The 2018 Ranked Choice Voting Election in Maine’s Congressional District No. 2

Maine’s recent congressional election – the first-ever federal poll in the U.S. to be held under Ranked Choice Voting (RCV) – took place against a backdrop of continuing opposition by the Republican Party to the recently introduced voting system. State GOP leaders not only called on their voters to rank just the party’s candidates, but sought as well a court ruling to prevent RCV from determining the outcome of the U.S. House of Representatives election in Congressional District No. 2, and subsequently a recount of all ballots in the election (later called off while it was underway).

Nonetheless, a significant minority of Republican voters in the district ignored party exhortations and indicated valid rankings for at least two candidates, while substantial minorities of non-GOP voters only gave a first preference to Democratic or independent candidates. At the same time, the number of voters who engaged in bullet voting – indicating a preference for just one candidate – constituted a minority of the voting electorate in CD-2. This is notable when one considers that in both the 2016 and 2018 RCV referendums held in Maine, a majority of voters in CD-2 rejected the switch from plurality voting.

Moreover, a federal judge first allowed the RCV count to take place, and subsequently issued a ruling upholding the constitutionality of the election in the congressional district, where incumbent Republican Bruce Poliquin obtained the largest number of first preference votes, but fell short of an absolute majority; he then lost the second and final round of counting to Democratic challenger Jared Golden, who prevailed with 142,440 votes (50.6%) to Poliquin’s 138,931 (49.4%) following the elimination of independent candidates Tiffany Bond and Will Hoar, whose second preferences were transferred to the remaining two candidates. The First Circuit Court of Appeals subsequently denied Congressman Poliquin’s motion for an injunction to prevent Golden from being declared the winner, and Poliquin – who wanted the election outcome determined solely by the first preference count, or by a re-run under plurality voting – dropped the lawsuit challenging RCV shortly thereafter.

The following table, based on a tally of 296,077 cast vote records in CD-2, published by Maine’s Secretary of State on his official website, shows the distribution of first preference votes for each candidate with at least a valid second preference for another candidate (“Preference”), or no second and successive preferences for a different candidate (“Bullet”); the “Other” category groups ballots with valid first preferences, but no valid second or successive preferences due to either overvoting on the second preference ranking – indicating a second preference for more than one candidate – or undervoting i.e. leaving blank more than one consecutive ranking beyond first preference while indicating preferences for other candidates, or a combination of both. State of Maine 2018 Ranked Choice Voting (RCV) Election Data has frequency counts for all 1,564 tallied preference combinations in the CD-2 election.

Candidate Bullet % Preference % Other % Total
Bond (I) 4,333 26.2 12,106 73.1 113 0.7 16,552
Golden (D) 51,423 39.0 79,551 60.3 1,039 0.8 132,013
Hoar (I) 2,120 30.8 4,713 68.6 42 0.6 6,875
Poliquin (R) 89,228 66.5 43,955 32.8 1,001 0.7 134,184
Total 147,104 50.8 140,325 48.5 2,195 0.7 289,624

There were 435 overvotes and 6,018 undervotes in the first preference count; the latter figure – which included 5,711 ballots undervoted on all available rankings – is noticeably lower than the reported number of blank ballots in the plurality-based 2014 and 2016 U.S. House elections in CD-2, and it is also lower than the number of blank ballots in the district for this year’s gubernatorial election in Maine, which was also carried out by plurality voting. At the very least, these numbers indicate the introduction of RCV did not bring about an increase in the number of blank or invalid ballots. In addition, the very low number of overvotes strongly suggests there was little voter confusion about the new electoral system.

Of the 147,104 voters in CD-2 who indicated valid preferences for just one candidate, 137,971 indicated only a first preference, including 315 cases with a second preference but no first preference, while an additional 9,133 voters indicated a valid first preference (or a valid second preference without a first preference), as well as additional preferences, but only for a specific candidate; a large majority of these – 7,706 voters – gave all five preference rankings to their chosen candidate. Under Maine’s RCV counting rules, these votes had the same effect as having indicated only a first preference for the selected candidate. However, while ballots with valid preferences for just a single candidate constitute a narrow majority of the valid first preference votes, they represent a minority of 49.7% of all votes cast in the district. By contrast, in both the 2016 and 2018 RCV referendums, CD-2 reported a majority of votes against RCV both among the valid and overall vote totals. Moreover, even among voters casting valid first preferences, those who indicated a first preference only were a minority of 47.6%.

Bullet voting for the two major-party candidates had no impact in the CD-2 election outcome, since their first preferences were tallied in the second count as preferences for continuing candidates. However, the 6,453 ballots with valid rankings for either Bond only or Hoar only made up the bulk of the 8,253 non-transferable votes in the second count (most of the remaining 1,800 votes in that group had valid rankings for both Bond and Hoar, but not for the other two candidates). It has been suggested that these voters were confused as to which candidates would make it to the second count, but a far more likely explanation is that they simply wished to support the independent candidates only and didn’t care for either of the two major-party candidates. In fact, their behavior is functionally the same as that of voters in traditional runoff systems casting a blank or invalid ballot in the runoff election, after the candidates they originally supported were eliminated in the first round of voting. Moreover, Bond and Hoar first preference voters had the lowest proportion of bullet voting, at just under two out of seven ballots cast for them.

The overall distribution of bullet votes and preference votes in the CD-2 election closely resembles the 2016 and 2018 RCV referendum outcomes, and it would seem this is not entirely a coincidence: in towns with more than ten voters, there were moderately strong correlations between bullet voting in 2018 and opposition to RCV in 2016 (0.62), as well as between preference votes in 2018 and support for the new electoral system two years earlier (0.64); when the correlations were calculated on the basis of valid votes only, both stood at 0.63.

In conclusion, neither all GOP voters in CD-2 ranked Congressman Poliquin only (nearly a third cast a preference vote) nor did all Golden voters (or those backing independent candidates Bond and Hoar) rank other candidates – this was the case with almost three out of eight non-Poliquin voters. There was little evidence of voter confusion, and casting a preference vote or a bullet vote may have been indicative of ongoing support for RCV or opposition to it, respectively; if so, the election outcome did not point to growing opposition to the newly adopted electoral system. Just as important, these findings should leave no doubt that RCV is not a clear-cut partisan issue.

Is AV just FPTP on steroids?

In debates over electoral systems in Canada, one often hears, from otherwise pro-reform people, that a shift to the alternative vote would be worse than the status quo. It is easy to understand why this view might be held. The alternative vote (AV), also known as instant runoff (IRV), keeps the single-seat districts of a system like Canada’s current first-past-the-post (FPTP) system, but replaces the plurality election rule in each district with a ranked-ballot and a counting procedure aimed at producing a majority winner. (Plurality winners are still possible if, unlike in Australia, ranking all candidates is not mandatory. The point is that pluralities of first or sole-preference votes are not sufficient.)

Of course, the claim that AV would be FPTP on steroids implies that, were Canada to switch to AV, the current tendency towards inflated majorities for a party favored by less than half the voters would be even more intensified. This is plausible, inasmuch as AV should favor a center-positioned party. A noteworthy feature of the Canadian party system is the dominance, most of the time, by a centrist party. This is unusual in comparison with most other FPTP systems, notably the UK (I highly recommend Richard Johnston’s fascinating book on the topic). The party in question, the Liberal Party, would pick up many second preferences, mainly from the leftist New Democratic Party (NDP) and so, according to the “steroids” thesis, it would thus win many more seats than it does now. It might even become a “permanent majority”, able to win a parliamentary majority even if it is second in (first-preference) votes to the Conservatives (who thus win the majority or at least plurality of seats under FPTP). The “steroids” claim further implies that the NDP would win many fewer seats, and thus Canada would end up with more of a two-party system rather than the multiparty system it has under FPTP.

There is a strong plausibility to this claim. We can look to the UK, where AV was considered in a referendum. Simulations at the time showed that the Liberal Democrats would stand to benefit rather nicely from a change to AV. While the LibDems are a third party, heavily punished by the FPTP electoral system even when they have had 20% or so of the votes, what they have in common with the Canadian Liberals is their centrist placement. Thus, perhaps we have an iron law of AV: the centrist party gains in seats, whether or not it is already one of the two largest parties. An important caveat applies here: with the LibDems having fallen in support since their coalition with the Conservatives (2010-15), the assumptions they would gain from AV probably no longer apply.

On the other hand, we have the case of the Australian House of Representatives, which is elected by AV. There, a two-party system is even stronger in national politics than in the FPTP case of the UK, and far more so than in Canada. (When I say “two party” I am counting the Coalition as a party because it mostly operates as such in parliament and its distinct component parties seldom compete against one another in districts.)

It is not as if Australia has never had a center-positioned party. The Australian Democrats, for example, reached as high as 11.3% of the first-preference votes in 1990, but managed exactly zero seats (in what was then a 148-seat chamber). Thus being centrist is insufficient to gain from AV.

Nonetheless, the combination of centrism and largeness does imply that Canada’s Liberals would be richly rewarded by a change to AV. Or at least it seems that Justin Trudeau thought so. His campaign promised 2015 would be the last election under FPTP. While he did not say what would replace it, he’s previously said he likes a “ranked ballot” and he pulled the plug on an electoral-reform process when it was veering dangerously towards proportional representation.

Still, there are reasons to be somewhat skeptical, at least of the generalization of the Australian two-party experience. The reasons for my caution against the “steroids” view are two-fold: (1) the overlooked role of assembly size; (2) the ability of parties and voters to adapt.

Assembly size is the most important predictor of the size of the largest party, disproportionality, and the effective number of seat-winning parties in countries that use single-seat districts. (It is likely relatively less important when there are two rounds of voting, as in France, but still likely the most important factor.) This is a key conclusion of Votes from Seats. It is thus important not to overlook the fact that Australia has an assembly size considerably smaller than Canada’s. In the book, Taagepera and I show that Australia’s effective number of seat-winning parties and size of largest parliamentary party are almost what we would expect from its assembly size, even if FPTP were used. (See also this earlier post and its comment thread; how close it is to expectation depends on how we count what a “party” is.) The data are calculated over the 1949-2011 period, and the effective number of parties has been just 1.10 times the expectation from the Seat Product Model (which is based only on assembly size when single-seat districts are used). Similarly, the average largest party has been 93% of the expected size (averaging 50.5%  of seats when we would expect 54.2%).

Thus we do not need to invoke the alleged steroids aspect of AV to understand the dominance of two parties in Australia. But this does not mean it would not make a difference in Canada. Consider that the current effective number of parties and size of the largest party in that country, averaged over a similar period, are also just about what we should expect. The multipartism, including periodic minority governments, that characterize Canada are not surprising, when you use the Seat Product Model (SPM). They are surprising only if you think district magnitude is all that matters, and that FPTP is FPTP. But it isn’t! An electoral system using the FPTP electoral rule with an assembly of more than 300 seats is a different, and more multiparty-favoring, electoral system than one with 150 seats. Replace “FPTP” in that sentence with “AV” and it is surely still true.

But what about the centrist party, the Canadian Liberals? Surely AV would work differently in this context, and the Liberals would be a much more advantaged party. Right? Maybe. If so, then it would mean that the SPM would be overridden, at least partially, in Canada, and the largest party would be bigger than expected, for the assembly size, while the effective number of parties would be lower than expected. Of course, that’s possible! The SPM is devised for “simple” systems. AV is not simple, as we define that term. Maybe the SPM is just “lucky” that the one country to have used AV for a long time has the expected party system; or it is lucky that country has the “correct” assembly size to sustain two-party dominance. (Australia is the Lucky Country, after all, so if the SPM is going to get lucky somewhere, it might as well be Australia.)

This is where that other factor comes in. While no one has a crystal ball, I am going to go with the next best thing. I am going to say that the SPM is reliable enough that we can predict that, were Canada to have AV, it would have an effective number of parties around 2.6 and a largest party with around 48% of seats. In other words, just about where it has been for quite some time (adjusting for the House size having been a bit smaller in the past than it is now). Note these are averages, over many elections. Any one election might deviate–in either direction. I won’t claim that a first election using AV would not be really good for the Liberals! I am doubting that would be a new equilibrium. (Similarly, back in 2016 I said my inclination would not be to predict the effective number of parties to go down under AV.)

Parties and voters have a way of adapting to rules. Yes the Liberals are centrist, and yes the Conservatives are mostly alone on the right of the spectrum (albeit not quite as much now, heading into 2019, as in recent years). But that need not be an immutable fact of Canadian politics. Under AV, the Liberals might move leftward to attract NDP second preferences, the NDP center-ward to attract Liberal and even Conservative second preferences, the Conservatives also towards the center. It would be a different game! The Greens and other parties might be more viable in some districts than is currently the case, but also potentially less viable in others where they might win a plurality, but struggle to get lower ranked preferences. The point is, it could be fluid, and there is no reason to believe scenarios that have the largest party increasing in size (and being almost always the Liberals), and correspondingly the effective number of parties falling. With 338 or so districts, likely there would remain room for several parties, and periodic minority governments (and alternations between leading parties), just as the SPM predicts for a country with that assembly size and single-seat districts.

As I have noted before, it is the UK that is the surprising case. Its largest party tends to be far too large for that huge assembly (currently 650 seats), and its effective number of seat-winning parties is “too low”. Maybe it needs AV to realize its full potential, given that the simulations there showed the third party benefitting (at least when it was larger than it’s been in the two most recent elections).

Bottom line: I do not buy the “FPTP on steroids” characterization of AV. I can understand were it comes from, given the presence in Canada of a large centrist party. I just do not believe Liberal dominance would become entrenched. The large assembly and the diversity of the country’s politics (including its federal structure) both work against that.

I agree with electoral reformers that PR would be better for Canada than AV. I also happen to think it would be better for the Liberals! But would AV be worse than FPTP? Likely, it would not be as different as the “steroids” claim implies.

IRV-MMP

What do folks think of this idea, proposed by Mark Roth in the thread on open-list MMP?

I do not believe it is entirely necessary to have two votes; though I don’t oppose the idea. Essentially I would have IRV-MMP. An instant runoff determines which candidate wins the local seat in each district. First preferences determine who receives the at-large seats. If a voter wants the Greens, but knows that they won’t win locally, a vote 1 Greens 2 Labor has the effect of supporting a winnable local candidate and helping the Greens secure seats in general. I would allow transfers to second (or lower) ranked parties should the first choice(s) of parties not reach a threshold. I would also be inclined to allow a List Party that isn’t running a candidate to appear on the ballot anyway; probably marked to indicate that the List cannot win the local seat. The candidates who lose in their local race would be selected to fill the at-large seats based on their personal vote counts. List order would only be a tiebreaker.

Decoy lists would technically be possible, but they would stick out like a sore thumb, require voter coordination to ensure that the “right” candidate gets the vote in the district level races, and would still need to front candidates in local races to have enough warm bodies.

As I say at the other thread in a comment of my own, I like it much better than the “AV+” idea of having two votes (one ranked-choice for local candidates and the other for list).

Early STV voting equipment

Voting technology is one obstacle to wider use of ranked-choice voting. Although groups like OpaVote have had open-source fixes for years, US jurisdictions tend to rely on commercial vendors. A decade ago, many of them resisited developing the technology. Now, of course, voters can “complete the arrow,” as is done in San Francisco, or bubble in a candidate-by-ranking matrix, as was done in Maine last week.

The challenges get thornier with STV elections. Due to the “multi-winner” nature of a race, there sometimes are very many candidates. That can result in confused voters and burdensome vote counts. Only in 1991 did Cambridge (MA) solve these problems by computerizing its electoral system. That could have happened as early as 1936, when many cities still were holding STV elections.

As it turns out, IBM had found a way to mechanize the voting process. George Hallett of the erstwhile Proportional Representation League writes:

Among the most persuasive arguments against P. R., in spite of their essential triviality, have been the objections that it required several days to get the result in a large election and that it required paper ballots and hand counting, both of which in plurality elections without the safeguards of a central count have acquired an evil reputation. In connection with the possible early use of P. R. in New York City, where these objectives would be stronger than ever psychologically, an effective answer to them has now been devised.

 

IBM’s system used standard, punch-card readers to count STV ballots at a rate of 400 per minute. According to Hallett, “the final result of a P. R. election in New York City can easily be determined by some time in the morning of the day after election.”

Voters would use a series of dials to rank candidates, one through 20. Then, as some will recall, the machine would record a voter’s votes when they pulled the lever to open the curtain. Opening the curtain punched the holes into the punch-card ballot.

Here is the quotation in its context (albeit a bit blurry):

Other features of the system were:

  • Precinct-based error correction. A voter could not give the same ranking to more than one candidate. Nor could a voter skip a ranking.
  • Freedom of choice. A voter could rank as few candidates as they wanted. They also could rank as many as they wanted. Although the machine was built for 20 rankings, there appears to have been accommodation for write-in and additional candidates. Finally, a voter could go back and change their mind about a ranking.
  • Early “cyber-security.” Now we worry about nefarious actors loading malware onto touchscreens. Back in the 1930s, however, the worry was that poll workers might stuff a ballot box or throw out ballots they did not like. IBM’s solution was simple. Poll workers would not have access to individual ballots. Once a voter voted, the ballot fell into a sealed container, only to be opened in the central-count location.

Why the machine did not catch on remains a mystery. IBM appears to have been pitching it to New York City in advance of the November referendum, which put STV into place from 1937 to 1947. Those passing by 41 Park Row could see a demonstration model at the Citizens Union office.

It is a shame that New York (and other cities) did not go with the system. According to Mott (1926), the average invalid-ballot rate in 19 elections to that point was 9.1 percent. My data reveal invalid rates of up to 18 percent (Manhattan and Brooklyn, 1941). Part of this was abstention altogether. Another part was the lack of interest in discerning voter intent, handling skipped rankings with compassion, and so forth. IBM’s machine, however, would have addressed some of those issues, all while educating voters at the same time that they voted.

Voter choice or partisan interest? The case of ranked-choice voting in Maine

Galvanized by the first ever ranked-choice-voting (RCV) win in a U.S. state, reformers just hours ago held a conference call to build their movement. Ranked-choice voting is a set of voting rules more kind to “outsiders” than our ubiquitous plurality system. Given the unusual strength of America’s two-party system, why do outsider-friendly electoral reforms ever win?

My answer is: a replacement institutional template, losing-party self-interest, and ruling-party disunity. In a recently published paper, I show how this logic can explain the spread of “multi-winner ranked-choice voting” (i.e., proportional representation or PR) in the first part of the 20th century. Losing parties and disgruntled ruling-party factions promote voting-system change in a bid for policy-making influence. Voting reform organizations supply the replacement template.

Does my answer also explain the RCV win in Maine? Is that enough to buy my argument? If the answers are “yes,” reformers would concentrate on jurisdictions with sizable out-parties and fractious ruling parties.

Americanist political scientists would also change the way they think about election “reform.” The dominant trend for more than a century has been to see party and reform as exclusive. Fifty years ago, we would have read about conflict between “machine politics” and “good government.” Now we read about “activists” versus “compromisers,” legacies of Progressivism, and reformer “process-obsession.” What if party itself were a critical reform ingredient? As Jessica Trounstine reminds us in her excellent book, Democratic boss Thomas Pendergast was more than happy to turn the model city charter (without PR) to his own “machine” ends in Kansas City.

Let’s see if my template-loser-faction model explains what just happened in Maine.

The template

“Maine has not elected a governor to a first term with majority support since 1966,” said Jill Ward, President of the League of Women Voters of Maine. “Ranked Choice Voting restores majority rule and puts more power in the hands of voters.” – quoted from FairVote.org

Efforts to enact RCV began in 2001.

The losing party

Circumstantial evidence suggests that, from 2001 until the 2014 re-election of Gov. Paul LePage (R), the Democratic Party either:

1) controlled a policy veto point via the governorship, or

2) did not expect “independent” voters’ ballot transfers under single-winner RCV to help elect its candidates.

How is 2014 different for Democratic Party expectations? If the rhetoric of the current governor is any indication, the Maine Republican Party has become more socially conservative. Perhaps it is now so socially conservative (in Democrats’ minds) that the Democratic Party thinks “independent” voters would rank its candidates over Republicans. Maybe Democrats are thinking: “If we had RCV, we wouldn’t be the losing party.”

The disgruntled, ruling-party faction

My hunch is that this is a group of fiscal conservatives, no longer at home in either state party. That doesn’t make them a disgruntled, ruling-party faction, but it might have made them willing to consider Republicans in earlier years. Consider:

  • Proponent of record for Question 5: An Act to Establish Ranked-choice Voting. Liberal on some economic issues, but supports consumption taxes and income-tax reduction.
  • Two-time independent candidate for governor. Liberal on the environment, ambiguous on economics, but not a conventional Democrat of yore. Endorsed independent candidate Angus King (over the Democrat) to replace outgoing Sen. Olympia Snowe, a famed “moderate” Republican.
  • One-time independent candidate for governor. Quits Democratic Party to run. Wanted Maine “to be the Free Enterprise State.”

Predictions and evidence

Last month I predicted that a coalition of regular Democrats and “the independents” would put RCV over the top. Republicans threw me a curve ball by endorsing RCV the very next day, but, as the proprietor of this blog has written, such endorsements can be strategic.

If I was right, Democrats and “the independents” should have voted for RCV, but the Republicans should not have.

Below I give a rough test of these hypotheses. Here are precinct-level results of the vote in favor of RCV by the vote for each major-party presidential candidate. (Vote shares are overall, not of the two-party vote.) This is preliminary. I only have data so far for 87 percent of precincts, the state has not released official results, and I have not looked at the correlation of RCV support with partisanship in other offices. I don’t yet have a way to get at behavior by “the independents.” Finally, I have not yet run an ecological inference analysis, but I plan to remedy all this later.

As you can see, Democrats seemed to like RCV, and Republicans did not, at least as revealed by presidential voting.

The role of uncertainty

Why don’t “the independents” simply join the Democratic Party if they dislike current Republican positions as much as the Democrats? This is what’s really interesting about the adoption and use of RCV. I argue that groups in reformist alliances do not plan to cooperate on all pieces of legislation. Let’s say Maine ends up with an “independent” governor or a sizable contingent of “independents” in its state legislature. I would not be surprised if we see them working with Democrats on some legislation (e.g., “social”), then with Republicans on other bills (e.g., taxes).

Why don’t Democrats foresee this possibility? Perhaps they recognize that single-winner RCV is not the same as PR. Consequently they may reason that “independents” will not become a bargaining force. Rather, “independent” ballots will bolster the position of Democrats in government.

Then why are “independents” going along with a reform that’s good for Democrats? Perhaps they disagree with Democrats on who’s likely to benefit from strategic voting. As Gary Cox reminds us, strategic voting depends in the end on voter expectations, shaped by elite messaging about precisely which party or candidate is “hopeless” under a given electoral system. The perception that RCV has made elections kinder to outsiders is important. If there really are many sincerely “independent” voters, “independent” candidates may get a toehold in government.

And that’s when things get interesting.

Partisan dynamics of support for AV

Maine voters will decide in November whether to use the alternative vote (AV) for all single-winner elections. (I’m not sure about the congressional-district Electoral College votes.) Why does AV have traction, and if it wins, how long can we expect it to last? I assume we need to examine the incentives of party factions. I assume these factions are fighting over a law-making veto point, which is identical to the office itself in a single-winner context. (You can see how I use these assumptions in a working paper on STV, which is the PR cousin of AV.)

Democrats (two factions: regular and insurgent) are the main AV supporters right now. Why? The current Republican governor won with 48 percent of votes in a three-way race. The independent candidate was probably an insurgent Democrat. The 8 percent of voters who supported him probably would have voted for the regular Democrat in his absence. Regular Democrats like AV right now because it would move insurgent Democratic ballots into their column.

Why didn’t regular Democrats like AV before now? As Marsha Mercer notes for Pew:

State legislators in Maine first introduced ranked-choice voting legislation in 2001, when the governor was an independent. They did again when the governor was a Democrat, and once more during the term of current Republican Gov. Paul LePage.

When the bills went nowhere, the Committee for Ranked Choice Voting-Maine collected 73,000 signatures for the citizen ballot initiative.

Regular Democrats didn’t like AV in 2010 because the insurgent Democrat led the regular Democrat, implying regular Democratic votes would have transferred to an insurgent Democrat, thereby creating an insurgent Democratic governor.

Why didn’t regular Democrats like AV in 2006? One answer is that they had the governor they wanted, so the “transaction costs” of a referendum campaign outweighed the benefits. But that’s a lazy explanation. Another answer is that AV would have helped elect a right-of-center — ugh, I hate that word — candidate. The lead insurgent Democrat was Barbara Merrill. She has a history of supporting corporations. So a rerun of the 2006 election under AV would have created a Republican governor.

Why didn’t regular Democrats like AV in 2002, just after the first introduction of AV legislation? Again, they had the Democrat they wanted, but that’s the lazy route. A glance at the vote totals shows the Green Party held the balance. “Ahh, the Green Party people would have ranked the Democrat second,” you say. That is not certain. The Green Party of the early 2000s had a reputation for “centrism” (that word again), meaning their votes may well have transferred to a Republican in an AV rerun of 2002. (I’ll let the Green Party explain its preferences.)

So regular Democrats like AV right now because they expect it to help them, not the insurgents. The minute AV elects an insurgent Democrat, regular Democrats will collude with Republicans to repeal AV.

Other lessons:

  1. Greens are not a genuine party of the left, at least not entirely. Many once came from the pain caucus.
  2. You like AV and want to keep it? Don’t run an insurgent Democrat who beats the regular Democrat in first-choice votes. (If you do, the regular will lose, their votes will transfer to the insurgent, Maine will have an insurgent Democratic governor, and the regular Democrats will be angry.)
  3. Are you an insurgent Democrat? Take over the Democratic Party (becoming a regular). That way you can win with AV if the former regulars run their own candidate, win with or without AV if they do not, and scream them down if they run their own candidate

Does AV mean higher or lower effective number of parties?

There may be a conventional wisdom among people who study comparative electoral systems that the Alternative Vote (also known as Instant Runoff or Majority Preferential) tends to suppress the effective number of parties, compared to plurality (First Past the Post, or FPTP). Or maybe it is just me, but I will admit to having such a notion. After all, Australia is a pretty strict two-party system, isn’t it?

The correct way to approach the question of whether AV means a higher or lower effective number of parties (N) than FPTP is to ask: What we should expect N to be, given the country’s seat product?

As explained by Taagepera (2007) and further elaborated and tested by Li and Shugart (2016), the seat product is a country’s mean district magnitude (M), times its assembly size (S). The Seat Product Model says that the effective number of seat-winning parties (Ns) tends to be the sixth root of this product: Ns=(MS)1/6.

The model is logical, not a mere product of empirical regression work, although regression tests confirm it almost precisely (Li and Shugart, 2016).

When all districts elect just one member, thus M=1, the Seat Product is just the assembly size, S. Hence we take the sixth root of S to get an expectation for Ns. What if we do this for Australia’s House of Representatives? We get an expectation of 2.31.

The actual Ns for Australia’s elections since 1984, the year S was increased from 125 to 148 (subsequently it has increased to 150, a minor change) is… 2.53. However, I believe that figure (I am using Gallagher’s Election Indices) treats the Coalition parties as one in elections before 2010.

In the two most recent elections, Ns has been 2.92 and 3.23. The notes to Gallagher’s Election Indices indicate that for these elections the Liberal Party, the Nationals, and the Liberal National Party of Queensland are treated as separate parties. In my opinion they should be so treated, although I suppose one could have a debate about that.

The actual mean is thus above the expectation for a hypothetical FPTP of the same size assembly. If we use the figure of 2.53, it is obviously not much higher than 2.31 (the ratio is 1.10). However, if we consider the value, at least in recent elections, to be around 3.0, it is about 1.30 times the expectation value.

Contrast this with the UK, where elections of the same period (1987-2010) have a mean Ns=2.30. This is just what we expect for FPTP, right? Not much over 2.0. Not so fast! The UK has a huge assembly, and with S=650 (aprpox., as it varies over the period), we should expect Ns=2.94. The UK actually has one of the more under-fragmented assemblies, according to the Seat Product Model, with this recent-period average being only 78% of expectation.

So how about Canada, where AV is one of the potential reforms being considered? Over a similar period (1984-2011) we get Ns=2.63. With S around 300 during this time, we should get Ns=2.59. So Canada pretty much nails the expectation of the model.

So, should we expect Ns to go down if Canada were to adopt AV, as (what I characterized as) the conventional wisdom would have it? Or should we expect it to go up?

I would not be inclined to say ‘down’. I will just leave it at that for now.

Against ranked-ballot systems: But why?

No, I am not against them, but Joseph Heath, writing at In Due Course, is a skeptic.

Heath’s main concern with “instant runoff” (IRV), also know as the alternative vote (AV), is that it does not guarantee a Condorcet winner.* This is true, and well known. It is one of several methods that will guard against a Condorcet loser, however. Of course, if you want to guarantee a Condorcet winner when there are three or more candidates, you might still use ranked-choice ballots (but a different counting rule), but that’s not what motivates me to write this response.

The motivation is the following passage, which comes after a stylized illustration based on the recent three-way race for Toronto mayor, after which Heath suggests that under his hypothetical distribution of preferences among the voters, Ford supporters would have been better off strategically giving first preferences to a different candidate. He then says:

So it is absolutely and categorically false to say that IRV eliminates the incentive for strategic voting. All it does is invert it. (This is something that everyone should know from history as well – in 2002 too many French voters failed to vote strategically in the Presidential election — which uses a run-off system — leading the Socialist Party candidate Lionel Jospin to be eliminated in the first round, forcing them all to vote for Jacques Chirac to keep the far right out of power.)

That’s some sleight-of-terminology there. Yes, France elects its president with “a runoff system”, but not with an instant runoff. It is a top-two runoff on a later date, and indeed, coordination failure on the left in 2002 prevented Jospin from beating LePen for the second slot. However, can anyone seriously doubt that with IRV/AV, Jospin would have been one of the last two standing when things reached the final count? I don’t think so.

I do not know who the Condorcet winner was in France in 2002. It might have been Jospin, and had approximately the same distribution of first preferences been rendered under an IRV system, then Jospin would have won–if he rather than Chirac was the Condorcet choice.

I agree with what I take to be Heath’s broader point that proponents of IRV oversell it at times. However, if one prefers a system that enhances the electability of a Condorcet winner, it is a pretty good choice. If one wants to guarantee a Condorcet winner, well, that is a different conversation, and one Heath does not enter in to. Most of the rules that would do so have their own pathologies. For single-winner contests, I’m still sticking with AV/IRV as being a good enough solution.

It is also possible that Heath would be happier with the Coomb’s Rule variant of AV.

________________

* A Condorcet winner is a candidate who would beat each of the others in a pairwise competition.

Imagining the 2015 UK general election under AV

What would the last UK election have been like be under the alternative vote (AV)? I was discussing this question with Henry Schlechta, and I thank him for bringing it up.

The 2011 attempt at electoral reform failed, but let’s imagine it had somehow succeeded, say, for example, if the referendum had turned the other way, or if the Liberal Democrats had succeeded in getting it passed without a referendum.

In the referendum, one of the challenges the Liberal Democrats faced was that they were seen as by far the main beneficiaries of the proposed change. Conservatives and Labour alike would have been expected to rank the Lib-Dems second, and in all projections of previous elections under AV, the Lib-Dems were estimated to gain about two-dozen seats on average, even becoming the official opposition had AV been in place in 1997.

However, according to the Electoral Reform Society’s report on the 2015 general election (which presents several projections of the results of the election had it been held under a different electoral system)[1], the 2015 general election would have hardly been any different under AV:

Party Seats under AV Difference from actual seats
Conservative 337 +6
Labour 227 -5
SNP 54 -2
Liberal Democrats 9 +1
Plaid Cymru 3
UKIP 1
Greens 1

In fact, not only would the Liberal Democrats have received just one seat more than under First Past the Post, the main beneficiaries apparently would have been the Conservatives, who were hell-bent on preventing the system’s adoption in the 2011 referendum campaign.

It would seem that the Lib-Dems lost so much support in 2015 that there would have been far fewer seats where they were among the top two parties in first-preference terms, thus being able to survive exclusion until the final round where they could benefit from Labour or Conservative lower preferences. Perhaps they also lost so much credibility that they would get fewer lower-preferences than in previous elections (I couldn’t find the full poll results on which the projection is based – I would be very grateful if someone else were able to share them with us).

Of course, this is just a projection, with some serious limitations. Firstly, it appears votes under FPTP were simply translated into first preferences. In reality, many voters who voted strategically under FPTP would use the opportunity given to them by AV to rank their real first preference first. Secondly, AV might incentivise parties to pursue different campaign strategies, and maybe even have an impact on manifestoes and candidates, as the need to get a majority of the vote would change what it takes to be elected in many constituencies.

In any case, there would probably have been more 1st-preference votes for smaller parties, including UKIP and the Lib-Dems. But the Lib-Dems might also have benefitted from second preferences from parties such as the Greens, who would usually get eliminated from the count first. Might this have evened out the effect of the Lib-Dem collapse in some constituencies and allowing the Lib-Dems to beat the Conservatives to the final ‘round’ on preferences, by getting second preferences from parties smaller than UKIP?

What would have become of the Conservatives’ campaign warning against a minority Labour government dependent on SNP support? Could it easily have become a campaign for voters’ second preferences?

In Scotland, would Unionist parties have recommended preferences to each other to block the SNP?

These are just a few of the questions that should be considered when constructing this alternate history, and I’d love to hear our readers’ thoughts on these as well as other potential changes to the campaign and results if the UK had adopted AV in 2011 (or any other country – there’s at least one that has been discussed here where the question is becoming increasingly relevant).


[1] Based on second- and third-preference polls. The methodology is stated in appendix of the report (page 33).

Party-preferential voting

I had missed most of the following discussion earlier, and as it occurred in a thread on Lower Saxony, it could easily have been missed by others as well.

I am going to reproduce two comments that describe interesting ideas for coping with thresholds.

Vasi:

there’s no reason you couldn’t make thresholds less discontinuous, by combining them with preferential votes.

The mechanics would be a tweaked STV, with the following differences:

1. Because of large district magnitude, a full preferential vote for all candidates would be impractical. Instead voters would list parties in order of preference, with the intraparty order being fixed as in List-PR.

2. To effect a threshold of T seats, a party would not be awarded its first seat until it has accumulated T quotas.

Chris:

Vasi, that sounds essentially like the NSW Legislative Council electoral system.

The election is by STV, with 21 vacancies at an election. One has the option of either voting below the line, by ranking at least 15 candidates, or voting above the line, by voting for one or more party tickets. Unlike federal Senate elections, voting for a party’s ticket does not result in a vote for a preset preference ranking of every candidate; instead, it only ranks the candidates of that party, in the order they appear on the ballot paper. Voters have the option of marking multiple parties above the line, unlike federal Senate elections. So, for instance, the Labor how-to-vote cards in the last election suggested that their supporters vote 1 Labor, 2 Greens above the line. That means they essentially ranked every Labor candidate, followed by every Greens candidate, and if all of them are elected or excluded, their ballot is exhausted.

The Australian group voting ticket essentially operates like closed-list PR, with the exception of in very large elections. The NSW Legislative Council used to use the same ticket style system that the federal Senate uses, but after the 1999 election resulted in a ballot paper the size of a tablecloth (almost 1 sq. m), and a candidate from the “Outdoor Recreation Party” got elected with 8,000 first preference votes (something like 5% of a quota), they changed the group-ticket system to the single party ticket system now in place.

Stephane Dion, the former Leader of the Official Opposition in Canada, also is advocating a version of party-preferential voting. though in ridings which would be only 3-5 seats (1 seat by AV in the territories). In a 4-seat riding, the threshold would be 25% + 1 vote. If all remaining parties are above the threshold, seats are awarded to them by largest remainder (I believe). If there are any parties remaining below the threshold, the party with the least votes is eliminated and their supporters votes transferred to their highest remaining preference. His system is OLPR, with each voter able to cast a preference vote for one candidate of his first-preference party.

I think the most proportional system possible would be party-preferential with a low threshold and a large district magnitude (the most proportional would obviously be a single national district). You could either exclude parties one-by-one (hopefully with block exclusions) until every party remaining was above the threshold, then distribute seats. Otherwise you could simply exclude all parties below the threshold and distribute their voters’ preferences to remaining parties. It avoids the huge numbers of voters wasting their votes by being below the thresold; for instance, even with a relatively low threshold of 3%, 19% of the valid votes in the May 2012 Greek election were cast for parties below the threshold. In this system, the only voters who do not have either a first preference or a transfer vote elect an MP are those who deliberately choose not to rank any parties that make it into parliament.

I also think a novel way to build a stronger government while remaining representative of votes would be to use preferential ballots, but with multiple thresholds. In a 120 seat legislature, 60 seats could be awarded to those parties above 2%, with voters below the threshold transferring to their highest placed remaining party. Then a further 30 seats could be given to those parties above 5% (including transferred votes), then a further 20 seats to those parties above 10%, and then the final 10 seats to the party which wins a majority by transfers. This means that even voters who vote below the threshold are represented, and parties with a decent amount of support have representatives in parliament, just not proportionally to their first-preference votes. You also get larger parties at the top, making a stable government more likely, but unlike supplemental member, or the Italian/Greek plurality-winner top up system, the larger bloc is distributed based on all voters’ preferences, retaining a much larger degree of proportionality than other semi-proportional systems.

I am not necessarily endorsing this concept, although I do find it very interesting. I would be interested in further discussion.

The thread has a lot of other interesting comments on the relationship of thresholds to democratic theory (particularly the last several comments posted as of 4 February). I re-posted the two comments above simply because they refer to proposals for an alternative way of coping with thresholds in electoral-system design.

(I did note the Dion proposal before.)

Northern Territory election 2012

The Northern Territory of Australia goes to the polls this Saturday. 25 MLAs are to be elected from SMCs by majority preferential voting. There is no second chamber. The numbers in the old parliament were 12 ALP, 12 CLP and 1 independent.

Because the territory is remote and thinly populated there is usually very limited polling. What we have suggests the Henderson labor government may actually be returned to office, which would be a first for Labor governments in some years.

San Francisco to repeal ranked-choice voting?

I am passing along the following action alert, dated 15 June, from Californians for Electoral Reform:

The San Francisco Board of Supervisors will, as early as June 26th, consider placing on the November ballot a charter amendment that will repeal RCV for all city-wide elections (Mayor, Sheriff, District Attorney, City Attorney, Treasurer, Assessor-Recorder, and Public Defender), replacing them with a two-round runoff. The first election will be held in September; if no candidate receives 65% or more of the vote (yes, it establishes a 65% winning threshold), the top two go on to a November runoff, seven or eight weeks after the first election. (RCV is kept for the Board of Supervisor elections.)

We need you to contact your Supervisor and tell them why you are opposed to it and ask them to vote against putting it on the ballot. We especially need people who live in District 5, Supervisor Olague’s district, to lobby her, as she may be the swing vote on this issue.

If you don’t know what district you live in, this link should help you.

Some talking points can be found at www.sfbetterelections.com

Additional things to note: San Francisco isn’t used to voting in September; turnout will be abysmally low.

Eight weeks isn’t enough time to certify the first election and get runoff ballots to military and overseas voters. (In some years, such as 2015, there will only be seven weeks between the elections.)

Every year in which we vote for Governor there will be *three* elections: the June statewide primary, the September San Francisco election, and the November general/City runoff election. Talk about voter fatigue, what about poll-worker fatigue!

Please note that the Board of Supervisors will NOT be taking public testimony at the meeting(s) where they consider this issue, as they took public testimony in yesterday’s Rules Committee hearing. Contacting your Supervisor personally is now the best way to make your voice heard. (In addition, you can write the Board, but please contact your Supervisor.)

Thanks,
–Steve Chessin
President, Californians for Electoral Reform
www.cfer.org

Ireland’s presidential election, 2011

Ireland has its presidential election today. The president is elected by an “instant runoff”–specifically, the same Single Transferable Vote system that is used for the Irish parliament, but given a single seat, the quota for election is 50%+1. Of course, this means it’s the Alterative Vote, electing the first candidate to reach a majority on either first preferences or transferred lower preferences of voters whose higher-preferred candidates have been eliminated from the count.

As noted in the Irish Times:

TODAY, FOR only the second time since 1938, a presidential election will ultimately be determined by the second, third and fourth preferences cast by voters…  

This year, unless the polls are seriously wrong, no candidate is likely to be within 10 percentage points of a simple majority on the first count. The election, with seven in the race spread out the way they appear to be, is certain to go to a second, probably a third, and possibly even fourth or fifth counts.

The Irish presidency is weak, within a premier-presidential system that is almost parliamentary. Yet I wonder if the current political upheaval could lead to a president asserting more influence for the office.

The Supplementary Vote alternative?

In reviewing the extensive entries about the recent UK electoral system referendum, which had been posted at the LSE blog on British politics over the past year, I was struck by Professor Patrick Dunleavy’s claim that the Supplementary Vote would have been a fairer system to propose. The Supplementary Vote (SV) is another form of “instant runoff,” and thus bears considerable affinity to the Alternative Vote (AV)–the option that actually was put up against the status quo in last week’s referendum (and was defeated resoundingly).

Under SV, voters are allowed to give just two preferences. If no candidate has a majority of first-preference votes, then all candidates in the contest other than those with the top two totals of first preferences are eliminated. Second choices are then redistributed among the top two, and the candidate with the most votes at this stage is elected.

This procedure differs from AV in at least two fundamental respects. First of all, under AV the voter may (or if we are talking about the Australian House of Representatives, must) rank all the candidates, not just two. Second, AV is a sequential elimination method. Rather than eliminating all but the initial top two, it eliminates the weakest candidate at each stage of the count, up until a winner has been determined. AV thus emulates, in one “instant” act of ballot-marking, the sequential processes used in many political parties and legislative institutions to choose a leader, rather than the two-round majority rules used to elect many national presidents, as does SV. ((Intermediate variants are also possible. For instance in some US jurisdictions that have adopted forms of instant runoff, voter preferences are restricted to the top three. The counting process in these jurisdictions typically otherwise follows the AV approach, rather than that of SV (as far as I know; there may be exceptions).

There is also the so-called Contingent Vote, which allows voters to express more than two preferences, but shares with SV the immediate elimination of all but the top two candidates if none has a majority. Thus the Supplementary Vote and the Contingent Vote arguably have more in common with one another than either has with the Alternative Vote, although the Contingent Vote is less likely to waste votes of those who gave their second preference to an already eliminated candidate. (A form of the Contingent Vote seems to be what Sri Lanka uses to elect its president.) ))

SV is used to elect the London mayor and those of some other English cities; it is used nowhere to elect legislators that I know of.

In a post from last summer, Dunleavy stated:

The London form of AV also has the great advantage that it creates a run-off between the top two candidates in a local constituency – only one of them can win.

I must admit to having a hard time understanding why this is a virtue for a ranked-ballot method. If the race has two clear front-runners and one or more also-rans, fine. The two systems would produce the same result, and such a result would be “fair” by almost any standard, I should think. However, if there is uncertainty about the top two as the election day looms, SV poses a real problem for voters. Not only must they know their two top preferences (easy enough), but they must also think strategically about them (potentially a challenge). If one’s first preference is likely to place third or lower, then one must be careful that one’s second preference will clear the initial top two. Otherwise, the second preference will not be counted, as it will have been wasted on an eliminated candidate. By contrast, under AV, if your second preferred candidate has been eliminated, the counters look to your third (and so on).

But isn’t SV a truer (instant) runoff, inasmuch as it is a top-two system? Here is where the instant-ness ceases to be an advantage. In an actual two-round system of majority runoff, the top two choices of the electorate as a whole are revealed before the voters must give their second preference (i.e. before voters whose initial choice was eliminated in the first round decide which of the two remaining candidates to support in the runoff). The risk of a relatively more extreme candidate winning would be greater under SV than under either AV or two-round majority: if more moderate voters lost their second preferences by tending to have cast them on the third-place candidate, relatively more extreme voters may elect, by less than 50%, their preferred candidate. I have no idea how common this would be in practice, but it seems it would be a not implausible effect of SV. In an actual second round, under this scenario, the voters can “update” and vote for the less extreme of the remaining candidates in the runoff. ((I recall this point being addressed somewhere in the political science literature; apologies for not having a citation.))

Given the substantial number of UK constituencies in which there is a close race between second and third, SV would seem to be clearly less fair than AV to voters confronted with strategic choices about estimating which two candidates are the top two of the electorate at large. (By the way, it is also the case that in presidential elections by two-round majority, the gap between the second and third candidates tends systematically to be less than that between the top two, suggesting that SV would indeed be a poor choice for these elections.)

I fail to see why a case could be made that SV is fairer, especially in any setting in which the “Duvergerian” tendencies of an existing FPTP (plurality) system have already broken down–which are pretty much the only situations in which it would be considered in the first place.