Americans of color are apparently moving to the right. Donald Trump is poised to make historic gains with nonwhite voters in November, according to recent polling, accelerating the racial depolarization of the electorate. Polls consistently show that Latinos are continuing a significant shift toward Trump, who already received a surge of Latino support in 2020. Many polls also show Black voters, especially Black men, moving away from Kamala Harris toward Trump. And some data shows Asian Americans drifting rightward, with support for Democrats dropping in younger generations especially. Although all three groups still heavily favor Harris, Trump appears to making real inroads, campaigning for nonwhite support to erode his opponent’s edge among groups at the heart of the Democratic coalition.

But Trump has a problem. Even if he wins over a large number of nonwhite voters, it may not help him secure the presidency. It may, in fact, make no meaningful difference in his quest for 270 electoral votes. Trump is, ironically, crashing into the same roadblock that has stymied Democrats for decades now: The Electoral College overvalues white votes at the expense of racial minorities, giving white voters considerably more influence over the presidential race. Nonwhite voters are distributed inefficiently—for purposes of the Electoral College’s arbitrary geography—in states that are highly unlikely to put either candidate over the top. This biased and anachronistic institution is therefore working as the Framers intended: It was designed to give Southern slave states an edge in selecting the president, and today it gives white Americans an unfair advantage over minority voters by inflating the value of their presidential votes. After harming Democrats for so long, this disgraceful feature of the system might actually hurt Trump on Election Day.

This devaluation of nonwhite votes should not be surprising, since the Electoral College was one of the Constitution’s several accommodations for slavery, and later served as a perpetuation of Jim Crow. A refresher: At the Constitutional Convention, Southern delegates opposed direct elections of presidents. Although the South contained some of the most populous states, a huge portion of the region’s population was enslaved and disenfranchised; Southerners therefore feared that direct elections would give free states a perennial edge. So Southern delegates demanded a compromise: States would vote for the president, with each state receiving the same number of votes as members of Congress. This solution gave slave states an outsize impact on the presidency, because these states had already secured extra votes in Congress by counting their enslaved residents as three-fifths of a person.

This overtly racist solution worked as intended, giving the South an unearned grasp on the presidency until the Civil War. And its advantage persisted even after the abolition of slavery, when Southern states used Jim Crow laws to disenfranchise Black residents. The Jim Crow regime continued to inflate the influence of white voters in the South—where Black Americans earned their states additional electoral votes but could not vote themselves.

Today the Electoral College still dilutes racial minorities’ political power, though in a different way: It elevates the impact of whiter states and the white voters within them. Nonwhite voters are distributed throughout the country in a pattern that happens to be inefficient for presidential candidates relying on their votes. In other words, racial minorities disproportionately live in states that are already guaranteed to vote for Harris or Biden in November—that is, fewer of them live in swing states.

A few examples drawn from current demographic data: Latinos make up about 40 percent of California’s population, and Trump appears to be making inroads with them there. But there is no chance that this trend will flip the heavily Democratic state toward Trump, so eroding Harris’ edge with California Latinos will not help him reach the White House. Latinos also make up about 40 percent of Texas’ population, yet Trump is essentially certain to carry the Lone Star State already, so additional votes from Texas Latinos won’t matter to him. Florida is an increasingly diverse state, and Trump may win an eye-popping number of nonwhite votes there. But he is overwhelmingly likely to win it no matter what, so this feat would not clinch him a second term.

The same dynamic applies to smaller states. Black voters make up about 36 percent of Mississippi’s population, for instance, but winning over some of these voters will not boost Trump’s overall chance of victory because he already has Mississippi locked down. Conversely, Asian Americans make up 10 percent of New Jersey and Washington state, but Harris has both in the bag. So it will be irrelevant to Trump if these Asian American voters defect to him.

There are, of course, exceptions to this dynamic: Arizona, Georgia, Nevada, and North Carolina all have large numbers of nonwhite voters, and all are swing states. But Harris’ clearest path to victory remains a sweep of the famed tipping-point states—Michigan, Pennsylvania, and Wisconsin—which historically vote the same way. If she carries this so-called blue wall, along with the single electoral vote of Omaha, Nebraska, she’ll win the whole election. Pennsylvania in particular has emerged as the probable tipping-point state. That’s why Harris and Tim Walz, her vice presidential nominee, are spending the final weeks with a campaign blitz in the Rust Belt. Current polling suggests that winning these states is pretty much mandatory. Any victories in the more diverse Sun Belt will probably be a bonus.

Adjusting each candidate’s vote share by demographic, as FiveThirtyEight allows, illustrates the problem for Republicans: Bumping up minority support for Trump does shockingly little to improve his odds. My former colleague Matt Yglesias has pointed out, for example, that Trump could improve 10 points with Hispanic voters and 20 points with Black Americans—but still lose to Harris if he does just 2 points worse with white people. Even if Trump does exponentially better with Hispanic voters than he did in 2020, he’ll lose if Harris shaves off a few points among whites. The Electoral College bias is exacerbated by the fact that white people still make up a sizable majority of the country. So Republican gains among nonwhites don’t count for much, especially when they’re offset by even minor Democratic gains among whites.

This political geography has also dictated a strategy that seems to prioritize white voters. Harris’ campaign is angling to capture as many suburban voters as possible, in part by appealing to former Republicans and Trump-loathing moderates. It is taking advantage of a shift away from the GOP among college-educated voters, who are mostly white, to make up for losses among racial minorities. Harris’ appearances with Liz Cheney; her much-touted endorsements from other Republicans; her reassurances to moderates and pivots toward the center: These tactics are all aimed at the white-skewing, higher-educated Americans who may well decide the election. In Michigan, Pennsylvania, and Wisconsin especially, the suburbs are rapidly growing and filling up with these voters. Nabbing them could easily counterbalance a national erosion of support from nonwhites. (Notably, winning whites in this region is exactly how Trump carried the Electoral College in 2016.)

In fairness, the three “blue wall” states are not entirely homogeneous; their urban centers are rightly celebrated for their diversity. Each of them is, however, whiter than most states, and whiter than the country overall: Michigan and Pennsylvania are just under 75 percent white, while Wisconsin is nearly 80 percent white. Harris’ campaign has concluded, rationally, that reaching the White House requires catering to persuadable white voters in these states. Trump’s appeals to racial minorities, by contrast, seem destined to run up his margin of “wasted votes”—votes that don’t help grow his margin in the Electoral College. Yes, he may receive far more support from racial minorities than any modern Republican candidate. But because of the way these voters are geographically distributed, this achievement may count for nothing.

This racial bias, while egregious, is not even the best reason to get rid of the Electoral College. The system’s more obvious and fundamental problem is that it weighs some votes more than others in the first place. For example, Wyoming has one electoral vote for every 192,284 residents, while California has one for every 732,190 residents—meaning that a presidential vote in Wyoming counts almost four times as much as one in California. That the institution’s skew favors white voters is merely an additional reason to abolish it. It also reveals the dead hand of enslavers strangling democracy today. Slave states wanted a system that would enshrine white supremacy in the election of a president. They have succeeded.

Harris is still favored to win the popular vote, even if she loses the Electoral College. In that critical respect, the institution still helps Trump more than it hurts him. If he cannot shore up robust support among white Americans, he will soon learn the limits of nonwhite support in a system designed to devalue their votes.