See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/228261355

Intraday Trading Patterns and Commonality in Liquidity

Article in SSRN Electronic Journal · September 2011

DOI: 10.2139/ssrn.1925826

CITATIONS READS
2 632

2 authors:

Mehmet F. Dicle Tarun K. Mukherjee

Loyola University New Orleans University of New Orleans
64 PUBLICATIONS 399 CITATIONS 76 PUBLICATIONS 1,929 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Tarun K. Mukherjee on 08 October 2020.

The user has requested enhancement of the downloaded file.

Intraday Trading Patterns and Commonality in Liquidity

Mehmet F. Dicle
Assistant Professor of Finance
Loyola University New Orleans
[email protected]

Tarun Mukherjee
James R. Moffett Chair in Financial Economics
Department of Economics and Finance
University of New Orleans
[email protected]

September, 2011

Electronic copy available at: http://ssrn.com/abstract=1925826

 Intraday Trading Patterns and Commonality in Liquidity

Abstract

A common thread in literature that connects liquidity commonality and intraday trading patterns
lies in the potential sources of these two phenomena: inventory risk and asymmetric information.
There is ample evidence in literature on trading patterns to suggest that both volume and return
volatility are highest during the first and last half hours of a trading day. Thus, one would expect
that liquidity commonality would be most concentrated during the intraday trading periods when
the trading volume and return volatility are most intense. Therefore, we hypothesize that the
degree of liquidity commonality is highest during these two trading periods. We present evidence
in support of this hypothesis.

Keywords: Liquidity, commonality, U-shape, intraday trading

Electronic copy available at: http://ssrn.com/abstract=1925826

 Intraday Trading Patterns and Commonality in Liquidity

Introduction

Commonality in liquidity is defined as the impact of market liquidity on individual

stock’s liquidity. A positive commonality coefficient means, as market liquidity increases on a

trading day so will the stock's liquidity. This concerted liquidity movement has important

implications for asset pricing, portfolio management and exploring reasons of market crashes.

Based on microstructure Chordia, Roll, and Subrahmanyam (CRS) (2000) identify two

potential sources of common liquidity movements. These two not mutually exclusive reasons are

inventory risk and asymmetric information. A principal factor that affects inventory risk is the

trading volume. According to CRS (2000, pg. 5), "Since trading volume is a principal

determinant of dealer inventory, its variation seems likely to induce co-movements in optimal

inventory levels which lead in turn to co-movements in individual bid-ask spreads, quoted depth,

and other measures of liquidity."

There is ample evidence in the literature (for example, Jain and Joh,1988 and McInish

and Wood, 1990) suggesting that trading volume of New York Stock Exchange is not uniform

during any given trading day. For example, Jain and Joh find that average trading volume is

highest in the first trading hour of the day and then declines monotonically until the fourth hour

before increasing in the fifth and sixth hour. In a similar vein, Wood, McInish and Ord (1985)

report that all positive returns are earned during the first and last thirty minutes of a trading day.

Several models (Scott and Whaley, 1990, Brock and Kleidon, 1992, Madhavan, 1992, Admati

and Pfleiderer, 1988, Chan, Christie, and Schultz, 1995, and Coughenour and Saad 2004) exist to

explain intraday trading patterns. Admati and Pfleiderer (1988) offer reasons as to why heavy

Electronic copy available at: http://ssrn.com/abstract=1925826

trading volume and narrowing of bid-ask spread simultaneously occur at the opening and close

of a trading day. They suggest that open and close fall just after and before the market when it is

difficult or impossible to trade. This situation leads to an increase in nondiscretionary liquidity

during the two periods, resulting in trading concentration by discretionary liquidity (as well as

informed) traders during these periods. Further, by concentrating their trades in the two periods,

discretionary traders minimize the adverse selection costs facing specialists resulting in the

simultaneous occurrence of heavy trading volume and narrow spreads. Supporting the Admati

and Pfleiderer (1988) prediction, Foster and Viswanathan (1993) show that asymmetric

information is higher at the beginning and end of the day. Blau, Van Ness and Van Ness (2009,

pg. 2) suggest: “Together with the ex ante theories explaining the intraday pattern of returns and

trading activity, stealth trading suggests that smaller trades will move prices when volume is low

because informed traders do not want to reveal their information to the market. During periods of

high volume, informed traders are able to increase their trade sizes because high volume

provides a sufficient disguise for their information."

A common link that connects liquidity commonality and intraday trading patterns are

inventory risk and information asymmetry. Thus one would expect trading patterns and liquidity

commonality to go hand-in-hand. In other words, liquidity commonality should be more intense

at the opening and closing of the market. This is the issue we embrace in this paper.

Our sample consists of intraday trading volume of NYSE composite index for September

2009 - June 2010 period. We partition each trading day into 13 half-hour periods (HHP).

Intraday trading patterns are consistent with those found in the literature. In determining

commonality in liquidity, we use both daily aggregate variables similar to previous studies and

HHP aggregate variables. Coughenour and Saad (2004) provide evidence that using intraday

2
aggregates of liquidity measures the level of liquidity commonality is significantly higher

compared to the level provided by the studies that use daily aggregated liquidity measures such

as Chordia, Roll, and Subrahmanyam (2000). There could be two reasons for this increased

commonality level. First, intraday trading patterns are market-wide and thus reasons by

themselves for common liquidity movements. Daily averaging of liquidity measures smoothes

intraday trading patterns in liquidity (Coughenour and Saad, 2004). Thus, the level of

commonality in liquidity evident in daily studies cannot be explained by the intraday trading

patterns. This means that intraday trading patterns cannot be the reasons by themselves for

commonality in liquidity. Second, the reasons of intraday trading patterns are the same as

commonality in liquidity and changes in these reasons are more pronounced intraday. Consistent

with our hypothesis, we find a high level of commonality in liquidity for the first and the last

HHPs of the day and a low level of commonality for the rest of the day. These results are in line

with our expectations and robust to size, sector and different data filters. These results are based

on publicly available 5-minutes interval intraday data.

A brief summary of literature about the intraday trading patterns as well as commonality

in liquidity is provided within the next section. The data, variables and the empirical models are

explained in the third section. Empirical results are followed by the concluding remarks. Tables

providing descriptive statistics and estimation results are included in the appendix.

Literature Review

Earlier studies providing evidence of commonality include Chordia, Roll, and

Subrahmanyam (2000) (hereafter CRS), Hasbrouck and Seppi (2001) and Huberman and Halka

(2001). Following their evidence, literature focus mostly on sources, whether commonality in

3
liquidity is priced, whether similar patterns exists for other markets and with other markets.

Acharya and Pedersen (2005), for instance, argue that such commonality is priced. There is also

ample evidence that commonality in liquidity is not specific to NYSE and in fact exists at higher

levels at international equity markets (i.e. Domowitz and Wang, 2002; Fabre and Frino, 2004;

Friederich and Payne, 2002; Galariotis and Giouvris, 2007; Brockman, Chung and Perignon,

2008).

International commonality evidence is important for several reasons. Commonality in

liquidity is argued to be due to institutional trading behaviors (i.e. trade portioning) or their

consequences (i.e. herding) as well as asymmetric information and inventory risks (i.e. CRS).

Coughenour and Saad (2004) argue that the liquidity co-movement is due to market maker

portfolios including multiple stocks and providing liquidity for them. Both CRS and Coughenour

and Saad (2004) provide evidence for their arguments using NYSE stocks. Fabre and Frino

(2004), on the other hand, use stocks listed on Australian Stock Exchange (ASX) which is an

order-driven market, and find significantly lower levels of commonality in liquidity. This

international evidence based on market structure differences between NYSE and ASX is a

supporting evidence for the inventory risk explanations for the existence of market-wide liquidity

movements.

Coughenour and Saad (2004) recognize the intraday liquidity patterns and control by

aggregating the variables for the first hour, for the last hour and for the rest of the day. Their

mean R2 is about 22% while it is about 1% for CRS. They attribute this difference to the

utilization of three intraday periods to calculate liquidity variables. Also the average level of

commonality is above 90% for Coughenour and Saad (2004) while this level is about 30% for

CRS.

4
 One of the important implications of common liquidity movements is the potential to

understand the channels in which market crashes may occur. Another important implication is

the impact of asymmetric information as well as institutional trade portioning on the overall

market. It is argued in the previous literature (ex. Barclay and Warner, 1993; Chakravarty, 2001;

Hansch and Choe, 2007) that informed traders portion their large trades which is usually referred

to as 'stealth trading'. This practice is in an attempt to keep their information hidden. Thus, if

there are too many informed trades that hide their information through stealth trading, then the

number of trades will be higher. Accordingly, Jones et al. (1994) suggest number of trades as a

measure of firm level asymmetric information.

The question then is whether the level of asymmetric information and level of common

liquidity movements are correlated. Admati and Pfleiderer (1988) argue that the "…discretionary

liquidity trading (as well as informed trading) will also be concentrated" at the open and the

close of the market. Foster and Viswanathan (1993) argue that the level of asymmetric

information is in line with the intraday trading patterns, specifically higher open and close

periods.

Wood, McInish and Ord (1985) is one of the earlier studies to show the intraday trading

patterns. Later studies also show similar U-shaped trading patterns (ex. Admati and Pfleiderer,

1988; Jain and Joh, 1988; Lockwood and Linn, 1990; McInish and Wood, 1990) for different

time periods. Wei (1992) argues that "…trading activity, price variability, the information

component, and the proportion of block trades are the highest in the first period." This is in

support of the Admati and Pfleiderer (1988)'s argument. Stoll and Whaley (1990) also argue in

favor of the information asymmetry to be one of the reason of the intraday trading patterns.

Accordingly, if the specialists have knowledge of 'order imbalances' at the market open, then

5
specialists would try to capitalize on this information asymmetry. Chan, Christie and Schultz,

1995 further argue that 'institutional factors' should be considered when testing "…the important

of information asymmetries in determining intraday spreads…".

The contribution of the present study is to provide evidence for the hypothesis that the

liquidity commonality and the intraday trading patterns have the same sources namely

information asymmetry and inventory risks.

Data and Empirical Model

The data for the study includes all securities included in the NYSE composite index for

the period between September, 2009 and June, 2010 at 5-minutes intervals (MI). Each

observation includes bid, ask, last trade and (aggregating) volume variables. Data source is

Yahoo! Finance1. Following McInish and Wood (1992), each trading day is split into 13 half-

hour periods (HHP). For each MI, return2 (
 ), proportional spread3 (
 )4, change in

proportional spread5 (∆
 ) and volume6 (  ) are calculated. While there are several

liquidity measures in the literature we use proportional spread because it is the intersecting

liquidity measure of commonality in liquidity and intraday trading patterns literatures. For each

HHP, total return (
 ), standard deviation of return (, ), mean proportional spread

(
 ), total change in proportional spread (∆
 ) and total volume (  ) are

calculated. In order to standardize the volume for each HHP across securities (  ), each

1
Yahoo! Finance is available through http://finance.yahoo.com/. Yahoo! Finance uses quote data for NYSE directly
from the exchange (http://help.yahoo.com/l/us/yahoo/finance/quotes/fitadelay.html).
2

 = ln (
 /
 )
3

 = ("# − % )⁄(("# + % )/2)
4
McInish and Wood (1992) refer to this liquidity measure as percentage bid-ask spread and Chordia et. al. (2000)
refers to it as proportional quoted spread.
5
∆
 = ln (
 /
 )
6
 =  − 

6
stocks' HHP volumes are divided by the total daily volume for the stock. For each day, total

return (
)*+ ), standard deviation of return (,)*+ ) and total change in proportional spread

(∆
)*+ ) are calculated.

Some of the most common filters are employed in this study as well. These include; bid-

ask spread cannot be more than 40% of the price; each observation has to have ask, bid, last trade

price and volume variables; each observation has to be in sequence; each day has to have average

price above $2. Observations after closing and out of sequence observations are excluded. In

order to obtain statistically viable degrees of freedom, we require stocks to have at least 40 days

of trading. In order not to contaminate the sequence of first and last HHP of the day, half day

trading days are excluded. After the filters there are 27,775,621 observations. For each stock

there are on average; 5.77 observations per HHP, 74.95 observations per day for 205 days. There

are, on average, 1693 stocks per day.

For the intraday trading patterns, mean proportional spread, total volume, total return and

standard deviation of return variables are used for each HHP and for each stock. Cross sectional

(across stocks) averages of these variables for each HHP are used to establish the intraday

trading patterns. Statistical significance of the intraday patterns is tested by comparing the mean

for a specific HHP to the mean for the following HHP. Thus, a difference in means t-test is

employed for the means for consecutive HHPs. Last HHP of the day is tested with the first HHP

of the same day. While the difference is tested to be statistically significantly different than zero,

it is also tested whether it is greater or less than zero to establish the pattern.

Commonality in liquidity is estimated following CRS as follows;

7
 1

∆
,-, = . + / 01, 2_∆
 + 4, 2_
,-+, 5 + 
,,-, + 7,-,

,-+,
=−1

Equation 1: Commonality in liquidity using daily variables

For Equation 1, "M_" refers to the equally weighted market variables excluding the

security for which the equation is estimated. Equation 1 is estimated for each security (i) within

the sample. Equation 1 is also estimated using the equally weighted sector variables. Based on

the evidence with the descriptive statistics about the intraday trading patterns, commonality is

estimated for each HHP to evaluate the effect of intraday trading patterns on the market level of

statistically significant commonality in liquidity. Equation 1 is amended as follows;

∆
)*+,, = .88, + / 9188,, 2_∆
,-+,88, + 488,, 2_
)*+:;,, <

=−1

+ 
,,-,88, + 7)*+,,

Equation 2: Commonality in liquidity using HHP variables

Equation 2 is estimated for each of the 13 HHPs. It is important to note that the time

series is based on a specific HHP across days. For instance, Equation 2 is estimated for the first

HHP of the day for all days.

Empirical Results

Table 1 provides the descriptive statistics across securities for HHP variables including

volume (  ), proportional spread (
 ), return (
 ) and standard deviation of

return (, ). Figure 1 provides the graphs for the cross-sectional means provided with the

Table 1. The reported graphs are in line with the previously reported evidence. The results

provided with the Table 1 and Figure 1 is evidence for the intraday trading patterns.

8
 Intraday pattern of volume is similar to the evidence provided by the previous studies (i.e.

Wood, McInish and Ord, 1985; Admati and Pfleiderer, 1988; Jain and Joh, 1988; Lockwood and

Linn, 1990; McInish and Wood, 1990; McInish and Wood, 1992). However, it is interesting to

note that the current data shows higher closing volumes compared to the opening volumes. "The

pattern in the last period can be explained by investors avoiding overnight risk exposure by

unwinding their positions before close." (Wei, 1992). Considering the financial distress during

the sample period, higher closing volumes compared to the previous studies provide supporting

evidence for the argument set forth by Wei (1992). It is also interesting to note that there is no

significant difference of intraday pattern of volume for different size deciles. Based on different

hypotheses about the intraday trading patterns (asymmetric information, stealth trading etc.),

different sized companies have similar volume patterns. This however is not the case for

proportional spreads and for standard deviation of returns. Size is a significant differentiating

factor for both. As companies get larger there seem to be flatter intraday proportional spread

pattern. This would suggest, for instance, that if specialists are capitalizing on the information

asymmetry during the market open, there is lower opportunity for larger companies.

For the entire sample, for the first HHP of the day, HHP volume (proportional spread) is

9.71% (0.271%) while this level is 25.70% (0.173%) for the last HHP of the day. For the rest of

the day, the average volume (proportional spread) is 5.88% (0.139%). The curvature for the

intraday trading activity and liquidity measures is evident in the sample.

Differences between consecutive HHPs are statistically significant at1% level for volume

and for proportional spread for almost all HHPs. The evident pattern in proportional spreads

disappears if daily averaging were to be used. For instance, for the entire sample, the

proportional spread for the first HHP is 0.271%, it is down to an average of 0.139% until 3:29pm

9
which is followed by an increase to 0.173%. However, average of proportional spread is 0.152%

for the entire day. This daily averaging, thus omitting the significantly different first and last

HHP of the day, should impact studies that use daily averaged trading activity and liquidity

measures.

The point of this study is to argue that first and last HHP of the trading day have trading

activity that is the main determinant for any commonality in liquidity evidence. While intraday

descriptive statistics present evidence of intraday trading patterns, they fall short of explaining

why commonality in liquidity should be related to these specific periods. We hypothesize that

daily change in liquidity is due to the change in liquidity for the first and the last HHP of the day

compared to the same HHPs of the previous day. This means, the reasons pertaining to the

change in liquidity impact liquidity during the first and the last HHP of the day. If this were to be

true, there should be two empirical consequences. First, there should be high standard deviation

of change in liquidity for the first and the last HHP across days. Also, there should be low

standard deviation of change in liquidity for the rest of the day. In other words, the U-shape in

trading activity should not shift (up or down) completely from one day to the next. Instead, while

the bottom of the U-shape has very low change, the tips of the U-shape changes (higher or lower)

across days. Evidence for this empirical consequence is provided with Figure 2. The standard

deviation of level of volume and proportional spread for the first and the last HHPs across days

is evident. In fact, similar pattern exists for return and standard deviation of return.

The second empirical consequence of our hypothesis is the expectation that the level of

commonality in liquidity would be different for different HHPs of the day. In other words, if

level of commonality in liquidity is estimated for each HHP across days, we should find that the

first and the last HHP have the highest commonality level. The rest of the day should have very

10
low levels of commonality. In order to prove this point, we initially provide evidence of

commonality using daily change in liquidity similar to previous studies. Then, the evidence is

provided to show that the first and the last HHPs are the main reasons for the reported

commonality in liquidity.

Equation 1 is estimated for each stock using daily variables for the commonality in

liquidity evidence. Table 2 provides the results for the Equation 1. Since we cannot report results

for each stock, the table provides the means of coefficients for the change in stock's proportional

spread with the change in market's proportional spread; commonality in liquidity. The table also

reports the percentage of the market that have positive commonality coefficient and percentage

of the market that have positive and statistically significant (at 5% level) commonality

coefficient; level of commonality. For the entire sample, 59.27% of the market have statistically

significant commonality in liquidity. This level is reported to be about 30% for the Chordia et. al.

(2000) and about 90% for the Coughenour and Saad (2004). Mean commonality coefficient is

0.9786 for the entire sample.

The size effect is evident in the Table 2. Smallest of size deciles have the lowest level of

commonality in liquidity. The level of commonality is increasing by size except for the largest

two deciles. This evidence is in accordance with the previously reported commonality in

liquidity evidence in the literature.

To evaluate the impact of the first HHP on the level of commonality in liquidity,

Equation 2 is estimated for each stock for each HHP (using HHP variables) and the results are

reported in Table 3 and Figure 3 for each of the size deciles. As expected, the first HHP has the

highest commonality level for the entire market (67.24%) and for each of the size deciles. For the

11
entire market, level of commonality is on average 21.61% for each HHP excluding the first and

the last HHP of the day. The level of commonality increases for 2:00pm-2:30pm HHP and again

for the last HHP of the day. Size deciles four and five have the highest level of commonality

(average for 10:00am-3:30pm trading). Smallest of the size deciles has the lowest level of

commonality.

Concluding remarks and implications

Explaining commonality in liquidity is important for portfolio management. Theories

explaining intraday trading activity and liquidity patterns also explain commonality. The

evidence of commonality based on HHP variables show that part of the commonality in liquidity

can be explained by the trading in the first and in the last HHP of the day regardless of size. It is

also evident that commonality in liquidity can also be explained, but to a lesser degree, by size.

However, considering that the smallest of the size deciles has the highest proportional spreads

for the first HHP of the day, the level of commonality is more likely to be explained by the

intraday trading pattern which is more evident for the smallest companies. This evidence would

also explain the size effect in the level of commonality. These results are very similar to the

results with the sectors instead of the market as well as the results with the filtered sample

instead of the entire sample.

We argue that inventory risk and information asymmetry explain common liquidity

movements as well as common intraday trading patterns. Using daily aggregates of liquidity

measures smooth the intraday patterns. Estimating liquidity commonality with intraday data,

leaving intraday trading patterns intact, we show the close relationship between liquidity

commonality and intraday trading patterns, suggesting common reasoning to explain both lines

of literature.

12
References

Acharya, V. V. & Pedersen, L. H. (2005). Asset pricing with liquidity risk. Journal of Financial
Economics , 77 (2), 375–410

Brockman, P., Chung, D. Y. & Perignon, C. (2009). Commonality in Liquidity: A Global

Perspective. Journal of Financial and Quantitative Analysis, 44, 851-882

Chordia, T., Roll, R. & Subrahmanyam, A. (2000). Commonality in liquidity. Journal of

Financial Economics , 56 (1), 3–28.

Coughenour, J. F. & Saad, M. M. (2004). Common market makers and commonality in liquidity,
73, 37-69

Domowitz, I. & Wang, X. (2002). Liquidity, liquidity commonality and its impact on portfolio
theory. Working Paper, Penn State University .

Fabre, J. & Frino, A. (2004). Commonality in liquidity: Evidence from the Australian Stock
Exchange. Accounting and Finance , 44, 357–368.

Friederich, S. & Payne, R. (2007). Dealer Liquidity in an Auction Market: Evidence from the
London Stock Exchange. Economic Journal, Royal Economic Society, 117 (522), 1168-
1191

Galariotis, E. C. & Giouvris, E. (2007). Liquidity Commonality in the London Stock Exchange.
Journal of Business Finance & Accounting , 34, 374–388.

Hasbrouck, J. & Seppi, D. J. (2001). Common Factor in Prices, Order Flows and Liquidity.
Journal of Financial Economics , 59, 383–411.

Huberman, G. & Halka, D. (2001). Systematic Liquidity. Journal of Financial Research , 24,
161–178.

McInish, T. H. & Wood, R. A. (1990). An Analysis of transactions data for the Toronto Stock
Exchange: Return patterns and end-of-day effect, Journal of Banking and Finance, 14, 41-
458

McInish, T. H. & Wood, R. A. (1992). An Analysis of Intraday Patterns in Bid/Ask Spreads for
NYSE stocks. Journal of Finance, 47 (2), 753–764

13
Appendix
Table 1 (Panel A): Descriptive statistics across securities for each half-hour period for proportional spread (sprd: mean proportional spread for an half-hour period) and
volume (vol: total volume for an half-hour period / total daily volume). Data includes all listed securities on NYSE for the period between September, 2009 and June,
2010. Data source is Yahoo! Finance (http://finance.yahoo.com/, Yahoo! Finance uses quote data for NYSE directly from the exchange,
http://help.yahoo.com/l/us/yahoo/finance/quotes/fitadelay.html). "t " is the t-score for the difference of means t-test for each half-hour period to the next one. For the last
period of the day, t-test is to the first half-hour period of the same day. "<>" (">")("<") is the statistical significance for the test that period's mean is equal to (greater
than) (less than) the next period's mean. Size is the market capitalization and the market is divided into deciles. Mean values are provided for the deciles. Mean and the
t-test are provided for the entire market. "*", "**", and "***" refer to statistical significance at 1%, 5% and 10% levels.

Panel A: Proportional Spread

HHP Size 1 Size 2 Size 3 Size 4 Size 5 Size 6 Size 7 Size 8 Size 9 Size 10 All t <> > <
9:30 - 9:59 0.718% 0.454% 0.358% 0.275% 0.226% 0.179% 0.160% 0.138% 0.119% 0.085% 0.271% 52.04 * *
10:00 - 10:29 0.470% 0.283% 0.218% 0.158% 0.127% 0.096% 0.089% 0.082% 0.068% 0.056% 0.165% 38.80 * *
10:30 - 10:59 0.435% 0.259% 0.195% 0.142% 0.115% 0.087% 0.082% 0.076% 0.064% 0.053% 0.151% 28.04 * *
11:00 - 11:29 0.425% 0.249% 0.190% 0.135% 0.110% 0.083% 0.078% 0.074% 0.062% 0.053% 0.146% 9.61 * *
11:30 - 11:59 0.416% 0.242% 0.185% 0.131% 0.107% 0.082% 0.077% 0.076% 0.063% 0.053% 0.143% 18.82 * *
12:00 - 12:29 0.412% 0.239% 0.180% 0.127% 0.103% 0.079% 0.074% 0.072% 0.061% 0.052% 0.140% 20.50 * *
12:30 - 12:59 0.404% 0.233% 0.178% 0.124% 0.101% 0.077% 0.073% 0.071% 0.059% 0.051% 0.137% 15.39 * *
13:00 - 13:29 0.397% 0.230% 0.173% 0.122% 0.099% 0.076% 0.072% 0.070% 0.059% 0.051% 0.135% 13.46 * *
13:30 - 13:59 0.390% 0.225% 0.170% 0.120% 0.098% 0.075% 0.072% 0.068% 0.058% 0.051% 0.133% 13.86 * *
14:00 - 14:29 0.384% 0.221% 0.165% 0.116% 0.096% 0.074% 0.070% 0.069% 0.058% 0.051% 0.131% 16.84 * *
14:30 - 14:59 0.382% 0.217% 0.161% 0.113% 0.093% 0.072% 0.069% 0.068% 0.057% 0.050% 0.128% 25.39 * *
15:00 - 15:29 0.376% 0.211% 0.154% 0.107% 0.090% 0.069% 0.066% 0.065% 0.055% 0.048% 0.124% -28.85 * *
15:29 - 16:00 0.458% 0.262% 0.205% 0.157% 0.134% 0.120% 0.115% 0.108% 0.098% 0.075% 0.173% -32.95 * *

Panel B: Volume
HHP Size 1 Size 2 Size 3 Size 4 Size 5 Size 6 Size 7 Size 8 Size 9 Size 10 All t <> > <
9:30 - 9:59 11.41% 9.78% 9.31% 8.70% 8.30% 8.47% 8.97% 9.75% 10.29% 12.19% 9.71% 33.96 * *
10:00 - 10:29 8.40% 7.97% 7.85% 7.73% 7.83% 7.96% 8.16% 8.45% 8.33% 9.08% 8.18% 77.31 * *
10:30 - 10:59 7.17% 6.83% 6.67% 6.64% 6.70% 6.79% 6.89% 7.13% 7.05% 7.48% 6.93% 60.04 * *
11:00 - 11:29 6.56% 6.13% 6.00% 5.94% 6.00% 6.09% 6.10% 6.27% 6.18% 6.69% 6.20% 45.83 * *
11:30 - 11:59 5.96% 5.65% 5.48% 5.50% 5.51% 5.59% 5.57% 5.66% 5.52% 5.94% 5.64% 54.34 * *
12:00 - 12:29 5.27% 5.05% 4.97% 4.91% 5.00% 5.05% 4.95% 5.00% 4.90% 4.94% 5.00% 43.42 * *
12:30 - 12:59 4.84% 4.68% 4.64% 4.60% 4.63% 4.63% 4.55% 4.58% 4.52% 4.45% 4.61% 2.28 ** **
13:00 - 13:29 4.76% 4.67% 4.59% 4.59% 4.63% 4.66% 4.56% 4.57% 4.51% 4.39% 4.59% -18.84 * *
13:30 - 13:59 4.91% 4.82% 4.79% 4.77% 4.82% 4.79% 4.74% 4.73% 4.66% 4.46% 4.75% -59.74 * *
14:00 - 14:29 5.39% 5.49% 5.48% 5.46% 5.48% 5.46% 5.40% 5.31% 5.24% 4.94% 5.36% -53.43 * *
14:30 - 14:59 5.83% 6.07% 6.08% 6.10% 6.07% 6.11% 6.02% 5.89% 5.84% 5.47% 5.95% -119.97 * *
15:00 - 15:29 7.00% 7.41% 7.71% 7.74% 7.82% 7.82% 7.64% 7.46% 7.38% 6.81% 7.48% -164.75 * *
15:29 - 16:00 22.85% 25.55% 26.56% 27.39% 27.27% 26.64% 26.51% 25.29% 25.67% 23.22% 25.70% 97.28 * *

14
Table 1 (Panel B): Descriptive statistics across securities for each half-hour period for return (ret: total return for an half-hour period) and standard deviation of return.
Data includes all listed securities on NYSE for the period between September, 2009 and June, 2010. Data source is Yahoo! Finance (http://finance.yahoo.com/, Yahoo!
Finance uses quote data for NYSE directly from the exchange, http://help.yahoo.com/l/us/yahoo/finance/quotes/fitadelay.html). "t " is the t-score for the difference of
means t-test for each half-hour period to the next one. For the last period of the day, t-test is to the first half-hour period of the same day. "<>" (">")("<") is the
statistical significance for the test that period's mean is equal to (greater than) (less than) the next period's mean. Size is the market capitalization and the market is
divided into deciles. Mean values are provided for the deciles. Mean and t-test are provided for the entire market. "*", "**", and "***" refer to statistical significance at
1%, 5% and 10% levels.

Panel A: Return
HHP Size 1 Size 2 Size 3 Size 4 Size 5 Size 6 Size 7 Size 8 Size 9 Size 10 All t <> > <
9:30 - 9:59 -0.025% -0.011% -0.012% 0.022% 0.025% 0.028% 0.026% 0.023% 0.019% 0.015% 0.011% 8.18 * *
10:00 - 10:29 -0.058% -0.030% -0.011% 0.002% 0.000% -0.005% -0.018% -0.009% -0.006% -0.005% -0.014% -4.20 * *
10:30 - 10:59 -0.022% -0.007% -0.001% 0.000% -0.001% -0.005% -0.006% -0.005% -0.005% -0.007% -0.006% 14.63 * *
11:00 - 11:29 -0.040% -0.039% -0.034% -0.034% -0.029% -0.021% -0.025% -0.028% -0.022% -0.023% -0.029% -3.92 * *
11:30 - 11:59 -0.035% -0.040% -0.030% -0.028% -0.021% -0.020% -0.016% -0.014% -0.017% -0.016% -0.024% -24.69 * *
12:00 - 12:29 0.008% 0.014% 0.017% 0.016% 0.010% 0.008% 0.005% 0.009% 0.006% 0.006% 0.010% 10.20 * *
12:30 - 12:59 0.003% 0.005% -0.004% -0.003% -0.002% -0.003% -0.004% -0.005% -0.006% -0.009% -0.003% -16.82 * *
13:00 - 13:29 0.023% 0.020% 0.019% 0.019% 0.017% 0.018% 0.017% 0.018% 0.016% 0.017% 0.018% 15.70 * *
13:30 - 13:59 0.008% -0.003% 0.001% 0.003% 0.002% 0.001% -0.004% -0.001% -0.004% -0.004% 0.000% 6.68 * *
14:00 - 14:29 -0.023% -0.016% -0.009% -0.005% -0.010% -0.005% -0.002% -0.002% 0.002% -0.003% -0.007% -2.19 ** **
14:30 - 14:59 0.000% 0.001% -0.003% -0.004% -0.009% -0.011% -0.007% -0.001% -0.008% -0.006% -0.005% 16.76 * *
15:00 - 15:29 -0.039% -0.027% -0.028% -0.034% -0.029% -0.024% -0.020% -0.023% -0.021% -0.022% -0.027% -23.04 * *
15:29 - 16:00 0.075% 0.053% 0.050% 0.039% 0.026% 0.025% 0.034% 0.009% 0.021% 0.021% 0.035% 6.21 * *

Panel B: Standard Deviation of Return

HHP Size 1 Size 2 Size 3 Size 4 Size 5 Size 6 Size 7 Size 8 Size 9 Size 10 All t <> > <
9:30 - 9:59 0.0078 0.0059 0.0052 0.0047 0.0043 0.0038 0.0037 0.0037 0.0032 0.0027 0.0045 78.17 * *
10:00 - 10:29 0.0054 0.0041 0.0036 0.0033 0.0030 0.0026 0.0026 0.0026 0.0022 0.0021 0.0031 78.46 * *
10:30 - 10:59 0.0045 0.0034 0.0029 0.0026 0.0024 0.0021 0.0021 0.0021 0.0018 0.0017 0.0026 61.31 * *
11:00 - 11:29 0.0041 0.0030 0.0026 0.0023 0.0021 0.0018 0.0018 0.0019 0.0016 0.0016 0.0023 38.44 * *
11:30 - 11:59 0.0039 0.0028 0.0024 0.0022 0.0020 0.0017 0.0017 0.0018 0.0015 0.0014 0.0021 52.68 * *
12:00 - 12:29 0.0036 0.0026 0.0022 0.0020 0.0018 0.0016 0.0015 0.0016 0.0014 0.0013 0.0020 46.81 * *
12:30 - 12:59 0.0034 0.0024 0.0021 0.0018 0.0016 0.0014 0.0014 0.0014 0.0012 0.0012 0.0018 7.68 * *
13:00 - 13:29 0.0033 0.0024 0.0020 0.0018 0.0016 0.0014 0.0014 0.0014 0.0012 0.0012 0.0018 14.52 * *
13:30 - 13:59 0.0032 0.0023 0.0020 0.0018 0.0016 0.0014 0.0014 0.0014 0.0012 0.0011 0.0017 -17.31 * *
14:00 - 14:29 0.0033 0.0024 0.0020 0.0018 0.0016 0.0014 0.0014 0.0014 0.0013 0.0012 0.0018 -21.94 * *
14:30 - 14:59 0.0034 0.0025 0.0022 0.0019 0.0017 0.0015 0.0015 0.0015 0.0013 0.0012 0.0019 -7.98 * *
15:00 - 15:29 0.0035 0.0026 0.0022 0.0020 0.0018 0.0015 0.0015 0.0016 0.0014 0.0013 0.0019 -33.12 * *
15:29 - 16:00 0.0045 0.0032 0.0028 0.0025 0.0025 0.0023 0.0023 0.0027 0.0025 0.0031 0.0028 -46.22 * *

15
Figure 1: Graphs for the means for each half-hour period across securities for volume (vol: total volume for an half-hour period / total daily volume), mean proportional
spread (sprd: mean proportional spread for an half-hour period), mean return (ret: total return for an half-hour period) and mean standard deviation of return. Data
includes all listed securities on NYSE for the period between September, 2009 and June, 2010. Data source is Yahoo! Finance (http://finance.yahoo.com/, Yahoo!
Finance uses quote data for NYSE directly from the exchange, http://help.yahoo.com/l/us/yahoo/finance/quotes/fitadelay.html). Size is the market capitalization and the
market is divided into deciles.
30.00% 0.800%

0.700%
25.00%

0.600%

20.00% Size 1 Size 1

Size 2 0.500% Size 2
Size 3 Size 3
Size 4 Size 4
15.00% Size 5 0.400% Size 5
Size 6 Size 6
Size 7 Size 7
Size 8 0.300% Size 8

10.00% Size 9 Size 9

Size 10 Size 10
All 0.200% All

5.00%
0.100%

0.00% 0.000%
9:30 - 10:00 - 10:30 - 11:00 - 11:30 - 12:00 - 12:30 - 13:00 - 13:30 - 14:00 - 14:30 - 15:00 - 15:29 - 9:30 - 10:00 - 10:30 - 11:00 - 11:30 - 12:00 - 12:30 - 13:00 - 13:30 - 14:00 - 14:30 - 15:00 - 15:29 -
9:59 10:29 10:59 11:29 11:59 12:29 12:59 13:29 13:59 14:29 14:59 15:29 16:00 9:59 10:29 10:59 11:29 11:59 12:29 12:59 13:29 13:59 14:29 14:59 15:29 16:00

Panel A: Volume Panel B: Proportional Spread

0.100% 0.0090

0.080% 0.0080

0.0070
0.060%

Size 1 0.0060 Size 1

0.040%
Size 2 Size 2
Size 3 Size 3
0.0050
Size 4 Size 4
0.020%
Size 5 Size 5
Size 6 Size 6
0.0040
Size 7 Size 7
0.000%
Size 8 Size 8
9:30 - 10:00 - 10:30 - 11:00 - 11:30 - 12:00 - 12:30 13:00 13:30 14:00 14:30 - 15:00 - 15:29 -
9:59 10:29 10:59 11:29 11:59 12:29 - 12:59 - 13:29 - 13:59 - 14:29 14:59 15:29 16:00 Size 9 0.0030 Size 9
Size 10 Size 10
-0.020%
All All
0.0020

-0.040%

0.0010

-0.060%
0.0000
9:30 - 10:00 - 10:30 - 11:00 - 11:30 - 12:00 - 12:30 - 13:00 - 13:30 - 14:00 - 14:30 - 15:00 - 15:29 -
9:59 10:29 10:59 11:29 11:59 12:29 12:59 13:29 13:59 14:29 14:59 15:29 16:00
-0.080%

Panel C: Return Panel D: Standard Deviation of Return

16
Figure 2: Graphs for the standard deviation, across days, of means for each half-hour period for volume (vol: total volume for an half-hour period / total daily volume),
mean proportional spread (sprd: mean proportional spread for an half-hour period), mean return (ret: total return for an half-hour period) and mean standard deviation of
return. Data includes all listed securities on NYSE for the period between September, 2009 and June, 2010. Data source is Yahoo! Finance (http://finance.yahoo.com/,
Yahoo! Finance uses quote data for NYSE directly from the exchange, http://help.yahoo.com/l/us/yahoo/finance/quotes/fitadelay.html). Size is the market capitalization
and the market is divided into deciles.
0.05000 0.00140

0.04500
0.00120

0.04000

0.00100
0.03500
Size 1 Size 1
Size 2 Size 2
0.03000
Size 3 Size 3
0.00080
Size 4 Size 4
0.02500 Size 5 Size 5
Size 6 Size 6
Size 7 0.00060 Size 7
0.02000
Size 8 Size 8
Size 9 Size 9
0.01500 Size 10 Size 10
0.00040
All All

0.01000

0.00020
0.00500

0.00000 0.00000
9:30 - 10:00 - 10:30 - 11:00 - 11:30 - 12:00 - 12:30 - 13:00 - 13:30 - 14:00 - 14:30 - 15:00 - 15:29 - 9:30 - 10:00 - 10:30 - 11:00 - 11:30 - 12:00 - 12:30 - 13:00 - 13:30 - 14:00 - 14:30 - 15:00 - 15:29 -
9:59 10:29 10:59 11:29 11:59 12:29 12:59 13:29 13:59 14:29 14:59 15:29 16:00 9:59 10:29 10:59 11:29 11:59 12:29 12:59 13:29 13:59 14:29 14:59 15:29 16:00

Panel A: Volume Panel B: Proportional Spread

0.00700 0.00160

0.00140
0.00600

0.00120
0.00500

Size 1 Size 1
Size 2 0.00100 Size 2
Size 3 Size 3
0.00400
Size 4 Size 4
Size 5 0.00080 Size 5
Size 6 Size 6
0.00300 Size 7 Size 7
Size 8 0.00060 Size 8
Size 9 Size 9
Size 10 Size 10
0.00200
All 0.00040 All

0.00100
0.00020

0.00000 0.00000
9:30 - 10:00 - 10:30 - 11:00 - 11:30 - 12:00 - 12:30 - 13:00 - 13:30 - 14:00 - 14:30 - 15:00 - 15:29 - 9:30 - 10:00 - 10:30 - 11:00 - 11:30 - 12:00 - 12:30 - 13:00 - 13:30 - 14:00 - 14:30 - 15:00 - 15:29 -
9:59 10:29 10:59 11:29 11:59 12:29 12:59 13:29 13:59 14:29 14:59 15:29 16:00 9:59 10:29 10:59 11:29 11:59 12:29 12:59 13:29 13:59 14:29 14:59 15:29 16:00

Panel C: Return Panel D: Standard Deviation of Return

17
Table 2: Results for the daily commonality in liquidity. Commonality is estimated by regressing total daily change in proportional spread for an individual stock on the
same variable for the entire market excluding the individual stock. Data includes all listed securities on NYSE for the period between September, 2009 and June, 2010.
Data source is Yahoo! Finance (http://finance.yahoo.com/, Yahoo! Finance uses quote data for NYSE directly from the exchange,
http://help.yahoo.com/l/us/yahoo/finance/quotes/fitadelay.html). Size is the market capitalization and the market is divided into deciles. Mean values are provided for
the coefficient for the market liquidity variable, its lag, its lead and total of these three coefficients (Msprd, L.Msprd, F.Msprd and total).Mean R-squared is also
provided (R2). Percentages are provided for the percentage of the entire market that have positive (+), and positive and statistically significant (at 5% level) (+ & Sig.)
market liquidity coefficient. Model is estimated for the entire sample."*", "**", and "***" refer to statistical significance at 1%, 5% and 10% levels.

Size Msprd + + & sig L.Msprd + + & sig F.Msprd + + & sig Total R2
1 0.4970 * 86.36% 28.98% -0.1151 * 40.34% 1.70% -0.1893 * 34.09% 1.14% 0.1926 * 0.071 *
2 0.5091 * 89.20% 28.98% -0.1704 * 35.23% 1.14% -0.1459 * 37.50% 1.70% 0.1928 * 0.064 *
3 0.6236 * 93.18% 34.66% -0.1376 * 37.50% 1.70% -0.1228 * 36.93% 0.00% 0.3632 * 0.070 *
4 0.7225 * 96.02% 42.61% -0.0428 47.16% 3.41% -0.1030 * 35.23% 1.70% 0.5768 * 0.078 *
5 0.9881 * 99.43% 65.71% -0.1203 * 41.14% 1.14% -0.0297 48.00% 1.71% 0.8381 * 0.092 *
6 1.3135 * 98.86% 77.27% -0.0219 52.84% 4.55% 0.0129 53.41% 1.14% 1.3044 * 0.126 *
7 1.3890 * 96.02% 80.11% 0.0470 53.98% 6.25% 0.0458 55.11% 6.82% 1.4817 * 0.146 *
8 1.3548 * 99.43% 79.55% 0.1179 * 56.82% 9.66% 0.1376 * 60.80% 7.39% 1.6104 * 0.146 *
9 1.4574 * 98.30% 85.23% 0.1841 * 65.34% 13.07% 0.1900 * 67.05% 9.66% 1.8315 * 0.161 *
10 0.9310 * 95.43% 69.71% 0.2501 * 74.29% 16.00% 0.2565 * 69.14% 16.57% 1.4377 * 0.147 *
All 0.9786 * 95.22% 59.27% -0.0010 50.46% 5.86% 0.0051 49.72% 4.78% 0.9827 * 0.110 *

18
Table 3: Results for the commonality in liquidity for each half hour period (HHP). Commonality is estimated by regressing total HHP change in proportional spread for
an individual stock on the same variable for the entire market excluding the individual stock. Data includes all listed securities on NYSE for the period between
September, 2009 and June, 2010. Data source is Yahoo! Finance (http://finance.yahoo.com/, Yahoo! Finance uses quote data for NYSE directly from the exchange,
http://help.yahoo.com/l/us/yahoo/finance/quotes/fitadelay.html). Size is the market capitalization and the market is divided into deciles. Percentages are provided for the
percentage of the entire market that have positive and statistically significant (at 5% level) (+ & Sig.) market liquidity coefficient. Model is estimated for the entire
sample (No filter).

9:30 10:00 10:30 11:00 11:30 12:00 12:30 13:00 13:30 14:00 14:30 15:00 15:30
9:59 10:29 10:59 11:29 11:59 12:29 12:59 13:29 13:59 14:29 14:59 15:29 16:00
Size Msprd Msprd Msprd Msprd Msprd Msprd Msprd Msprd Msprd Msprd Msprd Msprd Msprd
0.4576 0.4273 0.4238 0.7519 0.3273 0.5366 0.5538 0.5067 0.4458 0.5613 0.6617 0.8013 0.4368
1
27.27% 8.52% 5.68% 11.36% 6.25% 9.66% 6.82% 9.66% 8.52% 16.48% 13.07% 19.89% 15.91%
0.4812 0.9391 0.8032 0.8907 0.8021 0.9068 0.9833 0.8781 0.8504 0.8569 0.9717 0.9581 0.5024
2
32.39% 18.75% 20.45% 13.64% 13.64% 18.75% 17.61% 17.61% 13.64% 30.68% 21.02% 25.57% 25.00%
0.6693 1.0529 1.1839 0.9013 0.9923 1.0423 1.0134 1.0237 1.0880 1.0612 1.1205 1.1700 0.6077
3
56.82% 22.16% 31.82% 13.07% 14.20% 19.89% 18.75% 17.05% 22.16% 41.48% 27.27% 34.09% 28.98%
0.7988 1.0973 1.3158 1.1091 1.2644 1.2648 1.3213 1.2208 1.3414 1.2264 1.1007 1.3058 0.7175
4
65.34% 25.00% 33.52% 18.75% 21.02% 31.25% 27.84% 25.00% 32.39% 40.34% 26.70% 42.61% 38.64%
1.0137 1.2252 1.3407 1.2514 1.1497 1.1798 1.2075 1.2147 1.1745 1.1455 1.1293 1.1912 1.0013
5
77.14% 29.14% 37.14% 21.14% 19.43% 27.43% 24.57% 23.43% 22.29% 40.57% 30.29% 40.57% 56.00%
1.1405 1.0526 1.0602 1.1991 1.1472 1.0528 1.0052 1.1309 0.9565 1.0082 0.8955 1.1430 1.4457
6
82.95% 23.86% 28.41% 22.73% 16.48% 22.73% 17.61% 21.59% 19.89% 35.23% 25.57% 41.48% 75.00%
1.2723 0.9119 1.0135 0.9575 1.0087 1.1632 0.8382 0.9612 0.9824 1.0653 0.9336 0.9883 1.4797
7
82.39% 17.05% 23.86% 17.05% 17.61% 24.43% 19.32% 23.86% 21.59% 40.34% 30.11% 36.36% 75.00%
1.2929 0.8986 0.8892 0.8382 0.7923 0.8728 0.8097 0.8754 0.9582 0.8582 0.7506 0.8064 1.3653
8
83.52% 22.73% 27.27% 16.48% 13.07% 16.48% 17.61% 15.91% 22.16% 30.11% 25.57% 26.70% 72.73%
1.4609 0.7983 0.6897 0.7203 0.5582 0.8442 0.6789 0.6944 0.7240 0.7481 0.7042 0.6388 1.3321
9
86.36% 18.75% 18.75% 13.64% 6.25% 15.91% 14.20% 13.64% 15.34% 28.41% 26.14% 27.84% 73.86%
1.0787 0.6271 0.5623 0.5127 0.5501 0.5728 0.5685 0.4848 0.3622 0.7029 0.5728 0.6026 0.8528
10
78.29% 14.29% 14.86% 8.57% 10.29% 12.00% 13.71% 8.00% 6.29% 21.71% 29.14% 28.57% 56.57%
0.9665 0.9030 0.9282 0.9132 0.8592 0.9437 0.8980 0.8991 0.8885 0.9234 0.8841 0.9606 0.9742
All
67.24% 20.02% 24.18% 15.64% 13.82% 19.85% 17.80% 17.58% 18.43% 32.54% 25.48% 32.37% 51.76%

19
 Figure 3: Graph for the percentage of market that have positive and statistically significant (at 5% level) half hour period (HHP) commonality in liquidity by size. Commonality is estimated by regressing total daily change in
proportional spread for an individual stock on the same variable for the entire market (M) as well as for the sector of the individual stock (S). Data includes all listed securities on NYSE for the period between September,
View publication stats

2009 and June, 2010. Data source is Yahoo! Finance (http://finance.yahoo.com/, Yahoo! Finance uses quote data for NYSE directly from the exchange, http://help.yahoo.com/l/us/yahoo/finance/quotes/fitadelay.html). Size is
the market capitalization and the market is divided into deciles. Model is estimated for the entire sample.

90%

80%

70%

60% 80%-90%
70%-80%
50%
60%-70%
50%-60%
40%
40%-50%

30% 30%-40%
20%-30%
20% All 10%-20%
Size 9 0%-10%
10%
Size 7
0% Size 5

Size 3

Size 1

20