At some point, the age of the baby boomer in American politics will end. It’s simple demographics: Boomers keep getting older and older, and that means there are fewer and fewer of them. But, contrary to two recent news articles, the boomer political era hasn’t ended yet — and it won’t end next year, either.

We will start with an article from Pew Research, which notes that, for the first time, voters younger than the boomers outnumbered baby boomers (and those older) in votes cast in 2016.

There’s an important caveat to this that we’ve noted before. Unlike “baby boomer” — people born between 1946 and 1964 — there is no set definition for “millennial” (or “Generation X,” for that matter). Generations are mostly made-up marketing gimmicks, with the exception of the boomers for whom particular demographic boundaries exist. So Pew’s definition of “millennial” isn’t necessarily someone else’s, and that’s important to remember for analysis of how the generations compare.

That same issue affects another recent article. At CNN, political analyst Ron Brownstein predicts that 2018 will be the first election in which there are more millennials eligible to vote than boomers. “That transition,” he writes, “will end a remarkable four decades of dominance for the baby boomers, who have been the largest generation of eligible voters since 1978, when they surpassed what’s been popularly referred to as the Greatest Generation (or G.I. Generation) raised during the Depression.”

But then he raises a critical point, one also at the heart of the Pew analysis: Young people vote less consistently.

Pew’s combination of Gen Xers and millennials in its comparison to boomers is important because there are more millennials than Gen Xers — but Gen Xers vote a lot more heavily.

Take Gen Xers out of Pew’s calculus and boomers are still the biggest voting bloc.

In early 2015, we used data from Political Data to graph turnout by age in California the previous November. The result was a remarkable curve in which first-time voters cast ballots more heavily than those in their 20s, after which point turnout tracked upward along with age.

That was an off-year election, which matters. We pulled exit polling data from elections since 1976 and compared turnout in that year’s presidential or House elections with the composition of the population at large. In every election, the composition of the electorate contained a lower percentage of the youngest age group (usually 18- to 29-year-olds) than the population on the whole. In most elections, the second-oldest group turned out more heavily as a percentage of the electorate than of the population.

But notice the difference between the red years (presidential) and the green (off-year) ones. While in most presidential years, the oldest age group turned out as a lower percentage of the electorate than they constituted in the population, that wasn’t the case in off-year elections. (Arrows mark exceptions and extremes to this rule.)

This a critical point for Brownstein’s thesis: Millennials (however you define the group) aren’t going to vote more heavily than boomers next year because young people simply don’t vote as much, particularly in off-year elections! And they didn’t vote more than boomers last year, either.

Again: At some point, this will change, and millennials will call the shots in our political process. But for that to happen, they need to actually vote. And for that to happen, it seems, they need to get a little older.