1
Mergers and Acquisitions
Influence of the financial crisis on the method of
payment for European M&A’s
Abstract
This thesis examines how the investment opportunities of the acquiring firm affects the choice of payment method, how this effect changed during the crisis, and how the crisis itself
affected the choice of payment method for corporate mergers and acquisitions in Europe. This influence is tested using a sample of 418 mergers and acquisitions completed between January 2004 and June 2009 in twelve European countries. The results suggest that during the crisis having good investment opportunities decrease the probability to finance mergers and acquisitions with equity. Before the crisis this effect is found to be positive. Furthermore, a positive effect of the financial crisis itself on the probability to finance a merger and acquisition with equity is found.
Michiel Karssen 10264329 February 2015
BSc Business Economics - Finance & Organization University of Amsterdam
Bachelor Thesis
2
Statement of Originality
This document is written by Michiel Karssen who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
3
Acknowledgements
First of all I would like to thank my thesis supervisor, Ieva Sakalauskaite. I am especially thankful for the available time, advice, and great guidance during the writing of this thesis. Next, I would like to express my gratitude to all professors who helped me shaping my view on business economics during my bachelor. Finally, I would really like to thank my parents for giving me the opportunity to do this study and supporting me throughout this journey.
4
Table of Contents
Title and abstract 1
State of Originality 2
Acknowledgements 3
Table of Contents 4
1. Introduction 5
2. Literature review 7
3. Hypotheses and methodology 12
4. Data and descriptive statistics 18
5. Results 24
6. Conclusion 31
5
1. Introduction
The choice of payment for mergers and acquisition (M&A’s) plays an important role for an acquiring company. The payment method can influence the newly formed firm’s profits and future cash flows by having a significant impact on the ownership structure, financial leverage, and solvency. The research about the determinants of the choice of payment for M&A’s mostly focuses on firm and deal characteristics (Martin, 1996, Faccio and Masulis, 2005, Martynova and Renneboog, 2009). Martin (1996) was the first who argued that external determinants in the form of business cycle conditions might influence the method of
payment. For example, a high external shocks to fund availability. In addition, Garcia-Feijóo et al. (2012), state that industry and economic conditions also might influence the choice of payment. This thesis combines the influence of an external factor, the crisis, and firm and deal specific factors to investigate the payment method.
The existing literature classifies the payment method of mergers and acquisitions to three main payment methods. The acquiring firm can pay with only cash, only stock or a mix of these two methods (Martin, 1996, Faccio and Masulis, 2005, Martynova and Renneboog, 2009, Zhang, 2001, Swieringa and Schauten, 2007). Moreover, according to Swieringa and Schauten (2007), the choice between paying with cash and paying with equity is mainly the choice between debt-funding and equity-financing. They state that internally generated funds are regularly insufficient to finance cash payments in M&A transactions. Additional
borrowing in the form of debt is needed. Hence, this thesis will focus on the choice between equity and cash payments.
Numerous papers have focused on the effects of investment opportunities on the M&A payment method. Myers (1977), for example, linked the firm’s investment opportunities with the amount of debt a firm uses to finance its investments. He concluded that firms with better investment opportunities should finance their investments with more equity than debt. The effect of investment opportunities on the source of funding for specifically M&A investments has been in most detail investigated by Martin (1996). He found that it is one of the most important determinants of the payment method for M&A’s. Correspondingly with the
conclusion of Myers (1977), he stated that ‘the higher the acquirer’s investment opportunities, the more likely the acquirer is to use stock to finance an acquisition’. Furthermore, Whited (1992) explained that limited access to external cash funds induces acquiring firms to look for other sources to finance their investments. In the case of M&A’s this indicates equity
6
financing. Since not all investment opportunities can be exploited, the influence of the quality of these investment opportunities on the choice of payment for acquiring firms might change. In order to further explore the findings of Martin (1996) and Whited (1992), this thesis will investigate how the investment opportunities affect the payment method, and how this relationship is affected by huge shocks to cash availability, such as the crisis. As an indicator for investment opportunities, corresponding with the paper of Martin (1996), the Tobin’s Q (Brainard and Tobin, 1968) of the acquiring firm is used.
The crisis had a huge influence on the bank lending volume for the corporate sector (Ivashina and Scharfstein, 2010), reducing firms’ access to external cash funds. By increasing the cost of debt for acquiring firms and making them more constrained, the crisis might have affected the choice of payment for M&A’s through limiting access to finance, and the importance of investment opportunities in this decision.
Another contribution of this thesis to the existing literature is the focus on the twelve European countries that started using the Euro as their currency in 2002. Most research on the payment method uses data about transactions in the United States. Research on the payment method of European M&A transactions is limited. Faccio and Masulis (2005) were only the first to study European mergers and acquisitions.
The research question will be tested using a sample of 418 mergers and acquisitions completed in one of the twelve used European countries between January 2004 and June 2009.
The remainder of this thesis is organized as follows. First, the main existing theories about the method of payment and the influence of the crisis on the financial markets will be explained. Second, the hypotheses will be formed based on the literature. Thereafter, the methodology used to test the hypotheses is explained, along with the explanation of all the used variables. Third, the data sample used will be described and a small description of the distribution of the payment method in the sample is provided. Afterwards, the results of the estimated Probit model and the robustness tests will be reported. Finally limitations of this thesis and recommendations for future research will be reported.
7
2. Literature review
The first part of this section will provide an overview of the main theories in the existing empirical literature about the payment method for M&A transactions. Thereafter, the empirical papers about the influence of the crisis are discussed.
2.1 Theories
The existing literature regarding the payment method of M&A’s is wide-ranging. Before any theory was formed about the payment method, Modigliani and Miller (1958) stated that in a perfect market, the way an investment is financed should have no
consequences for the value of the firm. Nevertheless, in 1984, Myers stated that a firm would prefer internal financing, such as cash, over external financing, such as equity issuance for new investments. This would make them less dependent on the capital markets. This order of financing preference is called the pecking order theory.
More recent literature about the payment method also provides evidence that does not support the paper of Modigliani and Miller (1958). These papers show that the factors such as the debt/equity ratio, fraction of voting rights held by the managers and investment
opportunities of the acquiring firm affect financing preferences for M&A transactions. While most of the existing literature uses data on M&A’s in the United States, only a few papers focus on European M&A transactions. This section will start with empirical evidence from the US and will expand to research on European firms.
Research about the influence of investment opportunities on the source of financing is found about investments in general. Myers (1977) provides a theoretical model that predicts an inverse relation between a firm’s borrowing and the extent that the firm value depends on its investment opportunities. Issuing debt reduces the market value of a firm. Therefore a firm with good investment opportunities should use equity more often to finance its investments.
2.1.1 Method of payment in the United States
One of the leading papers for this thesis is the paper of Martin (1996), because the second hypothesis formed in section 3.1 is derived from its findings. Martin (1996) investigated the influence of the acquirer, target and deal characteristics on the method of payment of M&A’s for 846 acquisitions, completed during the years 1978-1988 in the United States. He concluded that the growth or investment opportunities, indicated by Tobin’s Q, of the acquiring firm are one of the most important characteristics determining the payment
8
method. Martin (1996) argued that acquirers with good investment opportunities are reluctant to borrow and therefore prefer equity financing. That is because borrowing might lead to possible constraints on debt-financing and thus less flexibility in the future when new investment opportunities arise. Martin (1996) found support for this explanation in a survey conducted by Pinegar and Wilbricht (1989). They surveyed Fortune 500 firms and found that ‘maintaining financial flexibility’ is one of the most important factors driving funding
decisions.
Whited (1992) tested the hypothesis that asymmetric information problems in capital markets affect financially constrained firms’ external cash availability and consequently their ability to exploit investment opportunities. The results of Whited (1992) indicate that financial constrains have a negative effect on the ability to exploit all investment opportunities.
Therefore, as stated in section 1, the lower cash availability induces acquiring firms to use other sources to finance their investments.
Furthermore, Martin (1996) found a significant relation between the probability of paying with cash and management ownership. Managers with ownership stakes between 5 and 25 percent are afraid of loss of control in case of an equity payment and therefore prefer to pay with cash. Using a sample of 209 domestic acquisitions completed between 1981 and 1983, Amihud et al. (1990) found that also evidence for a relation of managerial ownership and the payment method. A higher fraction of managerial ownership of the acquiring firm makes mergers or acquisitions more likely to be paid with cash. The findings by Amihud et al. (1990) are consistent with the finding of Stulz (1988). Stulz (1998) stated that managers with large ownership stakes are unwilling to pay with shares, because they are afraid of the loss of control. Amihud et al. (1990) differs from Martin (1996) by stating that the anxiety of loss of control is not restricted to certain ownership stakes.
Another variable that is frequently proved to influence the method of payment of M&A’s is firm size. For example, Martin (1996) found firm size to be an important
determinant of the financing decision. He derived his conclusion about the importance of firm size from the paper of Hansen (1987). Hansen (1987) found evidence for a greater likelihood of financing M&A’s with equity when the total assets of the acquiring firm are large relative to the total assets of the target. Hansen (1987) investigated the targets value with the
assumption of asymmetric information about the targets value. When this information asymmetry occurred, acquiring firms with greater total assets should be more eager to prefer equity as payment method to avoid the ‘lemon problem’ according to Hansen (1987). By
9
using equity financing, the relative bigger acquirer forces target shareholders to share the risk of overpayment by the acquirer. Therefore, firm size is assumed to have a significant
influence on the probability to finance an M&A transaction with equity.
2.1.2 Method of payment in Europe
The most relevant literature about the method of payments in Europe is written by Faccio and Masulis (2005). Therefore, most control variables used in this thesis are derived from this paper. Faccio and Masulis (2005) showed the influence of firm size, financial leverage, relative deal size and collateral on the payment method. Acquiring firms with high financial leverage prefer to pay with stock. High financial leverage is similar to a high
debt/equity ratio, which indicates that the firms experience debt-financing constraints. Faccio and Masulis (2005) found, contrary to Hansen (1987), evidence of a positive relationship between the acquiring firm’s size and the preference for a cash payment. They stated that larger firms experience lower cost of debt due to their lower probability of default. Lower cost of debt gives larger firms a greater ability to issue debt, which provides a greater ability to pay with cash. Their data sample included 3667 M&A’s in thirteen European countries, with the announcement date between January 1997 and 2000.
Following up on the paper of Faccio and Masulis (2005), Martynova and Renneboog (2008) also investigated the influence of several determinants of the payment method. They found evidence on the influence of variables such as control, financial leverage and collateral. The authors extended their research to the type of financing of cash-only deals. They
distinguish cash paid deals in either internal cash funded or external funded deals. Martynova and Renneboog (2008) indicate that the three main factors determining the method of
payment are the cost of capital, agency costs and the funding of payments. Their data included European M&A deals between 1997 and 2000.
Another follow up on the paper of Faccio and Masulis (2005) is the paper written by Swieringa and Schauten (2007). They analysed how factors related to this thesis affect the payment method choice of Dutch mergers and acquisitions. Consistently with the findings of Martin (1996) and Faccio and Masulis (2005), Swieringa and Schauten (2007) proved that large firms prefer to pay with cash and high-growth firms are more likely to pay with equity, compared to low-growth firms. They confirmed the statement of Faccio and Masulis (2005) that large acquiring firms have better access to debt markets, and therefore a greater ability to pay with cash. Swieringa and Schauten (2007) also confirmed the statement of Martin (1996)
10
that high-growth acquiring firms are unwilling to borrow. The data sample of Swieringa and Schauten (2007) contained 227 Dutch M&A’s during the period of January 1996 and
December 2005.
In conclusion, the main assumption of the relating research is that the choice of payment depends most on the relative gains and costs of paying with either cash or equity. Acquiring firms with high investment opportunities prefer to finance their M&A transactions with equity. On the other hand, another implication from the existing literature is the strong influence of the cost of debt on the choice of payment for mergers and acquisitions. The greater the ability to borrow, the more likely a firm is to pay with cash. Therefore, both external borrowing constraints and firm investment opportunities determine the chosen method of payment.
2.2 Financial crisis and the payment method
All previous research about the method of payment for mergers and acquisitions is based on the period before the crisis. Since this thesis wants to find what the influence of the financial crisis on the choice of payment is, the existing literature about the influence of the financial crisis will be investigated.
According to the Basel Committee on Banking Supervision (BCB, 2010), the rise in the cost of borrowing and the decrease of the availability of liquidity and credit during the crisis caused a contraction in the real economy. BCB (2010) stated that one of the main reasons the financial crisis in 2007 influenced the credit availability was the fact that banking sectors in many countries were strongly leveraged. This weakness in the banking sector caused that small losses were quickly translated to contraction of credit to the real economy and to higher borrowing costs. As stated in section 1, debt is one of the main components for cash payments. Therefore, changes in bank lending might have had an influence on the payment decision for M&A’s.
Gan (2007) found that the financial health problems of banks, caused by a decline in the asset markets, resulted in a decline in lending volume. He investigated how these financial health problems of banks affect the real economy by separating the impact of a loan supply shock and a demand shock. Banks are significantly related to the asset markets through their direct holding of stocks. The declining asset markets made banks less solvent. Gan (2007) argued that, corresponding with the paper of Stein (1998), asymmetric information between banks about their solvency, in combination with the decline of the asset markets affected the
11
interbank loan markets in a negative way. This decrease in credit availability for banks made banks less able to lend money to the corporate sector.
Ivashina and Scharfstein (2010) continued to investigate the influence of financial health problems of banks on the real economy. They stated that during the financial crisis bank lending started to decrease from mid-2007. This decline contracted lending for restructuring like LBO’s, share repurchases and M&A’s. According to the research of Ivashina and Scharfstein (2010), during the peak period of the financial crisis (Q4 of 2008), bank lending volume was 47% lower relative to the previous quarter and 79% lower relative to the start of the credit crunch (Q3 of 2007).
Campello et al. (2010) stated that firms with lower cash availability, which can be caused by a lower lending volume for example, tend to prefer equity financing over cash financing for their investments, such as M&A’s. They analyzed the decisions made by 1050 CFO’s in the U.S., Asia and Europe during the crisis. Firstly they determined whether the firm was financially constrained during the crisis. These results showed that firms became
significantly more constrained during the crisis. Thereafter, conditional on the requirement of being financial constraint, Campello et al. (2010) asked if the CFO’s preferred cash funding of investment opportunities when the debt availability was low. Another important finding of Campello et al. (2010) for this thesis is that differences, in for example determinants of the choice of payments, between firm’s financial situations became more significant during the crisis due to the financial constraints. Therefore, the effect of investment opportunities on the choice of payment might have changed during the crisis.
To further investigate the change in the effect of investment opportunities on the method of payment during the crisis, Mishkin (1992) provides a useful theoretical framework in understanding financial crises. He describes a financial crisis very precisely as follows: ‘A
financial crisis is a disruption to financial markets in which adverse selection and moral hazard problems become much worse, so that financial markets are unable to efficiently channel funds to those who have the most productive investment opportunities.’ In general,
Mishkin (1992) argues that a healthy economy requires a financial system that leads the funds to firms with the best investment opportunities. Financial crises interrupt this equilibrium. The information asymmetry between a bank and the borrowing firm increased, so indicators for financial health can become more important. This can imply that the importance and effect of having good investment opportunities for the choice of payment changed during a financial crisis.
12
Additionally to the lower external cash availability of acquiring firms, the financial crisis also affected internal cash levels. According to Campello et al. (2011), the contraction of the real economy caused declining revenues for many firms. To compensate for this decline, firms addressed their internal cash levels during the crisis and decreased the overall cash holdings.
Generally, the financial crisis reduced the bank lending volume for transactions like M&A’s. This decrease in bank lending for the corporate sector influenced the cost of debt and the availability of external cash for acquiring firms in M&A transactions. The financial crisis might have also affected the influence of investment opportunities on the choice of payment for M&A transactions.
3. Hypotheses and methodology
The first part of this section describes the hypotheses. Secondly, the methodology to test the hypotheses and all the variables included in the regression model are described.
3.1 Hypotheses
This thesis analyses the influence of the financial crisis on the payment method of M&A’s in twelve European countries. The European countries that are used in this sample all started using the Euro as their currency at the beginning of 2002. Based on the literature discussed in section two, two hypotheses are formed to measure the effect of the financial crisis. The hypotheses relate the method of payment to the presence of the crisis (H1) and investment opportunities indicated by Tobin’s Q (H2).
H1: The crisis years have a positive effect on the probability to pay with shares for M&A transactions
The main assumption regarding the method of payment in general is that the choice of payment depends on the relative gains and costs of paying with either cash or equity. The literature review in section 2.2 showed that the crisis caused a decline in bank lending (Ivashina and Scharfstein, 2010, Gan, 2007, Campello et al. 2010). This decrease negatively influenced the total availability of external cash and the cost of debt for acquiring firms in M&A’s. Besides to the effect on external cash, the crisis also influenced the internal cash levels in a negative way (Campello et al, 2011). This decline in internal cash level might have
13
made internal cash to be more valuable. Therefore, the gains from using cash instead of equity decreased during the crisis and more equity deals are expected.
H2.1: Tobin’s Q has a positive effect on the probability to pay with shares for M&A transactions
According to the results of Martin (1996), one of the most important factors
determining the payment method are the investment opportunities of the acquiring firm. He stated that the higher the acquirer's investment opportunities, the more likely the acquirer is to use stock to finance an acquisition. Therefore, hypothesis 2.1 focuses on this main
determining factor in order to find how investment opportunities affect the payment method. Correspondingly with Martin’s (1996) paper, the investment opportunities are indicated by Tobin’s Q in this thesis and are expected to increase the probability of financing a M&A transaction with equity
H2.2: The positive effect of Tobin’s Q on the probability to pay with shares for M&A transactions will decrease during the crisis
By combining hypothesis 1 and 2.1, hypothesis 2.2 describes the predicted change in the importance of investment opportunities for the determination of the payment method during the crisis.
Mishkin’s (1992) theoretical framework about financial crises suggests that the indicators for the financial health of the acquiring firm became more important during the crisis when accessing external funding. The investment opportunities of the acquiring firm are a good indicator for this according to Mishkin’s framework and Brainard and Tobin (1968). Nevertheless, Martin (1996) stated that borrowing might increase future constraints on debt-financing for acquiring firms with good investment opportunities. Acquiring firms with better investment opportunities prefer equity financing of their M&A’s. However, Ivashina and Scharfstein (2010) proved that the debt availability overall decreased during the crisis and the results of the survey by Campello et al. (2010) support this by indicating that firms became more constrained on average. According to Whited (1992) this induced the more constrained acquiring firms during the crisis to look for other sources to finance their investments.
14
prefer equity financing of their M&A transactions. If small Q firms start preferring equity financing as well, the average Tobin’s Q for equity payments will decrease. Consequently, the effect of investment opportunities on the probability to pay with equity is expected to weaken but still be positive during the crisis.
3.2 Methodology
In order to investigate the effect of the crisis and Tobin’s Q on the payment method of M&A’s in Euro countries the dependent variable, Method of Payment, is formed as a binary variable. This research focuses on the decision to pay with shares instead of cash, therefore the dependent variable equals one when the payment method includes some equity (equity-only or mixed) and zero when cash-(equity-only. Mixed payments and equity payments are merged because within mixed payments the acquirer does not always define the actual percentage of equity financing. The model that is used to determine the effects on the dependent variable is the Probit model.
3.2.1 Construction of the dependent variable
Within the existing literature about determinants of the method of payment, different approaches for the determination of the dependent variable are used. Faccio and Masulis (2005) used an Ordered Probit model, where Martin (1996) used a Logit model and Amihud et al. (1990) used a Probit model. The advantage of the Probit and Logit model is that the interpretation of the results is easier than when using an Ordered Probit model. Therefore these models are preferred over the Ordered Probit model. The difference between the Probit model and the Logit model lies in the distribution of the errors (disturbances). The Logit model errors are assumed to follow the standard logistic distribution, while the errors of the Probit model are assumed to follow the standard normal distribution (Stock and Watson, 2012). Since the standard normal probability distribution has thinner tails than the standard logistic probability distribution the Probit model is preferred.
15 3.2.2 The Probit model
The Probit regression is a nonlinear model specifically designed for binary dependent variables (Stock and Watson, 2012). The Probit model uses standard normal cumulative probability distribution functions (c.d.f.) to produce probabilities between 0 and 1, or stated as a formula:
Pr( Y = 1| X) = Φ(β0 + βiXi + εi)
Where indicates the standard normal c.d.f. The independently distributed error εi is
assumed to follow the standard normal distribution with mean zero and variance one. The parameters β0 and βi are typically estimated by maximum likelihood.
3.2.3 The regression function
The Probit regression will be executed three times. The first regression focuses on the effects during the whole sample period. The second and third regressions investigate the change in behavior of Tobin’s Q between the period before the crisis and during the crisis. Correspondingly, the following equation is formed to test the hypotheses:
Method of paymenti = β0+ β1 Crisis + β2Tobin’s Q + β3Crisis*Tobin’s Q + βiControl
variablesi + εi
Where εi are independently distributed robust errors and i indicates every deal individually.
Method of payment is a binary dependent variable with value zero for cash-only and one for mixed and equity-only paid deals.
3.2.4 Main variables of interest
The variable CRISIS is a dummy variable which equals one when the announcement of the deal was made during the financial crisis, and zero otherwise. The financial crisis is indicated as the period reaching from the third quarter of 2007 until the second quarter of 2009.
Ivashina and Scharfstein (2010) showed that bank lending started to fall in the third quarter of 2007. According to the National Bureau of Economic Research (2010), the financial crisis in the United States ended in June 2009. Consistently, the GDP growth rate in the twelve used European countries was firstly positive again in the third quarter of 2009 (Eurostat, 2014). The variable TOBINSQ is included to investigate the effect of investment opportunities on the
16
payment method of M&A’s. Tobin’s Q is computed by dividing the sum of the acquirer’s market capitalization and the total liabilities by the sum of the acquirer’s common stock and the total liabilities. Market capitalization, total liabilities and common stock are all derived at the beginning of the year of the announcement. The interaction term CRISIS*TOBINSQ is formed to investigate the change of the effect of Tobin’s Q during the crisis, corresponding to hypothesis 2.2. By including this interaction term the influence of Tobin’s Q on the
probability to pay with equity is influenced by whether the deal is announced during crisis or not. When CRISIS equals 0 the effect of Tobin’s Q is indicated by the coefficient of
TOBINSQ. However, when CRISIS equals 1, the effect of Tobin’s Q is indicated as the sum of the coefficient of TOBINSQ and the coefficient of the interaction term. In order to test the hypotheses, the effects can be determined by observing the significance and coefficients of the crisis dummy, TOBINSQ and the interaction term CRISIS*TOBINSQ.
3.2.5 Control variables
The used control variables are determinants of the method of payment derived from the existing literature. These variables can be divided to firm specific and deal specific control variables. Firm specific variables control for the characteristics of the acquiring firm and deal specific variables control for characteristics of the deal itself.
The variable FIRMSIZE will be used to control for the size of the acquiring firm. To determine the size of the acquirer, the natural logarithm of its total assets at the announcement date of the deal is used. As described in section 2, Faccio and Masulis (2005), Hansen (1987) and Martin (1996) stated that the percentage of equity in M&A payments is influenced by the size of the acquiring firm. Another control variable is CONTROL. Amihud et al. (1990) and Stulz (1998) found evidence of a negative influence of the fraction of voting rights held by the management of the acquirer on equity payments, because these managers are afraid of the loss of control when paying with equity. This thesis follows the relative, to Amihud et al. (1990) and Stulz (1998), easy approach to determine this fraction used by Swieringa and Schauten (2007). The variable CONTROL equals the amount of closely held shares divided by the number of common shares outstanding. Closely held shares are ‘shares held by officers, directors and their immediate families, shares held in trust, shares of the company held by any other corporation, shares held by pension/benefit plans and shares held by individuals who hold 5 percent or more of the outstanding shares’(Datastream).
17
By assuming that cash for cash payments is mainly obtained by debt (Swieringa and Schauten, 2007), the more leveraged a firm is the more constrained it is due to risk of default. Hence, the debt/equity ratio of the acquiring firm should have an influence on cash-only payments. Faccio and Masulis (2005) support this in their paper by showing that the higher the debt/equity ratio, the more likely the acquirer is to use equity as payment method. Since Tobin´s Q is computed by using the total liabilities, the effect of the debt/equity ratio can be captured by Tobin´s Q if there will not be controlled for this. To control for the effect of the debt/equity ratio, the variable FINANCIAL LEVERAGE is included. This variable is computed by dividing total debt of the acquirer at the announcement date by the total book value of its assets at the announcement date. To control for debt, this is not the only control variable needed. Faccio and Masulis (2005) stated that fewer tangible assets will decrease the debt capacity, since these assets can serve as collateral. The higher the cost of debt, the less eager acquiring firms will be to finance an M&A transaction with cash. This will influence the choice to pay with equity in a positive way. Therefore, the fraction of tangible assets relative to the total assets of the acquirer can be used as an indicator for the ability to borrow. The variable indicating this is COLLATERAL and is calculated by dividing the acquirer’s tangible assets in property, plant and equipment (PPE) by the book value of the acquirer’s total assets.
To control for deal-specific factors two factors relating to the deal size will be used. Firstly, DEAL SIZE is used as a control variable and is specified as the total value of the deal. Faccio and Masulis (2005) found that the deal value is positively correlated with the
preference of paying with equity instead of cash. This finding is consistent with the results of Martin (1996). Secondly, the variable, RELATIVE DEAL SIZE, is included to control for deal size. Corresponding with the paper of Faccio and Masulis (2005), RELATIVE DEAL SIZE is computed by dividing the deal value by the sum of the deal value and the market capitalization (at year end previously to the announcement date of the deal) of the acquirer. Deal size is part of the computation of relative deal size; therefore a high correlation of these variables is expected. To investigate the option of multicollinearity, the correlation of deal size and relative deal size along with other correlations between explanatory variables is tested in section 5.1.
18
4. Data and descriptive statistiscs
The first part provides a brief description of which databases are used in composing the sample. Then, the main characteristics of the data are provided.
4.1 Data
4.1.1 The sample
The sample includes only merger and acquisition transactions in twelve European countries. The acquirer and the target should both be located in one of those countries. This will make transactions better comparable due to a similarity of regulations. The European countries that are used in this sample are the countries that started using the Euro as their currency at the beginning of 2002. These countries are Austria, Belgium, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, Netherlands, Portugal and Spain. To have enough data from before and during the crisis, all transactions should be completed between January 2004 and January 2013. Additionally, the deal should not be announced later than the end of the financial crisis, the second quarter of 2009. Furthermore, the acquired stake should equal a minimum of five percent while the final stake of the acquiring firm after the transaction should equal a minimum of fifty percent. Following the main approach in the existing
literature, the payment method of the transactions should be either equity-only, cash-only or a mix of these two methods (Martin, 1996, Faccio and Masulis, 2005, Swieringa and Schauten, 2007, Zhang, 2001), Martynova and Renneboog, 2009). Deals containing any other payment method are excluded from the sample. In order to get access to financial data of the acquiring firm, the acquiring firm should be listed on a stock exchange in one of the Euro countries. There are no further restrictions on the target firm.
4.1.2 Data sources
From Zephyr data about the location of the acquiring and the target firm, the announcement date, the completion date, the deal value and information about the payment method is
collected. Data to determine Tobin’s Q and the control variables is collected from Datastream. These criteria result in a sample of 895 deals. After excluding deals with missing data on control variables or Tobin’s Q, the final sample consists of 418 deals. Two subsamples are constructed. The first with 298 deals announced before the crisis and the second with 120 deals announced during the crisis.
19
Figure II
Distribution of the Method of Payment before and during the Crisis
4.1.3 Distribution of the method of payment
In Figure I, the distribution of the method of payment in the sample is presented. Figure I shows that a large part of the M&A deals of the total sample are paid with cash-only (60%). The equity is the second favorable method of payment with less than 30% transactions (28%). Only 12% of the deals are paid with a mix of equity and cash.
Regarding the hypotheses formed in section 3.1 the main point of interest is the difference in distribution of the payment method before and during the crisis. Figure II shows this difference. The percentage of cash payments decreased from 64% to only 48% during the crisis. This decrease in cash paid deals caused a rise in equity paid deals of 12% and a rise of 5% of mixed paid deals. The change in distribution to equity payments during the crisis, supports hypothesis 1.
Figure I
20 4.2 Descriptive statistics
In order to provide a description of the used variables, two tables are formed. Table I summarizes the explanatory variables per payment method during the whole sample period.
Table II presents the descriptive statistics of the sample before and during the crisis.
As can be seen in Table I some variables differ across the different payment methods. When looking at the first main variable of interest, Tobin’s Q, the investment opportunities for acquirers using a mix of cash and shares (Tobin’s Q average of 1.711) are better than for cash and equity paid deals. Based on the literature (Martin, 1996) this was not expected. This deviation can be caused by the low amount of observations for mixed paid deals. When looking at the standard deviations, we see that the average of Tobin’s Q for mixed payments is also more volatile and can therefore be less accurate. The Tobin’s Q for the latter two payments methods is almost equal. 1.5757 for cash paid deals and 1.561 for equity paid deals.
Regarding control variables, relative deal size and firm size provide the biggest differences. Firm size is on average the greatest for cash paid (14.03) deals compared to equity (12.85) and mixed paid deals (12.33). Since deal value is almost equal across the payment methods, the lower average of the relative deal size for cash paid deals is reflected by this greater average of firm size. Consistent with the existing literature about the
debt/equity ratio (Faccio and Masulis, 2005), financial leverage is on average bigger for share paid deals (109.9) than for cash paid deals (102.8). Acquiring firms with higher financial leverage prefer equity financing since their access to cash is lower due to their financial constraint. Also consistent with the existing literature is the variable collateral. The debt Table I: Descriptive statistics for all explanatory variables sorted by payment method
The table presents a summary of the statistics per payment method during the sample period of January 2004 until June 2009. It shows the number of observations the mean and standard deviation per variable per payment method.
Mixed Cash Shares
N Mean St. Dev. N Mean St. Dev. N Mean St. Dev. Tobin's Q 49 1.711 0.898 250 1.575 0.727 119 1.561 0.680 Collateral 49 0.119 0.150 250 0.231 0.203 119 0.212 0.206 Relative deal size 49 0.206 0.192 250 0.0731 0.107 119 0.167 0.192 Financial leverage 49 92.71 114.4 250 102.8 122.0 119 109.9 143.9 Control 49 38.53 21.64 250 42.41 21.93 119 43.10 23.43 log(Firm size) 49 12.33 2.028 250 14.03 2.141 119 12.85 2.244 Deal value 49 1474438 7922323 250 243860.1 1030953 119 677722.9 4259688 Source: Datastream
21
capacity of the acquiring firm is rising with tangible assets. Therefore collateral is on average higher for cash paid deals. Control has minimum dissimilarities across the payment methods.
Table II: Descriptive statistics for all explanatory variables sorted by payment method for two subsamples: before and during the crisis
The table presents a summary of the statistics per payment method during the subsample period of January 2004 until June 2007 for subsample A and during the subsample period of July 2007 until June 2009 for subsample B. It shows the number of observations the mean and standard deviation per variable per payment method.
Subsample A: Descriptive statistics before the crisis
Mixed Cash Shares
N Mean St. Dev. N Mean St. Dev. N Mean St. Dev.
Tobin's Q 31 1.890 0.911 192 1.615 0.733 75 1.708 0.748
Crisis 31 0 0 192 0 0 75 0 0
Tobin's Q*Crisis 31 0 0 192 0 0 75 0 0
Collateral 31 0.140 0.166 192 0.255 0.213 75 0.230 0.216
Relative deal size 31 0.195 0.180 192 0.0742 0.110 75 0.168 0.197 Financial leverage 31 88.76 117.6 192 103.8 125.0 75 99.38 145.4
Control 31 35.46 20.63 192 42.39 22.02 75 40.44 23.61
log(Firm size) 31 12.41 2.349 192 14.19 2.219 75 12.70 2.291 Deal value 31 2215131 9931089 192 265132.8 1136477 75 289095.3 853247.9
Subsample B: Descriptive statistics during the crisis
Mixed Cash Shares
N Mean St. Dev. N Mean St. Dev N Mean St. Dev.
Tobin's Q 18 1.404 0.807 58 1.440 0.699 44 1.311 0.451
Crisis 18 1 0 58 1 0 44 1 0
Tobin's Q*Crisis 18 1.404 0.807 58 1.440 0.699 44 1.311 0.451
Collateral 18 0.0815 0.113 58 0.152 0.141 44 0.181 0.188
Relative deal size 18 0.225 0.214 58 0.0694 0.0999 44 0.164 0.184 Financial leverage 18 99.49 111.6 58 99.55 112.8 44 127.8 141.1 Control 18 43.81 22.90 58 42.49 21.82 44 47.64 22.67 log(Firm size) 18 12.21 1.359 58 13.51 1.778 44 13.11 2.161 Deal value 18 198799.3 665975.3 58 173440.1 555467.6 44 1340156 6915761 Source: Datastream
Table II shows some remarkable differences between the two subsamples. Firstly, looking at
Tobin’s Q, the Tobin’s Q is on average higher for firms before the crisis than during the crisis. Correspondingly with the existing literature (Campello et al., 2010) this indicates that firms had on average better investment opportunities before the crisis than during the crisis. The average of Tobin’s Q for cash and share payments during and before the crisis shows a possible deviation from hypothesis 2.2. Before the crisis Tobin’s Q is bigger for share payments (1.708) than for cash payments (1.615). This supports hypothesis 2.1 and Martin’s (1996) statement about investment opportunities and equity payments. However, during the crisis, Tobin’s Q is bigger for cash payments (1.440) than for equity payments (1.311). This
22
indicates that the relationship between investment opportunities and the method of payment might have changed during the crisis.
Regarding the control variables, collateral provides averages most contrary to the existing literature. As described in section 3, Faccio and Masulis (2005) stated that fewer tangible assets had a positive effect on equity payments. Therefore a lower average of collateral on share paid deals is expected. However, collateral is on average bigger for equity payments (0.181) than for cash paid deals (0.152) during the crisis. The relatively larger average of financial leverage for share paid deals during the crisis again supports the findings in the existing literature about the debt/equity ratio (Faccio and Masulis, 2005). In spite of that, financial leverage seems to matter less before the crisis. Cash paid deals (103.8) even have a higher financial leverage average compared to equity paid deals (99.38). Looking at the deal specific variables, the average deal value decreased during the crisis for all sorts of payments. Relative deal size is also decreased across all categories during the crisis. This decrease can be caused by the lower average of deal value.
4.3 Correlations between variables
Before the Probit regressions can be performed, the correlation matrix of the variables is presented in Table III.
The correlation between crisis and the method of payment of 0.1485 already suggests a positive relation between the equity and mixed paid deals and the crisis. Tobin’s Q seems to have a positive relation with the method of payment as well. Consistently with hypothesis 2.2 the interaction term appears to relate in a positive way to the method of payment.
As can be seen in Table III, only a few correlation coefficients are high enough for us to consider implications related to multicollinearity. First, the correlation between the deal value and relative deal value should be explored in more detail. A high correlation between these two variables could be expected since the deal value is part of the computation of the relative deal value. The correlation of 0.2603 seems high, but since it is still below 0.8 it is conceivable that the coefficients of relative deal value and deal value are still reliable.
Secondly, the correlation between collateral and firm size (0.4442) and financial leverage and firm size (0.2701) are above average. For these correlations the same holds as for the deal value and relative deal value correlation. Both are below 0.8 which makes it plausible to state that the correlations between the explanatory variables will not give any problems in the regressions.
23
MoP Tobin's Q Crisis log(Firm size) Control Fin.Lev Rel. Size Deal value Collateral
MoP 1 Tobin's Q 0.0203 1 Crisis 0.1485 -0.1725 1 log(Firm size) -0.2900 -0.1613 -0.0924 1 Control -0.0142 -0.0444 0.0691 -0.1636 1 Financial Leverage 0.0080 -0.1916 0.0312 0.2701 0.0335 1
Relative deal size 0.3311 -0.1368 0.0499 -0.1045 0.1199 0.0958 1 Deal value 0.0902 -0.0206 0.0164 0.2325 -0.0010 -0.0022 0.2603 1 Collateral -0.1140 -0.1448 -0.1914 0.4442 0.0675 0.1460 -0.0559 0.0455 1
24
Firstly, the results of the regressions in Table IV will be interpreted. Afterwards, two additional tests will be provided in order to assess robustness.
5.1 Empirical results of the Probit regressions
Table IV presents the results of the three regressions that were performed. The first column
contains the regression on the whole sample period, where the second and the third column include the regressions using data only on the period during the crisis and the period before the crisis, respectively. For each regression the same explanatory variables are used.
Regressions 2 and 3 do not provide any coefficients on crisis and the interaction term Tobin’s Q*crisis, due to collinearity. This is expected since crisis is respectively 1 and 0 for all
observations in these two regressions.
Table IV: Probit model explaining the choice of payment This table shows the performed probit regressions. The dependent variable is in all regressions the method of
payment, taking on the value 1 if the payment method is either equity or a mix of cash and equity and zero otherwise. Column 1 presents the regression performed on the total sample, where column 2 and 3 present the regressions performed on the period during the crisis(2) and the period before the crisis(3). The z-statistics are based on the pseudolikelihood with robust standard errors.
(1) (2) (3)
Dependent variable Method of Payment Method of Payment Method of Payment Observations 418 120 298
Coeff. z-Stat. Coeff. z-Stat. Coeff. z-Stat.
Tobin's Q 0.206** 2.00 -0.185 -0.94 0.180* 1.73 log(Firm size) -0.255*** -5.81 -0.230*** -2.91 -0.265*** -5.06 Control -0.00799** -2.49 -0.00366 -0.60 -0.0105*** -2.71 Financial leverage 0.000940* 1.72 0.00202 1.62 0.000431 0.75
Relative deal size 2.599*** 4.67 3.179*** 3.59 2.536*** 3.85
Collateral 0.707* 1.80 1.070 1.18 0.677 1.53
Deal value 1.36e-07 1.41 3.88e-08 1.51 1.49e-07 1.32
Crisis 1.129*** 3.14 Tobin's Q*Crisis -0.461** -2.12 Constant 2.414*** 4.01 2.710** 2.42 2.750*** 3.91 Pseudo R-squared 0.1751 0.1474 0.1930
***)Significance at the 1% level, **)Significance at the 5% level, * )Significance at the 10% level
25
Corresponding with hypothesis 1, column 1 shows a significant positive effect at the 1% level of the crisis towards the probability to pay with equity for M&A transactions in the Euro countries. This means that there has been a significant shift in the payment method since the start of the crisis in the third quarter of 2007. The positive sign shows that this shift was towards equity paid deals. In addition, Tobin’s Q also has a significant positive effect on the method of payment. Acquiring firms with a higher Tobin’s Q tend to prefer equity and mixed paid deals over cash deals.
Contrary to hypothesis 2.2 the sum of Tobin’s Q and the interaction term is negative (0.206 – 0.461 = -0.255) for the whole sample. This indicates that the effect of Tobin’s Q changed during the crisis. Martin’s (1996) statement that: ‘the higher the acquirer's growth opportunities, the more likely the acquirer is to use stock to finance an acquisition’ does not hold during the crisis. This is also shown by looking at Tobin’s Q in the second and third regression. While before the crisis Tobin’s Q has a positive effect on the preference of equity and mixed payments over cash payments, during the crisis this effect changed to a negative effect. The insignificance of the negative effect of Tobin’s Q in regression 2 can be caused by a lack of observations. Nonetheless, by keeping in mind the significance of the interaction term in regression 1 and the small amount of observations in regression 2 this negative effect is assumed to be reliable, despite the insignificance. This deviation from hypothesis 2.2 can be caused by the difference in access to external funds between high Q firms and low Q firms. One of the main conclusions from the literature review was that the choice of payment depends most on the relative gains and costs of paying with either cash or equity. From the theoretical framework of Mishkin (1992) it can be derived that high Q firms had easier access to external funds or that low Q firms had no access at all. This could indicate that the relative gains of using cash instead of equity might have changed as well. Apparently, it might be that the higher the investment opportunities (the higher Q) of the acquiring firm, the higher the relative gains for using cash and the more likely the acquiring firm became to use cash as method of payment during the crisis. On the other hand, the lack of access to external debt for low Q firms could have been driving the contrary result as well. According to Whited (1992) this lack of external debt caused low Q firms to finance their M&A’s with equity. The results imply that during the crisis, as a firm’s investment opportunities increased (high Q firms), firms were less likely to pay with equity since the sum of the coefficients of Tobin’s Q and the interaction term is negative. This at the same time means that as a firm’s Tobin’s Q
26
diminished (low Q firms) it became more likely that it indeed would fund its acquisition with equity.
Table IV suggests that of the control variables only relative deal size and firm size
have significant influence on the method of payment. Firm size is decreasing the probability to pay with equity, while the relative deal value is increasing this probability. This is
consistent with the findings of Martin (1996) and Faccio and Masulis (2005). They both found evidence for a negative relation of the acquiring firm’s total assets and the probability to pay with equity. Faccio and Masulis (2005) found evidence for a negative relation of the relative deal size with cash paid deals. During the crisis the relative deal size became more important in determining the payment method. As can be seen in column 2 and 3, the coefficient for the relative deal value increased (3.179) during the crisis relative to before the crisis (2.536). This corresponds to a positive relation of the relative deal size with equity paid deals. The variable control has a significant negative effect while computed during the whole sample period or before the crisis. This corresponds with the findings of Amihud et al (1990), who found evidence for a negative influence of the fraction of voting rights held by the management of the acquirer on equity payments.
Despite the effects of financial leverage found by Faccio and Masulis (2005), this thesis found no strong evidence of the influence of this variable on the payment method. Faccio and Masulis (2005) state that the higher the debt/equity ratio, the more likely the acquirer is to use equity as payment method. Only in the regression performed on the whole sample a small positive significant (at the 10% level) influence of financial leverage is found. If we look at the value of the coefficients regarding financial leverage, we see a difference between the two subsamples. The second column shows that the effect of financial leverage on the probability to pay with equity increased during the crisis. A highly leveraged firm was even more likely to prefer equity financing during the crisis than before.
One of the control variables that differ from the existing literature is the result for collateral. Martynova and Renneboog (2008) and Faccio and Masulis (2005) found evidence for a negative influence of the ability to borrow on preference of equity payments. Again, this thesis only finds limited evidence for on collateral in the regression on the complete sample. Contrary to the existing literature this thesis finds some evidence for a positive relation between collateral and the preference of equity payments. The explanation for this can be the different calculation of collateral in this thesis as compared to the existing literature.
27
both, the target and the acquiring firm in the papers of Faccio and Masulis (2005) and Martynova and Renneboog (2008). This thesis deviates from this computation because only the acquiring firm is restricted to being a listed firm. Hence, no information about tangible assets and total assets is available for the unlisted target. Like financial leverage, the effect of collateral on the probability to pay with equity increased during the crisis.
Comparing the results of the performed regressions to the hypotheses, evidence is found to support hypothesis 1 and 2.1. However, the results provide evidence that is contrary to hypothesis 2.2.
28
5.2 Robustness tests
The dependent variable used in the previous regressions equaled 1 for equity only payments and mixed payments. The amount of equity in the mixed payment transactions is hard to measure. Therefore, a mixed payment can consist of 99% of cash and 1% of equity. Since this thesis investigates the effect on the probability to pay with equity only, a regression on equity only and cash only would test the robustness of the previous regressions. Table V presents the results for this robustness test.
Table V: Robustness test on the performed regressions This table shows the performed Probit regressions adjusted for the mixed payment method. The dependent variable
is in all regressions the method of payment, taking on the value 1 if the payment method is equity only and zero otherwise. Column 4 presents the regression performed on the total sample, where column 2 and 3 present the regressions performed on the period during the crisis(5) and the period before the crisis(6). The z-statistics are based on the pseudolikelihood with robust standard errors.
(4) (5) (6)
Dependent variable Method of Payment Method of Payment Method of Payment Observations 369 102 267
Coeff. z-Stat. Coeff. z-Stat. Coeff. z-Stat.
Tobin's Q 0.178 1,58 -0.264 -1.24 0.144 0.210
log(Firm size) -0.233*** -5.04 -0.205** -2.44 -0.246*** 0.000
Control -0.00644* -1.87 -0.00204 -0.32 -0.00930** 0.026
Financial leverage 0.000910 1.60 0.00226* 1.70 0.000229 0.716
Relative deal size 2.340*** 4.17 2.885*** 2.97 2.334*** 0.000
Collateral 0.921** 2.22 1.395 1.44 0.873* 0.064
Deal value 1.02e-07 1.23 4.26e-08 1.61 1.07e-07 0.255
Crisis 1.231*** 3.15
Tobin's Q*Crisis -0.541** -2.30
Constant 1.924*** 3.03 2.183* 1.88 2.354*** 0.002
Pseudo R-squared 0.1555 0.1410 0.1585
***)Significance at the 1% level, **)Significance at the 5% level, * )Significance at the 10% level
By adjusting for mixed payments we see minor changes in the original regressions. The signs of the main variables of interest, Crisis, Tobin’s Q and Tobin’s Q remain the same. Only the Tobin’s Q is not significant anymore in all new regressions. This difference in significance can be caused by the fewer observations. Excluding mixed payments decreased the total observations by 49 which is equal to a decline of almost 12%. If we look at the pseudo R-squared, the values for the mixed payment adjusted regressions are lower in all cases. This can indicate that the original regressions include more precise information about the method
29
of payment, which can be caused by the amount of observations. In general, the robustness test of excluding mixed payment deals does not show major changes towards the original regressions.
As stated in section 2, the existing literature focused mainly on firm specific and deal characteristics. However, Garcia-Feijóo et al. (2012) showed that the influence of firm characteristics changes along industry conditions. They state that changes in industry conditions are driven by industry-specific shocks. The results of Garcia-Feijóo et al. (2012) indicate that industry characteristics caused a shift in the distribution of payment method over time.
To control for these industry characteristics in this thesis, an additional robustness test is performed. A multinomial industry effect dummy is created to indicate whether an M&A transaction is completed in a particular industry. The industries are categorized on their Standard Industrial Classification code (SIC-code). This code is derived from the Zephyr database for every deal. Table VI presents the included industries and their corresponding dummy indicator.
The robustness test is performed on the regression for the total sample. Regressions adjusted for industry effects for the periods before and during the crisis provided incomplete information, due to the smaller amount of observations.
Table VII presents the results of the Probit regression of the original model adjusted for
industry effects.
Table VI: Industries sorted by SIC Code
Industry SIC Code
Agriculture, Forestry and Fishing 100 - 999
Mining 1000 - 1499
Construction 1500 - 1799
Manufacturing 2000 - 3999
Transport, Electric, Communication, Gas and
Sanitary service 4000 - 4999
Wholesale trade 5000 - 5199
Retail trade 5200 - 5999
Finance, Insurance and Real Estate 6000 - 6799
Service 7000 - 8999
30
Table VII shows the remaining significance of Crisis, Tobin’s Q and the interaction term.
When controlling for industry effects, the impact of the financial crisis, Tobin’s Q and support hypotheses 1 and 2.1. Similar to the found change in section 5.2 of the effect of Tobin’s Q during the crisis, the interaction term has a negative sign which is different from hypothesis 2.2. Only when controlling for these effects Tobin’s Q becomes less significant (Significant at the 10% level instead of the 5% level). The pseudo R-squared is increased comparing to the original regression. This indicates that by controlling for industry effects the new regression is a slightly better fit of the model. This could be seen as that by adding industry effects there is more precise information regarding the method of payment.
Overall, adding industry effects or adjusting for mixed payments provide minor changes to the impact of Tobin’s Q and the financial crisis on the payment method.
Hypothesis 1 and 2.1 are still supported, while the results are still in contrast with hypothesis 2.2.
Table VII: Robustness test on the performed regressions adjusted for industry effects This table shows the performed Probit regressions adjusted for industry effects. The dependent variable is the
method of payment, taking on the value 1 if the payment method is equity only or mixed and zero otherwise. The z-statistics are based on the pseudolikelihood with robust standard errors.
(7)
Dependent variable Method of Payment
Observations 418 Coeff. z-Stat. Tobin's Q 0.192* 1.72 log(Firm size) -0.274*** -5.71 Control -0.00798** -2.37 Financial leverage 0.000536 0.84
Relative deal size 2.653*** 5.14
Collateral 0.720* 1.71
Deal value 1.35e-07* 1.91
Crisis 1.081*** 2.88 Tobin's Q*Crisis -0.437* -1.91 Constant 2.418*** 3.12
Industry effects YES
Pseudo R-squared 0.2064
31
6. Conclusion and limitations
6.1 Conclusion
Using data on 418 mergers and acquisitions in twelve European countries completed between January 2004 and June 2009, this thesis proves that the financial crisis of 2007 significantly affected the amount of equity payments for mergers and acquisitions in Europe. This thesis also provides some evidence of a positive influence of investment opportunities on equity payments. Moreover, the effect of the investment opportunities appears to have changed significantly during the crisis.
The existing literature has mainly focused on firm- and deal specific characteristics determining the method of payment for M&A transactions in the United States. This thesis contributes to the existing literature by focusing on an external factor, the crisis, next to firm- and deal characteristics. Furthermore, this thesis focused on European M&A’s since research on the payment method of European M&A transactions is limited. Another main contribution is the combination of the influence of the crisis and the investment opportunities of the acquiring firm.
The statement of Martin (1996) that ‘the higher the acquirer's investment
opportunities, the more likely the acquirer is to use stock to finance an acquisition’ is found to hold during the period before the crisis. However, during the crisis the effect of investment opportunities on the probability to pay with equity seems to have changed to a negative effect. This is in contrast with hypothesis 2.2 and the finding of Campello et al. (2010). They found that acquiring firms with less access to cash and better investment opportunities are likely to prefer to pay with equity. The dramatic decrease in bank lending during the crisis was shown to have reduced cash availability to acquiring firms by Ivashina and Scharfstein (2010). Furthermore, Whited (1992) proved that lower cash availability during the crisis induced the more constrained acquiring firms to look for other sources to finance their investments. This was expected to increase the probability to finance M&A’s with equity for low Q firms as well. As a consequence, the effect of investment opportunities on the probability to pay with equity was expected to weaken but remain positive.
The results contrary to this expectation can be explained by the higher relative gains of cash financing for high Tobin’s Q firms or by the increasing likelihood of low Q firms
preferring equity funding during the crisis. Campello et al. (2010) found that the differences between firms increased during the crisis. Regarding the results a main difference during the
32
crisis could be between high Q and low Q firms. High Q firms had easier access to external funds, compared to low Q firms, during the crisis (Mishkin, 1992). A possible result of this difference in cash access between high Q and low Q firms could be that the relative gains of using cash instead of equity for high Q firms have changed. Therefore, having better
investment opportunities during the crisis decreases the probability to fund M&A transactions with equity. Another possibility is that low Q firms had no access to external funds at all during the crisis and had to look for other sources of funding. Therefore, the lower a firm’s investment opportunities, the more likely they were to fund its M&A’s with equity. This supports the negative effect of having better investment opportunities on the probability to pay with equity.
The significant increase in equity payments during the crisis supports the existing literature on the payment method. Based on this literature the choice of payment depends most on the relative gains and costs of paying with either cash or equity. According to Swieringa and Schauten (2007), the choice between paying with cash and paying with equity is mainly the choice between debt-funding and equity-financing. Since the cost of debt increased during the crisis (Ivashina and Scharfstein, 2010, Gan, 2007, Campello et al. 2010), the increase of equity payments during the crisis was expected regarding the literature.
When controlling for industry effects and excluding the mixed payments from the sample, the results appear to remain significant. Only in the case of the exclusion of the mixed payments, the effect of Tobin’s Q became insignificant. The reason for this occurrence can be the decline in observations by 49, which equals almost 12% of the total sample.
Based on the results, it can be concluded that, even when controlling for industry effects and excluding mixed payments, there is a change in payment method for European mergers and acquisitions during the financial crisis of 2007. This thesis also provides
evidence for a positive influence of investment opportunities on the likelihood of paying with equity. However, most remarkable is the change found in effect on the payment method of the investment opportunities during the crisis.
33
6.2 Limitations of the thesis and possibilities for future research
The first limitation of this analysis was the data availability. After collecting data on the payment method for the mergers and acquisitions completed during the sample period, firm specific data for the acquirer was required. For most variables this data was not complete. After excluding all deals with incomplete data, the sample size decreased by more than 50% from 895 to 418. This smaller amount of observations caused some problems in determining the significance of variables. The robustness test with excluded mixed payments on the
difference between the two subsamples was not very reliable since the observations during the crisis decreased to only hundred.
Another limitation is the dependent variable used. The Probit model needs a two sided dependent variable. Therefore, mixed payments are indicated as an equity payment as well, which could bias the results. Faccio and Masulis (2005), Swieringa and Schauten (2007) and Martynova and Renneboog (2008), used Orderd Probit and Tobit models to determine the effects on the payment method. The approach used in this thesis deviates from the existing literature, because the Probit model is assumed to be better interpretable. Furthermore, as stated in section 5.2, excluding the mixed payments decreased the observations and influenced the significance of one of the main variables of interest.
The results of this thesis showed one major interesting point for future research. A good contribution could be the further exploration of the unexpected change in the effect of investment opportunities on the probability to pay with equity during the crisis. Campello et al. (2010) stated that firm’s financial situations differ more during the crisis. Hence, high Q firms and low Q firms seem to differ more during the crisis. The effects of the crisis on the payment method could also differ between high Q and low Q firms. Research on the different impact of the crisis for these firms could be interesting.
Furthermore, to explore the influence of the crisis on the payment method in more detail, the crisis could be investigated in combination with more specific variables than investment opportunities.
Finally, another suggestion for future research is the investigation of other crises. In order to fully understand the choice of payment during crises, the main similarities in determinants or changes in effects of determinants can be explored.
