New Insights into Coalition Negotiations—The Case of German Government Formation Open Access

For almost six months—171 days—following Germany’s legislative election on September 24, 2017, a caretaker government presided over the country, and a record was set for the length of time between an election and the formation of Germany’s government. On March 14, 2018, a new government was finally formed with Angela Merkel reelected as chancellor, leading a government of a renewed grand coalition of the Christian Democrats and the Social Democrats (the Christian Democratic sister parties—the Christian Democratic Union of Germany and the Christian Social Union in Bavaria [Christlich Demokratische Union Deutschlands or CDU and Christlich‐Soziale Union in Bayern or CSU])—and the Social Democratic Party (Sozialdemokratische Partei Deutschlands or SPD). This outcome was surprising, as SPD party leader Martin Schulz had ruled out the grand coalition’s renewal immediately after the election results were announced. The failure to form the “Jamaica” coalition—so named because of the colors adopted by the participating parties—CDU/CSU (black), Free Democratic Party (Freie Demokratische Partei (FDP)) (yellow), Alliance 90/The Greens (Bündnis 90/Die Grünen) (green)—was especially puzzling for political scientists and other observers who had expected these negotiations to succeed (Buck 2017). The failure could not be explained by established theories of coalition negotiations based on utility maximization, as the interests of these parties were much more aligned than those of CDU/CSU and SPD, although the alignments differed across policy sectors (e.g., they were closely aligned in railway policy (Döhler 2019), somewhat aligned in family policy (Gülzau 2019), and weakly aligned in refugee policy (Engler, Bauer‐Blaschkowski, and Zohlnhöfer 2019). Thus, the process of government formation in this case provides new insights into coalition negotiations that are only accessible through in‐depth interviews with those who directly participated in these talks. Data from interviews, public statements, and internal party documents shed light on the events that occurred during this period of government formation.

Immediately following the publication of election results, then‐party leader Schulz ruled out another grand coalition—defined as a partnership of the two long‐term dominating parties in terms of members and vote shares—between his SPD and CDU/CSU, which again had obtained the majority of votes. Thus, only one potential coalition remained (Czuczka 2017). To form a coalition that was composed of a majority in parliament and also politically compatible, CDU/CSU approached the FDP and the Greens to explore a potential Jamaica coalition. But the FDP unexpectedly declared these talks failed (Karnitschnig 2017), which left the German political system with a first‐time challenge of failed coalition negotiations on the federal level. A grand coalition or a minority government of CDU/CSU were the only available options. In the end, the “GroKo” (short for Große Koalition, German for “grand coalition”) came to an agreement and after the approval of the SPD party members, the new government (which had also been the old government before the general elections) elected Chancellor Angela Merkel on March 14, 2018. It is the fourth grand coalition in German history; all have been led by Christian democrats (Kurt Georg Kiesinger, 1966–1969; Angela Merkel 2005–2009 and 2013–2017). But unlike previous grand coalitions, this one presents a least likely case because in terms of party interests, a more rational alternative was at hand. Investigating why this coalition formed offers interesting insights into coalition negotiations today, both in Germany and beyond.

During the nearly six months of uncertainty surrounding the coalition’s formation, there were two distinct phases, each marked by changing actor constellations and negotiation venues. An analysis of the period reveals why one phase succeeded while the other failed. It highlights the important influence of mutual trust between the parties and of intraparty power games, factors on which traditional analyses of coalition formations do not primarily focus.

This article is organized as follows: While the second section formulates theoretically derived hypotheses on what outcome would have been expected in the exploratory talks, the third section emphasizes why these prominent theories on negotiation fail to explain the formation of the current German government. The fourth section then systematically analyzes the second phase of negotiations and puts forth the alternative theoretical explanations for its outcome. A conclusion summarizes the explanations and lists lessons learned.

Political scientists have long believed that coalition negotiations are driven by the traditional goals of political parties—achieving policy objectives, winning votes, and obtaining government offices (Döring and Hellström 2013). Parties seek to set policy agendas (Zohlnhöfer 2009), obtain votes in future elections, and maximize power by occupying posts in governmental bodies (Müller and Strøm 1999; Meyer and Schoen 2015). Parties’ goals are set according to partisan interests rather than the desires of individual politicians. However, in negotiations, politicians may seek to realize their own policy goals or aspirations for political power. The assumption that parties are unitary actors has been challenged in recognition of the fact that intraparty strength, especially the intraparty support of the party leader, rather than unitary party positions shapes coalition outcomes (Debus and Bräuninger 2008). Thus, intraparty support for the party leader appears as an important resource in coalition negotiations.

Within the traditional perspective of parties as unitary actors, scholars have identified a number of factors that make a party’s entry into a coalition more likely. Among these factors are the party's size, ideological positions, and degree of experience (Isaksson 2005). Martin and Stevenson (2001) found an incumbency advantage that makes entry into coalition government more likely for parties that already hold office—although this advantage diminishes when coalitions end in conflict or parties lose a substantial number of legislative seats. If either or both of these two outcomes occur, it is unlikely that these coalitions are renewed or that coalition talks are initiated—or if initiated succeed. Paradoxically, moderate electoral performance is not predictive of whether a party is likely to engage in coalition negotiations—this observation is not new, nor is it limited to the German case. Statistically, winning seats does not increase the likelihood of participation in government but losing seats decreases the likelihood, which makes it less important to be a winner of new seats than it is to be a loser of existing seats (Mattila and Raunio 2004).

What is common to all these perspectives is their assumption of rational behavior. Rationalist theories dominate many areas of political science in addition to the field of coalition building (Sunken and Schubert 2018). In negotiation, rationality especially requires refraining from posing threats that would decrease one’s own utility, as such threats are not credible (Janusch 2018). Smyrl (2002) showed that in negotiations, short‐term effects can be more important than long‐term rationalities, so that short‐term rationality (e.g., breaking off coalition negotiations to sustain the confidence of the electorate) wins over long‐term rationality (e.g., long‐term effects of a decline in party support in the aftermath of a break off of negotiations). In general, rational behavior is defined as an individual behavior that maximizes one’s own utilities or the benefits of the group that one represents. In party politics, rational behavior follows the goal of setting policy and maximizing the amount of offices occupied by the political party, and maximizing its votes.

Scholarly literature has only recently recognized the importance of actor‐level factors on the duration of time it takes for a government to form (Ecker and Meyer 2020). Helene Pedersen (2012) argues that when analyzing coalition negotiations, more attention should be paid to intraparty politics. Pedersen challenges the assumption of unitary party goals. Relatedly, coalition negotiation can be seen as a two‐level power game, in which a leader’s weak position within a party can be used as an advantage in negotiations if the leader persuades her negotiating partner that she lacks the power to gain support for agreements that are not in her party’s interests (Putnam 1988). In a similar vein, Moshe Maor (1995) argues that centralization of party structures is a disadvantage in coalition negotiations because party members who are disappointed by the negotiation outcome may no longer feel connected to the party. In decentralized parties, he argues, there are other organizational mechanisms that help the members to retain their party attachment. Corporate actors (among whom we count political parties) must maintain internal solidarity in negotiations while achieving external success (Cutcher‐Gershenfeld 2017). This article connects to this work by examining the role of party leaders’ individual characteristics in coalition negotiations and accounting for new outside influences such as social media on the formation of the recent German government. It thus focuses on the failure of the Jamaica talks and the successful formation of the grand coalition by highlighting the importance of intraparty power games through erosion of party unity (see Zittel and Nyhuis 2018).

Another factor that might prove helpful in explaining the irrational outcome (in terms of policy, obtaining government office, and vote seeking) is that of personal trust (Leach and Sabatier 2005). Fewer coalition partners and better knowledge of each other make intracoalition trust and therefore coalition formation between parties more likely (Bäck 2003; Skjæveland, Serritzlew, and Blom‐Hansen 2007). Trust is a result of a number of factors, including institutionalized long‐term communication and cooperation between negotiating parties, the reputation and credibility of each party, clear rules for beginning and ending cooperation, scrutinized perspectives, and jointly accepted sanctions (Saalfeld 1999; Irwin and Van Holsteyn 2004). Moreover, smaller negotiation units are more likely to trust each other than larger ones (Ostrom 1990). Over the last few years, the role of social media in political processes has strengthened the tendency toward personalization and more interactive modes of communication (Kuhn 2013). In these times, trust‐building mechanisms are even more important as communication is more far‐reaching and personal than before (Hegelich and Shahrezaye 2015; Sevin and Manor 2019).

According to established coalition theories, the parties most likely to form a coalition are those that

1. have not lost a large number of seats (are nonlosers of elections),

2. share ideological perspectives,

3. have large numbers of members,

4. have experience in government,

5. are decentralized, and

6. trust each other.

One would expect these parties to follow negotiation strategies that, at a minimum, enable them to implement their own policies to the greatest degree possible, satisfy the electoral base to ensure winning votes in future elections, increase the number of political offices held by party members, and/or enable party members already in office to maintain their seniority within the government. The final coalition between CDU/CSU and SPD presented the least likely case because the SPD had ruled out another grand coalition before the elections; their ideological intersections—although dependent on the policy area—were farther apart than those between CDU/CSU, FDP, and the Greens; and both parties had lost a significant number of votes and a significant amount of support. Considering the objectives of achieving policy goals, obtaining governmental office, and winning votes, Table One presents the potential payoffs for each party resulting from various coalitions. Words in italics denote goals that were realized and words in bold denote goals that were not realized. In the case of Jamaica, the outcomes are hypothetical; our predictions are based on opinion polls and policy preferences. In the case of the GroKo, actual outcomes are set forth. The outcomes of new elections cannot be predicted. The table emphasizes that in terms of rationality, the best outcome for all Jamaica partners would have been a Jamaica coalition. Surprisingly, the FDP quit the talks although the party would have gained in achieving policy goals, obtaining governmental office, and winning votes, as opposed to losing all.

Policy and office evaluations are based on the premise that government participation results in both for a given party—a premise derived from the previously elaborated theoretical considerations. Votes are assumed on the basis of opinion polls conducted by the independent “Research Group Elections” (Forschungsgruppe Wahlen 2019).

Drawing upon the factors linked to coalitions that are set forth above, this section takes a closer look at the political situation in Germany after the 2017 general election and how it led to the negotiation outcomes. Political outcomes in Germany are based on consensus; the country may be described as a consensus democracy (Strünck 2013). As a consequence, coalition negotiations have long been a part of German parliamentarism (Bandelow, Eckert, and Rüsenberg 2017).

Firstly, most of the parties in negotiation talks for a coalition government in 2017 were not “winners” (in terms of having won seats) of the general election; this presented a challenge to the party chairs and lead negotiators. CDU, CSU, and SPD won less votes and seats than they did in the general election of 2013. According to the reasoning set forth above, their loss of seats made engagement in a coalition government less likely. Moreover, the incumbency advantage did not seem to be in play, as the SPD’s resistance indicated a coalition conflict and of all parties entering parliament, the incumbent parties lost the most legislative seats relative to the number of seats won in the general election of 2013. The Greens increased their percentage of the vote but lost ground relative to the other parties, falling from the fourth strongest parliamentary party to the least strong among the seven parliamentary parties, garnering only 8.9 percent of the vote. A year before the election (in May 2016), political observers had predicted that the Greens would become the third strongest party with as much as 14 percent of the vote (Forschungsgruppe Wahlen 2016). Only the right‐wing Alternative for Germany (Alternative für Deutschland [AfD]) won more seats than the FDP; the AfD’s gains were largely responsible for the weak performance of the caretaker governing coalition of CDU/CSU and SPD (see Figure One).

Secondly, each party was experiencing an internal power struggle. Christian Lindner, the FDP’s chair, was the only actor with strong partisan backing. In 2013, the FDP did not reach the 5 percent of votes necessary to enter parliament; becoming the fourth strongest parliamentary party (out of seven) in 2017 was a major gain. Organizational restructuring after 2013 resulted in party stability and in support for Lindner as the leading figure. Angela Merkel, for whom the legislative period 2017–2021 will be the last period in office, was criticized from within the CDU, especially during the refugee crisis in 2015. Merkel had successfully fended off intraparty aspirants for the chancellorship in previous years. However, to appease her critics in the party, she put Jens Spahn—one of the strongest contenders for the chancellorship—in the cabinet.

As for the SPD, Martin Schulz became the party leader in March 2017, only months before the year’s general election. In the intervening months, three state elections (legislative elections in the German Länder) ended disastrously for the SPD. In two of the states, the SPD fell from its standing as the party of the Minister President and became the opposition party. In the CSU, a power struggle between the Minister President of Bavaria Horst Seehofer and Markus Söder over the party chairmanship and the Minister President position had smoldered and eventually overshadowed the negotiations (Luyken 2017). While the Greens did not suffer such a top‐level power struggle, some senior party members participating in the talks opposed forming the coalition and advocated becoming part of the parliamentary opposition. The party’s chairs, Katrin Göring‐Eckardt and Cem Özdemir, represented the wing of the party that advocated membership in a government coalition so they were not supported by the wing of the party that opposed such participation.

Before the start of negotiations to form a German government, the chair of the majority party usually invites the chairs of potential coalition partners—parties that may be of political and strategic benefit to the party—to exploratory meetings. If these talks are successful, the process of government formation continues. After the 2017 election, it quickly became clear who those potential coalition partners were, as the SPD had stated publicly that it would not join another grand coalition. Thus, the only parties left to form a politically compatible majority in parliament were CDU/CSU, the FDP, and the Greens.

The talks were the first attempt to form a Jamaica coalition on the federal level. All other attempts to bring these parties together occurred at the state level. Based on what occurred at the state level, it seemed very likely that a Jamaica coalition could form at the federal level. As recently as June 2017, a Jamaica coalition of CDU, the Greens, and the FDP had formed a government in Schleswig‐Holstein. Therefore, the public believed that a Jamaica coalition had a good chance of success; the confidence expressed by politicians and political scientists contributed to this view (Buck 2017). Although exploratory talks between party chairs are usually designed to resolve whether and not how they want to form a coalition, the Jamaica talks, which began on October 18, 2017, took another path. The talks resembled negotiations between parties that already had agreed to form a coalition; there were several concurrent negotiations at different party levels involving a variety of negotiators. Likewise, the partisan strategies and intraparty pressures resembled that of coalition negotiations. Arguing and bargaining became more complex with the involvement of more parties with diverging policy preferences, especially as Jamaica included parties on a spectrum that included left ecological (Greens), liberal economic (FDP), and Christian Democrats (CDU/CSU). At first, conflicts between the Greens and the CSU, and to a lesser degree between the Greens and the FDP, appeared as the major obstacles to forming a coalition. In particular, the parties disagreed about issues related to refugees, environmental policy, and economic policy (Luyken 2017). The challenge was to find policies on which they could all agree. Although the Greens publicly acceded to the other parties’ positions on climate and finance policy—notwithstanding the centrality of climate policy, along with animal welfare, to their agenda (Vogeler 2017)—other trade‐offs could not be negotiated.

As the FDP had the least amount of disagreement with the other parties over policy positions, it was surprising that the FDP was the party that finally ended the exploratory talks. Explanations based on ideological closeness, parties’ attempts to align policy positions with those of middle‐of‐the‐road voters, and election success do not, at first glance, explain the outcome of this first phase of negotiations. Upon further analysis, however, it is clear that the FDP’s policy preferences were not a significant factor in the coalition talks. As one FDP delegation member stated: “But if you have nothing at all—if three sit at the table and two work together against one, then it won’t work.” (interview by Nils C. Bandelow and Johanna Hornung with FDP delegation member on September 24, 2019) (see also Kubicki, Käfferlein, and Köhne 2019). Such a situation decreases trust. But why was that the case? Why did the other parties fail to take the FDP’s policy goals seriously and why did the FDP leave the coalition talks rather than continue to advocate for the advancement of a coalition agreement that reflected its policy objectives?

There are two possible reasons for this outcome. Firstly, the party leaders’ success was essential to the retention of their positions. The party chairs of CDU, CSU, and the Greens needed successful talks so they could stand up to their intraparty competitors. As the SPD had opposed a renewed grand coalition, the expectation was that a second round of elections would follow if coalition talks failed, a failure that would have ended the party leaders’ careers. In contrast, the FDP’s Lindner did not face intraparty opposition. He had the party’s strong backing and repeatedly emphasized the FDP’s indifference to holding new elections (Karnitschnig 2017). However, it does not seem that abandoning the talks was in the party’s interest. The alternative, a grand coalition, advanced policies that were significantly unaligned with the FDP’s objectives. Therefore, the FDP’s strategy should have been to fight for its policies. For example, the FDP and CSU wanted to reduce coal energy by three to five gigawatts, whereas the Greens wanted a reduction of eight to ten gigawatts; Merkel suggested that the parties compromise and agree to a reduction of seven gigawatts (Kirchner 2017). The grand coalition later convened a commission to make recommendations for the coal phase‐out; the commission recommended a reduction of twelve gigawatts by the end of the legislative period (Bundesregierung 2019). From the FDP’s perspective, this outcome is worse than any that might have resulted from a compromise within Jamaica. Additionally, the FDP is the only one of the parties participating in the exploratory talks that lost votes after the Jamaica coalition’s failure and that does not currently hold any governmental office. (The Greens also do not hold a governmental office.) The decision to declare the talks failed and leave them therefore seems to be influenced not by party objectives but personal gain. Senior FDP leaders were apparently motivated by a desire to hold their leadership positions; they expected that new elections would benefit them politically. They also believed that new elections would lead to a change in leadership of potential coalition partners, anticipating that the current leaders’ failure to build a coalition would result in their loss of power.

Secondly, the public’s considerable interest in the talks greatly exacerbated the difficulty of building mutual trust. Through social media—especially Twitter—and leaks, much confidential information about the progress of the talks and the actors’ preferences and behavior quickly became public. In contrast with the usual process of government formation, intraparty disputes and power struggles were visible to the public. Lindner did not want to enter a coalition that was hampered by the parties’ mistrust; in his letter to the FDP membership he wrote:

Moreover, Lindner did not want to join a coalition that would result in only partial realization of the FDP’s policy objectives. He was mindful of the fate of Guido Westerwelle, a former FDP chair. Westerwelle had achieved great results in the 2009 election and had joined a coalition with CDU/CSU but later was responsible for the FDP’s failure to gain any seats in parliament following the 2013 election, the first time the parliament lacked FDP representation. Westerwelle also had borne responsibility for unfulfilled coalition promises (Zohlnhöfer 2015). As noted above, Lindner had strong support within his party and did not need successful coalition talks to retain power. Hence, he placed his individual interest over the party’s interest. From the party’s perspective, leaving the talks did not maximize the party’s utility as compromising most certainly would have led to the FDP’s realization of at least some of its policy agenda and some governmental offices and—with positive communications to the public—votes of party supporters. Due to the lack of long‐term cooperation on the federal level and the destruction by social media and media leaks of the credibility and reputation of participating partisan members, trust could not be built and this—together with intraparty power struggles—seems to have led to the failure of the negotiations. However, it would have been more rational for the party to realize some policy objectives in a coalition agreement.

If these propositions are correct, the failure of the Jamaica talks is a logical consequence of the concurrency of intraparty power struggles and mutual mistrust; the latter was deplored by the participants in the aftermath of the talks (Delhaes and Kersting 2017). The power struggles and mistrust were exacerbated by an institutional design that resembled coalition negotiations rather than discussions on potential cooperation among a small group of party chairs. As a CDU member who participated in the talks argued, the scope of negotiations was a result of “vanity and mistrust of one’s own party leadership” (de Maizière 2019: 16), illustrating the conflict between personal and party interests.

After the Jamaica talks failed, the SPD again had a decision to make—whether or not to keep its promise not to participate in a CDU‐led government. President Frank‐Walter Steinmeier—a former SPD politician who took an unusually strong role as German head of state—pressured the party chairs both publicly and privately to uphold their political responsibility to form a functioning government. Steinmeier acted as the coalition’s formateur (Bräuninger et al. 2019) and the SPD’s party chair entered the exploratory talks on December 15, 2017.

Schulz emphasized that such talks were not guaranteed to lead to the formation of a government. He also thought it very important to follow a strict schedule and to exclude the public’s involvement by banning communication via social media—a lesson learned from the Jamaica experience. Pictures of politicians standing on the balcony of Berlin’s Parlamentarische Gesellschaft (building of the parliamentary society) had helped to form the public’s impression of the Jamaica negotiations; the public viewed those talks as much ado about nothing rather than constructive cooperation (Mudge 2017). In the 2017 talks, Schulz sought to shield the negotiations from anything that would impede the parties’ trust or otherwise negatively impact the talks, especially leaks on social media. Other than these considerations, the talks exploring a potential third grand coalition under Merkel proceeded in a manner similar to that of the Jamaica talks; the working groups were of comparable scope and size. Moreover, despite the parties’ stated desire to prevent document leaks, ongoing developments were often made public shortly after meetings ended¹. After only six days of talks (January 7–January 12, 2018), the parties jointly drafted a paper on the basis of which they were to enter coalition negotiations. However, it was necessary for the SPD’s party conference to approve the start of the negotiations. It did so on January 21, 2018 with approval from 56.4 percent of the party’s members.

The factors that account for the Jamaica talks’ failure also explain the success of the exploratory talks between CDU/CSU and SPD. The institutional design and organization of the CDU/CSU–SPD talks facilitated the parties’ trust and the parties had known each other for a long time as a result of participation in their collectively led government. Even though the SPD was not satisfied with parts of the jointly drafted paper, experienced partners are more likely to build mutual trust (Warwick 1996). Furthermore, as this constellation included the incumbent cabinet, the failure to reach an agreement probably would have resulted in a minority government or new elections. CDU/CSU’s party leaders opposed both alternatives; SPD’s party leader particularly opposed new elections, fearing the loss of his position and the power that went with it. Because Merkel, Seehofer, and Schulz all would have been replaced if new elections were held, they were motivated to achieve success in the negotiations (Martin and Stevenson 2001). However, the party’s interest—at least in terms of winning votes—was best met in failed coalition talks, as both parties’ supporters preferred new elections or a minority government over a grand coalition (ARD and Infratest dimap 2017).

Consequently, both parties were under pressure. Policy positions were clear and relatively easy to negotiate given their long‐term cooperation and greatly aligned policy goals. However, the importance of setting policy diminished notwithstanding the SPD’s conditions for entering a Merkel‐led government, which included the introduction of citizens’ health‐care insurance; improvements in pension, education, and housing policy; and a renewed demand for unlimited admission of refugees (Contiguglia 2017). In putting these topics on the agenda, the SPD tried to employ the lessons of the failed Jamaica talks for improving its own bargaining situation. But the final consultation paper addressed these policies only very marginally and vaguely (CDU/CSU and SPD 2018b), which again supports the argument that individual interests—much more than policy‐oriented party interests—were decisive in forming the government.

Vote seeking was likewise an unimportant factor in the negotiations. For the SPD, the insufficient integration of its policy agenda into the final agreement was problematic, triggering the major critique of the party base. SPD leaders—remembering the CDU’s failure to keep its promises to the SPD in the coalition agreement of 2013—feared that if another grand coalition formed and the CDU again failed to honor its commitments, the SPD would lose votes and continue to be weakened. The CDU’s earlier failure to keep its promises had led the SPD’s membership to mistrust the CDU’s representations.

After its historically low vote count in the 2017 general election (the SPD’s vote share was 20.5 percent), polls taken during the negotiations indicated a continuing decline in support of the SPD. Some polls even showed the SPD’s approval rate as lower than that of the far‐right AfD (Reuters 2018). For that reason, the younger generation, especially, tried to mobilize party members to stand up against another coalition with the Christian Democrats. This opposition endured even after the SDP’s vote at the party convention approved party leaders’ participation in coalition negotiations, and lasted until all party members approved the coalition treaty on March 4, 2018. Thus, despite the potential negative effects on vote seeking, the SPD proceeded with the negotiations—in large part because leaders prioritized individual self‐interest over partisan goods. Each politician wanted to secure his or her place in parliament. Office seeking (or office keeping) and the desire of every individual to ensure his or her political survival—along with increased pressure from President Steinmeier and the dire prospect of new elections—finally drove the parties to engage in talks.

After the party chairs had agreed to enter coalition negotiations and party members had voted—at their conventions—in favor of such negotiations, a coalition treaty was reached in only twelve days (January 26–February 7, 2018). After approval by the CDU/CSU membership at its convention on February 26, 2018 and approval by the SPD membership on March 4, 2018, the SPD finally joined a grand coalition that elected Chancellor Angela Merkel on March 14, 2018. Figure Two provides an overview of the negotiations.

Notwithstanding the announcement of both sides that a new grand coalition would not merely be a weiter so (“let’s have more of the same”), the coalition treaty suggests otherwise (CDU/CSU and SPD 2018a). Much of the treaty was similar to the coalition treaty of 2013 (CDU/CSU and SPD 2013), despite the SPD’s determination at its party convention to prioritize three policy objectives in the negotiations: (1) the end of the “two‐tier health system,” (2) the end of fixed‐term (temporary) employment contracts used by employers to save money on parental leaves and other benefits, and (3) family reunification for refugees (Nienaber 2018). The SPD’s efforts to advance the first goal resulted in the creation of commissions to examine doctors’ salaries in both the statutory health insurance (SHI) and private insurance schemes, out of which a holistic vision for the health policy sector might emerge (Hornung and Bandelow 2018; Bandelow, Hartmann, and Hornung 2018). A citizens’ insurance² that was promoted by then‐party chair Schulz during the exploratory talks—following the opinion of the SPD’s health‐care policy expert and vice chairman of the parliamentary party Karl Lauterbach—did not make it into the final treaty. This would have been a significant success that the Social Democrats could have highlighted for their members. Negotiations on fixed‐term employment contracts resulted in a ban on such contracts (except for justifiable reason) for the 2.5 percent of the workforce employed in firms with more than 75 employees. Smaller companies are excluded from this regulation and may continue to use fixed‐term employment contracts. Regarding refugee policy, the SPD and the CSU in particular disagreed on whether or not to impose a limit on the number of refugees to accept into the country; the issue previously had caused discord in the exploratory Jamaica talks. The parties reached a compromise on refugee numbers that both parties could present to their constituencies as partial policy successes (CDU/CSU and SPD 2018a). Therefore, as to the implementation of policy, SPD’s unitary party objectives were not fully met.

The SPD had announced that it would enter another grand coalition only when a majority of its members approved a coalition treaty. The party used this position to put pressure on the CDU/CSU in the negotiations not only to advance the party’s interests, but also to advance the party leaders’ personal office seeking goals. In advancing their personal interests, the relationships among the three leading social democratic politicians (Martin Schulz, Sigmar Gabriel, and Andrea Nahles) was important. Schulz and Gabriel had been friends for a long time (Göpffarth 2017). Schulz, acting to advance his own self‐interest rather than the party’s policy objectives, had supported Nahles—chair of the SPD’s parliamentary party—as the next party chair, opting to seek a role for himself as foreign minister. Schulz’s support of Nahles disadvantaged Gabriel, who had stepped back in favor of Schulz as candidate for chancellor following the poll results. With no prospects for a cabinet position, Gabriel perceived betrayed. The intraparty conflict was public knowledge and a significant concern to many SPD supporters: “[T]he turmoil at the top of the party has angered many SPD stalwarts, who accuse the party leadership of putting their own careers ahead of the party’s interests” (Karnitschnig 2018).

Successful coalition negotiations were important to Angela Merkel as she sought to assert her post as chancellor, especially against major detractors such as Jens Spahn (CDU), who ultimately took charge of the health ministry. For the CSU’s Horst Seehofer, who was fighting against Markus Söder for the role of Bavarian Minister President, a cabinet post in the new government and successful coalition negotiations were also essential for political survival.

The negotiations undertaken to form a grand coalition in Germany following the general election of 2017 hold lessons for understanding future government formation in Germany and beyond. It is necessary to note that the case study analyzed in this article presents a least likely outcome. Against this backdrop, it is ambitious to make predictions regarding future negotiation outcomes. However, this research does provide insight into the factors likely to affect the formation of future governments.

The successful formation of Germany’s recent grand coalition highlights the limitations of traditional explanations for the formation of coalition governments. Theories that view parties as unitary actors or focus on negotiation as a problem of delegation within parties—addressed by agency theory—do not adequately explain the formation of Germany’s fourth grand coalition. Such theories ignore the role of political leaders’ pursuit of individual interests and intraparty conflicts, which have become more visible to the public due to social media.

To an increasing degree, coalition negotiations turn on intraparty power games centered on politicians’ individual goals; the goals of parties negotiating as unitary actors are declining in importance (Zittel and Nyhuis 2018). During the process of government formation in Germany, the leadership of three of the five parties involved in the coalition changed during or shortly after the negotiations. In the SPD, Schulz resigned as party chair under partisan pressure following a conflict with his predecessor Sigmar Gabriel over ministry posts in the grand coalition. The party almost reached Hobbes’ natural condition—Schulz against Gabriel against parliamentary party chair Andrea Nahles; the party elites against the younger generation. The chairman of the CSU, Horst Seehofer, lost his post as Minister President in Bavaria in exchange for the less attractive ministry post in Berlin. The Greens changed their party chair following the failed Jamaica talks—Cem Özdemir, especially, has little chance of resuming party leadership—and climate policy has become a top priority for the party in the course of the 2019 elections for the European Parliament (Ceccobelli and Cotta 2016). The FDP’s support in polls decreased while the Greens profited immensely in the aftermath of the “Jamaica failure.” Christian Lindner’s interest had been to ensure his personal approval and support for the FDP. Although he succeeded in the first goal—he continued to be ranked among Germany’s top‐rated politicians (FOCUS‐Ranking 2019), polls showed a decline in the FDP’s approval as compared to the results of the last election. Nevertheless, a member of the FDP delegation stated, “I also know times when we were happy to have over 5%, and we’re not there at the moment.” (Personal interview with FDP delegation member on September 24, 2019). Thus, so the intra‐party backing for Lindner here is strong.

The process of government formation enables top‐level party personnel to gamble on their political posts and political survival. The intraparty backing of top‐level politicians therefore influences whether parties enter coalitions, as politicians act to advance their own interests and pursue party leadership roles and governmental positions even if their actions do not gain votes for the party or advance the party’s policy objectives. Paradoxically, policy seeking is also likely to remain crucial to gain party members’ approval.

Together with intraparty power games, the role of social media in building trust among coalition partners is substantial. While coalition negotiations in earlier decades largely took place behind closed doors, they are now visible to the public, with the possibility of each participant leaking important information—including the blaming and praising of individual actors—even as the negotiations unfold. However, mutual trust is essential for productive cooperation, particularly in political processes, and this does not refer only to interparty disputes but is equally related to intraparty politics. As a result, social media can substantially influence the success or failure of negotiation talks. Future research thus might benefit from an explicit focus on social media and its effects on the strategies of individual actors. In creating public pressure on individual actors, the media itself is likely an important factor in influencing negotiation outcomes.

In forecasting future political outcomes, a comparative look at Germany’s neighboring countries is worthwhile. France already has experienced the declining role of parties as unitary actors in its political system and the institutional structures provide an adequate response to this development through the strong role of the president. The election of Emmanuel Macron in 2017, in particular, confirms that established party structures are not needed for a successful presidential campaign. France certainly is more prone to such developments as a majoritarian democracy. But with an increasing individualization of the policy process and the specific role of social media in political communication, it becomes much harder to predict which coalition will govern in the aftermath of an election. This finding is especially relevant to other parliamentary democracies that increasingly face coalition governments or new coalition compositions, such as Great Britain or Italy, as new parties emerge and the importance of social media increases.