Covered Warrants in Taiwan, and Are Consistent with Our Hypothesis 3

Covered Warrants, Stock Returns and Trading Volumes: Evidence from

Taiwan

Chia-Ying Chan* and Ranko Jelic**

* London Metropolitan University, London EC3N 2EY, UK

**University of Birmingham, Birmingham Business School, Birmingham B15 2TT, UK

Corresponding author: [email protected]; tel. 44 (0) 121 4145991; fax. 44(0) 121 4146238

Abstract

Covered warrants are synthetic, rather than pure, financial derivatives listed on stock exchanges like any other listed security. We examine the Taiwanese IPO market for covered warrants, and impact of the warrants’ initiations on underlying stocks’ returns, systematic risk, volatility, and trading volume. The results suggest positive, and statistically significant, price and trading volume effects associated with warrant introductions. The effects are notably different for sub-samples of the first time and subsequently listed warrants. The results of Exponential Generalized Autoregressive

Conditional Heteroskedastic model (EGARCH) suggest that covered warrants introduction does not change the conditional distribution of underlying stock returns.

JEL classification: G; N2

Keywords: Covered Warrants, Synthetic Financial Derivatives, Stock Market in Taiwan

Acknowledgements

We would like to thank Mike Theobald, participants at the European Financial Management Association Meeting –

Doctoral Colloquium in London (2002), participants at seminars at University of Manchester (2003), Aston University

(2003), and University of Birmingham (2004), for their helpful comments. All remaining errors are ours.

Introduction

Covered warrants are options alike financial instruments, giving holders the right to buy or sell an asset at a specified price over a specified period. Unlike corporate warrants, covered warrants are not issued by companies on their own stocks, and no new stocks are issued upon exercise. The covered warrants, therefore, convey no information about the underlying companies’ future investment or financing policy. Unlike options, they are non-standardized contracts traded on stock exchanges like any other listed security.1 They also have much longer maturity (up to 2 years) than options. Finally, they contain a

‘covered agreement’ which stipulates that a certain amount of the underlying asset must be hold by the issuer to cover (hedge) his position. Due to the above differences, covered warrants are classed as synthetic, rather than pure, financial derivatives (e.g. options and futures).2 The status of synthetic derivatives suits trading bank mandates in many countries and allow them to issue and trade in warrants with different underlying assets

(e.g. stocks, market indices, commodities, interest rate futures, etc.) without complying to

1 In some markets they are also traded over the counter.

2 There is some confusion over terminology. For example, in the Hong Kong market

‘covered warrants’ are ‘issued by companies on their own shares and result in the issuance of new equity upon exercise’ while ‘ non-collaterized’ warrants are issued by third parties on a domestic company’s shares. Exercise of non-collaterized warrants results in re-allocation of existing shares rather than flotation of new corporate equity; (Banks, 1996).

In the HK literature the first type of warrants is called ‘equity warrant’ and the second type is called ‘derivative warrants.’ For the reminder of this paper, covered warrants will refer to those issued by a third party and as defined in the text.

3 sophisticated regulatory requirements and requiring a margin account. The banks normally target retail investors by selling smaller units of popular securities.

The initial public offering (IPO) market for financial derivatives, opposite to the IPO of equities received relatively little attention in the literature. One of possible reasons could be that after the unexpected introduction of ‘pure derivatives’ (e.g. traded options) the subsequent listing dates are common knowledge as they are predetermined (Chan and Wei

2001). Introductions of synthetic derivatives, however, are not predetermined, one-off events. Decisions about the warrants initiation, subsequent listings and different terms of the contracts (e.g. maturity) are at discretion of issuers (numerous investment banks rather than the exchange), subject to regulatory requirements. These decisions, therefore, are expected to produce some information (surprise) effect to investors.3 In this respect, the issues of covered warrants are more similar to IPOs and seasoned equity offerings (SEOs) of ordinary shares rather than to listings of ‘pure’ derivatives.

Leading international markets for covered warrants are in Germany, Switzerland, Hong

Kong, Taiwan, Australia, and Italy (Figure 1 about here). Germany is the biggest market with over 25,000 issued warrants. Notable absence of US and UK from the list could be explained by developed markets for pure derivatives, and the existence of many options alike financial instruments for retail investors in these countries. For example, US long term equity anticipation securities (LEAPS), and UK spread betting, together with

3 Throughout the paper terms introductions and initiations are used inter-changeable and correspond to warrants’ announcement dates as opposed to warrants’ listing dates.

4 contracts for differences (CFDs), are main covered warrants’ competitors.4 In spite of significant trading volumes and popularity of covered warrants there is a paucity of studies examining these contracts and their interaction with underlying shares. This is compounded further by a paucity of literature examining emerging markets for financial derivatives.

Figure 1 about here

In Taiwan, the government’s liberalization of the local markets and the internationalization of local corporations created conditions for the development of a derivative markets in

1990s. Initially, warrants on Taiwanese companies and the market index were issued in the offshore markets by leading investment banks such as Union Bank of Switzerland, Carr

Indosuez Asia, and Bankers Trust International. The Taiwan Stock Exchange (TSE) started to accept applications for the issue of covered warrants in June 1997, and the market has grown rapidly since. With 18 issuing banks, and more than 270 covered warrants, the market is the second largest in Asia (behind Hong Kong).5 The warrants are predominantly

American style, on a single underlying stock, with maturity of one to two years, where parties can choose either cash settlement or physical stock delivery.6 Exercise price, maturity, and other terms of covered warrants differ among issuers, over time. This further

4 The UK covered warrants market, however, opened in 2002, and recently reached 460 listed warrants issued by more than 40 banks.

5 This includes warrants on a single underlying stock, basket of securities, and other types of (exotic) covered warrants. Leading issuers are leading domestic (e.g. Yanda, Jihsun, and

Grand Cathay) and some foreign banks (Merrill Lynch).

6 Since 2003, the minimum maturity for warrants is 6 months.

5 implies existence of subsequently issued warrants, on the same underlying asset, by a number of different issuers. One of the interesting features of the Taiwanese market is that both, underlying stocks and covered warrants are listed and traded on the same stock exchange. The exchange, therefore, regulates issues and trading in both warrants and underlying assets. This is different from other covered warrant markets (e.g. Frankfurt) where warrants are, predominantly, traded over the counter.

Previous studies on Taiwanese covered warrants report apparent miss-pricing of covered warrants (Lee and Hsieh, 2000; Chung et al. 2002). Mixed evidence was reported on lead lag relationship between stock and warrants prices (Zane 1998; Young 1999; Lee et al

2000), and effects of warrants’ trading volumes on underlying shares volatility (Wang

2003; Yu and Tsai 2003). Hsu and Huang (2003) report a positive price effect on 9 underlying shares for covered warrant introductions during 2000-2001 period. The objective of this paper is to examine the impact of covered warrant introductions on underlying stocks’ returns, systematic risk, volatility, and trading volume, from the market’s inception in August 1997 to February 2003. We examine a larger sample, over a longer period, than any of previous studies on Taiwanese covered warrants market.7 We also contribute to the literature by applying more robust tests for post-listing changes in stock volatility and trading volumes. We model conditional distribution of the underlying

7 The above mentioned papers, except for Chung et al. (2002), were translated from

Chinese. The translated papers are, except for Yu and Tsai (2003), based on very limited samples. For example, Hsu and Wang (1999) examine 9, Zane (1999) examines 18, Young

(1999) examines 3, and Wang (2003) examines 20 companies.

6 stock returns as the EGARCH process, and measure changes in trading volumes by both, trading volume ratios and abnormal trading volumes. Finally, we investigate and compare effects of first and subsequent introductions on underlying shares.

We expect that differences between options and covered warrants, as well as institutional characteristics of Taiwanese market for covered warrants, may lead to different results from those reported in the previous literature on option and covered warrants listings and/or listing announcements (introductions) in other countries (Bollen, 1998;

Hernandez-Trillo, 1999; Draper, 2001; Chan and Wei, 2001; Chen and Wu, 2001).8 In particular, we expect a stronger price effect for stock exchange-listed covered warrants, than for derivative exchange traded options and OTC covered warrants, reported in previous literature. We also expect different price and trading volume effects for first and subsequent covered warrants’ listings.

The reminder of the paper is organized as follows. In the next section we discuss literature and hypotheses. In chapter 3, we discuss sample and methodology. The results are presented in section 4. The final section presents conclusions and suggests some areas for further research.

8 Draper and Chen and Wu examine effects around introduction dates, Bollen and

Hernandez - Trillo examine effects around listing dates, while Chan and Wei examine both, introduction and listing dates.

2. Previous research and hypotheses

Price effect

Numerous studies examined the price effect of the option listing on the underlying stocks.

Studies based on data from US, Swiss, and Norwegian markets report a positive permanent and statistically significant price effect.9 On the other hand, Watt et al. (1992), and Hamill et al. (2002), suggest a temporary positive price effect prior to, but negative price effect following the UK option listings. The negative price effect was reported for US (Bollen

1998, and Sorescu 2000) and Dutch markets (Kabir, 1997). 10 Finally, the findings differ depending upon whether they are based on the announcement date or the listing date. For example, majority of studies utilizing US data seem to suggest that the price effect is associated with the listing date but not with the announcement date.

The empirical evidence on price effect of covered warrants’ listings is scarce. Draper et al.

(2001), Chen and Wu (2001), and Chan and Wei (2001) all find positive price effect prior the covered warrant introduction, and negative price effect after the warrants introduction in Hong Kong.

Different trading and regulatory mechanisms from option exchanges, together with major role of issuers, before and after warrant listings, are main reasons why we expect

9 For US results see Skinner (1989), Conrad (1989), Branch and Finnerty (1981),

DeTemple and Jorion (1990), and Kim and Yong (1991). For Swiss market see Stucki and

Wasserfallen (1994). For Norwegian market see Gjerde and Settem (1995).

10 Sorescu reports mixed results: positive price effect from 1973-1980, and negative price effect since 1981.

8 significant, and more conclusive, evidence on price effects around warrants’ listings. For example, issuers would normally operate in the market for the underlying stocks before announcement of the warrant issue. The issuer might take a long position in underlying asset for two reasons: hedging purposes, and in attempt to drive the price up in order to achieve a better warrant issue price. Other investors might receive this private information

(information leakage) and start trading before the covered warrants issuance in order to gain the payoff from the reaction of the non-informed traders.11 In both cases, the positive pre-listing effect is expected. The issuers also actively participate in the secondary warrant market with a view to provide price support for their recently issued warrants.

In the primary market, banks compete by issuing different classes of warrants on the same underlying stocks. The subsequent issues sometimes sell at a premium compared to first time introductions.12 The ‘premium’ could be due to the fact that some warrants could be issued in balk volume and investors are thus able to buy a large amount at better terms than buying existing warrants from the secondary market (Chan and Wei 2001). We, therefore, examine price effects for the first and subsequent covered warrants listings on the same underlying stocks.

The literature on options listings (e.g. DeTemple and Jorion, 1990) suggests that early option listings might generate different results from later option listings. Similarly, Draper

11 Similarly, other market participants could anticipate that issuers may want to acquire the underlying stocks prior to delisting (exercise).

12 Lee and Hsich (2000) report that the offer prices for some warrants seem to be much higher than either historical or post-listing volatility would suggest.

9 et al. (2001) report a stronger price effect for the first time listed warrants in comparison with the subsequent listings on the same underlying stock in Hong Kong. We, therefore, test the following hypotheses:

H1: There is a price effect on underlying stocks around covered warrants announcement dates.

H2: The price effect is different for the sub-sample of first time listed covered warrants then for the sub-sample of subsequently listed warrants.

Volatility

The early empirical evidence on effects of US option listings on volatility of underlying assets is extensive and seems to conclusively suggest a reduction in volatility after the listings (Damodaran and Subrahmanyam, 1992). The results were echoed in studies for other option markets across the world (e.g. UK, Sweden, Finland, Japan).13 The reported results are explained by improved market completeness and information discovery, associated with the option markets, which ultimately result in better pricing efficiency for the underlying assets. The results of more recent US studies, however, are less conclusive.

Lamoureux and Panikkath (1994), Freund et al. (1994), and Bollen (1998) all find a lack of significant reduction in volatility. Mayhew and Mihov (2000) find that stock volatility increased in recent US option introductions. According to the authors, the increase in

13 For UK see, Watt et al. (1992), for Sweden see, Alkeback and Hagelin (1998), for

Finland see, Sahlstrom (2001), for Japan see Becchetti (1996) and Kumar et al. (1995).

10 volatility is due to the fact that stock exchanges deliberately choose to issue options on shares for which they anticipate an increase in volatility. Finally, evidence from studies which allow for systematic variation in expected returns and volatility suggest absence of any significant changes (Pierre, 1998).

Empirical evidence for covered warrants is inconclusive. Chen (1999) finds an increased stock volatility, whereas Draper et al. (2001) document lack of the significant effect on the stock volatility after covered warrant introductions in Hong Kong market. Hernandez-

Trillo (1999) investigated Mexican covered warrants and reports insignificant changes in conditional volatility, modeled by the GARCH model.14 We, therefore, test the following hypothesis:

H3: There is no significant change in volatility of underlying stocks after covered warrants announcement dates.

Systematic risk

The empirical evidence on changes in systematic risk after option listings is inconclusive.

Elfakani and Chaudhury (1995), and Gjerde and Settem (1995) report decrease in betas after option listings in Canada and Norway, respectively. On the other hand, numerous studies report lack of statistically significant changes in systematic risk for US underlying

14 It is worth pointing out that Hernandez-Trillo’s sample consists of both, stock and market index warrants.

11 stocks.15 The warrants listing, however, can increase stocks’ popularity and information transparency. Information should, therefore, be reflected in stock prices more quickly, especially where warrants and stocks are traded in the same market. The information flow between warrants and stocks may ultimately result in less clustered volatility over time and fewer outliers in the stocks’ return distribution. The above factors could affect the covariance risk and, ultimately, betas of the underlying shares. We test the following hypothesis:

H4: Systematic risk of underlying stocks, measured by ß, changes after covered warrants announcement dates.

Trading volume

Early empirical evidence of change in trading volume to option introduction in the U.S. seems to suggest lack of change in the market-adjusted trading volumes (Damodaran and

Lim, 1991). Some of more recent US evidence, however, suggests increase in trading volumes after option listings (Shastri et al., 1996; Wei et al., 1997). Gjerde and Setten

(1995) report similar results for Norwegian market, whilst Chamberlain et al. (1993) observe lack of any significant changes in trading volumes in Canada. Finally, Kumar et al.

(1995) report decrease in trading volumes in Japan.

15 See, Trennepohl and Dukes (1979), Klemkosky and Maness (1980), Whiteside et al.

(1981), Whiteside et al. (1983), Bansal et al. (1989), Conrad (1989), Skinner (1989),

Hadad and Voortheis (1991), Freund et al. (1994), and Schinski and Long (1995).

Covered warrants are expected to create better choice for investors and induce more hedging related trades from issuers, especially in the absence of any similar contracts. On the other hand, the availability of warrants, in an emerging market such as Taiwan, can move trades away from the underlying stocks towards warrants and, therefore, reduce trading volumes. The empirical evidence from other markets in the region suggests an increase in trading volumes around warrant listings. For example, Chen and Wu (2001),

Chan and Wei (2001), and Draper et al. (2001) find increase in trading volume around the warrant introduction in Hong Kong. Chan and Wei (2001) report an increase in trading volume on the warrants announcement days, due to increase in trading activities by both investors and issuers. The trading activities, however, drop during the first trading day. The effect on volume, therefore, was only temporary. There is also a difference in results for first time listings and subsequent listings. The results for the first time listed warrants seem to be consistent with the US early evidence of no change in the market-adjusted trading volumes (Draper et al., 2001). We, therefore, test the following hypotheses:

H5: The trading volume of underlying stocks changes around covered warrants announcement dates.

H6: The changes in trading volume of underlying stocks for first time covered warrants listings are different from the changes associated with the subsequent listings.

3. Data and methodology

Data

The Securities and Futures Commission (SFC) sets strict requirements for the issuance of covered warrants. For example, to qualify as an underlying stock, the companies have to be listed at the Taiwanese Stock Exchange (TSE) with a market value of at least TW$15 billion and the recent trading volume of at least 20% of total issued capital accumulated in the secondary market.16 The underlying companies should be of at least ‘BBB’ credit rating and with rather dispersed shareholders structure.17 The issuers are required to establish hedging positions in the underlying shares, convertible bonds issued by the underlying companies, or in covered warrants issued by the other issuers. The other relevant aspects of regulatory requirements are: i) exercise ratio of the covered warrants is, normally, 1 warrant to 100 shares; ii) issuers are required to sell out 80% of the issued amount to the market participants before the warrant listings; iii) the lowest issuing amount is 20 million units, and a single investor cannot hold more than 10 % of the total amount.

The TSE reviews the requirements every 3 months, and reveals both, lists of banks that qualify as issuers and lists of companies that qualify for underlying stocks.

We obtained a list of all covered warrants, during August 1997 to February 2003, from

TSE and Taiwanese Economics Journal’s (TEJ). Out of the population of 278 warrants,

222 were call covered warrants with a single stock as underlying asset.18 For the warrants’

16 The TW$/US$ exchange rates was 34.42, in 2003.

17 For example, they should have more then 10,000 shareholders or at least 5,000 shareholders holding between 1,000 and 50,000 shares.

18 Other warrants were on market index and on various baskets of securities.

14 announcement dates we use the dates of the SFC’s announcement of approved listings.19

We check for any official announcements made by the companies during the event window, on the TSE’s website. The companies with announcements of any important information, other than warrant issues, were removed from the sample.

We also require the underlying stocks to have daily price data for 261 days (from 150 days before introduction through 100 days after) available from TEJ database. In cases when two or more warrants, on the same underlying stock with similar exercise price and expiration date, were issued during the same trading month we selected only the earliest introduction to avoid overlaps in observations during the event window. The adjustments resulted in our sample of 146 call covered warrants, on 58 underlying stocks (Table 1). The sample, therefore, contains 58 first-time, and 88 subsequently listed call covered warrants.

Table 1 about here

Although the market value of traded warrants represents less than 1% of the market value of all TSE stocks, the respective percentage for trading volume is much higher. This indicates much higher trading volume for the warrants, after adjusting for the difference in market capitalization, than for other securities traded on the TSE. The largest sample

19 The alternative dates could have been dates of prospectus or listing dates. Our announcement dates are the earliest announcement dates and correspond to the event dates used in similar studies (e.g. Draper et al. 2001).

15 company’s market value was TW$ 631.89 billion and the smallest was TW$ 9.96 billion.20

The sample underlying stocks come from 13 different industries. Given the importance of high-tech industry in Taiwan, it should not be surprise that majority of sample companies are high tech stocks (74% of the sample). These companies are among the largest and with the most dispersed ownership structure. 21

Methodology

Returns and systematic risk

An event study methodology, based on the market model, is applied to examine the abnormal returns before and after the warrant listings (see MacKinlay, 1997). An estimation period was set as 150 days prior to introduction, whilst an event window was set as 10 days before and after warrants introduction. Abnormal returns are calculated as a difference between the actual and the market model returns. The abnormal returns are then cumulated over the event period to calculate the cumulative abnormal returns. The market model is also used to test whether systematic risk of underlying shares changes after covered warrant listings. The estimates of beta, generated by the market model, were subjected to Chow’s (1960) test for equality of the betas, before and after listings.

20 Based on the stock exchange statistics in year 2003.

21 The dispersed ownership is mostly due to the popularity of employee stock ownership plans (ESOPs) in Taiwan.

Volatility

Early studies on options listings assume constant expected returns and volatility of underlying stocks (Trennepohl and Dukes 1979; Bansal et al 1989; Jennings and Starks

1986; Detemple and Jorion 1990; Long et al. 1991). However, evidence suggests that both expected returns and volatility change systematically over time and that time varying returns and volatility need to be controlled for (Christie 1982; French et al. 1987; Conrad et al. 1989; Damodaran and Lim 1991). By examining how warrants’ listings affect the conditional distribution of security returns might help us understand changes in unconditional returns distribution. Previous studies on covered warrants utilize GARCH model to investigate the alternation of stock return conditional volatility (Hernandez-Trillo,

1999; Draper, 2001). In this study we employ the following EGARCH model:22

pqεε2 ′22t−1t−1equation 1 ln(σγti) =+ln( *ν) ∑∑βlogσt−1 +(αj−+γj) ij==11σπtt−−11σ

2 Where, ln(σ t′ ) represents return conditional volatility, ln(γ *ν ) represents long-term

2 average volatility, β logσ t−1 is the predicted conditional variance from last period,

q q εt−1 εt−1 2 represents scale of the information effect, and (γ ) represents sign ∑(α j − ) ∑ j j=1 σ t−1 π j=1 σ t−1 of the information effect.

EGARCH model has several advantages over GARCH model. First, there is no need to impose any non-negative constraints on the model parameters like in GARCH models.

22 The model is adopted from Nelson (1991), as presented in Petterson (2000).

Second, GARCH tend to enforce a symmetric response of volatility to positive and negative information shocks. It is, however, possible to have an asymmetric response to information shocks, where negative shocks increase volatility more than positive shocks of the same magnitude. For example, the negative shocks tend to be associated with a fall in the value of a firm’s stock and, therefore, trigger an increase in the firm’s debt/equity ratio

(leverage effect). This leverage effect ultimately leads to a higher risk for shareholders and an increase in volatility.

Trading volume

We examine daily abnormal trading volumes, during 21 days event window. 23 First, we calculate the daily stock turnover for each of underlying shares as:

TOit = Number of shares traded it / Number of shares outstanding it equation 2

The average estimated daily turnover for each of underlying shares, during estimation period is,

−11 equation 3 TOi= ∑ TOit/150 t=−160

The cross sectional daily turnover for day‘t’ as the average turnover for all underlying shares during the estimation period and the event window is,

23 We adopt the method introduced by Michaely et al. (1995), for examination of price reaction to dividend initiations and omissions.

1 N TO it equation 4 TOt = ∑ N i=1 TOi

The abnormal trading volume for each underlying share during the estimation period and the event window, in percentage terms is,

AVit = TOt - 1 equation 5

AV is the average cross sectional abnormal trading volume across the sample stocks during the estimation period is,

−11 ∑ AVt AV = t=−160 equation 6 150

The estimated standard deviation during the estimated period is,

−11 1 2 S.D. (AVt) = ∑ (AVAt − V) equation 7 149 −160

Finally, T-statistics is,

AV T = t equation 8 S.D. (AVt)

We also measure and compare trading volume ratios 100 days before (RVi,B ), and 100 days after (RVi,A ) listing 100 days before and after the warrants introduction. This method allows us to check for robustness of our results and to compare the results with the results reported in previous literature. The ratios are calculated as the average ratio of firms’ daily trading volumes (Vi, t-s and Vi, t+s) divided by total daily trading volumes of the market (TVm,

24 t-s and TVm, t+s ):

RVi,B = ( Vi, t-s/ TVm, t-s) *100 equation 9

RVi,A = ( Vi, t+s/ TVm, t+s) *100 equation 10

The Wilcoxon test is adopted to test for the differences in trading volumes before and after the introductions of covered warrants.

4. Results

The results reported in Table 2 suggest positive and statistically significant mean sample cumulative abnormal returns (CARs) prior to introduction (days –7, -6, -5, and –1).

Predominantly positive abnormal returns (ARs) before the announcements change to negative and statistically significant mean ARs after announcements (days 0, 1, 2, and 3).

CARs remain positive after announcements but statistically significant only on the day of announcements and on the following day. By closer examination, however, differences between two sub-samples become to emerge. For example, mean ARs before first time introductions are bigger than before subsequent introductions. The same holds true for the mean CARs. For example, CARs for the first time introduction are positive and statistically significant before the announcements (days 6, 5, 2, and 1), whereas mean

CARs for the subsequent introductions exhibit statistical significance only one day prior to the announcements. After announcements, average CARs for the first time introductions

24 Adopted from Draper et al. (2001).

20 remain positive (and statistically significant in days 1 and 2), while mean CARs for the subsequent introductions turn negative (but not statistically significant) (Table 2).

Table 2 about here

The positive average CARs for the total sample, as well as for two sub-samples during entire event window (Table 3 – Panel A), lend support to our hypothesis 1. The positive and significant CARs, for the total sample, are consistent with results reported in previous studies on option listings (Gjerde and Settem, 1995; Skinner, 1989; Mendenhall and Fehrs,

1999; Kim and Young, 1991; and Watt et al.,1992), but contradict the results on option introduction (Conrad 1989; DeTemple and Jorion, 1990). The results are consistent with

Chan and Wu (2001), Chan and Wei (2001), and Draper et al. (2001), who report positive price effect prior to the warrant introductions and negative after the introductions. Our results on positive cumulative abnormal returns throughout the event periods are consistent with the results reported for Hong Kong’s warrants (Draper et al. 2001).

The results of parametric and non-parametric tests for difference in the sample average

(mean and median) ARs and CARs, before and after announcements, are positive and statistically significant at 5% level or better. The differences between pre and post average

CARs are more evident in the sub-sample with subsequent introductions (results are statistically significant at 1% level or better) (Table 3 – Panel B). The above results highlight differences between two sub-samples and lend support to our hypothesis 2.

Table 3 about here

The results reported in Table 4 show the impact of covered warrant listings on changes in conditional volatility for underlying stocks. Measured by the difference in numbers of companies with the significant coefficient for the lagged conditional variance pre and post listings, the forecast ability of the post conditional volatility has only slightly improved

(increase from 72 to 76 companies). The results also indicate presence of the leverage effect of warrant listings (increase from 28 to 40 companies), and only a marginal changes in the sign of the effect (increase from 12 to 15 companies).25

Table 4 about here

Overall, the results of the EGARCH model indicate a lack of permanent changes in stock return variances after introduction of covered warrants in Taiwan, and are consistent with our hypothesis 3. This result is consistent with Pierre (1998), who also utilizes EGARCH model, and reports that conditional distribution of security return is unaffected by option introduction in the U.S. Similarly, Hernandez-Trillo (1999), utilizes GARCH model, and reports that introduction of covered warrants does not reduce Mexican stock return volatilities.

25 Given the importance of companies from high tech industry in Taiwan, we compare the results for the companies in high-tech industry with the results for the companies in other industries. The results for both, systematic risk and for conditional volatility do not seem to be statistically different from the results for the rest of the sample.

The results on changes in beta values after covered warrants introductions suggest that 12 companies experience statistically significant reduction in beta, while 10 companies experience a statistically significant increase in beta after warrants listings (Table 5). We find no significant differences in results for the first time and subsequent listings. Overall, the results do not lend support to our hypothesis 4. This is consistent with Yu and Tsai

(2003) who report a mixed evidence for 125 Taiwanese covered warrants, and Skinner

(1989) and Conrad (1989) who report lack of statistically significant changes in systematic risk after the introduction of options in US. The results, however, contradict Elfakaniand and Chaudhury (1995), and Gjerde and Settem (1995) who report an overall reduction in systematic risk in Canada and Norway, respectively.

Table 5 about here

The results reported in Table 6 suggest the existence of significant abnormal trading volume for the sample companies during the event window. The sub-sample with first time introductions, however, exhibits statistically significant daily trading volumes both before and after introductions. The sub-sample with subsequent announcements exhibits lack of statistically significant daily abnormal trading volumes from days 2-8 after the listing.

Table 6 about here

The average daily abnormal trading volumes are higher in the first time introduction than in subsequent introductions sub-sample during the entire event period (Table 7-Panel A).

The difference, in means and medians, is statistically significant at 1% level. The positive and statistically significant abnormal trading volumes are persistent before listings in both subsamples. The average abnormal trading volume, however, significantly drops after the announcements. The results in Table 7 – Panel B indicate reduction in sample average

(mean and median) daily trading volume (significant at 6% and 7%, respectively) after the introductions. The differences are more evident in the sub-sample for the subsequent listing (significance levels of 3% and 1% for mean and median, respectively). Overall, the results are consistent with our hypotheses 5 and 6, and with results of Draper et al (2001) and Chen and Wu (2001).

Table 7 about here

The relative trading volume ratios (both before and after listings) seem to be higher for subsequent listings (Table 8 – Panel A). The relative trading volume ratios for the sample, as well as two sub-samples, drop after introductions.26 The difference in the trading volume ratios, before and after introduction, is statistically significant for the total sample

(at 7% level for changes in median) and subsequent listings (at 10% level for changes in mean). A direct comparison of the two sub-samples (Table 8 – Panel B) suggests, however, that the difference is only statistically significant for median trading volume ratios after introductions (at 6% level of significance). This supports our results based on AVs.

Table 8 about here

26 This is consistent with recorded drop in average (mean and median) daily abnormal trading volumes reported in Table 7.

5. Conclusions

Our results suggest positive, and statistically significant, price and trading volume effects associated with the sample warrant initiations. The effects are notable different for sub-samples of the first time and subsequently introduced warrants. The differences between pre and post introduction’s average returns are more evident in the sub-sample with subsequent listings. After the listings, mean CARs for the first time listings remain positive and statistically significant, while mean CARs for the subsequent listings turn negative. The sample average abnormal trading volumes significantly drop after the listings. The drop in volume is, again, more evident in the sub-sample for subsequent listings. Finally, the results of EGARCH model suggests that covered warrants initiations do not change the conditional distribution of underlying stock returns.

The results provide investors with insights into underlying stock price and trading volume changes around warrants introductions. Prices increase few days before introductions and drop after. The price effect is present for both, first and subsequent issues, but ARs are bigger for the first time issues. Our results contradict findings, for US data, which suggest that the price effect is associated only with the listing dates and not with the listing’s announcement dates.

The ‘information’ effect is weak and only temporarily, and the prices increase only up to the first day after the warrants introductions (up to two days for the first time issues). The positive returns before introduction could be explained by trading (hedging) activities of warrant issuers. The positive returns immediately following the introduction suggest some

25 surprise (information) effect of the issues.27 The shift from the positive to negative returns could be due to disappearance of excess demand created by issuers who building up hedging positions before the issue. The positive effect on trading volume of the sample underlying shares is present trough the event window. The effect is, however, more prevalent for first time issuers. It starts ten days prior to the announcement and continues up to ten days after the announcements. For the subsequent issues, the positive effect on trading volume starts seven days before the announcement and finishes immediately, on the first day, after the announcements. The changes in trading volume, measured by both daily abnormal volumes and trading volume ratios, are more evident for subsequent issues.

The covered warrant market have provided important investment and educational roles for retail investors, and paved the way to introduction of Taiwanese option market in January

2003. An interesting extension of our work would be to investigate whether and how the interaction between the markets changes after the introduction of options market in 2003.

27 Chan and Wei (2001) also suggest a possibility of market manipulation by issuers by bidding up the price during the introduction date.

List of References

Alkebäck, P., Hagelin, N., 1998. The Impact of Warrant Introduction on the Underlying

Stocks, with a Comparison to Stock Options. Journal of Futures Markets 18, 307-328.

Banks, E., 1996. Asia Pacific Derivative Markets, Macmillan, London.

Bansal, V.K., Pruitt, S.W., and Wei, K.C.J., 1989. An Empirical Reexamination of the

Impact of CBOE Option Initiation on the Volatility and Trading Volume of the Underlying

Equities: 1973-1986. Financial Review 24, 19-29.

Becchetti, L., 1996. The Effect of Bond plus Equity Warrant Issues on Underlying Asset

Volatility: An Empirical Analysis with Conditional and Unconditional Volatility Measures.

Applied Financial Economics 6, 327-335.

Bollen, N.P.B., 1998. A Note on the Impact of Options on Stock Return Volatility. Journal of Banking and Finance 22, 1181-1191.

Branch, B., and Finnerty, J.E., 1981. The Impact of Option Listing on the Price and

Volume of the Underlying Stock. Financial Review 16, 1-15.

Chamberlain, T.W., Cheung, C.S., and Kwan, C.C.Y., 1993. Options Listing: Market

Liquidity and Stock Behavior: Some Canadian Evidence. Journal of Business Finance and

Accounting 20, 687-698.

Chan, Y.C., and Wei, K.C.J., 2001. Price and Volume Effects Associated with Derivative

Warrant Issuance on the Stock Exchange of Hong Kong. Journal of Banking and Finance

25, 1401-1426.

Chang, P.Y., Wu, C.L., and Lee, S.F., 1999. Impact of Financial Crisis on Taiwan from

South East Asian Financial Crisis. Chung Yuan University (translated from Chinese).

Chen, K.C., 1999. Pricing U.S. Exchange-traded Hong Kong Index Warrants, in: Flowers,

E. et al. (ed.), Interlocking Global Business Systems: the Restructuring of Industries,

Economics and Capital Markets, 93-119.

Chen, K.C., and Wu, L., 2001. Introduction and Expiration Effects of Derivative Equity

Warrants in Hong Kong. International Review of Financial Analysis 10, 37-52.

Christie 1982

Chow, G.., 1960. Tests of Equality Between Sets of Coefficients in Two Linear Regressions,

Econometrica 28, 591-605.

Chow, E., Lee, J,H. Liu, Y.J., and Chen, L,W., 2000. The Relationship between the Price of the Warrant and the Underlying Stocks in Taiwan. Development of Securities Market 12,

109-146 (translated from Chinese).

Chung, H., Lee, C.S., and Wu, S., 2002. The Effect of Model Errors and Market

Imperfections on Financial Institutions Writing Derivative Warrants: Simulation Evidence from Taiwan. Pacific-Basin Finance Journal 10, 55-75.

Conrad, J., 1989. The Price Effect of Option Introduction. Journal of Finance 44, 487-498.

Damodaran, A., and Lim, J., 1991. The Effects of Option Listing on the Underlying Stocks

Return Processes. Journal of Banking and Finance 15, 647-664.

Demodaran, A., and Subrahmanyam. M., 1992. The Effects of Derivative Securities on the

Markets for the Underlying Assets in the United States: A Survey. Financial Markets,

Institutions and Instruments 1, 1-21.

DeTemple, J., and Jorion, P., 1990. Option Listing and Stock Returns: An Empirical

Analysis. Journal of Banking and Finance 14, 781-801.

Draper, P., Mak, B.S.C., and Tang, G.Y.N., 2001. The Derivative Warrant Market in Hong

Kong: Relationships with Underlying Assets. Journal of Derivatives 2, 72-83.

Elfakhani, S., and Chaudhury, M., 1995. The Volatility Effect of Option Listing: Some

Canadian Evidence. Quarterly Review of Economics and Finance 35, 97-116.

French et al. 1987

Freund, S.P., McCann, D., and Webb, G.P., 1994. A Regression Analysis of the Effect of

Option Introduction on Stock Variance. Journal of Derivatives 1, 25-38.

Gjerde, Ø., and Sattem, F., 1995. Option Initiation and Underlying Market Behavior:

Evidence from Norway. Journal of Futures Markets 15, 881-899.

Haddad, M.M., and Voorheis, F.L., 1991. Initial Option Trading and Security Risk and

Return. Journal of Business Finance and Accounting 18, 903-913.

Hamill, P.A., Opong, K.K., and McGregor. P., 2002. Equity Option Listing in the U.K.: a

Comparison of Market-based Research Methodologies. Journal of Empirical Finance 9,

91-108.

Hernandez-Trillo, F., 1999. Financial Derivatives Introduction and Stock Return Volatility in an Emerging Market Without Clearinghouse: The Mexican Experience. Journal of

Empirical Finance 6, 153-176.

Honk Kong Fact Book, 2003. Turnover of Derivative Warrants by Major Markets. WFE

Monthly Statistics.

Hong Kong Stock Exchange, 2004. Hong Kong, World’s No. 2 Warrant Market. The

Exchange, 40-58.

Hsu, C.G., Huang, S.W., 2003. Impact of listing, issuance and expiration of covered warrant in Taiwan on the Underlying share. Industrial Finance 199, 77-92. (Translated from Chinese).

Jennings, R. H., Starks, L.T., 1986. Earnings Announcements, Stock Price Adjustment and the Existence of Options Market. Journal of Finance 41, 43-61.

Kabir, R., 1997. New Evidence on Price and Volatility Effects of Stock Option

Introductions. Working paper, Tilburg University

Kim, W.S., and Young, C.M., 1991. The Effect of Traded Option Introduction on

Shareholder Wealth. Journal of Financial Research 14, 141-151.

Klemkosky, R.C., and Maness, T.S., 1980. The Impact of Options on the Underlying

Securities. Journal of Portfolio Management 9, 12-18.

Kumar, R., Sarin, A, and Shastri, K., 1995. The Impact of the Listing of Index Options on the Underlying Stocks: The Evidence from the Listing of Nikkei Stock Average Options.

Pacific-Basin Finance Journal 3, 303-317.

Lamoureux, C.C., and Panikkath, S.K., 1994. Variations in Stock Returns: Asymmetries and Other Patterns, Working paper.

Lee, C.S. and Hsieh.W.L., 2000. Primary Market Pricing, Risk, and Issuing Strategies in

Taiwan Warrant Market. Fu-Hwa Journal, 1-36 (translated from Chinese).

Long, D.M., Schinski, M.D. and Officer, D.T., 1994. The impact of Option Listing on the

Price Volatility and Trading Volume of Underlying OTC Stocks, Journal of Economics and

Finance 18, 89-100.

MacKinlay, A.C., 1997. Event Studies in Economics and Finance. Journal of Economic

Literature 35, 13-19.

Mayhew, S., Mihov, V., 2000. Another Look at Option Listing Effects. Working paper,

Purdue University.

Mendenhall, D.H. and Fehrs, R.R., 1999. Option Listing and the Stock-Price Response to

Earnings Announcements, Journal of Accounting and Economics 27, 57-87.

Michaely, R., Thaler, R., and Womack, K., 1995. Price Reactions to Dividend Initiations and Omissions: Over-reactions or Drift? Journal of Finance 50, 573-608.

Nelson, D., 1991. Conditional Heteroscedasticity in Asset Returns: a New Approach.

Econometrica 59, 347-370.

Petterson, K., 2000. An Introduction to Applied Econometrics. MacMillan Press, 722-726.

Pierre, E.F., 1998. The Impact of Option Introduction on the Conditional Return

Distribution of Underlying Securities. Financial Review 33, 105-118.

Poon, P.S., 1994. An Empirical Examination of the Return Volatility –Volume Relation in

Related Markets: The Case of Stock and Options. Financial Review 29, 473-496.

Sahlström, P., 2001. Impact of Stock Option Listing on Return and Risk Characteristics in

Finland. International Review of Financial Analysis 10, 19-36

Schinski, M.D., Long, D.M. 1995. The Wealth Effects of Liquidity Gains from Option

Listing, in: Chen, A. H., (ed.), Research in Finance, 13, 203-217.

Shastri, P., Sultan, J., and Tandon, K., 1996. The Impact of the Listing of Options in the

Foreign Exchange Market. Journal of International Money and Finance 15, 37-64.

Skinner, D.J., 1989. Options Markets and Stock Return Volatility. Journal of Financial

Economics 23, 61-78.

Sorescu, S., 2000. The Effect of Options on Stock Prices: 1973 to 1995. Journal of Finance

55, 487-514.

Stucki, T., and Wasserfallen, W., 1994. Stock and Option Markets, The Swiss Evidence.

Journal of Banking and Finance 18, 881-893.

Trennepohl, G.L., and Dukas, W.P., 1979. CBOE Options and Stock Volatility. Review of

Business and Economic Research 14, 49-60.

Wang, Y, M., 2003. The Impact of Warrants Trading on Underlying Stock Return Volatility:

The Taiwan Evidence. Taiwan Finance Quarterly Journal 4-2, 65-79 (translated from

Chinese).

Watt W.H., Yadav P.K. and Draper. P., 1992. The Impact of Option Listing on Underlying

Stock Returns: The UK Evidence. Journal of Business Finance and Accounting 19,

485-503.

Wei, P., Poon, P.S., Zee, S., 1997. The Effect of Option Listing on Bid-Ask Spreads, Price

Volatility and Trading Activity of the Underlying OTC Stocks. Review of Quantitative

Finance and Accounting 9, 165-180.

Whiteside, M., Dukes, W. P., and Dunne, P., 1981. Announcement Impact on Securities of

Future Option Trading. Nebraska Journal of Economics and Business 20, 63-72.

Whiteside, M., Dukes, W. P., Dunne, P., 1983. Short Term Impact of Option Trading on

Underlying Securities. Journal of Financial Research 6, 313-321.

Young, C.W., 1999. A Causality Study of Warrants and Securities in Taiwan. Taiwan Land

Bank Finance Quarterly Journal 141, 51-68 (translated from Chinese).

Yu, H.C., Tsai, C.S., 2003. Impact of Covered Warrant Issuance on the Beta Risk and

Trading Volume of the Underlying Shares. Commercial Bank Quarterly Journal 26, 77-99

(translated from Chinese).

Zahari, Z., Briston, R. and Jelic, R., 2002. The Impact of Warrants upon the Underlying

Securities – A Survey and an Empirical Study of the Malaysian Derivatives Exchange,

Working Paper presented at the World Federation of Exchanges – Forum on Managing

Exchanges in Emerging Economies, Kuala Lumpur 2002.

Zane, W.D., 1998. Effects of Covered Warrant Issuance and Trading on the Underlying

Stocks. Securities and Finance Journal, 25-27 (translated from Chinese).

Figure 1 Annual Turnover of Listed Covered Warrants on World Stock Exchanges (2003) Source: Hong Kong Stock Exchange, “Exchange”, April 2004, p.41; Taiwanese Stock Exchange; WFE (World Federation of Exchanges) Monthly Statistics, December 2003.

Annual turnover of listed covered warrants on world stock exchanges (2003)

US$m 50,000 46,233

40,000 33,961 30,000

16,505 20,000 13,125 Annual turnover 10,039

10,000 3,480 1,829 1,655 808 546 222 185 59 39 33 20 2 1 0 e a s g ge a t ge d i c g e n lia n i iv o o s w o ian a ink sl ic sh a Bor hab wan han tr la Iri s e c tal onex ai c ai ls Av bour O ex Bor ar x I ur T us th Afr el m ch ng K a T He e M W s E s E Ex A T ner s h ux ie ct Ho or s Sou L wis B W De S pani SE S J Market

Table 1 Population and Sample Covered Warrants Stratified By Year of Listing Population consists of all warrants (including warrants on baskets of securities, individual shares, indexes, exotic warrants, etc.). Sample consists of first time and subsequently introduced warrants on individual stocks, from the market’s inception (August 1997), until February 2003. End of year market values in million US$. Exchange rates TW$/US$: 1997 (26.66), 1998 (33.44), 1999 (32.27), 2000 (31.23), 2001 (33.80), 2002 (34.58), 2003 (34.42). Source: Taiwan Stock Exchange (TSE) Fact Book, and authors’ calculations.

Year 1997 1998 1999 2000 2001 2002 2003

Population

- Number of all warrants 61452545759 36

- Number of stock warrants 6 8 34 42 57 58 17

- Market value of all warrants 57 384 1,905 4,772 836 2,190 3,480

- Market value of all TSE stocks 1,279,899 890,994 907,813 977,632 543,045 632,651 595,902

Sample

- Number of the sample warrants 5 7 22 27 37 35 13

- Number of first time listed warrants 4 3 12 6 11 14 8

- Market value of the sample warrants 33 282 1,320 2,247 530 1,180 346

- Market value of the sample stocks 40,812 182,898 306,244 247,565 194,689 169,145 78,875

Table 2 Daily Abnormal (ARs) and Cumulative Abnormal Returns (CARs) During the Event Window Total sample consists of both, first time and subsequently introduced warrants on the underlying stocks. Average (mean and median) AR and CAR, 10 days before listing announcements, and 10 days after listing announcements. The level of significance for the two-tailed T test of mean = 0 vs. mean ≠ 0 indicated as:*** significant at 1% level; ** significant at 5% level; * significant at 10% level. Unreported results for the Sign test of median = 0 vs. median ≠ 0 total sample abnormal returns confirm results of the T-test for days –4,-1,0,1,2,3, and 10.

Total sample First time introduced Subsequently introduced Day AR CAR AR CAR AR CAR -10 0.0510 0.0510 0.01981 0.01981 0.0715 0.0715

-9 0.0728 0.0875 0.08713 0.07574 -0.0086 0.0445

-8 0.0815 0.1185 0.14127 0.14341 0.0308 0.0541

-7 0.1527 ** 0.1789 ** 0.15981 0.20410 0.1566 0.1252

-6 0.0416 0.1786 ** 0.11964 0.23606** -0.0615 0.0844

-5 0.0328 0.1765 ** 0.23054*** 0.30961*** -0.0908 0.0400

-4 -0.1340 ** 0.1127 -0.19350 0.21350 -0.0281 0.0264

-3 0.0435 0.1208 0.04353 0.21510 0.0321 0.0360

-2 -0.0190 0.1076 0.07603 0.22815** -0.1177 -0.0053

-1 0.7234 *** 0.3308 *** 0.77051*** 0.46009*** 0.7463*** 0.2310**

0 -0.2251 *** 0.2476 *** 0.00772 0.44101*** -0.3670*** 0.1096

1 -0.2923 *** 0.1527 ** -0.26768** 0.34496*** -0.2608*** 0.0296

2 -0.1833 ** 0.0958 -0.24491** 0.26359** -0.1507 -0.0144

3 -0.1778 ** 0.0448 -0.33490*** 0.16441 -0.0390 -0.0233

4 -0.1314 0.0094 -0.04249 0.14787 -0.1919** -0.0720

5 -0.0285 0.0020 -0.09017 0.12063 -0.4733 -0.0816

6 -0.1279 -0.0291 -0.22468** 0.06253 -0.0843 -0.0996

7 0.0836 -0.0086 0.03354 0.06868 0.1229 -0.0680

8 0.0758 0.0090 0.10203 0.1023 -0.00484 -0.0663

9 0.1304 0.0739 0.10814 0.1081 0.1868** -0.0229

10 -0.1380 ** 0.0069 0.06570 0.0657 -0.1009 -0.0443

Table 3 Average ARs and CARs Total sample consists of both, first time and subsequently introduced warrants on the underlying stocks. Average (mean) ARs and CARs within the event window in percentage terms. Median returns reported in (brackets). P-values [in parentheses] indicate the level of significance for the differences in mean T-test) and median (Wilcoxon test) AR and CAR, before and after listing announcements, and between first and subsequently introduced warrants. Unreported results for Kruskal-Wallis test confirm reported results for differences between median AR and CAR before and after listing announcements. The level of significance for the two-tailed T, and Wilcoxon tests, for mean/median = 0 vs. mean/median ≠ 0 given as: *** significant at 1%; ** significant at 5%; * significant at 10%.

Panel A: Average ARs and CARs During the Event Window

N ARs CARs

Total sample 146 0.00 0.10 (0.33) (0.10) First time introduced 58 0.01 0.19 (0.03) (0.16) Subsequently introduced 88 -0.03 0.17 (-0.04) (0.03)

Difference (First vs. Subsequent) 0.04 0.02 (0.07) (0.13)

P-value for T-test [0.55] [0.0] P-value for Wilcoxon test [0.35] [0.0]

Panel B: Average ARs and CARs Before and After Introductions (Announcements)

Total Sample First time listed Subsequently listed N AR CAR N AR CAR N AR CAR

Before announcements 146 0.10 0.14 58 0.15 0.21 88 0.07 0.07 (0.05) (0.12) (0.10) (0.21) (0.01) (0.05)

After announcements 146 -0.09 0.05 58 -0.10 0.17 88 -0.12 -0.03 (-0.13) (0.01) (-0.09) (0.12) (-0.10) (-0.04)

Difference (before-after) 0.19 0.19 0.25 0.05 0.19 0.10 (0.18) (0.18) (0.19) (0.09) (0.11) (0.09)

P-value - T-test [0.03] [0.02] [0.01] [0.48] [0.05] [0.00] P-value – Wilcoxon test [0.04] [0.04] [0.01] [0.34] [0.06] [0.00]

Table 4 Changes in Conditional Stock Volatility After Warrant Introductions Number of equation’s 1 coefficients with statistical significance at 5% or better, as reported in Annex 1.

EGARCH First time introduced Subsequently introduced Total sample Before After Before After Before After 2 28 25 44 51 72 76 β logσ t−1

q εt−1 2 12 17 16 23 28 40 ∑(α j − ) j=1 σ t−1 π

q εt−1 6 4 6 11 12 15 ∑ ()γ j j=1 σ t−1

Table 5 Changes in Systematic Risk After Warrant Introductions Number of shares with significant and non-significant changes in beta (β) after warrant listings. Reported results for 5% significance level.

Changes in β Value Total Sample First time introduced Subsequently introduced Statistically significant 22 9 13 Statistically non-significant 124 79 45

Table 6 Daily Abnormal Trading Volumes (AVs) Mean AVs during the event window calculated as in equation 5 in percentage terms. The level of significance for the T- test for mean = 0 vs. mean 0, calculated as in equation 8, given as: *** significant at 1%; ** significant at 5%; * significant at 10%

Day Total sample First time introduced Subsequently introduced

-10 35.2393*** 64.0387*** 16.7810

-9 27.8127*** 45.5596** 16.4787

-8 33.9408*** 56.6421** 18.7720

-7 40.6174*** 72.6487*** 19.8785**

-6 42.6210*** 65.4313*** 28.0265***

-5 42.1836*** 67.7836*** 23.7906**

-4 52.6046*** 95.9081*** 23.0616**

-3 48.1356*** 80.3655*** 27.0058**

-2 39.3697*** 57.4203*** 27.4737***

-1 82.1184*** 108.2153*** 65.6949***

0 48.5283*** 76.0833*** 30.4310***

1 31.3507*** 45.3315** 22.0493**

2 34.7008*** 63.4676*** 16.3569

3 25.0876** 42.7602** 14.0190

4 22.8274** 41.4893** 10.6454

5 25.4037** 51.2322** 8.5096

6 22.4655** 45.7426** 7.3892

7 24.0925** 40.7438** 11.6378

8 44.9149*** 93.9734*** 13.6428

9 40.0195*** 71.2563*** 20.4340**

10 48.1009*** 102.3035*** 13.8311

Table 7 Average Abnormal Trading Volumes (AVs) Total sample consists of both, first time and subsequently introduced warrants on the underlying shares. Average (mean and median) AVs, 10 days before announcements, and 10 days after announcements. Median returns reported in (brackets). P-values [in parentheses] indicate the level of significance for the differences in mean (T-test) and median (Wilcoxon test) ARs, before and after introductions, and between first time and subsequently introduced warrants. Unreported results for Kruskal-Wallis test confirm reported results for differences between median returns. The level of significance for the two-tailed T, and Wilcoxon tests, for mean/median = 0 vs. mean/median ≠ 0 given as: *** significant at 1%; ** significant at 5%; * significant at 10%.

Panel A: AVs for First Time and Subsequently Introduced Warrants During the Event Period

First time introduced 58 66.11 (64.04)

Subsequently introduced 88 20.76 (18.77)

Difference (first minus subsequent) 45.35 (5.27)

P-value for T-test [0.0] P-value for Wilcoxon test [0.0]

Panel B: Average AVs Before and After Introductions

Total Sample First time introduced Subsequently introduced N AV N AV N AV

Before announcements 146 44.46 58 71.40 88 26.70 (41.40) (66.61) (23.43)

After announcements 146 33.41 58 61.31 88 15.36 (31.35) (51.23) (13.83)

Difference (before minus after) 11.05 10.09 11.34 (9.6) (10.05) (15.38)

P-value for T-test [0.06] [0.28] [0.03] P-value for Wilcoxon test [0.07] [0.17] [0.01]

Table 8 Relative Trading Volume Ratios Total sample consists of both, first time and subsequently introduced warrants on the underlying shares RViB stands for average (mean and median) relative trading volume ratio100 days before the warrant introductions; RViA stands for the average (mean and median) relative trading volume ratio 100 days after the warrants listings. Median RViB and RViA reported in (brackets). The ratio is presented in P-values [in parentheses] indicate the level of significance for the differences in mean (T-test) and median (Wilcoxon test) ratios before and after listings, and between first time and subsequently introduced groups.

Panel A: Average Relative Trading Volume Ratios Before and After Introductions

Total sample N Total sample N First time N Subsequently introduced introduced

RViB 146 1.11 (0.79) 58 0.97 (0.74) 88 1.17 (0.83)

RViA 146 0.99 (0.64) 58 0.93 (0.50) 88 1.00 (0.68)

Difference (RViB-RViA) 0.12 (0.15) 0.04 (0.24) 0.17 (0.15) P-value for T-test [0.12] [0.26] [0.10] P-value for Wilcoxon [0.07] [0.20] [0.15]

Panel B: Average Relative Trading Volume Ratios for the First Time and Subsequently Introduced Warrants

First time vs. Subsequent RViB First time vs. Subsequent RViA N N First time introduced 58 0.97 58 0.93 (0.74) (0.50)

Subsequently introduced 88 1.17 88 1.00 (0.83) (0.68)

Difference (First vs. Subsequent) -0.2 -0.07 (-0.19) (0.18)

P-value for T-test [0.20] [0.61] P-value for Wilcoxon test [0.13] [0.06]

Annex: Results for EGARCH model

Underlying Dates of EGARCH q q ε εt−1 2 t−1 ()γ j stocks warrant 2 ()α j − ∑ ln(γ *ν ) β logσ ∑ j=1 σ t−1 t−1 j=1 σ t−1 π introductions

ASUS 9/7/99 Coefficient 1.83 -0.14 -0.05 -2.11 Standard error 0.73 0.21 0.41 3.14 T-value 2.51 -0.64 -0.13 -0.67 3/16/01 Coefficient 2.98 0.00 -0.17 1013 Standard error 1.42 2267 0.55 13700 T-value 2.10 -0.06 -0.31 0.07 12/17/01 Coefficient 3.95 0.00 -0.72 1709 Standard error 0.31 1026 0.09 7823 T-value 12.78 -0.21 -8.00 0.22 1/3/03 Coefficient 2.08 0.27 0.04 -0.60 Standard error 1.14 0.22 0.52 0.70 T-value 1.83 1.22 0.08 -0.85 BODA 12/12/01 Coefficient 0.99 -0.19 0.58 -0.43 Standard error 1.34 0.21 0.57 0.79 T-value 0.74 -0.91 1.03 -0.55 CDH 6/11/99 Coefficient 3.31 -0.42 -0.83 0.27 Standard error 0.21 0.13 0.10 0.23 T-value 15.56 -3.18 -8.03 1.18 3/8/00 Coefficient 0.23 0.00 0.89 3554 Standard error 0.18 392.00 0.08 1013 T-value 1.28 1.91 10.62 3.51 5/16/00 Coefficient 4.35 0.26 -1.00 0.68 Standard error 0.02 0.02 7043 0.01 T-value 189.68 16.25 142.41 52.55 CH 8/20/97 Coefficient 3.24 -0.22 -0.25 -0.63 Standard error 1.51 0.24 0.58 1.07 T-value 2.15 -0.94 -0.43 -0.59 12/23/97 Coefficient 0.31 0.18 0.88 -0.21 Standard error 0.36 0.16 0.13 0.56 T-value 0.86 1.08 6.74 -0.37 7/7/98 Coefficient 0.41 0.45 0.81 -0.20 Standard error 0.23 0.17 0.11 0.18 T-value 1.81 2.66 7.53 -1.11 10/22/98 Coefficient 0.42 0.28 0.79 -0.42 Standard error 0.30 0.17 0.14 0.41 T-value 1.43 1.60 5.48 -1.03 4/28/99 Coefficient 0.09 0.00 0.94 -6355 Standard error 0.11 172.00 0.06 57356 T-value 0.85 0.10 15.13 -0.11 10/13/99 Coefficient -0.04 0.00 1.01 1469 Standard error 0.07 264.00 0.03 10986 T-value -0.59 0.12 38.38 0.13 9/29/00 Coefficient 4.58 0.29 -1.02 0.31 Standard error 0.14 0.07 0.03 0.12 T-value 33.09 4.03 -31.13 2.54 8/10/01 Coefficient 0.70 0.00 0.65 -9086 Standard error 0.43 304.00 0.21 47773 T-value 1.61 0.18 3.09 -0.19 12/27/01 Coefficient 2.11 -0.08 0.13 1.34 Standard error 1.32 0.24 0.54 4.42

T-value 1.60 -0.32 0.24 0.30 CHIB 10/9/01 Coefficient 4.38 0.00 -0.92 270.20 Standard error 0.26 0.00 0.06 1987 T-value 17.11 -0.12 -16.70 0.14 CHZ 3/28/02 Coefficient 0.23 -0.05 0.89 -3.36 Standard error 0.16 0.17 0.07 1.27 T-value 1.41 -0.30 12.18 -0.30 CS 11/26/02 Coefficient 0.08 0.00 0.94 3965 Standard error 0.07 337.00 0.05 25696 T-value 1.16 0.14 18.70 0.15 8/2/02 Coefficient 1.92 0.16 -0.31 1.37 Standard error 0.67 0.22 0.45 2.04 T-value 2.87 0.72 -0.70 0.67 CSC 1/13/03 Coefficient 0.18 -0.54 0.87 -0.18 Standard error 0.09 0.14 0.06 0.03 T-value 2.08 -3.91 14.67 -6.45 DONG 1/19/00 Coefficient 0.33 0.15 0.82 1.27 Standard error 0.34 0.38 0.19 3.46 T-value 0.98 0.39 4.41 0.37 9/28/00 Coefficient 0.09 0.00 0.96 -4524 Standard error 0.06 267.00 0.02 19308 T-value 1.47 0.22 39.49 -0.23 11/9/00 Coefficient 1.12 -4.29 -0.57 -0.70 Standard error 0.61 0.61 0.61 0.61 T-value 1.82 -6.99 -0.93 -1.14 10/24/01 Coefficient 3.08 0.44 -0.41 -0.15 Standard error 0.62 0.22 0.28 0.30 T-value 4.98 1.96 -1.48 -0.51 12/14/01 Coefficient 0.53 0.28 0.76 0.23 Standard error 0.79 0.16 0.35 0.41 T-value 0.67 1.76 2.16 0.56 3/6/02 Coefficient 0.29 0.11 0.87 2.14 Standard error 0.18 0.18 0.08 3.87 T-value 1.58 0.60 10.69 0.55 FM 1/28/00 Coefficient 1.19 -2.86 0.09 0.28 Standard error 0.09 0.10 0.05 0.05 T-value 13.16 -28.30 1.95 5.47 FPC 8/15/02 Coefficient 1.36 -0.12 -0.25 -2.38 Standard error 0.47 0.28 0.40 5.94 T-value 2.90 -0.42 -0.64 -0.40 HNC 12/30/02 Coefficient 1.97 0.21 -0.25 -0.62 Standard error 0.61 0.23 0.39 1.12 T-value 3.24 0.91 -0.64 -0.55 GY 3/27/02 Coefficient 2.53 -0.18 0.00 -0.88 Standard error 1.29 0.30 0.51 1.52 T-value 1.96 -0.62 0.01 -0.58 HBD 4/1/99 Coefficient 3.54 -0.17 -0.62 -0.18 Standard error 0.98 0.20 0.47 0.70 T-value 3.62 -0.84 -1.31 -0.26 6/11/99 Coefficient 2.72 -0.23 -0.41 0.60 Standard error 1.11 0.19 0.58 1.18 T-value 2.44 -1.20 -0.71 0.51 10/22/99 Coefficient 2.50 0.15 -0.03 0.34 Standard error 1.84 0.23 0.76 1.15 T-value 1.35 0.64 -0.04 0.29 2/19/00 Coefficient 3.78 0.46 -0.94 0.10 Standard error 0.12 0.09 0.04 0.03 T-value 32.56 5.38 -24.84 3.83

6/21/00 Coefficient 4.32 0.23 -0.83 0.49 Standard error 0.34 0.19 0.10 0.36 T-value 12.60 1.25 -8.70 1.37 4/16/01 Coefficient 0.03 -0.21 0.99 1.07 Standard error 0.02 0.02 6741 3816 T-value 1.74 -11.50 147.15 279.29 7/17/02 Coefficient 0.22 0.19 0.91 -0.30 Standard error 0.24 0.17 0.10 0.65 T-value 0.90 1.12 9.61 -0.46 HSCI 10/8/99 Coefficient 0.32 0.26 0.84 0.73 Standard error 0.17 0.13 0.08 0.48 T-value 1.88 1.98 10.12 1.53 3/24/00 Coefficient 4.56 0.43 -0.67 -0.18 Standard error 0.51 0.26 0.16 0.31 T-value 8.96 1.64 -4.28 -0.58 6/14/00 Coefficient 0.10 0.21 0.95 -0.33 Standard error 0.30 0.15 0.12 0.51 T-value 0.32 1.45 8.09 -0.66 10/18/00 Coefficient 0.13 0.45 0.94 -0.49 Standard error 0.13 0.21 0.06 0.27 T-value 1.03 2.17 16.98 -1.83 10/17/01 Coefficient -0.03 0.00 1.03 -576 Standard error 0.06 555.00 0.02 2.36 T-value -0.51 0.82 48.73 -0.87 HT 1/27/00 Coefficient 0.11 0.12 0.95 -0.14 Standard error 0.09 0.09 0.04 0.51 T-value 1.29 1.27 23.75 -0.27 HUA 1/14/03 Coefficient 4.23 0.07 -0.56 -2.35 Standard error 1.02 0.22 0.38 6.84 T-value 4.12 0.32 -1.50 -0.34 HUAY 4/4/02 Coefficient 3.27 -0.46 -0.25 0.09 Standard error 1.00 0.32 0.38 0.39 T-value 3.27 -1.40 -0.65 0.22 HUHI 1/8/03 Coefficient -0.09 0.00 1.05 -71.82 Standard error 0.05 1487 0.03 7.01 T-value -1.63 2.07 40.64 -4.22 KB 1/10/02 Coefficient 1.89 -0.26 0.32 -0.21 Standard error 2.65 0.34 0.96 1.13 T-value 0.71 -0.76 0.34 -0.18 KFC 3/20/01 Coefficient 0.36 0.68 0.79 -0.09 Standard error 0.18 0.22 0.11 0.14 T-value 1.98 3.10 7.30 -0.66 KH 8/20/97 Coefficient 2.96 -0.38 -0.53 0.96 Standard error 0.76 0.41 0.36 1.10 T-value 3.90 -0.93 -1.47 0.87 12/26/97 Coefficient 0.44 0.17 0.83 -0.80 Standard error 0.70 0.28 0.27 1.35 T-value 0.63 0.60 3.08 -0.59 1/18/00 Coefficient 1.78 0.15 0.26 0.86 Standard error 1.07 0.21 0.44 1.37 T-value 1.67 0.73 0.59 0.63 5/19/00 Coefficient 2.22 0.12 0.17 1.06 Standard error 1.81 0.21 0.68 1.99 T-value 1.23 0.56 0.25 0.53 6/18/01 Coefficient 2.25 -0.57 0.04 -0.52 Standard error 0.89 0.28 0.37 0.30 T-value 2.53 -2.00 0.11 -1.71 10/23/01 Coefficient 2.51 0.48 -0.10 0.01

Standard error 1.29 0.27 0.56 0.30 T-value 1.94 1.81 -0.17 0.05 08/01/20002 Coefficient 4.49 0.11 -0.80 1.30 Standard error 0.45 0.12 0.15 1.29 T-value 9.97 0.88 -5.24 1.00 KMT 1/10/03 Coefficient 0.17 0.16 0.93 0.07 Standard error 0.21 0.16 0.09 0.48 T-value 0.80 1.01 10.91 0.15 KPC 1/09/03 Coefficient 0.38 0.09 0.80 1.29 Standard error 0.39 0.15 0.21 2.35 T-value 0.98 0.60 3.73 0.55 KUA 3/27/02 Coefficient 0.21 0.33 0.90 -0.12 Standard error 0.22 0.19 0.10 0.31 T-value 0.98 1.71 9.20 -0.38 1/20/03 Coefficient 0.09 0.00 0.96 -4523 Standard error 0.11 312.00 0.04 40235 T-value 0.83 0.10 21.79 -0.11 LAI 3/6/02 Coefficient 2.08 -0.21 0.18 -0.19 Standard error 1.83 0.24 0.72 1.04 T-value 1.14 -0.85 0.25 -0.19 LT 3/4/98 Coefficient 0.91 0.26 0.66 -0.56 Standard error 1.02 0.18 0.38 0.51 T-value 0.89 1.51 1.74 -1.10 5/27/99 Coefficient 1.02 0.00 0.46 -6255 Standard error 0.19 344.00 0.19 0.19 T-value 5.26 2.14 2.36 -32386 10/20/99 Coefficient -0.14 0.01 1.05 2.78 Standard error 0.05 4786 0.02 2.54 T-value -2.75 2.59 49.47 1.10 1/13/00 Coefficient 1.83 -0.07 -0.11 0.91 Standard error 0.71 0.32 0.43 3.90 T-value 2.57 -0.23 -0.25 0.23 11/9/00 Coefficient 0.08 0.14 0.96 -1.61 Standard error 0.08 0.09 0.03 1.29 T-value 1.04 1.48 28.89 -1.24 4/12/01 Coefficient 2.50 -0.12 -0.07 1.02 Standard error 1.10 0.22 0.47 2.07 T-value 2.28 -0.55 -0.16 0.49 12/17/01 Coefficient 0.62 0.26 0.76 -0.34 Standard error 0.59 0.23 0.23 0.40 T-value 1.05 1.17 3.23 -0.83 2/19/02 Coefficient 3.44 -0.16 -0.30 -0.96 Standard error 1.21 0.24 0.45 1.76 T-value 2.85 -0.65 -0.67 -0.55 MC 12/16/98 Coefficient 0.10 0.27 0.95 -0.40 Standard error 0.11 0.11 0.05 0.25 T-value 0.95 2.46 19.78 -1.62 11/15/99 Coefficient 1.60 0.23 0.38 0.12 Standard error 1.67 0.28 0.64 0.66 T-value 0.96 0.83 0.61 0.18 MCHI 6/5/99 Coefficient 0.25 -0.17 0.86 0.35 Standard error 0.17 0.13 0.09 0.39 T-value 1.49 -1.31 9.56 0.90 2/13/01 Coefficient 2.51 0.88 0.09 0.05 Standard error 0.88 0.27 0.31 0.17 T-value 2.85 3.23 0.30 0.29 10/29/01 Coefficient 3.52 0.84 -0.78 -0.19 Standard error 0.10 0.10 0.10 0.10

T-value 35.26 8.51 -7.98 -1.88 3/26/02 Coefficient 4.78 -0.29 -0.85 0.21 Standard error 0.42 0.20 0.13 0.35 T-value 11.38 -1.46 -6.52 0.59 NSD 1/11/00 Coefficient 1.13 0.17 0.53 -0.27 Standard error 1.35 0.16 0.56 0.54 T-value 0.84 1.08 0.94 -0.51 NY 1/21/98 Coefficient 2.14 0.50 -0.17 0.02 Standard error 0.64 0.22 0.34 0.30 T-value 3.33 2.26 -0.49 0.07 3/31/99 Coefficient 2.08 0.00 -0.25 2931 Standard error 0.35 806.00 0.19 0.37 T-value 5.92 1.91 -1.37 929.22 10/14/00 Coefficient 1.68 0.03 0.11 10.04 Standard error 0.58 0.21 0.30 7.70 T-value 2.93 0.13 0.35 0.13 1/20/00 Coefficient 0.45 -1.21 0.65 -0.20 Standard error 0.01 0.03 9518 0.01 T-value 34.70 -43.70 67.85 -15.44 12/27/01 Coefficient 3.49 0.65 -0.45 0.02 Standard error 0.56 0.22 0.20 0.17 T-value 6.29 2.91 -2.23 0.10 PAOC 3/12/01 Coefficient 0.57 -0.14 0.67 0.82 Standard error 0.76 0.31 0.46 2.17 T-value 0.75 -0.44 1.46 0.38 1/22/02 Coefficient 0.78 0.20 0.65 0.99 Standard error 0.77 0.21 0.35 1.48 T-value 1.01 0.94 1.89 0.67 1/6/03 Coefficient 0.56 -0.10 0.74 -2.91 Standard error 0.31 0.15 0.15 3.59 T-value 1.77 -0.68 4.94 -0.81 SAMD 3/5/02 Coefficient 1.10 0.12 0.61 0.32 Standard error 1.74 0.23 0.61 0.94 T-value 0.63 0.51 1.00 0.34 SCIP 11/17/99 Coefficient 0.08 0.12 0.96 -0.10 Standard error 0.15 0.08 0.08 0.59 T-value 0.54 1.47 12.46 -0.18 6/28/00 Coefficient 1.85 0.14 0.16 0.59 Standard error 1.76 0.23 0.80 1.50 T-value 1.05 0.61 0.21 0.40 1/10/02 Coefficient 0.94 0.22 0.66 -0.14 Standard error 0.95 0.31 0.34 0.53 T-value 0.99 0.73 1.96 -0.26 1/8/03 Coefficient 2.60 -0.15 0.02 0.20 Standard error 1.59 0.20 0.60 0.98 T-value 1.63 -0.76 0.03 0.20 SML 6/9/99 Coefficient 3.05 -0.13 -0.47 -1.63 Standard error 0.76 0.25 0.37 3.18 T-value 4.00 -0.52 -1.27 -0.51 11/26/99 Coefficient 0.05 0.17 0.97 0.32 Standard error 0.10 0.15 0.05 0.45 T-value 0.54 1.16 17.73 0.72 SSC 1/10/03 Coefficient 0.19 -0.56 0.87 -0.41 Standard error 0.04 0.02 0.02 0.04 T-value 4.50 -33.29 53.96 -9.71 TATO 12/9/99 Coefficient 2.19 -0.34 -0.06 0.45 Standard error 0.73 0.30 0.34 0.50 T-value 3.02 -1.15 -0.16 0.89

TCC 12/20/02 Coefficient 3.67 -0.16 -0.95 0.42 Standard error 0.28 0.10 0.04 0.57 T-value 13.23 -1.53 -22.77 0.74 TCD 3/26/99 Coefficient 3.61 -0.21 -0.91 -0.47 Standard error 0.24 0.10 0.07 0.37 T-value 15.00 -2.09 -13.27 -1.26 7/21/99 Coefficient 2.51 -0.20 -0.27 -0.89 Standard error 0.70 0.26 0.34 1.50 T-value 3.59 -0.74 -0.77 -0.59 4/19/00 Coefficient 2.68 0.00 -0.37 -473 Standard error 0.59 4843 0.28 4102 T-value 4.51 0.11 -1.31 -0.12 2/1/01 Coefficient 0.42 0.16 0.84 -1.02 Standard error 0.34 0.15 0.13 1.09 T-value 1.26 1.13 6.73 -0.94 TD 8/21/97 Coefficient 0.55 0.15 0.71 -0.66 Standard error 0.49 0.14 0.26 0.82 T-value 1.13 1.09 2.69 -0.81 4/9/98 Coefficient 1.90 0.87 -0.32 0.23 Standard error 0.32 0.20 0.20 0.12 T-value 5.94 4.32 -1.61 2.01 6/15/98 Coefficient 1.69 0.78 -0.30 0.19 Standard error 0.46 0.22 0.30 0.15 T-value 3.69 3.56 -0.99 1.28 1/12/00 Coefficient 0.12 0.07 0.94 1.24 Standard error 0.16 0.19 0.08 4.16 T-value 0.76 0.37 11.73 0.30 8/14/01 Coefficient 0.17 0.32 0.92 -0.34 Standard error 0.16 0.13 0.07 0.33 T-value 1.11 2.52 13.16 -1.02 12/11/01 Coefficient 1.00 0.07 0.55 1.42 Standard error 1.00 0.23 0.44 4.61 T-value 1.00 0.30 1.25 0.31 3/4/02 Coefficient 3.88 0.10 -0.79 3.53 Standard error 0.32 0.13 0.09 4.60 T-value 11.99 0.75 -9.08 0.77 TDT 11/26/99 Coefficient -0.02 0.00 1.01 1796 Standard error 0.02 0.00 0.01 1037 T-value -0.90 1.45 110.62 1.73 12/5/01 Coefficient 2.84 -0.14 -0.02 0.14 Standard error 3.84 0.31 1.37 1.02 T-value 0.74 -0.45 -0.01 0.14 THU 1/16/03 Coefficient 1.54 0.14 0.31 -0.94 Standard error 1.21 0.29 0.54 2.48 T-value 1.28 0.47 0.58 -0.38 TN 1/11/00 Coefficient 0.50 0.22 0.72 1.26 Standard error 0.23 0.13 0.12 0.93 T-value 2.22 1.74 5.97 1.34 TONE 6/7/01 Coefficient 2.02 0.26 -0.01 1.17 Standard error 0.65 0.24 0.32 1.15 T-value 3.12 1.07 -0.02 1.02 1/17/02 Coefficient 2.02 0.26 -0.01 1.17 Standard error 0.65 0.24 0.32 1.15 T-value 3.12 1.07 -0.02 1.02 7/16/02 Coefficient 0.88 0.52 0.35 0.11 Standard error 0.37 0.21 0.26 0.26 T-value 2.38 2.49 1.34 0.42 12/19/02 Coefficient 1.98 0.18 -0.28 0.04

Standard error 1.26 0.24 0.79 0.84 T-value 1.58 0.77 -0.35 0.05 TS 7/7/99 Coefficient 2.20 -0.15 -0.26 -1.68 Standard error 0.65 0.16 0.36 1.92 T-value 3.36 -0.92 -0.72 -0.88 1/19/00 Coefficient 2.29 -0.28 -0.77 -0.17 Standard error 0.35 0.16 0.20 0.42 T-value 6.49 -1.74 -3.80 -0.40 10/3/01 Coefficient 2.80 -0.28 -0.48 -0.35 Standard error 0.77 0.21 0.40 0.59 T-value 3.65 -1.34 -1.20 -0.59 1/21/02 Coefficient 1.03 -0.18 0.53 -0.59 Standard error 0.93 0.22 0.41 0.95 T-value 1.12 -0.83 1.29 -0.62 1/13/03 Coefficient 1.64 0.24 0.11 -0.67 Standard error 0.89 0.26 0.48 1.16 T-value 1.84 0.91 0.23 -0.58 WAN 10/22/01 Coefficient 0.03 0.00 0.99 -3201 Standard error 0.05 403.00 0.02 18100 T-value 0.65 0.15 58.70 -0.18 WEI 12/11/01 Coefficient 0.76 0.17 0.73 -0.22 Standard error 1.69 0.26 0.60 0.68 T-value 0.45 0.64 1.22 -0.32 1/7/03 Coefficient 2.32 -0.33 0.11 -0.86 Standard error 0.83 0.35 0.31 0.82 T-value 2.78 -0.95 0.35 -1.06 XUCI 10/22/01 Coefficient 0.18 0.00 0.91 3596 Standard error 0.15 511.00 0.07 0.17 T-value 1.18 1.94 12.49 -21736 YACI 4/16/02 Coefficient 4.53 -0.26 -0.76 0.17 Standard Error 0.55 0.19 0.20 0.47 T-value 8.24 -1.38 -3.87 0.37 YSC 10/29/01 Coefficient 2.36 0.67 -0.32 0.35 Standard error 0.47 0.30 0.25 0.26 T-value 5.00 2.24 -1.29 1.31 YUDA 7/12/02 Coefficient 4.83 -0.13 -0.89 -0.26 Standard error 0.40 0.18 0.15 0.74 T-value 12.05 -0.76 -6.10 -0.35 YUF 3/24/00 Coefficient 0.28 0.20 0.87 0.88 Standard error 0.17 0.11 0.07 0.63 T-value 1.67 1.75 11.72 1.40 8/17/01 Coefficient 0.49 0.52 0.79 -0.10 Standard error 0.24 0.19 0.10 0.19 T-value 2.07 2.67 7.92 -0.54 12/13/01 Coefficient 1.72 -0.43 0.35 -0.55 Standard error 0.60 0.24 0.22 0.38 T-value 2.87 -1.81 1.58 -1.47 3/5/02 Coefficient 3.17 -0.17 -0.23 -0.74 Standard error 2.26 0.25 0.87 1.30 T-value 1.41 -0.68 -0.26 -0.57 YUS 8/17/01 Coefficient 2.61 -0.28 0.01 0.35 Standard error 1.60 0.21 0.61 0.61 T-value 1.63 -1.32 0.01 0.57 1/25/02 Coefficient 3.62 -0.25 -0.21 0.88 Standard error 1.11 0.24 0.37 0.93 T-value 3.27 -1.04 -0.57 0.94 YULG 2/13/03 Coefficient 0.40 -0.62 0.84 0.15 Standard error 0.30 0.47 0.12 0.24

T-value 1.33 -1.30 6.85 0.63 YYD 1/3/02 Coefficient 1.20 0.28 0.56 -0.07 Standard error 1.47 0.25 0.53 0.63 T-value 0.82 1.10 1.05 -0.11 ZMB 12/7/99 Coefficient 2.19 -0.15 -0.14 -0.01 Standard error 1.04 0.23 0.54 0.96 T-value 2.10 -0.68 -0.26 -0.01 10/24/00 Coefficient -0.04 0.04 1.02 -2.49 Standard error 0.08 0.07 0.03 5.29 T-value -0.50 0.66 30.16 -0.47 1/30/01 Coefficient 2.34 0.24 0.07 -0.76 Standard error 1.56 0.22 0.62 0.80 T-value 1.50 1.09 0.12 -0.95 9/25/01 Coefficient 2.71 0.02 -0.17 -6.66 Standard error 1.59 0.16 0.69 4.66 T-value 1.71 0.15 -0.25 -0.15 12/21/01 Coefficient 2.68 -0.34 -0.04 0.56 Standard error 0.93 0.26 0.35 0.65 T-value 2.87 -1.28 -0.10 0.85 3/7/02 Coefficient 2.54 -0.39 0.02 0.30 Standard error 1.06 0.26 0.41 0.40 T-value 2.41 -1.48 0.04 0.74 8/20/02 Coefficient -0.05 0.00 1.03 -1623 Standard error 0.05 412.00 0.02 4278 T-value -1.17 0.36 67.36 -0.38 ZSCI 1/9/02 Coefficient 3.25 -0.27 -0.83 -0.52 Standard error 0.36 0.15 0.15 0.41 T-value 9.10 -1.86 -5.48 -1.27 3/29/02 Coefficient 2.00 0.05 -0.08 3.35 Standard error 1.23 0.28 0.66 8.23 T-value 1.62 0.19 -0.12 0.18 ZYI 7/27/01 Coefficient -0.08 0.00 1.03 -50.50 Standard error 0.59 0.59 0.59 0.59 T-value -0.13 0.00 1.76 -86.02 3/13/02 Coefficient 3.07 -0.46 -0.38 -0.52 Standard error 0.53 0.24 0.24 0.46 T-value 5.78 -1.92 -1.62 -1.13