Jumps and Trading Activity in Interest Rate Futures Markets: The Response to Macroeconomic Announcements

• Date: 01 October 2013
• Author: George H. K. Wang,Robert I. Webb,Johan Bjursell
• DOI: http://doi.org/10.1111/ajfs.12028
• Published date: 01 October 2013

Jumps and Trading Activity in Interest Rate

Futures Markets: The Response to

Macroeconomic Announcements*

Johan Bjursell

Barclays Capital

George H. K. Wang

School of Management, George Mason University

Robert I. Webb**

McIntire School of Commerce, University of Virginia

Received 30 January 2013; Accepted 21 May 2013

Abstract

We apply nonparametric statistical procedures to extract jumps around scheduled macroeco-

nomic news in U.S. Treasury bond, U.S. Treasury note and Eurodollar futures prices from

2001 to 2004. Volatility and trading activity during announcement days with jumps versus

no jumps are also analyzed with computerized trade reconstruction (CTR) and time and sales

high frequency data. Several interesting results are obtained. First, while jumps often occur

during announcement periods, many jumps cannot be associated with macroeconomic news

releases. Second, volatility and trading volume are higher during announcement days with

jumps than announcement days without jumps. Furthermore, volatility returns to the pre-

announcement level faster following scheduled news releases with jumps than after announce-

ments without jumps. Third, we find that price and trading volume are adjusting simulta-

neously in the first 1-minute interval following the announcement. Thus our results do not

confirm that there exists a two-stage adjustment process for prices and trading volume in

interest rate futures following scheduled public news releases.

*Acknowledgments: This article has benefited from the comments of the Editor, two anon-

ymous reviewers, and participants at seminars at the following universities, research centers

and conferences: National Chengchi University, University of South Australia, Deakin

University, 2011 Financial Management Association Conference, 2011 Korean Finance Asso-

ciation-Taiwan Finance Association Joint Conference (Mirae Asset Global Investments Best

Paper Award), Hong Kong Baptist University, 2010 Financial Management Association

Europe Conference, 2010 Asia Pacific Association of Derivatives Conference, Capital Markets

Cooperative Research Centre (Sydney, Australia).

**Corresponding author: Robert I. Webb, McIntire School of Commerce, University of Vir-

ginia, 125 Ruppel Lane, Charlottesville, Virginia 22903, United States. Tel: +1-434-924-7570,

Fax: +1-434-924-7074, email: riw4j@virginia.edu.

Asia-Pacific Journal of Financial Studies (2013) 42, 689–723 doi:10.1111/ajfs.12028

©2013 Korean Securities Association 689

Keywords Realized variation; Bipower variation; Jump statistics; Interest rate futures; Macro

announcements; Price and trading volume patterns

JEL Classification: C32,G13

1. Introduction

How financial markets incorporate scheduled macroeconomic news releases into

prices is a major topic for academic researchers, regulators and financial traders.

The price behavior and trading activity around macroeconomic announcements

offer unique opportunities for researchers and regulators to evaluate market price

efficiency and operational efficiency in trading mechanisms across financial markets.

Financial traders formulate their trading strategies based on their expectations about

the information content of forthcoming announcements. This paper applies recently

developed nonparametric statistical procedures to extract jumps around scheduled

macroeconomic news in U.S. Treasury-bond futures, U.S. Treasury-note futures and

Eurodollar futures prices during the sample period from 2001 to 2004. We also

examine volatility and trading activity on days with a jump at the announcement,

announcement days with no jumps, and days without announcements and jumps.

Ederington and Lee (1993) examine the impacts of scheduled macroeconomic

announcements on U.S. Treasury-bond futures, Eurodollar futures and Deutsche

mark futures from 1988 to 1991. They find that macroeconomic announcements are

often responsible for observed day-of-the-week and time-of-the-day patterns of the

volatility in these markets. They report that most of the price adjustments are com-

pleted in about 1 minute following the release of the news whi le the volatility remains

higher for about 15 minutes. They find that the unemployment report (i.e., Employ-

ment Situation Report), consumer price index (CPI), producer price index (PPI), and

durable goods orders news releases have significant impacts on these markets.

Webb (1994) examines the impact of the forecast errors from scheduled

monthly announcements of the unemployment, CPI, PPI and durable goods orders

reports on U.S. Treasury bond futures prices and volatility using time and sales data

from 1989 to 1990. He discusses several instances of seemingly anomalous behavior

of futures prices in response to the news in the reports.

Flemming and Remolona (1999) document that the arrival of public informa-

tion (e.g., the unemployment situation, CPI, and PPI reports) in the 5-year U.S.

Treasury note cash market sets off a two-stage adjustment process for prices and

trading volume. In the first stage, the price change rises sharply with the reduction

of trading volume at the time of release of major announcements. In the prolonged

second stage, trading volume surges, volatility persists and bid-ask spreads remain

wide. Their intraday data are obtained from the interdealer government bond mar-

ket and span the period August 23, 1993 to August 19, 1994.

Using intraday data from the interdealer government bond market for the period

1991 to 1995, Balduzi et al. (2001) exami ne the impacts of seventeen macroeconomic

news releases on prices, trading volume and bid-ask spreads in the U.S. Treasury

J. Bjursell et al.

690 ©2013 Korean Securities Association

market. They document a persistent increase in volatility and trading volume after

the announcements. Price adjustments to news generally occur within 1 minute after

the announcement.

1

However, none of these papers have identified jumps around

macro news releases in interest rate futures markets.

Recently, several papers have applied nonparametric statistical procedures to

extract jumps around news events. For example, Barndorff-Nielsen and Shephard

(2006) apply their bipower varia tion procedure to estimate jumps on 10 years of

U.S. Dollar/Deutsche Mark (USD/DEM) and U.S. Dollar/Japanese Yen (USD/JPY)

exchange rate data around macroeconomic announcements. They relate jump days

with macro news release days. Andersen et al. (2007) identify jumps in S&P 500

futures prices, U.S. Treasury bond futures prices and exchange rates. They find

some evidence that jumps are linked to news events and demonstrate that the jump

component is less persistent than the corresponding continuous component of the

realized volatility.

Jiang et al. (2011) identify jumps in the prices of the 2, 3, 5, 10 and 30 year U.S.

Treasury notes and bonds traded on the BrokerTec electronic trading platform during

2005–2006. Their results show that jumps mostly occur during scheduled macroeco-

nomic announcements. They demonstrate that liquidity shocks play an important role

in creating jumps in the U.S. Treasury markets. Lahaye et al. (2007) analyze how news

arrival causes jumps and co-jumps in stock index futures, U.S. Treasury bond futures,

exchange rates and gold.

2

Their results show that many announcements do not create

jumps and many jumps are not related to announcements. Dungey et al. (2009) use

Cantor-Fitzgerald tick data sampled from 2002 to 2006 to identify jumps and

co-jumps in the term structure of the U.S. Treasury market. They find that jumps are

often associated with macroeconomic news and “co-jumping” often occurs at sched-

uled macroeconomic news releases. They also report that not all news releases generate

jumps and news surprises do not always generate jumps. However, they do not ana-

lyze trading activities around announcements conditioned on jumps in returns. Kim

and Lee (2011) examine the effects of six Korean and US macroeconomic news

announcements on the risk–neutral distribution (RND) of KOSPI 200 prices using

intraday option data from January 2003 to September 2009. They find that the RND

responds to most of the macroeconomic news announcements, implied volatility

tends to increase and RND tends to become less (more) negatively skewed after good

(bad) news announcements. However, they do not investigate the intraday trading

activity of KOSPI200 index options around macroeconomic announcements versus

non-macroeconomic announcement periods.

1

Green (2004) examines the information content of order flows in the U.S. government bond

market during macro announcements. Simpson and Ramchander (2004) investigate the

impact of macroeconomic news on the spot and futures markets, but they do not consider

jumps due to the impact of macroeconomic news releases.

2

According to Lahaye et al. (2007), two markets are considered to have co-jumps if there is a

significant jump in both markets within 10 minutes following an announcement.

Interest Rate Futures Markets

©2013 Korean Securities Association 691