What is wrong with the pay equality prediction?

Woman and man in office
Thomas Tolstrup | Getty Images

The headlines have spoken: Women shouldn't expect to receive the same pay as men until 2133, according to the World Economic Forum.

But wait. Last year, we were told that the gender gap would be closed by 2095 — nearly 40 years sooner. How did things get so much worse in just a year?

Well, you may be surprised to learn that that number — pay parity in 2133 — doesn't actually show up in the World Economic Forum report itself. It's a calculation that was made for the press releases that accompany the 400-page report.

While it is an interesting way to point out how slow progress has been, it probably won't mean anything in the real world — it's just too much of an extrapolation into the future to have much value. The number changes wildly with each new year of data — here are a few estimates of when we should reach economic equality in a few years of the report based on the same methodology:

Of course, projections should change as new data come in. But we only have 10 years of data, so every hiccup in the numbers will cause vast swings in the projected equality date. In this case, that new data pointed to slower progress this year. Next year it could swing the other way. It's a well known statistical issue — the further out you extrapolate from the known data, the more error you introduce.

The second problem is there is no reason to assume that reaching economic equality is a linear process. In fact, there are a lot of reasons to assume the opposite. After all, if the move toward equality were strictly linear, then we would expect women to be making twice as much as men in the century after that — but obviously cultural and structural forces intervene at certain parts of the process.

"We put this out there saying if the rate of change in the last 10 years continued in this way, then this could happen with all other things being equal," said Saadia Zahidi, the forum's head of employment and gender initiatives. "But obviously not all things will be equal."

In the real world, social change tends to come in bursts. Take a look at how quickly Americans changed their mind on gay marriage and other issues.

Not all things will be equal

Zahidi pointed out a number of factors that could help the economic participation and opportunity subindex reach 100 percent before 2133: Global technological and economic changes and shifts in business practices are already happening and could boost the rate of change in the near future. Not to mention, as standards change, each generation expects more from the workplace.

"There's certainly something to be said about a snowball effect in terms of women's aspirations," said Zahidi. "There are obviously a lot of changes coming in terms of millennials' viewpoints including both women and men and what they desire from work and family."

The rate of change also differs substantially between countries — many countries, like Guatemala, have low scores but have seen big gains in recent years, while others have been more steady. The economic participation and opportunity subindex takes into account more than just male and female earnings — it also includes wages for the same job, workforce participation and advancement in powerful roles. So it's hard to make projections based on how all those different rates will converge for the global total.

There's also a chance that the trend line will slow as it approaches 100, since exact equality is thwarted by more pernicious structural or cultural issues. Zahidi is optimistic that progress will not hit a plateau at that point — many of the top ranking-countries have continued to show progress even as they approach 100. Iceland, the most gender-equal country, has gained more than 10 percentage points in the last year and could hit 100 in a little over a decade.

"It remains to be seen if there is some natural plateau, if at some point equality of opportunity doesn't lead to equality of outcomes," said Zahidi.

So the final trajectory of economic equality will likely look much different than a line ending at 2133. It could be within our lifetimes, or maybe not — we'll have to wait and see. But we can be sure that using 10 years of data to project out more than 100 years in the future isn't going to be very accurate.