Distribution of Ownership, Short Sale Constraints, and Market Efficiency: Evidence from Cross‐Listed Stocks

• Published date: 01 September 2014
• DOI: http://doi.org/10.1111/fima.12073
• Date: 01 September 2014

Distribution of Ownership, Short Sale

Constraints, and Market Efficiency:

Evidence from Cross-Listed Stocks

Louis Gagnon and Jonathan Witmer∗

We investigate the interplay between the distribution of ownership, short sale constraints, and

market efficiency. Using minute-by-minute data during the period surroundingthe short sale ban

of 2008, we demonstrate that short sale restrictions cause price disparities among cross-listed

stocks when ownership in the stocks is distributed unevenly across the two markets. The stocks

tend to trade at a premium in the market where long sellers are relatively scarcer, whichreduces

the speed at which prices adjust to bad news. The premium is driven primarily by an increase

on the ask side of the market where ownership is thinner, is only evident when prices are moving

down, and disappears quickly.

The notion that short sale constraints act like “sand in the gears,” preventing arbitrageurs

from enforcing the “law of one price,” is already well established empirically, but there is no

consensus in the literature as to how a stock’s ownership distribution across competing trading

venues affects the way in which short sale constraints impact market efficiency. We address this

important question in the context of US cross-listed stocks around the temporary short sale ban

imposed by the Securities and Exchange Commission (SEC) on financial stocks from September

19, 2008, to October 8, 2008.1

US cross-listed stocks from Canada offer an especially convenient vantage point for this

investigation for three key reasons. First, ownership in Canadian cross-listed stocks is heavily

tilted towardCanadian domiciled investors. Figure1 demonstrates that domestic investors capture

more than 80% of all Canadian financial shares held by Canadian and US investors combined,

suggesting that the short sale ban was more binding on the United States due to the relative

scarcity of long sellers in the United States when compared to Canada. In addition, short interest

∗Louis Gagnon is a Professorof Finance and Distinguished Faculty Fellowof Finance in the School of Business, Queen’s

University in Kingston, Ontario. JonathanWitmer is an Assistant Director of the Research Teamin the Financial Markets

Department at the Bank of Canada in Ottawa, Ontario.

This paper was previously circulated under the title “Short Changed? The Market’s Reaction to the Short Sale Ban of

2008.” Louis Gagnon is gratefulfor financial support from Queen’s University’s Office of Research Services and from the

Social Sciences and Humanities Research Council (SSHRC). We express our sincere thanks to the anonymousreferee and

MarcLipson (Editor) for their invaluable insights and guidance, as well as Marcus Braga-Alves,Jean-S´

ebastien Fontaine,

Scott Hendry, Andrew Karolyi, St´

ephane Lavoie, Dan Li, Ingrid Lo, Nadia Massoud, Teodora Paligorova, Adrian Pop,

Maxwell Stevenson,Caroline Trevithick,as well as seminar participants at the Bank of Canada, HEC Montr ´

eal, McMaster

University, Wilfrid Laurier University, the 2009 INFINITI Conference on International Finance, the 2009 International

Banking, Economics, and FinanceAssociation Conference, the 2009 Northern Finance Association Annual Meetings, and

the 2009 Financial Management Association Annual Meetings for their helpful comments. We thank Katharine Gagnon,

Sharlene He, Nicholas Michalski,and Michael Portner-Gartke for their research assistance.All remaining errors are our

own. The views expressed in this paper arethose of the authors and no responsibility for them should be attributed to the

Bank of Canada.

1See “SEC Halts Short Selling of Financial Stocks to Protect Investors and Markets,” Washington,DC, September 19,

2008, http://www.sec.gov/news/press/2008/2008--211.htm.

Financial Management •Fall 2014 •pages 631 - 670

632 Financial Management rFall 2014

Figure 1. Institutional Ownership in Canada and the United States

The figure displays the proportion of North American institutional owners that are based in Canada for

each of our treatment group stocks as of June 30, 2008. Our treatment group includes the Bank of Montreal

(BMO), Bank of Nova Scotia (BNS), Canadian Imperial Bank of Commerce (CM), Fairfax Financial

Holdings (FFH), Manulife Financial (MFC), Royal Bank of Canada (RY), Sun Life Financial (SLF), and

Toronto Dominion Bank (TD).

in Canadian financial stocks was up to five times larger in the United States before the ban,

compounding the effect of the relative scarcity of long sellers on the US side. This is illustrated

in Figure 2, where we scale short interest in each country by the number of shares held by local

institutions. Moreover, the ban was imposed concurrently in Canada by the Ontario Securities

Commission (OSC) on Canadian financial stocks cross-listed in the United States, ruling out

regulatory arbitrage between the two trading venues during the ban.2Thus, the short sale ban of

2008 creates a controlled experimental setting where we can examine the interplay between the

distribution of ownership, short sale constraints, and market efficiency.

Diamond and Verrecchia (1987) posit that short sale constraints do not typically affect price

levels, but reduce the speed at which prices adjust to news, especially bad news.3From this

perspective, absent ownership considerations, one wouldnot expect the frequency or the severity

2When compared to other US cross-listed stocks, Canadian stocks present several advantages: 1) the stocks are cross-

listed in the US as ordinary shares rather than in the form of an American Depositary Receipt, which is a more complex

intermediated facility; 2) the differences between the two trading venues in terms of the legal environment, accounting

standards, and the degree of economic and financial market development are minimal; and 3) the stocks are traded

concurrently on two markets sharing the same trading hours in the same time zone. Refer to Gagnon and Karolyi (2010),

Section I, for an overview of the institutional features associated with US cross-listed stocks.

3Previous research has found support for the model’s theoretical predictions around earnings announcements (Reed,

2007; Boehmer and Wu, 2013) and in individualstocks’ cross-autocor relation with aggregate stock market returns (Bris,

Goetzmann, and Zhu, 2007; Chen and Rhee, 2010; Saffi and Sigurdsson, 2011; Beber and Pagano, 2013; Boehmer and

Wu, 2013).

Gagnon & Witmer rDistribution of Ownership, Short Sale Constraints, and Market Efficiency 633

Figure 2. Short Interest Scaled by Institutional Ownership in Canada and the

United States

This figure displays the short interest as a proportion of institutional ownership for our treatment group

stocks. For each stock, the number of shares shorted in each country is scaled by the total number of shares

owned by institutions headquartered in that country.

of violations of the law of one price between the twotrading venues to be affected during the ban.

After all, when stocks are traded in two marketsconcur rently, market participants can easily “look

over the fence,” to see the prices at which the stocks are trading in the other market and adjust

prices accordingly. However, if ownership is distributed unevenly across the trading venues, the

short sale constraint will be more binding in the market with a relative shortage of long sellers or

a relative over-reliance on short sellers (or both). Then it is eminently conceivable that violations

of the law of one price will arise simply because the two markets won’t be able to adjust to bad

news at the same pace. In other words, we see the uneven distribution of ownership in a stock

across trading venues as a legitimate and rational limit to arbitrage whose impact has not been

investigated previously. We predict that violations of the law of one price between the United

States and Canada during the ban will be neither larger nor smaller, on average, among Canadian

financial stocks. We further posit that the shortage of “long” sellers and the absence of “short”

sellers in the United States, where ownership is thin, will cause ask limit orders in that market

to dry up and a reduction in the speed at which traders can adjust prices to bad news relative to

Canada. We also predict a drop in the share of aggregate trading captured bythe United States as

this market faced a relative shortage of long sellers and short sellers were unable to fill the void

created by the short sale prohibition.

Weconduct our investigation by comparing the intraday (minute-by-minute) dynamics of price

differentials between two groups of stocks. We employ a treatment group consisting of US

cross-listed Canadian financials subjected to the short sale prohibition in both jurisdictions from