Intra-household Bargaining,Son Preference and the Status of Women in India

Abstract

Son preference prevails among Indian families. We test the hypothesis that women who bear sons experience an elevated status within the household, which translates into their increased role in decision-making. Using data from the Indian Human Development Survey, we find that women who have given birth to at least one son show greater participation in the household’s financial decisions as well as other decisions on a regular basis. After analysing 13 decision variables, we find that women who have given birth to at least one son enjoy higher overall bargaining power in the household. For example, in decisions concerning what to spend money on, having a son is equivalent to 6 years of formal education. We also show that the presence of a senior member in the household, however, often reduces the women’s household bargaining strength.

JEL Classification Codes: J11, J12, J13, D190

Keywords

Intra-household son preference decision

I. Introduction

The declining sex ratio (defined as the number of females per 1,000 males) and a strong son preference in most families (implying a general belief that sons are more valuable as offspring than daughters) in India are some of the most discussed topics in social science forums. The existence of a strong son preference in Indian society was empirically established (Arnold et al., 1996; Bhat & Xavier, 2003; Miller, 1981). Researchers have attributed cultural, economic, religious and social reasons to this strong son preference among Indian households (Bardhan, 1974; Basu, 1989; Miller, 1981; Sen, 1992; Sen & Sengupta, 1983).

This article tests whether giving birth to at least one son translates into empowerment for women in Indian households. Our hypothesis is that given the strong son preference in India, women who have sons are more likely to enjoy greater bargaining power in family matters, ranging from basic meal preparation to complex financial decisions. In addition, women with sons enjoy more respect from the elders of the household, especially in a joint family set-up, which is still prevalent in Indian society.¹ In this article we attempt to identify and quantify the preferential treatment given to women who have given birth to at least one son by examining an Indian data set at the family level. The household bargaining literature is relatively new. Recently Li and Wu (2011) established a connection between the issues of son preference and the overall bargaining strength enjoyed by women with sons using the data from China. Our literature review will focus more on India, in which a strong son preference, the dowry system, joint family system, caste and religion all play important roles.

The remainder of this article is organised as follows. Section II reviews the existing literature. Section III describes our data. Section IV discusses the econometric specifications. Our main empirical results are presented in Section V. Section VI concludes the article.

II. Literature Review

The reasons for son preference vary across different societies and are determined by diverse factors including geographical location, economic situation and religious beliefs and customs. The relationship between the economic well-being of a family and family size has been examined since as far back as the early theoretical work by Malthus. Similarly, Easter Boserup’s conceptual elaboration of the ‘economies of female survival’ argues that the nature of women’s participation in the labour force determines the extent of discrimination. In other words, the role of women’s labour, when translated from the domain of invisibility (household) to a visible economic role (participation in the labour force as an economic agent), establishes the socio-economic worth of her existence. However, this possibility of increased participation of women in the workforce gets sidelined when sociocultural beliefs dominate an economic rationale. A daughter is considered as a double loss in India, since the investment in her upbringing and education benefits the family that she is going to be married into more than the family in which she was born, coupled with the expenditure on her dowry that has to be incurred by her family (Basu, 1992). Investment in girls depends heavily on the economic condition of the family (Bhalotra, 2009). In India, sons, on attaining adulthood, are expected to provide economic support for their parents (Das, 1984; Lahiri, 1984; Miller, 1981; Operations Research Group 1990). In contrast, daughters represent a substantial economic burden in places where their parents are expected to provide a dowry to the families into which they marry. The other social order argument is that sons would assist the family financially through their earnings during their parents’ old age (Rangamuthia et al., 1997). Ethnographic evidences also suggest that scores of rituals that are considered to be essential for successful family life are performed by males, from birth to the lighting of the funeral pyre. Some other scholars have also suggested that the economic constraints of managing a family, coupled with the prevalence of the patriarchal system of Indian society, are the main reasons for this manifested gender bias (Agnihotri, 1996; Kishore, 1993). Also, regions with fertile land and heavy reliance on agriculture are more likely to exhibit a strong son preference, since having a male child is more beneficial and productive than a female child for families who have to undertake labour-intensive work in the fields.

Another important determinant of son preference pertains to the dowry that has to be paid at the time of a daughter’s marriage. Most families in India, especially in north India, follow the practice of dowry. Among some castes and communities, the amount of the dowry to be paid is quite substantial, often exceeding the average lifetime household income of agricultural labourers. Although the practice of dowry is illegal in India, it is still widely prevalent in the society, particularly in Hindu marriages. Hindus are usually rigid about having marriages in the same caste. The practice of dowry is especially important in marriages when women are married to wealthier men within the same caste group (Caldwell et al., 1983). Dowry can be seen as economic compensation made by the bride’s father to the groom’s family for accepting the girl in their household (Rao, 1993). The practice of dowry is, however, not so prevalent among the Muslims.

Another important determinant that plays an important role in fuelling the son preference in Indian society is religion. In India, there are two major religious groups, Hindus, who constitute around 80% of the total population, and Muslims, comprising around 13.4% of the total population of India (Census, 2001). Also, the fertility rate among Hindus is 2.59 while it is 3.4 among Muslims (NFHS-3, 2007). There are many explanations for this large difference in fertility among the two groups. The most commonly cited reason among scholars is the difference in contraceptive use, pregnancy termination methods and son preference (Iyer & Borooah, 2004). The use of contraceptives is very low among Muslims compared to other religious groups. It is estimated that only 37% of the Muslims use contraception as opposed to 49% among Hindus (International Institute for Population Sciences and ORC Macro International, 2000). This is because popular perceived religious notions among Muslims prohibit them from using contraception and terminating pregnancy, though some scholars have argued that the Muslim religion does allow birth control (Sikand, 1993). Even though there are differences in the fertility rate among the two religious groups, both religions directly or indirectly preach son preference. This is because, in both religions, the last rites of a person after his/her death are always performed by sons and women are not allowed to participate in any of these rituals (in most cases, they are not even allowed to enter the burial or cremation sites). Therefore, it is considered inauspicious and even a curse on a family if there is no boy in the family to perform the last rites of his elders. It is because of these retrograde practices that even religious leaders and priests offer newly wedded couples blessings for having sons.

The fertility rate in India has consistently declined over the last 15 years, though it is still quite high, mainly due to awareness drives initiated by the government, civil societies and NGOs. Unlike China, India has no family planning laws but only awareness drives to achieve population control. Alarmingly, however, the combination of a declining fertility rate and a strong son preference leads to another disastrous situation for not only the country but the Indian subcontinent as a whole – the declining sex ratio. Some researchers have linked the declining sex ratio with the government’s policy of propagating a two-child norm (Visaria et al., 2006). This norm has been introduced to help stabilise the population through a reduction in the fertility rate of the country. Nevertheless, this government policy, accompanied by the advent of modernisation and the consequent urbanisation and preference for the nuclear family system, has failed to reduce the age-old son preference in the country (Das, 1987; Kulkarni, 1999; Malhotra et al., 1995). According to a study conducted in one of the prosperous states of India, namely, Punjab, by Das Gupta (1987), when the size of the family in the popular social construct was large, the preference for a son was latent. However, the recent trend of promoting a small nuclear family and the consequent decline in fertility rates have obviously led to shrinking of the family size, which has brought the son preference to the forefront and consequently adversely affected the birth rate of girls, thereby leading to a drastic fall in the sex ratio. A few other regional studies also suggest that the cohort sex ratio at birth is masculine in some parts of country, particularly in northern and western India. After the introduction of sex determination technology and the selective abortion of girl babies, that is, the practice of female foeticide, advertisements began to appear on the walls of big and small cities by private doctors practicing such tests claiming, ‘Pay ₹500 (US$ 10) today rather than ₹500,000 (US$ 10,000) later.’ These attractive advertisements were specifically targeted at prospective families that would choose to abort the female foetus in order to avoid having to incur expenditure on dowry later (Mazumdar, 1994).

It is estimated that this epidemic (female foeticide) has cost more than 100million female lives all over the world (Sen, 1990) and around 35–37million lives in India alone (Dreze & Sen, 1996). A majority of the developed countries have a sex ratio in favour of girls. For instance, in the United States and Europe, there are 1,050 females per 1,000 males, whereas in India, there are just 933 females per 1,000 males (Census, 2001). At the regional level, many states present an even sorrier picture. States like Haryana and Punjab, two of the relatively rich states in India, for example, have less than 900 females per 1,000 males.

Das Gupta and Bhat (1997) examine the changes in juvenile sex ratios (0–4 years), the mortality sex ratio and fertility rates for the period of 1981–1991. They conclude that during the period of declining fertility in India, parents were not substituting prenatal for postnatal discrimination against girls but were actually combining these two strategies. The bias towards a male child in India thus appears to be intensifying. Some other researchers suggest that the low autonomy among women combined with strong son preference is the major factor responsible for the sustained decline in fertility, rise in neglect of girls and female infanticide in India (Basu, 1992; Das, 1987; Dreze & Murthi, 2001; Dyson & Moore, 1983; Kulkarni, 1999). This can be attributed to the existing sociocultural order in the country, which considers women inferior to men in the society. Sen (1990) introduced the concept of ‘missing women’, forcefully raising the concern for the missing girl children in the public domain at the international level. Subsequently, a UNEPA Report entitled ‘India Toward Population and Development Goal’ points out that 48million women were actually ‘missing’ in India. According to this report, 40–50million girls have gone missing in India since 1901, missing either because they were not allowed to be born or, if born, were killed immediately thereafter.

Although women in India have traditionally not been empowered enough to make major decisions independently, the social norms have evolved recently. The entry of women into the labour market and their heightened awareness about their rights brought significant changes. Women now not only participate in family matters to a much larger extent than before, but are at times even equally responsible for major decision-making at the household level. Anecdotal evidence suggests that women who have given birth to son(s) are treated better than those women who have not, even within the same household for the reasons discussed above. Earlier studies have claimed that women’s relative bargaining power in the household is generally dependent upon many factors such as income from employment (Folbre, 1984; Thomas, 1990), the amount of dowry that a woman brings to the household at the time of marriage (Zhang & Chan, 1999), and the assets held by the individual (Brown, 2009). However, it is difficult to distinguish the income effect from the bargaining effect for women (Behrman, 1997).

Studies on the role of women in terms of financial decisions for the household have confirmed that women are more sensitive to the needs of households than their personal needs when asked to make household decisions (Ashraf, 2009). Similarly, in the Philippines, women tend to commit to depositing more savings in joint accounts than to receive personal goods (Ashraf, 2009). Women also prefer to spend more on health and education than their husbands when asked to make these decisions (Thomas, 1990). Overall, these studies indicate that women are good managers and use the household resources efficiently.

While women’s income, the assets owned by them, and the dowry that they bring into the household are important determinants of the power enjoyed by them in the household, this is not all. Income-related measures have endogeneity issues (Li & Wu, 2011). Dowry and assets often affect the current consumption factor and cannot be treated as permanent and certainly not as the only factors. Li and Wu (2011) have proposed women’s contribution to the next generation as exogenous determinants which can influence her bargaining power in the household. They argue that given the strong son preference in China, women giving births to sons are more likely to have a greater say in the household and would be given preferential treatment over women without sons.

Youssef (1978) reports that a woman’s status is enhanced by marriage and children, which improves her bargaining position in the household. In a society characterised by a strong son preference, women not only acquire a superior status after giving birth to a son, but their sons also clearly act as protective shields or a sort of insurance against the threat of divorce or abandonment by their husbands. Given this argument, it is obvious that women themselves would also exhibit a high degree of son preference along with the other household members. It is for this reason that women are expected to continue their childbearing activities throughout their reproductive years to fructify the universal desire of giving birth to sons (Youssef, 1978).

III. Data

This article uses India Human Development Survey (IHDS) data, jointly administered by the University of Maryland and the National Council of Applied Economic Research in India. The IHDS is a nationally representative survey of 41,554 households, covering a total of over 215,000 individuals. The survey includes households from all the 33 states and all Union Territories of India. It covers a wide range of topics and includes health, education, employment status, marriage, gender relations, fertility, income and employment, among other things. The IHDS includes eight data files that can be merged with each other through a unique identifier.

During the course of this study, we have made extensive use of the Education and Health Questionnaire file, which includes a set of questions posed to ever-married women in the age group of 15–50 years. For the purpose of this research, we match data relating to ever-married women with their fertility history. Further, we have only considered women in the age group of 21–50 years for this article, primarily because the legal age of marriage in India is 18 years. The Education and Health Questionnaire has a detailed section on ‘Gender Relations’, wherein ever-married women are asked about their roles in decision-making within the household. For example, ever-married women are asked who in the family decides about the number of children that the couple should have, among other decisions. There are several questions, which explicitly highlight the women’s decision-making role in the household. However, the questions included in the survey are not limited to decisions taken within the household but extend beyond that to measure the overall bargaining strength and position of women in the household. Questions as to whether the respondent has a bank account or any property ownership documents in her name also directly measure the respondent’s position in the household. These responses provide us with direct measurements of women’s bargaining power in the household.

IV. Model

In order to carry out our analysis on the effect of having at least one son on the overall bargaining strength of women in the household, we resort to the following binary dependent variable specification:

Prob (Y = 1) = G (Having son(s), Income, Education, Age, Urban, other covariates),

where Y is the women’s role in the household decision-making process, ‘“Having son(s)’ is a dummy variable indicating whether the woman has at least one son. In this analysis, we are not interested in the number of sons born or the gender of the first-born child simply because in India there is no law to limit the number of children in a family. Families with a strong son preference can keep on trying until they have a son. In the model, we are controlling factors for various individual traits, like the age of the women, education of the ever-married women, the principal occupation of the household, the education level of the head of the household, the presence of senior citizen(s) in the household, rural–urban disparities and variables associated with region, socio-religious group and the income of the household. Most of the variables are self-explanatory. We are controlling for the presence of one or more senior citizen(s) in the household, irrespective of the latter’s gender. We define household member aged 65 or above as senior. As discussed earlier, the joint family system continues to prevail in India. Having a senior member in the family affects the decision-making capabilities of the eligible woman. Similarly, the educational attainment of the head of the household also matters. Women are more empowered in the households where the head of the household is educated. We also include dummy variables for socio-religious groups, taking into account the numerous cultural and religious practices prevalent within various households. A common perception is that the household in which the principal occupation of the members is service (which can also be a proxy for a wealthier and more educated head of the household) accords more freedom to women as compared to the household in which the principal occupation is agricultural labour (proxy for the economically backward class). However, past research has shown conflicting results on the relationship between wealth and son preference among the households (Pande & Malhotra, 2006).

Overall, we have 13 decision variables as the dependent variables for this study, all of which reflect different perspectives of the women’s bargaining strength in the household.

V. Results

We primarily use the logit model in estimation. We also use the probit model and the linear probability model (Ordinary Least Squares or OLS) for robustness check. The results from logit, probit and OLS are reported in the tables. We have run the regression for all 13 dependent variables.

Table 1 reports the summary statistics of the data used for this research.

Table 1.

Summary Statistics: Women with at Least One Son and Women with at Least One Child and No Son.

Variable	N	Mean	Std. Dev	Min	Max
(a) Women with At Least One Son
Income quintiles	25,592	2.893912	1.401822	1	5
Socio-religious groups	25,592	2.742576	1.593995	1	7
Women age	25,588	34.68739	7.197612	21	50
Women education (years)	25,228	4.250436	4.680592	0	15
Household’s principle occupation	25,592	3.098664	1.596344	1	6
Household head education	25,076	0. 9416972	1.455717	0	6
Presence of senior citizen in the household	25,592	0. 1960769	0.3970351	0	1
Urban	25,592	0. 3515552	0.4774652	0	1
Region	25,592	4.553532	1.910042	1	7
(b) Women with At Least One Child and No Son
Income quintiles	4,286	3.14419	1.413297	1	5
Socio-religious groups	4,286	2.687821	1.647055	1	7
Women age	4,286	31.14232	7.364445	21	49
Women education (years)	4,255	6.147356	5.119186	0	15
Household’s principle occupation	4,286	3.330845	1.594979	1	6
Household head education	4,206	1.201854	1.619514	0	6
Presence of senior citizen in the household	4,286	0.2027531	0.402097	0	1
Urban	4,286	0.4349043	0.4958023	0	1
Region	4,286	5.006066	1.860844	1	7

Source: The authors.

Table 2 reports the results for the decision variable ‘whether the woman decides the number of children to have’. Column 1 of the table presents the coefficients of the logit regression. Columns 2 and 3 report the marginal effects from the logit and probit regression, respectively. Column 4 reports the OLS coefficients (marginal effects in the linear probability model). These results indicate that women with sons are 1.84% less likely to participate in the decision pertaining to the number of children. Older and more educated women in the household exhibit some decision-making power. It can be concluded that the presence of senior members in the household negatively impacts the decision-making capability of younger women by 2.11% points and that women in the urban areas have greater say by 3.07% in these matters than do women in the rural areas. All the above estimates are statistically significant at 1%.

Table 2.

Women’s Participation in Decision-Making on ‘How Many Children to Have’.

	Logit Coefficients	Logit Marginal Effects	Probit Marginal Effects	OLS Coefficients
Having sons	–0.116***	–0.0184***	–0.0182***	–0.0189***
	(0.0430)	(0.00699)	(0.00705)	(0.00705)
Women age	0.0189***	0.00292***	0.00300***	0.00295***
	(0.00216)	(0.000333)	(0.000338)	(0.000338)
Women education	0.0156***	0.00241***	0.00255***	0.00245***
	(0.00421)	(0.000651)	(0.000656)	(0.000664)
Presence of senior citizen in the household	–0.140***	–0.0211***	–0.0211***	–0.0204***
	(0.0402)	(0.00589)	(0.00594)	(0.00585)
Urban	0.195***	0.0307***	0.0317***	0.0322***
	(0.0387)	(0.00618)	(0.00625)	(0.00634)
Per capita income dummies	Yes	Yes	Yes	Yes
Principle occupation dummies	Yes	Yes	Yes	Yes
Socio-religious group dummies	Yes	Yes	Yes	Yes
Head of the household education dummies	Yes	Yes	Yes	Yes
Geographic region dummies	Yes	Yes	Yes	Yes
N	28,250	28,250	28,250	28,250
R ²				0.032

Source: The authors.

Note: Column 1 reports the coefficients for logit regression. Columns 2 and 3 report the marginal effects from logit and probit regression, respectively. Column 4 is the coefficients from OLS regression. Heteroskedastic robust standard errors are in parentheses.

*p < 0.1, **p < 0.05, ***p < 0.01.

Table 3 presents the results for the decision variable what to spend money on. Women with sons are 2.88% more likely than women without sons to participate in this decision. Having a son is equivalent to 6 years of formal education for the woman. Also, older and more educated women show greater participation. All these marginal effects are statistically significant. Women in the rural areas enter into more frequent discussions with their husbands by 1.22% than women in the urban areas. This could be attributed to the fact that couples in rural areas spend more time together than their counterparts in the urban areas, since the former usually work together in the fields or run small businesses from their homes. Also, as compared to the poorest people, that is, people in the lowest income quintile, more affluent couples obviously indulge more in discussions regarding purchase decisions and what to spend the money on, with the gap between the discussions carried out by couples in the two quintiles being about 1.75%. These estimates are statistically significant at a 1% level of significance. However, having an elderly member in the household decreases the frequency of discussion among the couple by 1.04%.

Table 3.

Women’s Participation in Decision-Making on ‘What to Spend Money On’.

	Logit Coefficients	Logit Marginal Effects	Probit Marginal Effects	OLS Coefficients
Having sons	0.299***	0.0288***	0.0309***	0.0310***
	(0.0524)	(0.00549)	(0.00575)	(0.00583)
Women age	0.00461*	0.000408*	0.000419	0.000477*
	(0.00279)	(0.000247)	(0.000259)	(0.000275)
Women education	0.0552***	0.00488***	0.00511***	0.00507***
	(0.00547)	(0.000480)	(0.000502)	(0.000493)
Presence of senior citizen in the household	–0.114**	–0.0104**	–0.0108**	–0.0106**
	(0.0484)	(0.00452)	(0.00474)	(0.00479)
Urban	–0.135***	–0.0122***	–0.0129***	–0.0136***
	(0.0491)	(0.00448)	(0.00473)	(0.00501)
Per capita income dummies	Yes	Yes	Yes	Yes
Principle occupation dummies	Yes	Yes	Yes	Yes
Socio-religious group dummies	Yes	Yes	Yes	Yes
Head of the household education dummies	Yes	Yes	Yes	Yes
Geographic region dummies	Yes	Yes	Yes	Yes
N	28,331	28,331	28,331	28,331
R ²				0.032

Source: The authors.

*p < 0.1, **p < 0.05, ***p < 0.01.

The next decision variable is ‘what to cook on daily basis’. The decision may seem trivial to those not familiar with the Indian culture. In the traditional Indian society, the selection of dishes for the daily menu is often viewed as a proxy for exercising sustained control over the daily affairs of a household. Table 4 provides estimates for the dependent variable: whether the woman makes this decision. If a woman has a son, her role in this decision increases by 3.12%. The decision-making power also marginally increases with age and decreases with education. More educated women usually work outside the household and therefore participate less in daily cooking than women who are exclusively home-makers. When there is a senior member in a household, the decision-making power of the younger woman decreases by 12.4% points, which indicates the importance of this decision-making variable within the household given that usually the job of preparing meals is left to the younger woman in the household.

Table 4.

Women’s Participation in Decision-Making on ‘What to Cook on Daily Basis’.

	Logit Coefficients	Logit Marginal Effects	Probit Marginal Effects	OLS Coefficients
Having sons	0.180***	0.0312***	0.0325***	0.0329***
	(0.0410)	(0.00734)	(0.00745)	(0.00730)
Women age	0.0482***	0.00803***	0.00797***	0.00785***
	(0.00224)	(0.000366)	(0.000367)	(0.000351)
Women education	–0.0210***	–0.00349***	–0.00353***	–0.00351***
	(0.00401)	(0.000666)	(0.000679)	(0.000673)
Presence of senior citizen in the household	–0.670***	–0.124***	–0.126***	–0.123***
	(0.0345)	(0.00698)	(0.00695)	(0.00678)
Urban	0.167***	0.0274***	0.0275***	0.0269***
	(0.0395)	(0.00640)	(0.00646)	(0.00633)
Per capita income dummies	Yes	Yes	Yes	Yes
Principle occupation dummies	Yes	Yes	Yes	Yes
Socio-religious group dummies	Yes	Yes	Yes	Yes
Head of the household education dummies	Yes	Yes	Yes	Yes
Geographic Region Dummies	Yes	Yes	Yes	Yes
N	28,513	28,513	28,513	28,513
R ²				0.071

Source: The authors.

*p < 0.1, **p < 0.05, ***p < 0.01.

The prevalence of the parda (veil) system in India signifies a centuries-old tradition that prohibits Indian women from openly facing males, particularly men who are not part of the family. In fact, in many sections of the Indian society, especially in the rural areas, women are not even allowed to show their faces to elderly family members including the father-in-law, brothers-in-law (who are elder to the husband) and other distant male relatives. Various theories have been offered to explain this practice. These include the need to accord respect to elders, maintain the dignity and privacy of the women concerned, and often even superstitious beliefs that forbid women from exposing their faces to strangers. This restriction is especially rigid for pregnant women and mothers with new-born children. This practice is more rampant in rural areas, where a child’s sickness is often believed to be caused by somebody’s evil eye on the child. These beliefs are reinforced due to the strong son preference in India. Sons and their mothers become more precious for the household than the other family members. From the estimates in Table 5 we can see that women with sons are 4.70% more likely to follow the parda system than those without sons. This finding reflects the irony of the situation as the parda system is a veritable sign of suppression in modern India. On the other hand, there are indications that women who deliver sons are more likely to enjoy greater powers and privileges, and by implication they would therefore be less suppressed and more empowered. In this context, therefore, the parda system, while definitely symbolising suppression, to some extent, can also be viewed as a reflection of the protective instinct of the family that wishes to protect its women and their sons from ‘evil eyes’. As discussed above, since this is a religious and cultural phenomenon, it is followed by most Hindu and Muslim families. As compared to higher caste Hindus, however, Christian and Sikh women are less likely by at least 25% to follow the parda system. On the other hand, the practice of parda among Muslim women exceeds that of higher caste Hindu women by 43.4%. (These estimates are not reported in the table. However, they are available upon request.) The parda system shows a decline with an increase in age and the level of education among the women, and though the magnitude of the decline is less, the estimates are statistically significant. The education level of the head of the household also plays an important role. If the head of the household has more education, the wife in the family is less likely to follow this system. As discussed earlier, the parda system is, however, on the decline in modern India, with education playing a very important role in its downfall. If the household is headed by an individual who has completed at least the first year of college then the women in that household are 9.41% less likely to practice the parda system than women belonging to a household headed by an illiterate. Urban women are 15.0% less likely to practice the parda system than rural women.

Table 5.

Do You Practice Ghungat/Purdah/Pallu?

	Logit Coefficients	Logit Marginal Effects	Probit Marginal Effects	OLS Coefficients
Having sons	0.188***	0.0470***	0.0437***	0.0290***
	(0.0436)	(0.0109)	(0.0101)	(0.00691)
Women age	–0.0175***	–0.00438***	–0.00404***	–0.00274***
	(0.00212)	(0.000530)	(0.000497)	(0.000342)
Women education	–0.0421***	–0.0105***	–0.00974***	–0.00678***
	(0.00401)	(0.00100)	(0.000943)	(0.000673)
Presence of senior citizen in the household	0.0341	0.00852	0.00836	0.00885
	(0.0370)	(0.00922)	(0.00869)	(0.00608)
Urban	–0.605***	–0.150***	–0.139***	–0.100***
	(0.0389)	(0.00949)	(0.00894)	(0.00624)
Per capita income dummies	Yes	Yes	Yes	Yes
Principle occupation dummies	Yes	Yes	Yes	Yes
Socio-religious group dummies	Yes	Yes	Yes	Yes
Head of the household education dummies	Yes	Yes	Yes	Yes
Geographic region dummies	Yes	Yes	Yes	Yes
N	28,808	28,808	28,808	28,808
R ²				0.347

Source: The authors.

*p < 0.1, **p < 0.05, ***p < 0.01.

Regressions estimates for the question whether the woman undertakes the daily shopping in the household can be found in Table 6. Women with sons are 1.05% more likely to shop for the household, though statistically insignificant. Older women are more likely to do the daily shopping. More educated women participate in the family shopping less. If there are senior members in the household, the younger woman is 7.77% less likely to shop. Women in urban areas are 11.6% more likely to shop.

Table 6.

Whether You Do the Food and Vegetables Shopping in Your Household.

	Logit Coefficients	Logit Marginal Effects	Probit Marginal Effects	OLS Coefficients
Having sons	0.0448	0.0105	0.0101	0.00930
	(0.0384)	(0.00904)	(0.00883)	(0.00769)
Women age	0.0359***	0.00840***	0.00820***	0.00717***
	(0.00194)	(0.000451)	(0.000440)	(0.000379)
Women education	–0.0296***	–0.00692***	–0.00665***	–0.00564***
	(0.00372)	(0.000868)	(0.000845)	(0.000720)
Presence of senior citizen in the household	–0.325***	–0.0777***	–0.0754***	–0.0656***
	(0.0334)	(0.00811)	(0.00786)	(0.00685)
Urban	0.510***	0.116***	0.115***	0.0980***
	(0.0353)	(0.00780)	(0.00772)	(0.00692)
per capita income dummies	Yes	Yes	Yes	Yes
Principle occupation dummies	Yes	Yes	Yes	Yes
Socio-religious group dummies	Yes	Yes	Yes	Yes
Head of the household education dummies	Yes	Yes	Yes	Yes
Geographic region dummies	Yes	Yes	Yes	Yes
N	28,741	28,741	28,741	28,741
R ²				0.171

Source: The authors.

*p < 0.1, **p < 0.05, ***p < 0.01.

Another important indicator of the bargaining strength of women in the household is the leisure time that they spend with their husbands outside of the household. The question whether they could find time to go for outings with their husbands and children was asked. Table 7 provides the estimates for the above question. The results suggest that women who have given birth to at least one son are 0.58% more likely to spend leisure time with their husbands than women who do not have sons. This also depends, to a great extent, on the household income as it has been observed that women in the highest income quintile go for family outings 12.4% more as compared to women in the lowest income quintile. Similarly, older women get to spend less time on leisure activities than their younger counterparts. These estimates are statistically significant. An extra year of education increases the probability by 1.77%. The presence of the elderly in the household negatively impacts the time available for outings. As expected, urban women are 14.1% more likely to spend time with their husbands than rural women.

Table 7.

Do You and Your Husband Sometimes Go Out by Yourselves (or with Children) to Movies, Fair, Restaurants?

	Logit Coefficients	Logit Marginal Effects	Probit Marginal Effects	OLS Coefficients
Having sons	0.0231	0.00575	0.00560	0.00577
	(0.0373)	(0.00930)	(0.00903)	(0.00800)
Women age	–0.0218***	–0.00544***	–0.00535***	–0.00485***
	(0.00183)	(0.000456)	(0.000444)	(0.000400)
Women education	0.0711***	0.0177***	0.0174***	0.0158***
	(0.00348)	(0.000867)	(0.000846)	(0.000762)
Presence of senior citizen in the household	–0.0376	–0.00936	–0.00909	–0.00810
	(0.0323)	(0.00807)	(0.00787)	(0.00713)
Urban	0.573***	0.141***	0.139***	0.128***
	(0.0334)	(0.00803)	(0.00787)	(0.00744)
Per capita income dummies	Yes	Yes	Yes	Yes
Principle occupation dummies	Yes	Yes	Yes	Yes
Socio-religious group dummies	Yes	Yes	Yes	Yes
Head of the household education dummies	Yes	Yes	Yes	Yes
Geographic region dummies	Yes	Yes	Yes	Yes
N	28,753	28,753	28,753	28,753
R ²				0.120

Source: The authors.

*p < 0.1, **p < 0.05, ***p < 0.01.

Tables 8 and 9 provide estimates for the questions as to whether women have to seek permission from their elders for participating in other forms of social interaction like visiting the local kirana shop (grocery store; Table 8) and friends and family (Table 9). It is clear from the marginal effects that need to seek permission variable goes down with higher education and age for both forms of social interactions. Another important result which emerges from this analysis is that if there are senior members in the household, the women are more likely to seek permission to visit the kirana shop and their friends and families by 3.81% and 4.22%, respectively. These estimates are statistically significant at a 1% level of significance.

Table 8.

Whether Permission Is Required to Visit Kirana Shop.

	Logit Coefficients	Logit Marginal Effects	Probit Marginal Effects	OLS Coefficients
Having sons	0.0412	0.0102	0.0101	0.00577
	(0.0388)	(0.00962)	(0.00954)	(0.00800)
Women age	–0.0173***	–0.00429***	–0.00423***	–0.00485***
	(0.00193)	(0.000478)	(0.000473)	(0.000400)
Women education	–0.00627*	–0.00155*	–0.00153*	0.0158***
	(0.00368)	(0.000911)	(0.000904)	(0.000762)
Presence of senior citizen in the household	0.155***	0.0381***	0.0375***	–0.00810
	(0.0350)	(0.00856)	(0.00849)	(0.00713)
Urban	–0.0207	–0.00512	–0.00543	0.128***
	(0.0346)	(0.00858)	(0.00852)	(0.00744)
Per capita income dummies	Yes	Yes	Yes	Yes
Principle occupation dummies	Yes	Yes	Yes	Yes
Socio-religious group dummies	Yes	Yes	Yes	Yes
Head of the household education dummies	Yes	Yes	Yes	Yes
Geographic region dummies	Yes	Yes	Yes	Yes
N	23,982	23,982	23,982	28,753
R ²				0.120

Source: The authors.

*p < 0.1, **p < 0.05, ***p < 0.01.

Table 9.

Whether Permission Is Required from Husband or Elders in the Household to Visit Friend or Family.

	Logit Coefficients	Logit Marginal Effects	Probit Marginal Effects	OLS Coefficients
Having sons	0.0543	0.0100	0.0101	0.0105
	(0.0409)	(0.00760)	(0.00762)	(0.00746)
Women age	–0.0301***	–0.00549***	–0.00555***	–0.00546***
	(0.00200)	(0.000364)	(0.000368)	(0.000363)
Women education	–0.00280	–0.000510	–0.000526	–0.000492
	(0.00391)	(0.000713)	(0.000714)	(0.000704)
Presence of senior citizen in the household	0.241***	0.0422***	0.0421***	0.0412***
	(0.0378)	(0.00635)	(0.00638)	(0.00622)
Urban	0.0554	0.0101	0.00920	0.00968
	(0.0364)	(0.00658)	(0.00666)	(0.00676)
Per capita income dummies	Yes	Yes	Yes	Yes
Principle occupation dummies	Yes	Yes	Yes	Yes
Socio-religious group dummies	Yes	Yes	Yes	Yes
Head of the household education dummies	Yes	Yes	Yes	Yes
Geographic region dummies	Yes	Yes	Yes	Yes
N	28,346	28,346	28,346	28,346
R ²				0.039

Source: The authors.

*p < 0.1, **p < 0.05, ***p < 0.01.

Table 10 examines whether women seek permission to visit the local health centre. Table 11 investigates the decision ‘what to do when a child in the family falls sick’. Estimates from Table 10 suggest that when there is an elderly person in the household, the younger women are 5.09% more likely to seek the former’s permission to visit the health centre. We find that when a child in the family falls sick, the mother is 3.27% less likely to get the opportunity to make an independent decision about the course of action to be followed in the presence of senior members in the household. Independent decision-making instances increase with education and age. Urban women are more independent.

Table 10.

Whether Permission Is Required to Visit Local Health Center.

	Logit Coefficients	Logit Marginal Effects	Probit Marginal Effects	OLS Coefficients
Having sons	0.0497	0.00928	0.00934	0.0102
	(0.0403)	(0.00759)	(0.00765)	(0.00739)
Women age	–0.0416***	–0.00771***	–0.00781***	–0.00756***
	(0.00199)	(0.000365)	(0.000372)	(0.000358)
Women education	–0.0260***	–0.00481***	–0.00481***	–0.00476***
	(0.00378)	(0.000699)	(0.000709)	(0.000704)
Presence of senior citizen in the household	0.288***	0.0509***	0.0511***	0.0490***
	(0.0375)	(0.00630)	(0.00638)	(0.00614)
Urban	–0.121***	–0.0226***	–0.0237***	–0.0258***
	(0.0353)	(0.00665)	(0.00675)	(0.00684)
Per capita income dummies	Yes	Yes	Yes	Yes
Principle occupation dummies	Yes	Yes	Yes	Yes
Socio-religious group dummies	Yes	Yes	Yes	Yes
Head of the household education dummies	Yes	Yes	Yes	Yes
Geographic region dummies	Yes	Yes	Yes	Yes
N	28,752	28,752	28,752	28,752
R ²				0.066

Source: The authors.

*p < 0.1, **p < 0.05, ***p < 0.01.

Table 11.

Women’s Participation in Decision-Making on ‘What to Do If Child Falls Sick’.

	Logit Coefficients	Logit Marginal Effects	Probit Marginal Effects	OLS Coefficients
Having sons	–0.0595	–0.0127	–0.0126	–0.0127
	(0.0382)	(0.00819)	(0.00817)	(0.00791)
Women age	0.0248***	0.00524***	0.00528***	0.00507***
	(0.00187)	(0.000395)	(0.000397)	(0.000383)
Women education	0.0111***	0.00235***	0.00235***	0.00224***
	(0.00359)	(0.000758)	(0.000761)	(0.000752)
Presence of senior citizen in the household	–0.158***	–0.0327***	–0.0325***	–0.0317***
	(0.0343)	(0.00695)	(0.00697)	(0.00674)
Urban	0.0184	0.00388	0.00488	0.00461
	(0.0339)	(0.00718)	(0.00721)	(0.00726)
Per capita income dummies	Yes	Yes	Yes	Yes
Principle occupation dummies	Yes	Yes	Yes	Yes
Socio-religious group dummies	Yes	Yes	Yes	Yes
Head of the household education dummies	Yes	Yes	Yes	Yes
Geographic region dummies	Yes	Yes	Yes	Yes
N	28,578	28,578	28,578	28,578
R ²				0.052

Source: The authors.

*p < 0.1, **p < 0.05, ***p < 0.01.

The women in the survey were asked whether their names were on the ownership or rental papers for their homes. The results are reported in Table 12. Results suggest that women with sons are more likely to see their names appended on documents of house ownership, but these findings are not statistically significant. This finding may appear surprising at first glance. An important point to be noted is that the ownership of the household properties could be several years old, often purchased before having kids, and most households cannot be expected to buy new property frequently or to alter the ownership rights often. This question can, therefore, be answered more comprehensively and accurately only if additional data on the length of the ownership of the property were available. Women belonging to the higher income quintiles are more likely to find their names on the property documents in view of the fact that the families included in this quintile purchase properties more often than those in the lower income quintiles. It was thus observed that women belonging to the fourth and highest income quintiles would have greater opportunities by 2.45% and 6.29%, respectively, of having their names on the property documents in comparison to women belonging to the lowest income quintile. Further, older and more educated women would have greater instances of having property ownership than their younger and less-educated counterparts. Not surprisingly, the presence of senior citizens in the household has a negative effect by about 3.38% points on the power and authority exerted by the younger women.

Table 12.

Is Your Name on the Ownership or Rental Papers for Your Home?

	Logit Coefficients	Logit Marginal Effects	Probit Marginal Effects	OLS Coefficients
Having sons	0.00615	0.000792	–0.0000664	–0.000545
	(0.0488)	(0.00628)	(0.00649)	(0.00637)
Women age	0.0466***	0.00601***	0.00623***	0.00631***
	(0.00237)	(0.000301)	(0.000313)	(0.000322)
Women education	0.0182***	0.00234***	0.00231***	0.00229***
	(0.00455)	(0.000585)	(0.000606)	(0.000623)
Presence of senior citizen in the household	–0.278***	–0.0338***	–0.0355***	–0.0368***
	(0.0441)	(0.00504)	(0.00523)	(0.00537)
Urban	–0.0791*	–0.0101*	–0.0101*	–0.0123**
	(0.0428)	(0.00543)	(0.00565)	(0.00566)
Per capita income dummies	Yes	Yes	Yes	Yes
Principle occupation dummies	Yes	Yes	Yes	Yes
Socio-religious group dummies	Yes	Yes	Yes	Yes
Head of the household education dummies	Yes	Yes	Yes	Yes
Geographic region dummies	Yes	Yes	Yes	Yes
N	27,603	27,603	27,603	27,603
R ²				0.057

Source: The authors.

*p < 0.1, **p < 0.05, ***p < 0.01.

Table 13 reports the results for the question whether the woman has any cash in hand to spend on household expenditures. Results suggest that women who have given birth to at least one son are 1.09% more likely to have cash in hand for their personal and household expenditures. The presence of a senior member decreases the probability of making a decision by 3.74% points. In urban areas women are about 4.83% points more likely to have cash in hand to spend. Women in the fifth income quintile have 5.71% higher possibility of having cash in hand to spend on themselves or on their households.

Table 13.

Do You Yourself Have Any Cash in Hand to Spend on Household Expenditures?

	Logit Coefficients	Logit Marginal Effects	Probit Marginal Effects	OLS Coefficients
Having sons	0.0909*	0.0109*	0.0124**	0.0109*
	(0.0473)	(0.00579)	(0.00611)	(0.00631)
Women age	0.0152***	0.00177***	0.00195***	0.00188***
	(0.00246)	(0.000287)	(0.000299)	(0.000304)
Women education	–0.00943**	–0.00110**	–0.000885	–0.00118**
	(0.00463)	(0.000540)	(0.000568)	(0.000548)
Presence of senior citizen in the household	–0.300***	–0.0374***	–0.0384***	–0.0377***
	(0.0406)	(0.00539)	(0.00559)	(0.00568)
Urban	0.431***	0.0483***	0.0503***	0.0493***
	(0.0465)	(0.00498)	(0.00516)	(0.00549)
Per capita income dummies	Yes	Yes	Yes	Yes
Principle occupation dummies	Yes	Yes	Yes	Yes
Socio-religious group dummies	Yes	Yes	Yes	Yes
Head of the household education dummies	Yes	Yes	Yes	Yes
Geographic region dummies	Yes	Yes	Yes	Yes
N	28,828	28,828	28,828	28,828
R ²				0.061

Source: The authors.

*p < 0.1, **p < 0.05, ***p < 0.01.

The results on the question whether the woman’s name is on any bank account are reported in Table 14. Women with sons are 1.10% more likely to have their names on bank accounts than women without sons. Unlike property rights, it is much less cumbersome to open bank accounts with the names of the women or add their names to an existing account since this does not entail any lengthy legal process. Similarly, it was also found that women who belong to the two highest income quintiles are more likely to have bank accounts in their names. In fact, women in the fifth income quintile have 11.4% greater opportunity of having bank accounts in their own names. These possibilities increase with age and education levels of the women. The possibility of having one’s own bank account increases by 0.1% in the presence of a senior member in the household. It increases by 1.87% for urban women as compared to their counterparts in the rural areas.

Table 14.

Is Your Name on Any Bank Account?

	Logit Coefficients	Logit Marginal Effects	Probit Marginal Effects	OLS Coefficients
Having sons	0.0836*	0.0110*	0.0120*	0.0119*
	(0.0496)	(0.00641)	(0.00688)	(0.00670)
Women age	0.0457***	0.00615***	0.00641***	0.00613***
	(0.00249)	(0.000330)	(0.000349)	(0.000326)
Women education	0.127***	0.0171***	0.0180***	0.0192***
	(0.00460)	(0.000609)	(0.000645)	(0.000670)
Presence of senior citizen in the household	0.00745	0.00100	0.000213	0.000226
	(0.0432)	(0.00583)	(0.00615)	(0.00583)
Urban	0.137***	0.0187***	0.0207***	0.0269***
	(0.0431)	(0.00595)	(0.00632)	(0.00626)
Per capita income dummies	Yes	Yes	Yes	Yes
Principle occupation dummies	Yes	Yes	Yes	Yes
Socio-religious group dummies	Yes	Yes	Yes	Yes
Head of the household education dummies	Yes	Yes	Yes	Yes
Geographic region dummies	Yes	Yes	Yes	Yes
N	26,055	26,055	26,055	26,055
R ²				0.205

Source: The authors.

*p < 0.1, **p < 0.05, ***p < 0.01.

VI. Conclusion

In this article we tried to empirically investigate the impact of having son(s) on the bargaining power enjoyed by women in their households in the Indian society. We tested the hypothesis whether having son(s) increases the status of women in the household, given the strong son preference in India. We have used various household decisions in this study as dependent variables. We found that a woman with son(s) is more likely to have a significant say in household decisions such as the menus for daily meals and the purchases on which to incur daily expenditures. In addition, such women are also more likely to go for outings and shop more frequently than those without sons. Similarly, these women would have more cash in hand to spend and would have a greater likelihood of having bank accounts in their own names. We also observe that the presence of an older family member in the household is another important variable which determines a woman’s overall bargaining strength in the family. In most of the cases, the overall bargaining power of women diminishes with the presence of senior members in the household. These results raise important questions about the manner in which household resources are allocated among two generations living in the same household. Further, it would also be an interesting exercise to determine the gender of the senior member in the household in order to ascertain whether any power struggle is likely to occur in the household and if so, whether it is between women of the two different generations or just between the two generations, regardless of gender. We will leave this issue for a future study.

Footnotes

Acknowledgement

We would like to thank Sonalde Desai, Amaresh Dubey, Devendra Bajpai, Aditya Gupte as well as conference participants of IHDS User Conference (New Delhi) for their valuable comments and suggestions. The authors have no financial disclosures to make.

Declaration of Conflicting Interests

Funding

The authors received no financial support for the research, authorship and/or publication of this article.

Notes

ORCID iD

Abhinav Alakshendra

References

Agnihotri

(1996). Juvenile sex ratios in India: A disaggregated analysis. Economic and Political Weekly , 31(52), 3369–3382.

Arnold

, Choe

M. K.

, & Roy

T. K.

(1996). Son preference, the family building process, and child mortality in India (Working Papers, Population Series, no. 85). East-West Center, Honolulu.

Ashraf

(2009). Spousal control and intra-household decision-making: An experimental study in the Philippines. American Economic Review , 99(4), 1245–1277.

Bardhan

(1974). On life and death questions. Economic and Political Weekly , 9(32–34), 1293–1304.

Basu

A. M.

(1989). Is discrimination in food really necessary for explaining sex differentials in childhood mortality? Population Studies , 43(2), 193–210.

Basu

A. M.

(1992). Culture, the status of women, and demographic behavior . Clarendon Press.

Behrman

(1997). Intra-household distribution and the family. In Rosenzweig

, & Stark

(Eds.), Handbook of family and population economics . Elsevier Science B.V.

Bhalotra

(2009). Fatal fluctuations?—Cyclicality in infant mortality in India. Journal of Development Economics , 93(1), 7–19.

Bhat

P. N .M.

, & Xavier

A. J. F.

(2003). Fertility bias and gender decline in northern India. Demography , 40(4), 637–657.

10.

Brown

(2009). Dowry and intra-household bargaining: Evidence from China. Journal of Human Resources , 44(1), 25–46.

11.

Caldwell

J. C.

, Reddy

P. H.

, & Caldwell

(1983). The causes of marriage change in south India. Population Studies , 37(3), 343–361.

12.

Census. (2001). Indian population census 2001 . Retrieved June 27, 2011 from http://www.censusindia.net.

13.

Das

(1984). Sex preference pattern and its stability in India: 1970–80. Demography India , 13(1 and 2), 108–119.

14.

Das

(1987). Sex preference and fertility behavior: A study of recent Indian data. Demography , 24(4), 517–530.

15.

Das Gupta

(1987). Selective discrimination against female children in rural Punjab, India. Population and Development Review , 13(1), 77–100.

16.

Das Gupta

, & Mari Bhat

P. N.

(1997). Fertility decline and increased manifestation of sex bias in India. Population Studies , 51(3), 307–315.

17.

Dreze

, & Murthi

(2001). Fertility, education and development: Evidences from India. Population and Development Review , 27(1), 33–63.

18.

Dreze

, & Sen

(1996). India: Economic development and social opportunity . Oxford University Press.

19.

Dyson

, & Moore

(1983). On kinship structure, female autonomy, and demographic balance. Population and Development Review , 9(1), 35–60.

20.

Folbre

(1984). Market opportunities, genetic endowments, and intra-family resource distribution: Comment. American Economic Review , 74(3), 518–520.

21.

International Institute for Population Sciences (IIPS) and ORC Macro International. (2000). National family health survey–1 (NFHS–1), India, 1998–99 . IIPS.

22.

Iyer

, & Borooah

(2004). Religion and fertility in India: The role of son preference and daughter aversion (Working Papers in Economics). Cambridge.

23.

Kishore

(1993). May god give sons to all: Gender and child mortality in India. American Sociological Review , 58(2), 247–265.

24.

Kulkarni

P. M.

(1999). Gender preference and contraceptive prevalence: Evidences of regional variation. Economic and Political Review , 34(42–43), 3058–3062.

25.

Lahiri

(1984). Demand for sons among Indian couples by rural–urban settlement size. Demography India , 13(1 and 2), 120–132.

26.

, & Wu

(2011). Gender of children, bargaining power, and intrahousehold resource allocation in China. The Journal of Human Resources , 46(2), 295–316.

27.

Malhotra

, Vanneman

, & Kishor

(1995). Fertility, dimensions of patriarchy, and development in India. Population and Development Review , 21(2), 281–305.

28.

Mazumdar

(1994). Amniocentesis and sex selection (Occasional Paper No. 21). Center for Women’s Development Studies.

29.

Miller

B. D.

(1981). The endangered sex: Neglect of female children in rural north India . Cornell University Press.

30.

Operations Research Group (ORG). (1990). Family planning practices in India: third all India survey .

31.

Pande

, & Malhotra

(2006). Son preference and daughter neglect in India: What happens to living girls? ICRW.

32.

Rangamuthia

, Minja

K. C.

, Fred

, & Roy

T. K.

(1997). Son preference and its effect on fertility in India (National Family Health Survey (NFHS) Subject Reports (3)). International Institute for Population Sciences; East-West Center.

33.

Rao

(1993). The rising price of husbands: A hedonic analysis of dowry increases in rural India. Journal of Political Economy , 101(4), 666–677.

34.

Sen

(1990). More than 100 million women are missing. New York Review of Books , 37(20), 61–66.

35.

Sen

(1992). Missing women. The British Medical Journal , 304(6827), 586–587.

36.

Sen

, & Sengupta

(1983). Malnutrition of rural Indian children and the sex bias. Economic and Political Weekly , 18(19–21), 855–864.

37.

Sikand

(1993). Bogey of family planning and Islam. Observer of Business and Politics , 27-2-1993.

38.

Thomas

(1990). Intra-household resource allocation: An inferential approach. Journal of Human Resources , 25(4), 635–664.

39.

Visaria

, Acharya

, & Raj

(2006). Two child norm: Victimizing the vulnerable? Economic and Political Weekly , 41(1), 41–48.

40.

Youssef

N. H.

(1978). The status and fertility patterns of Muslim women. In Beck

, & Keddie

(Eds), Women in the Muslim world (pp. 69–99). Harvard University Press.

41.

Zhang

, & Chan

(1999). Dowry and wife’s welfare: A theoretical and empirical analysis. Journal of Political Economy , 107(4), 786–808.