This panel is entitled Monetary Policy and Banking.¹ That takes in a lot of territory.  Rather than try to cover the waterfront, I am going to start with a single premise: Financial stability is a necessary prerequisite for an effective monetary policy.  There is a critical chain of linkages from monetary policy to banking and onwards to the real economy.  Financial stability is a necessary condition for those linkages to operate effectively.  Thus, it is a necessary condition for monetary policy to be able to achieve its economic objectives.  

I will then outline briefly what I see as some of the major implications that stem from that premise.  These include that the central bank has a major role to play in ensuring financial stability and should evaluate the stance of monetary policy in light of problems in the financial system that may impair the monetary policy transmission mechanism.² 

In my mind, the biggest lesson of the financial crisis has been that monetary policy cannot work properly when there is financial instability.  When financial instability occurs, it disturbs market functioning and can also impair bank balance sheets.  The result can be disruption to the financial intermediation function with resulting constraints on the availability of credit for households and businesses.  This, in turn, can lead to further reductions in aggregate demand that put additional stress on the weakened financial system.  Obviously, this is not a favorable dynamic.

Financial instability can impact the conduct of  monetary policy via three major channels.  First, financial instability can generate a sufficiently large shock to aggregate demand that the central bank may encounter the zero lower bound constraint—the constraint that the monetary policy instrument, for example, the federal funds rate—cannot easily be pushed below zero. In such circumstances, it may not be easy to fully offset the shock through the pursuit of a more stimulative monetary policy. At the zero bound, the central bank is not powerless, and may turn to other monetary policy tools such as forward guidance and large scale asset purchases.  But these tools may not be as effective as lowering the short-term rate instrument.  In particular, the central bank may not be willing to use these nonconventional tools to the full extent necessary to provide the same degree of stimulus as it would provide if it could set interest rates at negative levels.  That might be because of uncertainty about how nonconventional tools will work or because of the potential costs associated with the use of such tools in terms of market functioning and the risks of future financial instability.   

Second, financial instability can impair the linkage between monetary policy and financial conditions.  The central bank may move to a much more accommodative monetary policy stance, but this may not lead to much improvement in financial conditions. We see this clearly in spreads and risk premia during periods of financial stress. Depletion of capital in the banking system or the interaction of frictions and market failures may also lead to constraints on the availability of credit following a financial shock.  As a recent U.S. example of the second issue, households with lower FICO scores have had difficulties in refinancing their mortgages or in obtaining new mortgages to purchase a home.  This stems from worries by lenders about put back risk—that such mortgages might be put back to them by Fannie Mae or Freddie Mac should the mortgages become nonperforming.  For any given reduction in mortgage rates relative to normal levels, this reduces the support for the housing sector and consumption.   Thus, another lesson of the crisis is that monetary policymakers need to be more attuned to how the condition of the financial system influences how monetary policy changes affect financial conditions. 

Third, the linkage between financial conditions and aggregate demand may also be impaired.  For example, if an asset bubble has caused a buildup of debt that now needs to be reduced significantly, then lower interest rates that ease financial conditions may not stimulate aggregate demand very much.  Separately, in a protracted zero bound episode, the monetary policy impulse may become attenuated over time.  Lower interest rates may make financial conditions easier, lifting wealth, and encouraging households to shift spending from the future to the present.  But when the future arrives, spending may then be lower as a consequence.  The linkage between financial conditions and economic activity may vary depending on how long the accommodative monetary policy regime has been in place.

So what are the implications that stem from the fact that financial instability can impair the efficacy of monetary policy?   I think there are at least three major implications:

• The central bank has a strong stake in preventing financial instability. 
• When financial instability occurs, the central bank has an important role in taking steps to mitigate that financial instability.  This includes ensuring that credible liquidity backstops are available and forcing banks to strengthen themselves, for instance by raising additional capital. 
• The stance of monetary policy needs to be judged in light of how well the transmission channels of monetary policy are operating.  When financial instability has disrupted the monetary policy transmission channels, following simple rules based on long-term historical relationships can lead to an inappropriately tight monetary policy. 

Let me now consider each of these implications in turn.  The first implication, that the central bank has a strong stake in preventing financial instability, has several elements.  The central bank needs to be willing to respond to limit financial market bubbles from developing in the first place. This includes not just paying attention to asset price bubbles, but also to related excesses in leverage and in short-term funding markets.  As I noted in a speech a few years ago, this is difficult to do in practice.³  After all, bubbles are difficult to identify in real time and the central bank’s policy toolkit to deal with bubbles may be limited.  However, this difficulty cannot be an excuse for inaction.  Using the bully pulpit, implementing macroprudential measures, or adjusting monetary policy can generate superior results compared to inaction.     

The central bank also has to ensure that the financial system is sufficiently robust and resilient so that if there are financial market shocks, the banking system can absorb any shocks and continue to perform its credit intermediation function.  This has many elements including tough capital and liquidity requirements for banks and strong financial market infrastructures.  It also includes taking action to eliminate sources of vulnerabilities whenever they may arise—for example, strengthening the triparty repo system, reducing the vulnerability of money market mutual funds to runs, or forcing derivatives contracts to be standardized and centrally cleared. These actions should be undertaken during “peacetime” rather than when a crisis is already underway.

Finally, the central bank needs to look over the horizon and identify market practices and processes that could make the financial system more vulnerable to shocks or that could amplify such shocks. As an example, a tougher response to the poor mortgage underwriting practices evident during the U.S. housing bubble would have been appropriate in my opinion.  Or more attention to bank compensation practices could have mitigated the incentives for excessive risk-taking.   

The second implication is that when financial instability does occur, the central bank needs to respond forcefully to mitigate such instability.  This includes ensuring that appropriate liquidity backstops are available.  This is necessary in order to reassure counterparties that solvent institutions will have access to liquidity so that they can repay their obligations regardless of the degree of illiquidity or stress in the financial system.  As we saw in the fall of 2008, without appropriate backstops, markets can seize up completely even when counterparties are perceived as solvent.  That is because market participants tend to hoard liquidity during times of stress and because counterparties may not engage with one another when they are uncertain whether others will do so.  Central bank liquidity backstops act as a coordinating device, solving the collective action problem faced by private market participants.

However, in the most serious stress episodes, credible liquidity backstops alone will not always prove sufficient.  It is also necessary to take steps to restore the banking system to health as soon as possible.  This means forcing banks to cut their capital distributions early and to raise new capital in a timely way even if such capital-raising results in the forced dilution of existing shareholders. 

With respect to bank capital, we saw two unwelcome behaviors during the crisis that central banks and other banking regulators need to lean against.  Banks were reluctant to cut their capital distributions because of worries that this would signal weakness.  Of course, these capital distributions just made the banks and the banking system weaker. 

Also, we saw that banks were reluctant to raise capital to guard against particularly bad states of the world because they thought this might unnecessarily dilute their shareholders should such bad states of the world not materialize.  Of course, the collective action of banks to not raise this additional capital made the banking system weaker than it would otherwise have been. This in turn made the possibility of the particularly bad states of the world more likely. In such circumstances—which the U.S. faced in 2009 and arguably the Eurozone faces today—it is essential for the authorities to force sufficient capital into the system. The subtlety here is that in a time of stress it is necessary to overcapitalize the banking system relative to the base case scenario in order to ensure financial stability and an ongoing supply of credit to the real economy in a stress scenario. Doing this makes the bad scenario less likely to materialize and the improvement in credit supply helps to engineer a stronger recovery than would otherwise materialize.⁴

The third implication is that the stance of monetary policy needs to be assessed based on the state of the financial system and the real economy.  How the monetary policy impulse is transmitted to the real economy is not immutable but changes depending on how monetary policy affects financial conditions and how financial conditions affect economic activity. 

During and following financial crises, problems in the financial system can impair the transmission of monetary policy to the real economy.  When this happens, policy may need to be more accommodative than otherwise in order to achieve its objectives. 

The experiences of both Japan and United States are cases in point.⁵  In retrospect, we know that following the collapse of the property bubble and the investment boom in Japan in the 1990s, the Bank of Japan (BoJ) did not follow a sufficiently accommodative monetary policy to prevent deflation.  Although the BoJ response was appropriate relative to the economic forecasts that prevailed at the time, those forecasts proved much too optimistic. Thus, policy was insufficiently accommodative with the benefit of hindsight.  To a lesser degree, the same critique also applies to the United States.  Despite an aggressive shift towards greater monetary policy accommodation in 2008 and 2009, and ongoing subsequent easing—which has supported a return to growth and helped to facilitate needed adjustments in housing and household balance sheets—the economic recovery has been consistently weaker  than forecast.   As a result, the Federal Reserve has fallen short of meeting its employment and inflation objectives.  This suggests that with the benefit of hindsight, U.S. monetary policy, though aggressive by historic standards, was not sufficiently accommodative relative to the state of the economy.      

In this regard, I would caution against the mechanical use of monetary policy rules following a financial crisis.  In the United States, the so-called Taylor Rule receives considerable attention as a guide to policy.  But there are a number of reasons to believe that using it mechanically would result in too tight of a monetary policy setting. While discussing the shortcomings of the Taylor Rule is beyond the scope of my remarks today, let me make one simple point.  If the transmission channels of monetary policy are impaired, then the equilibrium real rate associated with monetary policy neutrality is likely to be considerably lower than the 2.25 percent real rate assumption embodied in the typical Taylor Rule formulation.  Monetary policy needs to take this onboard, even before considering other issues such as risk-management at the zero bound.

I have argued today that the central bank needs to have a major role in ensuring that the financial system is sound.  Without this, the central bank cannot ensure that monetary policy will be effective in achieving its objectives. This is the important linkage of monetary policy to banking. 

Now some express concern that involvement in banking regulation and supervision may lead a central bank to bias the setting of monetary policy to try to help troubled banks or other institutions. This fear is misplaced. Problems in the financial sector are taken into account for monetary policy purposes only to the extent that they affect the transmission of policy to the real economy. There is a separate tool kit to address problems at the level of the banks themselves.

Others will argue that having the central bank involved in the oversight and regulation of the financial system puts too much power in the hands of one authority. Another concern is that such a consolidation of responsibilities is dangerous for the central bank because it could threaten the central bank’s independence with respect to monetary policy.  I think these are legitimate issues.  But the bigger risks, in my view, are a monetary policy that fails to accomplish its economic objectives because of financial instability, and difficulties that stem from the inevitably imperfect coordination that takes place between multiple agencies with divided responsibilities during a crisis.  In those cases, not only will the central bank own that bad outcome with the resulting negative consequences for its ongoing credibility and effectiveness, but also the economic consequences will be severe for its citizens.

To address concerns about central bank power, the best approach is to ensure accountability.  This means appropriate oversight by the legislature and ongoing transparency by the central bank about the motivations for its actions, what it can reasonably expect to accomplish, and an honest assessment of its successes and failures.  

Thank you very much for your kind attention.   

¹ Krishna Guha, Marc Saidenberg and others on my staff contributed to these remarks

² As always, what I have to say here today reflects my own views and not necessarily those of the FOMC or the Federal Reserve System.

³ See “Asset Bubbles and the Implications for Central Bank Policy”, Economic Club of New York, April 7, 2010.   

⁴ Ideally, as was the case in the U.S. in 2009, the authorities should require banks to meet their stress scenario capital ratio standards by adding equity capital, with no credit given for reducing assets. This leans against the incentive to deleverage to meet higher capital ratio requirements. 

⁵ See my recent speech “Lessons at the Zero Bound: The Japanese and U.S. Experience”, Japan Society, May 21, 2013.