One state, one interest? How a historic shock to the balance of power of the Bundesbank and the German government laid the path for fiscal austerity

Abstract Many economies in the Western world have been through a regime shift towards fiscal austerity since the 1970s. Existing scholarship ascribes trends in austerity to globalisation or the influence of a new economic paradigm. This paper develops a different approach by stressing the strategic intervention of central banks in governments’ fiscal decisions. It analyses archival documents from the German Federal Cabinet and the Bundesbank Council over more than two decades (1960–1981) and finds that the fiscal regime was shaped by changes in transnational institutions which the central bank used to strategically expand its institutional power within the larger macroeconomic framework. With the breakdown of the Bretton Woods system, the Bundesbank was able to greatly increase its power resources while the government’s powers diminished. The Bundesbank used its new powers to strategically ‘overreach’ into the fiscal sphere and ‘bargain’ with the government to achieve fiscal reforms. By shedding light on the interactions of global legal and economic developments and the micro-level strategies of state actors, the paper proposes a more complex view of the role of the state and brings state-actor strategies into our understanding of the grand shifts in economic policymaking.


Introduction
Austerity has shaped economic policy decisions in many advanced economies for more than four decades. Since the radical turn from Keynesianism to monetarism in the mid-1970s, a regime has prevailed which stresses balanced budgets, spending cuts and debt and deficit reductions. The persistence of these programmes since the late 1970s is striking given the incremental but continuous decline in real consumption, productivity and investment across developed economies and the growing levels of income and wealth inequality to which fiscal austerity contributes (Baccaro & Pontusson, 2016;Carlin & Soskice, 2009). Even during the Global Financial Crisis of 2007, the Eurozone crisis of 2009 and the unfolding COVID-19 crisis, many heads of government have been holding onto the general principles of fiscal restrictiveness, stressing that stimulus programmes have to be followed by consolidation.
Austerityan orthodox policy orientation which emphasises debt and deficit consolidation, limited interventions in the economy and a shift away from countercyclical fiscal policies (Armingeon, 2012, p. 343)has prompted considerable changes in the relation of the state with its citizens and the economy. States' revenue collection and government spending to GDP went through a rapid expansion in the post-war era, permeating almost all sections of society (Moene & Wallerstein, 2003). Austerity brought this development to a sudden halt. As the expansion of social transfers and services was scaled back (Pierson, 1998;Streeck, 2010) and infrastructure projects and public goods became more restricted (Ag enor & Yilmaz, 2017), inequality levels soared (Burkhauser et al., 2012) and democratic responsiveness deteriorated. While voters continuously supported spending programmes, this interest no longer found representation (Els€ asser et al., 2017). Austerity policies entail significant electoral and economic risks. First, many constituent groups disapprove of consolidation programmes. This is particularly a risk for social democratic parties whose constituencies are unevenly dependent on spending (Notermans, 2000;Scharpf, 1991). Second, balanced budget rules and fiscal consolidation can impair governments' ability to revive aggregate demand and economic growth (Carlin & Soskice, 2009;Stockhammer et al., 2019). The paper, therefore, asks why many governments performed this radical institutional shift towards austerity from the late 1970s.
While the existing scholarship focuses on the unitary state's response to external factors such as globalisation (Genschel, 2002;Lierse & Seelkopf, 2016) and ideas (Blyth, 2013b;Hay, 2016), I argue that we also have to account for the interaction of different actors within the state. Specifically, I hypothesise that central banks developed strategies to influence the policy outcomes of governments and used changes in the global legal and economic sphere to extend their influence within the domestic economic policy framework. I base my argument on insights from the critical juncture literature (Capoccia, 2015) as well as the international political economy (IPE), international relations (IR) and economic history literature which have already examined conflicts between central banks and fiscal actors and the role of institutional power of central banks for the outcomes of these conflicts (Goodman, 1992;Marsh, 1992Marsh, , 2009. To this, I add micro-level strategies, or specific strategic choices made by central bankers to extend their power position vis-a-vis the government. This micro-level analysis shows that institutional instruments are only one way how central bankers power within the larger macroeconomic framework. They also actively utilise opportunities in the political and economic context to bargain with the government. I introduce specific strategies used by central bank actors which have never been tested before and explore how they function on the ground. Tying together radical changes in the global economic and political sphere with micro-level strategies allows me to bring together insights from the international political economy (IPE) and the comparative political economy (CPE) literature.
The case study in this paper illustrates how changes in the global legal and economic sphere triggered a power struggle between two German state actors: the central bank and two Social Democratic administrations. This power struggle developed from the implementation of the DM float in 1973, which allocated new power resources to the Bundesbank. The central bank used these powers to strategically 'overreach' into the fiscal sphere and to 'bargain' with fiscal actors. Overreaching meant that the Bank implemented new monetary instruments, including monetary targets, to shape fiscal outcomes through signalling activities.
Bargaining entailed the use of institutional power and drawing on the political and economic context to bargain with the government over fiscal outcomes. In this second phase, the central bank raised interest rates for several months during a time in which the government depended on lower interest rates for domestic and foreign policy reasons. Then the central bank offered lower interest rates in exchange for the implementation of fiscal rules.
In order to test my argument, I am introducing a causal mechanism between the breakdown of the Bretton Woods system and fiscal austerity. For the first time in this research area, I elaborate on transparent process-tracing tests with assigned test strengths. Since there is a low prior confidence in the existence of the mechanismas the literature has not tested it yeta fruitful strategy is to test it on a most likely case. While this research design does not allow for the generalising of the mechanism across a larger population of cases, the test can be applied to a least likely case in the future (Beach & Pedersen, 2013, pp. 151-152). I conduct a test on Germany in the 1970s as it is a most likely case for power struggles won by the central bank. The independence of the Bundesbank during this time makes the power struggle and an ostensible win by the central bank likely. The case study relies on archival material collected from the archives of the Bundesbank and the German Cabinet between 1960 and 1981.
The paper proceeds as follows: The next section introduces fiscal austerity conceptually and presents the existing explanations in the literature. Section three suggests how this paper will add to the literature by investigating state actors' micro-level strategies. Section four lays out the methodological underpinnings of this paper and section five empirically traces the shift from expansive fiscal policies to austerity in the German case. The conclusion summarises the study's findings and draws lessons for political economy research.
State of the art: existing explanations of fiscal austerity I conceptualise austerity according to Pierson's (2002) notion of 'fiscal stress' -as a situation in which fiscal discretion is constrained through the simultaneous pressures of rising needs for spending and limited room for finance. Pierson focuses on the endogenous pressures of a post-industrial society in which stagnating productivity, ageing and unemployment lead to rising expenditure needs (Pierson, 1998(Pierson, , 2002; see also Mares, 2006). Structural accounts have added that the rising levels of sovereign debt resulting from these expenditure pressures simultaneously generated the threat of capital flight as financial markets constantly monitored the 'soundness' of fiscal decisions (Streeck, 2011, pp. 22-24;Streeck & Mertens, 2011). While this image of fiscal stress emphasises a rising trade-off between different types of government expenditure and the credibility of macroeconomic decisions in international markets, I highlight independent central banking as a source of fiscal stress.
Fiscal stress does not necessarily mean fiscal retrenchment. In fact, expenditures continued to rise in many OECD countries even at the height of austerity in the 1980s. However, the significant increases of post-war spending have slowed down. Table 1 displays the development in the growth rates of final consumption expenditure, social transfers and gross fixed capital formation in 16 OECD countries. 2 It shows that most OECD countries went through significant reductions in spending increases from the 1980s. In times of rising need for expenditures, this should cause significant fiscal stress for governments. Moreover, austerity affected the distribution of resources between different spending areas. Since welfare spending was particularly resilient against spending reductionsdue to its popularity and organised interests (Pierson, 2001(Pierson, , 2002, funds were increasingly shifted from discretionary public investment to welfare spending (Breunig & Busemeyer, 2012;Streeck & Mertens, 2011, pp. 36-37). Table 1 shows that the growth rate in gross fixed capital formationwhich entails infrastructure and housing expendituresdeclined at a much higher rate compared to the other two spending areas between 1980 and 2014. This has direct consequences for the supply-side of OECD economies. While continued welfare spending should have positive effects on the demand-side of economies (Stockhammer et al., 2019), reduced infrastructure spending has contributed to the secular stagnation of economies observed since the 1980s (Ag enor & Yilmaz, 2017).
The first literature strand to explain austerity is the globalisation literature. It stresses the role of capital mobility and international exchange rate speculation for fiscal decisions which were prompted by the rise of the Eurodollar markets. When the expansion of those markets peaked in the 1970s, governments increasingly  1974-1979 1980-1989 1990-2014 1974-1979 1980-1989 1990-2014 1974-1979 1980-1989 1990 Sources: Average social transfers growth rates are measured from social benefits other than social transfers in kind as percentage of GDP (Armingeon et al., 2017). General government gross fixed capital formation includes stock of fixed assets including residential and non-residential buildings, roads, railways, bridges, and airports and is at constant prices from AMECO database. General government final consumption expenditure growth is composed of government expenditure on non-market final goods and services at current prices from the World Bank database. I used data on 16 of 21 OECD countries selected on data availability. The table focuses on spending as opposed to budget balances or sovereign debt because of the different strategies available to countries to balance their budgets. 1 struggled with payments imbalances and exchange rate speculation which enforced policy change (Goodman, 1992;Goodman & Pauly, 1993). In order to signal sound policies to the markets, governments increasingly emphasised central bank independence, price stability and inflation control in the monetary realm (Schelkle, 2012, p. 35;Johnson et al., 2019, p. 548). Moreover, capital mobility forced governments to consolidate their budgets in the face of tax optimisation and rising bond yields on debt finance (Devereux et al., 2008;Lierse & Seelkopf, 2016). While this literature has convincingly shown that capital mobility contributed to radical changes in the fiscal and in the monetary sphere, the state is treated as a unitary actor. This contradicts recent findings in the central bank literature which demonstrate that structural pressures considerably differ in the two policy realms and that the two actors often clash over policy goals (Diessner & Lisi, 2020;Fern andez-Albertos, 2015). A second literature strand, the ideational literature, stresses that it is not economic developments alone which lead to policy change, ideas determine how policy actors view the economy and respond to crises (Blyth, 2002;Hay, 2016;Schmidt, 2010). The literature shows that in the 1980s, governments became increasingly convinced that inflation was the main crisis indicator in OECD economies and that fiscal consolidation and central bank independence were critical solutions (Blyth, 2013b;Blyth & Matthijs, 2017). Neoliberal and ordoliberal epistemic networks were critical for the dissemination of those ideas (Mirowski & Plehwe, 2009;Ptak, 2009;Slobodian, 2018). Bonefeld (2012) defines ordoliberalism as a German variant of liberalism, in which a strong state guarantees market-liberalism and helps to embed entrepreneurialism into society. Recent studies have shown how these notions contributed to austerity policies (Blyth, 2013a;Dellepiane-Avellaneda, 2015;Young, 2014). While this literature has made important contributions to our understanding of austerity, it has viewed ideational change as a unitary process spanning the entire state. Since we know from the central bank literature that ideas often considerably differ between central banks and governments (Bodea, 2013;Gray, 2007), an in-depth analysis of how policy approaches travel from one policy realm to another would enhance our understanding of the rise of austerity.
While the CPE literature delves deeper into the interaction of policy realms, it focuses on institutional continuity and, therefore, does not yet provide an explanation for the changes which contributed to austerity. Starting with the 'bringingthe-state-back-in' discussion in the 1980s, institutionalists explored the interests and the institutional culture in different institutions (Rueschemeyer & Evans, 1985;Skocpol, 1985). First, neo-Weberians examined the organisational logics and autonomous activities of state actors (Allen, 1989;Johnson, 1998). More recently, the varieties of capitalism (VoC) approach studied the complementarity of macroeconomic institutions (Carlin & Soskice, 2009;Soskice & Iversen, 2000). In the German case, institutional advantages arose out of the coordination of an inflationaverse central bank, wage-bargaining institutions and a non-discretionary fiscal policy (Franzese & Hall, 2000;Hall, 2018;Hall & Franzese, 1998). While these findings of the CPE literature help to understand general dynamics within the state, the conflicts of state actors in times of crisis have not been at the forefront of these approaches.
Finally, the IPE and economic history literatures have examined how a shift in the global sphere contributed to changing relations of fiscal and monetary actors.
While these findings are central to my understanding of the rise of austerity since the 1970s, a conceptualisation of how actors used change in the global sphere to strategically shape fiscal outcomes has not yet been developed. Putnam (1988) linked in his two-level games domestic politics and global developments. He finds that domestic actors can use pressures in the international realm to advance interests domestically. Others have investigated how states strategically used the liberalisation and financialisation of economies to generate positive economic outcomes (Goodman & Pauly, 1993;Helleiner, 1994). This literature also highlights the differing interests of state actors and how independent central banks use their institutional powers to shape the outcomes of conflicts with governments (Berger, 1997;Berger & de Haan, 1999;Goodman, 1991;Trampusch, 2015). In the German case, the Bundesbank held a sufficiently independent legal mandate that could maintain monetary stability even if this clashed with fiscal goals. This power was heightened after the collapse of the Bretton Woods system which freed the bank from the obligation to maintain dollar parity (Katzenstein, 1978;Marsh, 1992Marsh, , 2009). Thus, this conflict perspective is useful to analyse the interaction of global developments and domestic conflicts within the state. However, so far, the literature has not unpacked the specific acts, or micro-level strategies, through which actors deployed their power. This is the sphere where we can flesh out a more action-based approach of how actors use their power in specific political and economic circumstances.
In sum, the literature has highlighted four critical factors which contributed to austerity conceptualised as fiscal stress: economic pressures, ideational developments, coordination of institutions and independent central banking. I contribute to the final approach but ask what kind of specific acts central banks engaged in to influence fiscal actors.

Theorising critical junctures and within-state power struggles
In this section, I derive a mechanism from the critical juncture approach and the IPE, IR and economic history literature which helps to make sense of how the demise of the Bretton Woods system (X) is linked to austerity policies (Y). The mechanism elaborates change as a function of the strategic power struggles of different state actors (M). Capoccia (2015, p. 151) defines critical junctures as specific moments in institutional development in which an event or a series of events (typically exogenous to institutions) generates a phase of uncertainty among policy actors, which opens up an extended range of policy options after 'structural (economic, cultural, ideological and organisational) influences on political action are relaxed'. Critical junctures arise during unexpected events, from a crisis in the old institutional system or from the faltering of societal support for the old system (Capoccia, 2015). The demise of the Bretton Woods system is one such instance, in which a larger crisis of rising capital mobility, rising levels of inflation and payments imbalances shook up the existing institutional certainties and instigated radical institutional change (Blyth & Matthijs, 2017;Gray, 2007).
But what were the specific internal dynamics of radical institutional change? While the recent literature on the demise of the Bretton Woods system focuses on the conflicts of the state and private actors over whether exchange rates should float (Germann, 2014(Germann, , 2020Gray, 2007), once exchange rates floated, this institutional shift had an individual and additional impact on the power relations of state actors. Before the float, the enforcement capacity of central banks was curtailed because exchange rate parity mandated the pre-eminence of stable exchange rates over domestic monetary concerns. This impairment of central bank power was aggravated by rising capital mobility, which rendered interest rate increases ineffective: higher rates resulted in capital inflows, which brought interest rates down again (Goodman & Pauly, 1993). In the German case, an undervalued DM forced the Bundesbank to support the dollar over periods as long as two years, which arose in the later stages increasingly at the expense of price stability (Marsh, 2009, pp. 39-40). Conversely, the fall of the Bretton Woods system extended the scope for monetary tightening and freed the Bundesbank from the obligation to defend the parity of the DM (Johnson, 1998, p. 59;Marsh 2009). Thus, it must be expected that the float fundamentally shifted the power relations of fiscal and monetary actors in the German macroeconomic regime.
Changing power relations alone should not lead to radical institutional change, but they activate a simmering conflict between central banks and governments. The two-state actors generally share goals of macroeconomic stability, economic growth, appropriate levels of employment, credit and consumption, but prioritise these differently. Governments are deeply embedded in electoral politics. Irrespective of ideological orientation, successful parties depend to some extent on the support of workers, manufacturers and farmers who benefit from fiscal stimulation (Alesina & Tabellini, 2007;Goodman, 1991). Central banking, on the other hand, has increasingly become a technocratic policy field, insulated from the electoral arena (Lepers, 2018). Moreover, due to close ties with the inflation-averse financial community, central banks tend to support price stability (Blyth & Matthijs, 2017;Goodman, 1991). Accordingly, I expect that the latent conflict between the Bundesbank and the government was triggered by the breakdown of the Bretton Woods system.
Critical juncture accounts stress the role of contingency in moments of radical upheaval while acknowledging the limits to the range of possible choices posed by institutional power structures. Capoccia (2015, pp. 158-159) follows Berlin (1974, p. 176) in developing a context-based understanding of contingency, elaborated as 'what happened in the context of what could have happened'. This means that those actors who are in a powerful strategic position at the juncture (political leaders, bureaucrats, policymakers and judges) will be able to use their institutional power to steer outcomes (Capoccia & Kelemen, 2007;Thelen & Steinmo, 1992). Since the breakdown of the Bretton Woods system brought with it significant power-distributional consequencesthe turn from the Keynesian post-war institutions to monetarism instigated a shift from full employment to price stability-oriented policies (Blyth & Matthijs, 2017, p. 208;Goodhart, 2011) it has to be expected that officials sought to enhance their power position within the larger macroeconomic framework. And since the Bundesbank had just acquired additional institutional powers, I expect that the central bank steered the crisis outcome in a direction favourable to its own interests.
Specific strategies which allow central banks to shape fiscal decisions can be derived from inductive studies into central bank behaviour. Here authors have theorised that central banks will likely try to 'overreach' into the fiscal sphere by setting policy signals through monetary policy choices (Bodea, 2013;Diessner & Lisi, 2020). Since the policy spheres of monetary and fiscal policy are not neatly separable and often affect each other through spillovers (Fern andez-Albertos, 2015, pp. 226-227), conflicts arise as soon as the goals of the two actors differ (Way, 2000). These conflicts are particularly likely with an independent central bank that tries to curb the partisan motivations of expansionary monetary and fiscal actors (Lohmann, 2008). Central banks have attempted to solve this spillover problem by signalling monetary policy goals to fiscal actors. In the German case, the central bank introduced new monetary tools to set guidelines for fiscal outcomes (Berger, 1997;Bodea, 2013). I, therefore, expect that the Bundesbank selectively introduced new monetary instruments to strategically reach into the fiscal sphere.
A second strategy elaborated by this literature is the act of 'bargaining'. Elgie and Thompson (1998, p. 33) find bargaining occurs especially in times of conflict and crisis. It often begins with a phase in which each side informs the other about its policy goals, its plans and the means available to achieve these goals. Next, actors contemplate their strategic options depending on the means and policy options available to the other side. During the bargaining process proper, actors use their institutional power and the political and economic circumstances to influence each other's decisions (Elgie & Thompson, 1998, pp. 33-34). Different from 'overreaching', 'bargaining' entails the direct interaction of central bank officials with fiscal actors in cabinet or council meetings, where they flex their muscles by using the interest rate as a lever to get their desired reforms in a foreign policy sphere. Due to its level of independence, the Bundesbank is a prime example of an actor which can bargain with the government over the wider direction of fiscal outcomes (Marsh, 1992;Scharpf, 1991, p. 5). Thus, I expect that the Bundesbank will have used its enforcement capacity and the political and economic conditions to bargain for fiscal reform.
Complementing our larger understanding of critical junctures with specific micro-level strategies of state actors allows us to portray a more complex reality of radical institutional change. Instead of only viewing change as the function of uncertainty and existing institutional power (Capoccia, 2015), it allows us to examine how state officials actively use transitions in the global sphere to power. Instead of viewing state responses to internal and external pressures as coherent, it can shed light on how policy initiatives spread from one sphere of the state to another even if state actors disagree initially. The ability to analyse these strategies is particularly important for cases in which policies are transferred from a democratically accountable to a less democratically accountable policy sphereas was the case in the shift from fiscal to monetary responsibility for macroeconomic stability (Marsh, 1992;Notermans, 2000;Scharpf, 1991). Moreover, it allows us to make sense of how global shifts in economic and legal conditions have translated into policy change on the ground. We can, therefore, tie together the insights of two-level games (Putnam, 1988) and approaches which highlight the strategic interaction of state actors at the micro-level (Streeck & Thelen, 2005).
Two critical developments since the 1990s likely reinforced the pressures of central bank independence on fiscal outcomes. First, financial market integration has been nurtured by central banks as financialisation enhanced the central bank's ability to implement effective monetary policy (Braun, 2018;Walter & Wansleben, 2019). Since central banking became more effective but also heavily dependent on financialised markets, it is likely that this development further accelerated the pressure on governments to consolidate budgets. Moreover, the establishment of the European Central Bank (ECB) also likely made the pressures of central bank independence on fiscal policy more stringent. Even in the German case, a lowinflation country, the high interest rates set by the inflation-averse ECB have led to further stagnation in the economy (Baccaro & Pontusson, 2019).
However, more recent developments may point to an easing off of these pressures. While the literature on austerity has so far not elaborated on an endpoint to austeritysee Pierson's (2002) argument of permanent austerity -spending increases eased up in many countries in the 1990-2014 period (see Table 1). It is likely that the changes in the relationship of fiscal and monetary policy since the Global Financial Crisis of 2007 have contributed to this change. While the ECB's monetary innovation developed slowly and was accompanied by strict rules of fiscal discipline (Henning, 2017), other central banks rigorously shifted from rule-based policies to quantitative easing and forward guidance. Since quantitative easing entails explicit redistributive elements, including targeted funds for certain sectors explicitly denounced in the early years of independent central banking, it has likely allowed for more expansiveness (Fern andez-Albertos, 2015).
I introduce a four-step process with specific process-tracing expectations (PTE): PTE 1: The central bank gains enforcement capacity through the breakdown of the Bretton Woods system PTE 2: Due to step 1, the central bank officials 'overreach' into the fiscal sphere by strategically selecting new policy settings and instruments and signalling policy goals to the government PTE 3: Due to the limited effect on the fiscal realm in step 2, the central bank deploys its institutional power as well as the dependence of the government on lower interest rates to 'bargain' for fiscal reform PTE 4: Following steps 2 and 3, the government implements fiscal rules and an effective fiscal council whereby the central bank's interests become ingrained ( Figure 1).
From the central bank literature, I derive two scope conditions that trigger the causal mechanism. First, higher levels of central bank independence increase the probability of the power struggle occurring because independence ensures baseline powers to freely implement settings and instruments in the monetary realm (Goodman, 1992, pp. 2-5). Second, the partisan position of the government should affect the response of the central bank. The more left-leaning the government, the more likely are conflicts over price stability (Way, 2000).

Methods and operationalisation
I test the expectations of my causal mechanism using theory-testing process-tracing. Unlike approaches which focus on correlations, this method assesses whether a hypothesised causal mechanism can be found between X and Y (Beach & Pedersen, 2013;Collier, 2011). It is often combined with comparative case study methods in a least likely case design, which allows scholars to generalise beyond their case. However, this is a high-risk strategy when there is low prior confidence about the existence of a mechanism. Thus, a low-risk strategy is to test the mechanism on a most likely case (Beach & Pedersen, 2013, pp. 151-153). I am testing the mechanism on the most likely case of Germany. Power struggles between the central bank and government are 'most likely' here because the Bundesbank was one of the first and most independent central banks in the world, with a mandate that was particularly inflation averse. It, therefore, frequently clashed with the government over the direction of fiscal and monetary policy (Berger, 1997;Marsh, 1992Marsh, , 2009. I develop transparent process-tracing tests in two steps. First, following Beach and Pedersen's (2013, p. 56) suggestion to deduce the mechanism through a review of the existing empirical and theoretical work, I have developed (in the previous section) a within-state power-struggle mechanism, which allows me to flesh out the logical steps between X and Y. In a second step, I have formulated case-specific predictions about each step of the mechanism, followed by structured empirical tests (see Appendix A). For the structured empirical tests, I assigned values of uniqueness and certainty to each process-tracing evidence (PTE) and assessed the strength of the test (see Meissner, 2019 for an example). PTE 1 to 4 were assessed through smoking gun and hoop tests. In a smoking gun test, finding the predicted evidence updates our confidence in the validity of the hypothesis, while not finding the evidence does little to increase our confidence (Beach & Pedersen, 2013, pp. 103-104). Hoop tests, on the other hand, set the necessary criteria for a theorised causal mechanism. If the evidence does not confirm the test, the theoretical expectation is disconfirmed. If it is confirmed, it substantially supports the expectation (Beach & Pedersen, 2013, pp. 102-103;Collier, 2011, p. 826). In sum, if I do not find the expectations confirmed, I can be confident that the mechanism is not present in the case. If I do find them confirmed, the confidence in the PTE is increased.
Since process tracing is a method that involves 'the examination of "diagnostic" pieces of evidence within a case that contribute to the supporting or overturning of alternative explanatory hypotheses' (Bennett, 2010, p. 208), we also have to test alternative explanations to the mechanism (Bennett & George, 2005 chapter 9). For this purpose, I present a process-tracing table (in Appendix B) which introduces Figure 1. The within-state power struggle mechanism and its scope conditions. the hypotheses, expectations, and predicted observable implications or evidence used to test the explanatory power of alternative theories.
To ensure the accuracy of evidence used in process tracing, I collect observations from several independent sources and triangulate these (see Beach & Pedersen, 2013, p. 128). First, documents gathered from the Historical Archive of the German Bundesbankcitations from this archive are designated with the signature 'HA BBk'comprise the minutes of the Central Bank Council and the executive board of the Bundesbank, which were written for documentation and internal communication purposes. The executive board consists of the President, the Vice President and four other members. The Council consists of the executive board plus presidents of the state central banks. In both bodies, the government (often the Chancellor or the Finance Minister) can attend the meetings but is not always present. The second independent source is documents collected from the cabinet meetings of the government from the Bundesarchiv in Koblenz, 3 which are designated with the signature 'BArch'. Attending were the cabinet members and, at times, representatives of the Bundesbank (often the President, the Vice President or the Chief Economist). The types of documents used from the archive were the summaries of the cabinet meetings on economic and fiscal policy issues. These documents are also used for internal use by officials and the public after the retention period has passed. In sum, the documents analysed were written and published by two different institutional actors and each provides particular viewpoints of the respective actors.
The Bundesbank minutes are verbatim protocols while the cabinet documents are summaries of the discussions. While this limits my ability to establish full triangulation, I used two strategies to enhance the data basis. First, I read all relevant documents from the Bundesarchiv and selected those to the total population of analysed documents in which the interpretations, concerns, and requests of the two relevant institutional actors were clearly identifiable. Then, I compared the findings with historical research reports produced at the time in think-tanks and government-related research institutes. This also allowed me to fill blind spots where data was lacking from the cabinet. I collected all documents in a MaxQDA file and coded the relevant sections before evaluating the relevant factors. The total population of protocols is 159each spanning between 1 and 10 pageswhich adds up to roughly 600 pages of text. Finally, I translated all statements from German as closely as possible to the original wording and syntax while adjusting direct translation where it was necessary to enhance understanding.

Tracing the within-state power-struggles after the breakdown of the bretton woods system
It is helpful to first understand the legal constitution of the Bundesbank and the government before examining the power struggle. Both state actors were embedded in a larger institutional framework of the 'semi-sovereign state' (Katzenstein, 1978). Intending to devolve the power of the German state, the Allied forces had strengthened the power of state-level fiscal and monetary actors after the war. The Bank der deutschen L€ ander, established in 1948, combined a decentralised US-type federal reserve system with a centralised system similar to the Bank of England, in which Land presidents and a centralised directorate worked together (Marsh, 2009, p. 35). State central banks appointed the council, held all voting rights, earned profits and were accountable only to the Allied High Banking Commission (ABC) (Beyer et al., 2008). Despite several attempts to expand government influence over the bank, 4 state central banks refused to give up their position and the federal government never received significant veto powers (Berger, 1997, p. 435). 5 Indeed, article 12 of the Bundesbank Law, which established the Bundesbank in 1957, ruled out any intervention of the parliament or the cabinet in monetary decisions. Moreover, section 3 of the law recognised the pre-eminence of safeguarding the currency over supporting the economic policy of the government, authorising the Bundesbank to solely focus on price stability (Lohmann, 1994, p. 402;Trehan, 1988).
Despite this unusual level of independence of the Bundesbank, central bank power was limited during the 1970s for two reasons. First, the Bretton Woods system overruled the price stability mandate whenever the US-dollar parity of the DM had to be restored (Johnson, 1998, p. 59). Another factor which curtailed the central bank's power was powerful fiscal actors. Despite the weak position of the German federal government to directly influence monetary decisions, local fiscal actors held considerable powers. Within the larger framework of cooperative federalism, the federal state determined the overall direction of policies, but states decided how to administer or implement them (Prasad, 2006, p. 198). According to Article 109, section 1, of the Basic Law, states and municipalities held a 'selfgovernment guarantee' which allowed the independent administering of their budgets (Geske, 1983, p. 85). As a result, fiscal decisions were made in eleven state parliaments and 8,000 municipalities (Shonfield, 1965, p. 268;Zunker, 1972, p. 84). When Kiesinger's Grand Coalition implemented Keynesian fiscal instruments in Stability and Growth Act in 1967, this local fiscal power generated unusual expansionary pressures. The 'concerted action' (Konzertierte Aktion) added to this pressure. This forum in which the government, the central bank and unions and employers met to coordinate policies first helped to revive the economy through wage moderation and investment programmes. However, since 1969, wildcat strikes and wage increases increasingly contributed to excess liquidity (Mares, 2006, p. 143).
The globalisation literature argues that capital mobility, and capital flight in particular, are critical drivers of austerity. However, during the 1970s, capital outflows were not a problem in the German economy. FDI was not very mobile at that point in time -foreign-direct investment outflows averaged 0.5 percent of GDP between 1970and 1980(World Bank, 2021)while portfolio capital was characterised by excessive inflows into German markets. Moreover, these capital inflows principally caused problems for monetary actors, rather than fiscal actorstriggering conflicts between Bundesbank and administration. The Bundesbank had already struggled with trade imbalances since the transition to convertibility in 1958 leading to clashes with the government over the choice of rebalancing policies (Gray, 2007;Johnson, 1998, pp. 71-73), these conflicts became heightened in the 1970s, as the Bundesbank increasingly lost its power to stabilise price levels domestically (Scharpf, 1991, Chapter 11). The Mundell-Fleming trilemma predicts that states can at most achieve two out of three macroeconomic conditions: capital mobility, monetary autonomy, or fixed exchange rates. When actors have to secure parity with another currency in an exchange rates regime (as was the case in the Bretton Woods system until 1974), monetary policy autonomy can only prevail as long as capital is immobile (Goodman & Pauly, 1993). Thus, in the 1970s, when the expansion of the Eurodollar markets peaked (Scharpf, 1991, p. 298), the Bundesbank's autonomy became increasingly curtailed as so-called 'hot money' flooded the German capital market. The general upward pressure on liquidity was aggravated by the fact that to restore parity in the Bretton-Woods System, the Bank had to buy foreign currency when the DM reached the intervention point (HA BBK, 1971, B330 6160/2;Kitterer, 1999, p. 170).
The historical order of events demonstrates that the enforcement capacity of the central bank increased with the breakdown of the Bretton Woods system (PTE 1). Once the float was introduced, the Bundesbank implemented a new set of monetary instruments. After March 19, 1973, when the government decided to dismiss the dollar-parity of the Bretton Woods system, the German monetary regime underwent a radical transition from Keynesian full-employment goals towards a pragmatic monetarist approach (Beyer et al., 2008, p. 11). This shift was accompanied by institutional changes: the Bundesbank decided to let go of administrative and regulatory approaches to monetary policy and increasingly relied on openmarket operations. Pragmatically applying theorems from the 'monetarist counterrevolution' adherents, who argued that central banks should not engage in a fine-tuning of the economy but follow monetary targeting, it additionally implemented a quantitative target for money growth as a nominal anchor for inflation and expectations of price developments (Fratianni & von Hagen, 2001).
Another piece of evidence for PTE 1 is the statements of central bankers, which demonstrate the perception that the enforcement capacity of the central bank increased. Only after the float of the DM was implemented were central bankers free of the exchange rate obligation and held institutional capacities to enforce a more restrictive course of monetary policy (Johnson, 1998, p. 85). As one former central banker remembered, this 'gave the Bundesbank new scope for the control of domestic monetary conditions' and allowed actors to focus on domestic price stability (Issing, 2005, p. 330). The Bundesbank used its new freedom by implementing a restrictive course immediately following the implementation of the float (BArch, 02.02.1973 special session;HABBk, 15.02.1973, B330/6703/1). The council argued that decisive monetary tightness was now critical to dry out all channels of liquidity and started to raise the discount rates. As a result, the spread of interbank rates soared from 13% to 33% between March andAugust 1973 (HABBk, 12.04.1973, B330/6705/1;von Hagen, 1999, p. 687).
However, while implementing these price-stability instruments, central bank officials were not generally convinced of monetarist ideas. Instead, during the debates over solutions of the crisiscapital controls or the float of the DMin January 1973, the central bank reviewed two rival approaches to resolve inflation: a 'dirigiste' approach and monetarism. The dirigiste proposal envisioned greater central bank influence over selective credit controls, mandatory credit ceilings and a minimum reserve to actively steer credit availability in the economy (von Hagen, 1999, p. 686). Since the proposal found considerable support among a large group of central bankers, including the president of the Bundesbank, Karl Klasen (von Hagen, 1999, p. 687), it seems unlikely that ideational changes drove the shift to new policy instruments in the central bank at this point in time.
Instead, monetarist policy settings and instruments were suddenly endorsed on the basis of strategic considerations on how to enhance the power position of the central bank in the overall macroeconomic system (PTE 2). Evidence for the strategic policy decision can be seen in the deliberations of central bank actors behind closed doors, where both approaches were considered but after a few weeks, the preference tilted towards monetarism. In those weeks, a rising number of council members followed chief economist Schlesinger in his concerns that 'the Bundesbank's power position would be diminished' under dirigisme because the government could use credit ceilings to implement quotas or enforce the supply of central bank money or credit to the federal government. These types of obligations 'weakened the power position of the central bank' (HA BBk, 18.01.1973 As part of an effort to expand influence in the overall macroeconomic system, the central bank used its implementation rights in 1974 and 1975 to 'overreach' into the fiscal sphere and to signal the direction of monetary policy (PTE 2). That the target became explicitly defined as a weighted currency in circulation, along with the reserve requirements of bank liabilities so that fiscal policymaking was directly affected, indicates that the central bank used its institutional capacities to influence fiscal decisions. The reduction of the target was supposed to instantly enforce a reduction of the balances public authorities held with the central bank (Johnson, 1998, p. 86). Moreover, statements made by members of the council demonstrate that actors viewed the money supply target of central bank money (CBM) as a security for future monetary policy independence (PTE 2). They argued that the specific composition of CBM ensured that trade unions and fiscal actors would no longer 'underestimate the Bundesbank's fight against inflation and expand spending to counter unemployment' (BArch B330/7499/1; von Hagen, 1999, p. 690). The target was supposed to give an 'orientation to the economy as a whole', but to 'the social partners and the state in particular' (HA BBk, 1974, B330/ 7499/1). It 'clarified' the direction of monetary policy and 'disciplined' where monetary policy so far had been a 'toothless' measure, 'not taken seriously' (HA BBk, 1975, B330/7894;HA BBk, 1975, B330/7895). Even though the administration rejected the CBM, it had no legal lever to intrude on the implementation rights of the central bank as ministers argued in meetings with the council (HA BBk, 1974, B330/7499/2;HA BBk, 1975, B330/7895;Johnson, 1998, p. 94). The fact that the council designed the CBM in such a way that it would directly affect fiscal decisions indicates the strategic considerations of central bank officials vis-a-vis fiscal actors.
However, institutional power of the central bank was not sufficient to shape fiscal outcomes. Despite having implemented new monetary instruments to signal monetary goals to the government, the latter still prioritised spending increases in many areasin August 1975, for instance, it decided to increase unemployment benefits making it difficult to keep to low deficit levels (BArch, 1975, 106 th cabinet meeting). A second strategy had to be applied to achieve fiscal change. In this step, the central bank strategically used the dependence of the government on lower interest rates to bargain over fiscal solutions (PTE 3). Instances of bargaining predominantly take place in the context of conflicts (Elgie & Thompson, 1998, pp. 33-34), which, in the German case, materialised either in economic crises or after international agreements for stimulus packages. These were the economic and political contexts in which the government was particularly vulnerable to the interestrate weapon because the goals of the two actors diverged. An early instance of such acts of bargaining can be observed in 1973 during the first oil crisis. In its first step, the central bank informed the government about its goals and plans. It intended to bring inflation down and would raise interest rates soon. Then it argued that interest rates could be lowered 'if the net borrowing of the municipalities was kept under the 1972 level. Their credit taking could be reduced through the introduction of the debt lid' (HA BBk, 15.02. 1973, B330/6703/1). Since spending increased further during this crisis, the central bank raised the discount rate gradually from 2% to 7% between October 1972 and June 1973, and the economy slid into a deep recession until fiscal reforms were implemented (HA BBk, 06.10.1972, B330/6172/2;HA BBk, 30.05.1973, B330/6706/2).
In 1979, during the second oil crisis, more bargaining took place (PTE 3). Again, the Bundesbank divulged its goals. The governors urged the government to consolidate the federal budget as rising sovereign debt threatened its capacity to stabilise the current account. Moreover, the council argued that the position of the DM as a reserve and investment currency was under threat if capital continued to leave at the current rate (BArch, 1979, 126 th cabinet meeting;BArch, 1980, 5 th cabinet meeting). At the time, the DM held the status of a 'stability leader' and foreigners held 'very high amounts of financial wealth' in DM. This status of reserve currency had to be safeguarded by 'sizable increases in exports' (BArch, 1981, 42th cabinet meeting), achieved by a fiscal policy which enhanced 'prices, costs and competitiveness' (HA BBk, 1981, B300/11169).
That the Bundesbank used the dependence of the Schmidt administration on lower interest rates during the process of bargaining shows how state actors strategically use political and economic conditions for bargaining outcomes (PTE 3). Since Schmidt tried to achieve a simultaneous harmonisation of interest rates across the G7 countries and a domestic fiscal stimulus in 1979 and 1980 (further elaborated on in PTE 4), the administration was particularly vulnerable to higher interest rates. In a cabinet meeting in 1980, Schmidt laid out his plan to harmonise interest rates internationally and urged the Bundesbank to use 'all available measures to keep international economic relations intact' (BArch 1980, 5 th cabinet meeting). However, the Bundesbank rebutted that it was no longer in control of domestic interest rates as the upward pressures of international interest rates spilled over into the German economy. It was, therefore, of utmost importance that fiscal measures enhanced investor trust in the DM to limit capital flight (BArch, 1979, 126 th cabinet meeting;BArch, 1980, 5 th cabinet meeting).
The fact that Schmidt planned to implement yet another stimulus package, despite having experienced that the central bank punished these with interest-rate hikes since the mid-1970s, demonstrates the limits of the coordination argument. Instead of learning to adjust fiscal behaviour to monetary requirements, Schmidt tried to convince the council members that this crisis had significantly different features to the first oil crisis and needed to be resolved through the modernisation of the economy and the 'maintenance of the social consensus', which required lower interest rates (BArch 1980, 5 th cabinet meeting). The evidence surrounding this episode also speaks against the globalisation argument. Even though Germany experienced one of the rare current account deficits in 1980, Chancellor Schmidt was not at all worried about the economic consequences. He argued, in a meeting with Bundesbank representatives that a deficit would 'resolve itself' since Germany had been 'struggling with the effects of current account surpluses in the past' and had accumulated high levels of exchange reserves (BArch, 136 th cabinet meeting).
Evidence for the hypothesised process tracing segment (PTE 4) would be particularly strong if it showed that the government did not initially intend to implement austerity policies and changed its mind after the bargain with the Bundesbank. What speaks against this expectation is that the newly elected Chancellor, Helmut Schmidt, and his social-liberal governing coalition turned away from the countercyclical approach. This became particularly evident during the first G5 meetings organised by Schmidt and his French counterpart, Val ery Giscard d 'Estaing, in 1975 and1976, where Schmidt supported anti-inflation policies to counter exchange rate fluctuations (James, 1996, pp. 235-238;Romero, 2014, p. 117).
While this general shift implies that changing ideas contributed to fiscal policy change, the reform discussions in 1979 point against ideas as a sufficient explanation for the rise of austerity in the German case. In 1979, the administration supported a large fiscal stimulus package due to foreign policy and domestic economic goals. Only after bargaining with the Bundesbank, the Chancellor turned to austerity. Thus, this episode provides evidence for PTE 4. Since growth rates further decelerated in 1976, the newly elected American president, Jimmy Carter, proposed at the third G5 (now G7) meeting an interventionist 'locomotive' strategy. The locomotive suggested that a coordinated fiscal stimulus by the US, Japan, and Germany could revive growth in deficit economies. Concerned about the effects on domestic inflation, Schmidt first objected, but, in 1978, when the US had become dangerously exposed to international markets through its unilateral stimulation programme, agreed to implement additional fiscal stimuli. Another impetus for a stimulus came from the second oil crisis in 1979 which pushed the German economy into a deep recession and domestic demands for a more expansive solution emerged (Putnam, 1988, p. 428). The combined effects of the second oil price hike and FED chairman Paul Volcker's decision to curb expansionary monetary policy drove West German interest rates to new heights and the current account into a significant deficit position. The DM devalued vis-a-vis the dollar and capital flight reversed the previous concerns of over-liquidity (Goodman & Pauly, 1993, p. 63). In this situation, the SPD and its market-liberal coalition partner, the Free Democrats (FDP), returned to their traditional support of fiscal stimuli. They found that expansionary fiscal measures were the most effective instruments for reviving economic growth and to prevent an international conflict. They thus proposed a DM 6 billion loan-financed business cycle programme (BArch, 1979, cabinet meeting 126).
Despite the ambitions to implement another stimulus package, Schmidt had to scrap this initiative and, instead, implemented a strict debt lid to facilitate the enforcement of fiscal reforms. This change of mind followed bargaining with the Bundesbank. Between 1976 and 1981, the Bundesbank lowered the allowed inflation rate from between 4% and 5% to only 3.5%. The gradual tightening policies resulted in soaring interest rates in the following years (Trehan, 1988, p. 38). In response, the government set up a debt lid, which determined that spending and borrowing had to suit the current business cycle and external economic requirements (BArch, 1980, p. 7. Cabinet meeting;Geske, 1983, pp. 98-100), and reduced the planned spending levels for 1982 by DM 16 billion (BArch, 1981, p. 42. cabinet meeting). Bundesbank President Karl Otto P€ ohl argued in a cabinet meeting that he supported the decision: a lower level of net credit taking was necessary to 'support the capital market and facilitate the exchange rate discussion' (BArch, 1981, p. 42. cabinet meeting). A good piece of evidence for PTE 4the effect of bargaining on the fiscal decisions of the governmentis Chancellor Schmidt's statement in this meeting, which demonstrates that the government implemented the debt lid because of the interest rate pressures, and the prospect of the Bundesbank releasing those under lower spending, during moments of high vulnerability: Presenting his spending cuts to the cabinet, he highlighted 'the central importance of the development of interest rates' (BArch, 1981, p. 42. cabinet meeting).
In sum, the central bank deployed changes in the global legal and economic sphere strategically to achieve its policy outcomes in a foreign policy sphere. First, it used its implementation rights to 'overreach' into the fiscal sphere and signal monetary policy goals. Then, the Bundesbank 'bargained' with the government over fiscal reforms, using its institutional power and political and economic conditions to increase pressure on the government.

Conclusion
This paper asked the question as to why governments implemented austerity policies that radically reversed the expansionary nature of government spending in the post-war era. It explored this question by examining the rise of austerity in the 1970s in Germany using theory-testing process-tracing with transparent tests. In this paper, I developed the theoretical argument that central banks strategically used institutional power resources and opportunities in the economic and political context to shape fiscal outcomes.
The empirical findings support this argument: the breakdown of the Bretton Woods system enhanced the enforcement capacity of the Bundesbank. Becoming aware of this rise in power, central bank officials strategically considered new monetary instruments, depending on their ability to shape fiscal outcomes. They first used newly won implementation rights to 'overreach' into the fiscal sphere. The newly implemented monetary target signalled monetary policy goals for fiscal actors. When this did not yield the desired results, central bank actors turned towards acts of 'bargaining', in which they used their newly acquired institutional powers (interest rates) as well as political and economic conditions to raise the pressure on the government. The government, dependent on lower interest rates, implemented the reforms desired by the central bank.
This paper shows that an in-depth understanding of micro-level strategies can provide new insights into the dynamics of radical fiscal changes in the German economy. I show that the central bank's power went beyond institutional resources new settings and instruments of monetary policy. Actors also actively utilised opportunities in the economic and political context to nudge the government to conduct restrictive fiscal decisions. In moments in which the government was particularly vulnerable to interest rate increases, central bank actors successfully bargained for the implementation of fiscal rules.
As the empirical case study presented here is limited to developments in the German case, this article is designed as a starting point for further exploration of the strategic interaction of central banks and governments within the global legal and economic conditions. Since we know that the decline of the Bretton Woods system structurally distributed enforcement capacity towards central banks (Goodman, 1992, p. 59;Goodman & Pauly, 1993), it is possible that independent central banks, and those that were made independent in the 1980s and 1990s, used similar strategies to the ones observed in Germany. Future research would have to test whether other central banks used the technique of overreachif they possess freedoms of implementationand bargainingif they can weigh their own powers against those of the government depending on the political and economic context.
If it is true that changes in the global legal and economic sphere interact with state actor strategies, the analysis of power-struggles of fiscal and monetary actors needs a more prominent place in political economy research. Recently, the growthmodel literature has taken up the challenge of analysing economic developments from a demand-side view of the economy while also disentangling the interactions of several different private actors which correspond with developments in the growth model. Hall (2020), for instance, examines how electoral cleavages and growth models interact. In his understanding, liberalisation in the 1980s emerged from the interaction of a secular crisis of the previous model and a split in the working class between skilled and semi-skilled workers. Baccaro and Pontusson (2019) add that private actor coalitions can be forged through a dominant discourse of the national interest. In the German case, an undervaluation regime had kept interest rates at high levels since post-war times and suppressed the interest rate sensitive construction sector, housing, and mortgage finance, while distributing resources to the export sector. Ideology was critical for keeping this model intact. These approaches have made important contributions to a more complex understanding of the demand-side by disentangling the interests and ideologies of different actors in the economy. This complexity could be enhanced further if a more comprehensive understanding of state-actor strategies and the interaction of monetary and fiscal actors were included. Notes 1. Scandinavian economies have structural advantages to acquire balance their budgets as they collect higher revenue for reasons of size and revenue-effective tax mixes (Martin, 2014). The three spending categories in Table 1 are the most important government expenses for economic growth and political relations between the state and its citizens (Cusack & Fuchs, 2002). 2. I have organised these countries into three welfare state groups to show that changes applied to all types of states. There were only three exceptions to a continuous fall in spending -British final consumption expenditure and Norwegian and French social expenditurewhose growth rates increased over the entire period of observation.
3. These are the cabinet documents which the Bundesarchiv presents in 'the online edition' which can be found here: https://www.bundesarchiv.de/cocoon/barch/0000/ index.html. 4. To be sure, the Bundesbank's independence was only established by secondary law.
However, several studies show that reform initiatives to make the central bank more dependent were never seriously pursued. One critical factor was the Bank's popularity among the public. The Bundesbank forged close ties with constituents who believed that the post-war miracle had been built on a low rate of inflation (Berger, 1997;Goodman, 1991, pp. 58-59;Kreile, 1977). Whenever governments tried to intrude in the Bank's business, it used the press to blow the whistle and officials retreated from interference (Beyer et al., 2008). 5. The government's sole veto power was to postpone interest rate changes for 8-14 days (Berger, 1997, p. 435).