• No results found

Regression results show strong evidence of a private firm discount on the Dutch acquisition market and consequently support the universality of the PFD

N/A
N/A
Protected

Academic year: 2021

Share "Regression results show strong evidence of a private firm discount on the Dutch acquisition market and consequently support the universality of the PFD"

Copied!
43
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

The Impact of the Private Firm Discount Stock Price Performance: Theory and Evidence From Dutch Acquisition Targets (2001-2007)

Groningen, October 2009

Author: Thesis Supervisor: Methodology Instructor:

V.L.Benjamins dr. G. de Jong dr. P. Rao Sahib

S1322087 Faculty of Economics Faculty of Economics

Faculty of Economics Landleven 5, 9747AD Landleven 5, 9747AD Westergracht 99r, 2013ZN Groningen, Groningen,

Haarlem , The Netherlands The Netherlands The Netherlands Haarlem , The Netherlands The Netherlands The Netherlands V.L.Benjamins@student.rug.nl g.de.jong@rug.nl p.rao.sahib@rug.nl

(2)

Abstract

This study provides a conceptual framework and an empirical study to assess whether the universal. Differences between public and private targets such as illiquidity of shares, valuation problems and information risk, affect abnormal returns The sample contains 129 Dutch targets, bought by US, UK and Dutch acquirers, during the period 2001-2007. Regression results show strong evidence of a private firm discount on the Dutch acquisition market and consequently support the universality of the PFD. No evidence is found of payment effects, experience effects, or cross-border effects on the performance of private target acquisitions. However observations support the idea that (1) stock financing is used less for private targets, that (2) concentrated control results in cash payment preferences, and (3) foreign acquirers perform better in public acquisitions than in private acquisitions. Furthermore regression results show that UK acquirers perform significantly worse than national acquirers when buying a private target.

Keywords: Private firm discount, Target choice, Acquisition performance, Abnormal returns

V.L. Benjamins 1

(3)

TABLE OF CONTENTS

LIST OF TABLES 3

LIST OF ABBREVIATIONS 4

1. INTRODUCTION 5

2. THEORY AND HYPOTHESES 9

2.1 Mergers and Acquisitions .. ... 9

2.2 The Private Firm Discount (PFD) 11

2.3 Public and private firms in the Netherlands 17

2.4 Hypotheses 19

3. METHODS 22

3.1 Data Collection and Sample . 22

3.2 Discount measurement 23

3.3 Model Specification 24

3.3.1 The dependent variable CAR 24

3.3.2 The independent variables 25

3.3.3 The control variables 26

3.3.4 Model . 27

4. EMPIRICAL RESULTS 28

4.1 Sample Characteristics and Descriptive Statistics 28

4.2 Regression Results 30

5. CONCLUSIONS 34

REFERENCES 36

V.L. Benjamins 2

REFERENCES 36

APPENDIX ... 38

(4)

LIST OF TABLES

Table 1. Strategic motives of M&As 10

Table 2. M&A dimensions in performance literature 11

Table 3. Differences between public and private targets 12

Table 4. Empirical studies of 16

Table 5. Characteristics of BVs and NVs 18

Table 6. Conceptual model 19

Table 7. Descriptive statistics & Pearson correlations 28

Table 8. Regression results dependent variable CAR 31

Table 9. Robustness checks for industry, country, and year effects 33

APPENDIX A: Deal and firm size 38

APPENDIX B: The cross-border variable 39

APPENDIX C: The year variable 40

APPENDIX D: The industry variable 41

APPENDIX E: The country variable 42

V.L. Benjamins 3

(5)

LIST OF ABBREVIATIONS

AEX Amsterdam Exchange Index

AR Abnormal Return

BV Besloten Vennootschap (Dutch private limited liability company) BV Besloten Vennootschap (Dutch private limited liability company) CAR Cumulative Abnormal Return

CBS Centraal Bureau voor de Statistieken (Central Bureau for Statistics) DCF Discounted Cash Flow

DJIA Dow Jones Industrial Average DLM Discount for Lack of Marketability

EU European Union

EU European Union

FTC Federal Trade Commission LSE London Stock Exchange M&A Mergers and Acquisitions

Nace Nomenclature statistique des activités économiques dans la Communauté européenne (Statistical Classification of Economic Activities in the European Community)

NL The Netherlands

NL The Netherlands

NV Naamloze Vennootschap (Dutch public limited liability company) NYSE New York Stock Exchange

OLS Ordinary Least Square PFD Private Firm Discount

US United States

UK United Kingdom

V.L. Benjamins 4

UK United Kingdom

(6)

1. INTRODUCTION

Every year acquirers buy ten thousands companies globally. The total value amounts thousands of billion dollars, exceeding the GDP of several large countries (Cartwright &

Schoenberg, 2006). Academic research has given considerable attention to Merger &

Acquisition (M&A) activity. M&A literature studies acquisition waves, industry concentration, acquisition performance and more. M&A literature on performance can be distinguished in short-term (< 12 months) and long-term (12 months or more) performance.

Measurements on long-term performance are questionable, as it is difficult to isolate an acquisition event in a large stock price event window, whereas short-term abnormal returns are more reliable measurements (Kothari & Warner, 2004). M&A literature can also be distinguished in acquirer and target performance. The stock price performance for targets show positive results in literature, while no consensus is reached on the short-term performance for acquirers (Cartwright & Schoenberg, 2006). Many researchers have investigated the returns to acquirers, however with mixed findings (Kräusll & Topper, 2007;

Doeswijk & Floor, 2007). This study will contribute to the literature on M&A performance and the discussion on short-term abnormal returns to

Dutch acquisition targets.

Acquisition transactions involve billions of dollars, both in the purchase of huge companies and in the acquisition of ten thousands of smaller companies. On the US M&A , McConnell, and Stolin, 2006; Koeplin, Sharin, and Shapiro, 2000; Kooli , McConnell, and Stolin, 2006; Koeplin, Sharin, and Shapiro, 2000; Kooli , 2003). Acquirers have the choice to buy either a publicly traded or a privately traded target.

Private and public targets differ from each other in several aspects, like share transferability and disclosure requirements, resulting in a discount when buying a private target on the US M&A market. The universality of the PFD is unknown.

As matters of fact corporate characteristics on continental Europe, the Western European mainland, differ from the US. This makes US studies less universal. The US market for corporate control is more active than other markets (Capron & Shen, 2007).

Furthermore, the US is characterized by dispersed ownership and control (Barca & Becht,

V.L. Benjamins 5

Furthermore, the US is characterized by dispersed ownership and control (Barca & Becht, 2001), whereas companies on continental Europe have a high level of ownership concentration. According to Barca & Becht (2001), half of the non-financial companies listed on a stock exchange, have a single blockholder controlling 20-57% of the votes. The

(7)

ownership structure of Dutch firms is relatively concentrated (De Jong, Van der Poel, and Wolfswinkel, 2007). In half of the Dutch non-financial companies listed on the AEX, single blockholder control 43.5% of the votes. On the other hand, half of the non-financial US listed companies at the NYSE have single blockholders controlling only 5.4% of the votes (Barca

& Becht, 2001). Studies on corporate control suggest that the corporate characteristics on continental Europe may dominate around the world. This perspective increases the interest to study corporate assumptions like the PFD in the Netherlands, given the typical continental European characteristics of the Dutch market.

The Dutch market is characterized by liability firms and limited liability firms, determining their financial risk and type of tax to pay. The liability firms are sole traders, general partnerships, and limited partnerships. These legal forms pay income tax over their profits, and their personal assets are liable for obligations. The limited liability firms are the private limited liability companies (BV, Besloten Vennootschap) and the public limited liability companies (NV, Naamloze Vennootschap). Their shareholders are only liable for the amount that they have invested1. In 2008, the Netherlands had 580.700 firms registered at the

100.000 partnerships, and more than 160.000 limited liability companies. This thesis will analyze the limited liability firms, and test whether acquirers create more shareholder value when buying a BV compared to buying a NV.

The Dutch acquisition market is likely to differ from the US acquisition market and is therefore relevant to test whether the PFD is universal. This thesis analyzes the performance of acquirers buying the Dutch limited liability companies (BVs) during 2001-2007.

of acquirers buying the Dutch limited liability companies (BVs) during 2001-2007.

Acquisition activity in the Netherlands has flourished since the 1990s; therefore the period 2001-2007 represents a more mature acquisition market. This thesis includes acquirers from the UK, the US and the Netherlands, which are the most present acquirers on the Dutch

acquisition market. of shareholders

analyzed around acquisition announcements, similar to studies of Capron & Shen (2007), Faccio et al. (2006), and Fuller, Netter, and Stegemoller (2002). A replication of their studies applied to the Dutch case will show the target choice effect (a public or private target) on It makes it possible to test whether the PFD is a universal

V.L. Benjamins 6

It makes it possible to test whether the PFD is a universal phenomenon in terms of abnormal returns to shareholders. Furthermore, it offers the

1 http://www.kvk.nl/english/startingabusiness/030_Paperwork/legalforms/default.asp,

(8)

opportunity to test whether continental Europe reacts favourably to private target acquisitions. The Research Question of this thesis is:

- Do acquirers of Dutch private targets perform better than acquirers of Dutch public targets?

This question has various managerial implications: U

for any M&A practitioner. Every acquirer runs the risk of paying too much for its target, and every seller may accept a too low price due to, for instance, a lack of knowledge. Managers only want to buy targets that create shareholder value and therefore buy or sell only when the deal creates positive value for them (Cartwright & Schoenberg, 2006). To test the success of an acquisition, this thesis analyzes the created shareholder values. If around the acquisition announcement the stock price of the acquirer increases abnormally, it can be assumed that shareholder value is positive and the target is bought with a discount. If the stock price decreases, the shareholder value is negative and the acquisition is bought with a premium.

Empirical studies on M&A are stock market event studies, large-scale accounting data studies, and case studies. Most evidence on M&A performance comes from the finance literature, where stock market event studies focus on the M&A effects for stockholders of acquirers or targets. Among other things, event studies assume that financial markets are effi

examine the accounting data for firms before and after an acquisition to determine the changes associated with the merger. Financial measures such as profit margins and cash flow changes associated with the merger. Financial measures such as profit margins and cash flow - intensive performance approaches to study individual mergers (Pautler, 2001). This thesis is an empirical study using the stock market event approach. Stock market fluctuations are subtracted from stock price fluctuations

This thesis contributes to the literature on private firms and supports the private firm discount literature. Regression results show strong evidence of a private firm discount on the Dutch acquisition market and consequently support the universality of the PFD. No evidence is found of payment effects, experience effects, or cross-border effects on the performance of

V.L. Benjamins 7

is found of payment effects, experience effects, or cross-border effects on the performance of private target acquisitions. However observations support the idea that (1) stock financing is used less for private targets, that (2) concentrated control results in cash payment preferences, and (3) foreign acquirers perform better in public acquisitions than in private acquisitions.

(9)

Furthermore regression results show that UK acquirers perform significantly worse than national acquirers when buying a private target.

This thesis applies the following key concepts, in merger and acquisition literature an acquisition describes the transfer of ownership, whereas a merger is a legal procedure of combining one company with another. In this thesis, I refer to the acquired corporation as the target and the buyer as the acquirer. A public company is one that sells securities (stocks or bonds) to the general public. Acquirers can easily valuate public companies as their security price on the stock exchange is public. Public companies are subject to a relatively high number of laws compared to privately held companies, because of their impact on society (Cartwright & Schoenberg, 2006). Public companies are valued daily by stock price movements; private companies lack this market feedback. The value of a private company is therefore difficult to compute and justify.

The remainder of this thesis is structured as follows. Section two offers a literature review on M&A key concepts, and discussion of empirical studies on the PFD. Furthermore Dutch private and public firms are compared, and hypotheses on the Dutch private firm discount are developed. Section three will discuss the empirical method of this thesis. First it discusses the criteria used for the data collection, then it expands on the measure used, and finally it specifies the model used. In section four, the empirical results are summarized. First of all, the descriptive statistics and correlations of the variables are discussed, then the regression results are analyzed, and finally the robustness of the results are checked for industry, country, and year effects. The final section reviews the implications of the regression outcomes, as well as the contributions to academic literature.

regression outcomes, as well as the contributions to academic literature.

V.L. Benjamins 8

(10)

2. THEORY & HYPOTHESES

The literature review will discuss studies on M&A, the PFD, and corporate characteristics in the Netherlands. Furthermore, hypotheses will be developed on a Dutch private firm discount.

2.1 Mergers & Acquisitions (M&A)

M&A motives. Acquirers buying a target consider the deal to be a profitable investment. The acquisition is an alternative form of investment to enhance capacity, to obtain new knowledge or skills, to enter new product or geographic areas, or to reallocate assets. However, acquisitions are not just a business investment; they are part of a broader strategic plan (Pautler, 2001). The most popular strategic classification in M&A is the Federal Trade Commission (FTC) classification of mergers. The FTC classification identifies five acquisition types; namely, horizontal integration, vertical integration, product-extension, market-extension, and the conglomerate acquisition. These acquisition types can be grouped by motivation. When planning an acquisition, acquirers evaluate their target along their assets. M&A activity is motivated by synergy, the so called two-plus-two-equals five effect (Cartwright & Schoenberg, 2006). Synergy can be grouped in three types, namely operational, financial, and diversifying synergy. When synergy occurs, companies give each other some previously missing element. Elements can be physical assets like buildings or land, financial assets like debt-equity ratio or interest coverage, and intellectual assets like land, financial assets like debt-equity ratio or interest coverage, and intellectual assets like patents or brands.

Table 1 shows acquisition and motivation types. Horizontal integration occurs when expanding in businesses that are familiar. Vertical integration occurs when companies buy a supplier (vertical backward integration) or a customer (vertical forward integration) to achieve economies. Both the horizontal and the vertical integration are related acquisitions.

Acquirers achieve financial efficiencies by exerting market power, primarily reached through extended size form horizontal or vertical integration. Product-extension occurs when the acquirer and the target have functional relatedness in production or distribution. Market-

V.L. Benjamins 9

acquirer and the target have functional relatedness in production or distribution. Market- extension occurs when acquirer and target manufacture the same products, but sell them in different geographic markets. Both product-extension and market-extension are related acquisitions; acquirers achieve operational efficiencies by using resources more efficiently.

(11)

The same input amount produces a higher quantity of outputs, lowering the average cost of a product. Conglomerate acquisitions exist when the acquirer buys a non-related target. The acquirer searches for diversification economies, by acquiring targets with dissimilar earning streams. Earnings diversification lowers the variation in profitability and reduces the risk of bankruptcy. Conglomerate acquisitions diversify economies by lowering risk relatively to performance.

Table 1. Strategic motives of M&As

Synergy type Tool Acquisition type Relatedness

1. Financial efficiencies Market Power Horizontal integration Vertical integration

Related Related

2. Operational efficiencies Lower Average Cost Product-extension Market-extension

Related Related

3. Diversification economies Lower Risk Conglomerate Diversifying

Besides the above mentioned literature on synergy, alternative M&A motives exist (Pautlin, 2001). This literature indicates that managerial incentives may drive acquisitions. Managers may overdiversify, overemphasize growth, or just make bad acquisition decisions.

Performance. The acquisition failure rate in terms of performance is high. Acquisition managers report that only half of their acquisitions can be considered successful against the original objectives set (Cartwright & Schoenberg, 2006). Acquisition targets generally enjoy positive short-term results, however acquirers frequently experience share price underperformance (Kräusll & Topper, 2007; Doeswijk & Floor, 2007). Research attention exists on a broad range of managerial disciplines like financial, strategic, behavioural, operational and cross-cultural aspects. They try to explain the performance variance between individual acquisitions and overall acquisition underperformance.

The dominating research areas of M&A performance are strategic fit, organizational fit, and the acquisition process literature. These M&A dimensions contribute to the question why most acquirers underperform after an acquisition (see Table 2). M&A studies on strategic fit

V.L. Benjamins 10

most acquirers underperform after an acquisition (see Table 2). M&A studies on strategic fit state that acquisitions create value when resources can easily be exchanged, and knowledge transferred. These strategic attributes lead to higher acquisition performance. M&A literature on organizational fit state that acquisitions create value when there is cultural compatibility

(12)

between the acquirer and the target. Cultural fit leads to higher acquisition performance.

M&A literature on the acquisition process states that organizational learning is the value creating factor of an acquisition. Acquisition performance is dependent on the success of the integration strategy.

Table 2. M&A dimensions in performance literature

M&A Dimensions Tool Factors of value creation 1. Strategic fit Strategic attributes Resource sharing

Knowledge transfer 2. Organizational fit Cultural fit Cultural compatibility 3. Acquisition process Integration strategy Organizational learning

Target choice. Most existing M&A studies have focused on public target acquisitions bought by public acquirers (Capron & Shen, 2007). Literature analyzes target choice between related and unrelated targets. Related targets are classified along similar industries, where acquirers and targets share perceptions about strategic commonalities. In history, M&A waves were characterized by target choices. The 1960s were characterized by conglomerate acquisitions, where acquisition typically involved non-related industry targets. The 1980s were characterized by hostile acquisitions, where acquirers were mostly financiers, so non-related targets. The late 1990s, was the time of related M&A deals, acquisitions were typically made within industries both horizontal as vertical (Shleifer and Vishny, 2003). Since the last within industries both horizontal as vertical (Shleifer and Vishny, 2003). Since the last decade, target choice has been analyzed from a different perspective: Does target choice in terms of private and public targets affect acquisition performance?

2.2 The Private Firm Discount (PFD)

In literature strategic studies explain target choices and finance studies explain performance, this study combines both. How does target choice affect acquirer performance? Public and private targets

V.L. Benjamins 11

private targets

companies exhibit three important distinctive characteristics (Kraakman & Hansmann, 2004).

Private targets tend to be smaller, are less often listed on a stock exchange, and have less regulatory disclosure requirements. The last two characteristics imply higher risk to private

(13)

target acquirers. Unlisted targets are illiquid in share transferability and difficult to valuate.

They do not have observable share prices indicating their market value. Furthermore private targets have fewer disclosure regulations, resulting in a-symmetric information between acquirer and target. PFD literature argues that a discount is applied for privately traded targets to compensate for these risks.

Table 3. Differences between Public and Private targets

Public targets Private targets Risk Private Targets 1. Size Larger on average Smaller on average -

2. Listing Stock Exchange Often listed Seldom listed Share illiquidity Valuation problem

3. Regulatory Disclosure Many requirements Few requirements Information risk

Share illiquidity. Koeplin et al. (2000) state that the illiquidity of shares is the main difference between public and private targets. Shareholders of public targets have a stock exchange where they can sell their shares. Shareholders of private targets lack that facility, since private targets are seldom listed on a stock exchange. This makes it difficult for investors to sell their shares within a given time period. Investors lack the opportunity to rebalance their portfolio and allocate capital to alternative marketable assets (Kooli et al., 2003). Public targets have liquid shares, their marketability involve lower costs. The PFD reflects the reduced value due to the lack of marketability of private target et al., 2003).

Valuation problem. Private target valuation is challenging for finance professionals, as information is few available both in history and depth (Kooli et al., 2003). Unlike public targets, private targets have no observable stock price to serve as an objective measure for market value. For public targets the corporate control market serves as an asset valuation mechanism for potential acquirers (Capron & Shen, 2007). On the other hand, the valuation

V.L. Benjamins 12

methods for other private targets is unknown, making comparisons between firm values difficult (Koeplin et al, 2000). An alternative valuation approach is the discounted cash flow (DCF) method. The DCF method depends on cash flow and the use of appropriate risk

(14)

measures. However, by using the same discount rate for the risk of private targets as well as for public targets, one neglects the different firm risk levels.

Information risk. Public targets listed on the stock exchange have disclosure requirements, and therefore provide corporate disclosure to investors through regulated financial reports.

Investors are more targets have

high levels of corporate disclosure (Healy & Palepu, 2001). Acquiring private targets involve information asymmetry, since they have more freedom in what information they reveal. This increases the risk of not properly evaluating private target assets (Capron & Shen, 2007).

Investors value private targets lower and ask an acquisition discount for bearing the targets.

Several authors have empirically tested the private firm discount. Five reference studies are selected for PFD analysis (see Table 4); they support the existence of some sort of PFD.

However, the empirical results show great diversification in discount percentages. The discount ranges from 1.48% to 50%. This diversity in discount is most likely due to differences in samples, models and methodologies used. Koeplin et al. (2000) and Kooli et al. (2003) use financial multiples for their reference portfolio method and find discount rates between 17-50%. Faccio et al. (2006) and Capron & Shen (2007) use the abnormal return method areholders as measurement and find discount rates between 1.48-4.7%. All methodologies use the OLS method, analyze the nineties as time period, and four studies used the US as target market. The results from Faccio et al. (2006) period, and four studies used the US as target market. The results from Faccio et al. (2006) indicate that the PFD may not be a US phenomenon solely, encouraging further study of the European market.

A. Capron & Shen (2007) investigate acquirer returns and find that target choice (private/public) is not universal but depends and on the merging

between public and private targets? (2) Do private target acquisitions elicit a more positive stock market reaction than public target acquisitions, which on average, destroy value for

V.L. Benjamins 13

stock market reaction than public target acquisitions, which on average, destroy value for

target fits the theory? The sample contains 92 observations of public acquirers, including 52 public targets and 40 private targets. This small sample size is the weakness of their study.

(15)

Nevertheless their conclusions on the research questions are: (1) acquirers favour private targets in familiar industries, and turn to public targets to enter new business domains or industries with a high level of intangible assets; (2) acquirers of private targets perform better than acquirers of public targets on merger announcement, after controlling for endogeneity bias; (3) acquirers of private targets perform better than if they had bought a public target, and acquirers of public targets perform better than if they had bought a private target. To test these research questions, Capron & Shen use a methodology combining an acquirer return event study and an acquirer post-acquisition survey. The AR is calculated using independent variable for relatedness, acquisition experience, target scope, target intangibles, and target age. Furthermore, they controlled for a large number of variables: number of competing bidders, target size, target pre-merger profitability, target industry growth, US target, acquirer scope, acquirer pre-merger profitability, and US acquirer. Capron and Shen use a dummy for US target and US acquirer to point out the deviation of US markets for corporate control, compared to other countries around the world. The AR is tested for the event window [-20, +10], so calculated from 20 days before the acquisition announcement till 10 days after the announcement. Capron & Shen use shorter time windows, [-20, +5] and [-20, +1], to test for robustness. A shorter time window is better able to isolate the acquisition effects from potential other confounding effects. They found a PFD between 2.78% and 4.70%, dependent on the event window used.

B. Faccio et al. (2006) re-examine the listing-related effect of targets on a stock exchange -related effect, with insignificant -related effect, with insignificant negative returns for public targets and significant positive returns for private targets.

However the fundamental factors creating this effect are elusive. Their sample consists of 4,429 acquisitions by continental European firms over the period 1996-2001. Faccio et al.

calculate a 5-day announcement period CAR, by

the event window [-2, +2] to get the 5-day CAR. The dependent variable CAR is calculated using the following independent variables: UK acquirer, acquirer size, payment, and acquirer pre-announcement CAR. Furthermore, control variables are included to correct for AR

V.L. Benjamins 14

pre-announcement CAR. Furthermore, control variables are included to correct for AR size, hostile bids, and a dummy for announcement day. After running several regressions, Faccio et al. conclude that private

(16)

acquisitions of public targets do (-0.38%). Moreover they conclude that the listing-related effect is not due to a blockholder-creation effect or the effect explained by the agency theory. Finally, bigger

C. Fuller et al. (2002) study shareholder returns for acquirers that buy five or more public, private, or subsidiary targets within a short time period. They reason that since the same bidder chooses different target types and payment methods, any variation in returns must be due to target or acquirer characteristics. Fuller et al. collected 2,060 completed private target acquisitions over the period 1990-2000. The results of Fuller et al. indicate that the market views bids for private targets differently from bids for public targets

gain when buying a private targets and lose when purchasing a public targets. The AR is greater (1) when the target is larger and (2) when the acquirer offers stock. The CAR is significantly negative (-1%) for public targets and significantly positive (2.08%) for private targets.

D. Koeplin et al. (2000) indicate an objectivity problem when valuing private targets, as no observable stock price is available. They tested how large the PFD is, as percentage of similar public targets. They observe 192 acquisitions during 1984-1998; 84 transactions in the US and 108 transactions outside the US. Koeplin et al. identify a set of private target acquisitions.

For each private target transaction, a public target is matched according to industry, time, and size. They use financial parameters from sales and earnings to calculate the purchase price size. They use financial parameters from sales and earnings to calculate the purchase price multiples. Koeplin et al. state that this approach is a standard technique used by investment bankers in valuing acquisition candidates. Koeplin et al. include explanatory variables in their sample to account for differences in size and historical growth rates, between the private and public companies. Their results indicate a significant PFD. Private companies sell at a discount relatively to comparable public companies. The earning multiples show a 20-30%

discount when buying a private target. The discounts are even larger for foreign companies;

acquirers of non-US private targets show a 40-50% discount.

V.L. Benjamins 15

E. Kooli et al. (2003) examine how large the discount is for the lack of marketability (DLM) attached to private target valuations. They use a US sample containing 331 private

(17)

Table 4. Empirical studies of

A. Capron &

Shen, 2007

B. Faccio et al, 2006

C. Fuller et al., 2002

D. Koeplin et al., 2000

E. Kooli et al., 2003

Research Question

(1) What drives target choice? (2) Do private targets outperform public

Do listed targets create greater shareholder wealth to

How is acquirer shareholder return influenced by

How large is the PFD as a percentage of similar public

How large is the discount for lack of

marketability outperform public

targets? (3) Do acquirers gain when following their theory?

wealth to acquirers?

by

characteristics of the target and the bid?

similar public firms?

marketability (DLM) attached to private firm valuations?

Results (1) Private targets in familiar industries (2) Private targets perform better (3) AR dependent on search type.

Private targets create greater value, although not due to a blockholder creation effect or explained by the agency theory.

Positive AR when buying a private firm and negative for public firms.

The PFD is 20- 30% for US targets and 40- 50% for non- US targets.

The PFD is dependent on the growth and the size of private targets.

Sample Sample Observations Deal value Countries Industries

Time Period 92

-

US

Manufacturing

1988-92

4,429

> 5 mill $

W-Europe

All

1996-2001

2,060

> 1 mill $

US

All, except utility/ financial

1990-2000

192

-

US

All

1984-98

331

-

US

All

1995-2002

Methodology OLS OLS OLS OLS OLS

Dependent Abnormal returns Abnormal returns Abnormal Financial Financial Dependent

Variable

Abnormal returns Abnormal returns Abnormal returns

Financial multiple

Financial multiple

Independent Variables

- ownership - relatedness - experience (a) - scope (t) - intangibles (t) - age (t) - fit

- UK acquirer - size (a) - pre CAR (a) - country - payment

- ownership - ownership - ownership

Control Variables

- bidders - size - payment - pre-merger

-

- relatedness - size (t) - hostile bid

- payment - industries relatedness - rel. size

- industry - year - size

- industry - year - size

V.L. Benjamins 16

* (a) = acquirer, (t) = target profitability (a/t) - industry growth - U.S. (a/t) - scope (a)

- announcement day

(18)

acquisitions over the period 1995-2002. Kooli et al. support the existence of a PFD, and find a discount of 34%. From cross-sectional analysis however, it shows that the discount observed varies among target characteristics and industries. The PFD depends on the growth and size of private targets. Low-growth targets are more discounted than high-growth targets, and small targets are more discounted than large targets. The reference portfolio methodology that they use was constructed as follows. Kooli et al use matching procedures to control for the differences in characteristics between private and public targets. The private discount is estimated using financial multiples, such as sales, earnings, and cash flow multiples. The procedure assumes that the universe of public deals is more rationally priced than the ones of private targets and therefore can be used as benchmark. Thus, for each private target, a control portfolio of public companies is constructed. Kooli et al. rank the private and public targets by year, industry and size, to create a match between private and public targets.

Methods used. The papers from Koeplin et al. (2000) and Kooli et al. (2003) use a reference portfolio method, also called the accounting method, where direct performance measures capture the extent of PFD. Comparable samples of private and public targets are matched, based on size and industry. These studies measure the private target discount by comparing them to publicly sold targets, however horizontal comparisons between private targets themselves are not possible. Also the match of public targets bought in time, size and industry is not sufficient to control for all factors affecting the discount rate.

The papers from Capron & Shen (2007), Faccio et al. (2006), and Fuller et al. (2002) follow the abnormal return method, also called the stock market method. The stock market follow the abnormal return method, also called the stock market method. The stock market approach has methodological concerns when calculating long-horizon performance. It is difficult to isolate the acquisition announcement effect in a large event window. However, short-horizon studies measure performance around the announcement date of an acquisition and are more reliable. The short-horizon AR method can be con

Therefore this thesis analyzes abnormal

2.3 Public and private firms in the Netherlands

V.L. Benjamins 17

2.3 Public and private firms in the Netherlands

Dutch corporate law differentiates between different business associations such as partnerships (e.g. law firms), trusts (e.g. pension funds), companies limited by guarantee (e.g.

(19)

universities), and companies limited by shares (privately or publicly traded). Shareholders trade public targets on a regulated stock exchange, whereas private targets are more restricted on share transferability (Kraakman & Hansmann, 2004). Both public and private targets are characterized by separate legal personality. They have the right to sue and be sued. The shareholders have limited liability meaning that when the company is insolvent, shareholders financial liability is limited to their investment value in the companies.

The Dutch public and private companies have similar characteristics as for instance companies from the UK. However some differences exist, such as the difference in capital requirements. The standard authorised share capital for Public Limited Companies (Plc.) amounts £50,000 or , whereas for Dutch public companies (NV) it is lower, 45,000. In case of Private Limited Companies (Ltd.) there is no minimum share capital, whereas the requirement for a Dutch private company (BV) is 2 Besides the capital requirements, there are no odd characteristics of BVs compared to the reference studies. The limited liability companies in the Netherlands are in general the larger NVs (Naamloze Vennootschap), and the smaller BVs (Besloten Vennootschap). NVs have freedom in share transferability, whereas BVs have restrictions (Wooldridge, 2008). Another difference between NVs and BVs, is that NVs are obliged by law to publish their annual accounts and BVs not. Furthermore, BVs are less frequently listed on a stock exchange (De Jong, Kabir, Marra, and A. Röell, 2000). The Dutch NVs and BVs have the same differences in characteristics as the public and private firms discussed in the reference literature. Table 5 summarizes that BVs have illiquid shares compared to NVs, BVs do not have an observable stock price reflecting their market value like NVs do, and BVs are less transparent than NVs stock price reflecting their market value like NVs do, and BVs are less transparent than NVs because they have less disclosure requirements.

Table 5. Characteristics of BVs and NVs

Naamloze Vennootschap (NV) Besloten Vennootschap (BVs) Limited Liability

Legal form

Yes

Public Company

Yes

Private Company

Size Larger on Average Smaller on Average

V.L. Benjamins 18

2 http://www.companieshouse.gov.uk

Listed Often Listed Seldom Listed

Regulatory Disclosure Many Requirements Few Requirements

(20)

2.4 Hypotheses

This thesis will analyzes whether the PFD-phenomenon can be identified on the Dutch acquisition market, by testing whether there is a significant difference in CAR between public and private targets upon acquisition announcement. The Dutch market is characterized by BVs (private firms) and NVs (public firms). BVs are riskier targets than NVs, because of their share illiquidity, unclear market value, and information asymmetry. Therefore BVs are likely to sell at a discount (see Table 6, first hypothesis). The second hypothesis makes propositions about deal and acquirer characteristics, explicitly the payment method, acquisition experience, and cross-border transaction. These characteristics have well-known effects on CAR to shareholders when buying a public target. This thesis introduces additional implications when buying private targets. Because of the increased risk involved with BVs, it is likely that (1) acquirers earn higher CAR when buying private targets with stock than with cash, (2) acquisition experienced buyers earn higher CAR when buying private targets than less acquisition experienced buyers do, and (3) national acquirers earn higher CAR when buying a private target than cross-border acquirers do.

Table 6. Conceptual model

Dutch Private Targets Hypothesis 1 Hypothesis 2 BVs are sold at a

discount BVs are riskier then NVs due to:

1) Share illiquidity 2) Market valuation

Acquirers earn higher CAR when buying a private target when:

a) they pay with stock 2) Market valuation

3) Information asymmetry

a) they pay with stock b) they are more experienced c) they are national acquirers

PFD. US studies (Koeplin et al., 2000; Kooli et al., 2003; Fuller et al., 2002) show that private targets are sold at a discount to compensate for differences in characteristics between private and public targets. BVs are such private targets with illiquid shares, unclear market value, and information asymmetry. Whereas NVs listed on a stock exchange, signal the likelihood of surviving in the long run. The PFD is likely to exist on the Dutch acquisition

V.L. Benjamins 19

market, as the characteristics of Dutch BVs and NVs are similar to the private and public targets in the reference studies. Consequently, the first hypothesis expects the CAR of a target.

(21)

Hypothesis 1. olders is higher when buying BVs than buying NVs.

Acquisition characteristics such as the payment method, acquisition experience, and cross- own from the literature on public targets; however this thesis suggests that these characteristics have implications on private target acquisitions as well.

Payment method. A general consensus exists in literature on the payment method effect.

Public targets show a positive relation between cash payments and acquirer returns, and a negative relation between stock payments and acquirer returns. Shareholders profit from & Floor, 2007). Acquirers use cash payments when they assume public targets to be undervalued.

Therefore acquisitions financed with cash have higher returns upon announcement than acquisitions financed with equity (Kräussl & Topper, 2007).

When evaluating private targets, a different relation is expected between payment method and acquirer returns.

asset. They can use lower premium bids (i.e. buying at a discount) or use stock payments to cope with the problem (Capron & Shen, 2007). Coff (1999) state that the risk of overpaying targets with unclear market values can be reduced by using stock payments. Cash has a fixed value that does not depend on the outcome, whereas stock payments are more flexible.

value that does not depend on the outcome, whereas stock payments are more flexible.

Therefore pure cash payments are used when there is great confidence. Whereas a bid made with stock reveals that the bidder views their stock as overvalued (Fuller et al, 2002).

Acquirers may use their overvalued stock to buy less attractive projects. Faccio et al. (2006) and Fuller et al. (2002) support this reversed payment method effect for private targets.

Acquirers earn higher AR when buying public targets in cash, and when buying private targets in stock. These findings may indicate that acquirers use their overvalued stock to buy a relatively risky private target compared to a public target. Therefore this thesis expects that

shareholders is higher when the private target is paid in stock.

V.L. Benjamins 20

shareholders is higher when the private target is paid in stock.

Hypothesis 2a.

stock than when BVs are paid in cash.

(22)

Acquisition experience. Buyers follow a learning curve when making acquisitions.

Acquisition experienced buyers have improved their skills in screening potential targets with refined selection criteria, and price more accurately new targets based on prior experience (Capron & Shen, 2007; Fuller et al., 2002). Acquisition experienced buyers are better able to evaluate market value in comparison to less acquisition experienced buyers.

Therefore it can be expected that more acquisition experienced buyers earn higher CAR to ying private targets. In contrast, acquirers lacking acquisition

experience are likely to have when buying private

targets.

Hypothesis 2b. Acquisition experienced buyers of BVs earn higher CAR than inexperienced buyers do.

Cross-border deal. Lower AR are expected with cross-border transactions, therefore investors tend to be home biased (Faccio et al., 2006; Kräussl & Topper, 2007). Cross-border deals create challenges in due diligence and the integration of two firms (Capron & Shen, 2007). Acquirers involved with a cross-border acquisition have lower ability to price their targets accurately since foreign acquirers have less knowledge about a market then domestic acquirers do on private targets. The information asymmetry involved with the private target acquisition is higher for foreign acquirers, consequently lowering their performance when buying a BV. Therefore it can be expected that Dutch acquirers earn higher CAR to

areholders when buying a BV than foreign acquirers do.

areholders when buying a BV than foreign acquirers do.

Hypothesis 2c. National acquirers of BVs than cross-border acquirers do.

In order to test these hypotheses on the Dutch PFD a sample is constructed, a measurement ed, and a model is specified.

V.L. Benjamins 21

(23)

3. METHOD

The first part of this section will start with a data collection evaluation. The second part, introduces the measurement that is used to calculate the abnormal returns (AR) to acqui shareholders. The final part of the section specifies the model with a description of the variables used.

3.1 Data collection and sample

The acquisition data is obtained from the Zephyr M&A database, and meets the following selection criteria:

· The sample includes completed acquisitions of Dutch targets, and public acquirers from the UK, the US, or the Netherlands3.

· The announcements are made between January 1, 2001 and December 31, 20074.

· Some acquirers return several times in the sample, however, they are included only when acquiring different targets5 and not when buying a target in phases.

· A deal values at least 6.

· The sample excludes investment funds and private persons7.

With this procedure a sample of 131 acquisition deals is obtained.

3 The UK, US, and the Netherlands are the largest group of acquirers on the Dutch acquisition market available at the Zephyr database over the period 2001-2007.

4 This period represents a stable and mature period on the Dutch acquisition market. Acquisition activity in the Netherlands began to flourish since the 1990s.

5 The sample contains fourteen acquirers that buy two different targets divided among year, industry, and

V.L. Benjamins 22

5 The sample contains fourteen acquirers that buy two different targets divided among year, industry, and country of origin. Three Dutch acquirers are repeated acquirers of more then two targets. Ordina NV buys three targets in the sample, SNS Reaal Groep NV buys four targets, and Koninklijke KPN NV seven times.

6 From the reference studies it shows that this is the lower bound when analyzing private acquisitions.

7 Their characteristics may differ from other companies in the sample.

(24)

3.2 Discount measurement

In accounting and finance research, event studies apply to a variety of firm specific and economy wide events. Event examples include acquisitions, earning announcements, issues of new debt or equity, and announcements of macro-economic variables such as the trade deficit (MacKinlay, 1997). Event studies in the field of law and economics measure for instance the impact of a change in the regulatory environment on firm value. Acquisition events can be analyzed through the behaviour of stock returns (Capron & Shen, 2007; Faccio et al., 2006; Fuller et al., 2002). When using a random selection of stocks and a random selection of an event date to each stock, it can be assumed that all should have a normal or zero abnormal performance on average. Therefore stock prices reflect the effect of an event.

Kothari and Warner (2004) analyze the literature on event studies and conclude that long- horizon event studies are less reliable than short-horizon event studies. The impact of an event on security price performance is difficult to isolate in the long-term. Therefore, this thesis measures abnormal stock returns caused by an event over a short time period.

The normal return is the expected return without an event taking place. The abnormal return is the unexpected change of shareholders wealth associated with an event. For each

day within a given subtracted from the ,

giving the effect of an event. So the abnormal return (AR) to shareholders is given by subtracting the market return (rm) from the firm return (rf):

for market fluctuations. The Amsterdam Exchange Index (AEX) is used to calculate the normal returns for Dutch acquirers. For UK acquirers the FTSE100 London Stock Exchange (LSE) is used, and for US acquirers corrections are made using the Dow Jones Industrial Average (DJIA). These indices show stock price averages from the largest and most widely publicly traded companies . The firm return (rf) for day t = -1, the day before the announcement of the acquisition, is the percentage change in stock price from day t = -2 to day t= -1. Similar method is used to calculate both the firm as the market

V.L. Benjamins 23

from day t = -2 to day t= -1. Similar method is used to calculate both the firm as the market returns within the event window.

The CAR aggregates the subtraction of the daily stock market index returns (Rm) from

f) over a given event window. This thesis regresses over

(25)

different time windows, all beginning one day before the announcement of an acquisition to several days after the announcement. The abnormal returns (AR) aggregation over the event window results in the cumulative abnormal return (CAR). For event window [-1, +7] the formula is as follows:

The outcome of this formula shows how the financial market has valued the acquisition around announcement.

positive, this means that the acquirer has bought the target of a discount. In case of private targets, a positive CAR reflects the private firm discount. In case of a negative CAR, the target is bought against a premium.

3.3 Model specification 3.3 Model specification

This section specifies the model. The dependent varia

shareholders. The independent variables are target choice, payment method, acquisition experience, and cross-border transaction. Finally the model is controlled for several control variables, namely deal relatedness, the acquisition year, the acquisition industry, and the acquirer country.

3.3.1. The dependent variable CAR 3.3.1. The dependent variable CAR

T is the dependent variable of the regression model. The CAR includes only trading days. When the first day is Friday, the second day will be

Monday. Similar procedure is us country of origin when

no trade has taken place. The announcement date is obtained from the Zephyr database. The

f) are obtained using historical quote tools of financial investment communities as Yahoo Finance and Business Week Investing. The stock returns

are returns

V.L. Benjamins 24

are returns

(Rm) are found using the historical quotes of the Dutch website Euroinvestor. The AEX is used for Dutch acquirers, the DJIA is used for US acquirers, and the FTSE100 is used for UK acquirers.

(26)

3.3.2 The independent variables

The independent variables are selected from leading studies. The data for these variables are obtained from the Zephyr database.

Target choice. Abnormal returns are expected to be affected by target choice as shareholders value riskier targets lower (Koeplin et al., 2000; Kooli et al., 2003; Fuller et al., 2002). A binary variable is used for the target choice. When a public acquirer has bought a BV the variable values 1, and when a NV is bought the variable values 0. If the coefficient of the variable target choice is positive, this represents that the BV is related to higher CAR to

plicating a PFD on the Dutch acquisition market.

Payment method. Abnormal returns are expected to be affected by payment method as shareholders value stock payments higher in case of private targets (Capron & Shen, 2007;

Coff, 1999; Faccio et al., 2006; Fuller et al., 2002). Each deal has been categorized along their payment method. A deal paid for 100% in cash values 1, pure stock payments value 2, and 3 for mixed payments of cash and stock8.

Acquisition experience. Acquisition experience is expected to affect abnormal returns because of advanced valuation skills (Capron & Shen, 2007; Fuller et al., 2002). The acquisition experience is the sum of all completed acquisitions, that valued more than 1 million, in the five years prior to the announcement of the specific deal, and available at the million, in the five years prior to the announcement of the specific deal, and available at the Zephyr database. If the coefficient of the variable experience is positive, this indicates that more acquisition experienced buyers create higher CAR to their shareholders than less acquisition experienced buyers do9.

Cross-border transaction. Cross-border transactions are expected to affect abnormal returns since they involve greater challenges in target valuation (Faccio et al., 2006; Kräusll &

8 Additional dummy variables are introduced to allow further examination of the payment method, however the

V.L. Benjamins 25

8 Additional dummy variables are introduced to allow further examination of the payment method, however the little availability of pure stock payments prevented any significant outcome on payment method.

9 Additional variables were constructed for no experience, few experience (1-5 acquisitions), moderate experienced (6-10 acquisitions), and highly experienced (> 10 acquisitions) to further examine the role of acquisition experience. However, the outcomes did not differ from the original experience variable.

(27)

Topper, 2007). A binary variable is used to indicate whether the acquisition was a cross- border deal or not. The acquirers from the UK and the US value 1, the Dutch acquirers value 0. If the coefficient of the variable cross-border transaction is positive, this means that cross- border acquisitions create shareholders wealth to the acquirer.

3.3.3. The control variables

Additional factors can influence the relationship between target choice and abnormal returns.

Controls are made for the industry relatedness between target and acquirer, for the acquisition year, for the acquisition industry, and for l were coded as dummy variables to be able to examine for potential unusual events during a particular year or specific industry.

Deal Relatedness. To identify whether the core businesses of the acquirer and the target are related, the NACE industry code (the Statistical Classification of Economic Activities in the European Community) are analyzed at a two-digit level 10. A binary variable is used, that values 1 if target and acquirer are related by industry, and 0 if not. A positive coefficient of the variable relatedness, this shows that acquisitions made within industries create higher

siness.

Year variable. A year variable is included to control for unexpected events in time. If the coefficient of the variable is positive, this means that at the end of the time period CAR was coefficient of the variable is positive, this means that at the end of the time period CAR was higher than at the beginning of the period 2001-2007. Dummy variables are used to examine whether the years 2004, 2005, and 2006 have significant effect on the CAR. These years are more represented in the sample than other years.

10 Although the Zephyr database reports NACE industry codes at the four-digit level, it is common to analyze relatedness for a broader industry group. For example Koninklijke KPN is coded NACE 6420

V.L. Benjamins 26

relatedness for a broader industry group. For example Koninklijke KPN is coded NACE 6420 Telecommunications, but is analyzed in this thesis at the two-digit level NACE 64 Post and

telecommunications. Some firms are classified in several industries according to the NACE code classification, then the company activity descriptions of the Zephyr database are compared to value the acquirer and target as related or not.

(28)

Industry variable. Ten industries are constructed using the NACE industry code. The industries are oil & gas (0), manufacturing (1), building & machinery (2), food & beverage (3), pharmaceuticals (4), services & media (5), telecommunications (6), electricity & energy (7), real estate (8), and ICT services (9). Dummy variables are used to examine whether the building & machinery industry, the services & media industry, and the ICT services industry significantly affect the CAR, as these industries are most frequent present in the sample.

Country variable. The company country code allocation of the Zephyr database is followed to create country dummies, which proxy for unforeseen country characteristics. A dummy is used for acquirers of the UK and the US. If both equal zero the regression results refer to the CAR when bought by a Dutch acquirer.

3.3.4. Model

The final model to test the effect of target choice on the cumulative abnormal return to shareholders is as follows:

CAR = ß0 + ß1 (target choice) + ß2 (payment method) + ß3 (acquisition experience) + ß4

(cross-border transaction) + ß5 (relatedness) + ß6 (year) + ß7 (industry) + ß8 (country) +

V.L. Benjamins 27

(29)

4. EMPIRICAL RESULTS

4.1 Sample characteristics

This section begins with some general characteristics of the sample, followed by an analysis of correlations between the variables.

From the 131 deals in total, 98 of the 131 deals involves private target acquisitions. The sample contains mainly small deals with values between one

Appendix, Table A1). This sample characteristic is consistent with the structure of the Dutch corporate market. The twenty-five companies with the highest share turnover listed on Euronext Amsterdam are represented on the AEX. From these twenty-five public Dutch companies, the largest four amount for 50% of the AEX in terms of index weighting. These companies are Unilever (15%), Royal Dutch Shell (15%), KPN (10%), and ArcelorMittal (9%) (see Appendix, Table A2). The other twenty-one large public companies are already much smaller. A few large companies and many smaller companies characterize the Dutch market.

From the descriptive statistics table it shows that several variables in the sample correlate with each other (see Table 7). Relations are found between the cross-border variable and target choice, payment method, and the year variable. The industry variable is related to target choice and the year variable, and finally the country variable is related to the payment method, cross-border transactions, and the year of acquisition.

method, cross-border transactions, and the year of acquisition.

Table 7. Descriptive statistics and Pearson correlations

Variable Mean s.d. 1 2 3 4 5 6 7

1. Target choice 0.75 0.43

2. Payment 1.74 0.92 .04

3. Experience 5.58 6.86 .05 -.09

4. Cross-border 0.62 0.49 .18* .23** -.08 5. Relatedness 0.95 0.21 .13 -.02 .00 .06

6. Year 4.48 1.92 .05 -.03 .10 -.25** .04

7. Industry 5.16 2.78 -.21* .11 -.04 -.14 -.04 .21*

8. Country 2.08 0.83 -.16 -.23** .05 -.88*** .02 .20* .09

V.L. Benjamins 28

8. Country 2.08 0.83 -.16 -.23** .05 -.88*** .02 .20* .09

a n = 129.

b Statistically significant at *** p < .001 (two-tailed); ** p < .01; * p < .05.

Referenties

GERELATEERDE DOCUMENTEN

- Does general pessimism induced by Dutch financial news media (content), reflecting investor sentiment, have a negative effect on the Dutch stock market index AEX.. The

∆EQUITY stands for the change in outstanding equity, ∆DEBT stands for the change in outstanding long-term debt, SIZE stands for the natural logarithm of the

The overreaction hypothesis predicts that the average α p over the five years of the test- period should be positive for the extreme prior losers (portfolio 1) and

Cumulative abnormal returns show a very small significant reversal (significant at the 10 per cent level) for the AMS Total Share sample of 0.6 per cent for the post event

Figure 7 Conceptual model Environmental beliefs H3 Country of origin effect H5 Energy labels H2 Purchase intention green labeled cars Gas prices H1 Performance H4

The main variable tested in their model was the toehold; Walking and Edmister (1985) used the toehold as a measure of bargaining strength. Another variable used was the

Rodríguez López and Rubio Martín (2019) also study the marketability discount based on a multiple approach but use a regression model to determine the marketability discount..

Based on theory using sales as a matching criterion should be preferred above deal value matching and therefore the results indicate discounts for private companies