PHN-871 0 THE IMPACT OF MODERNIZATION ON THE MOTIVATION FOR FERTILITY CONTROL, EGYPT, 1979-80 1 by Richard Easterlin Eileen M. Crimmins Mohamed Aly Ahmed Samia Mohamed Soliman May 1987 Population, Health and Nutrition Department World Bank The World Bank does not accept responsibility for the views expressed herein which are those of the author(s) and should not be attributed to the World Bank or to its affiliated organizations. The findings, interpretations, and conclusions are the results of research supported by the Bank; they do not necessarily represent official policy of the Bank. The designations employed, the presentation of material, and any maps used in this document are solely for the convenience of the reader and do not imply the expression of any opinion whatsoever on the part of the World Bank or its affiliates concerning the legal status of any country, territory, city area, or of its authorities, or concerning the deliminations of its boundaries, or national affiliation. FILE COPY PHN Technical Note 87-10 THE IMPACT OF MODERNIZATION ON THE MOTIVATION FOR FERTILITY CONTROL, EGYPT, 1979-80 ABSTRACT Because of the importance of motivation as a determinant of Egyptian contraceptive use, this paper investigates which aspects of modernization significantly affect the supply and demand components of motivation, taking account concurrently of cultural influences. The rich data set on modernization provided by the Egyptian Fertility Survey makes possible a more thorough investigation than in most inquiries using data from the World Fertility Survey. Despite this, the results are generally consistent with those obtained in more limited inquiries. Modernization variables play a much more important role in explaining differences in demographic behavior than cultural variables, and among the modernization variables, education is by far the most important. The more refined form in which the education variable is used in the present analysis has the effect of reducing or eliminating effects previously attributed to other variables such as husband's occupation. Somewhat different effects are found for husband's versus wife's education -- the former has a bigger negative impact on the demand for children, the latter, a larger effect in delaying wife's age at marriage. An income variable adds surprisingly little to the analysis, and the results suggest that effects estimated for variables such as education, urbanization, and occupation are not seriously misstated when data for income are not available. Other modernization variables that have some significant effects are farm versus nonfarm occupation, urbanization, and wife's work before marriage. Extension of the analysis to regional differences in demographic behavior shows that these reflect chiefly the influence of modernization variables. Differences in natural fertility and the supply of children are fairly small among Egyptian regions but because the more modernized regions have lower demand for children they tend to have greater contraceptive use and a correspondingly smaller number of children ever born. The chief policy implication is that an education policy focusing on commensurate expansion of literacy for both sexes is most compatible with reducing fertility. The results indicate that when both spouses are literate there is a "reinforcement effect" tending to lower fertility, over and beyond the effect attributable to literacy of either spouse taken alone. Prepared by: Richard Easterlin, Eileen M. Crimmins, Mohamed Aly Ahmed, and Samia Mohamed Soliman Consultants to the World Bank May 1987 The Impact of Modernization on the Motivation for Fertility Control, Egypt, 1979-80 by Richard A. Easterlin, Eileen M. Crimmins, Mohamed A. Ahmed, and Samia Mohamed Soliman* The purpose of this paper is to identify theoretically and empiri- cally the way in which economic and social modernization in Egypt has shaped the motivation for fertility control and thus contributed to the fertility transition, defined as the adoption of deliberate family size limitation and associated reduction in fertility. An earlier paper demonstrated the major importance of motivation for fertility control in the fertility transition (Easterlin, Crimmins, and.Khodair 1985). Economic modernization involves a sustained rise in real output per head and wide-ranging changes in techniques of producing, transporting, and distributing goods, in the scale and organization of productive activities, and in types of outputs and inputs. It also embraces major shifts in the industrial, occupational, and spatial distribution of productive resources and in the degree of exchange and monetization of the economy. Social modernization encompasses major improvements in education and public health and changing status of women (Coleman 1968; Easterlin 1968; Ki-nets 1966; Lerner 1968; Inkeles and Smith 1974). If parents do not regulate their fertility, then they may have more children than they want -- the greater the prospective number of un- * The authors are grateful for the excellent research and typing assis- tance of Donna Hokoda. Financial support was provided by the World Bank and the University of Southern California. 2 wanted children, the greater the motivation for fertility control. Viewed in this way, the motivation for fertility control is the excess of the potential "supply" of children -- the number parents would pro- duce in the absence of fertility control -- over their "demand" for children, that is, their desired family size. The supply of children depends partly on a couple's natural fertility, the number of births they would have if fertility were unregulated,, and partly on the child survival rate, the proportion of births surviving to adulthood. Differences among couples in motivation may arise from differences either in supply or demand. 0As between two couples with the same desired family size, the couple with the greater potential supply of children will be more motivated for fertility control; as between two couples with the same potential supply, the couple with the lower demand for children will be more motivated. Thus, in general, motivation for fertility control varies directly with supply and inversely with demand. The various processes of modernization -- per capita income growth, occupational change, education, urbanization, and so on -- influence both the supply of and demand for children. In general modernization is thought to raise supply and lower demand, thus increasing the motivation for fertility control, although this need not always be the case (Bongaarts and Menken 1983, Lee and Bulatao 1983). But the causal linkages between specific modernization processes and the supply and demand components of motivation are not well understood. One of the principal reasons for this is that data are not usually simultaneously available on a number of different dimensions of modernization that would permit distinguishing the separate effect of each. One of the great advantages of the data collected in the first and second phases of 3 the Egyptian Fertility Survey is that it provides an unusual combination of detailed demographic information along with extensive background data relating to many aspects of the modernization process. Thus it becomes possible to explore in depth the links between modernization and the emergence of the motivation for fertility control. This is the primary concern of this paper. The adoption of deliberate fertility control depends, not only on motivation, but also on what are termed "regulation costs" -- attitudes toward and access to contraceptive knowledge and techniques. The pre- ceding paper investigated the relative importance of motivation and regulation costs in causing the adoption of deliberate control, employ- -ing a variety of measures of regulation costs. The present paper does not pursue further the topic of regulation costs, because the cost measures available at the national level appear to be significantly affected by contraceptive use, rather than vice-versa. For example, some of the most promising cost measures from a theoretical viewpoint, such as the price of pills or time and distance required to obtain pill supplies, are chiefly reported only by pill users. Hence, comparative data for non-users are not available, and the responses for users may reflect the effect of their decision to use, rather than their knowledge at the time the decision was made. The measure of motivation for fertility control used here is a new one, and still somewhat controversial. Empirical research has shown that typically it provides a better explanation of contraceptive behavior than alternative measures commonly used in scholarly studies, and this is true of Egypt (Easterlin and Crimmins 1985; Easterlin, Crimmins, and Khodair 1985). However, different motivation measures-, 4 including the present one, are usually positively correlated. Because of this, the significance of the present paper is not confined to the present supply/demand measure of motivation for fertility control. Rather, by identifying here the way in which modernization alters the demand for and supply of children, one is, in effect, gaining insight into the processes shaping the changing motivation for fertility control as captured to a varying degree of approximation in a whole set of motivation measures.. For example, one common measure of motivation is a question on whether the respondent wants any more children. The results of the present analysis can be looked upon as clarifying the mechanisms leading couples to. want no more children, a situation usually presumed to signify the presence of motivation for fertility control. The next section provides a brief elaboration of the theoretical approach that frames the empirical inquiry, and the following presents the principal empirical results. Finally, these results plus those from the preceding paper are used to illustrate how differential moderniza- tion in the five principal regions of Egypt has led to differential motivation for fertility control and different levels of contraceptive use and fertility. Theory In the past the most common approach to identifying links between modernization and fertility has been to regress fertility (as measured, say, by children ever born) directly on measures that reflect different aspects of modernization such as education, occupation, and rural-urban residence. Sometimes other possible fertility determinants, such as 5 cultural conditions, are included. This approach is illustrated in panel A of Figure 1. The development of "proximate determinants" analysis (Davis and Blake 1956; Bongaarts 1978) has led to growing recognition of the need to insert a new stage in the sequence, so that now fertility is usually seen as determined directly by a set of "proximate determinants," with moderni-ation, in turn, operating only indirectly on fertility through these determinants (see Fig. 1, panel B; cf. also Bongaarts 1978, p. 106). The proximate determinants comprise factors such as extent of exposure to intercourse, fecundability of a couple (including the effect of frequency of intercourse), duration of postpartum infecundability, spontaneous intrauterine mortality, sterility, and use of deliberate fertility control (contraception and induced abortion). The present approach can be thought of as a further evolution of these two approaches. It singles out one subset of proximate deter- minants, that relating to deliberate fertility control, and inserts still another set of variables (the supply and demand components of motivation plus regulation costs) between deliberate control and modernization (Figure 1, panel C). As applied to the fertility transi- tion, this approach thus sees the various modernization variables as impinging directly on supply ( including the proximate determinants other than fertility control), demand, and costs of regulation. These three factors, in turn, shape the trend in use of deliberate control and fertility. The previous paper traced empirically the linkages from demand, supply, and regulation costs to the use of deliberate fertility control and fertility behavior. This paper adds to these results the linkages . 6 between the various aspects of modernization and the demand for and supply of children. To illustrate the reasoning, two aspects of moder- nization are selected for fuller discussion, expansion of formal school- ing and urbanization. As is clear from the earlier brief description of modernization, these aspects are far from exhaustive, but they should clarify the theoretical conception. The essence of the approach is captured in Table 1. In the table the modernization variables are listed on the left hand side, and the variables immediately relevant to motivation for fertility control -- demand for and supply of children -- at the top, as column headings. An entry in a cell indicates that the specified item on the left influences the variable at the top in the direction shown. For example, the nega- tive sign in column 1 of row 2 indicates that, other things being equal, the process of urbanization tends to reduce the demand for children via its effect on the strength of preferences for children relative to goods. The effects shown in Table 1 are illustrative; no attempt is made to be exhaustive. A brief sketch of the reasoning underlying the specific cell entries follows. (For a lengthier discussion, see Easterlin 1978, and Nag 1983.) Education - One of the most pervasive factors influencing the motivation for fertility control is the growth of formal education (Cochrane 1979, 1983). As shown in Table 1 it operates on both the supply of and demand for children. With regard to supply, formal education improves health conditions by diffusing improved knowledge with regard to personal hygiene, food care, environmental dangers, and so on. It may also break down traditional beliefs and customs and thus undermine cultural prac- 7 tices, such as an intercourse taboo or prolonged lactation, which have had the latent function of limiting reproduction. In these ways it tends to enhance supply by raising natural fertility and/or increasing the survival prospects of babies; hence the positive signs in columns 4 and 5 of row 1. However, a possible offsetting effect arises from the tendency of increased education to lead to later age at marriage, there- by lowering natural fertility. Formal education tends to reduce the demand for children by shift- ing tastes in a manner unfavorable to children and decreasing the price of goods relative to children (see Lindert 1983). With regard to the relative price of children (row 1, col. 3), if better education improves the income-earning possibilities of women, then the alternative cost of the mother's time required in child-rearing is increased. While some offset to this. may be available, for example, through the help of other family members or domestic workers, there is probably some net positive effect on the cost of children and thus a tendency toward a reduction in the demand for children. In addition, compulsory education may increase the relative cost of children by reducing the possible contribution of child labor to family income. Tastes for children, that is, the intensity of the desires for children relative to goods, are affected negatively by education (row 1, col. 1), because children, and the life style associated with them, are essentially an "old" good, while education presents images of new life styles competitive with children. Also, education may lead to higher standards with regard to child care and rearing, creating greater emphasis on the "quality" of children at the expense of numbers. In these ways, education increases the subjective attractiveness of life 8 styles competitive with having more children, and thus tends to lower the demand for children. Urbanization - The process of modernization requires a redistribution of population from rural to urban areas that is accomplished in part by a vast increase in rural-urban migration. Urbanization reduces the demand for children by reducing tastes and lowering the price of goods relative to children (row 2, cols. 1 and 3). The mechanism of the effect via tastes is like that for education, by promoting antinatal life-styles. With regard to costs, the relative price of children of a given "quali- ty" is usually higher in urban areas than in rural (Lindert 1980, 1983; Cochrane 1983). A variety of factors are responsible for this. The price of food is higher in urban areas than in rural. Also farm chil- dren take less time away from a mother's paid work and contribute more time toward family work than do urban children. In both cases, this would raise the relative cost of children in urban areas compared with rural. Thus, the effect of urbanization of the population is increas- ingly to place the population in an environment where goods become relatively less expensive than children, and, other things being equal, correspondingly more attractive. With regard to supply, urbanization may influence the survival rate of children. In contemporary developing countries, however, the long- standing excess of urban over rural mortality has tended to disappear with the advent of public health measures. Hence no entry appears in the survival rate cell for urbanization in Table 1. Another mechanism through which urbanization might influence potential supply is by reducing lactation. As a general matter urban women tend to breastfeed 9 for a shorter period than rural. The result, other things constant, would be to enhance the natural fertility of women (row 2, col. 4). Both education and urbanization may lead to higher family income and through this affect the supply of or demand for children. The effect of income on demand and supply, however, calls for separate discussion as another aspect of modernization influencing motivation; accordingly, only the direct effects of education and urbanization on demand and supply are considered here. These illustrations with regard to two modernization processes should make clear the present approach. A given modernization process may influence the motivation for fertility control via its impact on the demand for children, supply of children, or both. If it affects both demand and supply, the effects could be either offsetting or reinforc- ing. The subsequent empirical analysis inquires into possible effects on both demand and supply of each of a variety of modernization processes. Empirical Links between Modernization Factors, Demand,, and Supply The first part of the empirical analysis is devoted to identifying the links from modernization to the demand for and supply of children, and, thus, to the motivation for fertility control. A critical concern in this analysis is to identify as precisely as possible the specific features of modernization that play the most important causal parts, and the role of modernization processes vis-a-vis cultural and other factors in shaping the demand for and supply of children. 10 Data and hethods Data - The data are a matched sample of wives aged 35-44 and their hus- bands. Both spouses are still in their first marriage. The matching of wives and husbands enables one to obtain the maximum amount of data on a couple, and the restriction to first marriages avoids complicating the analysis with questions of the causes of marital dissolution. Dependent variables - The dependent variables are the demand for chil- dren plus the variables shaping the supply of children: Cd, desired family size X1, age at marriage, years X X2, first birth interval, months X3,.second birth interval, months X4, not secondarily sterile (=1; others=0) X5, months breastfed, last closed interval x , proportion of pregnancy wastage X , proportion of child mortality where X' through X7 determine the supply of children. For the present purpose X' replaces XI, duration of marriage, in the preceding paper, since age at marriage is the behaviorally determined component of dura- tion of marriage. The above variables as constructed from the wives' data are used here, because wives' responses were found to have a some- what closer link to fertility behavior than husbands' (Easterlin, Crimmins, and Khodair 1985). As will be seen, the supply variables divide into two groups, one of which (X',5, 7), shows sensitivity to modernization processes and the other (X2" X3, X4, X6) shows relatively little sensitivity. Hence, the first set is placed in the forefront of the supply analysis. 11 Independent variables - Several recent papers have documented the importance of modernization for demographic behavior in Egypt (Deeb and Casterline 1983, Hallouda 1983, Kelley and Schmidt 1983, Nawar and Hobcraft 1983). The present analysis includes the variables identified as important in this work plus several others. Most notable of the additions is family income, a variable constructed from the intensive economic inquiry included in the second phase of the EFS (World Fertili- ty Survey 1980). The special value of this variable is that its inclu- sion enables one to distinguish the effects of education and occupation independently from income. In contrast, in previous studies it is un- clear whether the effects attributed to, say, education, work primarily through education as such or via the effect of education on income. Specifically, the modernization variables included here are (see Appendix Table A-I for complete definitions): Wife's education, years Husband's literacy (can read=1, other=0) Urban-rural residence (urban=1) Husband's occupation (professional or clerical, manual, other nonagricultural, self-employed farmer, agricultural employee) Husband's employment status (cash wage employee=1, other=O) Wife's work experience before marriage (modern sector, traditional sector, no work experience) Wife's work experience after marriage (classified the same as for before marriage) Family income quintile. Because of their high correlation (.64), the treatment of wife's and 12 husband's education is a problem, because when both are included in multiple regressions, they tend to cancel each other. The choice here was to use the best available measure of the wife's education, because the dependent variables are constructed from the wife's responses and in theory are more likely to be affected by the wife's education. Empiri- cally, this proved almost always to be the case -- considered separately wife's education had a higher t-statistic in multiple regression analy- sis than husband's. However, to retain .the possibility of effects from husband's education, husband's literacy was included along with wife's years of schooling. The modernization variables included here are distilled from a larger list. Among those examined but found to add little or nothing to the analysis are migrant status of both husband and wife and education of wife's father. In addition to the modernization factors several other variables are regularly included in the regression analysis: age of wife age of husband blood relationship of spouses (from 3=no relationship, to 0, if spouses are first cousins) region of residence (Lower Egypt=1, Upper=O) We take the last two as capturing cultural differences among families. It is possible, of course, that they too reflect the results of moder- nization processes not captured elsewhere in the analysis, but given the wide range of modernization variables investigated, it is not clear what the omitted processes would be. Religion was also investigated as a possible cultural variable, but added nothing to the analysis. 13 Excluded from the present analysis are various attitudinal or expectation variables collected in the EFS, through which modernization processes might be expected to operate, such as responses on the age at which children are useful, aspirations for children's education, and expectations of old-age support. One might expect, for example, that the higher the parents' education, the higher would be the parents' aspirations for education of their children, and the later the age at which they thought children are useful. Although exploration of such linkages is desirable and useful (Taha and Cochrane 1984), the interest here is in trying to identify the modernization phenomena that are the "prime movers" of demographic change, whether they work via changes in people's preferences, perceptions of cost oif children, or some other channel. Methodology - For each dependent variable two OLS regressions are pre- sented. The first is on all of the variables listed above. The inter- est in this is in identifying which aspects of modernization do and which aspects do not influence the dependent variable under considera- tion, and whether the cause-effect relationships are as expected. The second regression aims to identify the importance of the modernization variables alone; hence the age and culture variables are excluded. The individual dependent variables are taken up, in order, below. Empirical Results Desired family size - The empirical findings for this variable clearly indicate that both social and economic modernization decrease the demand for children -- social modernization via enhanced education; economic 14 modernization thru occupational change (a shift out of agriculture), and, over and above the effect of occupational change, via redistribu- tion of the population from rural to urban areas (Table 2, column 1). In the regression, husband's literacy has a stronger effect than wife's years of schooling -- which falls somewhat short of statistical signifi- cance -- but for the reasons just indicated one should probably not make much of this. Almost as interesting are the modernization variables that have no significant impact on family size desires -- family income, women's work, and men's employment status -- although three of the four signs on women's work are negative, in keeping with theoretical expecta- tions. On the side of cultural influences the most striking finding is the difference between Lower and Upper Egypt. Even after controlling for differences in economic and social modernization, residents of Lower Egypt want about two less children than those of Upper Egypt. The absence of significant associations with age is noteworthy too. Once one controls for a few salient features of modernization, there is no indication that younger men or women want fewer children than older. Determinants of supply - Previously, empirical studies for Sri Lanka and Colombia found that four of the supply variables studied here are large- ly insensitive to modernization -- first birth interval, second birth interval, NSS, and pregnancy wastage (Easterlin and Crimmins 1985, -2 chapter 4). The same appears to be true of Egypt. Examining the R in the last line of Table 2, one finds that the lowest values -- .04 and .05 -- are for the two birth interval variables. For NSS and pregnancy .2 wastage the R values are higher, around .10, but the subsequent dis- . 15 cussion points out that the significant associations with modernization variables may, in part, be a statistical artifact, rather than reflect- -2 ing substantive factors. If this is so, then the R for these two variables would probably be much like those for the birth interval variables. The implication is that modernization alters the supply of children chiefly via its impact on age at marriage, infant and child mortality, and lactaticn. For this reason, these variables are taken up first. Age at marriage of wife - The women being studied here were mostly married in the period around 1960, a time when marriage decisions in Egypt were mostly the product of contracts negotiated between the parents of the prospective bride and groom. The formation of these contracts was almost certainly influenced by fairly reliable information on the educational status of the prospective parties, as well as their work history and prospects. Hence, although some of the variables used here to explain age at marriage relate to attributes that post-date marriage -- and sometimes by a sizeable period (husband's occupation, family income) -- it is plausible to suppose that they are reasonable proxies for those that more immediately shaped the marriage decision. The results point to four aspects of modernization that influence age at marriage, all in a plausible way. First, the greater the woman's years of schooling, the later her age at marriage. Second, women whose husbands hold professional or clerical jobs are likely to marry later. Unlike the results for desired family size, however, there is no indica- tion that the shift out of agriculture or redistribution to urban resi- dence has a significant impact. Third, women who work in modern jobs 16 before marriage are likely to marry later. Finally, higher income results in earlier age at marriage. The last finding is, perhaps, initially startling, because the direction of effect of modernization on marriage is counter to that found for the other-modernization variables, a higher degree of modernization (higher income) being associated with earlier marriage. But, on reflection, it makes sense to suppose that, in a society in which pre-marital and extra-marital sexual relations are severely proscribed, those who can afford to, i.e., those with higher income, are likely to marry earlier. p Despite the sizeable difference between residents of Lower and Upper Egypt in family size desires, there is no difference in age at marriage, when other factors are controlled. However, the other cul- tural variable in the analysis, blood relationship of spouses, does show a slight effect, just falling short of statistical significance -- the more traditional cultural pattern, involving some blood relationship between spouses, is associated with a slightly lower age at marriage, about 0.2 years. The associations with wife's age at marriage of husband's age and wife's age are at first, perplexing and seemingly contradictory -- the sign on husband's age is negative and that on wife's age, positive. The explanation probably lies in the nature of the present sample. First, there is a truncation problem that may account for the association between wife's current age and age at marriage. Older women are more likely to have a later mean age at marriage, the association observed, because they have had more time in which to get married. Second, although by its nature the sample is restricted to wives between the ages of 35 to 44, no such restriction applies to husbands, and the latter, in fact, range from 29 to 64 years 17 in age. If there was a trend toward later age at marriage of husbands, then one might find that older husbands are matched, on the average, with wives who married younger. Proportion of child mortality - The regression results indicate that modernization reduces infant and child mortality via channels that are both expected and unexpected. The favorable impact of improved female education on child survival is in keeping with the usual results (Preston 1978). So too is the favorable impact of family income. Surprisingly, however, the favorable effect of income appearsrnot to be due to better water or sanitation facilities that typically go with higher income, but simply to an increase in the number of rooms. This is brought out if one uses specific standard of living variables in multiple regression analysis. Number of rooms as an independent varia- ble performs slightly better than family income, while water and sani- tation facilities have little effect on the results. Unfortunately, standard of living data for diet or nutrition are not available, so this linkage could not be explored. Somewhat unexpected is the finding that husbands who work for .others are more likely to have families with better child survival. Possibly this may reflect health insurance arrangements available in urban areas, especially Cairo and Alexandria, to the families of such men. More puzzling is the finding that women who work in agriculture before marriage have higher infant and child mortality. One might dismiss this on the grounds that the after-marriage variable for this category has the opposite sign, though not a significant one. But, as will be seen, there are other respects in which women's work in agricul- 18 ture, sometimes after as well as before marriage -- is associated with rather similar adverse fertility effects -- on miscarriages and secondary sterility -- and this pattern gives some credence to the results. Turning to the other variables, one .finds the usual regional effect -- lower mortality associated with residence in Lower Egypt (WFS 1980). However, there is no rural-urban difference in mortality, when region is controlled. The other cultural variable, blood relationship of spouses, shows a direction of effect similar to that for Lower versus Upper Egypt, in that the less traditional situation is associated with more favorable mortality. Specifically, where there is no blood relationship between spouses, child survival is better, though the association falls a little short of significant. Finally, the data indicates that older women have a poorer child survival record, though this may merely indicate that their children have had a longer time to die. Duration of breastfeeding - Studies of breastfeeding behavior have typically found that shorter breastfeeding occurs among more educated women and those living in urban areas (Jain and Bongaarts 1980), and the results here are the same. In addition, shorter breastfeeding appears to characterize women who work in modern jobs before marriage. No significant effects are found for variables other than modernization variables, although a few fail the test by only a small margin. First birth interval - Because this variable aims to measure a couple's natural fecundity, the regression analysis is confined to women who did 19 not contracept before the first birth. In fact, however, only two women were excluded from the analysis by this restriction. The results for the modernization variables suggest that where the husband or the wife is working for others for cash wages, the first birth interval is longer, although for the husband the relationship is just short of satisfying the criterion of significance (t ! 2.0). Possibly there is something in the nature of such work situations that lowers frequency of intercourse and thus the couple's fecundity. An alternative interpretation is that the couple used contraception prior to the first birth but did not report it. The results for the variables designated here as cultural -- resi- dence in Upper vs. Lower Egypt and blood relationship between spouses -- are interesting. In both cases the less traditional attribute -- resi- dence in Lower Egypt or no blood relationship between spouses -- is associated with a shorter birth interval. This result is consistent with the practice that among more traditional couples consummation of marriage is likely to follow less closely on the date of marriage. Second birth interval - The regression analysis for this variable is limited to the population not contracepting prior to the second birth, with a view, as in the case of first birth interval, to identifying natural fertility determinants. This time the number of women excluded is somewhat larger (26), though this amounts to only about 6 percent of the sample. For the modernization variables the only significant finding is that women who work in agriculture after marriage are likely to have a longer birth interval. Since there is no significant association be- 20 tween such work and length of breastfeeding (see column 4), it does not appear that lactation behavior accounts for this result, although for most women the breastfeeding duration reported is for a birth interval other than the second. Conceivably the result may be connected to the finding noted below that pregnancy wastage is significantly higher for women who work in agriculture after marriage. A higher incidence of pregnancy wastage would, in general, contribute to longer birth inter- vals, and, in particular, if these women were more likely to be working earlier rather than late in marriage, to a longer second birth interval. The only other variables significantly associated with second birth interval are age of wife and husband. Older women are likely to have a shorter second birth interval, perhaps reflecting a tendency toward shorter breastfeeding for these women (see column 4), though the relationship between breastfeeding and age of wife falls -slightly short of statistical significance. The wives of older men are likely to have a longer second birth interval, perhaps due to lower frequency of inter- course. Secondary sterility (NSS) - In the case of this variable, for which 0 = secondarily sterile and 1 = fecund, a positive coefficient means higher fecundity. Among the modernization variables, income is significantly positively related to fecundity, suggesting that something in the life style of higher income women -- perhaps better nutrition and housing, perhaps better medical care -- may maintain fecundity to a later age. Interestingly, wife's work in agriculture before marriage is negatively related to secondary sterility. Without a better idea of the circum- stances of this work it is difficult to speculate on the reason for this 21 association. Moreover, it is surprising to find that although this relationship is fairly strong, there is no evidence of such a relation- ship for women who engage in agricultural labor after marriage. The other two variables significantly associated with NSS are the wife's and husband's age. The relationship for wives is .no doubt biological -- older women being more likely to be menopausal. In the case of husband's age the relationship may reflect lower frequency of intercourse for older husbands. The NSS variable is constructed in part by classifying women as not fecund if they were not contracepting and had no birth in the previous five years. Thus, by this criterion, a couple with a relatively low frequency of intercourse might be classi- fied as secondarily sterile. -Pregnancy wastage - The only significant associations with pregnancy wastage are for two -- perhaps three -- modernization variables, and for all of these the size of t-coefficients is marginal. One suspects that the positive association with wife's education is a statistical arti- fact. Better educated women are more likely to recall miscarriages than less educated, partly by virtue of their better education, and partly because they have fewer conceptions, and these are more likely to be closer to the survey date. The same reasoning probably applies also to the slightly less than significant association found for couples in which the husband is a professional or clerical worker. The other significant association -- again a positive one -- is for wives who work after (and, possibly, also for those who work before) marriage in agriculture. Conceivably this might reflect the difficulty of carrying to full term for women engaged in relatively strenuous farm 22 labor, but, again, it might be a statistical artifact, reflecting the occurrence of miscarriages under conditions that are more likely to be recalled -- in this case by virtue of their place of occurrence, while away from home working in the fields. However, the fact that for NSS the before marriage variable for women's agricultural work was signifi- cant adds to the suspicion that there is an underlying substantive phenomenon. Explanation of Regional Demographic Differences The preceding section revealed the various ways in which different aspects of modernization influence the demand for children and the several determinants of supply of children in Egypt. The purpose of this section is to see to what extent the regression equations presented in that section plus those from the preceding paper linking demand and supply to fertility control and fertility (the "proximate determinants" and "use" equations) can account for differences in demographic behavior among the five major regions of Egypt -- Metropolitan (Cairo and Alexandria), Urban Lower, Urban Upper, Rural Lower, and Rural Upper. For each region the mean values of the independent variables are sub- stituted in the equations of Table 2 to obtain predicted values of the dependent variables, which are then compared with the actual values, when available. The predicted values of the dependent variables of Table 2 are, in turn, substituted in the proximate determinants and use equations estimated in the preceding paper to obtain predicted values of use of fertility control and children ever born for comparison with the actual values. The guiding idea is that analysis of the pattern of 23 demographic differences observed in the regional cross-section provides an approximation to the temporal pattern of change in Egypt. Thus, as . one goes from the least modernized region, Rural Upper, to the most modernized, Metropolitan, one observes in approximate fashion, the manner in which different aspects of modernization have shaped the transition to deliberate control and lower fertility. Regional variations in modernization, culture, and demographic behavior -- In Metropolitan Egypt about 88 percent of the sample wives report having used deliberate fertility control at some timUe. On the average, use of fertility control was initiated over 11 years prior to the survey date and the average number of children ever born was slightly under 6. At the other extreme, in Rural Upper Egypt, only 28 percent report ever using contraception, with an average time since first use of only 2 years, and a cumulative fertility of 7.2 children. Between these extremes, in order of descending use of contraception are Urban Lower, Urban Upper, and Rural Lower. Typically, children ever born increases as use of contraception decreases, but in the two rural regions it is just about equal (Figure 2). It is not uncommon to find that in the early stages of transition to fertility control, increased use of contraception is not accompanied by declining fertility (Easterlin and Crimmins 1985). As a general matter, Metropolitan Egypt is the most modernized region and Rural Upper, the least modernized. When the regions are ranked in order of contraceptive use the progression from least to most modern is by no means linear, however, and varies with the aspect of modernization under consideration (Figure 3). Clearly the modernization 24 variables are capturing different dimensions of the process of economic and social development, as one would hope. Similarly, the two variables used here to proxy cultural differences -- residence in Upper versus Lower Egypt and blood relationship between spouses -- give rather different impressions of cultural variation among regions. In what follows, the basic question is how well regional differences in moderni- zation and culture, of the type represented in Figure 3, can explain the demographic differences shown in Figure 2. Regional variations in determinants of the supply of children -- The regression equations of Table 2 predict fairly well regional differences in the factors shaping the supply of -children (Figure 4). For most of the supply determinants, regional differences make for greater supply in the more modernized regions. For example, in Metropolitan Egypt child survival is higher, breastfeeding shorter, birth intervals shorter, and secondary sterility less prevalent -- all of which would enhance the supply of children. The only significant factor operating in the oppo- site direction is age at marriage, but it is an important one. (Region- al differences in pregnancy wastage probably should be discounted as reflecting reporting rather than substantive differences.) Women in the more modernized regions tend to marry later, and, given that the average age of women in the five regions is fairly similar, this means that the duration of exposure to conception is correspondingly shorter. On balance, the shorter length of exposure in more modernized regions just about offsets the factors making for enhanced natural fertility, and regional differences in natural fertility tend to be negligible (Figure 5). However, because child survival is higher in the 25 more modernized regions, the supply of children tends to increase .with the degree of.modernization of a region, though not greatly. In contrast to the pattern for supply, regional differences in the demand for children vary inversely with the degree of regional moderni- zation. * Comparison of predicted and actual values of desired family size indicates that the regression equation captures quite well the factors responsible for regional variations in demand (Figure 5). In turn, lower demand for children in more modernized regions, coupled with slightly higher supply, yields considerably higher motivation for con- traception in the more modern regions. For example, in Metropolitan and Urban Lower Egypt, couples would produce, on the average, three more surviving children than they want, if fertility were not regulated. In contrast, in Rural Upper Egypt, the demand for children is so high relative to supply, that the typical couple has about one child less than they want. One would expect that the regional variations. in motivation for fertility control would result in corresponding variations in use, and this appears to be largely true. Whereas upwards of 80 percent of women in Metropolitan and Urban Lower Egypt are ever-users, having initiated use over 9 years ago, only 28 percent are users in Rural Upper Egypt and the time since first use is only about 3 years (Figure 2). There is no contradiction, of course, between the negative value of motivation in Rural Upper Egypt and the fact that over one-fourth of the sample has ever used contraception. The motivation value is an average for all parents and does not reflect the fact that some couples lie sufficiently far above the average to be positively motivated to use contraception. In the model used in this and the preceding paper, contraceptive 26 use is a function of costs of regulation, as well as motivation for fertility control. The results of the earlier paper indicated that the variable focussed on here, motivation, was considerably more important in explaining contraceptive behavior. No attempt has been made in the present analysis to model the principal variable previously used to measure regulation costs, number of contraceptive methods known by the respondent and reported without prompting, although the regional mean for this variable was, of necessity, used in deriving predicted fertili- ty and fertility control values from the proximate determinants and use equations. The omission here of regulation costs as a subject for analysis is because the growth of contraceptive knowledge, though depending partly on the progress of modernization, is also influenced significantly by family planning programs, and national data for such programs that could be used in the model are not available. In addi- tion, there is the problem already mentioned that the measure of regulation costs, number of methods known, partly reflects use of contraception rather than vice-versa. So as not to omit consideration of regulation costs altogether, however, the regional means on number of methods known are shown in Figure 5. It is striking that although Metropolitan Egypt -has the highest knowledge (and, by implication, the lowest costs) and Rural Upper Egypt, the lowest knowledge, contraceptive knowledge is uniform among the other three regions. In contrast, use of fertility control differs considerably among these regions, a difference that is predicted by the motivation measure except for Upper Urban Egypt. In general, the regional differences in motivation versus contraceptive knowledge appear consistent with the general point that motivation is more important than regulation costs in explaining contra- ceptive behavior. 27 The differences among regions in contraceptive use and fertility predicted by the present model are fairly similar to the actual values (Figure 5). The implication is that regional differences in moderniza- tion and, to a lesser extent, culture are basically responsible for the observed variations in contraceptive use and fertility. In the future, as modernization progresses throughout the different regions, one may anticipate the further growth of motivation for fertility control, continued expansion of contraceptive use, and additional fertility reduction. 28 Figure 1 Approaches to Analyzing the Impact of Modernization on Fertility A. Multivariate Regression of Fertility on Basic Determinants a Basic Determinants Children Ever Born B. Proximate Determinants Analysis Basic Determinants Proximate Determinants Children Ever Born C. Present Approach Proximate Determinants I I Regulation RC a M Deliberate Basic Determinants Demand. l Fertility Control Children Cd Variables Ever Born Supply, I Cn Other Proximate I Determinants a Basic determinants include modernization variables (education, urbanization, etc.), cultural factors (ethnicity, religion, etc.), and other determinants such as genetic factors. Table 1 Direction of Effect of Various Aspects of Modernization on Indicated Determinants of Motivation for Fertility Control () (2) (3) (4) (5) Demand. Supply, Cd Cn Nat- Sur- ural vival In- fer- pros- Aspect of modernization Tastes come Prices tility pects I. Growth in formal education - - + + 2. Urbanization - - + Table 2 Multiple Regression of Specified Variable on Hodernization and Other Variables, Egypt, 1979-80 Variablea Determinants of supply, Cn and Age. Proportion Months First Second Not Proportion Statistic Demand at of child breast- birth birth second- of Cd marriage Mortality feeding interval interval arily pregnancy sterile wastage A. Metric coefficient (t-statistic in parentheses) Modernization variables 1 Years of schooling (W) -.0875 .2227* -.0072* -.4589* -.4057 .2361 -.0046 .0059* (1.73) (3.81) (2.11) (2.35) (1.17) (.64) (.65) (2.49) 2 Literate (H) -.7800* -.5569 .0247 1.7414 -1.9041 .5690 .0399 -.0006 (2.57) (1.59) (1.21) (1.49) (.92) (.28) (.94) (.04) 3 Urban residence -1.1960* .0169 .0088 -3.3630* 1.1171 -2.1586 .0050 .0235 (3.41) (.04) (.37) (2.48) (.47) (.90) (.10) (1.42) Occupation (H) 4 Profess. or clerical .2226 1.4983* -.0156 -1.472 -4.2734 -6.6043 .1478 .0557 (.36) (2.12) (.38) (.62) (1.03) (1.55) (1.73) (1.93) 5 Nonfarm. manual .6345 .3123 -.0019 .6363 -1.1679 -3.9887 .0433 -.0079 (1.27) (.54) (.06) (.33) (.34) (1.20) (.62) (.33) 6 Sales or service .8392 .1442 -.0426 -2.5664 -2.3340 3.9082 .0521 -.0123 (1.62) (.24) (1.22) (1.28) (.66) (1.14) (.72) (.50) 7 Farmer 1.3035* -.3240 -.0572 -1.9091 2.7418 .1665 -.0658 .0002 (2.34) (.50) (1.52) (.89) (.72) (.04) (.84) (.01) 8 Cash wage worker (H) .1878 .4573 -.0870* -1.8766 4.6397 2.5421 -.0364 -.0050 (.51) (1.08) (3.53) (1.33) (1.87) (1.02) (.71) (.29) Work before marriage (W) 9 Modern -.6356 3.2427* -.0214 -5.6387* -7.7717 .6277 .0674 -.0149 (.93) (4.12) (.47) (2.15) (1.67) (.12) (.70) (.46) 10 Other work -.5235 -.2961 .0790* .0794 2.3332 -1.2252 -.2888* .0407 (1.03) (.50) (2.31) (.04) (.68) (.36) (4.07) (1.70) Work after marriage (W) 11 Modern .0494 -.5214 .0007 4.117 9.4204* -.1816 .0234 .0097 (.07) (.67) (.01) (1.59) (2.06) (.04) (.25) (.30) Table 2.(continued) Variable Determinants of supply, Cn and Age Proportion Months First Second Not Proportion statistic Demand at of child breast- birth birth second- of Cd marriage Mortality feeding interval interval arily pregnancy sterile wastage 12 Other work -.2105 -.3797 -.0522 .5025 2.2507 8.2287* .0087 .0496* (.41) (.64) (1.52) (.25) (.65) (2.44) (.12) (2.07) 13 Family income .0457 -.2871* -.0154* -.5516 -.8673 -.0817 .0406* .0024 (.42) (2.26) (2.08) (1.30) (1.16) (.11) (2.64) (.46) Other Variables 14 Resident Lower Egypt -1.9509* .1002 -.0504* -1.8167 -4.2597* -.0109 -.0026 .0131 (6.35) (.28) (2.43) (1.53) (2.04) (.005) (.06) (.91) 15 Not blood relation (HW) .0207 ..2229 -.0123 .7534 -1.3965* -.3890 .0033 -.0061 (.20) (1.91) (1.81) (1.93) (2.03) (.59) (.23) (1.29) 16 Age (W) -.0712 .2434* .0092* -.3749 .5801 -1.0601* -.0284* .0001 (1.31) (3.88) (2.51) (1.79) (1.57) (2.85) (3.75) (.05) 17 Age (H) .0268 -.2134* .0016 .1366 -.0204 .3338* -.0164* .0010 (1.08) (7.48) (.98) (1.43) (.12) (1.99) (4.76) (.90) 18 Constant 7.7147* 17.150* -.0376 32.298* 9.5494 53.2922* 2.4968* -.0015 (3.99) (7.66) (.29) (4.32) (.72) (4.02) (9.22) (.02) B. Summary statistics 19 Number of cases 443 443 443 441 417 443 443 443 .2 20 R .22 .32 .16 .14 .08 .06 .22 .12 .-2 21 R , modernization .15 .22 .11 .12 .05 .04 .10 .11 variables only a. I = husband, W = wife, no designation both. * Significant at .05 level or better. -31 . . . . .. .. .. . . ... . ... : .. . .. . .. . . - - - - - - -- - - ___....._. Fertillty Control and Fertility, Mean Values, 1 : :ve or Regioäöns-, .79/9 g--80. . .. :..... ....- . . . L. . . Eer --1 8 60 - rs . - .~1. ..... .... .± -- -- - - -t __ _ __ _ -- - - _ _ _ _ -40___ _ ~20 ... ; -.-- -- e r . --.-.- -.- -- - . -7 - .--ears r-r .7n Square --o--e-Inc ,C @_____ ... ..... . --1 ---- ---- ---- -- _______._ .. ............pa ~OSquares to the Inch 32 n: A . . . . .. . . . ~. ...E.7-- ~ i ure.3 __________.... 9..... _..._.. dScernization and Cultural Variables, Mean Values, . 1.....- 7iVe-Majör Regioh', ~1979-80 H-hsbh-..o..desigaton=both) - ftash Wage Wrker 0- (H) .co T n ...... ...r ~ . - . . - - -40 30 r ....... - %.Wit - Job Before ____ Marriage 1... jj- ... -1 u 1 01- - - ...na... . - - --------7(w---7 -O-ui -ntile- - -µ Rlood--e- - - - . - . ationship ___________ - _Between -~ - -- - ---------- -7- - 20 Squares to the Inch &. 弁’牌不華可下平不中汗 磬＿ 造葬韋舉絮雙還遲藝華擊訪討方中R壯于付礬乎一十州不不 粹粹〔哥· 自韭韭呂至至韭呂至兀藝呂至蒞：至症至韭過至手至鑒州寫寫浩論領計叩一他弁一方 如榆邐“ar。‘to lh&I&ch 34 . ............ kgura. 5 ........ .. -Fått and, tivätiOä9 9 and.Tärcllity,_Åctuial (--ý and Predicted Means, 71ve Major-Regions, 1979-80 Natural ty ------ -lertill number 77' L.of:-biýths . . ......... 7 .......... _'c. w Years [Supplvlxtiy' er-of First .-numb [.childtSný Use _8 /4 Z* 7 . 0 L-Dgmand --.(Cd) -t.- Milldren -Born . ten !ä ----- -- -------- ............ ...... -7_ _ 7-7-ý-7' 6 f- vätion ------ . .... .... ... .. ... .. .... .. . ... . Hf ...... Sumbet behil-dr ~7 e s ew- own ä_ _=1 i ------- ... .... 4~ RLýý U L M- 21P Squares tf) the Inch eo & 35 References Bongaarts, John, 1978. "A Framework for Analyzing the Proximate Deter- minants of Fertility," Population and Development Review 4:105-32. Bongaarts, John and Jane Menken, 1983.-"The Supply of Children: A Criti- cal Essay," in R. Bulatao and R.D. Lee, eds., Determinants of Fertility in Developing Countries: A Summary of Knowledge, New York: Academic Press. Cochrane, Susan H., 1979. Fertility and Education: What Do We Really Know? World Bank Staff Occasional Papers, no. 26. Baltimore: Johns Hopkins University Press. Cochrane, Susan H., 1983. "Effects of Education and Urbanization on Fertility," in R. Bulatao and R.D. Lee, eds., Determinants of Fertility in Developing Countries: A Summary of Knowledge. New York: Academic Press. Coleman, J.S., 1968. "Modernization, II - Political Aspects," in David L. Sills, ed., International Encyclopedia of the Social Sciences, 10:395-402, New York: Macmillan. Davis, Kingsley and Judith Blake, 1956. "Social Structure and Fertili- ty," Economic Development and Cultural Change 4(3)(April): 211-35. Deeb, Bothina El and John Casterline, 1983. "Contraceptive Use in Egypt," Central Agency for Public Mobilisation and Statistics & International Statistical Institute (WFS), International Conference on Fertility in Egypt, Cairo, Egypt, December 20-22. Easterlin, Richard A., 1968. "Economic Growth: An Overview," in David L. Sills, ed., International Encyclopedia of the Social Sciences, 4:395-408, New York: Macmillan. 36 Easterlin, Richard A., 1978. "The Economics and Sociology of Fertility: A Synthesis," in C. Tilly, ed., Historical Studies of Changing Fertility, Princeton: Princeton University Press. Easterlin, Richard A. and Eileen M. Crimmins, 1985. The Fertility Revolution, Chicago: University of Chicago Press. Easterlin, Richard A., Eileen M. Crimmins and Ibrahim Khodair, 1985. "Determinants of Fertility Control in Egypt, 1979-80," Washington, D.C.: World Bank. Hallouda, Awad Mokhtar, 1983. "Patterns of Reproductive Behaviour in Egypt," Central Agency for Public Mobilisation and Statistics & International Statistical Institute, Cairo, Egypt, June. Inkeles, Alex and David H. Smith, 1974. Becoming Modern, Cambridge, Mass.: Harvard University Press. Jain, Anrudh K. and John Bongaarts, 1980. "Socio-Biological Factors in Exposure to Childbearing: Breastfeeding and its Fertility Effects," in World Fertility Survey, World Fertility Survey Conference 1980, Volume 2, Record of Proceedings, London, July 7-11, 255-302. Kelley, Allen C. and Robert M. Schmidt, 1983. "Family Planning, Socio- economic Change and Population Policy in Egypt: An Exploratody Methodology," paper presented at the Seminar on Egyptian Population Policy, Cairo, Egypt, October 16-17. Kuznets, Simon, 1966. Modern Economic Growth: Rate, Structure, and Spread, New Haven, Conn.: Yale University Press. Lee, Ronald D. and Rodolfo A. Bulatao, 1983. "The Supply of Children: A Critical Essay," in R. Bulatao and R.D. Lee, eds., Determinants of Fertility in Developing Countries: A Summary of Knowledge, New York: Academic Press. 37 Lerner, D., 1968. "Hodernization, I - Social Aspects," in David L. Sills, ed., International Encyclopedia of the Social Sciences, 10:386-95, New York: Macmillan. Lindert, Peter H., 1980. "Child Costs and Economic Development," in Richard A. Easterlin, ed., Population and Economic Change in Developing Countries, 5-69, Chicago: Chicago University Press. Lindert, Peter H., 1983. "The Changing Economic Costs and Benefits of Having Children," in R. Bulatao and R.D. Lee, eds., Determinants of Fertility in Developing Countries: A Summary of Knowledge, New York: Academic Press. Mueller, Eva and Kathleen Short, 1983. "Effects of Income and Wealth on the Demand for Children," in R. Bulatao and R.D. Lee, eds., Deter- minants of Fertility in Developing Countries: A Summary of Knowledge, New York: Academic Press. Nag, Moni, 1983. "Modernization Affects Fertility," Populi 10(1):56-77. Nawar, Laila and John Hobcraft, 1983. "An Analysis of Determinants of Fertility in Egypt," Central Agency for Public Mobilisation and Statistics & International Statistical Institute, International Conference on Fertility in Egypt, Cairo, Egypt, December 20-22. Potter, Joseph E., 1983. "Effects of Societal and Community Institutions on Fertility," in R. Bulatao and R.D. Lee, eds., Determinants of Fertility in Developing Countries: A Summary of Knowledge, New York: Academic Press. Preston, Samuel H., 1978. "Mortality, Morbidity, and Development," Population Bulletin of ECWA, 15 (December), 63-75. Rosovsky, Henry and K. Ohkawa, 1961. "The Indigenous Components in the Modern Japanese Economy," Economic Development and Cultural Change 9(April):476-501. 38 Taha, Ibrahim Khodair and Susan Hill Cochrane, 1984. "The Determinants of Desired Family Size: A Causal Analysis for Policy," preliminary draft, Washington, D.C.: World Bank, July. World Fertility Survey, 1980. The Egyptian Fertility Survey 1980, Volume III, Socio-Economic Differentials and Comparative Data from Husbands and Wives, Central Agency for Public Mobilisation and Statistics & International Statistical Institute, Cairo, Egypt. 39 Table A-1 Definition and Measurement of Variables Variable EFS Question- Definition and measurement variable naire Wife's age at V109 Wives marriage Wife's years of X704 Wives education Husband's literacy Q113 Husband Two-category variable: 1=can read; O=cannot read. For those who did report their literacy, Q113 was set equal zero. Place of residence V702 Wives Two-category variable: 1=urban, O=rural. Husband works in V804 Wives Two-category variable: i=if professional or V804 = 1 or 2; O=otherwise. clerical position. (HJOB1) Husband is a skilled V804 Wives Two-category variable: 1=if or unskilled worker V804 = 8 or 9; O=otherwise. (HJOB2) Husband works in V804 Wives Two-category variable: 1=if sales or service V804 = 3 or 7; O=otherwise. (HJOB3) Husband is self- V804 Wives Two-category variable: 1=if employed in agric. V804 = 5; O=otherwise. (HJOB4) Husband works for V805 Wives Two-category variable: 1=if other & paid in cash V805 = 4; O=otherwise. (EMPEE) Wife worked before V708 Wives Two-category variable: 1=if marriage in a non- V709 Wives V708=1 or 2 or 3 or 7 or agricultural occu- 8 or 9 and V709=4 or 7; pation (except O=otherwise. domestic service) and she was self- employed or paid cash. (WWORKA1) Wife worked before V708 Wives Two-category variable: 1=if marriage in V709 WWORKAl=O; 0=if WWORKAl=1. agriculture (WWORKB) 40 Table A-I (continued) Variable EFS Question- Definition and measurement variable naire Wife worked after V710 Wives Two-category variable: 1=if marriage in prof. V711 V710=1 or 2 or 3 or 7 or 8 or 9, or cler. position and V711=5 or 8; O=otherwise. or in sales or as skilled or unskilled labourers and she was self-employed or paid cash. (WWORKA2) Wife worked after V710 Wives Two-category variable: 1=if marriage in V711 WWORKAl=O and V710 i0; agriculture (WWORKB1) O=otherwise. Family income QINC Economic Five category variable: quintile 1=1owest quintile to 5=highest quintile. Upper-lower region V701 Wives Two-category variable: 1=if of residence (UL) V701=1 or 2 or 3 or 5 or 7; 0=otherwise. Blood relationship S102 Wives. Four-category variable: 3=if between spouses $102=8 or 9; 2=if S102=7; 1= .(SI02) if 1102=5 or 6; O=if S102=1 or 2 or- 3 or 4. Age of wife in years VO1O Wives Age of husband in Q105 Husbands years. Table A-2 Mean, Standard Deviation, and Range for Variables in Regression Analysis Variable Mean Standard Range Number Deviation Maximum Minimum of Cases Children ever born 6.66 2.43 14 2 443 Years since first use of fertility control 6.88 6.76 26.33 0 443 Ever use of control .65 .49 1 0 443 Desired family size 4.70 3.07 20 1 443 Age at Marriage (W) 17.07 3.74 32 11 443 Number of methods Known (W) 1.89 1.09 6 0 443 Proportion ot child mortality .208 .198 1 0 443 Length of Breastfeeding 18.16 11.28 75- 0 First birth interval 24.36 19.19 117 7 441 Second birth interval 26.03 18.49 202 0 417 Proportion of pregnancy wastage .09 .13 . .57 0 443 Not secondarily sterile .76 .43 1 0 443 Years of schooling (W) 2.22 3.69 20 0 443 Literate (H) .29 .45 1 0 443 Urban residence .48 .50 1 0 443 Occupation (H) Profess. or clerical .16 .37 1 0 443 Nonfarm manual .29 .45 1 0 443 Sales or service .20 .40 1 0 443 Farmer .22 .41 1 0 443 Cash wage worker (H) .59 .49 1 0 443 Work before marriage (W) Modern .06 .24 1 0 443 Other work .10 .30 1 0 443 Work after marriage (W) Modern .07 .26 1 0 443 Other work .10 .30 1 0 443 Family income 3.34 1.35 5 1 443 Resident Lower Egypt .70 .46 1 0 443 Not blood relation (HW) 1.74 1.35 3 0 443 Age (W) 38.69 2.77 35 44 443 Age (H) 46.81 6.08 29 64 443 42 Table A-3 Regional Values for Figure 2 All Metro- Urban Urban Rural Rural Egypt politan Lower Upper Lower Upper % Ever Using 61 88 78 63 54 28 Years Since First Use 6.88 11.28 9.26 6.58 5.46 2.08 Children Ever Born 6.66 5.94 6.19 6.32 7.21 7.21 Table A-4 Regional Values for Figure 3 All Metro- Urban Urban Rural Rural Egypt politan Lower Upper Lower Upper Years of Schooling (W) 2.22 3.88 2.31 3.29 1.32 0.92 % Urban 49 100 100 100 0 0 % In Nonagricultural Occupations (H) 65 99.1 86 93 36 37 % In Lower Egypt 70 100 100 0 100 0 % Cash Wage Worker (H) 59 83 57 61 43 54 % With Modern Job Before Marriage (W) 6 13 9 12 2 0 Family Income Quintile 3.3 3.9 3.5 3.2 3.3 2.6 Blood Relationship Between Spouses (3=none) 1.74 1.93 2.07 1.83 1.76 1.21 43 Table A-5 Regional Values for Figure 4 All Metro- Urban Urban Rural Rural Egypt politan Lower Upper Lower Upper Age at Marriage, years Predicted 17.07 17.98 17.51 17.73 16.43 16.30 Actual 17.07 18.01 17.71 17.34 16.33 16.46 % of Children Surviving Predicted . 79 84 80 77 80 73- Actual 79 82 82 80 81 72 Duration of Breastfeeding, months Predicted 18.16 14.49 16.08 17.47 19.92 21.79 Actual 18.16 14.52 14.43 19.73 20.60 20.79 First Birth Interval, months Predicted 24.36 21.59 21.71 26.03 23.76 29.47 Actual 24.36 21.92 23.50 22.90 22.82 30.87 Second Birth Interval, months Predicted 26.04 23.31 23.20 23.41 28.15 28.46 Actual 26.04 22.33 24.56 24.28 28.47 28.15 Proportion Not Secondarily Sterile Predicted 0.76 0.83 0.81 0.84 0.71 0.68 Actual 0.76 0.83 0.86 0.76 0.68 0.72 Proportion of Pregnancy Wastage Predicted 0.09 0.12 0.10 0.09 0.08 0.06 Actual 0.09 0.12 0.11 0.09 0.08 0.07 44 Table A-6 Regional Values for Figure 5 Metro- Urban Urban Rural Rural politan Lower Upper Lower Upper Natural Fertility (N), Number of births 7.74 7.98 7.66 7.78 7.70 Supply (Cn), Number of children 6.50 6.41 5.86 6.21 5.64 Demand (Cd), Number of children Predicted 3.32 3.40 5.36 4.82 6.80 Actual 3.46 3.69 4.54 4.57 7.16 Motivation (Cn-Cd), Number of children 3.18 3.02 0.51 1.39 -1.16 Years Since First Use Predicted 9.06 8.29 4.14 5.05 1.88 Actual 11.28 9.26 6.58 5.46 2.08 Children Ever Born Predicted 6.21 6.58 6.96 6.93 7.39 Actual 5.94 6.19 6.32 7.21 7.21 Number of Methods Known (without prompting) 2.47 1.98 1.95 2.00 0.92