Why America’s 2-party system is on a collision course with our constitutional democracy

Andrew Harrer-Pool/Getty Images
This post is part of Polyarchy, an independent blog produced by the political reform program at New America, a Washington think tank devoted to developing new ideas and new voices.

There was a time, several decades ago, when America’s two-party system was praised for its moderation. Unlike European parliamentary democracies where “dogmatic ideological parties” of Europe thrived, America’s winner-take-all electoral system seemed to reward and therefore encourage parties and candidates with broad national appeal. No party, it was argued, could simply give up on half of the electorate. Similarly, no party could convincingly win a majority by putting forward extremist anti-system candidates far outside the mainstream.

Obviously something has gone wrong with this theory. Instead of being rejected as outside the mainstream, Donald Trump, an extremist anti-system candidate, simply redefined what “mainstream” is for almost half of the electorate.

And today, both American parties regularly forsake about half the electorate. Or even more than half, really.

Consider some basic numbers: Trump was the choice of 14 million people who voted in the Republican primaries. But in a nation where 230.6 million Americans are eligible to vote, that’s 6 percent of eligible voters. In the 2017 German election, the far-right populist Alternative for Germany (AfD) won 5.9 million votes. In a nation of 61.5 eligible voters, that’s almost 10 percent.

In short, when voters in both countries were given the full range of options, Donald Trump was less popular in the United States than the AfD was in Germany.

But in the German system, AfD can be kept out of power by other parties forming a coalition. In the United States, Trump’s 6 percent support gave him a major party’s nomination, which gave him instant legitimacy. And because he was a Republican candidate and because he wasn’t Hillary Clinton, 63 million Americans cast a vote for him — enough to catapult him to the presidency.

Sixty-three million is a lot. But that’s also just 27 percent of eligible voters nationwide. Likewise, 63 million of Americans voted for to send Republicans to the US House, also just 27 percent of the eligible voters. In many cases, these were not even affirmative votes for Republicans, but votes against Democrats.

I raise these numbers to point out that, contrary to claims that American political parties have to appeal broadly to win, they only need to win a quarter of the voting-age population to gain unified control of government in Washington, and their presidential nominee needs to win far less than that. Lest you think I’m picking on Republicans, the same was true (roughly) of Democrats in 2008.

Part of this is because unlike in Germany, where voter turnout hovers closer to 80 percent, American voter turnout is usually in the mid-50s in presidential elections, and closer to 40 percent in midterms (an international laggard). Many US voters don’t bother to vote because neither of the two parties appeals to them, or because they live in a safe state where their vote doesn’t matter, or because by comparative standards, there are significant hurdles to voting in the United States (such as more complicated registration, or voting being on a workday instead of on a weekend).

In short, there is nothing structural about a two-party system that guarantees moderate parties that have to appeal broadly. We just got lucky. Well, sort of — the past wasn’t so great either.

The problem with “moderate” parties

Back in 1950, when both major parties were broad and moderate with overlapping appeals, many of America’s leading political scientists wrote a report in which they bemoaned this state of affairs.

In a report, “Toward a More Responsible Two-Party System,” they saw two national parties that were but loose confederations of state and local parties, incapable of bringing forward coherent programs to the voters and carrying them out when they got into power.

Thus, instead of elections giving voters a meaningful choice between well-thought-out alternatives to pressing national political problems, voters encountered a muddle. What happened in Washington seemed to have little connection to what happened at the ballot box.

This seemed to the authors an untenable situation, one that would fail to generate legitimate responses to emerging domestic and foreign problems whose scale and scope demanded strong, responsible national parties. To accomplish this, they recommended a massive centralization of the two parties so that national party leaders develop carefully researched policy alternatives, which voters could then approve authoritatively through a simple majority vote.

And if the American political parties failed to heed their advice? The authors issued a dire warning:

If the two parties do not develop alternative programs that can be executed, the voter’s frustration and the mounting ambiguities of national policy might also set in motion more extreme tendencies to the political left and the political right. This, again, would represent a condition to which neither our political institutions nor our civic habits are adapted. Once a deep political cleavage develops between opposing groups, each group naturally works to keep it deep. Such groups may gravitate beyond the confines of the American system of government and its democratic institutions.

Assuming a survival of the two-party system in form though not in spirit, even if only one of the diametrically opposite parties comes to flirt with unconstitutional means and ends, the consequences would be serious. For then the constitution-minded electorate would be virtually reduced to a one-party system with no practical alternative to holding to the “safe” party at all cost.

In line with the political scientists’ advice, the two parties did indeed develop “alternative programs.” Without a doubt, we now have the clear choices the report’s authors recommended.

The problem, it turns out, was with the “can be executed” part. More coherent, non-overlapping parties did not give us “executable programs.” Instead, because our system of checks and balances and decentralized authority was designed specifically to prevent against the “tyranny of the majority,” polarized parties gave us gridlock, a steady erosion of procedural consensus, and mounting frustration.

Parties and candidates then channeled this frustration into increasingly bold and exaggerated promises about how they would fix that dysfunction. In particular, one party, the Republicans, spent the past decade with egregiously and cynically un-implementable promises about shrinking the federal government and repealing Obamacare, when they should have known better. That party is now more than “flirting” with unconstitutional means and ends. And the consequences are indeed serious.

While the prediction of the American Political Science Association’s report is spookily prescient, I’m less optimistic than the APSA committee that the American voters will pick the “safe” party, because I’m less confident that the American electorate is as “Constitution-minded” as the authors claimed.

Note that this report was published in 1950. That was before social scientists made two important discoveries about the American electorate — that Americans are not all that tolerant, and that partisanship is “a pervasive dynamic force shaping citizens’ perceptions of, and reactions to, the political world.”

Instead, scholars discovered that there was no deep, abiding support for civil liberties, tolerance, and procedural rule following among the electorate. Tolerance and respect for democracy instead depended on political elites, who were the keepers and teachers of the democratic traditions.

And if partisanship is so pervasive, and Americans are not deeply attached to tolerance and procedural “norms,” and we have a two-party system, and voters have deep frustration with how things are going, then it’s not a surprise how we got to this current moment, when the future stability of our democratic constitutional order is an open question.

Why “better elites” probably can’t save us now

So what do we do? An obvious suggestion is to get better elites, who respect tolerance and dissent and our liberal-democratic traditions. In particular, we can and should urge Republicans to push back against Trump’s reckless tearing down of our institutions.

But the problem with this strategy is that, as Arizona Sen. Jeff Flake has made clear, this appears to be a political career-ending strategy. In a highly partisan two-party system, in which Trump and his acolytes in conservative media set the talking points for the party, dissent is hard. Flake can stand up all he wants and decry Trump, but without a following, he’s nowhere politically.

In a multi-party proportional voting system, of course, Flake and other truly “Constitution-minded” conservatives could split with Republicans, form their own party, and still get elected, and probably have a pivotal role to play in most coalition government, even if they only captured 10 percent of the overall vote.

But of course, in a proportional voting system, Trumpist populism never would have taken over a major party in the first place. Instead, it would have had its own party all along. Obviously, such party could have caused some trouble — but probably not as much trouble as it is now causing as it takes over the Republican Party.

Why the two-party system is at the root of the problem

In the debate about whether democracy is in decline in the West, there’s some important cross-national variance. In a response to the widely discussed democratic decline findings of Roberto Foa and Yascha Mounk, Pippa Norris compared support for democracy across Western democracies and found whatever cohort decline existed, it was largely limited to Anglo-American democracies, which tend toward two-party systems.

By contrast, in parliamentary democracies with proportional voting, there has been no consistent erosion in support for democracy. As Norris argues by way of explanation, “parliamentary democracies with PR elections and stable multiparty coalition governments, typical of the Nordic region, generate a broader consensus about welfare policies addressing inequality, exclusion, and social justice, and this avoids the adversarial winner-take-all divisive politics and social inequality more characteristic of majoritarian systems.”

But there’s another piece of the puzzle that is relevant here.

A recent report I wrote with Larry Diamond and Joe Goldman, “Follow the Leader,” looked into Americans attitudes toward democracy and its alternatives. As part of the report, we broke down voters based on their attitudes on cultural issues and economic issues, and then looked at how different combinations of attitudes corresponded to attitudes toward democracy.

Those who were most consistently liberal across both dimensions were the biggest boosters of liberal democracy. But note that the highest rate of anti-democratic sentiments came from those in the “off” dimensions, individuals who hold mixes of views that neither party represents, or those who fell in the middle. (For a distribution of where voters fall in this mix, See figure 2 in my previous Voter Study Group report.)

We also looked at similar attitudes based on 2012-2016 voting pattern and found some notable patterns. Obama-Trump voters had among the highest levels of anti-democratic sentiments, as did voters who went from not voting for a major-party candidate in 2012 to voting for Trump in 2015. Together, these Obama-Trump and other-Trump voters make up 9 percent of the electorate, and almost half of them think a strong leader who doesn’t have to bother with Congress or elections is a good thing. (The “other” category includes third-party candidates or not voting at all.)

What explains this pattern? Well, to start: Some Americans are engaged in politics and some aren’t. Those who are engaged, and follow politics more closely, tend to be more attached to democratic and procedural norms, and tend to be more ideologically “coherent.” That is, they know what goes with what, at least as the two parties have defined it. And most of all, they tend to be loyal partisans.

Those who aren’t as engaged in politics have less attachment to the parties and are more adrift from the so-called “ideological coherence” of the two parties. It’s not surprising they would think democracy isn’t a great system and have weaker attachments to it. It sure doesn’t seem to represent them well. And they’re not socialized into its norms.

Causation is tricky here. Presumably, people who are less educated about politics are less likely to be loyal partisans and therefore know “what goes with what.” But presumably if they feel unrepresented, they might also think: Why bother to engage in the first place? So there’s a bit of a reinforcing feedback loop here, though I’d put the most emphasis on lack of engagement and education coming first.

What does this suggest as a policy response? The conventional wisdom is that we should spend more on civic education and do what we can to expand voting participation. All this sounds nice, and more civic education is always a good thing.

But we have to be careful here. If we’re asking for more engagement from people who feel unconnected to the parties and hold negative views of democracy, we run two risks.

The first risk is that we’re bringing more people with anti-democratic views into the electorate, which further increases the electoral power of an anti-system populist candidate who promises to blow it all up.

Maybe they’d eventually become socialized into the party system. But even so, this then raises a second problem — that we’d likely just produce more strong partisan voters. After all, the more politically engaged people become, the more they become strong partisans. This makes sense, since if you’re going to become involved in politics, you must think it matters who’s in charge. Once you pick a team, you tend to start engaging in motivated reasoning about politics, disregarding information that undermines your side. So a more engaged electorate becomes a more strongly partisan electorate. And since hyperpartisanship is an obvious danger to our political system, more public participation doesn’t solve that problem.

Do we then want an electorate that is less engaged and less well-educated, and in which the political parties are incoherent overlapping coalitions that don’t stand for much at the national level? Sure, it made for consensus-oriented moderate politics at the national level. But then we’re back at the problems the writers of the 1950 APSA report identified, in which voters didn’t really have much meaningful choice. And more significantly, whatever bipartisan consensus existed only existed because both parties took civil rights off the agenda. There’s also that pesky arrow-of-time problem, which only goes in one direction.

Why Democrats winning more elections won’t solve the problem — and might make it worse

This points to an even larger problem, one that won’t simply be solved by Democrats winning more elections as the party of the “constitutionally minded” (as the APSA report predicted one party would become). In some ways, if Democrats win landslides in 2018 and 2020, the problem will be even worse, because it would likely entail the continued Democratic conversion of educated suburban former Republicans most likely to support constitutional norms and hasten a partisan cleavage that would become more and more a fight over basic democratic institutions like voting rights and a free press, leaving Republicans more and more a party of Ted Cruzes and Steve Kings, of Roy Moores and Louie Gohmerts. And then things will really get ugly.

The obvious challenge then becomes how to shift the axis of political conflict back away from a battle over the nature of America and its political institutions, and to more of a non-existential “normal politics” argument over public policy and its implementation. The answer has to involve somehow scrambling the current party system, so that being a Democrat or being a Republican is not wrapped up in these fundamental zero-sum questions about the basis of American democracy.

This is why I’m an enthusiastic supporter of efforts to expand ranked-choice voting, which are gaining steam, and of more incipient efforts to move our elections away from zero-sum winner-take-all, single-plurality winner affairs toward proportional, multi-winner elections. This would give us a more fluid party system, more in line with our constitutional design.

This means changing our electoral institutions. I recognize this is a major undertaking, and broad electoral system change is never easy. But at this point, anything less seems like taking buckets to a flood when we know the levees have broken.

There are big, important conversations to have here on the best way forward. But first, we have to admit that we have a problem. And the problem right now is that the two-party system is trapped in a doom loop that it can’t get out of on its own without significant collateral damage.

Back to top ↑