Papers of the East-West Population Institute, no. 64 Socioeconomic and cultural aspects of marriage and fertility in urban Pakistan MehtabS. Karim X \7*S EAST-WEST CENTER HONOLULU HAWAII Socioeconomic and cultural aspects of marriage and fertility in urban Pakistan Mehtab S. Karlm Number 64 • December 1979 PAPERS OF THE EAST-WEST POPULATION INSTITUTE M E H T A B S. KARIM is a doctoral candidate in sociology, International Population Program, Department of Sociology, Cornell University. Library of Congress Cataloging in Publication Data Karim, Mehtab S 1947Socioeconomic and cultural aspects of marriage and fertility in urban Pakistan. (Papers of the East-West Population Institute ; no. 64) Bibliography: p. I. Marriage age-Pakistan. 2. Fertility, HumanPakistan. I. Title. II. Series: East-West Population Institute. Papers of the East-West Population Institute ; no. 64. HQ670.5.K37 304.6'3 79-28401 CONTENTS Preface v Abstract / 4 Present study Findings and discussion Summary and conclusions References 24 IV TABLES AND FIGURES Tables 4 1 Singulate mean age at marriage (SMAM): Pakistan, 192 1 - 7 2 2 Age-specific and total fertility rates, mean number of live births, and mean number of pregnancies, by current age of women: urban Pakistan, 1968-69 7 3 Total marital fertility rate, mean number of live births, mean number of pregnancies, and mean first pregnancy interval, by age at marriage: Pakistan. 1968-69 8 4 Mean number of live births by age at marriage and selected characteristics: currently married women in urban Pakistan 10 5 Correlation matrix of socioeconomic and demographic characteristics, for currently married women aged 1 5—44 in urban Pakistan: 1968-69 13 6 Multiple classification analysis of number of live births, using age at marriage as a predictor and duration of marriage (A) and first pregnancy interval (B) as covariates, for three age groups of currently married women in urban Pakistan, 1968—69 14 1 Multiple classification analysis of number of live births, using socioeconomic status as a predictor and age at marriage (A), duration of marriage (B), and first pregnancy interval (C) as covariates, for three age groups of currently married women in urban Pakistan, 1968-69 18 8 Multiple classification analysis of number of live births, using current residence and language as a predictor and age at marriage (A), duration of marriage (B), and first pregnancy interval (C) as covariates, for three age groups of currently married women in urban Pakistan, 1968-69 19 Figures 1 Mean age at marriage by year of marriage, among currently married women: Pakistan, 1968-69 5 2 Mean number of live births by age at marriage, unadjusted (A) and adjusted for the effect of duration of marriage (B), first pregnancy interval (C), and combined effect of (B) and (C): currently married women aged 15—44 in urban Pakistan 16 V PREFACE Thanks are due to Drs. J. Mayone Stycos, Peter C. Smith, and Samuel H. Preston for their valuable comments on an earlier draft of this paper. Support for the publication of this paper was provided by the Office of Population, Agency for International Development. ABSTRACT Using data from the 1968-69 Pakistan National Impact Survey and multiple classification analysis, this paper examines the relationship between age at marriage and fertility among 1,114 currently married women 15—44 years of age who were living in urban areas at the time of the survey. Mean age at marriage reported by the sample was very low (16 years), although among younger women it was somewhat higher (17 and a half). A verage cumulative fertility among older women (those aged 35—44 when interviewed) was 6. 7 births, but it was 29 percent lower among women who had married at age 19 or later than among those who had married at age 16 or earlier, even though women who had postponed marriage had a faster tempo of childbearing. Cumulative fertility differentials were found among socioeconomic groups, ethnic groups, and residential (metropolitan versus nonmetropolitan) groups who married at younger and older ages. Duration of marriage and first pregnancy interval proved to be important predictors of the inverse relationship between age at marriage and cumulative fertility. With duration of marriage controlled, cumulative fertility was 14 percent lower among late-marrying women than among those who had married early; with first pregnancy interval controlled, cumulative fertility was 32 percent lower among late-marrying women. Raising the mean age at marriage for Pakistani women to 19 from, the current legal age of 16 would have beneficial social, as well as demographic, effects. But because the early marriage age appears to be related, at least among some ethnic groups, to the economic liability of daughters, increasing the legal age at marriage must be accompanied by fundamental changes in women's roles. The fertility of a population is determined by a number of interrelated social, economic, and demographic factors. Especially in areas where fertility is high, female age at marriage has important implications for the study of fertility, because the length of the female reproductive period is largely determined by the age at which a woman enters marital union. Coale and Tye (1961) have pointed out that early marriage could be one of the major determinants of fertility; they suggest, however, that the fecundity of women married at a very early age may be impaired by premature pregnancy. Malthus recognized the importance 2 of the age at which women marry and suggested that delayed marriage was one important way to reduce population growth (Leasure, 1963). Davis (1955) identified certain values prevalent in non-Western societies, such as universality of marriage and emphasis on early marriage for women, along with the joint family system, as the major factors contributing to higher fertility. Though no significant differences in fertility have been reported among women living in joint or nuclear families in Pakistan (Karim, 1974), it appears quite likely that a low average age at marriage may account in part for Pakistan's high fertility. This paper attempts to relate the fertility of urban Pakistani women to their ages at marriage and to identify sociocultural and economic factors influencing this relationship. The religious support given to universal and early marriage in Pakistan, as well as such social values as segregation of the sexes, strong taboos against premarital sex, and a large-family norm may contribute to the low age at marriage as well as to the high fertility of women in Pakistan. Historically, a well-established inverse relationship between a woman's age at marriage and her fertility has existed throughout the world. In the recent past, however, this association seems to have become substantially weaker in the developed countries. For example, Glass and Grebenik (-1954) have emphasized the decreasing importance of age at marriage in predicting fertility in Britain. They argue that age at marriage and fertility are not tightly linked for the more recently married women owing to their adoption of birth control. But Busfield (1972) has shown that these differences persist even in recent marriage cohorts. Her study explores a number of reasons for this persistence. Differences in fertility by age at marriage have remained relatively large in societies where marriage is universal, average age at marriage is low, and the use of contraception is not very common. Numerous studies done in India have demonstrated an inverse relationship between age at marriage and fertility. Driver (1963), for example, relating the high fertility of central Indian women to their low age at marriage, reported that, at the completion of their childbearing period, women who married at age 1 8 or later were 30 percent less fertile than those who married before age 14. On the basis of a set of calculations using the stable population concept, Agarwala (1967) suggested that if female age at marriage in India were raised from 15.6 years to 19.3 years on the average, there would be a decline of 27 percent in the birth rate. Malakar (1972) reported a consistent inverse relationship between age at marriage and fertility in India based on different fcr- 3 tility measures. Although he indicated that no significant decline in fertility would result from raising age at marriage from 15 to 17 years, a substantial decline would occur, if it were raised to 19 or 21 years. Talwar and Seal (1974) reached results somewhat similar to those of Malakar. But they suggested that decline in fertility resulting from an increase in age at marriage might not be significant in the long run. On the other hand, Das (1967) did not find a significant impact of an increase in age at marriage on fertility and suggested the existence of a critical line below which the postponement of marriage does not have a bearing on completed family size. The results of studies conducted elsewhere generally suggest a significant impact of female age at marriage on fertility. Leasure (1963), comparing Bolivian and Turkish data, showed that a substantial decline in fertility would result if average age at marriage were increased to 27 or 28 years in those countries. Siffman (1967) noted high marital fertility rates of 8.5 and 9.1 among women in Armenia and Azerbaijan in the Soviet Union prior to the Second World War; and he related these rates to a low average age at marriage of about 17 years. Kim et al. (1974) considered age at marriage as the most important factor in explaining fertility decline in Korea during the past two decades among currently married women 4 0 - 4 9 years old. Palmore and Marzuki (1969), using West Malaysian survey data, suggested that early marriage leads to higher cumulative fertility whereas the nonexposure associated with multiple marriage tends to retard completed family size. They found that fertility, as well as age-at-marriage differentials by rural-urban residence, ethnicity, and education of women, shows a clear impact of age at marriage on completed family size. In Pakistan, despite the low average age at marriage and high fertility, little research has been done on the relationship between these two variables. Using available census data on marital status by age, Sadiq (1965) prepared estimates of singulate mean age at marriage (SMAM). He reported an increase of over four years in S M A M during the past half century, as shown in Table 1. A sudden increase in SMAM in the 1940s was mainly the result of social and political disturbances during the decade. But the increase in the 1960s seems to have been due to economic and social changes. (For a detailed discussion of trends in nuptiality in Pakistan see Karim, 1979.) On the basis of 1961 Census data, Ahmed (1969) estimated the effect on the crude birth rate of an assumed increase in the minimum age at marriage for females in Pakistan from the minimum legal age at marriage of 16 years to 20 years. He concluded that such an increase T A B L E 1 Singulate mean age at marriage (SMAM): Pakistan, 1921-72 Year SMAM 1921 1931 1941 1951 1961 1968 1972 15.8 15.7 16.4 17.9 17.6 19.6 20.0 SOURCES: 1921-61: Sadiq (1965), based on Population Censuses of Pakistan. 1968: Pakistan, Siatistical Division (1973). 1972: Pakistan, Census Organization (1974). would have resulted in a maximum fertility decline of 12 percent between 1961 and 1971. Furthermore, he suggested that the crude birth rate would show an upward trend between 1976 and 1991 and would decline afterward as small-sized cohorts resulting from the increase in age at marriage began reproduction. Ahmed, however, assumed a sudden shift in age at marriage from 16 to 20 years; therefore, his results may not be entirely realistic, because changes in age at marriage in societies like Pakistan occur gradually. In a more recent study, De Tray (1977) has pointed out some of the problems associated with studying the relationship between age at marriage and fertility in Pakistan. Nevertheless, his regression results indicate a decline of 8 percent in completed fertility from raising the average age at marriage from 15.2 to 19.2 years. PRESENT STUDY A review of the literature strongly suggests that the age at marriage has a direct bearing on fertility. Most of the studies dealing with the age-atmarriage and fertility relationship have, however, ignored the social and cultural aspects of the relationship. They are usually "model" oriented, explaining the magnitude of expected decline in overall fertility with increases in age at marriage. They have largely failed to mention that a shift in age at marriage is an important cultural change, a change in traditional behavior relating to one of the most fundamental institutions in the society. It is likely to be indicative of other changesiin attitudes toward marriage, as well as reproductive behavior. A rise in age at marriage is one of the major factors that may cause 5 a decline in fertility. This is especially true in societies like Pakistan, where not only are large-family norms prevalent but also contraceptive use is still very limited (Pakistan Population Planning Council, 1976). Age at marriage in Pakistan, as in many other modernizing societies, has shown an upward trend during the past decades. This increase is especially notable in urban areas (Smith and Karim, 1978). Figure 1 shows an increase of about two years in age at marriage for women married recently (four years prior to the survey) compared with those married for more than 20 years. It appears likely that, as in other countries, women who postpone their marriage for a few years will tend to have lower fertility. F I G U R E 1 M e a n age at m a r r i a g e b y y e a r o f m a r r i a g e , a m o n g currently married w o m e n : Pakistan, 1968—69 u-f Before 1950 1 1950 1 1955 r 1960 1 1965 Year of marriage S O U R C E : Pakistan, Family Planning Council (n.d.). 1 1970 6 Data analyzed here are from the National Impact Survey, conducted in 1968—69, which represented the first attempt in Pakistan to collect social and demographic data at the national level. The survey covered about 2,500 households. Interviews were conducted with 2,910 currently married women under 45 years of age. (See Pakistan, Family Planning Council, n.d., for details.) Care was taken to record accurately the ages of women at the time they started living with their husbands instead of the ages at which they were married. Nevertheless, the estimates of exact age at marriage may still have been biased by somewhat inaccurate age reporting as a result of memory lapses, especially among older women. Similarly, there were recall problems with answers concerning the total number of live births. As a result, some underreporting of cumulative fertility among older women was likely. Because of the possibility that reports of current age, age at marriage, and number of live births were more accurate among urban than among rural women, only data for urban residents are analyzed here. Urban women are frequently able to relate vital events to some wellknown and important occurrence in their lives owing to their greater exposure to the mass media and their relatively higher literacy. They may therefore provide more reliable information than their rural counterparts. The sample for this study consists of 1,114 currently married women of ages 15-44 who had married only once and were living in urban areas of Pakistan. Socioeconomic and demographic characteristics of the women selected for study are current age, level of education, husband's occupation, type of residence (metropolitan or nonmetropolitan). and mother tongue (as a proxy for ethnicity) because they are considered to be most likely to affect the relationship between age at marriage and cumulative fertility. In a multiple classification analysis, duration of marriage and first pregnancy interval are used as covariates to determine their effects on the relationship between age at marriage and cumulative fertility within three age groups. FINDINGS A N D DISCUSSION Relationship of fertility to demographic and socioeconomic characteristics An inverse relationship between age at marriage and fertility is indicated by the cumulative fertility measures in Tables 2 and 3, though 1 Respondents were asked: "What was your age when you first lived with your husband?" and number of years was recorded. Interviewers were instructed to probe for an estimate of each respondent's age using a date/event card and to specify the basis of each estimate. T A B L E 2 Age-specific and total fertility rates, mean number of live births, and mean number of pregnancies, by current age of women: urban Pakistan, 1968—69 Index of fertility Current age of women by age at marriage 15-19 2 0 - 24 25-29 30-34 35-39 40-44 All ages Age-specific fertility rale All ages at marriage 16 or less 17-18 19 or more .268 .281 .250 0 .395 .402 .420 .359 .322 .337 .253 .281 .200 .337 .225 .200 .181 .179 .192 .190 .095 .103 .046 .111 269.3 272.1 248.8 281.9 Mean number of live births All ages at marriage 16 or less 17-18 19 or more .6 .7 .4 0 2.0 2.6 1.8 1.0 3.6 4.1 6.3 6.6 6.8 4.1 7.1 7.4 6.5 5.9 4.2 4.7 3.1 2.3 5.1 5.4 5.0 4.1 Mean number of pregnancies All ages at marriage 16 or less 17-18 19 or more .7 .8 .5 0 2.3 2.9 2.1 1.2 4.0 4.6 3.4. 2.6 5.6 6.0 5.3 4.5 7.0 7.4 7.2 4.8 7.7 8.2 7.0 6.3 4.7 5.2 4.2 3.4 Number of women 112 215 273 201 187 126 1,114 Percentage 10.1 19.3 24.5 18.0' 16.8 11.3 100.0 SOU RCE: Pakistan, Family Planning Council (n.d.), for this and subsequent tables. 3.9 2.9 8 T A B L E 3 Total marital fertility rate, mean number of live births, mean number of pregnancies, and mean first pregnancy interval, by age at marriage: Pakistan, 1968—69 Fertility ratio (for specified current age groups) Age at marriage 16 or less 17—18 19 or more All ages Marital total fertility rate (15-44) 7.9 6.7 6.0 7.6 Mean number of live births (40-44) 7.4 6.5 5.9 7.1 Mean number of pregnancies (40-44) 8.2 7.0 6.3 7.7 Mean interval between marriage and first conception, in months (15-44) 19.7 13.3 9.6 16.7 Number of women 724 202 188 1,114 Percentage 65.0 18.1 16.9 100.0 the age-specific rates are irregular. An average of 1.5 additional live births or an average of 1.9 additional pregnancies was reported by . women aged 4 0 - 4 4 who married at age 16 or earlier as compared with those married at age 19 or later (Table 2). Similarly, there was a difference of 1.9 in the total marital fertility rate between women who married earlier and those marrying at later ages (Table 3). The findings indicate a reduction of 20 percent in the number of live births (in the age group 40-44), 23 percent in total pregnancies (in the same age group), and 24 percent in the marital total fertility rate among women married at age 19 or later compared with those married by age 16. However, in accordance with earlier findings (Rele, 1962), women marrying later tended, on the average, to have shorter first pregnancy intervals, averaging 9.6 months compared with 19.7 months for those marrying earlier. This difference of 10.1 months may be due partly to secondary sterility associated with early age at marriage. Thus, although the women in the sample tended to follow the large-family norm prevalent in Pakistani society irrespective of their age at marriage, those marrying later eventually ended up with an average of 1.5 fewer children than women marrying earlier. As will be shown, this is largely a result of a shortening of their reproductive period by an average of three years due to later marriage. Table 4 presents the mean number of live births for women by 9 selected demographic and socioeconomic characteristics within three broad age groups (15-24, 25-34, and 35-44 years) to allow for possible cohort effects. Age at marriage is shown to be an important predictor of fertility among women in each age group, and older women who reported marriage by age 16 show an excess of two live births over those married at age 19 or later. This finding suggests that, other things being equal, the average completed family size of the former group would have been 4.9 instead of 6.9 if their average age at marriage had been 19 instead of 16 years. In other words, those married at age 19 or later had 29 percent lower cumulative fertility than women married at age 16 or earlier. But women married at ages 1 7 and 18 had almost the same cumulative fertility as those married earlier, suggesting that age 19 may be a cutoff point in predicting fertility decline due to marriage delay. As seen in Table 4, educational level influenced age at marriage and fertility. Women with more education who married at age 19 or later had, on the average, 2.7 fewer live births than those who married at age 16 or earlier and had three or less years of schooling. Similarly, women married to men in higher status occupations (professional and managerial) tended to be less fertile if married at a later age than if married earlier (the difference being 3.1 children). But wives of unskilled workers did not always have the highest cumulative fertility. This finding suggests that economic status has a complex influence on the cumulative fertility of older women, who tend to have the highest fertility. Metropolitan versus nonmetropolitan residence may be one of the factors aiding changes in attitudes and thus leading to a higher mean age at marriage and to reductions in family size. Though women living in metropolitan areas reported higher cumulative fertility, those married at age 19 or later tended to have the lowest cumulative fertility: about two children fewer than the average woman and 2.7 fewer than their counterparts who married at age 16 or earlier. Ethnic groups in Pakistan differ not only in the languages they speak; they also have their own traditions and subcultures, including different marriage practices and customs. Ethnic groups speaking Urdu and Punjabi are largely cosmopolitan and have lived for centuries close to Hindu (Indian) culture. Consequently, many of their traditions and customs are Hindu in origin. On the other hand, the Sindhi, Pushto, and Baluchi speaking populations are more traditional and closer to the Arab and Iranian cultures. Arabs entered the Indian subcontinent through Sind in the seventh century A.D., and the first foundation of T A B L E 4 Mean number of live births by age at marriage and selected characteristics: currently married women in urban Pakistan To women married at age: To all women women marriage women <16 17—18 >19 Difference between mean number of live births for women married <16and>19 ([5]-[7]) (1) (2) (3) (4) (5) (6) (7) (8) (9) WOMEN 35-44 YEARS 313 100.0 15.9 6.6 6.9 6.7 4.9 2.0 29.0 Woman's education level <3 years of schooling >4 years of schooling 268 45 85.6 14.4 15.7 17.3 6.8 5.8 7.0 6.5 6.8 5.9 5.2 4.3 1.8 2.2 25.7 33.9 81 99 95 29.5 36.0 34.5 15.9 15.7 16.5 7.0 6.7 6.5 7.1 7.0 7.0 6.4 6.1 7.4 7.0 5.3 3.9 0.1 1.7 3.1 1.4 24.3 44.3 Residence Non metropolitan Metropolitan 229 84 73.2 26.8 15.7 16.4 6.5 6.8 6.7 7.4 6.5 7.2 5.1 4.7 1.6 2.7 23.9 36.5 Language Urdu Punjabi Sindhi, Pushto, Baluchi 98 147 65 31.6 47.4 21.0 15.6 15.9 16.2 6.9 6.8 6.0 7.1 7.2 5.9 6.9 6.2 7.5 5.1 4.6 5.4 2.0 2.6 0.5 28.2 36.1 8.5 WOMEN 25-34 YEARS 474 100.0 16.4 4.2 4.6 4.0 3.0 1.6 34.8 Woman's education level <3 years of schooling >4 years of schooling 391 83 82.5 17.5 16.0 17.9 4.2 4.1 4.7 4.9 3.9 4.4 3.0 3.2 1.7 1.7 36.2 34.7 Number of Age group and characteristics Husband's occupation Unskilled Clerical or skilled Professional or managerial Percentage Mean of age at Mean number of live births a Husband's occupation Unskilled Clerical or skilled Professional or managerial 125 172 136 28.9 39.7 31.4 15.7 16.4 16.7 4.3 4.4 4.3 4.8 4.9 4.5 3.6 4.0 4.1 2.1 3.2 3.7 2.7 1.7 0.8 62.8 34.7 17.8 Residence Nonmetropolitan Metropolitan 346 128 73.0 27.0 16.4 16.1 4.2 4.4 4.6 4.9 3.9 4.1 3.0 3.1 1.6 1.8 34.8 36.7 Language Urdu Punjabi Sindhi, Pushto, Baluchi 153 234 86 32.3 49.5 18.2 16.1 16.5 16.3 4.4 4.3 3.7 5.0 4.8 3.9 3.7 4.2 3.9 3.0 3.0 2.9 2.0 1.8 1.0 30.0 37.5 25.6 WOMEN 15—19 YEARS 327 100.0 16.1 1.5 1.8 1.4 0.9 0.9 50.0 Woman's education level <3 years of schooling >4 years of schooling 250 77 76.5 23.5 15.7 17.6 1.6 1.4 1.7 2.0 1.4 1.4 0.9 0.9 0.8 1.1 47.1 55.7 Husband's occupation Unskilled Clerical or skilled Professional or managerial 88 126 82 29.7 42.6 27.7 15.5 16.2 16.5 1.5 1.6 1.6 1.5 1.9 2.0 1.8 1.3 1.3 1.1 1.0 1.1 0.4 0.9 0.9 26.7 47.4 45.0 Residence Nonmetropolitan Metropolitan 257 70 78.6 21.4 16.1 16.2 1.5 1.5 1.8 1.7 1.5 1.1 0.8 1.3 1.0 0.4 55.6 23.5 Language Urdu Punjabi Sindhi, Pushto, Baluchi 101 134 89 31.2 41.3 27.5 15.8 16.9 15.5 1.6 1.6 1.3 1.9 1.9 1.4 1.2 1.5 1.4 0.8 1.1 0.2 0.9 0.8 1.2 59.7 42.1 85.7 NOTE: 110 women either failed to report their husband's occupation or reported him to be a farmer. They are therefore not included in the analysis of husband's occupation as an independent variable. a Percentage difference shown in col. 9 equals col. 8 divided by col. 5 X 100. 12 Arab and Islamic culture was laid there. Pushto and Baluchi speaking areas border on Iranian and Afghani areas, and people living there trace their origins to central Asia. One marriage custom widely practiced by the Sindhi, Pushto, and Baluchi groups is payment of a bride price. At the time of marriage negotiations, a bridegroom's family has to offer a bride price to the family of the bride as a security against divorce. On the other hand, among the Urdu and Punjabi speaking populations, who have a long affiliation and common ancestry with Hindus, it is customary for the bride to bring a dowry to the marriage from her family. This contrast in marriage practices may have a strong bearing on age at marriage and in turn on fertility within each marriage cohort. Though the differences in the mean age at marriage among the three ethnic groups do not seem to be great, a somewhat lower cumulative fertility is noticed among those speaking Sindhi, Pushto, and Baluchi. Little difference was reported between women married at earlier and at later ages, though the very high fertility of those married at age 17 or 18 is quite conspicuous. Punjabi speaking women reported 2.6 fewer live births if married at age 19 or later than if married at the earliest ages. Among Urdu speaking women, those marrying at age 19 or later also reported lower fertility than those marrying at age 16 and earlier. A similar trend was observed for the youngest cohort in this language group. Multiple classification analysis So far we have looked at the differences in the mean number of live births among women in three age categories, married at different ages and having certain socioeconomic characteristics. In what follows, a closer look is taken at these relationships with the help of multiple classification analysis (MCA). Two demographic variables are introduced as covariates. A n important aspect of the M C A technique is its capacity to show the effect of an independent variable, or predictor, on the dependent variable both before and after the effects of selected other variables, used as covariates, are taken into account. One useful statistic is the adjusted deviation from the grand mean (or the adjusted mean) for each category of the predictor. The adjusted means are estimates of wjiat the mean levels of the dependent variable would have been if the subjects under study had been equally distributed according to the control variables (Andrews et al., 1973). Cumulative fertility is a function not only of the current age of a woman and her age at marriage, but also of the span of her reproductive period. In the analy- 13 sis that follows the reproductive span is indexed by duration of marriage and first pregnancy interval. These variables are themselves only weakly correlated (see Table 5). T A B L E 5 Correlation matrix of socioeconomic and demographic characteristics, for currently married women aged 15—44 in urban Pakistan: 1968-69 Variable (9) 1 Age at marriage 2 Number of live births 3 Current age .32 .13 .08 .01 -.24 -.31 .01 -.09 0 -.09 .73 .05 -.06 .03 -.04 .19 .92 .06 -.16 - 0 2 -.07 .27 .06 - 1 7 -.10 -.06 -.04 .11 -.01 -.19 -02 .01 .24 4 Duration of marriage 5 First pregnancy interval 6 Residence 7 Linguistic group 8 Husband's occupation 9 Woman's education (8) (7) (6) (3) (4) (5) .02 .70 (2) (1) -.21 Table 6 indicates the clear downward trend in fertility with age at marriage among younger women (col. 4). But when duration of marriage is introduced as a covariate in the MCA analysis (col. 5), the fertility difference between late-marrying and all women is reduced by one live birth. Similar findings are reported for the other two groups of younger women. On the other hand, the effect of controlling on the first pregnancy interval is to increase the age-at-marriage differential (col. 6). When the additive effects of duration of marriage and first pregnancy interval are considered simultaneously, 15 percent of the variance in number of live births is explained. This finding reflects the fact that there is some association between the two variables and a corresponding overlap in explanatory power, because R with separate control for duration of marriage (.095) and for first pregnancy interval (. 105) sum to a higher value than the combined R when one takes into account both variables simultaneously. It may easily be seen that duration of marriage is an important factor in explaining the relationship between fertility and age at marriage, but first pregnancy interval does not affect this relationship to a great extent. However, duration of marriage and the first pregnancy interval are most important for the early years of childbearing, as shown in the third panel of Table 6. 2 2 Using duration of marriage and first pregnancy interval as constants, T A B L E 6 Multiple classification analysis of number of live births, using age at marriage as a predictor and duration of marriage (A) and first pregnancy interval (B) as covariates, for three age groups of currently married women in urban Pakistan, 1968—69 Current age and age at marriage Number 0) Mean number of live births (grand or catePercentage gory mean) 13) U) Current age 35—44 Age at marriage 16 or less 17-18 19 or more Eta or beta 313 103.0 6.62 227 47 39 75.5 15.0 12.5 6.90 6.66 4.92 474 99.9 4.23 296 85 93 62.4 17.9 19.6 4.68 3.97 3.03 2 R Current age 25-34 Age at marriage 16 or less 17-18 19 or more Eta or beta 2 R Current age 15-24 Age at marriage 16 or less 17-18 19 or more Eta or beta R 327 100.0 1.53 201 70 56 61.5 21.4 17.1 1.76 1.40 .91 2 NOTE: Percentages may not sum to 100 because of rounding. Unadjusted deviation from grand mean (4) Deviation adjusted for: (A) (5) (B) (6) (A) + (B) (7) .28 .04 -1.70 .22 .048 .09 .21 -.75 .10 .095 .34 -.11 -1.87 .24 .105 .15 .06 -.93 .12 .151 .45 -.26 -1.20 .30 .093 -10 .04 .27 .07 .276 .53 -33 -1.40 .36 .131 -.04 -02 .14 .03 .353 * .22 -.13 -.62 .22 .050 -.08 .11 .15 .07 .451 .21 -.12 -.59 .21 .052 -.07 .11 .11 .06 .479 15 Figure 2 shows the effect of age at marriage on cumulative fertility. It plots the mean number of predicted live births for each age group and age-at-marriage category as if the women had had identical marriage durations and first pregnancy intervals. Whereas, among older women, the unadjusted difference in the mean number of live births is 29 percent, this difference is reduced to 14 percent after duration of marriage is accounted for. It rises slightly (to 32 percent) when first pregnancy interval is held constant. The combined effect of both variables indicates a difference of 16 percent in cumulative fertility between women married at age 16 or earlier and those married at age 19 or later. The data also indicate that when duration of marriage is held constant, there is virtually no difference in cumulative fertility among younger women at whatever age they married. Rather, women who married later tended to show higher cumulative fertility during their early reproductive period, which may be due to their shorter first and subsequent pregnancy intervals. Nevertheless, women who married at age 19 or later ultimately ended up with a completed family size of about one child less than those married at age 18 or younger. The overall lower fertility among women who married at older ages, to the magnitude of 29 percent (and of 14 percent when duration of marriage is held constant), suggests.that if female age at marriage in Pakistan were to increase to about 21 years, a substantial decline in fertility would occur, as has happened in other countries, such as the Republic of Korea (Kim et al., 1974). Because a very high proportion of women in our sample were married at age 16, an analysis of completed family size of women who married at age 21 or later has not been attempted. But age at marriage seems to be rising; whereas 76 percent of the older women (35—44) were married by age 16, only 62 percent of the younger women (15—34) were married by that age. Though the future fertility trend of younger women may not be correctly predicted, it is evident that younger women who postponed marriage are experiencing lower fertility than those who married at early ages. Thus, because a higher proportion of younger women are marrying at later ages, the cumulative fertility of the younger cohorts may ultimately be less than 6.6 births per woman, the rate reported for older women who married at early ages. Contraception may also play an important role in reducing cumulative fertility, even among younger women who have still not adopted any family planning method (Pakistan, Population Planning Council, 1976). Earlier in this paper I noted a much lower cumulative fertility rate among women who married at age 19 or later, who had some higher F I G U R E 2 M e a n n u m b e r o f l i v e b i r t h s b y age at m a r r i a g e , u n a d j u s t e d (A) a n d a d j u s t e d f o r t h e e f f e c t o f d u r a t i o n o f m a r r i a g e ( B ) , f i r s t p r e g n a n c y i n t e r v a l (C), a n d c o m b i n e d e f f e c t o f ( B ) a n d (C): c u r r e n t l y married w o m e n aged 15—44 in u r b a n P a k i s t a n S O U R C E : Table 6. 17 education, and whose husbands were in professional and managerial occupations, than among those who were classified in the same educational and occupational categories but had married at age 16 or earlier (Table 4). Table 7 presents the MCA results of five combinations of women's educational level and husband's occupation. Very few women with husbands in unskilled occupations had more than three years of schooling, so they were combined with those whose husbands were in clerical and skilled occupations. The five categories are treated as a rough scale of socioeconomic status (SES) from low to high. As expected, women with higher SES and in the oldest age group tended to have 1.5 (or 21 percent) fewer live births than those with the lowest SES, and 1.3 fewer than the average woman in that age group. With age at marriage held constant, however, the difference is reduced to 0.9 live births (or 13 percent). Exactly the same results (0.9 fewer live births) are obtained by controlling for duration of marriage. But a slight increase (to 1.6 fewer live births with higher SES) is noticed by taking into account first pregnancy interval. The findings suggest that age at marriage and duration of marriage are both important factors in explaining fertility differentials among different socioeconomic groups. Seventeen percent of the variance is explained by the combination of the three demographic variables among the oldest group of women. There is a substantial increase in the explanation of variance (to 37 percent) for women aged 25—34 and 49 percent for those younger than 25 years. But the results for the younger women must still be interpreted with caution because they have not yet completed their families and their future fertility behavior is not known. Table 4 provided some evidence of a lower cumulative fertility, to the magnitude of 0.9 live births, within the Sindhi, Pushto, and Baluchi language groups as compared with those speaking Urdu and Punjabi. Yet the latter groups reported much lower cumulative fertility among women who postponed marriage. This difference may be explained by whether current residence was in metropolitan or nonmetropolitan areas, as shown in Table 8. Because of the fairly small numbers of women of the Sindhi, Pushto, and Baluchi language groups residing in metropolitan areas, they are all treated here as nonmetropolitan. In any case, those residing in metropolitan areas would have recently moved there, for the two largest cities of Karachi and Lahore are largely inhabited by Urdu and Punjabi speaking populations. That the highest cumulative fertility (7.8 births) was reported for Urdu speaking metropolitan women may be due to the fact that this T A B L E 7 Multiple classification analysis of number of live births, using socioeconomic status as a predictor and age at marriage (A), duration of marriage (B), and first pregnancy interval (C) as covariates, for three age groups of currently married women in urban Pakistan, 1968—69 Current age, years of schooling, and husband's occupation Number (D Mean number of live births Per(grand or catcentage egory mean) (2) (3) Current age 35—44 <3, unskilled <3, clerical/skilled <3, professional/managerial >4, unskilled/clerical/skilled >4, professional/managerial Eta or beta R Current age 25—34 <3, unskilled <3, clerical/skilled <3, professional/managerial >4, unskilled/clerical/skilled ^4, professional/managerial Eta or beta R 275 77 84 76 19 19 100.0 28.0 30/6 27.6 6.9 6.9 6.73 6.93 6.84 6.75 6.63 5.47 433 117 141 99 39 37 100.0 27.0 32.6 22.9 9.0 8.5 4.33 4.29 4.25 4.27 4.25 3.92 296 84 96 52 34 30 100.0 28.4 32.4 17.6 11.5 10.1 1.59 1.53 1.54 1.75 1.74 1.47 2 2 Current age 15—24 <3, unskilled <3 clerical/skilled <3, professional/managerial >4, unskilled/clerical/skilled >4, professional/managerial Eta or beta R ( 2 Unadjusted deviation from grand mean (4) Deviation adjus ted for: A (5) B) (6) (C) (7) (A)+(B)+(C) (8) .20 .11 .02 -.10 -1.26 .12 .014 .15 -.02 .03 .13 -.76 .07 .056 .15 .03 -.02 .06 -74 .07 .086 .24 .14 .01 -.31 -1.31 .13 .077 .18 .01 -.02 -.08 -.61 .06 .167 -.04 .08 .06 .08 -41 .06 .004 -.20 .04 .02 .35 .06 .07 .120 -.20 .02 .03 .26 .17 .07 .284 0 .10 .03 .03 -.49 .07 .018 -.10 .06 -06 .16 .08 .04 .365 -.06 -.05 .16 .15 -.12 .07 .005 -.18 -.08 .11 .36 .18 .12 .071 -.27 -.05 .27 .22 .22 .15 .460 -.11 -.04 .17 .17 -07 .08 .013 -.18 -.07 .29 .13 .09 .12 .492 T A B L E 8 Multiple classification analysis of number of live births, using current residence and language as a predictor and age at marriage (A), duration of marriage (B), and first pregnancy interval (C) as covariates, for three age groups of currently married women in urban Pakistan, 1968-69 Current age, current residence, and language Number (D Mean number of live births Per(grand or catcentage egory mean) (2) (3) Current age 35—44 310 100.0 Nonmetropolitan, Urdu Metropolitan, Urdu Nonmetropolitan, Punjabi Metropolitan, Punjabi Nonmetropolitan, Sindhi, Pushto, Baluchi Eta or beta 2 (C) (7) (A)+(B)+(C) (8) 6.65 6.23 7.83 6.91 6.35 -.42 1.18 .26 -.30 -49 1.15 .23 -.19 -.58 .87 .34 -.15 -29 1.22 .24 -.42 -.45 .94 .29 -.25 65 21.0 5.99 -66 -.58 -.54 -.70 -.57 .20 .19 .17 .20 .17 .039 .070 .114 .091 .172 .19 .25 .01 .21 .11 .21 .09 .17 .09 .01 .07 .25 .21 .27 0 .18 -.56 -.42 -.54 2 R (B) (6) 18.4 13.2 36.1 11.3 R Nonmetropolitan, Urdu Metropolitan, Urdu Nonmetropolitan, Punjabi Metropolitan, Punjabi Nonmetropolitan, Sindhi, Pushto, Baluchi (A) (5) Deviation adjusted Tor: 57 41 112 35 Eta or beta Current age 25-34 Unadjusted deviation from grand mean (4) 473 100.0 4.72 92 61 177 57 19.5 12.9 37.4 12.0 4.41 4.47 4.23 4.43 86 18.2 3.67 -.55 .13 .02 .05 .16 -.36 .13 .13 .10 .13 .08 .017 .142 .284 .027 .360 TABLE 8 (continued) Current age, current residence, and language Number (1) Mean number of live births Per(grand or catcentage egory mean) (2) (3) Current age 15-24 324 100.0 1.53 69 32 102 32 21.3 9.9 31.5 9.9 1.78 1.31 1.51 1.78 .25 -.22 -.02 .25 .19 -.32 .13 .39 .02 -.29 .11 .37 .24 -.20 -.01 .29 .03 -.31 .11 .30 89 27.4 1.34 -.19 -32 -.17 -.20 -.15 .13 .18 .13 .13 .12 .018 .098 .472 .027 .503 Nonmetropolitan, Urdu Metropolitan, Urdu Nonmetropolitan, Punjabi Metropolitan, Punjabi Nonmetropolitan, Sindhi, Pushto, Baluchi Eta or beta R 2 Unadjusted deviation from grand mean (4) Deviation adjusted tor: (A) (5) (B) (6) (C) (7) (A)+(B)+(C) (8) 21 group does not have the characteristics typical of a metropolitan population. Most of the women in this group married a generation ago, were new immigrants from central India having either rural or nonmetropolitan backgrounds, and came from traditionally large families (Driver, 1963). Moving to the city may have improved their living level but not affected their traditional fertility values, thus providing the condition for higher fertility. Moreover, with better medical facilities in the nation's capital than in rural areas, the chances of pregnancy loss were smaller for this group of women. Finally, as mentioned earlier, women in the metropolitan areas tend to have fewer recall problems when reporting the total number of live births. On the other hand, fewer medical facilities have been available to the other groups, particularly to the Sindhi, Pushto, and Baluchi language groups, who might have had more incidents of pregnancy loss and thus may have reported fewer live births. 2 The substantial difference in cumulative fertility among ethnic and residential groups is largely reduced among women in the middle age group (24—34), the first four groups reporting almost the same cumulative fertility. In the youngest age group (15-24), the fertility pattern is opposite to that of the oldest group; Urdu speaking metropolitan women reported the lowest fertility. They were followed by the nonmetropolitan Sindhi, Pushto, and Baluchi speaking groups; and highest fertility was reported by the nonmetropolitan Urdu and metropolitan Punjabi speaking groups. This finding suggests that if the present trend continues, Urdu speaking women in the largest city (Karachi) will eventually have lower cumulative fertility than women of other ethnic groups and areas. The apparent fertility decline among Urdu speaking metropolitan women may be due to a combination of factors affecting female age at marriage as well as fertility, such as better educational facilities, a higher standard of living, a higher proportion of women in the nonagricultural labor force, the housing shortage, greater exposure to the mass media, and the concentration of family planning activities during more than two decades. With an increase in female age at marriage, the enrollment in schools above the primary level, higher labor force participation in nonagricultural occupations, and a wider use of contraception, a decline in fertility among Pakistani women might be expected in the near future. This is indicated by the lower cumulative fertility of younger women, who married later, completed four or more years of schooling, and were living in metropolitan areas. 2 Karachi was Pakistan's capital until 1961. 22 SUMMARY AND CONCLUSIONS The well-established inverse relationship between age at marriage and cumulative fertility has been examined in this paper with data from the 1968—69 Pakistan National Impact Survey for currently married women 15-44 years of age who were living in urban areas. These women reported marrying at the very low mean age of 16 years, although among recently married women age at marriage was slightly higher (17.5). As a result of their lower age at marriage, women aged 35—44 at the time of the survey had a high cumulative fertility rate of 6.7 births. Twenty-nine percent lower cumulative fertility was reported by women who married at age 19 or later, compared with those married at age 16 or earlier. The latter group bore an average of 1.5 additional children, reported 1.9 additional pregnancies, and had a 1.7 higher total marital fertility rate. However, women who postponed marriage tended to reproduce faster, reporting a first pregnancy interval ten months shorter than that of women who married at earlier ages. Women with four or more years of schooling and husbands in professional and managerial occupations who married late had the lowest cumulative fertility. Cumulative fertility differentials among ethnic groups and between metropolitan and nonmetropolitan residents were also found among women who married at younger and older ages. The M C A results indicate that duration of marriage and first pregnancy interval are important predictors in explaining the relationship between age at marriage and cumulative fertility. After taking into account duration of marriage, I found cumulative fertility to be 14 percent lower among women who postponed marriage compared with those who married at early ages. When the first pregnancy interval was controlled, the cumulative fertility was 32 percent lower for late marrying women. Socioeconomic status appears to have had a strong bearing on cumulative fertility; the relationship was also affected by age at marriage and duration of marriage. Older, Urdu speaking women living in metropolitan areas had the highest average cumulative fertility; but younger, Urdu speaking women living in metropolitan areas reported the lowest cumulative fertility of all ethnic groups. Age at marriage and duration of marriage were again important predictors of cumulative fertility for each of the categories of social and economic characteristics. Though the data revealed an inverse relationship between cumulative fertility and age at marriage, it should be kept in mind that demographic variables are affected by social, economic, cultural, and other conditions, in Pakistan as elsewhere. These conditions have been 23 changing rapidly in recent years as a result of socioeconomic development, particularly in urban areas. Because the findings reported here mainly apply to older women who were not affected by this rapid change during their reproductive years, more research is needed on the impact.of these changes on younger age cohorts. Female age at marriage should be treated as an important aspect of population policy in Pakistan, where an integrated approach is needed to reduce high fertility. Increasing the legal female age at marriage from 16 to 19 years or more would have both social and demographic benefits. An average female age at marriage of 17 years is too low for a country in which marriage is universal. The values supportive of lower age at marriage have deeply rooted social and economic bases. Girls are considered a liability, for they are not expected to be economically productive; therefore parents desire them to marry early. It seems apparent that not only must the legal age at marriage be increased, but also basic changes in women's roles must be introduced if a decline in fertility is to take place in Pakistan. 24 REFERENCES Agarwala, S.N. 1967 Effect of a rise in female marriage age on birth rate in India. Proceedings, World Population Conference, 1965, vol. 2. New York: United Nations. Ahmed, Feroz 1969 The effect of raising age at marriage on fertility in Pakistan. Paper presented at the Annual Meeting of the Population Association of America, Atlantic City. Andrews, Frank M., James N. Morgan, and John A. Sonquist 1973 Multiple Classification Analysis. Ann Arbor: Survey Research Center, University of Michigan. Busfield, Joan 1972 Age at marriage and family size: social causation and social selection hypotheses. Journal of Biosocial Science 4:117-34. Coale, Ansley J., and C.Y. Tye 1961 The significance of age patterns of marriage on fertility in high fertility populations. MilbankMemorial Fund Quarterly 39:631 — 66. Das, Nitai Chandra 1967 A note on the effects of postponement of marriage on fertility. Proceedings, World Population Conference, 1965, vol. 2. New York: United Nations. Davis, Kingsley 1955 Institutional patterns favoring high fertility in underdeveloped areas. Eugenics Quarterly 2:33—39. De Tray, Dennis N. 1977 Age of marriage and fertility: a policy review. Pakistan Development Review 16:89-100. Driver, Edwin D. 1963 Differential Fertility in Central India. Princeton: Princeton University Press. 25 Glass, David V., and E. Grebenik 1954 The trend and pattern of fertility in Great Britain. Papers of the Royal Commission of Population, VI. London: HM Stationery Office. Karim, Mehtab S. 1974 .Fertility differentials by family type. Pakistan Development Review 13:130-44. 1979 Nuptiality transition in Pakistan and its correlates. Unpublished paper, International Statistical Institute, World Fertility Survey, London. Kim, Mo-Im, Rowland V. Rider, Paul A. Harper, and Jae-Mo Yang 1974 Age at marriage, family planning practices, and other variables as correlates of fertility in Korea. Demography 11:641-56. Leasure, J. William 1963 Malthus, marriage and multiplication. Milbank Memorial Fund Quarterly 41:419-35. Malakar,C.R. 1972 Female age at marriage and the birth rate in India. Social Biology 19:297-301. Palmore, James A., and Ariffin bin Marzuki 1969 Marriage patterns and cumulative fertility in West Malaysia: 1966-67. Demography 6:383-402. Pakistan, Census Organization 1974 Statistical Tables 1972. Pakistan Islamabad (Mimeographed). Pakistan, Family Planning Council n.d. National Impact Survey Report. Lahore: Training, Research and Evaluation Center. Pakistan, Population Planning Council 1976 Pakistan Fertility Survey: First Report. Islamabad. Pakistan, Statistical Division 1973 Population Growth Survey: 1968. Karachi. Rele, J.R. 1962 Some aspects of family and fertility in India. Population Studies 15:267-78. 26 Sadiq, Nasim M. 1965 Estimation of nuptiality and its analysis from the census data of Pakistan. Pakistan Development Review 5:229—48. Siffman, R.I. 1967 Age at marriage as a demographic factor in conditions of high fertility. Proceedings, World Population Conference, 1965, vol. 2. New York: United Nations. Smith, Peter C , and Mehtab S. Karim 1978 Urbanization, education and nuptiality in four Asian societies. Paper presented at the Conference on Comparative Fertility Transition in Asia, Tokyo, March 27-31. Talwar, P.P., andK.C. Seal 1974 Measurement of effect on fertility of shifts in age-at-marriage: a case study of India. Demography India 3:367-74. R E C E N T A V A I L A B L E PAPERS O F T H E EAST-WEST P O P U L A T I O N INSTITUTE No. 41 A method o f decomposing urban population growth and an application t o Philippine data, by Ernesto M . Pernia, December 1976, 26 + v pp. 42 Methodological difficulties encountered in using own-children data: illustrations f r o m the United States, by Ronald R. Rindfuss, February 1 9 7 7 , 1 7 + v pp. 43 The fertility of migrants to urban places in Thailand, by Sidney Goldstein and Penporn Tirasawat, April 1977, 49 + v pp. 44 The demographic situation in the Philippines: an assessment in 1977, by Mercedes B . Concepcion and Peter C. Smith, June 1977,75 + vii pp. 45 T h e demographic situation in Thailand, by Fred A r n o l d , Robert D . Retherford, and A n u r i Wanglee, July 1977, 35 + vii pp. 46 The role o f migration and population distribution in Japan's demographic transition, by Toshio K u r o d a , July 1 9 7 7 , 1 7 + v pp. 47 The recent fertility decline in the Chiang Mai area of Thailand, by Tieng Pardthaisong, February 1978, 36 + vii pp. 48 Spatial analysis of f a m i l y planning program effects in Taiwan, 1966—72, by Albert I. Hermalin, A p r i l 1978, 39 + vii pp. 49 Gains f r o m population control: results f r o m an econometric model, by Daniel B . Suits and Andrew Mason, A p r i l 1978, 22 + v pp. 50 The economic value o f children in Asia and A f r i c a : comparative perspectives, by Helen Ware, A p r i l 1978, 36 + v pp. 51 Rural-urban migration and social mobility: studies o f three South Korean cities, by Man-Gap Lee and Herbert R. Barringer, May 1978, 4 4 + vii pp. 52 Preliminary estimates of Indonesian f e r t i l i t y based on the 1976 Intercensal Population Survey, by Sam Suharto and Lee-Jay C h o , May 1978,21 + v pp. 53 Circulation in the context of total mobility in Southeast Asia, by Sidney Goldstein, August 1978, 69 + v pp. 54 Effects of program contraception o n fertility: a comparison o f three Asian countries, by Siew-Ean K h o o , September 1 9 7 8 , 5 8 + ix pp. 55 Population projections f o r planning and policy, by William Brass, September 1978, 16 + v pp. 56 Spatial fertility analysis in a limited data situation: the case of Pakistan, by Gary Fuller and Mohammad M . K h a n , O c t o b e r 1978, 2 0 + vii pp. 57 Infant and child mortality in Thailand: levels, trends, and differentials as derived through indirect estimation techniques, by John Knodel and Apichat Chamratrithirong, November 1978, 40 + vii pp. 58 Regression estimates of changes in fertility, 1955—60 to 1965—75, f o r most major nations and territories, by James A . Palmore, December 1978, 59 + vii pp. 59 Comparison of three acceptance strategies: a progress report, by Robert G . Potter, Frances E. K o b r i n , and Raymond L . Langsten, February 1979, 16 + vii pp. 60-A O n the nature of the transition in the value of children, by R o d o l f o A . Bulatao, March 1979, 104 + xvi pp. 61 Prediction of family planning and f a m i l y size f r o m modernity value orientations of Indian women, by Bishwa Nath Mukherjee, A p r i l 1979, 50 + v pp. 62 Issues in the comparative analysis of World Fertility Survey data, by Ronald Freedman, July 1979, 22 + v pp. 60-B Further evidence of the transition in the value of children, by R o d o l f o A . Bulatao, November 1979, 8 4 + vii pp. 63 Own-children estimates of fertility f o r Thailand based on the 1970 Census, by Robert D. Retherford, Chintana Pejaranonda, Lee-Jay C h o , Apichat Chamratrithirong, and Fred A r n o l d , November 1979, 52 + vii pp. E A S T - W E S T C E N T E R - o f f i c i a l l y known as the Center f o r Cultural and Technical Interchange Between East and W e s t - i s a national educational institution established in Hawaii by the U . S . Congress in 1960 to promote better relations and understanding between the United States and the nations o f Asia and the Pacific through cooperative study, training, and research. The Center is administered by a public, nonprofit corporation whose international Board of Governors consists of distinguished scholars, business leaders, and public servants. THE Each year more than 1,500 men and women f r o m many nations and cultures participate in Center programs that seek cooperative solutions to problems of mutual consequence to East and West. Working with the Center's multidisciplinary and multicultural staff, participants include visiting scholars and researchers; leaders and professionals f r o m the academic, government, and business communities; and graduate degree students, most of whom are enrolled at the University of Hawaii. For each Center participant f r o m the United States, two participants are sought f r o m the Asian and Pacific area. Center programs are conducted by institutes addressing problems of communication, culture learning, environment and policy, population, and resource systems. A limited number of " o p e n " grants are available to degree scholars and research fellows whose academic interests are not encompassed by institute programs. The U . S . Congress provides basic funding f o r Center programs and a variety o f awards to participants. Because of the cooperative nature of Center programs, financial support and cost-sharing are also provided by Asian and Pacific governments, regional agencies, private enterprise, and foundations. The Center is on land adjacent to and provided by the University of Hawaii. T H E E A S T - W E S T P O P U L A T I O N I N S T I T U T E , established as a unit o f the East-West Center in 1969 with the assistance of a grant f r o m the Agency f o r International Development, carries out multidisciplinary research, training, and related activities in the field of population, placing emphasis on economic, social, psychological, and environmental aspects o f population problems in Asia, the Pacific, and the United States. E A S T - W E S T P O P U L A T I O N I N S T I T U T E , published about eight times a year, facilitate early dissemination of research findings related to the demography o f Asia, the Pacific, and the United States. Annual subscription rate, $12. Single copies are available without charge to organizations and individuals engaged in demographic research or programs. Requests describing the nature of the research or program and the intended use o f the publications should be addressed to the Publications O f f i c e of the Institute. PAPERS O F THE East-West Population Institute East-West Center 1 777 East-West Road Honolulu, Hawaii 96848 Director Lee-fay Cho Publications O f f i c e r Sandra E. Ward Editor Elizabeth B. Gould Production Specialist Lois M. Bender Cartographer Gregory Chu