ISSN (Print) - 0012-9976 | ISSN (Online) - 2349-8846

A+| A| A-

Factors Contributing to the Declining Trend in Sex-Differentials in Mortality in India

The disappearance of "excess" female mortality tends to be attributed to a reduction in discrimination against females in the intra-family allocation of resources in access to healthcare and nutrition. The exploratory analysis in this paper suggests that the disappearance of "excess" female mortality in India is attributable to the process of demographic development and not to declines in discrimination in the intra-family allocation of resources.

SPECIAL ARTICLEEconomic & Political Weekly EPW august 9, 200861Factors Contributing to the Declining Trend in Sex-Differentials in Mortality in IndiaD JayarajThe disappearance of “excess” female mortality tends to be attributed to a reduction in discrimination against females in the intra-familyallocation of resources in access to healthcare and nutrition. The exploratory analysis in this paper suggests that the disappearance of “excess” female mortality in India is attributable tothe process of demographic development and not todeclines in discrimination in the intra-family allocation of resources.The author is grateful to S Subramanian for the paper has benefited immensely from the discussions with him and to Robert Cassen for useful comments on an earlier version of the paper.D Jayaraj (jayaraj@mids.ac.in) is with the Madras Institute of Development Studies, Chennai.It is often believed that nature has equalising tendencies. However, throughout human life nature appears to prefer inequality to equality. The preference for inequality probably starts at a very early stage – at embryonic stage – of human life. At conception, males appear to be favoured. It is likely that around 1301 [Huxley 1926] males are conceived for every hun-dred females. But this ratio probably drops sharply to around 105 at birth. This implies that more males than females die in utero. The observed mortality pattern in the developed countries sug-gests that excess male mortality is not just restricted to the embryonic stage but continues almost2 throughout the lifespan of a birth cohort. This observed excess mortality of males at every age, attributed to “biological” factors, is assumed to be “natural”. Males are physically stronger compared to females, but the latter are biologically stronger [Aaby 1992]. Since excess male mortal-ity at every age is accepted to be natural, the mortality differen-tials observed against females in most developing countries need to be explained. The explanation provided is often social or gen-der discrimination [see, for example, Kumar 1989; Sen 1990, 1992 and Aaby 1992] in the intra-family allocation of resources, particularly in the spheres of access to healthcare and nutrition. 1 IntroductionIn India, until the mid-1980s, the sex-differential in expectation of life at birth was against females. Since the mid-1980s the expectation of life at birth differential has turned in favour of females. This change has been taken to imply, as Sen (2003, p 1297) has concluded, “...female disadvantage in mortality has typically been reduced substantially...”. Such reductions, elsewhere, in mortality disadvantage of females are attributed to reductions in discrimination against females, particularly against young girls, in the intra-family allocations of resources [Klasen and Wink 2002]. They state, in the context of north Africa and west Asia, “The major factors accounting for this relative improve-ment [female life expectancy surpassing male life expectancy] appear to be improved female education, improved female employment opportunities, and generally rising prosperity in these middle-income countries, which lessens families’ incli-nation to ration scarce health and nutritional resources to the dis-advantage of young girls” (emphasis added) [Klasen and Wink 2002: 302]. In India, too improvements in employment and edu-cation of females appear to have accompanied the improvement in sex-differentials in mortality in favour of females. The Ninth Five-Year Plan report (Volume 2) indicated that: expec-tation of life at birth of females has registered steady improvements from 31.7 years in 1951 to 59.7 years in 1989-93 and overtook that
SPECIAL ARTICLEaugust 9, 2008 EPW Economic & Political Weekly62of males in the period 1989-93; female literacy rate went from 21.97 to 39.17 per cent and male literacy went up from 45.95 to 64.13 per cent between 1971 and 1991; and, female work participa-tion rate considerably improved from 14.2 per cent in 1971 to 22.3 per cent in 1991 and male work participation registered a marginal decline from 52.75 to 51.61 per cent in the same period. These fig-ures suggest considerable narrowing of the gender gap in literacy and employment opportunities. The document also stated, “In the field of health and demography, significant gains in respect of women’s health status have been recorded”. Based on these observed changes, the plan document claims that there has been “...a perceptible improvement in the socio-economic status of women in the country”. Accordingly, it could be inferred that the disappearance of “excess” female mortality in India is attributed to the improvement in the socio-economic status, an important com-ponent of which is declines in discrimination against females in the intra-family allocation of resources. In this context, the following question arises: Does the substan-tial reduction in “excess” female mortality in India connotes sub-stantial reduction in discrimination against females in intra-family allocation of resources in the spheres of access to healthcare and nutrition? Finding answers to this question serves as the basic motivation for the present paper. As a prelude to answering this question, the complexities attending sex-differentials in mortality are explored in Section 2. Attempt is made to identify the factors contributing to reduction in “excess” female mortality in India in Section 3. Concluding observations are made in Section 4.2 ExploringtheComplexitiesAs indicated earlier, in the introduction, the observed mortality dif-ferential against males in most developed countries is accepted to be natural. Mortality differential against females observed in most developing countries is attributed to social practices, particularly to discrimination in intra-family allocation of resources. In this connection, it appears that attributing “excess” female mortality to social factors – discrimination in intra-family allocation resources – is too simplistic. The complex interactions between different fac-tors, classified into five groups: biological, economic, social and cultural, environmental and behavioural [see Aaby 1992 and United Nations 1998] need to be taken into account in explaining the trend in “excess” female mortality. It needs to be stated here that the attempt in this section is to provide a glimpse of the complexities attending “excess” female mortality. Such attempt does not extend to providing a comprehensive review of the literature on the subject. It also needs to be stated that the discussion on biological superiority is conducted to highlight a, somewhat, neglected dimension – the variability that could arise in genetic endowment across birth cohorts which is likely to be an important source of mortality differen-tials – in the analyses of sex-differentials in mortality over time.Crucial to the analysis of mortality differentials is the assump-tion that females are biologically superior. Biological superiority of females arises due to differences in genetic endowments of males and females. Hence, it becomes essential to understand the factor that contributes to gender differences in genetic endowments. It appears that the differences in the combination of the chromo-somes that produce a male or a female offspring determine the biological superiority of females [see, for example, Huxley 1926]. While females have a pair of ‘X’ chromosomes, males have only one ‘X’ and the partner in the pair being a ‘Y’ chromosome. Huxley adds that the ‘Y’ chromosome is usually rudimentary in size and bears no hereditary factors. Hereditary factors are transmitted through ‘X’ chromosomes. Many hereditary factors are recessive when masked by the presence of the corresponding dominant fac-tor. To quote Huxley (1926; reprinted in 1937; p 56), “It is also a well known fact that many hereditary factors are recessive – that is, do not produce their effects if masked by the presence of their corresponding dominant factor”. Thus, the genetic factors/defects, while in females are masked by the presence of the corresponding dominant factor, in males they are not masked. For this reason, while males fall victim, females just turn out to be only carriers of the genetic factors. Indeed Huxley (p 58) says, “... their [genetic] effect will be almost exclusively masculine”. Accordingly, it is con-ceded that females are biologically superior. Genetic FactorsGenetic factors are classified as: (1) styled lethal, (2) semi-lethal, and (3) recessive factors. Genetic factors that make the survival of the embryos impossible are called styled lethal. Factors that allow the embryos to survive under the most favourable circumstances are termed as semi-lethal [Huxley 1926]. Recessive factors are those that reduce the general resistance but do not affect the via-bility. In females, since defective factors are masked by the pres-ence of their corresponding dominant factor, genetic factors could only be recessive. But in males the genetic factor could be either lethal or semi-lethal but cannot be recessive. It is important to note here that semi-lethal allows the embryos to survive only under the most favourable conditions, and this appears to be the natural source of mortality differential against males throughout the lifespan. This implies that survival disadvantage that arises from genetic factor depends on survival conditions faced by males at different age/stages of their life. Accordingly, the impact of genetic factors on mortality differentials experienced by a birth cohort are likely to depend on the survival conditions faced by the cohort from conception, and not only from birth, to the end of life. It is also possible that the observed mortality differential at a particu-lar age depends on survival condition faced by a cohort in the pre-vious age/stages of its life. To be more precise, survival conditions faced at embryonic/fetal stage may have a bearing on the observed mortality differentials of a birth cohort throughout its lifespan. In what follows, how survival conditions in utero impacts the observed mortality differentials across birth cohorts and through out the lifespan of a birth cohort are sought to be explained.Viability or circumstances of survival of fetuses in utero are con-ditioned, largely, by nutritional status of women during pregnancy. The importance of nutritional status of mothers to the survival of male fetuses could be deduced from Huxley (1926, p 58). He states: That [the absence of a ‘X’ chromosome] will mean that inherent incapac-ity to exist, or inherent lack of resistance to any unfavourable agencies, will be commoner in males than in females, and this should hold for other mammals besides man. This is actually so, as Dr Parkes of Univer-sity College has verified. He has taken the matter a step further, since he has found that poor nutrition (in mice) not only causes more deaths of embryos in utero (the number dying being ascertained by the difference
SPECIAL ARTICLEEconomic & Political Weekly EPW August 9, 200863between the number of young born and the number of corpora lutea, or traces of discharged ova, in the ovaries) – not only this, but it causes a shift towards femaleness in the sex-ratio of the young born, a dispropor-tionately large number of males dying as embryos. This discussion, by extension to other mammals including human beings, indicates that the poor nutritional status of women during pregnancy makes the circumstances of survival in utero unfavourable. It is likely that under such circumstances all male embryos with genetic defects die in utero. This implies that only the fittest of the males (i e, only those with no genetic factor) sur-vive from conception to birth when the nutritional status of preg-nant women is poor. For this reason, it is possible that among birth cohorts to women of poor nutritional status during pregnancy, the biological superiority of females gets neutralised at birth. This neu-tralisation is expected to have a bearing on the observed mortality differential of a birth cohort throughout its lifespan. Two Hypothetical ExamplesTwo hypothetical examples/scenarios are employed to bring out the impact of the neutralisation of the biological superiority of females at birth on the observed mortality differentials throughout the lifespan of a birth cohort. The first hypothetical example is con-structed on the assumption that the nutritional status of the gen-eral population and by extension that of women during pregnancy is extremely poor. Accordingly, it is also assumed that the rate of foetal loss is as high as 303 per cent. Given these specific assump-tions underlying the first example, the following are assumed to hold true for both the examples: (1) 130 males are conceived per every 100 females; (2) irrespective of the sex of the offspring 20 per cent of all offspring have a genetic factor; and (3) a total of 230 conceptions occur. A simple calculation, which incorporates the above assumptions, indicates that: (1) 26 male and 20 female foe-tuses inherit genetic factors; and (2) 69 fetuses, out of a total 230, are lost. All male embryos with genetic factors are expected to die in utero, as survival condition, determined by nutritional status of women, is not favourable. Accordingly, all 26 male embryos with genetic factors are lost before birth. Death of the other 43 embryos is not due to the presence of genetic factors. For this reason, it is assumed that the loss of the other 43 fetuses is distributed, respectively, between males and females according to their respective share in the total embryos that survive genetic factors. Specifically, of the 43 embryos lost, 22 will be males and 21 will be females. Thus, in this situation 82 males and 79 females will be born alive. Notice that (1)the sex ratio – which is defined as the number of females per 1,000 males – at birth is equal to 9634, and (2) more importantly only the fittest of the male embryos survive at birth. The later implies that genetic factors have a very little role in explaining the sex- differentials in mortality throughout the lifespan of this birth cohort. However, in this situation, there exists the possibility that more females than males are born with nutritional deficiency. Nutritional deficiency, to the extent that it is an important determinant of general resistance to diseases, may expose females to greater risks of ill-health (morbidity) and mortality due to infectious diseases. The second hypothetical example is constructed on the assumption that the nutritional status of the general population, and by extension that of women particularly during pregnancy, is vastly superior to those in the first example. To be more explicit, it is assumed that the extent of discrimination experienced by women in intra-family allocation of nutrition is the same in both examples. Further, it is also assumed that: (1) the reproductive loss is at 25 per cent, and (2) the genetic factors are reduced to semi-lethal in 20 per cent of the male embryos. The latter assump-tion implies that 20 per cent of the male embryos survive with genetic factors in utero.A simple calculation, based on these assumptions, indicates the following: (1) A total of 58 fetuses (25 per cent of 230 embryos) are lost. (2) Of the 58 embryos lost, 21 will be male embryos with genetic factors. (3) The loss of the other 37 embryos is due to fac-tors other than that attributable to the presence of genetic or sex specific factors. The last observation implies that the loss of the 37 embryos will be distributed between male and female fetuses according to their share in the total number of embryos that sur-vive genetic factors. Based on these assumptions, it could be shown that, of the 37 embryos lost, 19 would be males and 18 would be females. It may also be noted that, of the five male embryos that are assumed to survive the effect of genetic factors, one would be lost due to other factors. This implies that at birth four male embryos survive with genetic factors. Thus, in the present case, 90 males and 82 females would be born alive. Notice that (1) the sex ratio at birth that obtains in the second hypothetical situation at 911 is lower than that obtained in the earlier example, and (2) more importantly, four males, i e, 4.44 per cent of males survive with genetic factors at birth. This implies that the genetic endowment of the birth cohort born in the second hypothetical situation is differ-ent from that of the birth cohort of the first example.While in the first hypothetical example only the fittest of the males have survived at birth, in the second example 4.44 per cent of the males with genetic factor survive at birth. For this reason, it is expected that any unfavourable agency that may affect the via-bility of the live born at every age will exert greater impact on males than on females of the birth cohort of this example. Accord-ingly, the mortality differentials observed at every age of this birth cohort is expected to be against males. Thus, what happens in utero appears to have implications to the observed mortality dif-ferentials at every age of a birth cohort. In the present case, mor-tality risk at every age for males due to any unfavourable agency is likely to exceed that of females by a factor of 0.04444 per cent. Notice here that the superior nutritional status5 of women during pregnancy in the second example is attributed to reflect a sex-neutral improvement. Such improvement is shown to lead to (1) a fall in the sex ratio at birth, and (2) an increase in mortality dif-ferentials against males. This suggests that an increasing trend in the well-being of a population which is sex-neutral may induce an increasing trend in the observed “excess” male mortality over time. Thus, these examples suggest that improvements in sex-differentials in mortality in favour of females need not necessarily imply improvements in the intra-family allocation of resources in healthcare and nutrition in favour of them. Behavioural CausesApart from biological factors, there are other factors which affect sex-differentials in mortality. Behavioural factors, such as
SPECIAL ARTICLEaugust 9, 2008 EPW Economic & Political Weekly64smoking, alcohol consumption, attitude towards risk, violence, nature of work affect sex-differentials in mortality. More males than females die of such behavioural causes [see, Taket 1986 and United Nations 1998]. Employing data from 18 developing coun-tries Taket (1986) had identified that accident mortality is usually higher among young boys than girls. Mortality transition related factors too affect the observed mor-tality differentials. Aaby’s (1992) work suggests that mortality and health transition may either widen or narrow the observed levels of mortality differentials between males and females, particularly among children. For example, mortality differential due to measles appears to depend on nature and intensity of transmission of the disease. Aaby (1992) has observed in Guinea-Bissau that the case fatality rate (defined as the ratio of number of deaths to the number of cases who are ill) was around 22, 13, 42 and 31 per cent, respec-tively, when the transmission was from male-to-male, female-to-female, male-to-female and female-to-male. Notice here that the case fatality rate is the highest when the transmission is from male-to-female. In this connection, it is important to note that in a popu-lation where females are assigned household duties as their primary responsibility, they are likely to be confined largely to their home. Under such circumstances, it is likely that males who go out get infected from outside their home and pass on the infection to females. It may also be noted here that the extent and intensity of transmission depends on living conditions. Housing, particularly availability of number of rooms which determine the possibility of confinement of affected persons, and sanitary facilities appear to be important determinants of the extent and intensity of transmission. Natural CalamitiesFamines, wars and natural calamities are other factors which have a bearing on the observed mortality differential. Famines lead to higher mortality of males than females [Watkins and Menken 1985, Jannetta 1992, Pitkanen 1993, Pitkanen and Mieklke 1993 and Seminar Report 1999]. However, it is likely that males who survive the impact of a famine are likely to be better fit to survive the other causes of mortality than females who survive the famine. This implies that among the cohort that has survived a famine, mortality differentials against males are likely to be less pronounced. Wars, apart from contributing directly to excess male mortality, may also have an indirect effect. It is possible that those who serve as sol-diers in hostile environments acquire certain infections which might lower their resistance capacity to certain diseases. Thus, the presence/absence of war cohorts may affect the observed sex- differentials in mortality of a population over time. To sum up, the discussion in this section indicates the complex interplay of biological, behavioural, social and cultural, eco-nomic, and environmental factors in determining the observed sex-differentials in mortality. Hence, it is far too simplistic to attribute the observed “excess” female mortality in the developed countries, largely, to social factors. More importantly, the discus-sion indicates that the incidence of presence of genetic factors among males is likely to vary across birth cohorts. In this context, it needs to be noted that mortality comparisons tend to treat that all birth cohorts are homogeneously endowed with genetic fac-tors. Such treatment may result in either over or understating the impact of social discrimination in reducing “excess” female mortality over time [see, also, United Nations 1998]. While these complexities are recognised here, it is impossible to account for the impacts of all these factors on observed sex-differentials in mortality. For an exhaustive accounting of the impacts of all factors, reliable annual time series data on the sex ratio at birth, pregnancy wastage (spontaneous abortions and stillbirths), and deaths classified by age, sex and causes are required. It is well known that such data are hardly available for the developing countries, particularly for India. Hence, making use of the available data, in the next section, an exploratory attempt is made to understand the factors that contribute to the observed improvements in mortality differential in favour of females in India between 1971 and 2001.3 AnExploratoryAnalysisThis section offers an exploratory analysis of the factors con-tributing to the trend in sex-differentials in mortality.3.1 IdentifyingContributionsIt may be recalled that one of the questions sought to be analysed in this paper is does the substantial reduction in “excess” female mortality in India imply a substantial reduction in discrimination against females in intra-family allocation of resources? To answer this question one needs to know: (1) How much of the reduction in “excess” female mortality is due to change in age-mortality pat-tern?; (2) How much of the reduction is attributable to an increase in mortality due to behavioural factors?; (3) What is the impact of mortality transition on mortality differentials?; and (4) How much of the reduction in mortality differential could be attributed to the change in genetic endowment of birth cohorts that constitute the population at different points of time? The required data to find answers to these questions are, as noted earlier, simply not availa-ble for India. For this reason, an attempt, which is exploratory in nature, is made to identify the impact of behavioural factors and the impact of change in age-pattern of mortality. The causes of death could be broadly classified into two groups: behavioural and non-behavioural. As a first step towards identi-fying the impact of change in intra-family allocation of resources to change in “excess” female mortality, the contributions of behavioural and non-behavioural causes of deaths to sex- differentials in mortality need to be obtained. To this end, first the index of sex-differentials in mortality (ISDM) is defined as: ISDMt = (CDtf–CDtm) ...(1)where superscripts f and m stand for females and males subscript t represents time, and CD stands for the crude death rate. Notice thatISDM is just the absolute difference in crude death rates of males and females. While positive values of ISDM indicate “excess” female mortality, negative values indicate “excess” male mortality. Obviously, the value of ISDM equal to 0 indicates that there is neither “excess” female nor “excess” male mortality. It may also be noted here that CDf and CDm, are the numbers of female and male deaths per thousand, respectively, female and male populations. ISDM could be further decomposed into two components: (1) behavioural, and (2) non-behavioural. To decom-pose the differential, the crude death rates of behavioural and
SPECIAL ARTICLEEconomic & Political Weekly EPW August 9, 200865non-behavioural causes need to be estimated. The two types of crude death rates for females are defined as:CDtfb=(Dtfb/(Ptf)*1000 ...(2)andCDtfnb=(Dtfnb/(Ptf)*1000 ...(2’)where as indicated earlier,CD stands for crude death rate; super-script f, b and n represent, respectively, females, behavioural, and non (i e, nb denotes non-behavioural); subscript t denotes time;D signifies total deaths; and P stands for population. Simi-larly, the behavioural and non-behavioural crude death rates for males are defined as follows:CDtmb= (Dtmb/(Ptm)*1000 ...(3)andCDtmnb=(Dtmnb/(Ptm)*1000 ...(3’)Notice here that in equations 3 and 3’ superscript m represents males, and other notations are as defined earlier. Notice also that CDtf =CDtfb+ CDtfnb, and CDtm = CDtmb+ CDtmnb. Accordingly, it is easy to decompose the index of sex-differentials in mortality into that attributable to behavioural and non-behavioural causes.Given equations 2, 2’, 3 and 3’, the sex-differentials in mortality attributable to behavioural and non-behavioural causes are iden-tified as follows:ISDMbt=(CDtfb–CDtmb) ...(4)and ISDMtnb=(CDtfnb–CDtmnb) ...(5)where ISDMbt and ISDMnbt, respectively, are the absolute differ-ences in crude death rates or mortality differential between males and females attributable to behavioural and non-behavioural causes. The total change in sex-differentials in mortality between two points of time t and t’, denoted as ΔISDM, could be obtained as ISDMt, –ISDMt. Similarly, the changes in mortality differentials between two points of time t and t’ attributable to behavioural and non-behavioural causes, denoted respectively, as ΔISDMb and ΔISDMnb could be obtained as: ISDMbt’ – ISDMbt,and ISDMnbt’ – ISDMtnb. Given these three quantities, ΔISDM, ΔISDMb and ΔISDMnb, the contributions of behavioural and non-behavioural causes of mortality to the observed change in sex-differential in mortality could be obtained as: ΔISDMb/ΔISDM, and ΔISDMnb/ΔISDM. Notice here that an overwhelmingly large contribution by behavioural causes indicates the absence of theimpacts of biological and dis-crimination in intra-family allocation ofresources related factors on observed sex- differentials in mortality.In what follows the procedure employed to identify the contri-bution made by each age-group to the decline in mortality differen-tials in favour of females is detailed. For each age-group the female and male crude death rates at a point of time, t, are defined as:CDfit = Dfit/ Ptf ...(6)andCDmit = Dmit/Ptf ...(6’)where i represents age-groups, and other notations employed are as defined earlier. Given CDfit, and CDmit, notice that N NCDtf =ΣCDfit andCDtm = ΣCDmit. Where i stands for age-group i=1 i=1and N represents the total number of age-groups: equal to 15 inthe present analysis. Now for each age-group i the index of sex-differentials in mortality is obtained as: ISDMit = (CDfit – CDmit), and the overall index of sex-differentials in mortality NISDMt = Σ (CDfit – CDmit). Now it is easy to see that the contribution i=1 to “excess” female mortality by an age-group i is obtained as: ISDMit/ISDMt. The change in the ISDM in age-group i between two points of time t and t’ is obtained as: ΔISDMi = ISDMit’ – ISDMit. It is easy to see that the contribution to change in “excess” female mortality by an age-group i isΔISDMi/ΔISDM. Notice here that the contributions by different age-groups may indicate the sources of change in ISDM. For example, the contributions by the age-groups in the lowest end of the age-spectrum where there has been no significant improvement in the relative survival chances of females could be attributed to change in age-structure and to the relatively steep declines in mortality rates of males and females (sex-neutral improvement in survival) in these age-groups.Given the procedure adopted for identifying the contributions by behavioural and non-behavioural factors, and age-groups to ISDM andΔISDM, the results of the identification exercises are presented in the next sub-section. 3.2 Results and DiscussionData on all behavioural causes of deaths in India are difficult to obtain. However, data on accidental deaths and suicides, which probably accounts for the major share of behavioural causes of deaths in India, are available for the period 1969 to 2004. Such data provided by the National Crime Records Bureau, ministry of home affairs, government of India, are published in the annual publication titledAccidental Deaths and Suicides in India. It is pos-sible that there is considerable under-reporting of both suicides and accidental deaths to avoid possible police investigation into the causes of such deaths and criminal proceedings that could be initiated. However, this appears to be the only source of statistics that provides fairly reliable data on accidental deaths and sui-cides. Accordingly, this data set is employed to identify the con-tribution of accidental deaths and suicides (behavioural causes) to sex-differentials in mortality. It may be noted here that the attempt to identify the contributions of behavioural and non- behavioural causes of deaths toISDM and change inISDM is restricted to the years 1971 and 2001. Thus, in this exercise t and t’ stand, respectively, for the years 1971 and 2001. These years are chosen as 1971 and 2001, respectively, are the first and the latest census years in the period 1970 to 2004. The behavioural causes of mortality include accidental deaths and suicides. Accidental deaths comprise deaths due to road, rail, air, factory and other accidents; accidental falls and drowning; and, natural disasters and sudden deaths. It may also be noted that sudden deaths also include deaths due to sudden heart attacks, and sudden deaths due to abortion and child birth. Suicides accounted for between 25 and 32 per cent of the total behavioural deaths in the period 1971-2001. Road accidents emerge as the single most important cause accounting for around one-third of the accidental deaths. Now, for the two groups of causes, behavioural and non- behavioural, data on (1) mortality rates (crude death rates) of
SPECIAL ARTICLEaugust 9, 2008 EPW Economic & Political Weekly66males and females, and the index of sex-differentials in mortality (ISDM) for the years 1971 and 2001, and (2) the change in ISDM between 1971 and 2001 are presented in Table 1b. Per cent contri-butions by the behavioural and non-behavioural factors to ISDM and change in ISDM are also provided in Table 1b. The basic data employed to arrive at the estimates in Table 1b are presented in Table 1a. The numbers presented in Tables 1a and b are largely self- explanatory, and hence only the salient features are highlighted here. A simple calculation based on the numbers presented in Table 1a indicates that the share of behavioural deaths in total deaths for females and males, respectively, has gone up from 1.34 to 2.95 per cent and 2.27 to 5.60 per cent between 1971 and 2001. A similar simple calculation based on the numbers presented in Table 1a show that while behavioural deaths for females have gone up by a factor of around 220 per cent, that of males has gone up by a factor of around 247 per cent. The analysis of the trend, not reported here but will be made available on request, suggests that the share of mortality attribut-able to behavioural causes rose relatively more sharply after the mid-1980s, and such trend seems to continue. This suggests that the importance of behavioural causes in determining the obser-ved sex-differentials in mortality is likely to increase with time. The numbers in Table 1b indicate that in 1971 while behav-ioural mortality was against males, non-behavioural mortality was against females, and the latter swamped the effect of the former. Accordingly, at the overall level there was “excess” female mortality in 1971. However, “excess” female mortality due to non-behavioural causes has registered very steep declines between 1971 and 2001. Indeed, the decline has been so steep that these causes account for around 68 per cent of the “excess” male mortality in 2001. The numbers in the last row of Table 1b, show that while behav-ioural causes account for only around 9 per cent, non- behavioural causes account for 91 per cent of the change in ISDM between 1971 and 2001. This seemingly suggests that the decline in intra-family allocation of resources that affects the relative sur-vival chances of women vis-à-vis men has largely contributed to the disappearance of “excess” female mortality, and to the emergence of “excess” male mortality between 1971 and 2001. However, it needs to be recognised that to reach such a conclusion one needs to net out the decline in ISDM due to changes in genetic endowment of different sets of birth cohorts that constituted the populations in 1971 and 2001, and the change in age-distribution of mortality between 1971 and 2001. As noted earlier, it is difficult to identify the contribution of changes in genetic endowments of the sets of birth cohorts that constituted the populations in 1971 and 2001. However, it is possible to identify the contributions by different age-groups to the decline in ISDM between the two points of time. The contributions by different age-groups to the decline in ISDM, as suggested earlier, could pos-sibly be employed to infer on the sources of such decline in ISDM between 1971 and 2001.Contributions of Each Age-GroupTo identify the contributions by each age-group to the change in ISDM, age-group specific crude death rates need to be estimated for the years 1971 and 2001. This requires data on male and female populations, and the numbers of male and female deaths in each age-group. To esti-mate the numbers of male and female deaths in each age-group, data on (1) the total male and female populations enumerated by the population Census in 1971 and 2001, and (2)the age-distribution and the age-group specific mortality rates of males and females provided by the Sample Registration System (SRS) for the respective years are employed. Using the data from the two sources for the years 1971 and 2001, the population of males and females in each age-group have been estimated. For example in 1971, for age-group i, the male and female populations, respectively, have been estimated as: Pmi = (smi*Pm) and Pfi = (sfi*Pf). Notice here that smi and sfi, respectively, are the shares of male and female populations in age-group i as published in SRS for the year 1971, and Pm and Pf respectively, are the total male and female populations enumerated in the population Census of 1971. The same procedure has been employed to arrive at the age- distributions of the male and female populations in 2001. Making use of the estimated male and female populations in each age-group, and the age-group specific mortality rates (MRi) published in SRS, the numbers of male and female deaths in each age-group have been estimated for the years 1971 and 2001. For example, for age-group i in 1971, the numbers of male and female deaths, respectively, have been arrived at as Dmi = (MRmi/1000)*Pmi, and Dfi = (MRfi/1000)*Pfi. Using the data on Dmi, Dfi, Pm, and Pf for the years 1971 and 2001, the crude age-group specific death rates for the relevant years have been estimated employing equations 6 and 6’.Data on the age-structure, the age-group-wise estimated male and female populations for the years 1971 and 2001 are presented in Table 2a (p 67). In Table 2b (p 67), data on age-group-wise deaths and crude death rates are presented. Age-group specific mortality rates from SRS and the per cent change in mortality rates between 1971 and 2001 are provided in Table 2c (p 68). Information on the index of sex-differentials in mortality, the change in the index of sex-differentials in mortality between 1971 Table 1b: Data Relating to “Excess” Female MortalityYear/Period Behavioural Non-behavioural Index of ‘Excess’ Female Mortality Death Rates Death Rates MaleFemaleMaleFemaleISDMb ISDMnb ISDM1971 0.33340.206814.366615.1932 -0.1266(-18.08)0.8266(118.08)0.7000(100.00)2001 0.49310.23588.30697.7642-0.2573(32.16)-0.5427(67.84)-0.8000(100.00)Change in ‘Excess’ Female Mortality ΔISDMb ΔISDMnb ΔISDM1971-2001 -0.1307(8.71)-1.3693(91.29)-1.5000(100.00)Source: Estimated based on numbers presented in Table 1a.Table 1a: Population, Crude Death Rate and Number of Deaths, India (1971 and 2001)Year Population Crude Death Rate Estimated Total Deaths Behavioural Deaths(t) MaleFemaleMaleFemaleMaleFemaleMaleFemale (Pm) (Pf) (CDm) (CDf) (Dm) (Df) Dmb Dfb1971 283,936,614264,013,195 14.7 15.4 4,137,868 4,683,563 94,670 54,6062001 532,223,090 496,514,346 8.8 8.0 4,683,563 3,972,115 262,443117,082Total deaths of males and females for each year are estimated, respectively, as: Dm = (CDm/1000)* Pm and Df = (CDf/1000)*Pf. Notice that Dmnb=(Dm- Dmb) and Dfnb=(Df- Dfb).Source: Listed in Appendix 1.
SPECIAL ARTICLEEconomic & Political Weekly EPW August 9, 200867and 2001, and the contributions by each age-group to ISDM and change inISDM are presented in Table 2d (p 68). Before entering into a discussion on the features of the numbers, it needs to be pointed out that the SRS provide data on the shares of populations in each age-group corrected to a single decimal place. This correction has not led to any problem in 1971 as in the case of both males and females the age-group specific population shares add up to 100 per cent. In 2001, age-group-wise population shares, while for males add up to only 99.7 per cent, in the case of females they add up to 100.2 per cent. For males the dif-ference of 0.3 per cent (100-99.7) has been dis-tributed to each age group according to the share of each age-group in the sum total of age-group specific population shares at 99.7. To be more precise, for males, the corrected population share for age-group i has been obtained as: s*mi = smi+((smi/99.7)*0.3). Similarly, for females the corrected population share for age-group i has been estimated as: s*fi = sfi–((sfi/100.2)*0.2).The age-group specific corrected population shares of males and females, provided in Table 2a for 2001 have been employed in the analysis. Notice that these corrections do make a differ-ence. The published overall crude death rates of males and females, respectively, are at 8.8 and 8. Whereas the estimated overall crude death rates that corresponds to the corrected age-structures for males and females, respectively, are at 8.6968 and 8.1387. The differences between the pub-lished and the estimated absolute values of the crude death rates of males and females may appear insignificant. However, the estimate of ISDM that corresponds to the corrected age-structures at –0.5581 is around 30 per cent lower than that estimated employing the published fig-ures on crude death rates of males and females at –0.8. There is a case for publishing data accu-rate at least to three decimal places. Given the problem, as there is no other alternative, it is assumed that the estimates of “excess” female mortality based on the corrected age-structure is relatively more accurate.Data presented in Tables 2a to 2d are largely self-explanatory. For this reason only the impor-tant points that emerge from a glance at these numbers are highlighted here. The numbers in Table 2a indicate that the shares of both male and female populations in the first two age-groups (0-4 and 5-9) have registered sharp declines between 1971 and 2001. On the other hand one observes significant increases in the shares of populations in the age-groups beyond age 50. These changes are features of demographic development. Fertility has fallen, and expectation of life at birth has increased substantially in this period. Crude birth rate has fallen from 36.9 in 1971 [SRS 1970-75] to 25.4 in 2001 [SRS 2001]. Expectation of life at birth of person has gone up from 49.7 years [SRS 1970-75] at the beginning of the 1970s to 61.1 years by mid-1990s [Women and Men in India 2000]. The numbers in Table 2b and Table 2c indicate mortality rates have declined relatively more sharply in the first two age-groups 0-4 and 5-9 between 1971 and 2001. The decline in these age-groups is observed to be of the order of 60 per cent, far higher than that for any other age-group beyond age 9. The declines in the shares of populations and the relatively sharper declines in mortality rates in these age-groups have led to a decline in the contribution of these two age-groups to overall mortality in this period. Indeed, a simple calculation based on the numbers in Table2b show that the contribution by these age-groups to the total estimated deaths has declined from 54.29 per cent in 1971 to 28.81 per cent in 2001. The numbers presented in Table 2d show that between 1971 and 2001 “excess” female mortality has completely disappeared, indeed one observes “excess” male mortality in 2001 at Table 2a: Population Characteristics, India (1971, 2001)Age-group Share of Population Estimated Population 19712001 1971 2001 MaleFemaleMaleFemaleMalesFemalesMalesFemales0-4 14.4014.4811.73511.27740,886,87238,229,11162,457,47455,994,1335-9 14.2114.3310.53210.27940,347,39337,833,09156,051,57951,038,90010-14 12.7112.3111.93611.47736,088,34432,500,02463,525,12356,985,17915-19 9.719.1910.93310.18027,570,24524,262,81358,186,87750,543,37720-24 7.778.299.0279.38122,061,87521,886,69448,044,21146,579,19025-29 7.087.598.3258.28320,102,71220,038,60244,307,43941,128,43430-34 6.717.157.3227.58519,052,14718,876,94338,969,19337,659,77135-39 6.065.956.5206.88617,206,55915,708,78534,698,59734,191,10840-44 5.475.215.717 5.589 15,531,33313,755,087 30,428,00027,749,30545-49 4.333.994.7144.69112,294,45510,534,12625,089,75423,289,59550-54 3.863.613.711 3.593 10,959,953 9,530,876 19,751,50917,838,83955-59 2.502.39 3.009 3.094 7,098,415 6,309,915 16,014,73715,361,22260-64 2.342.462.2072.4956,644,1176,494,72511,744,14012,388,08265-69 1.27 1.341.9062.196 3,605,995 3,537,77710,142,667 10,901,51370+ 1.581.712.407 2.994 4,486,199 4,514,626 12,811,79014,865,699All 100.00100.00100.00100.00283,936,614264,013,195 532,223,090 496,514,346Source: Provided in Appendix 1.Table 2b: Age-group Specific Crude Death Rates, India (1971, 2001)Age-group Estimated Number of Deaths Crude Death Rates 1971 2001 1971 2001 MaleFemaleMaleFemaleMalesFemalesMalesFemales0-4 20,11,634 20,94,955 11,42,972 11,47,880 7.0848 7.9350 2.1475 2.31195-9 1,81,5631,85,3821,00,8931,02,078 0.6395 0.7022 0.1896 0.205610-14 75,786 68,250 82,583 68,382 0.2669 0.2585 0.1552 0.137715-19 52,383 72,788 87,280 96,032 0.1845 0.2757 0.1640 0.193420-24 66,186 94,1131,00,8931,16,448 0.2331 0.3565 0.1896 0.234525-29 60,308 86,166 1,15,1991,19,272 0.2124 0.3264 0.2164 0.240230-34 74,303 98,1601,28,598 94,149 0.2617 0.3718 0.2416 0.189635-39 94,636 94,2531,45,734 99,154 0.3333 0.3570 0.2738 0.199740-44 1,14,932 82,531 1,64,311 94,348 0.4048 0.3126 0.3087 0.190045-49 1,29,092 89,540 1,95,700 1,04,803 0.4547 0.3392 0.3677 0.211150-54 2,08,239 1,36,292 2,27,142 1,44,495 0.7334 0.5162 0.4268 0.291055-59 1,74,621 1,10,424 2,99,476 1,99,696 0.6150 0.4183 0.5627 0.402260-64 2,45,832 2,12,377 3,11,220 2,26,702 0.8658 0.8044 0.5848 0.456665-69 1,94,003 1,51,771 4,48,306 3,64,111 0.6833 0.5749 0.8423 0.733370+ 4,94,828 4,89,83710,78,343 10,63,418 1.7427 1.8554 2.0261 2.1418All 41,73,868 40,65,803 46,83,563 39,72,115 14.7158 15.4039 8.6968 8.1387Source: Based on data provided in Table 2a.
SPECIAL ARTICLEaugust 9, 2008 EPW Economic & Political Weekly68the overall level (see, columns 2 and 3 of the last row). The contri-bution of the age-group 0-4 to “excess” female mortality in 1971 was around 124 per cent, which has declined to –29.44 per cent in 2001. Notice here that in 2001 there is “excess” male mortality at the overall level. The youngest age-group 0-4 alone accounts for 55 per cent of the total improvement in sex-differentials in mortality in favour of females between 1971 and 2001. Infant mortality is the single most important component of mortality in the age-group 0-4. Infant mortality has been observed to be responsive to changes in fertility behaviour, particularly birth spacing [see, for example, Bhalotra and Soset 2007]; public spending on reproductive health-care [Bhalotra 2007]; and access to basic amenities or infrastruc-tural facilities such as electricity for lighting, safe drinking water, clean fuel, transport, and healthcare facilities [Jayaraj 2005]. Surprisingly, the age-group 70+ contributes to “excess” female mortality in both years 1971 and 2001. However, the female age-group specific death rates in both years 1971 and 2001 have been observed to be lower than that of males in this age-group. But, the female-to-male ratio of deaths at 990 in 1971 and 986 in 2001 in this age-group far exceed that of the female-to-male ratio of all deaths at 974 and 878 in the respective years 1971 and 2001. Notice, also that in 1971 only in this age-group there are more females than males, and in 2001 the female-to-male ratio in this age-group far exceeds that in the other age-groups. It seems possible that the population composition (sex-ratio) effect has swamped the relative survival advantage effect of females. This indicates that there is a case for decomposing “excess” female mortality into that attributable to population composition or sex-ratio effect, the mortality level effect, and the mortality differen-tial effect. Such attempt is beyond the scope of this paper. What is important to note here is the fact that the two age-groups 0-4, and 5-9, where there have been no significant declines in mortality disadvantage experienced by females, account for almost 60 per cent of the decline in ISDM between 1971 and 2001. The population shares in these age-groups have registered fairly sharp declines, and declines in mortality rate in these age-groups have been steeper compared to that for the other age-groups. These changes appear to be part and parcel of the process of demographic development induced, largely, by increase in expectation of life at birth and fall in fertility rate. The rest, nearly 456 per cent, of the improvements almost exclusively have been accounted for by the reproductive age-groups 15-19 to 45-49, where there have been substantial gains in relative survival advan-tage of females – the negative trend coefficients of the ratio of female-to-male mortality rates in these age-groups have been observed to be statistically significant. Declines in “excess” female mortality in the reproductive age-groups, it appears, could be attributed, largely, to declines in expo-sure to maternal mortality risks, and improvements in access to reproductive healthcare. The crude birth rate has fallen from 36.9 in 1971 to 25.4 in 2001. The share of births in higher orders (births in birth orders above 3) has declined from around 45 per cent in 1971 [Family Welfare Programme in India 1997-98], to 22 per cent in 2001 [SRS 2001]. Age-specific fertility rate in the age-group 15-19, where reproductive mortality risks associated with teenage pregnancy are more pronounced, has declined from 100.8 [SRS 1970-75] in 1971 to 47 in 2001 [SRS 2001]. The incidence of medical attention provided at birth by trained professionals (trained to attend birth) has improved considerably between 1971 and 2001. In 1971 only around 14 per cent of the total births in India were delivered in hospitals and medical institutions, and another 14.5 per cent of total births received assistance by trained professionals at home SRS (1970-75). On the other hand, in 2001, 26.3 per cent of the total births in India were delivered in hospitals and medical institutions, and another 18.2 per cent were assisted by trained professionals [sRS, Report 2001]. In this context, it is important to note that the increase in access to healthcare facilities run by the state appears to have contributed significantly to improvement in access to medical attention at birth. Information provided by the National Family Health Survey 1998-99 indicates that around 34 per cent of the births in the period 1996-97 to 1998-99 took place in institutions. Of the births delivered in institutions, 50 per cent Table 2c: Age-group Specific Mortality Rates, India(1971, 2001)Age-group Mortality Rates % Change in Mortality Rates 1971 2001 Between 1971 and 2001 Males Females MalesFemales Males Females0-4 49.20 54.80 18.30 20.50-62.80-62.595-9 4.504.901.802.00-60.00-59.1810-14 2.10 2.101.301.20-38.10-42.8615-19 1.903.001.501.90-21.05-36.6720-24 3.00 4.302.102.50-30.00-41.8625-29 3.00 4.302.602.90-13.33-32.5630-34 3.90 5.20 3.30 2.50-15.38-51.9235-39 5.50 6.004.202.90-23.64-51.6740-44 7.40 6.005.40 3.40-27.03-43.3345-49 10.50 8.50 7.80 4.50-25.71-47.0650-54 19.00 14.30 11.50 8.10-39.47-43.3655-59 24.60 17.50 18.70 13.00 -23.98-25.7160-64 37.00 32.70 26.50 18.30 -28.38-44.0465-69 53.80 42.90 44.20 33.40 -17.84-22.1470+ 110.30108.5084.17 71.54-23.69 -34.07All 14.7015.408.808.00-40.14-48.05Source: Provided in Appendix 1.Table 2d: Index of and Contributions to “Excess” Female Mortality, India(1971, 2001) Age-group Index of “Excess” Contribution to Index of Change in Contribution Female Mortality “Excess” Female Mortality Index of to Change in “Excess” Index of “Excess” FemaleMortalityFemaleMortality 19712001197120011971-20011971-20010-4 0.85020.1643123.5545-29.4426-0.685955.03545-9 0.06270.01609.1143-2.8702-0.04673.747110-14 -0.0084 -0.0174 -1.2207 3.1248-0.0090 0.725415-19 0.09120.029413.2544 -5.2712 -0.0618 4.957820-24 0.12340.045017.9278 -8.0556 -0.0784 6.291225-29 0.11400.023816.5618-4.2588-0.0902 7.237430-34 0.1101-0.0520 16.0009 9.3172-0.1621 13.007635-39 0.0237-0.0741 3.4440 13.2797 -0.0978 7.848940-44 -0.0922 -0.1187 -13.3953 21.2678 -0.0265 2.128445-49 -0.1155 -0.1566 -16.7841 28.0615 -0.0411 3.299850-54 -0.2172 -0.1358 -31.5585 24.3236 0.0814 -6.531955-59 -0.1968 -0.1605 -28.5912 28.7544 0.0363 -2.909260-64 -0.0614 -0.1282 -8.9196 22.9629 -0.0668 5.358965-69 -0.1084 -0.1090 -15.7524 19.5277 -0.0006 0.047670+ 0.11260.115716.3642-20.72130.0030-0.2444All 0.6881-0.5581100.00100.00-1.2463100.00Source: Based on data presented in Table 2c.
SPECIAL ARTICLEEconomic & Political Weekly EPW August 9, 200869(ie, around 17 per cent of total births) took place in public (i e, government hospitals/primary health centres) institutions. Notice here that the proportion of total births delivered in public institutions in the late-1990s exceeds the proportion of total births in institutions (including public and private medical institu-tions) in 1971. However, it is distressing to note that only around 55 per centof total births in 2001 received medical attention by trained professionals.Contribution of Age-Groups To sum up, the results presented in this paper indicate that the two age-groups (0-4 and 5-9) at the lowest end of the age-spectrum, and the reproductive age-groups (15-19, 20-24, 25-29, 30-34, and 35-39) account for almost the entire improvements in the relative survival advantage of females (ISDM) at the overall level between 1971 and 2001. The improvements accounted for by the reproduc-tive age-groups reflect reductions in mortality disadvantage against females. These reductions in mortality disadvantage, it appears, could be largely attributed to reductions in exposure to maternal mortality risks associated with fertility transition, and improvements in access to reproductive healthcare facilities pro-vided by the state run hospitals and primary healthcare centres. The contribution by the two age-groups (0-4, and 5-9) at the lowest end of the age-spectrum at 60 per cent to reductions in “excess” female mortality (change in ISDM) at the overall level is not even due to reductions in mortality disadvantage experienced by females in these age-groups – the ratio of mortality rates of females-to-males has remained more or less constant over time. The contribution is attributable to reductions in the shares of pop-ulations and gender neutral improvement in mortality in these age-groups. Accordingly, the reductions in “excess” female mortal-ity at the overall level do not signify reductions in discriminations in the intra-family allocation of resources in the spheres of healthcare and nutrition in India. These results also indicate the importance of disaggregated analysis of sex-differentials in mor-tality by age/age-groups and by causes of mortality to infer on the underlying process of change observed at the overall level.4 ConclusionsTill the mid-1980s, in India, the mortality differential was observed to be against females. The observed sex-differential in mortality was attributed to discrimination in intra-family allocation of resources, particularly in the spheres of access to healthcare and nutrition. By the mid-1980s, the mortality differential against females has disappeared. This disappearance of “excess” female mortality is attributed to a decline in discrimination against females in the intra-family allocation of resources. In this context, an attempt has been made to study the trends in the mortality rate and change in mortality differential disaggregated by age-groups. The complex interplay of biological, behavioural, social and cultural, economic, and environmental factors in determining sex-differentials in mortality has been highlighted. However, due to paucity of data required to disentangle the impacts of various factors, the attempt has been restricted to analysing the contribu-tions of behavioural factors, and change in mortality pattern across age-groups to the observed change in mortality differential in favour of females. It needs to be emphasised here that such an attempt is only exploratory in nature, and is neither exhaustive nor conclusive. However, the results of this exercise are found to be extremely instructive and useful.The age-groups 0-4 and 5-9 account for around 60 per cent of the declines in “excess” female mortality between 1971 and 2001. However, the ratio of female-to-male mortality rates has not reg-istered significant declines in these age-groups. This indicates that survival disadvantage experienced by females in these age-groups has not declined. The declines in mortality rates in these age-groups compared to other age-groups have been relatively steeper, and the shares of populations in these age-groups have registered perceptible declines. Accordingly, the contribution of these age-groups to overall mortality has declined between 1971 and 2001. For example, the share of the age-group 0-4 in the esti-mated total deaths which was around 50 per cent in 1971 has declined to around 26 per cent in 2001. Such a decline in the con-tributions by these age-groups to overall mortality is the source of 60 per cent of the improvement in mortality differential in favour of females at the aggregate level. Notice here that the decline in the shares of the population and the faster decline in mortality rates in the age-groups at the lower end of the age-spectrum are attributable to the process of demographic develop-ment. Thus, 60 per cent of the improvements in mortality differ-entials in favour of females at the aggregate level are attributable to the demographic development effect. More importantly, the results indicate that the trends in aggregate indices of relative well-being, particularly in the sphere of survival, hide more than what they reveal! There have been substantial reductions in “excess” female mortality in the reproductive age-groups. However, these reduc-tions, it appears, could be attributed largely to improvements in access to medical attention at birth provided by trained persons (trained to attend birth), and to declines in the incidence of expo-sure to maternal mortality risks between 1971 and 2001. Medical attention provided at birth by trained professionals has increased from 28.5 per cent in 1971 to 44.5 per cent in 2001. The improve-ment in access to medical attention at birth by trained profes-sionals, to a large extent, seems to be attributable to government investment in primary healthcare centres and government hospi-tals where healthcare is free. The crude birth rate, births to teen-age girls (in the age-group 15-19), and the incidence of higher order births (an indicator of physical exhaustion and nutritional depletion arising from excessive childbearing) have registered substantial declines in the period between 1971 and 2001. It is also important to note here that state interventions to reduce mortality (which includes provision of piped drinking water, an important intervention that probably reduces the incidence of diarrhoea – a major killer of infants) have a considerable impact on fertility and birth spacing [Bhalotra and Soset 2007]. Thus, the decline in “excess” female mortality in India does not neces-sarily indicate declines in “subtle” forms of discrimination in intra-family allocation of resources. The results, presented in the paper, are likely to be of interest to policymakers. The role of the state in expanding access to basic infrastructural facilities, particularly healthcare, to
SPECIAL ARTICLEaugust 9, 2008 EPW Economic & Political Weekly70reduce the effect of discrimination against females in intra-family allocation of resources is indicated. It appears that at least a part of the reductions in “excess” female mortality in the reproductive age-groups could be attributed to the increase in access to medical attention provided by trained professionals at birth through the state-run healthcare facilities. Notice also that “excess” female mortality still persists in the two reproduc-tive age-groups of 15-19 and 20-24, and as late as in 2001 more than 50 per cent of the total births in India have not received medical attention by trained professionals. These highlight the importance of maintaining pressure on the state to expand pub-lic healthcare facilities. The analysis also points to the impor-tance of attending to discrimination in survival in the early years of life, which requires identifying the causes of “excess” female mortality in the age-groups 0-4 and 5-9, and devising appropriate intervention strategies. Notes 1 While the exact magnitude of the primary sex ratio may be under considerable dispute, it appears that male preponderance at conception is accepted. 2 In most developed countries females had experi-enced excess mortality until 1930 in the age range 3-19. It is only towards 1940 that “excess” female mortality in every age/age-group had disappeared [see, Tabutin and Willems 1998]. It appears that there is a relationship between “excess” female mortality observed in the age-range 3-19 and the expectation of life at birth of a population. Such a relationship needs further probing. 3 There appears to be literature [see, for example, Goodkind 1996, Government of India 2003 and Visaria 2003] which suggests that discrimina-tion in infancy, particularly infanticide, has been substituted by foeticide. The result that indicates that mortality differential has not declined in the age-group 0-4, somewhat, con-tradicts this claim. 4 It is worth pointing out that the sex ratio at birth obtained in India towards the end of the 19th cen-tury appears to have had been even higher than 963. The National Sample Survey (1955) estimate – estimated employing the reverse survival method – suggests that the female-to-male ratio at birth at the turn of the 20th century was around 990, which probably indicates that only the fittest of the male embryos survived at birth in this period. 5 A large number of other factors could contribute to improvements in nutritional status of women during pregnancy [see, in this context, Jayaraj and Subramanian 2004]. Important among them are increase in age at marriage, reductions in reproductive burden achieved either through increase in age at marriage [Gopalan 1989] or by adopting deliberate birth control measures, improvements in the availability of food (caused by declines in the occurrences of famines and increase in productivity of land) and intra-family allocation of resources, particularly food alloca-tion in favour of females. 6 The contributions by the first two age-groups and the reproductive age-groups add up to more than 100 per cent as there are three age-groups which contribute negatively to the change in ISDM.ReferencesAaby, P (1992): ‘Lessons for the Past: Third World Evidence and the Reinterpretation of Developed World Mortality Declines’,Health Transition Review, 2 (Supplementary Issue), pp 155-81.Agarwal, D K, A Agarwal, M Singh, K Satya, S Agar-wal and K N Agarwal (1998): ‘Pregnancy Wastage in Rural Varanasi: Relationship with Maternal Nutrition and Socio Demographic Characteris-tics’,Indian Pediatrics, 35(11), pp 1071-79.Bhalotra, S (2007): ‘Spending to Save? State Health Expenditure and Infant Mortality in India’, Work-ing Paper No 07/168, Centre for Market and Pub-lic Organisation, Bristol Institute of Public Affairs, University of Bristol. Available at: http://www.bris.ac.uk/Depts/CMPO/workingpapers/wp169.pdf (accessed August 3, 2007).Bhalotra, S and A V Soset (2007): ‘Birth Spacing, Fertil-ity and Neonatal Mortality in India: Dynamics, Frailty and Fecundity’, Working Paper No 07/168, Centre for Market and Public Organisation, Bristol Institute of Public Affairs, University of Bristol. Available at: http://www.bris.ac.uk/Depts/CMPO/workingpapers/wp168.pdf (accessed August 3, 2007).Family Welfare Programme in India, 1997-98, Depart-ment of Family Welfare, Ministry of Health and Family Welfare, Government of India.Huxley, J (1926): Essays in Popular Science, Penguin Books, London (reprinted in 1937).Goodkind, D (1996): ‘On Substituting Sex Preference Strategies in East Asia: Does Prenatal Sex Selec-tion Reduce Postnatal Determination’, Population and Development Review, 22(1), pp 111-26.Gopalan, C (1989): ‘Women and Nutrition in India: General Considerations’ in C Gopalan and Suminder Kaur (eds),Women and Nutrition in India, Nutrition Foundation of India, pp 1-16. Government of India (2003): ‘Save the Girl Child’, Press Information Bureau, Government of India. Available at: http://pib.nic.in/release/rel_print_page.asp?relid=480 (accessed July 24, 2007). Planning Commission, Ninth Five-Year Plan (Vol 2): ‘Human and Social Development: 3.8 Empower-ment of Women and Development of Children’. Available at: http://planningcommission.nic.in/plans/planrel/fiveyr/9th/vol2/v2c3-8.htm (accessed July 27, 2007).Jannetta, A B (1992): ‘Famine Mortality in Nineteenth-Century Japan: The Evidence from a Temple Register’,Population Studies, 46(3), pp 427-43.Jayaraj, D (2005): ‘A Preliminary Analysis of the Rela-tionship between Generalised Deprivation and Infant Mortality in India’, CICRED Seminar on Mortality as Both a Determinant and a Conse-quence of Poverty and Hunger, Thiruvanan-thapuram, India, February 23-25, Available at: http://cicred.org/Eng/Seminars/Details/Semi-nars/Trivandrum2005/TrivandrumJAYARAJ.pdf (accessed August 3, 2007).Jayaraj, D and S Subramanian (2004): ‘Women’s Well-being and the Sex Ratio at Birth: Some Suggestive Evidence from India’,The Journal of Development Studies, 40(5), pp 91-119.Klasen, S and C Wink (2002): ‘A Turning Point in Gender Bias in Mortality? An Update on the Number of Missing Women’, Population and Deve-lopment Review, 28(2), pp 285-312.Kumar, G (1989): ‘Gender, Differential Mortality and Development: The Experience of Kerala’,Cam-bridge Journal of Economics, 13(4), pp 517-39. National Family Health Survey (NFHS-2) 1998-99: India, International Institute for Population Stud-ies, Mumbai, India and ORC Macro, Calverton, Maryland, USA, 2000. National Sample Survey (1955): Report on: Couple Fer-tility, Department of Economic Affairs, Ministry of Finance, Government of India, December.Pitkanen, K J (1993): Deprivation and Disease: Mortal-ity during the Great Finnish Famine of the 1860s, Publications of the Finnish Demographic Society, Helsinki.Pitkanen, K J and J H Mieklke (1993): ‘Age and Sex Differentials in Mortality during Two Nineteenth Century Population Crises’,European Journal of Population, Vol 9(1), pp 1-32.Seminar Report (1999): ‘The Demography of Famines: Perspectives from the Past and the Present’, Les Treilles, May 25-31. Organised by the IUSSP Com-mittee on Historical Demography and the ‘Fonda-tion Les Treilles’. Available at: http://www.iussp.org/Activities/scp-his/his-rep99.php (accessed July 9 2007).Sen, A (1990): ‘More Than 100 Million Women Are Missing’,New York Review of Books, 37(20). Avail-able at: http://ucatlas.ucsc.edu/gender/Sen100M.html (accessed October 19, 2006). – (1992): ‘Missing Women’, British Medical Journal, 304, pp 586-87. – (2003): ‘Missing Women – Revisited’ (editorial) British Medical Journal, 327, pp 1297-98.Sample Registration System 1970-75, Vital Statistics Division, Office of the Registrar General, India, Ministry of Home Affairs, 1982.Sample Registration System: Statistical Report 2001, Report No 1 of 2004, Office of the Registrar General, India. Tabutin, D, and M Willems (1998): ‘Differential Mor-tality by Sex from Birth to Adolescence: The His-torical Experience of the West (1750-1930)’ inToo Young to Die Genes or Gender?, United Nations, New York.Taket, A (1986): ‘Accident Mortality in Children, Ado-lescents, and Young Adults’,World Health Statis-tics Quarterly, Vol 39, No 3, 232-56. United Nations (1998): ‘The Extent and Cause of Female Disadvantage in Mortality: An Overview’ in Too Young to Die Genes or Gender?, United Nations, New York.Visaria, L (2003): ‘The Missing Girls’. Available at: http:// www.india-seminar.com/2003/532/532%20leela% 20visaria.htm#top (accessed July 24, 2007). Watkins, S C and J Menken (1985): ‘Famines in His-torical Perspective’,Population and Development Review, Vol 11, No 4, 647-676.Women and Men in India, 2000, Central Statistical Organisation, Ministry of Statistics and Pro-gramme Implementation, Government of India.Appendix 1Sources of DataData on Age-group Specific Mortality Rates, Per cent Distribution of Estimated Population by Sex and Age-group, and Overall Mortality Rates (Crude Death Rates) of Males, Females and Persons.Sample Registration System, 1970-75, Vital Statistics Division, Office of the Registrar General, India, Ministry of Home Affairs, 1982.Sample Registration System, 1976-78, Vital Statistics Division, Office of the Registrar General, India, Ministry of Home Affairs, 1981.Sample Registration System, Vital Statistics Division, Office of the Registrar General, India (for the various years since 1979 to 2004). Sample Registration System (1999): Compendium of India’s Fertility and Mortality Indicators: 1971-1997, based on Sample Registration System (SRS), Registrar General of India, New Delhi.Accidental Deaths and Suicides: Accidental Deaths & Suicides in India, National Crimes Records Bureau, Ministry of Home Affairs, Gov-ernment of India (for the years: 1996, 1997, 1998, 2002, 2003 and 2004). Male and Female Populations:Census of India, 1971, Union Primary Census Abstract, Part II-A(ii),V-I, Demography, Office of the Regis-trar General, India. New Delhi.Census of India, 2001, General Population Tables, A1 to A3, Part 1, Tables: India, States and Union Territories.

Dear reader,

To continue reading, become a subscriber.

Explore our attractive subscription offers.

Click here

Comments

(-) Hide

EPW looks forward to your comments. Please note that comments are moderated as per our comments policy. They may take some time to appear. A comment, if suitable, may be selected for publication in the Letters pages of EPW.

Back to Top