Every two years, the Census Bureau releases data breaking down voter turnout in federal elections by race and age. It’s one of the most comprehensive looks America gets at its process of choosing leaders and a close read of the numbers offers insights about the most important factor in that choice: who actually weighs in.

On Thursday, the bureau released its data for 2020. It shows, for example, that the downward trend in the density of the White vote in presidential elections has continued. In 2004, 79 percent of voters were White, according to the bureau. Last year, only 71 percent were.

Because the government separates race and ethnicity, the figures above are for non-Hispanic Whites. The bureau figures that more than 10 percent of voters in 2020 were Hispanic and another 12 percent Black, including Hispanic Blacks.

On the chart above, we’ve included the bureau’s estimates as well as estimates of the density of the vote by race compiled by Michael McDonald of the U.S. Election Project. Recognizing the way that the bureau’s data collection can overstate non-White participation (read more here or here), McDonald adjusts the data to better represent turnout.

What’s important to note here is that the density of the Black and Hispanic vote is lower than the density of those groups in the population overall. The country is 60 percent non-Hispanic White and 19 percent Hispanic, but you wouldn’t know it based on voter turnout.

Now consider the bureau’s measurement of turnout by age.

You can see that turnout was up from 2016 across the board; for nearly every individual year of age, turnout was higher in 2020 than in 2016. But it increased more among younger age groups. On average, turnout was up 8.6 points for the 18-to-24 age group, up 6 points among those 25 to 44, up 4.2 points for those 45 to 64 and up 3.7 points for those 65 and older.

These two bits of data — turnout by race and turnout by age — overlap. Older Americans are more heavily White than are younger Americans. So when we talk about non-White voters turning out less heavily and younger voters turning out less heavily, we’re often talking about the same voters.

Here’s average turnout for an age group vs. the density of the non-Hispanic White population in the group. Until you get to the 75 and older population, where turnout declines are in part a function of health, there’s an obvious correlation: Younger, more non-White voters turn out less.

Once we consider the correlation of age with race, that initial graph is obviously measuring something more complicated than just that Whites vote disproportionately. We know, for example, that voting can be habit-forming, meaning that those who’ve voted in the past are more likely to do so in the future. If you haven’t yet gotten into that pattern, perhaps because you’re newly registered, that has an effect.

So do other factors that correlate to youth. People who rent are less likely to vote than those who own houses, as the new census data show, and those who’ve lived in the same place for a long time are more likely to vote as well.

Why? In part because moving means having to find a new polling place or having to re-register to vote. People who own houses tend to have consistent registrations and tend to vote in the same place. And people who own houses tend to be older and are more likely to be White.

The bureau asked those who reported not voting (all of this data is self-reported, by the way, an important consideration) why they hadn’t done so. One of the response categories was that the respondent was too busy or had scheduling conflicts. Among those under 25, about 1 in 6 cited this reason. Among those 65 and older, only about 2 percent did.

Older people are more likely to be retired, of course, giving them fairly clear schedules relative to those still in the workforce. But it’s also the case that older people who are probably still working were less likely to cite scheduling complaints. Older workers are more likely to have more established careers and more stable schedules than are younger people who are relatively new to employment.

These patterns predate 2020 and they by no means capture the nuances in different turnout rates. But they are factors, ones that tend to get buried under top-line numbers.

Put another way, there are a lot of reasons people don’t vote and a lot of those reasons land more heavily on younger voters, a group that tends to be more densely non-White — and, generally, Democratic. This is a key element of the push from the left for things like mail-in balloting (to reduce the need to figure out a polling place) or expanded voting hours (to avoid schedule conflicts). Fixing the obstacles young voters face means more young voters and more Democratic votes.