RIVATE E QUITY VS . S TRATEGIC B UYER P

P

RIVATE

E

QUITY VS

. S

TRATEGIC

B

UYER

A

E

A

F

D

S

P

E

D

O

M&A

A

by

D

M

University of Groningen

Faculty of Economics and Business

MSc Business Administration – Corporate Financial Management

Thesis supervisor: Dr. Ing. N. Brunia

2

P

RIVATE

E

QUITY VS

. S

TRATEGIC

B

UYER

A

E

A

F

D

S

P

E

D

O

M&A

A

A

The share of private equity deals in overall M&A activity is determined by the ability to pay the highest premium. This study conducts research on factors that contribute to the ability to pay premiums. These factors are credit market conditions, private equity commitments, and mispricing in the stock market. Credit market conditions are measured by the cost of debt and the availability of debt (i.e. total debt per transaction, known as debt quantum). Private equity commitments are measured by the total annual fund commitments by investors. Stock market mispricing is measured by the S&P 500 price/earnings ratio. The dataset consists of all LBO/MBOs and strategic transactions, completed between 01/01/1997 and 30/06/2007 in the USA and Canada, and the European Union (27 member states). This yields a total of 16,531 transactions. I measure the share of private equity deals on a monthly basis. I find that an increasing cost of debt, overvalued stock markets and low private equity commitments reduce the share of private equity transactions. Increasing debt quantums, undervalued stock markets, and high private equity commitments increase the share of private equity transactions. Furthermore, strategic buyers appear not to be influenced by a changing cost of debt. Finally, I find significant differences between the variables of the North America, United Kingdom, and Continental Europe subsets.

JEL codes: G15, G24, G34

Key Words: strategic buyers, private equity, premiums, leverage, synergies, fund commitments, stock mispricing

D.M. Mostert

davidmostert@gmail.com

3

ACKNOWLEDGEMENTS

This thesis analyses the share of private equity deals in overall M&A activity. It marks the completion of my MSc Business Administration - specialization Finance, profile Corporate Financial Management at the University of Groningen.

Although the road travelled has proven to be difficult and demanding, in retrospect it has been an ultimately satisfying and rewarding experience. Private equity is known to be a difficult research topic, mainly due to the fact that data collection is quite challenging and often impossible. Hence, this study could not have been conducted without the data provided by PricewaterhouseCoopers.

In addition, I would like to express my gratitude to some people for contributing to the realization of this thesis. First, my thesis supervisor, Nanne Brunia, for his useful comments and ever so constructive criticism. Second, Donar van den Berg from Rabo Private Equity for his suggestions on my research topic. Third, Remco van Daal from PricewaterhouseCoopers Transaction Services for his guidance and advice during my internship. Finally, I owe many thanks to my parents and sister, for supporting me during the course of my studies in general, and my thesis in particular.

David Mostert

4

TABLE OF CONTENTS

I INTRODUCTION ... 5

II THEORETICAL FRAMEWORK ... 7

II.1 The ability to pay premiums ... 7

II.2 Credit market conditions ... 8

II.3 Private equity commitments ... 13

II.4 Stock market mispricing ... 14

5

I INTRODUCTION

According to research by Mergermarket (2008), gaining synergy effects is one of the key reasons for strategic buyers to do acquisitions. Once the acquired target is fully integrated, the acquirer can benefit from economies of scale (i.e. cost synergies) and/or generate more revenue than the two standalone companies would have been able to generate (i.e. revenue synergies). Hence, targets are more valuable than standalone values indicate. Strategic buyers often benefit from stock market mispricing (Shleifer & Vishny, 2003) and acquire targets with overvalued stock. Synergy effects and stock market mispricing therefore allow strategic buyers to pay premiums over the enterprise value of companies. Private equity firms only benefit from synergy effects when they employ a buy-and-build strategy (i.e. add-on acquisitions are merged with platform companies thus creating synergy effects). Private equity firms, however, often lever their deals up to 70%, whereas 30% is a more common level for strategic buyers (Axelson et al, 2007). As interest is tax deductible in most countries, leverage thus creates a valuable tax shield. Kaplan (1989) estimates that tax shields can explain 4% to 40% of a company’s value. Furthermore, fund commitments by (institutional) investors increase private equity transaction activity (Kaplan & Strömberg, 2008). Tax shields and high private equity commitments thus enable

private equity firms to pay premiums over the enterprise value targets.

Both types of buyers are therefore able to pay premiums. But under what circumstances is one type able to outbid the other? I analyse the influence of credit market conditions, the size of private equity commitments, and mispricing in the stock market on the monthly share of private equity transactions in overall M&A activity.

The dataset consists of all LBO/MBOs and strategic transactions between 01/01/1997 and 30/06/2007 in North America (USA and Canada) and the European Union (27 member states). I also perform tests on subsets, i.e. North America, the United Kingdom, and Continental Europe (i.e. European Union excluding the United Kingdom), as the private equity market is much more mature in North America and the United Kingdom than in Continental Europe. In addition, each of these regions has its own central bank, which is responsible for determining borrowing rates.

6

For the entire dataset, and the North America, United Kingdom, and Continental Europe subsets I find that an increasing cost of debt decreases the monthly share of private equity transactions. Alternatively, a decreasing cost of debt increases this share. Strategic buyers appear not to be affected by a changing cost of debt. The availability of debt, as measured by the debt quantum, has little explanatory power; it only has significant (positive) influence in the entire dataset and the North America subset. In each dataset overvalued stock markets and low private equity commitments reduce the share of private equity transactions, whereas that share increases when stock markets are undervalued and/or when private equity commitments are high. Finally, I find significant differences between the variables of the North America, United Kingdom, and Continental Europe subsets.

This thesis provides a useful framework to all parties involved in M&A. First, it enables both types of buyers to determine their competitive position based on changing market conditions. Strategic buyers know when to expect competition from private equity firms, and the latter know when they might not be able to outbid the first. Both types of buyers can adjust their strategy based on this information. Second, banks gain a better understanding of the implications of the conditions they set. Raising loan spreads or decreasing debt quantums might disable private equity firms to compete with strategic buyers. This reduces revenue generated from loans as private equity firms lever their deals to a larger extent than strategic buyers. Third, M&A advisors, including Transaction Services, can decide to invest more in relations with either one type of buyer, depending on market conditions.

7

II THEORETICAL FRAMEWORK

This theoretical framework is divided in four subsections. Section II.1 explains that the share of private equity transactions is determined by the ability of both types of buyers to pay premiums. Sections II.2 describes the influence of credit market conditions. Section II.3 discusses the influence of the private equity commitments. Section II.4 mentions the influence of stock market mispricing. Each subsection concludes with a hypothesis.

II.1 The ability to pay premiums

This study assumes that acquirers are able to complete transactions because they are able to outbid competitors. Other factors that may influence the bidding outcome are assumed equal for all bidders. As the intrinsic value of targets is equal, regardless of the type of buyer, the ability to pay in excess of the intrinsic value (i.e. premium) determines which buyer acquires the target. Consequently, the share of private equity deals in overall M&A activity depends on private equity firms’ ability to pay higher premiums than strategic buyers. The premium is subject to several factors, some are irrational (hubris), others can be explained. This paper examines three factors that may influence the share of private equity deals, where credit market conditions is divided in three sub-factors (figure I).

8 II.2 Credit market conditions

Section II.1 described that the share of private equity deals is determined by the ability to pay higher premiums than strategic buyers. This section explains how credit market conditions influence that ability. The use of bank loans increases the value of a company because (in many countries) the tax deductibility of interest lowers tax payments. Ross et al (2005) provide an example to illustrate this:

Unlevered Levered Earnings before interest and taxes (EBIT) 1,000,000 1,000,000

Interest 0 (400,000)

Earnings before taxes (EBT) 1,000,000 600,000

Taxes (35%) (350,000) (210,000)

Earnings after taxes 650,000 390,000

This example shows that leverage leads to lower tax payments. The difference between the cash flows (i.e. tax shield, 140,000) can be calculated as the tax rate (35%) times interest (400,000). The value of the tax shield of a company equals the present value of all future, annual tax shields. In his study on 76 management buyouts (MBOs) in the period 1980-1986, Kaplan (1989) finds that the tax shield can explain 4% to 40% of a company’s value and 21% to 143% of the premium paid. The estimated value depends on the rate buyout debt is repaid and the tax rate applied to the interest deductions.

Research by Standard & Poor’s (2008) on private equity deals in the United States of America and Europe yields average annual equity contributions of 33% for most years since 1999. Leverage is thus common practice in private equity transactions as it increases the value of the firm and thus the investment. This should not be different for strategic buyers, yet they tend not to lever their deals to the same extent as private equity firms. Axelson et al (2007) anecdotally remark that when you ask a typical CFO of a public firm about capital structure policy, he will express his need for financial flexibility and concern about distress costs, while a partner of a private equity firm will often say that they borrow as much as the banks are willing to lend them.

9

junior to debt, so they demand higher returns when debt levels increase. Furthermore, Ross et al (2005) mention that almost every industry has a debt level to which companies active in that industry adhere, regardless of what debt level would be optimal. In terms of transactions,Axelson et al (2007) find an average debt level of 14% and average net debt to EBITDA of 1.1x for transactions by strategic buyers in the USA and Europe between 1985-2006, compared to 67% and 5.4x for transactions by the 50 largest private equity funds in the same countries and period. Private equity firms try to maximize the benefits of the tax shield and do not have to answer to worrying shareholders or industry standards. Leverage thus enables private equity firms to pay premiums to a larger extent than strategic buyers. Strategic buyers, however, can benefit from economies of scale (i.e. cost synergies) and/or generate more revenue than the two standalone companies would have been able to generate (i.e. revenue synergies) once the acquired target is fully integrated. The target is thus more valuable than its standalone value indicates. Hence, depending on the size of the synergy effects, strategic buyers are able to pay a premium over the standalone value of the company. Research by Cools et al (2007) shows that the biggest synergies occur when the acquirer is at least twice as large as the target. Private equity firms usually do not benefit from synergy effects unless they employ a buy-and-build strategy where add-on acquisitions are merged with a platform company.

Filek, partner with PricewaterhouseCoopers' Transaction Services (USA) mentions that for the last two years prior to the economic downturn, strategic buyers were getting outbid by private equity firms as a result of favourable credit market conditions1. The leverage gains were thus often superior to the synergy benefits. Two important determinants of leverage, and to a lesser extent of synergy benefits, are the cost- and availability of debt.

The cost of debt

In addition to repaying the loan, borrowers have to pay interest. This cost of debt consists of two items: a base rate (usually LIBOR) and a premium, expressed as a number of basis points (bps) above the fixed rate. The base rate is in fact the amount of interest banks require when lending to one another. Therefore, to a large extent this rate depends on trust between banks. Events such as interest de/increases by central banks, or financial crises can cause substantial fluctuations. For example, overnight LIBOR lending rates increased from 310 bps to 643 bps on September 16th 2008, the day after Lehman Brothers filed for Chapter 112. The premium, or loan spread, is mainly determined by the risk profile of the borrower. Banks adjust their returns to the risk of default by the borrower.

When the cost of debt increases, debt becomes less attractive because it causes larger debt payments. In private equity transactions these payments may be too large for the target to bear once acquired. Private equity firms thus need to reduce the amount of debt financing they intend to request for

1

http://www.reuters.com

2

10

transactions compared to when the cost of debt was lower. Axelson et al’s (2007) dataset consists of all transactions between 1985 and 2006, completed by the 50 largest private equity funds in the USA and Europe. They find that increasing interest rates are associated with decreasing leverage. In order to prepare a competing offer, private equity firms need to contribute more equity. However, using less debt and more equity reduces the value of the tax shield and thus the Internal Rate of Return (IRR). Private equity firms aim for a certain minimum IRR; if this target cannot be met they often rather withdraw from a deal. Ljungqvist et al (2007) conducted research on investment data from 207 US private equity funds in the period 1981-2000. They find that private equity firms complete more transactions when credit market conditions are favourable to them. Hence, a larger cost of debt is associated with an absolute decrease in private equity activity.

When strategic buyers believe that an acquisition leads to synergy effects, these gains will be included in future cash flows. If the cost of capital increases, these cash flows will be discounted by a higher rate which causes synergy gains to decline. The influence of changes in the cost of debt of strategic acquisitions is seemingly small as many loans by strategic buyers are relatively small, or non-existing when transactions are paid in shares or with excess cash. Consequently, Axelson et al (2007) find that higher a cost of debt has little impact on leverage and pricing by strategic buyers.

In order to gain understanding of the influence of an increasing cost of debt on both types of buyers’ ability to pay premiums, consider the following example. Strategic buyer A has a debt-to-equity ratio of 1/4. The costs of equity and debt are both 5%, and the tax rate is 35%. The Weighted Average Cost of Capital (WACC)3 is thus 3/4*5 + 1/4*5*0.65 = 4.56%. Suppose the cost of debt increases to 6%. The WACC then becomes 3/4*5 + 1/4*6 = 4.73%. For the purpose of this example, assume that with a WACC of 4.56% synergy effects are 100. With the new WACC the synergy gains are 96.44 as the WACC increased by 3.56% (i.e. from 4.56% to 4.73%).

Private equity firm B has the same cost of equity and sees a similar increase in the cost of debt (from 5% to 6%). At a cost of debt of 5% the private equity firm is able to lever the deal for 75%. Hence, the WACC equals 1/4*5 + 3/4*5*0.65 = 3.69%. As the cost of debt increases 20% (i.e. from 5% to 6%), private equity firm B is unable to complete the transaction with 75% debt. Leverage decreases 20%, i.e. from 75% to 60%. The new WACC is 1/3*5 + 2/3*6*0.65 = 4.27%. Assume that at 75% the gains of the tax shield are equal to the synergy benefits of the strategic buyer: 100. As the WACC increases 15.71%, the gains of the tax shield reduce by the same percentage: from 100 to 84.29. The same increase in the cost of debt thus has a much larger influence on the transaction gains for private equity firms than for strategic buyers. There is, however, one pitfall to this example. As mentioned before, the cost of debt consists of the base rate and the loan spread. The base rate and the risk-free-rate,

3

11

which is included in the cost of equity, essentially co-move4. If the increase in the cost of debt from the example presented above is caused by an increase in the base rate, then the risk-free-rate increases by the same amount. Hence, the cost of equity also increases to 6%. The synergy benefits then reduce to 80, and the gains of the tax shield to 75.25. As a result, the difference between the gains of the two types of buyers is much smaller. In fact, if private equity firm B is able to sustain the debt level of 75% at a cost of debt of 6%, the gains of the synergy effects and tax shields are equal.5 Hence, besides the (difference between the) size of the synergy benefits and the tax shield gains, the ability to outbid the other type of buyer depends on the foundation of the change in the cost of debt. Strategic buyers are equally affected when the change is due to adjusted base rates, provided that leverage by private equity firms does not change. Changes in the loan spread only affect one specific (type of) buyer. Table I presents examples that influence the cost of debt of both types of buyers.

Table I

Examples of changes in the credit market that drive the cost of debt of both types of buyers

Base rate both Loan spread Loan spread types of buyer strategic buyer private equity

Actions by central banks X

State of the stock market X

Reassessment of risk of highly leveraged deals X

Examples of changes in the credit market that affect the cost of debt of strategic buyers and private equity firms. Actions by central banks include changing borrowing rates and providing money to banks. These actions should also affect the risk-free-rate which is included in the cost of equity. The state of the stock market changes the loan spread of strategic deals as it affects the risk of default by the borrower. A collective risk reassessment of LBOs affects the loan spread on private equity transactions for the same reason.

In conclusion, the fact that many strategic transactions are paid by excess cash or in shares should enable strategic buyers complete relatively more transactions than private equity firms when the cost of debt of private equity deals increases. Alternatively, a decreasing cost of debt of private equity transactions enables private equity firms to increase leverage. If the benefits of leverage outweigh synergy effects, private equity firms should be able to outbid private equity firms and thus complete relatively more deals. When strategic buyers use bank loans to complete transactions, they are often less affected by an increasing cost of debt, especially when the cost of equity is unaffected.

The above leads to hypothesis 1:

a. The share of private equity transactions declines when the cost of debt of private equity transactions increases

b. The share of private equity transactions is not influenced by the cost of debt of strategic deals

4

Table IX (appendix A) presents annual averages of the differences between the base rate and the risk-free-rate.

5

12

The availability of debt

The second factor that determines credit market conditions is the availability of debt. When liquidity in the credit market declines, banks most likely do not provide the same amount of debt to companies as in prosperous times. Standard & Poor’s (2008) confirms this by reporting decreasing loan volumes over the past year, but also when the internet bubble burst in 2002. Liquidity in the credit market can decrease to such an extent that debt financing is very hard or even impossible to get. Over the past decades we have witnessed several financial crises in the credit market (Standard & Poor’s, 2008):

- the subprime mortgage crisis in 2007-present - the internet bubble in 1999-2002

- the Japanese asset price bubble in 1991-2003 - the US savings and loans crisis in 1986-1995.

Though different causes, these crises all had similar effects: sudden sharp reductions in liquidity in the credit market (i.e. credit crunches). These crises are cases of extremes and reflect the ability of banks to provide loans. Besides the ability there should also be the willingness to provide loans. Banks set minimum interest coverage ratios6 before providing loans. This minimum, however, is among others subject to the state of the credit market (Standard & Poor’s, 2008).

When high interest coverage ratios are required, private equity firms may be forced to agree with lower debt levels than initially intended. As is the case with an increasing cost of debt, private equity firms need to contribute more equity in order to compete with strategic buyers.

The highly leveraged deal structure thus causes private equity firms to depend on bank financing in order to prepare a competing offer. It is therefore not surprising that Axelson et al (2007) find a relation between the availability of debt financing and deal activity by private equity firms. Axelson et al (2007) measure the availability of debt by the interest rates of transactions.

Strategic buyers are considerably less dependent of bank financing, as indicated by the debt levels mentioned earlier in this section. Their ability to pay premiums is therefore less affected by the ability of banks to provide loans. As a result of lower debt levels, requirements regarding interest coverage ratios are also less challenging than for private equity firms.

The above leads to hypothesis 2:

The share of private equity transactions declines when the availability of debt for M&A transactions decreases.

6

13 II.3 Private equity commitments

Private equity firms rely on commitments by limited partners to do acquisitions. Kaplan &Strömberg (2008) conducted research on private equity funds. Their dataset consists of funds in the USA, established between 1980 and 2007. They find that an influx of capital into private equity is associated with lower subsequent vintage year returns.7 This is due to the fact that private equity firms need to invest the fund commitments in order to generate returns for the investors. Private equity firms are thus willing to pay higher premiums to close the deal, which results in lower returns. Moreover, Kaplan & Strömberg (2008) find that commitments are positively and significantly related to private equity returns. As a result, we can identify a pattern in private equity activity (figure II). Once commitments turn out to be low, private equity firms are unable to pay high premiums which results in fewer transactions. Hence, strategic buyers benefit from decreasing competition which enables them to complete relatively more transactions. Alternatively, high commitments should enable private equity to pay high premiums and acquire relatively more companies.

Figure II:

Private equity commitment cycle

The above leads to hypothesis 3:

The share of private equity transactions declines when private equity commitments are low, and increases when commitments are high.

7

14 II.4 Stock market mispricing

Strategic buyers generally have two options in funding acquisitions: they either pay in cash (i.e. excess cash, if necessary complemented with debt financing) or in shares. Bradley & Sundaram (2004) performed empirical research on all completed public and non-public acquisitions made by publicly-traded US acquirers between January 1990 and March 2000. They find that 40% of all transactions were paid in cash only, while 25% only exchanged shares. The remainder was paid through a combination of cash and shares. In choosing between these methods, Koller et al (2005) mention that if the ability to pay in cash is not a constraint, the issue is whether the risks and rewards of the acquisition should be shared with the target’s shareholders. When the acquisition is paid in cash the acquirer’s shareholders bear the entire risk. On the other hand, if shares are exchanged the target’s shareholders will be exposed to risk too.

Research by Cools (2007) and Kummer & Steger (2008) has shown that M&A activity occurs in waves. Both studies identified six M&A waves over the past century8. Harford (2004) explains these M&A waves by economic, regulatory, and technological shocks. Shleifer & Vishny’s model (2003) opposes Harford’s point of view. This model shows that M&A waves are stock-market-driven and arise due to mispricing. Companies acquire targets and pay for the target’s hard assets by exchanging their overvalued stock. Shleifer & Vishny (2003) claim their model is able to explain each M&A wave and that it is consistent with empirical research by other authors.

Being overvalued therefore increases the ability to complete transactions. Most private equity firms, however, are not listed on stock exchanges and can therefore not benefit of being overvalued. Alternatively, strategic buyers may not be able to compete with private equity firms in times of undervaluation. Mispricing in the stock market thus influences the share of private equity deals in overall M&A activity.

The above leads to hypothesis 4:

The share of private equity transactions declines when stock markets are overvalued, and increases when markets are undervalued.

8

15

III DATA

To conduct this research I have been granted access by PricewaterhouseCoopers Debt Advisory (UK) to Reuters Loan Pricing Corporation’s Loan Connector. This database consists of all kinds of loans and provides information such as deal amounts (i.e. total debt), base rates, loan spreads, tranche amounts, maturities, purpose of the loan, industries, and country of the acquirers.

I select all transactions classified as “takeover” and “LBO/MBO”, completed between 01/01/1997 and 30/06/2007 in the European Union (27 member states) and North America (United States and Canada). As Loan Connector only shows strategic acquisitions with debt financing, I use Bureau Van Dijk’s Zephyr to obtain transactions without debt financing. All these 7,808 transactions are completed by strategic buyers. After merging with the Loan Connector data, 2,855 LBO/MBOs and 13,676 strategic acquisitions lead to a grand total of 16,531 transactions.

This study tests for influences on the share of private equity transactions in overall M&A activity. Figure III shows the shares of both types of buyers:

Figure III

Share of private equity transactions in overall M&A activity for the entire dataset.

The white line represents the share of private equity transactions in overall M&A activity The soft grey area represents strategic transactions, the dark grey area private equity deals.

This figure is based on all transactions by strategic buyers and private equity firms in North America (USA and Canada) and the European Union (27 member states) in the period 01-01-1997 - 06-30-2007.

Date P er ce n ta g e 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 1 -1 -1 9 9 7 1 -7 -1 9 9 7 1 -1 -1 9 9 8 1 -7 -1 9 9 8 1 -1 -1 9 9 9 1 -7 -1 9 9 9 1 -1 -2 0 0 0 1 -7 -2 0 0 0 1 -1 -2 0 0 1 1 -7 -2 0 0 1 1 -1 -2 0 0 2 1 -7 -2 0 0 2 1 -1 -2 0 0 3 1 -7 -2 0 0 3 1 -1 -2 0 0 4 1 -7 -2 0 0 4 1 -1 -2 0 0 5 1 -7 -2 0 0 5 1 -1 -2 0 0 6 1 -7 -2 0 0 6 1 -1 -2 0 0 7 Strategic buyer Private equity

Raw Data Sources Modifications Transformed to variable

All transactions after selection* Loan Connector, Zephyr Make distinction between deals from North America, United Kingdom and Continental Europe Share of private equity deals Make distinction between private equity deals and strategic transactions.

Divide private equity deals by transactions of both types of buyers.

Group transactions in monthly data and calculate share of private equity deals

Base rate Bloomberg terminal Match transaction date with Bloomberg LIBOR-GBP data

Group transactions in monthly data and calculate median base rate Add to loan spread

Loan spread Loan Connector Calculate weighted average of all loan spreads per deal

Group transactions in monthly data and calculate median loan spread Add to base rate

Debt tranche amount Loan Connector Sum all debt tranche amounts per deal and compare with total debt amount Debt quantum

www.x-rates.com Group transactions in monthly data and calculate sum of total debt both types of buyers Convert to local currencies in order to perform tests on regional subsets

Private equity commitments Private Equity Analyst Match annual commitment data with monthly transaction data Low commitments

EVCA, BVCA Calculate lowest quartile

Private equity commitments Private Equity Analyst Match annual commitment data with monthly transaction data High commitments

EVCA, BVCA Calculate highest quartile

S&P 500 price/earnings ratio Standard & Poor's website Match monthly S&P 500 price/earning ratios with monthly transaction data Undervalued stock markets Calculate lowest quartile

Match monthly S&P 500 price/earning ratios with monthly transaction data Overvalued stock markets Calculate highest quartile

* Sample selection criteria: This table provides an overview of what adjustments are made to the raw data in order to be transformed to the variables used in this study. - all takeovers and LBO/MBOs The raw data is collected from the sources indicated in the column "sources" and is based on the criteria mentioned left to this caption. - between 01-01-1997 and 06-30-2007 North America consists of the United States and Canada, Continental Europe consists of the European Union without the United Kingdom - in North America (USA and Canada) and

the European Union (27 member states)

Cost of debt (PE and SB) Table II

The share of private equity deals

The explained variable is the monthly share of private equity deals in overall M&A activity. Besides analyzing the entire dataset I decide to split the data into subsets and test for differences between the influences of each subset’s variables on the share of private equity deals. I split the dataset into North America (i.e. USA and Canada), the United Kingdom, and Continental Europe (i.e. the European Union (27 member states) without the United Kingdom). There are at least two reasons to split the data in these regions. First, private equity firms have been quite active in North America and the United Kingdom since the 1980s. These private equity markets are thus more mature than in Continental Europe (Strömberg, 2008). Hence, there are seemingly more opportunities for private equity firms to gain high returns in Continental Europe, which can lead to a larger share of private equity deals in overall M&A activity than in North America and the United Kingdom. Second, each of the aforementioned regions has their own central bank and currency.9 These central banks have their own policy in determining borrowing rates and supplying money to the banking system.

Table III presents an overview of all transactions for each type of buyer, per year, and per region. All regions have experienced some ups and downs along the way for both types of buyers: 2000 was a record breaking year, but once the dotcom bubble burst (in 2001/2002) overall M&A activity plummeted by 46% compared to 2000. Most share prices were overvalued during the internet bubble and were corrected by the market afterwards. The share of private equity transactions is considerably larger in Continental Europe than in North America and the United Kingdom, which seems consistent with the first reason to split the entire dataset in these regions.

The cost of debt

The first and second explanatory variables are the cost of debt of private equity deals and the cost of debt of strategic transactions. In the Loan Connector output each transaction consists of several debt tranches (term loan A, B, C, other loans etc.), each with its corresponding maturity, loan spread, and base rate (similar for each tranche). For the purpose of this research I calculate the weighted average loan spread for each transaction. Using the arithmetic mean would overstate the actual loan spread as Term Loans B, C etc and especially Other Loans usually have (much) higher loan spreads than Term loan A. Only taking the loan spread on Term Loan A (usually the largest tranche) would understate the actual loan spread. Loan Connector mentions which base rate is used for each transaction, but does not give the actual rate. From an interest rate parity10 perspective, I decide to use the three-month LIBOR-GBP for all transactions. These rates are collected from a Bloomberg terminal. The base rate plus the loan spread equals the cost of debt in basis points (bps).

9

North America: Federal Reserve System United Kingdom: Bank of England Continental Europe European Central Bank

10

Year N % N % N % N % N % N % N % N % 1997 957 87.6% 136 12.4% 756 86.4% 119 13.6% 172 96.1% 7 3.9% 29 74.4% 10 25.6% 1998 1,005 85.1% 176 14.9% 717 83.4% 143 16.6% 219 92.8% 17 7.2% 69 81.2% 16 18.8% 1999 1,102 84.5% 202 15.5% 678 82.3% 146 17.7% 267 91.4% 25 8.6% 157 83.5% 31 16.5% 2000 2,004 89.9% 226 10.1% 1,261 88.7% 161 11.3% 368 93.6% 25 6.4% 375 90.4% 40 9.6% 2001 1,516 90.8% 153 9.2% 1,102 93.0% 83 7.0% 229 87.4% 33 12.6% 185 83.3% 37 16.7% 2002 1,074 88.0% 147 12.0% 773 92.0% 67 8.0% 156 87.2% 23 12.8% 145 71.8% 57 28.2% 2003 999 81.8% 222 18.2% 746 86.2% 119 13.8% 141 79.2% 37 20.8% 112 62.9% 66 37.1% 2004 1,279 79.1% 337 20.9% 998 82.8% 208 17.2% 168 80.8% 40 19.2% 113 55.9% 89 44.1% 2005 1,376 77.6% 398 22.4% 991 80.4% 242 19.6% 224 81.2% 52 18.8% 161 60.8% 104 39.2% 2006 1,565 75.3% 513 24.7% 1,166 79.9% 293 20.1% 215 78.2% 60 21.8% 184 53.5% 160 46.5% 2007 799 69.8% 345 30.2% 593 72.9% 220 27.1% 100 66.7% 50 33.3% 106 58.6% 75 41.4% Total 13,676 82.7% 2,855 17.3% 9,781 84.5% 1,801 15.5% 2,259 86.0% 369 14.0% 1,636 70.5% 685 29.5% Note: 2007 up to June 30th

North America is the United States and Canada, Continental Europe is the European Union (27 member states) without the United Kingdom. This table presents an overview of the absolute number of transactions by strategic buyers and private equity firms, and each type of

buyer's share in overall M&A activity for the Entire dataset, the North America subset, the United Kingdom subset, and the Continental Europe subset.

Number of deals Table III

Private equity Strategic buyer Private equity

Strategic buyer Private equity Strategic buyer Strategic buyer Private equity

North America

Figure IV shows the median monthly costs of debt for private equity deals and strategic transactions expressed in basis points. For all but a few months the median monthly cost of debt of private equity deals is higher than the median monthly cost of debt of strategic transactions.

Figure IV

Median monthly cost of debt of both types of buyers for the entire dataset

Figure V

20

The debt quantum

The third variable is the debt quantum. I define the debt quantum as the total amount of debt financing available for all M&A transactions. It is measured as the monthly sum of total debt (i.e. sum of all debt tranches) per transaction. Loan Connector provides data on debt per tranche per transaction, expressed in British Pounds. Figure V shows the monthly debt quantums for both types of buyers combined in billion British Pounds (GBP bln). For the analysis of the subsets I converted the debt quantums to regional currencies (i.e. US Dollar and Euro) using monthly average exchange rates. Axelson et al (2007) measure debt availability by interest rates on transactions. Since I already used interest rates in the cost of debt variable I decided to focus on the debt quantum. The debt quantum can be seen as a measure of how much debt financing banks were able and/or willing to provide in that specific month.

Low / high private equity commitments

The Private Equity Analyst and the British- and European Venture Capital Associations (BVCA and EVCA) provide data on private equity fund commitments. Unfortunately it is unavailable on a monthly basis. Figure VI presents the annual commitments; table X (appendix B) contains all actual numbers. I construct two variables, one indicating low commitments and one for high commitments. The lower quartile of all annual commitments constitutes as low, the upper quartile as high, and the rest is neither high nor low. In order to match the annual commitments with the monthly data of the other variables, all months get the same label (low, high, neither) as the year they are in. The column with the totals is used for the entire dataset; for tests on the subsets I use local data.

Figure VI

Overview of annual private equity commitments for each region and in total

Annual private equity commitments per region and in total. North-America consists of the USA and Canada, Continental Europe equals the European Union (27 member states) without the United Kingdom. The commitments are in billions and local currency (US Dollar, British Pound, Euro respectively). The total is expressed in British Pounds; the commitments in North-America and Continental Europe are converted based on annual average exchange rates.

These numbers are commitments to all private equity types, the majority however is for LBOs. Sources: www.privateequityanalyst.com / www.bvca.org / www.evca.eu / www.x-rates.com

21

Under- / overvalued stock markets

As mentioned in section II.4, Shleifer & Vishny (2003) find that M&A waves are caused by mispricing of the stock market. A general measure of mispricing is the price/earnings (P/E) ratio. This ratio divides the current share price by the most recent earnings per share. Overvalued stock markets are associated with a high ratio, and undervalued stock markets with a low ratio. The Standard & Poor’s website11 provides monthly P/E ratios for the S&P 500 (500 largest listed companies in the USA). Unfortunately the FTSE in the United Kingdom does not provide such a ratio, nor is there one for Continental Europe. This, however, is not such a problem as a mispriced S&P 500 index will influence other stock exchanges as well, as many multinationals are listed on multiple stock exchanges. The S&P 500 P/E ratio should therefore have explanatory power on the United Kingdom and Continental Europe subsets as well. Figure VII presents the historical monthly S&P 500 ratios, the exact numbers can be found in table XI (appendix C). The lower quartile of all monthly P/E ratios constitutes as undervalued stock markets, the upper quartile as overvalued stock markets, and the rest is neither under- nor overvalued.

Figure VII

Monthly S&P 500 price/earnings ratios

11

22

Descriptive statistics

23

Table IV

Descriptive statistics of explanatory variables

Entire dataset Cost of debt PE Cost of debt SB Debt quantum Low High Under Over Observations 126 126 126 126 126 126 126 Mean 7.478 6.671 21.683 0.254 0.254 0.254 0.254 Median 7.635 6.703 17.555 0.000 0.000 0.000 0.000 Minimum 4.744 3.478 3.719 0.000 0.000 0.000 0.000 Maximum 10.264 9.843 77.391 1.000 1.000 1.000 1.000 Standard deviation 1.206 1.443 13.980 0.437 0.437 0.437 0.437 Skewness 0.144- 0.320- 1.644 Kurtosis 2.849 2.834 6.154 Jarque-Bera 0.553 2.301 109.000

North America Cost of debt PE Cost of debt SB Debt quantum Low High Under Over Observations 126 126 126 126 126 126 126 Mean 7.380 6.570 12.869 0.254 0.254 0.254 0.254 Median 7.485 6.550 11.006 0.000 0.000 0.000 0.000 Minimum 5.015 3.517 1.859 0.000 0.000 0.000 0.000 Maximum 10.297 9.923 46.658 1.000 1.000 1.000 1.000 Standard deviation 1.230 1.441 7.800 0.437 0.437 0.437 0.437 Skewness 0.183 0.145 1.236 Kurtosis 2.482 2.464 5.207 Jarque-Bera 2.107 1.952 57.664

United Kingdom Cost of debt PE Cost of debt SB Debt quantum Low High Under Over Observations 126 126 126 126 126 126 126 Mean 7.764 6.629 2.538 0.254 0.254 0.254 0.254 Median 7.602 6.520 1.590 0.000 0.000 0.000 0.000 Minimum 5.985 4.098 0.029 0.000 0.000 0.000 0.000 Maximum 14.330 10.706 15.662 1.000 1.000 1.000 1.000 Standard deviation 1.099 1.271 2.854 0.437 0.437 0.437 0.437 Skewness 1.876 0.498 2.145 Kurtosis 11.603 3.343 8.497 Jarque-Bera 462.506 5.832 255.212

Continental Europe Cost of debt PE Cost of debt SB Debt quantum Low High Under Over Observations 126 126 126 126 126 126 126 Mean 5.624 4.534 6.662 0.254 0.254 0.254 0.254 Median 5.597 4.500 3.946 0.000 0.000 0.000 0.000 Minimum 4.174 2.507 0.062 0.000 0.000 0.000 0.000 Maximum 8.309 8.121 40.925 1.000 1.000 1.000 1.000 Standard deviation 0.909 1.242 7.532 0.437 0.437 0.437 0.437 Skewness 0.523 0.683 2.148 Kurtosis 2.654 3.110 8.032 Jarque-Bera 6.369 9.868 229.808

These tables present descriptive statistics for the variables of i) the entire dataset, ii) the North America subset iii) the United Kingdom subset, and iv) the Continental Europe subset. The cost of debt of both types of buyers

is expressed in percentages, the debt quantum in billion British Pound, US Dollar, British Pound, and Euro respectively. The variables Low, High, Under, and Over are represented by dummy variables. For each variable one quartile is marked 1. The descriptive statistics are therefore equal for each variable and each dataset.

24

IV METHODOLOGY

We are interested in the relative shifts in transaction activity by both types of buyers. Therefore I use the share of private equity transactions as dependent variable:

RPE = PE / (SB + PE) (1)

where RPE is the share of private equity transactions, PE is the number of transactions by private equity firms and SB is the number of strategic acquisitions. RPE is in fact the white line in figure III. An RPE of 0.11 means that 11% all transactions for that specific period of time were made by private equity firms. The probability that a transaction is made by a private equity firm is thus 11% or 0.11. The most common model to estimate equations is the Ordinary Least Squares (OLS) method. However, one problem with OLS is that RPE is always between zero and one, but the explanatory variables can take any real value (except for the dummy variables). This means that there is no guarantee that the predicted values will be in the correct range unless complex restrictions are imposed on the coefficients (Rodríguez, 2007). The solution is to remove the range restrictions, and to model the transformed dependent variable as a linear function of the explanatory variables. This can be done in three steps. First, I change the probability of a private equity transaction πi into the corresponding odds πi/(1- πi). This removes the ceiling restriction as odds can take any positive value. Second, I take the (natural) logarithm (LN) of the odds, thus calculating the logit:

logit(πi) = LN [πi/(1- πi)] (2)

This removes the floor restriction. As the probability goes down to zero the odds approach zero and the logit approaches - ∞. When probabilities near one, the odds and logit approach + ∞. If the probability is 0.5 (i.e. equal shares for private equity firms and strategic buyers) the odds are 1:1 and the logit is zero. The final step is to construct a linear model, using the logit as dependent variable:

Logit = β0 + β1Xi1 + β2Xd2 + β3Xq3 + β4Xl4 + β5Xh5 + β6Xu6 + β7 Xo7 + µ (3)

25

To test whether differences between the coefficients of the variables of each region are significant, Seemingly Unrelated Regressions (SUR) are performed on an unrestricted system and on restricted systems. I impose restrictions on each variable of each region pair (i.e. the cost of debt of private equity transactions in Continental Europe and North America). The F-test statistics on the differences between the residual sums of squares (RSS) of the unrestricted and the restricted systems indicate whether the restrictions are supported by the data.

In order to calculate the exact influence of a variable on the share of private equity transactions we first need to transform the logit back to RPE, and then use the following equation:

RPE = 1/(1+e - logit

) (4)

where e (Euler’s number) represents the inverse of the natural logarithm.

Equation (3) is estimated with the Ordinary Least Squares (OLS) method. According to Brooks (2002), five assumptions are made in order to test hypotheses validly using OLS. These assumptions are the following:

E(ut) = 0 The average value of the errors is zero

This assumption cannot be violated when a constant term is included in the equation, which is the case in equation (3).

var(ut) = σ 2

< ∞ The variance of errors is constant (i.e. homoscedastic)

If heteroscedasticity is ignored, the standard errors of the variables could be wrong and any inferences made could be misleading. Consequently, I estimate equation (3) with heteroscedastic robust standard errors.

cov (ui,uj) = 0 The covariance between the error terms over time is zero (i.e. uncorrelated)

If the errors are not uncorrelated with one another, it would be stated that they are auto-correlated. The Durbin-Watson statistic indicates whether autocorrelation is present in the data. For this dataset the Durbin-Watson statistic should be between 1.653 and 2.347.

cov (ut, xt) = 0 The xt are non-stochastic

26

ut ~ N(0, σ 2

) The disturbances are normally distributed

A normal distribution is not skewed and is defined to have a coefficient of kurtosis of 3. The Jarque-Bera statistic has a critical value of 5.99 (with 5% significance). Any inferences drawn might be misleading when the disturbances are not normally distributed. However, for sample sizes that are sufficiently large, violation of this assumption is virtually inconsequential (Brooks, 2002).

An additional implicit assumption that is made using the OLS method is that the explanatory variables are not correlated with each other (i.e. multicollinearity). Near-multicollinearity arises when two or more explanatory variables show a considerable relationship. According to Brooks (2002) it can be measured by means of a correlation matrix. Table XII (appendix D) presents correlation matrices for all datasets. As can be seen, the numbers for the cost of debt for private equity deals and for strategic transactions indicates near-multicollinearity. I solve this issue by substituting the variable cost of debt

of strategic transactions by the difference between the cost of debt of private equity deals and strategic transactions. A significant negative coefficient for the variable difference indicates that the share of private equity transactions decreases when the difference between the cost of debt of strategic transaction and private equity deals increases. Alternatively, an insignificant coefficient means that the difference has no impact on the share of private equity deals. This implies that the cost of debt of strategic deals has no explanatory power on the share of private equity deals. Table XII shows that the correlation matrices with the new variable difference improved compared to the correlation matrices with the cost of debt of strategic transactions as variable.

27

V RESULTS

This section describes the statistical significance of the regression results of equation (3) on the entire dataset and subsequently on the subsets. Next, I elaborate on the economical significance of the regression results.

V.1 Statistical significance

Consistent with hypothesis 1a, an increasing cost of debt of private equity deals reduces the share of private equity transactions, as can be seen in table V. The variable difference turns out to have no explanatory power, as the coefficient is insignificant. This implies that the share of private equity deals is not influenced by the cost of debt. An additional method to test this statement is to separate the base rate (LIBOR) from the loan spread. Increasing base rates reduce private equity firms’ appetite to complete transactions. However, as LIBOR is more or less equal to the risk-free-rate, an increasing base rate would also increase the cost of equity and thus reduce the appetite by strategic buyers, even when bank loans are not used to acquire the target. The estimation results in table XV (appendix F) show that the coefficient of the base rate, based on all transactions, is negative and significant. This implies that an increase of the base rate has a stronger influence on private equity firms than on strategic buyers. Finally, I applied a direct test where the share of private equity deals is solely explained by the cost of debt of strategic transactions. Table XIV (appendix E) shows that the coefficient is negative, but insignificant. Hence, we cannot reject hypotheses 1b that the share of private equity transactions is not influenced by the cost of debt of strategic deals.

Hypothesis 2 also cannot be rejected, as the coefficient is of the debt quantum positive and significant. The ability of banks to provide loans thus affects the share of private equity deals, although the low value of the coefficient implies that the influence is not very strong. Axelson et al (2007) also find that the availability of bank financing influences transaction activity by private equity firms, although they used a different metric (i.e. interest rates).

28

Finally, mispricing in the stock market, as expressed by the price/earnings ratio of the S&P 500, indicates that an overvalued stock market is associated with a decreasing share of private equity transactions, and alternatively, an undervalued stock market leads to an increasing share of private equity deals. This is consistent with hypothesis 4, and with Shleifer & Vishny’s (2003) model, which predicts that strategic buyers use shares to acquire targets in times of overvalued stock markets.

This model seems to fit the data very well. Most explanatory variables are significant at 1%, and the Adjusted R2 is 85%. In addition, the F-statistic rejects the null hypothesis that all of the slope coefficients are equal to zero. Furthermore, the Durbin-Watson statistic (2.089) indicates that there is no evidence of autocorrelation (i.e. OLS assumption 3). Moreover, table III shows that the Skewness, Kurtosis, statistics slightly deviate from the expected values of 0.00 and 3.00. Hence, the Jarque-Bera statistic exceeds the critical value of 5.99, which indicates that the data is not normally distributed. However, according to Brooks (2002) this occurs often in empirical research based on financial data so I decide to ignore it. The section on methodology explains how I deal with multicollinearity.

Table V

Estimation results using OLS with heteroscedastic robust standard errors

Entire dataset North America United Kingdom Continental Europe

Constant -1.163* -1.621* -1.630* -0.700* (-10.796) (-6.751) (-6.277) (-2.649) Cost of debt -0.092* -0.051*** -0.066** -0.129* (-5.436) (-1.678) (-2.142) (-2.646) Debt quantum 0.008* 0.024* 0.006 0.000 (5.723) (4.758) (0.405) (0.543) Low commitments -0.560* -0.626* -0.212*** -0.800* (-9.560) (-8.883) (-1.717) (-8.786) High commitments 0.166* 0.311* 0.177** 0.312* (2.822) (4.080) (2.090) (4.486) Undervaluation 0.247* 0.152 0.767* 0.172*** (4.257) (1.220) (8.435) (1.958) Overvaluation -0.093*** -0.159*** -0.564* -0.132** (-1.723) (-2.654) (-4.474) (-2.071) Adj. R2 0.845 0.642 0.718 0.780 F-statistic 114.520 38.204 54.101 74.975 Durbin-Watson 2.089 1.754 2.121 1.866 Skewness -0.684 -0.665 0.256 -0.549 Kurtosis 3.777 4.851 3.818 3.841 Jarque-Bera 13.013 18.342 4.894 10.044

Note: T-statistics in parentheses * Significant at the 1% level ** Significant at the 5% level *** Significant at the 10% level

29

In order to search for differences between regions I divide the data into three subsets: North America, United Kingdom, and Continental Europe (27 member states excluding the United Kingdom). Table V shows that the Adjusted R2 values of the subsets are quite high, consistent with the entire dataset. The Durbin-Watson values do not indicate autocorrelation, and the statistics of the normality tests show that the disturbances are only normally distributed in the United Kingdom subset.

Table V shows that the influence of the cost of debt is largest in the Continental Europe. In each subset the difference between the cost of debt of strategic transactions and private equity deals is insignificant, as is the case in the entire dataset, so this variable is excluded from equation (3). The ability of banks to provide loans has little explanatory power, as the coefficient of the debt quantum is only significant in the North America subset. Private equity commitments have the strongest influence in the Continental Europe subset, both when commitments are low and high. The difference with the North America subset, however, is very small. The influence of mispricing in the stock market has a strong influence on the share of private equity transaction in the United Kingdom. Remarkably, in North America the influence of undervalued stock markets is insignificant.

As described above, when comparing the region subsets we see differences between the coefficients of each variable. To find out whether these differences are significant, I perform Seemingly Unrelated Regressions (SURs) on unrestricted- and restricted systems. Table VI presents the F-test statistics of these SURs.

Table VI

F-test statistics on residuals sums of squares of (un)restricted Seemingly Unrelated Regressions

Restriction on North America & North America & United Kingdom & United Kingdom Continental Europe Continental Europe

Cost of debt 0.019 4.707** 2.811 Debt quantum - - -Low commitments 8.981* 0.213 15.196* High commitments 1.745 0.189 3.964** Undervaluation - - 8.965* Overvaluation 24.090* 1.903 18.723* Note:

* Significant at the 1% level (critical value F-distribution: 6.694) ** Significant at the 5% level (critical value F-distribution: 3.858)

- Insignificant coefficients

30

I find the following based on tables V and VI:

Table VII

Implications of table V and VI

The influence of is significantly greater in than in

Cost of debt private equity Continental Europe North America

Low commitments North America United Kingdom

Low commitments Continental Europe United Kingdom

High commitments Continental Europe United Kingdom

Undervalued stock markets United Kingdom Continental Europe

Overvalued stock markets United Kingdom North America

Overvalued stock markets United Kingdom Continental Europe

Implications of tables V and VI

The cost of debt of private equity deals has a significantly greater influence in Continental Europe than in North America. The share of private equity deals in overall M&A activity is also considerably larger in Continental Europe. These two findings are consistent with (the inverse of) hypothesis 1a. The influence of (low) private equity commitments is significantly greater on the share of private equity deals in Continental Europe and North America than in the United Kingdom. The state of the stock market strongly affects the share of private equity transactions in the United Kingdom. The logic behind the above is beyond the scope of this study but I strongly recommend research on this topic. V.2 Economical significance

31

influence of private equity commitments is very strong in each dataset, except for the United Kingdom. Low commitments reduce the share of private equity transactions in Continental Europe by 9.65%. High commitments lead to an increase of 5.35%.

Finally, the influence of mispricing in the stock market is very strong in the United Kingdom: undervalued stock markets raise the share of private equity transactions by 10.02%, overvalued stock markets reduce this share by 4.34%.

Table VIII

Economic effects of the variables on the share of private equity transactions

North America United Kingdom Continental Europe

Median cost of debt 7.635 7.485 7.533 5.597

Cost of debt increase: 1 SD -1.36% -0.72% -0.66% -1.77%

Cost of debt increase: 2 SD -2.61% -1.41% -1.29% -3.41%

Cost of debt decrease: 1 SD 1.46% 0.76% 0.70% 1.90%

Cost of debt decrease: 2 SD 3.04% 1.55% 1.44% 3.94%

Median debt quantum in bln 15.458 10.554 2.490 3.537

Debt quantum increase: 1 SD 1.51% 2.40% 0.00% 0.00%

Debt quantum increase: 2 SD 3.15% 5.13% 0.00% 0.00%

Debt quantum decrease: 1 SD -1.40% -2.09% 0.00% 0.00%

Debt quantum decrease: 2 SD -2.68% -3.89% 0.00% 0.00%

Private equity commitments*

Low commitments -5.83% -5.82% -1.84% -9.65%

High commitments 2.24% 4.07% 1.76% 5.35%

Mispricing in stock market*

Undervaluation 3.41% 0.00% 10.02% 2.84%

Overvaluation -1.14% -1.76% -4.34% -1.99%

Note: 0.00% represent insignificant variables * By default on neither

This table presents the effects of in/decreases on the variables on the share of private equity transactions. As starting points I take the median values of the cost of debt and the debt quantum.

The median cost of debt and debt quantum in/decrease by one and two standard deviations.

The median debt quantums differ from those in the descriptive statistics as a result of the removal of outliers. The same goes for the cost of debt of private equity deals in the United Kingdom subset. Continental Europe consists of the European Union (27 member states) without the United Kingdom. North America equals the United States of America and Canada.

32

VI CONCLUSION

This section presents the main conclusions of this study, and discusses whether the hypotheses of section II should or should not be rejected. This section also reflects on the limitations of the study, and provides recommendations for further research.

Conclusions

Hypothesis 1a proposes that the share of private equity transactions declines when the cost of debt of private equity transactions increases. Both the entire dataset and the subsets support this hypothesis, as each coefficient has a negative sign and is significant. The influence on the share of private equity transactions is smallest in North America, and largest in Continental Europe. The difference is significant. The fact that Continental Europe has the largest share of private equity deals and lowest cost of debt reflects the estimation results.

Hypothesis 1b proposes that the share of private equity transactions is not influenced by the cost of debt of strategic deals. The variable “difference” is insignificant for each dataset, and the results in table VI amplify the suggestion that the cost of debt of strategic deals has no explanatory power. The results of hypotheses 1a and 1b are in line with Axelson et al (2007) and Ljungqvist et al (2007). Although these studies do not conduct research on the share of a specific type of buyer, they do find a negative relation between interest rates, leverage, and transaction activity by private equity firms. In my study leverage influences the share of private equity transactions as it increases the ability to pay premiums.

Hypothesis 2 proposes that the share of private equity transactions declines when debt quantums for both types of buyers decrease. The results of the entire dataset confirm this. However, based on the subsets there is little support for this hypothesis, as only the North America subset yields a significant coefficient. The influence of the availability of debt on the share of private equity transactions is therefore weak, at least as measured by debt quantums.

33

Hypothesis 4 proposes that the share of private equity transactions declines when stock markets are overvalued, and increases when stock markets are undervalued. Except for undervaluation in North America each dataset confirm this hypothesis. The influence on the share of private equity transactions is smallest in Continental Europe, and largest in the United Kingdom. The difference is significant. Shleifer & Vishny (2003) do not make a distinction in economic regions, nor do they conduct research on the share of private equity transactions, but they do find that overvaluation leads to concentrated M&A activity.

These results may not be shocking or irrational to people with knowledge on M&A. There is, however, no existing academic literature regarding this specific topic. This study is thus able to replace intuition with empirical evidence, based on an analysis of 16,531 transactions in over 10 years and multiple regions.

Limitations and recommendations

This thesis is subject to some limitations; these lead to recommendations for further research. First, I was unable to collect and/or match enterprise values for each transaction. Loan Connector does not provide enterprise values, and matching the data with Bureau Van Dijk’s Zephyr is impossible as Loan Connector often mentions project names instead of acquirer/target names. It would be very interesting to test for differences based on enterprise value segments. It is quite likely that the share of private equity deals in overall M&A activity differs between enterprise value segments. The same goes for industries, some industries are historically popular for each type of buyer, whereas others are not. Second, the variable debt quantum does not prove to be a proper measure for debt availability for the subsets of the United Kingdom and Continental Europe, as the coefficients of these subsets are insignificant. This is due to the fact that some months only contained a few transactions, and as result the debt quantum data sometimes fluctuates heavily, which is reflected by the signs of non-normality. General data on monthly bank loan volumes probably provide a better measure of the ability by banks to provide debt financing.

Third, private equity commitments are only available on annual basis. As all other variables are on a monthly basis, the effects of private equity commitments presented may be biased. Actual data on fund commitments is very hard if not impossible to obtain but would be very useful, as are IRRs on these commitments.

Fourth, this study is based on the period prior to the credit crunch and subsequent recession. The relevancy of the results of this study for future research depends on many factors, but primarily the future state of the credit market.

34

REFERENCES

Axelson, U., Jenkinson, T., Strömberg, P., Weisbach, M., (2007), “Leverage and pricing in buyouts: an empirical analysis”, Working Paper SIFR

Bradley, M. and Sundaram, A., (2004), “Do Acquisitions Drive Performance or Does Performance Drive Acquisitions?” Working Paper Thunderbird Research Center

Brooks, C. (2002), “Introductory Econometrics for Finance”, Cambridge University Press, Cambridge

Cools, K., Gell, J., Kengelbach, J., Roos, A., (2007), “The Brave New World of M&A: How to Create Value from Mergers and Acquisitions”, BCG Report

Han Kim, E. (1977), “A Mean-Variance Theory of Optimal Capital Structure and Corporate Debt Capacity”, Journal of Finance Vol 33 No.1, 1978 p45-63

Harford, J., (2004), “What Drives Mergers Waves?”, Journal of Financial Economics Vol. 77, 2005 p529-560

Kaplan, S. (1989), “Management Buyouts: Evidence on Taxes as a Source of Value”, Journal of

Finance Vol 64 No. 3, 1989 p611-632

Kaplan, S. and Strömberg, P., (2008), “Leveraged Buyouts and Private Equity”, Working paper 14207

Koller, T., Wessels, D., Goedhart, M., (2005), “Valuation: Measuring and Managing the Value of Companies”, Wiley, New York

KPMG, (2008), “Competing For Growth 2008”

Kraus, A. and Litzenberger, R., (1973), “A State Preference Model of Optimal Financial Leverage”,

Journal of Finance Vol. 28 No.4, 1973 p.911-922

Kummer, C., and Steger, U., (2008), “Why M&A Waves Reoccur: The Vicious from Pressure to Failure”, Strategic Management Review Vol.2 No.1, 2008 p44-63

35

Mergermarket, (2008), “Dutch Mid-market survey”

PricewaterhouseCoopers, (2008), “European Leveraged Finance Market Update Q3”

Rajan, R. and Zingales, L., (1995), “What Do We Know about Capital Structure? Some Evidence from International Data”, Journal of Finance Vol 50 No.5, 1995 p1421-1460

Rodríguez, A. (2007), “Logit Models for Binary Data”, Working Paper

Ross, S., Westerfield, R., Jaffe, J., (2005), “Corporate Finance”, McGraw-Hill, New York

Shleifer, A. and Vishny, R., (2003), “Stock Market Driven Acquisitions”, Journal of Financial

Economics Vol 70, 2003 p295-311

Standard & Poor’s, (2008), “LCD Leveraged Loan Review – US / Europe”

Strömberg, P. (2008), “The New Demography of Private Equity”, Global Economic Impact of Private

Equity Report 2008, 2008 p3-26

Websites

http://www.bvca.co.uk Accessed March 10th 14:30

http://www.economagic.com Accessed October 2nd 15:30

http://www.evca.eu Accessed March 11th 09:30

http://www.reuters.com Accessed October 31st 10:30

http://www.privateequityanalyst.com Accessed February 23rd 13:00

http://www.standardandpoors.com Accessed several times

36

APPENDIX

A

Table IX

Difference between the base rate and the risk-free-rate

Year Average 1997 0.46 1998 0.57 1999 0.56 2000 0.45 2001 0.31 2002 0.27 2003 0.20 2004 0.34 2005 0.36 2006 0.36 2007 0.39 Total 0.39

The TED spread is the difference between the three-month T-bill rate (i.e. risk-free-rate) and the three-month LIBOR (i.e. base rate). Source: Bloomberg software and calculations

37 B

Table X

Annual private equity commitments in each region and as a total

North America United Kingdom Continental Europe Total

1997 41.5 12.8 1.4 39.1 1998 61.9 7.1 9.6 50.7 1999 43.4 8.4 12.2 43.4 2000 79.6 16.3 21.4 82.0 2001 51.5 20.2 7.5 60.4 2002 43.1 15.4 3.5 46.5 2003 28.4 14.7 5.8 36.1 2004 57.4 6.1 18.0 48.8 2005 110.8 27.3 32.1 110.3 2006 148.8 34.3 61.6 156.5 2007 228.0 29.3 35.6 _167.4

Annual private equity commitments per region and in total. North America consists of the USA and Canada, Continental Europe equals the European Union (27 member states) without the United Kingdom. The commitments are in billions and local currency (US Dollar, British Pound, Euro respectively). The total is expressed in British Pounds; the commitments in North America and Continental Europe are converted based on annual average exchange rates.