Glenn Stevens: Financial regulation - some observations

Speech by Mr Glenn Stevens, Governor of the Reserve Bank of Australia, to the Federal Reserve Bank of San Francisco's Symposium on Asian Banking and Finance, San Francisco, 10 June 2014.

The views expressed in this speech are those of the speaker and not the view of the BIS.

Central bank speech  | 
10 June 2014
PDF version
 |  7 pages

I thank Alexandra Rush, Carl Schwartz and Penny Smith for assistance in compiling these remarks.


The Federal Reserve Bank of San Francisco has long looked to the west, to the Pacific Basin. The Center for Pacific Basin Studies was established here nearly 25 years ago, and it was a professional highlight for me to be a visiting scholar here at the time. It is a particular pleasure to return to San Francisco to take part in this Symposium on Asian Banking and Finance.

Regulation and the Asian jurisdictions

It is apt that the Symposium should focus on some of the regulatory challenges we are collectively seeking to address in the wake of the crisis of 2007-2008, and particularly welcome that it gives voice to partners from the Asian time zone.

More than ever, finance has been global over the past couple of decades. That has brought many benefits but also certain vulnerabilities. It has also brought challenges for the regulatory community. There is a prodigious effort underway to work together to produce a regulatory framework that is genuinely international. Yet many important responses to the crisis are, and will remain, national. As such, they are at least partly driven by the circumstances and imperatives of the nations concerned. Nothing else could realistically be expected, but ensuring that all those national responses dovetail into a coherent international framework, and one that preserves what is good about globalised finance, is a difficult job.

Moreover, the economic and financial importance of the Asian region is greater now than on past occasions when international regulatory standards were put together. So engaging Asia in the search for genuinely global approaches to regulation is important. To that end, the Financial Stability Board (FSB), and some of the standard-setting bodies that attend it, have adjusted their membership over the past five years in a way that recognises the importance of the Asian countries (and those of some other regions too). The FSB is currently undertaking a further review of the structure of its representation.

It is nonetheless sometimes the case that the Asian members, and perhaps some other emerging market members, cannot avoid the feeling that the agenda is still driven by the major advanced economies. To the extent that those countries - the G7, say - are used to working together, and given that the financial crisis was so strongly centred in the north Atlantic countries' financial systems, perhaps that is no big surprise. However, while that historical cooperation is a strength on which to build, it is also important that the system we build looks forward, and acknowledges that much of the future growth will be in Asia. It is also important that those of us in the Asian time zone continue to strengthen our capacity to engage effectively in the international groups of which we now have the privilege and responsibility of membership.

Major themes in regulation

There are many components to the international agenda for regulatory reform. The G20 is the body that has accepted, or perhaps asserted, high-level oversight over these efforts. The FSB is the body that has accepted the task of trying to ensure a coordinated effort, respecting the integrity and independence of the standard-setting bodies and addressing risks itself where required.

As you know, Australia has the responsibility of the G20 presidency this year. Australia is highly supportive of the way the Chair of the FSB has structured the efforts around four key themes. They are:

  • increasing the resilience of financial institutions, which in the main means implementation of the Basel III standards for banks
  • reforming markets for trading, settling and clearing derivatives
  • addressing risks to the financial system from certain types of "shadow banking"
  • addressing the "too big to fail" problem at a global level.

Highlighting these four themes doesn't mean other things are forgotten. But it is an attempt to prioritise, to focus energies and to use the opportunity of the Leaders Summit in Brisbane, Australia, in November as a focal point for our efforts to get some important things across the line this year.

Basel III

The Basel III process is well on track. Most countries are well placed to implement the new international capital and liquidity standards in line with agreed timetables. Remaining policy details are being ironed out, with the leverage ratio requirement recently finalised, and further proposals on the net stable funding ratio released by the Basel Committee on Banking Supervision (BCBS). The Committee is continuing its important work to address the excessive variability in the calculation of risk-weighted assets.

On the question of capital standards, it is frequently claimed that pressing major banks to hold more capital will impair economic growth. There are two points to make about this.

The first is that, in situations of acute capital deficiency, it is best to address the problem as quickly as possible. Consider the contrast between the United States and Europe over the past five years. In 2009, US authorities conducted a public stress-test exercise on the balance sheets of major US banks and insisted that, where the tests revealed weaknesses, banks strengthen capital positions. They could carry out this exercise confident that a backstop capital delivery mechanism, in the form of public injections of capital, was available if private capital was not forthcoming.

But it has taken much longer to get to this position in Europe. There were stress tests in earlier years but they do not seem to have had the same credibility in the eyes of markets as those undertaken by the Fed in 2009. Moreover, Europe did not have a European public capital delivery device; it was left to individual national governments to address any need for that.

It is apparent that, while American banks carried out balance sheet repair early, and have for some time been in a position to help the US economy in its recovery, European banks have, in aggregate, continued to seek to deleverage by lowering their risk-weighted assets. This cannot have helped the euro area's growth prospects.1

This is being rectified. Later this year the ECB will release the results of its Asset Quality Review, which should be a highly credible exercise. And Europe is in the process of building a single supervisory mechanism for large banks. Over time, though rather a long time, Europe is also committed to creating a single resolution capability and a Europe-wide capacity for injecting capital from a single resolution fund if needed. These are all important steps but the respective US and European experiences do, I think, show that prompt efforts to fix serious problems of capital deficiency are ultimately pro-growth.

What about in more stable times? Can banks with more capital support growth as well as those with less capital? Equity is more expensive than debt for banks, and so a financial structure with more equity does mean, other things equal, that the cost of intermediation to the community is higher. Even without appealing to propositions of a Modigliani-Miller kind, which would dispute this assumption, the question is whether this apparently higher cost is a serious impediment to growth. Bankers often claim it is. The empirical estimates published by the FSB and BCBS suggest the effect is small, particularly when compared with the costs of large financial crises.2

Stepping back from that debate for a moment, and this is my second point, we might reflect on the following question: did the highly leveraged expansion of some parts of the financial sector in the period prior to the crisis really add much, sustainably, to growth?

It is far from obvious that it did. It seems more likely, to me, that it was the other way around: a period of good global growth and, in particular, unusually stable growth, led to a rise in leverage. The reasons for that growth stability were mainly not, I suspect, things that happened in the financial sector. But people concluded that macroeconomic stability meant that higher leverage was safer than before. In acting on that conclusion, they inadvertently sowed the seeds of instability, since leveraged balance sheets left borrowers and lenders more exposed when an adverse shock occurred.

The feedback effects from economic growth, or lack of it, to the capital positions of financial institutions are powerful. The big question is not, in fact, what more demanding capital standards will do to economic growth. The question is: what will economic growth, or lack of it, do to banks' capital positions?

It is noteworthy in this context that the phase-in for the Basel III capital standards extends until 2019. One has to ask: how likely is it that we will go another five years without an economic downturn of some kind? Even if we do, even if we assume that growth in the United States and Europe extends to the end of this decade, by that stage these expansions would be fairly mature.

Given that, one would hope that by 2019 major financial institutions would not only have reached new international minima for capital, but would have risen above them. One would hope that balance sheets by that time would be at their strongest position for the cycle. This is a reason to go faster, rather than slower, in accumulating capital to higher minima, while one can. This point is of some relevance to the discussion in my own country at present.


On the second core regulatory theme of derivatives, efforts to achieve more reporting, more platform trading and more central clearing are running behind original timetables. That is in large part due to the complexity of the issues involved in crafting mutually consistent sets of regulations at the jurisdiction level, for a market that is highly globalised in operation. The regulations of the major jurisdictions - the United States and the European Union - have considerable extraterritorial effects. It is critical that there be a high degree of consistency if we are to avoid fragmentation and unnecessary cost. The smaller jurisdictions, such as my own, feel this acutely, but it is not just a concern for the smaller jurisdictions. Even large jurisdictions would bear costs from unnecessary fragmentation.

What is needed is a good understanding in the various jurisdictions of each other's regulatory arrangements, and of the areas in which they can give recognition to those arrangements. But it takes more than just verbal commitments to the concept of mutual recognition to achieve this. It requires confidence that the application and enforcement of sets of rules in other jurisdictions, which have the same intent as one's own but which differ in precise wording or form, will in fact produce broadly equivalent outcomes. It takes time to establish that confidence, based on careful analysis, lengthy discussion and building of trust.

Progress is now being achieved here, including between the two biggest jurisdictions. And the efforts of the authorities in the United States and the European Union to assess the regulatory arrangements in the various other jurisdictions, which is no small task, must be acknowledged. But we have a way to go yet. Progress could be improved if relevant international standards were the benchmark for assessments of regulatory equivalence, and if regulators clearly communicated the basis of their decisions and provided regulatory certainty by clarifying the scope for transitional relief in a timely manner.

Shadow banking

In respect of the third theme, there have been considerable efforts by the FSB and the standard-setting bodies to address the risks posed by shadow banking - those entities and activities beyond the perimeter of prudential regulation. As the crisis showed, this sector can be a source of systemic risk, especially in the jurisdictions where shadow banks account for a relatively large share of the financial system or are important in credit intermediation. Authorities also have to be alert to any new build-up in shadow banking risk resulting from the likely migration of some activities in response to tougher prudential regulation of banks. However, a balance needs to be struck. The jurisdictions where these problems are likely to be serious are few, and we do not want efforts to contain shadow banking activity unnecessarily to stifle genuinely productive intermediation and innovation.

The FSB, International Organization of Securities Commissions (IOSCO) and the BCBS have worked on addressing risks in several areas, such as money market funds, securitisation, banks' links with shadow banks and securities financing transactions (SFTs). Broad policy frameworks have largely been finalised in these areas, with work ongoing in relation to SFTs (such as an assessment of the impact of proposed recommendations). It has been important to retain a degree of flexibility in the policy recommendations, given that some shadow banking markets are relatively small and not likely to be a major source of systemic risk.

Too big to fail

On the fourth area, addressing the problem of "too big to fail" entities, a key area of work this year is to put forward a proposal for "gone-concern loss-absorbing capacity", or "GLAC", for global systemically important banks (G-SIBs). The idea is that, even though additional capital surcharges above the (now higher) standard Basel III minima make it less likely that a G-SIB could fail, they do not make it impossible. Such entities are sufficiently large and interconnected that an uncontrolled failure could easily cause systemic disruption. Therefore, it is argued that further loss-absorbing capacity is needed, to be called on at the point of non-viability, so as to allow vital functions to continue and non-critical operations to be wound down in a controlled way. This limits adverse spillovers to the system and the economy. Generally, this loss-absorbing capacity is to come from a "bail-in" of certain classes of private creditors, so as to avoid calling on the public purse for a "bail-out".

This is an appealing idea, though it comes with the caveat that, to my knowledge, it has not successfully been done for a major institution to date.3

There are several important issues to resolve in the process of crafting a proposal. Among them are the questions of what instruments count towards any GLAC, how much GLAC may be appropriate, and what measures can be taken to mitigate risks of contagion from GLAC holders bearing losses.

On the matter of what instruments should count, a key question is whether or not equity capital above regulatory minima should qualify.

The view against doing so is that equity "buffers" may turn out to be illusory in a stress situation. That is, the uncertainty over asset valuations may be such that a presumed equity buffer is not, in fact, there. The "illusory capital" problem is certainly not unknown in the annals of crisis management.

The alternative view, which, it is fair to say, is more widely held, is that not to count equity above minimum requirements as GLAC would act as a disincentive for banks to hold or maintain higher than minimum equity buffers. From a risk perspective, higher equity buffers are unequivocally a good thing. Such a disincentive would therefore be unhelpful. This view is held quite strongly in Asia where banks tend to have high levels of equity capital. Under this approach, concerns about "illusory capital" could be addressed by enforceable triggers to allow early intervention when excess capital runs down to a point that breaches the GLAC requirement.

Moreover, the more genuine equity a bank has, the more likely it can remain a going concern in the face of a given shock. It would still need to have a degree of recapitalisation, by a share issue, and perhaps would need to call on contingent capital instruments, triggered by some sort of capital threshold. But that would be a recapitalisation as a going concern, not a resolution as a gone concern. One would have to have a lot of confidence in a resolution regime not to prefer working with an injured but still living G-SIB, as opposed to resolving a "gone concern" one.

There could still be a tail event big enough to exhaust even a higher level of equity in a G-SIB, therefore requiring resolution. And we might be uncomfortable about how quickly equity might be burned through in such an event, and we do want to be sure we can resolve a G-SIB. So some GLAC is helpful in that situation. Unless we have draconian capital standards for equity, some requirement for bail-in debt in addition to equity may well be sensible for G-SIBs. But overall it is appropriate, in my judgement, to allow equity capital in excess of regulatory minima to be counted towards GLAC requirements.

Another important consideration is to mitigate the risk that imposing losses on holders of GLAC will trigger destabilising contagion among investors. Measures to protect against this risk could include: the subordination of all GLAC liabilities, to reduce creditor uncertainty about the position of GLAC liabilities in the hierarchy of claims; limitations on the term of GLAC liabilities, to guard against the potential for runs on GLAC; and measures to limit cross-holdings of GLAC by other financial institutions, to reduce the prospect of losses on GLAC holdings significantly weakening other parts of the financial system. We have seen that in times of uncertainty financial markets often shoot first and ask questions later, so the task of protecting against contagion should not be underestimated.

These and other issues on GLAC will be debated over the coming months, with the intent to put a proposal to the G20 Leaders Summit in Brisbane in November, which would then go to public consultation and a Quantitative Impact Study.

Other elements in addressing "too big to fail" come under the heading of international cooperation, of a kind that goes beyond high-level statements of good intentions.

By definition, the smooth resolution of a global systemically important bank would have to be an internationally cooperative one. Hence cross-border cooperation agreements, in which home and key host jurisdictions accept and agree to be bound by a framework setting out their obligations and rights, would be central. Settling these agreements is running behind the timetable originally envisaged, partly reflecting existing statutory hurdles to information sharing.

Recognition of resolution actions taken in another country in respect of a G-SIB, which may involve stays of various kinds, would also be key. This will require, among other things, legislation in a number of jurisdictions. More generally, legislatures need to ensure that the relevant regulators have a mandate to cooperate effectively and share information with their counterparts elsewhere in the world and the legal tools to take part in relevant resolution actions.

Limits to regulation?

So there are a number of issues on which we need to make further progress by the time of the G20 Leaders Summit in November. It is to be emphasised that the work ahead is mainly not so much new regulation as the fulfilment of commitments already made, and made some time ago.

What about beyond that? In particular, at what point does the agenda for better, and frankly more, regulation reach a natural conclusion? When does our focus shift from regulatory redesign, under the duress of a crisis, to operating the system in (hopefully) more normal times? And to evaluating how the strengthened regulations are working?

I'm sure that is a question the regulated institutions ask. It is one the regulatory community asks as well.

Central to addressing the question is the need to remember what we are trying to achieve. Put simply, we are applying the lessons learned during the crisis about the extent and nature of risk, the set of incentives that allowed it to build up, the channels of contagion for distress once the extent of risk became apparent, and the weaknesses in our collective capacity to manage a crisis when one emerged.

This does not equate to a desire to eliminate all risk. In fact, risk-taking is, to a point, good. We want it to occur. That is how society advances. A problem right now, arguably, is that there is not enough genuine entrepreneurial risk occurring - judging by the low levels of private capital spending in many advanced countries. (That is a different thing from saying that there is not enough financial risk-taking occurring - various people argue that, in some respects, there is too much of that.)

We do need, however, risk to be recognised, and we need to be clear about who bears it.

Standard capital requirements are about ensuring shareholders accept the risks run by the firms they own. With appropriate capital structures, shareholders in financial institutions should be able to expect a reasonable, risk-adjusted, return - remembering that banking is a leveraged business and that when things go wrong the shareholders' return can quickly become a large negative. With additional loss absorbency requirements for G-SIBs, we are asking shareholders to internalise some of the externalities very large and complex firms create for the system in the event of distress. And with the GLAC requirements, we are asking some of the other creditors to accept the risk that they may, in the face of a truly exceptional event, in effect become equity holders as part of a resolution of a G-SIB.

But the risk/funding cost curve is not linear, at least not over its entire length. At some point requiring more and more loss absorbency to be provided by the private sector will come at an increasingly steep price. There must be events sufficiently far into the tails of the relevant distributions that no private balance sheet can reasonably be expected, ex ante, to bear the associated loss, no matter what the price. Should such an event occur, policymakers will still face the decision of whether to close or support the institution. The policy task is not, in my view, one of ensuring that the probability of such an event is absolutely zero, but of making it acceptably low at an acceptable price.

Most particularly, we need to prevent a recurrence of a situation in which outcomes that are not, in fact, especially remote put the system in need of a public rescue.

It equally follows that we are not trying to extend the regulatory "perimeter" indefinitely. There will always be some risky activity around the fringes of the system, and there is nothing particularly wrong with that. Those who seek high returns, and are prepared to accept the risk, should be allowed to do so. There is value in that occurring.

The proviso is that as well as allowing such risk-takers to accept the rewards of their activities, we have to be able safely to allow them to wear the losses. We cannot have them generating significant spillovers to the parts of the financial sector that are publicly underwritten or that are key to the day-to-day functioning of the economy. Where such spillovers do exist, rules that serve to isolate the risk-taking activities, to keep them genuinely on the fringe, are in order.

This is perhaps the greatest regulatory challenge for the future: assessing when an activity that is technically outside the "perimeter" might be about to present a threat to overall stability. Devoting a lot of resources to ever greater refinements to the details of existing regulatory structures will not help us do that. Having the big things inside the perimeter about right should be good enough. After that, we need to make sure we devote adequate resources to keeping a general weather eye on the broader situation, beyond just the area illuminated around our current lamp post.

With the pace of technological change, it probably will not become easier to do that. The possible rise of "virtual" currencies, the potential for the distinction between regulated financial institutions and, say, telecommunications and technology companies to become blurred, to name just two developments, may pose challenges. They doubtless will not be the only ones. It probably will still pay dividends, though, to keep our antennae tuned for the nature of the promises made by the various players, the extent of leverage involved and the frequency with which we hear the cry that "this time is different".

Thank you for your attention. I wish you well in your deliberations.

1 This is not a new lesson: it was already evident from Japan's experience in the 1990s.

2 See BCBS (2010), "An Assessment of the Long-term Economic Impact of Stronger Capital and Liquidity Requirements", August. Available at <>; and BCBS and FSB (2010), "Assessing the Macroeconomic Impact of the Transition to Stronger Capital and Liquidity Requirements", Final Report, December. Available at <>.

3 There are examples from the recent crisis, and previous ones, where large banks have been resolved by imposing losses on their shareholders, junior bondholders and, in some cases, their senior bondholders. I am not aware, however, of any case where a bank returned from resolution as a privately owned entity, trading under its original name, having been recapitalised by private creditors alone.