Disparities Within: Occupational Inequality among Indian Women across Socio-demographic Groups

Background
Studies on gender inequality in India overwhelmingly emphasize male–female gaps, overlooking disparities within women. Caste, religion, location, and education not only shape whether women work, but also what kinds of jobs they access. This paper presents the first systematic investigation of within-women occupational inequality using nationally representative data spanning three decades.
Data
The analysis uses five rounds of the National Family Health Survey (NFHS, 1992–2020), restricting the sample to women aged 15 and above not currently enrolled in school or college. Ten occupational categories are harmonized across rounds and further collapsed into three groups: (i) Not working, (ii) Desired/formal jobs (administrative/managerial, professional/technical, clerical/sales), and (iii) Less desired/informal jobs (cultivator, agricultural and non-agricultural labor, skilled manual, service, residual).
Methods
The study combines descriptive, inequality, and econometric tools:
– Descriptive distributions track women’s occupational shares by caste, religion, location, region, education, and wealth.
– Mutual Information (MI) indices and their decomposition quantify occupational inequality, separating within- and between-group contributions.
– Index of Representation (IOR) compares a group’s share in desirable jobs with its population share, highlighting over- and under-representation.
– Multinomial logistic regressions estimate the conditional probability of women being in desired or informal jobs (relative to non-work), controlling for education, household assets, media exposure, time poverty, and spousal characteristics.
Results
– Descriptive trends show persistent concentration of women in agriculture and informal work, with formal jobs limited to a privileged minority.
– MI analysis reveals that between-group inequality—especially caste, religion, and rural–urban divides—explains a substantial share of occupational entropy.
– IOR values highlight underrepresentation of Muslims, disadvantaged castes, and rural women in desirable jobs, alongside overrepresentation of educated and wealthier women.
– Multinomial models confirm education as the strongest predictor of entry into formal work: women with higher education are nearly twelve times more likely to access desirable jobs than those with none. Female household headship and spouse’s formal employment also increase probabilities. In contrast, Muslim identity, low caste status, and rural residence significantly reduce the likelihood of formal work.

Contribution
This study is the first to apply MI indices, representation measures, and multinomial models to the question of within-women occupational inequality in India. By moving beyond aggregate male–female comparisons, it uncovers how structural barriers and social identities segment women’s labor market participation. Findings hold critical implications for policies seeking inclusive, identity-sensitive labor market reforms.”