Field notes on pay equity analysis. The first thing to settle is which “pay gap” someone is quoting, because two different numbers travel under the same name and people argue past each other for an hour before noticing.
The raw gap
Take the average (or median) pay of all men and all women in the company, and compare. Illustrative numbers: men average ₹18 lakh a year, women average ₹15.3 lakh. The gap is (18 – 15.3) / 18 = 15 percent. This number answers a question about the whole organisation: how is pay distributed between men and women? It says nothing about whether any two people doing the same job are paid differently.
Why can the raw gap be large without anyone being underpaid for their job? Because of who sits where. If most senior engineering leaders are men and most junior support staff are women, the averages diverge even if every individual is paid exactly by the book. That is a representation problem, a pipeline problem, and it is serious, but it is not a pay-setting problem. The fix is hiring, promotion and retention, not corrections in payroll.
The adjusted gap
Here you compare like with like. You control for the legitimate factors that explain pay: job level or grade, function, location, tenure, performance rating, perhaps critical skills. The method is usually a regression: pay is the outcome, those factors are inputs, and gender is added to see what it explains once the others are accounted for. If the gender coefficient is, say, minus 3 percent, women are paid about 3 percent less than men with the same level, function, location and so on.
You do not need a statistics team to start. A simpler version is a compa-ratio comparison: pay divided by the midpoint of the salary band, by gender, within each grade and function. If women in grade 5 average 0.93 and men 1.01, you have a lead worth chasing.
Where my opinion comes in
I think companies lean on the adjusted gap too comfortably. “After controlling for everything, it is only 3 percent” sounds reassuring, but each control variable can itself carry bias. If women are rated lower on average, controlling for rating hides the effect. If they are placed at a lower level for the same experience, controlling for level hides that too. The adjusted gap tells you whether pay decisions are consistent given the inputs, not whether the inputs were fair. Run it, but look at the inputs as well: starting pay, ratings distribution, promotion rate by gender.
The reverse error is just as common. A leader sees a raw gap of 15 percent and demands the money be “fixed” by raising pay of women across the board. That spends budget on the wrong problem and can create reverse inequities inside grades.
Doing the analysis in practice
- Clean the data: grades, locations, joining dates, ratings, total pay including variable and allowances. Most delays are here.
- Compute the raw gap, overall and by function and level.
- Compute the adjusted gap, and the compa-ratio by gender within grade.
- List outliers: individuals far below their peer group. Ignore groups with fewer than five or so people, where the numbers are noise.
- Review each outlier by hand for a real reason (rare skill, recent promotion, part-year service). If none exists, it is an unexplained gap.
- Decide corrections, budget them, and fix them in the next cycle or off-cycle.
What if you find an unexplained gap
Correct it, and say so internally. Quietly adjusting a few salaries while the rest of the organisation guesses invites rumours. Then find out where it comes from. In many companies the leak is the starting offer: a candidate negotiates, another does not, and the difference compounds through percentage-based raises for years. Fixes include offering from the band with a clear rule instead of negotiating ad hoc, not asking about current pay where you can avoid it, and reviewing offers above a threshold.
On the legal side, Indian law has long required equal remuneration for men and women doing the same work or work of a similar nature and prohibits discrimination in recruitment on the basis of gender; the provisions that came from the Equal Remuneration Act now sit in the Code on Wages. There is no general requirement in India to publish a gender pay gap figure, though listed companies may report related data in sustainability disclosures, and some global clients ask for it. Check the current requirements for your company type.
Last point. Gender is the usual cut, but the same method works for other groups, such as people hired from a merger, or by age or disability, as long as data privacy rules are respected in how you collect and use that data. A pay equity review is not a one-off. Repeat it every year after the increment cycle, because that is when gaps reopen.
Figures and names in the examples are illustrative.
Leave a Reply