Central banks and the challenges of the zero lower bound

Intervention by Benoît Cœuré, Member of the Executive Board of the ECB,
at the “Meeting on the Financial Crisis”,
hosted by the Initiative on Global Markets,
at the University of Chicago Booth School of Business,
Miami, 19 February 2012

Ladies and Gentlemen, [1]

It is a great pleasure for me to contribute to this final session on central bank communication and the zero lower bound.

Faced with the crisis, central banks had to enter new territory. We had to design strategies of intervention that had not been studied much.

I will expand on such strategies, starting from facts and then proceeding to policy approaches. Most of you will find nothing new in what I’m about to say. But a policy-maker’s perspective on facts and on policy approaches might help to establish a better understanding between the theory and practice of monetary policy, and on the specific features of the euro area.

Central banks in the crisis

Let me start with an overview of central bank actions during the crisis. Central banks responded to what turned out to be the worst financial crisis since the Great Depression by taking unprecedented measures, both “standard” and “non-standard”. “Standard” is what you learn in “Econ 1”: a central bank moves its policy rate to influence monetary conditions when an adjustment is needed and uses liquidity interventions only at the margin, to ensure that short-term money market rates remain close to their desired value. “Non-standard” includes any other action that a central bank can take to try and influence monetary conditions when the first avenue is impractical or insufficient. On the “standard” side, central banks have reduced their main policy rates, as well as the rates at which they remunerate reserve balances, to very low levels [2]. For example, in the euro area, the interest rate on the Eurosystem’s main refinancing operation currently stands at a historically low 1%. The rate at which the Eurosystem remunerates liquidity placed on its deposit facility is even lower, at a quarter of 1%. In these conditions, and reflecting large amounts of excess liquidity, the overnight interest rate (as measured by the EONIA) has remained – for some time now – below the main refinancing rate, and closer to the deposit facility rate.

On the “non-standard” side, central banks have been experimenting in various ways. The Fed, the Bank of England and the Bank of Japan have opted in various guises for quantitative easing. As for the ECB, our measures have been tailored to the specificities of the euro area. This means assigning banks the central role they play in credit intermediation, with around two-thirds of euro area firms’ external financing intermediated by banks in recent years. This means that any impairment to bank lending will have more serious consequences in the euro area than in other economies where market sources of funding are more important. Accordingly, the ECB’s policies have targeted banks’ funding conditions – by cutting the cost at which banks can borrow and by enhancing the security and stability of their funding over increasing longer horizons. With the exception of the covered bonds purchase programme, the ECB has avoided setting quantitative targets.

By providing liquidity to the banking sector, particularly in the first phase of the crisis, we have “liquefied” otherwise illiquid assets. This has avoided fears of fire sales and forestalled disorderly deleveraging, which in turn would have risked dragging the euro area economy into an ever deeper recession. I will come back later to the long-term implications of this crisis intervention.

In the second stage of the crisis, low liquidity in some segments of the government securities markets imposed further strains on banks’ funding conditions, due to the widespread use of these securities in bank secured lending and as liquidity buffers. Contagion forces quickly found in banks’ exposure to sovereign securities a formidable transmission and amplification channel, threatening the stability of banks and the whole financial system. In these circumstances, we, together with other central banks, have increased our lending to replace the withdrawal of private lending in inter-bank transactions and, sometimes, in the broader market for capital. These interventions have served to shore up savers’ confidence in the financial sector and, ultimately, of the currency. They are of course no substitute for the far-reaching reforms needed in fiscal governance and competitiveness, and for banks strengthening their capital basis.

Whether it’s the direct effect of an active policy of liquidity supply under quantitative easing or the collateral implication of banks’ hoarding behaviours in the euro area, abundant liquidity has gone hand in hand with the cost of borrowing in the money market drifting down to levels close to zero. A combination of historically low interest rates and stagnant or declining output levels has heightened concerns that central banks’ policies of liquidity provisions may be ineffective and futile, as banks amass “idle” cash rather than lend it out. Some observers fear that a liquidity trap as described by Keynes in the General Theory is looming: banks might be afraid to acquire exposures to any institutions apart from the central bank, in an environment in which interest rates can only be expected to increase. These views have led to various proposals, all of which aim to discourage banks from holding excess reserves [3].

In the euro area context, the fear of a liquidity trap seems to me largely unwarranted. First, even though euro area output growth is likely to have been very weak in the last quarter of 2011, there are tentative signs of a stabilisation in economic activity according to the most recent survey data. Looking ahead, we expect the euro area economy to recover very gradually in the course of 2012. Second, euro area inflation is likely to stay above 2% for several months to come, before declining to below 2%. And third, the substantial build-up of reserves on the liability side of the Eurosystem’s balance sheet simply reflects the accounting counterpart of large-scale liquidity provisions by the central bank. In itself, it does not convey information on the actual effect that such liquidity supply may have on the economy more broadly [4]. In fact, it does not provide information on the economic trades that any euro injected into the system might generate before returning to the central bank as a temporary cash buffer of the bank that receives the cash at the end of what might be a long transaction cycle. This transaction cycle can be an active contributor to the revitalisation of the economy.

The example of the recent first three-year LTRO conducted by the Eurosystem in late December 2011 corroborates this view. We have indications that the liquidity provided through this operation is indeed flowing, or will flow, between economic agents. [5] Also, banks borrowing from the Eurosystem are not necessarily the same as those that “park” money in our deposit facility. We have also seen interest rates declining all along the yield curve, initially in the short segment and then at longer horizons. Overall, these developments suggest that such non standard-type measures are likely to have contributed materially to preventing a disorderly de-leveraging of the banking sector as a whole and, through that channel, a far more serious credit crunch.

Let me now turn to the consequences of a protracted period of low or even negative interest rates.

Low interest rates – current and expected

I mentioned how abundant liquidity in money markets has compressed money market rates to levels that are considered a floor for “standard” policy instruments. One obvious question is: why should there be a floor, a “lower bound”, in the first place?

Some economists have called for monetary policy makers to push the interest rate at which banks can borrow from the central bank into negative territory [6]. From a technical point of view, there is in fact nothing that prevents central banks from paying negative rates for the security they offer to depositors, at least temporarily [7]. Given the costs associated with holding large amounts of banknotes, it is likely that significantly negative interest rates would be required to trigger a switch from money holding to investment in banknotes. So, there seems to be technical leeway.

At the same time, however, negative interest rate trading on, for example, Treasury bills, or – for all practical purposes – trading at interest rates not significantly different from zero certainly necessitates adjustment of market practices. [8] And there is a degree of hysteresis: a temporary situation of zero or negative interest rates can have long-lasting implications for banks and their trading incentives. In theory, everybody would love to borrow at zero or negative rates without going through the pain of finding a creditor happy to take the opposite side of the transaction. If the central bank offers this service systematically, banks can dismantle their trading platforms – which are costly to maintain – and become addicted to central bank credit. The Japanese experience is of course relevant here. [9] Within three months of the introduction of the zero interest rate policy, the volume of transactions declined by around one-half and low turnover persisted until end-2006 when the zero interest rate policy was discontinued. To be sure, it is difficult to disentangle the impact on money market turnover of zero interest rates from other important determinants, such as increased counterparty risk or ample liquidity provision by the central bank, or the uncertain macroeconomic environment that led to the central bank decision in the first place. Overall, historical evidence does not point unambiguously in one direction. But if a protracted period of zero or negative interest rates were to be experienced in the euro area, it would be particularly important not to lose the perspective, and the possibility, of restarting the interbank market at a later stage. The intermediation role taken by the central bank cannot, and should not forever take the place of money market activity [10].

Let me mention two further areas where risks could emerge from zero or even negative interest rates.

First, money market activity might suffer beyond the trading incentives effect. Important market intermediaries, such as money market funds, could be driven out of business, as their business model loses profitability, for both domestic and foreign investors with excess liquidity may shift their investments to alternative, more profitable market segments.

Second, zero or negative interest rates may produce adverse effects on the profitability of commercial banks and financial intermediaries more broadly. In a financial crisis this can result in a credit contraction. The reason is well known. Commercial banks may have to keep retail deposit rates unchanged to preserve their deposit base in a context of increased retail funding competition and substitution by banknotes [11]. But, with lending rates reduced, this hurts banks’ profitability, especially in countries where variable rates predominate. This impact could be particularly pronounced in the case of a widening of the policy rate corridor, with the main refinancing rate remaining unchanged and the remuneration of the deposit facility being adjusted to zero or below. Commercial bank equity debasement, in turn, can reduce the effectiveness of “non-standard” monetary policy and may even provide incentives for money creation to take place in the shadow banking system. Other channels – going in the opposite direction – are also conceivable, with search-for-yield considerations prevailing and leading to higher short-term profitability. All in all, the impact of zero rates on the profitability of banks remains uncertain and highly dependent, among other determinants, on parallel regulatory response. [12]

Overall, a switch to zero or negative interest rates bears some risks – mainly of a microeconomic nature – which would have to be weighed against potential benefits in terms of additional macroeconomic stimuli. It also has to be noted that such steps would be warranted only in the face of clear downward risks to price stability, which today are not present in the euro area. In particular, the discussion about deflation risks remains largely speculative.

Central bank communication and the zero lower bound

Some analysts argue that, when interest rates reach their effective lower bound, central banks should indicate their future intentions more precisely than they do under normal circumstances. Because bondholders are exposed to a risk of capital loss due to the uncertainty about the timing and magnitude of the rise in short-term interest rates that would accompany a future recovery, it is argued that central bank communication could provide insurance against that risk. This is consistent with the adoption of forward guidance by some central banks. It is also reflected in the recent decision by the Federal Reserve to release FOMC participants’ projections of the appropriate level of the target of the federal funds rate, including the likely timing of the first increase in the target rate [13].

Risks and mitigants

A central bank which promises to maintain interest rates low for a prolonged period entails some risks, a point which economic theory has started to acknowledge and address. Raghuram Rajan has been studying these risks for quite some time. The belief that interest rates will stay low for long periods might lead banks to make excessive liquidity promises and increase the future need for low rates, and thus sow the seeds of future crises. [14]

A second source of risk stems from the implications of low interest rates for balance sheet adjustment in the financial and non-financial sectors. While low policy rates may in the short term help to prevent a disorderly adjustment of balance sheets and provide some relief in terms of lower interest payments, they also weaken the incentives for repairing balance sheets in the first place [15].

Thirdly, when confronted with forward guidance, the public may fail to understand the conditionality of the information that is given by the central bank. Such a mismatch between the conditionality of the statement and the public’s expectations can put at risk the credibility and reputation of the central bank, in the event an ex post change in circumstances were to necessitate a change in policies.

Finally, the publication of an interest rate path, as a way to reinforce forward guidance, can pre-empt private sector expectations and reduce their information content, thereby crowding out valuable private information. Conversely, if financial market expectations differ from the path announced by the central bank, the latter might be perceived as having a credibility or reputation problem.

Overall, and as suggested by numerous studies [16], [17], the costs of an anticipated protracted low rate policy – in terms of fuelling risk-taking, promoting an overhang of non-performing loans and prolonging future imbalances – can be significant.

What are the mitigants in the euro area? I mentioned already our main approach to “non-standard” policy action: provision of unlimited liquidity over long-term maturities. I note here that this liquidity is provided through reverse repurchase agreements and not through outright purchases of securities, and at low but adjustable [18] interest rates. In other words, the Eurosystem provides liquidity upfront, while being clear from the start that the cost of that liquidity can only be determined ex post, as economic conditions unfold. There is no trade-off between ex ante commitments and ex post constraints to policies in our “non-standard” operations.

In the end, many of the risks we see in central banks’ crisis management practices can be greatly mitigated by an independent statute, a sharp quantitative definition of the price stability objective and a particular emphasis on monetary and financial indicators, all of which help to avoid pro-cyclicality whenever inflation poses less of a threat than financial instability. This latter dimension, for example, allows a more restrictive monetary policy to be pursued during a period of booming asset prices, even in an environment of relatively subdued inflationary pressures.

Concluding remarks

The financial crisis has posed challenges for us all. For its part, the ECB has responded to such challenges, by reducing its key interest rates to very low levels as well as by introducing a number of non-standard measures to support funding conditions for banks and thereby maintaining the transmission of its monetary policy. These measures have been tailored to the specificities of the euro area. The ECB’s standard and non-standard measures ensure that its stance can be adjusted in time to counteract risks to price stability, while addressing remaining impairments to the transmission mechanism. It remains however crucial that the implementation of these measures does not hamper the chances of reviving currently dysfunctional market segments, such as the interbank market. It is also crucial that governments and regulators in specific countries and sectors within the euro area address ongoing funding strains.

[1]I wish to thank Arthur Saint-Guilhem for his contributions to the speech. I remain solely responsible for the opinions contained herein.

[2]For a review of central banks’ experience with remuneration of reserves, see Bowman D., Gagnon E., and M. Leahy, 2010. “Interest on Excess Reserves as a Monetary Policy Instrument: The Experience of Foreign Central Banks”, Board of Governors of the Federal Reserve System, International Finance Discussion Papers, No 996, March 2010.

[3]Such proposals include a tax on excess reserves (see Sumner, S., 2009. “Comment on Brad Delong: Can We Generate Controlled Reflation in a Liquidity Trap?”. The Economists’ Vol. 6, Issue 4, Article 7), or a cap on the amount of reserves a bank would be allowed to hold (see Dasgupta, S. 2009. “Comment on Luigi Zingales: Why Not Consider Maximum Reserve Ratios?”. The Economists’ Vol. 6, Issue 4, Article 6).

[4]For a detailed exposition in the case of the US, see: Keister T. and J. McAndrews. “Why Are Banks Holding So Many Excess Reserves?”. Federal Reserve Bank of New York, Current Issues in Economics and Finance, Vol. 15, No 8, December 2009.

[5]There are indications that one of the main reasons why banks participated in the three-year LTRO was to front-load their funding needs for 2012. The increase of liquidity provision through the three-year LTRO was about €200 billion (as of the settlement day of the first three-year LTRO). This figure has to be contrasted with total banks’ maturing debt of around €210 billion for the first quarter of 2012.

[6]See Mankiw, N. G., 2009. “It May Be Time for the Fed to Go Negative.” New York Times, 19 April. See also W. Buiter, 2009. “Negative interest rates: when are they coming to a central bank near you?”, May 2009. The traditional literature on the zero lower bound focuses on monetary policy rules and offers limited insight into today’s developments since it does not consider the possible role of unconventional measures. See Coenen G., Orphanides A. and W. Wieland, 2004. “Price stability and monetary policy effectiveness when nominal interest rates are bounded at zero”. Advances in Macroeconomics, 2004, 4(1), 1187-1187.

[7]Negative interest rates may require significant adjustment in market practices, with potentially large disruptions attached. For example, one episode of negative interest rates on the US Treasury repo market in 2004 was associated with a sharp increase in settlement fails, to a point where the issue totally stopped clearing. See Fleming, M.J. and K.D. Garbade, 2004. “Repurchase Agreements with Negative Interest Rates”. Federal Reserve Bank of New York, Current Issues in Economics and Finance, Vol. 10, Number 5, April 2004. More recently, another episode of negative interest rates on the US Treasury market led to the implementation of a fails charge for the settlement of Treasury Securities, one example of an institutional adaptation required for markets to function at very low interest rates. See K. D. Garbade, F. M. Keane, L. Logan, A. Stokes, and J. Wolgemuth, 2010. “The Introduction of the TMPG Fails Charge for U.S. Treasury Securities”. Federal Reserve Bank of New York, Economic Policy Review, October 2010.

[8]An interesting example in this respect is the recent adjustment of the bidding rules for German Bubills in primary auctions. As of 9 January 2012 counterparties are required to bid in price terms instead of yields, which allows for the submission of above-par bids that reflect a negative yield. In addition, there is anecdotal evidence that some electronic trading systems for secondary market trading also require adjustments to cope with negative rates and it is possible that the proprietary systems of some market participants might need changes.

[9]See T. Fukui, “Challenges for Monetary Policy in Japan”, speech at the Spring Meeting of the Japan Society of Monetary Economics, Tokyo, 1 June 2003.

[10]See L. Bini-Smaghi, ‘Restarting a Market: The Case of the Interbank Market’, speech given at the ECB Conference on Global Financial Linkages, Transmission of Shocks and Asset Prices, Frankfurt, 1 December 2008.

[11]Unless a demurrage rate or regular stamping are imposed on banknotes, as proposed by Irving Fischer and others.

[12]See IMF 2005, Japan Article IV Staff Report, “Why is Japanese Banking Sector Profitability so Low?”, pp.54-64, August 2005.

[13]In December 2011 the FOMC decided to release participants’ projections of the appropriate level of the target of the federal funds rate in its Summary of Economic Projections, and to include the likely timing of the first increase in the target rate given the members’ projections of future economic conditions. The FOMC also adopted a precise quantification of the inflation objective of 2%.

[14]See Diamond D and R. Rajan, 2011, “Illiquid Banks, Financial Stability, and Interest Rate Policy”, mimeo, University of Chicago and NBER, April 2011.

[15]Experience from Japan in the 1990s has shown that low policy rates indeed fostered “ever-greening”, or rolling-over of non-viable loans. By practising this, banks in Japan could preserve capital, but at the cost of delayed restructuring. See Caballero R., Hoshi T. and A. Kashyap, 2008, “Zombie Lending and Depressed Restructuring in Japan,” American Economic Review, 98, pp. 1943–77. The Japanese episode also suggests that “evergreening” with low interest rates becomes even more pervasive in the case of lax banking regulation and forbearance.

[16]Several studies suggest that extensive monetary policy accommodation over a protracted period can impact on risk-taking and leverage via several channels. First, lower interest rates tend to be associated with a feedback loop between asset prices, leverage and risk premia triggered by pro-cyclical fluctuations in collateral valuations. Second, widespread use of value-at-risk methodologies for regulatory capital purposes may act as an amplification mechanism whereby rising asset prices lead to further investment in risky assets and a further narrowing of risk premia (see Adrian T. and H. Shin, 2011, “Financial Intermediaries and Monetary Economics,” in Handbook of Monetary Economics, edited by B. Friedman and M. Woodford (ed.), pp. 601-650; Borio C. and H. Zhu, 2008, “Capital Regulation, Risk-Taking and Monetary Policy: A Missing Link in the Transmission Mechanism?,” BIS Working Paper Series, No. 268). Third, risky portfolios can also result from a “search for yield” by investors who target a nominal return and thus might be induced to move out when risk-free nominal returns become too low.

[17]These channels have been illustrated in theoretical models, but there is also some empirical evidence suggesting a significant relationship between developments in the financial side of the economy and accommodative monetary policy. For evidence based on individual loan data, see Jimenez G., Ongena S., Peydró J.L. and J. Saurina (2007), “Hazardous Times for Monetary Policy: What Do Twenty-Three Million Bank Loans Say About the Effects of Monetary Policy on Credit Risk?,” CEPR Discussion Paper Series, No 6514; for evidence based on individual bank level data, see Altunbas Y., Gambacorta L. and D. Marqués-Ibañez (2010), “Does monetary policy affect bank risk-taking?,” BIS Working Paper Series, No 298. See also BIS (2010), “Low interest rates: do the risks outweigh the rewards?”, BIS Annual Report, pp. 36-46.

[18]The rate in the long-term liquidity operations of the ECB is equal to the average rate of the main refinancing operations over the life of the respective operation. Interest is paid when the respective operation matures.

Media contacts