The gender pay gap is close to parity among the youngest workers and widens steadily with age. Women aged 16 to 24 earn around 94 percent of what men the same age earn, those aged 25 to 34 about 89 percent, and from 35 onward the ratio settles into a band of roughly 77 to 83 percent. That shape means the divergence happens during a career rather than at the point of entry, which changes both how you should read the headline figure and what an individual can usefully do about it.
The shape of it
Break the headline gender pay ratio down by age and it stops being one number. Among workers aged 16 to 24 women earn roughly 94 percent of what men the same age earn, which is close enough to parity that entry-level pay-setting is clearly not where the story is. By 25 to 34 the figure is about 89 percent. From 35 onward it settles into a band of roughly 77 to 83 percent and stays there for the rest of a working life.
What makes that pattern so informative is what it rules out. An explanation resting on employers offering women less at hire has to account for near-parity at hire, and cannot. An explanation resting on occupational choice made at eighteen has to account for the divergence appearing fifteen years later, and struggles. The gap is produced during careers, in the decade or so where most of the compounding happens.
It is also worth noticing that the widening stops. The ratio does not keep falling into people’s fifties and sixties; it reaches a level and stays roughly there. Whatever causes the divergence largely finishes doing so by the late thirties, which narrows the window worth examining considerably.
Why compounding does most of the work
Pay raises are almost always expressed as percentages of current pay. That single convention means every difference that opens early is multiplied by every subsequent increase rather than merely sitting alongside it. Two people who start level and diverge by four percent in year three do not stay four percent apart in any meaningful sense. The gap grows in absolute terms every single year, even when both receive identical percentage raises for the rest of their careers.
Run it forward and the arithmetic is stark. Somebody starting at $60,000 and receiving 3 percent a year reaches about $80,600 after ten years. Somebody who fell 4 percent behind in year three and received the same 3 percent thereafter reaches about $77,400 — a gap of roughly $3,200 a year that neither person’s annual raise ever caused, because it was set years earlier and then compounded.
That is why a single moment of divergence matters so much more than it feels like it should at the time. A modest difference in a starting offer, one missed promotion cycle, or one year of raises taken at a lower rate does not stay modest. The percentage machinery of ordinary pay administration converts it into a permanent and growing gap without anybody making a further decision.
What actually happens in those years
The period from the late twenties to the late thirties is when several things coincide. It is when promotion into the management layer typically happens, and management is where the step changes in pay live rather than the gradual increases. It is when specialization decisions get made, and specializing into a well-paid niche is worth far more than any raise. And it is when career interruptions concentrate, which in practice fall unevenly.
An interruption costs more than the salary forgone during it. It removes the compounding for that period, it frequently means returning at or near the previous level rather than the one seniority would have reached, and it can miss a promotion cycle whose effects then compound for the following two decades. The visible cost is a year of pay; the actual cost is a permanently lower trajectory.
None of this requires anybody to behave badly for the pattern to appear in the data. It requires only three things: that raises are percentages, that promotions are competitive, and that time out of the market has a cost. Those are ordinary features of employment that nobody designed with this outcome in mind, which is precisely why the pattern is so persistent and so hard to attribute to anybody’s decision.
Why this matters for how you read the number
If you are in your twenties and the gap looks like somebody else’s problem, the age data is telling you something inconvenient. This is precisely the period when the decisions that produce it are being made. The gap arrives later and is caused earlier, which is an uncomfortable combination — the cost is invisible at exactly the point where it could still be avoided cheaply, and obvious only once avoiding it has become expensive.
And if you are past that window, the same data says something different but equally useful. The divergence has largely already happened, which means the remedy is a step change rather than an incremental one. Percentage raises will not close a gap that percentage raises created, because they preserve proportions by construction. Only a level change, a move, or a renegotiated base does that.
What an individual can actually do
Treat the starting number as the most consequential figure in any job you take, because it is the base that every future percentage multiplies. Negotiating a starting salary up by five percent is worth more across a decade than several years of above-average raises applied to a lower base. It is also a single conversation rather than a repeated one, which makes it the highest-return twenty minutes in the whole process.
Benchmark at least every two years rather than at moments of dissatisfaction. The whole mechanism runs quietly, so it is only visible against an external reference — the percentile range for your occupation in your area, or posted ranges for the role in a jurisdiction requiring them. Somebody who checks regularly notices a four percent drift in year three; somebody who does not notices it in year nine when it has become twelve.
Ask about the band rather than the raise. The top of your band is the ceiling on what percentage increases can ever reach, and if you are near it the productive conversation is about level and scope rather than performance. That is the step change the compounding argument implies, and it is a different conversation from the one most people prepare for.
And treat any planned career interruption as a moment to establish the return terms in advance and in writing, including level and salary basis on return. That is not a guarantee of anything, and it will not survive a reorganization. It does convert an assumption into an agreement, and the gap between those two is where most of the avoidable loss occurs.
This is general information about what the data shows rather than legal or financial advice about your situation. If you believe a specific pay difference between you and a comparable colleague is unlawful, that is a distinct legal question with its own tests and deadlines, covered separately in this section.
Common questions
How does the gender pay gap change with age?
About 94 percent for workers aged 16 to 24, about 89 percent at 25 to 34, then roughly 77 to 83 percent from 35 onward, where it stays.
What does that shape rule out?
Explanations resting on employers offering women less at hire, since pay is near parity at hire. The divergence is produced during careers rather than at entry.
Why does compounding matter so much?
Raises are percentages of current pay, so an early difference is multiplied rather than added to. It grows in absolute terms every year even when both people get identical percentage raises.
Can you show the arithmetic?
At $60,000 with 3 percent annual raises you reach about $80,600 after ten years. Falling 4 percent behind in year three and receiving the same raises gets you to about $77,400 — a $3,200 gap no annual raise caused.
What happens in the divergence years?
Promotion into the management layer, where the step changes in pay are; specialization decisions; and career interruptions, which fall unevenly and cost far more than the salary forgone.
Why does an interruption cost more than the pay missed?
It removes compounding for that period, often means returning at the previous level rather than the one seniority would have reached, and can miss a promotion cycle that then compounds for two decades.
What is the single most useful action?
Treat the starting number as the most consequential figure in any job. It is the base every future percentage multiplies, and negotiating it is a one-time conversation.
How often should I benchmark?
At least every two years, not at moments of dissatisfaction. The drift is only visible against an external reference, and someone checking regularly catches four percent in year three rather than twelve in year nine.