Central banking under test?
Speech by Andrew Crockett, General Manager of the BIS and Chairman of the Financial Stability Forum, at the BIS Conference on 'Monetary stability, financial stability and the business cycle', Basel, Switzerland, 28 March 2003.
Let me begin by asking you to think back to the mid-1970s. At that time, after the first round of oil price increases, inflation was in double digits, and rising. If you had told central banks then that one day they would function in a world in which inflation was low and stable, their anti-inflation commitment was highly credible and they had operational independence, they would probably have suspected you of smoking a prohibited substance. If, nevertheless, they had believed your forecast, they would surely have thought it must be the best of all possible worlds.
Well now we live in such a world. And it has yielded great benefits. It has laid the basis for better long-term economic performance. It has removed a major cause of arbitrary income redistribution. And it has "downgraded" what for so many years had been the main reasons why economic expansions came to an end, namely tighter monetary policy designed to quell rising inflation.
And yet, the new world has not quite turned out to be "the best of all possible ones". Central banks have had to face some considerable challenges. Bringing down inflation to low levels has raised the spectre of deflation, which is already a problem in one major country, Japan, and is increasingly debated elsewhere. Moreover, low inflation has not proved to be a guarantee of financial stability. Indeed, banking crises seem to have had just as large costs in terms of output forgone than in periods when inflation was higher. In short, the successful war against inflation has not yielded as big a "peace dividend" as we might have hoped.
Central bankers and academic observers are increasingly debating the underlying reasons why this has been the case. Why has monetary stability apparently not resulted in greater financial stability? What is the role in the process of the credit cycle and asset price volatility? Can asset price volatility and financial fragility interact to lead economies into a deflationary trap? And are there adjustments to central bank policy that would reduce these risks and enable a better combination of monetary and financial stability to be achieved?
The present conference is devoted to exploring the background to many of these questions. My own view is that the observed developments are not purely coincidental, but rather reflect the interaction between changes in the financial and monetary regimes. A crucial missing link in the story as told so far is that developments in the financial regime have arguably made the system more prone to booms and busts in financial asset prices with significant macroeconomic costs. But the monetary regime may have a subtle role to play too. Paradoxically, high anti-inflation credibility may sometimes allow economic expansions to proceed while masking those signs that, in earlier times, would have pointed to their unsustainable pace. Under these conditions, financial imbalances may be more prominent in business cycle dynamics and provide valuable signals of possible future pressures on the economy. A key question is how best to factor such imbalances into the policy framework.
Let me use these introductory remarks to flesh out this basic thesis. I will first review some salient changes in the economic landscape. I will then explore the implications of this evolution for the relationship between monetary stability, financial stability and the business cycle. I will finally draw some policy conclusions. In the process, I will try to touch on the themes that will crop up during the rest of the conference, and to develop a logical thread tying together the various contributions.
1. The evolving economic environment: some stylised facts
First, then, how has the economic landscape evolved? I will approach this question by highlighting the behaviour of four key "performance indicators", and noting two regime shifts.
The first performance indicator is the behaviour of inflation. Over the last twenty years or so the inflation picture has changed dramatically. Gone are the days of high and variable inflation rates. Since at least the early 1990s, much of the world has entered a period of relatively low and more stable inflation. Indeed, the disinflation phase has been so strong that some countries in Asia have actually been experiencing declines in the overall price level, while in a larger group "true" inflation may be close to zero or negative once technical biases in price indexes are taken into account. Greater inflation stability has also resulted in lower inflation persistence, in the sense that inflation rates now tend to exhibit mean-reverting behaviour. Inflation seems better anchored than in the past.
The second performance indicator is short-term output volatility. According to this metric, in large parts of the world the business cycle appears to have become more moderate since at least the mid-1980s. This trend has been particularly evident in the United States, but has been documented for a broader set of industrial countries. Still, there are exceptions. In particular, output volatility has tended to be higher in several of the economies experiencing widespread financial distress. In the industrial world, Japan is the clearest example; among emerging market countries, a number of East Asian economies seem to fall in the same category.
The third performance indicator relates to asset price booms and busts. Since the mid-1980s many countries have witnessed large medium-term swings in asset prices, typically accompanied by similar fluctuations in credit. Three major fluctuations can be discerned among industrial countries since the 1970s. They correspond to the early to mid-1970s, the mid-1980s to the early 1990s, and the second half of the 1990s to the present, with Japan not having taken part in the latest upswing following the bust in asset prices in the early 1990s. The amplitude of these asset price cycles does not appear to be getting any smaller.
The fourth performance indicator relates to financial crises. Since the 1980s the world has witnessed an increase in the incidence and severity of episodes of widespread financial distress. These have typically been associated with the bust phase of asset price and credit cycles, as the corresponding financial imbalances built up in good times unwound in a disruptive manner. Examples include the Nordic countries and Japan as well as some Latin America and, later on, East Asia. The resulting macroeconomic costs have been sizeable, with percentage point estimates of GDP forgone often running in the double digits. Even in more benign cases, the unwinding of financial imbalances resulted in considerable strains on the financial system and powerful "headwinds" for the real economy, as suggested by the experience of the United States, United Kingdom and Australia in the early 1990s.
What about the concurrent changes in the policy regime?
The first change has been financial liberalisation. Financial liberalisation made its first timid steps in the 1970s, and gathered momentum in the 1980s, both within and across national borders. By the 1990s the global shift from what has been called a government-led to a market-led financial system was largely complete, sustained by the ascendancy of free market philosophy, the recognition of the costs of financial repression, and technological innovation. Hence the rapid growth of the financial superstructure in the form of "financial deepening" within individual countries and tighter links across countries.
The second change in the policy regime has been the increased focus of central banks on a firm commitment to price stability. This commitment was in part the reflection of a stronger political and social consensus to bring inflation under control. Over time, it has been underpinned by changes in the institutional and operational framework. Hence the generalised trend towards giving central banks mandates more clearly focused on price stability; endowing them with a greater degree of "autonomy" to carry out these mandates; and the adoption of more structured approaches for pursuing price stability, in the form of inflation targeting regimes. The resulting stronger commitment has been instrumental in the successful fight against inflation. Among its various benefits, and as often underlined by central banks, it also seems to have helped better to anchor expectations around inflation objectives, as judged from evidence drawn from yield curves or surveys.
In some respects, this configuration of arrangements in the financial and monetary spheres has similarities with the one that prevailed in the gold standard and early interwar period. Looking beyond obvious differences in other respects, it was then that we saw for the last time the combination of liberalised financial markets with a monetary regime that, by design or implication, was seen as delivering price stability.
This suggests that there may be lessons to be learned from revisiting history. It is the reason why the first session of the conference takes us back to those days. Thus, David Laidler retraces for us the intellectual debate of that time over the relationship between monetary stability, financial stability and the business cycle, while Barry Eichengreen and Kris Mitchener explore how far the Great Depression can be seen as a credit boom that went wrong.
2. The evolving economic environment: an interpretation
How should one interpret the stylised facts I described a moment ago? Many interpretations are, of course, possible. The more sanguine ones would highlight the observed moderation in the business cycles as evidence of better policy and of the benefits of lower inflation. They would probably consider the larger medium-term swings in asset prices and credit, and the subsequent episodes of financial strains, as either the result of country-specific weaknesses in the financial infrastructure or of transitional problems in adjusting to a new financial environment.
There is, in my view, an important element of truth in these interpretations. Surely lower inflation has helped to moderate business fluctuations, not least by removing an obvious cause of the stop-go policies of the past. And no doubt transitional difficulties in the wake of financial liberalisation have contributed to the episodes of financial distress around the world; risk pricing and management skills had been stunted in the previous financially repressed environment. Nor is it hard to find weaknesses in the financial infrastructure of many of the countries that experienced financial strains. And the very shift from a high to a low inflation environment could have exacerbated risks, by clouding the distinction between nominal and real magnitudes. For example, some of the overly exuberant equity valuations and over-indebtedness may well have been the result of mistaking nominal for real declines in interest rates.
Nevertheless, it cannot be excluded that the changes run deeper. My conjecture is that the characteristics of the business cycle may have been evolving under the combined influence of changes in the financial and monetary spheres. What we have been witnessing may thus be the reflection of a more fundamental transformation in the economic environment. As such, it may also contain more informative clues about what might lie ahead. Let me elaborate.
In the financial sphere, a liberalised financial system can more easily accommodate, and reinforce, fluctuations in economic activity. In such a system, perceptions of value as well as the willingness to take on risk arguably become more important factors driving economic activity. And these factors move in close sympathy with business conditions, potentially amplifying fluctuations. Hence, in this sense, the highly procyclical nature of credit, asset prices and indicators of the pricing of risk, such as credit spreads. During booms, self-reinforcing processes can develop, consisting of higher asset prices, looser external financing constraints, possibly an appreciating currency, further capital deepening, rising productivity and higher measured profits. These processes operate in reverse during contractions.
Such processes are a natural component of business fluctuations. But under some conditions they may go too far. On these occasions, the system's in-built stabilisers might not prevent it from becoming over-stretched. Masked by seemingly benign economic conditions, financial imbalances and the associated distortions in the real economy would build up. Their subsequent unwinding could then cause serious strains on the economy, generating strong headwinds and, in worst-case scenarios, serious financial instability.
For a number of reasons, in business fluctuations of this kind, price pressures could well be more muted than otherwise. For one thing, these fluctuations are more likely to develop in the wake of positive developments on the supply-side. Such developments tend to reduce price pressures directly, by lowering production costs, as well as indirectly, by encouraging an appreciation of the currency and capital accumulation. Moreover, an unsustainable rise in asset prices could itself help to limit price increases. Accounting profits, boosted by financial returns, would be artificially enhanced, thereby allowing more aggressive pricing strategies. Think, for instance, of lower contributions to pension funds, or the pricing policies of Japanese firms in the boom years. Large financial gains by employees could also help to keep wage pressures in check. And by boosting tax revenues, asset price booms could limit public sector financing requirements.
Against this background, arrangements in the monetary sphere play a subtle role. The success in achieving and maintaining low inflation, and the resulting increased credibility of central banks, could further attenuate the inflation process. If so, excess demand pressures could show through more gradually into higher inflation. With inflation expectations better anchored around central bank objectives, and given a credible commitment not to accommodate price increases, prices and wages would be less likely to rise. There may also be a risk that the very belief in price stability would strengthen confidence in the sustainability of unsustainable booms, by removing a reason to bring expansions to an end. Moreover, with short-term price pressures under control, policy rates may fail to rise sufficiently promptly to restrain the build-up of financial and associated real imbalances in the economy. Paradoxically, the credibility of the anti-inflation commitment can thus cloud the picture of the risks facing the economy.
This analysis points to at least three conclusions.
First, nobody should underestimate the enormous costs that inflation can have for economic activity, or ignore the many channels through which inflation can contribute to costly financial instability. Obvious examples include blurring the distinction between real and nominal magnitudes, encouraging excessive investments in inflation hedges, masking over-extension in balance sheets, and dealing with the consequences of bringing inflation under control. However, it would be unwise to expect that low inflation will, by itself, secure the appropriate degree of financial stability.
Second, measures of financial imbalances can contain useful information about the future course of the economy that is additional to that culled from measures of short-run inflationary pressures, such as the behaviour of prices or of indicators of economic slack, notably output gaps. In fact, using inflation to infer the level of output slack can be especially misleading and potentially dangerous, bolstering confidence in the sustainability of expansion. The build-up of financial imbalances can result in a subtle change in the balance of risks, towards potential future economic weakness down the road.
Finally, monetary policy can unwittingly play an enabling role. The reason is simple. In today's fiat money regimes, there is no external anchor on credit expansion other than the reaction function of the monetary authorities. If that reaction function focuses exclusively on short-run inflation pressures and does not respond to the accumulation of financial imbalances, the system may lack a sufficiently effective external anchor in the monetary sphere to prevent it from becoming over-stretched.
The papers in the second session will, in their own way, explore issues that have a bearing on the analysis just outlined. The paper by Jeff Amato and Hyun Shin can be thought of as casting light on the "paradox of credibility", as it investigates how in a world without common knowledge, credible public signals can distort the information content that variables such as inflation may contain about underlying economic fundamentals. Mark Gertler, Simon Gilchrist and Fabio Natalucci consider the implications of the exchange rate regime for the impact of financial factors on the real economy, with particular attention to financial crises.
So much for diagnosis. What are the policy implications if financial factors are likely to play a more prominent role in the business fluctuations that will no doubt continue to be a part of the economic landscape?
The first, and most natural, line of defence against the build-up of financial imbalances is prudential policy. In recent years much has been done, nationally and internationally, to strengthen prudential frameworks. The New Basel Capital Accord is but the latest in a series of important measures designed to strengthen supervisory and regulatory safeguards. These steps have reinforced, and are in the process of hardwiring, the substantial improvements in risk management that have taken place in recent years. Together, these changes have no doubt greatly increased the robustness of the financial system to financial imbalances. It is no coincidence that in the present slowdown, the banking system has so far proved much more resilient than in earlier times.
Even so, there are reasons to believe that these steps might not, by themselves, be sufficient to guarantee the necessary degree of economic stability. Since I have discussed this in detail elsewhere, I will just make a few points here. One is that the destabilising processes I have described may have significant effects on economic activity even when they do not erupt into full-blown financial crises. In addition, they operate as much through open capital markets as through regulated financial institutions. With the increased ease with which risks can be shifted outside the banking system this aspect takes on added significance. More fundamentally, the procyclical behaviour of commonly used measures of risk points to inherent difficulties in assessing how system-wide, undiversifiable risk evolves during business fluctuations and in responding to it adequately.
Dealing with this issue would call for finding ways of building up the system's defences, or protective cushions, in a boom, as risk builds up, in order to be in a position to rely on them in a downswing, when risk materialises. Quite apart from being a very difficult exercise, such a "macroprudential" approach is not a natural part of the ingrained culture of supervisory authorities. Although considerable steps in this direction have been taken in recent years, there is still a certain reluctance to address financial instability through prudential instruments when its origin is perceived to be macroeconomic.
Put differently, there are limits to what prudential instruments can achieve. In part this has to do with the raw material on which they operate, namely perceptions of risks and valuation. But it also reflects the fact that such perceptions are intimately related to the availability of liquidity in the system, which affects the ability to translate them into hard funding. Here, the role of credit extension is crucial. But what anchors the credit extension process itself?
This brings us to the potential role of monetary policy. The argument so far points to a number of reasons why there may be a prima facie case for monetary policy to respond to financial imbalances, as they build up, even if short-run inflation pressures remain in check. First, in the absence of such a response, monetary policy can unwittingly accommodate the build-up of imbalances, raising the risk of larger economic costs later on. Second, the disruptive unwinding of imbalances can cripple the effectiveness of monetary policy itself. The financial headwinds associated with strains on balance sheets and, in extreme cases, the unappealing set of options once policy rates reach the zero lower bound are obvious cases in point. The paper by Mitsuhiro Fukao in the third session explores these issues in some detail. Finally, lowering rates when problems materialise but failing to raise them when they build up could promote an insidious form of "moral hazard", which could actually contribute to generating the problem in the first place.
Even accepting these basic principles, there are, of course, major practical problems in implementing a strategy of this kind. Is it really possible to identify imbalances early enough to take remedial action? By the time they are recognised with a sufficient degree of confidence, they may be about to reverse spontaneously, so that the economy could later be struggling under the joint effect of the unwinding and the lagged impact of the previous tightening. How could one calibrate a response? The response of an economy where imbalances are present is likely to be less predictable, and the imbalances may prove to be particularly unresponsive, shifting the brunt of the reaction to more interest-sensitive sectors. And how could such a tightening be justified to public opinion if inflation pressures remained subdued or non-existent? The political economy constraints are daunting. In the third session, the paper by Charlie Bean considers these issues in some detail.
These counter-arguments are powerful; but I think it would be unwise to rule out a monetary policy response altogether. First, obtaining reliable estimates of more traditional concepts, such as output gaps and growth potential, is not straightforward either. Moreover, recent work, including some done at the BIS, suggests that it may be possible over time to develop the necessary tools. The guiding principle is to move away from trying to identify asset price bubbles per se and to focus on financial imbalances, or the set of symptoms that are likely to foreshadow subsequent strains on the real economy. Excessive credit expansion is one of these.
Second, the objective of a monetary response would be to slow down the economy in the near-term to avoid a larger and more costly slowdown later on. Non-linearities are of the essence here. Expecting a surgical removal of imbalances is unrealistic, as they are inextricably linked to the rest of the economy. In fact, if they were not, they would hardly be of great concern. Therefore the conditions for an appropriate response of the economy to monetary tightening to deal with financial imbalances are not all that different from those applicable to a traditional policy action aimed at containing inflation. And market anticipation of such a policy reaction function might even help prevent the build-up of the imbalances in the first place.
Finally, the political economy constraints are not immovable. They presumably depend on how we think the economy works and on the perceived costs of alternative courses of action. While the conditions for a pre-emptive tightening may not be satisfied at present, the formation of a greater intellectual and political consensus on diagnosis could change that.
A final question is whether a pre-emptive tightening in response to the build up of financial imbalances, even if short-term inflation pressures are under control, is consistent with current monetary frameworks. I think that the answer is yes. Only minor modifications may be necessary.
A willingness of contemplate pre-emptive tightening would not require a redefinition of ultimate objectives. Assuming the costs in terms of the traditional objectives, such as inflation and output, is the correct way of thinking about the problem. But it should be recalled that even in strict inflation targeting regimes concerns with output performance are incorporated through the length of the horizon and the trajectory chosen to return the inflation rate to within its target range, following an external shock.
If my analysis is accepted, it would probably call for two refinements in the application of an inflation-oriented strategy. First, it would suggest looking at somewhat longer policy horizons than the one-to-two year period commonly used at present. While the precise timing of the unwinding of imbalances is rather unpredictable, the processes involved tend to be drawn out ones. One could think of a tightening implying an undershooting of the objective in the near term in order to avoid a larger undershooting further out in the future. Second, such an approach would call for greater weight being placed on the balance of risks to the outlook. Monetary policy would be attempting to avoid some of the less palatable outcomes.
To conclude, central banks have been extremely successful in their long war against inflation. And yet, just as they have been emerging victorious from one battlefront, another seems to have opened. They are finding themselves under test again. How they respond to this new challenge deserves careful attention and thought.
What framework do we need to address the challenge? Perhaps one in which central banks and prudential authorities realise that the problems they face are closely linked. A framework, that is, in which they seek ways of securing both monetary and financial stability, on a sustained basis, through an appropriate distribution of responsibilities and a mutually supportive use of the instruments at their disposal. Bearing in mind, at the same time, that the ever-changing nature of the challenges calls for vigilance and an on-going willingness to re-examine critically even well-honed convictions.
The objective of this conference is to explore in depth the issues that lie at the heart of the relationship between monetary stability, financial stability and the business cycle. It is to edge us closer to providing the basis for an appropriate policy response. That basis requires a stronger intellectual consensus on diagnosis and remedies than exists at present. It is only by addressing the questions head-on, with analytic rigour and a good sense of what is practicable, that such a consensus can be developed. Events such as this conference are ideal occasions to make progress.