Data on income distribution in the Philippines

DRAFT EPD INCOME DISTRIBUTION PROJECT DATA ON INCOME DISTRIBUTION IN THE PHILIPPINES V.V. Bhanoji Rao Division Working Paper No. 1980-4 August 1980 Economic and Social Data Division Economic Analysis and Projections Department Development Policy Staff The World Bank Division Working Papers report on work in progress and are circulated for Bank staff use to stimulate discussion and comment. The views and interpretations in the Working Paper are those of the author and may not be attributed to the World Bank or its affiliated organizations. DATA ON INCOME DISTRIBUTION IN THE PHILIPPINES This paper contains a brief review of the data on income distribu- tion for the Philippines. Following an evaluation of the data from the Family Income and Expenditure Survey (FIES) for 1970-71, the data on average income and average expenditures by income class were used to derive, from the distribution of expenditures by income class, a more representative distribution of income for rural, urban and all Philippines. This paper can be expanded as and when published data from the FIES 1975 become available. However, as is discussed in this paper, the quality of the data from the FIES 1975 is generally considered to be inferior to that of the FIES 1971. ECONOMIC AND SOCIAL DATA DIVISION ECONOMIC ANALYSIS AND PROJECTIONS DEPARTMENT AUGUST 1980 TABLE OF CONTENTS Page I. Introduction ............................................. 1 II. Sources of Data: A Brief Review ......................... 2 III. The 1971 and 1975 FIES Income Estimates .................. 4 IV. The 1971 Survey: A Description ........................... 6 V. The 1971 Survey Data: An Evaluation .................... 7 VI. Derivation of Income Distributions from Expenditure Distributions for 1970-71 ..............................11 VII. Implications of the Derived Distributions ................13 Bibliography ...................................................16 Annex Tables ...................................................17 DATA ON INCOME DISTRIBUTION IN THE PHILIPPINES I. Introduction The April 1980 edition of the World Bank's social indicator data sheet for the Philippines contains summary data on income distribution for 1961, 1971 and 1975. Data for the first two years were also reported in S. Jain (1975). The data for 1971 (actually May 1970-April 1971) were extensively /1 used by researchers in the Philippines and abroad- and were also reported in the World Development Indicatois for 1979 and 1980. It is now agreed by researchers (see Mangahas and Barros, 1979) and some Bank staff (see Chander and Dailly, 1979) that the 1975 survey did not cover the upper income groups adequately, and hence the data are unreliable. This fact was emphasized in /2 a recent Bank report (in draft),- which considered the data on expenditures, rather than those on income, when discussing the incidence of poverty. In view of the consensus against the reliability of the 1975 data, either the 1971 data have to be used or one has to await data from a survey planned for the early 1980s. This paper contains an evaluation of the 1971 data on income dis- tribution. Despite the limitations of the data on income, which are noted, income distributions were derived for rural, urban and all Philippines, using the respective distributions of expenditures. It is believed that the dis- tributions are fairly representative of actual distributions. /1 See Mangahas and Barros (1979).for a review of the research and an exhaustive bibliography. From a draft paper of Ullman (1974), one notes that the Bank has obtained some special tabulations for 1971. /2 World Bank, East Asia and Pacific Regional Office, The Philippines: Poverty, Basic Needs and Employment, (confidential draft) Report No. 2984-PH, 3 volumes, May 9, 1980. -2- /1 II. Sources of Data: A Brief Review- The sources of data for recent years are shown in Table 1. The Philippine Family Income and Expenditure Surveys (FIES) conducted at approximately five-year intervals since 1956 are the principal sources for income distribution in the Philippines. All the surveys contained a compre- hensive question on income. Nevertheless, the total personal income estimated from the surveys for the relevant years except 1975 amounted to only about two-thirds of the personal income derived from the national accounts.- In 1975, according to the estimates of Lim (1978), aggregate household income based on the FIES (40,060 million pesos) constituted only 44 percent of total personal income (88,955 million pesos) based on the national accounts. However, as will be shown in Section III below, the personal income estimated from the national accounts may be slightly less than the Lim estimate, so that the FIES estimate may equal 53 percent of the personal income from the national accounts. Even that proportion is still quite low. The National Demographic Survey (NDS), 1973 had only two questions on income, namely, the respondent's estimates of cash and non-cash income. Under-reporting of income was considered to be high, since the principal purpose of the survey was to obtain demographic information. The socioeconomic survey of 1975, like the FIES, collected compre- hensive data on income. Respondents also estimated their rents and home-consumed production. Average household income came to only 5,731 pesos, slightly less than the 5,840 pesos of the FIES 1975. Thus the socioeconomic survey also apparently underestimated income. /1 The discussion in this section is based on Mangahas and Barros (1979), which, as noted, contains a rather exhaustive bibliography, as well as a detailed discussion of income inequality issues pertaining to the Philippines. /2 See Lim (1978) for data and review. -3- Table 1: INCOME DISTRIBUTION IN THE PHILIPPINES-- SOURCES OF DATA FOR RECENT YEARS Source and latest Sample size year for which (No. of house- Comments data are available holds) Family Income and Expenditure 11,659 Conducted since 1956 Survey, 1971 and 1975 (1971) by the National Census and Statistics Office. National Demographic Survey, 8,434 Conducted in 1968 and 1973 1973 by the Population Institute of the University of the Philippines. Socio-Economic Survey, 1975 2,920 Conducted by the School of Economics and the Popula- tion Institute, University of the Philippines, in association with the Devel- opment Academy of the Philippines. Survey on Filipino Family 1,770 Conducted by the Philippine Household: Distribution Social Science Council of Income and Consumption (PSSC). Expenditure Patterns, 1974 -4- The survey on Filipino Family Households (PSSC Survey), which covered the period October 1973 - September 1974, included 1,770 households (out of an estimated total of over 6 million). The sampled households came from 18 areas which did not fully represent the nation as a whole. The esti- mated average household income derived from the survey was 8,901 pesos, much higher than that of any other survey, but 14 percent less than the average household expenditures of 10,185 pesos. Further, the estimate of personal income derived from the survey came to 70 percent of the personal income estimated from the national accounts. While this result gives some credence to the validity of the survey as compared with other surveys, the Gini ratio for income distribution was 0.62, much higher than the generally expected range of 0.4 to 0.5 for the Philippines and similarly placed developing coun- tries. In spite of its limitations, the PSSC survey is remarkable for its comprehensive coverage of income and the documentation of sampling and non- /1 sampling errors.-- III. The 1971 and 1975 FIES Income Estimates The brief account in Section II indicates that the FIES is about the best source of data on income distribution for the Philippines. However, for reasons noted below, only the data for 1971 were used in the evaluation presented in this paper. The Bank report on poverty, basic needs and employment in the Philippines, presented in May 1980, provides a comparison of the FIES estimates of income and expenditures for various years and compares them with the estimates derived from /1 The abovekaccount emphasizes quantitative aspects. For a review of the qualitative aspects of survey administration, data processing, etc., in general, see Chander and Dailly (1979) and Chander (1980). -5- the national accounts. The data for 1971 and 1975 from the Bank report are provided in Table 2. On its face, the 1971 survey appears to be better than the 1975 survey insofar as crude indicators of under-reporting can be used as a guide. Moreover, the 1975 data are not available in published form, nor are they readily available in unpublished form-- Thus the 1971 survey was used in the detailed description and evaluation presented here. Table 2: COMPARISON OF NATIONAL ACCOUNTS AND FIES ESTIMATES OF FAMILY INCOME AND EXPENDITURES, 1971 AND 1975 Item 1971 1975 National accounts personal income (mill. pesos) 34,790 76,257 National accounts personal consumption (mill. pesos) 30,778 67,644 No. of families (thousands) 6,347 6,860 National accounts average family income (pesos) 5,481 11,116 National accounts average family consumption (pesos) 4,849 9,861 FIES average family income (pesos) 3,736 5,840 FIES average family expenditure (pesos) 4,566 6,940 FIES income - national accounts income ratio 0.68 0.53 FIES expenditure - national accounts expenditure ratio 0.94 0.70 Source: World Bank, East Asia and Pacific Regional Office, The Philippines: Poverty, Basic Needs, and Employment: A Review and Assessment, Vol. III (Confidential draft), May 9, 1980, p. 28. /1 The Bank's regional department had access to unpublished data, supplied on a confidential basis. -6- /1 IV. The 1971 Survey: A Description- Reference Period: May 1970 to April 1971. Sampling procedure: Municipalities, enumeration districts and house- holds were selected in stages for each province and for urban and rural areas. Sample size: The household was the ultimate sampling unit. But income and expenditure data were for families. However, the disparity between "household" and "family" is not expected to be significant with respect to distribution of income and expenditures in summary form. Data collection procedure: Interviews were conducted in May 1971. A great deal of information was collected on the labor force, as well as on income and expenditures. For most households, data were collected in a single interview. Income and expenditures: The data on income and expenditures included both cash and non-cash items. The questionnaire from which income was obtained from different sources was quite long, based on a comprehensive income concept. Imputed rents on owner-occupied houses were included as part of income, as were most transfer payments. Annual family expenditures were also tabulated. Expenditures for food, beverages and tobacco were obtained for the week preced- ing the interview;/2 for several other regularly consumed items, average weekly or monthly expenditures were the basis for annual figures. For durable goods, actual expenditures for the 12-month period were included. Cars, jeeps and motor boats were not treated as consumption expenditures. Personal taxes were included in expenditures. /1 This section is essentially a summary of the detailed description in The Philippines, Bureau of the Census and Statistics, Family Income and Expendi- tures: 1971, Series No. 34 in The BCS Survey of Households Bulletin, Manila: 1973. /2 Thus seasonal factors may affect the data on food expenditures. -7- Reliability of data: The sample size of 11,659 households was fairly large, though it represented only about 0.2 percent of total Philippine house- holds. The main questions on reliability are likely to arise not from sampling, but from non-sampling errors. Data were obtained in one interview for a variety of items, all in very great detail, and where interviewers and respondents got tired, some inaccuracies might have been introduced in the reporting and recording. There was also the problem of lapses in recall with respect to income and expenditures in kind. Some differences in the basis of valuation of home- grown food, etc. might occur across respondents and interviewers. Most of these problems, however, are not peculiar to the Philippine survey, but are common to moqt surveys of family income and expenditures. V. The 1971 Survey Data: An Evaluation The published data on income and expenditures in the report, Family Income and Expenditures: 1971, are tabulated for all Philippines and its various geographic regions. Since the survey followed closely and was based on the sample frame of the 1970 census, not much would be gained by comparing the data from the two. For instance, data from the survey on the number of families by family size were used to arrive at the population covered by the survey--36.5 million persons, very close to the estimated population of about 37.8 million on May 1. As noted earlier, the survey did not cover personal income adequately, but seems to have covered personal expenditures quite adequately (see Table'2 above). The survey data on income and expenditures by income class for the Philippines as a whole and its rural and its urban areas respectively are shown in Tables A.1, A.2 and A.3 in the Annex. The information in Table A.1 indicates -8- that as many as 94 percent of the families spend more than what they earn! This finding is difficult to accept, even though it is generally expected that some families, especially in the lower income groups, can have considerable dissavings.-- The data in Table A.1 also indicate that the personal dissaving of Philippine families is 4.7 billion pesos, as against an estimate derived from /2 the national accounts of family savings of about 4 billion pesos.-- From Table A.1 it can be seen that the last three income groups experience a posi- tive savings and that their savings taken together total about 1.2 billion pesos. This is still very much less than the figure of 4 billion pesos. Thus even if the dissavings of the lower income groups are ignored, the total of savings for all income groups would still not come close to that computed from the national accounts. While this may be so, the least that can be done is to assume that for families in the first 10 or less income groups, the levels of expenditures by income class provided in Tables A.1 through A.3 are better indi- cators of income than are the reported average incomes themselves. There is one problem with using the average expenditure levels to show income distribution and for the computation of summary measures of inequ&lity. Consider, for instance, the data for rural Philippines in Table 3. /1 In India, for instance, 6.8 percent of the households have negative savings, according to a 1975-76 survey. See National Council of Applied Economic Research, Household Income and Its Disposition, New Delhi: The Council, January 1980, p. 197. /2 The figure is derived from the data in Table 2 (average family income of 5,481 pesos and average expenditures of 4,849 pesos give an average savings of 632 pesos per family). Expenditure in FIES included spending due to direct taxes. Thus, saving here is net of direct taxes. -9- Table 3: RURAL PHILIPPINES--AVERAGE FAMILY INCOME AND AVERAGE EXPENDITURES BY INCOME CLASS, 1970-71 Number of Average Average Income class families family family (pesos) ('000) income expenditures (peso) (peso) Less than 500 292 341 1,607 500 - 1,000 700 755 1,985 1,000 - 1,500 666 1,248 2,407 1,500 - 2,000 615 1,738 2,822 2,000 - 2,500 447 2,236 3,225 2,500 - 3,000 351 2,741 3,601 3,000 - 4,000 505 3,442 3,926 4,000 - 5,000 288 4,446 4,744 5,000 - 6,000 168 5,467 5,482 6,000 - 8,000 192 6,861 6,505 8,000 - 10,000 91 8,903 8,082 10,000 - 15,000 74 11,792 9,115 15,000 - 20,000 19 16,740 12,643 20,000 + 26 28,239 12,551 All classes 4,434 2,818 3,474 Source: Family Income and Expenditures, 1971. - 10 - For most of the income classes, the reported average income is very close to the midpoint of the income ranges for those classes. Moreover, the income classes are many. Hence, it can be assumed that intra-class inequality is not significant, and summary measures of inequality can therefore be computed based on average incomes. However, the average expenditures by income class are substantially larger than the average incomes by income class and it is not appropriate to assume that families with a given average expenditure level /1 belong to a narrow income or expenditure class.-- Thus, even though the data in Table 3 are useful for indicating the disparities in reported income and expenditures and provide a basis for suspecting a nearly universal downward bias in reported incomes, the data on expenditures in the table may not be used as a proxy for a straightforward distribution of expenditures by expendi- ture class as shown in Table A.4 in the Annex. The difference between the distribution of families according to average expenditure level when arranged by income class (as in Table 3) and the distribution of families when arranged by expenditure class (as in Table A.4) is demonstrated by the Gini ratios in Table 4 below. Table 4: CINI RATIOS FOR THREE TYPES OF INCOME/EXPENDITURE DISTRIBUTIONS, PUBLISHED DATA FOR THE PHILIPPINES . Gini Ratios Type of distribution Rura Ranioa Rural Urban Total Families by income class (Tables A.2, A.3) 0.4586 0.4542 0.4882 Families by expenditure class (Table A.4) 0.3576 0.3952 0.4110 Families by average level of expenditure, arranged by income class (Table 3) 0.2532 0.3139 0.3243 /1 In Table 3, for instance, in the income class 500 - 1000 pesos, there are 700,000 families with a reported average income of 775 pesos, almost equivalent to the class midpoint. However, the average expenditures of these families are 1,985 pesos, way above the upper limit of the income class. Thus it is not possible to assumd that the 700,000 families belong to a narrow expenditure group such as 500 - 1,000 pesos. - 11 - The Gini ratios for families when arranged by average level of expenditures (by income class) are unrealistically lower than the Gini ratios for families when arranged by expenditure class. Thus, if expenditure levels are to be used at all as proxies for true income levels, the distribution of families must be considered by expenditure class, from which the income distribution can be derived. The family income Gini ratios in Table 4 indicate that the rural ratio is slightly larger than the urban one. This is not generally expected. In this respect, too, the expenditure data appear to be more realistic than the income data. It is therefore considered appropriate that income distributions be derived from the expenditure distributions shown in Table A.4. VI. Derivation of Income Distributions from Expenditure Distributions for 1970-71 Table 3 shows that expenditures for the income classes through 5,500 pesos are consistently higher than income. It is assumed that income was underreported 'and that income and expenditures can be considered equal for those classes. Commensurate with this, in Table 5, which provides data on expenditure distribution, it is assumed that up to and including the expendi- ture class 5,000 - 6,000 pesos, the distribution serves as a proxy for income distribution. (In other words, there will not be a significant number of families who spend 5,500 pesos or less, but who have larger incomes). It was also assumed that for families with an average expendi- ture above 5,500 pesos, average income would be greater than average 11 The observations in the last group in Table 3 are not included, since the average expenditure level in that group appears to be rather low, in fact slightly lower than the average expenditures of the previous group. - 12 - expenditure, since in Table 3 average incomes exceed average expenditures beyond the 5,500 peso level. The following 4 pairs of observations from /1 Table 3 indicate this:-- Average income (Y) Average expenditures (C) 6,861 6,505 8,903 8,082 11,792 9,115 16,740 12,643 These 4 pairs of observations provide a linear expenditure equation with an r2 of 0.986: C (in thousands) = 2.4058 + 0.6033Y (in thousands). This equation is then used to find the level of average income for the families with C (average expenditures) above 5,500 pesos in Table 5.-- With regard to urban, families, average expenditures are considered to be a proxy for income through the average expenditure level of 7,000 pesos, instead of 5,500 pesos as in the rural areas. For families with average expendi- tures above 7,000 pesos, the relationship between C and Y is used to obtain /3 average income levels.-- The income distributions derived from the expenditure distributions are shown in Tables A.5 through A.7. Some of the implications of the derived distributions are considered in the next section. /1 The observations in the last group in Table 3 are not included, since the average expenditure level in that group appears to be rather low, in fact slightly lower than the average expenditure of the previous group. /2 It is almost certain that the 4 pairs of income-expenditure levels used to fit the relationship have some error. It is hoped that the errors are randomly distributed and affect both C and Y. This procedure for deriving the level of income from consumption is necessary since the expenditure distribution in Table 5 is the basis for deriving the income distribution. An alternative procedure would be to allocate estimated savings to the upper expenditure classes, but there is no way to formulate a basis for such estimation and allocation. /3 The readers may note that the relationship is based on rural data. Urban data are inadequate for the purpose (see Table A.3). - 13 - Table 5: FAMILY EXPENDITURE DISTRIBUTION, RURAL PHILIPPINES, 1970-71 Expenditure Mean Number of class (pesos) expenditures* families ('000) (pesos) Less than 500 250 13 500 - 1,000 750 208 1,000 - 1,500 1,225 443 1,500 - 2,000 1,750 658 2,000 - 2,500 2,250 604 2,500 - 3,000 2,750 533 3,000 - 4,000 3,500 769 4,000 - 5,000 4,500 439 5,000 - 6,000 5,500 248 6,000 - 8,000 7,000 284 8,000 - 10,000 9,000 115 10,000 - 15,000 12,500 89 15,000 - 20,000 17,500 22 20,000 + 25,960 9 All classes 3,487 4,434 * Midpoint of the class interval except for the last class, which is open-ended, for which the estimate was based on a pareto fit for the last two classes. - 14 - VII. Implications of the Derived Income Distributions The derived distributions in Tables A.5 (rural), A.6 (urban), and A.7 (national) provide the following information: Rural Urban National Average family income (pesos) 3,726 8,473 5,157 Average family savings (pesos) 238 1,473 610 Average savings rate (%) 6.4 17.4 11.8 Gini ratio 0.3941 0.4712 0.4705 The estimated average family income nationally is 5,157 pesos, 94.1 percent of the average derived from the national accounts (5,481 pesos from Table 2). Total family savings amount to 3.9 billion pesos, close to the 4 billion pesos expected on the basis of the national accounts. The Gini ratios appear reasonable. The urban ratio is significantly larger than the rural ratio, as expected, not the case in the published data. As for the decile shares, those from different sources are noted below: Family Percent of income received in 1970-71 Percentile Data in Data in Data in Annex Derived FIES S. Jain to WDR, 1980 in this paper lowest 20 3.6 3.9 3.7 5.2 second 20 8.1 8.0 8.2 9.0 third 20 13.3 12.9 13.2 12.8 fourth 20 21.0 21.2 21.0 19.0 top 20 54.0 54.0 53.9 54.0 - 15 - As anticipated, the shares reported in the FIES report, the Jain compilation and the annex to WDR, 1980 are close, since they were all based on the published data on income distribution of the FIES. The estimates derived in this paper attribute slightly higher income shares to the two bottom quintiles, while the share of the top quintile is unchanged. - 16 - BIBLIOGRAPHY 1. Chander, R. and J.P. Dailly. "Household Surveys in the Philippines." Draft paper. The World Bank, Economic Analysis and Projections Department, Washington D.C., 1979. 2. Chander, R. "Recent experience in household surveys in Malaysia and the Philippines." Mimeographed paper. The World Bank, Economic Analysis and Projections Department, Washington, D.C., 1980. 3. Jain, S. Size distribution of income: a compilation of data. The World Bank, Washington, D.C., 1975. 4. Lim, Antonio C. "An evalaution of income distribution data sources in the Philippines." Unpublished master's thesis. University of the Philippines, 1978 (quoted in Mangahas and Barros, 1979). 5. Mangahas, Mahar, and Bruno Barros. The distribution of income and wealth: a survey of PhilipDine research. Philippine Institute for Development Studies, Manila, July, 1979. 6. Ullman, Carmel J. "Income distribution in the Philippines." Draft paper. The World Bank, Development Research Center, Washington, D.C., 1974. - 17 - ANNEX: TABLES Table A.l: FAMILY INCOME DISTRIBUTION IN THE PHILIPPINES, 1970-71 (PUBLISHED DATA) Number Average Average Percent Income class of family family Pc in pesos families income expenditure families ('000) (pesos) (pesos) Less than 500 329 337 1,641 5.2 500 - 1,000 76-7 754 2,009 12.1 1,000 - 1,500 773 1,250 2,458 12.2 1,500 - 2,000 748 1,743 2-,951 11.8 2,000 - 2,500 611 2,245 3,364 9.6 2,500 - 3,000 517 2,744 3,766 8.1 3,000 - 4,000 795 3,446 4,307 12.5 4,000 - 5,000 475 4,452 5,182 7.5 5,000 - 6,000 316 5,455 6,114 5.0 6,000 - 8,000 403 6,866 7,397 6.4 8,000 - 10,000 226 8,909 9,107 3.6 10,000 - 15,000 235 11,994 11,594 3.7 15,000 - 20,000 71 17,092 15,832 1.1 20,000 + 81 31,746 18,915 1.3 All classes 6,347 3,736 4,479 100.0 Source: Family Income and Expenditures, 1971. - 18 - Table*A.2: FAMILY INCOME DISTRIBUTION, RURAL PHILIPPINES, 1970-71 (PUBLISHED DATA) Number Average Average Percent Income class of family family of in pesos families income expenditure families ('000) (pesos) (pesos) Less than 500 292 341 1,607 6.6 500 - 1,000 700 755 1,985 15.8 1,000 - 1,500 666 1,248 2,407 15.0 1,500 - 2,000 615 1,738 2,822 13.9 2,000 - 2,500 447 2,236 3,225 10.1 2,500 - 3,000 351 2,741 3,601 7.9 3,000 - 4,000 505 3,442 3,926 11.4 4,000 - 5,000 288 4,446 4,744 -6.5 5,000 - 6,000 168 5,467 5,482 3.8 6,000 - 8,000 192 6,861 6,505 4.3 8,000 - 10,000 91 8,903 8,082 2.0 10,000 - 15,000 74 11,792 9,115 1.7 15,000 - 20,000 19 16,740 12,643 0.4 20,000 + 26 28,239 12,551 0.6 All classes 4,434 2,818 3,474 100.0 Source: Family Income and Expendittires, 1971. - 19 - Table A.3: FAMILY INCOME DISTRIBUTION, URBAN PHILIPPINES, 1970-71 (PUBLISHED DATA) Number Average Average Income class of family family Percent in pesos families income expenditure families (1000) (pesos) (pesos) Less than 500 37 305 1,913 1.9 500 - 1,000 67 741 2,253 3.5 1,000 - 1,500 107 1,258 2,774 5.6 1,500 - 2,000 133 1,766 3,546 7.0 2,000 - 2,500 164 2,272 3,743 8.6 2,500 - 3,000 166 2,750 4,114 8.7 3,000 - 4,000 290 3,453 4,972 15.1 4,000 - 5,000 187 4,462 5,858 9.8 5,000 - 6,000 148 5,442 6,833 7.7 6,000 - 8,000 211 6,871 8,208 11.0 8,000 - 10,000 135 8,913 9,799 7.1 10,000 - 15,000 161 12,086 12,733 8.4 15,000 - 20,000 52 17,222 17,014 2.7 20,000 + 55 33,441 21,990 2.8 All classes 1,913 5,867 6,810 100.0 Source: Family Income and Expenditures, 1971 . - 20 - Table A.4: DISTRIBUTION OF FAMILIES BY EXPENDITURE CLASS, PHILIPPINES, 1970-71 Expenditure Mean class in expenditure Number of families ('000) pesos Rural Urban Total Less than 500 250 13 4 17 500 - 1,000 750 208 19 227 1,000 - 1,500 1,225 443 48 491 1,500 - 2,000 1,750 658 75 733 2,000 - 2,500 2,250 604 117 721 2,500 - 3,000 2,750 533 136 669 3,000 - 4,000 3,500 769 271 1,040 4,000 - 5,000 4,500 439 239 678 5,000 - 6,000 5,500 248 203 451 6,000 - 8,000 7,000 284 284 568 8,000 - 10,000 9,000 115 161 276 10,000 - 15,000 12,500 89 220 309 15,000 - 20,000 17,500 22 75 97 20,000 + * 9 61 70 All classes ** 4,434 1,913 6,347 * Estimates were obtained by fitting a pareto distribution for the last two classes. Estimated means are 25,960 pesos for rural and 31,300 pesos for urban Philippines. Based on these two, the mean for the Philippines as a whole worked out to 30,613 pesos. ** 3,487 pesos for rural families, 7,000 for urban, and 4,546 for all Philippines. Source: Family Income and Expenditures, 1971. -21- Table A.5: DERIVED FAMILY INCOME DISTRIBUTION, RURAL PHILIPPINES, 1970-71 Average family No. of families Percent of income '000 families in pesos 250 13 0.3 750 208 4.7 1,225 443 10.0 1,750 658 14.9 2,250 604 13.6 2,750 533 12.0 3,500 769 17.3 4,500 439 9.9 5,500 248 5.6 7,615 284 6.4 10,930 115 2.6 16,732 89 2.0 25,019 22 0.5 39,042 9 0.2 3,726 4,434 100.0 Note: The mean income levels were derived from the mean expenditure levels in Table A.4. Up to and including the average expenditure level of 5,500 pesos, average expenditures and average income were assumed to be equal. Above 5,500 pesos, average income (Y) was ob- tained from average expenditures (C) on the basis of the relationship C(000) W 2.4058 + 0.6033 Y(000) (See text for the data used to obtain the above relationship.) - 22 - Table A.6: DERIVED FAMILY INCOME DISTRIBUTION, URBAN PHILIPPINES, 1970-71 Average family No. of families Percent of income in pesos (1000) families 250 4 0.2 750 19 1.0 1,225 48 2.5 1,750 75 3.9 2,250 117 6.1 2,750 136 7.1 3,500 271 14.2 4,500 239 12.5 5,500 203 10.6 7,000 284 14.9 10,930 161 8.4 16,732 220 11.5 25,019 75 3.9 47,893 61 3.2 8,473 1,913 100.0 Note: The mean income levels were derived from the mean expenditure levels in Table A.4. Up to and including the average expenditure level of 7,000 pesos, average expenditures and average income were assumed to be equal. Above 7,000 pesos, average income was obtained in the same way as in the case of rural Philippines (see note, Table A.5). - 23 - Table A.7: DERIVED FAMILY INCOME DISTRIBUTION, ALL PHILIPPINES, 1970-71 Average No. of Cumulated Cumulated family families percent of percent of income "000) families income in pesos (%) (%) 250 17 0.26 0.01 750 227 3.84 0.53 1,225 491 11.58 2.37 1,750 733 23.13 6.29 2,250 721 34.49 11.25 2,750 669 45.03 16.87 3,500 1,040 61.42 27.99 4,500 678 72.10 37.31 5,500 451 79.21 44.89 7,308 568 88.16 57.57 10,930 276 92.51 66.79 16,732 309 97.38 82.59 25,019 97 98.91 90.00 46,755 70 100.00 100.00 5,157 6,347 - - Source: Derived from Tables A.5 and A.6. - 24 - Table A.8: INCOME SHARES BY FAMILY PERCENTILES, ALL PHILIPPINES, 1970-71 Cumulated family Cumulated income percentile share (%) (%) 10 2.0 20 5.2 30 9.3 40 14.2 50 20.2 60 27.0 70 35.5 80 46.0 90 61.5 95 74.9 99 90.8 100 100.0 Source: Obtained by linear interpolation from the cumulative percentages in Table A.7.