Greece today held what will almost certainly be its first of two general elections of 2023 (hence the ‘a’ in the title above). The incumbent New Democracy (ND) won a substantial plurality of the vote, around 41%, over Syriza’s 20%. By seats, ND has 146, which is 48.7% (BBC, Wiki). That’s five seats short of a majority. The advantage ratio (%seats/%votes) is relatively high (1.19), but the electoral system did not quite manufacture a majority. And therein lies the reason why there will be another election, most likely.
In most recent elections, Greece has used a bonus-adjustment system, whereby the plurality party automatically wins 50 seats before the remaining 250 are allocated proportionally among those clearing the legal threshold (3%). However, back in 2016, the Syriza government passed an electoral system reform that removed the bonus provision. Under the Greek constitution, an electoral reform can take place only upon the second election following passage (unless it is passed with a two-thirds majority). When ND won the 2019 election–158 seats on just under 40% of the vote, thanks to the bonus still being in place–it passed a reform to restore the bonus. However, its bonus could not take effect in the immediately following election, but the one after that. That subsequent election, using the bonus, can take place as soon as the parties “fail” to form a majority on this election. Presumably the ND will try really hard to fail, so that its preferred electoral system can be used, and it can get its parliamentary majority.
It is my understanding that the bonus provision is somewhat different in the new law, although I lack details. I was told by a correspondent that it is “graduated” somehow–that is, instead of a flat 50 seats for any plurality, the size of the plurality will affect the size of the bonus.
I am also unclear on whether the former provision still applies whereby a second election in close succession after an initial one is run under closed lists instead of the usual system of (mostly) open lists.
The relevant articles of the Greek constitution are as follows. Art. 54 states “The electoral system and constituencies are specified by statute which shall be applicable as of the elections after the immediately following ones, unless an explicit provision, adopted by a majority of two thirds of the total number of Members of Parliament, provides for its immediate application as of the immediately following elections.” Art. 37 sets a strict timetable for forming a government and makes another election automatic if conditions are not met. Excerpts: “…If no party has the absolute majority, the President of the Republic shall give the leader of the party with a relative majority an exploratory mandate… [followed by the second and third parties]. Each exploratory mandate shall be in force for three days. If all exploratory mandates prove to be unsuccessful… he shall attempt to form a Cabinet composed of all parties in Parliament for the purpose of holding parliamentary elections. If this fails, he shall entrust the President of the Supreme Administrative Court or of the Supreme Civil and Criminal Court or of the Court of Audit to form a Cabinet as widely accepted as possible to carry out elections and dissolves Parliament.”
Fruits and Votes 
Maybe I am just biased against the naked cynicism of both having a majority bonus and forcing an election to get one.
But is there any real logic to “60% of the people voted against us, let’s have another election so we can get a majority when 60% vote against us again.” In other words, could this backfire if the opposition organizes for it or are they so divided the plan will work?
LikeLike
It seems unlikely, unless there’s a drastic shift in voter preferences. Even if you add up the vote shares for SYRIZA, KKE and PASOK you don’t quite have enough votes to equal ND’s total at this election, and such an alliance seems…implausible, given recent relationships between the parties (an alliance between one of the most unreconstructed Communist parties in Europe and a centre-left party which was part of a coalition with ND up until 2015 might encounter a few internal strains).
LikeLiked by 2 people
MeRA25 got 3.44% in 2019, won 9 seats, and expected to get over the 3.0% threshold this time. If they had, the four left parties would have had more votes, and seats, than ND. Still, not enough to form a majority coalition, unless they held their noses and gave cabinet positions to the ultranationalist Greek Solution, which they could still do today if they were that desperate to avoid a second election.
LikeLike
Is there any logic? Yes, they want to win a majority, and have a path to it. Sometimes it is that simple.
Backfire? I can’t imagine it. While the analogy is not perfect, the second election would be like a honeymoon election: if anything, ND should do better in votes in a new election, because all expectations would be that they are winning anyway. I say this partly because I’d kind of expect turnout among non-ND voters to drop a bit.
LikeLike
Also, this is perhaps a pedantic point, but what else is this website for? It is not a “majority bonus” because it does not guarantee a majority (unlike the one formerly used in Italy). The prior version in Greece was just a guarantee of 50 extra seats, whether or not that was sufficient to get the plurality party over 50% of seats.
(As noted above, the new version is different, but I am not sure just how.)
LikeLike
If I understand it right, the plurality bonus is just a linear relationship between votes and bonus seats. If the plurality winner receives 25% of the vote, they get 20 bonus seats, which are deducted from the proportional seats. If they get 25.5%, it’s 21 seats… and so on until it is capped at 50 bonus seats at 40% of the popular vote. I’m not sure if this applies only to parties or electoral alliances though.
While not quite as asinine as the old formula, it still strikes me as pretty asinine that 50 seats could flip based on a single vote. If you are going to have a plurality bonus at all (big if), why not base it on the margin of that plurality?
LikeLike
Isn’t the amending formula for Greece to have an electoral law to take into effective after the next election seems problematic? Wouldn’t it be better to change it to require a new electoral has to be approve by simple majority + election + simple majority in order for it to go into effect, it seems like the politicians are playing games with the electoral system because they know they won’t win the next election, but the next election after that because they are playing charades with the electoral system.
Why didn’t Syriza when passing the then electoral law just make it as proportionate as possible?
LikeLike
The principle that a majority should not be able to change the electoral system for its own short-term benefit is a good one. Perhaps there are better ways to implement it, but Greece has the right idea here. I’d say that the bigger problem is it’s perhaps too easy to trigger a second election.
LikeLike
Regarding your question, Syriza did not have a majority, so that could have constrained what they could do. But I’d guess the explanation is one or both of the following (1) They just did the easiest reform, which would be keep the system as is minus the bonus clause; (2) they were worried about encouraging fragmentation on the left (a very, very real issue in Greece!) and expected to benefit from a little disproportionality even in losing, which they obviously expected anyway. No one really foresaw that ND would come so close to winning a seat majority even without the bonus clause being in effect!
LikeLike
My understanding of the situation: the reason the electoral system didn’t manufacture a majority is that it’s not a bonus-adjusted system at all, it’s straight PR*. Syriza passed a law introducing PR in 2016, but since they didn’t have a 2/3 majority, it didn’t apply in the 2019 election that elected the current ND government, instead applying in the 2023a election. However, 2023b will be held under the law bonus-adjustment system ND passed before 2023, also without 2/3, which is why it didn’t apply to the 2023a election, but it will apply to 2023b.
*you can tell by the fact that even the smallest party is overrepresented – 5.33% on 4.45% of the vote, for an advantage ratio of 1.2, almost exactly the same as ND. ND’s near-majority seems to be largely a function of the threshold. I think all the 2016 law did was eliminate the plurality bonus, making the electoral formula nationwide PR with a 3% threshold, with seats afterwards being distributed among the districts.
LikeLike
Is there any country other than Greece that uses a reinforce system of Proportional Representation?
LikeLike
We should not use the term “reinforced proportional” given that it is the plurality that is actually enforced, not proportionality!
Italy formerly used a bonus-adjusted system, but it is different. Search the blog for “bonus adjusted” and you will find previous discussions.
LikeLike
The over-representation of small parties that cross the threshold in a PR voting system compared to larger parties is an effect of the magnitude of the electoral system, which is often as high as in majoritarian voting system. The electoral bonus is something else. Greece seems like an anachronism these days, but historically there used to be more cases, particularly in South-East Europe, with Italy viewed as the inspirational example. It’s mostly forgotten stuff, but it was a way of coping with the effects of sudden democratization around there, of correcting in a way a crisis of governability : the tranzition from a census-based majority voting system, with a more limited, stratified, political participation that produced rather homogeneous Liberal elites, to a democratic, necessarly PR, voting system with increased participation and more heterogeneous political elites in the aftermath of World War One. I am a bit surprised it is still used and viewed as legitimate, nut Fascist or something like that, in Greece….
LikeLike
Most PR electoral systems do not overrepresent small parties, and if they do so as a result of a legal threshold, then all parties that clear are overrepresented. So the small parties are not overrepresented relative to other larger parties.
The exception would be Hare quota and largest remainder (with no legal threshold), which indeed can overrepresent small parties, because the cost of a seat via remainder is less than the cost via quota. This is presumably one of the reasons why HQLR is seldom used without a legal threshold (or remainder pooling, which obviates this issue). There are still a few cases of single-tier, no-threshold HQLR in Latin America, for instance El Salvador.
LikeLike
(also the Macanese divisor system where the divisors increase exponentially)
LikeLike
To the allotment of the remainder in D’Hondt proportional representation I was referring in my mind as offering an advantage to small parties that cross the threshold (didn’t know and still don’t know for sure if it was used in Greece; the information about the Greek elections provided here were surprising for me; but by the look of an advantage ratio somebody gave above it looks that it is D’Hont). The D’Hondt method is said to be the most proportional method of distribuiting the remainder of the fractions of votes over seats, but it is still slightly biased in favour of smaller parties. I reviewed the math in Wikipedia and it hasn’t changed. As far as I know it is also the most widely used PR electoral allotment method, often with a threshold. The votes below the threshold are redistributed proportionally to the prties tht cross it, of course, but the threshold – like the district magnitude – creates a source of distroportionality in the representation in the sense that the votes of the parties that don’t cross the threshold get redistrubuted to the parties that cross it.
LikeLike
P.S. D’Hondt disfavours parties around the threshold, the very small parties need a bigger share of votes per seat won, but does favor smaller – as in medium – parties. This is also an indirect effect of the implicit threshold of the district size. My fast remarks should not be taken as obscuring this detail, which is often stated simply as D’Hondt proportional representation favours big parties. Nevertheless, D’Hondt is still more equiproportional than the odd numbers divisors quoitient method of Saint-Lagué, which inflates even more the smaller-to-medium parties the gained seats per votes received ratio (sure at the expense of the large parties, but does nothing for the very small ones). It’s a matter of perspective.
LikeLike
Yes, that is exactly what’s provided in the 2016 law, under which all 300 Vouli seats were initially distributed by the largest remainder method of PR in last week’s election.
In addition, the number of national list seats was increased by three, from 12 to 15, while the total number of constituency seats was reduced by the same amount, from 288 to 285. Nonetheless, they continued to be distributed among qualifying party lists in a way such that it didn’t alter the nationwide upper-tier seat distribution. And while Greek expatriates were allowed to take part in the election, they had no constituency set aside for them.
LikeLike
JD, I think that is what the post says, no?
This election was held without a bonus adjustment, but the next one will have it again. Maybe I am missing the point you are making.
And, yes, the overrepresentation is almost entirely due to parties wasting votes below the legal threshold.
LikeLike
You’re right… I read the post too quickly and was mainly replying to some of the comments I saw.
LikeLike