This Time Wasn ’ t Different: Responsiveness and Responsibility in the Eurozone between 2007 and 2019

This article investigates the balance between responsiveness and responsibility in France, Germany and Spain between 2017 and 2019. Based on the party-politics literature, the article argues that during crises responsibility may temporarily increase, but that subsequently governments need to base their actions on responsive policy criteria in order to maintain citizens ’ trust. Yet, given the dominance of austerity ideas during this decade and the introduction of a more stringent European ﬁ scal framework, it is questionable whether responsiveness increased after the Eurozone crisis. Analysing original data from the yearly budget speeches of ministers of ﬁ nance, the article ﬁ nds that responsibility dominated budgetary policy-making in the early 2010s and that since 2014 responsiveness has increased, regardless of a country ’ s ﬁ nancial situation. Furthermore, during the crisis responsibility tended to be higher under centre-right cabinets. The article portrays the early 2010s as an episode of ‘ politics in hard times ’ during which governments chose their policy response based on their ideas.


Introduction
During the 2010s fiscal discipline shaped large parts of Europe's political economy (Blyth, 2013b;Matthijs, 2016;Seikel, 2016;Bremer and McDaniel, 2020;Schmidt, 2020), inspiring not only the response to the Eurozone crisis but also the reform of European economic governance (EEG) (Laffan, 2014;Wasserfallen et al., 2019). This dominance of the fiscal discipline doctrine has generated various reflections about the possible implications for national democracy (Crum, 2013;Maatsch and Cooper, 2017;Schmidt, 2020) and in particular for the balance between democratic responsiveness and institutional responsibilities (Mair, 2013(Mair, , 2014Bardi et al., 2014;Rose, 2014;Lefkofridi and Nezi, 2020). Following these reflections, the 'primacy of responsibility' witnessed in the early 2010s (Laffan, 2014) is sometimes not seen as a temporary crisis response, but rather as a structural by-product of European integration that reduces the room for responsiveness towards domestic democratic preferences (Alonso and Ruiz-Rufino, 2018;Ruiz-Rufino and Alonso, 2017;Schäfer and Streeck, 2013).
Compared with previous instances of 'politics in hard times' (Gourevitch, 1986), the homogeneity of the policies pursued by the various member states in the early 2010s seemed to make the Eurozone crisis a rather unique episode. While in the past international economic and financial crises generated different responses in different countries, this time, disregarding their political composition, governments decided to consolidate their budgets, even at the cost of neglecting demands for social protection by the growing numbers of unemployed citizens (Pontusson and Raess, 2012;Schäfer and Streeck, 2013). In public discourse, this policy response was regularly legitimized with arguments about responsibility, intended as compliance with (European) norms and procedures (Alonso and Ruiz-Rufino, 2018;Hutter and Kriesi, 2019;Karremans and Lefkofridi, 2020). The austerity-based response to the Eurozone crisis was in fact in many ways based on strengthening pre-existing European budgetary agreements (Verdun, 2015). The fiscal discipline-based response to the Eurozone crisis may thus not be the result of politics in hard times but may rather be embedded in the institutional commitments that governments have under the conditions of deepened integration (Scharpf, 1998(Scharpf, , 2015Rodrik, 2000).
In order to investigate this proposition, this article presents an analysis of original data from budget speeches delivered by finance ministers to their national parliaments in France, Germany and Spain between 2007 and 2019. The article starts with a discussion in sections 2 and 3 of how the legitimacy of representative democracy relies on a balance between responsiveness and responsibility, and how this balance may have been affected during and after the Eurozone crisis. Subsequently, section 4 presents the scheme for coding responsiveness and responsibility in budgetary discourse. Section 5 presents the research strategy, illustrating the operationalization of the explanatory and control variables, namely time, institutional constraints and government partisanship. Section 6 presents the empirical findings, showing how the balance between responsiveness and responsibility evolved over time, taking into account institutional budgetary constraints, partisan composition of government and country differences.
The findings disconfirm the idea that fiscal responsibility structurally overshadows responsiveness as main criterion for decision-making, and instead portray the Eurozone crisis as an episode of politics in hard times during which fiscal responsibility temporarily became the dominant policy paradigm. From 2014 onwards in all three cases taxation and spending policies were increasingly driven by governments' social and market-oriented preferences, disregarding the country's public deficit levels. Furthermore, the findings reveal an interaction effect between the crisis period and centre-right government partisanship, suggesting that the centre-right composition of various member state governments in the early 2010s may have triggered additional levels of fiscal responsibility. In the conclusion the article reflects on the broader implications of these findings.

I. The Relevance of the Balance between Responsiveness and Responsibility
Concerns about the balance between responsiveness and governmental responsibilities (Mair, 2013) originate from a well-established theory of party politics, according to which parties fulfil the unique function of combining political representation with the task of running the state apparatus (Sartori, 2005;Bardi et al., 2014). This unique combination is a distinctive feature of the legitimacy of representative democracy, which rests on a balance between responsive and responsible government. In this balance, the former ensures that the preferences of citizens are channelled into the policy-making process, while the latter ensures that governments act within the boundaries of existing norms and procedures and guarantee the long-term provision of public goods (Mair, 2014). In budgetary policy this translates into finding a balance between fiscal stimulus policies which, in the form of expenditure increases and tax reductions, respectively, meet the preferences of either welfare or liberal-oriented voters (Häusermann and Kriesi, 2015) and fiscal consolidation policies, which, in the form of either tax increases or expenditure reductions, contribute to reducing public deficit levels (Truchlewski, 2020).
The combination of the democratic representation of political preferences and the preservation of sustainable public finances is an important element of trust between citizens and governments. More precisely, after several years of responsiveness, governments build a 'reservoir of goodwill' among citizens , who become more willing to accept that in some circumstances public policy may not meet their preferences (Linde and Peters, 2020, pp. 292-293). By preserving the country's long-term interests, in turn, responsible governments ensure that citizens' demands can also be met in the future. In the short term, however, responsiveness and responsibility may sometimes be mutually exclusive, particularly during economic and financial crises, when the tasks of responding to social demands and fulfilling budgetary responsibilities are often diametrically opposed (Schäfer and Streeck, 2013;Bohle, 2014;Lefkofridi and Nezi, 2020). In these instances governments are confronted with the dilemma of which duties they should meet: their responsive duties or their responsible ones (Mair, 2013).
During the years of the Eurozone crisis the preservation of sustainable public finances became the main driver of governments' policies, and demands for social protection were for a considerable extent neglected (Pontusson and Raess, 2012;Schäfer and Streeck, 2013). For the legitimizing principles behind representative government, it is important that after a period of responsibility there is a return of political responsiveness, in order to replenish the trust between citizens and government. In the case of the Eurozone crisis, however, fiscal responsibility not only characterized the policy responses but was also institutionalized in legislation with a strengthening of the Stability and Growth Pact (Laffan, 2014). If responsibility is induced by the logic of European integration and supranational constraints (Mair, 2013;Scharpf, 2015;Ruiz-Rufino and Alonso, 2017), the return of responsive policy-criteria may thus be hampered. This raises the question whether the expected return of responsiveness also happened after the Eurozone crisis.

II. Responsiveness and Responsibility during and after the Eurozone Crisis
The contemporaneousness of the Eurozone crisis and the reform of EEG makes it difficult to disentangle whether the austerity-based policies of the early 2010s are to be understood as simply a crisis response or instead as a structurally led consequence of European budgetary agreements. Crises force governments to choose quickly among a limited set of policy approaches and thereby to define the socioeconomic policy paradigm for leading their countries out of the crisis (Gourevitch, 1986). In the case of the Eurozone crisis, the intergovernmental negotiations of late 2009 and early 2010 were characterized by a confrontation between a German-led coalition favouring a fiscal discipline approach vis á vis a coalition composed of mostly Mediterranean countries advocating fiscal solidarity (Lehner and Wasserfallen, 2019;Wasserfallen et al., 2019). During these negotiations the austerity-oriented ideas of the German-led coalition were more successful in spreading across the political elites of the Eurozone (Blyth, 2013a(Blyth, , 2013bMatthijs, 2016). In the political economy literature, this spread of the fiscal discipline approach has two potential explanations. On the one hand, it can be attributed to the ideas of the policy-makers in charge (Blyth, 2013b;Matthijs, 2016;Bremer and McDaniel, 2020). This explanation finds support in the plain observation that during the years of the Eurozone crisis the political composition of most governments leaned towards the centreright, particularly between 2011 and 2012. 1 During the crucial months for reforming EEG, the particular composition of these governments may have influenced the intergovernmental decision-making process. On the other hand, the spread of the fiscal discipline approach can also be attributed to the fact that it was in line with the pre-existing Maastricht criteria (Verdun, 2015). An alternative explanation resides therefore in arguments about the structural constraints derived from European integration (Schäfer and Streeck, 2013;Scharpf, 2015).
These alternative explanations have different implications for the long-term balance between responsiveness and responsibility. In the ideational line of argument, governments choose their policies according to their ideaswhich are at least partially shaped by their preferences (Bremer and McDaniel, 2020) and are therefore not constrained to switch to responsive policy criteria once the crisis is over. In the structural argument, responsibility is also bound to remain influential for governments' policy-making criteria in the long term, particularly when responsive policies would prevent a government from meeting European budgetary targets. In the party-politics scholarship, the balance between responsiveness and responsibility during and after the Eurozone crisis has mostly been analysed with single case studies of countries that were most severely affected by the crisis, like Greece and Ireland, emphasizing how European budgetary commitments constrain governments from being responsive to their electorates (see Mair, 2013;Lefkofridi and Nezi, 2020). Yet it is in precisely these cases that it is difficult to disentangle the logics of politics in hard times from EEG-related constraints.
This article, therefore, looks at countries that were affected by the Eurozone crisis in different degrees, but that were less harshly hit, like in the case of Greece or Portugal. More specifically, this article focuses on France, Germany and Spain. Besides capturing different degrees of exposure to the Eurozone crisis, this selection also captures the variety of political economies within the Eurozone. Recent literature has emphasized how the new EEG seems to favour the export-oriented economies of northern European countrieslike Germanywhich have an interest in domestic wage compression, and poses unfavourable conditions to the economic growth models of southern European countries, which have traditionally relied on domestic consumption (Johnston andRegan, 2018, 2016). Consequently, EEG may have negative democratic consequences in countries like France and Spain that tend to find it harder to match European budgetary targets with national economic objectives, and it may be less influential in Germany, where meeting European budgetary targets is largely in line with the country's economic strategy. This discrepancy seems to be also reflected in data indicating satisfaction with democracy, which tend to be more negative in southern European countries and more positive in north-western countries (Matthijs, 2017). The balance between responsiveness and responsibility needs therefore to be assessed in these different contexts.

III. Responsiveness Versus Responsibility: Conceptualization and Measurement
To grasp the distinction between responsiveness and responsibility it is necessary to look beyond the types of policies being pursued and to grasp the main criteria behind government action (Karremans and Lefkofridi, 2020). Responsive governments act in line with the political preferences of the parties in government, whereas responsible governments act in line with rules and procedures associated with running the state apparatus (see also Sartori, 2005). In the field of budgetary policies this translates into on the one hand, governments acting according to either centre-left or centre-right political preferences, and on the other complying with public debt and deficit rules (Schiavo-Campo, 2017). The balance between responsive and responsible criteria ensures that the direction of budgetary policy conforms to the political preferences of the ruling parties while at the same time ensuring the long-term stability of public finances.
Budgetary policy follows a yearly cycle in which, generally in autumn, finance ministers are expected to present the government's plans for the upcoming fiscal year to parliament. In France, Germany and Spain these speeches are given to the plenary session of the main legislative chamber and mark the beginning of a week of parliamentary debate on the government's proposed budget. These debates constitute the final and decisive stage of the budgetary policy process, in which the government needs to secure the support of its parliamentary majority (Bonafont et al., 2015). These speeches are therefore an occasion for the government to make clear political statements about its preferred budgetary policy course (Hübscher, 2019) to an audience that not only includes the parliament majority and opposition but, through media reporting, also extends to international stakeholders and local constituencies. The institutional regularity with which these speeches happen, in turn, allows a comparison between the levels of responsiveness and responsibility across years and countries (Karremans and Damhuis, 2020).
The analysis of budget speeches allows to code both the policies presented and the arguments with which these are justified. As anticipated, budgetary policies can be broadly classified into fiscal stimulus and consolidation measures. Justifications can be classified into arguments clarifying how policies serve (certain sections of) the society or economy, and arguments clarifying how the policies conform to national or international budgetary rules about debts and deficits. The combination of different types of justifications and different types of policies can finally be classified into different shades of responsiveness and responsibility. The following passage of the speech by the French finance minister Bruno La Maire from 17 October 2017 constitutes a typical example of a policy justification statement: To facilitate access to corporate capital, to give our entrepreneurs and our entire economy the means to better finance, we will remove the wealth tax. This is an example of a fiscal stimulus policy that is justified by a market-oriented argument favouring private entrepreneurship. Therefore, this is an example of responsiveness towards distinct neoliberal political preferences.
When consolidating the budgetfor example, through expenditure reductionsgovernments generally emphasize the importance of safeguarding the sustainability of public finances. In the following passage from 20 October 2008, for instance, the French finance minister Eric Woerth announces a policy package aimed at restricting public expenditure: The plan for financing the economy works mainly through loans guaranteed by the state. In a crisis, the state is the only actor to which the markets still agree to lend, the only one that the markets trust. This confidence depends on the structural recovery of our public finances, which only a determined action on spending can guarantee. (Eric Woerth, French Assemblée Générale, 20 October 2008).
In this case, the policy is justified with an argument about the importance of maintaining the confidence of financial markets, which will in turn guarantee the provision of liquidity in the future. As this confidence is dependent on the public deficit levels, the finance minister highlights the importance of maintaining stable public finances. This is therefore an example of a justification emphasizing fiscal responsibility. Table 1 summarizes the distinction between responsiveness and responsibility and their nuances, presenting different subcategories of the two central concepts. The sharpest responsive arguments are social and market-oriented justifications of fiscal stimulus policies. These are the instances in which, for example, increases in public spending are justified by arguments about the importance of meeting the needs of low-income groups, or in which tax reductions are justified by arguments about stimulating private entrepreneurship. These categories contain also mixed argumentsreflecting post-industrial ideological changes (Häusermann and Kriesi, 2015) in which tax reductions may be justified as meeting social objectives or, alternatively, spending increases may be presented as measures to support entrepreneurs or as rewards for meritocracy. Justifications referring to stimulus policies and focussing on economic growth are also considered as responsive. These arguments are considered as being responsive to the centre-left when referring to expenditure increases and to towards the centre-right when referring to tax reductions.
It is important to highlight that sometimesor oftenresponsiveness may manifest only in a rhetoric form. These are instances in which justifications focus on social or market-oriented goals, without, however, presenting policies that actually serve those goals. A first example are social and market-oriented justifications that do not refer to any specific policy, but are related only to the overall budget. In other cases, these responsive justifications may be used in relation to fiscal consolidation policies, for example, with claims that entrepreneurs are being partially sheltered from tax hikes, or that essential social services are exempt from expenditure cuts. These justifications are still coded as responsive for their focus, but they refer to policies that serve purposes of fiscal responsibility. The coding scheme presented in Table 1 therefore keeps track of whether the responsive justifications also present policies that serve their social or market-oriented goals.
Responsible justifications feature arguments about meeting financial commitments such as a balanced budget, and complying with international rules. Also for these justifications it is important to trace whether they actually present fiscal consolidation policies or whether they are used rhetorically. Just as social and market-oriented arguments may be used to justify fiscal consolidation policies, responsible justifications may be used to justify fiscal stimulus policies. Expenditure increases or tax cuts may, for example, be justified by the argument that they will not increase public deficit. In these instances, while their policies may potentially increase deficit levels, governments still try to present themselves as fiscally responsible. A rhetorical use of responsible justifications may also be related to the overall budget. In these instances, the justifications present the government as fiscally responsible without, however, referring to specific measures. The most unambiguous manifestations of fiscal responsibility are when statements about budgetary targets and international rules are used to justify fiscal consolidation policies. In some instances fiscal consolidation may be presented as the only alternative for ensuring a return to economic growth. In other instances the government may simply limit itself to explaining that these unpopular measures are needed to reduce deficit levels or to meet international commitments.
Finally, when the general policy is justified by arguments about macro-economic policies -for example, with the statement that the budget will help to improve the country's economic performancethese are categorized as neutral, in the sense that they do not portray the government's action either as responsive to political preferences or as conforming to budgetary rules. In the Supporting information Appendix 1 (Table S4), examples of each of the coding categories are provided.

IV. Research Strategy
Coding responsiveness and responsibility allows us to measure their relative variation over time. This allows us to test whether responsiveness and responsibility alternated with each other during the 2010s, or whether instead responsibility remained the main policy-making criterion in the second half of the decade also. To make a more precise assessment of whether during the 2010s fiscal responsibility was driven by governments' ideas or by structural constraints, this article carried out a statistical analysis in which the changes over time in the balance between responsiveness and responsibility were controlled for against institutional and political factors.  (4) 2017-19 as the period long after the Eurozone crisis. To make a more precise assessment of the changes over time in the balance between responsiveness and responsibility the analysis takes int account country-specific institutional and political factors.
Among the institutional factors, the analysis considers deficit levels and credit ratings. These represent a yearly variation in the budgetary or financial situation of the countries and therefore the institutional pressures to which governments are exposed. Deficit is captured by two indicators. The first indicates the actual deficit or surplus levels in the year preceding the budget year, which in most cases corresponds to the year in which the speech was delivered. As an example, a speech delivered in October 2012 presents the budget for 2013, and in the analysis it is associated with the deficit and surplus levels of 2012. Government deficit and surplus data is retrieved from the Eurostat database (2020), expressed as a continuous variable ranging between À11.3 per cent in Spain in 2010 to +1.9 per cent in Germany in 2019. The second indicator captures budgetary deficit excess and is expressed as a binary variable in which a value of 1 indicates a condition of excessive deficit and 0 indicates that public finances are within a certain threshold. For France and Spain the threshold is set at the Maastricht level of À3 per cent of GDP. For Germany the threshold is set at 0 per cent, because of its national debt brake law. This binary variable is intended to capture more directly the expected impact of institutional constraints. If at the time of delivering a budget speech governments are exceeding the deficit thresholds they should be under stronger pressure to beor at least to appear to be -fiscally responsible.
The credit-rating variable captures the lowest credit rating that the country was assigned by one of the leading rating agencies (Fitch, Moody's or Standard & Poor's) in the six months prior to the budget speech. Before the global financial crisis the three countries selected for this study relied on stable triple-A ratings, indicating the highest quality investment grade (see Supporting information Appendix 2, Table S5). During the Eurozone crisis west European countries became increasingly confronted with the risk of being downgraded to lesser quality investment grades. Among the three countries analysed in this article, only Germany was able to maintain its triple-A rating, while France was downgraded to a double-A from 2013 onwards and Spain was declared to have only an adequate payment capacity (triple-B rating) between 2012 and 2018. Based on the differentiation in investment grades used by rating agencies (Afonso et al., 2007) and given the rating range among the selected cases over the time period analysedcredit rating is coded as a continuous variable ranging from 3 to 0. 2 The highest value (3) corresponds to the highest quality investment grade (that is, triple-A ratings); 2 to a high quality investment grade (double-A); 1 to a strong payment capacity (single A); and 0 corresponds to an adequate payment capacity (triple-B; see also Supporting information Appendix 2, Table S5). As one of the aims of European fiscal coordination is to maintain high levels of market confidence across the Eurozone, from an institutional point of view, budgetary discourse is expected to be more fiscally responsible when a country has a low credit rating.
Following the observation about the centre-right composition of many governments during the years of the Eurozone crisis, the analysis also tries to grasp whether the centre-right composition of a government influenced the balance between responsiveness and responsibility. In the analysis, government ideology is coded as a binary variable with a value of 1 for governments that were exclusively composed of centre-right parties. Among the cases analysed in this study, these are the second Merkel government for Germany, the governments of the Sarkozy presidency for France and the first and second Rajoy governments for Spain. The remaining governments of the case selection (see Table 3) were assigned a value of 0.

V. Findings: Responsiveness and Responsibility in Germany, France and Spain (2007-19)
The coding scheme presented in Table 1 was applied to the yearly budget speeches delivered during three consecutive legislatures in France, Germany and Spain between the global recession of the late 2000s until 2019 (for the detailed list of speeches coded, see Supporting information Appendix 1, Tables S1-S3). For France, this meant coding the yearly budget speeches delivered under the presidencies of Sarkozy, Hollande and Macron (from October 2007 to October 2019). For Germany, these were the speeches from the second, third and fourth Merkel governments (January 2010 3 -September 2019). For Spain, the analysis covers the speeches from the second Zapatero government and the first two Rajoy governments (October 2008-April 2018 4 ). In total, 34 speeches 2 The actual ratings used by the agencies are more detailed than these investment grades. For example, Standard & Poor's adds plus (+) and minus (À) signs to its ratings, a minus signifying somewhat lower quality. To capture this more fine-grained variation, it is also possible to use a 17-point scale (Afonso et al., 2007). For the purpose of this paper, the main interest is whether downgrade to lower investment grades affects the levels of responsibility in a yearly budget speech. The OLS models presented in Table 4 are therefore based on the broader categorization of investment grades. A replication of the analysis using a 17-point scale is presented in Supporting information Appendix 2 (Table S6), showing that the main conclusions remain unchanged. 3 While they are generally given in autumn, in election years budget speeches are sometimes moved to the following winter or spring. The first budget speech of the second Merkel government was thus delivered in January 2010. 4 After 2015, Spanish budget speeches started occurring in spring (see Supporting information Appendix 1, Tables S1-S3 for precise dates of speeches).
were coded: 13 for France, 11 for Germany and 10 for Spain. The coding of these speeches generated a dataset of 3,963 policy-justification statements. Table 2 shows how the total number of observations (that is, policy justifications) is distributed across the coding categories.
The overall frequency of responsive and responsible justifications is skewed towards responsibility (33.61% vis á vis 42.24%). The more unambiguous responsive and responsible statementsthat is, justifications for fiscal consolidation with arguments about budgetary rules and social or market-oriented justifications for fiscal stimulus policiesconstitute a significant share of the overall discourse, namely 15.19 per cent and 13.02 per cent, respectively. The justifications referring to international commitments appear in relatively limited numbers, amounting to 8.47 per cent of the total number of justifications, most of which refer to the general action of the government and not to specific budgetary measures.
At first sight the relative frequencies of responsive and responsible justifications seem to corroborate the view that under the new EEG framework governments are inclined to pursue national budgetary policies according to fiscal responsibility criteria. The yearly distribution of justifications, however, shows that only the years of the Eurozone crisis were dominated by fiscal responsibility. In the subsequent years governments gradually started again to frame their budgets in responsive terms. Figure 1 shows the share of responsive, responsible and neutral arguments at each budget year, that is, the fiscal year that the budget speech refers to (thus, a budget speech given in autumn 2013 refers to the fiscal year of 2014). The yearly distribution of responsiveness and responsibility shows a predominance of the latter during the years of the Eurozone crisis. From autumn 2013 onwards governments start presenting their budgetary policies according to either social or market-oriented preferences. Figure 2 shows that the gradual decline of responsible discourse is common to all three countries. The decline is clearest in Germany, which features the highest shares of yearly responsible statements of the entire case selection; namely, around 65 per cent for the 2012 and 2014 budgets (which were presented on 6 September 2011 and 30 January 2014, respectively). In subsequent years there is a gradual decline every second year, reaching also the lowest yearly shares of responsible justifications of the entire case selection; namely, 19 per cent and 23 per cent, respectively, for speeches presenting the 2019  and 2020 budgets. The Spanish case most closely resembles the German pattern, registering its peak of yearly responsible justifications in 2012, and a gradual decline during the following years. The gradual decline over time is less clear in France, as it registers only three relatively high peaks of responsible justifications in the speeches for the 2009, 2013 and 2015 budgets, which were given in October of the previous year, respectively. Nonetheless, the French case also seems to contradict the proposition that under EEG national budgetary discourse is persistently dominated by the presentation of fiscal consolidation policies and arguments about public deficit levels.
The return to higher levels of responsiveness is also confirmed when looking at the distribution over time of the specific subcategories. Table 3 illustrates the distribution of policy-justification statements for each government, distinguishing also between the subcategories of responsiveness and responsibility. Besides showing that the most recent governments of the case selection are clearly the most politically responsive, the percentages reported in Table 3 also indicate that the increases in responsiveness include substantial increases in the unambiguous responsive categories; namely, those that justify fiscal stimulus policies with social and market oriented arguments. Note also the diverse forms that responsiveness seems to take in the three countries as, for example, under the Macron presidency there is a peak of market-oriented responsiveness, whereas in the German and Spanish cases the increases in responsiveness are almost entirely socially oriented. This suggests that governments have room to respond to distinct social demands.
The observed variation in the contents of policy justifications over time is an indication that under the new EEG, fiscal responsibility does not necessarily reduce responsiveness. However, this does not cancel out the possibility that the peaks of responsibility during the first half of the decade were institutionally driven by EEG rules, rather than by political choices. This would entail that, while in good economic times governments are under less pressure to stick to budgetary rules and may therefore be fiscally more responsive, in times of crisis EEG rules induce governments to respond with policies of fiscal consolidation, disregarding governments' political preferences. To establish whether the peaks of responsibility were driven by political or institutional factors, a more fine-grained analysis is needed accounting for differences over time in countries' fiscal situation and government partisanship.
Based on multivariate ordinary least squares regression models, Table 4 reports estimates of how the share of responsible justifications varies over the four time periods described in Section 5 (2007-8, 2009-13, 2014-16, 2017-19), controlling for institutional and political factors. The level of responsibility is calculated against the level of responsivenessexcluding the neutral argumentsand therefore indirectly also shows the variation of responsiveness (that is, when responsibility increases, responsiveness declines and vice versa).
Model 1 presents the association between the time period and the level of responsible discourse, net of country differences. The estimates show that between the first and second time periods, which correspond to the global financial crisis (reference category) and the Eurozone crisis , respectivelythere is no substantial variation in the levels of responsibility. These levels start declining in the first post-crisis period (2014-16) by almost 16 percentage points. The decline continues also in the last observed time period , in which the share of responsible justifications is estimated to be almost 29 percentage points lower than during the years of the global financial crisis.   For the two post-crisis time periods the estimates are statistically significant (P value < 0.01). Model 2 introduces three controls to account for variation over time in the institutional constraints: yearly surplus or deficit level, budgetary deficit excess and credit rating. The institutional constraint estimates show that net of the time period, surplus (or low deficit levels) and higher credit ratings are negatively associated with responsibility and thereforeby reflectionthey have a positive effect on responsiveness. This entails that when governments are in a positive budgetary situation and have high credit ratings they have more room for manoeuvre to be responsive. Budgetary deficit excess, by contrast, has a weak independent association with the balance between responsiveness and responsibility, which indicates that exceeding the prescribed budgetary thresholds does not have any additional effect on responsible discourse net of the time period. This is the case also when excluding the other two institutional controls from the model. Importantly, Model 2 estimates show that the over time decline of responsibility remains strong and statistically significant, also net of variation over time in countries' fiscal situation. This implies that the stark differences in the level of responsible discourse between the economically hard and good times are not explained by changes in the fiscally induced pressures to which the governments were exposed. Model 3 introduces the government partisanship control; namely, whether the budget is presented by a conservative government. The estimates show that the centre-right composition of a government is positively associated with the share of responsible justifications. Net of the time period and institutional constraints, centre-right partisanship in government is associated with a 5.5 percentage points higher level of responsibility. After accounting for government partisanship, the steepness of the decline over time in responsible justifications decreases, suggesting that the relatively high levels of responsibility during the global economic downturn and the Eurozone crisis were to some extent driven by the partisan composition of government. In order to capture the extent to which government partisanship mattered during and after the Eurozone crisis, Model 4 introduces an interaction between partisan composition and each of the four time periods, with the aim of disentangling whether the austerity of the early 2010s was related more to political preferences or to institutional factors. The estimates show that the effect of government partisanship is sizeable and statistically significant only during the Eurozone crisis period, with centre-right governments estimated to have a 9 percentage point higher share of responsible policy justifications compared with the non-centre-right executives (P value < 0.05). The estimate thus suggests that interaction between the crisis and centre-right government triggered additional levels of responsibility. The association between a centre-right government composition and responsibility levels is small and statistically nonsignificant in the post-crisis period. These findings thus corroborate the idea that the high levels of responsibility during the Eurozone crisis years were at least partially driven by political preferences.
The results of Table 4 show that responsibility was dominant only in a circumscribed period of time. Even though the peaks of responsibility are associated with deficit levels and credit ratings, the interaction analysis presented in Table 4 indicates that the focus on fiscal responsibility in the budgets presented between 2009 and January 2014 was also driven by the centre-right composition of governments. This suggests thatif during the years of the Eurozone crisis member states were governed by governments of different partisan compositionthe policy response might have taken a different form. These findings, therefore, weaken the proposition that EEG structurally induces governments to be fiscally responsible. The choice for or against fiscal consolidation, instead, seems to be at least partially dependent on the preferences of the incumbent government.

Conclusion
The findings indicate that the austerity of the early 2010s was a story of politics in hard times, during which the member states' governments agreed on a common policy approach of fiscal discipline. The extent to which this approach was adopted, the findings indicate, was influenced by the partisan composition of government. The findings of this article raise the impression that if during the early 2010s member states had not been governed by conservative cabinets, the austerity of the early 2010s might have taken milder or different forms. The findings therefore are in line with recent insights about how the fiscal consolidation policies of the early 2010s were political choices driven by economic ideas that took different forms under different political forces (Bremer and McDaniel, 2020). In addition, the findings suggest that the grip these ideas had on governments was temporally limited.
Since 2014 governments have increasingly presented fiscal stimulus measures to meet domestic socioeconomic demands. Furthermore, the post-2014 increase of responsiveness identified in this article is also aligned to a broad pattern that is observed across EEG, with developments such as the 'socialization of the European semester' (Zeitlin and Vanhercke, 2018), by which European country-specific recommendations are increasingly focused on social issues. An avenue for future research is to investigate the extent to which the social policies presented in national budgets align with these recommendations, and whether these proposals originate from domestic party platforms or EU policy reports. However, as even in the sphere of treaty-based budgetary constraints national governments seem to be able to cherry-pick EU policy guidelines (Eihmanis, 2018), it is likely that their arbitrariness is even greater in the sphere of social policy recommendations. Therefore, the partisan composition of national governments still matters substantially for the form that socioeconomic policies take.
On a final note, it is important to highlight that the most responsive governments of this case selection have been confronted with mass protests or vast societal discontent, often challenging the legitimacy of public authorities, like the gilet jaunes or more recent protests against proposed pension reforms of the Macron presidency in France. In Spain, the second Rajoy government was confronted with the threat of the Catalan separatists, which was largely about how resources were redistributed within the country. In Germany popular discontent was not as sharp, but all the same among the Social Democratic Party base a profound discontent persisted over the government's socioeconomic policies. This persistent discontent, in spite of a return of political responsiveness, calls for deeper reflections on the legacy of the austerity of the early 2010s. The high levels of fiscal responsibility of the 2009-13 period indicate that for more than three consecutive years debt and deficit levels were by far the main concerns for budgetary policy in Europe and that societal concerns were for a great extent neglected. The balance between responsiveness and responsibility is functional to generating among citizens a reservoir of goodwill, by which decisions that are not directly favourable become easier to accept (Linde and Peters, 2020). The extensive austerity pursued during and after the Eurozone crisis years has possibly drained this reservoir, potentially harming the chain connecting citizens to policy outputs. The consequences of the extensive austerity of the 2010s for the trust between citizens and governments are still to be reckoned with in the future.