Nonlinear exchange‐rate pass‐through in emerging markets

• Published date: 01 December 2019
• Date: 01 December 2019
• DOI: http://doi.org/10.1111/infi.12344

DOI: 10.1111/infi.12344

ORIGINAL ARTICLE

Nonlinear exchange-rate pass-through in emerging

markets

Francesca G. Caselli

|

Agustin Roitman

International Monetary Fund,

Washington, District of Columbia

Correspondence

Francesca G. Caselli, International

Monetary Fund, 700 19th Street, N.W.,

Washington, DC 20431, USA.

Email: fcaselli@imf.org

Abstract

This paper documents nonlinearities and asymmetries in the

transmission of exchange rate fluctuations to prices in a

panel of 27 emerging markets for the period 1990–2013.

Using local projection techniques, we find evidence of

asymmetry in the extent of exchange-rate pass-through

during episodes of appreciation relative to episodes of

depreciation. We then test for the presence of a threshold

that can lead to nonlinearities when the exchange rate

experiences a large depreciation. The findings suggest that

the pass-through becomes nonlinear when the exchange rate

depreciates by more than 24%. We also find that, in line with

the literature, the adoption of inflation targeting reduces the

degree of exchange-rate pass-through. These results are

robust to various econometric specifications that account

for the presence of outliers and potential endogeneity

concerns.

1

|

INTRODUCTION

Exchange-rate pass-through is usually assumed in the empirical literature to be both linear and

symmetric, but large exchange rate movements, such as those witnessed in recent years by emerging

markets, underline the importance of nonlinearities and asymmetries. The aim of this paper is to

estimate exchange-rate pass-through to domestic prices in emerging markets by focusing on three

specific dimensions: nonlinearities, asymmetries, and the monetary-policy regime. More specifically,

we test the following hypotheses: is exchange-rate pass-through symmetric during episodes of

appreciation and depreciation? Do nonlinearities exist when the exchange rate experiences a

particularly sharp depreciation? What is the effect of the adoption of inflation targeting on the

magnitude of exchange-rate pass-through?

International Finance. 2018;1–28. wileyonlinelibrary.com/journal/infi © 2019 John Wiley & Sons Ltd

|

1

279

International Finance. 2019;22:279–306. wileyonlinelibrary.com/journal/infi © 2019 John Wiley & Sons Ltd

The taper talk in the summer of 2013 triggered changes in the direction of capital flows, generating

sharp depreciations in emerging markets. The oil price decline coupled with geopolitical uncertainty in

2014 and, more recently, the US elections, had an impact on exchange rates in several emerging

markets. The degree of exchange-rate pass-through during such episodes of unusually large exchange-

rate movements has important implications for the conduct of monetary policy and directly affects the

expenditure-switching mechanism. Moreover, emerging markets have been changing their monetary

frameworks and/or exchange-rate regimes during the past two decades or have transitioned toward

inflation targeting. Countries operating under currency boards, or with currencies pegged to the US

dollar, have decided to introduce more flexibility and allow the exchange rate to fluctuate. To capture

the effect of such sharp exchange-rate adjustments or the transition between monetary-policy regimes,

a richer empirical framework than the standard linear exchange-rate pass-through model is needed.

This paper contributes to the literature by estimating state-dependent impulse responses for the

exchange-rate pass-through to consumer prices in a panel of 27 emerging countries over the period

1990–2013, using monthly data and local projection techniques. To test for the presence of

nonlinearities and asymmetries, we augment a standard linear model with interaction terms to capture

episodes of appreciation and large depreciation, and investigate whether inflation targeters have a

different degree of pass-through. We adopt a sequential approach. First, we study whether the

exchange-rate pass-through is asymmetric and find that inflation reacts differently if the exchange rate

appreciates or depreciates. Second, we analyze whether the response of prices during large

depreciations is different from that during average depreciations. We empirically test for the presence

of a threshold above which large depreciations can affect consumer prices more than proportionally.

The results suggest that the exchange-rate pass-through becomes nonlinear when the exchange rate

depreciates by more than 24%. We cannot argue for the presence of a unique threshold—rather, we

claim that a range of reasonable thresholds exists and that such nonlinearities cannot be ignored.

Moreover, we investigate the role of the inflation environment in the transmission of exchange-rate

shocks to prices and confirm that the adoption of inflation targeting reduces the degree of pass-through.

Our results are robust to various economic specifications and, while we acknowledge that it is very

difficult to claim causality in this context, we discuss the consequences of the potential endogeneity

that might characterize our estimation. To our knowledge, this is the first paper that systematically

analyses nonlinearities and asymmetries in the exchange-rate pass-through by exploiting local

projection techniques; the latter allow more flexibility than do standard vector autoregression (VAR)

models in investigating heterogeneous patterns in a panel setting.

Most empirical studies assume the exchange-rate pass-through to be both linear and symmetric.

However, many factors can generate directional and size asymmetries, and few papers have

comprehensively tested for the presence of nonlinearities and asymmetries in the transmission of

exchange-rate movements to prices. The paper that is most similar to ours is Bussière (2013), which

tests systematically for the presence of nonlinearities and asymmetries in the G7 economies, both with

panel and country-by-country estimation, and augments a linear model with a polynomial function of

the exchange rate and interacted dummies. It focuses on export and import prices and finds evidence of

both mechanisms, even though, in terms of magnitude, there is high heterogeneity across countries. We

build on that work by: (i) focusing on a large panel of emerging markets, where exchange-rate pass-

through is typically higher and where large exchange-rate movements can have a more significant

impact on prices;

1

(ii) using local projections, which allows us to characterize the dynamic and state-

dependent responses of prices to exchange-rate movements, exploiting the heterogeneity of our panel;

and (iii) instead of relying on polynomials, explicitly testing for the presence of a threshold via the

Hansen (1999) methodology (which aims to test for the presence of nonlinearities, comparing different

models with alternative thresholds). Frankel, Parsley, and Wei (2012) focus on developing countries

2

|

CASELLI AND ROITMAN

a

i

r

r

h

f

i

i

a

F

m

i

T

R

a

I

d

e

t

a

e

p

t

(

s

t

t

p

l

d

s

c

h

t

i

t

r

c

e

t

T

i

p

(

U

t

C

2

|

CASELLI AND ROITMAN

C

280