Does insider trading explain price run-up ahead of takeover announcements? ANGELO ASPRIS*, SEAN FOLEY and ALEX FRINO Finance Discipline, Faculty of Economics and Business, University of Sydney, Sydney, 2006, Australia Abstract Motivated by previous empirical findings which infer that a significant portion of the price run-up observed prior to takeovers is attributable to insider trading, we examine 450 takeovers of Australian firms between 2000 and 2009 to study the contribution of toehold timing to price run-up. We provide evidence that the acquisition of toehold stakes within thirty days of the formal takeover announcement generates significant pre-bid run-up in the target firm. Notably, an examination of price run-up for target firms where a toehold is established immediately prior to the takeover bid has a significant positive impact on the pre-bid run-up. The effects of this short-term toehold are able to explain a significant amount of the pre-bid run-up. This suggests that previous research may overstate the degree of insider trading implied by pre-bid price run-up. The authors would like to thank the participants of the second annual Finance and Corporate Governance Conference for their feedback, articularly the input of Bart Frijns, Robert Faff and Tom Smith. *Contact Author, Finance Discipline, School of Business, University of Sydney, 2006, Australia: Tel (+612) 9114 0553; Fax (+612) 9351 6461; Email: angelo.aspris@sydney.edu.au.
1. Introduction Previous event studies estimating the wealth effects of takeovers over the last 30 years have documented considerable evidence of abnormal price run-up in markets such as the US, UK, Australia and Canada (Dodd, 1980; Jarrell and Poulsen, 1989; Betton and Eckbo, 2000 and Bris, 2005). Two separate and competing explanations have emerged for this phenomenon. The first is that insiders trade in advance of a takeover bid, with this trading activity disseminating their knowledge to the broader market. 1 Using data on illegal insider trading, Meulbroek (1992) shows that almost half the pre-announcement stock price run-up in takeovers occurs on days in which insider trading occurs. The second competing theory, originally proposed by Jensen and Ruback (1983), suggests that this run-up is generated by market participants predicting takeover targets through legally obtained information channels. 2 The explanations for the price run-up therefore are still contentious and have implications for both market integrity and quality. This paper uses a unique dataset to examine the trading behavior of acquirers immediately prior to the announcement of a takeover offer. Change in substantial shareholding notices that detail bidder s purchases must be provided within two business days of the purchase. 3 These notices specify both the number of shares acquired and the date of acquisition. This allows us to examine one of the mechanisms by which private information becomes public, alluded to in Jensen and Ruback s (1983) market 1 Suggestions of insider trading in takeovers have typically relied on examining the extent to which share prices move ahead of regulatory announcements which issuers are required to make to the market. See Keown and Pinkerton (1981); Meulbroek (1992); Dubow and Monteiro (2006) and King (2009). 2 See also Jarrell and Poulsen (1989); Pound and Zeckhauser (1990); Sanders and Zdanowicz (1992). 3 Section 671B of the Corporations Act (2000) Cth. Requires any holder of greater than 5% of a public company to report this shareholding or changes to it within 2 business days of the change. 2
anticipation hypothesis. Abnormal returns prior to takeover bids have been shown in previous studies to exist even after factors such as rumours and the size of the toehold are accounted for. This study extends the existing literature by demonstrating that the observed pre-bid run-up is not robust to the additional inclusion of the timing of the toehold purchase as a potential signal to the market of the takeover bid. Previous theoretical (Bris, 2002) and empirical (Jarrell and Poulsen, 1989; Aitken and Czernkowski, 1992, Betton, Eckbo and Thorburn, 2008b) studies acknowledge an impact on price arising from the acquisition of a toehold. Empirical evidence regarding the impact of toeholds on run-up is both scant and inconclusive, with Betton and Eckbo (2000) finding a significant and negative relationship between toeholds and the pre-bid run-up, whilst Jarrell and Poulson (1989) show a significant positive relationship. This lack of consensus is likely driven by the absence of information regarding the timing of the accumulation of target shares by the bidder, a problem acknowledged and attributed to the lack of data availability by both studies. We provide a potential reconciliation of this lack of consensus by collecting data that allows us to differentiate between toeholds that are taken immediately prior to takeover bids (henceforth short-term toeholds ) and those that have been held for a significant period of time (long-term toeholds). In addition to bidder toehold information, the examination of takeovers in the Australian market, which requires firms to conform to continuous disclosure regulations, allows us to isolate any extraneous information (and potentially price-sensitive) events that could 3
distort wealth patterns associated with takeover bids. 4 Not only does continuous disclosure require the immediate release to the market of price sensitive announcements, the Australian Securitites Exchange (ASX) also flags each of these announcements as either price or non-price sensitive. The ability to control for price sensitive announcements provides further clarity on the causes of pre-bid price reactions by allowing potentially confounding events to be removed from the data set. This disclosure regime differs significantly from the Fair Disclosure regulations of the US, which mandate only periodic reporting, with additional news releases submitted voluntarily. Similar continuous disclosure requirements in other jurisdictions, notably the European exchanges, allows for the extension of this research to other developed markets. This paper identifies to what extent legal factors can explain price movements prior to takeover bids. Our ability to explain the pre-bid price run-up is of interest both academically and professionally. If the timing of the toehold decision has an impact on the run-up this could potentially explain the lack of consensus in the literature regarding the impact of toeholds. It could also cast doubt on the assertions of the works finding that unexplained pre-bid price run-up is solely attributable to insiders. This work also has implications for regulators interested in identifying the extent to which price sensitive announcements are anticipated. It is also relevant to market participants as it contributes to the literature on the optimal timing of the toehold decision and has potential implications for bidders considering acquiring a pre-announcement stake in the target company. 4 These regulations mandate that any price-sensitive information be released to the market as soon as companies become aware of it. 4
The remainder of this paper is structured as follows; Section 2 reviews the existing literature on toeholds and discusses previous explanations. Section 3 describes the data used. Section 4 outlines the research methodology. Section 5 discusses the results and Section 6 provides concluding remarks. 2. The causes of pre-bid run-up The stock price run-up observed in target firms prior to takeover announcements is a phenomenon that has endured through time and across market structures. 5 A survey by Bris (2005) documents an average price run-up of between 4% and 17% in over 4,500 deals across 52 countries. Despite the extensive evidence documenting this abnormal price behaviour, explanations for its occurrence remain elusive, with no study able to identify explanations for all of the abnormal run-up. The possibilities explored in the current literature consider insider trading (Keown and Pinkerton (1981), Meulbroek (1992)), rumours in the news media (Pound and Zeckhauser (1990), Sanders and Zdanowicz (1992)) and market anticipation (Jensen and Ruback (1983), Jarrell and Poulsen (1989)). Insider trading was originally proffered as an explanation for abnormal pre-bid price movements by Keown and Pinkerton (1981). The authors observe average pre-bid runups of 13% for over 200 US takeovers, which they argue is evidence of widespread 5 See Keown and Pinkerton (1981), Franks and Harris (1989), Betton and Eckbo (2000) and King (2009). 5
illegal trading on inside information. 6 This conclusion is supported by studies of prosecuted insider trading episodes undertaken by Meulbroek (1992) and Cornell and Sirri (1992). While Keown et al. (1981) rely on the extent to which share prices move ahead of public takeover announcements to form this opinion, Meulbroek (1992) uses data from illegal insiders trades to document that almost half of the observed pre-bid run-up was experienced on days in which insiders traded. Chakravarty and McConnell (1997), find support for this theory, identifying a positive and significant relationship between insiders trades and pre-bid price run-up. Jensen and Ruback (1983) propose an alternative explanation for the abnormal price movements observed prior to takeover events, arguing that the pre-bid run-up is a function of market anticipation derived from public information. The authors reason that public rumours of impending takeovers are traded upon by speculators, ultimately being reflected in share price movements prior to the official announcement. Anticipation can however be more than just deal specific, with takeovers shown to be clustered over time, further facilitating takeover target prediction. 7 Although Jensen and Ruback (1983) do not specifically address the mechanisms by which this information is translated into prices, Jarrell and Poulsen (1989), formally 6 Regulatory exchanges across the world have adopted this form of analysis to examine the degree of market cleanliness. For example, in the FSA Measurement of Market Cleanliness publication the pre-bid run-up is examined to show what degree insider trading exists in the marketplace (Dubow and Monteiro, 2006). 7 Merger waves or bubbles of financial activity as referred to by Brealey and Myers (1996) are depicted in a broad sample of studies across different regions. This can be a source of profit for arbitrageurs and other sophisticated investors if there are perceived trends (Mitchell and Pulvino, 2001). 6
examine the contribution of different forms of publicly available information on the share price run-up. In a study of 172 US takeovers, the authors show that there is an insignificant difference between the run-up for bids with and without prosecuted insider trading, contrary to the arguments put forward by Keown et al. (1981). They find that the primary predictors of pre-bid run-up are news-media rumours and the existence of toeholds greater than 50%. The importance of the news-media speculation as a mechanism by which private information becomes public is similarly shown in studies such as Sanders and Zdanowicz (1992), Pound and Zeckhauser (1990) and Aitken and Czernkowski (1992). These studies find that up to 40% of the pre-bid run-up can be explained by legal market anticipation following published rumours of potential consolidation between the target and acquirer. Further examining the role of market anticipation in explaining share price run-ups, Bris (2002) theorises that the acquisition of a toehold stake produces a potential information flow that can result in share price movements ahead of takeover announcements. Empirical evidence consistent with this theory is provided by Choi (1991) and Barclay and Holderness (1991), who find significant positive abnormal returns around the purchase of toehold stakes. Empirical evidence regarding this theory is however divided, with Betton and Eckbo (2000) showing that a negative relationship exists between the size of the toehold and the pre-bid run-up. Conversely, Jarrell and Poulsen (1989) find that the greater is the toehold as a percent of the targets shares, the higher is the pre-bid run-up. This is consistent with the theoretical model of Ravid and Spiegel (1999) who 7
argue that toehold purchases create rumors of an imminent takeover bid, resulting in runup in the targets share price. There are two potential reasons for the differing relationships found by Jarrell and Poulson (1989) and Betton and Eckbo (2000). One is due to the direction of causality, with Jarrell and Poulson (1989) using run-up as the dependent variable, whilst Betton and Eckbo (2000) use toehold size as the dependent variable. The other potential source of difference is that neither of these studies account for the timing of the toehold. Differences in the mix of both short and long-term toeholds could result in differences in the impact of toeholds on run-up. Indeed, the impact of short-term toeholds is briefly documented by Betton, Eckbo and Thorburn (2008b), finding that short-term toeholds facilitate information flows, leading to higher pre-bid price run-up. If toeholds signal to the market the potential for a bid, as Bris (2002) theorises, then the toehold acquisition timing will likely be a key determinant of the size of the run-up. There are several mechanisms by which a short-term toehold could induce a pre-bid runup. First, the upward sloping supply curve for target shares will result in an increase in target share price caused by purchasing pressure from the bidder (Stulz, 1988). This problem will be especially pronounced for less liquid targets. Second, any large unexplained trades could trigger speculation that the firm is a takeover target. Evidence from Jarrell and Poulson (1989) and Aitken and Czernkowski (1992) indicate that rumours have a significant impact on pre-bid run-up. Further, the information contained 8
in substantial shareholder notices 8 can also signal the intentions of the bidder to the market, triggering buying in anticipation of a takeover premium (Choi, 1991). Toeholds bought immediately prior to the bid should therefore have a larger signaling impact on the pre-bid run-up as compared to those toeholds which have been held for a longer period of time. These long-term toeholds will have already incorporated information about the probability of a future takeover, (Choi, 1991) and as such it is expected that the price run-up immediately prior to the announcement will be insignificantly different from the case where no toehold is purchased. The current literature has not been able to offer a clear explanation for the welldocumented abnormal price reactions to takeover announcements, with two potential explanations having been formulated, that of insider trading and market anticipation. This study confirms the role of market anticipation generated by media speculation as an important determinant of pre-bid price run-up and further examines the impact of toehold acquisitions and price-sensitive news announcements within a regulatory framework of continuous disclosure. This framework provides a unique opportunity to explore the effectiveness of the regime at minimising the level of insider trading around takeover announcements. If market anticipation is the cause of the documented pre-bid run-up in Australia, these elements should be able to explain the pre-bid run-up. However, if illegal trading is causing the abnormal price reactions, accounting for these factors will not fully explain the observed pre-bid price run-up. 8 Substantial Shareholder notices are required to be provided by the acquirer of greater than 5% of a public companies shareholding within two days under the Corporations Act (Cth) 2001. 9
3. Data The data set of takeovers is obtained from the Reuters Connect4 database, which provides a record of all deals involving ASX listed public companies from 2000 to 2009. Data collected from Connect 4 includes deal values, deal types and break fees. It also includes other relevant acquisition and accounting information. There are 852 takeover bids in this database for 668 unique target companies. Stock trading information and firm level data are collected via the Reuters intra-day interface provided by the Securities Institute Research Centre of Asia Pacific (SIRCA). This data is captured in real time from the Australian Securities Exchange Integrated Trading System (ITS). Toehold information has been collected manually from bidder statements and either form 603 notifications Becoming a substantial shareholder or form 671B Change in substantial shareholding notices, which are analogous to the US s.13(d) filings. As soon as an investor has beneficial ownership of greater than 5% of a publically listed company, they are required to complete one of these notices within two business days of the acquisition. As investors are not required to report holdings smaller than 5%, any toehold smaller than this 5% level will not be captured in our dataset, and is thus treated as if no toehold existed. 9 Of the target firms subject to takeover offers, 38% contain pre-announcement purchases by the acquirer. News releases used to adjust announcement dates are manually collected from the Factiva database and cross referenced with ASX announcements. In order for an announcement date to be adjusted, 9 There were twelve such observations in our dataset, with toeholds ranging between 0.43% and 4.9%. 10
both the acquirer and target must be named in the same release as a potential business combination. Of the initial 852 bids, any deal that represents a secondary bid for the same target firm, that is a competing or revised bid, is excluded from our sample in order to isolate the only the run-up effects of the first bid. We exclude 184 deals in this manner. A further 30 bids are removed as the bidder has pre-existing majority control via a toehold that is greater than 50%. Bids with more than 50 missing observations in the period (-250,0) are removed due to difficulties in estimating returns, which removes an additional 56 bids. Any bid with an announcement flagged by the ASX as price sensitive in the 30 days prior to the adjusted announcement date is removed so as to narrow the possibility of confounding effects in the price run-up. This resulted in the removal of an additional 132 observations. The final sample adjusted for missing data, multiple bids, existing control and sensitive announcements contains 450 takeover offers. <INSERT TABLE 1 HERE> Table 1 provides descriptive statistics for the final sample of 450 target firms. Information regarding the share price, deal value, deal premium, market to book ratio, earnings (EBIT), and the news adjustment process are all provided. The mean (median) value of the sample of deals is $AU1.28 billion ($AU113 million) indicating that a sizable fraction of the deals announced are for small-to-medium tier capitalisation firms. This inference is further supported with average (median) earnings of approximately 11
$AU62.38 million ($AU4.24 million) for takeover targets. The average premium of approximately 18% over the defined sample period is broadly consistent with 18% found by King (2009) in Canada and the 28% found by Jarrell and Poulson (1989) in the US. Of our sample of 450 announcements, only 76 (or 16.8%) required adjusting. The mean number of calendar days between the news-adjusted and formal announcement dates is 68.69 days across those announcements that were adjusted. This is a similar adjustment period to that reported by King (2009) who reconciles the announcement date by a period of 58 days for deals conducted on the Toronto Stock Exchange. Similarly, Aitken and Czernkowski (1992) show an average adjustment of 59.21 days on the ASX. Short-term toeholds are purchased on average up to 10 days prior to the bid, although the median of 1 day is evidence of frequent buying activity on the day before the formal announcement. In purchasing one day prior to the bid, acquirers are able to delay reporting their stake until after the bid has been made public. This deprives potential rivals of sufficient time to prepare competing bids, allowing firms to minimise the transmission of their intentions to the market. Long-term toeholds are held on average for just over a year. The mean of 404 days implies that these toeholds are often acquired with the intention of making a bid in the near future, either as an investment, to deter rival merger activity or as an option over anticipated future acquisition (Betton, Eckbo and Thorburn, 2009). 12
<INSERT TABLE 2 HERE> Table 2 provides information about the acquisition patterns of toeholds in Australian deals between 2000 and 2009. The deals are divided into two categories: short-term toeholds, which occur when the bidder acquires an incremental toehold stake within thirty days of the official announcement date; and long-term toeholds, which represent all toeholds taken outside 30 days of the official announcement date. If an acquirer holds a long-term toehold and subsequently increases their stake within the short-term period, triggering a change in substantial shareholding notice, this is considered to be a shortterm toehold, due to the additional information provided by the increased purchase of shares. 10 Table 2 indicates that the incidence of long-term toehold acquisitions is more than double that of short-term toeholds in the sample period. This potentially reflects the common use of standstill agreements at the start of negotiations, which prevent bidders from acquiring a stake during the period prior to the formal announcement (Bruner, 2004). There are many costs to toeholds documented in the literature that could explain the preference for long-term toeholds over short-term toehold acquisitions. These include market illiquidity that will increase the eventual purchase price (Ravid and Spiegel, 1999), the possibility for information disclosure which signals the bid to the market, creating an opportunity for competitors to prepare counterbids (Bris, 2002) and the potential for costly target board resistance (Betton, Eckbo and Thorburn, 2009). 10 There are 24 such takeovers in our data set that have been categorised as short-term. 13
<INSERT TABLE 3 HERE> Table 3 shows that over the ten year sample 38% of deals include some form of pre-bid toehold stake. This figure is marginally higher than comparative empirical studies, however the relative decline observed through time is consistent with the findings of Betton, Eckbo and Thorburn (2009) in the United States. 11 Despite the advantages of toeholds hypothesised by the literature, descriptive statistics show that the majority of Australian acquirers do not take a toehold. The advantages of acquiring a toehold include; cost advantages (Bishop, (1991)), avoiding the free-rider problem (Shleifer and Vishny (1986)), reducing price impact (Walkling (1985)) and the potential to deter rivals (Bulow, Huang and Klemperer (1999)). Whilst there are numerous potential benefits, all appear to be overshadowed by the risks inherent in toehold acquisition. To date, only one other study has examined the timing of toehold purchases. Betton, Eckbo and Thorburn (2008b) report that of the firms acquiring toeholds, 90% choose to acquire their stake over a period greater than six months prior to the launching of a bid. Table 3 reports data consistent with this finding, with the majority of toeholds acquired being long-term in nature. Their proportion of acquirers taking toeholds in their U.S. sample is significantly lower than reported in this study, where long-term acquisitions are found to account for over 69% of all toehold activity. The reluctance of firms to engage in short-term toehold activity indicates that the perceived costs are much greater than the 11 Jarrell and Poulson (1989) report that 58% of US bidders take toeholds, Betton, Eckbo, and Thorburn (2009) report 13% in a later sample, and King (2009) finds 19% in Canada. 14
perceived benefits of such activity. This could be due to the potential for short-term toeholds to act as a signaling agent to the market, increasing the takeover premium paid (Bris, 2002). <INSERT TABLE 4 HERE> Table 4 provides a breakdown of the size of toehold positions held by bidders in the sample. The majority of the toeholds held by both short and long-term acquirers are clustered around 15-20% of outstanding share capital, which is reflective of formal regulations limiting the size of a toehold to 20%. 12 In the US market, where such "fair price" provisions do not exist, Betton, Eckbo and Thorburn (2009) report a similar clustering of both short-term and long-term toeholds around the 10% level. The smaller size of short-term toeholds can also be partially attributed to the upward sloping-supply curve encountered by acquirers increasing costs with toehold size (Walkling, 1985). The size of long-term toeholds is also similar between samples, with Betton, Eckbo and Thorburn (2009) finding that 44% of long term toeholds are greater than 20% of the targets equity, and our sample showing that 46% of long term toeholds exceed 20%. The tendency for long-term toeholds to be greater than 20% is likely due to the need to acquire a significant enough position to deter rivals (Bulow, Huang and Klemperer (1999)). 12 Section 615 of the Corporations Act 2001 (Cth) prohibits an individual or company from exceeding 20% share ownership without launching a full takeover bid for all remaining shares. Section 618 provides an exception for "creeping" takeovers, whereby an acquirer may increase their shareholding above the 20% level in 3% increments every six months without launching a full takeover bid. 15
4. Research Design Standard event study methodology is used to examine whether the timing of the toehold placement has an impact on the pre-bid price run-up, with the date of the public takeover announcement being day 0. In order to limit the upward bias in the run-up driven by trading on rumours of the deal in the financial press, the event date 0 is adjusted to reflect the earlier of the public announcement of the takeover, or rumour mentioning both the takeover target and the acquirer as a potential combination. The announcement of a toehold will not of itself necessitate an adjustment, however it is possible that a toehold reported to the exchange results, on the following day, in a rumour of the potential takeover. In this case, the toehold will be considered short-term. If, however, the toehold is taken after the first rumour, this would be considered as a no-toehold event. This adjustment process is consistent with the previous work of Aitken and Czernkowski (1992) and Sanders and Zdanowicz (1992). Abnormal returns are calculated for a window around the adjusted announcement date for all the firms in the sample. Market model regressions are performed in the following way: t = -250,.,-100 (1) where refers to the return on day t for target firm i, and is the market return on day t. The residual defines the excess return for firm i on day t. 16
To calculate the pre-bid run-up we aggregate the daily excess returns from 30 days prior to the announcement to the day before the announcement, to compute the cumulative abnormal return per stock (CAR): (2) The reaction of traded volume is also considered due to its potential to signal the existence of informed traders within the market. A benchmark ( normal volume ) measure is calculated as the average daily volume for each target over the 150 days from -250 to -100. This Normal Volume is then used to standardize the abnormal turnover metric as follows: This measure is calculated for every target firm and cumulated for each event day from - 30 until day 0 to construct the cumulative abnormal turnover (CAT). Abnormal turnover has been used in studies such as King (2009) in order to help ascertain whether the observed pre-bid price movement is generated by insiders or by market anticipation. Figure 1 reports the CARs for the official announcement date, the news adjusted date and the announcement adjusted data set. Consistent with Aitken et al. (1992), mediapublished rumours account for over 20% of the observed pre-bid run-up. The results in Figure 1 show that prior to any adjustment, the run-up is significantly different from zero up to twenty-three days prior to the official release of the takeover, at almost 6% on day - 1. Once rumours in the market are controlled for, the impact on CARs is dramatically 17
reduced, becoming significantly different from zero at the 1% level seven trading days prior to the first public rumour or announcement, and amounting to just under 4.5% on day -1. A further source of potential bias in the run-up is from price sensitive announcements unrelated to the takeover. The additional removal of these from our sample accounts for a further 10% of the observed pre-bid run-up, with rumour and announcement adjusted returns of 4.1% on day -1. This reduction in the pre-bid run-up is potentially due to acquirers, anticipating the release of additional information, delaying their acquisition decision until after the release of scheduled announcements such as halfyearly results. This delay provides the acquirer with a kind of option, allowing bids to be delayed or abandoned if negative results are revealed. If this kind of bias occurs frequently, it would tend to inflate the pre-bid price run-up. <INSERT FIGURE 1 HERE> Our analysis uses the news and announcement adjusted event dates to assess the additional impact of toehold timing on the pre-bid run-up. The impact of short and longterm toeholds on the pre-bid run up is determined by performing a regression analysis with the following specification: (3) 18
Where; Run-up is the CAR of stock i between day -30 and -1; Turnover is the CAT in stock i as a % of clean period (-250,-100) turnover; Third Party is a dummy variable that takes 1 if a third party other than the bidder increases their substantial shareholding in stock i in the 30 days prior to the announcement, which could act as a spurious signal of an impending takeover, and 0 otherwise; ln(value) is a control variable for the size of the stock, being the natural log of stock i s market cap twenty days prior to day 0 ; LT is a dummy variable that takes 1 if a long-term toehold was acquired in stock i and 0 otherwise; ST is a dummy variable that takes 1 if a short-term toehold was acquired and 0 otherwise, with no-toehold forming the base case; and ST*Toehold and LT*Toehold are interaction terms between the timing and size of the toehold stakes, to account for potential size effects. In order to capture the potential for changes occurring by year, a fixed effects model is used with dummy variables for each of the calendar years, with 2000 being the base year. The results of this control for potential variation are reported in Table 5. Whilst a large body of empirical literature exists that documents the post-takeover announcement effects of bid characteristics such as whether the offer is in cash or stock, if the bidder is a public or private company and whether the bid was subsequently consummated, these characteristics will only become public knowledge after the announcement of the bid. As we are considering the run-up prior to the takeover announcement we have only considered factors contemporaneous with the period prior to the takeover bid. 13 13 Separate regressions controlling for these factors indicate that our results are robust to their inclusion. (available upon request) 19
5. Results Figure 2 presents the results of the target run-up analysis adjusted for toehold timing. These results provide preliminary evidence that a significant proportion of the takeover run-up can be explained by considering the point in time at which the toehold is acquired. Figure 2 reports that takeovers preceded by long-term toeholds, without confounding information announcements, have on average a 2.25% run-up in the thirty days leading up to the first rumour or announcement of a bid. This is statistically insignificant from zero at the 5% level. Choi (1991) argues that the acquisition of a toehold, which results in a positive valuation effect, reflects investors perception of an increased probability of a subsequent takeover. Since the takeover premium is largely impounded in the stock price at the time of the acquisition of the toehold, the insignificant run-up in this period is consistent with no further information leakage prior to the announcement of the bid. <INSERT FIGURE 2 HERE> Figure 2 shows a significant run-up for short-term toeholds in the 30 days prior to the first rumour or announcement of a takeover bid. This run-up of approximately 15% in the thirty days prior to the adjusted announcement date is statistically significant at the 1% level. Consistent with the theoretical predictions of Bris (2002), these results provide preliminary evidence that the purchase of a toehold in the target generates heightened market anticipation of a takeover bid, leading to increased abnormal returns. Whilst it is 20
not possible to disentangle the effects of market anticipation and the market impact costs of the toehold acquisition, the two are intrinsically related. Market participants attempting to identify takeover targets may use unusual price and volume movements to inform their purchasing behavior (Choi, 1991). If this is the case, large trading volume will drive market anticipation, which will in turn drive additional trading volume. This circularity describes the mechanism by which toehold activity can cause prices to anticipate takeovers. Table 5 presents the results of the regression analysis using different model specifications. There are several interesting results. First, Model 1 shows that the size of the toehold alone cannot explain the variation in the run-up. This is consistent with the lack of consensus in the empirical literature regarding the impact of toehold size on the pre-bid run-up. Second, Models 2, 3, 4 and 5 demonstrate that short-term toeholds are associated with significantly increased price run-up in the pre-announcement period. This finding indicates that the presence of short-term toeholds are associated with larger pretakeover run-ups relative to long-term and no toehold positions, consistent with the results of Betton, Eckbo and Thorburn (2008b). This increased run-up provides support for the theoretical predictions of Bris (2002) that toehold activity can signal a bid to the market, generating market anticipation. Third, while short-term toeholds have a significant and positive impact on the run-up, long term toeholds present no statistical difference from acquisitions without toeholds. These results provide evidence consistent with the empirical results of Choi (1991), that toehold acquisitions generate price increases immediately following their acquisition. As toeholds have the potential to 21
signal the increased probability of future takeover, this causes a permanent price increase. In targets that have long-term toeholds, the impact of the acquisition is impounded well before the run-up period, leading to little additional anticipation of the takeover bid. <INSERT TABLE 5 HERE> Focusing on the control variables, the measure of abnormal turnover reported in Table 5 is positive and highly significant across all three measure of run-up. This finding suggests a positive association between increased market activity and abnormal pre-bid price increases. The market anticipation hypothesis proposed by Jensen and Ruback (1983) posits that traders predicting a takeover bid will buy in anticipation of receiving the takeover premium. This buying activity will also increase the volume of shares traded, consistent with our results. As found in previous studies, the coefficient on the deal size variable is positive and significant, demonstrating that larger target firms have a greater pre-bid share price run-up. This may be due to the comparatively larger amount of capital required to purchase a larger company resulting in a more predictable takeover process. Finally, the interaction terms in Table 5, show that the size of the toehold, with respect to the proximity of its placement to the announcement date, does not have a statistically significant impact on the size of the run-up. This is consistent with the insignificance of the toehold size variable in models one and three. This result demonstrates that the majority of the impact toeholds have on run-up lies not in their size but rather in the information conveyed by their purchase. This finding provides a potential explanation for 22
the inability of previous empirical findings to agree on the impact of toeholds size on runup, with Betton and Eckbo (2000) finding a significant and negative relationship between toeholds and the pre-bid run-up whilst Jarrell and Poulson (1989) show a significant positive relationship. The failure of previous empirical studies to agree on this direction could be due to their non-consideration of the timing of the acquisition of a toehold, with evidence from Betton, Eckbo and Thorburn (2008b) finding results similar to those of this study for short-term toeholds. The results presented in Table 5 suggest that the absolute size of the toehold does not offer any additional explanatory power when the timing of the toehold is taken into consideration. 6. Conclusions This paper documents and analyses the abnormal price run-up of 450 Australian takeover targets throughout a ten year trading period. We identify and control for three potential causes of pre-bid run-up: media speculation, price sensitive announcements and toehold acquisitions by the bidder. We show that the activity generated from these three areas is able to explain a majority of the pre bid share price run-up. We specifically assess how the timing of the acquisition of a toehold stake affects the price run-up and show that takeover bids with short-term toeholds are associated with significantly larger share price run-ups. Our results provide new evidence on the impact of the toehold timing decision which can potentially explain the failure of previous studies to reach a consensus on the impact of toeholds on pre-bid run-up. This evidence suggests that once the timing of the toehold decision is accounted for, there remains no statistically significant pre-bid run-up. 23
It is possible that the attribution of pre-bid run-up to insider trading found in prior studies such as those of Meulbroek (1992) and King (2009) may require re-examining in light of this potential source of pre-bid run-up. Whilst we cannot rule out the possibility that insiders traded prior to any individual takeover event, we are able to demonstrate that their existence is not so prevalent as to induce widespread unexplained pre-bid price run-ups, as has been previously claimed. However, as insider trading is by nature surreptitious, without prima facie evidence on the trading activity of actual insiders, it is not possible to rule out arguments such as that proposed by Chakravarty (2001), that insiders trade stealthily to conceal their trades. If insiders actively try to conceal their trades for fear of prosecution, it is possible that they trade more frequently when they observe abnormal price and volume movements generated by toehold acquisitions or by market rumours. However, without data describing the trades of prosecuted insiders, it is extremely difficult to ascertain if this behavior is occurring. Despite this potential limitation, our research suggests that almost all of the pre-bid runup which has previously been attributed to insider trading can be explained by legal factors associated with market anticipation. 24
References Aitken, M. & Czernkowski, R., 1992. Information Leakage Prior to Takeover Announcements: The Effect of Media Reports. Accounting and Business Research, 23(89), p.3. Barclay, M.J. & Holderness, C.G., 1991. Negotiated Block Trades and Corporate Control. The Journal of Finance, 46(3), pp.861-878. Betton, S. & Eckbo, B.E., 2000. Toeholds, Bid Jumps, and Expected Payoffs in Takeovers. The Review of Financial Studies, 13(4), pp.841-882. Betton, S., Eckbo, B.E. & Thorburn, K.S., 2008b. Markup Pricing Revisited. SSRN elibrary. Available at: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1094946 Betton, S., Eckbo, B.E. & Thorburn, K.S., 2009. Merger negotiations and the toehold puzzle. Journal of Financial Economics, 91(2), pp.158-178. Bishop, S.R., 1991. Pre-Bid Acquisitions and Substantial Shareholder Notices. Australian Journal of Management, 16(1), pp.1-33. Brealey, R. & Myers, S., 1996. Principles of Corporate Finance 5th ed., New York: McGraw-Hill. Bris, A., 2005. Do Insider Trading Laws Work? European Financial Management, 11(3), pp.267-312. Bris, A., 2002. Toeholds, takeover premium, and the probability of being acquired. Journal of Corporate Finance, 8(3), pp.227-253. Bruner, R.F., 2004. Applied mergers and acquisitions, John Wiley and Sons. Bulow, J., Huang, M. & Klemperer, P., 1999. Toeholds and Takeovers. The Journal of Political Economy, 107(3), pp.427-454. Chakravarty, S., 2001. Stealth-trading: Which traders trades move stock prices? Journal of Financial Economics, 61(2), pp.289-307. Available at: [Accessed May 20, 2011]. Chakravarty, S. & McConnell, J.J., 1997. An Analysis of Prices, Bid/Ask Spreads, and Bid and Ask Depths Surrounding Ivan Boesky s Illegal Trading in Carnation's Stock. Financial Management, 26(2), pp.18-34. Choi, D., 1991. Toehold Acquisitions, Shareholder Wealth, and the Market for Corporate Control. Journal of Financial and Quantitative Analysis, 26(03), pp.391-407. 25
Cornell, B. & Sirri, E.R., 1992. The Reaction of Investors and Stock Prices to Insider Trading. The Journal of Finance, 47(3), pp.1031-1059. Dodd, P., 1980. Merger proposals, management discretion and stockholder wealth. Journal of Financial Economics, 8(2), pp.105-137. Dubow, B. & Monteiro, N., 2006. Measuring Market Cleanliness. Financial Services Authority Occasional Paper, 23. Franks, J.R. & Harris, R.S., 1989. Shareholder wealth effects of corporate takeovers: The U.K. experience 1955-1985. Journal of Financial Economics, 23(2), pp.225-249. Jarrell, G.A. & Poulsen, A.B., 1989. Stock Trading before the Announcement of Tender Offers: Insider Trading or Market Anticipation? Journal of Law, Economics, & Organization, 5(2), pp.225-248. Jensen, M.C. & Ruback, R.S., 1983. The Market for Corporate Control: The Scientific Evidence. Journal of Financial Economics, 11, pp.5-50. Keown, A.J. & Pinkerton, J.M., 1981. Merger Announcements and Insider Trading Activity: An Empirical Investigation. The Journal of Finance, 36(4), pp.855-869. King, M.R., 2009. Prebid Run-Ups Ahead of Canadian Takeovers: How Big Is the Problem? Financial Management, 38(4), pp.699-726. Meulbroek, L.K., 1992. An Empirical Analysis of Illegal Insider Trading. The Journal of Finance, 47(5), pp.1661-1699. Mitchell, M. & Pulvino, T., 2001. Characteristics of Risk and Return in Risk Arbitrage. The Journal of Finance, 56(6), pp.2135-2175. Pound, J. & Zeckhauser, R., 1990. Clearly Heard on the Street: The Effect of Takeover Rumors on Stock Prices. The Journal of Business, 63(3), pp.291-308. Ravid, S.A. & Spiegel, M., 1999. Toehold strategies, takeover laws and rival bidders. Journal of Banking & Finance, 23(8), pp.1219-1242. Sanders, R.W. & Zdanowicz, J.S., 1992. Target Firm Abnormal Returns and Trading Volume Around the Initiation of Change in Control Transactions. Journal of Financial and Quantitative Analysis, 27(01), pp.109-129. Shleifer, A. & Vishny, R.W., 1986. Large Shareholders and Corporate Control. Journal of Political Economy, 94(3), p.461. Stulz, R., 1988. Managerial control of voting rights: Financing policies and the market for corporate control. Journal of Financial Economics, 20, pp.25-54. 26
Walkling, R.A., 1985. Predicting Tender Offer Success: A Logistic Analysis. The Journal of Financial and Quantitative Analysis, 20(4), pp.461-478. 27
Table 1 Characteristics of Target Firms This table provides descriptive statistics for the 450 target firms that form our final sample. Share price is reported as at 20 days prior to the takeover announcement. Deal value is the offer price multiplied by the number of shares to be purchased. Premium is the difference between the bid price and the price 20 days prior to the announcement of the bid. Market to book is a financial ratio comparing the deal value to the book value of net assets. EBIT is the earnings before interest and taxes and is taken from the last available annual report. Days news adjusted reports the number of days that the announcement date has been modified due to rumours of a potential merger in the news media mentioning both the bidder and target. Variable Mean Median Standard Deviation Share Price ($) 2.73 0.94 7.50 Deal Value ($ millions) 1,281.90 113.45 8,497.31 Premium (%) 18.28 16.10 23.98 Market to Book 3.03 1.83 4.10 EBIT ($millions) 62.38 4.24 547.74 Days News Adjusted 68.69 11.00 176.23 28
Table 2 Toehold Timing This table contains descriptive statistics relating to the average time that toeholds are held for. Short Term toeholds are defined as those that are purchased within 30 days of the official announcement of the takeover bid. Long Term toeholds are defined as those that are purchased outside of 30 days of the announcement of a takeover bid. Observations is the number of short or long-term toeholds in our sample of 450 firms. Holding Period Observations Mean Median Standard Deviation Short Term (Days) 52 10.26 1.00 55.65 Long Term (Days) 118 404.95 367 373.46 29
Table 3 Toehold Time Series Distribution This table documents the number and distribution of toehold bids within our sample on an annual basis. Bids is the number of takeover bids that were launched in each year. Short-term toeholds is the proportion of bids for that year where a toehold was acquired within the 30 days of the announcement of the bid. Long-term toeholds is the proportion of bids in which a toehold stake was acquired outside of 30 days prior to the announcement of the bid. Toeholds is the proportion of bids that had a non-zero toehold prior to the announcement of the bid. Year Bids Short-Term (%) Long-Term (%) Toeholds (%) 2000 50 16.0 32.0 48.0 2001 52 9.6 34.6 44.2 2002 30 23.3 23.3 46.6 2003 44 9.1 38.6 47.7 2004 36 13.9 27.7 41.6 2005 34 11.7 14.7 26.4 2006 64 12.5 18.7 31.2 2007 64 6.2 28.1 34.3 2008 33 12.1 30.3 42.4 2009 43 7.0 11.6 18.6 Average 45 12.2 26.0 38.1 30
Table 4 Toeholds by size and type Table 4 provides a breakdown of the size of toehold positions held by bidders in the sample. Short-term is the number of toeholds acquired in each size range within the 30 days prior to the takeover bid. Long-term is the number of toeholds acquired outside of thirty from the announcement date. Toehold (%) Short-term Long-term 0-5% 8 4 5-10% 6 9 10-15% 10 19 15-20% 23 31 20 25% 2 13 25 50% 3 42 31
Cumulative Abnormal Return Figure 1 News Adjusted Returns Figure 1 displays the cumulative average abnormal returns for the entire sample. Official date indicates the event time relative to the public announcement of the bid. News-adjusted indicates the returns relative to the news adjusted announcement date. News & Announcement adjusted reports the returns relative to the news adjusted date after the removal of targets that had a price sensitive news announcement in the 30 days prior to the announcement date. 25.00% 20.00% 15.00% 10.00% 5.00% 0.00% -5.00% -30-28 -26-24 -22-20 -18-16 -14-12 -10-8 -6-4 -2 0 2 Event Days Official Date News-Adjusted News & Announcement Adjusted 32
Cumulative Abnormal Return Figure 2 Toehold adjusted returns Figure 2 outlines the cumulative average abnormal returns by toehold type. Day 0 is the news adjusted announcement date. Long-term indicates stocks where a toehold was acquired outside of 30 days from the announcement. Short-term indicates toeholds that were acquired within the 30 days of the takeover announcement. No toehold indicates those targets where the bidder did not acquire any shares prior to the takeover announcement. 35.00% 30.00% 25.00% 20.00% 15.00% 10.00% 5.00% 0.00% -5.00% -30-28 -26-24 -22-20 -18-16 -14-12 -10-8 -6-4 -2 0 2 Event Days Short-Term Long-Term No Toehold 33
Table 5 Regression results: The determinants of pre-bid run-up This table presents results of a standard regression analysis used to examine the determinants of price run-up prior to the announcement of a takeover bid. The dependent variable, Run-up, is the average CAR between day -30 and -1; Toehold is the proportion of target shares held by the acquirer immediately prior to the takeover bid; Long-Term Dummy and Short-Term Dummy are dummy variables that takes 1 if a long-term (short-term toehold) was acquired in stock i and 0 otherwise. The base case is no-toehold; log(value) is the natural log of the market cap of stock i as at day -1 ; Turnover is the cumulative abnormal turnover in stock i as a % of clean period (-250,-100) turnover; Third Party is a dummy variable that takes 1 if a third party increases their substantial shareholding in stock i in the 30 days prior to the announcement and 0 otherwise; Long-Term*Toehold and Short-Term*Toehold are variables that indicate the percentage of stock i shares owned by the bidder on the date of announcement. Standard Errors are reported in brackets. *** represents significance at the 1% level, ** at 5% and * at 10%. Variable Model 1 Model 2 Model 3 Model 4 Model 5 Intercept 0.038*** 0.033*** 0.030* -0.113-0.151 (0.011) (0.010) (0.011) (0.078) (0.086) Toehold 0.018-0.047 - -0.024 (0.043) - (0.047) - (0.031) Long-Term - -0.021-0.033-0.029-0.103 - (0.026) (0.023) (0.030) (0.040) Short-Term - 0.116*** 0.113*** 0.104*** 0.103** - (0.017) (0.027) (0.039) (0.040) Log(Value) - - - 0.008* 0.010** - - - (0.004) (0.004) Turnover - - - 0.015*** 0.014*** - - - (0.005) (0.005) Third Party - - - -0.001-0.001 - - - (0.080) (0.018) Long-Term*Toehold - - - 0.051 0.032 Short- Term*Toehold - - - (0.690) (0.099) - - - 0.012 0.007 - - - (0.070) (0.171) Sample Size, N 450 450 450 450 450 Adjusted R-Squared 0.018 0.043.058 0.078 0.093 Yearly Fixed Effects - - - - Yes 34