LSM- 2 wm OCT. 1980 Living Standard Measurement Studi\ Working Pafper No 2 Poverty and Living Standards in Asia An Overview of the Main Results and Lessons of Selected Household Surveys Pravin Visaria assisted by Sh ker0aldu- P. LSMS Working Paper Series No. 1. Living Standards Surveys in Developing Countries. No. 2. Pbverty and Living Standards in Asia: An Overview of the Main Results and Lessons of Selected Household Surveys. No. 3. Measuring Levels of Living in Latin America: An Overview of Main Problems. No. 4. Towards More Effective Measurement of Levels of Living, and Review of Work of the United Nations Statistical Office (UNSO) Related to Statistics of Levels of Living. No. 5. Conducting Surveys in Developing Countries: Practical Problems and Experience in Brazil, Malaysia, and the Philippines. No. 6. Household Survey Experience in Africa. No. 7. Measurement of Welfare: Theory and Practical Guidelines. No. 8. Employment Data for the Measurement of Living Standards. No. 9. Income and Expenditure Surveys in Developing Countries: Sample Design and Execution. No. 10. Reflections on the LSMS Group Meeting. No. 11. Three Essays on a Sri Lanka Household Survey. No. 12. The ECIEL Study of Household Income and Consumption in Urban Latin America: An Analytical History. No. 13. Nutrition and Health Status Indicators: Suggestions for Surveys of the Standard of Living in Developing Countries. No. 14. Child Schooling and the Measurement of Living Standards. No. 15. Measuring Health as a Component of Living Standards. No. 16. Procedures for Collecting and Analyzing Mortality Data in LSMS. No. 17. The Labor Market and Social Accounting: A Framework of Data Presentation. No. 18. Time Use Data and the Living Standards Measurement Study. No. 19. The Conceptual Basis of Measures of Household Welfare and Their Implied Survey Data Requirements. No. 20. Statistical Experimentation for Household Surveys: Two Case Studies of Hong Kong. No. 21. The Collection of Price Data for the Measurement of Living Standards. No. 22. Household Expenditure Surveys: Some Methodological Issues. Poverty and Living Standards in Asia An Overview of the Main Results and Lessons of Selected Household Surveys The Living Standards Measurement Study The Living Standards Measurement Study (LSMS) was established by the World Bank in 1980 to explore ways of improving the type and quality of household data collected by Third World statistical offices. Its goal is to foster increased use of household data as a basis for policy decision making. Specifically, the LSMS is working to develop new methods to monitor progress in raising levels of living, to identify the consequences for households of past and proposed government policies, and to improve communications between survey statisticians, analysts, and policy makers. The LSMS Working Paper series was started to disseminate intermediate products from the LSMS. Publications in the series include critical surveys covering different aspects of the LSMS data collection program and reports on improved methodologies for using Living Standards Survey (LSS) data. Future publications will recommend specific survey, questionnaire and data processing designs, and demonstrate the breadth of policy analysis that can be carried out using LSS data. LSMS Working Papers Number 2 Poverty and Living Standards in Asia An Overview of the Main Results and Lessons of Selected Household Surveys Pravin Visaria asgigted by Shyamalendu Pal The World Bank Washington, D.C., U.S.A. Copyright () 1980 The International Bank for Reconstruction and Development/THE WORLD BANK 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. All rights reserved Manufactured in the United States of America First printing October 1980 Third printing May 1986 This is a working document published informally by the World Bank. To present the results of research with the least possible delay, the typescript has not been prepared in accordance with the procedures appropriate to formal printed texts, and the World Bank accepts no responsibility for errors. The publication is supplied at a token charge to defray part of the cost of manufacture and distribution. The World Bank does not accept responsibility for the views expressed herein, which are those of the authors 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 presentafion of maferial, 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 concerning the legal status of any country, territory, city, area, or of its authorities, or concerning the delimitation of its boundaries, or national affiliation. The most recent World Bank publications are described in the annual spring and fall lists; the continuing research program is described in the annual Abstracts of Current Studies. The latest edition of each is available free of charge from the Publications Sales Unit, Department T, The World Bank, 1818 H Street, N.W., Washington, D.C. 20433, U.S.A., or from the European Office of the Bank, 66 avenue d'l6na, 75116 Paris, France. When this paper was first printed Pravin Visaria was an economist at the Sardar Patel Institute of Economic and Social Research, Ahmedabad, India. Library of Congress Cataloging-in-Publication Data Visaria, Pravin M. Poverty and living standards in Asia. (LSMS working paper, ISSN 0253-4517 ; no. 2) "October 1980." Bibliography: p. 1. Cost and standard of living-Asia. 2. Income-Asia. 3. Household surveys-Asia. 4. Quality of life-Asia. I. Pal, Shyamalendu. H. Title. III. Series. HD7048.V57 1985 339.4'1'095 85-12020 ISBN 0-8213-0030-X POVERTY AND LIVING STANDARDS IN ASIA: AN OVERVIEW OF THE MAIN RESULTS AND LESSONS OF SELECTED HOUSEHOLD SURVEYS TABLE OF CONTENTS Page No. LIST OF TABLES iii-viii LIST OF FIGURES ix PREFACE x-xii I. INTRODUCTION 1-12 II. COMPARABILITY OF DATA FROM INCOME AND EXPENDITURE SURVEYS WITH OTHER INFORMATION 13-31 III. INDICES OF INEQUALITY AND THE RANKING CRITERIA 32-46 IV. CHARACTERISTICS OF HOUSEHOLD HEADS OR POPULATION IN DIFFERENT DECILES 47-102 V. SALIENT RESULTS OF A MULTIVARIATE ANALYSIS OF HOUSEHOLD EXPENDITURE AND EARNINGS OF EMPLOYEES 103-125 VI. SOME OUTSTANDING RESEARCH ISSUES 126-132 ANNEX 1: A REVIEW OF SURVEYS RELEVANT TO STUDIES OF THE SOCIO-ECONOMIC CHARACTERISTICS OF DIFFERENT INCOME OR EXPENDITURE GROUPS 133-162 ANNEX 2: A CHECKLIST OF MAIN ITEMS ON WHICH INFORMATION IS AVAILABLE ON DATA TAPES USED FOR THE PROJECT 163-165 ANNEX 3: CONCEPTUAL FRAMEWORK OF THE DATA ON (i) EXPENDITURE AND/OR INCOME AND (ii) LABOR FORCE IN THE SURVEYS SELECTED FOR ANALYSIS 166-172 ANNEX 4: SAVINGS RATES FOR HOUSEHOLDS IN DIFFERENT DECILES BY PER CAPITA INCOME AND EXPENDITURE 173-176 ANNEX 5: INDICES OF INEQUALITY 177-178 ANNEX 6: SHARE OF TOTAL EXPENDITURE OR INCOME ACCRUING TO DIFFERENT DECILES 179-187 ANNEX 7: INDICES OF HOUSEHOLD AND PER CAPITA EXPENDITURE OR INCOME BY AGE GROUP OF HOUSEHOLD 188-198 - 11 - Page No. ANNEX 8: ESTIMATION OF ADULT EQUIVALENTS 199-201 ANNEX 9: AGE COMPOSITION AND THE RELATIONSHIP BETWEEN PER CAPITA EXPENDITURE AND VITAL RATES; AN EXPLORATORY EXERCISE FOR GUJARAT, 1972-73 202-206 ANNEX 10: STUDENT-POPULATION RATIOS 207-215 ANNEX 11: CRUDE LABOR FORCE PARTICIPATION RATES 216-224 - iii - LIST OF TABLES Table No. Page No. 1. Salient Features of (A) Surveys Analysed in the Project and (B) Those Assigned Low Priority 5 2. Surveys Examined but Not Considered Suitable for the Project 6 3. Peninsular Malaysia: Monthly Per Capita Income and Expenditure According to the Household Expenditure Survey (HES) and the National Accounts, 1973 17 4. Sri Lanka: Level and Distribution of Monthly (A) Per Capita Income and (B) Private Consumption Expenditure According to National Accounts and the Socio-Economic Survey, 1969-79 18 5. Taiwan: Distribution of Annual Per Capita Personal Income by Source and Private Final Consumption Expenditure, According to National Accounts and the Family Income and Expenditure Surveys of 1968 and 1974 19 6. Savings Rates According to Household Surveys of Three Countries 23 7. Indices of Inequality of Expenditure or Income with Alternative Ranking Criteria and Units of Aggregation i5 8. Percentage of Households in the Same Decile or Quintile According to the Alternative Criteria of Per Capita and Total Household Expenditure or Income 39 9. Indices of Inequality in Expenditure or Income per Equivalent Adult 44 10. PAv8tAga of ouseholdg in the Same Decile or Quintile According to the Alternative Criteria of Per Capita and Per Adult Equivalent Expenditure or Income 46 11. Average Size of Households in Different Deciles with Alternative Ranking Criteria 49 12. Single and Two-Member Households as Percent of Households in the Bottom and Top Deciles with Ranking in Terms of Per Capita Expenditure (PCE) 52 - iv - Table No. Page No. 13. Percentage of Females Among Household Heads in Different Deciles with Alternative Ranking Criteria 55 14. Monthly Household and Per Capita Expenditure or Income According to the Sex of the Household Head, in Various Countries/Regions 57 15. (A) Percentage of Females Among Heads of Households by (i) Age of Heads and (ii) Size of Households and (B) Marital Status Distribution of Female Heads in Different Countries/Regions 59 16. Percentage of Females Among All Persons in House- holds in Different Deciles, with Alternative Ranking Criteria 60 17. Age Dependency Ratios in Different Deciles with Alternative Ranking Criteria 65 18. Children Aged 0-14 as Percent of Total Population in Different Deciles with Alternative Ranking Criteria 68 19. Percentage of Illiterates (or Those with No Formal Schooling5 Among Persons Aged 15 and Over in the Bottom and Top Deciles of Households (with Ranking in Terms of Per Capita Expenditure) 76 20. Percentages of Children Not Going to School Among Persons Aged 5-19 in the Bottom and Top Deciles of Households Ranked According to Per Capita Expenditure 79 21. Labor Force Participation Rates of Population Aged 10 and over in Different Deciles According to Alternative Ranking Criteria, Both Sexes 86 22. Incidence of Unemployment Among Persons Aged 10 and Over (or 15 and Over) in Different Deciles According to Alternative Ranking Criteria, Both Sexes 90 23. Percentage of Households for Which Summary Variables had Zero or Undefinable Values 110 24. Number of Categories Considered for Each Variable in the Multiple Classification Analysis of Per Capita Expenditure of Households 113 -v - Table No. Page No. 25. Results of Multiple Classification Analysis of Per Capita Expenditure of Households 115 26. Results of Multiple Classification Analysis of Per Capita Expenditure in Rural Maharashtra Before and After the Exclusion of the Aged- Dependency Ratio from Among the Explanatory Variables 118 27. Results of Multiple Classification Analysis of Earnings of Employees 124 Annex 4: 1. Peninsular Malaysia: Ratio of Per Capita -Expenditure to Per Capita Income in Different Deciles of Households Ranked According to (A) Per Capita Income and (B) Per Capita Expenditure, All Areas, 1973 174 2. Sri Lanka; Ratio of rer Capita Expenditure to Per Capita Income in Different Deciles of Households Ranked According to (A) Per Capita Income and (B) Per Capita Expenditure, All Areas, 1969-70 175 3. Taiwan: Ratio of Per Capita Expenditure to Per Capita Income in Different Deciles of Households Ranked According to (A) Per Capita Income and (B) Per Capita Expen- ditura, All Areas, 1974 176 Annex 6: 1. Gujarat: Decile Limits and Percentage Share of Expenditure for Deciles of Households According to Alternative Criteria (and Shares for Corresponding Deciles of Population), 1972-73 180 2. Maharashtra: Decile Limits and Percentage Share of Expenditure for Deciles of Households According to Alternative Criteria (and Shares for Corresponding Deciles of Population), 1972-73 181 - vi - Table No1 Pa12 No. 3. Nepal: Decile Limits and Percentage Shares of Expenditure for Decile of Households Ranked According to Alternative Criteria (and Shares for Corresponding Deciles of Population), Eleven and Seven Towns, 1973-74 and 1974-75, Respectively 182 4. Peninsular Malaysia: Decile Limits and Income Shares for Deciles of Households Ranked According to Monthly Per Capita Income and Income Shares for Corresponding Decile of Population, 1973 183 5. Sri Lanka: Decile Limits and Percentage Shares of Income or Expenditure for Deciles of Households Ranked According to Alternative Criteria (and Shares for Corresponding Deciles of Population), 1969-70 184 6. Taiwan: Decile Limits and Percentage Shares of Income or Expenditure for Deciles of Households Ranked According to Alternative Criteria (and Shares for Corresponding Deciles of Population), 1968 185 7. Taiwan: Decile Limits and Percentage Shares of Income or Expenditure for Deciles of Households Rank@d Aceording to Altamntivg Criitgri (dA Shares for Corresponding Deciles of Population), 1974 186 Annex 7: 1. Gujarat: Decile Limits and Percentage Share of Expenditure for Deciles of Households Ranked According to Alternative Criteria (and Shares for Corresponding Deciles of Population), 1972-73 189 2. Nepal: Per Capita and Total Household Expenditure (in Nepali Rupees) by Age Group of the Head of Household, and (A) Eleven Towns, 1973-74 and (B) Seven Towns, 1974-75 190 3. Peninsular Malaysia: Per Capita and Total House- hold Expenditure (in Malaysian Dollars) by Age Group of the Head of Household, 1973 191 - vii - Table No. Page No. 4. Sri Lanka: Monthly Per Capita and Total Household Expenditure and Income (in Sri Lankan Rupees) by Age Group of the Household Head, 1969-70 192 5. Taiwan: Monthly Per Capita and Total Household Expenditure and Income (in N.T. Dollars) by Age Group of the House- hold Head, 1968 and 1974 193 Annex 8: 1. Weights for Estimating Equivalent Adult Consumers 200 2. Measures of Association Between Household Size and the Number of Adult Equivalents in Households for Some Data Sets 201 Annex 9: 1. Birth and Death Rate per 1,000 Population by Household Expenditures, Urban India, 1959-61 203 2. Birth, Death and Natural Increase Rates per 1,000 Population by Monthly Per Capita Expenditure, Rural and Urban India, 1963-65 204 3. Birth, Death and Natural Increase Rates (per 1000 Population) Expected for Different M.P.C.E. Deciles Without Any Difference in Sex- Age-Specific Rates, Rural and Urban Gujarat, 1972-73 206 Annex 10: 1. Student-Population Ratios by Sex and Age Group and Decile of Per Capita Expenditure, Rural and Urban Gujarat, 1972-73 208 2. Student-Population Ratios by Sex and Age Group and Decile of Per Capita Expenditure, Rural and Urban Maharashtra, 1972-73 209 3. Student-Population Ratios by Sex and Age Group and Decile of Per Capita Expenditure, Eleven and Seven Towns of Nepal, 1973-74 and 1974-75, Respectively 210 - viii - Table No. Page No. 4. Student-Population Ratios by Sex and Age Group,, and Decile of Per Capita Expenditure, Sri Lanka, 1969-70 211 5. Student-Population Ratios by Sex and Age Group a and Decile of Per Capita Expenditure, Taiwan, 1968 and 1974 -212 6. Student-Population Ratios by Sex and Age Group and Decile of Per Capita Income, Taiwan 1968 and 1974 213 7. Student-Population Ratios by Sex and Age Group and Decile of Total Household Expenditure, Taiwan 1968 and 1974 214 8. Student-Population Ratios by Sex and Age Group and Decile of Total Household Income, Taiwan 1968 and 1974 215 Annex 11: 1. Crude Labor Force Participation Rates for Different Deciles with Alternative Ranking Criteria (Males/ Females/Persons) 217 - ix - LIST OF FIGURES Figure No. Page No. 1. Average Household Size 50 2. Percentage of Females Among Household Heads 56 3. Percentage of Females Among All Persons 61 4. Age Dependency Ratios 66 5. Percentage of Children (0-14) Among All Members 69 6. Illiteracy Rates 75 7. Student Population Ratios by Sex for Age Group 10-14 80 8a-f Labor Force Participation Rate for Persons Aged 10 and Over 86a-f 9a-d Incidence of Unemployment Among Persons Aged 10 and Over 90a-d Annex 7: 1. Gujarat State: Average Household Size and the Index of Total and Per Capita Monthly Expenditure of the Household, in Rural and Urban Areas, 1972-73 194 2. Maharashtra State: Average Household Size and tha Indgy of Total and Per Capita Monthly Expenditure of the Household, in Rural and Urban Areas, 1972-73 195 3. Nepal: Average Household Size and the Index of Total and Per Capita Monthly Expenditure of the Household, in Eleven Towns (1973-74) and Seven Towns (1974-75) 196 4. Peninsular Malaysia and Sri Lanka: Average Household Size and the Index of Total and Per Capita Monthly Expenditure of the Household 1973 and 1969-70, Respectively 197 5. Taiwan: Average Household Size and Index of Total and Per Capita Annual Expenditure of the Household 1968 and 1974 198 Annex 11: la-e Crude Labor Force Participation Rates 220-224 Preface This paper seeks to provide a summary of the main findings of an analysis of selected household surveys undertaken in the joint ESCAP-IBRD Project on Income Distribution in Asia, undertaken by the Development Research Center of the World Bank. It sheds light on the socio-economic characteristics of different income and expenditure groups in the states of Gujarat and Maharashtra in India, 18 towns or development centers of Nepal, Peninsular Malaysia, Sri Lanka and Taiwan on the basis of surveys which were undertaken between 1968 nd 1974. Ot the bagia of thig information an attempt is made to suggest some guidelines for future surveys that may be undertaken to measure the living standards of the population in different countries. Six draft papers containing the basic material for much of what is presented here are available at the Development Research Center. Although they .are preliminary and cannot be quoted without clearance from the author, they are available on request to interested scholars. Following some revisions, the bulk of the material included in these papers will soon be made available for public distribution in the World Bank Staff Working Papers. The draft papers are listed below. Working Paper No. 1: "Living Standards, Employment and Education in Western India, 1972-73," 1977. Staff Working Paper No. 417: "Poverty and Unemployment in India: An Analysis of Recent Evidence," April 1980. - xi - Working Paper No. 3: "Size of Land Holding, Living Standards and Employment in Rural Western India, 1972-73," 1978. Working Paper No. 4: "Incidence of Poverty and Characteristics of the Poor in Peninsular Malaysia, 1973," 1980 Working Paper No. 5: "Some Aspects of Relative Poverty in Sri Lanka, 1969-70," 1980. Working Paper No. 6: "The Incidence of 'Absolute' Poverty in Sri Lanka, 1969-70," 1979. This research project has drawn heavily on the advice and cooperation of many friends and colleagues. The directors of statis- tical agencies in the various countries have been very generous in providing access to their data and, even more importantly, in helping clarify frequently ensuing questions. They include: Mr. P. B. Buch, former Director of the Bureau of Economics and Statistics in Gujarat, and subsequent, Director of the Gujarat Computer Centre, who provided exemplary leadership to a large team of co-workers in the arduous task of editing and tabulating the data for both Gujarat and Maharashtra; Mr. M. A. Telang and Mr. S. M. Vidwans, both former Directors of the Bureau of Economics and Statistics in Maharashtra State; Mr. B. B. Pradhan and Mr. B. P. Risal of the Research Department of the Nepal Rastra Bank; Mr. Ramesh Chander and Mr. Khoo Teik Huat, former and current Chief Statisticians of Malaysia; Mr. L. N. Perera and Mr. W. A. A. S. Peiris, former and current Directors of Census and Statistics in Sri Lanka; and Mr. T. T. Huang, the present Director of Statistics in Taiwan. - xii - Among other friends and colleagues who have given valuable advice and assistance at several stages of the project, special mention must be made of Sudhir Bhattacharya, V. M. Dandekar, M. L. Dantwala and Suresh Tendulkar, all in India; and Montek Ahluwalia, B. S. Minhas, Graham Pyatt and T. N. Srinivasan, all colleagues at the Development Research Center of the World Bank. Some useful conments and suggestions have been made by Surjit Bhalla, Oey A. Meesook and V. V. Bhanoji Rao at the Bank and Bryan Boulier at Princeton University. During the first two years of the project, from March 1975 to November 1976, the Statistics Division of ESCAP provided various facilities for the project; and Mr. T. V. Viswanathan and Mr. K. 0. Clark were very helpful. Several colleagues have assisted in the statistical work of the project. They include: Robert E. Sterrett, Jr., R. Murti Pemmarazu, Cynthia Hwa, Larry Swertloff, Shail Jain, Pramilla Burger, and most imp6rtantly, ghyamalendu Pal, whose high deAlcation and devotion to the project has resulted in a contribution that is acknowledged on the title page. As I acknowledge my deep gratitude to persons noted above, the responsibility for all errors remains solely mine. October 1980 Pravin Visaria I. Introduction The origin of the Joint ESCAP-IBRD project on the "evaluation of Asian data on income distribution" can be traced to the various meetings held to evaluate the feasibility of reorienting planning priorities towards "redis- tribution with growth." Among the statistical priorities for such a reori- entation, the first place was. assigned to "a clear identification of poverty groups and their economic and social characteristics" as "an indispensable component of policy diagnosis and design."- The project was simultaneously intended to implement the mandate of Income Distribution Division of the Development Research Center (IRC) which was to help improve the data base on income distribution in developing countries by evaluating the existing surveys as a basis for suggesting better procedures for data collection and analysis. In spite of the project's formal title, its primary focus has continued to be an identification of the poor and the key socio-economic and demographic characteristics of the poor in particular, and of different income and/or expenditure groups in general. The project was launched with the belief that many of the income and expenditure surveys conducted in the developing countries since the end of the Second World War had the potential to shed considerable light on the living standards of different groups of population; but for various reasons, thil p9tSRChi rtWin94 U~nexploited. An effort was to be made, therefore, to supply the missing stimulus for more purposive elaborate cross-tabulations and analyses beyond those customarily attempted to meet the immediate limited objectives under the given constraints. I/ C. L. G. Bell and John H. Duloy, "Statistical Prioritie," in Redistribution *itkGrowth- Hollis Chenery et al. (London: Oxford University Press, 1974), p. 237. - 2 - During the first year of the project/ , contact was established with the national statistical agencies and other interested scholars in 13 developing countries of the ESCAP region, namely, Bangladesh, Hong Kong, India, Indonesia, Iran, Korea (Republic of), Malaysia, Nepal, Pakistan, Philippines, Singapore, Sri Lanka,and Thailand2/ and in the economy of Taiwan. In each case, an attempt was made to identify a relatively recent survey with questions on income and/or expenditure, as well as information on relevant demographic and socio-economic characteristics (such as education, and labor force and employment particulars) of all members of the households (not only the heads of households). Designed to parallel a similar project on income distribution in Latin America (undertaken jointly by the Statistics Division of ECLA and the DRC), the project was launched jointly with the Statistics Division of ESCAP. The principal researcher of the project was based in Bangkok during the initial phase of the project (April 1975-October 1976) and in Washington.D.C. thereafter. The Statistical Offices of Afghanistan and Brunei had communicated an interest in obtaining the services of an expert to plan a study and/or a survey on the issues within the scope of the project. These requests were conveyed to the relevant authorities in the World Bank and the ESCAP, but could not be met within the mandate of the project. -3- A. Initial Aspirations A review of the initial phase of the project, conducted in early MarAk 17, led to a decision to work on the data for eight countries, Hong Kong, India, Malaysia, Nepal, Philippines, Singapore, Sri Lanka and Thailand,- and for the economy of Taiwan. Considering the two states of India (Gujarat and Maharashtra, with 1971 populations of 26.7 and 50.4 million, respectively) as distinct from the two town groups of Nepal (surveyed in two consecutive years), and counting the 1968 and 1974 surveys of Taiwan separ- ately, the project was designed to undertake an analysis of 12 data sets. In addition, there was some hope, subsequently belied, that we may be given access to the 1974 survey data for the Republic of Korea. For four countries (Indonesia, Iran, Pakistan, and Bangladesh), the available surveys were not considered suitable for the project. The main criteria that governed our selection of surveys for analysis can be summarized briefly. We generally preferred to work with national surveys, although for Nepal we accepted data sets for selected "towns," and in India, we decided to work with the data for two states because of the large size of the sample. We also preferred a survey conducted over an entire year rather than one conducted during only one or two months because in the former, the seasonal element could be averaged out for the aggregate sample (i.e., for the annual average, but not for the distribution). Likewise, based on an a priori hypothesis and earlier experience, and as confirmed in 1/ Hong Kong and Singapore were recognized as atypical of Asia. Yet we thought that a comparative analysis of the situation in the two "city-states" would be useful and interesting. the course of the project (discussed below), surveys which used a shorter reference period for collecting data on income or expenditure seemed preferable to those which used a longer reference period and encountered problems of recall and other response errors.-/ B. Revised Plans As the work on the data sets proceeded, the gross ubderestimation of the resource requirements of a project based on primary data became- evident. There were many painful reminders of the celebrated Murphyts Law. As a result, during a review of the project in April-May 1978, it was decided that the analysis of surveys for four countries (Hong Kong, Philippines, Singapore, and Thailand), which had been found deficient in the course of preliminary work, would be assigned low priority. Consequently, the scope of the project has been restricted to five countries (six regions and eight data sets). Table 1 lists some key features of (a) the surveys that have been worked on in the project, and (b) those that were assigned low priority in May 1978. Table 2 aummari2es the key featuras of surveys that were examined but not considered suitable for the project. Annex 1 reviews at some length the surveys that were considered for the project, and outlines our rationale for selecting and rejecting particular surveys.2 1/ An overriding practical criterion was the availability of survey data, at least on punch cards, which could be taped. It was felt that the punching and verification of data directly from the schedules would involve very high cost. 2/ Annex 1 also includes a brief discussion of a few data sets that were proposed to be taken up for analysis "later" (i.e., after the initially- selected surveys had been analyzed), subject to the availability of resources and in light of lessons learnt from the analysis of a few surveys. Table 1 SALIENT FEATURES OF (A) SURVEYS ANALYSED IN THE PROJECT AND (B) TROSE ASSIGNED L-OW PRIORITY Reference Area Title of the Survey Year Sample Size PrimaEy Emphasis (A) Surveys Analysed in the Project 1. India* (a) Gujarat 19/2-73 8978 State (5443 Rural) National Sample Survey (3535 Urban) Level and structure of 27th Round employment and (b) Maharashtra (State Samples only) 1972-73 16494 consumption expenditure State (5357 Rural) (11137 Urban) 2. Malavsia Household Expenditure 1973 7273 Expenditure, income (Peninsular) Survey and labour force, housing 3. Nep al (a) Eleven Towns 1973-74 4438 Expenditure, income Household Expenditurehousing (b) Seven Towns Survey 1974-75 2243 4. Sri Lanka Socio-Economic Survey 1969-70 9694 Expenditure, income, labour force, fertil- ity, housing 5. Taiwan Family Income and 1968 2776 m Income and expenditure Expenditure Survey 1974 5272 (B) Surve-ye or Data Sets InitiaUy Included in the Project but not AnaZeed 1. Hong Kong 1971 Population Census January 8546 Population, labour One Percent Sample 1971 force, housing 2. Korea (Republic) Employment and Income November 129,000 Level and structure of Survey 1974 employment and income 3. Philippines National Demographic Survey May 7237 Fertility, labour 1968 force and income 4. Singapore Traffic Survey of (a) Vehicle Owning and (a) Oct-Nov 1974 2645 Traffic patterns (b) Other Households (b) Jan. 1975 4869 5. Thailand Socio-Economic Survey 1971-73 8748 @ Income and expenditure Tabulations were done by the Cujarat Computer Center. Pouseholds (2,172) surveyed in Bangkok-Thonburi in 1971 have been excluded. Table 2 SURVEYS EXAMINED BUT NOT CONSIDERED SUITABLE FOR THE PROJECT Reference Sample Country Survey Year Size Remarks 1. Indonesia Cost of Living Surveys 1968-70 7,988 Data not edited; their quality (11 major cities) could not be assessed. The Central Bureau of Statistics wouLd need substaxtial technical assistance to compile the re- quisite tables. 2. Iran (a) Bank Merkazi Urban Expenditure Surveys 1973-74 5,000 Documentation avallable in Persian only!I (b) Iran Statis- Rural and Urban Expenditure 1972-73 Unknown QuaLity of fieldwork doubtful; tical Center Surveys 1973-74 documentation avallable in Persian onlyl 3. Pakistan Household Income and 1968-72 Around The information on tape did not Expenditure Surveys 7,000 include sex, age or education of each year workers or population. Occupa- tion and employment status cannot be interpreted without knowledge of other characteristics. 4. Bangladesh Household Expenditure Survey 1973-74 Unknown Same as for 3. 5. Philippines 1975 Integrated Census of Dec 1975 400,000 Quality of data was not known. Population and Its Economic It seemed unlikely that the data Activities, Phase II would be available for analysis during the life of the project. 1/ The English translations could not be obtained despite reminders. -7- C. Some of the Difficulties in Analysis of Primary Data The factors that compelled a revision of our initial plans include a certain amount of naivete about the time and resource costs of working with primary data on the computer. To name a few contributory factors, first the estimates of computer costs given to us by the "experts" whom we consulted had a range of 1:10. We adopted something like a median value and found that the estimate involved a gross understatement of the costs. Secondly, we did not make adequate allowance for the fact that even with the best programming assistance available (which is both difficult and expensive to obtain), almost every attempt at working with the computer is liable to errors, some human and some non-human (the latter include the so-called parity errors reported by the computer because of some dust or foreign particles on the tape). Thirdly, due allowance is seldom made for the turn-around time required in testing or executing any job. Fourthly, research projects seldom proceed mechanically according to the initial plan. The objectives and preferences regarding analytical procedures are almost always redefined due partly to the learning process involved in all research; thus the time schedule and budget suffer the consequences./ An equally (if not more) important factor contributing to the delays is the altogether different set of standards used to evaluate the consistency of primary data that becomes feasible when one has access to the computer.2/ Such consistency checks and editing are virtually impossible when the processing of data is attempted manually: The problem arises because If Any interesting finding based on any one data set stimulates a search for corresponding evidence in other data sets. Although exciting and useful, the process is not without costs. 2/ Some of the basic consistency checks have been noted in my paper entitled "Some Aspects of Relative Poverty in Sri Lanka, 19§9-70" (August 1979), Annex 1. the resources available to the countries or authorities simply do not permit the appropriate consistency checks when the data are not processed on the computer. This lesson was learnt in the course of our work on data that had been extensively tabulated (e.g., the Socio-Economic Survey of Sri Lanka 1969-70). To illustrate, due to punching or coding errors, even the number of sample households on the tape did not quite tally 1/ vith thA number raDortedly interviewed,- but such problems are either not caught or not corrected unless one processes data on the computer. It may be easily argued that even while working with a computer, one must tolerate a certain margin of error in such cases and reject some sample observations. However, many of the inconsistencies caused by such errors affect the data for several additional observations; although it is not easy to decide when one must halt the editing process, it obviously cannot go on indefinitely. The problems would be more manageable (although delays would still be unavoidable) if the agency, which collects the data and has access to the original records, were to undertake careful editing and checking to verify implauail attriag.1 1/ In the case of Nepal, a fortunate coincidence permitted a comparison of the number of records on the tape provided to us with an earlier count and led to the discovery that some 25 percent of the records for the largest "town" (Kathmandu) had accidentally been left out of the tape. The identification of the missing households, relevant records, and the creation of a new tape required a fair amount of patience and perseverance. 2/ For the state of Gujarat in India, the fact that the computer center editing the data had access to the original schedules within the same building, led to the adoption of rather stringent consistency checks and verification of almost all implausible entries. Of course, this highly commendable approach was necessarily costly in terms of computer time and required a departure from the initially envisaged time schedule. - 9 - In any case, a research project undertaking the analysis of primary data must make adequate allowance for the large amount of time and computer facilities required to "clean" the data, if unnecessary frustrations over the non-attainment of impossible targets are to be avoided. D. Objectives of the Project The remainder of this paper is intended to provide a summary of the findings of the research project from the analysis of eight data sets examined to date. Our analysis of the data for Nepal and the economy of Taiwan has not been as intensive as that for other data sets; therefore, the assertions relating to them must be viewed as somewhat tentative and subject to revision. The scope of the discussion is deter- mined by the contents of the schedules used in different surveys, which are summarized in a checklist form in Annex 2. Since the concepts and definitions used for some of the key items of information can differ considerably between countries, Annex 3 outlines at some length the salient features of the procedures used to collect information on the economic activities of households (and/or their members) and their income and/or expenditure. The extent of comparability of these procedures and the resulting data is discussed in the paper as a preliminary to the presentation of findings on each issue. To put the subsequent discussion in proper perspective, it should be noted that although the project was not designed to test any well-defined hypotheses or to arrive at specific policy prescriptions,L 1/ One must recognize the limitations of any effort to make policy recommendations for any specific country on the basis of a cross- sectional analysis of data sets for a few countries. The historical context and the larger socio-political background must necessarily receive adequate attention in policy formulation. -10- it was intended to shed some light on some major issues which we had previously identified (in March 1976). (It was presumed that a proper descriptive analysis of the prevailing conditions would provide the foundation for exploring various alternative strategies of intervention.) And an important preliminary to the examination of our data relating to these issues was their assessment through comparisons with national accounts data (wherever available) and an evaluation of the plausibility of reported savings rates, where the surveys had gathered data on both incoma And ampenditure. The issues we sought to examine are listed below: (1) The extent to which the target groups, identified according to the criterion of per capita expenditure or income, would be identical to those identified according to the criterion of aggregate household ampanditurA ar iicomaa (2) The extent to which a household's living standard or the level of per capita expenditure observed in surveys may be associated with a particular stage of the life cycle; (3) The proportion of "unemployable" persons, i.e., the very young and the old among the poor, which would determine the extent to which poverty can be ameliorated only through a policy of transfers or social assistance rather than through the creation of more employment or measures to raise the productivity of jobs; (4) The extent to which women are overrepresented among the poor; (5) The extent to which children in poor families are able to avail of schooling facilities; -11- (6) Differences in the educational attainment and skills of persons in working age groups according to the level of per capita expenditure; (7) Differences among income groups in the proportion of persons of working ages actually participating in economic activity; (8) The extent of association between open unemployment and poverty or living standards, and the extent to which incomes of the poor can be augmented through the creation of work opportunities in the public works programs, etc.; (9) The extent to which the inequality in living standards is explained by differences in the quality or characteristics of employment as reflected in the occupation, industry, and status of workers; (10) Housing conditions of different expenditure groups; (11) The consumption pattern of different expenditure or income groups; (12) The extent to which the ownership of assets rather than different earnings contribute to inequality. The first nine of these issues have been explored in considerable detail for most of the data sets, whereas the last three could be examined for only one or two data sets. Only the Sri Lankan and Nepal surveys had some detailed information on the housing characteristics of surveyed households; of these, the Nepal data have not been examined. The pattern of consumption has been examined in some detail for Sri Lanka and Peninsular Malaysia, with an emphasis on the implications of alternative ranking criteria; for Sri Lanka, the level of caloric intake implied by the reported quantities of food consumption has been explored. Our examin- ation of the contribution of assets to the disparities in living standards -12- is limited to rural areas of Gujarat and Maharashtra, for which data were available on the land "possessed" (i.e., owned land plus leased in land minus leased out land) by the respondent households. Overall, our study has demonstrated that the existing household surveys, conducted for different objectives, have considerable potential for shedding light on the living standards of the people and the disparities in them, but this potential often remains untapped. -13- II. Comparability of Data from Income and Expenditure Surveys with Other Information The evaluation of the quality of data collected in the income and/or expenditure surveys in different countries naturally entails the assessment of their comparability with similar information available from other sources. The possible comparisons (with due allowance for the time interval, if any, and other differences of methodology between various sources) include: (a) the consistency between the demographic (including school enrollment rates and literacy) and economic (relating to employment and unemployment) characteristics of the population reported by the survey and the population census; (b) differences between the survey estimates of the level and ditributioni (by soutca, a.., wages, property income, etc.) of personal income and those based on national accounts; (c) an assessment of the plausibility of savings rates implied in the survey responses, relative to alternative estimates. A. Data on Demographic and Economic Characteristics In the course of our analysis of various data sets, we have attempted all possible comparisons with the population census data. Typically, it is found that the surveys perform remarkably well in ob- taining estimates of demographic characteristics similar to those observed in the censuses, particularly when some allowance is made for any selection (i.e., restriction) of the survey sample from the universe (e.g., the exclusion of institutional private households, or in the Sri Lankan Socio- Economic Survey of 1969-70 the exclusion also of single member households).1 I/ Some plausible explanations are: (a) the stability of major demographic characteristics, (b) the successful use of random sampling procedures in surveys, and (c) the use of essentially the same or similar interviewers in surveys and censuses. -14- There does, however, appear to be some (unintentional) selection in surveys, of households with a higher level of literacy and education or school-enrollment ratios than reported by the population censuses. With respect to the economic characteristics of the population, such as (a) the laboF force participation rates, (b) the incidence of unemployment or (c) the structure of employment (i.e., the distribution of the employed by occupation, industry, and class of worker or status), one often observes relatively larger differences between surveys and censuses. It seems difficult to judge which source should be viewed ag the more reliable; but the lack of a uniform or comparable cQnceptual framework, and the possible seasonal variations in the level and pattern of economic activity, particularly in rural areas, seem to contribute to the observed differences. Even approximate adjustments for differ- ences in conceptual framework reduce the magnitude of differences in the estimates of economic characteristics.- One might venture to suggest that if the surveys and censuses were to collect data on the "usual" economic characteristics, which might be presumed to be more relevant to the understanding of the levels of income or consumption, in addition to the information on "current" economic activities of the reference week preceding the interview, the observed differences might shrink significantly..i 1/ For an illustration, see the discussion of "Comparability of the Socio-Economic Survey with the 1971 Census of Sri Lanka," Annex II to Some Aspects of Relative Poverty in Sri Lanka, 1969-70 (August 1979). 2/ The Indian census experience of collecting data on the economic activities of the population suggests, however, that unless the wording of questions and definitions are kept uniform, it becomes very difficult to compare the resulting data. The same lesson seems to emerge from a comparison of the Indonesian censuses of 1961 and 1971. -15- B. Comparisons with National Accounts Where our data sets relate to entire countries (rather than parts of a country), the survey estimates of the level and/or distribution (by source) of personal income or consumption can be compared to those available in the national accounts. The validity of such comparisons, however, is not beyond doubt even when the undistributed profits of corporations can be excluded from GDP at factor cost.11 There are differences in coverage because the institutional private households (e.g., residents of hostels or other institutions including perhaps the armed forces in camps or barracks) are excluded from the scope of surveys. Secondly, a part of the income of unincorporated enterprises that is reinvested for productive purposes (i.e., not consumed and not drawn upon) is probably excluded from the reported income of the household, unless an attempt is made to obtain a profit and loss account for the enterprise-- an extremely difficult proposition in any survey. The national accounts estimates normally make some adjustments for that purpose. More importantly, for these unincorporated enterprises, many of which can be termed the "unorganized" or informal sector of manufacturing units or service establishments (including hotels, restaurants, transportation and trade firms), the estimates of national income are obtained through some guestimates of the working force engaged in such activities and the, earnings per worker. The importance of these activities in various countries is naturally different, but the 1/ If the survey data refer to income after taxes, direct taxes should also be excluded from the GDP figure before attempting any comparison. -16- basic yardsticks are derived from surveys similar to income and expenditure 1/ surveys;- furthermore, since surveys are difficult to conduct and consicterab.1e time lapses before the results become available, the empirical basis of these yardsticks often becomes rather remote. In Sri Lanka, for example, items, whose current value is "assumed to grow at a trend rate" (related to the rate of growth of the population),formed 15 percent of the estimated GDP in1 1970.-/ Therefore, one must recognize the limitations of any comparisons of income and expenditure survey data with national accounts. With the caveats noted above, we can examine the comparability of the survey estimates of personal income and expenditure with national accounts aggregates for Malaysia, Sri Lanka, and the economy of Taiwan. Tables 3, 4, and 5 present the available information. The scope for comparing the Malaysian data is limited in relation to that for Sri Lanka or the Taiwan economy; nevertheless, the following conclusions, all consistent with general impressions, can be drawn from these tables. 1/ In Taiwan, one of the main purposes of the Family Income and Expenditure Survey is said to be "to provide the basic information for the estimation of national income accounts of private consumption." See: Jo-ho Hsu, "The Reliability of Income and Expenditure Data Collected in Taiwan's Three Household Income Surveys," (Taipei: Academia Sinica, August 1973), typescript. 2/ See: Arun Shourie, "National Accounts and the Regression Enthusiast, Sri Lanka: A Study," Economic and PoliticaZ Weeky, Vol. 9, No. 52 (December 28, 1974). Some 33 percent of the GDP was estimated indirectly (i.e., as a multiple of other items). In India, only about 60 percent of the estimated net domestic product for a particular year is based on direct information for that year; for the remaining 40 percent, estimation is indirect, from some benchmark estimates of output or value added, which are related to extrapolated figures of the number of workers in a given sector or activity. See: M. Mukherjee, "National Income Statistics," in Data Base of Indian Economy: Review and ReappraisaZ, ed. C. R. Rao, Vol. I (Calcutta: Statistical Publishing Society, 1972), pp. 84-100. -17- Table 3 Peninsular Malaysia: Monthly Per Capita Income and Expenditure According to the Household Expenditure Survey (HES) and the National Accounts, 1973 HES National Accounts Percentage Variable (M$) (M$) Difference Per Capita Income 67.6 100.q9a -33.0 Per Capita Expenditure 57.2 67.5Lb -15.3 /a The figure refers to per capita CDP at factor cost estimated by the Economic Planning Unit and cited in the World Bank Report No. 1177a-MA, Malaysia: Second Plan Performance and Third Plan Issues, Vol. 1 (1976), Table 2.5 in the Statistical Appendix. /b Th6 figure is based on the estimate of private consumption expenditure by the Inter-Agency Planning Group, cited in: Malaysia, the Treasury, Economic Report, 1974-75,(Kuala Lumpur, 1974),Table 1.1 of the Appendix. Note: The national accounts figures cited above are not strictly comparable to the HES data which are based on a sample of private households only. -18- Table 4 Sri Lanka: Level and Distribution of Monthly (a) Per Capita Income and (b) Private Consumption Expenditure According to National Accounts and the Socio-Economic gurvey, t060-76 - Income by Absolute Amount in Rupees Source/ Socio-Economic Percentage Expenditure National Accounts /a Survey /b Difference (a) Income Compensation of 35.1 25.3 -27.9 Employees (48.0) (53.6) Income from Unincorporated 34.7 21.2 -38.9 Enterprises (47.6) (44.9) Rent, Interest 3.2 0.7 -78.1 and Dividends (4.4) (1.5) Total 73.0 47.2 -35,3 (100.0) (100.0) (b) Private Consumption 58.5 50.6/-c -13.6 Expenditure /a Sri Lanka, Department of Census and Statistics, Statistical Abstract of Sri Lanka, 1973 (Colombo, 1973), p. 252. /b Excluding pension, remittances, cash allowances, other periodic cash receipts, and the imputed value of free rice rations. /c If certain expenditures had not been excluded, per capita consumer expenditure according to the surveywould have been Rs. 54.1, only 7.6 percent below the national accounts figures. Notes: 1. Figures in parentheses show income from a specific source as a percentage of the total. 2. The national accounts figures for 1969-70 have been obtained by adding one-sixth of the income from each source for 1969 to five- sixths of the corresponding income for 1970. The relevant pop- ulation of the mid-point of the survey period has been estimated by linear interpolation at 12.52 million. 3. All figures have been rounded independently. Table 5 Taiwan: Distribution of Annual Per Capita Personal Income by Source ard Private Final Consumptiorn Expenditure, According to National Accounts and the Family Income and Experditure Surveys of 1968 and 1974 1968 1974 Major Category of National Survey Percentage National Survey Percentage Income/Expenditure Accounts Data Difference Accounts Data Difference 1. Compensation of 5,094 3,342 -34.4 14,916 11,421 -23.4 Employees (56.3) (49.2) - (63.4) (60.5) 2. Income of Unincorpo- rated Enterprises -- including Agri., 1,705 2,382 +39.7 3,366 4,769 +41.7 Forestry, Fishing (18.8) (35.1) (14.3) (25.2) 3. Property Income 2,256 1,070 -52.6 5,263 2,697 -48.8 (24.9) (15.7) (22.4) (14.3) F*a 4. All Income 9,055 6,794- -26.6 23,545 18,887- -19.8 (100.0) (100.0) (1o.0) (100.0) 5. Private Final Con- 7,320 6,169b/ -15.7 18,234 (i) 16,079b/ -11.8 sumption Expenditure (ii) 15,129.s/ -17.0 a/ Excluding transfer receipts. b/ Total household expenditure as defined in the survey, including interest, taxes and transfers and comtributions. This figure is presented here only to perit comparison with the 1968 figure. c/ Total consumption expenditure, excluding interest, taxes and transfers and contributions. This figure cannot be estimated from the 1968 tape. g Eatimates of national income from secified major categories Ln 1968 and 1974 have been taken from page 110 and 92, respectively, of: [Taiwan] Directorate General of Budget, AccountiaS and Statistics, NationaZ Income of [Taiwan] : National Accounts in Taiwan for 1963-1977 and PreZimin=TiU Estimates of National Incone in -Paiomn,or9.978 (Taipoi, December 1978). The population figures for mid-1968 (13.68 million) and mid-1974 '(f571ln)hAare been taken froa (a) 1973 Taiwan DemograpFic Fact Book, 'aiarC(3 published by the Ministry of the Interior (Taipei, September 1974), p. 40 and (b) kbnthly Butletin of Statieties of '[ ai(e0 Vol. 2, No. 5 (May 1976), p. 1. -20- (1) The discrepancy between surveys and national accounts with respect to estimates of consumption expenditure is less (between 12 and 17 percent) than that with respect to income (20 to 35 percent). (2) Among different categories of income, the discrepancy is highest with respect to income from property and smallest with respect to wages and salaries. The extent of discrepancy with respect to income of unincorporated enterprises lies between that for the other two categories in Sri Lanka; the Taiwanese data suggest a rather surprising overestimation of the income of unincorporated enterprises in the survey, and illustrate the difficulty of distinguishing between the income of unincorporated enterprises and 2/ property income.- The relatively amall differences between ourveye and national accounts in the estimates of expenditure or consumption may partly reflect the lack of independence between them. However, expenditure seems to be easier to recall than income; also, it is probably more stable over time than income, at least for many of the rural agricultural households. It may also be more closely related to what the economists call "permanent income" (rather than "transitory" income). Often, the survey questions on expenditure 1/ The Directorate of Statistics in Taiwan considers the survey estimates of wages and salaries to be incomplete with respect to payments in kind. 2/ If the two sources are grouped together, the survey estimate of income from there is only 12.9 percent below the national accounts figure in 1968 and 13.5 percent below the corresponding figures for 1974. Note that the survey used income categories (apart from wages and salaries) of: (a) net income from (i) agriculture, (ii) forestry, and (iii) fishing; (b) business income; (c) professional income; (d) interest; (e) actual and imputed rent; (f) investment income; (g) transfer receipts and (h) miscellaneous. These categories have to be combined for comparisons with national accounts. The Directorate of Statistics in Taiwan has confirmed that while compiling national accounts, a part of the income of unincorporated enterprises is attributed to property. -21- are more detailed (than on income) in terms of the commodities and the number of days for which the quantity and value of consumption are recorded.1 The problems of imputing value for consumption out of "home-grown" stock or the collection of "free goods" and gifts or transfers in kind are, of course, common to both expenditure and income information. Before discussing these issues further, let us review briefly the savings rates implied in some of our surveys. C. Savings Rates Implied by Surveys When surveys collect data on both income and expenditure (as was true for Peninsular Malaysia, Sri Lanka, and Taiwan).Y, the plausibility of implied savings rates is examined. Before looking at our data on the subject, it should be noted that there is no real reason for a household to balance its receipts and expenditures for a short period of one month; and for expenditures on durable goods, a substantial proportion of households probably draw on their past savings. However, the Malaysian and Sri Lankan data relating to expenditure on durable goods are based on a reference period 1/ In Peninsular Malaysia, the survey households were visited on alternate days of the reference month (the original plans envisaged a daily visit). In Sri Lanka, expenditures on food, drink and tobacco were recorded Oeparately for each day of the reference wepk (implying seven visits for the week); in Nepal, the survey had planned six visits during the week. 2/ Even the surveys in Nepal have collected data on incomes, but the non- reporting of earnings of a substantial proportion of self-employed persons leads us to doubt the value of this information. Therefore only the expenditure data have been used in our analysis for Nepal. -22- of "one year," and our estimates of monthly expenditure on durable goods are the monthly averages for the relevant year. The resulting estimates of monthly expenditure and savings are "hybrids" in a sense, but there is no alternative. For Taiwan, both the expenditure and income data relate to a whole calendar year; and while the long reference period is known to introduce biases through recall lapse, there may be some cancellation of errors. Three tables in Annex 4 show the savings rates for different deciles of households, in Peninsular Malaysia, Sri Lanka, and Taiwan (1974 survey only), ranked according to per capita income as well as per capita expenditure. Households reporting negative and positive savings are distinguished within each decile, and dissavings as well as savings rates are shown separately for them as well as for the decile as a group. Consistent with a priori expectationS, among decilea ranled in terms of per capita incomes the proportion of dissaving ,households as well as the rate of dissaving does indeed vary inversely with decile, or the saving rate rises with decile. This is seen in Peninsular Malaysia and Sri Lanka where the bottom five and seven deciles, respectively, reported negative savings and also in Taiwan where even the bottom decile of households reported a small positive savings rate. With deciles based on per capita expenditure, it is difficult to detect a clear relationship between the savings rate and decile in Sri Lanka and Peninsular Malaysia; but in Taiwan the savings rate varies directly with PCE decile. A serious problem in evaluating the plausibility of savings rates implied by survey data, as shown in Table 6, is the lack of -23- Table 6 SAVINGS RATES ACCORDING TO HOUSEHOLD SURVEYS OF THREE COUNTRIES Reference Savings Rate Year (Percent) Malaysia (Peninsular) 1973 15.5 a/ Sri Lanka 1969-70 -0.3 b/ Taiwan 1968 9.8 1974 15.2 at The figure is based on the total sample of 7,273 households. If households with zero or negative incomes (not implausible during any particular reference month) are excluded, the relevant sample size declines to 7,212 households and the savings rate rises to 16.3 percent. b/ Certain categories of expenditures such as (a) "rates" or taxes on owned housing and expenditure on the repair and maintenance of housing units, (b) expenditure on the maintenance and repair repair of private vehicles, and (c) miscellaneous items, i.e., trade union and other professional contributions, cost of litigation, interest on consumer debt, gifts and donations, expenses on family weddings, births, funerals and other social or ceremonial occasions, and (d) motorcycles and scooters, and new or secondhand motor cars. If these items were not excluded, the dissaving rate would rise to 0.8 percent. c/ The figure represents the difference between all current expenditures and current receipts reported by households on our tape. The published reports for Taipei city and the rest of Taiwan, taken together, imply a savings rate of 15.0 percent for 1974. The 1968 Survey data imply a savingwrate of 8.4 percent according to John C. H. Fei et al., Growth with Equity: The Taiwan Case (New York: Oxford University Press, 1979), pp. 300, 301. -24- independence between these estimates and the figures used in national accounts, such as was noted above for Taiwan. Even for Malaysia, the planning authorities have probably used the results of the HES for their estimates of private household saving. The Sri Lankan survey data raise the interesting question whether the household sector as a whole can really dissave for an extended period of time. It is quite likely that the under- statement of incomes is an important contributory factor. Quite apart from any unwillingness to reveal one's "private" affairs to an employee of the government (except for wages and salaries hlch ara 6ftan "at" Acedisi to one's "grade" in the government or public sector), the absence of any tradition of careful accounting or record-keeping of receipts by different members of the family and of consumption out of "own produce" or stocks makes it difficult to obtain good estimates of income for any specific time interval. The problem becomes even more serious when the respondents are asked to recall income and expenditure for a whole year, as is done in Taiwan. The experience of a large scale survey conducted by an Asian country provides some illuminating information. The sample survey was spread over a full year and roughly one-twelfth of the total sample was visited several times during each month to gather considerably detailed data on income and expenditure during the month.1/ All the households were revisited also during the early part of the next year and were asked 1/ For many unincorporated enterprises or businesses, it would be based on the annual net income of the previous year. -25- to report their income for the previous calendar year. A comparison of the monthly income reported during the first visit with the average monthly income reported for the year as a whole (by the same household) showed a correlation of only 0.42; and the average monthly income according to the second round of the survey was about one-seventh lower than the average monthly income estimated from the first visit. (The discrepancy, in percentage terms, was almost the same for both urban and rural households, but a little higher for households where the household head was an employee than for households with the head as a self-employed or employer.) Admittedly, response errors also affect the expenditure data and can be illustrated by a US study of 237 families in the Chapaign-Urbana area (out of an initial aample of 556--with a response rate of 43 percent) con- ducted in 1951-52. Questions on the "nature and characteristics of durable goods purchased during the last six months"-f were asked of different adult members of the households, with care taken to ensure independence between the responses of different members. It was found that in 68 percent of the families, purchase of a durable good was not reported by one or more family members; the percentage was 35 with respect to the purchases of 1/ Robert Ferber, "On the Reliability of Responses Secured in Sample Surveys," Journal of the American StatisticaZ Association, Vol. 50, No. 3 (September 1955), pp. 788-810 The actual question was: "Has anyone in your household bought any durable goods since last (name of month approximately six months earlier), that is, items similar to those listed on card A?" Card A listed 50 different durable goods. -26- cars and motorcycles and 76 for the purchases of furniture and rugs.1/ The problem may be less serious in developing countries insofar as (a) purchases of durable goods may be important events because of their relative rarity, and (b) the household as a whole (if not also the.neighbors) gathers around an investigator asking such questions. However, the risk certainly exists when a survey like that of Taiwan uses a reference period of a year./ There is a serious dilemma in the choice of reference periods and the underlying duration of recall. Ideally, for studying the disparities in income or in living standards, we would like to focus on a relatively stable situation; and to minimize the influence of transient factors, a full year would seem to be an appropriate reference period. However, if the income or expenditure data for an entire year are collected in a single visit, the results are likely to be affected by serious problems of recall. An alternative is to undertake a "panel study," involving several revisits to the same household. However, panel studies tend to be more expensive than single-visit surveys, and except in the United States,- there are 1/ Items not reported by any members at all were naturally excluded. 2/ As in Japan, a small number of Taiwanese households are selected to maintain account books of expenditure and income. However, no study was made (at least up to 1976) of the comparability of data reported by households maintaining records with those who were only interviewed. The Japanese data from two sources are reported to show rather large discrepancies. 3/ In the United States, the Survey Research Center of the University of Michigan has been conducting a longitudinal survey of a sample of households to study the dynamics of their income and earnings. -27- few successful examples of such longitudinal surveys. More experimentation with panel surveys to obtain better information on disparities in income or expenditure seems highly desirable. From the point of view of income distribution studies, the question often arises as to whether the understatement of incomes occurs uniformly across all income levels, or is limited to upper income groups who may want to conceal their affluence to avoid taxes, etc. Frankly, we do not know. It is not unlikely that the interviewers get less cooperation from the rich and might, therefore, avoid interviewing them. (That is one reason why certain surveys, for example in Indonesia, exclude the rich from the frame of their sample.) At the other extreme, the poor who live in slums may be difficult to contact in terms of their availability at home; and/or the interviewers may be reluctant to visit the areas inhab- ited by the poor; and/or the sampling frame (unless it is prepared just before the survey by listing all households in a block) may be less up- to-date for areas at the periphery of an urban center. Further, it is dabatabl whathe tha pmr, Whe are interviewea, (a) overstate their poverty in the expectation of some assistance, or (b) understate it to appear better-off than they really are to preserve "self-respect." No definitive conclusion is possible. Most assertions on these issues are likely to be "opinions," based on the analyst's perceptions in a few cases, and the real situation probably varies within and across countries over time. Problems are compounded by the fact that interviewers in large sample surveys seldom have the time to establish a rapport with the -28- respondents such as is possible for the anthropologists undertaking intensive-type studies of a village or two through participant observation over a relatively extended period of time.1/ Even when the respondents report their income or receipts correctly, some problems can arise because of variations among households with respect to the importance of the receipts in kind or consumption out of their own produce or out of the stock of the family enterprise. The practices of different countries with respect to the prices used to impute the value of such receipts or consumption are not uniform; and it is likely that, depending on the spread between ex-farm or ex-factory prices and the retail prices, the estimate of income or consumption expenditure would be affected by the choice of one or the other. The underlying idea is to estimate the opportunity cost; the choice of the ex-farm price seems appropriate if the alternatives are seen to be: (a) to sell the produce, or (b) to consume it. If, however, the alternatives are viewed to be: (a) to produce oneself, or (b) to purchase from the market, the use of retail prices would seem appropriate. Note that the retail prices encountered by different households are not necessarily uniform; they depend on the frequency (and volume) and timing of the purchase as well 1/ There seems to be a consensus among survey researchers that the income of an individual or a household is a sensitive subject and questions about it are likely to have an adverse impact on the extent of cooperation or the response rate in a survey or study. The use of broad "intervals" to record income and some preference for relegating income questions to the end of the interviews are thought to minimize the risk of non-response. -29- as the quality of the item purchased. The available information on these issues is inadequate, although it is widely believed that, ceteris paribus, the poor pay more for an item of the same quality because they purchase smaller quantities more frequently than do the better-off; furthermore, the seasonal variations in the prices of foodgrains are unfavorable to them. The possibility of differences in the quality of items limits the usefulness of indirect comparisons of prices paid by different groups of households (estimated by dividing the reported value of purchases by the relevant quantities); yet, this area deserves additional research effort. On the whole, therefore, one is forced to rely on the premise that even when income or expenditure is understated, the relative ranking of households would not be affected to any large extent. And if somewhat broad groups are used, such as deciles of households, the likely errors in classification may be assumed to be small. The estimate of the incidence of poverty may not, therefore, be wide off the mark, even though the measures of inequality may be affected./ The validity of this assumption is obviously not yet firmly established. Nonetheless, the admittedly limited evidence reported earlier guggestg that it would be advisabla to dmphaaige the data on expenditure over those on income and to collect them for relatively short reference periods to reduce the problems of recall lapse. -30- There is no standardization, however, of the reference periods used for surveys in different areas. In our data sets, all the infor- mation for Gujarat and Maharashtra is based on a reference period of the last 30 days, whereas that for Nepal and Sri Lanka is based partly (i.e., for food, drink, and tobacco) on daily visits for a week, and partly (only for non-durable goods in Sri Lanka) on a reference period of a month. (As noted earlier, Sri Lanka used a reference period of a full year for recording expenditures on durable goods.) In the Malaysian survey, data on the consumption and purchase of non-durable items were collected for a full month with visits on alternate days (initially through daily visits); however, the expenditure on durable goods was recorded for the 12 months preceding the month of interview. (Subsequently, the expenditure on durable goods during the calendar year 1973 as a whole, as reported during the revisit during January-April 1974, was substituted for the data gathered during the month of interview.) The Taiwanese data are all based on a reference period of a full calendar year. These differences in reference periods introduce some non-comparability in the basic data; but given the state of the art, a cross-country study can only note it and cannot eliminate it through any arbitrary adjustments. We turn now to a discussion of the indices of inequality of expenditure and income according to alternative criteria for ranking house- holds (total vs. per capita income or expenditure) and the implications of taking account of the share of income received by percentiles of households or individuals within households. The question of ranking criteria is -31- relevant not only to the measures of inequality, but also to the iden- tification of the poor if and when one is forced to consider the bottom 40 percent of households as poor. -32- III. Indices of Inequatity and the Ranking Criteria One of the main objectives of our research was to examine the extent to which the poor who were identified according to the criterion of per capita expenditure or income (PCE or PCY) are identical to those identified according to the criterion of total household expenditure or income (THE or THY). If the differences in income levels are to be related to welfare, the simple adjustment for differences in household size and ranking of households according to their PCE or PCY seem to be clearly advantageous.-/ Yet much of the literature on income distribution in various countries is based on the ranking of households according to the size of their THY or THE rather than PCY or PCE./ In a compilation of income distribution data for 80 countries, published in 1975 by Tha W6rld BAtk, aim e6tntrias (Chila, Cost Rica, Dominican Republic, Honduras, Malaysia, and the United States) based at least / There may be some economies of scale in consumption which are overlooked in the per capita ranking. However, these economies are unlikely to exist with respect to all commodities and will seldom be so important that an increase in the size of the household would not raise the consumption requirements. The need for taking account of differences in the sex and age composition uf household members is examined below. 2/ One can speculate about the reasons for this practice. When the data sources collect information on income on the basis of one (or a few) question(s) in terms of rather broad class intervals, it is not possible to estimate per capita income with any reasonable precision. If the mid- point of a class interval is taken as the actual income of all households in that class, the households will be ranked essentially in terms of their household size (obviously with an inverse relationship between the latter and the estimated per capita income). The simple solution adopted by the statistical agencies is to present the data for households ranked according to their total household income. -33- one of their distributions on both per capita and total household income; further, for Tunisia, the available data pertained to only the distribution of individuals by per capita income.1/ The data based on these different criteria are compared without any recognition of the underlying difference. The recent United Nations guidelines on the income distribution statistics recommend tabulations mainly in terms of the (pre-tax) total household income and only two (out of 26) general tables on per capita basis for comparative purposes.- It is not adequately recognized that the two criteria lead to a significantly different ranking of the households, although this is not evident from the Gini coefficients or other measures of the concentration of income. 1/ Shail Jain, Size Distribution of Income: A Compilation of Data (Washington, D.C.: The World Bank, 1975). The distinction between per capita and total household income is not relevant when the income data relate to workers, income recipients, or the economically active population. In India, the National Sample Survey data provide information only on the distribution of population ranked according to per capita expenditure of the household; but these data were not included in the compilation by Shail Jain. Likewise, the Indonesian Socio-Economic Survey (SUSENAS) of 1969-70 had tabulated individuals according to their per capita expenditure. For the 1976 Survey, however, the published data show only the distribution of households according to total household expenditure, although the tables according to per capita expenditure are available. 2/ United Nations, Provisional Guidelines on Statistics of the Distribution o? Income. Consump+ion and Accumulaidm 6f Rmohod4 (Earion M, No. 61J (New York, 1977), pp. 58-59. Since the household is generally the production (or consumption) unit and the specific contribution (share) of individual members or workers is difficult to identify or measure, it is only natural that it should also be the basis of the reporting of income (or expenditure/ consumption). That is particularly true of the family farms and enterprises which account for a substantial proportion of the population in the rural areas of developing countries (and for consumption or expenditure, it is true of all societies). -34- A. Indices of Inequality To examine the effect of alternative ranking criteria on the indices of inequality, we have summarized in Table 7 three measures based on different data sets.- The three measures are outlined briefly in Annex 5. The underlying data are shown in Annex 6 and include: (a) the decile limits, (b) the percentage of total population included in each decile of households, (c) the average income (or expenditure) per house- hold and per person within each decile of households and (d) the share of total income (or expenditure) estimated to accrue to each decile of house- holds and to each decile of population. Following the practice in the literature, we have worked with deciles of households although it is possible to delimit deciles so that each decile has ten percent of the population. While the limits for deciles of population would naturally be different, the issues raised in this paper are not affected by the use of deciles of householas. Table 7 demonstrates the importance of the units of aggregation in addition to that of the ranking criteria. Even when households are ranked according to total household income or expenditure, different 1/ An earlier short version of this table was presented in a paper presented at a conference convened by the International Union for the Scientific Study of Population entitled, "Demographic Factors and the Distribution of Income: Some Issues." It has been published in the conference proceedings entitled, Economic and Demographic Change: Issues for the 1980's, Vol. I (Liege: Ordina Publications), pp. 289-320. (A slightly revised version is available as Reprint 129 in the World Bank Reprint Series.) The income inequality indices for Sri Lanka in the published volume are based on the data as collected by the survey and do not include any imputation for the value of the free rice ration supplied throughout the survey period. Almost 93.2 percent of all sample households had received the free rice ration; its imputed value formed nearly nine percent of the total expenditure (including it, but excluding certain items noted in a footnote to Table 4 above). All the inequality indices declined after the imputation of the value of the rice ration. Obviously, it is important to take account of the various kinds of subsidies, many designed to benefit specifically the low income groups, although the household surveys often do not provide the relevant information. Table 7 Indices of Inequality of Expenditure or Income with Alternative Ranking Criteria and Unite of Aggregation Ranking Criterion Peninsular for Unit of Gularat Maharaslitra Nepal Sri Lanka Taiwan Malaysia Deciles AgAregation Rural Urban Rural 17rban 11 Towns 7 Towns (1972-73) 1972-73) (1973-74) <1974-75) (1969-70) (1968) (1974) (1973) A B A B A B A B Cini Coefficients Per Capita Individuals .259 .261 .273 .331 .312 .303 .279 .331 .299 .321 .269 .307 .421 .502 Households .192 .143 .206 . 184 .223 .193 .204 .266 .189 .210 .164 .206 .340 .428 Total Households .328 .304 .354 . 371 .389 .367 .310 .357 .316 .330 .271 .301 .452 .518 Household Individuals .172 .148 .273 . 206 .201 .185 .194 .257 .205 .230 .178 .224 .335 .429 Kuznets Index Per Capita Individuals .194 .193 .201 .251 .235 .229 .208 .247 .224 .240 .200 .228 .317 .381 Households .143 .099 .151 . 134 .167 .147 .149 .196 .139 .155 .120 .152 .255 .323 Total Households .246 .224 .263 .280 .291 .277 .231 .267 .237 .247 .200 .223 .344 .396 Household Individuals .135 .117 .138 , 162 .152 .140 .145 .193 .161 .180 .135 .169 .253 .323 I Entropy Measure Per Capita Individuals .110 .118 .126 . 177 .164 .154 .131 .185 .148 .173 .228 .160 .283 .391 Households .060 .039 .073 .057 .082 .059 .075 .130 .059 .074 .152 .074 .181 .282 Total Households .169 .148 .204 .213 .244 .219 .155 .207 .159 .175 .223 .151 .311 .401 Household Individuals .049 .041 .062 .4078 .078 .067 .073 .124 .071 .090 .169 .089 .183 .295 Notes: 1. IWen two columns are shown for a country, Colemn A shows indices of inequality in exrenditure and Column B those for income. For other regions, only expenditure deciles have been demarcated, .nequality in expenditure has been estimated. 2. The Taiwanese data on total expenditure reported for 1968 and 1974 represent a sum total of (a) consumption expenditure and (bD) interest payments, loss on investment, income and othe-r taxes as well as transfer expenditures, For 1974, it is possible to exclude the latter categories, which formed about 5.9% of the reported total expenditure. As would be expected, the inequality indices for consumption expenditure in 1974 were all lower; they are shown below. Indices of Inequality in Consumption Expenditure in Taiwan, 1974 Ranking Unit of Gini Kuznets Entropy Criterion Aggregation Coefficient Index Measure THE Household .262 .193 .1L2 PCE Household .149 .110 .038 THE Individual .163 .126 .045 PCE Individual .256 .191 .19 -36- deciles account for different shares of the population; and the population receiving a particular share of income or expenditure differs from 10.percent of the total population. The need for taking account of this fact has been emphasized in the recent work of Professor 1/ Simon Kuznets.- Table 7 confirms that even when the households are ranked according to their total income or expenditure, aggregation in terms of individuals (i.e., due recognition of the percentage of population in each decile) implies significantly lower indices of inequality. In other words, inequality in per capita expenditure or income is lower than that in total household expenditure or income. This happens because, as shown in the next section, the average size of households generally tends to rise with THE or THY decile, and the lower deciles accounting for a small share 2/ of income also include a small percentage of individuals.- Even when we rank households according to their per capita income or expenditure, it is possible to look at the iftea ar -SMO-1nditura shareg of deciles of households and overlook the different shares of population in each decile. The rationale for such an approach would be to ensure comparability with the conventional measures of inequality, which look at the shares of income (or expenditure) of deciles of households (without 1/ Simon Kuzets, "Demographic Aspects of the Size Distribution of Income: an Exploratory Essay," Economic Development and Cultural Change, Vol. 25, No. 1 (October 1976), pp. 1-94. 2/ Thus our data seem to contradict the assertion of Robert Repetto that "if the basis of the (income) distribution is the number of individuals rather than the number of household units, and household income is converted to a per capita basis, then the degree of inequality is comparatively higher." See: Robert Repetto, Economic Equality and Fertility in Developing Countries (Baltimore: The Johns Hopkins University Press, 1979), p. 22. -37- paying attention to the population shares). Of course, our preferred approach is to rank households according to their per capita income and to take due account of the proportion of population in each decile, i.e., per capita ranking and aggregation in terms of individuals. With the per capita ranking of houaeh6lds, Atrattion in termn of individuals always implies greater inequality than aggregation in terms of households because of the generally inverse relationship between average household size and the PCE or PCY decile. A given percentage of expenditure or income accruing to deciles of households is shared by more than 10 percent of the population. If the differences between the percentages of population in different deciles are ignored, the inequality of expenditure or income appears to be much lower. A comparison of our preferred indices of inequality based on per capita ranking and aggregation in terms of individuals with the conventional indices (based on total household ranking and aggregation in terms of households) shows the former to be significantly lower. The sole exception is Taiwan in 1974, where the inequality in PCY, measured with due recognition of the share of population in each decile of households, appears slightly higher than the inequality in THY (with THY ranking and aggregation in 1/ 2/ terms of households).- - 1/ Note, however, that inequality of PCE in Taiwan in 1974, measured through our preferred approach, is slightly lower than that in THE, measured conventionally; and the difference is slightly larger when we limit our attention to consumption expenditure. 2/ As shown in the next section, the inter-decile range of variation in average household size in the 1974 Taiwanese data based on the THY ranking was smaller than in the data based on the PCY ranking; this observation was an exception to the general pattern seen in all other data sets. It is likely that the marked decline in fertility in Taiwan between 1968 and 1974 contributed to a reduced variation in the size of households. -38- One would not want to compare the inequality indices based on the THE or THY ranking and aggregation in terms of individuals with indices based on PCE or PCY ranking and aggregation in terms of households; such a comparison has little to recommend itself. Most of them are remarkably close to each other but there is no clear pattern in the direction of differences. B. Target Groups for Anti-Poverty Programs The ranking criterion is important for the indices of inequality and even more so for the identification of target groups for anti-poverty programs, particularly when the poor are identified in terms of the con- cept of relative poverty, i.e., such as the bottom 40 percent of households or households who received less than a given proportion (e.g., one-third) of the average per capita income in a country. The differences in the characteristics of the poor according to the per capita and total house- hold rankings will be highlighted throughout this paper. For the present, it should be noted that, as shown in Table 8, the percentage of households falling in the same decile is between 14 and 26 in the ten data sets for which PCE and THE rankings have been attempted, and between 20 and 28 in the four data sets of Peninsular Malaysia, Sri Lanka, and Taiwan. The percentage of households in the same quintile according to the alternative criteria is naturally larger; yet a majority of the households fall in a different quintile according to per capita and total household ranking. Let us consider the frequently used approach of considering the bottom 40 percent of households to be poor. If the ranking is in terms of the THY (or THE) criterion, in our data sets, the bottom 40 percent of households would cover between 49 and 75 percent of the poor households -39- Table 8 PERCENTAGE OF HOUSEHOLDS IN THE SAME DECILE OR QUINTILE ACCORDING TO THE ALTERNATIVE CRITERIA OF PER CAPITA AND TOTAL HOUSEHOLD EXPENDITURE OR INCOME Percentage of Percentage of Households in the Households in the Area Same Decile Same Quintile 1. GUJARAT 1972-73 Rural 17.4 (16.5) 32.9 (32.0) Urban 17.7 (16.1) 31.3 (29.0) 2. MAHARASHTRA 1972-73 Rural 15.9 (15.2) 32.1 (31.3) Urban 15.3 (14.7) 26.0 (25.4) 3. NEPAL Eleven Towns 1973-74 14.7 (13.5) 31.3 (29.7) Seven Towns 1974-75 14.5 (12.8) 29.4 (28.1) 4. PENINSULAR MALAYSIA 1973 22.5 (20.1) 40.0 (37.4) 1973* 26.2 (23.5) 44.9 (42.1) 5. SRI LANKA 1969-70 17.4 (16.2) 33.0 (31.6) 1969-70* 20.6 (19.2) 35.7 (34.3) 6. TAIWAN 1968 19.5 (17.7) 35.1 (33.3) 1968* 20.2 (18.3) 36.0 (34.3) 1974 26.5 (24.7) 40.1 (38.4) 1974* 28.3 (26.4) 43.4 (41.5) *Deciles in terms of income. Figures in parentheses show the percentage of total population accounted for by the households in the same decile or quintile. -40- according to the PCY (or PCE) criterion (and between 41 and 64 percent of the population of PCY/PCE poor households). If, on the other hand, the bottom 40 percent of households had been identified on the basis of PCY or PCE criteria, they would include the same (between 49 and 75) percentage of poor households in terms of the THY or THE criterion, but they will account for between 71 and 88 percent of the population of the THE/THY poor households. This happens because, as was noted earlier, the bottom deciles in terms of THE/THY have smaller-than-average household sizes, whereas the opposite is true with the PCE/PCY ranking. On the whole, the per capita ranking also provides a better coverage of the poor population according to the total household income or expenditure. The reclassification of households with the choice of the ranking criterion can be examined partly by looking at the relationship between the age group of the household head-/ and the level of total and per capita 1/ A precise identification of the household head in terms of the structure of authority in the family is seldom possible in a survey; most surveys accept what the respondents report. However, the data on the relation- ship of each member to the household head are extremely useful for evaluating the role of life cycle factors in income distribution (and also for verifying the reported information). Unfortunately, however, the available data tapes for Gujarat, Maharashtra, and Sri Lanka do not explicitly identify the household heads; in other words, the infor- mation on the relationship to the head of the household has not been transferred to the tapes. We have been assured that persons whose serial number was equal to one were the heads of households. In Peninsular Malaysia, 84 percent of the household heads were main earners in their households; that may partly be a statistical artifact because the income attributable to unpaid family helpers or non-labor inputs is generally recorded against the person reported as the head. In Taiwan, the surveys distinguished the "economic" household head from the "census" household head; but they were the same individuals in 94 percent of the households in 1968 and 85 percent of the households in 1974. -41- expenditure or income. The relevant data are presented in Annex 7. It is evident that the average household size tends to rise with the age of the head, at least up to the age group 40-44; it levels off or declines for older heads whose children eventually form separate households. Also, the relationship between THE or THY and the age of the household head generally forms an inverted-U shape curve, indicating the operation of life cycle factors.- However, the two tendencies offset each other; and the relationship between PCE and the age group of the head becomes relatively flat as in Gujarat, Maharashtra, and Nepal or somewhat U-shaped as in Sri Lanka. Households with young heads and a lower-than-average household size fall in the bottom deciles when THE or THY is used as the ranking criterion; but they often move to the top deciles when PCE or PCY is used to rank households. Such reclassification explains the small proportions of households falling in the identical decile under alternative ranking criteria, and the differences in the target groups for poverty-alleviation programs in terms of the PCE or THE (PCY or THY) rankings. More importantly, the life cycle effects would seem more significant in the distribution of total household income or expenditure than in the distri- bution of per capita income or expenditure. This hypothesis receivep some support from the results of multivariate analysis reported later in the paper. 1/ A similar and, in fact, more distinctly inverted-U shaped relationship is observed between the earnings of individual employees (wage or salary earners) and their age. The relevant data will be discussed later. -42- C. Estimation of Adult Equivalents and Alternative Indices of Inequality Since the size of the household varies with the age or the stage of the life cycle of the household head, it is logical to presume that the age composition of the poptlation in decile with a highat-than-Average household size might be significantly different from the average and particularly the age composition of persons in the upper deciles. (This presumption is confirmed by our data presented later.) Given the distinct age-related differences in the capacity to work or earn as well as what Kuznets calls "the volume of goods required to fulfill whatever may be considered acceptable or warranted needs,"-1/ the differences in per capita income or consumption do not seem to be good indicators of differences in welfare. Some form of standardization of household members into adult equivalents seems necessary. The FAO, the WHO, as well as some countries, such as India and the United States, have attempted to identify the sex-age-specific differences in caloric requirements, but no norms have been developed with respect to the requirements of other goods. The relationships between expenditure on food and other commodities and services, estimated from consumption expenditure data, can hardly be taken as a good index of the "needs" because they are obviously influenced by the existing level and distribution of income, wealth and varying customs among different socio- economic groups. However, the Bureau of Statistics in Taiwan has been 1/ Simon Kuznets, "Demographic Aspects of the Size Distribution of Income: An Exploratory Essay,"Economic DeveZopment and Cutural Change, Vol. 25, No. 1 (October 1976), p. 30. -43- using some weights to estimate adult equivalents, which provide a starting point. For India, some weights for the conversions of individuals into adult consumption units have been estimated on the basis of caloric requirements. These weights are shown in Annex 8. The Indian weights have been used to estimate adult equivalents for Gujarat, Maharashtra, and Nepal. For Peninsular Malaysia, Sri Lanka, and Taiwan, both the Indian and Taiwanese weights have been used. As shown in Table 2 of Annex 8, the correlation between the number of members and the number of adult equivalents in a household is very high, a little higher when the Indian weights are used than with the Taiwan weights. The indices of inequality in expenditure or income per adult equivalent (AEE or AEY) are shown in Table 9. On the whole, the inequality indices based on AEE or AEY are not much different from those based on per capita ranking. The use of adult equivalents estimated with the Taiwan weights lowers the inequality indices slightly more than those with the Indian weights, although the latter distinguish between five age groups above the age of 20, whereas the Taiwan weights assume persons of a given sex aged 21 and over to be equal. Pending further research and experimentation with different weights, it can be concluded that although the ranking of households in terms of per capita expenditure or income does not seem conceptually ideal, the resulting measures of inequality are not very different from those in terms of adult equivalents. -44- Table 9 INDICES OF INEQUALITY IN EXPENDITURE OR INCOME PER EQUIVALENT ADULT Gini Kuznets Entropy Area Variable Weights Coefficient Index Measure 1. GUJARAT, 1972-73 Rural A22 Tndan 0.262 0.196 0.113 Urban AEE Indian 0.260 0.193 0.116 2. MAHARASETRA, 1972-73 Rural AEE Indian 0.273 0.202 0.126 Urban AEE Indian 0.327 0.247 0.173 3. NEPAL Eleven Towns, 1973-74 AEE Indian 0.305 0.230 0.157 Seven Towns, 1974-75 AER Indian 0.294 0.222 0.144 41 SRI LANKAI 1969-70 AEE Indian 0.282 0.211 0 133 AEE Taiwan 0.269 0.200 0.121 AEY Indian 0.332 0.248 0 187 AEY Taiwan 0.316 0.235 0.172 5. PENINSULAR MALAYSIA, 1973 AEE Indian 0.422 0.317 0.283 AER Taiwan 0.407 0.306 0.265 AEY Indian 0.514 0.383 0.406 AET Taiwan 0.501 0.373 0.390 6. TAIWAN 1968 AEX Indian 0.297 0.222 0.146 AEE Taiwan 0.285 0.212 0.135 AEY Indian 0.319 0.238 0.171 AEY Taiwan 0.306 0.227 0.159 1974 AEE Indian 0.270 0.201 0.124 AEE Taiwan 0.258 0.191 0.112 AEY Indian 0.308 0.230 0.162 AET Taiwan 0.295 0.219 0.149 ARE: Expenditure per equivalent adult. ART: Income per equivalent adult. -45- Table 10 shows the proportions of households that fall in the same decile or quintile according to alternative rankings in terms of per capita and per adult equivalent expenditure or income. With the Indian weights, between 58 and 74 percent of the households (accounting for about 61 to 76 percent of the population) fall in the same decile according to the two criteria; the use of Taiwan weights lowers these percentages significantly. The weights chosen to estimate adult equivalents do have some influence on the ranking of households and the identification of appropriate weights must continue to receive some attention. In the next section we shall concentrate on the characteristics of households or population according to the PCE and THE (or PCY and THY) deciles of households. -46- Table 10 PERCENTAGE OF HOUSEHOLDS IN THE SAME DECILE OR QUINTILE ACCORDING TO THE ALTERNATIVE CRITERIA OF PER CAPITA AND PER ADULT EQUIVALENT EXPENDITURE OR INCOME Variable rercentage of Frcentage Qf for Weights Households in the Households in the Area Ranking Used* Same Decile Same Quintile 1. GUJARAT, 1972-73 Rural Expenditure Indian 62.4 (65.1) 80.1 (82.6) Urban Expenditure Indian 59.3 (62.1) 78.6 (80.6) 2. MAHARASHTRA, 1972-73 Rural Expenditure Indian 57.6 (60.9) 78.0 (80.7) Urban Expenditure Indian 63.4 (68.5) 81.5 (84.6) 3. NEPAL Eleven Towns 1973-74 Expenditure Indian 64.3 (67.5) 82.1 (84.3) Seven Towns 1974-75 Expenditure Indian 62.7 (66.3) 80.7 (82.8) 4. PENINSULAR MALAYSIA, 1973 Expenditure Indian 69.7 (73.0) 85.5 (87.2) Epnditurp Taiwan 57.7 (61.3) 79.9 (81.5) Income Indian 73.8 (76.4) 87.1 (88.7) Income Taiwan 64.8 (66.4) 82.5 (83.5) 5. SRI LANKA, 1969-70 Income Indian 64.7 (66.7) 82.2 (83.4) Income Taiwan 47.6 (51.0) 72.4 (74.6) Expenditure Indian 61.6 (63.8) 79.4 (80.9) Expenditure Taiwan 43.5 (46.6) 67.1 (69.0) 6. TAIWAN, 1968/ Income Indian 63.4 (60.8) 81.7 (78.4) Income Taiwan 49.5 (52.5) 73.2 (75.0) Expenditure Indian 60.7 (63.5) 80.2 (81.8) Expenditure Taiwan 46.5 (50.6) 70.4 (72.8) TAIWAN, 1974 Income Indian 60.7 (61.9) 79.9 (80.7) Income Taiwan 44.5 (47.2) 68.3 (70.2) Expenditure Indian 56.7 (58.3) 77.5 (78.4) Expenditure Taiwan 40.7 (43.5) 64.1 (66.2) *The weights are shown in Table 1 in Annex 8. Note: Figures in parentheses show the percentage of total population accounted for by the households in the same decile or quintile. -47- IV. Characteristice of Household Heads or Population in Different Deciles One of the main objectives of the present project has been to identify the socio-economic characteristics of different income and/or expen- diture groups. The possibility of such an analysis was a major criterion in the selection of surveys. The characteristics to be reviewed here include (a) the demographic characteristics such as average household size, the size distribution of househQlds and the sex and age composition of heads of households and of the population in different deciles; (b) the human capital endowments or the literacy and/or educational attainments of the adult population, and the school-enrollment ratios for children in school- going ages; and (c) the economic characteristics such as the level of labor force participation rates and the level and structure of employment and unemployment. A. Demographic Measures 1. Average Household Size The issue of household size is relevant to a major inconsistency between the widely held view that the poor have large families and the observation that the bottom deciles of households, based on the conventional ranking in terms of total household income (or expenditure), account for a small share (much less than 10 percent) of the population. It might be reasoned that analysts who talk about the large families of the poor consider the total number of children ever born to the poor women and not their surviving children; and, if the poor also experience a higher-than-average level of infant and child mortality, the differences in the number of -48- surviving children, between the poor and the better-off, would be smaller than with respect to the number of children ever born. But the inconsistency is much larger than can be explained by reference to this fact; and the ranking criterion seems especially relevant to its resolution. Table 11 and Figure 1 summarize the data on the average size of households falling in different deciles ranked by per capita and total household expenditure or income.1 As was noted earlier, the average size of the household varies inversely with per capita deciles and directly with total household deciles. With a few exceptions, the relationship seems to be monotonic, more so with the latter ranking than with the former. The per capita ranking confirms the widespread belief that the poor have typically larger households. The contrary impression conveyed by conventional data on the distribution of income seems to be due entirely to the use of an inappropriate ranking criterion. Except in Taiwan in 1974, the inter-decile range of differences in average household size with total household expenditure ranking exceeds that observed in per capita expenditure ranking. In Taiwan, the decline in average household size from 5.8 in 1968 to 5.3 in 1974 is associated with a clear decline in the inter-decile range of differences in average size; the standard deviation of household size (estimated on the basis of individual observations) decreased from 2.74 in 1968 to 2.26 in 1974. Quite probably this latter change, associated with the decline in the Taiwanese birth rate 1/ For reasons of space, figures in this section do not show the characteristics of population or households according to income deciles. rable 11 Average Size of Households in Different Deciles with Alternative Ranking Criteria PenineuZar Decile of Guarat Maharasht2a Nepal S2- Lanka Taiwan Malaysia Bouseholds Rural Urban Rural Urban 11 Towns 7 Towns (1972-73) (1972-73) (1973-74) (1974-75) (1969-70) (1968) (1974) (1973) A B A B A B A B A. BousehoLds Ranked According to Their -Per Capita Expenditure (PCE) or Income (PCY) 1. 6.8 6.7 6.3 6.6 6.4 6.3 7.2 7.3 7.5 7.7 7.1 7.1 6.2 5.9 2. 6.6 6.4 6.0 6.3 6.3 6. 3- 6.8 6.9 7.1 6.9 6.5 6.3 6.0 6.0 3. 6.2 6.3 5.6 6.1 6.1 5.9 6.4 6.4 6.5 6.6 6.1 6.0 6.0 5.8 4. 6.2 6.0 5.5 5.8 5.8 5.9 6.2 6.3 6.3 6.4 5.8 5.7 5.8 6.1 5. 6.0 5.8 5.8 5.4 5.7 5.2 5.9 6.0 6.2 6.3 5.4 5.5 5.9 5.8 6. 6.1 5.5 5.3 5.1 5.6 5.4 5.5 5.6 5.9 5.8 5.4 5.3 5.9 5.7 7. 5.5 5.0 5.2 4-.4 5.0 5.0 5.7 5.7 5.8 5.9 4.9 5.1 5.4 5.8 8. 5.5 4.2 4.9 4.0 5.0 4. 7 5.0 5.0 5.0 5.0 4.5 4.7 5.0 5.2 9. 5.0 3.9 4.5 3.2 4.5 4.2 5.0 5.0 4.3 4.3 4.3 4.1 4.3 4.1 10. 4.3 3.2 4.1 Z.7 3.5 2.8 4.6 4.6 3.6 3.6 3.5 3.6 3.5 3.6 41 All 5.8 5.3 5.3 4.9 5.4 5.2 5.8 5.8 5.8 5.8 5.3 5.3 5.4 5.4 B. Households Ranked According to Their- TotaL Expenditure (THE) or Income (THY) 1. 2.4 1.9 2.2 1.7 2.2 2.4 3.3 3.6 2.9 3.0 3.1 3.4 2.9 3.3 2. 4.0 3.5 3.5 2.8 3.4 3.2 4.3 4.5 4.6 4.8 4.5 4.6 4.0 4.3 3. 4.8 4.3 4.1 3.8 4.1 3.8 4.9 5.0 5.2 5.1 4.8 5.0 4.6 4.8 4. 5.1 4.8 4.7 4.5 4.2 4.3 5.4 5.5 5.4 5.5 5.2 5.2 4.9 4.9 5. 5.7 5.5 5.1 5.0 5.0 4.8 5.9 5.9 6.0 6.2 5.4 5.4 5.2 5.2 6. 6.3 6.0 5.4 5.5 5.4 5.3 6.3 6.1 6.4 6.0 5.6 5.6 5.6 5.5 7. 6.6 6.1 5.9 5.9 5.9 5.8 6.5 6.4 6.2 6.4 5.8 5.7 5.9 5.9 8. 7.0 6.5 6.4 6.2 6.8 6.1 6.8 6.7 6.8 6.7 6.0 6.0 6.5 6.3 9. 7.5 6.8 7.1 6.6 7.4 7.2 7.4 7.2 7.0 6.9 6.2 6.1 7.0 6.7 10. 8.9 7.7 8.8 7.4 9.4 8.9 7.6 7.7 7.8 7.6 6.8 6.6 7.6 7.2 All 5.8 5.3 5.3 4.9 5.4 5.2 5.8 5.8 5.8 5.8 5.3 5.3 5.4 5.4 Sampe BousehoLds 5560 3528 5314 11103 4393 2254 9664 2776 5256 7285 Notes: 1. When two columns are shown f4or an area Column A refers to deciles demarcated on the basis of expenditure data and Column B to those on the basis of income data. For other regions, only expenditure deciles have been demarcated. 2. All deciles are deciles of households and not of population. -50- AVERAGE MUSEHOLD SIZE RURAL GUJARAT URBAN GUJARAT RURAL MAHARASHTRA l0- URBAN MAHARASHTRA NEPAL 11I TOWNS NEPAL 7 TOWNS 4 2 10- 0 2 4 6 8 10 0 2 4 a a 10 0 2 4 6 8 10 PENINSULAR MALAYSIA 1973 SRI LANKA 1969-70 TALWAN 1974 10- Lm N e .DECILE HOUSEHOLDS RANKED BY PER CAPITA AND TOTAL EXPENDITURE . PCE .............. THE -51- which was of the order of 18.7 percent during the period from 1968 to 1974 (from 28.8 to 23.4 per 1000 population),- has been one of the factors contributing to the observed decline in the inequality of income over the past decade or so.2 An important finding is the tendency for the percentage of one-or two-member households to rise with decile when the ranking is in terms of per capita expenditure. As shown in Table 12, such small households together formed less than 7 percent of the households in the bottom per capita expen- diture decile and between 29 and 65 percent of all households in the top decile. Associated with these differences in the average household size and the size distribution of households by decile are differences in the proportion of children aged 0-14; as shown below, they form a much higher proportion of the population in the bottom deciles than in higher deciles. 1/ Taiwan, Ministry of the Interior, 1974 Taiwan-Fukien Demographic Fact Book ([Taiwan], Taipei, December 1975), pp. 986-987. The decline in the general fertility rate (number of children born per 1,000 women aged 15-44) was sharper (34.5 percent), from 116 in 1968 to 76 in 1974. The dependency ratio, i.e., persons aged 0-14 and 65 and over as a percentage of persons aged 15-64, declined by 19.5 percent (from 82 to 66) over the period. The decline shown in our survey data is only 14.3 percent because of differences between the age composition of population reported in the Fact Book and the surveys (particularly the 1974 survey). 2/ Preliminary results of interesting research by Susan Greenhalgh of Columbia University suggest that the definitions used by the Bureau of Statistics in Taiwan were likely to lead to a selective coverage of the dispersed members of the families of the poor households, defined as such in terms of total household income. The extent of such selection or of dispersal of family members may have changed between 1968 and 1974. -52- Table 12 Single-and Two-Member Households as Percent of Households in the Bottom and Top Deciles with Ranking in Terms of Per Capita Expenditure (PCE) Single-Member Households Two-Member Households Region Bottom Decile Top Decile Bottom Decile Top Decile Gujarat, 1972-73 Rural N 13.4 2.2 16.1 Urban 1.0 22.7 1.7 20.5 Maharashtra, 1972-73 Rural 1.6 18.7 2.3 14.7 Urban 2.2 46.0 1.8 14.7 Nepal Eleven Towns 1973-74 0.7 33.5 2.3 15.9 Seven Towns 1974-75 * 49.1 1.8 15.5 Peninsular Malaysia, 1973 1.5 25.9 5.5 18.0 Sri Lanka 1969-70 - - 2.1 16.2 Taiwan 1968 0.4 19.7 1.1 15.4 1974 N 14.9 1.1 15.1 Notes: N: Negligible. -: Not included in the sample *: None of the single-member households was in the bottom PCE decile. -53- 2. Age Distribution of Household Heads The decile-related differences in the size distribution of house- holds and their average size are naturally related to the age composition of heads of households. With per capita expenditure ranking, the bottom deciles include a higher-than-average proportion of middle-aged heads of households, whereas the top deciles include a high proportion of younger heads of households.1/ The differences are particularly marked in the age distribution of male heads of households; they are blurred when one looks at the distribution of all household heads together, because of the markedly different characteristics of female heads of households, examined next. 3. Are Women Overrepresented Among the Poor? Many recent discussions about the role of women in development seem to assume that women are overrepresented among the poor; some believe them to be "the poorest of the poor." Insofar as this assertion relates to the intra-family allocation of responsibilities and benefits, its validity cannot be assessed with the data gathered by surveys examined in this project. Limiting ourselves to inter-household comparisons, however, a rough assessment is possible by looking at the proportion of females among (a) heads of households, and (b) all persons in households falling in different deciles. 1/ As noted in Section III, theiaverage size of households generally has an inverted-U shaped relationship with the age group of heads of households. Data on the age distribution of heads of households by per capita expenditure decile are shown in the country-specific papers. -54- Table 13 and figure 2 show the percentage of females among household heads in different deciles according to the alternative criteria of per capita expenditure or income (PCE or PCY) and total household expenditure or income (THE or THY). With PCE or PCY ranking, one observes a small relative overrepresentation of females among headrs of households in the bottomh one or two deciles in urban areas of Gujarat and Maharashtra, Sri Lanka, Taiwan in 1968, and Peninsular Malaysia; but the overrepresentation does not seem to be such as would justify widespread concern. If, however, one looks at the deciles based on the criterion of total household expenditure or income, all data sets suggest a sizable relative overrepresentation of females among the heads of households in the bottom one or two deciles. In all probability, the alleged overrepresentation of women among the poor is inferred from data relating to the total income or expenditure of households. The reclassification of households, that occurs when Athr of tham ranking criteria ig replaced by per capita criteria, seems to be quite important for female heads of households. Table 14 shows the household and per capita expenditure and/or income for different data sets according to the sex of the household head.1 Except in Taiwan in 1974, the monthly THE or THY of households with female 2/ heads ranged between 48 and 86 percent of that of households with male heads.- But the average size of households with females as heads was also considerably 1/ Table 14 has been compiled for male and female heads of households as a group in each data set and not for each PCE/THE decile because of the relatively small number of sample households with female heads. 2/ In Taiwan in 1974, the total household income and expenditure did not differ significantly according to the sex of the head. Table 13 Percentage of Females Among Household Heads in Different Deciles with Alternative Ranking Criteria Peninsular Decile of GWu,arat Maharaahtra Wepal SrfiLanka Taiwian Malaysia Households Rural Urban Rural Urban 11 Towms 7 Towns (1972-73) (1972-73) (1973-74) (1974-75) (1969-70) (1968) (1974) (1973) A B A B A B A B A. Households Ranked According to Their Per Capita Ecpenditure (PCE) or Income (PCY) 1. 4.8 10.6 10.2 10.4 7.1 9.8 14.4 14.9 10.8 8.7 9.2 9.1 20.4 22.9 2. 4.5 9.2 13.6 9.4 6.4 11.1 13.4 12.6 8.7 7.9 11.3 10.6 16.8 16.4 3. 3.5 7.7 10.5 9.0 9.6 10.6 12.6 12.1 7.6 9.7 7.5 7.5 15.8 18.7 4. 4.7 9.3 9.0 5.9 9.1 9.1 10.2 10.1 7.5 7.2 10.6 8.6 18.2 16.8 5. 6.0 6.3 9.0 5.6 6.2 11.3 10.3 10.9 9.8 9.0 8.2 10.9 17.7 19.4 6. 5.3 5.5 11.5 7.0 10.7 8.4 9.7 11.9 8.0 8.3 10.6 10.2 21.8 20.4 7. 5.5 8.2 9.2 7.0 13.8 13.3 11.3 10.9 9.0 7.9 10.0 9.4 19.2 17.3 8. 5.9 9.5 11.6 6.9 12.3 10.7 12.8 10.3 8.2 9.0 12.7 11.3 16.2 16.6 Un 9. 6.7 8.0 10.3 7.8 11.6 7.6 11.6 12.5 7.9 8.3 11.8 12.6 19.0 18.2 10.. 9.5 6.3 9.1 10.7 13.0 4.4 12.1 12.3 9.3 10.8 11.9 13.4 15.5 14.3 All 5.6 8.1 10.4 8.0 10.0 9.6 11.8 11.8 8.7 8.7 10.4 10.4 18.1 18.1 B. Households Ranked According to Their Total Expendi-ture (THE) or Income (THY) 1. 27.3 37.0 46.3 22.5 29.7 30.0 25.7 25.3 21.4 18.5 17.3 16.5 38.0 36.9 2. 8.2 11.8 17.2 12.3 14.5 16.1 12.8 14.5 14.4 13.0 12.2 11.2 21.7 21.7 3. 5.0 7.2 10.9 9.6 10.5 11.9 13.7 9.9 8.3 9.5 7.6 8.6 19.4 19.5 4. 4.8 4.0 8.4 9.1 8.7 6.2 11.9 10.2 5.8 6.3 9.3 8.4 17.6 18.5 5. 5.3 5.2 6.2 5.3 8.4 8.0 9.9 9.9 6.5 5.0 7.8 9.5 15.5 16.0 6. 1.7 4.3 4.1 5.3 6.2 6.6 9.0 9.5 5.7 6.3 10.3 8.0 14.7 15.3 7. 1.4 3.4 3.8 4.2 6.1 7.1 9.0 10.8 5.4 8.1 8.6 11.4 14.7 13.5 8. 1.5 2.2 3.2 5.0 5.4 4.0 8.3 8.9 7.6 8.0 11.0 8.2 14.4 15.2 9. 0.6 4.4 2.6 2.8 6.1 4.4 10.5 10.7 4.7 4.6 8.8 9.9 13.5 14.3 10. 0.6 1.9 1.5 3.8 4.3 2.2 7.5 8.4 7.2 7.5 10.8 12.1 11.0 9.9 All 5.6 8.1 10.4 8.0 10.0 9.6 11.8 11.8 8.7 8.7 10.4 10.4 18.1 18.1 Notes: 1. When two columns are shown for an area Column A refers to deciles demarcated on the basis of expenditure data and Column B to those on the basis of incoone data. For other reeions. only exDendit:ure deciles have been demarcated. 2. All deciles are deciles of households and not of population. -56- Figure 2 PERCENTAGE OF FEMALES AMONG HOUSEHOLD HEADS RURAL GUJARAT URBAN GUJARAT RURAL MAHARASHTRA 40- * S URBAN MAHARASHTRA NEPAL 11 TOWNS NEPAL 7 TOWNS O GS 0 20. O 40 10 S 0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10 3 20 PENINSULAR MALAYSIA SRI LANKA TAWAN 1974 400. 30 20 0 2 4 6 8 10 2 4 6 810 0 2 4 6 8 10 DECILE OF HOUSEHOLDS RANKED BY PER CAPITA TOTAL EXPENDITURE PCE DECILE ............. . THE DECILE Table 14 fonthly flousehold and Per Capita Expenditure or Income According to the Sex of the Household Head, in Various Areas/Regions Peninsular Gujarat Maharaahtra Nepal Sri Lanka Taiwan Taivan Malaysia (Rs) CRs) (Re) (Rs) (NT$) (NTSI (f) variable/Sex of the Household Read - Rural Urban Rural Urban 11 Touns 7 Towns (1972-73) (1972-73) (1973-74) (1974-75) (1969-70) (1968) (1974) (1973) A B A B A B A B (A) Monthly Household Expenditure/Income (1) Male Heads 302.0 339.4 230.7 333.8 518.3 607.0 332.2 331.6 3048.1 3368.6 7205.3 8464.2 324.0 389.5 (if) Female Heads 150.7 189.3 109.8 217.8 334.6 364.2 273.3 270.0 2502.3 2887.7 6935.7 8417.8 240.5 256.9 (ti) All Heads 293.4 327.3 218.1X 324.5 500.0 583.7 325.3 324.3 3000.7 3326.9 7177.3 8459.4 308.9 365.5 (tv) Relative Level for Females, i.e.. (1t) as %of (1) 49.9 55.8 47.6 65.3 64.6 60.0 82.3 81.4 82.1 85.7 96.3 99.5 74.2 66.0 (B) Monthly Per Capita Expenditure/Income **I (1) Male Heads 50.3 61.7 41.2 65.7 92.4 114.1 55.6 55.5 511.9 565.8 1331.5 1564.2 57.3 68.9 (it) Female Heads 54.0 54.4 37.5 64.6 100.9 99.3 54.3 53.7 537.0 619.7 1419.5 1730.4 56.1 59.9 (Lit) All Heads 50.4 61.3 41.0 65.6 92.9 113.1 55.4 55.3 513.7 569.5 1339.8 1579.2 57.1 67.6 (tv) Relative Lavel for Females, i.e.. (ii) as a of (1) 107.3 88.2 90.9 98.4 109.2 87.0 97.7 96.8 104.9 109.5 106.6 110.6 97.9 86.9 (C) Average Household Size (f) Male Needs 6.0 5.5 5.6 5.1 5.6 5.3 6.0 6.0 5.4 5.6 (ii) Female Heads 2.8 3.5 2.9 3.4 3.3 3.7 5.0 4.7 4.9 4.3 (Hi) All Heads 5.8 5.3 5.3 5.0 5.4 5.2 5.9 5.8 5.4 5.4 (1v) Relative Level for Females, i.e., (ii) as %of (1) 46.4 63.8 52.3 66.3 59.1 68.8 84.1 78.3 90.3 75.9 (n) Number of Sample Households (1) Hale Heads 5217 3277 4759 10287 3954 2037 8466 2535 4695 5960 (it) Female Heads 343 268 555 816 438 216 1198 241 543 1313 (111) All Heads 5560 3545 5314 11103 4392 2253 9664 2776 5238 7273 Notes- 1. All figures bve been rounded independently. 2. Wan two columns are shows for an area , Column A refers to expenditure and Colunm B to income. 3. The Tains date peartain to "ceas huseaeld beads" and not "economic household heads". These terms are discussed elsehere in the paper. -58- lower than that of households with male heads. The two factors compensated one another and the monthly PCE or PCY of households with female heads was, at the most, 14 percent lower (and at its best, ten percent higher) than that in households with male heads. The factors underlying the lower average size of households with female heads are suggested by the data shown in Table 15. Between 60 and 80 percent of females who were reported as heads of households were widowed, divorced, or separated. The risk of widowhood certainly rises with age; and as a logical corollary, the percentage of female household heads also shows a general tendency to rise with the age group of heads. Also, the percentage of female heads generally declines with an increase in the size of households. An alternative approach to answering the question under discussion is to look at the percentage of females among all persons in households in different deciles. The relevant data are shown in Table 16 and Figure 3. According to them, even with PCE or PCY ranking, the percentage of females in the population in the bottom one or two deciles does exceed the average; the implied overrepresentation of women was the highest (between five and seven percent) in the bottom decile in urban Maharashtra and the seven towns of Nepal. With the alternative ranking in terms of THE or THY, one finds more marked overrepresentation of females in the bottom decile in rural areas of Maharashtra, Malaysia, and Sri Lanka, but in some other areas the bottom decile seems to have a sizable deficit of females. It is conjectured that lone male migrants to urban areas forming single-member households often Table 15 (A) Percentage of Females Among Reads of Households by (i) Age of leads and (i) Size of Households and (B) Marital Status Distribution of Female Heads In Different Areas Peninsular Guiarat Maharashtra Nepal Sri Lanica Taiwan Malaysia url ULrban Rural Urban 11 Towns 7 Towns (1972-73) (1972-73) (1973-74) (1974-75) (1969-70) (1968) (1974) (1973) (A) Percentage of Females Among Heads of HoLseholds (1) Age Group of Heads Less than 20 8.2 15.9 25.0 13.4 4.8 3.9 37.5 39.1 20.0 23.5 20-24 2.5 0.9 10.8 7.7 6.1 7.9 8.6 22.9 19.7 19.9 25-29 1.9 2.2 5.3 4.2 5.8 6.0 6.8 7.1 12.5 10.5 30-34 2.5 3.9 8.5 3.8 6.2 5.6 6.4 7.0 7.5 10.0 35-39 3.2 4.3 10.2 6.0 8.0 7.1 8.0 6.5 10.7 12.5 40-44 6.3 5.1 9.0 6.7 10.9 9.5 8.3 6.7 8.8 18.5 45-49 6.1 11.1 9.4 10.1 9. 9 13.8 12.3 7.8 8.3 18.7 50-54 7.0 12.5 11.3 11.3 15.4 11.6 13.1 8.6 9.4 21.6 55-59 6.1 9.9 12.9 11.0 12.7 16.4 15.4 11.6 12.9 26.1 60+ 11.6 20.1 13.4 13.3 15.3 16.7 17.4 14.5 15.9 24.4 All 5.7 8.1 10.4 8.0 10.0 9.6 11.8 8.7 10.4 18.1 (ii) Size of Households 1 48.1 37.1 60.6 16.9 36. 3 20.6 - 16.4 17.1 35.3 2 15.0 11.8 20.5 14.9 19.2 16.1 23.0 17.3 21.2 30.1 3 8.9 13.8 15.0 8.9 11.4 14.2 17.6 13.9 11.2 26.0 4 5.0 5.8 9.0 9.2 10. 7 8.8 14.7 10.6 10.8 18.3 5 2.9 4.3 6.3 6.2 7.0 9.1 9.9 11.0 9.4 15.4 6 2.7 5.5 4.3 3.4 5. 1 6.2 10.4 6.4 9.5 13.6 7 1.4 3.3 2.9 4.1 3. 1 6.8 9.5 5.0 7.5 11.6 8 1.2 3.9 1.6 3.1 2. 7 3.2 6.7 7.1 10.2 10.1 9 - 3.2 1.5 2.2 5. 7 3.4 8.0 4.7 10.8 11.7 10 or more - 2.3 0.8 3.3 1. 8 1.9 7.5 1.5 5.2 9.3 All 5.7 8.1 10.4 8.0 10.0 9.6 11.8 8.7 10.4 18.1 (B) Marital Status Distribution of Female Heads in Different Areas Marital Status Never married 3.4 3.6 3.2 14.1 4. 1 4.1 2.0 N/A N/A 6.9 Currently married 15.2 28.5 16.2 20.5 24. 7 36.9 24.4 N/A N/A 33.1 Widowed, Divorced or Separated 81.4 67.8 80.6 65.1 71.2 59.0 73.6 N/A N/A 60.0 All 100.0 100.0 100.0 100.0 100.0 100.0 100.0 N/A . 100.0 Sample Households (343) (355) (268) (816) (435) (216) (1198) (241) (543) (1313) Table 16 Percentage of Females Among All Persons in Households in Different Deciles, With Alternative Ranking Criteria Peninsular Decile of Gujarat Maharashtra Nepal Sri Lanka Tairwan Malaysia Households Rural Urban Rural Urban 11 Towns 7 Towns (1972-73) (1972-73) (1973-74) (1974-75) (1969-70) (1968) (1974) (1973) A B A B A B A B A. Householdo Ranked According to Their Per Capita Expenditure (PCE) or Zncome (PCI) 1. 50.7 49.8 49.3 49.9 48.3 47.7 49.6 48.8 49.4 48.6 48.5 .48.0 46.7 46.8 2. 51.6 50.4 51.0 50.7 50.3 50.5 51.1 50.2 50.0 49.8 49.8 49.8 48.8 48.6 3. 50.6 50.5 49.9 50.8 49.6 48.3 49.4 50.5 49.0 49.3 49.9 50.0 49.1 49.1 4. 51.4 51.1 47.4 52.7 49.7 50.0 50.6 49.8 47.7 49.2 49.7 49.8 49.2 48.7 5. 49.9 51.6 51.2 52.5 51.2 48.3 52.2 50.3 50.6 49.6 50.6 .51.6 49.1 48.9 6. 51.3 51.0 48.7 55.0 49.6 51.1 50.7 51.0 50.6 51.8 51.2 51.3 48.7 49.2 7. 51.5 50.3 51.7 55.5 50.4 50.5 50.4 51.1 50.9 49.5 51.3 51.7 48.8 48.8 8. 51.7 51.5 51.0 56.4 50.9 52.9 49.4 51.1 51.1 51.3 52.2 50.6 50.8 51.0 1 9. 50.2 51.9 50.8 57.9 52.9 57.2 51.3 51.5 52.0 52.3 51.8 53.0 51.4 50.8 10. 51.5 53.9 55.6 59.0 58.3 64.9 51.1 52.2 54.1 54.4 52.3 51.4 50.9 51.5 0 J All 51.1 51.2 50.2 53.4 50.8 51.2 50.6 50.6 50.3 50.3 50.5 50.5 49.2 49.2 B. Households Ranked According to Their Total Expenditure (THE) or Incone (THY) 1. 50.4 49.5 41.5 60.4 47.9 49.5 48.0 47.6 55.3 54.9 53.4 51.9 46.1 44.7 2. 51.1 50.9 48.5 56.9 49.4 51.1 50.2 49.0 51.7 50.9 51.2 51.4 47.1 48.5 3. 51.1 50.2 50.3 53.5 50.7 49.7 51.9 51.2 50.3 49.2 50.9 51.5 49.7 48.8 4. 51.6 52.1 50.1 51.9 51.0 52.5 50.5 49.5 48.8 50.2 49.6 49.7 49.5 49.4 5. 50.8 52.0 48.3 52.7 51.0 51.1 48.6 50.3 49.4 51.3 50.1 49.7 49.4 49.7 6. 50.7 48.9 50.8 53.1. 50.4 50.2 50.9 50.9 50.7 49.5 49.9 51.0 50.5 51.5 7. 52.0 53.2 51.2 52.3 51.8 51.2 50.7 50.9 50.2 50.7 51.0 50.3 50.6 49.2 8. 50.8 52.4 50.1 53.4 50.5 51.6 51.0 50.4 50.7 49.6 50.7 50.6 48.6 49.3 9. 50.0 50.6 51.8 52.9 50.3 51.8 50.8 51.8 49.8 49.0 49.7 50.2 50.2 50.0 10. 51.6 51.2 52.4 53.1 51.8 51.7 51.6 52.0 49.1 50.0 50.3 49.8 48.4 48.8 All 51.1 51.2 50.2 53.4 50.8 51.2 50.6 50.6 50.3 50.3 50.5 50.5 49.2 49.2 *Figures pertain to population aged 5 and o-ver only. Notes: 1. When two columns are shown for an area , Column A refers to deciles demarcated on the basis of expenditure data and Column B.to those on the basis of income data. For other regions, only expenditure deciles have been demarcatec 2. All deciles are deciles of housseholds and not of population. -61- Figure 3 PERCENTAGE OF FEMALES AMONG ALL PERSONS RURAL GUJARAT URBAN GUJARAT RURAL MAHARASHTRA 55 - 45 40* 35 30- l i f t i l l i I I 0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10 URBAN MAHARASHTRA NEPAL 11 TOWNS NEPAL 7 TOWNS 55- 50-. 46 40 * 35- 30- 0 2 4 6 a 10 0 2 4 6 8 10 0 2 4 6 8 10 PENINSULAR MALAYSIA SRI LANKA TAIWAN 1974 60 55 so - . . ....... ~.C 45- 40 35 30 fi I I I I I II I 0 2 4 6 B 10 0 2 4 B 8 10 0 2 4 6 B 10 .THE DECILE DECILE OF HOUSEHOLDS RANKED BY PER CAPITA AND TOTAL EXPENDITURE PCE DECILE -62- fall in the bottom decile in terms of THE or THY and lead to the observed dfidt of famal. On the whole, our data provide rather limited confirmation of the hypothesis that women are overrepresented among the poor. Moreover, the extent of overrepresentation appears to be relatively small. The possible explanations of the absence of any strong evidence in support of the hypothesis include the fact that we have not explored the sources of income of house- holds with a female head and not identified the extent to which they depend on transfers or borrowings from other households..- Since a majority of female heads of households are widowed, divorced, or separated, transfers 1/ Uhen households are ranked according to per capita expenditure or income, the lone migrants would probably fall in the top decile. Such single- member households of migrant males are presumed to dominate the top decile in urban Maharashtra and towns of Nepal. Unfortunately, the surveys included in the project have not collected any information on the migration status of the household or the individual members. Given the strong likelihood that the migrants continue links with households in their place of origin and undertake transfers, such information can be very useful and its collection deserves to be encouraged. The Indian national sample survey in 1972-73 recorded the migration status of the household during the preparation of the frame to select the sample; but it distinguished between only (a) "temporary in-migrant households" and others. Also, the information was not transferred to the schedule on which the information gathered during the detailed interview was recorded. In Taiwan, the data collection procedures require enumeration of out-migrants with the family at the place of origin, if their "economic livelihood" or budget is related or shared. However, the data available to us do not permit an identification of such households. 2/ If they rely on transfers on a long-term basis, one can question the validity of treating them as separate households for purposes of studying the size distribution of income. However, few surveys would be able to abandon the widely used definition of a household as the group of persons sharing a common kitchen and living together or having common arrangements for supplying basic living needs. -63- could indeed be an important source of their support (although the widowed, divorced, and separated females are known to report much higher rates of work participation than their never-married or married sisters in several states in India)*- Also, an examination of data for smaller groups of house- holds than deciles might show a different picture. It is equally likely that the widespread impression of an overrepresentation of females among the poor could be based on the data examined in terms of total expenditure or income of households. Of course, the smaller-than-average size of households with female heads denies them some of the economies of scale in consumption; and they may be vulnerable in terms of their assets. In terms of their living standards measured in per capita terms, however, they do not seem to be heavily overrepresented among the poor. 4. Differentials in Dependency Ratios or the Proportion of "Employable" Persons An important issue is the extent of variation in the proportion of "employable" or "working age" persons in different income or expenditure groups. If the poor contain disproportionate numbers of the very young and the very old, their income-earning capacity would be restricted even if public policy were to ensure full employment of persons in working ages. This issue has received little attention in the literature because of the paucity of data on the subject. 1/ This observation is based on a special tabulation and analysis of the 1961 census data of India for the five states of Bihar, Himachal Pradesh, Maharashtra, Punjab, and West Bengal. See: Pravin Visaria, "The Level and Nature of Work Participation by Sex, Age and Marital Status in India, 1961," paper presented at the Indian Census Centenary Seminar, New Delhi (October 23-29, 1972), mimeographed. -64- Table 17 and Figure 4 show the relevant data in terms of the age- dependency ratios (defined as the number of persons aged 0-14 and 60 and over--or 65 and over--per thousand persons in the working ages of 15-59-- or 15-64)1/ for different deciles in terms of both the per capita and total household expenditure or income.-/ With a ranking in terms of PCE or PCY, the age-dependency ratio varies inversely with decile, virtually monotoni- cally. In rural areas of Gujarat and Maharashtra, the dependency ratios of the bottom and the top deciles of households differ by less than a factor of 2:1; but in other areas, the ratio of the two ratios approaches (or even exceeds) 3:1. When households are ranked according to their total household expenditure or income there is no clear relationship between decile and the dependency ratio.A In other words, although the relationship between 1/ We have chosen age 60 as the upper limit of working ages in Gujarat, Maharashtra, Nepal, and Sri Lanka because the public sector employees usually retire before reaching that age. In Taiwan and Malaysia, on the other hand, the retirement age is around 65 years, and, therefore, ages 15-64 are taken as the working ages. 2/ The well-known errors of age reporting--preference for ages ending in 5 and 0 and, to a lesser extent, 2 and 8--do affect our survey data; but the smoothing of age data that is commonly used in the analysis of census data seems hazardous with the relatively small number of sample households or persons in each decile. The broad age groups help to mitigate the influence of these errors. Besides, the differentials follow a consistent pattern and are so large that the errors cannot be a major contributory factor. 3/ The inter-decile range of variation in dependency ratios is always smaller with the ranking of households in terms of total household expenditure or income than with per capita expenditure or income ranking. -65- !ab1e 17 Age Dependency Ratios in Different Deciles with Alternative Ranking Criteria Peninsular Lecile of ;ara z Maharmhtra Nepa? Sri Lank-- fe_wan Mfta=aia Bouseholds Rural Urban Rural Urban 11 Towns 7-Tons (1972-73) (1972-73) (1973-74) (1974-75) (1969-70) (1968) (1974) (1973) A B A U A U A B A. boueehosd Ranked According to Their Per Capita Expenditure (PCE) or Income (PC) 1. 1165 1150 1256 1205 1279 1235 1233 1319 1302 1303 1173 1195 1325 1244 1. 1220 080 1176 1022 1135 1123 1105 1138 1099 1101 996 1047 1175 113 3. 1089 924 1134 927 979 1073 927 1033 1071 1124 877 923 1116 1181 4. 933 847 1033 838 915 914 895 886 962 976 880 882 1038 1103 5. 891 883 961 741 787 815 833 846 829 842 761 692 1010 1031 6. 928 749 926 713 779 812 757 749 774 745 650 620 933 905 7. 885 639 909 575 669 819 721 707 728 724 570 571 831 820 6. 814 594 858 516 628 653 692 575 648 597 499 471 671 699 9. 707 549 803 405 877 592 572 608 527 510 439 452 523 510 10. 657 388 729 328 417 352 525 506 386 404 305 327 371 368 All 922 790 1002 762 821 855 831 831 846 846 725 725 901 901 B. Bousehods Rarked According to Their TotaZ Expenditure (TEE) or Income (THY) 1. 637 684 739 475 615 658 833 1005 734 815 806 870 850 951 2. 814 715 777 597 750 692 845 931 881 944 855 932 927 947 3. 928 759 919 738 900 811 872 919 973 917 869 923 1042 1051 4. 958 781 924 844 775 864 877 931 938 1004 925 892 977 1023 5. 973 845 998 879 921 913 910 941 1037 1014 885 862 982 1002 6. 990 921 1010 829 906 956 905 886 952 906 847 804 995 1000 7. 993 834 1043 862 865 967 837 815 849 866 747 674 923 977 E. 935 835 1074 805 825 837 799 777 767 758 560 628 968 915 S. 9;7 )92 1052 738 853 858 832 723 798 752 571 557 833 783 10. 863 680 1093 631 746 849 688 632 649 656 473 472 682 629 All 922 790 988 756 821 835 831 831 846 846 725 725 901 901 Botes: 1. Dependency ratios for Gujarat, Maharashtra, Nepal and Sri Lanka show persons aged 0-14 and 60 and over per 1000 persons aged 15-59. 2. Dependency ratios for Taiwan and Peninsular Malaysia show persons aged 0-14 and 65 and over per 1000 persons aged 15-64. 3. Dependency ratios for all deciles together for rural and urban Maharashtra differ slightly because some sample households were excluded during editing after the first runs in terms of PCE decile. 4. When two columns are shown for an area , Column A refers to deciles demarcated on the basis of expenditure data and Column B to those on the basis of income data. For other regions, only expenditure deciles have been demarcated. 5. All dicles are deciles of households and not of population. -66- -Igure 4 AGE DEPENDENCY RATIOS RURAL GUJARAT URBAN GUJARAT RURAL MAHARASHTRA 1400 1200 - 00 O ..**.. .. .. .. .. .. .* 1000 - . .......* rp eoo - % m * 4* 1400 W2 100 200 0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10 URBAN MAHARASHTRA NEPAL 11 TOWNS NEPAL 7 TOWNS 1400 1200 2 1000 s0Goo - -. 0 *, w* ~40 200 - I I I I I I I 0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10 D PCE eIYle .................................. ......THE Deciles PENINSULAR MALAYSIA SRI LANKA TAIWAN 1974 1400 1200- S1000 - .. C)r . ...... .... . *.*4 z4 CL 600 4 400 - 200- o 2 4 6 8 10 0 2 4 6 a 10 0 2 4 6 a t0 DICILE OF HOUSEHOLDS RANKID BY PER CAPITA TOTAL EXPENDITURE -67- household size and THE decile is exactly opposite to that between size and PCE decile, the same is not true of dependency ratios. Households falling in the upper THE deciles have a larger-than-average size but, except in rural Maharashtra, they do not necessarily report higher dependency ratios. In most of the data sets, the relationship between dependency ratio and THE decile assumes an inverted-U or.-V shape, implying that the households in middle deciles have a higher percentage of persons in dependent ages, i.e., mainly children aged 0-14, than the low or the high deciles. Table 18 and Figure 5 show the percentages of children aged 0-14 in the total population of households in different deciles according to alternative ranking criteria. It is evident that differences in the percentages of children follow the same pattern as do the age-dependency ratios shown in Table 17. With households ranked according to their-per capita income or expenditure, the bottom deciles have a much higher-than- average proportion of children. With the alternative ranking, the middle deciles have the higher-than-average proportions of children. On the basis of these data, the poor defined in terms of the PCY or PCE ranking seem to be handicapped in terms of the proportion of employable persons or persons in the working ages. They shoulder a high burden of dependency, largely of young children; and this constraint would persist even with full employment. Any short-run relief program could probably be organized best through social transfer schemes, designed to . Table 18 CHILDREN AGED 0-14 AS PERCENT OF TOTAL POPULATION IN DIFFERENT DECILES WITH ALTERNATIVE RANKING CRITERIA PeninsuZc-r Decile of Gujarat Maharashtra Nepal Sri Lanka Taiwan Ma.aysia Households Rural Urban Rural Urban 11 Towns 7 Towns (1972-73) (1972-73) (1973-74) (1974-75) (1969-70) (1968) (1974) (1973) A B A B A B A B A. Households Ranked According to Their Per Capita Expenditure (PCE) or Income (PCY) 1. 50.5 50.0 50.7 49.4 52.9 51.6 51.0 53.1 52.7 52.8 50.0 50.7 54.4 52.3 2. 48.2 45.9 48.9 45.5 48.0 48.4 47.6 48.9 49.2 49.6 47.1 48.3 50.7 51.3 3. 48.7 43.9 46.6 43.4 45.4 47.4 43.5 45.9 48.6 50.1 43.7 45.2 49.8 51.0 4. 43.2 41.1 44.0 41.0 44.2 43.0 41.6 40.9 46.9 46.9 44.6 44.6 47.4 49.3 5. 42.0 41.8 42.8 37.8 39.9 40.4 39.6 40.1 43.0 42.8 40.7 38.6 47.1 47.3 6. 43.1 37.0 41.6 36.5 37.1 39.8 36.1 36.3 41.0 40.3 37.2 36.4 45.0 44.3 7. 41.1 34.5 40.5 31.2 34.4 38.8 34.6 34.8 39.1 39.5 34.3 33.9 42.1 41.5 8. 38.2 32.5 37.5 28.7 32.6 34.2 32.6 28.4 37.1 34.6 31.2 29.4 35.8 37.6 9. 34.8 29.4 36.9 22.7 30.4 32.5 28.5 28.9 32.0 31.4 27.9 29.1 30.6 30.2 10. 32.1 22.4 36.0 18.1 21.0 21.9 27.3 23.7 24.8 25.8 20.8 21.6 23.7 23.8 All 42.9 39.3 43.1 37.9 39.7 41.4 39.1 39.1 43.0 43.0 39.4 39.4 44.1 44.1 B. Households Ranked According to Their Total.Expenditure (THE) or Income (THY) 1. 29.1 22.6 28.7 23.8 26.7 29.5 31.0 38.7 34.5 37.3 39.2 41.4 35.4 39.3 2. 29.8 36.2 36.5 31.9 38.1 35.0 39.2 42.6 44.5 46.1 43.4 45.7 43.9 44.2 3. 43.9 39.3 42.0 37.6 41.7 40.0 40.8 42.8 46.7 45.2 44.2 45.5 47.7 4.0 4. 44.7 39.9 42.2 41.0 40.0 42.7 40.5 43.2 45.8 46.9 45.9 45.0 46.4 47.9 5. 44.7 42.5 43.6 42.0 43.2 42.8 42.2 42.7 48.0 48.5 44.7 43.6 46.8 47.4 6. 45.3 43.9 45.3 41.0 42.9 45.1 42.7 41.6 46.4 44.9 43.2 42.2 47.2 47.5 7. 44.6 41.9 45.2 41.7 40.9 45.2 40.3 39.2 43.6 44.0 40.1 38.1 45.2 46.9 8. 43.5 40.6 45.7 39.9 40.1 41.7 38.6 37.2 41.0 40.9 35.2 36.2 46.7 45.7 9. 43.2 39.2 44.1 36.6 40.6 41.8 39.9 35.4 41.8 40.3 33.6 32.9 42.8 40.6 10. 41.3 35.5 44.6 33.1 37.7 41.1 34.4 31.9 36.4 36.5 29.5 29.4 37.7 35.9 All 42.9 39.3 43.1 37.9 39.7 41.4 39.1 39.1 43.0 43.0 39.4 39.4 44.1 44.1 Notes: 1. When two columns are shown for an area ,Colunn A refers to deciles demarcated on the basis of expenditure data and Column B to those on the basis of income data. For other regions, only expenditure deciles have been demarcated. 0 2. All deciles are deciles of households and not of population. -69- Figure 5 PERCENTAGE OF CHILDREN (0-14) AMONG ALL MEMBERS RURAL GAJARAT URBAN GUJARAT RURAL MAHARASHTRA 60 60 20 10 IURBAN MAHARASHTRA NEPAL 11 TOWNS NEPAL 7TOWNS O 60 50 40 - iu20 » 10 0 . 0 2 4 6 8 10 0 2 4 6 a 10 0 2 4 6 8 10 z PENINSULAR MA LAYSIA SRJ LANKA TAIWAN 1974 60 30 10 0 2 4 a a 10 0 2 4 6 6 10 0 2 4 6 a 10 ....... PCE Detiles DECI LE HOUSEHO LDS R ANKED BY PER CAPITA AND TOTA L EXPENDITURE 40.......... THE Dedit. -70- benefit the young children.1/ In the long-run perspective, however, the family planning programs aimed at lowering fertility would have a direct impact on the level of dependency ratioB of households in different deciles,2/ While the decile-related differences in dependency ratios are large and important, they can be at least partly offset by different behavior with respect to the extent of participation in economic activity. The latter differences are examined later in this paper by looking at labor force participation rates. We shall first look at the differences in educational attainment or the human capital endowment of adults, an important factor influencing the nature of economic activity. 1/ Insofar as the incidence of unemployment and/or underemployment is greater among the poor, programs to augment the work opportunities would also provide quick relief to the poor. 2/ Apart from the life cycle factors, differentials in fertility, mortality and the rate of natural increase might also be contributing to the observed large decile-related differences in dependency ratios (and the underlying age distribution of population). However, none of the data sets permits any estimation of differentials in the rates of natural increase. An exploratory exercise based 6n tha data for Cuj&rAt indicates that even if there were no decile-related differences in age- specific fertility and mortality rates, the observed differences in age distribution would imply rather significant differences in the crude birth and natural increase rates of different deciles; but the rate of natural increase would be positively related to decile. If the rate of natural increase is to show an inverse relationship with MPCE decile (as is sometimes postulated), the differentials in sex-age specific mortality and fertility must be large enough to neutralize the effect of differentials in age composition. Annex 9 reports briefly on the results of this exercise essentially to elicit further comment and reactions on this complex subject. -71- B. Human Capital Dimensions Human capital is the cumulative result of investment or expenditure that raises the skills, ability, and the earning capacity of individuals. Education, job training, health, and nutrition are all important dimensions of human capital. It is generally believed that the poor are poor in terms of their human capital endowments and include a much higher-than-average proportion of the illiterates than persons with a low level of education. Further, their handicaps are likely to be perpetuated because the children in poor households either do not enter schools or drop out after very few years of schooling. Our data provide an opportunity to assess the validity of these beliefs. Among the human capital dimensions examined in this study are the decile-differences in (a) illiteracy rates and the educational attain- ment of literates among persons aged 15 and over, and (b) the school enroll- ment rates of children and youth in ages 5-24. In the next section we shall assess the importance of differences in the educational attainment of (a) employee workers in determining their wage and salary earnings and (b) household heads in influencing the level of per capita income or expenditure of the household. The importance of literacy and/or education as a key element indicating the living standard of an individual is widely recognized (although illiteracy certainly does not mean ignorance or absence of some valuable skills). Among the surveys studied in our project, the Malaysian -72- HES and the Nepal surveys did not explicitly distinguish between illiterates and literates.- They identified persons with no formal schooling as a separate group while recording the highest education attained. The Sri Lankan Socio-Economic Survey had an explicit question on the subject, whereas the surveys in Indian regions and Taiwan recorded illiterates as a separate category in response to the question on educational attainment.-/ I/ According to the 1970 Census of Malaysia, 60.8 percent of the population aged 10 and over (72.1 percent of males and 49.6 percent of females) in Peninsular Malaysia were literate in some one language. An additional 4.8 percent were "semi-literate" (i.e., they could read but could not write). See: Malaysia, Department of Statistics, 1970 Population Census of Malaysia, General Report, Vol. 1 (Kuala Lumpur, 1977) p. 322. In Nepal, the 1971 Census had enumerated 87.5 percent of the population aged 15 and over (77.6 percent of males and 97.4 percent of females) as illiterate. See: Nepal, Central Bureau of Statistics, Population Census - 1971, Vol. II, Social Characteristic Tables, Part II (Kathmandu, 1975). 2/ Literacy rates for persons aged 15 and over in these areas were as follows. Area Year Source Males Females Persons Gujarat Rural 1971 Census 47.3 17.7 32.7 Urban 1971 Census 77.2 49.5 64.2 Maharashtra Rural 1971 Census 52.0 16.3 34.1 Urban 1971 Census 78.0 51.5 66.6 Sri Lanka 1971 Census 86.0 68.5 77.6 Taiwan End of 1974 Household 91.6 74.2 83.4 Register -73- Our data on school enrollment rates for ages 5-24 are a by-product of the recorded activities,of those classified as economically not active (i.e., neither employed nor unemployed). Implicitly, therefore, the data assign some priority to a person's classification as a worker rather than as a student; but presumably (although not explicitly) the time spent in work would influence the respondent's report on the interviewer's classification of an individual as a worker or tudent. The labor force participation rates for young children aged 5-9 are so low that, the bias introduced by our relying on their economic activity data is not significant. For ages 10-14, the participation rates in rural areas of Gujarat and Maharashtra or towns of Nepal are high enough to introduce some downward bias in our estimates of student-population ratios, particularly for the bottom deciles (because within the age group 10-14 and 15-19, participation rates seem to vary inversely with PCE decile). The effect of such a bias cannot be estimated but is believed to be relatively small. In Malaysia, the questions on economic activities were asked only of persons aged 15-64; therefore, the percentage of students in the important age group of 5-14 cannot be estimated. Further, as noted above in Section II, literacy rates based on surveys are generally higher, particularly for urban areas when a distinction is possible, than those reported by the nearest censuses, to a larger extent than would be expected on the basis of the time interval between the two operations. Quite probably, the surveys do seem to be somewhat selective -74- of the literate, although the extent of bias does not appear to be large and the inter-decile differences noted below are more or less real. 1. Illiteracy and Educational Attainment The data presented in Figure 6 confirm the expected inverse dSOciation between FCE (or PCY) decile and illiteracy or the percentage of persons with no formal schooling.-/ Figures for the bottom and top PCE deciles have been summarized in Table 19. In the two Indian states or in the towns of Nepal, where a majority of the literates essentially have only primary education, the percentage of persons with primary education rises with PCE deciles. Differences in the proportions of the primary-educated are higher among females than among males, and higher in rural areas than in urban areas. In relative terms, the differences in the proportions of persons with secondary (i.e., high school or matriculation) or higher educations are larger than those with respect to primary or middle school education. On the whole, the educational standards attained by the literates in the lower PCE deciles are markedly lower than of those in upper deciles.-/ An examination of the illiteracy rates by broad age groups (e.g., 15-29, 30-44 and 45-59)3/ in Gujarat and Maharashtra provides an 1/ The estimation of illiteracy rates or the proportion of persons with different levels of education for ages 15 and over serves as a proxy for age-standardization which would eliminate the effect of age dis- tribution on rates for different deciles. 2/ In Taiwan, the percentage of children aged 5-9 not going to school was slightly higher in the top decile than in the bottom decile of house- holds. The difference is small and is presumed to be due to random variation. 3/ The use of broad age groups helps to reduce the effect of errors in age reporting on the estimates. If seven years is taken as the con- ventional or modal age of entry into school, the three cohorts could have entered schools during 1922-34, 1935-50 and 1950-65. -75- Figure 6 ILLITERACY RATES RURAL GUJARAT URBAN GUJARAT RURAL MAHARASHTRA 100 90 80 ** ** * 70- so.1 40-4 1N \ 14 304 20- 8 0\ 10.. g to \ 40 zURBAN MAHARASHTRA NEPAL 11 TOWNS N EPA L 7 TOWNS 100 I 0 IL z 40 U1A1ANRSTR EA 11 TOWN I I IIJTW 10 10- PENINSULAR MALAYSIA SRI LANKA TAIWAN 1974 3 0. 20* 1 0 A. 0 2 4 6 8 1 0 2 4 6 8 0 2 4 6 8 10 DECILE HOUSEHOLDS RANKED BY PER CAPITA AND TOTAL EXPENDITURE MALES ...............*** FEMALES .-- -- -- -PERSONS -76- Table 19 Percentage of Illiterates (or Those with No Formal Schooling) Among Persons Aged 15 and Over in the Bottom and Top Deciles of Households (with Ranking in Terms of Per Capita Expenditure) Males Females Persons Bottom Top Bottom Top Bottom Top Area Decile Decile Decile Decile Decile Decile Gujarat, 1972-73 Rural 66.8 28.6 95.3 62.8 80.7 45.2 Urban 34.2 3.8 69.3 16.7 52.4 9.5 Maharashtra, 1972-73 Rural 63.1 25.3 91.9 66.2 78.3 44.0 Urban 33.7 4.9 71.8 10.6 53.3 7.2 Nepal 11 Towns, 1973-74 50.5 11.9 88.1 34.1 70.2 20.7 7 Towns, 1974-75 59.4 7.1 90.4 31.3 75.6 14.5 Peninsular Malaysia, 1973 38.1 9.3 24.7 4.2 49.3 35.1 Sri Lanka, 1969-70 48.5 4.4 41.6 11.2 30.4 7.7 Taiwan 1968 24.5 3.6 48.8 19*0 37.1 10.7 1974 15.4 1.3 36.3 9.4 26.1 5.2 -77- indication of the changes over time in the extent to which different cohorts have had access to, and/or have taken advantage of, opportunities for schooling and education.-- For each age group, one observes an inverse relationship between FCE decile and illiteracy rate, for both males and females. Also, within each decile (including the lowest), the younger cohorts report lower illiteracy than the older group; but the pace of progress appears to be slower among females, among lower PCE deciles and in rural areas than among males, upper deciles and urban areas. While the rural urban differentials are probably widened through migration of the literate and the better-educated to urban centers, the data confirm the acceleration of progress during the quarter century preceding the survey. A similar change seems to be occurring in Nepal as well, although our sample is drawn from centers where development effort might have been concentrated. Admittedly, since the PCE decile of households is partly a function of the life cycle, the inference of dynamic trends in literacy and education on the basis of data tabulated by age and decile in one year is subject to some error. However, the strong positive association between education and earnings, discussed later in the paper, leads one to believe that the observed differences are to a considerable extent independent of the life cycle factors. In any case, some more direct evidence on the changing relationship between PCE or PCY decile and literacy or education is provided by the school enrollment ratios examined next. 1/ The effects of improvement in access to schools are compounded with those resulting from the increase over time in the real income of parents in households with low per capita expenditure. -78- 2. School Enrollment Ratios The data on school enrollment ratios (available from most of the surveys), examined in this study are summarized in Annex 10. They confirm the expectation of a positive relationship between PCE decile and the school enrollment rates for ages 5-24. The differentials are larger for females than for males and larger in rural areas than in towns. The positive effect of government efforts to provide free and compulsory primary 1/ education- is seen in almost every data set in that even in the bottom PCE deciles, the percentages of children aged 10-14 and classified as students are higher than the proportions of literates among persons aged 15 and over. Since our school enrollment data are subject to some down- ward bias, the real degree of improvement might be somewhat higher. Table 20 focuses attention on children aged 5-19 not going to school (to facilitate comparisons with Table 19). Since literacy can be attained at any age and entry into schools is feasible without much difficulty at least up to age 14, one can focus attention on the age group 10-14. Figure 7 shows the school enrollment ratios for ages 10-14, by sex and decile of per capita expenditure. The data clearly suggest that the proportions of children not in school are lower than the proportions of illiterates (or those with no formal schooling) among persons aged 15 and over in both bottom and top deciles; but the improvement often appears to be greater in the top deciles and for boys, and relatively large inter- decile differentials seem to persist, particularly for girls. Apparently, 1/ In Gujarat and Maharashtra, primary education up to the seventh year is free (unless a household decides to send children to "private" schools). In Sri Lanka, the government decided in 1945 to provide free education "from Kindergarten to the University." In Taiwan, the regional government- prescribed compulsory education was for six years until 1968 and for nine years thereafter. Table 20 PERCENTAGE OF CHILDREN NOT GOING TO SCHOOL AMONG PERSONS AGED 5-19 IN THE BOTTOM AND TOP DECILES OF HOUSEHO"LDS RANKED ACCORDING TO PER CAPITA EXPENDTTURE Ages 5-9 Ages 10-14 Ages 15-19 Males Females Males Females Males Females Bottom Top Bottom Top Bottom Top Bottom Top Bottom Top Bottom Top Region Decile Decile Decile Decile Decile Decile Decile Decile Decile Decile Decile Decile Gujarat, 1972-73 Rural 77.3 46.1 91.4 49.1 66.8 21.6 86.4 SI.8 86.7 54.5 99.3 81.8 Urban 57.9 22.3 69.2 30.5 30.0 2.5 54.8 1-7 61.7 12.6 76.7 41.1 I Maharashtra, 1972-73 Rural 75.4 45.4 83.4 47.1 51.1 17.2 78.9 47.7 81.4 56.4 96.1 92.5 Urban 59.6 13.7 57.9 13.0 22.9 7.8 49.5 5.8 53.4 21.7 74.6 32.0 Nepal 11 towns, 1973-74 70.5 22.2 84.7 28.8 41.8 15.1 75.0 27.7 69.0 40.4 83.2 50.6 7 towns, 1974-75 82.5 18.2 93.0 50.0 65.3 24.2 85.3 20.7 87.1 42.6 94.9 71.0 Sri Lanka, 1969-70 29.7 10.2 3.2 18.1 24.8 17.7 39.0 21 .2 68.9 40.3 76.0 51.1 Taiwan 1968 29.4 31.6 32.6 18.9 26.0 1.9 35.7 10.9 84.7 16.2 92.1 30.9 1174 27.1 29.2 25.9 35.8 7.5 - 13.1 - 69.3 15.0 80.0 19.8 -80- Fig~ 7 STUDENT POPULATION RATIOS BY SEX FOR AGE GROUP 10-14 RURAL GWARAT URBAN GUJARAT RURAL MAHARASHTRA o- 10 70 - 40 30 - - 10 9 1C24 . 8 1. URBAN MAHARASHTRA NEPAL 11 TOWNS NEPAL 7 TOWNS 10 60 - .- - 10 40- 30- 20- 10- 0 2 4 6 B 10 0 2 4 6 8 10 0 2 4 6 8 10 UR C OFA UHDA ANEPA 1 ER CAA A E NTEL7 0 80 / 70 SCI. I. 40- 30 i 20. 10 0 4 Lo la 10 0 .2 4 6S 8$ 10 0 2 4 6S 8S 10 EIL O LANKAOLO RADBYPW CIAND 1968L TAIWANI1974 -81- quite a few parents continue to attach low priority to the education of daughters. Of course, "free" education normally implies only free tuition, and the opportunity cost of schooling of children continues to be felt by the parents or the household as a whole. Vence, the possibility of entry into the labor force leads the children from households in the bottom PCE or PCY deciles to drop out of school (except in Taiwanese economy, or the developed countries, where the government prescribes and is able to enforce the minimum age for leaving school and taking up employment). C. Economic Characteristics of the Population The economic characteristics of the population in different deciles considered in this section include (a) labor force participation rates, (b) the incidence of unemployment (or unemployed as a percent of labor force) and (c) the distribution of the employed by class of worker, occupation, and industry. The effect of using deciles based on different ranking criteria is discussed with respect to participation rates and the incidence of unemployment. The discussion of various characteristics of the employed in different deciles is limited to deciles of per capita expenditure or income.- 1/ Some caution is necessary in drawing inferences from any comparison of participation rates and measures of unemployment as well as other characteristics based on different data sets because of differences in the underlying concepts and definitions. A detailed discussion of these concepts and the extent to which the resulting data differ from other estimates is available in the draft papers listed in the Preface. On the whole, only the Malaysian data are based on the activities of respondents during a specified reference week (in the month in 1973 in which the particular household was surveyed). All other data pre- sented below relate essentially to the "usual status" of respondents. The surveys in Gujarat and Maharashtra also provide estimates of participation rates based on the activities of the reference week preceding the date of interview; but these rates are not significantly different from those based on the usual activities of the population, which are shown in the tables in this paper. -82- The determination of the economic activity status of an individual-- that is, an identification of him as in or outside the labor force and employed or unemployed during a specified period of time--requires the use of a rather arbitrary cut-off along a broad, nebulous spectrum. The difficulty is greater when the reference period is longer than say one week or fortnight, although for evaluating the relationship between economic activities and living standards, the "usual status" approach (i.e, iden- tification of the usual major activity of a person over a relatively long period such as a year) appears advantageous. Except for Peninsular Malaysia our data sets do indeed approximate the usual activity approach. Few studies have examined the relationship between the level of 1/ labor force participation and income or expenditure deciles;-- but there is a widespread impression that children and women in poor households participate in economic activity to a much greater extent than do their counterparts in better-off households. If this impression is correct, the labor force participation rates should vary inversely with decile. The criterion used to rank households has a clear relevance to this issue because of the large inter-decile differences in the proportions of persons outside working ages and the divergence in the pattern according to the ranking criterion used. If households are ranked by per capita 1/ Several studies have focussed on labor force participation rates of females, which tend to vary more over time and between countries than do the male participation rates. -83- expenditure, in the absence of age standardization, one would observe a positive relationship between crude (i.e., for all ages) participation rates and per capita expenditure decile at least for males because an over- whelming majority of males ageA 1 and over are normally reported to be economically active. For females, the relationship is likely to depend on many variables, including customs (or notions) about what constitutes work and whether it should be reported. With households ranked according to their total expenditure, however, labor force participation rates would be higher in the lower deciles, which have relatively lower dependency ratios according to Table 17. 1. Crude Labor Force Participation Rates According to our data, shown in Annex 11, crude labor force participation for both sexes together and for males and females separately generally vary inversely with decile when total expenditure is the ranking criterion.-/ The inter-decile differences in crude participation rates (for both sexes together) are smaller when ranked according to the per capita criterion than to the total household expenditure, except in Peninsular Malaysia and Sri Lanka, where the opposite is true. Furthermore, the crude participation rates do not always vary inversely with per capita expenditure decile. In some data sets, the positive relationship between male labor participation rates and per capita expenditure decile is partly 1/ In Peninsular Malaysia and in the economy of Taiwan in 1974, the crude labor force participation rates for both sexes together have a U-shaped relation- ship with total expenditure decile, probably because of a similar relationship between the percentage of children aged 0-14 and decile (see Figure 5 above). -84- offset by a mildly inverse or even slightly U-shaped relationship between female participation rates and decile. For both sexes together, the crude participation rates rise as the decile rises in urban areas of Gujarat and Maharashtra (essentially in the top four deciles), in Peninsular Malaysia and in Sri Lanka. In other cases, the decile differences in crude labor force participation rates for both sexes together are relatively small and show no clear pattern. This happens although the participation rates of children aged 10-14 or youth aged 15-19 show a roughly inverse relationship with per capita expenditure decile in all data sets except Taiwan. 2. Participation Rates for Ages 10 and Over An attempt has been made to examine the relationship between per capita expenditure decile and participation rates.after excluding from consideration the age groups 0-9 or 0-14 in order to minimize the effect of differences in age composition on estimated participation rates. Table 21 summarizes the labor force participation rates for males, females, and persons aged 10 and over (15 and over for Taiwan and 15-64 for Peninsular -85- . Malaysia) in different deciles of households.- As would be expected, the inter-decile range of variation in male participation rates is smaller than in female rates.- The labor force participation rates for males aged 10 and over based on the PCE or PCY criteria, shown in Table 21 and graphically in Figure 8, do not exhibit any consistent relationship with decile. The rates in rural Gujarat, eleven towns of Nepal, and Malaysia show no clear pattern; those for the seven towns of Nepal and for Taiwan seem to vary inversely with decile; and the rates for urban Maharashtra and Sri Lanka and, to some degree, those for urban Gujarat, seem to rise with decile. In rural Maharashtra, the rates seem to rise at first, as decile increases,and then fall. 1/ Conventional wisdom suggests that the poor attempt to mitigate the burden of dependency by having their children participate in economic activity; but with the growth of population and labor force over the past two decades or so, the direct contribution of these child-workers to the income or earnings of the family tends to be quite limited, at least in the densely populated countries. Perhaps that factor and the rise in school enrollment ratios, rather than the biases of enumerators and interviewers, explain the very low labor force participation rates reported in our surveys for ages 5-9 and 10-14 and shown below. Of course, the children might contribute to family earnings by taking care of siblings and permitting women to work. arat Maharashtra Nepal Sri Lanka Taiwan Sex/Age Rural Urban Rural Urban 11 Towns 7 Towns 1968 1974 Males 5-9 1.0 0.3 1.2 0.1 3.2 3.6 0.3 0.1 -- Males 10-14 19.4 4.5 20.1 5.0 16.0 25.7 6.3 12.3 1.6 Females 5-9 1.2 0.2 1.1 0.1 4.4 3.7 0.4 0.3 0.1 Females 10-14 20.3 1.5 22.7 2.8 14.3 35.0 3.5 20.1 3.9 Persons 5-9 1.1 0.2 1.1 0.1 3.8 3.7 0.4 0.2 -- Persons 10-14 19.9 3.1 21.4 4.0 15.2 30.1 4.9 16.1 2.7 2/ The per capita ranking indicates a smaller range of variation in female participation rates than does the total household ranking in eight sets; the reverse is true in six sets (four of them in Taiwan). -86- Table 21 Lät Set ee åttICIpto Batgen ef Fepilatq. Aged 18 and Over a Different ettle Aecordtet te Ateruttive "«kl.8 Critert., Bska Sona Pe~insula~. haette of _fMtkt Mtfamsat_ N;l Sr tka iam 96al,eid Boa.eolde Rux.ut tre.,.n Éural ?rbano 11 Towua 7 Twn <1912-73) ¯1972-73 <1971. 4> ~1974-) (19469-70) (1968) (1974) (1973) A. Eotsehod Pnked Aordi. to leis Per C.pita t.8'apen (PCE) or Z~s (PCT) 1. 77.6 63.6 72.2 63.1 76.9 84.3 39.7 57.6 90.2 90.6 87.8 8.8 90.3 83.4 2. 16.7 63.4 74.0 64.8 75.6 80.4 63.8 60.9 91.6 92.2 86.4 86.1 8.2 87.9 3. 77.1 67.4 74.3 64.4 75.8 79.0 64.2 64.7 90.7 90.2 44.1 84.3 86.9 87.3 4. 77.3 65.8 78.6 66.1 73.2 77.1 67.6 47.7 68.9 86.5 83.4 86.9 87.4 86.8 3. 77.1 63.2 76.9 65.6 74.4 76.8 67.3 48.4 86.4 83.1 84.1 81.2 85.9 87.7 6. 74.4 63.5 78.2 64.4 67.6 73.9 71.2 68.9 82.7 84.5 70.4 79.3 84.0 85.2 7. 73.6 69.7 77.1 66.9 72.9 76.3 71.2 71.5 81.6 83.8 79.2 78.6 84.1 82.3 8, 72.7 68.0 77.0 72.4 68.4 72.9 72.6 73.8 82.6 79.2 76.6 77.7 84.7 87.1 9. 73.4 67.2 79.2 73.5 69.6 68.4 73.6 75.3 79.3 80.2 74.7 78.1 86.0 88.3 la. 76.2 10.7 75.2 75.0 74.7 7. 71. 7.7 76.5 78.7 81.5 82.1 8858.3 All 75.6 66.2 76.2 67.1 72.9 76.3 48.1 68.1 83.2 85.2 82.1 82.1 86.7 86.7 5. Koseholds Rmked According to Mheir Total Ependitury (TU) or Znow~ (=E1) 1. 84.1 70.6 76.1 74.5 84.4 87.1 70.2 64.6 90.1 92.2 89..8 88 .4 86.2 2. - 85.1 77.6 81.0 77.3 86.2 83.3 73.2 69.6 94.0 93.8 88.9 87.1 89.6 86.3 3. 81.2 74.4 79.5 75.7 82.4 88.3 73.3 69.5 94.0 92.7 86.8 87.3 92.2 88.6 4. 79.4 73.2 80.4 71.8 82.0 85.3 69.0 68.7 91.0 92.0 87.2 88.3 90.2 90.2 3. 78.2 61.5 78.7 67.9 77.4 82.3 70.4 68.0 89.3 87.4 84.8 84.3 88.8 89.8 6. 76.0 67.2 76.9 63.4 77.2 77.3 87.8 68.3 87.0 88.3 84.4 83.3 87.3 87.7 7. 73.8 63.7 76.9 65.4 71.3 74.5 67.7 68.7 86.8 83.9 79.1 79.9 84.9 86.0 8. 71.6 61.4 74.7 62.0 69.8 73.7 67.6 69.7 81.0 8.0 78.4 78.3 84.7 86.2 9. 72.3 62.7 74.1 62.8 66.1 71.6 65.1 68.0 81.4 79.7 75.9 78.3 83.8 84.7 10. 69.7 42.3 71.4 63.8 62.8 65.3 43.8 66.2 72.6 75.3 73.2 76.4 84.0 84.7 All 75.6 66.2 76.2 67.1 72.9 76.3 68.1 68.1 83.2 85.2 82.1 82.1 86.7 86.7 4. Eousehold Ranked A-cording to ThUir Per Capita Ep.dttw The reported expenditure on marriages, funerals and festi- vals and government taxes has been excluded because only a few households had reported such expenditure. The inclusion of these items tended to distort the mean levels of expenditure. (b) The seven towns surveyed during 1974-75 have been kept dis- tinct from the eleven towns surveyed during 1973-74 because of significant inflation during the intervening period and the non-availability of any suitable price index numbers for an adjustment. -169- SRI LANKA Income 1. Income was defined as "receipts in cash or kind from work, property, transfers and other sources that contribute to the individual's or individual earner's spending power." 2. The households whose income was subject to considerable seasonal fluctuations were asked to report "average monthly income based on the total income over the year preceding the Survey." For others, the reference period was "last month." 3. Value was imputed for "home-grown" consumption, free goods and services and owner-occupied housing. The latter was valued at "prevailing market prices," adjusted for maintenance costs and taxes. 4. A distinct record was made of income from: (a) Wages, salaries and related receipts; (b) Profits from business enterprise or farm; (c) Rent, dividends, and interest; (d) Pension, remittances, and cash allowances; (e) Other periodic cash receipts; and (f) Non-monetary income. Expenditure 1. For food, drink and tobacco, data were .collected through daily visits on seven consecutive days (blown up to get monthly estimates). 2. For non-durable goods and services, data were collected for the calendar month immediately preceding the field survey. 3. For consumer durables, data were collected for the "preceding year" and then averaged to get monthly estimates. 4. No value was imputed for the free rice ration. However we have imputed the value of the free rice ration--two pounds per week per person--on the basis of the average price paid for rice purchased from the market. The average price was estimated separately for each district and the three sectors--urban, rural, and the estates. 5. "Whenever possible," a responsible member of the household was asked to maintain a daily record of consumption. Labor Force/Employment 1. For "regular" workers, the reference period was one month preceding the survey. 2. For persons in seasonal occupations, the "usual status" was recorded. -170- 3. Persons working temporarily or casually on a contractual basis were classified as employed if they had worked for at least 10 days during the month preceding the survey. 4. The unpaid family workers in household enterprises were classified as employed if they had worked for at least 10 days during the month or the season preceding the survey. 5, The unemployed were defined as persons aged 15 to 55, who were not classified as employed in terms of (a) to (d), and were seeking work (excluding persons "mainly engaged" in household work and students). ECONOMY OF TAIWAN Income 1. Income during the previous year was recorded with a detailed accounting of all receipts (both in cash and kind), consumption, changes in fixed assets such as land, house, or other construction and also in financial assets and liabilities. 2. Personal income was recorded for all wage or salary earners. (Wage and salaries paid in kind were valued at "market prices.") 3. Net farm or business income was recorded for the whole household. Expenditure 1. In collecting data on expenditure, the primary emphasis was on recording "annual expenditure." Expenditure during each month or period was asked for if the annual figure could not be obtained directly. 2. Consumption out of own produce, gifts or donations as well as relief or subsidiaed subsistence was to be valued according to "market prices." 3. The imputed rent for owner-occupied, allotted or assigned dwelling was estimated according to the rent paid for similar houses in the vicinity. (Such rent was. recorded under expenditure as well as income.) Labor Force 1. No explicit reference period was used for the questions on occupation; but a distinction is made between major and secondary occupations. 2. Data on the major and secondary occupations of the employed are available in terms of: (a) 12 industrial divisions; (b) 12 occupational divisions; (c) whether employed in (i) public corporations, (ii) other government sector, or (iii) the non-government sector; and (d) class of worker. -171- 3. Unemployed persons have been identified, but there is no information on the duration of their employment. NOTE: A small sample of 300 households was required to maintain detailed accounts of expenditure and income; but the potentially useful comparative analysis of the expenditure pattern based on accounts and that reported in interviews has not been attempted. PENINSULAR MALAYSIA 1. Income of all members of the family from all sources (identified distinctly) was recorded for (a) the r6ference month, and (b) the survey year.. (Information for the survey year was collected through a supplementary survey, conducted after the Household Expenditure Survey of 1973 was completed. 2. Value was imputed for payments received in kind or in the form of concessional or subsidized supply of goods/services. 3. A separate detailed record was made of property income and changes in assets and liabilities. Expenditure 1. Households were visited daily for one month to collect information on consumption expenditure by item, type of payment, quantity and value. (In some cases, the adult individuals in the households kept the daily records.) Rental value of owner-occupied dwelling was imputed. 2. Consumption of own-produced goods was valued at the prices "normally charged by the respondent" when he sold them. If the household did not sell the items, they were valued according to prices prevailing in the same locality. 3. Information was sought on savings, financial transactions and business expenses. 4. Expenditure on consumer durables was to be recorded for the "12 months prior to the reference month." Labor Force 1. Questions were asked to only persons aged 15-64. (There is some indication that income-earners outside the age group 15-64.were to be asked questions on labor force. But, in fact, few persons covered by the labor force survey were outside this age group.) 2. Information was collected for a fixed reference week in each month (one week prior to the survey week); in addition, data have been collected on the usual status. (The original schedule did not provide for the usual status questions; they must have been added later.) -172- 3. Questions permit identification of: (a) the employed and their (i) principal occupation; (ii) principal industry; (iii) principal 'status' or class of worker; (iv) hours of work; (v) reasons for less than 25 hours of work and the availability of such persons for additional work; (b) the unemployed and their (i) willingness to move; (ii) efforts made to find a job; (iii) duration of job search. -173- ANNEX 4 SAVINGS RATES FOR HOUSEHOLDS IN DIFFERENT DECILES BY PER CAPITA INCOME AND EXPENDITURE The three tables in this Annex show the savings rates for households in different deciles, ranked according to per capita income as well as per capita expenditure. They relate to Peninsular Malaysia, Sri Lanka, and Taiwan (1974 only). Table 1 PENINSULAR MALAYSIA: RATIO OF PER CAPITA EXPENDITURE TO PER CAPITA INCOME IN DIFFERENT DECILES OF HOUSEHOLDS RANKED ACCORDING TO (A) PER CAPITA INCOME AND (B) PER CAPITA EXPENDITURE, ALL AREAS, 1973 HOUSEHOLDS WITH NEGATIVE SAVINGS HOUSEHOLDS WITH POSITIVE SAVINGS ALL HOUSEHOLDS Percentage of Per Capita Per Capita Ratio of Percentage of Per Capita Per Capita Ratio of Number of Per Capita Per Capita Ratio of aL. households Expenditure Income Expenditure all households Expenditure Income Expenditure Sample Expenditure Income Expenditure Decile in this class (MS) (M$) to Income in this class (MS) (M$) to Income Households (M$) (M$) to Income (A) Per Capita Income 1 84.51 25.55 5.13 4.981 -15.49 10.61 12.74 0.833 762 23.14 6.36 3.638 2 61.11 29.25 19.57 1.495 38.89 16.79 19.87 0.845 720 24.44 19.69 1.241 3 .53.85 36.87 26.58 1.387 45.15 21.64 26.78 0.808 728 29.74 26.67 1.115 4 50.27 45.96 33.92 1.355 49.73 27.38 33.94 0.807 730 36.31 33.93 1.070 5 41.99 55.02 42.58 1.292 58.01 33.45 42.45 0.788 724 41.96 42.50 0.987 6 39.09 71.05 52.97 1.341 60.91 40.99 53.50 0.766 747 52.62 53.30 0.987 7 28.49 85.66 68.12 1.257 71.51 50.65 67.96 0.745 709 59.99 68.00 0.882 8 23.61 121.60 89.96 1.352 76.39 63.38 89.50 0.708 703 75.61 89.60 0.844 9 21.70 170.59 129.45 1.318 78.30 89.12 131.17 0.679 728 105.64 130.82 0.808 10 11.63 384.89 285.29 1.348 88.37 172.26 342.90 0.502 722 195.67 336.55 0.581 ALL 41.92 57.24 39.52 1.448 58.08 57.08 88.73 0.643 7,273 57.15 67.63 0.845 (B) Per Capita Expenditure 1 41.95 14.52 10.58 1.372 58.05 12.82 25.81 0.497 727 13.55 19.24 0.704 2 41.81 21.53 15.64 1.377 58.19 21.44 31.10 0.689 727 21.48 24.71 0.869 3 43.21 27.50 20.45 1.348 56.79 27.55 39.14 0.704 729 27.53 31.09 0.885 4 46.22 34.03 25.04 1.359 53.78 34.09 48.48 0.703 727 34.06 38.13 0.893 5 &4.44 41.62 28.18 1.477 55.56 41.31 58.05 0.712 729 41.45 44.82 0.925 6 41.99 49.79 36.31 1.371 58.01 50.19 74.02 0.678 724 50.02 58.31 0.858 7 43.41 61.02 42.04 1.451 56.59 61.59 86.90 0.709 728 61.33 66.79 0.918 8 42.78 76.75 53.44 1.436 57.22 77.23 116.99 0.660 727 77.01 88.45 0.871 9 35.67 104.87 68.75 1.525 64.33 104.87 195.60 0.536 726 104.87 146.69 0.715 10 37.72 237.53 157.76 1.506 62.28 220.93 342.11 0.646 729 227.86 265.11 0.859 ALL 41.92 57.24 39.52 1.448 58.08 57.08 88.73 0.643 7,273 57.15 - 67.63 0.845 Table 2 SRI LAkA: RATIO OF PER CAPITA EXPENDITURE TO PER CAPITA INCOME IN DIFFERENT DECILES OF HOUSEHOLDS IKED ACCORDING TO CA) PER CAPITA INCOME A0 (B) PER CAPITA EXPENDITURE, AL.. AREAS, 1969-70 E'CIT,V SA GS WMS r7 PMITAvs SA'IN29 ALU RMStci.S ?atio of ulatio Rural Guj_are t A. Households Ranked According to Per Capita Epcuditure (PCL) 1. up to 26.66 9.95 11.57 152.18 22.48 5.161 4.31 2. 26.66 - 32.36 9.99 11.24 193.65 . 29.55 6.60 5.72 3. 32.36 - 36.70 10.00 10.76 217.40 34.76 7.41 6.60 4. 36.70 - 41.47 9.98 10.66 242.58 3. 9g 8,25 7.42 5. 41.47 - 46.58 10.01 10.28 262.03 43.8 :95 8.29 6. 46.58 - 52.74 9.96 10.43 301.34 49.4S. 10.24 9.27 7. 52.74 - 59.95 10.0 9.53 310.52 56.28 I.64 10.45 8. 59.95 - 69.00 9.99 9.40 351.42 64.07 11.97 11.95 9. 69.00 - 86.97 10.00 8.58 380.52 76.15 12.97 14.21 10. above 86.97 10.07 7.56 518.81 118.73 17.81 21.78 All 100.00 100.00 293.29 50.37 l oueholds Ranked According to Total Fxpenditure of the MI)hoZl (TH) 1. up to 110.83 9.94 4.04 82.08 34.63 2.78 6.76 2. 110.83 -149.75- 10,01 6.82 130.44 32.86 4.45 6.99 3. 149.75 -181.09 10.00 8.25 166.85 3.7,3 5.69 7.43 4. 181.09 -213.74 9.98 8.81 197.35 3?.40 6.72 7.97 5. 213.74 -247.G4 9.98 9.78 230.46 >. 39, 7.84 8.62 6. 247.04 -286.69 10.03 10.83 266.,5 42.41 9.12 9.42 7. 286.69 -336.34 9.98 11.2,) 310.78 4".I5 10.57 10.39 . 336.34 -403.19 10.01 12.04 368.33 5 12.57 11.65 9. 403.19 -509.91 10.00 12.83 450.55 6 .30 15.36 13.43 10. above 509.91 10.07 15.30 724.80 81.96 24.90 17.34 All 100.00 100.00 293.29 52.37 Urban Cuýyrat C. House4olds Rq:k7 759./ 95.> 23.35 17.31 All 100.00 10.00 327.'1 1.1.33 TABLE 2 NEPAL: PER CAPITA AND TOTAL HOUSEIOLD EXPENDIT'RE (IN NEPALI RUPEES) BY ACE CROUP OF 3TIE HA!)D OF 0E.O:S 1973-74 AND_ (B) SEVEN TOWt?S,_197t-75 Age of Nu=ber Average Average Average Index of Index c.f Iousehold of Household 1ouschold Per Capita Household Per Capita Head H0urnholis Size xetnditure E r.iture E1epnditure Fxpenditure (A) Eleven 'o 3-74 All '4393 5.4 500.0 92.9 100.0 100.0 less than 20 62 3.5 334.6 95.8 66.9 103.1 20-24 228 3.4 418.7 122.9 83.8 132.3 25-29 486 4.0 406.0 101.2 81.2 108.9 30-34 595 4.8 451.0 94.5 90,2 101.7 35-39 646 5.4 487.5 90.5 97.5 97.4 40-44 599 5.9 528.5 90.1 105.7 97.0 45-49 474 5.9 513.9 86.7 102.8 93.3 o 50-54 409 6.0 591.2 98.8 118.2 106.4 I 55-59 292 6.2 580.5 93.6 116.1 100.8 60+ 602 6.3 545.2 86.5 109.0 93.1 (B) Sfvcn -ToLmr. 19?4-? All 2254 5.2 583.6 113.1 100.0 100.0 less than 20 76 2.8 397.2 143.1 68.1 126.6 20-24 203 3.3 475.3 142.0 81.5 125.6 25-29 268 4.0 483.9 121.2 82.9 107.2 30-34 321 4.9 608.1 125.1 104.2 110.7 35-39 309 5.4 544.6 101.8 93.3 90.0 40-44 283 6.1 644.4 106.1 110.4 93.8 45-49 246 6.1 651.6 106.2 111.6 94.0 50-54 199 5.9 617.5 104.9 105.8 92.8 55-59 110 6.0 . 676.3 313.6 115.9 100.5 60+ 239 5.9 651.4 110.7 111.6 97.9 Table 3 Peninsular Malaysia: Per Capita and Total Household Expenditure (in Malaysian Dollars) by Age Group of the Head of Household, 1973 Age of Number Average Average Average Index of Inder. of Average Average Index of Index of Household of Household Household Per Capita Household Per Capita Household Per Capita Household Per Capita Head Households Size Expenditure Bxentur Expenditure Expenditure Income Income Income Income All 7273 5.40 308.90 57.20 100.0 100.0 365.55 67.63 100.0 100.0 less than 20 81 1.90 162.64 85.60 52.6 149.6 191.88 100.92 52.5 149.2 20-24 352 2.83 216.14 76.37 70.0 133.5 274.27 96.93 75.0 143.3 25-29 713 4.08 266.40 65.29 86.2 114.1 323.70 79.31 88.6 117.3 30-34 870 5.28 328.45 62.21 106.3 108.8 382.93 72.46 104.8 107.1 35-39 934 6.26 301.56 48.17 97.6 84.2 348.74 55.76 95.4 82.4 40-44 896 6.38 323.96 50.78 104.9 88.8 441.32 69.22 120.7 102.4 45-49 797 6.55 376.93 57.55 122.0 100.6 440.80 66.98 120.6 99.0 50-54 782 5.85 348.05 59.50 112.7 104.0 402.30 68.75 110.1 101.7 55-59 589 5.51 304.69 55.30 98.6 96.7 327.50 59.48 89.6 87.9 60-64 650 5.24 317.91 60.37 192.9 60.7 354.63 67.66 97.0 100.0 654 609 4.32 248.08 57.43 80.3 57.4 282.63 65.45 77.3 96.8 Urban Areas All 2359 5.43 439.17 80.88 100.0 100.0 557.98 102.63 100.0 100.0 less than 20 28 1.64 193.21 117.81 44.0 145.7 217.64 132.47 39.0 129.1 20-24 128 2.49 258.33 103.75 58.8 128.3 352.21 141.33 63.1 137.7 25-29 234 3.70 353.36 95.50 80.5 118.1 435.08 117.43 78.0 114.4 30-34 311 4.93 441.89 89.63 100.6 110.8 525.53 106.61 94.2 103.9 35-39 290 6.04 422.70 69.98 96.2 86.5 506.87 83.90 90.8 81.7 40-44 286 6.37 463.75 72.80 105.6 90.0 786.57 123.38 141.0 120.2 45-49 276 6.61 537.17 81.27 122.3 100.5 631.88 95.09 113.2 92.7 50-54 235 6.16 535.71 86.97 122.0 107.5 686.49 111.41 123.0 108.6 55-59 189 5.89 442.06 75.05 100.7 92.8 509.08 86.37 91.2 84.2 60-64 204 6.00 473.43 78.90 107.8 97.6 562.82 93.80 100.9 91.4 65+ 178 4.86 381.60 78.52 86-9 97.1 456.02 93.84 81.7 91.4 Rural Areas All 4914 5.39 246.36 45.71 100.0 100.0 273.17 50.68 100.0 100.0 less than 20 53 2.04 146.49 71.81 59.5 157.1 178.26 87.48 65.3 172.6 20-24 224 3.02 192.03 63.59 78.0 139.1 229.74 76.01 84.1 150.0 25-29 479 4.26 223.91 52.56 90.9 115.0 269.28 63.14- 98.6 124.6 30-34 559 5.48 265.33 48.52 107.7 105.9 303.60 55.37 111.1 109.3 35-39 644 6.35 247.01 38.90 100.3 85.1 277.54 43.70 101.6 86.2 40-44 610 6.38 258.42 40.50 104.9 88.6 279.51 43.83 102.3 86.5 45-49 521 6.52 292.04 44.79 118.5 98.0 339.58 51.87 124.3 102.3 50-54 547 5.72 267.43 46.75 108.6 102.3 280.20 49.00 102.6 96.7 55-59 400 5.32 239.79 45.07 97.3 98.6 241.70 45.41 88.5 89.6 60-64 446 4.90 246.77 50.36 100.2 110.2 259.41 53.00 95.0 104.6 65+ 431 4.10 192.93 47.06 78.3 103.0 211.03 51.53 77.3 101.7 Table 4 Sri Lanka: Monthly Per Capita and Total Household Expenditure and Income (in Sri Lankarn Rupees) by Age Group of the Household Head, 1969-70 Age of Number of Average Average Average Index of Index of Average Average Index of Index of Household Sample Household Household Per Capita Household Per Capita Household Per Capita Household Per Capita Heads Households Size Expenditure Expenditure Expenditure Expenditure Income Income Income Income All 9664 5.9 325.2 55.4 100.0 100.0 324.2 55.3 100.0 100.0 Less than 20 7 5.0 413.1 82.5 127.0 148.9 480.7 96.0 148.3 173.8 20-24 175 3.3 211.0 63.2 64.9 114.0 191.3 57.3 59.0 103.7 25-59 635 4.1 239.9 58.1 73.7 104.9 217.3 52.6 67.0 95.3 30-34 949 5.1 284.5 56.3 87.5 101.5 271.4 53.7 83.7 97.1 35-39 1302 5.9 317.1 54.2 97.5 97.8 309.4 52.9 95.4 95.7 40-44 1219 6.7 342.8 51.4 105.4 92.8 338.9 50.8 104.5 92.0 45-49 1433 6.8 350.0 51.4 107.6 92.7 335.9 49.3 103.6 89.3 50-54 1011 6.5 357.8 55.3 110.0 99.7 360.2 55.7 111.1 100.7 55-59 986 6.1 357.4 58.8 109.9 106.0 368.9 60.7 113.8 109.8 c3+ 1947 5.5 325.1 59.3 100.0 107.0 344.5 62.8 106.2 113.7 Table 5 Taiwan: Monthly Per Capita and Total Household Expenditure and Income (in N.T. Dollars) by Age Group of the Household ead, 1968 and 1974 Age of Number of Average Average Average Index of Index of Average Average Index of Index of Household Sample Household Household Per Capita Household Per Capita Household Per Capita Household Per Capita Heads Households Size Expenditure Expenditure Expenditure Expenditure Income Income Income Income A. 1968 Survey All 2776 5.84 3000.7 513.7 100.0 100.0 3326.9 569.5 100.0 100.0 Less than 20 23 5.13 2054.1 400.4 68.5 77.9 2071.2 403.7 62.3 70.9 20-24 48 4.06 3264.9 803.7 108.8 156.5 3659.0 900.7 110.0 158.2 25-59 224 4.85 2577.2 531.6 85.9 103.5 2861.9 590.3 86.0 103.7 30-34 371 5.92 2985.2 504.1 99.5 98.1 3319.5 560.6 99.8 98.4 35-39 508 6.25 2819.3 451.4 94.0 87.9 3042.7 487.1 91.5 85.5 40-44 460 6.05 2962.0 489.6 98.7 95.3 3244.7 536.3 97.5 94.2 45-49 410 5.92 3250.2 548.8 108.3 106.8 3567.7 602.5 107.2 105.8 50-54 279 6.06 3469.2 572.0 115.6 111.4 3941.8 650.0 118.5 114.1 55-59 225 5.62 3087.7 549.6 102.9 107.0 3489.3 621.1 104.9 109.1 60+ 228 5.62 2856.4 508.4 95.2 99.0 3305.8 588.4 99.4 103.3 B. 1974 Survey All 5238 5.36 7177.3 1339.8 100.0 100.0 8459.4 1579.2 100.0 100.0 Less than 20 5 4.00 5084.2 1271.0 70.8 94.9 5425.8 1356.5 64.1 85.9 20-24 71 3.24 4928.3 1521.3 68.7 113.5 5662.7 1748.1 66.9 110.7 25-59 367 4.07 6073.3 1491.9 84.6 111.3 7210.5 1771.2 5.2 112.2 30-34 742 4.87 6508.5 1332.2 90.7 99.4 7730.7 1582.4 91.4 100.2 35-39 777 5.56 6746.0 1213.6 94.0 -90.6 7881.8 1418.0 93.2 89.8 40-44 896 5.74 7079.5 1233.4 98.6 92.1 8072.4 1406.3 95.4 89.1 45-49 821 5.61 7894.3 1407.7 11.0.0 105.1 9222.0 1644.5 109.0 104.1 50-54 598 5,23 8181.1 1565.5 114.0 116.8 9827.6 1880.6 116.2 119.1 55-59 426 5.58 7631.6 1366.6 106.3 102.0 9317.2 1668.4 116.1 105.6 60+ 535 5.83 7 W6.6 1266.6 102.9 94.5 8830.8 1514.3 104.4 95.9 -194- 7.0. 5.0 WCE ) 4-5 LL C 4.0 :> 3.5 - cc:LLJ U)3.0 :D 0 S cm CI M In .T In 10 RGE OF HOUSEHOLD HEÄD 20- 120.0 IN, 70.0 60.0 S 50.0- 40.0 ~ da C4 e.1 en1 M Ln. ~ I. 0 r4 0 In In 1 01 1 I 1 0 C4 2 Below 20 25-29 35-39 45-49 55-59 65+ 1968 ............... 1974 120 110 100 105 U, C) O 90 IZ 80 70 z so Below 20 25-29 35-39 45-49 55-59 65+ 1G0 0 00 0 70 Sio 20 -25-29 36-30 45-4 55-59 65+ AGE OF 14OUSEHOLD HEAD) -199- ANNEX 8 ESTIMATION OF ADULT EQUIVALENTS Table 1 below shows the two sets of weights or coefficients for different sex-age groups used to convert individuals into adult equiv- alents. The basis for the weights used by the Bureau of Statistics of Taiwan is not known. The Indian coefficients were calculated on the basis of the recommendations of a committee set up by the WHO and the FAO. They were used to standardize the estimates of caloric con- sumption from the data on quantities of different food items consumed by sample households during the month preceding the date of the survey, during July 1971-June 1972. The weight for males aged 13-15 is lower than that for ages 10-12 or 16-19 because of similar differences in the FAQ estimates of daily caloric requirements. The weights do not take account of the increased caloric needs of women during pregnancy and lactation. The estimation of equivalent adults in Maharashtra has required some arbitrary assumptions because the specific ages of children aged less than five years were not recorded on the tape. An average weight of 0.53 was estimated on the basis of the detailed age composition of the age group 0-4 reported by the 1961 Census of Maharashtra. This weight was applied to all children below five years of age in every household. To some extent, the usefulness of fine age-specific weights for estimating equivalent adults is limited by the errors in age reporting which are particularly.large in single-year age distributions. However, as shown below in Table 2, the correlation between the number of per- sons and the number of adult equivalents in households has been ex- ceedingly high - a little higher when the Indian weights were used than when the Taiwan weights were applied. -200- Table 1 Weights for Estimating Equivalent Adult Consumers India, 1971-72 Economy of Taiwan, 1974 Age Group Males Females Age Group Males Females 0 0.43 0.43 0-1 0.3 0.3 1-3 0.54 0.54 2-4 0.4 0.4 4-6 0.72 0.72 5-7 0.5 0.5 7-9 0.87 0.87 8-10 0.7 0.7 10-12 1.03 0.93 11-14 0.8 0.8 13-15 0.97 0.80 15-20 0.9 0.9 16-19 1.02 0.75 21 and 1.0 0.9 over 20-39 1.00 0.71 40-49 0.95 0.68 50-59 0.90 0.64 50-69 0.80 0.51 70+ 0.70 0.50 Sources: (1) For the Taiwan data, figures supplied by the Bureau of Statistics, Taipei, Taiwan. (2) The National Sample Survey, Draft Report No. 258/10, Calorie and Protein Content of Food Items Consumed Per Diem*Per Consumer Unit: All India, Rural, 26th Round, July 1971-June 1972, mimeo. -201- TABLE 2 Measures of Association between Household Size and the Number of Adult Equivalents in Households for Some Data Sets Household Size and Household Size and Adult Equivalents Adult Equivalents (Indian Weights) (Taiwanese Weights) Correlation Coefficient of Correlation Coefficient of Area Coefficient Determination Coefficient Determination GUJARAT, 1972-73 Rural 0.99 0.97 Urban 0.99 0.98 MAHARASHTRA, 1972-73 Rural 0.99 0.97 -- -- Urban 0.99 0.98 -- -- SRI LANKA, 1969-70 0.99 0.97 0.96 0.92 TAIWAN 1968 0.98 0.97 0.97 0.94 1974 0.98 0.96 0.96 0.91 NEPAL Eleven Towns 1973-74 0.99 0.98 -- -- Seven Towns 1974-75 0.99 0.98 -- -- PENINSULAR MALAYSIA, 1973 0.99 0.97 0.97 0.94 NOTE: All figures are independently rounded. -202- ANNEX 9 AGE COMPOSITION AND THE RELATIONSHIP BETWEEN PER CAPITA EXPENDITURE AND VITAL RATES: AN EXPLORATORY EXERCISE FOR GUJARAT, 1972-73 There is widespread interest in estimates of income-related differences in rates of natural increase because of their serious implications for the extent to which any benefits of redistribution of income would be neutralized by accelerated population growth. Most discussions of this subject focus attention on differentials in fertility. Differentials in mortality have been neglected as a subject for research in recent years and, therefore, it is seldom possible to obtain any direct information on differentials in the rates of natural increase. A priori, the acquired wisdom postulates an inverse relationship between economic status and fertility as well as mortality. Higher mortality among the poor would partly compensate for their higher fertility and therefore, the differentials in the rate of natural increase would be expected to be smaller than those in either fertility or mortality alone. However, since the birth rates are significantly higher than death rates, the absolute magnitude of differentials in the former must exceed those in death rates and one would expect to observe higher rates of natural increase among the low income groups than among the better off. Some evidence on the subject is available for India, where the National Sample Survey has tabulated, for some selected years, the data on births and deaths ("during the previous year") by the per capita expenditure of the households. These data have been used to estimate crude (and sometimes also the age-specific) birth and death rates. Tables 1 and 2 below show the available data for urban India during 1959-61 and for both rural and urban India during 1963-65. (The rates show an average situation during the year of survey and the previous year because the questions pertained to births and deaths during the year preceding the date of survey.) Surprisingly, the estimated urban birth and death rates for 1963-65 were higher than for 1959-61. The latter seem to involve a greater degree of underreporting of births and deaths than the more recent data. Yet, while the 16th Round data do essentially confirm the generally-postulated inverse relationship between per capita expenditure (a proxy for income and a good index of the level of living) and the birth and death rates, the 19th Round data, particularly those for urban India, fail to show any clear relationship. The rural data for 1963-65 suggest an inverse relationship between PCE and birth rates but an implausible direct relation between PCE and death rates. A possible explanation can be the reporting biases, particularly the recall lapse, which generally tends to be greater with respect to deaths than births (partly because the survivors of births are normally around). -203- Table 1 Birth and Death Rate per 1,000 Population By Household Expenditures, Urban India 1959-61 Per Capita Monthly Birth rate Death rate Rate of Household Expenditure Natural _(in rupees)__ _ _ _ _ _ _ _ _ __ _ _ __ _ _ _ All 33.00 7.82 25.18 0 - 7 41.20 8.08 33.12 8.- 10 43.21 9.10 34'11 11 - 15 39.45 .9.13 30.32 16 - 20 34.83 8.38 26.45 21 - 25 31.09 7.21 23.88 26 - 30 25.24 7.29 17.95 31 -- 50 22.04 5.74 16.30 51+ 14.47 4.05 10.)2 Sourct: India National Sample Survey, 1971. Tables with Notes on the Fertility and Mortality Rates in Urban Areas of India, Sixteenth Round, July 1960-August 1961, No.180, New Delh: Cabinet Secretariat, p. 99. Table 2 Birth, Death and Natural Increase Rates per 1,000 Population by Monthly Per Capita Expenditure, Rural and Urban India, 1963-65 Rural India Urban India Monthly Rate of Rate of Per Capita Birth Death Natural Sample Birth Death Natural Sample Expenditure Rate Rate Increase Households Rate Rate Increase Households All 37.32 14.75 22.57 39,469 36.77 10.00 26.77 11,987 0-11 44.28 10.03 34.25 3.396 37.35 12.76 24.59 456 11-15 41.45 10.58 30.87 6,006 44.78 7.29 37.49 1,018 15-21 38.01 13.38 24.63 10,443 35.66 11.47 24.19 2,199 J 21-28 36.01 16.06 19.95 8,980 39.06 7.87 31.19 2,356 28-43 33.16 17.88 15.28 7,463 41.88 6.63 35.25 2,975 43+ 32.30 21.81 10.49 3,181 26.90 10.24 16.66 2,983 Source: India, National Sample Survey, Report No. 186, Tables with Notes on Differential Fertility and Mortality Rates in Rural and Urban Areas of India; Nineteenth Round: July 1964-June 1965, Delhi: Manager of Publications, 1970, pp. 5-6. -205- However, an Important relevant factor is the differentials in the age composition of population in different PCE clases. Our analysis of the 27th Round data for Gujarat and Maharashtra States (based on the state samples) shows large differences between deciles, mainly in the proportion of children aged 0-14, which varies inversely with MPCE decile. The pro- portion of the elderly--those aged 60 or more--varies directly with NPCE decile; but since they account for only about five percent of the total population both in rural and urban areas, the inter-decile differences in the proportions of the young have a dominant role.l/ The differences are larger in the urban population than in rural areas. The proportion of population in the main working ages of 15-59 is much higher in the top deciles than in the lower deciles. Since these are also the ages of reproduction, a logical consequence would be inter-decile differences in birth rates even if the age-sepcific fertility rates were identical. Further, since the risks of death are associated with age, the differences in the age composition would produce differences in the crude death rates as well. The possible magnitude of likely differences can be estimated by weighting the age-specific birth and death rates for rural and urban Gujarat, reported by the sample registration scheme for 19702/ by the proportionate age composition of the population of different MPCE deciles. The results of such an exercise which may be termed "expected birth rates," (expected if there were no decile-specific differences in age-specific birth and death rates and only the age composition varied), are shown below in Table 3. It seems that the crude birth rates would tend to rise (rather than decline) with MPCE--both in rural and urban areas--because of differences in age composition alone (if there were no differences in age-specific fertility in different deciles). Under a similar assumption concerning mortality, there would be a modest inverse relationship between MPCE decile and the expected crude death rate in urban areas, but hardly any relationship in rural areas. Since the expected birth rates exceed, and show larger inter-decile differences than, expected death rates, the resulting rates of natural increase have a positive relationship with MPCE. If, therefore, one is to observe an inverse relationship between the rate of natural increase and MPCE, the decile-related differ- entials in age-specific mortality and fertility must be quite large; otherwise the differentials in age-composition according to NPCE are likely to neutralize the impact of differences in fertility and mortality associated with income or economic status. 1/ The age-distributions of population in different deciles (or quintiles of households in Gujarat and Maharashtra States have been shown in Pravin Visaria, Living Standards, Employment and Education in Western India, 1972-73, mimeographed, August 1980. 2/ When this exercise was undertaken, these rates were the latest readily available from the Pocket Book of Population Statistics (Gujarat, Ahmedabad: Directorate of Census Operations, 1973). -206- Table 3 Birth, Death and Natural Increase Rates (Per 1000 Population) Expected for Different M.P.C.E. Deciles Without Any Difference In Sex-Age-Specific Rates, Rural and Urban Gujarat, 1972-73 RURAL GUJARAT URBAN GUJARAT Rate of Rate of M.P.C.E. Natural Natural Decile Birth Rate Death Rate Increase Birth Rate. Death Rate Increase 1 37.5 17.8 19.7 31.1 13.8 17.3 2 40.0 16.7 23.3 32.0 13.4 18.6 3-4 41.6 17.4 24.2 34.7 13.7 21.0 5-6 43.7 16.8 26.9 36.3 13.4 22.9 7-8 42.5 17.0 25.5 39.1 12.9 26.2 9 47.1 15.1 32.0 40.3 12.6 27.7 10 44.1 17.2 26.9 39.4. 11.8 27.6 All 42.2 17.0 25.2 35.7 13.3 22.4 SRS Estimates 41.7 18.3 23.4 34.9 13.7 21.2 (197 0) RelatIve Level of Expected Rite RURAL GUJARAT URBAN GUJARAT Rate of Rate of M.P.C.E. Naturalu Packle Birth Rate Death Rate increase Birth Rate Death P.-;Le nacreas.e 1 89 105 78 87 104 77 2 95 98 92 90 101 83 3-4 99 102 96 97 103 94 5-6 104 99 107 102 101 .02 7-8 101 100 101 110 97 117 9 112 89 127 113 95 124 10 105 101 107 110 89 123 All 100 100 100 100 100 100 Notes: (a) For calculating expected crude birth rates, the age-specific fertility rates for five-year age groups estimated by the SRS (sample registration schime) for 1970 were first weighted by the age distribution reported by the 1971 Census in order to obtain raLes for ages 15-29 and 30-44. (b) The expected crude death rates -were first estimated separately for males and females in each decile on the basis of the S.S rates for 1970; the estimates for two sexes were then weighted according to the proportio,n of males and feimacs reported by Lbe 27th Round. -207- ANNEX 10 STUDENT-POPULATION RATIOS Tables in this Annex show the student-population ratios for ages 5-24 by sex and decile of per capita expenditure. As noted in Section IV of the paper, these data are a by-product of the information on the nature of activity pursued by individuals in respondent households, i.e., questions relating to labor force participation. The Household Expenditure Survey of 1973 in Peninsular Malaysia had asked questions on labor force participation only to individuals aged 15-64. The student-population ratios cannot, therefore, be estimated for Peninsular Malaysia for ages 5-14, the prime ages of school enrollment. For Taiwan in 1968 and 1974, student-population ratios have been estimated for deciles of per capita expenditure and income and of total household expenditure and income. -208- Table 1 STUDENT-POPULATION RATIOS BY SEX AND AGE GROUP AND DECILE OF PER CAPITA EXPENDITURE RURAL AND URBAN GUJARAT, 1972-73 Age/Sex Decile According to Per Capita Expenditure 1 2 3 4 5 6 7 8 9 10 All RURAL GUJARAT Males 5- 9 22.7 26.5 26.6 33.6 36.0 33.2 44.9 51.7 58.7 53.9 35.6 10-14 33.5 41.9 38.2 56.2 55.0 59.7 68.5 68.8 78.2 78.4 55.4 15-19 13.5 6.4 15.4 15.8 18.7 18.8 24.6 37.0 40.0 44.8 23.4 20-24 -- 0.5 4.8 2.3 3.5 4.1 4.7 6.6 3.9 13.2 4.4 Females 5- 9 8.6 10.2 19.2 20.3 24.1 23.1 37.2 32.9 40.9 50.8 23.5 10-14 13.6 10.2 20.1 26.9 32.3 34.8 39.1 38.0 37.4 48.1 28.1 15-19 0.7 6.2 0.8 3.9 3.6 5.7 4.2 13.8 8.4 17.8 6.4 20-24 0.9 -- 0.7 -- -- -- -- -- 1.9 2.0 0.6 Persons 5- 9 16.2 18.9 23.0 26.9 30.1 28.5 41.3 42.3 50.4 52.3 29.8 10-14 23.6 28.3 29.4 42.9 43.9 48.6 54.7 55.0 59.4 64.2 42.7 15-19 7.9 6.3 8.4 9.9 11.1 11.7 15.3 25.7 24.1 32.9 15.1 20-24 0.5 0.2 2.4 1.2 1.8 2.1 2.3 3.7 2.8 7.7 2.5 URBAN GUJARAT Males 5- 9 42.5 42.7 56.9 63.7 58.8 64.9 75.8 74.4 84.1 78.6 58.5 10-14 70.0 77.4 81.4 86.7 92.5 90.6 95.3 90.3 91.9 97.5 85.2 15-19 38.5 37.9 45.4 49.0 61.4 62.6 55.7 68.2 70.5 86.2 54.8 20-24 9.3 8.9 10.9 11.6 11.0 10.8 13.3 22.4 18.9 30.7 14.7 Females 5- 9 31.2 48.8 51.4 55.2 60.7 58.9 66.4 70.9 76.6 69.5 54.2 10-14 45.9 62.8 68.9 76.1 82.4 83.1 84.3 90.3 96.2 98.4 73.5 15-19 23.8 27.9 25.3 33.5 33.6 42.1 40.1 51.1 50.9 58.9 36.9 20-24 2.7 -- 3.5 0.9 2.0 10.9 4.1 15.1 16.5 11.7 6.7 Person 5- 9 37.2 45.7 54.3 59.6 59.6 61.6 71.2 72.8 80.4 73.2 56.4 10-14 58.4 70.1 75.7 81.9 87.8 87.2 90.4 90.3 . 93.8 97.9 79.8 15-19 31.5 33.1 34.2 41.7 48.8 53.5 48.3 59.8 65.8 74.0 46.3 20-24 6.0 4.7 7.5 6.7 6.3 10.9 9.4 18.8 17.7 23.1 10.9 -209- Table 2 STUDENT-POPULATION RATIOS BY SEX AND AGE GROUP AND DECILE OF PER CAPITA EXPENDITURE, RURAL AND URBAN MAHARASHTRA, 1972-73 Sex/Age DECILE ACCORDING TO PER CAPITA EXPENDITURE 1 2 3 4 5 6 7 8 9 10 All RURAL MAHARASHTRA Males 5- 9 24.4 29.8 31.9 38.4 41.0 36.8 44.7 45.4 53.3 54.6 37.8 10-14 48.7 56.5 56.2 62.8 66.1 59.8 61.6 64.8 72.3 82.8 61.9 15-19 18.6 17.7 30.4 20.1 26.5 26.7 29.8 37.9 27.8 43.3 27.6 20-24 1.8 4.1 0.7 3.2 3.0 5.6 8.4 8.2 10.9 9.2 5.8 Females 5- 9 16.6 17.4 19.9 20.0 25.4 36.0 31.5 33.2 43.2 52.9 26.6 10-14 21.1 23.9 21.3 31.6 36.2 30.0 29.9 38.5 46.0 54.0 31.5 15-19 3.9 1.5 2.1 2.8 -- 6.4 1.5 8.5 13.0 7.4 4.5 20-24 -- -- -- -- -- 0.4 -- 3.9 -- -- 0.4 Persons 5- 9 20.4 23.9 26.3 28.7 32.9 36.4 38.1 39.5 48.7 53.9 32.2 10-14 36.8 40.9 39.4 47.1 51.0 45.6 48.1 53.1 59.1 72.1 47.7 15-19 11.4 11.0 16.8 12.1 14.0 17.5 17.5 24.2 20.5 28.0 17.0 20-24 0.7 2.2 0.3 1.2 1.4 2.7 4.0 6.0 4.6 4.8 3.0 URBAN MAHARASHTRA Males 5- 9 40.9 52.8 63.8 60.2 66.7 73.8 80.4 81.8 86.3 86.3 62.7 10-14 77.1 81.5 90.1 89.5 90.7 94.6 94.4 94.6 96.0 92.8 88.1 15-19 46.6 44.6 59.0 61.8 66.4 71.3 70.9 71.4 73.1 77.6 62.7 20-24 12.2 15.5 17.1 12.8 23.3 17.5 29.0 25.0 29.1 35.1 21.8 Females 5- 9 41.9 49.7 55.3 63.3 63.7 73.7 75.6 83.2 83.4 88.8 59.9 10-14 50.7 67.6 75.5 79.9 85.3 87.5 93.5 93.2 96.4 92.9 76.9 15-19 25.5 25.4 23.7 34.7 42.7 41.4 -51.4 54.4 59.8 68.0 39.4 20-24 2.8 2.3 3.3 4.2 8.6 9.1 11.6 16.5 15.8 23.6 9.3 Persons 5- 9 41.4 51.2 59.7 61.7 65.1 73.8 78.2 82.4 85.0 87.3 61.3 10-14 64.5 75.4 82.9 84.9 88.2 91.5 94.0 93.9 96.2 92.8 82.9 15-19 36.5 35.4 41.9 50.1 55.1 57.9 62.2 63.0 67.4 72.9 51.9 20-24 7.4 9.2 10.3 9.0 16.3 13.6 21.7 21.6 23.3 30.0 16.1 -210- Table 3 STUDENT-POPULATION RATIOS BY SEX AND AGE GROUP AND DECILE OF PER CAPITA EXPENDITURE, ELEVEN AND SEVEN TOWNS OF NEPAL, 1973-74 AND 1974-75, RESPECTIVELY Sex/Age DECILE ACCORDING TO PER CAPITA EXPENDITURE 1 2 3 4 5 6 7 8 9 10 All ELEVEN TOWNS OF NEPAL, 1973-74 Males 5- 9 29.5 32.8 42.2 48.4 49.2 61.7 62.5 66.7 64.6 77.8 47.9 10-14 58.2 59.0 70.0 73.0 74.5 87.4 83.2 86.9 86.6 84.9 74.8 15-19 31.0 34.9 39.1 40.5 45.2 59.0 48.5 57.4 60.0 59.7 48.5 20-24 7.1 8.9 8.6 8.3 14.3 19.6 16.2 29.2 26.0 25.3 18.1 Females 5- 9 15.3 20.2 31.4 33.2 35.2 44.2 57.8 57.9 67.4 71.2 36.6 10-14 25.0 29.1 47.3 49.7 57.6 69.8 78.0 77.3 88.8 72.3 55.1 15-19 16.8 21.1 19.4 22.6 19.7 38.7 33.3 44.6 51.4 49.4 31.4 20-24 -- 1.0 1.7 4.8 4.6 8.0 11.2 14.3 13.0 11.9 7.2 Persons 5- 9 22.4 26.9 37.0 40.8 42.6 52.7 60.2 62.2 66.0 74.6 42.4 10-14 41.0 43.8 58.4 62.5 66.1 78.5 80.5 82.1 87.5 79.0 65.1 15-19 23.2 27.8 29.4 31.4 31.8 50.0 41.3 50.8 55.9 55.3 40.1 20-24 3.2 4.4 4.8 6.3 9.1 13.8 13.6 22.1 19.9 21.3 12.6 SEVEN TOWNS OF NEPAL, 1974-75 Males 5- 9 17.50 23.8 36.0 45.8 54.3 52.8 60.5 57.4 73.3 81.8 43.0 10-14 34.7 51.5 59.0 66.3 65.1 75.8 73.6 82.7 85.1 75.8 65.1 15-19 12.9 15.6 19.4 33.3 36.5 42.0 51.7 48.7 61.8 57.4 38.3 20-24 2.2 -- -- 4.4 8.8 6.0 8.2 13.7 17.4 24.7 10.0 Females 5- 9 7.0 21.9 25.7 29.8 39.4 37.4 41.8 62.5 72.0 50.0 32.4 10-14 14.7 26.0 25.3 31.5 45.1 48.6 53.1 55.9 75.0 79.3 39.9 15-19 5.1 3.2 10.9 8.8 17.9 12.9 18.0 23.9 37.9 29.0 16.1 20-24 -- -- -- -- 3.5 1.8 -- 5.8 2.3 16.1 2.4 Persons 5- 9 11.9 22.9 30.8 37.5 46.1 45.4 51.0 60.0 72.6 67.5 37.6 10-14 24.4 39.0 42.1 50.3 55.4 64.0 62.8 71.4 80.9 77.4 53.1 15-19 9.1 10.0 15.6 21.9 27.7 28.2 34.5 37.1 52.4 47.8 28.2 20-24 1.1 -- -- 1.9 6.1 4.1 4.6 10.4 11.5 22.2 6.5 -211- Table 4 STUDENT-POPULATION RATIOS BY SEX AND AGE GROUP AND DECILE OF PER CAPITA EXPENDITURE, SRI LANKA, 1969-70 Decie. of Per Capita Expenditure Sex/Age 1 2 3 4 5 6 7 8 9 10 All ALL ISIAND Males 5 - 9 70.3 77.4 78.7 77.3 79.4 79.5 80.2 80.5 83.5 89.8 78.0 10 - 14 75.2 82.8 78.4 85.0 86.5 89.0 91.4 85.5 86.3 82.3 83.2 15 - 19 31.1 34.3 38.7 42.6 43.6 39.0 49.3 46.6 56.9 59.7 42.9 20 - 24 5.8 5.4 4.1 3.6 6.4 6.8 4.9 7.8 9.6 18.7 7.3 Females 5 - 9 65.8 72.6 78.7 79.5 80.2 80.6 80.3 84.5 86.1 81.9 77.1 10 - 14 61.0 70.0 75.8 79.8 78.0 79.7 76.7 83.9 83.6 78.8 75.1 15 - 19 24.0 29.9 28.0 33.9 34.7 32.1 42.5 39.0 45.1 48.9 34.7 20 - 24 4.1 5.0 6.3 3.8 6.3 5.3 6.2 7.2 7.0 12.0 6.3 Persons 5 - 9 68.2 75.1 78.7 78.4 79.8 80.1 80.3 82.6 84.7 85.9 77.6 10 - 14 68.5 77.0 77.2 82.3 82.7 84.3 83.7 84.7 84.8 80.6 79.3 15 - 19 27.5 32.0 33.0 38.3 39.0 35.4 45.8 42.2 50.7 54.3 38.7 20 - 24 5.0 5.2 5.3 3.7 6.3 6.0 5.6 7.5 8.3 15.4 6.8 URBAN AREAS Males 5 - 9 77.1 83.3 84.2 82.2 87.0 83.1 84.1 94.9 89.4 96.1 84.3 10 - 14 81.1 88.3 89.3 90.0 90.5 91.0 84.6 91.6 75.9 75.2 86.1 15 - 19 37.1 42.8 44.7 55.6 51.7 59.8 55.2 70.6 70.1 55.4 53.2 20 - 24 4.5 8.4 8.6 8.8 10.9 9.5 12.0 23.3 21.6 26.1 13.1 Females 5 - 9 76.5 80.3 82.0 78.3 84.9 81.8 85.7 82.6 89.0 90.1 81.8 10 - 14 75.4 79.1 88.5 84.0 87.6 86.3 82.3 86.0 83.3 79.6 83.0 15 - 19 34.3 27.7 36.8 37.5 43.8 45.9 44.5 56.4 54.1 54.9 42.6 20 - 24 3.0 4.4 4.0 2.7 6.8 4.8 8.7 15.9 13.4 10.2 7.2 Persons 5 - 9 76.8 81.8 83.2 80.3 85.9 82.5 85.0 88.9 89.2 93.2 83.1 10 - 14 78.6 83.9 88.9 86.8 89.1 88.6 83.4 88.7 79.6 77.2 84.6 15 - 19 35.7 35.6 40.7 46.3 47.6 52.7 49.8 63.6 62.5 55.1 47.8 20 - 24 3.7 6.5 6.3 5.5 8.6 7.1 10.3 19.5 17.3 18.4 10.1 RURAL AREAS Males 5 - 9 72.3 81.1 81.0 76.2 80.7 78.0 75.2 82.4 80.6 88.5 78.5 10 - 14 75.5 85.6 79.0 85.0 87.4 89.2 91.5 85.8 86.9 90.1 84.3 15 - 19 32.7 36.6 37.4 43.2 48.9 42.4 47.4 45.7 50.5 60.9 43.4 20 - 24 5.4 5.9 5.3 4.2 5.1 5.3 4.9 5.6 6.8 15.5 6.4 Females 5 - 9 66.5 74.1 79.6 82.9 80.4 85.8 82.6 82.4 84.1 81.6 78.4 10 - 14 59.6 73.0 78.9 82.1 82.1 79.4 78.0 83.6 88.4 76.6 76.6 15 - 19 24.7 34.6 33.7 29.5 40.5 38.7 44.1 -38.8 45.8 45.2 36.7 20 - 24 4.2 7.8 4.6 8.2 5.3 7.4 8.0 7.2 7.0 11.0 7.1 Persons 5 - 9 69.6 77.7 80.3 79.6 80.6 81.8 78.8 82.4 82.3 85.5 78.4 10 - 14 67.8 79.8 78.9 83.5 85.1 84.6 84.6 84.7 87.7 83.1 80.6 15 - 19 28.7 35.6 35.4 36.6 44.4 40.6 45.6 41.8 48.0 53.0 39.9 20 - 24 4.8 6.9 4.9 6.2 5.2 6.4 6.5 6.5 6.9 13.2 6.8 ESTATE SECTOR Males 5 - 9 41.0 53.6 67.7 68.3 62.5 82.7 72.1 '79.4 76.0 77.3 62.6 10 - 14 55.1 69.7 70.5 73.1 74.7 83.3 71.1 76.9 84.8 58.3 69.8 15 - 19 14.6 14.9 20.3 17.6 20.0 25.0 18.4 22.0 41.4 20.5 20.0 20 - 24 4.5 - 1.9 1.6 3.0 1.8 1.8 7.3 9.1 2.3 3.2 Females 5 - 9 47.5 52.9 56.8 61.6 63.6 61.5 64.3 70.0 85.7 83.3 59.9 10 - 14 43.3 43.4 51.8 43.7 52.9 58.2 44.2 59.5 65.5 51.9 49.4 15 - 19 5.2 2.5 1.9 1.7 9.2 12.9 , 6.0 4.9 14.3 11.4 6.6 20 - 24 - - - - - 1.5 - - 3.0 - 0.5 Persons 5 - 9 44.3 53.3 62.8 64.7 62.9 72.1 68.2 74.3 81.7 80.0 61.3 10 - 14 49.4 58.9 61.8 59.1 64.7 69.9 56.7 66.2 75.8 54.9 59.9 15 - 19 9.7 8.1 11.9 9.2 14.8 18.2 10.5 13.4 25.4 15.7 12.8 20 - 24 2.5 - 0.9 0.8 1.7 1.7 0.9 3.2 5.4 0.9 1.8 -212- Table 5 STUDENT-POPULATION RATIOS BY SEX AND AGE GROUP AND DECILE OF PER CAPITA EXPENDITURE, TAIWAN, 1968 AND 1974 Sex/Age DECILE ACCORDING TO PER CAPITA EXPENDITURE 1 2 3 4 5 6 7 8 9 10 All TAIWAN 1968 Males 5- 9 70.6 64.1 68.2 71.1 69.2 72.0 64.8 65.9 74.1 68.4 68.6 10-14 74.0 79.6 83.1 85.5 89.8 91.8 88.0 89.3 93.2 98.1 85.6 15-19 15.3 14.0 21.4 33.3 38.2 52.0 51.4 58.2 65.6 83.8 43.1 20-24 6.9 - -- 3.8 7.9 18.6 22.7 20.0 26.8 42.9 17.1 Females 5- 9 67.4 66.3 65.4 69.0 77.0 78.9 70.3 67.1 73.9 81.1 70.2 10-14 64.3 71.2 69.4 73.7 71.2 84.0 83.6 94.1 90.6 89.1 76.6' 15-19 7.9 9.5 12.0 14.9 25.0 31.3 41.5 47.0 50.0 69.1 29.3 20-24 -- 1.9 -- 3.4 4.4 2.9 7.9 8.5 14.6 20.0 6.4 Persons 5- 9 69.0 65.1 66.8 70.0 73.0 75.0 67.3 66.5 74.0 73.4 69.4 10-14 69.2 75.6 75.8 79.3 81.5 87.7 86.0 91.6 92.0 94.0 81.1 15-19 11.3 11.6 16.5 23.0 31.4 41.9 46.5 52.4 57.6 76.8 35.9 20-24 2.9 1.2 -- 3.6 5.7 8.9 13.3 12.8 20.2 29.9 10.5 TAIWAN 1974 Males 5- 9 72.9 75.5 76.9 75.3 73.5 71.7 72.8 72.8 77.7 70.8 74.3 10-14 92.5 95.7 98.1 97.5 99.1 96.5 97.8 97.5 98.4 100.0 96.9 15-19 30.7 47.5 48.7 54.9 57.1 70.0 66.5 74.5 80.0 85.0 61.0 20-24 1.4 5.9 10.7 9.1 6.2 20.7 21.1 23.5 38.0 34.1 18.5 Females 5- 9 74.1 67.0 73.0 69.8 75.8 76.8 69.0 73.3 69.2 64.2 71.9 10-14 86.9 91.9 89.3 92.8 97.0 93.9 96.1 98.1 98.3 100.0 93.0 15-19 20.0 34.7 30.6 42.9 41.3 47.9 57.6 61.9 66.9 80.2 45.5 20-24 0.9 2.5 4.9 8.0 4.2 7.6 11.5 11.8 12.9 25.7 9.4 Persons 5- 9 73.6 71.6 75.0 72.7 74.5 74.0 71.2 73.1 73.7 67.8 73.1 10-14 89.7 93.7 93.8 95.4 98.1 95.3 97.1 97.7 98.4 100.0 95.0 15-19 25.1 40.4 39.1 48.5 48.9 59.3 62.2 69.0 74.21 82.7 53.3 20-24 1.1 3.9 7.1 8.3 4.9 12.2 15.0 16.3 23.3 28.8 12.8 -213- Table 6 STUDENT-POPULATION RATIOS BY SEX AND AGE GROUP AND DECILE OF PER CAPITA INCOME TAIWAN 1968 AND 1974 Sex/Age DECILE ACCORDING TO PER CAPITA INCOME 1 2 3 4 5 6 7 8 9 10 All Taiwan 1968 Males 5- 9 67.3 62.6 75.5 67.3 69.0 69.3 71.4 66.7 70.8 68.6 68.6 10-14 72.2 83.6 82.5 85.7 90.2 86.2 89.3 89.5 95.8 98.2 85.6 15-19 12.3 16.3 23.5 41.5 36.4 46.3 54.0 58.9 64.0 79.2 43.1 20-24 -- -- -- -- 14.3 16.7 25.0 32.4 22.0 37.2 17.1 Females 5- 9 68.8 67.3 62.8 73.7 75.7 73.7 68.5 80.3 67.9 73.2 70.2 10-14 59.1 71.2 73.4 74.8 76.9 81.6 87.3 87.8 90.3 91.5 76.6 15-19 4.5 5.2 14.4 16.0 32.7 32.4 37.9 40.2 56.5 69.1 29.3 20-24 -- 2.1 1.9 6.8 1.5 3.2 7.0 10.4 11.7 17.0 6.4 Persons 5- 9 68.1 64.8 68.8 70.6 71.8 71.3 70.0 73.3 67.4 70.7 69.4 10-14 65.8 77.2 77.9 80.5 83.0 84.1 88.4 88.6 93.2 95.1 81.1 15-19 7.8 10.6 18.9 28.0 34.5 39.1 45.1 50.3 60.2 74.3 35.9 20-24 -- 1.3 1.3 4.8 6.5 9.0 13.1 18.3 15.8 26.0 10.5 Taiwan 1974 Males 5- 9 73.4 76.6 73.7 75.9 72.4 73.9 71.1 76.9 72.1 77.8 74.3 10-14 92.2 96.7 98.6 96.5 97.8 98.1 96.5 98.0 99.2 100.0 96.9 15-19 35.9 45.7 53.1 50.3 62.3 64.9 68.8 72.3 78.8 82.8 61.0 20-24 4.3 6.1 11.3 11.1 14.3 14.4 21.7 25.5 28.2 35.9 18.5 Females 5- 9 73.4 69.9 74.0 66.2 77.9 74.3 68.4 72.0 74.0 64.0 71.9 10-14 88.3 91.6 90.1 93.1 93.3 95.2 95.7 97.4 99.1 100.0 93.0 15-19 24.6 34.9 33.5 38.3 47.9 47.0 50.3 62.2 63.9 82.6 45.5 20-24 1.0 3.7 3.4 7.5 9.9 6.0 10.3 11.8 14.9 20.7 9.4 Persons 5- 9 73.4 73.3 73.8 71.5 74.6 74.1 69.9 74.5 73.0 72.1 73.1 10-14 90.2 94.0 94.8 94.8 95.8 96.8 96.2 97.7 99.2 100.0 95.0 15-19 29.6 40.2 42.1 43.9 55.1 56.6 60.2 67.7 72.0 82.7 53.3 20-24 2.3 4.6 6.4 8.8 11.3 9.2 14.8 16.6 21.0 26.0 12.8 -214- Table 7 STUDENT-POPULATION RATIOS BY SEX AND AGE GROUP AND DECILE OF TOTAL HOUSEHOLD EXPENDITURE TAIWAN, 1968 AND 1974 Sex/Age DECILE ACCORDING TO TOTAL HOUSEHOLD EXPENDITURE 1 2 3 4 5 6 7 8 9 10 All TAIWAN 1968 Males 5- 9 65.3 64.4 65.2 68.6 70.5 73.6 63.9 72.6 67.6 70.6 68.6 10-14 83.7 80.5 80.8 81.3 80.6 81.3 84.7 91.4 89.5 94.1 85.6 15-19 19.4 13.0 16.4 27.1 30.1 36.6 41.0 46.9 54.9 73.4 43.1 20-24 8.3 5.3 -- -- -- 6.5 9.3 18.6 19.6 38.0 17.1 Females 5- 9 65.9 64.0 67.2 66.9 62.3 73.5 74.4 70.7 77.5 73.9 70.2 10-14 72.4 62.7 61.9 74.5 77.0 70.5 73.2 79.5 86.0 89.2 76.6 15-19 16.7 8.8 10.3 10.0 10.1 22.7 32.3 34.4 38.7 55.3 29.3 20-24 -- -- 2.1 -- 6.0 2.0 8.5 1.4 11.1 15.8 6.4 Persons 5- 9 65.6 64.2 66.2 67.8 66.4 73.6 68.9 71.7 72.5 72.2 69.4 10-14 79.5 73.0 71.3 77.7 78.8 76.3 78.8 85.3 87.8 91.6 81.1 15-19 17.9 10.5 12.9 18.6 20.4 29.4 36.7 41.0 46.7 63.7 35.9 20-24 2.9 1.5 1.4 -- 4.3 3.7 8.9 7.9 14.2 25.7 10.5 TAIWAN 1974 Males 5- 9 70.5 70.9 75.0 75.3 73.6 76.0 73.0 75.7 78.0 72.3 74.3 10-14 90.8 93.8 95.1 98.0 97.1 96.7 98.5 97.6 97.7 98.1 96.9 15-19 26.1 45.0 53.4 51.2 48.0 55.6 68.4 65.3 67.8 75.3 61.0 20-24 -- -- 6.7 8.5 18.2 13.6 17.1 18.8 26.6 29.6 18.5 Females 5- 9 69.0 66.7 70.3 67.5 76.5 73.9 77.9 72.0 73.4 68.6 71.9 10-14 86.8 85.2 89.3 93.0 90.2 93.5- 93.6 95.4 95.8 98.6 93.0 15-19 24.4 25.8 27.2 32.3 38.7 44.6 38.8 52.7 57.9 62.5 45.5 20-24 -- -- 1.1 7.4 5.6 8.0 5.4 11.9 14.0 17.4 9.4 Persons 5- 9 69.8 68.8 72.8 71.4 75.1 75.0 75.4 74.1 75.7 70.6 73.1 10-14 88.8 89.9 92.3 95.5 93.9 95.2 96.3 96.6 96.7 98.3 95.0 15-19 25.3 35.5 40.4 41.6 43.5 49.6 53.3 58.8 53.2 69.0 53.3 20-24 -- -- 3.2 7.8 9.8 10.1 9.8 15.0 18.8 22.0 12.8 -215- Table 8 STUDENT-POPULATION RATIOS BY SEX AND AGE GROUP AND DECILE OF TOTAL HOUSEHOLD INCOME TAIWAN 1968 AND 1974 Sex/Age DECILE ACCORDING TO TOTAL HOUSEHOLD INCOME 1 2 3 4 5 6 7 8 9 10 All Taiwan 1968 Males 5- 9 65.4 60.0 66.4 70.2 67.7 72.0 68.1 76.8 67.5 69.7 68.6 10-14 70.4 88.3 72.2 88.3 81.3 82.9 86.6 90.7 90.7 93.2 85.6 15-19 10.3 11.4 18.1 25.0 37.5 38.3 42.4 47.5 58.1 68.6 43.1 20-24 -- -- -- -- 4.5 5.4 11.4 22.7 15.6 37.6 17.1 Females 5- 9 68.5 70.5 61.1 68.7 66.1 65.6 76.5 74.1 75.5 73.6 70.2 10-14 62.5 59.0 68.5 73.0 73.0 73.4 80.3 76.5 86.6 88.3 76.6 15-19 6.7 2.6 10.7 28.2 11.4 18.7 35.3 36.4 41.8 55.0 29.3 20-24 -- -- 2.8 2.2 5.4 4.2 2.2 4.7 14.6 11.8 6.4 Persons 5- 9 67.0 64.8 63.5 69.5 67.0 68.8 72.1 75.5 71.8 71.4 69.4 10-14 67.4 75.4 70.4 79.9 77.4 78.4 83.4 83.3 88.8 90.7 81.2 15-19 8.5 5.7 14.1 21.4 24.4 28.6 38.9 42.5 49.5 61.3 35.9 20-24 -- -- 1.8 1.7 5.1 4.7 6.7 10.9 15.0 23.6 10.5 Taiwan 1974 Males 5- 9 71.8 68.7 81.3 71.1 72.5 75.1 72.0 78.8 76.0 73.8 74.3 10-14 90.2 93.2 96.4 98.1 98.6 95.4 97.6 98.3 97.1 99.5 96.9 15-19 28.6 55.4 51.3 53.3 51.3 60.0 64.3 66.1 63.2 76.4 61.0 20-24 -- 2.6 9.6 9.8 18.2 11.8 19.1 23.0 17.4 31.7 18.5 Females 5- 9 71.3 69.8 67.6 72.7 71.3 75.0 76.4 71.3 72.5 69.5 71.9 10-14 88.2 85.6 92.1 92.6 93.4 87.6 96.5 96.0 93.9 99.0 93.0 15-19 22.0 31.3 32.7 39.2 37.3 36.8 46.3 55.9 48.9 64.6 45.5 20-24 -- -- 1.1 5.7 9.1 11.1 8.3 11.0 10.0 17.2 9.4 Persons - 5- 9 71.5 69.3 75.1 72.0 71.9 75.1 74.2 75.5 74.3 71.8 73.1 10-14 89.2 89.7 94.4 95.6 96.0 91.8 97.1 97.2 95.5 99.2 95.0 15-19 25.2 42.9 42.2 46.4 44.1 48.4 54.9 61.2 56.3 70.4 53.3 20-24 -- 0.8 4.3 7.4 12.3 11.3 12.6 15.7 13.1 22.5 12.8 -216- ANNEX 11 Crude Labor Force Participation Rates Table 1 in this Annex shows the crude labor force participation rates (by sex) for different deciles of households according to alternative ranking criteria. As discussed in Section IV of the paper, the pattern of decile-related differences in crude participation rates is significantly different from that observed in participation rates of population aged 10 and over (or 15 and over), when the per capita expenditure or income criterion is used to rank households. This is particularly true of the participation rates of males. The data for PCE deciles are also shown in graphical form in Figure 1. Table 1 CRUDE LABOR FORCE PARTICIPATION RATES FOR DIFFERENT DECILES WITH ALTERNATIVE RANKING CRITERIA (MMLS) Peninsular Decile of Cularat2/ Maharashtra .3 Nepal Sri Lanka Taiwan Malaysia Households Rural Urban Rura Urban 11 Towns 7 Towns (1972-73) (1972-73) (1973-74) (1974-75) (1969-40) (1968) (1974) (1973) -A- B A B A B A. Households Ranked According to Per Capita Expenditure or Income 1. 50.0 41.8 55.7 49.2 47.6 53.8 39.8 36,9 45.5 46.1 44.6 43.6 36.4 34.7 2. 51.0 45.5 57.1 53.0 49.9 53.7 43.1 41.9 47.3 47.0 45.1 44.3 37.4 36.9 3. 50.4 47.0 57.8 52.9 51.3 53.6 47.6 44.3 49.1 47.2 46.5 44.7 37.9 36.6 4. 53.9 49.1 63.4 55.3 51.4 53.9 48.9 49.6 49.7 47.6 45.5 47.2 39.4 38.7 5. 55.2 45.8 63.5 56.5 53.6 58.7 49.9 50.0 49.0 49.1 49.3 47.0 39.3 39.0 6. 53.0 49.8 65.0 55.2 51.4 54.8 54.6 53.1 48.6 50.7 48.4 49.4 40.2 42.2 7. 53.6 53.8 63.6 59.6 55.8 56.2 54.7 55.9 50.0 50.7 50.4 51.1 44.3 42,9 8. 54.7 55.2 65.2 64.6 54.2 58.9 56.8 60.0 52.0 52.0 51.7 54.1 48.2 49.4 9. 56.5 56.2 67.2 68.0 57.1 56.7 59.3 61.2 55.1 55.7 53,5 54.9 54.6 56.4 10. 60.6 62.9 65.4 70.8 65.2 70.4 59.7 62.3 57.5 58.4 64.1 63.9 64.1 63.6 All 53.6 49.6 62.1 57.3 53.2 56.4 50.6 50.6 49.8 49.8 49.1 49.1 43.0 43.0 B. Households Ranked According to Total Expenditure or Income 1. 65.6 60.8 66.7 68.9 66.8 68.7 54.8 50.8 59.3 60.2 56.6 53.8 43.6 40.9 2. 59.4 57.5 67.4 67.7 60.6 62.7 52.6 49.2 52.2 50.1 49.3 47.1 44.5 40.0 3. 55.0 53.6 63,0 64,0 58.0 65.8 51.1 48.9 51.8 55.4 48.0 46.3 41.2 39.8 4. 53.8 50.8 65.7 59.8 58.5 59.3 50.0 47.4 53.0 49.3 46.9 47.8 42.1 42.4 5. 55.1 44.6 64.3 57.0 53.0 61.4 Soj 49.0 48.4 45.5 47.8 46.5 42.3 42.2 6. 52.5 50.2 62.2 54.4 53.6 56.0 49.5 48.7 47.9 51.6 45.6 46.3 41.2 39.9 7. 50.8 47.6 61.6 54.2 52.0 53.2 49.4 50.2 50.5 48.4 46.0 48.3 41.6 42.3 8. 51.6 46.2 61.6 52.8 52.1 53.8 51.4 53.6 48.3 48.1 49.7 49.0 40.6 42.8 9. 52.6 47.6 59.8 54.3 49.1 52.7 48.6 52.4 47.0 48.0 51.0 52.7 43.8 45.7 10. 50.0 49.9 58.3 56.0 48.1 50.5 50.8 53.5 47.0 48.3 51.8 52.9 48.0 49.5 All 53.6 49.6 62.1 57,3 53.2 56.5 50.6 50.6 49.8 49.8 49.1 49.1 43.0 43.0 Notes: 1. When two columns are shomm for an area Column A refers to deciles demarcated on the basis of expenditure data aAd Column B to those on the basis of incom data. For other regions, only expenditure deciles have been demarcated. 2. Data based an usual activity. 3. Relate to ageS five and over. Table 1 (Continued) CRUDE LABOR FORCE PARTICIPTION RATES FOR DIFFERENT DECILES WITH ALTERNATIVE RANKING CRITERIA (FEMALES) D l o23/_ Peninsular Decile of Gujarat V Maharashtra Nepal Sri Lanka Taiwan Malaysia Households Rural Urban Rural Urban 11 Towns 7 Towns (1972-73) (1972-73) (1973-74) (1974-75) (1969-70) (1968) (1974) (1973) A B A B A B A B A. Households Ranked According to Per Capita Expenditure or Income 1. 39.8 16.6 45.6 21.8 27.2 45.0 16.3 13.9 33.0 34.9 33.3 32.2 21.2 22.2 2. 40.3 12.8 50.6 18.6 30.5 42.6 15.4 15.7 36.2 36.3 32.7 29.3 22.3 20.9 3. 37.0 11.2 53.2 16.6 28.7 44.3 17.5 16.8 34.7 31.5 31.2 29.7 23.1 22.6 4. 39.4 12.8 50.1 12.3 27.6 39.0 17.7 19.4 31.2 30.9 26.8 25.9 23.4 22.0 5. 34.1 8.1 52.4 11.4 27.7 37.8 16.1 15.9 32.7 28.4 26.7 29.2 21.6 23.1 6. 38.0 8.4 50.4 10.5 22.6 39.2 17.8 17.9 28.7 29.2 29.1 29.5 25.0 23.1 7. 36.0 8.4 53.2 8.3 22.6 40.0 16.8 17.9 26.9 27.6 28.3 28.6 23.7 24.2 1 8. 34.1 9.5 50.5 10.1 20.5 41.2 17.8 18.1 24.1 29.9 27.9 30.7 27.4 26.1 - 9. 35.6 8.7 47.1 12.8 19.1 24.8 17.5 16.4 25.6 25.2 29.5 30.7 28.6 30.4 c 10. 31.9 10.5 44.6 16.8 22.6 25.0 21.8 23.9 25.7 25.6 33.3 35.9 35.1 38.3 All 36.9 11.0 49.9 14.4 25.4 39.5 17.3 17.3 30.6 30.6 30.0 30.0 24.5 24.5 B. Households Ranked According to Total Expenditure or Income 1. 55.7 33.8 68.1 42.9 45.9 54.3 23.5 21.4 36.2 38.3 34.2 33.6 31.9 27.9 2. 46.9 19.6 58.4 32.0 36.7 48.4 22.2 20.6 33.2 33.3 30.5 27.4 26.6 26.5 3. 43.2 15.5 56.8 24.1 30.9 50.4 19.2 19.6 34.1 34.4 29.7 28.4 24.1 24.3 4. 39.3 12.0 54.4 18.6 28.9 45.6 18.3 16.4 29.7 30.3 26.7 23.7 24.4 23.2 5. 38.9 12.6 52.3 15.3 31.4 37.2 17.5 16.7 31.1 29.6 25.7 25.6 22.6 21.2 6. 36.4 7.7 52.1 12.5 30.8 39.2 14.8 16.4 32.9 29.9 28.1 30.5 21.2 21.7 7. 34.5 6.8 50.1 8.1 24.4 41.8 15.7 15.5 28.4 27.9 29.3 30.2 23.8 21.4 8. 35.6 8.7 45.0 9.0 23.0 39.6 15.1 15.2 31.1 32.9 31.2 29.2 22.2 23.8 9. 31.5 8.3 44.4 8.8 17.6 39.0 15.2 15.1 30.3 29.8 30.7 33.4 23.5 25.0 10. 28.9 7.7 39.6 10.4 14.5 24.8 17.1 19.4 25.0 26.1 34.7 36.6 27.5 29.7 All 36.9 11.0 49.9 14.4 25.4 39.5 17.3 17.3 30.6 30.6 30.0 30.0 24.5 24.5 Notes: 1. When two columns are shown for an area Column A refers to deciles demarcated on the basis of expenditure data and Column B to those on the basis of income data. For other regions, only expenditure deciles have been demarcated. 2. Data based on usual activity. 3. Relate to ages five and over. Table 1 (Continued) CRUDE LABOR FORCE PARTICIPATION RATES FOR DIFFERENT DECILES WITH ALTERNATIVE RANKING CRITERIA (92SNS) Peninsular Decile of Gularat - Maharashtra 2/3 Nepal Sri Lanka Taiwan Malaysia Households Rural Urban Rural Urban 11 Towns 7 Towns (1972-73) (1972-73) (1973-74) (1974-75) (1969-70) (1968) (1974) (1973) A B A B A B A B A. Households Ranked According to Per Capita Expenditure or Income 1. 45.0 29.2 50.6 35.4 37.1 49.2 27.9 25.1 39.2 40.3 38.8 37.7 28.3 28.1 2. 45.8 29.3 53.9 36.0 40.1 48.2 29.6 29.0 41.7 41.6 38.9 36.8 29.6 28.6 3. 43.8 29.3 55.5 35.0 39.9 48.8 32.4 30.6 41.7 39.2 38.9 37.2 30.4 29.5 4. 46.9 31.3 56.4 35.0 39.4 46.4 33.5 34.7 40.0 39.1 36.1 36.5 31.3 30.1 5. 44.6 27.5 57.8 35.1 41.0 47.9 33.7 32.9 40.9 38.7 38.1 38.4 30.3 30.9 6. 45.7 29.5 57.5 35.1 36.9 47.1 36.5 36.0 39.7 40.3 39.0 39.7 32.4 32.5 7. 45.1 32.3 58.6 36.7 39.4 48.2 35.9 37.2 38.7 39.0 39.7 40.2 33.7 33.4 8. 44.8 33.0 58.0 40.8 37.6 50.6 37.1 39.3 38.4 41.2 40.3 42.5 38.0 38.0 a 9. 46.1 33.4 57.3 44.8 39.2 43.0 38.9 39.5 41.0 41.2 41.9 43.5 42.0 43.6 10. 46.7 38.7 56.2 48.7 47.5 54.5 41.2 43.9 42.9 43.5 49.4 50.3 49.8 51.3 All 45.4 30.8 56.1 37.3 39.5 48.2 34.1 34.1 40.3 40.3 39.6 39.6 33.6 33.6 B. Households Ranked According to Total Expenditure or Income 1. 60.7 47.3 67.5 58.6 55.9 61.4 38.5 35.3 49.0 50.3 46.2 44.1 37.3 33.7 2. 53.3 38.9 62.8 52.3 48.5 55.7 37.4 34.5 43.0 41.8 40.1 37.5 35.0 33.0 3. 49.2 34.6 59.9 45.5 44.6 58.0 35.8 34.6 43.0 44.7 39.0 37.6 32.6 31.9 4. 46.8 32.2 60.1 40.0 44.0 52.8 34.3 32.0 41.1 39.8 36.7 35.7 33.2 32.7 5. 47.1 29.2 58.1 37.3 42.4 49.6 33.6 32.8 39.7 37.8 36.8 36.0 32.3 31.6 6. 44.6 28.5 57.2 34.8 42.3 47.6 32.5 32.8 40.5 40.6 36.8 38.5 31.3 31.1 7. 43.0 28.5 56.0 32.2 38.7 47.7 32.8 33.1 39.5 38.3 37.8 39.3 32.8 31.7 8. 43.7 28.3 53.4 32.4 37.7 46.9 33.6 34.5 39.8 40.4 40.6 39.3 31.1 33.2 9. 42.0 28.2 52.4 32.9 33.4 46.1 32.2 34.4 38.6 38.7 40.8 43.1 33.7 35.4 10. 40.3 29.3 49.4 34.6 31.9 38.1 34.5 37.2 35.8 37.2 43.3 44.7 37.5 39.4 All 45.4 30.8 56.1 37.3 39.5 48.2 34.1 34.1 40.3 40.3 39.6 39.6 33.6 33.6 Notes: 1. When two columns are shown for an area Column A refers to deciles demarcated on the basis of expenditure data and Column B to those on the basis of income data. For other regions, only expenditure deciles have been demarcated. 2. Data based on usual activity. 3. Relate to ages five and over. -220- Figure la CRUDE LABOR FORCE PARTICIPATION RATES RURAL GUJARAT MALES FEMALES PERSONS 10 60 . W ...... 40-e.~;:;d.. 30- 20 z O Uj 0 10 1 06 8 1 0. MALES FEMALES PERSONS O 70 40_ 30 2 4.10..80. .1 20 10 rURBANIIIIARAT DECILE OF HOUSEHOLDS RANKED BY PER CAPITA AND TOTAL EXPENDITURE PCE DECILE .***.************* THE CECILE -221- Figure lb CRUDE LABOR FORCE PARTICIPATION RATES RURAL MAH4ARASHTRA MALES FEMALES PERSONS 70 60 ........ 50 40 -. 30- 20 S10 z 0 2 4 6 8 10 0 2 4 8 8 10 0 2 4 6 8 10 0 70 20 DEI. MALosmoosRESc>s PEAE P^iAADTT:LEPNIUERSN -222- Figure le NEPAL ELEVEN TOWNS, 1973-74 MALES FEMALES PERSONS 00 90 80 70- 60 0 2 4 6 8 10 2 4 6 8 10 2 4 6 8 10 NEPAL SEVEN TOWNS, 1974-75 MALES FEMALES PERSONS go - *. ** S70- 40. 20. 10. 0 2 4 6 8 10 0 2 4 8 8 10 0 2 4 6 8 10 'DELE OF HOUSEHOLDS RANKED BY PER CAPITA AND TOTAL EXPENDITURE CRUDE LABOR FORCE PARTICIPATION RATE (PERCENT1 S 888 8 i 3 t 8 8 I4 I . y e e tit m ez zz M > e)e m> *-I-- *. z o - - - I II i i I c .· ~s m mm m -J, o * ~*.fl >* bu -224- Figure IS CRVDE LABOR FORCE PARTICIPATION RATES TAIWAN 1968 MALES FEMALES BOTH SEXES 60-. 40 - ***** 50 - 8 20* m 10 21 1I I I I I I Ia I - 2 4 6 8 10 0 2 4 6 10 0 2 4 6 8 10 TAIWAN 1974 0 MALES FEMALES BOTH SEXES 70 5 50 .o 40 a.. 30 20 10 I I I ___II I I_ ___I I_ __I 0 2 4 6 8 10 0 2 4 6 8 10 2 4 6 8 10 DECILE OF HOUSEHOLDS RANKED BY PER CAPITA AND TOTAL EXPENDITURE E The WOrld Bank Publications Order Form SEND TO: YOUR LOCAL DISTRIBUTOR OR TO WORLD BANK PUBLICATIONS (See the other side of this form.) P.O. BOX 37525 WASHINGTON, D.C. 20013 U.S.A. Date Name Ship to: (Enter if different from purchaser) Title Name Firm Title Address Firm City State Postal Code_ Address Country _ _'lephone ( ) City State_ Postal Code___ Purchaser Reference No. Country Ilephone ( Check your method of payment. Enclosed is my O Check Ol International Money Order O Unesco Coupons O International Postal Coupon. Make payable to World Bank Publications for U. S. dollars unless you are ordering from your local distributor. Charge my E1 VISA [ MasterCard O American Express l Choice. (Credit cards accepted only for orders addressed to World Bank Publications.) Credit Card Account Number Expiration Date Signature O Invoice me and please reference my Purchase Order No. Please ship me the items listed below. Customer Internal Stock Number Author/ Title Routing Code Quantity Unit Price Total Amount s All prices subject to change. Prices may vary by country. Allow 6-8 weeks for delivery. Subtotal Cost $ 'Ibtal copies Air mail surcharge if desired ($2.00 each) $ Postage and handling for more than two complimentary items ($2.00 each) S 'Ibtal $ ISUM . Thank you for your order. DISTRIBUTORS ALGERIA FINL^ND NnINA RPuBC SIGAPORE, TM , SAENA Office des Pubications Universnires Akateeminen KlrUakauppa Pan Korea Book Corporation Information Publications OF W ORLD BANK Centrale de Ben-Aknoun P.O. Box 128 P Ox 101. Kwangwtamun Private, Ltd. OF ORW M NK Algiers SF-00101 Seoul 02-06 Ist FI, Pei-Fu IndusMnal 8dg. PUBLICATIONS T 0 - s 4 Helsinki 10 Tel: 723-1421 24 New Industrial Road PUBUCA TION Tel: 651-1221121 WJAlSintaoe15 Ca4ls Hirsch. SRL FRANCE MEMRB Tel 283-1798 Galeria Guemes World Bank Publications PO. Box 5465 SPAIN Florida 165, 4th Floor-Ofc 4531465 66 Avenue d*Irna Kuwait Mundi-Prensa Litios, SA 1333 Buenos Aires 75116 Paris Tel: 421843 Castello.37 Tel: 30-7122, 33-1787 Tel: 723-54-21 MALAYSIA 28016 Madnd ALSTRALA, PAPUA NEW GRMANY, FlDERAL Universiy of Malaya Cooperative Tel: 276-997 GUNA, FIN. SOLOMON RPU IC OF Bookshop. bmited SRI LANKAAN THE MALDIVES IRANDS, AND VANIATU UNO-Verlag P.O. Box 1127 Jalan Pantai Baru Lake House Sookshop Info-ine D-5300 Bonn I Kuala Lumpur 22-11 P.0. Box 244 Overseas Document Delivery Simrockstrasse 23 Tel: 565-000 100, Sir Chittampalam A Gardiner Box 50 GPO Tel: (0228) 2129 40 Mawatha Sydney. NSW 2001 GOC Colombo 2 Australia(NOE Tel: 1021282-1614 MEMRB San Lorenzo 153-11. Col. deValle Deleg. Tel: 32104 24, podamou Street Bento Juarez SWEDEN AUSTRIA Athens-11635 03100 Mexico City D.F. For single ies Gerold and Co. Tel: 7018082 7016965 Tel: 559-5211 ABCE Fntzes Kungl. Hovbokhandel A-l0ll Wien ~ ~ MOO~ Regevngsgatan 12 Box 16356 Graben 31 HON032 SON,OckALMOOCm Tel: (02221 52 22 35 Asia 2000 Ltd MEM4RB S132 tchl 6 Ff., 146 Prince Edward Road, W. 2 rue Molitre Racine Tel: 08 23 89 00 BANRAIN Kowloon Casablanca Fs npDme MEMRB Hong Kong Tel: 259461 Wennergren-Williams AB P.O.Box2210 Te: 3-052415Box 30004 Ma c 3 TT 3-805294/ THE NETHERIANDS 5-10425 Stockholm Tel: 232934 INDIA Medical Books Europe. 8V (MBE) Tel: 08-54 1200 USS Publishers Distrbutors Ltd Noorderwal 38, BANGLADESH Post Box 7015 7241 BL Lochem Micro Industries Development Assistance Ansan Road Tel: 05730-4866 Ubraine Pt Socity (ICAS Ne Deli 110026 rue Grenus S.POiety 800S NwDRo100 NEW ZEAAN Case postal 381 GTel:273601-3 R Hill and Sond. 1211 Geneva II Dhaka 10 First Main Road Private Sag Tel: 31-89-50 Tel: 326427 Gandhi Nagar Ne MarketI BILGI Bangalore 560009 Auckland TANANIA tra s Nations Unies e 29206 Tel 54-8119 Oxford Uniersity Press Fv u ln des Aape"n Chmbrs P.O BoxQ 736 P0. Box 5299 1060 Brussels 5 Wallace Street NIGERIA Dar es Salaam Tel: 021S3&51.69 Bombay 400001 University Press Limited Tel: 29209 Tel: 260562. 260827, 266971 Three Crowns Building Jericho BRAZIL 81-8, Chowinghee Lane Private Mail Sag 5095CTraD Publicacoes Tecnicas internacionais Ltda. Calcutta 700016 Ibadan Central Department Store Rua Peixoto Gomide, 209 Tel: 212451, 249473 Tel: 0-22 411356 3 Slon Ro 01409 Sao Paulo. SP 7/188, 1 CAL, Swarup Nagar NORWAY 236930-9 Tel: (01t 257-1640, 258-167,258-8442 Kampu 208002 Tanum-Kar Johan, AS. CANA A Tel: 81300 PO. Box 1177 Sentrun TRINIAD S T Le Diffuseuir 5 n Roa Oso Systematics Studies Unit CP. 85,1501 rue Amp&e Nungambakkarn Tel: (02) 80 12-60 55 Eastem Main Road Bch 501wsle Ou 6e Madras 600034 Crp Canada O48 5E Tel 476355 PAKISTAN Trinidad, West Indies Canaa 14 SEbMirza Book Agerrcy Tel: 1-1809 662-5654 Tel: 15149 641-1334 IDONESIA 65, Shahrah-e-Ouaid-e-Azam CHIL Pt Indira Limited P.O. Box No. 729 TUNISI n Editorial Renacimienro A. Sam Rauangi 37 Lahore 3 Socit6 Tunisierne de Diffusion Miraflores 354 Jakarta Pusat Tel. 66839 5 Avenue de Carthage Mao 3PO Box 181 Tunis Tel: 35244 Tel 881-018 PANAMA Tel: 255-000 COLOMIA IEdiciones Libreria Cultural Enlace Lida. TCPuisher Epamna 16 Hat Kitapevi AS. Apartado Aero 34270 2 North Fredrick Steet Panama Zone 1 469, Istikal Caddesi Bogota Tel: 23-5628 ogu an Tel: 245-0645 Tel: 74483S5il Tel: 234-957028 COSTARICA TALYPERU C Tr cA IA Comissioaria Editorial Desarrollo SA UGANDA Ubei rlo CS Wn%)n anoiSA Apartaclo 3824 Uganda Bookshop Calle 11-13 Via Lamarmora 45 Ica 242. OE. 106 P.O. Box 7145 Av. Fernandez Guell laP 552 Ual 5Kampala San Jose 50121 Florence Tl 85-0Tl 3S Tel: 303-285 LISA THE PHILIPPINES UNIT= ARAB EMIRATES cmw IOET COAST MEMRB ME NCentre d don et de Diffusion Akca National Book Store P.0 Box 6097 SP.Om 800 Sevc (ESDAu arm 70n74 Shaxjah Njicosia Of BAP 541 etro Manila1 Tel: 592233 Tel: 21-99333 Abidjan 04 PlateauiTl 0-21 Tel: 22-20-55 an 326-2 FR1111SLUIE IGO N ROe md IEsenBo evc oj~p oSamfutndsUmatur WAA Livrana Portugal MmnoLd DK-1970 Copenhagen V 37-3. Hongo 3-Chome, Bunkyo-Wu 113 T 20 Aton Hampshire GU 34 2PG Tel: 01-35 19 42 Tokyo Tel 328220 United Kingdom Tel: 103) 438-2611 SAUDI ARANA Tel: AMton 86848 BOYP, ARAB MMMIJC OF Jarir Book Store AdAhaOMRBA P.O. Box 3196irei elEt Gaila Street Ryadh-I 1471 eEt Cavo P.O.Box3143 Apdo 60.337 JaalSud rai Caracas 1060-A Tel: 758333 Amman Tel: 4773140 Tel 32 2301/3326 04 KENYA WESTERN SAMOA Aftica Book Service (EAI Ltd. Wesliey Bookshop P.O. Box 45245 P.O. Box 207 Naffobi Apia Tel: 23641/330272 Tel: 22 822 The World Bank Headquarters European Office Tokyo Office 1818 H Street, N.W. 66, avenue d'l&na Kokusai Building Washington, D.C. 20433, U.S.A. 75116 Paris, France 1-1 Marunouchi 3-chome Telephone: (202) 477-1234 Telephone: (1) 47.23.54.21 Chiyoda-ku, Tokyo 100, Japan Telex: WUI 64145 WORLDBANK Telex: 842-620628 Telephone: (03) 214-5001 RCA 248423 WORLDBK Telex: 781-26838 Cable Address: INTBAFRAD WASHINGTONDC