Institutions for stability - current and past experience
Speech by Andrew Crockett, General Manager of the BIS and Chairman of the Financial Stability Forum, at the first conference of the Monetary Stability Foundation, Frankfurt, 6 December 2002.
The past twenty years or so have seen a remarkable evolution in the intellectual underpinnings and institutional setting for central bank policy making. A broad measure of agreement has been reached on a number of central issues. At the same time, new issues have arisen on which vigorous debate continues.
In my remarks today, I will note, without entering into great detail, those aspects of central banking on which there is, at least for the time being, a settled consensus. Then I will spend the major part of my time outlining five issues on which views continue to diverge, and which are of key importance for monetary and financial stability.
There is nowadays no dispute that price stability ought to be central banks' principal goal. Nobody seriously believes any more that there is a long-run trade-off between output and inflation. Of course, this broad agreement does not extend to important matters of practical detail, such as the precise rate of inflation that should be considered optimal, and the time frame within which price stability should be restored following a shock. But the remaining matters of debate are, I submit, less significant than the central area of agreement.
There is likewise little dispute that central banks should be given the necessary autonomy to pursue a mandate for price stability. Again, there are technical disagreements about how far this autonomy should go (goal independence or instrument independence) but little difference on the basic principle.
Next, there is a broad consensus on the importance of transparency in monetary policy making. Transparency satisfies the requirements of democratic legitimacy in an open society and also, as empirical evidence is now confirming, can improve the effectiveness of monetary policy. Once again, the debate has moved on from the principle, now settled, to the practical aspects of implementation.
Lastly, there is a growing recognition of the value of an inflation-targeting framework. In its broad sense, this framework can be said to be used by central banks where the inflation objective is implicit (such as the Federal Reserve) as well as those where it is explicit. As is well-recognised, an inflation targeting framework involves much more than simply a numerical target for inflation.
Before leaving those aspects of central bank policy making where consensus has been reached, it is well to enter a cautionary note. There have also been moments of consensus in the past (think of the Phillips curve) where apparent regularities have crumbled under the policy weight placed on them. Central banks have been remarkably successful in the past two decades in restoring and maintaining price stability. But we should recognise that the future usually has surprises in store.
I turn now to five topics where there is much more current debate, and where lively disagreement seems likely to persist for some time to come. Not surprisingly, these topics are at the nexus of price stability and financial stability. For while central banks have, broadly speaking, been winning the long struggle against inflation, success on the front of financial stability has proved more elusive. The topics are:
Few would deny that asset price bubbles are disruptive. Nor that it would be a good thing if they could be recognised and deflated without adverse side effects. But that is a far cry from saying that central banks should make asset price developments an objective of monetary policy.
At one level, there is no dispute that asset price developments need to be taken into account in monetary policy decision-making. To the extent that movements in asset prices create wealth effects that feed back into demand and inflation, they need to be, and typically are, factored into central banks' forecasting models.
Beyond that, however, there is scepticism about the usefulness of paying disproportionate attention to asset price movements that might or might not turn out to be a bubble. This scepticism is based, in the first instance, on doubts about economists' ability to distinguish a price "bubble", which is by definition reversible, from a shift in relative prices which may be lasting. Are central banks better placed to make this distinction than the multiplicity of market players whose collective views are reflected in revealed prices?
Secondly, there are doubts about whether monetary policy actions by central banks can be effective in deflating a bubble without serious negative side effects. On this argument, a modest tightening of monetary policy may have little effect on asset prices and may even for a while support asset prices by increasing market confidence in the central bank's ability to manage the economy. A more substantial tightening on the other hand may provoke a recession.
Thirdly, there is a growing view that agile monetary policy can limit any negative consequences from the bursting of an asset price bubble. If monetary policy is shifted rapidly towards ease as asset prices fall (the case of the United States in 2001), then the underlying health of the real economy can be sustained.
More generally, many observers question whether central banks should allow other considerations to interfere with pursuing the goal of price stability. If asset price movements do not impede the achievement of the desired inflation objective, is it not confusing (and non transparent) to attempt to achieve additional objectives?
These are all powerful arguments. And if the potential negative effects of the emergence or bursting of asset price bubbles are limited, they may be powerful enough to carry the day. But this depends essentially on how serious bubbles prove to be, something on which the jury is still out. Japan, at one extreme, is still struggling to overcome the real economic costs of the aftermath of the financial bubble in the late 1980's. The United States, on the other hand, acted decisively to deal with the bursting of its bubble, and so far real economic activity has held up reasonably well.
The outcome of this debate will doubtless depend on what happens from now on. If the after-effects of bubbles turn out to be more serious and long-lasting than currently expected, then more attention will surely be given to preventive measures. Among these, in addition to prudential policies, will be the role of monetary policy.
It would in my view be wrong, however, to assume that addressing the buildup of financial imbalances is necessarily at odds with the pursuit of an inflation objective. Much depends on the time frame considered. Financial imbalances typically take time to be perceived and eventually reversed. The subsequent delay in the impact on the real economy only adds to the lags at work. These lags may thus turn out to be much longer than the two years or so usually considered relevant for monetary policy making.
One lesson from recent experience may be to take a longer-term view of the impact of financial imbalances on the future course of inflation and real economic activity. But what are the practical implications of attempting to do so? Recent research gives ground to believe that financial imbalances (unsustainable movements in asset prices) can frequently be inferred on the basis of accompanying developments in credit expansion. Thus, it is not unreasonable to conjecture that a policy aimed at limiting credit expansion could have some effect, at least at the margin, in containing the size of financial imbalances. What form such a policy should take is a matter of debate, however.
The role of central banks in financial supervision is a subject not just of academic, but even more of political debate. In the past several years, the tide has been running in the direction of removing supervisory functions from central banks.
There are several arguments in favour of this separation. One is that of concentration of power. Central banks having received wider responsibilities in the area of monetary policy, a quid pro quo could be seen as reduced responsibility in the area of financial supervision. A second argument stems from the increasingly integrated nature of financial activity. Supervision of banks, insurance companies and securities firms must be internally consistent, given the blurred nature of the boundaries between the different activities. And since it is not generally considered appropriate to extend central banks' supervisory responsibilities to other parts of the financial sector, it may appear logical to establish an independent integrated regulator to cover the whole of the financial industry. A third argument is that of focus. A central bank should not be distracted from its primary objective of achieving price stability by concerns of the consequences for the health of individual financial institutions.
This is not the end of the story, however. There remain a number of reasons why central banks are closely interested in the health of the banking sector. In particular, important channels by which monetary policy operate flow through the banking system. The health of banks' balance sheets shapes how they respond to monetary stimuli.
More generally, cutting the supervisory link between central banks and the banking system obscures the question of who is responsible for overall systemic stability. In fact, the origin of central banking, and the initial definition of central banks' role, is to be found in the responsibility for financial system stability. Typically, this stability is threatened by dynamic processes affecting the stability of financial markets and the institutions that deal in them. Central banks, with their authority and liquidity-creating powers, have historically been the bulwark between strain and panic in the financial system. As constituted at present, regulatory authorities could not easily play this role.
My conclusion, therefore, is that the consequences of the transfer of authority from central banks to independent supervisors may not yet have been fully thought through. This does not necessarily mean that the process should be reversed. But more consideration needs to be given to what role central banks should continue to play in financial stability, broadly conceived. It is hard to believe that the system as a whole will be stronger if this role is allowed to atrophy.
Part of what is needed is clearly close cooperation between central banks and regulatory authorities, and a free flow of information between the two. In addition it will be important to develop a common perspective on the causes of, and remedies for, financial instability. This leads naturally to the third main topic of my remarks today.
There is a tendency in some circles to equate responsibility for financial supervision with responsibility for financial stability. At its root, this confusion probably rests on a belief that if each and every financial institution were sound, the overall system would be sound. This belief, in turn, stems from the view that financial system crises typically arise when a single institution fails and provokes failures among its counterparties.
These views underlie the (in itself wholly commendable) search for improved risk management practices at individual financial institutions. In this connection, impressive progress has been made under the guidance of the Basel Committee on Banking Supervision. The fact that banking systems in most major countries have recently been able to weather without serious difficulty the recent series of debilitating shocks is testimony to the value of what has been achieved.
Nevertheless, if the focus is solely on institutional (or "microprudential") supervision, there is a danger of giving too little attention to other important risks at the macroprudential level. One is the risk inherent in the economic cycle. With the benefit of hindsight, we know that risks in lending are greatest at the peak of the cycle, ie, just when the economy is about to turn down. Yet this is the time when, using conventional indicators, risks appear to be smallest. Means must be sought to improve the capacity to measure aggregate risk through time, to match the strides that have been made in measuring relative risk at a point in time. Techniques such as stress-testing, pre-provisioning and through-the-cycle risk assessment have their role to play in this endeavour.
A second, and related, source of macroprudential risk lies in the conflict between individually rational behaviour and socially desirable outcomes. When economic prospects are good, it pays to lend, and when the economy turns down, prudent banks will try to cut exposures before their competitors. Of course, since banks are responding to the same external circumstances, herd behaviour is the inevitable result. The incentives facing banks' managements add to the amplitude of the financial cycle, and hence to volatility in the real economy.
"Procyclicality" has now become a widely debated topic among those concerned with financial sector regulation. Realistically, procyclicality can hardly be removed from financial behaviour. But it seems desirable to design supervisory mechanisms so as to avoid exacerbating the problem. How far it is possible, or desirable, to go in actively combating it remains an open question.
Following the Asian crisis, it was widely recognised that financial sector weakness could trigger, or intensify, crises of confidence in currencies. It was also recognised that preventing crises, insofar as that was possible, was far less costly than resolving them. As a result, the international community adopted the strategy of developing codes and standards of best practice in the financial sector, to provide a benchmark against which all countries, and the international community, could measure progress.
The codes and standards are now very extensive (more than 60) and have been grouped in twelve key areas by the Financial Stability Forum1. Fundamentally, however, they are of three main types: those related to the supervision of financial institutions (banks, securities' firms and insurance companies); those related to the financial infrastructure (accounting, auditing, corporate governance standards, payment and settlement systems); and those related to transparency (data dissemination, fiscal and monetary policy transparency).
Not surprisingly, there are a myriad of technical issues involved in developing and implementing these standards. Today, I want simply to draw attention to two generic issues. They are: whether the standards should be minimum requirements or best practice; and whether they should be globally applicable or adapted to the requirements of each jurisdiction.
There are of course arguments on both sides. Those who argue for minimum requirements point to the difficulty of quickly reaching consensus on best practice across a wide range of countries. And those who argue in favour of national discretion in the application of standards note that each country has its own economic, historical and cultural specificities.
Nevertheless, I believe that the best course is to aim for standards that are of as high a quality as possible, and can be adopted in as many jurisdictions as possible. A case in point is accounting and auditing standards. Recent experience has demonstrated how costly lapses in this area can be. It has also demonstrated that lapses can occur in all jurisdictions, however robust their standards were previously considered to be.
It is encouraging that the International Accounting Standards Board (IASB) and the Financial Accounting Standards Board (FASB) have now embarked on a convergence project, aimed at combining (and indeed improving upon) the best of their respective standards. The significance of this project for financial reporting, and financial stability more generally, can hardly be overstated. It would be extremely unfortunate and short-sighted if national jurisdictions failed to give it their support and refused to indicate their willingness, in principle, to sign up to a single internationally consistent set of accounting standards.
Other projects are also underway. IOSCO has recently issued draft guidelines on auditor independence, auditor oversight, and disclosure and transparency. In my view, these offer a solid basis for putting in place additional crucially important building blocks of the new financial architecture.
Finally, there is corporate governance. The OECD has embarked on a revision of its Corporate Governance Principles. The original principles were a valuable initial effort to draw together experience in OECD countries. What is now needed, however, is more "beef": in other words, an attempt to distil and make specific "best practice", rather than to describe existing practice.
A review of outstanding issues in improving institutional arrangements for financial stability would not be complete without some reference to the relative roles of the public and private sectors in tackling international financial crises.
The nature of international financial crises has been changing over time. The most troublesome issues nowadays are created when a crisis of confidence provokes an outflow of capital from a country. The loss of confidence usually has its roots in doubts about whether the country's government will be able to undertake the necessary measures to render its debt sustainable. Frequently, a vicious circle develops. Interest rates on outstanding debt rise, making it even less likely that the country can service its debt and leads to further capital outflows.
The issue is how this vicious circle can be broken, and a return to a "good equilibrium" achieved, with restored confidence, lower interest rates, and hence a sustainable debt service burden. The possible responses to a confidence crisis (which are not mutually exclusive) fall into three categories: domestic adjustment measures (particularly fiscal measures) aimed at generating additional financial resources for debt service; official assistance from international financial institutions, aimed at providing confidence-building external financial support; and the renegotiation of private sector debt, aimed at lessening the burden of debt service.
All three approaches may have their role to play, but also their drawbacks. Adjustment measures are clearly needed if the existing macroeconomic stance is unsustainable. But where market confidence has evaporated, the required scope of measures may be so great as to be politically and economically counterproductive. Official assistance, mainly through IMF programmes, can help, since both the finance and associated economic measures help build confidence. In recent years, however, the amounts required to restore confidence have tended to exceed what creditor governments have been willing to provide.
This is the reason why the third response, under the generic heading of "private sector involvement" has attracted increased attention in the public debate. Its intuitive appeal is undeniable. The private sector has typically been the main source of lending (or overlending); it is the loss of confidence on the part of the private sector that has usually provoked the crisis; and it is the private sector that, in principle, stands to gain the most from an orderly resolution of debt servicing difficulties.
Reflecting considerations such as these the IMF has proposed a "Sovereign Debt Restructuring Mechanism" that, if implemented, would replicate some, though by no means all, of the features of domestic bankruptcy provisions in national jurisdictions. The argument, greatly simplified, is that when debt servicing on the terms of the original contract has become impossible, it is better for all parties to have an orderly and predictable resolution.
Despite the appeal of this approach in principle, most issuers and holders of sovereign debt remain vigorously opposed to it. Their main argument is that by making debt restructuring more normal and acceptable, it would reduce lenders' willingness to provide credit, reducing the amount and raising the long-run cost of borrowing. They further argue that the scope of the problem has been exaggerated. The number of situations in which an agreed solution between lenders and borrowers cannot be reached without bankruptcy-like procedures is, in their view, small.
This debate looks set to continue for some time to come. Without entering in an analysis of the merits of specific proposals, I would only enter a note of caution on the wisdom of imposing a solution when the presumed beneficiaries (lenders and issuers) remain unpersuaded of its potential advantages.