WINNER S CURSE AND IPO INITIAL PERFORMANCE: NEW EVIDENCE FROM MALAYSIA

WINNER S CURSE AND IPO INITIAL PERFORMANCE: NEW EVIDENCE FROM MALAYSIA Ruzta Abdul Rahm Faculty of Economcs and Management Natonal Unversty of Malaysa, 43600, UKM Bang, Selangor, MALAYSIA E-mal: ruztaar@ukm.my Nor Azzan Che Emb Kullyyah of Economcs and Management Scences, Internatonal Islamc Unversty of Malaysa, 53100, Jalan Gombak, Kuala Lumpur, MALAYSIA E-mal: zanebbm@yahoo.com Othman Yong UKM-Graduate School of Busness Natonal Unversty of Malaysa, 43600, UKM Bang, Selangor, MALAYSIA E-mal: othmanyo@ukm.my - Abstract - Ths study examnes the wnner s curse hypothess n a sample of 384 IPOs lsted on Bursa Malaysa from 1999 to 2008. We use two alternatve measures to test wnner s curse; allocaton rate (ALLOCT j ) measures the probablty of gettng the allocaton whle prvate placement (D PRIVATE ) measures the nsttutonal nvestor nvolvement n the IPOs. The coeffcent of both ALLOCTj and D PRIVATE are sgnfcantly negatve consstent wth wnner s curse hypothess. Investors are more lkely to get (wn) most of ther subscrpton f the IPOs are overprced. Smlarly, nvestors dentfy the ssues that do not nvolve nsttutonal nvestors as rsky and overprced. Issuers/underwrters proactvely respond by delberately underprcng the IPOs to allure the unnformed nvestors nto the market. Keywords: Wnner s curse; Allocaton rate; Prvate Placement; Intal Publc Offerngs; Malaysan IPO market JEL Classfcaton: D53, D82, G02, G11 1. INTRODUCTION Intal returns of ntal publc offerngs (IPOs) have been accepted as a unversal phenomenon and a puzzle that researchers consder fndng ts explanatons a challenge that s too appealng to gnore. Also known as underprcng, the abnormal ntal return occurs when the IPO offer prce s set much lower than the 151

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. Among the most cted studes on IPO ntal performance s Loughran et al. (1994) whch fnds that IPO underprcng exsts n all 25 countres that are examned. The results of ths study substantate the evdence of underprcng that have been establshed n earler studes ncludng by Relly and Hatfeld (1969) and Ibbotson and Jaffe (1975) for the US, Chowdhry and Sherman (1996) for the UK, Dawson (1987) and Yong (1991) for Malaysa, and Fnn and Hngham (1988) for Australa to name a few. Among the wdely accepted explanatons of IPO abnormal ntal performance are the sgnalng models (Leland & Pyle, 1977), wnner s curse hypothess (Rock 1986), nvestors faddsh behavor (Aggarwal & Rvol, 1990) and cascades n the IPO market (Welch 1992). These explanatons are commonly mutually exclusve and rooted to the problem of nformaton asymmetry (Akerlof, 1970). Of man nterest to the present study s the wnner s curse hypothess whch posts adverse selecton problem as the man reason restranng unnformed nvestors from enterng the IPO market. The nformed nvestors are not lkely to be trapped or cursed wth the overprced IPOs because wth access to more and/or better nformaton, they can dstngush hgh from low qualty nvestment. To regan the nterest of unnformed nvestors, ssuers and underwrters proactvely respond by strategcally underprcng the new ssues. A recent study by Chowdhry and Sherman (1996) on the UK-style IPOs lends a strong support for ths proposton. They explan that ssuers and underwrters have two motves to underprce the IPOs.e., to reduce the adverse selecton problem and to reduce the probablty that the ssue fals (undersubscrbe) due to the leakage of adverse nformaton. A recent study by Amhud et al. (2003) on Israel IPOs fnds that underprcng s negatvely related to IPO allocaton. They argue that ths evdence s consstent wth Rock s (1986) explanaton on the adverse selecton ssue. Unnformed nvestors wn (receve a greater porton of) the IPOs when the new ssues are overprced and lose (receve a smaller porton of) the IPOs when the IPOs are underprced. Ths s because underprced IPOs are more lkely to be subscrbed by and/or allocated to the nformed nvestors. In Malaysa, the number of studes on IPO performance s qute mpressve (for example Dawson, 1987; Yong, 1991; Jelc et al., 2001; Hau Abdullah & Mohd, 2004; Abdul Rahm & Yong, 2010). However, studes that focus on wnner s curse (Yong, 2011) are stll relatvely scant. Ths study attempts to close the gap n the lterature by examnng the phenomenon of wnner s curse durng the more recent perod that spans from 1999 to 2008 and by offerng another measure of wnner s curse proposed by Amhud et al. (2003). These are done by employng a 152

unque data set for IPOs lsted on Bursa Malaysa where nformaton on oversubscrpton rato of the IPOs and prvate placement ssue are avalable. Malaysa also offers a great settng for testng the wnner s curse hypothess because as an emergng market, nvestors n the IPO market are more lkely to experence relatvely severe asymmetrc nformaton problem. The remander of the paper s organzed as follows. Secton 2 presents an overvew of the Malaysan IPO market followed by secton 3 whch dscusses the data and research methodology. Secton 4 reports and dscusses the fndngs and secton 5 concludes and draws mplcatons from the results. 2. MALAYSIAN IPO MARKET Companes gong publc n Malaysa can seek for lstng on the Man Board, Second Board and MESDAQ (Malaysan Exchange of Securtes Dealng and Automated Quotaton) of Bursa Malaysa, dependng on the pad-up captal. IPOs are offered n two broad classes; prvate placement s offered to nsttutonal nvestors whle the non-prvate placement whch conssts of publc ssue and offer for sale s offered to ndvdual or retal nvestors. IPOs are generally sold through a fxed-prce offer mechansm, unlke n the US where book-buldng mechansm mostly practced. Smlar to the UK and many Asan countres (Agarwal et al. 2008), ssuers and underwrters of IPOs n Malaysa adopt the UK far allocaton polcy where all applcatons wll be allocated randomly and equally. Also, there s no wthdrawal opton allowed once the allotment decson s publczed unlke the case for Tawanese IPO markets (Ln et al. 2010). Studes on Malaysan IPOs s poneered by Dawson (1987) who fnds the average ntal returns of 21 new ssues from 1978 to 1983 are about 166.7 percent. Fgure 1 to llustrate the trend of ntal returns over the 31-year perod (1978-2007) n Malaysa. The yearly ntal returns pror to 1990 are rough estmates, but the declnng trend s stll clearly observed. The average ntal returns durng the three sub-perods declne from 125.07 percent (1978-1989) to 94.91 percent (1990 1998) to a mere 31.99 percent (1999 2007). The md sub-perod of 1990-1998 shows large uncertantes n the IPO market whereas after 1999 the ntal returns reman consstently below the all-tme average of 83.08 percent. 153

Fgure-1: Yearly ntal returns of Malaysan IPOs, 1978 to 2007 Notes: Yearly ntal returns pror to 1990 are estmated from the weghted average ntal returns reported n Dawson (1987) and Loughran et al. (1994). Yearly ntal returns for 1990-1998 perod are extracted from Yong and Isa (2003) whle for 1999-2007 perod are from Abdul Rahm and Yong (2010). 3. DATA AND METHODOLOGY The data set conssts of 384 fxed-prced IPOs ssued n Malaysa from January 1999 untl December 2008 and lsted on ether the Man Board, Second Board or MESDAQ of Bursa Malaysa (prevously known as Kuala Lumpur Stock Exchange or KLSE). On 25 March 2008 the frst two boards are merged to form the Man market whle the MESDAQ s converted to the ACE market. Ths study ses 1999 as ts startng pont to reduce the nfluence of the 1997/98 Asan fnancal crss. Ths year (3 May 1999) also marks the begnnng of the mandatory share moratorum mposed on selected companes on Man Board and all companes on Second Board and MESDAQ. The study covers the perod untl 2008 when the effect of the 2007/08 US sub-prme crss s at ts foulest. However, the Malaysan stock market s not adversely affected by the crss, not to the extent of the Asan fnancal crss. Durng the 10-year perod covered n ths study, there are a total of 426 new ssues. The study excludes specal purposes and REITS IPOs. The selecton crtera leave the study wth a fnal sample of 384 IPOs, of whch 134 (35%) nvolve prvate placement ssue. The data s collected from varous sources ncludng the webste of Bursa Malaysa and Securtes Commsson of Malaysa, Datastream and company s prospectus. Intal return or underprcng of IPOs (IPORTN) s measured usng the two methods most commonly used n prevous studes (for examples, Abdul Rahm & Yong, 2010; Agarwal et al., 2008; Dawson, 1987; Yong, 1991). One s ntal return based on the offer to openng prce and the other s ntal return based on 154

the offer to closng prce on the frst day of tradng. We focus on the former whch s superor because t s least nfluenced by market noses. The man varable to explan ntal returns s the wnner s curse whch s operatonally measured based on the logstc transformaton of the allocaton rate (ALLOCT j ) adopted from Amhud et al. (2003) and alternatvely, the nsttutonal nvestor nvolvement n prvate placement ssue as proposed by Yong (2011). D PRIVATE s a dummy varable that takes a value of 1 when the IPOs nvolve prvate placement and 0 otherwse. In examnng the evdence of wnner s curse n the ntal performance of the IPOs, we take nto consderaton the nfluences of eght other varables: growth motve, company s sze, age and busness rsk, ownershp of top fve shareholders, offer sze, underwrter reputaton, and market condton. The econometrc equatons are represented as follows; IPORTN J 1ALLOCT J CV j, j 1 IPORTN J 1DPRIVATE J CV j, j 1 where α s the constant term, β s coeffcent estmates, IPORTN s the ntal return of the th IPO, ALLOCTj s the logstc transformaton of allocaton rate, D PRIVATE s dummy varable for prvate placement ssue, CV s the controlled varable j = AGE,, D MKT, Age s the age of the company pror to IPO n years, LnCOSIZE s the natural log total assets, LnGROWTH s the natural log of growth purposes/total proceeds, LnOFFSIZE s the natural log of (total unts ssued x offer prce), D UNDEWTR s the dummy varable for underwrter reputaton, LnOWNSHIP s the natural log of the top 5 shareholders ownershp, D MKT s dummy varable for market condton, OPERISK s operatng rsk or standard devaton of the EBITDA over three years perod pror to IPO, and s the error term. 4. RESULTS AND DISCUSSION Table 3 summarzes the profle of the IPO companes based on the average values from 1999 to 2008. In general, the offer sze of the IPOs suggests that the companes are sellng at only 26 percent of the value of ther total assets. On average, the companes total assets worth more than RM200 mllons. At a glance, ths s a good ndcaton that the IPOs are sellng at dscount, below the fundamental values. However, gven that on average the companes are makng only RM20 mllon operatng profts per year, the IPO nvestors are roughly 155

payng RM2.97 per rnggt of operatng ncome (P/E = 2.97). Wth a medan of 2 years and mean of 5 years, t seems that most companes that go publc are only n busness for few years. In addton, only 61.25 percent from the IPO proceeds are allocated for growth purposes. The top fve shareholders retan on average 59.21 stakes n the company. More than half of the IPOs are advsed by reputable underwrters. Slghtly more than one-thrd of the IPOs nvolve prvate placement and are offered durng the hot market. The average offer prce s RM1.11 whch s 28 to 30 percent lower than the prces on the frst tradng day. Ths explan the postve ntal returns (30.21%-31.65%) whch ndcate that the performance of IPOs n Malaysa durng ths perod s consstent wth fndngs of recent studes that report hgh but declnng ntal returns (Yong & Isa, 2003; Abdul Rahm & Yong, 2010). The oversubscrpton rato (OSR) s regstered at 31.40 tmes, mplyng that the IPO market s stll actve despte the declnng returns. Ths proxy of demand translates nto an allocaton rate of around 20 percent of the IPOs appled by subscrbers. Table-3: Profle of the Sample IPOs, 1999 to 2008 Items Mean Medan Mnmum Maxmum Offer Prce(RM) 1.11 1.00 0.13 4.80 Openng Prce (RM) 1.44 1.15 0.17 7.00 Closng Prce (RM) 1.42 1.11 0.12 7.50 IPORTN OPEN (%) 31.65 20.00-68.13 275.00 IPORTN CLOSE (%) 30.21 18.13-70.70 263.64 Oversubscrpton Rato, OSR (tmes) 31.40 16.50-0.89 377.96 ALLOC (0<(1/OSR) 1.0) 0.21 0.06 0.00 9.09 IPOs wth Reputable Underwrter (%) 52.34 100.00 0.00 100.00 IPOs wth Prvate placement 0.35 0.00 0.00 1.00 Hot Market Condton 0.50 0.50 0.00 1.00 Offer Sze (RM 000) 56,937 18,315 2,400 3,049,981 Proceeds for Growth (%) 61.25 64.13 0.00 100.00 Total Asset (RM'000) 218,378 86,999 6,401 6,585,143 Average 3-yr EBITDA b (RM'000) 19,184 10,533 1,362 578,033 Frm s Age (Year) 4.70 2.00 0.00 39.00 Ownershp of Top 5 Shareholders (%) 59.21 60.40 8.95 83.79 Notes: Issues of outlers due to human errors n the data collecton process has been taken care-off n the earler stage and corrected through z-score method. Table 4 shows results usng ALLOCT j and D PRIVATE alternatvely to test for the wnner s cure hypothess. Other than ALLOCTj and D PRIVATE, offer sze, age, and ownershp of the top fve shareholders are also sgnfcant n nfluencng the level of ntal returns. The sgns of these sgnfcant predctors are consstent wth the theoretcal ntuton. D MKT s nsgnfcant and gves a contradctng sgn, 156

ndcatng that ntal returns are hgher when the IPOs are ssued n the cold market. GROWTH whch represents the growth motve n IPO has the correct postve coeffcent, but the effect on ntal returns s weak. It suggests that nvestors are not partcularly concern about how the company wll be usng the proceeds from the IPOs. Smlarly, even though nsgnfcant, the negatve coeffcent on RISK EBITDA suggests that nvestors are less concern about the company s ablty to sustan operatng proft, probably because most nvestors enter the IPO market for short-term rather than long-term motve. Table-4: Regresson results, 1999 to 2008 Varables Exp. Wnner s Curse: ALLOCTj Wnner s Curse: D PRIVATE (Sgn) Coeffcent t-stats Coeffcent t-stats Intercept 116.7600 3.1543*** 190.1777 5.3792*** Wnner s Curse -ve -26.3144-4.4790*** -12.9733-2.3483** GROWTH +ve 2.9402 0.4150 5.9579 0.8708 Ln(RISK EBITDA ) +ve -0.8168-0.4614-0.2367-0.1168 Ln(OFFERSIZE) -ve -7.6350-2.3183** -7.8120-2.4748** Ln(AGE) +ve 4.5770 2.1189** 4.1041 1.8535* Ln(TA) -ve 2.8512 0.7471-0.4085-0.1213 OWNER5 -ve -0.3081-1.9173* -0.3320-1.9215* D WRITER -ve -4.4301-1.0720-3.3242-0.7655 D MKT +ve -0.5699-0.0856-2.6060-0.3357 Adjusted R 2 0.1491 0.0595 F-Statstcs 8.4555*** 3.6939*** Probablty 0.0000 0.0002 Durbn-Watson 1.4465 1.3178 Notes: ***, **, and * ndcate sgnfcance at 1%, 5%, and 10%, respectvely. The autocorrelaton and heteroscedastcty detected through Breusch-Godfrey Langrange multpler test are corrected wth Newey-West covarance estmator. VIF range s 1.033 to 4.077. 5. CONCLUSION Ths study examnes wnner s curse hypothess n 384 Malaysan IPO market ssued durng 1999-2008 perod. Wnner s curse s proxed usng two alternatve measures, ALLOCT j whch s proposed by Amhud et al. (2003) and D PRIVATE by Yong (2011). Overall, the results consstently ndcate the presence of wnner s curse ether n the form of ALLOCT j or D PRIVATE and lead us to draw the followng conclusons. Frst, the sgnfcant negatve ALLOCT j suggests that when an nvestor receves a hgh allocaton relatve to the number of IPOs that he/she subscrbes, then he/she has the tendency to end up beng cursed for wnnng overprced IPOs. Ths mples that n such cases, the nvestor may be better off wthdrawng than takng the rsk of ncurrng a larger loss. Ths wthdrawal opton (Ln et al. 2010) s unfortunately not avalable for nvestors of IPOs n Malaysa. 157

In short, nvestors must scrutnze all avalable nformaton pertnent to the IPOs n addton to that provded n the prospectus pror to makng decson to place an applcaton or subscrpton for the IPOs. Second, the negatve coeffcent on D PRIVATE suggests that when nsttutonal nvestors partcpate n the IPOs, unnformed nvestors seem to be more wllng to pay the hgh offer prce and accept lower ntal returns, probably because they are confdent about the vablty of the IPOs. On the contrary, n the absence of nformed nvestors, the unnformed nvestors need to be convnced wth larger ntal returns to make sure that the non-prvate placement IPOs are fully subscrbed. Other than ALLOCT and D PRIVATE, nvestors stand a better chance n IPO market by patronzng IPOs that are smaller and ssued by companes wth less ownershp concentraton and longer experence n the busness. BIBLIOGRAPHY Abdul Rahm, Ruzta and Othman Yong (2010), Intal Returns of Malaysan IPOs and Sharah-complant Status, Journal of Islamc Accountng and Busness Research, Vol. 1, No. 1, pp. 60-74. Agarwal, Sumt, Chunln Lu and S. Ghon Rhee (2008), Investor Demand for IPOs and Aftermarket Performance: Evdence from the Hong Kong Stock Market, Journal of Internatonal Fnancal Markets, Insttutons and Money, Vol. 18, pp. 176-190. Aggarwal, Reena and Petra Rvol (1990), Fads n the Intal Publc Offerng Market, Fnancal Management, Vol. 19, pp. 45-57. Akerlof, George A. (1970), The Market for Lemons : Qualtatve Uncertanty and the Market Mechansm, Quarterly Journal of Economcs, Vol. 85, pp. 488-500. Amhud, Yakov, Shmuel Hauser and Amr Krsh (2003), Allocatons, Adverse Selecton, and Cascades n IPOs: Evdence from the Tel Avv Stock Exchange, Journal of Fnancal Economcs, Vol. 68, pp. 137-158. Chowdhry, Bhagwan and Ann E. Sherman (1996), The Wnner s Curse and Internatonal Methods of Allocatng Intal Publc Offerngs, Pacfc-Basn Fnance Journal, Vol. 4, pp. 15-30. Dawson, Steven M. (1987), Secondary Stock Market Performance of Intal Publc Offers, Hong Kong, Sngapore and Malaysa: 1978-1984, Journal of Busness Fnance Account, Vol. 14, pp. 65-76. 158

Fnn, Frank J. and Ron Hgham (1988), The Performance of Unseasoned New Equty Issues-Cum-Stock Exchange Lstngs n Australa, Journal of Bankng and Fnance, Vol. 12, pp. 333-351 Ibbotson, Roger G. and Jeffrey F. Jaffe (1975), Hot Issue Markets, Journal of Fnance, Vol. 30, pp. 1027-1042. Jelc, Ranko, Brahm Saadoun, and Rchard Brston (2001), Performance of Malaysan IPOs: Underwrters Reputaton and Management Earnngs Forecasts, Pacfc-Basn Fnance Journal, Vol. 9, pp. 457-486. Koh, Francs and Terry Walter (1989), A Drect Test of Rock s Model of The Prcng of Unseasoned Issues, Journal of Fnancal Economcs, Vol. 23, pp. 251-272. Leland, Hayne E. and Davd H. Pyle (1977), Informaton Asymmetres, Fnancal Structure, and Fnancal Intermedaton, Journal of Fnance, Vol. 32, pp. 371-387. Ln, Denns K. J., Lanfeng Kao and Anln Chen (2010), Wnner s Curse n Intal Publc Offerng subscrptons wth Investors Wthdrawal Optons, Asa-Pacfc Journal of Fnancal Studes, Vol. 39, pp. 3 27 Loughran, Tm, Jay R. Rtter and Krstan Rydqvst (1994), Intal Publc Offerngs: Internatonal Insghts, Pacfc-Basn Fnance Journal, Vol. 2, pp. 165-199. Relly, Frank K. and Kenneth Hatfeld (1969), Investor Experence wth New Stock Issues, Fnancal Analyst Journal, Vol. 25, pp. 73-80. Rock, Kevn (1986), Why New Issues are Underprced?, Journal of Fnancal Economcs, Vol. 15, pp. 187-212. Welch, Ivo (1992), Sequental Sales, Learnng, and Cascades, Journal of Fnance, Vol. 47, pp. 695-732. Yong, Othman (1991), Performance of New Issue Securtes n Malaysa, Malaysan Accountant (June), pp. 3-6. Yong, Othman and Zad Isa (2003), Intal Performance of New Issues of Shares n Malaysa, Appled Economcs, Vol. 35, pp. 919-930. Yong, Othman (2011), Wnner s Curse and Bandwagon Effect n Malaysan IPOs: Evdence from 2001-2009, Jurnal Pengurusan Vol. 32, pp. 21-26. 159