Expected inflation and the sacrifice ratio

• DOI: http://doi.org/10.1111/infi.12340
• Published date: 01 December 2019
• Date: 01 December 2019

DOI: 10.1111/infi.12340

ORIGINAL ARTICLE

Expected inflation and the sacrifice ratio

Joseph P. Daniels

1

|

Sandeep Mazumder

2

|

David D. VanHoose

3

1

Marquette University, Milwaukee,

Wisconsin

2

Wake Forest University, Winston-

Salem, North Carolina

3

Hankamer School of Business, Baylor

University, Waco, Texas

Correspondence

Sandeep Mazumder, Wake Forest

University, Box 7505, Winston-Salem,

NC 27109.

Email: mazumds@wfu.edu

Abstract

Using inflation forecasts from the OECD Economic

Outlook as proxy measures of inflation expectations, we

examine the impact of inflation expectations on the sacrifice

ratio for 20 OECD countries. The regression analysis

considers four different empirical models of the determi-

nants of the sacrifice ratio typically found in the existing

literature. The impact of the level of inflation expectations is

negative and significant, implying that a higher level of

expected inflation is associated with lower sacrifice ratios.

This result is consistent with the theoretical role of nominal

wage and price rigidities in that reductions in wage and price

stickiness diminish the tradeoffs between disinflations and

output losses. Interaction effects indicate that higher levels

of expected inflation allow policymakers to pursue ‘cold

turkey’inflation reductions even more aggressively. The

effect of the change in inflation expectations is negative and

significant, implying that faster adjusting inflation expec-

tations are associated with lower sacrifice ratios.

1

|

INTRODUCTION

There have been a number of studies on th e determinants of the sacrific e ratio or the tradeoff

between output and inflati on. The sacrifice ratio was fi rst quantified by Ball (1994a) to examine the

output loss associated with a disinflationary episode. Early studies focused on the relationship

between the sacrifice rati o and the speed of the disinflati on, to determine whether a gra dual

disinflation or ‘cold turkey ’disinflation resulted in a lo wer output loss. Subsequent studi es

considered economy-wide c haracteristics such as the de gree of trade openness, centra l bank

independence, wage stickin ess, the exchange rate regime or inflation targeting, and exchange-rate

pass-through, to name a few.

International Finance. 2019;1–16. wileyonlinelibrary.com/journal/infi © 2019 John Wiley & Sons Ltd

|

1

307

International Finance. 2019;22:307–322. wileyonlinelibrary.com/journal/infi © 2019 John Wiley & Sons Ltd

Theoretical models on (1) price/wage stickiness; (2) policymaking credibility/commitment; and (3)

macroeconomic variability introduced via policy actions suggest channels that link inflation

expectations, as well as the speed of adjustment of inflation expectations, to the inflation-output

trade-off. Limiting consideration to price and wage stickiness, we would a priori assume an estimated

negative relationship between expected inflation and the sacrifice ratio.

Once policy credibility effects are taken into account, however, the predicted sign of this

relationship becomes less obvious, as endogenously determined inflation expectations and the

credibility of macroeconomic policies can move together or in opposing directions. Empirical

treatment of the influence of policy credibility would, therefore, yield an ambiguous prediction for the

relationship between expected inflation and the sacrifice ratio.

In contrast, policy-induced macroeconomic volatility, itself depending on policy credibility,

impacts the optimal degree of wage stickiness that leads to output losses from disinflations. Taking this

channel, by itself, into consideration suggests a positive relationship between expected inflation and

the sacrifice ratio. Overall, these three channels operating together through the macroeconomic policy

process suggest a likely positive relationship between expected inflation and the sacrifice ratio. Given

some degree of ambiguity, however, the relationship remains an empirical question, which is the

motivation for this study.

Using inflation forecasts from the OECD Economic Outlook as proxy measures of inflation

expectations, we examine the impact of the level of inflation expectations and changes in inflation

expectations on the sacrifice ratio for 20 OECD countries. The regression analysis considers four

different empirical models of the determinants of the sacrifice ratio typically found in the existing

literature. The first sets of regressions compare the impact of the level of inflation expectations

prevailing at the beginning of a disinflationary episode to the effect of initial, or peak inflation at the

beginning of a disinflationary period. The direct impact of the level of inflation expectations—or the

total impact in models with interaction terms—is negative and significant in all models, implying that a

higher level of expected inflation is associated with a lower sacrifice ratio. This result is consistent with

the theoretical role of nominal price and wage rigidities, in that reductions in wage and price stickiness

diminish the tradeoffs between disinflations and output losses. Interaction effects indicate that higher

levels of expected inflation, consistent with reduced wage and price stickiness, allow policymakers to

pursue ‘cold turkey’inflation reductions even more aggressively.

The role of the level of inflation expectations, however, differ from initial inflation when

considering a model that includes a variable for central bank independence. We explore this difference

by considering mediation effects via the Sobel-Goodman mediation test (Sobel, 1982). We find that

initial inflation is a mediator—that is, it carries some of the influence of central bank independence to

the sacrifice ratio. This suggests that a more independent central bank has the power to exploit the

inflation-output tradeoff more when the level of inflation at the outset of a disinflation episode starts

from a high level. Results for the level of inflation expectations imply that inflation expectations are not

a mediator.

A second set of regression models examines the role of changes in inflation expectations that occur

over a disinflationary episode rather than the level of inflation expectations. Consistent with the results

for the level of expectations, the effect of the change in inflation expectations is negative and

significant in all four regression models, implying that faster-adjusting inflation expectations are

associated with lower sacrifice ratios. Likewise, the interaction between changes in inflation

expectations and the length of the disinflation is negative and significant, indicating that faster-

adjusting expectations allow policymakers to pursue disinflations even more quickly. In addition,

changes in inflation expectations also have a mediation effect in models that include a measure of

central-bank independence. Since the level of expected inflation does not have a mediation effect,

2

|

DANIELS ET AL.

w

c

i

t

i

a

s

b

e

i

2

S

s

T

D

2

|

DANIELS ET AL.

D

308