On District-Ordered Lists: in reaction to Éric Grenier’s proposal

A month or two back, Éric Grenier from ThreeHundredEight.com, who is often cited on this blog when the discussion turns towards current electoral prospects in Canada, proposed an electoral reform to introduce PR in that country. The proposal suggests a retention of current electoral districts as a list-ordering mechanism: while seats would be allocated proportionally to parties within each province, voters would still cast a vote for one candidate in their district, and each party’s seats would be awarded to the party’s candidates achieving the highest shares of the vote in their districts.

To my knowledge, two countries use a very similar system: Slovenia and Romania. The Slovenian system is practically identical, with the difference that seats are allocated to candidates on the basis of the number of votes they receive in their district, rather than the percentage. The Romanian system is different in that it guarantees a seat to any candidate with more than 50% of a district’s vote, and guarantees that each district has (at least) one representative (with potentially more in the case of an overhang). Another substantially similar case is the German state of Baden-Württemburg, where half the seats are filled by the plurality winners of each district and the other half by the party’s ‘best loser’ candidates by district vote share. In the absence of another moniker (as far as I’m aware) I shall collectively call all these systems (including Grenier’s proposal) ‘District-Ordered List systems’. Of these, only systems where each district winner is guaranteed a seat (as in Baden-Württemburg’s ‘best loser’ scheme) may be considered to be a type of MMP.

Though I recognise that District-Ordered List systems may have some merit, they suffer from some serious disadvantages, especially with comparison to 2-vote MMP.

As the Jenkins Commission put it some years ago, “turning losers into winners” may be seen as problematic. This is even more the case in those systems which do not guarantee district plurality winners a seat, as the representative elected from a district may only have taken a small share of the vote while the first-, second-, third-, and even fourth- placed candidates return empty-handed. In Slovenia (and under the Grenier proposal), some districts, particularly deeply-divided ones, can be left with none of its candidates elected and no local representation at all, while other districts may see more than one candidate elected. Romania doesn’t have that problem as its system ensures that every district has its own MP, but those MPs are more likely to be second- or third-placed within their own district. Baden-Württemburg’s best-loser MMP eliminates most of these problems by ensuring each district’s first-placed candidate gets a seat.

Another limitation of District-Ordered List systems is that they (as proposed or currently used) feature a single vote, meaning voting for district representative is completely tied up with the party vote (unlike the most commonly used variant of MMP, which features two votes: a district vote and a list vote). While ordering lists by district is an attempt to bring individual accountability to the list component[1], having a single vote means that voters cannot assess district candidates independently from the parties. For example, if a voter identifies with a party but dislikes that party’s candidate in his district, he may be faced with a dilemma. That dilemma does not exist under 2-vote MMP, where a district vote for party A’s candidate should not, in principle, affect party B (or any party’s) final seat count, which is determined separately according to the number of list votes attained by each party. This should make individual, district-level accountability stronger under regular 2-vote MMP.

Another issue has been raised previously by Matthew Shugart, this blog’s chief planter:

“By rewarding a candidate for driving up his or her vote even in a district where the candidate has little realistic chance of winning, best-loser allocation exacerbates some of the worst features of FPTP. Whereas under [2-vote] MMP rules, the two big parties have a strong incentive to seek to position themselves near the center of the nationwide electorate to maximize their party vote, under the best-loser rule, they would be back to the old days of targeting districts and seeking to appeal to voters located in relatively closely contested districts. They would want to do so because, even if they are not likely to win such electorates, those are the electorates where their highest quality candidates have the best shot at entering parliament despite losing the district race.”

I don’t know to what degree this has been the case in the aforementioned countries using district-ordered lists[2], but the charge certainly makes sense, although I think it is a little overstated: under First-Past-the-Post, parties target marginal seats as a seat-maximizing strategy. Under any system with district-ordered lists, each party’s seat total is generally determined proportionally and independently of the number of districts won[3]. This means the general effect should be lesser than under FPTP, though it may still be just as strong when it comes to nomination strategies. Either way, Matthew’s argument underlines the point I try to make above: though 1-vote district-ordered list systems are supposed to introduce a certain individual accountability to the list, the fact that voters have but one vote means that to a large extent, voting patterns are party- rather than candidate-based. Safe and marginal seats can therefore be expected to exist by virtue of partisan rather than candidate support. To summarize, this not only means personal, district-level accountability will not be any stronger than under first-past-the-post, but may also impact parties’ nomination strategies at least as much as FPTP.

To some degree, one could address these issues with the best-loser MMP model by slightly modifying it to have two votes: one for district representative and the other determining each party’s overall share of seats, but with list seats going to the ‘best losers’ in the district tier. Voters could split their vote, thus generating district results that more closely reflect voters’ appreciation for the candidates rather than for the parties they are affiliated with (all the while having their list vote count fully for the calculation of the seat total of their preferred party). There would be less of a risk in nominating ‘quality’ candidates in less-than-safe districts, as the relative safety of a district would less be a function of partisan preferences and more of appreciation for the nominated candidates. Of course, this may work only to the degree that people understand and use their two votes in this way. If they would still overwhelmingly vote for their preferred party’s candidate, either because they do not understand that it does not impact their party’s total or for instance because they only value candidates based on their partisan affiliation, the improvement felt would be negligible.

Either way, this also raises a broader question: is pitting individual candidates against other individual candidates of other parties an effective way of holding candidates individually accountable, as under FPTP, MMP or any district-ordered list PR? Personally, I suspect systems where candidates face competition from other candidates of the same party, such as open-list systems or STV, may have more to offer on this score[4].



[1] For example, the Ontario Citizens’ Assembly described best-loser MMP as having “greater accountability [than regular 2-vote MMP] as voters would have a more direct vote for those who are chosen to represent them.”

[2] The evidence Matthew cites comes from Japan, which allows parties to order their lists based on success in districts, but does not make it mandatory.

[3] The exception is overhang, which under some systems may increase a party’s seat total beyond its proportional share, and which may or may not be a realistic prospect, depending on the electoral situation.

[4] Whichever is the case, I can think of only one system that offers a direct compromise between the two principles: open-list MMP.

The dismantling of SYRIZA?

The deal struck in marathon talks between Greece and its Eurozone partners (is that word still appropriate?) requires the Greek parliament to take further action–some as soon as this Wednesday, 15 July.

The parliamentary vote last Friday, which I understand as simply a broad authorization to conduct negotiations around the government’s official opening proposal for the weekend talks, is revealing. While it passed overwhelmingly, with 250 votes in the 300-seat parliament, the government was unable to deliver a majority. SYRIZA (the Coalition of the Radical Left) and its coalition partner, Independent Greeks (ANEL, a right-wing party that agrees with Syriza only on confronting the creditors) control 162 seats. However, only 145 members of the governing coalition supported the authorizing bill.

It is in the context of that vote–and the tougher ones to follow–that we see the importance of the accord PM Alexis Tsipras of Syriza struck with the opposition immediately after the nationwide NO to the creditors’ terms on 5 July. Without the support for the center-right New Democracy (ND) and other parties, the cave-in by Tsipras could not have passed.

It seems that Tsipras’s 5 July plebiscite on his government did indeed strengthen Tsipras’s bargaining position, but with the traditional Greek right, rather than with the EU as he had promised. It is as if Tsipras said to ND: “The voters are with me; back me up so I can implement your policy instead of the one I sold to the voters”. This is why I am calling it a “plebiscite”, not a “referendum”; I understand the latter term to mean a vote on specific proposed laws, whereas a vote re-delegating authority to the government to make decisions on the electorate’s behalf is classic plebiscitarianism.

It is unclear to me how much tougher the conditions agreed this weekend are, relative to the proposals of last week, which, in turn, were roughly as tough as what Greek voters supposedly rejected the weekend before. It is clear, however, that they include some very strict monitoring requirements that are not going down well in Athens.

Can the Tsipras government can survive the stress of all that is being required of it? Already SYRIZA–which is a “party” only due to requirements of the election law, but is really (as its name implies) a coalition of various tendencies–is experiencing divisions. Tsipras has begun to expel several rebels. The party’s rules require dissidents to resign their seats, although the party can’t compel them to leave parliament; they can become independent MP instead. The coalition partner, ANEL, has said it can’t support the new terms, although it will remain in Tsipras’s cabinet.

This would be a less than opportune time for an early election, so any new government may have to be constituted from within this parliament. The Greek constitutional provisions on the cabinet’s confidence (Art. 84) require an absolute majority of MPs present (“which however cannot be less than the two-fifths of the total number of the members”) to remove a PM, which would be hard to reach as long as a rump SYRIZA and ANEL remain together, unless some of the far-left fracture from SYRIZA were to vote with the right. If a new government needs to be formed, the provisions on government formation (Art. 37, in the chapter on the president’s role) provide clear priority to the largest party in parliament, which, thanks to the disproportionality of the Greek electoral system, would remain Tsipras’s unless 74 of its MPs (just short of half) bolt. That seems highly unlikely. Thus it would seem Tsipras can survive, but may end up heading a radically different government from the one that so resoundingly won its plebiscite just over a week ago.

Caving in Greece

No, this is not a travel-adventure post.

It looks like Greece’s governing Syriza Party has caved. Accepting €13bn of fresh austerity.

A note from the Guardian’s live blog at 13:29:

Syriza MPs have been telling our Helena Smith that the big no received in the referendum on Sunday was a “confidence vote” in Tsipras who like no other prime minister before now has the popular support to enforce such punitive measures.

In other words, what occurred last Sunday in Greece was not a policy referendum, nor was it a tool to give Greece more bargaining leverage to reject the European measures to which the voters thought they were voting a loud NO. Rather it was a plebiscite to bolster a weak domestic position. Which would have become far weaker, given the preferences of Greek voters had Tsipras led Greece out of the euro.

Greece–what next? (Yes, I asked that already)

As I type this, nearly three fifths of the votes have been counted in the Greek referendum, and the percentage voting no keeps creeping up, currently 61.3%. This is quite a decisive result, notwithstanding a consensus of several polls that it was too close to call. So now what?

One hint of a possible next political step comes in the Guardian live blog:

And there is even the possibility that Greece’s president, Prokopis Pavlopoulos, resigns to trigger early elections in an attempt to prevent Greece sliding out of the eurozone.

Interesting. It was the deadlock of the parliament last December over choosing a president that, under a provision of the constitution, triggered the early general election. Pavlopoulos is a former minister in conservative New Democracy (ND) governments, but was put forward as the new government’s candidate (after a farther-left rebellion inside SYRIZA against the initial candidate).

I find it fascinating that an almost completely powerless presidency becomes the focus of deadlocks and early elections for parliament. In December, the ND-PASOK government and the opposition (of which SYRIZA was the largest party) could not agree on a candidate. It takes 3/5 to elect a president, or else there is a new general election. But the new parliament can elect a president with a simple majority. Despite that, SYRIZA put forward an ND guy, who then won over 2/3. So ND’s last lever might be for its guy to resign as president and try to force a renewed deadlock over the presidency. Which is powerless–except when its occupant resigns!

Majoritarianism with little strategic voting in Andorra

The Pyrennean microstate of Andorra held a parliamentary election in March.

The principality’s General Council consists of 28 members elected by a majoritarian mixed-member system (MMM): half is elected proportionally from national party lists (largest remainders; the threshold is one Hare quota – 7.15%) and the other half, in parallel, by list plurality (party with most votes takes all seats) – 2 from each of the country’s seven parishes, with quite a bit of malapportionment.

The results were as follows:

Party District List PR Total
seats
Votes (%) Seats Votes (%) Seats
Democrats for Andorra 39.4 (38.6) 10 37.0  (36.3) 5 15
Liberal Party of Andorra 27.5 (26.1) 4 27.7  (26.7) 4 8
PS+VERDS+IC+I 23.6 (24.0) 0 23.5  (24.6) 3 3
Social Democracy and Progress 9.5   (11.3) 0 11.7  (12.5) 2 2
  • The figures in brackets exclude votes from the two parishes where one or more parties did not put forward a slate of candidates.
  • The Democrats and Liberals were allied with various local parties and independents in most parishes.

Curiously, though the tiers are clearly separate, the parties’ vote shares in the different tiers were almost identical (a total of 4.9% when adding up the differences, 4.7% when disregarding votes from districts with incomplete nominations). The same pattern largely holds across districts, and can be observed in previous election results. The only times there has been a significant difference between vote shares in the two tiers occured when certain parties did not nominate in a district.In the districts, the Democrats and Liberals were allied with different parties and independents.

These results run counter to the expectation of strategic voting in the district tier, where the smaller parties had little chance of winning seats. Such strategic voting (in one tier) is common in mixed-member systems. If we can call the pattern in Andorra’s results cross-tier contamination, it would seem to run from the list tier to the district tier rather than the other way. However, this may not the best way to look at it.

Maybe the lack of polls in this small country left voters with too little information to vote strategically? But they could still have learnt about the different parties’ chances from previous elections (of course, strategic voting elsewhere predates the proliferation of public opinion polls anyway). Andorran parties certainly seem to have nominated more strategically in the past, most notably in the previous election (2011) when only two parties competed in each district; this was more rare this time, but alliances between parties remain common.

It may not be of any great significance, but I think it’s a fascinating anomaly.

Greece: What next?

Anyone wish to enlighten–or just to speculate–as to what prompted Greek PM Alexis Tsipras to call a referendum to take place after the current financial lifeline is to run out?

I would have assumed–and may be totally wrong–that the package of reforms would pass parliament if he submitted it there. It would have been a spectacle to see a bloc of Syriza MPs vote against, and the measure pass with the help of New Democracy and Pasok, but that seems like what would have happened. Would Tsipras then be forced out by his own party?

I further would guess that the referendum itself will pass, given that Syriza itself won only 36% of the vote in the election in January (winning almost half the seats and able to lead a coalition due to the electoral system), and most of the other parties are more moderate (leaving aside the nazis of Golden Dawn, of course, who got around 7%). Of course, who knows what will happen to the outcome of the referendum if the week between now and then is as chaotic as now looks possible.

Did Tsipras call a referendum that may provoke a deeper, more devastating crisis for his country just to save his “radical left” credentials? That seems hard to believe, but I can’t figure out what the motivation could be otherwise.

I can’t help but think that amateurs running the Greek government have made a bad situation worse. Discuss…

Denmark election, 2015: Connecting election results and government under multipartism?

For better or worse, Denmark’s election yesterday is a clear example of the connection (or lack of connection?) between election results and likely governing alliances in a multiparty system.

The party of the (outgoing) Prime Minister, the Social Democrats, actually gained seats and remains the largest party by a 10-seat margin. However, Helle Thorning-Schmidt submitted her resignation because her “Red” bloc will have fewer seats than the (former) opposition.

Meanwhile, the core party of the traditional Danish right, the Venstre (liberals, not actually “left”) lost 13 seats, but will be part of the new government. In fact, its leader, Lars Løkke Rasmussen, is the most likely next PM. A Guardian headline even declares that he “wins slender victory”.

The big gainer in the election is the “populist” Danish People’s Party, which gained 15 seats and finished second. It is thus the largest party in the expected new governing “bloc” (if it is even accurate to call it that). However, its leader surely will not be prime minister, and may not even be in the government. More likely it will support a minority government* of the center-right.

So there we have it: the PM’s party gains seats, the largest party by a good margin will head the opposition, the second largest party will be an outside support party, and the PM will come from the party whose seats declined the most!

None of the above is meant necessarily as criticism: in a multiparty parliamentary system, the government is comprised of that set of parties that is at least tolerated by a majority of elected representatives, not necessarily the largest (or even second largest!) party. In general, I admire the Danish electoral, party, and governing systems. But an outcome like this does raise questions about the accountability mechanisms of this pattern of multiparty politics. At the very least, it offers a great teaching case–too bad the Danish could not hold this election a couple of weeks ago when I was indeed teaching a course for which this is highly relevant!

Finally, for fans of Borgen, some help from The Local Denmark.

_________
* I can’t argue with the first two sentences of that article: “It seems to be the season for shocking elections. Rarely has the job of political scientists been so interesting.”