Do Hedge Funds Provide Liquidity?

Evidence From Their Trades

FRANCESCO FRANZONI∗
University of Lugano and Swiss Finance Institute

ALBERTO PLAZZI†
University of Lugano and Swiss Finance Institute

ABSTRACT
The paper provides significant evidence of limits of arbitrage in the hedge fund sector. Using
unique data on institutional transactions, we show that the price impact of hedge fund trades
increases when aggregate conditions deteriorate. The finding is consistent with arbitrageurs’
withdrawal from liquidity provision following a tightening in funding liquidity. Compared to
other institutions, hedge funds display the largest sensitivity of trading costs to aggregate con-
ditions. We pin down this effect to a subset of hedge funds that are more exposed to funding
constraints because of their leverage, lack of share restrictions, asset illiquidity, low reputa-
tional capital, and trading style. Value-based trading strategies demand liquidity in bad times,
whereas momentum strategies provide liquidity. Lastly, a decrease in hedge fund trading inten-
sity predicts a widening of the bid-ask spread at the stock-level, while other institutions’ trading
activity does not seem to matter for market liquidity.

JEL classification: G20, G23

Keywords: hedge funds, limits of arbitrage, liquidity provision, trading costs, funding liquidity

∗ USI, Institute of Finance, Via Buffi 13 Lugano, 6900, Switzerland, email: francesco.franzoni@usi.ch.
† USI, Institute of Finance, Via Buffi 13 Lugano, 6900, Switzerland, email: alberto.plazzi@usi.ch.
We thank Tobias Adrian, Giovanni Barone-Adesi, Pierre Collin-Dufresne, Jens Jackwerth, Terrence Hendershott, Paul
Irvine, Albert Menkveld, Joanid Rosu, Youchang Wu, and seminar participants to the University of Konstanz, Norwegian
School of Economics, and USI Ifin Brown Bag Seminar for comments. We are particularly grateful to Dimitri Vayanos
and Andy Puckett for extensive insights. The authors gratefully acknowledge financial support from the Swiss Finance
Institute, Inquire Europe, and the Swiss National Science Foundation. The Internet Appendix can be downloaded at
http://www.people.usi.ch/plazzia/research.htm.
1 Introduction

In the literature that relates market liquidity to investors’ funding conditions (e.g. Brunnermeier

and Pedersen (2009)), a liquidity provider is the ultimate holder of an underpriced asset and

short-seller of an overpriced one. Hedge funds are the main candidates to play this role as they

closely resemble the ‘speculators’ in these models in terms of both trading strategies and reliance on

external finance. Consistent with this notion of liquidity provision, the typical hedge fund strategies

provide an anchor for mispriced securities by making portfolio decisions conditional on perceived

misvaluations. In this regard, hedge funds are different from the traditional market makers (e.g.

the specialists or, more recently, the high-frequency traders), which tend to hold zero inventories

at the end of the day. This paper focuses on this dimension of liquidity provision and studies

empirically the dependence of hedge funds’ stock trades on funding conditions.1

While there is evidence that hedge funds profit from liquidity provision, (e.g., Aragon (2007),

Sadka (2010), Jylha, Rinne, and Suominen (2012)), other studies point out limits to this ability. A

large body of theoretical literature posits arbitrageurs’ dependence on external finance which, at

times of market stress, becomes unreliable.2 Some recent empirical evidence confirms that hedge

fund performance is related to funding conditions (Teo (2011)) and that hedge funds’ withdrawal

from the market impacts stock liquidity (Aragon and Strahan (2012)). Given the importance of

liquidity provision for the smooth functioning of financial markets, it is crucial to understand how

hedge funds respond to a deterioration in aggregate conditions. Knowledge of the interplay between

funding conditions and liquidity provision is also key to identify hedge funds’ role in spreading

systemic risk (Boyson, Stahel, and Stulz (2010), Billio, Getmansky, Lo, and Pelizzon (2012)).

Prior work uses quarterly portfolio holdings to infer the trading behavior of hedge funds in equity

markets during the last financial crisis (Ben-David, Franzoni, and Moussawi (2012)). However,

liquidity provision is inherently a trade-level concept relating to how patiently a trade is executed.

Quarterly changes in portfolio holdings, therefore, cannot distinguish between long-term portfolio

reallocations and short-term variations in liquidity supply. The latter can be examined more
1 The same notion of liquidity provision is used in a recent paper by Anand et al. (2013), who also draw on institutional trading
data. These authors describe a tri-party market with liquidity demanders, intermediaries (i.e. specialists or high-frequency
traders), and long-term liquidity suppliers. Like us, these authors interpret the buy-side institutions as the long-term suppliers
of liquidity.
2 A brief selection of theoretical papers that model limits of arbitrage includes Shleifer and Vishny (1997), Gromb and Vayanos

(2002), Brunnermeier and Pedersen (2009), Gromb and Vayanos (2010), Gromb and Vayanos (2012).

1
appropriately by means of the trade-level data that we use in this study. Anand, Irvine, Puckett,

and Venkataraman (2013), like us, draw on institutional trading data to infer market participation

of institutional investors. We depart from their analysis mainly in our focus on the cross-sectional

dimension within the institutional sector. We contrast the behavior of hedge funds, whose trading

strategies are very sensitive to external funding, to that of other institutions (mutual funds and

pension funds), which are less dependent on changes in funding conditions due to a limited reliance

on leverage. Also, within the hedge fund sector, we study how the impact of funding variables

on liquidity provision interacts with fund level measures of financial constraints, such as: leverage,

redemption restrictions, asset illiquidity, and reputational capital. Lastly, while prior studies mostly

focus on the financial crisis, we show that the dependence of hedge funds’ liquidity provision on

funding constraints holds in ‘normal’ times as well.

Our data set contains trade-level observations for over eight hundred different institutions (hedge

funds, mutual funds, pension funds, and other money managers) during the January 1999 to Decem-

ber 2010 period. The data source is Abel Noser Solutions (aka ‘Ancerno’), a company specialized

in consulting services to institutional investors for trading costs analysis.3 Ancerno provides re-

searchers with data on the trading activity of its clients’ portfolio managers, under the agreement

that the names of the client institutions are not disclosed. However, the name of the institution

that manages the client’s portfolio is provided. This information allows us to identify eighty-seven

distinct hedge fund management companies.4 These firms appear to be highly representative of the

overall industry along several dimensions. In particular, we provide evidence that the hedge funds

in our sample are not statistically different from the other funds in TASS in terms of the exposure

to the main explanatory variables of this study.

The notion of liquidity that we state at the beginning guides our identification of liquidity pro-

vision/demand in the data. We expect liquidity demanders to trade impatiently and, consequently,

to have a positive price impact. The opposite is true for liquidity suppliers. As in Puckett and
3 Other recent studies using Ancerno data are Chemmanur, He, and Hu (2009), Puckett and Yan (2011), Chemmanur, Hu, and
Huang (2010), Anand, Irvine, Puckett, and Venkataraman (2013), Anand, Irvine, Puckett, and Venkataraman (2012).
4 In more detail, we identify hedge fund management companies by manually matching the managers in Ancerno with a list of

hedge fund management companies in the 13F mandatory filings and the Lipper/TASS database. The list of hedge funds in
the 13F filings is the same that is used in Ben-David, Franzoni, and Moussawi (2012) and Ben-David, Franzoni, Landier, and
Moussawi (2012). The institutions that report to both Ancerno and 13F are hedge-fund management companies as opposed
to the individual funds. For consistency, we aggregate the fund-level observations in TASS at the management company level.
When we use the wording ‘hedge funds’, we broadly refer to the firms that belong to this asset class rather than to the specific
funds within a management company.

2
Yan (2011), we compute price impact as the percentage difference between the execution price and

the volume-weighted average price (VWAP) for the same stock during the day, and express this

difference as a fraction of the VWAP. Our choice of benchmark is immaterial for our conclusions,

as the results hold also when we compute price impact using the Price at Market Open, as in

Anand, Irvine, Puckett, and Venkataraman (2013), or the Price at Order Placement, as in Anand,

Irvine, Puckett, and Venkataraman (2012). We also consider proxies for liquidity provision based

on reversal strategies (Lo and MacKinlay (1990) and Nagel (2012)) and trading style (Anand,

Irvine, Puckett, and Venkataraman (2013)). These alternative measures correlate strongly with

price impact, which is reassuring about the identification strategy.

Our analysis is organized around three hypotheses that formulate the predictions of the limits

of arbitrage theories. First, in the times series, we conjecture and test that hedge funds’ liquidity

provision depends on aggregate funding conditions. Second, we contrast hedge funds to the other

institutions in our data, as in a difference-in-differences approach. If hedge funds trading costs

are driven by funding constraints, rather than by a generalized increase in the price of liquidity

in bad times, we expect this effect to persist also when benchmarking to other institutions. We

motivate this conjecture by arguing that hedge funds’ funding needs are more sensitive to financial

conditions than those of other institutions. Third, we explore the heterogeneity within the hedge

fund sector. The prediction is that hedge funds are not equally exposed to funding liquidity.

Rather, the sensitivity to funding conditions depends on firm-level determinants that are related

to the availability and reliability of funding.

The results are easily summarized. In the time series, we find that liquidity provision dete-

riorates when funding conditions tighten. There is substantial variation in hedge funds’ implicit

trading costs over the sample period. Starting in 2002, trading costs declined unambiguously until

the first semester of 2007. In this period, hedge funds’ liquidity provision appears to be at its

highest. Then, price impact starts to increase towards the end of 2007, in correspondence with the

Quant Meltdown (Khandani and Lo (2011)). In the first semester of 2009, our measure unambigu-

ously suggests that hedge funds drastically increased their liquidity consumption. Liquidity demand

became less pronounced from the second semester of 2009 through the end of 2010. Consistent with

Brunnermeier and Nagel (2004), who show that hedge funds rode the Internet Bubble and then

switched to contrarian positions shortly before the downturn, we also observe an increase in trading

3
costs in the year 2000, which is then reversed in the 2001-2002 period. In a regression framework,

we find that hedge funds’ price impact increases following a drop in the stock market and increases

in the VIX, the TED Spread, and the LIBOR and we show that the relation holds outside of the

financial crisis as well. These variables are proxies for funding liquidity, either through the value

of collateral (related to the return on the stock market), the tightness of margins (related to the

VIX), or the cost of leverage (measured by the TED Spread and the LIBOR).5 Thus, the evidence

is consistent with a role for limits of arbitrage in hedge funds’ liquidity provision.

In the comparison with other institutions, hedge funds appear to be significantly more exposed

to changes in funding liquidity. While other institutions also experience an increase in trading

costs during the financial crisis, the price impact of their trades is only about 50% of that for hedge

funds in 2009Q1. The regression analysis suggests that there is no systematic relation between the

trading costs of other institutions and most of the funding liquidity proxies. This is not a result of

the fact that hedge funds trade in more illiquid stocks, as it holds also when controlling for stock-

and trade-level characteristics. Using the other institutions as a control group allows us to infer

that the increase in hedge funds’ trading costs results from their reduced liquidity provision in bad

times, rather than from a generalized surge in trading costs for all investors.

The analysis of the heterogeneity within the hedge fund sector gives further confirmation of the

conjecture that limits of arbitrage theories describe hedge fund behavior. We find that the exposure

of price impact to aggregate funding conditions is significant larger for funds with higher leverage,

more illiquid assets, lower reputational capital (as measured by fund age and past performance), and

lower restrictions to investors’ redemptions. These characteristics are related to hedge funds’ ability

to retain capital in bad times. As such, they are proxies for funding constraints. Furthermore,

only the funds that are classified as constrained by having positive leverage and low redemption

restrictions, display a relation between their trading costs and the aggregate funding liquidity

proxies. This result gives an important qualification to our previous findings. The statement that

hedge funds are more constrained than other institutions in their ability to provide liquidity in

bad times has to be confined to the subset of hedge funds with trading styles that magnify their

exposure to limits of arbitrage.

5 Thesemacro variables are used as proxies for funding liquidity in, for example, Hameed, Kang, and Viswanathan (2010),
Boyson, Stahel, and Stulz (2010), Garleanu and Pedersen (2011), and Nagel (2012).

4
 As a further validation of the limits of arbitrage conjecture, we show that, when funding con-

ditions deteriorate, the trades of more constrained funds are less profitable over the next five-day

horizon. This can happen if hedge funds respond to a tightening in funding liquidity by giving up

to profitable trading strategies or by aggressively seeking liquidity in the market. In either case,

this result suggests that binding constraints force hedge funds to deviate from their optimal trading

strategies.

Another dimension of the cross-sectional heterogeneity in the hedge fund sector that we explore

is related to hedge funds’ trading styles. For a hedge fund, its trading strategy is the ultimate

determinant of the availability of funding liquidity. For example, different trading strategies require

different levels of leverage to be profitable. A fund whose trading strategy requires higher leverage

may be obliged to forced liquidations in bad times. We focus on three popular trading styles in

the equity space: short-term reversal, momentum, and value. We study the hedge funds’ behavior

in terms of liquidity provision based on each of these dimensions, both unconditionally and as a

function of aggregate conditions. We find that momentum and reversal strategies provide liquidity

in bad times, while value strategies extract liquidity when funding conditions deteriorate. Given

that value stocks are to a large extent low-beta stocks, the latter finding seems consistent with

the arguments in Frazzini and Pedersen (2013) that arbitrageurs’ use leverage to exploit the high

expected returns of low-beta/value stocks. This leverage makes a betting-against-beta strategy

harder to hold on to in bad times and obliges the funds to forced liquidations, which have large

price impact and raise the trading costs.

To conclude, we study to what extent hedge funds’ trading matters for market liquidity. Our

data give us a unique opportunity to address this question. We regress market liquidity on hedge-

fund trading intensity at the stock level controlling for stock characteristic and aggregate conditions.

Our analysis shows that an increase in hedge fund trading in a given stock reduces the bid-ask

spread in the following week. The trading intensity of other institutions, instead, does not seem

to matter for market liquidity. Using direct evidence on hedge fund trading intensity at the stock

level, this finding complements the results in Aragon and Strahan (2012) that hedge fund market

participation affects stock liquidity.

Our paper relates to different strands of the literature. Some recent papers explore trading

activity of institutional investors using Ancerno data. Most closely related to our work, Anand,

5
Irvine, Puckett, and Venkataraman (2013) contrast the trading behavior and liquidity provision

of Ancerno institutions in calm times and during the 2007-2009 financial crisis. They study the

implications of institutional trades for stock price resiliency. Relative to their work, our contribution

is to investigate differences in liquidity provision between hedge funds and other institutions. We

also identify heterogeneity in liquidity provision within the hedge fund sector as a function of limits

of arbitrage. Puckett and Yan (2011) show that the intra-quarter trading activity of institutional

investors that is concealed by quarterly filings generates persistent and significant abnormal returns.

They also find that the most skilled funds experience higher trading costs and, therefore, demand

liquidity in their trades. Our evidence lines up well with their findings in that hedge funds are among

the institutions with higher interim activity and are found to be on average liquidity demanders.

Lipson and Puckett (2010) find that the institutions in Ancerno are on aggregate net buyers during

extreme market declines, but this does not come at the cost of negative ex-post returns. We

complement their work by documenting that funding liquidity shocks have a negative impact on the

short-term performance of the most constrained funds. In the hedge fund literature, Sadka (2010)

identifies a liquidity risk premium in hedge fund average returns. Our results possibly provide

the microstructure evidence on the dependence of hedge fund returns on liquidity risk. Closely

related to our paper, Teo (2011) shows that the performance of hedge funds with low restrictions

to redemptions is most impacted by redemptions when aggregate conditions deteriorate. Our

incremental contribution relative to this work is to show that limits of arbitrage impact performance

by restraining hedge funds’ ability to provide liquidity. Patton and Ramadorai (2012) uncover high-

frequency variation in hedge fund risk loadings. They use the daily series of financial variables that

are similar to our measures of aggregate conditions. Our results suggest that there is heterogeneity

in hedge funds’ exposures to aggregate conditions as a function of fund-level financial constraints.

Jylha, Rinne, and Suominen (2012) regress hedge fund returns on a measure of the returns from

providing immediacy and show that liquidity provision depends on, for example, restrictions to

redemptions. We find consistent evidence by measuring liquidity provision directly at the trade

level. Also related, Gantchev and Jotikasthira (2012) use Ancerno data to show that hedge funds

are active in providing liquidity to the market for corporate control when other institutions are

selling their stakes. Finally, with respect to the theoretical literature, our results are consistent

with the claim that arbitrageurs’ liquidity provision is subject to time-varying financial constraints

6
(e.g., Gromb and Vayanos (2002) and Brunnermeier and Pedersen (2009)).

The paper is organized as follows. Section 2 describes the structure of our trade-level dataset

and the identification of hedge funds. Section 3 presents our measure of liquidity provision and

examines the time-series dynamics of trading costs contrasting hedge funds to other institutions.

Section 4 contains the fund level analysis that relates liquidity provision and performance to cross-

sectional fund characteristics. Section 5 studies hedge fund liquidity provision as a function of

different trading strategies. Section 6 relates stock-level liquidity to hedge fund trading intensity.

Section 7 summarizes the robustness analysis which is then reported in the Internet Appendix.

Finally, Section 8 offers concluding remarks.

2 Data source and descriptive statistics

We begin with a description of the institutional trading data that is used in this study. Section 2.1

discusses the data source and the information available for each trade. Section 2.2 describes the

procedure to identify hedge funds. Section 2.3 provides summary statistics for the key variables in

the dataset. Finally, Section 2.4 addresses the potential concern of a sample selection bias.

2.1 Institutional trading data

Our data on institutional trades span the January 1, 1999 to December 31, 2010 sample period.

The data provider is Abel Noser Solutions, formerly Ancerno Ltd. (we retain the shorter name

of ‘Ancerno’). Ancerno provides consulting services for transaction cost analysis to institutional

investors and makes these data available for academic research with a delay of three quarters under

the agreement that the names of the client institutions are not made public.6 An advantage of An-

cerno data is that they contain a complete and detailed record of a manager’s trading history since

the manager started reporting to Ancerno. While institutions voluntarily report to Ancerno, the

fact that clients submit this information to obtain objective evaluations of their trading costs, and

not to advertise their performance, suggests that self-reporting should not bias the data. Indeed,

the characteristics of stocks traded and held by Ancerno institutions and the return performance of
6 Priorstudies that use Ancerno data to investigate the behavior of institutional investors include Chemmanur, He, and Hu
(2009), Goldstein et al. (2009), Chemmanur, Hu, and Huang (2010), Goldstein, Irvine, and Puckett (2011), Puckett and Yan
(2011), Anand, Irvine, Puckett, and Venkataraman (2013), and Anand, Irvine, Puckett, and Venkataraman (2012).

7
the trades have been found to be comparable to those in 13F mandatory filings (Puckett and Yan

(2011), Anand, Irvine, Puckett, and Venkataraman (2012)). Another appealing feature of Ancerno

is the absence of survivorship biases in that it also includes institutions that were reporting in the

past but at some point terminated their relationship with Ancerno. Finally, the dataset is devoid

of backfill bias, as Ancerno reports only the trades that are dated from the start of the client

relationship.

The data are organized on different layers. The lowest-level observational unit is the individual

trade. Information at the trade-level includes key variables such as: the transaction date and time

(at the minute precision); the execution price; the prevailing price when the trade was placed on

the market; the number of shares that are traded; the side (buy or sell); the stock CUSIP. Ancerno

argues that among the sell trades they are also reporting short sales, which are especially relevant

for hedge funds. We cannot, however, separate regular sales from short sales. At the upper level,

the trade belongs to a daily broker release which is also called a “ticket”. At the daily ticket level,

we use the following variables: the volume-weighted average price of all trades in the market for a

given stock (VWAP); the opening price for the traded stock. In the top layer, trades are part of

a unique order, which can span several days. Ancerno provides us with two variables (lognumber

and oNumber ) whose combination helps grouping trades belonging to the same order. We double-

check the accuracy of these identifiers using the algorithm proposed by Anand, Irvine, Puckett,

and Venkataraman (2012), and opt for the latter whenever there is discrepancy between the two

procedures. At the order level, among other information, we use: the date/time of order placement;

the date/time of the last trade in the order; the market price for the stock at the time of order

decision and placement.7 Depending on the analysis, we make use of variables from different layers

of the data.

2.2 Identification of hedge fund management companies

Ancerno obtains the data from either pension funds or money managers. In case the client is a

pension fund, the trades can originate from multiple money managers. Client names are always
7 Inthe early part of the sample, there are concerns about the reliability of time stamps. For example, until 2002, 78% of orders
appear to have placement time at open (9:30 am) and last execution time at close (4:00 pm), while this fraction is just 17%
for the entire sample. From conversations with Ancerno, it turns out that the time stamps for these orders may be inaccurate.
Excluding these observations from the empirical analysis on execution time does not alter our conclusions.

8
anonymized. However, the names of the companies that are managing the clients’ portfolios are

given. This piece of information allows us to identify hedge funds among the different management

companies.

An identifier denotes the trades originating from the same management company (the variable

managercode). Also, corresponding to the company identifier, we are given the name of the man-

agement company to which the trade pertains (the variable manager ). This variable is crucial for

our identification of hedge funds. We identify hedge funds among Ancerno managers by matching

the names of the management companies with two sources. The first source is a list of hedge funds

that is based on quarterly 13F mandatory filings. This source is also used in Ben-David, Franzoni,

and Moussawi (2012) and is based on the combination of a Thomson Reuters proprietary list of

hedge funds, ADV filings, and industry listings. The second source is the Lipper/TASS Hedge Fund

Database, which contains hedge-fund-level information at the monthly frequency. In the identifi-

cation process, we make sure to select exclusively “pure-play” hedge fund management companies,

that is, institutions whose core business is managing hedge funds. This is done by applying the

same criteria as in Brunnermeier and Nagel (2004) and by manual verification. In the Internet

Appendix, we provide further discussion of the structure of the Ancerno dataset and details on the

matching procedure with these two institutional data sources.

Ancerno does not provide reliable information on the identity of the individual fund that is

executing the trade within a fund management company. For this reason, we work on trades ag-

gregated at the hedge fund management company level. Compared to other institutional investors,

such as mutual funds, aggregation at the management company level tends to be less of a concern

for hedge funds as the number of funds per company is rather small - in the order of two on average

- and the returns of funds within the same company tend to be highly correlated (Ben-David, Fran-

zoni, Landier, and Moussawi (2012)). When there is no possibility of confusion, we will refer to

hedge fund management companies simply as to hedge funds. In the end, the matching procedure

allows us to identify 87 distinct hedge fund management companies that are present in Ancerno at

various times throughout the sample.8

8 Ina recent paper, Jame (2012) also uses Ancerno to identify hedge funds following a procedure that resembles our own. He
ends up with a sample of 74 hedge fund management companies, which is somewhat smaller than our sample, possibly because
we also use the management company names in TASS for the identification. Based on the reported summary statistics, our
sample is fairly comparable to his set of hedge funds.

9
 As a validation of our matching procedure, in the Internet Appendix, we assess the extent to

which the hedge fund trades in Ancerno relate to the trades that can be inferred from 13F filings.

We find that the trades in the Ancerno dataset capture a fair amount of variation in the quarterly

holdings of the institutions that file the 13F form.

2.3 Sample selection and summary statistics

Following Keim and Madhavan (1997), we filter the data to reduce the impact of outliers and

potentially corrupt entries. In detail, we drop transactions with an execution price lower than $1

and greater than $1,000. We also eliminate trades from orders with an execution time, computed

as the difference between the time of first placement and last execution of the order, greater than

one month. Together, these filters reduce our initial sample by less than 3%. The resultant sample

consists of nearly 6.4 million of transactions in U.S. equity.

Panel A of Table 1 contains summary statistics for a number of daily series that are constructed

from the final dataset. The first row reports the number of hedge fund management companies

that are reporting on a given day. This number is on average 21, and ranges from a minimum of

10 to a maximum of 34 managers. These managers are responsible for an average of 2,214 daily

transactions (second row), or 736 orders (third row), which implies an average of about three trades

per order. The distribution of the number of trades and orders per day is however highly skewed,

with a maximum of 17,308 and 11,525 respectively. The average and median execution time (fourth

row) is about 1 day, and ranges between 0.069 (about 27 minutes) and a maximum of about 7 days.9

The last four rows in the panel provide information on dollar volume. The average daily volume is

about $350 million, with a maximum of $2.5 billion recorded on May 6, 2010. Volume per trade is

on average $229 thousand, and varies between $24 thousand and about $1 million. Finally, we look

at whether volume per trade differs across buy and sell trades. Interestingly, the volume per sell

trades tends to be larger than the volume per buy trade (averages of $240 thousand versus $226

thousand, respectively). Hence, hedge funds appear to be less concerned about reducing the price

impact of their trades when it comes to sell trades, possibly reflecting the urgency of fire sales.

This is consistent with Keim and Madhavan (1995) who find that institutions tend to split more
9 These figures are about half to one day smaller than those reported in Keim and Madhavan (1995) for other institutional
investors over the 1991-1993 period. The differences may be due to the fact that our hedge funds are not directly comparable
to their sample of institutions, and also to the overall decline in execution time over the 2000s.

10
buy trades than sell trades.

In Panel B of Table 1, similar statistics are displayed for all non-hedge-fund institutions that

report to Ancerno. These institutions include mutual funds, pension plans, and other financial

institutions that do not classify as “pure-play” hedge funds. The large bulk of these other institu-

tions consists of mutual funds. There are on average 248 non-hedge-fund managers per day during

our sample period. The number of trades and aggregate trading volume are, therefore, much larger

than for hedge funds. However, the volume per trade appears directly comparable and varies in

a similar range as for hedge funds. This implies that differences in trading costs between the two

groups are not mechanically due to systematically different trade sizes.

2.4 Is the sample representative?

Next, we tackle the important question of whether our sample of hedge funds suffers from a selection

bias. If the companies in our data are selected on the basis of characteristics that correlate with

the explanatory variable of interest (funding liquidity), the inference that we make cannot be

generalized to the entire hedge fund sector. For example, one may legitimately conjecture that the

institutions that turn to Abel Noser Solutions for consulting services are those with lower trading

skill. As such, they may be more likely to suffer when aggregate funding conditions deteriorate.

Our first reply to this concern is that the hedge funds that we study are managers for Ancerno’s

clients. As such, they are not choosing to use Ancerno’s consulting services. Rather, it is the

Ancerno clients (e.g. pension funds) that ask the hedge funds to report their trades. This fact, in

our view, goes a long way in addressing the issue of self-selection.

Second, we provide statistical evidence that further spells the concern of a self-selected sample.

In particular, if the sample is selected, we should observe a difference in loadings on the explanatory

variables between funds in the sample and the other funds in TASS. Our explanatory variables

measure funding liquidity and are explained in more detail in Section 3. For the present purposes,

it is sufficient to test whether the impact of these variables on the returns of the Ancerno hedge

funds is stronger than for the hedge funds that are in TASS, but not in Ancerno.

We test for this selection bias by running fund-level regressions of returns on the Fung and

Hsieh (2001) risk factors plus each of the explanatory variables of interest (i.e. the funding liquid-

ity factors), using the fund monthly returns available in TASS in the 1999-2010 period. To ease the

11
economic interpretation of the differences across variables, the liquidity factors have been standard-

ized to mean zero and variance unity. In Panel A of Table 2, we report the average sensitivities for

the funds in TASS that report to Ancerno and for all other funds in TASS, as well as the p-value

for the null hypothesis that their difference is zero. For the exposure to the market (RM ), TED

spread, and funding liquidity factor (PC) we find no significant difference in the exposure between

the two groups. Just in the case in LIBOR do we observe a higher sensitivity of Ancerno funds to

funding shocks. For LIBOR (p-value 0.011) and VIX (p-value 0.053) we find that funds in Ancerno

are actually on average less affected by funding condition shocks. The difference are, however,

economically rather small as they imply that a two standard deviation increase in either VIX or

LIBOR would lead to an expected difference of about 12.4bps and 25bps, respectively, between the

two groups returns. We also consider the possibility that the sample size is negatively affecting the

power of our test. This would be of special concern if the distribution of loadings for the Ancerno

funds were much larger than that of the other funds in TASS. We therefore also report the cross-

sectional standard deviation of the loadings (std) in the second row, and the p-value for the null

hypothesis that the variance of funds in Ancerno is larger than those in TASS. Only in the case

of the exposure to the market do we find that the distribution of Ancerno loadings is significantly

wider (at 1.186%) than that of the other funds in TASS (Std of 0.895%).

To provide a visual impression of the distribution of these factor loadings, we plot in Figure 1

the kernel densities of the risk loadings to each of the five funding liquidity variables. The solid line

denotes all other TASS funds, while the dotted line is for the funds in Ancerno. As we can see, the

similarity between the two distributions is substantial. In particular, the sensitivities of Ancerno

funds appears to be neither significantly skewed toward higher sensitivities nor to be sampled from

the tails of the population of TASS funds. This further validates the evidence from our previous

statistical tests that the sample of Ancerno funds is not pre-selected.

We also examine whether the number and risk profile of funds varies systematically over time

thus biasing our inference. To that end, Panel B of Table 2 presents the results of regressing the

number of funds in the database (row labeled ‘# funds’) on the funding liquidity variables. As

we can see, no predictable patterns are observed as all coefficients are largely insignificant. In

the second row, we repeat the regression but replace the dependent variable with the average risk

loading with respect to each funding variable across the funds that are reporting to TASS in a given

12
month. This conditional analysis amounts to asking whether funds that are present in Ancerno are

systematically different (in terms of their exposure to funding conditions) in periods of tighter or

looser capital constraints. The results, reported in the row labeled ‘loading’, clearly demonstrate

that this is not the case.

As a final investigation of the sample representativeness, we contrast the characteristics of funds

in Ancerno versus TASS. Specifically, we look at the distribution of: monthly returns; percentage

flows; the AR(1) coefficient of returns based on 1-year rolling windows, as a measure of asset

illiquidity; age, as proxied by the logarithm of the number of months in which the fund appears

in TASS; the amount of Leverage in place; and logarithm of assets under management, AUM.

Panel C of Table 2 reports the average value of the characteristic for the two groups of ‘TASS’ and

‘Ancerno’ funds, as well as the p-value for their difference. We note that the differences in average

returns, flows, asset illiquidity, and leverage are largely statistically insignificant. They are also

economically quite small. Just in the case of Age and AUM do we see significant differences, with

funds in Ancerno being generally older and bigger.10 If anything, the fact that funds in Ancerno

tend to be more mature should make it harder to detect the impact of limits of arbitrage, as these

funds may arguably be more unconstrained than otherwise younger funds.

Overall, our analysis shows that the hedge funds in Ancerno do not load on funding liquidity

more strongly than other funds in TASS, nor that the number and characteristics of reporting funds

differs significantly over time and with respect to TASS. Hence, we are inclined to conclude that

our sample is representative of the hedge fund universe as far as the exposure to funding liquidity

is concerned.

3 Time-series variation in liquidity provision

3.1 Measuring liquidity provision

Our analysis is inspired by the literature on the limits of arbitrage (Shleifer and Vishny (1997),

Gromb and Vayanos (2002), Brunnermeier and Pedersen (2009), Gromb and Vayanos (2010),

Gromb and Vayanos (2012)). In this literature, liquidity provision is modeled as speculators’

ability to absorb temporary order imbalances and to smooth price fluctuations (see, e.g., Brunner-
10 These two characteristics tend to be, as expected, highly correlated.

13
meier and Pedersen (2009)). The speculator in these models profits from the price pressure which

is induced by a liquidity demanding trade. Our goal is to provide the empirical counterpart to this

theoretical concept.

The standard approach in the empirical market microstructure literature is to identify liquidity

provision with limit orders and liquidity demand with market order. This strategy is not feasible

in our context, as we do not observe the order type. To overcome this limitation of the data, we

follow prior literature that works with institutional trades (Keim and Madhavan (1997), Puckett

and Yan (2011), Anand, Irvine, Puckett, and Venkataraman (2013)) and capture liquidity provision

via a measure of price impact. Price impact is related to the impatience of trading. A liquidity

providing trade typically leans against the main order flow. If it is a buy trade it is likely located

in out-of-favor stocks, whereas, if it is a sell trade, it makes stocks available that the majority of

investors are trying to buy. For this reason, liquidity-providing trades are expected to have limited

or negative price impact.

In computing price impact, we need to define the benchmark price to which the transaction

price is compared. Lacking the observation of the bid-ask quote prevailing at the time of the trade,

we rely on the volume-weighted average of trading prices for the same stock during the day in

which the transaction occurred (VWAP). This measure is originally proposed by Berkowitz, Logue,

and Noser (1988), and used more recently by Puckett and Yan (2011) who, like us, draw on the

Ancerno data. Adopting this benchmark amounts to asking how well the trader did relative to the

average transaction during the same day.11

We construct hedge fund-level trading costs as the dollar volume-weighted (T C V W ) or equally-

weighted (T C EW ) average trading cost with respect to the VWAP across all trades within a given

day. The volume-weighted daily trading cost on day t for manager i is computed aggregating across

all trades j:
 
VW
X $V olj Pj − V W APj
T Ci,t = P × Sidej (1)
j j $V olj V W APj

and the equally-weighted daily trading cost is:

X 1  Pj − V W APj 
EW
T Ci,t = × Sidej (2)
Nt V W APj
j

11 See Hu (2009) for an analysis of the properties of price-impact measures based on the VWAP.

14
where Side equals 1 for a buy and -1 for a sell trade. Analogous measures are computed separately

for buy (subscript b) and sell (subscript s) trades.

Of course our choice of benchmark, as well as the strategy to use price impact as a measure

of liquidity provision, are to some extent discretional. To reassure the reader, in the Internet

Appendix, we show that the benchmark is immaterial for our conclusions, as the results can be

replicated computing price impact using the Price at Market Open, as in Anand, Irvine, Puckett,

and Venkataraman (2013), or the Price at Order Placement, as in Anand, Irvine, Puckett, and

Venkataraman (2012). In the Internet Appendix, we also consider proxies for liquidity provision

based on reversal strategies (Lo and MacKinlay (1990) and Nagel (2012)) and trading style (Anand,

Irvine, Puckett, and Venkataraman (2013)). These alternative measures correlate strongly with our

main proxy for liquidity provision, which is reassuring about the identification strategy.

3.2 Overview of trading costs

Panel A of Table 3 contains summary statistics for trading costs expressed in basis points (bps)

pooling all fund-day observations. Over the sample, the average volume-weighted trading cost

is about 8bps with a standard deviation of 66bps. Much of this variability, however, is due to

observations in the tail of the distribution, as demonstrated by the relatively small interquartile

range (about 45bps). Equally weighted trading costs are lower on average at 4bps. This difference

suggests that large trades denote liquidity demand on average. Each of the two variables is char-

acterized by a modest degree of time-series persistence, with first-order autocorrelations of about

0.10.

In order to offer a low-frequency view on the evolution of hedge funds trading costs, we display in

Figure 2 the aggregate volume-weighted (top plot, black bars) and equally-weighted (bottom plot,

black bars) series averaged over the quarter. Several considerations are in order. Trading costs

show a substantial increase from about 0.20% at the beginning of the sample to a high of 1% in the

early 2001, a period that is characterized by the rise and burst of the Internet bubble. This finding

resonates with the results in Brunnermeier and Nagel (2004) who argue that hedge funds were

taking strong bets on overvalued technological stocks during this period and then partly reversed

their position during the subsequent downturn. In our series, trading costs start decreasing from the

2001 peak and remain at their lowest levels until mid-September of 2007, when they start increasing

15
again to reach late 2001 levels. Interestingly, the equally-weighted series remains negative for the

vast majority of the 2005 to late 2008 period, suggesting that during the last crisis some hedge

funds were providing liquidity to the market while others were engaging in massive liquidations of

their positions. This evidence lines up with the findings in Ben-David, Franzoni, and Moussawi

(2012).

A question that we address in our subsequent analysis is whether hedge funds are different

from other institutions in their liquidity provision. As a preview, we construct analogous trading

costs series using the trades of all non-hedge-fund institutions that report to Ancerno. Looking at

Panel B of Table 3, it is immediately apparent that the average trading costs for these investors

are much lower than those of hedge funds on both a volume-weighted basis (1bp versus 8bps) and

an equally-weighted basis (-3.4bps versus 4.3bps). The time-series behavior of the trading costs

experienced by the two groups is also markedly different. In Figure 2, the white bars display the

quarterly average for the volume-weighted (top plot) and equally-weighted (bottom plot) series

of trading costs to other investors in Ancerno. These series appear to be much less sensitive to

aggregate conditions than those for hedge funds. For example, trading costs in the down market

of mid-2000 to mid-2001, which was characterized by relatively high interest rates, have similar

magnitude to those observed in the boom market mid-2004 to mid-2005, with much cheaper access

to credit.

3.3 Hedge funds’ liquidity provision and funding liquidity

The upshot of the limit of arbitrage theories is that the agency relationship between the fund

managers and their capital providers can divert arbitrageurs’ actions from pursuing profitable

trading opportunity. As a result, mispricing persists and market liquidity decreases.

Due to the leeway in their trading strategies, hedge funds play the role of prototypical arbi-

trageurs. These institutions, however, need to raise and maintain their capital in order to be able

to trade. Thus, hedge funds are constrained in their actions by the need to retain and attract

investors. In addition, hedge funds make intensive use of leverage in the form of borrowed capital,

short selling, and derivative positions. This fact exposes them to a close scrutiny by their brokers

and trading counterparties. These actors stand ready to call for additional margins in case of in-

creased risk of the hedge funds’ positions, a surge in the cost of capital, and a drop in the value of

16
the collateral that is posted by hedge funds. These considerations suggest that limits-of-arbitrage

theories may well describe hedge fund behavior. If this is the case, we should observe a decrease

in hedge fund liquidity provision following a decrease in funding liquidity, which is defined as the

availability of trading capital (Brunnermeier and Pedersen (2009)).

We test this conjecture by first studying the trading behavior of our sample of hedge funds

in the U.S. equity market. In doing that, we face the challenge of identifying truly exogenous

variation in hedge funds’ funding liquidity. Fund flows are certainly related to the availability of

trading capital. However, they are hardly exogenous as investors react to a rational anticipation of

future performance, which in turn is related to hedge funds’ liquidity provision. Thus, we choose to

measure funding liquidity using financial variables that proxy for the prevailing funding conditions.

Our assumption is that the evolution of these variables does not depend on hedge funds’ liquidity

provision in the future. Under this assumption, changes in aggregate conditions can be used as

exogenous sources of variation in funding liquidity.

Based on the findings in Hameed, Kang, and Viswanathan (2010) that liquidity supply by

financial intermediaries is positively related to market performance, we test whether the hedge

fund sector decreases liquidity provision following a decline in the stock market. These authors’

implicit assumption is that the financial intermediation sector has a positive net position in stocks.

Thus, losses in the stock market entail a deterioration of liquidity providers’ capital. In the context

of hedge funds, this assumption seems to apply as well. For example, Fung and Hsieh (2004) show

that the average hedge fund has a positive beta on the S&P 500. Even some of the funds in the self-

declared market-neutral style, appear to have a significant exposure to the market factor (Patton

(2009)).

The extent to which hedge funds can lever up their positions affects their ability to correct

mispricing and provide liquidity. Patton and Ramadorai (2012) show that hedge funds’ risk expo-

sures are significantly related to the TED spread (the three-month LIBOR minus the three-month

T-bill rate). To explain their finding they refer to Garleanu and Pedersen (2011), who argue that

the interest rate difference between collateralized and uncollateralized loans (or Treasury securi-

ties) captures arbitrageurs’ shadow costs of funding. Thus, we also use the TED spread to proxy

for systematic time-series variation in funding liquidity. Moreover, the leverage available to hedge

funds varies with the costs of borrowing. Thus, we use the LIBOR to proxy for the level of interest

17
rates.

Finally, Brunnermeier and Pedersen (2009) argue that the margins imposed by brokers to

arbitrageurs depend on the volatility of asset prices. In their paper, brokers set margins according

to Value-at-Risk models, for which volatility is the main input. Because we want to focus on

aggregate variables, we use the VIX to measure the impact of volatility on liquidity provision by

the hedge fund sector through the margin requirements channel.

To summarize, we formulate and test empirically the following

Hypothesis 1: Hedge funds’ liquidity provision in the stock market is positively related to aggre-

gate measures of funding liquidity. In particular, liquidity provision depends:

· Positively on the past performance of the market

· Negatively on the TED spread
· Negatively on the LIBOR
· Negatively on the VIX.

Panel C of Table 3 reports summary statistics for the four financial variables in Hypothesis

1. At the daily frequency, these variables are quite persistent, but they experience substantial

variation during the prolonged period we consider.12 In addition, we also use the first principal

component of these four variables, which we label PC and we refer to it as to the funding liquidity

factor. From the correlation matrix, we notice that PC is strongly negatively related to the market

return and positively to VIX and TED, and has a small positive correlation with LIBOR.

We test Hypothesis 1 by means of the following linear regression model:

T Ci,t+1 = a + bF undLiqt + φT Ci,t + ǫi,t+1 (3)

where T Ci,t+1 is hedge fund i trading cost on day t + 1, and F undLiqt denotes alternatively one

of the five funding liquidity determinants (RM , VIX, TED, LIBOR, or PC) measured on day t.

We include lagged trading costs to account for the small persistence in this series that is observed

in Table 3. To ease the discussion, here and throughout our subsequent regressions, we express

trading costs in basis points and standardize the variables in F undLiqt , so that the coefficient b

represents the expected change in trading costs (in bps) following a one standard deviation increase
12 Forexample, the VIX ranges from a minimum of 0.10 in November, 2006 to a maximum of about 0.81 in October, 2008.
Similarly, the LIBOR rate varies from 0.2% throughout 2010 to as high as 6.9% in the mid-2000.

18
in F undLiqt . The model is estimated by pooling all fund-day observations over the full 1999-2010

sample.13

The left-hand-side block of Panel A of Table 4 presents the estimates for b when trading costs

are volume-weighted. The five different specifications for F undLiqt are reported in columns (1)

to (5), respectively. Below the estimates, t-statistics based on robust standard errors clustered at

the date level are reported in parentheses. In the first specification, we relate liquidity provision

to RM . The coefficient on RM is found to be negative at -1.515, and statistically significant at the

1% level. Hence, a one-standard deviation negative shock to stock market valuations is associated

with weaker liquidity provision by hedge funds whose trading costs increase on average by about

1.5 basis points. In column (2), we look at the relation between liquidity provision and the VIX.

The coefficient is positive at 1.750, with an associated t-statistic of 4.79. A slightly stronger effect

is found for the TED spread, with a coefficient of 1.925 and a t-statistic of 5.12, and especially for

LIBOR whose effect is largest at 3.220, or 3.2 basis points. Given the unconditional mean of trading

costs of about 8 basis points (Table 3), these effects appear, economically speaking, quite large.

Finally, the loading on the funding liquidity factor PC is positive at 2.618, and highly statistically

significant.

In the right-hand-side block of the panel, the same analysis is repeated for equally-weighted

trading costs. Similarly to the volume-weighted results, all coefficients are significant and have the

expected sign: they are negative for RM , and positive for the other four variables. However, the

economic impact of the variables is generally weaker. We impute this reduction in significance to

the fact that the equally-weighted measure dampens the impact of large trades, which are mostly

relevant to measure liquidity demand.

In Panel B of Table 4, we look at the results for volume-weighted trading costs when these are

separately calculated for buy and sell trades. Owing to their ability to take short positions, which

are part of the sell trades reported to Ancerno, hedge funds are in the position of taking advantage of

both upturns and downturns. Thus, it is interesting to examine whether time-variation in liquidity

provision characterizes differently the two sides of the market. The statistical significance of the

funding liquidity determinants characterizes both buys and sells, but the coefficients for sell-side
13 As a robustness check, Internet Appendix D shows that the main findings hold also when using data aggregated at the weekly
frequency.

19
trading costs are substantially larger in absolute value for all variables but TED. The difference is

particularly large for RM (-0.830 versus -1.727) and for the VIX (1.162 versus 2.771).

Next, we investigate the extent to which our results are affected by the 2007-2009 financial

crisis. This is a legitimate concern due to the large variations experienced by the funding liquidity

variables during that period. To that end, we repeat our analysis separately for the crisis and

ex-crisis period. Following Anand, Irvine, Puckett, and Venkataraman (2013), we define the crisis

period as spanning the January 2007 to May 2009 sample. Table 5 reports the estimates of equation

(3) in the two samples. Regardless of the sample considered, all coefficients have the expected sign

and their magnitude and significance is comparable to the estimates appearing in Panel A of Table

4.14 These findings lend support to the view that time-variation in hedge funds’ liquidity provision

represents a pervasive behavior, and not a crisis-only phenomenon.

In order to measure the impact of the funding liquidity variables on liquidity provision it is

essential to keep ‘order difficulty’ fixed. That is, it is possible that a deterioration of aggregate

conditions determines also a shift in the characteristics of hedge fund trades that drives up trading

costs, without an actual change in the attitude towards liquidity provision. Here, we address

changes in order difficulty that are observable at the trade level. In the next subsection, we develop

an approach to control for aggregate changes in order difficulty that affect all traders. We modify

equation (3) to include appropriate controls for order characteristics that allow us to keep the

observable order characteristics fixed:

T Ci,t+1 = a + bF undLiqt + δ′ Zi,t + εi,t+1 (4)

where the vector Zi,t collects the following variables, which are partly inspired by prior literature

(e.g. Anand, Irvine, Puckett, and Venkataraman (2012)): Buy, a dummy that equals 1 for buy

trades, and 0 otherwise; Lagged Return, the stock return in the prior day; NYSE, a dummy that

equals 1 for stocks listed at the NYSE, and 0 otherwise; Inverse Price, the inverse of day-t stock

price; Relative Volume, the ratio between the number of shares traded to the average volume of the

stock in the prior 30 days; Amihud, the Amihud illiquidity ratio; Size and Book-to-Market, the stock
14 The effect of RM is somewhat stronger in the crisis period, while that of VIX is comparable at about 1.9. In contrast, the
coefficient on TED is more than twice as large in the ex-crisis period, while LIBOR is insignificant in the 2007-2009 sample.
Finally, the loading on the principal component PC is a large 4.304 in the ex-crisis period compared to a 2.881 figure in the
crisis sample.

20
market capitalization and book-to-market deciles. All variables are computed as volume-weighted

averages at the fund-day-side level. These controls ensure that our inference is not driven by the

fact that hedge funds trade different volumes or in stocks with different degrees of liquidity.

The estimates of equation (5) are reported in Panel A of Table 6. They do not modify our

conclusions as the slopes on the funding liquidity variables are not substantially impacted by the

inclusion of trade-level characteristics. The signs of the control variables line up with the findings

of prior literature and conform to expectations. Overall, the results support the prediction in

Hypothesis 1 that periods characterized by tighter funding constraints in the form of a reduction

in the value of collateral, a surge in aggregate uncertainty or in risk premia as captured by the

VIX, and costlier access to credit are accompanied by a decrease in hedge funds’ ability to provide

liquidity.

3.4 Hedge funds vs. other institutions

Our empirical approach to capture liquidity provision consists of measuring the price impact of

hedge fund trades. Price impact, however, is not only related to the net demand of liquidity in

a trade, but also to the cost of liquidity in the market. Put simply, the same liquidity-motivated

buy order of 10,000 shares of IBM could have different price impacts in different market conditions.

Likely, when aggregate conditions deteriorate, the price of liquidity is higher. These considerations

highlight an important empirical challenge in identifying variations in liquidity provision as a func-

tion of aggregate funding liquidity. The component of price impact due to arbitrageurs’ withdrawal

from liquidity provision, which we wish to identify, is subject to the confounding effect from the

increase in the cost of liquidity in bad times.

To separate out the liquidity-provision effect, we benchmark hedge fund trading costs to those

experienced by other investors in comparable market conditions, as in a difference-in-differences

approach. Besides hedge funds, our data cover other institutional investors, mutual funds and pen-

sion funds, which we use to control for market-wide variation in the cost of liquidity. The rationale

for our strategy is that the aggregate increase in trading costs is filtered out when comparing hedge

funds to other institutions.

For this set of investors to serve as a valid control group, however, it has to be the case

that their liquidity provision is subject to changes in funding conditions to a smaller extent than

21
for hedge funds. This seems to be a fair assumption. For example, mutual funds make only

limited use of leverage, if at all, which makes them less exposed to changes in borrowing conditions

and volatility. Also, the convexity of the flow-performance relation (Chevalier and Ellison (1997),

Sirri and Tufano (2002)) implies that mutual fund flows are relatively less sensitive to poor past

performance than hedge fund flows, for which the flow-performance relation has been found to

be linear (Li, Zhang, and Zhao (2011)) or even concave (Ding, Getmansky, Liang, and Wermers

(2009)). Related to this point, the hedge funds’ institutional clientele is likely to react more quickly

to changing performance than the retail investors in mutual funds because of a more structured

investment process and risk management considerations. Supporting these arguments, Ben-David,

Franzoni, and Moussawi (2012) show that during the recent crisis hedge funds engaged in more

important portfolio liquidations than mutual funds, and link this different behavior to the more

significant outflows from the hedge fund sector, the leverage, and the institutional clientele of hedge

funds.

Thus, the need to benchmark hedge funds’ trading costs to those of other institutions brings us

to formulate the following

Hypothesis 2: Capital availability for hedge funds is more subject to aggregate conditions than

for other institutions (mutual funds and pension funds). As a result, the sensitivity of hedge

funds’ liquidity provision to funding liquidity is stronger than for other institutional investors.

To test Hypothesis 2, we estimate the interacted model:

T Ci,t+1 = a + b1 F undLiqt + b2 HF × F undLiqt + δ′ Zi,t + εi,t+1 (5)

where HF is a dummy variable that equals 1 if institution i is a hedge fund, and 0 otherwise. The

coefficient b1 measures the expected impact of a one-standard deviation increase in F undLiqt on

the trading costs of investors that are not hedge funds. The loading on the interaction term, b2 ,

captures instead the additional effect of the same shock on hedge funds’ liquidity provision. This

is the coefficient of interest to us for testing Hypothesis 2. The controls in Zi,t are meant to keep

trade-level characteristics fixed, as discussed in relation to equation (4).

Estimates of the model in (5) are reported in Panel B of Table 6. The top row of the table

indicates the funding liquidity variable that is used in the corresponding column. A few facts are

22
noteworthy. First, the interaction term b2 is always statistically significant at the 1% level or better.

Consistently with Table 4, the coefficient is negative for the market in column (1), and is positive

for LIBOR, VIX, TED, and PC. Therefore, there is an additional significant impact of funding

liquidity shocks on hedge funds’ trading costs relative to those of other institutions. Second, the

estimates for b1 are positive for LIBOR but negative and significant for VIX, TED, and PC. This

implies that the trading costs to other institutions decrease in certain periods of market stress, when

they are acting as liquidity providers, consistent with Anand, Irvine, Puckett, and Venkataraman

(2013). Third, b2 is about an order of magnitude larger than b1 . Since b1 captures the baseline

effect for all institutions, we conclude that only a limited part of the sensitivity of hedge funds’

trading costs we documented in the previous section is due to aggregate fluctuations in the cost of

liquidity that affect other investors as well. For example, in the case of VIX the coefficient b1 is

negative at -0.813, but it becomes a positive 2.784 for the hedge fund interaction term. Overall,

the results are consistent with Hypothesis 2. Also, it appears that in periods of tight funding, as

measured by the VIX, the TED spread, and PC, other institutions than hedge funds shift towards

liquidity provision.

3.5 Characteristics of traded stocks and execution speed

Another way of capturing the reaction of hedge funds to capital availability is to look at changes

in the type of stocks in which they trade. Working on Ancerno data, Anand, Irvine, Puckett, and

Venkataraman (2013) show that in periods of market stress buy-side institutions trade more on

liquid stocks and that this flight-to-liquidity has first-order effects for pricing. If hedge funds are

acting as ultimate liquidity providers, then we would expect them to trade on less liquid stocks

that are neglected by other investors and offer appealing premia for liquidity. Alternatively, they

might decide to liquidate more volatile and illiquid stocks because they are costly to hold in terms

of margin requirements, as argued by Brunnermeier and Pedersen (2009). In either case, because

trading costs are on average higher for illiquid stocks, and their trading costs may very well increase

during market stress, the change in the composition of hedge fund trades may be responsible for

part of the time-variation in trading costs that we find.

To test for this composition effect, we again use a fully interacted model where the dependent

variable is now the volume-weighted average decile (VWAD) of the stocks traded by a given insti-

23
tutional investor on day t. The deciles are computed based on the whole universe of CRSP stocks,

separately along the following two relevant dimensions of liquidity: market capitalization (Size),

and the Amihud (2002) ratio, which measures price impact and, therefore, captures illiquidity. The

deciles are ordered from 1=LOW to 10=HIGH values of the corresponding variables. Thus, the

top Size and the bottom Amihud deciles contain the most liquid stocks. The following regression

model is then estimated:

V W ADi,t+1 = a1 + a2 HF + θ1 F undLiqt + θ2 HF × F undLiqt

+φ1 V W ADi,t + φ2 HF × V W ADi,t + υi,t+1 (6)

Here, the coefficient θ2 measures the additional tilt in the degree of liquidity of the stocks traded

by hedge funds compared to other investors, following the shock in F undLiqt .

Panel A of Table 7 presents the results for Size V W AD. The first row of coefficients reports

the estimates of the interaction term, θ2 . The coefficient on the market return, RM , is negative

and significant at the 5% level. The estimates of VIX, TED, LIBOR, and PC are instead all

positive, and they are statistically significant at the 1% level, with the exception of LIBOR. On

the aggregate, it appears that hedge funds react to periods of tightening constraints by tilting

their trades towards liquid stocks. This effect is not only statistically, but also economically large

if compared to the behavior of other institutions. For example, the loading on VIX is 0.018 for

non hedge funds investors, while it is 0.058 (0.040+0.018) for hedge funds. Similarly, the effect of

market returns is almost twice as large for hedge funds (-0.007 versus -0.016). Interestingly, the

coefficient on TED for the other investors is negative and significant, which suggests that positive

shocks to the TED spread tend to be followed by liquidity provision in smaller, neglected stocks.

Looking at the combined effect of the liquidity factor PC, the effect of funding liquidity appears to

be four times larger for hedge funds than for the other institutions reporting to Ancerno. Panel B

of Table 7 displays the results for the Amihud V W AD, which lead to similar conclusions. The fact

that hedge funds react to funding liquidity shocks by moving toward liquid stocks, and that they

do so much more aggressively than other institutional investors, is in stark contrast with the claim

that they act as liquidity providers of last resort.

We can delve deeper into hedge funds’ demand for immediacy by studying their trading speed.

24
The speed of order execution, measured by the time between order placement and completion, is

related to whether the order is placed in the form of a limit or market order. Market orders are

typically characterized by a shorter execution time than limit orders, as they serve the needs of

investors with a high appetite for trading immediacy. Viceversa, liquidity provision is generally

associated with limit orders. Our data do not provide information on the order type. Still, according

to this logic, if hedge funds respond to a tightening in funding constraints by demanding liquidity,

we should observe a reduction in order duration, which is a variable that we can measure.

We compute the duration of hedge funds orders as the span (in minutes) between the time of

placement of the first trade in an order and the time of execution of the last trade in the same

order. The regression equation for testing our conjecture is:

ExT imei,t+1 = a + βF undLiqt + γ ′ Zi,t + ui,t+1 (7)

where ExT ime is the logarithm of the average duration across all orders placed by hedge fund i

on day t + 1, separately computed for buy and sell trades. Zi,t denotes the following four cross-

sectional controls: Side, Relative V olume, Size V W AD, and a time trend (T rend) capturing the

dependence of execution time on the prevailing characteristics of the stock exchange, such as the

progressive introduction of automated trading (Hendershott, Jones, and Menkveld (2011)). The

timing of the variables is organized so that β measures the expected impact of the aggregate shock

in F undLiqt on the duration of future orders.

The estimates of equation (7) are presented in Table 8. Two aggregate variables, the VIX

and the principal component PC, stand out as having a statistically significant negative effect on

execution time. The coefficients on these variables imply that a deterioration in funding liquidity

is followed by a reduction in the duration of hedge funds orders of about 5 to 9 percent. Given an

average order duration of about 1 trading day or 390 minutes (Table 1), a one standard deviation

increase in PC and VIX is accompanied by a decrease in the execution time of about 20 and 35

minutes, respectively. Importantly, notice that a higher VIX predicts a decrease in execution time.

The direction of the effect is the opposite of what we would expect if the results were mechanically

driven by the drop in liquidity that typically follows an increase in VIX (Nagel (2012)). Rather,

they are consistent with a stronger demand for immediacy by hedge funds in stressed market

25
conditions.15

4 Cross-sectional analysis

In this section, we refine the identification of the limits of arbitrage by exploiting cross-sectional

heterogeneity in hedge funds and formulate our Hypothesis 3. Also, we study whether the shift in

hedge fund positions that follows from a shock in funding liquidity impacts the trading performance

of more constrained hedge funds. In principle, if arbitrageurs are obliged to move away from their

preferred trading strategies, their performance should suffer.

4.1 Liquidity provision and hedge fund characteristics

A number of hedge fund characteristics are likely to determine different sensitivity of liquidity

provision to changes in aggregate funding conditions. For example, higher leverage makes a fund

more exposed to changes in the cost of debt and in margin requirements. Then, we expect highly-

levered hedge funds to withdraw their liquidity provision more strongly in bad times.

The ease with which investors can redeem their capital makes a hedge fund more susceptible to

redemptions in case of poor performance or deteriorating aggregate conditions. According to this

argument, we should expect that the liquidity provision of hedge funds with lower share restrictions,

e.g. lower redemption notice period, is more sensitive to aggregate funding conditions. Consistent

with this conjecture, Jylha, Rinne, and Suominen (2012) find that hedge funds with longer lockup

periods have a higher propensity to supply immediacy. The strength of this prediction can be

attenuated by the fact that share restrictions are set endogenously to provide a buffer to funds

that invest in illiquid assets (Ding, Getmansky, Liang, and Wermers (2009)). If illiquid assets are

more exposed to aggregate conditions, funds with stronger share restrictions could actually end up

increasing their liquidity demand in bad times. Which effect prevails is ultimately an empirical

question.

Related to the last point, hedge funds with an important component of illiquid assets in their
15 The statistical and economic significance of the aggregate funding shocks in Table 8 holds also after the inclusion of the
control variables. Side is positively related to execution time, but not significantly so. The execution time is shorter for
bigger stocks, that are usually characterized by a higher degree of liquidity. This is reaffirmed by the estimate of about -0.10
for the coefficient on Size V W AD. The loading on RelativeV olume is positive and highly significant, that is, orders that
are large relative to the normal volume of a stock take longer time to be executed. Finally, the time trend is negative and
significant, and corresponds to a decrease in trading time of about 8% per year.

26
portfolios may be more likely to alter their provision of liquidity in the stock market if funding

conditions deteriorate. The logic is that a negative shock to the illiquid part of their portfolio may

force them to liquidate their more liquid positions, which qualifies as demand for liquidity. Manconi,

Massa, and Yasuda (2012) provide evidence for the bond market during the recent financial crisis

that is consistent with this story. Following Getmansky, Lo, and Makarov (2004), we measure the

illiquidity of a fund’s portfolio using the first-order autocorrelation of the fund returns. We expect

the liquidity provision of more illiquid funds to be more strongly related to aggregate funding

conditions.

The extent to which hedge funds can preserve their trading capital when facing adverse condi-

tions also depends on their reputational capital. An established hedge fund can more convincingly

negotiate the lending terms with brokers and prevent investors from leaving the boat than a young

fund. For similar reasons, funds with a shining track record are more credible vis-a-vis brokers and

investors than poor-performers. Thus, we expect the sensitivity of hedge funds’ liquidity provision

to aggregate conditions to be stronger for young and poor-performing funds.

To summarize, we formulate the following

Hypothesis 3: The sensitivity of hedge funds’ liquidity provision to aggregate funding liquidity

is related:

· Positively to leverage
· To redemption restrictions with an ambiguous sign
· Positively to asset illiquidity
· Negatively to fund age
· Negatively to past performance.

To test Hypothesis 3, we regress the measure of hedge fund liquidity provision (T C) on the

interaction between aggregate funding liquidity and measures of the hedge fund characteristics

that are meant to capture limits of arbitrage. The regression that is run on hedge fund-day level

data is:

T Ci,t+1 = α + β ′ F undLiqt + γ ′ Xi,m−1 + η ′ F undLiqt × Xi,m−1 + ǫi,t+1 (8)

where F undLiq is, as before, alternatively RM , VIX, TED, LIBOR, or PC. The vector X collects six

cross-sectional characteristics, X = {Leverage, Y oung, Illiquid, Bad, LowRed, N oLock}. Notice

27
that the fund characteristics are obtained from the monthly TASS data. Hence, they are constant

within a month and are based on prior month information (month m − 1). Leverage is the amount

of leverage in place; Y oung is minus the logarithm of the number of months in which the fund

appears in TASS; Illiquid is the decile of the distribution of first-order autocorrelation in returns;

Bad is minus the fund’s year-to-date performance; LowRed is a dummy variable that equals 1 if

the sum of redemption notice period and redemption frequency is lower than 120 days (the sample

median); and N oLock is a dummy variable that equals 1 if the fund has zero lockup period. Except

for the variables measuring share restrictions (lockup and redemption periods), for which the effect

is a priori ambiguous, the fund-level regressors are defined so that a higher score captures higher

expected limits of arbitrage. In equation (8), our main interest is on the slope η that captures

the differential effect of funding liquidity shocks, originating from fund-level characteristics, on the

implicit trading cost.

Table 9 reports the estimates of equation (8) for the different funding liquidity variables. The

regression also includes hedge-fund style dummies. The dependent variables is trading costs ex-

pressed in basis points. The bold face highlights the coefficients among the η estimates that are

consistent with Hypothesis 3 (with 10% significance).

In the first column, the prior-two-weeks market return plays the role of the funding liquidity

variable. For the fund-level characteristic to capture limits of arbitrage, we expect the slope on

the interaction to be negative. An improvement in funding conditions is expected to benefit more,

in terms of lower trading costs, the more constrained funds. The only significant interaction is for

the no-lockup period dummy. Its positive sign confirms the ambivalent prediction of Hypothesis

3 with respect to the effect of share restrictions. The evidence is consistent with the view that

share restrictions are set endogenously by funds that trade in illiquid assets. In this sense, an

improvement in market conditions benefits more the funds with a lockup period in place, which

invest in more illiquid assets. This interpretation of the role the lockup period is confirmed across

specifications.

In the other specifications of Table 9, an increase in the funding liquidity variable signals a

deterioration of aggregate conditions. Hence, the expected sign for the regressors in X is reversed

relative to column (1). In column (2), VIX measures funding liquidity. A deterioration of aggregate

conditions, as signalled by an increase in the VIX, increases more significantly the trading costs

28
of young and poor-performing funds. This is consistent with Hypothesis 3 as an increase in these

variables reflects a lower reputational capital for the fund. As for the negative and significant

sign on the interaction between LowRed and VIX, it confirms the evidence above about N oLock.

That is, funds with higher share restrictions also have more illiquid assets, which harms liquidity

provision in bad times.

In columns (3) and (4), the funding liquidity variables are the TED spread and the LIBOR,

which capture credit market conditions. As expected from Hypothesis 3, a tightening of credit

harms more strongly the trading costs of the more levered funds. So, leverage appears to act as

a fund-level limit to arbitrage in bad times. In the same two columns, low share restrictions (as

measured by LowRed) magnify the impact of a deterioration of aggregate funding conditions on

trading costs. Confirming the ambivalence of Hypothesis 3, share restrictions appear in this case

to ease the fund-level constraints by shielding the managers from outflows in bad times. In terms

of magnitude, the estimates in column (3) suggest that a one-standard deviation increase in the

TED spread, increases trading costs by about 14 bps more for funds with low share restrictions

(LowRed = 1). This seems like an economically important effect, given that the standard deviation

of T C is about 66 bps (see Table 3). The positive and significant estimates on the interactions with

Illiquid and Y oung confirm Hypothesis 3 (while Bad is consistent only in column 4) . Finally,

column (5), in which deteriorations in funding liquidity are measured by the funding liquidity factor

(PC), reiterates some of the previous results.

The analysis in Table 9 is overall consistent with the predictions of Hypothesis 3. Based on

this evidence, we ask the question of whether the larger effect of aggregate funding liquidity on

the trading costs of hedge funds relative to other institutions (see Table 6) can be relegated to

the subset of hedge funds with stronger limits of arbitrage. This is a relevant point, as it seems

implausible that the entire hedge fund sector is in a worse position than other institutions to provide

liquidity in bad times. Rather, one would expect the results in Table 6 to be driven by the more

constrained hedge funds.

To test this conjecture, Table 10 replicates the analysis of Table 6 for the two subsets of con-

strained and unconstrained funds. Based on the analysis in Table 9, a constrained fund is defined

has one that has both non-zero leverage and LowRed equal to one. The comparison between Panel

A (constrained funds) and Panel B (unconstrained funds) suggests that the differential impact

29
of the macro variables on hedge funds (slope on HF×F undLiq) is concentrated in the group of

constrained funds.

This result gives an important qualification to our previous findings. The statement that hedge

funds are more constrained than other institutions in their ability to provide liquidity in bad times

has to be confined to the subset of hedge funds with fund level characteristics that magnify their

exposure to shifts in funding conditions. The rest of the hedge fund universe can provide liquidity

at least as well as other institutions.

4.2 Performance and hedge fund characteristics

The analysis conducted so far is consistent with the view that hedge funds are forced by limits

of arbitrage to unwind their positions when funding conditions tighten. One potential alternative

description of the evidence, however, is that hedge funds are optimally timing the market and

closing down positions in anticipation of further deterioration of aggregate conditions. In support

of this alternative conjecture, one may argue that the lack of an explicit benchmark allows hedge

funds to change their positions promptly in the face of changing conditions.

To disentangle these two hypotheses, we study the ex-post performance of the trades carried out

by constrained hedge funds as a function of aggregate funding conditions. If limits of arbitrage are

playing a role, we should observe underperformance of the stocks that are traded by constrained

funds when aggregate conditions deteriorate, as these transactions represent a forced deviation

from an optimal path. Underpeformance can originate from two sources. First, these positions

are initially held by hedge funds because they are expected to be profitable. Closing them down

corresponds to a missed profit opportunity. Second, the evidence on trading costs and execution

time in Section 3 is indicative of rushed liquidations. Then, hedge fund trades are going to have a

price impact which reverts over the next days. In either case, we expect the stocks that are sold

by constrained hedge funds to increase in value, while the stocks that are bought (also with the

purpose of closing short positions) should drop in value, in the days that follow the transactions.

For each hedge fund-day, we compute the ex-post volume-weighted performance between trade

execution (using the execution price provided by Ancerno) and the closing of day t + 5 (using the

closing price in CRSP), for all the trades occurring on day t. The choice of a five-day horizon is

suggested by the consideration that, if these forced liquidations have a price impact, it is going to

30
revert quite soon. Returns of stocks that are sold enter the total return with a negative sign, as

in Puckett and Yan (2011). We then take the volume-weighted average of the five-day abnormal

returns of all the stocks that are traded on day t by fund i. The abnormal return is computed

relative to a market model, for which beta is estimated on the prior sixty days. We estimate the

following regression, in which the dependent variable is the abnormal return (in %) of fund i’s day-t

trades between execution and t + 5:

ri,t:t+5 = α + β ′ F undLiqt + γ ′ Xm−1 + η ′ F undLiqt × Xm−1 + ǫi,t:t+5 (9)

where F undLiq is alternatively RM , VIX, TED, LIBOR, or PC. The vector X collects the six

cross-sectional characteristics in the prior month (m − 1), X = {Leverage, Y oung, Illiquid, Bad,

LowRed, N oLock}, that are described in relation to equation (8). Once again, our interest is on the

vector η that captures the additional effect of funding liquidity shocks on the future performance

of a constrained fund.

Table 11 reports the estimates for the regression in equation (9). Returns are measured in

percent. The first five rows of the table display the estimates of η. The bold face is for the

coefficients that provide significant evidence in favor of the limits of arbitrage hypothesis. While

not all slopes in η meet statistical significance, it is interesting to notice that the significant ones

point in the same direction and are consistent with the limits of arbitrage hypothesis. For example,

from column (1), we infer that the stocks traded by more levered funds outperform the stocks

traded by less levered funds, when the market goes up. This seems to suggest that an improvement

of aggregate funding conditions eases the constraints on funds that have stretched their borrowing

capacity. The effect of leverage is consistent with limits of arbitrage also in columns (2), (4),

and (5), where the negative signs are due to the fact that the funding liquidity variables measure

a deterioration in aggregate conditions. Further confirmations for limits of arbitrage come from

Y oung and LowRed, in column (2), and from Illiquid in column (4). To gauge the economic

magnitude, we can focus for example on the estimate for Illiquid in column (4). A one-standard

deviation increase in the LIBOR, reduces the five-day abnormal return of a fund in the top illiquidity

decile by about 36 bps more than for a fund in the bottom decile of illiquidity (-0.04%×9=0.36%).

This seems like a significant number as it translates into a 18% decrease in annual market-adjusted

31
performance (-.36%×250/5 = 18%).16

Overall, the analysis in this section provides support of the theories that posit limits on liquidity

provisions as a function of investors’ ability to attract and retain capital. Our evidence suggests

that hedge funds that are more severely constrained are likely to perform relatively worse than

other funds when funding liquidity tightens.

5 Liquidity provision and hedge funds’ trading style

Another dimension of the cross-sectional heterogeneity in the hedge fund sector relates to hedge

funds’ trading styles. For a hedge fund, its trading strategy is the ultimate determinant of the

availability of funding liquidity. For example, different trading strategies require different levels of

leverage to be profitable. A fund whose trading strategy requires higher leverage may be obliged

to forced liquidations in bad times. Hence, when aggregate conditions deteriorate, funds pursuing

different strategies may be more or less able to provide liquidity to the market. Some funds may

even be obliged to demand liquidity in stressed markets.

Hedge funds report a broad classification of their trading style to data providers. However, these

categories contain little information, especially for hedge funds operating in the equity domain. The

majority of funds in TASS appear in the ‘long-short equity’ style, but this does not say much about

the kind of stocks they actually invest in. Our data give us the unique opportunity to obtain a

timely snapshot of the stocks in which hedge funds trade. Differently from the quarterly holdings

in the 13F filings, Ancerno reports each trade over the period in which the funds appear in the

data set. This fact puts us in a privileged positions, as a high-frequency perspective is crucial for

inferring hedge funds’ trading strategies.

We focus on three popular trading styles in the equity space. The first one corresponds to short-

term reversal strategies, which provide liquidity in stocks that experience temporary price pressure.

This style is popular, for example, among those high-frequency traders that exploit signals from

prices and volume. The second one is based on the momentum anomaly (Jegadeesh and Titman
16 Somereaders wonder whether the constrained funds’ trading performance is lower when funding conditions deteriorate because
they switch to more profitable opportunities that open up in other asset classes. According to this story, to be able to promptly
rebalance their portfolio, the constrained funds are willing to accept small losses on their equity investments. In Internet
Appendix G, we provide evidence that seems to rule out this alternative, as constrained funds suffer more from the point of
view of total returns than unconstrained funds when funding conditions tighten.

32
(1993)). The strategy consists of buying prior-year winners and shorting prior-year losers. The

third one is a traditional value strategy (Fama and French (1993)) going long in undervalued stocks

(high book-to-market) and short in overvalued stocks (low book-to-market). Each fund in a given

month ranked along each of these three dimensions. Then, we study the fund’s behavior in terms

of liquidity provision, both unconditionally and as a function of aggregate conditions.

In more detail, the definition of a short-term reversal strategy follows Nagel (2012). We assign

CRSP stocks to reversal deciles each day based on the average of cumulative returns in excess of

the equally-weighted market return over the past 1 to k days, for k = {1, 2, 3, 4, 5}. We assign

scores from -5 to -1 for the 1st to 5th deciles, and scores form 1 to 5 for the 6th to tenth deciles.

Next, for each hedge fund and month we compute the value-weighted average Score based on trades

in that month, taking into account the sign of the trades.17 For long-term momentum and value,

we reconstruct hedge fund holdings from their trades and use their inferred portfolios to rank the

funds along the two dimensions. Specifically, we cumulate each hedge fund’s trades starting from

the first month a fund appears in Ancerno. Since we do not observe the initial positions, we allow

for a burn-in period that is then discarded.18 We report results for a burn-in period of 24 months,

but the results are insensitive to shorter burn-in periods of, e.g., 6 months. Then, at the end of

each month, we assign CRSP stocks to deciles (LOW to HIGH) separately based momentum (the

return from month -12 to month -1), and book-to-market. We assign scores from -5 to -1 for the

1st to 5th deciles, and scores form 1 to 5 for the 6th to 10th deciles. Finally, for each hedge fund

and month we compute the value-weighted average Score, across the two dimensions, taking into

account the sign of the trades.19

We define dummy variables to classify funds that follow, respectively, a short-term reversal

(rev = 1), long-term momentum (mom = 1), or value (value = 1) strategy if their end-of-month

reversal, momentum, or book-to-market scores are above zero.20 This procedure is repeated at the

end of each month to capture changes in the trading style of the fund. We use this classification

to examine hedge funds’ trading costs and their sensitivity to funding conditions in the following
17 Specifically, we first compute the value-weighted average Score for buy and sell trades in that month, separately. We then
combine the two sides into Score = (ScoreBuy ∗ V olumeBuy − Scoresell ∗ V olumeSell )/(V olumeBuy + V olumeSell ). Puckett
and Yan (2011) apply a similar procedure to compute the return to interim trading. The resultant Score ranges from -5 to 5.
18 This is equivalent to assuming that the hedge fund would have turned over its portfolio entirely by the end of this period.
19 Specifically, we first compute the value-weighted average Score for buy and sell trades, separately. We then combine the two

sides as Score = (ScoreBuy ∗ V alueBuy − Scoresell ∗ V alueSell )/(V alueBuy + V alueSell ), where V alue is the product of
number of shares (holdings) times price at the end of the month.
20 Using the median instead of zero produces quantitatively similar results.

33
month, thus reducing the risk of endogeneity.

Table 5 reports the results in the daily pooled regression of hedge funds’ trading costs on their

trading style (rev, mom, or value), and its interaction with each funding liquidity variable. In

Panel A, we look at the level of the trading costs. The focus in this panel is on how costly the

different strategies are. In Panel B, we focus on the sign of trading costs. Here, the goal is to capture

switches between liquidity demand (positive trading costs) and liquidity provision (negative trading

costs). In the regression, we include time fixed effects to capture aggregate changes in the price of

liquidity, but the results are not sensitive to their inclusion.

The results in Table 5 can be summarized as follows. We note that i) funds following reversal

strategies exhibit significantly lower trading costs unconditionally. The economic significance of

the coefficient is large, as reversal funds have on average 11 basis points lower costs. Recalling that

the average cost for all type of funds is about 8 basis points, the evidence that reversal strategies

experience negative trading costs provides a validation of the trading cost variable as an effective

measure of liquidity provision. The same funds also experience a decrease in trading costs when

funding conditions tighten. This evidence suggests that liquidity providing strategies are even

more rewarding in bad times, when the price of liquidity increases. ii) Long-term momentum

funds have higher trading costs on average, consistent with the view that this strategy requires fast

execution. However, somewhat unexpected, the trading costs of momentum funds decrease when

funding conditions deteriorate. This seems related to the recent finding by Daniel and Moskowitz

(2013) that momentum strategies perform poorly only during market rebounds. So, possibly, in

down markets momentum managers are still not constrained by redemptions and can afford to

provide liquidity. iii) Value funds also experience somewhat higher trading costs unconditionally,

but the difference with other funds is not significant. Most important, however, is the fact that

the trading costs of value managers increase significantly when capital availability drops. This fact

is consistent with the fact that value strategies are often combined with the high use of leverage,

which makes them exposed to funding conditions. Specifically, value stocks tend to be low-beta

stocks which, according to Frazzini and Pedersen (2013), appear in arbitrageurs’ portfolios using

leverage to magnify these stocks’ high expected returns.

As a robustness check, we also experimented classifying funds into three groups based on the

distribution of the deciles, and obtain very similar findings. To the best of our knowledge, this is

34
the first study to links funds trading styles, as inferred from their trades, to trade-level liquidity

provision. Consistent with the limits of arbitrage hypothesis, we show that trading styles that are

more likely to be combined with leverage suffer higher trading costs in bad times.

6 Market liquidity and hedge fund trading

After pointing out that hedge funds’ liquidity provision is subject to aggregate funding conditions

and fund-level measures of financial constraints, we wish to study the extent to which hedge fund

trading activity is actually important for market liquidity. Aragon and Strahan (2012) provide

convincing evidence that the withdrawal of hedge funds from the market, as a result of Lehman

Brothers’ collapse, negatively impacted stock liquidity. Our data give us a unique opportunity to

directly observe hedge fund trades and to relate them to market liquidity at the stock-level.

We begin by constructing a measure of hedge funds’ trading intensity. For each fund in the

sample, we compute the total number of shares traded on a given stock in a given week. We

then normalize this quantity by the number of shares traded in the previous 12-week period, thus

obtaining abnormal turnover.21 Our proxy for hedge funds’ trading intensity (HF Tradeint) is the

average abnormal turnover across all hedge funds that trade a given stock in a given week.

As a measure of market liquidity, we follow closely Hameed, Kang, and Viswanathan (2010)

who focus on adjusted weekly bid-ask spreads. From CRSP, we compute the daily quoted spread

as the ratio of the difference between ask and bid closing prices to their midquote. We filter out

seasonalities as in Chordia, Sarkar, and Subrahmanyam (2005) using all available trading days in

the 1999-2010 period.22 The average of the resultant daily seasonal-adjusted spread for stock s in

week w is denoted ASP Rs,w .

We analyze the impact of hedge fund trading intensity on future bid-ask spreads by means of
21 We opted for a 12-week window in order to avail ourselves of a sufficient number of observations for computing a benchmark
volume for the same fund. To prevent unusually high or low activity from exerting undue influence, we trim the abnormal
turnover at the top and bottom 2.5%.
22 To be precise, we trim the series at the top and bottom 0.5% and we regress the quoted spread on day of the week dummies,

month dummies, a dummy for trading days that either precede or follow a holiday, a dummy for trading days after the
decimal system was introduced on 01/29/2001, and a year trend dummy. The adjusted spread is defined as the sum of the
intercept estimate and residuals.

35
the following model:

3
X 3
X 3
X
ASP Rs,w = const. + βk HF Tradeint w−k + ξk RM,w−k + ψk Rs,w−k
k=1 k=1 k=1
3
X
+ζ1 ST DM,w−1 + ζ2 ST Ds,w−1 + ζ3 T U RNs,w−1 + φk ASP Rs,w−k + us,w (10)
k=1

In addition to three lags of our variable of interest HF Tradeint, the regression includes the following

controls: three lags of the weekly market return RM as proxied by the CRSP VW index; three

lags of the weekly stock s’s return Rs ; the market and stock return volatility (ST DM and ST Ds ,

respectively) computed as the standard deviation of daily returns; the stock turnover T U RN

computed as the ratio between the weekly number of shares traded to total shares outstanding;

and three lags of the dependent variable to account for residual autocorrelation. These controls

are also used in Hameed, Kang, and Viswanathan (2010). In order to facilitate the economic

interpretation of the slopes, we standardize the regressors to have mean zero and unit variance.

The dependent variable is expressed in basis points.

Estimates of the model in (10) pooling all stock-week observations are presented in Table 13.

From column (1), we see that all lags of HF Tradeint enter the regression with a negative sign. In

particular, the loading on the prior week HF Tradeint equals -0.143, and it is strongly statistically

significant with a t-statistic of -2.73. The coefficients on the other lags are not significantly different

from zero. Hence, periods when hedge funds participate more actively in the trading of a given

stock are followed by a decline in spread. Although not large, the economic magnitude is of the

same order as for the return on the market, which Hameed, Kang, and Viswanathan (2010) point

out to be an important determinant of market liquidity. This evidence is consistent with hedge

funds participation having a beneficial impact on market liquidity.23

In the second column, we experiment an alternative specification where we add week fixed

effects and include only one lag of the regressors. Fixed effects represent another way of controlling

for unobserved aggregate determinants that could affect spreads, and that are not fully captured
23 We notice that the other variables in the regression have the expected impact on spreads. Market liquidity is higher following
a decrease in market return, a decrease in stock return, a decline in turnover, and an increase in idiosyncratic volatility. These
effects are consistent with the results in Hameed, Kang, and Viswanathan (2010), and with predictions of several theoretical
models of bid-ask spreads.

36
by market return and volatility.24 The coefficient on HF Tradeint is now smaller in magnitude at

-0.080, but still significantly different from zero at the 5% level. Finally, column (3) shows that

our conclusions remain unaltered if we conduct our analysis on changes in spreads, similarly to

Hameed, Kang, and Viswanathan (2010).

The results in the previous sections show that hedge funds liquidity provision is more sensitive

to changes in funding conditions than that of other institutional investors. Therefore, it is natural

to ask whether the trading intensity of hedge funds plays a special role in explaining the dynamics

of market liquidity or, instead, our measure is merely proxying for institutional investors’ partic-

ipation. We address this question by including the trading intensity of all other institutions as

additional determinant of future spreads (OI Tradeint), constructed using the same procedure.

In column (4) of Table 13, we see that the inclusion of OI Tradeint does not alter the significance

of hedge funds’ trading intensity. Interestingly, the third lag of HF Tradeint has now a significant

impact at -0.094 (t-statistic of -2.09). In contrast, none of the coefficients of OI Tradeint meets

statistical significance. In columns (5) and (6) we present the fixed-effect estimates. Notably, the

coefficient on OI Tradeint has now a negative sign, indicating that reduced trading activity of these

investors is also followed by higher spreads, but the coefficient has a t-statistic of just -1.

In sum, an increase in the trading intensity of hedge funds is accompanied by higher stock level

liquidity. This effect appears to persist up to three weeks. The fact that the impact remains econom-

ically and statistically significant even after accounting for stock-specific determinants, market-wide

observable, and unobservable controls, as well as for other investors’ trading intensity makes the

case for a causality link stronger. To the best of our knowledge, this is the first study to bring

direct evidence of an association between market liquidity and hedge fund trading activity at the

stock-level.

7 Robustness checks and additional analysis

We conduct a series of checks to ensure that our results are robust to alternative measures of

liquidity provision. Since these results are in general agreement with our main findings, we briefly

describe them in the text, and make details and corresponding tables available in the Internet
24 The coefficients on RM and ST DM cannot be estimated in the fixed effect specification as they are collinear.

37
Appendix.

7.1 Comparison with other benchmarks for trading costs

Our choice of VWAP as a benchmark to compute price impact is subject to a few caveats (see

Hasbrouck (2007), p. 148). For example, if a trade accounts for a large proportion of the daily

volume, the weighted average execution price of the trade is likely to coincide with the VWAP.

Also, on a day in which the price has been trending up, a buy transaction executed early on in the

day may appear as having negative price impact, while indeed the transaction is meant to chase a

trend and, as such, it is consuming liquidity.

To assess the robustness of our measure of liquidity provision to such considerations, we consider

valid candidates. That is, we recompute the price impact measure using two alternative bench-

marks. In one case, we benchmark the execution price to the price observed at the time the order

was placed. This may capture the price at which the trader wished to have execution, Anand,

Irvine, Puckett, and Venkataraman (2012). The second benchmark is the opening price of the day

on which the order was entered, and is used by Anand, Irvine, Puckett, and Venkataraman (2013).

Again, this is meant to serve as reference price for the manager’s decision to place the order. We

find that using these alternative series does not impact the inference. In the Internet Appendix,

we report estimates of the main equations in the paper using either series and show that our main

conclusions remain unaltered.

7.2 Comparison with other measures of liquidity provision

More generally, one may object to price impact as a measure of liquidity provision for speculators.

As an example, in Brunnermeier and Pedersen’s (2009) model, speculators that provide liquidity

to a temporarily underpriced security could very well decide to place a buy order at the ask price,

which is effectively a market order. Such a trade could still be denoted as liquidity providing to

the extent that speculators react to a perceived underpricing of the security following a temporary

order imbalance, but its price impact is positive.

As a first alternative to price impact, we focus on the dimension of liquidity provision that

underlies reversal strategies, such as those in Lo and MacKinlay (1990) and Nagel (2012). These

strategies capture the behavior of contrarian investors who provide liquidity in stocks that undergo

38
temporary price pressure. We construct a measure of proximity of hedge fund trades in Ancerno

to those predicated by reversal strategies, and find that it correlates strongly (about 50%) with

the negative of our price impact series. Hence, higher trading costs are associated with trades that

depart from reversal strategies. The second alternative measure of liquidity provision is inspired by

Anand, Irvine, Puckett, and Venkataraman (2013). It captures the intuition that if an institutional

order i on day t is in the same direction as the daily return on that stock, it is considered as

liquidity demanding, vice versa, orders with the opposite sign are liquidity providing. We document

a positive correlation, of about 30%, between T C and the measure in Anand, Irvine, Puckett, and

Venkataraman (2013). Importantly, we find that the two measures also display a long term decline

that ends with the financial crisis. This confirms that the trend in trading cost that is observed in

Figure 2 is not the mere effect of a steady decrease in trading costs, but is rather due to a deliberate

shift towards liquidity provision by hedge funds.

7.3 Switch between supply and demand

From the summary statistics in Table 3, hedge funds trades are characterized on average by positive

trading costs. One may wonder to what extent hedge funds change from being liquidity demanders

to liquidity suppliers, or whether they are systematically demanding liquidity but with time-varying

intensity. To answer this question, we construct a dummy variable that equals one if the daily

trading cost of a fund is positive, and zero otherwise. Its average over the period is 0.53. Thus,

nearly half of the trading costs fall below the VWAP therefore reflecting liquidity provision. Results

from probit regressions of this dummy on the funding liquidity determinants are in line with those

in Table 6. This evidence testifies of a true shift of the hedge fund sector from liquidity provision

to liquidity demand as funding conditions deteriorate.

8 Conclusion

This paper studies the behavior of hedge funds’ liquidity provision as a function of changes in

funding liquidity. We gain a privileged vintage point on hedge fund trading activity in the U.S.

stock market thanks to a data set of institutional trades, among which we are able to identify hedge

fund trades. The data cover the long time series from January 1999 to December 2010 for a sample

39
of eighty-seven different hedge fund management companies.

We find that hedge funds’ liquidity provision, as measured by the difference in the execution

price and the volume-weighted average price, varies significantly both cross-sectionally as well as

over time. Consistent with prior studies, hedge funds were heavily demanding liquidity during the

dot-com bubble and the last financial crisis. In calmer times, between 2002 and 2007, they were

instead more inclined to provide liquidity.

The data give strong support to several predictions that stem from the literature on the limits of

arbitrage. Hedge funds’ trading costs are systematically higher during persistent declines in market

valuations, consistent with a collateral based story that links the ability to provide market liquidity

to the value of the underlying assets. Trading costs also increase following periods of higher VIX,

TED spread, and LIBOR. We interpret this as the consequence of an increase in holding costs, e.g.

higher margins, occurring at times of higher uncertainty, which induces hedge funds to liquidate

their positions and demand liquidity. These effects are not simply due to time variation in the

cost of liquidity, as the trading costs of other institutional investors show substantially weaker (or

even opposite) relation to the same variables. Hedge funds also appear to move more aggressively

toward liquid stocks than other investors when funding constraints are tighter and to trade more

impatiently.

Funding liquidity shocks do not, however, affect all hedge funds equally. Quite the opposite,

the ability to steadily provide liquidity varies across funds as a function of fund characteristics that

capture relevant dimensions of funding constraints. In particular, the funding liquidity variables

have a stronger impact on young funds, leveraged funds, funds which invest in illiquid assets,

and funds with a poor recent performance. The same funds also experience a worsening of the

short-term returns to their trades.

Lastly, we provide novel evidence that a decrease in hedge fund trading intensity in a given

stock anticipates a future deterioration of stock-level liquidity. A similar effect is not found for

the other institutions in our data, suggesting that hedge funds are likely to be crucial providers of

liquidity, as pointed out by other studies (Aragon and Strahan (2012)).

40
References
Amihud, Yakov, 2002, Illiquidity and Stock Returns: Cross-Section and Time-Series Effects, Jour-
nal of Financial Markets 5, 31–56.
Anand, Amber, Paul Irvine, Andy Puckett, and Kumar Venkataraman, 2012, Performance of insti-
tutional trading desks: An analysis of persistence in trading costs, Review of Financial Studies
25, 557–598.
, 2013, Market crashes and institutional trading, Journal of Financial Economics, forth-
coming.
Aragon, George O., 2007, Share restrictions and asset pricing: Evidence from the hedge fund
industry, Journal of Financial Economics 83, 33–58.
, and Philip E. Strahan, 2012, Hedge funds as liquidity providers: Evidence from the
Lehman bankruptcy, Journal of Financial Economics 103, 570 – 587.
Ben-David, Itzhak, Francesco Franzoni, Augustin Landier, and Rabih Moussawi, 2012, Do Hedge
Funds Manipulate Stock Prices?, Journal of Finance , forthcoming.
Ben-David, Itzhak, Francesco Franzoni, and Rabih Moussawi, 2012, Hedge Fund Stock Trading in
the Financial Crisis of 2007–2009, Review of Financial Studies 25, 1–54.
Berkowitz, Stephen A., Dennis E. Logue, and Eugene A. Noser, 1988, The total cost of transactions
on the nyse, Journal of Finance 43, 97–112.
Billio, Monica, Mila Getmansky, Andrew W Lo, and Loriana Pelizzon, 2012, Econometric measures
of connectedness and systemic risk in the finance and insurance sectors, Journal of Financial
Economics 104, 535–559.
Boyson, Nicole M., Christof W. Stahel, and Renè Stulz, 2010, Hedge fund contagion and liquidity
shocks, Journal of Finance 65, 1789–1816.
Brunnermeier, Markus K., and Stefan Nagel, 2004, Hedge funds and the technology bubble, Journal
of Finance 59, 2013–2040.
Brunnermeier, Markus K., and Lasse Heje Pedersen, 2009, Market liquidity and funding liquidity,
Review of Financial Studies 22, 2201–2238.
Chemmanur, Thomas J., Shan He, and Gang Hu, 2009, The role of institutional investors in
seasoned equity offerings, Journal of Financial Economics 94, 384–411.
Chemmanur, Thomas J., Gang Hu, and Jiekun Huang, 2010, The role of institutional investors in
initial public offerings, Review of Financial Studies 23, 4496–4540.
Chevalier, Judith, and Glenn Ellison, 1997, Risk taking by mutual funds as a response to incentives,
Journal of Political Economy 105, 1167–1200.
Chordia, Tarun, Asani Sarkar, and Avanidhar Subrahmanyam, 2005, An empirical analysis of stock
and bond market liquidity, Review of Financial Studies 18, 85–129.
Daniel, Kent, and Tobias Moskowitz, 2013, Momentum crashes, Working paper.
Ding, Bill, Mila Getmansky, Bing Liang, and Russell Wermers, 2009, Share restrictions and investor
flows in the hedge fund industry, Working Paper, SSRN 891732.
Fama, Eugene F., and Kenneth R. French, 1993, Common risk factors in the returns on stocks and
bonds, Journal of Financial Economics 33, 3–56.
Frazzini, Andrea, and Lasse H Pedersen, 2013, Betting against beta, Journal of Financial Eco-
nomics , forthcoming.
Fung, W., and D. A. Hsieh, 2001, The risk in hedge fund strategies: Theory and evidence from
trend followers, Review of Financial Studies 14, 313–341.

41
Fung, William, and David A Hsieh, 2004, Hedge fund benchmarks: A risk-based approach, Finan-
cial Analysts Journal pp. 65–80.
Gantchev, Nickolay, and Chotibhak Jotikasthira, 2012, Hedge fund activists: Do they take cue from
institutional exit?, Working paper.
Garleanu, Nicolae, and Lasse Heje Pedersen, 2011, Margin-based asset pricing and deviations from
the law of one price, Review of Financial Studies 24, 1980–2022.
Getmansky, Mila, Andrew W Lo, and Igor Makarov, 2004, An econometric model of serial correla-
tion and illiquidity in hedge fund returns, Journal of Financial Economics 74, 529–609.
Gromb, Denis, and Dimitri Vayanos, 2002, Equilibrium and welfare in markets with financially
constrained arbitrageurs, Journal of Financial Economics 66, 361–407.
, 2010, Limits of arbitrage: The state of the theory, Annual Review of Financial Economics
2, 251–275.
, 2012, Financially constrained arbitrage and cross-market contagion, Working Paper, IN-
SEAD and LSE.
Hameed, Allaudeen, Wenjin Kang, and Shivesh Viswanathan, 2010, Stock Market Declines and
Liquidity, Journal of Finance 65, 257–293.
Hasbrouck, Joel, 2007, Empirical market microstructure: The institutions, economics, and econo-
metrics of securities trading (Oxford University Press).
Hendershott, Terrence, Charles M. Jones, and Albert J. Menkveld, 2011, Does algorithmic trading
improve liquidity?, Journal of Finance 66, 1–33.
Hu, Gang, 2009, Measures of implicit trading costs and buysell asymmetry, Journal of Financial
Markets 12, 418 – 437.
Jame, Russell, 2012, How Skilled are Hedge Funds? Evidence from Their Daily Trades, Working
paper.
Jegadeesh, Narasimhan, and Sheridan Titman, 1993, Returns to buying winners and selling losers:
Implications for stock market efficiency, The Journal of Finance 48, 65–91.
Jylha, Petri, Kalle Rinne, and Matti Suominen, 2012, Do hedge funds supply or demand immedi-
acy?, Working Paper, SSRN 1629773.
Keim, Donald B., and Ananth Madhavan, 1995, The anatomy of the trading process, Journal of
Financial Economics 37, 371–398.
, 1997, Transactions costs and investment style: an inter-exchange analysis of institutional
equity trades, Journal of Financial Economics 46, 265–292.
Khandani, Amir E., and Andrew W. Lo, 2011, What happened to the quants in August 2007?
Evidence from factors and transactions data, Journal of Financial Markets 14, 1–46.
Li, Haitao, Xiaoyan Zhang, and Rui Zhao, 2011, Investing in talents: Manager characteristics and
hedge fund performances, Journal of Financial and Quantitative Analysis 46, 59.
Lipson, Marc, and Andy Puckett, 2010, Institutional trading during extreme market movements,
Working paper.
Lo, A.W., and A.C. MacKinlay, 1990, When are contrarian profits due to stock market overreac-
tion?, Review of Financial studies 3, 175–205.
Manconi, Alberto, Massimo Massa, and Ayako Yasuda, 2012, The role of institutional investors in
propagating the crisis of 2007-2008, Journal of Financial Economics 104, 491 – 518.
Nagel, Stefan, 2012, Evaporating Liquidity, Review of Financial Studies 25, 2005–2039.

42
Patton, Andrew, and Tarun Ramadorai, 2012, On the high-frequency dynamics of hedge fund risk
exposures, Journal of Finance , forthcoming.
Patton, Andrew J, 2009, Are market neutral hedge funds really market neutral?, Review of Finan-
cial Studies 22, 2495–2530.
Puckett, Andy, and Xuemin (Sterling) Yan, 2011, The interim trading skills of institutional in-
vestors, Journal of Finance 66, 601–633.
Sadka, Ronnie, 2010, Liquidity risk and the cross-section of hedge-fund returns, Journal of Financial
Economics 98, 54–71.
Shleifer, Andrei, and Robert W. Vishny, 1997, The Limits of Arbitrage, Journal of Finance 52,
35–55.
Sirri, Erik R, and Peter Tufano, 2002, Costly search and mutual fund flows, Journal of Finance 53,
1589–1622.
Teo, Melvyn, 2011, The liquidity risk of liquid hedge funds, Journal of Financial Economics 100,
24–44.

43
 Table 1: Summary statistics for daily series

The table displays the following statistics: mean; standard deviation; minimum; 25th, 50th, and 75th percentiles; maximum.
The variables are: the daily number of trades and orders, the daily average execution time, the total daily dollar volume, the
daily volume per trade/buy trade/sell trade. The statistics are calculated for trades originating from hedge funds in Panel A
and from other institutions in Ancerno in Panel B. The sample is period is January 1999 to December 2010.

Panel A: Hedge funds

Mean Std Min p25 p50 p75 Max

Number of management companies 21 4 10 18 21 24 34
Number of trades 2,124 2,320 69 530 1,233 2,778 17,308
Number of orders 736 946 1 283 463 815 11,525
Execution time 1.021 0.585 0.069 0.576 1.017 1.353 6.905
Volume ($ millions) 346 308 12 141 252 434 2,534
Volume per trade ($ thousands) 229 131 24 140 195 293 1,128
Volume per trade, buy trades ($ thousands) 226 138 23 129 191 292 1,346
Volume per trade, sell trades ($ thousands) 240 154 23 137 202 306 1,907

Panel B: Other institutions

Mean Std Min p25 p50 p75 Max

Number of management companies 248 18 134 237 248 260 300
Number of trades 95,367 67,247 8,902 44,091 85,064 127,981 1,706,477
Number of orders 36,593 38,170 3,064 18,641 27,280 47,850 1,513,899
Execution time 1.117 0.401 0.243 0.806 1.104 1.391 3.040
Volume ($ millions) 15,057 5,991 2,248 11,182 14,310 18,156 125,234
Volume per trade ($ thousands) 213 116 10 120 174 290 1,394
Volume per trade, buy trades ($ thousands) 203 111 8 118 169 273 1,128
Volume per trade, sell trades ($ thousands) 226 127 13 128 187 307 1,729

44
Table 2: Analysis of Representativeness of Ancerno funds: Comparison of slopes and
alphas with TASS funds

In Panel A, for each hedge fund management company in TASS we estimate separate regressions of monthly returns on the
Fung and Hsieh (2001) risk factors and each of the five funding liquidity variables. These are the market return (RM ), the
change in VIX, TED, LIBOR, or the first principal component thereof (PC). When including RM or change in LIBOR we
exclude from the Fung and Hsieh (2001) factors the market and change in 10-year yield, respectively. The first row reports the
average loading on the liquidity variable for the funds that are in TASS but not in Ancerno (column ‘TASS’) and for the funds
that are in TASS and report to Ancerno (column ‘Ancerno’). The column ‘pvalue’ reports the p-value for the null hypothesis
that their difference is zero. The second row reports the cross-sectional standard deviation of the loadings for the two groups,
and the column pvalue reports the p-value for the null hypothesis that the variance of the Ancerno group is greater than that
of TASS. Panel B reports the slope (‘Coefficient’), Standard Error, and pvalue in the regression of either the number of funds
in Ancerno (row ‘# funds’) or the average sensitivity of the reporting funds (row ‘loading’) on a constant and each funding
liquidity variable. Panel C reports the average monthly return, percentage flow, first-order autoregressive coefficient of returns,
age as log of number of months reporting in TASS, leverage in place, and log assets-under-management (AUM) for the group
of funds ‘TASS’ and ‘Ancerno’ funds, and the pvalue for the null hypothesis that their difference is zero.

Panel A: Risk-Loadings

TASS Ancerno pvalue

RM Mean 0.768% 0.913% 0.229
Std 0.895% 1.186% 0.001

VIX Mean -0.007% 0.055% 0.053

Std 0.250% 0.235% 0.725

TED Mean -0.185% -0.151% 0.472

Std 0.376% 0.297% 0.990

LIBOR Mean -0.156% -0.029% 0.011

Std 0.385% 0.398% 0.335

PC Mean -0.089% -0.049% 0.201

Std 0.242% 0.186% 0.995

Panel B: Sample composition

Coefficient Std.Err. pvalue

RM # funds -0.385 0.349 0.272
loading 0.099 1.468 0.946

VIX # funds -0.353 -1.020 0.307

loading 0.298 0.412 0.470

TED # funds 0.195 0.345 0.573

loading -0.337 0.624 0.590

LIBOR # funds 0.365 0.344 0.291

loading -0.136 0.651 0.835

PC # funds 0.224 0.345 0.518

loading -0.386 0.432 0.373

Panel C: Hedge funds characteristics

TASS Ancerno pvalue

Return 0.005 0.007 0.139
Flows 0.019 0.015 0.314
AR(1) 0.031 0.025 0.761
Age 3.163 3.598 0.000
Leverage 40.365 52.634 0.195
AUM 17.250 17.826 0.010

45
 Table 3: Summary statistics for trading costs and funding liquidity determinants

The table reports the following statistics: mean; standard deviation; first-order autoregressive coefficient; minimum; 25th, 50th,
and 75th percentiles; maximum. The statistics are for trading costs to hedge funds trades in Panel A, for trading costs to other
institutions in Ancerno in Panel B and for the funding liquidity determinants in Panel C. Trading costs, expressed in basis
points and aggregated at the manager-day level, are computed as volume-weighted (superscript V W ) or as equally-weighted
(superscript EW ) averages, and separately for buy trades (subscript b) and sell trades (subscript s). The funding liquidity
determinants are the two-week return to the CRSP value-weighted index (RM ), the VIX, the TED spread, the LIBOR rate,
and the first principal component of these four variables (PC). The sample January 1999 and December 2010.

Panel A: Hedge funds

Mean Std AR(1) Min p25 p50 p75 Max

T CV W 8.250 66.208 0.097 -232.160 -15.425 5.187 30.266 265.607
T CbV W 6.628 72.625 0.082 -255.616 -19.711 4.235 32.424 279.016
T CsV W 8.455 77.586 0.096 -274.106 -19.101 4.934 33.941 305.932
T C EW 4.252 63.419 0.126 -224.560 -19.577 2.352 26.963 243.939
T CbEW 3.086 70.174 0.101 -249.247 -23.710 1.749 29.695 262.816
T CsEW 5.293 76.344 0.121 -270.903 -22.766 2.734 31.975 293.345

Panel B: Other institutions

Mean Std AR(1) Min p25 p50 p75 Max

T CV W 1.037 54.275 0.091 -207.201 -16.810 1.120 19.195 204.326
T CbV W -0.118 60.098 0.081 -229.880 -20.160 0.420 21.280 217.660
T CsV W 1.106 62.506 0.080 -239.140 -19.270 0.910 21.960 234.930
T C EW -3.416 52.986 0.120 -205.293 -21.995 -1.916 15.812 188.971
T CbEW -4.224 58.907 0.103 -227.249 -25.139 -2.450 18.221 202.666
T CsEW -2.754 62.309 0.103 -242.211 -23.950 -1.490 19.398 224.390

Panel C: Time-series determinants

Mean Std AR(1) Min p25 p50 p75 Max

RM 0.002 0.038 0.876 -0.267 -0.018 0.006 0.023 0.222
VIX 0.223 0.092 0.983 0.099 0.160 0.214 0.259 0.809
TED 0.005 0.005 0.963 0.000 0.002 0.003 0.005 0.046
LIBOR 0.032 0.021 1.000 0.002 0.013 0.028 0.052 0.069
PC 0.000 1.301 0.962 -1.773 -0.842 -0.326 0.475 11.721

Correlation matrix

RM VIX TED LIBOR

VIX -0.310
TED -0.196 0.497
LIBOR -0.067 -0.194 0.258
PC -0.605 0.813 0.805 0.131

46
 Table 4: Hedge funds’ liquidity provision and funding liquidity

OLS estimates of equation (3):

T Ci,t+1 = a + bF undLiqt + φT Ci,t + ǫi,t+1

where T Ci,t+1 is the trading cost of hedge fund i on day t + 1, and F undLiqt denotes alternatively RM , VIX, TED, LIBOR, or
PC measured on day t. In Panel A, trading costs are computed as volume-weighted average across all trades in columns (1) to
(5) and as equally-weighted average in columns (6) to (10). In Panel B, hedge funds trading costs are volume-weighted and are
computed separately for Buy (columns (1) to (5)) and Sell (columns (6) to (10)) trades. Each column uses a different funding
liquidity variable, F undLiq, which are defined as in Table 3. Below the coefficients, t-statistics based on robust standard errors
clustered at the date level are reported in parentheses. The estimates of the intercept and lag term are omitted. The sample
period is January 1999 to December 2010.

Panel A: All trades

Dep. Var.: Volume-weighted Trading Costs Equally-weighted Trading Costs

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
RM -1.515 -1.024
(-4.33) (-3.01)
VIX 1.750 1.319
(4.79) (3.55)
TED 1.925 1.397
(5.12) (3.78)
LIBOR 3.220 2.891
(10.84) (10.58)
PC 2.618 1.943
(6.76) (4.93)

Obs. 53,211 53,211 53,211 53,211 53,211 53,211 53,211 53,211 53,211 53,211
R2 0.01 0.01 0.01 0.01 0.01 0.02 0.02 0.02 0.02 0.02

Panel B: Trades by side

Dep. Var.: Buy-side volume-weighted Trading Costs Sell-side volume-weighted Trading Costs
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
RM -0.830 -1.727
(-1.78) (-3.17)
VIX 1.162 2.771
(2.50) (4.89)
TED 1.911 1.225
(3.70) (1.98)
LIBOR 2.137 3.058
(5.80) (7.39)
PC 1.956 2.850
(3.71) (4.40)

Obs. 42,961 42,961 42,961 42,961 42,961 42,282 42,282 42,282 42,282 42,282
R2 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01

47
 Table 5: Impact of 2007-2009 financial crisis

OLS estimates of equation (3):

T Ci,t+1 = a + bF undLiqt + φT Ci,t + ǫi,t+1

where T Ci,t+1 is the trading cost of hedge fund i on day t + 1, and F undLiqt denotes alternatively RM , VIX, TED, LIBOR,
or PC measured on day t. The regression is estimated separately on the ex-crisis (Panel A) and crisis (Panel B) periods. The
financial crisis period is defined as July 2007 to March 2009. Each column uses a different funding liquidity variable, F undLiq,
which are defined as in Table 3. Below the coefficients, t-statistics based on robust standard errors clustered at the date level
are reported in parentheses. All regressions include a lag term and an intercept (whose estimates are omitted). The whole
sample period is January 1999 to December 2010.

Panel A: Ex-crisis period

Dep. Var.: Volume-weighted trading cost

(1) (2) (3) (4) (5)

RM -1.027
(-2.76)
VIX 1.995
(5.32)
TED 5.325
(7.43)
LIBOR 3.393
(11.11)
PC 4.304
(8.69)

Obs. 47,539 47,539 47,539 47,539 47,539

R2 0.01 0.01 0.01 0.01 0.01

Panel B: Crisis period

Dep. Var. : Volume-weighted trading cost

(1) (2) (3) (4) (5)

RM -2.899
(-3.61)
VIX 1.887
(2.42)
TED 2.244
(2.82)
LIBOR -0.083
(-0.06)
PC 2.881
(3.53)

Obs. 5,642 5,642 5,642 5,642 5,642

R2 0.01 0.01 0.01 0.01 0.01

48
 Table 6: Hedge funds’ and other institutions liquidity provision and funding liquidity

Panel A presents OLS estimates of equation (4) on hedge fund trades:

T Ci,t+1 = a + bF undLiqt + δ′ Zi,t + εi,t+1
where T Ci,t+1 is the volume-weighted average trading cost on day t + 1, separately computed for buy and sell trades. Panel B presents OLS estimates of equation (5) pooling
hedge funds and other institution trades:
T Ci,t+1 = a + b1 F undLiqt + b2 HF × F undLiqt + δ′ Zi,t + εi,t+1

HF equals 1 if the institution is a hedge fund, and 0 otherwise. Each specification from (1) to (5) uses a different funding liquidity variable, F undLiq, which are defined as in
Table 3. The vector Zi,t collects the following controls for trade difficulty: Buy is a dummy that equals 1 for buy trades, and 0 otherwise; Lagged Return is the stock return
in the prior day; NYSE is a dummy that equals 1 for stocks listed at the NYSE, and 0 otherwise; Inverse Price is the inverse of day-t stock price; Relative Volume is the
ratio between the number of shares traded and the average volume in the prior 30 days; Amihud is the Amihud illiquidity ratio; Size and Book-to-Market are the stock market
capitalization and book-to-market deciles. All variables are computed as volume-weighted averages at the fund-day-side level. Below the coefficients, t-statistics based on robust
standard errors clustered at the date level are reported in parentheses. The sample period is January 1999 to December 2010.

Dep. Var. Panel A: Volume-weighted trading cost, only HF Panel B: Volume-weighted trading cost, HF and MF

F unLiq RM VIX TED LIBOR PC RM VIX TED LIBOR PC

(1) (2) (3) (4) (5) (1) (2) (3) (4) (5)
HF ∗ F undLiq -1.382 2.784 2.206 2.360 3.305
(-3.97) (7.63) (5.16) (9.69) (8.30)
49

F undLiq -1.414 2.280 2.665 3.155 3.332 0.381 -0.813 -0.947 1.021 -0.907
(-4.49) (6.52) (6.22) (13.54) (8.39) (3.13) (-5.93) (-7.23) (13.78) (-6.43)
Buy -1.420 -1.419 -1.383 -1.362 -1.445 -1.067 -1.064 -1.070 -1.071 -1.061
(-2.79) (-2.79) (-2.72) (-2.68) (-2.85) (-4.98) (-4.96) (-4.99) (-4.99) (-4.95)
Lagged Return -0.260 -0.287 -0.304 -0.300 -0.248 -0.208 -0.205 -0.201 -0.191 -0.213
(-2.32) (-2.59) (-2.73) (-2.68) (-2.25) (-5.94) (-5.90) (-5.73) (-5.49) (-6.11)
NYSE -7.267 -7.588 -7.331 -7.039 -7.577 -5.757 -5.694 -5.773 -5.652 -5.735
(-9.60) (-10.02) (-9.70) (-9.30) (-10.02) (-23.55) (-23.17) (-23.62) (-23.34) (-23.39)
Inverse Price 0.075 0.016 0.089 0.175 0.039 -0.019 0.006 -0.019 0.019 -0.006
(0.79) (0.17) (0.94) (1.83) (0.42) (-0.73) (0.22) (-0.70) (0.72) (-0.23)
Relative Volume 0.867 0.905 0.896 0.880 0.928 0.412 0.408 0.402 0.409 0.406
(1.74) (1.80) (1.79) (1.77) (1.85) (4.36) (4.33) (4.28) (4.36) (4.31)
Amihud 1.302 1.296 1.283 0.731 1.184 0.087 0.090 0.090 0.071 0.090
(0.67) (0.67) (0.67) (0.39) (0.62) (0.86) (0.89) (0.88) (0.71) (0.89)
Size 0.817 0.741 0.830 0.965 0.768 0.862 0.894 0.864 0.906 0.880
(6.06) (5.47) (6.16) (7.15) (5.69) (23.39) (23.91) (23.43) (24.72) (23.79)
Book/Market -1.415 -1.300 -1.432 -1.445 -1.327 -1.285 -1.334 -1.281 -1.292 -1.312
(-9.89) (-9.02) (-9.99) (-10.06) (-9.30) (-33.87) (-33.59) (-33.77) (-34.06) (-34.11)

Obs. 103,673 103,637 103,637 103,637 103,637 1,318,745 1,318,361 1,318,361 1,318,361 1,318,361
R2 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01
 Table 7: Stock type and funding liquidity

OLS estimates of equation (6):

V W ADi,t+1 = a1 + a2 HF + θ1 F undLiqt + θ2 HF × F undLiqt + φ1 V W ADi,t + φ2 HF × V W ADi,t + υi,t+1
where V W ADi,t+1 is the volume-weighted average decile of the stocks traded by institution i on day t+1. Decile assignment for
a given characteristic is based on the distribution of the universe of stocks in CRSP. The characteristics are market capitalization
(Size) in Panel A and the Amihud (2002) measure in Panel B. HF equals 1 if the institution is a hedge fund and 0 otherwise.
Each column uses a different funding liquidity variable, F undLiq, which are defined as in Table 3. Below the coefficients, t-
statistics based on robust standard errors clustered at the date level are reported in parentheses. The estimates of the intercept
and lag terms are omitted. The sample period is January 1999 to December 2010.

Panel A. Dep. Var.: Size VWAD

F undLiq : RM VIX TED LIBOR PC

(1) (2) (3) (4) (5)
HF × F undLiq -0.009 0.040 0.018 0.001 0.033
(-2.08) (8.40) (3.66) (0.17) (6.72)
F undLiq -0.007 0.018 -0.004 0.014 0.010
(-4.56) (10.86) (-3.15) (8.81) (6.40)
HF 0.838 0.879 0.841 0.837 0.863
(17.30) (17.98) (17.32) (17.26) (17.71)

Obs. 659,329 659,329 659,329 659,329 659,329

R2 0.562 0.563 0.562 0.562 0.562

Panel B. Dep. Var.: Amihud measure VWAD

F undLiq : RM VIX TED LIBOR PC

(1) (2) (3) (4) (5)
HF × F undLiq 0.007 -0.034 -0.015 -0.004 -0.028
(1.74) (-7.60) (-3.21) (-1.02) (-6.22)
F undLiq 0.007 -0.019 0.006 -0.013 -0.010
(4.40) (-11.08) (3.87) (-8.28) (-6.18)
HF 0.190 0.199 0.189 0.190 0.193
(17.64) (18.29) (17.62) (17.66) (17.89)

Obs. 659,329 659,329 659,329 659,329 659,329

R2 0.53 0.531 0.53 0.53 0.53

50
 Table 8: Execution time and funding liquidity

OLS estimates of equation (7):

ExT imei,t+1 = a + βF undLiqt + γ ′ Xi,t + ui,t+1

where ExT imei,t+1 is the log average execution time on day t + 1 for orders placed by hedge fund i. Execution time is
computed as the distance (in minutes) between the first time when an order is placed and last time when order is executed.
Data are aggregated at the date-institution-Side level. The vector X contains the following controls. Side equals 1 for buy
trades and 0 for sell trades. Relative V olume is the value-weighted ratio of the number of stocks traded to the average volume
of the stock in the prior 30 days. Size V W AD is the volume-weighted average decile of the stocks traded by the hedge fund.
T rend is a time trend expressed as fraction of year. Each column uses a different funding liquidity variable, F undLiq, which
are defined as in Table 3. Below the coefficients, t-statistics based on standard errors clustered at the date and institution-
Side level are reported in parentheses. The intercept estimate is omitted. The sample period is January 1999 to December 2010.

Dep. Var: Order execution time

(1) (2) (3) (4) (5)

RM -0.002
(-0.23)
VIX -0.090
(-3.06)
TED -0.025
(-0.89)
LIBOR 0.034
(1.14)
PC -0.053
(-1.97)
Side 0.040 0.040 0.040 0.040 0.040
(0.49) (0.48) (0.48) (0.49) (0.48)
Relative V olume 1.039 1.025 1.033 1.040 1.027
(8.92) (9.54) (9.16) (8.95) (9.28)
Size V W AD -0.106 -0.098 -0.105 -0.106 -0.103
(-6.54) (-6.09) (-6.48) (-6.55) (-6.33)
T rend -0.087 -0.086 -0.086 -0.082 -0.087
(-6.49) (-6.63) (-6.65) (-5.76) (-6.54)

Obs. 82,973 82,937 82,937 82,937 82,937

R2 0.15 0.16 0.15 0.15 0.15

51
 Table 9: Liquidity provision, funding liquidity, and hedge funds’ characteristics

OLS estimates of equation (8):

T Ci,t+1 = α + β ′ F undLiqt + γ ′ Xi,m−1 + η′ F undLiqt × Xi,m−1 + ǫi,t+1
where T Ci,t+1 is hedge fund i volume-weighted trading cost on day t + 1. Xi,m collects six cross-sectional characteristics.
These are the amount of leverage in place (Leverage); minus the age of the fund (Y oung); the decile of the distribution of the
first-order autocorrelation in returns (Illiquid); minus the year-to-date performance (Bad); a dummy variable (LowRed) that
equals 1 if redemption notice period plus redemption frequency is lower than 120 days (the median in the sample); a dummy
variable (N oLock) that equals 1 if lockup period is 0 and 0 otherwise. All cross-sectional controls are from the month preceding
that of the trade (month m − 1). Below the coefficients, t-statistics based on robust standard errors clustered at the date level
are reported in parentheses. All regressions include a lag, a intercept term, and style fixed effects (whose estimates are omitted).
Bold face denotes estimates whose significance (at the 10% level) makes them consistent with Hypothesis 3 in the text. The
sample period is January 1999 to December 2010.

Dep. Var.: Volume-weighted trading cost

F undLiq : RM VIX TED LIBOR PC

(1) (2) (3) (4) (5)
Leverage × F undLiq -0.001 -0.002 0.006 -0.001 0.003
(-0.62) (-0.97) (3.76) (-0.46) (1.99)
Y oung × F undLiq 0.369 2.387 4.694 2.038 4.096
(0.37) (2.19) (3.12) (1.96) (3.29)
Illiquid × F undLiq 0.096 0.026 0.503 0.303 0.133
(0.50) (0.12) (2.38) (1.68) (0.58)
Bad × F undLiq -4.486 6.693 -10.086 7.381 0.991
(-1.15) (1.65) (-2.39) (1.82) (0.24)
LowRed × F undLiq 0.476 -3.647 13.740 12.858 4.816
(0.32) (-2.43) (5.19) (10.23) (2.53)
N oLock × F undLiq 3.596 -1.397 -11.984 -3.777 -8.235
(2.49) (-0.94) (-5.68) (-3.01) (-4.57)
Leverage -0.015 -0.017 -0.015 -0.014 -0.016
(-7.81) (-6.55) (-7.53) (-6.20) (-8.33)
Y oung 3.126 3.150 3.948 -0.055 3.882
(3.83) (3.81) (3.97) (-0.06) (4.47)
Illiquid 0.080 0.115 0.282 0.170 0.139
(0.50) (0.70) (1.71) (1.06) (0.86)
Bad 7.947 7.537 10.254 5.887 7.458
(2.05) (1.91) (2.60) (1.52) (1.94)
LowRed 5.141 5.986 8.284 5.342 5.516
(4.28) (4.91) (5.74) (4.48) (4.44)
N oLock 0.779 0.535 -1.781 -1.128 -0.021
(0.65) (0.44) (-1.37) (-0.96) (-0.02)
F undLiq -1.475 11.700 22.480 9.932 21.365
(-0.33) (2.32) (2.97) (2.01) (3.61)

Obs. 26,659 26,659 26,659 26,659 26,659

R2 0.01 0.01 0.02 0.03 0.02

52
 Table 10: Constrained hedge funds’ liquidity provision and funding liquidity

OLS estimates of equation (5):

T Ci,t+1 = a1 + a2 HF + b1 F undLiqt + b2 HF × F undLiqt + φ1 T Ci,t + φ2 HF × T Ci,t + εi,t+1
where T Ci,t+1 is the volume-weighted average trading cost of institution i on day t + 1 on stocks belonging to the top market
capitalization decile. In Panel A, only trades from constrained hedge funds and other institutions are used. In Panel B, only
trades from unconstrained hedge funds and other institutions are used. Constrained hedge funds are defined as those reporting
to TASS and having positive leverage and a redemption notice period plus redemption frequency lower than 120 days (the
median value in the sample). HF equals 1 if the institution is a hedge fund and 0 otherwise. Each column uses a different
funding liquidity variable, F undLiq, which are defined as in Table 3. Below the coefficients, t-statistics based on robust standard
errors clustered at the date level are reported in parentheses. The estimates of the intercept and lag terms are omitted. The
sample period is January 1999 to December 2010.

Panel A: Constrained hedge funds and other institutions

Dep. Var.: Volume-weighted trading costs

F undLiq : RM VIX TED LIBOR PC

(1) (2) (3) (4) (5)
HF × F undLiq 0.318 2.086 1.735 3.209 1.966
(0.35) (1.96) (1.53) (4.19) (1.78)
F undLiq 0.094 -0.338 -0.347 0.260 -0.340
(0.63) (-2.41) (-2.04) (2.66) (-2.00)
HF 4.100 4.044 4.313 3.617 4.179
(5.32) (5.32) (5.23) (4.99) (5.31)

Obs. 393,908 393,908 393,908 393,908 393,908

R2 0.01 0.01 0.01 0.01 0.01

Panel B: Unconstrained hedge funds and other institutions

Dep. Var.: Volume-weighted trading costs

F undLiq : RM VIX TED LIBOR PC

(6) (7) (8) (9) (10)
HF × F undLiq -0.079 -1.384 1.473 -0.282 -0.011
(-0.06) (-0.90) (0.90) (-0.26) (-0.01)
F undLiq 0.094 -0.338 -0.347 0.260 -0.340
(0.63) (-2.41) (-2.04) (2.66) (-2.00)
HF 7.022 7.052 7.107 7.022 7.015
(6.88) (6.95) (6.73) (7.02) (6.80)

Obs. 390,015 390,015 390,015 390,015 390,015

R2 0.01 0.01 0.01 0.01 0.01

53
 Table 11: Performance, funding liquidity, and hedge funds’ characteristics

OLS estimates of equation (9):

ri,t:t+5 = α + β ′ F undLiqt + γ ′ Xm−1 + η′ F undLiqt × Xm−1 + ǫi,t+1:t+5
where ri,t:t+5 is the volume-weighted abnormal return (in %) to hedge fund i trades between days t and t + 5. Each column
uses a different funding liquidity variable, F undLiq, which are defined as in Table 3. The cross-sectional characteristics in X
are measured in the month prior the trade and are defined as in Table 9. Below the coefficients, t-statistics based on robust
standard errors clustered at the date level are reported in parentheses. The intercept estimate is omitted. Leverage amount
is divided by 100. Bold face denotes estimates whose significance (at the 10% level) is consistent with a significant relation
between fund-level limits of arbitrage and a deterioration in trade performance when funding liquidity tightens. The sample
period is January 1999 to December 2010.

Dep. Var.: 5-day Ex-post Performance

F undLiq : RM VIX TED LIBOR PC

(1) (2) (3) (4) (5)
Leverage × F undLiq 0.064 -0.039 -0.017 -0.050 -0.044
(3.33) (-2.00) (-1.43) (-2.04) (-3.13)
Y oung × F undLiq -0.070 -0.223 0.028 0.011 -0.063
(-0.61) (-1.91) (0.20) (0.10) (-0.46)
Illiquid × F undLiq -0.029 0.012 -0.021 -0.040 0.006
(-1.46) (0.72) (-1.04) (-2.18) (0.29)
Bad × F undLiq 0.028 -0.068 0.302 0.351 0.060
(0.06) (-0.17) (0.74) (0.72) (0.14)
LowRed × F undLiq 0.009 -0.267 0.019 0.065 -0.141
(0.06) (-1.76) (0.10) (0.49) (-0.78)
N oLock × F undLiq -0.182 0.227 0.073 -0.007 0.287
(-1.22) (1.56) (0.51) (-0.06) (1.71)
Leverage -0.009 -0.029 -0.009 0.008 -0.009
(-0.74) (-1.54) (-0.68) (0.45) (-0.76)
Y oung -0.030 -0.055 0.006 -0.013 -0.039
(-0.36) (-0.65) (0.06) (-0.13) (-0.43)
Illiquid -0.041 -0.047 -0.039 -0.041 -0.044
(-2.52) (-2.86) (-2.40) (-2.48) (-2.71)
Bad -0.429 -0.551 -0.361 -0.350 -0.528
(-0.93) (-1.17) (-0.79) (-0.78) (-1.17)
LowRed -0.168 -0.153 -0.161 -0.198 -0.166
(-1.39) (-1.28) (-1.21) (-1.57) (-1.37)
N oLock 0.115 0.108 0.105 0.120 0.096
(0.96) (0.89) (0.85) (1.00) (0.79)
F undLiq -0.187 -0.949 0.370 0.291 -0.196
(-0.36) (-1.83) (0.56) (0.60) (-0.32)

Obs. 26,613 26,613 26,613 26,613 26,613

R2 (%) 0.074 0.088 0.086 0.062 0.089

54
 Table 12: Liquidity Provision and Hedge Funds Trading Style

OLS estimates in the daily fixed effect pooled regression of hedge funds trading costs on trading style, and trading style
interacted with funding liquidity variables. Funds are classified as following either a short-term reversal (rev = 1), long-term
momentum (mom = 1), or value (value = 1) strategy if the value-weighted decile of the stocks in their portfolios is above 0,
following the procedure described in Section 5. The dependent variable is the daily volume-weighted trading cost T CV W in
Panel A, and a dummy that takes the value of 1 if trading costs are positive, T C+ , in Panel B. The sample period is January
1999 to December 2010.

Panel A. Dep. Var. : Volume-weighted trading cost, T CV W

F undLiq - RM VIX TED LIBOR PC

rev -11.282 -11.197 -11.163 -11.331 -10.962 -11.392
(-14.76) (-14.84) (-14.93) (-14.51) (-13.49) (-14.72)
rev × F undLiq 3.445 -2.955 -0.552 1.048 -2.922
(3.54) (-3.11) (-0.53) (1.22) (-2.88)

F undLiq - RM VIX TED LIBOR PC

mom 4.697 4.702 4.532 4.650 5.148 4.561
(7.10) (7.13) (6.67) (6.89) (7.35) (6.64)
mom × F undLiq 0.598 -2.527 -0.694 1.831 -1.711
(0.73) (-3.04) (-0.78) (2.36) (-1.89)

F undLiq - RM VIX TED LIBOR PC

value 0.759 0.758 0.936 1.149 0.682 1.189
(1.20) (1.20) (1.45) (1.74) (0.98) (1.79)
value × F undLiq -2.222 3.092 2.984 -0.247 3.778
(-2.72) (3.87) (3.28) (-0.34) (4.18)

Panel B. Dep. Var. : Dummied Volume-weighted trading cost, T C+

F undLiq - RM VIX TED LIBOR PC

rev -0.089 -0.088 -0.089 -0.088 -0.089 -0.089
(-14.26) (-14.23) (-14.25) (-14.17) (-13.22) (-14.26)
rev × F undLiq 0.011 -0.000 0.004 -0.000 -0.002
(1.91) (-0.06) (0.68) (-0.05) (-0.42)

F undLiq - RM VIX TED LIBOR PC

mom 0.059 0.059 0.058 0.058 0.060 0.058
(10.45) (10.45) (10.23) (10.30) (10.20) (10.22)
mom × F undLiq 0.002 -0.019 -0.011 0.004 -0.015
(0.34) (-3.58) (-2.01) (0.54) (-2.75)

F undLiq - RM VIX TED LIBOR PC

value 0.003 0.003 0.004 0.004 -0.001 0.004
(0.60) (0.60) (0.69) (0.83) (-0.20) (0.81)
value × F undLiq -0.006 0.008 0.010 -0.014 0.010
(-1.12) (1.70) (1.83) (-2.20) (1.97)

55
 Table 13: Liquidity and Hedge Funds Trading Intensity

OLS estimates of the pooled regression of future weekly adjusted bid-ask spread ASP Ri,w on lagged determinants. The spread
enters in levels in columns (1), (2), (4), and (5), while columns (3) and (6) use the weekly change (∆) in spread. HF Tradeint is
the trading intensity of hedge funds in Ancerno, as defined in Section 6, while OI Tradeint is the trading intensity of all other
institutions in Ancerno. RM and ST DM denote, respectively, the weekly sum and standard deviation of the daily returns to
the CRSP VW Index. Ri and ST Di denote analogous measures for company i returns. T U RNi is the ratio between the weekly
volume, computed as the sum of number of shares traded in each day of the week, to the number of shares outstanding. All
variables are standardized to mean zero and variance one. ‘F-test’ reports the p-value of the F-test for the null hypothesis that
the coefficients on HF Tradeint are jointly zero. Below the coefficients, t-statistics based on robust standard errors clustered at
the permno level are reported in parentheses. Each regression also includes three lags of the dependent variable. The estimates
of the intercept and lag terms are omitted. The sample period is January 1999 to December 2010.

Dep. Var.: ASP Ri,w ASP Ri,w ∆ASP Ri,w ASP Ri,w ASP Ri,w ∆ASP Ri,w
(1) (2) (3) (4) (5) (6)
HF Tradeint i,w−1 -0.143 -0.080 -0.085 -0.108 -0.074 -0.069
(-2.73) (-2.07) (-2.27) (-2.15) (-2.17) (-2.08)
HF Tradeint i,w−2 -0.026 -0.020
(-0.42) (-0.39)
HF Tradeint i,w−3 -0.047 -0.094
(-0.76) (-2.09)
OI Tradeint i,w−1 0.043 -0.075 -0.092
(0.43) (-0.90) (-1.10)
OI Tradeint i,w−2 -0.066
(-0.62)
OI Tradeint i,w−3 0.037
(0.37)
RM,w−1 -0.203 -0.179
(-2.01) (-2.50)
RM,w−2 -0.193 -0.176
(-2.34) (-2.97)
RM,w−3 -0.038 -0.005
(-0.50) (-0.09)
Ri,w−1 -1.592 -2.193 -2.275 -1.369 -1.957 -2.023
(-8.97) (-19.68) (-20.70) (-10.25) (-20.89) (-21.95)
Ri,w−2 -0.761 -0.708
(-6.26) (-6.45)
Ri,w−3 0.157 0.106
(1.34) (1.09)
ST DM,w−1 -0.103 -0.177
(-1.52) (-3.29)
ST Di,w−1 0.116 0.023 0.190 0.128
(0.78) (0.20) (1.47) (1.21)
T U RNi,w−1 -0.017 -0.031 -0.032 -0.026
(-1.30) (-2.00) (-1.40) (-1.87)
∆ST Di,w−1 0.501 0.411
(4.76) (4.52)
∆T U RNi,w−1 -0.026 -0.029
(-1.97) (-2.41)

Obs. 197,040 459,900 457,896 192,342 447,820 445,967

Fixed Effects No Week Week No Week Week
R2 0.96 0.95 0.26 0.97 0.96 0.26
F-test 0.04 0.04 0.02 0.01 0.03 0.04

56
 Figure 1: Sensitivity to Funding Liquidity for Funds in TASS and Ancerno

For each hedge fund management company in TASS we estimate separate regressions of monthly returns on the Fung and Hsieh
(2001) risk factors and each of the five funding liquidity variables. These are the market return (RM ), the change in VIX,
TED, LIBOR, or the first principal component thereof (PC). When including RM or change in LIBOR we exclude from the
Fung and Hsieh (2001) factors the market and change in 10-year yield, respectively. The figures report the kernel densities of
the loadings on the liquidity variables for the funds that are in TASS but not in Ancerno (solid line) and for the funds that are
in TASS and report to Ancerno (dotted line).

RM VIX TED
150

100
40

80
30

100

60
20

40
50
10

20
0

−.02 0 .02 .04 .06 −.01 −.005 0 .005 .01 −.015 −.01 −.005 0 .005 .01

LIBOR PC
100
80

80
60

60
40

40
20

20
0

−.015 −.01 −.005 0 .005 .01 −.01 −.005 0 .005 .01

57
 %
%

0.00
0.05
0.10
0.15
0.20
0.25

−0.10
0.00
0.10
0.20
19
19 99
99 19 q1
19 q1 99
99 19 q2
19 q2 99
99 19 q3
19 q3 99
99 20 q4
20 q4 00
00 20 q1
20 q1 00
00 20 q2
20 q2 00
00 20 q3
20 q3 00
00 20 q4
20 q4 01
01 20 q1
20 q1 01
01
20 q2 20 q2
01 01
20 q3 20 q3
01 01
20 q4 20 q4
02 02
20 q1 20 q1
02 02
20 q2 20 q2
02 02
20 q3 20 q3
02 02
20 q4 20 q4
03 03
20 q1 20 q1
03 03
20 q2 20 q2
03 03
20 q3 20 q3
03 03
20 q4 20 q4
04 04
20 q1 20 q1
04 04
20 q2 20 q2
04 04
20 q3 20 q3

58
04 04
20 q4 20 q4
05 05
20 q1 20 q1
05 05
20 q2 20 q2
05 05
20 q3 20 q3
05 05
20 q4 20 q4
06 06
Value−Weighted

20 q1 20 q1

Equally−Weighted
06 06
20 q2 20 q2
06 06
20 q3 20 q3
06 06
20 q4 20 q4
investors reporting to Ancerno. The sample is January 1999 to December 2010.

07 07
20 q1 20 q1
07 07
20 q2 20 q2
07 07
20 q3 20 q3
07 07
20 q4 20 q4
08 08
20 q1 20 q1
08 08
20 q2 20 q2
08 08
20 q3 20 q3
08 08
20 q4 20 q4
09 09
20 q1 20 q1
09 09
20 q2 20 q2
09 09
20 q3 20 q3
09 09
20 q4 20 q4
10 10
20 q1 20 q1
10 10
20 q2 20 q2
10 10
20 q3 20 q3

Other Inst.
Other Inst.

10 10
q4 q4
Hedge Funds
Hedge Funds
Figure 2: Hedge funds’ and other institutions trading costs

Quarterly volume-weighted (top plot) and equally-weighted (bottom plot) trading costs of hedge funds and other institutional