DOES WINNER S CURSE HYPOTHESIS EXIST IN EXPLAINING THE UNDERPRICING PHENOMENON OF MALAYSIAN SHARIAH-COMPLIANT IPOs?

Volume: 2 Issues: 6, pp.176-189 Internatonal ournal of Accountng, Fnance and Busness eissn: 0128-1844 ournal webste: www.afb.com DOES WINNER S CURSE HYPOTHESIS EXIST IN EXPLAINING THE UNDERPRICING PHENOMENON OF MALAYSIAN SHARIAH-COMPLIANT IPOs? * Zunadah Sulong 1, Nor Azzan Che Emb 2 and Mohd Rahm Arffn 3 1 Faculty of Economcs and Management Scences Unverst Sultan Zanal Abdn Gong Badak Campus, Terengganu, Malaysa. E-mal: zunadah@unsza.edu.my 2 Kullyyah Economcs and Management Scences Internatonal Islamc Unversty of Malaysa 3 Faculty of Islamc Contemporary Studes Unverst Sultan Zanal Abdn Accepted date: 2 December 2017 Publshed date: 15 anuary 2018 To cte ths document: Sulong, Z., Emb, N. A. C., & Arffn, M. R. (2017). Does Wnner s Curse Hypothess Exst In Explanng The Underprcng Phenomenon Of Malaysan Sharah-Complant IPOs? Internatonal ournal of Accountng, Fnance and Busness (IAFB), 2(6), 176 189. Abstract: The Sharah-complant status seems to be the maor concern of the Securtes Commsson (SC) even though n realty the Malaysan stocks market players and the publc lsted companes (PLCs) owners are domnated by non-muslms. The need for Sharahcomplant status seems more sgnfcant when the SC n 2004 ntroduced the pre-ipo Sharahcomplant lst for new ssues. Ths study ams to examne whether the short-run performance of Malaysan IPOs experence wnner s curse n a sample of 153 Sharah IPOs ssued by Malaysan companes between anuary 2005 and December 2014. The study uses two alternatve measures to test the presence of wnner s curse; allocaton rate (ALLOCT) or prvate place to nsttutonal nvestors (DPRIVATE).The negatve assocatons between both of these varables and ntal returns suggest the exstence of wnner s cursen the Malaysan IPOs regardless of Sharah status and level of Sharah complance. Ths reveals a phenomenon where the hgh ntal returns are purposely created by the ssuers (by offerng the IPOs at a deep underprcng) to allure the unnformed nvestors nto the markets to replace the lack of nterest from the nformed nvestors. Keywords: Intal Publc Offerngs (IPOs) Wnner s curse Underprcng Malaysan Sharahcomplant IPO Introducton Equty captal stll represents as the man source of external fnancng for many frms n Malaysa, next to bank borrowngs, partcularly because the bond market n the country remans relatvely very small. Therefore, the frms frst ssuances of equty to the publc are always treated as a crucal event for ssuers and underwrters to dentfy the extent of the success of the new ssues. As such, ths event whch s more commonly referred to as ntal publc offerngs (IPOs) s sgnfcant due to ts role as maor fund provders for lsted frms n the process of expanson. 176

Intal returns of ntal publc offerngs (IPOs), also known as underprcng, are referred to the abnormal ntal return occurs when the IPO offer prce s set much lower than the prce on the frst tradng day or when the prce on the frst day of lstng ncreases to a level much hgher than the offer prce. In Malaysa, even though there are qute a number of studes on IPO performance that have been conducted (Dawson, 1987; Yong, 1991; Wu, 1993; Arff & Shamsher, 1999; elc et al., 2001; Nur Adana & Kamarun, 2004; and Abdul Rahm & Yong, 2008), those that concentrates on the Islamc segment of the market s rather scant despte the fact that the emphass on Sharah-complant status and ts mpact towards the stock performance s on the rse (see for example, (Abdul Rahm, & Yong, 2012). The Sharah complance status means that the company s actvtes should be free from any form of nterest based transactons, unethcal elements and doubtful transactons. The Sharahcomplant status seems to be the maor concern of the SC even though n realty the Malaysan stocks market players and the publc lsted companes (PLC) owners are domnated by non- Muslms. The need for Sharah-complant status seems more sgnfcant when the SC n 2004 ntroduced the pre-ipo Sharah-complant lst for new ssues. The ncreasng statstcs on PLC wth Sharah-complant status (refer to Appendx A) traded n the Bursa Malaysan are consstent wth the rsk control elements n the Sharah status that attract both Muslm and non- Muslm nvestors. One of the most wdely cted fndngs that explan the underprcng phenomenon s the wnner s curse hypothess by Rock (1986). The study descrbes the adverse selecton problem as the man problem that restrans the unnformed nvestors from enterng the market. Ths leads to the underprcng of new ssues by ssuers n order to regan the nterest of unnformed nvestors n the IPOs. Other relevant studes on the wnner s curse nclude those by Baron (1982), Beatty and Rtter (1986), Balvers et al. (1988), Beatty (1989), and Levs (1990). A study by Chowdhry and Sherman (1996) on UK-style IPOs fnds that underwrters have two reasons for underprcng such as to reduce the adverse selecton problem and to reduce the probablty that the ssue wll fal due to the leakage of adverse nformaton. Based on the above scenaro, ths study also suspects that nvestors sentments play a role n the performance of Sharah complant IPOs partcularly n terms of the degree of underprcng. How far the ssues of adverse selecton that Rock (1986) descrbes as wnner s curse play a role n the IPOs aftermarket performance s among the ssues that ths study hghlghts. The study wll focus on whether the ntal performance of Sharah-complant IPOs experences the mpact of wnner s curse. The study uses the wnner s curse hypothess model on the sample of Sharah-complant IPOs ssued durng the perod of 2005 to 2014. Ths study mght contnue the effort of Rock (1986), Amhud et al. (2003) and Yong (2009) n explorng the exstence of the wnner s curse effect wth the Malaysan IPO samples. The remander of the paper s organsed as follows. Secton 2 presents past studes on Malaysan Sharah-complant IPOs and wnner s curse model. Secton 3 descrbes the research methodology, followed by secton 4 reports and dscusses on the fndngs. Secton 5 provdes the conclusons and mplcatons of the study. 177

Lterature Revew Lstng Structure of Malaysan IPOs Intal publc offerngs (IPOs) are defned as new ssues that are sold to nvestors for the frst tme. The new shares are traded n a prmary market. Due to the new ssues perceved to be sold based on the offer prce below the true values of the IPO, the actvty of flppng 1 on the frst day of tradng n order to realze quck return s a common stuaton n most share markets. In order to examne the post lstng performance, the structure of Malaysan share market tradng should be clarfed frst, as every country s shares market has ts own unque characterstcs, such as the Sharah-complant status and thrty percent of Bumputra (ndgenous people of Malaysa) mandatory requrement. In IPO tradng systems, there are three partes nvolved n IPO ssuance whch are the ssung frms, the underwrters and the nvestors. Bascally, the players n the stock market are dvded nto two categores: ndvdual nvestors (retal nvestors) and nsttutonal nvestors. Insttutonal nvestors are regarded as nformed nvestors whle retal nvestors are consdered unnformed nvestors. Investment bankng plays an mportant role n IPO ssuance. The bank acts as underwrter and helps to dstrbute a new securtes ssue n the prmary market. Normally, ther clents are the companes that wll seek the nvestment banker s advce on how to get captal or on how to rase funds through equtes. In order to sell the IPOs, there are three types of sellng mechansms beng used n share tradng. These are auctons, book buldng concepts and fxed-prced offers. Bascally, IPOs n Malaysa are sold usng the fxed-prce offer mechansm whle other country lke U.S.A uses bookbuldng mechansm. In Malaysa, the applcaton for IPO s submtted for approval to the Foregn Investment Commttee (FIC), and Mnstry of Internatonal Trade and Industry (MITI). It then s submtted to the SC for approval where t examnes the company forecast profts and dvdends. The SC lmts the market s role n the determnaton of the subscrpton prce. Most Malaysan IPOs fall wthn the category of three types of new ssues: publc ssue, offers for sale or a hybrd of publc ssue and offer for sale. Publc ssues refer to the IPOs offered to the publc for the frst tme. They are usually allocated to ndvdual nvestors better known as retal nvestors. The mplcaton of the ssuance of these new shares wll result n an ncrease n the pad up captal of the company. The publc ssues are known as prmary shares. Currently, companes gong for IPO must have 25% of shares n publc hands and have Bumputras holdng 30% shares under FIC rules 2. Offer for sale are bascally shares allocated to the orgnal shareholders (owners), but then offered to the publc to buy. The owner s dscouraged from sellng shares before the expraton of the lock-up perod. In Malaysa, the orgnal shareholders cannot mmedately sell and transfer the IPOs. The Securtes Commsson has already mposed the regulaton on the orgnal shareholder or the owner, pertanng to the sale of the shares that they own. Ths s somehow not mposed on the retal and nsttutonal nvestors. In Malaysa, the owner-managers need to face the lock-up perod of three years before they can proceed to sell ther shares n the secondary market. The lock-up perod bascally s an agreement between the underwrter and the ssung frm prohbtng the sale of shares by nsders for a certan perod of tme. The lock- 25 Flppng refers to the mmedate sale of IPO s after ts lstng process especally n the early days of tradng. 26 The new polces effectve from the year 2009 wll see the FIC scrapped the thrty percent rule for companes gong for IPO. 178

up perod averages around sx months n other countres such as the U.S.A. Lastly, prvate placement refers to the IPOs offered to nsttutonal nvestors. Past Studes on Malaysan Sharah-Complant IPOs The unqueness of the Malaysan Captal Market compared to the other countres n the Asan regon s the ssuance of Sharah complant equtes. These equty nstruments must comply wth the rules of the Sharah Advsory Councl (SAC). The two basc crtera for new IPOs to be lsted as Sharah-complant are that the prmary busness actvtes of the company must be halal and the fnancal management of the company must be free from rba (nterest) elements. At the moment, there s a lack of studes that focus on Sharah complant IPOs that deal wth underprcng, nvestors sentment, and the cold and hot market phenomenon. In general, studes on the shares market performance that traded on Sharah-complant counters versus conventonal counters have been ncreasng n Malaysa (see, Abdul Ghafar & Nur Azura, 2004; Abu Sufan et al. 2004; and Sadegh, 2008). The Sharah-complant IPOs performance s one of the new dmensons that the authors capture. Ths s part of the unque features of Malaysan IPOs. The study by Abdul Rahm and Yong (2008), however, fnd that the Sharah-complant status does not alter the pattern of the ntal return. The downward trend gves a mean ntal return of 31.99 percent as compared to advanced markets. Ths could have been caused by the Asan fnancal Crss and the FIC deregulaton on prcng restrant. Accordng to the authors, the Sharah-complant IPOs are also drven by demand (over-subscrpton rato) factors rather than (offer sze) factors. Other than that, ther study fnds a postve relatonshp between the ntal return and the type of offer (publc ssue). Wnner s Curse Hypothess: Nowadays, the basc fact about the underprcng of IPOs n the short run s well accepted but the cause of underprcng and the degree of underprcng n the IPO market s stll beng dscussed and s a source of debate around the world. In past studes by Baron (1982), Rock (1986), Beatty and Rtter (1986), Beatty (1989), and Levs (1990), one of the reasons commonly gven for underprcng s asymmetrc nformaton and the adverse selecton problem. Among the wdely cted fndngs that explan the underprcng phenomenon s the wnner curse hypothess of Rock (1986). The hypothess dvded the nvestors nto two groups known as nformed nvestors and unnformed nvestors. Asymmetrc nformaton leaves the nformed group wth nformaton on the new ssues (IPO) true values. Ths useful nformaton wll help the nformed nvestor to buy only the IPOs wth a hgher after-market prce compared to ther offer prce. On the other hand, the unnformed nvestor faces the possblty of subscrbng to IPOs that have a lower aftermarket prce compared to ther offer prce. The unnformed nvestor faces nformaton dsadvantage about the market value of new ssues. Of course, they have a chance to wn the large allocaton of new ssues when less proporton of nformed nvestors subscrbes the new ssues. The unnformed nvestors realze that they face the probablty of gettng the wnner s curse effect. Therefore, n order to ensure the success of ssuers offerng, they underprce the ssues n order to compensate nvestors. The study also descrbes the adverse selecton problem as the man cause of the underprcng of new ssues by the ssuer. 179

Among the supporters of ths hypothess are Chowdhry and Sherman (1996). They beleved that the polcy n share allocaton n a number of countres 3 ncludng Malaysa that they quoted favoured small nvestors. The polcy plays ts role n terms of the noton of farness. Ths s because wth the U.K system the underwrter s not gven as much freedom as underwrters n the U.S. system n terms of the allocaton of new ssues. Wth these strategc allocatons, the authors clam that the ssuers have a chance to choose a hgher offerng prce as there s no need for the ssuers to compensate nvestors for the adverse selecton problem. Such a polcy, accordng to them, reduces the wnner curse or adverse selecton problem for unnformed nvestors. The degree of underprcng tends to be greater wth the U.K system but s offset wth the polcy of favourng small nvestors. Other supporters of ths theory are Amhud et al. (2003). They conducted a study of adverse selecton to test the theory of wnner s curse on the Israel IPO market. They fnd that underprcng s negatvely related to the rate of allocaton to subscrbers and ths result s consstent wth the exstence of adverse selecton Rock (1986). Accordng to the study, unnformed nvestors earned slghtly less n ntal returns compared to nformed nvestors. In other words, the overprced IPOs for unnformed nvestors show that the demand from ths group for new ssues s hgh. Other supportng study s by Yong (2009). He has found the exstence of wnner s curse effect n Malaysan IPOs market usng the nformaton on 120 Malaysan prvate placement IPOs from 2001 to 2006. The result shows that unnformed nvestors prefer a hgher ntal return when the percentage of nformed nvestors enterng the market s less. The wnner s curse theory s, however, crtczed by Welch (1992) who ntroduced the cascade model or bandwagon effect. The study beleves that the potental nvestor wll mtate earler nvestors n makng purchasng decsons. Methodology The data set conssts of all Sharah-complant IPOs ssued from anuary 2005 untl December 2014 whch s lsted on the Man Market of Bursa Malaysa (prevously known as the Kuala Lumpur Stock Exchange or KLSE). As n 2009, the frst and second lstng boards were then called as the Man Market whle the MESDAQ was then ntroduced as the ACE Market. Ths study s fully aware on the effort by Bursa Malaysa and SC to upgrade the Malaysan equtes market n 2004. Prevously, the allocatons of new ssues for nsttutonal nvestor were group together under the publc ssue category. However, the prvate placement category has been separated from publc ssue category n 2004. Snce then, Malaysan IPOs are classfed nto offer for sale, publc ssue and prvate placement. Ths study also fnds that n 2004 SC has ntroduced the pre-ipo Sharah status. The data on subscrpton rato and the quantty of IPOs allocated to the nsttutonal nvestors and retal nvestors helps ths study to observe the deas n the theory of market sentments such as the wnner s curse hypothess. By observng the ntal return and allocaton rato, the study can gan a clear pcture on whether they are consstent wth the wnner s curse effect theory or otherwse. The other two man varables are extracts from the company s prospectus as they comprse the data on rato of growth funds to total IPO proceeds and IPOs earnng s rsk that 28 Chowdhry & Sherman (1996) dentfy that among the countres that tend to favour small nvestors are Hong Kong, Inda, Indonesa, Malaysa, Sngapore, Thaland, Bangladesh, and the U. K. These countres use the U.K method that places the Farness Rule n the allocaton of shares whch requres the far treatment between nvestors on the bass of order sze. Ths s n contrast wth the U.S system that gves the underwrters consderable dscreton n allocatng new ssues. 180

nfluence ntal returns of IPOs. The obectve s tested whle takng nto consderaton control varables ncludng ownershp structure, underwrter reputaton, company age, company sze, offer sze, and market condton. Data are obtaned from the Bursa Malaysa s webste and customer servce unt, Securtes Commsson webste and t s Islamc Captal Market Unt, company s prospectus and annual report, Investors Dgest, Star on-lne and OSIRIS database. The Sharah status of the company s determned based on the lst publshed by the Sharah Advsory Councl (SAC) of Securtes Commsson and Knowledge Centre of Bursa Malaysa. Man Varables Measurement The dependent varable n ths study s the ntal return of IPO or underprcng. In order to examne the mpact of wnner s curse effect on the under-prcng level, the study focuses on two alternatve measures of ndependent varables namely allocaton rate and prvate placement. The study also uses sx other ndependent varables as control varables whch nclude frm sze, frm age, offer sze, ownershp varables, underwrter reputaton and lastly, market condton. Dependent Varables: Intal Return of IPO (Underprcng) The dependent varable n ths study s the value of underprcng for IPOs. There are varous studes that focus on the determnant factors of short-run performance of the IPO and the mpact of those determnants on the underprcng level. The underprcng (ntal return) result helps ths study to develop a general pctures of the performance of the Sharah-complant IPOs. The standard measure to calculate ntal return n most studes s by calculatng the percentage change n prce on the frst day of tradng, between the offer prce and the closng prce (Rtter, 1998). Smlarly, the Malaysan researchers also used the offer-to-close as the measurement of ntal return (Dawson, 1987 and Yong, 1991). Ths method also apples n Agarwal (2008) who focuses on the closng prce and subscrpton prce on the frst day of tradng. In the sprt of Abdul Rahm and Yong (2008) and most of the current studes, two methods are used to calculate the underprcng on the frst day of tradng. The frst s the percentage change n prce from the offer prce to the openng prce, IPORTN P P OPEN, OFFER, 1 x100 (Eq. 1a) Where: P OPEN, = openng prce, and P OFFER, = offer prce The second s the percentage change n prce from the offer prce to the closng prce. Where: P CLOSE, = closng prce, and P OFFER, = offer prce IPORTN P P CLOSE, OFFER, 1 x100 (Eq.1b) 181

The advantage of usng both offer-to-open and the offer-to-close methods help the study to dentfy the pattern of underprcng n the begnnng and at the end of the tradng day. The study prefers to emphasze more on ntal returns offer-to-open (IPORTN OPEN ) results rather than ntal returns offer-to-close (IPORTN CLOSE ) results. The study agrees wth Yong [24] that IPORTN OPEN s the pure ntal return wthout the market element (nose). The outcomes of IPORTN CLOSE are stll presented for reference purposes. In a few IPOs ssuance, the study apples the weghted average offer prce method to calculate the offer prce due to dfference quotaton made for publc ssue and prvate placement ssue. Independent Varables: Allocaton Rate and Prvate Placement to Insttutonal Investor Ths study adapts and enhances the methods used n the Amhud et al. (2003) to test the exstence of wnner s curse by evaluatng the adverse selecton aspects n Malaysan Sharahcomplant IPOs. The evdence of adverse selecton s assessed by examnng the relatonshp between the allocaton rate to subscrbers and overprced IPOs. The allocaton s defned as the proporton of subscrber s order that s flled n the IPO. The data on recprocal of oversubscrpton rato s used to measure allocaton rate. The data on number of tmes oversubscrbe from 1999 to 2003 are obtaned from Investors Dgest. However, the data on number of tmes oversubscrbe from 2004 to 2008 are prepared by the Bursa Malaysa through event package as Bursa Malaysa stop publsh the Investors Dgest n the mddle of 2004. The logstc transformaton of the allocaton rate or ALLOCT as suggested by Cox (1970) s descrbed through the allocaton model below: ALLOCT = log (ALLOC + a) / (1- ALLOC + a) (Eq. 2a) Where,. a= 0.5/N and, N s the sample sze.. ALLOC = 1 / OVERSUBSCRIPTION RATIO = NO. OF UNITS ISSUED / NO. OF UNITS SUBSCRIBE (Eq. 2b) Ths study also adapts Yong (2009) method to address the exstence of the wnner s curse by measurng the mean of ntal return n the prvate placement and non-prvate placement ssues. The data s obtaned partly from Investors Dgest and the rest are from Bursa Malaysa lbrary and ts customer servce unt. The prvate placement allocatons accordng to Yong (2009) are ganed populartes snce 2001 onwards. Therefore, no prvate placements for IPOs are detected n 1999 and 2000. Ths study enhance the prevous method by creatng a dummy varable takng a value of 1 f the offer ncludes prvate placement, otherwse the value s 0. The varables are known as DPRIVATE. The number of samples n Table 3 below already exclude out the IPO s wth restrcted or specal ssues and IPO s under REITS category. Control Varables There are sx control varables that are used n ths study whch nclude company sze, company age, ownershp, offer sze, underwrter reputaton, and market condton. 182

Hypothess Testng H1: There s no evdence of wnner s curse n nfluencng the ntal returns of IPOs regardless of Sharah status and level of Sharah complance. To test the hypothess, ths study employs cross-sectonal multple regresson analyses that take the followng forms; IPORTN 1 ALLOCT CV 1, IPORTN 1 DPRIVATE CV 1, (Eq. 3a) Where; IPORTN = the regresson ntercept = the estmated coeffcents = (Closng Offer prce) / Offer prce or (Openng Offer prce)/ Offer prce. (Eq. 3b) ALLOCT = The logstc transformaton of the allocaton rate. DPRIVATE CV, = The dummy varable takes a value of 1 f the offer ncludes prvate placement. Otherwse, the value s 0. = Control varables whch are: Age Company sze Offer sze Ownershp Underwrter reputaton Market condton = The number of years between the frm s year of ncorporaton and the IPO lstng year. = The total assets that the company owns pror to lstng = The number of shares offered multpled by the offer prce. = Percentage of shares owned by fve largest shareholders after the new ssue to the total number of shares outstandng after the IPO lstng. = The reputaton of nvestment banks those are responsble for IPOs offerng. A dummy varable takes a value of 1 f the company has been advsed by one of the fve reputable underwrters and 0 otherwse. = Unusually hgh volume of offerngs and hgh lstng of new ssues. The dummy varable takes a value of 1 f t s greater than the average amount volume of offerng. Otherwse, the value s 0. = Error term 183

Result and Dscusson Correlaton Analyss Overall, based on the entre sample of IPO groups, the study fnds that there s a strong negatve correlaton between ALLOCT and both measures of ntal returns. Maorty of the sample shows hgher sgnfcance level (p < 0.01). The results as shown n Table 1 supports the study hypothess because the low ALLOCT s supposed to ndcate less presence of nformed nvestors. Therefore, the underwrters and ssuers see more needs to under-prce the new ssues to attract more nvestors (especally the un-nformed ones) nto the market. Ths strategy results n hgher ntal returns. At the same tme, the deeper underprcng also tends to create hgher demand whch translates nto hgher oversubscrpton rato (and thus lower ALLOCT). The negatve assocaton between under-prcng (ntal returns) and ALLOCT s consstent wth result by Amhud et al. (2003). As descrbed by Amhud et al. (2003), such a negatve assocaton s a sgn of the exstence of wnner s curse. Table 1: Correlaton analyss results among varables LNOFF OWNER Varables ALLOCT LNAGE LNTA DWRITER DMKT SIZE FIVE Panel A: Whole Sample IPORTN OPEN -0.3563** -0.1888** 0.0973* -0.1285** -0.1289** -0.0621 0.0069 Panel B: Non-Sharah Complant IPOs IPORTN OPEN -0.4618** -0.0864 0.0481-0.0104 0.0360-0.0174 0.3651* Panel C: Sharah Complant IPOs IPORTN OPEN -0.3437** -0.2016** 0.1069* -0.1394** -0.1466** -0.0658-0.0137 Notes: ALLOCT s the logstc transformaton of the allocaton rate derved from the number of tmes over subscrbe; LNAGE s the dfference between the year of lstng and year of ncorporated (n natural log); LNOFFERSIZE s the number of shares offered multpled by the offer prce (n natural log); LNTA stands for company sze. It s based on the consoldated total asset values that smplfed through natural log; OWNERFIVE represents ownershp concentraton whch based on the percentage of shares owned by top fve shareholders; DWRITER a symbol for underwrter reputaton. It s based on the top fve rankng of underwrter n-terms of volume and unts of shares sold posted n Bursa Malaysa webste (2007 and 2008). A dummy varable takes a value of 1 f the company has been advsed by one of the fve reputable underwrters and 0 otherwse; and DMKT s a dummy varable that represent market condton. Market condton s ndcated by unusually hgh volume of offerngs and hgh lstng of new ssues. The dummy varable takes a value of 1 f t s greater than the average amount volume of offerng, otherwse, the value s 0.The symbol ** represents correlatons that s sgnfcant at the 0.01 level (2- taled) and * s correlatons that sgnfcant at the 0.05 level (2-taled). Table 2: Descrptve statstcs of ntal return (offer-to-open) by year Yr N Mean Medan Std. Dev. Skew. Kurt. Mn. Max. Panel A : All samples 2005 79 27.104 24.000 16.520 1.230 2.506 0.000 73.913 2006 40 63.322 50.000 40.092 0.379-0.898-6.383 144.444 2007 28 24.244 3.648 50.912 1.316 0.450-26.667 133.333 2008 23 25.131 13.846 37.417 2.570 6.622-5.714 161.905 2009 14 45.555 39.773 35.131 0.724-0.151 0.000 140.000 2010 28 40.878 26.631 51.410 2.436 7.294-12.037 275.000 2011 28 18.859 7.977 41.832 1.824 5.271-66.667 194.118 2012 17 23.085 19.790 29.440 0.426 0.779-37.398 95.122 2013 17 37.127 19.091 42.254 1.247 0.047 0.000 121.429 2014 15-16.542-7.346 25.903-1.065 0.017-68.132 15.862 184

Yr N Mean Medan Std. Dev. Skew. Kurt. Mn. Max. Panel B: Sharah samples 2005 48 28.2598 24.0372 18.1951 1.1460 2.0850 0.0000 73.9130 2006 23 64.1642 52.5000 40.4358 0.3260-0.9410-6.3830 144.4444 2007 15 27.2919 6.0465 51.1654 1.2540 0.2260-26.6667 133.3333 2008 14 22.4330 13.8462 33.7285 3.0630 10.439-5.7143 161.9048 2009 8 47.1024 41.7857 36.2687 0.6200-0.3770 0.0000 140.0000 2010 8 38.4812 25.1316 49.3628 2.6350 9.0710-12.0370 275.0000 2011 12 18.4606 7.9545 43.2125 1.8710 5.3590-66.6667 194.1176 2012 8 24.1961 21.2121 30.0975 0.3630 0.7340-37.3984 95.1220 2013 7 39.8393 19.0909 43.3310 1.1230-0.3240 0.0000 121.4286 2014 10-14.7252-5.6000 24.3012-1.1260 0.4860-67.2727 15.8621 Yr N Mean Medan Std. Dev. Skew. Kurt. Mn. Max. Panel C: Non-Sharah samples 2005 24 23.867 24.000 11.592 0.758 1.561 10.000 41.667 2006 12 36.364 36.364 - - - 36.364 36.364 2007 11-21.482-21.482 - - - -21.482-21.482 2008 9 52.785 34.417 65.169 1.005-0.513 2.308 140.000 2009 6 30.080 34.058 14.775 0.197-1.167 13.333 50.000 2010 20 80.417 53.333 76.041 1.578 2.308 25.000 190.000 2011 16 21.645 11.435 32.386 1.050 0.344-16.667 84.211 2012 9 11.603 5.263 22.278 1.177 - -6.819 36.364 2013 10 10.000 10.000 14.142 - - 0.000 20.000 2014 5-25.627-12.500 37.696-1.377 - -68.132 3.750 Notes: 1. Intal return (underprcng) s calculated as the return from the offer prce to the openng prce n the frst tradng day. IPORTN P P OPEN, OFFER, 1 x100 Weghted average offer prce s use n a few cases when the retal nvestors offer prce dffers from nsttutonal nvestors offer prce. Results of Multple Regressons As shown n Table 3 and 4, the two equatons are dfferent by only one factor, namely ALLOCT (Amhud et al., 2003) or DPRIVATE (Yong, 2009) whch s alternatvely used to test the presence of wnner s curse. The nterpretaton of the results wth regard to both varables, ALLOCT and DPRIVATE, are the same. Panel B of Table 3 begns wth factors that explan ntal returns on non-sharah IPOs. Apparently, based on factors that have sgnfcant (α < 0.1) coeffcents, ntal returns based on offer-to-open (IPORTN OPEN ) of ths group are explaned by ALLOCT, DWRITER, and DMKT. In the meantme, the IPORTN OPEN on Sharah IPOs are also explaned by ALLOCT, but that s the only smlarty that the two groups show. As reported n Panel A of Table 3, other than ALLOCT, IPORTN OPEN s explaned by OFFERSIZE, AGE and OWNERFIVE, whch are completely dfferent than the other factors that nfluence IPORTN OPEN on non-sharah IPOs. The results n Table 3 have also hghlghted another mportant fndng pertanng ALLOCT, besde ts role as the man determnant factor n nfluencng IPO performance. ALLOCT has been consstently found to have negatve coeffcents n explanng the regresson analyses. As suggested by Amhud et al. (2003), such a negatve assocaton between ALLOCT and ntal 185

returns s an ndcaton of the exstence of wnner s curse effect. As prevously explaned when dscussng the correlaton analyss, the negatve assocaton suggests that ntal returns are most lkely to be hgher for IPOs that have lower ALLOCT. Lower ALLOCT suggests less partcpaton by nformed nvestors and accordngly underprcng has to be put n place to allure more unnformed nvestors nto the IPO market. Ths conecture s also supported by the results of usng prvate placement (DPRIVATE) as the alternatve measure of wnner curse as prvate placement represents subscrpton by nsttutonal nvestors who are consdered as nformed nvestors. As reported n Table 4, DPRIVATE consstently shows sgnfcant negatve coeffcents n all two groups of IPOs. Snce DPRIVATE s 1 when IPOs are offered through prvate placement and 0 otherwse, a negatve DPRIVATE coeffcent means ntal returns are hgher when the IPOs are not offered through prvate placement. In other words, the partcpatng nvestors of these IPOs are more lkely to be among retal nvestors who are generally consdered as the unnformed nvestors, subscrpton of whch are satsfed because the IPOs are not taken by the nformed nvestors. In short, such condton s referred as a wnner s curse. Table 3: Results of multple regressons for wnner s curse: ALLOCT IPORTN 1 CV S, ALLOCT 1 (3.4a) Panel A: Sharah Sample Panel B: Non-Sharah Sample Varables Exp. (Sgn) Coeffcent t-statstcs Coeffcent t-statstcs IPORTN OPEN IPORTN OPEN ALLOCT -ve -24.727-3.884*** -38.591-2.5569** Control Varables (CV) LNOFFERSIZE -ve -7.845-2.077** -7.294-0.586 LNAGE +ve 5.261 2.051** 4.692 0.564 LNTA -ve 1.678 0.424 10.346 0.803 OWNERFIVE -ve -0.375-2.139** 0.564 0.755 DWRITER -ve -3.395-0.827-35.658-2.076** DMKT (+/-)ve -2.107-0.308 34.815 2.002* INTERCEPT 135.658 3.074*** -48.168-0.394 Adusted R-Squared 0.147 0.195 F-Statstcs 7.626*** 1.969* F-Statstcs Probablty 0.000 0.084 Durbn-Watson stat. 1.461 2.295 VIF range 1.030-3.420 1.377-6.754 Notes: Autocorrelaton and heteroskedastcty n the analyss are dentfed through Breusch-Godfrey Langrange multpler test and corrected wth Newey-West covarance estmator. The adusted values are hghlghted by symbol #. The symbols ***, **, * denote sgnfcance at the 1%, 5% and 10% levels respectvely. 186

IPORTN 1 CV S, Table 4: Results of multple regressons for wnner s curse: DPRIVATE DPRIVATE 1 (3.4a) Sharah Sample (N=154) Non-Sharah Sample (N=132) Varables Exp. (Sgn) Coeffcent t-statstcs Coeffcent t-statstcs IPORTN OPEN IPORTN OPEN DPRIVATE -ve -15.401-2.623*** 5.101 0.266 Control Varables (CV) LNOFFERSIZE -ve -7.732-2.167** -11.595-0.833 LNAGE +ve 4.731 1.852* 4.630 0.493 LNTA -ve -2.098-0.608 11.315 0.747 OWNERFIVE -ve -0.383-2.027** 0.476 0.564 DWRITER -ve -3.557-0.837-23.565-1.283 DMKT (+/-)ve -6.121-0.784 50.727 2.809*** INTERCEPT 207.393 5.181*** 38.736 0.293 Adusted R-Squared 0.077 0.003 F-Statstcs 4.190*** 1.011 F-Statstcs probablty 0.000 0.456 Durbn-Watson stat. 1.379 2.158 VIF range 1.030-0.742 1.376-7.545 Notes: Autocorrelaton and heteroskedastcty n the analyss are dentfed through Breusch-Godfrey Langrange multpler test and corrected wth Newey-West covarance estmator. The adusted values are hghlghted by symbol #. The symbols ***, **, * denote sgnfcance at the 1%, 5% and 10% levels respectvely. Concluson and Implcatons It s noteworthy to hghlght that n ths study, ALLOCT has consstently showed a sgnfcant negatve relatonshp wth the ntal return performance whch s consstent wth the results of Amhud et al. (2003). In connecton wth the nterpretaton of ths result, t bascally shows the exstence of the wnner s curse effect n Malaysan IPOs, regardless of Sharah status. In short ths study provdes a strong evdence to reect the null hypothess regardng wnner s curse. Ths evdence also supports Rock s (1986) hypothess of adverse selecton (or the wnner s curse) whch argues that the nformed nvestors (nsttutonal nvestors) avod overprced IPOs. Therefore, the deeper underprcng s ntentonally created by underwrters and ssuers as a strategy to attract more nvestors (especally the unnformed ones) nto the market. For robustness, ths study also tests another alternatve measure of wnner s curse whch s DPRIVATE (Yong, 2009). The regresson results obvously shows that DPRIVATE has sgnfcant negatve coeffcent whch can be nterpreted as suggestng that the ntal returns are hgher when IPOs are not offered through prvate placement. In other words, ths result suggests that ntal returns (underprcng) are hgher when the IPOs are offered to nonnsttutonal or retal nvestors who are generally taken as unnformed nvestors. The fndngs on the consstent sgnfcant nfluence on ALLOCT and DPRIVATE should provde alarmng nformaton to the nvestors n general. The negatve assocatons between both of these varables and ntal returns suggest the exstence of wnner s curse. Ths reveals a phenomenon where the hgh ntal returns are purposely created by the ssuers (by offerng the IPOs at a deep underprcng) to allure the unnformed nvestors nto the markets to replace the lack of nterest from the nformed nvestors. Whle grabbng the hgh ntal returns mght 187

be an opportunty that the unnformed nvestors should be wllng to take, but at the same tme these nvestors must be effcent enough not to hold the assets for long-term nvestment. Ths s because the lack of nterest from the nformed nvestors suggests the poor qualty of the IPOs and n long term, the far value of the IPOs wll eventually be reflected by the market. Of man focus s the consstent sgnfcant role of ALLOCT n explanng ntal returns. Ths fndng mples the mportance of wnner s curse phenomenon n descrbng Malaysan IPO market as debated earler by Yong (2009). It also ndcates that Malaysan IPO market s no dfference than other IPO markets such as Israel as dscovered earler by Amhud et al. (2003). To a certan extent, snce the bass of ALLOCT measurement s the recprocal of the oversubscrpton rato, ths study ndrectly supports fndngs of prevous studes (Abdul Rahm and Yong, 2008; and Yong, 2008) whch clam that ntal returns are drven by the demand on IPOs. Acknowledgements Ths paper s part of a research proect on Envronmental, Socal and Sharah Governance (ESSG) of Malaysan Sharah Complant IPOs sponsored by the Mnstry of Hgher Educaton under the Fundamental Research Grant Scheme [FRGS/1/2014/SS05/UNISZA/02/1]. References Abdul Ghafar I. and Nur Azura S. (2004). An emprcal analyss of cash flow and nvestment fluctuatons usng Sharah frm-level panel data. MFA 6 th. Annual Conference Proceedngs: Revtalsng the Fnancal Market: The Tasks Ahead, 368-380. Abdul Rahm, R. and Yong, O. (2008). Intal returns of Sharah-complant IPOs n Malaysa. Captal Markets Revew, 16(2), 270-279. Abdul-Rahm, R., Che Emb, N. A., and Yong, O. (2012). Wnner's curse and IPO ntal performance: New evdence from Malaysa. Internatonal ournal of Busness and Management Studes, 4(2), 151 159. Abu Sufan, Mohd. Saharudn, H, and Hasnza. (2004). Gelagat Kemeruapan Pulangan Indeks Syarah dan Indeks Kompost Kuala Lumpur. MFA 6th. Annual Conference Proceedngs: Revtalsng the Fnancal Market: The Tasks Ahead, 381-399. Agarwal S., Lu C. and Ghon Rhee. S. (2008). Investor Demand for IPOs and Aftermarket Performance: Evdence from the Hong Kong Stock Market. ournal of Internatonal Fnancal Markets, Insttutons and Money, 18, 176-190. Amhud, Y. Hauser, S. and Krsh, A. (2003). Allocatons, adverse selecton, and cascades n IPOs: Evdence from the Tel Avv Stock Exchange. ournal of Fnancal Economcs, 68, 137-158. Arff, M. and Shamsher M. (1999). Regulatory effect as an explanaton for the excessve underprcng n an emergng market. Paper presented at the Graduate School, Unverst Utara Malaysa. Balvers, R., McDonald, B. and Mller, R. (1988). Underprcng of new ssues and the choce of audtor as a sgnal of nvestment banker reputaton. The Accountng Revew, 605-622. Baron, D. (1982). A model of the demand for nvestng bankng advsng and dstrbuton servces for new ssues. ournal of Fnance, 37, 955-976. Beatty, R. P. (1989). Audtor reputaton and the prcng of ntal publc offerngs. The Accountng Revew, (October), 693-709. Beatty, R. P. and Rtter,. (1986). Investment bankng, reputaton, and the underprcng of ntal publc offerngs. ournal of Fnancal Economcs, 15, 213-232. 188

Chowder B. and Sherman A. (1996). The wnner s curse and nternatonal methods of allocatng ntal publc offerngs. Pacfc-Basn Fnance ournal, 4, 15-30. Dawson, S.M., (1987). Secondary stock market performance of ntal publc offers, Hong Kong, Sngapore and Malaysa: 1978-1984. ournal of Busness Fnance Account, 14, 65-76. elc, R., Saadoun, B. and Brston, R. (2001). Performance of Malaysan IPOs: Underwrters reputaton and management earnngs forecasts. Pacfc-Basn Fnance ournal, 9, 457-486. Levs, M. (1990). The wnners curse problem, nterest costs and the under-prcng of ntal publc offerngs. Economc ournal, 100, 76-89. Nur Adana, H. and Kamarun N. T. M. (2004). Factors nfluencng the under-prcng of ntal publc offerngs n an emergng market: Malaysan evdence. IIUM ournal of Economcs and Management, 12(2), 194-212. Rtter,.R. (1998). Intal Publc Offerngs. Warren Gorham & Lamont Handbook of Modern Fnance. Edted by Denns Logue and ames Seward. Reprnted wth modfcaton n Contemporary Fnance Dgest, 2(1), 5-30. Rock, K. (1986). Why new ssues are underprced?. ournal of Fnancal Economcs, 15, 187-212. Sadegh, M. (2008). Fnancal Performance of Sharah-Complant Investment: Evdence from Malaysan Stock Market. Internatonal Research ournal of Fnance and Economcs, 20, 1450-2887. Yong, O., (1991). Performance of new ssue securtes n Malaysa. Malaysan Accountant, (une), 3-6. Yong. O. (2008). Intal performance of IPOs lsted on MESDAQ. Captal Markets Revew, 16(1), 27-43. Yong, O. (2009). Wnner s curse, bandwagon effect and sze effect n IPOs: Evdence from prvate placement IPOs n Malaysa. Dscusson Paper Seres. Faculty of Economc and Busness, 4, 2-23. Yong, O. (2010). Wnner s curse and bandwagon effect n Malaysan IPOs: Evdence from 2001-2009. Paper for presentaton at the Malaysa-Indonesa Internatonal Conference n Economcs, Management and Accountng (MIICEMA). U.K.M., Bang, Selangor, 25-26 November. Welch, I. (1992). Sequental sales, learnng, and cascades. ournal of Fnance, 47, 695-732. Wu, T.S., (1993). The long run performance of ntal publc offerngs n Malaysa. Captal Market Revew, 1(1), 61-80. 189