Thanks to shifting demographics, the Democratic Party could face a nightmarish scenario in 2020: losing the electoral vote but winning the popular vote by an even wider margin than it did three years ago.

How? As Dave Wasserman explained for NBC News, the nonwhite (and therefore generally more Democratic-voting) populations in the two largest states, California and Texas, have grown substantially since 2016. Adding a ton more Democrats to California isn’t going to change things, since Hillary Clinton won the state. Adding a ton more Democrats to Texas, which backed Trump? Maybe — except that the margin by which Trump won is still wide enough to withstand the surge. So: a lot more votes for a Democrat but potentially no change in result.

Sen. Bernie Sanders (I-Vt.), one of several people who hopes to be the Democrat earning those votes, responded on Twitter by calling for the abolishment of the electoral college. Getting rid of the electoral college, of course, would mean that the person who received more votes for the presidency would get to serve in that office. In 2016, that was Hillary Clinton.

On Twitter, the Weekly Standard’s Jay Cost responded to Sanders. Over two dozen tweets, he argued against abolishing the electoral college, centering his argument on two points. First, that the country was created in a way that gave states a role in every level of government. Second, that having an electoral college forces candidates to build “geographically broad coalitions.”

“[M]aybe gentry liberals on the coasts should stop bellyaching about the rules,” he wrote, “and figure out how the hell they lost Ashtabula County by nearly 20 (!) points.”

The complaints about the electoral college at this point, he wrote, are based on complaints about the Democrats suffering a disadvantage at the moment, which is obviously true to a large extent (particularly when coming from Sanders). But that said, neither of his two points is particularly compelling.

Take the Ashtabula County argument. Cost is referring to the county in Ohio, a state that backed Trump by a wide margin in 2016. Ashtabula backed Barack Obama in 2012 by 13 points, swinging widely to the right under Trump. Obama’s 5,500-vote victory became a 7,500-vote loss.

Of course, it may be the case that blue-collar conservatives should figure out how the hell they lost Los Angeles County by nearly 50 (!) points. As far as percentages go, the difference from 2012 was more modest: Obama won by 42 points. But in terms of votes? Clinton’s margin in the county was 400,000 votes wider than Obama’s. That’s more than four times the entire population of Ashtabula County.

Why is it more important to worry about 5,500 votes in northeast Ohio than 1.3 million in Los Angeles? Only because of where Ashtabula County is.

Let’s consider the wonkiness of geography in another way.

This map shows the results of 2016 as they stood, with counties color-coded according to how strongly they favored either candidate.

And this one shows the exact same map, county by county, with one difference: Because of how state lines are drawn in two places, in this one, Hillary Clinton wins the presidency.

Can you spot the difference? Well, in the second map, 10 counties in the western part of the Florida Panhandle are instead given to Alabama, extending that state’s coastline. Another 10 counties in the Michigan peninsula are merged onto Wisconsin.

With those two moves, Florida and Michigan suddenly turn blue and Clinton wins a 278 to 260 electoral college victory. (We’re ignoring the electoral vote Trump won in Maine since it doesn’t matter for our purposes here.)

Again, the only reason Clinton wins on this map is that state boundaries are shifted slightly. FiveThirtyEight has a tool allowing you to explore other shifts in state boundaries to see how the electoral vote would change. That article contains an important caveat: One can’t simply redraw state lines without assuming there would similarly be changes in how places would vote or where campaigns would target.

That said, it’s baffling to argue that where those state lines are drawn should be particularly meaningful in selecting a president for the entire country. Cost does make a good argument that considering geography is important in ensuring that states in the middle of the country feel as though they have a stake in what’s happening nationally. Let California and New York run everything, he argues, and see how long Nebraska is interested in being part of the United States. But he offers that argument in defense of the Senate that, even if the electoral college vanishes, would still exist.

His other argument, about how the Founders intended things to work, is to some extent an argument based on faith. Things have worked generally okay so far and we don’t know what will happen.

It’s important to remember, though, that the country doesn’t look the way it did at the time the Constitution was ratified. It doesn’t look the way it did a century ago.

We can look at this through the lens of demographics, in fact. Over the past 30 years, the percentage of the electorate that is white (and isn’t Hispanic) has fallen from 85 percent to about 73 percent in 2018. The 2018 electorate was 3 percentage points less white than the 2008 electorate that brought Obama to the White House.

As noted above, nonwhite voters tend to vote more heavily Democratic. In 2016, there was a correlation between the density of the nonwhite population in a state and how heavily Democratic it voted.

It’s not the case that there’s a necessary correlation between the density of the white population in a state and how reliably Democratic it is. But there’s a big political and cultural divide between a California (about 60 percent nonwhite) and an Ohio (about 80 percent white). Since 1990, the density of California’s nonwhite population has grown twice as fast as Ohio’s. On this metric, which overlaps with partisanship, the two states are much different than they used to be.

As the Pew Research Center has found, partisanship is the widest divide in American culture. Partisan identity informs views on a broad range of issues, including apolitical ones.

We can flip Cost’s question about the Senate around: How long would New York be willing to accept contributing far more votes to a presidential election only to see the electoral college vote go to a Republican? Especially given the extent to which New York exports tax dollars to the federal government.

To be somewhat flip: Is the location of Florida’s state line on the Gulf Coast really a critical determinant in who is handed the immense power of the presidency?