The Law and Norms of the European Central Bank as Sovereign Lender of Last Resort: Crystallising Endogenous Authority

Abstract

Authority of the European Central Bank (the Bank) over its operational norms in the eyes of market actors – Exogenous and endogenous authority and legitimacy – The reconciliation by the Bank and the Court of Justice of the EU (the Court) of the pre-existing norm and political-economic reality with Article 123 TFEU – Sovereign lender of last resort – Eurozone Crisis – Outright Monetary Transactions (OMT) – Public Sector Purchase Programme (PSPP) – Pandemic Emergency Purchase Programme (PEPP)

Introduction

In a 2015 EuConst editorial, the editors captured the post-Gauweiler landscape with the observation that ‘it seems that so far the [European Central] Bank manages to successfully speak two languages to its different audiences: lawyers and bondholders’.¹ The Bank has likely prevented, at least for now, a sovereign debt crisis resulting from the requirement to place the economies of the Eurozone into a medically induced coma to mitigate the effects of the coronavirus pandemic. The capacity of the Eurozone to withstand this shock to its constitutional framework has been enriched by the fact that the Eurozone enjoys a sovereign lender of last resort if not in original intention, then in accomplished fact. Yet, this significant turn of the screw in the governance of the Eurozone was achieved absent any meaningful change to the treaties.² This simultaneous bilingualism emerged because:

The Court is satisfied with the current design of the OMT programme, and with the focus of the Bank on the practical limits of the programme, in view of the prohibition of monetary financing. Bondholders, by contrast, are satisfied with the ‘whatever it takes’ message, combined with the absence of an ex ante quantitative limit to the size of the bond purchases. In fact, the room for manoeuvre which the Court of Justice grants the Bank in Gauweiler will only strengthen the perception of the bondholders that the Bank indeed can and will act as a lender of last resort.³

Prior to the judgment, the manner in which the capacity of the Bank to act as sovereign lender of last resort to Eurozone member states was manifested (a press conference and, subsequently, a press release) was legally unsettled and unsettling.⁴ But the intervention of the Bank succeeded and the Court itself sent a clear, strong signal to market actors that it would not undermine the sovereign lender of last resort armamentarium appropriated by the Bank to safeguard the euro.⁵ Prescient of subsequent empirical testing,⁶ the Court did not accept the premise that the spreads in sovereign yields were solely attributable to the economic fundamentals of the member states.⁷ Instead, the Court proceeded on the basis of the Bank’s diagnosis of a destabilising self-reinforcing disequilibrium.⁸ By implication, the Court proved itself attuned to the economic reality that the Bank required authority in the eyes of market actors to quash any doubt and, by extension, any test by the market of its program. Indeed, so successful was the Bank’s intervention that the policy of OMT never required formal activation. So clear was the signal of the Court that the EuConst editorial went on to predict that the then proposed (third) sovereign bond-buying programme, called the Public Sector Purchase Programme (PSPP), providing for quantitative easing would survive its inevitable legal challenge before the Court, which it did in the 2018 Weiss decision.⁹

Together these cleared the way for the announcement of the Bank, on 18 March 2020, of its Pandemic Emergency Purchase Programme (PEPP), in effect a blend of OMT and quantitative easing. It consists of a €750 billion corporate and sovereign bond purchasing programme, and relaxed the self-imposed restrictions on government bonds for the duration of the crisis.¹⁰ Notably, the Bank has suspended its promise to purchase no more than one third of any member’s available bonds and to purchase the securities in proportion to the member’s economy.¹¹

On 5 May 2020, the German Federal Constitutional Court delivered judgment in Weiss through which it sought to undermine the Bank’s ability to operationalise quantitative easing and, by extension, the Pandemic Emergency Purchase Programme.¹² It did so by attacking the European Court of Justice’s assessment of the proportionality of the Public Sector Purchase Programme. The approach of the European Court, claimed the German Federal Constitutional Court, ‘manifestly fails to give consideration to the importance and scope of the principle of proportionality’ and was not ‘tenable from a methodological perspective’, meaning that the European Court ‘manifestly exceeds the mandate conferred upon it’ the consequence of which is that ‘the CJEU Judgment itself constitutes an ultra vires act and thus has no binding effect [in Germany]’.¹³ Although the Bank satisfied the European Court of Justice of the legality of its actions (and should leave judicial disputes to judges), the decision raised the prospect that the Bundesbank might withdraw cooperation from the Pandemic Emergency Purchase Programme or future policy measures.

In effect, the German Federal Constitutional Court has queried the input legitimacy of the Bank’s decision-making.¹⁴ Likewise, in their assessment of the unconventional measures required to save the euro, constitutional scholars have, quite properly, focused on the input legitimacy of the political and economic reforms that were the sine qua non of financial assistance during the Eurozone crisis.¹⁵ This is important because of the way these reforms were introduced,¹⁶ the input challenge being how to reconcile the power of the Bank with basic principles of democratic accountability.¹⁷

Yet the treaties served as a vehicle via which a deliberate signal was communicated to market actors. Labelling (now) Article 125 TFEU the ‘no-bailout provision’ and referring to Article 123 as the ‘prohibition of monetary financing’ was not ‘merely a question of linguistic convenience’, it also ‘expresses a specific understanding of the provision and serves as a signal to the markets’.¹⁸ And the Bank, supported by the Court, signalled to actors participating in sovereign debt markets that it is now a sovereign lender of last resort.¹⁹ This is a profound constitutional change absent any, or any meaningful, change to the text of those legal provisions that enshrine the grand bargain struck by the member states in the Maastricht Treaty.²⁰ One view is that the text of the treaty fails to offer an adequate account of the thick network of rules, standards and shared understandings that together manifest the functional constitutional order of the Eurozone and the core changes thereto cannot be understood without reference to norm-governed practices.²¹ Unavoidably, capturing the relationship between the Bank and market actors requires us to embrace what, in the context of US constitutional law, Primus describes as ‘small-c’ constitutional analysis.²²

By focusing on output legitimacy, this article complements the work on the relationship between rights, values and legitimacy in the Economic and Monetary Union.²³ It provides a useful set of analytical tools that capture and critique, from a consequentialist perspective, how, and the extent to which, the Bank has emerged as a sovereign lender of last resort.²⁴ It applies the consequentialist literature developed by Ullman-Margalit and Sunstein on the expressive function of law in revising norms by combining it with the recent pathbreaking work by Krisch and Black on the endogenous appropriation of authority and, by extension, legitimacy of international institutions. Where Kilpatrick focused on the violence visited upon the Rule of Law through the conditionality attached to protection from the market,²⁵ this study considers the other side of the coin – the analytical and normative implications of the Bank saving market actors from a narrow, originalist interpretation of the treaties.

The argument is as follows. Historically, a state issues debt in a currency controlled by its central bank which it can force to act as lender of last resort at times of crisis thereby creating an ‘implicit guarantee’ to sovereign debt counterparties that the state will have access to the necessary liquidity to meet its commitments when the bond matures.²⁶ This is akin to a sticky norm that coordinated the behaviour of counterparties to sovereign debt contracts thereby eliminating the threat of a liquidity crisis capable of pushing a state into a self-reinforcing disequilibrium capable of undermining the economic fundamentals of that state. The Treaty of Maastricht was meant to change that as a constituent member no longer issued debt in a currency under its control. Furthermore, it established a single currency absent a transparent, effective sovereign lender of last resort to the member states of the Eurozone. This was an attempt to revise the pre-existing coordination norm through legal intervention. The lack of credibility surrounding aspects of the legal framework led to a disassociation between the norm embedded within Article 123 and, to an extent, Article 125 TFEU and that guiding market behaviour. In July 2012, the Bank removed any doubt about its operational norms: the Bank will, in effect, play the role of sovereign lender of last resort to the member states subject to conditions.²⁷ By intervening to exercise interpretive control over its governing norms, the Bank itself played an active role in constructing and maintaining its own authority and legitimacy in the eyes of market actors. This development was subsequently supported by the Court, the sole authoritative interpreter of the treaties, albeit with few if any realistic alternatives available to it. Yet, this approach raises a legitimacy paradox for the Bank.²⁸

The balance of this study is structured as follows. The next section traces work in (predominantly) US law and economics literature on decision-making and, in particular, the use of legal intervention to revise norms. This consequentialist literature sensitises us to the pre-existing norm and focuses on monitoring for gaps in compliance with the proposed revised norm. The section after that considers the lender of last resort norm that pre-dated the introduction of the Economic and Monetary Union – the norm targeted for revision by Maastricht.²⁹ It offers a rational reconstruction of the emergence of the sovereign lender of last resort norm and reviews the then prevailing macroeconomic and political-economic scholarship to gauge the credibility of the Maastricht provisions. The subsequent section considers the lived experience of those instruments by reference to sovereign bond yields. The article concludes with reflections aimed towards that extant economic shock.

Norm revision through legal intervention

Articles 123 and 125 TFEU contain information introduced to signal change to, inter alia, market actors regarding the future decision-making of the Bank. More particularly, they represent constraints upon the ability of the Bank to provide liquidity to the member states.³⁰ These constraints, like the euro, were sui generis and represented a departure from the pre-existing norm governing sovereign debt.³¹

In this sense, this study is concerned with the process of changing norms and, by extension, the decisions of actors.³² Ullman-Margalit draws a helpful distinction between norm change and norm revision.³³ She thinks of norm change as spontaneous natural evolution over time and, occasionally, over communities. Norm revision refers to a situation where a specific existing norm is deliberately targeted for change. That is to say ‘where the change is instituted intentionally by some social agency’.³⁴ In this section, we are concerned with one method of norm revision: through legal intervention.

Posner provides a useful working definition of a ‘norm’:³⁵

A social norm (‘norm’ for short) is a rule that is neither promulgated by an official source, such as a court or legislature, not enforced by threat of legal sanctions, yet it is regularly complied with (otherwise it wouldn’t be a rule).³⁶

Posner goes on to observe that norms constitute a source of law, an alternative to law and, crucially for present purposes, an antagonist to law. Providing a satisfactory account of the emergence of a norm is complicated by the fact that it may not have come into existence at an identifiable point in time; rather it is probably the result of a complex pattern of behaviour involving a substantial array of actors over an extended period of time.³⁷ Nevertheless, should a complex pattern of behaviour be reducible to a relatively simple, albeit abstract, description of strategic decision-making by market actors and the Bank, then the generation of certain types of norms can be usefully accounted for.³⁸ To borrow the language of game theory, some norms represent solutions to problems posed by strategic interactions.³⁹ However, as the challenges faced by actors are not static, neither are the solutions. And so, in the words of Ullmann-Margalit ‘[n]orms, as social institutions, have careers. They emerge, endure, pass away’.⁴⁰

In an effort to formalise the relationship between law and norms, a wealth of literature on the ‘expressive’ function of law in revising norms mushroomed in the 1990s in concert with the emergence of the ‘New Chicago School’ project.⁴¹ Perhaps the most succinct description of the process of revising norms through statements embedded in legal instruments is captured by McAdams’s information theory of law: ‘[…] law provides information; information changes beliefs; new beliefs change behaviour. Law is informative’.⁴² A commonly cited example of this dynamic in action is the dramatic revision of norms surrounding smoking: the enactment of laws criminalising smoking indoors in public spaces implies the default information that smoking in the company of others is obtuse and discourteous.⁴³ This information, in turn, affects decision-making and behaviour, whether in a prohibited area or not. However, claims to the effect that a legal instrument has affected or shall affect a norm in a particular manner can be empirically difficult to prove (and disprove). Indeed, a criticism of expressivist scholarship is its potential to generate causal assertions that are unfalsifiable. And so for Sunstein the expressive function of law is normative as well as descriptive or positive.⁴⁴ Sunstein suggests that legal endeavours aimed at changing behaviour be combined with a consequentialist approach: ‘[…] if legal statements produce bad consequences, they should not be enacted even if they seem reasonable or noble’.⁴⁵ In other words, Sunstein suggests that we focus upon the consequences of a legal enactment and limit sterile debate about what so-called ‘message’ it sends about society’s political preferences.

McAdams, like Sunstein, argues that, as states seek to manipulate focal points through legal instruments, they should adopt a consequentialist analysis:⁴⁶ attempts to revise norms through legal enactment should focus on the level of ‘compliance’ with the proposed new norm. Norms that perform a coordinating function, such as the lender of last resort norm, enjoy a propensity to be self-enforcing and enduring, even when considered inefficient by the relevant actors. Such norms are sticky because the ‘source of their effectiveness’, or the main motivation for complying with them, is their coordinating function; market actors place a higher premium on having a coordination norm than on the efficiency thereof and face difficulties coordinating their behaviour in an effort to revise that norm.⁴⁷

To summarise, some norms, including as we shall see the sovereign lender of last resort norm under consideration, emerge to solve coordination problems; however, by virtue of the systems of expectations that surround them, coordination norms are inherently sticky and difficult to revise. It usually takes a legal event to do so. In this instance, the Treaty of Maastricht. Yet, the interpretation of a legal intervention may be inextricably linked with a pre-existing norm. States can misjudge the level of compliance that their legal intervention, and the proposed revised norm embedded therein, will enjoy.

We turn now to elucidate the sovereign norm that pre-dated the introduction of the euro.

The lender of last resort function of central banks to sovereigns

The lender of last resort function of central banks long predates the modern focus on the transmission of monetary policy.⁴⁸ In fact, up until the mid-20th century, the key function of central banks was to act as a lender of last resort in times of crisis.⁴⁹ Whelan captures the importance of the lender of last resort function in stark terms: ‘[c]entral Banks were put on this earth to be lenders of last resort’.⁵⁰ In 1802 and 1873 respectively, Henry Thornton and Walter Bagehot both recognised this reality.⁵¹ Goodhart summarised Bagehot’s classical principles for guiding central bank intervention at times of crisis as follows: (1) lend freely; (2) at a high rate of interest; (3) on good banking securities.⁵²

Yet, the lender of last resort function of central banks neither emerged nor operated in an apolitical manner. It developed gradually across the world, leading Calomiris and Haber to argue that the function is a locus of political power and should be viewed as the outcome of a political bargain.⁵³ By the middle of the 19^th century, Britain and France had established operational lenders of last resort whereas the US, Canada and Australia did not do so until 1913, 1929 and 1959 respectively.⁵⁴ Even when established, lender of last resort powers and duties were, like central banks’ independence more generally, far from uniform. The operational independence and flexibility of the central banks in dealing with liquidity shocks in sovereign debt markets differed. Indeed, the heterogeneous legal structures and cultures of the central banks of the member states is a major theme in the authoritative ‘micro-history’ of the Maastricht negotiations offered by Dyson and Featherstone.

Notwithstanding these important cultural and legal distinctions, it was commonly the case that, prior to the introduction of the euro, each Member issued debt in a currency controlled by its central bank. According to De Grauwe this ‘create[d] an implicit guarantee to the bondholders that they will be paid out when the bond matures’ because their ‘central bank can be forced to provide all necessary liquidity to the sovereign’.⁵⁵ The implicit nature of this guarantee, combined with the heterogeneous development of central banking, imposes a methodological constraint on identifying the emergence of the sovereign lender of last resort norm by way of a historico-sociological account. This is because posing seemingly straightforward questions, such as ‘How will I identify a norm when I see one?’, fail to yield straightforward answers. In fact, in the 1990s the focus on the relationship between law and norms waned, in part, due to the challenge of historically or empirically diagnosing a norm.⁵⁶ To make headway, Ullman-Margalit, de-idealising the Carnapian model, offers a framework for the ‘rational reconstruction’ of the emergence of norms as structural solutions to coordination problems.⁵⁷ A structural solution, in the game-theoretical sense, offers ‘a description of the essential features of a situation in which such an event could occur’ not simply as a matter of mere logical possibility.⁵⁸ Rather, the account of the generation of the norm – in this study the lender of last resort norm – is as a solution to a coordination problem. That is, the norm emerged as a stabilising device from the interdependent expectations of market actors together with those of a central bank.

The emergence of the lender of last resort norm

Although this study centres on the sovereign lender of last resort function of the Bank, to elucidate that function we first consider a (stylised) model of banking before applying the underlying dynamic to sovereign debt markets mutatis mutandis.⁵⁹ This methodology is utilised for three reasons. First and foremost, to demonstrate that a lender of last resort is a necessary legal condition for sustainable financial systems, akin to the protection of property rights and the enforcement of contracts.⁶⁰ Secondly, to show that a state and its banking system are tied at the hip; a banking crisis can contribute to a sovereign debt crisis irrespective of fundamentals and vice versa (i.e. the ‘doom loop’ or ‘deadly embrace’).⁶¹ Finally, although we are particularly concerned with rollover sovereign debt crises associated with difficulties in refinancing maturing debt obligations, the choice to provide or withhold liquidity may imply contentious distributional implications. Far from a technical legal concern, the liquidity/solvency categorisation is a crucial site of governance and constitutional debate.⁶²

Banks act as intermediaries between depositors and borrowers. In theory, banks secure and invest the shorter-term savings of depositors by matching them with borrowers who generally require longer-term financing in order to invest in projects in advance of revenues (‘maturity transformation’). Short-term savers loan their funds to a bank such that those funds are available on demand (i.e. a deposit). Yet that bank may provide a long-term loan to a borrower (e.g. a 30-year mortgage). When the decision of an actor to remove her deposit is correlated with that of many other depositors an otherwise healthy bank may not be able to meet its short-term obligations (i.e. it lacks liquidity). A bank might try and secure a loan from another source which may arrest the dynamic or lead to contagion. In order to meet its short-term needs for cash, a bank may be forced to sell its assets, generally, long-term loan books (such as mortgages) at fire-sale prices. As the bank converts its long-term assets into cash in order to meet its short-term obligations, the discounted price it receives for those assets can diminish its balance sheet such as to convert its illiquidity into insolvency. Depositors withdraw their funds in order to convert them into a safer asset class (i.e. a flight to safety) thereby creating a self-fulfilling liquidity crisis (i.e. a bank run). The crucial point is that this dynamic can commence for a good reason or no good reason (i.e. a ‘coordination failure’) and, irrespective of the fundamentals of the Bank at the outset, lead to insolvency. Moreover, this coordination problem can be further exacerbated by market sentiments (i.e. ‘animal spirits’ or ‘fear and panic’).

Applying this dynamic to rollover debt crises, states enjoy a fiscal authority that chooses how much to consume and borrow through the issuing of sovereign debt.⁶³ According to De Grauwe, like a bank ‘[g]overments’ liabilities are liquid, while most of their assets are illiquid’, such as infrastructure or claims on taxation.⁶⁴ Therefore, governments may not be capable of generating cash from asset disposals, taxation or other means quickly enough to payout bondholders at maturity. This gives rise to fear of a rollover crisis where an adverse shift in market expectations, whether for a good reason or not, could restrict a fiscal authority’s ability to roll over debt, creating liquidity problems. The fiscal effort to meet those liquidity problems may have significant adverse macroeconomic implications (i.e. an economic contraction) that would feed back into investor’s expectations, thereby creating a self-enforcing dynamic that could ultimately lead to sovereign default. Moreover, in a significant recession, a state could find repayment of debt more costly, and default less costly, than otherwise, thereby further exacerbating market expectations of default.

Rollover debt crises are relatively rare and are primarily considered a hazard of fixed exchange rate regimes or of the issuing of debt in a foreign currency (usually the US dollar).⁶⁵ The reason that that such crises are rare is attributable to the fact that historically states issue debt in their national currency controlled by their central bank. Therefore, bondholders, like depositors, lend to a state aware that such a dynamic can be arrested by that state’s central bank providing the government with the necessary liquidity to rollover the sovereign debt and repay its loans. According to De Grauwe, this creates an ‘implicit guarantee for bondholders that they will be paid out when the bond matures’.⁶⁶ This not only eliminates the prospect of a liquidity crisis, but in doing so prevents a liquidity crisis pushing a sovereign into self-enforcing disequilibrium thereby deteriorating its economic fundamentals. As we shall see, a member state of the euro area issues debt in a currency it does not control, yet the loss of monetary autonomy was not met with the institutionalisation of a sovereign lender of last resort facility.

Although the foregoing model distinguishes between sovereign and banking lender of last resort activities, domestic financial institutions tend to have significant holdings of the debt of their sovereign. During the Eurozone crisis Ireland’s sovereign debt spreads were negligible until investors began to lose confidence in its banking system while Greece’s public finances contaminated its banking system.⁶⁷ The key point is that, regardless of the source of the initial economic shock, both the banking and sovereign debt spreads moved in lockstep following bad news, creating a ‘doom-loop’ or ‘deadly embrace’.⁶⁸

Central banks are burdened with the responsibility of lender of last resort functions because they enjoy a legal monopoly on printing legal tender.⁶⁹ Depending upon the nature and scale of the crisis, it is likely that a central bank is the only institution capable of performing the role of a lender of last resort to a banking system or sovereign. Therefore, the responsibility to act as lender of last resort is intertwined with the legal responsibility of a central bank regarding money creation. In the words of the current President of the European Central Bank:

As the sole issuer of euro-denominated central bank money, the Eurosystem will always be able to generate additional liquidity as needed […] [s]o by definition, it will neither go bankrupt nor run out of money.⁷⁰

The lender of last resort activity of central banks creates a free-rider problem whereby imprudent risks are undertaken on the belief that a bailout, under the guise of liquidity, will follow (‘moral hazard’). Furthermore, taxpayers may be saddled with costs that are not repaid in full (including any costs associated with the provision thereof). Views remain divided on the gravity of moral hazard. The Bagehot principles require central banks to lend at a penalty rate to avoid perverse incentives. Yet, there is no empirical evidence that the introduction of lender of last resort policies in Britain and France in the 1800s created moral hazard.⁷¹

Despite the antiquity of the Bagehot principles for guiding central bank intervention at times of crisis, there remains no consensus regarding the difference between what constitutes a ‘liquidity’ versus ‘insolvency’ issue during a banking crisis or sovereign debt crisis. Speaking to the banking context, Campbell and Lastra are correct to emphasise that the concepts are ‘fluid and dynamic’.⁷² Goodhart, however, simply considers it a myth to assume that a central bank (or indeed anyone) can tell the difference between the two at the capricious time that the valuation takes place and assistance is required.⁷³ Stiglitz points out that when market actors refuse to lend, they have made a judgement that they will not be repaid and so, unavoidably, central banks pit their judgement against that of the market.⁷⁴ The problems associated with volatile asset valuation extend mutatis mutandis to a state in the throes of a sovereign debt crisis.⁷⁵

To summarise, although there were important legal differences between the central banks, prior to the introduction of the euro, all member states provided an implicit guarantee that their central bank was capable of acting as a lender of last resort in respect of their sovereign bonds.⁷⁶ This study considers that expectation to have been a coordinating norm, or something akin thereto, that existed prior to the introduction of the euro, that served to eliminate a coordination problem: the threat of a sovereign debt liquidity crisis capable of pushing a state towards a bad equilibrium. This study does not delineate the precise scope of the norm and, in fact, acknowledges that the nature of the self-enforcing dynamic brings with it inherent diagnostic challenges regarding the liquidity/solvency distinction that represents a critical site of constitutional governance.

The proposed revision of the sovereign lender of last resort norm

There exists a wealth of academic literature examining from where the ideas that ultimately found expression in the minimalist architecture of the Economic and Monetary Union originally came.⁷⁷ Lastra is undoubtedly correct when she says that the project went well beyond normative economic and monetary institution building.⁷⁸ The mandate of the Bank, along with the powers that it may employ towards that end, are (now) enshrined in both the TFEU and the Protocol on the Statute of the European System of Central Banks and of the European Central Bank (‘ESCB Statute’). The Bank is expressly mandated to act ‘within the limits of the powers conferred upon it by the Treaties’.⁷⁹ Indeed, the Bank argues that its legal framework ‘ha[s] gained “constitutional” status’.⁸⁰

The functions and objectives of the Bank are enshrined in Article 127 TFEU. Paragraph 1 circumscribes the primary objective of the Bank to maintaining ‘price stability’ (i.e. the control of inflation). This stands in contrast with, for example, the dual mandate of the US Federal Reserve, which provides that it conducts its monetary policy in pursuit of full employment and stable prices.⁸¹

Insofar as a measure of flexibility is embedded within the mandate of the Bank, its primary objective is to be pursued without prejudice to the secondary objective of supporting the general EU economic policies with a view to contributing to the achievement of the EU objectives as laid down in Article 3 TEU. Although the primary objective of the Bank is to maintain price stability, ‘financial stability’ is an enumerated objective pursuant to Article 127(5) TFEU. However, this objective is to be pursued within the rule-based restrictions on its powers.

As we have seen, central banks are historically in charge of both monetary policy and lender of last resort functions.⁸² Not so with the Bank. Instead, the treaty instruments curtail the ability of the Bank to act as a lender of last resort to the member states of the Eurozone and the banking system. Article 123 TFEU prohibits, inter alia, the provision of any credit facility by the Bank to a member state. Further, the Bank is prohibited from purchasing a debt instrument directly from a member state. Accordingly, this article is widely interpreted as creating a prohibition on monetary financing, albeit a close reading of the text shows that it is more ambiguous. Though the subject of less judicial scrutiny, Article 125 TFEU is relevant insofar as it prohibits the Union from assuming a liability on behalf of a member state, or the member states doing so on behalf of one another, save for in a narrow exception relating to joint venture public projects (the ‘no bailout’ provision). The second indent of Article 18.1 of the ESCB Statute protects the Eurosystem from counter-party risk when dealing with private financial institutions by requiring that lending be based ‘on adequate collateral’. There is no explicit reference to a banking lender of last resort in either the treaty instruments or the ESCB Statute, however, credit was extended to financial institutions in Ireland, Greece, Cyprus and elsewhere during the crisis. This took place at a national level whereby national central banks provided emergency liquidity assistance to domestic credit institutions. Pursuant to Article 14.4 of the ESCB Statute, the Governing Council (by a majority of two thirds of the votes cast) may declare that the emergency liquidity assistance interferes with the objectives and tasks of the European System of Central Banks.⁸³ The Governing Council can attach conditions to their assent, as exemplified by the 2010 correspondence from the President of the Bank to the Irish Minister for Finance advising him that ‘[i]t is the position of the Governing Council that it is only if we receive in writing a commitment’ to, inter alia, fiscal consolidation and structural reforms agreed and overseen by the Commission, International Monetary Fund and Bank that ‘we can authorise further provisions of [Emergency Liquidity Assistance]’.⁸⁴

Fiscal coordination was distinctly less institutionalised. Absent from the instruments are a fiscal union, a banking union (with risk sharing), or an explicit sovereign lender of last resort. Instead, the Stability and Growth Pact set, under threat of sanction, an annual budget deficit of 3% of GDP and the stock of public debt of 60% of GDP that, according to Eichengreen (writing in 2003) ‘has no basis in economic logic’.⁸⁵ These rules are counterproductive as, per Eichengreen, ‘actually imposing the 2 per cent of GDP fines provided for by the pact would only aggravate the financial problems of a heavily indebted government and provoke the very debt crisis that the mechanism was designed to avert’.⁸⁶ Likewise, more recently Tirole has argued that ‘sanctions aggravate deficits at a time distressed countries have difficulties controlling them’.⁸⁷ From near inception the Commission sued the Council for its effective suspension of the excessive deficit procedure against France and Germany, to little avail.⁸⁸ In any event, Ireland and Spain enjoyed debt to GDP ratios of 40% and were running budget surpluses in the run-up to the crisis.

The absence of a fiscal union is important because:

A LOLR that provides assistance to sovereigns or to banks may unwittingly serve as a device for transferring resources from one sovereign nation to another, if the country whose banks or sovereign has borrowed from the LOLR later decides to default and exit the currency union.⁸⁹

By contrast, Mody argues that in the United States the sovereign lender of last resort activities of the Federal Reserve could leave it with a loss on the government bonds purchased and the US Treasury would have an obligation to use taxpayer funds to replenish the Federal Reserve’s capital.⁹⁰ There is no Eurozone treasury to absorb such losses and the other member states may well bear indirect fiscal consequences.⁹¹ Sims points out that in the event of a capital injection ‘Germany would bear a large part of the burden, and it would be clear that German fiscal resources were being used to compensate for Bank losses on other countries’ sovereign debts’.⁹²

De Grauwe provides a useful way of thinking about the intended effect of these legal instruments: member states are required to issue debt in a ‘foreign’ currency.⁹³ In the event of economic difficulty, the instruments imply that the member states will be forced to resolve their difficulties by way of internal devaluation, disposal of assets, procuring credit facilities from other member states, borrowing on international markets, etc. However, in a serious economic crisis, absent a sovereign lender of last resort, the instruments empower sovereign bondholders to force the default of the constituent member states. Although there is no express provision providing for withdrawal from the Eurozone other than exiting the EU,⁹⁴ Greek exit from the Eurozone (amongst others) was a widely assumed possibility during the crisis.⁹⁵ Indeed, policy responses that pre-supposed temporary ‘sabbaticals’ from the monetary union for the purposes of restoring growth and competitiveness were seriously considered.⁹⁶

As we have seen, these legal instruments attempted a revision of the pre-existing lender of last resort norm whereby the member states issued their sovereign bonds in a currency controlled by their central banks, thereby providing an implicit guarantee that their central bank could act as a lender of last resort in times of crisis.

Credibility of the constitutional architecture of the euro

Prior to its introduction, US economists such as Feldstein and Friedman, who were strong proponents of the single market project, were deeply pessimistic about the prospects for, or need for, a common currency for the common market.⁹⁷ Feldstein warned that it may well be possible to launch a single currency and perhaps even sustain one; however, it would result in longer-term negative economic and political costs including higher average unemployment. Friedman cautioned that a common currency did hold out the advantage of lower transaction costs and the imposition of external discipline on some member states, but it was nevertheless a dangerous idea unless there existed adjustment mechanisms capable of absorbing economic shocks in the absence of flexible exchange rates.⁹⁸

From a foundational macro-economic perspective, Robert Mundell’s ‘Theory of Optimum Currency Areas’, an eight-page article published over half a century ago, provides the framework for identifying the preconditions required for a smoothly functioning monetary union.⁹⁹ In particular, Mundell’s theory, when supplemented with the insights of Kenen, emphasises economic convergence, labour mobility, and fiscal integration.¹⁰⁰ This focus is a by-product of the requirement that the economic structures of the constituent states be similar which in turn would act to minimise asymmetric shocks. In 1992 Bayoumi and Eichengreen¹⁰¹ building on Mundell’s Theory of Optimum Currency Areas, considered the ability of the Economic and Monetary Union architecture to handle asymmetric shocks across the then proposed member states, drawing a strong distinction between shocks to ‘core’ and ‘periphery’ member states.¹⁰²

The Varieties of Capitalism literature developed analysis of the asymmetry of fiscal and monetary policy preferences between northern European export-orientated market economies and the southern demand-based European countries (plus Ireland¹⁰³).¹⁰⁴ The pursuit of dual growth strategies to complement differing institutional arrangements underpinned by differing monetary policy requirements, presaged conflict between the core and periphery. The Treaty was the lowest common denominator upon which the member states could agree during the negotiation process leading up to it.¹⁰⁵

The concerns raised by Theory of Optimum Currency Areas and Varieties of Capitalism scholarship regarding the minimalist legal architecture, in turn, highlight the important role that rational-expectations theory played in providing justification for it. In the wake of the stagnation in the 1970s, Keynesian policies were undermined. In particular, the long-term benefits of manipulating monetary policy were challenged by Chicago School economists such as Friedman, who argued that market actors gradually adapt their expectations based on past outcomes for inflation.¹⁰⁶ Appeals to rational-expectations economics helped justify the sparse legal framework. Market actors, in this case actors in the sovereign bond market, would account for the implications of the information contained in Articles 123 and 125 TFEU, along with the minimalist architecture of the Economic and Monetary Union more generally. They would adapt to the risk implied by the legal instruments – i.e. the absence of a sovereign or banking lender of last resort. Accordingly, they would individually and collectively coordinate their behaviour pursuant to the proposed new norm. In particular, the risks arising from member states issuing sovereign bonds in a currency that they did not control, and in the absence of a sovereign leader of last resort, would be priced into assets, in particular sovereign debt yields.

In sum, the member states effectively established a monetary framework without a clear, transparent sovereign or banking lender of last resort. Moreover, they guaranteed that this would be so by enshrining measures expressly circumscribing assistance by way of monetary accommodation. The Treaty of Maastricht implied two relevant pieces of information regarding risk-sharing. First, Article 123, and possibly 125, TFEU constrained the Bank from acting as sovereign lender of last resort to the member states so that market actors could force sovereign default. Secondly, there was no fiscal or banking union to share the risks associated with default and/or bailouts.¹⁰⁷ This information had the purpose of establishing a new norm: that the Bank would not act as a sovereign lender of last resort. The market actors were, inter alia, an audience for this proposed new norm and those actors were expected to adapt their decisions accordingly. However, the then prevailing US macroeconomic and Varieties of Capitalism frameworks doubted the ability of the euro to absorb an asymmetric shock and doubted the efficacy of the new rules that were inconsistent with the longstanding (even if varied) norms that there would be a lender of last resort.

The lived experience of the legal instruments

In theory the constitutional framework of the euro constrained the Bank from acting as a sovereign lender of last resort to the member states. In reality, from the announcement of the Economic and Monetary Union financial markets lent to periphery states at historically low interest rates which moved towards near convergence of spreads. So much so, the spreads in 10-year government bond rates were close to zero prior to the crisis; a phenomenon disassociated from changes to the economic fundamentals of the member states.¹⁰⁸ In effect, the interest rates of the member states decreased and converged upon that of Germany and stayed at that level until the global financial crisis.¹⁰⁹

In fact, Whelan demonstrates that market actors priced in almost no default risk in the pre-crisis years.¹¹⁰ Investment in the sovereign debt of periphery member states did not command a higher risk premium (relative to the core member states) following the introduction of the Economic and Monetary Union because, in the eyes of the bondholders, the chances of default were significantly diminished notwithstanding the legal provisions.¹¹¹ The convergence of interest rates suggests that the market did not view the no-bailout provision as credible. Assuming that the core member states were unlikely to enjoy the fiscal capacity, let alone democratic will, to bailout the large periphery member states such as Italy and Spain (let alone collectively) without risk of contagion in the absence of cooperation from the Bank, then this also undermines the credibility of Article 123 TFEU.

And so, prior to the global financial crisis, the decision-making of market actors is better explained through an expectation that the periphery member states would have the liquidity to meet their debt obligations at maturity, implying that the Bank would adhere to the pre-existing sovereign lender of last resort norm if necessary.

As the crisis began to unfold, it became apparent that the decision-making of the Bank was not being guided by the pre-existing sovereign lender of last resort norm but rather by that embedded within the treaty.¹¹² As spreads diverged, the Bank did not signal a ‘whatever it takes’ commitment to provide the necessary liquidity to the member states such as to interrupt the flight to safety. Between 2010-12 the periphery member states experienced a significant increase in their spreads relative to those of the core member states. According to De Grauwe and Ji the marked increase in the spreads of the periphery countries could not ‘be accounted for by fundamental developments, in particular by the changes in the debt-to GDP ratios’.¹¹³

This is not to say that the Bank took no action. For example, on 14 May 2010, the Bank announced its temporary Securities Markets Programme.¹¹⁴ Unfortunately, according to Mody and Nedeljovic, the Bank’s measures led to a reduction in median bond spreads but also to a stronger perception that spreads on Greek bonds would rise and spreads on Spanish, Portuguese and Irish bonds would also rise and remain elevated. They conclude that the market was largely unsure about the Bank’s strategy.¹¹⁵

As the decision-making of the market actors departed from the pre-existing lender of last resort norm, their coordination shifted toward a new focal point; the Bank would not act as sovereign lender of last resort. The effect: a cascading dynamic of self-perpetuating crisis or in common parlance: a bank-run dynamic and flight to safety.

What ‘it’ took – the exercise of interpretive control over operational norms

Through the provision of emergency liquidity assistance, national central banks acted as lenders of last resort to their banking systems during the crisis.¹¹⁶ From May 2010 periphery member states began receiving bailouts and, on 2 February 2012, the European Stability Mechanism was agreed, thereby institutionalising a permanent crisis resolution mechanism funded by bond sales and capital from member states.¹¹⁷ Nevertheless, the situation continued to deteriorate.

On 26 July 2012, Mario Draghi announced that the euro was irreversible and that ‘[w]ithin our mandate, the ECB is ready to do whatever it takes to preserve the euro. And believe me, it will be enough’.¹¹⁸ In August 2012, the Bank announced its policy of OMT.¹¹⁹ On 6 September 2012, a press release on decisions taken by Governing Council of the Bank set out the technical features of OMT.¹²⁰ The Bank announced that it would purchase the debt of distressed member states absent any ex ante quantitative limit. The sum and substance of these communications was to signal to the market that the Bank would act as a sovereign lender of last resort subject to conditions.¹²¹ Pursuant to Article 123(1) TFEU the Bank is not prohibited from purchasing government bonds in the secondary markets.¹²² Formally in doing so it is not directly providing credit to the member state. In reality, however, the Bank is providing liquidity to the holder of that bond, typically a financial institution within that member.¹²³ What are the analytical and normative implications for the legitimacy and authority of the Bank in the eyes of market actors?

Transnational organisations can and do exercise significant authority over a variety of audience actors in the absence of a traditional formal or ‘solid’ legal architecture and accoutrements.¹²⁴ To do so, Krisch argues that they must establish and maintain their own authority in the eyes of their audience. One manner in which they do so is through building perceived functional effectiveness. The ‘liquid’ nature of their authority renders unavailable to them the traditional tools for norm revision by threat of sanction (penalties, fines, etc.). Instead, they must appropriate for themselves an endogenous, as opposed to exogenous, authority based more on deference than on command and control. Put another way, scholars such as Renan remind us that the exercise of power can be augmented and constrained by norms of legitimate and respectworthy behaviour.¹²⁵ Krisch complements this by arguing that power can also be established and maintained through adherence to those norms.¹²⁶

The Bank does not fall neatly into the category of a purely ‘liquid’ institutional actor. The treaties provide a level of solidity not enjoyed by many transnational institutions, and implied departure from the pre-existing norm of a sovereign lender of last resort at a time of crisis.¹²⁷ Yet, treaty revision to reinstate that historical norm was not feasible.¹²⁸

But Krisch’s distinction between ‘liquid’ and ‘solid’ authority is not just binary. The metaphor and ancillary vocabulary has been extended by Black, herself developing and extending the work of Raz, to usefully capture the output legitimacy paradox arising from such interventions. The Bank sought to exercise control over its operational norms through an authority that was more ‘liquid’ than formally legal (solid).¹²⁹ The intervention enjoyed a high level of dynamism, was delivered in a relatively informal manner (a conference for investors in London), the features of OMT were communicated in technocratic terms (a press release) over a month later, and the exercise of authority was not (and could not be) accompanied by a formal sanction aimed at bondholders such as a fine or penalty. In the eyes of its audience (market actors), the commitment of the Bank to save the euro, along with its technical ability to do so via OMT, removed any doubt about the operational lender of last resort norm; the Bank would act as de facto lender of last resort, albeit one only operating on secondary bond markets and subject to certain conditions, despite what a narrow, originalist reading of Article 123 TFEU had, until the press conference and the subsequent judgment of the Court, suggested.¹³⁰ In effect, the Bank signalled its decision to be guided by the pre-existing lender of last resort norm so long as member states entered into a credible programme of adjustment thereby ensuring the medium-term viability of their debt obligations. The Bank assured market actors that they could trust in this commitment and they did,¹³¹ a trust later validated by the Court.¹³² The role of perceived authority cannot be understated because once a lender of last resort norm is considered credible the chances of it actually becoming engaged become significantly diminished.

The decision of the Court in Gauweiler was justified (in part) on the distinction that the Bank is not primarily (as it claims) acting as a sovereign lender of last resort through the promise of OMT.¹³³ The Bank claimed that OMT was not aimed at supporting the financing conditions of the member states but rather to unblock monetary policy transmission channels.¹³⁴ In this respect, the Court found that the objectives and instruments of OMT contributed to monetary policy and that any indirect effects on economic policy did not classify it as so. Furthermore, the Court placed weight on the fact that the purchases would not be directly from a member. Suffice it to say these were not ordinary times. It would have been better had the Bank not been obliged to push its mandate, but it was.

In any event, the Court clearly signalled that the ‘whatever it takes’ intervention of the Bank, the absence of any ex ante quantitative limit on the size of its bond purchasing powers, and the broad institutional discretion that the Bank assumed for itself, all survived legal challenge.¹³⁵ This suggests that, at least from the audience perspective of market actors, the Bank’s signal to the effect it would act as a sovereign lender of last resort, as per the historic norm, was credible and trustworthy, especially when endorsed by the Court.¹³⁶ In this way, the Court could be said to have acted as a solidifying agent rendering exogenous the endogenous authority and legitimacy of the Bank. Indeed, in doing what was perceived as necessary to save the euro, just like the Bank before it, the Court might be said to have played an active role in maintaining its own authority and legitimacy.

This brings into focus the danger implied by the German Federal Constitutional Court’s decision in Weiss. In describing the legal analysis of the Court of Justice as ‘not comprehensible’, one that ‘manifestly exceed[s] the judicial mandate conferred upon it’ and the judgment no more than an ‘ultra vires act that is not binding upon’ Germany, the Federal Constitutional Court too is sending a signal to market actors about the extent of the Bank’s capacity and, by extension, authority to market actors.¹³⁷ To be clear, the Bank’s Public Sector Purchase Programme is consistent with EU law as adjudged by the Court of Justice. Nevertheless, as the Bank operationalises the Pandemic Emergency Purchase Programme, and possible successor programmes, the prospect that one of the national central banks may find itself restrained by its domestic constitutional apex court as it carries out its obligations as a constituent of the European System of Central Banks, creates unhelpful uncertainty that may serve to undermine the authority of the Bank.

Conclusions

For all the legal ink spilt on the Eurozone crisis, perhaps its genesis story is one of a legal instrument that was in strong conflict with a prevailing norm rendering it ineffective at times of crisis. That historic norm had emerged to solve a problem.

In any event, the fact remains that prior to the ‘whatever it takes’ intervention there was a clear mismatch between the framing of important treaty provisions and political-economic exigencies required to sustain the currency created by those treaties. Market actors during the crisis did not trust that the Bank would act as lender of last resort. For the institution tasked with adhering to the legal norm, the European Central Bank, the welfare costs associated with doing so became overwhelming. The Bank persuaded actors to trust its ‘whatever it takes’ intervention, but that persuasion was inextricably bound up with the operation of norms through which the choices of market actors were channelled. But the Bank was also able to successfully speak a second language to another audience, persuading the Court that it would operate in a manner consistent with the treaty, as now interpreted. The Court was itself sensitive to the political-economic reality. By acting as a bulwark against an undermining of the Bank’s authority, the Court maintained and solidified not only the legitimacy of the Bank but of itself.

In the wake of this experience, the European monetary infrastructure is dealing with a profound shock arising from the economic consequences of the Covid-19 pandemic. The Eurocrisis required the Bank and Court to depart to a disconcerting degree from originalist readings of treaty provisions to accommodate the sovereign lender of last resort function of the Bank. Underwriting the expansion of the European Stability Mechanism, Coronabonds, sovereign debt restructuring and other unconventional measures now under discussion imply a further dissolution of the constraints imposed by the treaties, giving rise to legitimate concern.¹³⁸ This, in part, may have prompted the unconventional rhetoric of the German Federal Constitutional Court in Weiss whereby it, in effect, sought to assert for itself greater capacity to scrutinise and constrain the Bank’s crisis response.

But, just as importantly, there are constraints imposed by the hard core reality that the formal legal framework enshrined a profoundly inadequate lender of last resort function for the Bank. The fact that even policy measures necessary to save the euro or to promote growth – such as quantitative easing, pursued by central banks around the world – require the Bank to speak two languages does not bode well. Yet there is hope in the observation of Oliver Wendell Homes Jr. that ‘[t]he life of the law has not been logic: it has been experience’.¹³⁹ Therefore, as this latest shock brings political actors back together to find new compromises regarding the distribution of the burden of adjustment arising from it, they do so with the benefit of the authority that the Bank and Court have appropriated for themselves in the eyes of market actors. They also do so with the benefit of a greater sense of the strength and robustness of the trust placed in the legal instruments at their disposal.

This time, reform proposals cannot operate on a latent assumption that member states can, in uncomplicated ways, lay down new rules for market actors to follow. Cold water must be poured over calls for rule adherence or judicial review that is insensitive to the norms that guide the decisions of market actors. Poorly designed legal instruments and ill-considered judicial interventions that take inadequate account of prevailing norms and the exigencies of the circumstances function to destabilise the economic constitution of the Eurozone and the credibility of EU law more generally. Thus, the ability of member states to effectively solve coordination problems through credible commitments enshrined in law requires sensitivity to the dynamic authority, legitimacy, and capacity of those legal instruments and the institutions they establish. The construction of a functional constitutional framework that reinforces desirable norm-governed behaviour, yet manifests the capacity to revise targeted legal provisions, will guard against the erosion of the euro by the economic shock visited upon it by Covid-19.