The Role of Private Equity in Private Acquisitions - An Empirical Study to Private Equity Exits –

The Role of Private Equity in Private Acquisitions

- An Empirical Study to Private Equity Exits –

Koos Leisink 1434012

August 2009

Supervisor: H. Gonenc

University of Groningen Faculty of Economics and Business

The Role of Private Equity in Private Acquisitions

- An Empirical Study to Private Equity Exits –

Koos Leisink1

Abstract

In this paper, we compare bidder announcement returns for strategic listed bidders that acquire European private equity (PE) backed targets to bidder returns for non PE backed, private targets. We study the impact of a liquidity discount, information asymmetry, reputation, selling capability and length of PE involvement and find that PE backed targets suffer less from the need for liquidity than their non PE backed counterparts. We also find that when PE firms exit within 30 months after the buyout, bidder returns tend to be significantly higher. Apart from these hypotheses, we find that for cash-financed deals, bidders of PE targets show a lower announcement return, whereas the difference for other deals is insignificant. We also find that bidder returns for targets acquired by Anglo-American companies are significantly lower than returns for Rest of World companies.

Keywords: acquisitions, announcement returns, private equity, information asymmetry, liquidity discount, reputation, selling capability, length of PE involvement

JEL-classification: G14, G24, G34

1

1. Introduction

Under the efficient market hypothesis, share prices should adjust immediately after an acquisition announcement and the new price should reflect the combined value of acquirer and target. Several studies show that the efficient market hypothesis does not hold (e.g. Agrawal, Jaffe and Mandelker, 1992; Loughran and Vijh, 1997). Researchers have tried to explain this anomaly and have studied several deal and target characteristics. One of the characteristics that has been studied is the target type of firm, that is either public, private or a subsidiary (Kooli, Kortas and L’Her, 2003; Fuller, Netter and Stegemoller, 2002). However, no distinction has been made for firms that are private-equity backed.

This paper studies strategic bidder announcement returns for firms acquiring private targets. Hereby, a differentiation is made between private firms that are private equity (PE) backed and firms that are not. PE backed targets means an asset that was first acquired by a PE firm and sold (exited) to a strategic buyer. With non PE backed targets, we refer to assets owned by others than PE firms and that are sold to a strategic buyer. Notably, we only study strategic buyers that are listed. We hypothesise that bidder returns of PE backed targets differ from the returns of non PE backed targets. In other words, the market reaction to the announcement of an acquisition done by a strategic bidder differs when the target is sold by a private equity firm than when the target is sold by a non-private equity firm. An important explanation from the literature is the existence of information asymmetry. One of the most important differences between listed and unlisted targets is the public availability of information. When a listed company is subject to a takeover, all parties involved in the acquisition have sufficient information about the target and the difference between what the acquirer and the seller know about the target is relatively small. When a target is unlisted, the information gap between acquirer and seller can be much larger. It is only common sense to say that when information is absent, acquirers react risk-averse and adjust their expectations downwards. This information asymmetry theory by Hansen (1987) is supported and quantified by data from Officer (2007). However, some of these private targets are sold by PE firms whose main objective is to maximise their return. As PE firms buy as many companies as they sell, they perfectly know that the absence of information leads to lower prices. It looks only logical that PE firms will do anything within their reach to close the information gap. For example the repeatedly use of vendor due diligence in sales processes proves this expectation.

the easy access to the capital market. If a firm is in desperate need for liquidity, the capital market can provide cash in exchange for shares (on condition that the price is right). This access to liquidity is more difficult to private firms. Debt can be a solution to a cash problem, although this source has its limitations and can also be the cause of the need for liquidity. One of the other options for private firms is to sell a company or a subsidiary in order to receive cash. In a study by Officer (2007), the author finds that the proceeds of subsidiaries sold by private firms cause an increase of the cash level of more than 100%. Also for private owners other than companies, the proceeds from a sale can be very useful, for example to invest in other companies or funds. If we consider PE firms on the other hand, we see firms with very large funds which makes it unlikely that they exit to fulfil their cash need. Like listed firms, PE players typically have access to large cash pools. In other words, it does not seem likely that PE firms are forced to sell their assets in order to keep their operations going or that the cash is absolutely needed for another investment opportunity.

rest of the world. This result once more indicates that generalising results should be done with caution.

the VC-backed acquisitions with non-VC-backed transactions. In addition, Gompers and Xuan (2008) study US transactions of VC-backed companies, whereas this study contains a dataset of PE-backed targets in the European Union (UK excluded).

The results of this study are important to both academics and practitioners in the field of corporate finance as private equity is in a constant upward stream of both more and larger deals. Also, as the field of private equity is relatively unexplored as a result of its private nature, new studies and methodologies to this topic help to unravel this field of research. Strategic bidder returns can be a helpful proxy to deal with this private nature.

This study continues with a theoretical background and hypotheses. Sections 3 and 4 describe the data and methodology respectively and section 5 includes the results of the univariate analyses and OLS regressions. We draw conclusions and discuss this study in section 6.

2. Theoretical background

2.1 Public vs private targets

Post-acquisition performance has been widely studied in the financial literature. Several studies determine the short term effects (Chang, 1998; Fuller et al., 2002) and long term effects (Asquith, 1983; Agrawal et al., 1992; Loughran and Vijh, 1997) of acquisitions. Although the different findings do not provide an unanimous conclusion, most studies find positive target announcement abnormal returns, negative to slightly positive acquirer announcement abnormal returns and negative long-term abnormal acquirer performance. These negative abnormal returns are found up to five years after the acquisition (Agrawal et al., 1992; Loughran and Vijh, 1997).

analyses. A number of (non-mutual exclusive) theories explain (partially) the discount for unlisted targets and the positive market reaction to the announcement.

Officer (2007) argues that the need for liquidity is an important factor why private targets and subsidiaries are sold at a discount. Liquidity is important to corporations but there are just a few ways to access the cash pool. One possibility is the sale of an (unlisted) asset. This is an important source for cash since private firms cannot place a seasoned equity offering. An alternative could be an IPO of either a subsidiary or an entire stand-alone firm, but this comes at substantial costs (cost of offering securities and underpricing when the securities are sold in the market). Subsequently, Officer (2007) argues that sellers of private targets value the cash proceeds of the sale and – more important – that buyers and the market know this which results in a lower asset price due to this liquidity discount. Fuller et al. (2002) elaborate on this from an auction perspective. They state that private targets prefer to be sold in an auction with several bidders (like public targets), but most likely will be sold in a sell-process with just a few interested parties and negotiations around the sale price. This latter process is an advantage to bidders and likely results in lower prices. Officer’s (2007) data supports the liquidity argument as he finds a majority of deals financed with cash. In addition, his data on subsidiary targets further expresses the parent’s need for cash. Subsidiaries, on average, have only a small fraction of total assets in comparison to their parents, but the proceeds from the subsidiary sale contribute significantly to the parent’s cash level (over 100% of pre-bid parent cash level). Additional evidence comes from a ‘fire sales’ theory from Shleifer and Vishny (1992). These authors reason that the need for liquidity is probably highest when the firm is in financial distress. However, financial distress is often caused by economic downturns or industry-wide problems. This makes it less likely that industry peers (potentially the highest bidders) are bidding for the assets as they are distressed themselves. Officer (2007) does find evidence for this theory in the form of higher discounts when debt markets are tighter and when the price for alternative sources of liquidity are higher. Similar results are found by Schlingemann et al. (2002) who find that liquidity plays an important role in assessing if and which assets are sold in divestitures.

theory and confirms Hansen’s (1987) information asymmetry theory, in which he states that targets are acquired with stock when information is not complete. Because acquiring firms are afraid of overpaying the private target, part of the risk is transferred to the target’s shareholders. This information asymmetry results from non-transparency around private firms and seems to be present not only relatively (subject to the method of payment) but also in absolute terms. Chang (1998) supports this theory from a perfect competition perspective, as he argues that the NPV of an acquisition project and the announcement return to the bidder are zero. If competition is not perfect due to scarcity of information, a price discount for private targets is likely. Although there is not much quantative research on the absolute discount for information asymmetry, Officer (2007) finds proof for the information asymmetry theorem as a partial explanation to the pricing discount. He suggests that around 25% of the private target discount reflects the lack of public information. More evidence comes from Capron and Shen (2007) who find that bidder firms less often acquire private targets when information asymmetry is high.

Whereas illiquidity and information asymmetry seem to be the most likely explanations of the private target discount, some theories explain an additional post-acquisition performance when the target’s owner is paid with shares. Shleifer and Vishny (1986) argue that large minority shareholders can increase shareholder value if they monitor management. Since shareholders of private firms in general are less dispersed than owners of public firms, an acquisition financed with stock can create a monitoring blockholder in the public entity, causing a positive market reaction (Chang, 1998; Fuller et al., 2002).

2.2 Private equity

Private equity (PE) has been a constant field of interest ever since the LBO-wave of the 80s. An interesting characteristic of private equity firms is that they are private and often disclose little information about firms they manage. This makes it hard for researchers to study PE-backed entities and their performance and explains the limited number of studies that have been carried out in this field. Although private equity backed targets are part of a larger group of private firms, it can be argued that PE backed assets differ in a number of ways from its non-PE-backed counterparts. We argue that this could result in different announcement returns. To better understand the dynamics, we first describe private equity in depth.

these companies are collected among private investors who require high rates of return. These returns are typically realised using three mechanisms that are generally unusual to non-PE-backed private firms. First, the financial sponsor acquires a target already knowing that it will try to exit the company typically within three to five years. Second, management’s interests are often aligned with that of the PE player, as they are often offered equity participation. Third, the acquisition is financed with large amounts of debt, making the repayment of debt among the most important goals of the company. In order to realise the high returns, private equity targets often fit a similar profile that suits the concept best. First, the target has high and stable cash flows which are the basis for repaying debt. Second, the company holds attractive assets which can serve as collateral to obtain higher and better levels of debt. Third, there is room for cost cuts, for example in the form of operational costs, capital expenditure or working capital. Fourth, current management is competent and well-experienced. Fifth, an unconsolidated market can be the basis for a buy-and-build strategy. Although these characteristics are not per definition a must, PE-targets often suffice most of these criteria. Some other characteristics that are not per definition target-related, but what PE-companies are looking for include liquid debt and capital markets, positive market outlook and attractive entry (low) and exit (high) valuations.

Another characteristic of PE-companies is that they often have large funds which they invest in companies. These funds typically have duration of ten years, after which the initial investment plus returns have to be paid out. So except for the end of these ten years, PE companies do not necessarily need cash as the large funds often suffice these needs. This implicates that, from a liquidity perspective, PE companies are less likely forced to sell an unlisted target than are non-PE owners.

non-PE-backed counterparts on all fronts. This might be a proxy of superior acquirer performance when the target is PE backed.

Private equity firms always have to look for possible investors in order to use the proceeds for capital investments. As the number of private equity players is increasing, the battle for funds is becoming more competitive. If investors have to make a choice which private equity firms to fund, firm reputation can be an important factor to make this decision. Nahata (2008) finds that more reputable venture capital (VC) firms have easier access to investor funds, that their offers for targets are more likely to be accepted and that they acquire target equity at a discount. This suggests that a good reputation is beneficial at the entry stage. In addition, the author finds that more reputable VC firms exit through IPOs and acquisitions more often then less reputable firms. Demiroglu and James (2009) find better lending terms for reputable PE firms over firms with lower reputation. Cummings (2008) adds that high reputation can certify the quality of a VC asset as reputation is important to firms in the PE and VC field.

2.3 Venture Capital

Closely related to private equity is venture capital (VC). In fact, venture capital is a submarket in the larger private equity market. It is the investment of equity in typically less mature firms with undeveloped products. There are several similarities between PE and VC companies. They both invest in other companies and closely monitor management. Both types are looking for an exit in due course and typically have multiple investments at a time. However, differences are also numerous. The average PE target company has a mature business with stable cash flows, whereas VC target companies in general are start-up firms with high growth potential and unstable future cash flows. For the latter, it is often the intellectual property and human capital that is of most interest to VC companies. Furthermore, PE-backed companies are highly leveraged, whereas VC-backed companies’ underlying business is often too unstable to use debt financing. The risk of both types of investments is also different. PE-backed companies face the highest risk from the large debt levels and the corresponding default risk. VC companies face the highest risk from the uncertainty of its future cash flows.

Xuan (2008) compare three mechanisms that might reduce this information asymmetry. First, VC companies have a reputation to hold up since start-up companies are their core products. Hence, selling assets of which the underlying value is worth less than suggested cause damage in the long run. Second, the authors suggest that VC companies are able to bridge the asymmetric information. This is done through personal and professional relationships of VC companies with both acquirer and target. Third, the authors propose that the distance between acquirer and target is related to the degree of information asymmetry. Gompers and Xuan (2008) show that bridge building decreases information asymmetry and increases announcement returns.

In another study, Gompers and Xuan (2006) look at the role of venture capital (VC) firms in the takeover of private targets by public bidders. From a dataset of private firm acquisitions, the authors document a negative market reaction after the acquisition of a VC-backed target relative to the reaction after the acquisition of a private, non-VC-backed target. The authors suggest two possible explanations. First, the market could see VC firms as good negotiators, as they are experienced in exiting their investments. Second, firms whose value is based upon real options instead of tangible assets might suffer to a greater extent to adverse selection. This is then priced in the announcement returns.

2.4 Exits

As is common to private equity, funds look for an exit after a couple of years, often in the range of 3-5 years. This time scheme is typically used to execute the new business plan and optimally manage the company, without waiting too long for an exit, so it would become harder to make the requested annual returns. One of these options is an IPO, which gives the company a listing and the obligation to publicise annual financial reports. A downside of an IPO is the substantial costs related to bringing the company to the market.

and decrease information asymmetry, PE companies often provide potential buyers with vendor due diligence (VDD), an independent review of the targets operations and assets, provided by the seller.

Private equity is characterised by the timeline of investments, especially the exit that usually comes 3-5 years after the initial investment. However, since returns on investments are the most important thing to PE companies, the price at both entry and exit level (often expressed in entry multiple and exit multiple) should be optimal. PE companies will therefore make their entry and exit decisions partially based on the economic cycle. This implicates that PE companies invest at the beginning of the growth cycle and most often will try to exit the company in the same upturn. When the economy shrinks, PE companies are less keen on exiting their investment as they might suffer from low exit multiples. On the other hand, extending their PE involvement drives down the internal rate of return.

2.5 Hypotheses

Several studies document positive bidder announcement returns (Chang, 1998; Fuller et al., 2002) and lower acquisition multiples (Officer, 2007) for private targets and subsidiaries. Public targets seem to be more expensive which is expressed in negative or zero acquirer abnormal returns upon announcement. In this study, we discuss and explore bidder announcement returns of PE backed targets and compare these to bidder announcement returns of non-PE backed private targets. Although PE-backed companies are private by nature, it can be reasoned that certain characteristics cause bidder returns of PE-backed targets to differ from bidder returns of non-PE-backed private targets. This brings us to the first hypothesis:

H1: Bidder announcement returns of private equity-backed targets differ from bidder announcement returns of non-private equity-backed targets

counterparts is that private equity firms are more experienced in selling assets. We expect that their selling capability causes higher prices and therefore lower announcement returns. On the other hand, we expect the returns for PE share-paid deals to be higher than the non PE share-paid deals as a result of the minority blockholder theory. This theory holds that the market values the creation of such a minority blockholder. We add the expectation that the market reacts more positively towards the creation of a PE blockholder than that of a non PE blockholder. These theories lead to the expectation that bidder returns differ and that several theories can cause either higher or lower returns. We elaborate on these and other theories and create some more hypotheses.

Many studies name information asymmetry in relation to higher announcement returns for private targets. Officer (2007) explains and quantifies this existence of information asymmetry. The absence of information is most often negatively interpreted and causes lower transaction prices. We argue that PE firms are well aware of this fact and will try to close the information gap when exiting their private assets. This leads to the following hypothesis:

H2: As a result of information asymmetry, non-PE-backed targets sell at a discount relative to PE-backed targets causing lower CARs for bidders of PE targets than for bidders of non PE targets

An often studied deal characteristic is the method of payment. An important implication of the method of payment choice is post-acquisition ownership. If a bidder finances the deal with shares, the owner of the target becomes a shareholder in the bidders company. When the relative deal size and the portion paid with shares is substantial, the target owners can become a significant minority shareholder. With a substantial minority share often comes certain privileges such as requesting shareholder’s meetings and demanding to take a vote on certain decisions. Therefore, such a ‘block’ of shares can be valuable, also to other minority shareholders if at least interests are aligned blockholder (Shleifer and Vishny, 1986). Naturally, the willingness and ability to take up this role of minority blockholder and monitor management depends on the type of shareholder. We argue that a PE firm incorporating this role is much higher appreciated at announcement than a non PE firm:

Another characteristic that is often mentioned in relation to higher bidder returns of private targets is the liquidity discount. Officer (2007) finds that private targets are sold at a discount compared to public targets which he dedicates partly to the lack of access to liquidity pools. If assets owned by PE firms would suffer from a similar lack of liquidity, a similar discount can be expected. However, we argue that these assets do not directly need liquidity or that PE firms have sufficient cash pools to fulfil these needs. We therefore hypothesise that PE targets do not suffer from a liquidity discount causing lower bidder announcement returns:

H4: As a result of a liquidity discount, non-PE-backed targets sell at a discount relative to PE-backed targets causing lower CARs for bidders of PE targets than for bidders of non PE targets

An important characteristic of PE firms is their experience and ability in leading a sale process. In general PE players know how to set up an auction-type of sale process and provide all interested buyers with a vendor due diligence report. As buying and selling companies is their core business, PE players are experienced in negotiating profitable terms. Gompers and Xuan (2006) find that VC firms are better at negotiating high prices. In order to gather similar capabilities, a non PE firm can hire M&A specialists to give advise during the sale process on legal, financial and accounting matters. Although these advisers are expensive, they are involved in (almost) all large transactions. However, when deal size is small, extensive advise is often too expensive. When we compare PE players as a selling party with families and strategic parent companies, we expect to see a gap in selling capability, which is mediated by professional legal and financial advisers. As smaller parties less often hire professional advisers, we expect to see this gap increased as deal size becomes smaller in both absolute and relative sense:

H5: As a result of PE selling capability, bidder announcement returns for deals involving non-PE targets are relatively high compared to returns involving PE targets when (relative) deal size is small

Apart from better selling capabilities, some PE firms might benefit from firm reputation. As stated by Nahata (2008) and Demiroglu and James (2009), PE firms with a sound reputation have better entry characteristics than less reputable firms. Since both studies controlled for several characteristics (e.g. portfolio company quality), it can be argued that a more reputable firm has some advantages that lead to a higher price. In order to test this, we will look at PE firms and determine a group of high-reputable PE firms. For simplicity, we assume that the PE world is a survivor’s business in which the largest funds are on average more reputable than their smaller counterparts. If the more reputable firms are able to negotiate better exit terms, bidder announcement returns would be lower for firms acquiring targets from more reputable PE firms:

H6: Bidder announcement returns are lower for targets acquired from more reputable firms than for targets acquired from less reputable firms

Although this hypothesis might look somewhat similar to the selling capability hypothesis (H5), the difference between the hypotheses lies in the comparison of PE sellers and non PE sellers (H5) and the comparison of high reputable PE sellers and less reputable PE sellers (H6).

Another interesting issue is the length of the PE investment. In general, PE firms aim to exit their investment within 2,5 to 5 years. However, many cases show investments that are exited well before three years or PE firms that stay committed (much) longer than five years. There are numerous reasons for such deviations but the decision will most often be taken from a maximising return perspective. An exit outside the ordinary time frame could cause a different market return. For instance, early exits could indicate that a PE firm is taking its losses due to a bad investment but it could also be a quick and good return while other investment opportunities are waiting for PE investments. We hypothesise that the market will react differently to exits outside the usual time frame of 30 to 60 months:

H7: Bidder announcement returns for PE targets that were exited by the PE firm within the usual time frame of 30 to 60 months differ from bidder announcement returns of PE targets that were exited outside that time frame

3. Data

We use acquisition data from the database Zephyr. Two samples are created with PE backed target deals and non-PE backed target deals. Zephyr provides a function ‘exit’ which includes all deals were a PE firm exits an asset. We use this function to gather our first sample with PE backed targets. In order to gather the second sample with non-PE backed private targets, Zephyr provides the function to include only private targets. With these two samples, we start the rest of the selection process with removing all non-public acquirers. We include a restriction that the deal has to be announced between 2001 and 2008 and requires a minimum deal value of EUR 10m. The next step is to reduce the set of deals by removing all deals where the target is from outside the EU-15 and exclude deals with targets from the UK, ending up with 14 countries. For this dataset, announcement returns will be measured for a five day time fence (-2,+2) around the announcement. We exclude all deals that do not fulfil the following requirements:

• The bidder is a listed company at the time of deal announcement • The acquisition is announced between 01/01/2001 and 31/12/2008 • The minimum deal value is EUR 10m

• The target firm is based in one of the EU-152

countries excluding the UK • Acquisitions with a relative deal value lower than 1% are excluded • Acquisitions where the acquirer is a financial institutions or REIT3

are excluded • Acquisitions where the acquirer had an initial stake above 50% are excluded • Acquisitions where the acquirer ended up with a stake below 50% are excluded • Acquisitions where the acquirer bought less than a 50% stake are excluded • Acquisitions classified as reverse takeovers are excluded

• Acquisitions classified as venture capital deals are excluded

We exclude UK deals for two reasons. First, with London being the financial centre of Europe and the fact that the majority of PE companies reside in London, we expect that UK deals would disproportionately dominate the sample, especially for PE deals. Second, as there has been quite some research to private equity deals in Anglo-American cultures, we want to see whether certain effects exist in continental European markets. The thresholds of a minimum value and a

2

Hence, 14 target countries are included: Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, Netherlands, Portugal, Spain, Sweden.

3

relative deal value are included to make sure all deals have a relevant size both in absolute and relative terms. Although arbitrary, this approach is in line with several studies to announcement returns (Chang, 1998; Conn, Cosh, Guest and Hughes, 2005; Officer, 2007; Masulis, Wang and Xie, 2007). We exclude acquisitions by financial institutions and REITs as is common in similar studies. This includes secondary buyouts and other deals where the buyer is a private equity or venture capital firm. Also, deals where the acquirer bought less than a 50% stake are excluded. Inherently, deals where the acquirer ended up with a stake below 50% or deals where the acquirer had already a 50% stake are removed from the sample. We do this to make sure that a change of control is established and that a significant part of the target is sold in the deal. Besides, we would not like to see the results moderated by market reactions to insignificant stakes bought or sold. We exclude all reverse takeovers by checking the deal description of such a statement. Also, we check all deals with a relative deal size (deal size/acquirer size) higher than 100% in Mergermarket and look for any statements of a reverse takeover. Whereas we exclude deals with a small relative deal size (below 1%), we argue that deals with a relative deal size above 100% (target is larger than acquirer) have For our set of PE target deals, we take one extra step. We want to make sure all PE target deals are buyout deals and no venture capital (VC) deals. A typical feature of VC is that the VC firm has a stake from a very early stage, often based on real-options instead of relatively sure cash flows. We therefore set a barrier and manually look up all PE deals in Mergermarket. If we can not find proof of the actual ‘buy-in’, we exclude the deal. In addition, if Mergermarket classifies it as a VC deal rather than a PE deal, we exclude the deal.

The total sample includes 1,124 deals with targets across the 14 EU countries, of which 100 PE deals and 1,024 non-PE deals. Table I shows the target countries for the samples. The sub samples’ distribution is roughly in line with the whole sample, although there are relatively more PE deals in Finland, Germany and Sweden and less PE deals in Belgium, Italy and Spain.

in mean values is in 2007 could indicate lower deal values in 2007 with a few extremes on the upside.

For non PE deals, we see the highest peak during 2001. Although the sample contains only 21 deals and we should be careful with inferences, these high valued deals are probably result of high valuations during the internet bubble that burst during 2001. After 2001, the non-PE deal values dramatically decreased for a couple of years where after they slowly rose to smaller peaks in 2005 and 2007. For both sub samples, the sample size increased until 2007 with a decrease in number of deals for 2008. This is not surprising as the credit crunch slowed down M&A activity. The low number of deals in 2001 and to a lesser extent 2002 is probably the result of the databases Zephyr and Mergermarket, which might suffer from incomplete data in this period, causing deals to be excluded or deals that were not recorded at all. This is expected as Bergström et al. (2006) do find high market activity in 2001 and 2002. Although the relative small size of the sample of PE deals could cause a problem when drawing inferences, the distribution does not necessarily so.

Table I

Sample Target Country

Sample of private targets acquired by listed bidders announced between 2001 and 2008. All targets are located in the EU-15, except for UK targets, which are excluded from the sample. Bidders are firms from all over the world. The sample is divided in targets exited by private equity (PE) firms and firms previously owned by non-PE firms. Both sub samples show roughly the same division for target country deals, although there are relatively much Finnish, German and Swedish PE deals and relatively much Belgian, Italian and Spanish non-PE deals.

Table II

Sample Deal Size per Year

Mean and median deal size per year for the total sample and for both sub samples. PE deals are on average larger than their non-PE counterparts, although the yearly PE sub samples suffer from a low number of deals. The distribution in number of deals over the years is more or less the same for both sub samples. The deal values move more or less in opposite ways in the first few years, but move in the same direction afterwards.

Mean Median N Mean Median N Mean Median N

2001 602 105 22 84 84 1 626 125 21 2002 194 80 99 123 127 3 197 80 96 2003 140 45 109 284 224 5 133 39 104 2004 175 48 132 313 257 9 165 46 123 2005 289 55 190 303 140 20 287 50 170 2006 205 45 215 242 84 18 202 35 197 2007 291 40 224 476 131 27 266 33 197 2008 223 42 133 262 118 17 218 36 116 Total 236 49 1.124 324 118 100 227 44 1.024 All PE non PE Table III

Sample Deal Size per Industry

Number of deals per industry and type of deal for both target and bidder firms. Classification based on NAICS 2007 sector guidelines4. Additional the number of targets that are in the same industry as the acquirer for both type of deals. The results show no major differences in the distribution per industry. For both PE and non PE deals, most targets and bidders are in manufacturing, information and wholesale- and retail trade. Also, the percentage of deals in the same industry are roughly the same for PE and non PE deals.

Target Bidder Same

Industry Target Bidder

Same Industry Agriculture, Forestry, Fishing & Hunting 0 0 0 4 6 2 Mining, Utilities & Construction 6 4 3 97 121 79

Manufacturing 54 61 50 419 499 425

Wholesale & Retail Trade 13 6 5 104 66 47

Transportation & Warehousing 3 3 3 28 27 25

Information 10 12 8 177 170 123

Finance & Insurance 1 0 0 15 0 0

Real Estate, Rental & Leasing 1 1 1 16 15 8

Professional, Scientific & Technical Services 6 8 2 85 63 30 Management of Companies & Enterprises 1 0 0 0 0 0 Administrative, Support & Waste Management 1 1 1 33 25 14

Education Services and Health Care 3 3 3 18 15 18

Arts, Entertainment & Recreation 0 0 0 6 7 2

Accomodation & Food Services 1 1 1 15 10 10

Other Services 0 0 0 2 0 0

Public Administration 0 0 0 5 0 0

Total 100 100 77 1024 1024 783

Table III shows the distribution of PE and non-PE deals per industry, following a NAICS 2007 classification.5 As can be seen, most targets and bidders are in the manufacturing, information and wholesale & retail trade industry. The relative number of deals are more or less the same for PE and non-PE deals, although only a few PE targets and bidders are active in the mining, utilities & construction industry, which can be understand from the perspective of PE houses and the type of deals they are interested in. Furthermore, we find that the percentage of deals with target and bidder in the same industry is almost identical for PE and non PE deals.

4. Methodology

This research studies the bidder announcement returns of PE backed and non PE backed targets. The bidders in this research are strategic buyers (i.e. no private equity firms nor financial institutions) that are listed on a European exchange. This means that we can observe the share price of the bidder, especially around an announcement of M&A activity. The announcement of a merger or acquisition (M&A) is often the first signal to the market of such intentions. Until that moment, the market is not supposed to know about these intentions. Hence, when an announcement is made that the bidder has plans to acquire another company, the market translates this strategic decision in a share price movement. This is referred to as the announcement return. However, part of this return can be allocated to general market movements. Therefore, we measure the individual bidder’s return compared to the market return, which is called ‘abnormal return’ (or ‘cumulative abnormal return’ (CAR) because we measure the returns over a few days). Another important note should be made with regard to the dynamic between share price returns and target prices. In this study, targets are the assets that are actually being acquired by strategic buyers. If the seller of the asset is a PE firm, we say it is a ‘PE backed target’. If the seller of the asset is not a PE firm (instead, the seller is a family or another private company) we say it is a ‘non PE backed target’. As said, when a target is sold at a certain asset price - which is often referred to as ‘deal value’- the price and its underlying value determine whether the market reaction is positive or negative. This means that the higher a price (higher deal value), the lower the market return is and vice versa. A high announcement return might therefore be caused by a lower price, in other words the target was acquired at a ‘discount’. To sum up, when we say ‘return’, we are referring to the market reaction given by the share price. When we use the words ‘discount’ or ‘premium’, we refer to the price paid for the target. Apart from studying announcement returns ‘as is’, we also compare the returns to the relative prices paid. We do this

5

by using multiples. Multiples indicate the relative price of an asset, such as the deal value to sales or deal value to EBITDA. Although we study multiples only to a very little extent, the dynamic between announcement returns and multiples (like returns and prices) is important to be understood.

Extensive research has been conducted on the choice which method to use for measuring bidder announcement returns (Brown and Warner, 1980; MacKinlay, 1997). we use a standard event study methodology in which the cumulative abnormal return (CAR) is calculated for each acquirer in a five day time fence (-2, +2) around the announcement. The most important methods for such event studies are the market adjusted return and the market- and risk adjusted return. Although often employed, we do not use the market- and risk adjusted model. This model obtains the expected return for individual stocks while accounting for the systematic risk for individual stocks and the market return, using an ordinary least squares (OLS) regression. This expected return is calculated during an estimation period prior to announcement (often 240 days). There are two reasons why we do not use this method. First, our dataset consists of several multiple acquirers, which would lead in some cases to announcement periods that appear in estimation windows for other acquisitions. Obviously, this could lead to biased results if it occurs too often. Second, Brown and Warner (1980) find that results are not significantly improved when using risk-adjusted returns in the case of short-window event studies. The other widely used model is the market adjusted return method, in which the CAR is calculated for each acquirer in a five day time fence (-2, +2) around the announcement, adjusting for market fluctuations. Since this method does not use an estimation window to calculate the normal return, the chance of including returns from abnormal events is greatly diminished. These arguments are in line with studies of Fuller et al. (2002), Dong, Hirshleifer, Richardson and Teoh (2006) and Conn et al. (2005). Like aforementioned authors, we use this market adjusted return method and calculate the CAR using the following formula:

∑

where CARi is the cumulative abnormal return of acquirer i for the five days surrounding the acquisition announcement (-2, +2), and ARitgiven by:

R mt Rit

where ARitis the abnormal return of acquirer i at time t, calculated by subtracting the return of

the market index (Rmt) of the return of the individual acquirer (Rit). The market index used is the

value weighted market index return for that specific acquirer. After we calculated the individual CARs, we first have to take care of potential outliers. Like Fernandes and Ferreira (2008), we winsorize our top and bottom 1% for each subsample. This means that we set an upper (lower) limit at the top (bottom) 1%. All observations above (below) this limit are replaced by the CAR of the upper (lower) limit. Next, we can calculate the average CARs for our subsamples, given by:

∑

where CAR is the average cumulative abnormal return for all firms n in that sample. Next, in

order to see whether the CAR significantly differs from zero – which indicates an abnormal bidder announcement return relative to the market index – the t-statistic is tested following the formula for an independent one-sample t-test:

n CAR t /

σ

where t is the t-statistic, n is the number of deals in the sample and σ is the sample standard deviation. This test gives guidance on the significance of the potential abnormal returns relative to the market and assumes a normal distribution in each sample. In order to determine whether the bidder CARs of PE-backed targets significantly differ from the bidder CARs of non-PE-backed targets, we conduct a t-test assuming unequal sample sizes and variances. This test is given by:

PE non PE CAR CAR PE non PE S CAR CAR t − − − − = (5)

where CAR_PE is the average cumulative abnormal return for the sample of bidders of PE-backed

PE non PE non PE PE CAR CAR n n S PE non PE − − − − = + 2 2

_σ

σ

where σ_PE2 and σ_non2 _−PE are the variances of respectively the sample with bidders of PE-backed targets and the sample with bidders of non-PE-backed targets and n_PE and n_non−PE are the sample sizes of both groups.

It is possible that the market likes or dislikes the acquisition of a target from a PE seller for other reasons than the relative price paid. Therefore, we examine multiples in addition to the CAR analyses. We study deal value multiples over sales, EBITDA, EBIT and assets and compare them for PE deals and non-PE deals.

Table IV

Liquidity Index per Year and Country

The liquidity index is calculated by dividing the total deal value of announced deals by the total book value of companies in the market and generate an indicator variable to ‘1’ if the index is above its time-series median and ‘0’ otherwise. 2001 2002 2003 2004 2005 2006 2007 2008 Austria 1 0 0 1 0 1 1 0 Belgium 0 0 0 0 1 1 1 1 Denmark 0 1 0 1 1 0 0 1 Finland 1 1 0 0 1 0 1 0 France 0 0 0 1 0 1 1 1 Germany 1 1 0 0 0 1 1 0 Greece 0 0 0 0 1 1 1 1 Ireland 1 1 0 0 0 1 1 0 Italy 0 1 0 0 1 1 1 0 Luxembourg 0 0 0 1 0 1 1 1 Netherlands 0 0 0 0 1 1 1 1 Portugal 0 0 0 0 1 1 1 1 Spain 0 0 0 0 1 1 1 1 Sweden 1 0 0 0 0 1 1 1

public targets. For non-PE deals, we expect information asymmetry to be larger, hence our expectation of a higher abnormal return for share-paid deals compared to cash deals. We argue that if the results show such a pattern, information asymmetry is likely to cause a discount on non-PE targets and a higher return for its bidders.

Another impact of stock payments is the creation of a minority blockholder given that the relative deal size and portion paid with shares is substantial. This can be favourable for other investors, who may profit from the monitoring role of a minority. Research to corporate governance mechanisms showed that a strong minority blockholder is positively received by minority shareholders. This implicates that PE firms could become a minority blockholder in case they sell their assets for shares. Although this might not be directly influential on the price, we expect a higher market reaction when PE firms are to fulfil this role of monitoring blockholder, especially when relative deal size is high. To assess the fourth hypothesis, we test if PE deals paid with shares cause relatively high CARs compared to PE deals paid with cash. As for non-PE deals, we do not expect to find this relationship to the same extent because the market will appreciate their blockholding role to a lesser extent. We acknowledge that the tests of the second and third hypothesis are basically the same. It is therefore not unlikely that at least one of the hypotheses is rejected. On the other hand, if we include relative deal size in our analysis, we might be able to show a pattern that supports both theories.

To test the fourth hypothesis, whether non-backed targets sell at a discount relative to PE-backed targets, we create a liquidity index based on the ideas of Schlingemann, Stulz and Walkling (2002) and Officer (2007). Table IV displays the descriptive statistics for this index. For each year and country, we calculate the total sum of deal values for all deals announced in that year and country. Then, we divide this amount by the total book value of companies in the market for that same year and country. Next, like Officer (2007) we take the time-series median and give a ‘1’ to each year that the ratio exceeds the median, indicating a liquid market and give a ‘0’ otherwise. Unlike aforementioned studies, we do not calculate the index per industry but per target country. We choose a different method because our study focuses on more than one country for target companies. Although it is bidder announcement return that is measured, we calculate the index based on target country, as the hypothesis holds that it are the target owners who face the potential liquidity problem, hence it is their market liquidity that is of interest. Next, we divide both groups of bidders of PE-backed targets and of bidders of non-PE-backed target in subgroups, where we differentiate for deals in liquid periods (‘1’) and illiquid periods (‘0’), so we

step is to measure the difference in both groups, that is, we calculate t as in (5) by subtracting the

CARof bidders of non-backed targets in liquid markets from the CAR of bidders of PE-backed targets in liquid markets and do the same for both groups in illiquid markets. If these two values differ significantly, the hypothesis of a greater discount for private targets in times of illiquid markets is supported by the data in this test.

We test the selling capability hypothesis (H5) using subsamples based on relative and absolute deal size. We expect that PE firms have an advantage when selling assets because they are more experienced in the sale process. Non-PE owners often mediate this advantage by hiring professionals, such as investment bankers and lawyers. However, we expect that when deals are smaller, the role of such professionals becomes smaller, hence the PE advantage becomes larger. We argue that the pattern between PE deal bidder returns and non-PE deal bidder returns is relatively stable across the absolute- and relative deal size sub samples. That is, returns increase or decrease in a similar pattern when absolute and relative deal size become smaller or larger. Due to PE’s selling capability however, we expect to see a deviation from this normal pattern when absolute- and relative deal size are small and when the role of aforementioned professionals is diminished at most. For our relative deal size measure, we use relative deal size higher and lower than 10%. In absolute terms, we divide deals in three categories, EUR 10m-100m, 100m-500m and higher than 100m-500m. If we find an abnormal high CAR for non-PE targets compared to PE targets when relative and/or absolute deal value are small, we might conclude that our data supports the selling capability theory.

and 52bn. We compare the bidder returns of targets acquired from these reputable firms with the bidder returns of targets acquired from less reputable firms.

We test whether the length of PE involvement influences the bidder return in the seventh hypothesis. As mentioned before, we looked up the buy-in moment for PE deals in Mergermarket. We use this data in combination with the exit moment to calculate the number of months of PE involvement. We determine three groups to analyse this data. First, the group of PE deals with the usual length of involvement, 30 to 60 months. Next, we create a group of deals with PE involvement less than 30 months and a group with involvement over 60 months. We then compare the differences in returns.

After we perform the univariate analyses, we test the impact of all variables using a multivariate regression analysis. We include explanatory variables to test the hypotheses and control variables to control for deal specific characteristics that otherwise might influence our results. We conduct tests for three samples. First we examine the PE sample, then the non PE sample and finally we test the whole sample. We choose this structure to optimally capture the dynamics between both subgroups (PE and non-PE) and in the whole sample. For each sample we first test the explanatory variables individually and then jointly while controlling for several variables. The main regression for the PE sample is given by:

= PE CAR ε β β β β β β β β β α + + + + + + + + + + VARIABLES CONTROL i CASH US UK EXIT EARLY CASH HIGHREP DOMESTIC HIGHREP HIGHREP LOWDS LOWRDS CASH LIQPERIOD LIQPERIOD * / 8 7 * 6 * 5 4 * 3 * 2 1 (7)

variable for US/UK acquirers, a SAME INDUSTRY dummy for acquirers and targets that are in the same industry, LN(SIZE) which is the natural logarithm of the acquirer size given by the acquirers market capitalisation one week before announcement and RELSIZE which is the relative deal size calculated by deal size divided by the acquirer market capitalisation one week before announcement. We also control for YEAR and COUNTRY but do not tabulate the results in the tables. In addition to COUNTRY, we include country-specific variables, such as SHAREHOLDER RIGHTS, CREDITOR RIGHTS and MARKET BASED countries. Shareholder rights are based on the revised anti-director index by Djankov et al. (2006). Creditor rights are a measure of creditor protection against defaulting debtors (Djankov, McLiesh and Shleifer, 2008). Market based is a dummy for either countries that are market based or bank based. The main regressions of the non PE sample and the whole sample are:

= −PE non CAR ε β β β β β α + + + + + + VARIABLES CONTROL CASH US UK LOWDS LOWRDS CASH LIQPERIOD LIQPERIOD 5 * / 4 * 3 * 2 1 (8) = All CAR ε β β β β β β β β β β β β β β α + + + + + + + + + + + + + + + IABLES CONTROLVAR i CASH US UK EARLYEXIT CASH HIGHREP DOMESTIC HIGHREP HIGHREP NONPE LOWDS LOWRDS LOWDS LOWRDS NONPE LIQPERIOD CASH LIQPERIOD LIQPERIOD SHARES PE CASH PE PE * / 13 12 * 11 * 10 9 * * 8 * 7 * 6 * 5 4 * 3 * 2 1 (9)

With extra explanatory variables PE as a dummy variable for all PE deals, PE*CASH as a dummy for PE deals paid with cash and PE*SHARES as a dummy for PE deals where shares were involved in method of payment. Other additional explanatory variables are LIQPERIOD*NONPE and LOWRDS*LOWDS*NONPE, with are dummies for non PE deals in liquid periods and with low relative and absolute deal size respectively. We use the same control variables as in other regressions.

5. Results

5.1 Univariate analysis

positive return on the five day interval (-2, 2) and both sub samples show a positive return with a mean of 2,18% and 1,76% for bidders of PE and non-PE targets respectively. Bidders for PE-backed targets show a greater announcement return than bidders for non-PE-PE-backed targets although the difference is not significant.

Panel B shows the samples classified by method of payment. Transactions are classified as ‘cash’, ‘shares involved’ and ‘other’. Cash deals include only deals financed with 100% cash. Deals with shares involved include all deals that have a stock-component. ‘Other’ includes earn-outs and deferred payments or a combination of these with cash. Except for PE deals paid with cash, all returns significantly differ from zero. Contrary to existing literature on private targets, our overall sample shows a higher return for cash deals than for share deals. Even more interesting are the returns per sub sample. For PE deals, returns are lowest for deals paid with cash followed by deals paid with shares involved. This is counterintuitive as we reasoned that PE firms are able to diminish the information asymmetry, hence cash-paid deals should be higher valued. So far, we do not find support for the hypothesis that PE firms manage to decrease information asymmetry. Yet, another explanation might be relevant. Several studies (e.g. Fuller et al., 2002) argue that private firms are by definition closely held, which means that a transaction paid with shares creates a minority blockholder in the acquired company, conditional on a substantial relative deal size. Although this minority blockholder could be both the PE firm as well as the owner of a non-PE target, markets might react better to the creation of a PE firm as a minority blockholder because they relate PE firms to more industry expertise. In other words, announcement returns could be higher when PE firms become an important minority shareholder than when ‘normal’ owners become a minority shareholder. Our data does support this explanation as share-paid deals are higher valued when the owner is a PE company, both in absolute terms as relative to cash-paid deals. However, if we look at the last column of panel B, we see that the different bidder returns for share paid deals are not statistically significant. In addition, our sub sample of PE targets paid with shares is small and therefore probably not sufficient to draw unconditional conclusions. Another remarkable finding is the high return when deal-financing includes a deferred payment or an earn-out. It is possible that the market values this extended payment as a form of insurance if the asset does not turn out to be as valuable as expected.

return for cash transactions than for shares- and other transactions. This result differs from existing theory (e.g. Hansen, 1987) and empirical research (e.g. Fuller et al., 2002). To sum up, the data does not support the hypothesis that the information asymmetry is higher for non-PE deals. It is possible that PE players do know how to close the information gap, but that the market does not see this as they are not as involved as managers from bidding companies. Unfortunately, testing this theory is not within the scope of this research. Another possible explanation could be in the fact that most literature study the US and UK market. Although this research aims to study European targets, still a lot of acquirers are listed in the US or the UK. If US and UK markets react different than other European markets, than we might be able to find such results in this sample. We will later address this issue in depth.

Table V

Cumulative Abnormal Returns for PE and Non-PE Deals

Cumulative abnormal returns for bidders of European private equity (PE) backed and non-private equity backed targets in the period 2001 – 2008 with a minimum deal value of EUR 10m. Cumulative abnormal returns CARi are calculated

using the bidder announcement return Rit during the (-2, +2) period and subtract the country-specific market return Rit:

∑

All transactions are executed by listed buyers and include targets who reside in one of the EU-15 countries, with an exception for UK targets, which are excluded from the sample. Panel A gives abnormal returns for all acquisitions in the total sample and sub samples of PE deals and non-PE deals. The column on the right side gives the differences between the averages of PE deals and non-PE deals. Panel B distinguish between method of payment: cash, shares involved and other. Panel C distinguishes domestic versus cross-border deals. In panel D, results are shown for relative deal size. The first number is the mean value, with the median between parentheses and number of transactions reported below the median. *, **, *** indicates whether values significantly differ from zero at the 10%, 5%, 1% level respectively.

All PE non PE PE - non PE

All acquisitions 1,80%*** 2,18%*** 1,76%*** 0,43% (0,91%) (0,76%) (0,91%) -(0,14%) 1124 100 1024 Cash 2,18%*** 0,82% 2,33%*** -1,51%* (1,20%) (0,11%) (1,28%) -(1,17%) 415 41 374 Shares involved 1,45%*** 2,23%** 1,40%*** 0,83% (0,71%) (0,74%) (0,71%) (0,03%) 553 36 517 Other 1,97%*** 4,55%*** 1,53%*** 3,02%** (0,85%) (1,48%) (0,65%) (0,83%) 156 23 133

Panel A: All deals

Table V - continued

All PE non PE PE - non PE

Domestic 2,12%*** 3,78%*** 2,04%*** 1,75% (1,09%) (1,81%) (1,09%) (0,72%) 362 18 344 Cross-border 1,64%*** 1,83%*** 1,62%*** 0,22% (0,80%) (0,65%) (0,85%) -(0,21%) 762 82 680 <5% 0,42%** -0,66% 0,50%** -1,16%* (0,16%) -(0,71%) (0,28%) -(1,00%) 382 26 356 5% - 10% 1,21%*** 0,93% 1,24%*** -0,31% (0,69%) -(0,11%) (0,77%) -(0,88%) 217 19 198 10 - 20% 2,29%*** 1,93%** 2,33%*** -0,40% (1,54%) (1,76%) (1,54%) (0,23%) 196 16 180 >20% 3,48%*** 4,80%*** 3,30%*** 1,50% (2,05%) (2,45%) (1,99%) (0,46%) 329 39 290

Panel C: Domestic / Cross-border

Panel D: Relative Deal Size

Panel C shows the different returns for targets acquired by bidders from the same country (domestic) and targets acquired by bidders from other countries (cross-border). Columns 1 to 3 show the means and medians for domestic and cross-border deals for the whole sample and for both sub samples. All means are higher than zero at the 1% significance level, although the smallest sub sample has only 18 deals. We also find that domestic deals yield a higher return than cross-border deals, which is in line with the literature (e.g. Conn et al., 2005). In terms of mean values, PE deals face higher returns than non-PE deals both with domestic and cross-border deals. This difference is not significant at a 10% level (domestic PE and non-PE deal differences are significant at the 15% level) and contradict for cross-border deals by the median value sign. Another remark should be made that, if the influence of US and UK acquirers effects the results, as suggested before, the differences between domestic and cross-border deal returns might also be explained by the fact that roughly half of the cross-border deals were done by US and UK acquirers.

announcement return increases with the relative deal size. The returns significantly differ from zero at the 5% level for the whole sample with a relative deal size under 5%. This significance increases with the relative deal size to 1% confidence levels. Remarkably, PE deals do not always yield higher returns than non-PE deals. Although acquisitions with relative deal size of over 20% do show higher returns, the mean for PE deals with a relative deal size under 5% causes a negative stock reaction. Even more, in terms of median values, PE deals below relative deal size of 10% are negative. However, it is difficult to immediately draw inferences from these results, as the sample sizes are very small. Nonetheless, a possible explanation could be again the theory of the minority blockholder, that is, shareholders value the creation of PE minority blockholders in the acquirer, which can only be formed it the relative deal size is substantial.

Panel A of table V showed that bidder returns do not significantly differ when the target is either PE or non-PE backed. It is possible that the market reaction, as measured with the CAR, is similar in terms of percentages, but that the relative price paid is higher for PE targets. In other words, a higher price in combination with a similar market return implicates that investors see more value for money in the acquisition of PE targets. Multiples are a useful tool in examining this relative price paid. Table VI shows the multiples of the acquisitions in the sample. We use four type of multiples, deal value (DV) over sales, DV over EBITDA, DV over EBIT and DV over assets in order to obtain as much information as possible. All input to the multiples calculations were obtained from Zephyr. Unfortunately, not all accounting numbers were available. Furthermore, multiples that had a negative value (e.g. when EBIT was negative) or multiples higher than 100.0x (e.g. when EBIT was just above 0) were left out. Although the 100.0x level is arbitrary, a cut had to be made as some deal values reached as far as 10,000 times EBIT. The missing values and outliers that were removed cause an incomplete sample if we are to compare the data with the announcement returns. Therefore, we should be very cautious in drawing inferences from this data.

would combine the higher multiples with the abnormal returns – which do not differ significantly between PE targets and non-PE targets – a conclusion could be that bidders pay a relatively high price for non-PE targets, but that the market somehow values both transactions equally. This implicates that the market values non-PE targets higher than PE targets. This is even more true when looked at the significance of the signs, which are all highly significant, except for the sales multiple. However, two important remarks must be made with respect to above conclusions. First, the sample size of our multiple analysis differs from the sample size of the CAR analysis. Second, as can be seen between the parentheses in the last column, the median values do not support this theory. As mentioned before, non-normality can lead to wrong inferences from the mean values, hence the use of median values. Only two median values show the same sign for the difference between PE and non-PE targets. Also, the median values differ only marginally compared to the differences in mean values. Given the potential pitfalls, we do not draw conclusions from this analysis, although we can state that the data does not confirm a higher price paid for PE targets.

An interesting fact from table V was that cash deals showed a higher abnormal return than noncash deals. This is not in line with existing literature and needs some further attention. As mentioned, this higher return was due to the return pattern from non-PE targets, as PE targets did show a higher return from cash-paid deals. As reasoned, this might come from differences in acquire types in this study (all over the world) and in existing literature (mostly US and UK acquirers).

Table VI

Deal Value Multiples for PE and Non-PE Deals

Prices paid for European private equity backed and non-private equity (PE) backed targets indicated by deal value multiples over target company figures. Four different multiples are used: Deal Value over Sales (1), EBITDA (2), EBIT (3) and Assets (4). The four columns depict multiples for all deals, PE deals, non PE deals and the difference between the latter two respectively. The numbers display the average and median (between parentheses) multiples and number of deals (below the median) . *, **, *** indicates whether values significantly differ from zero at the 10%, 5%, 1% level respectively.

All PE non PE PE - non PE

Deal Value / Sales 3.6x 3.0x 3.7x -0.7x

(1.2x) (1.2x) (1.2x) (0.0x)

663 78 585

Deal Value / EBITDA 15.0x 11.4x 15.5x -4.1x***

(10.3x) (10.8x) (10.2x) (0.6x)

517 69 448

Deal Value / EBIT 19.7x 15.2x 20.4x -5.2x***

(13.5x) (13.1x) (13.6x) -(0.5x)

477 65 412

Deal Value / Assets 3.8x 2.1x 4.0x -1.9x***

(1.6x) (1.4x) (1.7x) -(0.3x)

Table VII shows abnormal returns divided per acquirer type. We distinguish acquirers from Anglo-American countries (US and UK) and acquirers from the rest of the world (RoW). Panel A shows the results for Anglo-American acquirers. The first remarkable result is shown in the first column. Unlike all deals presented in the first column of panel B of table V, CARs of stock-financed deals are more or less in line with CARs of cash-stock-financed deals. ‘Other’ deals show a higher return. If we separate PE and non-PE deals and look at the returns regardless of method of payment, we see that returns for non-PE deals are higher than returns for PE deals. This sign is identical in terms of the median value, although the difference is not significant. If we zoom in on the source of this lower return for PE deals, we see that the lower return is attributable to cash-paid deals. Both in terms of mean and median values, cash-financed deals show a lower return when the target is PE backed than when the target is non-PE backed. This difference is highly significant and supports our hypothesis that bidder returns differ when the target is either PE backed or non-PE backed. Contrary to this finding, bidder returns for PE deals are higher when the deal involves share payment or ‘other’ payment methods such as deferred payment or earn-out. Although our sample of US and UK acquirers of PE targets suffers from a small sample size, the difference with non-PE targets is significant at the 10% level when payment includes a deferred payments or an earn-out. When we compare the different payment method-returns of non-PE targets, we see no spectacular differences. Stock-involved transactions display a slightly lower return, which is inconsistent with the existing literature, which showed higher returns for such deals compared to cash-deals. In short, panel A shows results for Anglo-American acquirers and partly supports our first hypothesis that bidder returns of PE-backed deals show a different return than non-PE deals. We do not find proof for reduced information asymmetry, as stock-involved payments increase PE target bidder returns compared to non-PE target bidder returns. If we look at stock-financed deals compared to cash-financed deals, we see that stock-financed deals show a higher return when the target is PE owned, in relative and absolute terms. This supports our minority blockholder hypothesis, although we must be careful drawing inferences due to the small sample size.