ISSN (Print) - 0012-9976 | ISSN (Online) - 2349-8846

A+| A| A-

An Analysis with the Labour Force Survey in India

Occupation, Earning, and Gender

This study analyses the employment distribution of the working-age women by occupations across their activities in usual principal status in the Periodic Labour Force Survey for 2017–18 by taking into account the household-specific factors and workers’ personal characteristics by using a multinomial logit model. The study infers that gender differences in returns to schooling are in favour of female workers, but they earned less than male workers in almost every occupation and employment status. The effect of education is stronger in selecting high-paying jobs.


Gender differences in occupation have widely been considered as one of the important factors to contribute to earning differences by gender (Groshen 1991; Mac­pherson and Hirsch 1995; Blau et al 2013). Conventionally, supply-side analysis of the labour market uses differences in human capital accumulation between men and women as the major explanatory factor for gender differences in job choices and earnings (both wage and non-wage earnings). While the gender gap in educational attainment has been reducing over time and today, in many cases, the education gap has reversed in favour of women, gender wage gap has increased in different sectors. Thus, earning differences by gender cannot be explained fully by the differences in human capital variables like education and experience between men and women. These findings highlight the significance of understanding the causes and consequences of differences in employment distributions by gender as an area of research.

Most of the studies available in the early literature have att­ributed mainly the pattern of occupational segregation by gender by looking into the differences in human capital accumulation and examine the incidence of discrimination across occupations. In recent literature, gender differences in preference for job attributes has been suggested as a potential explanation for gender differences in occupational choice and earnings. Recent studies have also used gender differences in cognitive skills in analysing the differences in occupations and wages between men and women (Bertrand 2011). It is obs­erved that women have a comparative advantage in cognitive relative to manual skills, and these facts are likely to explain why women are absorbed more in cognitive-intensive jobs (Welch 2000).

This study examines the role of occupational segregation in explaining the gender earning gap in India. We measure occupational segregation with unit-level data from the Periodic Labour Force Survey (PLFS) 2017–18 and examine its implications in analysing gender differences in occupation and wage as well as non-wage earnings in India. The study finds out the possible sources of gender differences in occupation and earnings by estimating the relationship between occupational characteristics and occupational choice in probabilistic sense and occupation-specific earning equations.

In India, women’s participation in wage employment has been declining even during the high growth regime. Structural reforms towards more services sector growth and increase in tertiary education have failed to push women into the labour force, and female participation rates in the labour market dec­lined substantially. On the contrary, participation of women outside the labour market, particularly in domestic activities of different types, increased over the survey rounds on emp­loyment and unemployment in India. This observed fact is contrasting to the experiences of structural reforms and opening of the economy of other countries. The present study analyses the fall in female labour participation by looking into the changes in occupational structure and distribution of earnings between occupations by gender using PLFS data in India.

There are some social customs that may have some influence in making decisions on labour market participation, particularly by the women. In analysing the pattern of employment distribution, we hypothesise that social status in terms of ethnicity determines the type of activity performed by a person, given other factors remain the same.

Literature on Gender Wage Gap in India

A notable number of studies are available in the literature foc­using on gender–wage differentials in India (Mishra and Kumar 2005; Madheswaran and Attewell 2007; Das 2012; Sengupta and Das 2014, for example). Most of the studies observed that women earn less than men for roughly similar type of work, although the magnitude of the gender wage gap in these stu­dies differ depending on the period of study and the geogra­phical coverage of the data source and methodologies used. Some studies also have examined the gender wage gap at different locations of the wage distribution. Das (2019) investigated the wage gap between workers in temporary and permanent employment across the wage profile, and tested the relevance of the glass ceiling or sticky floor hypotheses by using the methodology developed in Machado and Mata (2005) and obs­erved that the wage gap between temporary and permanent workers is wider at the upper tail of the distribution both for men and women. Sengupta and Das (2014) have focused on gender wage discrimination across different social and religious groups and concluded that the presence of substantial wage differentials between men and women workers cannot be explained simply by the gender gap of human capital. Das (2012), in a study with unit-level data from the National Sample Survey 61st round, exa­mined wage inequality by decomposing Gini inequality ­index by the sex of workers in the shape of “within” and “bet­ween” components. Das et al (2009), by using micro-level information from household surveys, observed that there had been no improvement in the status of women in the Indian ­labour market almost at all age groups during the 1990s. Madheswaran and Attewell (2007) found that occupational discrimination was more pronounced than wage discrimination among workers in the Scheduled Tribes (STs) and Scheduled Castes (SCs).

Different empirical studies on the labour market in India exa­mined the wage gap and employment gap separately, but they did not take them together to analyse the gender wage differential in terms of employment differences between men and women. In this study, we attempt to explain gender wage gap in Indian labour market, as pointed out in the existing literature, by looking at employment distribution and factors affecting ­occupational choice by using PLFS data published by the National Sample Survey Office.

Employment Types in the Data

In Schedule 10.4 of the survey, activity status is classified into 13 groups consisting mainly of different forms of self-employment, wage employment, and other activities outside the lab­our market. Persons who are either employed or unemployed during the reference period together constitute the “labour force” and persons who are neither “working” nor “seeking or available for work” for various reasons during the reference period are considered to be “out of the labour force.” The persons under the second category are students, those engaged in dome­stic duties, rentiers, pensioners, recipients of remitt­ances, and so on. “Self-employed” are those who operate their own farm or non-farm enterprises or are engaged independently in a profession or trade; the self-employed are further cate­gorised into own-account workers, employers, and unpaid workers in household enterprises. Wage employment is divided into regular wage employment and casual employment. Regular wage workers are those who work in other’s farm, or non-farm enterprises of household, or non-household type and get salary or wages on a regular basis, not on the basis of daily or periodic renewal of work contract. This category not only inc­ludes persons getting time wage but also persons receiving piece wage or salary and paid apprentices, both full-time and part-time. On the other hand, a person working in other’s farm or non-farm enterprises—both household and non-household type—and getting wage according to the terms of the daily or periodic work contract is a casual wage labour.

Occupation types are defined in the survey by following the National Classification of Occupations (NCO) (2004) at the three-digit level. In this study, we have summarised workers’ occupation at the one-digit level in nine categories. In one-digit classification, NCO (2004) describes occupations in the following form: legislatures, and executives (NCO 1); professionals (NCO 2); technicians and associate professionals (NCO 3); clerks (NCO 4); service workers and shop and market sales workers (NCO 5); skilled agricultural and fishery workers (NCO 6); craft and related trades workers (NCO 7); plant and machinery operators and assemblers (NCO 8); and elementary occupations (NCO 9).

We analyse occupational segregation by gender by considering the distribution of workers across their employment status in each occupation. Employment status is categorised into self-employment and wage employment. Self-employment is further divided into own-account worker, employer, and unpaid household worker. Wage employment is again categorised into regular wage and casual wage employment. Casual wage emp­loyment is subdivided into employment in public activities and employment in private activities. Thus, we have six types of employment status in nine occupation groups.

Distribution of Employment

The report on PLFS clearly demonstrates that women’s participation is less than one-third of men’s participation in Indian labour market. Self-employment has been the dominating part followed by employment in casual wage and regular wage, both for men and women workers in rural India (Statement 15, annual report PLFS 2017–18). Within self-employment, a major part of the female workforce worked as unpaid family worker. As unpaid family workers contribute to economy’s production without receiving any pay, they are in more vulnerable situations. In urban economy, on the other hand, the major part of the workers are in regular wage employment and this part has been rising both among men and women workers, but at a higher rate for women over different survey rounds. In casual wage employment, a major part of the female workers are absorbed in domestic work activities like maids and cooks, beauty and wellness service activities, and in call centres. In most cases, their working conditions are alarming.

The proportion of female workers in regular salaried employment, who did not have any written job contract, increased in 2017–18 from their respective share in 2011–12 (Statements 19, 20, and 21, annual report PLFS, 2017–18). The situation of females employment of this type is more alarming in urban locations than in rural areas. The PLFS report demonstrates that, in 2017–18, more than half of the regular-salaried female workers did not have any social security benefits, and the incidence was more among rural women than among urban. Incidence of informalisation of employment is very high in any indicator of informalisation for female workers, and this share has been rising significantly over different survey rounds on employment and unemployment in India.

In the rural economy, the workers, both men and women, were absorbed mainly in agriculture, fishing, and in elementary occupations, and the shares were much higher for women than for men in those occupations (Statement 17, annual report PLFS, 2017–18). In the urban economy, on the other hand, the occupations are relatively scattered across the occupation groups, but still the major occupations were elementary occupations, particularly for women.

For the analysis, a two-way distribution of men and women workers by their occupation and employment status for 2017–18 has been constructed by using information of the sample households of PLFS (Table 1). The distribution of workers by employment status is not similar in every occupation among men and women. Legislatures and executives are mostly own-account workers—both among men and women. Professionals and technicians, on the other hand, are mostly regular wage employees. The shares of women in regular wage employment in these occupations are significantly higher than men. The distribution of persons in clerical jobs is roughly similar for men and women, and they are mostly in regular wage employment. Service-related work, like selling, is mostly own-account and regular wage payment in nature; the incidence of regular wage workers in this occupation is high among female workers. While the share of skilled agricultural workers among men is very high (more than 73%) in own-account work, more than 68% of women workers in this occupation are unpaid family worker. The majority of the craftworkers and trade-related workers among women are own-account type, but the distribution of male workers in this occupation in significantly different. While the dominating part of machine operators is regular wage workers, most of the workers in elementary work are in casual wage employment in the private sector, both for men and women.

(i) Measuring occupational segregation by gender: We measure Duncan and Duncan (1955) index to provide an objective measure of the occupational structures of males and females:

... (1)

Here, Mjk is the employment share of the males and Fjk is that of women in employment type j in occupation k. The ­index ranges between 0 and 1. If the distributions of men and women across occupations are identical, the index would be 0. If, on the other hand, all occupations are completely performed by either men or women only, the index will be 1.

Table 2 (p 61) presents the shares of men and women, and the ­gender-specific dissimilarity index by employment type calculated by using Duncan and Duncan (1955) in each of the nine ­occupational groups as mentioned in the sample data. Each occ­upation in India has been male-dominated. Men’s concentration is the highest among plant operators and related works, foll­owed by legislatures, executives, and trade workers. The distribution of men and women across employment types is roughly identical in occupational group clerks where 80% of clerical workers are men. Dissimilarity in distribution of workers by employment types between male and female is relatively less in elementary occupations, technicians, and the high-paid occupation like legislatures and executives. But, the dissimilarity between men and women in terms of employment status is the highest among skilled workers in agriculture, in which roughly three-fourths of the workers are men.

(ii) Analysing occupational segregation—multinomial logit estimation: To analyse occupational distribution and dissimilarity by gender, we have estimated the multinomial logit model of occupational selection, separately for men and women by ­taking skilled agricultural workers as the reference group where the dissimilarity in employment status between men and women is maximum. Multinomial logit regression is a multi-equation model similar to the multiple linear reg­ression model where the dependent variable is binary. For a limited dep­endent variable with nine occupation groups, the multinomial regression model estimates eight logit equations for eight occ­upations other than skilled agricultural workers, the reference group.


Uij = Xiγj + uij ... (2)

Here, Uij is the utility of individual i in choosing occupation type j, Xi is a vector of observed individual characteristics det­ermining the choice of individual iγj is the coefficient vector in occupation juij is the random error. The utility function is stochastic and a linear function of the observed individual characteristics. Individual will participate in the labour market in occupation j when Uij > Uik.

We estimate a set of coefficients, γjj = 1, 2, 3…8, corresponding to each occupation with respect to the base occupation group—skilled agricultural workers.

The reduced-form multinomial logit model captures how different variables affect the probability of an individual working in a given occupation, while treating the occupational choice as endogenously determined.

... (3)

The supply-driven differences like workers’ personal characteristics in acquiring human capital and job skill in group preferences and demand-driven constraints like pay structures and other employment benefits are likely to create differences in the estimated coefficients between men and women measu­ring the presence of occupational segregation due to limited access to some occupations for female workers (Demurger et al 2008).

We consider employment in different occupations—a multiple category—in binary form as a dependent variable and workers’ education, work experience, marital status, vocational training, along with dependency ratio in the household, social status, and employment status as independent variables. We estimate the multinomial logit model separately for men and women because the employment characteristics are different in these two worker groups. The estimated coefficients shown in Table 3 (p 62) provide the direction of change of predicted probability of employment in a particular occupation relative to the change in probability of employment as skilled agricultural worker due to the change in explanatory variables.

Education plays a significant and positive role in the selection of occupations; the higher the level of education, the higher is the chance in getting high-skilled jobs like legislatures and executives, professional and technicians; the effect of education is the highest for occupation group “professionals” both for male and female workers. However, the effect of education in choosing low-skilled occupations has been quite low. The effect of education—other factors remaining the same—in selecting occupations is stronger for women than that for men. The stronger effect of years of schooling in predicting the probability of entry into a particular occupation among women workers has significance in explaining the occupational segregation by gender.

The work experience measured by workers’ age is associated with an increase in the probability of entering into high-skilled high-paid jobs for male workers, while it has no significant effect among women. Dependency ratio within a household lowers the chances of entry into the labour market in any occupation relative to skilled agricultural work, and it has a strong effect and an even stronger negative effect among women for high-skilled jobs. Similarly, a worker has lesser chance in getting job in any occupational group, compared to skilled work in agriculture, if they are married. This negative effect is more prominent in high-skilled jobs for men and in low-skilled jobs for women. Vocational training has mixed effects on predicted probability of employment. In some occupations, like clerical jobs, vocational training has no significant effect.

We have taken STs as the reference category in finding out the employment gap across social groups. The gap is observed to be more prominent in high-skilled occupations among male workers, while it is more visible in low-skilled occupations among women. The positive coefficients for social group dummies for SCs, Other Backward Classes, and upper castes imply the higher chance of employment outside agriculture for these groups compared to the reference group—STs. The estimated results suggest that the upper-caste people have got more advantages in selecting high-skilled jobs as compared to lower-caste people. If we compare caste differences in job selection, it would be clear that women in lower-caste families are more vulnerable than that in upper-caste families. If we call the differences in employment outcome because of social status or any other circumstances on which an individual has no control at all as employment discrimination, we can say that women have been discriminated at the entry point of the labour market in multiple ways—due to gender difference as well as caste or religion.

The status of employment is not similar between men and women workers across different occupations. To estimate the differential effects across employment types in each occupation outside skilled agricultural works, own-account self-employment is taken as a reference group. It is observed that in legislatures and executives, the conditional probability to work as employers is significantly high among men, while in other high-skilled occupations like professionals and technicians, the probability of being employed as regular paid employee is higher for women. Employment as unpaid family workers is less likely than own-account worker in every occupation, but at a higher rate among men than women.

Earning Differences by Occupation

The PLFS report points out that women’s average earnings are 34% less than men’s in rural areas and 20% less in urban areas among the regular salaried workers. The gender earnings gap is higher among self-employed workers and the gap is even more in urban areas. These observed facts raise concerns over the job conditions even within the regular salaried women workers in the Indian labour market. Table 4 (p 63) shows the mean and standard error of weekly hours work, hourly earnings, and years of schooling for men and women among young-age and working-age people. Significant differences in work hours and hourly earnings have been observed between male and female workers in each type of employment. Average work intensity in terms of hours worked per week was less with higher standard errors for women than for men, irrespective of their activity types. Also, hourly earnings for female workers were significantly less than those for men in each type of employment by activity status; only exception was young-age female workers in regular wage employment who earned more on average than their male counterparts. Earning differences occurred by gender in every type of employment in the Indian labour market, despite no significant difference in education between men and women, and indeed, in some categories of employment like regular wage employment and the working status employer in self-employment among young-age people, women were more educated than men. The presence of the gender pay gap with no gender gap in education in each employment type ­implies a significant impact of other variables not related to ­productivity in determining earnings in different occupations in Indian labour market. We can relate these discriminatory factors in earnings with the similar factors in employment ­discrimination.


The mean hourly earning within a particular occupation is diff­erent across employment types and even in a particular type of employment status, it is different by the gender characteri­stics of the workers. In each occupation category, mean earning is higher among employers and regular wage workers as compared to other types of employment status (Table 5). The gender gap in earnings differs across occupations and in different employment status in a particular occupation. The gender earning gap is the highest among services workers working on a casual basis in the public sector followed by among the same employment status in the occupation group professional. In addition, the gender gap is highly prominent among employers in the occupational groups of clerks and professionals. In regular wage acti­vity status, the gender gap is maximum among trading workers foll­owed by skilled workers in agriculture and services workers. But, in the high-paying jobs like legislatures and executives, female workers earn higher pay than men working as employer or regular wage worker.

(i) Occupation-specific wage regression: We have estimated occupation-specific wage regressions separately for male and female workers:

... (4)

where wij represents the hourly earnings (wage and non-wage) for individual employed in occupation jx’ij is the vector of explanatory variables, βj is the cesponding coefficient vector, and εij represents random error in the wage regression equation for occupation j. The estimated coefficients of occupation-specific wage regressions for male and female workers are shown in Table 6 (p 64).

The variation in the conditional mean earning (in log form) after ignoring the effects of education and other covariates as measured by the estimated intercept across different occupations both for male and female workers can easily be explained by skill differences. There is no significant difference in the mean earning by gender after controlling for all covariates used in this study. But, we observe marked differences in returns to education between male and female workers in each occupation group. Surprisingly enough, the return to education is significantly higher for female workers than for their male counterparts in each occupation, while, as described above, female workers earned less than male workers on an average. The emp­loyer and regular wage worker earned more than the own-account worker in high-skilled jobs, and this earning gap, because of the differences in the employment status in similar occupation, was higher for a female worker than for a male worker.

These findings indicate that human capital plays a little role in explaining the earnings gap by gender, although it has a significant effect in explaining the employment gap in Indian labour market. Occupational segregation, because of gender along with ethnic and religion characteristics and other factors not related to workers’ productivity, may have a significant role in analysing the gender gap in earnings. Men are better compensated than women for more experience in all occupations, and at higher rate in low-paying occupations.

Summary and Conclusions

This study examines the role of occupational segregation in explaining the gender earning gap with unit-level data from the PLFS report (2017–18) in India. We measure occupational segregation and examine its implications in analysing gender differences in occupation and wage and non-wage earnings. Possible sources of gender differences in occupation and earnings are pointed out by estimating the relationship between occupational characteristics and occupational choice in probabilistic sense and occupation-specific earning equations.

The effects of personal and household-specific factors on predicted probability of occupational choice have been estimated by using the multi­nomial logit model. It is observed that additional years of schooling have signi­ficant positive effect on predicted probability of entering into a particular occupation categorised by skill level in PLFS. The effect of education is stronger in high-paid jobs like legislatures and executives, professional, and technicians. In addition, social factors like marital status and ethnic differences among the people have also made significant differences in employment and occupational status by gender. In legislatures and executives, the conditional probability to work as employers is significantly high among men, while in other high-skilled occupations like professionals and technicians, the probability of being emplo­yed as regular paid employee is higher for women.

The distribution of workers by employment status is not similar in every occupation among men and women. In this study, we calculate the shares of men and women, and the gender-specific dissimilarity index by employment type by ­using the Duncan and Duncan (1955) index in each of the nine occupational groups. Dissimilarity in the distribution of workers by employment types between men and women is relatively less in elementary occupations, technicians, and the high-paid occupation like legislatures and executives.

Marked differences in work hours and hourly earnings have been observed between male and female workers. The fact that women have higher levels of education but lower hourly wages than men in these activities implies a significant impact of other variables not related to productivity in determining earnings in different occupations. The gender gap in earnings is different in different occupations and in different employment status in a particular occupation. The study infers that gender differences in returns to schooling are in part driven by dissimilarities in the occupational structures.

The persistence of occupational segregation along with pay differences by gender across occupations suggests that gender gaps in occupation could account for, at least partly, the gender pay gap. Gender differences in risk-taking in job characteristics directly affect the occupational choice by men and women and, consequently, the gender gaps in earnings. Job characteristics like job security and other benefits, risks in earnings, fatal risk at the workplace, performance evaluation, and competitive compensation in different occupations and even within the same occupation across different sectors are different. It is, however, difficult to capture the differences in risk aversion in an empirical study because PLFS does not include direct measures of risk aversion. The empirical findings of this study have some serious implications. The study shows clearly the need for institutional policies in the form of subsidies and provisions for unpaid care work such as childcare facilities, and a social infrastructure that allows for easy mobility of female workers. The low labour force participation and earnings is clearly not associated with their productivity but other factors based on the existing social norms. The paper makes a case for active care policies that shift the burden of unpaid care work from women. Furthermore, it reveals that the labour market in India is imperfect and a lot of distortions persist, which may be the cause of the gender pay gap that is not explained by the productivity-related factors and requires institutional frameworks to address this rigidity.


Bertrand, M (2011): “New Perspectives on Gender,” Handbook of Labor Economics, Orley C Ashenfelter and David Card (eds), Amsterdam: Elsevier, Vol 4b, pp 1545–92.

Blau, F D, Peter Brummund and Albert Yung-Hsu Liu (2013): “Segregation by Gender 1970–2009: Adjusting for the Impact of Changes in the Occupational Coding System,” Demography, Vol 50, No 2, pp 471–92.

Das, P (2012): “Wage Inequality in India: Decomposition by Sector, Gender and Activity Status,” Economic & Political Weekly, Vol 47, No 50, pp 58–64.

— (2019): “Wage Penalty in Temporary Employment in India Evidence from Quantile Regression,” Economic & Political Weekly, Vol 54, No 40, pp 51–59.

Das, P, B Dasgupta and P K Biswas (2009): “Gender and Labour: Post-reform Scenario in India,” Post-reform Development in Asia—Essays for Amiya Kumar Bagchi, M K Sanyal et al (eds), Hyderabad: Orient BlackSwan.

Démurger, S, M Gurgand, Shi Li and X Yue (2008): “Migrants as Second-Class Workers in Urban China? A Decomposition Analysis,” Journal of Comparative Economics, Vol 37, pp 610–28.

Duncan, O and B Duncan (1955): “A Methodological Analysis of Segregation Indexes,” American Sociological Review, Vol 20, No 2, pp 210–17.

Groshen, E (1991): “The Structure of the Female/Male Wage Differential, Is It Who You Are, What You Do, or Where You Work?” Journal of Human Resources, Vol 26, pp 457–72.

Macpherson, D A and T H Barry (1995): “Wages and Gender Composition: Why Do Women’s Jobs Pay Less?” Journal of Labour Economics, Vol 13, pp 426–71.

Machado, J and J Mata (2005): “Counterfactual Decompositions of Changes in Wage Distri­bution Using Quantile Regression,” Journal of Applied Econometrics, Vol 20, No 4, pp 445–65.

Madheswaran, S and P Attewell (2007): “Caste Discrimination in the Indian Urban Labour Market: Evidence from the National Sample Survey,” Economic & Political Weekly, Vol 42, No 41, pp 4146–53.

Mishra, P and U Kumar (2005): “Trade Liberalisation and Wage Inequality: Evidence from India,” International Monetary Fund Working Paper, Nos 05/20, International Monetary Fund, Washington, DC.

Sengupta, A and P Das (2014): “Gender Wage Discrimination across Social and Religious Groups in India Estimates with Unit Level Data,” Economic & Political Weekly, Vol 49, No 21, pp 71–76.

Welch, F (2000): “Growth in Women’s Relative Wages and in Inequality among Men: One Phenomenon or Two?” American Economic Review, Vol 90, No 2, pp 444–49.


Dear Reader,

To continue reading, become a subscriber.

Explore our attractive subscription offers.

Click here

Updated On : 17th Jul, 2022
Back to Top