Countering Downward Bias in Inflation

GIC and Banco de México
February 27, 2020
Mexico City, Mexico

Charles L. Evans
President and CEO
Federal Reserve Bank of Chicago

The dual mandate bullseye, showing target of 2% inflation and 4.1% unemployment. We are near those targets, but slightly under both.

The views I express here are my own and do not necessarily reflect the views of the Federal Open Market Committee (FOMC) or within the Federal Reserve System.

Key Messages

Effective Lower Bound (ELB) risk leads to downward bias in inflation
When ELB drives down π < π* for an extended period, need to follow with some period of π > π* in order to establish E[π] consistent with symmetric target

Low Trend Growth and Low Neutral Interest Rates (r*)

US (percent)

Trend growth fell sharply from it's peak in 1999. However, for the last few years it has been rising. Estimated real r* has also fallen sharply since 1999, but has been more volatile since then. Currently it has ticked higher.

Advanced Economies (percent)

A line graph showing the rate of trend growth and the estimated real r* rate. Both fell sharply around 2009, but both have recently risen.

Estimates for Advanced Foreign Economies are GDP-weighted averages across the US, Canada, the Euro Area, and the UK using OECD estimates of GDP at purchasing power parity. Prior to 1995, Euro-Area weights are the summed weights of the eleven original euro area countries. Sources: Laubach and Williams (2003); Holston, Laubach, and Williams (2017); FRBNY

Undershooting Inflation Goals

Deviation from Central Bank Inflation Target

Source: Various statistical collection agencies from Haver Analytics

Conventional Monetary Policy Easing During Past Recessions

Federal Funds Rate (percent)

Average easing during recessions	500 bps
Current fed funds rate range	150-175 bps
Long-run neutral rate	250 bps

Line graph showing the Federal Funds Rate since 1980. It was at its highest in the early '80s, then was near zero for about 5 years. It has recently been lowered.

Source: Board of Governors of the Federal Reserve System from Haver Analytics

Fed Funds Rate and a Traditional Benchmark

Federal Funds Target Rate (percent)

Taylor Rule (1999):
r(t) = r*(t) + π(t) + 0.5[ π(t) – π* ] + 2[ u^LR(t) – u(t) ]

A line graph showing the Federal Funds Rate and teh Taylor Rule. They line up prettey well except for a period between 2009 and 2015, when the Federal Funds Rate was at 0 at the Taylor Rule dropped below 0.

r*(t) and u^LR(t) from Blue Chip Consensus Forecast.
Source: Board of Governors of the Federal Reserve System from Haver Analytics

Offsetting ELB Downward Inflation Bias

Heightened risk of ELB
- Downward bias in inflation
- Risk of E[π] < π*
To offset bias, likely need π > π* for some period of time so that:
- E[π] is firmly anchored at π*
- π = π* in the medium term
Embrace approaches aimed at these bias-adjusting outcomes

Outcome-Based Approaches

Overarching aim: achieve dual mandate goals
To do so, monetary policy must commit to:
- Provide extraordinary policy accommodation during and after ELB episodes
  - Prescriptions from simple rules (e.g., Taylor) are inadequate
- Generate periods of π > π* to offset ELB downward inflation bias
  - Recognize π > π* is required more than in non-ELB world
  - Convey to public that periods of π > π* essential to achieve dual mandate over long haul
  - The outcome of E[π] = π* is key
A number of ways to operationalize this

Example: State-Contingent Price Level Targeting

Core PCE Price Index

Source: Bureau of Economic Analysis from Haver Analytics and staff calculations

Example: Asymmetric Policy Response

Respond more aggressively when inflation below target than when inflation above target: Bianchi, Melosi, Rottner (2020)
Adjust the standard Taylor Rule
r(t) = r*(t) + π(t) + λ[ π(t) – π* ] + 2[ u^LR(t) – u(t) ]
- If π(t) < π*, larger λ
- If π(t) > π*, smaller λ

Evans’s view: Inflation objectives that have a point target, such as 2 percent, are easier to communicate than objectives defined by an inflation range. As I discuss next, using a range requires even more attention to asymmetry.

Example: Inflation Ranges [π^L < π < π^U* ]

Alternative #1: Harris (2016); Mertens and Williams (2019)
- Recognize that inflation will be driven to π^L when at ELB
- Aim for higher inflation π^U away from ELB to average π* over time.
Alternative #2: Bianchi, Melosi, and Rottner (2020)
- When inflation is in range, react less aggressively
- But set range asymmetrically about target
  - e.g., if π* = 2%, then π^L = 1.5%, π^U = 2.85%

Example: Inflation Ranges [π^L < π < π^U* ]

Alternative #3: Symmetric Range of Policy Indifference
- When inflation is in range, do nothing. Say we can go home—that’s good enough for government work

Example: Inflation Ranges [π^L < π < π^U* ]

Rejected

Alternative #3: Symmetric Range of Policy Indifference
- When inflation is in range, do nothing. Say we can go home—that’s good enough for government work

Won’t cure ELB downward inflation bias

Properties of Asymmetric Responses and Range Alternatives #1 & #2

Parameters can be set so that inflation will average π* over long periods of time
Do not require mechanical makeup for past periods of inflation away from target

Some Questions

Can policymakers credibly commit to pursuing the policies prescribed by some of these alternatives?
How will central banks communicate these strategies effectively?
How will the public react to protracted periods of π > π*?
- Will long-run inflation expectations move up? By how much?
What are the financial stability implications of the highly accommodative policies prescribed by the alternatives?

My Key Considerations

Focus on outcome-based strategies
- In the U.S., focus on the dual mandate
- When ELB drives down π < 2%, likely need follow with period of π > 2% to get inflation expectations consistent with target
Given ELB, any operational framework will need to use unconventional tools (e.g., QE, forward guidance)
- Effectiveness of these policies will influence the policy parameters of the alternative frameworks
Address potential financial stability risks with regulatory and supervisory tools
Credibility is key and essential for any operational framework

Countering Downward Bias in Inflation

Key Messages

Low Trend Growth and Low Neutral Interest Rates (r*)

US (percent)

Advanced Economies (percent)

Undershooting Inflation Goals

Deviation from Central Bank Inflation Target

Conventional Monetary Policy Easing During Past Recessions

Federal Funds Rate (percent)

Fed Funds Rate and a Traditional Benchmark

Federal Funds Target Rate (percent)

Offsetting ELB Downward Inflation Bias

Outcome-Based Approaches

Example: State-Contingent Price Level Targeting

Core PCE Price Index

Example: Asymmetric Policy Response

Example: Inflation Ranges [πL < π* < πU ]

Example: Inflation Ranges [πL < π* < πU ]

Example: Inflation Ranges [πL < π* < πU ]

Properties of Asymmetric Responses and Range Alternatives #1 & #2

Some Questions

My Key Considerations

Countering Downward Bias in Inflation—Slides

Example: Inflation Ranges [π^L < π < π^U* ]

Example: Inflation Ranges [π^L < π < π^U* ]

Example: Inflation Ranges [π^L < π < π^U* ]