Speech by Ignazio Angeloni, Member of the Supervisory Board of the ECB, at the ECB colloquium “Challenges for Supervisors and Central Bankers”, Frankfurt am Main, 22 March 2019
I am grateful to you all for being here today. It is a great pleasure for me to see so many of the people I have had the fortune to know and work with throughout my professional life.
My topic this morning is supervisory independence. The subject has attracted some attention lately, as has central bank independence, although in a different way.
When I first joined a central bank in the early 1980s, central bank independence was a new idea. The classic papers by Barro and Gordon and by Ken Rogoff were still circulating among specialists in working paper form. My employer, the Banca d’Italia, had just gained some independence from the Treasury (the so-called “divorce”) and was beginning to exercise it with some hesitation, as if it was taboo; I remember one superior telling me at the time that a central bank should not discuss its own independence, because it was “not polite”. By the way, this is why I have waited until the day before leaving to talk about it: I am trying to be as polite as I can. Since then, central bank independence has become commonplace; central bankers, a very polite lot, speak about it all the time. But now the consensus is waning. An early supporter, Larry Summers, recently wrote: “… I suspect that we may have moved beyond the conditions arguing in favour of peak central bank independence in the industrialised world”.
The cycle of supervisory independence followed that of monetary policy with a lag of about 20 years. The idea dates back to the 1990s and became firmer after the financial crisis. The Single Supervisory Mechanism (SSM), the youngest supervisor in existence today, has in its charter independence provisions that are essentially copy-pasted from the EU Treaty articles on monetary policy.
Short and simple, my point is that in the present state of affairs, supervisory independence is necessary but also very difficult to put into practice. To make it real, some changes are required. The process has been set in motion, but more needs to be done. In my remarks I will mainly refer to the European situation, though I think that my main points may have more general validity.
Supervisory independence is necessary
An early mentor, who is familiar to all of you, but whom I will not name, once told me that there are two mistakes an economist should never make when approaching a subject: reading the literature and looking at the data. I have learned a lot from him, but I never followed this particular advice. So let me start by mentioning some early contributions on the subject.
The earliest reference I can think of to regulatory independence, or rather to the problems arising from the lack of it, is by the Chicago economist George Stigler. In 1971, using the rudimentary econometrics of his time, he was able to demonstrate that regulators systematically lean in favour of the industry they regulate. They should therefore be put in a position to withstand pressure and capture by the industry.
What I find remarkable in this early contribution is not only its enduring relevance, but the fact that it completely ignores the other dimension of the problem, that of political capture. Stigler makes no distinction between the regulator and the state and assumes that the latter always pursues the collective good. So much faith in the regulatory function of the state is surprising in a founding father of the Chicago school, who even called himself an “unregulated economist”.
The divergence between daily politics and collective welfare did, however, feature prominently in the discussions about central bank independence over the next 20 years or so. As is well known, the crucial distinction here is in the respective time horizons of political decision-making and collective welfare: the time inconsistency problem. The focus was initially only on monetary policy, even though central banks were also heavily involved in banking supervision. It broadened to financial regulation in the 1990s, with the growth of financial markets, the new liberalised environment and the occurrence of various banking crises.
One of the first, if not the first, to argue in favour of independence of banking supervision was not an economist but an eminent lawyer, Rosa Lastra. In a book published in 1996, she wrote “… though independence is often advocated for monetary policy, I claim that it is also necessary for banking supervision”. Indeed, the time inconsistency argument applies forcefully to banking supervision as well. Supervisory policies are inherently forward-looking, affecting financial stability with long and uncertain lags. The role of expectations, including of how the supervisor will behave in the future, is no less important. There may be short-term gains from supervisory accommodation, while the costs of laxity materialise over time. And financial instability is no less harmful than price instability. I think this reasoning is quite compelling.
Two well-known economists, Alberto Alesina and Guido Tabellini, added further arguments. They showed that policy areas that are relatively technical and areas where vested interests play a major role should, all else being equal, be delegated to independent agencies. Banking supervision possesses both characteristics. Conversely, the presence of distributional implications from the policy – an argument often invoked for politicising supervision – does not affect the advantages or drawbacks of delegation in any particular way.
Today, supervisory independence is broadly accepted as a concept. The Basel Committee on Banking Supervision has placed it high up in its list of Core principles of effective banking supervision, which are endorsed also by the World Bank and the IMF. Recently, the European legislator granted the SSM a high degree of protection, using in Article 19 of the SSM Regulation a formulation that is virtually identical to that in Article 7 of the ECB Statute, dealing with monetary policy. The Regulation also prescribes to the SSM a high degree of accountability to several political authorities.
However, establishing a principle does not mean being able to easily put it into practice. I will mention three current obstacles pertaining to the legal, the analytical and the political dimensions.
As everybody knows, banking supervision is very legalistic. Supervisory decisions are legal acts, affecting private rights which are in turn protected by law. Those decisions are often challenged in court, becoming invalid and even turning into a liability for the supervisor.
What is less clear to most, and something I myself came to appreciate only over time, is that the assignment of certain duties to the supervisor does not in itself mean that they can actually be discharged. For example, the SSM Regulation lists the ECB’s powers as supervisor, but this does not mean that the ECB can actually do those things: it only means that it can use the law to do them. This distinction is important because two pieces of legislation may differ, especially in their interpretation.
Here is a concrete example. Article 16 of the SSM Regulation states, without a shade of ambiguity, that the ECB can require banks to apply specific provisioning policies. Recently, however, the ECB’s power to set provisioning calendars for deteriorated exposures was challenged by the European Parliament, on the grounds that calendars, being akin to rules, invade the prerogatives of the legislator. Prudential provisions could only be applied case by case, not across the board. Indeed, European law stipulates that the Pillar 2 requirements, of which prudential provisions are a part, can only be derived from an assessment of each bank’s specific risk. One can see here a case where the combination of law and interpretation severely limits the supervisor’s scope to independently set prudential provisions. I took this as an example because the controversy between the ECB and the European Parliament on provisioning for non-performing loans (NPLs) was extensively in the media, but the problem is more general.
A catch-22 situation arises here. Everybody agrees that independence must be counterbalanced by transparency and accountability. The ECB is under pressure from the industry and others to be transparent and even-handed in its Pillar 2 demands. Yet the more it is so, the more these demands end up looking like an additional layer of rules, on top of the Pillar 1 requirements set by law. A systematic and transparent Pillar 2 policy is easily mistaken for another rule, and as soon as this happens the supervisor meets the boundary of what it is legitimately allowed to do.
I see no way out of this conundrum but to assign to the supervisor a limited but meaningful degree of regulatory powers. The difficulty, of course, is where to draw the line. I suspect that part of the reluctance in allowing supervisors more leeway in the regulatory field stems from the fact that the goal of supervision is not well defined. When the goal is vague, independence can lead to arbitrariness. This leads me to the next point I would like to address.
Article 1 of the SSM Regulation stipulates that the ECB is assigned supervisory tasks “…with a view to contributing to the safety and soundness of credit institutions and the stability of the financial system…”. This definition is rather common; other authorities have it with some variations and the Basel Committee also refers to “safety and soundness” in the first of its Core Principles.
I think supervision may benefit from a bit more clarity on what “safe and sound” exactly means. It cannot mean that banks should be riskless: a bank is inherently risky and it would not perform its function if it wasn’t. However, one may appeal to the basic notion in finance that states that the risk of a portfolio in equilibrium under risk aversion is directly related to its return. This principle must be true for bank balance sheets as well. One could think of safety and soundness as a situation in which the bank moves along an efficient frontier between profitability and risk. The supervisor would then be asked to ensure that, for any given profitability level, the overall risk of the bank is minimised.
But who should choose the desired point in this efficient trade-off? In other words, who is entitled to decide the level of risk that society as a whole considers acceptable for its banking sector? Being linked to collective preferences, it seems logical to entrust this choice to the political level. I am not aware, however, of any country where the supervisor’s delegation arrangements are designed in this way. At present, aggregate bank risk levels in our democratic societies are chosen by an opaque mix of legislature (for the Pillar 1 component), global fora, treasuries and central banks (who together compose the “delegated authorities” which set the macroprudential buffers), and supervisors.
One obstacle to this approach is the difficulty of measuring bank risk. Insufficient research has been devoted to this problem. Central bankers have spent much more time analysing price stability than supervisors have spent researching bank risk, and the same is true for the academics engaged in the respective fields. I have noted that supervisors are much less inclined than central bankers to analyse in depth what they do and why. They say they have no time for that, but there is a cost to this unwillingness.
I find it difficult to imagine supervisory independence becoming firmly established and respected until the objective of banking supervision is defined more precisely. Here are some thoughts on how that could be done:
- Market-based risk measures are plentiful, for example historical volatilities, yield spreads, or derivative-based implied volatilities or bank credit default swaps. These measures, however, must be handled with care, because they do not measure actual risk, but perceived risk. Markets can be wrong, and targeting market prices injects volatility into the policy process. Market-based indicators can at best provide elements in a combination of several indicators.
- Quantity-based indicators, such as capital and liquidity ratios, maturity mismatches and NPLs, are more objective. However, they are more instruments than objectives. They are also inherently multidimensional: it is difficult, for example, to weigh capital against liquidity in defining the extent to which a bank is “safe and sound”.
- A quantification of the goal of supervision should not ignore the banks’ ultimate purpose, which is to transform liquid savings into sound credit to the economy. I have noted that this element is never considered explicitly in supervisory decisions; however, in deciding how much risk a bank is allowed to take, it is not irrelevant whether the bank is granting good credit to good borrowers or not. Candidate measures are the volume of intermediation, its quality and its price, the latter measured by intermediation spreads, corrected for cyclical effects. Naturally, this dimension suggests that macroprudential elements should be considered as well.
My overview on supervisory independence would be incomplete if it didn’t mention the challenge posed by politics. I refer in particular to the progressive strengthening, in Europe and elsewhere, of political forces that appeal directly to the “will of the people” and that call for restoring national sovereignty at the expense of international cooperation. Definitions and labels vary, ranging from populism to sovereignism, nationalism, and so on.
Populists typically regard the will of the people as being opposed to the interests of banks. National sovereignty seemingly contradicts the logic of the banking union, which implies transferring certain policy functions to the supranational level. Equally important, supervisory independence deviates from direct democracy, which populists favour.
What the populist vision fails to recognise, in my view, is that there is more coincidence between the interests of the people and the purpose of a well-intentioned supervisor than meets the eye. The question here is: who are the people? I cannot think of a better equivalent than the combined populations of bank depositors and taxpayers. Most citizens have bank deposits, and all citizens are taxpayers, at least insofar as they contribute to indirect taxes. Banking supervision is fundamentally about ensuring that depositors’ funds are safe and that taxpayers are not asked to foot the bill for bank failures they are not responsible for. One can go even further and say that most, if not all, supervisory actions aim to protect citizens at large from a variety of special interests. My experience over the past five years tells me that a European institution is better placed to provide this type of protection than national authorities which are, institutionally and physically, closer to the origins of those special interests.
A more insidious line of attack maintains that entrusting public policies to independent agencies represents a breach of democratic norms. According to some, the “technocratic slide” has gone too far in our liberal-democratic societies, to the point of becoming a risk. Central banks are cited as an example of unelected power with dubious legitimacy, and supervision falls in this category as well.
This critique is important and deserves a short digression. I am not convinced that the asserted risk outweighs the opposite one, of exposing certain policy functions – not only banking regulation, but central banking, competition, financial markets, the judiciary, and others – to daily political influence. Even before Tocqueville wrote about the “tyranny of the majority”, the founding fathers of the American constitution warned about the risks of excessive recourse to majority rule. Political equality consists of much more than assigning every citizen a vote (or multiple tweets!) and counting them. We should also never forget that policy decisions affect many more people than those who decide under majority rule. The voting minorities, the non-voters, the very young, the unborn, all have a natural right to be represented.
Is delegation to independent institutions, within boundaries set by statutes and accountability provisions, the best way to provide such representation? This debate is open. As one expert on democracy might have said, it could still be the worst system, after all the others.
When I started out at the ECB in 1998, the last thing I expected was that I would end my journey here talking about banking supervision. But here I am; even as an economist (or even worse: a macroeconomist) I became involved and very interested in banking supervision. A macroeconomist should never specialise too much, but I have learned along the way that one can retain a broad vision even when engaging in special responsibilities.
My reflection on supervisory independence leads me to think that supervision is at the crossroads: it can grow and consolidate or it can disappear, at least in the form we know. The essence of banking supervision lies in the scope it has to independently exercise its Pillar 2 powers. If this scope narrows, so does the role and the relevance of banking supervision. We can, of course, think of a world without Pillar 2: all requirements would be set by law, perhaps at a slightly higher level than they are now, and would apply to all without distinction. The supervisor would be like a street cop, checking compliance and imposing fines. I hope supervision does not become like this. Not only would it be uninteresting, it would be inadequate to address today’s and tomorrow’s banking reality, where rapid change and increasing integration with the rest of finance and with modern technologies require a good deal of flexibility and judgement.
Alternatively, the Pillar 2 powers can be better established and strengthened. To do so, supervisory independence needs to be preserved and enhanced. Supervisors should be allowed to enter the regulatory field in a controlled way. This requires the boundaries and the goals of supervision to be better defined. The supervisors themselves should take an active part in this reflection, shaping up their analytics and overcoming their natural shyness.
I remember that in my early years at the Banca d’Italia, an institution responsible at the time for both monetary policy and supervision, with the internal rivalries that you would expect, one colleague used to say, only half-jokingly: “microsupervision is for microbrains”. I disagree with that statement. Perhaps it should be amended as follows: banking supervision is for big brains, which, however, should exploit their full potential.
Thank you for your attention.