Are Acquiring Firms in Western Europe Exposed to a Leverage
Effect in Acquisition Announcement Returns?
An Empirical Investigation
Master Thesis University of Groningen
Faculty of Economics & Business MSc Finance
Name: Marcel Buytendijk
Student Number: 1369687
Supervisor: Dr. H. Gonenc
Are Acquiring Firms in Western Europe Exposed to a Leverage
Effect in Acquisition Announcement Returns?
An Empirical Investigation
ABSTRACT
This paper analyzes whether acquiring firms are exposed to a leverage effect in acquisition announcement returns by examining a sample of 307 acquisitions in Western Europe between January 1997 and January 2008. Two theories arise in literature to explain the leverage effect: The contingent claim view and the monitoring debt view. The contingent claim view predicts a negative relationship between leverage and acquisition announcement returns, while the monitoring debt view predicts a positive relationship between leverage and acquisition announcement returns. This study argues that a thorough analysis to the leverage effect should take into account the diversification assumption of the contingent claim view. Therefore, the total sample is divided in groups depending on whether the acquirer and acquisition are focused or diversified. The outcome does not raise evidence for the contingent claim hypothesis. Diversifying acquisitions show no interaction with leverage on bidder returns. Weak evidence is found for the monitoring hypothesis as only diversified firms engaging in focused acquisitions show a significant positive association with leverage on bidder returns. The results suggest that firms should not determine capital structure in order to influence announcement returns.
JEL code: G14, G30, G34
Table of Contents
1. Introduction ... 4
2. Literature Review ... 8
2.1 Acquisitions ... 8
2.2 Towards a succesful acquisition... 9
2.3 Capital structure ... 11
2.4 Empirical evidence... 13
2.4.1 Announcement returns to bidders ... 13
2.4.2 Other measures ... 16
2.4.3 Literature overview... 18
2.5 Diversification... 19
3. Data, Expectations and Methodology... 22
1. Introduction
Financial leverage1 can be described as the extent to which a firm relies on debt financing rather than equity (Ross, Westerfield and Jaffe, 2005). As Modigliani and Miller (1958) propose the irrelevance of capital structure, why do firms take on debt? The rationale is that issuing debt provides advantages to equityholders2. The term ‘leverage’ implies magnifying an outcome. In a financial framework it allows greater potential returns on equity than otherwise would have been available as the firm takes on debt to invest and earn a higher level of return than the rate of interest on the loan. However, for corporate investments it appears that on average the announcement returns for the bidder are zero. The acquiring firm has to pay the expected synergies in the form of a premium to the target firm (Bruner, 2001). Therefore, it can be questioned: What really is the leverage effect? Two theories arise in literature when predicting the influence of leverage on announcement returns. The contingent claim view and the monitoring debt view. The contingent claim view predicts a negative relationship between leverage and acquisition announcement returns. The monitoring debt view predicts a positive relationship between leverage and acquisition announcement returns.
The contingent claim model is developed to explain why diversified firms trade at an equity discount. However, it can also be related to acquisition announcement returns. Contingent claim’s view is based on option pricing theory. According to Mansi and Reeb (2002), who describe the contingent claim framework in their paper, equity is a contingent claim3: A call option on the value of the firm exercised in states where the value of the assets is greater than the value of the debt claim. It is known that only a few factors determine the value of an option. Among these factors is the underlying volatility of the option: The value of the option increases when risk increases. When considering the cash flows of a firm as the underlying volatility, a decrease in equity value after an acquisition then might not indicate a reduction in firm value, but instead represents the risk effects of imperfectly correlated cash flows. When a firm acquires another firm it might actually be diversifying its operations and by means of cross pledging enhancing its cash flow stability and reducing risk. Trough collateral and coinsurance factors, a firm’s cash flow stability increases and the risk of a company declines, consequently the value of a share declines. Tirole (2006) describes this theory extensively. An acquisition is only diversifying when the operations of the acquired firm are not perfectly
1
From now on this paper uses the term leverage to describe financial leverage.
2
It is necessary to explicitly state that the debt-equity dichotomy as presented in this paper does not do justice to the richness of claims encountered in the world of corporate finance. For simplicity matters this paper only distinguishes between ordinary debt and ordinary shares, consistent with other related literature.
3
correlated with the operations of the acquiring firm. Of course the shareholders who own a call option on the firm do prefer to engage in perfectly correlated projects, as they get the upside potential and do not face the downward risks (Tirole, 2006). Therefore this research will make a distinction between acquisitions in related industries and non-related industries.
According to the contingent claim view the valuation effect to shareholders is negatively related with leverage. Mansi and Reeb (2002), describing the contingent claim framework in their paper, state that the value of the option is the maximum of the difference between the value of the assets and the value of debt on one hand and zero on the other hand. A lower level of debt is therefore associated with a deeper-in-the-money4 option and consequently a lower impact of volatility on the value of this call option. In other words, firms with low leverage ratios are not sensitive to decreases in risk as they are already deep-in-the-money. Contingent claim view therefore predicts that the negative acquisition announcement returns are stronger with higher leverage.
Summarized, in a contingent claim framework, diversifying acquisitions lead to lower firm risk. This lowers shareholder value and increases bondholder value. The value effects to shareholders depend on the amount of leverage in the bidding firm, with higher leverage leading to lower shareholder returns. This theory is confirmed in Mansi and Reeb (2002), who find that the average diversified firm has about 12 percent equity discount, and as leverage increases the wealth loss to shareholders becomes substantially higher.
A contrary perspective on the influence of capital structure on acquisition announcements is given by the view describing debt as a monitoring device, predicting a positive relationship between leverage and announcement returns. In this framework, Jensen (1986) presents the control hypothesis, theoretically outlining the role of debt in motivating managers and their organizations to be efficient. According to Jensen the control effects of debt are twofold. First, debt reduces the agency cost of free cash flow by constraining the amount available for spending at the discretion of managers. In this way resources under managements control are reduced, making it more likely they will incur the monitoring of the capital markets when the firm must obtain new capital to finance investments. Second, if managers do not maintain their promise to make the interest payments, the creditors are given the right to take the firm into bankruptcy court. Both effects prevent managers from investing in uneconomic projects. Stulz (1990) elaborates further on the work of Jensen and shows in his theoretical model that financing policies, by influencing the resources under management’s control, reduces the agency costs of managerial discretion. Debt issues reduce the cost of overinvestment by preventing management from pursuing their own objectives. According to Lang, Stulz and Walkling (1991) debt serves a disciplining function and higher leverage is expected to result in higher
4
announcement returns as management is more closely monitored and has less cash flow to spend, making bad acquisitions less likely.
Summarized, debt provides a signal that the acquisition has a high probability to be value-creating due to monitoring effects. This will consequently be rewarded by the capital market, leading to positive announcement returns. This view is confirmed by Maloney, McKormick and Mitchell (1993), who empirically examine the leverage effect for US acquisitions and find that debt has a positive influence on decision making. They state that agency costs are a real phenomenon and that capital structure adapts to account for them.
Apparently it can go either way, contingent claim view predicting a negative relationship and monitoring debt view predicting a positive relationship. Two theories that both hypothesize a different effect. When examining the existing literature it appears that evidence supporting the contingent claim view is limited. In addition, the only paper that examines the effect of leverage in acquisitions (Maloney, McKormick and Mitchell, 1993) does not even mention it. This study argues that an examination of the leverage effect should not neglect the contingent claim view. However, to observe the effect predicted by the contingent claim view two assumptions must hold. First, the acquisition itself needs to be diversifying. This will not be the case when the operations of the acquired company are similar to the acquirer. Second, the diversification must have a potential impact on the value of a firm. This will not be the case when a firm is already diversified. Previous papers did not take these implications into account. Hence, it is not clear what actually is observed. To determine diversification exposure in acquisitions this research divides the sample in groups. Based on diversification literature a classification is made depending on whether the acquirer and acquisition are focused or diversified.
The aim of this paper is to empirically investigate the leverage effect in Western European acquisition announcements. The organisation is such that it tries to answer the following research question: Are acquiring firms in Western Europe exposed to a leverage effect?
The leverage effect is examined for a sample of 307 acquisition announcements between January 1997 and January 2008. The sample does not does not raise evidence for the contingent claim hypothesis. Diversifying acquisitions show no interaction with leverage on bidder returns. Weak evidence is found for the monitoring hypothesis as only diversified firms engaging in focused acquisitions show a significant positive association with leverage on bidder returns. The results suggest that firms should not determine capital structure in order to influence announcement returns.
2. Literature Review
This section consists of several parts. First an understanding of acquisitions announcements and its influence on shareholder value is provided. Second, the factors that are known to lead to a ‘winning’ acquisition are outlined. Next is described how firms arrive at certain capital structure levels. Subsequently empirical evidence of the leverage effect is presented. The final part discusses the diversification assumptions of the contingent claim view.
2.1 Acquisitions
Acquisitions are the largest type of corporate investment and are a big part of the corporate finance world. They occur daily and very often make it to the headlines of newspapers. If they do not deal with creating big companies from smaller ones, it is the other way around and break up companies. What is it that makes acquisitions such an attractive form of investment for companies? Apparently the underlying rationale is the synergy effect, one plus one makes more than two. Is this rationale actually valid? Let’s take a look at a grasp of some empirical evidence. Of course hundreds of papers are written about the impact of acquisitions, the scope of literature in this field is extensive. To give an overview of them would be both time consuming as well as beyond the research objective of this paper. However, a few papers are of interest to get an idea of the influence of acquisitions on bidder returns. Bruner (2001) gives an overview of literature relating to acquisition announcements in the US and Campa and Hernando (2004) is one of the few papers which describe acquisition announcements in Europe.
Bruner (2001) summarizes the evidence from 128 studies from 1971 to 2001. Studies with various research approaches are included; event studies, accounting studies, surveys of executives and clinical studies. Bruner finds that for the mass of research it appears that target shareholders earn sizable positive returns, bidders earn zero returns on the aggregate, and bidders and targets combined earn positive returns. Therefore one could say M&A does indeed pay as the combined firm earns positive returns, however the acquiring firm apparently has to pay the expected synergies in the form of a premium to the target firm, leading to disappointing returns to the bidder. Bruner notes that expected synergies, efficiencies and value-creating growth seem hard to obtain. If anything, value is created by focus, relatedness and adherence to strategy. Diversification, size maximization, empire building and managerial hubris destroy value. Executives therefore should approach acquisitions with caution.
decade have stimulated a significant restructuring of companies operating in the European Union, and particularly in those countries that belong to the euro area. In this paper 262 firms over the period 1998 to 2000 are examined. Consistent with the evidence that the authors document in existing literature, target firms earn positive cumulative abnormal returns of 9 percent in a one month window. Acquirers’ cumulative abnormal returns are zero on average.
One plus one indeed makes more than two, but only the target gains. Why is it that managers spend their valuable time on something that has been proven to result in zero abnormal returns? This topic has been occupying many researchers for years. Several possible explanations are available. Overconfidence (Malmendier and Tate, 2005) or managerial hubris (Roll, 1986) in acquisitions is one; managers overestimate the returns on their investments and underestimate the costs. Reducing idiosyncratic risk also provides incentives to engage in acquisitions; managers with higher equity ownership face higher idiosyncratic risk, diversification lowers that risk (Amihud and Lev, 1981). However, as the following paragraph shows, it may also be that an acquisition is based on solid empirical evidence.
2.2 Towards a succesful acquisition
Although generally speaking acquisitions lead to zero returns for the bidder, it has to be emphasized that this is on average. There are acquisitions that lead to succes and there are acquisitions that are value-destroying for the shareholders. This section presents evidence documented in literature concerning the factors that influence announcement returns.
Travlos (1987) finds that the method of payment impacts stock returns. Bidding firms suffer significant losses in acquisitions paid for with shares, but experience normal returns in cash offers. The main underlying rationale is assymetric information. The bidding firm will finance the acquisition in the most profitable way for the existing stockholders. When the managers of the bidding firm believe that their firm is overvalued by the market they will issue shares, if the opposite holds cash is used. This signalling effect causes stock financing to be received negatively by the market.
Jensen (1986) develops the free cash flow hypothesis. He states that firms with large amounts of free cash flow are more likely to engage in value destroying acquisitions rather than pay out to shareholders. Lang, Stulz and Walkling (1991), examining the free cash flow hypothesis, find that bidder returns are indeed negatively related to free cash flow. An increase in free cash flow equal to 1 percent of the bidder’s total assets is associated with a decrease in gains to the bidder of approximately 1 percent of the value of the bidder’s common stock.
between shareholders and managers. This control mechanism is reduced in effectiveness for firms with large amounts of cash.
Moeller, Schlingemann and Stulz (2004) propone the influence of firm size in acquisitions. The authors find that announcement returns for acquiring-firm shareholders is roughly two percentage points higher for small acquirers. They argue that large firms offer larger acquisition premiums than small firms and enter acquisitions with negative dollar synergy gains. The evidence is consistent with managerial hubris playing more of a role in the decisions of large firms.
Asquith, Bruner and Mullins (1987) address the positive relation of relative size with bidder returns. They argue that absolute dollar gains may be the same for firms making an acquisition, however when measured as abnormal returns, relatively large bidder gains may appear statistically insignificant. Adjusting for relative size overcomes this problem.
Masulis, Wang and Xie (2007) examine corporate control mechanisms in mitigating the manager-shareholder conflict of interest. They find that bidder vulnerability to the market for corporate control has a positive influence on returns. Masulis et al. state that corporate governance provides managers with the proper incentives to maximize shareholder value and hence these firms experience higher abnormal announcement returns.
Fuller, Netter and Stegemoller (2002) find that returns to acquirers of private targets are higher than returns to acquirers of public targets. They suggest that acquiring private firms equals purchasing assets in a relatively illiquid market. The valuation of those assets reflects a liquidity discount, resulting in a higher return to bidder shareholders. Doukas, Gonenc and Plantinga (2008) approaching this phenomenon from a contingent claim view attribute the listing effect to increased idiosyncratic uncertainty as private firms are subject to less analyst coverage.
Moeller and Schlingemann (2005) show that cross-border acquisitions result in significantly lower announcement returns relative to domestic acquisitions. According to the authors the cross-border effect can be addressed to country-specific factors. In particular they mention takeover-activity in the target country and the legal system. A possible explanation not mentioned by Moeller and Schlingemann is provided by Seth, Song and Pettit (2002) who argue that the value destruction in cross-border activity is due to risk effects5.
All of these factors can be taken into account by managers to assess whether an acquisition will be able to create value. Hence, an acquisition can be made on solid empirical expectations. This research contributes to the extensive literature about the factors that drive M&A success by examining the effect of bidder leverage.
5
2.3 Capital structure
Before analyzing the effect of leverage it is essential to understand what capital structure actually represents. Therefore, this section performs an investigation to the realization of a firm’s observed leverage ratio. How do firms arrive at certain levels of capital? This research does not intent to breathe new life into the capital structure puzzle. However, this section is important in creating an insight in the factors that determine a firm’s leverage ratio.
The theory of capital structure
According to one of the earliest works, Modigliani and Miller (1958), decisions concerning the financial structure have no impact on firm value. An increase in debt only offsets the wealth effects to the holders of shares. Hence, setting a leverage ratio merely reflects the amount of risk the owner of a firm is willing to take. Modigliani and Miller’s proposition one states that in a perfect and efficient market a firm cannot change the total value of its outstanding securities by changing the proportions of its capital structure: It does not matter how the pie is divided. This theory is rather counterintuitive, but has provided some basic fundamentals.
Based on proposition one by Modigliani and Miller6, Kraus and Litzenberger (1973) have developed the trade-off theory, in which both taxes and bankruptcy policies are considered in the determination of the optimal leverage ratio. In this theory firms seek to maximize the market value of the firm by trading off between the tax advantage of debt and the disadvantage of the bankruptcy costs that come along by taking on debt. Firms take on debt until a certain point is reached where bankruptcy costs exceed tax advantages.
Myers’ (1984) pecking order theory7 presents a different argument. He states that an optimal capital structure does not exist, instead each firm’s observed leverage ratio reflects its cumulative requirements for external finance. Firms are reluctant to raise external finance because of information asymmetry. As managers have more information investors demand a discount. Hence, internal finance is prefered. If external finance is required firms issue the safest security first. They will start with debt, then issue hybrid securities like convertible bonds and use equity financing as a last resort as this is most sensitive to information.
Stulz (1990) presents another perspective on the optimal financing policy of a firm by including agency costs and the role of debt as a monitoring device. In his model Stulz shows that financing policy matters as it reduces the agency costs of managerial discretion. According to Stulz managerial discretion has two costs, overinvestment- and underinvestment costs. Overinvestment costs occur when management pursues its own objectives and instead of maximizing shareholder value,
6
And notes in papers by Robichek and Myers (1965) and Hirshleifer (1966), where both mention that optimization of capital structure involves a tradeoff between the advantages and disadvantages of a marginal increase in leverage.
7
maximizes investment for empire building means. Underinvestment costs occur when managers cannot fund positive NPV projects with internal resources and management’s lack of credibility prevents them from being refinanced. In this theory, debt issues reduce the overinvestment costs but exacerbate the underinvestment costs, while equity issues reduce the underinvestment costs but worsen the overinvestment costs. Since both types increase one form of managerial discretion and decrease the other there is a unique solution for the firm’s optimal capital structure.
Baker and Wurgler (2002) present evidence of the effects of market timing on capital structure. They argue that capital structure is the cumulative outcome of past attempts to time the equity market. According to Baker and Wurgler there is no optimal capital structure. The market timing effect is persistent and current capital structure is strongly related to historical market values for the examined sample.
Firm-related factors
Despite the plausibility of the above theories, advanced econometric papers claim that this is not sufficient in understanding financial policy. Firm-related factors are also considered to be important in the determination of capital structure.
Rajan and Zingales (1995), Hovakimian, Opler and Titman (2001) and De Jong, Kabir and Nguyen (2008) find that tangible assets has a positive influence on the leverage ratio. These papers all explain this relationship by pointing at the collateral value of tangible assets. Because of agency conflicts bondholders risk being expropriated by stockholders, selling secured debt restricts the borrower to use the funds for a specific project. Otherwise, creditors will require more favourable terms, which in turn may lead such firms to use equity rather than debt financing (Titman and Wessels, 1988), resulting in lower leverage. Further, Myers and Majluf (1984) demonstrate that because of asymmetric information issuing securities is costly. Collateralization mitigates this asymmetry and therefore has a positive impact on the amount of debt in a firm’s capital structure.
According to De Jong et al. (2008) large firms tend to be more diversified and less prone to bankruptcy, suggesting that large firms should be more highly leveraged. Smaller firms are expected to be financed less by debt because of the relatively larger information asymmetry problem (De Jong et al., 2008). The opposite is found by Titman and Wessels (1988) for short-term debt ratios. According to the authors this might reflect the relatively high transaction costs small firms face when issuing long-term debt or equity. Despite these contrary predictions, research by Rajan and Zingales (1995), Titman and Wessels (1998), Hovakimian et al. (2001) and De Jong et al. (2008) all find evidence of a positive relationship,
which is debt. Second, when firms are profitable, they accumulate retained earnings and are relatively less levered, when firms are unprofitable they will not accumulate earnings and will automatically become higher leveraged.
De Jong et al. (2008) suggest that firms that are exposed to higher firm risk have lower leverage ratios. The reason is straightforward: Higher risk indicates higher volatility of earnings and higher probability of bankruptcy. Firms with low risk face lower costs of financial distress and are therefore able to take on higher levels of debt in their capital structure.
Rajan and Zingales (1995), Hovakimian et al. (2001) and De Jong et al. (2008) present the negative effect of growth on leverage. With respect to long-term debt Titman and Wessels (1988) state that equity-controlled firms have a tendency to invest suboptimally to expropriate wealth from the firm's bondholders. Firms in growing industries have more flexibility in their choice of future investments, therefore creditors are reluctant to issue long-term debt. In the same line of reasoning short-term debt might be positively related to growth as these firms are forced to substitute short-term financing for long-term financing.
As it appears, capital structure theory describes how firms choose their capital structure to optimal levels, while advanced econometric studies demonstrate that firm-related factors determine the debt-equity choice. Either way, this section has provided the reader with an understanding of what leverage actually represents and how firms arrive at certain levels of debt and equity. The next part will describe the empirical evidence of the effect of leverage in mergers and acquisitions.
2.4 Empirical evidence
Several studies have examined the role of the acquirer’s leverage in acquisitions. In order of relevance this section starts by presenting the results of the papers examining the influence of leverage in merger and acquisition announcement returns8. Following, the impact of leverage on other measures than shareholder returns is given. This section ends with a graphical overview of the literature.
2.4.1 Announcement returns to bidders
This paragraph addresses studies that have examined the impact of bidder leverage on acquisition announcement returns. Maloney, McKormick and Mitchell (1993) have based their research on this topic. Other studies only mention the effect of leverage in a broader research objective: Lang, Stulz and Walkling (1991), Harford (1999), Kang, Shivdasani and Yamada (2000), Moeller, Schlingemann and Stulz (2004), Cosh, Guest and Hughes (2006), Kang (2006) and Masulis, Wang and Xie (2007).
Maloney, McKormick and Mitchell (1993) is probably the most cited paper when it comes down to the role of acquirer leverage in acquisitions announcement returns. In their paper Maloney et
8
al. examine the influence of leverage in the US for three sets of data. The first set contains 428 mergers over the period 1962-1982. Here a significantly positive return at around 0.5 is found for different measures of leverage9, meaning that an increase in leverage with 1 point increases acquisition announcement returns with 0.5 percentage points. The second set examines an independent dataset containing all types of acquisitions. For this sample 389 purchases over the period 1982-1986 are included. Again a positive and significant relation is found between leverage and, as they call it, ´the market´s assesment of the quality of the managerial decision to acquire´. Implying that the market values the decision to acquire better when a firm has more debt in its capital structure. In this sample the coefficients for the different specifications of debt-equity ratios10 are at around 0.8. The third set consists of 1,158 firms that engaged in a financial restructuring between 1982 and 1989. After imposing the selection criteria 173 acquisitions by 44 firms remained. For this sample acquisition performance prior to the leverage increase is significantly negative, the change in performance following the restucturing is significantly positive and both of these effects are significantly related to the degree of leverage restructuring. Overall, the evidence from the paper by Maloney, McKormick and Mitchell is straightforward; their findings support the hypothesis that debt has a positive influence on managerial decision making. Higher leverage leads to better acquisitions and therefore higher announcement returns.
Lang, Stulz and Walkling (1991) use bidder’s leverage as a control variable in examining the free cash flow hypothesis for a sample of 101 successful tender offers between 1968 and 1986. They use several control variables. Each coefficient has the expected sign according to the literature. However, leverage measured by debt to total assets, is negative and insignificant, whereas in Maloney et al. it is positive and significant. When using a measure of leverage more similar to Maloney et al., namely long-term debt divided by the market value of equity, they find a positive insignificant coefficient. Though, Lang et al. argue that it is appropriate to include short-term debt in a measure for leverage. The theory by Maloney et al. describes the monitoring function of debt in the capital structure of a firm, this is especially valid for short-term debt as short-term debt forces managers to return to the capital markets frequently and managers will therefore automatically be more subject to monitoring.
Harford (1999) examines whether acquisitions by cash-rich firms are value decreasing. The sample includes 487 deals from 1977 to 1993. For several regressions Harford found leverage, defined as the book value of debt divided by the sum of the book value of debt and the market value of equity,
9
The following measures of leverage are employed: unadjusted long-term equity ratio; long-term debt-equity ratio net of economy-wide debt-debt-equity value in year prior to merger; long-term debt-debt-equity ratio net of debt-equity value for industry classification in which firm operated in year of merger.
10
to have an insignificant negative effect on returns with coefficients at around -0.003. Therefore the cash effect is not a leverage effect.
Kang, Shivdasani and Yamada (2000) examine the influence of bank relations on acquisition announcements returns by analyzing 154 Japanese mergers during 1977 to 1993. In a multivariate regression analysis they find that the returns display a strong positive association with the strength of acquirer’s relationships with banks. In their model they estimate bidder CAR’s as as function of total leverage and other relevant control variables. Leverage is measured as the the ratio of book value of total debt to the sum of the book value of total debt and the market value of equity. In the regressions Kang, Shivdasani and Yamada model leverage in groups above and below the sample median. Leverage above the median appears to be significant at a 1 percent level with a coefficient of 0,016. Evaluating the estimated coefficient at the mean suggests that a one standard deviation increase in total leverage results in CAR that is higher by 1.67 percent. The effect of leverage on investors’ ex ante valuations of merger announcements appears to be both statistically and economically important.
Moeller, Schlingemann and Stulz (2004) examine the size effect in acquisitions for a sample of 12,023 deals. A quite impressive dataset and according to the authors such a comprehensive sample has not been studied before. Leverage, measured as the firm’s total debt over the firm’s market value, is in particular relevant for this research as Maloney, McKormick and Mitchell found that small firms have a higher leverage than large firms. The size effect could thus be a leverage effect. Moeller, Schlingemann and Stulz find that both for small and large firms leverage has an insignificant positive sign. Small firms gain more in acquisitions and this is robust for leverage.
Kang (2006) examines the effectiveness of a firm’s monitoring mechanisms to prevent managerial opportunism in acquisitions under high environmental complexity and/or dynamism. The dataset consists of 100 acquisitions by US firms in 1995 only, which is to control for year effects. Leverage is measured as the ratio of long-term debt over total assets and has a positive coefficient, significant at the 10 percent level.
Cosh, Guest and Hughes (2006) investigate the relation between takeover performance and board share-ownership of the bidding firm. For a sample of 363 completed UK takeovers in the period 1985-1996 the authors also include leverage in their regressions. They state the importance of leverage in the following phrase: “Decision-making may also be constrained by the presence of bank-debt or leverage. To the extent that directors of corporations incur restrictions on their activity through the existence of covenants associated with leverage, or have their actions constrained because of the level of leverage which they have attained, then once again discretion to act in non-profit maximising ways may be inhibited”. They measure leverage as total debt divided by the sum of total debt plus the market value of equity. Acquirer leverage has a small positive influence on acquisition announcement returns, but on the long run, this effect is reversed and becomes significant at the 10 percent level.
(1990) they argue that higher debt levels reduce future cash flows and limit managerial discretion. Furthermore they refer to evidence of leverage as takeover protection. They define leverage as the book value of total debt divided by the market value of total assets. Using a sample of 3,333 completed acquisitions by 1,268 firms during the period between 1990 and 2003 Masulis et al. find that leverage has an insignificant positive effect on bidder returns.
2.4.2 Other measures
Besides the papers described above, several other papers as well mention the role of leverage in acquisitions for bidding firms, however on other measures than annoucement returns. Here, again one paper has described the impact of leverage extensively, which is Mansi and Reeb (2002). Other papers by Switzer (1996), Linn and Switzer (2001) and Martynova, Oosting and Renneboog (2006) only mention the impact of leverage as part of a broader research objective.
Mansi and Reeb (2002), researching the cause of diversification discount, present results consistent with the contingent claim hypothesis. For a sample of 2,856 firms from 1988 to 1999 they measure the impact of diversification on firm value, based on the excess value measure by Berger and Ofek (1995)11. This measure is calculated by taking the logarithm of the actual value divided by the imputed value. Where the actual value equals the sum of the market value of equity and the book value of debt and imputed values are stand-alone values for each individual business segment. The research objective of Mansi and Reeb is to explain the diversification discount, and one of their main research questions is whether this diversification discount to shareholders is related to firm leverage. While diversification has no impact on firm value12, the influence of leverage is consistent with the contingent claim hypothesis. They perform two regressions on firm value. In the first regression, which is based on the book value of debt, leverage and an interactive variable for leverage and diversification are significant at the 1 percent level, with coefficients of -0.006 and -0.010 respectively. For the second regression, which analyses a subsample of 696 firms between 1993 and 1997, the market value of debt is used, resulting in leverage being significant at the 1 percent level with a coefficient of -2.17613. In their research Mansi and Reeb find that excess firm value is insignificantly related to corporate diversification. They argue that while diversification reduces shareholder value, it enhances bondholder value, and overall there is no significant impact on firm value. Their main findings with respect to this research are that all-equity firms do not exhibit a
11
(
)
value imputed value actual EV _ _ ln = 12Firm value equals total debt plus the market value of equity.
13
diversification discount and that equity holder losses in diversification are related to firm leverage, consistent with the contingent claim hypothesis.
Switzer (1996) examines the change in operating performance of 324 mergers between 1967 and 1987. Controlling for several variables including leverage Switzer finds that the performance of merged firms improves following the combination. In the regression leverage is modelled as a dummy variable, which takes the value 1 if the leverage of the bidder was greater than the median leverage for firms in the sample. Bidder’s leverage is measured by the book value of total debt divided by the sum of the book value of total debt, plus the book value of preferred stock, plus the market value of equity. Switzer found a positive but insignificant influence of the bidding firm’s leverage on median industry-adjusted operating cash flow return on assets of the combined firm. This evidence can be translated to a measure based on announcement returns, as a positive association has been found in this research between abnormal returns around the merger and the changes in observed operating performance.
Linn and Switzer (2001) also examine the change in operating performance of firms which merge, but in contrast to Switzer (1996) they look at the relation between this change and the method of payment. Their sample consists of 413 mergers for the period between 1967 and 1987. Consistent with Switzer’s earlier paper leverage is modelled in the regression as a dummy variable, which takes the value 1 if the leverage of the bidder is greater than the median leverage for firms in the sample. Leverage is also measured in the same manner as Switzer (1996). The authors find that acquisitions with cash outperform those with stock. This result is not sensitive to the bidder’s leverage as the coefficient of leverage is positive but insignificant.
Martynova, Oosting and Renneboog (2006) investigate the long-term profitability of a sample of 155 European takeover deals between 1997 and 2001. To examine the influence of leverage14 the sample is divided into quartiles based on the bidder’s leverage before the acquisition, where after the differences in post-acquisition profitability of the combined firms across the sub-samples are tested on significance. Leverage is defined as the total book value of debt divided by the book value of total assets. The outcome is that higher levels of pre-acquisition leverage do not lead to higher post-acquisition profitability. For different measures of operating performance Martynova et al. find both positive and negative coefficient values of bidder’s leverage prior to the acquisition, however in neither case it was significant.
As can be seen, the results in the literature are quite inconclusive about the effect of leverage in acquisitions. On announcement returns a positive sign is observed more frequent, whereas for other measures, both a positive as well as a negative sign is observed.
14
2.4.3 Literature overview
In this section a graphical overview is presented of the literature. In table 1 an overview is given of literature that examines the effect of leverage on announcement returns. All information is outlined in the above section except the seventh and eight column. The seventh column gives the estimation window. In this period the normal parameters for the regression are estimated, this can be seen as a reference period. The eight column gives the event period. This is a period surrounding the event, the announcement, which is excluded from the estimation window in which the abnormal returns are measured, the cumulative abnormal announcement returns.
Table 1: Literature overview on announcement returns
In table 2 on the next page, literature describing the effect of leverage on other measures than announcement returns is shown. Interesting to note are the different measures each paper uses for the dependent variable as well as the estimation of the performance of a merger or acquisition.
Year Author(s) N Country Dependent
variable Period
Estimation Window
Event
Window Coefficient Significance 1991 Lang, Stulz and
Walkling 101 US Shareholder returns 1968-1986 -300, -60 -5, +5 -0,135 Not significant 428 1962-1982 0,5 Significant at 5 percent 1993 Maloney, McKormick and Mitchell 389 US Shareholder returns 1982- 1986 -365, -50 -2, 0 0,8 Significant at 1 percent 1999 Harford 478 US Shareholder returns 1977-1993 -370, -20 -5, +1 -0,003 Not significant 2000 Kang, Shivdasani and Yamada 154 JPN Shareholder returns 1977-1993 -220, -20 -1, +1 0,016 Significant at 1 percent 2004 Moeller, Schlingemann and Stulz 12,023 US Shareholder returns 1980-2001 -205, -6 -1, +1 0,0007 Not significant 2006 Kang 100 US Shareholder returns 1995 -250, -50 -1, +1 0,21 Significant at 10 percent 2006 Cosh, Guest and Hughes 363 UK Shareholder returns
1985-1996 N/A -1, +1 0,02 Not significant
Table 2: Literature overview on other measures Year Author(s) N Country Dependent
variable Period
Estimation of
performance Coefficient Significance
1996 Switzer 324 US Cash flow based measure of operating performance 1967-1987 5 year median operating cash flow
returns before acquisition versus 5 year after acquisition
0,026 Not significant 2001 Linn and Switzer 413 US See Switzer (1996) 1967-1987 See Switzer (1996) 0,051 Not significant Firm value based on the book value of debt -0,006 Significant at 1 percent 2002 Mansi and
Reeb 2,856 US Firm value based on the market value of debt 1988-1999 Regressing excess value on corporate diversification using leverage as a control variable -2,176 Significant at 1 percent 2006 Martynova, Oosting and Renneboog 155 EU EBITDA as measure of operating performance 1997-2001 3 year median profitability prior acquisition versus 3 year after acquisition
-0,135 Not significant
2.5 Diversification
Berger and Ofek (1995) estimate the impact of diversification on firm value using the excess value measure as described above. They analyze a sample of 3,659 firms for the period 1986-1991. Berger and Ofek compare the sum of the stand-alone values to the firm’s actual value and find a 13 to 15 percent average value loss from diversification. The value loss appears to be smaller when the segments of the diversified firm are in the same two-digit US SIC code15. Additional support for their conclusion is given by the fact that single-segment firms have higher operating profit than the segments of diversified firms. Where the authors find that overinvestment and cross-subsidization contribute to the value loss, they also document benefits of diversification. Higher debt capacity and the ability of multi-segment firms to immediately realize tax savings by offsetting losses in some segments against profits in others appear to be small, only 0.1 percent of sales.
Mansi and Reeb (2002) also explore the diversification discount. For a sample of 2,856 firms from 1988 to 1999 they argue that the discount stems from the risk-reducing effects of corporate diversification. Viewed in a contingent claim framework, diversification reduces shareholder value and increases bondholder value. The average diversified firm has about 12 percent equity discount and as leverage increases, the wealth loss to shareholders becomes substantially higher.
Denis, Denis and Yost (2002) investigate the effects of global diversification and industrial diversification on firm value between 1984 and 1997. For a sample of 7,520 firms they find that both global diversification and industrial diversification result in valuation discounts. Multivariate regressions point out a 20 percent industrial diversification discount and an 18 percent global diversification discount. Firms that are diversified both industrially as globally trade at a 32 percent discount.
Graham, Lemmon and Wolf (2002) criticize the method used by Berger and Ofek (1995) to examine diversification discount. They argue that the implicit assumption of prior literature that stand-alone firms are a valid benchmark for valuing the divisions of conglomerates can be misleading if there are systematic differences between those two. They examine a dataset of 356 companies over the period 1980 to 1995. The expectations of Graham, Lemmon and Wolf are confirmed in their paper. They find that units that are combined into firms trough merger or acquisition are priced at significant discounts. They further find that excess value is not reduced when a firm increases its number of business segments without making an acquisition. It is apparently not the diversification that destroys value according to the authors; it is the selection bias of firms that trade at a significant discount that reduces excess value in acquisitions.
Villalonga (2004) even finds a diversification premium. The author argues that the diversification discount documented in prior literature is a consequence of segment data. Villallonga uses the Business Information Tracking Series panel data between 1989 and 1996 as source claiming that it allows to constructs business units that are more consistently and objectively defined than
15
segments and therefore more comparable across firms. Using a sample that yields a discount in earlier work for segment data, Villalonga finds a diversification premium. With this paper Villalonga questions the received wisdom about the diversification discount.
Doukas and Kan (2006) explore the sources behind the documented diversification shareholder value loss in acquisitions. Analyzing a dataset of 612 cross-border acquisitions of US firms over the period 1992 to 1997 they find that bidders that are globally and/or industrially diversified trade at a discount. Viewed in a contingent claim framework diversification indeed decreases the value of shareholder equity, but it should increase bondholder value. Therefore the excess value measure might be downward biased as it uses the book value of debt instead of the market value, which would represent the increase in value. According to Doukas and Kan this is hypothesis is correct, shareholders lose in acquisitions and bondholders gain. They find that the value loss to shareholders from diversification is directly related to a firms’ leverage, near-all equity firms do not trade at a discount and the use of book value debt in estimating excess value produces a downward bias in globally diversified firm. Overall diversification does not destroy firm value.
As in each theory, the diversification discount hypothesis has opponents and proponents. The findings in literature suggest that diversification destroys value and firms that are diversified trade at an equity discount. On the other hand, some authors question this hypothesis and state that this discount is not caused by diversification but selection bias (Graham et al., 2002) or even that this discount does not exist and is a consequence of segment data (Villalonga, 2004). However, relevant for this research is that the existence of a potential discount might distort the effect of diversification on shareholder value as predicted by the contingent claim view.
Implications
As this paper examined whether acquiring firms are exposed to a leverage effect the diversification assumption of the contingent claim view should not be neglected. It is in particular expected to see the influence predicted by this theory when acquisitions are diversifying and do not trade at a discount. Therefore, to capture for diversification exposure the total sample will be divided in groups depending on whether the firm16 and the acquisition17 are focused or diversified.
16
This division for firms is common in diversification literature but is often labelled as single-segment bidders and multi-segment bidders (Mansi and Reeb, 2002; Denis, Denis and Yost, 2002; Doukas and Kan, 2006)
17
3. Data, Expectations and Methodology
This section will first discuss the data used for this research. The next part will describe the expectations of this study. Lastly, the methodology is outlined to analyze the leverage effect in Western European acquisition announcement returns to bidding firms.
3.1 Sample selection
To examine the leverage effect in Western European acquisitions announcements, a sample of 307 takeovers is collected from Zephyr for the period between January 1997 and January 2008. Zephyr is an electronic database covering all information about acquisitions deals around the world. Relevant return data and information about capital structure has been gathered using Thomson Datastream, a financial statistical database.
The criteria used for selection and refinement of the data can be seen in table 3. First of all, acquisitions announcements in Western Europe18 are selected, which results in 55,866 deals. In the next step all deals that are not completed are eliminated, where after 38,297 acquisitions remain. The inclusion of this step is based on literature. Lang, Stulz and Walkling (1991) only focus on successful offers to make it less likely that their estimates of the target and bidder gains are biased downward because the probability of success is less than one. Of course this argument is intuitive and straightforward, and if this mitigates any bias in this research it is reason enough to consider only completed offers. Data will be examined for an extensive time period; from 1/1/1997 to 1/1/2008, resulting in 37,567 deals. This provides this study with an up-to-date sample, which is relevant as the majority of research has been done in the previous decade. To gain control over the acquired company the bidder has to end up with a majority stake, after this criterion 21,261 announcements remain. Next, acquisitions with a deal value below 1 million Euros are excluded to ensure that the deal has a potential impact on share price, leaving 2,599 events. Consistent with related literature financials firms are excluded from analysis. According to Berger and Ofek (1995) this is because the valuation methods are not applicable to firms in the financial services industry. 1,399 acquisition announcements remain. To be certain that return data is available, the next two steps eliminate all non-public bidding firms and companies without an ISIN code19. After eliminating all observations lacking available data on the variables that are included in the methodological framework, the final sample consists of 307 firms. In appendix 1 an overview is given of the dataset.
18
Western Europe is classified as including the following countries: Austria, Belgium, Denmark, Finland, France, Germany, Italy, Liechtenstein, Luxembourg, Monaco, Netherlands, Norway, Portugal, San Marino, Spain, Sweden and Switzerland. Deals from the UK are excluded as it would bias the results. Mergers and acquisitions occur more frequent in the UK and therefore the dataset would consists of a disproportionate part of deals from one country.
19
Table 3: Selection of events
Criterion Events
Mergers and acquisitions in Europe 55,866
Completed deals only 38,297
Period from 1/1/1997 to 1/1/2008 37,567
Acquiring a majority stake 21,261
Minimum deal value 1 million Euro 2,599 Exclude financial institutions 1,399
Acquirer is listed 463
ISIN code available 356
Eliminating erroneus observations 307
As put forward in the introduction, the effect of leverage on acquisition announcement returns can be described by two theories. The contingent claim view and the monitoring debt view. To perform a thorough analysis to the effect of leverage the diversification assumption of the contingent claim view will be taken into account. This will be done by dividing the sample in groups in order to capture for diversification exposure. As mentioned in chapter 2.6 the classification is based on whether the firm and the acquisition are focused or diversified. 4-digits US-SIC codes are used to divide the sample20. This leads to the following classification with respect to the bidder: If the acquirer has multiple 4-digits US-SIC codes, the acquirer is diversified. If the acquirer has only one 4-4-digits US-SIC code it is labelled as focused. For acquisitions the following holds: If the acquirer and the target share any 4-digits US-SIC codes according to Zephyr, the acquisition is focused. If the acquirer and target have no common 4-digits US-SIC codes, the acquisition is classified as diversified. Table 4 provides the classification results and the corresponding group numbers:
Table 4: Classification of acquirers and acquisitions
Focused Acquisition Diversified Acquisition
Group 1 Group 2 76 deals 28 deals Group 3 Group 4 152 deals 51 deals Diversified Acquirer Focused Acquirer
As can be seen in the table group 1 consists of focused firms making focused acquisitions. Group 2 are focused firms engaging in a diversifying acquisition. Group 3 includes diversified firms making a focused acquisition. Group 4 are diversified firms making a diversifying acquisition. It appears that the majority of deals can be classified as focused, they account for 228 deals, whereas there are only 79 diversifying acquisitions. Of these 79 acquisitions, focused acquirers engage in 28 deals and diversified acquirers in 51 deals.
20
3.2 Expectations
The effect of leverage can be described by two theories. For the total sample the effect of leverage is difficult to predict as they both mention another story. Therefore, as shown in the previous paragraph, the total sample is divided into 4 subgroups. This section will discuss the expectations per group. As already explained, the contingent claim perspective predicts a negative relationship between leverage and bidder returns when acquisitions reduce risk. This relation should be observed for group 2 and group 4 firms as they engage in diversifying acquisitions. However, the effect is expected to be more pronounced in group 2 as these firms are not affected by the potential impact of diversification discount. Firms in group 1 and group 3 do not reduce risk as they engage in focused acquisitions. The level of risk will be neutral or increasing in this case. Viewed from a monitoring perspective, the disciplining function of debt will be relevant for these groups as debtholders do not like potential risk increasing investments. They will monitor managers to make sure that they do not invest in uneconomic projects. The monitoring view of debt predicts that a higher leverage would lead to better acquisitions and hence higher returns. Therefore, leverage is expected to be positively related to bidder returns for acquisitions in group 1 and 3.
3.3 Methodology
To examine the influence of leverage in Western European acquisition announcements a regression analysis will be performed. A regression is an attempt to explain movements in a variable by reference to movements in one or more other variables (Brooks, 2002). In very general terms Brooks states that a regression is concerned with describing and evaluating the relationship between a given variable and one or more other variables. In this research the announcement returns of the bidder can be seen as the dependent variable and leverage and the control variables as independent variables. The multivariate regression model has the following form:
∑
= + + = N j i ji j i i x y 1 ~ε
β
α
(1) where, iy = The announcement returns for security i;
i
α
= The intercept of the linear relationship between the return of security i and the independent variables x;j
β
= The slope of the linear relationship between the return of security i and the independent variables x;ji
x = The independent variables;
i
Dependent variable
In this study the dependent variable is the bidder’s three-day cumulative abnormal return measured by means of the market model by Brown and Warner (1985). This measure is used as dependent variable by the majority of papers examing announcement returns21. Extending the event window results in a spread of the impact at the announcement date and lessens the chance on significance. Therefore, the three-day cumulative abnormal announcement returns as dependent variable is employed. Maloney, McKormick and Mitchell (1993) also use the 3-day cumulative abnormal return as independent variable, however instead of measuring the days surrounding the announcement, they employ a three-day cumulative abnormal return measured at the announcement date. In this paper the former approach will be used. The abnormal returns are calculated by using the market model:
it mt i it a R R = +
β
+ε
~ (2) where, itR = The rate of return of security i on date t;
mt
R = The rate of return on the market index of the firms in the dataset.
First, the rate of return on each security is calculated:
) log( ) log( − −1 = it it it RI RI R (3)
Where RIit is the Return Index of security i on date t. Using logarithms in calculating daily stock return data prevents correlation in the distribution of stock returns (Akgiray, 1989). The same formula is employed to calculate the rate of return on the market index. As a next step the normal parameters in formula 2,
α
i andβ
i, are estimated by using an ordinary least squares regression. These coefficients are obtained over an estimation window of 120 to 5 trading days before the acquisition announcement. Now, the abnormal announcement returns can be calculated by benchmarking the daily rate of return of security i against its normal returns:) ˆ ˆ ( ˆit Rit
α
iβ
Rmtε
= − + (4) 21When the residuals are obtained from 5 days preceding the announcement to 5 days after, the cumulative average residuals can be calculated for one day before the announcement to one day after the announcement as the dependent variable:
1
ˆ + −
= t t
t CAR
CAR
ε
(5)At this point the 3-day cumulative average residuals are obtained. Hence, the dependent variable is known. To further examine the announcement returns for the total sample or per group, the abnormal returns are simply added and divided by the number of observations22. To test for significance the event study methodology (MacKinlay, 1997) provides the following test23:
) var( t t t
ε
ε
= ~ N(0,1) (6) Independent variablesThe dependent variable is known. Needless to say the main independent variable is leverage. Leverage is defined as the total book value of debt divided by the sum of total book value of debt and the market value of equity. This measure is used most frequent in prior literature. A robustness test will be performed as well for another measure of leverage, namely debt to equity. To examine the effect of leverage in acquisition announcements it is essential to control for variables that might influence acquisition announcement returns. This is to be sure that what is observed when analyzing the data is not caused by another factor. Section 2.2 already outlined these factors: Corporate governance mechanisms, firm size, the relative size of a company, corporate holdings of cash, free cash flow of the bidder, the method of payment, whether the acquired firm is private or public and whether the acquisition is domestic or cross-border.
This research hypothesizes the monitoring function of debt, therefore other corporate governance variables are not included. The remaining variables are modelled in the multivariate regression analyis. To proxy for firm size this research does not use Moeller, Schlingemann and Stulz (2004) measure of size. Using market capitalization would on forehand lead to a high correlation with leverage. Therefore, consistent with De Jong, Kabir and Nguyen (2008) size is reflected by the natural logarithm of total sales. Following Asquith, Bruner and Mullins (1982), relative size is defined as the natural logarithm of the transaction value divided by the equity market capitalization of the acquirer at the end of the fiscal year prior to the acquisition announcement. Similar to Harford (1999) 22
∑
= = N j jt t N 1 ˆ 1ε
ε
23richness of a firm is defined by the amount of cash on the balance sheet divided by sales. Free cash flow of the bidder equals EBITDA divided by the book value of assets according to Lang, Stulz and Walkling (1991). With respect to the method of payment, a dummy is included for deals paid in cash only and a dummy for deals paid in shares only. 121 deals are financed with cash and 26 with shares. A dummy is included if the target is private, which is the case for 286 firms. Further 149 cross-border deals are also assigned a dummy variable equal to 1.
As hypothesized earlier it is expected that the negative relation between leverage and bidder gains is different between focused acquisitions and diversifying acquisitions. To provide a test of the association of leverage and each particular group on bidder returns, interaction terms are included in the model.
Correlation
To overcome a bias of multicollinearity, pair-wise correlations between the explanatory variables are measured. When explanatory variables are highly correlated with each other, the standard errors might be biased upwards and hence hypothesis tests could yield inappropriate inferences (Brooks, 2002). Table 11 in appendix 2 reports Pearson correlation coefficients for the independent variables. From the matrix it can be seen that leverage and debt-to-equity have a high correlation coefficient. These variables are used interchangeably and will not be included in the same regression. Another high correlation is between size and relative size. One option in this case is to drop the variable. However, dropping a theoretically important collinear variable might result in omitted variable bias. As it appears that the observed multicollinearity does not cause a problem in the regression, both are included in the model. Robustness tests will be performed to check for potential biases.
Regression
The multivariate regression takes on the following form. For each group a separate regression is performed. ε β β β β β β β β β β β α ~ ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) * ( ) ( ) ( 11 10 9 8 7 6 5 4 3 2 1 + + + + + + + + + + + + = dummy border cross dummy n acquisitio private dummy payment cash dummy payment shares size relative cash flow cash free size dummy group leverage dummy group leverage CAR (7) Testing
To test whether leverage or other independent variables are significant in explaining the three-day cumulative abnormal announcement returns, a t-test statistic is calculated:
X
S X
4. Results
Prior to analyzing the results of the regressions, this chapter will start describing the data. In order to draw inferences it is essential to have an understanding of the characteristics of the sample. After a thorough examination of the data the results of the market model event study and the multivariate regression analysis are presented.
Descriptive statistics
The sample includes 307 acquisitions in Western Europe for the period January 1, 1997 to January 1, 2008. In tables 14 and 15 in Appendix 2 the number of acquisitions per country and year are shown. From table 14 it can be seen that acquisitions from 14 different countries are included in the analysis. The distribution is somewhat skewed. Austria, Belgium, Denmark, Luxembourg and Portugal are all represented with less than 10 acquisitions in the dataset, whereas for France, Italy and Sweden around 40 acquisitions are included. From table 15 it can be seen that the numbers of acquisitions per year are also not distributed equally troughout the sample. The first five years only account for about 9 percent of the total amount of acquisitions. In the following six years the remaining acquisitions occur.
The descriptive statistics of the dataset are presented in table 5 below. The mean leverage ratio for firms in the sample is 0.23. The minimum leverage ratio is zero, caused by 16 firms without debt in their capital structure24. The sample of firms analysed in this research has on average a smaller debt to equity ratio than in Maloney et al. (1993). As can be seen in row 2 the mean debt to equity ratio is 0.45, whereas in Maloney et al. (1993) this is 0.566. For the firm characteristics total debt, market value of equity, fixed assets, total assets, EBITDA, cash and sales it can be said that a few outliers are responsible for the relatively low median compared to the mean and the relatively high standard deviation compared to the mean. This is indicated by the high maximum values in column 5. With respect to the deal value the same line of reasonings holds. The risk of a firm, which is measured by the covariance of the firm returns and the market returns, has a mean value of 0.64. This means that on average a firm’s stock returns vary positively with the market for the examined sample. In the last two rows past and future growth of the firms are given. Past growth, measured by current sales divided by sales one year prior to the acquisition, is on average 2.42. The large standard deviation indicates that there are several firms who have considerably increased their sales. Analyzing the data, it appears that only 13 firms have more than doubled their sales over the period of one year. Lundin Petroleum from
24
