https://orcid.org/0000-0001-6263-4410
Figures
Abstract
Background
Understanding variations in women’s work participation at lower administrative levels, such as districts, is a missing link in identifying trends, patterns and variation that can offer insights into this long-term stagnation. We link data from the 2019–21 Indian National Family Health Survey and the 2011 Indian Population and Housing Census to generate estimates of women’s work within 640 districts of India, and to examine the spatial clustering of women’s work across these districts. We examine women’s work through three outcome variables, namely, district-level estimates of 1) percentage of women who worked in the past 12 months, 2) percentage of women who were self-employed in the past 12 months, and 3) percentage of women who earned cash in the past 12 months.
Results
Diagnostic measures confirm that our model-based estimates are robust enough to provide reliable district-level estimates of women’s work in India. Women’s work and cash earnings were lowest in the districts of the central, eastern, and northern regions, and highest in the southern region. Self-employment rates for women were generally low in Indian districts, except for districts in Himachal Pradesh and the north-eastern region.
Conclusions
Considerable spatial heterogeneity in women’s work has been found across 640 districts of India. Our study demonstrates that estimated percentage of women who worked in the past 12 months, estimated percentage of women who earned cash in the past 12 months and estimated percentage of women who were self-employed in the past 12 months all vary substantially at the district level. Having only state-level estimates may thus be inadequate to inform efforts to remediate low levels of women’s work in India. The insights from our current study may help in the formulation and implementation of targeted policies that increase opportunities for women to expand their paid work in India.
Citation: Srivastava S, Kumar K, McDougal L, Upadhyay AK, Hay K, Singh A (2025) Mapping women’s work in India: An application of small area estimation. PLoS ONE 20(2): e0317783. https://doi.org/10.1371/journal.pone.0317783
Editor: Nishith Prakash, Northeastern University, UNITED STATES OF AMERICA
Received: August 20, 2024; Accepted: January 3, 2025; Published: February 19, 2025
Copyright: © 2025 Srivastava et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: Authors have used NFHS-5 and Census data in this paper. The NFHS-5 data can be obtained on request from the Demographic and Health Surveys (DHS) website https://dhsprogram.com/methodology/survey/survey-display-355.cfm. The Census 2011 data may be downloaded without restrictions from the website https://censusindia.gov.in/census.website/. Authors did not receive any special privileges in accessing the data that other researchers would not have.
Funding: Authors are thankful to the Bill and Melinda Gates Foundation for providing financial support for this research under Investments INV-008648 and INV-047355. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Abbreviation: CI, confidence intervals; CV, coefficient of variation; DHS, Demographic and Health Survey; GLMM, Generalized Linear Mixed Models; IHDS, India Human Development Survey; IPI, India Patriarchy Index; LISA, Local Indicator of Spatial Association; LMICs, low- and middle- income countries; MGNREGA, Mahatma Gandhi National Rural Employment Guarantee Act; MoSPI, Ministry of Statistics and Programme Implementation; NFHS, National Family Health Survey; NSO, National Statistical Office; NSS, National Sample Survey; NSSO, National Sample Survey Organization; OBC, other backward classes; OLS, ordinary least square; PLFS, Periodic Labour Force Survey; SAE, Small-area estimation; SC/ST, scheduled caste/ tribes; SDGP, State Gross Domestic Product; SER, Spatial Error Regression; SES, Socio-economic status
Introduction
Work, especially among women, is often considered as a measure of economic empowerment [1]. It can provide an opportunity to improve women’s wellbeing and capabilities, which may help empower them [2, 3]. However, women’s work participation has remained low in India in the last three decades, with fewer than a quarter of women working in 2018 [4]. A comparison with other BRICS countries (an intergovernmental organization comprising Brazil, Russia, India, China, South Africa, Iran, Egypt, Ethiopia and United Arab Emirates) suggests that India ranks last among the BRICS countries [5]. India ranked second last among the G-20 countries in terms of women’s work participation [5]. Among south Asian countries, India has the lowest female work participation [5]. Given these statistics, women’s work participation in India is a matter of concern.
Another important concern is that the static and decreasing trend in female work participation in India. Statistics from the Periodic Labour Force Survey (PLFS) shows that female labour force participation remained static between 25% and 27% between 1990 and 2005. Subsequently, there was a further decline in female labour force participation to 25% in 2017–2018 [4]. This trend stands in stark contrast to India’s dramatic rise in female literacy from 34% in 1990 to 69% in 2022 [6], rise in daily wages of women workers from $1.2 in 1990 to $2.8 in 2012 [7], decline in fertility rates from 3.4 children per woman in 2005 to 2.0 children per woman in 2021 [8], and high average annual GDP growth of 8.8% [9]. This was also the period when the largest employment generation programme, Mahatma Gandhi National Rural Employment Guarantee Act (MGNREGA), was successfully rolled-out in India.
These trends have shown some instability in recent years, with an increase in female labour force participation following the COVID-19 pandemic [10]. According to the 2022–2023 PLFS, the female labour force participation rate was 37%, which, though an increase of 12 percentage points from 2017–18 estimates, remains much lower than male labour force participation rate (79%). Female labour force participation varied by urban-rural residence; women in rural areas (42%) were more likely to work compared with women residing in urban areas (25%). Female labour force participation rates also varied across the Indian states; the lowest being in Bihar (11%) and highest in Himachal Pradesh (63%) [10].
Past studies have used female participation in the labour market, the labour force and the workforce, for analysing women’s participation in economic activities. While studies have interchangeably used these terms, these denote different dimensions of work [11]. While labour force includes all types of employment, labour market participation excludes unpaid family workers who do not enter the market [12]. Similarly, labour force consists of both employed and unemployed, whereas workforce captures only those who are employed. A number of studies in India used the term laborforce or workforce to examine women’s work [13]. Some studies used paid employment as a measure of women employment [14, 15]. A study by Jose (2008) analysed the relationship between women’s paid employment and their autonomy. He disaggregated the employed women into those who are paid and unpaid for their work and finally defined paid employment in the form of either cash or kind or both. Research by Lahoti and Swaminathan (2016) explored the relationship between economic development and women labour force participation; analyzed paid and unpaid work separately. They included any paid (cash or in kind) economic activity including self-employment, wage employment, and casual labour in paid work. On the other hand, unpaid work referred to assistance in the operation of a family farm or enterprise, and that did not receive regular remuneration in cash or kind. In sum, past studies have used women’s work in last 30 days, paid work, unpaid work or self-employment to examine women’s economic participation in India. Taking cue from the past studies, we used three indicators of women’s work—women who worked in the past 12 months, women who earned cash in the past 12 months and women who were self-employed in the past 12 months–in our study. We used these three indicators because women’s work participation in general and self-employment and cash earnings in particular can go a long way in improving the status of women [16].
Estimates from the 2019–21 National Family Health Survey (NFHS–5) indicate that in the past 12 months, only 31% of women worked (excluding household work) and only 4% of women were self-employed. One in four women (26%) earned cash in the past 12 months [8]. However, these estimates varied substantially across Indian states. Estimates of the percentage of women who worked in the past 12 months ranged between 11% in Lakshadweep and 49% in Manipur; women who earned cash in the past 12 months varied from 11% in Lakshadweep to 45% in Telangana and; women who worked as the self-employed varied from 2% in Karnataka to 17% in Sikkim. While a number of recent national and international research has examined the phenomena of women’s work in India, these papers lack detailed analyses of women’s work at lower administrative levels, such as districts, in India [13, 17–19]. States in India are often big comprising of a large number of districts. For example, Uttar Pradesh has 75 districts (Integrated Government Online Directory 2024) [20]. Studies have shown that state averages often mask significant district-level heterogeneity in health and development indicators [21–23].
While there is an array of data sources on women’s work in India, there are substantial limitations to all of them. The most common sources are 1) the decennial population and housing census, 2) nationwide quinquennial surveys on employment and unemployment by the National Sample Survey Organization (NSSO) under the Ministry of Statistics and Programme Implementation (MoSPI) of the Government of India, 3) PLFS, and 4) other large-scale household surveys, such as National Family Health Survey (NFHS) and India Human Development Survey (IHDS). In the Indian census, any member of the household (usually the head of the household) responds to the questions related to work of its members, offering data on women’s work at national, state, and district-levels. Unfortunately, the most recent Indian census was conducted in 2011, and is thus severely outdated. The NSSO data on employment and unemployment is available up to only 2011–12 [24]. The nationwide Employment and Unemployment (E&U) surveys have been replaced by the PLFS, conducted by the National Statistical Office (NSO) of MoSPI. The first PLFS was conducted in 2017–18 [25]. While the most recent round of PLFS was conducted in 2022–23, estimates of women’s work can only be derived at the national and state levels. In PLFS, as with the Indian census, any member of the household responds about the work of the other members. IHDS also provides estimates of women’s work in the country. However, the estimates can be derived only at the national level, and the most recent estimates are available for 2011–12. The NFHS, among others, has emerged as a key source of data providing estimates of women’s work only at the national and state levels, with sample size limitations restricting reliable district-level estimates of women’s work. The most recent round of NFHS was completed 2019–21. Women in the age-group 15–49 directly responded to the questions on work in 2019–21 NFHS (henceforth referred to as NFHS-5). A comparison of NFHS-5 estimates of women’s work in past 12 months (based on women age 15–49) with the PLFS estimates (based on women age 15–59) for the same reference period were similar at 31% (NFHS-5) and 32% (PLFS) [8, 26]. NFHS-5 thus offers a recent, and reliable means of estimating prevalences of women’s work in India.
While NFHS-5 has some advantages over other data sources for recent estimates of women’s work in India, small sample sizes at the district-level render the direct estimation of women’s work at district-level less reliable. Small-area estimation (SAE) techniques have emerged as a methodologically sound way to derive reliable estimates of demographic and health indicators at the district level [21, 22, 27]. We, therefore, used area-level SAE techniques to derive reliable estimates of women who worked in the past 12 months, women who worked as self-employed in the past 12 months, and women who earned cash in the past 12 months for the 640 districts of India. We further used spatial models to identify the potential clusters of districts with high and low levels of women’s work. District-level estimates of women’s work may help in appropriately resourcing and targeting districts with unusually low levels of women’s work.
Materials and methods
Data
Our study is based on an analysis of secondary data, the fifth round of the National Family Health Survey (NFHS-5) and the 2011 Indian Population and Housing Census. NFHS is a large-scale, multi-round survey conducted in a representative sample of households throughout India. The NFHS-5 fieldwork was conducted in two phases: phase one from June 17, 2019 to January 30, 2020, and phase two from January 2, 2020 to April 30, 2021. NFHS-5 interviewed a total of 724,115 women aged 15–49, with a 97% response rate [8]. Of these women, only 108,785 were randomly selected for the women’s work module, which included questions related to women’s work in past 12 months and 7 days. The average number of women interviewed per district in the women’s work module was 153, ranging between 14 in Khandwa district and 260 in Nagaur district.
The 2011 Indian Population and Housing Census was conducted between February 9 and 28, 2011, and included information on demography (population characteristics), economic activity, literacy and education, housing & household amenities, urbanization, fertility and mortality, scheduled castes and scheduled tribes, language, religion, migration, disability and many other socio-cultural and demographic data of India [28].
Outcome variables
The three primary study outcome variables are district-level estimates of 1) percentage of women who worked in the past 12 months, 2) percentage of women who were self-employed in the past 12 months, and 3) percentage of women who earned cash in the past 12 months. The three outcome variables were derived from the following questions asked in NFHS-5:
Aside from your own housework, have you done any work in the last 12 months?
Women were coded as ‘worked in the past 12 months’ if they answered ‘yes to the above question.
NFHS-5 further asked
Do you do this work for a member of your family, for someone else, or are you self-employed?
Women were coded as ‘self-employed in the past 12 months’ if they answered yes for self-employment.
NFHS-5 also asked
Are you paid in cash or kind for this work, or are you not paid at all?
Women who received payment in cash only or in cash and kind were coded as earning cash.
Before asking these three questions, NFHS-5 provided a description of the work that women usually do apart from their own housework. For example, some women take up jobs for which they are paid in cash or kind and others sell things, have a small business or work on the family farm or in the family business.
We finally aggregated these indicators at districts-level to estimate the district-level prevalence of women who worked in the past 12 months, women who were self-employed in the past 12 months and women who earned cash in the past 12 months. The denominator for all the three outcome variables is all women age 15–49 interviewed for the women’s work module.
Auxiliary information
In SAE analysis, two types of variables are required: outcome and auxiliary variables. Outcome variables measure the outcome of interest, and are usually derived from survey data. Auxiliary variables capture specific sociodemographic factors, and are required from the entire population. Typically, these auxiliary variables are derived from census or administrative records [29]. We included the following district-level information available in the 2011 Indian Population and Housing Census as auxiliary variables: percentage of households who are scheduled caste/ tribes (SC/ST), percentage of Muslim households, percentage of female literate and gender gap in literacy, percentage of female headed households, mean household size, mean age at marriage for females, spousal age gap, percentage of male child born in last one year, percentage of male migration in past 12 months, percentage of women exposed to media, percentage of rural population, socio-economic percentile, and the state of residence. The choice of auxiliary variables for our models is guided by the social, economic, and demographic determinants of women’s work identified in the existing literature [30–32]. We tested for multicollinearity before including these variables in the SAE models. We also tried different combinations of auxiliary variables and selected that combination of variables which explained maximum variation in the outcome variables. Details about the definitions of auxiliary variables has been given in S1 Text.
Statistical analysis
SAE is a powerful technique for deriving estimates of variables of interest at administrative levels, such as districts in India, for which the sample size is not enough. SAE models combine data from a survey, from which the variables of interest are to be derived, with census or other administrative records that have information for all the residents of the administrative units for deriving small-area estimates. In SAE, the model-based estimates derive strength from the auxiliary variables [33]. Although SAE techniques are well established, its use has been limited in social and public health research [33]. Other details about SAE techniques are given in S2 Text. SAE techniques are classified into two broad types: unit-level and area-level random effects models. Unit-level models are used when auxiliary variables are available at the individual level, whereas area-level models are used when auxiliary variables are only available at aggregate (e.g., district) level [27, 29]. We adopted the area-level SAE approach, as auxiliary data were only available at the district-level.
We first derived district-level estimates of women who worked in the past 12 months, women who worked as self-employed in the past 12 months, and women who earned cash in the past 12 months, directly from the NFHS-5, accounting for survey weights; these estimates are henceforth called as direct survey-based estimates. These direct survey-based estimates were then linked to auxiliary variables using Generalized Linear Mixed Models (GLMM) with logit link functions to derive the model-based district-level estimates of the three outcomes for the 640 districts of India [29, 34]. The GLMM accounted for area-specific random effects, which provided strength to the model-based district-level estimates of the women who worked in the past 12 months, women who worked as self-employed in the past 12 months, and women who earned cash in the past 12 months [33].
Diagnostic measures
We used two types of diagnostic measures (model diagnostic and diagnostic for the small-area estimates) to assess the validity of the fitted GLMM models and the reliability of the model-based district-level estimates of women who worked in the past 12 months, women who worked as self-employed in the past 12 months, and women who earned cash in the past 12 months. The model diagnostic was used to verify the assumptions of the underlying model. Under the GLMM framework, random area specific effects are assumed to follow a normal distribution with mean zero and a constant variance. If the model assumptions are upheld, then the area (district) level residuals are expected to be randomly distributed and not significantly different from the line y = 0 [34].
The diagnostics for small-area estimates are used to validate the reliability of model-based small area estimates obtained from GLMM models. These diagnostics include a) bias diagnostic, b) coefficient of variation (CV), and c) 95% confidence intervals (CIs) of model-based estimates [34]. The bias diagnostic is used to examine the deviation of the model-based district-level estimates from the direct survey estimates to validate the reliability of the model-based district-level estimates. The CVs are used to assess the improvement in the precision of the model-based estimates over the direct survey-based estimates. Estimates with low CV are considered more reliable. The 95% CIs of the model-based estimates and direct survey-based estimates are compared to validate the robustness of the model-based estimates.
Spatial clustering of women’s work
We mapped the estimates of three indicators of women’s work to identify clustering of districts of high and low women’s work. We also mapped the estimates to examine whether the spatial distribution of one variable is associated with the spatial distribution of the other. By doing so, the estimates may be effectively used for policy analysis as the focussed interventions may be directed towards a group of neighbouring districts having low levels of women’s work. Mapping of estimates may help identify and target cluster of districts with low coverage of women’s work even in higher prevalence state contexts. Mapping of estimates may also help in identifying clusters of districts with high-level of women’s work in poor performing states. Identification of such clusters of districts may help in understanding what interventions may work for improving women’s work. We used univariate Local Indicator of Spatial Association (LISA) to obtain geographical clustering of women’s work across the 640 districts of India. We used bivariate LISAs to examine spatial association between district-level estimates of a) women who worked in the past 12 months and the women who worked as self-employed in the past 12 months and b) women who worked in the past 12 months and women who earned cash in the past 12 months.
LISA measures the correlation of neighbourhood values around a specific spatial location. It determines the extent of spatial non-stationery and clustering present in the data.
LISA shows high-high clustering (high prevalent districts surrounded by high prevalent neighbourhood), low-low clustering (low prevalent districts surrounded by low prevalent neighbourhood), and spatial outliers (low–high and high-low clusters). The high-high clusters are also called hot spots representing high prevalent district surrounded by high prevalent districts in terms of women’s work. The low-low clusters are termed as cold spots and characterized by low prevalent districts surrounded by low prevalent districts in terms of women’s work. The districts marked as not significant are those which are surrounded by districts with different patterns of women’s work. Bivariate LISA measures the local correlation between a variable and weighted average of another variable in the neighbourhood. Queen’s contiguity weight matrix was used for estimating LISA. District-level women’s work may be influenced by district-level characteristics, such as percentage poor, percentage of rural population, percentage of households who are SC/ST, etc. [35, 36]. These variables were all estimated using NFHS-5 data, and then used to conduct bivariate LISA between a) district-level estimates of poor households and the three estimates of women’s work, b) district-level rural estimates and the three estimates of women’s work, and c) district-level SC/ST estimates and the three estimates of women’s work. We coded the bottom two wealth quintiles (poorest and poorer) in NFHS-5 to estimate percentage of poor households in the district. District-level women’s work may be influenced by gender norms prevailing in the districts of India. To measure gender norms at the district-level, we estimated district-level India Patriarchy Index (IPI), proposed by Singh et al. 2021, for the 640 districts of India using NFHS-5 [37]. IPI included 5 dimensions—1) domination of men over women, 2) domination of older generation over the younger generation, 3) patrilocality, 4) son preference, and 5) socio-economic domination of men over women [37]. Higher index values indicate more rigid and restrictive gender norms. Please refer to Singh et al. (2021) for details about the IPI [37]. To examine the association between patriarchy and district-level women’s work, we estimated bivariate LISA between district-level estimates of IPI and the three estimates of women’s work.
We also estimated ordinary least square (OLS) regression and spatial error regression (SER) to examine the association between district-level characteristics and the three indicators of women’s work. Both the OLS and SER models examine the effects of various predictors on the outcome variables, however, spatial error models incorporate spatial dependence by the error term. The independent variables included in the regression models are percentage poor, percentage of rural population, percentage of households who are SC/ST, IPI, mean age of women, percentage of married women, percentage of women exposed to media, percentage of female headed households, percentage of female literate, percentage of Muslim households, mean household size, and mean number of children under the age of 5 years present in the household.
The LISA, OLS regression and SER were estimated using GeoDa version 1.12.1.161. SAE was carried out in STATA 16.
Results
Descriptive summary of women’s work
Direct-survey-based summary measures of women’s work at the national level are presented in Table 1. About 68% (n = 73,175) of sampled women lived in rural areas; 31% (n = 33,336) belonged to the scheduled caste or tribe and 81% (n = 87,179) were Hindu. A total of 18.4% (n = 19,904) belonged to the bottom 20% of wealth quintile; whereas 22.6% (n = 24,352) had no formal schooling. About 30.5% (95% CI: 29.9, 30.9) of women worked in the past 12 months, 3.5% (3.3, 3.7) women were self-employed in the past 12 months, and 25.4% (24.9, 25.8) of women earned cash in the past 12 months.
SAE model statistic
Auxiliary variables included in the models explained 74%, 76% and 56% of the variation in women worked in the past 12 months, women earned cash in the past 12 months and women were self-employed in the past 12 months respectively across the districts of India (S1 Table).
Diagnostics.
District-level residuals of women who worked in the past 12 months, women who worked as self-employed in the past 12 months, and women who earned cash in the past 12 months are randomly distributed, confirming the normality assumption. This suggests that the model-based estimates of indicators of women work are robust, with means closer to their expected values (S1 Fig). To compare the level of consistency of model-based estimates and direct survey-based estimates we compare the proximity of the 450 line (y = x) to the fitted regression line for both the estimates. The line of best fit was not significantly different from the line y = x at 5% level for the model-based estimates, indicating the consistency between the model-based and direct survey-based estimates (S2 Fig).
The CVs of the direct survey-based estimates are much larger than the model-based estimates (S3 Fig). Additionally, the fluctuations in the CVs of the direct survey-based estimates are considerably larger compared to those of the model-based estimates. This diagnostic indicates that the model-based estimates are more precise than the direct survey-based estimates. S4 Fig shows the 95% CIs of the direct survey-based estimates and the model-based estimates. The direct survey-based estimates have much wider 95% CIs compared to the model-based estimates, suggesting that the standard errors of the direct survey-based estimates are large and unreliable. The model-based estimates are more robust than the direct survey-based estimates. The diagnostics clearly show the power of the SAE techniques for producing unbiased, consistent, and reliable estimates of women who worked in the past 12 months, women who worked as self-employed in the past 12 months, and women who earned cash in the past 12 months in the 640 districts of India.
The model-based estimates of all the three indicators of women’s work varied considerably across the districts of India (Fig 1). Estimates of women who worked in the past 12 months varied from a minimum of 10.9% (95% CI: 10.4%, 11.5%) in Arwal district of Bihar to a maximum of 61.4% (95% CI: 60.0%, 62.7%) in Bijapur district of Chhattisgarh. The estimated prevalence of women who worked as self-employed in the past 12 months ranged between 0.7% (95% CI: 0.6%, 0.8%) in Gopalganj district of Bihar to 19.8% (95% CI: 18.8%, 20.8%) in east district of Sikkim. The estimated prevalence of women who earned cash in the past 12 months varied from as low as 8.4% (95% CI: 8.2%, 8.7%) in Gopalganj district of Bihar to as high as 55.9% (95% CI: 55.5%, 56.3%) in Nizamabad of Telangana.
Note: The base map can be found at https://globalsolaratlas.info/download/india.
District level estimates of women’s work.
We found considerable intra-state variations in the indicators of women’s work. For example, in Bihar, the state with the lowest coverage of women’s work in the past 12 months (17.5% (95% CI:16.5–18.6)), the district-level estimates varied from 10.9% (95% CI: 10.4–11.5) in Arwal to 27.0% (26.5–27.5) in Jamui. Bihar was followed by Uttar Pradesh in terms of prevalence; the prevalence of women who worked in the past 12 months in Uttar Pradesh was 20.7% (95% CI: 16.7–20.6). The lowest and highest prevalence of women’s work in the past 12 months in Uttar Pradesh were estimated for Azamgarh (14.8%, 95% CI: 14.5–15.0) and Aligarh (29.6%, 95% CI: 29.3–30.0) districts, respectively. The highest percentage of women who worked in the past 12 months was observed in Manipur (49.2%, 95% CI: 45.8–52.7), with prevalence ranging from 40.9% (95% CI: 40.0–41.9) in Imphal East to 61.0% (95% CI: 59.3–62.7) in Chandel. Manipur was followed by Telangana, with a prevalence of women’s work in the past 12 months of 48.2% (95% CI: 46.2–50.3), where the district-level prevalence ranged between 36.0% (95% CI: 35.7–36.3) in Hyderabad to 58.7% (58.3–59.1) in Nizamabad (Table 2).
Sikkim had the highest percentage of women who were self-employed (16.7% (95% CI: 12.9–20.6)). The estimates of women who were self-employed in Sikkim ranged between 7.2% (6.3–8.2) in west district of Sikkim and 19.8% (18.8–20.8) in east district of Sikkim. The state with lowest percentage of women who were self-employed in the past 12 months was Karnataka (1.9%, 95% CI: 1.3–2.6), with prevalence ranging from 1.2% (95% CI: 1.1–1.3) in Uttar Kannada to 3.5% (95% CI: 3.3–3.7) in Kolar. The past year prevalence of self-employed women in Gujarat was 2.0% (95% CI: 1.4–2.5). The lowest and highest prevalence women who were self-employed in the past 12 months in Gujarat were estimated for Amreli (0.7%, 95% CI: 0.6–0.8) and Vadodara (3.4%, 95% CI: 3.3–3.5) districts, respectively.
Intra-state heterogeneity was also observed in the prevalence of women who earned cash in the past 12 months. In Telangana, the state with the highest prevalence of women who earned cash in the past 12 months was 45.1% (95% CI: 42.7–47.5), the prevalence of women who earned cash in the past 12 months ranged between 34.9% (34.6–35.2) in Rangareddy and 55.9% (55.5–56.3) in Nizamabad. Manipur was next to Telangana in terms of women who earned cash in the past 12 months; the district prevalence ranged between 34.8% (95% CI: 33.8–35.8) in Thoubal to 49.9% (48.2–51.6) in Chandel. The lowest percentage of women who earned cash in the past 12 months was in Bihar (12.6%, 95% CI: 11.5–13.8), with prevalence ranging from 8.4% (95% CI: 8.2–8.7) in Gopalganj to 20.7% (95% CI: 20.2–21.1) in Jamui.
Spatial patterns of women’s work in India.
Univariate LISA maps shown in Fig 2 depict considerable spatial heterogeneity in the district-level estimates of women who worked in the past 12 months, women who worked as self-employed in the past 12 months, and women who earned cash in the past 12 months. High-high spatial clusters of women worked in the past 12 months were observed primarily in the districts of Andhra Pradesh, Chhattisgarh, Gujarat, Karnataka, Maharashtra, Telangana, Tamil Nadu and north-eastern states like- Arunachal Pradesh, Manipur, Meghalaya, Mizoram and Nagaland. A few districts of Madhya Pradesh, Jharkhand, Odisha and Rajasthan also formed high-high spatial clusters of women who worked in the past 12 months. In contrast, low-low clusters of districts of women who worked in the past 12 months were located primarily in Bihar, Haryana, Uttar Pradesh, West Bengal, and few districts of Assam, NCT of Delhi, Chandigarh, Jammu & Kashmir, Jharkhand, Odisha, Punjab and Rajasthan.
Note: The base map can be found at https://globalsolaratlas.info/download/india.
High-high spatial clusters of women who worked as self-employed in the past 12 months were found in the districts of Arunachal Pradesh, Himachal Pradesh, Manipur, Meghalaya, Mizoram, Nagaland and Sikkim. In contrast, low-low spatial clusters of women who worked as self-employed in the past 12 months were located in 105 districts primarily from Bihar, Gujarat, Karnataka, Punjab, Uttar Pradesh and few districts of Andhra Pradesh and Maharashtra.
High-high spatial clusters of women who earned cash in the past 12 months were found in the districts of Andhra Pradesh, Chhattisgarh, Karnataka, Maharashtra, Manipur, Puducherry, Tamil Nadu, Telangana, and a few districts of Gujarat, Kerala, Madhya Pradesh, Meghalaya and Odisha. In contrast, low-low spatial clusters of women who earned cash were located in 144 districts primarily from Bihar, Haryana, Jammu & Kashmir, Jharkhand, Rajasthan, Uttar Pradesh, and few districts of Assam, Himachal Pradesh and Uttarakhand.
Fig 3 shows bivariate association between district-level estimates of a) women who worked in the past 12 months and the women who worked as self-employed in the past 12 months and b) women who worked in the past 12 months and women who earned cash in the past 12 months. Districts with high estimated percentage of women who worked in the past 12 months and high percentage of women who worked as self-employed in the same reference period were found in Himachal Pradesh, and states from the north-eastern part of the country—Arunachal Pradesh, Manipur, Meghalaya, Mizoram, Nagaland and Sikkim. In contrast, districts with low estimated percentage of women who worked in the past 12 months and low percentage of women who worked as self-employed in the same reference period were found in Bihar, Punjab and Uttar Pradesh.
Note: The base map can be found at https://globalsolaratlas.info/download/india.
Districts with high estimated percentage of women who worked in the past 12 months and high percentage of women who earned cash in the same reference period were found in Madhya Pradesh, Chhattisgarh, Manipur, Gujarat, Maharashtra, Telangana, Andhra Pradesh, Karnataka and Tamil Nadu. On the contrary, districts with low estimated percentage of women who worked in the past 12 months and low percentage of women who earned cash in the same reference period were found in Bihar, Haryana, Jammu & Kashmir, Jharkhand, Rajasthan and Uttar Pradesh.
We also found cluster of districts in which percentage of women who worked in the past 12 months was high but the percentage of women who worked as self-employed was low. These clusters were found in Gujarat, Andhra Pradesh, Karnataka, three districts of Maharashtra (Parbhani, Hingoli and Jalna), Krishnagiri district of Tamil Nadu, and Nalgonda district of Telangana. In contrast, low percentages of women who worked in the past 12 months but high percentages of women who worked as self-employed in the past 12 months were found in the Hailakandi and Dima Hasao districts of Assam, Himachal Pradesh, Kishtwar district of Jammu & Kashmir, Lunglei district of Mizoram, Wokha district of Nagaland, Garhwal district of Uttarakhand, and central, south and north districts of NCT of Delhi. High percentage of women who worked in the past 12 months was associated with low percentage of women who earned cash in the past 12 months in Almora district of Uttarakhand, Jammu, Rajouri, Punch and Kargil districts of Jammu & Kashmir, and Rajkot and Kachchh districts of Gujarat. Low percentage of women who worked in the past 12 months was associated with high percentage of women who earned cash in the last 12 months in Bangalore district of Karnataka, Idukki, Ernakulum and Wayanad districts of Kerala, Sidhi district of Madhya Pradesh, Kalahandi district of Odisha, Karaikal district of Puducherry, and Chennai district of Tamil Nadu.
Fig 4 shows the bivariate LISA examining spatial association of the selected district-level characteristics with the three indicators of women’s work across the districts of India. Districts that had high poverty and high women’s work (measured in terms of both past year work and past year cash work) were found in Madhya Pradesh, Chhattisgarh, Gujarat, Karnataka, Maharashtra, Odisha, Meghalaya, Arunachal Pradesh, and Manipur. Districts that had low poverty and low women’s work were found in Jammu Kashmir, Punjab, Haryana, few districts of Rajasthan, West Bengal and Uttar Pradesh. We find distinct patterns when we examine spatial association between poverty and women who were self-employed in the past year. Districts from Arunachal Pradesh, Assam, Nagaland, Meghalaya, Manipur and Himachal Pradesh had high poverty and high percentages of women who were self-employed. In contrast, districts from Andhra Pradesh, Gujarat, Karnataka, and southern Punjab showed low levels of poverty and low proportions of women who worked as self-employed in past 12 months.
Note: The base map can be found at https://globalsolaratlas.info/download/india.
The findings further suggest that the districts marked by high levels of rural population and high women’s work and cash earning were concentrated in Madhya Pradesh, Chhattisgarh, Odisha, Arunachal Pradesh, Manipur, Telangana, Andhra Pradesh, Karnataka, Tamil Nadu, from the south region, Maharashtra and Gujarat from the west region. Bivariate LISA between percentage SC/ST and women’s work indicates that high-high clusters of women who worked in the past 12 months and percentage who worked for cash were predominantly found in Chhattisgarh and a few districts of Madhya Pradesh, Maharashtra, Gujarat, Meghalaya, Karnataka, Odisha, Arunachal Pradesh and Manipur. Low-low clusters of SC/ST and past year cash work were found in Uttar Pradesh, Bihar and selected districts of Rajasthan, West Bengal, Haryana, Jharkhand, and Jammu & Kashmir. High-high clusters of districts having higher percentages of SC/ST and self-employed predominantly belonged to North-eastern states and Himachal Pradesh. Low-low clusters of SC/ST and self-employment belonged to the districts from Bihar, eastern Uttar Pradesh, Gujarat, Karnataka and a few districts of Andhra Pradesh.
Interesting patterns emerge when we examine the bivariate LISA between IPI and women who worked in the past 12 months. Districts from Rajasthan, Haryana, Uttar Pradesh, and Bihar–states with high levels of patriarchy—were also the districts where fewer percentage of women worked in the past 12 months. In contrast, districts with low levels of patriarchy and high levels of women’s work were predominantly located in Tamil Nadu, Andhra Pradesh, Telangana, Maharashtra, Chhattisgarh, Meghalaya, Arunachal Pradesh and Manipur. A few districts of Madhya Pradesh bordering Maharashtra and a few districts of Odisha bordering Andhra Pradesh and Chhattisgarh formed a low patriarchy / high women’s work cluster.
When we look at women who worked for cash in the past 12 months, districts from Rajasthan, Haryana, Uttar Pradesh and Bihar formed high patriarchy and low women’s paid work clusters. In contrast, districts with low levels of patriarchy and high levels of women’s paid work were predominantly located in Tamil Nadu, Andhra Pradesh, Telangana, Karnataka, Maharashtra, Chhattisgarh, Meghalaya, Arunachal Pradesh and Manipur.
There were districts that, despite having high patriarchy, had high percentage of women who worked and earned cash in the past 12 months. Such districts were predominantly located in Maharashtra and Karnataka.
Results of OLS and SER are shown in Table 3. While mean age of women, percentage of women exposed to mass media, percentage of rural population and percentage of households who are SC/ST were positively associated with percentage of women who worked in the past 12 months in the OLS, mean number of children under the age of 5 years, percentage of females literate, percentage of Muslim households and percentage of poor households were negatively associated. India patriarchy index was negatively associated with women’s work in the past 12 months. In the SER model, percentage of women exposed to mass media, mean number of children under the age of 5 years and percentage poor households lost its statistical significance. Percentage of married women was positively associated with women’s work in the past 12 months in the SER. When it comes to women’s self-employment, percentage of women exposed to media, mean age of women, mean number of children under age of 5 years and percentage of SC/ST households were positively associated with self-employment. India patriarchy index and percentage of Muslim households were negatively associated with self-employment. In the SER, percentage of female headed households, percentage of married women and mean age of women and percentage of household who are SC/ST were positively associated with percentage of women who worked for cash. India patriarchy index, percentage of females literate and percentage of Muslim households were negatively associated with earning cash. The lag coefficient (λ) was statistically significant for all the three outcomes, indicating that the relationship between the three indicators of women’s work and selected district-level characteristics at the macro-level (districts) may be misleading if spatial clustering is ignored.
Discussion
Our study is the first of its kind to provide district level estimates of women’s work in the past 12 months across 640 districts of India. We used area-level small-area estimation models, NFHS-5 survey data, and 2011 Indian Census data to generate these estimates. This research fills a substantial gap in the estimation of women’s work in India, given that estimates of women’s work-related indicators obtained directly from NFHS surveys are not reliable due to small sample sizes. Several diagnostic tests demonstrated the utility of SAE for estimating district-level estimates of women who worked in the past 12 months, women who worked as self-employed in the past 12 months and women who earned cash in the past 12 months in India. Reliable estimates of women’s work at the district-level can help policymakers and program managers to formulate and / or target relevant appropriate policies at more local levels. Unlike other surveys such as, National Sample Survey (NSS), PLFS and IHDS in India, which often provide data on women’s work at the national and state/province level, with no or little data at the local levels, NFHS-5 provides a unique opportunity to estimate women’s work at district-level using SAE. Moreover, every surveyed woman directly responded to the questions related to women’s work in NFHS-5. In contrast, any member of the household reports about the work done by other members of the household in all the labour force surveys conducted in India. Therefore, the likelihood of these surveys failing to capture women’s work is higher compared with NFHS-5. NFHS-5 also provided a description of the work that women usually do apart from their own housework. For example, NFHS-5 informed women that some women take up jobs for which they are paid in cash or kind and others sell things, have a small business or work on the family farm or in the family business. Then the NFHS-5 asked women whether they have done any of these things or any other work. Given the way the questions related to women’s work were designed in NFHS-5, NFHS-5 is less likely to miss unpaid work or self-employment. For the reasons mentioned above, our analyses demonstrate how a timely household survey, which is not a typical labour force survey, can be effectively used to provide reliable estimates of women’s work at more local levels.
In many low- and middle- income countries (LMICs), the primary source of data on economic activity at local levels is the population census, which is usually conducted at long intervals. Other specialized labour force surveys may exist, but may not provide estimates at more local levels. In absence of local level data, many LMICs rely on household survey data. This analysis offers an example of how commonly available household survey data (e.g., the NFHS or the Indian implementation of the Demographic and Health Surveys) can be combined with national population census data to estimate key metrics such as women work force participation, intimate partner violence, and the digital and financial inclusion of women that have been limited to more aggregate assessments.
Low levels of women’s work persist in India. Only 31% of women worked in the past 12 months, 25% of women earned cash in the past 12 months and only 4% of women were self-employed in the past 12 months. We observed significant geographical disparities at the district level. There are several districts within states where women’s work was very low while women’s work in the state as a whole is high. For instance, Madhya Pradesh, where 35% of women worked in the past year, had the highest variation across the districts; the estimated percentage of women who worked in the past 12 months ranged from 24% in Sidhi district to 61% in Alirajpur district. This sub-state heterogeneity underscores the inadequacy of state-level estimates in understanding and responding to the stagnation of women’s work in India.
A key contribution of the present study is to identify low-low and high-high clusters of women’s work. Identifying low-low clusters of districts is important for strategic targeted policies to increase women’s work in such clusters of districts. Identifying high-high clusters of districts offers valuable insights for best practices and interventions to improve women’s work in low-low clusters of districts. The study found that women’s work and cash earnings were lowest in the districts of central, eastern and northern parts of India. Districts from Uttar Pradesh (from central part) and Bihar (from the eastern part) have the lowest workforce participation among women. One of the possible reasons for low level of women’s work in Bihar and Uttar Pradesh may be natural disasters like flood. In these states, agriculture is the primary sector where women work [38]. Studies show that Bihar is the top-ranking state whose State Gross Domestic Product (SDGP) was affected by floods during 1983–2011. Likewise, Uttar Pradesh also ranked in top five states whose SGDP was affected by floods during the same period. In the context of frequent exposure to flood disasters, female farmers are more likely than male farmers to face unemployment because of dependency on agricultural work and lack of alternate employment opportunities due to resource and mobility restriction [39, 40]. Districts of Haryana and Punjab, from the northern region also show low levels of women’s work in the past year. Both states are known for their predominately agrarian economy known for intensive mechanisation; such economies require less labour [41]. In addition, these states are marked by high patriarchy [37]. In our spatial models, high patriarchy was associated with lower levels of women’s work, women’s self-employment, and women earning cash in the past 12 months.
In contrast, districts from southern-, western-, north-eastern- parts of India and Chhattisgarh depict the highest level of women’s work and cash earnings. This finding is in line with the other studies which also shows high workforce participation in these region [42]. In Andhra Pradesh and Telangana, women’s participation in agricultural activity is high compared to male worker [43]. Unlike in north India, characterized by wheat cultivation, large average landholding size, mechanized agriculture, and more limited involvement of women in agriculture, in southern India, agriculture is characterized by more manual labour associated with wet-rice cultivation than cultivation of wheat and other crops, leading to high involvement of women in agriculture [44]. Malhotra, Vanneman and Kishor (1995) used female share of labour force and area under rice cultivation (on the assumption that wet-rice cultivation involves substantial input from women than cultivation of wheat and other crops) as a proxy measure of women’s work [45]. Districts from Chhattisgarh, a less developed state of India, show high levels of women engaging in paid work, plausibly explained by the higher participation of women in MGNREGA scheme in this state [46]. Moreover, the high percentage of SC/ST population in Chhattisgarh could also explain high women’s work in Chhattisgarh. Chhattisgarh has the highest percentage of SC/ST population after the north-eastern states of Mizoram, Meghalaya, Nagaland and Arunachal Pradesh. Studies show that SC/ST women are more likely to work compared with women belonging to other social groups, such as other backward classes (OBC) and those who do not belong to SC/ST or OBC [36]. Women who do not belong to SC/ST are more likely to face restrictive norms than SC/ST women [47]. Our regression results also confirm positive associations between percentage of SC/ST households and all three indicators of women’s work. The Chhattisgarh government has also launched special programmes, such as Naruva-Garuva-Ghuruva-Badi to improve rural livelihood in the state. Naruva-Garuva-Ghuruva-Badi was launched in 2019 with a focus on water management, composting for soil health, animal husbandry and sustainable agriculture on backyard kitchen gardens [48]. Such programmes are likely to improve opportunity for women’s work in the state. Districts from the southern, western, north-eastern parts of India generally show lower levels of gender inequality compared with districts from the northern and central parts of India [37].
Further, our study shows high-high clusters of self-employed women in districts of Himachal Pradesh and north-eastern states, such as Arunachal Pradesh, Meghalaya, Manipur, Mizoram and Sikkim; all these states are mountainous. Our findings are consistent with the findings of prior studies [49, 50]. Studies have shown that women play a critical role in mountain societies as a significant proportion of workforce in agriculture [51] and managing and sustaining their natural resources and environment [52]. Samal (1993) argued that hill agriculture is totally dependent on women [53]. Recent research conducted in Uttarakhand and Manipur also argued that farming in the hilly areas has always been the domain of women [54, 55]. Women in hilly areas are also involved in a number of other activities, such as forest-based handicrafts, agro-forestry, livestock production, horticulture, and post-harvest activities [53–55]. Apart from being hilly, these states, except for Himachal Pradesh, are tribal dominated. Percent ST ranges from 33% in Sikkim to 94% in Mizoram [28]. Additionally, men from the hilly districts often migrate to the plain districts in search of better job opportunities and livelihood [56, 57].
Women from Kerala, despite high levels of schooling, show low levels of women’s work, including self-employment. High unemployment rates for women in Kerala may largely be due to constraints in skill utilisation, higher wage rates and massive inflow of remittances from migrants to the Gulf countries [58–60]. Past studies have also found U-shaped relationship between household income and female work participation in India [61–64]. When household income is low and agriculture activity dominates, there is high participation of women in the labour force. As household income rises women start withdrawing from the workforce. This withdrawal can be attributed to the ‘income effect’. But as the female education level increases further, the value of substitution effect (for highly educated women it would be costly to give up a paid job and take up an unpaid household work) also increases. As the substitution effect becomes greater than the initial income effect, women start moving back to the paid labour force. The evidence on this hypothesis is mixed and varies with geographical contexts [1, 65]. There is no clear evidence on the U-shaped relationship between economic development and women’s work in Kerala [58, 66].
We also found spatial outliers. For example, bivariate LISA between IPI and women’s work in past 12-months showed several outliers, i.e., districts with high patriarchy and high level of women’s work in the past 12-months. These outliers are primarily from Maharashtra and Karnataka. There were 13 districts from Maharashtra that had high women’s work despite having high levels of patriarchy. Our study is not powered enough to explain why a high proportion of women in these districts are working despite high levels of patriarchy. A study of such spatial outliers may shed light on what factors mitigate the effect of patriarchy when it comes to women’s work. Future studies may consider qualitative methods to research spatial outliers.
Female literacy was negatively associated with two of the three indicators of women’s work in the spatial regression models. A huge majority of women in India are employed in the agricultural sector [67]. With increase in literacy, women tend to withdraw from agriculture related activities and start looking for opportunity in the organized sectors, such as manufacturing, non-manufacturing (mainly construction) and service sectors (modern services) [67]. Chatterjee et al. (2018) also find out that women with increasing educational attainment are more likely to move from family farm or as wage labour to salaried positions [18]. The study further showed that if all or most available jobs were salaried, we could see a positive relationship between women’s education and employment. However, such job opportunities are limited in India [18, 68]. Micro-level studies for India have indicated that women withdraw from labour force with increase in women’s schooling and improvement in the status of the families [17, 18, 68].
Another finding that deserves mention is the negative association between the proportion of district residents who are Muslim and all three indicators of women’s work. Past studies have shown that Muslim women in India face educational, economic, social, and political marginalization, contributing to their low workforce participation in India [69–71]. Another study from India noted that Muslim women are often concentrated in areas of low economic activity due to which their participation in work is limited [72].
A key strength of our study is the use of the most recent population representative household survey, conducted in India in 2019–21, with district-level granularity. None of the other sources of data on women’s work in India offer such granularity. NFHS-5 provided us a unique opportunity to apply area-level SAE techniques to derive reliable estimates of the three indicators of women’s work for the 640 districts of India. Our analysis shows that direct survey estimates for many districts are unreliable due to large standard errors. These are the districts with very small sample sizes. Despite small sample sizes, the model-based estimates of women’s work are reasonable and representative for these districts. The direct and model-based 95% CIs are very close for the districts with reasonably large sample sizes. Another strength of our study is the use of spatial regression models to identify the factors associated with three indicators of women’s work at the district-level. Classical statistical models, such as OLS, often ignore spatial autocorrelation present in the data. Ignoring spatial autocorrelation may lead to underestimation of standard errors resulting in variables that should be statistically insignificant otherwise becoming significant. Spatial models also allowed us to identify clusters that require special attention. These spatial clusters may not always be spread within a state but may also be spread across the borders of neighbouring states. We have several examples where spatial clusters consisted of bordering districts of two or three neighbouring states. Such examples call for policies and programmes implemented jointly by group of neighbouring states. Our study also offers a convenient way to track progress in women’s work at regular intervals at lower administrative levels (such as a district), which cannot be easily done with census data that are collected at an interval of 10 years. As our work is based on NFHS-5 (Demographic and Health Survey (DHS) equivalent in India), our study can be easily replicated in other LMICs where DHS surveys are regularly conducted.
Our study has a few limitations. First, COVID-19 related lockdowns disrupted the collection of data in NFHS-5. Approximately 70% of the fieldwork in NFHS-5 was completed before the COVID-19 pandemic hit India and the rest was completed after the COVID-19 related lockdowns were lifted. This should not be a problem for our estimations as our three indicators of women’s work are based on a 12-months recall period, i.e. whether the woman has worked in the past 12 months. There are no reliable means of estimating the effects of the COVID-19 pandemic on the indicators of women’s work at the district level with NFHS-5 data. While we were able to estimate indicators of women’s work, we could not examine other aspects, such as women’s work-related decision-making, women’s entrepreneurship, etc., due to unavailability of such information in NFHS-5. Despite these limitations, our estimates of women’s work help us in understanding the geography of women’s work in India. Our estimates of women’s work may be used for future research including association of women’s work with intimate partner violence, women’s empowerment, women’s ownership of assets, demographic behaviour, etc.
Conclusions
This study demonstrated the power of SAE technique to generate robust estimates of women’s work at the district level in India. Combining large scale survey data like NFHS-5 with 2011 population census allow us to produce reliable districts level estimates of women’s work and documented high intrastate variation in all three outcomes–percentage of women who worked in the past 12 months, women who earned cash in the past 12 months and women who were self-employed in the past 12 months. We also conducted spatial analyses and identified the cluster of districts with low and high women’s work. Our study indicates a clear north-south divide in indicators of women’s work in India. Our study demonstrates that women’s work varies substantially at the subnational level and having only state-level estimates may not be enough to address the issue of low women’s work in India. The insights from our current study may help in the formulation and implementation of policies that increase the returns on women’s work in India.
Supporting information
S1 Table. Percentage variation explained by auxiliary variables for each outcome variable (Results based on OLS).
https://doi.org/10.1371/journal.pone.0317783.s003
(DOCX)
S1 Fig. Model diagnostic plot showing the distribution of the district level residuals for women’s work in the past 12 months in India, 2019–21.
https://doi.org/10.1371/journal.pone.0317783.s004
(DOCX)
S2 Fig. Plots comparing the ordinary least square regression line (dash line) and y = x (solid line), India, 2019–21.
https://doi.org/10.1371/journal.pone.0317783.s005
(DOCX)
S3 Fig. District wise coefficient of variation for women’s work in the past 12 months in India, 2019–21.
https://doi.org/10.1371/journal.pone.0317783.s006
(DOCX)
S4 Fig. District wise 95% CI for women’s work in the past 12 months in India, 2019–21.
https://doi.org/10.1371/journal.pone.0317783.s007
(DOCX)
References
- 1. Sundari S. Economic Development and Female Labour Force Participation in Tamil Nadu: A Test of the ‘U’Shape Hypothesis. Economic Affairs. 2019;64(4):695–702.
- 2. Jose S. Women’s paid work and well-being in Rajasthan. Economic and Political Weekly. 2012 Nov 10;48–55.
- 3. Gill K, Pande R, Malhotra A. Women deliver for development. The Lancet, 2007 Oct 13;370 (9595):1347–1357. pmid:17933650
- 4.
World Bank. Labor force, female (% of total labor force)-India. Washington DC, World Bank; 2022. Available https://data.worldbank.org/indicator/SL.TLF.TOTL.FE.ZS?locations=IN.
- 5. India Spend Report. Rising Income Levels, Stability Linked to Declining Female Workforce Participation in India; 2016. Available https://archive.indiaspend.com/cover-story/rising-income-levelsstbility-linked-to-declining-female-workforce%20participation-in-india-84594.
- 6.
World Bank. Literacy rate, adult female (% of females ages 15 and above). Washington DC, World Bank; 2023. Available https://data.worldbank.org/indicator/SE.ADT.LITR.FE.ZS?end=2022&locations=IN&start=1989
- 7. International Labor Organization. India wage report: Wage policies for decent work and inclusive growth; 2018. Available https://www.ilo.org/sites/default/files/wcmsp5/groups/public/@asia/@ro-bangkok/@sro-new_delhi/documents/publication/wcms_638305.pdf.
- 8.
IIPS, ICF. National Family Health Survey (NFHS-5), 2019–21: India. Mumbai: International Institute for Population Sciences; 2021. Available http://rchiips.org/nfhs/NFHS-5Reports/NFHS-5_INDIA_REPORT.pdf.
- 9.
Bank World. GDP growth (annual %). Washington DC, World Bank; 2015.Available https://data.worldbank.org/indicator/NY.GDP.MKTP.KD.ZG?end=2022&locations=IN&start=1990.
- 10.
Government of India. Periodic Labour Force Survey (PLFS), 2022–23. National Sample Survey Office, Ministry of Statistics and Programme Implementation; 2023.
- 11. Jose S. Women, paid work and empowerment in India: A review of evidence and issues. Centre for Women Development Studies.2007.
- 12. Elson D. Labour Markets as Gendered Institutions: Equality, Efficiency and Empowerment Issues. World Development 1999 Mar 1;27(3): 611–627.
- 13. Desai S, Joshi O. The paradox of declining female work participation in an era of economic growth. The Indian Journal of Labour Economics. 2019 March 1; 62(1):55–71. pmid:32831498
- 14. Jose S. Paid employment and female autonomy in India: Issues and evidence. The Indian Journal of Labour Economics. 2008; 51(3): 397–408.
- 15. Lahoti R, Swaminathan H. Economic development and women’s labor force participation in India. Feminist Economics. 2016 April 2;22(2):168–195.
- 16.
World Development Report. Gender Equality and Development. Washington DC, World Bank Group; 2012. Available http://documents.worldbank.org/curated/en/492221468136792185/Main-report
- 17.
Deshpande A, Singh J. Dropping out, being pushed out or can’t get in? Decoding declining labour force participation of Indian women. Institute of Labor economics. 2021; IZA Discussion Papers no 14639. https://doi.org/10.2139/ssrn.3905074
- 18. Chatterjee E, Desai S, Vanneman R. Indian Paradox: Rising Education, Declining Women’s Employment. Demographic research. 2018 Jan; 38(31): 855–878. pmid:30899196
- 19. Deshpande A, Kabeer N. Norms that matter: Exploring the distribution of women’s work between income generation, expenditure-saving and unpaid domestic responsibilities in India. World Development. 2024 Feb 1; 174: 106435.
- 20.
Integrated Government Online Directory. 2024. Districts–Uttar Pradesh. Accessed from https://igod.gov.in/sg/UP/E042/organizations on November 14 2024.
- 21. Srivastava S, Kumar K, McDougal L, Kannaujiya AK, Sikarwar A, Raj A, et al. Spatial Heterogeneity in Intimate Partner Violence Across the 640 Districts of India: A Secondary Analysis of a Cross-Sectional, Population-Based Survey by Use of Model-Based Small-Area Estimation. The Lancet Global Health.2023 Oct 1;11(10): e1587–97. pmid:37734802
- 22. Singh A, Upadhyay AK, Kumar K, Singh A, Amoako Johnson F, Padmadas SS. Spatial heterogeneity in son preference across India’s 640 districts: an application of small-area estimation. Demographic Research. 2022 Jul 1; 47(26): 793–842.
- 23. Khan J, Shil A, Prakash R. Exploring the spatial heterogeneity in different doses of vaccination coverage in India. PloS one.2018 Nov 28; 13(11): e0207209. pmid:30485291
- 24. Government of India. Employment and unemployment situation in India, 2011–12. National Sample Survey Office, Ministry of Statistics and Programme Implementation; 2014. Available https://mospi.gov.in/sites/default/files/publication_reports/nss_rep_563_13mar15.pdf.
- 25. Government of India. Periodic Labour Force Survey (PLFS), 2017–18. National Sample Survey Office, Ministry of Statistics and Programme Implementation; 2019. Available https://mospi.gov.in/sites/default/files/publication_reports/Annual%20Report%2C%20PLFS%202017-18_31052019.pdf.
- 26. Government of India. Periodic Labour Force Survey (PLFS), 2021–22. National Sample Survey Office, Ministry of Statistics and Programme Implementation; 2022. Available https://pib.gov.in/PressReleasePage.aspx?PRID=1902104.
- 27. Srivastava S, Chandra H, Singh SK, Upadhyay AK. Mapping changes in district level prevalence of childhood stunting in India 1998–2016: An application of small area estimation techniques. SSM-Population Health. 2021 June 1; 14: 100748. pmid:33997239
- 28.
Office of the Registrar General, India. Census of India Website; 2011. Available https://censusindia.gov.in/census.website/
- 29. Amoako Johnson F, Chandra H, Brown JJ, Padmadas SS. Estimating district-level births attended by skilled attendants in Ghana using demographic health survey and census data: An application of small area estimation technique. Journal of Official Statistics. 2010; 26(2): 341–359.
- 30. Govindapuram S, Bhupatiraju S, Sirohi RA. Determinants of women’s financial inclusion: evidence from India. Annals of Public and Cooperative Economics. 2023 Mar; 94(1): 131–158.
- 31. Sarma M, Pais J. Financial inclusion and development. Journal of international development. 2011 Jul; 23(5): 613–628.
- 32. Sanderson A, Mutandwa L, Le Roux P. A review of determinants of financial inclusion. International Journal of Economics and Financial Issues. 2018; 8(3): 1–8. https://www.econjournals.com/index.php/ijefi/article/view/5458.
- 33. Rao JN, Molina I. Small area estimation. John Wiley & Sons; 2015 Aug 24.
- 34. Pfeffermann D, Sverchkov M. Small-area estimation under informative probability sampling of areas and within the selected areas. Journal of the American Statistical Association. 2007 Dec 1; 102 (480): 1427–1439.
- 35.
Mitra A, Okada A. Labour market participation in India: A region- and gender-specific study. Springer Singapore; 2018: 1–51. https://doi.org/10.1007/978-981-10-7143-0_1
- 36. Srivastava N, Srivastava R. Women, work, and employment outcomes in rural India. Economic and political weekly. 2010 Jul 10; 45(28): 49–63.
- 37. Singh A, Chokhandre P, Singh AK, Barker KM, Kumar K, McDougal L, et al. Development of the India patriarchy index: Validation and testing of temporal and spatial patterning. Social Indicators Research. 2021; 159: 351–377.
- 38.
International Labour Organization. Women at work: Trends 2016. International Labour Office- Geneva: ILO,2016. Available http://www.ilo.org/gender/Informationresources/Publications/WCMS_457317/lang—en/index.htm.
- 39. Erman A, De Vries Robbe SA, Thies SF, Kabir K, Maruo M. Gender dimensions of disaster risk and resilience: Existing evidence. Washington DC. 2021. Available http://hdl.handle.net/10986/35202
- 40. Chowdhury JR, Parida Y, Goel PA. Does inequality-adjusted human development reduce the impact of natural disasters? A gendered perspective. World Development 2021 May 1; 141:105394.
- 41.
Afridi F, Bishnu M, Mahajan K. Gendering technological change: Evidence from agricultural mechanization. Institute of Labor Economics.2020; IZA Discussion Paper No. 13712.
- 42.
Fernandez C, Puri H. A Statistical Portrait of the Indian Female Labor Force. Philippines: Asian Development Bank. 2023. https://doi.org/10.56506/BDXR3681
- 43. Rao S. Work and Empowerment: Women and Agriculture in South India. The Journal of Development Studies. 2011 Feb 1; 47(2): 294–315. pmid:21506302
- 44. Bardhan PK. "On life and death questions," Economic and Political Weekly. 1974 Aug 1;9 (32/34):1293–1304.
- 45. Malhotra A, Vanneman R, Kishor S. Fertility, dimensions of patriarchy, and development in India. Population and Development Review.1995 Jun 1; 21(2): 281–305.
- 46.
Government of India. Participation of Women Worker in Mahatma Gandhi National Rural Employment Guarantee Scheme. Ministry of rural development; 2023. Available pib.gov.in/PressReleaseIframePage.aspx?PRID=1942781
(2023).
- 47. Field E, Jayachandran S, Pande R. “Do Traditional Institutions Constrain Female Entrepreneurship? A Field Experiment on Business Training in India,” American Economic Review. 2010 May 1; 100(2): 125–29.
- 48. Bej D, Karmakar S, Verma CS. Narwa, Garwa, Ghurwa and Bari (NGGB) Program Development of Mardapotti Gram Panchayat, Kanker District: A Remote Sensing & GIS Approach. International Journal of Science and Research.2022;11(4): 589–593.
- 49. Government of India. Periodic Labour Force Survey (PLFS), 2020–21. National Sample Survey Office, Ministry of Statistics and Programme Implementation; 2021. Available https://pib.gov.in/PressReleaseIframePage.aspx?PRID=1833855.
- 50. Tripathi S. Determinants of employment and unemployment situation in India with Special reference to North Eastern states of India. 2016; Munich Personal RePEc Archive, 2116: 0–33. Available https://mpra.ub.uni-muenchen.de/71469/
- 51. United Nations. Food and Agriculture Organization. The State of Food and Agriculture. Women in Agriculture: Closing the Gender Gap in Development. Rome; 2011. Available 12C5112E3B7A2EDFC125784C0038AE91-Full_Report.pdf.
- 52. Singh S. Women, Environment and Sustainable Development: A Case Study of Khul Gad Micro Watershed of Kumoun Himalaya. Space and Culture India.2014 March 1; 1(3): 53–64.
- 53. Samal PK. The Status of Women in Central Himalaya: A Cultural Interpretation. Man in India. 1993; 73(1):87–95.
- 54. Joshi CS, Chaudhary T. Women Participation in agriculture sector in hilly regions of Uttarakhand. India Journal of Hill Farming. 2021; 34(1):109–114.
- 55. Devi SR, Singh LK. Women’s Role in Agriculture and Allied Fields in Manipur. Indian Journal of Hill Farming. 2015; 28(2): 154–157. Available https://epubs.icar.org.in/index.php/IJHF/article/view/54820
- 56.
Awasthi IC. Livelihood Diversities in Mountain Economy, Constraints and Opportunities. New Delhi: Concept Publishing Company.2012. Available https://lccn.loc.gov/2012323904
- 57.
International Centre for Integrated Mountain Development, (ICIMOD). Labour Migration and Remittances in Uttarakhand. Kathmandu; 2011. Availabe attachment_735.pdf.
- 58. Mathew SS. Falling female labour force participation in Kerala: Empirical evidence of discouragement? International Labour Review. 2015 Dec; 154(4): 497–518.
- 59. Zachariah KC, Rajan SI. From Kerala to Kerala via the Gulf: emigration experiences of return emigrants. In: Rajan S. I. (ed.). India Migration Report. 2015; 269–280.
- 60. Peter B, Sanghvi S, Narendran V. Inclusion of interstate migrant workers in Kerala and lessons for India. The Indian Journal of Labour Economics. 2020 Dec; 63(4): 1065–1086. pmid:33204053
- 61.
Goldin C. The U-Shaped Female labour Force Function in Economic Development and Economic History. In Schultz T. P., Investment in women’s Human Capital and Economic Development. Chicago. University of Chicago Press.1995; 61–90. https://doi.org/10.3386/w4707
- 62. Abraham V. Missing labour or consistent “de-feminisation”? Economic and Political Weekly. 2013 Aug 3:48(31): 99–108.
- 63. Klasen S, Pieters J. What explains the stagnation of female labor force participation in urban India?. The World Bank Economic Review. 2015; 29(3): 449–478.
- 64. Afridi F, Dinkelman T, Mahajan K. Why are fewer married women joining the work force in rural India? A decomposition analysis over two decades. Journal of Population Economics. 2018 July; 31(3):783–818.
- 65. Gaddis I, Klasen S. Economic development, structural change, and women’s labor force participation: A reexamination of the feminization U hypothesis. Journal of population economics. 2014 Jul; 27:639–681.
- 66. Hajong VR, Kakarlapudi KK. Recent trends in female labour force participation in Kerala: A comparative analysis. Kerala Economy.2023;4(2): 24–37.
- 67. Mehrotra S, Parida JK. Why is the labour force participation of women declining in India?. World Development. 2017 Oct 1; 98: 360–380.
- 68.
Das M, Desai S. Why are educated women less likely to be employed in India? Testing competing hypotheses. Washington, DC. Social Protection, World Bank. 2003 May 1; Discussion Paper No. 313. Available https://documents1.worldbank.org/curated/en/616871468756558333/text/278680SP0discussion00313.txt
- 69. Noorjahan . Educational backwardness of women among Muslim minority: an analytical study. International Journal of Advanced Research. 2023;11(2):1182–1191.
- 70. Sharma M. Social and political status of Muslim women in India: a general analysis. Research Review. International Journal of Multidisciplinary.2023 Jan 16; 8(1): 139–144.
- 71. Sarkar S, Sahoo S, Klasen S. Employment transitions of women in India: A panel analysis. World Development. 2019 Mar 1; 115: 291–309.
- 72. Alvi MF. Caste, religion and the labor force participation of women: evidence from India. Review of Social Economy. 2023 Jun 28:1–28.