Rethinking banking supervision and the SSM perspective

Speech by Ignazio Angeloni, Member of the Supervisory Board of the European Central Bank,
Conference on “The new financial architecture in the Eurozone”,
European University Institute,
Fiesole, 23 April 2015

Introduction: remembering TPS ^[1]

I am grateful to the organisers for inviting me to this conference, as it also gives me the opportunity to visit the European University Institute (EUI) one more time. Every return to the hills of Fiesole is pleasant and brings back memories. This time the memory is that of Tommaso Padoa-Schioppa: a long-time “friend of the EUI”, as your website reads. Tommaso’s association with EUI was a long one. In 1982 he met here Altiero Spinelli ^[2], whose professional and personal influence would become central to his life. After that, he came back regularly to lecture and give speeches; I am honoured to say that I joined him a few times. Following his untimely death in 2010, his personal archives were donated to the EUI and will be, in due course, available to scholars. Recently, the Institute has created a Chair named after him, which will help preserve his memory and continue his research.

15 years ago (in March 2000) Tommaso wrote an article that, though not being among his most cited ones, I always found remarkable. The title is “An institutional glossary of the Eurosystem”. ^[3] The word “glossary” is an understatement, and refers to the dictionary-like form of the article. The ambition of the piece is no less than to review, for the newly created ECB, some of the foundations of central banking developed over 30 or more years of research and debates: things like central bank goals, independence, transparency, accountability, and so on. The semantic expedient is used to convey two implicit messages: first, that those notions are not immutable but should be interpreted and applied according to the times and circumstances; second, that the euro and the ECB are such novel endeavours that in order to understand them one must, first and foremost, redefine the language.

This article came to my mind while we were preparing for the Single Supervisory Mechanism (SSM) and during its first year of activity. The crisis called into question many established wisdoms regarding financial policies: monetary policy strategies and operations, lending of last resort, banking regulation and supervision, crisis management, contagion control, state support to banking, and so on. The lessons are relevant especially for the euro area, which is, in fact, where the main institutional changes are taking place. As the dust settles, one feels the need for a systematic rethinking and redefinition of many of those common wisdoms that looked immutable to many of us until 2006. In the field of banking supervision, the SSM, newly created precisely to respond to some of those challenges, is ideally placed to put any new thinking into practice.

I do not plan to carry out a similar task in my intervention today. I lack the time, let alone the vision and the insight, to do for supervision what Tommaso did for monetary policy 15 years ago. I just want to offer some reflections triggered by our experience in starting the SSM, also in the light of the crisis and my earlier experiences in the ECB monetary policy function. The similarities and differences between the two policy areas are instructive. I will organise my arguments around a few main themes, starting with the scope of banking supervision, its goals or mission, then moving on to its independence, transparency and accountability, and concluding with the relation with its “host”, the central bank.

Banking supervision: drawing the boundaries

Let’s indulge, for once, in Tommaso’s habit of starting a discussion on a topic by examining first the literal meaning of its name. The Webster online dictionary, under “supervision”, refers to “watching and directing what someone does or how something is done”. The Latin origin ( super and visio) literally means “to view from above”. The English terms “surveillance” and “oversight” are etymologically equivalent, though sometimes they acquire different meanings in economic policy practice. The German Aufsicht and the French supervision convey the same idea, while in the Italian vigilanza the etymology is lost.

What does “seeing from above” mean? I think it would be presumptuous to assume that those who exercise supervision are always “superior” to those being supervised, professionally, morally or in other senses. It doesn’t hurt if they are, but we cannot count on this being always the case. The interpretation I favour is that, in order to watch and direct, the supervisor has to be in a position to observe the broader context in which the actions take place. It has recently become clear that this superior perspective is essential: it is impossible to gain a proper understanding of the “safety and soundness” of a bank, in any relevant case, if that picture does not include the links between that bank, other banks and the broader economic and financial environment. The banking system is nested in the broader global financial and economic system in a complex interconnected structure, with different layers. ^[4] Recently, for example, shadow banking is increasingly in focus, as are the potential risks from the insurance and pension funds sectors. The interconnection with banks should not be underestimated. Supervision focused only on the books and activities of individual banks is insufficient and potentially misleading. I will say something later on how the SSM tries to combine the micro- and macro-financial perspectives.

Observing supervisors at work, I have noticed a recurring tension in the way the limits between supervision and regulation are defined. In English, the term “banking regulation” encompasses both supervisory and regulatory functions, but in continental Europe we consider them distinct. Regulators (including lawmakers) are supposed to write the rules, while supervisors merely ensure they are observed. I have used this distinction myself at times, because it is easy to explain and to understand. But it is to some extent illusory, and there is a risk it may at times become a way to elude responsibilities. The line between the two functions has weakened further recently. Let’s consider the European example. In 2011, three “supervisory agencies” were created; in fact, they are not supervisors – in the strict sense of monitoring compliance – but EU secondary regulators, tasked with ensuring that European laws are transposed into national law and applied consistently across all countries in the Union. In the area of banking, the European Banking Authority (EBA) does this by issuing implementation rules and technical standards that Member States (notably their supervisors and banks) are supposed to apply. Conversely, the EU supervisory authorities – meaning the SSM for 19 Member States and the respective national authorities for the remaining 9 – not only check compliance, but actively contribute to shaping the rules within the EBA’s decision-making process.

More importantly, national and European banking laws and secondary regulations typically set minimum standards (for example, for capital and liquidity ratios), not specific levels that banks have to maintain. Nothing prevents banks from upholding standards above those minima, and supervisors in fact typically require sizeable additional margins as part of their so-called Pillar 2 evaluation and intervention process. The Pillar 2 process consists in examining all sources of risks to banks, in addition to those inherent in the determination of minimum solvency criteria; this includes other balance sheet features, like liquidity and maturity transformation, plus internal organisation, governance, controls, the sustainability of the bank’s business model in different economic scenarios, and so on. The Pillar 2 process includes also macro-prudential elements, according to European legislation (CDRIV and CRR), though it rarely plays a central part. Supervisors typically incorporate all these elements in a framework (we can call it the “supervisory model”) to ensure consistency. The framework is to some extent quantified, by means of scoring methods, and allows margins of subjective judgement. ^[5] The end result is the determination of prudential add-ons to the minimum capital requirements, as well as the identification of other actions that banks are asked to undertake, depending on individual conditions.

There is a general tendency for Pillar 2 processes to become more articulated, systematic and codified, hence more transparent and convergent towards international best practices. The ECB has developed a methodology for its own evaluation process, starting from the experience of its constituent national authorities, that will be applied this year for the first time to the banks it supervises directly.

An important warning should be made here. While supervision becomes more rich and systematic, acquiring some regulatory characters, efficiency and simplicity constraints become more pressing. Supervision should never become a further regulatory overlay. The burden of compliance for the industry is already high and should be kept under control. This is especially important in the SSM area, where new authorities have been created.

Mission

Let’s move one step further. If we regard supervisors/regulators as agents delegated by society to accomplish a mission, more questions arise. What is the goal of banking supervision and regulation, and who should set that goal?

As I am speaking in a university, let me say first that I am surprised to see so little research devoted to the goals of banking regulation. The comparable literature on monetary policy is endless; discussions on how to set, measure, pursue and justify the goals of monetary policy virtually never stop, acquiring new life at each turn of the economy – the latest crisis being no exception. In other policy domains, such as public budgets, taxation or labour markets, debates on the nature of the policy objectives are less intense, but still more active – it seems to me – than what we see happening in the field of prudential regulation and supervision.

I don’t think this depends on the questions to be asked being trivial, or the answers unimportant. It probably has to do with certain complexities that discourage both the academic and the practitioner in tackling the subject head-on, and especially in entering into exchanges with one another. Between the banking supervisor and the theorist in the same field there is more distance and less understanding than between their homologues on the monetary policy side. The complexities have to do, in part, with the confidentiality of the subject and the vested interests involved. Supervisors deal with companies that compete on the market and are therefore reluctant to release information. As a result, the supervisory process takes place largely in the shadows, and this makes both scrutiny and independent analysis more difficult. Another factor is that the policy process itself is difficult to define, involving a multiplicity of instruments used to attain a continuum of generically defined objectives. We are very far from the simple one-instrument, one-target, one-transmission mechanism environment that most monetary policy scholars are familiar with.

Those complexities and that distance have costs. To begin with, it is more difficult for supervisors to communicate with public opinion or other non-specialists. The absence of explicit analytical frameworks and well-articulated policy objectives makes it more difficult to explain, in non-technical terms, what needs to be done and when, and why occasional policy failures occur. The activity of supervisors, not generally visible to outsiders, falls suddenly under the spotlight when banking crises occur; when that happens, the supervisor is usually found guilty without appeal. This lack of visibility may explain, incidentally, a somewhat lower perceived attractiveness of the supervisory profession relative, for example, to core central banking functions. The latter are – wrongly, I think – considered more “glamorous” because they appear to be more scientific and make the newspaper headlines more often. I want to note here, in passing, that this perception did not prevent the ECB, last year, from conducting a very successful recruitment for its supervisory structures, attracting high talents both from the supervisory community and from the market.

A consequence of this communication gap is a widely held public misperception of what supervision is supposed to achieve. Most non-specialists probably think banking policies should prevent bank failures in all circumstances. This expectation is unfounded, of course. Banks are subject to market discipline like other private firms. Their specific safety arrangements are meant to protect not their shareholders, but other stakeholders, including creditors and users of banking services that are regarded as public goods (like payments), and ultimately the general taxpayer, who acts as a backstop.

Is this extra degree of complexity and opacity an inherent feature of banking supervision, or is it the fruit of inherited working practices that can and should be changed? Views differ. I would agree with those who think that certain risks faced by banking supervisors, and them alone, require to be treated with particular caution in public communication. At the same time, I am also convinced that much more work can and should be done to make supervisory practices more transparent and better understood by all.

Let me return to the initial question now, concerning the goals of supervision. The Core Principles of the Basel Committee state that the primary objective for banking supervision is to promote the “safety and soundness” of banks. ^[6] However, the meaning of “safety and soundness” needs to be clarified. For sure, it does not mean riskless. Banks are risky by definition; they cannot conduct their business otherwise. ^[7] I think those terms mean, generally speaking, that the risks borne by the taxpayer and by the creditors of the bank are appropriately contained and transparently disclosed. The exact extent and distribution of those risks, however, needs to be determined more precisely. ^[8]

Hanson, Kashyap and Stein ^[9] have proposed a definition of the goals of micro- vs. macro-prudential supervision, assuming, in accordance with neoclassical logic, that public policy intervenes only to correct market failures. In their view, micro-prudential policy should correct the distortion towards risk taking created by the safety net. Conversely, macro-prudential policy is meant to correct for other market failures, also giving rise to undue risk taking, generated by “systemic externalities”.

I find this argument useful but incomplete. If bank services provide positive externalities to society, then not all taxpayer risk should be removed. Over-regulation is also sub-optimal. At the same time, the correct balance between taxpayer risk and the involvement of other stakeholders (shareholders, creditors) depends on collective preferences. We are crossing the line between technical competence and politics. Political reasons help explain, for example, why after the recent crisis regulation has increasingly tended to protect the taxpayer at the expense of bank creditors. In Europe, the bail-in provisions of the Bank Recovery and Resolution Directive, which are particularly severe, are an illustration.

I draw three conclusions from this discussion.

First, regulation and supervision should aim at balancing the risks and benefits of banking, taking into account all externalities involved. The correct solution is not one in which all taxpayer risk is removed.

Second, society should be put in a position to express more explicitly its preference as to where that balance is located; at present this form of collective guidance is lacking almost everywhere. This requires appropriate public communication on the nature of banking, its risks and implications; difficult issues on which specialists are also divided.

Third, more research is needed. We need proxies for bank risk and stability, providing a yardstick for setting supervisory goals and measuring performance. More work is needed also on supervisory instruments, clarifying how they interact with each other and how they affect stability (the “transmission mechanism”). Finally, evidence is needed on the interconnections and feedbacks between banks and the economy. Advances have been made, especially by network and contagion analyses; important micro-data sets on interbank exposures are being developed. Supervisory practice should hopefully be able to make increasing use of those data.

Independence

In thinking about supervisory independence, drawing a parallel with central banking can be helpful.

It is generally accepted that, in order to be successful, monetary policy needs to be free from short-term political interference and delegated to a technically equipped agency, the central bank, formally bound to a clearly defined goal. To balance that independence, appropriate reporting obligations must exist (“accountability”), requiring the central bank to provide information on its actions (“transparency”). Delegation of complex policy tasks where time inconsistency problems arise has proved to be beneficial not only in monetary policy but also in other policy areas.

Now the question arises: does the same framework apply to banking supervision? Or is there something inherently different there that warrants specific arrangements? This question was not explored in detail until the late 1990s, when several countries (the United Kingdom being the most prominent example) decided to separate banking supervision from the central bank and entrust it to a separate agency. This is a bit surprising to me, because while differences exist between the two policy functions – as I shall argue – they are still relevant regardless of whether they are performed by the same institution or not.

I think that the criteria suggesting delegation to an independent agency suit banking supervision no less than monetary policy. First of all, banking supervision is highly technical and complex, requiring a mix of financial, accounting and legal expertise. Moreover, the potential conflict between short-term and long-term objectives (the time inconsistency problem) is likely to be relevant as well; for example, supervisory forbearance may help to protect confidence in individual institutions in the short run, if the supervisor enjoys a high degree of credibility, but is likely to be detrimental to such credibility – and to financial stability – over a longer horizon. In addition, banking supervision typically involves important vested interests, a further reason for separating policy from direct political control.

This notion has been recognised in recent years as a key component of bank supervisory practice. The Core Principles of the Basel Committee on Bank Supervision specifically mention the need for operational independence of supervisors without interference by government or industry. ^[10] The supervisor should have full discretion to decide when and if it needs to take action.

A high degree of transparency is especially suited to a new institution like the SSM, which has no track record. In particular, the EU Regulation establishing the SSM stipulates that the ECB should be bound in its decision-making process by Union rules and general principles on due process and transparency. In this context, the ECB is accountable towards the European Parliament and the Council. This includes regular reporting, and responding to questions by the European Parliament and the Eurogroup. ^[11]

Transparency and accountability

While the case for independence is relatively straightforward, specific circumstances make transparency and accountability in supervision especially delicate, requiring particular safeguards.

First, supervision is special in that it involves handling two types of information, one concerning the authority itself and its behaviour (proceedings, deliberations, internal thinking, strategy and methodologies, etc.), and the other concerning the supervised entities.

Second, supervisors typically obtain, in the exercise of their function, sensitive information about the situation of individual banks. Proprietary information generated within the bank may, if publicly known, affect its competitive position. The supervisor does not have the legal right to disclose such information; the obligation may fall on the banks, themselves, in certain cases. Banks typically trust that the supervisor will treat information confidentially, and this facilitates the flow of information between them. A similar situation is not typically seen in monetary policy.

Third, the supervisory process generates information on the soundness of individual banks – their solvency, liquidity, profitability, quality of internal governance, viability of business models, etc. Early disclosure, especially when the picture is not yet complete and any necessary countermeasures have still to be taken or planned, can be risky and counterproductive, endangering financial stability. This does not exclude, however, the publication of supervisory statistical data, along the lines of what is done in the United States, for example.

The supervisor in the central bank: cohabitation issues

I have to mention here an instance in which I disagreed with Tommaso. In an article written for an ECB conference, he argued that central banking and financial stability are linked because financial stability is in the “genetic code” of central banks. He referred to the historical role of central banks, in many countries, as guardian of stability. ^[12] At the time (2002), macro-prudential policies were not established yet, hence what he essentially meant was that central banks should be involved in banking supervision.

I objected, arguing that institutions evolve, just like biological species do (penguins was the example I used, that nowadays use their former wings as flippers); the habits of our ancestors should not impede that evolution when conditions change. My argument echoed the prevailing thinking of the time, with its preference for narrowly defined central banks disjoint from supervision. ^[13]

I felt rather happy about my objection at the time, but in fact events proved him right shortly after. The crisis produced a unification – or re-unification – of central banking and supervisory functions almost everywhere. The most documented case is probably that of the Bank of England, but the pattern is more general. Central banks that host supervisory functions normally are organised so as to maintain an orderly distinction between the two activities, in the form, for example, of separate decision making bodies, organograms, and a degree of attention and control over the extent and the modality in which information is exchanged.

The ECB is a case in point. The decision to locate the new single supervisor in the ECB was almost immediate. It was dictated by legal considerations but also by the will to strengthen the new authority from the outset with the experience and the established reputation of the ECB. At the same time, the legislators – notably the European Parliament – insisted that clear separation lines should be included in the charter.

The SSM Regulation specifically mentions that the supervisory tasks should be separated from the monetary policy function. Provisions are included in the SSM Regulation to bring this about. A separate Supervisory Board has been established, in which senior representatives from all participating Member States, four ECB representatives, and a Chair and Vice-Chair participate. This Board is responsible for planning and executing the supervisory tasks and for drafting all supervisory decisions. The Governing Council of the ECB formally adopts the decisions via a non-objection procedure. In the case that the Governing Council does object to a draft decision prepared by the Supervisory Board, an established mediation panel is activated to resolve the differences.

The operation of the Governing Council is completely differentiated as regards monetary and supervisory functions. Such differences include strictly separated meetings and agendas. Moreover, the supervisory units are organisationally distinct from and subject to separate reporting lines. The ECB has set up a formal “separation framework”, in the form of internal arrangements to distinguish the two functions, in particular with respect to professional secrecy and the exchange of information, ^[14] while at the same time ensuring that the organisational and information synergies are exploited.

These arrangements are working. The clear statutory mandates of the Supervisory Board and the Governing Council make a blurring of responsibilities unlikely. The two bodies interact regularly, with the respective roles clearly delineated. Moreover, regular joint meetings are convened to discuss financial stability and macro-prudential issues, where central banking and supervisory interests intersect. ^[15]

Conclusion

This speech has already been long enough; moreover, its intent was to present elements for reflection, not firm policy statements. Hence no specific conclusions are needed.

To the prevailing academic audience, I would like to remind that the banking and financial regulatory/supervisory universe is in a state of rapid evolution everywhere, with consequences that are important for the society as a whole. There are many interesting and important issues that call for more research. Do not misinterpret the distance that exists, at present, between supervisors and academics as a sign that your work is uninteresting or unimportant. It is not. And that distance needs to be narrowed.

Thank you for your attention.