Anomaly of the fertility decline in India's Kerala state : a field investigation

PHN Technical Notes RES 2 ANOMALY OF THE FERTILITY DECLINE IN KERALA *k. C. Zachariah July 1982 * This paper is one of a series issued by the Population, * * Health and Nutrition Department for the information and * * guidance of Bank staff working in these sectors. The * * views and opinions expressed in this paper do not neces- * * sarily reflect those of the Bank. * RES 2 * ABSTRACT Kerala's fertility has declined by more than 35% during the past 10 years. This was rather unexpected as the population of Kerala is still very poor with a per capita income less than $200. How did fertility decline in Kerala in such an unfavorable circumstance? A resolution of this anomaly is the object of this paper. The paper concludes that fertility decline in Kerala was caused by an optimum sequence of the commonly recognized determinants of fertility trends. The decline began with improvements in health and education, but it gathered momentum as a result of subsequent policy interventions--an official family planning program, land reforms, wage reforms, and other redistributive policies. TABLE OF CONTENTS Page List of Maps...................................................,,.... i List of Figures...................................................... -Ii List of Tables......... ...............**** *** ***................. iv Summary and Conclusions............................................... Part I: Background Chapter I: The Anomaly........................................... 1 Chapter II: Kerala State: Geographic and Socio-Economic Background................... ...................... 5 Chapter III: Kerala State: Demographic and Family Planning Background................................. 22 Part II: Fertility Survey and Results Chapter IV: World Bank Fertility Srvey........................... 54 Chapter V: The Fertility Trend................................... 67 Chapter VI: Ideal Family Size..................................... 116 Chapter VII: Marriage as a Factor in the Fertility Trend........... 151 Chapter VIII: Family Planning as a Factor in the Fertility Trend............................................... 168 Part III: Interpretations and Conclusions Chapter IX: Determinants of Fertility Decline: Conceptual Framework........................................... 204 , Chapter X: Determinants of the Fertility Decline: Empirical Evidence.................................. 220 * Chapter XI: Lessons from the Kerala Study......................... 269 Annex I: Statistical Tables........................................... 276 Annex II: "Trends and Determinants of Infant and Child Mortality in Kerala" by K.C. Zachariah and Sulekha Patel********************************************.................. 320. Bibliography****************a************************................ 355 - ii - List of Maps Page Chapter I: Map 1. India: Kerala, Fertility Survey Districts (IBRD 15780)............................... xv Chapter II: Map 2. Kerala: Distribution of Hospitals, by Taluk............ 12 Chapter III: Map 3. Kerala: Density of Population, 1981................... 25 Chapter IV: Map 4. India: Kerala Fertility Survey, Alleppey District (IBRD 15930)............................... 56 Map 5. India:.Kerala Fertility Survey, Ernakulam District (IBRD 15931).............. ................ 57 Map 6. India: Kerala Fertility Survey, Palghat District (IBRD 15929)............................... 58 List of Figures Page Figure 3.1 Family Welfare Program, Organizational Chart............ 38 Figure 5.1 Crude Birth Rate of the Three District of Kerala, 1960-79................................... 71 Figure 5.2 Fertility Patterns in Kerala and Selected Asian Countries............................. 77 Figure 5.3 Period-Specific Fertility Rates, Three Districts of Kerala, 1965-70, 1970-75 and 1975-805-8................... *.......o......... 80 Figurr, 5.4 Trend in Specific Fertility Rates, 1945-50 to 1975-80, Three Districts of Kerala................ 81 Figure 5.5 Cumulative Cohort (Birth) Fertility Rates, 1930-35 to 1960-65, Kerala........................... 84 Figure 5.6 Percent Excess Fertility of Sterilized Women over Non-Sterilized Women, Kerala*****************************5**............... 105 Figure 6.1 Percent Distribution of Women by Desired Number of Children by Distiict, Kerala.............. 118 Figure 6.2 Percent Distribution of Women by Desired Number of Children and Children Ever- Born, for the Three Districts, Kerala................ 126 Figure 6.3 Distribution of Women by Difference Between Parity and Desired Number of Children by Parity, Kerala.................................... 129 Figure 6.4 Percent of Women with Excess Fertility by Parity, Three Districts, Kerala...................... 130 Figure 6.5 Distribution of Women by Difference between Desired and Actual Family Size in Each Age Group, Kerala.................................... 132 Figure 6.6 Percent Distribution of Women by Age at Marriage and Fertility Status, Three Districts, Kerala***********. *************4......... 134 Figure 6.7 Excess Fertility Women by Family Planning Practice by District, Kerela......................... 145 Figure 7.1 Specific Fertility Rates, 1965-70 and 1975-80, and Fertility Decline Due to Marriage and Fertility Factors..............*****.*** .......... 167 - iv - List of. Figures (cont.) Page Figure 8.1 Partial Regression Between Years of Schooling and Birth Control Practice, Kerala, 1980............. 197 Figure 8.2 Distribution of Women by Marital Status and Family Planning Status, Kerala 1980.................. 20 Figure 8.3 Potential Fertility Reduction Due to Celibacy, Sterilization, and Family Planning Use, Kerala, 1980 ......................................... 203 Figur- 9.1 Impact of Social Reforms on Fertility in Kerala............................................... 208 Figure 9.2 Impact of Land Reforms, Agrarian, and other Reforms on Fertility Trends in Kerala................ 212 Figure 10.1 Recursive Model of the Relationship Between Education and Fertility, Kerala...................... 227 Figure 10.2 Birth Interval Following an Infant Death by Duration of Survival (months), Kerala................ 232 Figure-10.3 Sterilization and Family Planning Practice by Age, Kerala and Sri Lanka............................ 265 Figure 10.4 Sterilization and Family Planning Practice by Education, Kerala and Sri Lanka..................... 266 -v - List of Tables Page Table 2.1 Distance of Physical Separation to be Maintained Between Castes in Earlier Times............. 9 Table 2.2 Health Facilities in Kerala and India, 1979-80............ 13 Table 2.3 Percent Literate in Kerala (1941-81) and India (1901-81).. 14 Table 2.4 Educational Facilities in Kerala and India, 1979-80......................................... 16 Table 3.1 Population Growth in Kerala and India, 1901-81............ 23 Table 3.2 Population Size, Growth Rate and Density by District, Kerala State................................. 24 Table 3.3 Towns with a Population of 100,000 or more in 1981................................................ 26 Table 3.4 Distribution of Villages by Population Size, Kerala State and India, 1971........................... 27 Table 3.5 Crude Birth Rate in Kerala and India, 1931-78............. 28 Table 3.6 Age-Specific Fertility Rates in Kerala, 1958, 1971, and 1976....................................... 29 Table 3.7 Crude Death Rate, Kerala and India, 1931-78.............. 30 Table 3.8 Infant Mortality Rates for Kerala and India, 1951-79................................................ 31 Table 3.9. Budgeted Medical and Paramedical Staff, Family Welfare Program, June 30, 1978......................... 40 Table 3.10 Expenditures on Family Welfare, Kerala, 1952-80........... 42 Table 3.11 Rates of Incentives Payments in 1966-67 (Rs.)............. 43 Table 3.12 Rates of Incentive Payments in 1977 (Rs.)................. 43 Table 3.13 Achievement Record of Vasectomy, Tubectomy, IUD, and C.C. Users, Kerala, 1957-80........................ 48 Table 3.14 Number of Births Averted Due to Use of the Various Methods, Kerala, 1957-80....................... 49 Table 3.15 Mid-Year Population, Births Averted, and Birth Rate, Kerala, 1961-80.................................. 50 - vi - List of Tables (cont.) * Page Table 3.16 Couples Protected by Various Methods of the Family Welfare Program, Kerala, 1957-80................ 51 Table 3.17 Number of MTP Acceptors in Kerala, 1972-80................ 52 Table 4.1 Number of Households Covered by the Survey and Number of Respondents in Each Category Selected a nd Interviewedtw............................... 61 Table 5.1 Percent of Population in the Younger Age Groups by District, Kerala............................. 67 Table 5.2 Crude Birth Rate, Three Districts in Kerala, 1965-80........................................ 68 Table 5.3 Comparison of Survey and SRS Birth Rates, 1975-80......... 70 Table 5.4 Crude Birth Rate, Kerala and Selected Countries in Asia......................................... ....... 70 Table 5.5 Child Woman Ratio: Three Districts of Kerala............... 72 Table 5.6 Average Number of Children Ever-Born (Parity) by Age of Woman, Kerala................................ 73 Table 5.7 Average Number of Children Ever-Born by Duration of Marriage, Kerala........................... 74 Table 5.8 Specific Fertility Rates for Recent Periods, Three Districts Combined, Kerala....................... 75 Table 5.9 Age-Specific Marital Fertility Rates, Three Districts Combined, Kerala............................. 76 Table 5.10 Age and Period-Specific Fertility Rates for Five-Year Intervals from 1945 to 1980, by Five-Year Age Groups, Kerala........................... 82 Table 5.11 Marital Fertility Rates for Palghat, Ernakulam, Alleppey............................................... 86 Table 5.12 Mean Number of Children Ever-Born by Duration of and Age at Marriage, Kerala..................... 88 Table 5.13 Mean Number of Children Ever-Born by Years of Schooling of Mother and Duration of Marriage, Kerala....................................... 90 - vii - List of Tables (cont.) Page Table 5.14 Distribution of Married Women by Age at Marriage and Educational Attainment, and Standardized Fertility Rates, Kerala................................ 91 Table 5.15 Total Marital Fertility by Education, Krala............. 91 Table 5.16 Marital Fertility Rates by Age and Educational Attainment, Kerala..................................... 92 Table 5.17 Mean Number of Children Ever-Born by Duration of Marriage and Caste, Kerala.......................... 94 Table 5.18 Marital Fertility Rates by Age and Caste, 1965-70 to 1975-80, Kerala............................. 96 Table 5.19 Mean Number of Children Ever-Born by Duration of Marriage and Per Capita Household Expenditures, Kerala................................... 97 Table 5.20 Percent Decline in Fertility Rate by Per Capita Household Expenditure, Kerala................... 98 Table 5.21 Marital Fertility Rate by Age and Monthly Per Capita Household Expenditures, Kerala............. 99 Table 5.22 Mean Number of Children Ever-Born by Land Owned and Duratioa of Marriage, Kerala................. 100 Table 5.23 Marital Fertility Rates by Ownership of Land, Kerala................................................. 101 Table 5.24 Marital Fertility Rates by Age and Family Planning Status, 1965-70 to 1975-80, Kerala............ 103 Table 5.25 Marital Fertility Rates by Sterilization Status, Kerala......................................... 104 Table 5.26 Marital Fertility Rates by Sterilization Status Broad Age Groups, Kerala.......... .................. 106 Table 5.27 Summary Results of Regression of Fertility Measures on Demographic and Socio-Economic Variables.............................................. 109 Table 6.1 Distribution of Women by Desired Number of Children, Kerala.................................... 117 Table 6.2 Desired Number of Children by Actual Number of Children Ever-Born (Parity), by District, Kerala....................................... 120 v viii- Li!-t: of* Tables (cont.) Page Table 6.3 Average Desired Number of Children by Age, by District, Kerala.................................... 121 Table 6.4 Desired Number of Children by Selected Socio- Economic Characteristics, Kerala........................ 123 Table 6.5 Distribution of Women by Fertility Status, by District, Kerala.................................... 128 Table 6.6 Proportion of Excess Fertility Women by Socio-Economic Groups, Kerala.......................... 135 Table 6.7 Distribution of Women by their Number of Undesired Children, Kerala............................ 137 Table 6.8 Distribution of Undesired Children (Gross) by the Parity of their Mothers, Kerala................. 138 Table 6.9 Ever-Married Women by Parity or Surviving Parity and Excess Fertility Status, Kerala.................... 139 Table 6.10 Ever-Married Women by Undesired Children, Kerala.......... 140 Table 6.11 Distribution of Ever-Married Women by Fertility Status and Family Planning Use, Keraa................. 143 Table 6.12 Percent of Non-Users of Family Planning by Fertility Status, Age and Parity, Kerala............... 146 Table 6.13 Percent of Women Who are Sterilized or Use Conventional Family Planning Methods by Education and Excess Fertility Status, Kerala.......... 147 Table 6.14 Percent of Women Who are Sterilized or Use Conventional Family Planning Methods by Ownership of Land and Excess Fertility Status, Kerala............................I............ 148 Table 7.1 Distribution of Women 15-49 Years by Marital Status, Survey 1980 and Census 1971................... 151 Table 7.2 Distribution of Ever-Married Women by Marital Status, 1971 and 1980, Kerala......................... 152 Table 7.3 Percentage of Never-Married and Currently Married Women by Age, 1971 and 1980, Kerala................................................ 153 Table 7.4 Percent Widowed or Divorced by Age, 1971 and 1980, Kerala.......................................... 154 - ix - List of Tables (cont.) Page Table 7.5 Percent Widowed or Divorced by Socio- Economic Groups, Kerala................................ 155 Table 7.6 Median Age at Marriage by District, 1961, 1971, and 1980................................... 156 Table 7.7 Mean Age at Marriage by Marriage Dates and District, Kerala................................... 156 Table 7.8 Average Age at Marriage of Females by Age, Kerala................................................. 158 Table 7.9 Average Age at Marriage of Females by Education, Ke rala,..................................... 158 Table 7.10 Average Age at Marriage of Females by Caste, Kerala.......................................... 159 Table 7.11 Average Age at Marriage of Females by Per Capita Household Expenditures, Kerala.................. 159 Table 7.12 Average Age at Marriage of Females by Land Owned, Kerala................. .................. 160 Table 7.13 Summary of Regressions of Age at Marriage on Selected Socio-Economic Variables, Kerala........... 162 Table 7.14 Percent Married by Age, 1965-80, Kerala................... 163 Table 7.15 Marital Fertility Rates by Age 1965-70 to 1975-80, Kerala............... ........................ 164 Table 7.16 Analysis of Fertility Change due to the Marriage Factor 1965-70 to 1975-80, Kerala............. 166 Table 8.1 Percent of Women Who Know One or More Birth Control Methods, Selected Countries....... ........ 169 Table 8.2 Percent Distribution of Women by Number of Family Planning Methods Known, by District, Kerala..................................... 170 Table 8.3 Knowledge of Family Planning Methods by Demographic and Socio-Economic Characteristics of Women, Kerala...................................... 172 Table 8.4 Percentage of Ever-Married Women Using Conventional Family Planning Methods by District, Kerala................................... 174 -x - List of Tables (cont.) Page Table 8.5 Percentage of Users of Family Planning Methods by Age, Kerala.... ........................... 175 Table 8.6 Percentage of Users of Family Planning Methods by Parity, Kerala............................. 176 Table 8.7 Distribution of Ever-Married Women by Use of Family Planning Methods and Age at Marriage, Kerala................................ 177 Table 8-.8 Percentage of Users of Family Planning Methods by Educational Attainment of Women, Kerala..................... .................... 178 Table 8.9 Percentage of Users of Family Planning Methods by Per Capita Family Expenditures, Kerala ................................................ 178 Table 8.10 Percentage of Users of Family Planning Methods by Caste, Kerala............................... 179 Table 8.11 Percentage of Ever-Married Women Sterilized by District............................................ 180 Table 8.12 Percentage of Exposed Women Sterilized, Selected Countries..................................... 181 Table 8.13 Percentage of Women Sterilized by Age, Kerala............. 182 Table 8.14 Proportion of Women (or Husbands) Sterilized by Parity, Kerala............. ........................ 183 Table 8.15 Proportion of Women Sterilized by Education, Kerala................................................. 183 Table 8.16 Distribution of Ever-Married Women by Use of Family Planning Methods and Education, Kerala.......... 184 Table 8.17 Distribution of Ever-Married Women by Use of Family Planning Methods and Per Capita Household Expenditures, Kerala......................... 186 Table 8.18 Distribution of Ever-Married Women by Use of Family Planning Methods and Land Ownership, Kerala................................................. 187 Table 8.19 Distribution of Ever-Married Women by Use of Family Planning Methods and Age, Kerala................ 188 -xi- List of Tables (cont.) Page Table 8.20 Distribution of Ever-Married Women by Use of Family Planning Methods and Parity, Kerala........... 189 Table 8.21 Distribution of Ever-Married Women by Use of Family Planning Methods and Caste, Kerala............ 190 Table 8.22 Summary of Regression Analysis on Family Planning Variables, Kerala........................... 199 Table 10.1 Desired Family Size and Age at Marriage by Educational Attainment, Kerala....................... 222 Table 10.2 Knowledge and Practice of Family Planning Methods by Educational Attainment of Women, Kerala........................................ 223 Table 10.3 Percent of Women Who are Sterilized or Who Use a Conventional Family Planning Method by Educational Attainment and Fertility Status, Kerala.................... .................. 224 Table 10.4 Fertility Rates by Educational Attainment, Kerala........******************* .................... 225 Table 10.5 Partial Regression Coefficients of Selected Fertility-Related Variables on Years of Schooling (Wife), Kerala............................. 226 Table 10.6 Birth Rates and Death Rates in Kerala, by Regions, Variable Years, 1956-75..................... 229 Table 10.7 Average Birth Intervals Classified by Mortality Experience of Initial Birth, Kerala.................. 231 Table 10.8 Partial Regression Coefficients of Birth Interval (i and i+1) on Mortality Status of the Initial Birth (ith birth) and Socio-Economic and Demographic Variables, Kerala........................ 233 Table 10.9 Average Birth Interval Following a Twin Birth, Kerala........................................ 234 Table 10.10 Parity Progression Ratio by Birth Order, Kerala......... 236 Table 10.11 Additional Fertility (Partial Regression Coefficient) of Women with One Death Among the First N Children over Women with No Deaths, fontrolled for Age, Subsequent Deaths Among Children, and Socio-Economic Variables, Kerala..................... 236 - xii - List of-Tables (cont.) Page Table 10.12 Fertility Effect of Amenorrhea by Parity, Kerala........ 239 Table 10.13 Biological and Behavioral Effects of Child Death on Fertility by Parity, Kerala................. 241 Table 10.14 To be added............................................. 244 Table 10.15 Fertility Differentials by Land Holdings, Kerala........ 245 Table 10.16 Partial Regression Coefficient Between Land Holdings and Fertility Related Variables, Kerala............................................... 247 Table 10.17 Fertility Trend Among Women in Households with 10 Cents of Land or Less by Sterilization Status, Kerala....................................... 249 Table 10.18 Expectations about Education of Boys and Girls by Age and Education of Father and Mother, Kerala............................................... 251 Table 10.19 Distribution of Women by Reasons for Children's Possible Non-Achievement of Desired Educational Attainment, Kerala....................... 253 Table 10.20 To be added........................................... 254 Table 10.21 Children-s Help Around the House, Kerala................ 256 Table 10.22 Expectations of Future Economic Help from Children, Kerala..................................... 256 Table 10.23 Work Status of Unmarried Children 12 Years or Over by Age of Mother, Kerala..................... 257 Table 10.24 Expectations of Old Age Financial Support from Children, Kerala..................................... 257 Table 10.25 To be added..................... .................... 258 Table 10.26 To be added............................................. 260 Table 10.27 Measures of Fertility Decline and Births Averted by District and Time Period, 1965-80................. 263 Table 10.28 Estimate of Excess Sterilization Rate in Kerala......... 267 - xiii - ANOMALY OF THE FERTILITY DECLINE IN KERALA by K. C. Za.chariah IBRD 15780 To Mangalorg 76 770 INEDIA KERALA FERTILITY SURVEY DISTRICTS To Mer'aro Ksaragod -121 Hosdrug 1..( To Viraiondrop.t NNANORE R..NATAKA ..ToMy.r Coanonor. 0 Mommiantoddy •...'' TeIlichery 0 4oB Sulta Cadagra uto Prmr .M f'A 110n To Ooimcomund Tomo :Gudalur KZH1K DEf- . ilmbnu Wandur CcU®1 R Rm NADU ° To Coimbotor. S Mlappuram Tirur k-o,rA To Cobm.r Ponnoi 0 Ch-rhu uthy Survey Districts RoadsKunmlmÍ an lm -+--4--f Rai|ways o Cities ond Towns To Ud-molpet PeringalkuNh o District Capitals Ojr District Boundaries -. - State Boundaries rDUKK1 -- -- International Boundaries o Dvco-~..m 100 109 Chathlra... rm Bo.(gciry Errn e '-fLpuzha 0 Kalmkaruy . MThodupuzha OVazhathope. 0 Kulamav T M.d.ra K T TAY M M°°'"' Kottsyam 0 hka peradmb Th,,,O.p,.Ort.Wnid5~k.~.~I il~ L ... Chan erry E opoloyam Thsoa a oa rar d h o r Ban 050a th ,Wly .a f or ecovmnc ep ri Ar0 Ad.r j e C H I N A Q U I L P""lo, S Mwt.h PAKKSTAN HUAiTood Ar-no C Qil.n N Vm..UM( N D l A^ BANGL.ADESH N.yyar Trivpdrum Co 0 25 50 75 100N.ytik. I i | KILOMETERS Padmanbhapura cap. Comorin LANKA 760 770 JUNE 1981 Part I: BACKGROUND CHAPTER I THE ANOMALY Introduction It is widely believed that during the past 15 to 20 years, the birth rate in Kerala has fallen very sharply. The evidences are not unequivocal, but diverse and mutually supportive. The Sample Registration System (SRS) indicated that between 1966 and 1978 the birth rate in rural Kerala declined from 37 to 26 (per 1000 population) (Government of India, Vital Statistics Division, 1966-78). The 1981 census showed that the State's population increase fell from 26 percent in 1961-71 to 19 percent in 1971-81, or a drop of 0.6 in the annual percentage growth rate (Government of India, Registrar General and Census Commissioner, 1981 ,Paper 1, p. 53). The number of scholars in Standard I increased by 2.7 percent per year between 1960/61 and 1970/71, but decreased by 2.2 percent per year between 1970/71 and 1979/80 (Government of Kerala, Statistics for Planning, 1980, p. 252). As a corollary, between 1964 and 1976, more than a million married individuals were sterilized and about 1.4 million births were averted by the official family planning program (Government of Kerala, 1978, Table 13). Fertility declines of such magnitudes are not unprecedented in other societies. In fact, more rapid fertility declines have been observed in recent years in other Asian populations. Singapore, Hong Kong, Taiwan, and South Korea are well known examples in this region. Fertility declines in -2- these countries have been associated with high average national incomes per capita, rapid economic growth, high proportions of the .labor force in non- agricultural occupations, high female literacy and employment, high nutritional levels, low mortality, high national expenditures for family planning programs, and other measures indicating cultural change. Kerala does not conform to this socio-economic development pattern. The conditions in Kerala are not generally thought to favor a rapid fertility decline. It is one of the poorer states of India, with a per capita income less than that of India. Kerala compares very unfavorably with the Asian countries mentioned above--having a per capita income less than $200 compared with $3,300 in Singapore, $3,000 in Hong Kong, $1,400 in Taiwan,and $1,200 in South Korea. In 1971, about 55 percent of the labor force in Kerala were employed in the agricultural sector compared with 2 percent in Singapore, 3 percent in Hong Kong, 37 percent in Taiwan, and 41 percent in South Korea. Regarding nutrition, the rank of Kerala among the Indian states is uncertain. According to the Indian National Sample Survey (NNS), the per capita caloric consumption in Kerala was only 1,571 in 1973-74, or 70 percent of the requirement, two-thirds of the average consumption in India, 50 percent that of Singapore, 53 percent that of Hong Kong, and 55 percent that of Taiwan and South Korea (see Panikar, 1980; and World Bank, 1980, Table 22). With respect to female literacy and total average life expectancy, Kerala compares very favorably with other parts of India. Female literacy was 65 percent in Kerala as opposed to 25 percent in India in 1981; life expectancy was 65 years in Kerala compared with 54 years in India in 1978. -3- The South Asian "success" countries and Kerala are fairly similar in average female literacy and total life expectancy.1/ Kerala spends comparably less per capita on family planning than the success" countries, however (Goverment of Kerala, 1978, Appendix E, Table 1). This per capita expenditure was less than 9 US cents per year. Singapore, by comparison, spent 30 cents; Hong Kong, 10 cents; Taiwan, 11 cents: South Korea, 16 cents; and Malaysia, 20 cents (Nortmann and Hofstatter, 1978, Table 9). Thus, many conditions in Kerala are different from those in other countries which have experienced rapid declines in fertility, and it does not conform to the stereotype of other success stories in South Asia. If high per capita income and rapid economic growth are necessary conditions for a rapid fertility decrease, then Kerala's fertility would not have declined so rapidly. If urbanization and rapid industrialization are essential prerequisites, then Kerala would have to wait more years to experience a rapid fertility decline. If a high level nutritional intake is necessary, Kerala would not be a place where a rapid fertility decline is seen. Yet, fertility has indeed fallen in Kerala. This is the anomaly of fertility decline in Kerala. A resolution of the anomaly is the objective of this study. What are the causes of the fertility decline in Kerala? The question is very lucidly put by John Ratcliffe in a 1978 article in the International Journal of Health Services: / Expectation of life at birth in 1978 was 63 in Korea, 72 in Taiwan and Hong Kong, and 70 in Singapore (World Bank, 1980, Table 21). Female literacy among those aged 6 years and over reached 84 percent in Korea in 1970, 63 percent in Hong Kong in 1971, 81 percent in Taiwan in 1975, and 67 percent in Singapore in 1976 (Nortmann and Hofstatter, 1978, Table 3). -4- "Row did such a large population with levels of income and nutrition among the lowest in the world achieve levels of fertility and mortality substantially lower than surrounding populations with similar cultural backgrounds but higher incomes and caloric intakes? The answer to this question holds profound implications for all nations concerned with population, health, and social and economic development" (p. 124). Available data on fertility trends and characteristics for Kerala are not sufficient to answer questions on the determinants of fertility. In this study, therefore, special data have been collected through a sample survey. The design of the survey and its results are given in Part II of the report. Part III gives the interpretations and analysis of the data, and the principal conclusions drawn from them about the determinants of the fertility decline in Kerala. The remaining chapters of Part I give the socio-economic background of Kerala's population and a description of the demographic situation in the state based on available data. -5- Part I: BACKGROUND CHAPTER IT KERALA STATE: GEOGRAPHIC AND SOCIO-ECONOMIC BACKGROUND Introduction The fertility level and trends of a population are closely related to its physical environment and socio-economic milieu. This chapter, therefore, attempts to give a brief description of the physical characteristics of Kerala State, and its principal socio-economic developments which are relevant to a study of the determinants of fertility. Formation of the State Kerala State was formed in 1956 out of three separate administrative units: Travancore State in the south, Cochin State in the middle, and the Malabar region of the Madras Province on the north. Travancore and Cochin were princely states ruled by local Maharajas, while Malabar was part of British India. Tranvancore and Cochin together account for about 60 percent of the population of Kerala and 55 percent of its area. Trivandrum at the southern end of the state is the political capital of Kerala; it used to be the capital of Travancore State. Geographic Characteristics Kerala is a relatively small state on the southwest corner of the Indian subcontinent. It is a narrow strip of land extending from Kasaragod in the north to Trivandrum in the south (about 585 km.) and from the Western Ghats in the east to the Arabian Sea on the west (about 120 km. at the widest point). With an average rainfall of about 300 cms., an average temperature of -6- 900F (ranging from 70oF to 95o), and with 40 and odd rivers flowing from the Ghats to the sea, the land is fertile and perennially green with paddy fields, coconut groves, banana gardens, tea and rubber estates, tropical forests, and so on. A chain of back-waters connected by man-made canals runs parallel to the ocean. The layout of the Ghats, the course of the rivers, and the location of the lagoons have determined the configuration of the land into three natural regions; namely, the lowland, midland, and highland. Although the land is fertile and water is plentiful, Kerala has been a food-deficit area since the turn of the century. There are two reasons For this. First, the man-land ratio is very small. The cultivable land per person is only 0.21 acres. Secondly, not all cultivable land is used for food production. Although more area could be brought under grain cultivation, this could be done only at the expense of the more profitable commercial cultivation of coconuts, sugar cane, bananas, and other crops. As a result, Kerala is likely to remain a food-deficit state for years to come. The heavy rain and the fast flowing rivers provide ideal conditions for the production of hydro-electric power. Kerala is one of the few states in India where there is no power shortage. In fact, the state generates more hydro-electric power than it can currently consume and the excess is sold to neighboring states. Almost all villages in Kerala are electrified, although the proportion of households which have taken electric connections is relatively small in most villages. Kerala is noted for the cleanliness of its people and land. The abundance of clean water is an important contributory factor. The annual flood brought by the southwest and northeast monsoons flushes out the countryside every year, helping to keep the land and people clean and healthy. -7- The state is endowed with good transportation networks--waterways, rrilways, and roadways. The rivers and lagoons are navigable and provide the means for cheap inland water transport; the roadways (about 90,000 kms. or 230 kms. per 100 sq. km.) connect most villages and towns and are the principal means of intra-state transport; and the railways (886 km.) are the principal inter-state transportation link. The most popular,mode of travel in the state is the government bus service which is extensive and highly subsidized. In 1979/80 the state transport had a route length of 159,000 km. (4 km. per each sq. km. of the state), operated 224 million kms. (9 km. per person per year), and carried 706 million passengers or each person in the state 28 times a 1/ year. - One of the six international airports of India (Trivandrum) and two of the internal airports are located in Kerala. A major seaport of the country (Cochin) is also located in the.state. With the many transport facilities available, it is no wonder that Kerala's people are highly mobile, migrating to all parts of India and many countries outside India. External Influence on Internal Developments From the very beginning of its history, Kerala has had extensive contacts with outside culture, from within India and outside the country. These contacts have played a significant part in the socio-economic development of the state. From the other side of the Ghats came the caste and land tenure systems: from West Asia came Christianity, Islam, and Judaism; the Europeans also brought Christianity, a modern administrative system, plantation industry, modern education, public health, etc. The government-run bus service is the largest bus company in the state. There are, in addition, several privately owned local bus companies. -8- With the coming of the Aryans two or three centuries before the- Christian era came the caste system, a dominant socio-economic and political force even today. The system used to be so rigid that people of different castes could not come closer than a prescribed minimum distance. The following table (2.1) gives this minimum. 2- European contact began towards the end of the 15th century with the arrival of the Portuguese in 1498. The Portuguese were followed by the Dutch ( ), Fr-ench ( ), and later the British. The political connection with Britain ended with the Indian Independence in August 1947. The foreign contacts have left some permanent marks on the population of Kerala. The state has a fairly balanced religious composition with almost equal strength among the Nairs, Izawas, Christians, and Huslims. As a result of this equilibrium, no one caste group can profit at the expense of another. There are no permanent alliances between any caste groups; they shift their allegiances to suit their self interests. The instability of the elected governments in Kerala, the give and take policy of each administration, etc., can be traced to a large extent to this balanced multi- religion-caste composition of the population. The modern formal educational system in Kerala owes its origin to the British rule in India, and the Christian missionaries, who seeing a sizeable Christian population in the state concentrated their work in the fields of education and health instead of proselytizing. Fortunately, the 2/ One saving factor was that in case they came too close, the person who was polluted (upper class) could cleanse himself with a simple bath. Because of this, alot of people in Kerala must have been spending great amounts of time in taking baths! Unintentionally, this caste rule-keeping and resulting baths could have been one of the factors in the low mortality found in Kerala. -9- Table 2.1: Distance of Physical Separation to be Maintained Between Castes in Earlier Times - 10 - local rulers were helpful and by the end of the 19th century every village in Kerala had a primary school. 3/ The European influence on the development of the health system was similar to that on education. They provided the initial impetus; the indigenous tradition of scientific medicine and the wholehearted support of the Maharajas carried the development forward 4/. Special Features of Kerala's Recent Socio-Economic Development In the theoretical framework adopted by this research (see Chapter IX), it is argued that Kerala's fertility decline was caused as much by historical developments as by recent policy interventions. The degree and structure of the fertility decline that took place in Kerala in recent years would not have taken place without both its historical development and the recent policy interventions. Kerala's history prepared the ground whereby a fertility decline became possible and desirable to a majority of the people. The recent policy interventions actually precipitated the decline. What are these policy interventions? This section attempts to describe very briefly some of them -- health, education, political awareness, and land reform. The official family planning program, which is also one of the policy interventions, is described in the next chapter. - In 1817 the Maharani of Travancore announced that the state should defray the entire cost of the education of its people (see Nag, 1981, p. 32). "I take this opportunity earnestly to impress this fact on the minds of all my native subjects and to urge them to seek for themselves, for their children, for their friends and for their servants the great protection of vaccination. They will see the strength of my conviction in the fact that there is no member of my family who has not had this protection conferred at an early age," the Maharajah of Travancore said in 1865-66 at the ceremonial opening of the Trivandrum Civil Hospital (Nagam Aiya, V., 1906, pp. 524-525). - 11 - Health Historically, what most distinguished Kerala from the rest of India was its achievement in the fields of health and education. The extent of the difference in health is indicated by the infant mortality rate which, in 1980, was about 45 in Kerala and 125 in India. The chance of an infant death during the first 12 months of life in India today is about 3 times the corresponding chance in Kerala. Health conditions in Kerala have always been better than those in other parts of India, but recent developments have accentuated the differentials. Kerala has a tradition of medical and health treatment (Ayurvedic system). The physical environment in most parts of Kerala is healthy and has always been so, except for the malaria tracts in the hills (malaria has been eradicated). The rulers of the state have always taken an interest in health matters and committed sufficient public funds for both curative and preventive medicine. In addition, the foreign missionaries established private hospitals and trained health per.sonnel. But, the continued improvement in the health of the people has been tied to their educational achievement (as discussed below) and the health policies through which medical facilities were brought closer to the masses and within their means. Map 2.1 gives the location of hospitals by taluk (administrative division below District). The hospitals are evenly distributed over the state. The average service area of a Primary Health Centre is 232 sq. km. in Kerala (1979) and 563 in India; that of a sub-centre 21 sq. km. in Kerala and 63 in India (Table 2.2). The number of hospital beds per 100,000 population was 458 in urban areas of Kerala (263 in India) and 107 in the rural areas (12 in India). The easy access to medical facilities (due to the relatively small � �Т•° ТУ Т!� j 2•i�p 2.1. кеRА�А . • O1STA18UT1oN OF HO5PITALS, f!У 7ALUK • 'в • f, в в 1(1970/7i) • • • �'в•''l• •• . •, • г • • в 4„ .. • � � � уΡ в• ..•:~ • У : 1 • ... • � • : � ..:,� �. • • ' • . +r • '' • .. � ' ' • • в'' � • 4 •• i 7 5в • ... • • •• -• .. • • . �.. • • r'• в ••'• � в • • • • • • ' • • • . . • . . 7 • •�•+� • ••• ' • • • • ... �•••:� • • в� •в: �• • :•• ,� •в•• • � � �. ••.�: 4 • :• � • ' В ;, • .� • в .• :••. •� • Disстicu апд гaWks в;,в ;•�1• . в в . • . • 1. Gплапгюго mпria г••' ••• :• в i 10 • • в сrв,ров 9 • • • • г. н в Ч в i тЛ.по...пе• �х • ., с.ми�оr. �� • 17 в • s. ты��т r а • • • • ь нопЛ w.п+о в в ч ... ' ..• • • • ',а 11. Коктkоо• dimic М в'•. • •-• • n r. в.иии• • • .•... • 1� • 4 оу,�вмь Т • ... ` • •. sыт w.�..a so ' е �а вим.ое. . �s 12• в •••._ • • в • • , в • 111. Мвlвооиrьт disvw[ ... • • �S : � • в • 11 Еrма' 1l • в7{• •• -' •в I4 • �1 Т.гиг• 1в • • ' • • ' • •'• ' 1w иЛN а • • • • ••• • :�в � а, Мо 'пм: т 1 • • •• wвп Т •`IQ в • • .: в..'1• 1 '••' • • •' �• i7 в г �� • . 1V. Рарп•[dimiст е '.у • � • • • • �• в 15. Опаиам 1ь . �1 •у• в Z7 • 1t 1в�п 9 .• в• • •� в ,Т. •i иг . 6 �Е в 1R GЛ�п°г 1] • .в.•"..'•• • '.• , �.•,- • • V ТГlгhиг dlmlct '�. • ZS • • • • �А ТЛМвроiПV 13 • '. • в в 70. ircn�в 34 . .в • • ii: с :м�вла:. i •24': • :.. • 11 Мивипацигвп 1л • • • • • • • . ' •ц•,.'�'• V1. ЕгrикиlьтдiткТ '•?5 • в • : ••'в у ]< Gоип 17 •.•• • • /о° ]6 1 •:в • i • • ° М. л 9 •.а • • • . / �о° ]Т випwипипм б ' ° •I • .8 • . • Я. 5 ппвппом 1� 2L.в е�.•••.•' • � �.. . �.• _ t в • ]' • • 32 V�t. ldikk�tlmnп • ° r1�'� /'� xt оЛёа�w �.< в ' ' в •° • • :. • ° " в � • • �: ц и�п°йил�.< • 7t• .'г. • и. г..�.л.� s • • •,.: • 75 • • • • I V11i. Ко�пв�увтдiпгiег ♦ ••: � • � . •..'�,' в • • .. '., . й ы.м.т�� �а н• ••76 в•:�• •• У, а,.ввмn 1• • • � •• . п `сЛ..�w�....т а ' _ е . • ' 7i • S7 • 3• Ки��.саив�w 6 - •�' • • • • � • • • .. :�,в . • • ix. дiиоо•rd�m�s , •и• ,• • �2• • .� ; „ . • е •'• • • впо� ,пenw.w. ё • ,"••:..� ": • 4в • ii: i�w"й. � :о � •:i'•.�70.3 • � вц• сыь•.wьi оа. . t 1.3� ! •4< • • ••'. • • ц. 4мЛ.WС'��. . 6 • � •.. . � • : • •�в • • . �. • х. Ош1оп аiпгкг : , �7 в . ."• • . в �• ... I� •Т КиvпвУоовпь ; цв' :' • : w • i° ' С 13 •� в..•. •:. .. � ц, г l• • 49 ц. � в •��' S.: •.: ео. ��и.и. �о � • в57 в ' в • • %1. Tnr•гогит tliwriп • о •'• • •. а • в1 слп.имм • • в в ' • о х г .°�� i� в � я. ин.пимы. :. и 52• ' • s . • �.•�,� • 57• • .•: .• • • � . в,, • • fмвМ 1М� оГ А'О!/Уев ''° • •�0в' •. • • . . бМ.СГ tNr (/�. �I в I •.. , • �.,ь' 'в в farnMвrro/А1�вг • �: •'• ввмм гм 19lв, ••• в !5 • • • .'Рв� � faw/�аиго/RO(iву�/11 ' 1tiг�F. • iиГеГ и!1 1У9/. D,Чгкt тиге•гУ • Т,вц •.• • FPw1в•О•/10/Fп1М1М11П ............ TW1•bouM>.У ' . ватп г.гr 1Ч/. ;�п.�rvЙ,i, а_. . � :. �L.S. ` � • Tb° Ти ТТ° •п:..:аХ-,:..�:. • • лf' . и�:�r,,..,ачгм.«_� , _. . . --- �*��: ��;;;�х.•i�?'.'� Source: United Nations, De artmerдt of Economic and Social Affair�, 1975в ,�.`,,,. �;. ,. - .. - "=• �•'•-• �' ,,.:. Роиег�, Unemplo�rmenC and Development Policy, i�tew Уог�ь: Тд•iJ. �� й:'�-йл}ч�.. 13 Table 2.2: Health Facilities in Kerala and India, 1979/80 Number of Hospital Beds 'Kerala India (a) Urban 21,443 400$343 Rural 21,635 60,543 Total - 3,078 460,886 M Government 24,875 330,196 Private 18,203 130,690 Total 43,078 T6-0,886 Beds Per 100,000 Population Urban 458 263 Rural 107 12 [ealth Centers Primary Health Centres 167 5,499, Sub-Centres: Health 492 28,558 Family Welfare 1,333 20,765 1,992 54,822 Average Service Area Per Health Center (Sq. Km.) Primary Health Centre 232 563 Sub-Centres (either) 21 63 Health 79 108 Family Welfare 29 149 Persons Per Unit of Health Staff Doctor 2,385 2,674 Nurse 2,538 4,774 Miduife 3,068 4,947 Source: Government of India, Central Bureau of Health Intelligence, Directorate General of Health Services, 1981. Pocket Book of Health Statistics of India, 1980 Tables 2, 43, 50, -2-, 53, and 56. - 14 - service area and good transport facilities), the relatively low cost of the services, and the high demand for medical services (due to a higher educational attainment) are all important factors in the comparatively better health conditions in Kerala than other states in India. Education The contrast between Kerala and India is -equally striking with respect to educational attainment. In 1981, about 70 percent of the population of Kerala were literate compared with 36 percent in India. Among females, the rates were 64 percent for Kerala and 25 in India. These differentials are not merely a carry-over from historical differences, which were quite large. Kerala's improvement in the last decade was greater than India's (8.8 points compared to 6.8). Table 2.3: Percent Literate in Kerala (1941-81) and India (1901-81) Kerala India 1901 5.4 1911 5.9 1921 7.2 1931 9.5 1941 48.1* 16.1 1951 53.8** 16.7 1961 46.8 24.0 1971 60.4 29.5 1981 69.2 36.2 * Travancore State ** Travancore and Cochin States Source: Registrar General and Census Commissioner for India, 1981. Census of India, 1981, Provisional Population Totals, Series 1, Paper 1, India, New Delhi: Pearl Offset Press, Statement 11 and Provisional Population Table 9. Years 1941-61 for Kerala from Census reports; citation to be added. - 15 - The development of education in the state owes much to the importance.given it in the Kerala culture and the work done by the European missionaries. Historically, education was provided to the Kerala masses using temples as centers. Non-formal education was imparted to both men and women through puranic stories (Hindu epics) and devotional songs (see Nag, 1981, pp. 31-32). M1ore recently, the rulers of both Travancore and Cochin gave considerable importance to education by making the state pay the entire cost of primary education. These efforts of the rulers were supplemented by private efforts which began with the foreign Christian missionaries' establishment of church-managed schools and colleges. The first Christian college (at Kottayam) was established in 1813. The improvement in the literacy rate alone does not give a full picture of the educational "explosion" in Kerala since 1950. Data on the number of students at all levels, on the number of schools and colleges of all types are necessary to map this picture. The data in Table 2.4 give statistics on formal education; that is, the schools and colleges managed by the government, private persons, or organizations, and recognized by the appropriate government agencies (Kerala State education department or University). There are a large number of other educational institutions outside this group-- the tutorial colleges and the parallel" colleges. The parallel colleges in Kerala are a unique institution. While the tutorial colleges take students who fail in public examinations, the parallel colleges take students who fail to gain admission in a recognized institution. Such students can study in parallel colleges and take the same examinations which students from recognized institutions take. If they pass, they get the same degree or diploma as the regular students. Thus, through the system of parallel colleges, a higher level of education, - 16 - Table 2.4: Educational Faciliti-es-in Kerala and India, 1979/80 0 Kerala India Average Population served by a school: Lower Primary 3,440 Upper Primary 8,800 High School 14,300 College 130,000 Average Area served by one school (sq. La.): Lower Primary 5.6 Upper Primary 14.2 High School 23.1 College 210.0 Per Capita Expenditures (Rs.): Total Education 68 36 Primary and Secondary Schools Higher Education Enrollmen't Ratio (enrolled /total age group): 6-11 Years - Boys 103.4 100.2 Girls 102.2 65.9 Total 102.8 83.6 11-14 Years - Boys 91.8 52.0 Girls 85.3 21.1 Total 88.6 40.2 Source: To be added. -17- especially in arts subjects, is made accessible to people from all walks of life. Limiting the discussion to formal schools and colleges, Table 2.4 indicates that schools and colleges are more easily accessible to people in Kerala as compared to India. There are more schools and colleges per population in 'Kerala than India, and Kerala spends a much higher proportion of its revenue for education than India. Further, the state spends relatively more money on primary school than on higher level education as compared to India. Two aspects of these educational developments are particularly relevant to the fertility decline. First, education is wo widespread that there are very few illiterate men or women in the prime childbearing ages. Second, the supply of educated personnel has been far in excess of the demand so that the extent of educated unemployment is becoming progressively worse. Political Awareness Every aspect of life in Kerala is politicized. Newspaper-reading and discussions of political news are principal pastimes of the common person. Protest marches, indefinite strikes, hunger strikes, and other public protests, are as much a daily occurrence in Kerala as the monsoon rain in July. For some people, participation in processions and protest meetings are a fulltime job for which they get the minimum daily wage plus transportation costs. In fact, the huts built and reserved for hunger strikers around the government secretariat in Trivandrum are booked months ahead for the various planned hunger strikes. - 18 - Such politicizing has many disruptive consequences, but it is not without beneficial effect. As Mencher says: "There is no question that politicisation of people in Kerala has played a major part in affecting people's health (for the better)... Tn Kerala, if a PHC (Primary Health Center) was un-manned for a few days, there would be a massive demonstration at the nearest collectorate (collector is the.chief administrator of a district) led by local leftists, who would demand to be given what they knew they were entitled to. This has had the effect of making health care much more readily available for the poor in Kerala" (Mencher, 1980, pp. 1781-1782). This is true not only for health services, but also for all public services such as education, rural bus service, postal service, food distribution, construction and maintenance of roads and bridges, and even contraceptive supplies and family planning services. The political awareness and politicization of all issues, even those affecting only a small minority of the population, may be responsible'for a better delivery of government services than would otherwise be the case. Newspapers and political leaders have explained the services and benefits available, along with their eligibility requirements. And what people are eligible for, they obtain by application, appeal, or agitation. Land Reform and Other Redistribution Policies Undoubtedly the most significant development in the recent political economy of the state is the shift in political power from the rich to the poor. Along with this transfer of power, and as a consequence of it, there has been a considerable shift of wealth and income toward the lower strata of society. These changes have been brought about through land reforms; job securiLy; enforcement of minimum wages; pension benefits for agricultural laborers; reservation of seats in educational institutions, public offices - 19 - (including the legislature and the cabinet) for the economically- and socially-backward communities; etc. Land reform was, of course, the principal means by which a major redistribution of wealth was expected to be brought about. In recent years (post-1947 period), the first legislative measure regarding land reform was taken by the first communist government in 1957. It sought to give permanent tenure to all tenants, the right to purchase ownership of land by the tenants, and a ceiling on the total area of land that a primary family unit could own. The proposal aroused considerable opposition and, as a result of the agitation which followed, the government was dismissed. Therefore, the actual implementation of land reform took place at a much later date (1964, the Kerala Land Reforms Act). By 1969, tenancy was altogether completely eliminated and rental income from land was virtually abolished throughout Kerala. The 1969 amendments to the Land Reforms Act lowered the ceiling on land holdings to 10 acres, removed some of the exemptions from ceilings, and gave the "hutment duellers" the right to purchase about one-tenth of an acre of land per household at a nominal price. The impact of these land reforms on the distribution of wealth and income has not been fully evaluated. It is likely to be greater in the Malabar area then in the former Travancore-Cochin area. At one extreme, some of the rich Nair and Brahmin landlords were financially ruined and today live far below their former status and income, possessing only those tracts of land where they built their houses. At the other extreme,the hutment dwellers gained legal ownership to the land around their huts. Whether it has caused any real change in their income is, however, very questionable. Eviction from - 20 - the huts was banned in much of Kerala several years before the, Land Reforms Act5/. The land reforms created a large middle class group of owner/ cultivators. These people should have gained by the land reforms, as they were no longer required to pay rents to their former landlords. But agrarian 6/ reforms simultaneously raised the wages and other benefits of the workers - and increased the cost of cultivation very considerably. The Kerala middle class farmers are accustomed to having all their agricultural work done by hired laborers and therefore wages are an important component of the cost of cultivation.7/ At the same time the prices of many agricultural products are controlled by the Government. It is, therefore, doubtful that land reforms have actually increased the income of the middle class farming families. How have the agricultural laborers fared under land reform? They constitute almost one-third of the labor force (30 percent in 1971; see Kerala State, Director of Census Operations, 1972). Have they gained in status? Lost out? Or not changed at all? They have benefited in two respects. Most 5 In 1949 beFore the formation of the Kerala State, the Travancore State Assembly passed legislation banning the eviction of hutment dwellers. 6/ Average paddy field labor wage rates in Kerala and India (Kerala, Directorate of Economics and Statistics, 1980, Tables 4.11(c) and (d); and Rs./day Period Kerala India 1960-61 1.85 1970-71 5.09 1978-79 6.26 "...even on operational holdings of less than 0.5 acre about one sixth of the rural households depend on hired labour for the bulk of the agricultural work." (See United Nations, Department of Economic and Social Affairs, 1975. Poverty, Unemployment and Development Policy, New York: United Nations, p. 71). - 21 - of them used to live in houses located on land belonging to other people. The land reforms have bestowed on them permanent rights to their house sites and a piece of land immediately around their houses (10 cents in rural areas, much less in towns and cities). 8/ They can, therefore, live in their houses without fear of eviction and begin to cultivate fruit trees and vegetables in the land around their homes. They have also benefited by the Minimum Wage Act and other agrarian reforms which fixed wages and conditions of work for all agricultural operations. The gains from higher wages (per day) have been partly off-set, however, by the losses encountered due to a reduction in the number of days a wage earner can find work and a reduction in employment opportunities for other family members, especially grown-up children. "....t is only in Kuttanad (Alleppey District) that I have heard workers talk about time spent harvesting in terms of 45 minutes or 1.5 hours (rather than in terms of days or at least half days -as they do els-ewhere)" (Mencher, 1980, p.'1787). It is therefore doubtful that the increase in the minimum wage is fully reflected by the average annual household income of an agricultural laborer. 8/ 1 cent = 1/100ths of an acre. - 22 - Part 1: BACKGROUND CHAPTER III KERALA STATE: DEMOGRAPHIC AND FAMILY PLANNING BACKGROUND The demographic picture of Kerala is one of a very dense population scattered fairly uniformly over the state, evidencing few nodes of concentra- tion. The population is fairly well advanced in its demographic transition with a high but rapidly declining rate of population growth, a high average age at marriage, a high degree of family planning practice, relatively low and declining levels of fertility rates, very low and very slowly declining rates of mortality, and a high rate of population mobility. The population is fairly well advanced with regard to literacy and educat-onal attainment, but only moderately successful in its economic transition, with nearly half the labor force still depending on agriculture and a third on daily wages from agriculture. These characteristics of Kerala's demographic profile and the official family planning program are described in greater detail below. Population Growth Until recently Kerala's rate of population growth has always been higher than that of India (Table 3.1). Between 1901 and 1981 Kerala's popula- tion grew by 297 percent while that of India grew only 187 percent. But, during 1971-81 the Indian growth rate was substantially higher. The prospects for a further decline in the growth rate are high in Kerala; as a result, Kerala's proportional contribution to the population of India is likely to diminish in the years to come. - 23 - Table 3.1: Population Growth in Kerala and India, 1901-81 Annual, Population Growth Rate Year of Kerala Kerala India (1,000s) (Percent) 1901 6,396 1911 7,148 1.90 0.56 1921 7,802 .90 -.03 1931 9,507 1.98 1.04 1941 11,032 1.50 1.33 1951 13,549 2.08 1.25 1961 16,904 2.24 1.95 1971 21,347 2.26 2.22 1981 25,403 1.74 2.22 Source: Government of Kerala, Directorate of Economics and Statistics, 1980. Statistics for Planning 1980, Cochin: Text Book'Press, Table 1.1; and Government of India, Registrar General and Census Commissioner, 1981. Census of India, 1981, Provisional Population Totals. Series-1, Paper 2, New Delhi: Pearl Offset Press, Statement 3. Population Distribution The state is divided into 12 administrative districts, with Trivandrum the most southern and Wynad the most northern. The population growth rate, and density of these districts are shown in Table 3.2 (and following map). The districts vary in population size from 2.8 million in Quilon and Cannanora to half a million in Wynad. The density also varies substantially from 192 persons per sq. km. in Idukki to 1,244 in Alleppey. The main reason for this variation is differences in district topography. - 24 - Table 3.2: Population Size, Growth Rate and Density by District, Kerala State. Population Density Percent Growth District 1981 1981 1971-81 (1,000s) (pop./sq.km.) Trivandrum 2,591 1,182 17.85 Quilon 2,807 608 16.35 Alleppey 2,343 1,244 10.21 Kottayam 1,681 763 9.23 Idukki 971 192 26.85 Ernakulam 2,533 1,052 17.08 Trichur 2,437 804 14.48 Palghat 2,042 465 21.16 Malappuram 2,401 654 29.35 Kozhikode 2,243 969 23.12 Cannanore 2,800 565 25.24 Wynad 553 260 33.71 Kerala 25,403 654 19.00 Note: District populations do not add to state total of 25,403,000 due to rounding error. Source: Government of Kerala, Director of Census Operations,, 1981. Census of India 1981, Kerala, Series 10, Provisional Population Totals, Paper 1 of 1981, Cochin: Text Book Press, Table 1. 徊一‘一，一‘一”一’〕 26 Kerala is not a highly urbanized state. In 198.1, only 18.8 percent of its population 1:.ved in urban areas compared with 23.7 percent in India as a whole. There is no city in the state with a population of a million or more people. The largest urban agglomeration is Cbt.chin with a total population of 686,000 in 1MI; and six other towns have a popislation of 100,000 or more (Ta.ble 3 .3) . I/ Table 3.3: Towns with a Population of 100,000 or More in 1981 Cochin 686,000 Calicut 546,000 Trivandrum 520,000 Trichur 170,000 Alleppey 170,000 Cannanore 158,000 Palghat 118,000 Source: As in Table 3.1, Government of India, Registrar General and Census Commissioner, 1981, Paper 2, Appendix II, pp. 122-123. One of the distinguishing features of the demography of Kerala is the distribution of its rural population and absence of rural-urban differen- tials. There is no viilage system in Kerala as is commonly known in other states of India. The locality concept of villages is completely non-existent in Kerala. There is no clustering of houses; instead one sees a scattering of houses in every locale, with plenty of space inbetween the structures and very often compound walls (this, in spite of the high density). The viilages are 1/ "An urban agglomeration is by definition the continuous urban spread consisting of a core town and its adjoining urban outgrowths which may be either urban in their own right or rural." (See Government of India, Registrar General and Census Commissioner, 1981, Paper 2, p. 23.) .. ........ - ,- - ........ . . - 27 - administrative divisions, and large with respect to population as compared to the Indian norm. Table 3.4 gives the distribution of villages by size in Kerala and India. The contrast is very vivid indeed. Table 3.4: Distribution of Villages by Population Size, Kerala and India, 1971 Percent of Rural Size Population, 1971 Kerala India Less than 1,000 0.1 37.9 1,000-1,999 0.1 25.8 2,000-4,999 2.6 23.8 5,000-9,999 13.7 7.5 10,000+ 83.5 5.0 Total 100.0 100.0 Source: Government of Kerala, Director of Census Operation, 1972. Census of India, 1971, Series-9, Kerala, Part II-A, General Population Tables, New Delhi: Government of India Press, State Table A-III; and Government of India, Register General and Census Commission, 1971. Census of India, Part II-A, Census Tables on Population, New Delhi: Government of India Press,. Table A-III. Fertility The crude birth rate in Kerala was above 40 (per 1,000 population) before 1950, but it probably is below 25 now. The trend is given in Table 3.5. The birth rate started declining in the early Fifty's, but gathered momentum only in the second half of the 1960's. - 28 - Table 3.5: Crude Birth Rate in Kerala and India, 1931-78. (Births per 1,000 Population) Period Kerala India 1931-40 40 45 1941-50 40 40 19-51-60 39 *42 1961-70 37 41 1966 36 - 1967 34 - 1968 33 - 1969 32 - 1970 32 - 1971 32 37 1972 30 37 1973 27 35 1974 28 35 1975 28 35 1976 26 34 1977 25 33 1978 25 33 Sources: Government of India, Central Bureau of Health Intelligence, Directorate General of Health Services, 1981. Pocket Book of Health Statistics of India, 1980, Coimbatore: Government of India Press, Table 7; and Kerala estimates for 1931-40 to 1961-70 prepared by Government of India Census Actuary, while those for 1966-78 are from the Sample Registration'System (see Government of India, Vital Statistics Division, Ministry of Home Affairs, 1966-78. "Sample Registration: Bulletin", Volumes for 1966-78, New Delhi: Office of Registrar General). The fertility rates associated with these birth rates are shown in Table 3.6. The Total Fertility Rate (TFR) was about 5.2 in 1959 and 3.4 in - 29 - Table 3.6: Age-Specific Fertility Rates in Kerala, 1958, 1971, and 1976. 1958-59 1971 SRS % decline 1976 SRS % decline % decline NSS 14th (Rural) 1958-59 (Rural) 1971-76 1958-59 Age Group Round-Rural, to 1971 to 1976 adjusted rates 1 2 3 4 5 6 7 15 - 19 87.5 48.4 44.7 42.7 11.8 51.2 20 - 24 237.9 210.8 11.4 193.4 8.3 18.7 25 - .29 291.8 223.3 23.5 202.3 9.4 30.7 30 - 34 205.0 173.2 15.5 137.1 20.8 33.1 35 - 39 161.8 116.7 27.9 80.7 30.8 50.1 40 - 44 46.7 42.8 8.4 25.3 40.9 45.8 TOTAL 5.15 4.07 (21.0) 3.41 (16.2) (33.8) (TFR/% decline) Sources: Based on Government of India, Central Statistical Organization, National Sample Survey Report No. 89 (N.S.S. 14th Round); and Government of Kerala, Directorate of Economics and Statistics, 1980. Statistics for Planning 1980, Cochin: Text Book Press, Table 1.12. 1976, a decline of about 35 percent. 2/ In absolute terms, it is equivalent to a decline of 1 .8 children per woman. From 1959-71, the largest fertility decline was in the 15-19 age group, due largely to a decline in the proportion married. The varying declines noticed among other age groups should mostly be due to contraceptive practice during 1971-76. The fertility declines at older 2/ The Total Fertility Rate is the average number of children that would be born to a woman if, during her childbearing years, she had the average number of births reflected in the age-specific fertility rates of a given year. The age-specific rates are derived by dividing the number of births to women of a certain age by the number of women in that age group. - 30 - ages were very pronounced, and may reflect the increasing usage of steriliza- tion among these women. Mortality Kerala's death rate has always been much lower than that of India. The crude death rate was estimated to be about 25 in the 1930's and about 7 in 1979. Much of the- decline took place before 1970 (see Table 3.7). The largest ga-in over mortality was experienced in the infant ages. Reliable estimates of the Infant Mbrtality Rate (IMR) for earlier periods are lacking. 3/ The estimates for 1951-60 onwards are given in Table 3.8. It has been less than 50 in recent years. Table 3.7 Crude Death Rate, Kerala and India, 1931-78. Period Kerala India 1931-40 25.0 31 .2 1941-50 20.0 27 .4 1951-60 16.1 22.8 1961-70 13.5 19 .0 1971 9 .2 14.9 1972 9.4 '16.9 1973 8.7 15 .5 1974 8.0 14.5 1975 8.5 15 .9 1976 8.3 15.0 1977 7 .5 14.7 1978 7 .2 14 .2 Sources: As in Table 3.5. 3/ The Infant Mbrtality Rate is defined as the numberoof infant deaths per year per 1,000 live births during the year. - 31 - Table 3.8: Infant Mortality Rates for Kerala and India, 1951-79. Period Kerala India 1951-60 120 1965-67 74 - 1971 61 129 1975 57 140 1976 56 129 1977 50 129 1978 43 125 1979 44 125 Sources: As in Table 3.5, Government of India, Central Bureau of Health Intelligence, Directorate General of Health Services, 1981, Table 14; and Kerala estimates as in Table 3.5. The difference between India and Kerala with respect to the IMR is very large indeed. For every 1,000 births, 80 more babies died in India dur- ing the first year of life than in Kerala in 1979. To put it another way, a baby's chance of dying in India during the first year is almost three times that of a baby born in Kerala. Official estimates indicate that in the 1930's life expectancy at birth in Kerala was at the same level as in India, about 32-33 years; but in recent years, Kerala's life expectancy (65 years) is about ten years greater than that of India. Ever since data have become available in Kerala, female expectation of life at birth has been higher than male life expectancy. This is a great contrast with India where males have had a higher expectation of life at birth. This favorable female mortality differential is reflected in the sex ratio of the total population of the state. Kerala is one of the few states in India in which the number of females has been greater than the number of males. - 32 - Migration Migration is a relatively small factor affecting the size and distri- bution of population in Kerala; yet, migration is a very important aspect of the socio-economic life of the state. Until 1931, Kerala received more people than it sent out. Since World War II, the trend has reversed and Kerala is a state of net out- migration. The first census of Kerala (1961) showed that 624,000 Kerala-born people were living outside the state and 233,000 outside-born people were liv- ing in Kerala, leaving a net lifetime loss of 391,000. By 1971 these numbers increased to 943,000; 271,000; and 672,000 respectively. For more recent periods, emigration to West Asian countries has become an important component of migration from Kerala. Although the volume is not so large as the volume of out-migration to other parts of India (one estimate puts 187,000 Kerala migrants outside India and migrants outside Kerala, but within India), the economic and social impact of West Asian emigration is much larger. The remittances to Kerala alone are esti- mated to be about US$200 million annually. Internal migration in Kerala, even rural-urban migration is not very large in Kerala. This apparent immobility is more than compensated by the increasing volume of commuting, particularly daily commuting for work and studies. A very rough indication of it is given by the fact that the govern- ment bus transport service alone carried each person in Kerala an average of about six times a year in 1960, 15 times a year in 1970, and 29 times a year in 1980 (refer to Government of Kerala, Directorate of Economics and - 33 - Statistics, 1980, Table 12.7). Because of the cheap transportation service, commuting has replaced internal migration in the state. Family Planning Program The family planning program in Kerala has developed mostly along the lines of the program development in India. But, over the course of years the state. has evolved certain innovative strategies, like the mass camps, which were later adopted in other parts of the country. The program made a very modest beginning in 1955 with 11 Family Planning Clinics attached to medical institutions. The course of progress since then falls into four distinct phases-- namely, (1) a period of slow growth (1955-64), (2) a period of reorganization and provision of state-wide Family Planning Service Centres (1964-70), (3) the years of the mass steril- ization camps (1970-73), and (4) a period o. intensified-maternal and child health (MCH) services (1973 onwards). During the first phase of development, a state Family Planning Board was constituted (1958) to develop the policies and procedures of the program and ensure their implementation. During the period 1956-61, 70 Family Planning Clinics were opened in the state, with facilities for sterilization in 53 institutions. During the next four years (1961-1965), 50 more clinics started functioning, and Family Planning Centres were opened in 93 panchayats. In each of these centers, two social workers (one male and one female) were entrusted with family planning education work. The supervision of the program was vested in three new "Regional Family Planning Officers." Later, 20 more facilities were initiated and incentives for promoters and doctors were introduced. At various levels, committees were formed to help in the activities of the Family Planning Centres. In 1964, on the basis of the - 34 - recommendations made by the Mukerjeo Committee, a committee set up by the Government of India to review the organization and management of the Family Planning Program, a network of service units and a hierarchy of administrative posts were-established. A District Family Welfare Bureau at the District Headquarters, a State Family Welfare Bureau at the Directorate of Health Services, and a cell in the State Secretariat also started functioning. Under this new arrangement, urban centers were attached to various medical institu- tions. In rural areas, a Family Planning Centre was started in each Primary Health Centre (PHC). Under each PHC, sub-centers- one each per 10,000 population- were also established. It was during this period, that I.U.D. and oral contraceptive programs were started. The third phase, covering the period from the end of 1970 to April 1973, saw the conduct of mass sterilisation camps. These camps not only proved to be pace-setters for the country as a whole, but were responsible, in -large measure, for the popularity of sterilization in the state. Nine one-month-long camps were held between 1970 and 1973 in seven of the 11 districts in the state. The most ambitious and impressive in terms of achievement were the three camps in Ernakulam district (one of the districts in the present survey), where the target was set at 85,000 sterilizations, and the achievement was 93,254 sterilizations (110 percent of the target). Impressive performance was also witnessed in Palghat, another survey district, where achievement represented 101 percent of the target (the total sterilized was 10,083 and the target was 10,000). Although no camps were held in Alleppey, the third survey district, 25 percent of the acceptors in the second Ernakulam camp were from Alleppey. - 35 - An evaluation of the performance of the first two camps in Ernakulam revealed that the camps reached 22.5 percent of the couples in the district with two or more children. The camps demonstrated that what could normally be achieved during a five-year period of official program activites could be accomplished in a single one-month camp. The camps, however, had their repercussions on subsequent program performance. For Ernakulam District, the average sterilization operations per year during the two-year period before the camps was 9,400, but the average after the camps fell to 5,500. The drop was of short duration, however, last- ing only two years. The Kerala experience showed that camps were a practical means to increase family planning acceptance. The fourth phase of the historical development of the program began with the abandonment of mass camps in 1973. This period is marked by the intensification of the integrated Family Planning Program and MCH approach and mini-camps. Facilities for female sterilization were extended to rural and semi-urban areas. Arrangements were made to ensure that the benefits of the integrated MCH and nutrition programs reached the people in rural areas through the sub-centers and those in urban areas through the urban family planning centers. Some other features of the Family Planning Program, like the hospital post-partum program, were also introduced during this phase of development. The introduction of the Medical Termination of Pregnancy (MTP), which became legally permissible as of April 1, 1972, is another important develop- ment. The impact of MTP on the program is discussed later. In 1976, a new and simple technique of female sterilization known as mini-laporactomy was introduced in the state. This has become very popular as - 36 - the technique can be done at any time on a non-post-partum basis and without hospitalization. The oral pill program was started in 1968-69 as a pilot project, in a few selected centers. It now operates in 163 rural Family Planning and 42 urban Family Planning Centers, but the number of women who have accepted the pills was only 1,071 during 1975-76 and 271 in 1979-80. Oral pills are also distributed in nine other centers; that is, eight voluntary organizations and one local body. Another new development during this phase was the introduction of the multi-purpose Health Workers Scheme. According to this arrangement, the existing lower level staff in various programs are pooled and designated as Male and Female Health Workers. Each of these health workers covers a popula- tion of 5,000 to 7,000 and serves all the health needs of the people. The offer of integrated health services is effected both at the periphery and supervisory levels. The scheme has been introduced in two districts of the state. The conduct of Family Planning Fortnights, two weeks set aside once a year by the state government to promote the family planning program, has been a regular feature since 1966. Their observance helps to create an awareness of the problem among the workers and encourages enthusiasm about their work. Since 1974, the fortnights have also been periods of conducting mini-camps. These are held in each of the PHCs or combinations of PHCs by pooling resources. The State Ministry of Health guides and supervises the program in the state. The organizational setup is given in Figure 3.1. There is a State Cabinet Sub-Committee on Family Planning to develop policies and review - 37 - program progress in the state. In order to advise the government, there is a State Family Welfare Council consisting of official and non-official members with the Health Minister as Chairman. There is a cell in the Secretariat headed by a Deputy Secretary (under the overall guidance of the Secretary to Government, State Health Department) in charge of the scheme. Technical supervision is vested with the Director of Health Services at the State level, and with the District Medical Officer at the District level. The State Family Welfare Bureau (SFWB) and District Family Welfare Bureau (DFWB) are in direct charge of the work. The State Family Welfare Bureau is headed by an Additional Director of Health Services (Addl. D.H.S.), assisted by an Assistant Director of Health Services (ADHs) for Family Welfare and another Assistant Director for Maternal-Child Health (MCH), a Medical Officer for IUD, and a State Mass Education and Media Officer (SMEMO) for mass education work. In this mass education division, there is a Chief Health Education Officer and an Editor. For the current evaluation of the program, there is a Demographer with a cell under him- namely, the Demographic and Evaluation Cell. The Regional Training Centres and the Post-Partum Program receive direction from the Additional Director for the State Family Welfare Bureau. Administrative concerns are housed with an Administrative Assistant and accounting matters are handled by the Senior Accountant. The District Bureau is headed by the District Family Welfare Medical Officer (DFWMO), who is aided by an Adminsitrative Assistant, a Mass Education and Information Officer (MEIO) and a Statistical Cell having a - 38 - Figure 3.1: Family Welfare Program, Organizational Chart Family Welfare Council J-4 State Cabinet Subcommittee on Family Welfare Secretary to Government, State Department of Health Deputy Secretary for Family Welfare, The Secretariat Director of Health Services Additional Director, in charge of State Family Welfare Bureau Asat.Dir.of Asst.Dir. of Medical State ass Demographer Regional Post- Admin. Senior Health Services Health Services Officer Educ. and (Demog. and FamilA Partum Asat. Accountant for Family Wel- for Maternal- for IUD Media Officer Eval. Unit) Welfare Programs (Admin. fare Child' Health Training (13) and Centres Stores) T ~ -1 (2) 4 - _T Chief Health Editor Social Statis- Stores Auditor Education Scien- tician Officer Officer tist I, Statist. Press Ast. District Family Welfare Medical Officer, District Family Welfare Bureau Admin. Mass Education Statistical Mobile IUD Units Oral Contraceptive Asst. and Information Assistant, (2) Prograa Officer in charge of Statistical Cell Urban Family Rural Family District Extension Welfare Program Welfare Program Educators Centres (21) Centres (163) Subtentres (1,797) BEST COPY AVAILABLE -39- Statistical Assistant. The Mobile IUD Units and the Oral Contraceptive Program are also under the overall direction of the District Family Welfare Medical Officer. Delivery of the many family planning services is centered around the various types of activities undertaken by the subcentres, PHCs (rural and urban), in the Taluks (administrative subdivision of district above panchayat level), District Hospitals, and other hospitals attached to Medical Colleges. At the sub-centre level, couples are advised about the Family Welfare Program (FWP), and conventional contraceptives (CC) are distributed. The sub-centre is managed by an Auxilliary Nurse Midwife (ANM) who has both clinical and extension work training. She assists the Medical Officer in IUD insertions and follows-up the acceptors. During her routine house visits, she advises and guides mothers and children on nutrition, sanitation, immuniza- tion, and the importance of limiting family size. At the PHC level, besides distribution of CC, there are facilities for vasectomy operations and IUD insertions (in a limited number of PHCs, the tubectomy operation is also done). At the PHC, the Medical Officer provides the necessary leadership and ensures coordination of work among the staff. He also maintains liaison with agencies like the Community Development Blocks and ensures supplids of medicines and contraceptives are available. Organization of mini-camps, training camps, the conduct of vasectomy operations,.and attending follow-up services also fall within his responsibility. The female Medical Officer attached to the PHC assists the Medical Officer in developing and directing the MCR and FW needs of the community and in educating the public about the need for acceptance of Family Welfare. - 40 - Medical and para-medical staffing for the Family Welfare Program are shown in Table 3.9. With the increased attention to extension activities, the budget of the program has been increased from year to year. The expenditure during 1958-59 was just less than Rs 200,000. By L%4-65 (the year of major reorganization), the expenditure rose to Rs. 3,680,000. The expenditure over the years since 1964-65 has shown a steadily upward trend. The total expenditure for 1979-80 was Rs. 55,550,000 (or about U.S. $6 million). 4/ An examination of expenditures for the Family Welfare Program shows that over the years, the major portion of the budget has been spent on Table 3.9 Budgeted Medical and Paramedical Staff, Family Welfare Program, June 30, 1978. Medical Block Lady Aux. Family Officers Extension Health Nurse Welfare Educators Visitors Midwives Health Assists. Rural Family Welfare Program & Sub- centres 163 163 163 1,485 894 Urban centres 7 7 7 7 7 Total 170 170 170 1,492 901 Source: Government of India, Department of Family Welfare, 1979. Family Welfare Programme in India, Year Book 1978-79, New Delhi: Mass Mailing Unit, Department of Family Welfare, pp. 129-131. 4/ The January 1982 exchange rate was U.S. $1 = Rs. 9.198. - 41 - services rendered. During the period from 1959 to 1976, 85 percent of funding went for this purpose. 5/ The comparable figure for 1976-80 is 91 percent, so that administrative and other overhead costs have decreased while spending for direct program activities has increased. Detailed expenditures since 1952 are given in Table 3.10. Incentives Under the Family Welfare Program An incentive in the context of the Family Welfare Program is the direct payment of money or material by the government or another organization to an individual, couple, group, or institution for the purpose of promoting contraceptive practice. Incentives in the form of financial or material goods are given to the acceptors to compensate for loss of wages in the process of family planning (F.P.) acceptance. It is given to the motivators and service- rendering personnel, as well. The rate of incentives given to the acceptor, motivator, doctor, and others has varied. The payment of incentives to acceptors also has differed by the type of method adopted. Apart from the monetary incentives given to the acceptors of F.P. methods, incentives in the form of materials (such as transistor radios, cycles, lunkies, sarees, etc.) are also given. Voluntary organizations (the Lions Club, Rotary Club, etc.) 5/ This represents the percentage of expenditures made on all budgetary items excluding transportation, construction, and administration. - 42 - Table 3.10: Expenditures on Family Welfare, Kerala, 1952-80. (Rs.) Year Adminis- Rural ass Construc- Compen- Transport Others Total tration & Urban Educa- tion sation F.W. tion (Incen- Services tives) 52-59 11,125 65,466 117,206 .. .. .. .. 193,797 59-60 19,994 157,211 108,849 .. ** ** .. 286,054 60-61 7,885 234,882 75,099 .. .. .. 317,866 61-62 264 260,129 315 .. .. .. .. 260,708 62-63 2,594 447,102 5,790 .. .. .. 2,075 457,741 63-64 2,697 1,089 ,143 26,196 . .. 231,438 1,349,474 64-65 40,138 3,107,466 19,121 .. .. .. 490,121 3,656,846 65-66 134,303 5,918,797 65,572 .. - .. 10,012 6,128,684 66-67 63,916 8,134,551 459,972 .. .. .. .. 8,658,439 67-68 166,035 12,152,833 440,071 912,851 .. .. 459,960 14,131,750 68-69 193,303 15,373,027 448,263 1,496,031 .. .. 960,095 18,470,719 69-70 218,551 17,166,098 489,594 106,637 .. .. 947,580 18,928,460 70-71 218,650 19,996,686 789,603 808,696 .. .. 929,352 22,742,987 71-72 1,629,047 15,100,844 2,185,186 779,561 7,709,751 1,688,877 4,334,767 33,428,033 72-73 1,799,341 16,563,638 3,618,424 2,356,238 3,353,043 1,340,863 3,710,915 32,742,462 73-74 1,709,354 17,115,900 4,444,439 2,275,430 1,320,327 936,272 2,674,307 26,476,029 74-75 1,991,052 18,628,830 667,062 1,419,655 1,877,285 804,987 2,507,808 27,896,679 75-76 2,327,259 22,024,368 241,670 971,510 9,156,450 1,056,891 3,825,628 39,603,776 76-77 2,380,301 23,388,071 211,782 676,954 30,916,648 1,326,338 4,493,816 63,393,910 77-78 2,447,235 23,537,559 184,355 192,355 10,286,870 1,245,198 5,315,485 43,209,057 78-79 2,467,692 25,342,940 686,213 46,485 10,191,054 1,532,620 5,985,539 46,252 543 79-80 2,679,455 27,632,474 790,868 1,757,079 15,041,220 1,278,285 6,375,177 55,554,558 Source: Based on annual reports of the Government of Kerala, State Department of Health, State Family Welfare Bureau. - 43 - have also paid incentives to acceptors. The rate of incentive payments made in 1966-67 is given in Table 3.11. Table 3.11: Rates of Incentive Payments In 1966-67 (Rs.) Vasectomy Tubectomy I.U.D. Acceptor 21 29 5 Medical Officer 3 3 2 Anaesthetist .. 2 Promotor 2 2 1 Drugs, etc. 2 2 2 Service 2 2 1 Total Rs. 30 40 11 Source: Provided by the Government of Kerala, State Department of Health, State Family Planning Bureau. These rates were later revised. According to the revised rates, the compensation paid to acceptors in 1977 was increased, as follows (Table 3.12): Table 3.12: Rates of Incentive Payments in 1977 (Rs.) Vasectomy Tubectomy IUD Acceptors 70 70 6 Drugs 10 15 Diet 5 20 - Transport 5 5 1.50 Miscellaneous 10 10 0.50 Total Rs. 100 120 8.00 Source: As in Table 3.11. Prior to 1977, a motivation fee of Rs. 5/- was paid to the motivators for each case of sterilization. When the state contribution towards the mis- cellaneous purpose fund ceased, however, the payment of this motivation fee -44- was dispensed with. Later it was revived at the rate of Rs. 3/- per case of sterilization brought into the mini-camps or during a period of the Family Welfare fortnights. Again in July 1979, the rate of the motivation fee was modified and fixed at a rate of Rs.6/- for all cases of vasectomy promoted and Rs.3/- for all cases of female sterilization promoted in the camps. In 1981, payment of monetary incentives was made to the Family Welfare service personnel for each case of sterilization done in the camps. The following rates were established: Medical Officer - Rs.3/- Nurse/A.N.M.- Rs.1/- Nursing Assistant- Rs.1/- Stretcher Bearer - Rs.0.50 Barber (preparation)- Rs0.50 During the period of a special Family Welfare Campaign organized from January-March 1981, additional incentives were given to acceptors as well as motivators. Accordingly, each acceptor of sterilization was given Rs.10/- and each promoter of sterilization was paid a bonus of Rs.4/-; that is, the moti- vators were given a total motivation fee of Rs.10/- for each case of vasectomy promoted and Rs.7/- for each case of female sterilization promoted. Corporation Councils, Municipal Councils, and the Gurufayur Township Committee were also asked to pay incentives to all persons who underwent sterilizations in their areas at the enhanced rate of Rs.50/-. Voluntary organizations (like the Lions Club, Rotary Club, etc.) also paid monetary incentives or other materials in their working areas to the acceptors and motivators promoting Family Welfare acceptance. - 45 - In collaboration with the Labor Department a novel scheme called the Voluntary Promotor's Scheme, a part of the Employment Generation Scheme, was launched during the year 1979-80. Educated, unemployed, satisfied acceptors with the prescribed qualifications were selected as Voluntary Promoters. They were given incentives at the rate of Rs.10/- for each case of sterilization they promoted and Rs.5/- for each I.U.D. case promoted in the area allocated to them. This Scheme continued till 1980-81. The incentives paid to acceptors, motivators and the service- rendering personnel have greatly assisted the improvement of contraceptive practice in the state. Impact on Fertility A "cafeteria".approach is followed in the state for the delivery of supplies and services. Accordingly, a variety of methods are offered to the public who can choose the method they want after proper consultations. The methods consist of both male and female sterilization, IUD, CC (like condom, diaphram, jelly/cream tubes, foam tablets), and oral pills. A Teview of the progress of achievement in usage of the methods over the years shows that sterilization is the bulwark of the program. The progress of IUD acceptance has been checkered. The CCs other than condoms are practically unused. Pregnancy termination (MTP), though introduced as a nonfamily welfare measure in 1972, is becoming more and more popular. The number of sterilizations (both male and female) done during the initial years was small. In 1957 only 679 sterilizations were done. The number gradually rose and reached 21,904 in 1964. The years of mass camps (1970-73) recorded achievements of above 100,000. The highest number ever recorded during a single year was in 1976, when 266,000 sterilizations were -46- done. The years 1978 and 1979 saw a slow-down in performance. In 1980, however, the performance was much better; the provisional total for 1980 was about 104,000 sterilizations. Detaiiing the total number of sterilizations by sex of acceptor shows some interesting changes in proportion over the years. From 1957 to 1973, the number of vasectomies (male sterilizations) out-numbered tubectomies (female sterilizations), accounting for as much as 84 percent of total sterilizations in 1966. But since 1974, there has been a reversal of the trend, the number of tubectomies exceeding the number of vasectomies in all the years except 1976. During 1978-80, the number of tubectomies was more than five times the number of vasectomies. The recent spurt in tubectomies is due to a change in the mode of operation; namely, minilap- which makes the sterilization of the female very much simpler than it used to be. The cumulative number of steril- izations done through 1980 is almost 1 1/2 million. The IUD was introduced in 1965. After an encouraging initial start, acceptance was low by 1973. Some gains were seen from 1974 through 1976. Since 1977, however, the number of acceptors has again gone down to below 20,000 a year. The use of CCs is mostly confined to condoms (Nirodh). Of late, the use of other methods has been practically discarded. The number of various CCs distributed each year is converted to the number of CC users through a norm adopted by the government of India. The norm is applied by dividing the total distribution of condoms by 72, of diaphragms by 2, jelly/cream tubes by 7, and foam tablets by 72. These norms are based on the average number of products nec,ssary for a couple to have complete protection over a year. There has been a steady increase in the number of CC users since 1973. - 47 - Table 3.13 shows the record of achievement of the various Family Welfare methods since 1957. Several factors must be considered in estimating the number of births averted by acceptors of the different methods. These considerations include: the differential efficiency of the methods, the carry-over effect of some of the methods, the mortality of either spouse, the possibility of expul- sion or removal as in the case of IUDs, and changes in the age structure of acceptors of IUDs and sterilization (both male and female). The estimates given in Table 3.14 are based on the Age-Specific Marital Fertility Rates (ASMFR) obtained from the National Sample Survey (NSS) for 1959-60, and adjusted to correspond to a birth rate of 38.9 per 1,000 (estimated from the Census for 1951-60). These estimates show that by 1978, a total of 1,736,000, births had been averted; and by 1980, 2,238,000 births were averted in the state. The impact of the program on the crude death rate is given in Table 3.15, and the number of couples estimated to have accepted each method is given in Table 3.16. History and Development of Induced Abortion in the State Induced abortion was illegal in Kerala and punishable under the Indian penal code except under special circumstances. The law was changed by the Medical Termination Act of India, 1971, which was made effective from the 1st of April 1972. The intent of the Act and the rulemaking which followed was that pregnancy could be terminated for the failure of any contraceptive device or method, or due to economic or health reasons. The Act has allowed abortion to be used, in effect, for interrupting pregnancies and thereby as an aid to population control. - 48 - Table 3.13: Achievement Record of Vasectomy, Tubectomy, I.U.D., and C.C. Users-- Kerala, 1957-80. Year Vasectomy Tubectomy Total IUD CC Users Equivalent Sterilization Sterilization* 1957 521. 158 679 - - 679 1958 L,633. 1,507 3,140 - - 3,140 1959 4,132 2,236 6,368 - - 6,368 1960 3,079 1,953 5,032 - - 5,032 1961 3,578 2,939 6,517 - - 6,517 1962 4,182 2,916 7,098 - - 7,098 1963 10,395 2,830 13,225 - - 13,225 1964 17,938 3,966 21,904 - - 21,904 1965 36,102 6,532 42,634 23,062 - 50,321 1966 33,251 6,147 39,398 43,517 - 53,903 1967 49,489 10,504 59,993 36,887 - 72,288 1968 64,081 14,066 78,147 39,742 - 91,394 1969 42,578 17,982 60,560 36,866 - 72,840 1970 48,847 20,592 69,439 32,559 9,650 80,828 1971 76,141 24,192 100,333 19,521 7,824 107,274 1972 73,683 30,374 104,057 21,719 4,522' 111,548 1973 48,448 35,707 84,155 15,690 11,797 95,939 1974 15,056 41,174 56,230 26,039 14,104 72,746 1975 22,823 48,981 71,804 22,720 17,228 88,948 1976 186,037 80,434 266,471 24,961 8,939 275,288 1977 33,216 76,472 109,688 11,468 20,996 114,677 1978 14,283 73,256 87,539 10,930 21,672 92,386 1979 15,573 80,507 96,030 14,663 19,288 102,039 1980 15,439 89,088 104,527 19,174 17,778 111,906 Grand Total 860,505 674,513 1,475,028 399,518 153,798 1,658,288 * Sterilization plus one-third of IUD insertions and one-eighteenth of CC users. Source: Provided by the Kerala State Family Planning Bureau, Directorate of Health Services, Trivandrum. MTP is done only in approved medical institutions. The number of such institutions has been gradually increased. At present, 77 government and 51 private institutions are approved for the purpose. -49 - Table 3.14: Number of Births Averted Due to Use of the Various Methods, Kerala, 1957-80 Method Method Year Vasectomy Tubectomy Total IUD C.C. Total 1957 18 10 28 - - 28 1958 181 140 321 - - 321 1959 647 598 1,245 - - 1,245 1960 1,558 1,164 2,722 - - 2,722 1961 2,194 1,693 3,887 - - 3,887 1962 2,899 2,392 5,291 - - 5,291 1963 3,872 3,030 6,902 - - 6,902 1964 6,296 3,667 9,963 - - 9,963 1965 10,693 4,663 15,356 975 - 16,331 1966 18,397 6,105 24,502 6,106 37 30,645 1967 25,428 7,633 33,061 12,328 445 45,834 1968 35,684 10,195 45,879 15,693 1,337 62,909 1969 47,333 13,575 60,908 18,688 1,383 80,979 1970 53,694 17,704 71,398 20,101 1,515 93,014 1971 67,392 22,516 89,908 19,865 1,239 111,012 1972 81,221 28,303 109,524 17,781 972 128,277 1973 90,979 35,469 126,448 16,610 883 143,941 1974 107,534 43,760 151,294 15,255 1,721 168,270 1975 89,639 53,311 142,950 15,740 2,070 160,760 1976 105,899 56,288 162,187 15,493 3,190 180,870 1977 129,149 85,049 214,198 15,141 5,087 234,426 1978 125,416 101,259 226,675 12,839 3,502 243,016 1979 118,777 115,777 234,554 11,370 N.A. 245,924 1980 112,534 132,184 244,718 10,904 N.A. 255,622 Note: N.A.- Not applicable. Source: Kerala State, Population Research Centre, Directorate of Economics and Statistics, 1980. "Note on the Calculation of Births Averted and Couples Protected due to Family Welfare Programme in Kerala," Trivandrum: Population Research Centre (unpublished). - 50 - Table 3.15: Mid-Year Population, Births Averted, and Birth Rate, Kerala, 1961-80. Year Mid-Year No. of Hypothe- No. of Estimated Estimated Crude Popula- Births tical CBR Births Amount Birth Rate Net tion mid-year if not Averted of Decline of Effects of (000's) (Based on for Family in CBR Family Welfare CBR=38.9) Welfare Program 1 2 3 4 5 6 7 1961 170.15 678,134 39.86 3,887 0.23 39.63 1962 174.59 687,061 39.35 5,291 0.30 39.05 1963 179.03 695,987 38.88 6,902 0.39 38.49 1964 183.47 704,913 38.42 9,963 0.54 37.88 1965 187.91 713,839 37.99 16,331 0.87 37.12 1966 192.35 722,765 37.58 30,645 1.59 35.99 1967 196.79 731,691 37.18 45,834 2.33 34.85 1968 201.23 740,617 36.80 62,909 3.13 33.67 1969 205,.83 749,543 36.45 80,979 3.94 32.51 1970 210.07 758,469 36.11 93,014 4.43 31.68 1971 214.89 768,726 35.77 111,012 5.17 30.60 1972 218.61 791,913 36.22 128,277 5.87 30.35 1973 223.61 813,316 36.37 143,941 6.44 29.93 1974 228.66 834,719 36.50 168,270 7.36 29.14 1975 233.72 856,122 36.63 160,760 6.88 29.75 1976 238.14 875,432 36.76 190,870 8.02 28.74 1977 243.71 888,464 36.61 234,426 9.62 26.99 1978 246.86 901,496 36.52 243,016 9.84 26.68 1979 250.89 914,528 36.45 245,924 9.80 26.65 1980 254.80 927,560 36.40 255,622 10.03 26.37 Source: As in Table 3.14. - 51 - Table 3.16 Couples Protected by Various Methods of the Family Welfare Program, Kerala, 1957-80. Year Couples Protected Couples Protected Vasectomy Tubectomy Total IUD C.C.Users Total 1957 511 156 667 .. .. 667 1958 2,101 1,641 3,742 .. .. 3,742 1959 6,039 3,809 9,848 .. 9,848 1960 8,766 5,646 14,412 .. .. 14,412 1961 11,851 8,412 20,263 .. .. 20,263 1962 15,377 11,086 26,463 .. .. 26,463 1963 24,779 13,587 38,366 .. .. 38,366 1964 41,037 17,100 58,137 .. .. 58,137 1965 74,191 23,014 97,205 19,123 .. 116,328 1966 102,944 28,356 131,300 48,676 556 180,532 1967 146,129 37,793 183,922 62,640 5,055 251,617 1968 201,365 50,259 251,624 74,203 4,938 330,765 1969 232,622 66,498 299,120 79,432 5,987 384,539 1970 267,831 84,581 352,412 79,444 4,825 436,681 1971 327,918 105,740 433,658 69,053 3,912 506,623 1972 383,152 132,593 515,745 64,458 2,261 582,464 1973 410,785 164,115 574,900 56,717 5,899 637,516 1974 403,643 200,254 603,897 61,416 7,052 672,365 1975 403,320 243,193 646,513 59,705 8,614 714,832 1976 563,002 316,278 879,280 60,585 19,973 959,838 1977 566,640 383,726 950,366 50,550 13,331 1,014,247 1978 549,983 432,305 982,288 42,782 10,420 1,035,490 1979 534,976 511,981 1,046,957 40,697 N.A. 1,087,654 1980 519,919 586,999 1,106,918 43,193 N.A. 1,150,111 Note: N.A. - Not Available. Source: As in Table 3.14. - 52 - The Act provides that pregnancies which do not exceed 12 weeks may be terminated by one registered medical practitioner. Two medical practitioners are required if the length of pregnancy exceeds 12 weeks, but not 20 weeks. The Act also provides for the termination of pregnancy if its continuance poses a risk to Lhe Life of the pregnant woman or grave injury to her physical or-mental healTh.. As this provision includes the failure of contraceptives, MTP.may be: obtained on: request. The MTP program has been well received by the people, especially the women. From a mere 1,084 acceptors in 1972-73, the number rose to 32,597 in 1979-80, as shown in Table 3.17. During the course of eight years, over 150,000 persons have accepted MTP. Table 3.17: Number of MTP Acceptors in Kerala, 1972-80 Year Acceptors 1972-73 1,084 1973-74 4,244 1974-75 9,564 1975-76 19,969 1976-77 25,389 1977-78 1 30,834 1978-79 27,830 1979-80 32,597 Total 151,511 Source: As in Table 3.16. An examination of MTP acceptors by district over the years 1978-79 and 1979-80 indicates that the highest number has been in Trivandrum District - 28 percent of the total in 1978-79 and 22 percent in 1979-80. The three districts of Idukki, Palghat and Malappuram lag behind with less than -53- 1,000 MTP's. Canuanore District, where only 630 MTPs were done in 1978-79, had 2,473 in 1979-80. The district variations are accounted for by differ- ences in medical facilities, as well as the educational standards of the people. -54 Part IT: FERTILITY SURVEY AND RESULTS CHAPTER IV WORLD BANK FERTILITY SURVEY Introduction Our discussion in the previous chapter indicated that demographic data for Kerala are barely sufficient to describe the recent fertility trend in the state as a whole. They are inadequate for an in-depth analysis of the characteristics of the fertility decline or a detailed analysis of its determinants. To remedy this lacuna, the project collected its own data on the basis of a sample survey. Organization of the Study This study on the determinants of fertility decline in Kerala is part of a larger World Bank sponsored research project entitled "Case Studies of the Determinants of Fertility Decline in Sri Lanka and South India"(RPO 671-70). There are two other studies in the project, one in Sri Lanka which was carried out by the Department of Census and Statistics, Colombo, and the other in Karnataka carried out by the Institute for Social and Economic Change, Bangalore. All these studies have the same objective: an understanding of the determinants of fertility decline in a low-income population. They are complementary studies in the sense that a comparison of their experience would enrich the understanding of the determinants of fertility decline in each separate population. The socio-economic and family planning factors are slightly different among the three populations. The Kerala study was organized as a collaborative research project between the World Bank and the Kerala State Bureau of Economics and Statistics - 55 - (the Bureau). Within the limits set by the World Bank with respect to the obiectives of the project, the two institutions collaborated on the development and execution of the project. The survey instruments were prepared by the Bank and the Bureau. The field work was organized and carried out by the Bureau with the help of its staff and a team of field investigators recruited from outside. The cost of the study was met mostly from funds provided by the World Bank and the UNFPA, but the Bureau shared the expenses by providing vehicles, offices, and other logistic services. The research design involved the collection of survey data suitable for an analysis of the determinants of fertility decline. & brief description of the sampling procedure, questionnaire design, and other details of the survey is given below. Sampling As the main objective of the survey was a determination of the causes of the fertility decline (and not an estimation of the fertility level for the State as a whole), it was thought unnecessary to cover the entire state. Taking the budget into consideration, a decision was made to limit the study to three districts, one each from the three administrative divisions which form the present Kerala state. Palghat District was selected from the erstwhile Malabar region, Ernakulam District from the former Cochin State, and Alleppey District from the former Travancore State. The three districts were not selected at random, but on the basis of geographic location, fertility level, and field problems related to carrying out indepth interviews on family planning and related matters. (See the following maps.) Each district has its distinctive features. Palghat is the new "rice bowl" of Kerala. The economy of the district is marked by the IBRD 15930 EiR N A K IL A M 76°7 NOVEMBER 981 PAKISTAN -12° KARNATAKA 12°L BANGLADECH SHERTALAI I N D l A TAMIL ( OEZHUPUNNA NADU THAIKKATTUSERRY -1RY KERALA10°- THURAVOOR SOUTH . VAYALAR A WEST -9W 9* .. SRI LANKA fKKKOTH MAGALAM SHERTALA1 NORTH - -Internatonal boundories S75' 76- 77' ,HANNIRMUKKOM NORTH \NCIA KERALA FERTILITY SURVEY MARARIKULAM ALLEPPEY DISTRICT NORTH Survey areas ------ Survey ward boundaries AMBALAPUZHA Survey panchayat boundaries ...A Taluk boundaries ARYA D SOUTH, ----- District boundaries ALLEPPEY KAINAKARY K O TTA KA M K U T T A N A D PUNNAPRA VELIYANAD CHAMPAKULAM MALLAPPALLY RAMANKARY(*J* RAMAKART H l R U V A L-L A AMBALAPUZ HA KTAUH PURAMATTOM EZHUMATTOOR TH LAADY HIRUVALLA NEDUMPURAM HOTTAPU7ZHASSERY KADAPRA KARUVATTA PANDANAD\ THRIPPERUMTHURA MAþNAR RANMIJ.LA CHENGAN UR . Q / O ARIlPAD Qu 0/ PALLIPPAD CHERIYANAD ULAKK ZHA KARTHIGAPALLY /EZ HUVELI I MAVELIKARA KANNAMANGALMý M ELIKKARA MUTHUKULAM THUMPA MON 0 5 10 15 THEKKEKAR :A PATH1YOOR Kilometers BHARANIKKAVU PALAMEL KANDALLOOR A RALM THAMARAKULA Thts msap has been prepared by the World Bank's staff exclusively for the convenien:e K AY A KU LAM ... of the readers of the report to which It s attached. The denominations used and the . boundanes shown on th,s map do not imply. on the part of the World Bank and ts affilates., any judgment on the legal status of any terrntory or any endorsement or acceptance of such boundanes IBRD 15931 NOVEMBER 19811 T R C H U R A L W A Y E·· ·.. MALAYATTOOR ··-- NILESWARAM HENNAMANGALAM-: ANGAMALI KALADY MUNICIPALITY PARUR VENKUR -- KUZHIPPALLI MUN. MUDAKUZHA SRIMOOLANAGARAM . ERUMBAVOOR MUN. PINDIMANA EDAVANAKAD K U N N A T H U N D CHURNIKARA . KOTHAMANGALAM MUN. NJARAKKALCEA J KALAMASSERY KIZHAKKAMBALAMK THA NGAL M ELAMKUNNAPUZHA HRIKKAKKARA KUNNATHUNAD PAIPRA POTHANIKADU COCHINMAHVNU VADA'UCODU- MUVATTUPUZHA PUTENCUZPOOTHRIKA VALAKOM MN KANAYANNUR \ TRIPOONITHR MARADU MUN. THIRUMARADI PALAKUZHA • EDKKATTUVAYALKUTHATTUKULM... CHEELLANL CA '-EPPE-Y INDIA -KERALA FERTILITY SURVEY 75° 76 7 ERNAKULAM DISTRICT .. PAKISTAN KARNATAKA -12° 2 sANGLADESH Survey areas ----_ Survey ward boundaries r~ N D l A Survey panchayat boundaries TAMIL ¯ Taluk boundaries ---. District boundaries KERALA) Kilometers 0 2 4 6 8 10 12 14 16 ERFNAX6ULAM` .iesi éA'N4KUERA*.A Miles 0 2 4 6 8 10 "ýKRALA This map has been prepared by the World Banks staff exclussvely for the convemence of the readers of the report to which tt ts attached. The denomnatons used and th -9 9. - SRI LANKA boundaries shown on th.s map do not tmply. on the part of the world Bsk and ,ts affiliates, any judgment on the legal status of any terrttory or any endorsement or accptance of such boundaries. Y ---State b.undi- 75° 760 -.--Internotool bountdories INDIA KERALA FERTILITY SURVEY ... PALGHAT DISTRICT PUDUR Survey areas - Survey panchayat boundaries Taluk boundaries - - -*----District boundaries r --- State boundaries M N NARGHAT A LANAL.LUR KUNARAMPUTHUR TA M IL ARA URU Y N D MALA PPURA 4 ADAMPAZHYPURA KARIMBA SREEKRISHNAPURAM L ALLUR NKARA PAL G HAT Ll UUNDUPUUSR '- 0 TAP PA LAM MBALAPARA KERAL SERY u MUUTH L MANNUR PALGHAT VADAKARAPATHY %NGAL7RI KOTTAI IRAYIRI ELAPPULLY TTAPPALA VAN MKULUM *. .- PERUNGOT UKURI IKOZHIN A ARA LAKKADIPERUR : THIRUMITTACODE . ZHALMANNAN NALLEPALLY ERITHENPATHY H- UTHANUR KODUVAOR: TR / CHUR -THARUR CHITT . ERUMAUR A6 7' / ALATHUR PALLASNA PAKISTAN BANGLADES 12 KARNATAKA 12* A L AT H U R INDIA TAMIL NADU PALGHAT5 NADU KERALA IKERALA 0 5 10 15 20 25 . ~Kilometers m- SRI LANKA 9*9* This map has been prepared by the World Banks staff exclusvly for the convenience of the readers of the report to which it is attached, The denom,nations used and the boundaries shown on this map do not imply, on the part of the World Bank and its State boundaries affiliates, any judgment on the legal status of any teratory or any endorsement or ( --- International boundoies 7 acceptance of such boundaries 1,5* 7W 77' Il - 59 - predominance of agricultural laborers and the absence of any industry--either modern or traditional. Ernakulam District, with its modern port of Cochin and large number of modern industries, represents the emerging industrial society of Kerala. Alleppey is the old rice bowl of Kerala; its economic scene is dominated by agricultural laborers. But, quite unlike Palghat, the number of workers in traditional industries like coir (making products from coconut fiber) and fishing is substantial. This district was the center of industry, trade, and commerce of the former state of Travancore and has slipped from its prime position as a result of the emergence of Cochin as an important center on the west coast of India. Based on the available data from other sources (see Chapter III), Palghat, has relatively high fertility. The fertility level of the other two districts is low, with only slight differences between them. The performance of the family planning program in these districts based on service statistics also shows considerable variation in the rate and trend of acceptance, as well as in the mix of various methods. The household was the ultimate sampling unit. From each of these districts, a sample of 1,000 households was selected, as follows: For an earlier sample survey in the state (on Employment and Unemployment), 120 panchayat wards (100 rural and 20 urban) were selected by systematic sampling in each of the 11 districts into which the state was divided at that time 1/. For the fertility study, a sub-sample of 50 out of the 120 wards were selected, the number of rural and urban wards being in the same proportion as in the original sample. Thus 42 panchayat wards (rural) A panchayat is the smallest administrative subdivision into which a district is divided. - 60 - and eight wards of corporations and municipalities (urban) were selected from each of the three districts. From each of the 50 selected wards, 20 households were selected (the first 20 from the list of 25 households covered in the Employment Survey). This gave a sample of 1,000 households from each of the selected districts. The 1,000 selected households were expected to provide 1,000 ever- married women (EMW) for detailed investigation. During the course of the field work, however, it was found that the average number of EMW per household was 0.9. Hence, it was decided to increase the number of households per ward to cover all the 25 households previously selected for the Employment Survey. Even then, it was not possible to cover 1,000 EMW in any of the districts. Details regarding the number of households covered and the number of respondents is given in Table 4.1. Schedules Fertility and related data were collected from sample households using five schedules described below: I. Household Schedule, Part I: (8 Questions) -basic demographic information Household Schedule, Part II: (26 Questions) - information about housing, household assets, and distance to facilities II. Schedule on Ever-Married Women: - general characteristics of women (7 questions) - marriage history, attitudes (32 questions) - 2/ Not all questions were asked of every woman; the actual number depended on her status with respect to marriage, number of children, contraceptive use, etc. Table 4.1: Number of Households Covered by the Survey and Number of Respondents in Each Category Selected (S) and Interviewed (I) No. of Respondents Selected (S) & Interviewed (I) No. of House- Name of holds Covered EMW Schedule NMW Schedule NMW Schedule NMW Schedule District by the Survey (Part II) (Part III) (Part IV) (Part V) S I S I S I S I Palghat 1,011 968 907 236 209 196 176 318 293 Ernakulum 1,009 908 878 543 433 337 318 490 394 Alleppey 1,043 960 933 404 383 316 307 345 329 01 Total 3,063 2,836 2,718 1,183 1,025 849 801 1,153 1,016 Notes: EMW = Ever-married women; NMW = Never-married women; and NMM Never-married men. - 62 - - maternity history (34 questions) - contraceptive knowledge (23 questions) - fertility regulation (30 questions) - work history, attitudes about children's education, work, old age support, etc. (44 questions) husband's background (42 questions) III. Schedule on Never-Married Men (20-39 years): - socio-economic characteristics (14 questions) - attitude about marriage (13 questions) IV. Schedule on Never-Married Women (18-34 years): - socio-economic characteristics (14 questions) - attitude about marriage (13 questions) V. Schedule on Currently Married Men: - attitude about marriage (20 questions) In brief, the 300-odd questions provided a detailed picture of the demographic and socio-economic character of the household and the adult members living in it. The fertility data consisted of children ever-born by the date of birth of each event. These data also provided the date of death of each child not surviving to the survey date, thereby yielding information on infant and child mortality. With respect to family planning, the survey collected data on knowledge, ever-use, and current use of each family planning method. Information on marriage was principally the age at marriage of those who were married, and the ideal age at marriage for boys and girls according to both married and unmarried adult males and females. In addition, information was - 63 - collected on the reasons for early or late marriages, and attitudes toward marriage. Questions regarding socio-economic variables included housing, education, religion, caste and related social variables, occupation, industry, income, and so forth, of men and women. Field Organization The field investigation was carried out by a field staff consisting of 15 female and 3 male interviewers, supervised by the staff of the Bureau assigned for this work during the period of the survey. The composition of field staff was more or less similar in the three districts, as follows: 5 female interviewers (graduate, married) 1 male interviewer (graduate) 1 field editor (Bureau staff) 1 field supervisor (Bureau staff) At the headquarters in Trivandrum, the study was supervised by a full-time project officer and some supporting staff. The staff were trained in Trivandrum and in the district headquarters by Dr. V. Varma of the World Fertility Survey, London, and local staff in Trivandrum. The operational procedure for work was as follows: The Field Supervisor and the Interviewers met at the headquarters early in the morning each day and proceeded to the sample panchayat or ward fixed for the day's work. The Field Editor accompanied them on certain days. On arriving at the sample ward, the Field Supervisor (and the Editor, if present) identified the selected household. Once the correctness of the identification was ensured, one of the Female Interviewers was left in the - 64 - house to do the canvassing of the respondents. The next household was identified and another investigator was left there, etc. The Female Investigator canvassed the household using the household questionnaire, identified the eligible individuals for detailed interviewing, and administered the questionnaires for ever-married women and never-married women. The Male Interviewer visited the households later and interviewed the eligible male members. The entire field operation was supervised by the Field Supervisor. Besides identifying the selected households, he gave the Interviewers their household assignments, and assured that the completed questionnaires were returned to headquarters. Visiting all the samples in the earlier stages of the field work, the Supervisors placed an emphasis on observing the interviews as they were conducted, to assure that the correct interviewing techniques were being followed. These observations helped to raise the standard of interviews to the desired level. Later, the supervisory emphasis was shifted to verifying the correctness of the entries in the completed questionnaires by re-visiting the households and re-questioning the respondents. Findings were reported in a form designed for the purpose. A daily record sheet was kept for each Interviewer. This enabled the Field Supervisor to watch the progress of the work, locate and arrange revisits, and find cases of non-response. The completed questionnaires were received daily from the Interviewers by the Field Editor for editing. By using a list of consistency checks, the Field Editor checked the entries, located mistakes, and had them corrected. If the mistakes required a revisit to the household, the questionnaires were returned to the Interviewer for reuse. - 65 - A weekly meeting of the Field staff was arranged by the Field Supervisor on every Saturday. The progress of the field work was reviewed, with particular emphasis on completion of revisits. The errors found by the Field Editor were discussed and the program of field work for the next week was derived. The average time taken for canvassing respondents using Part I of the questionnaire was 39 minutes; Part II of the questionnaire (EMW), 86 minutes; and Part III (NMM), Part IV (NMW), and Part V (CMM) were 26, 32, and 22 minutes respectively. The field work which commenced in mid-January 1980, was completed by the end of June 1980. The post-enumeration survey was done by the Field Supervisors during the month of July 1980. Post-Enumeration Survey A post-enumeration survey (PES) was conducted by the Supervisors soon after the completion of the field work and after all the completed schedules had been received in the headquarters. The PES included 12 samples from each district and five households from each sample panchayat (60 households in each district). The sample panchayat and households were selected at the headquarters. Not all questions were fielded in the PES; the information collected included number of persons in the household, their age, marital status, age at marriage, number of children ever-born, dates of birth and of death (if any) of each child, current use of contraceptives, current employment, and a few other items. The results of the PES indicated that the data were of good quality. For example, 50 percent of the EMW reported the same age in the PES and in the original interview, 72 percent reported with a margin of error of - 66 - one year or less; 90 percent reported within a range of two years. Age at marriage was less accurate: 41 percent reported the same age, 66 percent within one year, and 85 percent within two years. Nearly three-fourths of the women reported their school grade accurately and fot only 3 percent of the women did the two reports differ by more than one year. Similarly, high reliability was noted with respect to children ever-born, current family planning practice, and other inquiries. - 67 - Part II: FERTILITY SURVEY AND RESULTS CHAPTER V THE FERTILITY TREND Information on fertility levels and trends is available from two questions in the survey: that regarding the age distribution of the household, and that regarding children ever-born by date of birth of each child, age of woman, and her age at marriage. Crude Birth Rate The age distribution of the household population can be manipulated to derive crude birth rates for recent years. The observed age data are strongly suggestive of a sharp decline in the birth rate in all the survey districts. Table 5.1 gives the proportion of the population in the three youngest age groups compared with a stable age distribution. The percentages are relatively low, indicating a fairly low fertility rate, close to a Gross Reproduction Rate (GRR) of 1.5. 1/ Table 5.1: Percent of Population in the Younger Age Groups by District, Kerala, 1980 Stable age Age Palghat Ernakulam Alleppey Total distribution/1 0- 4 11.5 9.4 8.6 9.8 12.5 5- 9 12.8 9.9 10.1 10.9 11.1 10-14 13.7 12.3 13.1 13.0 10.1 0-14 38.0 31.6 31.8 33.7 33.7 All ages 100 100 100 100 100 1/ Stable age distribution features a GRR of 2.0 and a mortality level of 18 (see Coale and Demeny, 1966). 1/ The -Gross Reproduction Rate is the average number of daughters a woman would have if she passed through her childbearing-years conforming to the age-specific fertility rates of a given year. - 68 - The proportion of population aged 0-4 years is lower than that aged 5-9 years, which in turn is lower than that in the 10-14 year age group. If the fertility pattern had been constant, the trerd would have been the opposite-- with each older age group showing a lower proportion of the total population (see the last column in the table). The trend of a decreasing proportion of the population with each successive age group among those aged below 15 years is a clear indication of a decline in the birth rate during the 15-year period prior to the survey. How great is the decline in the birth rate, and what is its current level? A simple calculation of the birth rate by the reverse survival method gives some approximate answers. The results are shown in Table 5.2. Table 5.2: Crude Birth Rate, Three Districts in Kerala, 1965-80 (per 1,000 population) Three Districts Palghat Ernakulam Alleppey Combined Period (1) (2) (1) (2) (1) (2) (1) (2) 1965-70 41.1 41.7 36.9 34.0 40.1 36.1 39.3 37.0 1970-75 33.8 33.7 26.2 25.2 26.5 25.0 28.7 27.8 1975-80 26.8 26.6 21.8 21.3 20.0 19.5 22.8 22.5 (1) Reverse survival method using a growth rate of 2 percent and mortality levels of (South model) 17.2, 18, and 18.8. (2) Reverse survival method .(all ages) using mortality alone (same as above). Estimates by this method are considered more reliable in this case. For the period 1975-80 the crude birth rate for the three districts together was only about 22.5 per 1,000 population, a rate supported by both the methods. Additional credibility to the rate is obtained by comparing the survey rate with-the Indian Government's Sample Registration System (SRS) rate, which is - 69 - entirely independent source of data for births. 2/ This comparison is given in Table 5.3, and graphically seen in Figure 5.1. There is little doubt that the crude birth rate of Alleppey, Ernaku- lam, and Palghat Districts of Kerala State is relatively very low. Their average, about 23 births per 1,000 population, is approximately the same as the birth rate of South Korea over the same period (see Table 5.4). It is slightly lower, however, than the rate for Kerala State as a whole. The SRS indicates that the state rate is about 8.5 percent higher. 3/ On this basis the Kerala birth rate estimated from the survey data should be about 25 births per 1,000 population. Child-Woman Ratio The ratios calculated by relating children 0-4 years of age to women 15-44 years, children 5-9 years of age to women 20-49 years, and children 10-14 years of age to women 25-54 years are given in Table 5.5. These are rough fertility indices for the periods 1975-80, 1970-75, and 1965-70 respectively. In 1975-80, the ratio is lowest for Alleppey and highest for Palghat, which is similar to the pattern shown by the crude birth rate. However, the degree of the difference is larger for the child-woman ratio than for the crude birth rate. 2/ See Government of India, Vital Statistics Division, Ministry of Rome Affairs, 1975-80. "Sample Registration: Bulletin", Volumes for 1975-80, New Delhi: Office of the Registrar General. 3/ That is, by comparing the SRS rate for Kerala with that of the three districts from the SRS source. - 70 - Table 5.3: Comparison of Survey and SRS Birth Rates, 1975-80 Survey SRS Rate Rate District (1975-80) (1978) Difference Ralghat 26.6 26.9 -0.3 Ernakulam 21.3 20.9 0.4 Alleppey 19.5 21.4 -1.9 Three districts 22.5 23.5 -1.0 Estimate for Kerala State 24.4 25.5 -1.1 Table 5.4: Crude Birth Rate, Kerala and Selected Countries.in Asia Birth Country Rate Period 3 districts of Kerala 22.5 1975-80 Kerala State 24.4 1975-80 Singapore 17 1978 Hong Kong 19 1978 South Korea 21 1978 Taiwan 21 1978 Sri Lanka 26 1978 India 35 1978 Note: Estimates for other countries taken from The World Bank, 1980. World Development Report, 1980, New York: Oxford University Press for the World Bank, Table 18. Figure 5.1: Crude Birth Rate of the Three Districts of Kerala, 1960-79 CBR I 46 4-46 Unadjustod 44 444 5-year moving averages SRS! Kerala : 40 .. O 36 -6 32 32 28 28 24 - 20 Source: "Kerala Fertility Survey 1980", sponsored by the World!Bank, UNFPA, and the ; 20 Bureau of Economics and Statistics, Trivandrum. 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 Year - 7'2 - The overall trend of the child-woman ratio is also similar to that of the crude birth rate. The child-woman ratio (three districts together) has declined by 44 percent compared to a decline of 39 percent in the crude birth rate over the period- 1965-70 to 1975-80. Similarly, in each of the. three districts, the decline in the child-woman ratio was.larger than the decline in thee crude birth rate. Table 5.5: Child-Woman Ratio: Three Districts of Kerala Children's Women-s Age Age Child-Woman Ratio (per 1,000) Period Group Group Palghat Ernakulam Alleppey 3 Districts 1975-80 0- 4 15-44 537 426 370 441 1970-75 5- 9 20-49 699 546 507 5§0 1965-70 10-14 25-54 822 799 742 787 As Percent of 1965-70 Ratio 1975-80 65 53 50 56 1970-75 85 68 68 74 1965-70 100 100 100 100 The crude birth rate and the child-woman ratio are not entirely independent of each other. They use roughly the same data. A somewhat more independent measure is the fertility rate calculated from children ever-born. Children. Ever-Born, by Age of Woman In the survey, ever-married women were asked the date of birth of each of their children in addition to their own dates of birth and marriage. The average number of children ever-born by age derived from this data is shown in Table 5.6. The average at ages 45-49 years is the completed fertility or the (cohort) Total Fertility Rate (TFR) of women born during 1930-35. - 73 - Table 5.6: Average Number of Children Ever-Born (Parity) by Age of Woman, Kerala Average Number of Children Ever-Born Age of 3 Districts Woman Palghat Ernakulam Alleppey Combined 15-19 0.53 0.44 0.23 0.47 20-24 1.31 1.45 1.21 1.22 25-29 2.36 2.20 2.06 2.21 30-34 3.32 3.04 3.05 3.15 35-39 4.50 3.60 4.27 4.11 40-44 4.74 5.37 4.73 4.93 45-49 5.37 5.27 5.22 5.28 Average for all ages 3.39 3.22 3.46 3.36 Standardized average 3.56 3.38 3.37 3.37 The completed fertility of women in the three districts of Kerala is 5.3. The variation between the districts is not very large. It is slightly higher in Palghat and lower in Alleppey, the same pattern of differences shown by the crude.birth rate. A similar pattern of difference is shown by the standard- ized average parity (the unstandardized average shows a different pattern). Thus, these parity data indicate that in the past, the TFR in Kerala was at least 5.3; that among the three districts studied, the fertility level has been highest in Palghat and lowest in Alleppey; and that the fertility differentials between the three districts have been relatively small. -74--' Children Ever-Born by Duration of Marriage All ever-married women in the survey were classified by their dura- tion of marriage and children ever-born. The average number of children in each duration group is given in Table 5.7. The completed fertility rate is slightly higher as the denominator does not include the unmarried women, Table 5.7: Average Number of Children Ever-Born by Duration of Marriage, Kerala Duration of 3 Districts Marriage Palghat Ernakulam Alleppey Combined (years) 0- 4 0.54 0.72 0.56 0.65 5- 9 1.75 1.89 2.02 1.90 10-14 2.88 2.75 2.78 2.81 15-19 .3.96 3.83 3.81 3.87 20-24 5.26 4.56 4.77 4.86 25+ 5.81 5.32 5.66 5.58 All 3.66 3.45 3.31 3.48 Standardized 3.57 3.36 3.48 3.48 but the pattern of variation by age and inter-district variation are some- what similar. The standardized average parity is highest in Palghat and lowest in Ernakulam. An average married woman has a completed fertility rate of 5.6 children. Fertility Rates Information on the date of birth of each child makes it possible to estimate specific fertility rates for different periods of time. The rates for the three most recent 5-year periods are given in Table 5.8. The TFR for 1975-80 is estimated to be about 3.685 which is higher than the rate in Sri Lanka, Singapore, Taiwan, Hong Kong, or South Korea for recent periods. - 75 - Table 5.8: Specific Fertility Rates for Recent Periods, Three Districts Combined, Kerala (per 1,000 population) Age Percent Decline Group 1965-70 1970-75 1975-80 1965-70 to 1975-80 15-19 52 37 22 58 20-24 269 219 176 35 25-29 313 253 235 25 30-34 242 200 146 40 35-39 189 121 100 47 40-44 (93) 70 47 49 45-49 (23) (17) 11 52 TFR 5.905 4.585 3.685 38 Note: The rates in,parentheses are very rough estimates based on trends in previous age groups. Specific fertility rates for 1975-80 together with the household age-sex composition yield an estimate of the crude birth rate of 25.3 which is higher than the estimate obtained by the reverse survival method or the SRS rate (see Tables 5.2 and 5.3). This indicates that the SRS under-enumerates 7 percent of the births in Kerala. Marital Fertility Rates The specific fertility rates given above were derived from marital fertility rates and proportions married. The marital fertility rates themselves are given in Table 5.9. For the period 1975-80 the Total Marital Fertility Rate (TMFR) is 5.71. The difference between the TMFR and TFR (5.71 - 3.68) of about two children per woman is the effect of postponement of marriage (all women were not married at age 15 years). - 7& -- Table 5.9: Age-Specific Marital Fertility Rates, Three Districts Combined, Kerala (per 1,000 population) Percent Age 1965-70 1970-75 1975-80 Decline 15-19 201 206 200 0.5 20-24 375 334 333 11.2 25-29 344 292 280 18.6 30-34 255 213 162 36.5 35-39 198 127 106 46.5 40-44 (99) 74 49 51.5 45-49 (24) (18) 12 50.0 TMFR 7.480 6.320 5.710 24 Fertility Patterns Figure 5.2 compares the age pattern of the fertility curve for Kerala with that of a number of other countries (World Fertility Survey data). In Kerala the fertility rate is maximum at ages 25-30 years. The difference between the rate at 25-30 years and those in the neighboring ages is substan- tial. This is in sharp contrast with the Sri Lanka pattern which is more or less flat at ages 20-25, 25-30 and 30-35 years. Even in the age group 35-40 the Sri Lanka rate is not far below the peak value. In the younger age groups (15-19 and 20-24), the pattern in Kerala is similar to that in Sri Lanka, Hong Kong and Singapore, but contrasts with that in Indonesia. At older ages, however, the Kerala pattern is quite similar to that in Indonesia, but different from that of Singapore and Hong Kong. In fact, at ages above 30 years the curve for Kerala and Indonesia are identical. This is somewhat surprising in view of the differences in the fertility levels of these populations (TFR 4.97 77 - Figure 5.2: Fertility Patterns in Kerala and Selected Asian Countries Kerala, 1978.(3.69) Hong Kong, 1970 (3.42) .2 __--_------------Singapore, 1978 (1.84) -- --- Indonesia, 1974 (4.97) 11 Sri Lanka, 1974 (3.361 40 --7- -·10 25 30 35 40 45 Note: Percent distribution of specific fertility rates by age. Source: World Fertility Survey data and the "Fertility Survey in Kerala, 1980", sponsored by the World Bank, UNFPA, and the Bureau of Economics and Statistics, Trivandrum. BEST COPY AVAILALE - 78 - in Indonesia and 3.68 in Kerala). The differences and similarities should be attributable to the causes of the fertility decline. Age at marriage, which plays a principal role in fertility decline at younger ages, is similar in Kerala and Sri Lanka. At older ages, family planning practice is the immediate cause of fertility decline. In this respect, Sri Lanka and Kerala differ considerably, both in the extent of use and the methods used. Hence, the age patterns of fertility in Kerala and Sri Lanka are quite similar at younger ages, but very divergent at older ages. The similarities between Indonesia and Kerala at ages above 30 years should be related to the similarity of the extent of family planning practice. Fertility Trends The TFR declined from 5.905 in 1965-70 to 3.685 in 1975-80. This is a decline of 2.2 children per woman, or 38 percent from the initial value during a ten-year period, averaging a 32 percent decline per year. This is an extremely rapid decline, but by no means unprecedented. The TFR declined by 5.9 percent per year during 1965-70 in Singapore, and by 5.4 percent per year in Hong Kong during the same period. These are small populations, however, compared to Kerala. In South Korea, which is larger than Kerala and has a strong family planning program, the rate of decline was 2.5 percent per year during 1965-76. In Sri Lanka, fertility declined by 3.0 percent during 1965-74. In Taiwan, the corresponding rate was 2.8 during 1965-78. Thus, a reduction of the TFR by about 3.2 percent per year for a large population like Kerala's is not unprecedented but noteworthy. Fertility in Kerala declined fairly uniformly during the period 1965-80. The percent decline was 22 during 1968-73 and 20 during 1973-78. - 79 - The decline has not been uniform across age and socio-economic groups, however. As Table 5.8 indicates, the fertility decline was smallest in the 25-29 age group and highest in the younger and older age groups. In the younger age group, the fertility decline was linked with the marriage factor to a great extent, and with family planning to a much lesser extent. At the older ages, the decline is almost entirely due to the family planning factor (see Chapter VII on Marriage). Figure 5.3 gives a comparative picture of the fertility changes between 1965-70 and 1970-75, and between 1970-75 and 1975-80. On the whole, the decline was slightly greater during the first 5-year period than the second 5-year period (1.3 children compared to 0.90 children). Except for those aged 30-34 years, the remaining age groups reflect this pattern. Thus, the decline due to family planning (at older ages) seems to have permeated the lower age groups in more recent years. Trends in Specific Fertility Rates Specific fertility rates can be calculated at five-year intervals from 1945-50 onwards for ages 15-19 years, from 1950-55 onwards for ages 20-24 years, etc. These are given in Table 5.10. They are also shown in Figure 5.4, along with the trend in marital-specific fertility rates. The specific fertility rates show a continuous decline in each age group from 1960 onwards. The trend in earlier periods is uncertain; the rates in Table 5.10 show some increase up to 1960-65. This could well be the real situation. On the other hand, the data for these periods may suffer from recall lapse and the observed increase could be due to data deficiencies. Therefore, we cannot give much importance to these increases. - 80 - Figure 5.3: Period-Specific Fertility Rates, Three Districts of Kerala, 1965-70, 1970-75 and 1975-80 Specific Fertflity : Rate .400 300 Fertility decline during 1968-72 .200 Fertility Deeline .957 During 1972-78 1970-75 *100 1975-80 15• 20 25 30 35 40 45 50 Age Source: The "Kerala Fertility Survey, 1980", sponsored by the World Bank, UNFPA, and the Bureau of Economics and Statistics, Trivandrum. Figure 5.4: Trend in Specific Fertility Rates, 1945-50 to 1975-80, Three Districts of Kerala .. ....iS i. ...1 .... ..i1 .. .. ... ~ ..... .. p ific Førti~lit t . ... All Wøen I Spge .c r t Itty am t -U r 1 -t.> L.å ar rpi i , .. i.....* Spe'itIc st9 lL rn c 9t5g80` i of 1975-0 'ite f r th i gi . fr s pi art - - r-- - - -r 300 0 . ... ... .. 20- ---- - - ------ 220 --- ~- -9 2 100 415 5 0 6(- i 6-7 ( -5 544 ,5 5 .5 i5-60 6 6ý 75 71-7 'tea s Y P.Ar, -. .. Source: "Kerala Fertility Survey 1980", World Bank, UNFPA, and Bureau of Economics and Statistics, Trivandrum. BES-. 00. I åLABLE Table 5.10: Age-and Period-Specific Fertility Rates for Five-Year Intervals from 1945 to 1980, by Five-Year Age Groups, Kerala Age of Women at time of birth of Age-Specific Rates (per 1,000) children 1975-80 1970-75 1965-70 1960-65 1955-60 1950-55 1945-50 15-19 22 37 52 62 93 76 82 20-24 176 219 269 302 301 269 -- 25-29 235 253 313 316 299 -- -- 30-34 146 200 242 260 -- -- -- 35-39 100 121 189 -- -- -- - 40-44 47 70 -- -- -- -- -- 45-49 11 -- -- -- -- -- -- Between 1960-65 and 1975-80, the specific fertility rates declined by 65 percent at ages 15-19 years, 42 percent at ages 20-24 years, 26 percent at ages 25-29 years, and 44 percent at ages 30-34 years. The decline is minimum at ages 25-29 years. The marital fertility rates show much smaller decreases. At younger ages there is no decline at all; the observed increases at earlier years could be statistical aberrations. At older ages, age-specific fertility rates and age-specific marital fertility rates behave in a similar manner. One special point to note is that the rates for 1960-65 and 1965-70 are practically the same in the age groups for which they are available. It would be fairly accurate to conclude that there was little decline in marital fertility before 1960. Cohort Fertility Fertility rates for all cohorts are shown in Figure 5.5. A word of caution is necessary in interpreting these figures. Cohort fertility is - 83 - complete only when women have passed through the entire childbearing period. In our sample, only one cohort has completed the childbearing period, the cohort born in 1930-35. For all the other cohorts, the experience is partial. At one extreme is the group of women born during 1960-65, that is those who were 15-19 years at the time of survey. For them, their fertility experience is measured for a maximum of five years since much of their reproduc- tive period is still ahead of them. We have assumed that the period fertility rate of women during 1975-80 will apply without change to all women who have. not completed their fertility period. This assumption clearly over-estimates the fertility rates of recent cohorts (as future fertility is likely to be lower) and thus under-estimates the fertility differences between older and younger women. A few firm conclusions emerge from this Figure. First, completed fertility has decreased consistently with increasing year of birth. It is highest for those born in 1930-35 and lowest for those born in 1960-65. Second, the actual differences between the various cohorts are likely to be more than what is shown in Figure 5.5, as cohort fertility of recent cohorts would be lower than that which the figure indicates. Third, there is a very substantial difference between the fertility performance of women born before 1945-50 and those born after that period. The year of Indian Independence (1947) seems to be the great divide. Fertility decline in Kerala started with the cohort of women born around the year of Independence. The decline was not, however, confined to these women. Those born earlier were also affected by the trend and contributed to the fertility decline when the 1945-50 birth cohort came to reproductive age (i.e., after 1960). Fourth, the year of marriage, rather than the year of birth, could be more relevant in fertility trends. For example, those who were born in 1945-50 would be 15-19 years in 1960-65. This - 84 - Figure 5.5: Cumulative Cohort (Birth) Fertility Rates, 1930-35 to 1960-65, Kerala TFR 6. . . . 1930-35 -- -1935-40 - -----1940-45 .5 ... .. 1945-50 4 -.- - -1950-55 - 955-60 1960-65 3 Observed - - - - Estimated : 15-19 20-2,4 25-29 30-34 35-39 40-45 45-49 Age Source: "The Kerala Fertility Survey,1980", sponsored by the World Bank, UNFPA, and Bureau of Economics and Statistics, Trivandrum. - 85 - is the age group in which most of the women got married at that time. It is also the time when land reforms introduced by the Communist government were implemented, and large scale efforts in the family planning program were initiated. Hence, those who were married in 1960-65 were the first marriage cohort affected from the very beginning of their married life by the vastly expanded family planning programs and the social changes accompanying land reform, as well as other legislative efforts by the Communist government. Therefore, fertility rates of this cohort and those following it are different from those of the previous cohorts. Fifth, the trend in incomplete fertility rates is not as systematic as the trend in completed fertility. Those born in 1935-40 had higher fertility rates at ages 20-24 years and 25-29 years than those born in 1930-35. Similarly, those born in 1940-45 had higher .fertility at ages 15-19 years than those born in 1935-40 or 1930-35. Whether these trends are real or mere statistical aberrations due to errors in the data or a small sample size is uncertain. Fertility by District Earlier we have seen that the 1975-80 birth rate (per 1,000 population) was 26.6 in Palghat, 21.3 in Ernakulam, and 19.5 in Alleppey. The child-woman ratio gave the same type of differentials: 822, 799, and 742 (per 1,000 women) respectively for the three districts. The standardized average parity and completed fertility rate (Table 5.6 and Table 5.7) showed a somewhat similar pattern. Palghat has the highest fertility and Alleppey has the lowest. The average for Ernakulam is slightly higher than Alleppey in Table 5.6, but slightly lower in Table 5.7. Table 5.11 gives period fertility rates for the three most recent 5-year periods. The pattern revealed by these rates is not very much - 86 - Table 5.11: Marital Fertility Rates for Palghat, Ernakulam, and Alleppey 1975-80 Age Palghat Ernakulam Alleppy 20-24 303 348 370 25-29 263 307 281 30-34 201 143 146 35-39 145 88 90 40-44 80 44 30 TMFR' 4.960 4.650 4.585 1970-75 20-24 326 368 310 25-29 289 295 294 30-34 232 206 203 35-39 161 140 90 40-44 88 67 67 TMFR' 5.480 5.380 4.820 1965-70 20-24 357 385 387 25-29 309 368 358 30-34 252 280 238 35-39 214 197 184 40-44 (117) (94) (137) TMFR' 6.245 6.626 6.520 Period Percent Decline 1968-73 12 19 26 1973-78 9 14 5 1968-78 21 30 30 different from what was seen in the previous paragraphs (the data are not independent). TMFR' 4/ is highest in Palghat and lowest in Alleppey. This was not the case 10 to 15 years ago when Ernakulam had the highest rate and Palghat the lowest. 4/ TMFR' = Total Marital Fertility Rate for ages 20-44 years only. - 87 - The overall fertility decline has been higher in Alleppey and Ernakulam (about 30 percent each) than in Palghat (21 percent). Much of the decline in Ernakulam and Alleppey took place in the early part of this period, especially in Alleppey, where the decline during 1968-73 was 5 times that in 1973-78. Similarly, much of the decline took place in the older ages, rather than the younger ages. Thus, at ages 20-29 years, fertility declined by 15 percent in Palghat and 13 percent each in Ernakulam and Alleppey. At ages 30-44 years, however, fertility declined by 27 percent in Palghat and 52 percent each'in the other two districts. Thus, the differentials in fertility decline between the districts were actually differentials in the decline at older ages. Pre- sumably, this finding relates to differentials in the impact of family planning and sterilization. Fertility Differentials and Trends by Socio-Economic Groups One method of understanding the determinants of fertility changes in a population is an examination of fertility differentials at a point in time or the trends in fertility differentials over a period of time. This is what is attempted here. Differentials are examined on the basis of cohort fertility (marriage cohort) and period fertility, and trends in differentials are examined on the basis of per.,_ fertility calculated for the various socio-economic groups. Fertility measures in this section are all for married women only. As the classification of the general population by socio-economic group and marital status is not available, it is not possible to convert the marital fertility rates to rates for the total population. For comparative purposes, however, we are reviewing marital fertility instead of the fertility of all women. - 88 The principal objective of our analysis will be to locate the groups which have high or low fertility rates and those groups which have shown a more than average or less than average change in fertility rates in recent years. Together with data on family planning use by these group, it should be possible to draw inferences about the determinants of fertility change in Kerala in recent years. Age At Marriage: Those who marry early have more children. This is shown in Table 5.12. On average, women who were married at ages 23 years or above have only 2.0 children compared with 5.0 for those who were married at ages below 15 years. The difference is 3.0 children per woman. Table 5.12: Mean Number of Children Ever-Born by Duration of and Age at Marriage, Kerala Duration Age At Marriage Sample (Years) (Years) Row <15 15-16 17-18 19-20 21-22 23+ Totals 0-4 .63 .50 .63 .68 .71 .67 373 5-9 1.70 1.91 1.91 2.02 1,92 1.66 450 10-14 3.00 2.90 2.84 2.63 2.92 2.68 443 15-19 4.46 4.08 3.96 3.68 3.68 3.03 370 20-24 6.16 5.05 5.16 4.35 4.22 3.56 390 25+ 5.99 5.73 5.41 5.35 4.76 3.44 653-_ Total 5.02 4.27 3.59 2.92 2.61 2.04 2,679 Standardized 3.84 3.57 3.49 3.31 3.19 2.59 Part of this difference is due to compositional factors. Those who marry early have longer periods of married life. The standardized - 89 - average (with respect to distribution of women by duration of marriage) varies from 3.8 for the early marrying women to 2.6 for the late marrying women. The difference is narrowed by standardization to 1.2 children per woman. That there is a real difference in the fertility performance of women who marry early and those who marry late is very clear by examining the averages within each category of marital duration. For duration 0-4 years, the variation is not systematic. At higher durations, consistent differ- ences appear. Thus, at duration 15-19 years, the averages are 4.5, 4.1, 4.0, 3.7 and 3.0. The differentials become more systematic as duration increases. From this table, it is difficult to answer whether age at marriage determines fertility, or whether the fertility preference of a woman has helped decide her age at marriage. Those who married at below 15 years are all older women. In the sample, there were only four cases with duration of marriage, less than five years. Similarly, those who married at ages above 23 years are all younger women. There were only nine women who were married at ages above 23 years of age 25 years ago. These nine women must have been a highly selective group with respect to their fertility preference. Had they married younger, there is no guarantee that they would have had more children than they actually produced. Education: The average number of children ever-born is relatively low for the better educated than for the illiterate (2.1 for women with ten or more years of school and 4.5 for women with no schooling) (Table 5.13). Part of the difference is due to compositional factors, such as duration of marriage. If the averages are standardized for marital duration, the difference decreases, from 2.4 children before standardization to 0.8 children after standardization. - 90 - Table 5.13: Mean Number of Children Ever-Born by Years of Schooling of Mother and Duration of Marriage, Kerala Duration Sample of Row Marriage None 1-4 5-9 10+ Totals 0-4 .53 .63 .67 .71 373 5-9 1.80 2.01 1.93 1.64 450 10-14 2.87 2.80 2.90 2.53 443 15-19 4.00 4.09 3.82 2.83 370 20-24 5.41 4.85 4.57 3.67 390 25+ 5.83 5.71 5.12 4.40 653 Total 4.49 3.80 2.94 2.13 2,679 Standardized 3.61 3.55 3.34 2.80 The completed fertility (children ever-born to women with 25 years of marital duration) of the highly educated is an average of 1.4 children less per woman than their counterparts with no schooling (4.4 and 5.8, respec- tively). The decrease in the average number of children with increasing education of the woman is very systematic in our sample. Much of the variation in fertility levels by educational attain- ment is due to differentials-in age at marriage, as Table 5.14 indicates. The expected differential (5.8-5.0) is smaller than the actual (5.8-4.4). There- A fore, marital fertility rates are different by educational group. Age at marriage and fertility levels within marriage contribute to overall fertility differentials by educational attainment. - 91 - Table 5.14: Distribution of Married Women by Age at Marriage and Educational Attainment, and Standardized Fertility Rates, Kerala Cumulative Age at No Marital Marriage Education 1-4 yrs. 5-9 yrs. 10+ All Fertility <15 132 98 75 4 309 6.02 15-19 359 511 553 92 .1515 6.02 20-24 108 183 294 124 709 5.04 25-29 15 20 41 49 125 3.43 30-34 2 7 5 5 19 2.09 35-39 - 1 1 - 2 1.05 40-44 - - - - 45-49 - - - - - Total 616 820 969 274 2679 Actual 5.8 5.7 5.1 4.4 5.6 Expected 5.8 5.7 5.6 5.0 5.6 Fertility rates by educational attainment for the three recent periods are summarized in Table 5.15 and given in greater detail in Table 5.16. Table 5.15: Total Mariial Fertility by Education, Kerala No Education 1-4 yrs. 5-9 yrs 10+ yrs. 1965-70 6.2 6.7 6.3 5.0 1970-75 5.4 5.1 4.6 5.7 1975-80 4.8 4.4 4.7 4.1 Percent Decline 68-73 13 24 27 -14 73-78 11 14 -2 28 68-78 23 34 25 18 - 92 - Table 5.16: Marital Fertility Rates by Age and Educational Attainment, Kerala Age No 1-4 yrs. 5-9 yrs. 10+ yrs. Group Schooling of school of school of schools 1975-80 20-24 266 335 354 299 25-29 266 263 306 254 30-34 210 155 143 149 35-39 154 77 94 94 40-44 63 52 38 19 TMFR 4.795 4.410 4.675 4.075 1970-75 20-24 353 322 330 354 25-29 279 297 286 339 30-34 240 195 202 228 35-39 141 130 106 136 40-44 77 80 55 94 TMFR' 5.450 5.120 4.620 5.755 1965-70 20-24 350 400 384 281 25-29 319 357 373 271 30-34 259 257 241 294 35-39 227 201 157 133 40-44 ( 92) (136) ( 97) ( 27) TMFR' 6.235 6.755 6.260 5.030 TFR' = Total Marital Fertility for ages 20-44 years. - 93 - There are significant differentials and trends in these data, and not all are according to expectation. The overall pattern is that fertility has fallen among women of all educational levels but the highest decline has taken place among those with 1-4 years of schooling and the smallest decline among the highly educated. There could be two main reasons for this pattern. First, fertility has been highest in the group with 1-4 years of schooling, and the scope for decline among them was, therefore, considerable. Second, as will be seen later, sterilization has been an important factor in fertility declines in Kerala and the propensity to obtain the sterilization has been greatest among those with 1-4 years of schooling. Another important aspect of this table is the fertility trend among the illiterate. Their fertility was lower than that of either the 1-4 or 5-9 years of schooling group in 1965-1970. It is possible that such a pattern existed in earlier periods. Normally for such groups, fertility rises before falling. But the trend shown in the table is clearly downward. This reflects the fact that the potential fertility increase in this group due to improvements in health and nutrition was avoided through other means--such as the birth control program, particularly the acceptance of sterilization. One might suggest that without an active family planning program featuring an incentive system, it is very unlikely that an increase in fertility among the women who had never gone to school might have been avoided. The consistent downward trend in fertility among the scheduled castes and among the economical- ly backward groups of the society (see later sections) reflect what may well be similar impacts of the family planning program, made attractive to the poorer section of the population through financial incentives. -94- The differentials in specific fertility rates by education have not been the same in all age groups. Below 30 years of age, fertility seems to rise with increased schooling, especially at low levels of education. At ages above 30 years, the pattern is one of a systematic decline in fertility with an increase in schooling. Caste and Religion: Marriage customs vary considerably from caste to caste in Kerala (see Chapter VII). Similarly, living arrangements for the married couple vary by religion and caste group. These differences may have a significant effect on the number of children born to a woman, or her fertility. The available data indicate such a variation. The average parity of the Muslims was 4.1 while that of the Nairs was only 2.9, a difference of 1.2 children per woman. The completed fertility of these two groups differs by 2.5 children. Between Nairs and Latin Christians, the difference in completed fertility is 2.7 children (Table 5.17). Standardization with respect to duration of marriage does not eliminate all the intercaste differences. Muslims have the highest average number of children (4.2), and Nairs have the lowest (2.9). Table 5.17: Mean Number of Children Ever-Born by Duration of Marriage and Caste, Kerala Caste Dura- Scheduled Syrian Latin Sample tion Caste/Tribe Nairs Ezawas Xians Xians Muslims Row (years) Totals 0-4 .42 .73 .63 .86 .77 .51 373 5-9 1.76 1.64 1.99 2.03 2.20 2.12 450 10-14 2.69 2.67 2.71 2.89 2.88 3.44 443 15-19 4.18 3.00 3.85 3.97 4.09 4.53 370 20-24 5.13 4.39 4.31 4.93 5.47 6.35 390 25+ 5.70 4.39 5.52 5.35 7.09 6.91 653 Total 3.56 2.86 3.38 3.58 3.56 4.07 2,679 Standard 3.47 2.94 3.37 3.51 4.04 4.23 - 95 - Intercaste differentials in incomplete fertility are not always consistent with the pattern of differentials of completed fertility. For example, at duration 10-14 years, Syrian Christians and Latin Christians have more or less the same average parity, but among older women (duration 25+) their average parities differ by 1.7 children. A similar comparison holds between Nairs and Scheduled Castes. Thus, the fertility trend has not been the same in all the castes: the differentials which existed in the past have undergone major change and new types of differentials have come into being. Table 5.18 gives a rough estimate of specific fertility rates of married women by age and caste for the three most recent 5-year periods (because of small numbers involved in some age groups for some castes, and a heavy dependence on the accuracy of the reported timing of the birth, these rates should be interpreted with some caution). According to this table, the Nairs have the lowest fertility rates in 1965-70 as well as in 1975-80. The Christians and the Muslims had the highest fertility in 1965-70 and the highest today. In between are the Ezawas and Scheduled Castes. In terms of fertility decline, the Ezawas lead in the amount (1.96 children per woman) and percent of decline (31 percent), and Nairs were the last (0.75 children per woman, and 15 percent in 10 years). On the whole, the &v-efrall decline has been similar in all castes and religious groups (except the. Nairs), with a decline of about 1.1 children during 1968-73, and 0.6 children. during 1973-78. The pattern of differences and trends at the older ages are indica- tive of the effect of deliberate fertility control through family planning and sterilization. In the 35-39 age group, the Ezawas have the highest - 96 - Table 5.18: Marital Fertility Rates by Age and Caste, 1965-70 to 1975-80, Kerala Age Nairs Ezawas Scheduled Christians Muslims (years) Castes 1975-80 20-24 412 335 276 .370 216 25-29 202 268 280 327 286 30-34 103 147 183 166 262 35-39 87 79 120 110 149 40-44 33 40 70 42 87 TMFR' 4.185 4.345 4.645 5.975 5.000 1970-75 20-24 291 324 309 330 384 25-29 255 304 283 330 290 30-34 195 183 220 217 264 35-39 79 130 126 147 174 40-44 34 76 76 97 74 TMFR- 4.270 5.085 5.070 5.605 5.930 1965-70 20-24 332 382 345 423 377 25-29 297 342 358 385 328 30-34 190 247 252 278 316 35-39 122 193 202 197 205 40-44 ( 46) ( 98) (118) (75) (120) TMFR' 4.935 6.310 6.375 6.790 6.730 decline (0.114 in the specific fertility rate), followed by the Christians and Scheduled Castes (about 0.085), Muslims (0.056) and Nairs (0.035). The low decline among the Nairs is attributable partly to their relatively low rates in the beginning of the period. To conclude, there have been considerable inter-caste fertility differentials occurring in Kerala. The differentials have narrowed substantially - 97 - in recent years as the CAste groups with original higher fertility have expe- rienced larger fertility declines since 1965. All the caste groups indicate the impact of birth control practice, either sterilization or contracep- tion, but some, such as the Ezawas, show a greater impact than others, such as the Nairs. Per Capita Household Expenditures: Per capita household expenditures have a very consistent and systematic negative relationship with fertility. As the per capita household expenditures increase, from Rs 30 or less to Rs 120 or more, the number of children per woman in households decreases from 4.0 to 2.6. A similar pattern is observed with respect to completed fertility (Table 5.19). Among women with at least Table 5.19: Mean Number of -Children Ever-Born by Duration of Marriage and Per Capita Household Expenditures, Kerala Duration Per Capita Expenditures of Sample Marriage Less than Row (years) Rs. 30 Rs. 30-59 Rs. 60-120 Rs. 120+ Totals 0-4 .80 .63 .67 .63 373 5-9 1.80 1.99 1.91 1.49 450 10-14 3.00 3.07 2.73 2.11 443 15-19 4.42 4.22 3.64 2.61 370 20-24 6.07 5.46 4.40 3.17 390 25+ 6.67 6.23 5.27 4.46 653 Total 4.03 3.77 3.34 2.62 2,679 Standardized 4.03 3.83 3.29 2.60 25 years of married life, the average number of children decreases from 6.7 to 4.5 with increasing expenditures. Even with respect to incomplete fertility, the inverse relationship holds rather systematically. Thus, the - 98 inverse relationship between fertility and per capita household expenditures seems to be quite robust, as is shown not only by cohort fertility, but also by period fertility measures presented in Table 5.20. Table 5.20: Percent Decline in Fertility Rate by Per Capita Household Expenditures, Kerala Rs. 0-59 Rs. 60-120 Rs. 120+ Percent Decline TMFR' 1968-73 16 24 11 1973-78 8 17 12 1968-78 23 38 22 Fertility Rate, 20-29 Years 1968-73 -2 7 2 1973-78 11 18 3 1968-78 9 24 4 Fertility Rate, 30-44 years 1968-73 21 34 33 1973-78 21 36 24 1968-78 38 67 48 Marital fertility rates declined systematically with per capita household expenditures at every age group and in each of the quinquinnial periods (Table 5.21). (The only exception is ages 20-29 years in 1965-70.) Within the same expenditure group, fertility has declined over time. As shown with respect to other characteristics, the overall decline has been largest in the intermediate group, in this case, the Rs. 60-120 group. Fertility decline in this group has formerly been nearly twice that of adjoining groups. The extent of fertility declines had been greater in the - 99 - Table 5.21: Marital Fertility Rate by Age and Monthly Per Capita Household Expenditures, Kerala 1975-80 Age 0-59 60-120 120+ 1975-80 20-24 354 312 306 25-29 307 268 198 30-34 202 121 116 35-39 141 69 61 40-44 73 38 6 TMFR' 5.385 4.040 3.435 1970-75 20-24 345 333 272 25-29 300 294 243 30-34 235 194 172 35-39 181 93 60 40-44 109 ' 57- 38 TMFR' 5.850 4.855 3.925 1965-70 20-24 373 413 244 25-29 354 348 282 30-34 292 238 193 35-39 248 192 99 40-49 (128) (106) ( 63) TMFR' 6.975 6.485 4.405 older 30 age groups (above 30 years) than in the younger age groups (below 30). Among women 30-44 years in the per capita expenditure group Rs. 60-120, fertility declined by 67 percent during the 10-year period 1968-78. Land Ownership: Ownership of land is another economic indicator with which fertility in Kerala shows a fairly consistant relationship. Table 5.22 classifies children ever-born by duration of marriage and ownership of land. - 100 - Table 5.22: Mean Number of Children Ever-Born by Land Owned and Duration of Marriage, Kerala Duration Land Owned (Cents) 1/ Sample of Row Marriage 0-5 6-10 11-4,9 50-99 100+ Totals (years) 0-4 .64 .60 .65 .63 .73 373 5-9 1.8L 2.06 1.85 2.03 1.80 450 10-14. 2.77 3.00 2.81 2.54 2.84 443 15-19 4.00 4.10 3.95 3.67 3.66 370 20-24 5.10 4.90 5.06 4.29 4.78 390 25+ 5.69 5.94 5.55 5.38 5.46 653 Total 3.70 3.50 3.51 3.21 3.44 2,679 Standardized 3.53 3.65 3.50 3.29 3.41 1/ One cent equals 1/100th of an acre. In'general, the less land a household has, the higher is the average number of children per woman in the household; but the differences are not large. The unstandardized average has a range of variation of 0.5 children and the standardized average has a range of only 0.24 children. Completed fertility also differs very little from one land ownership group to another; the extremes differ only by 0.56 children. The pattern is somewhat consistent, however. Households with 6-10 cents of land have the highest fertility and those with 50-99 cents have the lowest. The highest fertility is not observed in the lowest ownership class (0-5 cents), and the lowest fertility is not observed in the highest land ownership class (100+ cents). Differentials and trends by ownership emerge again in Table 5.23. In 1965-70, the fertility rate declined consistently with an increase in land owned oy the household. The TMFR- declined from 6.74 to 6.12. This pattern does not hold in 1975-80, when the highest fertility was observed in-the highest land ownership group and the lowest fertility was observed in an intermediate group (11-49 cents). - 101 - Table 5.23: Marital Fertility Rates by Ownership of Land, Kerala 0-10 11-49 50-99 100+ Age Cents Cents Cents Cents 1975-80 20-24 309 332 375 343 25-29 270 278 304 283 30-34 181 160 138 154 35-39 115 90 62 129 40-44 51 50 38 50 TMFR' 4.630 4.550 4.585 4.795 1970-75 20-24 326 346 278 358 25-29 296 291 259 306 30-34- 217 213 200 213 35-39 133 126 124 122 40-44 80 80 65 67 TMFR- 5.260 5.280 4.630 5.330 1965-70 20-24 365 365 417 374 25-29 348 344 346 339 30-34 260 247 230 269 35-39 234 207 195 157. 40-44 (141) (131) (102) (86) TMFR- 6.740 6.470 6.450 6.125 Percent Decline 1973-78 12 14 1 10 1968-73 22 18 28 13 1968-78 31 30 29 22 Thus, fertility trends have not been uniform among the land ownership groups. On the whole, there is an inverse relationship between land owned and percent decline in fertility. In the lowest group, fertility declined by 31 percent from 1968-78, while in the highest group the decline was only 22 percent. - 102 - Fertility Differences and Trends by Family Planaing Status Conventional Family Planning: Women practicing family planning have much higher fertility than those who do not use a family planning method.. The TMFR' for practicing contraceptors was 5.5 in 1975-80 compared with 4.9 for their non-contracepting counterparts. (This calculation excludes the sterilized women who are not included in either group.) The difference between the groups was 0.275 children per woman in 1965-70, which increased to 0.665 in 1970-75, and then decreased slightly to 0.600 in 1975-80 (Table 5.24). ' High fertility seems to have induced women to use family planning and their adoption led to a reduction in fertility in subsequent years. The narrowing of the differentials in 1975-80 and a slightly higher rate of ferti- lity decline among family planning users may be c'ted as evidence of the impact of family planning on fertility trends. Since the date of adoption of family planning by the women is not known, it is difficult to give a more precise evaluation of the effect of family planning on fertility trends. On the whole, the effect appears to be marginal. Sterilization: As in the case of women who adopted a family planning method, sterilized women have higher fertility than non-sterilized women. In fact, it appears that women accept sterilization because they have too many children. In 1965-70, TMFR' (excluding ages below 20 years and above 45 years) of sterilized women was 7.5 compared to 5.9 for the non-sterilized women (Table 5.25). The difference is 1.6 children per women. In 1975-80, steril- ized women continued to have higher fertility (4.7 compared with 4.2 for the nonsterilized), but the difference between the two groups had. narrowed down to 0.5 children per woman. In 1970-75, the differential was even smaller (0.4). - 103 - Table 5.24: Marital Fertility Rates by Age and Family Planning Status, 1965-70 to 1975-80, Kerala Not Age Group Practicing Practicing 1975-80 20-24 294 418 25-29 270 321 30-34 205 176 35-39 143 120 40-44 60 57 TMFR' 4.860 5.460 1970-75 20-24 280 305 25-29 286 341 30-34 210 237 35-39 130 147 40-44 73 82 TMFR' 4.895 5.560 1965-70 20-24 342 333 25-29 314 330 30-34 222 282 40-44 ( 79) ( 82) TMFR' 5.725 6.000 Thus, it appears that high fertility induced women to accept sterilization and, as a consequence, the fertility differentials between them and the non-sterilized women were reduced considerably. Although the TMFR' was higher among the sterilized, fertility rates at ages above 30 years have not been higher in recent periods. In each of the -104 recent 5-year periods, sterilized women have had higher fertility at ages below 30 years (Table 5.26). In 1965-70, even at older ages (above 30 years), steri- lized women had higher rates than non-sterilized women (see Figure 5.6). After L9TO., the reverse-was true-sterilized women had significantly lower fertility rates, by as much as 0.9 children in 1975-80. Thus, although sterilization has been selective of high fertility women, it has, over time, resulted in a reduction of about one child per woman. From 1965-70 to 1975-80, fertility declines among the sterilized were much higher (38.1 percent) than those among the non-sterilized (29.1 percent) or family planning practitioners (9.1 percent). Table 5.25: Marital Fertility Rates by Sterilization Status, Kerala Age . Not Sterilized Sterilized 1975-80 20-24 160 457 25-29 285 269 30-34 196 116 35-39 136 60 40-44 59 28 TMFR' 4.180 4.650 1970-75 20-24 287 419 25-29 302 281 30-34 219 207 35-39 134 112 40-44 75 71 TMFR' 5.085 5.450 1965-70 20-24 340 416 25-29 319 384 30-34 238 289 35-39 185 248 40-44 103 157 TMFR 5.925 7.470 - 105 - Figure 5.6: Percent Excess Fertility of Sterilized Women Over Non-Sterilized Women, Kerala Percent (Fertility of Sterilized -) 100 (Fertillity of Non-Sterilized +so- _x 1965-70 __30- 1. 2-0 -- *-- - -_ __-1_ _-7 .. .. * . ..: 40 1975-80 7--0 4 IZ......'20 * ~~ L.........-1. 0-24 25-29 30-34 35-39 40-44 Age Source: "The Kerala Fertility Survey, 1980", sponsored by the World Bank, UNFPA, and Bureau of Economics and Statistics, Trivandrum. BES _r____L - 106 - Table 5.26: Marital Fertility Rates by Sterilization Status, Broad Age Groups, Kerala Non- Age Sterilized Sterilized 20-24 years 1975-80 2.225 3.630 1970-75 2.945 3.500 1965-70 3.295 4.000 30-44 years 1975-80 1.955 1.029 1970-75 2.140 1.950 1965-70 2.630 3.470 Factors Related to Fertility: Regression Analyses In the previous section, differentials in fertility levels and trends were estimated based on cumulative as well as period-specific fertility rates. The analyses indicated which socio-economic groups had higher or lower fertility levels, as well as which groups experienced fertility declines and which groups did not. Our analysis yielded considerable insight regarding the factors underlying fertility differentials and trends. In this section, the same question is approached through multivariate analysis. Several measures of fertility are used. We began with cumulative fertility or children ever-born or parity. The results are shown in Statistical Annex Tables I.1 and 2. We then reviewed two period measures--number of births - 107 - per woman during five- and ten-year periods: 1965-75 (Annex Table 1.3), 1970-80 (Annex Table 1.4), 1965-70 (Annex Table 1.5), 1970-75 (Annex Table 1.6), and 1975-80 (Annex Table 1.7). Current fertility was measured by the number of births during the last 12 months before the survey (Annex Table 1.8). The last measure used was the open birth interval, which is inversely related to fertility levels and directly related to the degree of fertility control. Factors related to open birth intervals are given in Annex Table 1.9. Along with each of the nine dependent variables, ten independent variables were used. Two are demographic control variables, age and age at marriage. The third is a family planning variable (1 = sterilized or using a conventional family planning-method, 0 = otherwise). There is one cultural variable, caste (1 = Nair or Syrian Christian, 0 = others) and one socio-economic variable, years of schooling (0,1,2,...) of the woman. The five other variables are indicators of the family's economic conditions. These are: -- per capita household expenditures - land owned by the household -- roofing material (1 = tile, asbestos, tin, concrete; 0 = all other materials) -- source of water supply (1 = pipe; 0 = all others) -- toilet facilities (1 = flush; 2 = ESP stab; 3 = bucket 4 = pit; 5 = compound). In Annex Table I.1 all the ten independent variables had a statistical- ly significant relationship with parity. The total explained variance (R2 ) is 51 percent of which 40 percent was due to age. Of the remaining .11 percent, 8 percent was contributed by age at marriage, 0.8 percent by years of schooling, 0.6 percent by*caste, 0.4 percent by family planning, etc. - 108 - Age at marriage had a negative relationship with parity, the regres- sion coefficient is -0.013. An increase of one year in age at marriage decreases the parity by 0.16. Years of schooling also had a negative coefficient (-0.0753); an increase of five years in schooling reduces the parity by 0.38. Caste is negatively related. Nairs and Syrian Christians had significantly lower parity than all others taken together. The economic variables such as per capita household expenditures, land owned, roofing material, source of water supply, etc., had positive relationships. Other things being equal, those who were "better off" had more children than those who were economically disadvantaged. Family planning status also had a positive relationship. Those who were sterilized or used conventional family planning had more children than others. If we take the number of surviving children as a measure of fertility, the relationship does not change much (A4nex Table 1.2). The R2 decreased to 45 percent, of which 35 percent was due to the age factor. All the factors remained statistically significant and kept the same sign. We next considered the number of children born in a specific period of time. The regressions are given Annex Tables 1.3 to 1.7. A few relationships appeared consistently in all the regressions (see Table 5.27). One was that age at marriage is positively related to fertility. A second was that caste and education are negatively related. In regressions 1.3 and 1.4, age was negative- ly related, age at marriage was positively related, caste and education were negatively related, and toilet facilities were positively related. Thus, between cumulative fertility (regressions I.1 and 1.2) and period fertility (regressions 1.3 to 1.7), there was a reversal of the relationship with respect to age at marriage. 109 - Table 5.27: Summary Results of Regression of Fertility Measures on Demographic and Socio-Economic Variables Independent Regression Number Variables I II III IV V VI VII VIII IX Age + + - - 0, 0 - - + 2 (Age) 0 0 + Age at marriage - + + + + + - Family planning + + + - + 0 - - 0 Caste - - - - - - 0 0 Household expenditures + + 0 0 0 0 0 0 0 Land owned + + + 0 0 0 0 0 0 Years of schooling - - - - - 0 - - + Roof material + + 0 + 0 + 0 0 0 Source of water + + 0 0 0 0 0 0 0 Toilet facility + + + + + + 0 0 - NOTE: + positive significant relationship - negative significant relationship 0 no significant relationship Blank space indicates that the variable was not included in the regression. Dependent Variables in Regression I to IX I: Children ever-born II: Surviving children III: Number of births during 1965-75 IV: Number of births during 1970-80 V: Number of births during 1965-70 VI: Number of births during 1970-75 VII: Number of births during 1975-80 VIII: Number of births in 1979. IX: Open birth interval - 110 - Cumulative fertility was positively related with age as it should be, but period fertility was either negatively related or not significantly related at all. For births in a 10-year period, there was a significant negative relationship; for births in a 5-year period the negative relation- ship was statistically significant only for the period 1975-80 but not for the other 5-year periods. For the 1975-80 period, age was significant at the second degree level in the positive direction. Age at marriage was negatively related to cumulative fertility measures, but positively related (very consistent in all the five regressions) to period fertility. Women who married late had more children born during a fixed 5-year or 10-year period. For example for the 1975-80 period, the regression coefficient is + 0.00173. An increase of ten years in the age at marriage was associated with an.increase of 0.2 children born in that period, a very small difference indeed for a very large difference in age at marriage. Years of schooling were negatively related in all regressions, and in all but one they had a statistically significant association. In general, there was a decline in the value of the regression coefficient with time, and the degree of association between education and fertility appeared to weaken as shown below: Regression B- Period Coefficient Coefficient F-Value 1965-75 -0.04,956 -0.10,218 9.053 1970-80 -0.03,118 -0.07,344 6.847 1965-70 -0.03,339 -0.10,396 8.623 1970-75 -0.00,772 -0.02,684 0.830 1975-80 -0.01,387 -0.05,747 5.203 - 111 - An increase of five years in schooling was associated with a decrease of 0.25 children born during 1965-75, but only 0.16 children in 1970-80. The corresponding changes from 1965-70 to 1975-80 were from 0.167 to 0.069. Thus, education seemed to have lost some of its power in determining a woman's fertility level. Caste had an independent impact on fertility. The regression coefficient was negative and statistically significant in all the regressions. Nairs and Syrian Christians had lower fertility than all the other castes put together. In this case, as in the case of education, there was a gradual decrease in the value of the regression coefficient as shown below: Regression Period Coefficient 1965-75 -0.2250 1970-80 -0.1756 1965-70 -0.1309 1970-75 -0.1215 1975-80 -0.0915 The economic variables had, on the whole, small statistically significant relationships with the dependent variables. Household expenditures were positively related to parity, but their statistical significance disappeared when period specific fertility was considered. Per capita household expendi- tures were not statistically related to the number of births in any period, or even the open interval variable. A somewhat similar picture emerged with respect to land owned by the household. The amount of land owned by a household was positively related to parity, but it had no significant relationship with any of the period specific fertility measures. The situation was no different - 112 - with the roofing material variable or source of water supply. Women living in houses which had a tile, asbestos, tin, or concrete roof did not have significant- ly (statistically) more or less children during a given period of time in recent years than women who live in houses where the roof was made of leaves, grass, or other materials. Similarly women who lived in houses where the water supply was from pipes did not have significantly less or more children than those who lived in houses where the water supply was from a well or other sources. The type of toilet facility showed some significant relationship with fertility. Table 5.27 indicates that the relationship was positive, but the facilities were inversely ranked; flush had rank 1, ESP slab 2, bucket system 3, pit 4, compound, etc. 5. Thus, higher numbers or poorer facilities were associated with higher fertility. Hence, the relationship was indeed negative. The relationship of the family planning variable (1 = sterilized or using a contraceptive, 0 = all others) with fertility was positive in some regressions, but negative in others. We have mentioned that family planning is positively related to cumulative fertility. Those who have large families get sterilized or use contraceptives. The relationship was positive with respect to period fertility for 1965-75 (1.3) and for 1965-70 (1.5). It becomes negative for 1970-80 (1.4) and for 1975-80 (1.7), and non-significant (but plus) for 1970-75 (1.6). The coefficients are shown below: Regression Period Coefficient 1965-75 +0.1976 1970-80 -0.1441 1965-70 +0.2244 1970-75 +0.0290 1975-80 -0.1058 113 - Thus, adopters of family planning were a select group, selective of high fertility. But once they adopted family planning, their fertility for the periods after their adoption was significantly lower than that of non-adopters. Family planning appears to have contributed to a fertility decline in Kerala in recent years. For the births in 1975-80, the ten independent variables explained at most about 29 percent of the total variance. Of this, 26 percent was explained by age. Thus, the variance explained by socio-economic variables was relatively small. The percent distribution of the total variance explained by individual variables is given below: Total 100.0(29.2) Age 90.5- Family Planning 2.2 Toilet Facility 1.9 Age a5 Marriage 1.8 (Age) 1.5 Caste 1.4 Education 0.6 Others 0.1 Two other regressions were included in our analysis. One on births during the 12 months preceeding the survey indicated a significant rela- tionship with age (-), education (-), and family planning (-). None of the socio-economic variables other than education had statistically significant relationships. The last regression was on the open birth interval which was positively related to age and years of schooling and negatively related to age at marriage and toilet facilities. These regressions do not add much to the understanding of the factors underlying fertility, except that they give additional support to the relatively insignificant part which socio-economic - 114 - variables played in explaining fertility differences between women in Kerala. 2 In the case of the open birth interval, the R was as high as 47 percent, but of this 39.5 percent was due to age, and 6.3 percent was due to age at marriage. Education, which was statistically significant, contributed only 0.4 percent to 2 the R. None of the other soc.io-economic. variables were statistically significant. A summary of all the nine regressions (signs and statistical signi- ficance) is given in Table 5.27. The main conclusion which emerges is that, apart from demographic variables such as age and age at marriage, the principal variables which had a statistically significant impact on fertility were family planning, educational attainment, and caste. The years of schooling of the woman had an independent effect on fertility. The relationship of fertility with family planning is understandable, especially because a large proportion of the "family planning" couples are actually sterilized (wife or husband). The relationship with education is in the expected negative direction. The relation- ship remained statistically significant ever after controlling for age at marriage and family planning -- two variables through which education generally operates on fertility. It is possible that the better educated women used family planning more efficiently than less educated women or they (the more educated) had been practicing family planning for longer periods of time. An alternative explanation may be differentials (by education) in the length of time couples live together. Educated women usually have educated husbands who work away from home for longer periods than less educated husbands. - 115 - The factors underlying the negative relationship between fertility and caste are equally obscure. We have seen that the total variance explained by the regression was less than 30 percent. This means that about 70 percent of the variance remains to be explained. Some of the factors involved are certainly cultural and related to marriage, abstinence, childbirth, cohabita- tion after child birth, breastfeeding, etc. Many of the factors are also religion- and caste-related. It is no wonder therefore that fertility varies by caste and religious group. A detailed explanation of the differentials will require an analysis of the cultural practices associated with marriage and child rearing. Some effort in this direction will be made in the last chapter on determinants of fertility. - 116 - Part II: FERTILITY SURVEY AND RESULTS CHAPTER VI IDEAL FAMILY SIZE Introduction The previous chapter was concerned with fertility: the actual number of children ever-born. This chapter is concerned with the desired number of children. The analysis of the desired number of children, especi- ally in conjunction with the actual number of children, provides a measure of the prevalence of "undesired" children (or excess fertility) in the society, a measure of the prevalence of women with excess fertility (more children than desired), and a measure of the utilizai:ion of birth control methods by women who have completed their desired family size. This information should yield better insights regarding the determinants of fertility and family planning usage in Kerala. Desired Number of Children: Distribution All ever-married women in the sample were asked: "If you could choose exactly the number of children to have in your whole life, how many children would that be?" The distribution of women by their desired number of children (D) is given in Table 6.1. The desired number clusters around two and three with about two-thirds of the women preferring either of these numbers. The mode is three, which is the desired number of 37 percent of the women; the mean is 3.25 and the median 2.47. Inter-district variation in the distribution of the desired number of children is shown in Figure 6.1. For both Alleppey and Palghat the mode i, three, but for Ernakulam three is slightly less preferred than two (36.3 percent for three and 36.5 for two). The proportion of women who prefer three children is practically the same in all districts; the proportion who prefer two varies considerably (36.5 percent in Ernakulam, but only 21.6 percent in Palghat). - 117 - Table 6.1: Distribution of Women by Desired Number of Children, Kerala Desired numer Number of of children Women Percent 0 4 0.1 1 56 2.1 2 826 30.7 3 984 36.5 384 14.3 5 209 7.8 6 116 4.3 7 46 1.7 8+ 51 1.9 Unknown 18 0.6 Total 2,694 100.0 - 118 - Figure 6.1: Percent Distribution of Women by Desired Number of Children by District, Kerala Percen Alleppey 40- Ern ulum - . ~ AIl~ëp& 30 Palghat 20 10 0 1 2 3 4 5 6 8 Desired number of children Source: "The Kerala Fertility Survey, 1980", sponsored by the the World Bank, UNFPA, and Bureau of Economics and Statistics, Trivandrum. - 119 - Differentials in Desired Number of Children The desired number of children varies considerably by the demo- graphic characteristics of the women. It also varies with some of the socio- economic. variables.. A few of these differentials are discussed below. Parity: Parity has the highest correlation with the desired number of children (also referred to as desired family size). Table 6.2 gives the average desired family size by parity for the three districts separately and for all districts combined. In all the districts, the desired family size increases with parity. The difference between averages among the districts is not large (3.54 in Palghat and 3.07 in Ernakulam), but fairly consistent at all parities. Between parities 1 and 7, the increase in desired family size is 2.62 and the increase in parity is 6.0. .On average, an increase of one parity results in an increase of 0.44 in the desired number of children. Thus, of every 100 "undesired" children born, 44 become "desired" over the course of time. Age: As age and parity are highly correlated, the relationship between desired family size and age is similar to that wiLh parity. The average desired family size increases from 2.72 at ages 15-19 years to 3.97 at ages 45-49 years (Table 6.3). The increase is consistent from one age group to another and the inter-district differentials are also consistent at all ages. One peculiar feature of these data is that the desired family size of the 20-24 year old women is lower than that of their counterparts aged 15-19. - 120 - Table 6.2: Desired Number of Children by Actual Number of Children Ever-Born (Parity), by District, Kerala Parity Palghat Ernakulam Alleppey Combined Sri Lanka 0 2.82 2.03 2.35 2.44 2.50 1 2.63 2.17 2.03 2.29 2.30 2 2.58 2.34 2.28 3.38 2.65 3 3.06 2.91 2.96 2.97 3.30 4 3.59 3.35 3.32 3.42 3.94 5 4.01 3.94 3.82 3.92 4.65 6 4.34 4.36 4.45 4.38 5.22 7 5.09 4.80 4.68 4.89 5.64 8+ 5.63 4.86 5.44 5.35 6.63 All Parities 3.54 3.07 3.15 3.25 3.75 NOTE: The average desired number of children in a few other countries is shown below (WFS data): Korea 3.2 Pakistan 4.2 Thailand 3.7 Fiji /A.2 Nepal 4.0 Malaysia 4.4 Colombia 4.1 Dom. Republic 4.8 - 121 - Table 6.3: Average Desired Number of Children by Age, by District, Kerala Age Palghat Ernakulam Alleppey Combined 15-19 2.96 2.32 2.38 2.72 20-24 2.94 2.47 2.38 2.60 25-29 3.05 2.49 2.66 2.73 30-34 3.29 2.87 2.83 3.00 35-39 3.91 3.32 3.46 3.55 40-44 4.21 3.75 3.62 3.84 45-49 4.18 4.03 3.76 3.97 All ages 3.54 3.07 3.15 3.25 - 122 - Similar "reverse" patterns are observed in the parity data where the average for zero parity women is slightly higher than that of one-parity women. Thus, women who were yet to bear their first child desired more children than those who already had their first child. The childbearing experience of the first birth seems to depress the desire for additional children. Other differentials: Table 6.4 summarizes the variation in the average desired number of children by age at marriage, family planning practice, education, and owner- ship of land. Those who marry later have a lower desired family size than those who marry early; those who use conventional family planning methods have a lower desired number of children than those who are sterilized (hus- band or wife), or those who do not use any family planning methods. Edu- cated women desire fewer children than the less educated or illiterate. While the ownership of increasing amounts of land does not contribute a marked difference in the desired family size, there does seew a general tendency for women with less land to desire more children than women with more land. Factors Related to Desired Family Size Comparison of the average desired family size in various socio- economic groups is affected by differences in the age-parity composition of the groups compared. Valid comparisons can be made only after standardizing the averages with respect to parity, age, or both. An alternate approach is the multiple regression method which makes it possible to measure the influ- ence of one variable while keeping the other variables constant. The regression co-efficients of selected variables are given in Annex Tables I.10 and I.10A. - 123 - Among the variables included, seven had statistically significant (95 percent level) relationships with the desired number of children: parity, age, age at marriage, family planning status, education, ownership of land, and caste. The only variable which did not have a significant relationship is per capita household expenditures. Table 6.4: Desired Number of Children by Selected Socio-Economic Characteristics, Kerala (a) Age at marriage (years) (b) Current family planning practice Less than 15 4.07 Not practicing 3.35 15-16 3.59 Using conventional 17-18 3.28 family planning methods 2.93 19-20 3.03 Sterilized 3.28 23+ . 2.48 All women 3.25 All women 3.25 (c) Education (d) Land ownership 1/ Illiterate 3.80 Less than 5 Cents 3.32 1-4 years of school 3.39 Less than 10 Cents 3.27 5-9 years of school 3.02 10-49 Cents 3.25 10+ years of school 2.47 50-99 Cents 3.15 All Women 3.25 100+ Cents 3.27 All women 3.25 1/ One cent equals 1/100th of an acre. - 124 - Among those variables which had significant relationships, parity had the highest explanatory power. The R2 was 48.2 percent of which 42.9 percent (or 89 percent of the total) was contributed by parity. With every increase in parity the desired number of children increased by 0.446. Con- versely, desired family size decreased by 0.446 when parity was reduced by one. This relationship can be interpreted in at least two ways. First, desired family size is undergoing a downward trend. If this is true, since the average family size was higher in the past, average parity and average desired family are positively related. Second, there is no trend in the average family size. The increase in D with an increase in P is a change which takes place after the birth of an "undesired" child. When an additional birth takes place, whether it was desired or not before its conception, it becomes desired in 44 out of 100 cases. In actual practice, both these processes may be operating: the desired family size has been declining, and the birth of an undesired child in nearly half the cases changes the parents' expressed attitude about desired family size. The policy implications of these two possibilities (one in which desired family size is determined more by socio-economic variables, and the other in which it is determined by parity), may be very much different. The relation between age of woman and her desired family size is similar to that between parity and desired family size, but the regression co-efficient is much smaller. Age at marriage is negatively related to desired family size (it has a higher regression co-efficient than age). Women who marry later have a lower desired family size irrespective of their age, parity, education, etc. - 125 - There is a strongly negative relationship between desired family size and the number of children dead. The regression co-efficient is -0.3103. For each child's death, the mother's desired family size is lowered by 0.3. Education has a negative relationship with desired family size and land ownership has a positive relationship. The regression co-efficients are relatively small, however. Desired and Actual Number of Children: Excess Fertility In comparison to the distribution of the desired number of chil- dren, the distribution of actual number of children is much more dispersive. (See Figure 6.2.) The mode of the distribution of actual number (parity) is also two or three (according to district), but the percent of women with two or three children is only 35 compared with 67 for the desired. The averages are, however, closer. The arithmatic average of desired number of children is 3.25 and of parity, 3.47. Thus, an ever-married woman in Kerala has, on average, 0.27 children more than she desires. This average difference is, however, not a good measure of the excess fertility in the state. First, the sample of women includes those ranging in age from 15 to 50 years, in all marital durations. The average is a balance between the "excess fer- tility" of older women, and the "deficit fertility" of younger women and, hence, is not a good measure of excess fertility. Second, the average parity reflects not only the present fertility rates, but also past ferti- lity rates. On the other hand, the average desired number of children is a reflection of current conditions. A better comparison would be between the average desired number of children and the cumulative fertility rate based on current fertility rates. Third, when asked about desired number - 126 - Figure 6.2: Percent Distribution of Women by Desired Number of Children and Children Ever-Born, for the Three Districts, Kerala . . Desired no.- of children Percent - Children ever- 40_- born DESIRED 30 . . 20 ACTUAL-- 10 - 01 1 . 4 5 6 7 8 . uimber of Children Source: "Kerala Fertility Survey,1980", sponsored by the World Bank, UNFPA, and the Bureau of Economics and Statistics, Trivandrum. - 127 - of children, a woman's answer is likely to reflect the desired number of surviving children (net fertility and not gross fertility). Therefore, to estimate "excess fertility," the correct comparison should be between de- sired family size and net fertility, that is, the number of surviving children. Excess, Par, and Deficit Fertility The actual number of children ever born (P) to a woman is either more, less, or equal to the desired number of children (D). On this basis women can be classified into three groups: 1. Excess fertility women, that is, for whom P>D 2. Par fertility women, that is, for whom P=D 3. Deficit fertility women, that is, for whom P<D The distribution of the sample of women by these categories is shown in Table 6.5. Thus, 866 women, or nearly one-third of the total, have more children born than they desired; they have excess fertility and can be called excess fertility women. The incidence of excess fertility is fairly uniform in the three districts, but it varies considerably by age, parity, and other socio-economic characteristics of the women. A few of these differentials are discussed below: Excess Fertility by Parity: We have seen that the number of children a woman has is a major determinant of the number of children she desires. It is also a major determinant of her fertility status (excess, par, or deficit). Figure 6.3 shows the distribution of women by fertility status at each parity for the three districts combined and Figure 6.4 gives the proportion of excess ferti- lity women by parity in the three districts separately. - 128 - Table 6.5: Distribution of Women by Fertility Status,by District, Kerala Fertility Status Palghat Ernakulam Alleppey Total Excess fertility 305 261 300 866 Par fertility 261 -347 379 987 Deficit fertility 341 257 243 841 Percent Excess fertility 34 30 33 32 Par fertility 29 40 41 37 Deficit fertility 37 30 26 31 Total 100 100 100 100 - 129 - Figure 6.3: Distribution of Women by Difference Between Parity and Desired Number of Children by Parity, Kerala 100 100. 90 90 80 Deficit fertility 80 70 70 Excess Fertility 60 - -- U 50 50. - 40Fertility at par with 40 desired number of 30 30children 20 20 10 10 1 2 3 4 5 :6 7 8+ Parity Source: "Kerala Fertility Survey, 1980", sponsored by The World Bank, UNFPA, and Bureau of Economics and Statistics, Trivandrum. - 130 - Figure 6.4: Percent of Women with Excess Fertility by Parity, Three Districts, Kerala Palghat Ernakulam Alleppey 90 44 Ernakulam 80 70 Palghat ) .-60 50 40 Alleppey . 30- -20 S . 3 4 5 6 7 ...Parity Source: "Kerala Fertility Survey, 1980, sponsored by the World Bank, UNFPA, and Bureau of Economics and Statistics, Trivandrum. - 131.- There are practically no excess fertility women among those who have two children or less. The proportion of excess fertility women increases very sharply at parity three. Between parties three and four, the increase jumps to 34 percent. The increase decelerates sharply at higher parities as shown below: Increase in % of excess Parity fertility women Between 2 and 3 16.9 " 3 and 4 33.6 " 4 and 5 12.5 " 5 and 6 7.9 6 and 7 5.3 These figures indicate that the three parity women are the best target for family planning education in the sense that they are likely to yield the maximum results by way of new acceptance of family planning programs. Excess Fertility by Age: The relation between excess fertility and age is quite similar to that between excess fertility aud parity. The gradient of the curve is smaller for age than for parity, and the highest proportion of excess ferti- lity women in any age group does not exceed 60 percent. At ages 25-29 years, there are more deficit fertility women than there are either par or excess fertility women. However, in the next higher age group (Fig. 6.5), the major- ity of women have the number of children they desire, and only a very small proportion want more children. In the next age group (35-39 years), there is a reversal of ranks with the majority falling in the excess fertility group. At higher ages (40+), there is neither any change of rank nor any - 132 - Figure 6.5: Distribution of Women by Difference Between Desired and Actual Family Size in Each Age Group, Kerala 100 90 Desired parity in excess of actual parity 80 70 -60 o50 40 30 Desired and actual Actual parity in excess parity at par of desired parity 15 20 25 30 35 40 45 50 - -Age Source: "Kerala Fertility Survey, 1980", sponsored by the World Bank, UNFPA, and the Bureau of Economics and Statistics, Trivandrum. - 133 - substantial change in the level of the various categories. Among women who have completed their reproductive years, about 55 percent have excess ferti- lity, about 37 percent have the number of children they desire, and about 8 percent have fewer children than they desired. Excess Fertility by Ag.e at First Marriage: The proportion of excess fertility women decreases and the propor- tion of deficit fertility women increases with increasing age at marriage (Figure 6.6). These opposing trends cancel one another, leaving unaffected the proportion of women whose desired number of children equals their actual number. Age at marriage has no bearing on about 35 percent of the women whose desired and actual number of children are the same. Among women who marry at ages below 15 years, about 45 percent are excess fertility women, but among those who marry at ages above 23 years only 15 percent are in this category. These proportions are, however, not standardized for differences in age or parity distribution and hence cannot be considered the real effect of age at marriage (see below). Excess Fertility by Socio-Economic Variables: There is a strong negative relationship between education and incidence of excess fertility. The proportion of excess fertility women decreases from 46 percent among women with no education to 15 percent among women with ten years or more of schooling. The relationship between ownership of land and excess fertility is also negative, but the strength of the relationship is rather weak. Among women with little land, about 35 percent are excess fertility women. The proportion decreases only slightly (to 29 percent) when the ownership of land increases to one acre or more (Table 6.6). - 134 - Figure 6.6: Percent Distribution of Women by Age at Marriage and Fertility Status, Three Districts, Kerala Excess Fertility with desire Deficit fertility -100 -90 80 70 ---6-0--- - - *- Excess Deficit 40 -30 Par 20 10 (15 1e-16 17-18 19-20 21-22 23+ .Age at marriage Source: "Kerala Fertility Survey, 1980", sponsored by the World Bank and UNFPA, and the Bureau of Economics and STatistics, Trivandrum. 135 - Table 6.6: Proportion of Excess Fertility Women by Socio-Economic Groups, Kerala Percent of Excess Fertility Women Education No Schooling 46 1-4 years of schooling 37 5-9 years of schooling 24 10+ years of schooling 15 All Groups 32 1/ Land Ownership Less than 5 cents 35 Less than 10 cents 34 11-49 cents .34 50-99 cents 30 100+ cents 29 1/ One cent equals 1/100th of an acre. - 136 - Undesired Children The 866 excess fertility women have between them 1,929 undesired children. The distribution of these women by the number of excess children is shown in Table 6.7; and in Table 6.8, a distribution of the excess children by the parity of their mother. There are 81 women who each had five or more undesired children. The average was about 2.23 children. The 1,929 undesired children represent a fifth (20.8 percent) of the total number of children ever born. As seen in Table 6.8 the proportion of undesired children increases very sharply with parity, from about 6 percent at parity three to about 42 percent by parity 8+. Surviving Children and Desired Number of Children: Excess Net Fertility As mentioned earlier, if a woman is asked about the number of children she would like to have in her lifetime, her answer would most likely reflect the number of surviving children and not children ever-born. Therefore, a more correct comparison regarding excess children is that be- tween desired family size (D) and surviving number of children (Ps). The difference (Ps-D) is called excess net fertility. When Ps < P, the number of women with excess net fertility is less than or equal to the number of women with excess fertility, and the number of undesired children alive is less than or equal to the number of undesired children ever-born. These comparisons are given in Tables 6.9 and 6.10. The number of women with excess net fertility is reduced by about 400, nearly half of the excess fertility women. The number of undesired children is reduced by 49 percent. However, the average number of undesired children per woman remains fairly constant. Those women who have excess fertility or excess net fertility each have about 2.2 undesired children. - 137 - Table 6.7: Distribution of Women by their Number of Undesired Children, Kerala Number of No. of Undesired Number of Women Children 1 364 (42.0) 364 2. 233 (26.9) 466 3 120 (13.9) 360 4 68 ( 7.8) 272 5+ 81 ( 9.4) 467 Total 866 (100.0) 1,929 - 138 - Table 6.8: Distribution of Undesired Children (Gross) by the Parity of their Mothers, Kerala Parity No. of Total No. Percent of Undesired Undesired of Children Children Children Per Children Undesired Woman (gross) (gross) (gross) (1) (2) (3) (4) (5) 1 0 377 0.0 - 2 6 946 0.6 0.01 3 6 1,371 6.2 0.19 4 223 1,368 16.3 0.65 5 313 1,400 22.4 1.12 6 336 1,230 27.3 1.64 7 272 903 30.1 2.11 8+ 694 1,666 41.7 3.45 Total 1,929 9,261 20.8 0.78 a/ Excluding women of zero parity but including all other women (excess, par, and deficit fertility women). Note: Gross here means that the figure is not net of children to mothers who desire more children than they have. - 139 - Table 6.9: Ever-Married Women by Parity or Surviving Parity and Excess Fertility Status, Kerala Fertility Status P - D Ps - D Difference Excess fertility 866 459' -407 Par fertility 987 1,224 +237 Deficit fertility 841 1,011 +170 Total 2,694 2,694 - 140 - Table 6.10: Ever-Married Women by Undesired Children, Kerala Children Surviving Number of Ever-Born Parity Undesired Children Women Children Women Children 1 364 364 153 153 2 233 466 140 280 3 120 360 74 222 4 68 272 42 168 5+ 81 467 29 159 All 866 1,929 438 a/ 982 Average 2.22 2.24 a/ For 21 women the number of undesired children is not known. - 141 - Factors Related to Excess Fertility and Excess Net Fertility Two sets of regressions have been worked out, one using parity and other variables, and the other using age in place of parity (those with an A following- the table number). Similarly, two sets of samples were used, one including excess and par fertility women and the other excluding par fertil- ity women (Annex Tables I.11, I.11A, 1.12, I.12A, 1.13, I.13A). The percent of variance explained is quite high on the whole, as much as 62 percent in the case of excess net fertility (Annex Table 1.13) and 54 percent in the case of excess fertility (Annex Table 1.12). In both cases, the principal factor explaining the variance is parity, which accounts for 71 percent of the explained variance in the case of excess net fertility and 94 percent in the case of excess gross fertility. The second variable in terms of importance is the number of children dead. The relationship in both these cases is positive. Women with more children are likely to have excess fertility (0.44 for each additional parity, net or gross), and women with one or more children dead are more likely to have excess fertility (0.31 gross excess fertility, 0.51 net excess fertility for each child dead). Education has a positive coefficient (statistically significant) which implies that excess fertility increases with educational attainment (+0.028). This result is somewhat surprising in view of the trend in Table 6.6 where the proportion of women with excess fertility decreases sharply with education. But, the two results need not be contradictory. Educated women report more excess fertility than uneducated women. It may well be that educated women report their actual desired family size while the illiterates rationalize the number already born and thus minimize the degree of undesired children. Ownership of land and per capita household expenditures are nega- tively related. As the area of land owned by the household increases and - 142 - as the per capita household expenditures increase, the incidence of excess fertility decreases. Excess Fertility and Family Planning Use One of the major interests in the study of desired number of children is the relationship with the actual number of children and family planning utilization. Do excess fertility women practice family planning? How does their practice compare with that of other women? The relevant statistics are shown in Table 6.11. Excess fertility women practice family planning or are sterilized more often than deficit fertility women (54 percent compared with 22 per- cent), but less often than the par fertility women (54 percent compared with 60 percent). The difference is mainly with respect to sterilization, how- ever, and not conventional family planning. The percent of women who use conventional family planning methods does not vary significantly by the fertility status of women (17, 20, and 18 respectively for the excess, par, and deficit fertility groups). It is instructive to note that' even among the deficit fertility group, 18 percent use contraceptives, which must be for spacing. About 27 percent of the excess fertility women and 40 percent of.the par fertility women were sterilized, however, while the corresponding ,proportion among the deficit fertility women was only 4 percent. The pattern of family planning use does not change substantially if excess fertility is defined in terms of surviving children (see bottom of Table 6.11). A small but noteworthy difference is the percent of users among excess net fertility women. The proportion of users (family planning or sterilization) is higher among the excess net fertility women than among the par fertility women. This is not the case with excess (gross) fertility women. -143 - Table 6.11: Distribution of Ever-Married Women by Fertility Status and Family Planning Use, Kerala Excess Par Deficit Family planning fertility fertility fertility Total Non-users 394 392 655 1,441 Conventional method 149 19 155 503 Sterilization 323 396 31 750 Total 866 987 841 2,694 Percent Non-users 46 40 88 53 Conventional method 17 20 18 19 Sterilization 37 40 4 28 Total 100 100 100 100 Net Fertility Non-users 162 498 761 1,421 Conventional method 92 239 171 502 Sterilization 184 487 78 749 Total 438 1,224' 1,010 2,672 Percent Non-users 37 41 75 53 Conventional method 21 20 17 29 Sterilization 42 39 8 28 Total 100 100 100 100 - 44 - The three districts show some significant variation in the family planning practice among excess fertility women (Fig. 6.7). In Palghat, 71 percent of the excess fertility women have not taken any steps to prevent additional pregnancies. Their desire for fewer children has not resulted in any follow-up action to prevent pregnancy. Although the proportion of non- users is smaller in the other two districts, it is above 30 percent in both. Family planning programs have not served all the women who might have need of them -- particularly among those who report having more children than they desire. There is room for considerable increase in family planning practice and sterilization in these districts. Differentials in Family Planning Use Age and Parity: Younger women and women of lower parity appear to practice family planning more frequently than older women and those of higher parity. This is true irrespetive of the excess fertility status of the woman. The in- creased utilization of family planning at younger ages is highest among the par fertility women and lowest among the deficit fertility women (Table 6.12). Education: About 37 percent of illiterate excess fertility women were steril- ized or using a conventional family planning method compared with 76 percent of those with ten years of schooling. Even among the deficit fertility women, the practice of birth control is greater among the educated: 30 per- cent compared with 9 percent among the illiterate. However, the increase in the use of birth control methods with increasing education is greater among the excess and par fertility women than among the deficit fertility women Figure 6.7: Excess Fertility Women by Family Planning Practice by District, Kerala 90 N 80 70 60 18 28 50-1 NH 40 P-1 30 6 20 10 IN Non-practitioner Conventional family planning methods 06H 4 .terilized Source: "Kerala Fertility Survey, 1980", sponsored by the World Bank, UNFPA and the Bureau of Statistics and Economics, Trivandrum. - 146 Table 6.12: Percent of Non-Users of Family Planning by Fertility Status, Age and Parity, Kerala Excess Par Deficit Age fertility fertility fertility J 25-30 35 32 75 30-34 38 27 65 35-39 39 34 72 40-44 43 47 88 45-49 58 63 97 All ages 45 40 76 Excess Par Deficit Parity fertility fertility fertility 3 27 34 75 4 36 45 67 5 41 45 88 6 52 55 75 7 59 63 100 8+ 56 59 100 All Parities 45 40 70 - 147 - (Table 6.13). A comparison between par and deficit fertility women highlights the part which education may play in the increased adoption of birth control methods. Table 6.13: Percent of Women Who are Sterilized or Use Conventional Family Planning Methods by Education and Excess Fertility Status, Kerala Fertility Status Education Excess Par Deficit Illiterate 37 30 9 1 - 4 years 60 63 14 5 - 9 years 65 69 28 10+ years 76 72 39 Total 55 60 24 Land: Ownership of land is not a significant factor influencing the use of contraceptives, especially among excess fertility women. The percent of women who practice family limitation is more or less the same at all levels of land ownership (Table 6.14). - 148 - Table 6.14: Percent of Women Who are Sterilized or Use Conventional Family Planning Methods by Owner- ship of Land and Excess Fertility Status, Kerala Land Ownership Excess Par Deficit Less than 10 cents 53 57 23 10 - 49 cents 62 62 16 50 - 99 cents 51 66 24 100+ cents 51 60 27 Total 54 60 22 Note: .One cent equals 1/100th of an acre. Factors Related to Adoption of Family Planning by Excess Fertility Women We have seen that not all women with excess fertility (or unde- sired children) adopt family planning methods. What are the factors which lead to family planning usage by some excess fertility women? To answer this question six regressions have been worked out, three for women with excess or par fertility (Annex Tables 1.14, 1.16, 1.18) and three for those with excess net fertility or whose desired number of children is equal to the surviving children (Annex Tables 1.15, 1.17, and 1.19). Annex Tables 1.14 and 1.15 give regressions of the use of conven- tional family planning methods. Only two factors have statistically signi- ficant relationships whether P or Ps is used to define excess and par fertility. These are education and land ownership. Family planning is positively related to education (linearly) and negatively related to ownership of land. - 149 - Thus, highly educated women from a household with a small land holding have the highest chance of adopting family planning. Conversely, illiterate women coming from a household with large land holdings have the minimum chance of using a family planning method. The variance explained is 14.4 percent of which 13.6 percent is contributed by education and 0.5 percent by land ownership. The use of children ever-born or children surviving does not make any change in the conclusions. Other things being equal, an in- crease of one year of schooling increased by 5 percent the proportion using conventional family planning among women who have equalled or exceeded their desired number of children. The propensity to get sterilized is affected by several more variables in addition to education, but the total variance explained by these variables" is much smaller. The principal variable is not education, but parity as shown below: Variable R2 Percent of Total Parity 0.020 42 Education 0.018 38 Household Expenditures 0.007 15 Others 0.003 5 Total 0.048 100 Parity has a significant relationship at the first degree as well as at the second degree. The turning point is 3.6. At parities four and above, the higher the parity, the less the chance of getting sterilized. The relationship with education is also curvilinear. In this case, the turning point is 6.7 years of schooling. At seven years of schooling or at higher - 150 - levels, the higher the schooling the less is the chance of getting steril- ized. The economic variables are negatively related (statistically signi- ficant). The propensity to get sterilized decreases as the area of land owned by the household increases or the per capita household expenditures increase. Poorer women with excess fertility have-a higher chance of being sterilized than richer women with the same education, number of children, etc. If conventional family planning and sterilization are combined, the overall picture does not change substantially. Education emerges as the principal factor with a positive relationship at lower levels of educa- tion and a negative relationship at very high levels. Parity is second in importance, with a positive relationship at lower parities and a negative relationship at higher parities (above three). The negative relationship with ownership of land persists, but that with per capita household expenditures becomes statistically insignificant. Caste is not a factor in the adoption of birth control methods by excess fertility women. This is an important finding in the sense that cultural background plays a minor part in the adoption of family planning. It is the acquired traits such as education, parity, etc., which influence a woman's. decision to accept birth control methods once she achieves or exceeds her desired family size. - 151 - Part II: FERTILITY SURVEY AND RESULTS CHAPTER VII MARRIAGE AS A FACTOR IN THE FERTILITY TREND Marital Status Distribution According to the survey, 1,646 out of a total of 4,315 women aged 15-49 years were never-married, 2,669 were ever-married, 2,371 currently married, 145 widowed, 47 divorced, and the balance were separated or of unknown marital status. These figures imply that a little more than one-third (38 percent) of all women of reproductive age were never married and, therefore, had no chance for reproduction (Table 7.1). Table 7.1: Distribution of Women 15-49 Years by Marital Status, Survey 1980 and Census 1971, Kerala Survey 1980 Census 1971 Number Marital Status Number % (000s) % Never-married 1,646 38.1 1,468 28.3 Ever-married 2,669 61.9 3,720 71.7 Currently married 2,371 54.9 3,318 63.9 Widowed 145 3.4 255 4.9 Divorced 47 1.1 ) ) 147 2.8 Separated 66 1.5 ) Unknown 40 0.9 - - Total 4,315 100.0 5,188 .100.0 Between 1971 and 1980 there was a drastic increase in the proportion never-married: from 28 percent in 1971 to 38 percent in 1980. By 1980, 10 - 152 - percent more of the women in the reproductive ages were not exposed to pregnancy because of celibacy. Among the ever-married women, some were not currently married. These women also were not exposed to pregnancy. The proportion currently married (among all women) was 55 percent in 1980 compared with 64 percent in 1971. A crude estimate of fertility reduction due to changes in the marital status distribution. during 1971-80 is about 14 percent. Table 7.2: Distribution of Ever-Married Women by Marital Status, 1971 and 1980, Kerala Marital Status 1980 1971 Currently married 88.8 89.2 Widowed 5.4 6.9 Divorced 1.8) ) 4.0 Separated 2.5) Unknown 1.5 - Total 100.0 100.0 If the never-married women are excluded, the proportion of currently married has declined very marginally (Table 7.2), the proportion widowed has declined significantly, and the proportion divorced or separated has increased. The net effect of these changes on fertility would be, on the whole, very small. The pronatal effect of a decrease in widowhood is offset by the anti-natal effect of an increase in divorce and separation. - 153 - Variation by Age The overall change in the distribution of women in the reproduc- tive ages by marital status is closely reflected in most younger ages, but the pattern is different in some of the older ages (Table 7.3). Table 7.3: Percentage of Never-Married and Currently Married Women by Age, 971 and 1980, Kerala Never-Married Currently Married Age 1980 1971 1980 1971 15-19 91.5 81.0 7.6 18.1 20-24 55.4 32.7 40.2 64.1 25-29 18.9 9.3 74.8 85.5 30-34 10.0 5.3 82.8 86.8 35-39 10.6 3.7 81.8 85.2 40-44 3.7 3.5 81.7 78.8 45-49 5.9 3.1 76.8 72.9 All ages 38.1 28.3 54.9 63.9 The proportion of never-married women has increased in all age groups, especially at the 20-24 age group. The increase has not resulted in a decrease in the proportion currently married in-ages above 40 years, however. In all the other ages, the proportion of currently married women has shown a considerable decline. The change is from 64 to 40 percent in the 20-24 age group, making possible a fertility decline of 37 percent in this age group over this period due to the marriage factor alone. - 154 - Table 7.4: Percent Widowed or Divorced by Age, 1971 and 1980, Kerala Divorced Widowed or Separated Age 1980 1971 1980 1971 15-19 0.3 0.1 0.6 0.8 20-24 0.9 0.8 2.6 2.3 25-29 0.9 1.9 3.6 3.3 30-34 2.9 3.9 2.7 4.0 35-39 4.3 7.2 1.6 4.0 40-44 8.5 13.3 5.6 4.3 45-49 12.3 20.3 4.1 3.8 All ages 3.4 4.9 2.6 2.8 The proportion widowed has declined in all ages (Table 7.4). Thus, the effect of increase in the proportion never-married has been partly offset by a decline in proportion widowed. The incidence of widowhood and divorce varies somewhat irregularly by socio-economic group (Table 7.5). About 5.4 percent of ever-married women are widows and 4.3 percent are either divorced or separated. Age at Marriage Trends: The average age at marriage of women in Kerala has been quite high compared with other Indian states. For example, in 1971 the average age at marriage was 17.1 years in India and 20.9 in Kerala. The corresponding figures in 1961 were 15.9 and 19.9 respectively. - 155 - Table 7.5: Percent Widowed or Divorced by Socio-Economic Groups, Kerala Socio-economic % Divorced Group % Widowed or Separated Combined Education: No Schooling 6.3 6.4 12.7 1-4 6.8 4.8 11.6 5-9 4.1 3.3 7.4 10 + yrs. 3.7 0.7 4.4 Land(cents): 0-5 7.0 6.3 13.3 6-10 7.7 6.4 14.1 11-49 3.9 3.9 7.8 50-99 5.6 4.6 10.2 100+ 4.4 1.9 6.3 Household Expenditures: < Rs30 8.9 8.9 19.8 30-50 3.9 5.5 9.4 50-60 3.8 4.4 8.2 60-80 6.5 2.8 9.3 80-120 5.7 2.2 7.9 120+ 6.7 2.9 9.6 Not only has the average age at marriage in Kerala been high, it has been increasing in recent years. The age at which 50 percent of the total number of women get married (median age) calculated on the basis of the proportion single is shown in Table 7.6 for the three survey districts. The - 156 - high averages are confirmed, as well as a steady increase over time. Alleppey has the highest average age at marriage and Palghat the lowest. The different- ials which existed in 1961 remained in 1980. Table 7.6: Median Age at Marriage by District, 1961, 1971 and 1980, Kerala 1961 1971 1980 Palghat 18.6 20.1 21.7 Ernakulam 20.7 21.0 23.5 Alleppey 20.9 21.8 23.9 Kerala - 20.7 - More detailed information on trends in the average age at marriage are given in Table 7.7. All the marriages (2,680) were classified Table 7.7: Mean Age at Marriage by Marriage Date and District, Kerala Marriage Year All Districts Palghat Alleppey Ernakulam 1937-49 15.2 14.4 16.2 15.2 1950-54 17.2 16.5 17.8 17.2 1955-59 18.0 17.2 18.2 18.5 1960-64 18.5 17.8 18.6 19.2 1965-69 19.2 18.0 19.8 19.7 1970-74 19.7 18.6 20.0 20.3 1975-80 20.7 19.5 21.7 20.9 Overall Mean 18.8 17.7 19.2 19.4 No. of Women 2,680 899 920 861 - 157 - according to date of occurrence: 1975-80, 1970-75, etc.; and the average age of the women at the time of marriage was calculated for each quinquennial period. Every recent quinquennium shows a higher age at marriage than the previous one in all three districts. Palghat has the lowest average in each of the quinquenniums and Alleppey has the highest.. The difference between these two districts narrowed between 1935 and 1960 but has widened since then. Differentials: The average age at marriage varies somewhat by the socio-economic status of women. Younger women have, on the whole, a higher age at marriage (Table 7.8). This is partly due to an increasing trend in the age at marriage and partly to differences in the socio-economic composition of younger and older women. Illiterate women have a lower age at marriage (Table 7.9). The difference between them and those with at least ten years of schooling is as much as 4.3 years. Among the various religious and caste groups, Muslims , have the lowest age at marriage and Christians have the highest (Table 7.10). The Scheduled Castes have a slightly higher age at marriage than Muslims. Differentials according to ownership of land and household expenses are not very significant (Tables 7.11 and 7.12). Factors Related to Age at Marriage Several regressions were run to identify the factors related to age at marriage. Most of the independent variables turned out to have statistically significant relationships. A simple classification of marriages as "early" (at ages 20 or earlier) or late (later than 20 years) indicated that late marriage is - 158 - associated with caste, education, and year of marriage. All are positively related (Annex Table 1.20). Nairs and Christians have significantly higher proportions of women married late. Similarly, educated women and women who married more recently have higher proportions in the late marriage group. Table 7.8: Average Age at Marriage of Females by Age, Kerala Average age at. Age marriage (years) 15-19 16.8 20-24 18.7 25-29 19.6 30-34 19.2 35-39 18.7 40-44 18.1 45-49 18.5 All ages 18.8 Table 7.9: Average Age at Marriage of Females by Education, Kerala Average age at Education . marriage (years) No Schooling 17.5 1-4 years 18.3 5-9 years 19.1 10+ years 21.8 All women 18.8 - 159 - Table 7.10: Average Age at Marriage of Females by Caste, Kerala Average age at Caste marriage (years) Scheduled Caste/tribe 18.2 Nairs 19.6 Ezawas 18.8 Syrian Christians 19.8 Latin Christians 20.1 Muslims 17.0 Others 18.6 All women 18.8 Table 7.11: Average Age at Marriage of Females by Per Capita Household Expenditures, Kerala Monthly Household Average age at Expenditures marriage (years) (Rs) < 30 18.2 30-40 17.9 50-59 18.6 60-79 19.0 80-119 19.0 120+ 20.1 All women 18.8 - 160 - Table 7.12: Average Age at Marriage of Females by LanL Owned, Kerala Land owned Average age at (in cents) marriage (years) 0- 5 18.3 6-10 18.5 11-49 18.5 50-99 19.1 100+ 19.1 All women 18.8 - 161 - In Annex Tables 1.21 and 1.22 the actual age at marriage is regressed against five variables. The year of marriage (Ym) is used both at the first and second degree. Of the total variance explained by the regression (25 percent), 69 percent was contributed by 'Ym and an additional 4 percent by - 2 m. Both these contributions are highly statistically significant. As the significance of the quadratic factor indicates, the change in age at marriage was not the same in all years; it was larger in earlier years. Thus, between 1960-61, the average age at marriage was estimated to have increased by 1.94 months; but, between 1980-81, the estimated increase was only 0.45 months. Caste, education, and household expenditures (all positive) are also significantly related to age at marriage, while land ownership is not. The regression coefficient for education is +1.074; that is, for each yearly increase in schooling, the average age at marriage increases by 1.07 months. The increase is much smaller than observed in a one-way comparison between age at marriage by educational attainment (more than 4 months). In Annex Tables 1.23 and 1.24 the analysis is confined to women who have completed age 25 years. There is no difference in either direction of relationship or in the statistical significance of the variables. Caste, year of marriage, education, and household expenditures remain the principal factors affecting age at marriage of women, all positively correlated. The total explained variance increased to 35 percent (from 25 percent for all the sample), but year of marriage remained the major factor, followed by caste, household expenditures, and education, in that order. Ownership of land is irrelevant as a factor in age at marriage once the other factors are taken into consideration. - 162 - In Annex Table 1.25, the analysis is confined to women who married before age 25 years. All the variables except land are significantly corre- lated to age at marriage. The effect of calendar year is positive until 1970 and negative afterwards-. Education has a coefficient of 0.85 (an 8.5 month iacrease for a 10-year increase in educational attainment), which is lower than the coefficient when all marriages were considered. In Annex Table 1.26, the analysis is confined to marriages which took place in the 1970's. The coefficient of education is higher (a 13.7 month increase for a 10-year difference in education), but the coefficient for per capita household expenditures became insignificant. In terms of the variance explained by each variable, caste is the principal variable followed by per capita household expenditures. A summary of the results of the seven regressions is given in Table 7.13. Caste, year of marriage, years of schooling appear as statistically Table 7.13: Summary of Regressions of Age at Marriage on Selected Socio-Economic Variables, Kerala Dependent Annex I Table Number Variable 20 21 22 23 24 25 26 Caste + + + + + + + Year of Marriage + + + + + + + (Year of Marriage) 2/ Years of Schooling + + + + + + + Land owned 2/ 0 0 0 0 0 0 0 Per Capita Expenditures 0 + + + + + 0 Notes: + or -: Statistically significant at 5% level. 0: Insignificant at 5% level. Blank space indicates that the variable was not included in the regression. 20, 21, 22: All women. 23 and 24: Women aged 25+ years. 25 and 26: Age at marriage below 25 years. "2/" equals "squared". - 163 - significant variables in all the regressions, while ownership of land is not significant in any of the regressions. Per capita household expenditures are significant in all but one (26), in which only recent marriages are included. The inclusion of all these variables explains a maximum of 38 percent (R2 ) of the total variance in age at marriage. This is a fairly good specification, except that 20 percent of the 38 percent is explained by the time trend and not by the socio-economic variables directly. Marriage Factor in Fertility Decline The date of marriage of each married woman was obtained in the survey along with her date of birth. From these data, it is possible to reconstruct the proportion married at any point of time. This information is shown in Table 7.14. Table 7.14: Percent Married by Age, 1965-80, Kerala Percent Age 1980 1975 1970 1965 15-19 9.5 13.0 22.6 28.7 20-24 46.1 59.9 71.0 72.6 25-29 80.8 87.1 86.1 95.8 30-34 91.2 89.4 98.8 90.8 35-39 89.8 99.0 92.2 (96.6) 40-44 99.3 92.2 (94.4) (94.4) 45-49 92.2 (96.6) (96.6) (96.6) Note: The figures in parentheses cannot be estimated from the survey data. They are assumed. - 164 - Estimated marital fertility rates for 1965-70 to 1975-80 are shown in Table 7.15. The marital fertility rate for 1965-70 is 7.48 and the TFR is 5.91. If the proportion married during 1970-75 had been applied to the marital fertility rate of 1965-70, the TFR would have been 5.63. (See Table 7.16). The difference (5.91-5.63 = 0.28) is due to a decline in the proportion married. The actual TFR is 4.59. We may conclude the following: TFR 1965-70 (actual) 5.91 TFR 1970-75 (actual) 4.59 difference = -1.32 due to marriage factors -0.28 (21%) (5.91-5.63) due to fertility decline = -1.0h (79%) (5.63-4.59) Table 7.15: Marital Fertility Rates by Age 1965-70 to 1975-80, Kerala Age 1975-80 1970-75 1965-70 15-19 200 206 201 20-24 333 334 375 25-29 280 292 344 30-34 162 213 255 35-39 106 127 198 40-44 49 74 (99) 45-49 12 (18) (24) TFR for married women 5.71 6.32 7.48 TFR for all women 3.68 4.59 5.91 - 165 - Similar analyses for 1970-75 and 1975-80 give the following results: TFR (1970-75) 4.59 TFR (1975-80) 3.69 Difference -0.90 Due to marriage factor -0.36 (40%) (4.59-4.23) Due to fertility decline -0.54 (60%) (4.23-3.69) And for the 1965-70 to 1975-80 period, the overall decomposition is as follows: TFR 1965-70 5.91 TFR 1975-80 3.69 Difference -2.22 Due to marriage factor = -0.68 (31%) (5.91-5.23) Due to fertility decline = -1.54 (69%) (5,23-3.69) Thus, the marriage iactor played a significant role in the fertility decline during 1965-80. Roughly 31 percent of the fertility decline during this period was due to declines in the proportion married, and the balance of 69 percent was due to fertility declines within marriage. The effect of the marriage factor was greater in 1970-80 than in 1965-75. A more detailed analysis by age is given in Table 7.16. Figure 7.1 indicates that the effect of marital status on fertility was confined to the younger ages, principally the 20-24 age group (52 percent of the total decline) and 15-19 age group (22 percent of the total decline). There is no age giaup where there was any increase in fertility due to the marriage factor. -166&- Table 7.16: Analysis of Fertility Change Due to the Marriage Factor, 1965-70 to 1975-80, Kerala 1965-70 to 19/5-75 Expected Change Change rates if due to due to Age Actual fertility rates Decline marriage rates marriage fertility 1965-70 1970-75 alone changed factor factor 15-19 52 37 -15 36 -16 + 1 20-24 269 219 -50 246 -23 -27 25-29 313 253 -60 298 -15 -45 30-34 242 200 -42 240 - 2 -40 35-39 189 121 -68 189 + 1 -68 40-44 93 70 -23 93 - -23 45-49 23 17 - 6 23 - - 6 TFR 5.91 4.59 -1.32 5.63 -0.28 -1.04 1970-75 to 1975-80 70-75 75-80 15-19 37 22 -15 23 -14 - 1 20-24 219 176 -43 177 -42 - 1 25-29 253 235 -18 245 - 8 -10 30-34 200 146 -54 192 - 8 -46 35-39 121 100 -21 120 + 1 -22 40-44 70 47 -23 71 - 1 -22 45-49 17 11 - 6 17 - - 6 TFR 4.59 3.69 -0.90 4.23 -0.36 -0.54 1965-70 to 1975-80 65-70 75-80 15-19 52 22 -30 23 -29 - 1 20-24 269 176 -93 199 -70 -23 25-29 313 235 -78 289 -24 -54 30-34 242 146 -96 230 -12 -84 35-39 189 100 -89 187 - 2 -87 40-44 93 47 -46 95 + 2 -48 23 11 -12 23 - -12 TFR 5.91 3.69 -2.22 5.23 -0.675 -1.545 - 167 - Figure 7.1: Specific Fertility Rates, 1965-70 and 1975-80, and Fertility Decline Due to Marriage and Fertility Factors, Kerala Specific Fertility Rare - - -- - -- Per 1,000 Women - - -Percent Relative contribution of - --fkrital -factor to -ell -factors -- - -- - - 1I 300 ..-....Due ..to Marriage.Fac-r 90 -965 -70. 80 -- - - - 70 200 60 1975-80 --- 40 Due to Ter- 100 tility 30 Decline with- in Marriage 20 10 15 20 25 30 35 40 45 Age Source: "Kerala Fertility Survey, 1980", sponsored by the World Bank, UNFPA, and the Bureau of Economics and Statistics, Trivandrum. - 168 - Part II: FERTILITY SURVEY AND RESULTS CHAPTER VIII FAMILY PLANNING AS A FACTOR IN THE FERTILITY TREND Introduction All ever-ma-rried women wer-e: asked the question: "As you know, there are various methods a couple could resort to with a view to delay the next. pregnancy or. to avoid pregnancy. Do you know of, or have you heard of any of these ways or methods?" After recording their answers about the various types of family planning methods they knew, the methods which they did not mention were described to them one by one and they were asked whether they knew or had heard of any of them, and whether they had ever used any of them. Additional information about the first method used (if any), the number of children the woman had when she started to practice family planning, induced abortion, current practice of family planning, sterilization (wife or husband) and desire for additional children, etc., were obtained. The data on these are now being tabulated. The following section is based on limited data regarding ever-use of contraceptives, current use of contracep- tives, and sterilization of husband or wife. Knowledge of Family Planning Methods Married women in Kerala are aware of birth control methods, both conventional family planning methods and sterilization. Of the 2,694 women to whom this question was asked 2,383 answered (spontaneously, witnout probing) that they knew one or other of the birth control methods, while 311 or 12 percent of the total said they did not know of any method. Comparison with a few countries is given in Table 8.1, indicating that knowledge of birth - 169 - Table 8.1: Percent of Women Who Know One or More Birth Control Methods, Selected Countries Country % Kerala 88, Pakistan 75 Indonesia 77 Sri Lanka 90 Malaysia 92 South Korea 97 Source: Kerala Survey; World Fertility Survey and International Statistical Institute, 1978. Malaysian Fertility and F5mily Survey, 1974, Summary of Findings, Kuala Lumpur: Malaysian Department of Statistics; Republic of Korea, World Fertility Survey, and International Statistical Institute, 1978. Korean National Fertility Survey, 1974, Summary of Findings, London (?): World Fertility Survey (?); Population Planning Council of Pakistan, 1976. World Fertility Survey, Pakistan Survey First Report, Lahore: Population Planning Council of Pakistan; Sri Lanka, Department of Census and Statistics, 1978. World Fertility Survey, Sri Lanka, 1975, First Report, Colombo (?): Ministry of Plan Implementation; and World Fertility Survey data regarding Indonesia. - 170 - control methods among Kerala women is higher than in Pakistan or Indonesia, but lower than in some other countries of Southeast Asia, such as Korea and Malaysia. Knowledge of family planning methods is highest in Alleppey District (95 percent) and lowest in Palghat (80 percent), with Ernakulam falling in between. On average, an ever-married Kerala woman knows three birth control methods: 3.5 methods in Ernakulam, 3.2 methods in Alleppey and 2.1 methods in Palghat. Although fewer of the Ernakulam women knew about family planning methods compared to women in Alleppey District, those who knew, knew more methods than their counterparts in Alleppey. About 30 percent of the Ernakulam women knew five or more methods. Table 8.2: Percent Distribution of Women by Number of Family Planning Methods Known by District, Kerala No. of methods known Palghat Ernakulam Alleppey Combined 0 19.5 10.0 5.1 11.5 1 13.4 3.8 0.4 5.9 2 33.0 10.4 27.2 23.8 3 15.9 18.7 25.4 20.0 4 12.1 26.8 26.8 21.9 5+ 6.1 30.2 15.1 16.9 Total 100.0 100.0 100.0 100.0. (907) (865) (922) 2,694 Average 2.1 3.5 3.2 2.9 - 171 - Differentials in Knowledge of Family Planning Methods Knowledge about family planning methods is highest at ages 30-34 years, where all but 8 percent of the ever-married women knew about family planning methods (Table 8.3). It is lowest at ages 15-19 years, where about one-fifth had not heard of any family planning methods. Variation of knowledge about family planning methods by parity is an inverted J-shaped curve, with the highest degree of knowledge at parity three. The relationship with age at marriage is somewhat linear, with knowledge increasing along with age at marriage. Years of schooling have a very significant and positive impact on the knowledge of family planning. While about 25 percent of the women with . no education have never heard of any family planning method, all but 2 percent of the women with ten years or more schooling knew one or other of the birth control methods. At each level of schooling, an increase in the years of schooling results in an increase in the percentage of women who have knowledge about family planning methods. Practice of Conventional Family Planning Methods Of the 2,671 women covered in the survey, 908 have used conventional family planning methods at least once in their lifetime and 502 were using a method at the time of the survey. These figures imply that 406 of those who used a method in the past have given up or discontinued. It does not, however, imply that all these 402 women totally disregarded birth control practice; in fact, 217 of them were subsequently sterilized and did not require any family planning method. The following distribution gives the proportion of women who are exposed to the risk of pregnancy and those among them who use a family planning method. - 172 - Table 8.3: Knowledge of Family Planning Methods by Demographic And Socio-Economic Characteristics of Women, Kerala Age A B Education A B 15-19 81 2.1 No Education 75 1.9 10-24 86 2.9 1-4 years 89 2.7 25.-29 90 3.1 5-9 years 94 3.3 30-34- 92 3.2 10+ years 98 4.2 35-39 90 3.0 40-44 88 2.8 45-49 85 2.5 All ages 89 2.9 All years 89 2.9 Parity A B Land ownership A B 0 76 2.4 Less than 10 cents 87 2.7 1 86 2.9 10-49-cents 87 2.8 2 91 3.2 50-99 cents 91 3.0 3 94 3.3 100+ cents 90 3.2 4 89 2.8 5 91 2.8 6 88 2.7 7 82 2.5 8+ 89 2.5 All 89 2.9 All 89 2.9 A = percent of women who know at least one F.P. method B = average number of methods known 40 - 173 - 1. Total ever-married 2,671 100.0 - 2. Sterilized 746 27.9 - 3. Currently not married 298 11.1 - 4. Pregnant 31 1.2 - 5. Total exposed (5=1-2-3-4) 1,596 59.8 100 6. Ever users of F.P. who ;ere not subsequently sterilized 691 25.9 - 7. Current users of F.P. 502 18.8 31.5 8. Discontinuers of F.P. 189 7.1 9 . Unprotected 1,094 41.0 68.8 The total number of women who were exposed to pregnancy was 1,596 or 60 percent of the total. Among them, 502 use one form of family planning method. The balance of 1,094 were unprotected. They form 41 percent of all women, or 69 percent of the total exposed women. The practice of conventional family planning methods was limited to only about 32 percent of the exposed women. Variation by District The usage rate of conventional family planning methods is very low in Palghat District; only about 7 percent of the ever-married women used some form of contraception (Table 8.4). It is highest in Alleppey District with 26 percent of the ever-married women having used contraception. At younger ages, women in Ernakulam District are more prone to use conventional . family planning methods than women in Alleppey District. At older ages, the reverse situation prevails. This is partly due to the differential adoption rate of sterilization as a birth control method between these two districts. If the usage rate of'conventional family methods is calculated for non-steri- lized women, the two districts are very close: 38 percent in Alleppey and 37 percent in Ernakulam. The differentials by age remain unaffected however. The usage rate is significantly higher in Ernakulam than Alleppey at younger ages, and lower at older ages. - 174 - Table 8.4: Percentage of Ever-Married Women Using Conventional Family Planning Methods by District, Kerala Age Palghat Ernakulam Alleppey Total 15-19 3.6 18.2 7.7 7.8 20-24 5.7 30.9 26.7 21.1 25-29 9.5 28.1 25.5 21.3 30-34 6.6 23.9 28.5 19.5 35-39 8.1 17.8 27.2 18.2 40-44 8.1 26.2 22.5 19.1 45-49 2.1 .14.0 28.0 15.5 All ages 6.5 23.5 26.1 18.7 Family Planning Practice by Demographic and Socio-Economic Variables Age: Family planning practitioners are fairly uniformly distributed by age. The proportion of ever-users and current users remains fairly constant at about 35 percent of all ever-married women for ever-users, and about half that percentage for current users. Some variation is noticed when the sterilized women are removed from the denominator as shown in Table 8.5. Family planning practice increases with age till age 30-34 years and then decreases. This is the result of two conflicting trends. The practice of family planning is greater among the educated; and since the younger are more educated than the older, the percent of family planning accepters should be higher at younger ages. - 175 - Table 8.5: Percentage of Users of Family Planning Methods by Age, Kerala Family planning users as percent of: Age All Ever-Married Women Currently Married Non- Sterilized Women 15-19 7.8 14.4 8.9 16.5 20-24 21.1 32.1 25.7 39.0 25-29 21.3 36.6 32.7 56.1 30-34 19.5 37.6 39.9 77.0 35-39 18.2 35.3 36.0 69.8 40-45 19.1 36.0 37.3 70.1 45-49 15.5 29.6 26.3 50.2 All ages 18.8 34.0 30.9 55.9 Average age 33.9 34.4 - - Parity: The relationship between family planning usage and parity is somewhat similar to that with age. The highest utilization is at parity two, at which level about 45 percent of all ever-married women have ever used a method and 29 percent were currently using. If we exclude the sterilized women from the denominator, the proportions given in Table 8.6 are obtained. - 176.- Table 8.6: Percentage of Users of Family Planning Methods by Parity, Kerala Family planning users Family planning users Parity as percent of ever-married as percent of ever-married women non-sterilized.women Current Use Ever Use Current Use Ever Use 0 4.3 11.3 4.3 11.3 1 24.9 34.2 25.9 35.5 2 29.0 45.5 39.3 61.6 3 17.9 39.2 32.2 70.2 4 16.4 36.3 28.3 62.6 5 16.8 32.1 27.9 53.6 6 14.6 30.0 21.4 44.3 7 9.3 25.6 13.2 36.3 8 14.9 29.7 20.8 41.7 9 22.8 28.1 31.7 39.0 10 16.3 23.3 18.9 27.0 All 18.7 33.9 25.7 47.0 Average parity 3.2 3.3 - - Age at Marriage, Education, Economic, and Caste Differentials: With age at marriage the relationship is fairly linear (Table 8.7). Women who marry late tend to adopt family planning methods more often than women who marry early. Similarly, women who have more schooling also practice family planning in greater numbers than women who are illiterate. The increase in proportion currently practicing with an increase in years of schooling is very sharp: from 6.3 percent for no schooling to 40.3 percent for those with 10 years or more education (Table 8.8). The relationship with per capita household expenditures is also sharply linear; an increase from 7 percent to 33 percent from the lowest expenditure class to the highest (Table 8.9). - 177 - Table 8.7: Distribution of Ever-Married Women by Use of Family Planning Methods and Age at Marriage, Kerala Conventional F.P. Methods. Currently Using or Age at Current- Used in Ever Steri- Steri- Never Marriage ly using the past Used lized lized Used Total 15 36 40 76 75 111 181 332 15-16 82 77 159 174 256 205 538 17-18 117 110 227 220 337 243 690 19-20 105 93 198 147 252 185 530 21-22 73 51 124 68 141 108 300 23+ 90 40 130 66 156 108 304 All 503 411 914 750 1,253 1,030 2,694 15 10.8 12.0 22.9 22.6 33.4 54.5 100 -- 15-16 15.2 14.3 29.6 32.3 47.6 38.1 100 17-18 17.0 15.9 32.9 31.9 48.8 35.2 100 19-20 19.8 17.5 37.4 27.7 47.5 34.9 100 21-22 24.3 17.0 41.3 22.7 47.0 36.0 100 23+ 29.6 13.0 42.8 21.7 51.3 35.5 100 All 18.7 15.3 33.9 27.8 46.5 38.2 100 -178- Table 8.8: Percentage of Users of Family Planning Methods - By Educational Attainment of Women, Kerala Family planning users as percent of: Educational attainment All Ever Married Women Ever-Married Non-Sterilized Women Current Use Ever Use Current Use Ever Use No education 6.3 12.9 8.0 16.5 1-4 years 25.1 30.8 22.9 46.5 5-9 years 23.4 41.8 33.1 59.2 10+ years 40.3 62.5 49.6 76.8 All 18.7 33.9 25.9 47.0 Table 8.9: Percentage of Users of Family Planning Methods By Per Capita Family Expenditures, Kerala Family planning users as percentage of: Per capita household All Ever-Married Women Ever-Married expenditures Non-Sterilized Women (monthy) Current Use Ever Use Current Use Ever Use Below Rs. 30 6.5 13.0 8.2 16.6 30-49 13.7 25.0 18.8 34.2 50-59 16.2 29.7 24.0 44.1 60-79 16.1 34.8 23.6 50.9 80-119 27.2 44.7 38.9 63.9 120+ 33.4 58.3 40.9 71.2 All 18.5 33.8 25.9 47.0 - 179 - Caste is an important variable affecting family planning usage. Scheduled castes, scheduled tribes, Muslims, and Ezawas have relatively lower proportions of users, while Christians have the highest (Table 8.10). Table 8.10: Percentage of Users of Family Planning Methods by Caste, Kerala Family planning users as percentage of: Caste All Ever-Married Women Ever-Married Non-Sterilized Women Current Use Ever Use Current Use Ever Use Scheduled Caste/ Tribe 9.3 20.1 13.0 28.3 Muslims 10.4 21.7 13.3 27.7 Ezawas "14.8 28.7 21.9 42.7 Nairs 24.5 44.7 32.3 58.9 - Syrian Christians 32.1 55.5 48.0 83.1 Latin Christians 32.3 49.6 44.1 67.7 All 18.5 33.8 25.6 46.6 Sterilization In this section we define a woman as sterilized if either she or her husband is sterilized. Detailed information about who is sterilized (husband or wife), when he/she was sterilized, etc., were obtained in the survey. These data will be available at a later stage of processing. The total number of women who were sterilized (or whose husbands were sterilized) was 750 or 28 percent of the total ever-married women, and 31 percent of the currently married women. Thus, nearly a third of the exposed women in Kerala were sterilized (Table 8.11). - 180 - As in the case of conventional family planning methods, the three districts vary considerably with respect to the prevalence of sterilization. The overall pattern is very similar. In Ernakulam District, 37 percent of ever-married women were sterilized; but the corresponding percentage in Palghat District was only 15. In the younger age groups, Alleppey District has a higher proportion of sterilized women, but at ages above 25 years, the proportion in Ernakulam District was significantly higher. The sterilization Table 8.11: Percentage of Ever-Married Women Sterilized By District Age Palghat Ernakulam Alleppey All districts 15-19 - - 7.7 1.1 20-24 4.1 8.1 9.2 7.1 25-2,9 16.1 34.7 26.1 26.1 30-34 25.2 56.0 46.4 42.0 35-39 20.3 56.2 23.4 40.5 40-45 16.1 41.0 40.4 33.0 45-49 9.9 30.7 25.0 21.5 All ages 14.7 37.2 32.0 27.8 - 181 - Table 8.12: Percentage of Exposed Women Sterilized, Selected Countries 1/ Country Percent Bangladesh 1.0 Fiji 20.2 Indonesia 0.4 Jordan 2.8 Korea 6.6 Malaysia 4.9 Nepal 2.1 Pakistan 1.4 Philippines 7.1 Sri Lanka 12.9 Thailand 11.5 India Kerala 31.5 1/ Male and female sterilization for contraceptive purposes. rate in Kerala is substantially higher than in many other countries. A few comparisons are given in Table 8.12. Variation by Age: The majority of sterilized women, 48 percent of the total, were in the 30-39 age group (Table 8.13). The average age of a sterilized woman is 34.4. Thus, age is an important factor related to the propensity to accept sterilization. At ages 15-19 years, very few are sterilized. The proportion of women who are sterilized increases considerably with the wife's age, until ages 30-34 years. In this age group, about 42 percent of all ever-married women were sterilized (56 percent in Ernakulam District). At higher ages, the proportion decreases, partly due to the fact that the proportion of currently married women decreases at higher ages, but also because the propensity to be sterilized among currently married women also decreases. - 182 - By Parity: The relationship between sterilization and parity is also curvili- near with the highest proportion of sterilization among the three parity women. The propensity to be sterilized increases sharply between parities one and two and between two. and threer. The proportions remain somewhat stable at parities three, four, and five, and then.decrease (Table 8.14). The average parity of a sterilized woman is-about 4.1. The mode is three and the median parity is very close to the mode: 3.2. Table 8.13: Percentage of Women Sterilized by Age, Kerala Percentage of: Age Ever-Married Currently Married 15-19 1.1 1.3 20-24 7.1 8.0 25-29 26.1 28.6 30-34 42.0 46.2 35-39 40.5 '44.5 40-44 33.0 39.1 45-49 21.5 26.7 All 27.8 31.5 Table 8.14: Proportion of Women (or Husbands) Sterilized by Parity, Kerala Proportion sterilized Parity Among Ever-Married Among Currently Married 0 1 3.7 4.4 2 26.2 29.5 3 44.2 49.0 4 42.1 47.1 5 40.0 45.2 6 31.7 36.3 7 29.5 33.0 8+ 25.4 28.5 All 27.8 31.5 - 183 - By Education: Unlike the sharp positive relationship between educational attain- ment and conventional family planning practice, the relationship between education and sterilization is curvilinear with the highest proportion steri- lized among those with 1-4 years of schooling(Table 8.15). Among them, about 39 percent of the currently married women were sterilized. Part of the reason why women with more years of schooling were not sterilized was that they were using other fertility control methods (Table 8.16). But this alone is not the complete explanation. A better insight regarding the relationship will be obtained through regression analysis (see below). Table 8.15: Proportion of Women Sterilized by Education, Kerala Education As percentage of: All Ever- A,ll Currently Married Women Married Women No education 21.6 25.2 1-4 years 33.8 39.2 5-9 years 29.3 32.7 10+ years 18.5 20.3 All women 27.8 31.6 - 184- Table 8.16: Distributtion of Ever-Married Women by Use of Family Planning Methods and Educ:ation, Kerala Cur- Conventional F.P. Methods rently Using or Current- Used in Ever Steri- Steri- Never Education ly using the past Used lized lized Used Total Not Edu- 39 41 80 134 173 405 619 cated 1-4 125 129 254 279 404 292 825 5-9 228 180 408 286 514 281 975 10+ 111 61 172 51 162 52 275 All 503 411 914 750 1,253 1,030 2,694 Not Edu- 6.3 6.6 12.9 21.6 27.9 65.4 100 cated 1-4 years 15.1 15.6 30.8- 33.8 49.0 35.3 100 5-9 year 23.4 18.5 41.8 29.3 52.7 28.8 100 10+ years 40.3 22.2 62.5 18.5 58.9 18.9 100 All women 18.7 15.3 33.9 27.8 46.5 38.2 100 - 185 - Other Differentials: Tables 8.17 and 8.18 give the variation of the proportion sterilized by per capita household expenditures and ownership of land. With household expenditures, the relationship is curvilinear. Women in households with a per capita monthly expenditure of Rs 0-59 have the highest proportion sterilized. With land, there is very little variation by size of land holdings. Conventional Family Planning Use and Sterilization Conventional family planning methods and sterilization are alternate means of controlling fertility. If a woman is sterilized there is no need for her to use conventional family planning methods; if she uses conventional family planning methods efficiently, there is no need to resort to steriliza- tion.. There is, therefore, a need to examine both these methods simultaneously. Tables 8.7 and 8.16-8.21 give the distribution of all ever-married women by their family planning practice. About 46.7 percent of all ever- married women were either sterilized or use a conventional family planning method. In general, the relationship between socio-economic factors and family planning practice is quite similar to that with sterilization. For example, the highest practice of fertility limitation is seen at ages 30-34 years (Table 8.19), and awn- women who have three children (Table 8.20). The overall relationship is similar with respect to sterilization and conventional family planning methods. There are differences, however, by education (Table 8.16). With conventional family planning methods, the relationship is positive throughout, but with sterilization the relationship is curvilinear. When these two methods are combined, the relationship remains positive. Nearly 60 - 186~ - Table 8.17: Distribution of Ever-Married Women by Use of Family Planning Methods and Per Capita Household Expenditures, Kerala Cur- Conventional F.P. Methods rently Monthly Using Per capita or Household Current- Used in Ever Steri- Steri- Never Expenditures ly using the past Used lized lized Used Total < Rs. 30 14 14 28 46 60 141 215 30-49 109 89 198 214 323 381 793 50-59 55 46 101 111 166 128 340 60-79 82 95 177 160 242 171 508 80-119 124 80 204 137 261 115 456 120+ 105 78 183 57 162 74 314 6 5 11 12 18 26 43 Unknown All 495 407 902 737 1,232 1,030 2,669 < 30 6.5 6.5 13.0 21.4 27.9 65.6 100 - 30-49 13.7 11.2 25.0 27.0 40.7 48.0 100 50-59 16.2 13.5 29.7 32.6 48.8 37.6 100 60-79 16.1 18.7 34.8 31.5 47.6 33.7 100 80-119 27.2 17.5 44.7 30.0 57.2 25.2 100 120+ 33.4 24.8 58.3 18.1 51.6 23.6 100 All 18.5 15.2 33.8 27.6 46.2 38.8 100 - 187 - Table 8.18: Distribution of Ever-Married Women by Use of Family Planning Methods and Land Ownership, Kerala Conventional F.P. Methods Currently Using or Land Current- Used in Ever Steri- Steri- Never Ownership ly using the past Used lized lized Used Total < 5 cent 62 65 105 109 171 185 399 6-10 78 58 136 144 222 196 476 11-49 132 125 257 208 340 254 719 50-99 58 58 116 105 163 128 349 100+ 173 127 300 184 357 267 751 Total 503 412 915 750 1,253 1,030 2,694 < 5 cent 15.5 16.3 26.3 27.3 42.9 46.4 100 6-10 16.4 12.2 28.6 30.3 46.6 41.2 100 11-49 18.4 17.4 35.7 28.9 47.3 35.3 100 50-99 16.6 16.6 33.2 30.1 46.7 36.7 100 100+ 23.0 16.9 39.9 24.5 47.5 35.6 100 Total 18.7 15.3 34.0 27.8 46.5 38.2 100 --188 - Table 8.19: Distribution of Ever-Married Women by Use of Family Planning Methods and Age, Kerala Cur- Conventional F.P. Methods rently Using or Current- Used in Ever Steri- Steri- Neverl/. Age ly using the past Used lized lized Used 15-19 7 6 13 1 8 76 90 20-24 77 40 117 26 103 222 365 25-29 112 80 192 137 249 196 525 30-34 85 79 164 183 268 89 436 35-39 80 75 155 178 258 106 439 40-44 76 67 143 131 207 123 397 45-49 65 59 124 90 155 205 419 All Ages 502 406 908 746 1,248 1,017 2,671 15-19 7.8 6.7 14.4 1.1 8.9 84.4 100 20-24 21.1 11.0 32.1 7.1 28.2 60.8 100 25-29 21.3 15.2 36.6 26.1 47.4 37.3 100 30-34 19.5 18.1 37.6 42.0 61.5 20.4 100 35-39 18.2 17.1 35.3 40.5 58.7 24.1 100 40-44 19.1 16.9 36.0 33.0 52.1 31.0 100 45-49 15.5 14.1 29.6 21.5 37.0 48.9 100 All Ages 18.8 15.2 34.0 27.9 46.7 38.1 100 1/ Minimum estimate; some women who had ever used conventional family planning later were sterilized. - 189 - Table 8.20: Distribution of Ever-Married Women By Use of Family Planning Methods and Parity, Kerala Cur- Conventional F.P. Methods rently Using or Current- Used in Ever Steri- Steri- Never Parity ly using the past Used lized lized Used Total 0 10 16 26 -- 10 204 230 1 94 35 129 14 108 234 377 2 137 78 215 124 261 134 473 3 82 97 179 202 284 76 457 4 56 68 124 144 200 74 342 5 47 43 90 112 159 78 280 6 30 32 62 65 157 78 205 7 12 21 33 38 83 58 129 8 15 15 30 29 74 42 101 9 13 3 16 16 29 25 57 10+ ' 7 3 10 6 13 27 43 Total 503 411 914 750 1,253 1,030 2,694 0 4.3 7.0 11.3 --- 4.3 88.7 100 1 24.9 9.3 34.2 3.7 28.6 62.1 100 2 29.0 16.5 45.5 26.2 55.2 28.3 100 3 17.9 21.2 39.2 44.2 62.1 16.6 100 4 16.4 19.9 36.3 42.1 58.5 21.6 100 5 16.8 15.4 32.1 40.0 56.8 27.9 100 6 14.6 15.6 30.2 31.7 46.3 38.0 100 7 9.3 16.3 25.6 29.5 38.8 45.0 100 8 14.9 14.9 29.7 28.7 43.6 41.6 100 9 22.8 5.3 28.1 28.1 50.9 43.9 100 10+ 16.3 7.0 23.3 13.9 30.2 62.8 100 All 18.7 15.3 33.9 27.8 46.5 38.2 100 - 190 - Table 8.21: Distribution of Ever-Married Women by Use of Family Planning Methods and Caste, Kerala Cur- Conventional F.P. Methods rently Using ,or Current- Used in Ever Steri- Steri- Never Caste ly using the past Used lized lized Used Total Scheduled 30 35 65 94 124 165 324 Nair 92 76 168 91 183 117 376 Ezawa 86 82 168 188 274 225 581 Syrian xians 134 98 232 139 273 47 418 Latin xians 41 22 63 34 75 30 127 Muslims 37 40 77 71 108 207 355 Others 75 54 129 120 195 239 488 All 495 407 902 737 1,232 1,030 2,669 Scheduled 9.3 10.8 20.1 29.0 38.3 50.9 100 Nair 24.5 20.2 44.7 24.2 48.7 31.1 100 Ezawa 14.8 14.1 28.9 32.4 47.2 38.7 100 Syrian 32.1 23.4 55.5 33.2 65.3 11.2 100 Latin 32.3 17.3 49.6 26.8 59.1 23.6 100 Muslims 10.4 11.3 21.7 20.0 30.4 58.3 100 Other 15.3 11.0 26.4 24.6 40.0 49.0 100 All 18.5 15.2 33.8 27.6 46.2 38.6 100 - 191 - percent of women who have ten years of schooling have adopted some form of birth control compared with 53 percent among those with 5-9 years of schooling, 49 percent among those with 1-4 years of schooling, and 28 percent among the illiterate. With respect to age at marriage (Table 8.7), the positive relation- ship with conventional methods and the negative relationship with sterilization cancel one another, leaving a flat curve for the two methods combined. Ownership of land is not a relevant factor in family limitation practice (Table 8.18), but caste (Table 8.21) and per capita household expenditures (Table 8.17) were relevant. Scheduled Castes and Ezawas have a low utilization of conventional family planning techniques, but this was compensated by a higher sterilization rate among them. Christians have high rates for both types of birth control methods while Muslims have low rates. In general, the proportion of women who use conventional family planning methods among those practicing birth control increases with socio- economic status: the relative use of conventional methods is higher among the better-educated, the higher castes, those households with more land, and those with higher per capita monthly expenditures. Factors Related to Family Plaining In the previous sections, variation of family planning practice by age, sex, education, caste, etc. were examined. Most of these variables were related to family planning practice in the sense that the proportion using or practicing family planning varied by these factors. In this section, we shall examine their simultaneous relationship with the help of multiple regression, thus identifying and measuring more precisely the effect of those variables which have independent effects on family planning knowledge and practice. We shall examine the relationship of the following family planning variables: - 192 - Y1 = knowledge of family planning methods 0 = don't know any method 1 = knows at least one method Y = Number of methods known 0, 1, 2, Y3 = Ever-use of conventional family planning methods 0 = never- used 1 = ever used Y = Current use of conventional family planning methods 0 = not using 1 = using ,one or more methods Y = sterilization 0 = not sterilized 1 = sterilized Y = use of conventional F.P. methods or sterilization 0 = neither sterilized nor using a conventional family planning method 1 = sterilized or using a conventional family planning method Independent variables used in the regression are: V = Toilet facilities 1 to 5 inversely related to quality of-facility V6 = Area of land owned by household 6 = Years of schooling VS8 = (Years of Schooling)2 V = Parity 92 VS9 = (Parity)2 VY31 = Age AGES = (Age)2 - 193 - EX = Excess fertility (parity - desired family size) PCE = per capita household expenditures per month V = Caste 1 = Nair or Syrian Christian 0 = All others Knowledge of Family Planning Methods: Knowledge of family planning methods is measured by two variables: Yl = whether knows any method or not (0, or 1) Y2 = number of family planning methods known (0, 1, 2...) Annex Tables 1.27 and 1.28 give the regression between y, Y2, and the independent variables enumerated earlier. yl has a statistically sig- nificant relationship with the following: Education 2 positive (Education) negative Excess fertility positive Per capita expenditures positive Age, (parity) 2 positive (Age) , (parity) negative Knowledge of family planning methods has no statistically significant relationship with: Land owned Caste Toilet facility Education is the principal variable affecting the knowledge of family planning methods. The total variance explained by all the independent variables together is 8.3 percent, of which 5.2 percent (63 percent of the 2 total) was explained by education, and 7 percent of the total by (education) - 194 - The regression coefficients are +0.04258 and -0.00180 for education and (education)2 respectively. At lower levels of education, an additional year of schooling would increase the knowledge of family planning methods by 4.3 percent. At higher levels of education the improvement would be marginal,. however. 1/ Age (parity) is statistically related to knowledge of family planning methods, both linearly and at the second degree. At younger ages (lower parity), there is a positive relationship; knowledge of family planning methods increases with age. At higher ages (above 30 years) the relationship is negative; that is, family planning knowledge decreases with age. As explained earlier, the direct relationship at younger ages and inverse relationship at older ages are due to two conflicting trends. Family planning is relatively new to Kerala women, and younger women with better education are more likely to know about these methods than older women with less education. Yet younger women have less need to acquire knowledge of family planning. The need is greater among older women and, since they have been exposed to longer periods of family planning publicity, the knowlege of family planning methods should be greater among them. These conflicting trends result in the maximum knowledge among women at around 28 years of age. The relation between the number of family planning methods known (y2) and the set of independent variables is closer than with knowledge of family planning. The set of independent variables could explain about 21 1/ The parameters indicate that at educational levels of 12 years or more, additional years of education would reduce the knowledge of family planning. The regression is not valid at these levels of education. - 195 - percent (compared with only 8 percent for knowledge) of the total variance. Here again, the principal factor is education (explaining 87 percent of the total variance explained by all the variables). All the other variables excluding land owned by the household have statistically significant relation- ships as well: there is a positive relationship with education, excess fertility, caste, per capita household expenditures, and age; and a negative relationship with toilet facilities (actually a positive relationship as the variable is inversely ranked), (age)2 and (education)2 Caste had no significant relationship with knowledge of family planning, but it had a significant positive relationship with the number of methods known. Nairs and Syrian Christians knew more methods of family planning than the other castes and religions in Kerala. Ownership of land by the household was not a factor in the knowledge of family planning methods. Whether the family is landless or owns one acre or more is not very relevant in either the knowledge or the number of methods known. Conventional Family Planning Use: The principal variables influencing the use of conventional family planning methods are education and excess fertility (see Annex Table 1.29). As with knowledge of family planning methods, at lower levels of education, the relationship is positive; that is, the use of a family planning method increases with education. It is negatively related in the second power, howevero Education accounts for 56 percent of the total variance explained by all the variables, of which 55 percent is explained by the linear relationship and only one percent explained by the second degree. - 196 - The regression coefficients indicate a turning point at 16 years of schooling. Thus, although the second degree relationship is statistically significant, for all practical purposes it can be ignored: education and use of conventional family planning methods are positively related in the first degree for much of the educational range. The regression coefficient is 0.051068. At low levels of education, an increase of one year in schooling increases family planning use by 5.1 percent. With other variables remaining constant, the level of contraceptive use in Kerala at various levels of educational attainment would be as shown in Figure 8.1. Excess fertility accounts for about 30 percent of the explained variance. The larger the excess fertility, the greater the chance that a woman will be using a contraceptive. While excess fertility is not an operable variable to improve family planning use, it provides a condition under which the woman would be willing to accept a family planning method. The other variables which make a significant contribution are age and caste. Age has a positive relationship; the use of family planning increases with age. It is also positively related to caste status; higher castes tend to use family planning more often than lower castes. Sterilization: Most of the variables except caste and toilet facilities have a statistically significant relationship with the propensity to get sterilized. Those with positive relationships are the following (see Annex Table 1.30): Excess fertility, Age, and Education. - 197 - Figure 8.1: Partial Regression Between Years of Schooling and Birth Control Practice, Kerala, 1980 Percent 0Percent USing conven- tional family planning :40.-.. -----. . -30 Percent Sterilized .20 2 -.68 10 --12 Years Of Schooling . Source: "Fertility Survey in Kerala, 1980," Bureau of Economics and Statistics, Trivandrum, World Bank and UNFPA. - 198 - The chance of a woman getting sterilized (or her husband) is 0 positively related to the excess number of children she has over the number she desires. The regression coefficient is 0.03607. When the excess fertili- ty increasesby one child, the chance of sterilization increases by 3.6 percent. Age has a positive relationship only at the younger ages. The propensity to be sterilized increases with age until age 36 years when it starts decreasing. Similarly, education is positively related to sterilization at lower levels of education (5 to 6 years of schooling). At higher levels of school- ing, the more educated a woman, the less her chance of being sterilized (see Table 8.22). The principal economic variables are land owned and per capita household expenditures. The relationship with both these economic variables is negative. The chance of a woman becoming sterilized decreases with an increase in the area of land owned by the household (about 1.5 percent decrease for an increase of one acre in land owned). The same type of relationship holds with per capita household expenditures (3.2 percent decrease for an increase of Rs1OO in per capita expenditures). In this analysis, Christians and Nairs are grouped together into one group and all others into another group. With this classification, caste Eid not have a significant effect on sterilization. As far as propensity to be sterilized, it is immaterial whether a woman belongs to a high or low caste. This analysis does not show that caste distinction has no relevance to sterilization. If all the castes are separately treated, then caste shows a significant relationship, even when allowances are made for other variables. - 199 - Table 8.22: Summary of Regression Analysis on Family Planning Variables, Kerala 1 2 3 4 5 Sign of the regression coefficients Independent Variables 1. Education + + + + 2. (Education)2 - - - - - 3. Excess fertility + + + + + 4. Caste + + + 0 + 5. Land owned 0 0 - - 6. Per capita expenditures + + 0 0 0 7. Toilet facilities + + 0 0 + 8. Age + + + + + 9. (Age)2 - - 10. Parity + + + + + 11. (Parity)2 - - - Percent of R2 contributed by the factor 1. Education 67.1 87.0 54.8 1.3 29.4 2. (Education)2 8.7 0.4 0.9 11.3 4.6 3. Excess fertility 17.0 2.8 30.3 43.0 34.7 4. Caste 0.2 1.1 5.2 0.1 2.7 5. Land owned - 0.1 0.7 3.9 2.0 6. Per capita expenditures 2.4 0.6 0.7 1.0 0.1 7. Toilet facilities 0.4 1.5 1.4 0.7 1.1 8. Age 4.0 6.3 5.9 1.8 1.7 9. (Age)2 0.1 0.1 0.1 36.9 23.7 10. Parity (50.5) (3.8) 14.3 (26.1) (25.9) 11. (Parity) (11.6) (6.0) 14.9 (60.3) (46.6) Total 100.0 100.0 100.0 100.0 100.0 R2 0.077 0.210 0.147 0.123 0.177 Notes: A. 1 Whether knows any family planning method 2 Number of family planning methods known 3 Use of conventional family planning methods 4 Sterilization status 5 Sterilized or using conventional family planning method B. + positive relationship - negative relationship 0 insignificant at 5 percent level C. Parity and age are never together in the same regression - 200 - Use of Conventional Family Planning Methods or Sterilization: When both conventional family planning methods and sterilization are combined, all variables except per capita household expenditures become statistically significant. The principal factor is excess fertility which accounts for 35 percent of the explained -variance. Education accounts for 34 percent, 29 percent at the linear level and 5 percent at the second degree level (see Annex Table 1.31). The overall conclusion is that education is the most important factor in the knowledge and use of family planning methods. Of the total variance explained by the independent variables, education contributed 76 percent for knowledge of family planning methods, 87 percent in the case of number of family planning methods known, 56 percent in the use of conven- tional family planning methods, but only 12 percent in the case of steriliza- tion (see Table 8.22). In the latter case, the second degree negative relation- ship is more prominant than the first degree positive relationship (11 percent for second degree and 1 percent for first degree). The overall level of R2 is low, varying between 7 to 21 percent. Thus, there are quite a number of other factors which determine whether a woman uses a contraceptive or she or her husband is sterilized. Family Planning Use and Fertility With nearly half the currently married women either sterilized or using a conventional family planning method, family planning should have a significant effect on fertility in Kerala. What is the actual contribution of family planning.to Kerala's fertility reduction? - 201 - Figure 8.2 gives the distribution of women by marital status family planning status. The overall distribution is as follows: Total women, 15-49 years 1,000 Never married 379 Widowed, divorced 69 Pregnant 7 Sterilized 174 Conventional family planning users 117 Women exposed to pregnancy 254 Thus, all but 25.4 percent of the women were protected from preg- nancy, if those who used conventional family planning method used them efficiently. This does not necessarily mean, however, that all but 25.4 percent of the births were averted, as the fertilty rates are not the same at all ages. The relative contribution of the various factors to birth control is shown in Figure 8.3. The age-by-age analysis does not change the conclusion substantially; the proportion of births averted is more or less equal to the pro- portion of women protected. Summary results are reproduced below: Balance 1. Total fertility rate 9.53 100% 2. Reduction due to celibacy, widowhood, divorce, etc. 4.32 45 5.21 55 3. Reduction due to sterilization 1.63 17 3.58 38 4. Reduction due to family planning 1.14 12 2.44 26 If family planning were 100 percent efficient, the TFR in 1980 would have been about 2.44 but the actual rate is higher, about 3.2. Thus, conven- tional family planning has not been totally effective in controlling births; it may have been useful in postponing some of the births. 叫 203 Lý.Zure 8.3: Potential Fertility Reduction Due to Ce-libacy, Sterilization, and Family Planning Use, Kerala, 1980 Sp. F.R. 9.53 (TFR) Celibacy (4.32) 0.3 5.21(TFR) 0.2- 7amily Sterilizým±an Planning 3.58 (TFR) 0.1 (1.14) 2. 44 (TFR) 0 20 ý5 _30 35 40 43- 50 ;,Age Note: Figures in parentheses are the potential reduction in TFR due to that factor. This is equal to the area between the curves. Source: "Fertility Survey in Kerala, 198011, sponsored'by. the* World Bank, UNFPA, and the Bureau af Economics and~ Statistics, Trivandrum. - 204 - Part III: INTERPRETATIONS AND CONCLUSIONS CHAPTER IX DETERMINANTS OF FERTILITY DECLINE: CONCEPTUAL FRAMEWORK Conceptual Framework and Hypothesis A general hypothesis followed in this research is that decisions about family building--whether, when and whom to marry'; whether, when and how many children to produce, etc., -- are strongly influenced by the perceived economic advantag-ps and disadvantages. The increase in age at marriage and decrease in marital fertility observed in recent years in Kerala are due to the perceived disadvantages of the past practice of marrying early and raising a large family in the changed economic and social situation in the state. As a general hypothesis, few can find fault with this framework. The difficulty comes when specifics are enumerated: what economic advantages ,or disadvantages are associated with what socio-economic changes and 4.r which section of the population? Ratcliffe ( 1978, p. 140) concluded that: "As education and income levels rose as a consequence of distributive policies fertility and mortality declined in response. As viable investment opportunities other than children and occupational opportunities other than child- bearing became available, fertility declined in response. As the political economy was transformed from one founded on exploitation of the many by the few to one which began increasingly to assume welfare functions traditionally fulfilled by children, fertility declined -- and continues to decline -- in response." Thus, according to Ratcliffe, the major cause of fertility decline in the state was the improvement in the level of living of Kerala's poor. In other words, fertility and income are negatively related; as income of the large number of poor increased, their fertility decreased. As the poor constitute a large proportion of Kerala's population,any decrease of fertility among them would be reflected in the state's average rate. He gives much credit to the political - 205 - situation in the stat; the social and economic reforms brought about by the successive governments and the social justice which resulted from these reforms. He does not give any special credit to the family planning program. He seems to imply (although he does not state explicitly), that much of the fertility decline would have taken place even without the official family planning program. In a recent article on "The Lessons and Non-Lessons of Kerala" Joan Mencher (1980, p. 1787) has a different point of view. "...the decline in fertility among agricultural laborers need not be seen as an indication of any improvement in their quality of life. In the Kerala context, it can equally well be seen as a sign of greater poverty. There is some element of truth in what both Ratcliffe and Mencher say. The redistributive policies did play a part and some of the fertility decline in Kerala may be characterized as "poverty induced" or, more correctly, as due to a decrease in income. But this is not the whole story. We believe the decline is as much due to historical developments as to recent policy interventions. Historical developments prepared the ground whereby a reduction in fertility became not only possible, but also beneficial to most families. But these developments by themselves would not have produced the type of fertility decline 1/ that took place in the state over the past 15 years. They were precipitated by more recent policy interventions. Historical developments behind the fertility decline were numerous; but for analytical purposes,the essential elements in them may be characterized as a systematic erosion of the economic significance of the inherited characteris- tics of a person (his religion, caste, family ties, etc.) and the simultaneous 1/ The degree of decline as well as the nature of the socio-economic differen- tials of the decline. - 206 - enhencement of the economic significance of his personal attributes (education, training, experience, health, etc.) As Robin Jeffrey (1976, p.267) writes in The Decline of Nayar Dominance: "In the last half of the 19th century a society which had survived fundamentally unchanged for 700 years came un- hinged. A movement from inherited to achieved status began, a movement from the interdependence of castes to the compe- tition of individuals, from traditional authority to modern bureaucracy." One good illustration of this transition may be seen in the criterion for admission to the government service of the former Travancore and Cochin States. Government service has been and still is one of the most coveted occupations in the state. Earlier, these services involved not only the usual responsibilities of law and order, tax collection, etc., but also administra- tion of temples and related responsibilities. As non-Hindus and low caste Hindus were not allowed in Hindu temples, government service was restricted to high caste Hindus, Brahmins and Nairs in particular. And among the high caste Hindus, those who were close to the Maharajah had a much higher chance. Educa- tional attainment or other personal attributes were secondary to inherited characterisics. This practice began to change with the expansion of the government-s responsibilities away from temple-related functions to new areas such as education, health, public works, and so on; with the introduction of a minimum educational qualification for entry into government service; and with the overall bureaucratization of the administration. Today, the transition is almost complete; a person's chance of getting into goverment service depends almost entirely on acquired characteristics -- education, training and experience, health, etc. Therefore, a young person's lifetime income is determined more by his educational attainment, experience, and health than by the caste to which he belongs, or the family into which he was born,or - 207 - the land or other he owns or will likely inherit. The transition from inherited to achieved status began with the British influence on the administration of Tranvancone and Cochin States and with the development and spread of education and medicine by the foreign missionaries. The local Christians who were closer to the missionaries than the other communities benefited most from these changes. Through education, they began moving up the social and economic ladder. But, the other communities were not far behind in seeing what was happening to the Christians. Realizing the economic benefit which the Christian community was receiving from formal education, other communities began establishing schools and hospitals for their people. The result was a healthy competition between communities and a rapid development of education and health in the state as a whole. Unlike other parts of India where inter-religious and inter-caste rivalries produced deaths and destruction, communal rivalry in Kerala produced social progress and individual liberation. "Take away communal spirit and you will kill communal individuality, reduce all communities to one deal level and retard progress." (Dewan Roghaviah quoted in Progressive Travancere).2/ How did this transition induce a fertility decline? Chiefly through changes in the benefit-cost ratio of children (see Figure 9.1). The cost of bringing up children has increased relative to income; and benefits from children have decreased, adversely affecting the benefit-cost ratio. In addition, the chance of a newborn surviving to adulthood has increased several fold and the pressure on land resources increased considerably (Kerala exceeded the 1981 population density of India, 221 per sq. km.,as early as 1935). 2/ S. Ramanath Ayar,1923. Progressive Travancere, Trivandrum: The Anatha Rama Varma Press, p.43. - 208 - Figure 9.1: Impact of Social Reforms on Fertility in Kerala Social changes during the past several decades resulted iii a movement from: Inherited Status to - 4 Achieved Status Int:erdependence of- Castes- Competition of Individuals Traditional Authority ) Modern Bureaucracy Age at which economic Cost of children increased status of individuals was with added cost of education fixed increased from birth and benefits from children to time when education was decreased while they were in complete and job secured. school and unemployed. Increase in age at marriage. Decrease in desired family size and increase in use of D a contraception. Decrease in fertility. I - 209 - As a result, a large family size became less desired now than in previous years (see below for effects through the age at marriage). In the changed economic perspective, children of higher quality became not merely a desirable thing to have, but also a necessity for economic survival of the family, not only for the rich and well to do, but also for the poor and daily wage earners. Good health and a better education have become critical factors in determining the children's lifetime incomes and economic survival both for them and their parents. Earlier, if children were not educated, they could fall back on the family's land or connections to make a decent living. There was not much differential between the economic status of the educated and uneducated in the same family. Things have changed very much in recent years. Now it is not all that easy to make a living on the basis of cast,e or religion, cr from land alone, and in any case land reforms have made it impossible to hold onto more than five acres of land per person (rubber, tea, coffee plantations are exempted). For the poor, a good education formerly was beyond the reach of most and they had little hope of getting out of the circle of illiteracy, poor health, and poverty. With the recent political changes, and with all the social and economic reforms favoring them, they see a chance of getting out of the vicious circle. Education has become cheaper as high schools and colleges are within a commuting distance of most. With the highly subsidized and widely available transport system, most children can pursue a higher education while living in their homes. Although education has become cheaper, the poor and MI the lower class are now spending more on education than previously. This may appear paradoxical. The reason is simple. In the past, education was expensive - 210 - and few of the poor or the lower class could send their children beyond primary school,which was always free. Now, the cost of education has become lower, and many of the lower class are utilizing higher educational facilities open to them within easy reach. Therefore, their expenditures on education have increased. Thus, for the rich and the poor-alike, higher quality children have become not merely something desirable but a necessity to keep up the status quo. This means that parents spend more money on their children's education, food, clothes, and medical services. Children have become more expensive for all sections of the society. At the same time, benefits from children have decreased. As children began spending longer and longer periods in schools, their economic contributions to the family became less and less. Even without schooling, because of the tight employment situation (see below for reasons) the children-s ability to obtain paid employment has dimished. They have less chance now than previously to find a job outside the home and contribute to the family income. Thus, through an increase in cost and a decrease in benefits, the ratio of benefits to cost of children has decreased, adversely affecting the public's perception of the economic advantages of raising a large family. These historical developments have produced a long-term downward trend in the desired family size,which in turn would be expected to be reflected in the actual family size. This it was, but principally through delayed marriages and reduced childbearing due to the use of contraception among the well-to- do. It is doubtful whether the recent sharp decline in marital fertility which embraced all socio-economic groups was due solely to this long-term atti- tudinal change concerning family size. More likely, the recent decline was pre- cipitated by two policy interventions: - 211 - (1) land reforms,agrarian reforms and related redistributive policiessand (2) the official family planning program. It is no coincidence that the sharp fertility decline began with the implementation of land reforms and the strengthening of the official family planning program. Land Reforms, Agrarian Reforms and Related Policies As pointed out earlier, Ratcliffe gives consi4erable credit to the distributive policies; fertility decline was caused by an increase in the poor's income due to the redistributive policies of the government. We give much less importance to income, especially increases in income among the poor. But there are other elements of the land reforms and other redistributive policies which would have changed the fertility preferences of much of the Kerala population, causing them to accept effective fertility control. One such element is economic deprivation resulting from the land reforms- the loss of land, the loss of income from land, etc. Excess land (over the ceiling set by the government) was taken by the government and distributed among the landless. Others who did not possess any excess land but were having what they owned cultivated by tenants lost these lands for token compensation. The most numerous of all land transfer was of course that of house sites of the "hutment dwellers". Owners who had hutment dwellers on their land lost the house sites and some land around the house (about one-tenth of an acre in rural areas, less in urban areas) to the hutment dwellers who became legal owners of the property. Figure 9.2 gives a schematic representation of the impact of economic deprivation on fertility. For those who lost their source of income, the effect was thr.ough an increased cost of children relative to income. Persons who Figure 9.2: Impact of Land Reforms, Agrarian, and other Reforms on Fertility Trends in Kerala Agrarian and Other Political Refora:8 High Wages, Better Redistribution Reservations for Schedulpd Working Conditions of Land Caste/Tribes and Backward Castes Job Reservation, Educatipu Increase in Unem- Increase in Law About Economic Reservation, Housing, ployment,Especially Opportunity House Sites Deprivation Government Services of Among Children Cost of Women All Kinds Housing Problem For New Generation - Decrease in 'H_r4 Income c k Economic Deprivation a)_" k of High Caste Increase in Increase Persons per in Age at 0 U 4) 0House and Lack Marriage 4Increase in a of Privacy Relative Cost _) U) of Children Increase in Relative Cost , of Bringing up Children Reduction in Reduction in Desired Family Fertility Size - 213 - lost land or other income sources through land reforms were economically well-off people accustomed to giving their children good educations, food, and health services. The sudden loss of income did not change their desired behavior with respect to expenditures on their cildren's development. If some groups have lost land and other sources of income, others must have gained by these changes (e.g., former tenants and hutment dwellers). However, it is doubtful whether even they have gained substantially by the legal ownership. On the one hand, the risk involved in agriculture and the cost of cultivation (see below) has increased considerably. On the other hand, the price of agricultural products is controlled by the government. It is, therefore, doubtful whether the new owner-cultivators who have their farming done by hired labor in fact realized any significant increase in income due to the land reforms. In many cases,-there was a decrease in income. The case of the hutment dwellers is slightly different. There was no economic deprivation for them. Due to land reform, however, it became diffi- cult for their older children to get house sites;and it was only a matter of time before overcrowding became evident in the hutments. The result was a desire for fewer children (see below) even among the hutment dwellers who have benefited economically. A second element of the land reforms and agrarian reforms, and per- haps more important than economic deprivation, was the wage increase. Wage in- creases affect the fertility trend in two ways. First, the increase in wages adversely affected the benefit-cost ratio of children. As Ratcliffe writes in his article (1978, p. 139): "Finally, opportunities for child labour have been virtually eliminated due to enforcement of minimum wage and child labour laws... And when value of children as employable economic assets declines, so does fertility". - 214 - A similar story is told by Mencher (1980, p. 1787) for agricultural laborers: .....though the statutory minimum wage rates are high in Kerala, the number of days for which employment is avail- able for agricultural labourers is very small... In this situation where agricultural labourers are in a sense driven to the wall, I am: arguing that while people still want children, they-do not see-any economic advantage in having large families and-this--coupled with the.fact that with better medical facilities, the children that they have are likely to live to adulthood- makes even agricultural labourers (including the illiterate) amenable to family planning". Wages have been kept artificially high by the powerful trade unions in all economic sectors, especially agriculture, and a sympathetic government. As a result, the volume of employment has decreased considerably and employment of children and adolescents has disappeared altogether. The value of children as an employable asset has decreased. Second, the higher wages and better working conditions introduced by the agrarian reforms have increased the income potential of women, thereby increasing their opportunity cost. The income potenial of women has also in- creasesd through education, but this applies only to those who complete high school or college. Thus, as a consequence of agrarian reform, income from children has decreased and the opportunity cost of mothers has increased (income foregone). Both these tend to reduce the desired family size of the rich and the poor, especially the latter, and to make them amenable to family planning. The Role of the Official Family Planning Program A good deal of the credit for the sharp fertility decline in Kerala among the agricultural laborers and other economically backward sections of the population should go to the official family planning program. The socio-eco- nomic and demographic conditions created a situation in which a fertility decline -21.- was possible and desiable. But without the official family planning program, it is doubtful whether a fertility decline of the magnitude observed among the lower class would have taken place. It is not necessary to explain how a fertility decline is caused by family planning, especially in Kerala where the principal method of family planning has been sterilization. What is necessary is to show that without the government program, family planning practice in Kerala would not have been so widespread. This is done in a numbers of ways, principally through areal com- parisons, temporal comparisons, etc. (as below). Age at Marriage In Kerala, as in most other populations, fertility declined not only through family planning (a decline in marital fertility) but also through a de- crease in the proportion of married women at younger ages (an increase in the age at marriage). The factors underlying this increase in age at marriage are not very different from those underlying an increase in the use of contraceptives. There are,of course,differences in timing; socio-economic factors affected age at marriage first, even before they had any perceptible impact on marital fertility. In the use of contraceptives, both demand and supply factors were involved; in the case of age at marriage,demand factors were the only ones involved. In Kerala, marriages are arranged and represent economic and social contracts between two families rather than a personal agreement between a boy and girl. In these contracts, the social status and lifetime income of the boy and girl (and their families), particularly of the boy, have been and still are very important considerations. Earlier, it was easier to estimate status and income when the boy and girl were very young, as they depended on ascribed status and inherited wealth, particularly land. Therefore, marriages could be - 216 - contracted at very young ages. As described earlier, land is no longer the principal source of status and income. Nowadays, judgements are based essen- tially on acquired status, through one's education and career. Thus, the economic value of a person can be realistically assessed only after he (or she) has completed his (or her) studies and takes up a job. This results in marriage being delayed until one has finished school and been employed. In a period of increasing years of schooling and an increasing lag between completion of education and the securing of a job, the age at marriage is likely to increase. Not all people are in school at ages 15 or 16; many have completed their formal education before that age. For them, a slight increase in their number of years in school need not necessarily affect their age at marriage, but what they will be doing for a living in life will be known only when they are older, usually 20 or over. Even apart from this there have been other economic reasons for delaying marriages to more advanced ages in recent years. One such factor is the dowry system. This practice is very common in most social groups in Kerala and the amount paid as dowry has increased considerably in relation to household assets. In the past, much of the dowry came from inherited wealth; today, however, the sum usually comes from savings and/or borrowings, and consequently marriages are delayed until the girl's parents can amass sufficient funds or credit-worthiness to satisfy the bridegroom's family. For a boy, his chance of receiving a good dowry is much brighter when his economic worth is fully manifest; that is, when he completes his education and acquires a suitable job. - 217 - Young couples' views on housing arrangements after marriage are changing rapidly. Newlyweds used to stay with either of the parents, but many are becoming unhappy with this arrangement. More of them are working away from their parents' home and separate housing becomes a necessity. Choice or necessity, housing is an important consideration in age at marriage. Housing has become an increasingly difficult problem not only for the urban dwellers but also for the rural residents. In urban areas, there are housing shortages and rents are beyond the means of the lower and middle classes. In rural areas, the housing problem is caused by shortages of land and changes brought on by land reform. Formerly, landless persons used to be able to build a hut on a wealthy neighbor's land without much difficulty. This arrangement was mutually beneficial. The owner of the land got a watchman to look after his land and crops; the landless person got a place to stay rent-free. This is not as available an alternative as it once was because land is scarce and the land reform laws make it difficult for the wealthy neighbor to get back his land when he needs it. Thus, among both rich and poor, problems related to housing tend to delay marriage. Marriage decisions are often made by the parents, particularly the fathers. This is changing, but as this survey has shown, even now moct of the decisions are made by the parents. It may be economically advantageous for the father to delay his son's marriage, because by doing this he prolongs the use of his son's wages for family expenses and saves himself the difficult problem of solving the housing dilemma of the newlyweds. Earlier, as children were not in school, they could contribute to family income at an early age. Now they do so only at relatively later ages when they have finished their education. - 218 - Thus, the conceptual framework regarding the fertility decline in Kerala is as follows: Socio-economic reforms in the state over the past 100 years have eroded the economic value of inherited characteristics (caste, religion, etc.) and inherited wealth (land), and enhanced the economic value of personal characteristics (education and health). Fueled by inter-communal rivalry in developing human capital, the.state experienced a sharp decline in mortality and an accelerated improvement in educational attainment. Mortality has declined so much that 90 percent of children born today survive to adult- hood, and education has improved so much that nearly all women in the child- bearing ages are literate. The low mortality conditions and high educational attainment provide the necessary milieu, for a fertility decline. The low mortality enables couples to attain their desired (surviving) family size with fewer children ever-born. The high education attainment enables parents to make rational decisions about family size, taking into consideration mortality risks and their given socio-economic pespective; it has enabled couples to utilize family planning services effectively and limit or postpone births as and when they feel desirable; it has increased the cost of bringing up children as education has become a necessity for economic survival, and moreover, educated parents tend to give an even better education to their children; it has increased the opportunity costs of women and reduced the economic benefits from children; and it has increased the risk of unemployment especially at critical younger ages. While these longterm socio-economic changes produced the necessary milieu and caused the age at marriage to increase, the sharp decline in marital fertility was precipitated by more recent policy interventions-land reforms, agrarian reforms, and the official family planning program in particular. The land reforms and other redistributive policies operated on fertility in diffrent ways. The economic realignment has caused some socio- economic groups to suffer loss of land, status, income, etc., with no hope of immediate recovery, forcing them to make demographic adjustments toward later marriages and fewer children. The increase in wages has decreased the income of land owners, even those who have not lost any land, forcing them alsov to adopt fertility control. Wage increases have increased the cash income of female daily wage earners, raising the opportunity cost of their time spent in bringing up c4ildren. The increase in wages also has reduced employ- ment opportunities forcing children out of the job market and making them economically less valuable. Even the hutment dwellers who undoubtedly gained by the land reforms and wage legislation began to feel the need for birth con- trol as their children have found it impossible to obtain a place for a hut of their own. The official family planning program came at an opportune time to serve not only the manifest demand (mostly among the educated) but also the latent demand (among the poor and the lower class) for birth control. - 220 - Part III: INTERPRETATIONS AND CONCLUSIONS CHAPTER X DETERMINANTS OF THE FERTILITY DECLINE: EMPIRICAL EVIDENCE Introduction According to. the conceptual framework described in the previous chapter, the fertility decline in Kerala was caused by recent policy interventions aided by some very favorable longterm socio-economic changes. The policy interventions identified to have precipitated the fertility decline are the official family planning program, land reforms, agrarian reforms, and other redistribution policies. The favorable socio-economic milieu was created principally through improvements in the educational and health status of the people. The data most suitable to test this hypothesis are time-series on fertility rates and the identified causal variables. But the data at our disposal are limited to those collected in a one-time survey (in 1980). The fertility decline took place mostly during a 20-year period prior to the survey; the land reforms and other agrarian reforms were implemented mostly between 1960 and 1970; improvements in education and health have taken place over a much longer period extending well over a century. A one-time survey has severe limitations in reflecting the consequences of changes which took place over extended and varying periods of time. Bearing in mind these limitations of our data, we give below the available evidence on the determinants of the fertlity decline from the survey, separately for each determinant and then together in a single model. The principal background variables are education and health; these are discussed first. - 221 - Education It was hypothesized that in the changed socio-economic perspective in Kerala, educating one-s children has become a necessity; and as education has spread so did the desire for a small family size and the ability to achieve the desired family size. Education has increased the cost of rearing children, reduced their economic benefit, increased the opportunity cost of women, increased the risk of unemployment at younger ages, helped couples to make rational decisions about family size and use contraceptives more effectively. If this hypothesis is correct, education should be negatively related to the desired number of children. The changed attitude about family size should increase the age at marriage and/or increase family planning practice and thus reduce the number of children born. We shall take up the variables in that logical order. Desired Family Size: Table 10.1 gives the desired family size by educational attainment of the mother. As expected, educated women desire a smaller family size than the illiterate--a difference of 1.4 children between the illiterate and those with at least ten years of schooling. That this difference is not entirely due to differences in the age distribution of the illiterates and the educated is shown by the standardized averages which also decline systematically with an increase in years of schooling. Age at Marriage: The lower desired family size of educated women is achieved through later marriage and/or fertility control within marriage. The age at marriage should be positively related to years of schooling. Table 10.1 which gives this relationship, indicates that illiterate women begin married life 4.3 years earlier than women with ten years of schooling. Among recently married - 222 - Tabe LO.L: Desired Family- SLze, arrd, Age at Marriage- by Educational Attainment, Kerala Average Family Size Avarage Age at Marriage (years) Education Actual Standardized Actual Standardized No education 3.80 17.5 L8.1 1-4 yrs. 3.39 18.3 18.5 5-9 yrs. 3.02 19.1 18.7 10+ yrs. 2.49 21.8 21.6 Total 3.25 3.25 18.8 18.8 women, the difference is smaller (2.7 years among those married in 1975-80 compared with 7.2 years among those married before 1960), but adjustment for the year of marriage does not remove all differences. The standardized averages show a difference of 3.5 years. Thus, one avenue by which educational attainment reduces fertility is through a reduction in the period of exposure to childbearing. Can it account for all the differences? Apparently, no. Rough calculations show that over the period 1950-1980: - 55 percent of the educational differential in fertility was due to an increase in the age of marriage; - 45 percent was due to reduced fertility rates within marriage (see Table 5.14 in Chapter V.) Family Planning: Fertility reductions within marriage are achieved principally through contraceptives. If educated married women desire fewer children, they should normally know more about family planning and practice contraception more - 223 - often than illiterate women. Table 10.2 gives the necessary data for comparison. The overall relationship between family planning knowledge (use) and educational attainment is positive. With respect to sterilization, however, the relationship is curvilinear with a maximum sterilization rate among women with six years of schooling. (See Figure 8.1.). Factors other . than education (e.g. financial incentives), which are not postively correlated with it, are involved in determining whether a person accepts sterilization or not. This could intervene in the expected overall negative relationship between fertility level and educational attainment. Table 10.2: Knowledge and Practice of Family Planning Methods by Educational Attainment of Women, Kerala % of Users of Conventional F.P. Among Ever- % of Women Married Non- Percentage Sterilized Who Know at Sterilized Women Currently Least One Current Ever Ever-Married Married Education Method Use Use Women Women No education 75 8.0 16.5 21.6 25.2 1-4 yrs. 89 22.9 46.5 33.8 39.2 5-9 yrs. 94 33.1 59.2 29.3 32.7 10+ yrs. 98 49.6 76.8 18.5 20.3 TOTAL 84 25.9 47.0 27.8 31.6 Excess Fertility: Among women who have more than or equal to the number of children they desire, some 58 percent used some fertility control method. The degree of effort to avoid undesired pregnancies is much greater among the more educated, however (Table 10.3). Thus, among women with ten years or more schoolings, 76 percent use a birth control method while among illiterate - 224 - women only 37 percent practice contraception, even though they already have more children than they desire. Table 10.3: Percent of Women Who are Sterilized or Who Use a Conventional Family Planning Method by Educational Attainment and Fertility Status, Kerala Fertility Status (Percentage of Women More Children Children Born Less Children Born Then Equal'to Born Than Education Desired Desired Number Desired Illiterate 37 30 9 1-4 yrs. 60 63 14 5-9 yrs. 65 69 28 10+ yrs. 76 72 39 TOTAL 55' 60 24 Fertility: Several measures of fertility based essentially on birth history are given in Table 10.4. They all show, in essence, that the more a woman is educated, the fewer number of children she has or the lower her fertility. Among older women (with 25 years or more of marriage duration), illiterate women have 1.4 children more than the women with at least ten years of schooling. The difference has narrowed in recent years. The fertility rates in 1975-80 indicate a difference of only about 0.7 children. Comparison between the two periods, 1965-70 and 1975-80, indicates not only a decrease in the magnitude of the differentials, but also a change in the patterns of the relationship between education and fertility. This was - 225 - Table 10.4: Fertility Rates by Educational Attainment, Kerala Educational Mean No. of Children Completed Attainment Ever-Born Family TMFR- (20-44 of years) of Woman Observed Standardized Size 1965-70 1970-75* 1975-80* None 4.49 3.61 5.83 6.23 5.45 4.79 1-4 yrs. 3.80 3.55 5.71 6.76 5.12 4.41 5-9 yrs. 2.90 3.34 5.12 6.26 4.62 4.67 10+ yrs. 2.13 2.80 4.40 5.03 5.75 4.08 Difference between None and 10+ yrs. 2.36 0.81 1.43 1.20 -0.30 0.71 *The irregularities seen in the pattern of relationships in these periods may be explained by the very large number of sterilizations carried out in the camps in Ernakulam and Palghat Districts. As shown earlier, the highest sterilization rate was among husbands of women who had 5-9 years of schooling. perhaps inevitable in the Kerala context where sterilization has been the principal means of fertility control and the sterilization rate varies parabolically as educational level increases, with a maximum rate at six years of schooling. In a detailed analysis of the relationship'between fertility and education, it is necessary to take account of this "disturbance" caused by internal factors such as financial incentives, official pressure, etc. In the discussion above regarding the impact of years of schooling on fertility and fertility-related variables, we have presented one variable at a time. Although some of the averages have been standardized to take into account differences in demographic composition (age, year of marriage, etc.), they have not been stand,ardized for socio-economic differences. For example, - 226 - the observed difference between the fertility level of the literate and illiterate may be due to some third variable which is correlated with both 4 fertility and literacy. That this is not entirely the case, however, is shown by the partial regression coefficients given in Table 10.5. These coefficients have been estimated after controlling for available socio-economic and demographic variables (age, caste, household expenditures, land owned by household, etc.). Table 10.5: Partial Regression Coefficients of Selected Fertility-Related Variables on Years of Schooling (Wife), Kerala Regression Dependent Variables Coefficient F-Ratio 1. Desired Family Size (1,2,...) -0.02868 10.1 2. Age at marriage (years), Women 25+ years' +0.68935 5.5 3. Knowledge of Any F.P. Method +0.04258 75.3 (yes/no) 4. Use of Conventional F.P. Method +0.05106 38.3 (yes/no) 5. Sterilization Status, Husband or Wife (yes/no) Linear: +0.03460 23.7 Quadratic: -0.00282 20.1 6. Sterilization or F.P. Practice Linear: +0.06649 75.9 (yes/no) Quadratic: -0.00361 29.7 7. Fertility Measure Children Ever-Born -0.07539 33.1 Children Born 1965-70 -0.03338 8.6 Children Born 1970-75 -0.00772 0.8* Children Born 1975-80 -0.01387 5.2 *Statistically not significant - 227 - All the partial regression coefficients are in the expected direction, and all are statistically highly significant except the one with births during 1970-75. There is a faint negative association between the number of births to a woman in 1970-75 (0,1,2,3...), and her years of schooling, but the relationship is far from significant. As mentioned above, this is possibly due to the disturbing effect of camp-induced sterilizations which were more attractive to the illiterate and the less educated persons (given financial incentives, other economic benefits, favors from public officials, etc.) than to those whose wives had ten years or more of schooling. We can now give the relationship between education and fertility in a simple recursive model (Figure 10.1). Family planning is not introduced in this model because, in the Kerala context, the direction of the relationship between children ever-born and family planning practice is uncertain. Fertility influences decisions about the adoption of family planning, rather than family planning determining fertility level. The correlation between them, even after controlling for socio-economic variables, is generally positive. Figure 10.1: Recursive Model of the Relationship Between Education and Fertility, Kerala Y2 Age at Marriage !Og Education *Children Ever- 10 j5 (...%h ) Born in 1980 X1 Y3 Desired Family Size L Y1 - 228 - The least squares estimation of'the parameters gives the following results: 4 Y= 3.7812 - 0.1162x1 (R = 0.282) Y2 =19.2887 + 0.2693x1 - 0.5518Y1 (R = 0.321) Y3 2.4064 - 0.0789x1 + 0.9688Y1 - 0.0923Y2 (R = 0.683) The direct effect of education on children ever-born was -0.1151 with an F-ratio of 0.57, and the indirect effect through desired family size and age at marriage was -0.2089. Desired family size and age at marriage are not the only ways by which education could affect fertility. For example, we have not included the effect through infant mortality, which is negatively related with education of the mother. Thus, the present analysis underestimates the impact of education on fertility. Health Among the Indian states, Kerala is unique in its mortality trend; by comparison, its mortality rate has been and still is the lowest. The probability of a newborn baby dying during its first 12 months of life is three times greater in India than Kerala. This low mortality level and the recent rapid decline in mortality in Kerala has been proposed as one of the principal factors underlying the relatively much larger fertility decline in the state compared to the rest of India. Are there any empirical supports for this hypothesis? A crude indication of the possible mortality impact on the fertility decline in Kerala is provided by regional fertility and mortality data. As - 229 - shown in Table 10.6 the mortality decline began in the southern parts of the state (Travancore-Cochin) and later spread to its northern part (Malabar). It is no mere coincidence that the fertility decline also began earlier and is much larger in the southern parts. Table 10.6: Birth Rates and Death Rates in Kerala, by Regions, Various Years, 1956-75 Malabar Travancore-Cochin Year D.R. B.R. D.R. B.R. 1956 23 47 10.4 35 1967 37 32 1970 12 36 8.0 29 1975 8 31 6.0 25 Sources: There are several reasons to believe that a mortality decline should lead to a fertility decline. First, there are biological reasons. It is well-known that breastfeeding tends to delay the return of ovulation. An early infant death cuts short the period of breastfeeding, reduces the period of post-partum amenorrhea, shortens the interval between births, and thereby increases fertility. Second, there are behavioral reasons. Three prevailing hypotheses can be proposed in this category. First, the replacement hypothesis: Mortality reduction leads to fertility reduction because parents need to bear fewer children to achieve their desired family size. Second, the insurance hypothesis: Mortality reduction reduces the uncertainty of child survival leading to less "hoarding" of children and therefore a reduction in fertility. Third, the positive cost-benefit hypothesis: Mortality reduction decreases the cost of producing surviving children, increasing the demand for children and hence fertility. - 230 - Empirical data to demonstrate the longterm impact of mortality reduction on fertility is hard to obtain in a fertility survey. What is available are cross-sectional data on children ever-born and intervals between births of women classified by different child death experiences. These data can at best establish short-term relationships. Even for that, the methodologies are not completely "above board". With these type of data it is hard to prove whether the direction of causation is from mortality to fertility or from fertility to mortality. Our analysis of the determinants of infant mortality indicated that the probability of a child dying during the first 12 months is significantly associated with birth order (see Annex II of this report, Table II.10). Fertility has indeed an influence on mortality. Much of the refinements in the analysis given below are efforts to minimize the reverse effects of fertility on mortality. Evidence of Mortality Effect on Fertility: Evidence of the net effect of mortality on fertility is given below by examining the effect of child death on birth intervals and parity progression ratios. A. Birth Interval As mentioned in the introduction, one way infant deaths affect fertility is through shortening not only the period of lactation and hence the period of amenorrhea, but also the interval between the birth of a child who dies and a subsequent birth. Therefore, a logical step in showing the impact of mortality on fertility is to analyze birth intervals. This is done in Table 10.7. All births of a given order (for example, first births) are classified according to their mortality experience (for example, whether the child died d.ring the first month, died during the first to eleventh month, or survived the first 12 months); also, the average birth interval between - 231 - Table 10.7: Average Birth Intervals Classified by Mortality Experience of Initial Birth, Kerala Average Time Interval in Months Initial Birth Initial Birth Ended in an Ended in a Infant Death Initial Birth Birth Orders Death under between 1 and Survived First Intervals All Births One Month 12 Months 12 Months 1 and 2 32.5 21.7 28.2 33.5 2 and 3 33.9 25.0 31.9 34.4 3 and 4 32.3 22.6 26.7 32.9 4 and 5 33.1 22.1 28.9 33.7 5 and 6 31.8 22.1 22.2 32.7 Average all birth orders 32.7 22.7 27.6 33.4 Difference from 12 m6nths survival -10.7 -5.8 - that birth order and the next birth order is calculated (for example, between first and second birth) for each mortality experience. A comparison of the average birth intervals indicates how much influence the death of a child has on the next birth interval. As expected, the average birth interval is shortest when an initial child died during the first one month and longest when it survived the first 12 months. A neo-natal death shortened the birth interval on an average of 10.7 months, while a post-natal death reduced it by 5.8 months. (See also Figure 10.2.) The relation between birth interval and the mortality status of the initial birth is examined further in Table 10.8 in a multiple regression analysis. The mortality variable was assigned a value of 1 if the initial birth ended during a death in infancy (12 months) and 0 if the birth survived - 232 - Figure 1.2: Birth. Interval, Following an Infant Death by Duration of Survival (months), Kerala Birth Intervals (Months) 34 32 30 28 26 24 22 200 2 4 6 8 10 12 14 Period of Survival (Months) Source: "Kerala Fertility Survey, 1980", sponsored by the World Bank, UNFPA, and Bureau of Economics and Statistics, Trivandrum. - 233 - Table 10.8: Partial Regression Coefficients of Birth Interval (i and i+1) on Mortality Status of the Initial Birth (ith birth) and Socio-Economic and Demographic Variables, Kerala Birth Order Regression Coefficient With Interval Mortality Caste Age at R Status Marriage 1 and 2 -10.71 (46.3)* - -0.513 0.19 2 and 3 -6.90 (12.8) 3.1 - 0.13 3 and 4 -9.10 (22.6) - - 0.16 4 and 5 -9.09 (9.8) 3.7 - 0.16 5 and 6 -10.99 (14.3) - - 0.19 6 and 7 -10.51 (5.3) - 0.20 *The figures in parenthesis are F-Ratios.. Note: Several other socio-economic variables were included in the regression (land, household expenditures, etc.), but none of them showed a statistically significant relationship with birth intervals. The number of women at higher parities was relatively small; hence, higher birth intervals are not shown. infancy. The partial regression coefficlents are all statistically significant and negative, indicating that the death of a child decreases the next birth interval. The average decrease is 10.7 months for the first parity, 6.9 months for the second parity, and then an increase to about 10 months is noted for higher parities. B. Twin Births and Subsequent Birth Intervals If the death of a child in infancy reduces the next birth interval as shown in Table 10.8, for the same reasons the occurance of a twin birth should increase the next birth interval. Twin births are relatively infrequent, but the few cases that were available indicated that such an - 234 - increase is noted in Kerala (Table 10.9). The average birth interval for all births in Kerala was 32.7 months; but for births following a twin birth in which both twins survived infancy, it was 40.1 months. Among twin births, the subsequent birth interval was smaller when one of the twins died in infancy and smallest whem both the twins died. Table 10.9: Average-Birth Interval. Following A Twin Birth, Kerala Subsequent Birth Type of Birth Interval (months) Twin births in which: - both children died 25.6 - one of the twins died 27.3 - both survived 40.1 All births 32-.7 C. Parity Progression Ratio A more direct fertility-related variable than birth interval is the parity progression ratio: that is, what proportion of n-parity women moves on to (n+1) parity. The higher these ratios are, the higher is the fertility level. Parity progression ratios for women having experienced child death are compared with those of women not having experienced child death. Table 10.10 indicates that the ratios of women whose children have died are invariably higher than those of women whose children have survived (or survived the first 12 months). The additional fertility is about 17 percent among women with infant deaths and 22 percent among women with deaths of children at any age. The relation between additional fertility and parity is bell-shaped with the largest additional fertility occurring between birth orders 3 and 4. - 235 - Measurement of Additional Fertility: Although the total number of children ever-born has a major impact on the number of deaths among children (r=+0.564), children born after a given parity (say n) will have no direct influence on the number of deaths among the first n children. Table 10.11 gives the additional fertility of women with one death among the first n children over women with no such deaths. It shows that among women with at least three children (for example), one death among the first three children increases the average parity by 0.81 (Row 1). Part of the observed increase is due to a difference in the age distribution of the women compared. Standardization with respect to age distribution indicates that the increase in fertility would be only 0.64 if there were no difference in the age distribution (Row 2). The difference seen in the fertility of the two groups of women could as well be due to a variation in subsequent child death experience. Women with one child death under parity 3 will have a higher chance of another death above parity 3 than women without any deaths under parity 3. The third row of the Table gives regression coefficients after controlling for age and deaths among birth orders (n+1) and higher. For n=3, the coefficient is 0.31. Further standardization with respect to socio-economic variables reduces it to 0.28. This is the "independent effect" of a child death under parity 3 on subsequent fertility. Such calculations for other values of n shown in the Table indicate that additional fertility due to child death is relatively low for n-1, highest when n=2, and decreases for higher values of n. The additional fertility at n=4 (0.158) is out of line with those for n=3 and n=5. One very probable explanation is that this is an unusual cohort - 236 - Table 10.10: Parity Progression Ratio by Birth Order., Kerala Birth Parity Progression Ratio (per 1,000) Among order Women with or without child death'/ Women with or without deaths at infancy4/ (i) to prior to birth of the (i+1) child prior to birth of the (i+1) child (i+1) With Without Percent With Infant Without Infant Percent where (i)= Death Death Difference Death Death Difference 1 934 843 10.8 915 850 7.6 2 954 754 26.5 934 768 21.6 3 917 698 31.4 924 712 30.4 4 832 675 23.3 803 688 16.7 5 750 651 15.2 717 659 8.8 All 875 719 22.2 857 731 17.2 1/ All deaths among children. / Deaths during the first 12 months of life only. Table 10.11: Additional Fertility (Partial Regression Coefficient) of Women with One Death among the First N Children over Women with No Deaths, Controlled for Age, Subsequent Deaths among Children, and Socio-Economic Variables, Kerala Control n Variable 1 2 3 4 5 6 1. None 1.236 1.222 .808 .406 .455 .301 2. Age .608 .845 .636 .363 .414 .316 3. Age and Death among Higher Order Births (n+1 and .225 .410 .312 .128 .262 .160 over) 4. All Above plus Socio- Economic Variables a/ .143 .335 .279 .158 .301 .260 5. R .76 .73 .67 .63 .61 .58 6. Average Additional Parity b/ 2.674 2.006 1.641 1.467 1.212 1.028 a/ Land-owned, household expenditures, caste, education, etc. b/ Average parity of women with at least n children ever-born minus n. - 237 - which experienced the brunt of the sterilization camps. It is therefore advisable to omit it from further consideration. The average additional parity (weighted by parity) for all values of n except n=4 for the range 1 to 6 is about 0.251. Thus, each child survival reduces fertility (average parity) by 0.25. Decomposition of Additional Fertility--Biological and Behavioral Effects: Mortality affects fertility through biological as well as behavioral factors. How much of the total effect on fertility is due to biological factors (such as lactation) and how much is due to behavioral factors? The Kerala data show that both biological and behavioral factors are operating to some degree. Biological factors operate in all populations which practice breastfeeding, and in Kerala breastfeeding of varying lengths is common in all communities. Therefore, it can be taken for granted that some of mortality's impact on fertility is due to lactation. Regarding behavioral effects, two evidences can be cited. First, the partial regression coefficient of the sterilization status of a woman (whether or not the woman or her husband is sterilized) and the number of deaths among children ever-born to her is highly significant (t=5.8) and negative. The chance of a woman or her husband getting sterilized decreases by 0.069 with the death of one of the children. This is a behavioral response. A second evidence comes from the birth interval after a twin birth. The average birth interval immediately following the birth of a surviving twin is seven months longer than that of a surviving single birth. The arrival of two children together seems to modify a woman's behavior regarding the timing of the next birth. - 238 - The Kerala fertility survey did not collect information on lactation. Hence, it is difficult to arrive at a direct estimate of the average period of lactation-caused amenorrhea, and the biological effects of mortality on fertility. We have, therefore, borrowed amenorrhea parameters from other populations in this general region. In Sri Lanka, the average period of lactation-caused amenorrhea is about 14 months.'/ An Indian study showed that the corresponding period was 11.7 months.2/ A Bangladesh study gives an average of 17 months.3/ None of these studies mentioned any variation by age or parity. In this analysis, we assume that in Kerala the period of amenorrhea is constant at about 13 months for lactating women and 2 months for non-lactating women. Thus, the additional period of amenorrhea for lactating women is 11 months. The survey data showed that women with at least one birth had an additional 2.674 children (first row of Table 10.12). These additional children would have been born over a period of 114 months (see second row) if the birth interval were not affected by child deaths. Thus, the amenorrhea effect of the death of one child (in this case the first child) on subsequent fertility is: 11 X 2.674 = 0.258 114 1/ John Bongaarts, 1980. "The Fertility Inhibiting Effect of the Intermediate Variables", Population Council Working Paper No. 51, May, New York: The Population Council, Table 2, p.10. 2/ R.G. Potter, et al., 1965. "Applications of Field Studies to Research on the Physiology of Human Reproduction" in Mindel C. Sheps and Jeanne Clare Ridley (editors), Public Health and Population Change, Pittsburgh: University of Pittsburgh Press, pp. 377-399. 3/ Lincoln Chen, et al., 1974. "A Prospective Study of Birth Interval Dynamics in Rural Bangladesh", Population Studies 28(2), p.. 287. - 239 - A similar calculation for other values of n are given in Table 10.12. The last column gives the effect of amenorrhea. It varies from 0.258 to 0.185 with a mean value of 0.215. Table 10.12: Fertility Effect of Amenorrhea by Parity, Kerala Period in Period of which Amenorrhea Additional of Lactating Births Women minus Occurred that of Non- Parity Additional In(months) Lactating Additional Births Birth Women (months) due to Amenorrhea 1 2 3 4 5 1+ 2.674 114 11 (9)a/ .258 (.211)b/ 2+ 2.006 96 11 (10) .230 (.209) 3+ 1.641 86 11 (11) .210 (.210) 4+ 1.467 77 11 (11) .210 (.210) 5+ 1.212 68 11 (12) .196 (.214) 6+ 1.028 61 11 (13) .185 (.219) a! Alternate assumptions about period of amenorrhea in which younger women are assumed to have shorter periods of breastfeeding. b/ Corresponding to alternate assumption about period of amenorrhea given in Column 4. Column 2: from Table 10.11 Column 3: if Bi = birth interval between ith and (i+l)th births when the ith birth does not die and Ri - parity progression ratio from parity (i) to (i+1) then In = Bn (Rn + RnRn+1 + RnRn + 1Rn + 2 + **) Column 4: difference in period of amenorrhea of lactating and non-lactating women Column 5: = (4) X (2). (3) - 240 - The period of amenorrhea is likely to be shorter for younger women than older women. The main reason is that there is a sequential, downward trend in the length of.lactation by decreasing age. Breastfeeding is longer among older women than among the younger. In Sri Lanka, the average length of breastfeeding- was 12.1 months for first order births, 12.4 for second, 14.3 for third, 17.6 for fourth and 17.7 for fifth order births.4/ It was 12.3 months for women below 25 years, 12.5 months for women 25-34 years, 17.6 months for women 35-44 years, and 19.4 months for older women, with a mean of 17.0 months. Assuming a slight upward trend in the period of amenorrhea with age for Kerala (see figures in parenthesis in Column 4 of Table 10.12), but with the same mean, a new estimate of the fertility effect of amenorrhea is obtained (see figures in parenthesis in Column 5, Table 10.12). The mean is 0.212, more or less the same as the former estimate, but the variation over the parities has completely disappeared. In other words, the effect of amenorrhea on average parity is a constant of 0.21 of a child. Comparison of additional fertility due to amenorrhea (0.21) with total additional fertility shows that the effect of amenorrhea is relatively higher at older parities. The estimates for the first parity cannot be easily explained. It is possible that the period of amenorrhea is much smaller than what we have used (probably only 5 or 6 months) for every young women and that the biological effect has been overestimated. Our conclusions from the Kerala survey are that much of the mortality effect on fertility is biological in nature (Table 10.13). The behavioral response is maximum at parity 2 or 3, and it is negligible at parity 1 and very high parities. 4/ Mary M. Kent, 1981. "Breastfeeding in the Developing World: Current Patterns and Implication for Future Trends", June, Washington, D.C.: The Population Reference Bureau, Tables A-4 and A-5. - 241 - Table 10.13: Biological and Behavioral Effects of Child Death on Fertility by Parity, Kerala Parity n+ Biological Behavioral Total Behavioral as % of total I , 2 0.21 0.13 0.34 38 3 0.21 0.07 0.28 25 4 0.21 0.07 0.28 25 5 0.21 0.07 0.28 25 6 0.21 0.07 0.28 25 Land Reforms In the theoretical framework given in Chapter IX, we have cited land reform as one" of the recent policy interventions which would'have contributed to the fertility decline in Kerala. However, unlike Ratcliffe and others, we did not place much importance on the redistributive aspects of these reforms. For one thing, the land reforms did not result -a a large redistri- bution of land or income. An estimate of the excess land available for redistribution according to the Land Reforms Act (1963) was 114,800 acres (2.5 percent of the total) in the state as a whole; 2,330 acres in Alleppey District (0.7 percent of the totel); 2,180 acres in Ernakulam District (0.5 percent); and 19,310 acres in Palghat District (3.3 percent).5/ With respect to redistribution of income, some data are available on consumption. The Gini coefficient of private household consumption showed a slight increase (increase in inequality) between 1957-64 and 1966-74 (from 0.330 to 0.338), and the coefficient in Kerala is higher than in India (0.310 and 0.294 in India for the two periods respectively). Secondly, by the 5/ Calculated from Tables 12.4 (p.127), 17.3 (p.190), 17.5 (p.194), 17.7 (p.198) in Kerala State, Bureau of Economics and Statistics, 1976. Land Reforms Survey in Kerala 1966-67, 1968 Report, Trivandrum: Government Press. - 242 - Kerala pattern of birth control practice, increase or decrease of land ownership would not have made much difference on the fertility of the women in a household. As shown in Chapter VIII, the sterilization rate is negatively related to ownership of land and, conventional family planning use increases with the area of land owned. Overall, the practice of birth control (sterilization or conventional family planning) is negatively related to land ownership. Thus, redistribution of land from the rich to the poor would not have made much difference in the fertility of the poor; according to our regression (Table 8.22), redistribution could decrease contraceptive practice among the poor. Land reforms could contribute to fertility decline in several other ways, however. One such way is through economic deprivation due to loss of land and income from land. A second way is the increased housing pressure felt especially by the hutment dwellers, who as a result of the land reforms, found it increasingly difficult to find new house sites for their grown-up children. Empirical data to test these propositions are not readily available in the survey. The fertility survey obtained information on land transfers during a 10-year period, 1970-80. Total area gained or lost through sales, land reforms, etc., during this 10-year period were obtained. However, much of the land transfers resulting from land reforms took place before 1970 and, therefore, the survey could not have been able to record them. The reference period could not have been extended very much beyond 1970 because of possible serious recall lapse. However, the survey data are adequate to analyse the overall relationship between land ownership and fertility, and the results of this analysis could be used to arrive at conclusions regarding the impact of land redistribution on fertility trends. These are discussed after the evidences from the limited direct data on land transfers is given. - 243 - Land Transfers due to Land Reforms: The survey showed that - households had gained land during 1970-80 as a result of land transfers and - households had lost land for the same reason. Table 10.14 gives fertility rates of women who belonged to a household which had gains or losses (incomplete). Fertility and Land Ownership: In the above analysis ona reason why land transfers did not show any fertility impact is that transfers due to land reforms occurred outside the refereuce period 1970-80. However, even if we could capture all the land transactions in the survey it is likely that redistribution of land may not show any significant effect on the fertility decline in Kerala. Land ownership, whether it is inherited, bought, or received as a result of land reforms is not associated with the fertility decline in Kerala. This is shown below in two ways: (1) differentials which give the overall net differences and; (2) regression coefficients which give the independent effect of land ownership. A. Differentials Table 10.15 gives fertility rates (TMFR' for the age group 20-44 years) and trends by land holdings. In 1975-80, TMFR' for women in households with less than 10 cents of land was 4.63 while that of women with 100 cents of land (one acre) or more was 4.79. On the whole, the differentials are not large. Earlier (1965-70), however, women in large land-holding households showed considerably lower fertility. Over the period 1965-70 to 1975-80, fertility declined among all land holding households. The extent of the decline has been fairly uniform at all levels of land holdings, except those at the very top who showed a relatively lower rate of decline. - 244 - Open for Table 10.14 on Fertility re: Land Transfe, - 245 - Table 10.15: Fertility Differentials by Land Holdings, Kerala Land TMFR' for the period: Owned 1975-80 1970-75 1965-70 0-10 cents 4.63 5.26 6.74 11-49 " 4.55 5.28 6.47 50-99 " 4.58 4.63 6.45 100 + " 4.79 5.33 6.12 Percent decline in TMFR- 1973-78 1968-73 1968-78 0-10 cents 12 22 31 11-49 14 18 30 50-99 1 28 29 100 + 10 13 22 - 246 - These lifferentials in fertility level and trends are consistent with the patterns shown by age at marriage and birth control practice. The average age at marriage is fairly constant for all classes of land holdings (Table 7.12); the proportion of women (or her husband) sterilized does not vary by land holdings except for the group of women in households with one acre or more (Table 8.18). It is only the use of conventional family planning that showed some increase with.the size of the land holdings. B. Fertility and Land Holdings at the Household Level The relationship between land holdings and fertility rates is given for the individual household level in Table 10.16 where the partial regression of land holdings and fertility related variables are shown. The results of the regression analysis are consistent with the differentials given in Table 10.15. Land has a statistically significant positive association with children ever-born, children surviving, and the desired number of children. All these are life-time cumulative measures. It has a negative association with sterilization status. All the other fertility related variables (period specific) show no significant association with land holdings. These include: births in 1979, 1975-80, 1970-75, and 1965-70; age at marriage; knowledge of family planning methods; and current use of conventional family planning methods. If land holdings and fertility level are not correlated, as the survey data seem to indicate, it is very unlikely that the redistribution of land following the land reforms could have been a factor in the fertility decline in the state. Other Effects of Land Reforms: Although land redistribution may not have been a significant factor in the fertility decline of the majority of the population, especially of the low income groups (it could have been a factor for the rich landlords who 247 - Table 10.16: Partial Regression Coefficient between Land Holdings and Fertility Related Variables, Kerala Dependent Variable Independent Variable: Land Possessed by a Household B F Children ever-born +.000434 7.9 Children surviving +.000415 8.9 Number of births during 1965-75 +.000373 4.3 Number of births during 1970-80 x x Number of births during 1965-70 x x Number of births during 1970-75 x x Number of births during 1975-80 x x Number of births during 1979 x x Open birth interval x x Desired number of children +.000670 36.4 Age at marriage x x Knowledge of F.P. x x Current Use of F.P. x x Sterilization -.000147 12.9 depended on land for most of their income), there is another way in which land reforms could have affected the fertility of the low income groups. This is through housing problems among the hutment dwellers. Land reforms made it difficult for children of hutment dwellers to get a house site of their own and move out of the parents household. As a result, the number of persons per household could have increased and the need for limiting family size could be felt by everyone in the household. Most of the households that gained house sites, or those who are likely to seek new house sites have 10 cents or less of land. Women in such - 248 - households had the highest rate of fertility decline (31 percent for women in all the other groups combined) and their 1975-80 fertility level was close to the average for all women. Such a significant fertility decline among hutment dwellers can be attributed mostly to the official family planning program (see below), but without some basic changes in the desired family size such a decline is rather difficult even with a strong F.P. program. It is possibly here (by changing the desired:family size) that land.reforms played a role in the fertiliy decline through increased housing problems. Table 10.17 gives the fertility decline among the 10 cents households classified further by sterilization status. The sterilized group experienced a fertility decline of 35 percent while the non-sterilized group had a 20 percent fertility decline. As the timing of the decline among the sterilized group indicates, much of this decline--over 50 percent--could be attributed to the mass vasectomy camps. Whatever be the other intermediate variables--increase in age at marriage,6/ use of conventional family planning methods, etc.,--without some significant changes in the attitude toward family size, it is difficult to generate such a large decline. Land reforms, through increased housing problems, could have played a small part in precipitating a change in attitude. The above analysis did not prove or disprove whether land redistribution played any significant role in the recent fertility decline in Kerala. But the circumstantial evidences point very strongly to the conclusion that land redistribution and consequent income redistribution played little or no part 6/ The fertility measure used in this study (TMFR) is not entirely controlled for difference in age at marriage. When TMFR is measured for 1965-70 and 1975-80, the fertility experience used is not for the same group of women. The latter would include a larger proportion of younger women who were married more recently and hence at relatively older ages. Thus the decrease in fertility levels between 1965-70 and 1975-80 as measured here is caused not only by birth control but also by postponement of marriage. - 249 - in the decline. This, however, does not mean that land reforms did not play any part in the fertility decline. Land reforms were partly instrumental in lowering the desired family size of several sections of the Kerala population--among the rich landlords due to economic deprivation and among the poor due to increased housing problem. However, the most important demographic consequences of land reforms came because of a fundamental change in the attitude of the Kerala population toward landed property. Table 10.17: Fertility Trend among Women in Households with 10 Cents of Land or Less by Sterilization Status, Kerala TMFR Age Year Sterilized Non-sterilized Age 20-29 years 1975-80 3.145 3.110 1970-75 3.290 3.025 1965-70 3.820 3.420 Age 30-44 years 1975-80 1.475 1.890 1970-75 2.050 2.130 1965-70 3.315 2.810 Age 20-44 years 1975-80 4.620 5.000 1970-75 5.340 5.155 1965-70 7.135 6.230 Land was the principal means of income of a household and its most important economic asset. A person's economic worth was measured in terms of the land he or she owns. Land reforms changed all that. Land, particularly agricultural land, lost its prime position in the economics of the Kerala society and in its place came personal qualifications--education, health, etc. This basic change in the value system has had a much deeper and long-lasting effect on the demand for fertility control than the mere redistributions (if any) which were brought about by the land reforms. - 250 - Agrarian Reforms Apart from land reforms, there were other (agrarian) reforms which could have also been instrumental in Kerala's fertility decline. The principal among them was wage reform which brought about a sharp increase in wages, especially among females. Wage increase and strict enforcement of child labor laws had a disastrous effect on children's employment. As a consequence of wage increase, employment in the state lagged very much behind labor force growth. In the tight employment market where minimum wage legislations were strictly enforced, children were the ones who suffered most. When the value of children as an employable economic asset declined, so did fertility. Wage increases were relatively larger among females. As a consequence of this and also because of a general improvement in working conditions, wage labor became increasingly attractive to females and the economic cost of not working because of continuous childbearing became more noticeable. Empirical data to test whether the increasing cost of bringing up children has been a factor in Kerala"s fertility decline are derived from the question on educational aspirations for children. All women were asked: In the socio-economic context of Kerala, how much education should boys and girls receive? The average desired years of schooling is given in Table 10.18 classified further by age and educational attainment of mother (for girls) and father (for boys). These averages show that the educational aspirations of Kerala women has been increasing; the younger women have a higher aspiration for their children's education than the older ones even after controlling for their own educational attainement. -251- Table 10.18: Expectations about Education of Boys and Girls by Age and Education of Father and Mother, Kerala Question 636 Average Desired Years of Schooling for Boys Age of Mother Educational Attainment of Father ______ No Education 1-4 Years 5-'9 Years 10+ Years 15-19 2 0-24 25-29 3 0-34 3 5-39 40-44 Total Question 640 Average Desired Years of Schooling for Girls Age of Mother Educational Attainment of Mother _____________ No Education 1-4 Years 5-9 Years 10+ Years 15-19 2 0-24 2 5-29 30-34 3 5-39 40-44 45-49 Total - 252 - This increase in aspirations may reflect an increase in the cost of educating children. This conclusion is supported by an analyses of reasons given by the women for believing that their children may not achieve their educational aspirations for them. Table 10.19 indicates that lack of money is an important reason cited. How have aspirations for children affected fertility-related variables? This question is examined with the help of a multiple- regression (Table 10.20). We have obtained regressions on the following (dependent) variables. Dependent Variables Y = Age at marriage Y2= Desired family size Y3= Use of conventional family planning methods Y4= Births in 1975-80 Independent Variables X1 = Desired years of schooling X2 = Reason for non-achievement of desired (1 = financial reasons; 0 = all other) X3 Education of wife X4 = Age of woman 5 = Household expenditures X6 =Land owned X7= Caste Benefits from Children: In the conceptual framework, it was postulated that Kerala women have begun to control their fertility because the cost of bringing up children has increased and the benefit from them has decreased. To test the latter part of this hypothesis, women were asked: 1. Do your children between 8 to 12 years presently help on the farm, in the business or around the house? 2. Do any of the unmarried children over 12 years presently work for pay? - 253 - Table 10.19: Distribution of Women by Reasons for Children's Possible Non-Achievement of Desired Educational Attainment, Kerala For Boys For Girls Question 639 Question 643 Reasons As % of As % of No. of Applicable No. of Applicable Women Women Women Women 1. Lack of money 2. Needed at home 3. Not good at job 4. Got married 5. Other 6. Not applicable - 254.- Open for Table 10.ZO:Regression Results Re: FertiLity and Aspirations - 255 - 3. Do you expect that your children later on are likely to help the family for a while by working for pay and contributing to your household? 4. What means of financial support do you think you might have when you and your husband are old and your husband can no longer work? The following Tables (10.21-24) indicate that a high proportion of Kerala mothers (-%) receive help from children with work around the house, but relatively fewer of the younger mothers think that their children help them in this work. Very few of the unmarried children (12+ years) work for wages and contribute to the family income. Most women expect help from their children, especially at older ages. Fertility and Benefits from Children: Whether expectation about benefits from children affects fertility- related variables is tested by a regression model (Table 10.25). The dependent variables (fertility related variables) are as defined earlier. The independent variables are: Regression I, X1 = 0 if 2 in Q 647 1 if 1 in Q 647 but confined to women with (1) in 646. Independent variables X3 to X7 given previously. Regression II, X1 - 0 if 3 in Q653 = 1 if 1 or 2 in Q 653 confined to women who answered (1) in 650. Independent variables X3 to X7 given previously. Another aspect of expected benefit from children is old age security. It is argued that in developing countries, children are the main source of support at old ages and this is one reason why fertility remains high. Do Kerala women expect old age support from their children? Is this expectation changing? Does the changing expectations affect the fertility trend? - 256 - Table 10.21: Children's Help- Around tIhe Hous.e KLrala Question 642 % of Women Who Receive Help from Age Help Don't Help Total Children 15-19 20-24- 25-29 30-34 35-39 40-44 45-49 Total Table 10.22: Expectations of Future Economic Help from Children, Kerala Question 654 Don't No Age of Expect Expect Living Mother Help Help Uncertain Children Total 15-19 20-24 25-29 30-34 35-39 40-44 45-49 Total - 257.- . Table 10.23: Work Status of Unmarried Children 12 Years or Over by Age of Mother, Kerala Question 652 % of Women Whose Work for Don't Work Children Work Age Pay for Pay Total for Pay 15-19 20-24 25-29 30-34 35-39 40-44 45-49 Total Table 10.24: Expectations of Old Age Financial Support from Children, Kerala Question 655 Other Means of Support Expect Income from Age of Help from Family Farm Mother Children Savings Business Pension Others Total 15-19 20-24 25-29 30-34 35-39 40-44 45-49 Total - 258 - Opren for Thble 10.2'5: Results of Regression on Fertility vs. Children's Benefits. - 259 - About percent of the women answered that they expect financial support from their children at old age. However, fewer of the younger women expected support from their children. Thus, there is a change in the expectation, but has this change anything to do with the recent fertility decline? The following regression provides some clues to the answer (Table 10.26). Dependent variables as in previous list: Independent variables Regression I, X1 = 0 if 2 in 654 1 if 1 or 3 in 654. Exclude women who answered 4 in 654. Other Independent variables X3 to X7 Regression II, X1 = 0 if 2,3,4,5 in 655 = 1 if 1 in 655 Regression III, X1 - 0 if 2 in 657 - if 1 or 3 in 657 Official Family Planning Program Much of the fertility decline among married women occurs through family planning, conscious or unconsious. All family planning need not be the result of the effort of the official family planning program, however. In fact, in many populations much family planning practice has nothing to do with official policies or programs; they don't exist. According to our hypothesis, the official program in Kerala has had a major role to play in the fertility decline. Without the official program, the degree of the fertility decline would have been smaller, and the socio-economic background of the population which contributed greatly to the decline would have been different. What empirical evidence can be given to support our hypothesis? Analysts of the Kerala situation have different views on this matter. Gopinathan Nair (1974) concluded that fertility declines in Kerala began in - 260 - Open for Table 10.26: Regression Results on Fertility and Old Age Benefits - 261 - the early 1960's prior to the intensification of the family planning program. He thought that education and mortality declines had played a major part in the initial fertility decline. T.N. Krishnan (1976) attributed the decline in fertility in Kerala to changes in nuptiality rates brought about by sharply increased female literacy. Kurup and Cecil (1976) attributed the decline in the birth rate mainly to increased acceptance of the family planning program. As mentioned earlier, Ratcliffe (1978) did not think that family planning played any major role in the fertility decline. We agree with Gopinathan Nair that some declines in the birth rate and TFR (not TMFR) took place before the intensification of the official F.P. program. The cause of this decline was mainly a decrease in the nuptiality rate, not declines in the marital fertility rates (which in some age groups could have experienced some increase in the 1950's). A significant decline in the marital fertility rates began only in the 1960's, coinciding with the implementation of land reform and the intensification of the F.P. program. Methodologically, it is rather difficult to show what would have happened to Kerala's fertility trend without the official F.P. program. We have adopted three approaches to answer this question: - Comparison of the fertility level and trends with official family planning outputs; - Comparison of the Kerala and Sri Lanka experiences; and - Analysis of the characteristics of family planning users among whom fertility has declined. Before we compare the fertility decline with the official family planning output, it is worth recalling one interesting aspect of the relationship between family planning practice and fertility. The relevant partial regressions are reproduced on the following page. - 262- Partial Regression. B F 1. Children ever-born, 1980 +0.534 57.5 2. Children born, 1965-70 +0.224 15.8 3. Children born, 1970-75 +0.200* 0.4 4. Children born, 1975-80 -0.105 10.3 5. Children born, 1979 -0.058 16.0 *Statistically insignificant. These data show that, on the whole, family planning practice and children ever-bor= are positively related. The relationship goes from fertility to family planning and not the other way around. High fertility women tend to adopt birth control methods. However, when the relationship is examined separately for births which took place in specific periods, the direction of the association changes from a significant positive value for births in 1965-70, to an insignificant relationship in 1970-75, and again to a significant but negative relationship for births in 1975-80 and 1979. Thus, in recent years, the direction of the relationship seems to have been reversed, with birth control practice affecting fertility rather than the other way around. This change in the direction of the relationship between fertility and family planning practice is one bit of evidence that the stronger, more recent official family planning efforts have made some dent on fertility in Kerala. Fertility Decline and Family Planning Output: Table 10.27 shows the extent of the fertility decline due to acceptance of family planning methods (including sterilization) from the official program. The rank correlation coefficient between the official F.P. output in a district in a given period and the fertility level of that district in A that period is -0.7, which is statistically significant for the nine observations. - 263 - Table 10.27: Measures of Fertility Decline and Births Averted by District and Time Period, 1965-80 Palghat Ernakulam Alleppey Period TMFR' for Ages 20-44 years 1965-70 6.24 6.63 6.52 1970-75 5.48 5.38 4.82 1975-80 4.96 4.65 4.59 Birth Averted Due to Official F.P. 1965-70 1970-75 1975-80 Usirg the family planning acceptors by methods, an estimate of the births averted by the official family planning program were obtained and are compared with the actual decline in number of births. It appears that birth averted were greater than the actual decline in each district. Comparison of Kerala with Sri Lanka: Sri Lai4ka is similar to Kerala in its socio-economic development. The relevant indicators are very close, but those of Kerala are systematically lower. The population density is much higher in Kerala, but it has a safety valve which Sri Lanka does not have, Kerala is part of India, and Kerala people can freely migrate throughout India in search of economic opportunities. On the other hand, Sri Lanka is an island and any movement out of it i- restricted by the usual formalities of work permits, visas, exit permits, etc. On balance, we would expect that if there were no government - 264 - influence, these twor populations would have more or less the same family planning practice rates. If there were differences in their practice rates, especially if they are lower in Sri Lanka, these could be attributed to differences in the official family plannix' programs. Figures 10.3 and 1.0.4 give a comparison of sterilization and conventional family planning; prevalence rates in Kerala and Sri Lanka. The data for Kerala are from this- study and therefore refer to 1980. The data for, Sri Lanka are taken from the World, Fertility Survey for Sri Lanka and therefore refer to 1975. Although more recent data are available for Sri Lanka we have taken the 1975 rates as they are not affected by the recent, more active nature of the official family planning program. The rates in Figures 10.3 and 10.4 indicate that the two populations have very close prevalence rates for conventional family planning methods, but vastly different rates for sterilization. If Kerala had no strong official F.P. program offering the sterilization alternative, we expect that its sterilization rate would not have been very much different from that,in Sri Lanka. We attribute the difference in sterilization rates to the official F.P. program. What is the consequence of this difference in sterilization rates on fertility levels in Kerala during 1975-80? The expected consequence is a decline of 0.92 children per woman (TMFR) (Table 10.28). This was obtained by estimating the average number of children that would have been born to the women who were sterilized in Kerala in excess of the number in Sri Lanka and applying the fertility schedule of the sterilized to all married women (See Table 5.25, Chapter V) - The actual decline was 1.77, CSee 7/ Which set of fertility schedules will best measure the rates of the sterilized group is debatable. We have used the rates for the intermediate time period, a sort of average for the entire period under study, 1965-80. - 265 - Figure 10.3: Sterilization and Family Planning Practice by Age, Kerala and Sri Lanka Kerala - t-. -. - rLanka % Use - Steril±zation 10 50 ..40 3 Conventional FP - -20 - 35 20 25 30 35 40 45 50 ABe BEST COPY AVAILPABLE ' - 266 - Figure 10.4: Steril. ration and Family Planning. Practice, by Education, Kerala and Sri Lanka % Use Kerala 40 - -----kConveantl,nal 20-- -- 2 44 6 8 - 0 12 14 16 Level of Education BEST Copy .4ALA1 - 20GL- - 267 - Table 10.28: Estimate of Excess Sterilization Rate in Kerala Kerala Marital Sterilization Rate (%) Fertility Rate, (5) x (4) Age Kerala Sri Lanka Difference Sterilized, -70-'75 100 1 2 3 4 5 6 15-19 1.1 0.5 0.6 200 1.2 20-24 7.1 1.9 5.2 419 21.8 25-29 26.1 8.1 18.0 281 50.6 30-34 42.0 12.9 29.1 207 60.2 35--39 40.5 13.9 26.6 112 29.8 40-44 33.0 10.3 22.7 71 16.1 45-49 21.5 5.5 16.0 24 3.8 TOTAL 27.9 8.9 19.0 0.9175 Notes: 0.9175 = (1.2 + 21.8 + ...) x 5/1,000. Sterilization rates for Sri Lanka from Sri Lanka, Department of Census and Statistics, 1978. World Fertility Survey, Sri Lanka, 1975, First Report, Table 4.3.1, p. 485. Table 5.9, Chapter V). Thus about 52 percent of the marital fertility decline could be directly attributable to the official F.P. program and 48 percent of the decline would have taken place even if the government had not taken any initiative in birth control. Data from our survey indicate that about 30 percent of the fertility decline during 1968-78 was due to a decrease in the nuptiality rate (see Table 7. 16, Chapter VII). This has nothing to do with the government's family planning program; it is a response to socio-economic changes. Combining these two, we may conclude that about 36 percent of the decline in the TFR (70% x 52%) was due to the official F.P. program and the balance of - 268 - 64 percent (30% + [48%7 x 70T]) would have taken place because of the socio-economic changes, irrespective of the official program. Characteristics of Family Planning Acceptors: A third piece of evidence supporting the contribution of the official family planning program to Kerala's fertility decline is the socio-economic characteristics of the family planning and sterilization acceptors. In a .population undergoing demographic transition without government input through family planning services (particularly financial incentives), we would expect birth control practice.to be, on the whole, positively related to economic and social status. That is, the better educated, the richer, those with non-agricultural occupations, etc., would tend to control their births more than the illiterate, tl,e poorer, the farmers, etc. In Kerala, we did not see this pattern. Our data indicate that the prevalence rate of sterilization was highest among those with 1-4 years of schooling, not among those with more than 10 years of schooling (Figure 8.1); among Ezawa and Scheduled Caste communities, not among the Nairs (Table 8.21); and among women from households with a per capita monthly expenditure of Rs 50-59, not among those with monthly expenditures above Rs 120 (Table 8.17). We attribute this higher sterilization rate among the lower strata of society to the official F.P. program, especially with its economic incentives.8/ Thus, the Kerala Fertility Survey provided enough evidence to conclude the presence of an official dent on the marital fertility rate of Kerala women, especially women in the middle and lower stratas. In the absence of a public program, the fertility decline in Kerala would have been only two-thirds of the actual decline. 8/ In a study made by the Demographic Research Centre, Trivandrum, in Palghat District, it was reported that "of the total acceptors 51% preferred mass vasectomy camp for greater monetary renumeration, 30% for the services of expert doctors,..." (Kerala State, Bureau of Economics and Statistics and Demographic Research Centre, 1977, p.15). - 269 - Part III: INTERPRETATIONS AND CONCLUSIONS CHAPTER XI LESSONS FROM THE KERALA STUDY Kerala has experienced one of the sharpest fertility declines in the South Asia region. With levels of income and nutrition among the lowest in the world, Kerala has achieved fertility declines comparable to those in the most successful middle-income countries of the world. How did they do this? Has Kerala-s experience in fertility decline any relevance to population programs in other states of India? Since the fertility decline in Kerala was as much due to historical factors as to recent policy interventions, its experience is certainly not entirely replicable in other parts of India; yet, there are a few lessons that can be learned from the Kerala experience which will be useful in reorganizing the population programs elsewhere in India. An increase in age at marriage is a familiar prescription for fertility transition not only for India but also for most other developing countries. It is so basic and fundamental that it needs repetition over and over again. All populations which have undergone fertility transitions started with an increase in the age at marriage. Kerala is no exception. About 30 percent of the fertility decline in Kerala during 1968-78 could be attributed to fewer married women at younger ages. During 1975-79 the average age at marriage was 19.5, 20.9, and 21.7 for the three survey districts in Kerala. The difference between Kerala and all the other states of India is very large indeed. For example, if the proportions married in India (in 1971) were the same as those in Kerala, the TFR in India would be lower by 20 percent. This - 270 - is the direct consequence. Then, of course, there would be some indirect consequences through increased adoption of family planning methods, etc. Thus, the potential for fertility reduction in India through increased age at marriage is very significant, and is a means of fert±ity-reduction untapp-ed so far- in rural India. The-" rrrala study did not discover any easy method to achieviE.an increasie-in age at marriage. Among the manipulatable variables, female education showed the strangest positive association with age at marriage.!/ There seem not many other alternatives to ensure a high age at marriage for Indian women. The Kerala study indicated that about 60 percent of the fertility decline in the state during the past decade could be attributed to socio-economic factors and the balance of 40 percent to the official family planning program. Even if Kerala did.,not have an official family planning program, the fertility rate in the state would have declined by about 60 percent. The principal socio-economic factors are, of course, reduced infant and child mortality, and basic education, particularly female education. Kerala differs from other states of India in many respects, but notably in health conditions and education: Kerala India Percent literate, 1981 69.2 36.2 Female literacy, 1981 64.5 24.9 Infant mortality rate, 1977 50.0 129.0 Child death rate, 1977 5.0 22.0 It would, therefore, appear that in order to encourage an increase in age at 'marriage and small family norms in the other states of India, measures should first be taken to reduce infant and child mortality and increase female 1/ For Kerala, the partial regression was +1.07 months; an increase of one year in female education raises the age at marriage by 1.07 months. - 271 - education. Like increasing the age at marriage, these are not new policy measures, but often repeated ones and supported by most studies on the determinants of fertility. The Kerala experience adds one more to the list, but in addition it shows an alternate means to achieving these goals especially by the low-income families. One way to improve the health and educational level of low-income families is to provide them the means to earn and thus be in a position to buy better health and education. Kerala shows an alternate way--reduce the cost of these services and make them accessible to low-income families through a cheap transportation network. These are direct means of intervention to improve the level of education and health of the low-income families. And the experience of Kerala shows that it will work. Agrarian reforms including fixation of minimum wages, better working conditions, land ceilings, etc., seemed to have helped to increase the desire for small families and the demand for birth control in Kerala. The high minimum wages and better working conditions operated in two ways. High wages reduced the availability of jobs, particularly for children who, as a result, prolonged their studies and became more of an economic burden (food, clothes, school expenses) than a source of income (wages from child labor, and unpaid family work). High wages also increased the opportunity cost of the mother. Whether the Kerala experience of agrarian and other redistribution policies can be followed in other states in India depends on the political will of the people of these states to enforce a shift in political power from the rich to the poor without which all such reforms will remain paper reforms wi,th little effect on the life style of the people. In Kerala the official family planning program has played a significant role in the recent fertility reduction in the state. As mentioned above, 40 - 272 - percent of the fertility reduction could be attributed to the official program. Among low income groups the contribution of the official family planning program may be even higher. If the official program has succeeded in reducing fertility very sigA±flcantly in Kerala, why is that it has not been so successful in many other states? What is different about Kerala is 'not higher family planning inputs, at least as measured by budgetary allotments. The success in Kerala is more likely due to a more efficient delivery of services and a higher spin-off effect; that is, a higher interaction between family planning services and socio-economic conditions. The same amount of family planning services was more effective in Kerala than in other states because of the different socio-economic conditions in the state. Yet, Kerala women do not have a very strong view regarding the number of children they would like to have; this number is determined mostly by the, number of children they already have, which explains 90 percent of the "explained" variance of the desired number of children. Even when women find that they have reached or achieved the number of children they would like to have (as reported by them), not all of them (at least 20 percent of ever-married women) take steps to prevent additional births. It is, therefore, neither a strong manifest demand for family planning services, nor an expressed desire for fewer children, but rather an effective delivery of family services aided by financial incentives.in conjunction with a weak dormant demand which has made the family planning acceptance as high as it is in Kerala. A basic question which we cannot adequately answer is: would the family planning program in Kerala have worked if the population were like the rest - 273 - of the Indian population with respect to education, mortality, etc.? Two results from the Kerala study are relevant to answer this question. First, within Kerala, high income and economic status is neither a necessary nor a sufficient condition for acceptance of birth control methods. A strong family planning program like Kerala's (with economic incentives and emphasis on sterilization) can alter the normal relationship between fertility decline and socio-economic status. In fact, economic incentives are more valued by the poor, and hence, the family planning acceptance rate (from the official program) and fertility decline could be higher among them. In Kerala, they are indeed higher. Second, the inter-district comparison shows a similar pattern. Within Kerala, Alleppey and Ernakulam districts are more advanced than Palghat and the family planning acceptance rates are much lower in the latter district. In general there is a positive association between development and family planning acceptance. But the experience in other districts in India shows that there is no necessary relationship. The positive association can be broken with suitable family planning delivery. In conclusion, although the Kerala experience in fertility decline cannot be replicated in the other states of India because of the critical part which the historical developments played in its recent fertility decline, there are a few lessons which can be learned from its experience and which will be useful in reorganizing the Indian family planning program and working out a long-term strategy for ensuring a sustained fertility reduction in the other states of India. First, a government sponsored family planning program can make a substantial contribution (about 40 percent in Kerala during 1968-78) to the fertility reduction in a population, not only among the higher socio-economic strata, but more so among the lower strata. Within limits, the normal - 274 - positive association between socio-economic status and family planning acceptance can be broken by a good family planning delivery system aided by economic incentives, and the acceptance rate of a public family planning program among the lower socio-economic strata can be even higher than among the higher strata. Without a public program such a pattern in unlikely to develop. Second, the family planning program has been more effective in Kerala than in the other states of India mainly because of the higher spin-off effect between the program and socio-economic development. One lesson from this result is that the sequence in which the determinants of fertility are applied to a population is as important as the determinants themselves. In Kerala the determinants came in the right order--a reduction in infant mortality followed by or along with an increase in female education, followed by land reforms and other redistributive policies, and finally the official family planning program. The impact of Kerala's family planning program would have been much smaller and less lasting had the program been introduced before a substantial reduction in the infant mortality rate. In states where the IMR is still above 100 and where female literacy is still below 25 percent, the socio-economic determinants of fertility should have precedence over the family planning program. Third, in Kerala changes in norms about family size came through a reduction in IMR and the cost-benefit ratio of surviving children. These changes occurred over a long period of time but were accelerated Inadvertently through land reforms, minimum wages, etc. Kerala's experience thus shows that socio-economic policies can be used successfully to change norms about family size. - 275 - Fourth, an increase in age at marriage played a major role in Kerala's fertility decline. The potential for fertility reduction in the other states of India through increased age at marriage is very large indeed. This is a means of fertility reduction very much untapped in most north Indian states. - 276 - ANNEX I: STATISTICAL TABLES - 277 - List of Annex I Tables Page Table 1. Regression of Children Ever-Born on Socio-Economic Variables............................... Table 2. Regression of Surviving Children on Socio-Economic Variables............................... Table 3. Regression of Number of Births in 1965-75 on Socio-Economic Variables................................ Table 4. Regression of Number of Births in 1970-80 on Socio-Economic Variables............................... Table 5. Regression of Number of Births 1965-70 on Socio-Economic Variables............................... Table 6. Regression of Number of Births 1970-75 on Socio-Economic Variables............................... Table 7. Regression of Number of Births 1975-80 on Socio-Economic Variables............................... Table 8. Regression of Pregnancy Status on Socio-Economic Variables ............................... Table 9. Regression of Open Birth Interval on Socio-Economic Variables............................... Table 10. Regression of Desired Number of Children on Selected Demographic and Socio-Economic Variables.............................................. Table 10A. Regression of Desired Number of Children on Selected Demographic and Socio-Economic Variables.............................................. Table 11. Regression of Excess Fertility Women by Selected Demographic and Socio-Economic Variables............................................. Table 11A. Regression of Excess Fertility Among Excess Fertility Women by Selected Demographic and Socio-Economic Variables.............................. Table 12. Regression of Excess Fertility Among Excess and Par Fertility Women by Selected Demographic and Socio-Economic Variables........................... Table 12A. Regression of Excess Fertility Among Excess and Par Fertility Women by Selected Demographic and Socio-Economic Variables........................... - 278 - List of Annex I Tables (cont.) Page Table 13. Regression of Excess Net Fertility Among Excess Net Fertility and Par Fertility Women by Selected Demographic and Socio-Economic Variables............................... Table 13A. Regression of Excess Fertility Among Excess Net Fertility and Par Fertility Women by Selected Demographic and Socio-Economic Variables.............................................. Table 14. Regression of Conventional Family Planning Use Among Non-Sterilized Excess and Par Fertility Women, on Selected Demographic and Socio-Economic Variables........................... Table 15. Regression of Conventional Family Planning Use Among Non-Sterilized Excess and Par Net Fertility Women on Selected Socio-Economic Variables........................................... Table 16. Regression of Sterilization Among Excess and Par Fertility Women'on Selected Demographic and Socio-Economic Variables........................... Table 17. Regression of Sterilization Among Excess and Par Net Fertility Women on Selected Demographic and Socio-Economic Variables............... Table 18. Regression of Conventional Family Planning or Sterilization Among Excess or Par Fertility Women on Selected Demographic and Socio-Economic Variables........................... Table 19. Regression of Conventional Family Planning or Sterilization Among Excess or Par Net Fertility Women on Selected Demographic and Socio-Economic Variables........................... Table 20. Regression of Early or Late Marriage on Selected Socio-Economic Variables...................... Table 21. Regression of Age at Marriage on Selected Socio-Economic Variables............................... Table 22. Regression of Age at Marriage on Selected Socio-Economic Variables (All women and with ym)******************************************.**** - 279 - List of Annex I Tables (cont.) Page Table 23. Regression of Age at Marriage on Selected Socio-Economic Variables (Women 25+ years and y m)............................................... Table 24. Regression of Age at Marriage on Selected Socio-Economic Variables (Women 25+ years and ym)................................................ Table 25. Regression of Age at Marriage on Selected Socio-Economic Variables (Age at Marriage less than 25 years).................................... Table 26. Regression of Age at Marriage on Selected Socio-Economic Variables (Marriages in the 1970's at Ages below 25 years)..................... Table 27. Regression of Knowledge of Family Planning Methods on Selected Socio-Economic Variables.............................................. Table 28. Regression of Number of Birth Control Methods Known on Selected. Socio-Economic Variables............................................. Table 29. Regression of Current Use of Conventional Family Planning Methods Among Non- Sterilized Currently Married Women on Selected Socio-Economic Variables...................... Table 30. Regression of Sterilization Status on Selected Socio-Economic Variables...................... Table 31. Regression of Use of Conventional Family Planning or Sterilization on Selected Socio-Economic Variables............................... 二 立 >л . �в . А!}пех ТаЬ1е I 3: -�� _�_ ., � . -- (`"--`-~- wtauaeiaи о► l�uemie ot цитые кн tявs-тs аи восlассоиокгс чвякмг.вr . � „• -. ••ве♦деесвв•веегвг•ав•а МУЬ II►L1 п1 Уп[ YlsDи •ве•веявввв•в YARtAlбi�tвT 0 atreнo[иг vАпкыLЕ,. ia4 нииlЕп оi.fкпти• ки к+�{ то �+1! пtfiпc{{tои �к11 t rыкА{l[iiI сиТ1п[о Ом гтL► ниипtп а.� vtI А{t • v3i АОЕ дТ MApIAOL � ►а iАицт ►ЕАии1и` аТАТио iЕпо ап онЕ с ыт= млкл ап lтпкди сиRк{т1Ап Qп о1и[пнк{t Yr1 ам !(х►[н{t1 � ,, е- vь `�но оИиtо У{ rпв о► ltиао� ' vв пбаг илttпlAt, • Yt •UUпCi 0! МАТLп oU►►LY . У3 TOILlТ /aCILITт MN6+1вLS Ч' {,вд{11 ANA{,Уlк! Ов улпквМСS OF {ии 0► lдUАиf• М[АИ lаUАп[ ► п{оиАпЕ 1.ке2�ь псопсоllои io. fl+,ы+ss �е,ее!►i iв,asjto лoJUilro п{аwп[ о•Iмы пикоиА� tt+s. :к3l.slкlо I.тt�s+ {Тли0ыо Еплоп 1=аа!{1 овв.•вв.w•wоап.в тАп►ААIL{ УИ ТМВ dYU4TIDИ •в-..в.ьrв•ввв.в•а .+■..ь..ь...ь улпкАl�Е{ NpT км тЧЕ Е9иАТIдн .••.-•.+...■.. � tlлRjцlR i пlтл 1Т0 tппОп i I УАпIл{L1 1ЕТА кМ ►ARTIAL �а�[пАЧС[ / vs� ..tolтlвsе.,,! .а.вl�s! !.!!вт! 2�а,+а+ i�! ,l�oltoыE.os o.tгse+ {.ооо!! '1о fг1 � O,1+tfSj2 /�Вьдi� овооb0? f�ы! cAiT: .l.i2S{вяi ь{.оа+1{ l.о+!!{ ��аlt rв1 ,тоtнs{г.лs l.э+21в о.ооо и коаев rь ,кlзкыlr,+s l�о�•кя о.{!ое{ в�3.+ . r! .,в+sыоь�.,к •в.Iаак{ l.tьввt +.о�) Ув •вtS12+Sf 1.141l8 �,lавtl 7евв� � Yi o.ISeos12 олевк2е о.l++ и 2.sвк У1 .les+2{ев.,,я l,о+о+: о.{t+lв •,222 tСон{тАит1 а,lsaslе _ _ � . а[rЕир[иТ rдпдцltв, а!i нии{[п о► {епгиi аи lва� То t+тf . ---- . N оо lUиМыУ TлfL[ �, � rикл{�[ ии4тд►�1 е а вдиАп[ п{о сидиаЕ Ч и►Е1 п • rtTA ! tl7i лаЕ l,:+N• !,ll2{3 lвl:iак ьl,1+в+• •s1{Sт1{у{.0! ьов4�!!У ru ло[ Ат нАпtыnt е.�llвi l.Ial�a о.lо{в2 •!.оlиа .�{яlоа•е-оа oi »�и г► rлиl�е ►�лNиаиl етлти{ аlпо о� оиЕ о.1о2�� t.1ы �2 о.ооо•т •.оавеt e.I+1ssZ� о.о•ыs слете и►ап ап {чикли сипкlт'Аи ои Dтнlпиl{Е о.;аи1 о,а1и1 1.оlяку •1.t2s п .o.2alll+ aotl�+oo ri1 иЧ Ев'iыЕ{ ,,lqatь о.11к27 оsовlоо •1.Од2/1 ,iоттаев[•в3 О.оьilв ra �Аи0 ониtо о.lяаьв o,I11tь l,оо и+ а.оlс_: .1t33atlЕ.о1 _ o�of{I1 re ты ог ц и�о� i.вг+12 о.кны о.оо+�о •е,оl�I� •.в+3аrа{[•Ч •oll02!• ++ поо► илтtмкль o.11Ilo о,оьУ+t o,oot2t •.ояsяt a.ISt2+se o.o;�ss +а - {Оипсs ог ■л�сп 1и►г�а о.�ык2 о.�{м{ о.ооа+� l.оыlт о.кsовкlа Q:o��20 db - ТиIЕЕт ►лсть+Тт о.!квlк o.tыra a.oos+v {.I иы ,{•а+а2 и-!I о!оlор2 tС9и{ТвиТа ь.оs2 и + �-------�--'- _ -- - -- вв� сОпЯ[lвGlАи б0trldisRмtп w А Улдуl Oi в+,ilf!{ кв •RдиТiФ � 1Р-$ сОtт►кскЕит ;лыюоt {` tОиРУУдо. � . �i •af vis е3: f► САпt vи va w, � w r2 r3 ии � � � •as к.еоооа .e,s•!+е •l,аиtь l.оамlт •!, ussl ra.or:11 •.ьlьав -о.аl2ls � l.elatl/ в.о++яз а.яа„е .о.вкrк• r!к .a/S�e+a l,Чоо! •,2sвl1 •{.о{Ча 1,о+1►ь ♦.o+IS• •.о:м3 -o.oleel -о.ооваl -о.оlыl -о.о+Iво O.i+яf1 � eL� •в.оо+tь o.2sel1 1.!!оо! l.оуl�о е.(��г! {.аlы+ •!,lек+! о.2о31+ •{,o1�lt o.oo2as •о.о�•то rd.it2si ►► о.lави .а,{{оев •,!!{вв 1.!lоов •,i1ыs l.я�в+s -а,вкв2к о.2ывs •l.оак++ о,о21вI •о.1ьы 2 •1.t2t+1 � сА{Tt .erltS31 n.o�laь a.Il��! o.I1a+l 1.ооiв• {.1тl1t о.tоь21 о.вв11+ {.Iьso2 •а.lsав+ •о.tакьl •3.031+о rвт .o.or2l1 •.иI76 о,е{ы! {.tf++f i,1TSf1 i.laooi •,2e�t! о.ввli+ !,2{авТ l.i1lа2 •о.вsоы o.ol►02 � re l.ваьl. л.ке!{ -в,ооки н.окlая t.IOы к o,2вltr t.oooo{ l.2ятк1 o.INo1 .{.о•1{1 •!.►sы2 а.еа•кь � уо •o.eaals •е,lяlва l,tоат+ {,tыas о.ьоlк+ о.в•tso •.2I1TS I.oaao{ {,t+ts1 •о.оо22в •о.вrsтд •0.1ы ot �, то •.lISI1 •nb!!Nt •o.os2et -{.оыи o.tвslt f.2oa41 l.IOS{1 {.I+t1t t.{!о{{ l.о2 ы з •о.2+2{1 �.+I+ao � �• Уе . о.{в+1) .n.l/e+1 а,Ч2аt 1.ottвl -o.tlao• {.отsы -а,анlt •l.+и 2л о,ии: I.lоеао •в.I•2as - .oI ы! � У3 l.tова� .а�оltв+ .о,н+lt• •o.Iaьrt +1.2l1ьI •l.вsои •o.lsa2a •{.ввslв •о.2+2ц -o.Il2ьs J.ооаов •i.8a si дп •е.аа1'1 e.�+I,+ .�.�221ь .{,1:к+{ .e.o1l+{ l.olaol о.оаllв •l.Iы от {.1lны •о.оlьа+ •о.о{в1в }.ооlоо �'".а,,� � - � i ' � � R � � г � i 〕．&:!&”甲 ,.，巒甲l•．，•豐俗．,.！啼”•一，,,,，盧．盧I•l&．•〕，, .r，二，&l…響，•二l&．響。一，1．面二藝俗’面緬，? /t.,&&t，•〞．．•t.&.．•”·，..t&,..& 盧奮州甲•，物tl甲留I留▼j細，，盒呈t奮1•曾l唱v為j•啊驢•奮•蘊t，蘊】•個多1審，奮卹響晨 ‘•－一‘&,-－一”•開言自，&.,，開魚”1 tn”•一1響，1.，禹－．一•·•一‘．•－一•一•一•·••一”一•••••哺自蘊，油，,“離開他•，1.,,.,.&”●••一•一”•一•，”一， 跚藝t‘擊甚•．0．•】口．，•細，二g 參，彎！•甲••f！勿，,，•簪！甲寥簪l奮，留nOI留奮，9.,”•舌•言．,n6e響•婦盒擊nf。， ,,&,,&&，各？二，&,l,t,.．…l’個唱”01,,j。，j.,l,,l:.,.,n。勵， 彥當”,n••”，奮“•言開，ne總扣。”•．•變，&vl,，禹州響Ie昌1，。，.t口．t&．細】】dl盒1&, 織ljdn•．萬亂，一n寫，．聞留頗倘 Ivl挪萬么v”卹O響．& )m,&）個州t結鳥•良 嚇革”一n•。甲1，為 •，•“露d，寥開闢t．為 ，•l一甘寫”盧0珍n開甲寫tt奮衫開，開，I柚亂••0響•，闖言•t，】 奮“0開00留蘿實•n為，I•9&I&&,ljl、讓”響J jJ 奮，，言婦，&l，革舟，ft禹 •華，, 才弓，It禹 ．•二〝廳I常1 1 jvJ。抑，I”臘t禹”&l”〕一“noJ習蘊“0 01•1煙。萬l口，】1.,1,，戶 ．盧．,0總t,．侖。奮彎”屆。奮審州．〕個”間O遞藝，•”。奮1馴，留騙〝為a”論．“論J多口 11個11闢nl彎t可曾，抑寥 I••l,,1,，奮，，禹，,,，•．•二•二•”nl•，，開今，&,,jl朧，“。•二•，，二，。．。。二，,，二•．• 幫書•磚V留間VA:&l州月k勵瞥．“•·貿，f‘•，&&＿『一‘．一。。．。－一＿__ 一’&’一”&”一”一”&”不”&&&&&&&&&&&&&&&&&&.’、一可叮方而可不而可 一一一一一一一一一一一一一一一一· 〕：。．挪魚 …！一：&::：為：:”森，,::,::::--.－·竺：！•••：..-.．二二，:::,-,:::&m,1.,...:..．一竺！！蓄 ”州戶‘漁，&N‘鰥訌戚．叩．臘州，．物”黔！撇審 1甲奮補革日，閱州叩響奮禹 ;&.&,“蹤，添雜轎矓爍·蔔 j。徐？·，•／/＃當臘汗 k 口口口口口口嘆口■■•■■■■•么” � � ' Annex Tab1e I.7: - .. __-_ ._.:-: __ _ . д - � п�r►sчви� w• ицWеи ur в�вТиr 1r7f ru ии ын:р.-с•хМиiи�с ыльГвв�и ' � • . • в - - - - . rваlвrво t1[1 1 . � иt6kllti4и l1lE� :��- в[�еиоlМ[ УАиlввLL.• вlевв иимв[и 1К р[�1и• [и 1�[1 iY t+i/ � Мви[вi►е!!) Rh1[11(и ОИ •1е• иииУ[и [.� У! 1Y11e1 1а[1L11Y � r� да1 вFе - � абU N.(. f.вы,ьiи • � I►2 /вМ1L1 elaMИlиi вlвlие [LMO ди 4м( . � е м т! и•1и ив оУиlви сми[!+!►и и/ оlнсииlа[ � rв1 мм [ц[ив[• �в ♦вир иииl0 � 9в +и• Of eCNW1L � • +в А!Ю/ ив1QАlвL r] еиииС♦ Ot ыв[еи lивв�r � autlt►►e s а.5.е[в • .мц.®1о ив vиlвисс oi sим и1 виwиtи нааи оеи.иt 1 . А lридвЕ •.2+111 иЕ61ь[ Ч1ои 1i• re/.вь1/1 1в.1ь1�в i1в1вlвi � •ртиll[9 и /ривве л.а,[е! исllоw► <.ы . �1l.2,t/t /.вт1:±� 9i1N/Аво [ии0М i.i�ьSв � ................. Vввввi,,li !и 1м! teualt{1'+ ----••••••_•.-..•• •_.-•.-°-..• еви[.rtc/ иЧ[ tи Ои! tви.ltии +ви2иti i r[tв /1р lвдW1 0 1 rвиlArL[ веlа 1и 1ви U в► tUl[ива[[ / v1 .+.]s1ые.,! •.в1••в •.в ц еа •.ли дsе .,1l+е►]гс.ь[ •1в1вг]ь А.и1в► fs.�]• лito .о1т]в1!l.,9 в.►sтsт в.вевв! 1в.1вt д1т .111ИЧt.,] �•:otts в.lвв Ч тl.1�• 1 // .вв1lSвв21 .�,ei�t► i.0!]01 1/в1/f „ [веlе •в+1+\в:вв.,1 r/.l91ь• ввв31]I ►вв,S ' N де1 +в�ь16в21ь..А •в.ее+►в в.Чееь в.вв• � ti- в SA+4i►1.в1f !�lв1е9 I•в/вео !.lв9 � д/ .,[1в�2Ut.,l •1в1s1+1 f.NlЧ fвтвl " де �]т1 ]!►••• 1в►1,т► �.вllos l.вы � � +t •1►�e1u.,i !.иlи t,Nиt �.е!! [соивии[в �.1ева11 ` о[sеиУ[и[ мвв[�вд[,. в[�ив иими[w1 д в[и[мо !и 1�as !о !+в• ,, � оУИNАи1 Iввl! .. дви]а/le ми`[]1�[ в и lрУАМ[ 11вУ СмвиG! •1и�lt и р - YL1a . д9 1о1и[ rлсl►,[r +.t[в1в •.••s.o в.в•sьs•r •.иs►в .••тл]•r[.•т •.•1•++ + У!! ьб[ I•y1+il !�]N!! в�т►1+j.� •t.S1S►/ •.1Чц]Ie•в1 •ItNttь , ►с[е lв►2►]1 в.]i1e1 •.lи:е• ...sвt►f .►lттвilt.вь в,►sат r]t ас! лг иав[.n[ /в►тв]+ /,]1а1• в.евslв ✓ о.е11+• .1[1r+ы !•в� !ltвtl� м s.и1`r в►.ии1м, в[агие ]ио ов оие евf9.в! в�]лs ц s.еент �•/.�и]f .в.lеsеигl •!./ьвiь = Gвв[t иААв Аи 1[иI1М Lии1а1lви 011 01мЕАУ[Ot 1в13в11 в.]в+f• в�lвlв! .1I.1]/1♦ •.+13 Ч1е1•в1 •б�вS!►• � дв1 Им [ввЕмliо 1�Ifr1в в.20vе] •.lввт• .в.оN[� •.[a1S�]1l.Ч •/./в2в� дв двио О.г�2► !,1'r3S D.]и►] в.1в1е• -в.в,1Sв .1fo+r+o[•вf в.lвlаS +! rио sw lСм�а !.1l+и •.т+1s• /. н11►� в.в•е!• -.1 и r3ьт[-в1 -в.ls[:[ r. rnur ивц вlва /.1re[в t.t+�i1 в.lваls ••.я] и в .7]11гsы -!1 е:/[1i• +] цw вс[ w-а.[в /ии►+ •.}.в1• в�]+[11 в.ввевl в.в1s[• •1в N`в!-в1 вtв/е�• пом![•и11 - �.lоы11 д[ви •[1 •сеа ru 1• ен[r r.r дs w +ь „] ri ,[19е• !.е/мв .,.s[9ое •е.9втвд е.slлв• -в.•.вr[] -в.1]е[• •в.ви[• .в,в.►еr в�ив[• -в.в1 Н а в,в1i[• .У1►[� rf1 .в.ь19' ..lвово о.+аьs.0 в.Nвr• .е.вrовв !. ЧS+• ►.н 1+2 о.о9slв .е.1►elt /. н+•! -!.е]n!f -�.i:вsr ва[в .l.9/1в� ,,e+s•r 1.овеев l.вlма -в.1.иt в.вв!•r •.lьт.+ /.lsf►т -в.111Ч в.►еr1► -l.в2и: •i.1]ьs� д1] •.s1[вв s,вввrв е.в1+вS 1.оеввв а.1доlь •.11тть в.ввf3• в.l�вSа в.т�+]] •в.в]в]1 в.в/)1с •1.ьвь,1 2г .е.в•в[[ •,.в•.в. -в.••!a� в.1ае►• [.е•и. /.1в•s. •.11в1• -е.л[9:: е.111l1 •.в1ти l.в2t[! -:.�и:1 tвlц -е.12в1• в,веS+е ь.вЧN e�1t]т• t�1s►S• 1.вивв 1.1в►][ ♦.11 цt 1,3+1I: l.1ь1е! •о,lввь! .�.т и+► де1 .в.ыь+lе ,.ва1•т е.в.т.� •.вЧ►и e.1tr Н в.1+.1[ 1.вц аь в.11 И1 в.в►►+• •.2ьв1ь /.в•i•• •1.,►r.� д► .,,.вr1►в ,.вsslв •�/st►2 •�еlвsа -в.вllтт в.[1вf1 t.111►1 1.а Ч в в.т1и• ►.1119+ -в,Чьь1 •1.1+]+1 д1 в.е0вfь .,.1ьо1е •в.f11+е lвтS+т] f.I111I I.1+112 /�в! н ► /.21/еь !•веовв /вт0и • •.од►11 •I.1Ъ1в! дь •н,/i3er о.е/eef ев/евf• •в.е]ое1 •.ifтlв �.1►1е/ •,21ь1ь i.111S• в.твв•• 1.евЧ • ►.в1ва• •1.2�91д д] д. l.е1Ъа• .,.i]в]9 -в,в]ав1 •.вв1[• в.�т211 .в.l�ввв l.вь3И -в./�ra! •.в]в[1 •.►1ов• 1.ввев• •�в1lтli д! •,![safil .п,�гтвSь .о,отаsв .в�вЧь• •о,аl+]1 .в�т1lRf -l.rве►/ .е.1не! ./�fsfвe .в,2++19 •t.112ef �.1цоы � 1 � -j ' { � t • ' � � 1 � � _ �ееер•j •etSt•н ttоы 'е- tвtве•в° еtв[г•е- в�[1S�е• •г�tг•в- 1в•!►'е- м att•в• г+�tп'е- �о.�е•ь- у+len•е qp � a s�°g- еееве'[ tttse•г •г*•е•г sее.л•в° sz1e1 � ы м 1•е- ti•г�•а �еггв'е •впео'е- Iьгпе•ь sьгоп°е гр гген •в. t[tse•е еоеав•t •ееlг'9 et1+�•е ¢t11r•в ►1вs1•е гвгtn•• nле ы•р- агвее•е "sSы •о •r1.n•r. тл �ь1•д't• •teee'е •еоРг•е г►еее'[ tN[г'е н+е►'г +лtв['е [[е+�•е r,►rt•е e+ a r°е• sьtiг'.- оt1[в'в г. n[вtг•g. евие•е- в[Рrг•е [►•tг•е евеее•1 ц е.[•е г1сг1•е rrtte'р r.tsг•е sг*г•°е рlсгв•� 1[,,[е•е- +л � iьsllp•g- �г1е1'е 1еlrt'е вtвsт'е J1г+[°е вовее•1 Ss�et'е [гевs•е +.•s1•е +rs•е•а- ег,ео•r- иsгв�'п- 1�л ir«tt°е- е[г►1'в- etost•е •ьtы •е tts![•г tsвet•е neneo•[ rtSSt•г tstвP•е и[ев•а е,гее•ь [.►•v'в- ¢lсrэ � ��etr•[- 1¢.ке•е гвsсе•е ttt.t•в tгlse•е 1ге.г•е +[асг'• •"вев•[ аггее•е г�е1n•в вг1[в•• г.tьв•а- •а � �[•rj•е- nttг'е егвео•е- t.•г[•+ tetse•г e+гst•е [в[+г•е r¢г[1'в .ne•n°1 ыlsе•е- tsz.o•�- ц е+и°е г[" � t+t[e•i• н еве'е- +tиее'в [емlг°г- Sггго•v •et ы'г- ►гtее•е t1e1e•е вvt�e•д• евеев•t lвгв►'в есг�,т�"° в». уΡ.+тt•а- [вtве'е ессее°е 5вt1г•е- •lsre•е п tee°в- n►вев•в rtltn•в 9iгы•в• Увгвь•в вевев°1 ►вiиг•b- 1[л � i+trn'� с►гвв•е •гt.n°в- etttn'в [1►гв•n- •сtм•в- [еьrе•в• eat•�•е• еtне'в nсгвР'м •вв.t'а- еоепв•1 увл 4� . [в tв м вл еА 1вл э1t►) oI гtд г17о rt" vав � �в�иr'� ' eiel+r4'в ' (•1иrlеип)Б '��в ? •s'1ее1е te-мг+Ret' sыа•е еыы•1 влИ••е ie+1t•s л9лтг ь11.. ю 1�,�nпе � 1�вге'г' 1е`ltrrlг►[•• 9г1tе•н 1Sr•в•+ ье+[в•е ГN1t•е в.9ьI1гМ lПU9 в• � 1;ыв3е tг-ltSв.в/1' eilte•е гвnы•е ►1вР•°в ввS1�'е +пnи я�r еив rл w [�[iе�г- ►Н i+5lмtfP'• ¢t+t•°е- в[•е••r s1вt+'г See/t'л �эи.п ин•9 вл i�i�г ` ♦г•ieie• fe°lгег►г[г•• вiггl•е- глы'n .ьSle•е r+c1f'е вгiи)�tt ии ¢вл � � е,а[глг• 1е-[еее[сгг'• [•в и•е• iгРеетя егsге'е tsr/t•а э ц.� � 1j1ег'г- [е•Нlицtв°• Нsы•г- nгlы'е ;nete'г ttrti'! 7ип ьи пr;г tпtrie эиlаивlи в9lurf �I � •[•1►�е 1в°+г[е•е�►• lfел•• е[ыв•о s•вге'1 cs•S.°е �чrtrrи 1е эв+ ttл r� 1t111•г .r-DllвгSff' eStet°е• Sыег°е 1ь[+г•i tв4►i°1 а)иrппе игиоw ¢гf 1!+ е1•► � t�sre°г- гг-гы геt�Р•- ев+ег'е• ыеlв•е гв r г•е `еев;°е э�r в[е � 1i/er�e- t•-t•гР•вt1°- [вtвв°• [19ы•в t11де•в t+ue°в 11111�e• U9to1 [л � alip е и 11Аи[@ 1!+иtи9 nви 9МвПвf N ь�1/119пи э1tт1►гrл � � � I � ¢iprl iиьиипR { epir ц л)иrиl7r• •вА ••Tlг.lrve сиl0и)в7в гг ;, [ев°г •1r .е �°ввsв[•+[•• t1и.[еие�1 . � !е е •troe t ег�lгее• tл r�e•1 г п 1е°в ге.ге•е. e�.�..rrt,1•- м , г� • �ее°в ¢вгЧ°е lгвае'е f�'s ' � t1f•0 •efe'е ¢св+в11 в f tetle•N ва'е•StatS1�' iл N - 1S•°е ев►ее•е •г•ге••. ►•-.t►iвгtt'• iвр з Р [+в�Р в¢S[г'в вt+le�s- tlь-�вrrsclt°q )ttr� л[ [1 ввrРг°Ф ег n е е• [ь•+•1►в1f.'- вл ,г ,� fвt'1Р U вн•s •r.1e'в [••.tteresв• t6A 1+в•i н еее°i ¢►tв1•� е.•.1tefS+S° •)!е , 9гг°s [в•ее"е лсгв'е- [ь•1вог+lе[°- [tв � ва•°е •двев•г 1[�•и•в• гь•аегlее[!°' tл , ,I е 11чrу19п1 r.tlиеа и1 rttp г9вvlи+Ч i м ьпrвэ е1в .1tв в ti+1т hrл ..-.. ...---•.°°°--... 1 °°-•.••°•°°-- иntlvhe7 1и1 иt 1ли •7lеr[итв •-••°-----•-- •-•»•� -°° иnil.пл7 .и1 и1 е7lьвrьrд , � Sf�aг•• иимиэ льеоив[е 1litг•е г[¢tв•4•У °[tвь 9vno@в!ь евliг•е 2rтneo r оз[впге. • � евi,вс•ь0 вТвбе'1 еМб¢е•г1 °i1 ип►вйУрllи � ва+аг'е уивпвt ь „ ! /и.nne иеlи rlиrnni Mv иМ �8 17ив1иrл �t► ееев9.rь гыгг�в ь 7711/1nv � i1�,пв �*irw 10 17МППе tл . 1в1М)1ви /рПМ вП 1(иtи7е 1п tNA •А . Ф}и�п ли.7 +д еlеиглг• ии ¢►д , . t1t.э ;ип Мп пиэР мntelf 1иlмиs9i вllнвf �! • 7�r[м.и 1а f�r г1в аеи.пвв иrиbw �п �ar гэвв у„ tt� ° str9riь. гвiапl ти •.В и[Мипи в11► ип о9Чеп�г islaмtrsл '' I (в17 ипtеСtи7lи tnlr/r лэивив[ри +ил "вrcтtrrл 1и1ли7l7n 1 1Ri7 911.1ивл • ° S91вт1иvА :еlипмr:rl-п1��t Nn 3A111R l1�МУи7;.М.1 .7р Н)1CS7w�7y ---------___ ' ы � � а_[QE1, Хацц� М в. � � ° э • � _ . � � • . 1 у 化 Annex Table 1.10: REGRESSION OF DESIRED NUMBER OF CHILDREN ON SELECTED DEMOGRAPHIC AND SOCIO- ECONOMIC VARIABLES DEPENDENT VARIABLE.. V7 DESIRED SIZE r, VARIABLE(S) ENTERED ON STEP NUMBER 1_ V9 PARITY V32 AGE AT MARIAGE ve YRS OF SCHOOL V6 LAND OWNED PCE P C EXPEND USE BINARY VARIABLE FOR FAMILY PLANNING USE V48 NUMBER OF CHILDRED DEAD MULTIPLE R 0.69417 ANALYSIS OF VARIANCE OF SUM OF SQUARES MEAN SQUARE F R a EGRESSION 7. 2727.07029 389.ES147 352.481133 SDUA:E I SQUARE 00.4-:0T 5 AodUST D :ESIDUAL 2653. 2932.16112 1.10 23 STANDARD ERROR t.05130 ----------------- VARIABLES jt4 THE EQUATION ------------------ ------------- VARIABLES NOT IN THE EQUATION -------------- VARIABLE a FETA STO ERROR 8 F VARIABLE BETA IN PARTIAL TOLERANCE F vs 0.4458612 0.74205 0.01110 1612.811 V32 -.;459693E-02 -0.0442: 0.000561 18.294 ve 568065E-01 -0.0623 0.0071 2 B53 V6 -3268938E-03 0.05219 0-00009 12.538. PCE -,1537717E-03 -0.00548 0.00044 0.120 USE -0.467OB65 -0.15984 0.04304 117.761 V49 -0.3103020 -0.14465 0.02875 116.505 (CONSTANT) 2.488046 co DEPENDENT VARIAMI.E.. V7 OESIR90, SIZE SUMMARY FABLE VARIABLE MULTIPLE R R SQUARE RSO CHANGE SIMPLE A 0 DETA 0.4458612 0.74205 vs PARITY 0,55474 0.420G8 0.42868 0.65474 -0.04425 W32 AGE AT MARIAQE 0.65831 0-43337 0.00469 -0.29187 -.1459693E-02 ve YRS OF SCHOOL O 66105 0.43699 0.00362 -0.28296 -.2568065E-01 -o.o6235 va LAND OWmED 0.66435 0.44136 0.00437 0.04797 .326893BE-03 0.05219 PCE C EXPEN 0 66444 0.44 48 0.000612 -0.013060 -.415377617E-03 -0 0054 8 CIO) USE :INARY VARIABLE FOR FAMILY PLANNING USE 0:67759 0,45 113 o.00 4 -0. 7344 0. 670a 5 -0.15984 V48 NUMBER OF CHILDRED DEAD 0.69417 0.48188 0.02275 0.26794 -0.3103020 -G.18665 (CONSTANT) 2.488624 V7 vo V32 va V6 PCE USE V48 V31 V7 1.00000 0.65474 -0-29187 -0.28296 0,04797 -0.13060 -0.07344 0.26794 0.35485 V9 0 65474 1.00000 -0.34777 -0.32448 0-00554 -0.15141 0.12169 0.56470 0.62458 V32 -0.29187 _O .34777 1.00000 0-32025 0.05690 0.15783* 0.06526 -0.20902 -0.04711 Va -0-20295 -0.32448 0-32025 1.00000 0.24387 0.43973 0.19410 -0.29631 -0.26635 V6 0.04797 0.00554 0-05690 0.24387 6-00000 0.25312 0,00017 -0.05784 0.01169 PCE -0.13060 -0.15141 0.15783 0.43973 0.25312 1,00000 0.07917 -O.It355 0.10846 USE -0.07344 0.;2169 0.06526 O.t9410 0.00017 0.07917 1.00000 -0.06118 0.08752 V48 0-26794 0 6 70 -0.20902 -0.2963t -0.05784 -0.11355 -0.06118 1.000,00 0.35128 V31 0.35485 0:62458 -0-04711 -0.26635 0.01169 0.10846 9.08752 0.35120 1.00000 r9m Annex Table I.10A: REGRESSION OF DESIRED NUMBER OF CHILDREN ON SELECTED DEMOGRAPHIC AND SOCIO- ECONOMIC VARIABLES DEPENDENT VARIABLE.- V7 DESIRED SIZE VARIABLE(S) ENTERED ON STEP NUMBER 1.. V31 AGE V32 AGE AT MARIAGE Vs YRS OF SCHOOL V6 LAND OWNED PCE PC EXPEND USE BINARY VARIABLE FOR FAMILY PLANNING USE V48 NUMBER OF CHILDRED DEAD MULTIPLE R 0.49155 ANALYSIS OF VARIANCE OF SUM OF SQUARES MEAN SQUARE F R SQUARE 0.24162 REGRESSION 7. 1367.39475 195.34211 120.75021 ADJUSTED R SQUARE 0.23962 RESIDUAL 2653. 4291.85666 1.61774 STANDARD ERROR 1.27190 ----------------- VARIABLES IN THE EQUATION ------------------ ------------- VARIABLES NOT IN THE EQUATION -------------- VARIABLE I1 BETA STO ERROR 8 F VARIABLE BETA IN PARTIAL TOLERANCE F V31 .42790115-02 0.31364 0.00026 261.387 V32 -.73393791'02 -0.22251 0.00060 151.757 V8 -.2868115E-0 -0.06963 0.00900 10.147 V6 .6701433E-03 0.10700 0.00011 36.406 PCE -.3146814E-02 -0.11215 0.00056 32.140 USE -0.1726131 -0-05907 0.051t3 11.396 V41 0.1337643 0.08044 0.03124 18.328 (CONSTANT) 3.447237 DEPENDENT VARIABLE.. V7 DESIRED SIZE SUMMARY TABLE VARIABLE MULTIPLE R R SQUARE RSO CHANGE SIMPLE R 1 BETA V31 AGE 0.35485 0.12592 O.12592 0.35485 .4210IE-02 0.31364 V32 AGE AT MARIAGE 0.44922 0.20180 0.07588 -0.29187 -,7339179E-02 -0.22251 VS vRS OF SCHOOL 0.46332 0.21466 0.01286 -0.28296 -.2860115E-01 -0.06963 V6 LAND OWNED 0.47221 0.22298 0.00832 0.04797 .6701433E-03 0.10700 PCE P C EXPEND 0.48233 0.23265 0.00966 -0.13060 -.3146814E-02 -0.11215 USE BINARY VARIABLE FOR EAMILY PLANNING USE 0.48619 0.23638 0.00374 -0.07344 -0. 1726031 -0.05907 V48 NUMOER OF CHILDRED DEAD 0.49155 0.24162 0.00524 0.26794 0.1337e43 0.08046 (CNSTANT) 3.447237 CD mmm ft..V Annex Table 1.11: REGRESSION OF EXCESS FERTILITY WOMEN BY SELECTED DEMOGRAPHIC AND SOCIO- ECONOMIC VARIABLES . .. .. .. .. .. .. U L T I P L E R E G R E 5 S 1 0 N . . ..VA R IAB.LERIBL LISTI REGRE StdN LIST I P>D DEPENOENT VARIABLE.. EX S| VARlABLE(S) ENTERED ON STEP NUMBER I.. V9 PARITY V32 AGE AT MARIAGE V8 YRS OF SCHOOL V6 LAND OWNED PCE P C EXPEND V48 NUMBER OF CHILDRED DEAD MULTIPLE R 0.70555 ANALYSIS OF VARIANCE OF SUM OF SQUARES MEAN SQUARE F R SQUARE 0.49780 REGRESSION 6. 972.97810 162.16302 141.41437 ADJUSTED R SQUARE 0.49428 RESIDUAL 856. 981.59432 1.#4672 STANDARb ERROR f-07085 ----------------- VARIABLES IN THE EQUATION ------------------ ------------- VARIABLES NOT IN THE EQUATION -------------- VARIABLE B BETA STO ERROR 8 F VARIABLE BETA IN PARTIAL TOLERANCE F V9 0.5349740 0.70645 0.02112 641.668 V32 .2793658E-03 0.00700 0.00103 0.073 VS .3165179E-01 0.06447 0.01394 5.123 VS -.1422414E-03 -0.01590 0.00023 0.382 PCE .9091314E-04 0.00230 0.00108 0.007 V48 .5940433E-01 0.04396 0.03601 2.632 (CONSTANT) -1.075247 DEPENDENT VARIABLE.. EA SUMMARY TABLE VARIABLE MULTIPLE R R SQUARE RSO CHANGE SIMPLE R 0 BETA V9 PARITY 0.70237 0.49332 0.49332 0.70237 0.5349740 0.70645 V32 AGE AT MARIAGE 0.70250 0.49351 0.00019 -0.20779 .2793658E-03 0.00700 VS YRS OF SCHOOL 0.70423 0.49594 0.00243 -0.14809 .3165179E-0 0.06447 V6 LAND OWNED 0.70444 0.49623 0.00029 0.01874 -.1422414E-03 -0.01590 PCE P C EXPEND 0.70445 0.49625 0.00002 -0.10317 .909f314E-04 0.00230 V48 NUMBER OF CHILDRED DEAD 0.70555 0.49780 0.00154 0.29899 .5940433E-01 . 0.04396 (CONSTANT) - 1.075247 EA v9 V32 VS V6 PCE V48 V31 Ex 100000 0.70237 -0.20779 -0.14809 0.01874 -0.10317 0.29899 0.29969 V9 0.70237 ,.00000 -0.31456 -0.28037 0.03322 -0.17446 0.38940 0.38797 V32 -0.20779 -0.31456 1.00000 0.23294 0.00652 0.11501 -0.07634 0.09025 VS -0.14809 -0.28037 0.23294 1.00000 0.21903 0.40701 -0.30889 -0.17159 V6 0.01874 0.03322 0.00652 0.21903 1.00000 0.28779 -0.08314 0.02388 PCE -0.10317 -0.17444 0.11501 0.40701 0.28779 1.00000 -0.10675 0.17826 V48 0.29899 0.38940 -0.17634 -0.30889 -0.08314 -0.10675 1.00000 0.17847 V31 0.29969 0.38797 0.09025 -0.17159 0.02388 0.17826 0.17847 1.00000 Annex Table I.11A: REGRESSION OF EXCESS FERTILITY AMONG EXCESS FERTILITY WOMEN BY SELECTED DEMOGRAPHIC AND SOCIO--ECONOMIC VARIABLES DEPENDENT VARIABLE.. EX VARIABLE(S) ENTERED ON STEP NUMBER I.. V31 AGE V32 AGE AT MARIAGE Va YRS OF SCHOOL V6 LAND OWNED PCE P C EXPEND V48 NUMBER OF CHILDRED DEAD m4 MULTIPLE R 0.45649 ANALYSIS OF VARIANCE OF SUM OF SQUARES MEAN SQUARE F R SQUARE 0.20838 REGRESSION 6. 407.29324 67.88221 37.55442 ) ADJUSTED R SQUARE 0.20283 RESIDUAL 856. 1547.27918 1.80757 ) STANDARD ERROR 1.34446 ----------------- VARIABLES IN THE EQUATION ------------------ ------------- VARIABLES NOT IN THE EQUATION ------------- VARIA(LE B BETA STD ERROR B F VARIABLE BETA IN PARTIAL TOLERANCE F V31 .6232518E-02 0.31809 0.00064 94.121 V32 -.7851244E-02 -0.19672 0.00127 38.012 Va .3512407E-01 0.07154 0.01802 3.801 V6 .5465655E-03 0.06110 0.00029 3.638 PCE -.6343684E-02 -0.16072 0.00140 20.545 V48 0.2939626 0.21755 0.04402 44.599 (CONSTANT) 0.8130593 DEPENDENT VARIABLE.. EX SUMMARY TABLE VARIABLE MULTIPLE R R SQUARE RSQ CHANGE SIMPLE R B BETA V31 AGE 0.29969 0.08981 0.08981 0.29969 .6232518E-02 0.31809 V32 AGE AT MARIAGE 0.38134 0.14542 0.05560 -0.20779 -.7851244E-02 -0.19672 Va YRS OF SCHOOL 0.38340 0.14699 0.00158 -0.14809 .3512407E-01 0.07154 V6 LAND OWNED 0.38406 0.14750 0.00051 0.01874 .5465655E-03 0.06110 PCE P C EXPEND 0.40882 0.16714 0.01963 -0.10317 -.6343684E-02 -0.16072 V48 NUMBER OF CHILDRED DEAD 0.45649 0.20838 0.04124 0.29899 0.2939626 Q.21755 (CONSTANT) 0.8130593 Annex Table 1.12: REGRESSION OF EXCESS FERTILITY AMONG EXCESS AND PAR FERTILITY WOMEN BY SELECTED DEMOGRAPHIC AND SOCIO-ECONOMIC VARIABLES *e* ** . * * * * * * ve + * * * * * * UL TIPLE R E G R E S S 1 0 N * * * * * * * * * * * * VARIABLE LIST I REGRESSION LIST I DEPENDENT VARIABLE.. EX D VARIABLE(S) ENTERED ON STEP NUMBER 1.. Vs PARITY V32 AGE AT MARIAGE Va YRS OF SCHOOL V6 LAND OWNED PCE P C EXPEND V48 NUMBER OF CHILDRED DEAD MULTIPLE R 0.73422 ANALYSIS OF VARIANCE OF SUM OF SQUARES MEAN SQUARE F R SQUARE 0.53907 REGRESSION 6. 2282.03966 390.33994 - 357.49123 ADJUSTED R SQUARE 0.53757 RESIDUAL 1834. 1951.21835 1 06391 STANDARD ERROR 1.0314. ----------------- VARIABLES IN THE EQUATION ------------------ ------------- VARIABLES NOT IN THE EQUATION -------------- VARIABLE B BETA STO ERROR 8 F VARIABLE BETA IN PARTIAL TOLERANCE F V9 0.4442349 0.63305 0.01402 1004.545 (A V32 .1000988E-02 0.02771 0.00062 2.596 VB .2825455E-0 0.06369 0.00868 10.606 V6 -.4402870E-03 -0.06295 0.00012 14.129 PCE -.3691473E-03 -0.01210 0.00056 0.440 V48 0.3122668 0.20209 0.02941 112.735 (CONSTANT) -1.402164 SUMMARY TABLE ( VARIABLE MULTIPLE R R SQUARE RSO CHANGE SIMPLE R B BETA V9 PARITY 0.71017 0.50434 0.50434 0.71017 0.4442349 0.63305 V32 AGE AT MARIAGE 0.71090 0.50538 0.00104 -0.20178 .iOO98BE-02 0.02771 Va YRS OF SCHOOL 0.71097 0.50548 0.00010 -0.22840 .2825455E-01 0.06369 V6 LAND OWNED 0.71466 0.51073 0.00525 -0.04207 -.4402870E-03 -0.06295 PCE P C EXPEND 0.71466 0.51074 0.00001 -0.15519 -.3691473E-03 -0.01210 V48 NUMBER OF CHILDRED DEAD 0.73422 0.53907 0.02833 0.52014 0.3122668 0.20209 mCONSTANT) -1.402164 unan .5 . EX V9 V32 Ve V6 PCE V48 V31 EX 1.00000 -0.71017 -0.20178 -0.22840 -0.04207 -0.15519 0.52014 0.32156 V9 0.71017 100000 -0.32706 -0.34393 0.03423 -0.20795 0.53308 0.46388 V32 -0.20178 -0.32706 1.00000 0.30573 0.01494 0.83367 -0.19476 0.01084 VB -0.22840 -0.34393 0.30573 1.00000 0.23217 0.46251 -0.30988 -0.23944 V6 -0.04207 0.03423 0.01494 0.23217 t.00000 0.25619 -0.06380 0.03742 PCE -0.15519 -0.20795 0.13367 0.46251 0.25619 1.00000 -0.14092 0.1420R a a A 0 Annex Table I.12A: REGRESSION OF EXCESS FERTILITY AMONG EXCESS AND PAR FERTILITY WOMEN BY SELECTED DEMOGRAPHIC AND SOCIO-ECONOMIC VARIABLES * * * * * * * * * * * * * * * * * * * * * * M ULTIPLE REGRESSION * * * * * * * * * * * * VARIABLE LIST I REGRESSION LIST 2 DEPENDENT VARIABLE.. EX VARIABLE(S) ENTERED ON STEP NUMBER I.. V31 AGE V32 AGE AT MARIAGE VS YRS OF SCHOOL V6 LAND OWNED PCE P C EXPEND V48 NUMBER OF CHILDRED DEAD MULTIPLE R 0.58078 ANALYSIS OF VARIANCE OF SUM OF SQUARES MEAN SQUARE R SQUARE 0.33731 REGRESSION 6. 1427.89918 237.98320 155.58123 ADJUSTED R SQUARE 0.33514 RESIDUAL 1834. 2805.35883 1.52964 STANDARD ERROR 1.23679 ---------------- VARIABLES IN THE EQUATION ------------------ ------------- VARIABLES NOT IN THE EQUATION -------------- VARIABLE B BETA STO ERROR B F VARIABLE BETA IN PARTIAL TOLERANCE F Vq1 .4206347E-02 0.25013 0.00036 140.300 V32 -.4412819E-02 -0.12216 0.00073 36.256 V8 .3151064E-01 0.07103 0.01087 8.408 V6 -.7352709E-05 -0.00105 0.00014 0.003 PCE -.4447720E-02 -0.14581 0.00069 41.114 V48 0.6706959 0.43406 0.03184 443.667 (CONSTANT) -0.1236665 DEPENDENT VARIABLE.. Ex SUMMARY TABLE VARIAUtE MULTIPLE R R SQUARE RSQ CHANGE SIMPLE R a UETA V31 AGE 0.32156 0.10340 0.10340 0.32156 .4206347E-02 0.25013 V32 AGE AT MARIAGE 0.38150 0.14554 0.04214 -0.20178 -.4412819E-02 -0.12216 V8 YRS OF SCHOOL 0.39332 0.J5470 0.00916 -0.22840 .315iO64E-01 0.07103 V6 LAND OWNED 0.39427 0.15545 0.00075 -0.04207 -.7352709E-05 -0.00105 PPE P C EXPEND 0.42070 0.17699 0.02154 -0.15519 -.4447720E-02 -0.14581 V48 NUMBER OF CHILDRED DEAD 0.58078 0.33731 0.16031 0.52014 0.6706959 . 0.43406 (CONSTANT) -0.1236665 . BEST COPY AVAIL ABLE Annex Table 1.13: REGRESSION OF EXCESS NET FERTILITY AMONG EXCESS NET FERTILITY AND PAR FERTILITY WOMEN BY SELECTED DEMOGRAPHIC AND SOCIO-ECONOMIC VARIABLES * * * * MULTIPLE REGRESSION ** VAhABLE LIST I R,-GRtSSION LIST I DEPENDENT VARIABLE.. EX VARIABLE(S) ENTERED ON STEP NUMBER I.. V9 PARITY V32 AGE AT MARIAGE Ve YRS OF SCHOOL V6 LAND OWNED PCE P C EXPEND V48 NUMBER OF CHILORED DEAD MULTIPLE R 0.78829 ANALYSIS OF VARIANCE DF SUM OF SQUARES MEAN SQUARE F R SQUARE 0.62140 REGRESSION 6. 2496.79003 416.13167 450.80831 ADJUSTED R SQUARE 0-62002 RESIDUAL 1648. 1521.234t4 0.92308 STANDARD ERROR 0.96077 ----------------- VARIABLES IN THE EQUATION ------------------ ------------- VARIABLES NOT IN THE EOUATION -------------- VARIABLE B BETA STD ERROR 8 F VARIABLE BETA IN PARTIAL TOLERANCE F V9 0.4383133 0.60993 0.01420 953.263 V32 .966tO93E-03 0.02627 0.00061 2.525 VS .2758258E-01 0.06058 0.00858 10.323 V6 -.2885016E-0 -0.04024 0.00012 6.290 PCE -.4192775E-03 -0.01347 0.00055 0.590 V48 0.5098956 0.29216 0.03264 244.089 (CONSTANT) -1.368506 DEPENDENT VARIABLE.. EX SUMMARY TABLE VARIABLE MULTIPLE R R SQUARE RSO CHANGE SIMPLE R B BETA VS PARITY 0.74857 0.56036 0.56036 0.74857 0.4383133 0.60993 V32 AGE AT MARIAGE 0.74940 0.56160 0.00124 -0.21740 .966003E-03 0.02627 Va YRS OF SCHOOL 0.74944 0.56165 0.00005 -0.25387 .2758258E-0 0.06058 Vs LAND OWNED 0.75186 0.56530 0.00365 -0.04369 -.28B85016E-03 -0.04024 PCE P C EXPEND 0.75188 0.56532 0.00002 -0.16502 -.4192775E-03 -0.01347 V48 NUMBER OF CHILORED DEAD 0.78829 0.62140 0.05608 0.62273 0.5098956 0.29216 (CONSTANT) -1.368600 EX V9 V32 V V6 PCE V48 V31 EX 1.00000 *0.74857 -0.21740 -0.25387 -0.04369 -0.16502 0.62273 0.34257 V9 0.74857 1.00000 -0.33484 -0.35804 0.01505 -0.20978 0.57130 0.45549 V32 -0.21740 -0.33484 100000 0.32016 0.03081 0.14263 -0.19060 0.02519 VS -0.25387 -0.35804 0.32016 1.00000 0.24092 0.47369 -0.30260 -0.23306 V6 -0.04369 0.01505 0.03081 0.24092 1.00000 0.26455 -0.08374 0.03302 PCE -0.16502 -0.20978 0.14263 0.47369 0.26455 1.00000 -0.15538 0.15986 V48 0.62273 0.57130 -0.19060 -0.30260 -0.08374 -0.15538 t.00000 0.26520 V31 0.34257 0.45549 0.02519 -0.23306 0.03302 0.15986 0.26520 1.00000 . BEST COPY AVAILABLE Annex Table I.13A: REGRESSION OF EXCESS FERTILITY AMONG EXCESS NET FERTILITY AND PAR FERTILITY WOMEN BY SELECTED DENOGRAPHIC AND SOCIO-ECONOMIC VARIABLES seeswa* * * * * * * * * * * * * * * * * MULTIPLE REGRESSION * * * * * * * * * * *** * VARIABLE LIST I * REGRESSION LIST 2 DEPENDENT VARIABLE.. EX ) VARIABLE(S) ENTERED ON STEP NUMBER 1. V31 AGE V32 AGE AT MARIAGE V8 YRS OF SCHOOL V6 LAND OWNED PCE P C EXPEND V48 NUMBER OF CHILDRED DEAD MULTIPLE R 0.66926 ANALYSIS OF VARIANCE OF SUM OF SQUARES MEAN SQUARE F R SQUARE 0.44791 REGRESSION 6. 1799.73322 299.95554 222.§4125 ADJUSTED R SQUARE 0.44590 RESIDUAL 1648. 2218.29095 1.34605 STANDARD ERROR 1.16019 ----------------- VARIABLES IN THE EQUATION ------------------ ------------- VARIABLES NOT IN THE EQUATION -------------- VARIABLE B BETA STD ERROR 8 F VARIABLE BETA IN PARTIAL TOLERANCE F V31 .4203229E-02 0.24025 0.00036 135.864 V32 -.4521643E-02 -0.12297 0.00072 49.086 o V8 . .2936444E-Oi 0.06449 0.01086 7.314 V6 .1287329E-03 0.01796 0.00014 0.874 PCE -.4295962E-02 -0.13804 0.00069 38.408 V48 0.9339919 0.53515 0.03474 722.782 (CONSTANT) -.9774014E-01 DEPENDENT VARIABLE.. EX SUMMARY TABLE VARIABLE MULTIPLE R R SQUARE RSQ CHANGE SIMPLE R BETA V31 AGE 0.34257 0.11735 0.11735 0.34257 .4203229E-02 0.24025 V32 AGE AT MARIAGE 0.41046 0.16848 0.05112 -0.21740 -.4521643E-02 -0.12297 VS YRS OF SCHOOL 0.42481 0.18046 0.01198 -0.25387 .2936444E-01 0.06449 V6 LAND OWNED 0.42530 0.18088 0.00042 -0.04369 .1287329E-03 0.01796 PCE P C EXPEND 0.45363 0.20578 0.02490 -0.16502 -.4295962E-02 -0.13804 V48 NUMBER OF CHILDRED DEAD 0.66926 0.44791 0.24213 0.62273 0.9339919 ' 0.53515 (CONSTANT) -.9774014E-01 BEST COPY AVAILABLE Annex Table 1.14: REGRESSION OF CONVENTIONAL FAMILY PLANNING USE AMONG NON-STERILIZED EXCESS AND PAR FEPTILITY WOMEN, ON SELECTED DEMOGRAPHIC AND SOCIO-ECONOMIC VARIABLES * * * * ****. ***** . U T I P L E R E G R E S S I 0 N ***** .******* VA.ARLE LIST i DEPENDENT VARIABLE.. YI REGR ESSION LIST I VARIABLE(S) ENTERED ON STEP NUMBER I_ VS8 V9 PARITY VS9 V43 CASTE PCE P C EXPEND V6 LAND OWNED Va YRS OF SCHOOL MULTIPLE R 0.37699 ANALYSIS OF VARIANCE OF SUM OF SQUARES MEAN SQUARE F R SQUARE O.14212 REGRESSION 7. 34.16001 4.88000 26.43634 ADJUSTED R SQUARE O.13675 RESIDUAL 1117. 206.19199 0.18459 STANDARD ERROR 0.42964 ----------------- VARIABLES IN THE EQUATION ------------------ ------------- VARIABLES NOT IN THE EQUATION -------------- VARIABLE B BETA STD ERROR 8 F VARIABLE BETA IN PARTIAL TOLERANCE F V58 .3014371E-03 0.02460 0.00097 0.096 V9 . 3977917E-02 0.01996 0.02159 0.034 VS9 -.4448407F-03 -0.02538 0.00186 0.057 V43 -,9937383f-04 -0.00554 0.00506 0.039 PCE .1860635E-03 0.02196 0.00028 0.445 V6 -.1562028E-03 -0.07864 0.00006 7.126 Vo .45679478,01 0.35270 0.01020 20.073 (CONSTANT) 0. 120311# DEPENDENT VARIABLI.. VI SUMMARY TABLE VARIABLE MULTIPLE R R SQUARE RSQ CHANGE SIMPLE R B BETA VS8 0.34796 0.112107 0.12107 0.34796 .301437tf-03 0.02460 V9 PARITY 0.34871 0.12160 0.00053 -0.14863 .3977917E-02 0.01996 V59 0.34881 0.12167 0.00007 -0.13259 -.44484'97-03 -0.02538 V43 CASTE 0.34908 0.12186 0.00019 -0.02887 -.9937303i-03 -0.00554 PCE P C EXPEND 0.34934 0. 12204 0.00018 0. 18733 .1860635 -os 0.02196 V6 LAND OWNED . 0.35b96 0. 12671 0.00467 0.01535 -.156202hf-03 -0.07564 VS YRS OF SCHOOL 0.37699 0.14212 0.01542 0.36887 .4507947-01 0.35270 (CONSTANT) , 0. 1202796 V1 Vs VS9 V43 PCE V6 VB VS8 YI 100000 -0.14863 -0.13259 -0.02887 0.18733 0.01535 0.36887 0.34796 vs -0.14863 1.00000 0.96424 0.14700 -0.2676t 0.01768 -0.36017 -0.36585 VS9 -0.13259 0.96424 1.000O0 0.13798 -0.23338 0.03473 -0.31214 -0.31466 V43 -0.02887 0.14700 0.13798 1.00000 -0.00195 0.10971 -0.03750 -0.03511 PCE 0.18733 -0.26761 -0.23338 -0.00195 1.00000 0.2711a: 0.49272 0.50061 V6 0.01535 0.01768 0.03473 0.10971 0.27118 1.00000 0.23651 0.23377 VS 0.36887 -0.36017 -0.31214 -0.03750 0.49272 0.2365i 1.00000 0.S3527 0 VS8 0.34796 -0.36585 -0.31466 -0.03511 0.50061 0.23377 0.93527 1.00000 I BEST COPY AVAILABLE Annex Table 1.15: REGRESSION OF CONVENTIONAL FAMILY PLANNING USE AMONG NON-STERTLIZED EXCESS AND PAR NET-FERTILITY WOMEN ON SELECTED SOCIO-ECONOMIC VARIABLES * * * a * * * * * a * * * * * * * U T I P L E R t G R E S S 1 0 N * * * * * * * * * a ** * * VARIABLE LIST 1 REGRESSION LIST I DEPENDENT VARIABLE.. Y1 VARIABLE(S) ENTERED ON STEP NUMBER 1.. VS8 V9 PARITY VS9 V43 CASTE PCE P C EXPEND V6 LAND OWNED VS YRS OF SCHOOL MULTIPLE R 0.37993 ANALYSIS OF VARIANCE OF SUM OF SQUARES MEAN SQUARE F R SQUARE 0.14435 REGRESSION 7. 31.72319 4.53188 23.54540 ADJUSTED R SQUARE 0.13822 RESIDUAL 977. 188.04737 0.19247 STANDARD ERROR 0.43872 ----------------- VARIABLES IN THE EQUATION ------------------ ------------- VARIABLES NOT IN THE EQUATION -------------- VARIABLE a BETA STD ERROR 8 F VARIABLE BETA IN PARTIAL TOLERANCE F VS8 .1248580E-04 0.00102 0.00103 0.000 V9 -.9452902E-03 -0.00468 0.02328 0.002 V59 -.8129938E-04 -0.00459 0.00200 0.002 V43 .1725379E-03 0.00093 0.00559 0.001 PCE .2386067E-03 0.02771 0.00031 0.608 V6 -.1772966E-03 -0.08764 0.00006 7.728 vs .4921009E-01 0.37412 0.01107 19.753 (CONSTANT) 0.1414522 DEPENDENT VARIABLE.. YI SUMMARY TABLE VARIABLE MULTIPLE R R SQUARE RSO CHANGE SIMPLE R a BETA VP8 0.34621 0.11986 0.11986 0.34621 .1248580E-04 0.00102 w PARITY 0.34747 0.12073 0.00088 -0.15899 -.9452902E-03 -0.00468 Vs 0.34748 0.12074 0.00001 -0.14164 -.8129938E-04 -0.00459 V43 CASTE 0.34766 0.12087 0.00012 -0.03058 .4725379E-03 0.000" PCE P C EXPEND 0.34811 0.12118 0.00031 0.19659 .2386067E-03 O.O:... V6 LAND OWNED 0.35644 0.12705 0.00586 0.01535 -.1772966E-03 -0.08764 vS Y S OF SCHQOL 0.37993 0.14435 0.01730 0.37027 .4921009E-01 0.37412 (CONSTANT) -0.1414522 Yl V9 VS9 V43 PCE V6 Ve VS8 VI 0.00000 -0.15899 -0.14164 -0.03058 0.19659 0.01535 0.37027 0.34621 V9 -0.15899 1.00000 0.96402 0.15039 -0.27511 -0.01184 -0.38238 -0.38020 VS59 -0 14164 0 qfi4n7 VV~ qi~ . ~ ~ r - . rBEST CuPV AVAILABLE Annex Table 1.16: RECRESSION OF STERILIZATION AMONr EXCESS AND PAR FERTILITY WOMEN ON SMELCTED DEMOGRAPHIC AND SOCIO-ECONOMIC VARIABLES Sm . . * * * * e é * é * a . . . . . * e U L f~IT L:¯¯R~! i N 1 0 4 4 • 4 b * • * • • VARIASt Liey i OEPENDENT VARIABLE., Yi REIGR4(II0N LII7 8 IAhIsLEek ENTER9D ON sirP NUmeet iU j v V9 PARITY v*9 - - .....- V43 CA11T PCE C EXPEND V6 LAND OWNED va VAD Or SCHOOL iLmPLE R . ,21331.. . ... ANALM*II OF VARIANC 9 7 ..UN O UAREG.. . 4EAN I0UARE . 7 m 9QUARE .4491 REGRressoN 7. 19.944281 2.4917 12.49769 AJUBTEO R ISUANt fjeclo6 1t81DUAL .5, Ma3l2 * IANDAR <RR0R 8.41147 .29 VA4.ARL1 IN 1-c HmUA.TN •.......m•.O••mu. a**-.. VARIAGLet NOT IN 14f (UATION .•..••.... . VABIASLE 5 . BeTA SID 14404 1 7 VARSILt st1* 5N PA4TIAL TOLERANCE P Vag ..3963940E.i .29663 ka.nlI $.235 v9 ,50903515!.0 1.22560 #.82048 6.227 V19 _-_ ____ e.6349e.2 .. . •0.31 387 .Afilef IN :41 - . f63•.440373te2 -O.Gra9t 9.0443 96 PC£ 8 .77436t3c 8,808962 *.002 1 .363 VI ,51595891..,1 *.380 .0093 1,.348 (CONSTANTI 0.3000913 SUNAPy TAsth VAarAIL NULTIPLE a n gSUARi i@ CHANet SIMPLE A a BtTA N$$ l.0aga ,0036 0.0036 .0e' -.396394o.0t -0.298* N PARITT 8.9757 0.095a 8.00916 -..09661 .50903.5E-01 0.2256O vet *.äm 0.168 .007M -..J156 e.635942E2 M49.3831 43 C1EN*.I3304 4.1778 .00M1 .. 90.03999 4.4400373E02 -0.2241 PCer P.C.fxLND OP9 6 eeN §.02Ö6 0.00686 .045942 ..171436f.03 .0.0962 LAND ONNED -.0260 . 0.0153. -..0639 .9.170704-03 . .04759 V¥ yns o7 $CHOCK *.ll 0.4558 #.0k942 0.07155 .53595991-01 0.3820 (CONSTANT) . 0.3000913 va 9 v9 V43 PC9' V6 vI Va f0•.0966) •0.1.7 8..19992 .. .06 1.0000 ..966060333 .34444 .329 119 -.~ -. .0.11567 . 0.96616 - .00000~ -I*.@966 .0,11579 .04986 .e.30621 .0.28702 . 30 4.11578 t.10966 .0000 .0035 0.09682 .03032 -8.02565 -4.21003 .g519 8.00305 0000 4.25793 8.469 9,4874 4.03338 0,8986 .096 2 O.383 1.80008 0.8t3252 .23843 . oIl.34444 öl.30621 ft.0303 8.46289 4.23252 8.00000 9.93939 0.. ...9,. 0.2.,02 .0.02,.5 0...,6. ..,,,43 ..9,0,• -••• BEST COPY AVAILABLE 二 Annex Table 1. 18: REGRESSION OF CONVENTIONAL FAMILY PLANNING OR STERILIZATION AMONG EXCESS OR PAR FERTILITY WOMEN ON SELECTED DEMOGRAPHIC AND SOCIO-ECONOMIC VARIABLES . . . . . . . . . . . . . . . . . . . . . . . M U L T I P L E R E G R E S S 1 0 N . . . . . . . . . . . . . VARqAdLE LIST I REGREGSION LIST I DEPENDENT VARIABLE.. Y3 VARIABLE(S) ENTERED ON STEP NUM13ER I.. VS8 V9 PARITY VS9 V43 CASTE PCE P C EXPEND V6 LAND OWNED vs YRS OF SCHOOL MULTIPLE R 0.31132 ANALYSIS OF VARIANCE OF SUM OF SQUARES MEAN SQUARE F R SQUARE 0.09692 REGRESSION 7. 43.55817 6,22240 28.13383 ADJUSTED R SQUARE 0.09348 RESIDUAL 1835. 405.86234 0.22118 STANDARD ERROR 0.47030 ----------------- VARIABLES IN THE EQUATION ------------------ ------------- VARIABLES NOT IN THE EQUATION -------------- t VARIABLE BETA SID ERROR 8 F VARIABLE BETA IN PARTIAL TOLERANCE f VS8 -.2954307E-02 -0.21234 O OGO113 11.883 V9 .3377133E-01 0.04700 0.02009 2.825 VS9 -.4456404E-02 -0.21737 0.00178 6.292 V43 -.3955113E-02 -0.02033 0.00436 0.821 PCE -.322166JE-03 -0.03250 0.00026 1.580 V6 -.180520DE-03 -0.07923 0.00005 11.359 a 3:9 Va *0 321SE-0i 0-48353 0.00880, 62.966 4479- Lo DEPENDINT VAnIA§Lg.. V3 SUMMARY TABLE VARIABLE 14ULTIPLE R R SQUARE RSO CHANGE SIMPLE R BETA V58 0.21701 0.04709 0.04709 0.21701 -.2854307E-02 -0.21234 V9 PARITY 0.23518 0.05531 0.00822 -0.15702 .3371103E-01 0.14780 VS9 0.24321 0.05915 0.003a4 -0.06184 -.4456404E-02 -0.21737 V43 CASTE 0.24492 0.05999 0.00084 -0.04401 -.39651m-02 -0-02033 PCE P C EXPEND 0.24777 0.06139 0.0014f 0.07667 -.322166it-03 -0.03250 V6 LAND OWNER 0.25677 0.06593 0.00454 -0.03366 -.180620OL-03 .-0.07923 va YRS OF SCHOOL 0.31132 0.09692 0.03099 0.16878 -698321SE-01 O 48353 (CONSTANT) 0.3076479 V3 V9 VS9 V43 PCE V6 ve vse Y3 1-00000 -0.15702 -0.161154 -0.04401 0.07667 -0.03366 0.26878 0.21701 V9 -0,15702 1-0,0000 0.96606 0.11571 -0.21003 0-03338 -0 34444 -0.32910 VS9 -0-16184 0.96606 1.00000 0.10966 -0.18579 0.04986 -0 30621 -0.28702 W43 -0.04401 O.li57i 0.10966 1.00000 0.00305 0.09682 -0-03032 -0.02585 PCE 0.07667 -0.21003 -0.18579 0.00305 1.00000 0.25703 0.46289 0-411766 vo -0.03368 0.03338 0.04986 0.09682 0-25703 1.00000 0.23252 0.23843 va 0.26878 -0.34444 -0.30624 -0.03032 0.46209 0,23252 1.00000 0.93030 VS8 0-2170t -0.32910 -0.28702 -0.025115 0.48768 0,23843 0.93030 1-00000 BEST COPY AVAILABLE Annex Table 1.19: REGRESSION OF CONVENTIONAL FAMILY PLANNING OR STERILIZATION AMONG EXCESS OR PAR NET FERTILITY WOMEN ON SELECTED DEMOGRAPHIC AND SOCIO-ECONOMIC VARIABLES .. . * *****... * * * * * . .* MULTI PLE REGRESS ION ON* * ** R SINL REGRESSNJIJ'LiSr DEPENDENT VARIABLE.. Y3 VAZIABLE(S) ENTERED ON STEP NUMBER 1.. VSB V9 PARITY VS9 V43 CASTE PCE P C EXPEND V6 LAND OWNED V8 YRS OF SCHOOL MULTIPLE R 0.31293 ANALYSIS OF VARIANCE OF SUM OF SQUARES MEAN SQUARE F R SQUARE 0.09793 REGRESSION 7. 38.73432 5.53376 25.54222 ADJUSTED R SQUARE 0.09409 RE5IDUAL t647. 356.82501 0.21665 STANDARD ERROR 0.46546 ----------------- VARIABLES IN TOE EQUATION ------------------ ------------- VARIABLE6 NOT IN TtIE EQUATION ------------- VARIABLE B BETA STD ERROR B F VARIABLE BETA IN PARTIAL TOLERANCE F V5B -.2886717E-02 -0.22096 0.00085 11.524 . .2499980E-01 0.11087 0.02087 1.435 V41 -.3969366E-02 -0.19644 0.00184 4.641 V43 -- t798044E-0 -0.00927 0.00459 0. 154 PCf -.2587561E-03 -0.02650 0.00027 0.930 V4 -. 96O151E-03 -0.08715 0.00006 12.339 VS .6875325E-01 0.48126 0.00921 55.763 - (CPNSTANT) 0.4295529 0 DEPENDENT VARIAOLE.. SUMMARY TABLE VARIA1LE . MULTIPLE R R SQUARE RSO CHANGE SIMPLE R i BETA VSB 0.21247 0.04514 0.04514 0.21247 -,2886717E-02 -0.22096 V9 PARITY 0.23867 0.05696 0.01182 -0.17376 .2499980E-0i 0.11087 VS9 0.24519 0.06012 0.00316 -0.17691 -.3969366E-02 -0.19644 V43 CASTE 0.24613 0.06058 0.00046 -0.03931 -.1798044f-02 -0.00927 PCE P C EXPEND * 0.24841 0.06171 0.00113 0.08194- - 258758if-03 -0.02650 V6 LAND OWNED 0.25959 0.06739 0.00568 -0.03887 -, 196015-E-03 -0.08715 Va YRS OF SCHOOL 0.31293 0.09793 0.03054 0.26557 .607532BE-O1 0.45126 (CONSTANT) 0.4±395529 Y3 V9 VS9 V43 PCE V6 Ve V58 Y3 1.00000 -0.17376 -0.17691 -0.03931 0.08194 -0.03887 0.26557 0.21247 V9 -0.47376 1.00000 0.96588 0.11449 -0.20978 0.01505 -0.35804 -0.33591 V59 -0. 17691 0.96588 1.00000 0.10799 -0.18964 0.03262 -0.31919 -0.29360 V43 -0.0393 1 0.11449 0.10799 1.00000 0.00127 0.0854i -0.0434 1 -0.03098 PCE 0.08194 -0.20978 -0.18964 0.00127 1.00O00 0.26455 0.47369 0.4999t V6 -0.03887 0.01505 0.03262 0.08541 0.26455 1.00000 0.24092 0.24951 VS 0.26557 -0.35804 -0.319t9 -0.04341 0.47369 0.24092 1.00000 0.93027 VS8 0.21247 -0.33591 -0.29360 -0.03098 0.49991 0.24951 0.93027 1.00000 BEST COPY AVAILABLE Annex Table 1.20: REGRESSION OF EARLY OR LATE MARRIAGE ON SELECTED SOCIO-ECONOMIC VARIABLES s s * 4 * 4*4* *4* * * * ** MULTIPLE REGRESSION ** *. VARIABLE LIST 2 REGRESSION LIST I DEPENDENT VARIABLE.. Y2 VARIABLE(S) ENTERED ON STEP NUMBER 1.. V43 CASTE YM YEAR OF MARRIAGE V8 YRS OF SCHOOL V6 LAND OWNED PCE P C EXPEND MULTIPLE R 0.30389 ANALYSIS OF VARIANCE OF SUM OF SQUARES MEAN SQUARE F R SQUARE 0.09235 REGRESSION 5. 34.35455 6.87091 51.44078 ADJUSTED R SQUARE 0.09055 RESIDUAL 2528. 337.66321 0.13357 STANDARD ERROR 0.36547 ----------------- VARIABLES IN THE EQUATION ------------------ ------------- VARIABLES NOT IN THE EQUATION -------------- VARIABLE B BETA STD ERROR B F VARIABLE BETA IN PARTIAL TOLERANCE F V43 .7315931E-01 0.09033 0.01677 19.027 . YM .7936655E-02 0.20388 0.00082 93.3610 V8 .1360498E-01 0.12251 0.00274 24.615 V6 . -.6849785E-05 -0.00411 0.00003 0.043. PCE .7290557E-04 0.00966 0.00016 0.197 (CONSTANT) -0.4215111 DEPENDENT VARIABLE.. Y2 SUMMARY TABLE VARIABLE MULTIPLE R R SQUARE RSQ CHANGE SIMPLE R 8 BETA V43 CASTE 0.14106 0.01990 0.01990 0.14106 .7315931E-0O. 0.09033 YM YEAR OF MARRIAGE 0.28397 0.08064 0.06074 0.24821 .7936655E-02 0.20388 Ve YRS OF SCHOOL 0.30376 0.09227 0.01163 0.23254 .1360498E-01 0.12251 V6 LAND OWNED 0.30377 0.09228 0.00001 0.03587 -.6849785E-05 -0.00411 PCE P C EXPEND 0.30389 0.09235 0.00007 0.06834 .7290557E-04 0.00966 (CONSTANT) -0.4215111 Y2 YM Va V6 PCE V43 Y2 100000 0.24821 0.23254 0.03587 0.06834 0.14106 YM 0.24821 1.00000 0.35681 -0.00549 -0.05577 0.01258 Va 0.23254 *0.35681 1.00000 0.23611 0.41545 0.37908 V6 0.03587 -0.00549 0.23611 1.00000 0.25974 0.10692 PCE 0.06834 -0.05577 0.41545 0.25974 1.00OOO 0.22387 V43 0.14106 0.01258 0.37908 0.10692 0.22387 1.00000 I I I BEST Cori( "U jI~L Annex Table I. 21: REGRESSION OF AGE AT MARRIAGE ON SELECTED SOCIO-ECONOMIC VARIABLES * * * * * * *0 0 * * * * * * * MULTIPLE R E G R ES S I 0 N * * * * * * * * * ** VARIABLE LIST I 'REGRESSION LIST DEPENDENT VARIABLE.. V32 AGE AT MARIAGE I) VARIABLE(S) ENTERED ON STEP NUMBER 1.. V43 CASTE VM YEAR OF MARRIAGE VMS YEAR OF MARRIAGE SQUARED V8 VQS OF SCHOOL V6 LAND OWNED PCE P C EXPEND MULTIPLE R 0.49945 ANALYSIS OF VARIANCE DF SUM OF SQUARES MEAN SQUARE F R SQUARE 0.24945 REGRESSION 6. 1309493.86505 218248.97751 148.11867 ADJUSTED R SQUARE 0.24776 RESIDUAL 2674. 3940068.79123 1473.47374 STANDARD ERROR 38.38585 ----------------- VARABLES IN THE EQUATION ------------------ ------------ VARIABLES NOT IN THE EQUATION -------------- VARIABLE B BETA STO ERROR 8 F VARIABLE BETA IN PARTIAL TOLERANCE F I V43 13.21754 0.14207 1.69809 60.587 YM 6.477982 0.43581 0.71650 81.743 4 YMS -.3746716E-01 -1.05203 0.00566 43.772 VB 1.074090 0.08595 0.Z1586 15.160 V6 -.1976314E-02 -0.01045 0.00331 0.356 PCE .9177618E-01 0.10768 0.01656 30.696 (CONSTANT) -50.36676 DEPENDENT VARIABLE.. V32 AGE AT MARIAGE SUMMARY TABLE VARIABLE MULTIPLE R R SQUARE RSQ CHANGE SIMPLE R B BETA V43 CASTE 0.20905 0.04370 0.04370 0.20905 13.21754 0.14207 YM YEAR OF MARRIAGE 0.46518 0.21639' 0.17269 0.41994 6.477982 1.43581 VMS YEAR OF MARRIAGE SQUARED 0.47642 0.22698 0.01059 0.40603 -.3746716E-01 -S.05203 Ve YRS OF SCHOOL 0.49073 0.24082 0.01384 0.32043 1.074090 0.08595 V6 LAND OWNED 0.49075 0.24083 0.00002 0.05305 -.197634E-02 -0.01045 PCE P C-EXPEND 0.49945 0.24945 0.00862 0.15761 .917761BE-01 0.10768 (CONSTANT) -50.36676 V32 YM VMS V8 V6 PCE V43 V32 1.ooooO 0.41994 0.40603 0.32043 0.05305 0.15761 0.20905 YM 0.41994 1.00000 0.99440 0.36366 0.00629 -0.03677 0.02141 YMS 0.40603 0.99440 1.00000 0.36630 0.00813 -0.03383 0.01784 VS 0.32043 0.36366 0.36630 1.00000 0.23777 0.43987 0.37172 V6 0.05305 0.00629 0.00813 0.23777 1.00000 0.25063 0.10978 P(E 0.15761 -0.03677 -0.03383 0.43987 0.25063 1.00000 0.22481 ESTVROPABLE Annex Table I.22: REGRESSION OF AGE AT MARRIAGE ON SELECTED SOCIO-ECONOMIC VARIABLES (All women and with ym) * * * * * * , * * * * * * * * * * * * * * * * MULTIPLE R -E G R E S S I 0 N * * * * * * * * * * * * VARIABLE LIST I .REGRESSION LIST I DEPENDENT VARIABLE.. V32 AGE AT MARIAGE VARIABLE(S) ENTERED ON STEP NUMBER 1.. V43 CASTE YM YEAR OF MARRIAGE V8 YRS OF SCHOOL V6 LAND OWNED PCE P C EXPEND MULTIPLE R 0.48699 ANALYSIS OF VARIANCE OF SUM OF SQUARES MEAN SQUARE F R SQUARE 0.23716 REGRESSION 5. 1244997.49689 248999.49938 166.32859 ADJUSTED R SQUARE 0.23574 RESIDUAL 2675. 4004565.15939 1497.03370 STANDARD ERROR 38.69152 ----------------- VARIABLES IN THE EQUATION ------------------ ------------- VARIABLES NOT IN THE EQUATION -------------- VARIABLE B BETA STD ERROR B F VARIABLE BETA IN PARTIAL TOLERANCE F V43 13.84593 0.14882 1.70893 .65.644 YM 1.770446 0.39241 0.08485 435.388 Va 0.9740933 0.07795 0.27764 12.309 V6 . -.2104422E-02 -0.01113 0.00334 0.397 PCE .9126204E-01 0.10708 0.01670 29.876 (CONSTANT) 94.04788 DEPENDENT VARIABLE.. V32 AGE AT MARIAGE SUMMARY TABLE VARIABLE MULTIPLE R R SQUARE RSQ CHANGE SIMPLE R B BETA V43 CASTE 0.20905 0.04370 0.04370 0.20905 13.84593 0.14882 YM YEAR OF MARRIAGE 0.46518 0.21639 0.17269 0.41994 1.770446 0.39241 VB YRS OF SCHOOL 0.47815 0.22863 0.01224 0.32043 0.9740933 0.07795 V6 LAND OWNED 0.47817 0.22864 0.00001 0.05305 -.2104422E-02 -0.01113 PCE P C EXPEND 0.48699 0.23716 0.00852 0.15761 .9126204E-Ot 0.40708 (CONSTANT) 94.04788 V32 yM Vs V6 PCE V43 V32 1.00000 0.41994 0.32043 0.05305 0.15761 0.20905 YM 0.41994 1.00000 0.36366 0.00629 -0.03677 0.02141 Ve 0.32043 .0.36366 1.00000 0.23777 0.43987 0.37172 V6 0.05305 0.00629 0.23777 1.00000 0.25063 0.10978 PCE 0.15761 -0.03677 0.43987 0.25063 1.00000 0.22481 V43 0.20905 0.02141 0.37172 0.10978 0.22481 1.00000 I Ii CO NvNINOPB L Annex Table 1.23: REGRESSION OF AGE AT MARRIAGE ON SELECTED SOCIO-ECONOMIC VARIABLES (womnen 25+ years and y ) * see e * * * * * * * * * * * * * U L T I P E REGRESSION * * * * * * * * * * VARIABLE LIST I REGRESSION LIST DEPENDENT VARIABLE.. V32 AGE AT MARIAGE RGRES' NLS ) VARIABLE(S) ENTERED ON STEP NUMBER I.. V43 CASTE YM YEAR OF MARRIAGE YMS YEAR OF MARRIAGE SQUARED V8 YRS OF SCHOOL V6 LAND OWNED PCE P C EXPEND MULTIPLE R 0.59194 ANALYSIS OF VARIANCE OF SUM OF SQUARES MEAN SQUARE F R SQUARE 0.35039 REGRESSION 6. 1658115.94574 276352.65762 199.66652 ADJUSTED R SQUARE 0.34864 RESIDUAL 2221. 3074021.88019 1384.07109 STANDARD ERROR 37.20311 ----------------- VARIABLES IN THE EQUATION ------------------ ------------- VARIABLES NOT IN THE EQUATION -------------- VARIABLE B BETA STD ERROR B F VARIABLE BETA IN PARTIAL TOLERANCE F V43 12.10566 0.12594 1.79027 45.723 VM 2.813714 0.54195 0.76572 13.503 YMS -.7239638E-03 -0.01710 0.00627 0.013 Vs 0.6893508 0.05326 0.29385 5.503 V6 -.2142145E-02 -0.01099 0.00348 0.379 PCE .9582621E-0 0.11130 0.01707 31.531 (CONSTANT) 38.95649 DEPENDENT VARIABLE.. V32 AGE AT MARIAGE SUMMARY TABLE VARIABLE MULTIPLE R R SQUARE RSO CHANGE SIMPLE R B BETA V43 CASTE 0.20482 0.04195 0.04196 0.20482 12.10566 0.12594 YM YEAR OF MARRIAGE 0.57693 0.33284 0.29089 0.55166 2.813714 0.54195 YMS YEAR OF MARRIAGE SQUARED 0.57697 0.33290 0.00006 0.54870 -.7239638E-03 -0.01710 Va YRS OF SCHOOL 0.58409 0.34116 0.00826 0.34533 0.6893508 0.05326 V6 LAND OWNED 0.58410 0.34117 0.00001 0.05490 -.2142145E-02 -0.01099 PCE P C EXPEND 0.59194 0.35039 0.00922 0.15474 .9582621E-01 0.lt)30 (CONSTANT) 38.95649 V32 YM VMS V8 V6 PCE V43 V32 100000 0.55166 0.54870 0.34533 0.05490 0.15474 0.20482 YM 0.55166 1.00000 0.49322 0.37498 0.02643 -0.01144 0.06586 YMS 0.54870 0.99322 1.00000 0.38325 0.03096 -0.00643 0.06493 V8 0.34533 0.37498 0.38325 1.00000 0.23935 0.44816 0.38237 V6 0.05490 0.02643 0.03096 0.23935 1.00000 0.24101 0.09948 PCE 0.15474 -0.01144 -0.00643 0.44816 0.24101 1.00000 0.22485 A . *TBEST COPY AVAILABLE Annex Table 1.24: REGRESSION OF AGE AT MARRIAGE ON SELECTED SOCIO-ECONOMIC VARIABLES (Women 25+ years and ym) ea * M* * * * * * * * e * * * * * * * * * MULTIPLE REGRESSION * * * * * * * * * * * * * VARIABLE LIST I REGRESSION LIST I DEPENDENT VARIABLE.. V32 AGE AT MARIAGE VARIABLE(S) ENTERED ON STEP NUMBER 1.. V43 CASTE - YM YEAR OF MARRIAGE VS YRS OF SCHOOL V6 LAND OWNED PCE P C EXPEND MULTIPLE R 0.59194 ANALYSIS OF VARIANCE OF SUM OF SQUARES MEAN SQUARE F R SQUARE 0.35039 REGRESSION 5. 1658097.47998 331619.49600 239.70359 ADJUSTED R SQUARE 0.34893 RESIDUAL 2222. 3074040.34595 1383.45650 STANDARD ERROR 37.19485 ------------- VARIABLES IN THE EQUATION ------------------ ------------- VARIABLES NOT IN THE EQUATION -------------- VARIABLE B BETA STD ERROR B F VARIABLE BETA IN PARTIAL TOLERANCE F V43 12.11525 0.12604 1.78794 45.915 YM 2.726004 0.52505 0.09846 766.461 O Ve 0.6860305 0.05300 0.29238 5.506 V6 -.2149122E-02 -0.01103 0.00348 0.382 I PCE .9584156E-01 0.11131 0.01706 31.557 (CONSTANT) 41.56448 DEPENDENT VARIABLE.. V32 AGE AT MARIAGE SUMMARY TABLE VARIABLE MULTIPLE R R SQUARE RSQ CHANGE SIMPLE R a BETA V43 CASTE 0.20482 0.04195 0.04195 0.20482 12.11525 0.12604 YM YEAR OF MARRIAGE 0.57693 0.3a284 0.29089 0.55166 2.726004 0.52505 VS YRS OF SCHOOL 0.58408 0.34116 0.00831 0.34533 0.6860305 0.05300 V6 LAND OWNED 0.58409 0.34117 0.00001 0.05490 -.2149122E-02 -0.01103 PCE P C EXPEND 0.59194 0.35039 0.00923 0.15474 .9584156E-01 0.11131 (CONSTANT) 41.56448 V32 YM V8 V6 PCE V43 V32 1.00000 0.55166 0.34533 0.05490 0.15474 0.20482 YM 0.55166 1.00000 0.37498 0.02643 -0.01144 0.06586 VB 0.34533 0.37498 1.00000 0.23935 0.44816 0.38237 V6 0.05490 0.02643 0.23935 1.00000 0.24101 0.09948 PCE 0.15474 -0.01144 0.44816 0.24101 1.00000 0.22485 V43 0.20482 0.06586 0.38237 0.09948 0.22485 1.00000 Annex Table I. 25: REGRESSION OF AGE AT MARRIAGE ON SELECTED SOCIO-ECONOMIC VARIABLES (Age at marriage less than 25 years) a a * a N U L T I P L E RE G R E 3 3 I ON a * * * * * * * A VARIABLE LIST I REGRESSION LIST I EPENDENT VARIA8LE6, -V32 AGE AT MARRIAGE ARIABLE(s ENTERED O- aTEP NUMBER I.. PCE - PC EKPEND VO3 CASTE YM YEAR OF MARRIAGE YMS YEAR OF MARRIAGE SQUARED . . VA YRS OF SCHOOL V6 LAND OWNED ULTIPLE R 0.60q77 ArALYSIS UF VARIANCE OF SUM OF SQUARES MEAN SQUARE SQUARE 0.03012 REGRESSION 6, 1673830.76584 278971*79431 262077926 DJUSTED R SQUARE 0.34266 RESIDUAL 2528, 2683776.09427 1061,62029 TANDARD ERROR 32O54252 .-- -. . VARIARLES IN THE EQUATION - *****-* ***---- * ***** -** -** VARIABLES NOT IN THE EQUATION ************** ARIABLE - B BETA STD ERROR 8 F VARIABLE BETA IN PARTIAL TOLERANCE F CE .6812033Fe1 ,08340 0.01465 21,608 43 11,62319 0,13265 1,49652 6,323 M 16.72086 3,9b73 0.61593 136.987 MS aO.1191065 -3.57104 0.004860 595,189 - A S - 0.8477519 4,0?O754' 0.24479 11.993 6 -.1633212E.42 .0,00905 0,00295 0,306 co CONSTANT) .366.2398 EPENDENT VAGIABLE.. V32 AGE AT MARRIAGE SUMMARY TABLE ARIABLE MULTIPLE R R SQUARE R90 CHANGE SIMPLE R 8 BETA CE P C EXPEND 0.10734 0,01152 0.01152 0,10734 *6812033E.01 0*08340 43 CASTE 0,2049a 0.04200 0.03048 0.1941o 11.62319 0.13265 14 YEAR OF MARRTAGE 0,48785 0.23799 0.19599 0.44040 16072086 3.96873 MS YEAR OF t1ARRIAGE SQUARED 0'1741 0.38119 0014320 0.3975 R0.1191065 -3.57104 8 YRS OF SCH110i 01e1971 0,30404 0.00265 0.26590 0,8477519 0.070$4 '6 LAND ONED 0,61977 0138412 0.00007 0.03571 -,163322E.02 *0.0090 CJtISTANTI *36,2398 VARIABLE MEAM 9TANDAR6 DEV CASES V32 %J43 YN YMS Vp Vb PCE V32 217 4j.,1hA7 ?55 32 1.00000 n.10416 0,440400 0.39715 0,2859u 0.03571 0,10734 V43 (,41743 ?S35 43 0019416 9.0000O 9.01273 0.0n935 0,37070 0.106A4 0.2?350 YK Q.83,75 Ps t45 IH 0,44040 n.0Q1275* 1.11000o 0,994077 0.s%75 -0.00550 -0,05581 VMS 4 22. 441;9 12uA.t1a PWI MS 0,3975 4,0003S 0,99427 1,00000 0,35934 -0,00395 -0,05287 ve 4.ul)70 3.oSo V,s a u.2'59o m.3717o 0.35675 0.35934 1.00uoo 0.23611 o,1S48 V6 Ig,6S55 2?Q.74,,0 45 16 0.03571 M.10644 *000S5% -0.00395 0.23611 1.00000 0.25974 PCE 72.351 50.7h'-4 CE 0.10734 n,22315t -,05581 -U.o5211r 0,41548 u.25974 1.00000 tESTC8PY AVAILABLE Annex Table I. 26: REGRESSION OF AGE AT HARRAGE ON SELECTED SOCIO-ECONOMIC VARIABLES (marriages in the 1970's at ages below 25 years) * * * s i *~ . * ee *. * * *-a a 1 *** MU LIT IP Lit' R 8 t S 8 ON ** * * * * * a * .. VARIABLE LIST I REGRESSION LIST i DEPENDENT VARIAILE Via AU AT R IMARIAM VARIAILiS ENTERED-ON STEP NUMBER 10 PCE 0C EXPEND V43 CAST[ M tYEAR Of MARRIA2E ve VRS OF SCHOOL, V6 LAND OWNED MULTOPLE It 101638 ANALYSIS OF VARIANCE o SUN OF SQUARES MEAN SQUARE F 9 SQUARE 0;16010 REGRESSION S. 88044.56704 17608.913o1 ADJUSTED ft SGUARE' 0.1490 RESIDUAL! 664. 79tS681619§4 014.0431 STANDARD ERROR 30611315, *****f***m*mO** VARIARLES IN THE 10UATION *.**************** ********* VARIABLES NOT IN THE EQUATION *****...*.* VARIAjLti S aETA MID'ERROR P VARIABLE' BETA IN PARTIAL TOLERANCE F PCi; *807SEm.g g03014 0140611 0.633 . V4, 31P99I99 *,190s 1,31899 86.59i V" $.363953 IN1sts gs3a2*7 19,62 VS 1.365394 0e1401: 4036092 14,344 V6 .,9j%94679m#j *.00724. 0.0043* 0.043 (CDNSTANT) SIIeT 93 DEPENDENT VARIABL6 V31. A1E. AT MARIAfE: SUMMARY TABLE VARIABLE' MULTIPLE AR A SUARE RSG CHANGE SIMPLE A a BETA PCEI PC WIPEND 0.14360 0.01061 0.02041 0.14360 62077423FO 0.03014 V43 CASTE 0255t 0.046510 0.0444 023806 118.9893 0y1704 YM YEAR OIMARRTA6E 0:*63 4.0890S 0.01995 0.14286 1.363913 0*12913 vR VI OF SCHUOL 0.31631 0.10000 0.01900 0.25794 1,365344 0,14602 6ONTANTI LAND OWNED .31636 0000t 806004 800260 *,91194679*03 000724 17.7893 Via v43 VN vs V6 PC9 V32 1,0000 4,2380,6 0 1kt, 8 0aa 23794 0.07280 0.14360 V43 0.23806 1,0000 .0.00752 36559 0.13628 0.2?TT3 TH 0414288 *aool52 1.00000 0,04496 0.00651 00.0387 vs 0.23794 0.36559 0094946 1.00000 0,29938 0.47674 V6 0.07280 ,13628 0.0061 0029936 1.00000 0434811 CE' 0.14360 4.22773 0.0358? 0.41676 0.34811 1600000 .BEST COPY AVAILABLE k1 . . 4 Аппех ТаЫе I.27: REGRESSION OF ICNOWLEllGE OF I�'AMILY PLANNING METHODS ON ` SЕLГСТР.п SOCIO-ECONOMTC VARIABLES � ' 7 • , ,. 1• . ,'�в в е-� � ь•'а-*-s-r-r t-sтr-в-�r--а-r-тв-ж-м и<'6, Т I л L Е д Е 0 д Е i• I 0 Ч в в в в•• r• в• в•• V�kiд�LE l1ST 1 �� � нEGkESai�н LIST �' � „�Е►ЕнD[нт--rввI�Е .L-C ;---xF -�----�.-____. -...�_.._.. .. . _ _ . _ � .,�ддIА8lЕ1оУЕМТСА[0°(2Ч-в1Е�-иUм0[д-1 .�----11р]- сй3ТС �--- - •-- - . .. _. .. � ' --- - - . . _ . ..--- --- -- --- - ----- - - - - ' 1 . „ мuETiPIE R о,г�лге ANAEY9I9 uF vАдlАисЕ oi 9им oi sou►нES иЕАи ;;nuAUE Е ' ,.'R ]OUAR е�л3Т ECRESSIO -6; ----z2.2Q7o1 ._. ..° . 2,7872а 2о.2tдеу ,„ воJи�тЕо А sои.дЕ о,а�о2ь дaslouA` zseo. г�а.оол7� o,.ovsзs � � ..-ЕТ,►ноАвD EaROи ---o.sfle7o---- -----------•-�-----._..- - .. - . м ' � „ ...............«.. УАд1АдLЕ! IN ТНЕ EoU►TIOи .'.........'•••'•в - ......."...r R А И .••�.'... •' . . --- - vA I BLkE Dr Iи ТиЕ Eau�tluN -. о р, VA4IвBLE _ В_ ВЕТв- gT0 ERROR !-'- . F' YApIABIE ВЕТА IN PApT1AL TULERAк[Е i , .. у• .С25Вгц5Е�лl р.аа7а5 • о.ааУо1 75.г1; . р ,.-yi- � �аv�7оЧеЕаnl---о.37ог0 -O.o07ss---- д2.:г5� ,- Мgв .�ЪезЬ33iс.п2 �о.г5391 о.ооUл1 2o,2V5 „'УS0--�.явоо355�:пг--.о:г2оS+� о'.оорвУ i7.112---------°---°------ • -� . � . РСЕ .г)16126�.n] о.о5ц51 o.00ot1 а.3л1 ' •„-v1 .5а]д005F-л2-р.огьеУ -o�ooYS1--- '�YSI . „ Vi .,65899у��.,л5 �О.Ооа1! О.оа003 О.ооЬ _ w л„ ыАъ �.3ot2fгл�•n2 о.оо�►е7��- --- o.olaus о.оsь . tг�ЧаtАит; о�ьоtоаьl �' о . ----------- -____._ _ . ..- -� - - 0 . • • � . .. 1 . ., "- SUNMARY ТА®lE - - - „�ARIABLE----- - " `° --мU�TIPEE д- д sаиАиЕ R9o СнАисЕ sI„PLE д в Giтл , ,��D-'-Уд3'OF"-SCMO ое2271а'---'9.а51ь2' "' р.о51Ьг- --"' 0.227fa - - .Ч?5aZn5E-O1 - C.co1-S „ У4 рвдlТУ 0,257vQ о,олб5! о.оl4vD а.0у19ц .Ч917оЧ6Е-01 е.?1'сС , ..-иsо---- ---- - - о.г7зь2 -� о.а7иат а.ооаз5 а.ооул7 •.злзСзт.F-ог -_>�� , VSd 0.2а349 а.о80л5 O.0U578. 0.1Ь2Sл ..1ApoS5iE�a2 -�.�.'�- ,. РСЕ"�- --°Р с Е1сР[нD - -- р.2В72л о,авг5г O.DOtB7 0.11793 .271н�г5с'.а; U,;;.., , .. V1 TDILET FACBLTTT 0.2еВ1ч O.oH3aS а.оаа4+� •О.ооb7ц .5ц]и0с5Е�02 �.�:-�ц ,.-9b�'[aND-D'лЧ �Z8825-O.oE30V -`-' a.Da0a3 -�-' 0.а5г78 - � .�85дЧаи1F-05 -С,:С.1; ' ,. У41 сАуТЕ О.гОдгВ o.otl3,1 о.ооо02 0•11217 .3ц1212дЕ-02 а.;.�:.с7 � , .. ЕСJИ!lТАИГА- �----� -•--- --•- o.o0f0u61 ., i1 1 4 � �. � � - $ 6�ST G�P� A�1tA��� � � � ' , ' � q . ` �. ч � . ._ . . .- . . . ' . . 1 - r_ I ''�'i" Аппех ТаЫе Т.27, contlnued: �, ��• � - lсиог,�с,Елб.С oF FАг-l�►-Ч f�-iР�.,и �иG 02/20/it PAGE во � 1 • - - - - -- - - - - - � _ _ . . . ' . ---- ----- , : г1�Е 4Е9��вн� fCдE�ri�N о4ТЕ • 01/3l/е9) KENw6.A Fj�E NITK INCONE ' ,1 � � l+3r1LE PьLGy1r! цЕоРЕУ1 ЕЧиьк1 -- � - -- - - --- - - - - - -- -- -- - ................•в...•. миlТ 1rEE ЧЕGиЕааlом •в-ь•••-••••a.r-• 4Уьаtьеl.Еlаlr 1-- с ' R[GaESS[с]и l,1st 1 , DE�EvOF•п v�яlвэ�l.Е'..---кF►--- --------- ---- -- ---, ° --- - ---- � v�a1iBLE(51 ENTEREO ОИ •Т[Р иииаЕЧ- 9,,; -- Уь---- -- - LANp UЧЧЕО ' • - - --- --- ---- - • .- . - - .- _ _ -- 1 ' +JLт1�_Е �+ о.2•т2в •иalraiS Of У�R1ьЧСЕ - 9F bUM О6 еои►ЧЕе ►+Еьн onUвaE f � е s,a,�E --- -о.о,аее аесаеаslом---�--,--------�;----- ао.ьг)sv - --x�z9lvg ---------г3,длхла , •�3us1E3 а даи�аЕ о.а�lьь aEaloUвL тsTV� . 2°Т.ьtл)3 . --- а.о9ла4 - -- - ' Sгav�e�� Еад�д о.3л990 '- - -- - ----..-.. -. .- --- - - -- - - • ' - - --- - - - - • -- � - - - - - - ' ...------_--_--_-�-- Увд�в_ц_Es 1и гиЕ Ег►и�тгUи •---�--••---•--^�- --.......... v�R1�бLE� wрт 1и тнЕ Еои►1lии .............. � � Y1i1lBlE 6 EEf• аг0 ЕЧRОЧ ® i VAeIA®l[ ,ЕТА 1Ч iвЧТТвЕ ТОLЕвьЧСЕ f � V! .4j19025E� » о�й63l0 О.о049) 72.2о0 � � Е( .1�f1l7а2F-л1 о.1о351 - 0.004ог 19 е52 � - - -.. - -_ . . _. . - . vsi ..2�o5so��_,2 .omta5)9 о.ооо�д тt.оье , ' г:Е---- . .2)лтs3тЕ_,I-- "O.o5S5D---б.oo011--.b.351-r---------•----------- - - ---------•-- ----- � ,г3 ,итzтьгоЕ_лt О.огsао о.ро453 1.ое7 ., vzS ,Т�рSьо2с_п2 р,о1059 0�0l465 р.25Т �з`s -.zsoyls н _n3 •о.4тtа3 о.оооое е.9о3 � • л Уrз1 .�)�+�гьт�-лг о.4т5лз o.oosot в.�3! � , i Уь -,S1o255tE.�5 �D.oo221 О,р0ое! о�012 - �С�У�Т�'�Т) а.�2ь)39t ------ ---- --- -----------�.-_.._._.---- ------------ ---- ---------- ,,, _ r � . ' ►� . А ..------- -•----- -- ---- --- SUMN4aY ТлаLЕ , YrllaeLE -' - -�--- мUET1PLE в R SoUвRE м5о GNANGf aIMPIE Ч • еЕТ� ..-Ye�-TRS-Or7CM�)� 0;22719---p;oSlAd'--е.051л2 --о.г2719 --- .421oo25F•01 р,�д11� „ Ех с-.�г-�� �^ 0.25433 о�Оьбbе о.о13о) о.Оьлlе �1'о1119гЕ-о1 C.iSi�i . .. Уае- ---�.{i д-� --О.гь)t2 - o,ofl35 О.ооьь) р.le2la •.2ooa5olF»о2 -с,2.5гv ,� - _ гсЕ • С ExPfиD 0.27о5а о.о)згt о.ооlеь o.ll7a3 .2)оТг])E-oS о.с5��о � „ Уз- rа1ЕЕг-iьегЕтТУ------- _ ---- р,2т1l9 - -о.о7з55 о.ооо33 �о.о9ьг4 _ .4rгтozef-а2 о.егs»3 ,. УС! С�9ТЕ 0�2)!оь О,р)3ьV о.0001а O.1121) �Т�г5оП2f-о2 O,G1;�9 �! „'ьGts--t?�:-�'� D�2)j59---0,0)375--'О.Ооб06-�О.о0391 -�e2S0�TSlF•OS -G,.:7?»3 ;� „ vT11 ьtЕ !и УЕ3а! - -- - 0.2Т)27 р,о)ьее о�оа112 о.оо5аь ,1au7:drk•о1 с,��GэТ „-Vь------- t�HD ОынЕ0 - --0.27)2в р�о'lьдв 0.0о000 0•OS27e ..3102S51E.05 -о.есгеl „ iC�NlТ�Иг1 о.42ь1391 � • --- - ------ _- . --- - ----.._..---� . . . 1 " Уз vь ve уае У9 Уа9 Ех vr3! лсЕS У43 vcE �F� - - -- - - �- -- ---- -- - - ---- - � У3 i.ooooo :л.го4sо -о�44�оВ -о.4sа3в o.o5zte о.о5о�5 о.о114г -o.ozd93 .o.o24as -o.too9c -о.Ч1i�д -О.��л7� - Vь �-•-�•-•-•p.2o45or��1,pa0no - а�24zи2 'р.г51�ь 0.оа34) 0.02гi6 �а.о?96г 0.ро1д4 р.Олбае 0.129с4 o.IFг.C9 о,;=�7о „ VO .р.44)Dд n,гаге2 1�0о000 0.9]2в1 �O�S2169 •0,3ор94 +0.2рат4 -Ое25)е! .0�26ро1 0,166е6 О.s���Ь �.с211� i „-rSб - •o.os)зe-�;zstть-pтoзzet---t:ooooo-•O;sost9�-�-o�г)tio -0.1де5о .о.2гь)е .o�гsilr о,iг5�3 o,sa�so �,:-�!о „ У9 о.05220 � л,р0347 .О,Ъ23b0 -0.7о519 1-.рОрао р.94l3В О.до128 р.ЬIУЧ9 0.61дц4 .о.о5д30 -O,ffe:o 0,��;�ь 1. • У59---- --- о.о5о15 - л.о227ь °о�3оо99 -о,2)71Ь О.о413В l.Ооо00 А.Т32Ь3 р.5оо05 o.S3d5T _О.оауь3 '4.i1SCг �.�г4л! � • Ех о.оllаг .п.оz9ьа -р.2оо)4 -р.leeso о.еаl2е о.азtль t.ooooo O.ssгd4 о.sзltu o.ooo9s -o,cs;e[ о,:ео�а �Yrs1- --о.отл9з - п�ро)еа --o,tsie! --о.22ь�д О.ь1449 р.5�оо5 о,552ец l.oooop 0.9аг54 о.ов3гs ь,ь��з� с,�с:•о � ; iбЕ! -о.о2ци5 �л.ооо90 -о.гьооl -о.23ьf4 О.ьtеа4 p.5jd5) o.5z1t4 о.99г5ч 1,опооо о,огьо) р,са��� c..c��l .-VaS--rp.t9o5и-n;tг944 --о.3ьдаь -о;згs)з---О�о5д3а- -o.oasaf o.oooQS о.ов323 о.оаыоs 1.ооаао с,гз��5 »�,ttzi[ , iCE .D.4114е л,►ее49 о�3)7дВ 0•39д7ь •0.11ь0• •0.113о2 •0�о5рь7 О.Оег7) O.otloeo 0.23175 1.OJ;:•^• O.t1)�+3 . -' - - � , - --.. .. - ...... . .. . .» . �.... , ..,,. п.. ..n „�э� п . � �.? • ..•.,t 1 - � � - �Е�Т С�РУ �VAiLA�LE i - - . 5i . Аnпеи Таи1.е I,28e �tEG1i�SSTON 4F' NUMiiER OF B1RT1l Сс)NTRQi, bП?т�nпs ипУ�т.т_t flлт SELECTED SOCIO-ECONOMIC VARIARI.ES � г � ,,.,• �- е- е�• s s- о �-,r- а-r�r-y-Q-*-t-л--ж--.�-r-ц-М-R-ь---р- U� Т � У L Е N Е G R Е i• 1 U и •• ь• л•••••• а• ' V � ц 1 �"г.,. Е t 1 5 i 1 ' • PEOpE5S1�� L:ST 1 ..�L►[ЧРСИГУдR1'А01:[в: �гТ'o------KNOk-t' , ---- ---- ---- ---- . - ..- .:-vAp1Ae1.EtlI �мтЕиЕ®-ом-,ТЕ►-кине[e--vвe УЬ------ `АИр ониер . �. MuL71PEF в Ове5799 ANALYSI9 Of VARIANCE DF !Uи Of 50USpE! ИE�ii SOU►RE f ;..я-lоидд , п97. и[cяts9loH ---- 9в --- --- 14зI•оаоьо - es9esfnsl 7�,�с:�у �.. AOJUlTED R SOUARE ов2лТ0о pE�101lА� 25Т9в 540гв0229р 2eo94b2 ,. l1аиDАдр-ЕдАор 1.аа72е ----- --- ---_. _ . � _.._------------ ------- -- . . - �, .....°•°.°°•°°.°° YAN1AдLE! 1Ч ТИЕ EOUATION .......•а......'.° ......°°..•.. VAA1A8lE! ЧОТ 1Ч ТиЕ EOW Тt0н _..-.--...°°_- . „ _--___ -. _ _ .. ыАи1+еЕЕ-_ ` - в вЕтА ^ 6ro Еддор s � Увр1АВlЕ вЕТА ЕЧ РАЧтt.<_ тUI.EвRиCf i � Ув ' о.2lьлло2- oea7t19 о.о�31о -- �7вгое п-Ех-'- -- ,]534217E.91-- р�ои3]1 ------Ов0194) --- 4во72 • . У3 °.�07035оЕоп1 .о�оц5Т9 D•0?117 4в8Ь5 ,-Y�I-"-`'do94b72�2 o�o41N7---4:OetlYI 4eS4T ----•-- ------ . . .. _ - . - , PSE ,129vu14E•n2 Oeo5t0i о.ооо51 ь•цfе - ..-Уа! -•--• .во17ЬвЬVЕ°nt--•ов1о31Т---овоогоа •- ц�1еь „ АбЕS ..2610l41Е.л2 +ов974ьа Овбо479 Оц,26( 1 ., Уrз'. _ .. овlе2ьл77 - -- oe9e2qt---___-о.огаьг � 4зв51г - - vь .►обАьlьЕ.n1 o.ot526 о.оооl3 о.7оо ,- Н .-tt�ивТ"Амrц :awoolos _-__-,^---- - ' . (v . ( ,.- --------- ----•----- - - lимивдr Т�мLЕ • „ YAR1101C----. °-------- ---- -_--------- Ми�Т1Р�Е Ч р 90UвFE р�0 СЧАЧОС �iM�LE R А ьtТА . ;��►l�-Трs-nг--ьсми - овц2гог---ов1ег39 ---о•t6zзv - ов42ТоТ ов2lеььо2 с.::т:►Q . Еи 4.ц]39ь o.teal2 o.oo5v3 -ово3о3о .laз�2f7E.of с.г_Т�т ..'ыI - --TOILET 1•ACTL?Ту-- � - ---овц1752 ' 'о.191а2 0.4о310 °оогЧЬ7о `вЧ67а15СЕ°96 °(.�..ti7c ,� Vц) ев�ТЕ Овцц032 о.►97оВ о•о02ц6 ов21о9о o,14e72t2 C,t_t�1 . ,:. РСЕ�----�--Р G fx�ENO�------�,- --- ----0.44161 ' 0,195го • 0.04111 0.21а75 .- ы50 ,1290�17f.o2 G,;S�o; ,.-лсЕ Оварге9 0.1959a о.ооо70 ов39'Ъз °�417иеечЕ-оt :.:;:sт ;чагvl o�tчat4 ---овооотг - -овоеллц -- - • � ,. ыril - всЕ 1►1 УЕАяf о р57Ть •в261n3изЕ•о2 _, 7_,� „-ыь- - -- - EAND ОинЕ0 -------•- _ _- � o.2oosa 4•ое3зs -o.oso29 овlаг3оТ7 c,s;:.z н гС7И�ТАмТ1 --о�ц5799 - ов2о976 о�040г1 Овl2В2ц а10ЕСе1ЬЕ.03 (�.С15ьо . .. ------ -- ----------� -------------- - � •0•90о)оЪl -, -____------ -- ---- , ------�- - - - . , -- уа --- --•vь ц!- YS6 ---Т У9 - У39 Ех VY31 AGE) V41 РСЕ v1� � „ У3 teoo4oo .дв2оц5о то,аатоб .o.as7le oeoszza oeosofs о.оilаг •о•аг8v3 •ов42Ч4s •ов3ое54 -о.Ч11Чв .а.r�г�� ,. vь ------.о•2obso •��ле'вооооо .-oe2azA2-.. овzуlзе -- о,оо3ц7 о.о2zТь °о.о29ь2 о.оо7ер о.ооецв о.t2лЧ4 о.1еЕ�а о.(т: �- „ V! °4,ЧЧ)о® n�242Aj 1р0о0оо 4вv12в1 °ОвЪ23е9 •ов�о09а °0.20о7р °0в25761 •о.2ьол3 oei68P6 0.377�G �.-27'7 ,.-V30-.o;u41lп--rt;23176"--Я',e326t-1;oo0oo-�oe30S19 -.oe2T71ь •o.10s5D -ов22е79 •0в23(l4 0в32517 0,3оьiб 4,��:+� ,. vv о.о522е •лвоод4Т .о�,lгзь9 .o�3a514 ieooooo овчаt3в о.еоlге овл3ц4v овеlа4и •овоsы 4-о,1lог.о -�,:5i�� „ ыs+-- - - - o.osots--��л.отzТь -`ов)ооУц-- -о.:тУ1ь -- а•9a1je 1•ооооо o.7lts3 o.saoo5 o.s3h57 •oeoFbь3 -в.lsзле -�.� « �t • .- ЕХ ' Ово11А2 :A.029bj •овг00Тр •Ов16650 ов0012• 0.7)2Ь1 1воо000 Ов552Вр Ов5]124 os40o9S •О.б5ог+7 -�.л1С:+ �. ,. ыr)1 ---- -- •o.o2evt п■оо7ар- •о.25ТеЕ -•о.tгы • о•ь)ца9 o.saoos o.55znц 1.ооооо о.9ч2s4 о•ов3zз о.од�i7 -ь._5ст� ,. AtE! .о.о2а45 л�оолае •ов2ьооl •o.21t1a oebtayц o.s)в5j о�531zд о•9огs4 1,ооооо о.отеоl о,сЕоеа -с,�::.-. ,.-ыаf .овlчьs•-п;t�9а r-о;эеееь-о;�2s71--о.о5е3о -•o,oesel овоооч5 o.oeszз лвоТеоу (�ооооо о.�3(.7s и,�1с�.о о ogE .1ое411о6 л.,1еб4У 0„з77Аб 0•]9l3ь •о.llеоь •о•1l30i •овоSОь7 овоА21] о•Рв66о 0.2I17S 1•G4on4 й.г1�7� J • r � � ' � 6 ���� ���� r7i�������� Annex Table 1.28, continued: NU48ER OF METHODS KNOWN REGRESSION 02/20/61 PAGE 17 FILE KERALAMG (CREATInN DATE a 01/31/8t) KERALA FILE WITH INCOME ,-SUBFILC-- PALGHATI L9PPlYl - ERNAKI -L;--t- R E 0 R t 3 3 1 0 N il lb ft VARIABLE LIST I ::-Or 'END-Elf"A* Mt L VIC KNO , -r- -REGRESSIGN LIST 2 ':-yAJtI ABLE 40-ENTCREP-O", lf"UMBER--4,-,----V*-LAND-OWNED--- 4ULTIPLE R 0045754 ANALY31,1 OF VARIANCE Of DUN OF SQUARED MEAN SQUARE F 0. n"4 - -REGRESSION ..ADJUSTED A SQUARE 0.24689 RESIDUAL 2560. 5404004244 2.09490 ,.-ST&NDARO-ERROR 1%4q?38 - ----------------- WARIARLES IN THE EQUATION .................. .......... VARIABLE& NOT IN THE CGUATIUN -------------- VARIABLE BETA ERROR 0 Ir VARIABLE ALTA IN PARTIAL TOLERANCE 0.2123040 0,46172 0402301 85,165 w9- - 0.250642s: - -- 585?B.- 0.03543- -- 53,163 ..230037bL.4l .0 3 62 ": 1 2 0.00378 36:966 oio44?? 0-io0osi-- 4.976- .Q59sofisE.aI -0,04503 0,021ts 4171? 0.03539---4.06772-1,415 V36 ..26?7690E.42 -0.04371 0,00204 1,659 0.746 (COM3TANTJ 1,533 0 VARIABLE - MUL-TIPLE--R---R- 3QUARE-'R3Q GIMPLE R LIFT& we FR3-3F--3"0OL 0.42?01 o-SO239 oila239- 0642107 ve PARITY 0043640 -:19044 0.00805 -0.05333 0125664?5 0,16578 .0.08611 -.2300376E.01 -0.11622 .PCE P C WEND 0,45394 0.20606 0800304 002147s s1126d2IF-02 0.0407 .-V3- -TOILET FACI"Ty-- -005547- 0,20745-- 0,007119 - -002'1670 -.4s9306SE-01 -0.r4S03 va3 CASTE 0.4sby? 0.20864 0,00119 o.21o9h Q.t2519a2 o.cis39 V`38 ----a 2627890E-02 -0.0 6 3 71 V6 LAND OWNED 0,4S754 0:20934 0*00023 GoJ2824 :312606$E-03 o9ols" -CIC3NITANT) 1.563807 BEST COPY AVAILABLi 韭 一 l Annex Table T.30, continued: .3TERIC13ATION REGRESSION PAGE 15 FILE KERALANG (CRgATInN DATE a 01#431/51) XERALA FILL NITH INCOME -SUIFILC-- PALGRATI-ALEPPEYk-ERNAKI P-L--9'-R E 6 R 9 3 3 1 0 N VARIkAtE LIST I REGRESSION LIST 2 -'kTEP-NUmblw-t--#---V3-TOILEV-FAC ILI I MULTIPLE It 0.3185) ANALYSIS OF VARIANCE DF BUM OF 3QUAA13 MEAN SGUARE ..-"QUARE __ 0;,smo __ REGRE33ION----7-,-- 56,45602 ADJU3TED R SQUARE O-IiJ94 8,06ris - - RESIDUAL 2326. .-STANDARD ERROR-*! ------------ ---- WANIARLE3 IN THE EQUATION ........ w ......... ------------- VARIABLES NOT IN THE EQUATION .............. :,VARIABLE BETA SID ERROR 5 F VARIABLE BETA IN PARTIAL TOLERANCE F V9 0.1715000 0.89731 0.01130 234,231 VAS 0,40497 0600014 0,63166 0.002 ,.-V3t- ..I56b290E-4I-w-0.?5577 00121- 1670915 ..13a7572F.13 .0.06594 0.00004 10,34 2522614E.q2 .0.247st 0.000&3 20,119 V3 0.1246933E.4i .0.44181 0.006sl 3,597 -(CONSTANT) 6822394r--iZ ---- F-LIVEL OR TOLERANCE-LEVEL INSUFFICIENT FOR FURTHER COMPUTATION 4 -SUMMARY --TA ::-VARXABL- ULTIPLE' R-R'*30UA14E- 880 CHAW7,E-"- SIMPLE R BETA - ---------- - 19 YAR. 7305- Go litsooe ..V39 0,31470 0,09964 0906910 4407519 0.1566290E-01 -0.75617 N D-OWN E 0 go32068- 000214- 0#00380 .013275?ZE.03 -0.0.594 we VR3 OF 3CHOOL 0032304 0,10435 0800152 -0604561 .1060302E-01 - 13527-----o -1124it-- 0,00005-- 00,05274 a,28Z21i341.02 PCE P C EXPEND 0:336.So o"tain 0,00083 00.06144 W*3174627f.al -0.04107 (CONSTANT) Doi IT-&a i 00649"-a. 124 693 Sr.* I BEST COPY AVAILABL Аппех ТаЪ1е I.31: REGRESSION OF USE OF CONVI;NTIONAL FAMILY PLANNING OR STГRILIZATION ON S);L]GCT.ED S4CI0-EC�°IOT4IC ��ARIART,т"S 3 7 � '� �•-r°в-в в-в� ••-в-r-r-в-в-�-rв-a�r-д-�rw-r-x-1f-U��-T-�P l`Е• R f С R Е о i I 0 И t r в в в ь в r в• s е о YAHJt9lE 415Т 9 G - ..�С�ЕиD[иТ-1гАRТееС .е- � -sтtя - -•-----••----._.._ _...- ---.. . yE�Rl:ssluй �FSa i , УАд aelC(е1 ЕЧТ R!0-СИ°1Т ►-ЧUнеЕ . - --• .� • G 2 Е Е R .-ч-: РСЕ--Р-�°(xPEND--• - ; ни`Та►ЕЕ д о,а?о2е АИАЕУSа3 oF vАдlдиСЕ oF sин oi s�ивдЕе неви snu,ыE ' � �' ,.�я-]�:злА .Г►ьь EcRES9a0и �•--�9,-°--'- - 1а3.а37оВ - =1�и59ь8 :��,:;s11 ., `OJUSTEp R SoU�RE о.1�3ц5 REBIOUA� 2326. ре0е70ььl о�гоьь9 � ,. 0ТАирАар ЕдRОд�-�-�о�а5аь3 �- --- ---- -• - -- - � - • --�--- - - •-- � -------•-•-- --------•-- - - - - - ,. - :: ------•---•-•-•.. и�яliиLEe Ун ТиЕ ЕоивТа0и ...-.•.•......•... .••-••--.•... vAдjAULES ноТ !и ТнЕ Еои�таии ..•------....• � .. Yвat►ew.E а _ еет► _втр ениод е _ г • vАисдаЕЕ еЕтА iи РА9таА1. ти�f.д•исs �Р � ;. Ех ,sе�эь5ие-пt о.г uv7 o.ooeas ез.ра5 ,.ve -- �ььи9l�л�_�•t--о.аТзат о.оа7ьз-----7s.иТ9.. _.__ ._ . -: V58 o.SьlElov7Eanг •о.29о7Ь О.о00ьЬ 29,et2 � ,.�►сз �.bao33�1E-�t О.оs7г9 -O:oгtSS -т�sез-----.._. _ . . _.. . . . •,.vь .,15ь7о3аЕ-�3 .о.отlа9 о�ооаоц а3.153 _ vr�1 -...... с.ьогsаsо - в,7ьуое ---� о,оочlе iг3.91ь С :. AGE� , •оlагlЬ55Е+ег •t.b9�92 0в00013 �16.1о5 . .� v3 • - .,1иь]ьone-ni -- -о,еи7г3•---•-- • о.0оь92 ��ц7Т . PGE ,ilи9едлЕ.л; О.оау53 O.o00t7 о.ц3; � 1 „-rсзя'тr�tгт-'=гтьттТ ---------__----�--- -- --- - . w - - - - � - ^ ---^ - - � Н ' „ �сгСnысл. rA�6-оь.ьвв •�с,вг♦ �'--- --- - OD 1 ,. 9l'MHANY-TABLE ----�----�- - -- ;;-+гRAtAet • иЕтlРЕЕ-•R--и-аGиьиЕ-- ддо� снАисЕ--�siигЕЕ д -• е 5En ..У �iuT6q---o�o6t33--p�o6115 ов2р7ьр -----�5д89о5цС-01 -°- o.211V7 -- „ Уе Yas oF аснооЕ о.3зьsь о.11зzТ o.oss95 ;..о.tьаео .ьои9ааеЕ•оа o.c7St7 �^•Y33 О,Уе�зр---0�1213а •-- О,оо8оь г��•- о.125Эв . . -,3ь1обv7Е-о2 -о.а�о7е , У03 CAStE 0,35з]3 0.1гь12 о.0оцТе�.° о�l52т5 .ьго33тtf,оl o.osтro „-vb ��н0-ОамЕС 0.35w6---0.12959 -----0.o03e7� - -o.Of2o8 - . ь9 i .15ь7оз�Е.03 . -G.071 „ УУ3{ AGE Уи УЕАдS 0.3ьn1Ч о�(12ь0 0.003р1•� �0.150р5 о.1о21Ч5о 1.�о_.s • � ..�СЕ . Y1ti-a;1-14ц9-o;oalvp-�--0.92STЬ----•�1ц21о55Е•ог--�--� - •1.ьагSг � „ ч3 TOtIEт PдCiLtTY .О.ц2о1д 0.170р8 O.OOlQ9� `0.132о9 •.1аь3ооьЕ•о1 •O.C�7t3 � -РСЕ- -1 t-ExPENp - р�р202б------0®1iбь3 - 0.00о15•�- 0•оо17о .11u9dsoE-oS о.оао5т 4 � tCJNSTANT) -- •S.3e?572 , _-'"_ __--._"�"'__- _""-_ '_"'_._ . . ! � � , � �� _ У3 _-_ де _ ч• Уее У9 У99 Ех чУ31 вGES УУ3 РС'с 5т'c�F.� i ., У3 1.ооо00 •n�2l396 -ооцигll `о.р5lь2 O,ub377 о�оь031 о,0lц7е •о.03926 •0,O355t •о.lа2о1 •о.�19�о -O.f?:аа �- ,.-уь - - .о.гl3ее - 1.оооод - o.2ssa9--д.гьцs9- .р.оlыzо o.oosoo -o.oзvlv о.оов7ь о.оо7гз o.137z9 о.lаос5 -�.��: r: � :, че -о.ригll •�.гss19 t.ooaoo о.оаго7 -o.3saьt -o.sa9le -о.гlьs5 -о.г5цар .д.гss5а о.371зь o,sл.lz о.:•�-� .�-r3e---ro;a5aeг-л,20q59-9fl-9329t--y;p00oo---O�a33oaV- .д.2о8б! `0.2о15В -0.2217D -0.2:Sa2 0.32567 O.�1G2u й.:-�ia .� .. Уо ое06377 ,;а.дlи2б -о,35цb1 `0.33оца 1•.ООООо о.942дь о.8037о 0.65ц]Ь О.е39о9 •0.07170 •O.11o91 о.:2531 „-rs�- о.оьОза--'--'г.oosoo---o,згoae-�•o�zveet о,9агце t.oaooo о�73sьг о.sьоы о,5ьо+�1 •о.о9ьhь -o.1z7�5 0,�:��t ,. Ех о.оараь -п.о39t� .o,aless ••o,totse о.ео3То, о.73sы а.ооооо о.5ь775 о.sиьzь оо.оот«3 -O.uoz�c �.�-��- ,.�УУ3�-- .о.03егb -.,n,0o07b�-•ор254а9 .p.г2f1D О.еSать д�50о62 q,5E:i75 1.Ооооо о�9923о 0.o7bo1 o.oFaG2 о.1:с�а ,. AGEO -Ф.0�551 е,0о72Ъ �оо2555д •0.225а2 О,ь39о9 o.56oaQ 0.5а62Ь Ов99236 1,ооооо О.От3n1 о.GиSбо 6.1���0 „-vо3---о.а9г9t--п;а3г3в-оо3тtiе-о:32аь r-�•о;отiто-- -о.ооеьь -о.ооТь3 О.о7ваТ O.o73o1 t.onooo о.23Ч9о b.t`.д15 .. ►с[ -о.рl9ао n.looo5 о�3ееlг o.qaoza •o�13e9t -о.3гтрs •а.оьг�+о о.оеоог о.овsьь оwг39са 1.cnnпo с.с�17� 1 У-sт[дrг •o.t3zoa--.г,.oaaoe --•o�aaae9 -o;tts3+�- о.1ге3Т� 0.o32ot о.г42ьр д.l5ецs о.lгs7ь o.lsг7s о.оьlто 1.ги�еи 1 --- } ' � + � , ! ! � . � -•----•• ' � ���� � . : л � _ � , ` � Annex ТаЫе I.31, continued: , _ _ ,_ - - -. .-- -- -- -- - - _ - - - ; соv�Еитlач+� ►AиTLr р�►.иц Чс usE ЧECдESStoиs -_ огi23�е1 РАсЕ - t7 �,.�- _ � . .`г6Е КЕRлЕАиG (СЧЕлiI�и рАТЕ r и1/1t/el) кЕдлЕА 6jEE ддТН 1иСинЕ ° ---- -• --- - • - -- - - -- - рцснлll -■LFpPEYt-ERи.к1- -- ----- •-•- - -.. - ------- •----- ......---- - --------- - ,-�u3fTlE � :-.•. ь в..- :-i ■ ь-ь -г-1 •ь-.тв-3�-Эь--s-iс-'rs-м u(;-Т �Е• Р[- Е.. . R Е с• R Е-0-! i-о и ь в•. ■ ь- r ь в• ь ь а vARIAв1.E L1lт 1 • . дECRESStoи LtвT 2 ..-OI�EЧDENт-YГQT�i .C�CF . -' --- --- .. _ _..------- -- - -- - •------ - - --- -- • , °-- ..-rAwt+66EflГЕиТЕдЕр•`АЧ t1Ев--ЧО�ВЕR ,--D ; ��1Г-..-----------•---- --- -- . уиЕ?1PlE и а.3лэ51 ` +иА1,r585 UE �ARIAЧCE ' DF 'Uи OF SGUAдEa нЕви 5пивдЕ / . ,.-R lо.�АдЕ •t,,�D EcRESSrnи- �- --аv.ьгьг3�- ----• л:го3ге ------1Ч;vоlог- .. Аа�иаТЕо s�пилдЕ о:1Лге� рЕаэоw� 1ех1. г�t�7esuz о.1 п sa - .. еТаиDАдD Едд0и е�ав135 ---•- - ----- -• - - • -- .- ... -- --- -•-- •-- - . .. . ._------•- --- - ------- •-- - - _ --------• ---- •--- - - -- - -- -- -- �---- �_ -------•.....---- �nиl+pLE� ур тнЕ ЕаилТl�н ------------------ -------•...•• YAдtAeLEa нит Iи ТиЕ Еnиlьtгии ----------•.ь•- ,. YAetAOt.E , е НЕт+ ато ЕидпЧ и s чАикАв�.Е е�тА ки РАит�Аt.__ тоЕЕдАисЕ Е „ ve .вss5�чзt•.�1 О.зачаь о.опнгь зl,аие vч � ----- �vu�176i�.,! - 0.53iae .----_-O.u12oi --.- 5в.гь9 •--- --- -------- - ----- - -- - ---- - - - - - - - -- � .. vsл -.1�гьzт5t.,г -о.иоьз� a.uutlv 1s.1tt _ 1 ._-yeS �e91v'УФ3г.л1�h.09о10"-'-О;аё5�о�12.544- - . V1 -.�532ь1и.а1 .0„и5цлл б.Uо'J72 3.9и3 ' �''� -мл - ---- -_,7ьа5ьрц-ла- .p,0ui95 ---р.оа0ои-..--. -з.оиг •-- ---------- - ------ --- - .. . .. _ . _ . . F.. � , ►ЕЕ ,г35szteF-�y О,Оииго 9.OOU17 2.FЧ0 � � .. Уае • :.lопэзtие-�г' .�:о9ььs- ---- •о.оооlг '- . г:1ге -- -- - ----.. . _. _- л - �с�и�тлиг) -.аtевоеиЕ.лt ..-_ � • ---- • -• . ! .,----•- - , - SимиАяr ТвпЕЕ--. . --- . ----- - . _. . ..-vA41AELE---"---`- u�11PLE н -д snwRe -р�д силиеЕ --�1"+PIE Ч �--•-- n --- - nЕТв � ,:�'�`уд��Еи о;г9вQа-- о,звQзv - о:ов9тv О.гав9в �ил55ии3Е-о1- о.зеаэл • .. vv P.qtTr - О.sэгзе о.119и7 о.о21о9 о.огзьь ,9�tэlпзЕ-п1 • о.sз�ал .. vs• --- ------ р,Sлэп1 ---� р�lsг.зь- о.огtе9 ----.о,огsзl --- _��9гьгаЕ-п2 ---- .р�аоьll , va3 �лутЕ - О.s�ч72 О.еиаиг о.ооела о.2гц21 .е4lагvзЕ-п1 o.o9nlo .• v1 ------аоtЕЕт iACr[,т ', - о.s7tl3z -- - о.1и31з --о.ооzтl --•о.глlиl ------.lsзrь�с,Е-а1 • - ----- -о.оs+оь . - , vь �вчо ОлнЕр . о,звоlb о.lииь9 0.0а`5л O.uSbl! .,Т695иовЕ-Ои •О.ои1�5 ;. Уь3 • \� .sezos--v��1�+5ve-u;oolгr--y:llTтa ,2е5sа1вЕ-аэ о,оиаго �, u.�ej5l о.lи7оа o.0a112 о.2л95г •.1ou13fuE-02 -о.о9ло5 ..-tСонlТлиТl --------•-- --- --- -•--- --------- - -..-,,.оlвеоеиЕ.а1 --°----- - • - ,. -- - ' ♦ . ♦ _ . . -- - - i , , ���Т ��РУ A�At�.���.� 320 ANNEX II,. TRENDS AND DETERMINANTS OF INFANT AND CHILD MORTALITY IN KERALA by K.C. Zachariah and Sulekha Patel July 1982 - 321 - Abstract This paper estimates the trends in neo-natal, infant, and child mortality rates in three districts of Kerala and analyzes their socio-economic differentials and determinants. Inasmuch as the three sample districts are fairly representative of the state as a whole, the analysis given here is more or less valid for Kerala State. Estimates of trends and the analysis of determinants given in this paper are the first of their kind for Kerala; the data available from other sources (principally the SRS) are confined to recent data on the infant mortality rate (IMR). The IMR in Kerala is now about 50, which is less than half its level in the state from 1955-60 (107) or the current level in other parts of India (125). The neo-natal mortality is about 40 and the child mortality rate (1970-75) is 32. The IMR corresponds to an expectation of life at birth of about 65 years in the West Model Life Tables. The extent of mortality decline has been largest in the childhood ages (64% during 1958-73) and least during the first month after birth (32% during 1958-73). The IMR declined by 45 percent during the same period. Mortality rates are higher among children of mothers with no schooling, mothers who belong to scheduled castes, mothers in households with no protected toilet facilities, etc. Multiple regression analysis indicates that infant mortality and neo-natal mortality are significantly associated with mother's education (inverse relationship with education), caste (scheduled castes have higher mortality), year of birth (recent births have lower mortality), single or twin status (twin births are associated with higher mortality), and total household expenditures (higher expenditure households have lower mortality). The I-MR did not show a statistically significant association with land owned by the household, source of water supply to the house, toilet facility, or per capita household expenditures. On the other hand, child mortality showed a strong association with economic factors (per capita expendiures, land, water supply, toilet facilities, and total household expenditures) and a relatively weak (insignificant) association with caste, order of birth, and single or twin status of birth. In general, socio-economic factors explain only a small percent of infant mortality differentials at the household level. An absence of variables on hospital utilization, especially for ante-natal and post-natal care, is proposed in the paper as the main reason for the low explanatory power of the regression model used in this study. It is recommended that such variables be included in future research on the determinants of infant and child mortality. From the methodological point of view, the paper demonstrates that the often-noted positive association between the infant mortality rate and per capita expenditures in India is due to a circular causation -- an inverse relationship between income (expenditures) and mortality, and an inverse relationship between mortality and household size resulting in a positive association between mortality and per capita expenditures. The paper discourages the use of per capita expenditures in demographic analysis, particularly mortality analysis, - 322 - TABLE OF CONTENTS Page No. Abstract ................................................... Preface .................................................... Introduction ............................................... The Mortality Trend ........................................ Socio-Economic Differentials................................ Parity of the Mother ..................................... Caste .................................................... Education ................................................ Land Ownership ........................................... Sanitary and Water Facilities ............................ Household Expenditures ................................... Determinants of Infant and Child Mortality ................. Infant Mortality ......................................... Neo-natal and Child Mortality ............................ Anomaly of the Relationship between Per Capita Household Expenditures and the Infant Mortality Rate ...................................................... - 323 - LIST OF FIGURES AND TABLES Page No. Figure II.1: Infant Mortality Rate by Order of Birth, Kerala................................. Table II.1: Neo-Natal, Infant and Child Mortality Rates, 3 Districts of Kerala.................. Table 11.2: Estimates of Expectation of Life at Birth Based on West Model Life Tables and Observed IMR, and Child Mortality Rates: 3 Districts of Kerala..................................... Table 11.3: Neo-Natal, Infant and Child Mortality, by Parity of Mother, Kerala...................... Table 11.4: Neo-Natal, Infant and Child Mortality by Caste and Religion, Kerala ......................... Table 11.5: Neo-Natal, Infant and Child Mortality by Educational Atainment of Mother, Kerala ....................................... Table 11.6: Neo-Natal, Infant and Child Mortality by Household Ownership of Land, Kerala........... Table 11.7: Neo-Natal, Infant and Child Mortality by Toilet Facilities in the Household, Keralaa ...................................... Table 11.8: Neo-Natal, Infant and Child Mortality by Source of Water Supply, Kerala....................... Table 11.9: Infant and Child Mortality by Household Expenditures, Kerala ......................... Table 11.10: Probit Analysis on Infant Mortality, Kerala........................................ Table II.11: Ordinary Least Squares Regression on Infant Mortality, Kerala ............................ Table 11.12: Probit Analysis of Neo-natal and Child Survival Status of Each Birth, Kerala ................. Table 11.13: Distribution of Households by Total Household Expenditures in each PCE Class, Kerala........................................ Table 11.14: Distribution of Selected Households by Family Size, Children Ever-Born, and Children Dead of Household, Kerala ............................ - 324 - Preface This is the second in a series of reports on the World Bank/Bureau of Economics and Statistics (Trivandrum) research project on the determinants of fertility decline in Kerala. Other reports on the project are being prepared by the Bureau in Trivandrum. The Kerala study is part of a larger World Bank research project entitled "Determinants of Fertility Decline in Sri Lanka and South India" (World Bank Research Project No. 671-70) co-sponsored by the World Bank, the United Nation's Fund for Population Activities (UNFPA), and the Governments of India and Sri Lanka. The study in Kerala was carried out by the Bureau of Economics and Statistics, Government of Kerala, Trivandrum, under the direction of Drs P.A. Nair and R.S. Karup. Mr P.S. Gopinathan Nair was the officer-in-charge. Data for the study were collected from a random sample of 3,000 households from 150 villages selected at random from 3 selected districts: Palghat, Ernakulam and Alleppey. The field work was carried out during January-June, 1980. The study used 10 schedules to collect data. These are listed below: I. Household Schedule Ii. Ever-married Women, Marriage History III. Ever-married Women, Maternity History IV. Ever-married Women, Contraceptive Knowledge V. Ever-married Women, Fertility Regulation VI. Ever-married Women, Work History VII. Ever-married Women, Husbands Background VIII. Never-married Men, 20-39 Years IX. Ever-married Women, , 18-34 Years X. Currently Married Men, 20-39 Years This report is based on responses concerning the date of birth and the date of death (for those who died) of about 9,400 births born between 1950 and 1980 to ever-married women interviewed in the survey. Its main focus is to estimate infant and child mortality trends and differentials in the three survey districts and analyze their determinants. - 325 - Introduction Kerala is reported to have fairly good health conditions and low mortality rates. This is supported by the available mortality data. Among the various states of India, Kerala has the lowest death rate. In 1977, the crude death rate in Kerala was estimated to be 7.5 per 1,000 population and the infant mortality rate 50 per 1,000 live births; for India as a whole the corresponding rates were 13 and 129 respectively . Thus, the mortality rates in Kerala are indeed substantially lower than the rest of India. Other than this overall level, little is known about the mortality pattern in the state: how it varies from one region of the state to another, how it differs between various sucio-economic groups, etc. In other words, although we have fairly good estimates of the level of mortality in Kerala, little is known about its determinants. Even 4hat we know about the level comes from a single source, namely, the Sample Registration System (SRS), which is not uniformly reliable. Whether SRS death rates are reliable for Kerala is a question which has not been properly examined on the basis of independent sources of data. This paper attempts to fill some of the gaps in our knowledge about the mortality situation in Kerala. In particular, it provides an estimate of trends in the neo-natal, infant and child mortality rates in three districts of Kerala for recent periods, along with an analysis of their socio-economic differentials. These differentials are analyzed further using a multiple regression model to test and measure their independent contribution to the variation in the mortality among households. Several authors, analyzing Indian mortality differentials, have pointed out the observed positive relationship between the infant mortality rate and per capita household expenditures. This paper examines this - 326 - relationship in depth and tests one of the hypothesis put forth as an explanation for the peculiar relationship. Mortality Trend The Kerala fertility survey collected information on the data of birth of abaut. 9-,400 children bo,r-r between 1950 and 1980, their sur- vival status in 1980 (dead or alive), and the-date of death of those who died. These data are sufficient to estimate neo-nata. mortality, infant mortality and child mortality rates. Estimates are given in Table II.1. For the period 1975-80 the infant mortality rate (IMR) is estimated to be 46 which is less than the SRS estimate of 50 by 4 points. There is no reason why Palghat, Ernakulam, and Alleppey districts should be representative of Kerala state as a whole; the average for the three districts could indeed be lower than the State average. But the closeness of the two estimates for the 1970-75 period and the difference in 1975-80 make one wonder whether the estimate from the fertility survey is an underestimate for the more recent period. It probably is, but unlikely to be by a large margin, given recent estimates based on the SRS (Table II.1). In addition to the IMR, the survey gave information on the trend seen in the neo-natal mortality and child mortality rates (4 q). Declines were evident in all these rates, but the largest decline was in the child mortality rate, a lesser decline in the IMR, and the least decline in the neo-natal mortality rate. Thus, it appears that those mortality rates which are most influenced by environmental factors are the ones which show the greatest decline. The life expectations (based on the West Model of life tables 1/) 1/ See Ansley J. Coale and Paul Demeney, 1966. Regional Model Life Tables and Stable Populations, Princeton, New Jersey: Princeton University Press. - 327 - Table II.1: NEO-NATAL, INFANT AND CHILD MORTALITY RATES, 3 DISTRICTS OF KERALA FERTILITY SURVEY DATA SRS Period Neo-natal Infant Child Infant Mortality Mortality Mortality Mortality 1975-80 41 46 10 48.2 (1976-79) 1970-75 50 59 32 60.9 (1971) 1965-70 52 68 42 74.0 (1965-67) 1960-65 62 83 57 -- Before 1960 73 107 88 120 1951-60) All Years 55 73 47 Source: Kerala Fertility Survey. For SRS rates see the Directorate of Economics and Statistics, Kerala, 1980. Statistics for Planning, 1980. Table 1.14, p.4. - 328 - corresponding to the estimated infant and child mortality are given in Table 11.2. The expectation of life at birth in the three districts was 65.6 years during 1975-80. It increased consistently from about 53 years in- 1955-60 to 66 years in 1975-80, a 13-year increase in a 20 or 25-year period. This is not impossible, and is, in fact, the expected rate. Comparison between the estimate of life expectancy from the IMR and q, indicates that, for a given infant mortality rate, the child mortality in Kerala appears much too high. However, the decline in child mortality has been faster than the decline seen in the infant mortality rate and, therefore, the gap between the two is narrower now than in the past. Similar trends are observed in many other populations in recent years, Egypt for example. Socio-Economic Differentials Mortality rates vary by socio-economic characteristics of the household. A few of these are discussed below, and the isolation of their independent effects is done in the regression analysis which follows this section. Parity of the Mother: Children of women who have a large number of children tend to experience a higher mortality rate than children of women with fewer children. Thus, the IMR (during 1970-80) among children of mothers with 5 children or more was 73, while the IMR among children of mothers with 2 or less children was only 28. The relationship between parity and the mortality rate is consistently positive not only for the IR, but also for neo-natal and child mortality. Over the years (before 1970 and after 1970), the level of association has declined, but the pattern of the relationship between parity and mortality rate remains unaffected (see Table II.3), We are not able-to 329 - Table 11.2: ESTIMATES OF EXPECTATION OF LIFE AT BIRTH BASED ON WEST MODEL LIFE TABLES AND OBSERVED LMR AND CHILD MORTALITY RATES, 3 DISTRICTS OF KERALA Life Expectancy Based On: Period MR4 1 1975-80 65.6 - 1970-75 62.6 59.0 1965-70, 60.6 55.8 1960-65 57.4 51.8 Before 1960 52.6 44.9 - 330 - answer the question whether high parity caused high infant and child mortality. While this may be the case, the reverse causation (higher infant mortality causing higher fertili.ty) is also possible. A partial check on this question, referring to Figure II.1, indicates high parity may cause high mortality. The LMR is highest among first order births and birth orders of 6 and above; it is lowest among second order births. The decrease (over time) in the IMR is more noticeable among lower order births. In fact, among the 8 and 9 order births there is no decline at all. Caste: Scheduled Castes and Ezawas have relatively higher mortality rates than the other caste and religious groups (Table 11.4). The Scheduled Castes have the highest mortality rate. In 1970-80 their IMR and child mortality rates were twice as much as those of Christians, but the mortality decline among them has also been substantial. For example, the D4R among the Scheduled Castes is estimated to have declined from 142 (before 1970) to 76 (1970-80), a 46 percent reduction. Syrian-Christians have the lowest mortality rate. A comparison of their rates with that of the Nairs is particularly noteworthy. In the past, the Nairs had, on the whole, lower mortality rates, but in recent years the rates for Christians are lower. Thus, the rate of decline in mortality has been faster among Christians than among the Nairs. The reasons for this differential are not evident; perhaps the utilization of medical facilities, particularly for maternity, is higher among Christians than among the Nairs. - 331 - Table 11.3: NEO-NATAL, INFANT, AND CHILD MORTALITY BY PARITY OF MOTHER, KERALA Parity Neo-Natal IMR Child Mortality Before 1970-80 Before 1970-80 Before 1970-75 1970 1970 1970 <2 30 23 45 28 20 12 3 29 38 46 42 25 27 4 66 48 86 60 39 36 5+ 68 55 93 73 74 39 1 333 Table 11.4: NEO-NATAL, INFANT AND CHILD MORTALITY BY CASTE AND RELIGION, KERALA Caste Neo-Natal IMR Child Mortality Before 1970-80 Before 1970-80 Before 1970-75 1970 1970 1970 Scheduled 99 52 142 76 100 45 Nairs 43 38 66 47 45 36 Ezawas 77 48 96 59 53 40 Syrian Christians 56 33 69 35 45 7 Latin Christians 49 24 61 53 73 32 Muslims 45 27 74 38 78 34 -334 -. Educ ation: Infant and child mortality rates are inversely related to mother's education. In recent years, the chaao.e of a child dying. during the first year of its life is twice as high if its mother is not educated than if the mother has some schooling. Mortality had declined in all educational groups, but the maximum decline occurred among children born.to mothers who had 1 - 4 years of schooling (Table 11.5). It may not be a mere coincidence that it is exactly among this group of women that the fertility decline has also been at a maximum. Land Ownership: The relationship between land ownership and mortality is not consistent. A priori one would expect a negative relationship: children born in households with little land would experience higher mortality rates than children born in households with larger land holdings. This pattern is not always observed. For example, the OIR of 1970-80 shows a weak positive relationship but that of the period before 1970 shows the expected negative relationship (Table 11.6). This somewhat inconsistent relationship may be due to the nature of the two sets of data. Land ownership refers to ownership at the time of the survey, but infant and child mortality refers to a previous period: 1970-80 or before 1970. In Kerala, there have been some significant changes in land ownership patterns in recent years. Sanitary and Water Facilities: Households were classified in the survey according to whether the house had flush toilets, ESP slab toilets, or whether the members use a "compound", etc. for toilet needs. The mortality rates of children born in the different types of households show the expected differentials: households - 335 - Table 11.5: NEO-NATAL, INFANT, AND CHILD MORTALITY BY EDUCATIONAL ATTAINMENT OF MOTHER, KERALA Years of Neo-Natal IMR Child Mortality Schooling of Mother Before 1970-80 Before 1970-80 Before 1970-75 197 .. 1970 1970 No School 86 66 119 86 85 53 1 - 4 56 33 97 43 67 30 5 - 9 44 34 61 46 34 20 10+ 33 32 47 31 15 23 - 3.3.6.- - Table 11.6: NEO-NATAL, INFANT, AND CHILD MORTALITY BY HOUSEHOLD OWNERSHIP OF LAND, KERALA Land Owned By The Neo-Natal IMR Child Mortality Household Before 1970-80 Before 1970-80 Before 1970-75 (Cents) 1970 1970 1970 0 - 5 69 38 104 42 80 28 6 - 10 75 43 111 58 96 40 11 - 49 66 37 87 50 59 32 50 - 99 53 40 68 58 35 50 100+ 49 48 64 58 42 23 Note: 1 Cent = 1 of an acre. 100 - 337 - with protected toilet facilities (flush and ESP slab) have distinctly lower mortality rates than households which use "compounds" for toilet needs (Table 11.7). Source of water supply for the household is not a significantly discriminatory factor, however. The IMR in 1975-80 of households with pipes was 50, those with wells was 56, and those which used other sources for water supply was 44 (Table 11.8). Household Expenditures: Total household expenditures and per capita expenditures (total monthly household expenditures divided by the number of persons in the household) may have different relationships with mortality. There are problems with either usage as a measure of poverty. Therefore, we have provided both total as well as per capita expenditures (Table 11.9). The top panel of Table 11.9 gives the relationship between infant mortality and total monthly expenditures and the bottom panel gives the relationship with per capita monthly expenditures. On the whole, infant mortality decreases with an increase in household expenditures. The decline was more systematic in the past (before 1970, a period for which we have a larger sample size), than in the most recent period. The relationship with per capita household expenditures is less regular. For example, as per capita expenditures increases from below Rs.30 to above Rs. 120, the IMR decreases from 93 to 64, then increases to 74, and decreases again to 61. Decomposition of the rates for the periods before 1970 and after 1970 does not clarify the issue; irregularities are observed for both periods. Determinants of Infant and Child Mortality The previous section has shown that infant and child mortality rates have been, and remain, different for the various socio-economic -.338 - Table 11.7: NEO-NATAL, INFANT, AND CHILD MORTALITY BY TOILET FACILITIES IN THE HOUSEHOLD, KERALA Toilet Neo-Natal IMR Child Mortality Facility Before 1970-80 Before 1970-80 Before 1970-75 1970 1970 1970 Flush 37 38 62 46 25 26 ESP Slab 56 34 59 41 41 13 Compound, etc. 70 45 93 56 71 36 Table 11.8: NEO-NATAL, INFANT, AND CHILD MORTALITY BY SOURCE OF WATER SUPPLY, KERALA Source Neo-Natal IMR Child Mortality Before 1970-80 Before 1970-80 Before 1970-75 1970 1970 1970 Pipe 53 39 86 49 64 27 Well 64 44 84 56 57 33 Others 67 34 92 44 80 29 - 339 - Table 11.9: INFANT AND CHILD MORTALITY BY HOUSEHOLD EXPENDITURES, KERALA Monthly Household Infant Mortality Rate Child Mortality Expenditures Before 1970 All Before 1970 All 1970 -80 Periods 1970 -75 Periods Total Household ExpenditureG <Rs. 250 133 70 108 95 45 82 250 - 349 93 61 80 78 39 67 350 - 449 84 41 69 60 25 51 450 - 549 64 56 61 45 39 43 550 - 749 53 28 44 38 9 31 750+ 66 54 58 26 26 26 All Classes 85 53 73 61 33 54 Per Capita Household Expenditures <Rs. 30 128 47 93 47 25 58 30 - 49 100 57 81 67 41 59 50 - 59 80 41 64 80 36 69 60 - 79 75 72 74 68 36 60 80 - 119 80 32 66 49 16 42 120+ 59 65 61 17 23 18 All Classes 85 53 73 61 33 54 --340 - groups. They are higher among the Scheduled Castes than among Christians; higher among children of the illiterate than among children of the educated; higher among households with no protected toilet facilities, etc. What are the underlying factors of these differentials? We hypothesize that the mortality level among children born in a household'is determined by: 1) Mother's health, nutritional level, etc.; 2) Nutrition received by the child since birth; 3) Medical assistance received by the mother and child before, during, and after birth; 4) Sanitary conditions of the house and its locality. Each of these is affected by the economic status of the household, educational attainment of the parents, and cultural practices pertaining to childbearing and child rearing. The sanitary conditions of the house will be affected also by the general location of the structure, and sanitary facilities such as water supply, toilet facilities, etc., in the house. The preliminary tabulation of the Kerala data does not give measures of all factors. The available relevant variables in the preliminary tabulations are: - Mother's education: years of schooling (V8) - Land owned by the household (V6 - Toilet facility in the house (V3) - Water supply (V2) - Caste (V43 - Monthly household expenditures (V47 - Per capita household expenditures PCE - 341 - Infant Mortality: This analysis is confined to 8,921 births (out of a total of about 9,400 reported in the survey), which took place at least 12 months before the survey date. Thus, it excludes births in 1979 and 1980, to ensure that all births have the same (12-month) exposure period. This restriction, however, does not solve the problem related to the choice of the best dependent mortality variable. There are a number of choices: 1) Number of deaths (infant) per woman 2) Proportion of children ever-born who died 3) Dummy variable 0 or 1, defined as 0 if a birth ends up in a death during the first 12 months and 1 if it survives the the first 12 months Other alternatives are possible, especially functions or transformations of the above. In the analysis given below we have tried (2) proportion of infant deaths and (3) survival status of each birth. A probit analysis on the dummy variable (3) gave the results in Table 11.10.2/ The chance of a child surviving is significantly higher for recent births than for earlier births (positive association with year and month of birth), for single births than for multiple births (negative association with twin-single status), for births of a higher order (positive association with birth order) (see also Figure 11.1) 3/, for children of 2/ For a discussion of probit analysis, see Takeshi Amemiya, 1981. "Qualitative Response Models: A Survey," Journal of Economic Literature, December, 19(4): 1483-1536. 3/ The relationship between birth order and infant mortality is not linear. The IMR decreases with birth order when the birth order is small (up to order 5) and then increases at higher order births. 34z - Table II.10: PROBIT ANALYSIS ON INFANT MORTALITY, KERALA Dependerrt Variable: Probability of a Live Birth Surviving The First 12 months Independent Variables B-Coefficient t-Statistic Mother's Education (years) +0.04778 6.0 Caste 1/ +0.1937 3.4 Birth Status (single or twin) -0.8947 8.8 Year and Month of Birth +0.001326 5-.3 Birth Order - +1.4 Linear +0.1082 +3.6 Quadratic -0.0108 -3.3 Per Capita Expenditures - 0.1 Land Ownership 0.2 Water Supply 2/ - -1.1 Total Household Expenditures* +0.0004096 3.7 Toilet Facility* 3/ - -1.0 *From a different regression from which the variable per capita household expenditures was excluded. 1/ Scheduled Castes and tribes = 0; all others = 1. 2/ Pipe and well = 1; all others zero. A modification of this in which pipe = 1; all others zero did not affect the statistical significance. 3/ Flush and ESP Slab = 1; all others zero. - 343 - educated mothers than of illiterate mothers (positive assocition with years of schooling of mother), for births to higher caste women (positive associa- tion with caste), and for births in households with higher monthly expendi- tures (positive association with monthly household expenditures). The availability of piped water or well water does not appear to affect greatly an infant's survival probability. The same is true of toilet facilities and the amount of land owned by a household. It was expected that a protected water supply, protected toilet facilities and large land holdings would significantly reduce infant mortality. There is, however, no evidence for such relationships from the data. Probit analysis does not give a good estimate of the total variance explained by the regression. However, available information (e.g., R2 statistic = 0.030) indicates that the independent variables together explain only a very small percent of the total variance. It appears that the really critical variables which explain much of the variance in the probability of survival are not included in our analysis. The above conclusions are on the whole supported also by an ordinary least squares analysis of the proportion of infant deaths per woman (Table II.11). Years of schooling of mother, her caste, and the average year of occurrence of her births, all have statistically significantly negative relationships with infant mortality. Land, water supply, and toilet facilities have no significant relationship. In these respects, the conclusions from probit analysis and OLS are the same. However, there is a difference between the two with respect to the significance of household -344- Table II.11: ORDINARY LEAST SQUARES REGRESSION ON INFANT MORTALITY, KERALA Dependent Variable: Proportion of Children Who Die in the First 12 Months Independent Variables B-Coefficient F-Ratio 1) Average Year of Birth of all Children -0.06415 8.1 2) Years of Schooling (Mother) -0.13646 26.8 3) Caste -0.08458 16.3 4) Land - 0.1 5) Per Capita Expenditures - 3.8 6) Water Supply - 1.3 7) Toilet Facility - 0.7 8) Total Household Expenditures* - 1.0 *Different regression in which per capita expenditures was not included. -345- expenditures. In OLS, total household expenditures did not show a significant relationship while the probit analysis indicated a negative relationship with infant mortality. Both approaches gave a similar relationship - a statistically insignificant relationship - with per capita household expenditures, but the OLS estimate is only just marginally below the 95 percent level of statistical significance (94.9%). Two overall conclusions emerge from the above analysis: 1. Socio-economic factors explain only a small percent of the infant mortality differentials at the household level. 2. Economic factors, such as ownership of land, have a relatively smaller role to play in determining the probability of infant deaths in a household than social factors such as mother's education or her caste. If socio-economic factors cannot account for much of the infant mortality differentials, there must be other factors which do. One possibility is hospital utilization for pre-natal care, confinement, and post-natal care of the mother and child. Unfortunately, until data are collected and tested it will not be possible to confirm this. The relationship with economic factors versus social factors might also be explained in terms of the hospital utilization hypothesis. In Kerala, government hospital facilities are free to most people. They are widely distributed over the state and the transportation network is good and cheap. For example, out of the 150 sample villages (panchayats) in the survey, 137 had either a government hospital or dispensary, or a private hospital or dispensary. Therefore, the utilization of hospital facilities is determined more by the individual demand for such services than by their - 346 - supply or the ability to pay for services. Education is an important element in the demand. Hence, this may be one reason we noted the statistically significant relationship between mother s schooling and the infant mortality rate. In the Kerala context, the demand for and'utilization of hospital facilities are also determined by the degree of personal contact with hospitals or a dispensary staff - doctors, nurses, midwives, technicians, etc. Historically, some communities (Syrian Christians, for example) were over represented among hospital staff than others (Scheduled Caste, for example). Although this imbalance has changed considerably, it is still possible that hospital staff are disproportionately distributed among the religious and caste groups. The statistically significant association between caste and infant mortality may be explained by the differential utilization of hospital facilities, and this may be influenced by the differential personal contact one has with the hospital system. Neo-Natal and Child Mortality: Probit analysis similar to that done on infant deaths was done for neo-natal mortality (0 if a birth ends up in a death during the first month, and 1 if it survives the first month) and child mortality (0 if a birth ends up in a death during 1 to 4 years of its life, and 1 if it survives the first five years). The results are shown in Table 11.12. These results regarding neo-natal mortality are very similar to those regarding infant mortality, but those regarding child mortality are different. For neo-natal (and infant) mortality, order of birth, single or - 347 - Table 11.12: PROBIT ANALYSIS OF NEO-NATAL AND CHILD SURVIVAL STATUS OF EACH BIRTH, KERALA Independent Neo-Natal Survivorship (first month) Child Survivorship (1-4 Years) Variable Coefficient t-Statistic Coerticient t-Statistic Mother's education +.05219 6.2 +.03737 3.6 Caste +.1316 2.1 - 1.7 Birth Order +.02249 2.0 - -1.7 Single or Twin Birth -.9123 -8.8 - -0.9 Year of Birth +.007802 3.0 +.002406 6.8 Per Capita Expenditures - -1.6 +.001564 2.1 Land -0.1 +.0003671 2.4 Water Supply -1.1 +.1411 2.0 Total Household Expenditures* +.000-'669 2.4 +.0003343 2.3 Toilet Facility* -1.0 +.1768 2.5 * From a different regression from which PCE was excluded. - 348 - twin status of the birth, and caste of the mother are statistically significant; while per capita expenditures, land, water supply, toilet facilities, etc., are not. On the other hand, for child mortality, caste, order of birth, single-twin status are not significant; while economic factors such as per capita expenditures (higher expenditures mean higher survival probability), land (the larger- the size of. land holding, the higher is the survival probability), water supply, toilet facilities, etc., are very important. Anomaly of the Relationship Between Per Capita Household Expenditures and the Infant Mortality Rate The previous section has indicated a somewhat conflicting relation- ship between the infant mortality rate and per capita household expenditures (PCE). Probit analysis showed no statistically significant relationship between the two (t = -0.1); OLS analysis of all births indicated a marginally significant (significant at 94.9% level) positive relationship; and OLS of recent births (1975-79) gave a highly significant positive association (t = +2.5). 4/ There was no support for the expected negative association between PCE and IMR in any of the analyses. This is not the first study which has shown a non-negative relationship between per capita.expet.ditures and the infant mortality rate. In fact, a positive relationship seems to be the general pattern with Indian data. In a recent article N. Krishnaji has 4/ This is not presented above. - 349 - brought attention to this point rather dramatically. 5/ According to the nineteenth round of the NSS, the IMR was lowest (33 per 1,000) in the lowest expenditure class and highest (293) in the highest expenditure class. Such a "perverse" relationship between PCE and IMR is reported to have been observed in all NSS rounds since 1953-54. One of the explanations given for this peculiar relationship is the two-way causation between PCE and mortality: incou (expenditures) affecting mortality and mortality affecting family size and thus affecting per capita income or expenditures. Do the Kerala data support this hypothesis? 5/ N. Krishnaji, 1980. "Poverty and Family Size," in Social Scientist, Special Number November, pp. 22-35. Table II from this article is reproduced below: TABLE II DIFFERENTIAL MORTALITY RATES IN RURAL INDIA (1964-65) Monthly Per Capita Infant Mortality Expenditure Death Rate Rate (Rs) 0-11 10.03 32.94 11-15 10.58 71.71 15-21 13.38 122.29 21-28 16.06 152.37 28-43 17.88 153.13 43 and above 21.81 293.27 All groups 14.75 127.29 -~ 350 - Table 11.13 gives the distribution of households by total household expenditures in each PCE class. There are 178 households in the lowest PCE class, and all of them have total expenditures below Rs. 450 per month. There are 331 households in the highest PCE class (more than Rs. 120). One of these households belongs to the lowest (total) expenditure class (less than Rs. 150). It was included in the highest PCE class simply because there was only one member in the household. Thus, the rich households as measured by total expenditures are usually not classified as poor when status is defined on the basis of per capita expenditures. On the other hand, some poor households (based on total expenditures) sometimes get classified as rich when the criterion is changed from total to per capita expenditures. Thus, the "misclassification" is not symmetric: the chance of "poor" becoming "rich" is greater than the "rich" becoming "poor". Consequently the IMR of the "rich" (PCE basis) is over-estimated. A more vivid confirmation of this shift is provided in Table 11.14 where the poor households which became rich (P-> R) due to the change in criterion (see footnote at the bottom of the table) is compared with rich households which became poor (R-> P). In the first category (P-4 R), there is no household with more than 7 persons and the average size of the household was only 4.25. In the second category (R-> P), there is not even one household with less than 7 persons. The average size of the household is 11.43 persons. This perfect separation is quite accidental, but the large difference between their averages (by as much as 7 persons per household in this case) is not. The "poor" families became "rich" mainly because of the small size of their families and the "rich" became "poor" because of the very large size of their household. - 351 - Table 11.13: DISTRIBUTION OF HOUSEHOLDS BY TOTAL HOUSEHOLD EXPENDITURES IN EACH PCE CLASS, KERALA Per Capita Expenditures (PCE) Total Rs 30-49 50-59 60-79 80-119 120+ Total Expenditure <30 <150 25.8 3.7 1.3 1.3 0.2 0.3 3.2 150-249 55.1 36.4 11.4 8.6 4.1 0.9 17.4 250-349 12.4 38.9 28.2 27.9 8.8 2.4 23.1 350-449 6.7 13.1 29.1 21.4 22.9 5.4 17.9 450-549 -- 4.7 18.7 20.9 21.8 6.6 13.5 550-649 --- 2.2 4.0 11.7 16.9 12.3 8.2 650-749 --- 0.9 2.2 4.6 9.8 11.5 4.7 750-1499 -- - 5.1 3.6 14.7 45.3 9.9 1500 -- -- -- --- 0.8 15.1 2.0 100 100 100 100 100 100 100 Number of Households 178 678 454 548 490 331 2679 - 352 - - Table II,14: DISTRIBUTION OF SELECTED HOUSEHOLDS BY FAMILY SIZE, CHILDREN EVER-BORN, AND CHILDREN DEAD OF HOUSEHOLD, KERALA Household- Size No. Of Children Children Died Households Ever-Born (P-- R) (R--- P) (P-> R) (R-> P) (P-- R) (R-> P) 0 - - 53 18 365 1,35 1 4 - 73 26 93 32 2 42 - 121 43 49 9 3 85 - 117 17 14 7 4 160 - 78 18 6 1 5 157 - 34 12 4 - 6 69 - 32 7 - - 7 11 - 12 13 - - 8 - 15 6 10 - - 9 - 41 3 12 - - 10 - 37 1 2 - - 11 - 13 - 3 - - 12 - 25 - 2 - - 13 - 11 1 1 - - 14 - 11 - - - - 15 - 22 - - - - 16 - 5 - - - - Total 528 180 531 184 531 184 Average 4.28 11.16 2.88 3.84 0.52 0.41 (P-4 R) = Households in which total expenditures are low (less than Rs. 450), but per capita expenditures high (more than Rs. 60 per month). (R-> P) = Households in which total expendi.tures are high (more than Rs. 450), but per capita expenditures low (less than Rs. 60 per month). - 353 - The asymmetry of the shift is quite evident from this array also (Table 11.14). While 180 rich households became "poor" as many as 528 house- holds moved in the opposite direction. Thus, the criterion of PCE tends to increase the mortality level of the rich defined on the basis of total expenditures. Is the anomalous relation between PCE and IMR due to higher mortality among the poor and its effect on PCE? To answer this question an adjusted measure of per capita household expenditures (APCE) was calculated for each household as a ratio of total expenditures to the expected household size if none of the children had died; that is, APCE - Total Monthly Household Expenditures Household Size + Deaths Among Ever- Born Children APCE is thus independent of the number of deaths among the children. Using APCE, regression coefficients were re-estimated by the OLS and Probit Analysis. The OLS gave a statistically significant negative relationship (B = -0.080; F - 11.0;A R 2 0.5%) 6/ between proportion of children who died and APCE. 6/ It may be argued that the negative relationship is inevitable as deaths occur in the denominator of the independent variable: Expenditures APCE H.H. Size + D (All deaths) and in the numerator of the dependent variable: D (infant deaths which are 55% of D) P (Children ever-born) However, H.H. size is net of all deaths; and, therefore, when D is added, the denominator of APCE becomes independent of D. - 354 - The Probit Analysis also gave a statistically significant negative relationship between APCE and infant mortality (t = -7.9). Thus, the insignificant relationship between PCE and infant mortality became significant once the PCE was adjusted to make it independent of deaths in the household. Thus, the Kerala data support the circular causation hypothesis. The observed positive or non-significant relationship between PCE and IMR is due to the inverse relationship between income and mortality, and the inverse relationship between mortality and household size. These result in a direct relationship between mortality and PCE. From our work, per capita household expenditures do not appear to be an adequate measure of the economic status of a household in demographic analyses, particularly in mortality analyses. .. . .... . ..... . .... - 355 - BIBLIOGRAPHY Amemiya, Takeshi, 1981. " Qualitative Response Models: A Survey", Journal of Economic Literature, December, 19(4): 1483-1536. Ayar, S. Ramanath, 1923. Progressive Travancore, Trivandrum: The Anatha Rama Varma Press. Bongaarts, John, 1980. "The Fertility Inhibiting Effects of the Intermediate Variables", Population Council Working Paper No. 51, May, New York: The Population Council. Chen, Lincoln, et al., 1974. "A Prospective Study of Birth Interval Dynamics Rural Bangladesh", Population Studies 28(2): 277-297, July. Coale, Ansley J., and Paul Demeney, 1966. Regional Model Life Tables and Stable Populations, Princeton, New Jersey: Princeton University Press. Government of India, Central Bureau of Health Intelligence, 1981. Pocket Book of Health Statistics of India 1980, Coimbatore: Government of India Press. Government of In`ia, Central Statistical Organization, n.d. National Sample Survey Report No. 89 (N.S.S. 14th Round),New Delhi: National Sample Survey Organization, Ministry of Planning, Government of India. Government of India, Ministry of Health and Family Welfare, n.d. Family Welfare Programme in India Yearbook, 1977-78, New Delhi: Department of Family Welfare. Government of India, Department.of Family Welfare, 1979. Family Welfare Programme in India, Year Book 1978-79, New Delhi: Mass Mailing Unit, Department of Family Welfare. Government of India, Office of the Registrar General, Ministry of Home Affairs, 1979. Survey on Infant and Child Mortality, 1979, A Preliminary Report, New Delhi: Sunlight Printers. Government of India, Registrar General and Census Commissioner, 1971. Census of India, Part II-A, Census Tables on Population, New Delhi: Government India Press. Government of India, Registrar General and Census Commissioner, 1981. Census of India, 1981, Provisional Population Totals, Series-1, Papers 1 and 2, India, New Delhi: Pearl Offset Press. Government of India, Vital Statistics Division, Ministry of Home Affairs, 1966-80. "Sample Registration: Bulletin", Volumes for 1966-80, New Delhi: Office of the Registrar General. Government of Kerala, 1978. Report of the Committee for Evaluation of Family Welfare Programme (Kerala), Shoranur: Government Press. - 356 - Government of Kerala, Director of Census Operations, 1972. Census of India, 1971, Series-9, Kerala, Part II-A, General Population Tables, New Delhi: Government of India Press. Government of Kerala, Director of Census Operations, 1981. Census of India, 1981, Kerala, Series 10, Provisional Population Totals, Paper 1 of 1981, Cochin: Test Book Press. Government of Kerala, Directorate of Economics and Statistics, 1980. Statistics for Planning 1980, Cochin: Text Book Press. Jeffrey, Robin, 1976. The Decline of Nayar Dominance, New Delhi: Vikas. Kent, Mary Mederios, 1981. "Breastfeeding in the Developing World: Current Patterns and Implication for Future Trends", June, Reports on the World Fertility Survey (2), Washington, D.C.: The Population Reference Bureau, Inc. Kerala State, Bureau of Economics and Statistics, 1976. Land Reforms Survey in Kerala 1966-67, 1968 Report, Trivandrum: Government Press. Kerala State, Bureau of Economics and Statistics, and Demographic Research Centre, Trivandrum, 1977. "An Assessment of the Camp Performance and the Unprotected Couples in Palghat District after the Mass Camp, D.R.C. Paper No. 92, Trivandrum: Government Press. Kerala State, Director of Census Operations, 1972.' Census of India: Kerala, General Population Tables, Series 9, Part 11-A, New Delhi: Government of India Press. Kerala State, Population Research Centre, Directorate of Economics and Statistics, 1980. "Note on the Calculation of Births Averted and Couples Protected due to Family Welfare Programme in Kerala", Trivandrum: Population Research Centre (unpublished). Krishnaji, N.,1980. "Poverty and Family Size", in Social Scientist, Special Number, November, pp. 22-35. Krishnan, T.N., 1976. "The Demographic Transition in Kerala: Facts and Factors", Economic and Political Weekly 11:1203-1224. Kurup, R.S., and S. Cecil, 1976. "A Note on Birth and Death Rates in Kerala Sample Registration Report", Series No. 11, Demographic Research Centre, Trivandrum: Bureau of Economics and Statistics (October), mimeographed. Mencher, Joan P., 1980. "The Lessons and Non-Lessons of Kerala: Agricultural Labourers and Poverty", Economic and Political Weekly, Special Number, October, pp. 1781-1802. Nag, Moni, 1981. "Impact of Social Development and Economic Development on Mortality: A Comparative Study of Kerala and West Bengal", The Population Council, Center for Policy Studies, Working Paper No. 78, New York: The Population Council. r - 357 - Nagam Aiya, V., 1906. The Travancore State Manual, Trivandrum: Travancore Government Press. Nair, P.R. Gopinathan, 1974. "Decline in Birth Rate in Kerala: A Hypothesis about the Interrelationship between Demographic Variables, Health Services and Education", Economic and Political Weekly 9:323-336. Nortmann, Dorothy L., and Ellen Hofstatter, 1978. Population and Family Programs, Ninth Edition, New York: The Population Council. Panikar, P.G.K., 1980. "Inter-Regional Variation.in Calorie Intake", Economic and Political Weekly, Special Number, October, pp. 1803-1813. Population Planning Council of Pakistan, 1976. World Fertility Survey, Pakistan Survey First Report, Lahore: Population Planning Council of Pakistan. Potter, R. G., Mary L. New, John B. Wyon, and John E. Gordon, 1965. "Applications of Field Studies to Research on the Physiology of Human Reproduction", in Mindel C. Sheps and Jeanne Clare Ridley (editors), Public Health and Population Change, Pittsburgh: University of Pittsburgh Press, pp. 377-399. Ratcliffe, John, 1978. "Social Justice and the Demographic Transition: Lessons from India's Kerala State", International Journal of Health Services 8(1): 123-144. Republic of Korea, World Fertility Survey, and International Statistical Institute, 1978. Korean National Fertility Survey, 1974, Summary of Findings, London (?): World Fertility Survey (?). Sri Lanka, Department of Census and Statistics, 1978. World Fertility Survey, Sri Lanka, 1975, First Report, Ministry of Plan Implementation (March). World Bank, 1980. World Development Report, 1980, New York: Oxford University Press for the World Bank. World Fertility Survey and International Statistical Institute, 1978. Malaysian Fertility and Family Survey, 1974, Summary of Findings, Kuala Lumpur: Malaysian Department of Statistics. United Nations, Department of Economic and Social Affairs, 1975. Poverty, Unemployment and Development Policy: A Case Study of Selected Issues with Reference to Kerala, ST/ESA/39, New York: United Nations.