The Relation between Initial Returns and Audits by the Big Four Accounting Firms

Transcription

1 Appled Econocs and Fnance Vol. 4, No. 4; July 2017 ISSN E-ISSN Publshed by Redfae Publshng URL: The Relaton between Intal Returns and Audts by the Bg Four Accountng Frs Jen-Sn Lee 1, Yue L 2, Xn Hu 3, Q-An Lu 4 1 Departent of Fnance, I-Shou Unversty, Tawan 2 Insttute of Tawan Econoy, Econoc School, Nanka Unversty 3 Departent of Budget and Fnance, Changan Unversty 4 School of Econocs and Manageent, Guangx Noral Unversty, Guln, Guangx, , Chna Correspondence: Jen-Sn Lee, Departent of Fnance, I-Shou Unversty, Tawan. Receved: February 27, 2017 Accepted: Aprl 27, 2017 Avalable onlne: June 7, 2017 do: /aef.v URL: The research s fnanced by Collaboratve Innovaton Center for Chna Econoy. Abstract Ths paper anly explores the relaton between ntal returns and audts by the bg four accountng frs (the Bg Four) n Chna. The saple perod s fro January 2007 to Deceber 2012 (the new accountng standards n Chna s pleented after January 2007 for ntegratng wth the nternatonal standards), and selected 1,069 IPO frs lsted n the Shangha Stock Exchange and Shenzhen Stock Exchange n ths paper. Many prevous studes have proposed the Inforatonal Hypothess, whch states that the ntal returns of IPOs beng audted by the Bg Four are lower than those IPOs beng audted by other accountng frs. Oppostely, ths paper proposes the Snap-up Hypothess due to consder the IPOs n anland Chna are characterzed by three lows, : the low relablty of audts beng perfored by non-bg Four, low proporton of IPOs audts beng perfored by the Bg Four, and low ballng rato. These three lows features ndcate that the Snap-up Hypothess apples n the IPOs arket of anland Chna. In other words, the ntal returns of the IPOs beng audted by the Bg Four are hgher than those IPOs beng audted by other accountng frs due to the Bg Four have the superor reputatons. Ths paper further collects the tradng volues and the turnover rato on the frst day, and selects the Bg Four audted IPOs by snap-up tde. As above entoned, because the snap-up tde and rased stock prces on the frst-day lstng, nvestors ay purchase the shares when offerng and sell the on the frst-day lstng to obtan consderable profts. Keywords: four accountng frs, ntal return,inforatonal Hypothess, Snap-up Hypothess 1. Introducton Intal publc offerngs (hereafter IPOs) are a crtcal research topc n fnancal and econoc area. Prevous studes related to IPOs have focused anly on three aspects of IPOs, ncludng short-ter underprcng, hot-ssue arkets, and long-ter underperforance. Beatty (1989), Mchaely and Shaw (1995), Su and Flesher (1999), Chen and L(2004), Ch and Padgett (2005), Yang et al. (2007), X.Chang et al. (2008), Qu et al. (2013), and Hu and Feng(2013) have dscussed short-ter IPOs dscounts. In prevous studes, IPO short-ter underprcng has been found to be an nternatonal phenoenon. In the stock arkets of 39 countres, the average rates of ntal return (hereafter IR) of IPOs are postve, and IPOs underprcng n Chna s the hghest n the world (Tan, 2011). Because of the econoc developent of Chna and the rapd developent of the Chnese stock arket, nvestors and fr anagers have focused ncreasng attenton on the hgh underprcng n the Chnese stock arket 1. 1 On Nov. 30, 2013, Chna Securtes Regulatory Cosson ssued Opnons on Further Prootng the Refor of IPOs Syste, n whch the Cosson ponted out that t would strengthen restrctons aganst IPOs speculaton. 44

2 The IRs of IPOs are also referred to as the extent of underprcng. Ths paper explores the relaton between IRs and audts by the bg four accountng frs (the Bg Four) 2. Prevous studes, ncludng those by Rock (1986), Beatty and Rtter (1986), Teoh and Wong (1993), Beatty (1989), Mchaely and Shaw (1995), and Lee et al. (2011), have found that audts perfored by accountng frs wth superor reputatons ay provde nvestors wth ore relable fnancal nforaton as well as help to reduce asyetrc nforaton and nvestor rsk. Assocaton wth the Bg Four can sgnal to nvestors that nvestng n frs s safe, whch can lower nforaton asyetry, nvestor rsk, and IPO IRs. That s called Lower Inforaton Asyetry Hypothess n the lterature. Because the IRs are calculated by [(P 1 - P 0 ) / P 0 ], where P 1 and P 0 are the closng prce on the frst day of lstng and the offerng prce respectvely. It has to be notced that lower nforaton asyetry wll cause saller declne of offerng prce P 0 n the ssue arkets. Many eprcal studes have supported ths hypothess, ncludng the studes of Balvers et al. (1988), Beatty (1989), and Holland and Horton (1993), whch exaned IPOs n the USA and UK. As the table 1 of descrptve statstcs shows, the IRs of IPOs of the lsted copanes audted by Bg Four CPA frs are hgher, whch ay be caused by three lows characterstcs of Chna IPOs arket (the low relablty of audts perfored by accountng frs other than the Bg Four, low proporton of IPOs audts perfored by the Bg Four, and low ballng rato). Thus, ths study proposes Snap-up Hypothess. These three lows cause nvestors to snap up shares offered by frs audted by the Bg Four on the frst day of lstng and rase the closng prce on the frst day, causng the closng prce to be uch hgher than the offerng prce, ndcatng hgh IPOs underprcng. It s worth notng that Snap-up Hypothess wll rase the closng prce P 1 on the frst day n the secondary arkets. Therefore, the audt fr reputaton ay have the effects, whch are Lower Inforaton Asyetry Hypothess and Snap-up Hypothess, the forer s occurred when lowerng the declne of offerng prce n the ssue arket; and the latter s occurred when rasng the closng prceon the frst day n the secondary arket. The pror studes fnes the effect of Lower Inforaton Asyetry s donated n ost IPOs arkets, but ths study fnes the effect of Snap-up Hypothess s donated n Chna IPOs arket. Few eprcal studes have been conducted on ths hypothess. However, Chang et al. (2008) eprcally studed ths hypothess by exanng lsted frs n Australa. They found that the IRs on IPOs of frs assocatng wth hgh qualty audtors were generally hgher than those of frs assocatng wth coon audtors. These results support the Snap-up Hypothess. Snce 2004, scholars n anland Chna have perfored any studes on the relaton between IPOs underprcng and audtor reputaton for Chnese enterprses. However, the results of these studes have vared. Chen and L (2004) as well as Zeng and Sh (2011) found that audtng by accountng frs wth superor reputatons ay sgnfcantly lower the extent of underprcng. Moreover, Lu and Tan(2009) as well as Wang, Xn, and Yang(2009)found n ther studes on the Shangha and Shenzhen Stock Exchanges that IPOs of frs audted by the Bg Four were ore underprced. In other words, these studes support the Inforatonal Hypothess. However, the studes by Qu et al.(2013) on sall- and edu-szed enterprse(sme) boards concluded that IPOs of frs audted by accountng frs wth superor reputatons were not ore underprced. However, Hu and Feng(2013) found that hgh qualty audtors could ncrease the IRs of IPOs. Furtherore, Song,Tan, and Y (2014) dvded the IRs of IPOs nto two parts, underprcng fro ssue arkets 3 and overvaluaton fro secondary arkets, and found through eprcal study that the reputaton of underwrters postvely correlated wth IPO underprcng. The results of these eprcal studes thus support the Snap-up Hypothess. The theoretcal arguents and varous eprcal results regardng the relaton between IPOs underprcng and audtor reputaton for Chnese enterprses ake ths topc appealng for researchers. Although scholars dspute whch hypothess ore accurately reflects the relaton between IPOs underprcng and audtor reputaton, ths paper argues that n practce both hypotheses are true to an extent and produce correspondng effects. On ths bass, ths paper contends that the prary ssue s not whch of the two hypotheses s true but rather whch of the hypotheses s ore donant. Ths paper selects IPOs n anland Chna as the object of study and dscusses the relaton between IRs and audts by the Bg Four. 2 There were fve Bg Four jont ventures n Chnese Manland, PwC Zhongtan, DT Huayong, E&Y Huang, E&Y Dahua and KPMG Huazhen before However after the consoldaton of E&Y Huang and E&Y Dahua n Deceber 2008, the Bg Four jont ventures n Chnese Manland are PwC Zhongtan, KPMG Huazhen, DT Huayong and E&Y Huang, respectvely. 3 The anstrea s that the ntal return s also called IPOs underprcng, refers to the phenoenon that the close prce s hgher than the offerng prce on the frst day of lstng. Whle soe scholars thought that the ntal return and IPO underprcng are totally dfferent concepts. In ths paper we take the anstrea pont. 45

3 Ths paper uses 1,069 A-share IPOs n the Shangha and Shenzhen Stock Exchanges fro January 2007 to Deceber 2012 as eprcal saples to dscuss the relaton between IRs and audts by the Bg Four. Copared wth other ajor IPOs arkets, the IPOs arket n anland Chna s characterzed by three lows, whch s explaned as follows. Frst, copared wth other ajor IPOs arkets, the proporton of lsted frs audted by the Bg Four s low n the IPO arket n anland Chna. For exaple, aong the 1,069 IPOs frs n our study, only 40 were audted by the Bg Four, ndcatng a proporton of 3.74%. Ths proporton s extreely low copared wth other ajor stock arkets. For exaple, Lee et al. (2011) found that 383 of the 434 IPOs frs n Tawan were audted by the Bg Four, ndcatng a proporton of 88.25%. Addtonally, Frankel, Johnson, and Nelson (2002) found that 90.44% of lsted frs were audted by the Bg Four for IPOs n the USA. Furtherore, Fan and Wong (2005) found proportons of 88.32% n Indonesa, 74.86% n Malaysa, and 72.37% n Korea. Fnally, Chang et al. (2008) found a proporton of 46% n Australa, whch s lower than those n the aforeentoned countres but hgher than that n anland Chna. Second, copared wth those n other ajor IPO arkets, the fnancal reports of lsted frs n the IPO arket n anland Chna are less relable. The relablty of the fnancal reports of lsted frs n anland Chna has been wdely crtczed (Ca et al., 2008; Ln and Ln, 2009; Ferr and Lu,2010; the credt ratng and certfcaton center of the research nsttute of the Mnstry of Coerce, 2013, 2014). Accordng to the Apprasal Report on Fnancal Safety of Non-fnancal Lsted Frs n Chna n 2013(2014), jontly released by the credt ratng and certfcaton center of the research nsttute of the Mnstry of Coerce and the Econoy Web of Chna Revew (a onthly journal), 766 of 2,230(744 of 2,213) lsted frs were suspected offorgng fnancal stateents, representng 34.35%(34.32%) of the saples. Ths report also stated that false fnancal stateents of lsted frs n anland Chna are a serous proble. For exaple, 96 real estate frs n anland Chna (over 70%) are suspected of forgng fnancal records. Ca et al. (2008) asserted that fnancal stateents audted by the Bg Four are hghly relable, whereas those audted by other accountng frs are not suffcently relable. Thrd, the IPO arket of anland Chna tends to be a hot-ssue arket. The average ballng rato s very low (1.2%), and the nu ballng rato value s only 0.66%. In other words, only 6.6 of every 1,000 applcants succeed n subscrbng. Based on the prevous analyss, the characterstcs of the IPO arket n anland Chna can be suarzed as the three lows. These characterstcs are dfferent fro the features of the IPO arket n Australa studed by Chang et al. (2008). Would the effect and nfluence of the two aforesad hypotheses dffer n such a arket? Whch hypothess, the Inforatonal Hypothess or Snap-up Hypothess, would donate n the IPO arket of anland Chna? These ssues should be exaned through eprcal studes. Copared wth prevous studes on the effect of audtor reputaton n anland Chna, ths paper dffers n the followng ways. Frst, as stated prevously, ths paper attepts to deterne whch hypothess, the Inforatonal Hypothess or Snap-up Hypothess, donates n explanng the three lows of the IPO arket of anland Chna. Second, ths paper analyzes whether the donance of the Inforatonal Hypothess or Snap-up Hypothess under varous arket condtonshas dfferent effects. Ths paper perfors a robustness test based on the varous arket states defned by Lee et al. (2011, 2013). In addton to arket state, robustness tests are also perfored for factors such as offerng scale andexchange board dvson. Fnally, copared wth prevous studes, the saples chosen n ths paper are characterzed by the followng: (a). the saple perod covers any arket states; (b). a large saple sze s ncluded n ths study; (c). the saple perod occurs after the pleentaton of the new accountng standards. Specfcally, the saple perod n ths paper s May 2006 to August 2013 (88 onths n total), whch ncludes the bull (23 onths), bear (23 onths), and range-bound arkets (26 onths) 4. In addton, copared wth other studes, the saple perod n ths paper covers ore arket states and has alarger saplesze of IPO frs. For exaple, Gao (2010) studed only 217 IPO frs fro July 2006 to Aprl 2008, when the Shangha and Shenzhen Stock Exchange 300 Index was n a bull arket state. Moreover, Qu et al. (2013) studed only 223 IPO frs fro June 2006 to Deceber 2008, when the Shangha and Shenzhen Stock Exchange 300 Index was n a bull arket state. Addtonally, Hu and Feng (2013) studed only 517 nonfnancal frs fro July 2009 to Aprl 2011, when the Shangha and Shenzhen Stock Exchange 300 Index was n a bear arket state. New accountng standards were pleented n January 1, 2007 for anland Chna, and the focus of ths paper s the relaton between IRs and audts by the Bg Four. Ths paper analyzed saples after January 2007 to avod saple heterogenety caused 4 As stated above, the perod studed n ths paper s fro January 2007 to August 2013 (totally 88 onths). However, as t needs 8 onths to dstngush arket states, the actual perod studed s fro January 2007 to Deceber 2012 (totally 72 onths). 46

4 Appled Econocs and Fnance Vol. 4, No. 4; 2017 by varous accountng standards. Tan (2011) studed 1,377 IPO frs fro January 1992 to Deceber 2004; although ther saple was suffcently large, the saple perod dd not nclude the perod when the new accountng standards s were pleented. Song, Tan, and Y (2014) studedd frs fro Septeber 2006 to Deceber 2011; part of ther saple perod thus occurred before the pleentaton of the new accountng standards (such a perod ay cause sgnfcant heterogenety of saples and thus errorsn eprcal results). To ensure the robustness of the eprcal results, ths paper perfors the followng four robustness tests: eprcal e analyss usng arket-adjusted IR rather than raw IR as the dependent varable and analyses based on exchange board dvson, arket state dvson, and offerng scale dvson. The an fndngs are as follows: The IPOs n anland Chna are characterzed by the three lows, the low relablty of audts perfored by accountng frs other than the Bg Four, low proporton of IPOs audts beng perfored by the Bg Four, and low ballng rato. Furtherore, n anland Chna, the IPOs frs beng audted by hgh-credblty accountng frs are found to cause snap-up tde. Because of the prevously entoned snap-upp tde and rased stock prces on the frst day of lstng, nvestors ay purchase the shares of frs at the te of offerng and then sell the on the frst day of lstng to obtan consderable profts. Therefore, ths paper concludes that the Snap-up Hypothess s donant n explanng the features of IPOs n anland Chna. Ths concluson dffers fro the results of prevous studes (.e., the Inforatonal Hypothesss s donant n explanng the features of IPOs n anland Chna). However, H our paper does not deny the applcaton of the Inforatonal Hypothess to the IPOs arket of anland Chna, but we fnd that the Snap-up Hypothess s donant n explanng the features of ths arket. In addton, ths paper also attepts to perfor four robustness tests, whch show no sgnfcant dfferences n the eprcal results. The sectons of ths paper are arranged as follows: the second chapter ncludes a descrpton of the hypothess deducton and eprcal odel proposals; the thrd chapter ncludes nforaton on arket state dvson and data analyss; the fourth chapter dscusses the eprcal results as welll as the hypothess and robustness tests; andthe ffth chapter provdes the sgnfcant fndngs of ths paper and dscusses ther plcatons. 2. Hypothesss Deducton and Eprcal Models 2.1 Market State and Intal Return Lee, Jang, and Indro (2002) found that arket states could affect nvestor sentent and behavor. Specfcally, n bull arkets, nvestor sentents are postve, and nvestors expect hgh excess returns n the future. By contrast, n bear arkets, nvestors are non-postve and expect low excess returns n the future. Ljungqvst, Nanda, and Snghh (2001) found that IPO IRs vared aong arket states (cold- or hot-ssue). As prevously entoned, Lee et al. (2011)) studed the factors nfluencng the IRs of IPO frs n Tawan and found that the audtor reputaton effect was sgnfcant n range-bound arkets but not n bear or bull arkets. Furtherore, they found that arket oentu was hgher n bull arkets than range-bound and bear arkets. Wang and Zhang (2005) asserted that n the anland Chnese share arket, the general arket trend also nfluenced nvestor behavor. Therefore, n varous arket states, nvestors have varous nvestent sentents, and the effect of varous factors on IRs ay also vary. In ths paper, the pact of audtor reputaton on IRs n anland Chna s studed n varous arket states to test the robustness of the eprcal results. 2.2 Hypothess Deducton Prevous studes, ncludng those by Rock(1986), Beatty and Rtter (1986), Teoh and Wong (1993), Beattyy (1989), Mchaely and Shaw (1995), and Lee et al. (2011), have found that audts perfored by accountng frs wth superor reputatons ay provde nvestors wth ore relablee fnancal nforaton as well as help to reduce asyetrc nforaton and nvestor rsk. Assocaton wth the Bg Four can sgnal to nvestors that nvestng n frs s safe, whch can lower nforaton asyetry, nvestor rsk, and IPO IRs. Ths s referred to as the Inforatonal Hypothess or the audtor reputaton effect. Ths hypothess s suarzed n the flow chart shown n Fgure 1. Fgure 1. Flow chart of nforatonal hypotheses If the flow chart n Fgure 1 s supported, then the correspondng hypothess s as follows. Hypothess 1:The Inforatonal Hypothess, n other words, the hypothess that the IRs of IPOs audted byy the Bg Four s lower than that of IPOs audted by other accountng frs. However, as dscussed prevously, audts of IPOs n anland Chna are dstnct fro audts n ature arkets. Ca et 47

5 al.(2008) asserted that the relablty of fnancal reports of lsted frs n anland Chna has been crtczed. The credt ratng and certfcaton center of the research nsttute of the Mnstry of Coerce (2013, 2014) stated that false fnancal stateents of lsted frs n anland Chna are a serous proble. Ths paper also fnds that, n anland Chna, the proporton of lsted frs audted by the Bg Four s very low. Copared wth other ajor IPO arkets, the characterstcs of the IPO arket n anland Chna can be suarzed as three lows, the low relablty of audts perfored by accountng frs other than the Bg Four, low proporton of IPO audts perfored by the Bg Four (the IPO arket share of the Bg Four n anland Chna s only 3.74%), and low ballng rato (the average ballng rato only 1.2%). These three lows cause nvestors to snap up shares offered by frs audted by the Bg Four on the frst day of lstng and rase the closng prce on the frst day, causng the closng prce to be uch hgher than the offerng prce, ndcatng hgh IPO underprcng. Ths s referred to as the Snap-up Hypothess. Ths hypothess s suarzed n the flow chart shown n Fgure 2. Hgher relablty of fnancal reports Low ballng rato Snap-up effect on the frst of lstng Heghten the close prce on the frst day Hgher ntal return Fgure 2. Flow chart of snap up hypotheses If the flow chart n Fgure 2 s supported, then the correspondng hypothess s as follows. Hypothess 2:The Snap-up Hypothess, n other words, the hypothess that the IRs of IPOs audted by the Bg Four are hgher than those of IPOs audted by other accountng frs. 2.3 Eprcal Model Ths paper anly studes the pact of audts by the Bg Four on IRs of IPOs, wth reference to the eprcal odel desgned by Lee et al. (2011). The orgnal odel of Lee et al. (2011) contaned the varables of nsder stockholdng and block holder proportons. However, because no data are avalable on these two varables for anland Chna, these two varables are excluded fro our study. In addton, consderng the hgh proporton of ndvdual nvestors n anland Chna, whch ay be easly affected by arket sentent, a new varable, the average IR (AIR) of ndvdual shares, s added n ths paper. Our odel s as follows: IR + β 7 = α LN + + β β 8 1 AR SIZE 4 + β + e 2 R + β 3 AIR + β 4 M + β 5 LT + β 6 BR In the odel, IR s easured by raw IR (RIR) and arket-adjusted IR (MAIR) n the robustness test; represents varous arket states, ncludng bull, bear, and range-bound; AR4 s a duy varable representng audts by the Bg Four; R, AIR, M, LT, BR, and SIZE represent arket rsk, average IR of ndvdual shares, arket oentu, lstng days (days between the offerng day and lstng days), ballng rato, and sze of ssuance, respectvely. The data used here s cross-secton data. 2.4 Varables Intal Return, Raw Intal Return, and Market Adjusted Intal Return In the lterature, IR s anly expressed as raw IR (RIR) and arket adjusted IR (MAIR), respectvely. RIR represents the dfference between the closng prce on the frst day of lstng and the offerng prce, whereas MAIR represents the return after deductng the arket ndex return.the forulas for both are as follows: (1) RIR = [ (P t - P 0 ) / P 0 ] 100% (2) In ths equaton, P t and P 0 are the closng prce on the frst day of lstng and the offerng prce, respectvely. 48

6 MAIR = [ RIR - ( I t - I 0 ) / I 0 ] 100% (3) In ths equaton, I t and I 0 are the arket ndex on the frst day of lstng and the arket ndex on the offerng day, respectvely. The duy varable representng audts by the Bg Four (AR4): In the lterature, two an hypotheses have been proposed concernng the relaton between IRs and audts by the Bg Four. The two an hypotheses are the Inforatonal Hypothess and the Snap-up Hypothess.The duy varable AR4 s set to 1 for IPOs audted by the Bg Four and 0 for IPOs audted by other accountng frs Market Rsk Market rsk reflects future uncertanty. Derren and Woack (2003) as well as Lee et al. (2011) used the standard devaton of daly returns of the arket ndex between the offerng day and lstng day to easure arket rsk, fndng that the arket rsk was sgnfcantly and postvely correlated wth IRs. Wth reference to the arrangeent of Lee et al. (2011), ths paper uses the standard devaton of the daly returns of the arket ndex between the offerng day and lstng day to easure arket rsk, and the forula s as follows: R T t = 0 = ( r r) t D 2 (4) In Forula (4), R represents the arket rsk of fr, and t = 0, t = T, and r t represent the returns of arket ndex on the offerng day, the lsted day, and day t, respectvely. r represents the average value of the arket returns between the offerng day and lstng day, and D represents the nuber of tradng days fro the offerng day to the lstng day Average Intal Return of Shares The average IR of shares represents the average value of IRs on new shares durng a gven perod and reflects arket sentent and nvestor expectaton of IRs n the arket. In ths paper, ths varable s used as a reference for the pact of IRs of new shares. AIR = t = T = N t = 0 = 0 In Forula (5), IR represents the IRs of fr, and N s the total nuber of lsted frs between t=0 and t=t Market Moentu In the lterature, such as the studes of Kunz and Aggarwal (1994), Loughran and Rtter (2002), Lyn and Zychowcz (2003), Derren and Woack (2003), and Lee et al. (2011) a ore prosperous arket before an IPO has been found to ncrease the IRs. Lee et al. (2011) found that varous arket states have dfferent effects on arket oentu. In ths paper, arket oentu s calculated usng the ethod developed by Lee et al. (2011).The forula s as follows: IR N (5) T M = ( 1+ rt ) 1 t= 0 (6) In Forula (6), M represents the arket oentu of fr, and r t represents the arket ndex returns on day t 5. Forula (6) shows that when the arket ndex trends upward, the arket oentu (M) s postve, and when the arket ndex trends downward, the arket oentu s negatve. The data of the paper s cross-sectonal data and not panel data. Market return s not the sae for all frs Lstng Days, Ballng Rato, and the Sze of Issuance Lstng Days refer to the nuber of days between the offerng day and lstng day. The ballng rato reflects the hotness of new shares aong nvestors; a lower ballng rato ndcates hgher hotness. The sze of ssuance represents the 5 The forula of r s as follows: r = ( P -P -1 ) / P -1, there P -1 and P represents the arket ndex on the day -1 and day, respectvely. 49

7 offerng scale of the lsted fr and reflects the scale of fundng through lstng. In ths paper, wth reference to the arrangeent of Lee et al. (2011), the natural logarth of the sze of ssuance s calculated to tgate the proble of scale of varous ndependent varables. 3. Data and the Dstrbuton ofmarket States 3.1 Source of Data On February 15, 2006, the Mnstry of Fnance held a news conference on the release of accountng and audtng standards n Bejng, ncludng 39 accountng standards for enterprses and 48 audtng standards for certfed publc accountants. The accountng standards for enterprses have been pleented by lsted frs snce January 1, 2007 to consoldate the accountng syste nanland Chna wth the nternatonal systes. The audtng standards for certfed publc accountants have been pleented snce January 1, Consderng the change of these standards, ths paper uses frs lsted fro January 2007 to Deceber 2012 as saples. The data on the lsted frs are sourced fro the Tawan Econoc Journal and RESSET database. The Shangha and Shenzhen 300 Index s sourced fro the webste of the Shangha Stock Exchange. The rankng nforaton of accountng frs s sourced fro the webste of the Chnese Insttute of Certfed Publc Accountants. The saples n ths paper cover 12 ndustres. 6 After excludng a few saples wth ncoplete data, a total of 1,069 IPO frs are selected, ncludng frs lsted n the A-share arkets of Shangha and Shenzhen Stock Exchanges. 7 Therefore, 119 frs are lsted n the Shangha Stock Exchange, 950 are lsted n the Shenzhen Stock Exchange, the IPO of 201 frs was durng a bull-arket perod, and the IPO of 868 frs was durng a bear-arket perod. 3.2 Descrptve Statstcs and Correlaton Coeffcents Table 1 provdes a lst of statstcal values of an varables and shows that the IRs of IPOs n anland Chna have the followng four characterstcs: Frst, the value of IRs of IPOs n anland Chna s hgh: The RIR s up to 59.40%, the MAIR s up to 59.55%, and a hgh return of nearly 60% can be obtaned under average lstng days (LT) of 12. Second, the IRs of IPOs nanland Chna see to be trendng downward: the saple perod n ths paper s fro January 2007 to Deceber 2012 (ncludng A-share frs lsted n the Shangha and Shenzhen Stock Exchanges), and the perods studed by Chan et al. (2004) and Chen et al. (2004) are fro 1993 to 1998 and 1992 to 2000 (A-share arket), respectvely. The average IRs of the perods studed by Chan et al. (2004) and Chen et al. (2004)are 178%. The IRs of IPOs n anland Chna see to be trendng thus downward. Thrd, soe negatve IRs are found, but the rato of these returns s not hgh. Aong the 1,069 lsted frs studed, 144 frs had negatve RIRs, and 136 had negatve MAIRs, yeldng 13.47% and 12.72%, respectvely. These two ratos are not hgh, and the average values of both the RIR and MAIR are nearly 60%, ndcatng that IPOs nanland Chna are worthy of nvestor subscrpton durng the offerng phase. Fnally, the RIR and MAIR are slar, ncludng ther average values, edans, axu values, nu values, and standard devatons, and no statstcal dfference s found between varous RIR values and correspondng MAIR values. The average IR of IPOs audted by the Bg Four (AR4) s 79.98%, sgnfcantly hgher than the RIR of 59.4%, ndcatng the possblty of ncreasng IRs by assocaton wth the Bg Four. In other words, Hypothess 2 sees to besupported, and the Snap-up Hypothess s donant n the IPO arket of anland Chna. Moreover, n ters of the ballng rato (BR), the followng two features are found: Frst, the IPO arket of anland Chna tends to be a hot-ssue arket. The average BR s very low (1.2%), and the nu BR value s only 0.66%. In other words, only 6.6 of every 1,000 applcants can succeed n subscrbng. Second, the possblty of unsalable IPO shares s low n anland Chna. The axu BR n the saples s 65.52%, less than 1, ndcatng that unsalable IPOs resultng fro a lack of subscrbers n Chna s unlkely. In ths paper, the correlaton coeffcents aong ndependent varables and varance nflaton factor (VIF) are evaluated to avod ultcollnearty aong ndependent varables. Furtherore, the Pearson correlaton coeffcents aong the ndependent varables are all uch lower than 0.7 and nsgnfcant, and the VIF values are all uch lower than 10. Multcollnearty s thus not a proble n ths paper. Table 2 shows the correlaton coeffcents aong the ndependent varables. 6 No fr n the fnancal or nsurance ndustres was lsted fro January 2007 to Deceber No new B-share IPOs n the Shangha and Shenzhen Stock Exchanges are found fro January 2007 to Deceber

8 Table 1. Descrptve statstcs of varables RIR (%) MRIR (%) RIR4 (%) R (%) AIR (%) M (%) BR (%) LT (days) Mean Medan Max Mn S.D Note: RIR s rawir; MAIR s arket adjusted IR; RIR4 represents rawir of those fnancal reports audted by the Bg ( ) 2 rt r R t = = 0 Four; R represents arket rsk, the forula s as follows: D, R represents the arket rsk of fr, and t = 0, t = T, and r t respectvely represents the return of arket ndex on the offerng day, the lsted day and day t, r represents the average value of the arket return between offerng day and lsted day, and D represents the nuber of AIR = IR N tradng days fro the offerng day to the lsted day; AIR represents the average of IR, t = 0 = 0, IR represents IR of fr, and N s the total nuber of lsted frs between t=0 and t=t; M represents arket oentu, and the T M = ( 1 + rt ) 1 0 forula s as follow: t = ;BR s ballng rato; LTrefers to the nuber of days fro the offerng day to the lsted day; SIZE takes the natural logarth of the offerng sze. Table 2. Correlaton Coeffcent AR4 R AIR M BR LT SIZE AR4 1 R AIR M BR LT SIZE Note: In Table 2, AR4R, AIR,M,BR,LT and SIZErepresents duy varable representng audt by the Bg Four, arket rsk, average IR of ndvdual share, arket oentu, ballng rato, SIZE takes the natural logarth of the offerng sze. 3.3 Market State Dvson Method Pagan and Sossounov (2003) dvded arket states nto bull and bear arkets. Lee et al. (2011) and Lee et al. (2013) further developed ths dvson ethod and added range-bound arkets. In ths paper, the odfed dvson ethod s used, and the arket states are dvded nto the bull, bear, and range-bound arkets. The process s as follows. Step 1: Deterne Peak and Trough Peak and trough refers to the hghest and lowest pont durng a gven perod. Accordng to a study by Pagan and Sossounov, f the condtons n Forula (7) are et, then poston P t s referred to as peak: T t = T = N SIZE [ Pt 8,..., Pt 1 < Pt > Pt + 1,..., Pt + 8 ] (7) In the prevous forula, P t represents the arket ndex correspondng to perod t. Forula (7) represents arket ndex P t, whch should be hgher than the arket ndexes n the prevous eght consecutve onths and followng eght consecutve onths. If P t eets these condtons, the poston of P t s referred to as a peak. Slarly, f the condtons n 51

9 Forula (8) are et, then poston P t s called trough: [ Pt 8,..., Pt 1 > Pt < Pt + 1,..., Pt + 8 ] (8) In the prevous forula, P t represents the arket ndex of onth t. Forula (6) represents arket ndex P t, whch should be lower than the arket ndexes n the prevous eght consecutve onths and the arket ndexes n the followng eght consecutve onths. If P t eets these condtons, the poston of P t s referred to as a trough. For the purpose of deternng a peak and trough, the prevous and followng eght consecutve onths should be checked based on the asserton by Pagan and Sossounov (2003) that the cycle of the US stock arket should exceed 15 onths. Step 2: Deterne Bull and Bear Markets A bear arket refers to a contnuous downward arket trend of the arket wtha suffcent accuulatve drop and downward duraton. A bear arket occurs when the followng three condtons are et: Frst, the arket transtons fro peak to trough. Second, the accuulatve drop s ore than 20%n stock ndex values. Thrd, durng the perod fro peak to trough s not shorter than 4 onths. When the arket reaches the trough, the downward trend ends, and the arket rses. The 4-onth duratonal requreent was deterned by Pagan and Sossounov (2003), Lee et al. (2011), and Lee et al. (2013) wth reference to Halton (1919) and Dow Theory (developed by Charles Dow), and verfed by eprcal analyss of the US stock arket. Edwards et al. (2003) also used ths 4-onth duratonal requreent. Slarly, a bull arket refers to a contnuous upward trend of the arket, wth a suffcent accuulatve rse and upward duraton. A bull arket occurs when the followng three condtons are et: Frst, the arket transtons fro trough to peak. Second, the accuulatve rse s ore than 20%n stock ndex values. Thrd, the perod fro trough to peak s not shorter than 4 onths. When the arket reaches to the peak, the upward trend ends, and the arket drops. The prevous two steps were developed by Pagan and Sossounov (2003), wth two ajor odfcatons ade by Lee et al. (2011) and Lee et al. (2013): Frst, a range-bound arket was added; second, the classfcaton of a bull, bear and neutral arket are robust wth a wndow length of 8 to 10 onths. Step 3: Deterne Range-Bound Markets After deternng bull and bear arkets, the reanng parts that cannot be defned as bull or bear arkets are called range-bound arkets n ths paper. Katsenelson (2007) entoned that when stock prces were narrow n valuaton, the arket was range-bound. Usng Katsenelson s defnton, ths perod s defned as a range-bound arket n ths paper. 3.4 Results of Market State Dvson The saple perod n ths paper s fro January 2007 to Deceber 2012, and the saples are frs lsted n the A-share arket of the Shangha and Shenzhen Stock Exchange. The Shangha and Shenzhen 300 Index s thus used as a substtute for the arket ndex. 8 Accordng to the studes by Lee et al. (2011) and Lee et al. (2013), 8 onths s requred before and after the study perod; our arket ndex selecton perod s thus fro May 2006 to August 2013, coverng 88 onths. The bull arket spans ore than three ntervals and covers 31 onths, consttutng 35.62% of total saples. Thebear arket spans ore than three ntervals and covers 48 onths, consttutng 55.17% of total saples. The range-bound arket spans ore than one nterval and covers 8 onths, consttutng 9.19% of total saples. In the ntervals used n the eprcal odel, the bull arket covers 23 onths, consttutng 32% ofthe total saples, wth an average arket return of 10%. The bear arket covers 48 onths, consttutng 68% of the total saples, wth an average arket return of -5%.The eprcal odel contans no range-bound arket. The arket state dvson s shown n Fgure 2, and the nforaton on the arket state dvson s suarzed n Table The Shangha and Shenzhen 300 Index s a share arket ndex calculated usng 300 A-shares n the Shangha and Shenzhen Stock Exchanges. These saples cover approxately 60% of the arket value of these two arkets and are hghly representatve. 9 Here, the average return of the arket under varous arket states s equal to the average value of the arket returns of all onths covered by such states. 52

10 Fgure 3. Result of arket states dvson Note: Fg. 3 s a bar graph of arket states, the horzontal axs represents te (onths), and the vertcal axs represents the arket ndex (CSI 300 Index).In order to dvde the arket states, Fg.3shows the change of the arket ndex fro May 2006 to August 2013, but the arket ndex used n the eprcal odel s fro January 2007 to Deceber Thus the fgure dvdes eprcal saple and full saple wth lne of dashes. Wthn two dotted lnes are the eprcal part, whte s bearsh arket states, dark state s bullsh arket states, lght gray s the range-bound arket state. 4. Eprcal Results and Robustness Tests The eprcal results of ths paper are suarzed n Table 3, and Table 3 shows the eprcal results of the effect of Bg Four audtng (AR4). The results show that Hypothess 2 s accepted. In other words, the Snap-up Hypothess s donant n the IPO arket of anland Chna. Therefore, the IRs of IPOs beng audted by the Bg Four are hgher than those IPOs beng audted by other accountng frs. The result s as follows. Table 3 shows that the coeffcent of AR4 s 0.63, whch s sgnfcant at the 1% level. The coeffcent of AR4 s sgnfcantly postve, whch ndcates that audts by the Bg Four can sgnfcantly ncrease the IRs of the frs audted, showng that hypothess 2 s accepted. In other words, accordng to the saples n ths paper, we fnd that the Snap-up Hypothess s donant, ndcatng that the hypothess that the IRs of IPOs audted by the Bg Four are hgher than those of IPOs audted by other accountng frs s accepted. The prevous fndngs support that the Snap-up Hypothess s donant n the anland Chnese IPO arket, ndcatng that IRs ncrease f IPOs are audted by the Bg Four, whch have superor reputatons copared wth local accountng frs. Ths fndng contrasts wth the results of studes conducted n other IPO arkets, whch have ndcated that the Inforatonal Hypothess s donant, eanng that the IRs decrease f IPOs are audted by the Bg Four. Studes by Teoh and Wong (1993), Beatty (1989), Mchaely and Shaw (1995), and Lee et al. (2011) have reached ths concluson. Table 3. Eprcal Results Dependent Varable: RIR 4.32 C (10.04) c 0.63 AR4 (6.90) c 9.27 RISK (3.49) c 0.67 AIR (25.80) c 1.53 M (5.44) c BR (-1.86) a 0.01 LT (2.74) c SIZE (-10.34) c Saples 1069 R Note: 1. The t value. a, b, c represent the sgnfcance n 10%, 5%, and 1% levels, respectvely. 2. AR4, R,AIR,M,BR,Lt and SIZErepresents duy varable representng audt by the Bg Four, arket rsk, average IR of ndvdual share, 53

11 arket oentu, ballng rato, SIZE takes the natural logarth of the offerng sze. 3. The t-rato s reported n the parentheses. The an cause of the donance of the Snap-up Hypothess n the anland Chnese IPO arket ay be as follows: The audtng qualty of local accountng frs n Chna s low (Ln and Ln, 2009; credt ratng and certfcaton center of the research nsttute of the Mnstry of Coerce, 2013); coparatvely, the credblty of the Bg Four n Chna s hgh; the Bg Four perfor a low proporton of the IPO audts n Chna (arket share of the Bg Four n IPOs s only 3.74%); fnally, the ballng rato of IPOs n Chna s very low (only 1.20% n our study saples), whch causes nvestors to snap up IPO shares fro frs audted by the Bg Four on the frst day of lstng. Ths leads to ncreased closng prces on the frst day that greatly exceed offerng prces, causng hgh IPO underprcng. In ths paper, the tradng volues and turnover rato of the saples on the frst day of lstng are further collected, and the IPO frs audted by hghly credble accountng frs are found to cause snap-up tde. The average tradng volues (turnover rato) of IPO frs audted by the Bg Four n bull and bear arkets are 311,590,380 shares (15.01%) and 27,942,517 shares (7.53%), respectvely. These greatly exceed the average volues of IPO frs audted by other accountng frs n bull and bear arkets (19,854,475 shares (5.22%) and 17,692,299 shares (3.13%), respectvely). The forer average tradng volues are (2.88) and 1.58 (2.41) tes the latter average volues, respectvely. Ths ndcates that the occurrence of the snap-up tde n anland Chna. Because of the prevously entoned snap-up tde and rased stock prces on the frst day of lstng, nvestors ay actvely subscrbe shares of frs to be lsted at the te of offerng and then sell the on the frst day of lstng to obtan consderable profts. 4.1 Robustness Tests To test the robustness of the prevous eprcal results, n ths secton the followng robustness tests are perfored: eprcal analyss usng arket adjusted IR nstead of raw IR as the dependent varable, analyss based on stock arket dvson, analyss based on arket states dvson, and analyss based on offerng scale dvson. The test results are respectvely shown n Table 4, Table 5, Table 6 and Table 7. It can be seen that the test results are close to the eprcal results provded n prevous secton, wthout sgnfcant structural change. Thus, the concluson of ths study s robust. Table 4. Eprcal results usng arket adjusted IR as dependent varable Dependent varable MRIR 4.32 C (10.04) c 0.63 AR4 (6.90) c 9.27 RISK (3.49) c 0.67 AIR (25.80) c 0.53 M (1.88) a BR (-1.86) a 0.01 LT (2.74) c SIZE (-10.34) c Saples 1069 R Note: 1. Thstable replace raw IR wth arket adjusted IR as dependent varable on eprcal analyss 2. Defnton of the relevant explanatory varables refers the nstructons of Table The t-rato s reported n the parentheses. In Table 4, arket adjusted IR nstead of raw IR s used as the dependent varable, and the results show that the coeffcent of AR4 s 0.63, sgnfcant at the 1 per cent level. Ths result shows that hypothess 2 s accepted. Thus, when arket adjusted IR s used as IR, the eprcal results also supports the hypothess 2, just lke the analyss usng raw IR as the dependent varable. 54

12 Table 5. Eprcal results of dstngushng dfferent Lsted Exchange Dependent varable: RIR Market states Shangha Stock Exchange Shengzhen Stock Exchange C (2.84) c (12.00) c AR (2.41) b (4.86) c RISK (2.69) c (2.04) b AIR (2.46) b (25.63) c M (3.57) c (3.97) c BR (-2.88) c (-0.67) LT (-0.85) (3.24) c SIZE (-2.59) c (-12.28) c Saples R Note: 1. Thstable dvdes saples lsted n Shangha Stock Exchange and ShenzhenStock Exchangeon eprcal analyss.2. Defnton of the relevant ndependent varables refers to the nstructons of Table The t-rato s reported n the parentheses. In Table 5, the saples are dvded nto frs lsted n Shangha Stock Exchange and frs lsted n Shenzhen Stock Exchange, and the results show that both the saples fro Shangha Stock Exchange and the saples fro Shenzhen Stock Exchange support the Hypothess 2, just lke the results of Table 3. Ths ndcates that the eprcal results of ths paper are ndependent of the stock exchange used. Table 6. Eprcal results of dstngushng dfferent Market States Dependent varable: RIR Market States C AR4 RISK AIR M BR LT SIZE Bull Market Bear Market (3.13) c (9.18) c (2.03) b (7.76) c (1.53) (3.16) c (12.34) c (16.57) c (0.51) (0.38) (-3.29) c (-1.72) a (-0.80) (3.75) c (-2.81) c (-9.49) c Saples R Note: 1. Ths table dvdes the saples lsted n bull arket and bear arket on eprcal analyss. 2. The way we dvdng bull arket and bear arket n ths paper refers to [3.3 Market states dvson Method]. 3. Defnton of the relevant ndependent varables refers to the nstructons of Table The t-rato s reported n the parentheses. In Table 6, the saples are dvded nto frs lsted n bull arket and bear arket, and the results show that both coeffcents are postve and sgnfcant, supportng the Hypothess 2. Ths ndcates that the eprcal results of ths paper are ndependent of the arket states. In Table 7, the saples are dvded nto frs of dfferent scales, and the results show that all the three coeffcents are postve and sgnfcant, supportng thehypothess 2. Ths ndcates that the eprcal results of ths paper are ndependent of the scale of the lsted frs. 55

13 Table 7. Eprcal results of dstngushng dfferent offerng sze Dependent varable: RIR Offerng Sze Bg Scale Medu Scale Sall Scale C (-0.02) (3.52) c (5.28) c AR (2.59) c (2.63) c (5.73) c RISK (4.01) c (0.76) (0.03) AIR (11.71) c (15.31) c (13.04) c M (5.95) c (4.07) c (1.25) BR (-2.02) b (-2.66) c (-2.33) b LT (1.35) (2.00) b (2.07) b SIZE (-0.31) (-3.50) c (-5.21) c Saples R Note: 1. Ths table dvdes the saples to bg scale, edu scale and sall scale on eprcal analyss. 2. In accordance wth the offerng sze, we regard the bggest 30%offerng sze of all lsted frs as the bg scale, and ddle 40% as the edu scale, sallest 30%as the sall scale. 3. Defnton of the relevant ndependent varables refers the nstructons of Table The t-rato s reported n the parentheses. 5. Conclusons and Suggestons Ths paper anly exploresthe relaton between IRs and audts by the Bg Four. The saple perod s fro January 2007 to Deceber 2012 (the new accountng standards n Chna s pleented n January 2007 for ntegratng wth the nternatonal standards). The objects of ths study are 1,069 IPO frs lsted n the Shangha and Shenzhen Stock Exchanges. Many prevous studes have proposed the Inforatonal Hypothess, whch states that the IRs of ntal publc offerngs (IPOs) beng audted by the Bg Four are lower than those IPOs beng audted by other accountng frs. Oppostely, ths paper proposes the Snap-up Hypothess due to consder the IPOs n anland Chna are characterzed by three lows, : the low relablty of audts beng perfored by non-bg Four, low proporton of IPOs audts beng perfored by the Bg Four, and low ballng rato. These three lows features ndcate that the Snap-up Hypothess apples n the IPOs arket of anland Chna. In other words, the IRs of the IPOs beng audted by the Bg Four are hgher than those IPOs beng audted by other accountng frs due to the Bg Four have the superor reputatons. As above entoned, because the snap-up tde and rased stock prces on the frst-day lstng, nvestors ay purchase the shares when offerng and sell the on the frst-day lstng to obtan consderable profts. Ths paper further collects the tradng volues and turnover rato on the frst day, and selects the Bg Four audted IPOs by snap-up tde. For exaple, n the bullsh arket, the average volues of IPOsbeng audted by the Bg Four are 311,590,380 shares, whch s tes as large as others (19,854,475 shares). To deterne f the Snap-up Hypothess s supportedn the IPO arket of anland Chna, several robustness tests are perfored: eprcal analyss usng arket adjusted IR rather than raw IR as the dependent varable as well as analyses based on exchange board dvson, arket state dvson, and offerng scale dvson. There s no sgnfcant structural change occurs on the test results. The concluson regardng the donance of the Snap-up Hypothess s robust thus. As above entoned, because the snap-up tde and rased stock prces on the frst-day lstng, nvestors ay purchase the shares when offerng and sell the on the frst-day lstng to obtan consderable profts. References Balvers, R. J., McDonald, B., & Mller, R. (1988). Underprcngof New Issues and the Choce of Audtors as a SgnalofInvestent Banker Reputaton. The Accountng Revew, 63, Beatty, R. P. (1989). Audtor reputaton and the prcng of ntal publc offerngs. Accountng Revew, 64 (4), Beatty, R. P., & Rtter, J. (1986). Investent bankng, reputaton, and the underprcng ofntal publc offerngs, Journal of Fnancal Econocs,15, Ca, Y. Q., N, B. H., Huang, M. Z., & Lao, L. M. (2008). Audt fee and prcng n Chnese Manland: a eprcal study on the lsted copanes n Shenzhen Stock Exchange. Journal of Soochow Econoc and Busness, 63,

14 Chan, K., Wang, J., & We, K. C. J. (2004). Underprcng and Long-Ter Perforance of IPOs n Chna, Journal of Corporate Fnance, 10, Chang, X., Gygax, A. F., Oon, E., & Zhang, H. F. (2008). Audt qualty, audtor copensaton and ntal publc offerng underprcng. Accountng and Fnance, 48, Chen, G., Frth, M., & K, J. B. (2004). IPO underprcng n Chna s new stock arkets. Journal of Multnatonal Fnancal Manageent, 14(3), Chen, H. M., & L, D. (2004). A postve research on the hypothess of the certfed accountant off ces reputaton of short-run underprcng on the stocks of IPO n Chna. Audt and Econoy Research, 1, 26-29(n Chnese). Ch, J., & Padgett, C. (2005). The perforance and long-run characterstcs of the Chnese IPO arket.pacfc Econoc Revew, 10(4), Derren, F., & Woack, K. L. (2003). Auctons vs. bookbuldng and the control of underprcng n hot IPO arkets. Revew of Fnancal Studes, 16(1), Edwards, S., Bscarr, J. G., & Pérez de Graca, F. (2003). Stock arket cycles, fnancal lberalzaton and volatlty. Journal of Internatonal Money and Fnance, 22(7), Fan, J. P. H., & Wong, T. J. (2005). Do External Audtors Perfora Corporate Governance Role n Eergng Markets? Evdence fro East Asa. Journal of Accountng Research, 43, Ferr, G., & Lu, L. G. (2010). Honor thy credtorsbefore thyshareholders: are the profts of Chnese state=owned enterprses real? Asan Econoc Papers, 9(3), Frankel, R., Johnson, M., & Nelson, K. (2002). The relaton between audtors fees for nonaudt servces and earnngs anageent. The Accountng Revew, 77, Gao, Y. (2010). What coprses IPO ntal returns: Evdence fro the Chnese arket.pacfc-basn Fnance Journal, 18(1), Halton, W. P. (1919). Stock arket analyss,wall StreetJournal, 9 (August) (Reprnted n Rhea, R. (1932).The Dow Theory, Barron s, New York, pp ). Holland, K. M., & Horton, J. G. (1993). Intal PublcOfferngs on the Unlsted Securtes Market: The Ipact of ProfessonalADvsers. Accountng and Busness Research, 24(93), Hu, D., & Feng, Q. G. (2013). Inforaton envronent, audt qualty and IPO underprcng: based on the A-share IPO frs durng 2009~2011.Accountng Research, 2013, 2, Katsenelson, V. N. (2007). Actve Value Investng: Makng Money n Range Bound Markets, John Wley and Sons, Inc Publshng, New York, NY. Kunz, R. M., & Aggarwal, R. (1994). Why ntal publc offerngs are underprced: Evdence fro Swtzerland. Journal of Bankng & Fnance, 18(4), Lee, J. S., Kuo, C. T., & Yen, P. H. (2011). Market States and Intal Returns: Evdence fro Tawan IPOs.Eergng Markets Fnance and Trade, 47(2), Lee, J. S., Yen, P. H., & Chan, K. C. (2013). Market States and Dsposton Effect: Evdence fro Tawan Mutual Fund Investors. Appled Econocs, 45(10), Lee, W. Y., Jang, C. X., & Indro, D. C. (2002). Stock arket volatlty, excess returns, and the role of nvestor sentent. Journal of Bankng & Fnance, 26(12), Ln, C. J., & Ln, X. L. (2009). Does audtor reputaton affect equty prcng: the evdence fro IPO arketzaton? Accountng Revew, 49, Lu, W., & Tan, Z. Y. (2009). Eprcal Study on sgnal transsson of audtor reputaton. Frends of Accountng, 11, Ljungqvst, A., Nanda, V., & Sngh, R. (2001). Hot arkets, nvestor sentent, and IPO prcng. Journal of Busness, 79(4), Loughran, T., & Rtter, J. R. (2002). Why don't ssuers get upset about leavng oney on the table n IPOs? Revew of Fnancal Studes, 15(2), Lyn, E. O., & Zychowcz, E. J. (2003).The perforance of new equty offerngs n Hungary and Poland. Global 57