University of Groningen
Faculty of Business and Economics
The impact of stock market valuation on
mergers and acquisitions of European firms:
an empirical analysis
Student: Stefan Kamp
Studentnumber: 1459279
Supervisor: dr. A. Plantinga
ABSTRACT
This study explores the influence of the valuation of the stock market on the acquisition return performance of acquirers in Europe. We show that
acquisitions made in low-valuation periods significantly outperform the
acquisitions made in high-valuation periods. High-valuation acquisitions earn in the long-run significant negative abnormal returns, while low-valuation
acquisitions earn a positive return, although not significant. The results are obtained using a methodology in calculating long-run buy-and-hold abnormal returns that is controlled for the survivorship bias, rebalancing bias and skewness bias. By subdividing the sample in acquirer types on the basis of market-to-book values, we find that the negative buy-and-hold returns in the high-valuation period are predominately caused by the high market-to-book firms, or equivalently the ‘glamour firms’. This result is in line with the
1 Introduction
The profitability of mergers and acquisitions (M&A) has been extensively researched in the economic literature. Bruner (2001) summarizes the findings and research approaches of 130 M&A studies from 1971 to 2001. The author concludes that in common, target shareholders earn a positive market return, whereas bidders make on average zero return and targets and bidders combined make a positive return. In addition, academic research shows that tender offers opposed to mergers earn a higher post-acquisition return. For example,
Loughran and Vijh (1997) show that tender offers, which are more hostile, could generate greater wealth for the bidder by addressing the targets shareholders directly. Furthermore, with regard to the method of payment, cash offers earn a higher post-acquisition return than stock offers. The higher post-acquisition return for cash deals, can be explained by the fact that the option to pay in stock signals that the bidders` shares are overvalued (Yook (2003)).
Several studies have found that mergers and acquisitions occur in waves and that they are clustered by industry. Studies as Mitchell and Mulherin (1996) and Andrade et al (2001) try to explain the merger waves. They show that deregulation and innovation give rise to higher merger and acquisition activity. Rhodes-Kropf and Viswanathan (2004), and Dong, Hirshleifer, Richardson, and Teoh (2006) however, argue that the merger and acquisition waves are not fully explained by deregulation and innovation. They believe that the waves are
Recently, more economic literature researched the influence of stock price valuation on merger and acquisition activity. The common finding is that high market valuations tend to correlate with high merger and acquisition activity. For example, Jovanovic and Rousseau (2001) show that merger waves coincide with periods of high P/E ratios which occur during stock-market booms. Bouwman et al (2003) assume that not only the firm-specific valuation but the overall market valuation influences the managers’ acquisition decision and therefore merger performance. They find that in periods of market decline the firms making acquisitions outperform acquisitions made in periods of market booming.
In the behavioural finance literature several arguments can be found that may support the influence of firm-specific valuation and market valuation on the M&A performance. One argument is overconfidence behaviour in decision-making caused by the illusion of knowledge and the illusion of control (Nofsinger 2008), also called hubris. The effect of hubris is more pervasive on the level of the management because of the higher commitment the managers face due to the fact that their wealth, professional reputation and employability are at stake (Gervais and Odean 2001). This managerial irrationality is hard to overcome because arbitrage is difficult due to high transaction costs, idiosyncratic risk and the absence of learning possibilities since managerial decisions are often
infrequent, have noisier feedback and have longer delayed outcomes (Nisbett and Ross 1980). Managers of firms with high market-to-book values are more likely to experience high past returns and high past growth and are therefore more likely to be influenced by hubris driven decision-making. A second
managers to enhance their reputation in the labor-market. They state that the degree of herding behaviour depends on the outside opportunities and the compensation scheme of the manager. They also describe the so called sharing-the-blame effect in which managers are more likely to take a decision which is taken before by other managers and is perceived as smart in the labor market.
This paper extends the work of Bouwman et al (2003), which will be further discussed in the literature section. We add evidence on the return performance of European bidders with regard to the valuation of the stock market. If the valuation of the stock market has an influence on the acquisition decision and price reaction in the market we expect the acquisitions in the low-valuation periods to outperform the acquisitions made in the high-low-valuation period. We make this assumption on the basis that acquisitions in low-valuation periods are less subject to hubris, market sentiment and herding behaviour. The hypothesis to be tested is whether acquisitions made in periods of low-market valuation outperform the acquisitions made in periods of high-market valuation. Thereby, exploring the short-run as well as the long-run return performance of acquirers with regard to the valuation of the stock market and the method of payment. Furthermore we try to find to what extent the behavioural arguments explain the observed results by exploring the overconfidence argument. We study to what extent the data is conform the extrapolation hypothesis. The extrapolation hypothesis states that managers and the market are subject to hubris or overconfidence, regarding the firms that have experienced high past performance (Rau and Vermaelen (1998). Lakonishok et al (1994) find that firms with high market-to-book firms, also called ‘glamour firms’ are especially firms that have experienced high past stock returns and growth in cash flow and earnings. If the data is in line with the hypothesis we expect to find more
acquisitions of ‘glamour firms’ in the high-market valuation period because of the high past performance. Further, we expect, due to the market overreaction to the high past performance of the bidder firm, that acquisitions of ‘glamour firms’ underperform the acquisitions of low market-to-book firms also called ‘value firms’. Specifically, regarding the valuation of the –stock market, we expect that the underperformance of ‘glamour firms’ contribute to the hypothetical
The rest of the paper is structured as follows: Section 2 shows the related literature to our research; The data and methodology will be outlined in
respectively section 3 and 4. The results are presented in section 5 and section 6 gives some concluding remarks.
2 Literature
In this section we provide an overview of the findings of related literature. Because we will explore the short-run and the long-run return performance of European bidders as well as the extrapolation hypothesis, a distinction between these three research topics is maintained in the structure throughout this paper.
2.1 Short-run and long-run return performance
Bruner (2001) concludes that the findings on returns to acquirers at the announcement are mixed. One third shows value destruction, one-third shows value conservation and one-third shows value creation. Of the eleven studies summarized by Bruner (2001) eight find significant negative returns for acquirers in the post-acquisition performance. Gregory (1997) for example finds significant negative returns for UK acquirers in the two-year post-acquisition performance even when controlling for size effects in the benchmark.
Bouwman et al (2003) test on the basis of theoretical models and
empirical evidence if the stock market influences managerial acquisition making and therefore the announcement effects and long-run performance of the
acquiring firms in the US. They find significant positive returns in the
underperform. These findings are robust to differences in method of payment and the classification method used to classify the valuation of the stock market.
Travlos (1987) studies the return relationship between the different
method of payment options for bidding firms when announcing an acquisition. He finds that stakeholders of bidder firms in stock deals experience significant losses at the announcement of an take-over bid, while cash deal take-over
announcements earn normal returns. These findings are consistent with the information signalling hypothesis which states that a stock offer contains a negative signal that the stock of the bidder firm is overvalued. Furthermore, Huang and Walking (1987) find that cash deals earn significantly higher returns even when controlling for type of acquisitions and degree of resistance. Wansley et al. (1983) and Yook (2003) find similar results. Also, in the post acquisition Gregory (1997) finds that equity offers earn significant negative returns, while cash offers earn returns that are insignificantly different from zero.
2.2 Extrapolation hypothesis
The empirical literature on specific CEO overconfidence is limited and often focused on the US. Malmendier and Tate (2005) look at CEO`s personal portfolio transactions and perception of outsiders to measure overconfidence. They find that CEO`s are sensitive to cash flow especially when internal resources are scarce. The theory behind this is that overconfident managers overinvest when there are enough internal funds and are reluctant to invest when they need external funds. The empirical study of Brown and Sarma (2007) explores the influence of CEO overconfidence and CEO dominance on the acquisition decision. They find that overconfident CEO`s are more likely to make an acquisition.
Rau and Vermaelen (1998) find that in contrast to tender offers, mergers underperform in the long-run horizon. Furthermore, they show that this
results are consistent with the performance extrapolation hypothesis which states that managers of high market-to-book firms are subject to hubris and therefore over extrapolate the past performance of the bidder firm. Sudarsanam and Mahate (2003) also study the performance of different acquirer types with regard to their market-to-book value. They find that in the announcement period the stock market does not extrapolate the pre-bid return performance of
acquirers. In the three-year post-acquistion period the ‘value acquirers’ outperform the ‘ glamour acquirers’ caused by a revised reaction in the stock market. Also they find evidence that ‘glamour firms’ are more likely to make stock offers than equity offers. This finding suggests that the management is aware of the overvaluation of the firms equity and uses the overvalued stock to make beneficial deals.
3 Data
3.1 Sample description
M&A deal data is obtained from the databank Zephyr, a database of mergers and acquisitions. There are several selection criteria chosen to obtain a representative and sufficient sample:
-The type of deals are either mergers or acquisitions
-The deals are announced and completed before 31/12/2004
-The acquirer is a quoted company on an index in Europe
-The deal value is more than 40 million euro
-The acquirer acquires at least 50% of the shares of the target
-The method of payment is either in stock or cash or a mix of both.
The period under study covers a period from January 1997 to December 2004. The lower limit, January 1997, is chosen due to the availability of recorded events on Zephyr and the upper limit, December 2004, is taken as an upper line to be able to test the long-run performance of the acquisitions with three years of stock data beyond the date of the deal announcement.
The financial data is obtained from Datastream, a financial statistical database from Thompson. The daily stock prices as well as the monthly stock prices, the market equity and the market-to-book value data of the acquiring firms in the sample are recorded. Due to the lack of some financial data the sample is further constricted. Taken all the selection criteria in account, this study derives at a sample of 726 events.
3.2 Calculation of the valuation of the stock market
In order to relate the return performance of the M&A firms with the valuation of the stock market, we use the price/earnings (P/E) ratio of the European market, provided by Datastream, to make a classification. The Datastream market index of Europe constituents 2099 European securities. In the same way as Shiller (2005), we estimate the monthly actual P/E ratio of the European equally weighted index. Shiller (2005) calculated the S&P 500 actual P/E ratio in the period 1871-2006 using the averaged inflation-adjusted earnings per share, calculated over a period of ten year1. In adjusting the P/E ratio to real values, this study uses the consumer price index OECD - Europe (excluding high inflation countries)2. Using the actual P/E ratio of the European Market index, every month in the sample period is classified into low-, neutral- and high-valuation months regarding their P/E ratio. We follow the same procedure outlined in Bouwman et al (2003) to make this classification:
First, we detrend the monthly P/E ratio, to overcome the problem of qualifying too much recent months as high-valuation months caused by the upward trend of the European Market index. To make the P/E ratio stationary,
1
The data of Shiller is found on the authors website. Data file Shiller (2005): http://www.econ.yale.edu/~shiller/data.htm
2
we subtract the best straight line fit of the P/E ratio of the month in question and the five preceding years. Second, we qualify the detrended months as above (below) average if the P/E value is above (below) the average of the five preceding years. Third, the top half of the above average group is classified as high-valuation months and the bottom half of the below average group is
classified as low-valuation months. The rest of the months in the sample period are classified as neutral-valuation months.
The period of 96 months under study has, according to this classification, 26 months classified as high-valuation months, 23 months classified as low-valuation months and 47 months classified as neutral-low-valuation months (see figure 1).
In terms of the valuation of the stock-market, the 726 acquisitions in the sample are distributed as follows: 197 acquisitions in high-valuation periods; 327 acquisitions in neutral-valuation periods; and 202 acquisitions in low-valuation periods. An overview of the data sample can be found in table I. The data is partly similar with the notions made in the M&A literature. Acquisition frequency is not higher in high-valuation periods compared to low-valuation periods (197 high-valuation acquisitions against 202 low-valuation acquisitions). Though more value is spend on acquisitions during the high-valuation period (42% of total deal value on high-valuation acquisitions against 21% of total deal value on
low-valuation acquisitions). Furthermore, in high-low-valuation periods the method of payment in stock is more present (17%) than in low-valuation periods (10%).
4 Methodology
4.1.1 Calculation of short-run return performance M&A
We measure the abnormal returns in the announcement period using the market adjusted return model described by Brown and Warner (1985) with a three-day window (from one day prior to the event day up to one day after the event day). We calculate daily abnormal returns in the event window by
(
)
+1 mt it -1 1, 1 = R - R CAR − + ∑
(1)where
R
it is firm i’s daily stock return on date t andR
mt is the return for the equally-weighted Datastream European Market index on date t.Figure 1: Classification value state of market
0 5 10 15 20 25 30 35 1997,01 1998,01 1999,01 2000,01 2001,01 2002,01 2003,01 2004,01
Actual P/E ratio Straight-line fit
Table I
Descriptive statistics
This table reports on the classification of the M&A events into the valuation of the stock market (High-Valuation, Neutral-Valuation and Low-Valuation) and the method of payment (Cash, Stock and Mixed Payment). Furthermore the deal characteristics, deal value, and acquirer charesteristic, market value (price times shares outstanding) are presented. All data is obtained from Zephyr.
Number of acquisitions % Number of acquisitions Total Deal value (mln euro) % of total deal value Mean deal value (mln euro) Median Deal value (mln euro) Mean Market Value (mln euro) Median Market value (mln euro) All acquisitions 726 100% 380.957 100% 542 147 20.863 3.179 High-Valuation Acquisitions 197 27% 160.710 42% 816 231 24.843 2.634 Neutral-Valuation Acquisitions 327 45% 140.196 37% 448 136 20.932 3.742 Low-Valuation Acquisitions 202 28% 80.051 21% 415 125 16.849 2.534 Cash Acquisitions 530 73% 225.532 59% 440 151 24.145 4.523 Stock Acquisitions 97 13% 83.185 22% 867 147 11.840 2.300
Mixed Payment Acquisitions 99 14% 72.240 19% 760 126 12.044 973
High-Valuation
Cash Acquisitions 131 66% 82.993 52% 634 204 27.466 4.276
Stock Acquisitions 33 17% 37.465 23% 1135 245 14.368 2.686
Mixed Payment Acquisitions 33 17% 40.251 50% 1220 271 24.909 1.315
Neutral-Valuation
Cash Acquisitions 246 75% 103.036 73% 438 142 24.956 4.828
Stock Acquisitions 43 13% 20.288 14% 548 146 13.632 2.409
Mixed Payment Acquisitions 38 12% 16.871 12% 411 100 4.359 1.239
Low-Valuation
Cash Acquisitions 153 76% 39.502 49% 271 130 19.998 4.009
Stock Acquisitions 21 10% 25.432 32% 978 138 6.227 329
4.1.2 Calculation of long-run return performance M&A
Barber and Lyon (1997a) advocate to use buy-and-hold returns over cumulative abnormal returns when calculating the long-run stock performance. They report on three causes of misspecification which are more pervasive when cumulative abnormal returns are used. Firstly, the authors report on a
misspecification called the new-listing bias or survivor bias. When new firms are included in the reference portfolio subsequent to the formation period, a positive bias can originate caused by the fact that newly listed firms that are going public generally underperform the market resulting in a deflated long-horizon return (Ritter 1991). Secondly, the rebalancing bias is mentioned. This bias can come into existence when, to maintain equal weights, the reference portfolio is rebalanced periodically which can cause a negative bias due to an inflated long-horizon return. Thirdly there is the so called skewness bias which tells that the distribution of the long-run abnormal returns are positively skewed resulting in a negatively biased t-statistic.
To calculate the long-horizon return on the reference portfolios, we compound the returns on securities within the reference portfolio and then sum these returns across securities:
(
)
s s+ n it t=s i=11 + R
- 1
=
p sR
n
τ ∏
∑
(2)where
R
it is the return on security i in month t, s is the beginning period for the return calculation,τ
is the period of investment in months, and ns is thenumber of securities traded in month s. This method calculates a passive equally weighted investment return or a buy-and-hold return on all securities in the reference portfolio, and is not subject to newly listed firms nor monthly
rebalancing. The problem of the buy-and-hold method is the question what to do with the proceeds of investments of delisted firms subsequent to period s. The authors resolve this issue by assuming that the proceeds are reinvested in an equally weighted reference portfolio, which is rebalanced monthly. Missing monthly returns are filled in with the mean monthly return of firms comprising the reference portfolio in accordance with the size and market-to-book value of the security. The annualized returns of the reference portfolios can be found in table II.
Table II
Annualized returns of buy-and-hold reference portfolios
From the Europe market index provided by Datastream, 12 size/market-to-book portfolios are created. The table reports the annualized buy-and-hold returns over 1, 2 and 3 years of these portfolios for the period January 1997 to December 2004. In calculating the returns, equal investment in the constituent securities is assumed.
Size quartiles are created where quartile 1 (quartile 4) contains the smallest (largest) firms in size; and market-to-book tertiles are created where tertile a (tertile c) contains firms with the lowest (highest) market-to-book values.
1 year 2 years 3 years
Market-to-book Market-to-book Market-to-book
Size a b c a b c a b c
1 14,32% 13,05% 16,34% 17,11% 17,95% 19,01% 22,22% 23,77% 18,73%
2 11,09% 8,86% 3,24% 12,47% 8,23% 4,52% 14,88% 10,84% 5,38%
3 7,82% 5,29% 1,77% 9,07% 5,95% 2,06% 11,54% 8,09% 3,98%
The buy-and-hold abnormal returns are calculated using the compounded post-acquisition return of the acquirer and the buy-and-hold return of the
appropriate size/market-to-book reference portfolio:
(
)
(
)
∏
T it∏
T pt t=1 t=1 = 1 + R - 1 + R it BHAR (3)where
R
it as the month t simple return on a sample firm and Rpt is the month t reference portfolio return. To overcome the skewness bias the bootstrapped adjusted t-statistic is used. Johnson (1978) developed the skewness-adjusted t-statistic to deal with positively skewed distributions that cause a negatively biased t-statistic:γ
γ
2 1 1 = n + + 3 6n t S S (4) Where,σ
BHAR = (BHAR) S andγ
(
)
σ
∑
n i 3 i=1 3 BHAR - BHAR = ( ) n BHARwhere
γ
is an estimate of the coefficient of skewness and n S is the conventional t-statistic. Furthermore, significance is based on bootstrapped critical values. Sutton (1993) concludes that if the parent distribution is asymmetrical, a bootstrapped version of the skewness-adjusted t-statistic is preferred. The bootstrapped skewness-adjusted t-statistic reduces the type 1 error and is more powerful.4.2 Testing the extrapolation hypothesis
To test the extrapolation hypothesis we use the same data set and same categorization of the valuation of the stock market and method of payment as outlined in the data section to measure the return performance of the
described in the calculation of the long-run performance subsection, is used to calculate the buy-and-hold returns for the different acquirer types. In the period of January 1997 to December 2004, every year we rank the firms in the sample by their size and market-to-book values. We segment the sample into size
quartiles and within the size quartiles we further segment the sample into tertiles on the basis of the book values. The tertile with the lowest market-to-book values is categorized as ‘value acquirers’ (A), the middle tertile as ‘neutral acquirers’ (B) and the tertile with the highest market-to book values is
categorized as ‘glamour acquirers’. The BHAR and the corresponding skewness-adjusted t-statistics are calculated as in respectively, equation (3) and equation (4).
5 Results
5.1.1 Short-run M&A return performance
Table III shows the cumulative abnormal returns (CAR`s) at the
announcement of the M&A deals in the period January 1997 to December 2004. In panel A the results are segmented into the valuation of the stock market (High-Valuation, Neutral-Valuation and Low-Valuation) and the method of payment (Cash, Stock and Mixed Payment). All acquisitions earned a highly significant positive CAR of 0,36%. On the basis of valuation of the stock market, the results show that the positive CAR is driven by the highly significant positive CAR in high-valuation periods (CAR 0,76%) and neutral-valuation periods (CAR 0,35%).
acquisitions in valuation of the stock market with regard to the method of payment, this pattern is preserved. Though the CAR`s for cash deals in high- valuation periods and neutral-valuation periods are more positive than for stock deals.
Table III
Short-run cumulative abnormal returns (CAR)
This table reports on the short run cumulative abnormal returns of acquisitions made from January 1997 to December 2004. The results are further segmented into the classification of the M&A events as the valuation of the stock market (High-Valuation, Neutral-Valuation and Low-Valuation) and the method of payment (Cash, Stock and Mixed Payment). Furthermore, the table shows the CAR`s with the corresponding t-statistics.
Panel A: CAR segmented in categories
N CAR t All acquisitions 726 0,36% 4,85c High-Valuation Acquisitions 197 0,76% 4,53c Neutral-Valuation Acquisitions 327 0,35% 3,46c Low-Valuation Acquisitions 202 -0,01% -0,06 Cash Acquisitions 530 0,34% 4,43c Stock Acquisitions 99 0,61% 2,34b Mixed Payment Acquisitions 97 0,24% 0,92
High-Valuation
Cash Acquisitions 131 0,75% 3,82c Stock Acquisitions 33 0,61% 2,33b Mixed Payment Acquisitions 33 0,24% 0,92
Neutral-Valuation
Cash Acquisitions 246 0,32% 3,23b Stock Acquisitions 38 0,21% 0,54 Mixed Payment Acquisitions 43 0,64% 1,65
Low-Valuation
Cash Acquisitions 153 0,02% 0,13 Stock Acquisitions 28 0,21% 0,47 Mixed Payment Acquisitions 21 -0,48% -0,68
a
= significant at 10 percent level; b = significant at 5 percent level; c = significant at 1 percent level
Panel B: CAR mean differences
∆CAR t
High- minus Low-Valuation Acquisitions 0,77% 3,60c
Cash minus Stock Acquisitions -0,27% -1,27
High-Valuation Cash minus Low-Valuation Cash 0,73% 3,26b High-Valuation Stock minus Low-Valuation Stock 1,18% 1,75a
a
In panel B of table III the different CAR`s are tested on equality. We see that the observed CAR`s of the high-valuation periods and low-valuation periods differ significantly in the announcement period. High-valuation acquisitions earn a significant higher CAR than the acquisitions made in low-valuation periods. The CAR mean difference between the cash acquisitions and stock acquisitions is not significant which confirms the weak information asymmetry link between cash deals and stock deals in this sample. The difference of payment in either cash or stock between the high-valuation periods and low-valuation periods is significant. The price reaction on the announcement of cash acquisitions and stock
acquisitions is more positive in the high-valuation periods.
5.1.2 Long-run M&A return performance
Table IV reports on the post-acquisition buy-and-hold returns (BHAR) one year, two years and three years after the announcement in the period from January 1997 to December 2004. In panel A the results are segmented into the valuation of the stock market (High-Valuation, Neutral-Valuation and Low-Valuation) and the method of payment (Cash, Stock and Mixed Payment). The BHAR of all acquisitions combined are for all years negative although not significant. These negative returns are driven by the acquisitions made in the high-valuation periods which show negative BHAR`s significant at the 5 percent level during one year (BHAR1 -8,13%) and highly significant during two and three years (BHAR2 -14,88% and BHAR3 -11,71%). The low-valuation
acquisitions however, show positive BHAR although not significant. The BHAR`s in neutral-valuation periods lay in between the BHAR`s of the high-valuation periods and the low-valuation periods and are not significant. According to these findings, which are in line with the general hypotheses, acquisitions made in high-valuation periods destroy shareholder value, while acquisitions made in low-valuation periods create shareholder value.
The results segmented by method of payment show that cash offers earn a positive abnormal return while stock offers earn a negative abnormal return. This is, in contrast to the results in the announcement period, in line with the
look at the segmentation into valuation of the stock market, the distinction in abnormal returns between cash and stock offers is preserved except for the BHAR`s during two and three years in the high-valuation period. During these
Table IV
Buy-and-hold abnormal returns (BHAR)
This table reports on the 1 year, 2 year and 3 year post-acquisition buy-and-hold abnormal return of acquisitions made from January 1997 up to December 2004. The results are further segmented in the classification of the M&A events as the valuation of the stock market (High-Valuation, Neutral-Valuation and Low-Valuation) and the method of payment (Cash, Stock and Mixed Payment). Furthermore, the table shows the buy-and-hold returns with the corresponding skewness adjusted t-statistics. Significance is based on the bootstrapped critical values.
Panel A: BHAR segmented in categories
1 year 2 years 3 years
N BHAR1 t BHAR2 t BHAR3 t
All acquisitions 726 -1,99% -1,38 -2,69% -1,14 -2,84% -0,85 High-Valuation Acquisitions 197 -8,13% -2,24b -14,88% -4,32c -11,71% -3,48c Neutral-Valuation Acquisitions 327 -1,12% -0,56 0,08% 0,06 -3,09% -0,54 Low-Valuation Acquisitions 202 2,59% 1,15 4,73% 1,16 6,22% 0,92 Cash Acquisitions 530 2,38% 1,67a 5,06% 2,30b 6,71% 2,28b Stock Acquisitions 99 -17,13% -3,59c -23,07% -2,71c -24,08% -1,68a Mixed Payment Acquisitions 97 -10,45% -2,04b -24,22% -3,45c -33,31% -3,70c
High-Valuation
Cash Acquisitions 131 0,43% 0,14 -4,45% -1,14 -2,69% -0,68 Stock Acquisitions 33 -36,55% -4,88c -40,19% -5,05c -33,83% -4,16c Mixed Payment Acquisitions 33 -13,71% -1,36 -30,96% -2,18b -25,38% -2,38b
Neutral-Valuation
Cash Acquisitions 246 1,32% 0,67 5,22% 1,44 4,06% 0,81 Stock Acquisitions 38 -14,89% -2,09b -21,41% -1,77a -27,05% -1,69a Mixed Payment Acquisitions 43 -2,90% -0,39 -10,37% -0,96 -22,87% -2,12b
Low-Valuation
Cash Acquisitions 153 5,77% 2,89c 12,95% 3,67c 19,00% 3,51c Stock Acquisitions 28 2,73% 0,35 -5,16% -0,25 -8,56% -0,19 Mixed Payment Acquisitions 21 -20,76% -2,34b -42,01% -3,07c -67,17% -1,83a
a
= significant at 10 percent level; b = significant at 5 percent level; c = significant at 1 percent level
Panel B: BHAR mean differences
1 year 2 years 3 years
∆BHAR1 t ∆BHAR2 t ∆BHAR3 t
High- minus Low-Valuation Acquisitions -10,72% -2,67b -19,61% -3,68c -17,93% -2,26b
Cash minus Stock Acquisitions 19,51% 4,95c 27,87% 4,32c 30,79% 3,36c
High-Valuation Cash minus Low-Valuation Cash -5,34% 1,24 -17,40% -3,08c -21,69% -2,87c High-Valuation Stock minus Low-Valuation Stock -39,28% -3,75c -34,74% -2,10a -25,26% -0,84
a
years the BHAR`s for high-valuation cash offers are negative although not
significant (BHAR2 -4,45% and BHAR3 -2,69%). In common, the mixed payment BHAR`s lay in between the BHAR`s of cash and stock payment and show
negative abnormal returns (significant at 5 percent level in high-valuation periods during two and three years and in low-valuation periods during one and two years). Only for the mixed payment acquisitions made in neutral-valuation this finding does not hold.
In panel B of table IV the mean BHAR`s of the different segmented groups are tested on equality. The table reports that in all buy-and-hold periods the difference between high-valuation acquisitions and low-valuation acquisitions is significant. The difference in BHAR is significant at the 5 percent level after one year and three years and significant at the 1 percent level after two years. These results suggest that low-valuation acquisitions significantly outperform
acquisitions made in high-valuation periods. Also, when we look at the mean BHAR differences with regard to the method of payment, we find that cash deals significantly outperform stock deals in buy-and-hold periods. Looking at the mean difference of cash deals and stock deals between high-valuation periods and low-valuation periods, we see that the BHAR`s of cash deals differ
significantly at the 1-percent level after two and three years. For stock deals the BHAR`s differ only significantly at the 1-percent level after one year and at the 10-percent level after two years.
In sum, the data indicates that acquisitions made in low-valuation periods outperform the acquisitions made in high-valuation periods which is conform to our hypothesis. High-valuation acquisitions earn in the long-run significant negative abnormal returns, while low-valuation acquisitions earn a positive return although not significant.
5.2 Extrapolation hypothesis
the ‘value acquirers’ between the two valuation periods in panel A, we see that ‘glamour acquirers’ make more acquisitions compared to the ‘value acquirers’ (81 against 47). The higher acquisitiveness of the ‘glamour acquirers’ who are
associated with overconfident behaviour is in line with and Brown and Sarma (2007) who state that CEO overconfidence leads to higher acquisitiveness. Further the study of Malmendier and Tate (2005) finds that when internal funds are high, overconfident CEO`s will be more willing to make an acquisition. The
Table V
3-year buy-and-hold abnormal returns (BHAR) – overextrapolation hypothesis
This table reports on the 3 year post-acquisition buy-and-hold abnormal returns of acquisitions made from January 1997 to December 2004. The results are further segmented in the classification of the M&A events as the valuation of the stock market (High-Valuation, Neutral-Valuation and Low-Valuation), the method of payment (Cash, Stock and Mixed Payment) and acquirer type (A=’value acquirer’, B=’neutral acquirer’ and C=’glamour acquirer’. Furthermore, the table shows the buy-and-hold returns with the corresponding skewness adjusted t-statistics.
Panel A: BHAR segmented in categories
High-valuation acquisitions Low-valuation acquisitions n % of
Total BHAR t n
% of
Total BHAR t Total A 47 22,93% 2,64% 0,39 63 30,73% 18,02% 1,28 205 B 69 26,64% -20,99% -4,16c 81 31,27% -4,43% -0,53 259 C 81 30,92% -11,64% -2,11a 58 22,14% -3,58% -0,34 262
197 202 726
Cash acquisitions Stock acquisitions n % of
Total BHAR t n
% of
Total BHAR t Total A 144 70,24% 15,40% 2,11a 35 17,07% -10,65% -0,59 205 B 190 73,36% 0,49% 0,15 31 11,97% -26,58% -3,25c 259 C 196 74,81% 0,68% 0,19 33 12,60% -31,77% -2,10a 262
530 99 726
a
= significant at 10 percent level; b = significant at 5 percent level; c = significant at 1 percent level
Panel B: BHAR mean differences
∆BHAR t
High-Valuation
Glamour acquirer minus Value acquirer -14,28% 1,72 Low-Valuation
Glamour acquirer minus Value acquirer -21,60% 1,16
High-Valuation and Low-Valuation
Glamour acquirer minus Glamour acquirer -8,06% 0,75 Glamour acquirer minus Value acquirer 19,72% 1,98a
a
result that ‘glamour firms’ make more acquisitions in the high-valuation periods, when internal funds are high, compared to the low-valuation periods (30,92% against 22,14%) gives some support that ‘glamour firms’ are subject to
overconfidence.
Further we see that in the high-valuation periods the ‘glamour acquirers’ show a significant negative BHAR of -11,64% at the 10 percent significance level, whereas the ‘value acquirers’ show a positive BHAR of 2,46% although not
significant. For the low-valuation periods the difference between the acquirer types is still there although for all types with a higher BHAR. In Panel B we see that when taking both periods together the mean difference between the
‘glamour acquirers’ and ‘value acquirers’ is significant. Also, when we look at the method of payment, we see that ‘glamour firms’ perform worse than ‘value firms’ both in cash deals as well as in stock deals. The result that ‘glamour firms’ earn significant negative abnormal returns in the post-acquisition period independent of the method of payment is in line with Rau and Vermaelen (1998). In contrast to Sudarsanam and Mahate (2003), we find no indication that ‘glamour firms’ are more likely to make stock offers. In our sample, 12,60% of the acquisitions of ‘glamour firms’ is in stock while of the ‘value acquirers’ 17,07% of the
acquisitions is in stock.
To conclude, our data shows that ‘value acquirers’ significantly outperform the ‘glamour acquirers’ which is consistent with the extrapolation hypothesis. The ‘glamour acquirers’ earn significant negative returns of -11,64%. Further, there is an indication that ‘glamour firms’ make more acquisitions in high-valuation periods.
6 Conclusion
decision and market reaction on acquisitions. We explore to what extent this influence may result in different return performances between acquisitions made in high-valuation periods and acquisitions made in low-valuation periods.
In the announcement period, we find highly significant positive cumulative abnormal returns for high-valuation acquisitions, whereas the cumulative
abnormal returns for low-valuation acquisitions do not differ significantly from zero. The results in the post-acquisition return performance show that
acquisitions made in low-valuation periods outperform the acquisitions made in high-valuation periods. High-valuation acquisitions earn in the long-run
significant negative buy-and-hold abnormal returns, while low-valuation
acquisitions earn positive buy-and-hold returns although not significant. Conform to our hypothesis the market reassesses the acquisition quality and rewards acquisitions made in low-valuation acquisitions but punishes acquisitions made in high-valuation periods. The results segmented by method of payment are in line with the signaling hypothesis. We find that cash offers earn a positive buy-and-hold abnormal return while stock offers earn a negative buy-and-buy-and-hold abnormal return.
Furthermore, this paper tries to find the answer on the observed difference in return performance of acquirers between the stock market valuations by
testing the extrapolation hypothesis. The extrapolation hypothesis states that the market and the decision makers over extrapolate the past performance of a bidder firm caused by overconfident behaviour. The results indicate that the negative buy-and-hold returns in the high-valuation period are predominately caused by the high market-to-book firms also called ‘glamour firms’. Low
Literature:
Andrade, G., Mitchell, M., and E. Stafford, 2001, New evidence and
perspectives on mergers, Journal of Economic Perspectives, vol. 15, pp. 103-120 Barber, B. M., and J. D. Lyon, 1997a, Detecting long-run abnormal stock returns: The empirical power and specification of test statistics, Journal of Financial Economics, vol. 43, pp. 341–372.
Bouwman, C.H.S, Fuller, K., and Nain, A. S., 2003, The performance of stock-price driven acquisitions, Working paper
Brown, R., and N. Sarma, 2007, CEO overconfidence, CEO dominance and corporate acquisitions, Journal of Economics and Business, vol. 59, no. 5, pp. 358-379
Brown, S., and J. Warner, 1985, Using daily stock returns: The case of event studies, Journal of Financial Economics, vol. 14, pp. 3-31
Bruner, R.F., 2001, Does M&A Pay?, Applied Mergers & Acquisitions, Chapter 3
Dong, M., Hirshleifer, D., Richardson, S., and S.H. Teoh, 2006, Does investor misvaluation drive the takeover market?, Journal of Finance, American Finance Association, vol. 61(2), no. 4, pp. 725-762
Fama, E.F., and K.R. French, 1993, Common risk factors in the returns on stocks and bonds, Journal of Financial Economics, vol. 33, pp. 3-56.
Gervais, S., Odean, T., 2001, Learning to be overconfident, Review of Financial Studies, vol. 14, pp. 1‐27
Gregory, A., 1997, An examination of the long run performance of UK acquiring firms . Journal of Business. Finance & Accounting, vol. 24, No. 7&8, pp. 971-1002.
Huang, Y.S. and R.A. Walking, 1987, Target abnormal returns associated with acquisition announcement: payments, acquisition forms and managerial resistance, Journal of Financial Economics, vol. 19, pp. 329-50.
Johnson, N.J., 1978, Modified t tests and confidence intervals for
asymmetrical populations, Journal of the American Statistical Association, vol. 73, pp. 536–544.
Lakonishok, J., Shleifer, A., and R.W. Vishny, 1994, Contrarian
investment, extrapolation, and risk, Journal of Finance, vol. 49, pp. 1541-1578. Loughran, T., and A. Vijh, 1997, Do long-term shareholders benefit from corporate acquisitions? Journal of Finance, vol. 52, pp. 1765-1790
Lyon J.D., B.M Barber, and C. Tsai, 1999, Improved methods for tests of long-run abnormal stock returns, Journal of Finance, vol. 54(1), pp. 165-201.
Malmendier, U., and G. Tate, 2005, Does overconfidence affect corporate investment? CEO overconfidence measures revisited. European Financial
Management, vol. 11, pp. 649–659.
Mitchell, M.L., J.H. Mulherin, 1996, The impact of industry shocks on takeover and restructuring activity, Journal of Financial Economics, vol. 41, pp. 193-229
Nisbett, R., and L. Ross, 1980, Human Inference: strategies and shortcomings of social judgment, Prentice-Hall, Englewood Cliffs, NJ.
Nofsinger, J.R., 2005, Social mood and financial economics, The Journal of Behavioral Finance, vol. 6, no. 3, pp. 144, 150 & 151
Nofsinger, J.R., 2008, The psychology of investing, Pearson Prentice Hall, 3rd ed.
Rau, P.R., and T. Vermaelen, 1998, Glamour, value and the
post-acquisition performance of acquiring firms, Journal of Financial Economics, vol. 49, pp. 223-253.
Rhodes-Kropf, M. and S. Viswanathan, 2004, Market valuations and merger waves, Journal of Finance, vol. 59, no. 6, pp. 2685-2718
Ritter, Jay R., 1991, The long-run performance of initial public offerings, Journal of Finance, vol. 46, pp. 3–27.
Scharfstein, D.S. and J.C. Stein, 1990, Herd behavior and investment, American Economic Review, vol. 80-3, pp. 465-479.
Shiller, R., 2005, Irrational exuberance (2d ed.). Princeton University Press. ISBN 0-691-12335-7.
Shleifer, A., and R.W. Vishny, 2003, Stock market driven acquisitions, Journal of Financial Economics, vol. 70 pp. 295–311
Sudarsanam, P., and A. A. Mahate, 2003, Glamour acquirers, method of payment and post-acquisition performance: The UK evidence, Journal of
Sutton, C.D., 1993, Computer-intensive methods for tests about the mean of an asymmetrical distribution, Journal of the American Statistical Association, vol. 88, pp. 802–808.
Travlos, N.G., 1987, Corporate takeover bids, method of payment and bidding firm's stock returns, Journal of Finance, vol. 42, pp. 943-63.
Wansley, J.W., Lane, W.R., and H.C. Yang, 1983, Abnormal returns to acquired firms by type of acquisition and method of payment, Financial Management, vol. 11, pp. 16-22.
