在具有突然惊喜的有效下限学习货币政策策略（英）

Learning monetary policy strategiesat the effective lower bound withsudden surprises by Spencer Krane, Leonardo Melosi and Matthias Rottner Monetary and Economic Department May 2026 JEL classification: E52, C63, E31 Keywords:newframework,centralbank'scommunications, deflationary bias, inflation surprises,asymmetricaverage inflation targeting,imperfectcredibility, liftoff, Bayesian learning BISWorking Papers are written by members of the Monetary and EconomicDepartment of the Bank for International Settlements, and from time to time by othereconomists, and are published by the Bank. The papers are on subjects of topicalinterest and are technical in character. The views expressed in this publication arethose of the authors and do not necessarily reflect the views of the BIS or its membercentral banks. This publication is available on the BIS website (www.bis.org). Learning Monetary Policy Strategies at theEffective Lower Bound with Sudden Surprises∗ Spencer KraneFRB ChicagoLeonardo MelosiEuropean University Institute& CEPRMatthias RottnerBIS & DeutscheBundesbank May, 2026 Abstract We examine how private sector agents might learn a new monetary strategy intro-duced while policy rates are at their effective lower bound (ELB) in an environmentwith large inflationary and deflationary shocks. We consider the adoption of a newasymmetric average inflation targeting rule aimed at countering the disinflationarybias imparted by the ELB. The most crucial time for learning runs from when rateswould be near liftoff under the old strategy through early liftoff under the new rule.Recessionary shocks during this time could delay learning while large inflationaryshocks could outright stop it, inhibiting the ability of the new strategy to address thecosts associated with the ELB. Using the US post-Covid experience as an example,we also find that the monetary policy shocks can have important feedback on thelearning process. Keywords: New framework, central bank’s communications, deflationary bias, in-flation surprises, asymmetric average inflation targeting, imperfect credibility, liftoff,Bayesian learning.JEL classification: E52, C63, E31 1Introduction The past 20 years have featured major changes to the monetary policy environment. Withthe onset of the Great Recession and ensuing slow economic recovery, the effective lowerbound (ELB) on policy rates went from being an isolated curiosity to a prominent featureof the monetary policy environment. Equilibrium policy rates fell below where they werein the 1980s and 1990s in many advanced economies, making the ELB a threat with even amodest recession. Meanwhile, inflation remained stubbornly subdued for an entire decade. In response, central banks adopted new strategies aimed at offsetting the risks posedby the ELB. Notably, ELB risks were stated as an important motivating factor in theFederal Reserve’s revised statement on Longer-Run Goals and Monetary Policy Strategyit adopted in August 2020 and the ECB’s Monetary Policy Strategy Statement issuedin July 2021.For both central banks, these changes were made when policy rates werestill at the ELB during the Covid-19 pandemic. These economies recovered quite swiftlyfollowing the pandemic, but then a mix of large shocks suddenly caused inflation to reachlevels unseen in the past 40 years. In this paper, we investigate how private sector agents might learn such new monetarystrategies introduced at the ELB in a stochastic environment in which large inflation-ary and deflationary shocks may occur unexpectedly.The rule we consider is a form ofasymmetric average inflation targeting. The exercise highlights several important consid-erations in implementing such strategies emanating from interactions between the learningprocess and economic outcomes.First, the period around when the liftoff of rates fromthe ELB would have occurred under the old rule is key for learning the new rule. Second,shocks that either return the economy to the ELB or that cause a surge in inflation thatquickly satisfies the inflation averaging goal can curtail agents learning the new rule andprevent the economy from capitalizing on its long-run benefits.Third, monetary policyshocks– perhaps due to discretionary actions by the central bank–come with a set of costsand benefits due to their interaction with the learning process. The theoretical foundations for strategies aimed at countering the ELB center on re-search showing that under traditional monetary strategies the limits on reducing policyrates presented by the proximity of the ELB imparts a downward bias to inflation andinflation expectations relative to the central bank’s target.The expectational bias is afeature of the distribution of inflation over time; it’s not just an occurrence associatedwith a large negative shock to the economy.In turn, this bias impinges on the central bank’s ability to stabilize both output and inflation.1 This literature has proposed a range of alternative monetar