trend thereafter."' But the income data for Colombia used by Ranis and others are incomplete and of highly variable quality. This means the estimates of income distribution are not comparable and, therefore, cannot be used to estimate changes in the indexes of concentration over time. Instead, a detailed analysis must be made of the coverage and the quality of the surveys used to obtain the primary income data, and a decision must then be made on which information is comparable. Chapter 4 of this study follows such a methodology. I first analyze the quality and the coverage of the various household surveys and then compare the income distribution derived from similar surveys. These comparisons do not support the hypothesis of a worsening income distribution. To complement the analysis based on income surveys and censuses, it is advisable also to analyze the information on eamings obtained from periodic surveys of salaries and wages. This latter analysis should not, 6 WINNERS AND LOSERS IN COLOMBIA'S GROWTH however, be limited to wages in the manufacturing sector. This sector clearly generates the most abundant and best-known statistics, but it is highly probable that trends in manufacturing wages do not coincide with those in other sectors. Accordingly, in Chapter 2, the present study uses various wage series as a complement to analyses of changes in distribution based on data obtained from household income surveys. These data also do not reflect a worsening of the distribution. The real wages of the very poor-the landless agricultural laborers-increased rapidly in the decade, as did the wages of various categories of unskilled urban workers. The sluggish trend of real wages in large manufacturing establishments and among white-collar workers is therefore not typical. Furthermore, all the data point to little growth in real income for the poor in the first part of the decade and to rapid progress in their standard of living in the second part of the decade, after economic policy started to be consciously designed with distributional goals in mind. Income surveys and wage series show average conditions and changes for different categories of the population, but they do not show what has happened to the real incomes and to the economic welfare of families and individuals over time. For that reason, an attempt is made in Chapter 3 to follow the fortunes of a group of poor families in Cali through the entire decade. The real incomes for this sample of poor families increased by about 100 percent in the decade. Obtaining information on the wealthiest group of the income distribu- tion is difficult. On the basis of the survey conducted in Cali, however, and a comparison of the rates of increase in the incomes of the poor and the middle class relative to the national per capita income, it can be concluded that the richest 5 percent of the families certainly did not suffer any loss in their share of the national income during the last two decades. But it is also possible that this share did not increase as much as is commonly believed. In Chapter 5 an effort is made to measure the proportion of families in absolute poverty at various times during the decade. This exercise again shows a deteriorating situation in the first part of the decade and clear improvements after 1976. It thus appears that different sets of data provide a fairly consistent picture of short-term changes in income distribution, and none of the data sets used confirm a deterioration in income distribution during the 1970s. In Chapter 6 1 suggest some hypotheses that may explain why income distribution did not become less equal in the decade, and discuss some INTRODUCTION 7 policy implications of the Colombian growth experience of the 1970s. This last chapter is tentative, because the main objective of this stage of my research was to find out what had really happened with the income distribution in Colombia between 1965 and 1980. In summary, the main finding of this study is that in the decade and a half after 1964, income distribution in the country did not deteriorate, and the real income of the poorest families increased significantly. Many people have criticized the economic policies followed in the 1970s, asserting that they led to a worsening of the income distribution and did not benefit the poor. The data collected here suggest the contrary: that if economic policy had an impact on distribution, it was to favor the poor. It is possible, however, that factors other than official economic policy, such as high coffee prices and the rapid growth of coffee production, helped to avoid a deterioration in income distribution. Nevertheless, a case can be made for the hypothesis that official policies helped to channel some of the incomes from the export boom of the late 1970s toward low-income families. There is no question, however, that exogenous factors, such as decreasing rates of population growth and the adoption of new technology in agriculture, affected distribution significantly. This study does not try to determine how each of these phenomena and each policy affected income distribution. From a historical perspective, the finding that a resource-rich country that followed fairly orthodox economic policies in the 1970s was able to achieve rapid growth, with no deterioration in income distribution and without increasing its foreign debt, may be both unexpected and significant. Notes to Chapter 1 1. Consmgna, September 15, 1980. 2. El Tiempo, May 10, 1980, p. 15A. 3. ANDI (Asociacion Nacional de Industriales), ANIF (Asociacion Nacional de Instituciones Financieras), CAMACOL (Camara Colombiana de la Construccion), FEDEMETAL (Federacion Metalurgica Nacional), and FENALCO (Federacion Nacional de Comerciantes), "Documento Dirigido al Presidente de ta Republica," April 1981. 4. J. A. Bejarano, "Crecimiento, Distribucion y Politica Economica" (Paper presented to the Congreso de Economistas de la Universidad Nacional, Melgar, Mav 1980; processed). 5. Albert Berry and Ronald Soligo, Economic Policy and Income Distribution in Colombia (Boulder: Wesrview Press. 1980), p. 17. 6. Ibid., p. 1. 7. Gustav Ranis, "Distribucion del Ingreso y Crecimiento en Colombia," Desarrollo y Sociedad, no. 3 January 1980). 2 Changes in Incomes of Different Occupations ONE USEFUL WAY TO EXAMINE CHANGES in income distribution over time is to compare the trends in wages for workers in different occupations. This type of analysis gives some idea of the changing income of populations with different skill levels and demographic characteristics. Also, because such analysis identifies variations in income differentials among large groups of the population, it may enable one to recognize broad transformations in social structure. Furthermore, the data on incomes by occupation are often of better quality than the data from household surveys used to calculate general measures of inequality; this is so because wage data are often based on the accounting records of enterprises, whereas survey data are based on the general impression of a respondent (usually a housewife) on the income of various household members. Of course, wage data are not useful for analyzing trends in capital income, which in general applies to the wealthiest group in society. But if one is interested in seeing what has happened to the poor, wage series are a good starting point. This is particularly true if it can be shown that differentials in labor income between wage earners and independent workers of the same age and sex and with similar education are not large; and this seemed to be the case in Colombia in the 1970s. The Poor in the Rural Sector In 1964, Colombia's poorest group of workers were the landless rural laborers. Daily farm wages were much lower than the wages earned by any urban worker. Wages of unskilled workers in construction were 10 percent higher than agricultural wages in 1964, while the wages of blue-collar industrial workers were three times higher (see Table 1). Moreover, rural 9 Table 1. Trends in Average Daily Wages for Various Occupational Categories (constant 1954 Colonmbian pesos) Unskilled construction workers Unskilled and semiskilled Manufacturing industry workers Agricultural construction workers Year workers a Bogotab Four cities C in four cities d Blue-collare White-collare 1960 3.35 4.08 - - 8.55 25.82 1961 3.53 4.08 - - - 26.09 1962 3.82 4.28 - - 11.41 27.41 1963 3.86 4.08 - - 12.23 27.09 o 1964 3.89 4.28 - - 11.82 26.64 1965 3.87 4.64 - - 12.23 27.55 1966 3.88 4.00 - - 11.91 26.95 1967 3.82 4.40 - - 12.18 27.73 1968 3.76 4.68 - - 12.50 28.64 1969 4.03 - - - 12.82 30.18 1970 3.90 4.64 - - 13.82 34.09 1971 3.72 4.28 - 5.99 13.55 33.64 1972 - 4.70 4.61 5.64 1.314 32.27 1973 - 4.40 4.24 5.23 12.05 30.27 1974 - 4.50 4.33 5.26 11.45 28.73 1975 - 4.70 4.50 4.84 11.23 28.09 1976 4.75 4.60 4.39 5.15 11.50 28.00 1977 5.34 4.40 4.25 4.88 10.86 26.23 1978 5.83 5.00 4.81 5.68 12.05 27.41 1979 5.79 5.50 5.34 6.43 12.86 27.86 -Data not available. Note: Deflators: In all cases, cost-of-living indexes of blue-collar workers will normally be tised to deflate wage rates. For national wage series, the national index was used, while for city wages, the index for that city was used. Daily wages are compared to avoid overestimations or underestimations in the case of agricultural and construction workers, who are not paid monthly and who normally do not work the entire month. a. Cold and hot regions were weighted equally. For 1960-71, the figure is the daily wage of males, without food. For 1976-79, it is the daily agricultural wage without food. b. For 1960-71, the table shows the series of wages paid to helpers in a small construction firm. For 1972-79, the figure is the average wage for helpers, based on the index published by DANE for Bogota. c. Weighted average of the wage paid to helpers in Bogota, Medellin, Cali, and Barranquilla. Assumes the following weights: 55 percent for Bogota, 20 percent for Medellin, 19 percent for Cali, and 6 percent for Barranquilla. d. Weighted average of wages paid to joumeymen and helpers in Bogota, Medellin, Cali, and Barranquilla. A weight of 55 percent was assigned to the wage of helpers and 45 percent to that of joumeymen. e. Assumes twenty.two days of work a month. Source: Departamento Administrativo Nacional de Estadistica (DANE). 12 WINNERS AND LOSERS IN COLOMBIA'S GROWTH laborers had benefited little from national economic growth during the three previous decades. The available statistics suggest that between 1935 and 1964 the purchasing power of agricultural workers earning daily wages did not increase. In 1964, the majority of the poor were in the rural sector. For that year, the average income of an employed person in the seventh decile of the rural population was less than the average income of an employed person in the second decile of the urban population. Put in another way, 71 percent of the economically active people with incomes of less than 3,400 pesos per year in 1964 lived in the rural area, and this included 58 percent of the rural labor force.! Clearly, not all of this large number of rural poor were workers on daily wages. Owners of small farms who derived an important part of their income from work outside their farms also had very low incomes. Rural poverty, of course, is not confined to Colombia. In most of the developing countries, poverty, underemployment, and even malnutrition are concentrated in the rural sector. In fact, the development process consists precisely in the creation of remunerative employment outside agriculture and in the improvement of the productivity of agricultural workers. Only when the rural population ceases to grow in absolute terms-because of a decline in the rural birth rate and the absorption of the increase in the rural labor force by sectors other than agriculture-can the real incomes of rural workers be expected to begin rising. W. Arthur Lewis's model of a dual economy implies stagnation of agricultural daily wages in real terms as long as there is redundant labor in the countryside./ According to this model, agricultural wages can increase in real terms only when the excess supply of rural labor has been absorbed; this probably occurs when the rural population ceases to grow or begins to decline. Historical experience seems to confirm this type of model. Ohkawa and Ranis, for example, show that in Japan agricultural wages did not begin to rise significantly until after 1917, when the agricultural labor force began to fall in absolute terms.3 Furthermore, a number of researchers have concluded that income distribution in developing countries begins to improve only when there is no more excess labor.4 For this reason, income distribution in the Republic of Korea and in Taiwan, where nonagricultural employment has expanded rapidly, is better than the average for economies with a similar level of per capita income. Thus, it is important to analyze the movement of agricultural wages in Colombia: an increase in those wages might reflect a decline in underemployment, and therefore the beginning of a development phase in which income distribution could improve. Data from Colombia's Departmento Administrativo Nacional de Estadistica (DANE) show a slight upward trend in real agricultural wages CHANGES IN INCOMES 13 during the 1950s and 1960s and a more substantial improvement in the late 1970s. Figure 1 shows the trend for agricultural wages for males between 1945 and 1971, in four departments (provinces) and in the nation as a whole. Antioquia and Caldas, both coffee-producing areas, were high-wage departments in the 1940s, and Antioquia had experienced some industrial- ization by that time. Boyaca and Narinio, by contrast, were both poor rural areas, but Boyaca, because it was close to Bogota, was influenced by the rapid urban growth of the capital city. On the semilogarithmic scale, the slope of the curve for Antioquia, the most industrialized area in 1948, is near zero until 1970; for Narinio and Boyaca, the two departments with the lowest wage levels at the start, the slope is positive, but, for Caldas, it is less so. The higher rates of growth for Boyaca and Narino suggest a small decrease in the amount of excess labor in these two highly populated and poor departments during this period. Table 2 displays the annual average rate of change in real agricultural wages for the nation and for the seven departments; it also compares these with rates of change in national per capita income. The table shows that between 1953 and 1969 agricultural wages increased at about the same rate as national per capita income, which means that rural workers roughly maintained their relative position in the income distribution.5 Since the rate of change of income per employed worker is higher than the per capita figure-because the population grew faster than the labor force-rural workers probably had a slightly lower share of total income in 1969 than in 1953. After 1970, agricultural laborers probably improved their relative position, particularly when one considers that the labor force grew more rapidly than the total population. This means that the agricultural wage with no food provided, shown in Table 1, grew roughly as fast as national income per worker. The agricultural wage with food provided, which appears on Table 2, grew much faster than national income per worker. Unfortunately, DANE discontinued its statistics on agricultural wages in 1969 (those for 1970 and 1971 are estimates). It did not resume its survey until 1976, and the new series uses a slightly different methodology. As Figure 1 shows, the 1976 wage level is significantly higher than that of 1969. The question, therefore, arises whether the higher level was due merely to the change in the sample, or whether it reflects a real improvement in the standard of living of rural workers. Analysis of the methodology used for the two series leads to the conclusion that they are not very different. The series ending in 1969 showed the "maximum," "minimum," and "most frequent" wage paid in each municipality, based on information provided by a well-informed person in that municipality. The departmental wage was the most commonly observed of the most frequent wages for the municipalities in Figure 1. Trends in Real Agricultural Wages for Males in Colombia, 1948-79 (semnilogarithmic scale) 10 9 9 3___ ____ __;_____\-* ____ ______ ______ ______ ____ ____________............; ....Nac__________ - Caldas Nacional 4 - '. - *- ~ - Caldas Anoquiaco ::,' *, ~~~~~~~~~~~~~~~~ * ~~~~~~~~~~Nacional Narrfio "'* rN ._ - "Antioquia 1' 2 . Naroiio ... ------- ...........-*- 0.9 0.8 i I t I 1 t I 1 I t -1 I I I I 1945 1950 1955 1960 1965 1970 1971 1976 1978 1979 Year Note: These figures are for wages without food in foLir cold-climiiate departments arid in the nation as a whole. The semilogarithmic scale shows rates of growth. Source: Departamento Administrativo Nacional de Estadistica (DANE). Table 2. Annual Average Rate of Change of Real Agricultural Wages (percent) Department 1953-691 1953-77b 1960.69a 1964-79b 1969-76c 1969-76b 1976-79d National 1.52 3.47 1.49 4.60 4.43 6.66 12.02 Antioquia -0.32 3.09 -0.75 4.63 6.63 8.80 14.03 Boyaca 2.88 4.03 1.84 3.98 6.06 5.89 5.47 Caldas -0.18 2.86 - 4.05 4.75 7.92 15.70 Magdalena 4.41 3.75 7.74 4.08 -1.14 2.70 12.29 Narinio 1.76 3.72 2.26 5.27 8.15 6.94 4.18 Santander 1.23 2.40 0.36 2.31 3.86 4.31 5.37 Valle -0.31 2.39 0.28 3.46 4.52 6.86 12.50 Rate of change of per capita income 1.59 2.50 2.72 3.67 4.10 4.12 4.19 -Data not available. Note: The rate of change is calculated by solving for I in the formula VF = VP (1 + iO), where i = rate of growth per year of the series, VF = wage in the last year, VP wage in the first year, and n = number of years. The figures for the series through 1969 refer to the wages of males. All figures are for cold climate with food supplied. The table includes different periods to allow the reader to see rates of growth for the whole period and also for each of the wage series, since it may be that they are not comparable. a. DANE wage series, old methodology. b. Wage series constructed from the two different DANE sets. c. Change between the two DANE wage series. d. DANE wage series, new methodology. 16 WINNERS AND LOSERS IN COLOMBIA'S GROWTH that department. The series beginning in 1976 is based on information from the Caja Agraria (an agricultural bank) on the most frequent wages in each municipality, but the departmental wage is the average of the most frequent wages paid in the municipalities.6 The second series includes wages in both farming and stockraising, but eliminates the distinction between wages of males and those of females. Because the dispersion of the "most frequent wages" by municipality tends to be slight, the average wage and the most frequent wage by province are probably similar. In theory, because of greater dispersion, the difference between the two wages could be greater for the national figures, but Figure 1 does not show an unreasonable growth of national wages compared with wages in the selected provinces. The methodological changes in the two wage series would, therefore, appear not to be so fundamental as to prevent their comparison.7 Nonetheless, because the increase from the end of one series to the beginning of the other is significant, an effort was made to obtain independent wage data for 1969-76 to verify whether the increase was real.' Data collected from the books of a few farms confirm the trends shown in Table 2. Real agricultural wages on farms in Boyaca and Cundinamarca increased substantially between 1969 and 1976, while the rate of growth was low or negative for farms in Valle. Table 2 also shows that rural wages grew faster in Boyaca than in Valle. Nevertheless, the direct farm data are too scanty to confirm or contradict the hypothesis that real rural wages grew during the period DANE discontinued its survey. More wage series from individual farms would have to be reconstructed to fill the gap in DANE wage information during 1969-76. What is clear is that since 1976 the incomes of rural workers have risen significantly; this is shown in both the DANE series and the data from individual farms. Between 1976 and 1979, agricultural wages in the country rose three times faster than national per capita income: the national farm wage index rose by 12 percent, while the national per capita income index rose by only 4 percent. Thus, in the second half of the 1970s, the income differential between rural and urban workers declined. In other words, the poorest group of workers in the society increased its standard of living more than the average. However, the increase in farm wages in the two poorest departments-Boyaca and Narifio-was much lower than the national index. Figure 1 reflects several other phenomena that deserve mention. First, the large increases in farm wages in Caldas and Antioquia, beginning in 1976, may bear some relation to increases in coffee production. Second, the data from farms suggest that during the 1970s wages increased less in CHANGES IN INCOMES 17 departments where they were already high, such as in Valle. Third, the rural minimum wage has increased rapidly since 1973, and it may be that this has contributed to raising the level of real wages. The farm data show that wages paid tend to increase when the minimum wage is raised. In addition to wages, the number of days worked also affects the incomes of rural laborers. Because of Colombia's topography, regions, sometimes fairly close to each other, experience peaks in labor demand at different times of the year, and there is evidence of much short-term geographical mobility of rural labor as a result. The improvements in communications have probably led to an increase in the number of days worked per year. At present, however, there is no empirical evidence on changes in the number of days worked. According to economic theory, changes in the minimum wage can increase the income of workers when monopsony conditions are present. These conditions are possible in the rural sector, where a few landowners can set the level of wages for their workers. In this case, an increase in the real minimum wage-as long as it is not so high as to cause unemploy- ment-could reduce the profits of the landowner and increase the earnings of the worker. Thus, one cannot reject the hypothesis that increases in the real minimum wage could have contributed to the rapid increases in real wages. In fact, the data from individual farms show that in general the minimum wage legislation has been effective. These data show that in all cases during 1975-79, the wages of unskilled workers, who often earn the minimum wage, increased more than those of overseers and skilled workers. This phenomenon, however, is consistent with a general decrease in wage differentials between skilled and unskilled workers observed at all levels of the economy. It is, therefore, likely that factors other than the minimum wage legislation were affecting the income of unskilled rural workers. The rural minimum wage in real terms increased 35 percent between November 1974 and January 1980. If there had not been some labor scarcity, such an increase in minimum wage would have led employers not to comply with the minimum wage legislation; there is no evidence that this occurred. Given the weak bargaining power of individual rural workers, what probably happened was that the minimum wage legislation acceler- ated wage adjustments that would have taken place more gradually. The government, pressured by the unions, proposed large increases in rural minimum wages, which the modem sector farmers did not oppose at the bargaining table because they had to pay wages higher than the minimum wage to keep their laborers. This raised the wages of laborers working in smaller and more traditional farms. At the end of the decade, when 18 WINNERS AND LOSERS IN COLOMBIA'S GROWTH minimum wages started approaching market wages on modem farms, pressure by employers against continued increases in rural minimum wages by the Salary Council intensified; thereafter, increases in real rural wages became much less rapid. Minimum wage management, therefore, may have had some influence on the income of some rural workers. The increase in real agricultural wages during the 1970s may mark the beginning of an improvement in the standard of living of the poorest groups in Colombian society. Not only did the wages of farm workers increase, but most likely so did the incomes of small landowners because many of them are also daily workers. In fact, stagnation in the incomes of small rural landowners is incompatible with significant increases in agricultural wages; otherwise the smallholder would become a daily worker. Because these two groups make up a significant proportion of the poor in Colombia, their economic improvement over the last decade has major implications for the national welfare. Occupations of the Urban Poor Myths abound concerning who are the poor in the cities of Colombia. Some believe that recent arrivals from rural areas constitute the largest group of urban poor. Many studies, however, have demonstrated that immigrants are not poorer than persons born in the city, nor do they face a greater likelihood of being unemployed.9 Others believe that many of the poor are in the informal sector of the economy, defined as the group consisting of self-employed persons, employees of small enterprises, and workers in trade and services. Mohan's 1980 study, however, shows that only construction workers and persons in domestic service are paid significantly less than workers in the manufacturing industry (see Table 3). Unquestionably, construction workers are the most poorly paid urban group next only to those in domestic service. Unfortunately, official statistics of their income were not published before 1972, and, although Urrutia and Berry constructed a daily wage series based on official statistics, the data for the 1960s do not seem to be sufficiently reliable.'° Accordingly, Fedesarrollo, a research institute, obtained wage series directly from two construction firms in Bogota (one small firm and one large organized firm) to compare the data with official figures for the 1970s. Table 4 shows the series of daily wages at constant 1954 prices for the two construction firms for 1959-79. Table 5 presents the DANE series for the total construction labor force and by skill categories for the 1970s. CHANGES IN INCOMES 19 Between 1960 and 1970, real wages for construction workers increased, according to both the series in Tables 1 and 4, but the increase was not constant. The real wage reached its highest level in the middle of the decade. In the 1970s, the daily wage (in Bogota, Medellin, and Cali) increased only at the end of the decade and concurrently with a decline in urban unemployment. In Barranquilla, however, the average wage in construction still has not regained its 1972 level, although the series shows an upward trend beginning in 1978. Since construction wages in Barranquilla were higher than those in other cities in 1972, labor mobility may explain why the wage decrease was so marked between that year and 1975. In general, the real wages of helpers and journeymen increased more than those of masters. During both the 1960s and the 1970s, the daily wage of construction workers rose less than national per capita income, implying that workers in the construction sector lost out relative to other sectors. It is not certain, however, what happened to the annual income of construction workers in the 1970s. If the reduction in unemployment meant more days worked in a year, the rise in annual income would be greater than the increase in the daily wage. If the average period of employment did not increase, however, this means the real incomes of construction workers rose at an annual rate of only 2 percent during the decade, while national per capita income increased by about 3.5 percent. The Fedesarrollo data for the two firms in Bogota shows that the large firm had greater annual increases in labor cost than the smaller firm. When the two firms are weighted equally, however, the behavior of the average wage in Table 4 is similar to the DANE statistics in Table 5. Within the construction industry during the 1970s, the salaries of helpers rose at a faster rate than those of masters and journeymen. This was true both for the large firm and for the weighted average of the two firms, as well as for the DANE figures for all cities except Barranquilla. Over the two decades, the income differential between helpers and masters in the small firm dropped from 3.90 to 2.91, while that between masters and journeymen together and helpers decreased from 1.94 to 1.90. It can thus be concluded that the incomes of unskilled and semiskilled workers in construction rose more than the average for that sector. This phenomenon is found in most urban statistics. Moreover, it is consistent with the results of the analysis of poor and wealthy families in Cali presented in Chapter 3. That analysis shows that the incomes of secondary workers in poor families-who, like helpers in the construction industry, are generally young-have risen more rapidly than the incomes of primary workers. 20 WINNERS AND LOSERS IN COLOMBIA'S GROWTH Table 3. Monthly Earnings and Distribution of Workers by Occupation in Bogota, 1973 and 1977 (earnings in current Colombian pesos) 1973 Males (M) Females (F) Earnings All ratio Occupation (lLO code) Mean Y Percent Mean Y Percent (percent) F/M Professional and technical 6,090 8.3 2,740 8.0 8.2 0.45 (1-19) (1.18) (1.68) Administrative and manager 9,412 2.1 3,692 0.5 1.6 0.39 (20-29) (0.98) (1.03) Clerk and typist 2,136 10.3 1,760 17.4 12.7 0.82 (30-39) (1.42) (1.24) Sales managers and proprietors 3,338 7.8 1,736 3.8 6.5 0.52 (40-41) (1.92) (2.55) Other sales 2,059 8.6 743 7.9 8.4 0.36 (42-49) (2.00) (0.98) Service workers (excluding maids) 1,416 7.1 807 14.0 9.4 0.57 (50-53; 55-59) (2.5) (2.37) Maids 537 0.5 364 30.4 10.6 0.68 (54) (0.90) (1.64) Agriculture workers 2,698 1.9 3,032 0.2 1.3 1.13 (60-69) (3.02) (2.24) Production supervisors 1,325 5.1 851 3.5 4.6 0.64 (70) (1.29) (0.68) Production workers 1,279 27.2 810 14.1 22.7 0.63 (71-94; 96, 97) (1.15) (1.71) Construction workers 968 10.8 487 0.1 7.2 0.50 (95) (1.29) (0.83) Transport workers 1,389 8.3 1,408 0.0 5.5 1.01 (98) (0.77) (0.82) Other 701 2.1 597 0.0 1.4 0.85 (0.73) (0.70) Total 2,166 100.0 1,043 100.0 100.0 0.48 (1.95) (2.13) Number of workers (thousands) 436 223 659 Number of workers responding in sample 37,455 17,456 53,911 No information (percent) 12.0 12.42 -Data not available. n.s. Not significant Note: Coefficients of variation are in parentheses. Because of rounding, the percentage figures may not correspond to the totals. Mean Y is mean income. CHANGES IN INCOMES 21 1977 Males (M) Females (F) Earnings All ratio Mean Y Percent Mean Y Percent (percent) F/M 11,921 12.0 6,197 9.0 10.9 0.52 (0.87) (0.87) 13,993 4.6 5,309 0.7 3.1 0.38 (0.69) (0.47) 4,390 11.7 3,600 18.8 14.4 0.82 (0.87) (0.55) 7,734 7.7 3,470 5.2 6.8 0.45 (1.22) (1.68) 4,794 8.9 1,911 9.0 8.9 0.40 (1.10) (1.09) 3,055 7.9 1,971 17.1 11.4 0.64 (0.63) (0.97) 1,731 0.1 2,279 24.4 9.3 1.32 (0.30) (0.32) 8,033 1.7 3,038 0.5 1.2 0.38 (1.20) (1.34) 6,314 1.0 4,795 0.3 0.7 0.76 (0.92) (0.50) 3,151 28.2 1,896 15.0 23.2 0.60 (1.37) (0.60) 2,226 8.2 1,800 n.s. 5.1 0.81 (0.62) (0.0) 3,558 7.5 - - 4.7 - (0.78) 2,345 n.s. n.s. n.s. 0.4 - (0.41) 5,402 100.0 2,828 100.0 100.0 0.52 (1.27) 733 442 1,175 3,371 2,039 5,410 Source: Rakesh Mohan, The People of Bogota Who They Are, What They Earn, Where They Live, World Bank Staff Working Paper no. 390 (Washington, D.C., 1980). Data for 1973 are based on the population census sample and for 1977 on the EH-15 household survey. Table 4. Average Real Daily Wages in Two Construction Firms in Bogota, 1959-79 (constant 1954 Colombian pesos) Small firm Large firn Weighted- Year Helper Joumeyman Master Totalb Helper Joumeyman Master Totalb Helper Journeyman Master Totalb 1959 3.39 6.80 13.20 5.84 - - - - - - - - 1960 4.06 6.53 12.67 5.90 - - - - - - - - 1961 4.09 6.14 12.87 5.80 - - - - - - - - 1962 4.28 6.16 14.27 6.10 - - - - - - - - 1963 4.07 5.97 14.06 5.90 - - - - - - - - 1964 4.26 6.25 14.70 6.20 - - - - - - - - 1965 4.62 8.00 14.20 7.00 - - - - - - - - 1966 4.00 6.89 14.81 6.30 - - - - - - - - 1967 4.38 6.58 13.70 6.30 - - - - - - - - 1968 4.66 7.77 12.94 6.80 - - - - - - - - 1969 C -c -c _ 1970 4.62 6.60 14.30 6.40 1971 4.14 6.15 14.12 6.00 4.08 6.64 19.16 6.70 4.11 6.39 16.64 6.35 1972 5.21 7.15 17.65 7.30 4.20 6.52 18.52 6.60 4.70 6.83 18.08 6.90 1973 4.31 6.18 20.13 6.70 4.12 5.76 17.24 6.10 4.21 5.97 18.68 6.40 1974 4.91 6.36 17.34 6.80 4.44 6.12 18.04 6.50 4.67 6.24 17.69 6.60 1975 _c _C _C _C 4.48 6.08 17.48 6.50 - - - - 1976 3.99 5.40 12.53 5.40 4.64 6.36 17.60 6.70 4.31 5.88 15.06 6.00 1977 4.82 6.33 12.65 6.20 4.64 6.28 17.40 6.60 4.73 6.30 15.02 6.40 1978 5.13 7.96 15.40 7.40 5.40 7.44 20.04 7.70 5.26 7.70 17.72 7.60 1979d 5.30 8.18 15.47 7.60 5.84 7.40 20.92 8.00 5.57 7.79 18.19 7.70 -Data not available. Note: Wages are deflated by the blue-collar consumer price index for Bogota: base 1954- 55 100. a. Assuming equal participation for each of the two firms in the total. b. The DANE weighting system was used (47 percent helper, 43 percent joumeyman, and 10 percent master). c. The small firm did no construction in that year. d. For the small firm, the figure is the average for the first quarter; for the large firm, it is the average for January. Source: Construction company records. Table 5. Average Real Daily Wages in the Construction Industry in Four Colombian Cities, by Category and Total, 1972-79 (constant 1954 Colombian pesos) Annual rate of change 1972-79 City and category 1972 1973 1974 1975 1976 1977 1978 1979 (percent) Bogota Total 6.9 6.5 6.5 6.7 6.5 6.0 6.9 7.8 1.8 Master 18.6 18.2 18.7 18.1 16.4 15.2 17.6 18.3 -0.2 Joumeyman 6.8 6.5 6.5 6.6 6.3 5.6 6.7 8.0 Z.3 Helper 4.7 4.4 4.5 4.7 4.6 4.4 5.0 5.5 2.3 Medellin Total 6.2 5.7 5.6 5.7 5.6 5.5 6.5 7.1 1.9 Master 13.0 12.5 12.0 11.5 11.5 11.0 13.6 13.6 3.5 Joumeyman 6.5 6.1 5.7 5.6 5.5 5.5 6.7 7.3 1.4 Helper 4.6 4.2 4.3 4.5 4.4 4.4 5.0 5.5 2.6 Cali Total 6.5 6.1 6.0 6.1 5.7 5.6 6.4 7.1 1.3 Master 22.0 19.7 18.3 17.1 14.6 13.9 15.7 15.9 -4.5 Joumeyman 7.0 6.4 6.3 6.3 5.8 5.8 6.7 7.6 1.2 Helper 4.0 3.7 3.7 4.0 3.7 3.7 4.2 4.8 2.6 Barranquilla Total 7.9 7.2 7.1 6.0 6.3 5.9 6.8 7.1 Master 13.4 12.3 13.1 12.3 12.6 12.1 17.7 14.3 0.9 Journeyman 8.7 7.1 8.1 6.4 6.6 6.1 7.3 7.6 -1.9 Helper 5.7 4.6 4.9 4.2 4.6 4.1 4.8 5.0 -1.8 Note: "Total" uses the DANE-weighted average for master, joumeyman, and helper. The amount in pesos was calculated on the basis of the labor cost index of DANE, using average wages by category of personnel in each municipality for December 1979. In the total, weights of 10 percent, 43 percent, and 47 percent were assigned, respectively, to the three categories. These figures were then deflated by the blue-collar consumer price index for each city. Source: DANE. CHANGES IN INCOMES 25 The DANE figures in Table 1 indicate that since the second half of the 1970s the agricultural wage has exceeded the wage of construction helpers in the four cities studied. In Cali, the agricultural wage was higher than the average wage for all construction workers in 1977 and 1978.11 Finally, data from the Social Security Institute (iss) on insured persons working in construction indicates a reduction in real wages during 1972-76 and a rising trend in 1976-79. However, the absolute level of earnings in these data is much higher than the level reported by DANE and the sample of firms, because the persons covered are permanent employees of the construction companies and include white-collar workers, architects, and some older and well-trained workers. As a rule, laborers are hired by subcontractors and are not covered by social security. In the case of services and trade, there are no official statistical series of wages covering at least a decade. Accordingly, a series had to be prepared on the basis of the records of selected firms. Table 6 presents information on wages for three retail trade stores-a supermarket, a parts store, and an appliance store. The data for the supermarket show gains in the real monthly wages of unskilled personnel, but the growth rate is lower than that of national per capita income. Between 1962 and 1978, the real incomes of cashiers rose at an average annual rate of 1.2 percent, those of packers did not rise, those of markers rose by 1.1 percent, and those of cutters rose by less than 1 percent a year. Thus, all of the groups lost out in relative terms. Furthermore, there were significant increases during the 1960s and a real decline in the 1970s. Wages in the other stores in the table behaved similarly. Increases in real wages were small or nil during the 1970s, and in all cases were less than the rise in national per capita income. In retail trade, as in agriculture, construction, and services, wages increased more in the second part of the 1970s than in the first; in fact, from 1970 to 1975, real wages declined. Also in trade, as in construction and manufacturing, wage differentials by level of training narrowed over the decade. DANE also has data on earnings in retail trade beginning in 1975. According to its sample, such earnings rose by 17 percent in real terms between 1975 and 1979, representing an average annual increase of 3.6 percent (see Table 7). In the service sector, five firms-a restaurant, a laundry, two hotels, and a liquids-handling plant-were surveyed (see Table 8). The series for four of the firms show stagnation or very loW growth in wages. They also indicate deterioration in the first part of the decade and an improvement beginning in 1974-75. The exception is the hotel in Girardot, which is more of a social club. 26 WINNERS AND LOSERS IN COLOMBIA'S GROWTH Table 6. Average Real Monthly Wages Paid by Three Retail Trade Stores, 1962-79 (constant 1954 Colombian pesos) Supermarket a Parts store b Year Cashier Packer Marker Cutter Salesman Repairman Secretary 1962 211 208 242 259 - - - 1963 249 220 249 246 - - - 1964 257 203 276 257 - - - 1965 267 214 288 266 - - - 1966 251 192 269 275 - - - 1967 255 192 285 304 - - - 1968 265 207 303 302 - - - 1969 264 212 306 289 - - - 1970 273 218 315 287 419 168 279 1971 281 204 328 279 383 191 298 1972 275 182 321 273 535 191 298 1973 252 151 289 250 238 1II 222 1974 247 197 240 239 250 115 209 1975 240 191 279 260 266 131 208 1976 245 196 277 280 409 133 212 1977 232 185 262 260 387 168 215 1978 255 204 289 286 372 163 207 1979 - - - - 327 190 233 -Data not available. Note: Wages are deflated by the blue-collar and white-collar consumer price index for Bogota: base 1954-55 = 100. a. The figures correspond to the basic wage of a person holding the position in a large supermarket chain in Bogota. b. The data correspond to the basic wage for the position in a store in Boyaca with fewer than ten employees. CHANGES IN INCOMES 27 Appliance store' Watchrnan Cleaner Messenger Technician Cashier Secretary Managerd Average' 154 181 132 374 219 264 753 275 170 181 131 343 235 272 928 294 158 176 141 318 247 265 946 291 153 153 144 345 216 244 819 269 144 156 162 347 208 266 901 284 142 138 170 331 216 239 784 269 157 155 158 376 227 223 831 277 150 149 149 386 257 221 895 266 150 164 155 321 236 241 930 279 188 193 169 388 252 285 878 300 c. The data correspond to the average basic wage for all identical positions in the Bogota branches of a large appliance store. Most of the positions have been held by the same persons, and for this reason the basic wage implicitly takes account of the seniority factor. d. In most years, the position has been held by a university graduate in most branches. e. Simple average of wages of workers with a fixed wage in all branches in Bogota. Source: Company records and Fedesarrollo computations. 28 WINNERS AND LOSERS IN COLOMBIA'S GROWTH Table 7. Average Real Monthly Salaries and Wages Paid to White-Collar Workers in Retail Trade, 1975-79 (constant 1954 Colombian pesos) Occupation 1975 1976 1977 1978 1979 Merchandise in general 248 260 255 282 306 Food and beverages 319 327 336 362 395 Clothing and footwear 265 294 279 293 310 Pharmaceutical products 247 236 241 279 300 Furniture and household appliances 437 456 483 500 470 Construction materials 341 421 409 391 386 Vehicles and spare parts 450 507 512 540 514 Fuels 297 359 388 396 397 Unclassified merchandise 377 363 349 348 352 Total 323 340 343 365 378 Note: Wages aTe deflated by the consumer price index: base 1954-55 100. Source: DANE monthly statistical bulletins and Fedesarrollo computations. Tables 6, 7, and 8 show, in general, that incomes in the services sector are much lower than those in trade. In both sectors, however, earnings are higher than the minimum wage. For the manufacturing industry, data from institutional sources such as DANE, the Social Security Institute, and the Business Opinion Survey of Fedesarrollo, as well as from individual firms, were used to obtain as consistent a picture as possible of the trends in earnings (wages plus fringe benefits). In general, both white-collar and blue-collar workers in the manufacturing sector enjoyed an increase in real wages during the 1960s. The DANE Monthly Survey of Manufacturing shows gains of about 20 percent for both groups (see Figure 2). The annual survey shows the same trend, although, because of different methodologies and definitions, the order of magnitude is not the same. During the 1970s, in contrast, white-collar and blue-collar workers suffered such a decrease in their real wages that by the middle of the decade they were earning less in real terms than they had ten years earlier. The decline is clear in the DANE monthly survey as well as in the annual survey and in the data of ISS and Fedesarrollo (see Table 9). Toward 1975-76, this trend began to be reversed and increases were observed in salaries and wages, although the level of the early 1960s was not regained until the end of the 1970s. Even in 1979, however, industrial wages had a lower purchasing power than at the beginning of the 1970s. CHANGES IN INCOMES 29 The DANE annual survey, which covers both wages and fringe benefits, shows that fringe benefits contributed significantly to improving the situation of wage earners during a period when basic salaries and wages were declining in real terms."2 While the salaries and wages of executives and technical personnel and of white-collar workers in the manufacturing industry rose by 16 percent between 1964 and 1970 and fell by 18 percent between 1970 and 1976, total income (wages plus fringe benefits) rose by 28 percent and fell by 9 percent in the respective periods. The situation is similar in the case of production workers and laborers: salaries and wages showed a gain of 12 percent between 1964 and 1970 and an equal reduction of 12 percent between 1970 and 1976, whereas total earnings rose by 25 percent in the first of those periods and fell by only 5 percent in the second. Consequently, the difference between total labor income and salaries and wages has been widening. Salaries and wages of executive and technical personnel and of white-collar workers accounted for 80 percent of total income in 1962, 71 percent in 1970, and 64 percent in 1976, whereas the percentages for blue-collar workers were 79 percent, 70 percent, and 64 percent, respectively, in the same years. In summary, all the data confirm the sluggish growth of wages in the manufacturing industry during the 1970s. But certain distinctions need to be made. First, if fringe benefits are added to the basic wage, total earnings perform better (see Table 10). Second, even in manufacturing, the poorer, unskilled workers did better than the skilled workers (see Table 11). Third, wages and total earnings have increased faster in small enterprises than in large ones (see Tables 12 and 13). All this has meant a decrease in income differentials in industry and relative gains on the part of the poorer workers. Finally, wage dispersion from industry to industry has decreased, particu- larly at the expense of the highly paid petroleum workers. The Urban Middle Class The middle class is difficult to define. A sociological definition would encompass a quite heterogeneous category. In surveys conducted in Japan, for example, the majority of the population declares that it belongs to the middle class. In Colombia, even persons in the highest decile of the income distribution believe that they belong to the middle class. A strict economic definition of the middle class would include persons who have incomes close to the national average. Thus, common usage includes in the middle class population groups different from those that an economist might include in that category. 30 WINNERS AND LOSERS IN COLOMBIA'S GROWTH Table 8. Average Real Monthly Wages Paid by Service Firms, 1965-78 (constant 1954 Colombian pesos) Laundry Hotel in Bogotad Restaurant Basic Wage and Year basic wage a wage b benefits c Laundry e Housekeepinge Averagef 1965 - 177 211 - - - 1966 - 148 175 - - - 1967 - 144 169 - - - 1968 136 146 178 - - - 1969 130 156 184 - - - 1970 142 142 167 - - - 1971 147 147 169 - - - 1972 131 131 151 140 138 138 1973 119 119 135 123 130 129 1974 119 119 137 115 129 127 1975 117 120 136 114 124 123 1976 122 124 140 126 139 137 1977 109 125 137 118 133 130 1978 126 140 154 133 149 145 -Data not available. Note: Wages are deflated by the blue-collar consumer price index for Bogota: base 1954-55 100. a. Average wage of a sample of fifteen persons in a total population of twenty-five (waiters, cashiers, cooks, kitchen helpers, and so forth). b. Average wage of the total population of twenty-four (drivers, pressers, launderers, spot removers, markers, counter personnel and so on). c. Includes vacations, family allowance, transportation allowance, bonuses, and severance pay. d. The data correspond to the basic wage plus night and holiday bonuses; the conversion from daily to monthly wage is based on twenty-five working days per month CHANGES IN INCOMES 31 Hotel in Girardotg Handling of liquids' Laundry Basic Real Cleaner worker Maid Waiter Cook Bellhop Porter Averageh wage income' 108 78 109 163 108 77 100 168 - - 92 65 91 137 91 65 83 138 212 340 105 74 99 139 96 74 109 143 200 338 104 84 102 152 102 84 111 154 189 354 113 107 108 262 113 105 118 167 197 326 123 156 124 268 124 150 160 184 202 325 127 190 156 265 144 185 192 212 242 345 137 180 183 270 183 179 212 218 219 344 162 193 194 257 204 193 203 218 189 410 153 180 202 242 210 180 188 203 190 363 145 157 184 286 185 156 178 197 179 314 162 198 173 252 180 198 201 202 181 338 143 153 153 223 157 153 164 175 186 280 153 132 137 213 137 153 141 165 188 265 e. These data are based on the average of a sample observed over eight two-week periods. f. Average wage in the laundry and housekeeping unit, weighted by the number of employees in each unit g. The data correspond to basic wages plus overtime for a single person who has performed the job. h. Simple average of a sample of twenty-four persons. i. Average for the population of a firm engaged in the treatment of liquids, with plants in Bogota, Cartagena, and Buenaventura. j. Basic wage plus overtime, bonuses, and allowances. Source: Records of firms surveyed. 32 WINNERS AND LOSERS IN COLOMBIA'S GROWTH Figure 2. Trends in Wages Paid in Manufacturing Industry and in National Per Capita Income, 1962-79 Percent 180 170 - I National income 160- ,*"* 1450 _ r4 120 . I Wages of blue-collar workers 0~~~~~~~~~~~~~~~ 1962 1964 1966 1968 1970 1972 1974 1976 1978 1979 Source: DANE, Monthly Survey of Manufacturing, Fedesarrollo computations. Ya Table 9. Indexes of Average Real Wages of Workers in the Manufacturing Industry, 1970-79 (Base 1970 = 100) Source 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 Annual survey Blue-collar 100 100 98 89 86 85 88 86 105 101 White-collar 100 103 98 89 87 81 82 79 89 87 Monthly sample Blue-collar 100 98 95 87 83 81 83 79 87 93 White-collar 100 99 94 89 84 82 82 77 80 82 Social Security Institutea Total 100 95 91 89 92 89 97 90 99 - Business Opinion Surveyb Blue-collar - - 100 92 87 90 85 84 94 93 White-collar - - 100 94 88 90 86 83 86 85 -Data not available. a. Social Security wage services underestimate wage increases for the reasons given in note 8 of this chapter. b. The Business Opinion Survey of Fedesarrollo was started in 1972, and the variations refer to the December/December averages (1972 = 100). Source: DANE annual survey and monthly survey; Social Security Institute, Infornmes Estadisticos, various numbers; Fedesarrollo, Encuesta de Opinon Empresarial. 34 WINNERS AND LOSERS IN COLOMBIA'S GROWTH Table 10. Percentage Change in Real Wages and Real Total Earmings of Blue-Collar and White-Collar Workers in the Manufacturing Industry, 1964-76 Employment categories 1964-70 1970-76 1964-76 All white-collar workers' salaries 28 -18 5 Technical and managerial (White-collar) employees Salaries 16 -18 -4 Total eamings' 28 -9 17 Blue-collar workers Wages 12 -12 -1 Total eamingsa 25 -5 19 a. Total eamings include wages plus fringe benefits. Source: DANE. In Colombia the middle class generally is considered to comprise white- collar workers and persons with a secondary education. This definition is quite consistent with economic reality. In fact, the 1977 DANE survey showed that the average income of a worker in Bogota was Col$4,434 a month, very close to the average income for persons with a secondary education." In other words, persons with a secondary education have incomes close to the average and can therefore be considered typically middle class. Because the distribution of income is not normal, however, the mean income is higher than the mode.14 This suggests that the middle class defined in terms of education, or average income, probably would include people from the seventh to the ninth deciles of the income distribution.5 In 1977, as shown in Table 3, the incomes of sales agents, production supervisors, and office workers approached the average income per worker. Thus, with a definition based on average income, those groups might be typical of the middle class. But the characteristics of middle-class workers can be defined in greater detail on the basis of tabulations of a 1978 household survey. Table 14, which displays the distribution of income by occupation for males in Bogota, shows an average income of about Col$10,000 per month. The average income in several occupations approximately equals the national average: teachers; accounting clerks, bank employees, and cashiers; business managers; and merchants. In other words, the middle class could include white-collar workers, merchants, and teachers."6 Some blue-collar occupa- tions, however, also account for substantial proportions of the labor force in the seventh to ninth deciles, which we have defined as the middle class. CHANGES IN INCOMES 35 Table 11. Percentage Change in Real Remuneration of Labor by Branch of Economic Activity, 1962-79 Branch of activity 1962-69 1970-79 Manufacturing industry blue-collar workers 12.0 -6.9 Industries in which unskilled workers are predominant' Food (20) 22.2 -0.4 Tobacco (22) -18.0 -23.8 Footwear and clothing (24) 1.1 9.9 Wood industry (25) 0.5 -6.0 Wooden fumiture (26) 0.4 -11.0 Leather and leather products (29) 20.5 -8.1 Metal products except machinery (35) 5.5 -1.2 Averagea 9.5 0.5 Industries in which skilled workers are predominanta Petroleum products (32) -0.8 -26.2 Beverages (21) 0.6 -8.0 Textiles (23) 23.1 -12.9 Paper and paper products (27) 33.2 6.2 Rubber products (30) 27.1 -0.6 Basic metals (34) 37.8 -25.8 Averagea 29.8 -11.0 Manufacturing industry white-collar workers 10.0 -18.3 Total constructionb 4.9 13.0 Master 0.2 -1.6 joumeyman 7.1 17.6 Helper 7.9 17.0 Financial sector Professional 18.1 -20.6 Nonprofessional (with 15 years of experience) 43.9 -7.0 Nonprofessional (with 10 years of experience) - 40.1 Government sector Professional (Ministry of Finance) -21.6 -32.9 Executive - -10.6 Primary school teacher - 2.7 Secondary school teacher - -5.4 Manager of electrification plant - -37.4 Head of personnel - -12.4 Lineman - 10.0 National government white-collar workers 1.0 - Departmental government white-collar workers 1.0 Municipal white-collar workers 7.0 -Data not available. a. Weighted average based on participation in employment. Based on DANE manufacturing survey, 1968. b. 1962-70: Data from small construction firm; 1972-79: DANE, Bogota. The periods are different from those covered in the case of manufacturing. Source: DANE, annual survey, and records of firms. Table 12. Average Real Monthly Salaries and Wages in the Manufacturing Industry by the Number of Persons Employed in the Enterprise, 1963-76 (constant 1954 Colombian pesos) Number employed 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 Fewer than 5 119 109 106 109 140 109 144 151 168 151 139 147 161 166 5-9 136 122 120 116 139 132 158 158 157 177 164 164 153 155 10-14 168 153 150 141 174 166 175 180 176 173 158 157 160 163 15-19 216 203 209 187 193 199 192 191 196 184 169 167 162 167 ce 20-24 230 215 218 209 211 207 207 197 193 203 176 177 173 174 25-49 247 232 246 232 234 234 222 230 227 216 197 191 189 194 50-74 283 277 276 257 264 263 267 265 262 245 224 219 214 214 75-99 291 273 293 284 305 296 274 297 293 292 253 252 232 251 100-199 340 336 342 327 333 338 334 352 339 316 294 278 265 281 200 and more 360 365 379 372 395 393 386 427 433 420 377 362 356 364 Note: The wage and salary figures are deflated by the national consumer price index: base 1954-55 = 100. Source: DANE, Annual Survey of Manufacturing, and Fedesarrollo computations. Table 13. Average Real Monthly Salaries, Wages, and Employee Benefits in the Manufacturing Industry by the Number of Persons Employed in the Enterprise, 1963-76 (constant 1954 Colombian pesos) Number employed 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 Fewer than 5 127 117 113 120 160 118 161 170 191 174 166 170 203 216 5-9 150 135 133 128 155 149 184 182 183 211 194 200 188 199 10-14 190 174 170 162 201 192 208 214 212 208 191 194 202 212 15-19 249 236 246 220 228 236 233 234 238 226 210 209 205 220 20-24 266 254 258 246 253 249 257 243 239 253 218 223 224 230 25-49 293 279 295 282 290 287 283 288 287 276 250 247 248 265 50-74 346 343 346 321 332 330 356 343 338 322 294 194 295 303 75-99 358 343 362 259 393 386 369 382 389 391 346 345 320 358 100-199 420 429 436 434 451 455 461 476 469 439 412 390 388 420 200 and more 490 493 518 515 561 562 582 647 629 633 579 566 572 594 Note: The figtLires used here are deflated by the national consumer price index: base 1954-55 100. Source: DANE, Annual Survey of Manufacturing, and Fedesarrollo computations. Table 14. Distribution of Income by Occupation for Males in Bogota, 1978 (income in 1978, Colombian pesos a month) Fourth Fifth Sixth Seventh Eighth Ninth Tenth Occupation decile decile decile decile decile decile decile Total Architects and engineers Row percentage 0 1.53 0 6.60 21.98 60.82 100 Column percentage 0 0 0.39 0 1.26 4.69 10.52 2.07 Average labor income - - 0 - 11,230 23,596 32,146 25,517 Average other income - 0 - 0 2,629 4,837 3,518 Medical doctors and dentists Row percentage 0 0 5.10 1.47 6.20 38.28 47.22 100 Column percentage 0 0 1.32 0.43 1.20 8.52 8.52 2.79 Average labor income - - 4,008 - 13,713 20,056 35,202 25,451 Average other income - - - - 0 96 2,320 1,133 Accountants Row percentage 12.60 0 0 5.09 18.05 11.47 52.79 100 Column percentage 4.02 0 0 1.41 3.50 2.44 9.09 2.66 Average labor income 9,052 _ _ 11,477 12,157 13,303 34,149 23,473 Average other income 2,565 - - 0 1,606 4,249 4,237 3,368 Jurists Row percentage 1.87 0 0 2.25 4.62 19.10 70.11 100 Column percentage 0.49 0 0 0.51 0.73 3.32 9.90 2.18 Average labor income 8,000 - - 14,300 8,811 16,550 39,453 31,767 Average other income 0 - - 0 0 0 2,277 1,596 Teachers Row percentage 6.02 6.43 8.93 7.21 14.36 23.41 26.40 100 Column percentage 2.67 2.60 3.06 2.78 3.37 6.90 8.32 3.69 Average labor income 5,139 5,501 6,991 9,569 8,103 13,655 23,364 12,717 Average other income 79 387 0 0 973 1,166 1,899 959 Managers and executives Row percentage 1.58 0 7.95 0.77 10.56 20.63 55.98 100 Column percentage 1.25 0 4.85 0.53 5.07 10.85 23.88 6.59 Average labor income 2,517 - 10,710 6,750 12,967 17,218 34,939 25,423 Average other income 2,676 - 0 0 2,444 1,468 4,005 2,845 Bookkeepers, bank employees, cashiers, and others Row percentage 6.40 7.61 11.34 13.23 16.70 21.97 17.74 100 Column percentage 5.67 6.13 7.76 10.21 8.99 12.95 8.48 7.38 Average labor income 6,333 4,818 6,771 6,768 7,080 8,156 11,918 7,701 Average other income 0 0 332 231 513 410 1,159 450 Mailmen and messengers Row percentage 14.65 21.36 10.26 18.80 16.69 4.39 1.12 100 Column percentage 4.67 6.22 2.53 5.22 3.23 0.93 0.19 2.66 Average labor income 5,392 2,523 3,247 3,029 3,120 3,193 2,915 3,399 Average other income 0 0 0 303 0 0 0 57 '.o Managers in the commercial sector Row percentage 10.93 6.70 11.41 11.14 20.15 32.91 3.48 100 Column percentage 1.54 0.86 1.24 1.37 1.73 3.09 0.26 1.17 Average labor income 3,097 8,667 5,581 5,946 14,560 17,581 29,167 12,021 Avcrage other income 0 0 0 0 0 895 0 294 Merchants and shop owners Row percentage 8.60 7.09 8.12 7.77 12.95 15.20 11.77 100 Column percentage 12.97 9.77 9.46 10.20 11.86 15.25 9.58 12.37 Average labor income 3,649 5,130 5,991 7,263 9,262 14,535 39,136 10,380 Average other income 841 552 175 1,054 1,621 337 1,937 722 Salesmen and trade employees Row percentage 8.69 11.26 15.04 9.88 11.07 8.96 7.05 100 Column percentage 14.87 17.59 19.58 14.71 11.51 10.20 6.51 14.26 Average labor income 3,635 4,104 4,132 5,779 8,337 8,870 23,814 6,085 Average other income 185 362 162 436 291 - 2.873 374 (Table continues on the following page) Table 14 (continued) Fourth Fifth Sixth Seventh Eighth Ninth Tenth Occupation decile decile decile decile decile decile decile Total Cooks, waiters, and others Row percentage 16.55 6.82 19.14 20.16 7.81 0 0 l00 Column percentage 2.93 1.10 2.62 3.11 0.84 0 0 1.47 Average labor income 3,975 4,092 4,583 5,281 3,883 - - 4,374 Average other income 0 1,291 0 0 0 - - 88 Security guards and night watchmen Row percentage 9.92 10.66 10.33 8.35 12.47 5.35 0.58 100 Column percentage 8.92 8.75 7.21 6.55 6.82 3.Z0 0.28 7.50 Average labor income 3,770 5,464 4,648 5,928 4,920 3,792 75,650 4,682 Average other income 244 913 832 250 298 3,175 0 466 Farmers and cattle ranchers Row percentage 0 0 8.38 0 16.65 15.39 25.28 100 o Column percentage 0 0 1.00 0 1.56 1.58 2.10 1.28 Average labor income - - 6,280 - 17,375 17,549 48,648 18,537 Average other income - - 0 - 0 1,423 8,085 2,454 Agricultural workers Row percentage 12.42 0 35.93 8.47 0 11.29 0 100 Column percentage 0.54 0 1.20 0.32 0 0.33 0 0.36 Average labor income 3,201 5,073 4,500 - 3,500 - 3,772 Average other income 0 - 0 0 - 0 - 0 Foremen, supervisors, and overseers Row percentage 2.80 8.98 19.89 22.56 27.56 13.17 0 100 Column percentage 0.56 1.64 3.07 3.92 3.34 1.75 0 1.66 Average labor income 5,810 6,369 7,946 7,130 7,665 16,384 - 8,408 Average other income 0 0 0 0 0 629 - 83 Food industry workers Row percentage 11.99 14.87 4.55 13.24 12.79 7.04 0 100 Column percentage 3.79 4.29 1.11 3.65 2.46 1.48 0 2.63 Average labor income 3,592 4,028 5,209 4,742 7,830 5,927 - 4,422 Average other income 79 0 0 0 0 0 _ 10 Tailors, dressmakers, upholsterers, and others Row percentage 5.94 25.34 18.19 4.78 11.45 4.47 0 100 Column percentage 2.09 8.15 4.95 1.47 2.45 1.05 0 2.94 Average labor income 3,168 3,633 5,948 5,587 5,024 5,004 - 4,060 Average other income 0 76 0 0 0 0 - 20 Cabinetmakers and woodworking machine operators Row percentage 8.02 19.57 15.62 9.58 19.34 0 0 100 Column percentage 3.18 4.86 3.29 2.27 3.20 0 0 2.27 Average labor income 5,490 2,695 3,385 6,305 5,594 - 4,077 Average other income 280 256 320 0 22 - - 168 Metalworkers Row percentage 8.55 9.32 11.06 13.01 18.01 9.10 4.16 100 Column percentage 5.34 5.32 5.33 7.08 6.84 3.78 1.40 5.20 Average labor income 3,343 3,190 4,570 6,028 5,650 7,286 15,417 5,074 Average other income 0 0 486 68 0 334 0 109 Machinery fitters, machine assemblers and installers, precision instrument mechanics, and watchmakers Row percentage 13.09 12.41 10.30 16.44 13.96 8.03 4.44 100 Column percentage 14.71 12.73 8.94 16.10 9.54 6.00 2.70 9.37 Average labor income 4,203 4,529 5,535 4,802 7,193 10,782 13,048 5,606 Average other income 37 0 273 76 3,865 0 1,790 696 Electrical workers and the like Row percentage 18.04 16.16 20.84 7.61 17.47 3.55 0 100 Column percentage 7.74 7.12 6.91 2.84 4.55 1.01 0 3.58 Average labor income 3,762 5,099 6,685 8,241 5,778 6,381 5,324 Average other income 0 137 324 687 845 569 - 313 (Table continues on the following page) Table 14 (continued) Fourth Fifth Sixth Seventh Eighth Ninth Tenth Occupation decile decile decile decile decile decile decile Total Plumbers, welders, and metal and structural metal preparers and installers Row percentage 8.26 8.41 13.61 16.47 23.84 2.78 1.32 100 Column percentage 3.07 2.85 3.90 5.33 5.38 0.69 0.26 3.10 Average labor income 3,576 5,451 6,548 4,475 5,254 6,503 10,000 4,740 Average other income 0 0 0 0 0 982 5,000 93 9 Total Row percentage 8.33 9.13 10.79 9.57 13.71 12.52 15.44 100 Column percentagea too 100 too too too 100 100 100 Average labor income 4,265 4,393 5,593 5.976 7,854 13,118 31,434 10,072 Average other income 312 250 218 264 918 767 3,101 314 -Data not available. a. The column figures don't add uip to 100 because certain occuLpations are not included in the table. For example, stonecutters and carvers were not included because they were represenited only in the second decile. Source: DANE, 1978 houseliold survey. CHANGES IN INCOMES 43 These include a proportion of skilled workers, such as plumbers; welders; metalworkers; fabricators and assemblers of metal structures; adjusters, assemblers, and installers of precision machinery and instruments; watch- makers; mechanics; overseers, supervisors, and foremen; and even workers in the food industry. In short, the middle class includes some skilled industrial workers. In addition, these three deciles also include a proportion of workers who in general would be assumed to be part of the upper-middle class. Examples are certain doctors, accountants, directors or executive personnel, and farmers and stockmen residing in Bogota. In summary, if the middle class is defined as the group of workers with incomes close to the urban average, it would include those falling between the seventh and ninth deciles of the family income distribution. The typical members of this group are white-collar workers, merchants, and teachers, but the group also includes large numbers of skilled workers, middle-level managers, and some professionals. Most professionals are, of course, in the highest decile of the distribution and are part of the upper class, even if they do not so consider themselves. As Table 14 shows, 61 percent of the architects and 70 percent of the lawyers are in the highest decile of the distribution. The decline in the real incomes of white-collar workers in the manufacturing industry was noted earlier. Apparently, the income of blue- collar workers and other unskilled persons rose faster than the incomes of white-collar workers during the latter part of the decade. Table 11 divides the industrial sector into groupings of skilled and unskilled industrial workers to permit an analysis of trends in the incomes of those skilled workers who, on the basis of household survey data, may be part of the middle class. The clothing, food, and wood industries are characterized by high proportions of unskilled labor. The opposite is true of the paper, petroleum products, and basic metals industries. The real incomes of the two groups behaved differently. In the 1960s, the real incomes of the skilled groups increased more than those of the unskilled. During this period, there was also a high return on education, as a number of studies on investment in human capital show. During the 1970s, the returns to education decreased along with the real incomes of skilled workers and white-collar personnel.'7 From 1970 to 1979, the real incomes of blue-collar workers in industries requiring a high skill level fell by 11 percent, not much less than the 18 percent decrease observed for white-collar workers in industry. In short, the 1960s were good for the middle class and the 1970s were not. For the period as a whole, the lower segments of the middle class 44 WINNERS AND LOSERS IN COLOMBIA'S GROWTH Table 15. Average Real Monthly Wages in Govemment Agencies, 1960-79 (constant 1954 Colombian pesos) District office Ministry of Finance Public Works Dept. (unskilled) Average for Year Professionala Messenger' Cleanera a department1 Retiredc Actived 1960 1,545 184 124 - - - 1961 1,613 234 175 - - - 1962 1,531 229 171 - - - 1963 1,271 173 130 - - - 1964 1,663 154 154 - - - 1965 1,561 220 185 - - - 1966 1,445 267 225 - - - 1967 1,318 247 208 - - - 1968 1,329 254 230 656 - - 1969 1,230 230 209 610 - - 1970 1,211 227 206 583 - - 1971 1,159 218 198 557 207 187 1972 1,278 200 182 647 209 192 1973 1,110 171 155 629 199 189 1974 1,045 205 169 553 195 180 1975 1,021 195 161 539 187 178 1976 898 173 143 521 202 192 1977 852 205 163 551 191 182 1978 825 221 185 575 202 189 1979 813 227 210 527 196 185 -Data not available. Note: The wage figures are deflated by the consumer price index for blue-collar and white-collar workers in Bogota and Medellin as applicable. Base: 1954-55 = 100. a. Basic wage. Source: Up to 1967, Miguel Urrutia and Mario Arrubla, Compendio de Estadisticas Historicas de Colombia (Bogota: Universidad Nacional, 1970). For 1968-79, records of the Ministry of Finance. b. Simple average of basic wages of all Direccion General de Presupuesto employees, excluding the director. c. Basic wage in August taken from budget for the year. Average for thirty-one persons. Conversion factor daily to monthly: twenty-five. d. Basic wage in August, taken from budget for forty persons. Daily converted to monthly. CHANGES IN INCOMES 45 Boyaca electricity company Teachers in Antioquiae Manager Head of personnel Lineman Primary' Secondaryf Salaryg Incomeh Salary' Incomeh Salary' Incomeh - - 1,792 - 508 - - - - - 2,268 - 567 - - - - - - - 530 - - - - - 2,462 - 685 - - - - - 2,363 - 788 - - - - 1,990 - 680 - - - - - 2,006 - 645 - - - 383 504 2,338 2,481 700 930 392 420 364 478 2,398 2,712 683 1,028 364 452 365 496 2,616 2,903 785 1,065 343 424 371 518 2,413 3,070 - - 345 431 360 513 2,183 - 713 885 358 439 322 490 2,186 2,729 779 1,039 346 404 397 498 2,480 2,733 687 876 298 383 357 605 2,396 2,892 603 767 334 372 309 530 2,049 2,517 - - 354 374 293 498 1,633 2,009 702 - 277 325 330 498 1,669 2,011 708 788 289 358 375 470 1,637 - 688 - 309 338 e. Total population taken into account. Source: Records of Fondo Educativo Regional, Antioquia. f. Monthly average per teacher, including bonuses. g. Including basic wage, plus cost-of-living and entertainment allowances. The figure shown is the statutory salary for that job. h. Including overtime and allowances. Excluding the following payments: retirement compen- sation and bonuses, living allowance, medical care, dental services, opthalmological services, payment of electricity charges, and so forth. i. Including basic wage plus cost-of-living allowance. The data reflect the actual earnings of a person holding that job. Source: Records of the government agencies. 46 WINNERS AND LOSERS IN COLOMBIA'S GROWTH enjoyed greater gains than the upper segments, if one considers that white- collar workers have higher incomes than skilled workers. Data on the financial sector confirm this conclusion. The personnel of one bank studied had smaller income gains than skilled industrial workers, particularly during the 1970s. Within that bank, personnel below the professional level also enjoyed relatively larger wage increases than more skilled individuals. Table 15 presents some data on the monthly wages in the public sector. Although professionals working in the Ministry of Finance suffered a clear decline in salaries, unskilled personnel in the same ministry saw some improvement in their income. During the 1970s, the real incomes of messengers, cleaning staff, and blue-collar workers in the public works department in the federal district remained constant. In general, that part of the middle class comprising government white-collar workers lost out in relative terms; the loss was much more pronounced for skilled per- sonnel. Wage surveys conducted in industries by consulting firms show that the upper-middle class also did not have large real wage gains. Although there are problems in trying to compare the data over time, these series confirm the decline in the relative position of the middle class. Table 16 shows a reduction in the real incomes of white-collar workers and supervisors in industry, particularly during the 1970s. Table 17 indicates a general stagnation in the wages of higher-level white-collar workers in 1977-79, a period when the wages of agricultural daily workers and the incomes of unskilled workers (including the unskilled white-collar workers in Table 17) were rising. Table 18, based on a different source, shows the same trends. Little information is available on the incomes of the upper class. Nonetheless, a number of indicators suggest that the earnings of this group rose rapidly during the 1970s. One such indicator is the rapid expansion of the supply of luxury housing and the sharp rise in real prices for such housing. Increases in automobile sales, shown in Table 19, also seem to reflect the rising incomes of the highest decile of the distribution. Since the income of the middle class did not rise during the decade, it may be assumed that the majority of the automobile purchases were made by the upper class.'8 The number of automobiles per 100 households in the wealthiest sector of Bogota rose by 8 percent between 1972 and 1978.29 In the second wealthiest area, the ownership of automobiles per 100 families decreased. The data on changes in automobile ownership, therefore, confirm the rapid rise in the incomes of the wealthiest families and the stagnation or decrease in the incomes of the middle class. In some lower- Table 16. Trends in the Monthly Remuneration of Selected Executive, Production, and Shop Personnel, 1961-79 (constant 1954 Colombian pesos) July November July First half Second half March Occupation 1961 a 1962b 1968c 1972d 1976e 1979f Beginning engineer 1,065 1,140 1,077 - - - Engineer (with two years' experience) 1,432 1,500 1,474 - - - Head of productiong 2,569 1,996 2,098 2,803 2,144 2,196 Head of engineering, 2,459 2,236 2,994 3,057 2,087 1,867 Head of industrial relations 1,866 - 2,428 3,134 2,706 2,787 Head of industrial relations (plant) - - - 1,703 1,642 1,622 Production manager - - - 2,533 2,162 2,062 Financial manager - - - 3,767 3,541 3,220 Head of accounting section - - - 1,143 964 865 Produiction supervisor' 659 _ 646 1,228 1,091 926 Maintenance shop supervisor1 664 769 728 1,209 785 546 -Data not available. Note: The figures are deflated by the consumer price index for white-collar workers: base 1954-55 100. The price index for Cali was used for the 1961, 1962, 1968 data; for the other years, the national consumer price index was used for the month of the survey, or the average for the year or six-month period, as appropriate. a. Eleven firms surveyed. b. Seven firms surveyed. c. Eight firms surveyed. d. Thirty-two firmns surveyed. e. Sixty-six firms surveyed. f. Fifty-eight firms surveyed. g. From 1972, sanme as head of production departmnent. h. From 1972, same as head of plant engineering and subsequently head of project engineering. i. In 1972, same as general production supervisor and in 1976-79 same as production supervisor. j. Beginning in 1972, same as maintenance supervisor. Source: Asesorias y Servicios Indtistriales. Information for years before 1972 is based on a suimmary of salary and wage surveys made in 1961, 1965, and 1968 for firms interested in the preparation of salary scales in the Cali area. From 1972 onward, the results correspond to a semiannual survey of salaries and benefits. 48 WINNERS AND LOSERS IN COLOMBIA'S GROWTH Table 17. Average Monthly Salaries for Selected Executive and Office Positions, 1977 and 1979 (Colombian pesos) 1977a 1979b Current Constant Current Constant Position prices prices' prices prices' Accounting supervisor National market average 20,395 1,283 - - Food and beverages 25,567 1,608 27,478 1,173 Textiles and apparel 20,911 1,315 26,891 1,148 Metalworking 19,946 1,254 27,982 1,195 Pharmaceuticals and chemicals 24,675 1,552 32,550 1,390 Other manufacturers 17,834 1,122 24,475 1,173 Manufacturing 20,451 1,286 28,105 1,200 Marketing chains 20,250 1,273 - - Banks and financial institutions 19,833 1,247 General accounting assistant National market average 5,444 342 - - Food and beverages 5,701 358 9,791 418 Textiles and apparel 5,723 359 8,277 353 Metalworking 5,351 336 8,407 359 Pharmaceuticals and chemicals 6,855 431 9,810 419 Other manufactures 6,456 406 10,429 445 Manufacturing 6,021 379 9,345 399 Marketing chains 5,019 316 - - Banks and financial institutions 4,408 277 General cashier National market average 6,875 432 - - Food and beverages 8,458 532 13,871 592 Textiles and apparel 8,687 546 11,673 499 Metalworking 9,331 587 12,831 548 Pharmaceuticals and chemicals 8,401 528 11,941 510 Other manufactures 8,807 554 11,365 485 Manufacturing 9,019 567 12,409 530 Marketing chains 10,355 651 - - Banks and financial institutions 5,227 329 Payroll officer National market average 7,464 469 - - Food and beverages 7,378 465 12,246 523 Textiles and apparel 7,868 495 11,268 481 Metalworking 7,056 444 10,829 462 Pharmaceuticals and chemicals 7,226 454 12,335 527 Other manufactures 8,063 507 13,968 596 Manufacturing 7,592 477 12,228 522 Marketing chains 8,049 506 - - Banks and financial institutions 6,493 408 CHANGES IN INCOMES 49 1977a 979' Current Constant Current Constant Position prices prices' prices prices Receptionist National market average 3,999 251 - - Food and beverages 4,419 278 6,722 287 Textiles and apparel 4,049 255 6,059 259 Metalworking 4,054 255 5,985 255 Pharmaceuticals and chemicals 4,479 282 6,443 275 Other manufactures 4,559 283 7,307 312 Manufacturing 4,330 272 6,553 280 Marketing chains 3,372 212 - - Banks and financial institutions 3,483 219 Internal messenger National market average 2,647 166 - - Food and beverages 3,574 225 6,510 278 Textiles and apparel 3,250 204 5,056 216 Metalworking 2,733 172 5,039 215 Pharmaceuticals and chemicals 3,337 210 5,108 218 Other manufactures 3,392 213 5,515 235 Manufacturing 3,281 206 5,377 230 Marketing chains 2,390 150 - - Banks and other institutions 2,276 143 Cleaning worker National market average 2,526 159 - - Food and beverages 3,408 214 5,366 229 Textiles and apparel 2,849 179 4,959 212 Metalworking 2,915 183 5,479 234 Pharmaceuticals and chemicals 2,965 186 5,050 216 Other manufactures 3,463 218 5,484 234 Manufacturing 3,113 196 5,252 224 Marketing chains 2,024 127 - - Banks and financial institutions 2,495 157 a. Results of a national salary survey conducted in 122 firms located in four departments. Information as of May 31, 1977. b. Results of a national salary survey conducted in 132 firms located in five departments. Information for the first six months of 1979. c. Deflated by the national consumer price index for white-collar workers: base 1954-55 = 100. Source: Specialized consulting firm. 50 WINNERS AND LOSERS IN COLOMBIA'S GROWTH Table 18. Average Real Monthly Remuneration to Labor in Selected Executive and Office Positions, 1977 and 1979 (constant 1954 Colombian pesos) 1977" 1979b Basic Total Basic Total Position wage' remunerationd wage' remunerationd Executive positions' Financial director 3,696 4,110 3,890 4,138 Plant manager 3,482 3,870 3,702 3,872 Production manager 2,092 2,342 2,373 2,526 General and cost accountant 1,654 1,828 1,707 1,823 Office positions Bilingual executive secretary, management 942 1,047 1,010 1,087 Spanish-speaking secretary, Class 1 580 649 538 580 Stenographer-typist 396 445 406 439 Receptionist 358 396 342 371 Messenger 247 276 296 320 Note: Figures are deflated by the national consumer price index for white-collar workers: base 1954-55 = 100. a. The survey covered twenty companies, of which twelve operate in the medical products market and the remaining eight in mass consumption products. At the management and administration levels, fifty-two positions were surveyed. The twenty companies had over 7,800 employees, of whom 35 percent were members of trade unions. The survey was conducted in 1976 on salaries and extra-legal benefits that would be in effect in 1977. b. In 1979, eighteen companies were included. The survey was conducted in May 1979, and the results were projected through December. c. Based on the average of thirteen months' basic wages, including the legal bonus for services. d. Including the basic wage plus extra-legal benefits. e. In 1979, a large percentage of those holding executive positions had company cars (especially the financial director, plant manager, and marketing manager). In some cases, moreover, club memberships were paid for the employee, together with a percentage of the basic monthly wage as "permanent allowances." (The percentage ranged from a low of 10 percent to a high of 50 percent.) Source: Private consulting firm. Table 19. Number of Private Automobiles Sold in Colombia, 1970-80 1970 1971 a 1972 1973 1974 1975 1976 1977 1978 1979 1980b Total 4,972 11,586 16,612 22,304 28,954 25,047 24,684 28,752 35,355 32,651 19,746 Domestic 3,851 10,703 15,668 20,413 27,093 23,238 23,347 27,275 33,342 30,670 16,177 Imported 1,121 803 944 1,819 1,861 1,809 1,337 1,477 2,013 1,981 3,569 Percentage of total imported 22.5 6.9 5.7 8.2 6.4 7.2 5.4 5.1 5.7 6.1 18.1 Mid-size automobiles Domesticc 355 1,826 5,692 8,620 9,858 8,090 9,979 12,183 16,195 13,749 10,439 Importedd 350 228 294 974 684 1,016 600 612 474 426 2,248 ,=, Large-size automobiles Domestic Imported' 75 104 228 218 186 209 150 326 309 169 Luxury automobiles Domesticf 1,364 2,467 1,703 1,324 1,790 1,987 1,830 2,127 2,834 2,403 1,080 Importedg 725 556 526 673 926 610 522 679 1,163 1,236 1,152 a. Renault began production in 1971. b. Data for the first six months. c. Dodge 1800, Renault 12. d. Alpine, Fiat 125. e. Ford, Datsun, BMW. f. Dart, Coronado, Ford. There are no large-size automobiles produced domestically. g. Mercedes, BMW. Source: Compania Colombia Automotriz. 52 WINNERS AND LOSERS IN COLOMBIA'S GROWTH middle-class and middle-middle-class sectors of the city, ownership of automobiles increased. This supports the other evidence presented in this chapter, which suggests that the income differential within the middle class may have been reduced during the decade. The next chapter, which analyzes the income trends of the upper class and the lower class in Cali, also confirms these impressions about the incomes of the highest decile of the distribution. Conclusion All of the information compiled on incomes by occupational categories points to the conclusion that the middle class lost part of its share of national income during the 1970s. It is even possible that the real incomes of this group declined. Nonetheless, to be certain of this latter conclusion, one would have to analyze the participation of members of middle-class families in the labor force. The increased participation of females may have permitted families to maintain their standards of living during the period. While the gap between the income of poor agricultural workers and small farmers, and other workers probably decreased, among urban workers the poor and the unskilled did better than the skilled and those employed in the formal sector. Notes to Chapter 2 1. Miguel Urrutia and Albert Berry, La Distribucion del Ingreso en Colombia (Medellin: Editorial la Carreta, 1975), tables 1 and 2 of chapter 1. 2. W. Arthur Lewis, "Development with Unlimited Supplies of Labour," The Manchester School, vol. 27, no. 2 (1954), pp. 139-92. 3. Kazuchi Ohkawa and Gustav Ranis, "On Phasing' (Paper presented at the Conference on Japan's Historical Development Experience and Contemporary Developing Countries, Tokyo, 1978; processed). 4. John C. H. Fei and Gustav Ranis, for example, have concluded this from their studies of the Republic of Korea and Taiwan. The theory is summarized in Ohkawa and Ranis, "On Phasing." 5. Although it would be better to compare growth in wages with growth in income per worker, there are no data on the economically active population for the noncensus years, and, given the instability of labor force participation rates, it was felt unwise to estimate those figures, Nevertheless, it is clear that between 1950 and 1970, population grew faster than the labor force. In that case, average income per employed worker grew faster than the agricultural wage in that period and, therefore, agricultural workers probably saw their relative income position deteriorate. In contrast, between 1965 and 1980, the labor force grew more rapidly than population, so the rate of growth per employed worker was lower than the 3.67 percent growth in per capita income (see the income figure at the bottom of column 4 in Table 2). It is, therefore, very clear that rural laborers improved their relative position in the period 1965-79 (while population grew 52 percent during CHANGES IN INCOMES 53 1965-80, the labor force grew 64 percent). The 1965 figure comes from International Labour Office (uLO), Labour Force Estimates and Projections 1950-2000 (Geneva, 1977), p. 34. The 1980 figures are from La Economica Colombiana en la Decada de los Ochenta (Bogota: Fedesarrollo, 1979), p. 43 and p. 70. 6. Municipal wages are weighed by the population figures of each municipality, once the population of the major city in the district is subtracted. The average wage of populous municipalities has, therefore, a larger weight in the departmental average. 7. If female wages are considered in the calculation of the most frequent wage by municipality, the new methodology would lower the wage level. If the distribution of municipal wages by department is skewed and dispersion is large, average wages will be higher than most frequent wages. This would mean that the second methodology may produce higher wage levels. 8. The agricultural wage series published by the Social Security Institute was analyzed and discarded. A small proportion of employers in the rural sector affiliate their workers to icsS (Instituto Colombiano de Seguros Sociales), because icss has few clinics in the rural sector. An exception might be Valle, where the big sugar plantations have been traditionally affiliated and where ICSS has clinics in many of Valle's towns. The employers originally affiliated to icss were the high-wage, large-scale, and capital-intensive enterprises. As icss services penetrated into the countryside, smaller and lower-wage employers joined, such as the companies producing flowers for export, which usually employ women at the minimum wage. Because of the entrance in the 1970s of smaller firms into the tcSs network, the social security wage data would appear to have a serious bias toward underestimating wage trends. 9. Helena Jaramillo, "Determinants of Income Differentials after Migration" (New Haven: Yale University, 1978; processed); and Rakesh Mohan, The People of Bogota: Who They Are, What They Earn, Where They Live, World Bank Staff Working Paper no. 390 (Washington, D.C., 1980). 10. Urrutia and Berry, La Distribucion. 11. This is not surprising when it is considered that agroindustry has pushed up agricultural wages in the department of Valle. In 1977, for example. Gomez-Angel discovered that in the squatter district of Cali the percentage of poor families who have electrical appliances was about the same as that for 479 families on a sugar plantation in the rural area of Valle. This suggests similar income levels. See Jairo, Gomez-Angel, "Traditional and Emerging Pattems in the Social Organization of a Large Estate in the Cauca Valley" (Ph.D. dissertation, Louisiana State University, August 1979). 12. The DANE annual survey defines salaries and wages as the fixed or ordinary remuneration that the worker received during the month in return for services rendered before deduction of items withheld at the source, such as social security, union dues, and the like. Employee fringe benefits comprise compulsory or voluntary payments other than salaries and wages made by employers to their employees. 13. Mohan, The People of Bogota. In July 1983, one U.S. dollar was roughly equal to 80 Colombian pesos (Col$). 14. Most income distributions are skewed, as opposed to normal, because there are many more people in the lower incomes than in the higher. This means that the most frequent income (mode) is in the low-income ranges, but the few high-income earners increase the average so that it is usually quite a bit highet than the modal income. 15. The DANE household survey for December 1978 found that average per capita income in Bogota was Col$7,010 a month, while the average income of the eighth decile of the distribution of per capita income per household was Col$6,851 a month. 16. The data on salaries and wages of white-collar workers in industry would also place them in these intervals of the distribution. 17. See Franqois Bourguignon, "Probleza y Dualismo en el Sector Urbano de las Economias en Desarrollo: El Caso de Colombia," Desarrollo y Sociedad (anuary 1979), and Bernardo Kugler, Alvaro Reyes, and Maria Isabel de Gomez, Educacion y Mercado de Trabajo Urbano en Colombia: 54 WINNERS AND LOSERS IN COLOMBIA'S GROWTH Una Comparacion entre Sectores Modemo y no Moderno, monograph 10 (Bogota: Corporacion Centro Regional de Poblacion, 1979). 18. According to Pachon, 73 percent of the households in the highest decile had an automobile in 1978, and 22 percent of those families had two or more cars. Only 40 percent of the families in the ninth decile and 20 percent in the eighth decile had an automobile. See Alvaro Pachon, "Automobile Ownership in Bogota and Cali" (The World Bank City Study Workshop, Bogota, February 13, 1980; processed). 19. Ibid., tables 1-6 and 201. The average monthly income in this section of Bogota in 1978 was Col$31,030 per family, whereas the city average was Col$13,150. 3 Changes in Real Income of a Set of Rich and Poor Families in Cali WAGE SERIES SHOW AVERAGE CHANGES in the incomes of different types of workers, but do not show what happens to the standard of living of individuals through time. It is possible, for example, for real wages in construction and manufacturing to remain constant while individuals improve their standards of living by moving from the construction sector to the industrial sector. Although changes in the production structure, which affect distribution through occupational mobility, can be captured by comparing estimates of the national income distribution, obtaining countrywide income data of comparable quality is not easy. An alternative is to analyze occupational histories of individual households. This is difficult, however, in a country with high or variable inflation, such as Colombia, because people tend not to remember the exact date in the past when they earned a certain nominal wage.' Fortunately, however, we have found income data for a sample of poor and rich families in Cali based on a set of surveys of these families conducted throughout the 1970s.2 This data set is exceptional, for as the World Development Report 1978 states: "Very few [household surveys] follow the fortunes of individuals and families through time, or disaggregate the household to examine the well-being of women, children, and the elderly."3 The surveys were carried out as part of a study on nutrition by the Fundacion de Investigaciones de Ecologia Humana of Cali. They included information on income levels, quality of housing, education, and other socioeconomic indicators for a group of families living in one of the poorest squatter districts (Union de Vivienda Popular) of Cali in 1970. In addition, the researchers looked into the incomes and welfare indicators of a group of high-income families as a control group.4 55 56 WINNERS AND LOSERS IN COLOMBIA'S GROWTH The 1970 survey identified low-income families with children who showed signs of malnutrition. These children subsequently scored much lower in tests of cognitive ability than children from the sample of families at a high socioeconomic leveL Nonetheless, the research and action program designed by the Fundacion de Ecologia Humana demonstrated that food supplements and psychological stimulation in special day-care centers for poor children enabled seven-year-olds who had been part of this program for four years to score measurably higher in tests of ability than children who were in the program for only one year, though these scores were still lower than those achieved by children of families at a high socioeconomic leveL Table 20 shows the general cognitive ability of the children in the sample and suggests the high cost to one's personal development of being born poor. It also indicates how programs to offset the disadvantages resulting from malnutrition and the lack of stimulation in childhood can help equalize the opportunities of individuals. Characteristics of the Sample For the purposes of this report, the important aspect of the research is that detailed information on the incomes and living conditions of a group of both low-income and high-income families in Cali was compiled in various years during a decade. Because it was possible to locate most of the Table 20. General Cognitive Ability of Rich and Poor Children in Cali, 1970 (Average scaled scores) Average age (in months) Number of Social group of children children 43 49 63 77 87 High income 28 -0.11 0.39 2.28 4.27 4.89 Low income Treated for 4 periods 50 -1.82 0.21 1.80 3.35 3.66 TTeated for 3 periods 47 -1.72 -1.06 1.64 3.06 3.35 Treated for 2 periods 49 -1.94 -1.22 0.30 2.61 3.15 Treated for I period 90 -1.83 -1.11 0.33 2.07 2.73 Note: The higher the test scores, the better they are and the more cognitive ability they reflect. A treatment period is equivalent to a school year at the day-care center. Source: Harrison McKay and others, "Improving Cognitive Ability in Chronically Deprived Children," Science, vol. 200 (April 21, 1978), p. 275. CHANGES IN REAL INCOME IN CALI 57 same families in 1980, a final survey of most of them was carried out ten years later, to complete the history for a ten-year period. The low-income families are a sample of the population in a district that was one of the poorest neighborhoods of Cali in 1970. The families were selected from a group with children born between June 1 and November 30, 1967; from this group, the children selected were those with the lowest weight and height, those with symptoms of clinical malnutrition, and those from families with the lowest incomes. The families were at the lowest quintile of the income distribution. The high-income families were within the highest 5 percent of the distribution.5 Table 21 compares selected characteristics of the high-income and low- income families in 1970. Clear differences are seen in the nutritional level of the children (as indicated by their weight and height), the education of their parents, and family income. Both the high-income families and the low-income families were relatively young in 1970: the average age of the father was thirty-seven and that of the mother thirty-one, while the number of family members under fifteen years of age was 4.8 for the poor and 2.4 for the rich. Table 22 shows the number of families for which comparable informa- tion is available for each year. The 1980 survey was conducted especially for this study in the months of April and May. Table 21. Characteristics of Families in the Sample, 1970 Low- income High-income Variable group group Height of child as a percentage of normal for age 90 101 Weight of child as a percentage of normal for age 79 102 Per capita family income as a percentage of per capita family income of the rich 5 100 Per capita expenditure for food as a percentage of expenditure in rich families 15 100 Number of family members' 7.6 5.2 Number of members under 15 years of age 4.8 2.4 Number of playrooms and bedrooms per child 0.3 1.6 Average age of father 37 37 Average age of mother 31 31 Average years of education of father 3.6 14.5 Average years of education of mother 3.5 10.0 a. From Table 23. Source: McKay and others, "Improving Cognitive Ability in Chronically Deprived Children." Table 22. Number of Families for Which Usable Information Is Available, 1970-80 1970 1974 1976 1980 High Low High Low High Low High Low Item of information income income income income income income income income Total income 37 176 35 159 34 171 34 152 00 Expenditures for food 37 175 35 162 33 173 37 168 Income of principal eamer 37 176 35 161 32 171 28 152 Total families in survey 37 183 35 184 34 184 37 186 Note: Not all of the information from the available surveys could be used for each year because data are missing in some categories. In some cases, the expenditure reported for food is greater than total income, or figures are suspect (outliers). Such cases have been exclided for the analysis of the data. Source: Fundaci6n de Investigaciones de Ecologia Humana (FIEH), Cali; and Fedesarrollo, Bogota. CHANGES IN REAL INCOME IN CALI 59 Trends in the Real Income of Two Social Classes The real family income and the per capita income of both socio- economic groups rose rapidly at a rate exceeding that of the national per capita income (see Table 23). The differential in family income between the rich and the poor improved to the benefit of the latter. Because of the greater increase in the number of persons per family in the low-income group, however, real per capita income increased less than family income, and less than the real per capita income of the wealthy families.6 This illustrates how differences in family size between the two social classes favor the real consumption level of wealthy persons. However, the increase in family size may have been the means by which the poor added to their family income through the financial contribution made by all family members able to work. Table 23. Average Monthly Real Incomes of Rich and Poor Families in Cali, 1970-80 (constant 1970 Colombian pesos) Annual rate of increase, 1970-80 Income group 1970 1974 1976 1980 (percent) High income Nominal family income 7,831 17,866 37,070 82,688 26.6 Real family incomea 7,199 11,028 13,070 14,034 6.9 Nominal per capita income 1,513 3,489 7,596 16,705 27.0 Real per capita incomea 1,395 2,153 2,678 2,835 7.3 Low income Nominal family income 748 1,654 2,562 9,865 29.4 Real family income' 728 929 909 1,546 7.8 Nominal per capita income 98 199 303 1,156 28.0 Real per capita incomea 96 112 107 181 6.5 Income differentials Family (5 - 1) 0.09 0.09 0.07 0.12 Per capita (7 - 3) 0.06 0.06 0.04 0.07 Real national per capita incomeb 428 498 539 613c 3.7 a. For purposes of the present study, the base used was December 1970-100. Average real income is the nominal monthly income of each family, deflated by the Cali blue-collar consumer price index for the month in which the survey was made. b. The figure used is that for December of the year closest to the month in which the survey was conducted. The base was 1970. Source: Banco de la Republica, national accounts. The annual income was divided by twelve to obtain a monthly estimate. c. Fedesarrollo estimate. Source: FIEH and Fedesarrollo sample. 60 WINNERS AND LOSERS IN COLOMBIA'S GROWTH One reason for the increase in family income has been the rise in the average number of persons working per family, especially among the poor (see Table 24). This increase is closely related to a significant rise in the participation of women and secondary workers in the labor force; moreover, the increase in participation has been greater in the low-income group.7 Finally, the data show that the increase in the number of women working did not contribute much to the rise in the family income of rich families because earnings of females in the high-income groups are low compared Table 24. Effect of Participation in the Labor Force on Incomes of the Rich and the Poor in Cali, I970-80 Income group 1970 1974 1976 1980 Average number of persons working per family High income 1.46 1.37 1.38 1.54 Low income 1.33 1.37 1.63 2.10 Average number of females working per family High income 0.32 0.29 0.38 0.56 Low income 0.24 0.28 0.46 0.68 Females working as a percentage of all females over 12 years of age High income 23 19 24 24 Low income 16 14 22 23 Average number of males working per family High income 1.13 1.08 1.00 0.97 Low income 1.07 1.08 1.16 1.42 Males working as a percentage of all males over 12 years of age High income 89 84 85 58 Low income 70 51 50 50 Average income of females as a percentage of average income of males High income 55 36 30 25 Low income 41 39 50 65 Average real income of head of household (Col $)a High income 5,652 9,728 10,230 12,360 Low income 646 760 695 916 Average real family income (Col $)a High income 7,199 11,028 13,070 14,034 Low income 728 929 909 1,546 Income of head of household as a percentage of total income High income 78 88 78 88 Low income 88 81 76 59 a. Deflators are Cali cost-of-living indexes for blue-collar workers. Source: FIEH and Fedesarrollo sample. CHANGES IN REAL INCOME IN CALI 61 with those of men and have decreased relatively throughout the decade. Among low-income families, in contrast, not only did the participation of females in the labor force increase, but their relative incomes rose as well. The effect of participation in the labor force is shown in Table 24, in which it can be seen that the increase in the real incomes of heads of households in the high-income group was 119 percent during the decade, but was only 42 percent in the low-income group. During the same period, real family income rose 94 percent in the high-income group and 112 percent in the low-income group. It is clear, then, that the economic improvement of poor families has been due to the expansion of employment opportunities and to the higher eamings of secondary workers. One of the most significant phenomena in the decade for the groups studied in Cali was the change in the participation of females in the labor force. Not only did female employment increase, but the gap between wages of females and males narrowed at lower skill levels.' The increased participation of women in the labor force could theoretically have been due to increasing poverty and male unemployment forcing women to enter the labor market. The increasing real wages of women and the decreasing income differential between the income of women and men, however, suggest that increasing labor demand was the determinant of higher participation rates. The hardship on both the working women and their children brought about by the increased participation in the Labor force was probably not severe, because the families were large and there were other women in the household who could take care of the children and the housework. As Table 24 shows, by 1980 only 23 percent of females over the age of twelve from poor families were in the labor force. Changes in Spending for Food Changes in food consumption are a good indicator of changes in the welfare of low-income families. Table 25 shows how expenditure for food has changed among the two population groups studied. Both family groups increased their consumption of food during the period. This indicator is an excellent nutrition index for families at a higher socioeconomic level, inasmuch as the average number of persons per household declined. At the lower level, however, the average number of persons per household rose from 7.6 to 8.5 over the decade. As may be seen, real per capita expenditure for food rose among the low-income group, but the increase was concentrated in the second part of the decade. 62 WINNERS AND LOSERS IN COLOMBIA'S GROWTH Table 25. Changes in the Expenditure on Food by Rich and Poor Families in Cali, 1970-80 (Colombian pesos) Income group 1970 1974 1976 1980 High income 1. Nominal expenditure on food 2,584 4,942 7,876 18,946 2. Nominal family income 7,831 17,866 37,070 82,688 3. Real expenditure on fooda 2,350 2,656 2,401 2,794 4. Real per capita expenditure on food 452 521 490 570 5. Row I as a percentage of row 2 32 28 21 23 Low income 6. Nominal expenditure on food 579 1,055 1,702 5,423 7. Nominal family income 738 1,609 2,539 10,583 8. Real expenditure on fooda 564 541 558 798 9. Real pet capita expenditure on food 74 65 66 94 10. Row 6 as a percentage of row 7 79 65 67 51 Note: The income data in this table are slightly different from that in Table 23 because the data on food expenditures were incomplete for some families where there were usable income data. This table, therefore, coveTs a smaller number of families than are covered in Table 23. a. Real expenditure on food is nominal expenditure divided by the food price index for the month of the survey. Source: FIEH and Fedesarrollo sample. Engel's law states that as a family's income rises, a smaller proportion of it is spent on food. Moreover, a number of studies made in Latin America suggest that urban families who spend more than 65 percent of their income for food can be considered very poor and below the poverty line. The decline in the proportion of expenditure on food in both groups in Cali confirms that real income has been rising, and the decline from 79 percent to 51 percent for the low-income group suggests that families are moving out of absolute poverty levels. The table also shows, however, that the real food expenditures of the poor deteriorated in the early 1970s and improved substantially in the last part of the decade. Rapid increases in food prices and the pressure to consume other goods, such as public services, may explain the decrease in food expenditures in the first part of the decade (see Figure 3). Other Indicators of Welfare The quality of housing and the access to public services are also good indicators of welfare. In the case of the residents of the squatter district of CHANGES IN REAL INCOME IN CALI 63 Figure 3. Indicators Reflecting the Physical and Hygienic Quality of Dwellings of a Poor Neighborhood in Cali, 1970-80 Percent 100 ~~~~~~~~~~~~~Potable water supply 90 _ / p-~~~ ~ ~ ~ ~ ~ SewSerage 80 _ 70 _ 60 - sO _ 40 - 30_ 20_ 10_ DiTT floors 0 ?S ~~~~~I I 1970 1974 1976 1980 Year Source: Fundacion de Investigaciones de Ecologia Humana (FIEH)-Fedesarrollo sample. 64 WINNERS AND LOSERS IN COLOMBIA'S GROWTH Cali, the improvement in these indicators has been significant (see Figure 3). In 1970, no family had sewerage, but this service was available to 90 percent of the families by 1974 and 99 percent by 1980. Connections to potable water supply rose from 6 percent to 100 percent during the same period. The decrease in the proportion of houses with dirt floors (42 percent to 9 percent) is another important improvement from the standpoint of health. The proportion of houses with brick walls also rose. Connection to the electricity system implies an increase in comfort, which is reflected in the fact that about 80 percent of the families were cooking with electric stoves in 1980 compared with 90 percent who cooked with oil in 1970. Table 26 shows that as income rose, the poor families in Cali began to consume durable goods.9 In 1976, for example, only 25 percent of these families owned refrigerators; by 1980, this proportion had risen to 48 percent. Ownership of a refrigerator is especially important from the standpoint of nutrition and comfort. In general, the consumption capacity of families like these seems to have increased over time, and they no longer constitute a negligible part of the market for domestic industry. Govemment programs have made an obvious contribution to the improvement in the standard of living of this district. The local govemment has brought electricity, water, sewerage, and other services to the neighbor- hood at comparatively low prices, subsidized through higher connection charges and service rates in higher-income districts. Because the value of these services is greater than their cost to the user, this has contributed to an improvement in welfare.10 The effect on health and nutrition of greater access to potable water and sewerage is clear. Electricity not only affords greater convenience but also, because of the time saved in performing Table 26. Percentage of Poor Families in Cali Possessing Selected Durable Goods, 1976-80 Percentage increase Item 1976 1980 1976-80 Refrigerator 25 48 93 Radio 82 83 2 Sewing machine 35 55 56 Television 31 72 132 Record player 16 27 73 Othera - 91 - -Data not available. a. Including electric stoves, irons, and blenders. Source: FIEH-Fedesarrollo sample. CHANGES IN REAL INCOME IN CALI 65 household chores, provides the possibility for additional family members to seek employment. Single-Parent Families During the decade, a high percentage of both rich (21 percent) and poor (24 percent) families broke up. Average family income generally was lower in families where the father had left the household. Incomes of both the poor and the rich single-parent families were lower than the overall average for the respective social group (see Table 27). Moreover, the rate of annual increase in the total incomes of the single-parent rich families was lower than the average for all rich families. Although single-parent poor families have lower incomes than the average for all poor families, their real incomes rose more rapidly than the average among poor families because the earnings of women and secondary workers in this social class grew so rapidly.'1 In other words, because the incomes that increased most were those of wives and because secondary workers were able to find employment, single-parent poor families saw an improvement in their relative situations during the decade. This phenomenon probably helped to alleviate significantly problems of extreme poverty. Table 27. Average Monthly Incomes of the Rich and the Poor Single-Parent Families in Cali, 1970-80 (Colombian pesos) Annual rates of growth Income by social group 1970 1974 1976 1980 (percent) Nominal family income High income 7,210 8,800 16,250 46,989 20.6 Low income 466 1,257 2,241 8,961 34.4 Real family incomea High income 6,637 5,403 5,610 7,975 1.8 Low income 461 688 795 1,401 11.8 Family income of single-parent family as percentage of average family income High income 92 49 44 57 Low income 62 76 87 91 a. Real family income is nominal income deflated by the Cali consumer price index for blue- collar workers (low-income families) and white-collar workers (in the case of high-income families). Source: FIEH and Fedesarrollo sample. 66 WINNERS AND LOSERS IN COLOMBIA'S GROWTH Occupational Mobility Analysis of the employment history of the sample does not suggest much upward or downward mobility. There are cases such as that of the head of household who declared in 1970 that his occupation was "whatever turns up"; in 1974 he was a warehouse worker, in 1976 a fruit seller, and in 1980 a tradesman. Most individuals, however, remain in the same occupational category over time. In other words, the rise in income in the low socioeconomic group resulted more from increases in the number of family members in the labor force and from higher earnings than from the occupational mobility of the father or the mother over time. Table 28 uses a broader definition than usual to measure mobility. In some studies, a person is considered to have ascended on the occupational scale only when he moves from laborer to clerk, or from self-employed worker to owner. In the present case, a person was considered to have ascended if he moved from caretaker to production worker, or from street peddler to industrial worker. Even with this broad criterion, however, it is clear from the table that there was little occupational mobility. Table 28. Measurements of Occupational Mobility, 1970-80 (number of workers) Social group Males Females High income Upward mobilitya 2 1 Downward mobility 1 0 Total workersb 29 12 Low income Upward mobility 30 5 Downward mobility 6 1 Total workersb 129 68 Note: The measurement is rather subjective. A street peddler who has become a miner is considered to have ascended, as has a warehouseman who has become a production worker; a bricklayer who becomes a doorman is not regarded as having changed occupational level, nor a street peddler who becomes a "trader." Although the occupations of people at the various sample dates were looked at to get an idea of the occupational history of individuals, the figures in the table reflect changes in occupation between 1970 and 1980. a. Movement from professional to managerial status. b. Persons working in 1980. Source: FIEH and Fedesarrollo sample. CHANGES IN REAL INCOME IN CALI 67 Comparison of Income Data from the Survey and Wage Series Table 29 compares the income data from the sample with information from DANE on salaries and wages of manufacturing, construction, and agricultural workers in the Cauca Valley and in Cali. In 1970, the incomes of heads of households in the poor neighborhood were higher than those of agricultural daily workers, but lower than those of construction workers Table 29. Comparison of Average Nominal Monthl) Salaries and Wages in Selected Industries and Incomes of Heads of Selected Households in Cali, 1970-80 (nominal Colombian pesos) Sector 1970a 1974b 1976c 1980 Manufacturing industry (Cauca Valley) Executive and technical personnel 6,683 9,940 16,484 37,600d Clerks 2,650 4,242 5,876 13,400 Laborers 1,294 2,054 3,191 7,900 Apprentices 517 1,119 1,973 Construction (Cali)e 932 1,203 1,758 5,600f Agriculture (Cauca Valley) 435 - 1,650 4,3509 Sample (Cali) heads of household. High income 6,165 15,771 28,756 72,803 Low income 653 1,361 1,949 5,861 -Data not available. Note: The comparison is between the FIEH-Fedesarrollo sample of families with the information collected from DANE surveys. a. For industry, average monthly salaries and wages are for 1970; for construction, the average wage is for December 1971; for agTiculture, the average daily wage is for 1970 (the wage is for men in hot-climate areas, with no food provided). b. For industry, average monthly salaries and wages are for 1974; for construction, the average wage is for the first four months of 1974. c. For industry, average monthly salaries and wages are for 1976; for construction, the average wage is for the months of May, June, July, and August. d. The estimate is based on the data from the monthly manufacturing survey for March 1980. e. Average for masters, journeymen, and helpers. Calculated from DANE data using average wages in Cali for each category in December 1979, assigning them weights of 10 percent, 43 percent, and 47 percent, respectively, in the total, and applying this value to the index of construction labor costs in Cali. f. Average for March and April 1980. g. Daily agricultural wage (in hot climate, with no food provided), during the fourth quarter of 1979. Source: DANE, Annual Survey of Manufacturing (various issues) and Monthly Statistical Bulletin, FIEH-Fedesarrollo sample. 68 WINNERS AND LOSERS IN COLOMBIA'S GROWTH or industrial laborers. This was to be expected, and the comparison confirms the high quality of the survey information. The workers in the sample are concentrated in occupations in the urban informal sector, which should generate higher earnings than farm work to justify rural-urban migration; however, incomes are lower than those of occupations in the formal sector of the economy, such as industry or construction, although the latter is quite similar to urban informal employment. By the end of the decade, the incomes of the low-income sample were higher than those of construction workers. The incomes of the high-income sample were close to the average income declared by industrial executives and technicians in 1970. At the end of the decade, however, most of the highly skilled persons in the sample were above the levels of industrial executives, which placed the group at the top of the distribution. Also significant is the reduction in the dispersion of income among the working class. The differential between daily wages in agriculture and those in construction declined over the 1970s, as did the differential between rural wages and income earned in the urban informal sector (represented here by the low-income sample). The period also saw a closing of the gap between industrial laborers and informal sector workers.12 This reduction in income differentials within the working class indicates a decrease in the dualism in the economy, which in turn may be related to the lessening of protection for the formal sector of the urban economy. During the 1970s, the real incomes of the upper-class group rose at a rapid rate of 9 percent a year (see Table 30). Over that period, the real incomes of the heads of households from the poor neighborhood rose at an average annual rate of 3.5 percent, which was greater than the increase in the incomes of construction workers and manufacturing industry em- ployees (the real incomes of the latter did not improve), but less than the rise in agricultural wages. These results help to confirm the hypothesis that there was a redistribution of income toward the two extremes of the population during the decade. The incomes of the poorest groups-the landless campesinos and female urban workers in the informal sector-increased faster than per capita income, and at rates matching those for professionals and business- men. Consequently, industrial wages paid in the formal sector of the economy have lost ground in relation to wages in the informal sector or to wages of unskilled workers and campesinos. Heads of poor urban households, however, probably did not improve their relative position in the income distribution because their incomes grew, according to Table 30, at about the same rate as the per capita national income. CHANGES IN REAL INCOME IN CALI 69 Table 30. Comparison of Average Real Monthly Salaries and Wages in Selected Industries and Incomes of Heads of Selected Households in Cali, 1970-80 (constant 1970 Colombian pesos) Annual rate Sector 1970 1974 1976 1980 of increase Manufacturing industry Executive and technical personnel 7,035 5,616 6,290 6,530 -0.7 Clerical workers 2,790 2,397 2,240 2,325 -1.8 Laborers 1,360 1,087 1,130 1,270 -0.7 Apprentices 545 592 700 Construction 785 730 630 900 1.7 Agriculture 458 - 569 700 4.3 Sample of household heads High income 5,652 9,728 10,230 13,393 9.0 Low income 646 760 695 916 3.5 -Data not available. Note: The comparison is between the FIEH-Fedesarrollo sample of families with the information collected from DANE surveys. Source: Table 29, deflated by the consumer price index for Cali (1970=100). In absolute terms, although the average incomes of heads of lower-class households rose during the period, in 1980 they remained below the average for industrial laborers. In contrast, for the upper class represented in the sample, the average income in 1970 was lower than that of the executive and technical personnel in industry, but was twice as high in 1980.13 This means that in 1980 the groups of the sample remained at the extremes of the distribution. Conclusion The research conducted in Cali documents the movement of income levels and welfare indicators for two groups of families at the extremes of the income distribution in the 1970s. This exercise leads to several important conclusions. * During the 1970s, income was redistributed toward those at the top and the bottom of the income distribution. The real per capita incomes of both the rich and the poor families rose at an average rate exceeding that of the per capita national income. Similarly, the incomes of heads of 70 WINNERS AND LOSERS IN COLOMBIA'S GROWTH households in these groups rose during the period when real salaries and wages in the manufacturing industry were declining or growing very slowly. * The real family incomes of both the poor and the rich increased significantly. The rise in the incomes of the poor seems to be due primarily to increases in the average number of family members who work and in the eamings of the wife and secondary workers, whereas among the rich, the increase in the income of the head of household explains most of the rise in total income. * Single-parent families in both social classes have lower incomes than the average for all families in their class, but the income gains for single- parent poor families were substantial. This phenomenon probably helped to decrease the incidence of absolute poverty because many of the poorest families are of this nature. None of these trends is inconsistent with the evidence presented in Chapter 2 on wage trends. On the contrary, the data on the large gains in earnings of unskilled and informal-sector workers in a poor area of Cali are consistent with the evidence on real wage increases for agricultural workers since there is probably much mobility between the two sectors. However, the low growth rate of earnings of workers in the formal sector is clearly not a good proxy for the earnings of the poor or of workers in the informal sector during the 1970s. Even within the informal sector, the earnings of secondary workers grew faster than those of household heads. All evidence, then, points toward a decrease in the dispersion of labor earnings in the economy. The data from the Cali sample also suggest that the commonly available statistics on wages in large industries may be misleading and that the stagnation in earnings of this sector, which in fact represents the middle class, may be consistent with an improvement in the standard of living for the lower deciles of the income distribution. The evidence from the wage series together with the income data from the Cali sample indicate that during the 1970s there was a reduction in poverty and that the gap between the formal and informal sectors diminished. It may be that this happened to some extent at the expense of the middle class. This deterioration in the relative position of the middle class may have profound political implications. To what extent, for example, has this phenomenon affected the trends toward a decrease in voter participation, the creation of urban political fringe movements that do not reject violent tactics, and a general dissatisfaction among intellectuals with the per- CHANGES IN REAL INCOME IN CALI 71 formance of the country's economic and political institutions? These questions are difficult to answer here. In the next chapter, evidence from household surveys will be shown to confirm the general trends described in this chapter. It is, however, clear from a careful analysis that the income distribution in Colombia has not become more skewed during the country's successful economic develop- ment effort of the last fifteen years. This runs counter to the common impression, because the people who write about Colombian economic policy belong to precisely the class that has seen no clear improvement in its standard of living. Notes to Chapter 3 1. It is for this reason that occupational histories are regarded as unreliable. See Bernardo Kugler, Alvaro Reyes, and Maria Isabel de Gomez, Educacion y Mercado de Trabajo Urbano en Colombia: Una Comparacion entre Sectores Modemo y no Modemo, monograph 10 (Bogota: Corporacion Centro Regional de Poblacion, 1979). 2. Cali is Colombia's third largest city and the capital of the department of Valle del Cauca. 3. World Bank, World Development Report 1978 (New York: Oxford University Press, 1978). 4. Harrison Mckay and others, "Improving Cognitive Ability in Chronically Deprived Children," Science, vol. 200 (April 21, 1978). 5. The average family income of the wealthiest 10 percent in Cali in 1968 was Col$4,900 a month; see Philip Musgrove, Con.sumer Behavior in Latin America (Washington, D.C.: Brookings Institution, 1978). The 1968 figure is equivalent to Col$5,494 a month in 1970 when correction is made for increases in the cost of living over two years. This figure compares with an average family income of Col$7,831 for the high-income group in our sample for that year. The rich families were chosen from a group of families headed by professionals, who had babies bom between June and November 1967. 6. In the upper class, the average number of persons per family declined from 5.2 in 1970 to 4.9 in 1980; in the lower class, it increased from 7.6 to 8.5 persons. The average number of children under fifteen years of age fell by 17 percent in both cases, although the total number of children in poor families continued to be roughly twice that in rich families. These findings may indicate that in low-income families children remain at home for a longer period after they have reached working age; this in tum may be associated with the need for a larger number of persons working in order to augment household income. 7. In both social classes, the total number of females working has increased more than that of males. This is reflected in an increase both in the average number of females working per family and in the rate of participation of females (see Table 24). In the case of males, the average number of workers per family rose in the lower class but fell in the upper class, while the rte of male participation declined in both cases. 8. According to the data from the Social Securitv Institute, the average salary of females covered by the programs of that institution was 68 percent of the average salary of males in 1970. By 1978, the proportion had risen to 74 percent. 9. Information on durable goods is available only for 1976 and 1980. Furthermore, the data only allow comparison of the possession of refrigerators, radios, sewing machines, television sets, and record players. In 1980, questions were included about electric stoves, irons, blenders, and other appliances, but comparable information is not available for the earlier period. 72 WINNERS AND LOSERS IN COLOMBIA'S GROWTH 10. The value of one gallon of water to the user is greater than the cost charged per gallon by the public utility company. This is evidenced by the fact that in districts without a water supply, families purchase water from trucks at prices much higher than those charged by the water supply system. 11. As may be seen in Table 24, the share of the income of the head of household in total family income for rich families increased from 78 percent in 1970-71 to 88 percent in 1980, whereas among the poor that share was 88 percent in 1970-71 and fell to 59 percent in 1980. 12. It can be argued that it is not correct to compare changes in average wages through time with changes in the incomes of individuals, because the latter reflect increases in age and experience, which the average wage data do not reflect in a rapidly growing population. Experience probably explains why in 1980 the incomes of the household heads of rich families are above the average incomes of executives. Experience, on the other hand, affects much less the wages of rural laborers or low-skilled self-employed workers. For that reason, it does not seem inadequate to compare the average wages of unskilled workers with the income of the heads of poor households in our sample. 13. This shows the high slope of the earnings cutve of managerial personnel between the ages of thirty-seven and forty-seven. Casual observation suggests that this type of eamings curve, in fact, reflects the job history of upper-class professionals in Colombia. 4 The Evidence from Household Surveys ESTIMATES OF INCOME DISTRIBUTION in most countries generally are based on data from household surveys. Income data from these surveys are often unreliable, however, and it is particularly difficult to compare distributions based on surveys carried out at different times and with different methodologies. In the case of Colombia, a simple comparison of concentration ratios derived from different household surveys is not valid because of large variations in the quality and coverage of the surveys. This chapter presents the results of different household surveys and attempts to derive a distribution for 1971-72 comparable to that calculated by the author in 1964. In addition, an effort is made to compare income distributions from surveys of similar quality at the start and at the end of the decade of the 1970s. Between 1970 and 1980, DANE carried out twenty-seven household surveys (EH). At the start of the decade, some of these were expenditure surveys, and consumption and income information was gathered with some care.' Other surveys dealt primarily with the labor force; their questions concerning income were fewer and more general, and for that reason they underestimated income more than the expenditure surveys.2 Even the higher-quality household surveys appear to underestimate income significantly if one compares the resulting income distributions with data from national accounts, which in turn probably underestimate the true level of per capita income in Colombia. The more disaggregated income data from the surveys indicate that the underestimation of rural income is quite high and that the information from the urban areas covers primarily labor income. Household surveys tend not to capture income from capital and land in the countryside or part of capital income in urban areas. Because of the large proportion of owner-operated farms in Colombia, the 73 74 WINNERS AND LOSERS IN COLOMBIA'S GROWTH degree of underestimation of rural incomes in household surveys is substantial. Table 31 shows estimates of the degree of underreporting of income in five household surveys and in the population census. Labor force surveys clearly have a higher degree of underestimation than expenditure surveys such as that of 1971, and the census contains the least reliable data. Furthermore, a survey such as EH-4 (1971) certainly cannot be compared with EH-19 (1978). The degree of underreporting of rural incomes also suggests the need to generate a rural income distribution with statistical material other than that obtained from household surveys. In addition to the official household surveys, two private surveys were also carried out during the decade. The first was a national income and expenditure survey in 1974; the second was a 1977 survey carried out in four cities by the Centro de Estudios para el Desarrollo Economico (CEDE) of the Universidad de los Andes for a research project on poverty and employment.3 The information from the latter probably gives quite accurate data on labor income of the poor in the largest cities, but the data on the incomes of the upper five deciles of the distribution seem to have serious errors of underestimation. Finally, because the rural income data derived from household surveys were unreliable, an effort was made to find agricultural production and income information from production surveys. Although there have been various attempts to measure income and production at the farm level, these studies are often confined to small geographical areas. The one national survey carried out by DANE had so many problems of coverage and inconsistencies that it was never published.4 Furthermore, in the 1970s, DANE did not carry out any agricultural production surveys such as those on which Berry and Padilla based the rural income distribution of 1960.5 Nevertheless, an interesting agricultural production survey was carried out by the Colombian Central Bank in 1972 to determine how credit was distributed in the countryside and what relation existed between credit availability and productivity. This survey is used here to estimate an agricultural income distribution similar to that used in the Urrutia-Berry distribution of income for 1964.6 Changes in Income Distribution between 1964 and 1972 The earliest reliable income distribution for Colombia is for 1964. The methodology used to arrive at that distribution is described in detail in the 1970 work by Berry and Padilla and in the 1975 study by Urrutia and Berry. Table 31. Percentage of Personal Income Accounted for in Various Household Surveys, 1970-78 EH-1 EH-3 EH 4b EH-5 Census Selowsky EH-19 Incomea June 1970 April 1971 July 1971 November 1971 1973 1974 April 1978 Rural income 48 36 81 40 20 50 36 Urban income - - 77 - - - 70 Total income - - 78 51 51.4 - 65 -Data not available. Note: Ell stands for household surveys. a. Total personal income was derived from national accounts. b. Expenditure surveys. Source: For EH-4 and EH-19, special tabulations of the DANE surveys. For EH-1, Eii-3, EH-5, and Census, the source is DANE; 1974 data is from Marcelo Selowsky, Who Benefits from Govemment Expenditure? A Case Study of Colombia (New York: Oxford Universiry Press, 1979). 76 WINNERS AND LOSERS IN COLOMBIA'S GROWTH This section will first summarize the statistical sources used to generate the 1964 distribution as a basis for understanding the methodology used for the 1972 estimate. Because rural household surveys were unavailable in the 1960s, the rural income distribution was calculated from four sources: the population census, data on value-added for farms of different sizes based on an agricultural production survey, information on wages of agricultural laborers, and data on the distribution of farms by size based on the 1960 census of agricultural production units. In other words, the agricultural income distribution reflects closely the distribution of land, corrected for the difference in average value-added by size of farms. To this distribution were added the distribution of income of landless laborers and the wage income of owners of small plots who work part of the year for wages. To obtain a rural income distribution, a special estimate of nonagricultural income in rural areas was added to the agricultural distribution.7 The urban income distribution for 1964 was derived primarily from the urban labor force surveys of CEDE, particularly those published in a 1969 study by Isaza and Ortega, but also those carried out in small cities in 1963V The incomes of agricultural workers living in the cities were subtracted from the urban distribution because this income was presumably captured at the farm level. Various additional adjustments included the imputation of income paid in kind for workers in domestic services and the increase of labor income by the equivalent of minimum legal fringe benefits, a type of income not clearly identified in the labor force surveys. The methodology used for obtaining a rural income distribution in 1964 clearly covers income from capital and land. It may, in fact, overestimate the concentration of income in agriculture by assigning all of the income from a large farm to one owner, even though large farms often support wholly or in part the families of various brothers and sisters. Furthermore, the value-added method of estimating net income probably leaves out some minor costs of production that may add up to significant amounts. For all these reasons, not only is the concentration of income calculated for the rural sector of Colombia in 1964 higher than any of the estimates that are based solely on household surveys, but also the Gini coefficient is higher than that found for the rural sector of many other countries.9 The main reason for the higher rural inequality found for 1964 is that the methodology used probably covers most capital income and rents from land, whereas most household surveys seriously underestimate these types of income. To calculate a comparable agricultural income distribution for 1972, we used the value-added data by farm size derived from a Banco de la EVIDENCE FROM HOUSEHOLD SURVEYS 77 Republica survey on agricultural credit.'" Information from this survey, in combination with data from the 1970-71 agricultural census, data on daily wages of agricultural workers, and labor force data from the 1973 population census, allowed us to calculate an agricultural income distribu- tion with a similar methodology to that used by Berry to obtain the 1964 distribution." Table 32 shows the agricultural income distribution for 1972. To obtain the rural income distribution, the income distribution of nonagricultural rural workers based on the DANE July 1971 household survey (EH-4) was added to the figures in Table 32, after adjusting for the difference in dates of the 1972 Banco survey and the EH-4 household survey. This rural distribution for 1972 is summarized in Table 33.1 Like the 1964 distribution, the one for 1972 probably overestimates to some extent the concentration of rural incomes because of an exaggeration of capital income in large farms and because all the capital income of large farms is assigned to only one person. In addition, in 1972 large farms seem to be overrepresented in the sample used. The Gini coefficient for the rural distribution of 1972 is 0.63, while that for 1964 was 0.58. As Table 34 shows, however, our estimate shows much more concentration of income than is shown by the Selowsky survey of 1974 or the DANE expenditure survey of 1971. Clearly, these last two estimates do not cover a significant part of capital income and may underestimate significantly the income of farmers who derived their income from their own land. The urban income distribution for 1972 was also calculated with the methodology used in 1964. The EH-4 (1971) household expenditure survey was used to derive the urban distribution for the economically active population. Because the EH-4 data are for family income, it was necessary to use the original data to obtain an urban income distribution for each economically active person. Persons deriving their income from agriculture but living in cities were taken out of this distribution because presumably their income is recorded in the rural distribution. The resulting distribution was then adjusted for population growth, changes in income per capita, and prices for 1971-72 to arrive at a 1972 urban income distribution, which, when added to Table 33, produced the national income distribution found in Table 35. The urban distribution for 1972 has a Gini coefficient of 0.54 compared with 0.48 obtained in 1964. Although for 1972 no adjustment was made for the incomes paid in kind for domestic service, and such an adjustment would decrease income dispersion, it still appears that in the urban sector the income distribution deteriorated between 1964 and 1972. Table 36 Table 32. Distribution of Personal Income Obtained from Agricultural Work, 1972 Total income Accumulated Accumulated Annual income Number of persons Percentage (millions of Percentage percentage percentage (Colombian pesos) (thousands) of persons Colombian pesos) of income of persons of income 0-2,000 46.5 2.24 80 0.16 2.24 0.16 2,001-4,000 133.6 6.43 434 0.85 8.67 1.01 4,001-6,000 170.8 8.22 906 1.77 16.89 2.78 6,001-8,000 248.2 11.95 1,791 3.50 28.84 6.28 8,001-10,000 538.4 25.91 4,883 9.55 54.75 15.83 10,001-12,000 195.7 9.42 2,137 4.18 64.17 20.01 12,001-14,000 78.3 3.77 1,053 2.05 67.94 22.06 14,001-16,000 182.8 8.80 2,685 5.25 76.74 27.31 16,001-22,000 123.4 5.94 2,414 4.72 82.68 32.03 22,001-28,000 114.0 5.49 2,948 5.77 88.17 37.80 28,001-36,000 47.8 2.30 1,519 2.97 90.47 40.77 36,001-48,000 64.3 3.09 2,745 5.37 93.56 46.14 48,001-60,000 30.9 1.49 1,640 3.21 95.05 49.35 60,001-72,000 12.7 0.61 829 1.62 95.66 50.97 72,001-84,000 12.9 0.62 985 1.93 96.28 52.90 84,001-96,000 4.9 0.24 441 0.86 96.52 53.76 96,001-108,000 13.0 0.62 1,317 2.58 97.14 56.34 108,001-120,000 1.7 0.08 204 0.40 97.22 56.74 120,001-144,000 8.7 0.42 1,056 2.06 97.64 58.80 144,001-180,000 7.8 0.38 1,297 2.54 98.02 61.34 180,001-240,000 13.3 0.64 2,812 5.50 98.66 66.84 240,001-360,000 14.9 0.72 4,272 8.35 99.38 75.19 360,001-480,000 0.6 0.03 232 0.45 99.41 75.64 480,001-600,000 4.6 0.22 2,393 4.68 99.63 80.32 600,001-720,000 2.2 0.10 1,609 3.15 99.73 83.47 720,001-840,000 1.9 0.09 1,396 2.73 99.82 86.20 840,001-960,000 0.7 0.03 607 1.19 99.85 87.39 960,001 or more 3.1 0.15 6,452 12.61 100.00 100.00 Total 2,077.7 100.0 51,135 100.0 Note: The figures do not include family workers. Source: Clara Elsa de Sandoval and Miguel Urrutia, "Distribucion del Ingreso Proveniente de la Actividad Agropecuaria en Colombia" (Bogota: Fedesarrollo, November 1980; processed). Table 33. Rural Income Distribution, 1972 Accumulated Total income Accumulated Annual income Number of persons Percentage percentage (millions of Percentage percentage (Colombian pesos) (thousands) of persons of persons Colombian pesos) of income of income 0-4,000 263.0 10.65 10.65 654 .i5 1.15 4,001-6,000 233.2 9.44 20.09 1,208 2.13 3.28 6,001-10,000 838.4 33.94 54.03 7,122 12.56 15.84 10,001-12,000 195.7 7.92 61.95 2,137 3.77 19.61 12,001-14,000 124.5 5.04 66.99 1,601 2.82 22.43 14,001-16,000 229.5 9.29 76.28 3,412 6.02 28.45 16,001-22,000 144.9 5.87 82.15 2,827 4.99 33.44 22,001-28,000 157.4 6.37 88.52 3,993 7.04 40.48 28,001-36,000 61.4 2.49 91.01 1,918 3.38 43.86 36,001-48,000 79.6 3.22 94.23 3,366 5.94 49.80 48,001-60,000 33.5 1.36 95.59 1,778 3.14 52.94 60,001-72,000 12.7 0.51 96.10 829 1.46 54.44 72,001-84,000 16.3 0.66 96.76 1,247 2.20 56.60 84,001-96,000 5.3 0.21 96.97 477 0.84 57.44 96,001-120,000 15.3 0.62 97.59 1,579 2.79 60.23 120,001-144,000 9.1 0.37 97.96 1,105 1.95 62.18 144,001-180,000 7.8 0.32 98.28 1,297 2.29 64.47 180,001-240,000 14.8 0.60 98.88 3.174 5.60 70.07 240,001-360,000 14.9 0.60 99.48 4,272 7.54 77.61 360,001-480,000 0.6 0.02 99.50 232 0.41 78.02 480,001-600,000 4.6 0.19 99.69 2,393 4.22 82.24 OO 600,001-720,000 2.2 0.09 99.78 1,609 2.84 85.08 720,001-840,000 1.9 0.08 99.86 1,396 2.46 87.54 840,001-960,000 0.7 0.03 99.89 607 1.07 88.61 960,001 or more 3.1 0.13 100.02 6,452 11.38 99.99 Total 2,470.4 100.02 56,683 99.99 Source: Total agricultural distribution obtained from the Banco de la Republica farm sample for 1972, plus the income distribution of nonagricultural workers obtained from EH.4 (1971). This last distribution was transformed to 1972 pesos using a consumer price index, a population growth rate, and an estimate of growth in income per capita between the two dates. 82 WINNERS AND LOSERS IN COLOMBIA'S GROWTH Table 34. Various Rural Income Distribution-Studies, 1964-74 (accumulated percentage of income) Accumulated percentage of economically active Urrutia- EH-4 Urrutia- Deciles by persons who Berry EH-4 adjustmenta Sandoval per capita Selowsky receive income 1964 1971 1971 1972 family income 1974b 10 0.9 1.5 1.1 1.2 10 3.0 20 4.2 3.8 3.4 3.6 20 7.6 30 8.5 8.0 7.0 5.5 30 13.8 40 13.5 13.2 11.5 9.5 40 20.9 50 19.0 19.0 16.0 15.0 50 29.8 60 26.0 26.0 22.0 18.5 60 38.6 70 34.0 35.0 29.5 25.0 70 47.5 80 45.0 46.5 38.5 32.5 80 59.4 90 59.5 60.0 52.0 43.0 90 73.5 95 - 71.5 63.0 53.0 95 81.9 100 100.0 100.0 100.0 100.0 100 100.0 Gini coefficient 0.58 0.49 0.56 0.63 0.32 a. The adjustment consisted of assigning all of the 19 percent calculated underestimation of rural incomes to the top 10 percent of the population. This is an extreme assumption, but it is based on the supposition that the household survey does not capture capital income. b. Accumulated by deciles of per capita family income. Source: For EH-4, DANE; Miguel Urrutia and Albert Berry, La Distribucion del Ingreso en Colombia (Medellin: Editorial la Carreta, 1975); de Sandoval and Urrutia, "Distribucion del Ingreso Proveniente"; Selowsky, Who Benefits from Govemment Expenditure? shows that the seventh to the ninth deciles of the population-that is, the middle class-and the tenth decile improved their positions at the expense of the lower six deciles. The national income distribution changed little between 1964 and 1972 (see Table 37). If anything, the top and the bottom of the distribution gained a little, but the changes are insignificant This is reflected in similar Gini coefficients-0.57 for 1964 and 0.58 for 1972. Since the 1972 agricultural distribution may overestimate the income from large farms, it is possible to state with some confidence that the distribution of income did not become more unequal between 1964 and 1972. By themselves, however, both the rural and the urban distributions were less equal in 1972, so that any partial analysis would suggest a concentration of income. But a closing of the gap between rural and urban incomes avoided a worsening in the national income distribution. Table 35. National Income Distribution, 1972 Accumulated Total income Accumulated Annual income Number of persons Percentage percentage (millions of Percentage percentage (Colombian pesos) (thousands) of persons of persons Colombian pesos) of income of income 0-4,000 567.4 10.32 10.32 1,128 0.91 0.91 4,001-6,000 580.3 10.55 20.87 2,640 2.12 3.03 6,001-10,000 1,162.8 21.15 42.02 9,410 7.55 10.58 10,001-14,000 623.4 11.34 53.36 6,706 5.38 15.96 14,001-16,000 522.2 9.50 62.86 7,147 5.73 21.69 16,001-22,000 564.1 10.26 73.12 10,030 8.05 29.74 22,001-36,000 610.7 11.11 84.23 15,341 12.31 42.05 36,001-48,000 337.6 6.14 90.37 12,076 9.69 51.74 48,001-60,000 127.9 2.33 92.70 5,972 4.79 56.53 60,001-72,000 67.8 1.23 93.93 3,718 2.98 59.51 72,001-84,000 88.0 1.60 95.53 5,714 4.58 64.09 84,001-96,000 29.8 0.54 96.07 2,219 1.78 65.87 96,001-120,000 58.9 1.07 97.14 5,185 4.16 70.03 120,001-144,000 36.4 0.66 97.80 3,917 3.14 73.17 144,001-180,000 42.6 0.77 98.57 5,697 4.57 77.74 180,001-240,000 20.5 0.37 98.94 4,120 3.31 81.05 240,001 or more 58.3 1.06 100.00 23,629 18.96 100.01 Total 5,498.7 100.0 124,648 100.01 Source: De Sandoval and Urrutia, "Distribucion del Ingreso Proveniente." 84 WINNERS AND LOSERS IN COLOMBIA'S GROWTH Table 36. Urban Income Distribution of the Employed Population Excluding Farmers Living in Cities, 1964-1971 (percent of total income in decile) Decile of economically active population 1964 1971 1 0.9 1.1 2 3.3 1.4 3 4.3 2.8 4 5.0 3.7 5 5.5 4.5 6 7.0 6.5 7 8.0 8.5 8 11.0 11.0 9 14.5 17.0 10 40.5 43.0 Source: Urrutia and Berry, La Distribucion del Ingreso en Colombia; and de Sandoval and Urrutia, "Distribucion del Ingreso Proviente." Table 37. National Income Distribution, 1964 and 1972 (percent of total income in decile) Decile of economically active population 1964 1972 1 1.1 0.9 2 1.4 2.1 3 2.8 3.0 4 3.7 3.5 5 4.5 5.0 6 5.5 5.5 7 8.0 7.0 8 10.0 10.5 9 15.0 13.5 10 48.0 49.0 Source: Urrutia and Berry, La Distnbucion del Ingreso en Colombia; and de Sandoval and Urrutia, "Distribucion del Ingreso Proveniente." EVIDENCE FROM HOUSEHOLD SURVEYS 85 Table 36 is consistent with the wage series analyzed in Chapter 2, which showed that the 1960s had been a period of rapid gains in real earnings for the middle class. This is reflected in the increasing share of total urban income received by the top four deciles of the urban distribution. Table 38 on rural income distribution is also consistent with the observed increase in agricultural wages, because in absolute terms the income of the lower deciles increased significantly, although less rapidly than that of the wealthy landowners and rural entrepreneurs. The rapid rate of growth in Colombia's rural areas, where most poverty was concentrated in 1964, explains why the national income distribution did not deteriorate despite growing inequality in both the rural and the urban areas. Alternative Comparison of Changes in Income Distribution in the 1960s An alternative to the 1964-72 comparison made earlier is to compare the EH-4 national expenditure survey, which is of high quality, with the 1964 income distribution. Table 39 shows that the Gini coefficients in these two distributions are similar, confirming the impression that the distribution of income did not become more unequal. However, the methodology for obtaining rural incomes is quite different in the two distributions; the 1964 methodology produces higher incomes for large farmers. Table 39, however, does show that the middle class received a greater share of total income in 1971 than in 1964. This had been suggested by the wage series presented in Chapter 2. The table also supports the results of the 1964-72 comparison discussed earlier, in the sense that it shows no deterioration in the distribution of income over the decade of the 1960s. Changes in Income Distribution between 1971 and 1978 No survey of agricultural production is available for the end of the 1970s and, therefore, the 1964 methodology cannot be used. The only comparison possible is between the distributions derived from household surveys. Because the only national survey at the end of the decade is DANE's EH-19 for April 1978, that survey must be compared with one of similar quality at the beginning of the decade. Table 38. Changes in Agricultural Income by Decile, 1960-72 Berry 1960 1972 estimate Increase in annual Decile of economically Percentage Average income Percentage Average income Average income at per capita income active population of income (1960 Colombian pesos) of income (1972 Colombian pesos) constant 1960 prices (percent)a 1 2.24 865 1.5 3,691 1,079 1.9 2 2.87 1,110 2.0 4,922 1,439 2.2 3 3.34 1,290 3.0 7,383 2,259 4.3 4 3.73 1,440 3.2 7,876 2,303 3.9 5 4.21 1,625 3.8 9,351 2,734 4.4 co 6 4.68 1,807 4.5 11,073 3,238 4.9 7 5.78 2,232 5.0 12,304 3,598 4.0 8 7.90 3,060 7.0 17,225 5,036 4.2 9 12.77 4,940 10.8 26,576 7,771 3.8 10 52.48 20,270 59.2 145,697 42,601 6.4 Total 100.0 3,830 100.0 24,611 7,196 5.4 a. Annual geometric growth rate, 1960-72. Source: Albert Berry and Alfonso Padilla, "La Disrribucion de Ingresos Provenientes de la Agricultura en Colombia, 1960" Universidad Nacional, Documento de Trabajo (Bogota: CID, January/March 1970); and de Sandoval and Urrutia, "Distribucion del Ingreso Proveniente." EVIDENCE FROM HOUSEHOLD SURVEYS 87 Table 39. Comparison of Two Income Distribution Estimates, 1964 and 1971 1964 1971 Decile of the Percentage Percentage Percentage Percentage economically of rural of total of rural of total active population income income income income 1 1.4 1.1 1.5 1.1 2 3.1 1.4 2.0 1.4 3 3.6 2.8 3.7 2.8 4 3.9 3.7 5.8 3.7 5 4.5 4.5 6.0 4.5 6 5.5 5.5 7.5 6.5 7 6.0 8.0 8.0 8.5 8 8.0 10.0 12.0 11.0 9 13.0 15.0 15.5 17.0 10 51.0 48.0 38.0 43.5 Gini coefficient 0.55 0.57 0.49 0.57 Source: Urrutia and Berry, La Distribucion del Ingreso en Colombia; and DANE (EH-4). Table 40. Income Distribution from Labor Force Surveys, 1971 and 1978 (accumulated percent) EH.3, April 1971 EH-5, November 1971 EH-19, April 1978a Employed Employed Employed population Income population Income population Income 36.6 7.0 36.9 7.5 - - 66.0 23.9 65.7 24.9 20 4.8 80.3 37.6 79.0 38.3 40 15.0 90.8 52.9 90.8 53.0 50 21.4 100.0 100.0 100.0 100.0 60 28.8 80 46.3 90 60.9 100 100.0 -Data not available. a. DANE did not publish the results of EH-19. The Corporacion Centro Regional de Poblacion tabulated this income distribution for Fedesarrollo, plus distributions by household and household income per capita for urban and rural areas as well as for Bogota and the whole nation. Source: DANE. 88 WINNERS AND LOSERS IN COLOMBIA'S GROWTH EH-3 (April 1971) and EH-5 (November 1971) are labor force surveys similar to EH -19. Table 40 shows that in the first two surveys, the top three deciles of the distribution have a similar share of total income. Table 40 compares the three distributions and suggests that income concentration decreased between 1971 and 1978. The lowest four income deciles seem to have improved their participation in the national income as was suggested in the study of the rich and the poor in Cali. The participation of the upper 10 percent of the population, on the contrary, decreased according to the surveys. Because the 1978 survey appears to cover a greater proportion of the national income, one would expect the participation of the tenth decile to be higher in 1978 than in 1971, if one assumes that underestimation is higher in the upper deciles."3 On the contrary, the degree of underestimation in general decreased from 1971 to 1978; therefore, the decrease in participation of the top two deciles strongly supports the hypothesis that income distribution improved in this period. Another possible comparison is between the Selowsky survey of 1974 and the 1978 distribution. Table 41 shows this comparison. Once again, the data are not really comparable, but, nevertheless, they indicate that income distribution has improved. Between 1970 and 1974, urban industrial wages deteriorated; therefore, the urban income distributions shown in Table 41 for the early 1970s show no improvement. In the second part of the decade, when real industrial wages improved, the urban distribution also improved. Even though the household surveys point to an improvement in the income distribution during the 1970s, their large underestimation of rural incomes casts some doubt on their overall reliability. This suggests that one should focus attention solely on the results of the surveys for the urban areas. Table 42, which displays the urban income distributions in EH-4 and EH-19, once again suggests an improvement: the income of the top urban quintile decreased from 60 percent to 53.7 percent in 1978, while that of the bottom two quintiles increased from 9 percent in 1971 to 15 percent of total income. The participation of the fourth quintile, which represents the middle class, decreased from 19.5 percent to 17.5 percent. Table 42, therefore, is consistent with the findings reported in Chapters 2 and 3 on the basis of quite different sources of data. One last survey-EH-21, a joint DANE and World Bank survey of December 1978-offers further evidence of an improvement in the urban income distribution during the decade. This survey of the income distribution for Bogota is of very high quality and quite comparable with EH-4 (1971). Table 43 compares the data for Bogota in these two surveys EVIDENCE FROM HOUSEHOLD SURVEYS 89 Table 41. Gini Coefficients for Various Income Distributions, 1964-78 Distribution Urban Rural Total Urrutia-Berry (1964)a 0.48 0.55 0.57 DANE EH-4 (1971)a 0.54 0.49 0.57 Sandoval-Urrutia (1972)' 0.54 0.62 0.58 Selowsky (1974)b 0.54 0.42 0.50 DANE EH - 19 (1978) c 0.45 0.44 0.47 a. Distribution of income among economically active persons. b. Distribution of income among households. c. Distribution of income among employed persons. Source: Tables 33, 34, 35, 36, 37, and 39. Table 42. Urban Income Distribution by Quintiles of Family Income per Capita, 1971 and 1978 (percent of total income) Quintiles of family EH-4 EH-19 income per capita 1971 1978 1 2.5 4.8 2 6.5 10.2 3 11.0 13.8 4 19.5 17.5 5 60.0 53.7 Source: DANE. Table 43. Income Distribution among Income Eamers in Bogota, 1971-78 (percent of total income in decile) Decile of the economically EH-4 EH-19 EH-21 active population 1971 April 1978 December 1978 1 1.6 1.8 2.74 2 1.5 3.5 3.42 3 2.4 4.4 3.73 4 3.0 5.0 4.21 5 4.0 5.5 5.14 6 6.0 6.6 6.20 7 8.5 7.6 7.76 8 12.0 9.6 11.21 9 18.0 14.6 17.61 10 44.0 41.4 37.98 Source: DANE. 90 WINNERS AND LOSERS IN COLOMBIA'S GROWTH and shows a considerable improvement in income distribution. Again, the first five deciles increase their participation, and both the middle class and the upper class (tenth decile) register relative losses. A 1980 study by Mohan, Garcia, and Wagner estimates that EH-21 covers as much as 86.4 percent of income generated in Bogota.14 Therefore, the income distribu- tion for December 1978, because of its high level of coverage, should have a high income participation of the upper deciles, because high income coverage in a survey usually implies little underestimation of high incomes. For that reason, the improvement in income distribution shown in Table 43 cannot be ascribed to changes in the quality of the survey data. In summary, the data from all sources point to an improvement in the income distribution between 1971 and 1978. Conclusion Comparing income distributions derived from different household surveys is always risky because of differences in coverage and in the quality of the income data. Nevertheless, an analysis of several such distributions carefully controlled in terms of similar data and similar methodologies does not support the common belief that income has become more con- centrated in Colombia in the last fifteen years. On the contrary, the distribution remained constant in the 1960s and improved in the 1970s. Many phenomena contributed to this result. Unquestionably, the improvement in the income of rural laborers and small farmers avoided the deterioration of the distribution in the 1960s and helped to diminish income dispersion in the 1970s. Also, government policy in the 1970s consciously attempted to improve distribution. Interestingly, the greatest opposition to the development model adopted by the government in the 1970s came from the employer federations, which were also the groups that promoted the belief that income concentration increased in the decade. Finally, the household survey data are consistent with the wage series by occupation analyzed in Chapter 2 and the analysis of family incomes in Cali described in Chapter 3. All the evidence suggests that the middle class did well in the 1960s and lost ground in the 1970s and that the poorest families improved their standards of living, particularly in the latter half of the 1970s. Notes to Chapter 4 1. These were EH-2 (1970), which was never published; EH-4 (1971); and EH-6 (1972). 2. Only EH-1 (1970), EH-3 (1971), EH-5 (1971), and EH-19 (1978) were national surveys, and EVIDENCE FROM HOUSEHOLD SURVEYS 91 only these have rural information. The other surveys have covered four cities in some cases and seven cities in others. 3. See Marcelo Selowsky, Who Benefits from Government Expenditure? A Case Study of Colombia (New York: Oxford University Press, 1979); and Ulpiano Ayala and Nohra Rey de Marulanda, Empleo y Pobreza (Bogota: CEDE, Universidad de los Andes, July 1978). 4. This was the 1972 Oficina de Planeacion del Sector Agropecuario (oPsA)-United States Agency for International Development (USAID), Muestra de Ampliacion of the 1970 agricultural census. 5. Albert Berry and Alfonso Padilla, "La Distribucion de Ingresos Provenientes de la Agricultura en Colombia-1960," Universidad Nacional, Documento de Trabajo (Bogota: CID, January/March 1970). 6. Miguel Urrutia and Albert Berry, La Distribucion del Ingreso en Colombia (Medellin: Editorial la Carreta, 1975). 7. This involved calculating the income of rural artisans in each department and assigning incomes to the other nonagricultural rural workers who appeared in the 1964 population census. These incomes were derived from a rural survey found in Rafael Prieto, Bill Hanneson, and Marco Reyes. Estudio Agronomico de la Hoya del Rio Suarez (Bogota: CEDE, 1965). 8. Rafael Isaza and Francisco Ortega, Ensuestas Urbanas de Empleo y Desempleo (Bogota: CEDE, 1968). 9. The Gini coefficient measures the degree of concentration of an income distribution. The coefficient is I when all income is controlled by one person, and 0 when everyone has the same income. For a simple explanation of the Gini coefficient, see Urrutia and Berry, La Distribucion, pp. 53-55. 10. Rafael Prieto and others, Fuentes y Usos de Recursos Financieros en el Sector Agropectario de Colombia (Bogota: Banco de la Republica, 1976). 11. For 1972, the same agricultural wage series as that used by Berry in 1960 was utilized. Therefore, the possible upward bias that would result from using the new agricultural wage series does not arise. 12. For details of how this rural income distribution for 1972 was calculated. see Clara Elsa de Sandoval and Miguel Urrutia, "Distribucion del Ingreso Proveniente de la Actividad Agropecuaria en Colombia" (Bogota: Fedesarrollo, November 1980; processed). The distribution probably overestimates the concentration of rural incomes because of an overrepresentation of large farms. This problem probably explains why the methodology produces an estimate of rural income 10 percent higher than that obtained from national income data. 13. Table 31 shows that EH-19 accounts for 56 percent of rural income, while EH-3 and EH-5 account for 36 and 40 percent of rural income. The earlier surveys therefore underestimate income to a greater extent. 14. Rakesh Mohan, M. W. Wagner, and Jorge Garcia, Measuring Urban Malnutrition and Poverty: A Case Study of Bogota and Cali, Colombia, World Bank Staff Working Paper no. 447 (Washington, D.C., 1981). 5 Changes in Colombia's Urban Poverty OVER THE YEARS, there has been a gradual decrease in rural poverty in Colombia. The real wages of rural laborers have increased, and there has been some real gain in the incomes of the rural poor between 1964 and 1972. Yet, Colombia is fast becoming an urban nation.' Therefore, it is important to determine what has happened to the urban poor in the recent past because, in the future, urban poverty could be a more serious problem than rural poverty. Although the growth of the Colombian economy has been characterized as satisfactory, it has not been particularly impressive; thus, any worsening of the income distribution could possibly mean increases in the absolute levels of urban poverty.2 In countries with much more rapid growth, such as Brazil in the 1960s, a worsening of income distribution is less likely to increase poverty levels. Although real per capita income rose by 23.1 percent between 1958 and 1968 and by 58.3 percent between 1968 and 1978, an acceleration of inflation between 1970 and 1974 seems to have led to decreases in real urban wages. This phenomenon has led many observers in Colombia to postulate that poverty increased in the 1970s. This chapter attempts to identify the urban poor and to ascertain the extent to which urban poverty has increased or decreased. Earlier Studies on Poverty Levels Any study of poverty comprises several stages. The first is to define the poor. One approach is to define poor families as those whose level of welfare does not meet a set of absolute standards established more or less 93 94 WINNERS AND LOSERS IN COLOMBIA'S GROWTH arbitrarily. Another approach is to consider poor families as those whose welfare does not meet a set of relative conditions, also established arbitrarily, in the context of the population analyzed. Absolute standards are usually established on the basis of family expenditures on food. This is so because specifying nutritional standards is easier than specifying minimum standards for housing and clothing, not to mention those for other components of expenditure whose behavior is less stable and systematic. Table 44 shows various standards proposed in terms of expenditures for food, expressed in 1978 Colombian pesos. The differences are quite large: the highest estimate is five times greater than the lowest. To frame these standards within the context of the Colombian economy, it is useful to compare them with data on average income, average food expenditure, and the minimum wage. This is done in columns 2, 3, and 4 of the table, where the proposed standards are divided by the indicators just mentioned. These results will be used to determine the absolute standard that will serve as the basis for analyzing the characteristics and trends of urban poverty. Relative standards of poverty are no less arbitrary because they can hardly be based on assessments of basic needs, which are usually specified in absolute terms. The degree of arbitrariness is reflected in the wide range of income-distribution percentiles within which various proposed standards fall. In some academic studies, for example, only the first two deciles are identified as the poor. The development plan "Para Cerrar la Brecha" and the report Hacia el Pleno Empleo by the International Labour Organisation (ILO) propose that economic policy should favor the first five deciles.3 In the 1984 study by Mohan and Hartline, the poor are those making up the first three deciles.4 The income distribution for 1964 indicates that those at the upper limit of the fifth decile have an income approximately four times greater than those at the upper limit of the first decile. This ambiguity naturally reflects the fact that poverty is not a binary variable but rather a problem of degree. Previous studies have offered various estimates of absolute poverty, which is usually expressed as the percentage of persons living below an arbitrary poverty line. Using criterion A of Table 44, Bourguignon found that 59 percent of Colombian families were poor in 1974.5 Applying alternative C, Altimir calculated that 38 percent of urban families and 54 percent of rural families were poor in 1970; this gave a national average of 45 percent.6 According to Mohan and Hartline, alternative D would place 30 percent of the families of Bogota below the poverty line in 1977.7 These findings obviously are not comparable because they are based on different estimates of minimum food expenditures. CHANGES IN COLOMBIA'S URBAN POVERTY 95 Table 44. Alternative Standards Proposed for Minimum Monthly Food Expenditure per Adult Person (1978 Colombian pesos) Normative monthly Proportion of food expenditure Proportion of average food Proportion of (1978 Colombian pesos) average income expenditure minimum wage Alternative (1) (2) (3) (4) A 1,705 0.72 2.36 0.66 B 822 0.35 1.14 0.32 C 696 0.29 0.96 0.27 D 535 0.22 0.74 0.21 E 341 0.14 0.47 0.13 Source: For column 1: Altemative A is taken from DANE (1973). Altemative B is taken from tabulation prepared by Aquiles Arellano for a study on family poverty using figures from the Family Budget Survey conducted by the Centro de Estudios para el Desarrollo Economico (CEDE) of the Universidad de los Andes in 1967-1968. Altemative C is the estimate given in Oscar Altimir, The Extent of PovertN in Larin America: A Summarv (Santiago, 1978; processed), p. 24. According to that estimate, the annual minimum diet would have cost US$85 in 1970. This figure was converted to pesos by taking the average buying quotation for the exchange certificate in that year (Col$18.45). Altemative D was calculated by Jorge Garcia, and also is mentioned in Rakesh Mohan and Nancy Hartline, The Poor of Bogota: Who They Are, What They Eam, Where They Live, World Bank Staff Working Paper no. 635 (Washington, D.C., 1984), p. 6. Alternative E results from a calculation of the minimum food expenditure that would purchase a basket that meets the nutritional standards of the Instituto Colombiana de Bienestar Familiar. The calculation gives weight to the original estimates for the seven cities (Bogota, Cali, Barranquilla, Medellin, Bucaramanga, Pasto, and Manizales) according to population. All figures have been converted to 1978 Colombian pesos in accordance with the food price index prepared by DANE for the blue-collar consumer. For column 2: This column shows the arithmetic ratio of the minimum expenditure estimates of column I to monthly per capita income in 1978. The Latter was obtained by dividing total national income for 1978, as shown in the Banco de la Republic national accounts, by the population estimate given in U.S. Department of Commerce (1979), table 1, page 5, for 1978 (25,673,000). (For full details, see Mauricio Carrizosa, "Determinantes de las Ingresos y ta Pobreza en Colombia" [Bogota: CEDE, 1981; processed.]) The monthly per capita income thus calculated is CoL$2.379. For column 3: This column shows the arithmetic ratio of the minimum expenditure estimates of column I to estimated average monthly expenditure for food (Col$722). To obtain the latter, the proportion of food expenditure in household consumption was calculated first by using the input- output matrix for 1976 prepared by DANE. This proportion was 36.8 percent and was multiplied by "private consumption expenditure" for 1978, given in Banco de la Republica national accounts, to obtain estimated total food expenditure. Per capita expenditure is based on the population estimates for 1978 mentioned in the preceding note. For column 4: In this column, the estimates of minimum food expenditure are divided by the monthly minimum wage. The daily minimum wage in effect in May 1978 (CoL$86) was multiplied by thirty days to obtain the monthly figure of Col52,580. 96 WINNERS AND LOSERS IN COLOMBIA'S GROWTH No studies attempt to measure relative poverty in Colombia. Some investigations of inequality, however, do provide relative quantitative indicators that can be used to analyze the problem of relative poverty. Specifically, as shown by Sen, the Gini coefficient is a special case of the measure of poverty.8 The Sen measure of poverty P can be expressed as follows: P = H [I + (1 - 1) G] where H is the percentage of the population considered poor, I is the average gap between the normative income and the income of the poor expressed as a percentage of the normative income, and G is the Gini coefficient for the population studied. If the entire population is included in P (that is, H = 1) and per capita income is taken as the norm, then I = 0 and consequently P = G. The terms of reference of this special case of Sen's measure-namely, consideration of the total population and the adoption of average income as the norm-clearly implies that it must be interpreted as a measure of relative poverty. If the foregoing interpretation is accepted, studies that estimate the Gini coefficient could be used to determine what has happened to relative urban poverty over time. Unfortunately, the available studies do not make possible any comparison of the Gini coefficient of family income per capita or that of income per equivalent consumer through time. These measures would be the most useful ones for our purposes. Table 34 in Chapter 4, however, includes various Gini coefficients for comparable distributions of income eamers. These coefficients suggest a constant level of relative poverty between 1964 and 1972 and a decrease in relative poverty in the 1970s. Finally, it is useful to summarize the available findings on the characteristics of poverty. In two recent studies, Musgrove and Ferber analyze the problem of identifying the poor.9 Some of their findings are useful for the analysis developed below. * Poverty can be identified on the basis of income from work, because poor families receive no significant income from other sources. This finding is important because the published tabulations of the DANE labor force surveys distribute employed persons according to income from work. Consequently, using these distributions to identify the employed poor and also to compare absolute poverty eliminates the risk of major error. * Educational level is the classifying variable that best distinguishes households by income level. This finding will help in the use of tabulations that combine educational level with other variables to identify which of these other variables can be associated with poverty. CHANGES IN COLOMBIA'S URBAN POVERTY 97 * In the Colombian cities (Bogota and Medellin) studied, the sectors of construction and domestic service show high percentages of poverty. This finding suggests that the movement of average income of these poor groups should be analyzed. Some argue that our knowledge of poverty would be more complete if the analysis took into account the supposed segmentation in labor markets-between the formal and informal sectors, the modern and nonmodem sectors, the marginal and traditional sectors, and so on. Two studies of the Colombian case show no results that would confirm the validity of such distinctions in the 1970s. Kugler, Reyes, and Gomez state: When we control for sex, education, and experience, the difference in average income from work [among sectors] disappears. In general, no significant differences are observed. Not even the fact that the modern sector offers its employees fringe benefits more frequently than the nonmodem sector seems to affect differentially the average incomes of persons with identical levels of education and experience.10 Bourguignon's 1979 study of Bogota yields a similar result: When the most common determinants of income are used correctly the income differential between the modem and traditional sectors decreases greater. It falls from approximately 100% to 20%, and on average, the lower limit in the confidence interval of 95% for that differential does not exceed 10%.01 The distinction between traditional and modern is often used to identify poverty groups. The studies mentioned, however, do not support the use of these classifications to identify trends in poverty. Even if relatively poor persons are found in the informal, marginal, or nonmodern sectors, that fact is not very useful for analyzing changes in the incomes of the poor since these categories usually are not presented in income tabulations. In any event, the studies mentioned suggest that it is not worth the effort in Colombia to try to classify the population according to modem and traditional sectors to study poverty trends. Mohan's 1980 paper gives a rather complete characterization of the poor in Bogota, based on the DANE surveys EH-8 (March 1975) and EH-15 (June 1977).12 The results, some of them surprising, are summarized and discussed below: * Among poor households, the largest ones are relatively less poor. Mohan finds, however, that the number of household members per worker is constant beginning with a family size of five. Actually, as Carrizosa shows, 98 WINNERS AND LOSERS IN COLOMBIA'S GROWTH the critical variable of the family structure for purposes of an analysis of poverty is the number of employed persons per household member.13 * Persons without education are probably poor at any age. Those with primary education also are likely to be poor, but the probability of poverty is greater for those under fourteen years of age, between thirty-five and forty-four, and over sixty-five. In this analysis, we distinguish these two educational categories to identify some other characteristics of poverty. * Migrants are not particularly poor when compared with nonmigrants. The majority of the poor are persons who have been in Bogota more than ten years. Thus, it is inappropriate to state that migrants are the poor; a more reasonable hypothesis is that migration has a depressing effect on urban wages. * The rate of unemployment is inversely related to income leveL This finding can be expected to the extent that the unemployed tend to receive less income. In this analysis, we calculate rates of unemployment by educational level; the result is that the pattern of unemployment based on permanent income appears to be somewhat different. (It is assumed that education is a proxy for permanent income, and, therefore, unemployment is not as clearly inversely related to education as one would expect.) * The rate of participation-that is, the economically active population divided by the total population over five years of age-is directly related to the income level. * The index of concentration of poverty by occupation shows an overrepresentation of poverty among workers in services, manufacturing, construction (males), and trade (females). * Poor males tend to work more hours a week than rich males; the opposite is true of females. * Poverty is high in the manufacturing, construction, retail trade, transportation and communications, and personal and domestic services (males) sectors. * At all educational levels, unemployment among the poor is greater than among those who are not poor. The methodology used to reach this last conclusion presents a problem. By definition, an unemployed person receives less income; he certainly does not receive income from work. As a result, he tends to be included among the "poor," so that, on this basis alone, the poor would show higher rates of unemployment than those who are not poor. When Mohan classifies persons by educational level alone, however, no systematic relationship is observed between educational level and rate of unemploy- ment. The 1977 EH-15 survey shows higher than average unemployment CHANGES IN COLOMBIA'S URBAN POVERTY 99 rates for the more educated females. Through the high school level, it also shows higher than average rates for the more educated males, whereas university graduates have an unemployment rate virtually equal to that observed for persons without education. A 1975 survey also found greater unemployment among the more educated females, except that high school graduates had a higher rate than university graduates. In the case of males, the relationship between educational level and unemployment rates is inverse. On the assumption that education is a variable that better classifies permanent income, the relationship between poverty and unemployment is not obvious.'4 We shall examine this point in greater detail in the section that follows. Several characteristics of the poor warrant attention. There is the relationship between education and poverty. Earlier studies on income functions and on the characteristics of the poor show the expected inverse relationship between poverty and education. Taking into account that with time there may be a decrease in the retums from education, one can investigate the extent to which education contributes to the eradication of poverty.'" Another characteristic is the age-poverty profile, which can be used to determine whether poverty is a permanent state or a transitory condition. Mohan's 1980 study and other studies on income profiles by age suggest that poverty may be a permanent condition for those affected. In the present study, we offer some additional evidence on this aspect. In addition, there is a fairly general agreement that poverty should be analyzed in a family context. For this reason, we complement our analysis of the distribution of employed persons with some evidence on family income per member. Urban Poverty The quantification of poverty presented here refers to different aspects of the urban sector. In general, the study is based on the data of Bogota alone; of the four largest cities (Bogota, Cali, Medellin, and Barranquilla); of three medium-size or small cities (Bucaramanga, Manizales, and Pasto); and of the urban sector as a whole. The study of Bogota alone contains the most data, with annual observations beginning in 1970 as well as some information dating back to the 1960s. For the group of the four largest cities, relatively good statistical information is available, which permits calculations on poverty. For the three medium-size or small cities, evidence is available only for the period 1973-80. Finally, indicators have been constructed for the urban sector as a whole for individual years over a fairly long period (1964-78). Table 45. Employed Poor in the City of Bogota, 1965-80 Normative Reference Price monthly Percentage Rate of Rate of Year period Survey index income of poor participation open unemployment (1) (2) (3) (4) (5) 1965 September CEDE 79.3 383 13.7a - 15.3 1967 Full year CEDE-PRESFAM 100.0 500 17.1 1970 June DANE El 1- 127.9 640 28.6 35.7 11.9 1971 July DANE EH-4 136.2 681 35.5 33.3 9.4 1973 October DANE CENSO 194.7 974 45.2a _ - 1974 June DANE EH-7 236.3 1,182 41.5 9.5 1975 March DANE EIIABE 293.8 1,469 46.4 _ 8.2 1976 June DANE EH-1 345.2 1,726 41.9 36.4 6.5 1977 June DANE EH-15 483.6 2,418 40.7 37.2 6.1 1978 September DANE EH-20 536.8 2,684 29.3 35.8 5.1 1979 June DANE EH-23 664.2 3,321 27.1 37.0 4.4 1980 Marcih DANE Eli-26 781.6 3,908 19.7 38.8 6.0 -Data not available. a. Unreliable data. Source: For column 1: Price index is the blue-collar index for Bogota prepared by DANE, Revista del Banco de la Republica, various issues. The base year 1967 = 100. For column 2: Normative monthly income: sec text. For column 3: Within each distribution of employed persons, the number situated in the interval below norm was aggregated. To this sum was added a proportion of the interval that includes the value of the poverty norm. For column 4: Rate of participation is the economically active population with primary eduication or without education, divided by population over five years of age with primary education or withouLt edtucation. For column 5: Rate of unemployment is unemployed persons with primary education or without education divided by employed persons with primary education or without education. For 1965, the figure used is the average of the rates for that year in the construction branch, reported by Rafael Isaza, "Ocupacion y Desocupacion en Bogota," Empleo y Desempleo en Colombia (Bogota: CEDE, Universidad de los Andes, 1968). For 1965: Richard Nelson, T. Paul Schultz, anid Robert L.. Slighton, Structural Change in a Developing Economy; Colombia's Problems and Prospects (Princeton, N.J.: Priniceton University Press, 1971), p. 146. Primary source is unpublished data of CEDE, Universidad de los Andes. For 1967: Rafael Prieto, "Gasto e Ingreso Familiar Urbano en Colombia," Ensayos ECIEL, nio. 4 (August 1977), p. 115. This is a family distribution. To make it comparable with the distribution of employed persons, the norm of Col$500 was adjusted to Col$1,000, because each family has an average of two employed persons. This assumption is based on the fact that the rate of participation of uineducated persons is nearly 0.35, and average family size is six to seven persons. For 1970-80: DANE. 102 WINNERS AND LOSERS IN COLOMBIA'S GROWTH The first results presented are those for the percentage of the poor in the population. A poverty line of about Col$500 a month per employed person was selected for them in 1967. This is equivalent to Col$2,780 in 1978 pesos. The highest alternative for the normative food expenditure shown in Table 44 represents 61 percent of this amount; this figure is approximately equal to the minimum wage established in May 1978. In selecting this figure, the upper limit of the lowest income bracket in the tabulations of the DANE surveys (Col$500 a month in 1970) was taken into account. The resulting standard would be too high for the analysis of per capita family income. Nonetheless, it seems reasonable to use this upper limit in the analysis of the incomes of economically active people because employed persons have dependents to support.'6 It can also be used to make comparisons over time, which probably do not depend significantly on the absolute standard adopted. The results for Bogota are shown in Table 45. By limiting the analysis to that city, it is possible to estimate a series for a fifteen-year period from 1965. As seen in the table, information was available for nearly all years of the 1970s and for two years of the preceding decade. To express normative income in current-year pesos for the corresponding reference period, the blue-collar consumer price index was used for the city of Bogota. To aid in interpreting the figure for the percentage of poor, the last two columns of the table show the rates of participation and of unemployment for persons without education or with primary education. According to the series obtained, the percentage of poor seems to increase from 1965 to 1975 and then decline until March 1980, the latest date for which information is available. As was mentioned in Chapter 4, the census income data for 1973 are very unreliable, and underestimation of incomes is substantial. The absolute levels of poverty are therefore clearly overestimated in 1973 in the table. The income figures for 1965 and 1967, in contrast, seem suspiciously high. The CEDE-PRESFAM income data give higher average incomes than those obtained in the other surveys. In 1968 constant pesos, the 1967 PRESFAM average income is Co1$25,024 a year, whereas the EH-4 (1971) survey has only Col$21,498 a year after four years of high per capita income growth at the national level. The PRESFAM data also give a better income distribution than other sources. For all these reasons, this 1967 data source is not comparable with other income sources, although in fact this survey may have income data of good quality. Also, the PRESFAM survey is not a random sample, and the assumptions used to derive the distribution from a stratified sample may create serious biases. In summary, the poverty trends CHANGES IN COLOMBIA'S URBAN POVERTY 103 Table 46. Rate of Participation in Bogota, 1970-80 Rate of participation (percent) Population without Reference Total education or with Year period Survey population primary education 1970 June DANE EH-1 54.5 35.7 1971 july DANE EH-4 49.3 33.3 1972 September DANE EH-6 46.2 1973 October DANE CENSO 46.1 1974 June DANE EH-7 51.9 1975 March DANE EH-8E 50.3 1976 June DANE EH- 1 51.0 36.4 1977 June DANE EH-15 49.8 37.2 1978 September DANE EH-20 50.5 35.8 1979 June DANE EH-23 51.6 37.0 1980 March DANE EH-26 56.4 38.8 -Data not available. Source: The total rate of participation is taken from DANE, Boletin Mensual de Estadistica no. 345 (1980), pp. 71 and 74. The rate of participation of the population without education or with primary education is taken from Table 45. The level of the two rates is not wholly comparable because the first covers the population over ten years of age, whereas the second covers the population over five years of age. of the 1960s may be a result of changes in the quality of the data and may not reflect changes in real poverty levels. It is also probably misleading to compare the EH-1 1970 survey with other income surveys. The 1970 survey was probably the best and the most thorough of these surveys, and it appears to have much less income underestimation. A basic limitation of the foregoing exercise is that income distributions for employed persons have been used. It is possible that the observed increase in the proportion of poor people in some years simply reflects a trend toward more participation in the labor force by members of poor families. Table 46 shows the rate of labor force participation in Bogota for the total population and for the population over five years of age without education or with primary education. Actually, the rate of participation of the total population declined somewhat between 1970 and 1979, while the percentage of the poor employed was rising. Consequently, at least part of the increase in the percentage of employed persons living below the poverty line could be attributed to a greater relative participation in the labor force by persons with a low level of education. This and other 104 WINNERS AND LOSERS IN COLOMBIA'S GROWTH problems of interpretation can be attributed to the fact that distributions of employed persons are not the most suitable measure of family welfare. Another problem with calculating changes in the degree of poverty on the basis of the income of the employed is that this leaves out the problem of unemployment, and one positive aspect of the period analyzed is the behavior of the unemployment rate, which decreased significantly. This trend is shown in Table 47 for the total economically active population and for the group with no education or with only primary education. Table 48 adds the uneducated unemployed to the poor; it therefore shows the percentage of the economicalFy active population that is below the poverty line. It can be seen that the inclusion of unemployed persons does not substantially change the poverty trends shown in Table 45 for Bogota. The unemployment series presented in Table 47 reveals one change: the unemployment rate for uneducated persons is higher than the average rate in 1963-67 and lower in the 1970s. On the average, unemployment of uneducated persons in 1963-66 was 11 percent, while the overall rate was 8.7 percent. The average figures for 1977-80 were 5.4 percent for the uneducated and 7.5 percent overall. The change in the pattern of unemployment can be seen better in Table 49, which shows the coefficient of concentration of employed persons without any education or with only primary education in Bogota. This coefficient is defined as the ratio of the proportion of unemployed persons without education or with primary education in the total number of unemployed to the proportion of the labor force without education or with primary education in the total labor force. It is a measure of the "overrepresentation" of unemployment of the poor. As shown in the table, throughout the 1970s, unemployed persons without education or with primary education got progressively more underrepresented. In other words, their participation in unemployment was progressively less than their participation in the labor force. The change probably occurred at the end of the 1960s. The corresponding index of concentration for blue-collar workers and nonfarm production workers was 1.64 for 1963, 1.49 for 1964, 1.48 for 1965, 0.85 for 1974, and 0.97 for 1978.'7 In other words, it appears that during the 1960s unemployment was concentrated among the poorest groups, whereas during the 1970s the poor represented a comparatively smaller proportion of the unemployed. Table 48 indicates a sharp decline in the percentage of the poor after 1977, with the percentage in 1980 being significantly lower than that observed at the beginning of the 1970s. Correspondingly, data on the incomes of unskilled workers indicate that poverty was less extensive in 1978-80 than during the 1960s. Those incomes rose considerably in the CHANGES IN COLOMBIA'S URBAN POVERTY 105 Table 47. Rate of Unemployment in Bogota, 1963-80 Rate of unemployment Employed persons without Reference education or vvith Year period Survey Overall primary education 1963 June CEDE 8.7 9.4 1964 June CEDE 7.2 11.6 1965 June CEDE 8.8 13.0 1966 March CEDE 10.1 12.9 1967 June CEDE 12.7 - 1970 June DANEEH-1 13.1 11.9 1971 July DANE EH-4 9.2 9.4 1972 September DANE EH-6 7.4 - 1973 October DANE CENSO 10.7 - 1974 June DANE EH-7 11.1 9.5 1975 March DANE EH-8E 10.2 8.2 1976 June DANE EH-iI 8.4 6.5 1977 June DANE EH-15 7.8 6.1 1978 September DANE EH-20 6.7 5.1 1979 June DANE EH-23 6.1 4.4 1980 March DANE EH-26 9.5 6.0 -Data not available. Source: The overall unemployment rate is taken from DANE, Boletin Mensual de Estadistica no. 345 (1980), pp. 83, 86. The unemployment rate for uneducated persons or those with primary education is taken from Table 45. For 1963-67, the unemployment rate of uneducated persons is based on employment, in the construction sector as reported in Rafael Isaza, "Ocupacion y Desocupacion en Bogota," Empleo y Desempleo en Colombia (Bogota: CEDE, Universidad de los Andes, 1968). Table 48. Percentage of Poor in the Economically Active Population of Bogota, 1970-80 Reference Year period Survey Percentage of poor 1970 June DANE EH-1 31.4 1971 July DANE EH-2 37.7 1974 June DANE EH-7 44.9 1975 March DANE EH-8E 45.8 1976 June DANE EH-11 41.4 1977 June DANE EH-1S 40.4 1978 September DANE EH-20 29.5 1979 June DANE EH-23 27.4 1980 March DANE EH-26 20.0 Source: Unemployed persons without education or with primary education were added to the employed "poor" and the resulting figure was divided by the total economically active population. The primary data sources are indicated in Table 44. 106 WINNERS AND LOSERS IN COLOMBIA'S GROWTH Table 49. Coefficient of Concentration of Unemployed Persons without Education or with Primary Education in Bogota, 1970-80 Reference Index of Year period Survey concentrationa 1970 June DANE EH-1 0.91 1971 July DANE EH-4 1.03 1974 June DANE EH-7 0.86 1977 June DANE EH-15 0.78 1980 March DANE EH-26 0.63 a. The index of concentration is 1 = di/fi, where di is unemployed persons without education or with primary education as a percentage of total unemployment, and fi is the economically active population without education or with primary education as a percentage of the total economically active population. Source: DANE. Table 50. Real Monthly Incomes of Selected Low-Income Groups in Bogota, 1976-80 (constant 1976 Colombian pesos) Reference Service Nonfarm laborers and Construction Year period Survey workers production workers workers 1976 June EH- ll 316 545 489 1980 March EH-26 44t 676 694 Source: DANE. Real incomes are arrived at by using the blue-collar consumer price index late 1970s, exceeding the maximum levels reached in earlier years. The same cannot be said about the incomes of skilled blue-collar workers (those working in manufacturing, for instance); for these, the recovery in real wages in the late 1970s was not sufficient to regain the levels reached earlier in the decade. But these workers clearly cannot be included among the poor. Table 50 presents additional evidence on the incomes of the poor. As Musgrove and Ferber (1978 and 1979) and Mohan and Hartline (1984) have shown, service workers and nonfarm laborers, as well as construction workers, have comparatively high indexes of poverty. Average incomes have been calculated for these groups in Bogota for 1976 and 1980. Although the differences in the quality of the data make it difficult to compare absolute income levels, real incomes do appear to have improved. CHANGES IN COLOMBIA'S URBAN POVERTY 107 The consistency of the data from labor force surveys can be analyzed by using the Bogota data. One way of doing this is to compare the "most frequent income" bracket shown in the surveys for construction workers with the nominal monthly wage reported to DANE by construction firms. It can be reasonably assumed that the most common worker in the construction sector is the unskilled laborer. Table 51 presents such a comparison and, in general, the two sets of statistics are similar. It can be concluded, therefore, that for this income level the survey data are accurate. The study of the survey data for the four largest cities together complements the analysis made for Bogota. Table 52 shows the high percentage of the poor observed in 1974-77 and the subsequent decline in this indicator. It also shows the decrease in unemployment that occurred during the decade as well as the increase in the rate of participation. Table 53 presents the degrees of poverty for three small- or medium- size cities (Bucaramanga, Manizales, and Pasto). The trend for the proportion of the poor is similar to that observed for the four largest cities, although the percentage is higher. This is explained largely by the structure of the economically active population by educational level. In the four largest cities, the proportion of persons without education or with primary education in the total labor force is 0.46 percent; the corresponding figure for Bucaramanga, Manizales, and Pasto is 0.51 percent. The unemployment and participation in the three cities show the same pattern noted earlier- namely, a decrease in unemployment and an increase in participation. The results for the urban sector as a whole, which are shown in Table 54, confirm what has been said already. Poverty increased at the start of the 1970s and decreased at the end of the decade. Another way of identifying trends in urban poverty is to examine the rates of participation and unemployment by age and educational level. Table 55 shows these rates for the urban sector as a whole. They indicate cycles of participation and employment over the life span. The poor participate in the labor force more at the beginning of the life cycle, rather than investing in education. Later, they participate relatively less as they find job opportunities that are comparatively less remunerative. Unem- ployment characterizes young workers regardless of educational level; later, at the usual productive ages, it affects all groups except university graduates. As noted in the analysis of the data for Bogota, the pattern of unemployment according to educational level changed during the 1970s. At the end of the decade, not only was unemployment lower, but it was also distributed more widely within the labor force. Thus, it no longer had such a particular impact on the poor as it did during part of the 1960s. Table 51. Most Frequent Incomes in the Construction Sector and Monthly Labor Costs for Apprentices, 1973-80 (Colombian pesos) Monthly income bracket Reference Modal bracket in the survey according to construction Year period Survey for the construction sector costs (DANE) 1973 October DANE-CENSO 500-1,000 742-1,012 1975 March DANE EH-9 - 1,175-1,603 1976 June DANE EH- l 1 1,000-1,500 1,372-1,870 1977 Juine DANE EH-15 1,500-2,000 1,694-2,311 1978 September DANE EH-20 2,500-3,000 2,400-3,272 1979 June DANE EH-23 3,000-4,000 3,138-4,279 1980 March DANE EH-26 4,000-5,000 3,914-5,337 -Data not available. Source: These survey brackets were found to have the largest number of employed persons among the distributions of employed persons in the construction sector. The other brackets are defined as the range betweeni what a construction apprentice would eam for twenty-two days of work per month and what he would earn for thirty days of work per month The income of a construction apprentice was calculated on the basis of the indexes published in various issues of the DANE Boletin Mensual de Estadistica, which take as the base value a daily wage of Col$25.70 in December 1971. Table 52. Poverty in Bogota, Cali, Medellin, and Barranquilla, 1974-80 Price index Normative Reference monthly Percentage Rate of Rate of open Year period Survey Original base 1967 income of poor participationa unemployment 1974 June DANE EH-7 862.4 236.7 1,184 44.4 - 11.9 1975 October DANE EH-9 1,100.5 302.0 1,510 42.9 - 9.7 1976 June DANE EH-1 l 1,261.8 346.3 1,732 43.6 35.1 9.1 1977 June DANE ElI-15 1,794.3 492.4 2,462 45.2 36.3 8.5 1978 September DANE EH-20 2,010.4 551.7 2,758 34.0 36.2 6.5 1979 June DANE EH-23 2,459.0 674.8 3,374 29.8 37.2 7.3 1980 March DANE EH-26 2,883.3 791.2 3,956 21.4 38.5 8.5 -Data not available. a. Rate of participation: economically active population with primary education or without education, divided by population over five years of age with primary education or without education. Source: Price index: weighted average of blue-collar indexes for Bogota, Cali, Medellin, and Barranquilla prepared by DANE, Revista del Banco de la Republica, various issues. The weights are based on the population of each city in 1973. Normative monthly income: see text Percentages of poor, rate of participation, and rate of unemployment: see Table 45. Income data 1974-80: DANE. Table 53. Percentage of Poverty in Bucaramanga, Manizales, and Pasto, 1975-80 Price index Normative Reference Base monthly Percentage Rate of Rate of Year period Survey Original 1967 income of poor participation unemploymnent 1975 October DANE EH-9 1,191.7 311.0 1,555 48.8 - 10.0 1976 March DANE E11-10 1,302.5 339.9 1,700 53.4 37.0 9.0 O 1977 September DANE EH-16 1,927.7 503.0 2,515 50.5 38.3 7.7 1978 September DANE EH-20 2,171.6 566.7 2,834 40.9 36.1 6.2 1979 September DANE EH-24 2,822.8 736.6 3,683 40.7 36.7 6.0 1980 March DANE EH-26 3,174.7 828.5 4,142 30.9 38.5 6.3 -Data not available. SOurce: Price index: weighted average of the blue-collar indexes for Bucaramanga, Manizales, and Pasto prepared by DANE, Revista del Banco de la Republica, various issues. The weights are based on the population of those cities in 1973. Normative monthly income: see text. Percentage of poor and rate of participation: see Table 45. Table 54. Poverty in the Urban Sector of Colombia, 1971-78 Norrnative Reference monthly Percentage Rate of Rate of Year period Survey Price index income of poor participation unemployment 1971 July DANEEH-4 136.1 680 42.42 31.5 10.9 1972 September DANE EH-6 159.9 800 47.04 - - 1974 November DANE EH-8 257.3 1,286 51.30 32.6 9.2 1978 June DANE EH-19 567.8 2,839 41.95 35.1 6.5 -Data not available. Source: Price index: total for blue-collar workers in seven cities prepared by DANE, Revista del Banco de la Republica, various issues. Normative monthly income: see text. Percentage of poor, rate of participation, and rate of unemployment: see Table 45. Distribution of eniployed persons: population with primary education or without education (total over five years of age, economically active, and unemployed). 1971: I)ANE, Encuesta Nacional de Hogares-Fuerza de Trabajo 1971 (Bogota: July 1976). 1972: DANE, Emenesta Nacional de Hogares, 1972 (December 1977), p. 58. This is a distribution of recipients; to obtain the proportion of employed persons below the poverty line of Col$800 per month, the ratio of employed persons to poor income recipients observed in EH-4 (1.64) was multiplied by the proportion of poor income recipients observed in EH-6. 1974-78: DANE, Table 55. Rates of Participation and Unemployment by Age and Educational Level, 1978 Educational level 10-11 12-14 15-19 20-29 30-39 40-49 50-59 60-69 70-79 None Rate of participation 0.08 0.45 0.44 0.54 0.60 0.56 0.52 0.37 0.18 Rate of unemployment 0.24 0.05 0.14 0.07 0.04 0.01 0.06 0.02 0.03 Primary Rate of participation 0.01 0.11 0.56 0.66 0.64 0.62 0.54 0.40 0.19 Rate of unemployment 0.03 0.11 0.14 0.10 0.03 0.03 0.04 0.04 0.10 ,>, Secondary Rate of participation 0.03 0.22 0.68 0.70 0.68 0.56 0.39 0.26 Rate of unemployment 0.24 0.18 0.11 0.04 0.04 0.02 0.05 0.00 University Rate of participation 0.16 0.61 0.92 0.93 0.81 0.41 0.55 Rate of unemployment 0.43 0.12 0.01 0.00 0.00 0.00 0.00 Note: The rate of participation is defined as the economically active population) in each age-educational level group, divided by the population over five years of age in that group. The rate of unemployment is defined as the ratio of unemployed persons to the economically active population in each group. Source: The original data are from Phase 19 (une 1978) of the DANE household survey. CHANGES IN COLOMBIA'S URBAN POVERTY 113 Table 56. Poverty by Employment or Unemployment in the Urban Sector, 1978 Distribution of Percentage of employment and the total Percentage of the unemployment economically active Coefficient of poor in category among the poor population concentration Category (1) (2) (3) (2 - 3) Unemployed 50.4 6.5 7.6 0.86 Employed 59.5 93.5 92.4 1.01 Total 58.8 100.0 100.0 100.0 Source: DANE, Boletin Mensual de Estadistica no. 326 (September 1978), pp. 39 and 57. The "poor' are defined as those who have a primary education as a maximum. Table 56 classifies the labor force on the basis of employed and unemployed persons. The first column shows that approximately one-half of the unemployed are comparatively "poor" (persons with primary education as their highest level). The second column shows that of the total number of "poor" in the labor force, 6.5 percent are unemployed. The third column indicates that 7.6 percent of the total labor force is unemployed. In other words, unemployment among the comparatively "'poor" is lower than overall unemployment. Consequently, the last column, which shows the arithmetic ratio of the second column to the third, gives an index of concentration of unemployment among the poor of slightly less than one. All of this means that unemployment is not confined to the poorest groups. It does not mean that unemployment is not a particularly serious condition for them. It obviously is, because, unlike persons in the higher income strata, the poor do not have supplementary incomes to offset the lack of income from work during periods of unemployment. Conclusion The study of poverty levels in this chapter depends greatly on the quality of the income data in the surveys used. As has been repeatedly explained, coverage and levels of underestimation vary among surveys. But because basic labor income is reported more accurately than other types of income, the observed changes in the percentage of wage earners below an absolute poverty line may approximate real changes in poverty levels for this group. The high concentration of poverty among the self-employed, on the other hand, may reflect more the problem of capturing this income in the surveys 114 WINNERS AND LOSERS IN COLOMBIA'S GROWTH Table 57. Poverty by Occupational Level in the Urban Sector, 1974 Percentage of Percentage of poor at poor accounted Percentage of occupational for by each total number Coefficient of Occupational level occupation of employed concentration group (1) (2) (3) (4) White-collar workers 37.9 27.0 36.6 0.74 Blue-collar workers 62.1 25.9 21.4 1.21 Employers 18.6 1.3 3.7 0.35 Self-employed workers 57.0 28.4 25.6 1.11 Unpaid family workers 100.0 6.7 3.4 1.97 Domestic workers 97.8 17.3 9.1 1.90 Total 51.3 100.0 100.0 1.00 Note: The total of column 2 is grearer than 100 because of errors in the published statistics. Source: DANE. than real poverty (Table 57). Therefore, the poverty trends in this section probably show the direction of change. However, small differences in the coefficients should not be considered significant. The problem posed by the income data is well illustrated by the information from the PRESFAM 1967 survey. This is probably the most carefully done of all the surveys, and it shows both less poverty and a better income distribution.1" Although the sample may have underrepresented the poor, it is very probable that careful questions concerning the income of the poor eliminates substantial underreporting of the income of the self- employed. But if we ignore problems of levels and concentrate on trends derived from the data of similar dubious quality, we get a picture of growing urban poverty from 1970 to 1975 and decreasing poverty in the second part of the decade. This picture coincides with the data on Cali at the microeconomic level presented in Chapter 3 and the data on real wages by occupational group. A 1981 study on nutrition levels by Mohan, Wagner, and Garcia offers some additional corroborating evidence. It concludes from comparable household survey data that malnutrition in Bogota and Cali decreased from about 30 percent to 15 percent of the population between 1973 and 1978.15 From this analysis, it does appear that urban poverty decreased in the 1970s. It is clear, however, that in the first part of the decade, poverty increased temporarily. In the following chapter, some hypotheses will be CHANGES IN COLOMBIA'S URBAN POVERTY 115 suggested to explain the trends in poverty and income distribution shown by the data analyzed up to now. Notes to Chapter 5 1. At the time of the 1964 census, 52.1 percent of the population lived in municipal capitals, while at the time of the 1973 census, the figure was akteady 63.1 percent. 2. Gustav Ranis, "Distribucion del Ingreso y Crecimiento en Colombia," Desarrollo y Sociedod, no. 3 January 1980), p. 91. 3. See Departamento Nacional de Planeacion, Para Cerrar la Brecha (Bogota: Banco de la Republica, 1975); and Intetnational Labour Organisation (ILO), Hacia el Pleno Empleo (Geneva, 1970). 4. Rakesh Mohan and Nancy Hartline, The Poor in Bogota: Who They Are, What They Eam, Where They Live, World Bank Staff Working Paper no. 635 (Washington, D.C., 1984). 5. Francois Bourguignon, "Pobreza y Dualismo en el Sector Urbano de las Economias en Desarrollo: El Caso de Colombia," Desarrollo y Sociedad Uanuary 1979). 6. Oscar Altimir, "The Extent of Poverty in Latin America: A Summary" (Santiago: Comicion Economica para America Latina [CEPAL], 1978; processed). 7. Mohan and Hartline, The Poor in Bogota. 8. Amartya Sen, "Poverty: An Ordinal Approach to Measurement," Econometrica, vol. 44, no. 2 (1976). 9. Philip Musgrove and Robert Ferber, "Finding the Poor," Revieuw of Income and Wealth, vol. 24, no. 3 (September 1978); and Musgrove and Ferber, "Identifying the Urban Poor: Characteristics of Poverty Households in Bogota, Medellin, and Lima," Latin American Research Review, vol. 14, no. 2 (1979). 10. Bernardo Kugler, Alvaro Reyes, and Maria Isabel de Gomez, Educacion y Mercado de Trabajo Urbano en Colombia: Una Comparacion entre Sectores Moderno y no Modemo, monograph 10 (Bogota: Corporacion Centro Regional de Poblacion, 1979), p. 48. 11. Bourguignon, "Pobreza y Dualismo," p. 69. 12. Rakesh Mohan, The Peopte of Bogota: Who They Are, What They Earn, Where They Lite, World Bank Staff Working Paper no. 390 (Washington, D.C., 1980). 13. Mauricio Carrizosa, "Determinantes de los Ingresos y la Pobreza en Colombia" (Bogota: CEDE, Universidad de los Andes, 1971, processed). 14. In the human capital approach, education is a determinant of lifetime income. In Colombia it has been shown that education and age explain a large proportion of income differentials; therefore, education may be a good proxy for lifetime income. 15. Francois Bourguignon, "The Role of Education in the Urban Labor Market during the Process of Development" (Paper presented at the Sixth World Congress of the Intemational Economics Association, August 1980, Mexico City, processed). 16. In Colombia, on the average, one income eamer supports himself and about two more people. Therefore, the Col$2,780 of 1978 would generate an income of close to Col$926 per capita. If 65 percent is spent on food, the food espenditure amounts to about Col$600 per month, which is between estimates C and D in Table 44. These two estimates seem the most realistic minimum of food expenditure levels. 17. The data for 1963-65 are based on figures reported in Rafael Isaza, "Ocupacion y Desocupacion en Bogota," in Empleo y Desempleo (Bogota: CEDE, Universidad de los Andes, 1968), pp. 123, 135. The data for 1974 and 1978 are based on tabulations of the EH-8E and EH-19 surveys of DANE. The denominator of the index is the percentage of employed nonfarm workers in the total number of employed persons. 116 WINNERS AND LOSERS IN COLOMBIA'S GROWTH 18. In 1967, urban unemployment in Colombia was higher than in any other year between 1961 and 1980, and the economy grew at one of the lowest rates of that period. 19. This is not strictly an estimate of those actually malnourished. It is an estimate of the proportion of people who do not have an income that is adequate to cover their food and nutrition needs under existing food and consumption habits and prices. 6 Some Hypotheses on the Determinants of Changes in Income Distribution in Colombia THE PURPOSE OF THIS RESEARCH PROJECT was to determine how the standard of living of various classes of the Colombian population had changed during the 1960s and 1970s. A similar effort had been carried out for the 1940s and 1950s in another publication.' No systematic effort has been made to determine empirically the causes of the changes in income distribution. This chapter, however, offers some hypotheses that may explain the observed changes and that may, in the future, be empirically tested. Labor Force Trends in the Countryside One of the most significant changes in Colombia in the 1960s was that the rural labor force ceased to grow.2 As the demand for agricultural products continued to increase, rural labor productivity per capita began to improve markedly. In the 1960s, the lack of growth in the supply of labor in rural areas probably reduced the degrees of underemployment in agricul- ture, but in the 1970s the real income of agricultural laborers started to grow because of two mechanisms. First, agricultural prices increased somewhat more rapidly than the general price level, and small farmers and landless workers received part of this gain. Second, productivity per worker increased, as output expanded while the supply of labor contracted. Because landless laborers were the poorest group in Colombian society, their improved income led to a decrease in the proportion of families below the poverty line and to an improvement in the income distribution. 117 118 WINNERS AND LOSERS IN COLOMBIA'S GROWTH The improvement of wages in the countryside was also transmitted to unskilled workers in the cities. Migration from rural to urban areas is a function of rural-urban income differentials and urban unemployment rates. As rural incomes improved, rural-urban migration could only continue if wages in the cities also improved or unemployment decreased. The empirical evidence suggests both things occurred. The urban labor demand absorbed all the natural labor growth of both rural and urban areas, while also absorbing part of the pool of urban unemployed workers that existed at the end of the 1960s. This tight labor market in the countryside seems to have led to some decrease in the wage differential between agricultural laborers and urban unskilled workers, but the income data analyzed in previous chapters show that the earnings of unskilled urban workers also increased in real terms. Despite gains in labor productivity in agriculture, labor demand did not decrease in the rural areas. In the first place, agricultural exports increased. The most dramatic case was that of coffee, the production of which almost doubled in the decade of the 1970s. This increased production required substantial investment in preparing land and planting the new coffee variety. The new variety (caturra) has increased the output per laborer, but during the time that the old plantations were being renewed and new lands prepared for caturra production, the demand for rural labor was high. Furthermore, coffee producers invested a high proportion of their windfall gains from the high world coffee prices of 1975-78 in renewing their plantations and planting caturra. In summary, the high investment in new coffee technology in the 1970s demanded much rural labor, and this was one determinant of higher rural wages. Because world demand for coffee grows slowly, the expansion of coffee production could not continue after Brazil recovered from the 1975 frost. The increase in labor demand associated with the introduction of the caturra technology, therefore, cannot be expected to continue. The weakening of this component of rural labor demand might be expected to affect rural incomes negatively. The export of cut flowers also grew extremely rapidly during the 1970s. This is an activity that makes intensive use of female labor. In the rural areas where flowers are grown for export, family incomes have increased substantially by the incorporation of women into the labor force. Economic policy during the period of high coffee prices favored rural labor income in various ways. First, minimum wages in the countryside were increased rapidly when the labor market was tight. Second, part of the increases in world prices were passed on to coffee producers, in a country DETERMINANTS OF CHANGES IN INCOME DISTRIBUTION 119 where most production is still in the hands of small and medium-size producers. Third, the supply of nonagricultural imported products was allowed to increase to avoid the increases in prices of consumer goods that would have occurred had rural incomes increased and import growth been restrained. During the coffee boom, Colombian industry was functioning at full capacity utilization, and a tight import policy would have led to rapid increases in the prices of nonfood consumer products. During the 1970s, the export of illegal drugs (marijuana and cocaine) also increased dramatically. Such exports may have gone from negligible figures to about US$500 million in 1977-78.3 Colombia became an important exporter of cocaine, but since the cocaine paste was not produced locally, but imported from Bolivia and Peru, this line of business only produced very high profits to a few underworld entrepreneurs and their laborers; it had little impact on labor demand. The industry was also capital-intensive; it used expendable airplanes, laboratories, ships, and armaments, but very little labor. While marijuana was grown in Colombia, most of the value added was produced in the United States where it was sold. The rural labor force involved in growing marijuana is small, located mostty in the non-coffee-producing north coast of the country, thus coincidentally increasing labor demand in an area where the coffee boom had little impact. In summary, the illegal drug exports of the 1970s probably had little impact on labor incomes, but they did produce some very high capital incomes for a very small group of people. This phenomenon probably helps to explain why the average incomes of the top 5 percent of the income distribution probably grew quite rapidly in the decade. In cities like Barranquilla and Medellin, there was indeed a feeling that much of the trade in luxury apartments and automobiles was dominated by people tainted by these illegal activities. Labor Force Trends in the Cities Although urban labor supply increased rapidly in the 1970s, economic growth was sufficiently rapid in labor-intensive sectors to absorb both the increase in the labor force and part of the high levels of unemployment common in the late 1960s. The decline in unemployment, especially among the poor, clearly reduced the proportion of families living below the poverty line. But the increasing participation of women in the labor force probably contributed even more to the improved welfare of the lower deciles of the population. The data from Cali show not only that women 120 WINNERS AND LOSERS IN COLOMBIA'S GROWTH increased their labor force participation, but that the income of these additional unskilled workers grew at much higher rates than the average for all occupations. Many of the changes in labor supply have to do with the decrease in the rate of population growth between the 1950s and the 1970s; that rate declined from about 3.3 percent to 2 percent a year. Even so, the labor force continued to grow rapidly in the 1970s because participation rates increased and the labor force had to absorb the postwar baby boom. The question that must now be addressed is what factors explain the rapid increase in labor demand in the 1970s. Sectoral Growth in the 1970s It is not easy to understand why labor demand grew so rapidly in Colombia in the 1970s. In fact, one report in 1970 had projected growing unemployment levels even on the assumption of rapid rates of economic growth.4 Some of the possible causes for a high labor demand in the countryside have been outlined earlier. It is more difficult to pinpoint the reasons why urban labor demand was able to absorb the natural rate of growth of the urban labor force, plus the rural people who migrated to the cities and part of the stock of those unemployed during the 1970s. Three factors may help to explain this phenomenon. For various reasons, sectors with low capital- output ratios grew faster than the sectors with high capital-output ratios; the latter grew faster in the 1960s. Some service sectors with low capital-labor ratios also grew rapidly. In the government sector, for example, the policy was to keep wages constant in real terms, or even to decrease them, while increasing the amount of coverage of services by increasing the number of employed people. This policy was followed in areas like education and the police. Finally, macroeconomic policy favored sectors with lower capital-labor ratios. A series of economic policy reforms starting in 1967 seem to have contributed to a structural change in the economy that helped to create employment. In that year, the government initiated a policy of opening up the economy, which had various positive effects. The institution of a crawling peg, in which the exchange rate is devalued by a few cents against the dollar two or three times a week, eliminated both the overvaluation of the Colombian peso and the periodic exchange-rate crises common in the 1960s. This gradual approach to devaluation depoliticized the exchange DETERMINANTS OF CHANGES IN INCOME DISTRIBUTION 121 rate, allowed a real depreciation of the Colombian peso between 1967 and 1975, and avoided the revaluation that would have occurred in a floating- exchange-rate regime when, after 1975, foreign exchange earnings in- creased dramatically because of temporarily high coffee prices. The move toward a long-run equilibrium exchange rate encouraged labor-intensive exports and reduced the bias against agriculture created by import-substituting policies. The latter effect, well documented by Jorge Garcia, allowed the performance of the agricultural sector to improve.5 Because agricultural production in Colombia is very labor intensive, this change in policy helps explain the improvement in agricultural incomes. Another factor that increased the rural demand for labor was the technological change in coffee production. During the 1970s, the adoption of the new high-productivity caturra variety caused coffee production almost to double. Although this new technology produced more coffee per acre and per worker, its profitability led to large investments in new plantations and the preparation of land for these plantations. The rural demand for labor, therefore, grew rapidly during this investment boom. The government policy of transferring part of the high intemational coffee prices during the decade to the coffee growers assured them a cash flow for carrying out these investments. Export subsidies and an adequate exchange rate stimulated production and exports of items like cut flowers, which generated substantial labor demand in the rural areas around Bogota, Medellin, and Cali. In the cities, the labor-intensive sectors also appear to have grown more rapidly than those with high capital-labor ratios. Small industry seems to have grown more rapidly than large industry. Truck and bus transport, and other labor-intensive sectors, also grew more rapidly than the average. Within large industry, however, there is no clear evidence to show a faster growth of labor-intensive sectors. In contrast, the elimination of import controls and periodic exchange- rate crises increased the degree of capital utilization in industry, and this clearly promoted employment growth. In the 1960s, there was excess capacity in industry because entrepreneurs would overinvest in imported capital equipment when the balance-of-payments situation was good, knowing that they would not be allowed to import when the balance of payments was negative. Furthermore, with import controls, an enterprise that could show excess capacity could convince the import authorities not to give licenses for capital equipment to competing industries or licenses to importers of competing foreign goods. The upward adjustment of the exchange rate between 1967 and 1973 also raised the prices of imported capital equipment and favored domestic labor. In addition, easier access to 122 WINNERS AND LOSERS IN COLOMBIA'S GROWTH import licenses made possible a decrease in inventories of industrial inputs and spare parts. This phenomenon also lowered capital-output ratios. In summary, the gradual move away from overvalued exchange rates in the 1970s probably favored employment, and this improved the income of the poor. The Impact of the Supply of Education The shifts in the middle of the income distribution were probably determined by the supply of education and inflation. In the 1960s and 1970s, Colombia made a great effort to increase the supply of educated manpower. According to the World Development Report 1980, the number of children enrolled in primary school rose from 77 percent of the normal primary school age group in 1960 to 103 percent in 1977.6 Progress in secondary school and university education was even more spectacular. The number of youths enrolled in secondary school went from 12 percent of the age group in 1960 to 39 percent in 1977, and the percentage of the population twenty to twenty-four years of age enrolled in higher education rose from 2 percent to 9 percent. This increase in the supply of educated manpower had to lower the rate of return to education, and there is some empirical evidence to confirm this hypothesis. Francois Bourguignon reported earnings functions for various years in the decade 1965-75 and concluded that the returns to education declined significantly between the mid- 1960s and the mid-1970s.7 This was particularly true for university education. In relative terms, individuals with university education may have lost between 30 percent and 45 percent of the advantage they had during the 1960s over those who had completed only primary school. For secondary education, the corresponding relative loss was between 25 percent and 30 percent.8 Using different data, Kugler, Reyes, and Gomez also identified the trend toward a decrease in the returns to education.9 The rapid increases in the supply of educated manpower may explain the sluggish growth in the real earnings of white-collar workers. The stagnation in the real income of this category of workers explains part of the loss of participation in total income of the middle-class deciles (seventh to ninth deciles) of the population."0 The fact that these groups saw their relative position improve in the 1960s and deteriorate in the 1970s coincides with the trends in the rates of return to education. Another factor that contributed to stagnant earnings and increased unemployment among educated workers in the 1970s was the slow growth of the government sector. Because government employs a large proportion DETERMINANTS OF CHANGES IN INCOME DISTRIBUTION 123 of highly educated workers, a growth of government services should increase the demand for this type of labor and, therefore, also increase the real eamings of the middle class. From 1967 to 1971, central government expenditures, in real terms, on wages and salaries of government employees grew by 42 percent; from 1971 to 1980, however, such expenditures grew by only 7 percent. The consolidated central govemment sector (the various ministries, decentralized institutes, and central government-owned enter- prises) had a less drastic, but still significant, decrease in the rate of growth of expenditures on wages and salaries. From 1967 to 1971, such expenditures grew by 52 percent in real terms, while, from 1971 to 1979, they grew by only 23 percent. The Effect of Growing Inflation Rates on Income Distribution The acceleration of inflation in the first part of the 1970s also had negative effects on white-collar employees and skilled workers in modern industry. From the late 1940s to the late 1960s, a process of collective bargaining was developed in Colombia, which led to the determination of wages in modern-sector enterprises on the basis of negotiated agreements signed for two-year periods."1 This institutional framework determined that when the rate of inflation accelerated in the early 1970s, real wages started to decline because of a lag in the adjustment of wages to inflation.2 By 1975, however, an institutional change had occurred and collective bargaining agreements were being negotiated every year. In the public sector, before 1975, wage increases required legislation to be passed by Congress, and such laws were passed only every two or three years; yearly increases began only in the middle of the 1970s. For these reasons, in both the private and the public sectors, real wages decreased when inflation accelerated. Another institutional factor that affected real wages during a period of accelerating inflation was the practice of infrequently changing the minimum wage."3 Since 1963, the minimum wage has been lower than the wage actually paid in the modern sector in the urban areas. Between 1963 and 1974, the minimum wage declined in real terms, and, when inflation accelerated in the early 1970s, the minimum wage was not used to avoid lags between negotiated wages and inflation. Starting in 1976, however, the minimum wage has been adjusted more frequently, and this has had some influence on wage bargaining agreements since unions and employees take into consideration the increase in the minimum wage when bargaining. 124 WINNERS AND LOSERS IN COLOMBIA'S GROWTH In summary, the institutional arrangements prevalent in the Colombian labor market in the 1960s were not well adapted to cope with an acceleration in inflation. When this happened, the real wages and salaries of modem sector workers suffered. In the urban traditional sector, where wage contracts either do not exist or are short term, wages kept up with inflation better. But if institutional factors explain the decrease in real wages in the early 1970s, why did modem sector workers not recover lost ground during the rest of the decade? One possibility is that in all societies income differentials are very rigid and that, even if the underlying demand and supply conditions warrant a decrease in the differential between white-collar and unskilled workers, social traditions and social structure make such a change difficult. The acceleration of inflation, on the other hand, may have made such a change imperceptible, and, once made, underlying market forces made it impossible to return to the status quo ante. The Disappearance of Dualism in the Labor Market Another major structural change in the Colombian economy in the 1970s was the apparent fading of the dual economy characteristics of the labor market. In the 1960s, technological and labor market dualism appeared to be an important feature of the economy, and the two studies written in the decade that dealt with issues of income distribution used a dual economy model to analyze the trends in income distribution. On the basis of data for the period 1958-65, Nelson, Schultz, and Slighton concluded that the "basic hypothesis is that the introduction of a modern sector on top of the traditional craft sector, and the widening of the productivity gap between the two, present the potential for a growing inequality of the distribution of income."14 They had forecast that the major beneficiaries of this development were likely to be the members of the eighth and ninth deciles of the income distribution. In short, what is commonly identified as the "middle" and "lower-middle" class was expected to increase in relative size and increase its share of total income. Exactly the opposite took place, however. This group's share of income declined in the 1970s and the real salaries of middle-class employees and skilled workers in modem industry decreased in some cases and increased little in other cases. One possible hypothesis is that the degree of dualism in the economy diminished. The second study, by Urrutia and Berry, also hypothesized the existence of surplus labor and of traditional and modem sectors to explain the DETERMINANTS OF CHANGES IN INCOME DISTRIBUTION 125 historical trend in income distribution in Colombia. This analysis anticipates that a worsening of the income distribution is particularly probable during the first part of the development process. Afterwards, when the economy emerges from the labor-surplus condition, the distribution is likely to improve."5 The recorded increase in agricultural wages in the 1970s would be consistent with the transition to a development phase with a less dualistic economy. It is possible then that as the economy emerged from the labor- surplus condition, dualism and imperfections in the labor market de- creased. The labor market studies carried out in Colombia before 1970 seem to offer substantial evidence of dualism. For example, Nelson, Schultz, and Slighton found that "the indirect evidence is very strong that the inequality of the distribution of wages outside agriculture has been widening in Colombia over the past fifteen years or so. Most of this change appears explicable in terms of a dual economy hypothesis.""16 In the previous chapter, in contrast, a decreasing gap was discovered between the incomes of workers in the modern and traditional sectors and among workers with different skill levels. This would point to a decrease in dualism in the labor market. This trend is confirmed by a large set of research results that find no evidence of dualism in the labor market in the 1970s. Recently, a number of studies have suggested that when a correction is made for skill level, eamings in the modem and traditional sectors are similar. In fact, Kugler, Reyes, and Gomez found that when they disaggregated the demand for labor between the modern and nonmodern sectors on the basis of size of establishment, public or private status, and other hiring and employment characteristics, there were income differences among sectors. Average wages in the modern sectors are higher than average wages or incomes in the nonmodern sectors. But when adjustment is made for sex, schooling, and experience, the difference in average labor income disappears. The study, therefore, concludes that in 1975 there was sufficient mobility and interaction of markets in Colombia to prevent intersector differentials in labor eamings in urban areas.17 Bourguignon reached the same conclusion in two recent studies."8 In the first, which is based on 1974 data, he concluded that: * Urban poverty is not limited to the traditional sector. * The "residual" nature of the traditional sector seems to be far from a fact. * It is not certain that the modem-traditional dichotomy corresponds to a real segmentation of the urban labor market.9 126 WINNERS AND LOSERS IN COLOMBIA'S GROWTH Finally, Fields also investigated the dualism hypothesis for the urban labor market of Bogota and concluded that it apparently is not highly segmented.20 In summary, a substantial number of investigators, using diverse methodologies, have concluded that the urban labor market in Colombia was quite competitive during the 1970s. The decreasing segmentation of the labor market would help explain some decrease in the dispersion of labor income in the urban areas during the decade. The decrease in segmentation was brought about by the elimination of excess labor in agriculture and by decreasing levels of urban unemployment and under- employment. Import Substitution and Distribution Another feature of the Colombian economy in the postwar period that was thought to contribute to the deterioration of income distribution was the import substitution policy. This hypothesis is at the core of the analysis by Nelson, Schultz, and Slighton. In another study, Urrutia and Berry also show that a period of worsening of the distribution of income in the 1950s coincided with the period of the most rapid import substitution. The attempt to promote exports through exchange-rate policy starting in 1967 implied a moving away from import substitution policies. As the exchange rate became less overvalued, import controls were eased, and average import tariffs were decreased in the early 1970s. By 1973, most import licenses were being approved, and some imports were allowed in sectors that had been completely protected for several decades. The gradual shift away from import substitution promoted some labor- intensive exports and decreased monopoly rents previously generated by quantitative import controls; this decreased the earnings of the top 5 percent of income recipients. As international coffee prices increased, different taxes on coffee exports were raised to avoid excessive increases in agricultural land rents. Thus, before 1975, the exchange rate was devalued in real terms to promote exports and employment.2" After 1975, weekly nominal devaluations were continued, although the real exchange rate was allowed to revalue to some extent, despite very large surpluses in the current account of the balance of payments. Exchange-rate policy, therefore, put downward pressure on income from capital invested in import-substituting industries before 1975; it also forced such industries to strike hard bargains with their workers, who, at the end of the 1960s, earned much higher wages than average. This policy probably DETERMINANTS OF CHANGES IN INCOME DISTRIBUTION 127 dampened the growth of income of families in the two top deciles of the income distribution. Exchange-rate policy during the coffee and drug export booms after 1975 dampened the windfall gains of these exporters through different forms of export taxes and avoided a further deterioration of urban industrial wages by preventing an appreciation of the exchange rate that would have been too harsh on import-substituting industries. The gradual dismantling of import controls, however, kept profits in modern- sector industries at reasonable levels. The Impact of Monetary and Credit Policy During the 1970s, greater freedom was introduced in the capital market. It is hard to say what effect this policy had on income distribution. This greater freedom, however, made credit available to individuals and small firms that could not give bankers impressive guarantees as collateral on their loans. At higher interest rates, bankers found these customers attractive, while when interest rates were controlled and there was excess demand for credit, bankers would ration credit to large customers, with vooQ allarantees, to minimize risk and administrative costs. Small firms and medium-size commercial farmers benefited from liberalization because they could have cheaper and better access to institutional credit. Financial market liberalization, therefore, benefited these economic actors and decreased the profits of large modern-sector enterprises that had access to subsidized credit when such funds were being rationed. This effect of credit liberalization should improve income distribution. Financial liberalization also led to higher interest rates on deposits. This would presumably benefit high-income families, because they are the ones who have higher rates of saving. The impact on income distribution of higher interest rates on deposits is, however, more complicated. Persons with high incomes in Colombia (the top 5 percent of the income distribution) had access to the international financial market and to the curb market; therefore, real interest rates did not increase for their investments. The higher interest rates probably did increase the returns to capital of middle-class families and compensated to a small extent the low growth in real labor income of this sector of the population. To the extent that a freer capital market brought about a better allocation of credit and helped bring down capital-output ratios, financial liberalization also may have contributed to greater employment growth and to an improvement in the incomes of low-income workers. In summary, it is not at all clear that higher interest rates benefited the richest groups in society. 128 WINNERS AND LOSERS IN COLOMBIA'S GROWTH Many people in Colombia identified financial liberalization with income concentration through high profits for financial intermediaries owned by a small group of people. In 1982, however, several of the fastest growing financial groups collapsed; their owners went to jail, and it became clear that the profits of the financial sector, after deductions for bad debts, were not terribly impressive. Although substantial research on the impact on income distribution of greater competition and fast growth in the financial sector is needed, the assumption that many people in Colombia have made-namely, that financial liberalization concentrates income-should not be accepted uncritically. The Impact of Fiscal Policy In Colombia, surprisingly, fiscal policy may have some positive impact on income distribution. Urrutia and Berry offered some evidence suggesting that in the 1960s the tax system was slightly progressive and that government expenditures were redistributive.22 For example, the income of the first decile of the economically active population was twice as high after taxes, government expenditures, and transfers as before, while the share of the top decile was reduced from 48 percent to about 42 percent.23 In his study of the incidence of government expenditures, Selowsky concluded that in education, health, and public services such as water and electricity, government programs benefit the poor in a more than proportional way.24 The improvement in public service levels in the poor barrio of Cali explained in Chapter 3 confirms this. In 1974, after these studies were made, however, the Colombian fiscal system was significantly reformed to make it even more progressive. A thorough and ambitious tax reform was carried out to make taxes more redistributive and to increse the share of government expenditure in the GNP.25 Furthermore, the Development Plan of 1975 determined a shift in government expenditure toward programs that would directly benefit the poorer half of the population. The proportion of expenditures on education was increased, and ambitious nutrition and integrated rural development plans were initiated.26 Although for this research project we attempted to evaluate the impact of some of these new programs on poverty, it still is too early to tell whether they have been as effective as originally planned. There is some evidence, however, that the nutrition plan has been effective in the department of Cauca. Statistically significant evidence shows that for similar populations there are fewer undernourished children in areas where the plan has operated than in those with no program.27 DETERMINANTS OF CHANGES IN INCOME DISTRIBUTION 129 Table 58. Labor Income as a Percentage of GDP in Various Sectors, 1962-78 (current prices) Agricultural Manufacturing Government Year sector industry Construction Trade services 1962 33.9 36.3 71.4 19.5 100.0 1963 35.6 36.5 73.6 18.3 100.0 1964 30.5 38.1 73.8 18.3 100.0 1965 34.3 38.1 74.4 18.3 100.0 1966 32.3 37.8 75.6 18.3 100.0 1967 32.1 38.9 77.4 18.3 100.0 1968 30.2 40.1 78.1 21.5 100.0 1969 31.3 39.9 81.1 21.7 100.0 1970 30.1 41.8 80.6 18.9 100.0 1971 29.7 40.8 78.7 18.6 100.0 1972 27.1 40.2 75.7 16.2 100.0 1973 23.9 35.2 73.7 14.0 100.0 1974 25.5 30.4 73.9 11.9 100.0 1975 23.8 31.2 74.9 12.1 100.0 1976 21.1 31.6 74.5 10.7 100.0 1977 22.9 31.6 73.8 10.9 100.0 1978 25.3 30.9 73.0 12.9 100.0 Source: Banco de ta Republica national accounts. In summary, redistribution through the fiscal system may have been a factor in the improvement in the income distribution after the tax and public expenditure reforms of 1974-75. Distribution of Income from Capital The reduction in the dispersion of urban labor income did not lead to an improvement in the distribution of urban income because of an apparent increase in the income share claimed by capital. The national accounts show this increase in capital income as a proportion of national income (see Table 58). This indicator does not really reflect the share of capital because it includes the incomes of small farmers and self-employed workers, which consist mostly of earnings from labor. Because the incomes of smallholders and self-employed workers in cities rose faster than the overall average in the 1970s, part of the apparent increase in the share of capital as shown in the national accounts was due to increases in labor income of small capitalists. The apportioning of the gross domestic product (GDP) between labor and capital in the Colombian national accounts is fraught with metho- 130 WINNERS AND LOSERS IN COLOMBIA'S GROWTH dological problems. It is clearly not realistic for the labor share in the trade sector to be at 10 percent, as shown in Table 58 for the 1970s. In Colombia, commerce is dominated by small family shops, and one would not expect 90 percent of GDP in the sector to be generated by capital. I do not believe statistical data on labor and capital share as calculated in Colombia have any value, and for that reason have not used this data source in this study. Nevertheless, since many of the salary and wage series show lower growth rates than the national per capita income series, it might be concluded that the share of capital in national income increased somewhat. Because the wealthiest 20 percent of all urban families receive most of the capital income, excluding the imputed rental value of housing, an increase in the share of capital would tend to intensify the concentration of income.28 There is, however, no reliable information on trends in income from capital available. Conclusion Although the empirical analysis of income trends in Colombia during the last two decades is a difficult task, two conclusions can be drawn. First, the overall distribution of income may have improved at the end of the 1970s. Second, it appears that absolute poverty was reduced, because the incomes of poor families rose during the 1970s at a rate matching or exceeding that of per capita income. An independent study of malnutrition by Mohan, Wagner, and Garcia concludes that malnutrition decreased in Bogota and Cali between 1973 and 1978.29 That study reports that probably 25 percent of the cities' population was undemourished because of low income in 1973 and only 12 percent in 1978. Data by occupations also confirm the evidence on increases in the real incomes of the poorest groups. Real agricultural daily wages increased in the 1970s, especially during the latter half of the decade. The real wages of construction workers and other unskilled urban workers also increased. Wages of blue-collar and white-collar workers in the manufacturing industry rose at a slower rate. Although the total income of industrial workers, including employee benefits, seems to have risen between 1965 and 1980, the growth rate was lower than that of national per capita income, which implies a decline in the relative position of this group in the distribution of income. The situation of white-collar workers, and of the middle class in general, is even less encouraging. The absolute level of their DETERMINANTS OF CHANGES IN INCOME DISTRIBUTION 131 real income may have declined, and unquestionably they suffered a loss in relative economic status. These negative findings on the incomes of certain groups of urban workers may be corrected partially when family income is considered. The reduction in unemployment and the increase in the participation of females in the labor force may have increased family income, even if income per worker did not increase for these groups of society. Because the incomes of the wealthiest families probably rose at least as rapidly as national per capita income, as did the income of poor families, the relative position of the so-called middle class clearly worsened. In the years before 1964, in contrast, that class benefited most from the economic development process. The sharp change in the relative position of this group has profound political implications. Finally, while in most developing countries the so-called middle class has great influence on policy, the democratic nature of Colombian politics seems to give this class less weight in decisionmaking. Since elections have to be won, govemment programs and subsidies do not benefit exclusively the top three deciles of the income distribution, and economic policy is not always geared to favor the modern urban sectors. Politics, therefore, may have something to do with the improvement of the Colombian income distribution in the 1970s. Slow income growth for middle-class workers, however, can lead to frustration in a politically strategic group of society. One optimistic conclusion to this study could be that the wage differentials in favor of middle-class workers were adjusted in the 1970s, and that in the future the differences in income growth among different classes will be less marked. Income distribution would improve through movement of people from low-income to higher-income occupations. Notes to Chapter 6 1. Miguel Urrutia and Albert Berry, La Distribucion del Ingreso en Colombia (Medellin: Editorial la Carreta, 1975). 2. The number of persons employed in agriculture increased only from 2,087,949 in 1951 to 2.509.428 in 1964. This figure decreased to 1,546,517 in the 1973 census. 3. There are few serious estimates of the drug trade. These figures come from the only academic publication on the subject. Because the authors assume rather high figures for the capital flight in 1977, when there appeared to be, on the contrary, large illegal influxes of illegal capital, I would guess the drug trade was even smaller. See Roberto Junquito and Carlos Caballero, "Illegal Trade Transactions and the Underground Economy of Colombia," in Vito Tanzi (ed.), The Underground Economy of Colombia in the United States ctnd Abroad (Lexington, Mass.: Lexington Books, 1982). 4. International Labour Organisation (ILO), Hacia el Pleno Empleo (Geneva, 1970). 132 WINNERS AND LOSERS IN COLOMBIA'S GROWTH 5. See Jorge Garcia, "The Impact of Exchange Rate and Commercial Policy on Incentives to Agriculture in Colombia: 1953-58" (Washington, D.C.: International Food Policy Research Institute, November 1980; processed). The rate of the growth of agriculture has been as follows: 1955-60 2.3 1970-75 4.0 1960-65 2.1 1975-80 5.1 1965-70 4.1 Also see La Economia Colombiana en la Decada de los Ochenta (Bogota: Fedesarrollo, 1979), p. 128; and Cuentas Nacionales 1970-79 y Estimacion PIB 1980 (Bogota: Banco de la Republica, 1981). 6. Because children younger and older than those from the bracket defined as containing school-age children can go to school, the children in school can exceed the number of children in the primary-school-age bracket. See World Development Report 1980 (New York: Oxford University Press, 1980). 7. Francois Bourguignon, "The Role of Education in the Urban Labor Market during the Process of Development: The Case of Colombia," (Paper presented at the Sixth World Congress of the Intemational Economics Association, August 1980, Mexico City; processed). 8. Ibid., p. 11. 9. Bemardo Kugler, Alvaro Reyes, and Maria Isabel de Gomez, Education y Mercado de Trabajo Urbano en Colombia: LUna Comparacion entre Sectores Modemo y no Modemo, monograph 10 (Bogota: Corporacion Centro Regional de Poblacion, 1979). 10. If the middle class is classified in terms of education, the size of that class would have increased and it would dominate more deciles of the income distribution. 11. For the development of this process, see Miguel Urrutia, The Development of the Colombian Labor Movement (New Haven: Yale University Press, 1969). 12. The rate of growth of the national consumer price index for employees was as follows: 1970-71 10.1 1961-62 5.5 1971-72 13.2 1962-63 24.5 1972-73 19.6 1963-64 15.7 1973-74 23.3 1964-65 3.8 1974-75 22.1 1965-66 16.8 1975-76 20.1 1966-67 8.9 1976-77 31.3 1967-68 7.6 1977-78 18.9 1968-69 7.0 1978-79 23.9 1969-70 7.3 1979-80 24.9 13. The dates of changes in the minimum wage and its level in real terms (1969=100) are as follows: August 1962 134.6 August 1976 90.3 January 1963 184.4 January 1977 101.9 July 1969 100.0 August 1977 84.3 April 1972 98.7 November 1977 106.3 December 1973 96.8 May 1978 107.5 November 1974 107.3 January 1979 131.0 January 1980 131.7 See Coyuntura Economica, April 1979 and April 1980. 14. Richard Nelson, T. Paul Schultz, and Robert L. Slighton, Structural Change in a Developing Economy; Colombia's Problems and Prospects (Princeton, N.J.: Princeton University Press, 1971). 15. Urrutia and Berry, La Distribucion del Ingreso en Colombia, pp. 21-22. 16. Nelson, Schultz, and Slighton, Structural Change in a Developing Economy, p. 153. DETERMINANTS OF CHANGES IN INCOME DISTRIBUTION 133 17. Urban areas are defined as cities with population exceeding 30,000. See Kugler, Reyes, and Gomez, Educacion y Mercado de Tabajo Urbano en Colombia. 18. Francois Bourguignon, "Pobreza y Dualismo en el Sector Urbano de las Economias en Desarrollo: El Caso de Colombia," Desarrollo y Sociedad, no. 1 Uanuary 1979); and Bourguignon, "The Role of Education." 19. Bourguignon, "Pobreza y Dualismo," p. 69. 20. Gary Fields, How Segmented Is the Bogota Labor Market? World Bank Staff Working Paper no. 434 (Washington, D.C., January 1980). 21. Urrutia and Berry, La Distribucion del Ingreso en Colombia, p. 32. 22. Ibid., p. 202. 23. Ibid., pp. 208-09. 24. Marcelo Selowsky, Who Benefits from Gotemment Expenditure? A Case Study of Colombia (New York: Oxford University Press, 1979). 25. For an evaluation of this tax reform, see Malcolm Gillis and Charles E. McLure, La Reforma Tributaria de 1974 (Bogota: Banco Popular, 1977). 26. Expenditures on education went up from 13 percent of the total central govemment expenditure in 1970 to 18 percent in 1980. 27. Mario Ochoa, "El Plan Nacional de Alimentacion y Nutricion en el Cauca" (Bogota: Fedesarrollo, 1980; processed). 28. Philip Musgrove, Consumer Behatior in Latin America (Washington, D.C.: Brookings Institution, 1978), chapter 2. 29. Rakesh Mohan, M. W. Wagner, and Jorge Garcia, Measuring Urban Malnutntion and Poverty: A Case Study of Bogota and Cali, Colombia, World Bank Staff Working Paper no. 447 (Washington, D.C., 1981). References Ayala, Ulpiano, and Nohra Rey de Marulanda, Empleo y Pobreza. Bogota: Centro de Estudios para el Desarrollo Economico (CEDE) Universidad de los Andes, July 1978. Bejarano, J. A. "Crecimiento, Distribucion y Politica Economica." Paper presented at the Congreso de Economistas de la Universidad Nacional, Melgar, May 1980. Processed. Berry, Albert, and Alfonso Padilla. La Distribucion de Ingresos Provenientes de la Agricultura en Colombia, 1960. CID, Universidad Nacional, Documentos de Trabajo, January/March 1970. Berry, Albert, and Ronald Soligo. Economic Policy and Income Distribution in Colombia. Boulder, Colo.: Westview Press, 1980. Bourguignon, Francois. "Pobreza y Dualismo en el Sector Urbano de las Economias en Desarrollo: El Caso de Colombia." Desarrollo y Sociedad, January 1979. -_____ "The Role of Education in the Urban Labor Market During the Process of Development: The Case of Colombia." Paper presented at the Sixth World Congress of the International Economics Association, Mexico City, August 1980. Processed. Cordoba, Polibio. "Analisis Econometrico de Distribucion de Ingresos." Departamento Administrativo Nacional de Estadistica (DANE), July 1972. DANE, Boletin Mensual de Estadistica. Various issues. . Encuesta Nacional de Hogares 1970. Bogota, 1971. . Encuesta Nacional de Hogares: Fuerza de Trabajo 1971. Bogota, July 1976. . Encuesta Nacional de Hogares 1974: Demografia, Educacion y Fuerza de Trabajo en Bogota, Medellin, Cali, Barranquilla. Bogota, 1976. . Encuesta Nacional de Hogares 1972. Bogota, December 1977. . Encuesta Nacional de Hogares: Fuerza de Trabajo en Cabeceras Municipales 1974. Bogota, 1977. 135 136 WINNERS AND LOSERS IN COLOMBIA'S GROWTH -______ Encuesta Nacional de Hogares 1975: Fuerza de Trabajo en Bogota, Barranquilla, Cali, Bucaramanga, Medellin, Pasto, Manizales. Bogota, 1977. Departamento Nacional de Planeacion, Para Cerrar la Brecha. Bogota: Banco de la Republica, 1975. Fields, Gary. How Segmented Is the Bogota Labor Market? World Bank Staff Working Paper no. 434. Washington, D.C., January 1980. Garcia, Jorge. "La Programacion Industrial y el Arancel Extemo Comun: Un Impuesto al Sector Agricola del Grupo Andino." Coyuntura Economica, July 1981. -_____. "The Impact of Exchange Rate and Commercial Policy on Incentives to Agriculture in Colombia: 1953-58." Washington, D.C.: Intemational Food Policy Research Institute, November 1980. Processed. Gillis, Malcolm, and Charles E. McLure. La Reforma Tributaria de 1974. Bogota: Banco Popular, 1977. Intemational Labour Organisation (ILO). Hacia el Pleno Empleo. Geneva, 1970. Isaza, Rafael, and Francisco Ortega. Encuestas Urbanas de Empleo y Desempleo. Bogota: CEDE Universidad de los Andes, 1969. Isaza, Rafael. "Ocupacion y Desocupacion en Bogota." In Empleo y Desempleo en Colombia. Bogota: CEDE, Universidad de los Andes, 1968. Jaramillo, Helena. "Determinants of Income Differentials after Migration." New Haven: Yale University, 1978. Processed. Kugler, Bemardo, Alvaro Reyes, and Maria Isabel de Gomez. Educacion y Mercado de Trabajo Urbano en Colombia: Una Comparacion entre Sectores Moderno y no Moderno, monograph 10. Bogota: Corporacion Centro Regional de Poblacion, 1979. McKay, Harrison, and others. "Improving Cognitive Ability in Chronically Deprived Children." Science, vol. 200, April 21, 1978. Mohan, Rakesh. People of Bogota: Who They Are, What They Earn, Where They Live. World Bank Staff Working Paper no. 390. Washington, D.C., 1980. Mohan, Rakesh and Nancy Hartline. The Poor of Bogota: Who They Are, What They Do, Where They Live. World Bank Staff Working Paper no. 635. Washington, D.C., 1984. Mohan, Rakesh, M. W. Wagner, and Jorge Garcia. Measuring Urban Malnutrition and Poverty: A Case Study of Bogota and Cali, Colombia. World Bank Staff Working Paper no. 447. Washington, D.C., 1981. Musgrove, Philip. Consumer Behavior in Latin America. Washington, D.C.: Brookings Institution, 1978. Musgrove, Philip, and Robert Ferber. "Finding the Poor." Review of Income and Wealth, vol. 24, September 1978. - "Identifying the Urban Poor: Characteristics of Poverty Households in Bogota, Medellin, and Lima." Latin American Research Review, vol. 14, no. 2, 1979. Nelson, Richard, T. Paul Schultz, and Robert L. Slighton. Structural Change in a Developing Economy: Colombia's Problems and Prospects. Princeton, N.J.: Princeton University Press, 1971. REFERENCES 137 Ochoa, Mario. "El Plan Nacional de Alimentacion y Nutricion en el Cauca." Bogota: Fedesarrollo, 1980. Processed. Okhawa, Kazuchi, and Gustav Ranis. "On Phasing." Paper presented to the Conference on Japan's Historical Development Experience and Contemporary Developing Countries, Tokyo, 1978. Prieto, Rafael, Bill Hanneson, and Marco Reyes. Estudio Agronomico de la Hoya del Rio Suaret. Bogota: Centro de Estudios para el Desarrollo Economico (CEDE)-Corpora- cion Autonoma Regional de la Sabana (CAR), 1965. Prieto, Rafael, and others. Fuentes y Usos de Recursos Financieros en el Sector Agropecuario de Colombia. Bogota: Banco de la Republica, 1976. Prieto, Rafael, "Gasto e Ingreso Familiar Urbano en Colombia." Ensayos ECIEL, no. 4, August 1977. Ranis, Gustav. "Distribucion del Ingreso y Crecimiento en Colombia." Desarrollo y Sociedad, no. 3, January 1980. Sandoval, Clara Elsa de, and Miguel Urrutia. "Distribucion del Ingreso Proveniente de la Actividad Agropecuaria en Colombia." Bogota: Fedesarrollo, November 1980. Processed. Selowsky, Marcelo. Who Benefits from Government Expenditure? A Case Study of Colombia. New York: Oxford University Press, 1979. Urrutia, Miguel. The Development of the Colombian Labor Movement. New Haven, Conn.: Yale University Press, 1969. Urrutia, Miguel, and Albert Berry. La Distribucion del Ingreso en Colombia. Medellin: Editorial La Carreta, 1975. Weisner, Guillermo. "Cien Afios de Desarrollo Historico de los Precios de la Tierra en Bogota." Revista Camara de Comercio de Bogota, nos. 41 and 42. Bogota: Corporacion Centro Regional de Poblacion, 1980. World Bank, World Development Report 1980. New York: Oxford University Press, 1980. I Index Absolute poverty, 6, 70, 93, 94, 102, 113, Centro de Estudios para el Desarrollo 130 Economico (CEDE), 74, 76, 102 Agricultural production, 74, 76, 117-119, Cocaine, 119 121 Coffee: prices. 7, 118, 121, 126; production, Agriculture: development and, 12; technology 7, 13, 16, 118-19, 121. See also Caturra and, 7, 118, 121. See also Income Colombian Central Bank, 74 distribution, agricultural; Labor force, rural; Communications, 17, 98 Real wages, agricultural; Wages, landless Concentration index, 104, 113 agricultural/rural landowner/unskilled rural Consigna (Bogota), 3 Agudelo Villa, Hemnando, 4 Construction workers, 9, 18-25, 55, 67, 68, Altimir, Oscar, 94 97, 98, 106, 107, 130 ANIF. See National Association of Financial Consumer goods, 64, 119. See also Institutions Automobile sales Automobile sales, 46, 52, 119 Consumer price index, 102, 132n Consumption leveLs, 59, 61, 62, 64 Balance of payments, 121, 126 Credit, 74, 77, 127 Banco de la Republica, 76-77 DANE. See Departmento Administrativo Bejarano, J. A., 4 Nacional de Esradistica Berry, Albert, 4, 5. 18, 74, 77. 124, 126, 128 Democracv 3, 131 Bourguignon, Francois, 94, 97, 122, 125 Departmento Administrativo Nacional de Brazil, 118 Estadistica (DANE), 12, 13, 16, 18, 25, 28, 34, 53n, 67, 73, 74, 85, 88, 96, 97, 102, Cali (Colombia), 6, 19, 25, 55-70, 71n 107; Monthly Survey of Manufacturing, 28, Capital equipment, 121 29 Capital income, 9, 73, 76, 77, 119, 129-30 Development. See Agriculture, development Capital market, 127-128 and: Economic policies; "Para Cerrar la Capital-output ratios, 120, 122, 127 Brecha" Carrizosa, Mauricio, 97 Development Plan (1975), 128 Caturra, 118, 121 Domestic industry, 64 CEDE. See Centro de Estudios para el Domestic service workers, 18, 76, 77, 97 Desarrollo Economico Drug exports, 119, 127 Censuses, 5, 74, 76, 77, 115n Dual economy, 12. 124-26 139 140 WINNERS AND LOSERS IN COLOMBIA'S GROWTH Durable goods, 64, 71n Household expenditure, 77 Household surveys, 5, 6, 9, 34, 55, 73-90, Economic growth, 3, 7, 12, 71, 93, 120. See 114 also Income distribution Housing, 62, 64; luxury, 46, 119 Economic policies, 4-5, 7, 17, 90, 94, 118, 120-22, 123, 126, 127-29, 131 icss. See Instituto Colombiano de Seguros Education: employment and, 104, 107, 122; Sociales enrollment rate, 122, 132n; income and, ILO. See International Labour Organisation 56, 57, 96, 97, 98, 99, 103, 115n, 122-23, Imports, 119, 121-22, 126 125; malnutrition and, 56; middle class Import substitution, 121, 126-27 and, 34, 122, 132n; policies, 120, 122-23, Income distribution, 53n, 88, 131; 128; unemployment and, 98-99, 102, 104, agricultural, 74, 76-77, 82; changes in, 74, 107, 113, 122 90, 117-31; deterioration of, 3, 4-5, 6, 77, Electricity, 64 90, 126; disaggregated, 73; improvement of, Elite, 3, 6. See also Capital income; High- 6, 7, 12, 18, 68, 69, 85, 88, 117, 119, 122, income families 130; inequality of, 85, 125; maintaining, Employer federations, 4, 90 13; middle class and, 29, 34, 43, 71, 82, Exchange rate, 120-21, 122, 126, 127 85, 88, 90, 124, 130-31; national, 55, 77, Export boom (1970s), 7, 118, 119, 121, 126, 82, 85, 88; poverty and, 94, 117. See also 127 Capital income; Rural sector; Urban sector; Export subsidies, 121 Wages Income per capita, 5, 12, 53n, 73, 77, 130; Family size, 59, 97-98 increase in, 59-60, 69, 93; national, 6, 13, Farm size, 76, 77 16, 19, 25, 59, 68, 69, 130, 131; Fedesarrollo (research institute), 18, 19; underreporting of, 73-74, 88, 90, 114 Business Opinion Survey, 28 Income surveys, 5-6, 7, 114. See also Ferber, Robert, 96, 106 Household surveys Fields, Gary, 126 Industrialization, 13 Finance, Ministry of, 46 Inflation, 4, 55, 93, 123-24 Financial liberalization, 127-28 Informal sector, 18, 68, 70. See also Domestic Flower exports, 118, 121 service workers; Retail trade sector; Self- Food expenditure, 61-62, 94, 102 employed Foreign debt, 7. See also Balance of payments Instituto Colombiano de Seguros Sociales Fringe benefits, 28, 29, 76, 97, 130 (icss), 53n Fundacion de Investigaciones de Ecologia Interest rates, 127 Humana, 55, 56 Intemational Labour Organization (ILO), 94 Isaza, Rafael, 76 Garcia, Jorge, 90, 114, 121, 130 GDP. See Gross domestic product Japan, 12, 29 Gini coefficients, 76, 77, 82, 85, 91n, 96 GNP. See Gross national product Kugler, Bernardo, 97, 122, 125 Gomez, Maria Isabel de, 97, 122, 125 Govemment. See Ecomomic policies; Public Labor: child, 3; demand for, 17, 118, 119, sector 120, 121, 124; rural, 117-19, 121, 126, Govemment expenditures, 128-29 131n; surplus, 124, 125, 126; urban, Gross domestic product (GDP), 129, 130 119-20. See aso Education; Participation Gross national product, (GNP), 128 rate; Rural sector; Urban sector; Wages Labor market segmentation, 97, 125-26 Hacia el Pleno Empleo (ILO), 94 Land, 73, 76 Hartline, Nancy, 94, 106 Lewis, W. Arthur, 12 High-income families, 55, 56-62, 68, 69-70, Liberal party, 4 71n, 88, 90, 127, 130, 131 Lopez Michelsen, Alfonso, 3, 4-5 INDEX 141 Lower-income families, 3, 4, 5, 6, 7, 18, 55, Real wages, 4, 5, 6; agricultural, 12-13, 16, 56-77, 90, 93, 97, 102, 106, 127, 130. See 18, 93, 117, 130; constant, 55, 120; also Poverty decrease in, 25, 28, 43, 46, 52, 93, Lower-middle class, 124 123-24; of government employees, 123; increase in, 6, 7, 12-13, 17, 19, 25, 28, Macroeconomic policy, 120 43, 46, 59, 61, 68, 69-70, 93, 106, 118, Malnutrition, 56, 114, 116n, 130 124, 130; industrial, 88 Manufacturing sector, 6, 18, 25, 28, 29, 43, Refrigerator ownership, 64 55, 64, 67, 68, 70, 98, 106, 130 Regional differences, 3, 17 Marijuana, 119 Retail trade sector, 25-28, 34, 98, 130 Middle class, 3, 4, 6, 29, 34, 43-52, 70, 71, Reyes, Alvaro, 97, 122, 125 127, 131. See also under Education: Income Rural sector, 3, 9-18, 73-74, 76-77, 82, 85, distribution 88, 117-19, 128. See also Agriculture; Migrants. See Rural-urban migration Poverty, rural; Wages, landless agricultural/ Modem sector, 97, 124, 125, 127, 131 rural landowner/unskilled rural Mohan, Rakesh, 18, 90, 94, 97, 98, 99, 106, Rural-urban migration, 68, 98, 118 114, 130 Musgrove, Philip, 96, 106 Salaries. See Wages National Association of Financial Institutions Salary Council, 18 (ANIF), 4 Savings, 127 National expenditure survey, 85 Schultz, T. Paul, 124, 125, 126 Nelson, Richard, 124, 125, 126 Secondary workers. 60, 61, 65, 70 Nutrition, 55, 56, 57, 61. 64. 94, 114, 128 Self-employed, 113-14, 129 Selowsky, Marcelo, 77, 88, 128 Occupational mobility, 55, 66, 125, 131 Sen, Amartya, 96 Ohkawa, Kazuchi, 12 Service sector, 18, 25, 28, 34, 98, 106. See (Ortega Francisco, 76 also Domestic service workers Ortega, Francisco, 76 Sewerage, 64 Padilla, Alfonso, 74 Single-parent families. 65, 70 "Para Cerrar la Brecha," 94 Slighton, Robert L., 124, 125, 126 Participation rate (in labor force), 98, 102, Social Security Institute (ssi), 25, 28 103, 107, 119-20, 131 Soligo, Ronald, 4, 5 Pastrana, Misael, 3, 4-5 Squatter districts, 55, 62, 64 Peso, 120-21 ssi. See Social Security Institute Petroleum workers, 29 Standard of living, 13, 16, 18, 52, 55, 70, 117 Politics, 3, 4, 71, 131 Poor. See Absolute poverty; Lower-income Tax system, 128, 129 families; Poverty Teachers, 34 Population: growth, 13, 52-53n, 77; decline, Technology, agricultural, 7, 118, 121 7, 12, 120 Tiempo, El (Bogota), 4 Poverty: levels, 93-94, 103; line, 102, 103, Traditional sector, 97, 124, 125 104, 113, 117, 119; measure of, 96; Transportation sector, 98, 121 programs on, 128; relative, 94, 96, 97; rural, 93,94, 117; unemployment and, 113; Underemployment, 12, 126 urban, 93, 94, 97, 98, 99-114, 125. See Unemployment, 19, 98-99, 104, 107, 113, also Absolute poverty 116n, 119, 120, 126, 131 PRESFAM, 102, 114 Unions, 17 Public sector, 46, 120, 122-23 Universidad de los Andes, 74 Public services, 62, 64, 120, 128 Upper class, 46. See also Elite; High-income famlties Ranis, Gustav, 5, 12 Upper-middle class, 43 142 WINNERS AND LOSERS IN COLOMBIA'S GROWTH Urban sector, 3, 18-52, 70, 73, 74, 76, 77, 102, 118, 123, 132n; municipality, 13, 16, 82, 85, 88, 90, 115n, 130-31, 133n. See 53n; rural landowner, 12, 16-17, 18, 52, also Middle class; Poverty, urban; 73-74, 76, 77, 85, 90, 129; rural nonfarm, Unemployment; Wages, unskilled urban! 12, 76; semiskilled, 19; skilled, 17, 29, 43, white collar 46, 52, 69, 106, 123; surveys of, 5-6, 9, 13, Urrutia, Miguel, 18, 74, 124, 126, 128 16, 18, 19, 25, 28, 29, 34, 46, 102-03, 107, 114; unskilled rural, 17, 25, 46, 52, Value added, 76, 119 67, 68; unskilled urban, 6, 9, 16, 17, Voter participation, 70 18-29, 43, 46, 52, 67-68, 69, 70, 104, 106, 107, 118, 120, 124, 127, 129, 131; Wages: blue collar, 28, 29, 34, 43, 46, 67-68, white collar, 4, 6, 25, 28, 29, 34, 43, 46, 69, 106, 130; by department, 13, 16-17; 68, 69, 123, 124, 130, 131. See also Fringe differentials, 125, 131; of females, 16, 52, benefits; Real wages 53n, 60-61, 65, 68, 70, 97, 98, 118, Wagner, M. W., 90, 114, 130 119-20, 125, 131; inflation and, 123-24; Water supply, 64, 72n landless agricultural, 6, 9, 16, 25, 52, 53n, World Bank, 88 67, 68, 76, 85, 90, 117-18; of males, 13, World Development Report 1978, 55; 1980, 16, 61, 97, 98, 125; minimum, 17-18, 94, 122 The full range of World Bank publications, both free and for sale, is described in the Catalog of Publications; the continuing research program is outlined in Abstracts of Current Studies. Both booklets are updated annually; the most recent edition of each is available without charge from the Publications Distribution Unit, Department B, The World Bank, 1818 H Street, N.W., Washington, D.C. 20433, U.S.A. Miguel M. Urrutia, formerly a consultant at The World Bank, is vice-rector, Development Studies Division, at the United Nations University in Tokyo., i - 019 520468 9