Erkki Liikanen: How to improve financial stability and resilience of systemically important financial institutions after the crisis?

Speech by Mr Erkki Liikanen, Governor of the Bank of Finland and Chairman of the Highlevel Expert Group on reforming the structure of the EU banking sector, at the EBA Policy Research Workshop "How to regulate and resolve systemically important banks", London, 14 November 2013.

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

Central bank speech  | 
19 November 2013

Accompanying slides can be found on the Bank of Finland's website: Slides

[Slide 2 - Outline]

[Slide 3 - Heeding the lessons of the crisis]

Heeding the lessons of the crisis

[Slide 4 - Substantial crisis impact on the real economy]

The vast scale of the Global Financial Crisis and its ramifications have been the main motivator of the regulatory reform agenda; endorsed by the G-20.

When the crisis began, the resulting economic downturn for a while resembled the pace of developments in the Great Depression.1 Only the liquidity support operations by central banks at different stages, and the explicit state guarantees to banks that were announced in different countries, changed the course of developments, compared to the 1930's.

Paradoxically, the success in averting the worst that could have happened, may be making some of us forget the lessons from the crisis too soon, or having us downplay the need for true reforms.

However, it is crucial not to forget the crisis, and to properly implement the reform agenda.

Have we done enough to solve the too-big-to-fail problem?

In response to the crisis, a number of international and EU wide regulatory reforms have been launched.

The capital requirement framework has been strengthened, new liquidity requirements have been proposed and the recovery and resolution directive will enable orderly "failure" of banks.

Now we need to assess whether we have done enough? I will argue that the regulatory reforms have taken us a long way, but not far enough. Banks will still be too complex to manage, supervise and resolve.

Below I will argue that structural reforms are an essential complement to the new resolution regime, because structural reforms are meant to reduce banks' complexity and interconnectedness. We need these reforms to make orderly resolution of systemically important financial institutions a credible option, even in the midst of a systemic crisis. Credibility of the resolution regime will be key to truly end the too-big-to-fail problem and get market discipline back to work.

Before exploring these arguments further, let me start by reviewing the main lessons of the crisis a bit further.

Excessive financial expansion in the run-up to the crisis

[Slide 5 - Rapid growth in the EU banking sector]

One important lesson from the crisis has been that we should better understand the point when and why the expansion of the financial sector turns excessive. Some recent research suggests that an excessively large financial sector may even turn from being a driver of additional economic growth to being an impediment to growth.2

Surely, large-scale financial crises are very bad for the long-run accumulation of economic output, even if some crises might have a silver lining in the sense of promoting necessary "creative destruction".3 It is essential that banks with unsustainable strategies are allowed to fail in an orderly fashion so that viable banks may compete and prosper on a level playing field.

Especially when the financial sector grows rapidly into large proportions, the risk of a serious crisis is heightened.4

[Slide 6 - Shifts in focus of operations as illustrated by shifts in assets structures]

Focus on trading-related activities, the rise of new complicated financial products, and increasing reliance on short-term market funding by financial institutions are symptoms of growing risks, as we saw in the years before the crisis.

This time, the ultimate risk stemmed largely from the US housing market, but it had been transferred also onto European banks' balance sheets.5 This took place with the help of securitizations and the related new financial intermediaries, generally dubbed as the shadow banking sector.

As a result, the number of counterparty relationships and the ensuing counterparty risks increased. The global financial network became highly interconnected and complex.

It is evident that many of these developments prior to the crisis did not benefit the real economy in a sustainable manner. Economic recovery from the crisis has been slow.6

The benefits of risk-taking were private whereas the enormous costs of the crisis became the responsibility of the society as a whole.

Regulatory and macroprudential failures and problems with management

[Slide 7 - Increased leverage as illustrated by shifts in funding structures]

Why did the financial sector and its risks grow so much? And why were they funded by increasing short-term leverage which added to the vulnerability?

There were global macroeconomic reasons behind the crisis developments, but also failures of regulatory and macroprudential policies.

Global imbalances provided a fertile ground for the development of financial excesses. Moreover, the monetary policy stance was accommodative in much of the early 2000.

There may also have been a false sense of security resulting from the successful track record of monetary and macroeconomic policies.7 As a result, the markets placed their trust in central banks' ability to take care of liquidity if markets were to be hit by serious shocks.

These expectations may also have contributed to imprudent risk-taking which, once the crisis started unfolding, led to real losses.

These experiences underline all the more the need to develop macroprudential supervision and policy. Macroprudential tools have to be applied when necessary to complement other policy areas in curbing excesses in the financial sector.

Banks' capital requirements and liquidity buffers did not keep pace with the growing risks.

Serious short-comings in risk measurement within the Basel framework became apparent once the crisis hit. This was symptomatic in the slower growth of risk-weighted assets relative to total assets in the years preceding the crisis.

There was also a lack of sufficient systemic view of financial risks. Macroprudential tools and mandates to deal with such risks were largely missing.8

Another important reason concerns market expectations of public support if a large-scale crisis hits. Leading European banks are very large in an international comparison, especially in relation to the size of their home countries.

[Slide 8 - Implicit subsidies remain high]

As a result, many large and complex financial institutions appear to have benefited from cheap funding, thanks to such too-big-to-fail expectations.9

Cheap funding which is insensitive to balance sheet risks may easily spur risk-taking, especially when banking competition gets tougher. Sheer ignorance by bank managements of the true risks may also have played a role.10

[Slide 9 - Market discipline vs. structure of banks - a "chicken-and-egg" problem to be solved]

The too-big-to-fail assumption weakened market discipline, as did the opacity of large, complex financial institutions. As a result the banking sector has not been renewed or restructured in a way that might have happened in a world with efficient market discipline. Inefficient business models have not been outdated in the same way as they have been done in the corporate sector, where the conglomeration trend of the 1980s reversed already some time ago. The market indications of a diversification discount appear to be too weak.11

The challenge at hand is thus to implement regulatory reforms and policies which help steer the financial sector and banks towards a size and structure, which in a better way support economic growth while maintaining financial stability. No one really knows what the right size of the financial sector or banks is. But at least what we can do is to remove any perverse incentives which could lead to an excessive growth of the sector and its risks. Eliminating the too-big-to-fail assumption by making resolution of systemically important financial institutions a credible option lies at the core of the solution.

Finally, there was also problems in the management culture of banks. The culture of the customer focused retail and commercial banking is inherently different from the one of the transaction based trading oriented banking.12 The transaction-oriented management and compensation culture of trading spread to more traditional banking. This provided incentives to risk-taking.

The former Citigroup architect, Mr. John Reed, has recently described how this "cultural mixing" had an impact on risk-taking. In traditional banking risk officers might easily say "no" if they judged a proposed new loan not good enough. However, if the loan officer's pay is linked to making new loans, as increasingly happened, then saying "no" becomes much more difficult in practice.

Last days events and currency exchange manipulation incidents have once more brought the issue of corporate culture of banks to the surface.

[Slide 10 - Policy reforms]

Regulatory and macroprudential policy reforms

[Slide 11 - Financial market liberalisation and governance must be complements]

Before the crisis the benefits of deeper financial integration were pursued via market liberalization. Improving the governance of financial markets was lagging behind, leaving the system vulnerable to crises.

We now understand that liberalization and governance are complements. We need them both to reap the benefits of financial integration in a sustainable manner.

[Slide 12 - Focus of current regulatory reforms]

Much has already been done to improve financial stability and resilience of financial institutions after the crisis, particularly of those seen as too systemically important to fail.

In an ideal world the externality imposed on the society by large systemically important banks might be dealt with by a single optimally designed regulatory instrument. However, in the real world subject to substantial uncertainties concerning the nature of risks and their measurement, it is better to rely on several regulatory instruments.13

First, Basel III agreement which is implemented here in Europe through the capital requirement regulation and directive will strengthen banks' capital and liquidity positions. This is a major step to improve banks' resilience, but will also reduce incentives to take excessive risks across different banking operations. Basel III also tackles the too-big-to-fail problem by imposing an extra capital requirement on international, systemically important financial institutions.

Liquidity requirements which previously did not exist bring additional important resilience as liquidity typically dries up when a systemic crisis starts.

Note that capital requirements also help maintain liquidity in the markets: when banks have higher capital buffers to absorb losses, there will be less need for fire sales of assets. Fire sales seriously undermined market liquidity during the crisis.

Second, new recovery and resolution tools have been constructed. The key instrument is to make it possible to bail-in bank debt in a crisis, in order to avoid bail-outs with taxpayer money. Moreover, the resolution authority is given great powers to restructure a bank when the point of likely to fail has been reached.

This will be a key in restoring private market discipline on banks, and thus to reduce the too-big-to-fail problem. The importance of implementing the Bank Recovery and Resolution Directive, upon which political agreement was achieved in the summer, cannot be overstated.

In Europe bank resolution (Single Resolution Mechanism, SRM) is part of the larger aim of forming a true banking union. The establishment of the SRM will facilitate the resolution of cross-border banks. The other element of the banking union is the Single Supervisory Mechanism in connection with the ECB.

The banking union project also includes a third element; the prospect of creating a common deposit insurance system. However, the first priority ought to be to harmonize the national systems, which currently vary greatly.

There are also a number of regulatory initiatives that intend to increase transparency in the financial markets. For example, accounting standards and disclosure practices are reviewed, and banks are urged to improve risk management and corporate governance practices.

Supervision and regulation of derivatives markets and the shadow banking system are revised, including moves to central counterparties which will help address the contagion risks embedded in long counterparty chains in the OTC markets.

[Slide 13 - Authorities should foster the stability of the entire financial system]

Macroprudential analysis and tools are being developed mainly on the national basis, but also in accordance with Basel III (and its application in Europe; the capital requirement directive) for the purpose of setting counter-cyclical capital buffers. Moreover, the European Systemic Risk Board has published guidelines on the implementation of macroprudential tools across the EU member states.

This is a highly important but challenging area of work where we will see a lot more research to be done.

The challenge is to find indicators which are able to warn early enough of looming vulnerabilities. Potential candidates are at least market prices, credit-to-GDP ratios, and changes in banking sector liabilities as emphasized by professor Shin of Princeton University.14

Finally, structural reforms of banks are already taking shape in some jurisdictions, such as the USA, the UK.

These were the elements of the regulatory reform which were on the table as we in the High-level expert group set out to evaluate the need for structural reform in the EU banking sector. The question we were encountered with was whether these regulatory reforms will take us far enough.

[Slide 14 - HLEG proposal for structural banking reform in Europe]

The work of the High-level expert group

In our deliberations we considered two avenues as possible ways forward. In the two avenues we put different emphasis on the most promising measures to end the too-big-to-fail problem; capital requirements, resolution regimes, and structural reform.

[Slide 15 - Two avenues as possible ways forward were considered]

In the first avenue, additional, non-risk-weighted capital requirements on trading activities and credible recovery and resolution plans for banks would have been the main instruments.

We acknowledged that the measurement of risks inherent in trading assets is prone to a significant "model risk". Robust capital requirements which do not rely on complicated models are one way to tackle this issue (as are limits on risk concentrations and counterparty exposures). Avenue 1 was based on this approach.

The possibility of structural measures did enter Avenue 1, but only as a conditional instrument. The idea was that if a bank was not able to prove that the required recovery and resolution plans were credible, separation of trading activities was to be imposed by authorities.

In the second avenue, by contrast, any significant trading activities would be required to be separated from retail deposit banking. The separation proposal outlined in Avenue 2 was based on the notion that capital requirements are not by themselves sufficient to limit excessive risk-taking incentives induced by deposit insurance if risks are difficult to measure and risk profiles can be changed rapidly, as in trading activities.

Avenue 2 also acknowledges the risk of heterogeneous application of tailor-made separation based on the credibility of the recovery and resolution plans of individual banks. Thus avenue 2 includes uniform separation ex ante in order to facilitate resolution of large and complex banks without public funds and hence reduce the too-big-to-fail problem.

Structural reform proposal for Europe

[Slide 16 - Rationale for mandatory separation]

Our Group came to the conclusion that although the regulatory reforms already implemented or in progress take us a major step forward, not enough has been done to eliminate the too-big-to-fail problem. Structural measures aimed at the largest banks and banks with significant trading activity are thus needed to complement them.15

I will also argue that structural reforms support the creation of the banking union.

More precisely, we found that the benefits of separation are the following.

Our Group's proposal focuses much on limiting the spillover of benefits from safety nets such as the deposit insurance to outside the core banking activities thus reducing the incentives for banks to become too-big-to-fail in the first place. Separation will restrict risk-taking opportunities of a deposit bank. The blending of the commercial and investment banking management cultures will also be reduced, which reduces the risk-taking incentives in traditional banking, evident in the years preceding the crisis.

Moreover, limiting the spillover of benefits from safety nets to outside the core banking activities will reduce risk-taking incentives and incentives for excessive growth in the trading entity. This is because the funding of the trading entity will be subject to more effective market discipline in the absence of insured deposits.

Reducing complexity by means of separation facilitates monitoring by outside stakeholders such as shareholders, bank creditors and other market participants, thus further reinforcing market discipline.

Reduced complexity by means of separation also has benefits for stakeholders closer to the bank. In the Group we argued that it would facilitate the task of the management and the board as steering effort to the right direction by means of incentive schemes, for example, is easier in a less complex organisation, where the organisational units are more homogeneous.

A more primitive benefit of separation is that reduced complexity makes it easier to understand the bank. This is a line of thought that the head of ECB's financial stability division, Ignazio Angeloni, aired recently by stating that "... very often [the CEO's] don't know what's going on, and they don't understand because it is very, very complicated".16

The supervisors face similar challenges. I think it is safe to say that the simplification of the structure of the largest, most complex European banks would facilitate the task, not only of the national supervisors, but also of the Single Supervisory Mechanism.

Finally and perhaps most importantly, I would once more like to emphasize that structural reforms are essential to make resolution of systemically important financial institution ex ante credible, even in the midst of a systemic crisis. If there is even a slightest doubt that the resolution tools will not be applied, the adverse implications of the too-big-to-fail assumption on market discipline, funding costs and risk-taking incentives will prevail.

A forthcoming Journal of Finance paper by Gandhi and Lustig show that a "tail-risk subsidy" can explain differences in equity returns of large and small banks.17 Interestingly this "tail-risk subsidy" grew dramatically as the Glass-Steagall was dismantled.

As previously noted, we have often seen that market discipline ultimately enforces structural changes which promote efficiency in non-financial industries. Unfortunately, the evidence suggests that the financial industry may be trapped in a "bad equilibrium" of weak market discipline and inefficient structures.

Hence a push from structural regulatory reforms is needed to move the financial industry to a "good equilibrium" with strong market discipline and efficient structures.

[Slide 17 - HLEG proposal]

Thus we in the Group proposed the separation of core commercial banking which is allowed to take insured deposits from trading and market making activities which are not allowed to take deposits.

Financial linkages between the separated entities need to be limited in order to make the separation effective.

Trading and the bulk of market making should be moved together to the separate entity. This is because it is very difficult to make the distinction between the two and any rules would be prone to circumvention.

Separation does not touch the universal banking model because the deposit bank and the trading entity are still allowed to be part of the same group. Moreover, mandatory separation would only apply to the largest institutions, consistent with the aim of tackling the too-big-to-fail problem.

[Slide 18 - Stylised classification of structural proposals: activities and strength]

On capital requirements

Another important issue which has been much discussed also recently is that regulations need to be robust enough. Robustness of capital requirements is needed because there will always be errors in risk measurement, and managerial incentives to manipulate risk-weights downwards will always be there.18

One answer is to add capital requirements which are not based on risk-weights, to complement the risk-based ones. An important example is the leverage ratio requirement of Basel III.

Our Group also considered, although did not directly propose, additional robust capital requirements on trading activities. If these were calibrated so that the robust requirements formed the bottom and the risk-based requirements came on top of that, then banks' incentives to keep on improving their risk measurement technologies might be better maintained.

The Basel Committee on Banking Supervision is still considering the merits of introducing the standardised approach as a floor or surcharge to the models-based approach.19 Of the proposals for how the market risk framework could be altered recently presented by the Basel Committee, I particularly welcome two key features. First, the suggestion of revising the standardised approach to be sufficiently risk-sensitive thus creating a credible alternative to the internal models. And second, the requirement of mandatory calculation and publication of the standardised approach by all banks, which would improve transparency and comparability.

Note that separation of trading-based risks from deposit banking can also be seen as a form of robust regulation. If potential errors in the measurement of trading risks may prove fatal to core banking services, it may be better to move the risks elsewhere altogether. This is what separation does: it compartmentalizes risks within the banking group, a bit like firewalls help restrict the spreading of fire.

A combination of structural reforms with capital and liquidity buffers may be a cost-effective way of regulation in the long-run.

Taking into account both public and private costs, bank equity capital is not as costly as one might deduce from banks' preference to use very high levels of debt financing. This point has been carefully scrutinized by Admati and Hellwig (2013).20

However, it is challenging to measure precisely how much bank equity would be socially optimal, although the crisis itself strongly suggests the levels of capital were far less than optimal before the crisis.

Hence, a combination of higher capital and liquidity requirements together with restrictions on deposit banks' riskiest activities might be a balanced choice of regulation.

Structural reforms may also contribute to successful macroprudential policies. This is because structural measures add simplicity and transparency of financial institutions, and reduce their interconnectedness. As a result, the identification of good macroprudential risk indicators becomes easier, and they can be measured with higher precision. The effectiveness of the respective macroprudential tools improves accordingly.

As a counter-argument to separation, some have argued that by separation we lose risk diversification benefits. However, the crisis itself suggests that we may have overestimated these benefits while underestimating the effects of risk contagion which separation helps to reduce.

Regarding banks' recovery and resolution, our group also made a proposal of additional designated bail-in instruments. These could be triggered in a relatively early phase of a bank's financial distress to provide for recovery. Their holding would be restricted to investors outside the banking sector in order to avoid contagion. Similar ideas have recently been aired also by others.21

In comparison with other structural proposals, our group's proposal settles in the middle in terms of narrowing down the activities of the deposit bank and in terms of the strength of separation.


[Slide 19 - To conclude: On the role of banking]

Banks serve central functions in a well-functioning economy. Their role in providing finance to the real sector is particularly important in Europe.

The regulatory reform agenda to prevent future banking crises is very ambitious, and it is essential to take it into completion. There are short-term adjustment costs to the banking sector, but as societies we must keep our focus on the long-run benefits.

As regards the potential structural reforms in Europe, good implementation will be the key to achieve the intended goals.

International coordination of the structural measures as called for by BIS and IMF would be highly welcome.

1 Rajan, R. (2013), A step in the dark: unconventional monetary policy after the crisis. Andrew Crockett Memorial Lecture, BIS, 23 June 2013.

2 Cecchetti and Kharoubi (2012), Reassessing the impact of finance on growth, BIS Working Paper 381.

3 Schumpeter (1994), "Capitalism, Socialism and Democracy", London.

4 See e.g. Claessens, Kose, Laeven, and Valencia (2013), "Understanding Financial Crises: Causes, Consequences, and Policy Responses", CAMA Working Paper 05/2013.

5 Shin (2012): "Global Banking Glut and Loan Risk Premium", IMF Economic Review, 60.

6 See e.g. Claessens, Kose and Terrones (2013), "The Recent Global Financial Crisis: How Similar? How Different? How Costly?", in S. Claessens et al., eds., Financial Crises, Consequences, and Policy Responses, forthcoming, IMF.

7 This refers to the era which is known as the Great Moderation.

8 See e.g. Borio (2003), "Towards a macroprudential framework for financial supervision and regulation?", CESifo Studies, 49(2) and the June 2013 IMF-report "Key aspects of macroprudential policy" on work to be done in the field.

9 Davies & Tracey (2014), "Too big to be efficient? The impact of implicit funding subsidies on scale economies in banking", Journal of Money, Credit and Banking (forthcoming).

10 Rajan, R. (2013), A step in the dark: unconventional monetary policy after the crisis. Andrew Crockett Memorial Lecture, BIS, 23 June 2013.

11 Laeven and Levine (2007), "Is there a diversification discount in financial conglomerates?", Journal of Financial Economics 85(2).

12 Boot and Ratnovski (2012), "Banking and trading", IMF Working Paper 12/238.

13 Cf. Richardson (2012), "Why the Volcker Rule is a useful tool for managing systemic risk?" A white paper. New York University Stern School of Business.

14 Shin (2013): "Procyclicality and the search for early warning indicators", in Claessens Kose, Laeven, and Valencia, eds., Financial Crises, Consequences, and Policy Responses, forthcoming, IMF.

15 Recall that the Group proposed an initial two-step threshold approach where the bank would advance to the second stage if its assets held for trading and available for sale exceed (1) a relative examination threshold of 15-25% of the bank's total assets or (2) an absolute examination threshold of EUR100bn.

16 ECB official prepares to perform triage on Europe's banks, interview with Ignazio Angeloni in International New York Times 21.10.2013.

17 Gandhi and Lustig, "Size Anomalies in U.S. Bank Stock Returns" doi: 10.1111/jofi.12119.

18 Cf. Admati, A. and M. Hellwig (2013): The bankers' new clothes: What's wrong with banking and what to do about it. Princeton University Press.

19 Basel Committee of Banking Supervision, Fundamental review of the trading book: A revised market risk framework, Consultative Document, October 2013.

20 They emphasize that their arguments do not require that the so called Modigliani-Miller capital structure irrelevance theorem holds literally. There are factors why banks have relative advantages from debt financing. But regardless of those, when bank debt to equity ratio (ie, leverage) increases, so does the risk of the holders of bank debt and equity (unless public support expectations distort this mechanism). Consequently, the risk premium on both bank debt and equity should go up. Conversely, reducing leverage as a result of higher equity capital requirements should make the risk premium on equity (and debt) decline.

21 See Paul Tucker's keynote in the Bank of Finland-SUERF conference in Helsinki, 13 June, 2013.