Mergers and acquisitions: Do interest
rates influence the method of payment
of mergers and acquisitions in the U.S.?
Name: Paul Koopman
Student number: 10642366
BSc Business Economics – Finance & Organization
Bachelor thesis, University of Amsterdam
Statement of Originality
This document is written by Student Paul Koopman who declares to take full responsibility for the contents of this document.
I declare that the text and 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.
Abstract
This paper analyzes the relation between interest rates and the method of payment of mergers and acquisitions. The hypothesis of this paper is that an increase of the interest rate results in a lower probability of cash-financed mergers and acquisitions. Two probit regression models are used to predict this effect. The sample consists of mergers and acquisitions that took place in the period of 1997 – 2015 in the United States. As explanatory variables the effective Federal Funds rate and the Wu-Xia Shadow rate are used. The Wu-Xia Shadow rate and the Federal Funds Rate Squared are in line with the hypothesis. However, the Federal Funds rate is predicted in contrast.
Table of Contents
1. Introduction ... 5
2. Literature review and underlying theory ... 7
2.1 Capital structure ... 7
2.2 Interest rates ... 7
2.2.1 Interest rates and capital structure ... 8
2.3 Mergers and Acquisitions... 9
2.3.1 Payment methods of M&A ... 9
2.3.2 Preferred methods of payment ... 10
2.4 Previous studies ... 11
2.5 Hypotheses ... 13
3. Methodology and data description ... 14
3.1 Methodology ... 14
3.1.1 Probit Regression Model ... 14
3.1.2 Regression model ... 15
3.1.3 Summary of variables ... 16
3.1.4 Dependent variable ... 16
3.1.5 Explanatory and control variables ... 17
3.2 Data description ... 19
4. Results and analysis ... 20
4.1 Descriptive statistics ... 21 4.2 Correlation matrices ... 22 4.3 Results analysis ... 24 4.4 Robustness check ... 25 5. Conclusion ... 27 References ... 29
1. Introduction
Mergers and acquisitions (hereafter M&A) cover a big part of the U.S. economy. In 2015 alone, all transactions involving M&A are estimated at roughly 3.8 trillion USD (Bloomberg, 2016). This makes M&A one of the most important decisions managers can make. Because these decisions are so important for the future of a company, the method of financing M&A has a severe impact on this future. Several factors have influence on the way firms prefer to finance their takeovers. It is argued that the interest rate is such a factor. Some studies prove that the interest rate has a relation with the way firms finance their takeovers, while others do not. After the financial crisis of 2007, the monetary policies of governments changed dramatically. As the monetary policy executed by the Federal Reserve Bank creates interest rates levels which are historically low, it is relevant to examine how firms adjust the decision in financing mergers and acquisitions.
Mergers and acquisitions can be financed in different ways, for example with cash, shares, debt and loan notes1. Borrowing debt or issuing loan notes are seen as receiving cash now in exchange for a future repayment of this cash in addition with interest payments. Because the opportunity costs in using cash or leverage is equal to each other, this paper mainly focuses on two different methods of payment: cash-financing and equity-financing.
The relation between interest rates and the method of payment of an investment is a widely researched topic. Marsh (1982) finds a significant, and negative, effect of the interest rate on the use of cash instead of equity: firms borrow less in times of high interest rates. In addition, Taggart (1977) does also find a significant effect of the interest rate on the corporate financing decisions of firms. The conclusion is that firms are more likely to issue equity when the price of their shares is relatively high, and issue debt (which makes an investment cash-financed) when the interest rates are low. However, in contrast to Marsh (1982) and Taggart (1977), Choe, Masulis and Nanda (1992) do not find evidence that the effect of the interest rate is significant on the use of cash or equity within an M&A or other investments.
1
This paper will analyze the effect of the low interest rates on the method of payment of M&A. The research question is: do interest rates influence the method of payment of mergers and acquisitions? The main variables of interest are the Federal Reserve Effective interest rate and the Wu-Xia Shadow rate. The Wu-Xia Shadow rate is a rate which resembles the effective Federal Funds rate in an environment where the interest rate can go below the lower zero bound. In addition, several other factors will be included as control variables. These control variables consist of business cycle variables and firm-specific variables. At last a robustness test is added for industry specific characteristics of the acquirer. The hypothesis is as follows: when the interest rate is low, the use of cash and debt will be more likely. Because different studies have shown different outcomes, it is interesting to examine if the historically low interest rates have an effect on the payment method they use within an important investment like mergers and acquisitions.
The method used in this paper is a probit regression model. The probit regression model is a model with a binary dependent variable. The probit model will predict the influence of the effective Federal Funds rate and Wu-Xia Shadow rate on the probability that a firm will finance their M&A with either cash or equity.
The dependent variable consists of cash-financed M&A if at least 50 percent of the deal value is financed with cash. If at least 50 percent of the deal value is financed by offering shares, it will be regarded as equity-financed. The regression models will contain two explanatory variables, and nine control variables. A dataset of 6,293 observations will be used. All announcement and completion dates are in the period of 1997 – 2015.
The results from this research are somewhat peculiar. The Wu-Xia Shadow rate and the effective Federal Funds Rate Squared are in line with the hypothesis. This means that the Wu-Xia Shadow rate and the effective Federal Funds Rate Squared predict an increase of cash-financed mergers when the rates are lower. However, the effective Federal Funds rate is predicted in contrast. This might be due to a non-linear effect of the interest rate, or due to theory not being accurate in times of a (post) financial crisis.
The structure of this thesis will be as follows. First, the related literature and underlying theories will be explained, followed by the hypotheses. Then the methodology and data will be explained. Hereafter, the results and the interpretation of these results are stated. The last paragraph will contain the conclusion.
2. Literature review and underlying theory
As stated above, there is already some related literature about the method of payment of an investment. In this section, an overview of the main empirical literature of this topic will be provided. First, an overview of capital structure will be given, followed by interest rates and mergers and acquisitions. Hereafter, recent studies are reviewed. At last, the hypotheses from the literature will be stated.
2.1 Capital structure
Modigliani and Miller (1958) provided the first rigorous theory about methods of payment, cost of capital and capital structure. They stated that in a perfect capital market the different way of financing investments has no effect on a firm. Myers (1977) stated that in such a market, issuing risky debt will lower the market value of a firm due to a suboptimal future strategy. The shareholders will absorb the loss in market value. However, there are various frictions in markets which can be affecting payment decisions. For example, adding corporate taxes on equity changes the way firms choose their level of leverage in capital structures and in finance decisions of investments, as the cost of capital changes. An increase of leverage will increase the deductibility of interest payments on tax payments which increases the value of the firm. But after a certain level of leverage, it will result in higher distress costs and higher chance of default. The cost of default consists of transaction costs of the reorganization or liquidation of a company (Myers, 1977). A result of this cost of default is that this cost affects the future cash flows and future investment decision making, which has impact on the way firms perform.
2.2 Interest rates
Interest rates can be regarded as the ‘price’ of borrowing money. Merton (1974) states that the risk structure of the interest rates of corporate debt depends on different factors. One of them is the rate of return required on debt. It is essential for firms to be aware of how much the interest payments are for borrowing certain amounts of money. For Central Banks interest rates are a commonly used tool for shaping monetary policy. As the financial crisis hit the United States, from December 2008 the Federal Reserve Bank modified the Federal Fund Rate to a percentage which reached the zero bound to maintain economic growth.
When the interest rate reached the zero bound, the effect of the interest rate to maintain the monetary policy becomes ineffective. It is possible for people and firms to hold currency at a zero level of interest rate. This makes it impossible for central banks to adjust the short term nominal interest rate at a lower point (Black, 1995). Wu and Xia (2015) built a model which estimates Federal Fund rates in a period of zero level interest rates. This interest rate is crossing the lower bound in a hypothetical way with the use of predicted forward rates. Due to these estimates, the Wu-Xia Shadow rate made it possible to describe yield curves below zero. The Wu-Xia Shadow rate could be interpreted as the effective Federal Funds interest rates if it was possible for the Federal Reserve Bank to adjust to negative interest rates. In Figure I the effective Federal Funds rate and the Wu-Xia Shadow rate are showed. It can be seen that from January 1997 – December 2008 the effective Federal Funds rate is equal to the Wu-Xia Shadow rate. From January 2009 – November 2015 the Wu-Xia Shadow rate is predicted below the lower bound.
Figure I
where the Effective rate is the effective Federal Funds interest rate and the Wu-Xia rate is the Wu-Xia Shadow Rate. The vertical axis shows the interest rate in percentages, the horizontal axis the timeline in years.
2.2.1 Interest rates and capital structure
Low interest rates have an effect on the way firms finance their activities and investments due to the cost of capital. Also the capital structure of firms matters for future cash flows and decision making of a firm. The cost of capital influences firms in which levels of debt and equity they can and should issue to reach the optimal capital structure. The cost of capital is
often calculated with the Weighted Average Cost of Capital method (WACC-method). The WACC-method calculates an appropriate discount rate for valuating investments (Miles and Ezzel, 1980). In Equation 1 the WACC-method is stated.
(1)
where MV is market value, is the cost of equity, the cost of debt and the corporate tax rate. The cost of debt is regarded as the interest rate lenders are requiring for the bonds or bills their issuing. Accordingly, a change in the interest rate would affect the present value of investments. Theoretically, this should have an effect on the way managers finance their investments like mergers and acquisitions.
2.3 Mergers and Acquisitions
Mergers and acquisitions are a kind of investment where an acquirer (buyer) merges with or takes over a target (seller). Mergers results in two different parties merging with each other, while having no explicit owner. A takeover (acquisition) results in one firm ‘owning’ the other one. The reason why firms merge or take over another firm is to create value. This is also regarded as synergy. Synergy means that two different companies can create more value together than when they operate apart from each other. There are two different kinds of synergies: cost reduction and revenue enhancement. Cost reduction synergy occurs due to the effects of economies of scale and economies of scope. Revenue enhancement arises when firms expand to new regions or markets which will result in a larger customer network (Berk and DeMarzo, 2014).
2.3.1 Payment methods of M&A
As stated in the introduction, there are different kinds of payment methods to finance a merger or acquisition. For example: cash, shares, debt, and loan notes. In this paper there will be a separation between cash-financed M&A and stock-swap (equity) M&A. A discussion about the differences of the two methods is stated below.
Cash-financed M&A are investments in which the transaction is financed directly with cash and can be regarded as a simple transfer of ownership of a company in exchange for cash. Cash means internally generated cash, but also issued debt and loan notes. Because the opportunity cost of leverage and cash are the same, it can be regarded as one method of
payment. The main distinction from equity financing is that with cash financing the investors bears the entire risk of the deal. The risk consists of uncertainty if value of the premium and synergies will materialize (Rappaport and Sirower, 1999).
Stock-swap M&A are investments which are paid by shares of the acquirer. This is also called as equity financing. An acquirer will swap shares of their company for shares of the targets company with a pre-arranged exchange rate. Bidders can be uncertain though, about the value of the target. Therefore equity financing forces shareholders of the target to share the risk that the bidder might have paid too much for a company (Martin, 1996). Also shareholder returns are affected by the choice of stock payment. It is found that investors of acquiring companies are worse off in stock transactions than they would be in cash transactions (Rappaport and Sirower, 1999).
2.3.2 Preferred methods of payment
Firms and managers do prefer certain payment methods over another. This is the so called Pecking Order Theory (Myers, 1984). The pecking order theory states that firms prefer financing with internally generated funds rather than debt, and financing with debt rather than with equity. In addition, in the choice of payment managers are assumed to be better informed than potential investors (Myers and Majluf, 1984). This means that managers may refuse to finance their activities with internally generated funds like cash instead of equity due to lose of control when issuing more equity. This can affect investment decisions and future cash flows of the firm.
In addition to the Pecking Order Theory, two theories about preferred methods of payment are to be mentioned. Wansley, Lane and Yang (1983) find that stock payments are most of the time tax deferred investments. Any proceeds of capital gains realized by the investors of the target firm are not taxed. Levying tax will only happen when the stock investors receive is sold. Investors are aware of this, so the acquiring firm has to pay a higher price in case of a cash payment to compensate them.
The choice in payment methods does also give a signal to the investors. Travlos and Papaioannou (1991) are referring to Jensen and Ruback (1983), which states that a cash payment means that managers believe their firm has an undervalued share price, while
stock payments means otherwise. This results that investors interpret a cash payment as a good signal and managers can react to this behavior.
2.4 Previous studies
Martin (1996) has researched motives behind the method of payment of M&A in the United States. The regression method was a logistic regression analysis on factors which could affect the method of payment in corporate takeovers. Martin (1996) used four different regressions with different dependent variables. The factors consist of bidder characteristics, ownership structure, ability to finance with cash and the mode of the takeover. In addition, several business cycle variables were added. The final sample contained 846 M&A which all took place in the United States and of which the transactions were completed in 1987 - 1988. In the results, Martin (1996) finds that many characteristics are significant and thus influencing the method of payment. The interest rate has a positive coefficient and is significant in the regression with stock and mixed financing results. This means that if the interest rate is increasing, firms are more likely to finance with stock instead of a stock/cash mix. Tobin’s Q has also a positive coefficient, which means that firms with better growth opportunities are less likely to use cash instead of shares to finance their takeover. The amount of cash flow, which is calculated by dividing the cash balance with amount paid for the takeover, has a negative coefficient. This means that firms with greater cash flows are more likely to finance with cash. Both, Tobin’s Q and the amount of cash flow, are significant. Characteristics as financial leverage and relative size of the deal are not significant.
Faccio and Masulis (2005) analyzed the M&A payment decisions in Europe in the period of 1997 – 2000. They find, consistent with earlier studies, that several characteristics influence the method of payment decisions of firms. Faccio and Masulis (2005) used a two-boundary Tobit estimator, with an interval of [0, 100]. The study is primary focused on control threats of the acquirer in relation to the method of payment. They find that acquirers with more corporate control prefer cash financing more than stock financing, in particular when the target shareholders are more concentrated. In addition, the characteristics leverage and relative size have both a negative coefficient. This is interpreted
as that when leverage or relative size increases, the acquirer is more likely to finance M&A with stock.
Choe, Masulis and Nanda (1993) analyzed the impact of business cycle factors on the use of common stock to finance their investments. Using a linear regression model with data from 1971 – 1991, they find that interest rates are positive, but always statistically insignificant on the frequency of equity payments. This means that the interest rate has no significant effect on the use of common stock in this regression. In contrast to interest rates, it is found that the amount of equity offers is positively related to economic growth. This means that when the economic business cycle is in an expansionary phase, equity offers are more likely.
Marsh (1982) used a logit model to analyze a sample of 748 issues of debt and equity. A sample period of 1959 – 1970 is used from United Kingdom companies only. The final model contained debt target variables, firm specific variables and market condition variables. The results conclude that market conditions do influence the choice of payment, issuing equity or debt. The interest rate is statistically significant and consistent with the hypothesis that an increase of the interest rate decreases the amount of debt used.
2.5 Hypotheses
Using the earlier studies, two different hypotheses are formed. The first hypothesis is about the relation between the effective Federal Funds interest rate and the method of payment. It is shown that different studies have different results about the significant effect of interest rates. The hypothesis can be stated as:
: 0, this means in words that a change in the interest rates is not affecting the method payment a firm chooses in a merger or acquisition significantly.
: ≠ 0, which means in words that a change in the interest rates is affecting the method payment a firm chooses in a merger or acquisition. From the theory and previous studies the expectation is that an increase in the interest rate will decrease the probability that firms will finance their M&A with cash.
In addition to the first hypothesis, a second hypothesis about the Wu-Xia interest rate is composed. The Wu-Xia Shadow interest rate is a not commonly used interest rate. As stated above, it predicts the Federal Reserve Bank interest rate when it reaches the lower bound. The second hypothesis can be stated as:
= 0, this means in words that a change in the Wu-Xia Shadow rate is not affecting the method payment a firm chooses in a merger or acquisition significantly.
≠ 0, this means in words that a change in the Wu-Xia Shadow rates is affecting the method payment a firm chooses in a merger or acquisition. The expectation regarding the Wu-Xia Shadow rate is in line as the effective interest rate: an increase of the Wu-Xia Shadow rate results in a lower probability of cash-financed M&A.
To test the hypotheses on significance, two-sided tests are performed. Since there is a difference between significance levels, a 5 percent significance level is regarded as ‘significant’ as most of statisticians and researchers use this level of denoting significance (Stock and Watson, 2007).
3. Methodology and data description
In this section the methodology and data will be explained. First the methodology of previous studies will be stated. Thereafter the probit regression and the regression model will be explained. At last, the construction of variables and data are set out.
3.1 Methodology
The methodology is composed by a selection of previous studies. As stated in the literature review, Martin (1996) used a logit regression model, Faccio and Masulis (2005) a two boundary Tobit-model, Choe, Masulis and Nanda (1993) a linear regression model and Marsh (1982) uses also a logit model. Dependent variables of these studies were all binary, which means it is measured as a probability from [0 , 1].
The linear regression model is relatively easy to use, but it is not appropriate because of the shortcomings of this model. In this model probabilities of the linear regression might go below zero or exceed one. Besides this, the linear regression model is not able to capture nonlinearity of the true population (Stock and Watson, 2007).
In order to prevent these problems, other researchers use the logit model. The logit model is a cumulative standard logistic model. The logit model has similar outcomes as the probit model. The main motivation to use the logit model is that it was easier to calculate. Nowadays, this restriction is no longer relevant due to more efficient computations. Additionally, the probit model has slightly thinner tails than the logit model, which means the outcomes differ as the probability reaches a level of 0 or 1. Because of these reasons, the model of this research will be a probit model (Stock and Watson, 2007).
3.1.1 Probit Regression Model
The probit regression is a nonlinear model for binary dependent variables where the dependent variable, Y, is a binary dependent variable. The probit model is:
) (3.1)
where is the probability that Y = 1, given the values of the regressors Xk. is the cumulative standard normal distribution function. The part
distribution model. This part explains the effect of a change in Xk to a change in the probability of Y. The probit coefficients of this model are estimated by the maximum likelihood estimator (MLE). The MLE produces estimators which are efficient, consistent and normally distributed in large samples (Stock and Watson, 2007). To prevent heteroskedasticity and serial correlation, all standard errors are predicted with the robust option.
As these values are predicted with the probit regression model, the values cannot be interpreted as straightforward as coefficients of linear regression models. After retrieving the values of the coefficients, it is necessary to compute the marginal effect of the explanatory variables (Hoetker, 2007). Since the explanatory variables are continuous variables, the marginal effect can be interpreted as a change of one unit of the explanatory variable, influences the probability of the dependent variable by X. X is the marginal effect computed by Stata.
3.1.2 Regression Model
The regression functions are stated in equations 3.2 and 3.3. These regressions will be inserted in the probit equation 3.1 to analyze the dataset. The difference between equation
3.2 and 3.3 regression is the explanatory variable. The first regression model, 3.2, is stated
below:
(3.2)
In regression 3.2 the explanatory variable is the effective Federal Funds interest rate, while the main explanatory variable in 3.3 is the Wu-Xia Shadow rate. The variable INTERESTRATE2 is the squared variable of the variable INTERESTRATE to capture non-linear results. Due to multicollinearity between the effective Federal Funds interest rate and the Wu-Xia Shadow rate, it is necessary to estimate two different regression models. The second regression model is stated below:
(3.3)
where the method of payment is the binary dependent variable, and the WUXIARATE is the explanatory variable. While the INTEREST variable is also included as a squared variant, this is not possible for the WUXIARATE because of the negative values. The other variables are control variables. The construction and the expected relation will be explained in detail in the next sections.
3.1.3 Summary of variables
A summary of the variables used in the empirical regression is stated in Table I. In addition, a short description is given. The binary dependent variable is the method of payment. All variables and construction of these variables are explained in detail in section 3.1.4 and
3.1.5. Data sources are described in detail in section 3.2. Table I
Variables Source* Description
METHOD OF PAYMENT Zephyr Dummy variable: 1 if cash or 0 if equity financed INTERESTRATE FED Monthly effective Federal Funds rate in %
INTERESTRATE2 FED Squared monthly effective Federal Funds rate in %
WUXIARATE FED Monthly Wu-Xia Shadow rate in %
ECONOMICGROWTH YahooFinance S&P500 daily returns
TOBINSQ Datastream Opportunity of firm to grow (market-to-book ratio) LEVERAGE Datastream Financial leverage of company in %
CONTROL Datastream Closely held shares hold by insiders in % CASHFLOW Datastream Free cash flow per share
lnSIZE Zephyr logarithmic function of the deal value RELATIVESIZE Zephyr Relative deal value to market capitalization CROSSBORDER Zephyr Dummy: 1 if target is cross border, 0 if otherwise CRISIS - Dummy: 1 if completed during crisis, 0 if otherwise *Source: the source is the data source from where all the data is retrieved. For the variable CRISIS the period is retrieved from Duchin, Ozbas and Sensoy (2010).
3.1.4 Dependent variable
The dependent variable, METHOD OF PAYMENT, is a binary dependent variable. The value it can obtain is either 0 or 1. In this model cash financed M&A are given the value of 1, the equity financed M&A are given a value of 0. The dataset also contains mixed method of payments, which means that different methods are used in a single takeover. For the sake of
simplicity and the application of the probit model these observations are divided with a border of 50 percent. When a takeover is at least financed with 50 percent with either cash or shares, it is used as an observation. From the final dataset 4,831 cash financed mergers and acquisitions are retrieved. 1,462 takeovers were financed with equity for at least 50 percent. This makes a full dataset of 6,293 observations.
3.1.5 Explanatory and control variables
The main explanatory variables are the effective Federal Funds interest rate (INTERESTRATE) and the Wu-Xia Shadow rate (WUXIARATE). The INTERESTRATE is the effective monthly Federal Funds rate of the Federal Reserve Bank. It is announced at the last business day of the month. As stated in the introduction, previous results differ from each other about the significance of the interest rate. The expectation is that a decrease of the interest rate results in more cash finance M&A because of the cheaper and easier access to external funds. In Figure I (Section 2.2) the effective Federal Funds interest rate and Wu-Xia Shadow rate are presented. Since the INTERESTRATE contains only positive values, an interest rate squared (INTEREST2) is added to the regression model to capture a non-linear effect of the effective Federal Funds interest rate.
The Wu-Xia Shadow rate is introduced into the regression model (Equation 3.3) as WUXIARATE. The Wu-Xia Shadow rate is a rate which resembles the FED rate in an environment where the interest rate can go below the lower zero bound. The federal funds rate is the primary instrument for the Federal Reserve to have influence on the monetary policy. From December 2008 on the Federal Reserve has decreased the federal funds rate to a level near zero, as it does not go further below. The Wu-Xia Federal Funds rate uses forward rates to estimate the term structure of the federal funds rate. With these estimations, shadow interest rates are produced which go below the lower zero bound (Wu and Xia, 2015). The expected relationship with the method of payment is similar to the Federal fund rate: an increase in the interest rate leads firms to financing M&A more frequently with equity. The Wu-Xia Shadow rate is only predicted from January 2009 – November 2015. Because the values before this period are equal to the effective Federal Funds rate, the values from January 1997 – December 2008 are equal to the values of the effective Federal Funds rate.
Choe, Masulis and Nanda (1993) found a relation between the amount of equity offers and the economic growth. The expectation of the relation with method of payment is in line with the results of Choe, Masulis and Nanda (1993): an expansion of the economic environment results in a higher chance of firms financing with equity. In this research the controlling variable for economic growth is found in the variable ECONOMICGROWTH. The data used are the daily returns of the S&P500 index.
Besides business cycle factors, firm specific factors are also influencing the ways firms finance their M&A. Martin (1996) found a relation between Tobin’s Q and the method of payment. Tobin’s Q is formulated as a ratio between the market value and the replacement costs of the assets (Brainard and Tobin, 1968). In practice Tobin’s Q is a ratio which determines when a firm has a better opportunity to grow. To calculate this ratio the summation of the market value of equity and market value of liabilities are divided by a summation of the book value of equity and the book value of liabilities. If the value of the ratio exceeds 1, this means firms have a better growth opportunity. The control variable for the Tobin’s Q is formulated in the regression as TOBINSQ.
In addition, Martin (1996) used a variable to explain the explanatory power of the amount of financial leverage a firm acquired. If the leverage ratio is higher, firms are less likely to obtain external funds because of the increasing risk. To control for this risk, the variable LEVERAGE is used and is defined as the leverage ratio in percentages. The computation is a summation of the short term debt and long term debt divided by the total assets of a company (Datastream).
Amihud, Lev and Travlos (1990) found a relation between corporate insiders and the method of payment. The empirical results conclude that firms with managerial ownership in the acquiring firm are more likely to finance a takeover with cash. The main theory behind this is that if firms finance with equity, this will dilute their ownership and has an increasing effect on the risk of losing control. Corporate control will be controlled via closely held shares. Closely held shares are shares which are in held by insiders. Shares held by insiders include shares held in trust, by officers, immediate families, other companies held by the acquirer, pension/benefit plans and individuals who hold at least 5 percent of the total outstanding shares. The closely held shares ratio (variable CONTROL) is calculated by dividing the number of closely held shares by the common shares outstanding (Datastream).
Jensen (1986) argues that firms with large amounts of free cash flow are more likely to finance a takeover with cash. Free cash flows are excess cash flows which are required to finance new projects when discounted with the cost of capital. This consists of high cash flows, but also as a sufficient debt capacity. The control variable CASHFLOW is calculated by the free cash flow per share. The free cash flow is calculated by the Funds from Operations minus the summation of the Capital Expenditures and Cash Dividends Paid (Datastream).
The size of the deal also has an impact on the method of payment. Faccio and Masulis (2005) used two different variables to analyze the relation. The first one, SIZE, is the value of the deal. Because the difference between the deal values differs a lot, the logarithmic function of the deal value is used. In addition to SIZE, RELATIVESIZE is added. RELATIVESIZE is the relative size of the deal value. The computation of the relative deal size is the same as Faccio and Masulis (2005): the deal value is divided by the sum of the market capitalization and the deal value. It is found that a higher deal value results in more equity financed M&A.
From Faccio and Masulis (2005) a third control variable is found necessary to include: CROSSBORDER. It is argued that investors could have a home country bias reflected in greater transaction costs for foreigners, exchange rate risk, lower liquidity and a higher chance of less access to firm specific information. This variable is defined as a dummy variable and has been given the value of 1 if the target is outside the U.S., and 0 if otherwise.
As the dataset has observations from 1997- 2015 some takeover announcements dates found place in the financial crisis. The financial crisis has a negative effect on the supply of external finance (Duchin, Ozbas and Sensoy, 2010). Investments were declining significantly. Firms with low cash reserves or high debt levels had the highest decline. The variable CRISIS is introduced to the regression to measure the effect of these effects. The value of 1 is given for announcements dates which found place from August 2007 – June 2009, and the value of 0 if otherwise.
3.2 Data description
The data for the method of payment variable is obtained from Zephyr database. The first limitation to the observations is that the deal type is either a merger or acquisition.
Secondly, the deal status must be completed or assumed to be completed. Thirdly, acquirers are addressed in the United States. At last, acquiring companies are listed at any stock exchange. The method of payment, value of the method of payment, the deal value, announcement- and completion date and the primary addresses of the acquirers (US) are obtained from Zephyr. In addition, ISIN- and SIC codes are retrieved. ISIN (International Securities Identification Number) codes are identifying codes for every listed company. SIC (Standard Industrial Classification) codes are codes which classify companies in different industries, which are used for the robustness check. There are no additional limitations to the target company. The time period when the merger or acquisition is completed is from January 1997 until December 2015. From Zephyr 12,058 observations were retrieved. Due to lack of information of the deal value 3,786 observations are left out. In addition, 20 deals were exactly financed for 50 percent with cash, and 50 percent with shares. Due to biased results, these deals are also excluded from the dataset.
Data for the main explanatory variables (effective Federal Funds interest rate and Wu-Xia Shadow Rate) are obtained from the Federal Reserve Bank of Atlanta. To control for the economic growth, historical returns of the S&P500 are obtained from Yahoo Finance. In addition, Datastream database is used to obtain firm specific information: Tobin’s Q, leverage ratio, management control and the free cash flow. All the data is taken from the database on the announcement date. The reason why the announcement date is taken is because it represents the date when the final decision is made. It is important that the explanatory variables are from the same date as that decision date. From the 8,252 observations the outliers, errors and missing values from Datastream are removed. This makes the total observations 6,293.
4. Results and analysis
In this section the results and data analysis are stated. First an overview of descriptive statistics from the sample will be presented in Table II. In addition, correlation matrices will be indicated for the two different regression models. In section 4.3 the results will be stated, and section 4.4 provides a robustness check.
4.1 Descriptive statistics
In Table II the descriptive statistics are given for each variable used in one of the two empirical regression models. The descriptive statistics consist of all the 6,293 observations. The outliers and errors from the dataset are already removed as stated in section 3.2.
Table II Quantiles
Variable Observations Mean Std.Dev. Min 0.25 Median 0.75 Max
METHOD OF PAYMENT 6293 0.77 0.42 0 1 1 1 1 INTEREST 6293 2.12 2.15 0.04 0.10 1.38 4.09 7.07 INTEREST2 6293 9.15 12.61 0 0.01 1.90 16.73 49.98 WUXIA 6293 1.54 2.80 −2.99 −1.11 1.38 4.09 7.07 ECONOMICGROWTH 6293 1347.68 325.44 666.79 1129.00 1273.34 1470.08 2126.06 TOBINSQ 6293 2.18 1.86 0.05 1.18 1.62 2.40 20.98 LEVERAGE 6293 22.31 20.66 0 5.40 18.75 33.03 200.74 CONTROL 6293 18,48 20.65 0 2.05 12.00 27.25 99.32 CASHFLOW 6293 0.87 2.92 −21.85 0.13 0.84 1.75 18.91 lnDEALSIZE 6293 18.00 1.99 8.29 16.71 17.93 19.26 25.22 RELATIVEDEALSIZE 6293 0.36 1.77 0 0.01 0.04 0.13 26.85 CRISIS 6293 0.14 0.34 0 0 0 0 1 CROSSBORDER 6293 0.20 0.40 0 0 0 0 1
The variables INTEREST, INTEREST2, WUXIARATE, LEVERAGE and CONTROL are stated in percentages. The variables METHOD OF PAYMENT, CRISIS and CROSSBORDER are dummy variables. From Table II it can be seen that the lowest effective interest rate reached 0.04 percent, which means it stays just above the zero percent border. The Wu-Xia Shadow rate has a value of -2.99 percent as the lowest rate. The effective Federal Funds rate and Wu-Xia rate do have the same median, three-quarter quantile and maximum value. This is due to the fact that the predictions of the shadow rate from January 1997 – December 2008 are the same as the effective Federal Funds rate. Because the values are rounded at two decimals, certain values can be recalculated with little deviation.
4.2 Correlation matrices
Two different regression models are used to analyze the effects of the effective Federal Funds rate and Wu-Xia Shadow rate. As the correlations of the regression models differ, two correlation matrices are stated on the next page, Table III and Table IV. Table III includes the explanatory variables INTEREST and INTEREST2, while Table IV includes the explanatory variable WUXIARATE. The correlation coefficients are providing information about how two variables are linearly related to each other.
From Table III it is showed that the variables INTEREST and INTEREST2 are highly correlated. This makes sense because the INTEREST2 is the squared value of INTEREST. In addition, the variables CRISIS and ECONOMICGROWTH are correlated with the variable INTEREST. This is as expected because the Federal Reserve Bank will lower the interest rate in an event of crisis or low economic growth. The results presented in Table IV show somewhat the same results as Table III. The Wu-Xia Shadow rate has a high correlation with CRISIS and ECONOMICGROWTH. The remainders of the correlation coefficients are not surprisingly large which do not result in a problem.
Regarding the main explanatory variables of both tables, INTEREST, INTEREST2 and the WUXIARATE, all have a negative correlation with the method of payment of a merger or acquisition. This is corresponding with the theory and literature: when the Effective Federal Funds interest rate (or Wu-Xia Shadow rate) increases, the amount of cash-financed M&A is presumed to be lower. One limitation to this conclusion is that no causality is measured in the correlation coefficients, which makes this effect not strong in explanatory power.
Table IV: Correlation matrix
Explanatory variable: Wu-Xia Shadow Rate
Variable METHOD OF PAYMENT WUXIARATE ECOGROWTH TOBINSQ LEVERAGE CONTROL CASHFLOW lnDEALSIZE RELDEALSIZE CRISIS CROSSBORDER
METHOD OF PAYMENT 1 WUXIARATE −0.0573 1 ECONOMICGROWTH −0.0126 −0.3627 1 TOBINSQ −0.1209 0.0801 −0.0086 1 LEVERAGE 0.0952 −0.0543 0.0736 −0.0652 1 CONTROL −0.0890 0.1130 −0.1072 0.0809 0.0695 1 CASHFLOW 0.1577 −0.0337 −0.0045 −0.0737 −0.0633 −0.1549 1 lnDEALSIZE 0.0136 −0.0597 0.1223 −0.1092 0.1303 −0.2886 0.1717 1 RELATIVEDEALSIZE −0.1205 −0.0165 0.0089 0.0713 0.0329 0.1382 −0.1355 −0.0470 1 CRISIS 0.0210 0.2471 −0.0483 −0.0508 −0.0118 0.0587 0.0415 −0.0507 −0.0139 1 CROSSBORDER 0.1104 −0.0063 −0.0257 0.0624 −0.0214 −0.0501 0.0469 −0.0649 −0.0126 −0.0143 1 Number of observations: 6,293.
Table III: Correlation matrix
Explanatory variable: Effective Federal Funds Rate
Variable METHOD OF PAYMENT INTEREST INTEREST2 ECOGROWTH TOBINSQ LEVERAGE CONTROL CASHFLOW lnDEALSIZE RELDEALSIZE CRISIS CROSSBORDER
METHOD OF PAYMENT 1 INTEREST −0.0697 1 INTEREST2 −0.0901 0.9627 1 ECONOMICGROWTH −0.0126 −0.1830 −0.0548 1 TOBINSQ −0.1209 0.0882 0.0972 −0.0086 1 LEVERAGE 0.0952 −0.0420 −0.0282 0.0736 −0.0652 1 CONTROL −0.0890 0.0847 0.0551 −0.1072 0.0809 0.0695 1 CASHFLOW 0.1577 −0.0402 −0.0356 −0.0045 −0.0737 −0.0633 −0.1549 1 lnDEALSIZE 0.0136 −0.0323 0.0033 0.1223 −0.1092 0.1303 −0.2886 0.1717 1 RELATIVEDEALSIZE −0.1205 −0.0150 −0.0100 0.0089 0.0713 0.0329 0.1382 −0.1355 −0.0470 1 CRISIS 0.0210 0.2079 0.1579 −0.0483 −0.0508 −0.0118 0.0587 0.0415 −0.0507 −0.0139 1 CROSSBORDER 0.1104 −0.0115 −0.0087 −0.0257 0.0624 −0.0214 −0.0501 0.0469 −0.0649 −0.0126 −0.0143 1 Number of observations: 6,293.
4.3 Results and analysis
In Table V the results of the regressions are presented. In total four different regressions are predicted. The main explanatory variables of regression models I and II are INTEREST and INTEREST2. The main explanatory variable of regression models III and IV is the WUXIARATE. The difference between model I and II and between III and IV is the addition of variables for robustness checks. This will be explained in the next section. Most of the control variables are significant at a level of 1 percent. Two control variables are not significant in regression model I and II: ECONOMICGROWTH and CRISIS.
When focusing on regression models I and III, it is showed that INTEREST, INTEREST2 and WUXIARATE are significant at a level of 1 percent. The INTEREST has a positive coefficient of 0.1669, while the INTEREST2 and WUXIARATE both have negative coefficients: respectively -0.0355 and -0.0292. As all the three variables are strongly significant, this means that both H0 hypotheses are rejected. As stated in the methodology, values predicted with the probit regression model cannot be interpreted as straightforward as interpretations of linear regression models. The marginal effects are computed and stated below.
The positive value of INTEREST means that the chance of a cash-financed M&A is higher with an increase in the effective Federal Funds interest rate. This is in contrast with what previous studies showed and the hypothesis stated above. Marsh (1982) and Taggart (1977) found a negative value, which means that the interest rate has a negative impact on the probability of the cash-financing method of payment. The results show a positive value of the coefficient of INTEREST: 0.1669. The marginal effect of INTEREST is computed at 0.0498. This implies that an increase of 1 percent in the effective Federal Funds rate, increases the probability of a M&A being cash-financed with 4.98 percent.
INTEREST2 has a negative coefficient value of -0.0355. The marginal effect calculated with this value is: -0.0104. This implies that an increase of 1 percent in the effective Federal Funds rate decreases the probability of a M&A being cash-financed with 1.04 percent. The variable INTEREST2 is created to capture the nonlinear effect of the interest rate. The negative value implies a decrease of cash-financed M&A when the interest rate increases. This is in line with the expectations, which stated that an increase of the interest rate results in less cash-financed M&A.
From regression model III the coefficient of the WUXIARATE is predicted. The WUXIARATE has a negative coefficient value of -0.0292. As stated above, the Wu-Xia Shadow rate is constructed from the effective Federal Funds rate, with an adjustment in the years 2009 until 2015. The negative value of the coefficient implies that an increase of the interest rate results in less cash-financed mergers and acquisitions. The marginal effect computed is -0.009, which means that an increase of 1 percent in the Wu-Xia Shadow rate will result in an increase of 0.9 percent in the probability that a M&A is financed with equity. This is in line with the expectations.
4.4 Robustness check
A robustness check is included to test the robustness of the two regression models. The robustness check consists of an industry specific fixed effects variable to control for industry specific characteristics. García-Feijóo, Madura and Ngo (2012) researched the effect of these characteristics on the method of payment in takeovers. They find a considerable effect across different industries, with regards to levels of free cash flow, financial leverage, relative size of the acquirer and equity overvaluation. To test the robustness of the empirical models, the SIC-codes for each acquiring firm are determined. As stated in the section 3.3, SIC-codes are industry specific codes which identify the primary business executed by the acquiring firm. The codes range from 0100 – 9999. Codes for Agriculture, Forestry and Fishing are for instance the codes 0100 – 0900, and for the Mining industry 1000 – 1400. In total ten different industries are distinguished and added to the regression equations. The results of this robustness check are stated in Table V: regression model II and IV. From these regressions is it shown that the Pseudo-R2 is increasing with almost 100 percent while the explanatory variables do not differ much. The explanatory power of the regression model including robustness check is thus significantly higher.
*** denotes significance at 1 percent, ** denotes significance at 5 percent, * denotes significance at 10 percent. Sample restrictions and construction of variables are elaborated in section 3. Regressions (1) and (3) are regressions with the empirical models 3.2 and 3.3. Regressions (2) and (4) are regressions including robustness checks.
Table V: Results empirical probit regressions
Binary dependent variable: Method of Payment (1 if cash-financed, 0 if equity-financed) Robust standard errors are reported in parentheses
Variables (I) (II) (III) (IV)
INTEREST 0.1699 0.1690 − − (0.0365)*** (0.0379)*** − − INTEREST2 −0.0355 −0.0360 − − (0.0060)*** (0.0062)*** − − WUXIARATE − − −0,0292 −0.0331 − − (0,0073)*** (0.0075)*** ECONOMICGROWTH 0.0001 0.0001 −0,0002 0.0002 (0.0001) (0.0001) (0,0001)*** (0.0001)** TOBINSQ −0.0641 −0.1133 −0.0670 −0.1162 (0.0098)*** (0.0111)*** (0.0010)*** (0.0113)*** LEVERAGE 0.0083 0.0082 0.0082 0.0082 (0.0011)*** (0.0011)*** (0.0011)*** (0.0011)*** CONTROL −0.0046 −0.0053 −0.0044 −0.0051 (0.0009)*** (0.0010)*** (0.0001)*** (0.0010)*** CASHFLOW 0.0708 0.0747 0.0704 0.0741 (0.0075)*** (0.0079)*** (0.0074)*** (0.0078)*** lnDEALSIZE −0.0300 −0.0284 −0.0331 −0.0316 (0.0107)*** (0.0113)** (0.0106)*** (0.0112)*** RELATIVEDEALSIZE −0.0600 −0.0614 −0.0607 −0.0622 (0.0118)*** (0.0123)*** (0.0118)*** (0.0122)*** CRISIS 0.0423 0.0257 0.1029 0.0879 (0.0566) (0.0582) (0.0565)* (0.05820) CROSSBORDER 0.4675 0.3035 0.4599 0.2959 (0.0500)*** (0.0534)*** (0.0496)*** (0.0530)*** _cons 1.1109 1.0035 1.5419 1.4026 (0.2171)*** (0.4376)** (0.2055)*** (0.4397)***
INDUSTRY FIXED EFFECTS NO YES NO YES
Pseudo-R² 0.0759 0.1448 0.0702 0.1392
5. Conclusion
One of the most important decisions of managers is the decision to take over another company. The method of payment of such a merger or acquisition has a great influence on the way the acquiring firm is performing in the future. Interest rates do affect managers these financing decisions. From the underlying literature it is showed that the interest rate influences the capital structure of firms. This paper has researched the effect of historically low interest rates on the financing decisions firms take in merger and acquisitions.
In total four different probit models have been estimated. The binary dependent variable of all models is the method of payment, constructed by given cash-financed (if the total deal value is at least financed for 50 percent with cash) takeovers a value of 1, and equity-financed takeovers a value of 0. The sample consists of 6,293 mergers and
acquisitions. 1,462 observations are equity-financed takeovers, 4,831 observations are cash-financed. All announcement and completions dates are in the period of 1997 – 2015, and addressed in the United States.
Model I predicts, with 1 percent significance, that the effective Federal Funds rate has a positive effect on the method of payment. This means that an increase of the interest rate results in a higher probability of cash-financed takeovers. This is in contrast with the hypothesis. To control for a nonlinear effect of the effective Federal Funds rate, an interest rate squared is added to the model. This value has a negative value, which is also significant at 1 percent.
In model III the main explanatory variable is the Wu-Xia Shadow rate. The
coefficient predicted is significant at 1 percent, and has a negative value. This is in line with the expected hypotheses that a decrease in the interest rate increases the probability of cash-financed investments. Model II and IV replicates the models announced above, but with an added variable for robustness checks. This robustness check consists of industry specific fixed effects variable. The main outcome of regressions II and IV is that the Pseudo-R2 increases by almost 100 percent, while the coefficients do not differ much.
The overall results from this research are somewhat peculiar. The Wu-Xia Shadow rate and the interest rate squared are in line with the hypothesis. This means that the Wu-Xia Shadow rate and Federal Funds Rate Squared predict an increase of cash-financed mergers when the rates are lower. However, the effective Federal Funds rate is
predicted in contrast. This might be due to a non-linear effect of the interest rate, or due to theory not being accurate in times of a (post) financial crisis. Another explanation could be the fact that in times of interest rate levels which are near zero the effect to the method of payment is different.
One limitation is encountered during this research. This regards the lack of data about the credit rating of firms. Adding the credit ratings as a control variable could control for the easiness of firms issuing leverage or equity.
References:
Amihud, Y., Lev, B., & Travlos, N. G. (1990). Corporate control and the choice of investment financing: The case of corporate acquisitions. The Journal of Finance, 45(2), 603-616.
Baigorri, M. (2012, January 5) 2015 Was Best-Ever Year for M&A; This Year Looks Good Too. Retrieved from http://www.bloomberg.com/news/articles/2016-01-05/2015-was-best-ever-year-for-m-a-this-year-looks-pretty-good-too
Berk, J. B., DeMarzo, P. M., & dawsonera. (2014). Corporate finance (Third edition, Global ed.). Boston, Mass. ; London: Pearson.
Black, F. (1995). Interest rates as options. The Journal of Finance, 50(5), 1371-1376. Brainard, W. C., & Tobin, J. (1968). Pitfalls in financial model building. The American Economic Review, 58(2), 99-122.
Choe, H., Masulis, R. W., & Nanda, V. (1993). Common stock offerings across the business cycle: Theory and evidence. Journal of Empirical finance, 1(1), 3-31.
Duchin, R., Ozbas, O., & Sensoy, B. A. (2010). Costly external finance, corporate investment, and the subprime mortgage credit crisis. Journal of Financial Economics, 97(3), 418-435. Faccio, M., & Masulis, R. W. (2005). The choice of payment method in European mergers and acquisitions. The Journal of Finance, 60(3), 1345-1388.
García-Feijóo, L., Madura, J., & Ngo, T. (2012). Impact of industry characteristics on the method of payment in mergers. Journal of Economics and Business, 64(4), 261-274.
Hoetker, G. (2007). The use of logit and probit models in strategic management research: Critical issues. Strategic Management Journal, 28(4), 331-343.
Jensen, M. C. (1986). Agency cost of free cash flow, corporate finance, and takeovers. Corporate Finance, and Takeovers. American Economic Review,76(2). Jensen, M. C., & Ruback, R. S. (1983). The market for corporate control: The scientific evidence. Journal of Financial economics, 11(1), 5-50.
Marsh, P. (1982). The choice between equity and debt: An empirical study. The Journal of finance, 37(1), 121-144.
Martin, K. J. (1996). The method of payment in corporate acquisitions, investment opportunities, and management ownership. The Journal of finance, 51(4), 1227-1246. Merton, R. C. (1974). On the pricing of corporate debt: The risk structure of interest rates. The Journal of finance, 29(2), 449-470.
Miles, J. A., & Ezzell, J. R. (1980). The weighted average cost of capital, perfect capital
markets, and project life: a clarification. Journal of Financial and Quantitative Analysis, 15(3), 719-730.
Modigliani, F., & Miller, M. H. (1958). The cost of capital, corporation finance and the theory of investment. The American economic review, 48(3), 261-297.
Myers, S. C. (1977). Determinants of corporate borrowing. Journal of financial economics, 5(2), 147-175.
Myers, S. C. (1984). The capital structure puzzle. The journal of finance, 39(3), 574-592. Myers, S. C., & Majluf, N. S. (1984). Corporate financing and investment decisions when firms have information that investors do not have. Journal of financial economics, 13(2), 187-221.
Sirower, M., & Rappaport, A. (1999). Stock or cash. Harvard Business Review, 77(6), 147-158.
Stock, J. H., & Watson, M. W. (2007). Introduction to econometrics. Boston: Pearson/Addison Wesley.
Taggart, R. A. (1977). A model of corporate financing decisions. The Journal of Finance, 32(5), 1467-1484.
Travlos, N. G., & Papaioannou, G. J. (1991). Corporate acquisitions: Method of payment effects, capital structure effects, and bidding firms stock returns. Quarterly Journal of Business and economics, 3-22.
Wansley, J. W., Lane, W. R., & Yang, H. C. (1983). Abnormal returns to acquired firms by type of acquisition and method of payment. Financial management, 16-22.
Wu, J. C., & Xia, F. D. (2014). Measuring the macroeconomic impact of monetary policy at the zero lower bound (No. w20117). National Bureau of Economic Research.
