The impact of the US stock market opening on price discovery of government bond futures

• Published date: 01 July 2019
• Author: Ivan Indriawan,Feng Jiao,Yiuman Tse
• Date: 01 July 2019
• DOI: http://doi.org/10.1002/fut.22015

Received: 30 October 2018

|

Revised: 9 April 2019

|

Accepted: 9 April 2019

DOI: 10.1002/fut.22015

RESEARCH ARTICLE

The impact of the US stock market opening on price

discovery of government bond futures

Ivan Indriawan

1

|

Feng Jiao

2

|

Yiuman Tse

3

1

Finance Department, Auckland

University of Technology, Auckland,

New Zealand

2

Dhillon School of Business, University of

Lethbridge, Lethbridge, Alberta, Canada

3

Finance and Legal Studies Department,

University of Missouri‐St. Louis, St. Louis,

Missouri

Correspondence

Feng Jiao, Dhillon School of Business,

University of Lethbridge, 4401 University

Drive Lethbridge, AB, Canada T1K 3M4.

Email: feng.jiao@uleth.ca

Abstract

We examine price discovery in sequential markets for the 10‐year US Treasury

note, German bund, and UK gilt futures over the period 2010–2017. We find

that price discovery increases after the opening of the US stock market. Order

flows in the bond futures markets are more informative for permanent price

changes in the 30‐min period after the US stock market opens. A placebo test

using US statutory holidays confirms our findings. A cross‐market analysis

suggests that the increased price discovery in the bond futures is related to

returns and net order flows of the US stock market.

KEYWORDS

government bond futures, information share, price discovery, sequential markets

JEL CLASSIFICATION

C32, G14, G15

1

|

INTRODUCTION

Cross‐asset correlations are important for the optimal allocation of financial assets and risk management. Considering

the broad spectrum of theoretical and empirical questions related to the links between the stock and bond markets, it is

not surprising that a large number of studies have addressed distinct aspects of this issue. Studies on the stock–bond

return find substantial yet time‐varying returns comovement (Addona & Kind, 2006; Baele, Bekaert, & Inghelbrecht,

2010; Connolly, Stivers, & Sun, 2005; Connolly, Stivers, & Sun, 2007; Yang, Zhou, & Wang, 2009). Other studies reveal

that innovations in the stock and bond market volatility and liquidity are significantly correlated (Chordia, Sarkar, &

Subrahmanyam, 2005; Fleming, Kirby, & Ostdiek, 1998; Goyenko & Ukhov, 2009). Furthermore, Underwood (2009)

documents that aggregate order imbalances in the US stock market contain important information about the intraday

returns of the US Treasury market.

Our paper adds to this growing literature by focusing on the cross‐market links of price discovery between the stock

and government bond futures markets. In particular, this study attempts to provide evidence on how intraday variations

in the US stock market can affect price discovery in the international government bond futures markets. In the US stock

market, information is incorporated into prices rather differently at different times of the day. During regular trading

hours, the intraday price discovery in the stock market usually exhibits a U‐shaped pattern, reaching peaks during the

opening and closing minutes. The US stock prices are also more efficient and informative during trading hours than

after hours. The preopening and postclosing periods generate noticeable, but inefficient, price discovery (Barclay &

Hendershott, 2003). As price discovery fluctuates in the US stock market throughout the day, one would expect to find

that price discovery in the Treasury futures markets also varies. Given its size and interconnectedness, it is important to

investigate whether the US stock market is a vital determinant of price discovery in government bond markets.

J Futures Markets. 2019;39:779–802. wileyonlinelibrary.com/journal/fut © 2019 Wiley Periodicals, Inc.

|

779

To examine the role of the US stock market on price discovery in international government bond futures, we explore

three distinct but related questions. First, do the prices of US Treasury futures become more informative when the

regular trading session starts in the US stock market? Second, how does the opening of the US stock market affect price

discovery in international government bond futures? Third, what happens to the cross‐market information content in

the stock and bond futures markets after the US equity market opens?

The intuition behind the sources of the stock‐market‐opening effect on bond futures price discovery could be subtle.

In a perfectly efficient securities market with complete information, asset prices change instantaneously in response to

the release of new information. As the government bond futures are trading continuously round‐the‐clock, all public

information should already be reflected in the price of the bond futures even before the US stock market opens, hence

the stock‐market‐opening effect might be barely observable. There are good reasons, however, for believing that

the gradual release of private information could produce a market‐opening effect. In government bond markets, private

information originates in the fact that agents in different markets possess heterogeneous abilities to understand the

current state of economic fundamentals, interpret economy‐wide indicators, and estimate the effect of the release of

public information on security price dynamics (Brandt & Kavajecz, 2004; Jiang & Lo, 2014; Pasquariello & Vega, 2007).

The stock‐market‐opening effect can be attributed to private information in two ways: The cross‐market learning

effect and the cross‐market hedging effect. First, although private information is not observable to other market

participants, investor beliefs and attributes can be inferred through their trading activities. For example, order flows in

the stock market have been shown to reveal supplementary information about the macroeconomic fundamentals that

are relevant for valuing assets in the bond markets (Underwood, 2009). Investors in the bond futures market would

monitor and infer information from the trades in the stock market.

1

As the order flows in the stock market become

more informative after the equity market opens,

2

bond futures investors can learn more information from the

stock market, resulting in more informed order flows in the bond futures market as well. The second source of the

stock‐market‐opening effect is the spillover of private information caused by cross‐market hedging (Fleming et al.,

1998). Consider an investor with a diversified portfolio consisting of stocks and bonds. Her private expectations about

stock returns directly change her demand for stocks. Indirectly, portfolio rebalancing would lead to a change in her

demand for bonds, eventually generating order flows in both markets. Through cross‐market hedging, private

information is conveyed simultaneously through the order flows in both stock and bond futures markets. Hence, an

increase in price discovery in the stock market would coincide with a corresponding increase in price discovery in bond

futures markets.

Our empirical design follows an event‐study approach. Specifically, we investigate the price discovery of government

bond futures in the 30‐min intervals immediately before and after 9:30 a.m. (Eastern Time), when the US stock market

opens. We choose the 1‐hr event window (from 9:00 a.m. to 10:00 a.m.) during the opening period for three reasons.

First, during regular trading hours, variation in price discovery and informed trading is relatively modest. However, a

structural break in the informational efficiency of stock prices occurs at the opening and the closing (Admati &

Pfleiderer, 1988; Brandt & Kavajecz, 2004). These exogenous shifts make it possible to investigate the spillover effect of

the US stock market on global government bond markets. Second, we choose the US market opening (9:30 a.m.) instead

of the closing (4:00 p.m.) because major European stock markets (e.g., London & Frankfurt) are still open. This allows

us to evaluate the incremental effect of the US stock market. Third, our choice of the event window leaves out the effect

of economic news, which is often released at 8:30 am.

We measure price discovery in markets that is sequential in time (e.g., 9:00–9:30 a.m. vs. 9:30–10:00 a.m.).

3

This

setting is different from the case of parallel markets, where trading takes place simultaneously in multiple venues.

4

In a

parallel‐market study, international comparisons of price discovery are often limited to one or 2 hr of overlapping

trading times. These few overlapping hours may lead to bias against the earlier market traders because they cannot

learn from past price movements, unlike the newly arrived traders in the later market.

1

The “learning by trading”effect was studied by Grundy and McNichols (1989), Dow and Gorton (1993), and Leach and Madhavan (1993). These studies conclude that uninformed traders infer

information from the trading activities of informed traders.

2

Barclay and Hendershott (2003) provide the intuition behind the increase in price discovery in the stock market after the regular trading session begins. They argue that, compared to the preopening

period, “larger volumes of liquidity trade facilitate the price discovery process and result in more price discovery and more efficient prices during the trading day”(p. 1043). In this paper, we treat the

increase in price discovery in the stock market at the opening as given. Instead, we focus on the cross‐market spillover effect of the increase in price discovery.

3

Because the trading volume and dealer quoting behavior might differ across trading hours, studying price discovery across different times of the day is necessary to achieve a clean separation of price

innovations from different markets (see, e.g., Barclay & Hendershott, 2003; He, Lin, Wang, & Wu, 2009).

4

Studies in government bond futures, including Mizrach and Neely (2008) and Fricke and Menkhoff (2011), employ common price discovery measures, such as the Hasbrouck (1995) information share

(IS) and the Gonzalo and Granger (1995) permanent‐transitory decomposition. See Baillie, Booth, Tse, and Zabotina (2002) for detailed model descriptions.

780

|

INDRIAWAN ET AL.