Toward a Flexible Average-Inflation Targeting Strategy

I was fortunate to have been at the Fed’s research conference at the Federal Reserve Bank of Chicago last week, part of its Review of Monetary Policy Strategy, Tools, and Communications. We were treated to an excellent program that Vice Chair Clarida organized. To me, one paper stood out above all others, the paper by Lars Svensson that set out a flexible AIT strategy.1In my view, a sufficiently flexible AIT-based makeup strategy is a leading candidate for an alternative to the prevailing flexible inflation targeting strategy.2I believe  Svensson’s flexible AIT strategy includes the ingredients of a revised strategy that could come out of the  Review. 

Go Directly to Page 24 

My presentation skips over the first 23 pages of Svensson’s paper. There he discusses his preference for his forecast targeting approach over simple policy rules as a guide for policymakers’ decisions. He also reviews price-level targeting (PLT) and temporary price-level targeting. We have heard many discussions of those strategies already. 

I start with his discussion of average inflation targeting on page 24. While we have also already read many discussions of AIT, Svensson’s presentation is notable for at least two reasons. First, he provides an excellent discussion of the technical aspects of applying AIT in practice. Second, and most importantly, he talks about the added flexibility, specifically the discretion, that can be applied in a flexible AIT strategy.  

The First Question 

Svensson is responding to the first of three questions that the Chair and Vice-Chair have told us to motivate the  Review: Can the Federal Reserve best meet its statutory objectives with its existing monetary policy strategy,  or should it consider strategies that aim to reverse past misses of the inflation objective? 

Strategies that seek to reverse past misses are called makeup strategies. Svensson defines makeup strategies as commitments to make up part or all of the inflation undershoots or overshoots. The second question posed for the Review is about tools, how to implement the strategy. And the third is about communications. I will write about those in later commentaries as well. 

The Review Came in Response to the Zero Nominal Bound (ZNB) Problem3 

1 Lars E. O. Svensson, “Monetary Policy Strategies for the Federal Reserve,” prepared for the conference Monetary Policy  Strategy, Tools, and Communication Practices, Federal Reserve Bank of Chicago, June 4-5, 2019.  2 Lars and I have had an ongoing debate about whether the Fed’s “dual mandate” strategy is equivalent to “flexible inflation targeting” as practiced elsewhere. But FOMC participants increasingly refer to the FOMC’s strategy as flexible inflation targeting, so I do so here. 

3 Many refer to the effective lower bound (ELB) rather than the zero lower bound or zero nominal bound (ZNB) since, in principle, the effective lower bound may not be zero. However, I don’t expect the FOMC will consider negative policy rates as part of the Review, so I refer to the ZNB throughout. 

Some central banks do regular reviews, but the FOMC’s current review is motivated by the recent experience and  focuses on how the monetary policy framework should be adapted to the “new normal.”4 The key feature of the new normal from the perspective of monetary policymakers is the decline in the real neutral fund’s rate, a  decline that is expected to persist. That decline has reduced the “buffer” between the nominal neutral rate and the ZNB, limiting how much the fund’s rate can be cut in a downturn. When I was at the Board, the buffer was believed to be about 450 basis points. Now the buffer is closer to 250 basis points.  

Given that the FOMC has lowered the fund’s rate 400-500 basis points in past recessions, this decline in the size of the buffer suggests that the Fed will more frequently find itself constrained by the ZNB. When it was previously at the ZNB, the Committee used LSAPs and more aggressive forward guidance as new “tools” but did not have a well-defined strategy to guide them. The overriding question is how to adapt the monetary policy to be more effective at the ZNB. Of course, the search for such a strategy may well end with one that the  Committee would also want to apply in normal times. 

While LSAPs and aggressive forward guidance did allow the Committee to lower the unemployment rate to below the estimated NAIRU, after nearly eight years the Committee has not achieved its inflation objective on a sustained basis. So revisions to the strategy are about achieving the inflation objective, though still in the context of the Fed’s dual mandate. 

The Prevailing Strategy: Flexible Inflation Targeting 

The prevailing strategy of the FOMC and other central banks in developing economies, flexible inflation targeting (IT), calls for policy to push inflation back to the 2% objective if it deviates from it, while also promoting maximum (sustainable) employment. It is “flexible” because there is both an employment and an inflation objective. The FOMC calls its inflation objective “symmetric,” meaning that, when inflation deviates from its objective in either direction, policy should aim to bring it back to the objective. The FOMC would continue to use the word symmetric in any revised strategy, but under a flexible AIT strategy, it would have a much more policy-relevant meaning. 

Guidelines for the Review 

The Chair and Vice-Chair have set out two features of any revised strategy: It must be evolutionary, not revolutionary, and it should have a makeup feature. While the Chair and Vice-Chair have not shared with us which options fall into the revolutionary and evolutionary categories, but I don’t believe the FOMC would ever adopt a full-offset commitment strategy.5 Those include price-level targeting (PLT) and, perhaps, temporary price-level targeting.  

Average-Inflation Targeting 

Svensson appears to favor PLT in principle. Under PLT, a central bank sets a target for the price level that increases at the Committee’s target inflation rate. This yields an average inflation rate equal to the central bank’s inflation objective. It works well in models that assume full credibility. But PLT is a full-offset commitment strategy and, as I noted above, the FOMC would never accept such a revolutionary commitment strategy. Svensson in the end prefers AIT, as he sees it as more evolutionary and flexible. Svensson defines  AIT as a central bank that has “a target for average inflation over a longer period than a year.” 

The implementation of AIT calls for a “look-back” period. The look-back period, assumed by Svensson—for  “simplicity and concreteness”—to be five years, is the period over which the average inflation rate is calculated for the purpose of judging how well the FOMC’s inflation objective is being met. He notes that  AIT, like IT, lets bygones be bygones: “With a 5-year average-inflation target, 5-year inflation is a bygone  after 5 years.” That’s not true in PLT and is an important source of flexibility in AIT. The implicit policy rule    

4 The Bank of Canada does framework reviews every five years, for example. 

5It would have been constructive for the Chair and Vice-Chair to tell us this at the beginning of the review because it would have allowed participants and outsiders to be more creative in developing and studying strategies that would be acceptable to the Committee. 

in AIT is to respond to the deviation between average inflation over the five-year look-back period and the inflation objective and to use policy to bring the five-year inflation rate (which evolves over the five years) to the inflation objective.  

PLT, in contrast, is sometimes described as having an infinite look-back period. Consider the Great Recession.  The look-back period would begin when inflation fell below 2% and, until the price level returned to the path consistent with 2% average inflation, the FOMC would still be making up for that initial shortfall. PLT, to be sure, is intended to reduce the cumulative undershoot of the inflation objective, but the offset period could nevertheless be very long, the cumulative undershoot could be very large, and the return to the target price level could require inflation either well above 2% for some time or modestly above 2% for a longer time. That is not something to which a central bank would likely commit. The FOMC would not. 

Here is how Svensson describes the flexibility in his flexible AIT strategy (italics my emphasis): 

Average inflation targeting as described here is quite flexible. It allows for some weight on both the annual and the multi-year average inflation target…The choice of the appropriate weights on stabilizing average inflation, annual inflation, and employment—in order to best correspond to the Federal  Reserve’s “balanced approach” and “equal footing”…has been left open in this discussion. In models,  the choice of optimal weights can be quite complicated… 

An Escape Clause for Overshoots? 

Under flexible average inflation targeting, it is clear-cut that “Any undershooting of annual inflation for a  couple of years [s] would normally be followed by some overshooting, in order to stabilize average inflation  around the target.” When Svensson turns to overshoots, he notes that, likewise, “If annual inflation has overshot the target for a couple of years, symmetric average-inflation targeting requires that—all else being equal—annual inflation should be brought down to undershoot the target for a couple of years” (italics our emphasis). 

However, in his discussion of overshoots Svensson then goes on to express some discomfort with making up for all overshoots—discomfort he doesn’t express for undershoots. He says, for example, “there are  situations, such as negative supply shocks, when tightening may be rather undesirable and when it may be  justified to allow a permanent shift in the price level.” He provides a couple of real-world examples and  concludes, “it needs to be considered whether or not some explicit escape clause should be included in  average-inflation targeting.”  

While Svensson does not go this far, the logic of having an escape clause in the case of a negative supply  shock would seem to apply to an extent to any overshoot, whatever its source. In the case of an undershoot,  makeup calls for temporarily higher inflation along with a decline in the unemployment rate. In the case of overshoots, makeup calls for temporarily lower inflation along with an increase in the unemployment rate above the NAIRU, perhaps for a considerable period.  

As a result, the Committee might want the added flexibility to respond asymmetrically to overshoots and undershoots. In that case, “makeup” would apply to undershoots, but perhaps there would be only a return to the inflation objective in the case of overshoots, as is currently the case under flexible inflation targeting.  Of course, that would take us some distance away from achieving an average inflation rate at the Committee’s inflation objective. That’s the price of flexibility. 

Temporary or Permanent?  

There is at least one more question about a makeup strategy: Should it be implemented only in extraordinary circumstances—that is, at the ZNB—or in normal times as well? Svensson argues, as have others, that if AIT  is not implemented in normal times, it may not be well understood and credible when it is needed most, at the ZNB. Action trumps words! 

But there may be more widespread support among participants for an AIT strategy if it is only to be implemented at the ZNB. If only implemented at the zero bound, that would leave a sharply asymmetrical response to undershoots and overshoots, as I suggested might be appropriate in any case. But that would also make the flexible AIT strategy more evolutionary than revolutionary. In effect, it would be a more aggressive threshold-based forward guidance at the ZNB, with an (average) inflation threshold. But markets would know, in advance of reaching the ZNB, that the FOMC would pursue such a strategy.  

The Flat Phillips Curve and Make-up Strategies 

Lurking in the background of discussions of alternative strategies is the challenge associated with the very flat Phillips curve. That is, after all, why it has become so challenging for central banks around the world to meet their inflation objectives. If it’s hard to get to 2%, wouldn’t it be even harder to get inflation to, say,  2½% for a few years? And would a strategy that was based on doing so even be credible?  

Svensson believes that make-up strategies can mitigate the challenges associated with a flat Phillips curve:  “With inflation expectations entering the Phillips curve, higher inflation expectations would through this channel independently increase inflation. This may be extra advantageous if the Phillips curve is now flatter  and inflation is less responsive to slack in the labor market and other relevant markets.” That’s fair, but it’s not clear at all how much this would help address the problem of a flat Phillips curve in practice.  

Bottom Line 

Svensson made a really important contribution to the Review in his presentation at the June conference in  Chicago. Svensson’s flexible AIT strategy is flexible enough to incorporate, in effect, the FOMC’s dual mandate, its symmetric inflation objective, and its “balanced approach” strategy—all the key features in its statement of strategy—while adding, of course, a makeup property, which could be asymmetric. And the AIT  could be temporary, only at the ZNB, or implemented in normal times as well. That makes it, in my view, a  plausible candidate—indeed, perhaps a leading candidate—for a revised strategy coming out of the Review!  If not the end, at least a very good beginning. 

Like this article?

Share on linkedin
Share on Linkdin
Share on facebook
Share on Facebook
Share on twitter
Share on Twitter
Share on pinterest
Share on Pinterest

Our Newest Content

Would you like a preview of our newest, members-only commentaries? Simply click the button below to see some of our newest commentaries and request a free trial today.


Please fill out the form below for a trial of our services.