With just over two weeks to go, it is tempting to preemptively dismiss Germany's parliamentary elections as a non-event. Indeed, chances are that investor apathy is quite a rational response to what has been an exceedingly boring campaign, even by German standards. How high those chances are, however, is surprisingly hard to say.
Statistically speaking, German elections are actually quite hard to forecast. As in the U.S., this is not due to flaws in polling. German pollsters have quite a good record in getting the results right within a percentage point or two, especially if you take polling averages of high-quality polls. To be sure, the electorate has gotten more volatile in recent decades. As in other countries, many German voters have become less attached to class or religious identities, less likely to organize in trade unions, and less likely to strongly identify with any one particular party. The share of late deciders has trended upwards. Still, voters tend to decide in fairly predictable ways.
In particular, it would now take a minor miracle to change the news flash you are likely to see on election night. The Christian Democratic Union (CDU), led by Chancellor Angela Merkel, and their Bavarian sister party, the Christian Social Union (CSU), have had a steady and very consistent lead since May. The Christian Democratic alliance is well ahead of its coalition partner, the center-left Social Democratic Party (SPD). Both polling data and historic voting patterns have put CDU/CSU ahead by about 15 percentage points for much of the summer. Take the prediction site Signal&Rauschen. Its forecast combines recent quality-weighted polling data with voting patterns in national and state elections of the past 20 years. This is worth keeping an eye on to put any last-minute polling swings in perspective. As of September 5th, Signal&Rauschen has CDU/CSU at just over 37% of the popular vote and 14% ahead of the Social Democrats.
Looking back even further, the combined share of votes of CDU/CSU has averaged 42.6% for all elections in the 1949-2013 period. This is a touch misleading, as the data until 1990 only covers West Germany. All of the West's traditional parties have at times had difficulties in the former communist East. Partly in response, the CDU has steadily moved towards the left, much to the dismay of its Bavarian sister party CSU. Nowadays, it is probably more accurate to describe the CDU as a centrist, rather than center-right party.
Analyses of party manifestos suggest that the Social Democrats have also moved to the left, particularly over the past 10 years. This pivot has been a lot less successful. In 2013, the SPD won only 25.7%, compared to a post-war average of 35.9%. Its low point so far was 2009, at just 23%. It may well test this low on September 24th. Martin Schulz, the Social Democratic candidate for the chancellorship, has had a tough time mounting a credible challenge. Hopes of winning back voters from the environmentalist Greens and the Left (Die Linke), a far-left party, have failed to materialize. As the junior partner in the coalition, the Social Democrats seem to get part of the blame for joint decisions, but seemingly not much credit by those who approve of the outgoing government. Its mishap-prone campaign has not helped matters either.
Source: Bloomberg Finance L.P., Bloomberg average of polls from GMS, INSA, Forsa, Ipsos, Infratest Dimap/ARD, YouGov and Kantar Emnid; as of 9/5/17
Based on the polling average shown above, one plausible scenario is another grand coalition with Merkel as chancellor and a somewhat smaller parliamentary majority. However, there may be a lot of resistance from Social Democratic Party activists against joining another Merkel-led government. The Christian Democrats have ruled out entering a coalition with either the Left or the far-right, anti-immigration Alternative for Germany (AfD). Apart from another grand coalition, this leaves the possibility of cobbling together a three-party coalition with the Liberals (FDP), a market-friendly and socially liberal party, and the environmentalist Green Party. Earlier this year, such a coalition came to power in the northern state of Schleswig-Holstein. It would be much harder to pull off nationally, notably because of the clear preference among Green Party activists for left-wing policies. After the 2013 elections, coalition talks between the Christian Democrats and the Greens barely got off the ground. Despite the slightly more pragmatic current Green Party leadership, it is difficult to see them agreeing to lower, rather than higher taxes, a key demand of the Liberals.
Instead, many Liberals and Christian Democrats are hoping for a narrow majority without the Greens. Such coalitions have been frequent in Germany's post-war history, most recently from 2009 to 2013. That last spell in government ended in an electoral meltdown for the Liberals at the 2013 elections, which saw them get booted out of the Bundestag. With memories of that defeat still fresh, the Liberals may play hard to get this time around. Its platform includes moderately lower taxes and a slower progression towards fiscal integration in the Eurozone. But, it is very hard to put a probability on either of this occurring. This is due to several quirks in the German political system and voting arrangements.
Casual observers tend to describe Germany as a system of proportional representation, with a five-percent threshold and a domineering lower house of Parliament, the Bundestag. This is misleading on all three counts.
Unlike voters in most countries, Germans get two votes when electing the Bundestag. The first vote is for an individual candidate in each of the country's 299 constituencies. Candidates are elected by plurality to represent the respective constituency in the Bundestag. Whoever gets the most votes wins a so called "Direktmandat". This is best translated as direct seat, as opposed to the rest of the seats in the Bundestag allocated via proportional representation. The concept of direct seats invites an interesting thought experiment. Imagine that the next Bundestag was only made up by directly elected constituency candidates, akin to a Westminster-style first-past-the-post system. Assuming no change in voting behavior, CDU/CSU could hope for well above 80% of seats in the Bundestag. A look at constituency-level forecasts such as the ones by election.de illustrates just how dominant the Christian Democrats have become across the country.
In addition to the first vote for determining which constituency candidate will get the direct seat, German voters get a second vote to choose a party. Voters can combine their two votes in any way they want. For example, you might give your (first) vote to the local Social-Democrat candidate, because you like her personally. You can still vote for the Greens – or any other party – with your second vote. Technically, the second vote is for the favored party's regional (state) list.
Once all the first and second votes are counted, the candidates who won the direct seats are then joined by candidates from the regional party lists. We will spare you the details behind the somewhat complicated calculations. For our purposes, there are only two things you need to know. First, a party needs to clear one of two hurdles, in order for its second vote total to count. It needs to receive more than 5% of second votes nationally or win at least three direct seats (i.e. get a plurality of first votes in at least three constituencies). And second, seats to the Bundestag are allocated until the percentage share of seats closely matches the percentage share of second votes of the parties that qualify for proportional representation (i.e. excluding second votes for parties that do not qualify).
Sounds confusing? Well yes, which is why these finer details are often misunderstood. The upshot is that numerous uncertainties lurk beneath those steady polling numbers. A look back at recent history helps drive home this point. In 2013, CDU/CSU came fairly close to an absolute majority in the Bundestag, winning 311 out of 631 seats, or 49.3% with just 41.5% of (second) votes. That's because in that election, a record 15.7% of (second) votes had gone to parties, notably the AfD and the Liberals, that failed to qualify for proportional representation as they just fell short of 5% nationally and failed to gain a single direct seat.
In 1994, by contrast, a similar CDU/CSU 41.4% share of votes translated into just 43.8% of seats in the Bundestag (294 out of 672 seats). The difference was partly due to just a few thousand votes each in several key constituencies in East Berlin. In four East Berlin constituencies, the Party of Democratic Socialism (PDS) was able to win a plurality. The PDS was the successor party to the former communist rulers of East Germany and one of the predecessors of the Left. Because of its strength in those four constituencies, it was able to participate in the proportional distribution of seats, thereby boosting the strength of its delegation to 30. Chancellor Helmut Kohl's coalition with the Liberals saw its majority sharply reduced and was unable to get much done during its subsequent term in office.
Fast forward to 2017, and an interesting asymmetry emerges for modelling purposes. Among the four smaller parties, only the Left has a solid shot at securing more than three direct seats. This is due to its strength in a handful of East German constituencies, mostly in Berlin. Of course, this is no guarantee of success. In 2002, it held on to just two constituencies, one short of the three needed to get representation in line with its 4% share of the national vote. And in any case, the Left has lately seen its poll numbers go up to 10%. Unlike the other three smaller parties, however, its Bundestag delegation could probably survive a last-minute mood swing taking its national share below 5%.
The Liberal's vote is too evenly distributed throughout the country to put it in contention for any direct seats. The AfD's limited track record makes it hard to predict its distribution of votes. It will probably struggle to win a single direct seat, though two of its most prominent figures, Frauke Petry and Alexander Gauland, are trying hard in Saxony and Brandenburg.
This leaves the Greens. They are defending just one direct seat in Berlin, with a first-time candidate, Canan Bayram, and not considered to be in serious contention anywhere else. They have also been trending downwards in recent pools, and are currently at 6.5%. Like any polling snapshot, this should be taken with a grain of salt. Models that include past voting behavior or rolling averages produce higher figures. That said, the trend clearly appears to be unfavorable, and even many erstwhile Green sympathizers have been dissatisfied for a while. It is therefore not implausible that the Greens are now just a normal polling error away of falling short of 5%. If that is the case, they risk being reduced to either having just one seat in the Bundestag or none at all. This would most probably allow the Christian Democrats and the Liberals to secure a majority. Conversely, and much less plausibly, it would take a last-minute collapse in the vote of both the AfD and the Liberals for a left-wing coalition of the Social Democrats, the Greens and the Left to get anywhere near a parliamentary majority.
Which leads to an interesting question: how much would it matter to financial markets if, on election night, Ms. Merkel and the Liberals found themselves with a comfortable majority, perhaps due to a combination of some last-minute gains and a surprising Greens shortfall? The most likely answer is, not all that much. National elections in Germany tend to be a lot less consequential than in the United States, France, or the United Kingdom. Historically they have not had much of a market impact, as we have previously pointed out. For example, there is no clear evidence that markets have particularly cared about the relative strength of left- versus right-wing parties.
The limited market impact of national elections is no mere historical fluke. Power in Germany is widely distributed. Most importantly, Germans will only determine the composition of the Bundestag, the parliament's lower house. By contrast, the balance of power in the upper house, the Bundesrat, is not on the ballot on September 24th. Voting strength in the Bundesrat is continuously determined by regional elections in each of the federal states (Bundesländer). States also play a critical role in implementing national policy. Members of the Bundesrat are designated representatives of state governments (not legislatures), and can be recalled or replaced at any time by state governments. No matter who "wins" elections to the Bundestag, and no matter what coalition emerges after national parties are done haggling, the new chancellor has generally been well advised to work with the opposition – or risk blockages in the Bundesrat. Indeed, some scholars have long argued that Germany can be thought of as in a permanent (formal or informal) grand-coalition state.
Lately, the realities of governing Germany have gotten even messier. Most states find themselves governed now by coalitions of two, sometimes three parties. This makes it hard to obtain thean affirmative majority required for many areas of legislation in the Bundesrat. All plausible coalitions that might emerge from the elections to the Bundestag are well short of the 35 votes needed for Bundesrat approval. For example, the grand coalition of CDU/CSU and Social Democrats has a mere 16 out of a total 69 votes in the Bundesrat. It commands 6 for Bavaria, where the CSU governs alone, plus 10 for Mecklenburg-Vorpommern, Saarland and Saxony, led by coalitions of CDU and Social Democrats. To be sure, both parties are in plenty of other state governments, often as senior partners. But, as each state casts all votes jointly, they would still need to get their junior coalition partners in each of those states on board.
Sources: Bundesrat, Deutsche Asset Management Investment GmbH calculations; as of 9/5/17
*The Bundesrat has a total of 69 votes, 35 votes are therefore needed to achieve an affirmative majority
The Bundesrat can slow down decision-making on other matters, too, beyond legislation requiring affirmative approval. This was a frequent complaint by both Gerhard Schröder and Helmut Kohl, Ms. Merkel's predecessors. Moreover, representation in the Bundesrat is not proportional to population. Instead, small states have disproportionally more votes in the Bundesrat. Partly as a result, even relatively small parties such as the Greens have quite a bit of veto power via the Bundesrat. The move of the Christian Democrats to the left reflects not just political maneuvering, but the realities of governing, given the frequent need for Bundesrat approval. Despite some modest reforms to the workings of the Bundesrat in 2006, the need for consensus building remains strong in Germany's federal system.
All this makes sharp policy changes unlikely, but also a touch unpredictable. If Ms. Merkel's Christian Democratic alliance does as well as forecast, the vast majority of CDU/CSU members of parliament will enter the Bundestag via direct seats. Unlike most of their peers in other parties, they would be able to claim the democratic legitimacy that comes from having carried a plurality in their own constituency. Combine that with the Liberals' skepticism on Eurozone rescue packages, and a large win may actually make Ms. Merkel's task of governing the country harder, rather than easier. In the longer term, even bigger challenges loom. If the AfD becomes a permanent fixture in the German parliament in the way the Left already is, it will be increasingly hard to build functioning coalitions. Admittedly, that's a big if. So far, the AfD has mainly shown a knack for factional infighting. That said, German voters – not to mention investors – may be well advised to enjoy the country's boring politics while it lasts. It is unlikely to last forever.
3. Even Germans tend to find them a little confusing and there has been a fair amount of tinkering over the years aimed at ensuring proportional representation. That's trickier than you think. For an overview of advantages and disadvantages of various methods one might consider, see Olga Birkmeier, Kai-Friederike Oelbermann, Friedrich Pukelsheim and Matthias Rossi (2012) "A Gentle Combination of Plurality Vote and Proportional Representation for Bundestag Elections", pp. 151- 171., in Dan Felsenthal and Moshé Machover M. (eds) Electoral Systems. Paradoxes: Assumptions, and Procedures, Studies in Choice and Welfare. Springer.
4. For polling data beyond the headline numbers, see, for example, Thomas Petersen of Insitut für Demoskopie Allensbach, "Grün mit Grauschleier", published March 29, 2017 in Frankfurter Allgemeine Zeitung and available online at http://www.ifd-allensbach.de/uploads/tx_reportsndocs/FAZ_Ma__rz2017_Gru__ne.pdf
6. Schmidt, Manfred (2003), "Germany: the grand coalition state", in Colomer, Joseph, Political institutions in Europe (2nd ed.), London New York: Routledge, pp. 55–93; an updated version from Feb. 2015 is available online: http://www.uni-heidelberg.de/md/politik/personal/schmidt/schmidt__grand_coalition_state__4th_edition.pdf