Quote dynamics of cross‐listed stocks

• Published date: 01 June 2021
• Author: Bart Frijns,Ivan Indriawan,Alireza Tourani‐Rad
• Date: 01 June 2021
• DOI: http://doi.org/10.1111/irfi.12289

ORIGINAL ARTICLE

Quote dynamics of cross-listed stocks

Bart Frijns | Ivan Indriawan | Alireza Tourani-Rad

Department of Finance, Auckland University

of Technology, Auckland, New Zealand

Correspondence

Ivan Indriawan, Department of Finance,

Auckland University of Technology, Private

Bag 92006, 1020 Auckland, New Zealand.

Email: ivan.indriawan@aut.ac.nz

Abstract

We develop a model to assess the quote dynamics of

stocks listed in multiple markets. This model allows us to

explain the price formation mechanism and the degree of

information spillover. We show that this model can be

transformed to assess the dynamics of the spreads, the

efficient price, and the market's relative premium for

cross-listed stocks. Applying our model to a sample of

64 Canadian companies listed in the United States and

Canada, we document strong intermarket competition

among liquidity providers; prices mainly adjust to trades in

their respective market, suggesting some degree of infor-

mational frictions; and U.S. trades have a greater price

impact than Canadian trades. We further find that the

U.S. market is informationally dominant due to its more

competitive quote-setting behavior and larger incorpora-

tion of informational shocks.

KEYWORDS

cross-listings, error-correction model, market microstructure,

quote dynamics

JEL CLASSIFICATION

C32; G15

1|INTRODUCTION

Market microstructure theories suggest that information can be inferred from quotes (e.g., Demsetz, 1968; Huang &

Stoll, 1994; Jang & Venkatesh, 1991) and trades (e.g., Bagehot, 1971; Copeland & Galai, 1983; Glosten & Milgrom,

1985). Quotes are informative because they reflect the information of liquidity providers while trades are informative

Received: 15 May 2018 Revised: 26 September 2019 Accepted: 23 October 2019

DOI: 10.1111/irfi.12289

© 2019 International Review of Finance Ltd. 2019

International Review of Finance. 2021;21:497–522. wileyonlinelibrary.com/journal/irfi 497

because when investors trade, part of their information is revealed to the market. The relation between these infor-

mation sources and subsequent price movements are the basis for many microstructure theories (see,

e.g., O'Hara, 1995).

Empirical investigations of market microstructure theories have predominantly taken place in a single market

setting (e.g., quote revisions as a function of spreads, price impact of trades, price reactions to order flow, etc.). How

these theories apply in a multimarket setting has received considerably less attention. In a multimarket context, the

main focus has been on price discovery at the intraday level (e.g., Eun & Sabherwal, 2003; Pascual, Pascual-Fuster, &

Climent, 2006), spillover (Levine & Schmukler, 2006) and commonality (Gagnon & Karolyi, 2009) at daily frequencies.

To date, limited research has been carried out on examining the impact of trades, order flow, and spreads in a multi-

market setting at a high frequency. Our work fills this gap by examining the mechanisms that link prices in multiple

markets using transaction-level data.

In this paper, we examine how quotes in one market reflect valuable information for liquidity providers in

another market, and how trades in one market affect quotes in the other market. We build on the framework of coi-

ntegrated quotes in a single market and extend it to a multimarket setting where bid and ask prices are modeled

jointly.

1

We use this model to show how information affects bid and ask prices in two different markets and assess

the degree of information spillover between them. From a methodological perspective, we extend the work of Engle

and Patton (2004) and Escribano and Pascual (2006) to explain price dynamics in a multimarket context. A unique

feature of our model is that it can further be developed to extract an implied vector autoregression (VAR) for various

microstructure variables, such as the bid-ask spreads, the change in price midpoint, and the difference in midquotes

across markets. These variables reflect the cost of trading, the efficient price, and the market's relative premium

when trading cross-listed stocks, respectively. To the best of our knowledge, our paper is the first to examine quote

and spread dynamics in a multiple market context.

We apply our model to a sample of Canadian firms that are cross-listed in the United States.

2

We find that

quote-related information directly affects prices in both markets, indicating strong intermarket competition among

liquidity providers. For trade-related information, we observe that signed order flows, defined as the aggregated

number of trades in buy and sell directions, have some impacts on quotes across markets, whereas individual

trades (either buys or sells) do not. This observation suggests that there is some degree of informational frictions

between the two markets. These are novel findings in the market microstructure and cross-listing literature, and

cannot be observed in a single market setting. Further analysis shows that trade-related information plays a

greater role in the United States than in Canada, leading to a greater impact of the U.S. trades on the implied

efficient price and on their relative premium. Our findings highlight the informational dominance of the

U.S. market which is observed through more competitive quote-setting behavior and faster incorporation of infor-

mational shocks.

The results of our paper provide new evidence on the role of trade-related information for intermarket trading.

In particular, we test the importance of microstructure variables, such as trade direction (Glosten & Milgrom, 1985;

Jang & Venkatesh, 1991), trade size (Barclay & Warner, 1993; Chakravarty, 2001; Easley & O'Hara, 1987), trade

duration (Dufour & Engle, 2000; Easley & O'Hara, 1992), and signed order flow (Kyle, 1985) as information signals

across market. While these studies provide important insights on the impact of market microstructure variables in a

single market, it is not clear what their impact is across markets. For instance, we document that liquidity providers

compete with each other by adjusting their bid-ask spreads. We also document that individual trades have minor

impact across markets, even for the most fungible cross-listed assets such as the Canadian stocks, whereas signed

order flow (aggregated number of trades) contains more information and has stronger impact across markets. These

findings, which were not previously investigated, are important given the sheer number of stocks cross-listed glob-

ally. The NYSE alone reported that by the end of 2018, 514 non-U.S. stocks from 46 countries were listed on the

exchange, out of a total of 6,400 listings.

3

One hundred twenty of these stocks are from Canada and another

115 stocks from Latin America where trading hours coincide with the U.S. exchanges. These statistics emphasize the

importance of identifying and understanding the cross-market information transmission.

498 FRIJNS ET AL.