This conference has examined in detail how policy authorities on both sides of the Atlantic have reacted to the financial crisis. I won’t consider in depth the events of the past two years. Let me just say that central banks on both sides of the Atlantic have responded swiftly and decisively, especially since September of last year, working very closely together, even to the point of coordinating some of their actions. Indeed, you doubtless remember that the first interest rate reduction in the easing cycle, on 8 October 2008, was a coordinated move by a number of major central banks.
Market interest rates are now at very similar levels. For instance, the money market interest rates at the twelve-month horizon both in the US and the euro area are currently just below 1.3%. 
Central banks on both sides of the Atlantic have also resorted to a number of non-standard measures to provide additional support and stimulus to their respective economies. The choice and design of those measures reflects the structural characteristics of those economies. The non-standard measures implemented by the ECB have focused primarily on banks, as banks are the main source of funding in the euro area economy. In the US, however, market-based financing plays a more important role.
The ECB’s framework for non-conventional measures comprises five main building blocks: meeting in full the banks’ liquidity requests at a fixed rate; expanding the list of assets eligible as collateral; lengthening the maturities of long-term refinancing operations; providing liquidity in foreign currencies; and, since July, supporting the financial market by making outright purchases of covered bonds.
These measures are by now well known. We consider that they have achieved the desired objectives, in particular by reducing the spreads in the money market, flattening the yield curve over the short-term horizon and restarting the market for covered bonds – an important instrument through which banks finance themselves.
So much for recent measures. Today, though, I’d like to look ahead and talk about a topic which has been widely discussed at recent international meetings: the exit strategy. This strategy sets a course for progressively reversing the policy measures that were implemented in recent months, i.e. the interest rate levels and non-standard instruments.
Let me make it quite clear: this is not – yet – the time to implement the exit strategy. We clearly stated our view last week that the level of interest rate is appropriate, and we decided to conduct another one-year refinancing operation at the MRO rate (1%) with full allotment.
But it’s critical to have a well thought-out exit strategy, because markets have to realise that the current policy is temporary and will be reversed when it is no longer appropriate, in particular when risks to price stability re-emerge and conditions in financial markets have improved. As far as the euro area is concerned, the primary goal of monetary policy remains price stability; and the ECB will do whatever is needed to carry out its mandate. Inflation expectations over the medium term have to remain well anchored. This is essential to ensure that long-term interest rates remain low, thereby supporting the economic recovery.
Having an exit strategy is thus an act of responsibility for a central bank.
Drawing up an exit strategy entails two questions above all: how? and when? I won’t touch upon the how – the instruments – as both Chairman Bernanke and President Trichet have examined this particular question recently. 
Instead, I will address the issue of the when from an analytical point of view. The concrete implementation will depend on many imponderables, not least the prospects for economic recovery and for the stabilisation of financial markets – and those prospects are fraught with uncertainty. I intend to discuss a few issues that – in my view – central banks will have to consider when taking the decision to exit. This will show how complex and difficult that decision is going to be.
My analysis does not pretend to be exhaustive, but in my view the following five issues demand our full attention.
One issue is reversibility. Once the decision has been taken, we can’t easily go back on it. This seems pretty obvious, but it has some deeper implications. Let me explain.
If the decision to exit is taken too late and monetary expansion continues for too long, fresh seeds of instability in the financial sector are sown. On the other hand, if the decision is taken too early, the economic recovery might be undermined. Mistakes in the timing, both premature and tardy, (what I would call type I and type II errors) have been made in the past.
The solution is to do it right, obviously.
Some might think that these errors are not so worrying because they can be remedied while the exit policy is being implemented. If the decision to exit has been taken too early, for instance, the central bank might slow things down. If, however, the decision has been taken too late, the exit can be speeded up. It sounds nice, but is a bit too easy.
Why? Because the decision to exit is bound to have an impact on agents’ interest rate expectations throughout the whole yield curve, and lead to substantial portfolio reallocations. This is desirable, and inevitable, as the purpose of the exit strategy is indeed to change the behaviour of the private sector. But if the adjustment in expectations is large, it can lead to disruptions that affect financial stability and in turn jeopardise the sustainability of the whole exit strategy. In particular, any late exit which needs to be accelerated could produce unexpected shocks, for instance in the form of losses in the fixed income market which could hamper attempts by financial institutions to progressively regain access to the funding market. The effects of such a delay might lead to a prolongation of the non-standard measures which were supposed to be phased out. Remember the 1994 episode, when a mere 25-basis-point increase in the Fed funds rate gave rise to major losses in the bond market (reflecting the fact that it was probably a bit too late). The losses then constrained the future moves, which had to be relatively more gradualistic.
The second issue relates to the uncertainties surrounding the analytical framework that central banks have at their disposal in order to decide on the optimal interest rate path. Ideally, the stance of monetary policy should be calibrated in such a way as to avoid an implicit easing of monetary conditions as the economy recovers. This implies that the interest rate should be raised as the economy starts growing above its potential rate and as the output gap starts closing. The problem with this framework is that it is very difficult to estimate potential growth and the size of the output gap, especially after a shock like the one experienced over the last two years. Looking at the past, the major policy mistakes have been made as a result of overestimating potential growth and the size of the output gap, in particular during a recovery. This has led monetary policy to be overly activist and has delayed the tightening phase, thus fuelling financial instability.
The two-pillar strategy of the ECB has helped to reduce this type of risk, as monetary and credit aggregates have been used as indicators of a strengthened economy, while inflationary pressures were still subdued. In the current phase, we might have to take into consideration the fact that the ongoing process of de-leveraging in the financial sector, which is likely to take place for some time as a result of the correction of the previous credit bubble, might affect the signalling content of money and credit aggregates.
To be sure, the tightening phase cannot wait until inflation materialises, but will have to precede it. This is even more the case this time around, given the very low level to which interest rates have been reduced, and the distance from the steady state level that will need to be attained. The latter is in itself very difficult to determine, given the uncertainties about some of the longer-term developments in the economy. This will, as usual, present some comunication challenges, such as having to answer the question commonly put to central bankers under these circumstances: “Why are you raising interest rates if there is no inflation?”
The pace of the tightening will obviously depend on the underlying conditions, in particular the speed of the recovery. Given the uncertainties, it would be inappropriate to commit to any specific path ex ante.
Another issue to take into account is the exit strategy from fiscal policy. I will not dwell on this issue, which is undoubtedly a challenging one per se. But there is a link with the exit from monetary policy and it will have to be considered. In particular, the more delayed the fiscal exit, ceteris paribus, the more the monetary policy exit might have to be brought forward. Indeed, given the level of the debt accumulated in most advanced economies, any delay in the fiscal exit is likely to have an effect on inflation expectations, and may even disanchor them. This is a risk that monetary policy cannot take, as it would undermine its overall strategy.
On the subject of inflation expectations, allow me to digress briefly. Inflation expectations are an important indicator for central banks, as they reflect their credibility in achieving price stability. However, recent experience suggests that central banks should not be too complacent, even when expectations appear to be anchored, as the private sector might not always be a good predictor of future developments. Over the previous cycle, inflation expectations seemed to be well anchored, but partially concealed the build-up of asset market instability and longer-run threats to price stability. Well-anchored inflation expectations are a necessary condition for an appropriate monetary policy, but not a sufficient one.
Let me touch on a final issue, related to financial stability. The exit strategy for the interest rate policy will be defined on the basis of the primary objective: price stability. In this respect, financial stability can only be a secondary objective. If it were given the same priority as price stability, the latter would obviously be compromised.
The exit from the non-standard measures is likely to be linked to the state of the financial markets, and in this respect can partly be disconnected from the interest rate policy. Given the design of the non-standard instruments implemented by the ECB, the exit from those instruments can take place before or after the interest rate decision, without major effects on it. However, if at the time of the exit a number of financial institutions are still addicted to central bank liquidity, the transmission channel of monetary policy might be impaired.
It is not the central bank’s task to continue providing liquidity to financial institutions which are not able to stand on their own feet, once the turmoil is over. It is the responsibility of the supervisory authorities, and ultimately of Treasuries, to address the problems of these addicted banks as soon as possible, through recapitalisation and restructuring, as appropriate, and to ensure that all banks in their jurisdictions can stand on their own feet even without the central bank’s facilities.
Let me conclude.
I have tried to identify some of the issues which will have to be considered in implementing the exit strategy on both sides of the Atlantic. My observations have been purely analytical, and do not aim to provide any insight into specific monetary policy decisions over the next few months. My purpose was to explain parts of the analytical framework in which central banks will have to operate, on this and that side of the Atlantic.
Just like the entry, the exit will also call for close interaction between the monetary authorities. This might not necessarily mean coordinated action or similar measures, given the different situations. However, the challenges ahead are quite similar.
Ultimately, if we want to avoid ‘being too late’ or ‘being too early’, the only solution is to ‘be right on time’. And this is what we are committed to doing.
 On Friday 4 September 2009, three, six and 12 month rates in the US (LIBOR) stood at 0.31%, 0.71% and 1.29%. The corresponding rates in the euro area (EURIBOR) were 0.80%, 1.07% and 1.28%.
 See the speech by ECB President Jean-Claude Trichet, “The ECB’s exit strategy”, Frankfurt, 4 September 2009, and testimony by Federal Reserve Chairman Ben S. Bernanke before the Committee on Financial Services, US House of Representatives, Washington, D.C., 21 July 2009
Reproduction is permitted provided that the source is acknowledged.