Giving and receiving: Exploring the predictive causality between oil prices and exchange rates

• Author: Jose E. Gomez‐Gonzalez,Jorge Hirs‐Garzon,Jorge M. Uribe
• Date: 01 March 2020
• DOI: http://doi.org/10.1111/infi.12354
• Published date: 01 March 2020

International Finance. 2020;23:175–194. wileyonlinelibrary.com/journal/infi © 2019 John Wiley & Sons Ltd

|

175

Received: 7 May 2018

|

Accepted: 29 April 2019

DOI: 10.1111/infi.12354

ORIGINAL ARTICLE

Giving and receiving: Exploring the predictive

causality between oil prices and exchange rates

Jose E. Gomez‐Gonzalez

1

|

Jorge Hirs‐Garzon

2

|

Jorge M. Uribe

2

1

Escuela Internacional de Ciencias

Económicas y Administrativas,

Universidad de La Sabana, Chia,

Colombia

2

Department of Economics, Universidad

del Valle, Cali, Colombia

Correspondence

Jorge M. Uribe, Department of

Economics, Universidad del Valle, Cali

76001, Colombia.

Email: jorge.uribe@correounivalle.edu.co

Funding information

Spanish Ministry of Economy, Grant/

Award Numbers: ECO2015‐66314‐R,

ECO2016‐76203‐C2‐2‐P

Abstract

We study the dynamic connectedness and predictive

causality between oil prices and exchange rates. Our

sample includes six important oil‐producing and six net

importing countries. Our results show that for the first set

of countries, oil prices are net spillover receivers from

exchange rate markets. Similarly, there is evidence of

bidirectional Granger causality, which is detected for

longer time periods from these countries’exchange rates to

oil prices. In contrast, for the second set of countries, oil

prices are net spillover transmitters, and the causality is

stronger from oil prices to exchange rates, mainly in the

aftermath of the Global Financial Crisis. However, even for

this group of countries, there are long periods of time for

which exchange‐rate markets transmit spillovers to oil

markets. Overall, oil markets are net receivers of shocks

during most of the sample period, thus providing evidence

in favor of the oil‐financialization hypothesis.

KEYWORDS

emerging market economies, foreign exchange returns, oil price, time‐

varying causality

JEL CLASSIFICATION

C22; G01; G12

1

|

INTRODUCTION

This paper studies the relation between oil prices and exchange rates for 12 countries. For six

countries, extracting oil is an important economic activity, and the other six are net oil

importers. The sample includes exchange‐rate floaters with different levels of economic

maturity (i.e., emerging and developed economies). Our contributions to the literature are

threefold. First, we examine the dynamic multivariate relation between oil prices and exchange‐

rate returns and measure the connectedness in foreign‐exchange (FX) markets from a global

perspective. We compute the total and directional connectedness indicators using forecast error

variance decomposition with vector autoregressions following the method developed by

Diebold and Yilmaz (2012, 2014). Second, we study the bidirectional causality between oil

prices and exchange‐rate returns following the method proposed by Hurn, Phillips, and Shi

(2016). This method, which is based on rolling‐window estimations of traditional Granger

causality tests, allows for the endogenous detection of changes in causality over time. Thus, it

enables the linking of directional changes in predictive causality to specific global economic

conditions, such as the recent Global Financial Crisis (GFC). Third, our sample includes both

major oil‐producing and oil‐importing countries, thereby allowing for the identification of

different connectedness and causality relations between oil prices and the exchange rates for

these sets of countries. Our analysis focuses on the connectedness between FX markets and oil

and extends the results of previous studies emphasizing that focus only on exchange rate

spillovers (Baruník, Kočenda, & Vácha, 2017; Chuliá, Fernández, & Uribe, 2018).

Spillovers are frequently modeled using correlation or covariance models (see Engle, 2002

and Engle & Kroner, 1995). Within this traditional framework, conclusive statements regarding

the association between a specific variable and the entire system cannot be given. This issue can

be addressed using different measures of market association that appear in the systemic‐risk

literature (e.g., Acharya, Pedersen, Philippon, & Richardson, 2017; Adrian & Brunnermeier,

2016). However, these measures, in turn, present the drawback of being unidirectional (i.e.,

they quantify the association between individual variables and the system in one direction or

another but not simultaneously). In this context, Diebold and Yilmaz (2012, 2014) propose a

unified framework for measuring spillovers and dependencies. Their framework allows for the

computing of several spillover indexes, including pairwise and system‐wide, in a mutually

consistent way. In our approach, we use both of these spillover measures and dynamic Granger‐

type causality tests to analyze the predictive causation.

The main advantage of our empirical approach is that it allows us to consider the dynamic

interplay between oil prices and exchange rates without making ad hoc assumptions on the

direction of the causality. With our approach, we are able to measure the directional and net

spillovers and evaluate the direction of causality at each point in time.

The relation between oil prices and exchange rates has recently regained interest in

academic and policy circles. Part of this renewed interest has been motivated by the surge of

sharp movements in oil prices occurring since 2000, particularly in the aftermath of the GFC.

As the literature has shown, the volatility of oil and financial markets has significant effects on

macroeconomic stability, especially for countries that are producers and consumers of crude oil

and its derivatives. Notably, exchange‐rate and oil‐price volatility are closely linked in countries

in which oil production represents an important share of their total output. This has led to the

term “commodity currencies”that was introduced by Chen and Rogoff (2003). Understanding

the dynamic linkages between oil and exchange rates is therefore important both for investors

and policymakers.

While most studies assume exogenous oil prices and test for their effect on exchange rates, a

new strand of the literature has come to challenge this approach. Recent studies on commodity

markets point out that increasing participation of institutional investors, such as hedge funds

and pension funds, has led to a higher dependence of prices on investors’preferences and

behavior. Indeed, the exposure to a variety of instruments, including futures, options, and

176

|

GOMEZ‐GONZALEZ ET AL.