DO FUNDS FOLLOW POST-EARNINGS ANNOUNCEMENT DRIFT? Ali Coskun Bogazici Universiy Umi G. Gurun Universiy of Texas a Dallas RACT Ocober 2011 Absrac We show ha acively managed U.S. hedge funds, on average, rade on he pos-earnings announcemen drif anomaly more aggressively han muual funds. Boh muual and hedge funds ha acively rade on drif anomaly face higher arbirage risk. However arbirage risk reduces muual funds' willingness o buy high-sue socks wih high reurn volailiy, bu no hedge funds. Keywords: Pos-Earnings Announcemen Drif; Arbirage Risk Daa Availabiliy: The daa used in his sudy are publicly available from he sources indicaed in he ex. Elecronic copy available a: hp://ssrn.com/absrac=1584693
1. Inroducion We invesigae wheher hedge fund managers rade on he pos-earnings announcemen drif (hereafer PEAD). Like muual funds, hedge funds are acively managed pools of money ha hold invesmen posiions in publicly raded securiies. Unlike muual funds, however, hedge funds are largely unregulaed, are subjec o far less oversigh by regulaory bodies, and enjoy a greaer level of flexibiliy in higher-risk invesmen sraegies ha involve aking undiversified posiions. Using hand colleced daa on hedge fund holdings and Ali e al. (2008b)'s approach o measure he covariance of fund rading decisions wih SUE of he securiies, we show ha boh muual and hedge funds rade consisen wih he PEAD anomaly. Our resuls indicae ha hedge fund managers follow a more aggressive invesmen sraegy based on he PEAD anomaly compared wih muual fund managers. We show ha boh muual and hedge funds acively rading on he PEAD anomaly face higher arbirage coss, measured by high reurn volailiy. However, arbirage risk affecs muual and hedge fund invesmen decisions differenly. Arbirage risk reduces muual fund managers willingness o buy high- SUE socks wih high reurn volailiy, bu i does no have he same effec on muual funds. Our findings are consisen wih argumens ariculaed in Shleifer and Vishny (1997), who provide an insigh o evaluae he arbirage risk differeniaion beween muual and hedge funds. In heir sudy, Shleifer and Vishny examine he professional arbirage whose fundamenal feaure is an agency relaionship ha separaes he professional/arbirageur who possesses highly specialized knowledge from ouside sakeholders who give heir money o he professional o inves. Invesors may infer from a poor, shor-erm reurn ha he professional is no compeen and may wihdraw heir money. Shleifer and Vishny (1997) refer o his invesor responsiveness o performance as performance-based arbirage. Alhough volaile markes are aracive for arbirage because high reurn volailiy is associaed wih more frequen exreme prices, muual fund professionals avoid hese volaile posiions due o he fear of a fuure ouflow of funds in he case of a possible adverse price movemen. Performance-induced money ouflows are more aggressive for muual funds han for hedge funds. Hedge funds have conracual 2 Elecronic copy available a: hp://ssrn.com/absrac=1584693
resricions on wihdrawals, meaning ha invesors are no allowed o pull ou money for one o hree years. Because money is locked in heir funds, hedge fund professionals are less concerned abou he shor-erm reurn volailiy associaed wih heir invesmen sraegies such as PEAD. The remainder of he paper is organized as follows: Secion 2 describes he sample and variables. In Secion 3, we explain he research design and provide he main empirical resuls demonsraing wheher muual and hedge funds rade on he anomaly. We presen concluding remarks in Secion 4. 2. Daa 2.1 Hedge Fund Sample We use hand-colleced hedge fund sockholding daa in our research. Hedge fund holding daa are very difficul o consruc because hedge funds are no required o disclose heir performance. Exising daabases repor only hedge funds ha volunarily repor o daa vendors. This inroduces hree sources of biases: self-selecion, survivorship, and backfilling (insan hisory). A self-selecion bias exiss because funds wih good performance are more likely o choose o be included in a daabase. The survivorship bias arises in he following wo cases: (1) daa vendors sop disclosing informaion abou liquidaed funds and repor only informaion abou surviving funds; (2) successful hedge funds sop volunary disclosures afer reaching a arge asse size because hey would no longer need o coninue adverising. The backfilling bias occurs when hedge funds choose o ener a daabase only afer a good performance, which means ha a fund s pas performance would impose a selecive view of he fund s posiion. The hand-colleced daase we use comes from he mandaory quarerly holding disclosures made o he SEC. Using mandaory disclosures parially addresses he problems of self-selecion, survivorship, and insan hisory biases in he daa. 1, 2 1 Agarwal and Naik (2005), Fodor e al. (2009), Sulz (2007) discuss he imporance of daa-relaed issues in hedge fund sudies. They argue ha, oher han self-selecion and survivorship biases, measuremen errors in he daabases may undermine he 3
We follow Aragon and Marin (2009) in consrucing he hedge fund holding sample. The sample covers he period from he firs quarer of 2001 unil he las quarer of 2007. The sample conains 8,192 fund-quarer observaions of 621 unique hedge funds. The average marke value of shares held by he funds is $321,972,109 hroughou he sample period. There is a seady increase in he number of unique hedge funds over he sample period. The number of disinc hedge funds in he sample increased from 163 in 2000 o 564 in 2007. In conras, he number of unique securiies held by hedge funds remains roughly he same over he sample period. The hedge funds bough shares of 3,939 differen securiies in 2001; 3,768 differen securiies in 2002; 4,118 differen securiies in 2003; 4,406 differen securiies in 2004; 4,537 differen securiies in 2005; 4,572 differen securiies in 2006; and 4,639 differen securiies in 2007. 2.2 Muual Fund Sample We merge wo daabases (he CRSP Survivor-Bias-Free U.S. Muual Fund Daabase and he Thomson Financial CDA/Specrum U.S. & Canadian Muual Funds Holding Daabase) o creae he muual fund sample. Because he former daabase does no provide deailed informaion abou fund holdings, he wo daabases are usually combined following Wermers (2000). Afer merging hese wo daabases, we delee quarerly observaions in which funds have oal ne asses of less han $1 million and where he oal marke value of repored holdings is eiher less han half or more han 150% of he oal ne asses. The merged sample conains 29,886 fund-quarer observaions of 1,584 unique muual funds. The number of differen muual funds in he sample has a declining paern hroughou he sample period. The number of disinc muual funds is 1,465 for 2001; 1,425 for 2002; 1,320 for 2003; 1,178 for findings in he hedge fund lieraure. Specifically, hey argue ha hedge funds may use sale prices for illiquid securiies, which leads o difficulies in calculaing values. In his sudy, we focus on fund rading sraegies and do no use reurn and fee informaion provided by daa vendors. Thus, sale price bias is no an issue in his sudy. 2 Fodor e al. (2009) documen ha hedge funds do no pursue a rading sraegy consisen wih he PEAD anomaly. The findings presened in his sudy differ from hose of Fodor e al. (2009) for several reasons. Fodor e al. (2009) uses daa ha hedge funds volunarily repor. Such daa has several biases (e.g. survivorship and backfilling biases). In conras, we use handcolleced hedge fund sockholding daa from he mandaory quarerly holding disclosures made o he Securiies and Exchange Commission, and his daa parially addresses he problems of survivorship and insan hisory biases in he daa. 4
-2004; 1,085 for 2005; 966 for 2006; and 935 for 2007. The number of differen securiies held by he muual funds remains roughly he same over he sample period. In Table 1, we presen several common characerisics of hedge and muual fund sample in order o enhance he level of comparison of he wo samples. The marke value of securiies held by he funds is approximaely $1.145 billion for muual funds and $322 million for hedge funds. The mean number of socks held in a muual fund is 228, and he mean is 244 for a hedge fund. The median numbers for a muual and hedge fund are 154 and 115, respecively. INSERT TABLE 1 HERE 2.3 SUE Invesing Measure (SIM) Our variable o measure he exen o which a muual or hedge fund manager rades consisen wih exploiing he PEAD anomaly relies on he level of covariance of fund manager rading decisions wih SUE of he securiies. Formally, we compue i by adding he produc of sandardized SUEs wih acive changes in fund porfolio weighs over N, he oal number of unique socks held by he fund i in quarer and -k, i.e. following Ali e al. (2008b), we define SUE Invesmen Measure (SIM) as follows: SIM i, N j 1 ( w i, w~ i, k SUE ( SUE) ) ( SUE) (1) where SUE E E 4 mean E and E -4 are he quarerly earnings per share repored during quarer and -4 3, mean is he ime series mean of earnings surprise (E - E -4 ), 3 We follow Bernard and Thomas (1990) in compuing, SUE. Specifically, we use forecas errors from a seasonal random walk wih drif and is scale i by esimaion-period (preceding eigh quarers) sandard deviaion. A minimum of four valid observaions during he esimaion period is required in order o compue he ime-series mean and sandard deviaion. 5
σ is he ime-series sandard deviaion of earnings surprise (E - E -4 ), µ (SUE ) and σ (SUE ) are he cross-secional mean and sandard deviaion of SUE, w~ ( wi, i, k ) is he deviaion of fund i s porfolio weigh on sock j in quarer from he passive porfolio weigh. A high value of his measure indicaes ha fund managers acively buy socks wih high unexpeced earnings and/or sell socks wih low unexpeced earnings. If a fund manager deliberaely rades on PEAD, hen he or she increases holdings of high-sue socks and sells low-sue socks. This sraegy leads o a high SIM for he fund. To conrol for he passive weigh changes due o price changes and ake only he deliberae aciviy of he fund manager ino consideraion, from porfolio weigh on sock j in quarer, we deduc w~ (passive weigh) which is defined as follows: i, j, k w R w~ (1 ) i, k (2) ) k k, N w j j k R 1, (1 k, where w -k is he porfolio weigh a -k and R -k, is he reurn for sock j from quarer -k o. In addiion o SIM, following Ali e al (2008b), o beer capure he degree o which he fund managers inenionally rade on he anomaly, we also calculae he rolling averages of over quarer -3 o quarer. We refer o hese rolling averages as SIM4. This procedure reduces he likelihood of a fund being classified as one ha is acively invesing in drif anomaly, if i exhibis a high SIM value in a quarer jus by chance. To calculae PEAD for he 2001 2007 sample period, we ake all publicly raded socks wih valid SUE observaions a he end of each calculaion quarer, Q0. There are 182,151 sock-quarer observaions in his main sample. We sor socks based on SUE in order o form equal-weighed decile 6
porfolios. For each of he subsequen four quarers (expressed as Q1, Q2, Q3, and Q4), we calculae reurns o he equal-weighed decile porfolios. 4 We also creae wo subsamples from he main sample. The subsamples conain all he socks acually held by muual and hedge funds. For he sample period, here are 95,083 and 68,559 sockquarer observaions in hese subsamples, successively. Similar o he porfolio creaion procedure pursued for he main sample, we assign socks ino equal-weighed decile porfolios based on SUE decile breakpoins calculaed earlier for he main sample. In Table 2, Panel A, we presen he average values of SUE and four successive quarerly porfolio reurns (reurns for Q1, Q2, Q3, and Q4) for each equal-weighed sock deciles for he main sample. The average SUE value for he boom decile porfolio (PORT1) is -1.62, and he value for he op decile porfolio (PORT10) is 1.57. For Q1, he difference beween average reurns of PORT10 (1.82) and PORT1 (-0.30) porfolios is 2.12 and is saisically significan wih a -value of 4.85. The difference says posiive and remains highly significan for Q2 wih a difference of 1.79 and a -value of 2.38. In Q3, he posiive difference is slighly significan a he 10% significance level, and in Q4 he paern reverses and a negaive bu insignifican difference is derived. 5 INSERT TABLE 2 For he subsample of socks held by muual funds, he reurn spreads beween PORT10 and PORT1 decile porfolios are 1.85 in Q1, 1.73 in Q2, 1.31 in Q3, and -0.10 in Q4, as seen in Table 2, Panel B. The firs wo spreads, as expeced, are posiive and significan (-values are 3.51 and 2.54, respecively). The hird spread is posiive bu insignifican (-value is 1.58), and he fourh one is negaive and insignifican (-value is -0.70). The analysis using a hedge fund subsample (Table 2, Panel C) 4 Following Shumway (1997), we use he CRSP delising reurn as he reurn for he remaining days in he quarer if a securiy is delised during ha quarer. If he delising reurn is missing and he delising is performance-relaed (i.e., delising codes are 500, 520, 580, 584, or beween 551 and 574) hen he delising reurn is aken as -30%. If delising codes are anyhing oher han he above-lised codes and he delising reurn is missing, we assign zero o he delising reurn. 5 To avoid violaing he assumpion ha errors should no be serially correlaed, we use heeroskedasiciy and auocorrelaion consisen errors, and all repored -saisics in his sudy are compued following he Newey-Wes procedure wih a lag of four quarers. 7
exhibis similar resuls. In Q1, he reurn spread is 1.35 wih a -value of 2.05. Q2 also presens a posiive and significan reurn spread wih a value of 1.41 (-value is 2.29). Q3 has a posiive and insignifican reurn spread, and Q4 has a posiive and insignifican reurn spread. The spreads are 1.69 and 0.66, in ha order, wih -values of 1.54 and 0.38, respecively. Evidence in Table 2 resuls collecively sugges ha following pos earnings announcemen drif using all socks and using socks held by muual and hedge funds was profiable during he sample period. 3. Research Design and Empirical Resuls 3.1 Trading Based on he PEAD Anomaly In Table 3, we es wheher fund managers on average rade on he drif anomaly. To do so, we compue ime-series averages of he cross-secional disribuional saisics of SIMs and SIM4s. Table 3, Panel A, shows ha for muual funds, he mean and median of SIM is 0.94% and 0.30%, respecively. The mean of SIM is significan wih a -value of 11.38. The mean and median of SIM4 are 0.83% (=5.86) and 0.33%, respecively. SIM and SIM4 values in his able sugges ha muual fund managers, on average, rade consisen wih he PEAD anomaly. We ge sronger resuls for hedge funds han for muual funds. The mean and median of SIM is 1.31% (-value=4.34) and 0.43%, respecively. The mean of SIM4 is significan (-value=3.25) wih a value of 1.26%. The median value of SIM4 is 0.42%. On average, hedge funds have higher SIM and SIM4 values compared wih muual funds. The mean difference beween muual fund SIM and hedge fund SIM is -0.37 wih a -value of -1.79. For SIM4, he mean difference of -0.43 is also significan (-value=-2.05). These resuls sugges ha he covariance of boh muual and hedge rading decisions wih SUE of he securiies is posiive, indicaing ha boh ypes of fund managers rade on he PEAD anomaly and ha hedge fund managers are more aggressive when hey are involved in PEAD anomaly based invesmen sraegies. INSERT TABLE 3 HERE 8
3.2 Persisence of Trading Based on he PEAD Anomaly To argue ha fund managers acively rade wih he drif anomaly in mind and ha he documened resuls are no merely he consequences of a lucky guess, we furher examine he persisence of he rading based on his anomaly. High-SIM4 funds, which inenionally use he PEAD anomaly in heir decisions, should experience high SIM values in he subsequen quarers (quarers Q1 hrough Q4 are considered). We sor funds ino deciles based on heir SIM4 values a Q0. Top-decile (DECILE10) funds have he highes SIM4 values. Also, we selec one-enh of he funds wih SIM4 values closes o and cenered around zero and refer o hem as INACTIVE funds. In, Table 4, Panel A, we documen SIM values (in Q1 hrough Q4) of each of he muual fund deciles. For he muual funds, he op-decile fund managers rade persisenly on he PEAD anomaly. The average SIM values are 6.11, 6.08, 5.95, and 5.81, respecively, for he subsequen four quarers. The values are significanly greaer han zero in all quarers suggesing ha managers of high SIM4 funds rade deliberaely on he drif anomaly. INSERT TABLE 4 HERE In Table 4, Panel B, we conduc he same analysis for hedge funds. According o he repored values, op-decile funds persisenly follow an invesmen sraegy ha is consisen wih he PEAD anomaly. In boh panels of Table 4, we observe a decline in he value of SIM4 in Q1 for highes-sim4 decile funds when we consider he subsequen values of SIM in Q1 hrough Q4. A smaller decline in value for hedge funds (Table 4, Panel B) compared wih he muual funds (Table 4, Panel A) demonsraes ha hedge fund managers are more likely o rade on he PEAD anomaly and ha he number of fund managers who persisenly rade on he drif anomaly is higher for hedge funds. In oher words, he higher he value decrease from SIM4 o subsequen SIM values, he higher he number of funds ha have higher SIM4 values by chance. In he analysis we presened up o his poin, we do no conrol for oher well documened anomalies such as momenum, size, and book-o-marke. If reurns o hese anomalies are correlaed wih SUE sraegy, omiing covariance of a fund s rading sraegies wih hese anomalies may affec our 9
resuls. We run he following Fama-MacBeh regression in order o conrol for he influence of oher marke anomalies on our resuls: SIM k 10 i 1 DECILE i, 0 BM 4 SIZE4 MOMENTUM 4 e i (3) where he dependen variable, SIM k, is SIM values for subsequen quarers Q1 o Q4. DECILE i,0 is he indicaor variable ha akes a value of one for each decile i creaed a quarer Q0 and a value of zero for oher deciles. e i is he error erm. This model allows us o ge a differen inercep for each decile. To conrol for hree marke anomalies (price momenum, size, and book-o-marke), we creae hree invesmen measures, BM 0, SIZE 0, and MOMENTUM 0 as follows: X i, N j 1 ( w i, w~ i, k X ) ( X ) ( X ) (4) where X is he book-o-marke raio (BM) or he log of marke capializaion (SIZE) a Q0, or sock reurns 12 monhs prior o Q0 (MOMENTUM). µ (X ) and σ (X ) are he cross-secional mean and sandard deviaion over all socks included in he sample wih valid SUE observaions during he calendar quarer of earnings announcemen. w w~ ) is calculaed in he same way as in Equaion (2). ( i, i, k Afer compuing hese variables, we ake he four-quarer rolling averages and calculae BM4, SIZE4, and MOMENTUM4. Resuls in able 5 shows ha, afer conrolling for he confounding effecs of oher invesmen syles, he coefficiens of boh muual and hedge funds in he op SIM4 deciles, DECILE9 o DECILE10, are significanly posiive from Q1 o Q4. INSERT TABLE 5 HERE 3.3 Esimaed Transacion Coss We examine he impac of marke fricions on he PEAD anomaly by considering wo proxies for ransacion coss. We refer hese proxies as he indirec measures of esimaed ransacion cos. The firs measure is he quarerly cross-secional percenile rank of he Amihud (2002) illiquidiy raio 10
(ILLIQUIDITY). The daily illiquidiy raio is he raio of daily absolue sock reurn o is dollar rading volume. The illiquidiy measure is a weighed average over he formaion quarer, Q0, across socks raded by a fund. The absolue dollar value of fund rades during Q0 is used as he weigh. The second measure of ransacion coss is INVPRICE. INVPRICE is he inverse of he sock price a he beginning of he quarer. We use percenile rank scores of illiquidiy raios for ILLIQUIDITY insead of raw measures. To compue he rank score on hese given characerisics, we sor all he socks separaely by ILLIQUIDITY during Q0 and hen assign a rank score on each characerisic, where rank lies beween one (low) and 100 (high). NASDAQ and NYSE/ALTERNEXT define rading volume differenly. Thus, we compue percenile ranks for socks raded on hese wo sock exchanges separaely. Table 6 repors he average indirec ransacion cos esimaes for each SIM4 muual fund decile for he fund-ranking quarer (Q0) and he subsequen quarer (Q1). We find a U-shaped relaionship beween ILLIQUIDITY/INVPRICE and SIM4 deciles for boh quarers. According o hese wo measures, boh low- and high-sim4 muual funds are acively engaged in cosly rading. Also, aking hese wo ransacion cos measures ino consideraion, we find ha he highes SIM4-decile muual funds have higher ransacion coss han INACTIVE funds and ha hese differences are significan in almos all cases. In Q0, he differences are 0.33 wih a -value of 2.59 and 0.004 wih a -value of 17.79 for ILLIQUIDITY and INVPRICE, respecively. In Table 6 we also show he ransacion cos esimaes for each SIM4 decile of hedge funds. The U-shaped relaionship beween he ransacion esimaes and SIM4 deciles of hedge funds is also observed in Table 6. For Q0 and Q1, highes SIM4-decile funds show higher values compared wih INACTIVE funds; hese differences are significan in all cases. In Q0, he differences are 1.39 wih a - value of 8.49 and 0.008 wih a -value of 18.47 for ILLIQUIDITY and INVPRICE, respecively. The significan differences are 0.72 (-value=4.43) for ILLIQUIDITY and 0.009 (-value=18.89) for INVPRICE in Q1. INSERT TABLE 6 HERE 11
3.4 Arbirage Risk Ali e al. (2008b) show ha muual funds ha deliberaely follow he PEAD anomaly in heir invesmen sraegies have significanly higher reurn volailiy and a less diversified invesmen porfolio han INACTIVE funds. In his secion, we invesigae wheher funds following PEAD anomaly are exposing hemselves o higher arbirage risk. We use four fund characerisics as proxies for arbirage risk. The firs wo characerisics are he oal number of socks held by a fund (NUMBER) and he Herfindahl-Hirschman index of porfolio weighs (HHI). HHI is calculaed as he sum of he squares of he porfolio weighs across socks expressed as a percenage held by a fund a he beginning of Q1. These wo measures capures he componen of arbirage risk due o lack of diversificaion. Boh measures are calculaed a he beginning of Q1. We use wo measures o measure arbirage risk due o high reurn volailiy. Our firs measure is he annualized average sandard deviaion of daily reurns of socks held by a fund (STDEV) in Q1, and he second measure is annualized average sandard deviaion of idiosyncraic daily reurns of socks held by a fund (IDRISK) during Q1. IDRISK is he annualized sandard deviaion of he residual obained from he regression of daily sock reurns on daily marke reurns as well as hree-lagged and hree-lead marke reurns. Boh sock and marke reurns are adjused for risk-free rae. The STDEV and IDRISK of a fund are he averages of STDEV and IDRISK of socks held by a fund, weighed by fund sockholding weighs. A minimum of 20 valid observaions during he quarer is required in order o compue he averages of STDEV and IDRISK. Table 7, Panel A, shows ha he highes SIM4-decile muual funds hold significanly fewer socks han INACTIVE funds. The ime-series means are 63.85 and 147.34 for DECILE10 and INACTIVE funds, respecively. Smaller sock holdings and a significanly lower Herfindahl-Hirschman index for DECILE10 muual funds (6.31% for DECILE10 funds and 4.47% for INACTIVE funds) indicae ha he funds ha mos aggressively follow he drif anomaly based invesmen sraegy are less diversified. Furhermore, STDEV and IDRISK values of DECILE10 funds are significanly higher han hose of INACTIVE funds. The differences are 1.32% wih a -value of 18.70, and 2.11% wih a -value 12
of 24.33, respecively. These four measures sugges ha an acive invesmen sraegy based on he PEAD anomaly is exposed o higher arbirage risk. INSERT TABLE 7 HERE Table 7, Panel B, shows ha similar o DECILE10 muual funds, highes SIM4-decile hedge funds show a lack of diversificaion in heir sock holdings and a higher sock reurn volailiy in heir porfolios compared wih INACTIVE hedge funds. NUMBER is 199.53 for DECILE10 funds as opposed o 280.89 for INACTIVE funds, HHI is 4.59% as opposed o 3.05%, STDEV is 37.30% as opposed o 34.36%, and IDRISK is 32.79% as opposed o 29.57%. All of he differences are saisically significan. An imporan oucome from Table 7 is he difference beween he muual and hedge fund invesmen sraegies. Muual funds follow an invesmen sraegy ha is more focused. The wo columns eniled NUMBER and HHI in Table 7 sugges ha muual funds, on average, carry fewer socks in heir porfolios and ha heir sock holdings show a higher concenraion in erms of he HHI. 3.5 Porfolio Allocaion In his secion, we invesigae porfolio allocaion decisions aken by highes SIM4-decile (DECILE10) funds based on hese marke fricions. When we plo he averages of aggregae porfolio weighs and weigh changes of highes SIM-decile funds (DECILE10 funds) and, for comparison, hose of INACTIVE funds for each SUE decile, we find ha DECILE10 funds under weigh socks in he lowes SUE deciles and overweigh socks in he op SUE deciles relaive o INACTIVE funds (Table 8). These findings indicae ha he aggressiveness of DECILE10 funds in following he pos earnings announcemen drif anomaly as an invesmen sraegy. We furher examine porfolio allocaion decisions of DECILE10 funds by picking all he highes SUE-decile (PORT10) socks held and bough by funds in he main sample, and hen allocae hese socks ino wo groups. The firs group conains PORT10 socks hose ha are owned or bough by he highes SIM-decile funds and he oher group has he remaining socks hose ha are no owned or bough by DECILE10 funds. We compare wo indirec rading cos measures as well as he reurn 13
volailiy and idiosyncraic reurn volailiy of hese wo socks in Table 8. Panel A shows ha relaive o PORT10 socks no owned or bough by DECILE10 funds, PORT10 socks ha are owned or bough by DECILE10 funds are significanly liquid (ILLIQUIDITY, INVPRICE) and have lower reurn volailiy (STDEV, IDRISK). Furhermore, Table 8, Panel B shows ha PORT10 socks owned or bough by DECILE10 funds are significanly liquid compared wih PORT10 socks no owned or bough by DECILE10 funds. Similar resuls for muual and hedge fund ransacion coss are no surprising. As sophisicaed invesors, fund managers avoid rading coss for higher profis. Conrary o he resuls presened in Table 8, Panel A, for muual funds, however, PORT10 socks ha are owned or bough and PORT10 socks ha are no owned or bough by DECILE10 hedge funds do no exhibi significan differences in erms of reurn volailiy. The skill required of hedge fund managers is o produce alpha reurns (Sulz (2007)). Alpha can be explained as he performance of he invesmen sraegy ha canno be explained by he risk arising from exposure o common marke movemens. Thus, hedge fund professionals swich more easily o highly volaile markes as long as hey have higher alpha and inves in sraegies ha may ake ime o prove profiable. The resuls presened in Table 8 suppor he Shleifer and Vishny (1997) argumen ha hedge funds do no shy away from invesing in socks wih high idiosyncraic volailiy. INSERT TABLE 8 HERE 4. Concluding Remarks In his sudy, we explore wheher hedge fund managers rade deliberaely on he PEAD anomaly. We use hand-colleced hedge fund daa compiled from he mandaory quarerly holding disclosure forms submied o he SEC by he hedge fund managers. We find ha acively managed U.S. hedge and muual funds follow invesmen sraegies based on he PEAD anomaly. Our findings show ha U.S. hedge funds are more aggressive in heir drif anomaly based rading sraegies han are muual funds. We find ha acive funds coninue o employ 14
heir invesmen sraegies based on he drif anomaly for he subsequen quarers and following years. More specifically, rading on he PEAD anomaly persiss for he op en percen of funds ha mos acively rade on he drif anomaly. We documen ha his rading sraegy is robus o conrolling for oher fund-invesmen sraegies, such as size, book-o-marke, and momenum. We also documen ha hedge and muual fund managers are sophisicaed in heir invesmen decisions. They ake marke fricions ino consideraion and allocae heir rading in he presence of hese marke fricions. Boh muual and hedge funds ha more acively rade on drif anomaly have lessdiversified sock holdings and face higher volailiy of sock reurns in heir sock holdings. These feaures of securiies held in fund porfolios induce arbirage risks and could have a diminishing effec on he fund managers moivaion o exploi drif anomaly more aggressively. Managers of boh ypes of funds have similar concerns abou ransacion coss. They aemp o minimize ransacion coss in heir rading decisions. However, here is a significan difference beween muual and hedge fund managers arbirage risk preferences in heir porfolio allocaion decisions. Muual fund rades are more in line wih minimizing arbirage risk. A likely reason for his behavior is ha muual funds are more prone o invesors money wihdrawals if fund performance is emporarily poor, because hese invesors may infer fund managers abiliy based on shor-erm performances (Shleifer and Vishny 1997). In conras, in hedge funds, invesors funds are locked because of conracual consrains. Thus, hedge funds do no shy away from invesing in socks wih high volailiy, if hese socks have he poenial o provide high abnormal reurns in he long run. 15
REFERENCES Agarwal, V., and N. Y. Naik. 2005. Hedge Funds. Foundaions and Trends in Finance 1 (February): 103 170. Ali, A., X. Chen, T. Yao, and T. Yu. 2008a. Do Muual Funds Profi from he Accruals Anomaly? Journal of Accouning Research 46 (March): 1 26.. 2008b. Profiing from he PEAD: Muual Fund Trades, Marke Fricions, and Marke Efficiency. Working Paper, Universiy of Texas a Dallas, Kansas Sae Universiy, Universiy of Rhode Island, and Universiy of Iowa. Amihud, Y. 2002. Illiquidiy and Sock Reurns: Cross-Secion and Time-Series Effecs. Journal of Financial Markes 5 (January): 31 56. Aragon, G. O., and J. S. Marin. 2009. A Unique View of Hedge Fund Derivaives Usage: Safeguard or Speculaion? Working Paper, Arizona Sae Universiy, and Carnegie Mellon Universiy. Bernard, V. L., and J. K. Thomas. 1990. Evidence Tha Sock Prices Do No Fully Reflec he Implicaions of Curren Earnings for Fuure Earnings. Journal of Accouning and Economics 13 (December): 305 340. Burch, T. R., and B. Swaminahan. 2002. Earnings News and Insiuional Trading. Working Paper, Universiy of Miami, and Cornell Universiy. Cao, B., D. Dhaliwal, A. C. Kolasinski, and A. V. Reed. 2007. Bears and Numbers: Invesigaing How Shor Sellers Exploi and Affec Earnings-based Pricing Anomalies. Working Paper, McKinsey and Company, Universiy of Arizona, Universiy of Washingon, and Universiy of Norh Carolina. Fama, E. F., and J. MacBeh. 1973. Risk, Reurn, and Equlibrium: Empirical Tess. Journal of Poliical Economy 81 (May): 607 636. Fodor, A., D. Lawson,and D. R. Peerson. 2009. Do Hedge Funds Arbirage Marke Anomalies? Working paper, Ohio Universiy, Gonzaga Universiy, and Florida Sae Universiy. Grinbla, M., S. Timan, and R. Wermers. 1995. Momenum Invesmen Sraegies, Porfolio Performance, and Herding: A Sudy of Muual Fund Behavior. American Economic Review 85 (December): 1088 1105. Newey, W., and K. Wes. 1987. A Simple Posiive Semi-Definie, Heeroskedasiciy and Auocorrelaion Consisen Covariance Marix. Economerica 55 (May): 703 708. 16
Shleifer, A., and R. W. Vishny. 1997. The Limis of Arbirage. Journal of Finance 52 (March): 35 55. Shumway, T. 1997. The Delising Bias in CRSP Daa. Journal of Finance 52 (March): 327 340. Sulz, R. 2007. Hedge Funds: Pas, Presen, and Fuure. Journal of Economic Perspecives 21 (Spring): 175 194. Wermers, R. 2000. Muual Fund Performance: An Empirical Decomposiion ino Sock-Picking Talen, Syle, Transacions Coss, and Expenses. Journal of Finance 55 (Augus): 1655 1695. 17
Table 1. Comparison Table for Fund Characerisics Muual Funds Hedge Funds Number of Funds 1584 621 Number of Fund-Quarer Observaions 29,186 8,192 Marke Value of Holdings 1,144.68 321.97 Number of Socks Held - Mean 228 244 Number of Socks Held - Median 154 115 Noe: This able repors summary saisics for muual and hedge funds samples. 18
Table 2. SUE Porfolios and Reurns Panel A. All Socks STOCK DECILE SUE Q0 RET Q1 RET Q2 RET Q3 RET Q4 PORT1-1.62-0.30 0.80 0.31 1.36 PORT2-1.04-1.28-0.86 0.01 0.05 PORT3-0.67-1.24-1.17-0.84-0.28 PORT4-0.36-1.14-0.90-0.45-0.92 PORT5-0.12-0.84-0.90-0.90-0.62 PORT6 0.16 0.03-0.51-0.27-0.56 PORT7 0.43 0.53-0.09-0.05 0.09 PORT8 0.67 0.76 0.67 0.01 0.11 PORT9 1.01 0.84 1.11 0.64 0.32 PORT10 1.57 1.82 2.59 2.00 1.34 PORT10-PORT1 2.12 1.79 1.69-0.02 (4.85) (2.38) (1.69) (-0.64) Panel B. Muual Fund Sockholdings STOCK DECILE RET Q1 RET Q2 RET Q3 RET Q4 PORT1 0.07 0.82 0.52 1.19 PORT2-1.03-0.69-0.01-0.08 PORT3-1.00-1.15-0.74-0.39 PORT4-1.15-0.82-0.46-0.90 PORT5-0.90-0.93-0.89-0.56 PORT6 0.04-0.39-0.19-0.62 PORT7 0.46-0.08-0.02 0.05 PORT8 0.63 0.56 0.07-0.02 PORT9 0.76 1.05 0.43 0.52 PORT10 1.92 2.55 1.83 1.09 PORT10-PORT1 1.85 1.73 1.31-0.10 (3.51) (2.54) (1.58) (-0.70) Panel C. Hedge Fund Sockholdings STOCK DECILE RET Q1 RET Q2 RET Q3 RET Q4 PORT1-0.81-0.21-0.74 0.16 PORT2-1.42-1.11-0.31-0.53 PORT3-1.25-1.35-1.01-0.3 PORT4-1.16-1.04-0.56-0.84 PORT5-0.83-0.86-0.76-0.57 PORT6 0.11-0.27-0.24-0.43 PORT7 0.50-0.07-0.38 0.11 PORT8 0.35 0.34-0.11-0.14 PORT9 0.26 0.46 0.16-0.01 PORT10 0.54 1.20 0.95 0.82 PORT10-PORT1 1.35 1.41 1.69 0.66 (2.05) (2.29) (1.54) (0.38) Noe: This able repors mean SUE and reurns in he subsequen four quarers afer porfolio formaion, for each SUE-sored decile wihin he main sample, he muual fund subsample, and he hedge fund subsample for he sample period 2001 2007. SUE represens forecas errors from a seasonal random walk wih drif and is scaled by esimaion-period (preceding eigh quarers) sandard deviaion. A minimum of four valid observaions during he esimaion period is required in order o compue he ime-series mean and sandard deviaion or earnings. The main sample conains all he securiies lised in CRSP/COMPUSTAT merged daabase wih a valid quarer-end price of no less han $1 and wih valid SUE observaions a he end of each calculaion quarer. The muual fund subsample conains all he socks acually held by muual funds in he main sample. The hedge fund subsample conains all he socks acually held by hedge funds in he main sample. The able also repors he sock reurn differences beween he op (PORT10) and boom (PORT1) SUE sock deciles. All he -saisics presened are compued following he Newey-Wes procedure wih a fourquarer lag. 19
Table 3. Descripive Saisics of SUE Invesing Measures (SIMs) Panel A. Muual Fund 5% 25% MEAN MEDIAN 75% 95% STD SIM -20.15-3.44 0.94 0.30 4.62 24.41 14.67 (11.38) SIM4-20.74-5.33 0.83 0.33 6.64 23.8 13.9 (5.86) Panel B. Hedge Fund 5% 25% MEAN MEDIAN 75% 95% STD SIM -30.77-8.50 1.31 0.43 11.19 34.92 18.93 (4.34) SIM4-29.96-8.28 1.26 0.42 10.29 34.48 18.37 (3.25) Noe: This able repors he cross-secional summary saisics for he SUE invesing measures, SIM and SIM4, across all funds for he sample period 2001 2007. The summary saisics include mean and median values; values a he 5h, 25h, 75h, and 95h perceniles, along wih he sandard deviaion. SIM is he covariance beween acive weigh changes and cross-secional sandardized SUEs. All he - saisics presened are compued following he Newey-Wes procedure wih a four-quarer lag. 20
Table 4. Persisence of SUE Invesing Measures (SIMs) Panel A. Muual Fund SIM4 DECILE SIM4 SIM Q1 SIM Q2 SIM Q3 SIM Q4 DECILE1-22.84-2.20-2.04-2.01-2.26 (-11.93) (-4.41) (-4.31) (-4.07) (-4.38) DECILE2-9.67-1.22-1.28-1.37-1.16 (-8.30) (-3.22) (-3.38) (-3.50) (-2.96) DECILE3-5.11-1.24-1.10-1.00-0.95 (-5.26) (-3.67) (-3.06) (-2.70) (-2.45) DECILE4-2.45-0.63-0.55-0.68-0.71 (-2.56) (-2.07) (-1.58) (-1.87) (-1.91) DECILE5-0.56-0.10-0.29-0.25-0.06 (-0.73) (-0.38) (-0.99) (-0.74) (-0.16) DECILE6 1.24 0.59 0.75 0.43 0.45 (0.80) (1.98) (2.09) (1.19) (1.14) DECILE7 3.35 1.49 1.56 0.78 1.21 (3.73) (4.25) (4.05) (1.88) (2.83) DECILE8 6.46 2.58 2.66 2.25 2.34 (4.79) (6.10) (6.82) (5.56) (6.07) DECILE9 11.74 4.63 4.81 4.54 4.06 (8.97) (8.06) (8.23) (7.56) (7.21) DECILE10 26.00 6.11 6.08 5.95 5.81 (13.14) (9.31) (8.92) (8.99) (8.10) INACTIVE 0.00 0.20 0.56 0.34 0.40 (-0.20) (0.18) (0.32) (0.16) (0.30) Panel B. Hedge Fund SIM4 DECILE SIM4 SIM Q1 SIM Q2 SIM Q3 SIM Q4 DECILE1-38.18-6.86-6.67-6.15-6.44 (-15.55) (-3.23) (-3.46) (-2.82) (-3.21) DECILE2-24.07-4.22-4.34-4.12-3.88 (-15.55) (-3.28) (-2.73) (-2.81) (-2.38) DECILE3-11.7-3.59-3.53-3.76-3.22 (-13.26) (-3.01) (-2.84) (-2.06) (-1.62) DECILE4-5.11-2.36-2.42-2.68-1.96 (-10.26) (-2.19) (-1.90) (-2.56) (-1.53) DECILE5-1.28-0.57-0.40-0.57-1.02 (-3.97) (-0.82) (-0.47) (-0.52) (-1.00) DECILE6 1.80 1.35 1.26 1.00 0.32 (4.70) (1.15) (0.92) (0.69) (0.23) DECILE7 5.99 2.11 1.91 2.56 2.10 (8.14) (1.45) (1.72) (1.84) (1.38) DECILE8 12.88 3.01 3.94 3.31 3.32 (11.47) (2.26) (3.19) (2.48) (2.28) DECILE9 25.37 8.22 7.43 7.98 7.47 (13.99) (5.50) (4.07) (4.21) (3.68) DECILE10 42.53 12.38 11.62 12.77 11.55 (12.11) (5.23) (5.61) (5.90) (5.17) INACTIVE -0.02-0.56 0.12 0.17-0.77 (1.15) (-0.52) (0.07) (0.08) (-0.62) Noe: This able repors average SIMs of funds during he subsequen four quarers afer he fund ranking quarer for each SIM4-sored decile. This able also repors SIM4 in he fund ranking quarer. Top-decile (DECILE10) funds have he highes SIM4 values. Ten percen of he funds wih SIM4 values closes o and cenered around zero are named as INACTIVE funds. All he -saisics presened are compued following he Newey-Wes procedure wih a four-quarer lag. 21
Table 5. Regression of Subsequen SIM Values Muual Fund Hedge Fund SIM1 SIM2 SIM3 SIM4 SIM1 SIM2 SIM3 SIM4 DECILE1,0-2.736-4.011-3.462-2.819-2.831-0.032-2.159-3.976 (-1.41) (-2.15) (-1.84) (-1.57) (-1.61) (-0.02) (-1.91) (-1.25) DECILE2,0-2.212-1.734-1.902-1.762-2.629-1.866-2.418-2.803 (-1.71) (-1.51) (-1.65) (-1.57) (-2.06) (-1.68) (-1.61) (-1.56) DECILE3,0-1.905-2.078-1.890-1.514-1.557-2.214-2.858-1.799 (-1.45) (-1.79) (-1.33) (-1.28) (-1.31) (-1.52) (-1.54) (-1.12) DECILE4,0-1.385-1.273-1.240-0.121-0.821-2.048-2.247-1.207 (-1.72) (-1.58) (-1.49) (-0.08) (-0.58) (-1.64) (-2.09) (-1.19) DECILE5,0-0.720-0.754-0.850-0.751-0.329-0.374-1.198-0.369 (-1.07) (-1.14) (-1.02) (-0.60) (-0.38) (-0.52) (-1.04) (-0.31) DECILE6,0 0.060 0.249-0.810-2.851 1.210 1.857 2.054 0.933 (0.11) (0.29) (-0.93) (-1.01) (1.64) (1.59) (1.23) (0.79) DECILE7,0 0.812 0.762 0.118 0.690 0.694 1.274 0.573 0.757 (1.66) (0.75) (0.14) (0.64) (0.53) (0.99) (0.39) (0.38) DECILE8,0 1.629 1.966 1.302 2.176 0.890 2.187 0.551 1.855 (1.81) (2.71) (1.51) (2.61) (0.38) (1.73) (0.30) (1.21) DECILE9,0 3.124 3.777 3.662 3.616 2.921 5.597 4.533 5.040 (3.69) (8.64) (6.37) (4.06) (0.70) (2.24) (2.34) (1.77) DECILE10,0 4.077 4.484 4.685 4.716 9.193 11.091 9.804 12.003 (6.01) (5.57) (6.00) (5.41) (2.82) (3.26) (1.70) (2.51) BM4-1.182-7.188-14.019-14.040 18.160 42.470 105.35 67.400 (-2.54) (-2.07) (-3.38) (-3.93) (0.44) (0.49) (4.02) (1.11) SIZE4-0.223-0.190-0.253-2.481-1.213-0.721-1.283-0.599 (-1.83) (-1.83) (-2.30) (-3.10) (-0.72) (-0.67) (-2.08) (-0.57) MOMENTUM4 63.670 66.190 32.080 32.930 38.530 12.760 36.641 19.840 (4.44) (6.65) (4.86) (1.89) (1.17) (0.64) (2.86) (0.95) # Observaions 18,542 16,961 15,711 14,572 4,868 4,373 3,963 3,571 Adj. R 2 0.12 0.13 0.12 0.12 0.06 0.07 0.06 0.05 Noe: This able repors he esimaed coefficiens from Fama-Macbeh regressions of funds SIMs in he subsequen four quarers on dummy variables for SIM4 deciles in fund ranking quarer and BM4, SIZE4, and MOMENTUM4. DECILEi,0 is he indicaor variable ha akes a value of one for each decile i creaed a quarer Q0 and a value of zero for oher deciles. Conrol variables are for hree marke anomalies (price momenum, size, and book-o-marke) and are creaed similar o SIM4. All he -saisics presened are compued following he Newey-Wes procedure wih a fourquarer lag. 22
Table 6. Transacion Cos Esimaes for Funds Muual Fund Hedge Fund Quarer (Q0) Esimaes Quarer 1 (Q1) Esimaes Quarer (Q0) Esimaes Quarer 1 (Q1) Esimaes SIM4 DECILE ILLIQUIDITY INVPRICE ILLIQUIDITY INVPRICE ILLIQUIDITY INVPRICE ILLIQUIDITY INVPRICE DECILE1 24.19 0.046 24.15 0.046 26.55 0.049 26.42 0.049 DECILE2 23.87 0.042 23.77 0.042 22.66 0.046 22.33 0.044 DECILE3 23.95 0.041 23.95 0.041 21.60 0.042 21.44 0.043 DECILE4 22.68 0.04 22.49 0.041 23.55 0.047 23.27 0.046 DECILE5 18.23 0.037 18.36 0.037 18.96 0.038 19.02 0.039 DECILE6 16.79 0.035 16.87 0.036 18.10 0.037 18.70 0.039 DECILE7 20.07 0.038 19.90 0.038 19.77 0.043 19.58 0.042 DECILE8 22.00 0.04 22.15 0.041 20.97 0.046 21.38 0.047 DECILE9 21.73 0.041 21.96 0.041 21.64 0.046 22.38 0.049 DECILE10 21.10 0.043 21.03 0.044 20.51 0.047 19.96 0.049 INACTIVE 20.77 0.039 20.82 0.040 19.12 0.040 19.24 0.04 D10 - INACTIVE 0.33 0.004 0.21 0.004 1.39 0.008 0.72 0.009-2.59-17.79-1.61-15.79-8.49-18.47-4.43-18.89 Noe: This able repors he average indirec ransacion cos esimaes for each SIM4 muual fund decile for Q0 and Q1. These measures of esimaed ransacion cos are he quarerly crosssecional percenile rank of he Amihud (2002) illiquidiy raio and he inverse of he sock price a he beginning of he quarer. All he -saisics presened are compued following he Newey- Wes procedure wih a four-quarer lag. 23
Table 7. Diversificaion and Arbirage Risk Proxies Panel A. Muual Fund SIM4 DECILE NUMBER HHI (%) STDEV (%) IDRISK (%) DECILE1 75.93 6.74 37.64 32.03 DECILE2 84.98 5.83 37.07 31.72 DECILE3 103.70 5.14 36.21 31.03 DECILE4 122.33 4.34 35.26 29.91 DECILE5 128.78 3.93 34.63 29.28 DECILE6 133.58 4.10 34.36 29.23 DECILE7 119.57 4.43 35.15 29.66 DECILE8 99.33 5.15 37.03 31.36 DECILE9 91.94 5.75 37.60 32.18 DECILE10 63.85 6.31 36.07 30.52 INACTIVE 147.34 4.47 34.75 28.41 D10 - INACTIVE -83.49 1.85 1.32 2.11 (-10.57) (6.02) (18.70) (24.33) Panel B. Hedge Fund SIM4 DECILE NUMBER HHI (%) STDEV (%) IDRISK (%) DECILE1 164.39 4.75 37.68 33.07 DECILE2 324.54 3.81 37.72 33.14 DECILE3 239.18 3.48 36.87 32.15 DECILE4 288.18 3.15 36.11 31.44 DECILE5 410.60 3.10 34.62 29.88 DECILE6 417.40 2.63 34.19 29.53 DECILE7 408.60 2.53 35.58 30.96 DECILE8 284.85 3.13 35.26 30.66 DECILE9 257.12 4.49 36.23 31.71 DECILE10 199.53 4.59 37.30 32.79 INACTIVE 280.89 3.05 34.36 29.57 D10 - INACTIVE -81.36 1.54 2.93 3.22 (-9.71) (5.42) (22.51) (28.97) Noe: This able exhibis ime-series means of fund characerisics for each SIM4-decile. Toal number of socks held by a fund (NUMBER) and he Herfindahl-Hirschman index of porfolio weighs (HHI) are presened as measures of diversificaion in hedge fund porfolios. Boh measures are calculaed a he beginning of he subsequen quarer afer he fund ranking quarer. The Herfindahl-Hirschman index of porfolio weighs measures he fund porfolio concenraion and is calculaed as he sum of squares of he porfolio weighs across socks expressed as a percenage held by a fund. Two measures are addressing high reurn volailiy of socks held by he funds. They are he annualized average sandard deviaion of daily reurns of socks held by a fund (STDEV), and he annualized average sandard deviaion of idiosyncraic daily reurns of socks held by a fund (IDRISK). IDRISK is he annualized sandard deviaion of he residual obained from he regression of daily sock reurns on daily marke reurns as well as hree-lagged and hree-lead marke reurns. Boh sock and marke reurns are adjused for risk-free rae. The STDEV and IDRISK of a fund are he averages of STDEV and IDRISK of socks held by a fund, weighed by fund sockholding weighs. All he -saisics presened are compued following he Newey-Wes procedure wih a four-quarer lag. 24
Table 8. Transacion Coss and Arbirage Risk Characerisics on Porfolio Allocaion Panel A. Muual Fund Highes SUE-Decile Socks Owned vs. No Owned by DECILE10 Funds ILLIQUIDITY INVPRICE STDEV (%) IRISK (%) Owned 21.75 0.04 39.92 29.32 No Owned 22.63 0.04 49.04 34.18 Owned - No Owned -0.86 0.00-8.80-4.86 (-3.64) (-2.05) (-6.97) (-4.36) Highes SUE-Decile Socks Bough vs. No Bough by DECILE10 Funds ILLIQUIDITY INVPRICE STDEV (%) IRISK (%) Bough 21.51 0.05 42.31 28.45 No Bough 22.27 0.04 52.19 33.12 Bough - No Bough -0.76 0.00-8.90-4.67 (-2.14) (0.77) (-7.79) (-2.15) Panel B. Hedge Fund Highes SUE-Decile Socks Owned vs. No Owned by DECILE10 Funds ILLIQUIDITY INVPRICE STDEV (%) IRISK (%) Owned 10.15 0.04 36.94 31.32 No Owned 10.88 0.04 36.05 30.76 Owned - No Owned -0.74-0.01 0.89-0.56 (-1.97) (-1.78) (1.21) (-1.41) Highes SUE-Decile Socks Bough vs. No Bough by DECILE10 Funds ILLIQUIDITY INVPRICE STDEV (%) IRISK (%) Bough 12.82 0.04 35.65 29.11 No Bough 13.29 0.04 35.78 28.62 Bough - No Bough -0.47-0.01 0.13-0.49 (-1.73) (-1.66) (0.58) (1.55) Noe: This able repors he average ransacion cos measures and arbirage risk measures for he socks in he highes SUE-sored decile (PORT10) ha are owned or bough by DECILE10 funds, and hose for PORT10 socks no owned or bough by DECILE10 funds for he sample period 2001 2007. All he -saisics presened are compued following he Newey-Wes procedure wih a four-quarer lag. 25