Currency Carry Trades: The Role of Macroeconomic News and Futures Market Speculation

• DOI: http://doi.org/10.1002/fut.21778
• Published date: 01 November 2016
• Date: 01 November 2016

Currency Carry Trades: The Role of

Macroeconomic News and Futures

Market Speculation

Suk-Joong Kim*

This paper investigates carry trade opportunities in major currencies against the US dollar over

the period January 2, 1999 to December 31, 2012. There is evidence of significant Australian

dollar (AUD), Euro, and Japanese yen (JPY) carry trades during noncrisis periods. The AUD

(JPY) was an investment (a funding) currency, and the Euro was both. However, cross currency

carry trades were not present. For the AUD and JPY, carry trades were more likely with low

volatility and volume. Macroeconomic news that appreciate (depreciate) the AUD (the JPY)

also stimulated the AUD (the JPY) carry trades. However, there is no evidence of meaningful

and consistent impact of macroeconomic news on the EUR carry trades. For weekly horizon

investigations, net long futures positions in the AUD promoted carry trades in the AUD and

JPY. However, net long positions in the JPY only managed to reduce JPY carry trade probability.

© 2016 Wiley Periodicals, Inc. Jrl Fut Mark 36:1076–1107, 2016

1. INTRODUCTION

One of the building blocks of the theory of internationa l finance is interest rate parity in

both covered and uncovered forms. The interest rate parity conditions suggest the

equilibrium relationsh ips between the movements of the exchange rate between two

national currencies and t heir interest rates over a sp ecified holding period. The c urrency

with a higher nominal interest rate is expected to depreciate against the other currency by

roughly the same amount as the inte rest rate differential in the a bsence of transaction

costs. However, empirical evidence thus far suggests that the interest rate parity is an

unreliable measure of futur e exchange rate movements. No t only is the magnitude of

exchange rate movement sug gested by an interest rat e differential incorrect, even the

actual direction is often op posite to what is predicted . Indeed, higher yielding cur rencies

tended to appreciate over the re levant holding periods rat her than to depreciate (Burnsid e

et al., 2007). The most likely ex planation is a combination of a higher real interest rate a nd

Suk-Joong Kim is Associate Professor and Chair of Discipline at Discipline of Finance, The University of

Sydney Business School, The University of Sydney, Sydney, NSW, Australia. I wish to thank an anonymous

referee and the editor (Robert Webb) for invaluable comments and suggestions that improved the quality of

this paper. All remaining errors are my own. I also wish to acknowledge a research grant from the University of

Sydney Business School.

JEL Classification: E44, F31, G15

*Correspondence author, The University of Sydney Business School, The University of Sydney, Sydney, 2006 NSW,

Australia. Tel: þ61-2-9114-0940, Fax: þ61-2-9351-6461, e-mail: sukjoong.kim@sydney.edu.au

Received September 2015; Accepted December 2015

The Journal of Futures Markets, Vol. 36, No. 11, 1076–1107 (2016)

© 2016 Wiley Periodicals, Inc.

Published online 22 February 2016 in Wiley Online Library (wileyonlinelibrary.com).

DOI: 10.1002/fut.21778

a risk premium in the high yieldin g currency. For instance, ex cept for only a handful of

occasions, real interest rat es in the Australian Dollar (AUD) have been considerably hi gher

than the corresponding US rates in recent times.

1

Higher risk premia attached to emerging

market interest rates and cu rrencies (most notably the R ussian Ruble, the Brazilia n Real,

and the Indian Rupee to name a few) were some of the factors that encouragedcarry trades

into the emerging market curre ncies until the abrupt revers als in the mid-2013 sparked by

the talks of a reversal of the US Quantit ative easing.

Persistent interest rate differentials not offset by corresponding exchange rate

movements present an opportunity for currency investors willing to take unhedged

speculative posit ions—borrowing in l ow yielding currenc ies (funding currenc ies, e.g., the

Japanese Yen (JPY), the Euro, and the USD) and investing in higher yielding currencies

(investment currencies, e.g., most emerging market currencies and commodity currencies

such as the AUD and the New Zealand Dollar). This strategy, known as the currency carry

trade, has emerged as an alternative asset class in a portfolio of investments (e.g., Das,

Kadapakkam, & Tse, 2 013; Lustig, Roussa nov, & Verdelhan, 201 4). However, this is i n

direct violation of the uncovered interest rate parity (UIP). For instance, the AUD has

been one of the more impo rtant investment c urrencies due to the per sistent positive

interest rate differentials against the USD and other major currencies, especially against

the JPY, since the early 2000s. The risk takers speculate that ex post exchange rate

changes would not completely offset ex ante risk premia thought to be included in higher

yielding currencies, leading to sufficient rewards for taking the risk. For instance, Coudert

and Mignon (2013) report that sovereign default risk, as measured by sovereign Credit

Default Swap spre ads, contributes t o carry trade profits during boom periods. On the

other hand, there is a significant downside risk to this approach. The downside risk or

crash risk of carry trade strategies relates to unexpectedly high levels of exchange rate

volatility leading to a high likelihood of significant losses from the expo sed positions. Th e

literature reports evidence of profitable carry trade strategies, however. Burnside,

Eichenbaum, and Rebelo (2008) report significant gains from carry trades where carry

trades diversified over a number of currencies improve the typical Sharpe ratio of hedge

funds by more than 50% . Colavecchio (2 008) finds significant carry trade opportunities

involving the JPY. M ollick and Assefa (2 013) report eviden ce in support of carry t rades

involving the JPY an dS wiss Frac. Jurek (2014) reports profitable carry trades involving the

G10 currencies where a one-third of the excessreturnswasduetothecrashriskpremia.

Kim (2015) reports that the probabilities of carry trades into the AUD depend not only on

the usual measures of market uncertainty but also on the AUD order flows and Australian

macroeconomic news.

In addition, a number of researchers report spillover effects from carry trade activities.

Cheung, Cheung, and He (2012) find the JPY carry trades have varying degrees of impact on

the stock market returns in Australia, Canada, UK, Mexico, and New Zealand. Similarly,

Fung, Tse, and Zhao (2013) find that casality runs from carry trade activities to Asian stock

markets. Lee and Chang (2013) report similar results for the USD carry trades. They find that

the G10 currency carry trade returns Granger cause stock market returns but not vice versa,

and that the impact is larger during bull markets. In contrast, Fong (2013) reports optimism

in the stock market seems to lead to currency carry trades by hedge funds. Tse and Zhao

(2012) find significant correlation but no causality between the US stock returns and carry

trades, however, the stock market volatility Granger causes carry trades.

1

World Bank reports in their World Development Indicators that the real interest rate in Australia has always been

higher than the US since 1990. The only exceptions are 1998, 2000, 2007, and 2009. For example, the Australian

real interest was 6.49% compared to only 1.75% in the US in 2013.

Currency Carry Trades 1077

This paper investigates the existence and the time varying nature of carry trade

relationships among major currencies, the Euro (EUR), the JPY, the GBP, and the AUD,

against the USD and against each other at daily and weekly horizons over the period

January 2, 1999 to December 31, 2012. This extends Kim (2015) where only the AUD is

considered in the analysis.

Secondly, the potential determinants of the carry trades involving the AUD, EUR, and

JPY against the USD over daily horizon are investigated. In particular, underlying market

(financial markets in general and foreign exchange market in particular) conditions and

macroeconomic news that may foster (or trigger an abrupt unwinding of) the carry trades are

examined.

Thirdly, the role of specul ative position taking in curr ency futures markets is

investigated as a determinant of the carry trade activities over weekly horiz on for the AUD

and JPY.

The important findings of this paper are summarized as follows. First, there is strong

evidence of carry trade activities involving the USD (against the AUD, EUR, and JPY at daily

horizon, and the AUD and JPY at weekly horizon). However, there is no such relationship for

the cross exchange rates at either horizon. The evidence suggests that carry trades reversed

during various episodes of financial market turmoil such as the period surrounding the 9/11

event in 2001, the Global Financial Crisis (GFC) period of 2008–9 and the Eurozone crisis

period of 2010–11. Carry trade probabilities for the Euro and JPY are less straight forward to

interpret, however. On the other hand, the AUD has been more consistently in a carry trade

regime than the other currencies.

Second, for daily horizon, the carry trade probabilities for the AUD and JPY are

significantly higher in the periods of low exchange rate volatility and trade volume. In

addition, higher AUD carry trade probabilities are observed when there are 1) unexpectedly

low Australian inflation and unemployment rates, unexpected interest rate hike from the

RBA and 2) positive AUD order flows. The JPY carry trade probability is higher when there

are 1) unexpectedly low machine orders and Tanken figures in Japan, and unexpectedly high

retail sale growth and the Fed policy rate news in the US and 2) when there are negative JPY

order flows. For the EUR carry trades, there is no consistent evidence of the carry trade

probability being influenced by market activity variables or macroeconomic news.

Third, for weekly horizon, carry trades probabilities can be explained by net speculative

position taking by futures market traders. Net long positions in the AUD by noncommercial

traders significantly increased not only the AUD but also JPY carry trade probabilities. On the

other hand, net long positions in the JPY only influenced the JPY carry trade probability. They

significantly lowered the JPY carry trades.

The outcome of the investigation provides significant insights into the time varying

patterns of the carry trade probabilities of major currencies and the reaction functions

of carry traders. Knowledge gained from this investigation is useful for various types of

participants in the foreign exchange market. These include fund managers who would benefit

from knowing the day to day probabilities of crash risks in order to implement an appropriate

currency hedge. Furthermore, this information would be an important factor (i.e., as a signal

of change in the systemic risk) in the respective central banks’policy formulation to achieve

system stability. The rest of this paper is organized as follows. Section 2 presents the

discussions on the modeling strategies and the investigation results of carry trades using a

Markov regime shifting approach. Section 3 presents the empirical methodologies and

estimation results of the determinants of the AUD, EUR, and JPY carry trade probabilities at

daily horizon. Section 4 provides some robustness estimation results including weekly

horizon estimations, further subsample analyses and aggregated news investigations.

Conclusions are offered in Section 5.

1078 Kim