The Effect of IPO Lockup Agreements on Stock Prices An

The Effect of IPO Lockup Agreements on Stock Prices: An Empirical Analysis on the Taiwan Stock Exchange Dar-Hsin Chen Chun-Da Chen Lloyd P. Blenman Feng-Shun Bin * Dar-Hsin Chen is an associate professor at the Graduate Institute of Finance and Department of Information and Financial Management, National Chiao-Tung University. Chun-Da Chen is an instructor at the Department of Banking and Finance, Tamkang University, Taiwan. Lloyd P. Blenman is an associate professor at the Department of Finance and Business Law, University of North Carolina at Charlotte, USA. Feng-Shun Bin is an assistant professor at the Department of Business Administration, University of Illinois at Springfield, USA. ** Please address all correspondence to Lloyd P. Blenman, Department of Finance and Business Law, University of North Carolina – Charlotte, Belk College of Business, Charlotte, NC 28223, USA. Email: lblenman@email.uncc.edu. Tel: 704-687-2823. Fax: 704-687-6967. Abstract We examine the performance of Taiwanese IPOs surrounding lockup expiration. We find some evidence that investors in IT-IPOs could suffer significant wealth losses, in the days just prior to the lockup expiration. Non-IT IPO investors are not similarly affected. However, contrary to previous research, we find significant return reversals for both IT and non-IT stocks. After the lockup expiration dates trading volumes are also abnormally high. The returns to IT firms and firms with larger market capitalizations, exhibit greater sensitivity to lockup expiration dates than those of non-IT firms and small-cap firms. JEL classification: G12; G14; G15 Keywords: IPOs, Lockup Period, Abnormal Returns, Event Studies. 1 The Effect of IPO Lockup Agreements on Stock Prices: An Empirical Analysis on the Taiwan Stock Exchange 1. Introduction Typically, when an initial public offering (IPO) or seasoned equity offering (SEO) takes effect, only a fraction of the shares outstanding is issued to the public, whereas insiders (e.g., the executive officers, block shareholders, and founders) own the remaining shares. Shares owned by insiders are subject to liquidity restrictions within a specified period after the IPO or SEO date. This restricted period is commonly named as a “lockup period” in the US or as a “lock-in agreement” in the UK. There is no uniform rule regarding the length of the lock-up period in the US, which can range from a month to several years.1 When the lockup period expires, US insiders are then allowed to sell their shares up to the volume limits in accordance with SEC Rules 144 or 701. 2 The date of an IPO lockup expiration is disclosed in the IPO prospectus. In an efficient capital market, the possible effects of allowing the insiders to sell their holdings freely should be incorporated into the IPO performance prior to the expiration of lockup period, since the non-restricted shareholders (i.e., non-insiders) can trade freely in anticipation of the increased sell-offs at the lockup expiration. However, if any significant price reactions are observed at the expiration, the semi-strong-form EMH may be violated, indicating that the stock market prices fail to reflect a publicly-anticipated event before its occurrence. In this paper, we employ event-study techniques to find out whether the IPO trading behaviors (price and volumes) are “normal” surrounding the expiration of lockup periods. Previous studies on the US IPO market have investigated the possible price impact surrounding a 1 SEOs usually have shorter lockup periods than IPOs. Information relating to lock-ups is contained in the SEC S-1 registration statement and the 424 filing (prospectus). 2 Rule 144 generally applies to offers and sales of securities relating to corporate insiders and buyers of private placement securities, while Rule 701 applies to offers and sales of securities pursuant to certain compensatory benefit plans and contracts relating to compensation. 2 single IPO lockup expiration date. We extend the study by further examining the valuation impact of multiple lockup expiration dates on the performance of Taiwanese IPOs before and after the expiration of lockup periods. To our knowledge, published research has mainly focused on US IPO lockup expiration effects and we have found no work on emerging market IPOs. So we attempt to fill in this gap through our contributions in this study. The rest of this paper proceeds as follows. Section 2 reviews the relevant literature on IPO “unlock” periods and the structure of Taiwan stock exchange. Section 3 outlines the hypotheses and data used in the study. Section 4 analyzes the methodology and empirical results of price performance surrounding the IPO lockup period expiration. Section 5 examines trading activities around the expiration of lockup period and employs a regression analysis to identify factors that might affect stock returns and trading volumes. Section 6 summarizes our findings. 2. Background Reviews 2.1 Literature on Lockup Regulation Information asymmetry models developed by Welch (1989), Allen and Faulhaber (1989), and Grinblatt and Hwang (1989) state that informed issuers signal the intrinsic value of IPOs to uninformed investors by deliberately under-pricing and retaining shares. Existing studies of IPO transactions in stock markets, like those of (Ritter, 1991; Krigman, Shaw and Womack, 1999; and Aggarwal, Krigman, and Womack, 2002), generally document the persistence of initial under-pricing. Some other recent findings (e.g., Mohan and Chen, 2001) suggest that the length of the lockup period conveys material information that is pertinent to the risk of an IPO. Specifically, if information asymmetries are high, the IPO should have a longer lockup period because it allows for more time for private information to be transferred to the public. Field and Hanka (2001). study abnormal returns and trading volume surrounding the lockup expiration for IPOs in the US market. They document a three-day cumulative abnormal return of -1.5% along with a trading volume increase of 40% on average when lockups on US 3 IPOs expire. The losses in market values and the increases in trading volumes, appear to be significantly larger if the US IPO-issuing firms are heavily financed by venture capital funds. They also find, that at lockup expirations, venture capital funds sell far more aggressively than firm executives and other shareholders. Ofek and Richardson (2003) provide similar results for the internet-sector IPOs. Bradley et al. (2001)’s empirical work also indicates that lockup expirations are accompanied by significant short-term decline in stock prices for new issues, and those IPOs with large venture capital stakes suffer considerably greater value losses; whereas Ofek and Richardson (2000) find a permanent price decline in addition to an volume increase following lockup expirations. As for pre-lockup-expiration performance, Keasler (2001) observes negatively significant abnormal returns prior to the lockup expiration date, whereas the post-expiration abnormal returns are statistically insignificant. Moreover, IPO issues with longer lockup periods tend to accumulate greater negative returns surrounding the expected expiration of lockup period. To date, only one corresponding study has been conducted on a non-US market; that is Espenlaub, Goergen, and Khurshed’s (2001) survey of IPOs in the UK. They observe that while the coverage scale of lockup agreements is much larger in the US, the UK provides greater diversity, including the allowance for separate lockup agreements corresponding to different categories of shareholders. Specifically, out of the total 188 UK sample firms, merely 54 set specific calendar dates for lockup expirations in “absolute” terms. The rest of UK firms set the lockup expiration date in firm-specific “relative” terms, such as “30 days after the publication of annual reports”. In the UK the average periods are close to two years (561 days), while in the US the vast bulk of lockup periods are for six months (180 days). In addition, the estimated proportion of locked-up shares in the UK is approximately 50%, much lower than in many US IPOs with lockup provisions. For technology firms, they find significantly negative cumulative abnormal returns around the expiration date, which appear to persist thereafter. However, they find little 4 evidence of such abnormal returns for non-technology sample UK firms. As Espenlaub, Goergen, and Khurshed (2001) suggest, information asymmetries and agency problems between insiders and new investors are more substantial for the technology sector than for the other industries, possibly causing the difference in abnormal performance between groups. 2.2 Lockup Regulations and The Taiwan Institutional Framework The legal lockup restrictions on Taiwan’s IPO markets are very strict, and considerably different from American and British regulations. The major stock exchange in Taiwan is the Taiwan Stock Exchange (TSE), and Table 1 presents the number of firms, trading volume, and trading value for securities listed on the TSE during the 1990-2001 period. In a US IPO, pre-existing shareholders rarely sell off the majority ownership of the company. Ofek and Richardson (2000) observe that in the US, approximately 15-20% of the shares are issued to the public and the remaining 80-85% shares are thus subject to the lockup constraints. In a Taiwanese IPO, at least 50% of shares in the firm must be sold to the public, and the insiders must place their shares under custody at the Taiwan Securities Central Depository Corporation (TSCD) for at least two years. Even after this two-year period, insiders can only sell up to 20% of those shares every six months. This implies that it would take no fewer than five different “unlock” intervals for insiders to liquidate their stakes entirely. There are no private negotiations with underwriters to permit early termination of lockup agreements. 3. Hypotheses and Data IPO-issuing firms in Taiwan cannot freely select the optimal length of the lockup period. We therefore cannot directly test whether a linkage exists between the length of lockup period and the degree of information asymmetry in the Taiwan equity market. Instead, our study focuses on two main hypotheses as follows: 5 Null Hypothesis 1: The average abnormal return, surrounding the expiration of a lockup period, should equal zero for IPOs. Null Hypothesis 2: There should be no abnormal trading activity (e.g., volume) for IPOs, in the post expiration lockup period. Null Hypothesis 3: Price returns should remain normal as lockup expiration dates approach. Tests of these hypotheses constitute a practical examination of the EMH, which predicts that the expiration of this period should not be associated with any unusual movements in price and/or volume. We also examine if the Taiwanese IT-sector IPOs are more responsive to lockup expirations than non-IT sector IPOs. Null Hypothesis 4: After the lockup expiration period, no differences in trading activities between IT firms and non-IT firms should exist. The investing public may view abnormal trading activities after the expiration of a lockup period as an indicator of insider confidence. Specifically, the heavy volume of insider sales that immediately follows the lockup expiration is perceived as a sign of low insider confidence. This is interpreted to be a bad signal regarding the firm’s prospects. On the contrary, if there is no abnormal change in insider trading volume following the lockup expiration, it should be viewed by the investing public as a signal of high insider confidence and thus, a good indicator of future firm value. The selected sample consists of 127 Taiwan-incorporated firms issuing an IPO on the Taiwan Stock Exchange (TSE), during the period from January 1, 1995 to December 31, 1999. So for each IPO sample firm, the first “unlock” date must fall in the period between January 1, 1997 and December 31, 2001. Out of the 127 sample firms, 48 firms with their first 2-digit stock identification code of 23 are classified as IT companies. 6 We then use the Taiwan Economic Journal (TEJ) database to identify their IPO dates. We screen the data to ensure that no significant information events occur around the lockup expiration dates. The stock return and trading volume for each firm and for the whole TSE market are also collected from the TEJ database. Descriptive data about the sample are presented in Table 2 with further details. 4. The Effects of Expiration of Lockup Periods on Stock Returns The event-study methodology is utilized to analyze the effects of IPO lockup period expirations on stock prices, in line with Mikkelson and Partch (1986, 1988). The abnormal returns are calculated around the date when the lockup period ends. defined as: In specific, the abnormal return (AR) for stock i on day t is ARit Rit ( i i Rmt ) (1) where Rit is the return on stock i on day t, and Rmt is the return on the Taiwan Stock Exchange stock market index on day t . The coefficient of i and i are estimated based on the market model by regressing Rit for the 120-day period [-150, -21], that is, 150 trading days before the event date, defined as the lockup expiration date (TLED) to 21 trading days before TLED on Rmt . We also calculate the cumulative abnormal return (CAR) for each individual firm i, covering 20 trading days pre-(TLED) to 20 days post-TLED. Cross-sectional average of abnormal returns (AARs) and cross-sectional cumulative average abnormal returns (CAARs) are then estimated to investigate for their statistical significance. The results for the event study are reported in Table 3. The daily average abnormal returns (AARs) for (i) all 127 firms in the sample (ii) the 48 IT firms and (iii) the 69 non-IT firms are calculated over the [-20, +20] period relative to TLED 0, respectively. The results show that returns on stocks gradually decline at least 20 trading days prior to TLED. The signs of average abnormal returns are mostly negative for both the whole sample and the IT-sector sub-sample. 7 However, the IT firms are generally more responsive to the lockup expiration date events, and the negatively significant AARs for the IT sector at pre-event days -18, -13, -9, and -3 indicate that unrestricted investors, anticipating impending selling pressure from insiders, apparently liquidate their holdings prior to TLED. The AARs immediately following the unlock day are still negative but statistically insignificant. After TLED has passed, the AARs gradually turn positive. For example, the AARs for IT firms are 0.70% on day 11, 0.96% on day 15, and 0.84% on day 16. results are statistically significant. The cumulative average abnormal returns (CAARs) in Table 3, measure IPO price performance surrounding TLED and indicate a more robust selling pressure for the IT sector. For the 21-day event interval [-20, 0], the CAAR is a significant -3.94% (t = -1.872) for the IT firms on average, while the CAAR for the all-firm sample is also negative (-1.29%) but not statistically significant (t = -0.811).. The CAARs over various event windows are further summarized in Table 4. We find that the CAARs for the various pre-event windows are all negative for both IT firms and all firms, but only the IT sector provides a finding that is statistically significant (for the [-5, 0] interval, CAAR = -2.14% with t = -2.297). Moreover, we also observe that the size of the negative CAARs for All these the across-event period decrease as we expand the event interval. For example, during [-10, 10] and [-20, 20] periods, CAARs for the IT sub-sample are -2.50% and -1.78%, respectively, and those for the all-firm sample are –0.33% and +0.52%, respectively. These four statistics are statistically insignificant. The CAARs in the post-event periods are also analyzed. The CAARs for the [0, 5] window are negative (significant at the 10% level for the IT sub-sample, but insignificant for the full sample). As for the [11, 20] period, however, the CAARs are significantly positive for both the IT sub-sample (at the 1% level) and the full sample (at the 5% level), suggesting that the value effects at lockup expirations may only be temporary phenomena. 8 Figure 1 provides a comparison of time series plot of the CAARs around the lockup expiration days for all firms, the IT-sector and non-IPO sector IPOs. The figure clearly exhibits that the downward drifts of all sub-categories of CAARs gradually turn upward after lockup periods expire. However, there still exists a difference in the price reactions to lockup expirations between the IT sub-sample and the full sample, with the IT sector losing relatively more value on average. Such results reinforce Field and Hanka’s (2001) findings that the average unlock-day returns tend to be more negative for high-tech firms. We however find no evidence that the abnormal returns around lockup expirations are permanent for Taiwanese IPOs as was reported for the case of US IPOs. The transitory pattern of the returns around TLED hold for both IT and no-IT sector IPOs. Based on an analysis of the overall sample, we find no significant evidence of value loss on or around TLEDs, as hypothesized. If we narrow our focus to the IT sector, the results do not support null Hypothesis 1, especially for the short time intervals around TLED. However, as we expand the time horizon, the negative abnormal returns diminish. Furthermore, the different patterns of abnormal returns around TLED indicates that IT firms are more sensitive to lockup expirations than the market average, possibly due to the greater information asymmetries and greater agency costs in the IT industry. 5. Trading Activities around TLEDs and Cross-Sectional Analysis 5.1 Trading Activities around TLEDs To test Hypothesis 2, we analyze the abnormal trading volumes for 41 days, centered on TLED. Abnormal daily trading volume is calculated relative to each stock’s pre-unlock average daily trading volume, as: 9 Abnormal Volume AVi ,T Vi ,T 1 150 Vi ,t 130 t 21 1 (2) where Vi ,T is the trading volume for stock i on day T . volumes of each stock from t 150 to t We first compute the average 0 ). 21 relative to TLED ( t For t 20 to t 20 , we calculate for each day the ratio of daily volume to its mean (which is obtained earlier, based on the period [-150, -21]). We subtract one from the ratios and average across firms to get an estimate of abnormal volume AVi ,T across each day surrounding TLED. The results for the average abnormal volume for each day are reported in Table 5. Figure 2 also provides a graphical illustration of the average abnormal trading volume around TLED. As Table 5 shows, abnormal trading volumes around TLED are mostly positive. We present the cumulative abnormal trading volumes on Table 6. An interesting observation is the “lead” effect of lockup expiration on trading volumes. That is, “abnormally” heavy sales occur even before the lockup expires and before the price begins to drop. For the full sample, the cumulative abnormal trading volumes during the three pre-lockup event windows [-5, 0], [-10, -1] and [-20, -11] are all significantly positive, supporting the conjecture that unrestricted investors are aware of the lockup release and liquidate their holdings prior to the event day. Therefore, the negative returns prior to the unlock day are simply a consequence of unrestricted stock sales. For the IT sector, the lockup expiration effect on trading volumes in the pre-event period is not as evident. The corresponding cumulative abnormal trading volume is statistically significant merely for the [-10, -1] window. Moreover, in the post-event period, we find that a statistically significant increase in trading volume does not take effect until 10 trading days after the unlock day, and such a volume effect lasts for the rest of the post-event window. This abnormal increase in trading volume occurs in the same period of return reversal (from negative to positive) that we observe in Table 4. The post-event window [11, 20], in Table 5, indicates that most of the trading days have 10 positively significant volume increases for both the IT sector and the full sample. The cumulative abnormal trading volume for the IT sub-sample is a positive 4.07% (significant at the 1% level), which exceeds 2.89% for the full sample (also significant at the 1% level), as presented in the last row of Table 6. One possible reason for the phenomenon of active post-event trading along with price reversal is that some investors might view the restricted investors’ liquidations (when the market selling pressure reaches its short-term peak) as a repurchase opportunity, thereby causing a rebound in both share price and trading volume some days after the lockup expiration. Overall, our findings based on Taiwan IPOs do not support the hypothesis that the expiration of a lockup period has no volume effects. We find TLEDs can actually trigger a boost in trading volume considerably with either a lead or a lag. Since TLED is a publicly-known event, the empirical evidence implies that the investing public, fearing the price drop at lockup expiration, do not wait to sell. They choose to unload their holdings (or even sell short) before that date and thus depress the price. However, the evidence also suggests that post-TLED, some market participants view the sell-off as a potential opportunity for buying at bargain prices. This activity explains the price reversals that occur post-TLED. Consistent with previous studies, we observe that new public IT firms tend to encounter a somewhat larger selling pressure beginning from 10 trading days before the unlock day, but their average magnitude of trading volume reversal is also relatively stronger beginning from 10 trading days after TLED. 5.2 Cross-sectional Determinants of the Abnormal Volume and Return Next, we employ a cross-sectional regression analysis in order to identify possible determinants of IPO abnormal performance. To examine the joint effects of market behaviors and firm characteristics on the IPO performance around lockup expirations, we estimate the following two regressions by using the two-stage least square method (2SLS): 11 CAR = 0 + 1 ( IT ) + 0 2 Log( Size )+ 2 Log( 3 Log( CATV ) 3 (3) (4) Log( CATV ) = + 1 ( IT ) + Size )+ ( CAR ) The endogenous variables in Equation (3) and (4) are the CAR and the natural logarithm of cumulative abnormal trading volume (CATV), respectively, for each sample firm over the across-event windows [-20, 20], [-5, 5], the pre-event window [-20, -11], and the post-event window [11, 20]. Additional explanatory variables include an indicator variable whose value is set to one if the firm is in the information-technology industry category (IT), plus the natural logarithm of the market capitalization of firm equity (Size). The results are presented in Table 7. The 2SLS regression estimates of Equation (3) (see Panel A of Table 7) indicate that on average, the CARs are more negative for firms in the information technology sector. The negative coefficients for the IT dummy are statistically significant for event windows [-5, 5] and [-20, -11], though they are rather insignificant for window [-20, 20] or [11, 20]. That is, 10 trading days prior to and 10 trading days surrounding the lockup expiration, investors in IT-sector IPOs on average suffer a greater wealth loss compared with other IPO investors. The effect of firm size on abnormal return is also sensitive to the selection of specific event windows. We find weak evidence supporting a negative correlation between CAR and Size over the pre-event window [-20, -11]. This suggests that firms with larger capitalizations might suffer relatively larger value declines prior to the lockup expiration. Moreover, abnormal IPO price performance is also positively associated with abnormal trading volumes, as we find that the coefficients are significantly positive for three of the four event windows. Field and Hanka (2001) in their study report that abnormal returns were positively related to a firm size variable for the entire sample group they analyzed. The size variable was positive and significant for non-venture capital firms and negative and insignificant for venture capital firms. The regression results of Equation (4) (see Panel B of Table 7) show that IT, Size and CAR largely explain the abnormal trading intensity over various event windows (with the only 12 exception of the CAR variable for the interval [-20, -11]). Among all IPO issues, firms in the IT industry are relatively more heavily traded, as are larger firms both in the pre-and post-event periods. Based on such evidence over our sample period, we conclude that IPO firms in the IT sector follow a more violent trading pattern than the other IPO firms generally do. The returns are also abnormal and null Hypothesis 3 is rejected. . 6. Conclusions We find that in the Taiwan IPO market, IT-IPO investors suffer a significant wealth loss just prior to the lockup expiration. Non-IPO stocks are not similarly affected. In this pre-event period the selling pressure also considerably increases, for all classes of IPOs, as is indicated by abnormally heavy trading volumes. After the lockup expiration, the overall IPO market experiences a strong price reversal on average, again along with abnormally active trading. Such a phenomenon suggests that investors view price declines around lockup expiration dates as opportunities to repurchase the stocks. In comparison with other industries, IT firms in Taiwan follow a trading pattern that is relatively more sensitive to lockup expiration events, with larger magnitudes in both abnormal returns and abnormal transaction volumes. Firm-specific factors, such as the size of market capitalization, might also play a significant role in determining the magnitude of IPO market anomalies. We find that firms with larger capitalizations might suffer relatively larger value declines prior to the lockup expiration. Our findings in the Taiwanese IPOs differ from those reported in other studies mainly because the abnormal returns for Taiwanese IPOs are apparently smaller in size and transitory, as compared with their US and UK counterparts. 13 References Aggarwal, R., L. Krigman, and K. L. Womack, 2002, Strategic IPO Underpricing, Information Momentum, and Lockup Expiration Selling, Journal of Financial Economics 66, 105-137. Allen, F. and G. Faulhaber, 1989, Signaling by Underpricing in the IPO Market, Journal of Financial Economics 23, 303-323. Bradley, D., B. Jordan, I. Roten, and H. Yi, 2000, Venture Capital and IPO Lockup Expiration: An Empirical Analysis, Journal of Financial Research 24, 465-93. Espenlaub, S., M. Goergen, and A. Khurshed, 2001, IPO Lock-in Agreement in the UK, Journal of Business Finance & Accounting 28, 1235-1278. Field, Laura C. and Gordon Hanka, 2001, The Expiration of IPO Share Lockups, Journal of Finance 56, 471-500. Grinblatt, M. and C. Hwang, 1989, Signaling and the Pricing of New Issues, Journal of Finance 44, 393-421. Keasler, T. R., 2001, The Underwriter’s Early Lock-up Release: Empirical Evidence, Journal of Economics and Finance 25, 214-228. Krigman, L., W. H. Shaw, and K. L. Womack, 1999, The Persistence of IPO Mispricing and the Predictive Power of Flipping, Journal of Finance 54, 1015-1044. Mikkelson, W. H. and M. M. Partch, 1986, Valuation Effects of Security Offerings and the Issuance Process, Journal of Financial Economics 15, 31-60. Mikkelson, W. H. and M. M. Partch, 1988, Withdrawing Security Offerings, Journal of Financial and Quantitative Analysis 23, 119-134. Mohan, N. J. and C. R. Chen, Information Content of Lockup Provisions in Initial Public Offerings, International Review of Economics and finance 10, 41-59. Ofek, E. and M. Richardson, 2000, The IPO Lock-up Period: Implication for Market Efficiency and Downward Sloping Demand Curves, Working paper, New York University. Ofek, E. and M. Richardson, 2003, DotCom Mania: The Rise and Fall of Internet Stock Prices, Journal of Finance 58, 1113-1138. Ritter, J. R., 1991, The Long-run Performance of Initial Public Offerings, Journal of Finance 46, 3-27. 14 Welch, L., 1989, Seasoned Offerings, Imitation Costs, and the Underpricing of Initial Public Offerings, Journal of Finance 44, 421-450. 15 Table 1. Number, Trading Volume and Value for Securities listed on the TSE during the 1990-2001 Period.a Year 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 a Number of firms 199 221 256 285 313 347 382 404 437 462 531 584 Trading Volume (100 Million shares) 2,323.1 1,759.4 1,075.9 2,046.8 3,512.4 2,673.0 3,507.4 6,542.0 6,120.1 6,780.6 6,308.7 6,064.2 Trading Valueb (NT$100 million) 190,312.9 96,827.4 59,170.8 90,567.2 188,121.1 101,515.4 129,075.6 372,411.5 296,189.7 292,915.2 305,265.7 183,549.4 Sources: Stock Exchange Fact Book (various issues). The Taiwan Stock Exchange. b The exchange rate is averagely NT$34.5 per US dollar as of May, 2003. 16 Table 2. Financial Summary of IT and Non-IT Firmsa (in thousands) IT Firms (N=48) Book value of asset Book value of equity Market value of equity Revenue Years from establishment a. Non-IT Firms (N=79) NT$45,856,152.71 NT$6,280,202.46 NT$10,059,506.00 NT$7,744,141.58 23.26 NT$15,082,161.70 NT$9,277,502.30 NT$31,884,136.97 NT$13,962,802.93 12.36 This table presents the financial characteristics between IT and non-IT firms based on financial statements at the end of lockup year. 17 Table 3.a Return Performance Prior To the Expiration of Lockup Periods IT Firms (N=48) Average Abnormal Return -0.48% (-0.960) -0.65% (-1.159) Non-IT Firms (N=79) Average Abnormal Return 0.20%( 1.171) All Firms (N=127) Average Abnormal Return Cumulative Average Abnormal Return Cumulative Average Abnormal Return -0.48% (-0.960) Cumulative Average Abnormal Return 0.20%( 1.171) Event Days -20 -19 -18 -17 -16 -15 -14 -13 -12 -11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 -0.06% (0.334) -0.06% (0.334) -1.13% (-1.498) -0.16%(-0.811) 0.04%( 0.255) -0.35% (-1.352) -0.40% (-0.720) 0.04%( 0.300) -0.19% (-0.941) -0.60% (-1.131) -0.51% (-1.735)* -1.64%(-2.225)** 0.00%( 0.160) -0.51% (-0.827) -2.16%(-2.340)** -0.06%( 0.013) -0.02%( 0.266) -0.23% (-0.499) -0.83% (-1.229) -0.10% (0.004) 0.11% (0.085) 0.43% (1.078) 0.70% (1.958)* -0.74% (-1.513) -0.16% (-0.506) 0.26% (0.745) 0.80% (2.123)** -0.18% (-0.418) -0.19% (-0.416) -0.58% (-1.409) -0.16% (-0.647) -0.41% (-1.317) -2.25%(-2.091)** 0.07%( 0.382) -2.14%(-1.875)* -1.72% (-1.328) 0.01%( 0.412) 0.30%( 1.168) 0.05%( 0.409) 0.05%( 0.542) 0.35%( 0.943) 0.23%( 0.642) 0.01% (0.304) -0.82% (-0.963) 0.04% (0.377) -0.78% (-0.725) 0.35% (1.584) -0.43% (-0.073) 0.19% (0.669) -0.24% (0.169) -1.02% (-0.550) -0.12%(-0.678) -1.76% (-1.023) -1.92% (-1.131) 0.02%(-0.036) 0.08%( 0.313) 0.25%( 0.594) -0.27% (-0.959) -0.51% (-0.161) 0.33%( 0.662) -0.01% (-0.065) -0.52% (-0.173) -1.66% (-0.853) -0.42%(-1.771)* -0.10%( 0.097) -0.16% (-0.938) -0.68% (-0.448) -0.85% (-0.204) -0.15%(-0.568) -0.25%(-0.071) 0.21% (0.857) -0.48% (-0.181) -1.04% (-0.312) -0.49%(-1.369) -0.74%(-0.448) -0.37% (-1.337) -0.85% (-0.545) -1.22% (-0.412) 0.15%( 1.419) -0.59%(-0.052) 0.02% (0.863) -0.83% (-0.294) -1.80% (-0.762) -0.01%( 0.143) -0.60%(-0.014) -0.23% (-0.754) -1.06% (-0.479) -1.97% (-0.899) -2.38% (-1.192) 0.37%( 1.065) 0.03%( 0.179) -0.23%( 0.253) 0.17% (0.442) -0.89% (-0.353) -0.20%( 0.289) -0.14% (-0.669) -1.02% (-0.505) 0.00%( 0.295) -0.29% (-1.606) -1.31% (-0.869) 0.04%( 0.371) 0.20% (1.107) -1.11% (-0.592) -1.09%(-2.689)*** -3.47% (-1.792)* 0.20%( 0.060) 0.46% (1.330) -0.53% (-1.291) -0.41% (-1.013) -3.01% (-1.439) 0.04%( 0.367) -3.54% (-1.692)* 0.11%( 0.177) -3.94% (-1.872)* 0.17%( 0.177) 0.15%( 0.402) -0.13% (-0.654) -1.24% (-0.723) 0.33%( 0.431) -0.05% (-0.483) -1.29% (-0.811) 18 Table 3. (Continued) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 0.10% (0.411) -0.29% (-0.236) -3.84%(-1.741)* -4.12% (-1.752)* 0.28%( 1.310) 0.05%( 0.055) 0.08%( 0.413) -0.09%( 0.056) 0.04%( 0.102) 0.18%( 0.620) 0.34%( 1.437) -0.08%(-0.596) 0.14%(-0.356) 0.06%( 0.838) -0.25%(-0.710) -0.11%(-0.982) -0.08%(-0.086) -0.05%(-0.321) 0.05%(-0.106) -0.08%( 0.246) 0.29%( 1.119) 0.61%( 0.700) 0.66%( 0.696) 0.74%( 0.766) 0.65%( 0.761) 0.69%( 0.767) 0.87%( 0.871) 1.21%( 1.127) 1.12%( 0.997) 1.26%( 0.915) 1.32%( 1.051) 1.07%( 0.909) 0.95%( 0.724) 0.87%( 0.699) 0.82%( 0.634) 0.87%( 0.608) 0.80%( 0.640) 1.09%( 0.813) 0.21% (1.286) -0.08%(-0.101) -1.07%(-0.518) -1.15% (-0.528) -0.54% (-1.293) -4.67%(-1.979)** -0.35% (-1.232) -5.02% (-2.186)** -0.21% (-0.867) -5.23% (-2.313)** -0.07% (0.003) -5.31% (-2.269)** -0.17% (-0.398) -5.48% (-2.303)** 0.10% (0.806) 0.44% (1.007) 0.52% (1.172) 0.70% (1.992)** 0.07% (0.370) 0.13% (0.455) -0.42% (-1.066) -5.38% (-2.114)** -4.94% (-1.894)* -4.42% (-1.653)* -3.72% (-1.275) -3.66% (-1.191) -3.52% (-1.095) -3.94% (-1.260) -2.98% (-0.778) -0.15% (-0.470) -1.31% (-0.613) -0.19% (-0.714) -1.49% (-0.743) -0.06% (-0.452) -1.55% (-0.818) 0.08% (0.491) 0.15% (0.889) -0.01% (0.026) 0.25% (0.339) 0.23% (1.382) 0.11% (0.665) -1.47% (-0.708) -1.32% (-0.527) -1.33% (-0.513) -1.08% (-0.443) -0.85% (-0.187) -0.74% (-0.067) -0.05% (-0.547) -0.79% (-0.161) 0.00% (0.212) -0.79% (-0.122) -0.19% (-0.909) -0.98% (-0.274) 0.39% (1.629) -0.59% (- 0.001) 15 0.96%(2.785)*** 16 0.84% (2.711)*** -2.15% (-0.321) 17 18 19 20 a. 0.27% (1.861)* -0.32% (- 0.307) 0.36% (1.758)* 0.04% (0.588) 0.62% (1.072) 0.40% (0.899) 0.52% (0.992) 0.46% (1.424) 0.65% (1.650)* -0.39% (-0.749) -0.36% (-0.673) -1.68% (-0.086) -1.04% (0.179) -1.42% (0.059) -1.78% (-0.047) 0.53%( 2.601)*** 1.62%( 1.219) 0.58%(3.066)*** -0.11%(-0.691) 0.41%( 1.369) 1.51%( 1.094) 1.92%( 1.295) -0.21% (-1.005) 0.12% (0.666) This table presents the abnormal returns surrounding the lockup period expiration date t = 0. The z-statistics test the Abnormal return is computed as the difference between the observed and expected returns. Expected return is generated from the standard market model regression. equal to zero. null hypothesis that the average abnormal returns or the cumulative average abnormal returns are Z-statistics are in parentheses. *** Statistically significant at 1%.** Statistically significant at 5%.* Statistically significant at 10%. 19 Table 4.a Cumulative Average Abnormal Returns for Various Event Windows Event Window Across the event IT Firms (N=48) Non-IT Firms (N=79) All Firms (N=127) [-20, 20] [-10, 10] [-5, 5] [-1, 1] Prior to the event -1.78% (-0.047) -2.50% (-1.228) -3.43% (-2.667)*** -0.83% (-1.092) 1.92% (1.295) 0.99% (0.820) 1.29% (1.194) 0.57% (0.961) 0.52% ( 0.992) -0.33% (-0.108) -0.49% (-0.698) 0.04% ( 0.086) [-5, 0] [-10, -1] [-20, -11] After the event -2.14% (-2.297)** -1.62% (-1.262) -1.92% (-1.131) 0.93% (0.827) -0.17% (-0.094) 0.33% (0.662) -0.23% (-0.760) -0.72% (-0.850) -0.52% (-0.173) [0, 5] [1, 10] [11, 20] a. -1.70% (-1.727)* -0.48% (-0.198) 2.64% ( 2.815)*** 0.54% (0.862) 0.99% (1.227) 0.60% (0.771) -0.31% (-0.381) 0.44% ( 0.846) 1.37% ( 2.338)** This table represents the cumulative abnormal returns surrounding the lockup period expiration date t = Abnormal return is computed as the difference between the observed and expected returns. The z-statistics test the null 0. Expected return is generated from the standard market model regression. hypothesis that the cumulative average abnormal returns are equal to zero. Z-statistics are in parentheses. *** Statistically significant at 1%.** Statistically significant at 5%. * Statistically significant at 10%. 20 Table 5.a Abnormal Trading Volume Surrounding the Expiration of Lockup Periods IT Firms (N=48) Abnormal Trading Volume 0.34 (1.89)* 0.14 (0.85) 0.16 (1.19) -0.07 (-0.67) 0.00 (-0.03) 0.11 (0.74) 0.09 (0.59) 0.23 (1.45) 0.14 (0.98) 0.21 (1.14) 0.35 (1.81)* 0.27 (1.35) 0.29 (1.48) 0.20 (1.42) 0.20 (1.14) 0.35 (1.74)* 0.10 (0.74) 0.27 (1.36) 0.43(2.01)** 0.13 (0.88) -0.08 (-0.59) Cumulative Abnormal Trading Volume 0.34 (1.89)* 0.48 (1.48) 0.64 (1.46) 0.56 (1.08) 0.56 (0.93) 0.67 (0.97) 0.76 (0.98) 0.99 (1.09) 1.13 (1.13) 1.33 (1.20) 1.68 (1.37) 1.95 (1.43) 2.23 (1.49) 2.43 (1.53) 2.63 (1.56) 2.98 (1.63) 3.08 (1.60) 3.35 (1.64) 3.77 (1.73)* 3.90 (1.70)* 3.82 (1.61) Non-IT Firms (N=79) Abnormal Trading Volume 0.12(0.52) 0.38(1.38) 0.15(0.93) 0.28(1.07) 0.30(1.17) 0.16(0.69) 0.21(1.02) 1.02(1.33) 0.15(0.70) 0.22(1.06) 0.30(1.22) 0.13(0.65) 0.16(0.73) 0.37(1.12) 0.39(1.62) 0.22(1.26) 0.12(0.77) 0.14(0.90) 0.08(0.64) 0.32(1.59) 0.07(0.58) Cumulative Abnormal Trading Volume 0.12(0.52) 0.50(1.12) 0.65(1.15) 0.94(1.16) 1.24(1.24) 1.39(1.18) 1.60(1.20) 2.62(1.46) 2.77(1.42) 2.99(1.42) 3.29(1.46) 3.42(1.43) 3.58(1.45) 3.94(1.51) 4.34(1.59) 4.56(1.63) 4.68(1.62) 4.82(1.62) 4.90(1.62) 5.22(1.66)* 5.30(1.65)* All Firms (N=127) Abnormal Trading Volume 0.20 (1.28) 0.29 (1.59) 0.15 (1.36) 0.15 (0.88) 0.18 (1.13) 0.14 (0.92) 0.16 (1.18) 0.72 (1.49) 0.14 (1.01) 0.22 (1.47) 0.32 (1.88)* 0.18 (1.27) 0.21 (1.35) 0.31 (1.44) 0.32 (1.95)* 0.27 (2.03)** 0.11 (1.03) 0.19 (1.54) 0.21 (1.86)* 0.25 (1.81)* 0.02 (0.16) Cumulative Abnormal Trading Volume 0.20 (1.28) 0.49 (1.63) 0.65 (1.66) 0.80 (1.48) 0.98 (1.49) 1.12 (1.44) 1.28 (1.46) 2.00 (1.72)* 2.15 (1.69)* 2.36 (1.72)* 2.68 (1.82)* 2.86 (1.82)* 3.07 (1.88)* 3.37 (1.95)* 3.69 (2.05)** 3.96 (2.12)** 4.07 (2.11)** 4.26 (2.14)** 4.48 (2.18)** 4.72 (2.21)** 4.74 (2.17)** Event Days -20 -19 -18 -17 -16 -15 -14 -13 -12 -11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 21 Table 5. (Continued) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 0.09 (0.62) 0.05 (0.40) 0.13 (0.79) 0.03 (0.24) 0.09 (0.75) 0.08 (0.71) 0.10 (0.74) -0.03 (-0.25) 0.18 (1.37) 0.21 (1.70)* 0.08 (0.75) 0.17 (1.21) 0.31 (1.89)* 0.10 (0.90) 3.91 (1.59) 3.96 (1.56) 4.08 (1.56) 4.11 (1.54) 4.21 (1.53) 4.29 (1.53) 4.39 (1.53) 4.36 (1.48) 4.54 (1.50) 4.75 (1.53) 4.83 (1.53) 5.01 (1.55) 5.31 (1.61) 5.41 (1.61) 0.14(0.71) 0.33(1.01) 0.09(0.52) 0.18(0.94) 0.09(0.50) 0.44(1.33) 0.55(1.43) 0.82(1.18) 0.13(0.98) 0.23(1.18) 0.32(1.52) 0.10(0.75) 0.17(0.96) -0.01(-0.08) 0.07(0.51) 0.04(0.30) 0.22(1.20) 0.65(1.83)* 0.47(1.33) 0.13(0.76) 5.44(1.63) 5.77(1.62) 5.86(1.60) 6.04(1.60) 6.13(1.56) 6.57(1.56) 7.12(1.64) 7.94(1.77)* 8.07(1.77)* 8.30(1.78)* 8.61(1.82)* 8.72(1.81)* 8.89(1.80)* 8.88(1.77)* 8.94(1.76)* 8.98(1.74)* 9.21(1.76)* 9.86(1.85)* 10.33(1.91)* 10.46(1.91)* 0.12 (0.90) 0.22 (1.08) 0.11 (0.84) 0.12 (0.97) 0.09 (0.76) 0.31 (1.45) 0.38 (1.56) 0.50 (1.15) 0.15 (1.54) 0.22 (1.71)* 0.23 (1.67)* 0.13 (1.29) 0.22 (1.75)* 0.03 (0.39) 0.29(2.58)** 0.26 (2.25)** 4.86 (2.14)** 5.08 (2.11)** 5.19 (2.10)** 5.31 (2.08)** 5.40 (2.04)** 5.71 (2.03)** 6.09 (2.10)** 6.59 (2.20)** 6.73 (2.21)** 6.96 (2.23)** 7.18 (2.26)** 7.31 (2.27)** 7.54 (2.28)** 7.57 (2.25)** 7.86 (2.30)** 8.12 (2.33)** 0.65(3.31)*** 6.06 (1.73)* 0.63(2.88)*** 6.69 (1.84)* 0.61(3.00)*** 7.30 (1.95)* 0.63(3.06)*** 7.93 (2.04)** 0.58 (2.61)** 8.51 (2.10)** 0.32 (2.07)** 8.82 (2.12)** 0.37(2.65)*** 8.49 (2.40)** 0.64(2.75)*** 0.51 (2.17)** 0.20 (1.67)* 9.13(2.52)*** 9.64 (2.61)*** 9.84 (2.63)*** a. This table presents the cumulative abnormal returns surrounding the lockup expiration day t = 0. Abnormal daily trading volume is also calculated relative to each stock’s pre-unlock average daily trading volume, as: AVi ,T Vi ,T 1 150 Vi ,t 130 t 21 1 , where Vi ,T is the trading volume for stock i on day T . The t-statistics test the null hypothesis that the average abnormal trading volume equals zero. level. T-statistics are in parentheses. *** Significant at the 1% level. ** Significant at the 5% level. * Significant at the 10% 22 Table 6.a Cumulative Abnormal Trading Volume for Various Event Windows Event Window Across the event IT Firms (N=48) Non-IT Firms (N=79) All Firms (N=127) [-20, 20] [-10, 10] [-5, 5] [-1, 1] Prior to the event 8.82 (2.12)** 3.41 (1.46) 1.57 (1.20) 0.16 (0.35) 10.46 (1.91)* 5.31 (1.66)* 1.79 (1.12) 0.54 (1.13) 9.84 (2.63)*** 4.59 (2.12)** 1.71 (1.55) 0.38 (1.17) [-5, 0] [-10, -1] [-20, -11] After the event 1.19 (1.37) 2.57 (1.82)* 1.33 (1.20) 0.96 (1.18) 2.23 (1.47) 2.99 (1.42) 1.04 (1.75)* 2.36 (2.18)** 2.36 (1.72)* [0, 5] [1, 10] [11, 20] a. 0.30 (0.48) 42.59 (1.54) 4.07 (3.03)*** 0.91 (0.86) 3.00 (1.49) 2.17 (1.47) 0.68 (0.98) 57.91 (2.14)** 2.89 (2.76)*** This table presents the cumulative abnormal trading volumes surrounding the lockup expiration date t = 0. The t-statistics test the null hypothesis that the cumulative average abnormal returns are equal to zero. T-statistics are in parentheses. *** Significant at the 1% level. ** Significant at the 5% level. * Significant at the 10% level. 23 Table 7.a 2SLS Model of Abnormal Returns and Volume Around the Unlock Day Event Window (-20, 20) Event Window (-5, 5) Event Window (-20, -11) Event Window (11, 20) Variable Panel A: Equation (3): CAR as the endogenous variable Constant IT Log(Size) Log(CATV) Panel B: Equation (4): Log of CATV as the endogenous variable Constant IT Log(Size) CAR a. 0.13 (0.06) -0.38 (-1.53) -0.06 (-0.57) 0.14 (1.77)** 2.27 (1.37) -0.61 (-3.17)*** -0.14 (-1.67)** 0.11 (2.05)*** -1.33 (-0.70) -0.45 (-2.02)*** 0.03 (0.29) 0.09 (1.32) 0.51 (0.25) 0.07 (0.27) -0.07 (-0.70) 0.14 (2.03)*** -7.22 (-3.14)*** 1.35 (5.10)*** 0.78 (7.59)*** 0.18 (1.77)** -9.96 (-3.77)*** 1.46 (4.71)*** 0.83 (7.08)*** 0.30 (2.05)*** -8.13 (-3.38)*** 1.36 (4.92)*** 0.75 (6.99)*** 0.16 (1.32) -9.42 (-3.73)*** 1.44 (4.97)*** 0.81 (7.16)*** 0.23 (2.03)*** Sample is 127 Taiwan-incorporated firms conducting an IPO at the Taiwan’s main stock market during the period from January 1, 1995 to December 31, 1999. This table represents the 2SLS regression results where CAR and the log of cumulated abnormal trading volume (CATV), respectively, over 4 event windows (-20, 20), (-5, 5), (-20, -11), and (11, 20) are as the endogenous variable. parentheses. *** Statistically significant at 1%. ** Statistically significant at 5%. * Statistically significant at 10%. T-statistics are in 24 Figure 1. Return Performance Surrounding the Expiration of Lockup Periods Cumulative abnormal average returns (CAARs) for 127 firms that provide IPOs between January 1, 1995 and December 31, 1999, of which 48 firms are classified as the IT sector. as the first day when the 2-year lockup periods end. Event dates t = 0 are defined 3% 2% 1% 0% CAAR -1% -20 -2% -3% -4% -5% -6% Event Days All Firms IT Firms Non-ITFirms -15 -10 -5 0 5 10 15 20 25 Figure 2. Abnormal Trading Volume Surrounding the Expiration of Lockup Periods Cumulative abnormal trading volumes for 127 firms that provide IPOs between January 1, 1995 and December 31, 1999, of which 48 firms are classified as the IT sector. first day when the 2-year lockup periods end. Event dates t = 0 are defined as the 1.2 Abnormal Trading Volume 1 0.8 0.6 0.4 0.2 0 -15 -10 -5 0 Event Days 5 10 15 20 -0.2 -20 All Firms IT Firms Non-IT Firms 26 This document was created with Win2PDF available at http://www.daneprairie.com. The unregistered version of Win2PDF is for evaluation or non-commercial use only.