Ladies and Gentlemen, 
It is a great pleasure to attend once again the Geneva Conference and discuss the important issue of exit from non-conventional monetary policies. Let me start with an important word of caution. We are still struggling with the euro area crisis. Yesterday’s decisions by the Governing Council of the ECB on reduction of the key interest rates, on modalities of liquidity provision and on possible ways forward to enhance the availability of credit to the broad economy should support prospects for a recovery later this year. But current underlying economic conditions still imply that the monetary policy stance will remain accommodative for as long as needed.
That said, exit from non-conventional policies will come one day, and lessons drawn from the crisis times will help us chart out our path. The debate on exit has chiefly concentrated on strategies: that is, on monetary policy reaction functions on the way out of these policies and over the new steady state. But it is fair to say that – with only few exceptions – participants in the debate have been less active in discussing the operating framework of monetary policy. Here, a few interesting questions remain unanswered. Will central banks go back to their pre-crisis operating framework once conditions return to normal? Or should some of the novel instruments remain in their toolbox ready for use, or even as standing instruments replacing old practices?
Let me start by mentioning the three main developments that, in my view, took place during the crisis:
First, central banks collectively moved towards instruments that can effectively disentangle interest rate decisions from decisions concerning the size of their balance sheets. One can debate whether this constitutes genuine bifurcation of instruments. And, with Tinbergen in mind, one could also conjecture that, in this way, the traditional assignment of the interest rate instrument to the price stability objective could be reinforced by an independent lever which, through its control over the volumes of central bank credit, would independently promote stable financial conditions. Here, other central banks had to innovate more than the ECB. The Fed, for example, acquired the authority to pay interest on reserves and this placed a floor – or, more precisely, a lower grey zone of values – on the level which the Fed funds rate can take on in daily overnight trading. The ECB, by contrast, entered the crisis with a corridor system and an elastic operating framework which did not need much tweaking to offset shocks. But we did shift from a competitive auction system, in which quantities are controlled by the central bank, to one in which we fix the price and take the quantities as endogenous in liquidity-providing operations. And this meant at some point that the overnight rate was left to drift down to the floor of the corridor, where it still stands. This, in turn, implies that the overnight rate has become more independent of the volume of excess liquidity.
Second important development during the crisis: central banks in a large part of the world engaged in non-conventional measures aimed at absorbing liquidity risk and duration risk from the market. Those central banks engaged in quantitative easing have focused on duration risk and have targeted term premia. The ECB has primarily concentrated on liquidity risk. As I said, we decided to fully accommodate banks’ demand for liquidity in an elastic manner. We also expanded the list of eligible collateral, so that banks could more easily liquefy their balance sheets and mobilise assets that had become scarcely tradable for liquidity purposes with us. The ECB took care of duration risk (or in our case funding risk), indirectly, to the extent that we replaced banks’ intermediate-maturity wholesale market borrowing with our longer- term refinancing operations. As a result, the duration risk vis-à-vis banks that markets did not want to bear migrated, to a certain extent and temporarily, to our balance sheet.
Finally, the third important development during the crisis was that central banks had to substitute for the sudden disruption of interbank market activity and became de facto the “money market intermediary” of last – and sometimes first - resort. For the ECB this was facilitated by the broad range of counterparties accepted in monetary policy operations and its broad collateral framework. Other central banks had to innovate in order to step up their intermediation role by resorting to specific targeted facilities outside their standard operating environment.
Modalities and parameters of intervention were adapted to local circumstances, such as the depth and liquidity of financial markets and the degree of intermediation in the financing of the economy, which differ considerably across major economies. Some would argue that it is likely that most of the new instruments may not be needed to the same extent once money market conditions - and more generally the transmission channels of monetary policy- return to normal. But is it so likely? This immediately brings up the question that the panel is posing: “Where to exit to?”. We may want to ask ourselves which elements of our current mode of operation have served us well, and not only as a crisis management tool, and may therefore survive as a more permanent feature of the new steady state.
So, let me share some thoughts on what remains unresolved about the issues that I mentioned at the start. I see four open questions, and I want to be very honest: I will not give answers. First question: the size and degree of symmetry of the interest rate corridor. Second question: the duration and maturity of central bank operations. Third question: the mode of the liquidity provision to banks. And fourth, the shifts in amplitude and quality of collateral pools, and the risk management implications associated with these shifts.
Let me start by discussing the interest rate corridor. As I mentioned already, the ECB entered the crisis with a tradition of conducting policy through a corridor – others would say, a “channel” – system. The ECB’s corridor is determined by the two overnight standing facilities: the marginal lending facility and the deposit facility. These standing facilities were set symmetrically around the main policy rate charged on central bank credit at the weekly refinancing operations. Prior to the crisis, the quantity provision at our weekly operations was calibrated in such a way that, given banks’ liquidity needs, the overnight money market rate would settle close to the policy rate. So, the corridor was symmetric de jure and de facto: liquidity volumes were just sufficient to validate the mid-point of the corridor as the price of liquidity to which market overnight rates would gravitate on average.
At the onset of the crisis, we moved quickly from a system of liquidity provision geared towards filling the system’s liquidity deficit towards a set-up in which banks operate under aggregate excess liquidity. So, following the shift to a fixed-rate full allotment mode of liquidity provision in October 2008, overnight money market activity rapidly occupied the lower half of the corridor, and in fact, with few interruptions, the overnight rate has been close to the floor of it ever since. For all practical purposes, the relevant portion of the corridor for daily market activity has shrunk to the space existing between the rate on the main refinancing operations – on the upside – and the rate on the deposit facility – on the downside.
Should the corridor be brought back to its standard width – spanning the entire distance between the marginal lending rate and the deposit facility rate – not only de jure but also de facto? In other words: should we go back to the standard quantity-based technology of liquidity provision which was founded on the principle of balanced liquidity conditions? Or should we maintain excess liquidity and operate the system through parallel adjustments to the deposit facility rate and the rate on the main refinancing operations?
I am asking this question, because observers – Marvin Goodfriend , for example – have been arguing for a system where the monetary authorities adjust a policy floor – the rate paid on reserves or the deposit facility rate – and the system is saturated with excess liquidity, so that overnight market rates are kept close to the floor.
Let me expound the merits and drawbacks of the discussed system, again upon the premise that I do not have a quantitative scale to weigh the ones against the others.
First advantage: the volatility of the overnight rate would likely be squeezed to minuscule numbers. True, our overnight rate has been more – rather than less – volatile since it moved to the floor of the corridor. But it is also true that excess liquidity conditions have been unsteady and uncertain in the past. Here, I imagine a scenario in which the liquidity supply would be kept steady and reasonably predictable. A more stable overnight rate, it has been argued, could enhance the transmission of policy shifts – adjustments to the deposit facility rate – throughout the term structure of money market rates.
Second advantage: the central bank would maintain one important acquisition from the crisis period: that is, its capacity to disentangle interest rate decisions from decisions concerning the scale of its own liquidity operations in order to remain resilient to large-scale liquidity shocks. This possibility may come in handy at the time of exit, in situations when inflation risks would call for a tightening of policy, while concerns about the fragility of the banking system would suggest prolonging the horizon for easy access to central bank credit. But even in the steady state – after the exit – the new system would be marginally more resilient to large liquidity shocks and acute liquidity stress in the money market. Banks would receive funding certainty, as their liquidity needs would be met. But there are drawbacks as well. First, for the ECB at least, this would require banks to be willing to hold reserves in excess of their needs. In a normal situation, absent any financial stress, this is difficult to ensure in a system (like ours) in which liquidity injections are the consequence of counterparties’ – not the central bank’s – decisions. Second, importantly, a system of permanent excess liquidity could potentially distort signals of liquidity and credit risk in the money market, as money market activity could be compressed – or suppressed altogether. I will come back to this point in a short while.
Let me turn to the second issue, the duration and maturity of central bank operations. In our case, before the crisis, the bulk of liquidity was provided through operations with a maturity of one week. We also provided (pre-set) amounts through longer-term operations, but to a lesser extent and with a maximum maturity of three months. As an alternative, one may conceive of frameworks in which the menu of maturities open for bank borrowing would be much richer. For example, there could be merits in matching the more structural long-term trends in the autonomous liquidity factors, most notably the growth in currency in circulation, with corresponding instruments of structural liquidity supply at longer maturities. Expanding the range of maturities following the exit would probably go hand-in-hand with the decision to return to competitive auctions. But such a steady state framework would be nevertheless more robust to renewed bouts of perceived funding risk in the market.
Let me turn to the third – very much related – issue: the most appropriate mode of liquidity provision and the desirability of market versus central bank intermediation.
Before the crisis, the intermediation role of the ECB was limited to filling the structural “liquidity deficit” of all the banks vis-à-vis the monetary authority. The pre-crisis weekly liquidity provision through variable rate competitive auctions was thought to be best suited to reveal banks’ “true” liquidity demand, incentivise interbank transactions, and enhance market scrutiny of banks’ credit standing, beyond the quality of the collateral that they could pledge in our lending operations.
But, during the crisis, we have moved to a system in which the central bank is de facto the “market intermediary” of last – and sometimes – first resort. This has further weakened banks’ incentives to trade liquidity in the market. Do we need them to trade in the very short-term market in the first place?
Well, this is a fundamental issue for optimal mechanism design. It is not entirely clear whether an interbank market is strictly needed from a welfare perspective. The key function of markets is price discovery. But here the price to be discovered is not only the starting point of parties’ negotiations, but also the end-point of the market tâtonnement: it is the price pre-set by the central bank! So, some would argue: why do we need a market if the price is pre-determined? Of course, I am over-simplifying. The welfare function of the secured money market probably differs and is less important than that of the unsecured money market, where the market exercises a type of discipline that a central bank backstop would not be able to provide.
Let's not forget moral hazard considerations. In stressed systemic conditions, banks may well be deprived of access to interbank credit for reasons that have little to do with their own merit of credit. These are conditions that require central bank interventions to replace the market, when this fails. But in normal times, banks which are kept out may be those that deserve credit restrictions. In this case, a central bank taking away market pressure would distort the market, not repair a market imperfection.
Also, allowing the price to be determined by competitive forces may have the collateral merit of revealing market information which is certainly of great value to the central bank. Maintaining an active market for liquidity may be an important pre-condition for the central bank to monitor important variables, such as the evolving market standing of single institutions and the dynamic process by which banks evaluate and price liquidity risk.
Finally, let me briefly address the fourth and last issue: the collateral framework. The crisis has underscored the importance of the collateral framework for monetary policy implementation. I believe there is still much to learn in this area. Ideally, any decision on the collateral framework under a new steady state should be based on a thorough examination of the impact of collateral valuation and haircuts on the size and composition of banks’ balance sheets, and the ramifications for the stability of the financial sector as a whole. Such an examination cannot be conducted in isolation from a review of collateral rules followed in private transactions, which could be affected by the on-going regulatory overhaul, in particular the migration of over-the-counter derivatives to clearing houses.
The ECB’s operational framework, conducting monetary policy operations as temporary refinancing operations in a corridor system, has not only served us well. It has also provided us with enormous flexibility during the crisis. By adjusting the price, amount, maturity, allotment procedures and eligible collateral for our operations we have been effective in containing funding risks of banks and in addressing distress in the money market. In a sense, the system can well be rolled back to the previous steady state, with the comforting expectation that this flexibility is a built-in permanent mechanism that restores itself again, elastically, in case of need. Yet, some of the questions I posed may be worth reflecting upon. And I am grateful to the organisers of this conference for stimulating discussion of these issues.
Before turning the floor to the other panellists, let me briefly add some further points on the interactions between exiting non-conventional monetary policy and financial stability.
Exiting the crisis mode could pose risks to financial stability in the current environment, which is marked by a prolonged period of low yields and reduced volatility, making it challenging for the financial sector to properly accommodate interest rate risk. The exit could be associated with a steepening of the yield curve as expectations of low short-term rates are reversed and central banks reduce their holdings of long-term securities. Uncertainty on the path of exit from unconventional policies may trigger a rise in volatility at the long end of the yield curve, exposing banks and investors to substantial losses. These effects would be more pronounced if the speed of interest rate adjustment were to exceed market expectations.
At the same time, delaying the exit from “crisis” monetary policies beyond what central banks’ reaction function would warrant could also entail risks to financial stability, by inducing a further build-up of the very same exposures that render exit more challenging in the current environment. In particular, a protracted period of low interest rates and ample liquidity could compromise the market mechanisms in efficiently allocating resources, hence encouraging the roll-over of loans to non-profitable businesses and weakening incentives for balance sheet repair. Prolonging low interest rates for a long time might raise the possibility of sudden shifts in market expectations and significant re-pricing of risks once indications of monetary policy tightening materialise.
Of course, these risks should also not be overrated at a time when financial risks are currently highly correlated with overall macroeconomic risks. Exit is very likely to take place in an environment where credit has recovered and risks have receded in the economy, making the financial sector more robust to interest rate shocks. The current situation of subdued bank lending and outlook for growth in the euro area is a case in point.
Yet, monetary policy-makers have to be aware of these challenges and need to guard against the risk of financial dominance by keeping a clear focus on their respective mandate. For the ECB, our “hardwired” focus on price stability provides a clear guidepost that has proved effective in anchoring inflation expectations over time. And the prominent role given to money and credit developments in our strategy ensures a more symmetric reaction to financial forces over the cycle.
At the same time, it is important that central banks continue to place particular emphasis on managing expectations. The ECB has all the instruments at hand to ensure a smooth exit if risks to medium term price stability should materialise. Our credit operations are temporary and the market has all the information to date to anticipate their unwinding and make adequate adjustments to prepare. If liquidity conditions should prove too lax while central bank credit is still abundant, we have available instruments of liquidity re-absorption to tighten money market. Some of these instruments have been tested successfully in the course of the crisis.
Finally, let me stress that a sound financial system is a necessary condition for an orderly exit, hence the importance of a swift implementation of the banking union in the euro area. This should include not only the Single Supervisory Mechanism but also a Single Resolution Mechanism, with a single resolution fund based on contributions by the industry and a common, last-resort fiscal backstop. This also includes the assessment of banks’ balance sheets that will be undertaken in the transition to the Single Supervisory Mechanism.
I thank you for your attention.
I wish to thank Massimo Rostagno, Tobias Linzert and Arthur Saint-Guilhem for their contributions to these remarks. I remain solely responsible for the opinions contains herein.
Marvin Goodfriend, “Interest on Reserves and Monetary Policy”, FRBNY Economic Policy Review, May 2002.
Reproduction is permitted provided that the source is acknowledged.