Journal of Financial Economics: Douglas Cumming, Sofia Johan, Dan Li

Journal of Financial Economics 99 (2011) 651–671
Contents lists available at ScienceDirect
Journal of Financial Economics

journal homepage: www.elsevier.com/locate/jfec
Exchange trading rules and stock market liquidity$

Douglas Cumming a,n, Sofia Johan a,b, Dan Li a
a
York University, Schulich School of Business, 4700 Keele Street, Toronto, Ontario, Canada M3J 1P3
b
University of Tilburg, Tilburg Law and Economics Centre (TILEC), Postbus 90153, 5000 LE Tilburg, The Netherlands
a r t i c l e in fo abstract
Article history: We examine stock exchange trading rules for market manipulation, insider trading, and
Received 8 January 2009 broker–agency conflict, across countries and over time, in 42 stock exchanges around
Received in revised form the world. Some stock exchanges have extremely detailed rules that explicitly prohibit
9 December 2009
specific manipulative practices, but others use less precise and broadly framed rules.
Accepted 31 January 2010
We create new indices for market manipulation, insider trading, and broker–agency
Available online 5 November 2010
conflict based on the specific provisions in the trading rules of each stock exchange. We
JEL classification: show that differences in exchange trading rules, over time and across markets,
G12 significantly affect liquidity.
G14
& 2010 Elsevier B.V. All rights reserved.
G18
K22
Keywords:
Market manipulation
Liquidity
Insider trading
Broker–agency conflict
Law and finance
1. Introduction
$
We are indebted to the Paolo Baffi Research Center, IIROC and
Stock exchanges around the world invest considerable
Capital Markets Cooperative Research Center for financial support. We
owe thanks to Michael Aitken, Paul Irvine, Elisa Di Marco, Benoit Mario manpower, technological effort, and financial resources to
Papillon, Thomas Poeppe, and Frank Sensenbrenner for helpful com- curb market manipulation and promote market efficiency
ments and suggestions, as well as the seminar participants at the and integrity (Aitken and Siow, 2003; Avgouleas, 2005;
Canadian Law and Economics Association Annual Conference (Septem- Comerton-Forde and Rydge, 2006). It is widely regarded
ber 2007 and September 2009), the SMARTS Conference on Surveillance
(February 2008), the Bursa Malaysia (April 2008), the University of
that securities law (La Porta, Lopez-de-Silanes, Shleifer, and
Amsterdam (September 2008), Bocconi University (October 2008), York Vishny, 1998; La Porta, Lopez-de-Silanes, and Shleifer,
University (January and December 2009), the University of British 2006; Jackson and Roe, 2009) and market microstructure
Columbia/University of Calgary National Center for Business Law (March (Harris, 2002; Harris, Aitken, Cook, and McInish, 2009) play
2009), the DeGroote-Regulation Services Conference on Microstructure
an important role in the development of stock markets
(November 2008), the Australasian Banking and Finance Conference
(December 2008), the American Law and Economics Association (May around the world. Despite these important developments
2009), Finlawmetrics (June 2009), and the Northern Finance Association in the literature, a dearth of attention has been paid to the
(September 2009). differences across exchanges with respect to the treatment
n
Corresponding author. of market manipulation within their trading rules.
E-mail address: douglas.cumming@gmail.com (D. Cumming).
URLS: http://ssrn.com/author= 75390 (D. Cumming), http://
In this paper, we show international differences in
ssrn.com/author = 370203 (S. Johan), http://ssrn.com/author=868088 trading rules for stock or equity exchanges, and we
(D. Li). examine the impact of market integrity rules on the
0304-405X/$ - see front matter & 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.jfineco.2010.10.001
652 D. Cumming et al. / Journal of Financial Economics 99 (2011) 651–671
performance of equity marketplaces. Specifically, we advantage of inevitable loopholes and, if so, more detailed
study the differences in regulation across 42 exchanges exchange rules could have a negative effect on liquidity.
worldwide during the time period of 2006–2008 and then Although exchanges do not amend their rules very
proceed to investigate whether integrity-related exchange frequently, amendments to rules are instituted over time.
trading rules matter for market liquidity. For the purposes Most notably for European exchanges, in November 2007
of this article, trading rules refer to the rules and the Directive on Markets in Financial Instruments (MiFID)
regulations that regulate the activities within a stock became effective and thereby gave rise to more detailed
market and the conduct of its participants, namely, the rules and more transparent investor protection for the
exchange and the members of the stock exchange who European exchanges. Although some European exchanges,
agree to be bound by such rules and regulations. such as the London Stock Exchange, already had in place
We create new indices for trading rules pertaining to trading rules that were analogous to the new rules in MiFID,
market manipulation, insider trading, and broker–agency others such as the Austrian exchange had significantly less
conflict for these 42 stock exchanges in both developed detailed rules prior to MiFID. Because the introduction of
and emerging markets. For the purposes of this article, MiFID affects only the countries of the European Union, it
market manipulation refers to the trading practices that creates a natural experimental setting in which to assess the
distort prices and enable manipulators to profit at the impact of exchange rule restrictions on trading activity. In
expense of other market participants. Insider trading this article, we exploit this setting to shed light on our
refers to acting on material nonpublic information. research question by examining the dynamics of the market
Broker–agency conflict refers to the actions that brokers liquidity measures between the two groups of exchanges
could take while acting as the agent of a client that around the introduction of MiFID. Because MiFID is
benefits the broker (or some other affiliated party) at the introduced as a major part of the European Union’s Financial
expense of the client or the market more generally. Some Services Action Plan (FSAP) instead of as a result of one
stock markets such as Nasdaq have extremely detailed single jurisdiction’s need to improve regulation, endogeneity
rules that explicitly prohibit specific manipulative prac- issues that relate rule changes to market outcomes are
tices and broker–agency conflict, as well as rules that are minimized in our experimental setting.
designed to curtail the presence of insider trading. For The data presented in this article show a strong and
example, Nasdaq’s rules provide detailed provisions robust effect of trading rules on liquidity. Detailed trading
regarding wash trades, pre-arranged trading, fictitious rules are positively associated with velocity and negatively
orders, giving-up priority, churning, front-running, and a associated with volatility and bid-ask spreads. We show this
variety of other types of practices that constitute market effect with panel data that vary across time and countries by
manipulation. Other exchanges are less precise and have considering a variety of robustness checks that include, but
broadly framed rules regarding what constitutes market are not limited to, fixed effects modeling and difference-in-
manipulation or broker–agency conflict. differences tests. To isolate the influence of the trading rules,
In view of the significant differences in the way trading we also control for a number of plausible factors that could
rules regulate market manipulation, insider trading, and effect trading activity based on prior academic works,
broker–agency conflict across countries and over time, it is including exchange institutional features (Röell, 1992),
worth considering whether these differences matter. To market microstructure aspects (Stoll, 2000), and interna-
this end, in addition to showing the differences in trading tional differences in securities regulation (La Porta, Lopez-
rules and developing new indices of market surveillance, de-Silanes, and Shleifer, 2006; Jackson and Roe, 2009),
we examine whether the differences in trading rules can among other things. The effect of rules on liquidity is robust
help to explain the differences in liquidity among to controls for economic, legal, and institutional differences
exchanges. Specifically, we examine whether a correlation across exchanges that might have been correlated with
between trading rules exists and a series of liquidity country specific differences in the drafting of trading rules.
measures that include velocity, volatility, and relative bid- This strong evidence is due to the fact that exchanges that
ask spread. The primary function of a marketplace is to specifically recognize and prohibit certain acts in the
provide liquidity to market participants. The effectiveness marketplace enhance investor confidence. As well,
of an exchange is affected by its rules that regulate security exchanges with more specific rules invariably have residual
transactions. We consider two competing hypotheses catch-all clauses that explicitly outline the spirit of the rules
regarding the impact of trading rules on liquidity. On the and regulations and prohibit a vaguely defined ’’any other
one hand, one can argue that vague regulations create type of manipulative activity’’ such that (arguably) there is
inefficiency as investors and traders are not clear as to scant scope for exploiting potential loopholes.
which activities are acceptable and which ones are in A few recent articles are closely related to our own. La
breach of the rules. Detailed rules, therefore, could give rise Porta, Lopez-de-Silanes, and Shleifer (2006) and Jackson
to greater investor confidence, provide greater dissemina- and Roe (2009) show that securities law matters for
tion of knowledge about prohibited conduct, and facilitate facilitating stock market development in 49 exchanges
invigilation of such rules, which in turn might reinforce around the world. Aitken and Siow (2003) provide a
investor confidence in the marketplace. As a result, these ranking of exchanges based on efficiency and integrity.
rules might help to improve trading activity, reduce Cumming and Johan (2008) provide survey evidence that
uncertainty, and decrease trading cost. On the other surveillance technology and information sharing facilitate
hand, one could argue that detailed regulations create market integrity. Hail and Leuz (2006), Daske, Hail, Leuz,
inefficiency as investors and traders are able to take and Verdi (2008) and Lampert, Leuz, and Verracchia
D. Cumming et al. / Journal of Financial Economics 99 (2011) 651–671 653
(2007) show that stronger securities law, accounting Insider trading can take many different forms, two of
rules, and stricter enforcement mechanisms lower firms’ which involve brokers using the information of a client
cost of capital. The findings in these articles are consistent order: client precedence and front-running. Client pre-
with a broader literature on the importance of securities cedence refers to brokers violating the time priority of
regulation and market surveillance for market efficiency client orders. A client precedence rule is violated during
and integrity.1 More generally, our article contributes to insider trading when a broker initiates a trade on his own
the general question of the value of broadly framed versus account shortly ahead of the execution of a client’s order,
specific rules in regulating markets and society (see, e.g., with the client’s trade being executed at a worse price.
Ferguson and Peters, 2003; Stevenson, 2005). Front-running likewise refers to brokers trading ahead of
We provide a novel source of information for under- a client’s order. In the case of front-running, upon receipt
standing international differences in stock exchanges. We of a large client order, a broker trades shortly prior to a
show that stock exchange trading rules, which specify in client’s order with the expectation that the client’s order
detail rules that pertain to market manipulation, facilitate will move the price. Front-running can also involve
trading activity. The implication is that an exchange’s brokers who, after receiving a client’s order, take the
trading rules are an important source of international opposite position to the client’s order in the market
differences in stock markets. This information is very without the client’s knowledge and then, immediately
transparent and readily visible for use in future research. thereafter, the same broker crosses with the same client
An index of exchange trading rules is provided herein. off-market at a profit.
The paper is organized as follows. Section 2 describes Other forms of insider trading can involve the use of
stock exchange trading rules and the creation of an index material nonpublic information about the company being
for exchange surveillance. The data are introduced in traded. Trading rules can mitigate the presence of this
Section 3. Section 4 presents multivariate analyses of the form of insider trading by prohibiting trading ahead of the
relation between the exchange surveillance index and public release of research reports created by brokerages,
trading activity. Concluding remarks follow in the last as well as seperating research and trading departments at
section. brokerages (commonly referred to as Chinese walls).
Rules that limit affiliation between exchange members
2. The indices of exchange trading rules and hypothesis and member companies, or between members and their
investment company securities, mitigate the flow of
In this section, we explain forms of market manipula- information that could be material and nonpublic. Rules
tion, insider trading, and broker–agency conflict, and we can also provide details with respect to the nature of
build indices for stock exchange trading rules. Rules can communication between brokerages and the public by
be broken down into one of three types: rules designed to regulating how the flow of material nonpublic informa-
mitigate insider trading, rules designed to limit market tion is released. Further, trading rules sometimes limit
manipulation, and rules designed to limit broker–agent brokerage ownership, the extent to which brokerages can
conflicts. influence or reward employees of others, or ban anti-
intimidation or coordination activities (e.g., to stop people
from reporting illegal activities). These restrictions can
2.1. Indices of exchange trading rules
have the effect of limiting the flow of material nonpublic
information.2
We define the indices in this section.
2.1.2. Market Manipulation Rules Index

2.1.1. Insider Trading Rules Index Market manipulation rules encompass price manipula-
Insider trading refers to a market participant who acts tion, volume manipulation, spoofing, and disclosure
on material nonpublic information. Although rules prohi- manipulation.
biting insider trading in general are commonplace around
the world, specific regulations governing market partici-
2.1.2.1. Price Manipulation Rules Index. Price manipulation
pants with respect to insider trading differ significantly
can be carried out in many different ways and take many
across exchanges.
forms. One common way is where one broker (or col-
luding brokers) enters purchase orders at successively
1
See, e.g., Aggarwal (2001), Aggarwal and Wu (2006), Allen and Gale higher prices to create the appearance of active interest in
(1992), Allen and Gorton (1992), Carson (2003), Clayton, Jorgensen, and a security, which is also termed ramping or gouging. This
Kavajecz (2006), Comerton-Forde and Rydge (2006), Comerton-Forde
can also take the form of pump and dump schemes
and Tang (2007), Daouk, Lee, and Ng (2006), DeMarzo, Fishman, and
Hagerty (2005), Gerard and Nanda (1993), Hillion and Suominen (2004), whereby exchange participants generate a significant
Jarrow (1992, 1994), Kumar and Seppi (1992), La Porta, Lopez-de- increase in price and volume for a security and carry out a
Silanes, and Shleifer (1999), La Porta, Lopez-de-Silanes, Shleifer, and
Vishny (1997, 1998, 2002), Mahoney (1999), Merrick, Naik, and Yadav
2
(2005), Ni, Pearson, and Poteshman (2005), Pagano, Randl, Röell, and In some countries the probability of detection of insider trading is
Zechner (2001), Pagano, Röell, and Zechner (2002), Peng and Röell low and, even upon detection and prosecution, the ensuing fines are also
(2009), O’Hara and Mendiola (2003), Pirrong (1993, 1995a, b, 1999, low (see, e.g., Bhattacharya and Daouk, 2002, 2009). We consider
2004), Pistor, Keinan, Kleinheisterkamp, and West (2003), Pistor and Xu separate variables for insider trading laws around the world (e.g., Beny,
(2003, 2005), Prichard (2003), Reiffen and Robe (2007), Romano (1993, 2005, 2007), among others, but those variables did not materially impact
2001, 2002) and Vitale (2000). the results presented here.
quick flip, and the securities are then sold (often to retail price, but we consider wash trades to more significantly
customers) at the higher prices. Another similar type of affect volume. As such, wash trades are categorized as
price manipulation takes the form of pre-arranged trad- part of volume manipulation.
ing. Pre-arranged trades involve colluding parties simul-
taneously entering orders at an identical price and 2.1.2.3. Spoofing Manipulation Rules Index. Spoofing, also
volume. Because pre-arranged trades avoid the order known as ‘painting the tape’, is a form of market manip-
queue, they can influence the price of a security. ulation that involves actions taken by market participants
Price manipulation can be carried out through dom- to give an improper or false impression of unusual activity
ination and control, and it can take the form of corners or or price movement in a security. Some trading rules have
squeezes in cross-market activity. Corners and squeezes very general statements of prohibition toward actions that
involve shortages in one market that can affect the price give rise to a false appearance. Other exchanges more
of a cross-market security. A corner involves securing explicitly indicate ways in which false appearance might
control of the bid- or demand-side of both the derivative be created, which includes fictitious orders, giving up
and the underlying asset, and the dominant position can priority, layering of bids-asks, and switches. The more
be exploited to manipulate the price of either. A squeeze general act of entering fictitious orders involves entering
involves taking advantage of a shortage in an asset by orders on one side of the market, then completing orders
controlling the demand-side and exploiting market con- on the other side of the market and deleting the original
gestion during such shortages in a way that creates order after the trade occurs. Giving up priority refers to
artificial prices. Another related form of manipulation deleting orders on one side of the market as they approach
includes mini-manipulations whereby trading in the priority and then entering the order again on the same side
underlying security of an option is carried out to of the market. Layering of bids-asks refers to traders or
manipulate its price so that the options become in-the- brokers that stagger orders from the same client reference
money (Merrick, Naik, and Yadav, 2005). at different price and volume levels to give the misleading
Price manipulation can also be carried out to take impression of greater interest in the security from a more
advantage of market setting whereby brokers cross-order diverse set of exchange participants and could be viewed as
at the short-term high or low to effect the volume being carried out for the purpose of manipulation. Switches
weighted average price or to set the price in one market involve deleting orders on one side of the market as they
for the purpose of a cross in another market. approach priority and then entering the order again on the
Three different forms of price manipulations refer to a opposite side of the market. These distinctions are some-
specific time period: marking the open with regard to the what subtle—nevertheless, these different scenarios are
opening of the market, marking the close with regard to explained in detail in some exchange trading rules.
the closing of the market, and trades to manipulate prices
at end of the month, quarter or year. The opening session 2.1.2.4. False Disclosure Rules Index. Distinct from insider
can be subject to particular types of manipulation subject trading rules, some rulebooks include information per-
to the rules for entering bids and asks in the pre-opening taining to false disclosure. For instance, market partici-
session. Similarly, end-of-day trades could be geared pants could actively distribute false or misleading
toward manipulating the closing market price of the information that has the effect of distorting the market-
security, and exchanges often specifically prohibit this place. Alternatively, there can be a failure to disclose
type of act. Financial record keeping among companies information such as the mandatory disclosure of owner-
provides incentives to manipulate share prices around the ship interests when they reach threshold level. This latter
end of the month, quarter and year that depend on the form of manipulation is commonly known as parking or
governance specific to the company. warehousing.
Overall, we refer to trading rules pertaining to price
2.1.2.2. Volume Manipulation Rules Index. Volume manip- manipulation, volume manipulation, spoofing, and false
ulation can take two primary forms: churning and wash disclosure as the market manipulation rules. Below, we
trading. Churning refers to the excessive trading of a stock aggregate these rules to form separate indices for each,
to inflate its volume thereby giving rise to the false which we refer to as subcomponent indices. Then we
impression that positive investor sentiment exists for the combine them in their sum total to form the Market
stock. While we recognize that the churning of client Manipulation Rules Index, one of the three primary legal
accounts could be carried out by traders or brokers to indices we are creating in this paper. These indices are
generate commission fees, given that the end result of considered separately from insider trading rules and
churning is to manipulate markets, and that the central broker–agency conflict rules, which form the other two
motivation of traders or brokers in churning both house primary indices.
accounts and client accounts is to manipulate the
appearance of volume, we deem churning as a form of 2.1.3. Broker–Agency Conflict Rules Index
volume manipulation. Brokers act on behalf of clients but can do so in ways
Wash trading means having the same client reference that are against client interests. This type of principal
on both sides of a trade. While there is no beneficial agent problem could arise from the failure of the broker to
change in ownership, wash trades have the effect of obtain the best price for a client (commonly known as a
creating a misleading appearance of an active interest in a breach of a trade through obligation). the broker charging
stock. We realize that wash trades can indirectly effect excessive fees, or the broker acting in ways that are
generally detrimental to client interests such as by 2.2. Testable hypotheses

investing in securities that do not match the risk-return
profile of the client (referred to as breach of the know- Following market microstructure literature, we focus
your-client rule).3 As well, brokers could use the on the three major measures of liquidity: velocity,
exchange’s name improperly in marketing their services volatility, and bid-ask spread. Trading velocity is defined
or carry out other forms of improper or unethical sales as the domestic share turnover per domestic market
and marketing efforts. For broker–agency conflict rules, capitalization (World Federation of Exchanges, 2006–2008).
we use information explicitly indicated in the rules of the High velocity is associated with the high turnover of
exchange, not guidelines from professional associations stock, which means that shares change hands more
such as the Chartered Financial Analysts ethics guidelines frequently, implying a more liquid market. Following
and the like. Roll (1988), Massimb and Phelps (1994), and Madhavan,
Table 1 outlines the different types of manipulation Porter, and Weaver (2005), higher volatility implies
described in stock exchange trading rules. The trading rules lower liquidity. Bid-ask spread is the compensation for
for a stock exchange are drafted with varying degrees of providing immediacy, which a trader receives for the
specificity as they outline the exchange membership risks that he could have to unwind his position at a loss
requirements, listing requirements, trading rules and reg- in the future. A smaller spread is associated with higher
ulations, and especially trading practices that are prohibited. liquidity (Amihud and Mendelson, 1988; Massimb and
Each of the different rules for insider trading, market Phelps, 1994).
manipulation, and broker–agency conflict described in the Below we test the proposition that explicit rules
exchanges’ trading rules are weighted equally in the indices pertaining to insider trading, market manipulation and
used in this paper.4 The Insider Trading Rules Index has ten broker–agency conflict enhance investor confidence,
items. The Market Manipulation Rules Index encompasses a mitigate abuse, and thereby facilitate trading activity. In
total of 14 items: price manipulation (seven items), volume other words, the central hypothesis is that vague
manipulation (two items), spoofing (three items), and false exchange trading rules do not provide adequate guidance
disclosure (two items). The Broker–Agency Conflict Rules and information for investors or traders, who are not
Index contains five items. However, certain rules could be sufficiently clear as to which activities are unacceptable.
relatively more important, but we do not have enough As such, investors and traders are less likely to trade in the
degrees of freedom to treat each rule separately. Plausible market for fear that it is more likely to be manipulated.
adjustments to different weightings do not materially The competing hypothesis is that detailed regulations give
change the empirical results reported below. rise to loopholes that investors and traders can take
The Insider Trading Rules Index and Market Manipula- advantage of, thereby creating inefficiencies and lowering
tion Rules Index also consider securities regulation trading activity. A priori, our expectation is that the latter
provisions when they are specific about the regulations effect is outweighed by the former, as detailed regulations
pertaining to trading on stock exchanges. Our analyses of can signal to market participants that exchanges actively
securities codes reveal a couple of cases in which the monitor and enforce regulations pursuant to investor
trading rules were more detailed in securities law. In protection (see, generally, Laffont and Tirole, 1993). We
China, we use the rules of the China Securities and therefore expect that detailed trading rules enhance
Regulatory Commission for the Shanghai and Shenzhen velocity and reduce stock market volatility and bid-ask
exchanges. In Canada, the pertinent rules are found in the spreads.
Universal Market Integrity Rules, which come from an In addition to differences in levels of rule detail across
independent regulatory body known as the Investment countries, in our analysis we make use of a material change
Industry Regulatory Commission of Canada. By contrast, to trading rules across countries due to the Directive on
the Broker–Agency Conflict Rules Index does not consider Markets in Financial Instruments. In November 2007,
professional association rules, such as that of the MiFID, an European-wide harmonization directive, became
Chartered Financial Analysts Code of Professional Con- effective. Because the timing, motivation, and content of
duct. The reason for this exclusion is that the exchange MiFID were not instigated by any one specific European
members are not obligated to be a part of these different exchange or European country, but at the European Union
professional associations to trade on the exchange. level, this legislative change can be regarded as exogenous,
thereby providing a useful test of causality between rules
and liquidity. MiFID became effective on November 1,
2007. While an earlier directive, the Market Abuse
Directive (MAD), was introduced in 2004, the appropriate
3
In the US, the breach of trade through obligation was released measures were not in place in 2004 across member states
under Regulation National Market System (Regulation NMS) and
for a number of reasons. First, surveillance data from
published in the Federal Register in June 2005. Regulation NMS was
phased in over many months in 2007, see http://www.Nasdaqtrader. Cumming and Johan (2008) indicate exchanges in 2004
com/content/marketregulation/regnms/regnms_factsheet.pdf. Assigning and 2005 did not adopt/implement the provisions in MAD
the value one or zero to trade through in the US during this period does in a meaningful way. Second, articles in MiFID cover many
not materially influence our results. aspects of MAD, and state that provisions are needed to
4
The equal-weighting is consistent with the approach used in most
law and finance studies, such as those by La Porta, Lopez-de-Silanes,
ensure that MAD principles are in place by November 1,
Shleifer, and Vishny (1998) and La Porta, Lopez-de-Silanes, and Shleifer 2007 (see, for example, Article 25 in MiFID). The draft
(2006). provisions in MiFID in 2004 even made this point, so
Table 1
Definition of variables.
This table defines the variables, including the trading rule indices.
Variable Definition
Insider trading rules

Front-running A dummy variable equal to one if the trading rules explicitly prohibit a broker’s house or employee
account from buying or selling in a period shortly prior to significant buying or selling by a client.
Client Precedence A dummy variable equal to one if the trading rules explicitly prohibit a broker from violating the time
priority of client orders.
Trading ahead of research reports A dummy variable equal to one if the trading rules explicitly prohibit brokers with proprietary access
to research reports from trading ahead of the release of the research report.
Separation of research and trading A dummy variable equal to one if the trading rules specify that research departments and trading
departments must have a Chinese wall separating these departments.
Broker ownership limit A dummy variable equal to one if the trading rules specify maximum ownership limits for brokerages
or employees with respect to any given security.
Restrictions on affiliation A dummy variable equal to one if the trading rules specify limits or restrictions on affiliation between
exchange members and member companies.
Restrictions on communications A dummy variable equal to one if the trading rules specify limits or restrictions on brokerages’
communications with the public.
Investment company securities A dummy variable equal to one if the trading rules specify restrictions or bans on the trading of
members’ own or affiliated investment company securities.
Influencing or rewarding employees of A dummy variable equal to one if the trading rules specify bans on any means of influencing or
Others rewarding employees of other members or member companies.
Anti-intimidation/coordination A dummy variable equal to one if the trading rules specify bans on any form of intimidation of or
coordination with other members or member companies.
Insider Trading Rules Index Sum of dummy variables for Front-running, Client precedence, Trading ahead of research reports,
Separation of research and trading, Broker ownership limit, Restrictions on affiliation, Restrictions on
communications, Investment company securities, Influencing or rewarding the employees of others,
and Anti-intimidation/coordination.
Price manipulation rules
Marking the open A dummy variable equal to one if the trading rules explicitly prohibit the placing of purchase orders at
slightly higher prices or sale orders at lower prices to drive up or suppress the price of the securities
when the market opens.
Marking the close A dummy variable equal to one if the trading rules explicitly prohibit the buying or selling of securities
at the close of the market in an effort to alter the closing price of the security.
Misleading end of the month/quarter/year A dummy variable equal to one if the trading rules explicitly prohibit transactions executed at a
Trades particular date to establish gains or losses or conceal portfolio losses or true positions in connection
with end of the month, quarter or year.
Intraday ramping/gouging A dummy variable equal to one if the trading rules explicitly prohibit the execution of a series of
trades over a short time period that generates a price movement over that period in which it is
unusual, given the trading history of the security.
Market setting A dummy variable equal to one if the trading rules explicitly prohibit market setting by crossing in the
short term, high or low. For example, this could be done to set the VWAP (volume weighted average
price) or cross market (setting the price in one market to justify crossing in the follow-on market).
Pre-arranged trades A dummy variable equal to one if the trading rules explicitly prohibit pre-arranged trades within an
extremely short time period whereby the client broker and another broker enter a bid and ask for the
same volume and price, which then generates a trade between the two brokers for the whole of the
volume. The volume of the order must be significant given the trading history of the security.
Domination and control A dummy variable equal to one if the trading rules explicitly prohibit a broker or client from
generating significantly greater price changes in a security, possibly for corners (securing control of
the bid/demand-side of both the derivative and the underlying asset, and explaining the dominant
position to manipulate the price of the derivative or the asset), squeezes (taking advantage of a
shortage in an asset by controlling the demand-side and exploiting market congestion during such
shortages in a way as to create artificial prices), and mini-manipulations (trading in the underlying
security of an option to manipulate its price so that the options become in-the-money).
Price Manipulation Rules Index Sum of dummy variables for marking the open, marking the close, misleading end of the month/
quarter/year trades, intraday ramping/gouging, market setting, pre-arranged trades, and domination
and control.
Volume manipulation rules
Churning A dummy variable equal to one if the trading rules explicitly prohibit excessive buying and selling of
stocks by a trader such as a broker to generate large commission fees (in the case of churning client
accounts) or the appearance of significant volume (in the case of churning house accounts or churning
client accounts).
Wash trade A dummy variable equal to one if the trading rules explicitly prohibit the same client reference on
both sides of a trade.
Volume Manipulation Rules Index Sum of dummy variables for Churning and Wash trade.
Spoofing rules
Giving up priority A dummy variable equal to one if the trading rules explicitly prohibit brokers from giving up priority,
such as entering a bid-ask for a significant quantity at a price away from priority and then both
canceling this order as it approaches priority, and re-entering the order shortly thereafter at a price
level further away from priority.
Table 1 (continued )
Variable Definition
Switch A dummy variable equal to one if the trading rules explicitly prohibit brokers from entering fictitious
orders, such as entering a significant quantity at or close to priority, then completing a trade on the
opposite side of the market, and thereafter deleting the original order shortly after the completion of
the opposite order.
Layering of bids/asks A dummy variable equal to one if the trading rules explicitly prohibit brokers from layering bids-asks,
such as stagger orders from the same client reference at different price and volume levels, with the
intent of giving a false or misleading appearance with respect to the market for the security.
Spoofing Rules Index Sum of dummy variables for Giving up priority, Switch, and Layering of bids-asks.
False disclosure rules
Dissemination of false and misleading A dummy variable equal to one if the trading rules explicitly prohibit the dissemination of false or
information misleading market information.
Parking or warehousing A dummy variable equal to one if the trading rules explicitly prohibit hiding the true ownership of
securities by creating a set of fictitious transactions and trades.
False Disclosure Rules Index Sum of dummy variables for Dissemination of false and misleading information and Parking or
warehousing.
Market Manipulation Rules Index Sum of Price Manipulation Rules Index, Volume Manipulation Rules Index, Spoofing Rules Index, and
False Disclosure Rules Index.
Broker–agency rules
Trade through A dummy variable equal to one if the trading rules explicitly prohibit the completion of a client’s order
at a price inferior to the best posted bid or ask; e.g., the market maker who received the order is
unable or unwilling to fill it at the best posted bid or ask price and, hence, the trade is instead executed
at the market maker’s price.
Improper execution A dummy variable equal to one if the trading rules explicitly prohibit brokers from charging fees for
completing a client order, which are unwarranted given the circumstances.
Restrictions on member use of exchange A dummy variable equal to one if the trading rules specify restrictions on exchange members’ use of
name the exchange name.
Restrictions on sales materials and A dummy variable equal to one if the trading rules specify restrictions on exchange members’ nature
telemarketing of sales and telemarketing.
Fair dealing with customers A dummy variable equal to one if the trading rules specify details with respect to the ’’know your
client rule’’ that requires brokerages to not make trades that do not fit within the clients interest, no
delays in the handling of client orders, and the like.
Broker–Agency Index Sum of dummy variables for Trade through, Improper execution, Restrictions on member use of
exchange name, Restrictions on sales materials and telemarketing, and Fair dealing with customers.
Surveillance, efficiency of judiciary, and
investor protection indices
Surveillance Index The principal component of single market surveillance and cross market surveillance. Source:
Cumming and Johan (2008). Available for a subset of countries, and provided contingent on
maintaining confidentiality and anonymity as exchanges do not want market participants to know all
of the things they do and do not look for in their surveillance.
Efficiency of the Judiciary Index Assessment of the efficiency and integrity of the legal environment. Scale from zero to ten; with lower
scores, lower efficiency levels. Source: La Porta, Lopez-de-Silanes, Shleifer, and Vishny (1998).
Investor Protection Index The principal component of Anti-Director Rights, Disclosure Requirements, and Liability Standards.
Sources: La Porta, Lopez-de-Silanes, and Shleifer (2006), Spamann (2010).
Market statistics
Velocity The ratio between the turnover of domestic shares and their market capitalization. The value is
annualized by multiplying the monthly moving average by 12, according to the following formula:
monthly domestic share turnover/month-end domestic market capitalization. Only domestic shares
are used to be consistent. Source: World Federation of Exchanges (2006–2008).
Volatility Volatility of each firm is calculated using the prior 60-month returns. For each exchange, the volatility
is the firm market capitalization weighted average of the volatilities of the firms consisting of average
and then annualized. Firm market capitalization is equal to the product of month-end shares
outstanding and stock prices, or the mid-price of bid and ask prices, if stock price is missing. The
monthly return, share outstanding, and stock price are from Thomson Reuters Datastream.
Bid-ask spread The relative quoted spread of each firm is the difference between the monthly bid and ask prices,
divided by the mid-point price. For each exchange, the relative quoted spread is the firm market
capitalization weighted average of the spreads of the firms that its major index consists of. Firm
market capitalization is equal to the product of month-end shares outstanding and stock prices, or the
mid-price of bid and ask prices, if stock price is missing. The bid and ask prices, shares outstanding,
and stock prices are from Thomson Reuters Datastream.
Log (market capitalization) Log of domestic market capitalization in millions of US dollars in the same period relative to the
measure of liquidity. Market capitalization is from World Federation of Exchanges (2006–2008).
Hybrid exchange A dummy variable equal to one for exchanges with both floor trading and limit order book.
Log (volume) Log of total value of shares trading in millions of US dollars in the same period relative to the measure
of liquidity. Total value of share trading data is from World Federation of Exchanges (2006–2008).
Log (number of trades) Log of total number of trades in thousands in the same period relative to the measure of liquidity.
Numbers of trades are from World Federation of Exchanges (2006–2008).
Log (1+ MSCI) Log of one plus the MSCI index in the lagged period relative to the measure of liquidity
Log (GDP) Log of gross domestic product (GDP) per capita in the lagged period relative to the measure of liquidity
investors in 2004 would expect adoption of MAD at the capitalization.5 If the month-end spread is missing due
time of MiFID. Third, principles in MAD were added to market close, the nearest prior day with a non-missing
to/clarified in MiFID for the implementation and definition spread is used for the calculation.
of conduct to ensure that MAD was legally effective. Hence, Following Röell (1992), we define a hybrid exchange-
while the legal situation in Europe is not perfectly dummy variable to control for the impact of market
delineated over time, we felt that the market adoption of structure on liquidity. This dummy is set to equal one for
these principles was best reflected (and best perceived by exchanges with both floor trading and limit order book
the market to have been reflected) with the November 1, and zero otherwise. Hybrid markets have the potential to
2007 change. We expect that the substantial details enhance liquidity if market makers compete with limit
provided in MiFID enhanced investor protection and orders to execute market orders (O’Hara, 2001). But the
facilitated liquidity, as tested below. overall effect could depend on the composition of
investors on the exchange. Seppi (1997) argues that large
institutional investors and small retail investors prefer
3. The sample and summary statistics
hybrid markets, while mid-size investors prefer pure limit
order markets. The microstructure data considered are
In this section, we describe our data and present the
listed in Table 1 and are collected from various sources
summary statistics and the results of comparison studies.
including the exchange webpages, directories, and hand-
books, as well as the World Federation of Exchanges
3.1. The sample (2006–2008) and Yahoo Finance. We do not have data on
the composition of investors on each exchange, but we do
Our sample is made up of 42 stock exchanges that use exchange-dummy variables to control for unobser-
are members of the World Federation of Exchanges vable differences across exchanges.
(2006–2008) and whose trading data are included in Surveillance data are used from Cumming and Johan
commonly used data sources such as Thomson Reuters (2008). Cumming and Johan surveyed 25 exchanges
Datastream. The sample covers Argentina, Australia, around the world to ascertain the extent of single- and
Austria, Bermuda, Brazil, Canada, China (Shanghai and cross-market surveillance. The data were obtained
Shenzhen), Chile, Colombia, Egypt, France, Germany, confidentially for the period 2004–2005 because a
Greece, Hong Kong, India (Bombay and the National Stock would-be manipulator could trade in ways that could
Exchange of India), Indonesia, Ireland, Israel, Italy, Japan, not be detected if precise information about surveillance
Jordon, Korea, Malaysia, Mexico, New Zealand, Norway, activity was available.6 The data are based on an equally
Peru, Philippines, OMX (Sweden, Finland, Denmark), weighted index that adds one every time a different type
Singapore, Slovenia, Spain, Sri Lanka, Switzerland, Taiwan, of single- and cross-market manipulation is monitored.
Thailand, Turkey, the United Arab Emirates, the UK, and We use the principal components of domestic- and cross-
the US (Nasdaq and NYSE). Trading rules for these stock market surveillance to mitigate collinearity associated
exchanges are found on the each exchange’s webpage, with other microstructure and country-specific variables
with the sole exception of China, where the pertinent of interest.
trading rules for the Shanghai and Shenzhen exchange are We also acquire a series of law and finance indices
found on the China Securities and Regulatory Commission from La Porta, Lopez-de-Silanes, Shleifer, and Vishny
webpage. (1998) and La Porta, Lopez-de-Silanes, and Shleifer
The definitions of the variables used in the analyses are (2006) and Spamann (2010), which include efficiency of
provided in Table 1. From the World Federation of the judiciary, anti-director rights, and liability standards.7
Exchanges (2006–2008), we use the annualized monthly Other legal indices were considered, but they did not
trading velocity values for February 2006–October 2008, impact the empirical tests reported below and are
the period considered by this study. The domestic market
capitalization at the end of each month, monthly total
value of share trading, and data for the total number of 5
The percentage quoted bid-ask spread is winsorized at the 5% level
trades for each stock exchange are also obtained from the due to the existence of negative quoted spreads in Datastream (on select
World Federation of Exchanges. Volatility and bid-ask days due to timing of reported values). For consistency, volatility and
spreads for each exchange are based on the stocks that velocity are likewise winsorized at the 5% level. Our analyses are robust
to winsorizing at the 1% level and robust to not winsorizing.
make up the exchange’s main index and are obtained 6
Comparable data for the 2006–2008 time period were not
from Thomson Reuters Datastream. Firm volatility is the forthcoming from the exchanges in our sample, and as such this
annualized monthly volatility calculated using the prior surveillance data represent a proxy for the monitoring activity of the
60 months of returns. Volatility of each marketplace is the 2006–2008 time period in which we have trading rule data and changes
over time from MiFID.
month-end firm capitalization weighted average of 7
Where these variables are not defined for a particular country, we
the firm volatilities. Following Chordia, Roll, and Sub- use the average value of the legal origin family for that particular
rahmanyam (2002), we compute the value-weighted country. As a robustness check in the empirical tests provided below, we
quoted spread for each exchange. The percentage quoted exclude these countries from the data and find the results do not
bid-ask spread of each is the difference between the ask materially change. Further, our results pertaining to the trading rule
indices are invariant to use of the La Porta, Lopez-de-Silanes, Shleifer,
and bid price divided by the midpoint of the bid and and Vishny (1998) and La Porta, Lopez-de-Silanes, and Shleifer (2006)
ask price, which is then averaged to calculate the indices versus the Spamann (2010) index for anti-director rights and
spread of the exchange, weighted by firm market other updates, as well as other variables from Jackson and Roe (2009).
therefore excluded for conciseness.8 Although we do have Table 3 indicates the summary statistics for all of the
information on surveillance, we do not have data on country-years in the data. The average monthly velocity in
enforcement of the trading rules that we analyze in this the sample is 95.1%, and the median is 70.9%. The range is
article. Nevertheless, our understanding from our data 11.5–218.3%. One standard deviation in velocity is 63.2%.
sources for surveillance in Cumming and Johan (2008) is The average annualized market capitalization weighted
that enforcement is highly correlated with surveillance volatility is 31.9%, with a range between 21.1% and 48.5%.
because otherwise exchanges would not bother to carry The average bid-ask spread is 0.75%, with a minimum of
out surveillance. To further proxy enforcement, we use 0.10% and a maximum of 7.4%. The number of non-
prior indices of enforcement such as efficiency of the missing country-month observations for velocity is 1,363
judiciary. In other work, note that La Porta, Lopez-de- due to the coverage by the World Federation of
Silanes, and Shleifer (2006) finds evidence that private Exchanges.9 The number of country-month observations
enforcement facilitates the development of stock markets, with non-missing volatility data is 1,319 and the number
while Jackson and Roe (2009) find stronger evidence on of non-missing spread data is 988 due to Datastream
the value of liability standards and public enforcement coverage.10 Table 3 also provides summary statistics for
(see also Roe, 2006; Jackson, 2007). Unlike La Porta, the legal indices and MSCI and GDP per capita.
Lopez-de-Silanes, and Shleifer (2006), Jackson and Roe Table 4 provides a comparison of mean and median
(2009) employ detailed resource-based measures such as tests of volume in relation to different cut-off values,
budgets per GDP and staffing per population to study which are the median value of the legal indices. Panel A
enforcement. These enforcement measures differ signifi- reports differences in means and medians of velocity for
cantly across countries, but not over time. We consider all the full sample of all country-years in the data. The data
of the indices in the La Porta, Lopez-de-Silanes, and indicate velocity is significantly greater for higher values
Shleifer (2006) and Jackson and Roe (2009). Inclusion or in the Insider Trading Rules Index. The average (median)
exclusion of these indices does not materially affect the velocity is 111.9% (72.2%) for exchanges with three or
conclusions regarding our trading rule indices. But more insider trading rules and is 80.8% (55.7%) for
inclusion of time-invariant indices in our multivariate exchanges with values of zero or one in the Insider
regressions limits the number of exchange-dummy vari- Trading Rules Index. Similar results are observed for
ables that can be included to control for unobservable differences in the Market Manipulation Rules Index. The
differences across countries. We report regressions with average (median) velocity is 113.4% (63.6%) for exchanges
exchange-dummy variables. with five or more market manipulation rules and is 68.9%
To control for the influence of market-specific changes, (54.5%) for exchanges with four or fewer market manip-
we draw from a series MSCI Global Standard Index from ulation rules. These differences in means and medians are
Morgan Stanley Capital International’s webpage. Also, we statistically significant at the 1% level. Likewise, the
include year-dummy variables in our multivariate analyses. subcomponents of the Market Manipulation Rules Index
show statistically significant differences at the 1% level for
price manipulation, volume manipulation, spoofing, and
3.2. Summary statistics
false disclosure. We compare the results with an Investor
Protection Index, which is the principal component of
Table 2 provides summary statistics of the trading rule anti-director rights, disclosure requirements, and liability
variables collected for this paper. Three primary legal standards (La Porta, Lopez-de-Silanes, and Shleifer, 2006;
indices are introduced: the Insider Trading Rules Index, see also Spamann, 2010). Unlike the results for our Insider
the Market Manipulation Rules Index, and the Broker–Agency Trading Rules Index and market manipulation indices, no
Conflict Rules Index. The Market Manipulation Rules significant differences emerge for the Investor Protection
Index consists of four subcomponents: the Price Manip- Index either in terms of means or medians. Nevertheless,
ulation Rules Index, the Volume Manipulation Rules the La Porta, Lopez-de-Silanes, and Shleifer (2006)
Index, the Spoofing Manipulation Rules Index, and Investor Protection Index does show significant differ-
the False Disclosure Rules Index. These indices are ences at the 1% level for values of three or more versus
summarized in Table 2 for the pre- and post-MiFID values of two or less. This latter result indicates that the
periods for 2006–2008. The indices are created by trading rules indices in this paper might be correlated
summing up the number of specific provisions in the with other legal differences across countries, such as the
exchange trading rules in each country. In the post-MiFID Investor Protection Index, and hence in our empirical
period the Insider Trading Rules Index varies from a assessment of trading rules on trading velocity we control
low value of zero (for a number of exchanges listed in for other legal and economic differences across countries.
Table 2) to ten (for Nasdaq). The Market Manipulation
Rules Index varies from a low value of zero (for Chile,
Peru, Philippines, and Turkey) to 13 (for London, NYSE,
Euronext Paris, and Slovenia). The Broker–Agency Conflict 9
Velocity data from the World Federation of Exchanges are
Rules Index varies from a low value of zero (for a number available monthly from February 2006 for most countries except Jordan
of exchanges listed in Table 2) to five (for Nasdaq). where data start in January 2008.
10
For example, Datastream does not cover the bid and ask price for
Argentina, Bermuda, China, Chile, Columbia, India (both Bombay and the
8
See, e.g., supra note 2 and accompanying text. Extra details and National Stock Exchange of India), Israel and Peru. The bid and ask price
empirical tests with additional indices are available upon request. data types for the Canadian market begin from November 27, 2006.
660
Table 2
Trading rule indices.
This table summarizes the index values for the trading rules for each exchange, as defined in Table 1. Values are presented for post-Directives on Market in Financial Instruments (November 2007–October
2008; in brackets are values for January 2006–October 2007). The *, ** and *** denote significance at the 10%, 5%, and 1% level, respectively.
Exchange Price Manipulation Volume Spoofing Index False Disclosure Market Insider Trading Broker–Agency
Index Manipulation Index Index Manipulation Index Index Index
D. Cumming et al. / Journal of Financial Economics 99 (2011) 651–671

English legal origin
Australia 3 (3) 1 (1) 2 (2) 0 (0) 6 (6) 2 (2) 0 (0)
Bermuda 2 (2) 1 (1) 1 (1) 1 (1) 5 (5) 2 (2) 2 (2)
Bombay 0 (0) 1 (1) 1 (1) 1 (1) 3 (3) 2 (2) 3 (3)
Canada 7 (7) 2 (2) 3 (3) 0 (0) 12 (12) 2 (2) 1 (1)
Hong Kong 3 (3) 2 (2) 1 (1) 1 (1) 7 (7) 0 (0) 0 (0)
India NSE 3 (3) 1 (1) 1 (1) 1 (1) 6 (6) 3 (3) 3 (3)
Ireland 7 (1) 1 (0) 3 (1) 1 (0) 12 (2) 2 (0) 0 (0)
Israel 2 (2) 0 (0) 0 (0) 1 (1) 3 (3) 1 (1) 0 (0)
London 7 (6) 2 (2) 3 (3) 1 (1) 13 (12) 3 (2) 0 (0)
Malaysia 0 (0) 0 (0) 1 (1) 1 (1) 2 (2) 7 (7) 2 (2)
Nasdaq 5 (5) 1 (1) 3 (3) 2 (2) 11 (11) 10 (10) 5 (5)
New Zealand 2 (2) 0 (0) 1 (1) 1 (1) 4 (4) 3 (3) 3 (3)
NYSE 6 (6) 2 (2) 3 (3) 2 (2) 13 (13) 7 (7) 3 (3)
Singapore 3 (3) 1 (1) 2 (2) 1 (1) 7 (7) 2 (2) 2 (2)
Sri Lanka 2 (2) 1 (1) 0 (0) 1 (1) 4 (4) 4 (4) 2 (2)
Thailand 4 (4) 2 (2) 1 (1) 1 (1) 8 (8) 1 (1) 0 (0)
Average English 3.50 (3.06) 1.13 (1.06) 1.63 (1.50) 1.00 (0.94) 7.25 (6.56) 3.19 (3.00) 1.63 (1.63)
legal origin
French legal origin

Argentina 2 (2) 0 (0) 1 (1) 0 (0) 3 (3) 3 (3) 1 (1)
Brazil 0 (0) 0 (0) 1 (1) 0 (0) 1 (1) 1 (1) 0 (0)
Chile 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)
Columbia 2 (2) 0 (0) 0 (0) 0 (0) 2 (2) 0 (0) 0 (0)
Egypt 2 (2) 0 (0) 0 (0) 0 (0) 2 (2) 0 (0) 0 (0)
France 7 (3) 1 (0) 3 (1) 2 (1) 13 (5) 2 (0) 0 (0)
Greece 7 (1) 1 (0) 3 (1) 1 (1) 12 (3) 3 (2) 0 (0)
Indonesia 1 (1) 0 (0) 1 (1) 1 (1) 3 (3) 2 (2) 1 (1)
Italy 7 (0) 1 (0) 3 (1) 1 (1) 12 (2) 3 (1) 0 (0)
Jordan 1 (1) 0 (0) 1 (1) 0 (0) 2 (2) 5 (5) 3 (3)
Mexico 3 (3) 1 (1) 1 (1) 1 (1) 6 (6) 2 (2) 0 (0)
Peru 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)
Philippines 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)
Spain 7 (0) 1 (0) 3 (1) 1 (1) 12 (2) 4 (4) 0 (0)
Turkey 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)
Average French 2.60 (1.00) 0.33 (0.07) 1.13 (0.60) 0.47 (0.40) 4.53 (2.07) 1.67 (1.33) 0.33 (0.33)
legal origin
German legal origin
Austria 7 (1) 1 (0) 3 (0) 1 (0) 12 (1) 2 (0) 0 (0)
Germany 7 (0) 1 (0) 3 (1) 1 (0) 12 (1) 3 (2) 0 (1)
Korea 4 (4) 2 (2) 2 (2) 1 (1) 9 (9) 3 (3) 2 (2)
Shanghai 2 (2) 1 (1) 1 (1) 1 (1) 5 (5) 2 (2) 0 (0)
Shenzhen 2 (2) 1 (1) 1 (1) 1 (1) 5 (5) 2 (2) 0 (0)
Slovenia 7 (5) 1 (1) 3 (1) 2 (1) 13 (8) 3 (2) 0 (0)
Switzerland 7 (2) 1 (1) 3 (1) 1 (1) 12 (5) 3 (2) 1 (1)
Taiwan 2 (2) 0 (0) 0 (0) 0 (0) 2 (2) 0 (0) 0 (0)
Tokyo 1 (1) 0 (0) 1 (1) 0 (0) 2 (2) 1 (1) 0 (0)
Average German 4.33 (2.11) 0.89 (0.67) 1.89 (0.89) 0.89 (0.56) 8.00 (4.22) 2.11 (1.56) 0.33 (0.44)
legal origin
Scandinavian legal origin

OMX 7 (2) 1 (1) 3 (2) 1 (1) 12 (6) 5 (4) 2 (2)
Oslo 7 (2) 1 (1) 3 (1) 1 (0) 12 (4) 4 (3) 0 (0)

Average 7.00 (2.00) 1.00 (1.00) 3.00 (1.50) 1.00 (0.50) 12.00 (5.00) 4.50 (3.50) 1.00 (1.00)
Scandinavian
legal origin
Tests of means
English versus 0.05 2.58** 0.23 1.94* 0.68 1.58 2.98***
civil law ( 12.96***) ( 15.19***) ( 12.19***) ( 12.83***) ( 15.91***) ( 8.98***) ( 13.59***)
English versus 0.91 0.92 0.54 0.62 2.59** 1.56 1.40

French (16.34***) (23.06***) (14.70***) (13.09***) (20.53***) (9.75***) (13.89***)
English versus 0.83 0.27 0.28 0.13 0.71 1.11 1.40

German (6.24***) (6.04***) (8.92***) (7.83***) (8.30***) (8.94***) (12.33***)
English versus 3.70*** 0.15 1.57 0.00 4.77* 1.34 0.62

Scandinavian (9.65***) (1.53) (0.00) (4.85***) ( 3.02***) (6.23***) (3.55***)
French versus 1.71* 0.64 0.81 0.48 3.23*** 0.47 0.00

German ( 9.08***) ( 11.85***) ( 5.81***) ( 3.43***) ( 1.97**) ( 9.58***) ( 1.66*)
French versus 4.53*** 0.82 2.09** 0.64 7.27*** 2.95*** 0.67

Scandinavian ( 16.18***) ( 66.30***) ( 11.13***) ( 1.24) ( 18.30***) ( 15.92***) ( 4.11***)
German versus 2.70*** 0.13 1.22 0.13 3.84*** 2.51** 0.66

Scandinavian (1.05) ( 6.86***) ( 6.43***) (0.65) ( 18.58***) ( 3.04***) ( 3.39***)
661
Table 3
Descriptive statistics.
This table presents statistics for the full sample of country-month observations in the data. The data span the months from February 2006 to October
2008, and the exchanges are listed in Table 2. The full number of country-months in the data is 1,386, but for some variables there are missing data.
Velocity data from the World Federation of Exchanges are not available for Jordan prior to 2008. Bid-ask-spread data from Datastream are missing for
Argentina, Athens, Bermuda, Chile, Columbia, India Bombay, India National Stock Exchange of India, Israel, and Peru. Volatility data are missing for
selected months for Athens, Bermuda, and Bombay, India. Surveillance data are available for select countries from Cumming and Johan (2008) as
indicated in Table 1. World Federation of Exchanges has incomplete data on the number of trades for Bermuda, Italy, Japan, Jordan, and Singapore for
select months and similarly for a few observations for market capitalization and volume.
Variable Mean Median Standard deviation Minimum Maximum Number of observations
Velocity 0.910 0.709 0.632 0.115 2.183 1,363

Volatility 0.319 0.295 0.083 0.211 0.485 1,319
Bid-ask spread 0.007 0.004 0.007 0.001 0.074 988
Insider Trading Rules Index 2.249 2 2.109 0 10 1,386
Market Manipulation Rules Index 5.212 4 4.024 0 13 1,386
Price Manipulation Rules Index 2.600 2 2.259 0 7 1,386
Volume Manipulation Rules Index 0.680 1 0.710 0 2 1,386
Spoofing Rules Index 1.238 1 1.011 0 3 1,386
False Disclosure Rules Index 0.695 1 0.585 0 2 1,386
Broker–Agency Rules Index 0.872 0 1.260 0 5 1,386
Surveillance 24.867 23.307 11.378 6.659 42.994 759
Efficiency of the judiciary 7.815 8 1.994 2.5 10 1,386
Investor Protection Index 2.321 2.272 0.838 0.686 3.775 1,386
Log (market capitalization) 12.819 12.821 1.843 7.619 16.625 1,361
Hybrid exchange 0.359 0 0.480 0 1 1,386
Log (volume) 9.974 10.146 2.720 0.000 15.168 1,361
Log (number of trades) 7.416 7.616 2.722 1.897 13.278 1,271
Log (1 +MSCI) 0.001 0.009 0.056 0.371 0.138 1,386
Log (GDP) 9.487 10.149 1.324 6.565 11.304 1,386
Panel A of Table 4 also considers differences in the B that use MiFID country data support our central
indices for the subset of European exchanges for which hypothesis for the Market Manipulation Rules Index and
MiFID applies. The results are broadly consistent with those all of its subcomponent indices, but not the Insider
reported in elsewhere in Panel A, with a few exceptions. Trading Rules, Broker–Agency Conflict Rules, and Investor
The mean and median values of velocity for the Market Protection Indices. Panel B shows that average and
Manipulation Rules Index are not significantly different, as median volatility is smaller after November 2007 for
are the mean and median values for price. The velocity is MiFID exchanges and those differences are significant at
significantly lower for exchanges with a higher value in the the 1% level and that the non-MiFID exchanges observe a
False Disclosure Rules Index, the opposite of that of the slight increase in volatility but this change is insignificant.
whole sample. Nevertheless, the differences in means and Again, these findings support our central hypothesis.
medians for the Broker–Agency Conflict Rules Index and the Table 4, Panel C provides comparison results for the
Investor Protection Index are statistically significant. bid-ask spread. The data indicate that the Insider Trading
Panel C of Table 4 compares velocity for the pre- and Rules Index, the Market Manipulation Rules Index, and its
post-MiFID time periods (pre-November 2007 and post- subcomponent indices are associated with a smaller bid-
November 2007) for both the exchanges affected by MiFID ask spread for the full sample. The subset of exchanges
(‘MiFID exchanges’) and exchanges not affected by MiFID affected by MiFID provides inconsistent results. Only a
(‘non-MiFID exchanges’) in the data. The data indicate higher Volume Manipulation Rules Index is associated
that for the MiFID exchanges both average and median with a lower bid-ask spread. Panel C also shows the
velocity were significantly higher after MiFID (average of comparison of pre- and post-bid-ask spreads. It indicates
126.7% and median of 137.3%) than before MiFID (average that no significant change exist in spreads for MiFID
of 113.7% and median of 127.6%). For the non-MiFID exchanges, but a significant increase is evident in spreads
exchanges, both average and median velocities were not after November 2007 for non-MiFID exchanges.
statistically different pre- and post-November 2007 Taken together, these statistics show there is a
(averages were 78.0% and 82.7%, and medians were material effect of MiFID on liquidity measures, which is
56.3% and 67.9%, respectively). consistent with our hypothesis that MiFID has a positive
Table 4, Panel B presents the differences in means and impact on market liquidity. The evidence of changes in
medians of volatility for the full sample and the liquidity in the post-MiFID period also indicates the need
subsample in which MiFID applies. The results are to assess difference-in-differences tests in the econo-
generally consistent with our hypotheses. Both mean metric tests. As well, the differences over time suggest a
and median tests of the full sample show that market need to control for market factors in assessing the
capitalization weighted volatility is significantly lower for determinants of liquidity.
exchanges with higher exchange rule indices, except for Table 5 presents a correlation matrix for the main
the Broker–Agency Conflict Rules Index. The tests in Panel variables used in the multivariate tests provided in
Table 4
Comparison tests.
This table presents the comparison of mean and median tests for velocity (Panel A), volatility (Panel B), and the bid-ask spread (Panel C) for different cut-off values of the indices defined in Table 1. Each Panel
considers all exchanges in the data set, the subset of exchanges for which the Directive on Markets in Financial Instrument (MiFID) applies, and Part 3 considers pre- versus post-MiFID for the subsample of
MiFID and non-MiFID exchanges. The *, ** and *** denote significance at the 10%, 5%, and 1% level, respectively.
Panel A: Comparison tests for velocity
Statistics Insider Market Price Volume Spoofing Index False Broker Investor
Trading Index Manipulation Index Manipulation Index Manipulation Index Disclosure Index Agency Index Protection Index
42 r2 44 r4 42 r2 41 r1 41 r1 41 r1 40 r0 42.27 r 2.27
All countries
Number of observations 445 918 678 685 537 826 198 1,165 384 979 90 1,273 526 837 660 703

Mean 1.119 0.808 1.134 0.689 1.105 0.783 1.216 0.858 1.331 0.745 1.614 0.860 0.864 0.939 0.896 0.923
Standard deviation 0.722 0.557 0.636 0.545 0.598 0.622 0.501 0.638 0.557 0.582 0.660 0.600 0.661 0.612 0.607 0.655
Median 1.354 0.648 1.019 0.495 0.930 0.530 1.124 0.612 1.365 0.552 1.623 0.668 0.678 0.739 0.754 0.664
Difference in means 8.01*** 13.86*** 9.50*** 8.90*** 16.93*** 11.44*** 2.11** 0.76
Difference in medians p o0.00*** po 0.00*** p o0.00*** p o 0.00*** p o0.00*** po 0.00*** p o 0.00 ***
p = 0.93
Statistics
Subset of MiFID exchanges

Number of observations 171 225 249 147 207 189 33 363 186 210 24 372 87 309 132 264
Mean 1.425 1.002 1.169 1.209 1.173 1.197 1.498 1.156 1.299 1.083 0.770 1.211 1.366 1.133 1.354 1.100
Median 1.449 0.869 1.319 1.448 1.323 1.354 1.532 1.313 1.365 1.182 0.752 1.336 1.352 1.292 1.457 1.210
Difference in means 8.09*** 0.57 0.57 8.68*** 3.93*** 3.83*** 4.84*** 5.84***
Difference in medians p o0.00*** p = 0.98 p =0.98 p o 0.00*** p o 0.0*** po 0.00*** p= 0.02** p o 0.00***
Statistics Non-MiFID countries MiFID Countries
Post-MiFID Pre-MiFID Post-MiFID Pre-MiFID
Pre-MiFID versus Post-MiFID

Number of observations 358 609 144 252
Mean 0.827 0.780 1.267 1.137
Standard deviation 0.644 0.618 0.58 0.537
Median 0.679 0.563 1.373 1.276
Difference in means 1.13 2.24**
Difference in medians p= 0.329 p = 0.016**
Panel B:Comparison tests for volatility
Insider Market Price Volume False Broker Investor

Spoofing Index
Trading Index Manipulation Index Manipulation Index Manipulation Index Disclosure Index Agency Index Protection Index
Statistics
42 r2 44 r4 42 r2 41 r1 41 r1 41 r1 40 o =0 42.27 o = 2.27
All countries
Mean 0.302 0.328 0.300 0.337 0.281 0.344 0.296 0.323 0.275 0.336 0.240 0.325 0.318 0.319 0.283 0.355
663
Median 0.282 0.296 0.271 0.338 0.260 0.342 0.259 0.301 0.250 0.331 0.239 0.307 0.321 0.293 0.261 0.366
Difference in means 5.38*** 8.35*** 15.49*** 4.26*** 14.08*** 29.59*** 0.23 17.45***
Table 4. (continued )
664
Difference in medians p o 0.00*** p o 0.00*** p o 0.00*** p o 0.00*** p o 0.00*** po 0.00*** p o 0.00*** p= 0.28
Mean 0.286 0.284 0.272 0.309 0.256 0.318 0.226 0.291 0.277 0.292 0.241 0.288 0.319 0.274 0.296 0.278
Median 0.261 0.273 0.256 0.284 0.240 0.287 0.226 0.276 0.247 0.276 0.240 0.274 0.296 0.258 0.288 0.271
Difference in means 0.30 10.91*** 10.91*** 15.31*** 2.28** 9.99*** 5.19*** 7.36***
Difference in medians p =0.26 p o 0.00*** p o 0.00*** p o 0.00*** p o 0.00*** po 0.00*** p o 0.00*** p o0.00***
Non-MiFID Countries MiFID countries

Statistics
Pre-MiFID versus post-MiFID

Mean 0.337 0.329 0.261 0.298
Median 0.32 0.32 0.245 0.276
Difference in means 1.24 5.98***
Difference in medians p= 0.446 p o =0.000***
Panel C: Bid-ask spread
Insider Market Price Volume Spoofing False Broker Investor

Trading Index Manipulation Index Manipulation Index Manipulation Index Index Disclosure Index Agency Index Protection Index
Statistics
42 r2 44 r4 42 r2 41 r1 41 r1 41 r1 40 o =0 42.27 o = 2.27
All countries
Mean 5.648 5.446 5.635 5.406 5.715 5.337 5.909 5.423 5.841 5.338 6.544 5.419 5.454 5.544 5.465 5.557
Median 5.572 5.368 5.603 5.301 5.841 5.256 6.066 5.264 5.936 5.186 6.834 5.334 5.446 5.471 5.330 5.497
Difference in means 2.91*** 3.51*** 5.87*** 5.86*** 7.54*** 9.97*** 1.32 1.41
Difference in medians p o0.00*** p o 0.00*** p o 0.00*** p o0.00*** p o 0.00*** p o0.00*** p =0.28 p = 0.14
Mean 5.858 6.026 5.808 6.113 6.008 5.895 6.699 5.879 5.934 5.972 5.884 5.959 5.924 5.964 5.748 6.072
Median 5.547 6.372 5.603 6.483 6.380 5.683 6.679 5.685 6.055 6.311 6.060 6.207 5.600 -6.368 5.771 6.276
Difference in means 1.77* 1.67* 1.67* 10.28*** 0.39 0.25 0.43 5.52***
Difference in medians p = 0.02*** p = 0.10* p o 0.00*** p o0.00*** p = 0.67 p = 0.80 p =0.86 p o 0.00***
Non-MiFID countries MiFID countries

Statistics
Pre-MiFID versus post-MiFID

Mean 5.119 5.356 5.927 5.969
Median 4.972 5.037 6.066 6.281
Difference in means 2.89*** 0.41
Difference in medians p = 0.005*** p = 0.953
Table 5
Correlation matrix.
This table presents Pearson correlation coefficients for the full sample of country-months in the data. Correlations in absolute value greater than 0.05,
0.06 and 0.07 are statistically significant at the 10%, 5% and 1% level, respectively.
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14)
(1) Velocity 1.00

(2) Volatility 0.04 1.00
(3) Bid-ask spread 0.51 0.22 1.00
(4) Insider trading rules index 0.40 0.29 0.13 1.00
(5) Market Manipulation Rules 0.37 0.36 0.24 0.49 1.00
Index
(6) Price Manipulation Rules Index 0.29 0.36 0.19 0.35 0.96 1.00
(7) Volume Manipulation Rules 0.26 0.12 0.21 0.30 0.79 0.69 1.00
Index
(8) Spoofing Rules Index 0.44 0.38 0.28 0.57 0.91 0.83 0.63 1.00
(9) False Disclosure Rules Index 0.33 0.27 0.20 0.63 0.63 0.48 0.50 0.53 1.00
(10) Broker–Agency Rules Index 0.15 0.09 0.04 0.73 0.24 0.10 0.22 0.29 0.49 1.00
(11) Surveillance 0.11 0.29 0.10 0.35 0.39 0.37 0.31 0.34 0.33 0.28 1.00
(12) Efficiency of the Judiciary 0.15 0.53 0.24 0.33 0.37 0.35 0.23 0.43 0.21 0.29 0.18 1.00
(13) Investor Protection Rules Index 0.04 0.28 0.13 0.34 0.24 0.15 0.33 0.27 0.18 0.45 0.13 0.44 1.00
(14) Log (1+ MSCI) 0.01 0.01 0.10 0.04 0.13 0.15 0.03 0.12 0.05 0.00 0.01 0.02 0.01 1.00
(15) Log (GDP) 0.39 0.54 0.44 0.18 0.44 0.44 0.22 0.52 0.21 0.07 0.23 0.65 0.11 0.04
Section 4. As predicted, a strong positive correlation exists specifications. The first two models in each panel present
between trading velocity and the Insider Trading Rules the difference-in-differences regressions. The difference-in-
Index (0.33), the Market Manipulation Rules Index (0.38), differences regression assesses the impact of MiFID by using
its subcomponent indices, and the Broker–Agency Conflict a control group to subtract other changes at the same time
Rules Index, all of which are statistically significant at the of MiFID, assuming these other changes were identical
1% level. GDP per capita is likewise highly positively between the treatment group (the European countries that
correlated with trading velocity (0.39). Volatility is also were subject to MiFID) and the control group (the other
highly correlated with trading volume, number of trades, countries). We run the following regression:
and market capitalization. Similarly, volatility and bid-ask
Y ¼ b0 þ b1 Treat þ b2 After þ b3 TreatAfter þ bi Controlsi þ e
spread are significantly negatively correlated with the
ð1Þ
trading rule indices, which is consistent with our central
hypothesis. We add an indicator variable (After) that is equal to one
in November 2007 and every month thereafter and zero in
all prior months. We also include an indicator variable
4. Multivariate analyses (Treat) that is set to one for exchanges subject to MiFID.
The interaction variable labeled Treat*After is the key
In this section, we consider whether MiFID has an variable in our experiment. Under the null hypothesis that
impact on market liquidity and which legal factors are MiFID encourages trading activity, and reduces market
most closely associated with cross-sectional differences in volatility and bid-ask spread, we expect the coefficient
liquidity, while controlling for other economic and institu- associated with this variable to be statistically positive in
tional determinants of trading activity. We consider each Panel A but statistically negative in Panels B and C.
exchange-month from February 2006 to October 2008 as a In Model 1 of Panel A, the coefficient associated with
separate observation (1,363 observations), with considera- the interaction variable is 0.087, significant at the 1% level,
tion to differences in rules over time as well as difference- which means that, after November 2007, MiFID exchanges
in-differences regressions, exchange-dummy variables, and experienced an incremental increase of 8.7% in velocity
exchange fixed effects. We cluster standard errors by compared with non-MiFID exchanges. The effect is
exchange (as in Petersen, 2009). Two-way clustering by estimated at 4.7% in Model 2 of Panel A, when control
year and exchange did not materially affect the reported variables for surveillance and enforcement are added.
results. Also, we considered other approaches for treating Similar results are observed in Panels B and C. After MiFID
standard errors for panel data sets (e.g., Bertrand, Duflo, is put into force, MiFID exchanges see a decrease of 4.1%
and Mullainathan, 2004), which we found to be robust. For (Model 1) to 3.8% (Model 2) in volatility, significant at the
each regression, we control for economic factors including 1% and 5% level, respectively. Panel C shows that MiFID
market capitalization, market conditions (MSCI Index), decreases bid-ask spreads in Europe by a significant 6
GDP, and exchange institutional features (Röell, 1992), as (Model 2) to 20 (Model 1) basis points. The evidence is
well as exchange and year dummy variables. robust to exchange-dummy variables for each exchange
Panels A–C of Table 6 examine the effect of MiFID and to pick up other exchange-specific factors that can
exchange trading rules on velocity, volatility, and bid-ask influence market liquidity. As well, the results are robust
spread, respectively. In each of the three panels we present to controls for changes in economic conditions over time
10 identical regressions to show robustness to alternative in the MSCI Index, differences in market capitalization,
666
Table 6
Regression analysis of market liquidity. This table presents Ordinary Least Square panel regressions of the determinants of market liquidity in the cross-section across countries. Variables are as defined in
Table 1. Standard errors are clustered by country (Petersen, 2009). Panels A–C present regressions for the velocity, volatility, and bid-ask spread, used as the dependent variables, respectively. The first two
regressions in each panel use difference in differences estimates for the effect of the Directives on Markets in Financial Instruments (November 2007). Model 3–Model 10 in each panel presents a regression
with the different exchange rules indices, and the final column presents a regression with multiple indices simultaneously. The *, ** and *** are statistically significant at the 10%, 5% and 1% level, respectively.
Panel A: Velocity
Independent variables Model 1: Model 2: Model 3: Insider Model 4: Market Models 5: Broker– Models 6:
Difference-in-differences Difference-in-differences Trading Rules Index Manipulation Rules Index Agency Rules Index Indices Jointly
Coefficient t-statistic Coefficient t-statistic Coefficient t-statistic Coefficient t-statistic Coefficient t-statistic Coefficient t-statistic
* ***
Constant 3.264 1.837 2.360 3.106 3.283 1.381 1.859 1.024 2.570 1.542 39.946 5.986***
Difference-in-differences variables

Treat 0.115 2.867*** 4.382 5.155***
After 0.019 0.858 0.055 2.719***
Treat*After 0.087 1.798* 0.047 2.475**
Trading rules
Insider Trading Rules Index 0.069 1.577 0.028 0.483
Market Manipulation Rules Index 0.013 1.847*
Price Rules Index 0.002 0.221
Volume Rules Index 0.236 3.074***
Broker–Agency Rules Index 0.226 6.917***
Enforcement
Surveillance 0.012 2.315** 0.015 1.019 0.058 5.069*** 0.017 1.230 0.046 5.775***
Efficiency of the Judiciary 0.650 3.697*** 0.300 0.695 0.154 2.133** 0.202 3.325*** 0.266 2.490**
Investor Protection Index 5.416 7.150***
Microstructure control variables
Log (market capitalization) 0.162 2.220** 0.096 3.105*** 0.236 11.643 0.215 2.646*** 0.205 2.569** 0.206 12.386***
Hybrid exchange 0.034 0.915 0.350 1.964** 0.473 0.992 0.018 0.351 0.034 0.726 0.030 0.799
Market condition control variables
Ln(MSCI) 0.177 2.648*** 0.001 0.008 0.185 2.019 0.150 2.198** 0.158 2.333** 0.161 1.948**
Dummy 2007 0.018 0.712 0.004 0.266 0.078 2.073 0.019 0.476 0.026 0.677 0.019 0.529
Dummy 2008 0.094 2.190** 0.002 0.049 0.091 1.053 0.120 1.369 0.141 1.778* 0.119 1.446
Country control variables
Log(GDP) 0.437 3.150*** 0.010 0.101 0.373 1.289 0.550 2.879*** 0.551 3.038*** 0.509 2.521**
Exchange dummy variables Yes Yes Yes Yes Yes Yes
Model diagnostics
Number of observations 1,361 757 757 757 757 757
Adjusted-R2 0.971 0.959 0.955 0.963 0.963 0.964
Panel B: Volatility
Independent variables Model 1: Model 2: Model 3: Insider Model 4: Market Manipulation Models 5: Broker– Models 6:
Difference-in-differences Difference-in-differences Trading Rules Index Rules Index Agency Rules Index Indices Jointly
Constant 0.026 0.082 0.716 2.692*** 0.622 1.571 0.589 1.505 1.070 1.855* 0.566 0.163
Treat 0.048 2.593*** 0.110 0.158
After 0.006 0.832 0.016 1.400
Treat*After 0.041 3.113*** 0.038 2.155**
Trading rules
Insider Trading Rules Index 0.028 2.874*** 0.010 0.842
Market Manipulation Rules Index 0.006 3.708***
Price Rules Index 0.015 2.238**
Volume Rules Index 0.024 0.585
Broker–Agency Rules Index 0.039 1.440
Enforcement
Surveillance 0.004 0.869 9.042E 04 0.325 0.003 1.128 0.007 1.757* 0.004 1.040
Efficiency of the Judiciary 0.010 0.066 0.013 1.400 0.021 3.267*** 0.062 2.773*** 0.031 0.688
Investor Protection Index 0.004 0.010
Log (market capitalization) 0.034 1.857* 0.064 3.139*** 0.061 2.897*** 0.061 3.003*** 0.065 3.255*** 0.061 3.082***
Hybrid exchange 0.023 5.766*** 0.039 5.295*** 0.038 1.014 0.039 1.041 0.019 0.474 0.037 5.417***

Ln(MSCI) 0.081 3.594*** 0.103 4.833*** 0.089 4.763*** 0.093 5.246*** 0.086 4.252*** 0.098 5.512***
Dummy 2007 0.018 3.905*** 0.033 3.994*** 0.034 5.977*** 0.034 6.069*** 0.043 6.390*** 0.034 4.004***
Dummy 2008 0.002 0.111 0.017 0.861 0.032 3.420 0.029 3.272*** 0.055 4.313*** 0.029 1.597
Log(GDP) 0.014 0.297 0.038 0.416 0.039 1.377 0.033 1.161 0.098 1.668* 0.037 0.398
Exchange dummy variables Yes Yes Yes Yes Yes Yes
Model diagnostics
Number of observations 1,294 757 757 757 757 757
2
Adjusted-R 0.875 0.851 0.846 0.851 0.838 0.856
Independent variables Model 1: Model 2: Model 3: Insider Model 4: Market Manipulation Models 5: Broker– Models 6:
Difference-in- Difference-in- Trading Rules Index Rules Index Agency Rules Index Indices Jointly
differences differences
Constant 1.312E 02 0.400 8.431E 03 2.062 7.135E 03 0.385 5.332E 03 0.305 5.440E 03 0.272 1.707E 02 0.916
Treat 1.563E 02 2.790*** 1.678E 02 0.432
After 2.049E 03 3.213*** 1.131E 03 2.765
Treat*After 2.045E 03 1.959* 6.019E 04 1.667*
Trading rules
Insider Trading Rules Index 3.140E 04 2.360** 4.771E 04 2.694***
Market Manipulation Rules Index 6.937E 04 0.160
Price Rules Index 2.108E 05 0.205
Volume Rules Index 3.840E 04 1.262
Broker–Agency Rules Index 3.597E 04 1.095
Enforcement
** ***
Surveillance 4.236E 04 1.299 2.149E 04 2.471 2.063E 04 2.585 2.330E 04 2.878*** 3.135E 04 2.300**
Efficiency of the Judiciary 4.315E 03 0.503 4.438E 04 0.874 8.536E 04 2.220** 4.639E 04 0.584 3.571E 04 0.674
Investor Protection Index 2.246E 03 1.231
Log (market capitalization) 3.370E 03 2.968*** 1.880E 03 2.567** 9.087E 04 4.097*** 6.628E 04 2.700*** 8.262E 04 3.044*** 1.386E 03 3.080***
Hybrid exchange 2.281E 03 0.411 5.086E 03 12.702*** 6.884E 03 2.322** 6.115E 03 2.201** 7.116E 03 2.592*** 6.677E 03 2.560**
Log (number of trades) 3.811E 04 0.319 3.887E 04 1.434 1.664E 04 0.708 4.861E 04 2.749*** 2.355E 04 0.651 4.241E 04 0.916
Volatility 5.053E 03 0.871 2.535E 03 1.157 5.399E 04 0.185 2.840E 04 0.077 3.791E 04 0.115 1.346E 03 0.374
667
market structure, and GDP, among other variables
t-statistic
0.530
0.900
considered, but not included for conciseness. The results
0.984
1.502
provide strong support for our hypothesis that the rules
Indices Jointly
Models 6:
introduced by MiFID help to enhance market liquidity.
0.801
615
Yes
1.518E 03
2.007E 04
In Models 3–5 in each panel, we assess the impact of
3.665E 04
3.272E 03
Coefficient
each trading rule index on market liquidity measures
separately. The liquidity measure is regressed on each
index separately along with a group of control variables.
Model 6 considers a number of different indices jointly in
t-statistic
1.058
1.003
0.085
1.107
the same regression. We do not include every trading rule
Agency Rules Index
Models 5: Broker–
index jointly in the same regression due to collinearity;

rather, we only include the Insider Trading, Price, and
0.801
615
Yes
1.743E 03
4.328E 04
3.553E 05
2.961E 03
Volume Manipulation Rules Indices in the same regression.

Coefficient
Model 3 reports the result of the regression on the

Insider Trading Rules Index. Consistent with our hypoth-
esis, trading velocity is 0.069 higher when there is an
increase in the Insider Trading Rules Index by one (Model
Model 4: Market Manipulation
t-statistic
3, Panel A), but this effect is marginally insignificant.

1.486
0.926
1.498
0.050
Likewise, volatility is 2.8% lower when the Insider Trading

Rules Index increases by one (Model 3, Panel B), and this
Rules Index
effect is significant at the 1% level. Panel C, Model 3

0.800
615
Yes
indicates that a one point increase in the Insider Trading

1.631E 03
5.235E 04
1.917E 05
3.231E 03
Coefficient
Rules Index value is associated with a 3 basis point

decrease in the bid-ask spread, and this effect is
significant at the 5% level. One likely explanation for this
result is that stricter rules discourage insider trading,
which decreases the informational content of spread as
t-statistic
1.069
1.024
well as the spread itself.

0.129
1.400
Trading Rules Index
Model 3: Insider
In Panel A, Model 4 the Market Manipulation Rules

Index is positively associated with velocity, and this effect
0.801
615
Yes
1.851E 03
3.939E 04
is significant at the 10% level in Model 4. In separate

5.267E 05
3.227E 03
Coefficient
regressions (not reported), we examine separately the

Volume Manipulation Rules Index and the Price Manip-
ulation Rules Index. The effect of the Volume Manipula-
tion Rules Index shows greater economic significance
t-statistic
0.444
0.927
1.090
0.938
(0.210) relative to that of the Market Manipulation Rules

Difference-in-
Index (0.013) and the Price Manipulation Rules Index

differences
Model 2:
(0.021), but all effects are statistically significant at least

0.804
615
Yes
6.799E 04
3.484E 04
8.999E 04
5.262E 03
the 10% level. When these variables are combined

Coefficient
together in Model 6, the Volume Manipulation Rules

Index is statistically significant at the 1% level and the
other trading rule indices are insignificant. The Spoofing
and False Disclosure Rules Indices (in models not
t-statistic
0.519
1.796*
0.356
0.381
explicitly reported) are statistically insignificant, and the

Difference-in-
Broker–Agency Conflict Rules Index in Model 5 is

differences
Model 1:
0.728
negative and significant. This latter result is counter-

921
Yes
1.558E 03
2.153E 04
5.614E 04
4.559E 03
Coefficient
intuitive. Regardless, it is insignificant in more parsimo-

nious specifications that were considered.
The Broker–Agency Conflict Rules Index is statistically
insignificant in Panels B and C. One likely explanation
for this result is that traders or brokers are bound by
rules of conduct formulated by professional associations
Exchange dummy variables
that vary from country to country, span different

Number of observations
countries, and have different affiliations within each

Independent variables
country. A second explanation is that brokers play less

Model diagnostics
Table 6 (continued )
of a role as a financial intermediary offering advice in

Dummy 2007
Dummy 2008
some countries, such as China; we consider this possibi-

Adjusted-R2
lity by excluding such countries and find the results to be

Log(GDP)
Ln(MSCI)
very similar. Importantly, this finding highlights the

fact that the significance of the Insider Trading Rules
Index and the Market Manipulation Rules Index (and its
subcomponents) is not merely a spurious indication that A number of our additional control variables for market
detailed rule drafters are more likely to reside in certain microstructure, market conditions, and other country
countries, as the details regarding broker–agency conflict factors are statistically significant. For example, velocity
are not statistically related to proxies for liquidity. and volatility are greater for our various proxies for
In Panel B, Models 3–5, volatility is negatively exchange size, and bid-ask spreads are smaller. However,
associated with the Market Manipulation Rules but not when we add or subtract other microstructure variables
with the Broker–Agency Conflict Rules. Even in separate (e.g., such as those in Panel C) in any of the Models 16 and
regressions with all of the Market Manipulation Rules country or market variables, we do not find different
considered separately (not explicitly reported), all coeffi- results pertaining to our findings regarding trading rules.
cients are significant at the 1% level except the False In summary, Table 6 indicates that a variety of
Disclosure Rules Index, which is significant at 10% level. specifications that include exchange -dummy variables,
The results are stronger for the Volume Manipulation exchange fixed effects, exchange institutional features,
Rules Index (4.3% economic significance). But in Model 6 and surveillance applications, among other things, show
when indices are combined in the same regression, the that trading rules facilitate trading velocity, reduce
Price Manipulation Rules Index is the one that remains market volatility, and reduce trading cost across
statistically significant and shows a 1.5% drop in volatility exchanges and over time.12 In general, MiFID is an
for a one point increase in the index. important factor in terms of improving market activities
In Panel C, Models 3–6 show that, among the different as evidenced by our difference-in-differences regressions.
indices, the Insider Trading Rules Index is most strongly Considering different indices in the same regression, the
statistically and negatively related to the bid-ask spread. Volume Manipulation Rules Index is most closely con-
An increase in the Insider Trading Rules Index by one nected to trading velocity, the Price Manipulation Rules
point reduces the bid-ask spread by 3.14 basis points in index is most closely connected to volatility, and the
Model 3 and 4.77 basis points in Model 6 and these effects Insider Trading Rules Index is most closely connected to
are significant at the 5% and 1% levels, respectively. bid-ask spreads. Overall, insider trading rules and market
In all models, except Model 1, we add the surveillance manipulation rules are important for facilitating trading
variable to control for the potential influence of the activity and decreasing volatility.
detection of illegal-trading behaviors on the inferences (as
in Cumming and Johan, 2008). In addition, we include a
variable for the Efficiency of the Judiciary as a proxy for 5. Conclusions
enforcement. Further, in Model 6, we add the Investor
Protection Index (defined in Table 1). The latter two In this paper, we contribute to the literature on
variables are more closely related to the enforcement of international differences in stock exchanges by examining
corporate governance rules against corporate self-dealing the effect of trading rules on liquidity as represented by
and the expense of outside shareholders and, as such, are velocity, volatility, and bid-ask spread. Building on prior
imperfect proxies for the enforcement of secondary work on mandatory disclosure and delegation between
trading rules. We consider other variables from La Porta, private and public enforcement of securities laws (La
Lopez-de-Silanes, and Shleifer (2006) and Jackson and Roe Porta, Lopez-de-Silanes, and Shleifer, 2006; Hail and Leuz,
(2009), but they do not materially impact our analyses 2006; Jackson and Roe, 2009) and surveillance technology
pertaining to the trading rules variables of interest. That and information sharing in ex post enforcement (Cum-
the inclusion of these control variables in the regression ming and Johan, 2008), we consider the interaction
has no effect on the statistical significance of the Treat*- between rule specificity in stock exchange trading rules
After interaction term, although the magnitude of the and stock exchange trading activity.
changes is reduced in Panels A and C and slightly reduced
in Panel B; see Model 1 versus Model 2. The coefficient
(footnote continued)
of surveillance variable generally shows a significant
Inclusion of this variable influences the coefficient estimates for the
improvement in liquidity, consistent with Cumming and Investor Protection Index, but does not influence the other variables of
Johan (2008). The negative and significant coefficient for interest in our analyses. Similarly, the inclusion or exclusion Criminal
the efficiency of the judiciary supports La Porta, Lopez-de- Enforcement Index (La Porta, Lopez-de-Silanes, and Shleifer, 2006) and
Silanes, Shleifer, and Vishny (1998) and La Porta, Lopez- additional variables from Jackson and Roe (2009) are immaterial to our
inferences drawn from the trading rules indices.
de-Silanes, and Shleifer (2006). The Investor Protection 12
In addition to our cross-sectional and time series specifications in
Index coefficient estimate in Model 6 in Panel A is Table 6, in a prior version of this article we reported cross-sectional
sensitive to the other included variables. For example, if regressions (42 observations) following nearly identical methods as in La
we exclude the Efficiency of the Judiciary variable then Porta, Lopez-de-Silanes, and Shleifer (2006). Also, we consider the use of
instruments as in La Porta, Lopez-de-Silanes, and Shleifer (2006)
the Investor Protection Index is positive and significant.
(including legal origin variables, as well as the repudiation index
Regardless, the addition or subtraction of these or other reported in La Porta, Lopez-de-Silanes, Shleifer, and Vishny, 1998).
variables does not materially impact our reported results Further, we present partial regression plots for some of the trading rule
for the trading rule indices.11 indices. Those results are available on request and also consistent with
the findings reported here. As well, we consider other robustness checks
not presented herein, such as excluding various exchanges from the
analyses, two-step regression methods to account for missing observa-
11
The Public Enforcement Index (La Porta, Lopez-de-Silanes, and tions, and subsamples with different time periods. Again, those checks
Shleifer, 2006) has a 0.72 correlation with the Investor Protection Index. show results that are consistent with the findings presented herein.
We employ a sample of 42 exchanges around the Clayton, M.J., Jorgensen, B.N., Kavajecz, K.A., 2006. On the presence and
world and find that stock exchange trading activity is market-structure of exchanges around the world. Journal of
Financial Markets 9, 27–48.
most closely related to trading rules specificity in regard Comerton-Forde, C., Rydge, J., 2006. Market integrity and surveillance
to insider trading and market manipulation but is not effort. Journal of Financial Services Research 29, 149–172.
statistically related to rules pertaining to broker–agency Comerton-Forde, C., Tang, K., 2007. Anonymity, front-running and
market integrity. Journal of Trading 4, 101–118.
conflict. The reasoning behind this finding is that insider Cumming, D., Johan, S.A., 2008. Global market surveillance. American
trading and market manipulation rules provide clarity Law and Economics Review 10, 454–506.
regarding prohibited manipulative trading practices and Daouk, H., Lee, C.M.C., Ng, D.T., 2006. Capital market governance: how do
security laws affect market performance? Journal of Corporate
are of direct and central importance to the conduct of
Finance 12, 560–593
market participants. By contrast, broker–agency conflict Daske, H., Hail, L., Leuz, C., Verdi, R.S., 2008. Mandatory IFRS reporting
rules are typically subject to extraneous rules from around the world: early evidence on the economic consequences.
governing bodies and professional associations for brokers Journal of Accounting Research 46, 1085–1142.
DeMarzo, P.M., Fishman, M.J., Hagerty, K.M., 2005. Self-regulation and
(such as the Chartered Financial Analysts Institute). The government oversight. Review of Economic Studies 72 (3), 687–706.
connection between trading activity and insider trading Ferguson, M.F., Peters, S.R., 2003. But I know it when I see it: an
and market manipulation rules is robust to concerns economic analysis of vague rules. Unpublished working paper.
University of Cincinnati and Kansas State University, Cincinnati,
about endogeneity, difference-in-differences specifica-
OH, and Manhattan, KS.
tions, and alternative control variables. Specifically, we Gerard, B., Nanda, V., 1993. Trading and manipulation around seasoned
observe the material impact of the MiFID rule changes on equity offerings. Journal of Finance 48, 213–245.
all dimensions of liquidity. Although it is difficult to Hail, L., Leuz, C., 2006. International differences in the cost of equity
capital: do legal institutions and securities regulation matter?
isolate precisely the components of trading rules that Journal of Accounting Research 44, 485–531
matter the most, it is noteworthy that we do observe a Harris, F.H.D.E.B., Aitken, M.A., Cook, R.M., McInish, T.H., 2009. Market
close connection between the Volume Manipulation Rules design and execution costs for matched securities worldwide.
Institutional Investor Journal 1, 38–76.
Index and trading velocity, the Price Manipulation Rules Harris, L., 2002. Trading and Exchanges: Market Microstructure for
Index and volatility, and the Insider Trading Rules Index Practitioners. Oxford University Press, Oxford, UK.
and bid-ask spreads. The results indicate trading rules are Hillion, P., Suominen, M., 2004. The manipulation of closing prices.
Journal of Financial Markets 7, 351–375.
an important source of information to consider in
Jackson, H.E., 2007. Variation in the intensity of financial regulation:
explaining the differences in trading activity among stock preliminary evidence and potential implications. Yale Journal on
exchanges around the world. Future work could look to Regulation 24, 253–291.
the exchange trading rules as a source of international Jackson, H.E., Roe, M.J., 2009. Public and private enforcement of
securities laws: resource-based evidence. Journal of Financial
differences in stock exchange efficiency and integrity. Economics 93, 207–238.
Jarrow, R.A., 1992. Market manipulation, bubbles, corners and short
squeezes. Journal of Financial and Quantitative Analysis 27, 311–336.
References
Jarrow, R.A., 1994. Derivative security markets, market manipulation,
and option pricing theory. Journal of Financial and Quantitative
Aitken, M., Siow, A., 2003. Ranking equity markets on the basis of market Analysis 29, 241–261.
efficiency and integrity. In: Skeete, H. (Ed.), Hewlett-Packard Kumar, P., Seppi, D.J., 1992. Futures manipulation with cash settlement.
Handbook of World Stock, Derivative and Commodity Exchanges Journal of Finance 47, 1485–1502.
2003, pp. xliv–lv. La Porta, R., Lopez-de-Silanes, F., Shleifer, A., 1999. Corporate ownership
Aggarwal, R.K., 2001. Regulatory infrastructure covering financial around the world. Journal of Finance 54, 471–517.
markets. Brookings–Wharton Papers on Financial Services, 55–76. La Porta, R., Lopez-de-Silanes, F., Shleifer, A., 2006. What works in
Aggarwal, R.K., Wu, G., 2006. Stock market manipulations. Journal of securities laws? Journal of Finance 61, 1–32
Business 79, 1915–1953. La Porta, R., Lopez-de-Silanes, F., Shleifer, A., Vishny, R., 1997. Legal
Allen, F., Gale, D., 1992. Stock price manipulation. Review of Financial determinants of external finance. Journal of Finance 52, 1131–1150.
Studies 5, 503–529. La Porta, R., Lopez-de-Silanes, F., Shleifer, A., Vishny, R., 1998. Law and
Allen, F., Gorton, G., 1992. Stock price manipulation, market micro- finance. Journal of Political Economy 106, 1113–1155.
structure, and asymmetric information. European Economic Review La Porta, R., Lopez-de-Silanes, F., Shleifer, A., Vishny, R., 2002. Investor
36, 624–630. protection and corporate valuation. Journal of Finance 57, 1147–1170.
Amihud, Y., Mendelson, H., 1988. Liquidity and asset prices: financial Laffont, J.J., Tirole, J., 1993. A Theory of Incentives in Procurement and
management implications. Financial Management 17, 5–15. Regulation. MIT Press, Cambridge, MA.
Avgouleas, E., 2005. The Mechanics of Regulation of Market AbuseOxford Lampert, R., Leuz, C., Verracchia, R.E., 2007. Accounting information,
University Press, Oxford, UK. disclosure, and the cost of capital. Journal of Accounting Research 45,
Beny, L.N., 2005. Do insider trading laws matter? Some preliminary 385–420.
comparative evidence. American Law and Economics Review 7, Madhavan, A., Porter, D., Weaver, D., 2005. Should securities markets be
144–183. transparent? Journal of Financial Markets 8, 266–288.
Beny, L.N., 2007. Insider trading laws and stock markets around the Mahoney, P.G., 1999. The stock pools and the Securities Exchange Act.
world: an empirical contribution to the theoretical law and Journal of Financial Economics 51, 343–369.
economics debate. Journal of Corporation Law 32, 237–300. Massimb, M.N., Phelps, B.D., 1994. Electronic trading, market structure,
Bertrand, M., Duflo, E., Mullainathan, S., 2004. How much should we and liquidity. Financial Analysts Journal 50, 39–50.
trust difference in difference estimates? Quarterly Journal of Merrick Jr., J.J., Naik, N.Y., Yadav, P.K., 2005. Strategic trading behavior
Economics 119, 249–275 and price distortion in a manipulated market: anatomy of a squeeze.
Bhattacharya, U., Daouk, H., 2002. The world price of insider trading. Journal of Financial Economics 77, 171–218.
Journal of Finance 57, 75–108. Ni, S.X., Pearson, N.D., Poteshman, A.M., 2005. Stock price clustering on
Bhattacharya, U., Daouk, H., 2009. When no law is better than a good option expiration dates. Journal of Financial Economics 78, 49–87.
law. Review of Finance 13, 577–627. O’Hara, M., 2001. Overview: market structure issues in market liquidity.
Carson, J.W., 2003. Conflicts of interest in self-regulation: can demutua- In: Market Liquidity: Proceedings of a Workshop Held at the Bank for
lized exchanges successfully manage them? Unpublished working International Settlements, BIS Papers, no. 2, April, Basel, Switzerland,
paper 3183. World Bank Policy Research, Washington, DC. pp. 1–8.
Chordia, T., Roll, R., Subrahmanyam, A., 2002. Order imbalance, O’Hara, M., Mendiola, A.M., 2003. Taking stock in stock markets: the
liquidity, and market returns. Journal of Financial Economics 65, changing governance of exchanges. Unpublished working paper,
111–130. Cornell University, Ithaca, NY.
Pagano, M., Randl, O., Röell, A.A., Zechner, J., 2001. What makes stock Pritchard, A.C., 2003. Self-regulation and securities markets. Regulation
exchanges succeed? Evidence from cross-listing decisions. European 26 (1), 32–39.
Economic Review 45, 770–782. Reiffen, D., Robe, M.A., 2007. Ownership structure and enforcement
Pagano, M., Röell, A., Zechner, J., 2002. The geography of equity listing: incentives at self-regulatory financial exchanges. Unpublished work-
Why do companies list abroad? Journal of Finance 57, 2651–2694. ing paper. American University, Washington, DC.
Peng, L., Röell, A., 2009. Managerial incentives and stock price Roe, M.J., 2006. Legal origins, politics, and modern stock markets.
manipulation. Center for Economic and Policy Research (CEPR), Harvard Law Review 120, 460–527.
Washington, DC. Discussion Paper DP7442. Röell, A., 1992. Comparing the performance of stock exchange trading
Petersen, M.A., 2009. Estimating standard errors in finance panel data systems. In: Fingleton, J., Schoenmaker, D. (Eds.), The Internationa-
sets: comparing approaches. Review of Financial Studies 22, 435–480. lisation of Capital Markets and the Regulatory Response. Graham
Pirrong, S.C., 1993. Manipulation of the commodity futures market and Trotman, London, pp. 167–177.
delivery process. Journal of Business 15, 335–370. Roll, R., 1988. The international crash of 1987. Financial Analysts Journal
Pirrong, S.C., 1995a. Mixed manipulation strategies in commodity 44, 19–35.
futures markets. Journal of Futures Markets 15, 13–38. Romano, R., 1993. In: The Genius of American Corporate Law. AEI Press,
Pirrong, S.C., 1995b. The self-regulation of commodity exchanges: the Washington, DC.
case of market manipulation. Journal of Law and Economics 38, Romano, R., 2001. The need for competition in international securities
141–206. regulation. Theoretical Inquiries in Law 2, 1–179.
Pirrong, C., 1999. The organization of financial exchange markets: theory Romano, R., 2002. The Advantage of Competitive Federalism for
and evidence. Journal of Financial Markets 2, 329–357. Securities Regulation. AEI Press, Washington, DC.
Pirrong, S.C., 2004. Detecting manipulation in futures markets: the Seppi, D., 1997. Liquidity provision with limit orders and a strategic
Ferruzi soybean episode. American Law and Economics Review 6, specialist. Review of Financial Studies 10, 103–150.
28–71. Spamann, H., 2010. The ‘antidirector rights index’ revisited. Review of
Pistor, K., Keinan, Y., Kleinheisterkamp, J., West, M., 2003. The evolution Financial Studies 23, 467–486.
of corporate law: a cross-country comparison. Journal of Interna- Stevenson, D., 2005. Toward a new theory of notice and deterrence.
tional Economic Law 23, 791–871. Cardozo Law Review 26, 1535–1594.
Pistor, K., Xu, C., 2003. Incomplete law: a conceptual and analytical Stoll, H.R., 2000. Friction. Journal of Finance 55, 1479–1514.
framework and its application to the evolution of financial market Vitale, P., 2000. Speculative noise trading and manipulation in the
regulation. Journal of International Law and Politics 35, 931–1013. foreign exchange market. Journal of International Money and
Pistor, K., Xu, C., 2005. Governing stock markets in transition economies: Finance 19, 689–712.
lessons from China. American Law and Economics Review 7, World Federation of Exchanges, 2006–2008. Annual report and statistics.
184–210. /http://www.world-exchanges.org/S.

Journal of Financial Economics: Douglas Cumming, Sofia Johan, Dan Li

Uploaded by

Copyright:

Available Formats

Journal of Financial Economics: Douglas Cumming, Sofia Johan, Dan Li

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Journal of Financial Economics: Douglas Cumming, Sofia Johan, Dan Li

Uploaded by

Copyright:

Available Formats

Journal of Financial Economics 99 (2011) 651–671

Contents lists available at ScienceDirect

Journal of Financial Economics

Exchange trading rules and stock market liquidity$

2.1.2. Market Manipulation Rules Index

generally detrimental to client interests such as by 2.2. Testable hypotheses

Insider trading rules

D. Cumming et al. / Journal of Financial Economics 99 (2011) 651–671

French legal origin

Scandinavian legal origin

D. Cumming et al. / Journal of Financial Economics 99 (2011) 651–671

English versus 0.91 0.92 0.54 0.62 2.59** 1.56 1.40

English versus 0.83 0.27 0.28 0.13 0.71 1.11 1.40

English versus 3.70*** 0.15 1.57 0.00 4.77* 1.34 0.62

French versus 1.71* 0.64 0.81 0.48 3.23*** 0.47 0.00

French versus 4.53*** 0.82 2.09** 0.64 7.27*** 2.95*** 0.67

German versus 2.70*** 0.13 1.22 0.13 3.84*** 2.51** 0.66

Variable Mean Median Standard deviation Minimum Maximum Number of observations

Velocity 0.910 0.709 0.632 0.115 2.183 1,363

Panel A: Comparison tests for velocity

D. Cumming et al. / Journal of Financial Economics 99 (2011) 651–671

Subset of MiFID exchanges

Statistics Non-MiFID countries MiFID Countries

Post-MiFID Pre-MiFID Post-MiFID Pre-MiFID

Pre-MiFID versus Post-MiFID

Panel B:Comparison tests for volatility

Insider Market Price Volume False Broker Investor

Non-MiFID Countries MiFID countries

Pre-MiFID versus post-MiFID

D. Cumming et al. / Journal of Financial Economics 99 (2011) 651–671

Panel C: Bid-ask spread

Insider Market Price Volume Spooﬁng False Broker Investor

Non-MiFID countries MiFID countries

Pre-MiFID versus post-MiFID

(1) Velocity 1.00

D. Cumming et al. / Journal of Financial Economics 99 (2011) 651–671

Adjusted-R2 0.971 0.959 0.955 0.963 0.963 0.964

D. Cumming et al. / Journal of Financial Economics 99 (2011) 651–671

Panel C: Bid-ask spread

market structure, and GDP, among other variables

index jointly in the same regression due to collinearity;

Volume Manipulation Rules Indices in the same regression.

Model 3 reports the result of the regression on the

3, Panel A), but this effect is marginally insigniﬁcant.

Likewise, volatility is 2.8% lower when the Insider Trading

effect is signiﬁcant at the 1% level. Panel C, Model 3

indicates that a one point increase in the Insider Trading

Rules Index value is associated with a 3 basis point

well as the spread itself.

In Panel A, Model 4 the Market Manipulation Rules

is signiﬁcant at the 10% level in Model 4. In separate

regressions (not reported), we examine separately the

(0.210) relative to that of the Market Manipulation Rules

Index (0.013) and the Price Manipulation Rules Index

(0.021), but all effects are statistically signiﬁcant at least

the 10% level. When these variables are combined

together in Model 6, the Volume Manipulation Rules

explicitly reported) are statistically insigniﬁcant, and the

French versus 4.53* 0.82 2.09 0.64 7.27* 2.95* 0.67

German versus 2.70* 0.13 1.22 0.13 3.84* 2.51** 0.66