Running Off the Competition

“You can’t confuse votes with love. They’re not the same thing.”

Winona Ryder as Vinni Restiano;
Show Me a Hero, Episode 5
written by David Simon & William F. Zorzi

Rep. Arthur “Doc” Corvese has a bill, H6622, that would establish a “preliminary election” in place of the current partisan primaries. This method of election is more typically known as “top-two” or on Wikipedia, the nonpartisan blanket primary (though it is neither nonpartisan nor a primary). Because of most American states’ hostility to political parties, this is effectively an American version of a two-round election.

On Jan. 27, the House of Representatives held a hearing where former Mayor of Providence Joe Paolino and former chair and executive director of the Democratic Party Guy Dufault spoke at length about why this was a good reform.

At the heart of this reform seem to be four spoken and unspoken points:

1. This ensures that candidates are elected with the support of the majority of voters.
2. This ensures the “best” candidate wins.
3. More people will turn out for the primaries
4. It will keep progressives from winning primaries. (This is the unspoken point by the bill’s proponents.)

So let’s take ’em point by point.

Mandating 50%+1 victories doesn’t result in voter support

Paolino spoke at length about this, at one point saying, “quite frankly, I would’ve been able to get more things done if I had over 50% of the people supporting me in the City of Providence.”

To which I have to ask: would you?

Paolino is equating winning a majority of the vote with support, and in a two-way race, you can’t make that determination.

What Paolino is referring to is “mandate politics” – the idea that winning an election with a majority conveys a mandate from the voters. It is most frequently invoked at the presidential level, but evidence for its existence is thin. However, in our current system, where a candidate can win with a plurality and there are often alternatives for voters dissatisfied with the leading parties candidates, someone who is able to win a majority of the vote is at least able to say that and have it mean that voters did it of their own free will.

Under top-two and (all two-round elections), voters don’t have freedom of choice. Two-round elections are designed to produce a majority, by their very binary nature. There are no other options for consideration. Because of that, unless it’s an exact tie, one candidate always ends up with majority and one candidate always gets a minority of votes.

That’s not a mandate, it’s an exacerbation of the usual complaint with American elections: that they’re lesser-of-two-evils decisions. Indeed, we see that in famous examples of this problem; in 1991 in Louisiana where corrupt former governor Edwin Edwards defeated former KKK Grand Wizard David Duke in the second round (memorable bumper stickers “Vote for the Crook: It’s Important), and in 2002 in France when Jacques Chirac (under corruption investigations) defeated the far-right Jean-Marie Le Pen (memorable poster “Vote for the Crook, Not the Fascist”). Plenty of two-round elections result in this; candidates winning with large majorities but who are ultimately disliked by majorities of voters.

So just because you mandate that elections can only be won with majorities does not mean that winning candidates actually have majority support.

This is not a method for picking the “best” candidates

Guy Dufault focused much of his testimony on how this would lead to only the “best” candidates facing off. But we can look at the examples above and see that’s not really true.

Math-inclined electoral reformers are obsessed with this idea of a “Condorcet winner” – the concept (named after the French mathematician the Marquis de Condorcet) basically posits that in most elections, there is one candidate who will beat all the others were the election held one-on-one. So in Paolino’s 1984 special election victory, we would call him a Condorcet winner if he would’ve beaten either Fred Lippitt, Keven McKenna, or Emmanuel Torti were he facing each alone for the mayoralty.

Electoral methods that provide for such a winner this are called “Condorcet methods” (and yes, it’s possible to not have that winner, but it’s rare). The problem is that two-round elections are not Condorcet methods.

Go to the 1991 Louisiana gubernatorial race. It’s probably fair to say that incumbent governor Buddy Roemer was the “best” candidate in the race and likely Condorcet winner (meaning he would’ve beaten either Duke or Edwards in a one-on-one). Unfortunately, Roemer abandoned the Democratic Party and switched to the Republican Party shortly before the election, resulting a split among Republican voters, who had a majority of first round votes. Duke and Edwards each received close to a third of the vote in the first round, resulting in a second round featuring two very polarizing candidates.

The result was that Republicans, who had a majority of voters in the first round, failed to win the election, and instead one of America’s most famously corrupt politicians was re-elected governor.

The mistake Dufault makes is that he has confused “winning votes” with being “best” – but they’re not the same thing. In many first round elections, the “best” candidate does not triumph, but rather the candidates with the most passionate followings.

Consider a hypothetical where there is an outright fascist and an outright communist (to pick two extremes), who both oppose democratic government, running in a race. They are not popular, but they have followings of about 15% to 20% of voters. Since they announce first, and there are a large number of undecideds, something like five pro-democracy candidates get into the race. All them are sober, serious people who see a chance to win against an unpopular opponent and poll roughly the same, about 10% to 15% of the vote. They’re all projected to lose, but they also all have a good chance at winning, so none of them bend to pressure to drop out. They all lose, and in the general, the electorate has no choice but to vote for one of the anti-democracy candidates.

This is a coordination problem, and it’s pretty common in elections. You might know it best from the 2016 Republican presidential primaries. Despite the party’s hostility to Trump, Republicans were unable to coordinate on a candidate they preferred, and Trump won the nomination without the majority support of the party. You don’t have to be the best candidate to win an election with more than two candidates, you just have to maintain a dedicated section of support while your opponents flail around you.

More people may turnout, but how many and does it matter?

One of the most effective ways we have of turning out voters is to have competitive elections, and this is true of primary elections as well as general elections. There is some evidence that there are modest turnout effects caused by the top-two system.

However, Dufault makes an assertion (at around minute 22 of the hearing recording): the adoption of top-two in California and Washington (in 2012 and 2008, respectively) has led to increased voter turnout in those states’ primaries.

And that’s not really true. In fact, turnout collapsed in both primaries (2012 and 2014) following its adoption in California, and only bounced back up in 2016 to levels similar to those preceding the reform. Washington state, which has much higher turnout versus other states mostly due to its robust vote-by-mail system, also didn’t see real impacts. Instead turnout bounces around quite incredibly for primary elections (ranging from 2020’s high of 54% in the primary to 2017’s low of 24.37%).

This won’t necessarily help moderates win

There is an unspoken part of this, which is that leaders in the Democratic Party are growing more fearful of progressive challengers as the Democratic Party as a whole moves leftwards. Corvese, often fingered as the Ur-DINO in the Democratic caucus, might be particularly worried about it. Paolino, who often seems to be on what is today the right wing of the party, also likely is fretting about the rise of a strong progressive faction within the Party, as it has the potential to put people hostile to both his political and business interests in positions of power.

Many of top-two’s proponents desire primary results that produce candidate of a moderate persuasion, and it’s often listed as a way to accomplish that. The assumption is also likely that a moderate incumbent who might otherwise be defeated in a Democratic primary could go on to win a general election against their progressive primary opponent.

There’s quite a lot of research which finds mixed effects on moderation. That said, a New America report finds that top-two wasn’t particularly effective at arresting polarization in California. Why not? Well, there weren’t enough same-party general elections, political actors adjusted to the new rules, in the same-party general elections that did occur the opposition party’s voters abstained, voters were unable to tell who was moderate and who was extreme without distinguishing party labels, and there really aren’t that many moderate voters and the belief that independents are moderates is wrong.

All of these will come into play in Rhode Island. My Republican followers on Twitter often point this out to me: that they don’t consider the Democratic primary for governor a battle between conservative and progressive and moderate factions, but rather a dizzying array of liberals they don’t really distinguish between. You can call this “tribalism” – but it’s also how voters behave. And the introduction of top-two doesn’t change that behavior, we’ll still just have two parties.

If the leaders of the Democratic Party think they can rely on inertia to save them from a progressive wave, they may end up sorely mistaken. What top-two may instead bring about are generals that result in progressive victories as Republican voters stay home and left-wing independents who sit out primaries turnout.

At its heart, this fundamentally misunderstands party primaries

It feels weird to write the above header, since so many of the people advocating for this change are current or former party leaders, but in suggesting this, it seems like they don’t understand what parties and primaries are for. The purpose of primaries is to select a party nominee. They were invented to stymie the power of bosses at conventions. They are not supposed to act as first rounds in two-round elections.

To be fair, very few people seem to think about things in this way. Much of the current fight over district committees and party endorsements also comes out of the mistaken belief that primaries are supposed to work as a de facto first round played on neutral ground; instead of a party’s selection process in which the party itself is an actor!

That misunderstanding of the purpose of primaries has led to internecine squabbling over every minor adjustment to procedure that results in just a set of unrelenting headaches and negative press for the Democratic Party. And Corvese’s bill isn’t going to lessen those, because it doesn’t remove a bunch of other provisions that it seemingly now complicates. It also can punish a party for being successful in their recruitment; too many Democratic candidates in the first round can split the vote so severely that even a district with a majority of Democratic voters could face the prospect of two Republicans facing off in the general.

The job of running the RI Democratic Party is already a nightmare (look at the parade of people who have served as its executive director), I don’t understand why Corvese (who is the Party’s secretary) would make it worse. And that’s basically what top-two does, it makes our primaries worse and makes operating as a political party worse, without guaranteeing any of the benefits that its supporters say it will.

There are better solutions out there, but I honestly don’t think our party leadership is willing to consider them, because those solutions mean potentially giving up power and that’s just unconscionable to them.