It’s not hard to identify many challenges for monetary policy in the wake of the crisis. I’ve chosen to focus on two for my presentation—the role that financial stability concerns should play in monetary policy making and the provision of forward guidance to markets about future movements in policy interest rates. Although they may appear to be two different and unrelated topics, by the time I’m finished I hope to convince you that they interact in some important ways, which central banks should take into account as they adjust monetary policy regimes to what we have learned in the build up to and recovery from the crisis.
Financial stability and monetary policy. In the years leading up to the crisis, I was in the camp of those skeptical that monetary policy should deviate from its focus on medium term price stability and high employment to counter a potential build up in financial risks or asset bubbles. I doubted the ability of central bankers to spot bubbles and imbalances in a timely way and of the efficacy of a relatively modest adjustment in policy interest rates to counter any build up in risks; and I was concerned about the costs in economic activity and inflation undershoots of a major policy adjustment to correct or forestall a possible increase in financial vulnerability. I still do not believe that low interest rates were a major contributor to the housing bubble in the United States or to the marked deterioration in credit standards, build up in leverage, and increased reliance on short-term funding in the financial sector that made the housing correction so painful and so prolonged. But they may well have played some role, and certainly given the costs of the crisis and slow recovery, it’s worth thinking about under what circumstances monetary policy might in the future give more explicit weight to financial stability concerns.
Moreover, people have legitimate worries about future risks to financial stability that might be building right now as a consequence of a prolonged period of very low interest rates. Unlike the period leading up to the crisis of a few years ago, we do not see increases in leverage and maturity mismatches in the financial sector, or rapid growth of credit to households and businesses. In fact, among lenders, leverage has come down, reliance on short-term funding has been reduced considerably, and credit to private nonfinancial sectors is growing very slowly. But very low bond yields, compressed credit spreads, and easing nonprice credit terms in a number of markets have raised questions about whether market participants are adequately allowing for tail risks—for surprises in the path of rates or the ability of borrowers to service their debts. Policy rates are likely to be low for some more time, given inflation tracking below target in many jurisdictions, output and employment still weak relative to estimated potential and downward revisions to potential GDP growth. Investors may not have made the required adjustments to their expectations for returns and be extending out risk curves in search of higher yields. Concerns about emerging asset mispricing have led to resistance by some central bankers to further ease when inflation has run below target and to calls by some policy makers and market observers for earlier rate hikes than might be justified by medium term outlooks for inflation and employment relative to objectives.
But there are costs as well as benefits from “leaning” with monetary policy against potential threats to financial stability that should be taken into account when judging whether a firming in policy to counter stability risks is in the public interest. In particular, leaning against such risks is likely to lead to undershooting of medium-term targets for inflation and output. We can think of the policy objective as trying to minimize the present value of deviations of inflation from its objective and output from full employment. Raising rates to deal with financial stability entails a tradeoff between a high probability of missing the targets at least a little in the near to medium term for reducing the low probability of a major or bigger shortfall later. That judgment entails lots of difficult elements, including the size of the rate adjustment required to be effective and its short-run costs.
It has been argued by Jeremy Stein among others that monetary policy has an advantage as a financial stability tool because it “gets in all the cracks”—it affects markets broadly. It can counter risk build ups without requiring the authorities to identify and take effective regulatory action when some of those risks may be occurring in markets or institutions for which the authorities do not have the regulatory tools or oversight roles. But the broad effect of monetary policy is a minus as well as a plus. Increases in interest rates will damp borrowing and spending in sectors not subject to overshooting in prices or credit. For example, raising interest rates to damp housing inflation and borrowing would also tend to reduce exports, business investment and other interest sensitive sectors that are not experiencing problems.
So I still support using regulatory policy first wherever possible to counter emerging financial stability risks. Importantly and constructively in the wake of the crisis, there has been a renewed interest in macroprudential regulation aimed at the externalities of risk build up. Such policies include structural measures to build safer systems less vulnerable to shocks, and countercyclical measures to bolster resilience and make credit more expensive in good times with buffers that can be drawn down when the cycle turns.
But we need to be aware of the current limits of our knowledge of the effectiveness of macroprudential policy. It is untested in the past few decades in highly developed financial centers of globally integrated markets; international cooperation and coordination will be required. The calibration of macroprudential tools to address risks is untested in these markets as are their effects on the macro economy that might have to be compensated for by monetary policy. And some countries have done a much better job than others at constructing institutional structures for coherently and transparently coordinating and explaining the interactions of monetary and macroprudential policies. I’m afraid the U.S. has still some ways to go in this regard.
Given the uncertain effects and limits on the applicability of macroprudential polices, monetary policy may need to be used to address financial stability risks, but the burden of proof on steering away from medium-term inflation and output targets for monetary policy should be reasonably high.
Forward guidance for interest rates. Offering considerable guidance to markets about the determinants of the likely path of its policy interest rate in the future is a potentially useful tool for the monetary policy authority. It is perhaps even a necessary tool when the policy interest rate is already at or very close to zero. In those circumstances, at the zero lower bound, market misunderstanding or misinterpretation of the central bank’s intentions might be especially costly in that the policy makers don’t have the option of lowering their target to change or offset the effects of market expectations of a more rapid increase in the policy rate than they consider appropriate.
At the ZLB, then, markets should have a deep understanding both of the strategy of the central bank—how it intends to pursue its targets—and of the central bank’s evaluation of the outlook–since the implementation of the strategy will depend on forecasts. Market participants, naturally, tend to extrapolate from past patterns of behavior, so forward guidance becomes particularly helpful when the central bank sees the economy as unlikely to follow past patterns—for example a slow rather than rapid rebound from a deep recession—or the central bank itself is not intending to follow its past reactions to incoming data or projections. That latter might be the case to compensate for time spent at the ZLB that has prevented desired easing or because the central bank wants to speed up the recovery from a deep recession. Note that neither of these imperatives—more information about strategy and forecasts—dictates the publication of a specific expected path for short-term interest rates.
It’s important for central banks and market participants to recognize the limits of forward guidance. In particular, central banks and others should not have a great deal of confidence in the ability of the central bank to predict the path of economic activity or inflation. Unanticipated developments and external events—for example problems in export markets or changes in fiscal policy or supply constraints for energy or other commodities– can throw the forecast off. More fundamentally, the experience of the past few years has underlined just how limited is our understanding of the most basic relationships in our economies. Those relationships include the effects of policy interest rates and other central bank instruments on financial conditions; the effects of financial conditions on demand and economic activity; the effects of activity on employment; and the effects of employment and activity on inflation.
The central bank can’t be more certain about future policy than knowledge allows. In its communication about future policy it should emphasize these limits and uncertainty. It can commit to a broad strategy dependent on progress towards its objectives; it can try to spell out how it will gauge such progress and how it would react to signposts along the way; but only rarely can it or should it commit to a particular path of rates. Markets should be encouraged to react to new information on the economy, inflation, and financial markets based on their knowledge of the central bank’s reaction function. And the central bank needs to maintain its flexibility to react to new information. For the most part, a central bank can commit to a strategy that to be sure may differ from its usual strategy, but not to a particular path of the policy interest rate.
Forward guidance in the time dimension, spelling out a path for the policy rate, can be troublesome for policy and for financial stability.
Forward guidance and financial stability. Much of the discussion of the intersection of monetary policy and financial stability has focused on the level of rates—whether low rates have resulted in excessive risk-taking through a “reach for yield” channel or through inducing greater leverage and maturity transformation. I want to focus on another dimension at that intersection—not the level of rates but the potential for rates to vary from their anticipated path. Perceived certainty around the path of rates—that the central bank won’t be deflected from an announced path except in extreme circumstances—itself can contribute to an unwarranted buildup of risk positons. Certainty can encourage leverage and maturity transformation—for example if participants do not see risk in the short-term rate leg of the carry trade. Certainty about future policy can contribute as well to damped volatility over all, and that volatility could then spike if changing circumstances induce the central bank to deviate from the expected path, with potential consequences for financial stability. And unwarranted certainty about short-term rates will damp market reaction to incoming news when those reactions could be stabilizing for the economic system.
My regrets about the monetary policy we followed at the Federal Reserve in the 2004-06 period leading up to the crisis center less around the level of rates—I see them as at most a little too low with 20-20 hindsight—and more about the promise of raising them “at a measured pace.” That promise was made for good reason–to avoid the markets building in more rapid tightening than we thought would be appropriate to the particular economic circumstances, which we thought was likely as markets anticipated a repeat of patterns of past tightening episodes. But that phrase came to mean a fairly automatic and predictable 25 basis point tightening each FOMC meeting and in so doing acted as a constraint on the FOMC, and gave market participants scope to engage in risky behavior based on certainty about the path for the federal funds rate.
Central banks today in many economies emerging from prolonged slumps following the great recession face a similar situation. They are concerned that the bond market may build in too steep a rise in interest rates too soon to support the desired path of returning to high employment and inflation rates rising to targeted levels. But we now see the potential cost of the buildup of financial fragilities if a path is specified and it is too predictable and not adjusted to new information.
So in my view, central banks should steer clear of commitments to explicit time paths for their target interest rate. They should explain their strategies for achieving their objectives, the economic variables they will be looking at to gauge progress toward those objectives and as best they can the values of those variables to which they will be paying particular attention, recognizing that the relationships among these variables is poorly understood and can shift over time. And they should emphasize uncertainty about the likely path of rates—the date of first tightening, the subsequent trajectory, and the ultimate value of the longer-term steady state rate. In her recent speeches and testimonies, Chair Yellen has indeed been highlighting both the “data dependency” of policy actions and the uncertainty around Federal Reserve forecasts. This is a constructive approach to forward guidance and should help reduce the financial stability risks that can be associated with a prolonged period of very low interest rates and intensified guidance about future actions.