IMPACT OF MACROECONOMIC ANNOUNCEMENTS ON INTEREST RATE FUTURES: HIGH‐FREQUENCY EVIDENCE FROM AUSTRALIA
Published date | 01 September 2013 |
Date | 01 September 2013 |
Author | Lee A. Smales |
DOI | http://doi.org/10.1111/j.1475-6803.2013.12015.x |
IMPACT OF MACROECONOMIC ANNOUNCEMENTS ON INTEREST RATE
FUTURES: HIGH‐FREQUENCY EVIDENCE FROM AUSTRALIA
Lee A. Smales
Curtin University, University of New South Wales
Abstract
I investigate the behavior of Australian interest rate futures around the release of major
scheduled macroeconomic announcements. The adjustment to new information occurs
quickly with the majority of the reaction complete within 30 seconds. The period
immediately before the announcement exhibits high volatility, low levels of volume, and
wide bid–ask spreads. In the 30 seconds following the scheduled announcement there is a
sharp increase in price volatility, significant positive correlation in returns, high levels of
trading activity, and large adjusted returns. The reaction is stronger in shorter maturity
contracts, and in the period surrounding the 2007–2008 financial crisis.
JEL Classification: G10, G14, G15
I. Introduction
I explore the price discovery dynamics of Australian interest rate futures in the period
around major scheduled macroeconomic announcements. High‐frequency data analysis
has become the norm in seeking to explain the economics of such price movements.
Indeed, Fleming and Remolona (1997) suggest that daily data are not sufficient to capture
the market reaction cleanly, or for capturing the microstructure aspects of the price
adjustment mechanism.
The identification of what constitutes a “major”economic release has developed
a literature of its own, and there is evidence of some instability in the market effects of
various macroeconomic announcements. Fleming and Remolona (1997) find that the
behavior of the fixed‐income market has a flavor‐of‐the‐month aspect in which different
announcements are regarded as important in different periods. For instance, studies from
the late 1970s to the mid‐1980s document a significant impact of money supply
announcements, while more recent studies, such as Cook and Korn (1991) and Krueger
(1996), solidify the importance of the nonfarm payrolls number in the monthly
employment report. Ederington and Lee (1993) note seven announcements, including
employment, as having a significant effect on U.S. equity markets.
I am very grateful to my referee, Christopher J. Neely, for his insightful feedback during the review process. I
gratefully appreciate the comments from the associate editor, Louis H. Ederington, which improved the scope and
conciseness of the article. I also thank seminar participants at the University of New South Wales and the 2012
Financial Markets and Corporate Governance conference for useful comments and suggestions. All remaining errors
are mine.
The Journal of Financial Research Vol. XXXVI, No. 3 Pages 371–388 Fall 2013
371
© 2013 The Southern Finance Association and the Southwestern Finance Association
Market efficiency would suggest that the expected portion of an announcement
should have no effect on asset prices, and recent studies (Anderson et al. 2003; Evans and
Lyons 2005) implicitly incorporate this assumption by focusing solely on the surprise
component of macroeconomic announcements. Kuttner (2001) and Fatum and Scholnick
(2008) find that asset returns and volatilities respond primarily to the surprise
component—measured by the difference between the actual announced figure and
the expected figure taken as the median of survey data—in macroeconomic data
announcements. Chen, Jiang, and Wang (2013) note that announcement surprises have a
significant impact on the trading activities and returns of U.S. equity futures.
If markets are efficient, news should be quickly impounded into market prices.
However, if the price adjusts slowly to the new information, it may be possible to earn
excess returns based on the initial market reaction to the news release, thus violating semi‐
strong‐form efficiency. Several studies (e.g., Urich and Wachtel 1981; Kim and
Sheen 2001; Han and Ozocak 2002) find support for at least semi‐strong‐form market
efficiency in the sense that prices react swiftly to news, and profit opportunities disappear
quickly as a result. Ederington and Lee (1993, 1995) find that efficiency holds in the
Treasury bond and foreign exchange markets, and Becker, Finnerty, and Kopecky (1996)
investigate the Treasury bond market and, based on conditional expectations, find
significant price adjustments, but only for Consumer Price Index, nonfarm payrolls, and
the trade balance.
The nature of an efficient market limits the length of the predictable return
response to news but does not place such restrictions on the impact of news on volatility. It
is therefore possible that the volatility response differs from the return response, and as a
result, another branch of analysis has focused on the effect of macroeconomic
announcements on asset price volatility. The general finding is that news releases result in
a rapid increase in return volatility; although the majority of the effects are relatively
short‐lived and subside within the first minute, there are varying degrees of volatility
persistence for periods ranging from 15 minutes to several hours—this is possible even if
the market is efficient. This may be explained by an increase in trading activity as market
participants rebalance their portfolios in light of the news, information traders enter the
market, and liquidity traders benefit from the elevated interest. Fleming and Remolona
(1997) reason that sharp price moves, and the accompanying increase in volatility, are
attributable to changes in expectations shared by investors, and surges in trading activity
are due to a lack of consensus on prices. Ederington and Lee (1993) suggest that volatility
may remain high for some period after an announcement because full information only
arrives gradually owing to the length of some reports and the analysis required in forming
a conclusion on the likely effect of the data release.
In addition, there is mounting evidence that asset prices respond differently to
macroeconomic news depending on market conditions as well as the general state of the
economy, with an impact on both returns and volatility. McQueen and Roley (1993) were
the first to note this effect in the equity market, and Roley and Sellon (1995) find the
relation between news and long‐term rates is likely to vary over the business cycle as
market participants alter their views on the persistence of policy actions. Cook and Korn
(1991) hypothesize that this is due to policy anticipation; the market reaction depends on
how participants expect the central bank to move monetary policy in response to the news,
372 The Journal of Financial Research
To continue reading
Request your trial