19 March 2008

SCOTUS Approves Washington Primary System

The U.S. Supreme Court concluded yesterday that the election system adopted in Washington State under ballot initiative I-872 is not constitutionally flawed.

How Washington's System and similar systems work

Washington State's political primaries under I-872 are similar to the electoral processes of Louisiana, France and Denver among other jurisdictions. Indeed, calling the first stage a primary is almost a misnomer. All candidates for the office run, giving themselves whatever party ID they feel most connected to, and the top two vote getters advance to the next round, even if they come from the same party or not major political party at all.

The only differences between these election systems, and those of Washington State are that (1) candidates who win a majority in the first round must still face voters in a second round in Washington but not in the other systems, and (2) candidates are free to adopt a party identification without political party approval (just like voters in states where registered voters declare a party identification).

Why did SCOTUS reach this conclusion?

The U.S. Supreme Court held that the system in general was valid because Washington State is not presuming to select a political party's nominee against its wishes, thereby burdening the party's freedom of association. Instead, in what I believe is a correct view on the part of the U.S. Supreme Court (and not just on an ipse dixit basis, I said so in opposition to a post by Markos when it went to oral arguments), the "primary" in this situation is simply one step of the general election process and does not claim to speak for a political party by selecting its nominee. The process is a government process, not a party process.

The U.S. Supreme Court addressed the constitutional concerns about stealth candidates using party designations by ruling that the law made it clear that these identification were subjective political identifications of the candidates only, and did not imply the approval or sanction of the political party identified with. Basically, SCOTUS said that in this system, the party label is more like saying you are a fan of a political party on your MySpace page, than it is a claim that you are a member in good standing or official candidate of that party. While SCOTUS noted that a confusing ballot design could create the impression in voters minds that a party specifically approved a candidate, implimentation of the law in a manner that places a disclaimer on a ballot, for example, could overcome this flaw and avoid the constitutional issues associated with what amounts to the trademark rights of political parties in the use of their names to market a candidate.

Colorado implications

The distinction between two part general elections and party primaries made in this case protects jurisdictions like Denver in Colorado which have two part elections.

The misappropriation of party label issues raised in the Washington case could also arise in Colorado.

In Colorado, it is possible, for example, for someone who has registered to vote as a Democrat to run for public office as a Democrat, using the petition process, even if party officials believe that the candidate is a stealth Republican and don't want that person using their party label. Indeed, one candidate in the House District 6 Democractic Party primary race has been accused by a candidate who has withdrawn from that race of doing just that this year.

This SCOTUS ruling is a strong precedent in support of the argument that this Colorado process is not facially invalid under the United States Constitution, even if it could be applied in a way the unconstitutionally burdens a political party's right to associate under the First Amendment if applied inartfully.

Notably, Colorado has closed primary elections (August 12, 2008 this year) in which only registered Democrats may vote. As a result, it is likely that a court would hold that any constitutional burden on the right to freedom of association for political parties in Colorado imposed by allowing voters registered with a political party to run for office in that party despite the objection of party officials would be insignificant and harmless. This is because, continuing the example above, even if someone who is not a "true Democrat" gets onto a Democratic primary ballot, the winner chosen entirely by voters in that jurisdiction affiliated with the Democratic Party will always, by definition, be a "true Democrat," because these affiliated voters are the ultimate, final and conclusive judges of that somewhat metaphysical question.

Why adopt a system like the one used in Washington?

Washington's system has a number of virtues.

1. It puts third party candidates on an equal footing with major party candidates procedurally. Facial neutrality based upon viewpoint as expressed in political affliation is desirable from a First Amendment perspective.

2. It avoids the tendency of a traditional two party system to disfavor political moderates. The tendency in traditional two party systems arise from the fact that one must cater to the more partisan group of voters to win a primary, before trying to win over moderate voters for the general election. This favors candidates who are slightly more moderate than the political center of their own parties, but far from the overall political center. In contrast, the Washington system has no bias in favor of, or any bias against, any part of the political spectrum.

3. It prevents independents and members of a political party that is weak in a jurisdiction from being disenfranchised at the only step of the process that matters. The reality of gerrymandering and political geography is that there is a dominant political party in the vast majority of political juridictions.

4. In situations where no candidate receives a majority, this system requires only voters whose favorite candidate does not make the second round to consider a second choice, and allows those disappointed voters to consider fewer alternatives in the second round. This greatly reduces the amount of information that voters need to cast an informed vote compared to an instant runoff voting system, where every single voter needs to consider their preferences from within the entire candidate pool in advance of the first round. A related improvement in this situation is that many voters have a hard time engaging in clear thinking about hypothetical second choices.

5. In situations where a candidate receives a majority in the first round, this system honors the fact that turnout may be greater and more informed in a second round. Notably, in the Colorado General Assembly, whose rules are typical, bills generally receive two substantive rounds of consideration on the floor (a second reading and a third reading), suggesting that there is value to allowing voters to reconsider their votes in the first round even for sophisticated voters.

6. It reduces the benefits to be obtained from gerrymandering. Under a traditional two party system, a party strengthens its position by creating safe 55-45 districts that can elect strong partisans while effectively shutting out lagging party members in the district. In the Washington State system, this kind of gerrymander is likely to produce a more moderate member of the leading party in that district.

Spoiler Effects In the Washington System

The Washington system is not immune to spoiler effects, but it is more resistent to them. In a traditional two party system in a competitive district (i.e. those in which each party has roughly equal numbers of voters who usually support them), even a weak third party candidate run can have major spoiler effects.

The classic case is Bush v. Gore in Florida in 2000. Bush ran as a Republican and drew conservative supporters as a result. Gore ran as a Democrat and drew liberal supporters as a result. But, third party candidate Ralph Nader also ran under the Green Party label and drew supporters who tended to be more liberal than the typical Democrat. Even though Nader got a small percentage of the total vote, his candidacy turned a reasonably secure Gore win into race close enough to be determined to be a Bush win in a very close race whose outcome was disputed (thereby flipping the entire Presidential election because the race was close on an electoral vote basis nationally), because (1) Nader supporters were drawn almost entirely from voters who either wouldn't have voted at all, or, in significant numbers, would would have voted for Gore if Nader had not been on the ballot, and (2) the race between Gore and Bush was very close in Florida.

If a Washington State style voting system, in contrast, in a Bush v. Gore type election, both Bush and Gore would easily become the two finalists in the second round, despite significant third party voting, and Gore would have had an edge in the second round one on one race undiluted by support for ideologically similar third party candidates.

There are several kinds of spoiler effects that the Washington system can produce, none of which is terribly worse than the existing system.

Dominant Party Cases

One set of cases involve a jurisdiction with a dominant party, and a leading opposition party. In this situation, the dominant party will certainly gets its most widely supported candidate on the ballot. But, spoiler effects may determine whether the other candidate on the ballot is the runner up in the dominant party, or the candidate from the leading opposition party.

If the dominant party has more than two-thirds support and has only two candidates, both will make the second round ballot and the leading opposition candidate will be eliminated in the first round. But, if a dominant party with more than two-thirds support has several candidates, the top candidate from the leading opposition party may get more votes than the runner up to the leading candidate from the dominant party.

Similarly, even if the dominant party has less than two-thirds support, if multiple candidates from the leading opposition party all run for office, but two or more run for the office in the dominant party, this may cause the runner up in the dominant party, rather than the strongest candidate in the leading opposition party may make it to the second round.

In these cases spoiler effects toggle the system between outcomes that serve different desirable goals for the political system.

If a dominant party sends two candidates into a second round election, other voters may be able to get the more moderate of the two dominant party candidates elected with support from moderate members of the dominant party, even if the less moderate candidate had greater support within the dominant party. This result shifts the political scene to the middle and thus more fairly represents the ideology of the jurisdiction in the overall political system, a result which cannot happen in a traditional two party system that tends to elect less moderate candidates to office in jurisdictions with dominant parties.

On the other hand, if division in the dominant party allows a leading opposition candidate to make it to the second round (which is what traditional two party system always does), the leading opposition party candidate will usually lose in the general election to the single dominant party candidate. But, the leading opposition party will have a greater ability to get out its message and have that message taken seriously by presenting itself as a party that is viable enough to make it into a second round election. This may help sustain a multiple party system, something that is particularly a concern if a single party is the dominant party in an entire U.S. state or nation.

Spoilers in Competitive Cases

Different and more exotic scenarios can give rise to spoiler effects in jurisdictions where there is a leading party and a lagging party which are both similar in electoral support (the analysis can hold up even if that support is virtually identical due to turnout quirks on any given election day).

In this case, the risk that the system will send two candidates from the leading party to the second round, rather than one candidate from the leading party and one candidate from the lagging party to the second round is more of a worry.

But, normally, the only way this can happen is if the leading party has precisely two or more evenly matched candidates, and the lagging party either has an equal number of evenly matched candidates to the leading party, or a greater number of candidates than the leading party.

This situation cannot come up if there are three or fewer candidates in the race from all parties combined. Also, it cannot come up if the leading party has two or more candidates but one of those candidates has a very strong lead over the others (e.g. due to being an incumbent) unless the lagging party is deeply splintered.

As a result, the worrisome case where one party gets two second round candidates because the other party unduly splits the first round vote will almost never happen in any situation other than a race for an open competitive seat in which none of the candidates running really stands out as exceptionally better than the others to that candidate's usually partisan base of supporters.

Such races are much less common than dominant party jurisdiction races. Typically, you can count the number of open competitive seat races in a given state or in the federal system in a given year on the fingers of one hand, and it isn't unusual for a state or the federal system as a whole, to have no such races at all.

Open competitive seat races in which multiple candidates from both major parties have near equal support are a minority of these races. Usually, the races that attract the most candidates are open seats in jurisdictions with a dominant party, where a first round win will usually translate automatically into a final round win. Experience also shows that there are very often clear front runners and underdogs in multiple candidate races in at least one of the major plitical parties in an election.

But, there are reasons to be not too concerned even if quirky effects of having multiple candidates in both races sends only one party's candidates to the second round.

First, the second round winner is likely to run for re-election as an incumbent the next election cycle, making spoiler effects in that election cycle unlikely, and because these jurisdictions are, by definition competitive, the party out of power has a reasonable attempt of securing it in the next election cycle, if the second round winner isn't attentive to the needs and desires of his or her constituents.

Second, if neither party has any clear standout candidate in the first round, the candidate who got the most voters in the first round is still probably the most likely to win, and the risk that however wins will turn out to be grossly incompetent relative to the other candidates in the race is rather low, and in that case, that candidate is likely to be replaced quickly anyway.


The design of the Washington State system is operationally easy to operate with existing technology and easy for voters to understand.

The Washington State system does not put nearly so much pressure on individual voters to vote strategically, rather than voting their true preference from the available choices, as the current system.

The Washington State system is fair to third party candidates, making it possible for multi-party, rather than two party systems to evolve over time in states that use it.

The Wasington State system is not systematically biased against moderates, while the current system is institutionally biased against them.

The Washington State system makes the process more relevant and representative in jurisdictions with dominant parties. This could have an important effect on turnout and produce more moderate representatives from "purple jurisdictions" where one major political party has a notable lead in support.

The main disadvantage of this system is that it further weakens the coherence and relevance of major political parties whose clout has been steadily chipped away at since progressive reforms like government run primary elections were instituted across the nation in the late 19th century and early 20th century. But, to a great extent this is already a done deal, and viability this system gives to third parties could compensate for the incoherence that would otherwise follow for the two major parties which already stuggle to maintain broad somewhat incoherent pre-election coalitions in the current system.

It might be better to give political parties greater control over their "brand image" than the Washington State system permits (e.g. requiring candidates who cannot receive approval from a political party to run as independents), but SCOTUS rightly identifies this as a minor rather than a constitutional class flaw in the Washington State system.


Anonymous said...

Washington's Top 2 will be complete death of minor party candidates. Do not expect any minor party or indepedent candidates on the November ballot in the state of Washington except in Pierce County where they will be using ranked choice voting.

Andrew Oh-Willeke said...