Simulating and dealing with financial instability: Challenges for central banks
Speech by Lucas Papademos, Vice President of the ECBat the conference on “Simulating financial instability: stress-testing and financial crisis simulation exercises”European Central Bank, Frankfurt am Main, 12 July 2007
The topic of this conference “Simulating financial instability” may, at first sight, seem odd for a conference organised by a central bank whose tasks include contributing to the safeguarding of financial stability. But we have all come to realise that in order to achieve the objective of preserving financial stability, we must further enhance our understanding of the factors and processes that lead to financial instability. To this end, we must improve our analytical tools for monitoring and assessing financial imbalances and potential risks that can adversely affect the stability and efficiency of the financial system. We must also strengthen the institutional arrangements that can help prevent the emergence of financial instability and, if a crisis occurs, can help to manage it effectively and mitigate its impact on the financial sector and the real economy.
To achieve these aims, it is useful – indeed necessary – to further develop analytical tools for modelling key processes, components and the functioning of the financial system and for stress-testing its resilience to shocks. Moreover, it is also essential to conduct crisis simulation exercises to test the response and preparedness of institutional arrangements to a hypothetical but plausible “virtual reality” situation of financial instability. These are the two, seemingly rather technical, but very practical and policy-relevant topics that we are trying to address in this conference in the light of past experience and ongoing developments in the financial system.
In my remarks, I will focus on three pertinent issues. First, I will elaborate on the role and significance of stress-testing and simulation exercises within the framework for safeguarding financial stability, with special emphasis on the tasks and responsibilities of central banks for preserving financial stability. And I will also explore how we could achieve synergies between these two areas of work.
Second, I will highlight some of the changes that are taking place in the financial system and the macro-financial environment and assess their implications for the analytical, operational and practical aspects of the financial stability framework, and specifically for the two tools we have been discussing today, stress-testing and crisis simulation exercises.
Third, I will point to some broader policy issues concerning the effectiveness of the institutional arrangements for financial stability.
II. The framework for safeguarding financial stability: the role of stress-testing and crisis simulation exercises
Let me start with some observations on the current “state of play” with regard to the framework for safeguarding financial stability. This framework has analytical, operational and institutional components. Since the early 1990s, significant progress has been made towards enhancing the precision and rigour of analysis and the effectiveness and efficiency of institutional arrangements. This progress reflects the increasing importance of financial stability issues as a result of the remarkable development, growth and integration of financial markets as well as the episodes of market turbulence experienced over this period.
Nevertheless, despite the progress made, it is fair to say that the financial stability framework has not yet reached “a steady-state”, but is still evolving. There are several reasons for this. First, the quantification of the objective of financial stability is not straightforward. Second, the analytical component of the framework (involving appropriate indicators, models and methodologies) for monitoring and assessing the dynamics of the financial system and of the factors and potential risks that can affect its stability is inherently more challenging than is the case for the monetary policy framework. Third, the continuous and striking transformation of the financial system poses additional challenges for the analytical and institutional components of the financial stability framework.
What are the roles of stress-testing and simulation exercises in the framework for safeguarding financial stability? Can these tools help address some of the implications for financial stability arising from the ongoing transformation of the financial landscape?
As you know, the framework for financial stability includes the performance of various functions and can be usefully considered as comprising three main stages concerning crisis prevention, crisis management and crisis resolution. Crisis prevention consists of the performance of both micro-prudential supervision and central banking functions. The latter include financial stability monitoring and assessment, which relies on macro-prudential analysis. Crisis management comprises the set of tools and policy actions that public authorities may use if and when a financial crisis occurs in order to contain its impact. And crisis resolution relates to the arrangements for the orderly winding up of a failing institution and the protection of the rights of creditors, notably depositors.
Within the financial stability framework I have just described, both stress-testing and crisis simulation exercises can support the coupling of the macro and the micro-prudential perspectives to financial stability and the preparedness of authorities to address a financial crisis. Macro stress-testing is one of the activities aimed at crisis prevention, while financial crisis simulation exercises are geared towards the testing and enhancement of crisis management and resolution procedures.
For central banks, key activities contributing to the performance of their financial stability tasks are (i) the identification and monitoring of sources of risk and vulnerability in the financial system and (ii) the analysis and assessment of internally generated imbalances and externally induced disequilibria due to shocks. Stress-tests are particularly useful for risk monitoring and assessment as they make it possible to quantify the likely impact of shocks, which helps to rank risks by their importance and allows assessment and surveillance to be more focused. Moreover, stress-tests can help provide early warning signals and thus contribute to the forward-looking dimension of financial stability monitoring and assessment.
Crisis simulation exercises are also forward-looking instruments as they aim to predict how the crisis management and crisis resolution arrangements will function in practice on the basis of crisis scenarios. Moreover, simulation exercises of a cross-border financial crisis can demonstrate the challenges to achieving the appropriate degree of cooperation between national authorities in order to address the cross-border systemic implications of a crisis, to the benefit of all those affected.
Crisis simulation exercises may also be designed to test the overall effectiveness of financial stability arrangements, including the consistency of the principles and procedures underlying crisis prevention, management and resolution. In particular, some of the simulation exercises organised thus far have helped to bring to light the existence of legal, regulatory and other obstacles to smooth cooperation between authorities. Some of the implications of these obstacles cannot be fully anticipated ex ante but, if the factors underlying such obstacles are replicated in a simulation exercise, their potential impact on the overall effectiveness of crisis management and resolution processes can be demonstrated, and this can help enhance preparedness for handling a crisis.
A key question worth exploring is whether useful synergies between macro stress-testing and crisis simulation exercises can be identified and achieved. This is an important issue as work has generally been conducted separately in the two fields and the current juncture could provide a good opportunity to examine potential complementarities and synergies. Let me say a few words on two such types of synergy.
The first type of synergy which can be achieved between stress-testing and simulation exercises concerns what we may call the design process On the one hand, crisis simulation exercises can benefit from stress-testing models, as the results obtained from such models can be used as input for the design of a crisis scenario to test institutional arrangements. Stress-testing models may be particularly useful in assessing whether a particular scenario will generate outcomes that are sufficiently severe as to have systemic implications. Of course, the use of model-based scenarios to replicate a financial crisis is subject to limitations. Nevertheless, such scenarios – supported by the outcomes of stress-tests – are useful to the extent that they allow policy-makers to consider the propagation of shocks and the evolution of a crisis and thus identify ex ante alternative channels of shock transmission and appropriate policy responses.
Conversely, stress-testing models can also benefit from simulation exercises as the design of a simulation exercise involves a mapping of the potential systemic linkages that may spread the propagation of a crisis and increase its severity. These linkages may include the relationships between vulnerabilities embedded in a bank’s balance sheet, the links to the liquidity needs of that bank and of the banking system as a whole, as well as the repercussions for payment systems. The thought processes underlying the design of a simulation exercise can provide valuable input for stress-testing models through the identification of additional indicators and factors which may be usefully considered.
The second type of synergy which can be achieved relates to the information provided by macro stress-testing and simulation exercises for policy-makers. While macro stress-testing certainly helps to quantify the effects of various shocks on the macro-financial environment and to model their transmission to the financial system using internally consistent scenarios, it is based on some simplifying assumptions. One such assumption is that it usually does not incorporate the responses of either financial market participants and financial institutions or of central banks and supervisors. Crisis simulation exercises can provide useful information on the effectiveness – as well as the potential limitations – of the institutional arrangements intended to safeguard financial stability when faced with a severe crisis scenario. Combining the outcomes of stress-testing and simulation exercises allows policy-makers to obtain a broader and more realistic picture of the degree of resilience of the financial system and of the effectiveness of institutional arrangements with respect to crisis scenarios.
III. The changing financial system and macro-financial environment: analytical issues and policy challenges for financial stability
Thus far, I have spoken at some length about analytical tools, simulation exercises and crisis scenarios. But there is one scenario which does not involve a crisis, namely the future development of our own efforts to better understand the changes that are taking place in the financial system and the macro-financial environment; to discern what they mean for financial stability and for the financial stability framework; and to assess their potential impact on the substance of our central banking task of safeguarding financial stability, as well as on the two specific tools of stress-testing and simulation exercises.
III.1. The changing financial system
We have all been witness to fundamental and far-reaching changes in the financial system. I would like to highlight three and point to their policy implications. The first is the rapidly increasing financial integration in Europe and globally, both across borders and across financial sectors. This is leading to a blurring of the boundaries of markets and of the separation between the activities of financial institutions. This, in turn, implies that a comprehensive financial stability assessment must take into account, monitor and analyse the cross-sectoral links and internationalisation of financial systems. Moreover, detecting and assessing risks and vulnerabilities, containing threats and managing a potential crisis in an increasingly integrated financial environment require appropriate sharing of information and effective cooperation between central banks and other responsible authorities across jurisdictions.
Second, financial innovation has engendered the emergence and rapid growth of new and complex financial instruments and the expansion of over-the-counter (OTC) derivatives markets and has fostered the growing presence of other, non-bank financial intermediaries that are very active in both traditional and OTC markets. Many of these other financial institutions are highly leveraged and most are lightly regulated. These developments have facilitated the spreading and redistribution of risks across sectors and have contributed to the completeness and efficiency of financial markets. One important consequence of recent financial innovations, advances in risk management techniques and the associated phenomenal growth in the market for credit risk transfer (CRT) instruments is that they are altering the traditional roles and business models of different types of financial intermediaries. There are good arguments supporting the view that, on the whole, these developments strengthen the shock-absorption capacity of the financial system. But there are also concerns that under certain circumstances, and in response to sizeable − and possibly correlated − shocks, the financial system may face a real and challenging “stress-test” of shock amplification. One thing is clear. The complexity of some of the new financial instruments, the cross-sectoral redistribution of risks and the opaqueness of the transactions of a growing number of non-bank financial institutions have made the modelling of the financial system as well as the monitoring and assessment of risks much more difficult.
The third, and final, feature of the financial market landscape I would like to highlight is a consequence of the previous two and their interaction with global macroeconomic factors and monetary conditions. Financial liberalisation and integration, innovation and consolidation combined with the global distribution of saving and investment, and global monetary conditions have contributed to a macro-financial environment characterised by low market volatility, low risk premia and abundant financial market liquidity. The potential links between, on the one hand, structural change and micro factors in the financial system and, on the other hand, macroeconomic factors and monetary conditions are not sufficiently well understood and deserve further analysis. For example, although financial market liquidity and monetary liquidity are distinct concepts, there are links between the two which could have implications for the level of risk and term premia, the dynamics of asset prices and the conduct of central bank policy. These observations support the view – recently also stressed by Malcolm Knight and Jaime Caruana – that, in performing the task of safeguarding financial stability, the micro and macro-prudential approaches should be increasingly combined, in order to obtain a broad-based and robust assessment.
What are the other implications of these changes in the financial system and the trends in the macro-financial environment for the analytical, institutional and operational aspects of the financial stability framework, and, specifically, for stress-testing and crisis simulation exercises? Are there any broader implications for central bank policy? These are the questions I would like to briefly address next.
III.2. Challenges for stress-testing and crisis simulation exercises
We have made impressive progress in the fields of both macro stress-testing and crisis simulation but, in the light of the substantive changes in the financial system I mentioned before, I would like to raise three issues concerning: first, the limitations of model-based scenarios; second, the availability of data; and third, the design of crisis simulation exercises.
First, overcoming the limitations of model-based stress-tests is an analytical challenge. Obviously, models can only provide a stylised and limited picture of reality. This limitation is especially acute in the case of a financial system which is evolving, becoming more complex and multi-dimensional. Stress-tests often focus only on part of the financial system and fail to produce sufficiently “severe” results, because many of the typical characteristics of financial crises, such as contagion, non-linearities and second-round effects, are not accounted for. These characteristics are probably becoming more relevant as a result of the ongoing transformation of the financial system. Further advances in the modelling of the manifold linkages within the system are crucial in order to generate stress-tests that simulate crises of systemic importance – for these are particularly relevant to us as central bankers. In addition to improved modelling of several interactions and contagion channels within a given financial system, we need a better understanding of the links between the financial system and the macroeconomic environment. We thus need to bring together both micro and macro aspects and incorporate them into stress-testing and simulation exercises.
Second, a more practical but equally essential issue is the availability of data. Any model can only be as good as the data fed into it. And undoubtedly, model-based stress-tests are demanding in terms of data. The challenges on that front are manifold. Sufficient historical data are often lacking, and even when relatively long time series are available, changes in the macroeconomic or regulatory environment may limit their usefulness. Cross-country analysis is especially complicated because data availability differs substantially across countries.
However, perhaps the most relevant data problem relates to confidentiality. I accept that in most cases there are very good, and indeed legal, reasons for preserving confidentiality, particularly when dealing with information on individual banks. But the scarcity of data seriously limits modelling efforts. The existing data difficulties are augmented by the growing importance of new financial instruments that are traded over the counter, and by the increasing activities of non-bank financial institutions. A key question is how we can obtain sufficient data to help improve our detection and understanding of risks and vulnerabilities without imposing restrictions and burdens which would unnecessarily hamper innovation and efficiency in financial markets.
I am not making this point to justify or rationalise the seemingly insatiable appetite of researchers and specialists for ever more data. I am making it because effective risk assessment and crisis prevention require a certain minimum degree of realism in the scenarios that are developed. Financial institutions, supervisors and central banks should explore ways to enhance disclosure of adequate and relevant information so that potential flash points in the financial system and the functioning of shock transmission mechanisms can be identified in a timely manner. That said, in calling for adequate and relevant information, we must bear in mind the risk that requests for more detailed disclosure could also result in undesirable outcomes.
Third, an operational aspect is the design of crisis simulation exercises The crisis scenario should be sufficiently realistic, but also manageable for everyone involved. At the same time, it should serve to effectively evaluate the institutional arrangements. As anyone who has designed or participated in a simulation exercise can testify, combining these three objectives is not easy. And there are inevitably limitations. This is particularly important when evaluating such exercises ex post. After all, the evolutionary biologist Thomas Huxley reminds us that “what you get out depends on what you put in; and the grandest mill in the world will not extract wheat flour from peascods”.
III.3. Policy implications for financial stability institutional arrangements
Finally, what are the broader policy issues and challenges stemming from the rapidly evolving and integrating financial system for the institutional component of the framework for safeguarding financial stability? And how can we effectively address them?
These questions do not have short and straightforward answers, but I would like to briefly stress a few points. First, increasing financial integration clearly requires enhanced cooperation between authorities across countries, not only in supervision and crisis prevention, but also in crisis management and resolution. Second, and in order to achieve that objective in an effective and efficient manner, the arrangements for crisis management and resolution should be consistent with the arrangements for crisis prevention, including financial stability monitoring and assessment. This entails a need to explore thoroughly how to strengthen cooperative arrangements for financial stability especially, but not exclusively, in the increasingly integrated European financial system which, however, is characterised by institutional arrangements in which authorities have an essentially national responsibility and accountability. Third, institutional arrangements should be such as to minimise moral hazard and preserve constructive ambiguity about the terms and timing of a possible public intervention. This is essential for reinforcing market discipline and fostering appropriate incentives. Fourth, it is important to address various legal and technical issues, for example the removal of legal obstacles limiting the transferability of assets across borders, which can facilitate the management of a potential crisis affecting cross-border banking groups. In this context, I am pleased to say that much has been learned from crisis simulation exercises in the Eurosystem on how to address some of these issues and on the importance of removing obstacles that can hinder the effective performance of our financial stability tasks. Finally, strengthening the links and achieving synergies between the micro and macro-prudential approaches to financial stability is becoming increasingly important for a more robust assessment of risks to financial and macroeconomic stability.
It could be observed that a number of the issues I have just raised are direct, rather obvious, implications of the ongoing changes in our financial system and macroeconomic environment. I would consider it very positive if agreement could be reached on the above points. But the design and implementation of the appropriate institutional arrangements for financial stability are less straightforward. Fortunately, a lot of pertinent work has been undertaken by authorities and other fora in order to enhance the effectiveness of the financial stability framework of the European Union (EU). Interestingly, some of that work started as a follow-up to an EU-wide simulation exercise of financial instability.
IV. Concluding remarks
After this rather dense presentation of various analytical and policy issues related to our topic “Simulating financial instability”, let me conclude by sharing with you an interesting observation on this very theme. Originally, simulation had rather negative connotations. In the dictionary of the English language of 1775 it is noted that “a Deceiving by Words, is commonly called a Lye, and a Deceiving by Actions, Gestures, or Behaviour, is called Simulation”. But, even with such associations surrounding the word, Francis Bacon, the English philosopher, could still see some utility in simulation when he mused about a good shrewd proverb of the Spaniard, ‘Tell a lie and find a truth’. As if there were no way of discovery, but by simulation”. This brings us back to our discussion today, for we seek, through simulating financial instability, to discover the “truth” or at least enhance our understanding about the underlying factors and processes that may lead to instability in today’s complex financial system, and about the most effective institutional arrangements and policies to prevent and deal with financial instability.
Thank you very much for your attention.