The EU’s Recovery and Resilience Facility: An Exceptional Borrowing Instrument?

ABSTRACT The Recovery and Resilience Facility (RRF) authorises the European Commission to borrow up to €672.5 billion to aid member states’ economic recovery from COVID-19. Some scholars see such funding as unprecedented. Others see a tight link with earlier borrowing instruments. By comparing the EU’s pandemic facility to eleven such instruments created between 1952 and 2021, this article shows that the RRF is familiar in some respects but novel in others. Viewed through a historical institutionalist lens, the RRF shows signs of layering, but limited evidence of displacement or path dependence. Over the last seven decades, member states have added to earlier instruments, we show through process tracing, but they have rarely been locked into institutional choices. The RRF’s strict time limit is consistent with this finding. The RRF will not become permanent, our analysis suggests, but borrowing is now part of the EU’s toolkit.


Introduction
Faced with an acute economic slowdown when COVID-19 hit, EU member states agreed in June 2020 that the European Commission could borrow up to €672.5 billion in 2018 prices to provide grants and loans to member states to support digital and green investment and bolster resilience (European Council 2020).The result of four days of negotiations, Next Generation EU's Recovery and Resilience Facility (RRF) was hailed by the European Council (2020) as an 'exceptional response' to 'extreme circumstances'.Some scholars concur with this assessment.Vanhercke and Verdun (2022, 204) describe this borrowing instrument as 'unprecedented' in size and scope, while Joan Miró (2022, 38) suggests that the EU has never before 'borrowed to finance expenditures of, let alone transfers to, member states'.Others find ample precedent.For Rehm (2022), the RRF is merely the latest addition to a European financial assistance regime, which began with the European Community's (EC) response to the 1973 oil shock and moved up a gear during the euro crisis.
Comparing the RRF to eleven European borrowing instruments created between 1952 and 2021, this article argues that the EU's pandemic fund is neither entirely new nor a simple replica of what came before.The RRF, it shows, is the latest in a series of supranational borrowing instruments which can be traced back to the establishment of the European Coal and Steel Community (ECSC).Compared to these antecedents, the EU's pandemic fund is sizeable but it is neither the first supranational borrowing instrument nor the most supranational.Nor has the RRF led to a permanent borrowing instrument, thus far, despite calls for further EU bond issuance in response to the war in Ukraine.
Theoretically, this article builds on and challenges historical institutionalist analyses of the EU as a borrower (Gocaj and Meunier 2013; Rehm 2022).A pivotal theory for understanding European integration (Bulmer 2013), historical institutionalism is well suited to analysing the continuities and discontinuities between the RRF and earlier borrowing instruments.Through process tracing, we challenge the idea of the EU's pandemic fund as a punctuated change driven by a critical juncture that will give rise to path dependence; our analysis of legal texts, speeches, press releases and press coverage show that the RRF is better understood as an instance of layering borne of compromise between governments which sought to defend the status quo and those which sought more ambitious change.The EU has rarely been locked into one set of institutional choices over borrowing instruments and such flexibility seems likely to prevail in relation to the RRF, we conclude.
The remainder of this article is divided into six sections.The first explores how scholars have sought to theorise the EU as a borrower, focusing in particular on historical institutionalist accounts.The second considers the RRF as an instance of layering on top of earlier supranational borrowing instruments.The third considers the institutional flexibility shown by member states in switching between intergovernmental and supranational modes of borrowing.The fourth puts the time limit employed under the RRF in historical context and the fifth looks for signs of path dependence in how EU borrowing instruments responded to the war in Ukraine.The final section summarises our key findings and considers the EU's future as a borrower.

Theorising the EU as a borrower
Our focus in this article is on cases where EU member states, or a sizeable subset thereof, empower pan-European bodies to raise funds on international capital markets to provide grants, loans or guarantees to public or private actors.As such, we do not consider EU financial institutions or instruments which offer such support without recourse to borrowing.A case in point is the Neighbourhood, Development and International Cooperation Instrument, which is financed via the EU budget.Nor do we consider central banking operations such as the European Central Bank's emergency liquidity assistance.Our definition also excludes the Medium-Term Financial Assistance Facility, which was funded via credits from member states rather than borrowing.This leaves us with eleven borrowing instruments created between 1952 and 2021 (see Table A1 in the Annex to this paper).
Borrowing is a core state power which has traditionally been seen as off limits to the EU because of its association with the 'core functions of sovereign government' (Genschel  and Jachtenfuchs 2014, 2).Although there is a growing literature on borrowing instruments created during the euro crisis and COVID-19 pandemic, there have been few attempts to explain the EU's evolution as a borrower over the longer-term.Within the literature on this topic, historical institutionalism is the most commonly employed analytic approach, perhaps because it is drawn to the kinds of crises that encourage borrowing (Verdun 2015, 221).This theoretical choice is also consistent with historical institutionalism's increasing importance in the study of European integration over the last two decades (Christiansen and Verdun 2020).Other schools of new institutionalism, from rational choice institutionalism to discursive institutionalism, have also increased in popularity.But historical institutionalism's status as a baseline theory of European integration is recognised even by its critics (Moravcsik 2018).
Historical institutionalism gives rise to several hypotheses.This article tests two, which are particularly prominent in existing studies regarding the EU as a borrower. 1The first, which follows Gocaj and Meunier (2013), posits that critical junctures will give rise to novel institutional choices which subsequently lock-in specific patterns of borrowing.The second, which follows Rehm (2022), hypothesises that layering and displacement will drive the evolution of EU borrowing instruments.
Critical junctures occur when sudden events privilege one radical policy choice over others (Capoccia 2016, 101).Path dependence captures the difficulty of changing course once such choices have been made, even when they produce undesirable outcomes (Fioretos, Falleti, and Sheingate 2016, 11).The ability of financial markets to make or break European policies underlies Gocaj and Meunier's conception of critical junctures.Where financial markets lose confidence in 'piecemeal, timid measures', European leaders will face instantaneous pressure to consider more radical institutional choices, the authors argue (Gocaj and Meunier 2013, 243).But, having made their decision, EU policymakers will find it difficult to revisit this earlier array of choices, leading to suboptimal outcomes.A key reason for such path dependence, Gocaj and Meunier (2013, 248) posit, is the 'sunk political costs' of chosen borrowing instruments.This is one explanation for why temporary pan-European public financial institutions become permanent, they contend (Gocaj  and Meunier 2013, 243).
Historical institutionalism is interested in incremental as well as radical change, with the latter traditionally explained through some combination of displacement, layering, drift and conversion (Mahoney and Thelen 2010).Displacement occurs when new institutions override existing ones.Layering sees new institutions built on top of earlier ones.Drift describes a situation in which rules are undermined or rendered obsolete by changing contexts.Conversion occurs when existing institutions are repurposed.Rehm (2022)  offers the most sophisticated treatment to date of why pan-European public financial institutions are likely to experience incremental change despite the recurrence of critical junctures.Drift is unlikely to occur in this policy domain, he argues, since periodic economic crises will either reinforce existing borrowing instruments or create permissive conditions for change.Member states tend to prefer layering since existing instruments have typically shown their utility in past crises and are rarely worth erasing (Rehm 2022,  15).Displacement is likely, however, given the potential for financial or legal constraints to drive member states to create new borrowing instruments even though such choices tend to be unintentional (Rehm 2022, 5).
Verdun, for one (Verdun 2015, 228), questions the explanatory power of displacement in relation to pan-European public financial institutions.Instead, she sees member states as engaged in deliberate acts of copying, whereby they borrow from the design of earlier pan-European public financial institutions or EU institutions more generally. 2n what follows, we look for evidence of layering, displacement and path dependence in the RRF by comparing it with the ten other borrowing instruments created by the EU and EC between 1952 and 2021.We begin by comparing the RRF to earlier supranational instruments before considering the pandemic facility's relationship with intergovernmental modes of borrowing.We then examine the RRF's time limits and look for evidence for path dependence after the instrument's establishment.

The Recovery and Resilience Facility as an instance of layering
For Buti and Fabbrini (2023, 667), the economic effects of COVID-19 gave rise to a 'governance discontinuity' in which member states chose the 'constrained supranationalism' of the RRF over the 'unconstrained intergovernmentalism' of the ESM.Although this argument is not couched in historical institutionalist terms, it captures well the idea of the pandemic as a critical juncture which favoured modes of borrowing that were previously beyond political reach.As a public health emergency which presented decision-makers with a wide-range of policy choices during a relatively short period of time, the COVID-19 crisis exemplified the idea of a critical juncture (Capoccia and Kelemen 2007).Yet, this does not mean that this period of flux produced punctuated change.On the contrary, our analysis of earlier supranational borrowing instruments suggests that the policy choices made during the pandemic were continuous rather than discontinuous and consistent with the practice of layering.
In comparing the RRF to its antecedents, we are mindful that the terms intergovernmental and supranational are contested.To most scholars of European integration, the Council of the EU is the archetypal intergovernmental institution, but to some its intensive working methods have a supranational quality (Beyers and Dierickx 1997).For the avoidance of doubt, we define intergovernmental institutions as bodies which are directed by, and accountable to, national governments.Supranational institutions are defined as bodies whose leadership is appointed by national governments but thereafter operationally independent of them.These definitions refer, of course, to ideal types.
The intergovernmental and supranational character of EU institutions is determined not only through hiring and firing rules but also by policy-making autonomy.Applied to EU borrowing, we think of highly supranational borrowing instruments as giving the Commission a high degree of autonomy over the disbursement of grants, loans and guarantees, the conditions attached to such assistance, and programme monitoring.Where member states have a say on such issues, the means through which they reach agreement is a key consideration.For Weiler (1981), unanimity is an expression of intergovernmentalism, while qualified majority voting evokes supranationalism.
Seen in these terms, the RRF counts as a sizeable supranational EU borrowing instrument which is nevertheless subject to significant intergovernmental oversight.Under Next Generation EU, the Commission is authorised to borrow up to €672.5 billion in 2018 prices (€723.8 in current prices as of June 2022) to finance up to €360 billion in loans and €312.5 billion in grants under the RRF. 3 As D'Erman and Verdun (2022, 3) note, the European Commission has never before been authorised to borrow on this scale, much less to do so for the purpose of providing grants as well as loans.Each member state seeking financial assistance from the RRF is required to prepare a Recovery and Resilience Plan, which the Commission assesses. 4Where this assessment is favourable, the Commission is responsible for proposing the award of all grants and loans requested and for disbursing payments based on the fulfilment of agreed milestones and targets.While these provisions give the Commission a major role in the RRF, the Economic and Financial Affairs Council (ECOFIN) must authorise assessment and award decisions by means of a Council implementing decision adopted by a qualified majority vote. 5The Commission has more autonomy over disbursement.However, even here the Commission must first seek the opinion of the Economic and Financial Committee, an advisory body of ECOFIN consisting of top national treasury officials. 6Moreover, in exceptional circumstances, a member state can refer its concerns over 'serious deviations' from milestones and targets to the President of the European Council.In such cases, the Commission will not proceed with disbursement until the matter has been 'exhaustively discussed' by heads of state or government, with a delay of up to three months. 7n historical institutionalist terms, the RRF is layered on supranational borrowing instruments which date back to the earliest days of the European Communities.Such layering does not equate to copying, it should be stressed, since these borrowing instruments differ in the degree of autonomy they delegate to supranational institutions.Established in 1952, the ECSC Loan Facility was a highly supranational instrument which entitled the High Authority to borrow without predetermined limits to provide loans to member states. 8The disbursement of loans to increase production, lower production costs or promote the marketing of products required a unanimous vote in the Special Council, ensuring intergovernmental oversight of the ECSC Loan Facility. 9But no such approval was required for High Authority loans to enterprises for investment and programme monitoring. 10The High Authority, meanwhile, enjoyed complete autonomy over the conditions attached to loans and programme monitoring.
The RRF is closer in institutional design to the Community Loan Mechanism, on which several other supranational borrowing instruments were also layered (Rehm 2022).Established in 1975 in response to the international economic crisis, the Community Loan Mechanism authorised the Community to raise funds on capital markets to provide up to $3 billion in loans to member states facing serious balance of payments difficulties due to high petroleum prices. 11Although the Commission was not formally authorised to borrow on behalf of the European Community, it did so in practice, before this role was codified by the Council of Ministers in 1981. 12Decisions on the granting of financial assistance, and conditions attached to Community loans, were taken by ECOFIN by unanimity. 13The Commission was not given a formal role in programme monitoring until 1981.In 1988, the Medium-Term Financial Assistance Facility and Community Loan Mechanism were merged into the Balance of Payments Assistance Facility, which was subject to a qualified majority vote for decisions on the granting of financial assistance. 14ayering occurs, Thelen and Mahoney (2010: 16-17) posit, where defenders of the status quo are powerful enough to resist radical change but not to stop those seeking the creation of additional rules and institutions on top of existing ones.Such contestation is not discernible in the creation of all supranational borrowing instruments; the Euratom Loan Facility and Macro-Financial Assistance, for example, were the product of a broad consensus among member states over the need for a coordinated response to the energy crisis of the 1970s and the fall of Communis t rule in Central and Eastern Europe in 1989.The establishment of a supranational borrowing instrument during the euro crisis was altogether more divisive.
During negotiations over the Maastricht Treaty, Germany, the Netherlands and the UK strongly opposed the creation of a fiscal capacity for the embryonic economic and monetary union (EMU) (Dyson and Featherstone 1999, 732).In consequence, euro area members were prohibited from seeking support under the Balance of Payments Assistance Facility, with the result that Latvia, Hungary and Romania received swift assistance when the global financial crisis unleashed destabilising macroeconomic imbalances, but Greece did not.In February 2010, as risk premia on Greek debt rose, euro area heads of state or government agreed to 'determined and co-ordinated action, if needed, to safeguard financial stability in the euro area' but this meant adding new borrowing instruments rather than reopening old ones (Council of the European Union 2010).Three months later, EU member states agreed to create the Greek Loan Facility, an ad hoc arrangement which relied on the pooling of bilateral loans rather than the issuance of common debt. 15This move was quickly followed by the establishment of the supranational European Financial Stabilisation Mechanism (EFSM), through which the Commission was authorised to borrow up to €60 billion on behalf of the EU to finance loans to any member state experiencing a 'severe economic or financial disturbance caused by exceptional occurrences beyond its control'. 16The German government reluctantly agreed to the establishment of this instrument (Howarth and Schild 2021, 215), while insisting that the bulk of financial assistance be provided by intergovernmental borrowing instruments in the form of the European Financial Stability Facility (EFSF) and European Stability Mechanism (ESM) (see next section).
The creation of the RRF also pitted member states which favoured the status quo against those who sought more radical policy solutions.In March 2020, the heads of state or government of Belgium, France, Italy, Luxembourg, Spain, Portugal, Greece, Slovenia and Ireland signed an open letter to European Council President Charles Michel calling for the creation of a 'common debt instrument issued by a European institution . . . to counter the damages caused by this pandemic' (Michalopoulos 2020).Dutch Prime Minister Mark Rutte quickly declared his opposition to coronabonds, as the proposal came to be known, and insisted that he could not 'foresee any circumstance' in which he would change his mind (Rios et al. 2020).He was backed by the leaders of Austria, Finland and Sweden and German Chancellor Angela Merkel, who acknowledged the need for solidarity but ruled out the idea of 'common debt' (Reuters Staff 2020).
Defenders of the status quo defeated the coronabonds proposal but they remained under intense public pressure to act amid growing media criticism of EU member states for failing to show solidarity as the pandemic worsened (McGee 2020).In May 2020, ECOFIN responded by establishing Support to mitigate Unemployment Risks in an Emergency (SURE), which authorised the European Commission to borrow up to €100 billion primarily to help member states with the costs of work furlough schemes. 17ignificant though this step was, discussions about a more ambitious policy response were already underway by this point, as the outlook for the EU economy steadily worsened (European Commission 2020).
A proposal by French President Emmanuel Macron and German Chancellor Angela Merkel in May 2020 to create a €500 billion borrowing instrument backed by the EU budget paved the way for the RRF. 18The Frugal Four of Austria, Finland, the Netherlands and Sweden initially pushed back against this proposal, but without Germany they found it difficult to prevent the layering of new borrowing instruments on top of existing ones (Rankin and Oltermann 2020).Rutte drove the hardest bargain at the four-day meeting of the European Council in July 2020, but he threw his support behind the RRF after other leaders agreed to use the majority of funds raised for loans rather than grants and to install an emergency brake mechanism to be activated in the event that a member state seriously deviated from the milestones and targets contained in its Recovery and Resilience Plan (De la Porte and Dagnis Jensen 2021).An increase in the Netherlands' rebate under the Multiannual Financial Framework further sweetened this deal.
EU member states could have pursued a far bolder response to the COVID-19 pandemic by issuing common EU debt backed by joint and several guarantees.Conversely, they could also have repurposed the Balance of Payments Assistance Facility and the EFSM to help member states to cope with the economic effects of COVID-19.Instead, they added a new supranational borrowing instrument to existing ones as a compromise between member states which preferred the status quo and those that advocated for more radical change.Faced with a critical juncture, in other words, they chose layering over both more conservative and radical alternatives.

Institutional flexibility in EU borrowing instruments
That the RRF is merely the latest in a series of supranational borrowing instruments lends weight to historical institutionalist claims of layering (Rehm 2022).But we see limited evidence of path dependence in the EU's evolution as a borrower (Gocaj and Meunier's 2013), given member states' willingness to switch back and forth between intergovernmental and supranational borrowing instruments.This institutional flexibility was apparent during the COVID-19 pandemic, which saw member states consider both types of borrowing instrument before creating the RRF.
Member states' decision to establish the European Investment Bank (EIB) in 1958 rather than extending the High Authority's borrowing powers to the European Commission offers the earliest example of such flexibility.The EIB is sometimes referred to as a supranational body, but it was created as a strongly intergovernmental institution and remains so to this day (Clifton, Díaz-Fuentes, and Gómez 2018, 733).The Bank's Board of Governors, which determines borrowing and lending policy, consists of one minister from each EU member state.Each member state also appoints a member to the Board of Directors which meets monthly and makes decisions on borrowing and lending.Although the treaty stipulates that members of the Board of Directors be chosen from 'among persons of indisputable independence and competence' and 'be responsible only to the Bank', directors mostly come from national finance ministries and retain such positions alongside their work for the EIB. 19he EIB's institutional design was the subject of intense debate in the Messina Conference (1955) and Val Duchesse Conference (1956), which together paved the way for the Treaty of Rome.Italy wanted an intergovernmental development fund which could help poorer regions cope with the competitive pressures of the European Common Market, but West Germany wanted a supranational investment bank which could borrow at lower interest rates to support profitable projects (Coppolaro 2010).Despite these differences, neither side seriously considered allowing the Commission to borrow, despite their willingness to allow the High Authority to do so.Viewed in historical institutionalist terms, there was no obvious path dependency here: sunk costs mattered little and there was no legal or fiscal imperative, thus ruling the European Commission out as a borrower.
In the end, West Germany and Italy each secured concessions over the EIB's institutional design.West Germany largely got its way on the Bank's functions, which focused on the development of the Common Market and empowered the EIB to provide loans and guarantees rather than grants.Italy had wanted ECOFIN to determine the general direction of EIB lending (Coppolaro, 2010).It achieved a close approximation of this aim, in as much as this role was entrusted to a Governing Board made up of the same member state ministers who participate in ECOFIN.
In contrast to the supranational EFSM, the EFSF and ESM are strongly intergovernmental.Neither are EU bodies and both are accountable to euro area members, which serve as their shareholders.The EFSF General Meeting of Shareholders and the ESM Board of Governors consist of one minister per member state, while high-ranking national officials make up their Boards of Directors. 20Representatives of the Commission and the ECB attend these decision-making bodies but only as observers.Decisions on whether to approve financial assistance are taken by EFSF shareholders, who decide by unanimity. 21The ESM takes the same decisions by mutual agreement, which in practice means a unanimous vote, although the ESM Treaty allows these decisions to be taken by qualified majority vote in exceptional circumstances. 22or Gocaj and Meunier (2013), the similarity between the EFSF and the ESM is evidence of path dependence.However, member states' willingness to create a supranational instrument, the EFSM, during the euro crisis speaks to their institutional flexibility despite pressure from financial markets.That the EFSM relied on a controversial reading of Article 122 of the Treaty on the Functioning of the European Union, allowing the EU to provide financial assistance to a member state facing difficulties 'caused by natural disasters or exceptional occurrences beyond its control', lends weight to the view that displacement drove the creation of the intergovernmental EFSF and ESM.And yet, at odds with this historical institutionalist interpretation, is the fact that the EFSF and ESM both faced legal challenges which were by no means unforeseen (European Stability Mechanism 2019, 74 and 236).
The EFSM's total borrowing capacity was limited to €60 billion, a figure determined by the 'margin available under the own resources ceiling for payment appropriations', leaving it with nowhere near the resources required to reassure financial markets that the euro crisis was under control. 23While this constraint helps to understand the urgency of creating the €440 billion EFSF and €500 billion ESM, the historical institutionalist concept of displacement does not explain why the European Commission was not given a more prominent role in the governance of the EFSF and the ESM.Although both instruments were underpinned by international law treaties, the same is true of the European Bank for Reconstruction and Development, which counts the EU as a shareholder and includes Commissioners and Commission officials in its senior decision-making bodies. 24A more plausible explanation for the EFSF and ESM's institutional design is that member states were determined to maintain a tight grip on euro crisis funds (Schimmelfennig 2015).Unusually for a pan-European public financial institution, four member states (Germany, the Netherlands, Finland and Estonia) can agree to ESM capital raising only with the approval of national parliaments (Howarth and Spendzharova 2019, 903).
The sunk costs of creating the ESM would, from a historical institutionalist perspective, have made it the obvious borrowing instrument to turn to when the COVID-19 pandemic threatened further fiscal difficulties for the euro area.Indeed, this is precisely what euro area finance ministers envisaged in April 2020 when they agreed to establish the Pandemic Crisis Support, an ESM precautionary credit line to help finance certain costs linked to the pandemic (Eurogroup 2020).Given the ESM's readiness, Germany argued against the creation of new borrowing instruments in the early months of the pandemic (Howarth and Quaglia 2021, 1562; Tesche 2022, 485).But Italian Prime Minister Giuseppe Conte was uneasy about accessing ESM funding, in part, because this borrowing instrument had been the subject of sustained criticisms from Matteo Salvini, a Eurosceptic former Deputy Prime Minister who had portrayed the euro crisis fund as an attack on national sovereignty (Fonte and Jones 2019).
The speed with which other member states swung behind the creation of the supranational RRF was not an act of displacement.The ESM faced no immediate financial or legal impediments to supporting member states during the pandemic.Rather, the RRF's creation provides further evidence of institutional flexibility when it comes to the design of borrowing instruments.

Time limits on EU borrowing instruments
EU member states' willingness to switch back and forth between intergovernmental and supranational modes of borrowing challenges historical institutionalist claims of path dependence surrounding the creation of the RRF.So too do the relatively strict time limits to which the RRF is subject.A Council regulation on the EU's own resources adopted in December 2020 makes it clear that Next Generation EU is an 'extraordinary and temporary' measure established for the 'sole purpose of addressing the consequences of the COVID-19 crisis'. 25In managing the RRF, the Commission can undertake 'no new net borrowing' after 2026. 26It is also required to begin repaying the principal on funds borrowed before the end of the 2021-2027 period and to repay all liabilities incurred by the last day of 2058. 27Although the Council regulation underpinning the EU's own resources could be revised to permit borrowing over a longer time period, such a move would have been difficult to justify without further lengthy lockdowns.
The regulation underpinning the RRF makes clear that all grants and loans financed through such borrowing be made available to member states no later than 31 December 2023. 28Although the Commission can propose amendments to the facility 'where appropriate', the regulation's requirement that an independent evaluation report on the RRF be published by 20 February 2024 assumes that the facility will have stopped providing grants and loans by well before this point. 29The RRF regulation also calls for an independent ex-post evaluation report by 31 December 2028, the assumption being that the EU's pandemic funds will have long since been allocated and spent. 30Taken together, such time limits constrain the Commission's powers as a borrower in new ways, further challenging the presence of path dependence in the design of the RRF.
Borrowing instruments created before the 1990s came without time limits.A partial exception was the ECSC Loan Facility, which was indirectly bound by the founding member states' decision to conclude the Treaty Establishing the European Coal and Steel Community for a period of fifty years. 31The result is that the ECSC expired on 23 July 2002.And yet, even then, the ECSC Loan Facility had a financial afterlife following the European Council's decision to transfer the net assets of the European Community, which amounted to nearly €2 billion, to a new Research Fund for Coal and Steel to be managed by the European Commission (2022). 32This fund also guaranteed outstanding ECSC loans, which retained their triple A rating as of 2022 (Moody's 2022).
The European Investment Bank faced no such treaty-based time limits on its borrowing activities, the Treaty Establishing the European Economic Community having been concluded for an indefinite period; nor did the Community Loan Mechanism, despite member states' insistence that their obligations in relation to the Medium-Term Financial Assistance Facility be limited to a specific period.The Medium-Term Financial Assistance was initially agreed upon for the period 1972 to 1976, on the understanding that it would then be automatically renewed every five years once the European Community had reached the second stage of the Werner Plan for Economic and Monetary Union. 33This plan was for all intents and purposes in abeyance by 1975, in response to which member states agreed to extend the Medium-Term Financial Assistance to 1980.Two-year extensions followed biennially between 1980 and 1986 reflecting member state reluctance about committing national resources to an instrument, which they accepted was necessary nonetheless.
The Community Loan Mechanism, in contrast, was established in 1975 for an indefinite period, albeit subject to a (US) $3 billion ceiling on the total amount of loans that it could grant.In 1982, ECOFIN agreed not only to raise this ceiling to ECU 6 billion but also to revisit the functioning of the instrument within five years to ensure, inter alia, that this ceiling met the needs which led to the creation of the mechanism.This led to the loan ceiling being increased to ECU 8 billion in 1986 before member states decided in 1988 to merge the Medium-Term Financial Assistance Facility and Community Loan Mechanism into the Balance of Payments Assistance Facility.Like the Community Loan Mechanism, this Facility was established for an indefinite period, albeit subject to fluctuating loan ceilings. 34ime limits for EU borrowing instruments became commonplace in the 1990s, but the RRF by no means copied them.Regulations underpinning Macro-Financial Assistance, a borrowing instrument established to support third countries facing balance of payments crises, typically provide loans over two and a half years, while allowing the Council to extend this 'availability period' based on a Commission proposal.The multiple rounds of Macro-Financial Assistance provided to Ukraine confirm that the borrowing instrument, while not indefinite in its use, was recurring.Such support was scaled up after the Russian invasion in February 2022, with the EU agreeing to provide Ukraine with €18 billion in Macro-Financial Assistance + funding in November 2022.
Both the EFSM and EFSF were intended to be temporary instruments.The EFSF's Articles of Incorporation made clear that no new financing programmes or loan agreements could be agreed upon after 30 June 2013.The EFSM had no such sunset clause, but the regulation underpinning the mechanism required reviews every six months on whether 'the exceptional occurrences that justify the adoption of this Regulation' continued.In December 2010, the European Council agreed to create the ESM as a permanent borrowing instrument and, in so doing, decided that both the EFSM and EFSF would be wound down in June 2013.The EFSF engaged in no new lending after this date, although the facility will not be wound down until its obligations are fully repaid (Howarth and  Spendzharova 2019, 895).The EFSM's time limits proved less constraining, however, as evidenced by EU finance ministers' decision to approve a €7.16 billion bridging loan to Greece in June 2015 under the EFSM. 35This decision proved especially controversial in the UK, which secured a commitment in Prime Minister David Cameron's so-called new settlement with the EU that the EFSM would never be used again.This settlement never entered into force because of the UK's referendum vote to leave the EU in June 2016, with the result that the EFSM remained 'in place' at the time of writing in May 2023.
As with the RRF, SURE was intended to be a temporary instrument but its time limits were less strict.The Council prohibited SURE's use 'once the COVID-19 emergency has passed' and set 31 December 2022 as the latest date on which the Council could make financial assistance available. 36The Council was empowered to extend SURE in six month increments, where the Commission concluded that a severe economic disturbance caused by the COVID-19 pandemic persisted.However, no extension was agreed.

The RRF and path dependence
In a survey of 111 leading experts on Economic and Monetary Union conducted in 2020, two thirds of respondents expected the EU's COVID-19 response to pave the way for a permanent EU borrowing capability (Begg 2023, cited in; D'Erman and Verdun 2022, 4).The first major test of path dependence occurred in February 2022 when Russia invaded Ukraine, exposing the EU to a profound geopolitical shock as well as a sudden spike in energy prices.Had the RRF constrained future institutional choices, we would have expected member states to extend or replicate this supranational borrowing instrument.However, this had not happened to any significant degree at the time of writing in May 2023 despite the best efforts of some political actors.
On 28 February 2022, EU foreign ministers agreed to provide €500 million in assistance to Ukraine under the European Peace Facility.Although this facility was not new, it was the first time it had been used to provide lethal equipment to a third country.However significant this step may have been in the EU's evolution as a military actor, it made no difference to borrowing.Although French President Emmanuel Macron floated the idea of financing military expenditure through EU borrowing at an informal gathering of national leaders in Versailles in March 2022, Dutch and German opposition to this idea meant that no mention was made of financing in the meeting's communique (Smith-Meyer 2022).By April 2023, the EU had provided €4.6 billion in aid to Ukraine under the European Peace Facility, but this assistance was exclusively financed through 'off-budget' contributions from member states.The war's economic impact on the EU, though it was profound, has also had a limited institutional effect thus far.By Autumn 2022, higher food, energy and commodity prices had pushed the rate of consumer price inflation in the EU to 9.3%, adding to public concerns over a cost-of-living crisis (European Commission 2022).When the German parliament approved a €200 billion energy relief plan in October 2022, two European Commissioners published an op-ed calling for a 'European budgetary response' modelled on the EU's reaction to COVID-19 (Breton and Gentiloni 2022).A borrowing instrument inspired by SURE 'to help Europeans and industrial ecosystems' could provide a first step, they suggested.German Finance Minister Christian Lindner quickly poured cold water on the idea, arguing that: 'More far-reaching proposals based on the SURE program cannot be justified at this point in time' (Tamma and Von der 2022).
A modest change to the RRF was enacted in February 2023 when the Council of Ministers agreed that member states could add a chapter to Recovery and Resilience Plans dedicated to the so-called REPowerEU initiative.Adopted ten months earlier, REPowerEU sought to end EU dependence on Russian fossil fuels by diversifying energy supplies and investing in Europe's clean energy transition.The Council of Ministers presented the RRF as 'the main source of financing for REPowerEU', but its claims that €255 billion would be provided for this purpose glossed over the fact that the pandemic fund's borrowing ceiling and time limits would remain unchanged; instead, member states would be allowed to direct unused funding towards the EU's new energy objectives in the remaining time available (Council of Ministers 2023).Taking account of the difference between EU and member state borrowing costs, the Centre for European Reform estimates that the RRF will contribute around €24 billion to REPowerEU (Cornago and Springford 2023).Seen in historical institutionalist terms, the war in Ukraine has produced a minor modification in the EU's pandemic fund rather than enduring institutional change.
To Fabbrini (2023), a more meaningful instance of path dependence associated with the EU's COVID-19 response can be seen in EU economic aid for Ukraine.In February 2022, the EU authorised up to €1.2 billion in loans for Ukraine through a new emergency Macro-Financial Assistance (MFA) programme.A further €14 billion in loans followed, before the European Parliament and Council agreed in December 2022 to offer Ukraine up to €18 billion in loans under a new MFA+ instrument.The MFA+ 'follows in the footsteps' of the RRF, Fabbrini (2023, 10) argues, by requiring Ukraine to sign a Memorandum of Understanding and to respect the principles of democracy and the rule of law as a condition for funding.Significant though the scale of EU funding for Ukraine has been since the Russian invasion, we would question such claims of path dependence.MFA loans were subject to similar stipulations long before the RRF was established, challenging the idea that the latter locked in the former.

Conclusion
In December 2021, Paolo Gentiloni was asked by a Member of the European Parliament whether the RRF could become a permanent borrowing instrument.The EU's decision to allow the European Commission to borrow up to €672.5 billion to help member states' economic recovery from the coronavirus 'was conceived as a one-off', the Commissioner for Economy insisted (Strupczewski 2021).However, he also noted that this did not mean that 'this kind of methodology to raise common resources for a common goal could never be used again in the EU'(Strupczewksi 2021).Gentiloni's intervention captures well the image of the EU's pandemic fund as being both unprecedented and a model for the Union's response to future crises.Our article gives pause for thought about such claims.
By comparing twelve European borrowing instruments created between 1952 and 2021, we challenge the idea that the RRF is an unprecedented institutional choice which is likely to constrain future developments in this domain.Although the European Commission has never before been authorised to borrow on such a scale, our findings show that the RRF is merely the latest in a series of supranational borrowing instruments which date back to the ECSC Loan Facility.This does not mean, however, that the EU has or will be locked into supranational borrowing instruments.The establishment of the RRF exemplifies the EU's tendency to switch between intergovernmental and supranational modes of borrowing, our findings demonstrate, as does the limited impetus for prolonging or replicating the EU's pandemic fund in response to the war in Ukraine.
Seen in historical institutionalist terms, our findings chime with Rehm (2022), who emphasises the importance of layering in relation to EU borrowing instruments, while challenging claims of path dependence (Gocaj and Meunier 2013).Although it bore all the hallmarks of a critical juncture, the pandemic exposed deep divisions between member states which championed bold steps and those which favoured the status quo.In keeping with Mahoney and Thelen (2010), the latter saw off radical plans for coronabonds while agreeing with the former to build the RRF on top of existing borrowing instruments.The relatively strict time limits placed on the RRF as part of this compromise have thus far proved impermeable, we have shown, with REPowerEU counting as a minor revision to the EU's pandemic fund rather than evidence of path dependence.
The apparent absence of path dependence does not mean that EU borrowing will fizzle out as the pandemic recedes.Among the questions not considered in this article is whether large-scale EU borrowing will move the Union any closer to having a safe asset and what this might mean for future borrowing decisions.Nor do we consider what impact the RRF will have on the EU's own resources given the need to repay EU pandemic borrowing by 2058 at the latest.Even if such effects materialise, we expect future borrowing to be characterised by a similar degree of institutional flexibility.While EU member states may eventually choose to layer another supranational borrowing instrument on top of existing ones, our findings suggest that they will not be bound to do so.