“Joe Biden can win back working-class white voters!”

“Bernie Sanders’s moment is here — the country is ready for Medicare-for-all!”

“Any Democrat will beat Trump!”

“All the Democrats have moved so far left that Trump is going to win again!”

For the next year or so, you’re going to hear people say things like this all. the. time.

People are obsessed with the question of which candidates can or can’t win and why. But when people talk about electability, they often conflate issues such as ideology, identity/demographics and “likability.” This is the second installment in a series that tries to untangle those issues and to determine which candidates really do perform best according to these different metrics. The first piece, which talks about how well candidates performed in the past, is here. Today, we’ll tackle ideology, specifically whether there is a specific mix of policy positions that a candidate can take to maximize their chances of winning the presidency.

Ideology is a much more complicated issue than our last subject: Defining a politician’s beliefs and pinning down what the electorate as a whole believes are less straightforward than measuring their margin of victory in previous races. When people talk about ideology, they often unconsciously smuggle complicated theories of how politics works into seemingly simple statements. So I’m going to talk through a few assertions about the 2020 election that will surely become cliches (if they aren’t already) and take apart what each one gets right or wrong.

Scenario One: Amy Klobuchar and Beto O’Rourke are the most electable 2020 candidates because they’re the most moderate.

What this gets right:

This line of logic is based on a intuitive view of ideology: that Republicans and Democrats can be placed in a one-dimensional line from most liberal to most conservative and that the politicians who are closer to the center of that spectrum have a better chance of winning over independents and swing voters.

It’s not hard to think up examples that support this. Barry Goldwater and George McGovern, who are often seen as far-right and far-left, respectively, lost in landslides in 1964 and 1972, respectively. Bill Clinton and Dwight Eisenhower, who seemed more moderate at various points in their presidencies, won handily.

Ideology isn’t the only thing that matters in this view — very liberal or conservative presidents can win under the right conditions. Ronald Reagan was very conservative, and he won partially because Americans thought Jimmy Carter was performing poorly. Similarly, Barack Obama ran as a liberal and he won partially because George W. Bush had presided over a recession and unpopular wars. This moderation-is-good theory would have predicted Reagan and Obama’s wins based on the state of the country — but it would stipulate that a more moderate candidate would have won by more.

This statement also correctly identifies Klobuchar and O’Rourke as relative moderates — that is, they’re sufficiently in step with the party to win the primaries but they’re somewhere to the right of the other potential primary candidates. They’re both pro-choice, pro-LGBT and generally in favor of expanding government in an effort to solve social problems.

That being said, Klobuchar voted for half of Trump’s initial Cabinet nominees (more than Kirsten Gillibrand, Cory Booker, Sanders, Kamala D. Harris or Eliabeth Warren), hasn’t been totally on board with abolishing the Immigration and Customs Enforcement Agency and didn’t support Sanders’s push for Medicare-for-all. In my colleague George F. Will’s words, she’s “liberal enough to soothe liberals without annoying anyone else.”

And O’Rourke, despite becoming a sort of liberal folk hero in 2018, hasn’t been purely progressive, either. He has been vague on Medicare-for-all and was part of the New Democrats (not the Progressive Caucus) in the House. His House DW-NOMINATE score (a political science metric that places legislators on a left-right scale based on how often they vote together) also puts him roughly in the most liberal third of the chamber.

Saying that candidates such as Klobuchar and O’Rourke are more electable than their left-leaning competitors rests on an important assumption: that true moderates, such as Sen. Joe Manchin (D-W.Va.), who are conservative by the standards of their party but stand in the middle of that national spectrum, couldn’t (and shouldn’t) win a Democratic Party primary. This statement is sort of like a spin on the Buckley Rule — arguing that Democrats should support the most left, viable candidate who can win.

What this gets wrong:

The problem with this line of thinking is that we don’t really know that the most moderate candidate is most electable. We’ve only had 18 presidential elections in the modern era (post-World War II). That’s not a lot of data to work with. Moreover, some recent political science research suggests that, after controlling for other factors, candidates who were ideologically extreme lost at most one to two points of vote share due to their ideology. If that research is correct (and reasonable people could make a different case), that’s not a huge bonus. If you’re a progressive Democrat who believes Trump is going to lose no matter what, you might not feel any need to compromise on an O’Rourke or a Klobuchar in the primary.

The other problem with this mode of thinking is that ideology isn’t really a one-dimensional left-right scale. The liberal vs. conservative dichotomy is a useful shorthand in some situations, but politicians can position themselves differently on social issues, identity/culture issues and economic issues. Emphasis also matters — Hillary Clinton and Sanders arguably differed more on emphasis and style than on substance. Calling a candidate liberal or moderate arguably oversimplifies ideology.

Scenario Two: Elizabeth Warren and Bernie Sanders are the most electable because they have bold, progressive economic ideas.

What this gets right:

This statement gets two big things right. First, it implies that ideology is more than a one-dimensional left-right space. Second, it gets at a big lesson from the 2016 election — that new approaches sometimes work. We can illustrate how using Trump’s campaign.

During the 2016 election, voters generally, and perhaps counterintuitively, thought Trump was closer to the political center than Clinton. In a Gallup poll, 35 percent of respondents said that Trump was somewhat or a lot more conservative than they were, and 49 percent of voters said Clinton was somewhat or a lot more liberal than they were. Moreover, people thought Trump was more moderate than anyone the GOP had nominated in decades.

But Trump definitely wasn’t one of those patrician moderate Republicans from New England. He combined an intense focus on immigration and race, departures from conservative orthodoxy on Social Security, mixed messaging on LGBT issues and promises to conservatives on judges and Mike Pence as vice president. He adopted conservative positions on some issues and pulled in a different direction on others. That let him look comparatively moderate on the one-dimensional left-right scale when he was actually chucking major tenets of movement conservatism and trying out a categorically different political approach.

Trump’s new method — which some have dubbed conservative populism — allowed him to win both the Republican primaries and the general election. Obviously not everyone loves it, but trying something different clearly worked for him.

Sanders and Warren supporters might claim that their candidate would break the mold in a different way. Sanders or Warren could focus on class-based appeals, advocate radical progressive economic ideas, balance that with traditional Democratic positions on race and blow through the primary and general elections. Trump has moved to the right on economics since becoming president (his major legislative accomplishment is a very conservative tax reform bill), so he may be vulnerable to exactly that sort of strategy.

What this gets wrong:

The problem with this argument is that we don’t know what will land with the public ahead of time. Most experts saw Trump’s new strategy and policies and promptly concluded he’d lose both the primary and the general elections. He didn’t. And candidates who break from the old mold (such as Goldwater) sometimes lose big.

Maybe Sanders and Warren can leverage popular progressive policies to win the next election. Or maybe not — maybe Trump will effectively cast the policies as too expensive or successfully shift the issues to something more favorable to him. Or maybe Howard Schultz will run hard in response to a Warren or Sanders nomination and end up splitting the Democratic base.

My point here isn’t to evaluate the merits of the policies. It’s to say that we don’t really know if something new is going to work out before someone tries it.

Scenario Three: Joe Biden is the real moderate — look at his clear record!

What this statement gets totally sideways:

This might sound a lot like the first statement, but Biden’s “moderation” is distinct from O’Rourke’s or Klobuchar’s. And this one isn’t really right or wrong — it sort of gets at Biden from an awkward, sideways angle.

In 2016, the political scientist Julia Azari pointed out that many recent presidents and presidential candidates were governors, vice presidents or people with little legislative experience. Holding those positions allowed them to talk broadly while evading thorny specifics. Vague candidates are adaptable candidates.

For the past 10 years, Biden has been vice president (from 2009 to 2017) and then a private citizen (2017 to now). That means Biden might be able to take advantage of this vagueness. Granted, he has a lot of baggage that he’ll have to deal with in both the primaries and the general election (if he makes it there), so he’s not completely flexible. But it is possible to imagine a reinvented, post-vice-presidency Biden gesturing at progressive ideals while strategically maintaining flexibility on some specifics.

Harris might also benefit from this sort of vagueness. Harris is new to national politics — she was elected senator only in 2016, so she has some ability to talk about her ideas without having to defend a long record of contradictory votes. She’s not totally baggage-free (not all progressives love Harris’s record as California attorney general), but in a general election she might be able to broadly talk about some policy goals without getting snagged on every specific.

Scenario Four: Every top-tier Democrat is electable, so Democrats don’t have to worry.

What this gets right:

This statement is unintentionally correct. High-level factors that neither candidate can control (such as economic growth and public opinion of the president) set the stage for presidential elections. If a really “electable” candidate challenges the president when the economy is growing and people are happy, that candidate will probably still lose. Conversely, if a president’s approval rating has dropped through the floor and a recession just happened, then the party that’s out of power can probably take the White House as long as it doesn’t nominate someone truly awful.

So far none of the top-tier Democrats seem to have truly disqualifying baggage. So, under the right conditions, any of them could beat Trump — regardless of ideology.

What this gets wrong:

Every Democrat is electable, but they do have to worry. If every Democrat is electable, it’s also true that none of them is a safe bet.

We don’t know what political conditions will look like in 2020. Maybe a recession will be on, or maybe the economy will be going strong. Trump seems unwilling to abandon his base-focused strategy, but there’s no law of physics that says he couldn’t try something new. The political scientist Phil Klinkner recently put it well

Every top-tier Democratic candidate may be electable, but they’re also all at the mercy of forces behind their control.

Read more: