Skip to content

What happened in the recent German election?

Divya Schäfer writes:

I am not a statistician, but a historian by training. However, I have a lay interest in election analyses, among other topics covered in your blog.

I live in Germany. Yesterday, the federal elections were concluded, as you may have heard. I was wondering if you could share your views on it on your blog?

A bit of background: Many of the results were expected, but some were surprising. Chancellor Merkel’s party was the largest, as expected. On the other hand, particularly unexpected was the extent to which the major parties were weakened and the strong performance of small parties, particularly the radical right-wing Alternative for Germany (AfD), which ended with 12.6%. It is strongest in eastern Germany. This is seen as a reaction to Merkel’s refugee policy, but probably has deeper roots.

Germany’s overall prosperity has also meant that the Left parties have found it difficult to push their message as strongly as they would have liked. According to some graphs they showed on TV yesterday, economic concerns like employment played a smaller role for voters than in previous elections. At the same time, there are many who feel left out. The German electoral landscape has changed, and analysts have offered various explanations for this. Is this speaking too soon? Maybe these elections represent only a blip rather than a rightward shift?

Germany has a proportional representation system, explained here.

Studies of German voting behaviour have been analysed in many scholarly works, which suggest both a generational effect as well as a life-cycle effect (sorry, I could not find English language resources for this). I think this has something to do with the proportional representation system. German voters change their choices from election to election more often than American voters, a phenomenon known as “Wählerwanderung”. This article (link here) from the Financial Times has some nice graphs on the German elections.

We see, for instance (based on fairly accurate exit polls), that the AfD has taken votes from all sorts of sources. Mostly it has mobilised non-voters, but has also taken substantial chunks from the CDU/CSU (Conservatives, also Merkel’s faction) and the SPD (Social Democrats). Even the Linke (Left Party) has lost as much as 11% of its voters to the radical right-wing party.

I have no idea; I know nothing about German politics beyond what I read in the newspapers. But it can be a good idea to blog on things I know nothing about. So feel free to comment, everyone!


  1. Malte says:

    This should give you enough time to educate yourself on German politics as we might have re-elections in a few weeks. I hope this doesn’t happen, but it’s one of two possible outcomes, the other being successful coalition negotiations between conservatives (CDU), ultra-conservatives (CSU), economic liberals (FDP), and the Green Party.

  2. Mathis says:

    German here (if that matters). One thing I feel is not mentioned often enough: People are not necessarily fed up with mainstream parties per se; a less extreme reading of the situation is that a substantial chunk of center-right voters felt that Merkel would win anyway. Nothing at stake, and not much of a difference to be made by voting CDU. So voting AfD makes sense as a signal to CDU leadership not to move too far to the left / progressive spectrum.

    • Quartz says:

      Indeed I’ve read somewhere that 60% of the AfD votes were intended as protest votes, not as support to the party itself. Even I, as an immigrant, have seriously considered it (despite disagreeing with roughly 40% of their program).

      • Terry says:

        Opposition to Merkel’s immigration policies runs around 50% +- depending on how the question is worded.

        How many parties besides AfD oppose those policies? What proportion of seats do those parties hold? Is AfD the only serious way for a voter to express opposition to those policies?

  3. John Hall says:

    Germany has a mixed member proportional system, closer to New Zealand’s than Italy’s. It’s a great system. We should adopt it here in the US and break up the two-party system.

  4. Max says:

    TBH I was not very surprised by the results. I really think the (justified) outrage about the AfD getting into parliament can sometimes be confues with the feeling of surprise. But while it is indeed an outrageous situation, I think it was pretty obvious that it would come to it. And I really think (quite bitterly) that compared to other recent global political events, we didn’t get it tooo bad.

  5. Terry says:

    Sounds like a pretty wild electoral system in Germany. Or maybe it just sounds funny in translation.

    … what you get is both one of the fairest and most complicated systems around. It only recently became completely fair. Germany’s Constitutional Court ruled in 2009 that the voting system used up through the 2009 general election was actually unconstitutional. Then the first fix offered up by the Bundestag was thrown out as well. It was only in February of this year that the country finally got a new system that conforms with the country’s constitution.

    For decades they have been using a voting system that wasn’t even constitutional. (Did anybody notice this?)
    But in February, it became COMPLETELY fair.

    • Björn says:

      It used to be that a small (in terms of vote share) party needed (slightly) more votes per parliamentary seat than a large one, even though seats were meant to be proportional to the vote share as per three constitution. The discrete nature of seats makes simplistic pre -computer-age algorithms hard too design. That is what was fixed, unsurprisingly the large parties took a while to do it.

  6. Terry says:

    These systems with a lot of small parties seem like they should be rather unstable. If forming a government depends on bringing into the government just enough minor parties, does that make the system inherently unstable? Do voters act strategically in ways that make the system unstable? Is a vote for a pivotal minor party more influential than a vote for a large loosing party?

    Disclaimer: I know almost nothing about parliamentary politics.

    • Björn says:

      To overcome some off the instability in the Republic of Weimar (between the two world wars), there is a 5% minimum vote share you need to get to get into parliament (you may still win individual seats unless you win three and get in as per your vote share, after all) it’s largely worked pretty well, but this election is the first time there are two fringe parties In parliament that are not considered acceptable coalition partners. As aresult coalitions are harder to form. It certainly was easier when parties were willing to partner with all other parties and the parties in the middle just went left or right depending on what worked.

      And yes, voters try to be strategic, but usually you just get your 2nd (main) vote counting for the national totals. Some minor stuff is that you might help a coalition partner for your main choice get over the 5% hurdle rather than voting for your true preference. I suppose this time around the voters caused the largest parliament yet at an additional extra cost of ~200 million euros, apparently.

  7. Richard McElreath says:

    One thing I see getting left out of much discussion of the election returns is demography: There is a substantial sex ratio imbalance in the former East, with as few as 70 women per 100 men in some parts (near where I live). This appears to be the usual pattern, where women migrate to cities more than men, leaving excess men in poorer, rural areas. Some portion of difference in voting patterns East-West is a result of demographic differences. But not all of it, for sure.

    • Guive says:

      Do we know if in Germany the men’s vote and the women’s vote differ in systematic ways? In the US they do now but they didn’t always.

      • Julia Rohrer says:

        They do, in particular when it comes to the new right wing party AfD: ~16% of the men but “only” 9% of the women voted for the AfD according to ZDF. This converges with findings from the German Socio-Economic Panel study, in which 2/3 of AfD supporters are male (report:

        AfD seems to be a distinctly male party. 205 of their direct candidates were men, only 30 women. As it happens, the party leader used to be a woman, Frauke Petry, but she left the AfD right after the election.

        As far as I know, the other parties don’t show such a strong gender disbalance, though as far as I recall, more women vote for the Green party.

  8. Emil says:

    The results of the election are not really surprising. All results are within the confidence regions of probabilistic predictions produced by researchers at

  9. Christopher says:

    An interesting side note: YouGov used, similar to the UK some months ago, MRP and published their results at

    Similar to other pollsters, they overestimated the two big parties (CDU and SPD) and underestimated AfD (right-wing populists), Greens (ecological/left-wing) and FDP (liberal democracts).

    I think, this might less be a statistical problem but has two other causes related to German politics:
    (1) About one third of the voters were undecided until Saturday [citation needed].
    (2) The last four years the government coalition was a “big coalition” (Große Koalition, GroKo) by CDU and SPD and in the last months the pollsters predicted that only this “big coalition” would have a majority. So at lot of media attention was on the smaller parties. Some voters thus tried to vote strategically to allow for a different coalition as big coalitions have quite a bad reputation.

    Unfortunately, some media outlets are already repeating the story of failed polls, as they did in the weeks leading up to the election.

  10. isopar says:

    Could anyone here define “populist”? Or “right-wing” for that matter? The definitions I see in the media depend on the acceptance of a particular political position as “true” and “normal”. I can see the value of that in the case of the very old debates – say, taxation, where “let’s tax the rich more” populist and left-wing. But the new things, like millions of Middle Eastern men in their 20s (men with extremely right-wing views on life, Jews and everything) walking into your welfare state, they are new enough for the “normal” and “true” to be a debatable matter.

    • isopar says:

      “The result makes kingmakers of both the FDP and the Greens, both of which have played the role in the recent past but neither of which now has enough support on its own to give Merkel a majority.

      FDP leader Christian Lindner, an ambitious 38-year-old who preaches an ultra-hard line on Europe and has unsettled the German political establishment, said he was open to coalition talks with Merkel but that Germany needed a change of course.

      The Greens’ Katrin Goering-Eckardt said: “We will see if there can be cooperation.”

      A three-way tie-up of Merkel’s conservatives, the FDP and the Greens – known as a “Jamaica” coalition because the black, yellow and green colours of the three parties match the Jamaican flag – is widely seen as inherently unstable.”

    • Andrew says:


      I associate right-wing parties with the following set of policies: In the economy, support for management over unions in labor relations, lower taxes on the rich, less business regulation, and higher military spending. On social issues, nationalism, support for particular ethnic and religious groups, and support for traditional religious and military values. Not all these things always go together, and I don’t know exactly what are the policy and rhetorical positions of the right-wing party in Germany. When moving from right-wing to far-right or Fascist, I think of flat-out racism, appeals to violence, and specific associations with violent far-right political organizations of the past. Again, this characterization is not perfect, and it’s particularly difficult with organizations such as Putin’s in Russia which has nationalism, militarism, anti-gay rhetoric (I guess that’s a sort of modern update of support for traditional religious values), and some sort of accommodation with powerful business interests, but in a post-Communist tradition with connections to Stalin who would have to be considered to be on the extreme left, politically.

      • isopar says:

        Putin is a good example. “Right-wing” movements in Europe and the US was traditionally nationalistic, and that means anti-“Russian”, but Putinism itself is a right-wing ideology. Those in Europe and the US who oppose “putinism” are opposing an “external enemy”, an “outsider”, in that sense their rhetorics is “traditionally right-wing” . Similarly, Islamism is a far-right ideology which happens to be “foreign”, so those who oppose Islamist immigration are opposing the far-right ideology. Islamism aside, the way things are now, mere self-identifying as a “Muslim” (as opposed to self-identifying as “Jewish”) would probably position you far to the right of Putin on most of the daily issues.

    • There is a nice fivethirtyeight piece that does a good job at explaining the german parties:

      Regarding the AFD, you can think of them as a german version of Trumps populism. They are anti-immigrant, deny climate change, have higher support from low-income voters and in low-income regions.

  11. Divya Schäfer says:

    I have a question regarding methodology. Recently published data suggests that AfD voters (right-wing populist party) are mostly those with vocational training, not a university education and mostly workers or self-employed.

    This would suggest that they are mostly “working class” or atleast non-elite. But other surveys (also linked in the article above) have suggested otherwise: that they are economically comfortable. One reason I can think of that might partly account for this discrepancy in findings between various studies is placing emphasis on different criteria: education/professional status/income. The above-cited data, crucially, do not take account of actual income. You could be self-employed without a high level of education but still earn well.

    Is there anything else that could account for a differential reading of the socio-economic status of a particular set of voters?

  12. A few observations from my side:

    See,_2017 for a graphical overview of opinion polling for the last 4 years. The first 2 years show little variation, then things get more interesting:

    Between the fall of 2015 and the end of 2016, roughly 1 Mio. refugees entered Germany (overall population ~80 Mio.), due to a kind of open-border policy established by Angela Merkel and most parties. Accomodation for refugees has been set up all over the country, leading to both a large wave of support and backlash (fueled by sexual assaults on new years eve 2015/16). See for detailed rundown of events. These events mostly correlate with the rise of polling values for the AFD. Since the end of 2016, the borders have been closed and the public debate has mostly moved on to other topics.

    In January 2017, the social democrats (SPD) nominated a new leader, Martin Schulz. The time plot in the first wikipedia article shows one of the prettiest examples I’ve ever seen for a convention bounce, with SPD polls rising +10% in a short span of time and eventually falling back to their original value (more or less).

    Compared to the 2013 election, all three governing parties (CDU, their bavarian counterpart CSU and SPD) lost ~ 20% of their voter share. There are several explanations, with a direct attribution of overall share of the loss being difficult:
    – Their share in the refugee crisis mentioned above (might be estimated by the amount of voters the AFD received)
    – Angela Merkels election taktik, often called “asymetric demobilization” in the media. She appears to mostly evade direct confrontation – see for more details. This worked rather well in the last two elections. In this election, the public TV debate between Merkel (CDU) and Schulz (SPD) was horribly dull, while the debate with the five smaller parties (CSU, FDP, Green, Left and AFD) received much praise. As Mathis and Christopher both wrote above, there has been little incentive for supporters to vote for the big parties.
    – General disappointment with the performance of the governing parties.

Leave a Reply