Dealing with financial systemic risk: the contribution of macroprudential policies

Panel remarks by Jaime Caruana, General Manager of the Bank for International Settlements, Central Bank of Turkey/G20 Conference on "Financial systemic risk", Istanbul, 27-28 September 2012.

Abstract

There are important two-way interactions between macroprudential policy and other areas of public policy. These interactions put a premium on cooperative institutional frameworks that recognise the complementarities between policy actions. This means that, within a single jurisdiction, macroprudential authorities should be independent and should focus primarily on mitigating systemic risk while recognising that other policies will have an impact on the same objective. Cooperation between macroprudential policies across national borders starts from the high level set by various international regulatory standards and is improving with the explicit macroprudential frameworks recently introduced for countercyclical capital buffers and the higher loss absorbency requirements for systemically important banks. Greater cooperation, however, does not mean that we should disregard that individual policies have specific objectives and that some hierarchy of action is necessary.

Full speech

Let me thank the Central Bank of the Republic of Turkey and the G20 presidency of Mexico for having invited me to attend such an interesting conference addressing the topic of financial systemic risk. In my remarks today, I would like to explain how macroprudential policies can greatly contribute to dealing with systemic risk and fostering financial stability. I will highlight a few key issues that we should focus on in order to make this effective.

1.    Trend towards strengthening the macroprudential orientation of policy

The term "macroprudential" has gained currency in policy discussions during the past four years. Indeed, the recent financial crisis has given rise to a general trend towards strengthening the macroprudential orientation of policy in countries with very diverse institutional frameworks and financial structures. This is very welcome: recent experience has taught us that we need to be more focused on addressing system-wide risk, and this is precisely what macroprudential policy is all about.

Macroprudential frameworks may be new, but mainly in the sense of becoming explicit. Many countries have been using prudential instruments to address system-wide vulnerabilities without making reference to macroprudential policies. For example, variable ceilings for loan-to-value (LTV) ratios have been used repeatedly in Hong Kong and other Asian economies to slow down frothy mortgage lending and ensure that banks do not overexpose themselves to property risk. Nevertheless, the more recent introduction of formal structures brings to the fore issues of definition, delineation of responsibilities and governance.

In my remarks today, I would like to underscore a critical aspect of macroprudential policy and to offer a word of caution.

  • The critical aspect I am referring to is the strong two-way interactions between macroprudential policy and other areas of public policy. These interactions put a premium on cooperative institutional frameworks that recognise the complementarities between policy actions, both within and across jurisdictions. This is a particularly important issue at the national level; but cooperative frameworks are also essential at the international level, requiring both sufficient information-sharing and reciprocity.
  • The word of caution is that we should be mindful that individual policies have specific primary objectives and that some hierarchy of action is necessary.

Let me explain.

 

2.    Macroprudential policy is not the only area of policy that influences systemic risk

Many other policies can affect the resilience of the financial system and its ability to provide valuable services to the economy. Quite apart from microprudential policy, the influence of monetary, fiscal and tax policies, of financial reporting standards and of legal frameworks is also very strong.

  • For instance, prolonged periods of low policy rates affect leverage, encourage financial market participants to take on risks and may at times fuel asset price bubbles.
  • Conversely, instruments and actions aimed at mitigating and managing systemic risk can have very important effects on the macroeconomy and thus impinge on the objective of other policies.
  • For example, tightening capital requirements to protect banks from the build-up of systemic risk during a credit boom can also cool down credit expansion and, by extension, aggregate demand.
  • To be sure, a more stable, more resilient and less procyclical financial system will improve the effectiveness of monetary and other policies.

So there are externalities in the interaction of different policies: there can be positive complementarities when the policies are mutually reinforcing, but also negative spillovers when one policy weakens the effectiveness of another. Hence, there is a need for coordination. This is true both within a given jurisdiction and across borders.

 

3.    The need for coordination within a single jurisdiction

Let me talk first about coordination within a single jurisdiction. The interactions between policies suggest a few principles for instrument design and deployment.

One such principle is that macroprudential policy instruments should be in the hands of an independent authority with the explicit objective of maintaining financial stability. This is important, for two reasons: the lack of precise measurement to quantify this objective, and policymakers' inevitable reliance on judgment in pursuing it.

Measurement presents a serious challenge for the design, governance and accountability of macroprudential policy. There are no readily available and widely accepted metrics of systemic risk to help calibrate instruments or gauge policy performance, even ex post, with much precision. And it is notoriously difficult to answer the counterfactual question of how things would have evolved had an alternative action plan been adopted. As a result, more than ever, policy needs to rely on a significant degree of judgment.

One telling example relates to anticyclical policies. All anticyclical policies have to work with real-time information that is incomplete and imprecise. Decisions rely on judgment to interpret the multitude of inputs. This is not unique to macroprudential policy, but it is particularly evident in this case: current technology provides far less in the way of robust quantitative models to guide macroprudential policy in addressing both the time and the cross-sectional dimensions of systemic risk.

As regards the time dimension, only recently have researchers been attempting to be specific about what the financial cycle is and how to characterise it. A few features are worth mentioning:

  • It is possible to identify a well defined financial cycle that is best characterised by the co-movement of medium-term cycles in credit and property prices.
  • Such financial cycles are longer and more severe than business cycles. The duration and amplitude of the financial cycle has increased since the mid-1980s: financial cycles last, on average, around 16 years; but when considering only cycles that peaked after 1998, the average duration is nearly 20 years, compared with 11 for previous ones.
  • Peaks in the financial cycle are closely associated with systemic banking crises (henceforth "financial crises" for short).
  • Finally, the financial cycle and the business cycle are different phenomena, but they are related. Recessions associated with financial disruptions tend to be longer and deeper.

As regards the cross-sectional dimension of systemic risk, we are also uncertain about how best to map the systemic importance of financial institutions onto their size, the extent and density of their links to others, and the uniqueness of their economic function. The need for judgment, combined with the need to resist powerful political economy pressures, puts a premium on operational independence. Pressures may be high because the future rewards of macroprudential policy actions tend to be uncertain, difficult to quantify and distant in the future, whereas the costs are immediate and can be easily exaggerated.

Operational independence is easier to achieve if the relevant authority has a clear mandate. And it has to go hand in hand with accountability and clarity of communication. Policymakers need to be transparent about how policy decisions relate to their mandate and to their economic assessments. This helps anchor the public's expectations and the holding of the authorities to account. From this perspective, it is key to ensure the adequate involvement of the central bank. One may even argue that it is preferable for the central bank to be the macroprudential authority.

A second principle is that the control over instruments should be commensurate with the objective of managing systemic risk. Not many tools are purely macroprudential. The vast majority are simply prudential tools tailored for use from a macroprudential perspective through adjustments to their design and calibration.

Capital requirements are a key tool but are not sufficient. They need to be complemented with other instruments and more intrusive supervision. Given that we have to deal with human behaviour that is imperfectly understood, combining instruments is more promising than relying on a single one. Liquidity requirements and instruments such as loan-to-value ratios or limits on exposures have all been used and proven effective. In addition, explicit resolution plans are also important: they address the source of the problem, as they reduce the costs of (disorderly) failure. More generally, all tools are inadequate in the absence of effective and at times intrusive supervision: the incentives for regulatory arbitrage are simply too powerful.

This means that there is a need for coordination in the use of various instruments, through both the sharing of information and the communication of assessments. Moreover, this should be supported by a framework that allocates responsibilities and accountability clearly.

 

4.    The international dimension

Let me now turn to the international dimension. As long as we have open financial systems, risks in one country can affect others. Similarly, macroprudential policy can have spillovers across borders. To what extent does this call for formal coordination?

Countries are developing their own policy frameworks to deal with the cross-sectional and the time dimensions of systemic risk. They are introducing arrangements to assess the banks that are systemically important from a domestic perspective. They are also introducing policy measures linked to rough indicators of banks' systemic significance. I would argue that, despite being the new kid on the block, macroprudential policy is one of the economic policy areas in which international coordination has gone furthest.

To be sure, we started from a very good basis, namely the existing international regulatory framework for markets and institutions. A number of independent international committees have proposed, and countries around the globe have adopted, minimum prudential standards for banks and market infrastructures. And, importantly, more recently there have been concerted efforts to promote consistent implementation across jurisdictions. The Basel Committee on Banking Supervision has conducted significant work in this area under the leadership of Stefan Ingves.

For my part, I would simply like to highlight two examples of coordination in the macroprudential area: as regards its time dimension, the design of the countercyclical capital buffers; and, as regards the cross-sectional dimension, the imposition of capital surcharges for systemically important banks.

  • The countercyclical buffer is intended to counterbalance the procyclical behaviour of banks by building up buffers in good times that can absorb losses in times of stress. It is a prudential instrument calibrated to achieve a macroprudential objective. Critically, the level of the buffer depends on the state of the financial cycle in a given jurisdiction. The framework allows for a large degree of judgment and tailoring to local circumstances - there is no one-size-fits-all solution. It also provides for international reciprocity: supervisors of foreign banks should apply the same surcharge on these banks' exposures as the supervisor in the host jurisdiction demands of the local banks. This levels the playing field and addresses regulatory arbitrage.
  • A similar degree of coordination applies to the treatment of systemically important banks. The Basel Committee and the Financial Stability Board have developed a framework to assess the banks that are globally systemically important (G-SIBs). By necessity, the assessment of capital surcharges and their application to those banks comprise a joint decision at the international level, since the relevant system is global. Furthermore, the proposed framework to deal with the banks that are systemically important from a domestic perspective (these are more numerous than the G-SIBs) sets out principles that govern the interaction between the assessment and actions of a bank's host supervisor and those of its home supervisor. Cooperation is built into the framework.

Macroprudential policy may be a recent addition to the toolbox of policymakers, but it already embeds international cooperation. I believe that this approach to international cooperation is a good one. It fully recognises international spillovers while preserving national room for manoeuvre in applying agreed principles. Coordination is advanced through information-sharing, common minimum standards and reciprocity.

 

5.    The hierarchy of action

Let me now close by offering a cautionary remark concerning the interaction between macroprudential policy and other policies. As I noted earlier, macroprudential policy may have macroeconomic effects. Attempts to mitigate the financial cycle are likely to influence the business cycle. Prudential tools may affect credit and asset price dynamics and, by extension, aggregate demand.

Because of that, it is essential to ensure that the hierarchy of policy tools is clarified. Macroeconomic management should first rely on macroeconomic tools (monetary and fiscal policies) before asking for help from macroprudential policy. Financial stability is already a large enough job for macroprudential policy. It should remain focused on its main objective rather than trying to smooth the business cycle. The temptation to bend prudential tools away from their primary objective of financial stability to tackle shorter-term macroeconomic fluctuations can be quite strong. Given measurement uncertainties, the case for doing so is less compelling than it appears. It is in situations like these that the independence and accountability of macroprudential frameworks are particularly valuable.

Moreover, financial stability is too big a burden to rest exclusively on prudential and macroprudential policies; it needs the cooperation of other policies: a more symmetrical monetary policy across the financial cycle, fiscal policies that create additional space in financial booms, etc.

Finally, let me finish on a positive note. Despite our limited knowledge about the impact of macroprudential policies, there is significant room for effective action - for at least three reasons:

  • First, potential policy conflicts are usually exaggerated. It seems likely that, in most circumstances, macroprudential policy and monetary policy will be complementary, tending to support each other instead of conflicting. It is important to realise that the financial cycles that matter for prudential policy are of a much lower frequency than business cycles. This suggests that, most of the time, monetary policy should be able to treat macroprudential policy developments as a relatively slow-moving background. However, it also requires monetary policy to keep an eye on developments over longer horizons in order to take into account the effects of the gradual build-up and unwinding of financial imbalances. This longer horizon diffuses some of the possible tensions between monetary policy and macroprudential decisions.
  • Second, there is already a growing body of research and experience that has led to significant progress being made both conceptually and operationally, for instance in the design and calibration of macroprudential tools.
  • Third, some tools and indicators seem to produce reasonable results - certainly better than doing nothing. In particular, the credit gap indicator embedded in the Basel III countercyclical capital buffer seems to provide good guidance for action. Simulations indicate that following this indicator would help to produce meaningful action (eg raising capital) at an early stage, before the beginning of a financial crisis. For example, the United States and the United Kingdom would have started setting aside more capital in 1999, and the 2.5% buffers would have been completed by 2002 and 2006 respectively, ie well before the financial crisis. Spain would have started even earlier, in 1997 (with the 2.5% buffer completed in 1999). Of course, the indicator would not have worked so well in some other countries. For instance, in the case of the Netherlands, it would have peaked too early compared to the evolution of the financial cycle; nonetheless, healthy buffers would have been built. Also, the credit gap indicator has proved to be noisy for some large emerging market economies such as Brazil and Turkey. To be sure, this indicator can be supplemented with the information coming from the analysis of other indicators.

These are just a few examples of the possibilities of one of the instruments of macroprudential policies. They illustrate the potential but also the need to continue to work on how the macroprudential approach can be formalised and applied to different institutional frameworks in a way that strengthens other policies and mitigates systemic risk.