At the conclusion of the first round of voting for the French National Assembly, with projections showing a majority for newly elected President Nicolas Sarkozy’s UMP ranging from nearly two thirds to more than three fourth of the seats, I posed the question of whether France needed a new electoral system. With the party system fragmented, yet now dominated by two large and relatively moderate parties, the old majority-plurality two-round system no longer seemed to be serving the country well. When the electoral system was adopted in 1958 (and used for every election since then, except for 1986), there were no real “giants” in the fragmented party system and not even two clear blocs. One of the largest parties was a doctrinaire Communist party. In that context, a system that very quickly realigned the parties into two major blocs and led to the under-representation of the Communists was a reasonably good choice. (The realignment was also aided greatly by the adoption of direct, two-round majority presidential elections beginning in 1965, although it was already in evidence before then.)1
In light of the projections arising from the first round, I suggested (perhaps rather shockingly) that even first-past-the-post would be an improvement over the current system, given the tendency of the two-round system to over-exaggerate the lead of the largest bloc, and within the bloc, its largest party. In other words, the 2007 elections were about to show, at the conclusion of the second round, that the electoral system no longer served the purposes for which it had been devised, but rather was just inflating the dominance of the president and his party. (Once again, we have to look beyond the electoral system itself: this dominance is greatly aided by the change in the presidential term from seven to five years, which effectively guarantees that a newly elected president will have an immediate National Assembly during his “honeymoon.”2)
Well, a funny thing happened on the way to the UMP’s two-thirds or three-fourths majority: the voters corrected it! Given a chance in the second round to exaggerate or trim Sarkozy’s majority, the voters (or, rather, the 60% who bothered to vote) cut it. The majority will still be large enough to empower Sarkozy’s government (apparently minus Alain Juppe, a former prime minister to whom Sarkozy had already given the environment and energy portfolios, but who lost his parliamentary seat). The UMP won 313 seats, or 54.2% of the 577-seat National Assembly. However, under several alternative electoral systems, the majority might have been bigger.
Le Monde has a nice interactive graphic that allows one to see estimated seat totals for the various parties under several alternative electoral systems. (Even though it is in French, you do not need to read French to understand the graphics.) It just so happens that the FPTP scenario is 421 seats for the UMP, or 73%. Even a PR system is estimated to give the UMP 333 (58%) of the seats. The estimate under German-style MMP with a 5% threshold is 317 (55%).3
Now it is worth noting that any such simulation is to be taken with a grain of salt. It is a “ceteris paribus” exercise; it assumes voter behavior would not change under the different system. In fact, the shape of the vote would have been different under any of these systems. Nonetheless, the exercise is a reminder that, for a given vote distribution, a two-round system is not necessarily as disproportional, once the voters have gone to the polls a second time, as it might appear from a projection of the first round.
I would still conclude that a PR system (whether MMP or all-list) would be better (surprise!), for the reason I articulated before.
A proportional system would have the advantage of confirming the strong position of the president’s party, while making the assembly election matter for the precise shape of the coalition the president builds.
After all, even with this diminished (relative to projections) majority for the UMP, it is still a single-party president-dominated majority. And such an outcome was never in doubt, making the election largely an exercise in coronation rather than choice. And that is why the turnout was so low–at 60% (both rounds), I believe it was the lowest for a legislative election under the French Fifth Republic. Given the new electoral cycle of Assembly elections shortly after presidential, as long as the electoral system is practically guaranteed to generate a majority, there will appear to many voters to be little at stake in the parliamentary contest. A PR system, and the different patterns of alliance building it would induce, would re-energize French parliamentary elections and be much more consistent with the premier-presidential (semi-presidential) model France has, in which the president dominates policy-making only to the extent that the voters (and the electoral system) permit him or her to do so.
____
1. In 1958, after Charles de Gaulle had been chosen as president by the National Assembly, France held its first election under the new Fifth Republic, under the two-round system (majority required to win a district’s sole seat in the first round, but a plurality suffices in the second). The largest parties in votes percentages nationally were Gaullists (20.6), Conservatives (20), Communists (18.9), Socialist (15.5), and Popular Republican (11.1). However, thanks to the electoral system, the seats percentages were, respectively: 42.6, 28.6, 2.2, 9.5, 12.3. (Yes, 2.2% of the seats for Communists, despite 18.9% of first round votes!)
In 1962, the second election under this electoral system, but still before the first direct presidential election, the leading vote-winners were: Gaullist (33.7), Communist (21.9), Socialist (12.4), Conservative (11.5), Popular Republicans (7.8). Their seats, respectively, were 49.5 (!), 8.8, 13.7, 6.9, 8.0.
Over the subsequent four elections, the trend of Gaullist dominance of the right and Socialist dominance of the left would accelerate. Socialist dominance of the left would accelerate even more after the election of President François Mitterrand and his exercise of his right to dissolve parliament and call a “honeymoon” election in 1981. The electoral system had worked to consolidate two blocs, each dominated by relatively moderate parties, replacing the earlier fragmentation and strong Communist Party. Mission accomplie.
2. “Cohabitation,” in which a president from one ideological bloc must appoint a premier of the opposing bloc because the latter controls the Assembly, happened only at the elections that occurred five years into each of Mitterrand’s terms (1986 and 1993) and again in 1997 when President Jacques Chirac gambled on an early dissolution of the National Assembly and wound up with a Socialist majority. Other elections have resulted in pro-presidential majorities, including Mitterrand’s two dissolutions after his own election in 1981 and reelection in 1988 (mission accomplie!) (Some of these majorities have not been dominated by the president’s own party or have required post-election cooperation by centrists from outside the two blocs.)
3. The list-PR simulation assumes proportionality with each department serving as a multi-seat district and a 5% (district-level) threshold. This would appear to be identical to the system used in the one PR system of the Fifth Republic, in 1986 (when the Socialists changed the system to conserve their own expected losses and to hand the new right-wing majority a parliament that would include the far-right National Front; again, mission accomplie). Under this variant of PR, many smaller parties with regional concentration do better than under the MMP simulation requiring a party to win either some (I assume, as in Germany, 3) single-seat districts or 5% of the national list vote in order to win seats from the list tier. The MMP simulation has seats for only the UMP, Socialists, and François Bayrou’s new centrist MoDem. The latter party gets 61 seats under MMP, compared to only 28 under departmental PR (and 3 in the actual majority-plurality system)–no wonder Bayrou supports MMP! Also worth noting is that the National Front (4.3% of first-round votes) wins 5 seats under departmental PR, and none under any of the others (though, of course, it might well clear 5% of the votes if the latter were sufficient for representation, in which case it would win at least 30 seats).
More on the French Elections……
Scion grafted by PoliBlog (TM): A Rough Draft of my Thoughts — 20 June 2007 @ 10:43
Le Monde’s projections on German MMP are wrong, since Germany first deducts small party votes, and the PCF and FN both have more than 5% after that is done.
And MMP with a 3% threshold would give seats to the Greens as well. The result on the first-round votes as cast would be UMP 47.23%. Rather than have to rely on the FN, the UMP would have had to form a coalition with Bayrou’s MoDem.
Of course this is academic. Voters would have voted differently. The PRG, which ran in alliance with the PS, would have formed a bloc with them. Still, I doubt the UMP would have had a majority.
Seed planted by Wilfred Day — 21 June 2007 @ 04:36
The newly elected french president Nicolas Sarkozy has installed a committee to propose constitutional changes (“comité de réflexion et de proposition sur la modernisation et le rééquilibrage des institutions de la V° République”).
Some issues Sarkozy puts forward are linked with the de facto “presidentialisation”. Since 2002 president and parliament are elected for synchronous terms of five years and the presidential elections precede the parliamentary elections. Parliament should be more powerful and the relationships between president, parliament and the prime minister should be rethought, and a formal recognition of “the opposition†(like in Westminster-style parliaments?). But Sarkozy doesn’t want a US-style separation of powers between president and parliament.
Last but not least, the committe should also think about a ‘bit’ of PR for national elections (“l’introduction d’une dose de représentation proportionnelle au niveau national aux élections sénatoriales ou législatives.”)
(His speech in French)
Seed planted by Bancki — 26 July 2007 @ 07:45
French Ministers may – nay, must – hold parliamentary seats now?
Seed planted by Tom Round — 26 July 2007 @ 20:49
A french MP who becomes minister, has to leave parliament (art. 23 Constitution). No reelection is needed: a supplementary candidate takes the seat until the next election (art. LO 176-1 Code Electoral).
But Sarkozy wants that a dismissed minister can return to his seat. “Cela permettrait d’avoir plus de renouvellement dans la composition des gouvernements.” (It allows having more government reshuffles)
Seed planted by Bancki — 27 July 2007 @ 05:26
Prof Todd Zywicki has argued that this was one advantage of the pre-Seventeenth Amendment US system, with Senators being chosen by State Legislatures – that eminent Senators could be persuaded to serve as Cabinet officials in the administration, knowing that their home State legislature could easily re-elect (or re-appoint) them to the Senate if they fell out with the President.
Now, Senators who resign or are dismissed from the Cabinet have to kiss a thousand babies and spend a few million to get their old seat back.
Seed planted by Tom Round — 30 July 2007 @ 02:28
The “comité de réflexion et de proposition sur la modernisation et le rééquilibrage des institutions de la V° République”, presided by ex-PM Balladur, has presented its report on october 29th.
Let me read it first, but it looks interesting.
http://www.comite-constitutionnel.fr/
Seed planted by Bancki — 16 November 2007 @ 07:56
We discussed the Balladur Report some at TDP. It almost looks like a call for MMP.
Seed planted by Jack — 17 November 2007 @ 18:45
I doubt that 30 seats out of 577 would result in a workable version of MMP. It sounds more like a (small) parallel component to me. The actual language (recommendation #62 in the report) seems awfully vague to me. But then so is my knowledge of French, so I might well be missing something.
Seed planted by Bob Richard — 17 November 2007 @ 20:10
It doesn’t say anything that mind-blowing – just “20 to 30 seats to ensure the representation of minority political formations.”
I don’t recall why I saw MMP over MMM in the crystal ball. Maybe I asked a stupid question, to which the answer is no. =)
Seed planted by Jack — 18 November 2007 @ 04:56
I’ll leave out the details on the semi-presidential system in France and stick to the ‘representativeness of MPs’:
Some proposals of the Constitutional Committee are endorsed by president Sarkozy (accoring to his response):
– Senators are indirectly elected, mostly by local councillors or their respresentatives. The voting rights (the amount of “grands élécteurs”) should be apportioned between the municipalities according to population (article 24.3) – until now, small (rural) municipalities are overrepresented in the departmental electoral colleges.
– National Assembly electoral districts should be redistricted every 10 years on the public advice of an independent commission (article 25.3) – the current map (‘Loi-Pasqua’) dates from 1986, based on population figures of 1982.
Representation of frenchmen living abroad: the Committee wanted to stick to the exiting solution of 12 indirectly elected senators and advised against out-of-country voting for the National Aseembly, but Sarkozy insists.
Sarkozy wants some issues to be discussed further :
– the ‘cumul’: many MPs also serve in lower institutions (40% is also mayor), how can the do both jobs at a time and without conflicts of interest? The Committee proposes a ban on the cumul between MP and a lower executive mandate (eg. mayor), Sarkozy doesn’t make a choice.
– the electoral system (single-member two-round majority at present) would remain a statutory issue and would not be mentioned in the constitution:
The committee proposed 20-30 compensatory MPs for underrepresented parties having more than 5% nationally.
Sarkozy wants to keep one-party-majorities in the National Assembly and favors proportionality in the Senate (besides the existing indirectly elected senators?).
Seed planted by Bancki — 21 November 2007 @ 13:48
France is up to a new apportionment and districting project.
1° The last districting act (1986 on the basis of the 1982 census) foresaw a new ‘redécoupage’ after the second following census (1999); but two elections (2002 and 2007) have been held on the 1986 map; the constitutional council wants also a new districting, but has no power other than asking it politely.
On the basis of the last available census (1999) and ignoring the seats for special overseas territories, the disparity has grown to 2-to-11 (Lozère 2°: 34.400 – Val-d’Oise 2°: 188.134) and, inside a single département (Var) to 2-to-5 (1°: 74.027 – 6°: 180.368).
2° The total number of seats remains 577, but two novelties ask for a reapportionment.
Untill now, seven seats are set aside for special overseas territories (Nouvelle Calédonie (2), Polynésie Française (2), Wallis et Futuna, Mayotte, Saint-Pierre-et-Miquelion). In the new apportionment, two extra seats will be set aside for Saint-Martin and Saint-Barthélémy who were recently detached from Guadeloupe département. Twelve others will be part of a new scheme for out-of-country-voters (there are already 12 senators representing French abroad). In sum there will remain 556 (was 570) seats to be apportioned to the 104 départements.
The idea to have some 30 MPs elected by PR is not mentioned any more.
The next step is an official set of inhabitant figures, to be delivered in the first days of 2009.
A constitutional amendment, being debated in parliament these days, sets a new rule : an independent commission will give a public advice on apportionment and districting proposals.
After the legislative districting is done, the ‘cantons’ (the basic units for districting within the département) will be remodeled too, because they are also the single-member-districts for the departmental councils (‘conseil général’) and they suffer from an even more severe population disparity due to passive gerrymandering. (e.g. in Var: Comps-sur-Artuby 1.109 – Fréjus 50.536 inhabitants = 1-to-45-disparity)
(Source of figures: Michel BALINSKI, Le suffrage universel inachevé)
Seed planted by Bancki — 20 June 2008 @ 11:27
That’s extraordinary that France has held elections recently on a 1986 apportionment! I wonder how many additional seats that bias has contributed to the right, compared to what a fair apportionment would have netted on the same votes.
Thanks very much, Bancki, for this update. It is seeds like these that keep the virtual orcahrd fruitful!
Regarding the proposal for an independent commission, do you have any details on the current process of apportionment and boundary delimitation?
Seed planted by MSS — 20 June 2008 @ 12:54
> “the constitutional council wants also a new districting, but has no power other than asking it politely”
Is this because, IIRC, the CC can only strike down new Acts immediately after their enactment, and only if a large quorum of MPs challenge it? Wasn’t Chirac talking about extending the scope of “constitutional review” (we can’t really call it “judicial review”) in the French system? I heard Prof Blandine Kriegel speak in Australia (mid-2000) and she mentioned the presidential commission on the “modernisation of the state”, which was considering moving the CC in a more (ahem) “American” direction.
(Actually, now I think of it, that’s probably what killed the idea…)
Seed planted by Tom Round — 20 June 2008 @ 17:00
A constitutional reform bill is currently being debated by the French parliament. It is the result of the modernisation committee that you mentioned. The bill has just passed the National Assembly, but is now being amended by the Senate. One element of the reform, which the Senate does not seem to be opposing, is the opportunity for citizens to use the courts to petition the Constitutional Council. However, this will still only happen very indirectly. If the courts uphold the petition, they will pass on the case to the Council of State or the Court of Cassation who will have to decide that there is a need for the Constitutional Council to make a decision. All of this means that if a pro-government redistricting act was not sent to the Council before it became law, it will now be possible for a citizen to take a case after the law is passed. In practice, though, a redistricting act would be one that the opposition would definitely send to the Council before it became law. This happened in 1986. Bear in mind, though, that even if the Constitutional Council is not quite as egregiously party political as it was in the early 1980s, the right still has a big majority on it.
Seed planted by Robert Elgie — 21 June 2008 @ 11:39
> “I wonder how many additional seats that bias has contributed to the right, compared to what a fair apportionment would have netted on the same votes.”
Even if the initial plan was biased, the passive gerrymander is not neccesarily also pro-right.
In fact, it seems to be the other way round! If you sort the 570 ‘circonscriptions’ in the 100 départements (and not 104 as in my earlier seed) according to population (1999 census):
* 190 smallest, most overrepresented:
103 right – 2 MoDem – 85 left
* 190 median:
114 right – 76 left
* 190 largest, most underrepresented:
124 right – 66 left
Seed planted by Bancki — 24 June 2008 @ 09:40
Gerrymander, or malapportionment? Please, let’s keep the concepts separate. I see them mixed up a lot. Obviously, they are similar.
Malapportionment is unequal ratios across districts of voters to representatives.
A gerrymander is the drawing of district lines to favor the election of specific parties or types of candidates. That is, there can be zero malapportionment yet much gerrymandering (as is the case in the US House, ignoring the small overrepresentation of the smallest states).
I have always heard of France having malapportionment. Is there also an issue of gerrymandering?
Seed planted by MSS — 24 June 2008 @ 12:23
Gerrymandering STRICTO SENSU is difficult to prove.
With ‘passive gerrymander’ I only meant ‘deliberately not adjusting, so that growing areas get underrepresented and declining areas get overrepresented’.
How should I call this phenomenon more correctly?
Can the seat-vote-equation be used now that France moves from a two-block to a two-party-landscape (right-left -> UMP-PS)
Seed planted by Bancki — 25 June 2008 @ 03:12
That’s malapportionment (and indeed, calling it “passive” in this case is sensible). Indeed, gerrymandering can be hard to prove, but it is by definition deliberate, whereas malapportionment may be, but also may simply be the result of population shifting faster than the recalibration of voter:representative ratios. (Of course, not recalibrating regularly is itself very likely a political decision.)
As for the s-v equation, it can be used with any number of parties (the effective number is one of the parameters) and with any electoral system. I do not recall what the parameter is for runoff systems, and arguably there is no good theoretical tool for majority-plurality, given that it is almost sui generis. However, that does not preclude making a model for France alone: We have a lot of elections, and we would be able to know the average s-v relationship. Thus we would be able to see how much any given election deviates from the norm, for a given percentage of the vote for a given party. (But Bancki’s question allueds to the difficult problem: do we use party votes and seats, or those of the blocs?).
Seed planted by MSS — 25 June 2008 @ 12:03
The Monster of Loch Ness of french electoral system debates is back: François Hollande (PS presidential candidate) says in his manifesto (engagement n° 48) he’ll introduce a bit of proportionality (“une part de proportionnelle”) in the lower chamber of parliament.
Seed planted by Bancki — 07 February 2012 @ 07:58
I believe introducing some element of proportionality was part of the electoral deal with Europe Ecologie-les Verts. Now that the latter are flailing in the presidential contest…I’ll believe it when I see it, put it that way.
Seed planted by DC — 07 February 2012 @ 09:47
Would France be better off changing it’s 2 round system for the lower house to an Australian Preferential Voting system?
How about for the election of the President as well? That would have made it less likely for Le Pen (2002 Presidential election) to have gotten 2nd place after the final distribution of preferences.
Seed planted by Suaprazzodi — 09 February 2012 @ 01:57
Preferential voting would certainly end the room for manipulation through leaving races in the Nat’l Assembly, not to mention saving a significant amount of taxpayers’ money.
Seed planted by JD — 09 February 2012 @ 03:44
I think Sarkozy would prefer FPTP, it was their initial proposal for the 2014 territorial (departements+regions) elections.
Seed planted by Bancki — 09 February 2012 @ 14:21
One compromise might be to use AV for the first round only. Exclude the lowest candidates one by one and transfer their ballots until the candidate(s) remaining have over 2/3 of the votes. Then put the last two survivors head to head, a week or fortnight later. (Or declare the higher candidate of the final two elected outright if s/he is supported by more than- say. 40% of all eligible voters).
I was surprised by how many AV opponents emerged in last May’s UK referendum debate who disliked AV but were happy with the idea of a second ballot (or Supplementary Vote). Something like the above would satisfy these people while avoiding the “Le Pen with 18%” problem.
One problem with introducing preferential (number-ranking) voting to European elections is that many nations- incldg France, IIRC – vote not by marking your (one and only) ballot-paper but by selecting one party’s or candidate’s ballot and placing it in the official envelope.
Seed planted by Tom Round — 10 February 2012 @ 00:12
Mathematically, once you get two candidates that each have over a third of the vote, you get your runoff. You could use AV in the first round, in districts when no one has gotten a majority of the first preference votes, until you get there.
The question is what happens if a candidate is pushed over the 50% mark, while preferences are being redistributed to see who is going to be in the second round. Do you still have the second round in that case? My guess would be no.
Then why not go ahead and use AV to decide the whole thing? The use AV to decide who gets into the runoff only comes into play if no candidate gets a majority outright on the first preferences AND you don’t have two candidates who have gotten a third of the vote on first preferences AND counting second preferences pushes two candidates over the one third threshold but does not give any candidate a majority. I think the Le Pen thing was a rare goof-up in an otherwise sound system, as shown by the fact that subsequent elections indicate that the effect on French politics was minimal to non-existent.
Seed planted by Ed — 10 February 2012 @ 01:42
The French don’t really have any tradition of numerical ranking in their elections-I thnk one of the many iterations of PR under the Fourth Republic had a provision for a flexible list, where candidates could be written in or struck out from the pre-ranked list.
Sarkozy and the UMP are considerably less enamoured of FPTP than they once were-the results of the last few local and regional elections were disastrous for the governing party and would have been even worse under a one round system. The controversy around that initial suggestion was quite interesting as it ilustrated how deep in the French political imagination the idea of the two round ballot goes-as deep as the idea of FPTP in the Anglo-Saxon countries, even if the modern form of both systems is far from ancient. Not even the addition of what they call a “dose” of proportionality was enough to convince the satellite parties of the centre and right to support the reform.
Incidentally, I think the proposals from the two major parties for an element of proportional representation seem to be more for a paralell alocation than compensatory.
Seed planted by DC — 10 February 2012 @ 08:32
I think that if the French are unwilling to try preferential voting, a good compromise might be a two-round (semi-)proportional system. MPs would be elected from open lists, and the second round would be among the country’s top two parties from the first round. The party that wins the second round gets a regional bonus (eg: +25% of seats in all regions were it had both a plurality in the first round and a majority in the second)
Seed planted by JD — 12 February 2012 @ 02:42