Acquisitions and Performance: A Re-Assessment of the Evidence
Michael Bradley and Anant Sundaram* February 2006
* The authors are, respectively, F.M. Kirby Professor of Investment Banking and Professor of Law at Duke University, Durham, NC 27708 (
[email protected]) and Faculty Director of Executive Education at the Tuck School of Business at Dartmouth College, Hanover, NH 03755 (
[email protected]). We are grateful to Robert Burnham, Mohanraman Gopalan, Kate Dhoolyabhandit, and Vivek Sriram for superb research assistance. We thank David Denis, Diane Denis, Robert Hansen, Kose John, David Robinson, and especially Robert Whaley for their comments. We are also grateful to the participants in the Finance Seminar series at the Tuck School of Business at Dartmouth College. An earlier version of the paper was presented at the 2005 Western Finance Association meetings. A previous version of the paper was titled “Do Acquisitions Drive Performance or Does Performance Drive Acquisitions?”
Acquisitions and Performance: A Re-Assessment of the Evidence
ABSTRACT We document that a portfolio of acquiring firms significantly outperformed market benchmarks during the 1990s, and that frequent acquirers outperformed infrequent acquirers. This out-performance reflects superior stock price performance that occurs before, not after, acquisition announcements, implying that it is good performance that begets acquisitions rather than the reverse. In addition to this pre-acquisition stock price run-up, in the vast majority of cases, we observe a statistically and economically significant positive market reaction to the acquisition announcement itself. Further, contrary to recent studies, we find that acquirer size is not the most important determinant of the market reaction to an acquisition announcement. Instead, the target organizational form – i.e., whether the target is public or non-public – dominates all else. In addition, the size of the target and the medium of exchange are at least as important as acquirer size. Our empirical results lead us to conclude that the widely accepted attributions of “hubris” and “agency costs” to the motivations of the managers of acquiring firms are perhaps overstated, since they apply only to a small minority of cases where the target is relatively large and publicly traded, and stock is used as the sole medium of exchange. A substantial portion of M&A activity is consistent with shareholder value-maximizing behavior.
I. Introduction Mergers and acquisitions (M&A) are among the most extensively researched phenomena in financial economics. Comprehensive summaries of the traditional M&A research chronicle evidence from the 1970s (Jensen and Ruback (1983)), the 1980s (Jarrell, Brickley, and Netter (1988)), and the 1990s (Andrade, Mitchell, and Stafford (2001)). On average, target shareholders realize significant gains, while acquirer shareholders suffer slight losses; combined acquirer and target shareholder wealth increases significantly; acquirers do worse when the medium of exchange is stock rather than cash, both in the short-term and long-term; and the ability of acquirers to capture takeover gains is elusive – value is difficult for acquiring firms to create.1 The most commonly accepted explanations for these findings are managerial hubris and agency costs. More recent scholarship has led to more nuanced findings. Target organizational form (i.e., whether the target is publicly traded or not, and if not, whether the target is a subsidiary or is privately-owned), acquirer size, and the overvaluation of acquirers’ stock are all shown to affect the returns to acquiring firms’ stockholders.2 In much of this newer research, the agency costs view still predominates. For example, interpreting the finding of strongly positive market reactions to announcements of non-public-target acquisitions, especially those of larger targets using stock as the medium of exchange, Chang (1998), and Fuller et al. (2002) argue that the certification and monitoring benefits from such acquisitions work to lower agency costs. The fact that a single seller, presumably with superior information, is willing to accept the acquirer’s stock
1
For a summary of the U.S. evidence from a managerial perspective, see Bruner (2001); for the non-US evidence, see Denis and McConnell (2003).
2
Hansen and Lott (1996) were the earliest to note the effects of organizational form on announcement effects; see also Chang (1998); Fuller, Netter, and Stegemoller (2002); and Moeller, Schlingemann, and Stulz (2004a). Moeller, Schlingemann, and Stulz (2004a; 2004b) examine the role of size. Jensen (2003), Shleifer and Vishny (2003), Ang and Chen (2003), Malmendier and Tate (2003), Dong, Hirshleifer, Richardson, and Teoh (2003), and Rhodes-Kropf and Viswanathan (2003) examine the links between high valuations and the propensity of firms to make acquisitions.
2
as compensation for the firm’s assets sends a positive signal to the market. From this perspective, the selling entity provides a “certification” function regarding the value of the acquirer’s stock. In addition, holding a significant block of the acquirer’s stock after the acquisition gives the seller the incentive to “monitor” the subsequent behavior of the acquiring firms’ managers and thus reduces agency costs. In interpreting their findings that large acquirers account for much of the negative announcement effects observed in the data, Moeller et al. (2004a, p. 23) hypothesize that “….agency problems in large firms lead them to make poor acquisitions.” These authors assert that once the size of the acquiring firm is taken into account, other factors such as the medium of exchange and the organizational form of the target have little effect on acquiring firm returns. Jensen (2003) argues that overvalued stock leads to over-confidence and that this, in turn, leads managers to waste corporate assets on over-priced acquisitions. This explanation is an adaptation of his more general arguments concerning the agency costs of free cash flows to the case of M&A (e.g., Jensen (2000)).3 However there are those who view similar evidence and conclude that the managers of acquiring firms are in fact acting in the best interest of their stockholders. Hansen and Lott (1996) argue that in the case of a public firm acquiring a public target, a fully-diversified shareholder cares about the combined gains from the takeover, and not the division of the gains between the target and acquiring firms’ stockholders. However, when the target is non-public, the acquirer’s shareholders care only about the acquirer’s returns, since the target’s stock is not a part of their portfolios. Thus, the fact that announcements of acquisitions of non-public-targets elicit positive market responses – suggesting that the target management has agreed to sell the target assets at a discount – is consistent with acquirers acting in their shareholders’ interests. Shleifer and Vishny
3
In an empirical analysis, Malmendier and Tate (2003) find support for the Jensen (2003) view. They show that ‘overconfident’ managers are more likely to undertake acquisitions, and that the negative market reaction is significantly stronger for overconfident managers as opposed to ‘rational’ managers.
3
(2003) argue that rational, better-informed managers use their overvalued stock to acquire real assets and thereby take advantage of windows of opportunity offered by temporary market inefficiencies. According to their thesis, the inevitable value loss resulting from the acquisition is less than what would have transpired had the firm done nothing, since the market would eventually correct such overvaluations. By redefining the post-event performance benchmark in this second-best sense, i.e., comparing how the stock would have performed had the managers done nothing, rather than how the stock performed relative to some market benchmark, Shleifer and Vishny (2003) is also consistent with the notion that managers of acquiring firms are acting in their shareholders’ interests.4 Consulting firms selling M&A advice also claim that M&A transactions are consistent with managers acting on behalf of their shareholders. They argue that the relative size of the target and acquisition frequency matter. For instance, a study undertaken by the consulting firm McKinsey & Co. purports to find that ‘….while the average merger or acquisition destroys value for the acquirer, deals carried out by companies that undertake them strategically and often actually do create value’ (emphasis added) and, that ‘Like good surgeons, the best are the busiest, and the busiest are often the best’ (Frick and Torres (2002)). Another study, conducted by the consulting firm Bain & Co. purports to find that ‘…frequent acquirers that build skills and experience through a host of small deals come out on top’ and that, therefore, ‘executives should make a big deal about doing a lot of little deals’ (Harding and Rovit (2004)). A study by the Boston Consulting Group (Cools, King, Noonan, and Tsusaka (2004)) concludes that highly acquisitive firms in their sample outperformed firms that made few or no acquisitions by 29% during the course of a decade. Typical arguments advanced are that smaller acquisitions are easier to integrate; more likely to be in related businesses; more likely to benefit acquirers from 4
Ang and Chen (2003) find direct support for the Shleifer and Vishny argument. They conclude that overvaluation increases the probability of the firm becoming an acquirer, the use of stock as the medium of exchange and, although merged firms perform poorly afterwards, they do not perform any worse than a matched sample of firms.
4
learning-by-doing; less likely to be done for reasons of hubris and empire-building; and more likely to enable acquirers to exploit information asymmetries.5 The classic example of a firm pursuing a strategy of acquiring many small, related firms and using its stock as the medium of exchange is that of Cisco during the 1990s.6 Interestingly, while this debate continues, M&A remains (arguably) the most important means by which firms grow. For instance, from January 1980 to January 2000 there were nearly 70,000 completed M&A transactions worldwide, for a total value approaching $9 trillion.7 Based on the “accepted wisdom” sketched out above, many would conclude that these trillions of dollars are being spent unwisely. One might then wonder why managers continue to pursue growth through mergers and acquisitions. As Andrade et al. suggest (2001, p. 118), ‘….. in an efficient economy, there would be a direct link between causes and effects, [and] mergers would happen for the right reasons...’ The challenge of identifying and understanding some of these ‘right’ reasons – assuming they exist – remains. Perhaps it is indeed the case that managerial hubris and agency costs are the primary motivations for corporate acquisitions, or perhaps these are just partial explanations. It is possible that there are links between the types of acquisition strategies pursued, the characteristics of bids, the nature of targets, the medium of exchange and the relative size of the target that may
5
Asquith, Bruner, and Mullins (1983), and Schipper and Thompson (1983) examined announcement effects of acquisition programs announcements, and find a positive acquirer market effect; Fuller, Netter, and Stegemoller (2002) examined dynamic effects in a study of frequent bidders making several acquisitions within a 90-day period and find inconclusive results. Haleblian and Finkelstein (1999) study announcement effects associated with a set of 449 acquisitions, and find that there is a U-shaped relationship between a firm’s acquisition experience and acquisition performance implying learning-bydoing. Moeller et. al (2004b), however, provide evidence that firms that have the largest losses associated with acquisition announcements are serial acquirers who were previously successful in their acquisition efforts. 6
For an analysis of Cisco’s acquisition strategy see, Inkpen, Sundaram, and Rockwood (2000).
7
Based on analysis of completed transactions from the M&A database of the Securities Data Corporation, there were 68,621 M&A transactions worldwide during these two decades, for a total value US$8,845 billion. Of this, the 1980s accounted for about one-fifth of the total number and value, while the 1990s accounted for four-fifths.
5
shed light on the apparent conundrum of why we observe so many corporate acquisitions, in light of the available evidence. Using a large, comprehensive sample of acquisitions during the 1990s, we examine this core issue of whether the predominant behavior that is displayed by managers in undertaking acquisitions is consistent primarily with hubris and agency costs, or whether it is consistent with managers acting in the interests of their shareholders. Specifically, we examine the relation between the pre-acquisition and post-acquisition performance of acquiring firms; their acquisition strategies over time; the characteristics of the targets they acquire; and the returns realized by their stockholders. We base our empirical analysis on a data set of all completed acquisitions of public and non-public targets made by publicly-traded US acquirers between January 1990 and March 2000 listed in the SDC Mergers and Acquisitions database. The data consist of 12,476 completed acquisitions worth a total of $2.8 trillion undertaken by 4,116 firms. We conduct event-study analyses of the acquisitions in our sample, and time-series analyses of the acquiring firms’ stock prices. The remainder of the paper is organized as follows. In Section II we summarize our main empirical findings. Section III presents the detailed empirical analysis. Section IV interprets the findings, and suggests avenues for further research.
II. Main Empirical Findings We find that an equally-weighted “prescient” portfolio, consisting of the stock of all firms that became acquirers during the 1990s outperformed market benchmarks. (A valueweighted portfolio performed at least as well as market benchmarks.) The fact that acquiring firms outperformed the market during a decade that witnessed the largest bull market in history suggests that the managers of acquiring firms may well be acting in their stockholders’ interests. Sorting acquirers on the basis of acquisition frequency and the relative size of the target reveals
6
that frequent acquirers significantly outperformed infrequent acquirers, as did acquirers of relatively smaller targets. The buy-and-hold return to a portfolio of frequent acquirers of relatively small targets was over 130% higher than the return to a portfolio of infrequent acquirers of relatively large targets during the decade of the 1990s. While these findings appear to lend credence to the claim made by consulting firms that “doing a lot of little deals” is a value-creating strategy, a closer examination of the data reveals that acquisitions are made by firms whose stocks have performed well. In other words, performance drives acquisitions. Stocks of firms that undertook acquisitions in the 1990s outperform the market by 23% during the year prior to the acquisition. This superior performance is substantially higher for those firms that use stock as the medium of exchange and for those acquiring targets that are not publicly traded. Firms acquiring non-public targets using stock as the medium of exchange outperform the market by nearly 40% in the year prior to the acquisition, while firms acquiring public targets using stock outperform the market by about 20%. However, post-acquisition excess returns in the year following the acquisition are essentially zero, with no difference between acquirers of public versus non-public targets, nor between those using stock versus cash as the medium of exchange. We analyze the roles of size, organizational form, medium of exchange, and preacquisition run-up in determining the market reaction to acquisition announcements. Consistent with past research, we find that the market reacts negatively to announcements of public-target acquisitions and positively to acquisition announcements of non-public-target acquisitions.8 The average announcement effect is negative for public targets in 10 out of the eleven years studied and positive for non-public targets in all eleven years.
8
“Non-public” targets predominantly consist of two categories: privately owned firms, and subsidiaries or divisions of publicly owned firms. Since all the results are identical for these two sub-groups (indeed, slightly stronger for the latter sub-group), we do not analyze them separately in the results presented here. Detailed results of this analysis are available upon request.
7
This difference in the market reaction to the organizational form of the target holds after controlling for factors such as the relative size of the target, the size of the acquirer, the medium of exchange, and the pre-announcement run-up in the price of the shares of the acquiring firm. In terms of the relative impact on the average announcement return, we find that the organizational form of the target dominates all else. The market reacts negatively to the announcement of a stock-based acquisition of a public target, and the larger the relative size of the target, the larger the negative reaction; the market reacts positively to the announcement of an acquisition of a nonpublic target, and the larger the relative size of the target, the larger the positive reaction. The average market reaction to the announcement of a public-target acquisition is negatively related to the pre-announcement run-up of the acquirer’s stock. However the average market reaction to the announcement of a non-public-target acquisition is positively related to the pre-announcement run-up in the price of the shares of the acquiring firm. While acquirer size does matter – larger acquirers realize greater losses – the effects of the medium of exchange and the relative size of the target are at least as important. Categorical analysis of announcement effects sorted by quintiles of the relative size of the target, medium of exchange (cash, stock, or mixed), and target organizational form (public versus non-public) reveals that only three out of thirty categories have negative announcement effects: the negative returns correspond to acquisitions in which the relative size of the target is at least 25% or more if mixed financing is used, and 10% or more if stock is used. In the other 27 categories, the announcement effect is zero or positive: the market reaction is significantly positive in 16, and not significantly different from zero in 11 out of these 27 categories. The three categories with negative announcement effects account for only about 5% of the number of acquisitions in the sample. Moeller et al. (2004a) argue that a few acquisitions made by large acquirers account for the bulk of the losses suffered by the stockholders of acquiring firms, and that value losses generated by these firms more than outweigh the gains from all the other acquisitions combined. 8
However, this is not the case for our sample of firms during the 1990s. First, we find that, in the aggregate, the excess market value change in the acquirers’ stock price during the five days surrounding acquisition announcements is $110 billion, or about 4% of the value of all targets acquired. This figure masks the dichotomy arising from the organizational form of the target: acquirers lose $112 billion from acquisitions of public targets (5.7% of the value of all public targets), but they gain $222 billion from the acquisitions of non-public targets, which is a remarkable 26.1% of the aggregate value of all non-public targets acquired. Second, the three categories with significantly negative announcement effects (referred to above) collectively account for only $33 billion in losses. Our empirical evidence allows us to draw a number of inferences. First, portfolios of acquiring firms outperformed the market during the 1990s, which is contrary to the conventional wisdom. But, our evidence indicates that good stock-price performance precedes acquisitions. Thus, the practitioner advice from the major consulting firms that ‘firms should do a lot of little deals’ is suspect. Second, while the pre-acquisition run-up is higher for firms that use stock as the medium of exchange, it is substantially higher for firms that acquired non-public targets. Third, we do not observe significant post-acquisition negative drift in the stock price of acquirers regardless of the medium of exchange or target organizational form. Fourth, the traditional conclusion that ‘targets win and acquirers do not lose,’ which is widely attributed to managerial hubris, overvaluation and agency costs, is much too simplistic. While these are no doubt factors that afflict some acquisitions, the vast majority of acquisitions result in positive market reactions. The majority of M&A activity involves non-public targets, and these acquisitions result in statistically and economically significant positive announcement effects regardless of target or acquirer size, or mode of payment. Even in the case of public targets, negative announcement effects are observed only when the target is relatively large and when stock is used as the medium of exchange. And for this group, the aggregate value losses are
9
small compared to the total vale of targets acquired, as well as the aggregate value gains from all other acquisitions. Finally, the fact that acquisitions follow performance provides a relatively simple and straightforward explanation for the prevailing evidence that acquisitions are pro-cyclical and occur in ‘waves’ in specific industries as well as in the aggregate: if acquisitions follow good performance, we should observe a larger number of acquisitions in industries that experience good performance or, in the aggregate, acquisitions should occur during periods of high market valuations. Our evidence also calls into question the sweeping generality that much of the M&A activity is driven by managerial hubris and agency costs. It is difficult to conclude that the same group of acquirers is afflicted by agency problems with respect to one type of acquisition (public targets), and not the other (non-public targets). This is especially true since, in the aggregate, the number of the latter far dominates the number of the former.9 In sum, our evidence leads us to conclude that the majority of M&A activity is consistent with managers acting in their stockholders’ interests.
III. Detailed Analysis A. Description of the Data Our data are taken from the Securities Data Corporation (SDC) mergers and acquisitions database of Thompson Financial and includes all completed acquisitions of US targets made by publicly listed US firms, with deal values and acquirer name or ticker symbol available for the period January 1990 through March 2000. The sample contains all mergers and acquisitions in
9
As we will see, non-public target acquisitions account for four-fifths of all acquisitions. Also recall that the firms acquiring non-public targets had a higher pre-acquisition stock price run-up (compared to those acquiring public targets), and yet, the market reacts unambiguously positively to announcements of such acquisitions.
10
which the deal value is disclosed and includes LBOs, tender offers, and acquisitions of remaining interests (‘Deal Type’ items 1, 3, 4, and 11 in the SDC M&A database).10 Our initial data set contained 15,213 acquisitions undertaken by 5,020 U.S. firms publicly listed on the NYSE, ASE, or NASDAQ. The unavailability of ticker symbols from SDC, relevant data from CRSP and Compustat, and the elimination of acquisitions with deal values less than $100,000 reduced the final sample to 12,476 acquisition events, undertaken by 4,116 unique companies. Tables 1 through 5 summarize the characteristics of the events and firms in our sample. Focusing on Table 1, nearly 70% of the acquisitions were undertaken in the period 1995-1999, when the number of acquisitions averaged nearly 1,800 per year, compared to an annual average of less than 700 during the period 1990-94. The peak year was 1997, with 2,256 acquisitions. Roughly 80% of the acquisitions involve non-public targets. Only 20% of the sample represents acquisitions of publicly traded targets. These proportions are roughly constant throughout the decade. Related acquisitions (defined as acquisitions in which the target and acquirer have the same two-digit primary SIC code) account for about 60% of the sample, a proportion that fell slightly during the latter half of the 1990s. Two-fifths of all acquisitions are made using cash as the sole medium of exchange, while one-quarter used stock as the sole medium of exchange. Many have pointed out (e.g., Holmstrom and Kaplan (2001)) that the role of cash as a medium exchange fell during the 1990s and that of stock increased substantially. During the course of the decade, the proportion of cash-only deals in our sample fall from about 52% of all deals in 1990 to 39% in 1999, while the proportion of stock-only deals increase from less than 19% in 1990 to over 28% in 1999, consistent with the trend noted in the literature. The proportion of acquisitions made by NYSE-listed acquirers fall from about half in 1990 to about one-third in 1999, while the proportion for NASDAQ-listed acquirers increase from about two-fifths to three-fifths of all
10
The sample excludes other transactions that are also included in the SDC M&A database, i.e., spin-offs, recapitalizations, carve-outs, self-tenders, repurchases, exchange offers, privatizations, and minority stake acquisitions.
11
acquisitions. This is consistent with the rise of technology-dominated industries during the latter half of the 1990s. ASE-listed acquirers account for only 5% of the acquisitions in our sample. Table 2 contains summary statistics on the dollar value of the deals in our database. The total dollar value of all acquisitions over the period is $2.8 trillion. Average deal values and acquirer market values steadily rise during this period. In 1990, the average acquisition was for slightly less than $100 million, but by 1999, the average exceeded $400 million. The average market value of acquirers was about $2.1 billion in 1990, growing to $8.3 billion in 1999. Reflective of this trend, the period 1995-99 accounts for nearly 84% of the value of all acquisitions, despite the fact that only 70% of the number of all acquisitions occurred in this period. Although publicly-traded targets account for less than 20% of the acquisitions, they account for nearly 70% by value. The proportion of public-target acquisitions increased from about half in 1990 to over three-quarters in 1999, while those of non-public targets fell from about half in 1990 to about one-quarter by 1999. Over 70% of acquisitions by value are related, with no discernible pattern throughout the 1990s. Although cash-only acquisitions were 40% by number, they account for only 17% by value; stock-only acquisitions were only 25% by number, but account for over 45% by value. Acquisitions made by NYSE acquirers account for 78% of all acquisitions by value, while those by NASDAQ acquirers account for one-fifth. To summarize, the data on acquisition dollar-values show considerable skewness in that stock-based acquisitions of publicly traded targets, undertaken by NYSE-listed acquirers, are substantially larger than cash-based acquisitions of non-public targets, undertaken by NASDAQ-listed acquirers. As one might expect, the dollar-value measures are dominated by large NYSE firms, acquiring large public targets. The data on numbers tell the opposite story. The modal transaction in our database is an acquisition done by a NASDAQ-listed firm, for a relatively small, non-public target, using cash as the medium of exchange. Table 3 shows the distribution of acquirers by the number of acquisitions made during the sample period. While the average acquirer made 3 acquisitions during this period, about two12
fifths made only one acquisition. In fact, 83% of all acquirers made four or fewer acquisitions and 75% made three or less.11 Tables 4 and 5 break down by public- and non-public-target acquisitions, respectively. Of the 2,305 acquisitions of publicly traded targets (Table 4), over 70% are related; nearly half are carried out with stock as the sole medium of exchange; and over half are carried out by NYSElisted bidders. The average value of the deals for public targets is about $850 million, and the average acquirer size is $7.2 billion. Thus the average target was approximately 12% of the size of the average acquirer in the case of publicly listed targets. The proportion of stock used in public-target acquisitions increase; the proportion of NYSE-listed acquirers decline; and the proportion of NASDAQ-listed acquirers increase. The sample of non-public-target acquisitions (Table 5) tells a different story. Only about 55% are in businesses related to that of the acquirer’s. While cash-only deals account for 23% of public-target acquisitions, they account for 45% of non-public-target acquisitions. Stock-only deals comprise 50% of the public-target acquisitions but only 20% of the non-public-target acquisitions. The relative number of NYSE and NASDAQ acquirers is also reversed. About 55% of non-public-target acquisitions are undertaken by NASDAQ-listed acquirers, whereas only 38% are undertaken by NYSE-listed acquirers. The average value of a non-public-target acquisition is one-tenth that of the average public-target acquisition, roughly $84 million. The average acquirer size is about one-third that of a public-target acquirer, so the average non-public-target is about 3.5% of the size of the average acquirer. In sum, compared to public-target acquisitions, nonpublic-target acquisitions are less related; use less stock and more cash; are more technologydriven; and are much smaller, at about one-tenth the size.
11
A comparison of our sample with that of Fuller et al. (2002), which also covers roughly the same period and is from the same data source, suggests that their sample of ‘frequent acquirers’ (defined as those firms making five acquisitions or more) is approximately one-fifth of ours.
13
B. Firm-Level Analysis Figure 1 Panel A plots a Value Index (VI) for three portfolios over the decade of the 1990s: an equally-weighted portfolio of the acquirers in our sample, the equally-weighted market portfolio provided by CRSP, and a benchmark portfolio consisting of the returns to a matchedsample to the acquiring firms based on the Fama-French two-factor model. We revised this third portfolio each month according to the number of firms entering or leaving the sample and changes in each acquirer’s size and market-to-book ratio. We calculated a Value Index (VI) for each of the firms in our sample. The VI for each portfolio was calculated as: T
⎛
VI P,T = ∏ ⎜ 1 + t =1
⎝
1 Nt
⎞
Nt
∑ (R i =1
i,t
)⎟
⎠
(1)
where VIP,T is the Value Index for portfolio P at time T, Nt is the number of securities in the portfolio with a valid return reported on CRSP on day t, and Ri,t is the return to the ith security in the portfolio on day t. Thus, at any point in time, VI measures the value of $1 invested in the indicated portfolio on January 1, 1990 and held through date T. Note that VI minus 1 is the holding period return from January 1, 1990 through date T. The data in Figure 1 Panel A indicate that an equally-weighted portfolio of acquiring firms significantly outperformed the market portfolio, as well as the Fama-French portfolio over the decade of the 1990s.12 The VI for acquiring firms began to diverge from its benchmarks in 1994, which corresponds to the beginning of the merger wave of the mid-1990s. Over the entire period, $1 invested in the portfolio of acquirers grew to about $6, whereas $1 invested in either of the two benchmark portfolios grew to about $4. Figure 1 Panel B plots the time series of the VI for the three portfolios constructed on a value-weighted basis. For each portfolio, we calculated
12
Of course it would have been impossible to invest in this portfolio in real time, since one would have to be sufficiently prescient to know, on January 1, 1990, which firms would engage in M&A activity over the next 10 years.
14
T
⎛
Nt
⎞
t =1
⎝
i =1
⎠
VI P,T = ∏ ⎜ 1 + ∑ w i,t (R i,t ) ⎟
(2)
where wi,t is the value weight of security i relative to the portfolio of N firms with valid returns for month t. The data in Figure 1 Panel B indicate that the return to the portfolio of acquirers outperformed the value-weighted market portfolio over the decade of the 1990s. However, returns to the value-weighted portfolio of acquirers are not different from the value-weighted Fama-French portfolio. The statistics reported in Table 6 confirm the impressions from Figure 1. Table 6 reports the results of time series regressions of the monthly return to the portfolio of acquiring firms on the market portfolio and the Fama-French portfolio. The regressions were performed on both equally-weighted and value-weighted portfolios. The intercept is 0.4% in both equally-weighted regressions and both are statistically significant, indicating that the portfolio of acquirers realized an average monthly excess return of 0.4%. The compounded return over the period is equal to 1.004131 – 1.0 = 69%. Although only one-quarter of the size of the estimate in the equally-weighted regression, the constant in the value-weighted-market regression is statistically significant, again indicating superior returns to acquirers. However, the constant in the value-weighted Fama-French portfolio is not significantly different from zero. The attenuation in the significance of the constants in the value-weighted regressions indicates that size is an important factor in affecting the returns to acquiring firms. However, as our subsequent analysis will show, size isn’t everything. As we will see, the medium of exchange, the relative size of the target and target organizational form are important factors as well. In fact we find that target organizational form is the most important factor in determining the return to acquiring firms. In order to test the “consulting firm” prescription that it is more profitable for a firm to grow through a large number of small acquisitions rather than through a small number of large 15
acquisitions, we divide our sample of firms into two groups: “frequent” and “infrequent” acquirers. “Frequent” acquirers are firms that acquired four or more targets13 between January 1990 and March 2000, and “infrequent” acquirers are firms that acquired fewer than four. Table 7 reports data for these two portfolios sorted into quintiles of the relative size of targets (RST) acquired over the period.14 We measure RST for each acquirer by dividing the sum of all deal values made by the firm by the sum of the firm’s market value at the time of each transaction, thus giving us a value-weighted average. For each “frequent/infrequent” portfolio and each RST quintile, Table 7 reports the average monthly return, the Value Index (VI) of an equally-weighted portfolio (calculated according to Equation (1) above), and the average size of the acquiring firms. The data show that for every quintile of RST, the average monthly return is greater for the portfolios of frequent acquirers. Although the differences in the average monthly returns are not significant between frequent and infrequent acquirers – the t-statistics of the differences are all around 1.0 – the difference between the average return to the portfolio of frequent acquirers in the smallest RST quintile is significantly greater than the average return to the portfolio of infrequent acquirers in the largest RST quintile (t = 2.82). The average monthly return for all acquirers in the smallest quintile of RST is significantly greater than the average monthly return of all firms in the largest RST quintile, which seems consistent with the proposition that it is more profitable to acquire small rather than large targets. The same relation holds for the two sub-samples as well – the average return is higher for firms in the smallest RST quintile than in the largest quintile. However, the difference is statistically significant only for the sub-sample of infrequent acquirers.
13
We have also replicated the analysis categorizing ‘frequent acquirers’ as firms doing (1) three or more acquisitions, and (2) firms doing five or more acquisitions. The results using these two cut-off points are identical to those presented in the text.
14
We also ranked each sub-sample by its own quintiles. Since non-public targets are significantly smaller than public targets, this alternative methodology yields even stronger results.
16
The data in Table 7 show that the Value Index (VI), which is reported in the second line of each cell, is higher for frequent acquirers for every RST quintile. The VI for the portfolio of frequent acquirers is on average 37% higher than that for the portfolio of infrequent acquirers. Finally, the data show that the average size of acquirers (the third entry in each cell) is greater for frequent acquirers for every quintile of RST. In three of the five quintiles the t-Statistics of the difference is in excess of 3.0. This suggests that larger firms make more frequent acquisitions, and that the relative size of their targets (relative to their own size) is smaller than that of the smaller acquirers. It also suggests that larger firms perhaps became so by making more frequent acquisitions. As an additional test of the proposition that frequent acquirers earn greater returns than infrequent acquirers, we performed a time-series regression analysis of the returns to frequent acquirers on the returns to infrequent acquirers for each of the RST quintiles. Table 8 reports the summary statistics of these regressions. The data indicate that in every RST quintile, the constant is positive, and significantly greater than zero (t > 2.0) in three of the five quintiles. More importantly, the constant is significant (t = 3.39) in the regression that compares the returns to the portfolios of the “smallest- frequent” and the “largest-infrequent” acquirers. To summarize, we find that acquiring firms that pursue a strategy of acquiring a large number of relatively small firms outperform those pursuing a small number of large acquisitions. Thus we find a positive association between returns and acquisition frequency, and a negative association between returns and relative target size. But, demonstrating a relation is not equivalent to establishing causation. It may be that firms earn abnormal returns by frequently acquiring relatively smaller firms or it might be the case that firms that experience an increase in their stock prices engage in more acquisitions of relatively small firms.
17
C. Event Analysis We employe the standard event-study methodology to measure pre-announcement, announcement, and post-announcement returns to acquirers. We calculate excess returns based on Market-Adjusted Returns (MARs) and Cumulative Market-Adjusted Returns (CMARs):
MAR i,t = R i,t − R M,t
⎛ 1 ⎞ Nt MAR P,t = ⎜ ⎟∑ (R i,t − R M,t ) ⎝ N t ⎠ i=1
(2)
(3)
T
CMAR P,T = ∑ (MAR P,t )
(4)
t=1
σ P,T = σ T
(5)
where Ri,t is the return to the ith firm on day (month) t; RM,t is the return to the equallyweighted market portfolio on day (month) t; Nt is the number of firms in the portfolio on day (month) t; T is the end of the accumulation period as well as the number of periods; and σP is the time-series standard error of the MAR, estimated from the returns data prior to the event window. We center our data on the day or month of the acquisition announcement as appropriate, and calculate an average abnormal return to acquiring firm-events relative to this event date.
1. Time Series Analysis
Figure 2 Panel A plots the monthly CMAR to the portfolio of all acquisitions in our sample from 12 months before the announcement month through 12 months after. The CMAR from month –12 to month –1 is 19.01%, with a t-statistic of 16.30. (See Table 9.) The size and
18
significance of this pre-announcement run-up imply that firms with higher market valuations are more apt to pursue acquisitions. In other words, acquisitions are the result of good performance, not the cause. While the returns to the portfolio of all events show a slight negative postacquisition drift (t = –2.24), it is only –2.48% for the post-event year. Figure 2 Panel B plots the monthly CMARs to four portfolios based on the organizational form of the target (public or non-public) and the medium of exchange (cash-only or stock-only). Summary statistics for these portfolios are reported in Table 9 as well. While all four portfolios reveal a significant pre-acquisition run-up in stock price, the run-up of firms acquiring targets with stock is significantly greater than the run-up of firms acquiring targets with cash. And this is true for acquirers of both public and non-public targets. Note that none of the sub-sample portfolios experiences a significantly negative post-acquisition drift. Figure 3 Panel A plots the daily CMAR for the entire sample from day –20 through day +20. Supporting statistics are presented in Table 10. Both the pre- and post-announcement CMARs are statistically significant. The pre-announcement run-up from day –20 through day –3 is 0.55%, and the post-acquisition drift from day +3 through day +20 is –0.84%. The daily CMARs to the four portfolios based on the organizational form of the target and medium of exchange are plotted in Panel B of Figure 3. Supporting statistics are presented in Table 10. Stock-based acquisitions show a significant pre-announcement run-up over the 20 days prior to the acquisition announcement, and cash-based acquisitions show a significantly negative postacquisition drift over the 20 days after the announcement. Table 11 reports the average CMAR for event days –2 to +2, the average announcement period return, for all acquisitions by year and by target type. This return for the entire sample is significantly positive in each of the eleven years reported, and is 1.45% (t = 17.27) over the entire sample period. The average CMAR for the sample of public-target acquirers is negative in ten of eleven years, and is statistically significant in four. The overall average CMAR for this subsample is –0.71% (t = –3.89). In contrast, the return for non-public acquirers is positive and 19
statistically significant in all eleven years and is 1.95% (t = 20.66) for the entire sample period. In all eleven years, the average CMAR is greater for non-public-target acquisitions than for publictarget acquisitions, and in eight of the eleven years, the difference is statistically significant. The difference for the entire period is 2.66% (t = 12.31).
2. Cross-Sectional Analysis
Prior research has shown that the medium of exchange, acquirer size, relative size of target, organizational form of target, and prior performance of the acquirer’s stock all affect how the market reacts to an acquisition announcement.15 Table 12 presents the results of a crosssectional regression analysis of the impact of each of these variables on the market reaction to the announcement of the acquisitions in our database. The data are organized under three general headings: all targets, public targets and nonpublic targets. The dependent variable in all the regressions is the five-day CMAR centered on the announcement date. The results of Regression (1) indicate that the average announcement effect is significantly positive (4.49%, t = 15.51). The data also show that the announcement effect is negatively related to acquirer size, (Size, t = –9.25) and a dummy for whether the target is publicly traded, (Public, t = –9.50). Thus, while the size of the acquirer is an important determinant of the return to an acquisition announcement, so too is the organizational form of the target.
15
On the role of medium of exchange, see, for example, Asquith, Bruner, and Mullins (1987); Huang and Walking (1987); Travlos (1987). On the importance of the other variables, see Footnote 2 (also, our own results in a previous section have shown a significant role for the pre-acquisition run-up). While some prior research has found a role for relatedness, the overall evidence is mixed – our analysis shows that relatedness does not matter much, and in order to keep the presentation simple, we drop it from the analysis presented here. Others have shown that deal attitude (i.e., hostile or friendly; see, for example, Schwert (2000)) and bid competition (see, for example, Bradley, Desai, and Kim (1988)) matter. Since hostile takeovers of public targets fell dramatically during the 1990s, and in the case of non-public targets, less than 0.1% of the acquisitions in our data are defined as “hostile” by SDC, we drop it from analysis; similarly, since only about 2% of the acquisitions in our data are competed bids, we also drop this variable from analysis. None of the results presented here change if these variables are included in the regressions.
20
Regression (2) contains three additional independent variables: the relative size of the target (RST), a dummy variable equal to 1 if the medium of exchange is stock (Stock) and an interaction term that is the product of RST and the stock dummy variable (RST*Stock). Although the addition of these variables reduces the size of the three coefficients in Regression (1), they all have the same sign and are all statistically significant. The estimated coefficients on the three additional independent variables indicate that the announcement effect is positively related to the relative size of the target and is greater for stock- as opposed to cash-based transactions. The interaction term is significantly negative, which implies that the loss to acquirers using stock as the medium of exchange is greater, the greater the relative size of the target. This result is consistent with the notion that the more stock issued by the acquiring firm, the greater the negative signal sent to the market. This result also suggests that the more stock that is issued, the greater the dilution of the acquirer’s stock. In Regression (3) we include the 12-month run-up in the acquirer’s stock price prior to the acquisition announcement month.16 The announcement period return is positively related to the 12-month run-up. The addition of this independent variable does not change the signs or the statistical significance of the other independent variables. The dependent variable in Regressions (4) through (7) is the announcement period return to the acquirers of public targets in our sample. In all regressions, the constant is positive and significant, and the coefficient on acquirer size is negative and significant. The coefficients for RST and Stock are both negative and significant, indicating greater dilution for stock-based acquisitions of public targets, and that this dilution is greater the greater the relative size of the target. Regression (6) confirms this finding. The announcement period return is negatively related to RST and the RST*Stock interaction variable. Regression (7) includes the 12-month run-up as an additional independent variable. Note that the addition of this variable does not affect the signs
16
The results are essentially the same if we use the 6-month rather than the 12-month run-up.
21
or the significance of the independent variables in Regression (4). The coefficient on the 12month run-up is negative and significant. Regressions (8) through (11) report the results for the sub-sample of non-public-target acquisitions. Again we see that the constant in each of these regressions is positive and highly significant. The coefficient on acquirer size is negative and significant, suggesting that acquirer size matters independent of target organizational form. Perhaps the most interesting result in these regressions is the positive relation between the announcement return and the relative size of the target. (Recall that the announcement return and the relative size of the target are negatively related in the sample of public targets.) In addition, the announcement return for the sample of non-public targets is positively related to the use of stock, and as shown in Regression (10), the return is higher the larger the relative size of the target. Finally, note that in Regression (11) the announcement return for the portfolio of non-public targets is positively related to the 12-month pre-acquisition return. These results are exactly the opposite to the results we obtained for public targets. In sum, using stock as the medium of exchange to acquire a public target reduces the announcement return to the acquirer, and the reduction is greater the greater the relative size of the target. In contrast, using stock as the medium of exchange to acquire a non-public target results in a positive announcement return, and the returns are more positive the larger the relative size of the target. In Table 13 we examine further the difference in the announcement returns to public- and non-public-target acquisitions classified by acquirer type and medium of exchange. The data in the table report these differences for three types of acquirers: firms that acquired both public and non-public targets, firms that acquired only multiple public or multiple non-public targets, and firms that acquired only a single public or a single non-public target. For all three subsets, we see that the announcement returns are greater for acquirers of non-public targets than for acquirers of public targets, when stock is the medium of exchange. The differences are statistically significant for the first two subsamples (t = –8.25 and t = –5.07, respectively). These results suggest that 22
target organizational form and medium of exchange generate the observed differences independent of the characteristics of the acquiring firm. It is also worth noting that all the announcement returns are positive and when the medium of exchange is cash, and that the announcement returns for public-target acquisitions are significantly negative for the first two subsets and not different from zero for the third. Finally, it is interesting to note that of the 12,476 acquisitions in our sample, only 275 involve single acquisitions of public targets. This implies that the vast majority of empirical research involving mergers and acquisitions is based on firms that make multiple acquisitions. Table 14 presents a categorical analysis of the average CMAR by type of target, and medium of exchange, sorted by relative target size quintiles. This analysis complements the regression analysis reported in Table 12. For public targets acquired with cash, there is a nearly monotonic increase in the average CMAR as RST increases. In contrast, for stock-based public targets, the larger the RST, the more negative is the announcement return. The largest gain to the stockholders of acquirers of public targets is in the fifth RST quintile of cash-based acquisitions and the largest loss for this group is in the fifth quintile of stock-based acquisitions. Also, the average CMAR is monotonically decreasing in the fifth quintile as we move across the medium of exchange columns from cash, to mixed financing, to stock for public target acquisitions. In the case of non-public targets, the average CMAR increases monotonically as RST increases, for both cash-based and stock-based acquisitions. Also, it is monotonically increasing in the fifth quintile as we move across the medium of exchange columns from cash, to mixed financing, to stock deals. The largest gain to the stockholders of acquiring firms is 5.67% (t = 5.85), which corresponds to the acquirers in the largest RST quintile for non-public targets. The largest loss to acquirer stockholders is –2.74% (t = –7.61), which corresponds to stock-based acquisitions of public targets in the largest RST quintile. The data reported in Table 15 measure average CMAR to acquiring firms in “excess” dollar terms. The excess dollar value changes are calculated by multiplying the pre-offer market 23
value of each acquiring firm by its CMAR. ∆AMV is the average excess dollar change in the value of the acquiring firms, and ∆AMV/ΣAMV is the change in the excess dollar value of acquirers as a proportion of the sum of all acquirer values in that category. Similar to the data in Table 14, the data in Table 15 are presented by the type of target and the medium of exchange, sorted by quintiles of the relative size of the target. Table 15 reveals that the percentage dollar losses to the acquirers of public targets increase with the relative size of targets, whereas dollar gains to the acquirers of non-public targets increase with the relative size of the target. Moreover, the second-highest percentage loss to the acquirers of public targets is in the largest RST quintile in which stock is used as the medium of exchange. Likewise, the second-highest percentage gain to the acquirers of non-public targets is in the largest RST quintile for stock-based acquisitions. Perhaps the most important results are in the last row of the table, where the total dollar change and the total dollar return are reported for all acquisitions, public-target acquisitions, and non-public-target acquisitions. The data indicate that the total dollar change in the market value of acquirers of public targets is negative, whereas the total dollar change for the acquirers of nonpublic targets is positive. However note that the total dollar gain to the acquirers of non-public targets is almost twice as large as the total dollar loss to the acquirers of public targets. Thus, in the aggregate, the acquisitions in our sample generated a net gain of approximately $111 billion to the stockholders of acquiring firms during the 1990s.
D. Additional Empirical Analysis We conducted a number of additional empirical tests, the details of which we do not report here because they are diagnostics confirming the validity of the results reported above and the conclusions discussed below. The results of these empirical tests are available upon request.
24
We replicated all of our firm-event analyses with three-day, instead of five-day returns, and with beta-adjusted returns. The results are similar to those presented here. We examined the excess returns to the publicly traded targets in our sample and find that they realize significantly positive announcement returns on the order of those reported in the literature. In order to determine whether there is a learning-by-doing effect from serial acquisitions, we examine the average CMARs by acquisition order (i.e., the first acquisition, versus the second, versus the third, etc) and find no statistical difference in the results for these sub-samples. We replicated this analysis for the subset of frequent acquirers in our sample (with ‘frequent’ defined as three or more, four or more, and five or more) and only for firms that came into existence after December 1989. We find no statistical difference in the results for these sub-samples. We ran logistic regressions to examine whether pre-announcement run-ups are associated with the acquisition of public versus non-public targets, and the choice of stock versus cash as the medium of exchange. The logistic regressions indicate that the size of the pre-acquisition run-up is significantly positively related to the number of stock-based acquisitions of non-public targets. These results are consistent with our findings reported in Figures 2 and 3, as well as in Tables 9 and 10. Similar to the analysis in Moeller et al. (2004b), we examined the aggregate dollar-loss to the ‘large loss’ acquisitions in our sample (defined as those acquisitions in which the excess dollar loss to the acquirer was $1 billion or more), and find that the total dollar-loss is over $300 billion.17 We examined the aggregate excess market value-changes to the largest acquirers in our sample and find that the 5-day excess value-loss to the largest 100 acquirers totals $388 billion. As a direct test of the Moeller et al. (2004a) finding that acquirer size is the dominant variable in explaining market reactions to acquisition announcements, we compared the average market reaction to the announcement of acquisitions of similarly-sized public and non-public targets 17
Interestingly, however, when sorted by the dollar value of the deal size, aggregate value losses are much smaller in magnitude. The aggregate value loss to the largest 100 deals in our data is a much more modest $60 billion, and the loss to the largest 50 deals, $59 billion.
25
undertaken by ‘large’ acquirers. We made these comparisons defining target size relative to the acquiring firm and in absolute terms. We find that the opposite market reactions to the organizational form of the target – negative for public targets and positive for non-public targets – strongly persist, and these results hold even after controlling for the medium of exchange. In order to abstract from the effects of the market ‘bubble’ of the late 1990s, we re-ran all the regressions in Table 12 with year fixed effects. The dummy variable for the year 1999 is significantly positive in some of the regressions, but including year fixed effects does not materially change the size or statistical significance of the estimated coefficients.
IV. Conclusions In this paper we synthesize a number of empirical and theoretical issues concerning the returns to acquiring firms. Since our data set includes the near-universe of completed acquisitions by publicly traded US acquirers during the 1990s (limited of course by the coverage and accuracy of the SDC database) we are able to re-examine, confirm, or contradict a number of theories regarding the determinants of acquirer returns. We generate many new results, and suggest avenues for future M&A research. Among the key new findings is that the portfolio of stocks of acquirers over the decade of the 1990s outperformed market benchmarks. Moreover, acquirers that pursued a strategy of growing via many small acquisitions significantly outperformed those that grew via acquisitions of a few large targets. We then addressed the larger question: Is this superior performance due to acquisitions or because acquirers experienced superior performance prior to their acquisitions? Our examination of the pre- and post-announcement stock prices of acquiring firms clearly indicates that acquisitions follow good performance. We find a significant share-price run-up for acquiring firms in the year prior to an acquisition announcement. Another major finding is that
26
there is a significantly greater run-up for acquisitions in which the medium of exchange is stock rather than cash, and acquisitions in which a non-public target is acquired.18 Consistent with the literature, we find a negative announcement effect for public-target acquisitions, and a positive effect for non-public-target acquisitions.19 Here, we are able to extend the findings in previous studies – which have only looked at small samples (Hansen and Lott (1996); Chang (1998)) or particular types of acquirers (Fuller et. al (2002)) – and conclusively show that target organizational form is the most important variable that affects the market response to acquisition announcements. The announcement effect is positive for acquirers of nonpublic targets and is greater for stock transactions than cash. Also, for both the cash and the stock sub-samples of non-public targets, the larger the relative size of the target, the more positive the announcement effect. For public-target acquisitions, the announcement effect is positive for all cash deals and larger for larger cash deals. In contrast, the average announcement effect of an allstock deal is negative, and more so the larger is the transaction. The negative announcement returns to acquiring firms that is so well-documented in the M&A literature is due to no more than a few hundred acquisitions (out of our sample of nearly 12,500 firm-events) of large, publicly-traded targets that use their own stock as consideration in the transaction. We also show that these results are determined by the organizational form of the target form (public v. nonpublic) and the medium of exchange (cash v. stock) rather than the characteristics of the acquirer. The negative announcement returns to acquiring firms that is so well-documented in the M&A literature is due to no more than a few hundred acquisitions (out of our sample of nearly 12,500
18
We cannot overlook the fact, however, that the announcement of an acquisition occurs at the peak stock price for the acquirer. One can only speculate on what the stock prices might have done had the firms not made the acquisitions that they did. Would the positive trend in the stock price have continued unabated if not for the acquisition, or would it have eventually flattened off? Put differently, would the stockholders of firms that made acquisitions in the 1990s be better off had their firms not engaged in these transactions? Unfortunately, our data cannot answer this very important question. 19
The particularly strong positive excess returns to stock-based non-public target acquisitions by publicly traded acquirers is similar to findings of market reactions to private equity placements (as opposed to seasoned equity offerings; see, e.g., Hertzel and Smith (1993)), and the first-day stock price increase of IPOs (i.e., first-time equity issues done by previously non-public firms).
27
firm-events) of large, publicly-traded targets that use their own stock as consideration in the transaction. The traditional conclusion that ‘targets win and acquirers do not lose’ because of hubris or agency costs is perhaps much too simplistic. The vast majority of acquisitions result in positive announcement effects, and a sizeable proportion of these deals (four out of five) involve nonpublic targets. Often, the same firm acquires both public and non-public targets. As we have seen, non-public-target acquisitions result in statistically and economically significant positive announcement effects regardless of target or acquirer size, mode of payment, or prior stock price run-up. It thus appears unlikely that a firm that is redolent of agency problems reveals such behavior with just one type of acquisition that it makes (those of large public targets using stock), and not with many, if not most, of the other acquisitions it makes in its lifetime. The 5-day excess value-change for acquirer shareholders is economically significant. The total gain in excess value around acquisition announcements during the 1990s is $110 billion. But this aggregate amount masks an important dichotomy. The acquisitions of the non-public targets in our sample generated an aggregate gain of $222 billion, whereas the acquisitions of public targets in our sample generated an aggregate loss of $110 billion. In conjunction with the finding that the vast majority of acquisitions produce positive acquirer announcement effects, the results on market-value gains raise the possibility that many managers are driven by shareholder valueoriented behavior in their pursuit of acquisitions. In a related vein, we find that during the 1990s, acquirer size is not the dominant variable affecting the market response to an acquisition announcement. Factors such as the medium of exchange and relative target size are at least as important. In fact, the most important variable affecting the market reaction to an acquisition announcement is the organizational form of the target. Our findings regarding aggregate acquirer gains and the relative role of acquirer size are
28
inconsistent with the conclusions reached by Moeller et al. (2004a, 2004b).20 We note, however, that their sample of acquisitions spans both the 1980s and the 1990s, while ours only spans the 1990s. Additionally, while hubris, overconfidence, and overvaluation (Roll (1986); Jensen (2003); Shleifer and Vishny (2003)) may be attributes that afflict some acquirers, our overall evidence is more nuanced. It is unlikely that non-public-target acquisitions, which are only onetenth the size of public targets, are being driven by these considerations to the same degree as public-target acquisitions. And non-public-target acquisitions dominate the M&A universe. Our finding that performance drives acquisitions is consistent with the hypothesis of “stock-market driven” acquisitions espoused by Shleifer and Vishny (2003). However, their analysis does not apply to non-public-target acquisitions. We find that the market reacts differently to stock transactions depending on target organizational form. If the acquisition is of a publicly traded target and the consideration is stock, the market reacts negatively, whereas if the acquisition is of a non-public target – or a division of a publicly traded firm – the market reacts positively. We also do not find the negative post-acquisition drift predicted by their model, although we have not applied the latest statistical methods to detect long-run abnormal returns. Our findings raise questions regarding the explanation that certification and monitoring benefits in stock-financed non-public-target acquisitions reduce agency costs (Chang (1998); Fuller et. al (2002)), or the explanation that a non-public seller might be willing to cash out at a discount for liquidity or diversification reasons (Fuller et al. (2002)). It is well-known that in many large stock-based public-target acquisitions the former target managers continue their
20
Other findings in our study strengthen this conclusion. Recall we find an inverse relation between acquirer size and relative target size, and that larger acquirers tend to be more frequent acquirers. Yet we saw that it is precisely this portfolio of acquirers – those acquiring relatively smaller targets, and frequently – that outperformed all other portfolios and market benchmarks during the 1990s. Moreover, our analysis of market reactions to target organizational form holding constant relative and absolute acquirer and target size as well as medium of exchange – where the market reacts positively to non-public target acquisitions and negatively to public target acquisitions – is further supportive of this conclusion (although we do not report this analysis as a table).
29
involvement in the acquiring firm and continue to hold stock in the combined firm after the acquisition is completed. Unless the managers of large public-targets are consistently worse at evaluating an acquirer’s stock, or in monitoring the post-acquisition behavior of the managers of acquiring firms than the managers of non-public targets, the certification/monitoring hypothesis cannot explain the different market reaction to the acquisition of a public versus a non-public target. The explanation that the positive acquirer returns from non-public-target acquisitions may result from the seller wanting to cash out for liquidity or diversification reasons – and hence is willing to accept a discount – is also unsatisfactory. As we see in the data, approximately onethird of non-public targets are divisions of publicly-traded firms that are being sold off, and illiquidity or risk aversion cannot be an issue for this group of sellers – yet, the market reaction to the sell-off of a division is similar to those of private sellers. Our results are not entirely consistent with the Hansen and Lott (1996) argument of the role of investor diversification in market reactions to acquisitions of non-public firms. If their explanation were true, we should not observe similar market reactions to acquisition announcements for subsidiaries and divisions, as we do for acquisition announcements of private-owner targets. Presumably, market valuations of the cash flows from subsidiaries and divisions should already be reflected in the diversified investors’ portfolios. Yet, we observe similar market reactions to both categories of non-public-target acquisitions. The larger puzzle in our data is why the market reacts negatively only to the small subsample of stock-financed acquisitions of large publicly traded targets, and why managers continue to pursue such acquisitions. Signals of overvaluation cannot be the answer to why the market reacts the way it does, given the opposite signs in market reactions to public-target versus non-public-target acquisition announcements. Perhaps this puzzle can be partly explained by Mitchell, Pulvino, and Stafford (2004) who find that roughly one-half of the negative announcement effects for acquirers of public targets in stock-based transactions reflect downward price pressure caused by short-selling undertaken by risk arbitragers. As to why managers might 30
continue to pursue such acquisitions, the explanation could be that their behavior is consistent with shareholder value-orientation if we accept Shleifer and Vishny’s (2003) “second-best” benchmark of how the acquirer’s stock would have performed in the absence of the acquisition. The overall positive market reaction to non-public-target acquisition announcements may perhaps be the result of no more than the likelihood of greater information asymmetries. Given the greater information asymmetries between the buyer and a non-public seller, the buyer will seek a discount to compensate for a potential ‘lemons problem.’ There is likely to be an adverse selection discount, consistent with the arguments made by Rock (1986). This discount would accrue to the buyer as a premium. This evidence is also consistent with Hansen (2001) who argues – consistent with empirical evidence – that sellers in private company auctions actively limit the number of bidders and the amount of information they provide to bidders, despite considerable evidence from the auctions literature that a seller who wishes to maximize sale price should do the reverse. The resulting lower sale price would also accrue to the buyer as a premium. Finally, our findings provide a fairly straightforward explanation for merger waves (Rhodes-Kropf and Viswanathan (2003); Rhodes-Kropf, Robinson, and Viswanathan (2003)), and for M&A waves in particular industries (Mitchell and Mulherin (1996)). If it is true that higher market valuations occur in waves, then, since higher valuations drive acquisitions, we should not be surprised to find that aggregate levels of acquisition activity are pro-cyclical with valuation waves. This is also as predicted by Shleifer and Vishny (2003). A similar argument applies if higher stock valuations are industry-specific rather than market-wide phenomena. To the extent that higher market valuations are associated with specific industry groups at specific points in time, we should expect those industry groups to engage in greater levels of acquisition activity. Our study suggests a number of avenues for further research. Given that the M&A landscape is dominated by acquisitions of non-public targets, and given the persuasive evidence 31
that the market reacts very differently to this group of acquisitions, researchers should devote more time to the study of non-public-target acquisitions. It would be interesting to explore whether the private benefits to managers who engage in acquisitions are different between publictarget and non-public-target acquisitions – perhaps there is a difference in the compensation structures in acquiring one type versus the other. It is also possible that the auction mechanism for non-public companies works to the advantage of bidders (Hansen (2001)). The auction mechanism for non-public companies is still not fully understood, and is worthy of more research. It is possible that non-public-target acquisitions are more likely to be seller-initiated than buyerinitiated transactions, which would work to the advantage of buyers. Finally, it is possible that the acquirer gains from acquisitions of non-public targets represent the weaker bargaining position and valuation capabilities of less sophisticated target companies, which are perhaps more likely to be private than public. Conversely, the negative reactions to large, stock-based public-target acquisitions may be the result of higher-quality M&A advisors helping to bargain for a higher target premium. All of these are potentially fruitful avenues for future research.
32
References
Andrade, Gregor, Mark Mitchell, and Erik Stafford, 2001, New evidence and perspectives on mergers, Journal of Economic Perspectives 15, 103-120.
Ang, James, and Yinmei Chen, 2003, Direct evidence on the market-driven acquisitions theory, Florida State University Working Paper, May.
Asquith, Paul, Robert Bruner, and David Mullins, 1983, The gains to bidding firms from mergers, Journal of Financial Economics 11, 121-139. Asquith, Paul, Robert Bruner, and David Mullins, 1987, Mergers returns and the form of financing, Proceedings of the Seminar On the Analysis of Security Prices 34, 115-146.
Bradley, Michael, Anand Desai, and E. Han Kim, 1983, Synergistic gains from corporate acquisitions and their division between the stockholders of target and acquiring firms, Journal of Financial Economics, 21, 3-40. Bruner, Robert, 2001, Does M&A pay? A survey of the evidence for the decision-maker, Darden Graduate School of Business Working Paper #01-23, University of Virginia. Chang, Saeyoung, 1998, Takeovers of privately held targets, methods of payment, and bidder returns, Journal of Finance 53, 773-784.
Cools, Kees, Kermit King, Chris Neenan, and Miki Tsusaka, 2004, Growing through acquisitions: The successful value creation record of acquisitive growth strategies, The Boston Consulting Group, May.
Denis, Diane, and John McConnell, 2003, International corporate governance, Journal of Financial and Quantitative Analysis, 38, 1-36.
Dong, Ming, David Hirshleifer, Scott Richardson, and Siew Hong Teoh, 2003, Does investor misvaluation drive the takeover market? Ohio State University Working Paper.
Frick, Kevin, and Alberto Torres, 2002, Learning from high-tech deals, The McKinsey Quarterly 2002, 1-7.
Fuller, Kathleen, Jeffrey Netter, and Mike Stegemoller, 2002, What do returns to acquiring firms tell us? Evidence from firms that make many acquisitions, Journal of Finance 57, 1763-1793.
33
Haleblian, Jerayr, and Sydney Finkelstein, 1999, The influence of organizational acquisition experience on acquisition performance: A behavioral learning perspective, Administrative Sciences Quarterly 44, 29-56.
Hansen, Robert G., and John Lott, 1996. Externalities and corporate objectives in a world with diversified shareholders/consumers, Journal of Financial and Quantitative Analysis 31, 43-68.
Hansen, Robert, G., 2001, Auctions of companies, Economic Inquiry 39, 30-43.
Harding, Dan, and Sam Rovit, 2004, The mega-merger mouse trap, The Wall Street Journal, February 17.
Hertzel, Michael, and Richard Smith, 1993, Market discounts and shareholder gains for placing equity privately, Journal of Finance, 48, 459-485.
Holmstrom, Bengt, and Steve Kaplan, 2001, Corporate governance and merger activity in the US: Making sense of the ‘80s and ‘90s, Journal of Economic Perspectives 15, 49-68.
Huang, Yen-Sheng, and Ralph Walkling, 1987, Target abnormal returns associated with acquisition announcements: Payment, acquisition form, and managerial resistance, Journal of Financial Economics 19, 329-350.
Inkpen, Andrew, Anant Sundaram, and Kristin Rockwood, 2000, Cross-border acquisitions of US technology assets, California Management Review 42, 50-71.
Jarrell, Gregg, James Brickley, and Jeffrey Netter, 1988, The market for corporate control: The empirical evidence since 1980, Journal of Economic Perspectives 2, 49-68.
Jensen, Michael, 1986, Agency costs of free cash flow, corporate finance, and takeovers, American Economic Review 76, 323-329.
Jensen, Michael, 1988, Takeovers: Their causes and consequences, Journal of Economic Perspectives 2, 21-48.
Jensen, Michael, 2003, The agency costs of overvalued equity, Harvard Business School Working Paper.
Jensen, Michael, and Richard Ruback, 1983, The market for corporate control: The scientific evidence, Journal of Financial Economics 11, 5-50. 34
Malmendier, Ulrike, and Geoffrey Tate, 2003, Who makes acquisitions? CEO overconfidence and the market’s reaction, Stanford University Business School Working Paper, March 15.
Mitchell, Mark, and Harold Mulherin, 1996, The impact of industry shocks on takeover and restructuring activity, Journal of Financial Economics 41, 263-290.
Mitchell, Mark, and Harold Mulherin, 1996, The impact of industry shocks on takeover and restructuring activity, Journal of Financial Economics 41, 263-290.
Moeller, Sara, Frederik Schlingemann, and Rene Stulz, 2004a, Do shareholders of acquiring firms gain from acquisitions? Journal of Financial Economics (forthcoming).
Moeller, Sara, Frederik Schlingemann, and Rene Stulz, 2004a, Wealth destruction on a massive scale? A study of acquiring firm returns in the recent merger wave, Journal of Finance (forthcoming).
Mueller, D., 1987, Mergers and market share, Review of Economics and Statistics 67, 259-267.
Rhodes-Kropf, Matthew, and S. Viswanathan, 2004, Market valuation and merger waves, Journal of Finance, 59.
Rhodes-Kropf, Matthew, David Robinson, and S. Viswanathan, 2003, Valuation waves and merger activity: The empirical evidence, Journal of Financial Economics (forthcoming).
Rock, Kevin, 1986, Why new issues are underpriced, Journal of Financial Economics 27, 473524.
Roll, Richard, 1986, The hubris hypothesis of corporate takeovers, Journal of Business 59, 197216.
Schipper, Katherine and Rex Thompson, 1983, Evidence on the capitalized value of merger activity for acquiring firms, Journal of Financial Economics 11, 85-119.
Schwert, William, 2000, Hostility in takeovers: In the eyes of the beholder? Journal of Finance, 55, 2599-2640.
35
Shleifer, Andre, and Robert Vishny, 2003, Stock-market driven acquisitions, Journal of Financial Economics, 70, 295-312.
Singh, Harbir, and Mauricio Zollo, 1997, Learning to acquire: Knowledge accumulation mechanisms and the evolution of post-acquisition management practices in the banking industry, Wharton Financial Institutions Center Working Paper 97-10, University of Pennsylvania.
Travlos, Nickolaos, 1987, Corporate takeover bids, methods of payment, and bidding firms’ stock returns, Journal of Finance 42, 943-963.
36
Table 1 Summary Statistics on Acquisition Activity and Acquisition Characteristics of US Acquirers: January 1990 – March 2000 Data are from the Securities data Corporation’s (SDC) Mergers and Acquisitions database and consist of 12,476 acquisitions undertaken by 4,116 publicly listed US acquirers from 1/90 through 3/00 for which deal values are available. ‘Public target’ is as defined by SDC. ‘Related’ is defined as an acquisition in which the acquirer and target have the same two-digit primary SIC code. ‘Cash only,’ ‘Stock only,’ the exchange in which the acquirer is listed (NYSE/NASDAQ), and deal value are as from SDC. Acquirer market value is the beginning-of-month market value of acquirer’s equity during the month in which an acquisition took place, from the Compustat database. Year Number of Targets
Number of Public Targets
Number of Related Acquisitions
Number Cash Only
Number Number of Stock NYSE Only Acquirers
Number of NASDAQ Acquirers
Avg. Deal Value ($ million)
Avg. Bidder Mkt. Value ($ million)
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000*
329 405 614 853 1188 1222 1608 2256 2195 1525 281
67 80 103 133 222 265 304 375 393 320 43
262 269 406 554 700 758 884 1081 1121 863 148
172 156 219 346 470 471 636 933 1005 589 96
61 95 158 201 272 342 402 491 460 432 97
167 191 268 322 427 442 664 1050 965 544 75
134 180 302 454 683 710 871 1117 1086 916 192
98.7 104.5 86.3 112.4 104.5 158.8 171.1 313.9 391.9 400.8 382.1
2113.7 1496.2 1410.9 1353.5 1532.9 1409.3 2087.2 4500.1 3420.7 8264.2 16029.3
199900
12476
2305
7046
5093
3011
5075
6645
225.2
3304.5
*
Data only through March 2000
Table 2 Summary Statistics on Dollar Value of Acquisition Activity and Acquisition Characteristics of US Acquirers: January 1990 – March 2000 (In US$ Billions) Data are from the Securities data Corporation’s (SDC) Mergers and Acquisitions database and consist of 12,476 acquisitions undertaken by 4,116 publicly listed US acquirers from 1/90 through 3/00 for which deal values are available. ‘Public target’ is as defined by SDC. ‘Related’ is defined as an acquisition in which the acquirer and target have the same two-digit primary SIC code. ‘Cash only,’ ‘Stock only,’ the exchange in which the acquirer is listed (NYSE/NASDAQ), and deal value are as from SDC. Acquirer market value is the beginning-of-month market value of acquirer’s equity during the month in which an acquisition took place, from the Compustat database.
Year
All Targets
Public Targets
Non-public Targets
Related Stock Only Cash Only NYSE Bidder Acquisitions Acquisitions Acquisitions Acquisitions
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000*
32.47 42.30 52.96 95.91 124.10 194.05 275.14 413.27 860.30 611.18 107.36
15.89 27.13 25.47 63.54 71.70 136.17 183.40 253.75 663.44 460.86 56.42
16.59 15.18 27.49 32.37 52.40 57.88 91.74 159.52 196.87 150.32 50.93
23.62 32.58 37.48 63.44 77.65 121.02 164.51 245.09 700.71 457.56 71.49
11.18 13.91 13.78 26.59 26.39 79.73 90.88 182.94 476.43 294.32 60.63
10.54 7.21 13.99 16.97 50.43 48.08 47.46 77.54 104.68 78.21 22.55
199000 2809.04
1957.77
851.29
1995.15
1276.76
477.64
*
Data only through March, 2000
37
27.57 33.73 42.43 63.71 85.72 155.66 208.45 317.04 751.50 459.25 47.64
2192.70
NASDAQ Bidder Acquisitions 4.17 7.31 9.27 20.13 28.82 36.47 64.65 89.22 102.99 148.46 59.22
570.73
Table 3 Acquisition Frequency of Publicly Listed US Acquirers: January 1990 – March 2000 Data are from the Securities Data Corporation’s (SDC) Mergers and Acquisitions database and consist of 12,476 acquisitions undertaken by 4,116 publicly listed US acquirers from 1/90 through 3/00 for which deal values are available.
Number of Acquisitions 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 – 62
Number of Bidders 1748 958 533 331 184 155 97 56 62 37 22 28 16 16 12 7 8 6 5 3 35
Percent of Bidders 40.5% 22.2% 12.3% 7.7% 4.3% 3.6% 2.2% 1.3% 1.4% 0.9% 0.5% 0.6% 0.4% 0.4% 0.3% 0.2% 0.2% 0.1% 0.1% 0.1% 0.8%
38
Cumulative Percent 40.5% 62.7% 75.0% 82.7% 86.9% 90.5% 92.8% 94.0% 95.5% 96.3% 96.9% 97.5% 97.9% 98.2% 98.5% 98.7% 98.9% 99.0% 99.1% 99.2% 100.0%
Table 4 Summary Statistics On Acquisition Activity and Characteristics of Public Target Acquisitions Data are from the Securities data Corporation’s (SDC) Mergers and Acquisitions database and consist of 12,476 acquisitions undertaken by 4,116 publicly listed US acquirers from 1/90 through 3/00 for which deal values are available. ‘Public target’ is as defined by SDC. ‘Related’ is defined as an acquisition in which the acquirer and target have the same two-digit primary SIC code. ‘Cash only,’ ‘Stock only,’ the exchange in which the acquirer is listed (NYSE/NASDAQ), and deal value are as from SDC. Acquirer market value is the beginning-of-month market value of acquirer’s equity during the month in which an acquisition took place, from the Compustat database. Year Number of Targets
Number of Public Targets
Number of Related Acquisitions
Number Cash Only
Number Number of Stock NYSE Only Acquirers
Number of NASDAQ Acquirers
Avg. Deal Value ($ million)
Avg. Bidder Mkt. Value ($ million)
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000*
329 405 614 853 1188 1222 1608 2256 2195 1525 281
67 80 103 133 222 265 304 375 393 320 43
40 58 83 108 159 189 195 257 278 216 29
24 12 18 34 51 63 72 59 74 73 13
24 35 54 57 120 145 138 198 207 158 23
39 47 57 70 119 135 168 207 206 168 17
25 27 41 57 92 117 131 155 172 142 20
237.1 339.1 247.3 477.8 323.0 513.9 603.3 676.7 1688.1 1440.2 1312.2
3314.6 2349.8 2157.7 3083.0 2707.5 3180.1 4296.0 5910.2 7972.1 20477.0 20788.2
199900
12476
2305
1612
493
1159
1233
979
849.3
7192.4
*
Data only through March, 2000
Table 5 Summary Statistics On Acquisition Activity and Characteristics of Non-Public Target Acquisitions Data are from the Securities data Corporation’s (SDC) Mergers and Acquisitions database and consist of 12,476 acquisitions undertaken by 4,116 publicly listed US acquirers from 1/90 through 3/00 for which deal values are available. ‘Public target’ is as defined by SDC, and ‘Non-public target’ is a target that is not publicly traded. ‘Related’ is defined as an acquisition in which the acquirer and target have the same two-digit primary SIC code. ‘Cash only,’ ‘Stock only,’ the exchange in which the acquirer is listed (NYSE/NASDAQ), and deal value are as from SDC. Acquirer market value is the beginning-of-month market value of acquirer’s equity during the month in which an acquisition took place, from the Compustat database. Year Number of Targets
Number of Non Public Targets
Number of Related Acquisitions
Number Cash Only
Number Number of Stock NYSE Only Acquirers
Number of NASDAQ Acquirers
Avg. Deal Value ($ million)
Avg. Bidder Mkt. Value ($ million)
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000*
329 405 614 853 1188 1222 1608 2256 2195 1525 281
251 321 510 718 960 956 1301 1878 1801 1205 238
144 208 322 444 540 568 687 892 842 647 119
142 144 201 312 414 407 562 872 931 516 83
34 58 104 142 152 197 273 293 253 274 74
127 143 210 252 303 307 495 801 758 376 58
101 150 261 395 590 593 739 961 914 774 172
65.6 47.2 53.9 45.1 54.6 60.5 70.5 82.9 108.2 124.8 214.0
1877.2 1298.6 1259.3 1036.7 1262.5 919.7 1575.1 2217.5 2394.4 5028.9 15169.5
199900
12476
10139
5413
4584
1854
3830
5650
83.4
2422.5
*
Data only through March 2000
39
Figure 1 Value Indexes for Equally-Weighted and Value-Weighted ‘Prescient’ Portfolios of Acquirers, Market Portfolio, and Fama-French Benchmark Portfolio: January 1990 – October 2000
T ⎛
⎞
1 N
∑ (R ) , where T is the last The value index (VINDX) for each portfolio is calculated as VINDXT = ∏ ⎜⎜1 + i,t N t = 1⎝
⎠
i=1
month in the calculation (October 2000), N is the number of firms in the portfolio, and Ri,t is the return to the ith security in the portfolio on month t. The ‘prescient’ portfolio of acquirers is created as follows: A firm that made at least one acquisition anytime during the period 1/90 – 3/00 is added to the portfolio as of January 1990 (if already listed) or the month when it first became listed during this period; an equally-weighted (value-weighted) portfolio is created, and rebalanced every month thereafter, for the period 1/90 – 10/00. The market portfolio is the equally-weighted (valueweighted) CRSP index. The Fama-French benchmark is an equally-weighted (value-weighted) portfolio of firms matched every month by size and market-book ratio with the firms in the acquirers’ portfolio, and similarly rebalanced. A. Equally-Weighted Portfolios
$8
$7
$6 Acquirers $5
$4
Market FF
$3
$2
$1
$0 Jan90
Jul90
Jan91
Jul91
Jan92
Jul92
Jan93
Jul93
Jan94
Jul94
Jan95
Jul95
Jan96
Jul96
Jan97
Jul97
Jan98
Jul98
Jan99
Jul99
Jan00
Jul00
Jan98
J u l98
Jan99
J u l99
Jan00
J u l00
B. Value-Weighted Portfolios
$6
$5 A c q u ir e r s FF $4 M a rk e t
$3
$2
$1
$0 Jan90
J u l90
Jan91
J u l91
Jan92
J u l92
Jan93
J u l93
Jan94
J u l94
Jan95
J u l95
40
Jan96
J u l96
Jan97
J u l97
Table 6 Regressions of the Monthly Returns to the ‘Prescient’ Portfolio of Acquirers on the Market Portfolio and the Fama-French Benchmark Portfolio: January 1990 – July 2000 The dependent variable is the monthly return to a ‘prescient’ portfolio of acquirers. The independent variables are the monthly return to the market portfolio and the monthly returns to a matched Fama-French portfolio. The ‘prescient’ portfolio of acquirers consists of all firms that that made at least one acquisition anytime during the period 1/90 – 3/00. Firms are added or subtracted from this portfolio as they are added to or deleted from the CRSP database. Monthly returns are calculated to equally-weighted and valueweighted portfolios of: (1) acquiring firms, (2) the CRSP Market portfolio, and (3) the returns to a matched Fama-French 2-factor portfolio. t-statistics in parentheses ( ** significant at 5%, *** significant at 1%) Dependent Variable – Return to Portfolio of Acquirers Equally-Weighted Intercept
0.004 (3.90)***
0.004 (3.40)***
Value-Weighted 0.001 (2.02)**
0.000 (0.36)
Equally-Weighted Market
0.992 (52.99)***
Fama-French
0.928 (41.88)***
Value-Weighted Market
1.005 (112.19)***
Fama-French N Durbin-Watson Statistic Adjusted R2
0.995 (99.98)*** 131 1.79 0.96
Dependent Variable: Mean Variance Independent Variable: Mean Variance
131 2.04 0.93
131 1.87 0.99
0.0154 0.0025 0.0103 0.0024
0.0077 0.0027
41
131 1.97 0.99
0.0101 0.0017 0.0083 0.0017
0.0101 0.0017
Table 7 Average Monthly Return, Value Index, and Average Acquirer Size for ‘Prescient’ Portfolios of All Acquirers, Frequent Acquirers and Infrequent Acquirers, and by Quintiles of Relative Size of the Target (RST)
T The value index for acquirer i is calculated as VINDXi = ∏ 1 + R i,t , where T is the last month in the calculation t =1 (October 2000), the first month is January 1990, and Ri,t is the return to the ith acquirer in month t. An acquirer is a firm that made at least one acquisition during the period 1/90 – 3/00. The number of acquisitions is from the SDC Mergers and Acquisitions database and consists of 12,476 acquisitions undertaken by 4,116 publicly listed US acquirers from 1/90 through 3/00 for which deal values are available. RST is the value-weighted average of the ratio of the acquisition value to the market value of the acquirer in the month prior to the acquisition of all acquisitions undertaken by Firm i during this period. Frequent Acquirers are defined as firms that made five or more acquisitions during this period, and Infrequent Acquirers are defined as firms that made four acquisitions or less during this period. t-statistics in parentheses (*** significant at 1%, **significant at 5%)
(
RST Quintile
All Acquirers
Smallest Quintile (RST ≤ 0.0431) Avg Monthly Return 1.72% Value Index 7.89 Avg Acquirer Size ($MM) 4,263 N 822 2nd Quintile (0.0431 < RST ≤ 0.0902) Avg Monthly Return 1.66% Value Index 7.37 Avg Acquirer Size ($MM) 1,657 N 825 3rd Quintile (0.0902 < RST ≤ 0.1661) Avg Monthly Return 1.48% Value Index 5.96 Avg Acquirer Size ($MM) 1,312 N 822 4th Quintile (0.1661 < RST ≤ 0.3021) Avg Monthly Return 1.52% Value Index 6.28 Avg Acquirer Size ($MM) 1,214 N 824 Largest Quintile (RST > 0.3021) Avg Monthly Return 1.33% Value Index 4.93 Avg Acquirer Size ($MM) 781 N 824 Smallest minus Largest Avg Monthly Return t-statistic Value Index Avg Acquirer Size ($MM)
0.39 2.21*** 2.96 3,482
)
Frequent Acquirers
Infrequent Acquirers
1.96% 10.97 10,085 131
1.67% 7.29 3,146 691
0.29% 3.68 6,939
1.57
1.83% 9.43 2,901 183
1.60% 6.67 1,305 642
0.23% 2.76 1,597
0.92
1.61% 7.16 2,760 131
1.44% 5.49 837 691
0.17% 1.67 1,922
0.82
1.63% 7.20 2,624 167
1.50% 6.01 855 657
0.13% 1.19 1,769
0.64
1.62% 6.89 1,802 62
1.30% 4.74 698 762
0.32% 2.15 1,104
1.07
0.34 1.00 4.08 8,283
0.37 2.06*** 2.55 2,448
Smallest Frequent minus Largest Infrequent Avg Monthly Return 0.66 t-statistic 2.82*** Value Index 6.23 Avg Acquirer Size ($MM) 4,262
42
Difference
t-statistic
3.25***
3.22***
3.14***
1.72
1.08
Table 8 Time Series Regression of Returns to Frequent Acquirers On the Returns to Infrequent Acquirers, by Quintiles of the Relative Size of the Target (RST) The dependent variable is the monthly return to a ‘prescient’ portfolio of frequent acquirers. The independent variable is the monthly return to a ‘prescient’ portfolio of infrequent acquirers. An acquirer is a firm that made at least one acquisition during the period 1/90 – 3/00. The number of acquisitions is from the SDC Mergers and Acquisitions database and consists of 12,476 acquisitions undertaken by 4,116 publicly listed US acquirers from 1/90 through 3/00 for which deal values are available. RST is the valueweighted average of the ratio of the acquisition value to the market value of the acquirer in the month prior to the acquisition of all acquisitions undertaken by Firm i during this period. Frequent Acquirers are defined as firms that made five or more acquisitions during this period, and Infrequent Acquirers are defined as firms that made four acquisitions or less during this period. t-statistics in parentheses (*** significant at 1%, **significant at 5%) RST Quintile
Intercept
Slope
Adjusted R2
D.W.
Smallest Quintile (RST ≤ 0.0431)
0.560% (3.29)***
0.836 (26.62)***
0.86
1.75
2nd Quintile (0.0431 < RST ≤ 0.0902)
0.649% (2.88)***
0.738 (18.58)***
0.72
1.86
3rd Quintile (0.0902 < RST ≤ 0.1661)
0.485% (2.50)***
0.783 (21.71)***
0.78
1.59
4th Quintile (0.1661 < RST ≤ 0.3021)
0.432% (1.63)
0.862 (21.68)***
0.78
1.80
Largest Quintile (RST > 0.3021)
0.461% (1.52)
0.890 (14.62)***
0.62
1.77
Smallest Frequent On Largest Infrequent
0.798% (3.39)***
0.889 (18..92)***
0.86
1.98
43
Table 9 Pre-event and Post-event Cumulative Monthly Market-Adjusted Announcement Period Returns (CMMAR) −1
Pre-event CMMAR =
∑(
RP.t − RM ,t
t=−T
)
T
(
and Post-event CMMAR = ∑ RP.t − RM ,t t=1
) where
RP,t is the
monthly return of an equally-weighted portfolio of firm-events for month t, RM,t is the monthly return of the equally-weighted CRSP index for month t, and zero is the event month. Data are from the Securities data Corporation’s (SDC) Mergers and Acquisitions database and consist of 12,476 acquisitions undertaken by 4,116 publicly listed US acquirers from 1/90 through 3/00 for which deal values are available. ‘Public target’ is as defined by SDC, and ‘Non-public target’ is a target that is not publicly traded. z-statistics in parentheses (*** significant at 1%; ** significant at 5%); the test statistic for monthly analysis is calculated as {CMMAR ÷ [σCMMAR√M]}, where M is the number of months. σCMMAR is calculated using the average excess returns over the market returns to the respective portfolios from Month –60 to Month –14.
Pre-event CMMAR
Post-event CMMAR
Month [–12 to –1]
Month [+1 to +12]
All Acquirers
19.01% (16.30)***
– 2.48% (–2.14)***
Acquirers of Public Targets with Stock Acquirers of Public Targets with Cash
22.89% (14.20)*** 8.47% ( 6.05)***
– 1.69% (–1.15) 1.68% ( 1.20)
Difference: Stock v. Cash for Public
14.42% ( 6.75)***
– 3.37% (–1.57)
Acquirers of Non-Public Targets with Stock Acquirers of Non-Public Targets with Cash
39.05% (12.40)*** 12.00% (10.97)***
– 3.63% (–1.15) – 0.53% (–0.49)
Difference: Stock v. Cash for Non-public
27.05% ( 8.11)***
– 3.10% (–0.93)
44
Figure 2 Pre-event and Post-event Cumulative Monthly Market-Adjusted Announcement Period Returns (CMMAR) −1
Pre-event CMMAR =
∑ (RP.t − RM ,t )
T
(
and Post-event CMMAR = ∑ RP.t − RM ,t
t=−T
t=1
) where
RP,t is the
monthly return of an equally-weighted portfolio of firm-events for month t, RM,t is the monthly return of the equally-weighted CRSP index for month t, and zero is the event month. Data are from the Securities data Corporation’s (SDC) Mergers and Acquisitions database and consist of 12,476 acquisitions undertaken by 4,116 publicly listed US acquirers from 1/90 through 3/00 for which deal values are available. ‘Public target’ is as defined by SDC, and ‘Non-public target’ is a target that is not publicly traded. A: All Events ALL FIRMS 25%
20%
15%
10%
5%
0% -12
-9
-6
-3
0
3
6
9
12
Event Month
B: Indicated Events
50%
Non-Public Stock 40%
30%
Public Stock
20%
Non-Public Cash 10%
Public Cash 0% -12
-9
-6
-3
0 Event Month
45
3
6
9
12
Table 10 Pre-event and Post-event Cumulative Daily Market-Adjusted Announcement Period Returns (CDMAR) −1
Pre-event CMMAR =
∑ (RP.t − RM ,t )
T
(
and Post-event CMMAR = ∑ RP.t − RM ,t
t=−T
t=1
) where R
P,t
is the daily
return of an equally-weighted portfolio of firm-events for Day t, RM,t is the daily return of the equallyweighted CRSP index, for day t, and T = 20. The window of days [–2,+2] is the event window. Data from the Securities data Corporation’s (SDC) Mergers and Acquisitions database and consist of 12,476 acquisitions undertaken by 4,116 publicly listed US acquirers from 1/90 through 3/00 for which deal values are available. ‘Public target’ is as defined by SDC, and ‘Non-public target’ is a target that is not publicly traded. z-statistics in parentheses (*** significant at 1%; ** significant at 5%); the test statistic is calculated as {CDMAR ÷ [σCDMAR√D]} where D is the number of days. σCDMAR is calculated using the average excess returns over the market returns to the respective portfolios from Day –250 to Day –21.
Pre-event CDMAR
All Acquirers Acquirers of Public Targets with Stock Acquirers of Public Targets with Cash
Day [–20 to –3]
Day [+3 to +20]
0.55% (3.66)***
– 0.84% (–5.60)***
1.48% ( 3.97)*** –0.33% (–1.78)
– 1.09% (–2.91)*** – 0.82% (–4.47)***
1.81% ( 4.35)***
Difference: Stock v. Cash for Public Acquirers of Non-Public Targets with Stock Acquirers of Non-Public Targets with Cash
Post-event CDMAR
1.60% ( 3.42)*** –0.80% (–1.63) 2.40% ( 3.54)***
Difference: Stock v. Cash for Non-public
46
– 0.27% (–0.65) – 0.46% (–0.99) – 0.49% (–1.00) 0.03% ( 0.04)
Figure 3 Pre-event and Post-event Cumulative Daily Market-Adjusted Announcement Period Returns (CDMAR) Pre-event CMMAR =
−1
T
t=−T
t=1
∑ (RP.t − RM ,t ) and Post-event CMMAR = ∑ (RP.t − RM ,t ) where RP,t is the daily
return of an equally-weighted portfolio of firm-events for Day t, RM,t is the daily return of the equallyweighted CRSP index, for day t, and T = 20. The window of days [–2,+2] is the event window. Data from the Securities data Corporation’s (SDC) Mergers and Acquisitions database and consist of 12,476 acquisitions undertaken by 4,116 publicly listed US acquirers from 1/90 through 3/00 for which deal values are available. ‘Public target’ is as defined by SDC, and ‘Non-public target’ is a target that is not publicly traded. A: All Events
3%
2%
1%
0%
-1%
-2%
-3% -20
-15
-10
-5
0
5
10
15
20
Event Day
B: Indicated Events
5%
4% Non-Public Stock 3%
2%
Public Stock 1%
Non-Public Cash 0%
-1% Public Cash
-2% -20
-15
-10
-5
0 Event Day
47
5
10
15
20
Table 11 Acquirer Announcement Effects for All Acquisitions, Public Target Acquisitions, and Non-public Target Acquisitions by Year of Acquisition: 1990 – March 2000 Acquisitions data are from the Securities data Corporation’s (SDC) Mergers and Acquisitions database and consist of 12,476 acquisitions undertaken by 4,116 publicly listed US acquirers from 1/90 through 3/00 for which deal values are available. Announcement effects are measured by the Cumulative Daily Market Adjusted Returns (CDMARs) during the five days surrounding the day of the announcement (Day –2 through Day 2). Abnormal returns are measured by the difference between the return to the security and the return to the CRSP equal-weighted index. ‘Public target’ and ‘non-public target’ are as defined by SDC. tstatistics in parentheses (*** significant at 1%; ** significant at 5%). Year
All
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 20001
1.17% 1.40% 1.73% 1.25% 1.32% 1.25% 1.85% 1.54% 1.05% 1.63% 2.72%
All Years
1.45% (17.27)***
1
Public Targets (2.58)*** (3.08)*** (5.12)*** (4.41)*** (5.34)*** (5.04)*** (8.69)*** (8.84)*** (5.08)*** (5.17)*** (3.01)***
0.18% –1.12% –0.60% –1.50% –0.03% –1.30% –0.05% –0.50% –0.90% –1.20% –0.60%
Non-public Targets
( 0.16) (–1.33) (–0.96) (–2.94)*** (–0.03) (–3.42)*** (–0.12) (–1.31) (–1.90)** (–2.11)*** (–0.36)
–0.71% (–3.89)***
Data only through March.
48
1.35% 2.01% 2.20% 1.77% 1.61% 1.97% 2.28% 1.94% 1.47% 2.38% 3.32%
(2.65)*** (2.76)*** (5.75)*** (5.56)*** (7.93)*** (6.67)*** (9.33)*** (10.05)*** (6.44)*** (6.56)*** (3.25)***
1.95% (20.66)***
Difference (t-Statistic) 1.17% 3.13% 2.80% 3.26% 1.58% 3.28% 2.33% 2.44% 2.36% 3.61% 3.91%
(1.02) (2.76)*** (3.11)*** (4.22)*** (2.50)*** (5.51)*** (4.31)*** (5.26)*** (4.40)*** (4.71)*** (1.56)
2.66% (12.31)***
Table 12 Regression Analysis of Acquirer Announcement Effects for Day [–2,+2] for All Acquisitions, Public-Target Acquisitions and Non-public-Target Acquisitions The dependent variable is the Day [–2,+2] announcement effect. The independent variables are the natural logarithm of acquirer market value (SIZE), a dummy equal to one (and zero otherwise) if the target is publicly traded (PUBLIC), the relative size of target as measured by the deal value divided by acquirer market value prior to the announcement (RST), a dummy equal to one (and zero otherwise) if the consideration is all stock, excess returns to acquirer stock during the year prior to the acquisition month (12M RUNUP), and an interaction term RST*STOCK. Acquisitions data are from the Securities data Corporation’s (SDC) Mergers and Acquisitions database and consist of 12,476 acquisitions undertaken by 4,116 publicly listed US acquirers from 1/90 through 3/00 for which deal values are available. Data on mode of payment, and “public target” and “non-public target” are from the SDC database. Announcement effects are measured by the Cumulative Daily Market Adjusted Returns (CDMARs) during the five days surrounding the day of the announcement (Day –2 through Day +2). Abnormal returns are measured by the difference between the return to the security and the return to the CRSP equal-weighted index. t-statistics in parentheses (*** significant at 1%; ** significant at 5%).
Regression Constant SIZE PUBLIC RST STOCK RST*STOCK 12M RUNUP Adjusted R2 F-statistic
(1)
All Targets (2) (3)
(4)
Public Targets (5) (6)
(7)
(8)
Non-public Targets (9) (10)
(11)
0.0449 0.0265 0.0320 0.0423 0.0403 0.0346 0.0421 0.0169 0.0173 0.0175 0.0221 (15.51)*** (8.21)*** (9.55)*** (5.13)*** (4.84)*** (4.30)*** (5.03)*** (4.66)*** (4.77)*** (4.84)*** (5.87)*** –0.0043 –0.0024 –0.0036 –0.0046 –0.0047 –0.0047 – 0.0042 –0.0016 –0.0015 –0.0015 –0.0029 (4.61)*** (4.64)*** (4.65)*** (4.14)*** (2.89)*** (2.69)*** (2.63)*** (5.12)*** (9.25)*** (4.94)*** (7.29)*** –0.0212 –0.0282 –0.0233 (9.50)*** (11.63)*** (9.71)*** 0.0442 0.0406 –0.0280 –0.0208 –0.0132 –0.0311 0.0695 0.0628 0.0621 0.0672 (12.08)*** (10.94)*** (5.16)*** (2.92)*** (2.04)** (5.66)*** (17.45)*** (14.84)*** (14.89)*** (16.59)*** 0.0059 0.0032 –0.0167 –0.0120 –0.0126 0.0085 0.0022 0.0036 (4.56)*** (2.55)*** (3.40)*** (3.54)*** (0.80) (1.46) (2.51)*** (1.34) –0.0277 –0.0306 –0.0161 –0.0325 0.0518 0.0561 (3.94)*** (4.35)*** (1.58) (4.09)*** (4.69)*** (5.82)*** 0.0064 –0.0232 0.0119 (5.61)*** (6.43)*** (3.74)*** 0.02 118.35
0.03 76.40
0.04 64.84
0.02 18.51
0.03 14.51
49
0.02 17.14
0.04 21.93
0.04 130.18
0.04 103.34
0.04 137.58
0.05 108.00
Table 13 Acquirer Announcement Effects for Public Target Acquisitions and Non-public Target Acquisitions, and Mode of Payment (Cash Only or Stock Only) for Firms Acquiring Both Public and Non-public Targets, Firms Acquiring Only Multiple Public or Only Multiple Non-public Targets, and Firms Acquiring Only One Public or Only One Non-public Target: 1990 – March 2000 Acquisitions data are from the Securities Data Corporation’s (SDC) Mergers and Acquisitions database and consist of 12,476 acquisitions undertaken by 4,116 publicly listed US acquirers from 1/90 through 3/00 for which deal values are available. Announcement effects are measured by the Cumulative Daily Market Adjusted Returns (CDMARs) during the five days surrounding the day of the announcement (Day –2 through Day 2). Abnormal returns are measured by the difference between the return to the security and the return to the CRSP equal-weighted index. ‘Public target’ and ‘non-public target’ are as defined by SDC. tstatistics in parentheses (*** significant at 1%; ** significant at 5%). Panel A: Firms Acquiring Both Public and Non-public Targets Non-public Targets Public Targets
Difference
(Public minus Non-public)
Announcement Effect: Cash Only Number of Observations
0.92% 399 (2.56)***
1.10% 1708 (6.88)***
–0.18%
Announcement Effect: Stock Only Number of Observations
–1.71% 968 (6.58)***
1.69% 814 (5.28)***
–3.40%
Difference (Cash minus Stock)
2.63% (5.92)***
–0.59% (1.65)
(0.46)
(8.25)***
Panel B: Firms Acquiring Only Multiple Public or Only Multiple Non-public Targets Non-public Targets Difference Public Targets (Public minus Non-public)
Announcement Effect: Cash Only Number of Observations
2.57% 30 (1.95)*
1.32% 2414 (7.76)***
Announcement Effect: Stock Only Number of Observations
–1.90% 55 (2.11)**
3.30% 778 (6.73)***
Difference (Cash minus Stock)
4.47% (2.80)***
–1.98% (3.82)***
1.25% (0.94) –5.20% (5.07)***
Panel C: Firms Acquiring Only One Public or Only One Non-public Target Non-public Target Difference Public Targets (Public minus Non-public)
Announcement Effect: Cash Only Number of Observations Announcement Effect: Stock Only N Number of Observations Difference (Cash minus Stock)
1.43% 63 (2.20)***
1.18% 469 (3.03)***
0.11% 63 (0.17)
1.61% 263 (2.21)***
1.32% (1.44)
–0.43% (0.52)
50
0.25% (0.33) –1.50% (1.53)
Table 14 Categorical Analysis of Acquirer Announcement Effects for Day [–2,+2] for Acquisitions of All, Public, and Non-Public Targets by Size of Targets and All Stock v. All Cash v. Mixed Mode of Payment Acquisitions data are from the Securities data Corporation’s (SDC) Mergers and Acquisitions database and consist of 12,476 acquisitions undertaken by 4,116 publicly listed US acquirers from 1/90 through 3/00 for which deal values are available. Data on mode of payment are from the SDC database. All Stock and All Cash refer to acquisitions where 100% of the consideration was in the form of stock and cash, respectively; Mixed refers to the sum of acquisitions in SDC categories ‘Some Stock,’ ‘Some Cash,’ and ‘Combination of Stock and Cash.’ Announcement effects are measured by the Cumulative Daily Market Adjusted Returns (CDMARs) during the five days surrounding the day of the announcement (Day –2 through Day +2). Abnormal returns are measured by the difference between the return to the security and the return to the CRSP equal-weighted index. ‘Public target’ and ‘non-public target’ are as defined by SDC. t-statistics in parentheses (*** significant at 1%; ** significant at 5%). Relative Size Quintile (RS)
Range Cash
Smallest Quintile
All Targets Mixed Stock
Announcement Effects for Day [–2,+2] Public Targets Non-public Targets Cash Mixed Stock Cash Mixed Stock
RS ≤ 2%
CMAR t-stat N
0.29% 0.14% (2.73)*** (0.63) 1542 596
0.39% (1.96)** 653
0.56% (2.00)** 148
–0.24% (0.49) 41
0.12% (0.48) 136
0.26% (2.36)*** 1391
0.16% (0.70) 555
0.46% (1.92)* 513
2nd Quintile
2% < RS ≤ 4%
CMAR t-stat N
0.31% 0.46% (2.01)*** (1.72) 826 474
0.70% (2.37)*** 480
0.27% (0.77) 50
0.39% (1.11) 41
0.37% (0.88) 130
0.31% (1.94)** 771
0.47% (1.62) 432
0.82% (2.16)*** 348
3rd Quintile
4% < RS ≤ 10%
CMAR t-stat N
0.63% 0.60 0.42% (4,50)*** (3.11)*** (1.80)* 1130 991 684
0.13% (0.38) 103
–1.07% (1.71) 94
–0.34% (1.10) 218
0.68% (4.53)*** 1025
0.76% (3.62)*** 893
4th Quintile
10% < RS ≤ 25%
CMAR t-stat N
0.87% 1.44% 0.33% (4.53)*** (6.79)*** (1.09) 949 954 562
1.15% (2.40)** 102
0.32% (0.78) 136
–1.36% (3.78)*** 250
0.74% (4.11)*** 841
1.58% (6.87)*** 816
1.46% (3.84)*** 310
RS > 25%
CMAR t-stat N
2.40% 1.99% 0.03% (8.20)*** (8.00)*** (0.07) 645 1188 632
2.01% (2.72)** 90
–0.82% –2.74% (2.25)*** (7.61)*** 325 415
2.43% (7.59)*** 555
3.10% (9.99)*** 864
5.67% (5.85)*** 218
Largest Quintile
51
0.78% (2.52)*** 465
Table 15 Categorical Analysis of Acquirer Market Value Changes for Day [–2,+2] for Acquisitions of All, Public, and Non-Public Targets by Size of Targets and All Stock v. All Cash v. Mixed Mode of Payment Acquisitions data are from the Securities data Corporation’s (SDC) Mergers and Acquisitions database and consist of 12,476 acquisitions undertaken by 4,116 publicly listed US acquirers from 1/90 through 3/00. Data on mode of payment are from the SDC database. All Stock and All Cash refer to acquisitions where 100% of the consideration was in the form of stock and cash, respectively; Mixed refers to the sum of acquisitions in SDC categories ‘Some Stock,’ ‘Some Cash,’ and ‘Combination of Stock and Cash.’ ∆AMV is the dollar change (in US$ millions) in excess acquirer market value calculated as AMV* CDMAR, where AMV is the acquirer market value prior to the announcement and CDMAR is the Cumulative Daily Market Adjusted Returns during the five days surrounding the day of the announcement (Day –2 through Day +2). Abnormal returns are measured by the difference between the return to the security and the return to the CRSP equal-weighted index. ∆AMV/ΣAMV is the change in AMV as a percent of ΣAMV, the sum of AMVs (in US$ millions) of all firm-events. ‘Public target’ and ‘non-public target’ are as defined by SDC. Relative Size Quintile (RS)
Range Cash
Smallest Quintile
RS ≤ 2%
∆AMV ∆AMV/ΣAMV N 2nd Quintile 2% < RS ≤ 4% ∆AMV ∆AMV/ΣAMV N 4% < RS ≤ 10% ∆AMV 3rd Quintile ∆AMV/ΣAMV N 4th Quintile 10% < RS ≤ 25% ∆AMV ∆AMV/ΣAMV N Largest Quintile RS > 25% ∆AMV ∆AMV/ΣAMV N All Quintiles ∆AMV ∆AMV/ΣAMV All Pmt Modes and All Quintiles ∆AMV ∆AMV/ΣAMV
67875 0.56% 1542 8750 0.38% 826 19136 0.82% 1130 1562 0.21% 949 –39398 –2.74% 645 57745 0.31%
All Targets Mixed Stock 20660 0.20% 596 –2175 –0.21% 474 13285 1.67% 991 12222 2.79% 954 1390 0.73% 1188 45382 0.35% 110625 0.28%
52
Acquirer Market Value Changes Public Targets Cash Mixed Stock
8026 12241 0.22% 0.35% 653 148 8297 4787 0.93% 0.47% 480 50 –2242 –9602 –0.20% –0.92% 684 103 11434 –1355 0.96% –1.31% 562 102 –18017 –45886 –1.40% –3.67% 632 90 7498 –39795 0.09% –0.57%
– 27307 –0.69% 41 –14300 –2.82% 41 163 0.05% 94 1374 0.55% 136 –942 –1.05% 325 –41012 –0.80% –111586 –0.71%
9742 0.91% 136 –1378 –0.42% 130 –6240 –1.20% 218 582 0.08% 250 –33845 –3.58% 415 –30779 –0.87%
Non-public Targets Cash Mixed Stock 55634 0.65% 1391 3783 0.31% 771 28738 2.23% 1025 2897 0.45% 841 6488 3.41% 555 97540 0.82%
47967 0.72% 555 12125 2.20% 432 13122 2.90% 893 10848 5.71% 816 2332 2.33% 864 86394 1.09% 222211 0.91%
–1716 –0.07% 513 9675 1.72% 348 3998 0.67% 465 10852 2.15% 310 15468 4.35% 218 38277 0.83%
