Grand Coalitions and Democratic Dysfunction: Two Warnings from Central Europe

Abstract

Are ‘grand coalitions’ – coalitions that include the two largest parties in a parliamentary system – good or bad for democracy? This article analyses that question in light of the recent rise of populist parties that large mainstream parties may try to exclude from government by forming grand coalitions with other large mainstream parties. I call this the ‘sterilization’ logic and note that mainstream parties’ ability to do this varies widely. Where parties have previously used grand coalitions primarily according to a ‘clientelistic’ logic – for example, Austria – mainstream parties may well be unable to rally the party system against insurgent populist challengers. Where mainstream parties have used grand coalitions exceptionally and for major institutional ‘renovations’ – for example, Germany – the grand coalition remains a viable option for responding to insurgent challenges, though this strategy is also quite risky. This article considers the implications for democracy by tracing recent developments in the context of the euro crisis and the European refugee situation. The major empirical referents are Austria and Germany – two countries with extensive experience and literature about grand coalitions – but the article draws out the implications of the analysis for other European parliamentary contexts.

Why have ‘grand coalitions’ of the two largest parties in a democratically elected parliament recently been seriously considered in places where they have rarely or never been used in the past, such as Ireland, Spain and Sweden? ¹ Why are they losing viability in places that once regularly used them, such as Austria, Belgium and the Netherlands? ² And are grand coalitions good or bad for democracy? These questions have drawn little scholarly attention. The coalition-formation literature takes little note of grand coalitions, usually preferring to see them mostly as a subset of ‘minimal winning coalitions’ that include any party (large or small) whose exclusion would cause the government’s parliamentary support to drop below 50 per cent (Miller and Müller 2010: 332; Müller 2008). Meanwhile, in part because both the usual ‘office seeking’ and ‘policy seeking’ hypotheses do a poor job explaining German grand coalitions (Spier 2013), a German-language literature has addressed grand coalitions focused primarily on the German and/or Austrian cases (Haas 2007; Stöss 2001).

Caught between indifference on one side and parochialism on the other, the grand coalition (GC) phenomenon deserves closer study because GCs are popping up in new places, while pressures build for them in places that have never (or rarely) used them. ³ For example, in addition to countries such as Austria, Finland, Belgium, Luxembourg and the Netherlands – where they have been common – Italy also tried a short-lived GC for six months in 2013. Next in line could be countries like Ireland (which, like Sweden, is trying a minority government precisely to avoid a GC), Spain (which used new elections in 2016 to avoid a GC), or even France (depending on how the Front National does in the 2017 National Assembly elections). Germany’s current federal GC is the country’s second in the last three governments, after having had only one in its first 56 years. ⁴

Overall, while GCs can be formed for many reasons, the rise of populist parties are challenging conventional party systems, exhausting and discrediting many classic coalition formulas, and shrinking the room for manoeuvre of large parties (along with their vote share). In a post-Brexit Europe, the seriousness of the populist challenge now seems a given. The question is whether the purported cure of a GC is sometimes worse than the populist disease. The answer depends, in part, on how extensively the cure is used: based on the German and Austrian experience, a coherent fear is that the more it is tried, the less it works. This is relevant for the 26 post-Second World War European democracies that have never or rarely used a formal GC of the two largest parties and for the 11 that have used them frequently. ⁵

In understanding the implications of GCs for the health of our democracies, this article foregrounds party choices. In most parliamentary systems, parties have critical roles in the expression of popular will and in the conception and execution of public policy. While institutional theories dominate extant accounts of coalition formation (e.g. Strøm and Müller 2003), my purpose is instead to enquire about the implications of a particular kind of coalition for democracy. Thus, where some structural accounts of contemporary Europe equate coalition choice of any kind with rearranging the deck chairs on the Titanic (e.g. Streeck 2014), this article assumes party choices over coalitions still matter for democracy.

These issues are also highly relevant to this Special Issue and the concerns indicated in the introductory article (Jones and Matthijs 2017). As the traditional democratic institutions of party coalitions appear less functional, new strategies arise. One strategic temptation is the GC, where the centrist parties close ranks, either to attempt major policy tasks or to guard against insurgent parties. Cognizant of this temptation, this article distils two warnings from the long Central European experience with GCs in Austria and Germany. The Austrian warning – relevant for states that have very often used coalitions of their two largest parties, such as Belgium, the Netherlands, Finland, Iceland, Latvia, Lichtenstein, Luxembourg and Switzerland – is that GCs used for one task may undercut their utility for other, arguably more important, tasks. The German warning – relevant for the European democracies that have rarely or never used GCs but who may be tempted to do so by rising populist challenges – is that the cure can sometimes be worse than the disease.

THREE LOGICS OF GRAND COALITIONS

This article distils three logics that stand behind large parties’ choices for GCs. ⁶ Briefly, the first logic I call ‘renovation’, and it sees large parties entering GCs together primarily to pursue policy goals unreachable in smaller coalitions, especially those goals requiring constitutional supermajorities. The second logic is ‘clientelism’ and sees GCs as a chance for large parties to channel benefits to their constituencies and to enjoy benefits themselves. The third logic I term ‘sterilization’, and it sees GCs as a strategy for large parties to limit the room for manoeuvre of (often new) parties unacceptable to one or both of them. ⁷ Broadly, European populism of both the left and right has made sterilization relatively more important in recent years, despite centrist parties’ well-grounded fears that short-term efforts to exclude populists from government may help them increase their medium-term support in parliament. Despite this danger, both the long-running euro crisis and the more recent refugee flows have pushed the sterilization logic closer to the forefront of large party choices without, of course, eliminating the other two logics.

The rise of populist parties (see Lochocki 2016; Mudde 2007) helps us understand both the heightened frequency of GCs in countries that have long used them – such as Germany and Austria – and also the rise of GCs in countries with little prior experience. ⁸ Broadly, the next section shows that Austrian parties’ long-successful instrumentalization of GCs for clientelistic purposes undercuts their capacity today to use GCs for sterilization: the party the traditional centre parties have tried to exclude is now larger than both and will probably leave them with no majority in the Nationalrat, even when in coalition together. Germany, meanwhile, retains (for now) the sterilization option but must confront the fact that the tool is a short-term instrument that also stimulates outsider challenges in the medium and long term (Haas 2007). Countries contemplating ‘first use’ of GCs should thus be closely attuned to the democratic dysfunctions that can come with its extended use. When the GC is a short-term tool for government exclusion of populists, it can easily give those populists a medium-term electoral boost, paving the way for their parliamentary inclusion. ⁹

Seen this way, the GC is a proxy for the ability of a democratic party system to generate enough agreement that the two largest forces can govern together, generally across the major socioeconomic cleavage. As noted, this article focuses on logics of party choices for GCs and the implications of such logics for democracy. Both states initially used GCs for renovation, but Austria did this at the federal level while Germany did it only at the state level. ¹⁰ Quite soon, Austrian politics turned the formal GC into a formal (and legal) system of clientelism. Germany saw less clientelism overall and, crucially, did not embed clientelism in federal GCs. As voter reactions against Austrian clientelism became a major cause of populist strength, it has long since lost utility as a tool for sterilization. In Germany, however, GCs remain a viable sterilization option, at least for the short term.

After developing these points, I then look at the euro crisis and the more recent refugee situation in both countries. The overall finding is that an important change in political conditions – the substantial rise of populist parties in parliament – has increased the importance of the sterilization logic across both major cases, but particularly in the refugee crisis. Smaller changes are seen in the other two logics, and these latter changes are not so clearly linked to the rise of the populist challenge.

THE AUSTRIAN WARNING: HOW CLIENTELISM OBSTRUCTS STERILIZATION

Post-war Austria quickly regained limited self-rule in 1945, and Austrian GCs have been common. Since 1945, Austria has had a grand coalition government for 40 of the subsequent 71 years. In every case, the same two parties were involved: the Austrian People’s Party (ÖVP) and the Socialist Democratic Party of Austria (SPÖ), whose joint domination of Austrian Second Republic politics eventually helped spark a populist reaction whose rise can no longer be checked by the cartel of insiders (Katz and Mair 2009).

Austria’s interwar history left ample reasons to doubt that the two large parties could work well together. Austria’s First Republic was formed in 1918 by a coalition of the Social Democrats and the Christian Social Party, predecessors to the SPÖ and ÖVP, respectively. Yet conflict between these two parties led to violence in 1927 and a brief civil war in 1934 before the German annexation in 1938 (Heinisch 2002: 9–10). At war’s end, Austria’s First Republic legacies and its Nazi experience naturally informed its Second Republic, including a widespread hope that deep divisions in society could be overcome. Thus, an initial logic of renovation was important early in the post-war period. The early grand coalitions were charged with rebuilding a nation after a devastating world war and taking key economic decisions, including nationalizing the oil industry, (re)building the social state and ensuring fair prices for farmers.

International pressure was also a factor in Austrian consensus rule. Austria was still occupied during the 1945 election, when the ÖVP won a majority but formed a cabinet including members of the socialist and communist parties (Müller 2003: 92). Inviting the communists into the government was partly a nod to Moscow, softening the USSR’s stance towards Austrian independence. But building domestic consensus helped placate all Austria’s foreign occupiers (Kuhn 2000: 26; Pelinka 1998: 24). The 1949 election saw the initial manifestation of the sterilization logic, as the new GC allowed the main parties to drop both the League of Independents (VDU), a favourite vehicle of former Austrian Nazis, and the communists, who also did not return to the government in 1949.

If both the sterilization and, especially, the renovation logics played some early role, however, clientelism was the primary logic that cemented the GC. Austria’s corporatist ‘social partnership’ between employers and unions had an analogue at the level of the two main parties, the so-called ‘Proporz’ system. Representatives from these parties, along with representatives from unions and employer organizations, were assigned positions in nationally owned financial and industrial institutions (Aichholzer and Willmann 2014; Katzenstein 1985; Rocca 1984: 22). ¹¹ Using their consistent involvement in government, these same parties soon built an empire of state and quasi-state structures in which both political camps placed supporters as employees and attempted to steer policy to benefit their clientele (Schermann and Ennser-Jedenastik 2014: 570). Thus, growing clientelism initially helped limit the appeal of other parties, and Austrian GCs routinely held over 90 per cent of parliamentary seats in the 1950s and early 1960s.

After 1966, however, one-party majority governments ruled Austria for nearly two decades. ¹² Yet, as GCs disappeared, the Proporz system was not fundamentally challenged. It continued to dispense jobs, rent-controlled apartments and government contracts to supporters of both major parties, largely irrespective of which was in power. One result was the rise of the Austrian People’s Party (FPÖ), the roots of which are complex but which criticized the way the two big parties divided the spoils and controlled the agenda (Art 2011; Mudde 2007). That said, in 1983 – before Jörg Haider became its chair – the FPÖ even formed a coalition with the SPÖ. ¹³

The 1987 election saw the first revival of the grand coalition, this time with the SPÖ in the strongest position and thus with the chancellorship. ¹⁴ At the same time, the FPÖ broadened its criticism of the two main parties to attack not just clientelism but now also immigration. Even with the FPÖ’s rise, however, the two Volksparteien still received a combined 84 per cent of the vote in 1987, and the various SPÖ–ÖVP GCs lasted until 2000. In 2000, the FPÖ’s vote jumped sharply to 27 per cent, slightly more than the ÖVP (Rose 2000: 31). When the SPÖ, with 33 per cent of the vote, failed to form a minority government, the ÖVP formed a short-lived coalition of the second and third largest parties with the FPÖ.

Once back in opposition, the FPÖ sustained its attack on the Proporz system and the clientelism of the Volksparteien. By 2006 their combined vote shares had fallen to 70 per cent, and another grand coalition resulted (Fallend 2008: 902). By 2013, ÖVP and SPÖ combined vote shares slipped to only 51 per cent of seats in the current parliament. Further erosion now leaves them, together, well short of a majority in opinion polls. Meanwhile, the largest refugee flows since the Second World War have brought Syrian, Iraqi and other asylum-seekers and migrants through Greece, the Balkans and into Austria. Though no federal election is due before 2018, the FPÖ has profited handsomely in regional elections, doubling its vote to over 30 per cent in Upper Austria and coming close to catching the SPÖ in traditionally red Vienna. ¹⁵ In May 2016, the FPÖ candidate came within a few tenths of a per cent of winning the presidential election, in which the candidate of the Greens prevailed (the ÖVP and SPÖ candidates having been eliminated in the first round with 11 per cent each). ¹⁶

Summarizing, the Austrian case displays all three logics of GC formation. The first two logics – renovation and clientelism – dominated the first two decades of the Second Republic. Post-war Austrian governments faced unusually stiff policy challenges, particularly in extricating Austria from the limited sovereignty of the occupation but also in overcoming a civil war in the not so distant past and sketching the basic rules of the mixed economy (Miller and Müller 2010; Pelinka 1998). By no means can we claim such challenges obliged Austria to choose a GC, but they certainly seem consistent with the ‘renovation’ logic. GCs also allowed frequent constitutional changes in the first two post-war decades but also again in the early 1990s GCs, when Austria underwent major institutional changes to prepare for EU membership in 1995 (Müller 2008).

Developing at roughly the same time – but lasting far longer – we also have the ‘clientelistic’ logic in play. As the two main parties laid down the rules of the game, they also built an entire ecosystem of Proporz institutions that gave the two main parties durable influence over state and quasi-state institutions (Pelinka 1998). In the short term, this system actually had a centripetal effect, generally undercutting incentives for new parties that would have no access to the spoils of this cartel. Indeed, so effective was the initial Austrian system in retaining votes in the two large parties that each eventually found ways to govern Austria on their own. In the long term, however, clientelism – which prevailed even after GCs gave way to single-party governments – paved the way for the rise of challenger parties from outside this consensus. These parties took a long time to break through, but when they did – particularly the FPÖ in the early 1980s – they sought to remove this institutional glue that benefited the two traditional parties of government (Kudrna 2015). ¹⁷

Against the backdrop of this long two-party dominance, the ‘sterilization’ logic played an initial role only in 1949 with the VDU. Much later, when the FPÖ (a descendant of the VDU) radicalized around issues of immigration after the mid-1980s, the sterilization logic once again played some role. The SPÖ first broke a coalition with the FPÖ over Haider’s rise to party leadership, and both parties manoeuvred to minimize FPÖ influence, though not very successfully (Art 2011). Eventually, the ÖVP also attempted a coalition with the FPÖ before returning to two subsequent GCs. As we will see below, the sterilization logic has subsequently grown in importance as the euro crisis and European refugee situation have both fed the FPÖ. Yet the long use of the clientelistic logic has so eroded the GC’s justification that it is not clear whether today’s Austrian GCs can command enough parliamentary seats to form an effective barrier to radical parties. As I develop further below, the overuse of clientelism has left Austrian GCs an ineffective sterilization tool. This pattern should serve as a warning to other European states that have long relied on GCs, though the clientelist danger seems weaker to states that may be contemplating ‘first use’ of a GC.

THE GERMAN WARNING: COULD THE STERILIZATION CURE BECOME WORSE THAN THE POPULIST DISEASE?

The German case is quite different and more complex, as states play much more central roles in German than in Austrian politics. GCs at the German federal level have occurred just three times (including the 2013–17 government), and both previous GCs ended after a single term. On the other hand, GCs were used early and often at the state level after the Second World War, later tapering off in the 1960s and totally disappearing in 1970, only to reappear after 1990, first in post-communist eastern Germany but then also in several western states. Early post-war state-level GCs were driven primarily by ‘renovation’ logics in a shattered society, while the ones since 1990 are driven far more by ‘sterilization’ logics as new parties perceived as extreme enter the party system from both left and right.

Germany went nearly 20 years before its first federal GC. The Allied powers did not allow national elections until 1949. Many state elections had occurred by then, and many had produced GCs, often in oversized ‘renovative’ coalitions that included nearly every party in the state parliament (Haas 2007; Stöss 2001). Yet both the Christian Democratic Union (CDU) and the Social Democratic Party (SPD) ruled out a federal grand coalition in advance of the 1949 elections. Though the election disappointed both large parties by putting nine other parties (11 total) in the Bundestag, CDU leader Konrad Adenauer nevertheless cobbled together a four-party ‘small’ coalition with a single-seat majority – famously, his own vote. Subsequent Bundestag elections returned more stable majorities for CDU-led small coalitions, although Adenauer frequently toyed with the idea of GCs – a move supported by the left wing of the CDU – and also a way of keeping pressure on his usual coalition partner, the Free Democrats (FDP). ¹⁸

Yet despite the relative paucity of federal GCs, a major scholarly tradition sees Germany as a kind of de facto GC (Katzenstein 1985; Schmidt 1996, 2015). Schmidt calls Germany a ‘grand coalition state’ primarily because of the logic of bicameralism and federalism, which requires a government to work very closely also with the upper house (Bundesrat) on roughly 60 per cent of bills. German governments often do not enjoy a majority in the upper house (states vote as a bloc, and any state ruled by a constellation other than the one that controls the Bundestag cannot be counted on to support the government). ¹⁹ According to this institutional logic, even normal small ruling coalitions in Germany need the opposition’s de facto assent to get its agenda through parliament. Moreover, the federal level depends on the states for policy administration. Despite this need for cooperation, however, German corporatism never took on the clientelist dimensions of the Austrian version, in part because the consistent junior coalition partner FDP could block the dynamic developing in Austria under the early GCs.

Thus, when in December 1966, a German GC was formed at the federal level under Kurt Kiesinger of the CDU, the primary motive was neither clientelism nor sterilization but renovation. ²⁰ In the recession of the mid-1960s, Germany’s CDU–FDP government had initially trusted the economy’s capacity to right itself. But as unemployment grew and production stagnated, the governing parties split over economic matters, and the CDU turned to the only other party in parliament, the SPD. The new grand coalition controlled 95 per cent of the Bundestag seats in 1966 and operated for the rest of the legislative term, ²¹ ushering in tax reforms and Germany’s first significant experiment with Keynesian economics. The economy performed reasonably well during the first GC. Unemployment dropped, industrial orders and the exchange rate rose, and prices remained stable (Zundel 1968: 368). Continuing the renovation logic, several constitutional changes were enacted, especially in foreign policy towards the East. These included opening diplomatic relations with Romania and Bulgaria and recognizing the Oder–Neisse border line (Garton Ash 1994; Kundnani 2015). ²²

While sterilization concerns were not paramount in the major parties’ 1966 coalition considerations, some state elections had seen candidates of the radical right National Democratic Party (NPD) elected to state parliaments. ²³ The quick rise of the NPD rattled the major parties, and some doubts arose about what had hitherto been taken for granted – that the Federal Republic had nothing in common with the Weimar Republic, whose democratic centre had been eroded by extremists of left and right. ²⁴ Memories of the economic crisis of the late 1920s were not the least of the factors that pushed the Social Democrats into the first federal grand coalition. Of course, the mid-1960s recession was far milder (and turned out to be much shorter) than that in 1929 and, in contrast to the Weimar Republic, the population overwhelmingly accepted the new state and parliamentary democracy. But the memories were still raw, and the affluent society, unaccustomed to crises, reacted with alarm to the recession (Zundel 1968: 368).

If the rise of the right was, at best, a minor force pushing for a GC, did the resulting GC cause the further rise of extremism? The evidence is not clear-cut. Bolgherini and Grotz claim that sparking radicalism is an unjustified allegation against the first GC (2010: 148; also Hildebrand 2006). The NPD’s rapid rise preceded the onset of the GC, though it continued for a time after 1966. At a certain point, it seemed almost inevitable that the NPD would have a substantial bloc in the Bundestag after the 1969 election (Mudde 2000: 25–6). That it did not – it reached just 4.3 per cent in the 1969 election – can be attributed in part to the government’s threat of a ban on the party, which may have discouraged some voters, and the party’s own internal battles, which were temporarily papered over by electoral success in several state elections. On the left, this period also saw the rise of Germany’s ‘extra-parliamentary opposition’, part of which later morphed into the Green Party but other parts of which turned to domestic terrorism (Karapin 2007).

Thus, if GCs can act as a tool for large parties to seal off (‘sterilization’) rising radical parties from government, it is also possible (and widely believed likely in Germany) that GCs can stimulate the growth of extremist parties in the parliament itself (Lohse 2014; Münchau 2016). This intuition, with major implications for popular and procedural democracy, arises not just from the experience of the late 1960s but from the post-unification experience as well. As noted, GCs at the state level disappeared entirely between 1970 and 1990 but soon became very common in eastern Germany after 1990. ²⁵ Here, sterilization appears to have been a major motivation. Using only the post-1990 coalitions across all 16 German states, Spier tests a number of propositions about the cause of GCs and finds the strongest evidence for the presence of parties in the legislature that other parties reject as suitable coalition partners – the sterilization logic (2013: 501–5). Thus, GCs have become frequent in eastern Germany due in no small part to the durable rise of the Left Party, while GCs in western Germany are linked primarily to the rise of right parties, especially in Berlin, Baden-Württemberg and Bremen.

Meanwhile, after a 35-year hiatus, the 2005 election led to a federal GC when neither the CDU nor SPD could win a parliamentary majority with their usual allies. The 2005 grand coalition was not the parties’ preferred choice but was perceived as the only viable option to build a stable cabinet (Bolgherini and Grotz 2010: 1–2). A CDU–FDP–Green coalition was ruled out by both the CDU and the Greens. ²⁶ On the left, the sterilization logic appeared when the SPD ruled out a coalition including the Left Party (Green participation would also have been required). ²⁷ Renovation impulses were similarly modest, as the only major challenge embraced by the two parties was a constitutional reform of federalism. While limited, this reform did reduce the percentage of bills requiring Bundesrat approval and thus did soften the underlying institutional incentives that tend to push German parties towards an explicit or implicit GC (Moore and Jacoby 2010; Schmidt 1996, 2015). ²⁸

Finally, a GC was formed again in 2013. Here, the renovation logic that drove federal coalition choice in 1966 and contributed in 2005 had shrunk further. The two parties’ coalition agreement actually set out few major new tasks, although the ongoing euro crisis may have led the CDU to anticipate the need for further constitutional changes to advance fiscal, economic or banking union proposals then under consideration. Still, it was the combined weight of populist challenges on both left and now right that most complicated any choices besides the GC. Put differently, populist success helped drive the FDP from the Bundestag and deprived the CDU of its preferred partner. Indeed, the post-2013 GC represents a strong supermajority of 80 per cent (503/630 seats) in the Bundestag, in part because two parties (the FDP and the new Alternative for Germany or AfD) were close to but under the 5 per cent threshold. With their voters thus unrepresented in parliament, the remaining four parties each received seats well above their vote shares. ²⁹ The CDU could have chosen any of the remaining three parties to make a majority and unsurprisingly approached the SPD rather than the Greens or Left Party. Data from Forschungsgruppe Wahlen in Figure 1 show the evolution of voter preferences for the three most common federal coalitions. The GC – indicated by the white triangles – was the choice of a plurality of voters in both autumn 2005 and again in autumn 2013 (i.e. just prior to the respective elections). This was, however, a far cry from the much stronger voter assent in 1967, when 72 per cent of voters preferred it to the single available alternative, with a further 10 per cent then opposed and 18 per cent undecided.

Figure 1 Coalition Preferences (1998–2016) Source: Forschungsgruppen Wahlen.

In summary, the German–Austrian contrast is instructive at both federal and state levels. In Austria, GC-bred clientelism eroded support for the major parties and now leaves them unable to secure a majority that can contest radical parties using a traditional GC. German clientelism is both more limited than in Austria and is not embedded in an analogue of the Proporz-cum-grand coalition system. Instead, Germany’s federal GCs have been used briefly and more for renovation, at least until 2013. As a result, so far, they still retain a capacity to adopt a sterilization logic, which the SPD does vis-à-vis the Left Party and the CDU could do if the AfD grows stronger still. At the state level, the experience has been in between these two extremes, with both renovation and sterilization options remaining intact in recent years. Where the parties in the Austrian GC now enjoy an aggregate support of 51 per cent of the seats and around 46 per cent of the projected vote, the German governing parties currently have 80 per cent of the seats and roughly 56 per cent of the projected vote. ³⁰ The right populists, FPÖ and AfD, are at 34 per cent and 13 per cent, respectively.

With Austrian sterilization having reached the end of the road, the drama in Germany will be to see if the GC can shrink the AfD below 5 per cent by the autumn 2017 elections. To understand the striking surge of both parties, we need now to look back on the enormous stress tests of the euro crisis and the refugee crisis.

GRAND COALITION POLICY DEBATES I: GERMANY, AUSTRIA AND THE EURO

There is a reasonable fear that grand coalitions will dampen debate about important issues, which links to concerns in the introduction to this Special Issue about both ‘representativeness’ (openness of democracy) and ‘accountability’ (accord with popular views) (Jones and Matthijs 2017; Mounk 2014). ³¹ That fear about dampened debate can be illustrated with the case of the eurozone crisis, which posed an unprecedented challenge in terms of institutional renovation, both domestically (e.g. fiscal stimulus, banking rescue and subsequent regulation) and internationally (e.g. bailouts, banking union, fiscal compact). Both Germany and Austria quickly developed a working consensus between the two largest parties, a consensus that remained intact even when, in the German case, the formal GC was soon replaced. Crudely, both countries’ euro crisis policies were highly popular (accountable) but low on representativeness (openness to a range of views) (Jones and Matthijs 2017).

The crisis hit the 2005 German GC near the end of its term and, contrary to government rhetoric and popular belief, Germany’s initial stimulus packages were in fact well above that of the OECD average (Vail 2014). Moreover, while the 2009 elections brought a CDU–FDP coalition that tried a more market liberal approach to the subsequent euro crisis, the CDU soon decided to work with the SPD in a kind of implicit grand coalition. The FDP was slowly marginalized in this policy area (Zimmermann 2014). Instead, with growing defections from both CDU and FDP backbenchers, Chancellor Angela Merkel’s euro-rescue policies increasingly relied on the votes of the SPD (and Greens), who extracted concessions in the form of a private sector involvement in bailouts, a financial transactions tax and labour market measures for Southern European youth. Thus, from a policy perspective, the already orthodox German approach to the euro crisis would have been even more orthodox without the SPD’s participation (Jacoby 2015).

The cross-party differences were small, with debate dampened in intensity and truncated in range. The 2013 Bundestag election virtually ignored the euro crisis, and the subsequent coalition agreement simply formalized the implicit GC that already existed. The CDU and SPD had little reason to discuss the crisis, which made it difficult for smaller parties to highlight the issues. The CDU routinely asserted successful German management of the crisis, insisting that the crisis was caused by excess debt and regulation in certain eurozone member states and that the solution was less debt and more growth-enhancing structural reforms in those states. This argument worked well with most voters, and the opposition SPD concluded it was better off echoing the broad CDU line. Thus, the euro crisis never achieved a central place in the SPD’s electoral strategy in 2013.

Meanwhile, the CDU knew many of its voters were unconvinced by the wisdom of bailing out Greece, Portugal and Ireland and of preparing a large fund for the potential eurozone rescues. Many CDU leaders were also unnerved by the European Central Bank’s more aggressive monetary policy, including its programme for Outright Monetary Transactions announced in summer 2012. Eventually, the CDU–FDP coalition faced an electoral challenge to its crisis-management strategy not from the SPD but from the AfD, which barely missed the 5 per cent hurdle for entry into the 2013 Bundestag. While AfD drew votes from several parties, its initial leadership was mostly former CDU and FDP officials. ³² Thus, the CDU had few incentives to address the euro crisis, since it was vulnerable to a challenge from an ordoliberal direction. ³³

The two electoral programmes confirm that the euro crisis was largely depoliticized. In the SPD’s 120-page 2013 electoral programme, the word Eurokrise (euro crisis) appears just once, and there is no separate discussion of the euro (SPD 2013). The CDU gave the euro crisis issues even shorter shrift (CDU/CSU 2013). Its electoral programme emphasized that the primary adjustment to the euro crisis – characterized as a ‘debt crisis’ (e.g. CDU/CSU 2013: 3, 8, 9) – had to come from the hardest-hit states themselves, and not from European initiatives. Together, the two big parties mostly kept euro issues off the campaign agenda. Of course, the AfD lambasted them for saying little about the euro crisis (as did the Left Party from an anti-austerity perspective (Die Linke 2013)). But with the two largest parties resolutely committed to playing down the euro crisis, neither received much traction. Euro issues also played a very modest role in the formal coalition agreement between the CDU and SPD, mostly hewing to the CDU’s programmatic themes (Jacoby 2014).

In popular terms, this policy had succeeded handsomely. Indeed, the CDU – the party most identified with it – nearly won an absolute majority in the Bundestag, falling just five seats short. ³⁴ Both parties that raised serious doubts about Germany’s bailout policies from a liberal perspective failed to reach 5 per cent, further evidence that German voters were not much troubled. In other words, the GC articulated only a narrow range of policy, and yet German voters still showed very little anxiety about this range of options. This strange constellation is underpinned by Germany’s dominant role in the eurozone reforms – reassuring to voters – and because Germany also benefited in the short term from the combination of extremely low interest rates (flight to safety) and a weakening euro, which made exports boom and unemployment shrink. ³⁵

Unlike in Germany, Austria’s GCs have remained in place throughout the crisis, and, with no equivalent of the FDP to placate, there has been even less open contention over rescue policies. Austrian economic performance has generally tracked Germany’s – for example, a sharp contraction in 2009 followed by better-than-eurozone-average growth and employment figures ever since. As in Germany, the social democratic tradition has made scattered criticisms of the government’s overreliance on a pro-austerity policy mix for crisis-hit states. ³⁶ But, as in Germany, these criticisms remain half-hearted and have virtually no effect on government policy. As in Germany, it has been the right-wing populists who have most challenged the government line. The state of Carinthia, governed by an offshoot of the FPÖ, challenged the government bill establishing Austrian participation in the European Stability Mechanism (ESM) before the Austrian Constitutional Court. The national FPÖ referred in the parliament to the ESM laws as a ‘coup d’état’ (Jaros 2014: 29).

Thus, the two core countries had very similar policies and very similar domestic responses under their GCs, despite the very different historical trajectories developed above. So far, the stress test of the euro crisis has been passed, as voters in both countries largely accepted their government’s policies. True, Germany is far larger, but since EU bailout guarantees are proportional to GDP sizes, the average Austrian citizen was roughly as exposed to the bailout as the average German citizen. And in any event, little money was paid out. The key to the relatively mild populist response – mostly confined to court actions – was that both GCs maintained a common position among the two parties and used the EU to push adjustment costs primarily onto Southern European debtor states (Matthijs and Blyth 2015). Austrian populists in the FPÖ made modest gains, and German populists in the AfD rose to (not quite) 5 per cent, but the test was passed. Both governments managed to be accountable, albeit without representing much of the range of opinion.

In both states, populist critics found relatively few footholds. They could not point to large internal differences among the GC parties, nor to large actual bailout costs (as opposed to guarantees that might or might not be called upon in the future). In the subsequent and much harder stress test of the refugee case, both of these features would change. And an important theoretical point from the comparative populism literature would find empirical validation: when debate on economic issues narrows, debate may be displaced into identity politics, in which populists have significant advantages – advantages that are best exploited when centrist parties begin to disagree with one another (Lochocki 2016).

GRAND COALITION POLICY DEBATES II: AUSTRIA, GERMANY AND REFUGEES

After the fear that GCs dampen debate, a second major worry with GCs is that their centrism will spark the growth of extremist parties. German parties fear this possibility and often invoke Germany’s first federal GC, when the NPD doubled its vote and the extra-parliamentary opposition got its start. Worrisome evidence is even stronger at the state level. Here, the post-unification resurgence in GCs has been partly a reaction to parties with whom the SPD and CDU would not go into coalition (‘sterilization’), but GCs also seem to spur the rise of such parties in electoral terms. In eight of the ten state-level cases investigated by Haas (2007), the PDS/Left Party increased its vote share in the next election while in six of ten the radical right profited electorally (though by a smaller margin). Thus, efforts to bar populist parties from government through GCs can strengthen their electoral hand in the German context. The broader fear is that sterilization – even where still available – is a wasting asset. Parties, of course, worry about this, and it makes the choice for sterilization a difficult one. As the recent Irish, Swedish and Spanish cases referred to earlier indicate, centrist parties often go to great lengths to avoid GCs if they can find another way.

This dilemma played out in autumn 2015 as Austria and Germany responded to the European refugee issue. The two states began on the same course, and Germany’s GC largely stayed this course, even while divisions within the GC helped the new populists push identity politics to centre stage. ³⁷ In Austria, similar divisions occurred, but the better-established populists surged far ahead of the GC parties, helped bring down the head of government and forced a major shift in asylum policies. This will culminate in the end of the Austrian GC in the next election – and perhaps for good. Unlike with the euro, this stress test proved too much for at least one government.

The dramatic refugee situation had intensified for European governments in summer 2015, as the brutal civil war in Syria and the rise of new human trafficking routes contributed to a sharp surge in refugees coming into Europe. In September, Merkel decided – largely unanticipated by prior policy steps and absent cabinet approval or even a statement to parliament – to accept into Germany the thousands of refugees gathered in chaotic conditions around Budapest’s Keleti train station. Though she apparently consulted no other European leaders, Merkel did consult with then-Austrian Chancellor Werner Fraymann (Stone 2016). By November 2015, up to 6,000 asylum-seekers arrived in Germany daily; about a million came in 2015 (Focus Online 2016).

Meanwhile, the Austrian GC initially pursued a very similar policy, taking in about 90,000 asylum-seekers, the highest number in Europe on a per capita basis. But while both GCs started shoulder to shoulder, Austria’s GC soon changed course radically and openly. And while populist movements in both countries benefited from the issue, Austria’s far more entrenched populists benefited more, consistent with the argument that long ‘overuse’ of the GC for clientelism has weakened its utility for sterilization. Unlike on the euro, both populations were sharply divided, with many Austrian and German citizens signalling a ‘welcoming culture’ and others reacting with anxiety, especially as large numbers of refugees arrived in hundreds of towns and villages across both countries.

As public sentiment grew, the Austrian GC sought to limit the flow. In January 2016, it tried to cap annual asylum inflows at 37,500. A related proposed limit of 80 asylum-seekers per day through the southern route contributed to the Balkan governments’ decision to close the ‘Balkan Route’, a decision the Austrian government helped implement by building some new border fences. Once challenged in court on the daily and annual limits, the Austrian government then passed an emergency law that allowed officials to process asylum requests at the border. ³⁸ The border process is supposed to determine – often within hours – if grounds exist for returning the applicant to a safe third country. As it continued to lose popularity – even after the SPÖ chancellor resigned in May 2016 – the GC adopted many FPÖ positions (Siebenhaar 2016). While Merkel has condemned the Austrian policy change, there seems little doubt that the closing of two major routes into Central Europe had dramatically lessened the pressure on her government through summer 2016 (Karnitschnig 2016).

Like the FPÖ, the AfD also soon surged in national polls. Languishing around 5 per cent in the polls as the euro crisis receded from the headlines, the AfD had expelled the ordoliberal economists who founded the party. The prescient shift to new leadership intent on exploiting German fears about integration of asylum-seekers and the possibility of domestic terrorism took place shortly before the largest surge of refugees. The AfD then did very well in the three state elections in March 2016, gaining 12 per cent in Rhineland-Palatinate, 15 per cent in Baden-Württemberg and 24 per cent in Saxony-Anhalt. In all three states, both major parties telegraphed their intention not to enter a coalition with the AfD before the elections. In Austria, as noted earlier, the FPÖ, which had taken a slight lead in the opinion polls already in 2013, saw that lead grow as the refugee situation intensified and nearly won the presidency in May 2016.

But while the Austrian GC shifted policy to a much more restrictive track and backed the change with clear messaging to voters, the German GC made more modest changes, despite a sharp rise in voter concern. Infratest/Dimap polls from September 2015 and February 2016 document the erosion of support for the GC’s policies. Figure 2 shows that in September 2015, only 28 per cent of German voters were satisfied or very satisfied with the government’s policy, with only Green voters even approaching a majority of support (45 per cent). February 2016 results (Figure 3) indicated only 18 per cent of voters felt the German government had the asylum and refugee situation ‘under control’. Despite subsequent policy shifts and an accord with Turkey that helped lead to a sharp drop in refugees, support for the GC has recovered only modestly. ³⁹

Figure 2 September 2015: Satisfaction with Asylum and Refugee Policy Source: Infratest Dimap.

Figure 3 February 2016: The Government Has the Refugee Situation Under Control Source: Infratest Dimap.

The German GC passed two omnibus asylum bills (Asyl I and Asyl II), in September 2015 and January 2016, respectively. ⁴⁰ But these had no upper limits on asylum and were not perceived by voters as a major shift. In part, this was due to intra-coalition disputes that cost the GC credibility with many voters. The most serious challenges to GC policy have come from the CSU (the CDU’s Bavarian sister party), which has blocked unrelated bills to protest the government’s refugee policy. It has also threatened to challenge that policy before the constitutional court. ⁴¹

Thus, while populist gains were roughly comparable in the two states across the two cases, the overall results were quite different. While the German AfD went from just under 5 per cent in autumn 2013 to the low to mid teens in November 2016, the Austrian FPÖ went from 20 per cent to 34 per cent. The absolute jump in Austria was not much higher, and yet roughly the same combination of centrism and GC infighting proved more toxic to GC fortunes in Austria. Austrian policy underwent sharp changes, German policy much less so. Despite the changes, the Austrian chancellor fell over refugee policies; the German chancellor did not. Further, while the GC parties’ centrism appeared not to hurt them on the euro – where intra-coalition differences were few – their infighting likely handed populist challengers even more advantages on the refugee question. In Austria, where the GC was already structurally weak from years of populist attacks, this was enough to spark major policy changes and, admittedly more speculative, bring the end of the era of grand coalition politics firmly into view.

CONCLUSIONS

Populist parties increase the temptation for grand coalitions by weakening existing parties and leaving them fewer choices over government formation. But GCs also generate centrist politics that can further spur the growth of populists. These tendencies suggest GCs should be used with caution. In Austria, where GCs were institutionalized around a clientelist logic – jobs, apartments and public contracts for supporters – they eventually undercut the conditions for their own reproduction once the fiscal scope for public employment and housing shrank. Time will tell if the SPÖ and ÖVP are similar to the case of the Netherlands and Belgium where GCs were a way-station on the road to near-total collapse for the Christian Democrats. ⁴² Certainly, the ‘Austrian warning’ is that the two traditional parties may not be large enough to form GCs alone in the future. In Germany, where GCs were used more often for renovation – after the Second World War and after reunification at the state level – their utility for sterilization seems more durable, but for how long? Thus, if populism is a ‘disease’, the ‘German warning’ is that grand coalitions are a risky cure.

This article made this case in two steps: (1) sketching three analytically distinct logics that guide party choices for GCs; (2) providing evidence from two national cases about the democratic implications of GCs during the euro crisis and refugee crisis. In the broader European context, the concept of the grand coalition is really a question about whether parliamentary systems can address core challenges on the backs of their two largest parties. The article argues that this is still possible but may be more likely in states that have not already been using GCs for a long time. The German states eschewed GCs for decades in the 1970s and 1980s, returning to them mostly as a tool to marginalize new parties after 1990. Yet data show that GCs do seem to lead to a higher voting for both left and right populists in the next election. At the national level in Germany, the long-feared tendency of GCs to produce extremism may finally be showing signs of coming true after the 1-2 punch of the euro crisis and the refugee inflows. Still, the extraordinary refugee politics in Germany underscore the substantial potential GCs have for institutional renovation and policy dynamism. The problem is that they seem to have shaky popular support from voters. Meanwhile, the euro crisis led to massive renovation to the European economic architecture that while constitutionally contested inside Germany, retains a high level of popular support.

There remains an unsolved chicken-and-egg problem in that GCs can both cause problems for democracy (by stimulating the rise of new populist parties) and also be a party strategy for dealing with those very problems (by excluding those parties from government). We saw that the German AfD arose in response to policies launched by one GC and flourished by criticizing the policies of a second GC. GCs can spur the rise of populist parties, and at the same time, we can also clearly see that GCs are sometimes an attempt by the large parties to exclude rising populist parties from governing, what this article calls ‘sterilization’. For some, this is simply a self-defeating response, as each GC further increases the scope for extremist success among voters. And yet it is not clear that democracy would better flourish if parties chose to include such parties in the government. After all, this is what Finland is trying with the Finns Party, which is indeed the country’s second largest.

Finally, this article assumed that contemporary European parliamentary democracy is either healthy enough to survive GCs – where these are shown to have deleterious effects on democracy – or to benefit from GCs – where these are shown to provide a medicine more powerful than the underlying ailments. In short, it assumes our democracies are vulnerable, but they are not beyond repair. This cannot be taken for granted. For example, Streeck (2014) has argued that Western democracies found various ways to ‘buy time’ but have now turned into ‘consolidation states’ in which the state–creditor relationship has priority over the state–citizen relationship. ‘Renovation’ is reduced to guaranteeing the state’s creditworthiness. In this scheme, GCs are simply a manifestation of an underlying reality that no state dares defy creditors and that all states seek the benefits – primarily low interest rates – that come with the reassurance of such creditors. Thus, democracy is deeply compromised. That said, though the damage may be caused by other forces – especially the defection of capital from the post-war settlement – GCs (like technocratic governments) are complicit in this long-term damage. Indeed, it hardly matters which parties run the game, and Streeck – in an interview – referred to the SPD and CDU as ‘two wings of a national unity party, with the Greens eagerly waiting to join them as the vegetarian subdivision of the Christian Democrats’ (Streeck 2015). While this article starts from a more optimistic premise, it cannot directly refute this negative one.