CURRENCY RISK PREMIUM AND U.S. MACROECONOMIC ANNOUNCEMENT

• Author: Ou Hu,Ding Du,Xiaobing Zhao
• Date: 01 December 2016
• DOI: http://doi.org/10.1111/jfir.12111
• Published date: 01 December 2016

CURRENCY RISK PREMIUM AND U.S. MACROECONOMIC

ANNOUNCEMENT

Ding Du

Northern Arizona University

Ou Hu

Youngstown State University

Xiaobing Zhao

Northern Arizona University

Abstract

In this article, we test whether the currency risk premium in the U.S. equity market is

particularly higher on prescheduled U.S. macroeconomic announcement days. Our

empirical analyses find supporting evidence. Our results help strengthen recent

conditional tests on currency risk and suggest that the currency risk premium in the U.S.

equity market is driven by U.S. macroeconomic conditions (e.g., U.S. monetary policy).

macroeconomic conditions (e.g., U.S. monetary policy).

JEL Classification: F31, G15

I. Introduction

With deviations from purchasing power parity (e.g., Rogoff 1996), currency risk can

affect both multinationals and domestic firms through its effects on sales, costs, or

competition. Therefore, a conventional economic wisdom (e.g., Adler and Dumas 1984)

is: U.S. stocks should be exposed to currency fluctuations, and U.S. equity investors

should be compensated for bearing currency risk. Early unconditional tests (e.g., Jorion

1990, 1991) find that U.S. stocks are generally not sensitive to currency fluctuations and

that there is no significant relation between currency risk and mean excess returns

across U.S. stocks. More recent conditional tests (e.g., Francis, Hasan, and Hunter 2008)

suggest that currency risk is indeed priced in the U.S. equity market and that the currency

risk premium depends on U.S. macroeconomic conditions (e.g., U.S. monetary policy)

(see also Bodnar and Gentry 1993; Bartov and Bodnar 1994; Chow, Lee, and Solt 1997;

Allayannis and Ihrig 2001; Dominguez and Tesar 2001; Williamson 2001; Bodnar,

Dumas, and Marston 2002; Kolari, Moorman, and Sorescu 2008; Bartram, Brown, and

Minton 2010; Balvers and Klein 2014; Du and Hu 2014).

A different recent strand of research focuses on prescheduled macroeconomic

announcement days to test conditional asset-pricing models. Intuitively, the trade-off

Part of this research was conducted while Ding Du was visiting the Robert H. Smith School of Business,

University of Maryland at College Park.

The Journal of Financial Research Vol. XXXIX, No. 4 Pages 359–388 Winter 2016

359

© 2016 The Southern Finance Association and the Southwestern Finance Association

RAWLS COLLEGE OF BUSINESS, TEXAS TECH UNIVERSITY

PUBLISHED FOR THE SOUTHERN AND SOUTHWESTERN

FINANCE ASSOCIATIONS BY WILEY-BLACKWELL PUBLISHING

between state variable risk and asset returns underlying standard asset-pricing theories

(i.e., risk premiums) should be higher on prescheduled macroeconomic announcement

days because important information about the state of the economy is revealed at such

times. Empirically, Savor and Wilson (2013, 2014) find that the market risk premium in

the domestic capital asset-pricing model (CAPM) is higher on U.S. macroeconomic

announcement days; Du and Hu (2015) document parallel evidence for the world market

risk premium in the international CAPM.

In the present article, we extend Francis, Hasan, and Hunter (2008), Savor and

Wilson (2013, 2014), and Du and Hu (2015) by examining whether the currency risk

premium is higher on prescheduled U.S. macroeconomic announcement days. If the

currency risk premium is causally driven by U.S. macroeconomic conditions as Francis,

Hasan, and Hunter imply, it should be higher on prescheduled U.S. macroeconomic

announcement days because important information about the state of the U.S. economy

is revealed at such times.

To empirically test our conjecture, we construct a currency factor-mimicking

portfolio and apply the Savor and Wilson (2014) methodology to daily U.S. equity

returns. To control for the effects of market risk emphasized by Savor and Wilson, we

focus on a two-factor model that includes both marketand currency risk factors. We find

that the currency risk premium is statistically and economically higher on U.S. macro-

economic announcement days even with the presence of the market factor, implying that

the currency risk premium in the U.S. equity market is driven by the state of the U.S.

economy. Our results are robust in subsamples and to alternative sets of test assets as

well as to alternative specifications to construct the currency factor-mimicking portfolio.

Our results are also robust to errors-in-variables and possible misspecification biases.

The present article adds to the literature by bridging the currency risk literature

(e.g., Jorion 1990, 1991; Francis, Hasan, and Hunter 2008) with the macroeconomic

announcements literature (e.g., Savor and Wilson 2013, 2014; Du and Hu 2015). We

show that at times when investors expect to learn important information about the state of

the economy (i.e., on macroeconomic announcement days), they demand higher returns

to hold currency-sensitivity assets. Thus, currency risk matters, conditionally.

II. Data

Currency-Sensitivity Portfolios

If a risk factor is priced in the equity market, mean excess returns of assets should vary

systematically with the exposure to this factor. Therefore, a standard approach in

empirical asset pricing is to construct diversified portfolios based on the exposure to a

factor. To focus on the effects of currency risk on excess returns, we construct 25

diversified value-weighted currency-sensitivity portfolios. In the same spirit of Savor

and Wilson (2014), the currency sensitivity for a stock at the beginning of a month is

estimated with the prior one year of daily data. To control for the effects of the market

risk emphasized by Savor and Wilson, we use the following two-factor model:

ri;t¼aiþbi;mrm;tþbi;fx

rfx;tþei;t;ð1Þ

360 The Journal of Financial Research