Donald Trump has between a 25 percent and 44 percent chance of becoming president—I’ll pause while you fix yourself a stiff drink—although he doesn’t crack 45 percent of the vote in prominent poll aggregators. One reason for this is an unusually strong collective showing by the Libertarian candidate Gary Johnson and the Green Party’s Jill Stein. Millennial voters, in particular, have the sound judgment to solidly disapprove of Trump, but are disproportionately attracted to third-party candidates. I think voters of any age who don’t want Trump to be president, but are considering a vote for Johnson and Stein, are making a mistake. But voters should not be faced with this dilemma in the first place.   

Given the American electoral system as it actually exists, it’s an easy call for people on the left of the American political spectrum. Gary Johnson had a terrible record as governor of New Mexico and has a mostly right-wing agenda. Jill Stein is, I suppose, to the left of Hillary Clinton, although she’s such an ill-informed crank I’m not sure how meaningful the description is.

But the bigger problem is that supporting any Green candidate for president is all downside and no upside. The only possible effect Stein could have on the presidential election is to attract enough votes to allow Donald Trump to win, which would have horrible material consequences for countless important issues: civil rights and liberties, economic equality, the environment, women’s reproductive freedom, and on and on.

The idea that without third-party challenges the major parties will just take their supporters entirely for granted, and hence that third parties are necessary for major change to occur, sounds plausible in theory but is egregiously wrong in practice. Conservatives didn’t capture the Republican Party by mounting vanity general election candidates against the establishment. The Social Security Act of 1935, the Civil Rights Act of 1964, and the national right to same-sex marriage were not the result of mass defections from the Democratic Party and effective third-party challenges from the left. Presidents inevitably disappoint their allies, and as Bernie Sanders recently noted, politics has to continue after Election Day. But some presidents are amenable to pressure from progressive groups and some aren’t, and activists who know what they’re doing do what they can to elect the former. In 2016, Hillary Clinton is in the former category and Donald Trump the latter.

Still, why should the consequences of voting your first choice be so potentially perverse? It is a function of the electoral system. To get a state’s Electoral College votes, a candidate does not need a majority, only one more vote than the runner-up. These simple plurality electoral systems have become increasingly discredited among liberal democracies, for good reason.

Plurality systems effectively ignore highly pertinent information. They treat all voters as having no preference between the candidates they don’t mark as their first choice, when we know that in most cases that isn’t true. (A person voting for Jill Stein will, in all likelihood, prefer Hillary Clinton to Donald Trump.) Because of strategic voting (e.g., liberals of all stripes voting to keep Trump out of office), plurality elections tend to produce acceptable, majority-supported winners—but not always. Electoral systems that take this information into account—and hence prevent the spoiler effect of third-party candidates—are available and could be instituted.

One common proposal, advanced by groups such as Fairvote, is “instant runoff (IRV)” or “ranked candidate” voting. Under this system, voters would rank as many candidates as they choose in order of preference. If no candidate gets a majority of first-choice votes, the candidate with the fewest first-choice votes is disqualified. That candidate’s supporters’ second choices are added to the remaining candidates until there is a winner. This system generally avoids classic spoiler situations like 2000, in which people who voted for Ralph Nader ended up throwing the election to George W. Bush, even though most Nader voters would have preferred Al Gore.

Another system, preferred by some scholars of electoral systems, is approval voting. Approval voting allows voters to vote for as many candidates on the ballot as they wish. The system has two obvious advantages over IRV: It is both simpler and more reliable. It allows voters who disagree with me about the value of voting for third parties to express themselves, while eliminating the problem of this expression leading to the election of a voter’s least preferred candidate. “Third parties would not win unless they really built up more support than the more established parties,” observes Jameson Quinn, a board member of the election-reform nonprofit Electology.org, “but in the meantime, they’d have a positive impact on the debate, not just a negative impact of occasional spoiled elections.”

While approval voting is preferable to IRV, either would be major improvements over plurality voting. But what are the chances that a decent voting system could be used to determine the results of the Electoral College?

One major barrier is the unfortunate decision of the framers to make presidential elections a state matter, rather than a federal one. Article II gives the states the power to determine the electors that formally choose the president. As a result, even a Democratic Congress could not pass legislation mandating approval voting for the Electoral College. At most, Congress could incentivize states to change their systems. Eliminating plurality voting for president will require either action by state legislatures or a constitutional amendment.

Both options would be extremely cumbersome, and meeting Article V’s standard for ratifying a constitutional amendment is nearly impossible. State-by-state reform will not be easy—but we shouldn’t assume it is impossible.

Unlike, say, voter ID laws or early voting periods, plurality voting does not confer an obvious inherent advantage on one party or the other. Granting that the Libertarian Party cross-cuts political cleavages and does not draw from the Republican vote to the same extent that the Green Party draws from the Democratic vote, it looks like Gary Johnson’s may be leading the most successful small-party challenge in 2016—and largely to the detriment of Republicans. And while the South is often a barrier to progressive reforms, if anything non-plurality voting is more prominent in jurisdictions below the Mason-Dixon line.  

Electoral reform won’t be easy, but 2016 shows it is a necessity. We can only hope that it will happen before a major Electoral College malfunction, and not in response to one.