Editor’s Note: These comments on the paper “The Most Dangerous Idea in Federal Reserve History: Monetary Policy Doesn’t Matter” (PDF) by Christina D. Romer and David H. Romer were presented at the American Economic Association Annual Meeting on January 5, 2013.
This is a very interesting and thought-provoking paper. I recommend it highly, not only to all who are interested in Federal Reserve history, but also to those seeking an interesting perspective on recent monetary policy.
The Romers ascribe lack of forceful enough action in the 1930s and 1970s to excessive policymaker humility about the power of monetary policy to address the most important problem of the time. Another way to characterize the same thing is that the timidity of action reflected policymakers perceptions of the relative effect of policy on output and inflation—and hence on the balance of costs and benefits of policy actions. In the 1930s policymakers saw little benefit for lowering unemployment of further ease and large potential costs in terms of higher inflation. In 1970s they saw little effect on inflation from tighter policy but large potential costs in terms of lost output. We can see in retrospect that these assessments were incorrect and based on misunderstandings of underlying economics—of how much slack there was in the economy, the relationship of slack to inflation, the effect on slack of policy actions, and the importance of expectations in shaping the response of the economy to policy actions.
The Romers raise the question of whether recent monetary policy has also been influenced by insufficient confidence in the power of policy actions to deliver net benefits in terms of reducing unemployment. They are concerned that the Federal Reserve has not acted as forcefully as it could have or should have over the past few years because of misperceptions about the relative costs and benefits of actions. An important handicap in making such a judgment is that we are in uncharted waters—both in the behavior of the economy and the nature of the monetary policy response. We don’t have the luxury of 20-20 hindsight to determine what misunderstandings might have been unnecessarily holding back policies. We are in the middle of the situation now and are highly uncertain about what forces are holding back economic growth. That uncertainty is compounded by the use of new policy tools. We have little or no empirical basis for making judgments about the effects—costs or benefits—of large scale asset purchases, huge increases in the FR balance sheet, and increasingly detailed interest rate guidance.
My comments are framed around the following points: We need to be humble about what we know as economists under any circumstance—but especially now; but policymakers can’t let that humility freeze them into inaction, and we need to think about techniques to move forward in such circumstances; then we can examine how the Federal Reserve’s approach to policy has stacked up relative to these techniques for operating under uncertainty.
The last 10 years have highlighted just how little we know about how the economy works. Both policymakers and private agents should have been more humble in the years leading up to the crisis. The so-called “great moderation” led to complacency and overconfidence in both the private and public sectors. There was overconfidence about the self-correcting properties of private markets—how self-interest would limit financial and economic fluctuations. We overestimated the extent to which new financial instruments had led to the diversification of risk across markets and participants. We were overconfident about ability of monetary policy to steer the economy— to avoid major bouts of inflation or deflation and all but the mildest recessions and to counter financial collapses. Especially egregious was tendency to ignore longer-term financial and credit cycles and the intricacies and nonlinearities of the financial sector and its interactions with the real economy. We missed the significance of the buildup of leverage in both the financial and nonfinancial sectors, and their tendency to reinforce each other. And both the private and public sectors paid insufficient attention to tail risk in asset prices and its interaction with leverage and maturity transformation.
Our understanding and the models that embody it not only didn’t see the crisis coming, we didn’t see or fully understand the dynamics of the financial system and its interactions with the economy during the crisis, and we don’t understand why the recovery has been so slow. To gauge our lack of understanding on this last point, I looked at the projections of FOMC participants in June 2010—the last time I contributed to them. All the key relationships were off. We’ve seen much less growth than expected then, despite a more expansionary monetary policy; the unemployment rate is only a little higher than expected, even with the much lower growth, so that relationship was off as well; and inflation has turned out significantly higher than expected, even with less growth and a slightly higher unemployment rate.
A major complexity in understanding recent developments is the interaction of the shortfall in demand with needed structural shifts and realignments of demand and production and the recovery of balance sheets—the layering of a shorter-term business cycle on top of longer-term structural shifts. We have only limited guidance from history about how this is going to play out in context of the current US financial system and the widespread adjustments in the global economy.
And to this relatively mysterious economic structure, the Federal Reserve is applying unprecedented policy actions. Efforts to drive down longer-term interest rates with unconventional policy actions are natural extensions of the short-term rate policy adjustments of more normal times and should work through the same policy channels, but the new actions are naturally very hard to calibrate. What effects are purchases or interest rate guidance having on financial conditions? What effect are changes in financial conditions having on spending and inflation? To what extent does it matter whether long-term rates are reduced by a change in expectations because of guidance versus a change in the term premium because of purchases? Will there be costs, especially down the road on exit? There is very little history to use to judge the costs and benefits of these actions.
But uncertainty about the economy doesn’t necessarily have to imply caution or paralysis in policy. For example, one could simply adjust policy to make the desired outcome the most likely—even if the distribution of possible outcomes was relatively flat and not well understood. But as the Romers highlight, the natural human tendency is to hold back when uncertain about the system and the effects of policy action—about the balance of costs and benefits to bolder action. How can policymakers counter such an impulse? We can draw on past experience with policymaking for a few suggestions.
The first technique is to use “risk management”. Weigh the costs and benefits of missing to one side or the other of your inflation and output objectives; identify the most important problem relative to those objectives—the one that would reduce public welfare the most if policy were overly cautious; and take some chances to deal with that problem. This approach is especially attractive if the likelihood and cost of missing other objective is seen as relatively small. So, this would imply extra effort to deal with unemployment in the 1930s relative to inflation and with inflation in the 1970s relative to unemployment. Of late, high unemployment or low output would appear to embody a greater cost and risk to public welfare than rising inflation, given the evident slack in the economy and low inflation; of course, this could change in the future.
Second, learn and revise as you go. Policymakers need to be ready to update their understanding as they gain experience with the prevailing economic situation and the effects of policy, which should help gradually clarify the costs and benefits of their policy choices. This implies flexible policy implementation and a willingness to revise the policy setting and the plan going forward. In my view, a time of great uncertainty about underlying economic structures is not well suited to policy rules, especially rules based on recent economic data. The implications of incoming data for future economic performance—when changes in policy will have their effect– will need to be updated frequently, and lessons learned about the effects of policy settings on the economy. These circumstances would seem to embody the classic case for gradualism in policy implementation—for slowly altering the stance of policy as you learn about its effects. It is critical under these circumstances to make the framework for actions and revisions very clear to the public so agents and markets understand how the central bank is processing information and can anticipate its actions as well as possible, increasing the odds that spending and pricing decisions will reinforce central bank action.
How do we evaluate recent monetary policy experience in the context of heightened uncertainty and the possible techniques to deal with it? The Romers have three issues with the language of Federal Open Market Committee participants that they fear may be indicative that the Committee is not being as bold as it should be. First is the argument made by the FOMC that the benefits of added easing are limited and declining—and relatedly that monetary policy is “no panacea” for the problems of the economy. Second, they highlight the argument that there are potential costs to unconventional policies that may be significant and rising. And third, they cite the gradual evolution toward more aggressive policies as proof that the Federal Reserve should have been more aggressive to begin with.
In my view the FOMC has in fact followed the prescriptions for overcoming inertia and excessive caution. With respect to risk management it has clearly identified long-term unemployment as its major concern—so long as inflation is not likely to stray very far from its objective. Especially in the criteria it has put forth for considering when to begin tightening policy, it has shown a willingness to take risks on the inflation side to make progress on unemployment. Partly this reflects the level of unemployment and inflation relative to objectives, but part of the reasoning surely is that we know more about how to control inflation than how to boost employment with monetary policy, so if inflation overshoots persistently the FOMC can be confident it has the tools to bring it back down. Notably, the quotes the paper views favorably from former chairmen about taking bold actions are about controlling inflation—not about raising output.
I would judge that the actions of the FOMC evidence a learning process as well. Policy has evolved as the FOMC learned about the evolution of the economy—just how sluggish the recovery would be; and as it has learned about inflation—just how little inflation and inflation expectations have been affected by unconventional policies and increases in the Federal Reserve’s balance sheet, an oft stated worry of those opposed to bold action. Consequently, it has become more willing to use its unconventional tools over time, even when its forecast has only changed marginally, consistent with gradualism in the face of uncertainty. Moreover, it has increasingly spelled out its policy approach in public. Some of the added discussion has been tactical—using dates of possible first tightening to guide rate expectations down—but some has helped to elaborate the strategy better, including the January 2012 document about the framework for policy and the added information in December about the economic conditions that would guide its consideration of first tightening.
Moreover, I don’t think you can dismiss the possibility that as the operations scale up, the benefits of added unconventional measures are limited and declining, or that the costs are increasing, as reflected in the comments of FOMC officials. On the benefit side, the marginal effect of lower longer-term interest rates in bringing increasing amounts of consumption and investment from the future to the present may well decline the deeper into the pile of possible future projects you dig. And the income effect of lower rates damping the spending of savers may gather force relative to the substitution effect inducing greater saving the longer rates are expected to be quite low. On the cost side, the greater the purchases the higher the risk that market functioning may be impaired or that the eventual unwinding of the portfolio will have adverse consequences for financial stability. And the larger the portfolio, the more questions are raised about the fiscal risk the Federal Reserve is taking and about complications when it’s time to exit—to raise rates and roll back the portfolio.
Importantly, these concerns have not stopped the Federal Reserve from taking increasingly bold actions over time. But in this context, warnings from the Federal Reserve that monetary policy is “no panacea” for what ails the economy are understandable and justified. Can monetary policy restore full employment before the benefit and cost lines cross? And even if it can, wouldn’t the return to full employment be faster and more certain if other government policies were working in the same direction?
Finally, people point to the Volcker disinflationary episode as an example of how humility about what we know about the economy doesn’t have to imply timidity in action, and how forceful action can overcome pessimism about the effectiveness of policy. Monetary policy under Volcker’s leadership was decisive, bold, and effective at promoting the longer-run welfare of the US economy. For sure, Paul Volcker didn’t believe we knew much about how the economy and inflation worked—for example he largely ignored staff forecasts in light of their poor track record over the 1960s and 1970s. But he knew inflation had been very costly and could be reduced with monetary policy, with a one-time cost in output of unknown dimensions, but long-run gains. And he was willing to innovate in the conduct of policy in order to get that done.
It’s important to recall, however, that, his action followed years of failed efforts to tame inflation with policy gradualism and nonmonetary policies. Like the current Federal Reserve, he talked a lot about other governmental policies that he thought might make his efforts more difficult; he worried and spoke out forcefully on the risks he saw from the fiscal and current account deficits of that time; he’s not present to speak for himself, but I suspect he would be comfortable with the “no panacea” language the current Federal Reserve uses. And he was willing to back off in the fall of 1982 with inflation still above his objective but recession deepening and financial stability threatened. That is, he was willing to update his cost/benefit calculation as circumstances changed.
In the current situation the FOMC has said unemployment is very costly; it has signaled it is willingness to take risks on inflation to boost resource utilization; it has innovated in numerous ways; but it can’t and shouldn’t ignore its calculation of costs and benefits and adjust policy as needed. In sum, I’m not sure it’s justifiable to argue that the FOMC has missed its “Volcker moment” or has been less forceful than warranted in real time policymaking.