Differences in Pre- and Post-Natal Expenditure across Female- and Male- Headed Households in India: A Non-Parametric Analysis

Abstract

This article investigates whether pre- and post-natal expenses are different across male- and female-headed households in India, using data from the 71st round of National Sample Survey. Results from fractional logit models provide evidence that female-headed households, compared to male-headed households, had 30 per cent increased likelihood of spending on postnatal care as a proportion of overall consumption expenditure, while no significant results were observed for prenatal expenditure. Results from Heckman two-part model also show that the decision on how much to spend on pre- or post-natal expenses is related to the gender of the household head. Overall, our results suggest that the gender of the household head may provide additional context in understanding expenditure patterns related to pre- and post-natal expenses in India.

Keywords

Pre- and post-natal expenses female-headed households fractional logit two-part model

Introduction

Investments on pre- and post-natal care have direct implications on lowering the incidence of premature births, low birth weight, infant mortality, maternal mortality and cognitive disorders in children, to name a few (Kattula et al., 2014; Turienzo et al., 2016). In a country like India, which ranks low in Maternal and Child Health (MCH) indicators—albeit there has been some modest progress achieved in recent years—effective pre- and post-natal investments assume an even more significant role, which can potentially help the country meet its SDG commitments related to MCH (Ghosh & Ghosh, 2020). Currently, the government provides most of the important pre-, post- and delivery care services free of cost, particularly, to the economically downtrodden; in addition to encouraging and incentivising the use of these services. However, a significant portion of the population across the socio-economic strata continues to incur huge out-of-pocket expenses related to maternal care (Akhtar et al., 2020; Goli et al., 2018; Leone et al., 2013; Yadav et al., 2021). The average out-of-pocket expenditure per delivery in public health facilities is INR 3,197 (NFHS-4, IIPS & ICF, 2017). While several factors in the literature are shown to be associated with the choice of uptake of pre- or post-natal care and the related expenses, we focus on one of the less explored elements, which in many ways are distinctive to India—the gender of the household head (Barman et al., 2020; Singh et al., 2012a; Yadav & Kesarwani, 2016). We explore this association between headship and pre- and post-natal expenditure using the National Sample Survey (NSS) 71st round.

Household Headship in India

In a patriarchal society such as India, women, as individuals, might not necessarily possess or exercise the autonomy to make individual or household decisions, including maternal care and the related expenses (Rajaram, 2013). These decisions typically rest with the ‘head’, who generally makes decisions on behalf of the household. The NSS data that we use in this study also defines ‘head’ as the person who is perceived by the members of the household to make decisions on behalf of the household members, which we use to measure household autonomy. While it is common in the literature to use educational attainment to measure autonomy, we choose gender of the household head because although education can be regarded as a gender equaliser with regards to human capital (which, in turn, translates to skills and labour market competency), it might still not necessarily guarantee autonomy in household decision-making; and, evidence for this claim exists both for developing and developed countries (Acharya, 2008; Albert & Escardíbul, 2017).

Women’s autonomy specifically relates to their power and agency within a household and encompasses decision-making autonomy, which is more closely represented by headship within a household as opposed to individual characteristics such as the level of educational attainment (Becker et al., 2006; Bloom et al., 2001; Ghuman et al., 2006). Hence, instead of individual-autonomy measuring variable like education, we make use of the variable that directly measures household decision-making authority from the survey—which is the ‘head’ of the household—and examine if gender of the household head is related to pre- and post-natal expenditure.

Copious evidence exists on the association between gender of the household head and healthcare outcomes (Ahmed et al., 2010; Doctor, 2011; Goebel et al., 2010; Haidar & Macau, 2009; among many others), although none for pre- or post-natal expenditure in India, to the best of our knowledge. It is well-documented in the literature on how expenditure and investment related to children’s welfare and long-term productive investments such as education and healthcare tend to be greater in households where woman is the decision maker (Handa, 1996; Khan & Khalid, 2010, 2012; Ogundari & Abdulai, 2014; Shroff et al., 2011).

We draw upon these inferences to hypothesise that pre- and post-natal expenditures differ across male- and female-headed households, and employ non-parametric estimation techniques to test our claim. We treat pre- and post-natal separately because (a) the factors that influence the decision for the uptake of each is dissimilar (Singh et al., 2012), and (b) much of the short- and long-term implications associated both with usage patterns and the frequency of pre- and post-natal care are different (IIPS, 2007–08).

Our study also attempts to advance the literature by employing a fractional logit model to gain insights on whether the fraction of pre- and post-natal expenditure as a proportion of total household expenditure is related to headship. In this context, while absolute value of expenses on goods and services are undoubtedly important to examine, relative expenses provide a better context to understand the choice patterns across competing expenses, given a household’s income or expenditure constraints. Understanding what proportion of household expenditure is spent towards pre- and post-natal care offers greater clarity on a household’s expenditure choice rather than merely knowing how much was spent.

To understand the choice influencers in even greater detail, we employ a two-part Heckman hurdle model that isolates the effect of headship on expenditure choice (if it exists) based on ‘participation’ and ‘consumption’ decisions. In other words, results from this model shed light on whether the effects of gender (of the household head) on the decision to spend on pre- and post-natal care is different from how much to spend on the same, once a positive amount is achieved. In summary, the study addresses the following questions:

Whether the gender of the household head is related to pre- and post-natal expenses, and if this relationship is different from overall household expenditure

Whether the fraction of pre- and post-natal expenditure (as a proportion of overall household expenditure) is greater among female-headed households compared to their male-headed counterparts

Whether the influence of gender on the decision to spend on pre- and post-natal care is different from how much to spend on the same

While a normal OLS regression is employed to address (a) above; a fractional logit model is employed to estimate (b); and a two-part (hurdle) model (modified Heckman model), to investigate (c). Results from our study affirm all of the above three hypotheses.

Methods

Data Source and Sample

For our analysis, we use data from the 71st round of the NSS (2014) titled, ‘Social Consumption: Health’. The 71st survey is the first of the NSS data gathered on health information post the Janani Shishu Suraksha Karyakram (JSSK), a significant nationwide programme instated in 2011, aimed at promoting safe pregnancy practices; hence, making data from this NSS round the most relevant and appropriate for our study. The survey covers all Indian states, achieving a total sample size of 65,932 households (333,104 individuals) and collects data on prevalence rate of morbidity among various age-sex groups and households’ health expenditures. It covers ailments under which medical care was sought, extent of use of government-provided services, expenditures incurred on treatment received by public and private sectors. MCH is one of the sub-components that collects data on (a) the extent of use of public and private services for childbirth, (b) the related costs, including expenses on pre- and post-natal care. Besides health information, the survey collects data on demographic and socio-economic characteristics of the individuals and households.

In this study, we make use of the following 4 (out of the 10) relevant sub-components or block datasets in the survey: (i) Block one and two: identification of sample household, (ii) Block three: household characteristics (includes household consumption expenditure), (iii) Block four: demographic information of household members and (iv) Block 11: pre- and post-natal care for women. The datasets were merged using a combination of household and individual ID numbers on a ‘one is to many’ basis; and since the component on MCH covers only women respondents, our final merged dataset contains only female respondents in the reproductive age group of 15–49 years. The final merged dataset gave us a sample size of 88,554 records.

Measuring Household Head

According to the NSS, a household head is the person formally in charge of the management of the household. He or she need not necessarily be the principal earning member of the household, but the customary head of the household decided on the basis of tradition (NSSO, 2001). It is left to the members of a household to decide upon whom they consider to be the head of the household. Since sex of the household head was not asked directly in the survey, we derived it by juxtaposing responses to the following two questions: (a) sex of the survey respondent and (b) relationship to the head. However, responses to these two questions still leaves few households which cannot strictly identify the gender of the household head.¹ Our dataset only includes those records wherein the gender of the household head is strictly distinguishable. This process reduced the sample by more than half from 88,554 to 41,123 (and of these households, approximately about 6% were headed by women). Roughly a sixth had expenses related to pre- and post-natal expenses, which further reduced our sample size to 6,273 households.

Measuring Expenditure Variables

The outcome variables used in this study include (a) monthly household consumption expenditure,² (b) pre- and post-natal expenditure (separately)³ and (c) fraction of pre- and post-natal expenditure—each calculated as the ratio of (b) to (a). Following the literature, we exclude outliers—greater than three times the mean—in the household expenditure; and cap the fraction of pre- and post-natal expenditure to one and consequently, our sample size marginally reduced by 2 per cent to 6,172 units, which was finally used in the analysis (Figure 1).

Other Covariates

Besides individual characteristics, which are used as controls in our model, we also control for variables that are commonly used in this literature (Bonu et al., 2009; Goli et al., 2016; Subramanian & Kawachi, 2004). Social and economic context, environmental and geographical factors in addition to individual biological factors are known to influence health behaviour (Giddens, 1986; Pebley et al., 1996; Sibley & Armbruster, 1997; Stokols, 1992). Community level factors include, for example, availability to health services, public health infrastructure, economic development, which can influence maternal healthcare outcomes. Hence, we examine the relationship between headship and pre- and post-natal care by accounting for factors at the individual, household and community levels.

Table 1 provides an overview and measurement of the variables used for the analysis. The covariates are classified as individual, household and community; and they were converted to binary form for ease of interpretation, following the literature. Note that an index was created to take the value of one if the household had at least two of the following: access to latrine, drainage facility, drinking water and non-conventional source of cooking fuel; and zero otherwise. Table 1 provides a snapshot.

Table 1.

Variable Measurement

Dependent Variables
Household consumption expenditure	Continuous (treated for outliers)
Prenatal expenditure	Continuous (treated for outliers)
Postnatal expenditure	Continuous (treated for outliers)
Prenatal expenditure as a proportion of total household consumption expenditure	Ratio
Postnatal expenditure as a proportion of total household consumption expenditure	Ratio

Independent Variables
Individual Characteristics		Household Characteristics		Community Characteristics
Age	Continuous	Female headed household (FHHH)	1 = FHHH (If Sex is Female and Relationship to head is self or Sex is Male and Relationship to head is spouse) 0 = Others	Type of area	1 = Rural area 0 = Urban area
Marital status	1 = Single 0 = Ever married (Includes divorced, widowed)	Male headed household (MHHH)	1 = MHHH (If Sex is Male and relationship to head is self or sex is female and relationship to head is spouse)
Caste	1 = SC/ST 0 = Others	Source of livelihood in rural areas	1 = Agriculture 0 = Others
Religion	1 = Hindu 0 = Others	Source of livelihood in urban areas	1 = Salaried/wage earner 0 = Self-employed
Education	1 = Illiterate 0 = Some form of literacy	Household size	Continuous (Treated for outliers)
Institutional delivery	1 = Yes 0 = No	WASH: Household facilities (water, sanitation and cooking fuel)	1 = Presence of any two of these facilities: latrine, drainage, access to drinking water and LPG as source of cooking fuel 0 = No access

Source: The authors.

Statistical Software

All data were analysed using Stata 14 (StataCorp, 2015).

Results

Profile Differences across FHHH and MHHH

Female headed households (FHHH) constitute 2.5 per cent of the study sample. While the overall consumption expenditure is greater in Male headed households (MHHH) compared to FHHHs by 18 per cent, there are no significant differences across the households on prenatal (9%) or postnatal expenditure (6%) as shown in Table 2. The distribution of the expenses across the households do not vary either, as shown by comparable values of the coefficient of variations (that measures the spread of the distribution).

Table 2.

Expenditure Profile by Gender of the Household Head

	Entire Sample		FHHH			MHHH			p-Value*
	Mean	SD	Mean	SD	CV	Mean	SD	CV	p-Value*
Overall HH	6916.4	3897.2	5647.9	3163.6	0.56	6949.1	3,909	0.56	.0001***
Prenatal exp	2888.3	4171.4	2609.9	3077.4	1.18	2895.5	4195.8	1.45	.4002
Postnatal exp	1689.6	3033.3	1592.8	1759.4	1.10	1692.1	3059.1	1.81	.6873
N	6,172		155			6,017

Source: The authors.

Notes: *Two-sample pooled t-test.

***Significant at the 1% level.

Even with regards to other covariates, in general, there are no significant differences across FHHHs and MHHHs as shown in Table 3, either on the mean levels of the variables or the spread of the variables (standard deviations), except for percentage of workforce engaged in salaried jobs in urban areas, which is much greater in FHHHs (91%) compared to MHHHs (62%).

Table 3.

Summary Statistics of Variables across Male- and Female-Headed Households

Variable	MHHH		FHHH
Variable	Mean/%	Std. Dev.	Mean/%	Std. Dev.
Caste (SC/ST vs. others)	0.35	0.48	0.25	0.44
Religion	0.72	0.45	0.61	0.49
Age	27.48	5.07	28.00	5.10
Marital status (married vs. single)	1.00	0.00	0.99	0.08
Education	0.24	0.42	0.26	0.44
WASH	0.65	0.48	0.61	0.49
Household size	4.55	1.45	3.66	1.38
Rural livelihood	0.73	0.45	0.70	0.46
Urban livelihood	0.62	0.48	0.91	0.29
Institutional delivery	0.89	0.31	0.85	0.36
Rural (vs. urban)	0.57	0.50	0.69	0.46

Source: The authors.

Note: *Except for age and household size, all other variables are binary and hence can be interpreted in percentages.

Results from Linear Regressions

Association between gender of household head and each of prenatal, postnatal and overall household expenditure

While the raw numbers presented above do not provide any evidence for differences in expenditure across FHHHs and MHHHs, we employ a set of linear regression (OLS) models—controlling for other pertinent variables—that could potentially influence expenditure levels to investigate the relationship between gender of household head and (a) overall household consumption expenditure, (b) prenatal expenditure and (c) postnatal expenditure. The model takes the following form:

Y_{h} = β_{0} + β_{1} F H H_{h} + β_{2} {\sum_{h = 1}^{n}}_{​}^{​} X_{h} + \in,

(1)

where $Y_{h}$ is the amount of expenditure incurred on each of overall consumption, pre- and post-natal care. $F H H_{h}$ denotes whether the household is headed by a woman; and X is a vector of variables that includes individual, household and community characteristics as explained in Table 1.

We employ three variants for each of the expenditures:

Model 1: Base model with only the household head’s gender

Model 2: Additionally, controlled for individual characteristics such as caste, age, marital status, education and religion

Model 3 (Fully-specified model): In addition to the covariates in Model 2, controlled for other household and community characteristics

Note: We follow the same modelling sequence on fractional logit model and the two-part model.

In the base model, we find notable relationship between gender of the household head and household expenditure, which is statistically significant as well (Table 4). Compared to MHHHs, FHHH’s overall consumption expenditure is lower by 1,301 Indian Rupees (18%). The economic and statistical significance is robust even after controlling for individual, household and community characteristics—suggesting no confounding or mediating effects from other relevant variables. The difference in expenditure values between the base model and the fully controlled model is less than 2 per cent, confirming the robustness of the model specification. Results from this table provide evidence that overall consumption expenditure is significantly lower in FHHHs compared to MHHHs.

While pre- and post-natal expenses are also lower in FHHH compared to MHHH, the effects are only a fraction compared to the results from overall household expenditure presented earlier. Note that the household expenditures are monthly values, while pre- and post-natal expenses are measured through the entire delivery event. Regardless, the values are not statistically significant to make a compelling argument on pre- or post-natal expenditure differences across the two sets of households.⁴

Table 4.

Relationship between Overall Consumption, Pre- and Post-Natal Expenditures and Gender of the Household Head Using OLS Regressions

Expenditure	Model I			Model II			Model III
Expenditure	Overall Expenditure	Prenatal	Postnatal	Overall Expenditure	Prenatal	Postnatal	Overall Expenditure	Prenatal	Postnatal
Female head	–1,301***	–293	–100	–1,407***	–254	–116	–1,330***	–281	–106
Caste				–1,084***	–1,132***	–302***	–504***	–879***	–173**
Religion				195*	544***	57	271***	570***	73
Age				165***	29***	25***	142***	36***	20***
Marital status				–2,095	220	–946	–1,178	430	–749
Education				–3,121***	–1,370***	–553***	–1,142***	–963***	–330***
HH facilities							1,384***	661***	278***
Household size							–3,924***	965	–1059*
Rural livelihood							–264**	–240	–44
Urban livelihood							720***	312***	183*
Institutional delivery							519***	710***	83
Type of area (rural)							–1,402***	–496***	–329***

Source: The authors.

Note: ***Significant at the 1% level. **Significant at the 5% level. *Significant at the 10% level.

Results from Fractional Logit Model

Association between headship and pre- and post-natal expenditure as a proportion of overall household expenditure

To the best of our knowledge, this is the first study to employ fractional logit model to understand the effects of gender on the fraction of household expenditure towards pre- and post-natal care. Hence, the functional form is slightly altered from Equation (1), and the underlying distribution is logistic instead of standard-normal, as in an OLS. Therefore, the expected value of Y, which is modelled as share of pre- or post-natal expenses as a proportion of overall consumption expenditure is given by the following⁵:

E (Y_{h} | X) = \frac{\exp (X β)}{1 + \exp (X β)} .

(2)

Fractional logit models offer the flexibility of using dependent variables that have fractional values or proportions, yet have bounded values on either extreme. In our analysis, the fraction of pre- and post-natal expenses are bounded by zero and one—implying that the predicted values of Y will also be bounded between those two values.

The odds ratios for gender variables across the three model variants remain stable and retain their respective statistical significances, implying no confounding or mediating effects of other variables on pre- or post-natal expenses (Table 5). Although FHHHs have a greater likelihood of spending on prenatal expenses (about 20 percentage points) compared to their male-headed counterparts, the values are not statistically significant to warrant a systematic explanation. On the other hand, FHHHs have greater odds of spending on postnatal expenses (by 30 percentage points), which is statistically significant. Once again, the odds ratios for gender variables are robust to model specification. The results suggest that the gender of the household head is related to postnatal but not prenatal expenditure, which is consistent with the results from other studies in the literature (Mistry et al., 2009).

Table 5.

Relationship between Pre- and Post-Natal Expenditure as Proportions of Overall Consumption Expenditure and Gender of the Household Head Using Fractional Logit Models

Expenditure	Model I		Model II		Model III
Expenditure	Prenatal	Postnatal	Prenatal	Postnatal	Prenatal	Postnatal
Female head	1.19	1.33**	1.18	1.31*	1.18	1.30*
Caste			0.76***	0.97	0.77***	0.96
Religion			1.13***	0.99	1.11***	0.98
Age			0.98***	0.99**	0.98***	0.98***
Marital status			1.49	0.41	1.41	0.39
Education			0.78***	1.03	0.84***	1.01
HH facilities					1.04	0.91*
Household size					2.36***	0.89
Rural livelihood					1.01	1.01
Urban livelihood					1.04	1.01
Institutional delivery					1.27***	0.99
Type of area (rural)					1.03	0.99

Source: The authors.

Note: ***Significant at the 1% level. **Significant at the 5% level. *Significant at the 10% level.

Results from Two-Part Model

Effects of Headship on Participation and Consumption Decision Related to Pre- and Post-Natal Expenditure

As it has been discussed earlier, a two-part model is employed to understand whether the choice to spend on pre- and post-natal care is any different from the choice of how much to spend on the same once a non-zero value of expenditure is achieved. In effect, the results from a two-part model would help understand the two choices on pre- and post-natal expenditure by households—namely the participation and consumption decisions. The two-part model uses a logit function for participation and a generalised linear model (with a gamma link function) for the consumption decision. While the first part of participation decision flows directly from the logit model as in Equation (2), the consumption decision takes the following form:

E (Y | Y > 0, X) = \exp (X α) .

(3)

Combining (2) and (3);

\begin{matrix} E (Y | X) = p r o b (Y > 0 | X) * p r o b (Y | Y > 0, X) \\ = \frac{\exp (X α) * \exp (X β)}{1 + \exp (X β)} + \frac{\exp (X (α + β)}{1 + \exp (X β)} . \end{matrix}

(4)

As with previous model, we begin with only the interest variables and sequentially add other controls.

Prenatal Expenses

For prenatal expenses, although odds ratios pertaining to participation decision, whether to spend or not, is lower in FHHHs compared to MHHHs; they are not statistically significant (Table 6). However, once a positive spending threshold is achieved, gender of the household head is positively associated with the consumption decision—that is, how much to spend on prenatal care, and this value is 18 percentage points higher in FHHHs compared to MHHHs. It is important to note that this difference is consistent with the results obtained from the fractional logit model. The results are robust to model specification; as the values of odds ratios or their statistical significance do not change across the three model variants.

Table 6.

Relationship between Prenatal Expenditure as Proportions of Overall Consumption Expenditure and Gender of the Household Head Using a Two-Part Heckman Hurdle Model with Gamma Link Function for GLM Estimation

Prenatal Expenditure	Model I		Model II		Model III
Prenatal Expenditure	Part-1	Part-2	Part-1	Part-2	Part-1	Part-2
Female head	0.76	1.20*	0.71	1.20*	0.81	1.18**
Caste			0.59***	0.80***	0.65***	0.80***
Religion			0.87	1.15***	0.93	1.13***
Age			1.01	0.98***	1.00	0.98***
Marital status			1.00***	1.24	1.00	1.10
Education			0.49***	0.87***	0.59***	0.92***
HH facilities					1.58***	0.97
Household size					1.35	2.34***
Rural livelihood					1.41**	0.97
Urban livelihood					0.92	1.04
Institutional delivery					1.05	1.27***
Type of area (rural)					0.59***	1.08

Source: The authors.

Note: ***Significant at the 1% level. **Significant at the 5% level. *Significant at the 10% level.

Part-1 represents the participation decision and Part-2, the consumption decision.

Postnatal Expenses

As with prenatal expenses, even for postnatal expenses, the participation decision, whether to spend or not, is not statistically significant to make claims on the association between the gender of the household head and postnatal expenditure, while the consumption decision is; and the results are robust to the full specified model (Table 7). FHHHs have a greater likelihood of spending more on postnatal care compared to MHHHs—by 26 percentage points, which is once again consistent with the results obtained from the fractional logit model.

Table 7.

Relationship between Postnatal Expenditure as Proportions of Overall Consumption Expenditure and Gender of the Household Head Using a Two-Part Heckman Hurdle Model with Gamma Link Function for GLM Estimation

Postnatal Expenditure	Model I		Model II		Model III
Postnatal Expenditure	Part-1	Part-2	Part-1	Part-2	Part-1	Part-2
Female head	1.22	1.30**	1.13	1.29*	1.24	1.26*
Caste			0.54***	1.04	0.54*	1.03
Religion			0.87	1.01	0.85	1.00
Age			1.01	0.98***	1.00	0.98***
Marital status			1.00	0.38	1.00	0.37
Education			0.78*	1.06	0.76***	1.03
HH facilities					0.87	0.93
Household size					3.17*	0.78
Rural livelihood					1.46**	0.97
Urban livelihood					0.92	1.02
Institutional delivery					0.84	1.01
Type of area (rural)					0.56***	1.03

Source: The authors.

Note: ***Significant at the 1% level. **Significant at the 5% level. *Significant at the 10% level.

Part-1 represents the participation decision and Part-2, the consumption decision.

Discussion

This article modestly adds to the existing knowledge in the gender and maternal health space by examining the association between the gender of the household head on each of pre- and post-natal expenses, employing fractional logit models that we believe have not been handled in prior studies. We treat pre- and post-natal care separately because the factors that influence the decision for the uptake of each is dissimilar (Singh et al., 2012); also, the short- and long-term risks associated with each of pre- and post-natal care usage patterns and frequency are not the same (IIPS, 2007–08).

While studies in the past, such as Skordis-Worrall et al. (2011), Bonu et al. (2009), Mohanty and Srivastava (2012), have analysed the relationship between overall maternal expenditure and socio-economic factors in general, they have been limited by small sample size (running into issues of generalisation), or have run into challenges of using surveys fraught with recall issues or inaccuracies in specifying the nature of expenditure (Goli et al., 2016). To circumvent these shortcomings, we use data from India’s NSS 71st round on health survey, which has not been explored to address the question we have in our study. This particular data set serves to be crucial and pertinent to understand this research question, given that this survey was the first health survey from NSS to be administered post the JSSK policy aimed at overhauling the MCH space in India.

Findings from our analysis show that although FHHHs have significantly lower overall consumption expenditure compared to MHHHs, the levels of pre- and post-natal expenses are not. At the same time, results from the fractional logit model points that FHHHs spend relatively greater proportion of their expenses on postnatal care compared to MHHHs. Results from the two-part model, however, provides evidence that both for pre- and post-natal care, the participation decision, whether or not to spend, was not associated with the gender of the household head, whereas, how much to spend (once a positive amount is spent) was.

Before we situate the findings, and generalise our results, the following data constraints must be recognised. We use self-reported responses on headship: hence, de-jure and de-facto differences cannot be strictly isolated. Also, like many other studies in the literature, our study uses cross-sectional data, and hence our results cannot throw light on the changes in the pre- or post-natal expenditure patterns or how gender is related to them over time. Since NSS does not follow the same households over time, a panel analysis is not possible, and a repeated cross-section would weaken the already low sample size of FHHHs. Also, eliciting information on households’ income or expenditure has structural issues related to accuracy and reliability—households typically tend to underreport income and overstate their expenses. These data-related issues, however, we believe, should not take away the significance or consistency of our findings, or affect the generalisability or usability of our results.

The results from our OLS, fractional logit and the two-part models taken together seem to align with the (empirically supported) claim that women, if given the power to make decisions for the household, tend to spend and invest relatively more on human capital and family welfare (Blumberg, 1988; DeGraff & Bilsborrow, 1993). Consequently, in female-headed households—where women are the typical decision makers—expenditures related to health and education are higher compared to male-headed households (Donkoh & Amikuzuno, 2011; Furuta & Salway, 2006; Handa, 1996; Khan & Khalid, 2010, 2012; Panda, 1997; Seebens, 2009). Specifically, our results corroborate with other studies (Ahmed et al., 2010; Furuta & Salway, 2006; Sado et al., 2014; Thapa & Niehof, 2013; to name a few) that provide evidence that empowering women’s households is associated with increased pre- and post-natal healthcare spending and utilisation.

In summary, our results show that gender of the household head may provide additional context for understanding pre- and post-natal expenditure patterns in India; and consequently, policies related to the uptake of pre- and post-natal care could be customised based on gender of the household head. Further, maternity schemes such as the Janani Suraksha Yojana could integrate components of outreach and training pertaining to empowering women on decision-making (within and outside of the household) which, in turn, could help in effective decision-making related to maternal and child healthcare. Such policy measures are likely to translate to increased awareness of safe maternal practices and the rights and facilities that women could avail in this regard, specifically leading to increased uptake of pre- and post-natal services in India. One natural extension of our study would be to investigate the presence of non-random differences in female- and male-headed households that can help to distinguish need- and choice-based expenditure in FHHH, which we leave for future research to examine.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

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

Notes

References

Acharya

(2008). Women’s education and intra-household autonomy: Evidence from Nepal. Journal of Development and Social Transformation, 5(1), 5–12.

Ahmed

, Creanga

A. A.

, Gillespie

D. G.

, & Tsui

A. O.

(2010). Economic status, education and empowerment: Implications for maternal health service utilization in developing countries. Plos One, 5(6), e11190. 10.1371/journal.pone.0011190

Akhtar

M. D. A.

, Ahmad

, & Chowdhury

I. R.

(2020). Measuring socio-economic inequality in self-reported morbidity in India: Decomposition analysis. Review of Development and Change, 25(1), 89–111. 10.1177/0972266120916317

Albert

, & Escardíbul

J.-O.

(2017). Education and the empowerment of women in household decision-making in Spain. International Journal of Consumer Studies, 41(2), 158–166. 10.1111/ijcs.12326

Barman

, Saha

, & Chouhan

(2020). Impact of education on the utilization of maternal health care services: An investigation from National Family Health Survey (2015–16) in India. Children and Youth Services Review, 108, 104642. 10.1016/j.childyouth.2019.104642

Becker

, Fonseca-Becker

, & Schenck-Yglesias

(2006). Husbands’ and wives’ reports of women’s decision-making power in Western Guatemala and their effects on preventive health behaviors. Social Science & Medicine, 62(9), 2313–2326. 10.1016/j.socscimed.2005.10.006

Bloom

S. S.

, Wypij

, & Gupta

M. D.

(2001). Dimensions of women’s autonomy and the influence on maternal health care utilization in a north Indian city. Demography, 38(1), 67–78. 10.1353/dem.2001.0001

Blumberg

R. L.

(1988). Income under female versus male control: Hypotheses from a theory of gender stratification and data from the Third World. Journal of Family Issues, 9(1), 51–84. 10.1177%2F019251388009001004

Bonu

, Bhushan

, Rani

, & Anderson

(2009). Incidence and correlates of ‘catastrophic’ maternal health care expenditure in India. Health Policy and Planning, 24(6), 445–456. 10.1093/heapol/czp032

10.

DeGraff

D. S.

, & Bilsborrow

R. E.

(1993). Female-headed households and family welfare in rural Ecuador. Journal of Population Economics, 6(4), 317–336.

11.

Doctor

H. V.

(2011). Does living in a female-headed household lower child mortality? The case of rural Nigeria. Rural Remote Health, 11(2), 1635. https://pubmed.ncbi.nlm.nih.gov/21486098/

12.

Donkoh

S. A.

, & Amikuzuno

J. A.

(2011). The determinants of household education expenditure in Ghana. Educational Research and Reviews, 6(8), 570–579. https://academicjournals.org/article/article1379768521_Donkoh%20and%20Amikuzuno.pdf

13.

Furuta

, & Salway

(2006). Women’s position within the household as a determinant of maternal health care use in Nepal. International Family Planning Perspectives, 32(1), 17–27. 10.1363/3201706

14.

Ghosh

, & Ghosh

(2020). Maternal health care in India: A reflection of 10 years of National Health Mission on the Indian maternal health scenario. Sexual & Reproductive Healthcare, 25, 100530. 10.1016/j.srhc.2020.100530

15.

Ghuman

S. J.

, Lee

H. J.

, & Smith

H. L.

(2006). Measurement of women’s autonomy according to women and their husbands: Results from five Asian countries. Social Science Research, 35(1), 1–28. 10.1016/j.ssresearch.2004.06.001

16.

Giddens

(1986). The constitution of society: Outline of the theory of structuration. University of California Press.

17.

Goebel

, Dodson

, & Hill

(2010). Urban advantage or urban penalty? A case study of female-headed households in a South African city. Health & Place, 16(3), 573–580. 10.1016/j.healthplace.2010.01.002

18.

Goli

, Moradhvaj

Rammohan

Shruti

, & Pradhan

(2016). High spending on maternity care in India: What are the factors explaining it? Plos One, 11(6), e0156437. 10.1371/journal.pone.0156437

19.

Goli

, Rammohan

, & Moradhvaj. (2018). Out-of-pocket expenditure on maternity care for hospital births in Uttar Pradesh, India. Health Economics Review, 8(1), 1–16. 10.1186/s13561-018-0189-3

20.

Haidar

, & Kogi-Makau

(2009, Feb). Gender differences in the household-headship and nutritional status of pre-school children. East African Medical Journal, 86(2), 69–73.

21.

Handa

(1996). Expenditure behavior and children’s welfare: An analysis of female headed households in Jamaica. Journal of Development Economics, 50(1), 165–187. 10.1016/0304-3878(96)00008-9

22.

International Institute for Population Sciences (IIPS), & Macro International. (2007). National Family Health Survey (NFHS-3), 2005–06, Volume I. https://dhsprogram.com/pubs/pdf/frind3/frind3-vol1andvol2.pdf

23.

International Institute for Population Sciences (IIPS), & ICF. (2017). National Family Health Survey (NFHS-4), 2015–16: India. https://dhsprogram.com/pubs/pdf/FR339/FR339.pdf

24.

Kattula

, Sarkar

, Sivarathinaswamy

, Velusamy

, Venugopal

, Naumova

E. N.

, Muliyil

, Ward

, & Kang

(2014). The first 1000 days of life: Prenatal and postnatal risk factors for morbidity and growth in a birth cohort in southern India. BMJ Open, 4(7), e005404. 10.1136/bmjopen-2014-005404

25.

Khan

, & Khalid

(2010). Household consumption patterns in Pakistan: Evidence from household data. NUST Journal of Business and Economics, 3(1), 1–14.

26.

Khan

A. H.

, & Khalid

(2012). Consumption patterns of male and female headed households in Pakistan: Evidence from PSLM 2007-08. The Pakistan Development Review, 51(4), 465–478. https://pide.org.pk/research/consumption-patterns-of-male-and-female-headed-households-in-pakistan-evidence-from-pslm-2007-08-2/

27.

Leone

, James

K. S.

, & Padmadas

S. S.

(2013). The burden of maternal health care expenditure in India: Multilevel analysis of national data. Maternal and Child Health Journal, 17(9), 1622–1630. 10.1007/s10995-012-1174-9

28.

Mistry

, Galal

, & Lu

(2009). Women’s autonomy and pregnancy care in rural India: A contextual analysis. Social Science & Medicine, 69(6), 926–933. 10.1016/j.socscimed.2009.07.008

29.

Mohanty

S. K.

, & Srivastava

(2012). Out-of-pocket expenditure on institutional delivery in India. Health Policy and Planning, 28(3), 247–262. 10.1093/heapol/czs057

30.

National Sample Survey Organisation, Ministry of Statistics & Program Implementation. (2001). Concepts and definitions used in NSS. Government of India. https://mospi.gov.in/documents/213904/0/concepts_golden.pdf/e98fc072-8660-edd9-f179-ce95674f4ca5?t=1615539414160#:~:text=The%20wide%20variety%20of%20subjects,enterprise%20surveys%20(4)%20Village%20surveys

31.

Ogundari

, & Abdulai

(2014). Determinants of household’s education and healthcare spending in Nigeria: Evidence from survey data. African Development Review, 26(1), 1–14.

32.

Panda

P. K.

(1997). Female headship, poverty and child welfare: A study of rural Orissa. Economic and Political Weekly, 32(43), 73–82. https://www.epw.in/journal/1997/43/review-womens-studies-review-issues-specials/female-headship-poverty-and-child

33.

Pebley

A. R.

, Goldman

, & Rodriguez

(1996). Prenatal and delivery care and childhood immunization in Guatemala: Do family and community matter? Demography, 33(2), 231–247. https://pubmed.ncbi.nlm.nih.gov/8827167/

34.

Rajaram

(2013). Women’s autonomy, maternal and child healthcare in India: Are the women and children from poorest households particularly disadvantaged? Review of Development and Change, 18(1), 19–35. 10.1177%2F0972266120130102

35.

Sado

, Spaho

, & Hotchkiss

D. R.

(2014). The influence of women’s empowerment on maternal health care utilization: Evidence from Albania. Social Science & Medicine, 114, 169–177. 10.1016/j.socscimed.2014.05.047

36.

Seebens

(2009). Child welfare and old-age security in female headed households in Tanzania (Discussion Paper No. 3929). Institute for the Study of Labor (IZA). 10.2139/ssrn.1329567

37.

Shroff

M. R.

, Griffiths

P. L.

, Suchindran

, Nagalla

, Vazir

, & Bentley

M. E.

(2011). Does maternal autonomy influence feeding practices and infant growth in rural India? Social Science & Medicine, 73(3), 447–455. 10.1016/j.socscimed.2011.05.040

38.

Sibley

, & Armbruster

(1997). Obstetric first aid in the community—Partners in safe motherhood a strategy for reducing maternal mortality. Journal of Nurse-Midwifery, 42(2), 117–121. 10.1016/S0091-2182(97)00022-0

39.

Singh

, Padmadas

S. S.

, Mishra

U. S.

, Pallikadavath

, Johnson

F. A.

, & Matthews

(2012). Socio-economic inequalities in the use of postnatal care in India. PLoS One, 7(5), e37037. 10.1371/journal.pone.0037037

40.

Singh

, Yadav

, & Singh

(2012). Utilization of postnatal care for newborns and its association with neonatal mortality in India: An analytical appraisal. BMC Pregnancy and Childbirth, 12(1), Article 33. 10.1186/1471-2393-12-33

41.

Singh

P. K.

, Rai

R. K.

, Alagarajan

, & Singh

(2012). Determinants of maternity care services utilization among married adolescents in rural India. PLoS One, 7(2), e31666. 10.1371/journal.pone.0031666

42.

Skordis-Worrall

, Pace

, Bapat

, Das

, More

N. S.

, Joshi

, Pulkki-Brannstrom

A.-M.

, & Osrin

(2011). Maternal and neonatal health expenditure in Mumbai slums (India): A cross sectional study. BMC Public Health, 11(1), Article 150. 10.1186/1471-2458-11-150

43.

StataCorp. (2015). Stata Statistical Software: Release 14. StataCorp LP.

44.

Stokols

(1992). Establishing and maintaining healthy environments: Toward a social ecology of health promotion. American Psychologist, 47(1), 6–22. 10.1037//0003-066x.47.1.6

45.

Subramanian

S. V.

, & Kawachi

(2004). Income inequality and health: What have we learned so far? Epidemiologic Reviews, 26(1), 78–91. 10.1093/epirev/mxh003

46.

Thapa

D. K.

, & Niehof

(2013). Women’s autonomy and husbands’ involvement in maternal health care in Nepal. Social Science & Medicine, 93, 1–10. 10.1016/j.socscimed.2013.06.003

47.

Turienzo

C. F.

, Sandall

, & Peacock

J. L.

(2016). Models of antenatal care to reduce and prevent preterm birth: A systematic review and meta-analysis. BMJ Open, 6(1), e009044. 10.1136/bmjopen-2015-009044

48.

Yadav

, & Kesarwani

(2016). Effect of individual and community factors on maternal health care service use in India: A multilevel approach. Journal of Biosocial Science, 48(1), 1–19. 10.1017/s0021932015000048

49.

Yadav

, John

, Menon

G. R.

, & Allarakha

(2021). Out-of-pocket payments for delivery care in India: Do households face hardship financing? Journal of Health Management, 23(2), 197–225. 10.1177%2F09720634211011552