Reproductive transitions and women’s status in Indian households

Abstract

Using panel data, this study tracks the impact of reproductive transitions on women’s status in the household in India. Here, status refers to the social benefits that women experience by meeting societal expectations related to childbearing. The analysis shows that becoming a mother is associated with increased freedom of movement and access to enabling resources. The adoption of permanent contraception—a common life course event marking the end of childbearing in India—is associated with increased freedom of movement but has no association with changes in access to enabling resources. Household decision-making, another dimension of women’s status examined in the paper, is less dynamic over time and there is limited evidence of its association with reproductive transitions. The findings illustrate the tight linkages between household power dynamics and the life course in the South Asian context, and highlight the centrality of women’s role as mothers in determining their social position.

Introduction

Researchers have long recognized the importance of the life course in South Asia, where age, marital status, reproductive status, and other life course factors are key determinants of an individual’s position both in the household and in larger society. Extended family systems such as India’s, where family members of different generations co-reside, tend to allow for significant mobility within the household hierarchy (Deshmukh-Ranadive 2004). For an Indian woman this often means that her status in the household will increase as she passes key milestones. In India’s patrilocal extended families, a new bride enters her husband’s home as the lowest-status woman in the hierarchy and her behaviour may be policed by the senior women of the household (Das Gupta 1995; Jeffery and Jeffery 1996; Rao 2014). When she becomes a mother, her status in the household is expected to increase, as motherhood is believed to solidify her position in the marital home (Riessman Kohler 2000; Dube 2001; Nahar and Richters 2011; Bhambhani and Inbanathan 2018). The eventual marriage of her sons means that she can become a mother-in-law, a transition which may be accompanied by an expansion of her household decision-making power and autonomy (Das Gupta 1995).

Whereas viewing household power dynamics primarily through the lens of the life course when studying India is quite common for social anthropologists, this perspective is less pervasive in the way that demographers have previously approached the topic. This is because, until recently, almost no demographic panel surveys collecting data on women’s status in the household have followed households over time. Without panel survey data, demographers are limited in their ability to document life course effects. In the past decade the availability of panel data that track women’s status has increased, creating new opportunities for researchers to examine the intersection of the life course with household power dynamics.

This paper uses data from India’s first large-scale panel study, the India Human Development Survey (IHDS), to measure how women’s status in the household changes over the life course, with a focus on how two reproductive transitions—the beginning and ending of childbearing—affect women’s status. A significant body of research in demography has documented a strong link between women’s status in the household and fertility outcomes such as total fertility, contraceptive behaviour, child survival, and the use of maternal health services (Abadian 1996; Hindin 2000; Bloom et al. 2001; Dreze and Murthi 2001; Ghuman 2003; Shroff et al. 2011; Cunningham et al. 2015). Most of this literature has focused on how women’s status can predict fertility behaviour, but researchers have acknowledged that fertility behaviours may also affect a woman’s status in the household. Until panel data became available, however, it was impossible to track the impact of reproductive transitions on status properly. The current study extends the previous literature and is unique in using a fixed effects panel research design, using multidimensional measures of women’s status, and examining the impact of different reproductive transitions. This paper documents how the transition into motherhood and the adoption of permanent contraception (which signals the end of childbearing) impact women’s status in Indian households, as measured by both their agency and their access to enabling resources.

Women’s status in the household

Women’s status in the household has received less attention in the literature in recent years, as researchers have shifted their focus to studying the related but distinct topic of women’s empowerment. The same survey questions are often used to measure both women’s status and women’s empowerment and, as a result, the terms are sometimes used interchangeably in the literature. However, as panel data become increasingly available, there is more urgency to articulate the distinction between status and empowerment clearly. Researchers should use caution when interpreting increases in autonomy over the life course, as these patterns are quite distinct from the process of empowerment. The term ‘status in the household’ provides a more accurate way to think about life course effects.

Women’s empowerment is the process of women enhancing their ability to make strategic life choices in their own interest (Kabeer 1999). It may be conceptualized as either an individual or a societal characteristic, but most researchers agree that it involves individuals overcoming barriers (both internal and external) to the enactment of their agency (Gram et al. 2019). There are three dimensions to empowerment discussed in the literature. The first dimension is resources, both material and social; these resources serve as a precondition for the enactment of agency, the second dimension (Kabeer 1999). Agency is the ability of individuals to define their goals and act on them (Ibrahim and Alkire 2007; Allendorf 2012; Gammage et al. 2016). Achievements, the third dimension, are the outcome of individuals having the resources to enact their agency in order to improve their well-being or reduce inequalities (Kabeer 1999).

Women’s status in the household is a related concept. The distinction between the two derives from the means through which women attain these benefits. Status is the value that a group or individual has in a society when judged based on a set of cultural beliefs about what is valuable (Weber 1968; Ridgeway 2014). High-status individuals enjoy privileges such as greater respect or prestige. Status hierarchies also structure and legitimate the way that power is distributed in a society (Ridgeway 2014). Like social status more broadly, a woman’s status in the household is determined by her conformity to external social and cultural values (Kabeer 1999). Women gain status by meeting external social expectations. This is quite different from empowerment, which involves actualization of women’s internal goals. If women are viewed as less than men within a cultural value system, then it is possible for women’s status in the household to be antithetical to women’s empowerment (Kabeer 1999). For example, a woman may reap status benefits in her household for bearing sons or bringing a large dowry to the marital home. These benefits may even include greater opportunities for her to enjoy autonomous decision-making. However, the means through which she gained these benefits are disempowering because they are built on a cultural framework wherein women are undervalued or viewed as a burden to their family (Kabeer 1999).

Women’s status is also multidimensional due to different social and institutional constraints on their access to prestige and power (Mason 1986; Malhotra and Schuler 2005; Agarwala and Lynch 2006; Ibrahim and Alkire 2007). For example, in South Asia, a woman may have significant decision-making power in her household over topics such as parenting and household expenses, but face severe restrictions on her freedom of movement due to social norms on the seclusion of women (Desai and Temsah 2014). For this reason, it is often useful to examine each dimension separately or in clusters of related domains (Pratley and Sandberg 2018).

Studying women’s status in the household is important because it illuminates the social structures of power and the cultural values that are present in a specific social context. A significant body of research has outlined the importance of context in determining how empowerment should be defined and measured (Malhotra and Schuler 2005; Agarwala and Lynch 2006; Rocca et al. 2009; Richardson 2018a; Donald et al. 2020). Understanding women’s status in the household is an important prerequisite to studying empowerment because it reveals the constraints that women work within in order to maximize their life choices (Kandiyoti 1988).

Motherhood and women’s status

A large literature on kinship and the family in Asia has stressed the importance of motherhood as a social role which can confer status within the household and larger society. South Asian women often report that bearing children is part of their social or religious duty (Säävälä 2001; Mehta and Kapadia 2008; Nahar and Richters 2011; Bhambhani and Inbanathan 2018). The status benefits associated with motherhood function at least partially through the woman’s relationships: first, by solidifying the bond with her husband and marital family, and second, through the creation of the maternal bond with her children, who will support her in old age (Mehta and Kapadia 2008; Nahar and Richters 2011; Rao 2015). Bearing sons also ensures that a woman will one day become a mother-in-law to her son’s wife, a position which often comes with significant power in the household. Women who are unable to fulfil the social expectation of motherhood due to infertility or who choose not to have children often report facing social stigma, diminished autonomy, desertion, and violence (Riessman Kohler 2000; Mehta and Kapadia 2008; Lee-Rife 2010; Nahar and Richters 2011; Rao 2014; Bhambhani and Inbanathan 2018).

Some previous studies using demographic survey data have attempted to quantify motherhood effects on women’s status in the household. Using cross-sectional data, Jejeebhoy and Sathar (2001) demonstrated that having a son was correlated with higher autonomy in the Punjab district of India but having a daughter was not. Kishore and Spears (2014) also used cross-sectional data to document that for urban Indian households, a male first birth increased the likelihood that the household used clean cooking fuel, an effect that they link with increased status for mothers of sons. Using retrospective data from one state in India, MacQuarrie (2009) found evidence that women in the childbearing stage of life enjoyed greater agency and less fear of domestic violence than those who had yet to make the transition into motherhood. Only a few studies, to my knowledge, have examined the impact of motherhood on women’s status using panel data and none were based in the Indian context. Using panel data, Li and Wu (2011) found that having a firstborn son improved the decision-making power and nutritional intake of women in China. Using panel data from Egypt, Samari (2017) found that Egyptian women experienced status benefits after the birth of each child.

Permanent contraception and women’s status

Like entry into motherhood, the end of childbearing is an important reproductive transition, especially in the South Asian context where young women often make an abrupt exit from their childbearing stage of life through the adoption of permanent contraception. Sterilization through tubal ligation is the dominant family planning method in India and is increasingly used by younger women with ever lower completed fertility (IIPS and ICF 2017). According to the latest National Family Health Survey data from India, 36 per cent of married woman in India are sterilized (IIPS and ICF 2017). Sterilization is now part of the reproductive life course of many women in India. Because marriage often happens early and childbearing begins shortly after marriage, most women will have completed their desired fertility by their mid- to late twenties. As a result, the adoption of permanent contraception also happens at a young age. In 2015–16, the median age at sterilization was only 25.7 (IIPS and ICF 2017).

Permanent contraception has a complicated history in India. From 1975 to 1977, the Indian government ran a forced sterilization campaign, which primarily targeted men. During that time, 8.3 million sterilizations occurred (Haub and Sharma 2006). The public backlash after the programme was intense. Since then, the government has attempted to rebrand family planning initiatives and in 1996 instituted a ‘no quota’ system for permanent contraception; however, the Indian government continues to promote sterilization as a way for families to achieve their desired family size, including through financial incentives (Donaldson 2002). As a result, 82 per cent of these surgeries take place in a public sector facility (IIPS and ICF 2017).

Today most sterilizations in India are carried out on women rather than men. Only 0.3 per cent of married women report that their husband’s sterilization is their birth control method (IIPS and ICF 2017). Reversible forms of contraception are unpopular and sometimes stigmatized in India. Only 4 per cent of married Indian women aged 15–49 report using the pill and 12 per cent report using condoms (IIPS and ICF 2017). Although relatively widespread in its use, sterilization is more common among poorer households (McNay et al. 2003; De Oliveira et al. 2014). Some women have been sterilized in large government camps set up in impoverished rural areas, a practice which has been criticized by many activists and was recently ordered to stop by the Supreme Court of India (The Guardian 2016). This occurred after several public scandals involving deaths following botched surgeries at these camps (Reuters 2014).

Despite these issues, rural women have been found to actively seek out sterilization to limit their fertility and these women often report favourable impressions of the practice (Säävälä 1999). Data suggest that women are often the primary or joint decision maker for contraceptive decisions in their family, although husbands or in-laws are the primary decision maker in about one-third of households (Char et al. 2009; Makade et al. 2012). However, researchers have expressed concern about coercion, given data which show that Indian women often have limited information on the risks before consenting to the procedure (Jadhav and Vala-Haynes 2018). Sterilization regret is usually reported at around 5 per cent in India (Ramanathan and Mishra 2000; Singh et al. 2012).

With permanent contraception increasingly becoming a common life course event for women in India, an emerging literature has begun to examine its effects on women’s lives. The current evidence is mixed regarding the effect of permanent contraception, if any, on women’s status in the household. Rao (1997) found that sterilized women in South India were more likely to suffer intimate partner violence, perhaps due to fears of infidelity. This would suggest that permanent contraception has a negative effect on women’s status in the household.

Other evidence points towards a positive relationship between the adoption of permanent contraception and women’s status in the household. Early studies suggested a possible link between sterilization and improved spousal communication and freedom of movement (Vlassoff 1991). Säävälä (2001) found evidence from their ethnographic study in rural Andhra Pradesh that sterilized women were able to reap the benefits of ‘post-procreative’ identity well before they would have normally terminated their childbearing years (at the time of menopause). Post-procreative women were seen as devoid of the ‘female liability’ stemming from women’s role in reproduction and fears over paternity if women’s sexuality was not controlled (Säävälä 2001, p. 202). Pallikadavath et al. (2015) found that young sterilized mothers enjoyed more autonomy than non-sterilized mothers of similar socio-economic backgrounds. However, they also found that women who used modern, temporary forms of contraception, such as the pill or intrauterine device (IUD), experienced even higher autonomy levels than sterilized women. MacQuarrie (2009), using retrospective data from Madhya Pradesh, found evidence of increased decision-making and greater freedom of movement at the end of the reproductive life course (marked by the adoption of permanent contraception).

Permanent contraception could improve marital relationships as well. At the most basic level, it allows couples to enjoy sexual activity without fear of pregnancy. Säävälä (2001) found that many villagers at their ethnographic field site viewed sterilization as a sacrifice that women made so that their husbands would not be required to bear the burden of fertility control through vasectomy, condom use, or abstinence. Interviewed women felt that making this sacrifice could improve their relationship with their husband.

As Indian women advance through the life course, their status in the household is expected to increase. Säävälä (2001) theorized that this happens like a stepladder, with women’s social prestige increasing first as they transition from girlhood to womanhood, then from unmarried to married, then as they transition into mothers, and finally to mother-in-law, the highest position possible. A post-procreative transition associated with the adoption of permanent contraception may also exist between motherhood and the mother-in-law position. The analysis that follows tests the relationship between women’s status and two reproductive transitions: motherhood and the adoption of permanent contraception. Examining each dimension of status separately also permits a test of which dimensions of status are impacted by reproductive transitions, if any.

Data and methods

This paper uses data from the IHDS, which is jointly organized by researchers from the University of Maryland and the National Council of Applied Economic Research (NCAER) in New Delhi. The IHDS is the first large-scale, nationally representative panel survey of households in India covering these topics (Desai et al. 2019). The survey is conducted in all states and union territories of India except for the territories of Lakshadweep and the Andaman and Nicobar Islands. Two waves of the panel have been completed: IHDS I, conducted in 2004–05, and IHDS II, conducted in 2011–12. The sample was drawn using stratified random sampling, which sampled 1,503 villages and 971 urban blocks (Desai et al. 2019). IHDS II reinterviewed 83 per cent of the households surveyed in IHDS I, including households that split apart between waves if the resultant households were located within the same village or town. This re-contact rate not only meets desired re-contact thresholds but is relatively high, considering the seven-year gap between survey waves (Kristman et al. 2004).

The ‘Eligible Women’ module, used in this analysis, is a separate questionnaire conducted with one woman per household who was aged 15–49 in IHDS I and had ever been married. Women were surveyed in the local language, usually by a female interviewer. The IHDS team attempted to follow up with all respondents, regardless of their current age. In total, they were able to reinterview 76 per cent of the initial IHDS I respondents. The rest were lost to follow-up, died, moved far away, or refused to participate in the second wave. Women who left the sample were slightly more likely to come from privileged households by education, caste, and economic position. The sample used in this analysis is restricted to only those women who are married at both survey waves, by excluding about 8 per cent of women who are separated, divorced, or widowed during at least one of the waves. Finally, casewise deletion is used to adjust for missingness on any independent or dependent variable in the analysis (17.8 per cent of the sample). This leaves a total sample of 19,263 women for the analysis.

Measures of women’s status in the household

The analysis that follows focuses on measures of agency and access to enabling resources, two dimensions of women’s status in the household that are captured well by the IHDS data. Rather than creating an overarching index of women’s status which combines the individual measures, I examine each measure independently. Creating a single-scale measure of women’s status has several drawbacks including measurement error issues and the fact that it would hide the multidimensionality of the concept (Pratley and Sandberg 2018). Furthermore, Cronbach’s alpha tests showed weak internal consistency, indicating that the variables captured unique dimensions of women’s status. A scale would be inappropriate in this context. By examining each variable separately, I shed light on which dimensions of a woman’s status in the household are impacted by reproductive transitions and which are not.

Summary statistics on the dependent variables used in the analysis can be found in Table 1. There is significant change between the survey waves in women’s responses to the status questions: for each dependent variable, between 13.9 and 82.1 per cent of respondents changed their response. Previous research in Egypt has found evidence that measures of women’s agency have measurement invariance, proving their reliability (Cheong et al. 2017). The time trends observed in the IHDS data may therefore be interpreted as evidence that women’s status in the household changes over time rather than pointing towards measurement error. Women’s status in India has been shown to have strong durability, meaning that improvements in status are retained over time rather than being temporary (Akter and Chindarkar 2019).

Table 1.

Descriptive statistics on variables used in analysis of women in India, IHDS I and II

	Mean in IHDS I	Mean in IHDS II	Percentage changing status between IHDS I and II			Total number of women who change status between IHDS I and II
	(Percentage unless otherwise indicated)		Acquired status (or increased)	Left status (or decreased)	No change		Total women
Dependent variables
Women’s agency
Can go alone to health centre	68.6	74.3	20.4	14.7	64.9	6,753	19,263
Can go alone to friend’s or family member’s house	70.4	80.7	22.7	12.4	64.9	6,748	19,263
Has most say in cooking decisions	74.0	73.9	17.1	17.1	65.8	6,578	19,263
Participates in large purchase decisions	71.0	80.6	22.6	13.0	64.4	6,856	19,263
Has most say in large purchase decisions	7.8	8.3	7.2	6.7	86.1	2,679	19,263
Has access to cash for household expenses	82.0	93.5	16.0	4.5	79.5	3,952	19,263
Access to enabling resources
Name on bank account1	47.3	62.6	29.3	13.9	56.8	2,498	5,786
Name on housing papers2	15.0	17.5	13.8	11.2	75.0	4,489	17,956
Spousal communication score (0–6 scale)3	3.5	3.8	46.5	35.6	17.9	15,806	19,263
Independent variables
Reproductive status
Mother	94.2	98.7	4.6	0.1	95.3	902	19,263
Uses permanent contraception4	40.7	66.3	25.5	0.0	74.5	4,919	19,263
Controls
Does paid work5	21.7	27.2	13.8	8.2	78.0	4,226	19,263
Does unpaid work6	40.1	45.0	16.2	11.3	72.5	5,303	19,263
Absent husband	3.2	4.9	3.1	1.4	95.5	860	19,263
Senior married woman in household7	63.6	74.8	14.2	3.1	82.7	3,314	19,263
Other married man in household8	26.9	25.1	11.3	13.0	75.7	4,673	19,263
Per capita household consumption (rupees)	18,548	26,309	70.9	29.1	0.0	19,263	19,263

¹Excludes households which reported having no bank account in at least one wave (70.0 per cent of households).

²Excludes households which reported that they had no housing papers in at least one wave (6.8 per cent of the sample).

³Captures the intensity of discussion with husband on three topics: daily events at work, household finances, and community happenings.

⁴Indicates that the respondent has adopted permanent contraception through either tubal ligation or a hysterectomy.

⁵Indicates that the respondent worked at least 250 hours of wage or salary work in the past year.

⁶Indicates that the respondent worked at least 250 hours in unpaid animal husbandry, family business, or agricultural work in the past year.

⁷An indicator which specifies that the respondent is the only or oldest married woman in the household.

⁸An indicator that there is another married man present in the household besides the respondent’s husband.

Source: India Human Development Survey I (2004–05) and II (2011–12).

Measures of women’s agency

The IHDS captures several dimensions of women’s agency. One of the most important dimensions of women’s agency in the South Asian context is freedom of movement. Derne (1995) found in his ethnographic study in Banaras, India, that restrictions placed on women’s physical mobility create a system where men’s authority is repeatedly reaffirmed. Restrictions on a woman’s movement also impact her ability to build and maintain social and economic networks, as well as her ability to participate in the economy and civic life (Klugman et al. 2014).

Two dichotomous variables in the IHDS capture respondents’ freedom of movement. Women are asked if they can go to the health centre alone and if they can go to the home of a friend or family member alone. A third question asks if she is able to go to the grocery store alone, but this variable is excluded because of its high non-response rate. Regressions run on this variable showed similar effects. The percentage of women who could go to the health centre alone increased from 68.6 to 74.3 per cent between the survey waves (Table 1). Similarly, the percentage of women who could go to the house of a friend or family member alone also increased from 70.4 to 80.7 per cent.

Decision-making power is one of the most common measures of women’s agency used in the literature. Researchers have noted that a distinction should be made between different domains of decision-making. For example, in South Asia, gender norms may allocate certain domains of decision-making, such as purchasing food or making decisions regarding children, to women, whereas other domains are strictly off limits to them (Rao 2014). Furthermore, in many households, men may make large policy decisions but delegate the implementation or management of decisions to women (Kabeer 1999; Richardson 2018b). Important distinctions can also be made regarding levels of involvement in decision-making, with researchers often making different decisions on whether to count joint decision-making as part of women’s agency (Gram et al. 2019). In order to address these issues relating to both to the domain and level of decision-making power, the analysis that follows examines each variable separately.

Four dichotomous variables capture women’s decision-making power in the IHDS. Respondents are asked who participates in decision-making and who is the main or primary decision maker in the household for different types of decisions. The first domain of decision-making relates to meal planning in the household. Since almost all respondents report that they participate in cooking decision-making, I focus on whether the respondent reports that she has the ‘most say’ in cooking decision-making. As shown in the descriptive statistics, about 74 per cent of all women reported that they were the main cooking decision maker in each survey wave, but there is significant movement into and out of this status.

The remaining decision-making variables are related to control over household finances. In IHDS I, 71.0 per cent of respondents reported that they participated in decision-making in their household on large purchases, such as buying a TV. By IHDS II, 80.6 per cent of respondents were now participating in large purchase decision-making. Few respondents in either survey wave reported that they were the primary financial decision maker for large purchases. However, the proportion of women reporting that they were the main large purchase decision maker increased from 7.8 to 8.3 per cent between IHDS I and II. A final variable captures whether the respondent has access to cash for household expenditure. In IHDS I, only 82.0 per cent of respondents reported access to household cash, whereas in IHDS II, 93.5 per cent of respondents did so.

Measures of access to enabling resources

The second category of outcomes in this analysis focuses on women’s access to enabling resources. Enabling resources, while not necessarily valuable when they are not paired with agency, serve as the precondition for individuals to enact agency (Kabeer 1999). The analysis focuses on measures of women’s access to economic resources and access to the intangible resource of a good communicative relationship with their spouse.

The IHDS contains measures of whether the respondent’s name is on their housing papers (title, lease, deed, etc.) and whether her name is on a bank account. When women gain property rights to their home, they have been shown to enjoy increased participation in household decision-making, increased community involvement, higher self-esteem, and more respect from their husbands (Datta 2006; Allendorf 2007; Baruah 2007). Women’s increased access to banking has been linked to increases in employment and expanded freedom of movement (Bruhn and Love 2011; Field et al. 2016).

Emotional bonds created by healthy, loving family relationships can be considered a type of intangible resource that improves a woman’s status in the household. Women may be more willing to assert themselves and ensure their families listen to them if they have strong emotional bonds with family members. Mumtaz and Salway (2009) observed this phenomenon in their fieldwork in rural Pakistan. There, they found that women made a distinction between a mazboot aurat (strong woman), who is deeply and securely embedded in her marital home, and a kamzoor aurat (weak woman), whose ties are weak (p. 1351). Allendorf (2012) found that the marital bond is especially important in determining a woman’s decision-making power in the household. Husbands with a close bond with their wives may be more likely to heed their advice and delegate them control over household decisions. Thus, a high-quality relationship with their husband can be considered a resource which women may use to enact greater agency.

Descriptive statistics on respondents’ access to enabling resources can be found in Table 1. Analyses using the banking and housing papers variables exclude those households which do not have the resource in at least one of the survey waves. This helps to adjust for the expansion of access to these resources in India over the time period. In total, 70 per cent of households did not have a bank account and 6.8 per cent did not have housing papers in at least one of the survey waves. Between IHDS I and II, the percentage of respondents whose name was on a bank account, for those households with bank accounts, increased from 47.3 to 62.6 per cent. For those women who lived in households with proof of legal tenancy, the percentage with their name on those documents increased from 15.0 to 17.5 per cent between IHDS I and II.

The IHDS also collects data on communication between the respondent and her husband, capturing an important dimension of a quality marital relationship, which can be considered an enabling resource for women. Women report whether they sometimes, never, or often speak with their husband about: (1) things that happen at work or on the farm; (2) what to spend money on; and (3) things happening in the community, such as elections or politics. From these three questions, a spousal communication scale could be created, with values ranging from ‘0’ (low communication) to ‘6’ (high communication). The Cronbach’s alpha score for the scale was 0.74, reflecting high internal consistency. Table 1 shows that the mean spousal communication score increased from 3.5 to 3.8 between IHDS I and II, with 46.5 per cent of women showing a higher spousal communication score in the second wave relative to the first wave.

Independent variables

The main independent variables of interest are the reproductive transitions into and out of the childbearing stage of a woman’s life. Motherhood is defined as having at least one living child. Most women in the sample had already become mothers by the time of the survey. As shown in Table 1, 4.5 per cent of respondents became mothers between IHDS I and II. Motherhood is nearly universal in India, with only 1.3 per cent of women reporting that they were childless in IHDS II. The transition out of the childbearing stage of life through the adoption of permanent contraception is defined by the respondent reporting that tubal ligation or hysterectomy is their primary or secondary contraceptive method. Between IHDS I and II, 25.6 per cent of respondents adopted permanent contraception. By IHDS II, a significant majority, 66.3 per cent, had adopted permanent contraception.

Control variables

In addition to these independent variables capturing the reproductive transitions of women in the sample, the models include controls. The fixed effects method used in this analysis examines variation only within individuals, so fixed characteristics such as region and caste are controlled for in the analysis. However, controls are included for labour force participation, household composition, and household consumption because these are three dimensions of women’s lives which may have changed between IHDS I and II.

Resource theory suggests that joining the labour force will increase a woman’s agency, since power dynamics are shaped by contribution to household resources (Agarwal 1997; Haddad et al. 1997; Malhotra and Mather 1997; Luke et al. 2014). Unpaid labour, such as working on a family farm, is not expected to have the same effects on women’s agency, since it does not contribute to household financial resources (Agarwal 1997; Rao 2011). The models in this paper control for changes in the respondent’s work participation, which is defined as working at least part-time in the past year, or for a minimum of around five hours per week. Paid work is defined as any work for wages or salary. Unpaid work includes animal husbandry, family farming, or contributions to a family business for which the worker is not compensated financially.

Household-level factors may also impact a woman’s status in the household. Numerous studies have identified a link between household composition and women’s status in the household (Jejeebhoy and Sathar 2001; Allendorf 2013, 2017; Debnath 2015). Although the IHDS lacks a perfect measure of the presence of in-laws in the household, a series of dichotomous variables can capture changes in the overall household composition. The first variable created indicates if the respondent is the senior (oldest) married woman in the household. If no older married woman is listed in the household roster, or the respondent is the only married woman in the household, then she is classified as the senior woman. Second, if the husband of the respondent is listed as non-resident, then the respondent is classified as having an absent husband. Husband absence is due primarily to labour migration. A third variable captures whether the household includes any married men other than the respondent’s spouse. In most cases, this variable captures the presence of the respondent’s father-in-law, but in some cases, there may be some other married man in the household, such as a brother-in-law.

Finally, the models include a control for changes in household per capita consumption (logged). The IHDS collects data on household consumption from a series of questions about spending on frequently purchased items.

Fixed effects method

The analysis uses individual fixed effects logistic and linear regression models to capture how changes in a woman’s life are associated with changes in her status. The advantage of the fixed effects approach is that each individual woman is treated as her own control. In the analysis that follows, the ‘motherhood effect’ denotes the effect of a change in motherhood status (transition into first birth) on changes in women’s responses to the status measures. Similarly, the ‘permanent contraception effect’ is the effect associated with the adoption of permanent contraception on a woman’s status. In this sense, we can think of entry into motherhood and the cessation of childbearing as life course events. What is being measured is the relationship between those events and changes in a woman’s reported freedom of movement, decision-making power, and access to enabling resources.

The fixed effects approach focuses only on the within-individual variation, which is the change in outcome variables within each woman over time. Fixed effect models do not include between-individual variation. Between-individual variation, such as the variation between individuals of different castes or religions, may be confounded with unobserved characteristics of the respondents, thus introducing bias into the measures (Allison 2009). Fixed effects regressions are considered more robust because they eliminate this form of bias. In addition, the focus on within-individual variation allows this analysis to avoid issues that have arisen in comparative studies, where the different cognitive or semantic meaning of questions across different settings has led to measurement issues (Ghuman et al. 2006). This method is especially attractive for examining changes from a life course perspective because it focuses on the changes related to events within an individual’s life rather than comparing individuals in a cross-sectional analysis. Being the largest longitudinal data set in India that covers these topics, the IHDS provides the first opportunity to examine within-woman changes in women’s status systematically over time in India.

Fixed effects logistic regressions are used for most of the outcome variables because they are dichotomous, whereas a fixed effects linear regression is used for the spousal communication scale. With dichotomous outcome variables, the cases contributing to the coefficients in a fixed effect logistic regression come only from those individuals who change on the outcome variable. This means that the cases include only those women who are acquiring the status or leaving the status, even though the whole sample is used in the analysis. The number of cases for each regression is different and corresponds to the total number of women who change between IHDS I and II for that variable, as listed in Table 1. The weights provided by the IHDS team are used in the analysis to help adjust for representativeness of the sample. Robust standard errors are also used.

Results

The first set of regression results are presented in Table 2, where Models 1 and 2 examine the association between reproductive transitions and the measures of freedom of movement using individual fixed effects logistic regressions. Controlling for household composition, work participation, and household consumption, motherhood is associated with a 78 per cent higher odds of being able to go to the health centre alone and a 51 per cent higher odds of being able to go to the home of a friend or family member alone. The adoption of permanent contraception is associated with 40 per cent higher odds of the respondent reporting that she can travel alone to the health centre and a 34 per cent higher odds that she can visit the home of a friend or family member alone. These results are statistically significant at the 0.01 level or higher, providing evidence that these reproductive transitions are associated with increases in freedom of movement.

Table 2.

Odds ratios from six individual fixed effects logistic regressions on measures of agency among women in India

	Freedom of movement		Decision-making
	(1)	(2)	(3)	(4)	(5)	(6)
Variables	Can go alone to health centre	Can go alone to friend’s or family member’s house	Has most say in cooking decisions	Participates in large purchase decisions	Has most say in large purchase decisions	Has access to cash for household expenses
Reproductive status
Mother	1.78*** (1.30–2.45)	1.51** (1.10–2.05)	1.96*** (1.40–2.75)	1.21 (0.86–1.70)	1.64 (0.77–3.49)	0.98 (0.60–1.60)
Uses permanent contraception	1.40*** (1.18–1.67)	1.34*** (1.13–1.60)	1.26* (1.05–1.50)	1.17+ (1.00–1.38)	1.00 (0.77–1.31)	1.15 (0.89–1.49)
Controls
Does paid work	1.02 (0.85–1.23)	1.17+ (1.00–1.37)	0.95 (0.80–1.13)	1.09 (0.93–1.28)	1.07 (0.85–1.34)	1.31* (1.03–1.69)
Does unpaid work	1.15+ (0.97–1.35)	1.12 (0.96–1.30)	1.19* (1.01–1.41)	0.95 (0.82–1.10)	1.01 (0.76–1.34)	1.06 (0.87–1.30)
Absent husband	2.33*** (1.52–3.57)	1.27 (0.89–1.82)	1.89*** (1.31–2.75)	1.28 (0.86–1.91)	3.91*** (2.16–7.08)	1.83* (1.07–3.15)
Senior married woman in household	1.35* (1.07–1.69)	1.12 (0.91–1.38)	2.51*** (1.94–3.26)	1.52*** (1.23–1.88)	1.82** (1.20–2.76)	1.66** (1.21–2.28)
Other married man in household	0.70*** (0.58–0.83)	0.67*** (0.56–0.79)	0.51*** (0.41–0.63)	0.72*** (0.60–0.87)	0.81 (0.60–1.09)	0.79* (0.62–1.00)
Per capita household consumption (ln)	0.74*** (0.65–0.83)	0.78*** (0.70–0.88)	0.94 (0.83–1.07)	1.09 (0.97–1.23)	0.97 (0.82–1.16)	1.12 (0.94–1.34)
IHDS II	1.48*** (1.33–1.63)	1.99*** (1.80–2.19)	0.72*** (0.64–0.80)	1.47*** (1.33–1.62)	1.02 (0.88–1.18)	2.99*** (2.58–3.46)
Cases1	6,753	6,748	6,578	6,856	2,679	3,952
N women in sample	19,263	19,263	19,263	19,263	19,263	19,263
Percentage of sample in cases2	35.1	35.0	34.1	35.6	13.9	20.5

^***p<0.001

^**p<0.01

^*p<0.05

⁺p<0.1

¹The estimation procedure does not incorporate information about women who did not change their response to the question. ‘Cases’ refers to the number of women who changed their response to that specific outcome variable between IHDS I and II.

²Displays the percentage of the total women in the sample who experience a change on the outcome variable and corresponds to the number of women who contribute to the fixed effects logistic regression as ‘cases’.

Notes: Robust 95 per cent confidence intervals in parentheses; Constant IHDS I weights applied to data. Fixed effects logistic regression modelling does not produce coefficient constants. See Table 1 notes for definitions of variables used in analysis.

Source: As for Table 1.

The IHDS II variable in all models presented in Table 2 displays the remaining changes over time not captured by the independent variables. Even controlling for all independent variables in the analysis, a statistically significant unexplained increase in the odds of a respondent reporting most statuses remains. The IHDS II indicator variable may be capturing the effect of ageing, since all respondents age seven years between the two survey waves. It could also be capturing some period effects, such as changes in social norms over time.

The ‘cases’ listed for all models in Table 2 reflect the number of women who experience a change in the dichotomous outcome variable. These cases, which constitute 35.1 and 35.0 per cent of the sample for Models 1 and 2, respectively, are the numbers of women who are contributing to the regression coefficient results because their response on the outcome variable changes between IHDS I and II. As discussed in the ‘Fixed effects method’ subsection, the fixed effects logistic regression framework examines only changes within individuals, which reduces the cases to only those who experience the change.

Models 3, 4, 5, and 6 in Table 2 depict the fixed effects regression results for the household decision-making measures. The odds that mothers report being the main cooking decision maker in their household are 96 per cent higher than those who are not mothers, controlling for the other variables included in the model. There is no statistically significant relationship between motherhood and whether a woman participates in large purchase decisions, is the main large purchase decision maker, or reports access to cash.

Model 3 in Table 2 shows that the odds of reporting being the main cooking decision maker are 26 per cent higher among women who adopted permanent contraception. This result is statistically significant at the 0.05 level. Adopting permanent contraception is also associated with a 17 per cent higher odds of a woman reporting that she participates in large purchase decision-making in the household, but this result is only marginally significant (Model 4). There is no statistically significant relationship between the adoption of permanent contraception and the odds of a woman reporting that she is the main decision maker on large purchases or has access to household cash.

Overall, the results in Models 3, 4, 5, and 6 of Table 2 provide limited evidence that reproductive transitions are associated with increases in household decision-making for women. While these results are robust for non-financial decision-making (cooking), there is little evidence that reproductive transitions impact financial decision-making. In addition, there is less change on these outcome variables between survey waves, suggesting that household financial decision-making is less dynamic over time than other status measures.

Table 3 depicts the results from fixed effects regressions examining the impact of reproductive status transitions on women’s access to enabling resources. Models 7 and 8 (Table 3) are from individual fixed effects regressions showing access to property rights and financial institutions. These regressions are conducted only on the subsets of households that have a bank account and housing papers in both survey waves, which excludes 70.0 and 6.8 per cent of the sample, respectively. Mothers have 127 per cent higher odds of reporting that their name is on their family’s housing papers than women who are not mothers (Model 7). The odds of mothers having their name on a bank account are 76 per cent higher; however, this result is only marginally significant (Model 8). The spousal communication scale results shown in Model 9 are from a linear fixed effects regression. Motherhood is associated with an increase of 0.25 on the 0–6 communication scale, controlling for other variables in the model. The adoption of permanent contraception is not associated with any statistically significant associations with women’s access to enabling resources.

Table 3.

Odds ratios and linear coefficients from three individual fixed effects regressions on access to enabling resources among women in India

	Enabling resources
	(7)	(8)	(9)
Variables	Name on housing papers (odds ratio)	Name on bank account (odds ratio)	Spousal communication score (0–6) (coefficient)
Reproductive status
Mother	2.27** (1.38–3.73)	1.76+ (0.94–3.28)	0.25+ (−0.00-0.51)
Uses permanent contraception	1.05 (0.84–1.30)	1.03 (0.75–1.43)	−0.10 (−0.21–0.02)
Controls
Does paid work	1.16 (0.95–1.41)	1.62** (1.15–2.26)	0.12* (0.02–0.23)
Does unpaid work	0.95 (0.80–1.13)	1.20 (0.91–1.57)	0.23*** (0.14–0.32)
Absent husband	1.12 (0.70–1.81)	0.94 (0.34–2.60)	−0.48*** (−0.75 to −0.20)
Senior married woman in household	1.02 (0.74–1.40)	1.01 (0.71–1.42)	−0.08 (−0.23–0.06)
Other married man in household	0.96 (0.78–1.18)	0.76* (0.59–0.98)	−0.15** (−0.26 to −0.05)
Per capita household consumption (ln)	0.94 (0.83–1.07)	1.19+ (0.98–1.44)	0.20*** (0.13–0.28)
IHDS II	1.24*** (1.11–1.40)	1.49*** (1.25–1.77)	0.19*** (0.12–0.25)
Constant	–	–	1.38*** (0.61–2.15)
Cases1	4,489	2,498	–
N women in sample	17,9562	5,7863	19,263
Percentage of sample in cases4	25.0	43.2	–

^***p<0.001

^**p<0.01

^*p<0.05

⁺p<0.1

¹The estimation procedure of the fixed effect logistic regression does not incorporate information about women who did not change their response to the question. ‘Cases’ refers to the number of women who change their response to that specific outcome variable between IHDS I and II.

²This regression excludes 6.8 per cent of households which report that they have no housing papers in at least one survey wave.

³This regression excludes 70.0 per cent of households which report not having a bank account in at least one survey wave.

⁴Displays the percentage of the total women in the sample who experience a change on the outcome variable and corresponds to the number of women who contribute to the fixed effects logistic regression as ‘cases’.

Notes: Robust 95 per cent confidence intervals in parentheses; Constant IHDS I weights applied to data. Fixed effects logistic regression modelling does not produce coefficient constants. See Table 1 notes for definitions of variables used in analysis.

Source: As for Table 1

There are also some interesting results from the control variables in the models presented in Tables 2 and 3. Participation in the paid and unpaid labour force are both associated with some positive effects on women’s status in the household but overall the effects are small and often only marginally significant. Most of the status measures impacted by work force participation are directly linked to earning income, for example having access to cash or a bank account. For the agency measures examined in Table 2, the largest status changes are associated with household composition. The absence of the husband, due primarily to outmigration, is generally associated with increased freedom of movement and decision-making. The presence of another married man who is not her husband (such as her father-in-law) in the household is associated with lower women’s status on all measures. Being the senior married woman in the household is associated with higher status for several of the measures tested. Higher household consumption is associated with a decrease in freedom of movement for women.

Testing the role of children’s sex in motherhood effects

Statistically significant motherhood effects are found for six of the nine women’s status variables tested in Tables 2 and 3. Given the previous literature on son preference and women’s status, we might suspect that these motherhood effects would be enjoyed only by women with sons. Analyses presented in the supplementary material examined whether it was only mothers with sons who enjoyed a motherhood effect on their status. There was no evidence of differential motherhood effects associated with becoming the mother of one or more sons compared with becoming the mother of one or more daughters on freedom of movement, access to enabling resources, and most measures of decision-making (Tables A1 and A2). In only one of the nine status variables tested was a statistical difference by sex of children found. Having a son was associated with increased odds of the respondent reporting that she was the main financial decision maker in her household. The coefficient indicating that the respondent had a son was statistically different at the 0.05 level from the coefficient indicating that the respondent had a daughter. This finding fits with previous evidence from India suggesting that mothers of sons may enjoy greater influence over household finances (Kishore and Spears 2014). Without considering the sex of children born, the motherhood effect on being the main financial decision maker is not statistically significant (Model 5, Table 2). For the remaining measures of women’s status, including the other decision-making variables, motherhood effects do not vary by the sex of children born.

Discussion

Using the first large-scale panel study following women in India, this paper has demonstrated that women’s status is dynamic over time. Depending on the measure tested, between 13.9 and 82.1 per cent of women in the sample reported a change in status between IHDS I and II. Changes in status showed some associations with entry into motherhood and the adoption of permanent contraception. The findings also provided further evidence of the multidimensionality of women’s status. By examining measures of each dimension of women’s status in the household separately, the analysis revealed significant variation in the amount of change over time and whether those changes were associated with reproductive transitions.

The first dimension of women’s status examined in the analysis was women’s agency. There were positive and significant associations between freedom of movement and both the transition into motherhood and the adoption of permanent contraception. Household decision-making experienced less change over time; this dimension of women’s agency was more fixed. There were positive and statistically significant associations between both reproductive transitions and non-financial decision-making. In terms of financial decision-making, there was no evidence of motherhood effects and limited evidence of a permanent contraception effect. Adoption of permanent contraception showed a marginally significant association with participation in large purchase decision-making. This effect was smaller and less robust than that found by MacQuarrie (2009) using retrospective data. While there were no statistically significant motherhood effects on financial decision-making, analysis in the supplementary material suggested that mothers with sons did show increased odds of reporting that they were the main financial decision maker. This was the only variable for which there were statistically different effects by sex of children born.

Access to enabling resources was the second dimension of women’s status examined in this analysis. The transition into motherhood showed a statistically significant association with gaining access to the resources of a bank account and a marginally significant association with obtaining property rights and a communicative marital relationship. Analyses presented in the supplementary material revealed that these motherhood effects on access to enabling resources were not dependent on the sex of children born. There was no evidence of an impact of permanent contraception on women’s access to enabling resources.

The individual fixed effects models used in the analysis were advantageous because they removed much of the omitted variable bias that is present in regular regression modelling. However, as with all research, there were limitations to the analysis. It is possible that some key time-variant variable that could be behind some of the results was omitted from the model. It is also possible that the effects go in the opposite direction, with improved status leading to changes in motherhood status or the adoption of permanent contraception.

Another limitation to this project comes from the constraints of the data. First, we did not observe a very large share of the women experiencing the transition into motherhood. In India, childbearing often begins quite soon after marriage, leaving the window between marriage and first birth very short. Because the sample was drawn from ever-married women aged 15–49, over 94 per cent of the sample had already had at least one birth by the first wave of the survey. Despite this limitation, there were enough women experiencing the motherhood transition in the sample to capture statistically significant effects in the models. In addition, the large gap between IHDS I and II and the small number of childless women in IHDS I made it difficult to parse out nuanced effects associated with the sex of children born. In future, it would be advantageous to track women’s status in the household with a sample that included more recently married women who had not yet had their first birth and to collect status data at more frequent intervals. There will be numerous opportunities to extend the literature on women’s life course as more panel data become available.

This analysis was also constrained in its ability to uncover the mechanisms and causes of the associations documented. Future research is needed to uncover the meaning of these reproductive transitions for women’s lives. This is especially true for permanent contraception. There is growing evidence that this contraceptive practice is becoming a part of the life course for millions of Indian women. Social scientists know little about what this means for Indian women, their families, and their communities. This is an important area for future research.

Overall, these findings demonstrate that fertility behaviours should be viewed not only as an outcome of women’s status in the household; fertility behaviours may also affect a woman’s status. As demographers, it is essential that we recognize the complicated and bidirectional relationship between fertility and a woman’s status in the household. The results can be interpreted through the lens of motherhood’s central role in Indian women’s lives. Motherhood was positively associated with most measures of status in the household. Furthermore, it is only through becoming a mother that a woman has the chance to eventually become a mother-in-law. Becoming the senior married woman in the household, often through becoming a mother-in-law, was associated with large and statistically significant increases in freedom of movement and decision-making power for most measures tested. Even the permanent contraception effects measured in the analysis can be linked back to motherhood. Families adopt permanent contraception when they have reached their target fertility in terms of the sex composition and number of children (Edmeades et al. 2012; Calhoun et al. 2013). One interpretation of the permanent contraception effects is that they are capturing the status rewards that women enjoy for meeting their family’s desired fertility.

The findings of this paper highlight the centrality of women’s bodies and their reproductive contributions in determining the way that they are perceived and treated in society. While these effects may look like empowerment, researchers should exercise caution in interpreting life course effects in this way. While motherhood may have its social rewards, its absence for women who are unable to or choose not to have children can lead to stigma, diminished autonomy, desertion, and domestic violence (Riessman Kohler 2000; Mehta and Kapadia 2008; Nahar and Richters 2011; Bhambhani and Inbanathan 2018). The highest levels of women’s agency are associated with modern forms of reversible contraception rather than sterilization, particularly when these methods allow women to make independent contraceptive decisions regarding the timing and spacing of births (Goldin and Katz 2002; Pallikadavath et al. 2015; Silverman et al. 2016). Rather than as a process of empowerment, the associations between reproductive transitions and measures of women’s status documented in this analysis should be interpreted as evidence of the tight linkages between household power dynamics and the life course in the South Asian context.