President Trump with Commerce Secretary Wilbur Ross and Attorney General William P. Barr at the White House on Thursday. (Jabin Botsford/The Washington Post)

After pushing for the 2020 Census to include a question on citizenship, President Trump abruptly and unexpectedly gave up on that effort Thursday afternoon.

Instead, he offered an executive order that orders government agencies to consolidate existing tools and data to allow the administration to better estimate how many immigrants in the United States haven’t yet been naturalized. The order includes a number of rationales for the need for such a determination, including this one:

“[I]t may be open to States to design State and local legislative districts based on the population of voter-eligible citizens. ... [B]ecause eligibility to vote depends in part on citizenship, States could more effectively exercise this option with a more accurate and complete count of the citizen population.”

That sounds fairly mundane, though an accompanying insistence that courts have not ruled out such a process hints at how it might be controversial.

In fact, it’s one of the most controversial parts of the entire fight. Opponents of adding the question to the census argued that this was precisely the government’s goal: to make it easier for states to redistrict according to the population of eligible voters, instead of the population overall. When Trump tacitly acknowledged that he wanted to include the question to facilitate redistricting, it was rightly identified as an admission that his opponents were right all along.

Why does it actually matter? Why do Republicans, including Trump, think there’s a benefit to drawing district lines based on voters instead of population? For two reasons: First, it would disproportionately exclude groups who tend to vote more heavily Democratic, making heavily Democratic districts less compact and therefore leading to fewer such districts. Second, it would water down the power of those constituents by packing more people into districts and therefore giving legislators bigger constituencies to have to manage.

We can illustrate how this works with a visual example.

Take this made-up, perfectly square state. It’s home to two types of people, some green and some purple. They live in various places, but the majority of the population lives in a city at the central northern portion of the state.

(Philip Bump/The Washington Post)

In this state, there are 75 purple people and 53 green people. So that’s 128 people in total.

Let’s assume the state is drawing four congressional districts. Each, then, should have 32 people (or 128 divided by four).

Here’s one way to draw those districts.

(Philip Bump/The Washington Post)

We’ve broken down the resulting demographics from that distribution. As happens in reality, the densely packed city results in a lopsided distribution of one group in Districts 1 and 2.

Here, each group is the majority in two districts. We can draw the lines differently, of course. Because about 60 percent of the population is purple, we can draw lines that make purple the majority in three districts instead of two.

(Philip Bump/The Washington Post)

Anyway, you get the point. This is how drawing district lines work.

But now let’s look at the scenario the Trump administration proposes: Focusing only on the population that’s eligible to vote. That means no one under the age of 18 and no one who isn’t a citizen.

So we take the same state and identify the people who are either minors or immigrants who haven’t been naturalized.

(Philip Bump/The Washington Post)

We made this all up, of course. We could have created a scenario in which the changes were more subtle. But let’s use this example to make our point.

We’ve assumed that there are far more noncitizen purple people than green. We’ve also made more purple people minors. This is intentional; because we’re loosely analogizing to the real world, it’s worth accommodating for how Hispanics in the United States tend to be younger than other groups, immigration status notwithstanding.

The result? While purple people make up nearly 60 percent of the population, in this scenario they make up less than 50 percent of the people eligible to vote.

When we draw the lines now, we have about 18 voting-eligible people per district (75 divided by four). Here’s one way that might shake out.

(Philip Bump/The Washington Post)

Notice that here we have two green-majority districts and one split district. We could have drawn these another way, certainly, but it’s inherently easier to draw more green-friendly districts because the population difference with the purple group has been erased by focusing on potential voters instead of population.

What’s more, the overall population of the districts varies widely. When we drew districts based on population, the goal was 32 people per district. Here, we instead have two districts with total populations over 32 and two with populations well below that mark, though each district has about the same number of voting-eligible people. The two districts with populations below 32 are both green-majority.

This example isn’t completely detached from reality. In June, after documents surfaced showing a Republican operative predicting a benefit to his party from implementing a system like this, we looked at the population in each congressional district and noted how Republicans might benefit.

One response to this would be to note that race and party aren’t the same thing. If we’re using green and purple as proxies for racial groups, how does that overlap with partisan politics?

The answer is straightforward. As we’ve written previously, race is often a very good proxy for party — at least among nonwhites. Pew Research Center data show the gap in party identification by racial groups. By identifying and reducing the power of black and, to a lesser extent, Hispanic residents, you’re disproportionately reducing the power of Democrats.

(Philip Bump/The Washington Post)

There’s a toxicity to using race as a way to build your own power or undercut the power of others — regardless of which party does it. The overlap of partisanship and racial identity adds a particularly problematic layer to the increase in partisan polarization and the increase in hostility toward the other side of the political spectrum.

But the appeal of building power is often much stronger than concerns about the effects of how you do so.