Do Currency Markets Absorb News Quickly? Martin D. D. Evans∗ Richard K. Lyons 30 July 2004 Abstract This paper addresses whether macro news arrivals affect currency markets over time. The null from macro exchange-rate theory is that they do not: macro news is impounded in exchange rates instantaneously. We test this by examining the effects of news on subsequent trades by end-user participants (such as hedge funds, mutual funds, and non-financial corporations). News arrivals induce subsequent changes in trading in all of the major end-user segments. These induced changes remain significant for days. Induced trades also have persistent effects on prices. Currency markets are not responding to news instantaneously. Corresponding Author Richard K. Lyons Martin D.D. Evans Haas School of Business Department of Economics U.C. Berkeley Georgetown University Berkeley, CA 94720-1900 Washington, DC 20057 Tel: 510-642-1059, Fax: 510-643-1420 Tel: 202-687-1570, Fax: 202-687-6102 [email protected] [email protected] faculty.haas.berkeley.edu/lyons georgetown.edu/faculty/evansm1 ∗ Journal of International Money and Finance, forthcoming. Respective affiliations are Georgetown University and NBER, and UC Berkeley and NBER. We thank the following for helpful comments: Carol Osler, Michael Melvin, an anonymous referee, and seminar participants at the 8th International Conference on Macroeconomic Analysis and International Finance. Both authors thank the National Science Foundation for financial assistance, which includes funding for a web clearinghouse for micro-based FX research (at faculty.haas.berkeley.edu/lyons, and at george-town.edu/faculty/evansm1).1Do Currency Markets Absorb News Quickly? This paper addresses an important null hypothesis within exchange rate theory: that macro news is impounded in exchange rates instantaneously. As an empirical matter, this is typically understood to mean a matter of seconds, or perhaps minutes, but certainly contained within the day of news arrival. This is consistent with the views of currency dealers as well: roughly 70% of the dealers surveyed by Cheung and Chinn (2001) respond that the effects of announcements are absorbed by the market within 1 minute. We test this hypothesis by examin-ing the effects of news on subsequent currency trades by end-user participants (such as hedge funds, mutual funds, and non-financial corporations). We find that news arrivals induce subsequent changes in trading behavior in all of these major end-user segments. These induced changes in trading remain significant for days. Induced trades also have persistent effects on prices. These findings provide strong evidence that currency markets are not responding to news instantaneously. Our basis for pursuing whether induced trades might prolong the absorption of news comes from recent empirical work demonstrating a tight link between signed transaction volume (order flow) and signed exchange rate changes.1 This link is not predicted by macro exchange rate theory, but it is predicted by an alternative modeling framework from microstructure finance. In this micro-based framework, transactions play a central, causal role in price determination (see, e.g., Glosten and Milgrom 1985, Kyle 1985). The causal role arises because transactions convey information that is not common knowledge. In this paper, we address whether the tight link between price adjustment and order flow that exists in general might also be playing a role in how currency markets absorb news. Most of the existing literature linking exchange rates to news is event-study based, and does not address how transaction quantities respond to news (i.e., it addresses the link between news and price in isolation). This literature has two branches. The first addresses the direction of exchange-rate changes (first moments) and the second, later branch addresses exchange-rate volatility (second moments). A common finding of the first branch is that, at least at the daily 1 This evidence is from both micro (i.e., single marketmaker) and macro (marketwide) studies. On the micro side, see, e.g., Lyons (1995) and Bjonnes and Rime (2004). On the macro side, see, e.g., Evans and Lyons (2002a,b), Payne (2003), Osler (2004), and Bjonnes, Rime, and Solheim (2004). Order flow is the cumulation over time of signed trades, where trades are signed according to whether the initiating side is buying or selling. (The marketmaker posting the quote is the non-initiating side.)2frequency, directional effects from scheduled macro announcements are difficult to detect because they are swamped by other factors affecting price. Intraday event studies do find statistically significant effects, particularly for employment and money-supply announcements (Andersen et al. 2003).2 The second, later branch of this literature—which focuses on news effects on volatility—is partly a response to early difficulty in finding news effects on first moments.3 This work finds that arrival of scheduled announcements does indeed produce the largest exchange-rate changes. Nevertheless, the ability of these fundamentals to account for overall volatility changes is lower than that of less fundamental factors such as time-of-day effects and ARCH (Andersen and Bollerslev 1998). A more recent literature has emerged that addresses the currency market’s response to news as a joint quantity/price response (Carlson 2002, Danielsson et al. 2003, Evans and Lyons 2003). Carlson (2002), for example, takes a case-study approach and analyzes a single macro announcement arrival. He finds that market characteristics were affected for hours following the arrival. (For example, liquidity remained significantly below normal—and below its ex ante state—for about 2 hours.) The case-study approach leaves open the question of how systematic these prolonged market effects are, and whether they might extend beyond the day of news arrival. The work of Danielsson et al. (2003) and Evans and Lyons (2003) provides less intraday resolution than Carlson, but these papers do examine multiple news arrivals over periods of months, and thereby do provide a sense for whether quantity responses to news are systematic. (Both papers find, for example, that roughly half of the transmission of news to prices actually operates through induced order flows.) Nevertheless, limited sample sizes only months