UNDERPRICING AND EX ANTE UNCERTAINTY IN IPOS: EVIDENCE FROM THE TUNISIAN STOCK MARKET

Busness Excellence and Management Jerb, A. UNDERPRICING AND EX ANTE UNCERTAINTY IN IPOS: EVIDENCE FROM THE TUNISIAN STOCK MARKET Ahmed JERIBI Unversty of Sfax, Sfax, Tunsa ahmedjerb07@yahoo.fr Abstract In our paper, we use a sample of 33 IPOs lsted on the Tunsan Stock Exchange (TSE) from 1994 to 2012 n order to study how the ex ante uncertanty affects the total and the nvoluntary underprcng. We fnd that the most underprced frms are IT frms and those whch choose the fxed prce mechansm and have the lttle ssue sze and the lowest number of days between the prospectus regstraton and the frst day of the subscrpton perod. We also fnd that only the frm s sector and the number of days between the prospectus regstraton and the frst day of the subscrpton perod affect nvoluntary underpcng. Keywords: Intal publc offerng (IPO), Ex ante uncertanty, Total underprcng, Involuntary underprcng. 1. INTRODUCTION Many studes argue that the frm s valuaton uncertanty s related to the level of the nformaton asymmetry relatve to the company. Rtter (1984) argued that rsker IPOs wll be more underprced than less-rsky ones. The rsk can reflect the complexty of the prcng problem. The author ntroduced the changng rsk composton hypothess. Rock (1986), Beatty and Rtter (1986), Lowry et al. (2010) and Jerb and Jarbou (2015, b) assume that frms whch have uncertan prospects are nherently dffcult to value. They postulate that underprncng s the effcent reasponse to ths valuaton problem. Lowry et al. (2010) assume that the complexty of the IPO prcng lmts the ablty of the underwrters to precsely value IPOs. Ths complexty s related to the market-wde factors and the frm s-specfc factors. Ljungqvst (2006) categorzed ex ante uncertanty proxes nto four groups: company s characterstcs, offerng characterstcs, prospectus dsclosure, and aftermarket varables. Jerb et al. (2014) decomposed total underprcng nto voluntary and nvoluntary underprcng. In ths paper, we use 33 Tunsan IPOs between 1994 and 2012 and examne how the frm s valuaton uncertanty measured by the ex ante uncertanty proxes affect total underprcng and nvoluntary undeprcng. 65

Busness Excellence and Management Jerb, A. Ths paper can have an mportant contrbuton to the lterature. Ths contrbuton s the study of the relatonshp between the ex ante uncertanty and the nvoluntary underprcng. The rest of ths study s organzed as follows: Secton 2 brefly revews the exstng lterature; Secton 3 presents the sample and the methodology; Secton 4 presents the emprcal results; Secton 5 concludes. 2. LITERATURE REVIEW Beatty and Rtter (1986) argue that there s an equlbrum relaton between the expected underprcng and the ex ante uncertanty regardng the frm s value. They demonstrate that there s a monotonous relatonshp between expected underprcng and the uncertanty of nvestors about ts value. They fnd that there s a postve relatonshp between the ex ante uncertanty about IPO share value and the underprcng. Ths postve relatonshp makes t s dffcult for nvestors to estmate the ntal returns on a hgh-rsk IPO. The authors predct that the frm s sze s negatvely assocated wth underprcng. They assume that small frms and small ssues are more volatle and uncertan. Mchaely and Shaw (1994) confrmed the predcton of Beatty and Rtter (1986) and show that the ssue sze has a sgnfcant role n explanng the degree of underprcng. The results of Mok and Hu (1998) suggest that a bgger ssue s less rsky than a smaller one. Beatty and Rtter (1986), Rtter (1991), Lowry and Schwert (2004) and Lowry et al. (2010) fnd that IT and younger frms are the rskest and consequently the most underprced. Jerb and Jarbou (2015, b) examne the role of the ex ante uncertanty n the pre-lstng Tunsan IPO process. Frst, they test the relaton between the delberate prce dscount level and the ex ante uncertanty proxes. They fnd that hgher ex ante uncertanty does not always lead to a hgh dscount level. Second, they tested the relaton between the subscrpton rato and the delberate prce dscount level. They fnd that sgnalng through the dscount level sn t credble for nvestors. Thrd, they test the relaton between the nvestors demand durng the subscrpton perod and the ex ante uncertanty proxes. They fnd that the level of uncertanty and nformaton asymmetry surroundng the IPO frms s the most mportant determnant of the nvestors demand durng the subscrpton perod. They argue that a frm can reduce the ssue sze, the number of days between the prospectus regstraton and the frst day of the subscrpton perod and the subscrpton perod n order to stmulate the nvestors demand. Jerb et al. (2014) examne the total and nvoluntary underprcng. They defne nvoluntary underprcng as the dfference between the closng prce and the far value estmate, dvded by the far value estmate. They fnd that the total and nvoluntary underprcng are explaned by the nvestors demand and the underwrter s reputaton. Jerb and Jarbou (2015, a) determne the mnmum prce requred by 66

Busness Excellence and Management Jerb, A. an nvestor allowng for recoverng the ntal nvestment, nformaton costs, transacton costs, and the offsettng of shortfall. They expect that the ntal return of an IPO share n the Tunsan market s postvely related to the followng factors: the number of non-nsttutonal nvestors who partcpate durng the subscrpton perod, the subscrpton rato of nsttutonal nvestors, the expected rate of return by nvestors, the gap between the closng date of the subscrpton perod and the day followng the announcement of the subscrpton result, the gap between the announcement of the subscrpton result and the frst lstng day, the number of tradng days, the cost of nformaton and the transacton costs. However, t s negatvely related to other determnants, such as the dscount level, the number of shares allocated for a non-nsttutonal nvestor and the number of offered shares, whch are allocated to nonnsttutonal nvestors. 3. SAMPLE AND METHODOLOGY 3.1 Sample The ntal sample s smlar to that of Jerb and Jarbou (2015, b). It conssts of 33 IPOs collected from the TSE durng the perod from 1994 to 2012. The data used n our study come from dfferent sources. Thus, the lstng prospectuses of the perod (2006-2012) are extracted from the Fnancal Market Councl (FMC) webste and whereas those of the perod (1994-2005) are drawn manually from the lbrary of the FMC. The share prces of the perod (1998-2012) are extracted from the TSE webste and those of the perod (1994-1997) are drawn manually from the archve of the TSE. 3.2 Methodolgy Rock (1986) and Beatty and Rtter (1986) argue that underprncng s the effcent response to the complexty of the IPO prcng. Beatty and Rtter (1986) use the frm s sze and age and the ssue sze as ex ante uncertanty proxes. Loughran and Rtter (2002), Lowry and Schwert (2004), Lowry et al (2010) and Jerb and Jarbou (2015, b) use the frm s sector as ex ante an uncertanty proxy. They fnd that IT frms are the rskest and are the most requested. Rock (1986), Loughran et al (1994) and Chowdhry and Sherman (1996) fnd that the fxed offer prce mechansm s assocated wth the most underprcng level. Lee et al. (1996) decompose the perod between the offcal regstraton of the prospectus and the begnnng of exchange tradng nto three sub-perods: the perod between the offcal regstraton of the prospectus and the openng of the ssue to subscrbers, the subscrpton perod and the perod between the ssue closng and the begnnng of exchange tradng. They fnd that underprcng s postvely related to these three perods. Lowry et al. (2010) suggest that the valuaton problem for IPO frms wth uncertan prospects s related to the frm-specfc factors and the market-wde factors. Ljungqvst (2006) 67

Busness Excellence and Management Jerb, A. categorzes ex ante uncertanty proxes nto four groups: the company and the offerng characterstcs, the prospectus dsclosure, and the aftermarket varables. In our study, we focus on the company s characterstcs and the offerng characterstcs by the use of the same varables as Jerb and Jarbou (2015, b). Frst, we study the relaton between the underprcng and the proxes for ex ante uncertanty regardng the company s characterstcs. The regresson model (1) s specfed as follows: U 0 1LOG( MV ) 2LOG( AGE) 3IT 4 ALTERNATIV E (1) Where; U s the total underprcng. It s defned as (Closng Offer) /Offer, where Closng and Offer are the frst lstng day closng and the offer prces, respectvely. LOG (MV) s the natural logarthm of the market value of the equty defned as the total number of IPO shares multpled by the offer prce (nflaton adjusted to the 2012 value). LOG (AGE): The natural logarthm of the frm s age (n number of years) at the tme of the IPO. IT: Dummy varable whch equals one f the frm s n a hgh-tech ndustry, and zero otherwse. ALTERNATIVE: Dummy varable whch equals one f the IPO frm wll be lsted on the alternatve market, and zero otherwse. Second, we study the relaton between total underprcng and the proxes for ex ante uncertanty regardng the offerng characterstcs. The regresson model (2) s specfed as follows: U 0 1OPF 2LOG( IS) 3SD 4 N (2) Where; U s the total underprcng. It s defned as (Closng Offer) /Offer, where Closng and Offer are the frst lstng day closng and the offer prces, respectvely. LOG (IS): The natural logarthm of the ssue sze whch s defned as the total number of shares offered to the publc multpled by the offer prce; OPF: Dummy varable whch equals one f the IPO frm chooses the fxed prce mechansm (OPF), and zero otherwse. SD: The number of the subscrpton days. N: The number of days between the prospectus regstraton and the frst day of the subscrpton perod. 68

Busness Excellence and Management Jerb, A. Jerb et al. (2014) dvde total undercng nto voluntary and nvoluntary underprcng. The delberate prce dscount s the voluntary underprcng. They defne nvoluntary underprcng as the dfference between the closng prce and the far value estmate, dvded by the far value estmate. Jerb and Jarbou (2015, b) studed the relaton between the delberate prce dscount and the proxes for ex ante uncertanty. In our study, we use the same steps as when studyng the relatonshp between the total underprcng and the proxes for ex ante uncertanty and we study the relatonshp between the nvoluntary underprcng and the proxes for ex ante uncertanty. Frst, we study the relaton between the nvoluntary underprcng and the proxes for ex ante uncertanty regardng the company characterstcs. The regresson model (3) s specfed as follows: IU 0 1LOG( MV ) 2LOG( AGE) 3IT 4 ALTERNATIV E (3) Where; IU s the nvoluntary underprcng. It s defned as (Closng FV) /Offer, where Closng and FV are the frst lstng day closng prce and the far value, respectvely. LOG (MV) s the natural logarthm of the market value of the equty defned as the total number of IPO shares multpled by the offer prce (nflaton adjusted to the 2012 value). LOG (AGE): The natural logarthm of the frm s age (n number of years) at the tme of the IPO. IT: Dummy varable whch equals one f the frm s n a hgh-tech ndustry, and zero otherwse. ALTERNATIVE: Dummy varable whch equals one f the IPO frm wll be lsted on the alternatve market, and zero otherwse. Second, we study the relaton between the nvoluntary underprcng and the proxes for ex ante uncertanty regardng the offerng characterstcs. The regresson model (4) s specfed as follows: IU 0 1OPF 2LOG( IS) 3NSD 4 N (4) Where; IU s the nvoluntary underprcng. It s defned as (Closng FV) /Offer, where Closng and FV are the frst lstng day closng prce and the far value, respectvely.log (IS): The natural logarthm of the ssue sze whch s defned as the total number of shares offered to the publc multpled by the offer prce; OPF: Dummy varable whch equals one f the IPO frm chooses the fxed prce mechansm (OPF), and 69

Busness Excellence and Management Jerb, A. zero otherwse. NSD: The number of the subscrpton days. N: The number of days between the prospectus regstraton and the frst day of the subscrpton perod. 4. EMPIRICAL RESULTS Table 1 presents the descrptve statstcs for a full sample of 33 Tunsan IPOs. We show that the total underprcng (25% ) s lower than what was found by Gana and El Ammar (2008), Ben Slama et al. (2011), Jerb et al. (2014) and hgher than the one found by Ben Naceur and Ghanem (2001), Mefteh Rekk and Boujelbene (2013) and Kanoun and Taktak (2013) on the Tunsan fnancal market. For the newly lsted securtes and durng the frst three tradng days, the allowable range s determned by applyng a percentage fluctuaton ± 18% compared to the reference prce of the day, no addtonal percentage fluctuaton s allowed. Durng these three sessons, only one fxng s programmed by sesson. However, these specal rules cease to be applcable when a course s set durng one of the frst two sessons. When the ±6.09% prce lmt s appled to all lsted securtes, the IPO shares aren t allowed to be freely traded untl the prce s determned to be n the range of the allowed prces. Wth the agreement of the CMF, the underwrters can legally support the IPO prces n the Tunsan market by sgnng a contract of lqudty for a determned perod. The underwrter guarantees the regulaton of the securty prce by purchasng or sellng. The nvoluntary underprng level (12%) s lower than the total underprcng. Ths can be explaned by the hgh delberate dscount level (voluntary underprcng) that IPO frms practce. TABLE 1 DESCRIPTIVE STATISTICS OF THE SAMPLE Varables Mean Medan Max Mn Std.dev U 0,25 0,16 0,90-0,02 0,27 IU 0,12 0,03 0,80-0,25 0,25 MV 106170,58 44447,44 615707,2 9315,03 136099,2 AGE 29,06 21,56 106,01 1,24 24,15 IT 0,06 n/a n/a n/a n/a ALTERNATIVE 0,12 n/a n/a n/a n/a OPF 0,24 n/a n/a n/a n/a IS 19361,56 10252,83 123141,47 2000 23764,75 SD 13,48 12 30 5 5,83 N 14,39 12 60 6 10,59 Ths table reports descrptve statstcs for the whole sample of 33 Tunsan between 1994 and 2012. The data are collected from the CMF database, TSE database and IPO prospectuses. U: The total underprcng. It s defned as (Closng Offer) /Offer, where Closng and Offer are the frst lstng day 70

Busness Excellence and Management Jerb, A. closng and the offer prces, respectvely. IU: The nvoluntary underprcng. It s defned as (Closng FV) /Offer, where Closng and FV are the frst lstng day closng prce and the far value, respectvely. MV: The market value of the equty. It s defned as the total number of IPO shares multpled by the offer prce (nflaton adjusted to the 2012 value). AGE: The frm age (n number of years) at the tme of the IPO. IT: Dummy varable whch equals one f the frm s n a hgh-tech ndustry (computer equpment, electroncs, and general IT), and zero otherwse. ALTERNATIVE: Dummy varable whch equals one f the IPO frm s lsted on the alternatve market, and zero otherwse. OPF: Dummy varable whch equals one f the IPO frm chooses the fxed prce mechansm (OPF), and zero otherwse. IS: The ssue sze. It s defned as the total number of shares offered to the publc multpled by the offer prce (nflaton adjusted to the 2012 value). SD: The number of the subscrpton days. N: The number of days between prospectus regstraton and the frst day of the subscrpton perod. We fnd that only 6% of the frms are n a hgh-tech ndustry. Ths small percentage by the means that ths ndustry s relatvely new n Tunsa and ths sector requres strong nvestment n research and development. However, the Tunsan market s relatvely small due to the small populaton sze and the low Tunsan GDP. We also fnd that 12% of the frms that make up our sample are lsted on the alternatve market. Indeed, ths market whch s reserved to small and medum-szed companes was launched n 2008, whch means that t s relatvely new. IPO frms can choose between the four followng sellng mechansms: drect regstraton, mnmum sale prce, open offer prce, and fxed-prce offerng. When the captal of the IPO frm s suffcently dffused to the publc, the drect regstraton mechansm (procedure) allows the TSE to proceed to the drect regstraton of the shares on the market (prncpal and alternatve) n order to be traded wth an ntal prce fxed by the TSE. If the frm s lsted on the alternatve market, the fracton of nsttutonal shareholdngs retaned one year before the IPO should be at least at the level of 20%. A notce emanatng from the TSE must be publshed at least two days before the frst lstng day. It must specfy the date of the frst lstng and the ntal prce. The mnmum sale prce mechansm allows the pre-ipo shareholders and underwrters to sell a quantty of securtes at a mnmum prce the day of the ntroducton. The TSE centralzes the purchasng orders submtted by the underwrter. The orders wth a lmted prce are only accepted and those whch abnormally exceed the mnmum sale are elmnated. The TSE selects and classfes the purchasng orders accordng to ther prce lmt then satsfes these orders completely or partally and fxes the ntroducton lsted prce whch s sngle and corresponds to the lower prce lmt of the last served order. Tunsan ssuers can also use a modfed bookbuldng procedure called the open offer prce that conssts n makng a number of securtes avalable to the publc by fxng a prce range whch can be modfed by the underwrter, provded that he has the agreement of the TSE. In order to determne the 71

Busness Excellence and Management Jerb, A. fnal prce, the leadng underwrter collects the orders ssued by the nsttutonal nvestors and uses the bookbuldng technque. In a fxed-prce offerng, a fxed number of securtes s offered at a fxed prce. The Stock Exchange centralzes the orders submtted by the underwrters. The only accepted orders are those whch are presented at the offer prce. From table 1, we can see that 24% of the IPO frms choose the fxed prce mechansm (OPF). TABLE 2 PEARSON CORRELATION COEFFICIENTS BETWEEN INDEPENDENT VARIABLES LOG (MV) LOG(AGE) IT ALTERNATIVE OPF LOG(IS) SD N LOG(MV) 1 LOG(AGE) 0,1961 1 IT -0,1778 0,1072 1 ALTERNATIVE -0,2172-0,2935* 0,2948* 1 OPF 0,2351 0,0896 0,4490*** 0,6565*** 1 LOG(IS) 0,9085*** 0,3033* -0,2616-0,2015 0,2693 1 SD -0,1284 0,1954 0,0228-0,1445 0,0138 0,0104 1 N -0,0491 0,2285 0,5507*** 0,075 0,3991** -0,0488 0,1395 1 N =33 Ths table reports the results of Pearson correlaton coeffcents between ndependent varables. The base sample ncludes 33 IPOs lsted on TSE between 1994 and 2012. The data for regressons are collected from the CMF database, TSE database and IPO prospectuses. LOG (MV) s the natural logarthm of the market value of the equty defned as the total number of IPO shares multpled by the offer prce (nflaton adjusted to the 2012 value). LOG (AGE): The natural logarthm of the frm age (n number of years) at the tme of the IPO. IT: Dummy varable whch equals one f the frm s n a hghtech ndustry, and zero otherwse. ALTERNATIVE: Dummy varable whch equals one f the IPO frm s lsted on the alternatve market, and zero otherwse. OPF: Dummy varable whch equals one f the IPO frm chooses the fxed prce mechansm (OPF), and zero otherwse. LOG (IS): The natural logarthm of the ssue sze whch s defned as the total number of shares offered to the publc multpled by the offer prce (nflaton adjusted to the 2012 value). SD: The number of the subscrpton days. N: The number of days between prospectus regstraton and the frst day of the subscrpton perod. *** the correlaton s sgnfcant at the 1% level (blateral); ** the correlaton s sgnfcant at the 5% level (blateral); * the correlaton s sgnfcant at the 10% level (blateral). In order to dentfy the potental problems of multcolnearty among the ndependent varables of our study, we establsh a correlaton matrx. Table 2 presents the results of Pearson correlaton coeffcents between the ndependent varables. The results reveal sgnfcant relatonshps among many ndependent varables wth low correlaton coeffcents. However, we fnd that the correlaton coeffcent 72

Busness Excellence and Management Jerb, A. between LOG (MV) and LOG (IS) s very hgh. We also show that the correlaton coeffcent value s mportant (65.65%) for OPF and ALTERNATIVE. TABLE 3 TOTAL UNDERPRICING AND EX ANTE UNCERTAINTY Varables Exp sgn regresson (1) regresson (2) regresson (3) Constant 0,8480 2,1201 1,0711 (0,9989) (2,4400) (1,1700) LOG(MV) - -0,0332 (-0,6976) LOG(AGE) - -0,0086 0,0144 (-0,1962) (0,3232) IT + 0,2014 0,4302* (0,9322) (1,7129) ALTERNATIVE + 0,0994 0,0840 (0,6177) (0,5430) OPF + 0,2561** (2,1722) LOG(IS) - -0,1069* -0,0425 (-2,0286) (-0,7464) SD - -0,0041-0,0038 (-0,5262) (-0,4645) N - -0,0089* -0,0104* (-1,8787) (-1,9266) R2 0,0963 0,2292 0,2260 Adj R2 0,0328 0,1191 0,0474 Ths table reports the results of ordnary least squares (OLS). The base sample ncludes 33 IPOs lsted on TSE between 1994 and 2012. The data for regressons are collected from the CMF database, TSE database and IPO prospectuses. Dependent varable s the total underprcng. It s defned as (Closng Offer) /Offer, where Closng and Offer are the frst lstng day closng prce and the offer prce, respectvely. LOG (MV) s the natural logarthm of the market value of the equty defned as the total number of IPO shares multpled by the offer prce (nflaton adjusted to the 2012 value). LOG (AGE): The natural logarthm of the frm age (n number of years) at the tme of the IPO. IT: Dummy varable whch equals one f the frm s n a hgh-tech ndustry, and zero otherwse. ALTERNATIVE: Dummy varable whch equals one f the IPO frm s lsted on the alternatve market, and zero otherwse. OPF: Dummy varable whch equals one f the IPO frm chooses the fxed prce mechansm (OPF), and zero otherwse. LOG (IS): The natural logarthm of the ssue sze whch s defned as the total number of shares offered to the publc multpled by the offer prce (nflaton adjusted to the 2012 value). SD: The 73

Busness Excellence and Management Jerb, A. number of the subscrpton days. N: The number of days between prospectus regstraton and the frst day of the subscrpton perod. * Sgnfcant at the 10% level. **Sgnfcant at the 5% level. ***Sgnfcant at the 1% level. Table 3 presents the results of the OLS model regressons. In regresson (1), we test the relaton between the total underprcng and the proxes for ex ante uncertanty regardng the company s characterstcs (frm sze, age, ndustry and market). We fnd that there s no sgnfcant statstcs relatonshp between total underprcng and the proxes for ex ante uncertanty regardng the company characterstcs. In regresson (2), we test the relaton between the total underprcng and the proxes for ex ante uncertanty regardng the offerng characterstcs. The coeffcent on the OPF s postve and sgnfcantly dfferent from zero at 0.05 level. Ths result ndcates that frms whch choose the fxed prce mechansm are more underprced than others. Ths result confrms that of Reber and Fong (2006), Agarwal et al. (2008) and Jerb and Jarbou (2015, b) who found that IPO frms whch choose the fxed prce mechansm are the most over-subscrpted. The coeffcent on the ssue sze s negatve and sgnfcantly dfferent from zero at 0.1 level. Ths result ndcates that frms whch have the lttle ssue sze are more underprced than others. Ths result confrms that of Beatty and Rtter (1986) and Jerb and Jarbou (2015, b). We also fnd a negatve and sgnfcant relatonshp (at the 0.1 level) between the underprcng level and the number of days between prospectus regstraton and the frst day of the subscrpton perod. Ths last result s consstent wth the predcton of Lee et al. (1996) and the result of Jerb and Jarbou (2015, b). However, t s nconsstent wth the result of Jerb and Jarbou (2015, a). In regresson (3), we test the relaton between total underprcng and the proxes for ex ante uncertanty. We fnd that the varable TECH has a postve and sgnfcant effect (at the 0.1 level) on the underprcng. Ths result s smlar to that of Zheng and L (2008), Lowry et al (2010) and Jerb and Jarbou (2015, b). From table 3, we conclude that only the proxes for ex ante uncertanty regardng the offerng characterstcs affect underpcng. The most underprced frms are IT frms and those whch choose the fxed prce mechansm and have the lttle ssue sze and the lowest number of days between the prospectus regstraton and the frst day of the subscrpton perod. TABLE 4 INVOLUNTARY UNDERPRICING AND EX ANTE UNCERTAINTY Varables Exp sgn regresson (1) regresson (2) regresson (3) Constant 0,6513 1,7112** 0,9050 (0,8087) (2,0369) (1,1063) LOG(MV) - -0,0295 (-0,6537) LOG(AGE) - -0,0076 0,0191 (-0,1828) (0,4805) 74

Busness Excellence and Management Jerb, A. IT + 0,2621 0,5507** (1,2784) (2,4536) ALTERNATIVE + -0.0083-0,0305 (-0,0549) (-0,2205) OPF + 0,1769 (1,5517) LOG(IS) - -0,0890-0,0382 (-1,7461) (-0,7501) SD - -0,0045-0,0049 (-0,6078) (-0,6701) N - -0,0086* -0,0128** (-1,8787) (-2,6502) R2 0,0865 0,1908 0,3059 Adj R2 0,0440 0,0751 0,1458 Ths table reports the results of ordnary least squares (OLS). The base sample ncludes 33 IPOs lsted on TSE between 1994 and 2012. The data for regressons are collected from the CMF database, TSE database and IPO prospectuses. Dependent varable s the nvoluntary underprcng. It s defned as (Closng FV) /Offer, where Closng and FV are the frst lstng day closng prce and the far value, respectvely. LOG (MV) s the natural logarthm of the market value of the equty defned as the total number of IPO shares multpled by the offer prce (nflaton adjusted to the 2012 value). LOG (AGE): The natural logarthm of the frm age (n number of years) at the tme of the IPO. IT: Dummy varable whch equals one f the frm s n a hgh-tech ndustry, and zero otherwse. ALTERNATIVE: Dummy varable whch equals one f the IPO frm s lsted on the alternatve market, and zero otherwse. OPF: Dummy varable whch equals one f the IPO frm chooses the fxed prce mechansm (OPF), and zero otherwse. LOG (IS): The natural logarthm of the ssue sze whch s defned as the total number of shares offered to the publc multpled by the offer prce (nflaton adjusted to the 2012 value). SD: The number of the subscrpton days. N: The number of days between prospectus regstraton and the frst day of the subscrpton perod. * Sgnfcant at the 10% level. **Sgnfcant at the 5% level. ***Sgnfcant at the 1% level. Table 4 presents the results of the OLS model regressons. In regresson (1), we test the relaton between the nvoluntary underprcng as entrepreneural nvoluntary wealth loss and the proxes for ex ante uncertanty regardng the company s characterstcs (frm s sze, age, ndustry and market). Our results suggest that there s no sgnfcant statstcs relatonshp between total nvoluntary and the proxes for ex ante uncertanty regardng the company characterstcs. In regresson (2), we test the relaton between the total nvoluntary and the proxes for ex ante uncertanty regardng the offerng 75

Busness Excellence and Management Jerb, A. characterstcs. We fnd that only the varable N has a negatve and sgnfcant effect on the nvoluntary underprcng level at 0.1 level. In regresson (3), we test the relaton between the nvoluntary underprcng and the proxes for ex ante uncertanty. We fnd that the varable IT has a postve and sgnfcant effect (at the 0.05 level) on the nvoluntary underprcng. We also fnd that there s a negatve and sgnfcant relatonshp (at 0.05 level) between the nvoluntary underpcng level and the number of days between the prospectus regstraton and the frst day of the subscrpton perod. From table 4, we conclude that only the frm s sector and the number of days between the prospectus regstraton and the frst day of the subscrpton perod affect nvoluntary underpcng. 5. CONCLUSIONS In ths paper, we test how ex ante uncertanty affects total and nvoluntary underprcng. Frst, we test the relaton between the total underprcng and the ex ante uncertanty proxes. We fnd that only the proxes for ex ante uncertanty regardng the offerng characterstcs affect underpcng. The most underprced frms are IT frms and those whch choose the fxed prce mechansm and have the lttle ssue sze and the lowest number of days between prospectus regstraton and the frst day of the subscrpton perod. Second, we test the relaton between nvoluntary underprcng and the ex ante uncertanty proxes. We fnd that only the frm s sector and the number of days between the prospectus regstraton and the frst day of the subscrpton perod affect nvoluntary underpcng. Fnanly, we can conclude that the level of uncertanty and nformaton asymmetry surroundng the IPO frms s the most mportant determnant of the total and the nvoluntary underprcng. However, Jerb and Jarbou (2015, b) found that the level of uncertanty and nformaton asymmetry surroundng the IPO frms are the most mportant determnants of the nvestors demand durng the subscrpton perod. REFERENCES Aggarwal, S., Lu, C. and Rhee, R. (2008). Investor demand for IPOs and aftermarket performance: Evdence from the Hong Kong stock market. Journal of Internatonal Fnancal Markets, Insttutons and Money, 18 (2), 176 190. Ben Slama, S., Boudrga, A. and Boulla, N. (2011). Determnants of IPO underprcng: Evdence from Tunsa. The Internatonal Journal of Busness and Fnance Research, 5 (1), 3-32. Benvenste, L.M.and Spndt, P.A., (1989). How nvestment bankers determne the offer prce and allocaton of new ssues? Journal of Fnancal Economcs 24 (2), 343 361. 76

Busness Excellence and Management Jerb, A. Beatty, R. and Rtter, J.R., (1986). Investment bankng, reputaton, and the underprcng of ntal publc offerngs. Journal of Fnancal Economcs 15 (1), 213 232. Ben Naceur, S.and Ghanem, H., (2001). The Short and Long-run Performance of New Lstngs n Tunsa. Internatonal Revew of Fnance 2 (4), 235 246. Chowdhry, B.and Sherman, A., (1996). The wnner's curse and nternatonal methods of allocatng ntal publc offerngs. Pacfc-Basn Fnance Journal 4 (1), 15 30. Gana, M.R.and El Ammar, A., (2008). Intal underprcng and transfer of shares on the Tunsan stock exchange. Journal of Corporate Ownershp and Control 5 (3), 434 444. Jerb, A.and Jarbou, A., (2014). Explanng delberate IPO prce dscount: Evdence from the Tunsan stock market. Internatonal Journal of Busness and Emergng Markets 6 (2), 121 138. Jerb, A.and Jarbou, A., (2015). (a). The mnmum prce requred by nvestors n IPOs. Decson Scence Letters, 4 (1), 63-76. Jerb, A.and Jarbou, A., (2015). (b). The role of the ex ante uncertanty n the pre-lstng IPO process: Evdence from the Tunsan stock market. Internatonal Journal of Economcs and Busness Research. In press. Jerb, A., Jerb Masmoud, M., and Jarbou, A. (2014). Voluntary and Involuntary Underprcng n IPOs. Global Busness and Management Research: An Internatonal Journal, 6 (2), 122-138. Kanoun, S.and Taktak, S., (2013). Informaton asymmetres, ssuers ncentves and underprcng n emergng market: Some evdence from Tunsan IPO frms. Internatonal Journal of Economcs, Commerce and Research 3 (5), 83 94. Lee, P.J., Taylor, S.L. and Walter, T.S., (1996). Expected and realzed returns for Sngaporean IPOs: ntal and long-run analyss. Pacfc-Basn Fnance 4 (2), 153 180. Ljungqvst, A.,(2006). IPO underprcng. In: Eckbo, B.E. (Ed.), Handbook of Emprcal Corporate Fnance. North-Holland. Loughran, T.and Rtter, J.,(2002). Why don't ssuers get upset about leavng money on the table n IPOs. Revew of Fnancal Studes 15 (2), 413 444. Loughran, T., Rtter, J.R.and Rydqvst, K., (1994). Intal publc offerngs: nternatonal nsghts. Pacfc Basn Fnance Journal 2 (2-3), 165 19. Lowry, M.and Schwert, W., (2004). Is the IPO prcng process effcent? Journal of Fnancal Economcs 71 (1), 3 26. Lowry, M., Offcer, M. and Schwert, W., (2010). The Varablty of IPO Intal Returns. Journal of Fnance 2 (1), 425 466. Mefteh Rekk, Y.and Boujelbene. Y., (2013). Tunsan IPOs underprcng and long-run underperformance: Hghlght and explanaton. E3 Journal of Busness Management and Economcs 4(4), 93-104 Mchaely, R.and Shaw, W., (1994). The prcng of ntal publc offerngs: tests of adverse-selecton and sgnalng theores. Revew of Fnancal Studes 7 (2), 279 319. Mok, H.and Hu, Y., (1998). Under-prcng and aftermarket performance of IPOs n Shangha, Chna. Pacfc-Basn Fnance Journal 6 (5), 453 474. 77

Busness Excellence and Management Jerb, A. Reber, B.and Fong, C., (2006). Explanng msprcng of ntal publc offerngs n Sngapore. Appled Fnancal Economcs 16 (18), 1339 1353. Rtter, J., (1984). The hot ssue market of 1980. Journal of Busness 57 (2), 215 241. Rtter, J., (1991). The long-run performance of ntal publc offerngs. Journal of Fnance 46 (1), 3 27. Rock, K., (1986). Why new ssues are underprced. Journal of Fnancal Economcs 15 (1-2), 187 212. Zheng, S.X.and L, M., (2008). Underprcng, ownershp dsperson, and aftermarket lqudty of IPO stocks. Journal of Emprcal Fnance 15 (3), 436 454. 78