7 Toeholds and the bidder shareholder wealth effects of takeover announcements.

Abstract:

This paper provides new evidence regarding the bidder shareholder wealth effects of takeover announcements by utilizing a robust method not previously employed in the area, namely Barber and Lyon's (1997) control firm approach. Thereafter, the study offers the first Australian evidence of the impact of toeholds on bidder abnormal returns. Finally, it considers whether this effect differs between single-bidder and multiple-bidder takeover contests. Using a dataset of 122 takeover announcements made by Australian listed companies between 1997 and 2004 inclusive, no significant bidder abnormal returns are documented in response to takeover announcements on average. However, examination of the cross-sectional variation in these wealth effects reveals a significantly positive association between the presence of toeholds/toehold size and bidder abnormal returns. Furthermore, immediate market reaction is consistent with Bukart's (1995) and Singh's (1998) Overbidding Hypothesis, which predicts a less pronounced toehold impact for rivaled bidders. Nonetheless, market expectation of overbidding for toehold bidders proves incorrect in the long-run post-announcement period, where the positive impact of toeholds does not differ between different models of takeover contests.

Keywords:

TAKEOVER; BIDDER SHAREHOLDER WEALTH EFFECTS; TOEHOLDS.

1. Introduction

Given the substantial volume of takeover activity in recent decades, it is not surprising that a large body of research has been dedicated to examining whether takeovers create value for participating firms' shareholders. Empirical studies focusing on the effects of takeover announcements unanimously document positive market reaction to target firms (Bugeja & Walter 1995). However, evidence regarding the impact on bidding firms is largely mixed, with studies finding positive (Jarrell & Poulsen 1989), negative (Dodd 1980), and even insignificant (Raad & Wu 1994) bidder abnormal returns.

We argue that this lack of consensus may well be a result of the flaws inherent in standard event study methodologies. This motivates us to re-examine the area, providing robust evidence by utilizing the control firm approach advanced by Barber and Lyon (1997). While Barber and Lyon (1997) propose this approach for use in the detection of long-run abnormal performance, we argue that matched companies represent unbiased benchmarks against which to evaluate bidding firms' wealth changes in the current context. Moreover, this approach is susceptible to neither the biases induced by prebid runup in the standard market model, nor the new-listing, rebalancing and skewness biases associated with the 0/1 market model.

Using a dataset of 122 takeover announcements made by 110 Australian listed companies between 1997 and 2004 inclusive, we also provide the first Australian evidence of the impact of toeholds on bidder abnormal returns. (1) We further contribute to the literature by considering this impact in two different types of bidding contests, namely single-bidder and multiple-bidder takeovers. While theory suggests possible different toehold effects conditional upon the presence of competing bidders (see Bukart 1995; Bulow, Huang & Klemperer 1999), an explicit attempt to empirically model this potential difference has been elusive.

Testing indicates the absence of significant overall bidder abnormal returns in response to takeover announcements. Specifically, while significantly negative abnormal returns are documented under the 0/1 market model, bidder abnormal losses disappear when we employ the bias-free method advanced by Barber and Lyon (1997). This suggests that the negative market reaction documented elsewhere may be an artefact of methodological flaws.

The lack of market response to takeover announcements in aggregate is not surprising given the possibility for cross-sectional variation in wealth effects. Univariate analyses show that bidders with toeholds earn significantly higher abnormal returns compared to their non-toehold counterparts. Multivariate analyses further support a positive association between toehold ownership and bidder gains from takeovers. Moreover, the larger the size of toeholds, the more favourable market response is observed upon takeover announcements.

Immediate market reaction further suggests that the toehold-abnormal return relationship is a function of the presence of rival bidders. In particular, while significantly positive toehold impact is documented for unrivalled bidders, the attractiveness of toeholds in multiple-bidder contests appears to be dampened by the increased probability of overbidding by the toehold bidder. However, the overbidding expectation proves incorrect in the long-run post-announcement period, where toehold benefits are not affected by the presence of bidding rivals.

The remainder of this paper is structured as follows: section 2 overviews the existing literature pertaining to the bidder shareholder wealth effects of takeover announcements, paying particular attention to their relationship with pre-takeover toehold formation; section 3 describes the research design; section 4 presents and discusses the results; and, finally, section 5 concludes the paper.

2. Prior Literature

2.1 The Bidder Shareholder Wealth Effects of Takeover Announcements

Two strands of theory have evolved to explain why takeovers occur, each with different implications for the bidder shareholder wealth effects. Under the Neoclassical Welfare Theory, takeovers are welfare-maximizing activities undertaken by managers whose interests are aligned with those of shareholders. The Agency Theory (Jensen 1986, 1988) and the Hubris Hypothesis (Roll 1986), on the other hand, stray from the premise that manager-specific objectives or managers' hubris drive takeover decisions. While the Neoclassical Welfare Theory predicts positive gains to bidder shareholders, the latter suggests that takeovers result in bidder shareholders' value loss. (2)

The surge in takeover activity stresses the importance of empirically ascertaining whether takeovers create value for bidder shareholders. To do so, many previous studies focus on the bidder wealth changes surrounding takeover announcements. These studies are predicated upon the assumption that the market is of semi-strong form efficiency where share prices reflect instantaneously and unbiasedly all new information affecting their intrinsic value (Fama 1970). If unexpected takeover announcements carry the information about future gains as predicted by the Neoclassical Welfare Theory, they should result in positive abnormal bidder returns. In contrast, the Managerial Theory envisages negative abnormal returns to acquiring firms upon takeover announcements.

Table 1 summarizes key empirical findings on the bidder shareholder wealth effects of takeover announcements. Noticeably, these findings vary significantly depending on time periods, windows examined, geographical locations, and in some cases, event study methods (see Leeth & Borg 2000; Draper & Paudyal 1999).

The Australian literature is largely consistent with takeover announcements having no significant impact on bidder shareholders' wealth (Walter 1984; Bellamy & Lewin 1992; Bugeja & Walter 1995).

Mixed evidence is documented in the US market. Studying the early takeover wave from 1919 to 1930, Leeth and Borg (2000) provide no conclusive evidence as to whether takeover announcements change bidder shareholders' value. (3) While Asquith, Bruner and Mullins (1983) and Jarrell and Poulsen (1989) find small but significantly positive bidder abnormal returns upon takeover announcements, negative announcement impact is documented by Dodd (1980), Servaes (1991) and Walker (2000). Interestingly, Morck, Shleifer and Vishny (1990) and Raad and Wu (1994) find no significant bidder abnormal returns in response to takeover announcements.

Similar puzzling results are obtained with the UK and Europe-based studies. Franks and Harris (1989) find positive but insignificant UK bidders' abnormal returns for the announcement month. Meanwhile, a large and significantly negative takeover announcement effect is documented by Firth (1980) for UK bidders. Studies conducted on European countries report either positive (Campa & Hernando 2004) or insignificant (Goergen & Renneboog 2004) bidder abnormal returns.

Overall, the existing takeover literature is characterized by the lack of consensus regarding the impact of takeover announcements on acquiring firms' equity value. Caution, however, must be exercised in interpreting the above findings. The aggregating effects in calculating abnormal returns may conceal the fact that the market may react differently to takeovers announced by bidders or groups of bidders with different characteristics (Jarrell & Poulsen 1989). Several variables have been proposed to explain the cross-sectional variation in the bidder shareholder wealth effects of takeovers. One such variable is the pre-takeover formation of toeholds, the literature on which is summarized in section 2.2 below.

2.2 Toeholds

Prior theoretical studies seek to explain the impact of toeholds in two types of takeover contests, namely single-bidder and multiple-bidder takeovers. We now review toehold theories related to each type in turn.

The Single-Bidder Model of takeovers examines the wealth effects of toeholds conditional upon the absence of rival bidders. Under this model, owning a toehold increases bidder gains by improving the probability of takeover success, decreasing the bid premium, and reducing the costs of the takeover process. Four main hypotheses have been proposed to explain the impact of toeholds in single-bidder contests, namely the Free-Rider Solution Hypothesis, the Signaling Hypothesis, the Cost-Related Hypothesis, and the Demand-Supply Hypothesis.

The Free-Rider Solution Hypothesis originates from the works of Grossman and Hart (1980) and Shleifer and Vishny (1986) and is further developed by Hirshleifer and Titman (1990). In Grossman and Hart's (1980) single bidding model, each target shareholder's decision to tender shares is assumed to be inconsequential to the takeover outcome. As shareholders fully enjoy the benefits from a successful takeover, they are inclined to retain their shares and free-ride on the realization of synergistic gains. Such rational profit-maximizing decision of individual shareholders might lead to the failure of what would otherwise be a profitable takeover attempt.

Shleifer and Vishny (1986) argue that the ability of the toehold bidder to profit on the value increase of shares held prior to making the offer provides a solution to the free-rider problem. Central to the model of Shleifer and Vishny (1986) is the assumption that target shareholders are not fully informed and can only estimate from the toehold size the value of wealth improvements brought about by the bidder. In equilibrium, this expected value is determined by the bidder whose marginal gains on toeholds equal the marginal loss on the subsequently acquired shares. At the expected value, target shareholders will tender. The model predicts that a low-toehold bidder signals small value improvements and is required to offer a large premium to induce target shareholders to sell their shares. Consequently, a negative relationship between bidders' toeholds and bid premia, and a positive impact of toeholds on success probability are expected. The same conclusions are drawn under Hirshleifer and Titman's (1990) model. Yet unlike Shleifer and Vishny (1986), Hirshleifer and Titman (1990) allow bid premium to be determined independently from the size of bidder toeholds.

While the Free Rider Solution Hypothesis analyzes the impact of toeholds which are taken as given, the Signaling Hypothesis of Chowdhry and Jegadeesh (1994) treats toeholds as endogenously determined. Under the Signaling Hypothesis, target shareholders determine the price at which they tender their shares based on their expectations of synergistic gains brought about by the acquisition. Those expectations are formed on the basis of signals from the size of toeholds. A larger toehold signals larger gains, inducing target shareholders to demand a higher premium while increasing the probability of bid success.

The Cost-Related Hypothesis identifies toeholds as the main source of bidders' gains from takeovers. Bishop (1991) and Kyle and Vila (1991) contend that larger toehold size is associated with higher bidder profits in two ways. First, the target share price appreciation at the takeover announcement means an increase in the value of shares previously held by the bidders. Second, the more shares accumulated by the bidder at a pre-bid price which is lower than the offer price, the lower the total takeover outlay incurred by the bidder.

The Demand-Supply Hypothesis is based on the assumption that the supply of tendered shares is upward sloping (Walking 1985; Stulz 1988; Stulz, Walkling & Song 1990). Under this hypothesis, a large toehold reduces the bidder's net demand for outside shares, thus decreasing the bid premium he has to offer (Stulz 1988). The smaller demand for tendered shares also implies greater negotiating power for the bidder (Hirshleifer & Titman 1990) which further improves the probability of bid success (Walkling 1985).

Notwithstanding the consensus regarding the wealth effects of toeholds in single-bidder takeovers, contradicting toehold hypotheses are advanced for multiple-bidder contests. Under the Overbidding Hypothesis, the toehold bidder has an incentive to bid aggressively, expecting to sell the toehold at high enough prices to the winning rival. This may lead to the unwanted outcome in which the toehold bidder faces the winner's curse of overpaying for the target (Burkart 1995; Singh 1998).

In contrast, Bulow, Huang and Klemperer (1999) contend that toeholds act to deter rivals, thus increasing the probability of bid success. They propose the Rival Deterrence Hypothesis which suggests larger toeholds should increase the initial bidder's aggressiveness. The rival bidder, observing this, bids more conservatively for fear of facing the winner's curse. This, in turn, reduces the winner's curse for the initial bidder, enhancing his chance to win the contest at a favorable price.

To study the impact of toeholds on bidder gains from takeovers, previous studies focus on two aspects. First, there is the relationship between toeholds and bid premia and the probability of takeover success. Second, the association between toeholds and the bidder shareholder wealth effects of takeover announcements. While different theoretical arguments are advanced for the wealth effects of toeholds in single-bidder and multiple-bidder contests, no such distinction is made in empirical studies of toeholds.

Evidence thus far largely supports the idea that toeholds benefit bidders by facilitating lower bid premia and higher probability of bid success. Betton and Eckbo (2000) study 2335 takeover bids in 1353 US tender offer contests from 1971 to 1990, documenting larger toeholds associated with smaller offer premia and larger bidder returns. A negative relation between toeholds and bid premia is also backed by Bris (1998), Asquith and Kieschnick (1999) and Goldman and Qian (2005). In addition, the probability of takeover success is found to increase with the size of toeholds (Walkling 1985; Goldman & Qian 2005; Choi 1991). To our knowledge, except for the work of Franks and Harris (1989), there has been no other empirical evidence of a positive relationship between toeholds and bid premia, Consistent with the Overbidding Hypothesis, Kaplan and Stein (1993) find a positive correlation between large toeholds and overpayment. Yet their sample is limited to management buyouts which may not be representative of takeover decisions.

The substantial wealth of support for a positive impact of toeholds on bidder gains mentioned above suggests a positive association between toeholds and bidder abnormal returns in response to takeover announcements. Surprisingly, empirical results are inconclusive regarding such an association. Franks (1978) finds significantly positive cumulative abnormal returns for US bidders with pre-offer toeholds. Positive toehold effects are also documented in emerging markets by Farinha and Miranda (2003). On the other hand, Dumontier and Petitt (2002), Franks and Harris (1989), and Van Hulle, Vermaelen and De Woulters (1991) find no relationship between toehold size and bidders' abnormal returns in France, the UK, and Belgium respectively. In Australia, the evidence is limited to Canil and Rosser (2004) who find significantly negative abnormal returns surrounding takeover announcements made by toehold bidders. By omitting non toehold bids, however, their study does not directly test the impact of the presence of toeholds on bidder abnormal returns observed upon announcements of takeovers.

3. Research Design

3.1 Methodology

3.1.1 Bidder Abnormal Return Calculations The existing literature employs various event study methodologies to examine the bidder shareholder wealth effects of takeover announcements. As shown in table 2, results obtained from different methodologies may differ even for the same study (Walker 2000; Leeth & Borg 2000). The table also reveals large variations in event windows examined.

In order to ensure the robustness of our results, we employ two event study methodologies to assess the bidder wealth effects. Given the popularity associated with the use of the market model, and to facilitate comparison with results documented in previous studies, we commence by utilizing the 0/1 market model to calculate bidder abnormal returns. (4) However, the 0/1 market model may introduce new listing, rebalancing and skewness biases resulting from the use of the market index as a reference portfolio (Barber & Lyon 1997). Consequently, we recalculate bidder abnormal returns using Barber and Lyon's (1997) control firm approach which addresses these biases. In addition, we employ various event windows that are commonly utilized in prior literature, namely the [-1,+1], [-1,0], [0,+1], [-2,+2], [-2,0], [0,+2], [-5,+5], [-5,0], [0,+5], [-60,+60], [-60,0], and [0,+60] windows.

Abnormal returns are initially assessed for the aggregate sample of takeover announcements. Then, to preliminarily test the impact of toehold ownership on the bidder wealth effects of these announcements, we assess the significance of the difference in average abnormal returns accruing to the toehold and non-toehold samples using the standard two-sample t-test.

3.1.2 The 0/1 Market Model Under the 0/1 market model, daily bidder abnormal returns (ARs) are calculated as follows:

A[R.sub.i,t] = [R.sub.i,t] -[R.sub.m,t] (1)

Where [AR.sub.i,t] is the abnormal return for bidding firm i on day t;[R.sub.i,t], is the actual continuously compounded return for bidding firm i on day t ; and, [R.sub.m,t] is the actual continuously compounded return on the market on day t.

The overall market reaction for each firm over the event window [[T.sub.1],[T.sub.2]], herein referred to as the cumulative abnormal return (CAR), is calculated as:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (2)

The average market response to a sample of n takeover announcements, herein referred to as the sample average cumulative abnormal return (CAAR), is defined as:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (3)

3.1.3 The Control Firm Approach Following Barber and Lyon (1997), we define bidder abnormal returns as the difference between the bidders' and control firms' returns. We initially identify all Australian listed companies with market capitalisation values ranging from 70% to 130% relative to the bidders. Among these firms, control firms are the ones with the closest price-to-book ratios to those of the bidders. (5) Matching of firms is conducted based on accounting information as at the last reporting day preceding takeover announcements. Note that each control firm is uniquely associated with an event firm, while any event firm appearing more than once in the dataset may have more than one control. As a result, 120 control firms are matched against 122 takeover announcements, with two control firms being used twice. In table 3, we present descriptive statistics for the two matching criteria: market values and price-to-book ratios for our event and control samples. Results from a two sample t-test shows no significant difference on average in either size or growth between the event and control firms.

Barber and Lyon (1997) define daily abnormal returns as:

[BLAR.sub.i,t] = [R.sub.i,t] - [R.sub.c,t] (4)

Where [BLAR.sub.i,t] is Barber and Lyon's (1997) abnormal return for bidding firm i on day t; [R.sub.i,t] is the actual discrete return for bidding firm i on day t ; (6) and, [R.sub.c,t] is the actual discrete return for control firm c on day t.

Barber and Lyon's (1997) cumulative abnormal return (BLCAR), calculated for each bidder over the event window [[T.sub.1], [T.sub.2]], is defined as:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (5)

Barber and Lyon's (1997) sample average cumulative abnormal return (BLCAAR) across the event window is calculated as:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (6)

While BLCARs do not suffer from new listing, rebalancing and skewness biases associated with the use of the market index, ignoring compounding effects results in BLCARs being a biased predictor of long run buy-and-hold abnormal returns. In light of this, we also consider Barber and Lyon's (1997) buy-and-hold abnormal returns (BLBHARs), calculated as follows:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (7)

Where [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] is Barber and Lyon's (1997) buy-and-hold abnormal return for bidding firm i over the event window [[T.sub.1],[T.sub.2]]; [R.sub.i,t] is the actual discrete return for bidding firm i on day t ; and, [R.sub.c,t] is the actual discrete return for control firm c on day t.

Thereafter, the Barber and Lyon (1997) average buy-and-hold abnormal return (BLABHAR) for each sample of n announcements is calculated as:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (8)

3.1.4 Multivariate Regression Exploring the impact of toeholds using univariate analyses ignores the potential confounding effects of other factors on the bidder shareholder wealth effects of takeover announcements. Therefore, in order to draw more reliable inferences, we study the impact of toeholds in a multivariate regression framework, where we control for other firm and bid specific characteristics. The characteristics we control for are: offer outcome, method of payment, existence of competition, industry relatedness, relative target size, and bidders' free cash flows. The rationale for the examination of these variables is briefly discussed below.

Offer Outcome: Arguably, offer outcome should affect the magnitude of bidder abnormal returns. Berkovitch and Narayanan (1993) argue that unsuccessful offers should result in bidder shareholder wealth losses due to the costly takeover process. Consistent with this, a number of US studies document significantly larger abnormal returns for successful bidding firms compared to those that are unsuccessful (Asquith, Brunet & Mullins 1983; Dodd & Ruback 1977; Bradley, Desai & Kim 1983).

Method of Payment: The idea of method of payment impacting on the bidder shareholder wealth effects stems from the Signaling Hypothesis suggested by Myers and Majluf (1984). Under this hypothesis, managers only issue stocks if they believe that their firm is currently overvalued. The use of cash financing, therefore, sends a positive signal about the bidding firm to the market (Hansen 1987). The hypothesis predicts larger bidder abnormal returns in response to takeover announcements where cash is the medium of exchange.

Competition: It is widely accepted that competition among rival bidders reduces returns to bidding firms' shareholders. Bradley, Desai and Kim (1988) argue that competing bidders tend to bid the offer price up to the point where all the gains from acquisitions accrue to the targets, leaving the successful bidders with at most zero gain. De, Fedinia and Triantis (1996) find evidence supporting this argument by documenting significantly negative abnormal returns to successful bidders in multiple-bidder contests as opposed to the negligible returns to those in single-bidder contests.

Industry Relatedness: Singh and Montgomery (1987), among others, argue that acquiring targets operating in related industries generates larger gains for bidders given the potential for economies of scale, greater synergy and complementary resources. In accordance with this line of reasoning, and with the supporting evidence documented by Walker (2000) and Doukas, Holmen and Travlos (2002), we expect to find higher abnormal returns accruing to bidders who acquire related targets.

Relative Target Size: Kuehn (1975) is among the first to suggest a negative relationship between target firm size and bidder abnormal returns. He argues that acquiring smaller targets involves lower acquisition and incorporation costs, resulting in higher bidder gains from takeovers. Conversely, Asquith, Bruner and Mullins (1983) contend that greater target relative size is associated with larger bidder abnormal returns. In fact, unless the investment in the target is sufficiently large relative to the bidder, the expected value improvements from the merger will realize insignificant changes in the bidder's share price. Consistent with this prediction, Asquith, Bruner and Mullins (1983), Jarrell and Poulsen (1989), and Loderer and Martin (1990) find evidence that US bidder abnormal returns increase with relative target size. Similar results are reported for the UK and Canada (see Franks & Harris 1989; Eckbo & Thorburn 2000).

Bidder's Free Cash Flows." The Free Cash Flows Hypothesis advanced by Jensen (1988) posits that in the presence of agency problems, firms with large free cash flows tend to waste resources on value-reducing investments including takeovers. As a result, we expect to find a negative relationship between the size of bidders' free cash flows and abnormal returns.

The regression model is formally presented as follows:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (9)

Where: [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] = abnormal return for bidding firm i over the event window [[T.sub.1], [T.sub.2]];

THLD = dummy variable (equal to 1 if the bidder holds a toehold in the target on the takeover announcement day and 0 otherwise);

CON = a dummy variable (equal to 1 if the toehold formation is announced within the [-10,0] window and 0 otherwise);

COMP = dummy variable equal to 1 if there are competing bidders and 0 otherwise;

SUCCESS = dummy variable equal to 1 if the takeover is successful and 0 otherwise;

CASH = dummy variable (equal to 1 if cash is the only method of payment and 0 otherwise);

FCF = dummy variable equal to 1 if there are competing bidders and 0 otherwise;

RELATED = the size of bidder's free cash flows as at the last reporting day preceding takeover announcement, calculated as (bidder's net profit after tax +depreciation/ amortisation expense-dividends)/market capitalisation; and,

LNRSIZE = the relative target size as at the last reporting day preceding takeover announcement. (7)

We examine the toehold impact on bidder abnormal returns via the inclusion of the dummy variable THLD and the interactive variable COMP*THLD. The coefficient on THLD ([[beta].sub.1]) represents the toehold effects uncontaminated by the presence of bidding rivals and is expected to be positive under the Single-Bidder Model of Shleifer and Vishny (1986). Meanwhile, the coefficient on COMP*THLD ([[beta].sub.3]) captures the impact of competition on toehold ownership desirability. Under Bukart's (1995) and Singh's (1998) Overbidding theory, [[beta].sub.3] is expected to be significantly negative, implying reduced toehold attractiveneness when rival bidders are present. Conversely, the Rival Deterrence Hypothesis of Bulow, Huang and Klemperer (1999) predicts [[beta].sub.3] to be insignificant, a notion consistent with the idea that owning a toehold is beneficial regardless of the presence of competition.

To further examine the impact of toeholds on bidder abnormal returns, we also include two control variables, which we form by interacting THLD with CON and SUCCESS. The interaction variable CON*THLD takes into account the fact that the timing of toehold formation may influence the relationship between toeholds and bidder gains. For deferred bids, market reaction to toehold information on the takeover announcement day may just represent adjustment of prior beliefs (Bishop 1991; Jarrell & Poulsen 1989). As such, we would expect the impact of toeholds to differ significantly between concurrent and deferred bids. Supporting this, Canil and Rosser (2004) examine market reaction to bids announced by concurrent and deferring bidders, documenting significantly lower abnormal returns to the later.

The control variable SUCCESS*THLD is formed to capture the market's preference with regard to offer outcome for toehold bids. Arguably, bidders tend to purchase a toehold before announcing takeovers for two main reasons. First, they hope to enjoy the benefits of toeholds in the form of lower takeover outlay, lower takeover premia or higher probability of bid success. Alternatively, they may wish to profit from selling the toeholds at favourable prices to rival bidders or to subsequent bidders should the bid become unsuccessful. As a result, the coefficient on SUCCESS*THLD will be positive should the market see the former purpose as creating more wealth than the latter, and vice versa.

As a sensitivity test, we rerun regression (9) where we replace the dummy variable THLD with the continuous variable TOESIZE. We define TOESIZE as the proportion of target shares held by the bidders upon takeover announcements. We expect takeover announcements to create a greater price impact the larger the size of the toeholds.

3.2 Data

We identify takeover announcements from ImageSignal Database, Connect4 and the Australian Financial Review. The initial dataset comprises 489 takeover announcements made between 1997 and 2004 inclusive by 394 bidders where at least one party to the transaction is an Australian listed company. We exclude all announcements where: (i) The target is not listed on the ASX; (8) (ii) Share price information for the bidder is missing; (iii) Accounting information for the bidder is missing; (iv) The bidder holds existing interests in excess of 50% of the targets' outstanding ordinary shares prior to launching the bid; (9) (v) The bid is conducted jointly by two or more bidders; (vi) The announcement is a revised takeover offer following an initial failed or withdrawn bid; and, (vii) The bidder makes multiple bids over a twelve-month period. (10)

Applying the above filters, our dataset is reduced to 122 takeover announcements made by 110 companies listed on the Australian Stock Exchange. A breakdown of the observations removed by each filter is provided in table 4.

We collect share prices from Thomson Financial Datastream and financial statement data from Aspect FinAnalysis. Deal values, offer outcome and SIC codes are collected from the SDC Database. The presence of toeholds on the announcement date and the method of payment are determined from takeover announcement texts. (11) Toehold announcement dates are identified from forms 603 Becoming a Substantial Shareholder. These forms are obtained via two sources: Aspect DatAnalysis database of shareholder notices for forms disclosed from 1999 onwards; and the Australian Financial Review for forms released prior to 1999. The toehold announcement date is defined as the earliest date the market is deemed to be informed about the toehold formation, either via press release of substantial shareholder notices, or via initial takeover announcements. (12)

Table 5 presents the breakdown for the dummy variables employed in multivariate analyses. Noticeably, more than half of the takeover announcements in our sample are made by bidders with ownership interests in targets' outstanding shares upon launching the bids (toehold bids). For these 69 toehold bids, the average toehold size at takeover announcements is 16.73%; 45 lead to successful takeovers; and, 16 bids are made in the presence of rival bidders. Additionally, more than half (39) of the toehold bids are concurrent, where toehold formation is disclosed on the same day or within the 10-day period leading up to takeover announcements. A summary of the timing of toehold formation is provided in table 6.

Table 7 summarises the descriptive statistics for the two continuous variables included in the study, namely relative target size and bidders' free cash flows. The correlation between the continuous variables of -0.1814 indicates a negative relationship between relative target size and bidders' free cash flows. Yet the small absolute magnitude of the relationship suggests that multicollinearity does not seem to be a problem.

4. Empirical Results

4.1 Univariate Results

Table 8 presents the results of abnormal returns calculated on the entire sample of takeover announcements. Consistent with previous Australian studies (Walter 1984; Bellamy & Lewin 1992; Bugeja & Walter 1995), we document insignificant bidder abnormal returns for the short event windows surrounding takeover announcements. Further, while the 0/1 market model suggests significant longer term bidder abnormal losses, this finding is not supported by Barber and Lyon's (1997) method. In light of the biases inherent in the 0/1 market model, we place heavier weighting on results obtained under Barber and Lyon's (1997) method. On this basis, we conclude that bidders on average do not earn significant abnormal returns in response to takeover announcements.

To preliminarily examine whether toehold ownership affects the magnitude of bidder gains from takeovers, we assess the difference in market reaction to announcements made by two groups of bidders. The toehold group includes bids in which bidders own toeholds in the targets at the announcements of takeovers. The non-toehold group, in contrast, comprises bidders who do not own a toehold in their targets when launching the bid. Results from this comparison are presented in Table 9.

The market reaction is consistent with the idea that owning a toehold increases bidder gains from takeovers. In fact, over the [0, +2] window, average abnormal returns enjoyed by toehold bidders are 2.3%- 2.5% higher than those accruing to the non-toehold ones. This result is statistically significant regardless of the method of abnormal return calculation employed. Although a longer term positive toehold impact is not documented under the 0/1 market model, it is present under the Barber and Lyon's (1997) approach. In particular, the latter method implies the difference in the sixty-one-day abnormal returns ranging from 10% to 12%. Taken together, these findings preliminarily suggest that the market views toeholds as beneficial for the bidders. Moreover, the toehold impact is not fully incorporated into share prices immediately upon takeover announcements.

4.2 Multivariate Results

While univariate results suggest possible toehold impact on the bidder abnormal returns, it is important to examine whether this impact exists after controlling for other explanatory factors. We achieve this via the fitting of multivariate regression models, where bidder abnormal returns are modeled as a function of toeholds and other firm and bid specific characteristics. Further, to account for potential heteroskedasticity in the sample, we employ White's (1980) standard errors.

A summary of statistically significant findings from multivariate regressions across twelve event windows are presented in table 10 below. In the interest of brevity, we only present detailed results for the [0,+1] and [0,+60] windows in tables 11 and 12 respectively. These windows are representative of the other short and long windows examined.

4.2.1 Toeholds Table 10 confirms univariate results reported earlier by documenting a significantly positive coefficient on THLD over the [0,+2] window. This short run positive association between the presence of toeholds and the magnitude of abnormal returns for unrivaled bidders is also found over the [0,+1] window (see table 11). Further, table 12 reveals the persistence of the toehold impact over the longer event window of [0,+60], indicating incomplete market reaction at the takeover announcements. Our evidence is in accordance with theoretical predictions under the Single-Bidder model of Shleifer and Vishny (1986) and Hirshleifer and Titman (1990). Specifically, facing no competition in the bidding process, toehold ownership should benefit bidders in terms of cost reduction and greater probability of bid success.

Results over the [0,+1] window shows different toehold effects in single-bidder and multiple-bidder takeovers. The significantly negative estimate for [[beta].sub.3] suggests that owning a toehold is less beneficial for bidders facing competition from bidding rivals. Immediate market reaction seems consistent with Bukart's (1995) and Singh's (1998) Overbidding Hypothesis. Specifically, the market expects the presence of competing bidders to reduce the attractiveness of toehold ownership through a higher probability of overbidding. The negative market reaction to higher potential overbidding costs, however, is mediated by the positive reassessment of the value of the bidder's existing toeholds. In fact, we document insignificant toehold impact on the magnitude of abnormal returns for rivaled bidders, evidenced by the fact that [[beta].sub.1] + [[beta].sub.3] is not significantly different from zero (13).

Results over the [0,+60] window, however, do not strongly suggest a difference in the toehold impact between rivaled and unrivaled bidders. In fact, while [[beta].sub.1] is still positive and significant, [[beta].sub.3] is no longer significant at the 5% level under the Barber and Lyon's (1997) buy-and-hold method.

Table 12 further shows that the toehold effect also depends on the bid outcome. The coefficient on SUCCESS*THLD is strongly negative and significant under both the control firm approach and the 0/1 market model. Given this, the market appears to prefer toehold bids which are unsuccessful, in which case toeholds can be sold to rival bidders or to subsequent bidders for the same targets. In other words, the gain from buying low and selling high of toeholds seems to surpass that derived from using toeholds to facilitate a successful takeover.

4.2.2 Other Findings Our results indicate that the magnitude of bidder abnormal returns depend on other factors apart from toehold ownership. These factors include bid outcome, method of payment, existence of competition and bidders' free cash flows. Specifically, we document superior performance of successful bidders following takeover announcements. Consistent with Asquith, Bruner and Mullins (1983), we conclude that unsuccessful costly takeovers are perceived as bad news. The market duly adjusts its valuation of takeovers over time when the probability of success becomes more evident.

We also find that all-cash bidders experience smaller abnormal returns in the sixty-day period before the announcements (see table 10) but fare better than bidders issuing stock in the period that follows (see table 12). The post-announcement positive cash impact corroborates the Signaling Hypothesis of Myers and Majluf (1984) and the Australian empirical evidence of da Silva Rosa, Izan, Steinbeck and Walter (2000). The negative cash effect in the pre-announcement period, however, is consistent with Bujega and Walter (1995). It also supports Martin's (1996) view that bidding firms' growth opportunities, as indicated by their pre-takeover performance, could affect their method of payment decision. In particular, bidding firms with higher growth opportunities prefer the use of stock over cash and debt due to the greater investing flexibility associated with equity financing.

Examination of table 10 further reveals the positive impact of the existence of competition on the bidder abnormal returns over a variety of short and long windows. While we do not depend on these results obtained from the 0/1 model, a similar observation is made under the [-60,+60] window when wealth effects are calculated using the Barber and Lyon's (1997) buy-and-hold method. Our results do not support Bradley, Desai and Kim's (1988) prediction that competition should reduce bidder gains by inducing higher offer prices. Nevertheless, we argue that the existence of bidding competition might signal the value of the target. Specifically, good targets tend to attract takeover offers from multiple bidders who vie for their control. As such, bidders who engage in multiple-bidder contests, viewed by the market as more likely to be undertaking worthwhile investments, will enjoy higher abnormal returns.

While tables 11 and 12 show no evidence of free cash flows affecting the bidder shareholder wealth effects of takeover announcements, Barber and Lyon's (1997) method supports a positive relationship over the [-60,0] window (see table 10). This finding is inconsistent with Jensen's (1986) proposition that managers have an incentive to squander the firm's resources in wealth decreasing takeovers. Rather, it suggests that takeovers are value-creating to bidder shareholders. The market consequently rewards more for takeovers announced by bidding firms with greater agency problems indicated by higher levels of free cash flows.

4.3 Sensitivity Test Results

In addition to suggesting that bidder gains depend on the presence of bidder's toehold ownership, previous theoretical work also implies that toehold size should affect the magnitude of these gains. To ensure that our results are not sensitive to different measures of the toehold variables, we re-perform the multivariate tests where the toehold impact is measured by toehold size. Results from these sensitivity tests are reported in tables 13 and 14 for windows [0,+1] and [0,+60] respectively. Results for the other windows are available from the authors upon request. We find that our original conclusions remain largely unchanged. The only notable exception is [[beta].sub.3] being insignificant over the [0,+60] windows under both the Barber and Lyon's (1997) cumulative and buy-and-hold approaches. This may be consistent with the Rival Deterrence Hypothesis of Bulow, Huang and Klemperer (1999). In particular, owning a toehold may actually lower the probability of overpayment by inducing conservative bidding from rival bidders. As such, owning a toehold may be as beneficial in a multiple-bidder takeover as it is in a single-bidder takeover.

5. Conclusion

This study is motivated by: (a) the lack of consensus in the extant literature regarding bidder abnormal returns in response to takeover announcements; (b) an absence of Australian research focusing on toehold impact; and, (c) inadequate modelling of toeholds in studies elsewhere. Employing the new methodology advanced by Barber and Lyon (1997) on a dataset of takeover announcements made by Australian listed companies between 1997 and 2004, we document no significant association between takeover announcements and bidder shareholders' wealth. The negative bidder abnormal returns documented under the 0/1 market model are not supported by the bias-free Barber and Lyon's (1997) method. This may suggest that the negative market reaction documented in some previous studies could be an artefact of the flaws associated with the use of the market model.

Our results also indicate a positive association between toehold ownership/ toehold size and abnormal returns accruing to bidders engaging in single-bidder takeovers. Moreover, immediate market reaction seems consistent with the idea that the impact of toehold on bidder abnormal returns is dependent on the presence of rival bidders. Specifically, while significantly positive toehold impact is documented for those engaging in single-bidder takeovers, the attractiveness of toeholds in multiple-bidder contests is eliminated by the increased probability of overbidding by the toehold bidder. This differential toehold impact may explain the finding of insignificant toehold impact in some previous studies focusing on the short term announcement effects (see, Franks & Harris 1989). However, market expectation of overbidding for toehold bidders is incorrect in the long run, where toeholds prove beneficial regardless of the presence of rival bidders.

This paper is based on Hanh Le's Honours thesis at the Australian National University, supervised by Emma Schultz. Our tremendous thanks go to Tom Smith for insightful comments and suggestions. We would also like to thank Terry Walter, Sian Owen and participants at the AFAANZ 2006 conference for helpful comments on earlier drafts of this paper. All errors are our own.

(Date of receipt of final transcript: June 19, 2007. Accepted by David Gallagher & Garry Twite, Area Editors.)

References

Asquith, D. & Kieschnick, R. 1999, 'An examination of initial shareholdings in tender offer bids', Review of Quantitative Finance and Accounting, vol. 12, pp. 171-88.

Asquith, P., Bruner, R.F. & Mullins, D.W. 1983, 'The gains to bidding firms from merger', Journal of Financial Economics, vol. 11, pp. 121-39.

Barber, B. & Lyon, J. 1997, 'Detecting long-run abnormal stock returns: The empirical power and specification of test statistics', Journal of Financial Economics, vol. 43, pp. 341-72.

Bellamy, D.E. & Lewin, W.M. 1992, 'Corporate takeovers, method of payment and bidding firm's shareholder returns: Australian evidence', Asia-Pacific Journal of Management, vol. 9, pp. 137-50.

Berkovitch, E. & Narayanan, M. 1993, 'Motives for takeovers: An empirical investigation', Journal of Financial and Quantitative Analysis, vol. 28, pp. 347-62.

Betton, S. & Eckbo, B.E. 2000, 'Toeholds, bid jumps and expected payoffs in takeovers', Review of Financial Studies, vol. 13, pp. 841-82.

Bishop, S.R. 1991, 'Pre-bid acquisitions and substantial shareholder notices', Australian Journal of Management, vol. 16, pp. 1-33.

Bradley, M., Desai, A. & Kim, E. 1983, 'The rationale behind interfirm tender offers: Information or synergy?', Journal of Financial Economics', vol. 11, pp. 183-206.

Bradley, M., Desai, A. & Kim, E. 1988, 'Synergistic gains from corporate acquisitions and their division between the stockholders of target and acquiring firms', Journal of Financial Economics, vol. 21, pp. 3-40.

Bris, A. 1998, 'When do bidders purchase a toehold? Theory and tests', working paper, Yale University, USA.

Bugeja, M. & Walter, T. 1995, 'An empirical analysis of some determinants of the target shareholder premium in takeovers', Accounting and Finance, vol. 35, pp. 33-60.

Bulow, J., Huang, M. & Klemperer, P. 1999, 'Toeholds and takeovers', Journal of Political Economy, vol. 107, pp. 427-54.

Burkart, M. 1995, 'Initial shareholdings and overbidding in takeover contests', Journal of Finance, vol. 50, pp. 1491-515.

Campa, J.M. & Hernando, I. 2004, 'Shareholder value creation in european M&Ss', European Financial Management, vol. 10, pp. 47-81.

Canil, J.M. & Rosser, B.A. 2004, 'Toehold', African Finance Journal, vol. 6, pp. 37-55.

Choi, D. 1991, 'Toehold acquisitions, shareholder wealth, and the market for corporate control', Journal of Financial and Quantitative Analysis, vol. 26, pp. 391-407.

Chowdhry, B. & Jegadeesh, N. 1994, 'Pre-tender share acquisition strategy in takeovers', Journal of Financial and Quantitative Analysis', vol. 29, pp. 117-29.

Da Silva Rosa, D., Izan, H.Y., Steinbeck, A. & Walter, T. 2000, 'The method of payment decision in Australian takeovers: An investigation of causes and effects', Australian Journal of Management, vol. 25, pp. 67-94.

De, S., Fedinia, M. & Triantis, A. 1996, 'Effects of competition on bidder returns', Journal of Corporate Finance. Contracting, Governance and Organization, vol. 2, pp. 261-82.

Dodd, P. & Ruback, R. 1977, 'Tender offers and stockholder returns', Journal of Financial Economics, vol. 5, pp. 351-73.

Dodd, P. 1980, 'Merger proposals, management discretion and stockholder wealth', Journal of Financial Economics, vol. 8, pp. 105-37.

Doukas, J.A., Holmen, M. & Yravlos, N.G. 2002, 'Diversification, ownership and control of Swedish corporations', European Financial Management, vol. 8, pp. 281-314.

Draper, P. & Paudyal, K. 1999, 'Corporate takeovers: Model of payment, returns and trading activity', Journal of Business Finance and Accounting, vol. 26, pp. 521-58.

Dumontier, P. & Petitt, B.S.P. 2002, 'Determinants of returns of acquiring firms around tender offer announcements: Evidence from French control-oriented and parent-subsidiary offers', paper presented at EFMA 2002 London Meetings, London, UK.

Eckbo, B. & Thorburn, K. 2000, 'Gains to bidding firms revisited: Domestic and foreign acquisitions in Canada', Journal of Financial and Quantitative Review, vol. 35, pp. 1-25.

Fama, E.F. 1970, 'Efficient capital markets: A review of theory and empirical work', Journal of Finance, vol. 25, pp. 383-417.

Farinha, J. & Miranda, F. 2003, 'Run-up, toeholds, and agency effects in mergers and acquisitions: Evidence from an emerging market', working paper, Faculdade De Economica, Universidade do Porto.

Firth, M. 1980, 'Takeovers, shareholder returns, and the theory of the firm', Quarterly Journal of Economics, vol. 94, pp. 235-60.

Franks, J.R. & Harris, R.S. 1989, 'Shareholder wealth effects of corporate takeovers: The UK experience 1955-1985', Journal of Financial Economics, vol. 23, pp. 225-49.

Franks, J.R. 1978, 'Insider information and the efficiency of the acquisitions' market', Journal of Banking and Finance, vol. 2, pp. 379-93.

Goergen, M. & Renneboog, L. 2004, 'Shareholder wealth effects of European domestic and cross-border', European Financial Management, vol. 10, pp. 9-45.

Goldman, E. & Qian, J. 2005, 'Optimal toeholds in takeover contests', Journal of Financial Economics, vol. 77, pp. 321-46.

Grossman, S.J. & Hart, O.D. 1980, 'Takeover bids, the free-rider problem, and the theory of the corporation', Bell Journal of Economics, vol. 11, pp. 42-64.

Hansen, R.G. 1987, 'A theory for the choice of exchange medium in mergers and acquisitions', Journal of Business, vol. 60, pp. 75-95.

Hirshleifer, D.S. & Titman, 1990, 'Share tendering strategies and the success of hostile takeover bids', Journal of Political Economy, vol. 98, pp. 295-324.

Jarrell, G.A. & Poulsen, A.B. 1989, 'The returns to acquiring firms in tender orders: Evidence from three decades', Financial Management, vol. 18, pp. 12-9.

Jensen, M.C. 1986, 'Agency costs of free cash flows, corporate finance and takeovers', America Economic Review, vol. 76, pp. 323-29.

Jensen, M.C. 1988, 'Takeover: Their causes and consequences', Journal of Economic Perspective, vol. 1, pp. 21-48.

Kaplan, S.N. & Stein, J.C. 1993, 'The evolution of buyout pricing and financial structure in the 1980s', Quarterly Journal of Economics, vol. 108, pp. 313-52.

Kuehn, D. 1975, Takeovers and the Theory of the Firm, Macmillan, London.

Kyle, A. & Vila, J.L. 1991, 'Noise trading and takeovers', Rand Journal of Economics, vol. 22, pp. 36-53.

Leeth, J.D. & Borg, J.R. 2000, 'The impact of takeovers on shareholder wealth during the 1920s merger wave', Journal of Financial and Quantitative Analysis, vol. 35, pp. 217-38.

Loderer, C. & Martin, K. 1990, 'Corporate acquisitions by listed firms: The experience of a comprehensive sample', Financial Management, vol. 19, pp. 17-33.

Malatesta, P.H. & Thompson, R. 1985, 'Partially anticipated events: A model of stock price reactions with an application to corporate acquisitions', Journal of Financial Economics, vol. 14, pp. 237-50.

Martin, K.J. 1996, 'The method of payment in corporate acquisitions, investments opportunities, and management ownership', Journal of Finance, vol. 51, pp. 1227-46.

Morck, R., Shleifer, A. & Vishny, R. 1990, 'Do managerial objectives drive bad acquisitions?', The Journal of Finance, vol. 45, pp. 31-48.

Myers, S. & Majluf, N. 1984, 'Corporate financing and investments decisions when firms have information that investors do not', Journal of Financial Economics, vol. 13, pp. 187-221.

Raad, E. & Wu, H.K. 1994, 'Acquiring firms' stock returns: Methods of payment, change in leverage, and management ownership', Journal of Economics and Finance, vol. 18, pp. 13-24.

Roll, R. 1986, 'The Hubris hypothesis of corporate takeovers', Journal of Business, vol. 59, pp. 197-216.

Servaes, H. 1991, 'Tobin's Q and the gains from takeovers', Journal of Finance, vol. 46, pp. 409-20.

Shleifer, A. & Vishny, R.W. 1986, 'Large shareholders and corporate control', Journal of Political Economy, vol. 94, pp. 461-88.

Simmonds, D.P. 2004, 'The impact of takeover offer timing on the measurement of Australian bidder gains: 1976 to 1995', Australian Journal of Management, vol. 29, pp. 1-60.

Singh, H. & Montgomery, C.A. 1987, 'Corporate acquisition strategies and economic performance', Strategic Management Journal, vol. 8, pp. 377-86.

Singh, R. 1998, 'Takeover bidding with toeholds: The case of the owner's curse', Review of Financial Studies', vol. 11, pp.679-704.

Stulz, R. 1988, 'Managerial control of voting right: Financing policies and the market for corporate control', Journal of Financial Economics', vol. 20, pp. 25-54.

Stulz, R., Walkling, R.A. & Song, M.H. 1990, 'The distribution of target ownership and the division of gains in successful takeovers', Journal of Finance, vol. 45, pp. 817-33.

Van Hulle, C., Vermaelen, T. & De Woulters, P. 1991, 'Regulation, taxes and the market for corporate control in Belgium', Journal of Banking and Finance, vol. 15, pp. 1143-70.

Walker, M.M. 2000, 'Corporate takeovers, strategic objectives, and acquiring-firm shareholder wealth', Financial Management, vol. 29, pp. 53-66.

Walkling, R.A. 1985, 'Predicting tender offer success: A logistic approach', Journal of Financial and Quantitative Analysis, vol. 20, pp. 461-78.

Walter, T.S. 1984, 'Australian takeovers: Capital market efficiency and shareholder risk and return', Australian Journal of Management, vol. 9, pp. 63-95.

White, H. 1980, 'A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity', Econometrica, vol. 48, pp. 817-38.

by

Hanh T. Le [dagger]

Emma Schultz [section]

([dagger]) Stern School of Business, New York University. Email: [e-mail] 

([section]) School of Finance and Applied Statistics, College of Business and Economics, Australian National University

(1.) Toeholds are commonly defined as the bidder's pre-takeover equity interests in the target (see Franks & Harris 1989).

(2.) Readers are advised to consult Berkovitch and Narayanan (1993) for an excellent review of takeover theories.

(3.) Their results are sensitive to methods utilised to calculate abnormal returns. When the 0/1 market model is considered, bidders earn small but significant abnormal returns of 1%. The standard market model, however, yields bidder abnormal returns insignificantly different from zero.

(4.) The use of the 0/1 market model is argued to be superior to the standard market model for this type of study. Specifically, bidders are found to perform strongly in the period preceding takeover announcements (Malatesta & Thompson 1985). This strong performance, if transitory, should result in upward biased market model expected returns and hence downward biased abnormal returns (Leeth & Borg 2000; Simmonds 2004). The 0/1 market model does not rely on past performance to estimate expected returns, thereby providing a more appropriate measure of bidder abnormal returns in response to takeover announcements.

(5.) Instead of matching firms based on size and book-to-market ratios as proposed by Barber and Lyon (1997), we select control firms on the basis of size and price-to-book ratios. The use of price-to-book ratios is determined by data availability. As price-to-book ratios are merely the inverse of book-to-market ratios, we argue that the choice between the two does not affect the matching results.

(6.) Discrete returns are calculated as

[R.sub.t] = ([P.sub.t] - [P.sub.t-1]/[P.sub.t-1]) where [R.sub.t] = the discrete return on day t; [P.sub.t] = the total return index on day t; and, [P.sub.t-1] = the total return index on day t - 1.

(7.) Relative target size is calculated as the natural logarithm of the ratio of target's market capitalisation to bidder capitalisation. Where target's market capitalisation information is unavailable, the value of the deal as reported by SDC is used as a proxy for target size.

(8.) As this study focuses on the relationship between toeholds and the bidder shareholder wealth effects of takeover announcements, we exclude all takeovers where shareholder information prior to takeover announcements cannot be ascertained. Section 710(4) of the Corporations Law mandates the disclosure of substantial shareholding only when the target is listed on the ASX.

(9.) The 50% controlling interest is the first trance of corporate control in Australia, allowing the holder to pass an ordinary resolution.

(10.) If a bidder announces bids for more than one target over a twelve-month period, only the announcement for the first target is included in the final sample.

(11.) A toehold is deemed to be present given the bidder's existing interests in the target's outstanding ordinary shares immediately prior to the takeover announcement.

(12.) For a number of cases where toeholds are formed just prior to takeover announcements, the two-lagged days allowed by the Corporations Law for the lodgment of substantial shareholder notices result in toeholds being announced subsequently to the takeover announcements. As takeover announcements usually contain information regarding the existing level of the bidder's stake in the target, for those cases, the takeover announcement date is treated as the announcement date of toehold formation. For the one case where pre-announcement toehold size is smaller than 5%, no toehold announcement is required by law and thus the takeover announcement date is also the date of toehold announcement.

(13.) For the [0,+1] window, the Wald coefficient test of the significance of [[beta].sub.1] + [[beta].sub.3] yields F-statistics of 0.0587, 0.0381, and 0.0311 in the regressions with dependent variables being the 0/1 market model CARs, BLCARs, and BLBHARs respectively. Consequently, the test fails to reject the hypothesis that the presence of toeholds has no significant impact on abnormal returns for bidders facing competition.

Table 1
Empirical Evidence on the Bidder Shareholder Wealth Effects of
Takeover Announcements

Table 1 presents the findings from event studies that examine the
bidder shareholder wealth effects of takeover announcements. Study
shows the reference for the study. Period studied shows the years over
which the study is conducted. Return interval shows the frequency over
which abnormal returns are computed. Event window shows the period
over which cumulative abnormal returns are calculated where 0 is the
date, week or month of the takeover announcement. CAAR is the
cumulative average abnormal return for bidding firms over the event
window. Country shows the geographic location from which the sample in
the study is drawn.

Study                         Period      Return
                              Studied     Interval

Leeth and Borg (2000)         1919-1930   Monthly
                              1919-1930   Monthly
Asquith, Bruner and Mullins   1963-1979   Daily
  (1983)
Jarrell and Poulsen (1989)    1960-1969   Daily
                              1970-1979   Daily
                              1980-1989   Daily
Dodd (1980)                   1971-1977   Daily
Servaes (1991)                1972-1987   Daily
Morck, Shleifer and Vishny    1975-1987   Daily
  (1990)
Radd and Wu (1994)            1981-1986   Daily
Walker (2000)                 1980-1996   Daily
Franks and Harris (1989)      1955-1985   Monthly
Firth (1980)
  successful bidders          1969-1975   Monthly
unsuccesful bidders           1969-1975   Monthly
Draper and Paudyal (1999)     1988-1996   Daily
                              1988-1996   Daily
Walter (1984)                 1966-1972   Weekly
Bellamy and Lewin (1992)      1980-1988   Daily
Bugeja and Walter (1995)      1981-1989   Daily
Campa and Hernando (2004)     1998-2000   Daily
Goergen and Renneboog (2004)  1993-2000   Daily

Study                          Number of        Event Window
                              Observations

Leeth and Borg (2000)             466               [0]
                                  466               [0]
Asquith, Bruner and Mullins       214              [-1,0]
  (1983)
Jarrell and Poulsen (1989)        106             [-10,+5]
                                  140             [-10,+5]
                                  159             [-10,+5]
Dodd (1980)                       126              [-1,0]
Servaes (1991)                    384        [0, delisting day]
Morck, Shleifer and Vishny        326              [-l,l]
  (1990)
Radd and Wu (1994)                105             [-1, 0]
Walker (2000)                     278             [-2,+2]
Franks and Harris (1989)           51               [0]
Firth (1980)
  successful bidders              434               [0]
unsuccesful bidders               129               [0]
Draper and Paudyal (1999)         256             [-1,+1]
                                  256             [-1,+1]
Walter (1984)                     368               [0]
Bellamy and Lewin (1992)          210            [-10,+10]
BugejaandWalter (1995)             78             [-60,+1]
Campa and Hernando (2004)         262             [-1,+1]
Goergen and Renneboog (2004)      142             [-2,+2]

Study                         CAAR           Country

Leeth and Borg (2000)          0.17 (a)      US
                               1.27 ** (b)   US
Asquith, Bruner and Mullins    0.9 ***       US
  (1983)
Jarrell and Poulsen (1989)     4.4 ***       US
                               1.22 ***      US
                              -1.1           US
Dodd (1980)                   -1.16 ***      US
Servaes (1991)                -1.07 **       US
Morck, Shleifer and Vishny    -0.7           US
  (1990)
Radd and Wu (1994)            -0.29          US
Walker (2000)                 -0.84 *        US
Franks and Harris (1989)       1.00          UK
Firth (1980)
  successful bidders          -6.3 ***       UK
unsuccesful bidders           -6.0 ***       UK
Draper and Paudyal (1999)     -1.14 **       UK
                              -0.64          UK
Walter (1984)                 -0.3           Australia
Bellamy and Lewin (1992)       1.60          Australia
BugejaandWalter (1995)        -1.8           Australia
Campa and Hernando (2004)      0.44          Europe
Goergen and Renneboog (2004)   1.18 ***      Europe

Note: *** Significant at 1% level;

** Significant at 5% level;

* Significant at 10% level;

(a) CAAR is calculated using the standard market model; and

(b) CAAR is calculated using the 0/1 market model.

Table 2
Summary of Event Study Methodologies

Table 2 summarises the event study methodologies employed in the
extant literature to examine the bidder shareholder wealth effects of
takeovers announcements. Study refers to the reference for the study.
Method refers to the event study methodology employed to examine
wealth effects. Window refers to the period for which cumulative
abnormal returns are calculated. Return Interval shows the frequency
for which returns are computed. CAAR shows the reported average
cumulative abnormal bidder return over the event window.

Study                          Method                    Window

Dodd (1980)                    Market model              [-1,0]
Asquith, Bruner and            Beta control            [-480,+20]
  Mullins (1983)                 portfolio
Walter (1984)                  Market model             [-100,0]
                               Market model               [0]
Jarrell and Poulsen (1989)     Market model             [-2,+1]
                               Market model            [-10,+20]
Morck, Shleifer and            Market model             1-11+11
  Vishny (1990)
Servaes (1991)                 Market model        [0, delisting day]
Walker (2000)                  Market model             [-2,+2]
                               Size Matched firm        [-2,+2]
Leeth and Borg (2000)          Market model                0
                               0/1 market model            0
Goergen and Renneboog (2004)   Market model              [-1,0]
                               Market model             [-2,+2]
                               Market model             [-40,0]
                               MM                      [-60,+60]

Study                          Return Interval     CAAR

Dodd (1980)                    Daily               -1.16 **
Asquith, Bruner and            Daily                9.20
  Mullins (1983)
Walter (1984)                  Weekly              30.7 **
                               Weekly              -0.3
Jarrell and Poulsen (1989)     Daily                0.70 **
                               Daily                1.96 ***
Morck, Shleifer and            Daily               -0.7
  Vishny (1990)
Servaes (1991)                 Daily               -1.07 **
Walker (2000)                  Daily               -0.84 *
                               Daily               -0.77
Leeth and Borg (2000)          Monthly              0.17
                               Monthly              1.27 **
Goergen and Renneboog (2004)   Daily                2.98 ***
                               Daily                3.18 ***
                               Daily                0.64
                               Daily               -0.26

Note: *** Significant at 1%;

** Significant at 5%; and,

* Significant at 10%.

Table 3
Descriptive Statistics of Market Capitalization and Price-to-Book
Values for 122 Takeover Announcements and Control Firms

Table 3 presents descriptive statistics for two matching criteria:
market capitalization and price-to-book values for the samples of 122
takeover announcements and their control firms. Market capitalisation
and price-to-book values are obtained from the reporting period
immediately preceding the takeover announcement date.

                      Market capitalisation
                          (A$ million)

                  Event Sample   Control Sample

Mean                   1821.06          1806.70
SD                     4167.16          4329.78
Min                       1.20             0.95
25th percentile          65.19            56.20
Median                  338.52           300.98
75th percentile        1898.18          1766.46
Max                   34702.88         33391.80

                          Price to book

                  Event Sample   Control Sample

Mean                      2.32             2.21
SD                        2.18             2.17
Min                       0.26             0.27
25th percentile           1.12             1.03
Median                    1.66             1.57
75th percentile           2.86             2.37
Max                      18.24            19.25

Table 4
Summary of Data Filtering Process

Table 4 summarises the data filtering process applied to each year for
the period 1997-2004 to arrive at the final sample of 122 takeover
announcements. Year represents the year in which the takeovers are
announced. Initial sample shows the number of takeover announcements
made each year that involve Australian companies as either bidders or
targets. Final Sample shows the number of announcements each year
included in the final sample of takeover announcements.

Year                1997   1998   1999    2000   2001

Initial Sample       47     55     60      75     73

Unlisted or           0      0      7       3      3
foreign targets

Inadequate share     20     26     25      21     23
price information

Inadequate           17     to     13      11     17
accounting
information

Presence of prior     2      3      1       9      3
controlling
interest

Joint takeover        0      0      2       2      5
announcements

Announcements         0      1      3       2      0
of subsequent
bids

Multiple              2      3      0       5      3
announcements
by the same
bidder

Final sample          6     12      9      22     19

Year                2002   2003   2004   Total

Initial Sample       50     67     62      489

Unlisted or           3      4      7       27
foreign targets

Inadequate share     22     22     16      175
price information

Inadequate            6     19      6       99
accounting
information

Presence of prior     4      3      4       29
controlling
interest

Joint takeover        0      1      2       12
announcements

Announcements         1      1      1        9
of subsequent
bids

Multiple              1      0      2       16
announcements
by the same
bidder

Final sample         13     17     24      122

Table 5
Descriptive Statistics for Dummy Variables

Table 5 summarises the classification of dummy variables used in the
study. Variable refers to the abbreviated name of the variable of
interest. Variable Value refers to the value for the dummy variables
(equal to 1 if the attribute(s) specified for the variable is (are)
present and 0 otherwise). Number of observations represents the number
of observations corresponding to each variable value category.
Proportion of sample refers to the proportion in the sample size
contributed by each variable category.

Variable   Variable Value    Number of     Proportion of
                            Observations      Sample

THLD             1               69           56.56%
                 0               53           43.44%
CON              1               39           31.97%
                 0               83           68.03%
SUCCESS          I               70           57.38%
                 0               52           42.62%
CASH             1               55           45.08%
                 0               67           54.92%
RELATED          1               65           53.28%
                 0               57           46.72%
COMP             1               24           19.67%
                 0               98           80.33%

Table 6
Timing of Toehold Formation

Table 6 provides a breakdown of the timing of toehold formation for
the takeovers announced by bidders having relevant interests in
targets' outstanding ordinary shares immediately prior to takeover
announcements. Timing of toehold formation refers to the number of
calendar days between the initial announcements of toehold formation
and announcements of takeovers.

  Number of calendar days between
    initial toehold announcement       Number of Observations
     and takeover announcement
                 X

                 0                               33
  0 < X [less than or equal to] 5                4
  5 < X [less than or equal to] 10               2
 11 < X [less than or equal to] 100              7
100 < X [less than or equal to] 500              16
500 < X [less than or equal to] 1000             5
               >1000                             2

Table 7
Descriptive Statistics for Continuous Variables

Table 7 presents the descriptive statistics in relation to non-dummy
explanatory variables used in the study. Variable refers to the
abbreviated name of the variable of interest. LNRSIZE is the natural
logarithm of the ratio of the target's to the bidder's market
capitalisation. FCF refers to the level of the bidder's free cash
flows standardised by the bidder's market capitalisation. TOESIZE is
the proportion of target shares owned by the bidders upon takeover
announcements.

Variable   Sample    Mean     Standard      25th
            Size              Deviation   Percentile

LNRSIZE     122     -1.6805    1.6457      -2.7013
FCF         122      0.0104    0.2664       0
TOESIZE     122      0.095     0.0941       0

Variable     Median       75th
                       Percentile

LNRSIZE      -1.4846     -0.5271
FCF           0.0325      0.0741
TOESIZE       0.0949      0.199

Table 8
The Overall Bidder Shareholder Wealth Effects of Takeover
Announcements

Table 8 presents the results of estimating the overall bidder
shareholder wealth effects of takeover announcements. Average wealth
effects are measured using the 0/1 market model cumulative average
abnormal returns (CAARs), Barber and Lyon's (1997) cumulative average
abnormal returns (BLCAARs) and Barber and Lyon's (1997) average
buy-and-hold abnormal returns (BLABHARs). CAARs are the averages of
individual bidders' cumulative abnormal returns (CARS) over specific
event windows, where CARS are the sums of daily abnormal returns
(ARs), where ARs are calculated as the difference between the actual
daily bidders' returns and the market returns: [AR.sub.i,t] =
[R.sub.i,t] - [R.sub.m,t]. BLCAARs are the averages of individual
bidders' BLCARs over specific event windows, where BLCARs are the
sums of daily bidders' Barber and Lyon's (1997) abnormal returns
(BLARs), where BLARs are calculated as the difference between actual
daily bidder returns and actual daily returns on the control firms:
[BLAR.sub.i,t] = [R.sub.i,t] - [R.sub.c,t]. BLABHARs are the averages
of individual bidders' Barber and Lyon's (1997) buy-and-hold abnormal
returns (BLBHARs) over specific event windows, where BLBHARs are
calculated as

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII].

Continuously compounded returns are employed in the 0/1 market model
calculations, while discrete returns are used to calculate the BLCAARs
and BLABHARs. t-statistics are shown in parentheses.

Model        0/1 Model     Barber and Lyon (1997)
Window         CAAR
                            BLCAAR      BLABHAR

[-1,+1]       0.0016         0.0068      0.0069
             (0.2780)       (1.0081)    (1.0211)
[-1,0]        0.0000         0.0073      0.0075
            (-0.0091)       (1.1617)    (1.2044)
[0,+1]        0.0007         0.0028      0.0027
             (0.1404)       (0.4559)    (0.4428)
[-2,+2]      -0.0011         0.0051      0.0054
            (-0.1546)       (0.6198)    (0.6641)
[-2,0]       -0.0007         0.0100      0.0097
            (-0.1225)       (1.4404)    (1.4374)
[0,+2]       -0.0014        -0.0017     -0.0015
            (-0.2250)      (-0.2203)   (-0.1953)
[-5,+5]      -0.0095        -0.0046     -0.0038
            (-1.0847)      (-0.4461)   (-0.3924)
[-5,0]       -0.0006         0.0041      0.0043
            (-0.0996)       (0.5240)    (0.5526)
[0,+5]       -0.0098        -0.0055     -0.0056
            (-1.3359)      (-0.6214)   (-0.6601)
[-60,+60]    -0.0861        -0.0205     -0.0060
            (-2.8629) **   (-0.6102)   (-0.1712)
[-60,0]      -0.0162        -0.0088     -0.0184
            (-0.8378)      (-0.3173)   (-0.6748)
[0,+60]      -0.0708        -0.0085     -0.0098
            (-3.6766)      (-0.3277)   (-0.3915)

Note: ** Significant at 1% level; and
* Significant at 5% level

Table 9
Differences in Abnormal Returns for Bidders with and without
Toeholds

Table 9 presents the results of testing the difference in abnormal
returns between the toehold and nontoehold bidders. Average wealth
effects are measured using the 0/1 market model cumulative average
abnormal returns (CAARs), Barber and Lyon's (1997) cumulative average
abnormal returns (BLCAARs) and Barber and Lyon's (1997) average
buy-and-hold abnormal returns (BLABHARs). Mean difference is
calculated by subtracting the average abnormal return accruing to the
non-toehold bidders from that obtained by the toehold bidders. The
null hypothesis is that the two groups exhibit no significant
difference on average in the wealth effects of takeover announcements.
t-statistics of significance are reported in brackets under the mean
difference.

Model          0/1 Model          BLCAAR            BLABHAR
Window      mean difference   mean difference   mean difference
               (t-stat)          (t-stat)          (t-stat)

[-1,+1]         0.0180            0.0143            0.0134
               (1.5190)          (1.0501)          (0.9901)
[-1,0]          0.0145            0.0111            0.0096
               (1.3495)          (0.8887)          (0.7726)
[0,+1]          0.0185            0.0193            0.0194
               (1.7684)          (1.5426)          (1.5579)
[-2,+2]         0.0247            0.0253            0.0242
               (1.7519)          (1.5308)          (1.5008)
[-2,0]          0.0108            0.0038            0.0024
               (0.9948)          (0.2817)          (0.1791)
[0,+2]          0.0288            0.0377            0.0381
               (2.3620 *)        (2.4765 *)        (2.4784 *)
[-5,+5]         0.0100            0.0208            0.0128
               (0.5267)          (1.0025)          (0.6578)
[-5,0]          0.0101            0.0089            0.0062
               (0.7746)          (0.5624)          (0.3932)
[0,+5]          0.0148            0.0281            0.0261
               (0.9688)          (1.5448)          (1.4928)
[-60,+60]       0.1001            0.1054            0.1269
               (1.6213)          (1.4851)          (1.8027)
[-60,0]         0.0398            0.0121            0.0002
               (1.0332)          (0.2142)          (0.0043)
[0,+60]         0.0752            0.1095            0.1145
               (1.8821)          (2.0916 *)        (2.2210 *)

Note: ** Significant at 1% level; and
* Significant at 5% level.

Table 10
Statistically Significant Results from Multivariate Regressions

Table 10 summarizes the statistically significant results of fitting
equation (9) to jointly test the impact of explanatory variables on
the magnitude of the bidder shareholder wealth effects of takeover
announcements across 12 event windows. All regressions are undertaken
using Ordinary Least Square method with White's (1980) standard
errors. The notation used in this table is defined as follows:
variable is the abbreviated name for the explanatory variable
employed in multivariate regression analysis; THLD is a dummy variable
equal to 1 if the bidder owns a toehold in the target on the takeover
announcement day and 0 otherwise; CON is a dummy variable equal to 1
if the toehold formation is announced within the window [-10,0] and 0
otherwise; COMP is a dummy variable equal to 1 if there exist
competing bidders) and 0 otherwise; SUCCESS is a dummy variable equal
to 1 if the bid is successful and 0 otherwise; CASH is a dummy
variable equal to 1 if the proposed method of takeover payment is
solely cash and 0 otherwise; RELATED is a dummy variable equal to 1 if
the bidder and the target have the same 2-digit SIC code and 0
otherwise; LNRSIZE is the relative target size variable, calculated as
ln (the target's market capitalization/ the bidder's market
capitalization); FCF is the bidders' free cash flows variable,
calculated as (net profit after tax +depreciation +amortization-
dividends)/market capitalization. A, B, and C refer to the models
where the dependent variables are the 0/1 market model cumulative
abnormal returns, Barber and Lyon's (1997) cumulative abnormal returns
and Barber and Lyon's (1997) buy and hold abnormal returns
respectively; ++ denotes positive significance at 1 % level; + denotes
positive significance at 5% level;--denotes negative significance at
1% level; denotes negative significance at 5% level.

Variable       [-1,+1]   [-1,0]    [0,+1]    [-2,+2]   [-2,0]

THLD                               ++A       +A
                                   +B
                                   +C
CON*THLD

COMP*THLD                          --A
                                   -B
                                   -C
SUCCESS*THLD

SUCCESS

CASH

COMP                     +A                            ++A

RELATED
LNRSIZE
FCF

Variable       [0,+2]    [-5,+5]   [-5,0]    [0,+5]

THLD           ++A
               +B
               +C
CON*THLD                           -B
                                   -C
COMP*THLD      -A

SUCCESS*THLD

SUCCESS

CASH

COMP

RELATED
LNRSIZE
FCF

Variable       [-60,+60]   [-60,0]   [0,+60]

THLD           +C                    ++A
                                     ++B
                                     ++C
CON*THLD

COMP*THLD      -A                    -A
               -C                    -B

SUCCESS*THLD                         --A
                                     --B
                                     --C
SUCCESS        ++A                   ++A
               +B                    ++B
               +C                    +C
CASH                       -B        +A
                           -C        ++B
                                     +C
COMP           ++A         +A        +A
               +C
RELATED
LNRSIZE
FCF                        ++B
                           ++C

Table 11
Results of Multivariate Regressions Over the [0,+1] Window

Table 11 summarizes the results of fitting equation (9) to jointly
test the impact of explanatory variables on the magnitude of the
bidder shareholder wealth effects of takeover announcements over the
[0,+1] window, where day 0 is the takeover announcement day. All
regressions are undertaken using Ordinary Least Square method with
White's (1980) standard errors. The notation used in this table is
defined as follows: 0/1 Model CAR, BLCAR, and BLBHAR refer to the
regressions where the dependent variables are the 0/1 model cumulative
abnormal returns, Barber and Lyon's (1997) cumulative abnormal returns
and Barber and Lyon's (1997) buy and hold abnormal returns
respectively, calculated over the [0,+1] window; THLD is a dummy
variable equal to 1 if the bidder owns a toehold in the target on the
takeover announcement day and 0 otherwise; CON is a dummy variable
equal to 1 if the toehold formation is announced within the window
[-10,0] and 0 otherwise; COMP is a dummy variable equal to 1 if there
exist competing bidder(s) and 0 otherwise; SUCCESS is a dummy variable
equal to 1 if the bid is successful and 0 otherwise; CASH is a dummy
variable equal to 1 if the proposed method of takeover payment is cash
and 0 otherwise; RELATED is a dummy variable equal to 1 if the bidder
and the target have the same 2-digit SIC code and 0 otherwise; LNRSIZE
is the relative target size variable, calculated as ln (the target's
market capitalization/ the bidder's market capitalization); FCF is the
bidders' free cash flows variable, calculated as (net profit after tax
+depreciation +amortization-dividends)/market capitalization.
t-statistics are presented in parentheses under the coefficients.

Variable             0/1 Model CAR   BLCAR          BLBHAR

Constant              -0.0263         -0.0247       -0.0251
                     (-1.7276)       (-1.3189)     (-1.3504)
THLD                   0.0436          0.0500        0.0497
                     (-2.695) **      (2.2629) *    (2.2617) *
CON*THLD              -0.0018         -0.0079       -0.0078
                     (-0.1197)       (-0.4533)     (-0.4536)
COMP*THLD             -0.0476         -0.0541       -0.0535
                     (-2.6883) **    (-2.2116) *   (-2.1708) *
SUCCESS*THLD          -0.0259         -0.0305       -0.0304
                     (-1.2625)       (-1.1725)     (-1.1705)
SUCCESS                0.0112          0.0174        0.0178
                      (0.6435)        (0.8793)      (0.9039)
CASH                   0.0072         -0.0012       -0.0013
                      (0.6280)       (-0.0798)     (-0.0872)
COMP                   0.0221          0.0302        0.0303
                      (1.5838)        (1.5484)      (1.5358)
RELATED                0.0117          0.0039        0.0036
                      (1.0125)        (0.3026)      (0.2736)
LNRSIZE                0.0009         -0.0015       -0.0017
                      (0.2221)       (-0.2830)     (-0.3158)
FCF                   -0.0120          0.0110        0.0108
                     (-0.6672)        (0.3563)      (0.3472)
Adjusted [R.sup.2]    -0.0171         -0.0335       -0.0332
[R.sup.2]              0.0670          0.0519        0.0522
F-statistic            0.7971          0.6075        0.6114

Note: ** Significant at 1% level; and
* Significant at 5% level.

Table 12
Results of Multivariate Regressions Over the [0,+60] Window

Table 12 summarizes the results of fitting equation (9) to jointly
test the impact of explanatory variables on the magnitude of the
bidder shareholder wealth effects of takeover announcements over the
[0,+60] window, where day 0 is the takeover announcement day. All
regressions are undertaken using Ordinary Least Square method with
White's (1980) standard errors. The notation used in this table is
defined as follows: 0/1 Model CAR, BLCAR, and BLBHAR refer to the
regressions where the dependent variables are the 0/1 model cumulative
abnormal returns, Barber and Lyon's (1997) cumulative abnormal returns
and Barber and Lyon's (1997) buy and hold abnormal returns
respectively, calculated over the [0,+60] window; THLD is a dummy
variable equal to 1 if the bidder owns a toehold in the target on the
takeover announcement day and 0 otherwise; CON is a dummy variable
equal to 1 if the toehold formation is announced within the window [-
10,0] and 0 otherwise; COMP is a dummy variable equal to 1 if there
exist competing bidder(s) and 0 otherwise; SUCCESS is a dummy variable
equal to 1 if the bid is successful and 0 otherwise; CASH is a dummy
variable equal to 1 if the proposed method of takeover payment is cash
and 0 otherwise; RELATED is a dummy variable equal to 1 if the bidder
and the target have the same 2-digit SIC code and 0 otherwise; LNRSIZE
is the relative target size variable, calculated as ln (target's
market capitalization/ bidder's market capitalization); FCF is the
bidders' free cash flows variable, calculated as (net profit after tax
+depreciation+amortization-dividends)/market capitalization.
t-statistics are presented in parentheses under the coefficients.

Variable             0/1 Model CAR       BLCAR          BLBHAR

Constant              -0.2936         -0.2668         -0.2580
                     (-4.5414) **    (-3.7705) **    (-3.8172) **
THLD                   0.2050          0.3079          0.2653
                      (2.6335) **     (3.3658) **     (3.3888) **
CON*THLD               0.0293         -0.0147         -0.0044
                      (0.6678)       (-0.2538)       (-0.0844)
COMP*THLD             -0.1908         -0.2488         -0.1979
                     (-2.5894) *     (-2.1316) *     (-1.5541)
SUCCESS*THLD          -0.2655         -0.3044         -0.2497
                     (-3.4515) **    (-3.1397) **    (-2.6306) **
SUCCESS                0.2357          0.2004          0.1772
                      (3.6728) **     (2.7485) **     (2.3776) *
CASH                   0.1124          0.1688          0.1319
                      (2.4997) *      (2.9896) **     (2.4630) *
COMP                   0.1569          0.0469          0.0339
                      (2.5198) *      (0.4620)        (0.3076)
RELATED                0.0389          0.0370          0.0409
                      (1.0293)        (0.7780)        (0.8705)
LNRSIZE                0.0103         -0.0097         -0.0177
                      (0.7906)       (-0.5346)       (-1.0147)
FCF                    0.0184         -0.2578         -0.1618
                      (0.1850)       (-1.8502)       (-1.0060)
Adjusted [R.sup.2]     0.1611          0.1748          0.1193
[R.sup.2]              0.2305          0.2430          0.1921
F-statistic            3.3244 **       3.5634 **       2.6392 **

Note: ** Significant at 1% level; and
* Significant at 5% level.

Table 13
Sensitivity Test Results Over the [0,+1] Window

Table 13 summarizes the sensitivity test results of
fitting equation (9) to jointly test the impact of
explanatory variables on the magnitude of the bidder
shareholder wealth effects of takeover announcements over
the [0,+1] window, where day 0 is the takeover announcement
day. All regressions are undertaken using Ordinary Least
Square method with White's (1980) standard errors. The
notation used in this table is defined as follows: 0/1
Model CAR, BLCAR, and BLBHAR refer to the regressions where
the dependent variables are the 0/1 model cumulative
abnormal returns, Barber and Lyon's (1997) cumulative
abnormal returns and Barber and Lyon's (1997) buy and hold
abnormal returns respectively, calculated over the [0,+1]
window; TOESIZE is the percentage of target shares owned by
the bidder as at the takeover announcement date; CON is a
dummy variable equal to 1 if the toehold formation is
announced within the window [-10,0] and 0 otherwise; COMP
is a dummy variable equal to 1 if there exist competing
bidder(s) and 0 otherwise; SUCCESS is a dummy variable
equal to 1 if the bid is successful and 0 otherwise; CASH
is a dummy variable equal to 1 if the proposed method of
takeover payment is cash and 0 otherwise; RELATED is a
dummy variable equal to 1 if the bidder and the target have
the same 2-digit SIC code and 0 otherwise; LNRSIZE is the
relative target size variable, calculated as 1n (the
target's market capitalization/ the bidder's market
capitalization); FCF is the bidders' free cash flows
variable, calculated as (net profit after tax +depreciation
+amortization-dividends)/market capitalization. t-statistics
are presented in parentheses under the coefficients.

Variable             0/1 Model CAR       BLCAR          BLBHAR

Constant                -0.0222         -0.0241         -0.0246
                       (-1.5205)       (-1.3333)       (-1.3700)
TOESIZE                  0.2383          0.3022          0.3024
                        (2.7985) **     (2.7375) **     (2.7468) **
CON * TOESIZE           -0.0113         -0.0261         -0.0267
                       (-0.1308)       (-0.2673)       (-0.2739)
COMP * TOESIZE          -0.3220         -0.4070         -0.4064
                       (-3.3939) **    (-3.4101) **    (-3.3770) **
SUCCESS *               -0.0387         -0.0847         -0.0843
TOESIZE                (-0.3292)        (0.6025)       (-0.6015)
SUCCESS                 -0.0013          0.0061          0.0066
                       (-0.0783)        (0.3250)        (0.3507)
CASH                     0.0074         -0.0016         -0.0018
                        (0.6179)       (-0.1134)       (-0.1214)
COMP                     0.0255          0.0378          0.0382
                        (1.9676)        (2.0729) *      (2.0743) *
RELATED                  0.0107          0.0038          0.0035
                        (0.9513)        (0.3140)        (0.2835)
LNRSIZE                  0.0018         -0.0007         -0.0009
                        (0.4337)       (-0.1359)       (-0.1662)
FCF                     -0.0099          0.0133          0.0131
                       (-0.05828)       (0.4619)        (0.4518)
Adjusted [R.sup.2]       0.0283          0.0188          0.0199
[R.sup.2]                0.1086          0.0999          0.1009
F-statistic              1.3521          1.2316          1.2459

Note:
** Significant at 1% level; and
* Significant at 5% level.

Table 14
Sensitivity Test Results Over the [0,+60] Window

Table 13 summarizes the sensitivity test results of
fitting equation (9) to jointly test the impact of
explanatory variables on the magnitude of the bidder
shareholder wealth effects of takeover announcements over
the [0,+60] window, where day 0 is the takeover announcement
day. All regressions are undertaken using Ordinary Least
Square method with White's (1980) standard errors. The
notation used in this table is defined as follows: 0/1
Model CAR, BLCAR, and BLBHAR refer to the regressions where
the dependent variables are the 0/1 model cumulative
abnormal returns, Barber and Lyon's (1997) cumulative
abnormal returns and Barber and Lyon's (1997) buy and hold
abnormal returns respectively, calculated over the [0,+1]
window; TOESIZE is the percentage of target shares owned by
the bidder as at the takeover announcement date; CON is a
dummy variable equal to 1 if the toehold formation is
announced within the window [-10,0] and 0 otherwise; COMP
is a dummy variable equal to 1 if there exist competing
bidder(s) and 0 otherwise; SUCCESS is a dummy variable
equal to 1 if the bid is successful and 0 otherwise; CASH
is a dummy variable equal to 1 if the proposed method of
takeover payment is cash and 0 otherwise; RELATED is a
dummy variable equal to 1 if the bidder and the target have
the same 2-digit SIC code and 0 otherwise; LNRSIZE is the
relative target size variable, calculated as 1n (the
target's market capitalization/ the bidder's market
capitalization); FCF is the bidders' free cash flows
variable, calculated as (net profit after tax +depreciation
+amortization-dividends)/market capitalization. t-statistics
are presented in parentheses under the coefficients.

Variable             0/1 Model CAR       BLCAR          BLBHAR

Constant               -0.2820         -0.2450         -0.2387
                      (-4.5963) *     (-3.5503) **    (-3.6319) **
TOESIZE                 1.1248          1.5816          1.3463
                       (2.5926) *      (3.1183) **     (3.2636) **
CON * TOESIZE           0.2587         -0.0421          0.0037
                       (1.0435)       (-0.1345)        (0.0131)
COMP * TOESIZE         -0.9599         -1.1065         -0.8847
                      (-2.4619) *     (-1.8291)       (-1.4060)
SUCCESS * TOESIZE      -1.2813         -1.6505         -1.4170
                      (-2.9653) **    (-3.2003) **    (-2.9867) **
SUCCESS                 0.2019          0.1846          0.1717
                       (3.3337) **     (2.7103) **     (2.5136) *
CASH                    0.1162          0.1810          0.1447
                       (2.5417) *      (3.1646) **     (2.6388) **
COMP                    0.1339          0.0069          0.0036
                       (2.2926) *      (0.0752)        (0.0361)
RELATED                 0.0359          0.0321          0.0369
                       (0.9688)        (0.6605)        (0.7647)
LNRSIZE                 0.0119         -0.0085         -0.0166
                       (0.8916)       (-0.4691)       (-0.9558)
FCF                     0.0207         -0.2526         -0.1570
                       (0.2236)       (-1.9113)       (-1.0170)
Adjusted [R.sup.2]      0.1563          0.1634          0.1090
[R.sup.2]               0.2260          0.2325          0.1826
F-statistic             3.2418 **       3.3633 **       2.4804 *

Note:
** Significant at 1% level; and
* Significant at 5% level