a,~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Transcription

1 Public Disclosure Authorized '~~~~~~7 ~ j; ~ ~ ~ ~ ~ ~ 1 tk, ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~A a,~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Public Disclosure Authorized NX - ~ ~ ~ ~ ~ ~ ~ ~ ijla.nfl-rvii losure Authorized Public Disclosure Authorized

2 CURRENCY EUVENS Average Exchange Rate Year (Pesos per US$)

3 FOR OMCLIAL USE ONLY THE PHILIPPINES: RECENT TRENDS IN POVERTY, EMPLOYMENT AND WAGES Table of Contents Page No. EXECUTIVE SUMMARY... v I. INTRODUCTION ti) The Background. 1 (ii) Objectives of the Present Report.. 1 II. FAMILY INCOME AND EXPENDITURE DATA A. Sources of Data... 2 (i) Income/Expenditure Data Prior to (ii) Data Sources and Data Availability, B. Limitations of the ISH... 3 (i) The Income Concept in ISH... 3 (ii) Underreporting of Income..3 (iii) Differential Implications of Family Income and Family Expenditures.. 5 (iv) Ranking by Family Income, Family Expenditure and Per Capita Expenditure: Illustration of Quantitative Differences.6 III. TRENDS IN POVERTY AND INEQUALITY.. 8 A. Trends in Poverty Incidence 8 (i) Poverty Trends, 'ii) Poverty Trends, (iii) Poverty, 1979 and 1983: A Panel Study of 3,294 Households.. 12 (iv) The Poverty-Employment Linkage.14 B. Trends in Inequality (i) Family Income Inequality (ii) Rural-Urban Inequality This report is based on the findings of a reconnaissance mission which visited the Philippines in September The mission consisted of Bhanoji Rio (mission leader) and Sudhir Anand (consultant). The report was discussed with the Government in March Subsequently, in June 1985, written comments from the Government were received and these were taken into account in the preparation of the final report. This document has a resticted distribuon and may be used by recients only in the peorfnce of thek ofic dutis Its cuntents may not otbaise be dicosed witout Wod Bank authizaton.

4 IV. TRENDS IN EMPLOYMENT AND WAGES A. Employment Trends (i) The Data Situation (ii) Employment and Unemployment (iii) Sources of Employment Expansion B. Trends in Real Wages.... * (i) The Data (ii) The Trends, (iii) Wages and Poverty Thresholds V. RECOMMENDATIONS (i) Monitoring Poverty (ii) Data Collection and Dissemination (iii) Statistical Organization REFERENCES ANNEXES: 1. Integrated Survey of Households The 1975 Integrated Census of the Population and Its Economic Activities The 1981 Census of Agriculture and Fisheries Sources of Employment and Wage Data Poverty Lines in Earlier Philippine Research Poverty Lines Based on Rice, Other Food and Non-Food Expenditures Statistical Appendix TABLES IN TEXT 1. Comparison of National Accounts and ISH Income Estimates, Comparison of National Accounts and FIES Income and Consumption Expenditure Estimates, Poverty Estimates from Income and Expenditures Distributions Poverty Lines and Poverty Incidence, Poverty Incidence by Main Source of Income Shifts in Poverty Group , Urban and Rural Areas Quintile Income Shares Population, Labor Force, Employment, Unemployment and Underemployment, Underemployment Among the Poor and Nonpoor by Occupation and Industry, Fourth Quarter, Index of Average Daily Basic Wage Rates of Laborers in Industrial Establishments in Metro Manila Daily Wage Rates of Laborers in Agricultural Operations,

5 - iii illustrative Comparison of Poverty Threshold Earning Levels and Average Daily Wage Rates, TABLES IN ANNEXES ANNEX I 1. Summary Information on ISH Surveys - ISH Family Income Data Status of ISH Tabulation ANNEX 4 1. Sources of Data on Employment, Employment Estimates from the Household Surveys and Establishment Surveys 1978 and Comparative statistics on Employment Status of Persons 15 Years Old and Over, Based on Weekly and Quarterly Reference Periods, Third Quarter, ANNEX 6 1. Rural and Urban Population, 1980 and Retail Price of Rice, 1975, 1978 and 1983 by Region Regional Consumer Price Indices Price Indices for Deflating the 1983 Poverty Line to 1975 and Poverty Lines, Comparison of Selected Poverty Lines.. 75 STATISTICAL APPENDIX 1. Per Capita Income and Consumption, Family Income Distribution Data, Third Quarter, Family income Distribution Data, Fourth Quarter, Family Income Distribution Data, Third Quarter, Family Income Distribution Data, Fourth Quarter, Family Income Distribution Data, Third Quarter, Family Income Distribution Data, Fourth Quarter, Family Income Distribution Data, Third Quarter, NCSO and World Bank Data on Metro Manila Income Distribution, First Quarter, World Bank Data on Metro Manila Household Expenditure Distribution World Bank Data on Metro Manila per Capital Expenditure Distribution * Poverty Incidence Trends, , Ban'k Poverty Report, and Abrera Study Regional Poverty Lines and Poverty Incidence, Poverty Incidence, , Among a Panel of 3,294 Families Percentage Distribution of Families with Specified Income Increase Between 1979 and Poverty Incidence by Level of Education 1979 and Income Shares of Bottom 20% and Top 20% of Families in Selected Countries Estimates of Average Annual Family Income, Rural and Urban GDP Per Worker in Agriculture and Nonagriculture,

6 20. Estimates of Ratios of Agricultural and Nonagricultural Incomes Per Worker, Terms of Trade in Agriculture, Estimates of Number of Households, Population and Average Household Size, o s Poverty Incidence Among the Employed Persons, Structure of Employment, c.*...e Employment Change by Sectors in Urban Areaso...s Employment Change by Sectors in Rural Areas... *... * Direct Government Employment in Relation to Total Employment, 1976 and a Employment Growth by Category of Workers Employment Change by Major Regions es Unemployment Rates by Sex, Metro Manila and Other Areas, 1976 and Occupation by Three Principal Educational Levels, 1976 and Average Daily Basic Wage Rates of Laborers in Industrial Establishments in Metro Manila, By Occupation, Nominal Daily Wage Rate Changes in Metro Manila, Legislated Minimum Wage and Reported Wage, Estimated Number of Households and Employed Persons by Rural and Urban, 1976 and Projected Labor Force,

7 v - THE PHILIPPINES: RECENT TRENDS IN POVERTY, EMPLOYMENT AND WAGES EXECUTIVE SUMMARY Introduction 1. Despite rapid economic growth during the 1970s, the Philippines has been noteworthy a country having a relatively high level of poverty. A 1980 Bank study _ noted that in 1975, 452 of the Philippine families were estimated to be living in poverty. Real wages, furthermore, showed a persistent downtrend during the 1970s, despite increases in real per capita income. The report also pointed out that the distribution of income changed little during The purpose of this review, which is the outcome of a reconnaissance mission which visited Manila in September 1984, is to make a quick survey of recent trends in poverty, employment and real wages, evaluate the relevant data and make recommendations to improve the data base. The Trends in Brief 3. This report points to both adverse and favorable trends affecting living conditions of the people of the Philippines in the past ten to twelve years. One adverse trend was the decline in the real wage, which was also associated with an increase in underemployment, relatively high rate of growth of unpaid family workers in the rural areas and the failure of industry to absorb additions to the urban labor force. The other adverse trends include an increase in family income inequality and a widening of rural-urban income differentials. The major favorable trend was a reduction in the percentage of families living in poverty in both rural and urban areas. The apparent contradiction between the wage and poverty trends can in part be explained by the following observations. First, in rural and urban areas, even though real wages declined, families headed by wage-earners generally had lower levels of poverty incidence. Thus, a decline in real wages still left most of the families with total family incomes above the poverty line. Second, demographic changes point to a reduction in average household size from 6 in 1971 to 5.6 in 1980, while the increase in female labor force participation has resulted in an increase in the average number of employed persons per household from 1.8 in 1971 to about 2 in These two changes lowered the dependency burden from 3.3 in 1971 to 2.8 in Despite these observations, the deficiencies in the data base, both in regard to wages and in respect of family incomes may also be at the root of the apparent contradictory trends in wages and poverty. 1/ Aspects of Poverty in the Philippines: A Review and Assessment, Report No PH, December 1, 1980.

8 - vi - The Data Situation 4. Since 1976, the National Census and Statistical Office collected data on family incomes as part of the Quarterly Integrated Surveys of Households (ISH). The earlier Family Income and Expenditure Surveys (FIES), obtained data on both incomes and consumer expenditures. Expenditure distributions are generally preferred to income distributions for poverty analysis. However, since ISH did not collect expenditure data, poverty analysis in the present report had to be based on income data. 5. A second limitation of the ISH is that the data officially published and tabulated so far are mostly based on income per family, which does not adjust for family size differences. This limitation is partly surmounted in this report by obtaining a few special tabulations on the distributions of families by per capita income class. 6. Employment and wages data also have some limitations. The coverage and concepts of labor force and employment have changed in 1976 with the launching of the ISH. Data on rural wages cover only four crops - rice, corn, coconut and sugarcane. Urban wage data were limited to Metro Manila and the collection of the data on a comparable basis was discontinued in Trends in Poverty Incidence A. Major Findings 7. The analysis in this report is based on poverty lines that provide the minimum required food and nonfood expenditures. For the benchmark year 1983, separate poverty lines were computed for rural and urban areas and the thirteen regions of the country. Special poverty level food and nonfood weights were used to extrapolate the 1983 rural and urban poverty lines to earlier years. 8. There was a decline in the incidence of poverty in the early 1980s compared to the early 70s. The percentage of the families below the poverty line declined from 57% to 47% ( ) in the rural areas, from 35% to 28% in the urban areas and from 51% to 41% in the country as a whole. Furthermore, during , the level of poverty did not change significantly despite worsening economic conditions and declining per capita incomes. The number of familes in poverty, however, increased from 2.5 million in 1971 to 2.8 million during in the rural areas, and from 3.3 million to 3.6 million nationally. In the urban areas, the number remained constant at about 0.8 million. In 1984, the poverty situation probably worsened because of the very high rate of inflation and the negative economic growth rate. 9. The highest rates of poverty incidence (between 50% and 51%) were found in Western and Central Visayas. Relatively high rates (around 40%) were also observed in Ilocos, Cagayan Valley, Bicolt Western Mindanao and Northern Mindanao. Rates of around 30% were found in Central Luzon, Southern Tagalog, Southern Mindanao and Central Mindanao. This finding is broadly similar to that of the Bank's 1980 study.

9 - vii A little over half (54%) of the urban families derived their main livelihood from the non-agricultural sector by providing skilled and unskilled labor. Poverty incidence among these families was relatively Low at 16Z. The next numerically large group comprises a fifth of the urban families deriving income predominantly from dividends, rents and pensions. This group also had a relatively low poverty incidence (18.5%). Among the families who were ownaccount workers in non-agricultural sectors (18% of the total), poverty incidence varied from 17% to 24%. Finally, farming, fishing and Livestock in the urban areas were characterized by relatively high rates of poverty incidence (between 45% and 50%); but only 8% of the total urban families depended on these activities in In the rural areas, relatively high rates of poverty incidence occurred among the families dependent on corn, coconut, other crops (notably sugarcane) and fishing, and among agricultural labor families. Even among the families dependent on manufacturing and those deriving livelihood from dividends, rents and pensions, the poverty incidence rates were close to 40% to 50%. Non-agricultural labor was a source of livelihood with relatively low poverty incidence (26%). This group, however, accounted for only a quarter of the total rural families. 12. The characteristics of a group of 3,294 families in the 1979 ISH which was also covered in the 1983 ISH were studied. Poverty incidence among these families declined marginally from 43% in 1979 to 40% in Of the poor in 1979, about 40% were able to raise themselves from poverty by 1983, while 60Z continued to remain poor. Meanwhile, of the nonpoor in 1979, 25% fell below the poverty line by 1983 while 75% continued to be nonpoor. Thus, while the overall poverty incidence changed only marginally between 1979 and 1983, there was substantial mobility between the poor and the nonpoor groups. Some 27% of the rural families and 23Z of the urban families were poor in 1979 as well as in Within the families referred to in the previous para, among those headed by persons with elementary education, regardless of urban or rural location, poverty incidence was a high 50% in Rural-urban differentials become greater, however, as one moves higher on the educational ladder. For instance, within rural and urban families headed by high school graduates, the respective rates of poverty incidence were 37% and 28%; and within the rural and urban families headed by college graduates, the respective rates were 23% and 14%. These rural-urban differentials as well as the fairly high absolute levels of poverty incidence may be in part due to the possible excess supply of high school and college graduates. 14. Data on the distribution of the employed population by earnings levels were available for By suitably converting the per capita poverty line into an earnings threshold, it was found that 34Z of the urban earners and 40% of the rural earners were poor (on the basis of earnings from work -- not counting pensions, rents, dividends and similar unearned income accruals). Data by sector of activity point out relatively high poverty incidence in all the services sectors in urban areas and in agriculture, manufacturing and distributive trade in rural areas.

10 - viii - Trends in Inequality 15. During the period , according to the available evidence, there was little change in the income distribution, and the bottom 40% of the families received a share of income of around 12Z. Between 1971 and , however, there was a slight worsening of income distribution. The share of the bottom 40% declined from 12% in 1971 to less than 10% during The shares of all the first four quintiles declined similarly, while the share of the top 20% increased from 53% in 1971 to 59% in The Philippines compares unfavorably to certain other Asian countries such as Hong Kong, Japan, Korea and Taiwan where the share of the bottom 20% of the families is between 6% and 9% compared to 3.9% in the Philippines. Income distribution in the Philippines is very similar to Malaysia and more or less similar to some of the Latin American countries. 17. As in the case of family income inequalities, there was a deterioration in the rural-urban income differentials. The rural-urban per capita income ratio declined from 0.75 in 1975 to about 0.5 in the period Income per worker in agriculture was about half of the income per worker in nonagriculture in This deteriorated to a third in the early 1980s. This was brought about, not by a deterioration in agricultural productivity, which in fact increase during , but by a deterioration in the agricultural terms of trade, by about 30% during This deterioration was partly the result of Covernment policies which tended to raise the prices of agricultural inputs and kept down output prices. It was also in part due to a relative deterioration in the prices of the country's export commodities. Domestic agricultural price policies are presently under review by the Government. Employment Trends 18. The average annual labor force growth rate during was a high 3.6%. This reflects both a rapid growth of the working age population, and a rising labor force participation rate especially in the case of the females. Between 1971 and 1982, labor force participation rose from less than 60% to 64%. The result has been a downward pressure on real wages and growing levels of underemployment. In the Philippines, as in many other developing countries, unemployment rates do not indicate the well-being of the population. In fact, low unemployment rates -- such as the 4.1% of the Last quarter of indicate that economic difficulties force many people to do some work and earn some income. The underemployment rate (percent of the labor force working less than full-time as well as working full-time, but wanting additional work), on the other hand, is a better indicator of the labor market situation. The rate was on the increase since 1978 and reached its peak, 30X of the labor force, in the last quarter of 1983, compared to 14% in The share of agriculture in total employment declined from 602 in 1965 to 50% in 1971 and then averaged 52-53% during Manufacturing share remained around 19-20% since Services share increased from 22% in 1965 to about 30% in 1971 and stayed at that level in the 70s and the early 80s. Thereewas thus virtually no relative shift from agriculture to industry in the past one and a half decades.

11 -ix Urban employment growth during (2.8Z p.a.) slackened compared to the growth rates during (4.9Z p.a.) and (3.6Z p.a.). Urban employment increase largely came from the services sector. This sector accounted for 93Z of the total job increase during , compared to 731 during and 641 during In general, industry failed to absorb the increase in the urban labor force. Recent economic difficulties especially reduced the tempo of employment expansion in Metro Manila. The region accounted for only 5X of the employment expansion during as against its share of 161 in the population increase. 21. Rural employment expanded at an annual rate of 5.1X during as against 4.71 during and a mere 0.61 during The very high rural employment growth rate during was partly due to the expansion of rural non-agricultural activities during and partly due to the drift of labor from the urban areas to the rural areas during the economic recession of In the rural employment increase, the largest shares were from agriculture (94Z during and 59X during ). During , however, rural services accounted for 281 of job expansion. This dominance of agriculture and services in recent years was, of course, accompanied by large increases in underemployment. In 1982, over a fifth of the total work force was unpaid family workers, the large majority (88X) being found in the rural areas. This category of workers grew at an average annual rate of 9.6% during , a clear indication of lack of regular jobs for wage and salary workers. 22. By occupation and industry, the evidence clearly points to relatively higher levels of underemployment among the poor than among the nonpoor. A general expansion of work opportunities will thus help alleviate poverty. Trends in Real Wages 23. Real wages of skilled and unskilled workers exhibited a declining trend in Metro Manila during During the trend continued except that workers in some of the skilled occupations appeared to have experienced modest growth in real wages. In agriculture, wage data were available in respect of workers in palay, corn, sugarcane and coconut. In general, real wages declined during , increased during and then declined again. 24. The increase in the agricultural wage rate during could be due to various factors. Relatively high rates of agricultural CDP growth occurred during This period also saw a stable overall CDP growth rate, relatively high investment growth rates, the favorable effects of the commodity boom of the earlier years, and special rural infrastructure programs. The incidence of underemployment was low during 1977 and 1978, an indication of a tightening of the rural labor market assisted by the movement of labor from rural to urban areas. Finally, during inflation rates were low. All these factors helped to bring about an increase in the agricultural real wages during The economic environment deteriorated in subsequent years, slack developed in the labor demand, and inflation started to erode the real value of the nominal wage rates.

12 25. In the urban areas, employment expanded at 4.9% during and 2.81 during but real wages either declined or stagnated. The healthy overall economic growth environment of the first period and part of the second was counter-balanced by excessive labor supply (indicated by the 70 to 90% service sector share in employment growth) and the result was only a modest increase in the money wage. Thus when the inflation rate was high during , real wages declined. They stagnated when the inflation rate was low during The deterioration in demand for Labor and the high inflation rates of the post-1978 period led to declining real wages. 26. Given the very high inflation rate and the negative rate of economic growth of 1984, real wages have probably dropped significantly in both rural and urban areas in the past year. Sustained economic growth and at most modest inflation over relatively long periods of time can provide a basis for growth in real wages. The process can be greatly aided by a slower labor force expansion. However, action on the population front today will produce benefits only some 15 to 20 years hence. For , however, the employment problem is of a serious magnitude, with annual additions to the labor force amounting to almost a million people. Monitoring Poverty R. Recommendations 27. At present, little direct attention is paid by the Government to monitoring poverty or the impact of development programs on poverty and employment. Because of the importance of this issue in future development, it is recommended that poverty monitoring should be undertaken on a formal and regular basis by the Covernment. To facilitate this, urgent attention is needed to streamline and improve data collection. Data Collection and Dissemination 28. Household Incomes and Expenditures: The practice of holding more than two quarterly ISH resulted in the accumulation of a backlog of unprocesse data. Both from the view point of quick data processing and to affect improvements in the survey operations, only one or at most two ISH per year should be conducted. Once every two or three years, the ISH should be expanded to collect family income and expenditure data. Tabulations should be prepared by per capita income and per capita expenditure classes in addition to tabulations by family income and family expenditure classes. 29. Data from the Agricultural Census should be processed quickly and MCSO should publish important national and regional tables on a comparable basis for 1971 and Labor Force and Employment: Each year in the ISH, labor force and employment data must be collected as at present. Establishment surveys should concentrate on obtaining data on output and value added rather than on employment and earnings. The Hinistry of Labor and Employment should have full access to ISH labor force and employment data and should discontinue its

13 - xi - household surveys. BAEcon should conduct, once in two years if not annually, comprehensive socioeconomic surveys, rather than the present rice and corn surveys, to obtain detailed information on the employment and related economic conditions in the rural economy. 31. Wage Rates and Prices: The Ministry of Labor and Employment should collect quarterly data on wage races in non-agricultural sectors in key cities. The proposed BAEcon rural surveys would provide rural wage rate data with an expanded coverage. NCSO should prepare and publish rural and urban consumer price indices for the general populatiod as well as low income groups. 32. Statistical Organization: The problems of data collection and disspmination are not limited to household expenditure, employment and wages. There is a need to review the entire statistical system to suggest ways to ensure the timely collection and dissemination of stacistical information without duplication of efforts. Such a review should address all relevant aspects including staffing, salaries and budget.

14 THE PHILIPPINES: RECENT TRENDS IN POVERTY, EMPLOYMENT AND WAGES I. INTRODUCTION (i) The Background The Bank's poverty report 11 on the Philippines was issued in It contained poverty incidence data based on a 1975 family income and expenditure survey (FIES) and detailed analysis of poverty based on a 1971 FIES. The present information on the poverty problem is thus quite dated, yet the problems of unemployment, poverty and living standards remain crucial. While no data of the FIES type have been collected during the past several years, some preliminary tabulations on household income distribution have been published for the period on the basis of "integrated surveys of households" launched in Published data from these surveys contain distributions of families by family income class for the different regions. These surveys do not provide data on family expenditures, and hence cannot yield results comparable with the FIES 1975 and FIES One can, however, try to evaluate the family income distributions from the surveys of and then make some inferences on recent trends in poverty. The Bank's 1980 poverty report, in the chapter on employment and wages, pointed out that the period was marked by a tendency for the real wages to decline. This was mostly attributed to the relatively high growth rate of Labor force (about 3X p.a.). The report also pointed out that in most sectors, value added per worker stagnated. Based on these two observations, it has drawn two implications: steps should be taken to slow labor supply growth and policies should be geared to increase employment and productivity. On the employment front, the report thus took an essential first step, namely, to signal the trends of very high labor force growth, stagnating productivity and declining real wages. These trends need to be updated, especially at the present time when the economy is going through a major crisis and the adjustment and recovery programs will have important implications to employment, productivity and real wages. (ii) Objectives of the Present Report Against the background described in the earlier section a reconnaissance mission visited the Philippines in September This report which is the outcome of the mission has the following objectives: (a) evaluation of the currently available post-1975 data on employment, wages and income distribution, with respect to coverage and quality of data; 1/ World Bank, Aspects of Poverty in the Philippines (1980).

15 -2J (b) a brief analysis of available data focusing on trends in employment/ unemployment, real wages, income distribution and numbers below the poverty line, pointing out significant changes, if any, in the situation since 1975; and (c) recommendations for improving the quality and coverage of the data. II. FAMILY INCOME AND EXPENDITURE DATA A. Sources of Data (i) Income/Expenditure Data Prior to 1976 In 1957, 1961, 1965, 1971 and 1975 the National Census and Statisti7al Office (NCSO) conducted family income and expenditure surveys (FIES).2 The survey of 1971 obtained data for the 12 month period of May 1970 to April 1971 and was conducted in May Income data for 1975 were available from the Integrated Census of the Population undertaken by NCSO and expenditure data from the Survey on Family Expenditures." The income data in 1975 were obtained in December 1975, the questions referred to annual income and the int?me questions were less definite and detailed in comparison to the 1971 FIES.- Expenditure data for 1975 were actually collected some five months after the income survey, thus with a relatively long recall period. (ii) Data Sources and Data Availability, After 1975, NCSO did not conduct a comprehensive family income and expenditure survey. Instead, starting in 1976 an integrated household survey (ISH) was held each year in selected quarters.' Income data from the surveys of 1976 through 1978 have not yet been cleaned; data for 1979 are being cleaned and data for 1980 through 1983 are cleaned but not yet processed (as of October 1984). At present, data are available on the aggregate distributions for 1980 through In addition, the Bank mission obtained income 2/ For a brief description of these and other related surveys, see Terasaki (May, 1977). 31 These two surveys correspond to the former FIES, but there are two differences compared to the earlier surveys: (1) field operations on income and expenditure were undertaken separately, and (2) expenditure survey used a sub-sample of the integrated census. 4/ See World Bank Poverty Report (1980). 5/ rd Quarter, st, 3rd and 4th quarters, all four quarters, all four quarters, 1980, 1981 and 1982: 3rd and 4th quarters, and 1983: 1st, 3rd and 4th quarters.

16 -3- data on about 3,300 families for 1979 as well as 1983 thus providing a basis for finding out what happened to the income of each family during (i) The ncome Concept in ISH B. Limitations of the ISH The ISH income concept is quite comprehensive. Total family income includes primary income (earnings) and receipts from other sources received by all family members during the reference quarter as participants in any economic activity (either as laborer, proprietor, capitalist or entrepreneur) or as recipients of transfers, pensions, grants, etc. Primary income includes wages and salaries, commissions, tips, bonuses, family and clothing allooances, and net receipts/profits derived from the operation of family-oriented enterprises/activities and the practice of a profession or trade. Imputed rental values of owner-occupied houses, interests, royalties, and profits from sale of stocks and bonds, rentals, including landowners' share of agricultural products, pensions, supports, gifts, winnings, proceeds from insurance, etc., constitute the receipts from other sources. Also included in the estimated total family income are the receipts from family sustenance activities which are not considered as family-operated activities (see Annex I for more details). (ii) Underreporting of Income (a) ISH-NationaL Accounts Comparison: The fact that the income concept is quite comprehensive does not automatically ensure a high degree of accuracy in reporting by respondents. There are hardly any igernal consistency checks that can be built into the various income questions.- The result is, more often chan not, a high degree of income underreporting deliberately as well as because of recall lapse and absence of accounting and bookkeeping practice by most families. In addition, when income data is gathered as part of a large questionnaire covering various other socio-economic aspects, the income question may not receive adequate attention by the respondent as well as the enumerator. One method employed to find the degree of undercoverage in ISH income is the comparison of ISH and national accounts data. For the period 1980 through 1983, the comparison indicated that ISH income fell short of personal income from national accounts 7o the extent of half to one-third, about the same as the pre-1980 record._ In contrast to this high apparent 6/ For instance, when detailed consumer expenditure data are also collected, a check can be made by comparing the two data sets. The respondents can be asked to explain the sources of other revenue if expenditure is high relative to income. Where income is quite high relative to expenditure, questions can be raised on the use of savings. 7/ See also the works of Mangahas cited in the references.

17 -4- undercoverage of income, the degree of undercoverage of consumption expenditure in the family expenditure surveys relative to personal consumpti in national accounts varied from 6% to 30Z in the pre-1980 FIES surveys. These results are not very different from the situation in other developing couitries. The fact that a comparison between the survey estimates of family income and personal income from national accounts provides a basis to assess the degree of underestimation in the surveys does not mean that an across-theboard adjustment of survey data is desirable and can be defended, for, there is no way one can safely make the heroic assumption of an equal degree of under-reporting by one and all of the thousands of families in the survey.- In any case, while such an adjustment will aff8 9 t the poverty incidence numbers based on an absolute poverty cut-off, it will not make any difference to the income distribution. Table 1: COMPARISON OF NATIONAL ACCOUNTS (NA) AND ISH INCOME ESTIMATES, N.A. Personal Income (P Bln) ISH family income (P Bln) a/ ISH-NA Income Ratio Sources: NEDA and NCSO a/ Based on 3rd and 4th quarter data. 81 It is quite plausible that personal consumption is understated in the national accounts since in all likelihood private investment has been overstated as evidenced by the very high ICOR's in recent years. 9/ For an example of such adjustments, see Daly et al, Income Distribution, Food Consumption and Poverty, Staff Paper No. 7, Ministry of Agriculture of the Philippines, June / An across-the-board adjustment of the incomes of all families would reduce, for instance, the poverty incidence of 59% in 1971 (based on the poverty line in the 1980 Bank report and the family income distribution) to 35.2%. Similarly, the poverty incidence in 1975 will decrease from 64.4% in 1975 to 23.3%. The poverty trend stands reversed because of a relatively higher degree of undercoverage of income in the 1975 survey.

18 -5- Table 2: COMPARISON OF NATIONAL ACCOUNTS (NA) AND FIES INCOME AND CONSUMPTION EXPENDITURE ESTIMATES, Ratio of FEIS to NA Income Ratio of FIES to NA Expenditure Source: World Bank, Aspects of Poverty in the Philippines, Vol. II, p (b) A Comparison of NCSO and World Bank Income Distribution Data for Metro Manila for First Quarter 1983: Metro Manila income data for the first quarter are available not only from the NCSO, but also from a household survey conducted by the University of the Philippines researchers under the auspices of the Bank's urban research group. From the NCSO data, the estimated average family income was P 3,860 per month. The World Bank survey estimate was P 4,184, about 8% more than the NCSO estimate. Another implication of the two data sets also may be noted. Based on a Metro Manila povlj>y line of P 695 per month per family for the first quarter of 1983,- the NCSO distribution indicated that 7% of the families were below the poverty line, whereas the World Bank data indicated a 4% poverty incidence. The implication is an under reporting of incomes in the NCSO survey in the very low income groups. (For instance, below a family income of P 1,000/month, NCSO indicated 18% of the families as against 10.8% in the World Bank survey.) (iii) Differential Implications of Family Income and Family Expenditures The Bank reporc issued in 1980 adopted a poverty line of 912 per capita or P 5,147 per family for the year When the family poverty line was applied respectively to the family expenditure and income distributions, the percentages of families below the cut-off were 53.2% and 64.4Z. This is because of the phenomenon that family income estimates were relatively more downward biased than expenditure estimates. Hence, to find the extent of poverty incidence, one prefers data on family expenditure distributions to family income distributions. The ideal data, however, are distributions of families by per capita expenditure classes. Such data are to a large extent, free from the bias in poverty incidence measures based on family expenditure distributions, where the bias arises due to the high degree of positive association between family expenditure and family size. Thus, when a per capita poverty line is defined and converted to a per family poverty line based on the average family size, and when the per family poverty line is applied to a family expenditure distribution to measure the incidence of poverty, the measure is biased on two 11/ For the methodology of this poverty line, see Annex 6 of this report.

19 counts. First, there will be some families below the cut-off with family size less than the average and second, there will be some families above the cutoff with family size larger than the average. As a result, some famili', below the cut-off and some above it, may be wrongly identified as poor _-/ and the net effect may be upward or downward bias on overall poverty incidence. (iv) Ranking by Family Income, Family Expenditure and per Capita Expenditure: Illustration of Quantitative Differences (a) National Data: The data in Table 3 illustrate the degree of difference between the poverty incidence numbers when a given poverty line is applied to family expenditure and income distributions. 12/ Ranking by family (or household) expenditure and ranking by per capita expenditure will not identify the same families below and above the appropriate poverty cut-offs. For instance, in the South Asian data sets, covering the period 1968 to 1975, Visaria found that only 14 to 28Z of the households were assigned to the same deciles when ranked by perhousehold and per-person criteria. Large families tend to have large total expenditures and smaller average expenditure and thus the particular households which get classified as poor will be almost entirely different between the two types of rankings. See Visaria, P., Poverty and Living Standards in Asia, LSMS Working Paper No. 2, World Bank, The disparities in the results of income distribution analysis from household income and per capita income rankings has been pointed out in Cautam Datta and Jacob Meerman, Household Income or Household Income Per Capita in Welfare Comparisons, W.B. Staff Working Papers, No. 378, World Bank, For example, the paper shows that the distribution of benefits from public education is very inegalitarian if household income is used as a basis for analysis. The reverse occurs when per capita income is used. In a recent paper, Lipton had reviewed the evidence and concluded that the ranking by outlay per household should be considered useless for poverty studies and that the ranking by outlay per person should be clearly preferred. See Michael Lipton, Demography and Poverty, World Bank Staff Working Papers, No. 623, World Bank, 1983.

20 -7- Table 3: POVERTY ESTIMATES FROM INCOME AND EXPENDITURE DISTRIBUTIONS Percent below poverty threshold Difference Source and Poverty Line Year A B B-A World Bank Report a/ Threshold P 500/annum/capita in Abrera study a/ Food Threshold P 520/annum7capita in Abrera Study al Total Threshold: P 877/annum/capita in a/ See Annex 5. A: Based on expenditure B: Based on Income The differences in the poverty incidence rates based on income and expenditure data tend to be relatively high in the middle of the income distribution, tapering off towards the end. There is no way one can derive an adjustment formula to proceed from one distribution to the other. The differences, however, could be quite substantial and significant. Clearly, one cannot combine poverty measures from expenditure and income distributions and assess trends. One has to consider either of them only, but, with the full understanding that expenditures tend to be better covered than incomes. (b) World Bank Data for Metro Manila for First Quarter 1983: The ready availability of the data set at the Bank helps to assess the differences in poverty incidence based on distributions of families by family income class, by family expenditure class and by per capita expenditure class. The poverty line per family in the First Quarter of 1983 for Metro Manila was P 695 per month (see Annex 6). World Bank data on the distribution of families by family income class indicate a poverty incidence of 4%. The family expenditure distribution, on the other hand, indicates a poverty incidence level of 7.5%. Next, when the poverty line is taken on a per capita basis, (F 131), poverty incidence amounted to 10.62, which is, of course, a relatively more realistic figure. (c) Concluding Observation: Comparative poverty incidence estimates obtained from family income and family expenditure distributions of NCSO indicated that the estimates from the income distributions were smaller than those from expenditure distributions. This will perhaps be the case as long as incomes are severely under reported relative to expenditures; but, even this cannot be taken for granted as evidenced in regard to Metro Manila. A more

21 -8- important point, however, is the invalidity of estimates based on the ranking of families by family incomes and family expenditures. From the World Bank data on Metro Manila on the average family size by family (monthly) expenditure class it was found that from "less than P-167" up to P-2,500, the average family size increased very rapidly from 2 to about 7. From P 2,500 to l'p 16,664 and over," it increases rather slowly from 7 to 8.3. The poor families would then appear to be those with the small average family sizes. On the other hand, when households are ranked by per capita expenditure, the household size declines as per capita expenditure increases and the poor families are those with the highest average household size, a result almost diametrically opposed to the one produced by a ranking of families by family expenditure. The conclusion is that distributions by per capita incomes and expenditures are the right data set for poverty studies. III. TRENDS IN POVERTY AND INEQUALITY Note on the Validity of the Results. As stated in Section II the data available for recent years ( ) are distributions of families by family income class (ISH data), and not by per capita income class. Similarly the data for 1975 and earlier years were 3 7ased on rankings of families either by family income or family expenditure.- The observations in Sections Ati) and A(ii) below are thus of limited significance in regard to poverty incidence data for any particular year. They could, however, be used as indicators o' broad trends over time. In addition, sections A(ii), and A(iii) below provide an analysis based on per capita income distributions drawn from the ISH of 1983 (Third Quarter). i) Poverty Trends, A. Trends in Poverty Incidence A study by AbreraJ±' found an increase in poverty incidence from 34% to 48% from 1965 to 1971 when the poverty threshold was based only on food adequacy, but found poverty incidence to be about constant at a high 70% when a much higher threshold was considered. A high threshold naturally inflates poverty incidence. On the one hand, it looks odd to infer a much larger poverty incidence in the Philippines than in, say, India where about 40 to 50% of the people have been found to be in absolute poverty. On the other hand, as the vast majority (such as 702 or 80Z) get classified as poor, such a measure is unlikely to be useful for policy decisions and actions with a sharp poverty focus. 13/ So far, NCSO has never published data by per capita expenditure or per capita income. However, the Bank poverty report (1980) obtained bivariate distributions of families by family expenditure and family size, thus overcoming the limitations of not having per capita expenditure distributions. 14/ See Abrera (1976).

22 -9 - The Bank's 1980 poverty report used a poverty line which roughly identified the bottom 40% in 1971 as belonging to the poverty group. Between 1971 and 1975 quite apart from a slight increase in the overall and rural incidences of poverty there was one notable trend in the data in the report. It was the jump in the urban poverty incidence from 29% in 1971 to 40% in Such a major change in a relatively short period did not occur in the evidence presented by Abrera for 1965 and (ii) Poverty Trends (a) Note on the Poverty Line: For the purposes of the analysis in this report, relatively conservative poverty lines were used. In 1983 the annual poverty lines were P 7,363 for a rural family and P 10,584 for an urban family. These amounts of purchasing power were just enough to provide the needed food and non-food items (See Annex 6). Special poverty level food and non-food weights were used to deflate the 1983 poverty lines and obtain the lines for earlier years. (b) Rural-Urban Poverty Trends: Several scholars have noted the inadequacy of the incom 5oncept and other limitations of the 1975 data compared to the data for Whether for this reason or not, one can see the conspicuous increase in the rural and urban poverty incidences in 1975 compared to 1971 (Table 4). If one ignores the 1975 data in Table 4, and considers 1971 and early 1980s, a modest and plausible decline in rural and urban poverty incidence is noticed in the early 1980s compared to There was thus some trickle down of growth to the lower income groups. The "fluctuations" are not worthy of any special interpretation since sampling errors, differences in the degree of coverage etc., can cause such minor variations. There is yet another interesting way to pick a data set from Table 4. This is by looking at broad data comparability of the income from the sample surveys and the income from national accounts. The last column of the table provides the ratio of the income estimates from the two sources. In 1971 and 1982, personal income in the household surveys amounted to a little over two-thirds of the personal income from the national accounts. The comparison of 1971 and 1982 confirms the reduction in rural and urban poverty incidences in the early 1980s compared to / See, for instance, the Bank's Poverty Report (1980) for a concise summary of the differences. A major problem was that the number of households estimated by the 1975 FIES was somewhat low, 6.8 million as against 7.2 million covered by the 1975 integrated census. See also Annex 2 for limitations of income data.

23 Table 4: POVERTY LINES AND POVERTY INCIDENCE, a/ Poverty lines Poverty incidence (per family) (Z of families below Survey to Year b/ pesos/annum poverty line) c/ NA income Rural Urban Rural Urban Rural and ratio Urban ,341 2, ,570 5, ,537 7, ,384 8, ,892 9, ,363 10, /a See Annex 6 for details of methodology relating to the poverty lines. 7i third quarter data. 7& Extreme poverty may be denoted with reference to an income level that is just enough to buy the food that can provide the recommended calories. This level of income equals the value of 16kg of rice per head per month (see Annex 6). The incidence of extreme poverty declined from 24.9% in 1971 to 20.1X in 1980 and 15.1% in 1983 in the rural areas and from 7.9Z in 1971 to 6.1% in 1980 and 5.9% in 1983 in the urban areas. The number of poor families in the country as a whole increased from 3.3 million in 1971 to 3.6 million in In the rural areas, the number increased from 2.5 milion to 2.8 million. In the uban areas, the number remained constant at 0.8 million. In an earlier para, it was noted that distributions of families by per capita income class (in contrast to per family income class) were obtained for the third quarter of Based on these distributions, rural and urban poverty incidences were 43% and 20% respectively, as against 45% and 26% noted in Table 4. The relatively large differential in regard to the urban areas must be due to the relative predominance of young families, with a small family size who get classified as poor when poverty incidence is measured on the basis of the distribution of families by family income class. Thus, the "true" poverty incidences, especially in the urban areas, will be less than the rates indicated in Table 4. However, it is expected that the conclusion on the general decline in poverty incidence in the early 1980's compared to the early 1970's is still valid. (c) Regional Variation in Poverty in 1983: Poverty lines and the rates of poverty incidence for the thirteen administrative regions are given in Table 13 of Annex 7 based on the rankings of families by per capita income. Metro Manila had the least poverty incidence of 11.2% in The highest rates of poverty incidence (about 50X) were found in Western and Central

24 Visayas. Relatively high rates (around 402) were also observed in Ilocos, Cagayan Valley, Bicol, Western Mindanao And Northern Mindanao. Central Luzon, Southern Tagalog, Southern Mindanao and Central Mindanao had rates around 30%. While these trends are broadly consistent with general perceptions, the incomes data could be subjecc to differential degrees of under-reporting and this could affect the regional comparison. (d) Poverty Incidence by Main Source of Livelihood, 1982 and 1983: Tabulations on the distribution of families by per capita income classes were specially prepared by NCSO for the rural and urban areas and by main source of livelihood for the third quarcers of 1982 and From these distributions, poverty incidence rates were computed by main source of livelihood. The results are given in Table 5. In the urban areas, where a third of the Philippine families live, the incidence of poverty was 19.2Z in 1982 and 19.7% in The distribution of families by main source of livelihood changed little in 1983 compared to A little over half (54%) of the urban families derived their main livelihood from the non-agricultural sector by providing skilled and unskilled Labor. Poverty incidence among these families was relatively low at 16%. The next numerically large group comprising a fifth of the urban families derived income predominantly from dividends, rents and pensions. This group also had a relatively low poverty incidence (18.5%). Among the families (18% of the total), who were own-account workers in non-agricultural sectors, poverty incidence varied from 17% to 24%. Farming, fishing and livestock in the urban areas were characterized by relatively high rates of poverty incidence (between 45% and 50%); but only 8% of the total families depended on these sectors. Among the rural areas, where two thirds of the country's families lived, significant annual shifts in the patterns of livelihood occur. For instance, for 9.9% of the country's total families (or 15Z of the total rural families) in 1982, the main source of income was palay farming. In 1983, only 5.2% of the total families (or 7.7% of all rural families) belonged to that group. On the other hand, families deriving major part of income from coconut increased from 3.7% of total families in 1982 to 4.9% of total families in A number of factors such as yield and price changes by crop determine whether the main source of income of a family stays the same or changes over short periods of time. In view of this, it is not possible to identify a fixed set of families as palay or coconut families, since no such clear-cut set exists year after year. In spite of this limitation, some labels are quite useful in order to associate source of livelihood and level of poverty incidence. The data in Table 5 indicate that relatively high poverty incidence (50% to 60%) occurs among the families dependent on corn, coconut, other crops (notably sugarcane) and fishing and among agricultural labor families. Even among the families dependent on manufacturing and those deriving livelihood form dividends, rents and pensions, the poverty incidence rates were close to 40% to 50%. Non-agricultural labor was a source of livelihood with relatively low poverty incidence (26%). It does seem that sustained poverty alleviation in the rural areas requires major structural changes in the rural economy, bringing about shifts in sources of livelihood from agriculture to nonagriculture.

25 -12- Table 5: POVERTY INCIDENCE BY MAIN SOURCE OF INCOME Third quarter Third quarter of 1982 of 1983 Poverty Z of Poverty Z of Region and main incidence total incidence total source of income (x) families (Z) families Urban Farming Fishing, livestock, forestry Manufacturing Wholesale and retail trade Other services Dividends, rents and pensions Non-agricultural labor Rural Farming - Palay Corn Coconut Other crops Agricultural labor Fishing Livestock, poultry, forestry Manufacturing Wholesale and retail trade Other services Dividends, rents and pensions Non-agricultural labor (iii) Poverty, 1979 and 1983: A Panel Studs of 3,294 Households (a) The Data: A household-by-household comparison of the filled-in questionnaires for the third quarters of 1979 and 1983 was undertaken by the NCSO staff and a total of 3,294 households, each with the same name and Location of the head of household in 1979 and 1983, were identified. For each household, household size, household income, head's occupation and head's education were analyzed for 1979 and The results are reported below. (b) Data are not a Representative Sample: The average household sizes in 1983 should be around 5.2 for rural and 6.3 for urban. The panel showed 5.6 for rural and 5.7 for urban. In the urban areas, relatively larger households were left out of the panel. The average family income in the third quarter of 1983 from the total sample was P 3,090 for rural and P 7,983 for urban. The panel estimates were P 3,224 for rural and P 5,796 for urban. Clearly, the urban data were least representative of the national sample,

26 biased downward both in relation to family size and average income and hence would over-estimate urban poverty. Despite this limitation, some insights can still be gained on the progress of the group of families over a period of four years. (c) Overall Poverty Incidence and Mobility from Poor to Non-Poor: Among the families under consideration, poverty incidence declined marginally from 45Z in 1979 to 421 in 1983 in the rural areas, and from about 40% to 35Z in the urban areas (Table 6). Further analysis of the poor and non-poor families revealed a very important feature. Among the urban poor in 1983, 652 were poor in 1979 also, but, as many as 35Z were not poor in Similarly, in rural areas, among the poor in 1983, 64% were poor in 1979 also while the remainder of 36X were not poor in Some 27% of the rural families and 23Z of the urban families were poor in 1979 as well as Nationally, poverty incidence declined from 43% in 1979 to 40% in 1983 among the families under discussion. Of the poor in 1979, about 40% were able to raise themselves from poverty by 1983, but, 60% continued to remain poor. At the same time, 25% of the nonpoor in 1979 fell below the poverty line by 1983 whereas 75% continued to be nonpoor. Thus, while the overall poverty incidence changed only marginally between 1979 and 1983, there was substantial mobility between the poor and the nonpoor group. Incomes of different families increased by different magnitudes during Among the urban and rural families respectively, 27% and 30% had zero or negative nominal income growth and a further 21% each had positive income growth below the inflation rate (58% from third quarter 1979 to third quarter 1983). Thus, a half of thelorban and a half of the rural families had experienced real income increases. Such volatility of income changes goes hand in hand with a high degree of mobility from poor to non-poor and non-poor to poor categories for a significant proportion of families. In such circumstances, reliable poverty analysis requires data on distribution of families by per capita expenditure since consumption expenditure is a reasonable proxy for "permanent" income especially for the low income groups. Despite this valid observation, a hard-core of rural and urban poor, those who were poor in 1979 as well as 1983, could be identified from the per capita income distributions and they made up for 27% of the families in the rural areas and 23% of the families in the urban areas. Further data on household characteristics are available at the NCSO and tabulation and analyses work on them would be useful for policy purposes. 16/The Development Academy of the Philippines collected data on socio-economic status and perceptions from 500 heads of households in 1981 (March- Nay). Considering the past three years (that is, ), 42% felt that their lives improved, while 31% responded that their lives deteriorated. See Mangahas (December 1981). These findings are comparable to the ones noted above.

27 - 14 Table 6: SHIFTS IN POVERTY GROUP, , URBAN AND RURAL AREAS Poor Families Nonpoor Families Status in in 1983 in 1983 Total Region 1979 No ('000) Z No ('000) Z No ('000) Z Urban Poor Nonpoor Total , Rural Poor , Nonpoor , Total , , Source: Panel of 3,294 families from ISH, 3rd Quarter Surveys (Data processed at the World Bank). (d) Poverty Incidence by FamiLy Head's Educational Level: Among the families headed by persons with elementary education, regardless of the location (urban or rural), poverty incidence was 52% in 1979 and 47Z-49Z in Urban-rural differentials begin to emerge as one moves up from high school to college and higher levels, with urban poverty incidence lower than rural poverty incidence by level of education. Among the families headed by high school and college qualified persons, poverty incidence rates respectively were 37X and 23% in 1983 in the rural areas and 28% and 14% in the urban areas. These are fairly high rates of poverty incidence and could be attributed in part to the possible excess supply of such graduates relative to needs. A related issue is the quality of education. tiv) The Poverty - Employment Linkage (a) A Note on the Data: The ISH tables provide data on earnin s of the wage and salary workers and the net income of own account workers. 7' The data are provided by occupation, industry, education and number of days worked in the refernece quarter. Such data for 1976, 1977 and 1978 are available in the ISH Bulletins on Labor Force. Data for are not yet fully tabulated, but, when they become available, some insight can be obtained on the relative positions of different occupational groups, sectors and educational 17/ Of the three categories of "employment status," the third, namely, unpaid family workers do not have any explicit earnings of their own.

28 categories. In this section, the linkages between poverty and emyhyment are analysed on the basis of the data for the fourth quarter of (b) From per Family to per Earner Poverty Lines: The data on earnings are for Hence the per family poverty lines for the third quarter of 1983 in Annex 6 were first extrapolated backwards to 1978 fourth quarter, by using rural and urban price indices. To convert the per-family 9 overty lines to per-earner, an average of 2 earners per family was assumed.9 The earnings thresholds that signify poverty in the fourth quarter of 1978 were P 750 per urban and P 543 for rural. (c) Poverty among Earners by Urban and Rural: Among the urban erners, 34Z were poor in The proportion was 40% in the rural areas. The urban poverty incidence among families would be less than the 34% indicated by poverty among the earners because incomes from interest, dividends, rents and transfers, which are not included in the earnings could be relatively more significant in the urban areas than in the rural areas and also possibly because of relatively higher under-reporting of earnings. Rural poverty incidence among families may be more than the incidence among earners because of the possible existence of a relatively large number of families living on meagre "other income" without participating in employment and therefore without any earning members. (d) Poverty among Earners by Occupation and Industry. In the urban areas, poverty incidence was 42% among the sales workers and 41% among the service workers. These two categories make up for a third of the urban earners. Another third of the earners were the occupational group of production workers, transport operators and laborers and 30% of this group had earnings below the threshold. In the rural areas, sales workers, service workers and agricultural workers respectively had poverty incidence rates of 45Z, 49Z and 43%. Data by sector of activity point out relatively high poverty incidence in the services sectors (distributive trade, community services and personal services) in urban areas and in agriculture, manufacturing and distributive trade in the rural areas. In fact, rural manufacturing had a poverty incidence of 52% and distributive trade had 45% as against the 42% in agriculture. Perhaps, in the contemporary rural setting, landlessness drives people into some petty manufacturing and trade activity with relatively high poverty incidence. The urban areas also attract the landless where since "petty manufacturing" can be very expensive to launch, the avenues open are distributive trade and personal services. 18/ NCSO, Integrated Survey of Households Bulletin: Labor Force, Third Quarter 1978, Fourth Quarter 1978, Manila, September / The average number of employed persons per family was about 2 during

29 -~~ ~ ~ ~ ~ ~ ~ (i) Family Income Inequality B. Trends in Inequality (a) The Trend. The income shares of the bottom 40Z were 12.6% in 1956, 12.1% in 1961, 11.5% in 1965 and 12.0% in The overall Gini coefficients were respectively 0.48, 0.50, 0.gJ,and / The data for the different years were not strictly comparable,- and hence one might at best state that there was no strong evidence to conclude that inequality narrowed or widened during (b) The Trend. Data on quintile shares for 1971, 1975 and are given in Table 7. The first impression one gets is that the data for 1975 are out of tune with the rest since the degree of income undercoverage was the highest and since the income shares of the bottom 20% and the bottom 40% were relatively very high. Thus one might wish to ignore the 1975 data in the inter-temporal comparison. Second, the data for the period could be treated as a composite, since the quintile shares mostly exhibit fluctuations over a stable trend. Table 7: QUINTILE INCOME SHARES, a/ Shares in total income of families belonging to Ratio of ISH personal Year Bottom Bottom Bottom Bottom Top income to National 20a 40Z 60% 80% 20% Accounts estimate [1975] rd Quarter rd Quarter rd Quarter rd Quarter a/ Computed by straightforward interpolation of the cumulative family income distributions. From first and second, the third point emerges which is based on a comparison of the income shares of 1971 on the one hand and the shares on the other. The income shares of the bottom four quintiles declined and the share of the top 20% increased. One might also note that the relatively poor 20/ See Sta. Romana III (September 1976) 21/ See Terasaki (May 1977).

30 economic performance of the periodl_/ tended to depress the income shares of the middle and upper income groups and had the effect of inflating the shares of the lower income groups. If this is allowed for, the final conclusion that emerges is a widening of income inequality in the past decade ( ) of fairly high rates of economic growth. It was widely acknowledged that the periq was especially characterized by rapid agricultural sector growth- - at an average of about 5% a year - and the Leading sector in this growth was rice, the growth of which was assisted by irrigation, high yielding seeds and credit. The 24/ Government also undertook agrarian reforms in the rice and corn sectors.- It must now be concluded that all these had noc in anyf2/ helped to reduce the degree of inequality in the family income distribution. _ (c) Growth, Poverty and Inequality. In retrospect, it might be said that the growth record of the 70s in the Philippines was of mixed welfare consequences. To the extent that poverty incidence declined during the period both in the rural areas and the urban areas, growth did trickle down to the lower income groups. However, the trickle down was quite inadequate to improve the income shares of the bottom quintiles of the population. It was so because rapid economic growth over a short period of time can hardly be accompanied by structural changes, productivity improvements and employment shifts that lead to sustained increases in real wages and enhanced share of labor income in national income which bring about a reduction in overall income inequality. (The estimated share of wages and salaries in family income in the Philippines declined from 44% in 1971 to 39Z in 1982). {d) Philippines and Other East Asian Countries. Data on quintile shares are available for 1970 or 1971 for Hong Kong, Japan, Korea (Rep.), West Malaysia and Taiwan (1972). In all countries except West Malaysia, the income share of the bottom 20% of the families was between 5.6Z-8.8Z. In West MaLaysia it was 3.8Z, about the same as in the Philippines (3.9Z). The share of the top 20% varied from 37.2% in Taiwan to 49% in Hong Kong with Japan and Korea falling in between. Malaysia's top 20% had 56% share and in the 22/ See World Bank, The Philippines: An Agenda for Adjustment and Growth, Report No PH, November GNP growth slackened to 3.4% in 1981 and less than 2% in 1982 and 1983 as against the average of 6% during the 1970s. 23/ See H.T. Oshima (1983). 24/ See World Bank, Aspects of Poverty (1980). As Mangahas (December 1981) pointed out, there was a little of every thing (e.g. land reform, family planning, dispersal of industry, food self-sufficiency, rural credit, price control etc.) that could have led to reduction in inequality and poverty, but the actual outcome was different in regard to inequality. 25/ As data from the 1981 Agricultural Census, now being processed, become available, a full assessment of inequalities within the agricultural sector and trends in inequality could be undertaken.

31 Philippines it was 54X. The Philippine and Malaysian distributions are closer to those of the Latin Amercian countries than East Asian countries. It is well known that land reform has played an important role in reducing income inequities in Japan, Korea, and Taiwan. It does not appear to have as yet been the case in the Philippines, despite the well publicized land reform program affecting rice and corn farms. (ii) Rural-Urban Inequality (a) Income and Productivity Differentials. In a recen htudy, NEDA examined the changes in the rural-urban income differential L 6 1 The ruralurban per capita income ratio was 0.75 in 1975, as obtained from the basic data from the 1975 FIES. For the period , NEDA could not directly obtain rural and urban mean income estimates since the ISH tabulations were not available. Instead, NEDA used the NCSO average family inc2me estimates by region and classified some regions as rural and some as urban.- f On this basis, it was found that the rural-urban per capita income ratio declined from 0.75 in 1975 to 0.63 in 1980 and 0.60 in NCSO had since brought out the ISH household income data by rural and urban. During , rural income per head was no more than half of urban per capita income and there was no clear upward/downward tendency in the past four years. This ratio, however, was much less than the 0.75 observed in The evidence presented here suggests a deterioration of rural-urban differentials in the 80's compared to Another way of assessing the rural-urban differential is by looking at the proxies of agricultural and non-agricultural incomes, that is, GDP originating in each of the sectors. The productivity differential at current prices deteriorated in the 80s compared to GDP per person at constant prices increased during in both sectors but stagnated since then in both sectors. Average productivity in agriculture was a constant proportion of the productivity in non-agriculture. The ratio was a low one-third; but, the ratio of sectoral productivities would not be exactly equal to the ratios of sectoral per worker incomes, since the ratio of ov2pall personal income to total CDP in 1975 was about 0.78 in the Philippines,- and it could be different between the two sectors. Sectoral income-gdp ratios are not available. It may be assumge that in agriculture the ratio will be higher than the national average,2 and equals to, say, With an agriculture 26/ NEDA, Some Aspects of Rural-Urban Welfare Differential, (August 1984). 27/ The regions with rural population comprising over 70% of the total as rural areas. Regions III, IV, NCR and VII were thus classified as urban and the other nine regions were classified as rural. 28/ 0.78 in 1975 and 0.75, 0.76, 0.75 and 0.76 respectively in the four years 1980 through / In general, depreciation and other deductions as percent of CDP are likely to be relatively lower in agriculture.

32 share in GDP of 0.29, the income-gdp ratio in non-agricultural sector for 1975 can be derived as ( x 0.90) e 0.71 = This leads to a ratio of agric 0lIural income per worker to non-agricultural income per worker of On the basis of computations such as this, it was found that the agricultural income per worker at current prices which was almost half of the non-agricultural income per worker in 1975 deteriorated to a third by the early 1980s. It is thus clear that the agriculture to non-agriculture income differentials indeed widened and could have set the basis for the deterioration of the rural-urban income differentials noted earlier. This, however, happened even though the ratio of real (constant price) productivity remained constant. There are thus the following trends that need an explanation: (1) the real productivity improved in 1980, compared to 1975 in both sectors and remained stagnant during ; (2) the real productivity ratio remained constant throughout; and (3) the differentials in sectoral income per worker, and along with it rural-urban income disparity, deteriorated. Leaving aside the productivity stagnation, all other trends imply a deterioration in the agricultural terms of trade. (b) Terms of Trade in Agriculture. Since 1975, agricultural terms of trade deteriorated. The implicit price deflator for value added in agriculture, relative to the price index of agricultural inputs declined by 28.8% during During the same period, the agriculture to non-agriculture relatives of implicit price deflators of gross value added declined by between 25% and 30%. Relative to rural consumer prices, the implicit price deflator for agriculture declined by 33%. Over prolonged periods of modern economic growth, as labor moves from agriculture to industry, as rural real wage rates increase, and as mechanization improves agricultural productivity, one would expect a secular decline in agricultural terms of trade relative to industry. However, the post-1975 reduction in the agricultural terms of trade in the Philippines was not part of any secular trend brought about by the conditions of modern economic growth. Agricultural productivity improved or stagnated, agriculture continued to absorb labor at a positive rate, and rural real wages declined or stagnated. These resulted largely due to Government policies that affected agricultural prices and incomes on the one hand and policies that tended to discourage labor-intensive manufacturing and construction on the other. Prices of the Philippine export commodities also deteriorated compared to import prices. As pointed out by the recent NEDA study 31/ on rural-urban welfare differential, the deterioration in the agricultural terms of trade could be attributed largely to government intervention policies which tended to raise prices of agricultural inputs but kept down the prices of agricultural outputs. Such intervention policies involved the imposition of price controls, tariffs and taxes, and subsidies, which taken together, were biased 30/ (0.37) x (0.90 e 0.73) where 0.37 is the productivity ratio. 31/ NEDA (August 1984).

33 against the agriculture sector and in favor of the manufacturing sector. These policies were generally aimed at promoting industrial production, raising government revenues, and protecting consumer welfare. The cost of these policies, however, seems to have been the relative deterioration of farmers' incomes and the hampering of the growth of agricultural production. What is more, the overvalued exchange rate, the relatively low real interest rates, high effective protection and the various other industrial incentives promoted capital intensive production (for domestic use mostly) in the nonagricultural sectors. Large public corporations with access to huge amounts of domestic and external credit exacerbated the process of inefficient technological choices. Labor often found its way into low productivity service occupations rather than into manufacturing and construction. In addition, labor moved from the villages to the towns and into the urban services sectors. IV. TRENDS IN EMPLOYMENT AND WAGES (i) The Data Situation A. Employment Trends (a) Household Surveys. Household surveys in the Philippines conducted at different times in each year since 13ig provide labor force data on a continuous and fairly consistent basis. Between the old household surveys of and the post-1976 integrated household surveys, there were two principal differences: the first was in the age cut-off of the population covered (10+ in the former surveys and 15+ in the ISH) and the second was the reference period (one week in the former surveys and one quarter in the ISH). The first change is justified on basis that most children in the group are outside the labor force - and there is no point in going below the legal minimum age of 15 for entry into the labor market. The consequences of the second change are a slight increase in the overall employment rate and a decline in the unemployment rate. Thus, for instance, for the third quarter of 1976, the former survey concepts and ISH concepts respectively provided labor force participation rates of 59.5% and 60.5%, emplo.yment rates of 93.7% and 94.8% and unemployment rates of 6.3% and 5.2%. (b) Change in the Reference Period. The reference period change was quite an issue in the academic circles in the Philippines. Most academics contend that the ISH reference period of one quarter reduced the unemployment 32/ See Annex 4 for sources of data. 33/ According to the 1975 Integrated Census of Population, of the 5.6 million persons aged 10-14, 0.5 million or about 9% had "gainful occupations." If such trends continued in recent years, the cut-off of 15± might have excluded between 2-3% of the labor force.

34 rate to a misleadingly low level.341 While this is true, data on unemployment are hardly useful as indicators of economic well-being of the people in the developing countries where problems of low productivity, lack of skills and lack of full-time work are the principal dimensions of the employment problem. ISH provide data on the number of days worked in the reference quarter, crosstabulated by earnings, occupation, and education. Improvements in collecting these data by providing intensive training to the enumerators may be a better strategy rather than simply reverting to a one-week reference period. (ii) Employment and Unemployment (a) Labor Force Participation. The overall labor force participation rate has been on an increasing trend. For instance, in , the rate was around 60%, whereas in it was around 64%. Much of the increase was due to changes in the age structure in the case of males and increases in age specific participation rates in the case of females. The average of the annual labor force growth rates during was 3.6%. (b) Unemployment and Underemployment. In a country like the Philippines with predominance of rural activities, petty trade, small manufacturing and various types of personal and community services, establishments and households where production takes place are largely informal in organization structure and personnel practices. Firms (and households) cannot afford to come in and go out of business frequently as in the industrialized countries. Unemployment rates, therefore, do not move in unision with the upswings and downswings in economic activity. In fact, if anything, low open unemployment rates indicate the need for relatively more and more people to do some work and help families tide over an economic crisis, as in the case of 1983 (when the lowest unemployment rate of 4.1% was recorded for the fourth quarter - see Table 8). The inadequacy of employment opportunities is better reflected in the lack of full-time work when in fact people like to do more work and when people have full-time work and yet want more work presumably to earn a little more income to meet their family and personal obligations. The ISH classify people who want additional work by the number of days they have worked in the 34/ See for instance the University of the Philippines Workshop Report (June 1984) which says "In 1976 the reference period was changed from the past week to the the past quarter. Moreover employment was expanded to include such activities as home gardening and raising poultry, as long as there was a harvest. These changes effectively increased the numerator of the employment rates. The base of the employment rate, which is the labor force, was also changed. Where before the labor force included the population 10 years old or over, who were either employed or looking for work, the cutoff age was changed to 15 years, thus reducing the size of the denominator. All these changes tend to reduce the unemployment rate figure. What is additionally vexing is that, since data based on the older definition are no longer being collected, researchers no longer have a basis for comparing employment performance before and after 1976."

35 Table 8: POPULATION, LABOR FORCE, EMPLOYMENT, UNEMPLOYMENT AND UNDEREMPLOYMENT, /a 1976/a (Qr 3) (Average) (Average) (Average) (Qr 4) (Qr 4) (Qr 4) (Qr 4) Population 15 + ('000) 25,811 31,376 24,837 25,695 26,688 27,918 28,929 30,023 30,978 31,907 Labor force ('000) 12,911 16,244 15,018 15,328 16,502 17,536 18,543 19,005 19,698 20,521 Employed ('000) 12,246 15,427 14,238 14,547 15,652 16,802 17,746 18,017 18,614 19,671 Unemployed ('000) , Underemployed ('000) 1,862 1,634 3,628 2,897 2,324 n.a. 3,875 5,180 5,445 6,259 Visibly/b ('000) ,178 1,588 1,236 n.a. 1,928 2,827 2,813 3,405 Invisibly/c ('000) 1, ,450 1,309 1,088 n.a. 1,947 2,353 2,632 2,854 L.F.P.R. (%) As X of L.F. Underemployment (%) n.a ,5 Visibly (%) n.a Invisibly (M) n.a Unemployment (X) Population aged b Working less than 40 hours a week (60 days/quarter) and wanting additional work. EWbrking 40 hours or more (60 days/quarter) and wanting additional work. Source of Basic Data: NCSO Household Surveys.

36 reference quarter. Using a "full-time norm" of 60 days per quarter, those who want additional work (the underemployed) can be classified into the two categories of visibly underemployed (those who worked less than 60 days) and invisibly underemployed (those who worked 60 days or more). The highest underemployment rate was noted in the fourth quarter of 1983, the worst manifestation of the then emerging economic crisis. The rate, shgever, was increasing since 1978, after a temporary drop in 1977 and 1978._ (c) Employment Structure: Between 1965 and 1971, there was a reduction in the share of agriculture in total employment from about 60Z in 1965 to 50% in During the years , the sector share was around 52-53Z. Manufacturing share changed very little since 1965 remaining around 19-20%. Services share increased from 22% in 1965 to about 30% in 1971 around which level it continued during There was thus virtually no relative shift from agriculture to industry in the past one and one half decades. (iii) Sources of Employment Expansion (a) Urban and Rural Rates of Employment Growth. Employment growth during three sub-periods was as follows: : 3.6% p.a. in urban and 0.6% p.a. in rural, : 4.9% p.a. in urban and 4.7% p.a. in rural; and : 2.8% in urban and 5.1% in rural. While import substituting industrialization provided the impetus for urban employment growth during , the relatively high overall rates of economic growth and particularly the high rate of agricultural sector growth provided the impetus to urban as well as rural employment growth during In the period , however, employment grew at an apparently very high rate in rural areas compared to the urban areas which contradicts past experience. Two reasons may be advanced to explain this. First, as noted earlier, the rural (nonagricultural) sector activities were particularly at a peak level during Second, the lack of momentum in urban labor absorption, especially after 1980, resulted in a swelling of the numbers in rural employment without much increase in full-time employment. (b) Sectoral Contribution to Employment Expansion. During , in the urban employment expansion, 64% came from services. It was 73.5% during and 93.2% during Industry failed to absorb the growing urban labor force. Rural employment growth patterns differ between the three subperiods. During , 56.6% of the increase in total employment was from services, 30.5% from manufacturing and 24.5% from construction. Agricultural absorption, however, was negative, and as a result, overall rural employment growth was only 0.6Z p.a. The period was characterized by relatively 35/ NEDA, in its recent study of rural-urban differentials pointed out that rural employment expanded at a higher rate during than urban employment contrary to the usual trend. "It was specifically between that the promotion of rural-based non-farm activities was given emphasis, and rural infrastructure, such as rural roads, bridges and rural electrification and water supply were intensified to bolster employment activities in the rural areas." (NEDA, August 1984), p. 41.

37 large rate of labor absorption in agriculture. Some 94Z of the total employment expansion was due to agricultural employment increase. During , agriculture contributed 58.7% and services contributed 27.6Z to rural employment expan. n, accompanied, of course, by large increases in underemployment.3- tc) EmpLoyment Expansion by Class of Worker. The increase in employment during can be further analyzed by class of worker. Of the increase in urban employment 29% were wage and salary workers, 41% own account workers and as much as 30% unpaid family workers. In the rural employment growth, 44% was made up of unpaid family workers and 34% own account workers. In 1982, some 22% of the work force (or a total of 4.1 million) was unpaid family workers (3.6 million in rural areas). The category increased by 77% in six years or an extremely high rate of 9.6X p.a., a clear indication of lack of regular jobs as wage and salary workers and lack of small business opportunities to be own account workers. It also signals the existence of serious underemployment. (d) Employment Expansion by Major Region. Metro Manila, Luzon, Visayas and Mindanao are considered here as the major regions. During , Metro Manila accounted for 5% of the total employment increase, even though its contribution to national population increase was around 16%. Luzon and Mindanao shares in employment and population increase were about equal (35% and 34% in Luzon and 30% and 29% in Mindanao). Visayas accounted for about 30% of the employment increase, larger than the 21% share in the population increase. The relatively small share of Metro Manila in employment growth compared to other regions and compared to its own share in population growth would only imply relatively large open unemployment in Manila. This, of course, was validated by an 11% male unemployment rate and 16% female unemployment rate in 1982 in Metro Manila as against less than 2.7% male unemployment rate and 7.5Z female unemployment rate in the rest of the Philippines. The relatively low unemployment rates in the regions outside Metro Manila to a large extent mask the high levels of labor underutilization. Ce) Education and Occupation. Compared to most middle income countries, the Philippines historically had very high enrollment rates in secondary schools and higher educational institutions. For instance, in 1981, Philippines enrollment in higher educational institutions was 26%, which was surpassed 3n7y by four countries out of a group of 60 middle income countries.- Normally, education has a positive effect on income distribution and poverty alleviation; however, if there is a mismatch between 36/ Most researchers (e.g. Tidalgo 1976 and 1982) on Philippine employment trends have noted the labor saving nature of technology adoption in agriculture, industry and construction and the resultant expansion of service sector employment (as a last resort). While the comment on technology adoption may be true, the labor absorption patterns differ between rural and urban areas. 37/ See World Development Report, 1984, pp

38 education and occupation and if the educational levels are not adequately used, such a result may not accrue. The percentage of college level work force occupying professional, technical and related occupations was slowly declining in the Philippines (during it declined from 34.5% to 31.5Z) and this broadly supports the view that the highly educated work force has not been adequately utilized. The issue of appropriate utilization of educated manpower assumes even greater significance in the Philippines where higher education continues to expand. (f) Underemployment, Occupacion, Sector and Poverty. The ISH data of the fourth quarter of 1978 are used to asses the differences in the levels of underemployment among the poor and non-poor within the high poverty occupations and sectors. The data are given in Table 9. By occupation, the evidence clearly points to relatively higher levels of underemployment among the poor. A general expansion of work opportunities will thus help alleviate poverty. The same signal comes from the data by industrial sector. However, in the urban services sector, which has absorbed increasingly higher percentages of people in the past decade, poverty cannot be significantly alieviated by increasing the work opportunities, for, the percent underemployed was low and equal both among the poor and the non-poor in services. Poverty in this case was the result of low-productivity and low rewards. This comment applies to some extent to the service sector in the rural areas also. Another important aspect is the high and equal (60%) rate of underemployment in the rural agricultural and manufacturing sectors among the poor in contrast to the relatively low rate in the service sector poverty group. Household manufacturing of some sort was the alternative to agriculture in the rural areas and sufficient full-time work did not exist in both sectors. Thus, poverty alleviation in rural areas requires a major expansion of work opportunities both in agriculture (where the majority are now occupied and who presumably possess the skills for agriculture) and non-agricultural sectors. Si) The Data B. Trends in Real Wages A brief note on the sources of data and recent developments in regard to data collection is given in Annex 4. The following points summarize the situation. First, data on agricultural wages, obtained by the Bureau of Agricultural Economics were limited to labor force in rice, corn, coconut and sugarcane. The Bureau is continuing this activity. Second, the Central Bank, until 1980 collected wage rate data in Ketro Manila from selected establishments. This activity was abandoned when the Government assigned the responsibility to NCSO. Third, the responsibility was shifted to the NWC in 1983 and the first (preliminary) set of wage rates for March 1983 were available recently. Problems of comparability and limited coverage apply to both rural and urban wage series. (ii) The Trends, (a) Metro Manila, : The Central Bank series indicate a declining trend in che real wage rate of the skilled and unskilled workers

39 -26- Table 9: UNDEREMPLOYMENT /a AMONG THE POOR AND NONPOOR BY OCCUPATION AND INDUSTRY, FOURTH QUARTER, 1978 Urban Area Rural Area Among the Among the Among the Among the Occupation/Industry poor non-poor poor non-poor Occupations Sales workers Service workers Agricultural, forestry and fishery workers Production workers, transport operators and laborers Industries Agriculture, forestry, fishing Manufacturing Wholesale and retail trade Community and personal services /a Percent of persons working less than 60 days. Source of Basic Data: NCSO, ISH, th Quarter.

40 (Table 10). In actual fact, the drop was highest between (51% for skilled workers and 53Z for unskilled workers). Then followed a period of near v7stancy of the real wage ( ) with further declines in 1979 and _ These trends were brought about by steady increses in nominal wages and different rates of inflation during different periods. The inflation rates in Metro Manila in 1971, 1972, 1973 and 1974 were 15%, 10, 14% and 34% respectively. These are high rates compared to 1975 (8%), 1976 (6%), 1977 (8%) and 1978 (7%). The situation worsened again in 1979 (19%) and 1980 (19%). (b) Metro Manila, : The Central Bank data for 1980 and the NWC data for 1983 (March) for a set of ten occupational categories indicate that for some groups real wages have declined in March 1983 compared to These groups were electricians, foremen, masons and common laborers. The other groups (carpenters, compositors, latheman, linotypists, painters and plumbers) had higher real wages in March 1983 compared to (c) Agricultural wages, : The BAEcon data on nominal wage rates in palay, corn, sugarcane and coconut are converted into real wage rates by applying the consumer price index for areas outside Metro Manila (Table 11). (The NCSO does not prepare a rural CPI). Real wage rates in Palay declined during , increased during and declined again in 1979 and 1980, with little change in 1981 and Real wages in the corn sector declined marginally between 1970 and 1974, increased during and then declined until The increase and the subsequent decline are broadly true of wages in sugarcane and coconut sectors also. In general, the real wage rates in the early 80s were lower than in the early and especially mid-7us. (d) A Tentative Explanation of the Pattern of Real Wage Movements: The decline in rural and urban real wages during , increase in rural areas and constancy in urban areas during and the decline Aft-r that (up until3j?82/83) needs some explanatory hypothesis if not a precise explanation., Before attempting an explanation, however, some of the limitations of the data need to be mentioned. First, problems and errors may be present in the series on nominal wages especially since the Government agencies collecting the data as well as the reporting establishments are 4qrecious of Government minimum wage guidelines issued from time to time. 4 U1 There is, however, no simple way of evaluating the accuracy of the wage da:s (except by commissioning an independent wages survey). Second, the index number problem can vitiate the real wages significantly. Assuming that data limitations have 38/ These trends were shared by a number of occupational groups (skill categories). See Table 32, Annex 7. 39/ It is useful to note that the situation to be explained is not one of a simple declining trend. 40/ Fortunately no statistical relationship has been found between the legislated and the reported wages series (see Table 34, Annex 7).

41 -28- Table 10: INqDEX OF AVERAGE DAILY BASIC WAGE RATES OP LABORERS IN INDUSTRIAL ESTABLISHMENTS IN METRO MANILA, ( ) Nominal Real Period SkLlled Unskilled Skilled Unskilled Source: Central Bank.

42 -29- Table 11: DAILY WAGE RATES OF LABORERS IN AGRICULTURAL OPERATIONS, Consumer price Daily wage rates (pesos/day) index outside Palay Corn Sugarcane Coconut Manila Nominal Real Nominal Real Nominal Real Nominal Real 1972 = Source: BAECON.

43 not affected the noted real wage patterns, they can be explained by the trends in labor supply, labor demand, and the overall macro-economic environment. Throughout , labor supply (in an ex ante or potential sense) was expanding rapidly (ex post average annual growth rate was 3.6%). In such circumstances, relatively sluggish demand for labor tends to pressure money wages down, though money wages may not in fact decline for reasons such as setting wages equal to or slightly higher than previous levels and the probable practice of adhering to the (increasing) trend indicated by the legislated minimum wage levels. Under this scenario, whether or not real wages increase will depend on the inflation rate, determined largely outside the labor market. If the demand for labor is buoyant, money wages will be under an upward pressure. Then also the real wage movement depends on the inflation rate except that a macroeconomic environment conducive to stimulating labor demand is more likely to go hand in hand with a relatively modest rate of inflation. During and , rural employment expanded at high rates (4.7% and 5.1% respectively), but real wages increased only for a short while during Relatively high rates of agricultural GDP growth accrued during This period also saw a stable overall GDP growth rate, relatively high investment growth rates, the favorable effects of the earlier commodity boom, and special rural infrastructure programs. The incidence of underemployment was low during 1977 and 1978, an indication of some tightening of the labor market aided by rural to urban migration. Finally, during inflation rates were relatively low. The economic environment deteriorated in subsequent years, slack developed in labor demand, and inflation started to erode the real value of tnie nominal wage rates. The situation was exacerbated by pressures on rural labor supply due to movement of labor from urban to rural areas. In the urban areas, employment expanded at 4.9% during and 2.8% during The healthy overall economic growth environment of the first period and part of the second ( ) was counterbalanced by ruralurban migration and excess labor supply (indicated by the 70Z to 90% service sector share in employment growth) and the result was only a modest increase in the money wage. The relatively high inflation rate during contributed to the real wage decline and the low inflation rate of the period helped to keep the real wage rate constant. The deterioration in the demand for labor and the high inflation rates of the post-1978 period led to declining real wages. Sustained growth and modest inflation over relatively long periods of time ensure the best conditions for growth in real wages via growth in labor demand (cutting down underemployment), and increases in productivity. The process can be greatly aided by a slower (rather than a faster) labor force expansion which requires action on the population front today to derive benefits some 15 to 20 years hence.

44 (iii) Wages and Poverty Thresholds Trends, : In the previous sections, it has been noted that during , poverty incidence has declined, but, underemployment has increased and real wages have generally declined. In order to "reconcile" (at least partly) the noted trends, data on the poverty thresholds and wage rates are brought together (Table 12) and their implications are noted below. In the rural areas, wage rates are shown for laborers in corn farming where historically, the rates have been lower than in rice, coconut and sugar farming. The corn labor wage rates were lower than the rural poverty threshold wage levels in both 1971 and The situation changed in the subsequent years and the wage rates were higher than the poverty thresholds. In the urban areas (Metro Manila), the wage rates for common laborers (the lowest paid occupational category) were higher than the poverty threshold throughout The situation of falling real wages could thus coexist with a situation of declining poverty incidence because the wage rates were higher than the poverty thresholds. In addition, and most important, the levels of poverty incidence were in fact very low among the wage earner families in the urban areas and nonagricultural labor families in the rural areas. This could be also due to the significance of non-wage incomes for some of the families. However, the data in Table 12 do point out one disturbing trend in both urban and rural areas. It is the rapidly narrowing gap between the wage rate and the poverty threshold in recent years. It is quite likely that this reduction reflects the pressure of the expanding labor force on the job market and the slackening of employment growth in the last couple of years, indicated by the recent increases in underemployment. Prospects, 1984 and Beyond: With the economy registering negative growth in 1984 and inflation running at a very high rate, a worsening of employment, wages and poverty was almost certain. The situation for the period also is not very bright. With labor force expanding by 2.1 million persons in-the three years and economic growth projected at only 3% p.a., unless steps are taken to minimize the adverse effects of the slow growth on the lower income groups, the poverty situation will continue to worsen The Government anticipates a 4.4% annual growth rate in agriculture 4 which if realized will go a long way to prevent poverty escalation in the rural ares. As for the urban areas, with projected growth rates of 2.5% in industry and 2.3% in services, the situation is less comforting. Between , the annual addition to the labor force averages 0.8 million. This increases to a million persons between To address the employment problems of such magnitudes, relatively high rates of economic growth, fairly high rates of investments and economic policies favoring labor-intensive development have to be pursued. 41/ Updated Philippine Development Plan, , NEDA, September 1984, p. 22.

45 Table 12: ILLUSTRATIVE COMPARISON OF POVERTY THRESHOLD EARNING LEVELS AND AVERAGE DAILY WAGE RATES, Monthly Poverty poverty threshold line per daily earning Average daily wage Wage to poverty family (P) level (F) /a rate (P) threshold ratio (Z) Rural Metro Manila laborers common Year Kural Urban Rural Urban in corn/b laborers Rural Urban n.a n.a n.a n.a n.a n.a /a Computed by deflating the monthly poverty line by an assumed number of 26 working days and the average number of employed persons per family. The average number of employed persons was 1.9 and 2.1 respectively in 1976 and 1982 in rural ares and 2.0 and 1.8 in 1976 and 1982 respectively in the urban ars. For the period covered in the above table, the following averages were assumed: Rural: Urban: 1.9 for 1971 and 1975, 2.0 for ; and 2.1 for for 1971 and 1975; 1.9 for ; and 1.8 for /b Wages are about the lowest in corn compared to rice, sugar and coconut.

46 As for the next century, since the new entrants to the labor force are yet to be born, the country can opt for promoting population control seriously and effectively now so as to reduce the intensity of the future labor force explosion. (i) Monitoring Poverty V. RECOMMENDATIONS Monitoring poverty is a serious matter and the Government cannot ignore this responsibility. The Government, however, was in favor of monitoring the progress of the bottom 30%. Since there is no way of identifying a fixed group of families or persons as the bottom 30%, exercises on monitoring their progress will become controversial and complicated. From the standpoint of long-run political and social stability, the most desirable course is for the Government to establish simple poverty lines (such as those suggested in Annex 6), monitor both poverty incidence and the poverty focus in its development programs and projects. To facilitate this, streamling and improving the data collection are recommended. (ii) Data Collection and Dissemination (a) Household Incomes and Expenditures: The practice of holding more than two quarterly ISH should be reviewed since data processing and the publication of data can not be handled expeditiously when more than two quarters are covered in a year (see Annex 1). Collecting household income data in each and every quarterly ISH did not really serve any purpose. Instead, once in every two or three years, data on incomes, expenditures and savings should be obtained in detail by expanding the scope of the ISH for the concerned years. Tabulations should be provided by per capita income and per capita expenditure classes in addition to tabulations by family income or family expenditure classes. The 1981 Census of Agriculture and Fisheries is an important source of information on agricultural and fishery output. The data refer to The previous agricultural census was conducted in The two censuses together provide valuable data not only to assess the changes over a decade in rural levels of living, but also to find the sources of agricultural growth and its distributive benefits. The NCSO should process the 1981 census data quickly and publish important national and regional tables on a comparable basis for 1971 and (b) Labor Force and Employment: The present efforts to strengthen the ISH (see Annex 1) should be continued and fully implemented. Each year labor force and employment data must be collected as at present. However, an ISH should be conducted only once a year or atmost twice a year. Establishment surveys should concentrate on obtaining data on output and value added rather than on employment and earnings. The Ministry of Labor and Employment should have timely and full access to ISH labor force and employment data and should discontinue its household surveys. BAEcon should conduct, once every two years, comprehensive socio-economic surveys, rather than the present rice and

47 corn surveys, to obtain detailed information on the employment and related economic conditions in the rural economy. (c) Wages and Prices: The Ministry of Labor and Employment should collect quarterly data on wage rates in the non-agricultural sectors in key cities. The proposed BAEcon rural surveys would provide rural wage rate data with an expanded coverage. As for prices, the CPI's are produced for the regions and the country, but not by rural and urban classification. The price data routinely collected by NCSO are sufficient to prepare reasonable rural and urban price indices. Such indices should be prepared for the general population as well as for the low income groups such as agricultural laborers in the rural areas and the industrial workers in the urban areas. (iii) Statistical Organization The problems of data collection and dissemination are not limiced to household expenditure, employment and wages. In cases like national accounts, some weaknesses are known to exist. The weighting pattern for consumer price index is outdated. Trade data are said to be off from partner data. It is common knowledge that data on public enterprises are limited. Vital registration is inadequate for estimating birth rates and death rates. The list can be long. The need is for a comprehensive review of the entire statistical system with a view to recommending measures to ensure the timely collection and dissemination of statistical information without duplication of efforts by different agenices. Such a review should go into all relevant aspects including staffing, salaries and budget.

48 REFERENCES 1 Abrera, Ma. A.S., "Philippine Poverty Thresholds" in M. Mangahas (ed)., Measuring Philippine Development, Manila: Development Academy of the Philippines, Bautista, Romeo M., The Development of Labor Intensive Industry in the Philippines, (Unpublished) July Bautista, R., Power J., et al, Industrial Promotion Policies in the Philippines, Manila: Philippine Institute for Development Studies, Berry, A., "Income and Consumption Trends in the Philippines, ", Review of Income and Wealth, September Daly, Rex F. Merlinda D. Ingco and Clemencia H. Padrinao, Income Distribution, Food Consumption and Poverty, Staff Paper No. 7., Agricultural Policy Analysis Staff, Ministry of Agriculture, Quezon City, June Dutta, G., and Meerman, J., Household Income or Housebold Income Per Capita in Welfare Comparisons, Staff Working Paper No. 378, World Bank, ILO., Sharing in Development: A Programe of Employment, Equity and Growth for the Philippines, Geneva, Jain, Shail, Size Distribution of Income: A Compilation of Data, World Bank, Lal, Deepak, Wages and Employment in the Philippines, World Bank, October Lipton, M. Demography and Poverty, Staff Working Paper No. 623, World Bank, Mangahas, Mahar, Quizon, J.B. and Lim, Antonio, A Critique of NCSO 1975 Family Income and Expenditure Survey (Unpublished) 12. Mangahas, Mahar, "Income Inequality in the Philippines: A Decomposition Analysis," in Income Distribution, Employment and Economic Development in Southeast and East Asia, the Japan Economic Research Center, Tokyo, July Mangahas, Mahar, "What happened to the Poor on the Way to the next Development Plan?" Development Academy of the Philippines, Manila, December NEDA, Some Aspects of Rural-Urban Welfare Differential, Manila, August 1984.

49 Oshima, Harry T., "Sector Sources of Philippine Postwar Economic Growth: The overall record in Comparative Perspective" Journal of Philippine Development, First Semester, pp Power, J.P. and Sicat, G.P., The Philippines-Industrialization and Trade Policy, Oxford University Press, Rao, V.V. Bhanoji, Poverty in Indonesia, : Trends, Associated Factors and Research Issues, World Bank, January Republic of the Philippines, Five-Year Philippine Development Plan, , Manila, September Republic of the Philippines, Five Year Philippine Development Plan, , Manila, May Republic of the Philippines, Updated Philippine Development Plan , Manila, September, Sta. Romana III, Leonardo, Income Inequality in the Philippines: A Survey of Evidence, Working Paper, Income and Assets Distribution Research Project, Institute of Economic Research, Hitotsubashi University, Tokyo, September Tan, E.A., and Holazo, V., Measuring Poverty Incidence in a Segmented Market, Discussion Paper No , Quezon City: School of Economics, University of the Philippines, December Terasaki, Yasuhiro, Overviews of Philippine Income Distribution Studies (I): Data Availability, Working Paper, Income and Assets Distribution Research Project, Institute of Economic Research, Hitotsubashi University, Tokyo, May Tidalgo, R., "Labor Absorption in the Philippines, " Philippine Economic Journal Vol. 15, Nos. 1 & 2, Tidalgo, Rosalinda, and Esguerra, Emmanuel F., Philippine Employment in the 1970s, Manila: Philippine Institute of Development Studies, July University of the Philippines, An Analysis of the Philippine Economic Crisis, A workshop Report (by a panel of 10 economics professors), June Visaria, P., Poverty and Living Standards in Asia, LSMS Working Paper No. 2, World Bank, World Bank, The Philippines: Priorities and Prospects for Development, 1976.

50 World Bank, Aspects of poverty in the Philippines: A Review and Assessment, Vol. II: Main Report, Report No PH, December 1, World Bank, World Development Report, World Bank, The Philippines: An Agenda for Adjustment and Growth, Report No PH, November 1984.

51 -38- ANNEX 1 Page 1 THE PHILIPPINES: RECENT TRENDS IN POVERTY, EMPLOYMENT AND WAGES Integrated Survey of Households (i) Introduction In November 1976, the regular quarterly labor force survey conducted by the National Census and Statistics Office (NCSO) was 'integrated' with the semi-annual agricultural survey conducted by the Bureau of Agricultural Economics (BAEcon) to form a single survey called the Integrated Survey of Households (ISH). (The integration was enabled by Presidential Letter of Instruction No. 521 and NEDA Memorandum Circular No ) The ISH was planned with technical assistance from NEDA's Inter-Agency Committee on Survey Design, which advised on sampling, the formulation of concepts, the definition of terms, and the questionnaire design. The ISH is a joint undertaking of various government agencies. The principal agencies involved are the National Economic and Development Authority (NEDA) which is the coordinating agency, the Bureau of Agricultural Economics(BAEcon) and the NSCO, which are the operating agencies. The ISH was envisaged as a continuing quarterly survey of non-institutional households. The ISH was designed to serve the following objectives: (a) to measure the level of employment, unemployment and underemployment in the country; (b) to gather data on employment generation, production and productivity of family-operated agricultural and non-agricultural activities; (c) to gather data on rice and corn production, household rice stock, and inventory of livestock and poultry; and (d) to gather data on both individual and family earnings and income from other sources. (ii) ISH Questionnaire Design The ISH questionnaire is structured into several parts: Part I - Data on Individuasq Part II-A - Data on Own Family-Operated Activities Part II-B - Data on Family-Sustenance Activities not Considered in Part II-A Part III - Data on Crops, Livestock and Poultry. The data on individuals (Part I of the Questionnaire) covers demographic characteristics (age, sex, marital status, highest grade completed, and usual occupation), and economic characteristics (various items pertaining

52 _q -9 _ANNEX 1 to employment and unemployment). In addition, receipts from dividends, pensions and transfer are also obtained in Part I. Data of own family-operated activities (Part II-A of the questionnaire) refers to information on production and net receipts from various economic activities (fishing, manufacturing, etc.). Similar data are obtained in Part II-B on the family-sustinance activities not considered in Part II-A. Data on crops, livestock and poultry was obtained in the final part of the questionnaire for processing by BAEcon. (iii) Instructions for CoLlecting Employment Data in ISH For a few selected questions, an edited extract of the instructions for obtaining employment data from respondents to the ISH questionnaire is given below. It shows that the data collected (if reliable) can be extremely useful to study labor utilization, intensity of employment, unemployment (labor-time not at work but available for work), and visible underemployment. The survey also allows the employment data to be linked with income Tea-rnings). Did the respondent have a job or business or unpaid work on family farm or business, even for only one hour at any time during the past quarter? ExpLain to the respondent the meaning of work; ask probing questions to ascertain existence of a job or business, or unpaid work on family farm or enterprise during the reference quarter. Work is any activity of a person done for pay, profit or without pay on a family farm or enterprise or an activity done by a farm operator or member of his family on another's farm on 'exchange labor' arrangement. (Work for pay, work for profit, work without pay on family farm or enterprise were defined.) Examples of activities which are not considered as work are: housekeeping in own home; building, repairing or painting own house; volunteer work; activities of students in civic action; and begging or gambling. What was his primary job or business during the past quarter? What other job or business did he have during the past quarter?: Primary job/business is any gainful activity of a person which is permanent and a full-time job, lasting for 1 year or longer or had lasted or was expected to last for one year or longer, regardless of whether he had a job/business or not during the reference quarter. If a person has two permanent jobs consider the one where he worked more hours as his primary job. If, however, these two jobs have equal hours of work, consider as primary job the one where he derives more income. Other jobs/businesses are any gainful activities of a person which may be permanent or temporary, full-time or not, aside from his primary job. No. of Full Days Worked: For the purposes of this survey a full days work is based on a person's normal prescribed working hours per day whether it be 4 hours, 5 hours, 10 hours, as long as he was paid a full day's pay. Normal Working Hours per Day: For the purposes of this survey, number of normal working hours per day refers to the usual or prescribed

53 -40- ANNEX 1 Page 3 working hours of a person for a full day's work in his job/business. (iv) Instructions for Collecting Income Data in ISH Adapting the ISH Manual for Field Workers (revised August 1978), the instructions for obtaining income data from the respondents for a sample of items are given below. Wage/Salary During the Past Quarter: These items must be asked for all persons with entry for primary and secondary jobs except for the following: unpaid family worker on a family-operated activity, and operators or managers of a family-operated activity who do not receive salaries or wages either in cash or in kind from the family business or enterprise. Cash salary or wage includes basic compensation received in cash, cash commissions, tips, bonuses and family and clothing allowances in cash. Salary and Wage in kind includes those compensations received by an employee in the form of goods such as rice, corn, fish or any other form of payment not in money, and rental value of housing quarters provided to an employee. The amount to be entered must be the total wage and salary earned (even if not received during the quarter) before deductions of contributions for insurance, retirement and medicare premiums, withholding tax and other contributions reflected in the payroll. Take note that total wage and salary include basic pay plus commissions, tips and allowances as defined earlier. However, salesmen and brokers who worked on commission basis should report only the net amount received after deducting all expenses incidental to their jobs as salesmen or brokers. Reteipts from non-economic activity during the past quarter: This particular secuion aims to collect data on receipts from non-economic activity as well as earnings of persons less than 10 years old, either in cash or in kind during the reference quarter. These receipts are dividends, interest and roylaties and profits for sale of stocks and bonds; rentals, including landowner's share of agricultural products and rental value of owner-occupied house; and pension, retirement benefits, support, gifts, winnings, proceeds from insurance, etc. Data on Own Family-Operated Activities: There are five blocks comprising Part Il-A of the Questionnaire. These are: (A) fishing; (B) manufacturing and processing; (C) service-type activities and construction; (D) forestry and logging, hunting and trapping, mining and quarrying; and (E) farming, gardening and livestock and poultry raising. All five blocks ask for information on production, cost of operation, net receipts from the particular activity, and labor inputs of family and non-family members during the reference quarter. Information on the extent of production and disposal of produce are also gathered. The instructions issued to the enumerators are illustrated below in respect of fishing activity. Block A - Fishing: There will be entries in Block A - Fishing, if any member of the family did some fishing or gathered live shells and seaweeds for sale or for consumption, on own account during the past quarter. This includes deep-sea and off-shore fishing and fishpond/pens and seaweed farm etc. Fishing enterprise managed and operated by cooperatives and corporations

54 -41 - ANNEX 1 Page 4 are not considered family operated activities, and therefore they are to be excluded. "Managers, "encargados, "caretakers" and other workers having direct supervision in a fishing activity who are paid in cash and/or in kind as employees should not have reports in this block. A lessor or one who rents out fishpond, fishing boat or other fishing equipment as a regular enterprise, is not considered an operator of a fishing activity but rather operator of a rental business. The activity will be reported in Block C (Service type Activities and Construction.) Total catch, harvest or gathering (in kg): Enter the total catch, harvest or total quantity gathered by type of fishing used during the reference quarter in kilograms. In the case of fingerlings and bangus or shrimp fry, report the number gathered. The report for the total catch, harvest or gathering must include the quantity consumed by the household, those sold or processed for sale, those given away or paid for debt, and those given as share to workers. (Conversion to Kilograms: If the report on total catch is given in units like "tiklis"(basket),"banyera"(tub),"kahon"(box), etc., ask the average weight in kilos that the unit contains. Multiply the weight given by the number of units or containers to obtain the figures in kilograms. When a respondent knows the total value of catch but he does not have any idea about the weight of the catch, simply divide the total value of catch by the average selling price per kilogram during the quarter, to estimate the weight in kilograms.) Total value of catch, harvests, or gathering (Pesos): Report the total value of catch, harvest or gathering for each type of fishing activity in pesos. "Value" includes the total amount of money received by the family from the sale of its catch, harvest or produce plus the imputed cost of the catch, harvest or produce, which was consumed by the family, paid for debt, given as share to workers, processed for sale and/or given away free. The imputed value should be based on the market price prevailing in the locality at the time of the catch, harvest or produce. Net receipts in cash and in kind: Total net receipts refer to excess of the total value of catch during the reference quarter over the cost of operation. To compute the total net receipts in cash or in kind, deduct all the expenses incurred in cash and the imputed value of expenses in kind from the total value of the catch. If operator had non-family member business partners, operator's share of net receipts (percent): Ask the respondent if the operator of the fishing activity had a business partner who is not a family member. If so, inquire about the percentage share of the operator from the net receipts both in cash and in the form of goods. It should be stressed here that if the operator had a partner who is not a family member, the report on family production, distribution of produce and cost of operation shall be those of the whole enterprise and not the percentage of the operator alone. Data on Family-Sustenance Activities: Part II-B deals with household activities not considered as family-operated activities. These are also sources of income which are not included in Part II-A. Some examples of these activities which could augment the family income are: raising one or two