A supporter of German Chancellor Angela Merkel holds posters reading “For one Germany in which we will love to live and live good” before an election rally Sept. 6 in Torgau, Germany. (Reinhard Krause/Reuters)

German citizens head to the polling booths Sunday to vote for the German parliament, the Bundestag. A majority is expected to vote for the Christian Democratic Union (CDU), returning the current Chancellor Angela Merkel to office for her fourth term.

But CDU isn’t expected to win an outright majority, and will therefore have to build a coalition government. So which other parties will gain enough seats to join Merkel in governing?

Here’s what our team of political scientists from Mannheim, Berlin and Zurich have concluded from our forecasting project zweitstimme.org.

As of Friday, two days before election day, our final forecast for the party vote shares is as follows: the allied CDU and Bavaria’s Christian Social Union (CDU/CSU) at 35.7 percent, Social Democrats at 22.4 percent, the Left at 9.5 percent, the Greens at 7.8 percent, the Free Democratic Party at 9.3 percent, and Alternative for Germany (AfD) at 10.8 percent. We discuss the uncertainties about those point predictions below.

The Christian Democrats will lose seats

Merkel’s leadership has generally gotten good media coverage around the world. Nevertheless, we predict that the CDU (and its junior Bavarian party, the Christian Social Union, or CSU) will together lose about five percentage points of the vote from the previous election. Despite this loss, we forecast that Merkel will nevertheless be in a strong position to choose coalition partners.

Currently, the center-right CDU/CSU is governing in a coalition with the Social Democrats (SPD), the center-left party. But during this campaign a new possible set of partners has been discussed: the Jamaica coalition, called so because the combination of the three parties’ colors are the same colors as the Jamaican flag. CDU/CSU, the lead partner, is symbolized by black. The Greens, which focus on ecological, economic and social sustainability, obviously are green. The remaining partner would be the Free Democratic Party, or FDP, a small party dedicated to civil liberties and human rights that is represented by the color yellow and has long worked in coalition with the CDU/CSU.


How certain are we about the predicted vote shares?

Pollsters have called a number of recent elections incorrectly. Our group has put a great deal of effort into calibrating the uncertainty estimates of our simulations and communicating them in a transparent manner. Each party’s bar in the graph above indicates the range of vote shares that it has a 5-in-6 chance of winning. To end up in this range is very likely but still not certain.

Our predicted vote shares imply that all of the six parties are likely to get at least 5 percent of the national vote, which is the minimum required to get seats in the Bundestag. If our predictions are correct, the Free Democrats (FDP) will reenter parliament after falling below that threshold at the last election.

The most significant change in the parliamentary landscape isn’t one that would change the government, however. We predict that the right-wing populist Alternative for Germany (AfD) party will get enough votes to gain Bundestag seats for the first time — and might even become the third-biggest party. In fact, according to our simulations, the AfD has a more than 2-in-5 chance of becoming the third-biggest party in parliament.

How likely is a majority of seats for various coalitions?

We predict that the CDU/CSU-SPD have a 100 percent chance of winning enough seats to be a majority coalition; the Jamaica coalition has an 88 percent chance; CDU/CSU-FDP, 26 percent; CDU/CSU-Greens, 16 percent; SPD-Greens, 0 percent; SPD-FDP-Greens, 2 percent; SPD-Left Party-Greens, 2 percent.

Other conceivable options for the CDU/CSU with only one additional partner, such as a coalition with the FDP (26 percent) or the Greens (16 percent), are considerably less likely to gain a majority of seats given our predictions. According to our simulations, they have no viable way to form a majority coalition against Merkel’s Christian Democrats.

Merkel, by contrast, will have two viable options. First, she could continue the current coalition government between CDU/CSU and the SPD, which will gain a majority of seats for sure. Second, the Jamaica coalition has a current 88 percent probability of winning a majority of seats.

By contrast, Merkel’s main challenger on the left, the Social Democratic Party’s Martin Schulz, will have very few options for forming a government. This past spring, the SPD had a sudden surge in the polls, but that has vanished. All potential SPD-led coalition options have a probability of less than 2 percent of actually gaining a majority of seats.

The new political map

You can see how dominant the CDU/CSU will be in another aspect of our forecast: the predicted winners of the 299 electoral district seat races. German voters use a two-vote ballot for the election of their federal parliament (Bundestag). The first vote is like voting for the U.S. House of Representatives. It is a vote for a local candidate in the district, within a single-member district system, with the person who gets the most votes winning one of these 299 district seats. Half the seats in the Bundestag are allocated in this way.

More important for the election outcome is the second vote, which is a vote for a party list. These proportionally allocate the total number of seats in the Bundestag to each party according to the share of second votes it garners nationally. This proportion includes the district seats already won by candidates through the first vote.

If a party won more electoral district seats through the first votes than it should be allocated through the second votes, the party can keep those seats — and the Bundestag will then expand, because other parties will be given additional compensatory seats. When this happens, the number of seats within the Bundestag will grow.

That’s what we predict for Sunday’s election. We predict that the CDU/CSU will win a large number of districts, even though we predicted the party will win only a moderate national vote share of 35.7 percent of second votes. Those two things together will increase the number of seats within the next Bundestag. With a 5-in-6 chance, the number of seats in the next Bundestag will increase from its current 631 to a total of between 646 and 773 seats.

A larger parliament not only increases the burden for taxpayers; it also decreases the efficiency of day-to-day policymaking.


When inking the map in the color of the party of the likely electoral district winner, the electoral map is a study in black: 218 districts have a probability of 65 percent or higher of electing the CDU/CSU candidate, while only 34 districts have a chance of 65 percent or higher of electing an SPD candidate. The CDU/CSU will most likely win all districts in the southern states. With few exceptions, North Rhine-Westphalia’s industrial center, the Ruhr region, is the only safe spot left for SPD candidates.

The Left Party, which used to have viable candidates in some electoral districts, particularly in the states belonging to the former German Democratic Republic, or East Germany, has also lost ground. Only three districts in Berlin will almost certainly elect Left Party candidates.

Despite the fact that Merkel gets another term, the political landscape in Germany will most likely be a very different one starting next Monday. Given our predictions, we expect more parties in a larger parliament, including a right-wing populist party that will use its new publicity to strengthen the party’s extremist agenda. It has to be seen whether the established parties successfully work together across the aisle of government and opposition to counter this extremist agenda. Nevertheless, more parties in the Bundestag will make governing harder.

Notes on methods: Our dynamic forecast is a hybrid of fundamentals (conceptually similar to the classic “Chancellor Model”) that define the environment of the campaign on the one hand, and pre-election opinion polls on the other. For more details on methodology, see our research paper and a summary on our website. District specific forecasts are featured on a dedicated page. Code and data are available on GitHub.

Thomas Gschwend is a professor of political science at the University of Mannheim.

Simon Munzert is a lecturer in political data science at the Hertie School of Governance in Berlin. Find him on Twitter @simonsaysnothin.

Marcel Neunhoeffer is a PhD candidate at the Graduate School of Economic and Social Sciences at the University of Mannheim. Find him on Twitter @mneunho.

Sebastian Sternberg is a PhD candidate at the Graduate School of Economic and Social Sciences at the University of Mannheim. Find him on Twitter @GFCC_Forecast.

Lukas F. Stoetzer is a postdoctoral researcher in the political science department of the University of Zurich. Find him on Twitter @lstoetze.