The dynamic effect of macroeconomic news on the euro/US dollar exchange rate

• Author: Walid Ben Omrane,Robert Welch,Xinyao Zhou
• Date: 01 January 2020
• Published date: 01 January 2020
• DOI: http://doi.org/10.1002/for.2615

SPECIAL ISSUE ARTICLE

The dynamic effect of macroeconomic news on the euro/US

dollar exchange rate

Walid Ben Omrane

1

| Robert Welch

1

| Xinyao Zhou

2

1

Goodman School of Business, Brock

University, St. Catharines, Ontario,

Canada

2

Schulich School of Business, York

University, Toronto, Ontario, Canada

Correspondence

Walid Ben Omrane, Goodman School of

Business, Brock University, 1812 Sir Isaac

Brock Way, St. Catharines, ON L2S 3A1.

Canada.

Email: wbenomrane@brocku.ca

Abstract

This study investigates the impact of 70 US and EU macroeconomic news

announcements on euro/dollar returns and volatility from November 2004 to

April 2014. We use regime smooth transition regression to endogenously define

recession and expansion. Our sample period includes the US mortgage crisis

and EU sovereign debt crisis. Most news is unstable as its effect varies between

these economic states. There are asymmetrical effects between recession and

expansion states for both US and EU news, with most US news having a larger

impact and nearly double the number of significant EU announcements. Vol-

atility increases for over 85% of news coefficients, with more than half still

being significantly different between states.

KEYWORDS

financial crisis, foreign exchange rate, high‐frequency data, macroeconomic news

1|INTRODUCTION

This study investigates unstable US and EU macro news

effects between recession and expansion states within

2004–2014 which includes both the US and European cri-

ses and focuses on several questions. Can domestic and

foreign macroeconomic news explain foreign exchange

return and volatility? Are some categories of macroeco-

nomic news announcements more prominent than

others? Finally, are currency return and volatility

response to macro news influenced by different economic

recessions such as US mortgage and EU sovereign debt

crises? These questions are pursued via the endogenous

estimation of both crisis periods, followed by the return

and volatility response to macro news over these expan-

sions and recessions.

We find that a specific news announcement can have

asymmetric effects on FX return and volatility in these

expansion and recession periods. In other words, the same

announcement is context specific as its effect is contingent

on the corresponding economic state. This variability in

exchange rate response raises significant concerns over

currency exposure, which corporations involved in inter-

national trade have to manage effectively.

The main contribution of this paper relates to the

unstable macroeconomic news effect between specific

recession and expansion states around US mortgage and

EU sovereign debt crises. Specifically, these differences

affect the euro/US dollar returns and volatility in magni-

tude and sign—in other words, differences in degree and

kind. Our results reveal the context‐specific nature of

news shocks over time. These effects are based on the

standardized news surprise component.

We examine the effect of US and EU macro news

announcements around US and EU crisis subperiods,

each of which includes an expansion and a recession

state. US news has nearly double the number of EU sig-

nificant news effects (coefficients) on return and volatil-

ity, and larger return magnitudes (absolute value) of

these surprise effects. Over 94% (70%) of US (EU) news

have at least one state with a significant return coefficient

and 12 (9) US news switches signs between (within) sub-

periods. Volatility increases for over 85% of US and EU

significant news announcements and, consequently, only

Received: 8 August 2018 Revised: 8 February 2019 Accepted: 26 May 2019

DOI: 10.1002/for.2615

Journal of Forecasting. 2020;39:84–103.wileyonlinelibrary.com/journal/for© 2019 John Wiley & Sons, Ltd.84

4 (2) US news switch signs between (within) subperiods.

However, half the news still has significantly different

magnitudes between states within a subperiod.

In general, there is a striking contrast of these effects

as they vary between and within subperiods with respect

to their states as endogenously defined by the US and EU

crises. US context‐specific examples include New Home

Sales appreciating the US dollar in both expansions (no

effect in recessions), while Construction Spending appre-

ciates (depreciates) the US dollar in the first (second)

period expansion. Retail Sales only affects the first period

where it appreciates (depreciates) the US dollar in expan-

sion (recession), while Minutes of the FOMC Meeting has

the opposite effect only in the second period. One inter-

esting EU sign reversal is ZEW Survey Current Situation,

which appreciates (depreciates) the US dollar in the first

period expansion (second period recession). Although less

common, even volatility exhibits this asymmetric behav-

ior; for example, Personal Spending decreases volatility

in both recessions but increases it in the second period

expansion, while Business Inventories decreases volatility

in the first period recession but increases it in the second

period expansion.

We also find asymmetrical aggregate news effects. For

example, US news tends to appreciate the US dollar in the

first period expansion and second period recession. EU

news tends to appreciate the euro in US and EU crisis

periods. Most US news volatility reductions occur in the

first period recession, while for EU news it is in the sec-

ond period expansion. It is clear that news announce-

ments have a profound and dynamic effect on euro/US

dollar returns and volatility between states both within

and between subperiods.

Overall, our findings suggest that investors and man-

agers of multinational corporations are exposed to this

state‐dependent FX risk. Traders and institutional asset

managers whose portfolios include international assets

should also consider the time‐varying impact of macro-

economic news on exchange rate return and volatility

when designing strategies to improve the risk manage-

ment and return performance associated with their inter-

national transactions.

We distinguish our study from prior investigations in

the following ways. First, we examine the EUR/dollar for-

eign exchange return and volatility response to extended

categories of domestic, foreign, and unscheduled macro-

economic news. Second, we consider a more recent and

longer time period including both US mortgage and EU

sovereign debt crises, which are endogenously estimated

using a smooth transition regression model. Third, we

examine the sensitivity of return and volatility response

to US and EU macro news over the estimated recession

and expansion periods.

The reminder of the paper is organized as follows. Sec-

tion 2 presents a literature review. Section 3 explains the

methodology, and the Appendix provides technical

details about the embedded models. Section 4 describes

the data. Section 5 presents the empirical results, and Sec-

tion 6 concludes.

2|LITERATURE REVIEW

Previous research investigates the stable effects of macro-

economic news on foreign exchange (FX) rates (Ander-

sen, Bollerslev, Diebold, & Vega, 2003; Bauwens, Ben

Omrane, & Giot, 2005), but more recently the focus has

been on unstable effects for both FX returns (Ben

Omrane & Savaser, 2016; Fratzscher, 2009; Melvin & Tay-

lor, 2009; Swanson & Williams, 2014) and volatility (Ben

Omrane & Savaser, 2017; Laakkonen & Lanne, 2009,

2013; Pearce & Solakoglu, 2007).

1

Meese and Rogoff

(1983) show that the major macroeconomic models

linking the exchange rate to macroeconomic fundamen-

tals are not able to compete with a random walk, partic-

ularly for shorter time horizons. Subsequently, two lines

of research try to explain this puzzle.

One line of research changes the time frame from

“macro”to “micro”.

2

The seminal work of Evans and

Lyons (2002) and Lyons (2001) provides a “hybrid model”

that incorporates macroeconomic fundamentals and a

microstructure factor–order flow which can explain 60%

of daily changes from their 5‐minute frequency Deutsche

Mark/dollar data set. Evans and Lyons (2008) found that

scheduled macro news could affect exchange rates

directly by less than 10% of total variance, while the indi-

rect impact was about 20% for daily frequencies. Ander-

sen et al. (2003) and Andersen, Bollerslev, Diebold, and

Vega (2007) provided evidence of instantaneous impact

of news on 5‐minute exchange rates and adjusted for

intraday seasonality pattern effects on volatility. One

advantage of high‐frequency data is the reduction of

other confounding news effects, as argued by Dominguez

and Panthaki (2006, p. 170).

Applying this microstructure approach, Bauwens et al.

(2005) found that news impact on euro–dollar volatility

1

Neely (2011) and Neely and Dey (2010) provide a useful survey of

announcement effects on returns and volatility.

2

The other line of research remains focused on macroeconomic models.

Sarno and Taylor (2002) believed that a change of policy regime and the

agents' heterogeneity of beliefs on related macro information would

impact model specification. Supporting evidence was provided by

Cheung and Chinn (2001), who found that US Forex traders frequently

changed their weights on specific macroeconomic indicators. Rossi

(2006) provided robust evidence on parameter instability. Bacchetta

and Wincoop (2004, 2006, 2011) found that parameter instability may

be caused by expectations and “scapegoat”effects.

BEN OMRANE ET AL.85