Crisis and paradigm change in the European semester: from austerity to investment-oriented policy ideas

ABSTRACT The policy-responses to the COVID-19 pandemic have fuelled expectations about potential paradigmatic change in European Economic Governance (EEG). Building on existing scholarship on policy paradigms, we develop testable expectations about the steps preceding paradigmatic change, which we explore on the basis of three types of data. First, we show how public spending in three key Eurozone countries between 2008 and 2021 follows predicted patterns of the punctuated equilibrium model. Second, we show that governments’ justifications for their annual budgets reflect a gradual change in policy-ideas between 2009 and 2020, following the expected three orders of change. Third, we show how this gradual change is also present in the European Country-Specific Recommendations and in the bureaucratic logics within the European Commission. Our findings reconcile three strands of scholarship on policy change and have implications for our understanding of the European integration process and for future research on economic policy-making within EEG.


Introduction
In the last two decades, public budgeting in Member States (MS) of the European Union (EU) has been subject to several major shocks, from the Great Recession and the sovereign debt crisis, first, and of the COVID-19 pandemic, later.In the context of the Economic and Monetary Union (EMU), both crises have demanded countries to pursue a common framework to design their policy response, agreed at the supranational level by European institutions.Consequently, economic and budgetary choices of MS have been driven and shaped by common economic ideas (Bremer and McDaniel 2020;Matthijs 2016;Pontusson and Raess 2012;Schmidt 2020).The two crises, though, have not been handled with the same recipes.While the response to the sovereign debt crisis consisted in a reinforcement of the austerity ideas entrenched in the EU's treaties (Verdun 2015), the COVID-19 pandemic has triggered extensive high-level debates about the possibility of reforming the key rules of European Economic Governance (EEG), which in fact has just started. 1 The case of the COVID-19 crisis seems thus to reflect what scholars of paradigm shifts identify as an external shock that offers a window of opportunity for a change of paradigm.
According to existing scholarship on paradigm change (Baumgartner 2013), crises are not the only moments in which change happens, but are rather a momentum in which previous series of small incremental changes get translated into an entire new paradigm.For decades, public policy scholars have been engaged in developing theories to explain paradigmatic shifts, particularly with regards to public budgeting because of its primary role for the functioning of the state.Despite the different intellectual traditions and methodological choices, scholars have reached very similar conclusions about rules dominating policy changes (Baumgartner 2013).In both qualitative and quantitative scholarships, it has been established that public budgeting follows a punctuated equilibrium in which large changes are concentrated in a very few years, mostly in concomitance with a dramatic shock.To the extent that the COVID-19 crisis represents a momentum of paradigm change, was this preceded by a series of small incremental changes, as predicted by existing theories on paradigm change?
In this paper, we explore this question, aiming to provide a threefold contribution.Firstly, while existing studies on the potential paradigmatic change brought by the COVID-19 pandemic tend to focus on ongoing events (e.g.Ladi and Tsarouhas 2020), in this study we systematically apply existing theories about paradigmatic change and test them in the context of EEG between 2008 and 2021.Secondly, we contribute to the literature on paradigmatic change by combining three different traditions of comparative research (Hall 1993;Padgett 1980Padgett , 1981)), Jones and Baumgartner 2005) to study paradigmatic change in public budgeting at three different levels of observation, namely public spending, policy-justifications in public discourse, and ideas among bureaucrats.Thirdly, our findings offer an important contribution to the literature on European integration, suggesting that the EU is subject to the same competition of economic ideas that have been observed in the past at the state level (Hall 1993), which renders the course of future integration more uncertain and open to many paths.
Our paper studies public budgets of France, Germany, and Italy.Because of their size and political economic differences (Johnston and Regan 2018), the focus on these cases allows us to capture the extent to which economic ideas are shared by some of the most powerful governments that often stand in opposing coalitions in European intergovernmental negotiations (Lehner and Wasserfallen 2019).In Section 2, we theorize how budgetary policy needs to be based on an underlying broad strategy on how to maintain long-term sustainability and that radical changes to this strategy are relatively rare.Subsequently, in Section 3 we explore how under the EMU, governments have settled on a common strategy for long-term financial sustainability and how the related policyideas circulate among key actors in EEG, namely the European Commission, the European Central Bank and the European Council.Comparing these ideas over time, we observe major differences between 2020 and the pre-Covid-19 period, suggesting that a paradigmatic change has indeed taken place.
To validate this conjecture, in Section 4 we develop three testable hypotheses from existing theoretical scholarship on paradigmatic change and, after having presented the research strategy in Section 5, in Section 6 we proceed in a three-step analysis.First, we show how public spending data for the three countries over the 2008-2021 period reflect the predictions of the Punctuated Equilibrium theory (henceforth PE), according to which actual budgetary changes are concentrated in a limited number of years.Secondly, studying policy-justifications for the annual budgets, we show how governments' ideas about fiscal sustainability changed, moving from the goal of low debts and deficits during the sovereign debt crisis, gradually emphasizing the need for higher economic growth between 2014-2019, towards portraying public spending as potentially beneficial for long-term fiscal sustainability in 2020.Thirdly, based on a coding of the European Semester's country-specific recommendations and on recently published in-depth qualitative studies on the inner-workings of the Commission (Mérand 2021;Schmidt 2020), we trace how austerity ideas gradually disappeared from the macro-economic policyprescriptions, which became instead gradually more oriented towards socio-economic investments.In the concluding section, we discuss the relevance of our findings.

Theories of policy-change in public budgeting
The budget is a document that defines the policies that the government aims to implement in the next year (and, usually, in the following three or five years depending on the national design) with concrete financial commitments.In shaping the budget, governments have the possibility to introduce policies that are congruent with their political preferences, but they must have at the same time a strategy for ensuring the long-term sustainability of public finances (Yarrow 2008).Broadly speaking, there are two main ways in which governments can guarantee such sustainability.Governments can either pursue a more investment-oriented approach involving high public spending with the promise of high fiscal returns, or a more neo-liberal approach involving low taxation with the promise of enough fiscal returns to finance the necessary expenditures (Ólafsson et al. 2019).While adaptations to the budgetary policy-mix may change in the short-term according to governments' political preferences, changes in the overall strategy for ensuring longterm financial sustainability are likely to require longer periods to materialize, because they are narrowly linked with the functioning of the domestic economy.In fact, the economic preferences of the governments of the Eurozone during intergovernmental negotiations are overall consistently aligned on more expansionary and more restrictive budgetary policy-approaches, disregarding the political composition of the cabinet (Wasserfallen et al. 2019).In other words, even though in pure ideological terms left parties may favour an expansionary approach and right parties a more restrictive approach, at the stage of policy-making, the strategy for long-term budgetary sustainability is a state interest that often transcends partisan differences (Johnston and Regan 2018).Hence, the approach by which governments use the budget to sustain the economy and ensure long-term financial sustainability is what can be called a paradigm (Hall 1993;Ólafsson et al. 2019).
As highlighted by Baumgartner (2013), existing research that documents how and to what extent policy changes in budgetary approaches can be divided in three strands of scholarship (Table 1).Firstly, research clustered around the PE theory is currently recognized as 'the most serious attempt to formulate a formal explanatory model of policy and budget dynamics' (Jensen, Mortensen, and Serritzlew 2016, 227), being able to elucidate on the magnitude and frequency of budgetary changes in multiple contexts and budget categories.Research on the PE theory has demonstrated that large budgetary changes (i.e.punctuations) occur within a very little time span as consequence of an erroraccumulation process which causes growing pressure for change 2 or because of an external shock which demands immediate action.In subsequent years, adjustments are hyper-incremental (i.e.tiny and adaptive) and punctuations will happen again only after a long period of stasis.While the PE explains that during processes that bring about punctuations the conditions for a change of paradigm are met, this theory doesn't elucidate on whether changes of paradigm successfully happen in those moments.
A second strand of scholarship focuses on frames and policy-ideas (Hall 1993;Karremans 2021b) and was developed from the insight of cognitive sciences which argues that the way people frame an issue in their minds has a substantial impact on their actions and choices in that domain.It also builds on the insight that policy-makers' learn about policy-failures and adapt their interventions accordingly (Dunlop and Radaelli 2016;Hemerijck 2012).In order to understand policy-makers' cognitive frameworks and their learning capacity, the study of policy-makers' discourse has become an established practice in political economy and public policy analysis, particularly for understanding their ideas and for trying to explain policy change (Schmidt 2010).This strand of research identifies acknowledged moments of policy change, and shows how these were preceded by different discourses, and therefore ideas.In the context of the Eurozone crisis, the relevance of austerity-ideas has been identified to have played a key role in shaping policy-responses (Blyth 2013;Bremer and McDaniel 2020;Karremans 2021b).
A third research avenue involves looking more in depth into the budgetary process, focussing on bureaucratic levels.Padgett (1981) described the US budgetary decisionmaking process as a three-levels hierarchical model.Each level is characterized by a particular organizational logic, which shapes how decisions are made.As we move from the lowest to the highest level -that is from programs (1) to agencies (2), then from agencies to the president (3) -the cultural and professional norms change and are responsible for specific budgetary choices and outcomes.While the lowest level deals with the administrative determination of 'proper' budgetary allocations, and norms determine which program are controllable and which are not, more important budgetary changes require the intervention of the second level (Office of Management and Budget = II order change, changes in the policy instruments themselves used to achieve shared policy goals = Revisions to policy that must be approved by a higher political authority 3 New policy problems identified as major priorities = III order change, shift in the goals themselves = Availability of resources when the right bureaucratic logic and evidence being adopted by higher authorities, random timing Source: authors' own elaboration from Baumgartner (2013).
or cabinet secretary level), which have the authority for deciding about the spending priorities among budget programs.A real policy change, that is, a change of the goals happens only if the right bureaucratic logic about the macroeconomic determination of total federal spending is adopted by higher authorities with the president's support and succeeds in overcoming the power of the status quo.Not all ideas can successfully bring about a paradigm change, as these may not have enough strength to replace the previous policy idea (which is not perceived to have failed) and may not have reached the highest bureaucratic level.
As summarized in Table 1, the different scholarships on budgetary (and, more in general, policy) changes share the evidence that there are regularities in the way in which policies change (Baumgartner 2013;Hall 1993).All the three strands of research find that large reorientations in public budgeting are preceded by periods of small incremental change.They all concur that public budgeting is a 'sticky' process, in which past legacies keep structuring contemporary choices until a new momentum of change.In European countries' budget, the COVID-19 pandemic surely represented a shattering event that endangered the stability of European public finances and that therefore had the potential to cause a paradigmatic change.Following current debates about the innovative, and to some extent, revolutionary aspects of the economic policy-decisions taken since March 2020, we conjecture that the COVID-19 triggered the third and highest level of policy-change from Table 1, creating a situation in which 1) new resources are made available by high authorities, 2) the goals of public budgeting change and 3) new policy problems are identified as major priorities.
Even though this scholarship is based on research on autonomous governments, these insights also offer the possibility to establish the extent to which there has been paradigmatic change within EEG.Before empirically analysing the steps leading to a potential paradigm change, it is however first necessary to clarify how EEG as a polity 1) makes resources more or less available for national budgets, 2) defines the goals of national budgetary policy and 3) delineates the policy-problems that governments must address as major priorities.This allows us to map whether the actions taken since the outbreak of the COVID-19 pandemic represent a substantial change of budgetary approach on all three dimensions.Incidentally, as for MS public budgeting is no longer a purely national affair, we need to take into account the developments within EEG, which considerably shapes national policy decisions.

Budgetary ideas and policy since the wake of Maastrichtaastricht
The Eurozone is an area in which national governments cannot autonomously decide on their long-term financial sustainability strategy.With the signing of the Maastricht Treaty in 1992 and the creation of the EMU, national budgetary policies started to be limited by supranational requirements and coordinated between countries and European institutions within the multilevel governance.The integration privileged the monetary side, creating the supranational and independent European Central Bank and later introducing a common currency, with clear guidelines with regards to the dilemma between fiscal expansion and fiscal restraint.The benchmarks of a deficit-to-GDP ratio below 3% and a debt-to-GDP ratio below 60% clearly pushed for the latter, inducing countries wishing to enter the EMU to pursue policies of fiscal consolidation (Rotte and Zimmermann, 1998), guarded by the European 'external constraint' (Dyson and Featherstone 1996;Moschella 2017).The shared ideas within the European policy community about what constituted 'good' budgetary policy thus started to be explicitly based on a recipe of fiscal restraint.
The enshrinement of these ideas into binding commitments happened through a gradual process that reached its climax during the sovereign debt crisis of the early 2010s (Verdun 2015).The rules of the Maastricht Treaty were further developed and reinforced in 1997 with the Stability and Growth Pact (SGP), a set of rules to pursue sound national budgetary policies and reinforce their coordination.For several years, though, the SGP's framework for economic governance remained in the background of the national budgetary policy-making processes, as European institutions had no 'teeth' to sanction MS that would not meet the prescribed thresholds.This became particularly evident between 2003 and 2005 when 12 MS -including France and Germany -ran up excessive deficits compared to the limit set by pact (Saurugger and Terpan 2016).What was still missing during the early 2000s was an institutional framework that allowed enforcing rules about the agreed budgetary goals.
In the context of the sovereign debt crisis, the SGP was reformed in ways that granted the European Commission the faculty to begin the procedure against those MS breaching (or at risk of doing so) the SGP parameters.With the introduction of the European Semester in 2010, a common timeline was set up for all MS to submit and revise their national budgetary plans according to the European guidelines (Dunlop and Radaelli 2016).In this framework, the Commission gained a powerful position especially on economic policy surveillance and coordination, increasing the breadth and depth of its supervisory powers (Bauer and Becker 2014).In parallel, the European Council was formally recognized as the predominant actor in economic governance (Puetter 2014) by convening almost once a month to discuss and coordinate those national policies put in place to face the consequences of the crisis.To endure the effective functioning of the European Semester, the Union adopted the Six-Pack in 2011, the Treaty on Stability, Coordination and Governance (TSCG, also referred to as Fiscal Compact) in 2012, and the Two-Pack in 2013.These new instruments, in turn, were also accompanied by a substantial increase in the stringency of national fiscal rules, entailing that governments had to keep low deficit levels not only because of European agreements but also because of national legislation (Doray-Demers and Foucault 2017).The Fiscal Compact overcame the excessive deficits rule by enshrining the balanced budget rule into national primary laws, thus considerably reducing MS' discretion (Laffan and Schlosser 2016) and amplifying the commitment of countries to shared fiscal rules.
The framework for EEG that resulted out of the reforms introduced between 2009 and 2013 was thus essentially a strengthening of the principles for national fiscal policy laid out in the 1992 Maastricht Treaty.Even though in the specific domain of social policy there were initiatives for finding new ways to invest in human capital (Zeitlin and Vanhercke 2018), the 2009-2013 reforms of EEG had the main aim of reinforcing 'the credibility and enforceability of the EMU's rules-based economic coordination regime' (Laffan and Schlosser 2016), disallowing de facto fiscal expansion (Guter-Sandu and Murau 2022).One of the main goals of these reforms was to gain confidence of financial markets.The governance framework came therefore to be considered as an expression of a neoliberal paradigm, targeted at promoting fiscal restraint even in those countries with a long tradition of high public deficits, 'forcing' them to commit to tight fiscal rules.Under this framework, national governments were allowed to increase expenditures only to the extent that these were backed by sufficient revenues.As in the European integrated market economy increasing taxation may be counterproductive for the economy as it may chase away investors, this paradigm severely reduced the room for new spending programmes (De Grauwe 2016).
In the following years, despite the already-evident shortcomings (IMF 2012; OECD 2014) and serious damages inflicted to debtor states (e.g.low growth and rising inequality in the long-run (Perez and Matsaganis 2018), both the surveillance and the enforceability of the rules-based coordination regime were strengthened (Laffan and Schlosser 2016).Meanwhile, some countries started to ask for more flexibility in the application of the parameters, that led on some occasions to a fierce confrontation with the Commission (for instance, in Italy under Renzi and Conte I governments; see Brunazzo and Della Sala 2016;Cavalieri 2023).Again, while the Commission, following direction from the European Council, issued in 2015 a revised guidance about the implementation of the SGP (Council of the European Union, 2015), these new ideas did not alter the existing goals and objectives of EEG, which remained grounded on fiscal discipline-based ideas (Schmidt 2020).

Budgetary ideas and policy after the pandemic
The COVID-19 pandemic and the related economic recession, by contrast, triggered a new type of fiscal response, different from the response to the Great Recession.First, the Commission activated the general escape clause of the SGP, which allows MS to temporary deviate from the deficit and debt parameters in order to face the huge fiscal response demanded by the effects of the pandemic.New resources have been made available by means of the Commission, mainly through a massive Pandemic Emergency Purchase Program (PEPP) -initially amounted to € 750 billion, further raised up to a total of € 1850 billion (ECB, 2023) -and a recovery plan called Next Generation EU (NGEU).These resources are not only massive funds never made available before to MS but constitute a real novelty, as they are a brand-new policy instrument in the EEG framework.The power of the European Commission to borrow money on the financial markets, and to transfer funds to the MS represents a considerable rebalancing of the two elements of EMU, namely the economic and monetary aspects (D'Erman and Verdun, 2022;Fabbrini 2022).On the side of economic policy, the pandemic crisis seems to have triggered a further integration of fiscal affairs, with the Commission being able to borrow money for NGEU on the financial markets, transfer these resources to MS as grants and repay these funds through tax revenues (Fabbrini 2022, 197).These resources can be used by national governments under the condition that they use them for specific goals. 3These goals are substantially different from those laid out in the Maastricht Treaty (no fiscal consolidation for confidence in financial markets, but economic growth and improvement of health and social infrastructures).This was perfectly exemplified by the communication of the Commission that acknowledged the many transformations happened since the Maastricht Treaty and that '[n]ew challenges such as the green and digital transitions and energy supply issues will require [. ..] major reforms and investments for years to come' (Dombrovskis, in November 2022).Changes affected not only policy problems -now identified in healthcare and societal well-being (Brooks and Geyer 2020; Wolff and Ladi 2020) -but, most importantly, the overarching principle of the EU, which moved from responsibility to solidarity (Ioannidis 2020).

Theorizing paradigmatic change in the Eurozone
When looking at the resources made available for public budgeting, and goals and problems spelled by key actors in EEG, it emerges that the policy responses agreed by European and national actors to the COVID-19 pandemic tick all the boxes of what public policy scholars (Baumgartner 2013;Hall 1993) define as paradigmatic change.In the pre-Covid-19 period, national budgeting in EU MS was subject to an austerity-based paradigm, enforced by the Maastricht Treaty.Even though it was formally established with the Treaty of Maastricht, this paradigm came into its full effect only from 2009 onwards, when the doom-scenario of sovereign defaults led the Union to endow itself with new resources and Eurozone countries into adopting a common fiscal approach to effectively implement the neo-liberal recipes laid out in the existing treaties (Verdun 2015).
All in all, while under the Maastricht Treaty European rules have mostly acted as a brake on national public spending, the post-Covid-19 European agreements not only allow for more flexibility, but also introduced resources for MS to invest in the long-term economic growth, both in the form of flexibility in deficit rules and in the provision of new funds.Our argument does not foreclose that pre-2020 EU institutions could at times recommend governments to pursue policies that would involve increases in social or economic spending.Existing research on the European Semester has documented how the Commission has also acted as a promotor of social services during the 2010s (Verdun and Zeitlin 2018).We argue however that, before 2020, expenditure increases were deemed appropriate only if the overall balance remained within the prescribed budgetary thresholds.In the aftermath of the COVID-19 pandemic, instead, this cognitive framework has changed, with a shared European consensus that expenditures may temporarily exceed revenues if they ensure fiscal returns in the long-term.
Figure 1 graphically illustrates the timeline for the argument about paradigmatic change in budgetary policy in Europe.The austerity-based paradigm was established in 1992 with the Maastricht Treaty and entered with full force in national policyagendas in 2009, in the context of a strengthening of the EEG framework.Based, on the one hand, on requests of flexibility from some (especially Southern) MS and, on the other hand, on the recognition about the shortcomings of austerity within the Juncker Commission (Schmidt 2020), from 2014 small incremental changes in budgetary approach are introduced, but still within the overall framework of maintaining the prescribed budgetary thresholds.The migration crisis of the 2014-2016 period also put some pressure on European public finances, but not to the extent to trigger a change in the overall budgetary approach, arguably because, unlike the sovereign debt crisis and the pandemic crises, it did not pose a direct threat to long-term financial sustainability.Only the advent of the COVID-19 and its economic consequences was forceful enough to trigger a new paradigm that shifted the focus of the budgetary approach towards investment, and that relied on principles of fiscal sustainability often theorized by scholars of social investment (Esping-Andersen 2002; Hemerijck 2012; Hemerijck, Ronchi, and Plavgo 2023).
Our conjecture is that the post-Covid-19 paradigmatic change followed the patterns theorized by existing scholars of policy-change (Baumgartner 2013;Hall 1993).This entails that it was preceded by small incremental changes.To further explore the validity of our argument, based on the theoretical propositions summarized in Table 1, we developed the following hypotheses: H1: In terms of public spending output, we expect that the biggest changes are concentrated in moments of crisis, thus in 2009-2010 (the moment in which each country started to address the consequences of the crisis) and in 2020.Instead, 'normal times' (2014-2019) are characterised by minor adjustments.

H2:
In terms of policy-ideas, we expect governments to have a radically different discourse about their budgetary choices in 2009 compared to 2020, with austerity being dominant in the former, and investment-oriented ideas being dominant in the latter.In addition, we expect the years 2014-2019 to feature a smoothening of austerityled ideas, in which the goals of consolidated public finance and confidence of financial markets are maintained, but the instruments to achieve them start changing.

H3:
In terms of bureaucratic logic, we expect the 2014-2019 period to feature Commission-bureaucrats to gradually promote less austerity-based policies and gradually encourage governments to support their domestic economies and societies, but still within the budgetary space allowed by the framework of the Stability and Growth Pact.
To investigate these hypotheses, we rely on diverse sources of empirical evidence and methodological approaches, derived from the scholarship presented in Table 1.The next section presents the case selection strategy, data and methodology used, while Section 6 conducts the empirical analyses.

Research strategy
Although the economies of European MS are increasingly coordinated through the reinforced EEG, differences among them remain.Choices about the budgetary approach are affected by a combination of government preferences -mostly between responsiveness and responsibility and partisan alternation (Karremans 2021b) -, the structure of the domestic economic system and the situation of national accounts, which limit the set of potential budgetary decisions.
We chose to investigate the three largest economies in the Eurozone, namely Germany, France and Italy, which are also very different countries in several respects, particularly concerning the budgetary choices and the stability of national accounts.As shown in Figure 2, the three countries share a similar trend for all the parameters considered but for public debt, which decreased only in Germany between 2014 and 2019 while in the other two countries remained stable in the same period.However, even in Germany the indebtedness raised with the outbreak of the Covid-19.At general level, both the Great Recession and the pandemic deteriorated all the macroeconomic indicators, although the latter seems to have had a more shattering impact.However, the difference in the starting condition of the three countries is evident and it is particularly interesting when comparing the fiscal and the primary balance of the three countries.While Italy displays always a negative fiscal balance, close to the French one and even worse after the pandemic, its situation considerably improves when looking at primary balance (which is the fiscal balance net of interests), where the country reaches surpluses close to the German ones.This happens because Italy needs to allocate about 30% of its annual budget for payments on public debt (Cavalieri 2023)  constrain the set of budgetary choices Italian governments could make, especially on the side of long-term investments.For these reasons and because of the budgetary decisions made during the period in analysis, we can have an ample view of the budgetary policy of MS within the framework of the EEG.
To test our expectations about incremental and paradigmatic changes, we rely on different sources of data which allow us to explore both the budgetary outcome in terms of expenditure and policy ideas and goals, as well.Specifically, for H1 we use public spending data from Eurostat, categorized into 10 macrocategories and 66 micro-categories of expenditure according to the Classification of the Functions of Government (COFOG) scheme.The dataset provides the amount of general government expenditure that each country in Europe allocates for the different budget categories.Using deflated values of public expenditure, we calculate the year-to-year percentage change (Jones and Baumgartner 2005) of each macro and micro category.
For H2, we rely on the coding of policy-justifications in governments' annual budgetary plans, following the same approach used in recent studies (anonymized for peer review).In presenting their budgetary plans, governments inform the parliament about the criteria by which they strike the balance between expenditures and taxation.Our analysis focusses on these passages, coding as evidence of an 'austerity ideas' those passages in which policies of fiscal consolidation are justified with arguments about low debt and deficits, and coding as 'investment ideas' those passages in which expenditure increases are justified with arguments about economic growth and future fiscal returns.These two categories thus capture the two alternative budgetary approaches that we theorized in Section 2, with the former representing a strategy for financial sustainability based on the short-term reduction of the difference between expenditure and revenue levels, and the latter representing a strategy in which short-term imbalances in the budget are functional towards generating future fiscal returns and should therefore be seen as an investment.
In order to grasp the bureaucratic logics by which the Commission recommends socioeconomic policy interventions to its MS (H3), we direct our focus on the Country-Specific Recommendations (CSRs) that in the framework of the European Semester were sent to MS between 2011 and 2020.We consider these documents to be representative of the bureaucratic exchanges between key actors of EEG (Haas et al. 2020).We complement this analysis with insights from recently published ethnographic studies of the Commission at work (Mérand 2021;Schmidt 2020), which provide information about the logics by which the Commission and its bureaucrats sought to make countries comply with the criteria of the SGP, in times in which European budgetary rules were increasingly unpopular.Our coding of the CSRs follows largely the same logic as that of policy-justifications.We coded 'austerity ideas' those statements in which governments were recommended to reduce debt and deficit in order to ensure financial sustainability, and as 'investment ideas' those statements that recommended expenditure increases in order to generate future growth and fiscal returns.In addition, we gave higher value to statements appearing at the top of the list of recommendations, assuming that these are the important actions that governments are expected to take.A more detailed description of the coding scheme used for policy-justifications and CSRs is available in the Supplementary Material.

Distribution and magnitude of budgetary Changes
To begin with, we measure changes in the budgetary policy of France, Germany, and Italy in order to empirically verify H1.Pooling together budgetary changes across time and budget categories, we first show a comparable shape of the distribution of yearly modifications in the three countries (Figure 3), which confirms the PE model. 4Precisely, the theory predicts that the highest number of observations is concentrated in the central peak of the distribution, where yearly changes are very close to 0. This status quo of tiny adjustments remains in place for a very long time, when medium-sized rational changes (which should be located in the shoulders of the distribution) are almost absent.As a matter of fact, the median change in France, Germany and Italy is 1.07, 1.73 and 0%, respectively (see Table S1 in Supplementary Material).Then, the PE explains that pressure caused by an error-accumulation process or a crisis leads to budget punctuations (visible in the fat tails), which seems precisely what we can see in Figure 3.The tails of each country's distribution gather massive policy cuts -the minimum value is −100% in France and Germany, and −77.13% in Italy -and huge Figure 3. Distribution of year-to-year percentage of budgetary changes in macro and micro categories in France, Germany, andItaly (2007-2021).Note: the right side of the distributions is artificially bounded at +100% (with all changes higher than the threshold recoded as +100%) to prevent the distributions from being extremely right skewed because of very high positive changes (see Table S2 in Supplementary Material).The left side is naturally bounded at −100%, indicating the cancellation of the program.Source: authors' own elaboration, data from Eurostat (available at: https://ec.europa.eu/eurostat/web/government-finance-statistics/data/database; viewed 23/03/2023).
budgetary increases -the maximum value reaches 1319.81 (France), 423.64 (Germany) and 1107.89%(Italy) (see Table S1 in Supplementary Material).This is precisely what the leptokurtic distribution of budgetary changes of the three countries demonstrates: there are several huge budgetary modifications that interrupt a long pattern of tiny adjustments.
The distributions in Figure 3 point out that several huge changes occurred in the period under investigation, which reach and overcome the threshold of 100% change. 5 However, budgetary changes, and punctuations in particular, are not all the same (Cavalieri 2023;John and Bevan 2012).Among the identified extreme modifications, the highest number − 75% in France and Germany, and 60% in Italy -are concentrated in crisis-periods, as illustrated by the categories 'Public health services' (202.28%increase in France and 423.63% increase in Germany in 2020), 'General economic, commercial and labour affairs' (172.87%raise in Germany and 156.62 raise in Italy in 2020), 'Civil defence' and 'Unemployment' (respectively 110.87 and 108.55% increase in Italy in 2020).Instead, in the period 2014-2019, there is a lower number of extreme changes (see Table S2 in Supplementary Material) which affected mostly 'residual' categories.All macro categories have a micro one labelled 'not elsewhere classified (n.e.c.)' and contains expenditure that can't be included in any other micro categories of that group.Following Eurostat codebook, we consider them as 'residual' categories (see Supplementary Material for detailed information).These categories represent a tiny part of the total budget and the related shifts are mostly the result of inconsistent labelling within government departments, and thus do not reflect real changes in expenditure (Cavalieri 2023, 239-247).A very few actual changes, not in residual categories, happened also in non-crisis-period (see Supplementary Material).In fact, it is possible that major changes happen because of other reasons than pressure caused by a crisis (e.g.domestic pressure on that issue, policy-makers' preferences; Jones and Baumgartner 2005;Karremans 2021b).However, we show that while these modifications seem to happen randomly, in crisis-period major changes occur in all three countries and concentrated in 2020.In this regard, the pandemic and the economic crisis show a different pattern, with the former gathering huge changes in the same moment both in France, Germany and Italy and the latter showing differences across countries in the timing of response and categories affected by spending modifications.
We also conjectured that the Great Recession didn't set up a new policy paradigm about the management of public budgeting, while the COVID-19 did.A closer look to annual adjustments informs us that extreme, actual, changes are concentrated in 2020, not during the Great Recession.This can be explained by the considerably different type of crisis and crisis-response (Garzia and Karremans 2021; see also Section 3).The economic crisis demanded severe cuts, a choice that political actors usually try to avoid implementing or try to mask by engaging in across-the-board cuts (Cavalieri, Russo, and Verzichelli 2018) which limit the risk to incurring electoral backlash.Our data (Figure 3 and Table S1 in Supplementary Material) confirm that in the three countries policy makers preferred small and evenly distributed cuts across several budget categories to severe and concentrated, thus more visible, cuts.Instead, the COVID-19 pandemic allowed decisionmakers to considerably boost expenditure in several budget categories (Table S2 in Supplementary Material), as shown by the many huge positive changes in Figure 3.This is in line with the PET, which tells us that negative changes are less punctuated than positive ones (Breunig and Jones 2011).As the main expenditure changes are concentrated in the years of the COVID-19 pandemic and in the previous years we observe only minor policy adjustments, we find corroboration of our first hypothesis about the punctuated nature of paradigm change in EEG.

Policy justifications for national annual budget plans
The analysis of the distribution of budgetary changes provides useful insight on the magnitude of changes and on the occurrence of both tiny modifications and punctuations.However, to prove that only the COVID-19 crisis triggered a change of paradigm in the national budgetary policy of MS within the framework of the EEG, we need to understand the intentions of decision-makers (or what they advertise as such), so to uncover which policy idea drove the allocation of expenditure.To do that, we study the budgetary plan of each country in each year considered, either as annual speech of the finance minister in front of the parliament (France and Germany) or as a budget plan drawn by the government (Italy) (see Table S3 in Supplementary Material).In each of these, the government justifies its decisions about the allocation of expenditure for the next year.Figure 4 shows the overtime trend of three categories of policy-justifications in each country.
The categories contain decisionmakers' statements about their budgetary choices in terms of impact on the country's financial stability and macro-economic   Germany, andItaly (2009-2020).Source: author's own elaboration, data from annual budgetary plans of the three countries.N = 4600 policy justifications (see Supplementary Material, Table S3, S4, S5).
performance.The 'austerity paradigm' contains statements that refer to fiscal consolidation policies, thus to the idea that financial stability should be achieved by reducing public deficits in the short term.The 'investment paradigm', instead, advocates for expenditure increases with the conviction that financial stability can be achieved by investing in the economy in order to generate future fiscal returns.The patterns in the three countries are in line with recent analyses showing that fiscal consolidation gradually disappeared from the policy agendas of national governments from 2014 onwards (Karremans 2021b).The disappearance of fiscal consolidation initiatives is most visible in Germany, arguably because, in addition to having some important changes in the composition of the cabinet, from 2014 onwards it started performing fiscal surpluses (Karremans 2021a;Rixen 2019).
Nonetheless, the presence of this same trend in fiscally troubled countries like France and Italy suggests that it is not only linked to Germany's improved macroeconomic conditions.While between 2010 and 2013, fiscal consolidation seemed to be the only game in town, from 2014 onwards investment policies were increasingly presented as a strategy for economic growth and financial stability.Austerity ideas still remained consistently present but started visibly declining even further from 2017 onwards.By autumn 2020, they seemed to have disappeared from national budgetary discourses regardless of the financial situation of the country.
Beyond the further quantitative decline in the presence of austerity ideas, the budgetary discourse after the outbreak of the pandemic also seems to mark a qualitative change.During the 2014-2019 period, when justifying expenditure increases in financial terms, the argument was that the thresholds prescribed in European budgetary rules would not be exceeded.When, for example, the Italian government justified its expenditure increases of 2019 -which included for instance the very generous social assistance scheme Reddito di Cittadinanza -it asserted that it was committed to economic growth and at the same time at reducing the level of public debt: ' . . . the Government confirms the fundamental objectives of its action: gradually reducing the growth gap with the European average and, at the same time, the debt/GDP ratio (. ..).For the following years, the [government's] stability programme traces a path of public finance that gradually reduces the public deficit to 1.5 percent in 2022 (. ..).The programmatic objectives outlined in the programme are in line with the [European] Stability and Growth Pact's dictation'.(Italian Finance Ministry, DEF, 2019) In the post-pandemic discourse, instead, we observe a different logic of how to combine investments in economic growth and reduction of public debt.When presenting its pandemic response package in spring 2020 -where the government asked the parliament to approve a € 55 billion of deviation for 2020, only a month later the approval of the additional € 25 billion of deviation 6 -the Italian government emphasized that its priorities were to invest in innovation, in the environment and in social sustainability, and that the public debt levels would be brought back to the European average within a decade (Giuseppe Conte and Roberto Gualtieri, Rome, 24 April 2020).The temporal relation between investments in economic growth and the prospects for debt reduction have thus visibly changed.Similarly, across other European governments, from 2020 we observe expenditure increases being presented as investments that will generate economic growth and therefore fiscal returns (Garzia and Karremans 2021, 111).In the speech This gradual change in how the European Commission monitors national budgets has also been documented by studies closely following the operations within the key Commission directorates.In the context of the European Semester, from 2014 onwards the Commission's Directorate-General of Employment and Social Affairs started being increasingly involved in formulating CSRs (Zeitlin and Vanhercke 2018).Particularly under the Juncker Commission, the European bureaucratic machinery started to think more creatively about how more fiscal flexibility could be granted to MS within the clauses of the SGP (Schmidt 2020, 113).In autumn 2014, Juncker's advisors started drafting communications that would be later sent to the Parliament, with the explicit aim to make the best possible use of the flexibility built within the rules of the SGP, without changing these rules (Mérand 2021, 105).From 2014 onwards, thus, the Maastricht paradigm started being interpreted with more flexibility also within the Commission's offices.In line with our third hypothesis, also among EU officials we find a gradual shift of ideas from being austerity-based to becoming investment-oriented.

Conclusion
The comprehension of dynamics of stability and change is fundamental to understanding how policies evolve over time and on which occasions.This is even more important and helpful in public budgeting, where decision-makers must plan future choices a few years in advance.prescriptions (2011-2020).Note: bars for each year do not sum to 100% as CSRs do not address only these two broad categories.Here we show only the paradigm we are interested in for the analysis.N = 621 grammatical sentences recommending a policy-action (see Supplementary Material, Table S6).Source: authors' own elaboration, data from European Country Specific Recommendations.
In this paper, we followed Baumgartner's suggestion and tried to reconcile the three most important scholarships on policy changes (i.e. the PET [Jones and Baumgartner 2005], Hall's, 1993 theory of the different orders of change, and Padgett (1981) Serial Choice Model).Focussing on national public budgets, we used different sources of data to integrate with each other these three main scholarships that have so far remained somewhat unrelated.We studied the budgetary policy of the three largest Eurozone countries (France, Germany, and Italy) that also tend to have contrasting positions during negotiations about European economic strategies.We focussed on the period ranging from the Great Recession to the COVID-19 pandemic and investigated whether the COVID-19 crisis triggered a change of paradigm in national budgetary policies, concerning the choice between austerity-or investmentoriented policies.
After having established that the post-Covid-19 policy-initiatives feature the key relevant characteristics of a paradigm change, we developed three theory-based hypotheses about the mechanisms leading up to a redefinition of the foundational logics by which under EEG governments pursue long-term financial sustainability.Based on three different sources of empirical data, we found confirmation of all three hypotheses, entailing that the post-Covid-19 economic governance approach features the trademarks of what scholars of public policy describe as paradigm change.Firstly, using public expenditure data, we empirically demonstrated that the budget policies of the three countries between 2007 and 2021 were characterised for the most part by small hyperincremental adjustments and that major changes happened only in crisis-moments, especially after the outbreak of the COVID-19 pandemic.Secondly, adding the analysis of policy-justifications for the annual budgets in each country, we proved that the small adjustments of the 2014-2019 period were accompanied by a gradual metamorphosis in the policy ideas behind spending choices.Slowly, the austerity paradigm that distinguished the EEG started to be challenged by new investment-oriented policy ideas that were already in place on the eve of the pandemic and gained momentum when the crisis burst.Thirdly, we traced a similar pattern of ideational change within the European Commission and in the Country-Specific Recommendations adopted by the European Council.
Our findings are relevant for two main reasons.Firstly, they show that multilevel policymaking within EEG is following the same regularities that public policy scholars have identified for national economic policy-making.Interestingly, we find the same gradual change of ideas in national budget plans and in European Country-Specific Recommendations, entailing that the gradual paradigm change has occurred within both supranational and intergovernmental institutions of EEG.This finding sheds a new fresh light on the question of how far European integration has reached, as it shows that supranational institutions and national governments are by now already well integrated and paradigmatic changes are possible only when policy ideas are installed by the EU into national policies.Using Padgett's framework, in the European continent, today, the highest bureaucratic level that must intervene in order to witness a paradigm shift is not anymore within national borders but is the European economic institutions.
Secondly, and relatedly, our analysis shows that the patterns of punctuated policy change apply also to supranational institutions.This insight opens new avenues for future research on the EU, as it indicates that traditional tools and theories of public policy scholarship can go a long way in explaining decision-making dynamics in a polity that is often considered as too unique to be comparable with other systems of governance.

Figure 1 .
Figure 1.Timeline of ideas in the EEG and hypothesized changes of paradigm.Source: authors' own elaboration.Note: the dot-dashed line from 2014-2015 to 2020 indicates the gradual interest into the investment paradigm, which however failed in those years, as explained in section 3.1.

Figure 5 .
Figure 5.European country Specific recommendations:'austerity' and 'investment'-oriented policyprescriptions (2011-2020).Note: bars for each year do not sum to 100% as CSRs do not address only these two broad categories.Here we show only the paradigm we are interested in for the analysis.N = 621 grammatical sentences recommending a policy-action (see Supplementary Material, TableS6).Source: authors' own elaboration, data from European Country Specific Recommendations.

Table 1 .
Classification of policy changes according to the most relevant scholarships.