Trading Aggressiveness and its Implications For
Market Efficiency∗
Olga Lebedeva†
April 20, 2012
Abstract
This paper investigates the empirical relationship between an increase in trading aggressiveness after earnings announcements and the speed of price adjustment. An increase in trading aggressiveness allows for quicker price changes,
which can be beneficial, if the majority of aggressive orders are informative and
push the price in the direction of its new equilibrium level. However, abnormal
trading aggressiveness can also slow down the adjustment process, if aggressive
orders are mostly used by uninformed investors to trade on their heterogeneous
beliefs. In this case, quick price changes in different directions just increase intraday volatility and the probability of price overshooting. Empirical findings of this
study suggest that the latter negative effect overall dominates, and it is especially
harmful for illiquid stocks. Adjustment time of these stocks has even increased as
compared to the period, when aggressive orders were not available on the market.
JEL classifications: G14, G18, G19
Keywords: Trading Aggressiveness, Price Adjustment, Market Efficiency,
Regulation NMS, Sweep Order, Earnings Announcement
∗
I thank Alex Edmans, Roni Michaely, Ernst Maug, Miguel Ferreira, Pedro Santa-Clara, Amy
Edwards, Albert Kyle, Klaus Adam and Erik Theissen for insightful comments on this paper. I
am particularly grateful for prolific discussions to participants at research seminars in Wharton and
Mannheim as well as participants of 2011 FMA European Conference Doctoral Consortium and 2011
FMA Annual Meeting Doctoral Consortium.
†
University of Mannheim, 68131 Mannheim, Germany. E-mail: lebedeva@corporate-financemannheim.de. Tel: +49 621 181 2278.
1
1
Introduction
This paper analyzes the influence of abnormal trading aggressiveness on the speed of
price adjustment after earnings announcement releases. An investor is trading aggressively, if she prefers quicker execution of her limit order over a better execution price.
Such a situation is most likely to arise, when investors expect immediate changes in the
value of a stock and the speed of order execution is of primary importance. Two recent
examples of abnormal trading aggressiveness on the market are the Flash Crash day
(May 6, 2010), when the Dow Jones Industrial Average index (DJIA) dropped by more
than 1,000 points in less than one hour, and the release of erroneous information about
United Airlines bankruptcy on Bloomberg terminals on September 8, 2008. In both
of these events, traders switch to the most aggressive orders on the market as soon as
they realize that it is better for them to have their order executed immediately, even
at an inferior price.1 Waiting for execution at the best quoted price in such moments
is costly, since the best quote may change by a large amount within the next second.
Periods immediately after corporate information releases also belong to the category
of events, when it pays off to act quickly. New information makes investors revise
their beliefs, which leads to a subsequent increase in trading aggressiveness.2 Which
implications does abnormal trading aggressiveness have on the speed of price adjustment
after a corporate information release? Higher execution speed of an aggressive order
ensures that a larger portion of this order, as compared to a standard limit order, is
executed per one time interval. Thus, aggressive trading enables quicker price changes
over relatively short time intervals. Quicker changes are beneficial if aggressive trading is
informative and, thus, pushes stock’s price more quickly towards its new fundamental
value. In contrast, if aggressive trades are mostly submitted by uninformed traders,
who are equally likely to buy or to sell, quick price changes in different directions
just increase intraday volatility and the probability of price overshooting. An abnormal
increase in intraday volatility makes it then harder for a price to stabilize around its new
equilibrium value.3 Empirical results of this study show that the negative effect of an
1
As documented by Chakravarty et al (2011b) for the Flash Crash day and Lei and Li (2010) for
the false announcement of United Airlines bankruptcy.
2
For example, the recent study by Chakravarty et al (2011a) documents an abnormal increase in
trading aggressiveness on the day of an earnings announcement release.
3
Fleming and Remolona (1999) analyze a two-stage adjustment process in the U.S. Treasury market
upon arrival of macroeconomic announcement releases. They identify the first stage as an almost
immediate price reaction with a reduction in trading volume. The second “stabilization” stage lasts
for more than an hour with abnormal price volatility, trading volume and bid-ask spreads. Brooks,
Patel and Su (2003) examine the price adjustment process in U.S. equity markets after unanticipated
2
increased trading aggressiveness overall dominates, and abnormal trading aggressiveness
is especially harmful for stocks with low liquidity levels. Importantly, adjustment time
of illiquid stocks with abnormal trading aggressiveness is significantly longer than in the
period, when aggressive orders were not available. The negative effect gets, however,
reduced as aggressive trades become more informative and move stock’s price in the
correct direction.
Since the efficient market hypothesis of Fama (1965), theoretical papers, e.g. French
and Roll (1986), treat public information as such that is already incorporated into prices,
before any investor can trade on it. In contrast with the theory, prior empirical studies
document abnormal trading volumes and abnormal volatility after information releases,
indicating that price adjustment does not happen immediately.4 The main reason for
the prolonged price adjustment process is that investors have heterogeneous beliefs and
different rates of information processing, which induces trading among investors until
they reach the consensus.5
Prior empirical studies on the speed of price adjustment after information releases
are diverse. Patell and Wolfson (1984) is the pioneer study to examine the speed
of price adjustment after earnings announcements. They find that abnormal stock
returns dissipate within five to ten minutes immediately after an announcement release,
whereas intraday volatility and serial correlation in returns persist for several hours.
Jennings and Starks (1985) extend their study by additionally controlling for differences
in information content and show that the adjustment process lasts longer for larger
earnings surprises. Further studies about the speed of price adjustment after earnings
announcements investigate the effect of firm and report characteristics (Defeo (1986),
Damodaran (1993)), compare announcements made within and outside trading hours
(Francis, Pagach, and Stephan (1992)), and examine differences across different market
structures (Greene and Watts (1996), Masulis and Shivakumar (2002)).6
events. They find that the initial price reaction lasts around 20 minutes after an information release
and that prices experience some reverse changes over the following two hours.
4
Bamber, Barron, and Stevens (2011) provide an extensive literature review on trading volume
around corporate information releases.
5
Later theoretical studies incorporate heterogeneity of investors in their models of stock trading by
assuming that either investors have differences in interpreting an announcement (Holthausen and Verrecchia (1990), Harris and Raviv (1993)) or that they are differentially informed (Kim and Verrecchia
(1991), He and Wang (1995), Hong and Stein (1999)).
6
In addition, there are several studies that analyze the speed of price adjustment after different
types of announcements. Ederington and Lee (1993) look at price adjustment of interest rate and
foreign exchange futures after scheduled releases of macroeconomic news. Further studies examine
intraday stock price reaction to shifts in dividend policy (Gosnell, Keown, and Pinkerton (1996)),
releases of analyst’s opinions (Busse and Green (2002)) and unanticipated events (Brooks, Patel and
3
Whereas it is important to know how quickly price adjusts per se, it is even more
interesting to examine how investor trading process directly influences price adjustment.
There exists a vast literature that investigates investor trading around information
releases.7 Surprisingly, the overlap between this literature and the price adjustment
literature is relatively small. To the best of my knowledge, there exist only two studies,
which analyze in more detail how exactly the price adjusts to its new level. Woodruff
and Senchack (1988) find that stocks with large positive earnings surprises experience
quicker adjustment than stocks with large negative earnings surprises. They further
show that there occurs a large number of smaller trades after positive earnings surprises
and relatively few, but larger, trades after negative earnings surprises. However, they
do not establish the causal relationship between differences in trading process and the
speed of price adjustment. Ederington and Lee (1995) examine short-run dynamics
of price adjustment in interest rate and foreign exchange futures markets. The main
finding of their study is that prices adjust in a series of small price changes, and not in
few large price jumps, which also suggests that there is an intensive trading going on
immediately after an information release.
This paper extends the previous literature on implications of investor trading for
the price adjustment process. Specifically, it addresses the question how trading aggressiveness of investors influences the speed of price adjustment after an earnings
announcement release. Trading aggressiveness is measured as a proportion of the total
volume, executed through aggressive orders within a particular interval of time. To
differentiate aggressive orders from non-aggressive ones, I use a new order type, an intermarket sweep order (ISO), which represents the most aggressive trading instrument
on U.S. equity markets.8 If an order is marked as intermarket sweep, a trading venue
has to give this order an immediate execution, even if this leads to a trade-through of
the best quoted price.9 Since an ISO is marked as such at the time of its submission, I
can ex post observe investor preferences for the speed of order execution.
Su (2003), Coleman (2011)).
7
Lee (1992) is among the first to examine differences in clustering of small and large trades around
earnings announcements. Recent studies examine informativeness of institutional (Ali, Klasa, and Li
(2008)) and individual trades (Kaniel, Saar, and Titman (2008), Kaniel et al (2012)) around earnings
announcements. Also, Sarkar and Schwartz (2009) document a post-announcement increase in twosided trading, especially when news surprises are large.
8
An intermarket sweep order (ISO) was introduced with the implementation of the Regulation
National Market System (Reg NMS) by the Securities and Exchange Commission (SEC) in October
2007 to allow institutional investors to trade large blocks quickly.
9
Chakravarty et al (2010) provides an excellent overview of ISO characteristics and their use on
current financial markets.
4
Major findings of this paper are as follows. First, I show that ISO trades have
higher intraday price impact than non-ISO trades and the difference in price impact
between ISO and non-ISO trades is larger for illiquid stocks. This finding implies that
aggressive trades indeed contribute to larger price changes per time interval, especially
for illiquid stocks. The reason behind is that illiquid stocks have a thin order book
with a lower number of shares quoted at each price, and, therefore, it is even easier for
aggressive orders to move prices of these stocks. Further, I investigate intraday changes
in trading aggressiveness on an earnings announcement day and document a significant
15% increase in the proportion of ISO volume in the first fifteen minutes after an announcement release. Afterward, the proportion of volume traded with aggressive orders
steadily decreases, but it continues to deviate significantly from its base level until the
end of the trading day. Additional analysis reveals that the post-announcement jump in
trading aggressiveness can be explained by a significant increase in the proportion of ISO
sell volume after negative earnings surprises. This result suggests that investors trade
more aggressively when confronted with negative news. For positive news releases, ISO
trades are largely uninformative in the first two hours after an announcement release,
with large increases in the proportion of ISO volume in both trading directions.
Finally, this paper establishes the link between investor trading aggressiveness and
the speed of price adjustment after an earnings announcement release. The length
of the price adjustment period is defined as the number of 10-minute intervals from
an announcement release until the interval, in which intraday volatility of one-minute
midpoint returns is no longer abnormal.10 Results of difference-in-differences analysis
suggest that the relationship is monotonically increasing for liquid stocks and exhibits a
U-shape for illiquid stocks. With excessively low trading aggressiveness, price changes
of illiquid stocks are not sufficiently quick, which slows down the adjustment process.
For liquid stocks, this effect is not observed, since their price changes are relatively
quick due to a large number of investors trading these stocks.
Importantly, excessively high increases in trading aggressiveness slow down the price
adjustment process significantly: if the proportion of ISO volume increases by 100%
10
I prefer the volatility criterion over measuring abnormal returns or serial correlation in returns,
since it covers both stages of price adjustment, the initial price reaction and the subsequent stabilization
period. Note that this paper takes a short-run perspective on the adjustment process and assumes that
the adjustment process ends within two days after an information release. There is a separate literature
on the post-earnings announcement drift (PEAD), which documents that the full adjustment of prices
may take up to one year after an announcement release. See, for example, Ball and Brown (1968),
Bernard and Thomas (1989, 1990), Foster, Olsen, and Shevlin (1984), Garfinkel and Sokobin (2006).
The short-run perspective is the most suitable for this study, since abnormal trading aggressiveness is
only observed in short periods after an information release.
5
from its base level, adjustment time increases on average by fifty minutes for liquid
stocks and by almost two hours for illiquid stocks. Consistent with prior expectations,
this overall negative effect of aggressive trading is more pronounced after positive earnings surprises, when aggressive trading is largely uninformative. After negative earnings
surprises, when the majority of aggressive trades is submitted in the direction of the
earnings surprise, it is less pronounced and no longer statistically significant.
This paper contributes to the literature in several ways. Following the pioneering work of Chakravarty et al (2010), it sheds light upon the use and characteristics
of intermarket sweep orders (ISOs) on the current financial markets. In addition to
Chakravarty et al (2011a), who analyze changes in market breadth and daily trading
aggressiveness on an announcement day, I investigate intraday changes in the use of
aggressive orders. Further, I examine informativeness of ISO trades by testing whether
the proportion of ISO volume increases more in the direction of the earnings surprise
right after an announcement release.
This paper also contributes to the on-going debate on the efficiency of financial
markets. To the best of my knowledge, this is the first study to analyze how abnormal
trading aggressiveness influences the speed of price adjustment after an information
release. Prior empirical studies by Woodruff and Senchack (1988) and Ederington and
Lee (1995) also examine the relationship between trading process and price adjustment
after information releases, but their studies concentrate mainly on trade size and transaction frequency. The main focus of this study, however, is the effect of investor trading
aggressiveness, revealed in their preference for the speed of order execution, on the price
adjustment process.
The remaining part of the paper is organized as follows. Section 2 provides details
of the relevant institutional framework and develops main hypotheses of this study.
The construction of the data set is described in Section 3. Section 4 analyzes the
use and characteristics of aggressive orders in the base period and around earnings
announcements. Section 5 investigates how abnormal trading aggressiveness affects the
speed of price adjustment after an announcement release. Section 6 briefly concludes.
6
2
Institutional Background and Hypothesis Development
2.1
Overview of the Reg NMS and Intermarket Sweep Orders
On August 29, 2005, the Securities and Exchange Commission (SEC) adopted a new set
of rules to modernize US equity markets and to promote their efficiency. This set of rules
is known as the Regulation National Market System (the Reg NMS) and aims to improve
competition not only among individual trading venues, but also among individual orders
sent to different markets. Due to technical difficulties in implementation of several
changes required by this new regulation the full compliance with the Reg NMS was
first achieved in October 2007.11
The most important change, introduced by the Reg NMS, was the adoption of the
“Order Protection rule” (Rule 611), which requires execution of any incoming order at
the best available price. The best available price is defined as the lowest ask or the
highest bid price quoted over the previous one second among all equity trading venues
in the US. If a venue, to which a limit order is sent, does not currently quote the best
price, it has to re-route the order to the venue with the best price.12
This rule guarantees intermarket protection of orders against possible execution
at inferior prices (“trade-throughs”).13 Such a protection is needed for both liquidity
demanders and liquidity suppliers in the current markets with a high speed in quotation
changes. For a liquidity demander, it is important that his order is executed at the best
available price. The best bid and offer quotes may change over very short time intervals
between the submission of an order and its execution. Thus, it is almost impossible
to control ex-post whether an incoming order was executed at the deteriorated best
quoted price, or whether it simply obtained an inferior execution price. A liquidity
supplier is always interested in a faster execution of his order, since he is facing a
tradeoff between an expected profit from trading at the desired price and a potential
11
See Regulation NMS, SEC Release No. 34-51808.
Among other important changes is the so-called “Access Rule”, which requires transparent access
to price quotations on all trading venues and regulates maximum amount of fees for every quotation
access. “Subpenny Rule” requires the minimum tick size of $0.01, unless a stock price is below $1. The
major change to Market Data Rules and Plans is providing an opportunity for individual markets to
distribute their market data independently (in addition to providing it to Plans, such as, for example,
Consolidated Tape).
13
A trade-through occurs when the best available bid or the best available offer quotation is ignored,
that is “traded-through”. According to the Trade-Through Study of the SEC (SEC Release No. 3451808), around 7.9% of the total volume traded on NASDAQ and around 7.2% on NYSE was traded
through prior to the adoption of the Reg NMS.
12
7
risk of order non-execution. Regular trade-throughs of outstanding limit orders would
discourage liquidity suppliers and potentially lead to decreased market liquidity.
The Order Protection Rule cares mainly for interests of retail investors. The bestprice execution guarantee increases investor confidence and decreases retail investor’s
search costs for the market with the best available price. Further, protection of the
best-priced limit orders minimizes investor transaction costs, since the number of tradethroughs automatically declines. The reduction of investor transaction costs is also
beneficial for listed firms. With smaller transaction costs, investors will demand a
lower expected return on equity and the cost of capital for these firms will decrease.
Although appealing to retail traders with a long-term investment horizon, for whom
the speed of order execution is not of the highest priority, the “Order Protection Rule”
is less attractive for short-term traders and institutional traders with large blocks.
Imagine an institutional investor, or a dealer representing this investor, who would like
to sell 3,200 shares at a price not lower than $10.67. For simplicity, suppose there exist
only two trading venues: A and B. Figure 1 shows a current bid side of a limit order
book for each of two venues. The first column shows currently quoted bid prices and
the second column indicates a number of shares available at each price.14 Trading venue
A currently quotes the highest bid price at $10.75. Suppose an investor submits her
order to A. However, A’s depth at the best available quote is too small to get this order
fully executed. In this example, only 500 shares can be sold at the best price. Without
trading venue B, the next 2,000 shares of the order would get executed at $10.70 and
the remaining 700 shares at $10.67. However, it is quite seldom that a stock is traded
on a single trading venue or that a single trading venue continuously quotes the best
price. In the current example, venue B quotes the next best bid price at $10.73. Under
the “Order Protection Rule” the outstanding part of the order (2,700 shares) has to be
re-routed by venue A to venue B. After an execution of 500 shares at $10.73 on venue
B, the remaining part (2,200 shares) has to be re-routed to A again. Unfortunately,
re-routing takes time and the best bid offer may change while the order is being rerouted. Thus, the execution of large-sized orders under the “Order Protection Rule”
takes longer time and might end up executing at an inferior average price, as compared
to the case, when the whole order had been executed at a single venue.
14
Note that total depth at each price is equal to the sum of a number of shares quoted at this
price and a cumulative number of shares quoted above this price for the bid side of the book (or,
equivalently, below this price for the ask side of the book). Thus, total depth for P = $10.70 on a
trading venue A equals 2,500 shares (500 shares quoted at P = $10.75 and additional 2,000 shares
quoted at P = $10.70).
8
Figure 1: Bid Side of Limit Order Book
Venue A
Price
Shares
$10.75
$10.70
$10.67
500
2,000
3,500
Venue B
Price
Shares
$10.73
$10.66
500
3,000
To avoid such situations, the “Order Protection Rule” makes an exemption for a
specific order type, an intermarket sweep order (ISO).15 An ISO is a marketable limit
order (Immediate-or-Cancel) and it provides an opportunity for institutional investors
to trade large block positions quickly. Specifically, when an intermarket sweep order
arrives at a particular trading venue, it is executed as if this venue were a single one.
An intermarket sweep order simply “walks down” a limit order book until either the
order is completely filled or the limit price of the order is reached (the outstanding
part of an ISO is then canceled). Importantly, no re-routing is required, even if some
parts of the order are executed at inferior prices as compared to the best national bid
offer. However, this opportunity of quick execution comes at an additional cost. To
comply with principles of the “Order Protection Rule”, an investor submitting an ISO
order is obliged to send additional limit orders, also marked as ISO, with the same limit
price to all other venues quoting the stock. The size of these additional ISOs should
equal to the number of shares, quoted at prices better than the limit price at the time
of submission of an ISO order. Therefore, an ISO order represents not a single order,
but rather a series of marketable limit orders with the same limit price sent across all
trading venues quoting the stock.16
To give an example, suppose that an institutional investor again would like to sell
3,200 shares at the limit price of $10.67 on venue A, but she chooses an ISO to get
the quickest execution. According to the new regulation, in addition to the limit order
routed to A, she has to submit simultaneously an additional ISO with the same limit
price for all shares quoted at prices better than $10.67 on venue B. Currently, superior
price quotations include 500 shares at $10.73. Thus, she sends an additional ISO for 500
15
Paragraphs (b) 5 and (b) 6 of Rule 611, Regulation NMS, SEC Release No. 34-51808.
Paragraph (b)(30) of Rule 600 gives a formal definition of an intermarket sweep order as a limit
order, which satisfies the following requirements: (1) when routed to a trading venue, the limit order
is identified as an intermarket sweep order; and (2) simultaneously with the routing of the limit order
identified as an intermarket sweep order, one or more additional limit orders, as necessary, are routed
to execute against the full displayed size of all protected quotations with a superior price.
16
9
shares at a limit price of $10.67 to venue B. Since trading venues can recognize both
orders as intermarket sweep, they instantaneously execute them against outstanding
orders up to a limit price of $10.67. If current bid quotes do not change, an investor
will instantly sell 3,200 shares on venue A and additional 500 shares on venue B. The
new best price will drop to $10.67 on venue A. If some better priced quotations are no
longer available by the time of order execution, then these orders (or their outstanding
parts) get automatically canceled, and an investor has still fulfilled her obligations with
respect to the “Order Protection Rule”.
In principle, an intermarket sweep order does not differ from a normal limit order,
if there is only one trading venue. Additional benefit of quicker execution of these
orders is only apparent in presence of several trading venues. Note that in addition
to explicit costs of submitting additional orders and monitoring all venues for better
priced quotations, there is also an implicit inventory cost associated with intermarket
sweep orders. The obligation to extract all shares available at better prices from other
venues makes it hard to control the total size of an ISO order. In the previous example,
an investor ended up instantly selling 3,700 shares in both venues. Thus, the only
instrument investors can use to control, at least approximately, the total size of an ISO
order is adjusting its limit price. Upon an execution of an ISO order, an investor will
sometimes get more shares and sometimes less as compared to what she is actually
willing to trade, which results in additional inventory risks. Since ISOs can only be
used by professional investors, these additional inventory costs should not, however,
play a major role.
In this paper I measure trading aggressiveness of a stock as the proportion of volume
traded with ISO orders over a particular time interval.17 ISOs represent the most
aggressive orders available on the current U.S. equity market. With their ability to
sweep liquidity almost instantly up to a particular price level, ISOs produce a higher
price impact per one infinitely small trading interval, as compared to the standard
limit order. This higher price impact per trading interval is a result of an overall
quicker execution of an intermarket sweep order: if both a standard limit order and
an aggressive limit order of an identical size are submitted in t, then a larger part of
the aggressive order, as compared to the standard order, will be executed in t + 1. To
illustrate this point, consider the previous example. Figure 2 summarizes the number
17
Chakravarty et al (2010) find that the use of ISO orders in the post-Reg NMS world is non-trivial
and amounts up to 46% of all trades or 41% of trading volume in their sample.
10
Figure 2: Price Impact per Time Interval
Limit order of 3,200 to A
ISO of 3,200 to A
t
Action
Shares
executed
Average
Price
Best
Price
after
Price
Impact
Shares
executed
Average
Price
Best
Price
after
Price
Impact
1
Execution
500
$10.75
$10.73
$0.02
3,200
$10.701
$10.67
$0.08
2
Re-routing
3
Execution
500
$10.73
$10.70
$0.03
4
Re-routing
5
Execution
2,200
$10.69
$10.67
$0.03
Total
3,200
$10.705
3,200
$10.701
$0.08
$0.08
of shares executed and the price impact of both orders in each trading interval.
Note that in t=1 the price impact of the standard order equals only $0.02, with 500
shares executed at the best available price, whereas the price impact of the aggressive
order, $0.08, is fully realized, since the total size of the order (3,200 shares) is immediately executed. If the limit order book did not change over time, the cumulative price
impact of both orders would be the same after t=5. The standard limit order just takes
longer time to execute because re-routing between two different exchanges is needed
to get execution at the best price. Importantly, additional ISO orders sent to other
exchanges to execute against all shares quoted at better prices (in this example, an
additional ISO order of 500 shares sent to venue B) will not change the aggregate price
impact, since all simultaneously sent ISO orders have the same limit price ($10.67 in
our example). As expected, the average execution price of the standard order ($10.705)
is slightly better than that of the aggressive order ($10.701). The reason for the lower
average execution price of the aggressive order in this example is the trade-through of
500 shares quoted at $10.70 on venue B.18 19
18
500 shares at $10.70 are, in fact, executed with the second ISO order sent to venue B, as required
by the regulation. The total average execution price of both ISO orders (3,700 shares executed) is still
slightly worse than the average execution price of the standard limit order (3,200 shares).
19
In the perfect world, where the limit order book would not change between the time of the order
submission and the time of its execution, institutional traders could perfectly split the total desired size
of their order between different exchanges. In the previous example, a trader, wishing to sell exactly
3,200 at the limit price of $10.67 could send an ISO order for 2,700 shares to venue A and an additional
ISO order for 500 shares to venue B. He would then get a total of 3,200 shares at the average price
of $10.705, which equals the average execution price of the standard order. Still, the price impact of
$0.08 would be fully realized at t=1, as two orders are instantly executed at both venues.
11
2.2
Hypothesis Development
Implications of trading with aggressive orders for the speed of price adjustment. Prior empirical studies document an increase in trading aggressiveness following
company’s information releases.20 Which implications does it have on the speed of price
adjustment towards its new fundamental value? The major difference between aggressive intermarket sweep orders and standard limit orders is quicker execution of ISOs,
which leads to a larger price change per one trading interval. Therefore, the higher
proportion of aggressive orders in the order flow will enable quicker price movements
within relatively short time intervals.
Quicker price movements are beneficial for the price adjustment if the majority of
traders are “informed” in the following sense: they have already correctly processed new
information and know the true fundamental value of a stock. They will then purchase
the stock if it is undervalued or sell the stock if it is overvalued, pushing stock’s price
towards its new fundamental value. In this case, an increase in trading aggressiveness
may speed up price adjustment due to a quicker movement of the price in the correct
direction.
However, quicker price movements may also slow down the convergence process,
if the majority of aggressive traders are “uninformed”, in the sense that they do not
observe the true fundamental value of a stock and can only form their subjective beliefs about it. Some “uninformed” investors will purchase the stock and push the price
temporary upwards, whereas the other “uninformed” will sell the stock and push the
price downwards. As the stock price continuously experiences quick upward changes,
followed by quick downward changes, it is constantly over- and undershooting its true
fundamental value. Thus, aggressive trading by “uninformed” investors with heterogeneous beliefs produces unnecessary abnormal volatility and makes it harder for a price
to stabilize on its new level.
Overall, the positive effect of quicker price movements towards the new fundamental
value should dominate in situations with the higher proportion of informed traders,
whereas the negative effect of increased intraday volatility should dominate when the
majority of aggressive traders are uninformed and have heterogeneous beliefs about the
true value of a stock.
20
Chakravarty et al (2011a) report an increase in the proportion and volume of ISO orders after
earnings announcements; Lei and Li (2010) document an increased use of ISO orders after an erroneous
information on bankruptcy announcement of United Airlines on September 8, 2008.
12
Liquid vs illiquid stocks. Does the influence of trading aggressiveness on the
speed of price adjustment differ for stocks with high and low liquidity? Since illiquid
stocks have lower depth of the limit order book at each price level (their limit order book
is “thinner”), price impact per share traded is overall higher for these stocks. Assume
that price impact for a liquid stock equals σ per 100 shares traded and price impact for
an illiquid stock equals x · σ, with x > 1.
Importantly, the difference in price impact per infinitely small trading interval between an aggressive order and a standard order will be higher for an illiquid stock than
for a liquid stock. To illustrate this point, suppose that two orders, a standard limit
order and an aggressive intermarket sweep order, both for 100 shares, are submitted at
t. The aggressive order is executed completely already in t + 1, due to its immediate
execution, and produces the full price impact of σ for a liquid stock and x·σ for an illiquid stock. However, the standard limit order is executed only partially. The executed
part of the standard order equals the number of shares available at the best price, y1 ,
with y ≥ 1 and produces a lower price impact of y1 · σ for a liquid stock and y1 · x · σ
for an illiquid stock. The following table summarizes different levels of the price impact
per trading interval for liquid and illiquid stocks:
Price impact per trading interval
Liquid
Limit order
1
y
·σ
σ
Aggressive order
Illiquid
1
y
·x·σ
x·σ
Obviously, the difference in price impact per trading interval between an aggressive
order and a standard order will be always higher for illiquid stocks, as long as x > 1. The
effect of the aggressive order is larger on the price of an illiquid stock, since the larger
number of shares will be executed in one trading interval and, additionally, the price
will change by a larger amount per each traded share. Basically, the effect of a “thinner
book” for illiquid stocks is additionally multiplied with the effect of “faster trading”
with aggressive orders, so that an aggressive order “goes faster through a thinner limit
order book”. Therefore, positive and negative effects of increased aggressive trading on
the speed of price adjustment after an information release should be more pronounced
for illiquid stocks.
13
3
Data and Sample Construction
3.1
Earnings Announcements Sample
Investigating immediate price reactions after information releases is only possible
if announcements happen within trading hours. For this reason, it is necessary to
have exact time stamps of announcement releases. Since the full compliance with the
Reg NMS was achieved in October 2007, there are only few years of trading data
available. Thus, only information releases that happen regularly can provide a large
enough sample to analyze a broad cross-section of firms in this short period of time.
Earnings announcements are the most natural choice for this study, since this is the
most common type of information release for any stock. Therefore, examining the speed
of price adjustment around these regular announcements is of higher importance than
examining the speed of price adjustment around one-time events, such as a tender offer
announcement or a merger announcement.
Data source for earnings announcements is the Institutional Brokers Estimate System (I/B/E/S) database. I collect announcements between January 2006 and December
2009 that happen within trading hours of US equity trading exchanges (9:30 a.m. to
16:00 p.m. EST). Each record has an exact date and a time stamp (up to a minute).21
Further, each firm is required to exist in the intersection set of I/B/E/S and CRSP.
Table 2 provides details of the sample construction.
[Insert Table 2 approximately here]
The initial sample comprises 10,334 announcements of 3,361 firms. I omit 647
announcements when a stock is not traded on the announcement day, out of which 148
announcements are made on the weekend and the remaining announcements are either
for stocks that stopped trading or for which trading is halted on the announcement
day. Intraday transaction data is not available for 265 firms, which excludes another
967 announcements. I further eliminate very illiquid stocks, for which the closing price
is less than $5 on an announcement day. Excluding days with multiple announcements
21
Bradley et al (2011) show that earnings announcement time stamps reported during the daytime
by I/B/E/S might be delayed. However, over years this time difference becomes lower: for example,
I/B/E/S time stamps and exact announcement time on newswires differ only in 30% of all cases in
2007. Importantly, using delayed time stamps rather leads to underestimation of intraday price jumps
and, if anything, biases my sample against finding significant differences in adjustment time across
stocks.
14
and announcements, with less than 40 days of trading data previously available, leaves
5,944 announcements and 2,307 firms. Finally, to ensure that differences in results
between the pre-Reg NMS period and the post-Reg NMS period are not driven by
differences in samples of underlying stocks, I require that each stock in the sample has
at least one announcement in each of regulation periods. The final sample consists of
3,613 announcements and 675 firms, out of which 1,818 announcements happen prior
to the adoption of the Reg NMS and 1,795 afterward.
One of the most critical requirements for the data set construction is that an announcement should happen within trading hours. Out of 6,536 firms, for which I/B/E/S
reports earnings announcement releases over 2006-2009, 3,175 firms never announce
within trading hours. The remaining 3,361 firms constitute the initial sample, out of
which 58 firms release their earnings information exclusively within trading hours and
3,303 announce both within and outside the trading hours. Overall, firms announcing
both within and outside trading hours are smaller than firms announcing only outside
trading hours, with the median market capitalization of $239 mln and $482 mln, correspondingly (results not tabulated). Even though there is a bias towards smaller firms,
the initial sample still covers more than 50% of all firms with earnings announcement
releases. The final sample does not differ significantly from the initial sample. Table 3
displays summary of main firm characteristics in the final sample and the initial sample.
All variable definitions are given in Table 1.
[Insert Table 3 approximately here]
Since the most illiquid stocks with closing prices below $5 are excluded from the
final sample, the median firm in this sample has a slightly larger market capitalization
of $256 mln, as compared to $239 mln of the median firm in the initial sample. As
expected, the median firm in the final sample is slightly more liquid than the median
firm in the initial sample, as measured by the daily relative spread and the daily Amihud
measure. Overall, the final sample is representative for a broad cross-section of firms
in the US economy.
3.2
Intraday Transaction and Quote Data
The two major variables of interest in this study, trading aggressiveness and price
changes within a day, require the use of intraday transaction data. The source for
these data is the NYSE Transaction and Quote database (TAQ). In the first step,
I extract data on the number and trading volume executed with intermarket sweep
15
orders (ISO) and standard limit orders (non-ISO) for each stock in the final sample
on an announcement day as well as 40 trading days preceding an announcement. The
base period consists of 38 trading days preceding an announcement day, starting on day
-40 and ending on day -2. I collapse transaction-by-transaction data over 15-minute
intervals and extract the number of trades and traded volume in each 15-minute interval
separately for ISO and non-ISO orders. ISO orders are marked with the code “F” in
the condition field of the TAQ database.
To identify the direction of a trade, I use Lee and Ready’s (1991) algorithm: for
each trade I assign the bid (Bt ) and the ask quote (At ) prevailing one second before the
trade took place.22 The midpoint price (Qt ) is calculated as the average of the prevailing
t
). Trades with the transaction price (Pt ) above the
bid and ask quotes (Qt = At +B
2
midpoint price (Pt > Qt ) are identified as buyer-initiated transactions and those with
the transaction price below the midpoint price (Pt < Qt ) as seller-initiated transactions.
If the transaction price is equal to the midpoint price, the current transaction price is
compared with the previous transaction price. If Pt < Pt−1 , I consider a trade to be
seller-initiated; if Pt > Pt−1 , I consider it to be buyer-initiated. Should the two prices
be equal, the trade is left as unclassified.
In addition, I calculate the average 5-minute price impact and the average effective
and quoted relative spread over 15-minute intervals separately for ISO and non-ISO
orders. The quoted relative spread for a transaction is defined as the difference between the corresponding ask and the corresponding bid, scaled by the midpoint price
(RelSprt = (At − Bt )/Qt ). I set observations with RelSpr > 0.5 to missing values.
The effective relative spread of each transaction is calculated as twice the absolute
difference between the transaction price and the midpoint price, scaled by the midpoint price (Ef f Sprt = 2 |Pt − Qt | /Qt ). Observations with Ef f Spr > 0.5 are also
set to missing values. The price impact of each trade after 5 minutes is defined as
P rcImpt = 2 |Qt+5 − Qt | /(Qt · wt ), where Qt+5 represents the midpoint price for a
stock after five minutes (300 seconds) and wt is the size of a transaction (in shares).
Intraday one-minute returns are computed from the closing midpoint price for each
minute from the TAQ Consolidated Quotes database. Closing midpoints better serve
the purposes of the price adjustment analysis, since they exclude the bid-ask bounce,
which is present in transaction prices.
22
Henker and Wang (2006) consider this procedure to be more appropriate compared to the classical
Lee and Ready (1991) five-second rule. Bessembinder (2003) tries zero- to thirty-second delays in
increments of five seconds and does not find any differences in the results.
16
4
Trading Aggressiveness in the Base Period and around
Earnings Announcements
Definition of Trading Aggressiveness. I define trading aggressiveness as the proportion of total volume, traded with aggressive intermarket sweep orders within a particular time interval (the proportion of ISO volume, %ISO V olume). Daily trading
aggressiveness is, therefore, the proportion of daily volume, executed through ISO orders. Intraday trading aggressiveness is measured as the proportion of ISO volume over
a respective time interval within a day, for example 10 minutes, 15 minutes, 1 hour
etc.23 In the remainder of the paper, terms “trading aggressiveness” and “trading with
aggressive orders” are used interchangeably.
The median proportion of ISO volume in my sample is 36%. However, the variation
is quite significant with 22% of volume traded through ISOs for firms in the lowest
decile and 56% in the highest decile (not tabulated).
Trading characteristics of aggressive orders. In the first step, I analyze differences in characteristics of intermarket sweep orders (ISO) and standard limit orders
(non-ISO) in the base period and in hours immediately following an earnings announcement release.
Since the base period is rather short (38 days) and proportions of the number of
trades and of their volume are not normally distributed, I create an empirical distribution for these variables with a bootstrapping procedure. I draw with replacement one
observation from the base period that happened between the time of an announcement
release and the end of the trading day for each stock-announcement and repeatedly
calculate the mean across all stock-announcements in this bootstrapped sample. The
empirical distribution for each variable is based on 1,000 bootstrapped samples. Table
4 summarizes differences in trading characteristics between ISO and non-ISO orders
in the base period and on the event day. Columns 2 and 3 display the mean of the
bootstrapped distribution for ISOs and non-ISOs from the base period, correspondingly. Columns 5 and 6 report the cross-sectional mean of respective variables starting
from an announcement release until the end of the trading day (16:00 p.m. EST). The
last two columns display the difference-in-differences (or simply the difference between
the base period and the event day for variables, calculated as proportions) and test its
significance with a standard t-test.
23
The proportion of the total number of trades, executed through ISOs, is highly correlated with the
proportion of ISO volume (correlation coefficient of 93%). Main insights of the paper do not change,
if the proportion of ISO trades is used to measure trading aggressiveness.
17
[Insert Table 4 approximately here]
The proportion of trades, executed with aggressive orders, %T rades, equals 40.8%
in the base period. It increases significantly by 4.6% on an announcement day in
hours immediately following an information release. The proportion of total volume,
executed with aggressive orders, %V olume, is lower than %T rades in the base period,
only 37.8%, but it also significantly increases to 42.4% on an announcement day. The
reason for the lower proportion of ISO volume is an overall smaller size of ISO orders.
An average size of an ISO in the base period equals 176 shares, as compared to 256
shares of a standard limit order. The size of an ISO is smaller, since the TAQ database
records not a cumulative size of all ISOs, sent simultaneously across all exchanges, but
rather a size of each individual order, sent and executed on a particular stock exchange.
Interestingly, ISOs are used approximately equally for purchases and for sales, with sales
slightly dominating purchases both in the base period and on an announcement day.
The proportion of purchases, executed with ISOs, %P urchases, and the proportion of
ISO sales, %Sales, both increase significantly by around 4% on an announcement day.
Overall, ISO characteristics in my sample are similar to ISO characteristics in the
Chakravarty et al (2010) sample, a pioneering study on the use of intermarket sweep
orders.24 In the later study, Chakravarty et al (2011a) also document an increase of
5.5% in the proportion of ISO volume on an earnings announcement day. In addition to
the daily trading aggressiveness, as analyzed by Chakravarty et al (2011a), I examine
changes in the use of aggressive orders on the intraday level, in trading hours immediately before and after earnings announcement releases. Intraday analysis can help
better understand whether aggressive trading persists during the whole announcement
day and in which hours investors trade most aggressively.
Figure 3 displays mean percentage deviations in the proportion of ISO volume
throughout an announcement day. Deviations from the bootstrapped means are measured in 15-minute intervals relative to the 15-minute interval with an earnings announcement release (interval 0). The dashed line shows the 1% significance level for
the mean percentage change in the proportion of ISO volume, equal to 3.8%.
24
The proportion of ISO trades is 46% and the proportion of ISO volume equals 41% in their sample.
The average size of an ISO order equals 178 shares and is also significantly smaller than the average
size of a standard limit order.
18
[Insert Figure 3 approximately here]
Trading aggressiveness experiences a jump of up to 15% (%ISO V olume = 43.47%)
in the first fifteen minutes after an information release. Afterward, it steadily decreases,
but never drops below the 1% cutoff value till the end of the trading day. In the periods
prior to an announcement release, deviations are significant in 60% of all cases and rarely
drop below zero.
Why does trading aggressiveness increase on an announcement day? The reasons
for this increase may be twofold. First, investors have different rates of information
processing. Those investors who are able to process new information more quickly try
to exploit their advantage. Second, trading by uninformed investors might also become
more aggressive due to scarce liquidity supply and reduced market breadth around
earnings announcements. Chakravarty et al (2011a) provide empirical evidence in
support of the second explanation. The 15% jump immediately after a release, however,
indicates that there might be an additional pressure from traders with quicker rates of
information processing immediately after an information release. Since the influence
of trading aggressiveness on the speed of price adjustment depends on whether the
majority of aggressive trades comes from informed or uninformed investors, the detailed
analysis of an overall informativeness of ISOs follows in the next section.
Further results from Table 4 show that the effective relative spread, Ef f Spr, is
marginally lower for ISOs in the base period, 1.73%, as compared to 1.82% for a nonISO order. However, it does not differ significantly from the non-ISO effective spread
on the event day: the ISO effective spread increases to 1.84% in hours following an
information release, whereas the non-ISO spread practically remains on the same level.
One explanation for this finding might be that trading with ISOs is more informed in
the base period, so that aggressive traders can target better execution prices in the
base period.25 As liquidity around information releases reduces, all traders, including
uninformed ones, become more aggressive, and the effective relative spread increases
accordingly to the level of a non-ISO order.
The price impact of ISO trades, P rcImp, is higher than the price impact of standard
limit orders, and even more so in hours following an information release (the differencein-differences equals 0.11% and is statistically significant at the 5% level). This finding
is important, since it provides the first supportive evidence for the assumption that ISO
trades have higher price impact per trading interval than non-ISO trades, made earlier
25
Chakravarty et al (2010) provide empirical evidence that ISO trades have significantly larger
Hasbrouck (1995) information share, as compared to non-ISO trades, during normal trading periods.
19
in the hypothesis development section of the paper. The multivariate analysis of the
intraday price impact of ISO trades follows later in this section.
Determinants of ISO volume. The previous analysis shows that ISOs constitute
a considerable proportion of trading volume, they have smaller size and higher price
impact than standard limit orders. The next step is to examine, which stock characteristics lead to higher rates of aggressive trading, both in the base period and around
earnings announcement releases. Panel A of Table 5 analyzes determinants of ISO volume in a multivariate setup. Models (1) and (2) report results of cross-sectional OLS
regressions for the base period with the mean proportion of daily volume, executed
through aggressive orders as the dependent variable (%ISOvol ). Models (3) and (4)
examine factors that influence percentage deviations in the proportion of ISO volume
after an earnings announcement release from its base period level (∆ISOvol). All regressions include year-fixed effects. Models (3) and (4) additionally include weekdayand daytime-fixed effects.
[Insert Table 5 approximately here]
Results from Models (1) and (2) show that traders use aggressive orders mostly for
stocks with a larger market capitalization during normal trading times. This finding is
intuitive, since intermarket sweep orders provide an opportunity for large traders, such
as institutional investors, to trade large blocks quickly. Thus, trading with aggressive
orders should be especially intensive in large firms, which usually have a higher percentage of institutional ownership. Further, investors trade higher proportion of their
volume with ISOs in NASDAQ-listed stocks.26 Importantly, the proportion of ISO volume does not differ significantly in the base period for stocks with different liquidity
levels, as measured by the daily average of intraday quoted relative spread (Model 1)
and the Amihud measure (Model 2). The situation changes to the opposite on an announcement day: trading aggressiveness increases more for illiquid stocks, whereas the
size of the firm does not play a role any longer (Models 3 and 4). Also, in line with prior
expectations, trading aggressiveness is higher after larger earnings surprises, as proxied
by an absolute stock return in 24 hours after an announcement release (EarnSurp).
Results from Table 5 suggest that there is an important difference between the use
of ISO orders in normal trading periods and around earnings announcement releases.
26
There is no clear explanation to this fact. However, prior study by Chakravarty et al (2011b) also
documents an overall higher proportion of ISO volume for NASDAQ-listed stocks.
20
In the periods of normal trading ISO orders are mostly used to trade large blocks of
orders in the stocks of large firms (with higher institutional ownership), which exactly
justifies their introduction to equity markets. Around earnings announcement releases,
however, traders mostly use ISO orders to extract all available liquidity in stocks with
already scarce liquidity supply. To investigate this issue closer, I report the percentage
of ISO trades and ISO volume in hours following an earnings announcement release
separately for liquid and illiquid stocks (Panel B of Table 5). The stock is classified as
liquid if its daily quoted relative spread is above the median for all stocks in the sample
in the base period, and it is classified as illiquid otherwise. The construction of Panel
B is similar to the construction of Table 4. Indeed, the proportion of ISO volume is
higher for illiquid stocks (43.3%) than for liquid stocks (41.6%) in post-announcement
hours and the difference is statistically significant at the 1% level.27
Overall, findings from the previous analysis suggest that traders increase their trading aggressiveness on an announcement day, especially for illiquid stocks. This result
provides additional support for Chakravarty et al (2011a) explanation that uninformed
traders are induced to switch to ISOs around information releases, as a result of scarce
liquidity supply. Obviously, there is an endogeneity issue between the intraday use of
ISOs and the intraday liquidity of the stock. As liquidity supply reduces around information release, all traders become more aggressive, which drives even more liquidity
suppliers away from the market. The analysis of this issue is, however, beyond the
scope of this paper.
Intraday analysis of the effective relative spread and price impact. Previous results from Table 4 confirm that ISO trades have an overall higher price impact
per trading interval than non-ISO trades. However, I expect the difference in intraday
price impact between ISO trades and non-ISO trades to be higher for illiquid stocks
than for liquid stocks. Remember that an aggressive order can go faster through a thinner limit order book of an illiquid stock, which produces an even higher price change,
as compared to the price change after the same aggressive trade in a liquid stock. The
last line in Panel B of Table 5 confirms this prediction, since the difference in intraday price impact between ISO and non-ISO trades for illiquid stocks is higher than
the corresponding difference for liquid stocks by 0.45% and is statistically significant
at the 1% level. The difference-in-differences value for the effective relative spread is
27
The higher proportion of ISO volume for illiquid stocks is the result of the significantly larger
proportion of ISO trades for these stocks. The differences in size of ISOs cannot be the reason for
the higher proportion of ISO volume for illiquid stocks, since the relative size of ISO trades (Size of
ISO/Size of non-ISO) is significantly smaller for these stocks.
21
insignificant, which implies that aggressive orders are not executed at better prices on
an announcement day for both stock types.
Table 6 additionally investigates differences in the effective relative spread and the
intraday price impact between ISO and non-ISO trades in a multivariate setup. One
observation represents a 10-minute trading interval for a stock. All models represent
panel OLS regressions and include firm-, year-, daytime- and weekday-fixed effects. In
addition, I control for the inverse of the mean stock price in a 10-minute period, which
is mechanically related to the two dependent variables, for total volume executed within
a 10-minute trading interval and for the listing exchange of a stock.
[Insert Table 6 approximately here]
Model 1 shows that there exist no significant differences in the effective relative
spread of ISO and non-ISO trades in the base period, since the coefficient on the
indicator variable ISO is not significant after including control variables. The coefficient
on the interaction between ISO and Event, denoting an announcement day, is also not
significant, which confirms univariate results that effective spreads of ISO and non-ISO
trades do not differ significantly in post-announcement hours. These results hold if I
additionally control for liquidity of a stock with an indicator variable Illiquid (Model
2). All control variables have expected signs.
In line with prior findings, the intraday price impact is higher for ISO trades in
the base period and even more so on an announcement day. However, the latter effect
disappears if I control for liquidity of a stock. This means that an additional increase
in the intraday price impact on an announcement day is driven by illiquid stocks, which
is consistent with univariate results from Tables 4 and 5. After controlling for liquidity
as well as other control variables, the additional intraday price impact of an aggressive
order constitutes 0.187% for an illiquid stock, which is statistically and economically
significant (e.g., 1.87 cent for a stock with a price of $10).
Informativeness of ISO trades after earnings announcements. Trading
aggressiveness increases significantly in hours following earnings announcement releases
and the price impact per time interval of aggressive intermarket sweep orders is higher
than the price impact of standard limit orders. Does an increased use of aggressive
orders in post-announcement hours represent informed or uninformed trading? This
question is important, since the effect of trading aggressiveness on the speed of price
adjustment depends on informativeness of ISO trades. The effect is positive (higher
trading aggressiveness speeds up price adjustment) if the majority of ISOs are submitted
22
by informed investors, who trade in the direction of the new equilibrium level. The
effect is negative (higher trading aggressiveness slows down price adjustment) if ISOs
are mostly submitted by uninformed investors, who are equally likely either to buy or
to sell a stock.
I analyze informativeness of ISO trades by testing whether the proportion of ISO
volume increases more in the direction of the earnings surprise. For positive earnings
surprises, aggressive trading is more informative if the proportion of ISO buy volume
(%ISOBuyV ol = ISO buy volume/T otal buy volume) is higher than the proportion of
ISO sell volume (%ISOSellV ol = ISO sell volume/T otal sell volume). For negative
earnings surprises, the opposite relationship should hold. I measure an earnings surprise
as a 24-hour stock return after an announcement release. The results do not differ
materially, if CAR(0;1) or I/B/E/S analyst earnings forecasts are used to measure
earnings surprise. In case of analyst earnings forecasts, I lose around 50% of observations
in my final sample due to missing data on I/B/E/S.28
In the first step, I examine the imbalance between the proportion of ISO buy volume
and the proportion of ISO sell volume on an announcement day. Figure 4 displays both
proportions for each 15-minute event interval, relative to the 15-minute interval with an
earnings announcement release (interval 0). The dashed line marks the event interval
0.
[Insert Figure 4 approximately here]
Panel A shows the imbalance in proportions of ISO volume for positive earnings
surprises. The proportion of ISO buy volume is on average higher than the proportion
of ISO sell volume in one hour preceding an announcement release. However, in one
hour after an announcement both the proportion of ISO sell volume and the proportion
of ISO buy volume increase, with a slightly higher increase for ISO sales than for ISO
purchases. The proportion of ISO buy volume starts dominating only in the second hour
after an information release and continues to dominate until the end of an announcement
day. These preliminary results suggest that the majority of ISO trades are mostly
uninformative in the first hour after a positive earnings announcement release. As time
28
Chakravarty et al (2011a) estimate the information share of Hasbrouck (1995) for ISO trades and
find that the information share of ISO trades decreases with the higher increase in trading aggressiveness around earnings announcement releases. This finding suggests that more uninformed investors
switch to ISOs around earnings announcements. My analysis is complementary to theirs, since I investigate the imbalance of ISO volume and, in addition, differentiate between positive and negative
earnings surprises.
23
passes by, traders correctly process the information from an announcement release and
ISO trades increase their informativeness.
The situation is different for negative earnings surprises (Panel B). The proportion
of ISO sell volume experiences a jump of up to 3% (from 47% to 50%) already in the
first fifteen minutes after an announcement and significantly dominates the proportion
of ISO buy volume for at least 3 hours after an announcement release. Thus, investors
react quickly to the negative news and increase their aggressiveness on the sell side
almost immediately.
Table 7 examines informativeness of ISO trades separately for subsamples of liquid
and illiquid stocks. In addition to positive and negative earnings surprises, I differentiate
between large positive surprises and large negative surprises. An earnings surprise is
defined as large if a 24-hour stock return is above its median for positive earnings
surprises and below its median for negative earnings surprises. The first column shows
the number of hours since an announcement release. The remaining columns report the
difference in means between an increase in the proportion of ISO buy volume and an
increase in the proportion of ISO sell volume for the corresponding hour:
∆ = ∆Buy − ∆Sell ,
where
and
∆Buy = %ISOBuyV olEvent − %ISOBuyV olBase
∆Sell = %ISOSellV olEvent − %ISOSellV olBase .
T-statistics of the two-tailed t-test with the null-hypothesis of a difference in means
equaling zero are in parentheses below each coefficient.
[Insert Table 7 approximately here]
The first interesting observation is that investors increase their trading aggressiveness on average in the correct direction: they trade a higher proportion of ISO buy
volume if an earnings surprise is positive (∆ > 0), and a higher proportion of ISO sell
volume if an earnings surprise is negative (∆ < 0). As expected, the differences in
proportions are higher for larger earnings surprises.
Consistent with Figure 4 (A), ISO trades are quite uninformative in the first hour
after a positive announcement release for both liquid and illiquid stocks. For illiquid
stocks, the difference is even negative, which means that investors increase their aggressiveness more in the direction of sell orders. Over time aggressive trading becomes more
informative and the differences attain their peak in the third hour after an announcement release for both types of stocks. However, all coefficients are not statistically
24
different from zero. The results are mostly the same for large positive surprises, except
that an increase in the proportion of ISO buy volume is even higher and becomes statistically significant at the 10% level for liquid stocks after 4 hours since an announcement
release.
For negative earnings surprises, the results differ strongly between subsamples of
liquid stocks and illiquid stocks. ISO trades are largely uninformative for liquid stocks
(the differences are quite small and not significant), and they are strongly informative
for illiquid stocks for up to 5 hours after an announcement release (the differences are
negative and significant either at the 5% or at the 1% level, also for large negative
surprises). These findings suggest that a significant jump in the proportion of ISO sell
volume immediately after an announcement release, observed in Figure 4 (B), is mainly
driven by an increase in trading aggressiveness of illiquid stocks.
Overall, even though investors increase their trading aggressiveness mostly in the
correct direction, ISO trades are largely uninformative for positive earnings surprises
and are strongly informative for negative earnings surprises, but only in the subsample
of illiquid stocks.
5
Trading Aggressiveness and The Speed of Price Adjustment After Earnings Announcements
This section examines the central question of this paper: how does an increase in trading aggressiveness after an earnings announcement release influence the speed of price
adjustment to its new equilibrium level? Remember that trading aggressiveness is measured as the proportion of ISO volume within a particular time interval. For “the speed
of price adjustment”, however, there exists no precise definition. Theoretically, the
speed of price adjustment can be measured as the difference in time (e.g., in seconds)
between an announcement release and the time when the price reaches its new equilibrium level. Since the new equilibrium price level is unobservable, the most important
empirical problem is to define the time period when the price ends its adjustment process. In the next step, I provide a detailed definition of the end of the price adjustment
process for this study.
Definition of the end of the price adjustment process. I consider that price
adjustment ends in the period, in which volatility of one-minute closing midpoint returns is no longer abnormal. The volatility criterion is more appropriate for this study
than, for example, measuring abnormal returns or abnormal serial correlations in price
25
changes, since this is the only criterion that captures two stages of price adjustment,
the initial price reaction as well as the subsequent period of price stabilization. The
initial price reaction is mostly driven by orders of informed traders, and an increase in
trading aggressiveness of these informed traders can potentially make this first stage
shorter. At the same time, an increase in trading aggressiveness of uninformed traders,
which happens due to scarce liquidity supply, increases intraday volatility of the stock
and the probability of price overshooting, which makes it harder for a price to stabilize
around its new equilibrium level.
Prior studies by Patell and Wolfson (1984) and Jennings and Starks (1985) find
that the initial price reaction is rather short, since abnormal returns dissipate in five to
fifteen minutes after an earnings announcement release. However, abnormal volatility of
intraday returns persists for several hours and can even extend to the following trading
day. The recent study by Brooks, Patel and Su (2003) provides similar evidence
for unanticipated events with abnormal returns lasting for 15 minutes and abnormal
variance for at least three hours after an event.
The exact procedure to determine the end of the price adjustment process is as
follows. First, I calculate volatility of one-minute closing midpoint returns within each
10-minute interval for an event window, which includes event days 0 to 2, and for the
base period with event days -40 to -3. I numerate all intervals in the event window
relative to the first 10-minute interval immediately after an announcement (interval 0).
The numeration is consecutive for all days in the event window. For example, if an
announcement time was 3 p.m. on day 0, then a period from 9:30 a.m. till 9:40 a.m.
on the next day is numerated as period 7. Such a numeration procedure is needed,
since for each announcement I have to identify a unique number of 10-minute interval,
in which volatility of one-minute closing midpoint returns is no longer abnormal. For
some of these announcements, however, volatility deviates abnormally not only on an
announcement day, but also for up to the following two days. I use a non-parametric
test, proposed by Smith et al (1997), to define whether volatility is abnormal in each
of the 10-minute periods after an information release. Volatility is considered to be
abnormal if it exceeds the median volatility in the same 10-minute period calculated
over days [−40; −3].29
The period, in which the price ends its adjustment, is defined as the first 10-minute
29
To increase the power of the test, I compare volatility in each of 10-minute intervals in the event
window to six 10-minute intervals that lie within the same hour over all non-event days. Thus, if a
stock was traded over all 10-minute intervals in the base period, one 10-minute interval in the event
window is compared to 6 intervals per day * 38 days in the base period = 228 ten-minute periods.
26
period, for which:
- volatility is no longer abnormal
- volatility in the preceding hour is abnormal (more precisely: volatility in the
preceding six 10-minute intervals is abnormal in more than 50% of all cases)
- volatility in the following hour continues to stay on a normal level (more precisely:
volatility in the following six 10-min intervals is normal in more than 50% of all cases).30
Univariate results. Panel A of Table 8 displays the distribution of the length of
price adjustment periods across pre- and post-Reg NMS periods, separately for subsamples of liquid and illiquid stocks. One unit corresponds to a ten-minute interval.
Thus, the median length of a price adjustment period for a liquid stock prior to the
Reg NMS is 18*10=180 minutes or 3 hours after an announcement release. After the
Reg NMS the median adjustment time for liquid stocks has significantly decreased by
forty minutes (4 intervals), as reported by a non-parametric Mann-Whitney test. Surprisingly, the median length of a price adjustment period for an illiquid stock has even
increased by almost an hour in the post-Reg NMS period (from 5 hours and 40 minutes
to 6 hours and 30 minutes), even though this difference is not statistically significant.
Given a significant increase in an overall electronic trading in recent years, it is even
more surprising that the length of the adjustment period did not decrease for illiquid
stocks. Further, the standard deviation of the length of a price adjustment period has
also increased for these stocks, which implies that the adjustment process has become
quicker for some illiquid stocks in the post-Reg NMS period and slower for the other
ones.
[Insert Table 8 approximately here]
To investigate this issue more closely, I break all announcements into terciles of trading aggressiveness (TA1 - TA3) in the post-Reg NMS period. TA3 includes announcements with the highest increases in trading aggressiveness on an event day, whereas
30
This paper takes a short-run perspective on the speed of price adjustment, that is, I assume
that price adjustment ends shortly (up to two days) after an information release. There is, however,
empirical evidence that stocks earn on average positive (negative) abnormal returns after positive
(negative) earnings surprises up to three quarters after an announcement release - the tendency called
“post-earnings announcement drift” (PEAD). Hirschleifer et al (2008) provide a recent survey on
findings of the PEAD literature. In that sense, the full price adjustment may last up to one year
since the time of an information release. However, since I am interested in the effect of a temporary
increase in trading aggressiveness on the price adjustment process, the short-run perspective is the
most suitable for this study.
27
TA1 includes stocks with the lowest increases.31 To compare the change in the length
of an adjustment period between two regulation regimes, I also assign “pseudo”-terciles
of trading aggressiveness for all announcements in the pre-Reg NMS period. For this
purpose, I calculate the median TA tercile for each stock after the Reg NMS and assign
this TA tercile for all announcements of this stock that happen prior to the Reg NMS.32
Panel B of Table 8 displays the mean number of 10-minute periods, until the price ends
its adjustment, for each of TA terciles. The last two rows report the significance level
of the t-test on the equality of means across different terciles of trading aggressiveness.
Consistent with previous results from Panel A, the price adjustment process is
quicker in the post-Reg NMS period for each TA tercile in the sample of liquid stocks.
The relationship between an increase in trading aggressiveness and the speed of price
adjustment is positive, but rather weak - the difference between adjustment time for
announcements in the first and the third tercile constitutes 20 minutes and is not statistically significant.
Whereas changes in trading aggressiveness on an announcement day barely have any
influence on the average speed of price adjustment of liquid stocks, there is a striking
U-shaped relationship for stocks with low liquidity levels. Illiquid stocks with moderate
changes in trading aggressiveness (TA2) do not experience a significant change in their
adjustment time, as compared to the pre-Reg NMS period. Surprisingly, the price adjustment process slows down for illiquid stocks with the largest decreases (TA1) and the
largest increases (TA3) in their trading aggressiveness in the post-announcement hours.
For stocks with excessive increases in their trading aggressiveness (TA3), adjustment
time increases by 1 hour and 20 minutes, and this difference is statistically significant at
the 5% level. The two-hour difference in means between the second and the third tercile
of trading aggressiveness in the post-Reg NMS period is also statistically significant at
the 1% level.
Figure 5 depicts the relationship between the mean length of the price adjustment
period, measured as the number of ten-minute time intervals until the price stabilizes
on its new level, and the mean change in the proportion of ISO volume.
31
Remember that trading aggressiveness is measured as the change in the proportion of ISO volume
traded after an announcement release relative to its mean in the base period (∆ISOvol). Although on
average trading aggressiveness increases in hours after the release, changes in the proportion of ISO
volume can take negative values for some stocks in the TA1 sample.
32
Normally, there is almost no within-stock variation in liquidity and trading aggressiveness and I
can correctly assign the “pseudo”-TA tercile for almost 90% of all stocks in my final sample. I exclude
the remaining 10% from the univariate analysis in Table 8, but include these observations later in my
multivariate analysis.
28
[Insert Figure 5 approximately here]
Overall, the patterns are consistent with those reported in Table 8. For liquid stocks,
the relationship is always increasing, with slightly quicker price adjustment for stocks
with higher negative deviations in the proportion of ISO volume on an announcement
day. For illiquid stocks, the relationship has a U-shape, with the slower adjustment
process for stocks with higher absolute deviations in the proportion of ISO volume.
The adjustment period is the longest for stocks with excessive increases in their trading
aggressiveness on an announcement day (above 50%).33
Findings so far show that large increases in trading aggressiveness may be harmful
for the speed of price adjustment after an earnings announcement release, especially in
the sample of illiquid stocks. Thus, negative effects of trading aggressiveness (higher
abnormal volatility and an increased probability of price overshooting) dominate in
case when traders increase the proportion of ISO volume by more than 50% from its
base level. This result suggests that large increases in aggressiveness are mainly driven
by uninformed traders, who switch to aggressive orders as soon as liquidity supply
decreases - however, this result still has to be checked in the multivariate analysis.
A more surprising finding is that decreases in trading aggressiveness have different
implications for liquid and illiquid stocks. Higher decreases in trading aggressiveness
are beneficial for liquid stocks, but they slow down the adjustment process of illiquid
stocks. Why? Remember that aggressive orders induce quicker price changes per one
time interval. Since liquid stocks are by definition traded more frequently, their prices
adjust more quickly, and thus, there is no great need for even quicker price movements
of these stocks. Therefore, only very small proportions of ISO volume can positively
influence their adjustment process. Illiquid stocks are infrequently traded, so they can
profit from quicker price movements, especially with moderate increases in the use of
ISO orders. If trading aggressiveness considerably decreases, price changes per one time
interval are relatively small, which slows down the movement of a stock price towards
its new equilibrium level.
Regression analysis. To control for other factors that might influence the length
of the adjustment period, I estimate negative binomial regressions with the number
of 10-minute intervals required for price adjustment as the dependent variable. The
identification strategy is a difference-in-differences analysis, since I am interested in the
33
Stocks with excessive increases in their trading aggressiveness (∆ISOvol > 0.5) are not just
outliers in the sample, since 222 stocks out of 839 illiquid stocks belong to this group. Only the group
of stocks with large negative decreases in their trading aggressiveness (∆ISOvol < −0.5) has slightly
less than 200 observations.
29
effect of changes in trading aggressiveness on the speed of price adjustment after controlling for the differences in the speed of price adjustment in the pre-Reg NMS period.
For this reason, all regressions include earnings announcements from the pre- and the
post-Reg NMS period. Since each stock in the final sample has at least one announcement in each of the regulation periods, the results are not influenced by differences in
the underlying stock samples.The main regression specification is as follows:
#Intervalsi = αi · Xi + θi · Zi + εi ,
where Xi is the vector of explanatory variables and Zi is the vector of control variables. Xi includes the following variables: Post Reg, which equals 1 if an announcement
happened after the adoption of the Reg NMS, and 0 otherwise; Illiq, which equals 1
if the relative spread of the stock is above the median of all stocks in the sample, and
0 otherwise; the interaction of the previous two variables Illiq · P ost Reg; the change
in the proportion of ISO volume for liquid stocks, Liq · ∆ISOvol; the positive (negative) change in the proportion of ISO volume for illiquid stocks, Illiq · |∆ISOvol|∆>0
(Illiq · |∆ISOvol|∆<0 ). I examine separately the influence of positive and negative
deviations in the proportion of ISO volume on the length of the adjustment period
for illiquid stocks, since the relationship between trading aggressiveness and the speed
of price adjustment might be U-shaped for these stocks (as suggested by univariate
results). For liquid stocks, the relationship is monotonically increasing, and the differentiation between positive and negative changes is not necessary.
The vector of control variables Zi consists of the mean turnover on event days [0; 2],
Turnover, the mean intraday volatility of one-minute returns in the base period, Intr
Volat, and an average size of a firm, proxied by the log of its market capitalization at
the beginning of the base period, LnMCap. I expect the coefficient for Turnover to be
negative, since more frequently traded stocks converge quicker to their equilibrium level.
The higher intraday volatility of a stock in the base period, the harder it is for a price to
stabilize on its new level. Thus, I expect a positive coefficient for this variable. Finally,
I expect a negative coefficient on LnMCap, since stocks of large firms are normally
traded by a larger number of investors, which contributes to the quicker convergence
of the stock price. I also include year-fixed effects and control for the weekday and the
time of an announcement.
[Insert Table 9 approximately here]
Model (1) of Table 9 presents the results for all announcements in the final sample.
The benchmark value of the dependent variable equals the mean length of the price
30
adjustment period for liquid stocks prior to the adoption of the Reg NMS (around 6
hours, according to Panel A of Table 8). All coefficients should be interpreted as relative
changes in the length of the price adjustment period from this benchmark value.34 For
example, the coefficient on Post Reg equals -0.15 and is statistically significant at the
10% level. The coefficient of -0.15 means that the length of the price adjustment period
has decreased by e−0.15 − 1 = −0.14 or 14% from the benchmark value for liquid stocks
in the Post-Reg NMS period (from 6 hours to 5 hours and 20 minutes). As expected,
the price adjustment period is significantly longer for illiquid stocks by approximately
34%. It does not differ significantly between two regulation subperiods.
Consistent with univariate results, the relationship between trading aggressiveness
and the speed of price adjustment is positive and significant at the 5% level for liquid
stocks in the post-Reg NMS period. A 100% increase in the proportion of volume traded
with ISOs increases the number of 10-minute periods until the price fully adjusts by
22% or by approximately 50 minutes.35 Further, multivariate results also confirm the
U-shaped relationship between changes in trading aggressiveness and the length of the
adjustment period for the sample of illiquid stocks. The higher the absolute value of
the deviation in the proportion of ISO volume (positive or negative), the longer it takes
for a stock to converge to its new level. A 100% change in the proportion of ISO volume
slows down the adjustment process by 27% for positive changes and by 28% for negative
changes, or approximately by 2 hours from its mean value of 7 hours if ∆ISOvol = 0.
This change is significant not only statistically, but also economically.
As expected, price adjustment is significantly shorter for stocks with a larger turnover
during the event window and longer on average for stocks with higher intraday volatility
in the base period. However, the coefficient on LnMCap is close to zero and insignificant. The insignificant coefficient means that the size of the firm does not add any
explanatory power if there is an explicit control for liquidity of the stock.
Consistent with univariate results, excessive increases in trading aggressiveness slow
down the adjustment process, especially for illiquid stocks. Previously formulated hypotheses suggest that such a negative effect of trading aggressiveness should dominate
if investors, who submit intermarket sweep orders, are largely uninformed and have
heterogeneous beliefs. In such a case, aggressive trades produce very quick upward
34
General interpretation of any coefficient β in negative binomial regressions is as follows: for one
unit change in the explanatory variable, the difference in logs of expected counts of the response
variable is expected to change by β.
35
The mean number of 10-minute periods until adjustment equals 25 if ∆ISOvol is 0. The expected
22% change increases this number to 30 for ∆ISOvol = 1.
31
price changes following purchase transactions and very quick downward price changes
following sale transactions. An increase in aggressive trading increases then the probability of price overshooting and intraday volatility of the stock. To test this hypothesis,
I split the total sample in subsamples of positive earnings surprises and negative earnings surprises. Remember that aggressive trading is largely uninformative after positive
earnings surprises for both liquid and illiquid stocks, and it is informative after negative earnings surprises only in the sample of illiquid stocks. Thus, I expect the negative
influence of excessive trading aggressiveness to be stronger in the sample of positive
earnings surprises, especially for illiquid stocks.
Model (2) presents results for positive earnings surprises. The most striking result is
that coefficients on changes in trading aggressiveness increase in their value, especially
for illiquid stocks. For positive earnings surprises, a 100% change in the proportion of
ISO volume slows down the adjustment process of illiquid stocks by 41% for positive
changes and by 58% for negative changes. So, an extreme decrease in trading aggressiveness can be even slightly more harmful, than an extreme increase, in the situations
when majority of investors are uncertain about the new fundamental price for an illiquid stock. A large decrease in trading aggressiveness may suggest that the majority of
investors are uninformed and they prefer to stay out of the market, which increases the
probability that the trading process freezes out completely. As ISO trades are slightly
more informative for big positive earnings surprises, the negative effect of trading aggressiveness becomes smaller for illiquid stocks, and completely disappears for liquid
stocks (Model 3). In contrast, the length of the adjustment period even decreases for
liquid stocks, since aggressive trades become strongly informative after 4 hours since an
announcement with a big positive earnings surprise (see Panel A of Table 7). However,
this decrease is not statistically significant.
For negative earnings surprises, there is still a negative effect of trading aggressiveness on the speed of price adjustment of illiquid stocks, however, it is rather small and
not statistically significant (Models 3 and 4). This finding confirms prior expectations
that the negative influence of trading aggressiveness is lower when ISO trades are on
average more informative. For liquid stocks, the negative effect still dominates, since
aggressive trading is largely uninformative after negative earnings surprises for these
stocks. This effect is especially high after big negative surprises and increases the length
of the adjustment period for liquid stocks by 39% or 1.5 hours, which is statistically
and economically significant.
32
Robustness checks. Since the average price of illiquid stocks, $20.1, is lower than
the average price of liquid stocks, $32.8, larger deviations in intraday volatility of oneminute returns of illiquid stocks on an announcement day might be just mechanical.
Therefore, the lower price of illiquid stocks could bias adjustment time upwards and
overestimate the influence of trading aggressiveness on the speed of price adjustment
of these stocks. To account for the price level of illiquid stocks, I use an alternative
definition of intraday volatility. Intraday volatility is now calculated as the maximum
price change (in cents) within each 10-minute interval:
Intr V olatt = Pmax, t − Pmin, t .
Apart from the new definition of intraday volatility, the identification procedure of
the end of the price adjustment process remains the same. Models (1)-(3) of Table 10
repeat the analysis of Table 9 with the modified dependent variable. Model (1) reports
results for the total sample, Model (2) analyzes the subsample of positive earnings
surprises and Model (3) examines the subsample of negative earnings surprises. The
influence of trading aggressiveness on the convergence time of illiquid stocks is consistent
with previous results: excessive increases and decreases in trading aggressiveness have
a negative effect in the sample of positive earnings surprises, and do not have any
significant effect in the sample of negative earnings surprises. For liquid stocks, the
negative effect is, however, smaller and loses its significance.
[Insert Table 10 approximately here]
Finally, I analyze whether main results change if the illiquidity of a stock is proxied
by the daily Amihud (2002) measure.36 An indicator variable Illiq now equals 1 if
the mean Amihud measure of the stock in the base period is above the median of all
stocks in the sample, and equals 0 otherwise. Models (4) to (6) of Table 10 display
the corresponding results.37 Again, main results are consistent with those of Table 9.
The negative influence of trading aggressiveness on adjustment time of illiquid stocks
decreases in the sample of negative earnings surprises, but remains still significant at
the 10% level.
Overall, findings in this section show that excessive increases in trading aggressiveness in the post-announcement hours slow down the price adjustment process of
36
The Amihud (2002) measure is defined as the ratio of the daily absolute return to the
dollar trading volume on that day: Illiqi,t = |Ret|i,t /Dollar V olumei,t .
37
The dependent variable is based on the standard definition of intraday volatility.
33
all stocks. For illiquid stocks, the relationship is more pronounced and exhibits a Ushape: the adjustment process becomes slower following not only large increases, but
also large decreases in the proportion of aggressive trades. If trading aggressiveness decreases, price changes per one time interval of illiquid stocks are just too small, which
prevents the stock price to move towards its new equilibrium level. The negative effect of excessive increases in trading aggressiveness is more surprising, since quicker
price movements induced by aggressive trades should contribute to a quicker convergence process. However, this negative effect can be explained by the fact that excessive
increases in trading aggressiveness are mostly driven by uninformed traders, who submit a large number of aggressive orders in different directions. Since aggressive orders
produce larger price impact per one time interval, trading by uninformed investors unnecessarily increases intraday volatility and slows down the stabilization of the price
around its new equilibrium level. In situations when aggressive trading becomes more
informative and the majority of ISO trades are submitted in the direction of earnings
surprise, the negative effect of excessive trading aggressiveness is reduced and becomes
largely insignificant.
6
Conclusions
This paper analyzes how abnormal trading aggressiveness after earnings announcement
releases influences the speed of price adjustment of stocks in U.S. financial markets.
Trading aggressiveness is measured as the proportion of volume, traded with the most
aggressive limit orders, intermarket sweep orders, over a particular time interval. Intermarket sweep orders represent an exemption to the “Order Protection Rule” of the
Regulation National Market System and are executed more quickly than other limit
orders, but possibly at an inferior price. They produce larger intraday price impact
and contribute to quicker price changes per time interval.
The major result of this study is that excessive trading aggressiveness after earnings
announcements is overall harmful for the speed of price adjustment, especially for illiquid stocks. As compared to the pre-Reg NMS period, when aggressive orders were not
yet available, the adjustment time has increased for illiquid stocks with large deviations
in the proportion of ISO volume immediately after an earnings announcement release.
This finding can be expained by the fact that aggressive trades are mostly conducted
by uninformed investors, who trade on their heterogeneous beliefs. Since uninformed
investors are equally likely to buy or to sell, quick price changes in different directions
34
increase intraday volatility and make the price stabilization process more difficult.
The findings in this paper suggest that excessive use of intermarket sweep orders
produces adverse effects on the adjustment process of illiquid stocks after information
releases. Thus, market efficiency for these stocks can be even further reduced in situations when traders become too aggressive - something, that needs to be taken into
account by stock exchanges and market regulators, if they are interested in the promotion of accurate and transparent prices.
35
References
Ali, Ashiq, Sandy Klasa, and Oliver Zhen Li, 2008, Institutional stakeholdings and
better-informed traders at earnings announcements, Journal of Accounting and Economics 46, 47 – 61.
Amihud, Yakov, 2002, Illiquidity and stock returns: cross-section and time-series effects,
Journal of Financial Markets 5, 31–56.
Ball, Ray, and Philip Brown, 1968, An empirical evaluation of accounting income numbers, Journal of Accounting Research 6, 159–178.
Bamber, Linda Smith, Orie E. Barron, and Douglas E. Stevens, 2011, Trading volume
around earnings announcements and other financial reports: Theory, research design,
empirical evidence, and directions for future research, Contemporary Accounting Research 28, 431 – 471.
Bernard, Victor L., and Jacob K. Thomas, 1989, Post-earnings-announcement drift:
Delayed price response or risk premium?, Journal of Accounting Research 27, 1 – 36.
, 1990, Evidence that stock prices do not fully reflect the implications of current
earnings for future earnings, Journal of Accounting and Economics 13, 305 – 340.
Bessembinder, Hendrik, 2003, Issues in assessing trade execution costs, Journal of Financial Markets 6, 233–257.
Bradley, Daniel, Jonathan Clarke, Suzanne Lee, and Chayawat Ornthanalai, 2011, Information disclosure and intraday price discovery: Evidence from jumps, Working
Paper, AFA 2011 Denver Meetings.
Brooks, Raymond M., Ajay Patel, and Tie Su, 2003, How the equity market responds
to unanticipated events, Journal of Business 76, 109–133.
Busse, Jeffrey A., and T. Clifton Green, 2002, Market efficiency in real time, Journal
of Financial Economics 65, 415–437.
Chakravarty, Sugato, Pankaj Jain, Thomas McInish, and James Upson, 2011a, Trading
aggressiveness and market breadth around earnings announcements, Working Paper,
University of Texas at El Paso.
36
Chakravarty, Sugato, Pankaj Jain, James Upson, and Robert Wood, 2010, Clean sweep:
Informed trading through intermarket sweep orders, Working Paper, Purdue University.
Chakravarty, Sugato, James Upson, and Robert Wood, 2011b, The flash crash: Trading
aggressiveness, liquidity supply, and the impact of intermarket sweep orders, Working
Paper, Purdue University.
Coleman, Les, 2011, An exploratory analysis of factors influencing initial market response and media reports following shock corporate events, Financial Review 46,
313–336.
Damodaran, Aswath, 1993, A simple measure of price adjustment coefficients, Journal
of Finance 48, 387–400.
Defeo, Victor J., 1986, An empirical investigation of the speed of the market reaction
to earnings announcements, Journal of Accounting Research 24, 349–363.
Ederington, Louis H., and Jae Ha Lee, 1993, How markets process information: News
releases and volatility, Journal of Finance 48, 1161–1191.
, 1995, The short-run dynamics of the price adjustment to new information,
Journal of Financial and Quantitative Analysis 30, 117–134.
Fama, Eugene F., 1965, The behavior of stock-market prices, Journal of Business 38,
34–105.
Fleming, Michael J., and Eli M. Remolona, 1999, Price formation and liquidity in the
U.S. treasury market: The response to public information, Journal of Finance 54,
1901–1915.
Foster, George, Chris Olsen, and Terry Shevlin, 1984, Earnings releases, anomalies, and
the behavior of security returns, Accounting Review 59, 574–603.
Francis, Jennifer, Donald Pagach, and Jens Stephan, 1992, The stock market response
to earnings announcements released during trading versus nontrading periods, Journal of Accounting Research 30, 165–184.
French, Kenneth R., and Richard Roll, 1986, Stock return variances: The arrival of
information and the reaction of traders, Journal of Financial Economics 17, 5–26.
37
Garfinkel, Jon A., and Jonathan Sokobin, 2006, Volume, opinion divergence, and returns: A study of post–earnings announcement drift, Journal of Accounting Research
44, 85 – 112.
Gosnell, Thomas F., Arthur J. Keown, and John M. Pinkerton, 1996, The intraday
speed of stock price adjustment to major dividend changes: Bid-ask bounce and
order flow imbalances, Journal of Banking and Finance 20, 247–266.
Greene, Jason T., and Susan G. Watts, 1996, Price discovery on the NYSE and the
NASDAQ: The case of overnight and daytime news releases, Financial Management
25, 19–42.
Harris, Milton, and Artur Raviv, 1993, Differences of opinion make a horse race, Review
of Financial Studies 6, 473–506.
Hasbrouck, Joel, 1995, One security, many markets: Determining the contributions to
price discovery, Journal of Finance 50, 1175–1199.
He, Hua, and Jiang Wang, 1995, Differential information and dynamic behavior of stock
trading volume, Review of Financial Studies 8, 919–972.
Henker, Thomas, and Jian-Xin Wang, 2006, On the importance of timing specifications
in market microstructure research, Journal of Financial Markets 9, 162–179.
Hirshleifer, David A., James N. Myers, Linda A. Myers, and Teoh Siew Hong, 2008, Do
individual investors cause post-earnings announcement drift? Direct evidence from
personal trades., Accounting Review 83, 1521 – 1550.
Holthausen, Robert W., and Robert E. Verrecchia, 1990, The effect of informedness
and consensus on price and volume behavior, Accounting Review 65, 191–208.
Hong, Harrison, and Jeremy C. Stein, 1999, A unified theory of underreaction, momentum trading, and overreaction in asset markets, Journal of Finance 54, 2143–2184.
Jennings, Robert, and Laura Starks, 1985, Information content and the speed of stock
price adjustment, Journal of Accounting Research 23, 336–350.
Kaniel, Ron, Shuming Liu, Gideon Saar, and Sheridan Titman, 2012, Individual investor trading and return patterns around earnings announcements, Journal of Finance 67, 639 – 680.
38
Kaniel, Ron, Gideon Saar, and Sheridan Titman, 2008, Individual investor trading and
stock returns, Journal of Finance 63, 273–310.
Kim, Oliver, and Robert E. Verrecchia, 1991, Trading volume and price reactions to
public announcements, Journal of Accounting Research 29, 302 – 321.
Lee, Charles M.C., 1992, Earnings news and small traders: An intraday analysis, Journal of Accounting and Economics 15, 265–302.
Lee, Charles M. C., and Mark. J. Ready, 1991, Inferring trade direction from intraday
data, Journal of Finance 46, 733–746.
Lei, Adam Y.C., and Huihua Li, 2010, Still in the air: A clinical study of order strategies
in response to erroneous information, Working Paper, SSRN eLibrary.
Masulis, Ronald W., and Lakshmanan Shivakumar, 2002, Does market structure affect
the immediacy of stock price responses to news?, Journal of Financial and Quantitative Analysis 37, 617–648.
Patell, James M., and Mark A. Wolfson, 1984, The intraday speed of adjustment of stock
prices to earnings and dividend announcements, Journal of Financial Economics 13,
223–252.
Sarkar, Asani, and Robert A. Schwartz, 2009, Market sidedness: Insights into motives
for trade initiation, Journal of Finance 64, 375–423.
Smith, Brian F., Robert White, Michael Robinson, and Nason Richard, 1997, Intraday
volatility and trading volume after takeover announcements, Journal of Banking and
Finance 21, 337–368.
Woodruff, Catherine S., and A. J. Jr. Senchack, 1988, Intradaily price-volume adjustments of NYSE stocks to unexpected earnings, Journal of Finance 43, 467–491.
39
Figure 3: Changes in Proportion of ISO volume on an Announcement Day. This figure
displays mean percentage deviations in the proportion of ISO volume throughout an announcement
day. Deviations from the bootstrapped means are measured in 15-minute intervals relative to the
15-minute interval with an earnings announcement release (interval 0). The dashed line shows the 1%
significance level for the mean percentage change in the proportion of ISO volume, equal to 3.8%.
40
Figure 4: Trading imbalances in ISO volume on an Announcement Day. This figure displays
the mean proportion of ISO buy volume, defined as ISO buy volume/T otal buy volume, and the mean
proportion of ISO sell volume, defined as ISO sell volume/T otal sell volume, throughout an announcement
day. Both proportions are measured in 15-minute intervals relative to the 15-minute interval with an
earnings announcement release (interval 0). The dashed line marks the event interval.
A. Positive Earnings Surprises
B. Negative Earnings Surprises
41
Figure 5: Adjustment Time and Trading Aggressiveness. This figure depicts the relationship
between the mean length of the price adjustment period and the mean change in the proportion of ISO
volume after an information release, separately for samples of liquid and illiquid stocks. The length of
the price adjustment period is measured as the number of ten-minute time intervals until volatility of
one-minute midpoint returns is no longer abnormal.
42
Table 1: Variable Definitions. This table defines all variables in this paper and indicates
databases, used for their construction.
Variable
Description
Source
1/P
The inverse of a stock price (in $)
TAQ
∆ISOvol
Change in the proportion of daily volume, executed through
TAQ
aggressive intermarket sweep orders (ISO), after an announcement release relative to its mean in the base period
Adjustment
The number of 10-minute intervals from an announcement
Own
period
time until the interval, in which intraday volatility of 1-
calculations
minute returns (Intr Volat) is no longer abnormal
Amihud
Amihud’s measure of illiquidity, defined as the ratio of the
CRSP
daily absolute return to the dollar trading volume on that
day (Amihud, 2002).
Big Neg Surp
1, if a 24-hour post-announcement return is negative and
TAQ
below the median of all negative earnings announcements;
zero otherwise
Big Pos Surp
1, if a 24-hour post-announcement return is positive and
TAQ
above the median of all positive earnings announcements;
zero otherwise
EarnSurp
An absolute value of a stock return in 24 hours after an
TAQ
announcement release
EffSpr
The effective relative spread, calculated as twice the abso-
TAQ
lute difference between the transaction price and the midpoint price, scaled by the midpoint price (Ef f Sprt =
2 |Pt − Qt | /Qt ). Observations with Ef f Spr > 0.5 are set
to missing values
Eventi ,
1 for observations on event day i, where i is calculated as
i [−2; +2]
Current Day - Announcement Day
Illiquid (Illiq)
1, if the relative spread of a stock is above its median value
I/B/E/S
TAQ
of all stocks in the sample.
Intr Volat
Intraday volatility, measured as the standard deviation of
one-minute closing midpoint returns (in %)
43
TAQ
Variable
Description
Source
ISO
1, if a transaction was marked as intermarket sweep, and
TAQ
zero otherwise
Leverage
Market leverage, defined as the ratio of total liabilities to
Compustat
the sum of total liabilities and market capitalization of the
company.
Liquid (Liq)
1, if the relative spread of a stock is below its median value
TAQ
of all stocks in the sample.
LnMCap
Natural logarithm of market capitalization
CRSP
MCap
Market value of equity (in million $)
CRSP
Nasdaq
1, if a stock is listed on Nasdaq, and zero otherwise
TAQ
Neg Surp
1, if a 24-hour post-announcement return is negative; zero
TAQ
otherwise
Pos Surp
1, if a 24-hour post-announcement return is negative; zero
TAQ
otherwise
Post-Reg NMS
1 after the final implementation of the Regulation NMS (October 2007), and zero otherwise
Pre-Reg NMS
1 before the final implementation of the Regulation NMS
(October 2007), and zero otherwise
Prc
Stock price (in $)
CRSP
PrcImp
The measure of price impact of each trade after 5 minutes,
TAQ
defined as P rcImpt = 2 |Qt+5 − Qt | /(Qt ∗ wt ), where Qt+5
is the midpoint price of a stock after five minutes and wt is
the size of a trade
Proportion of
The ratio of the number of intermarket sweep orders to the
ISO trades,
total number of orders executed in a particular time interval
TAQ
%Trades
Proportion of
The ratio of the volume, executed through intermarket sweep
ISO volume,
orders, to the total volume traded in a particular time inter-
%ISOvol
val
Proportion of
The ratio of the number of purchase transactions executed
ISO purchases,
through intermarket sweep orders to the total number of
%Purchases
purchase transactions in a particular time interval
44
TAQ
TAQ
Variable
Description
Source
Proportion of
The ratio of the volume of purchase transactions executed
TAQ
ISO buy volume,
through intermarket sweep orders to the total volume of pur-
%ISOBuyV ol
chase transactions in a particular time interval
Proportion of
The ratio of the number of sale transactions executed
ISO sales,
through intermarket sweep orders to the total number of
%Sales
sale transactions in a particular time interval
Proportion of
The ratio of the volume of sale transactions executed through
ISO sell volume,
intermarket sweep orders to the total volume of sale trans-
%ISOSellV ol
actions in a particular time interval
RelSpr
Intraday relative spread, defined as the difference between
TAQ
TAQ
TAQ
the ask and the bid, scaled by their average; observations
with RelSpr>0.5 are set to missing values.
RelSpr (daily)
Daily relative spread, defined as the difference between the
CRSP
closing ask and the closing bid, scaled by their average; observations with RelSpr (daily)>0.5 are set to missing values.
ROA
Return on assets, defined as the ratio of operating income
Compustat
after depreciation to average total assets of the current year
and the previous year.
Size
Size of a transaction (in shares)
TAQ
TAi
ith tercile of trading aggressiveness (TA1 - the lowest tercile
Own
of trading aggressiveness and TA3 - the highest tercile of
calculations
trading aggressiveness)
Total Assets
Total assets (in mln $)
Compustat
Total Liabilities
Total liabilities (in mln $)
Compustat
Turnover
Average daily traded volume, divided by the number of
CRSP
shares outstanding
Volatility
Annualized standard deviation of daily stock returns over
CRSP
the calendar month
Volume
Total volume traded within a 10-minute interval (in shares)
45
TAQ
Table 2: Sample Construction. This table shows the sample selection of earnings announcements
in US firms that happened within trading hours (from 9:30 a.m. till 16:00 p.m. EST) over 2006-2009.
Data source for dates and times of earnings announcements is the Institutional Brokers Estimate
System (I/B/E/S) database. Each firm is required to exist in the intersection set of I/B/E/S and
CRSP.
Criteria
Announcements
Lost
obs.
Firms
Initial sample
10,334
Stock was traded on an announcement
day
9,687
647
3,273
Intraday transaction data available on
TAQ
8,720
967
3,008
Closing price not less than $5
6,126
2,594
2,334
Not more than one announcement per
day
6,040
86
2,322
Trading data exists for previous 2
months
5,944
96
2,307
At least one announcement before and
one announcement after the Reg
NMS, out of which:
3,613
2,331
675
- Before the Reg NMS
1,818
675
- After the Reg NMS
1,795
675
46
3,361
Table 3: Sample Distributions. This table displays distributions of firm characteristics in the
final sample (Columns 1-3) and the initial sample (Columns 4-6). See Table 1 for the exact definition
of all variables.
Final
Total Assets (in mln $)
Total Liabilities (in mln $)
MCap (in mln $)
Prc (in $)
ROA
Leverage
RelSpr (daily)
Amihud
Volatility
Turnover
Initial
N
Mean
50%
N
Mean
50%
666
666
675
675
654
666
675
675
675
675
8672
5990
2670
26
0.07
0.54
0.01
0.95
0.44
0.006
686
507
256
21
0.05
0.55
0.00
0.04
0.41
0.003
3300
3300
3361
3361
3054
3290
3361
3361
3361
3361
6220
4309
2254
20
-0.01
0.46
0.01
1.93
0.59
0.007
477
274
239
14
0.04
0.41
0.01
0.06
0.53
0.005
47
Table 4: Differences in Characteristics of ISO and non-ISO Orders. This table summarizes
differences in trading characteristics between intermarket sweep orders (ISO) and standard limit orders
(non-ISO) in the base period and after earnings announcement releases. Columns 2 and 3 display the
mean of the bootstrapped distribution for ISOs and non-ISOs from the base period, correspondingly.
Columns 5 and 6 report the cross-sectional mean of respective variables starting from an announcement
release until the end of the trading day (16:00 p.m. EST). Column 4 and Column 7 show the p-value
of the t-test for the null-hypothesis that the difference in means between ISOs and non-ISOs equals
zero. * denotes statistical significance at the 10% level, ** denotes statistical significance at the 5%
level, *** denotes statistical significance at the 1% level. The last two columns display the differencein-differences (Column 7 - Column 4) and report the p-value of the two-tailed t-test for the hypothesis
that it equals zero. For proportions, they display simply the difference between the base period and the
event day. All variables are calculated from the intraday transaction data in the NYSE TAQ database.
See Table 1 for the exact definition of all variables.
Base Period
ISO
%Trades
%Volume
%Purchases
%Sales
Size
EffSpr, %
PrcImp, %
40.8
37.8
40.5
41.3
176
1.73
1.29
Event Day
Non- (2)-(3)
ISO
ISO
45.4
42.4
44.4
45.5
180
1.84
1.47
256 ***
1.82 *
1.20 **
48
NonISO
Diff-in-Diff
(5)-(6)
226 ***
1.86
1.27 ***
(7)-(4)
4.6
4.6
4.0
4.2
33
0.08
0.11
***
***
***
***
***
**
Table 5: Determinants of ISO trading volume. Panel A examines the determinants of ISO
volume in the base period and after earnings announcement releases. Models (1) and (2) report results
of cross-sectional OLS regressions with the mean proportion of daily volume traded with aggressive
orders in the base period as the dependent variable (%ISOvol ). The dependent variable in Models
(3) and (4) is the percentage deviation in the proportion of ISO volume on an announcement day
from its value in the base period (∆ISOvol). See Table 1 for the exact definition of all variables.
P-values of the two-tailed t-test with the null-hypothesis of a coefficient equaling zero are reported in
form of asterisks to the right of each coefficient. * denotes statistical significance at the 10% level, **
denotes statistical significance at the 5% level, *** denotes statistical significance at the 1% level. I
also report the number of observations (N ) and R2 for each regression. Panel B summarizes differences
in trading characteristics between intermarket sweep orders (ISO) and standard limit orders (non-ISO)
after earnings announcement releases for stocks with different liquidity levels. The layout is the same
as in Table 4.
Panel A: Determinants of ISO volume
LnMCap
RelSpr
Amihud
EarnSurp
Nasdaq
Constant
N
R-squared
Weekday FE
Daytime FE
Year FE
(1)
%ISOvol
0.021 ***
0.0009
(2)
%ISOvol
0.020 ***
(3)
∆ISOvol
0.003
1.95 **
-0.0000
0.141 ***
0.173 ***
1771
0.25
No
No
Yes
0.142 ***
0.184 ***
1771
0.25
No
No
Yes
0.53 **
-0.002
-0.142
1768
0.02
Yes
Yes
Yes
(4)
∆ISOvol
-0.008
0.01 ***
0.56 **
0.004
-0.018
1768
0.02
Yes
Yes
Yes
Panel B: Liquid vs Illiquid on Event Day
Liquid
ISO
%Trades
%Volume
Size
EffSpr, %
PrcImp, %
45.0
41.6
163
0.57
0.64
Illiquid
Non- (2)-(3)
ISO
ISO
45.9
43.3
200
3.43
2.51
200 ***
0.56
0.62
49
Non- (5)-(6)
ISO
257 ***
3.38
2.04 ***
Diff-in-Diff
(7)-(4)
0.9 *
1.8 ***
-20 ***
0.05
0.45 ***
Table 6: Regression Analysis of the Effective Spread and Price Impact. This table presents
results of panel OLS regressions with the effective relative spread as the dependent variable for Models
1 and 2 and the price impact as the dependent variable for Models 3 and 4. One observation represents
a 10-minute trading interval for a stock. All regressions include firm-, year-, daytime- and weekdayfixed effects. See Table 1 for the exact definition of all variables. P-values of the two-tailed t-test with
the null-hypothesis of a coefficient equaling zero are reported in form of asterisks to the right of each
coefficient. * denotes statistical significance at the 10% level, ** denotes statistical significance at the
5% level, *** denotes statistical significance at the 1% level. I also report the number of observations
(N ) and R2 for each regression.
ISO
Illiquid
ISO · Illiquid
Event−2
Event−1
Event
Event+1
Event+2
ISO · Event−2
ISO · Event−1
ISO · Event
ISO · Event+1
ISO · Event+2
1/P
Volume
Nasdaq
N
R-squared
Weekday FE
Daytime FE
Year FE
Firm FE
(1)
EffSpr
-0.003
0.000
0.062
0.049
-0.015
-0.034
0.010
-0.012
0.020
0.001
-0.003
7.504
-0.000
1.413
1489118
0.33
Yes
Yes
Yes
Yes
***
***
***
***
***
***
(2)
EffSpr
-0.003
0.664
0.001
0.003
0.064
0.051
-0.011
-0.030
0.010
-0.011
0.021
0.001
-0.003
6.768
-0.000
1.352
1489118
0.34
Yes
Yes
Yes
Yes
50
(3)
PrcImp
0.053 ***
***
***
***
**
***
***
***
0.004
0.064
0.091
0.004
-0.000
-0.007
0.002
0.036
0.003
-0.014
19.374
0.000
1.197
1489118
0.31
Yes
Yes
Yes
Yes
***
***
**
***
***
***
(4)
PrcImp
0.016
0.623
0.187
0.007
0.066
0.097
0.009
0.005
-0.007
0.004
0.028
-0.000
-0.015
18.593
0.000
1.122
1489118
0.32
Yes
Yes
Yes
Yes
***
***
***
***
***
***
***
***
Table 7: Informativeness of ISO Trades. This table reports differences in means between an
increase in the proportion of ISO buy volume and an increase in the proportion of ISO sell volume
on an announcement day, separately for liquid stocks (Panel A) and illiquid stocks (Panel B). The
first column shows the number of hours since an announcement release. Columns 2 and 3 report
results for positive surprises and large positive surprises, respectively. Columns 4 and 5 report the
corresponding results for negative surprises. An earnings surprise is measured as a 24-hour stock return
since an announcement release. It is classified as large if a 24-hour stock return is above its median
for positive earnings surprises and below its median for negative earnings surprises. T-statistics of
the two-tailed t-test with the null-hypothesis of a difference in means equaling zero are in parentheses
below each coefficient. P-values are reported in form of asterisks to the right of each coefficient. *
denotes statistical significance at the 10% level, ** denotes statistical significance at the 5% level, ***
denotes statistical significance at the 1% level.
Panel A: Liquid Stocks
Hour
1
2
3
4
5
Pos Surp
Big Pos Surp
Neg Surp
Big Neg Surp
0.47%
(0.49)
0.86%
(1.09)
0.90%
(1.23)
0.79%
(1.15)
0.55%
(0.82)
0.80%
(0.69)
0.74%
(0.76)
1.22%
(1.43)
1.43%
(1.79)
1.45%
(1.84)
-0.45%
(-0.45)
-0.13%
(-0.17)
-0.33%
(-0.44)
-0.13%
(-0.19)
0.03%
(0.05)
-0.51%
(-0.43)
-0.11%
(-0.12)
-0.39%
(-0.45)
-0.17%
(-0.20)
-0.27%
(-0.34)
*
*
Panel B: Illiquid Stocks
Hour
1
2
3
4
5
Pos Surp
Big Pos Surp
-0.73%
(-0.34)
0.91%
(0.44)
1.72%
(0.90)
1.27%
(0.73)
0.86%
(0.50)
1.85%
(0.59)
3.61%
(1.26)
3.76%
(1.42)
3.74%
(1.53)
3.03%
(1.28)
Neg Surp
-6.13% **
(-2.12)
-6.51% ***
(-2.76)
-4.95% **
(-2.21)
-6.03% ***
(-2.95)
-5.67% ***
(-2.78)
51
Big Neg Surp
-5.21%
(-1.43)
-7.11% **
(-2.41)
-7.32% **
(-2.59)
-7.00% ***
(-2.75)
-6.86% ***
(-2.70)
Table 8: The Length of the Price Adjustment Period: Summary Statistics and Univariate
Analysis. Panel A of this table presents distributions of the length of the price adjustment period,
separately for samples of liquid and illiquid stocks, in the pre- and the post-Reg NMS periods. See Table
1 for the exact definition of all variables. 1 unit corresponds to a 10-minute interval. The significance
level of the non-parametric Mann-Whitney test on the equality of medians of pre- and post-Reg NMS
distributions is reported in row 3. Panel B displays the mean adjustment time across different terciles
of trading aggressiveness (TA1 - TA3) in the post-Reg NMS period. Trading aggressiveness is measured
as the change in the proportion of ISO volume traded after an announcement release relative to its mean
in the base period (∆ISOvol). TA1 includes stocks with the lowest increases in trading aggressiveness
on an announcement day and TA3 includes stocks with the highest increases. The announcements in
the pre-Reg NMS period are divided in “pseudo” - TA terciles, which equal the median TA tercile in
the post-Reg NMS period. Columns 4 and 7 report the significance level of the t-test on the equality
of means between pre- and post-Reg NMS periods for liquid and illiquid stocks, correspondingly. The
last two rows report the significance level of the t-test on the equality of means across different terciles
of trading aggressiveness. * denotes statistical significance at the 10% level, ** denotes statistical
significance at the 5% level, *** denotes statistical significance at the 1% level.
Panel A: Pre-Reg NMS vs Post-Reg NMS
Liquid
5%
25%
50%
75%
95%
Mean
Std
Pre-Reg
NMS
Post-Reg
NMS
0
6
18
45
94
29.33
30.23
0
5
14
34
89
25.37
28.40
Illiquid
MW-test
**
Pre-Reg
NMS
Post-Reg
NMS
1
11
34
79
108
44.52
36.59
1
10
39
84
110
47.80
38.51
MW-test
Panel B: Liquidity and Aggressiveness Terciles
Liquid
TA1
TA2
TA3
Pre-Reg
NMS
Post-Reg
NMS
28
29
29
23
25
25
Illiquid
t-test
**
TA1-TA3
TA2-TA3
Pre-Reg
NMS
Post-Reg
NMS
47
43
46
50
42
54
***
52
t-test
**
Table 9: Length of the Price Adjustment Period: Difference-in-Differences Analysis. This
table presents results of negative binomial regressions, including observations from the pre- and the
post-Reg NMS periods. The dependent variable in each model is the number of 10-minute periods until
the price adjusts to its new equilibrium level. Model (1) reports results for the total sample. Models
(2) to (5) present results separately for (big) positive and (big) negative earnings surprises. See Table
1 for the exact definition of all variables. P-values of the two-tailed t-test with the null-hypothesis of
a coefficient equaling zero are reported in form of asterisks to the right of each coefficient. * denotes
statistical significance at the 10% level, ** denotes statistical significance at the 5% level, *** denotes
statistical significance at the 1% level. I also report the number of observations (N ) and the p-value
of the Likelihood-ratio test with the null hypothesis that the dispersion parameter is zero.
# 10-min Periods
Post Reg
Illiq
Illiq· Post Reg
(1)
Total
(2)
Pos
Surp
(3)
Big
Pos
Surp
(4)
Neg
Surp
(5)
Big
Neg
Surp
-0.15 *
-0.10
-0.59 ***
-0.18
-0.42 **
0.29 ***
-0.10
0.30 ***
-0.14
0.13
0.29 ***
0.17
0.11
-0.10
-0.03
-0.03
0.17
0.33 *
Liq·∆ISOvol
0.20 **
0.22 *
Illiq· |∆ISOvol|∆>0
0.24 ***
0.34 ***
0.29 **
0.14
0.09
Illiq· |∆ISOvol|∆<0
0.25 **
0.46 ***
0.32
0.07
0.12
Turnover
-7.53 **
-10.41 **
-7.95
-5.16
-8.52
Intr Volat
0.06 *
0.05
-0.03
0.09
0.02
LnMCap
0.00
0.01
-0.01
0.00
-0.02
3,613
0.000
Yes
Yes
Yes
1,822
0.000
Yes
Yes
Yes
910
0.000
Yes
Yes
Yes
1,791
0.000
Yes
Yes
Yes
895
0.001
Yes
Yes
Yes
N
P(Chi-Squared)
Weekday FE
Daytime FE
Year FE
53
Table 10: Length of the Price Adjustment Period: Robustness Checks. This table presents
results of negative binomial regressions, including observations from the pre- and the post-Reg NMS
periods. The dependent variable in each model is the number of 10-minute periods until intraday
volatility is no longer abnormal. Models (1)-(3) use an alternative definition of intraday volatility,
which is measured as the absolute price range in each 10-minute interval. Models (4)-(6) use the
standard definition of intraday volatility (standard deviation of one-minute closing midpoint returns),
but differentiate between liquid and illiquid stocks with the Amihud (2002) illiquidity measure. Models
(1) and (4) report results for the total sample. Models 2 and 5 present results for positive earnings
surprises, and Models 3 and 6 - for negative earnings surprises. See Table 1 for the exact definition
of all variables. P-values of the two-tailed t-test with the null-hypothesis of a coefficient equaling zero
are reported in form of asterisks to the right of each coefficient.
(1)
Total
# 10-min Periods
Post Reg
Illiq
-0.14 *
0.27 ***
Price Range
(2)
(3)
Pos
Neg
Surp
Surp
-0.11
-0.17
0.23 **
0.33 ***
(4)
Total
Amihud
(5)
Pos
Surp
(6)
Neg
Surp
-0.16 *
-0.12
-0.19
0.31 ***
-0.13
-0.20
-0.10
Liq·∆ISOvol
0.14
0.15
0.12
0.22 **
0.22 *
0.20
Illiq· |∆ISOvol|∆>0
0.23 ***
0.34 ***
0.14
0.25 ***
0.34 ***
0.18 *
Illiq· |∆ISOvol|∆<0
0.23 *
0.44 **
0.07
0.31 ***
0.47 ***
0.17
-4.79
Intr Volat
0.10 ***
LnMCap
0.00
3,613
0.000
Yes
Yes
Yes
N
P(Chi-Squared)
WeekdayFE
DaytimeFE
YearFE
-7.40
-2.10
0.09 *
-7.01 **
-0.12
0.30 ***
Illiq· Post Reg
Turnover
-0.10
0.31 ***
-9.70 **
-0.12
-4.83
0.13 **
0.06
0.04
0.09
-0.01
0.01
0.01
0.01
0.00
1,822
0.000
Yes
Yes
Yes
1,791
0.000
Yes
Yes
Yes
3,613
0.000
Yes
Yes
Yes
1,822
0.000
Yes
Yes
Yes
1,791
0.000
Yes
Yes
Yes
54