France 2017: Round 4 (honeymoon elections and presidentialization matters!)

Today is the fourth round of the French 2017 election process–that is, the runoffs of the honeymoon assembly election.

Following round 1 (the first round of the presidential election), I used a formula (from Shugart and Taagepera, 2017, Votes from Seats) to “predict” what the round 3 (first round, assembly election) vote percentage would be for the party of first-round leader Emmanuel Macron (on the safe assumption he would win the second round). I pegged it at 29%, based only on Macron’s first-round vote and the elapsed time between then and the scheduled date of the assembly first round.

In the actual voting, La Republique En Marche! (LREM) got around 32%, although I believe that also includes some small vote share for MoDem (which was part of a pre-election coalition). In any case, I won’t quibble about an error of ±3 percentage points. At the time, various commentators were fretting over how “weak” EM would be, what with an untested party and Macon’s having come from seemingly nowhere. Some folks even were wringing their hands over possible cohabitation. It did not take long for polls to catch up with the institutional reality, which is that honeymoon elections matter. The voting result was highly predictable.

Where I went well off the rails was in questioning whether a plurality of votes of around 30% in the first round could translate into an assembly majority. I noted that similar percentages of the vote in previous first rounds in France had translated into around half the seats, but that a safer prediction might be for Macron’s party to be just short. I was not worried about a “weak” presidency, but I thought some degree of post-electoral bargaining would be necessary.

Well, that was silly. I somehow forgot that our assumptions about how votes translate into seats in France are based on the “textbook” French V party system, whereby there are many parties, but two dominant blocs. In such a setting, a leading party (such as a just-elected president’s) with around 30% of the vote would be just far enough ahead of both its allies and the leading party of the opposing bloc so as to translate into a solid majority of seats for the alliance, but not necessarily for the leading party itself. The bloc of the loser of the second round, in the “textbook” party system, is not so far behind the president’s bloc. Therefore, you get a clear pro-presidential majority, but not a knock-out.

Two things should have given me pause. First of all, that the second round presidential candidate was of the National Front, so 2002 would be a better guide than, say, 2012. In 2002, the party of the second major bloc (i.e., the Socialists, whose presidential candidate had finished third) suffered terribly from the honeymoon cycle, and of course, the FN assembly candidates did poorly for lack of allies. This allowed just 33% of the first-round votes for the newly elected president’s party to translate into more than 62% of the seats.

Second, and more to the point, the party system of France 2017 has collapsed badly. Thus being at only 30% of the votes makes you a dominant player in what is, for the time being, a one-bloc system. If you are the centrist party in a two-round system, it does not matter that you lack allied parties in a bloc; what matters is that you have no opposing parties that combine for a coherent bloc against you. Seat projections, issued on the day of the first round of the assembly election, suggested that LREM could get over 400 seats. Some even say 475 (out of 577). LREM candidates will win by default, because in relatively few districts will there be active coordination against them. Moreover, turnout is (predictably) low today.

The following screen shot from Henry Schlechta on Twitter, shows just how dominant the LREM is in today’s runoffs. In other words, don’t let 32% of the first-round votes fool you (as it did me). With different opponents in different districts, from different political camps, there is no reason not to expect a massive majority.

Now that everyone seems to accept that LREM will have a big majority, the concerns (expressed in various news media stories) has shifted to how difficult it may be to govern with a party full of novices. Such concerns are also misplaced. That the party is full of novice politicians makes it more, not less, likely that it will stick to Macron even when times get tough. They have nowhere else to go. They owe their nominations and assembly seats to Macron. France 2017 is presidentialization on steroids!. And, remember, honeymoon elections matter.

24 thoughts on “France 2017: Round 4 (honeymoon elections and presidentialization matters!)

    • I just saw that on France24. So there might have been some district-level coordination just to preserve the presence of an opposition. Or else one of the projections (last week’s or the current one) is wrong.


  1. From the European Tribune website:

    “Poll closed at 20:00 CEST. First estimates are in:

    La République En Marche 355 seats
    Les Républicains 125 seats
    Parti Socialiste 34 seats (PS + allies 49 seats)
    France Insoumise & PCF 30 seats
    Front National 8 seats

    It is a landslide all right, but not the tsunami wave that was predicted last week where estimates were giving Macron’s LREM up to 450 seats.”

    Big win, but not a “rewrite the Constitution”style win, and Macron owns whatever the French government does in the next five years.

    The Socialists are pretty screwed, they look like they will wind up with the same fate of the Radicals, unless they just merge with Macron’s party.


  2. Some records were broken in this conclusion to a remarkable election year in France:

    Abstention, at 57%, a new record
    223 women have been elected deputies, the most ever
    75% of the new assembly are new deputies, also a record
    For the first time ever, 3 Corsican nationalists have been returned to the assembly

    LRM has 308 seats by itself, with an allied centrist paty Modem winning 42


  3. Small caveat on this point:

    “That the party is full of novice politicians makes it more, not less, likely that it will stick to Macron even when times get tough. They have
    nowhere else to go. They owe their nominations and assembly seats to Macron.”

    I think this is probably right only to the extent that these folks aspire to political careers. If they do, then, yes, they rode in on Macron’s coattails and are likely to do his bidding. But it’s also possible that many have joined his merry band in a fit of enthusiasm and advocacy that will wane as the reality of life as a politician sets in, and many may choose to return to their normal lives rather than run for a 2nd term. In that case, he has little with which to motivate them other than policy reforms, and “anti-establishment” is a pretty weak basis for policy consensus.


  4. Yes, that’s possible. I guess I assume the drug of national politics will sweep most of them up, and they’ll want to ride those coattails a second time in five years. Of course, if the presidency turns out really bad (perhaps for reasons out of Macron’s control), then they’d be looking for the exit options. But under such conditions much of that would be true with an established party, too. I have to assume that many Socialist deputies started making other plans a while ago.


  5. I took a look at the differentials between the first round of the 2012 and 2017 Presidential elections in France, and the first round of the 2012 and 2017 legislative elections, and I’m not sure if I agree with the thesis of the original post.

    In 2012 the Socialist presidential candidate took 28.3% of the vote in the first round and went on to win the presidency. In the legislative elections, the Socialists took 29.3%, a gain of exactly one percent.

    There is a case if you combine the Socialist vote with that of allied parties. This went from 30.6% for the first round in the presidential elections, if you combine the votes of the Socialist and Green candidate (the parties were allied for the legislative elections). The Socialists and their allies drew 39.8% of the vote in the first round of the legislative elections.

    However, the main party on the right, the UMP, took 27.2% in the first round of the presidential election and 27.1% in the first round of the legislative election that year. But their allies for the legislative election provided another 6.2% of the vote, none of them having run any presidential candidate, so their vote also increased.

    The big increase in the Main Party + Allies in the legislative first round as opposed to the presidential first round may be just due to the fact the small parties who have made alliances with the big parties generally don’t run candidates for president, the Greens being exceptional in this regard, and their supporters use the first round of the presidential elections to scatter their votes among several minor candidates with no real party affiliation. Otherwise, its remarkable how closely the first round presidential and legislative percentages track.

    Was this the case in 2017? Macron got 24% in the first round of the presidential party, and his instant party got 28.2% in the first round of the legislative election, an increase of 4.1%. He had one ally who was not just absorbed into REM, the MoDems, who did run a presidential candidate, seeing their vote increase between the presidential and legislative elections from 1.2% to 4.1%. So there is a honeymoon or bandwagon effect, though fairly small, not as great as the usual difference in people saying they voted for the winning presidential candidate in the US and the actual percentages (something that happens after every US presidential election, a higher percentage say they voted for the winning candidate than who actually did).

    For there parties, Les Republican’s first round legislative vote was 15.8%, a drop from the 20% first round vote scored by their presidential candidate (their legislative allies got 3.8%). The Socialists first round legislative vote was 7.4%, actually higher than the 6.4% obtained by their presidential candidate in the first round.

    For the other parties, the Front National legislative vote percentage was 13.2%, compared to the 21.3% obtained by their presidential candidate in the first round, and actually lower than the 13.6% they got in the 2012 legislative elections. I’m not sure what the reasons behind the increase, but I doubt 8% of the electorate voted for Le Pen and then voted for Macron’s party, it could be a higher abstention rate or something to do with the slim chances of FN legislative candidate prevailing in the second round (though they increased their seat total this year). Their first round legislative percentage was lower (by 4.3%) than their first round presidential percentage in 2012 as well.

    The Communists backed, as they did in 2012, Jean Luc Melanchon in the presidential race, but ran a separate slate of candidates in the legislative elections from France Insoumise (Rebel France), the party he created. In 2012, Melanchon got 11.1% of the vote in the first round, but his front with the communists got only 6.9% of the first round legislative vote, a similar drop to that of the Front National. In 2017, Melanchon got 19.6% of the first round legislative vote, and in the first round of the legislative elections, France Insoumise got 13.7%, and the Communists 4.3%, which combines for 18% and actually tracks the vote of the presidential candidate both parties backed reasonably well.


    • I should add that I didn’t discuss second round results above, but in the second round a voter has a choice of either abstaining, or making the best of two alternatives that may both be unpalatable. As centrists, Macron and Macronistes benefit from this dynamic. I think the first round percentages are better indicators of popular support.


    • “the MoDems, who did run a presidential candidate, seeing their vote increase between the presidential and legislative elections from 1.2% to 4.1%”. Bayrou, the MoDem leader, actually pulled out of the presidential race in Macron’s favour. Lassalle, the actual candidate, does not appear to be supportive of Macron; he ran against an En Marche candidate for the Assembly, so I think it’s more accurate to speak of Macron’s ‘base’ from the Presidential elections as 24%. Given the nature of the En Marche-MoDem alliance, in which neither party ran against the other, it would be fair to add the support of those parties together to get the vote share for the alliance (which is 32%)

      As for Melenchon, I don’t know where you’re getting your figures from. The Interior Ministry site states that France Insoumise got 11% of the vote and the PCF 2.7%, meaning a roughly six-point drop from the presidential result.


      • I’m also getting my figures from the French ministry of the interior. But when I transcribed the data a few days ago, they definitely had higher percentages for FI and the PCF. The figures for the the other parties don’t seem to have changed.


    • Mathew, certainly the fact that the ‘centrist’ (sic) party was one of just two candidates in most of the round 2 contests was a massive help to winning seats. So much so that I had expected (as did most pundits) that they would get 400+ seats. (My thoughts here following round 1 ( and round 2 ( election).

      But really, the overwhelming issue from this election is the final turnout collapse. It starts to make comparisons with past elections using vote-share percentages pretty abstract. Question for you to mull: are we reaching the point where the more meaningful analysis and comparisons between party performances ought be made using absolute registration-share stats, rather than vote-share stats?

      Also, I think what just happened is pretty clear evidence that France’s unique two-round, threshold-rule system is a poor one, really only suited to the two-major-left-and-right party systems. I’m an Australian, so I wonder how the French election would have gone using preferential voting in a single round. Would not the turnout have been higher if all the candidates had been on such a ballot, and if non-REM voters had had more incentive to bother showing up?

      As for the future, what happens when c25% of the typically-voting French population that just sat out the legislative elections returns to voting (assuming they do!)?

      I am not confident that REM will hold together for five years, lacking a core of policy principles to unite its members. Also, with the current numbers in parliament, every conceivable bloc of around 30 REM members will soon learn that they can threaten to defect on policy issues and bills. To go five years without splits and acrimony looks unlikely.


      • ps – Matthew, not Mathew – sorry!
        also ps – I am part way through transcribing into tables the data from the Interior ministry site also. You couldn’t by any chance point to an online location where that has been done? Would save me many hours.


  6. OK, I see where my error was. The 13.8 percentage was the combined vote of FI and the PCF.

    Because I was doing a comparison of 2012 and 2017, I added the two together for 2017 in my notes. In both elections, Melanchon’s presidential candidacy was backed mostly by the same parties. In 2012, the PCF remained as part of the front for the legislative elections. But in 2017, they ran a separate ticket. So to so see how the “LF” 2012 change, I needed to keep the PCF added in. I then inadvertently double counted the Communists. There was definitely a fall off between the presidential elections and legislative elections, similar to that of the FN. This was probably due to abstentions.


  7. Pingback: The Salvadoran result 2018: The electoral cycle counts! | Fruits and Votes

  8. Pingback: Colombia 2018: Counter-honeymoon elections and presidential primaries | Fruits and Votes

  9. Pingback: El Salvador presidency 2019–a president from a different party? | Fruits and Votes

  10. Pingback: Ukraine honeymoon election today | Fruits and Votes

  11. Pingback: Presidentialization | Fruits and Votes

  12. This in Politico today (well, yesterday for Alain, Henri, et moi):

    ‘… It may seem crazy to sack a popular prime minister. But Philippe’s’ popularity has become part of the problem. France has performed reasonably well in containing the coronavirus pandemic – not brilliantly but not badly. And yet voters and large parts of the media believe that the country’s performance has been disastrous. They blame the government. Or rather, they don’t blame the government – they blame Macron….’

    John Lichfield, Elyséeology: “Macron considers appointing a new prime minister – himself: The incumbent, Edouard Philippe, is outshining the French president.” Politico (22 June 2020),

    Which confirms an observation by the satirical weekly magazine Le Canard enchainé: “A Prime Minister must have no merits…” – quoted 34 years ago by the late Professor George Winterton in Monarchy to Republic: Australian Republican Government (1994, p 63).


    • Is there a semi-presidentialist country where the President nominates himself as the Prime Minister? If parliament is oppose to that, then the President will nominate a Prime Minister that parliament is not oppose to, or the parliament will elect a prime minister. That would be kind of cool that a country goes from Presidentialist to Parliamentary.


  13. Pingback: The Georgia 2020-21 Senate runoffs | Fruits and Votes

  14. Pingback: France’s toxic combo of institutions finally bites | Fruits and Votes

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.