How Job Ad Language Shapes India’s Gender Wage Gap
India’s gender wage gaps begin earlier than we think – at the application stage, shaped by the language of job ads
Sugat Chaturvedi: IIT Delhi
Kanika Mahajan: Ashoka University
Zahra Siddique: University of Bristol
SDG 5: Gender Equality | SDG 8: Decent Work and Economic Growth
Ministry of Labour and Employment | Ministry of Women and Child Development
As online job portals increasingly mediate access to urban employment, job ads function as a first layer of labour market allocation. When their language is systematically gendered, inequality is produced early and at scale – not only through overt discrimination, but through how roles are described, which skills are emphasised, and which working conditions are highlighted. These choices influence who applies and which jobs different candidates perceive as viable.
This early sorting has measurable consequences. Women typically apply to lower-paying jobs than comparable men, creating wage differences before experience, performance, or bargaining come into play.
Gendered Signals in a Modern Labour Market
Online job portals are often perceived as more meritocratic than traditional recruitment networks, but they are not inherently free from discrimination and can still reflect or reproduce biases in hiring. In particular, explicit gender signalling remains widespread. Data from a large Indian online job portal covering nearly 158,000 job ads and 6.45 million applications posted between mid-2018 and early-2020 suggest that nearly 8 percent of ads explicitly specify a gender in their job description, despite operating in a legal grey zone under existing labour law.
Because wages are displayed in most postings, it is possible to observe how these preferences align with pay. Ads that explicitly prefer women offer the lowest posted wages on average, followed by ads preferring men; postings without any gender specification pay the highest.
This pattern is not random. When roles require higher skills or offer higher pay, employers are less likely to specify a preferred gender, since doing so would shrink an already limited applicant pool. Explicit gender requests therefore cluster in lower-wage, lower-skill jobs.
How Language Encodes Gender and Job Value
The signals embedded in job descriptions are revealing. Textual analyses from job ads suggest that female-coded ads tend to emphasise interpersonal traits, appearance, routine computer or accounting tasks, design and beauty services, and flexibility features such as work-from-home. Male-coded ads, by contrast, stress programming, analytics, hardware, finance, leadership responsibilities, and forms of inflexibility such as night shifts, rotational hours, travel, or fieldwork.
Some of these patterns reflect compensating differentials – the tendency for jobs to pay differently based on advertised work conditions. For example, jobs with female associated hard skills tend to offer lower posted salaries and attract a higher share of female applicants. On the other hand, less flexible jobs offer higher salaries but receive fewer applications from women.
When these trade-offs are encoded at the recruitment stage, wage gaps are no longer just about later choices or negotiations, but are built into initial job matches that shape early experience and, over time, affect future pay.
Early Sorting and the Wage Gap at Application
So how large is this effect at the point of application? After accounting for age, education, and location, women apply to jobs that pay about 3.7 percent less than those comparable men apply to. This gap is large enough to be comparable to the wage difference between entry-level jobs and jobs that already require some experience. For example, within the same occupation and state, jobs that require one to two years of experience advertise wages about 7.4 percent higher than entry-level jobs. Seen against this benchmark, the application-stage gap is economically meaningful – roughly half the size of the wage difference associated with modest experience requirements.
A decomposition of this gap helps identify how much leverage the language of job ads has. Roughly 45 percent reflects occupational and geographic sorting, as men and women apply to different jobs and locations. Explicit gender requests explain an additional 7 percent of the application-stage wage gap, and when subtler gendered language is included, recruitment signals together account for roughly 17 percent. This indicates that changing how jobs are advertised can influence outcomes at the application stage, even though it cannot eliminate gender wage gaps on its own.
Gendered Language and Employer Misperceptions
Job ad language also offers insight into how employer expectations align with applicant behaviour. Words associated with flexible arrangements such as work-from-home, and with certain service-sector skills – such as basic accounting, office support, and design or beauty services – are treated by employers as female-oriented and are also linked to a higher share of women applicants.
For soft skills and personality traits, however, the alignment weakens. Attributes that recruiters often code as “female”, such as being punctual or keen, are associated with a lower share of women applicants. By contrast, traits more commonly coded as “male”, including being motivated, passionate, or creative, are associated with a higher share of women applying. This suggests that some elements of the candidate profiles embedded in job descriptions reflect recruiter stereotypes that do not fully correspond to observed application behaviour.
These patterns have implications for employers as well. Within the same occupation and state, job ads that explicitly specify women receive about 50 percent fewer applications relative to comparable gender-neutral ads. But in this smaller pool of applicants, the share of women rises by about 15.5 percentage points. Explicit male preferences are associated with marginally smaller reductions, while the share of women falls by about 9.5 percentage points.
Taken together, gendered language reshapes the composition of applicant pools, while also narrowing them – an effect that matters for both matching and recruitment efficiency.
What Policy and Practice Can Change
The policy question is not whether to regulate every adjective in a job description, but how to ensure that recruitment practices remain gender-equitable.
India’s legal framework already prohibits gender-based discrimination in recruitment, with limited exceptions linked to safety or the nature of work. But, both public and private online platforms operate with little scrutiny. A pragmatic response would apply existing law more consistently across public and private recruitment, rather than create new regulation.
Platforms are the primary lever for change. Labour authorities could issue directions requiring job portals to flag explicit gender requests and prompt employers to justify or revise them.
Where inflexible hours, travel, or night work are genuinely required, employers should address safety and logistics directly rather than treating lower female take-up as evidence of unsuitability.
Digital recruitment also enables low-cost monitoring that can reinforce these shifts. Large platforms can periodically publish anonymised diagnostics on gendered wording in the job ads posted on their platform. Such transparency can nudge behaviour without heavy-handed enforcement.
India has invested heavily in women’s education – preventing early career sorting from reproducing gender gaps is central to realising its returns.
Authors:
The discussion in this article is based on the authors’ research published in Labour Economics (Volume 96). Views are personal.