The post-operating performance of acquirers who take over financially distressed firms, measured by synergies and the return on sales

(1)

The post-operating performance of acquirers who

take over financially distressed firms, measured

by synergies and the return on sales

Abstract

This thesis looks at the effect of financially distressed targets on the post-operating performance of the acquirer, using acquirers from the Unites States and from Europe. The time period that has been observed is 1978 until 2015. The sample contains 6459 deals, of which 336 targets are in financial distress. Financial distress is measured by three bankruptcy measures, namely Altman’s Z-score, Ohlson’s O-score and the Zmijewski score. The model that has been used to test the hypotheses is the random effects model and the variables that have been used to test the

performance is sales revenue divided by total assets (revenue synergies), SGA expenses divided by sales revenue (cost synergies) and EBIT divided by sales revenue (operating performance). The performance will be compared with acquirers that took over non-distressed targets. This research found that the revenue and cost synergies increased when acquirers took over distressed firms, while is decreased for acquirers that took over non-distressed firms. However, the operating performance has increased for acquirers of non-distressed targets, while it decreased for acquirers of distressed targets.

Keywords: M&A, Financial distress, synergies, sales revenue, SGA expenses, return on sales.

Name: Anne-Lotte Meier

Student number: 10190546

Supervised by: Mr. dr. Vladimir Vladimirov Amsterdam Business School

MSc Business Economics, Finance track Course: Master’s Thesis Finance

Number of ECTS: 15 points Amsterdam, August 2016

(2)

Statement of originality

This document is written by Student Anne-Lotte Meier, 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)

1. Introduction

Mergers and acquisitions (M&A) have increased significantly in importance in the corporate finance world since the 1990s (Gerchak and Gupta, 2002, p. 517; Fuller et al, 2002). Every day, in the economic newspapers, you will find articles about M&A. It will therefore come to no surprise that there has already been done a lot of research regarding this subject.

The existing literature did already research to the performance of M&A as homogeneous group (Cao et al., 2016; Healy et al, 1992) and the performance in abnormal returns when looking at different motives or different targets (Amit et al., 1989; Fuller et al, 2002). Further, the timing of mergers has already been investigated multiple times (Thijssen; 2008; Harford, 2005; Lambrecht, 2004). In addition, synergies are recently also a popular way to assess the performance of merges and acquisitions (Lambkin and Rahman, 2015; Devos et al., 2009; Thijssen, 2008; Gerchak and Gupta, 2002). Another interesting topic is mergers as way to restructure financially distressed and bankrupt targets. Does the performance of these distressed targets improve? Mostly it does (Balcaen et al, 2012; Jostarndt and Sautner, 2008; Hotchkiss, 1998). However, still very little is unambiguously known about taking over financially distressed targets, especially when looking at the improvements for the acquirer after the deal.

In the past, a firm in financial distress is mainly seen as dangerous or risky, which can only do harm. In particular debt holders are affected when the firm is unable to meet their debt obligations (Balcaen et al., 2012, p. 952; Berk and DeMarzo, 2011, p. 509). But also customers, employees and suppliers are affected (Damodaran, 2006, p. 612; Opler and Titman, 1994, p. 1015). Taking over such a firm, could be improving for these target stakeholders, but what about the stakeholders of the acquirer? Clark and Ofek (1994) found negative performance results for the bidders, using 5 measures. Also Bugeja (2015) found that financial distress has a negative affect on the takeover success. The results of Amit et al., with respect to the abnormal returns of the bidding firm, were negative as well. However, some researchers found positive results when buying a firm in financial distress (Khatami et al., 2015; Jory and Madura, 2009; Hotchkiss and Mooradian, 1998).

A recent example in The Netherlands is the big department store, Vroom & Dreesman. Cool Investments, the owner of multiple fashion shops, has signaled that it would be interested in taking over the financially distressed firm, but eventually, they decided not to do it and V&D got bankrupt. This example raises some further questions. Why was the acquirer interested in the first place and why did the acquirer eventually

(5)

decide to quit? These questions are worthwhile to be examined. Understanding the motives of these acquirers is especially important for managers, employees, investors, shareholders and bondholders (Kahya et al, 1996, p. 699). Investors and shareholders can, for example, profit from the increase in the acquirer’s share price. Managers and employees can become part of a bigger and growing company, which can increase their responsibilities and salaries. And bondholders have more chance getting back their lent money. But it could also be interesting for society. Why was it still possible to rescue that bankrupt shoes shop around the corner, but is your favorite sports shop going down the drain.

Until now, previous researchers had little success in explaining the relation between distressed targets and the improvements for the acquirer and its stakeholders. Therefore, this thesis will take another look at this puzzle, with the aim to figure out if taking over a financially distressed target will have a positive effect on the post- performance of the acquirer. The research question is: Does a firm in financial distress have a positive effect on the post-operating performance of the acquirer? And if yes, will this improvement be higher than when it had taken over a non-distressed firm?

The post-performance will be measured by revenue synergies and cost synergies, which will arise, for example, due to selling new products, engaging in new geographic areas, spreading costs as results of being large and the elimination of overlapping expenses. This could eventually lead to a higher return on sales, as measured by EBIT divided by sales revenue. Lambkin and Rahman (2015) used these measures as well, but they didn’t look at targets in financial distress. Further, the exact measures are also a bit different as can be read in the methodology section. Financial distress will be measured using three bankruptcy measures, namely, Altman’s Z-score, Ohlson’s O-score and the Zmijewski score. These measures have also successfully been used by previous

researches (Bugeja, 2015; Amit et al., 1989). A dummy variable will be included into the regression model to indicate that the target firm is in financial distress or not.

Additionally, an entity specific time dummy will be included to indicate the time period before and after the deal.

The data will be retrieved from the Thomson One database and Datastream. In order to answer the research questions, this thesis will look at panel data regressions, specifically, the random effects model, which means that multiple entities can be studied over time, in this thesis multiple M&A deals. Eventually, 6459 deals are obtained, of which 336 target firms are measured as financially distressed. This thesis will focus on acquirers from the United States and from Europe. Targets can be established anywhere

(6)

else in the world. In addition, the time period that will be observed is from 1978 until 2015.

This thesis found that the revenue and cost synergies increased for the acquirer of a financially distressed firm, while it decreased for acquirers of non-distressed targets. This could be explained due to the unexploited efficiency opportunities in distressed targets. However, eventually, the return on sales decreased for distressed targets, while it increased for acquirers of “healthy” firms. This could be explained by higher costs during the production process when taking over a distressed firm or more experienced

managers and employees in the “healthy” target. It must be kept in mind that the results are not statistically significant.

This thesis has the following structure. The first chapter will cover the literature review, which is followed by the methodology that will be used. Thereafter, the data will be discussed and some descriptive statistics will be shown. At last, the results are presented and a conclusion will be formed, together with some limitations and suggestions for further research.

(7)

2. Literature review

As mentioned above, there has already been done a lot of research, especially with respect to the post-performance of mergers and acquisitions in general (Cao et al., 2016; Lambkin and Rahman, 2015; Fuller et al., 2002; Healy et al., 1992). However, there exist some papers about financially distressed targets as well. A lot of these articles focus on finding the best and most successful way to restructure the target (Erel et al., 2015; Hotchkiss and Mooradian, 1998; Clark and Ofek, 1994). This section will first give an introduction about M&A, which is followed by an explanation about synergies. And at last, financial distress will be discussed, how can it be measured and why could a financially distressed target be attractive?

2.1 Mergers and Acquisitions

According to Berk and DeMarzo, mergers and acquisitions (M&A) can be classified into “the market for corporate control”. They explain that mergers and acquisitions are actually two different methods that can be used to change the ownership and control of a company. A firm, or group of investors, buys another firm, or two existing firms merge and become one new firm (2011, p. 891). This thesis will use both mechanisms

interchangeably and refer to it as takeover or acquisition, just like Berk and DeMarzo did (2011, p. 891). However, it will namely focus on acquisitions.

In this market there are periods in which M&A transactions occur quite often, during economic expansions, but also periods in which they are less popular, during economic recessions. This phenomenon is called merger waves (Berk and DeMarzo, 2011, p. 891; Thijssen, 2008, p. 1703; Lambrecht, 2004, pp. 41-42). The peak periods, with heavy activity, took place in the 1960s, 1980s, 1990s and 2000s (Berk and DeMarzo, 2011, p. 892; Harford, 2005, p. 530).

Acquirers often pay a premium to the shareholders of the target firm, which is an additional amount to the premerger value of the target shares. This additional amount depends mostly on the extra expected value that can be created as result of the takeover, whereby synergies are usually mentioned as explanation (Berk and DeMarzo, 2011, p. 894; Clark and Ofek, 1994, p. 547; Healy et al., 1992, p. 136).

However, the synergies are not the only motive for a takeover. Other reasons are market power, managerial reasons, diversification, restructuring a financially distressed target, and tax savings (Cao et al., 2016, p. 374; Bugeja, 2015, p. 362, Berk and DeMarzo, 2011, p. 894-898; Lambrecht, 2004, p. 43; Gerchak and Gupta, 2002, p. 517).

(8)

As already explained in the introduction, this thesis investigates the post-operating performance of an acquirer, which is measured by the synergies that may be created.

2.2 Synergies

Synergy is the extra economic value that is created as result of a takeover, which could not be achieved as firms independently (Berk and DeMarzo, 2011, Damodaran, 2008, p. 541; Lambrecht, 2004, p. 43; Gerchak and Gupta, 2002, p. 517; Bradley et al., 1988, p. 4). Devos et al. and Damodaran point out that synergies can be divided into two groups: operational synergies and financial synergies. Operational synergies, as the name already implies, are related to changes in the operation of the firm. They include for example, economies of scope and scale and increasing market power. On the other hand, financial synergies are more related to tax reductions, diversification and increased debt capacity (2009, p. 1180; 2006, p. 542). This thesis will mainly look at operational synergies, which can be subdivided into two groups: revenue synergies and cost synergies (Lambkin and Rahman, 2015, p. 25; Berk and DeMarzo, 2011, p. 894; Devos et al., 2009, p. 1181). If eventually both values were successfully achieved, a logical result would be higher returns on sale and increased profitability (Lambkin and Rahman, 2015, p. 28; Hoberg and Phillips, 2010, pp. 3774-3775; Devos et al., 2009, p. 1181; Bradley et al., 1988, p. 4) 2.2.1 Revenue Synergies

Lambkin and Rahman state that revenue synergies, also called product market synergies by Hoberg and Phillips (2010, p. 3773), can be generated by entering new product markets and new geographic areas, whereby they refer to economies of scope (2015, p. 26; Berk and DeMarzo, 2011, p. 895; Damodaran, 2006, p. 542). Hoberg and Phillips claim that this is possible when both the target and acquirer have complementary assets. They also mention that as acquirer you need to differentiate yourself from your rivals in order to gain customers and revenues, which can be accomplished by taking over a target with other experiences, skills and technologies (2010, pp. 3774-3775; Devos et al., 2009, p. 1183; Bradley et al., 1988, p. 4). Berk and DeMarzo also point out that expertise is needed to distinguish yourself from your competitors. Finding people with this talent can be hard, therefore, hiring them by buying another firm is a good solution (2011, p. 895). Further, being able to take over your competitors can result in monopoly gains. However, this is limited due to regulations of the government (Berk and DeMarzo, 2011, p. 896; Devos et al., 2009, p. 1184). If these revenue synergies are achieved, existing products can be sold to more customers, or more diverse products can be sold to the existing

(9)

customers, which increases sales (Lambkin and Rahman, 2015, p. 28; Hoberg and Phillips, 2010, p. 3775; Lambrecht, 2004, p. 43).

2.2.2 Cost Synergies

According to Berk and DeMarzo, cost synergies can be created by taking over another firm and thereby reducing the operating costs, as result of economies of scale (Devos et al., 2009, p. 1181; Damodaran, 2006, p. 542). They state that due to being large, costs can be spread over a higher production volume (2011, p. 895). In addition, Lambkin and Rahman claim that due the efficiency gains, a firm needs less assets to achieve their preferred goals (2015, p. 28). These cost savings can arise due to the elimination of duplication, for example overlapping employees, sale of surplus assets, and other redundant resources (Lambkin and Rahman, 2015, p. 28; Berk and DeMarzo, 2011, p. 894).

2.3 Financial Distress

Financial distress is a concept without an unambiguous or clear definition and is therefore hard to measure (Bugeja, 2015, p. 363; Balcaen et al., 2012, p. 958; Oplet and Titman, 1994, p. 1016). Balcaen et al. (2012) state for example that a firm in financial distress is still able to survive, but experience difficulties in repaying debt, due to high leverage (p. 952). They lack revenues in order to cover operating expenses, cost of debt and taxes (p. 958). This definition has also been expressed by Damodaran (2006, p. 612). Berk and DeMarzo simply claim that these firms are unable to meet their financial obligations (2011, p. 509). If one of the following signs can be found within a firm, it is, according to Kayha et al., classified as financially distressed: “1. Debt default, 2. Debt renegotiation attempts with creditors and financial institutions, and 3. Inability to meet fixed payment obligations on debt” (1996, p. 708).

Previous researches have used different measures to be able to classify the firms as financially distressed, as there is still no single measure that is accepted worldwide. Bugeja made use of multiple measures, for example, negative earnings and operating cash flows, negative retained earnings and negative working capital. In addition, he used the Altman Z-score, the Ohlson O-score and the Zmijewski score (2015, p. 363). Amit et al. did use the Altman Z-score as well (1989, p. 147) and Lawrence et al. (2015) also looked at the O-score. Other researchers worked with interest coverage shortfalls (Jostarndt and Sautner, 2008, p. 2190), the Wall Street Journal Index (Kahya et al., 1996, p. 707; Clark and Ofek, 1994, p. 544) or companies that were filed for Chapter 11 (Hotchkiss and Mooradian, 1998, p. 244).

(10)

This thesis will apply the bankruptcy scores of Altman, Ohlson and Zmijewski, which will be further elaborated in the next subparagraph.

2.3.1 Financial Distress Measures

The first measure is the Z-score, developed by Altman (1968), which is often used by researchers to measure financial distress (Bugeja, 2015, p. 367). With 94%-95% of the sample being correctly predicted, it is claimed to be an extremely accurate measure (Altman, 1968, p. 609). Altman came to his idea to develop the Z-score, because there was a gap between the traditional ratio analysis and the more accurate statistical techniques used these days to assess the corporate performance (1968, p. 589). Altman chose to work with the multiple discriminant analysis (MDA), a statistical technique, which can be used to make groups when the dependent variable is qualitative, for example bankrupt and non-bankrupt (1968, pp. 591-592). The discriminant function has the following format: 𝑍 = 𝑣₁ 𝑥1+ 𝑣2 𝑥2+.. . . + 𝑣𝑛 𝑥𝑛. By applying individual ratio’s (xi), one single

Z-score will be obtained. The ratio’s and coefficients (vi) were determined using a sample of

66 firms, whereby half of the firms were bankrupt, according to Chapter X. The final model is defined in the methodology of this thesis (1968, pp. 592-593).

The second measure is the Ohlson O-score (1980). This measure is seen as an improvement compared to the Altman Z-score and correctly measures bankruptcy in 96.12% of the cases (Ohlson, 1980, p. 121). Ohlson started his research due to problems using the MDA, which is used by Altman. Three of the problems, according to Ohlson (1980, p. 112), are:

1. “Certain statistical requirements imposed on the distributional properties of the predictors”.

2. “The output of the model is a score which has little intuitive interpretation, it is basically an ordinal ranking device”.

3. “Certain problems related to the matching procedures, which criteria tend to be somewhat arbitrary”

Therefore, he chose to use the conditional logit analysis (p. 111), with a sample of 105 bankrupt firms. The logit model that has been used to obtain the maximum likelihood estimates (β1, β2, ….) is: 𝑙(𝛽) ≡ ∑_{𝑖 ∈ 𝑆}_𝑖log 𝑃(𝑥_𝑖, 𝛽)+ ∑_{𝑖 ∈ 𝑆}_𝑖log(1 − 𝑃(𝑥_𝑖, 𝛽)), whereby the

first group, S1, is classified as bankrupt and the second group, S2, as non-bankrupt. The

estimates are then obtained by solving: 𝑚𝑎𝑥

𝛽 𝑙(𝛽) (Ohlson, 1980, pp. 117-118). The final

(11)

The last measure that will be discussed is the Zmijewski score (1984). Zmijewski developed this model, because previous models did not account for choice-based sample biases and sample selection biases when using nonrandom datasets, which could lead to biased estimators (1984, p. 59). According to Zmijewski (1984), a choice-based sample arise when the dependent variable is observed first and thereafter the sample will be selected, which lead to oversampling of bankrupt firms (p. 60), and sample selection bias arises when only complete datasets are used, which lead to bankrupt firms being

excluded (p. 62). An available technique that is more appropriate with regard to these issues is the probit or bivariate probit equation (p. 65). To find the best and unbiased parameters, he used a sample of 40 bankrupt firms and 800 non-bankrupt firms (pp. 65-68). This final model is also defined in the methodology part of this thesis.

Agarwal and Tafflet did research to the accuracy of these bankruptcy models and found out that, despite the criticism of the accounting-based models, they are not

outperformed by the recently used option-based models. They even found more significant benefits with respect to the accounting-based models (2008, p. 1550). 2.3.2 Targets in Financial Distress

As explain in the beginning of the paragraph, firms in distress have difficulties paying off their debt, as they do not have enough resources. According to Opler and Titman (1994), these firms are mostly highly leveraged and often harm stakeholders such as suppliers, employees and customers (p. 1015). Customers don’t want to buy anymore and strong competitors try to gain market share by advertising and lower their prices. This is called customer- and competitor-driven losses (p. 1017), which is also confirmed by

Damodaran (2006, p. 612). Erel et al. claim that these firms often have financial constraints and thus do not have access to capital to make positive investments.

Financially distressed firms have a couple of options trying to solve this, namely, restructure its operations voluntarily, restructure themselves via bankruptcy court or become part of another corporation, an acquisition (Jostarndt and Saunter, 2008, p. 2188; Damodaran, 2006, p. 812; Hotchkiss and Mooradian, 1998, p. 241; Clark and Ofek, 1994, p. 541). A takeover seems to be an effective way to restructure the targets (Erel et al., 2015; Damodaran, 2006, p. 612; Hotchkiss and Mooradian, 1998 p. 260; Jensen, 1991). However, what are the reasons for an acquirer to take over a financially distressed firm?

As discussed in the synergy paragraph, there are enough reasons to acquirer another firm, such as growth options due to new products, geographic areas, skills, technologies and monopoly gains, but also cost savings, due to efficiency gains. Khatami et al. found that targets in financial distress have more unexploited opportunities for the

(12)

acquirer and create more chances for future growth (2015, p. 119). Further, firms in financial distress are mainly forced to liquidate their assets at “bargain basement prices” (Jory and Madura, 2009, p. 749; Damodaran, 2006, p. 612; Schleifer and Vishny, 1992). Hotchkiss and Mooradian state that this discount can be about 45% less than in the case of non-bankrupt targets (1998, p. 243). In addition, Berk and DeMarzo claim that if stockholders are unhappy about the current situation and their executive managers, they are willing to sell their stock at discounts, compared to a more viable situation with capable leadership. A strong firm can take advantage of these discounted prices and replace the inefficient managers (2011, p. 896; Jory and Madura, 2009, p. 750; Kahya et al., 1996, pp. 704-705). These synergies, discounted prices and an inefficient

management are good reasons to takeover a financially distressed firm (Jory and Madura, 2009, p. 758; Kahya, 1996, p. 715). Another reason could be tax advantages after the takeover, due to the depreciation of new assets and the carryforwards of losses taken over (Berk and DeMarzo, 2011, p. 500; Clark and Ofek, 1994, p. 548; Amit et al., 1989). Previous researchers also point out that these targets are mainly small and young, which do not pay taxes yet and are easier to restructure (Berk and DeMarzo, 2011, p. 499; Clark and Ofek, 1994, p. 561).

3. Methodology and Hypothesis

In this chapter the model will be discussed to test the hypothesis and to answer the research question.

3.1 Methodology

First, in order to answer the research question and assess the performance of the acquirer after the takeover, the model that is used will be explained, called the random effects model. Further, the measures to determine if the target is in financial distress will be discussed.

3.1.1 Panel Data Model

In this research, the difference in post-performance of the acquirer will be examined when it is taking over a financially distressed firm compared to a firm outside financial distress. To measure this difference, a panel data regression will be used. The reason to use a panel data regression is that the performance, the dependent variable, can be studied over time (Stock and Watson, 2012, pp. 389-390). According to dr. J. C. M van Ophem (personal communication, August 11, 2016), the two panel data models that can

(13)

be used are the fixed effects model and the random effects model. Further, he stated that in most cases the fixed effects model will be used, whereby the entity and time fixed effect variables are automatically deleted from the model, which means that only the variables that vary across entities and over time will be kept. This is also confirmed by Stock and Watson (2012, pp. 396-399). However, in this thesis, the independent variable of interest is constant over time. Therefore, the random effects model will be used instead, whereby those variables will not be removed from the model (dr. J. C. M van Ophem, personal communication, August 11, 2016). This model uses the generalized least squares (GLS) technique to compute the unknown parameters of the model, compared to the ordinary least squares (OLS) technique by the fixed effects model.

As explained before, the two main types of synergies, according to Berk and DeMarzo (2011), are revenue enhancement and cost reduction. First of all, they state that, due to combining two firms, extra products can be added or new markets can be entered, which can result in more customers and subsequently increase revenues (economies of scope). This is also stated by Lambkin and Rahman (2015). Therefore, the first dependent variable that will be used to compare the performance after the deal is Sales revenue, measured as sales revenue divided by total assets, which can also be called the asset turnover (Kahya et al., 1996, p. 704; Healy et al., 1992, p. 153). This variable will measure how much revenue will be earned with respect to the total assets available to the acquirer, before and after the deal.

In addition, Lambkin and Rahman point out that cost synergies, which may result due to savings in personnel or sale of surplus assets, arise from economies of scale. This is measured by the ratio of selling, general and administrative (SGA) expenditure to sales revenue in each year and will, in this thesis, be used as dependent variable in the second model (2015, p. 28). This is the second variable that measures the performance after the deal. Eventually, if the revenues have increased and the expenditures have reduced, this could lead to operating return improvements, which, in the last model, will be measured by the return on sales: EBIT/sales revenue, like has been done by Lamkin and Rahman as well (Cao et al., 2016, p. 380; Balcaen et al., 2012, p. 960; Hotchkiss and Mooradian, 1998, pp. 253-254; Clark & Ofek, 1994, p. 550; Healy et al., 1992, p. 139). According to Cao et al., accounting measures, like return on assets and return on sales, are the most useful measures in order to study synergies (2016, p. 379). In addition, Hotchkiss and

Mooradian (1998) point out that return on sales is good measure due to the fact that it is not influenced by the different accounting treatments used by firms.

(14)

The regression models look as follows: (1) 𝑆𝑎𝑙𝑒𝑠 𝑟𝑒𝑣𝑒𝑛𝑢𝑒 𝑇𝑜𝑡𝑎𝑙 𝑎𝑠𝑠𝑒𝑡𝑠 𝑖𝑡 = 𝛽0+ 𝛽1 𝐹𝑖𝑛𝑎𝑛𝑐𝑖𝑎𝑙 𝑑𝑖𝑠𝑡𝑟𝑒𝑠𝑠𝑖+ 𝛽2 𝐴𝐹𝑇𝐸𝑅𝑖𝑡 + 𝛽3 𝐹𝑖𝑛𝑎𝑛𝑐𝑖𝑎𝑙 𝑑𝑖𝑠𝑡𝑟𝑒𝑠𝑠𝑖∗ 𝐴𝐹𝑇𝐸𝑅𝑖𝑡 + 𝛽4 𝐶𝑎𝑠ℎ 𝑎𝑐𝑞𝑢𝑖𝑟𝑒𝑟𝑖𝑡 + 𝛽5 𝐶𝑎𝑠ℎ 𝑡𝑎𝑟𝑔𝑒𝑡𝑖𝑡+ 𝛽6 𝑇𝑜𝑡𝑎𝑙 𝑎𝑠𝑠𝑒𝑡𝑠 𝑎𝑐𝑞𝑢𝑖𝑟𝑒𝑟𝑖𝑡 + 𝛽7 𝑇𝑜𝑡𝑎𝑙 𝑎𝑠𝑠𝑒𝑡𝑠 𝑡𝑎𝑟𝑔𝑒𝑡𝑖𝑡+ 𝛽8 𝑅𝑒𝑙𝑎𝑡𝑖𝑣𝑒 𝑠𝑖𝑧𝑒 𝑡𝑎𝑟𝑔𝑒𝑡𝑖𝑡 + 𝛽9 𝐸𝑚𝑝𝑙𝑜𝑦. 𝑛𝑟. 𝑎𝑐𝑞𝑢𝑖𝑟𝑒𝑟𝑖𝑡+ 𝛽10 𝐸𝑚𝑝𝑙𝑜𝑦. 𝑛𝑟. 𝑡𝑎𝑟𝑔𝑒𝑡𝑖𝑡 + ∑ 𝛾𝑗𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 𝑎𝑐𝑞𝑢𝑖𝑟𝑒𝑟𝑗 48 𝑗=1 + ∑ 𝛿𝑗𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 𝑡𝑎𝑟𝑔𝑒𝑡𝑗 48 𝑗=1 + 𝜀𝑖𝑡 (2) 𝑆𝐺𝐴 𝑒𝑥𝑝𝑒𝑛𝑠𝑒𝑠 𝑆𝑎𝑙𝑒𝑠 𝑟𝑒𝑣𝑒𝑛𝑢𝑒𝑖𝑡 = 𝛽0+ 𝛽1 𝐹𝑖𝑛𝑎𝑛𝑐𝑖𝑎𝑙 𝑑𝑖𝑠𝑡𝑟𝑒𝑠𝑠𝑖+ 𝛽2 𝐴𝐹𝑇𝐸𝑅𝑖𝑡 + 𝛽3 𝐹𝑖𝑛𝑎𝑛𝑐𝑖𝑎𝑙 𝑑𝑖𝑠𝑡𝑟𝑒𝑠𝑠𝑖∗ 𝐴𝐹𝑇𝐸𝑅𝑖𝑡+ 𝛽4 𝑅𝑒𝑙𝑎𝑡𝑖𝑣𝑒 𝑠𝑖𝑧𝑒 𝑡𝑎𝑟𝑔𝑒𝑡𝑖𝑡 + 𝛽6 𝐸𝑚𝑝𝑙𝑜𝑦. 𝑛𝑟. 𝑎𝑐𝑞𝑢𝑖𝑟𝑒𝑟𝑖𝑡+ 𝛽7 𝐸𝑚𝑝𝑙𝑜𝑦. 𝑛𝑟. 𝑡𝑎𝑟𝑔𝑒𝑡𝑖𝑡 + ∑ 𝛾𝑗𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 𝑎𝑐𝑞𝑢𝑖𝑟𝑒𝑟𝑗 48 𝑗=1 + ∑ 𝛿𝑗𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 𝑡𝑎𝑟𝑔𝑒𝑡𝑗 48 𝑗=1 + 𝜀𝑖𝑡 (3) 𝑅𝑒𝑡𝑢𝑟𝑛 𝑜𝑛 𝑠𝑎𝑙𝑒𝑠_𝑖𝑡 = 𝛽0+ 𝛽1 𝐹𝑖𝑛𝑎𝑛𝑐𝑖𝑎𝑙 𝑑𝑖𝑠𝑡𝑟𝑒𝑠𝑠𝑖+ 𝛽2 𝐴𝐹𝑇𝐸𝑅𝑖𝑡 + 𝛽3 𝐹𝑖𝑛𝑎𝑛𝑐𝑖𝑎𝑙 𝑑𝑖𝑠𝑡𝑟𝑒𝑠𝑠𝑖∗ 𝐴𝐹𝑇𝐸𝑅𝑖𝑡 + 𝛽4 𝐶𝑎𝑠ℎ 𝑎𝑐𝑞𝑢𝑖𝑟𝑒𝑟𝑖𝑡 + 𝛽5 𝐶𝑎𝑠ℎ 𝑡𝑎𝑟𝑔𝑒𝑡𝑖𝑡+ 𝛽5 𝑅𝑒𝑙𝑎𝑡𝑖𝑣𝑒 𝑠𝑖𝑧𝑒 𝑡𝑎𝑟𝑔𝑒𝑡𝑖𝑡 + 𝛽7 𝐸𝑚𝑝𝑙𝑜𝑦. 𝑛𝑟. 𝑎𝑐𝑞𝑢𝑖𝑟𝑒𝑟𝑖𝑡+ 𝛽8 𝐸𝑚𝑝𝑙𝑜𝑦. 𝑛𝑟. 𝑡𝑎𝑟𝑔𝑒𝑡𝑖𝑡 + ∑ 𝛾𝑗𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 𝑎𝑐𝑞𝑢𝑖𝑟𝑒𝑟𝑗 48 𝑗=1 + ∑ 𝛿𝑗𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 𝑡𝑎𝑟𝑔𝑒𝑡𝑗 48 𝑗=1 + 𝜀𝑖𝑡

3.1.2 The Dependent and Independent Variables

The first dependent variable, Sales revenue to Total assets, also called asset turnover, is applicable to the acquirer and isavailable in the three fiscal years before and after the deal. It measures the efficiency with which an acquirer is able to generate revenues from their assets. The second dependent variable is the ratio of Selling, general and

administrative (SGA) expenses to Sales revenue, which is also applicable to the acquirer and available in the three fiscal years before and after the deal. The last dependent variable, Return on sales, is the ratio of earnings before interest and taxes to sales revenue, again available in the three fiscal years before and after the deal and also a measure for the acquirer. These measures will be studied before and after the deal to find

(15)

out if there will be an improvement at all when the acquirer is taking over a firm. In the models, this is indicated by the variable AFTER, which is a dummy variable that equates 1 if it is one of the years after the deal and 0 if it is one of the years before the deal. Erel et al. have used this variable as well (2015, p. 301). Further, the main idea of this thesis is to figure out if the Return on sales of the acquirer will improve (more) when it is taking over a financially distressed firm instead of a firm outside financial distress, due to an increase in Sales revenue and a decrease in the SGA expenditure to Sales revenue. To indicate the financial situation of the target, the Financial distress dummy variable has been added to the model. This variable equates 1 if the target was in financial distress in the year before the deal and 0 if the target was “financially healthy” in the year before the deal (Clark and Ofek, 1994, p. 558; Amit et al., 1989, p. 149). How the variable Financial distress is

measured, is explained in the fifth subparagraph. In addition, the interaction variable Financial distress * AFTER is included in the model to capture the combination effect of the two individual dummy variables (Stock and Watson, 2012, p. 318). Erel et al. did it in their research as well (2012, p. 305).

3.1.3 Interpretation of the Independent Variables

First, when looking at the two independent indicator variables, Financial distress and AFTER, and their interaction term, Financial distress * AFTER, we can conclude that there are only 3 coefficients that can be computed, while there are four situations that can be interpreted. These four situations are:

1. Before acquiring a “healthy” firm 2. After acquiring a “healthy” firm 3. Before acquiring a distressed firm 4. After acquiring a distressed firm

By comparing and testing some of these situations, the research question can be answered.

When looking at the models (1), (2) and (3), the coefficient that describes situation 1, before taking over a “healthy” target, is dropped out of the model. This is to prevent the regressions from perfect multicollinearity due to the dummy variable trap (Stock and Watson, 2012, p. 243).

The first coefficient that can be computed is therefore β1, which describes, in each

model, the increase or decrease in the mean ratio of the dependent variable, while comparing the situation before taking over a financially distressed target with the

(16)

situation before taking over a “healthy” firm (3 compared to 1).1_β₁_{is not of main interest,}

as this thesis is interested in this difference after the takeover. The sign of this coefficient, in each of the three models, can be positive or negative, because there is no clear reason to determine that an acquirer in one of the two situations earn more revenues or face more SGA expenses than the other. This also applies to the return on sales.

The second coefficient is from the variable AFTER, β2. This variable can be

interpreted as the increase or decrease in the mean ratio of the acquirer’s sales revenue, SGA expenses and return on sales, while comparing the situation after acquiring a “healthy” target with before (2 compared to 1).1_{The sign of this coefficient will be}

positive with respect to the dependent variable Sales revenue to Total assets and negative with respect to SGA expenses to Sales revenue, because it can be expected that taking over “healthy” firm will create more (revenue and cost) synergies and therefore increases the revenues and decreases the SGA expenses of the acquiring firm with respect to before (Lambkin and Rahman, 2015, p. 28; Hoberg and Phillips, 2010, p. 3775; Lambrecht, 2004, p. 43). This will subsequently result into a positive sign in the third model (3), as the return on sales will logically improve.

At last, the coefficients β1 + β2 + β3, will, in each model, represent the increase or

decrease in the mean ratio of the dependent variable, while comparing the situation after taking over a financially distressed firm with before taking over a non-distressed firm (4 compared to 1).1_{This interpretation is not something this thesis is interested in.}

This research is actually interested in comparing situation 4 with 3, to study the difference in the dependent variables while comparing after taking over a financially distressed firm E(Yi| β1i = 1, β2i = 1, β3i = 1) with before E(Yi| β1i = 1, β2i = 0, β3i = 0), and

situation 4 with 2, to study the difference in the dependent variables while comparing after acquiring a distressed firm E(Yi| β1i = 1, β2i = 1, β3i = 1) with after acquiring a

non-distressed firm E(Yi| β1i = 0, β2i = 1, β3i = 0).1

The coefficients that belong to this first comparison are therefore β2 + β3, and the

coefficients that belong to the second comparison are β1 + β3. To know if these differences

are statistically significant, a chi2_{-test will be performed, which tests the null hypotheses}

β2 + β3 =0 and β1 + β3 =0against the alternative hypotheses β2 + β3 ≠ 0 and β1 + β3

respectively. If the alternative hypotheses are true, we can conclude that the differences are statistically significant at a 1%, 5% or 10% level (Stock and Watson, 2012, pp. 268-269). The sign of β2 + β3 together is expected to be positive for the dependent variable

Sales revenue to Total assets, because the new resources, products, production lines,

(17)

geographic areas and/or extra market share that arise from taking over a financially distressed firm, will increase the revenues of the acquirer after the takeover (Lambkin and Rahman, 2015; Hoberg and Phillips, 2012; Berk and DeMarzo, 2011). The sign is expected to be negative when looking at the dependent variable SGA expenses to Sales revenue, because due to economies of scale, which arise from taking over a distressed target, the SGA expenses will decrease (Lambkin and Rahman, 2015; Berk and DeMarzo, 2011; Devos et al, 2009). At last, the sign is predicted to be positive with regard to the dependent variable Return on sales, because higher revenue and lower SGA expenses result into higher earnings before interest and taxes (Lambkin and Rahman, 2015). When looking at the sign of β1 + β3, with respect to the different dependent variables, we

expect it to be positive for a Sales revenue to Total assets, negative for SGA expenses to Sales revenue and again positive for Return on sales. This is due to the fact that distressed targets, which are mainly small, are easier to restructure and take over (Berk and

DeMarzo, 2011; Clark and Ofek, 1994). In addition, the assets of distressed targets are now deployed by more capable and more efficient managers, while healthy firms bring assets that were mainly already deployed by capable and efficient managers (Berk and DeMarzo, 2011, p. 896). Therefore, the distressed firms will create more value for an acquirer, due to the growth opportunities than a healthy target will do (Jory and Madura, 2009, p. 750).

3.1.4 Control variables and Interpretation

As can be seen in the regression models, a couple of control variables are added to the regression equations, for example firm and industry characteristics.

The first one, Cash acquirer, is defined as the amount of cash divided by total assets of the acquirer in each fiscal year. The more cash available, the more access to capital, which can be used in the process of making (new) products and the daily operations of the firm (Cao, 2016, p. 373; Bugeja, 2015, p. 370; Erel et al., 2015, p. 290). Therefore, the sign of this coefficient will be positive, because it can increase revenues. The second variable, Cash target, is also defined as the amount of cash divided by the total assets, but in this case with respect to the target firm. This variable is mainly

available for the three years preceding the deal. When taken over and being restructured, holding an amount of cash provides access to capital, which can be used for the further business and improvements of the merged firm (Bugeja, 2015; Balcaen et al., 2012, p. 960). This means that the sign of this coefficient will be positive. The two cash variables are added in the first (1) and third model (3).

(18)

The variable Relative size target is defined as the total assets of the targets divided by the total assets of the acquirer, like has been done by Cao et al. (2016). This control variable will be used in all the three models. When looking at the first model (1), it can be stated that the bigger the target, the more experience, knowledge and assets will flow over to the acquiring firm, which can result in more revenues (Rahman and Lambkin, 2015; Balcaen et al., 2012, p. 952; Kahya et al., 1997, p. 704). The coefficient will therefore have a positive sign. When looking at the second model (2), Rahman and Lambkin (2015) point out that a bigger firm with more employees will face more salary costs, but on the other side, this bigger firm is better able to spread the SGA expenses due to economies of scale. In addition, they state that, most of the time, duplicate costs will disappear after the acquisition. Therefore size, being or becoming large, will have a negative effect on the SGA expenses to Sales revenue. The sign of the coefficients is predicted to be negative. With respect to the third model (3), the bigger the target firm, the more revenues the acquirer can create and the lower the SGA expenses will be, thus the higher the return on sales. A positive sign will be expected.

In the first model (1), the dollar amount of total assets, with respect to the acquirer and the target, has been included as the 4th_{and 5}th_{control variable, Tot. asset}

acquirer and Tot. asset target. Balcaen et al. (2012), Hotchkiss and Mooradian (1998) and Opler and Titman (1994) did the same. Kayha et al. claim that the more assets a firm have, fixed or variable, the more products can be produced and the more revenue will be earned (Lambrecht, 2004, p. 42). The coefficients will therefore have a positive sign.

The 6th_{and 7}th_{control variable is Employ. nr. target and Employ. nr. acquirer}

respectively, which is defined as the natural logarithm of the number of employees. It can be included as a firm growth variable (Kahya et al., 1996, p. 705), but the number of employees also has a positive impact on the production and thus revenue, a negative impact on the SGA expenses, because it leads to more salary, and again a positive impact on the growth opportunities and thus operating and profitability improvements (Erel et al., 2015, p. 301; Rahman and Lambkin, 2015; Hotchkiss and Mooradian, 1998, pp. 257-258). A positive, negative and positive sign is predicted respectively.

Further, this thesis control for the effects of omitted variables that differ across industries,by adding the fixed effectsvariables, Industry acquirer and Industry target. Both variables are binary and divided into the 48 Fama-French industries (Fama and French, 1997, p. 179). In addition, an extra industry group has been added, group 0, which contain the SIC codes that were not assigned to one of the 48 groups. Therefore, these models have 49-1 intercepts for the acquirer industry and 49-1 intercepts for the target industry. In both cases, one of the industries is excluded from the model to prevent

(19)

them from perfect multicollinearity. The classification of the industries is based on the four digit SIC codes and an overview of the 48 industries can be found in Appendix 1. 3.1.5 Financial Distress Measures

As explained before, the target firms are divided into two groups, financially distressed targets and targets outside financial distress. To be able to classify the firms into the two groups, three bankruptcy prediction models are used, as Bugeja did as well in his

research (2015, p. 363). The measures were already explained in the literature review: 1. Altman’s Z-score

2. Ohlson’s O-score 3. Zmijewski’s score

If at least two of these measures suggest that a target firm is in financial distress the year before the deal, this thesis assesses the target as financially distressed. As mentioned before, the variable Financial distress is a binary variable, which thus equates 1 if at least two of these measures meet their benchmark.

3.1.5.1 Altman’s Z-score

The discriminant function of the Z-score is modeled as follows (Bugeja, 2015; Altman, 1968):

𝑍 = 1.2𝑥1+ 1.4𝑥2+ 3.3𝑥3+ 0.6𝑥4+ 0.999𝑥5, where

x1 = Working capital/total assets

x2 = Retained earnings/total assets

x3 = Earnings before interest and taxes/total assets

x4 = Market value equity/book value of total debt

x5 = Sales/total assets

According to Altman (1968), if the Z-score is higher than 2.99, it means that the company is not classified as bankrupt. However, if the Z-score is lower than 1.81, the company is. The area between 1.81 and 2.99 is called the “grey area”, “because of the susceptibility to error classification” (p. 606). Eventually 2.675 is chosen as benchmark to divide the firms into two groups. This critical value had the least misclassified firms in his research (p. 607). Bugeja and Amit et al. have used this benchmark as well. (2015, p. 147; 1989, p. 377)

(20)

3.1.5.2 Ohlson’s O-score

The Ohlson bankruptcy model is, as explained before, a logit model, which avoids the known problems of the Z-score model. The model is presented as follows (Ohlson, 1980, p. 112):

𝑂 = −1.32 − 0.407𝑥1+ 6.03𝑥2− 1.43𝑥3+ 0.0757𝑥4− 1.72𝑥5− 2.37𝑥6− 1.83𝑥7

+ 0.285𝑥8− 0.521𝑥9

whereby the accounting ratio’s, 𝑥_𝑖, are (Ohlson, 1980, pp. 118-119; Bugeja, 2015, p. 368): x1 = Size = log(total assets/GNP price-level index)

x2 = Debt-to-assets ratio = total liabilities/total assets

x3 = Working capital/total assets

x4 = Current liabilities/current assets

x5 = Dummy, indicating negative net assets (= 1 if total liabilities > total assets)

x6 = Return on assets (ROA) = net income/total assets

x7 = Funds from operations/total liabilities

x8 = Dummy indicating loss in the past two years (=1 if net income was negative)

x9 = Change in net income = (NIt – NIt-1) / NIt-1

Due to the lack of the GNP price-level index in the different databases, the consumer price index has been used as approximation. The consumer price index is the price index that is most frequently used. According the website of The World Bank

(http://data.worldbank.org/indicator/FP.CPI.TOTL), it is the price of a basket of goods and services in an economy to the average consumer, while the GNP price level include all final products produced.

In the end, the model is applied to the sample and the results are transformed to compute the implied probability. The transformation has been done with this formula (Lawrence et al., 2015, p. 2047; Ohlson, 1980, p. 118):

1 1 + 𝑒(−𝑂)

If the implied probability is bigger than 0.5, the target is defined as financially distressed (Bugeja, 2015, p. 377).

3.1.5.3 Zmijewski’s score

For the bankruptcy measure of Zmijewski, three ratios will be examined using the probit model (Bugeja, 2015, p. 368; Zmijewski, 1984, pp. 65-66):

x1 = Return on assets (ROA) = net income/total assets

(21)

x3 = Current ratio = current assets/current liabilities

These ratios are implemented in the following equation: 𝑍𝑚 = − 4.336 − 4.513𝑥1+ 5.679𝑥2+ 0.004𝑥3

To compute the probability of default, the Zm-score has been transformed using the

normal distribution function: 𝜙(𝑍_𝑚) = 𝑃(𝑍 ≤ 𝑍𝑚). Again, if this probability is higher

than 0.5, the target firm will be classified as financially distressed (Bugeja, 2015, p. 377).

3.2 Hypothesis

The research question that will be answered in this thesis is: Does a firm in financial distress have a positive effect on the post-operating performance of the acquirer? And if yes, will this improvement be higher than when it had taken over a non-distressed firm? Post-performance is measured by return on sales, revenue and cost synergies.

After studying the exiting literature, this thesis hypothesizes that:

1. Taking over a target in financial distress will create value (revenue and cost synergies) for the acquirer, which result in higher return on sales.

2. This created value will be higher when buying a firm in distress, compared to a firm outside distress.

Those two main hypotheses can be divided into smaller expectations:

1.1. Taking over a financially distressed firm will increase the sales revenue of the acquirers, due to the fact that the acquirer is able to redeploy the combined assets more efficiently and now engage in new product markets and geographic areas (Lambkin and Rahman, 2015, p. 28; Berk and DeMarzo, 2011, p. 895; Hoberg and Phillips, 2010, p. 3809; Devos et al., 2009, p. 1208; Jory and Madura, 2009, p.750; Damodaran, 2006, p. 460; Bradley et al., 1988, p. 4).

1.2. Taking over a financially distressed firm will reduce the selling, general and administrative expenses of the acquirer, due to efficiency gains and

economies of scale. This is possible, for example, because the acquirer is, after the takeover, able to eliminate overlapping operating expenses, such as salaries and marketing costs, while still producing the same amount of products (Lambkin and Rahman, 2015; Berk and DeMarzo, 2011; Devos, 2009, p. 1208; Damodaran, 2006; Hotchkiss and Mooradian, 1998, pp. 260-261).

(22)

1.3. Higher revenues and lower SGA expenses after the takeover logically result into higher return on sales for the acquirer, which is measured by earnings before interest and taxed divided by sales revenue (Hotchkiss and Mooradian, 1998, p. 260).

2.1.

This created value for the acquirer (higher revenues, lower SGA expenses and higher returns on sales), will increase more when it takes over a distressed firm versus a non-distressed firm. According to Berk and DeMarzo and Clark and Ofek (1994), this is due to the fact that distressed firms are often cheap, young and small, which is easier to restructure or acquire (2011, p. 897). In addition, Berk and DeMarzo (2011) state that distressed firms are mainly run by more incapable and inefficient managers, compared to “healthy” firms. This means that an acquirer is better able to redeploy the less efficiently deployed resources of the distressed firm than the already efficiently deployed assets of the non-distressed firm and thus extra revenue can be created. This is also confirmed by Jory and Madura (2009) and Kahya et al. (1996). In addition, this also applies to the increase in efficiency with respect to eliminating overlapping costs. The same amount of products can be produced with less and cheaper resources (Damodaran, 2006, p. 460; Hotchkiss and Mooradian, 1998, pp. 260-261; Kahya et al., 1996, p. 704). At last, the higher revenues and lower SGA expenses will logically results into higher returns on sales, due to the unexploited efficiency opportunities a distressed target will bring (Khatami et al, 2015, p. 115; Hotchkiss and Mooradian, 1998, p. 260).

(23)

4. Data and Descriptive Statistics

In this chapter, the data sources will be listed together with the requirements that were met to obtain the data. In addition, the descriptive statistics will be shown.

4.1 Data

The primary database that has been used to find M&A deals is the Thomson One database. It contains a broad range of financial content from annual reports, as well as information about M&A deals. From this database, 14,632 deals were gathered, together with deal-related information, for example, the deal value, the deal year and firm nations. A couple of requirements were met during the sample selection:

1. Deal type: mergers and acquisitions.

2. Company status: both the acquirer and target are publicly listed. 3. Percentage of final stake: 50% to 100%.

4. Deal status: completed.

5. Geographical area acquirer: Europe and the Unites States. 6. Time period: 1978 – 2015.

The reason that this thesis has only selected acquirers and target that are publicly listed, is to be able to access as much as financial information as possible. The advantage of a listed company is that they are obligated to disclose their financial and non-financial data. Further, like has been chosen in the research of Lambkin and Rahman (2015, p. 29), the minimum percentage of final stake that an acquirer is holding is 50%, which means that it will have control over the future business of the target. Moreover, all the deals are

completed, so that it is possible to measure the performance of the acquirer after the takeover. Another requirement that has been made is that this thesis will only contain acquiring companies from Europe and the United States. Targets, on the other hand, can come from anywhere else in the world. The total time period that will be examined is 1978 until 2015. The time period in which the deals can occur is 1981-2012, which makes it possible to gather enough information about the performance of the acquirer before and after the acquisition. Like Lambkin and Rahman has been done, this thesis will use (financial) information from three years before the deal and three years after the deal. The deal year itself will be excluded (2015, p. 29; Jory and Madura, 2009, p. 750).

The second database is Datastream. Datastream is an extensive database, which delivers economic data about companies and countries worldwide. It has been used to retrieve information about the performance of the target before the takeover, is it in financial distress or not, and about the performance of the acquirer before and after the takeover. Jostarndt and Sautner have used this database as well (2008, p. 2191).

(24)

Almost all variables that are needed in the research models are found in this database. The only two variables that are gathered from another source are industry and the consumer price index. The SIC codes that has been used to create the 48 different industries are obtained from the Thomson One database, while, as explained in the methodology, the 48 industry groups are listed by Fama and French (1997). The consumer price index has been acquired from the website of The World Bank (http://data.worldbank.org).

At the end, after going through the whole dataset and dismissing all the deals that contain errors, and therefore cannot be used, 6459 deals remain.

4.2 Descriptive Statistics

This chapter presents overview of the summary statistics. Table 1 provides the statistics of all M&A deals. Table 2 and 3 show the statistics of acquirers and “healthy” targets before and after the deal respectively. Table 4 and 5 demonstrate the statistics of acquirers and distressed targets before and after the deal respectively. At last, table 6 contains the cross-correlation matrix.

Table 1.

Descriptive statistics of all M&A deals in the period 1978 – 2015.

Variables Obs. Mean Median SD Min Max

Sales revenue 42014 0.742 0.620 0.658 0.010 3.379

SGA expenses/Sales revenue 29729 0.308 0.228 0.287 0.027 1.495

Return on sales 40674 0.231 0.141 0.301 0.011 1.654

Financial distress 45213 0.052 0 0.222 0 1

AFTER 38754 0.500 0.500 0.500 0 1

Cash acquirer 41969 0.124 0.062 0.160 0 1

Cash target 21955 0.164 0.067 0.216 0 1

Total assets acquirer 42079 3,400,000 3,231,207 47,100,000 42,430 188,000,000

Total assets target 22189 3,337,294 319,823 7,954,453 9,844 32,800,000

Relative size target 21320 0.645 0.188 4.798 0 532.640

Ln(Employ. number acquirer) 42146 8.525 8.780 2.351 0 14.604

Ln(Employ. Number target) 40654 6.566 6.531 2.022 0 14.207

Industry acquirer 45213 29.904 34 15.392 0 48

Industry target 25836 29.722 34 15.253 0 48

Notes: Sales revenue is measured as the sales divided by total assets, also called asset turnover. It has been winsorized at a 1% level. SGA expense divided by Sales revenue is winsorized at a 2.5% level. Return on sales is defined as EBIT divided by sales and is winsorized at a 2.5% level. Financial distress and AFTER are both dummy variables, whereby Financial distress indicates if a target firm is in financial distress or not and AFTER indicates the period before or after the deal. Cash acquirer and Cash target are both computed as cash divided by total assets. Total asset acquirer and Total assets target are both defined as the dollar amount and they are winsorized at a 5% level. Ln(Employ.number acquirer) and Ln(Employ.number acquirer) are both the natural logarithm of the employee numbers. Relative size target is defined as the total assets of the target divided by the total assets of the acquirer. Industry acquirer and Industry target are both dummy variables that indicate to which Fama-French industry group the acquirer and target belong.

(25)

The first column of table 1 presents the number of observations of each variable. The maximum number of observations is 45213, because there are 6459 deals and seven years per deal to be observed, namely three years before the deal, the deal year itself and 3 years after the deal. The target variables, Cash target, Total asset target, Relative size target and Industry target, have much less observations, this is due to the fact that most of the targets have only observations before the deal and in the deal year. The variable AFTER has less observations, because it indicates the period before and after the deal, which is three year before and three years after the deal, so the deal year has been excluded. The acquirer variables, are all above the 40,000 observations, except for SGA expenses to Sales revenue, which has a lot of missing variables.

The second column shows the mean of each variable. Here can be seen that the mean of Sales revenue for all M&A deals is 74.2%, which is quite high. The mean of SGA expenses/Sales revenue is 30.8% and the mean of Return on sales is 23.1%.

The third column shows the median of each variable. Again the variable Sales revenue scores quite high. Here can also be seen that there are mainly non-distressed firms in the sample, as the variable Financial distress has a median of 0. Further, most M&A deals take place in the Fama-French industry Business Services.

The fourth column displays the standard deviation, which ranges from 0.160 to 15.392, plus two large amounts for the variables Total assets acquirer and Total assets target.

The last two columns report the minimum and maximum observation of each variable. From these columns can be seen that Financial distress and AFTER are indeed indicator variables, as they range from 0 to 1. However, the variables Cash acquirer and Cash target are not indicator variables, but they cannot be bigger than one, as cash can never be higher than the total assets.

Column 1 in table 2 shows the number of observations for all “healthy” targets and for the acquirers that are going to take over the “healthy” targets. When looking at the variable Financial distress, AFTER and industry, it can be seen that the maximum number of observations is 18369, which means that there are 18369/3 years before the deal = 6123 healthy targets.

In the other columns can found that the average asset turnover for an acquirer that takes over a “healthy” target is 76.6% before the deal, the average SGA expenses to Sales revenue is 31.7% and the average Return on sales is 23.6%. The average number of employees for a “healthy” targets, before the deal, is 6.542, and the maximum number is 12.751. Most of the “healthy” targets are from the Business Service Fama-French industry

(26)

group. The same applies to their acquirers. The indicator variables Financial distress and AFTER are 0 in columns 2 until 6, because there are no financial distressed firms in this table and it indicates the period before the deal.

Table 3 shows the summary statistics of the same acquirers and targets as in table 2, but now it indicates the period after the deal. This is shown by the variable Financial distress, which is still 0, and the variable AFTER, whereby the median is 1, as well as the minimum and maximum value.

Further the average Sales revenue of the acquirer after the takeover is 71.7%, which has been decreased compared to before the deal. The average SGA expenses to Sales revenue is 29.8% and the average Return on sales is 22.6%, which both went down as well. The number of employees for the acquirer increased from 8.366 to 8.671 and the amount of total assets increased from $20,100,000 to $27,300,000.

Table 2.

Descriptive statistics of acquirers and “healthy” targets, before the deal, in the period 1978 – 2015.

Variables Obs Mean Median SD Min Max

Sales revenue 17483 0.766 0.651 0.682 0.010 3.379

Return on sales 16812 0.236 0.148 0.300 0.011 1.654

Financial distress 18369 0 0 0 0 0

AFTER 18369 0 0 0 0 0

Cash acquirer 17468 0.141 0.066 0.182 0 1

Cash target 16409 0.174 0.070 0.223 0 1

Total assets acquirer 17511 20,100,000 2,342,790 43,900,000 42,430 188,000,000

Total assets target 16567 2,805,341 291,614 7,172,586 9,844 32,800,000

ln(Employ. number acquirer) 16307 8.366 8.601 2.398 0 14.557

ln(Employ. Number target) 15540 6.542 6.467 2.008 0 12.751

Industry target 18369 29.943 34 15.242 0 48

Notes: Sales revenue is measured as the sales of the acquirer divided by total assets of the acquirer, also called asset turnover. It has been winsorized at a 1% level. SGA expense divided by Sales revenue is information of the acquirer only and it is winsorized at a 2.5% level. Return on sales is defined as EBIT of the acquirer divided by sales of the acquirer and is winsorized at a 2.5% level. Financial distress and AFTER are both dummy variables, whereby Financial distress indicates if a target firm is in financial distress or not and AFTER indicates the period before or after the deal. Cash acquirer and Cash target are both computed as cash divided by total assets. Total asset acquirer and Total assets target are both defined as the dollar amount and they are winsorized at a 5% level. Ln(Employ.number acquirer) and Ln(Employ.number acquirer) are both the natural logarithm of the employee numbers. Relative size target is defined as the total assets of the target divided by the total assets of the acquirer. Industry acquirer and Industry target are both dummy variables that indicate to which Fama-French industry group the acquirer and target belong.

(27)

Table 3.

Descriptive statistics of acquirers and “healthy” targets, after the deal, in the period 1978 – 2015.

Sales revenue 16476 0.717 0.594 0.639 0.010 3.379

Return on sales 16015 0.226 0.138 0.294 0.011 1.654

AFTER 18369 1 1 0 1 1

Cash acquirer 16451 0.111 0.058 0.142 0 1

Cash target 2761 0.136 0.059 0.196 0 1

Total assets acquirer 16496 27,300,000 4,553,136 50,600,000 42,430 188,000,000 Total assets target 2787 6,387,746 644,453 11,000,000 9,844 32,800,000

Industry target 18369 29.943 34 15.242 0 48

Notes: The same notes as in table 2 are also applicable in this table.

Table 4 and 5 present the summary statistics for the financially distressed targets and their acquirers, before and after the deal respectively. In the first column of each table can be seen that the total number of observations is 1008. This means that there are 1008/3 years before the deal = 336 distressed targets in this sample. The variable Financial distress is now equal to 1, as shown by the mean, minimum and maximum value. The second column of table 4 reveals that the average asset turnover for the acquirer is 94.3% before the deal. In addition, the average SGA expenses to Sales revenue is 31.9% and the average Return on sales is 20.4%. In the second column of table 5 can be seen that these percentages all decreased: 92.4%, 28.8% and 18.4% respectively. Further, the average number of employees of a distressed target is 7,010 before the deal. The average number of employees of the acquirer has increased from 8.631 to 8.907 after the deal. When looking at the difference in total assets for the acquirer, this has increased from $14,300,000 to $20,100,000. When looking at the median of the industry variable, in both tables can be found that distressed targets, and their acquirers, are mainly from the Communication Fama-French industry group.

Table 6 shows the cross-correlations between the dependent variables and the

independent variables. The lowest correlation is between the dummy variables, namely 0.000, which indicate the weakest relation. The strongest relation is between the target and acquirer industry variables. Other strong relations are also between the same target and acquirer variables, for example cash, total assets and the number of employees.

(28)

According to the table, the relations between the dependent variables and the Financial distress or AFTER indicator variable are also quite weak.

Table 4.

Descriptive statistics of acquirers and distressed targets, before the deal, in the period 1978 – 2015.

Sales revenue 940 0.943 0.798 0.673 0.010 3.379

Return on sales 920 0.204 0.116 0.321 0.011 1.654

AFTER 1008 0 0 0 0 0

Cash acquirer 941 0.125 0.079 0.138 0 1

Cash target 970 0.116 0.061 0.158 0 1

Industry target 1008 25.688 32 14.886 0 48

Notes: Sales revenue is measured as the sales divided by total assets, also called asset turnover. It has been winsorized at a 1% level. SGA expense divided by Sales revenue is winsorized at a 2.5% level. Return on sales is defined as EBIT divided by sales and is winsorized at a 2.5% level. Financial distress and AFTER are both dummy variables, whereby Financial distress indicates if a target firm is in financial distress or not and AFTER indicates the period before or after the deal. Cash acquirer and Cash target are both computed as cash divided by total assets. Total asset acquirer and Total assets target are both defined as the dollar amount and they are winsorized at a 5% level. Ln(Employ.number acquirer) and Ln(Employ.number acquirer) are both the natural logarithm of the employee numbers. Relative size target is defined as the total assets of the target divided by the total assets of the acquirer. Industry acquirer and Industry target are both dummy variables that indicate to which Fama-French industry group the acquirer and target belong.

Table 5.

Descriptive statistics of acquirers and distressed targets, after the deal, in the period 1979 – 2015.

Sales revenue 897 0.924 0.748 0.670 0.010 3.379

Return on sales 890 0.184 0.105 0.286 0.011 1.654

AFTER 1008 1 1 0 1 1

Cash acquirer 899 0.106 0.072 0.116 0 1

Cash target 201 0.111 0.053 0.159 0 1

Industry target 1008 25.688 32 14.886 0 48

(29)

Table 6.

Cross-correlations.

Notes: “Tot.” refers to Total. “acq.” refers to acquirer. “tar.” refers to target. “Rel.” refers to Relative. “Ln” refers to natural logarithm. “Emp.” and “Employ” refers to Employees. “nr.” refers to number. “ROS” refers to the variable Return on sales. and “Financ. distr.” refers to the variable Financial distress. ***, ** and * indicate the significance levels at 1%, 5% and 10% respectively.

Sales

revenue expense SGA ROS Financ. distr. AFTER acquirer. Cash target Cash Tot.assets acquirer Tot.assets target Rel size target Ln(Emp. nr. acq) Ln(Emp. nr. tar) Industry acquirer Industry target

Sales revenue _1.0000 SGA expense -0.2552*** 1.0000 ROS _{-0.3650*** 0.5896*** 1.0000} Financ.distr _{0.0639*** -0.0042 -0.0281*** 1.0000} AFTER _{-0.0358*** -0.0347*** -0.0186*** 0.0000} _1.0000 Cash acquirer _0.0056 _{0.4959*** 0.2567*** -0.0137*** -0.0913*** 1.0000} Cash target _0.0110 _{0.3811*** 0.1588*** -0.0591*** -0.0597*** 0.4962*** 1.0000}

Tot. asset acq. -0.2394*** -0.1529*** -0.0362*** -0.0303*** 0.0756*** -0.1720*** -0.1039*** 1.0000

Tot. asset tar. -0.1820*** -0.1439*** -0.0374*** -0.0433*** 0.1550*** -0.1586*** -0.1622*** 0.6008*** 1.0000

Rel. size target _0.0014 _{0.0098 0.0407*** 0.0042} _{0.0130* 0.0575*** -0.0145** -0.0504*** 0.0919***} _1.0000

ln(Emp. nr. acq.) 0.1890*** -0.3715*** -0.3938*** 0.0236*** 0.0646*** -0.2544*** -0.1142*** 0.4654*** 0.3358*** -0.0918*** 1.0000

ln(Emp nr. tar.) 0.2020*** -0.2963*** -0.2945*** 0.0496*** -0.0012 -0.1860*** -0.2794*** 0.2667*** 0.4333*** 0.0642*** 0.6172*** 1.0000

Industry acq. -0.1906*** -0.0525*** 0.0775*** -0.0454*** 0.0000 -0.1012*** -0.1214*** 0.1707*** 0.1251*** -0.0200*** -0.1201*** -0.0781*** 1.0000

The post-operating performance of acquirers who take over financially distressed firms, measured by synergies and the return on sales