Operational Value Creation in Secondary Buyouts in the Nordics

Transcription

1 Operational Value Creation in Secondary Buyouts in the Nordics An Empirical Study of Private Equity Owned Portfolio Companies Master s Thesis M.Sc. Economics and Business Administration (Cand. Merc.) Finance and Strategic Management (FSM) Authors: Mads Rosenvang Rasmussen Nikolai Bårup Enggård Supervisor: Morten Holm, Department of Accounting and Auditing Characters (incl. spaces): Number of pages (incl. tables and figures): 94 Date of submission: September 17 th, 2014 Copenhagen Business School, September 2014

2 Abstract Purpose: The high frequency of secondary buyouts means that these transactions have spurred the interest of academic scholars. However, because secondary buyouts are a relatively new phenomenon the body of literature on this topic is still young and concentrated among a limited number of empirical contributions. With our master thesis we therefore intend to contribute to the current academic debate about the value creation proposition for secondary buyouts. Method: Based on a novel, hand- collected sample of 63 secondary buyouts in the Nordics, we compare the operational performance improvements of the respective portfolio companies during the first round buyout and the secondary buyout. We measure operational performance by breaking down ROIC into its components of profit margin and turnover rate of invested capital for which we develop four testable hypotheses. In that regard, we compare operational performance improvements directly for the same portfolio company over time and engage in an analysis at the deal level. Findings: After adjusting for industry specific effects we find statistically significant results, which show that SBOs offer higher working capital improvements than LBOs measured by working capital to revenue. With regards to revenue growth, profitability and fixed asset utilization we cannot reject that SBOs are similar to LBOs. Our findings therefore suggest that secondary buyouts are not inferior deals in terms of operational performance improvements, compared to first round buyouts. Originality: Our master thesis is of interest to both private equity professionals and academic scholars because we contribute to the limited understanding of operational value creation in secondary buyouts with empirical results and, secondly, because we focus only on secondary buyouts in the Nordics. Keywords: Private Equity, Leverage Buyouts, Secondary Buyouts, Operational Value Creation

3 Table of Contents 1. Introduction Scope and delimitations Definitions and abbreviations Thesis structure About Private Equity and SBOs Private equity as an asset class Private equity investments The limited partner The general partner Private equity funds and fundraising The acquisition phase Value creation Divestment phase The exit choice The European and Nordic private equity market at a glance Literature Review Financial Engineering Governance Engineering Operational Engineering Differences between LBOs and SBOs Theoretical drivers of SBOs Motivation for engaging in SBOs Empirical findings Empirical findings on operational performance in leveraged buyouts Empirical studies on SBOs Framework and study design Return on invested capital Study design Hypothesis development Data collection and applied method Sample selection Data collection

4 Specifying the relevant years of the observations Accounting data Measuring performance Computation of performance measures Variables Growth Profitability Working capital management Fixed asset utilization Adjusting for the industry effects Statistical Method Choice of significance test Hypotheses testing Results and analysis Descriptive Statistics Test results Total sample Subsamples The country of the portfolio company The country of the private equity firm The industry of the portfolio company The form of the financial statements used to collect data Realized deals and unrealized deals Discussion Prerequisites for the discussion Interpreting statistically insignificant results Fundamental reasoning Superior working capital management Potential industry differences Limitation and biases of the method applied Sample period Length of periods evaluated Consecutive buyouts

5 Use of the transaction year Accounting data used Measuring changes Statistical power Conclusion Future research and implications References Appendix 1 Nordic private equity activity in numbers of portfolio companies traded Appendix 2 Lists of company and deal specific information Appendix 3 Test assumptions and algorithms Appendix 4 - SPSS output

6 1. Introduction Since the 1980s leveraged buyouts have become an increasingly important type of investment with a record high amount of capital committed to private equity, relative to the overall stock market, prior to the global financial crisis (Kaplan and Strömberg, 2009). Despite the freeze down of the capital markets in 2008 and a following drop in the overall buyout activity 1, the long- term trend is that private equity has increased in importance as an investment opportunity. Jensen (1989) suggested, in the wake of the private equity, that organizations relying on leveraged buyouts would be superior as these firms would benefit from concentrated private ownership with active governance. Moreover, highly leveraged capital structures, co- investment by the management, and performance- based compensation in the private equity owned companies are expected to align the incentives of the owner and management. The competitive advantage of private equity firms, in terms of value creation, is commonly attributed to three levers: Financial-, governance-, and operational engineering with all of the three sources having an effect on the performance of the target company (Freelink and Volosovych, 2012). A large number of studies have examined the value creation potential of leveraged buyouts and have found evidence supporting improvements in efficiency and operational performance of portfolio companies owned by private equity (e.g. Bull, 1989; Kaplan, 1989a; Smith, 1990; Muscarella and Vetsuypens, 1990; Lichtenberg and Siegel, 1990; Gou et al., 2011). Most of the literature covering operational performance improvements in the buyout companies have covered the U.S, Europe or the UK while fewer studies have investigated smaller regions like the Nordic countries (e.g. Bergström et al. 2007). The increase in LBO activity over the past decades has created a sub- market for secondary LBOs where a portfolio company is sold from one private equity owner to another. After the dot- com bubble in 2000/01 traditional exit routes for private equity investments, such as trade sales to a strategic buyer or Initial Public Offerings (IPOs) in the stock markets, were not easily accessible (Sousa, 2010). The cold IPO markets and the need of private equity firms to exit their investments in order to return money to their investors is one of the main factors that initiated the wave of 1 Latest numbers from European Venture Capital Association (EVCA) shows that the buyout activity dropped drastically after The development of the buyout market will be covered later. 4

7 secondary buyouts, which peaked in 2007 prior to the global credit- and financial crisis (Sousa, 2010). Nikoskelainen and Wright (2007) show that leveraged buyout transactions exited through IPOs and trade sales are earning higher returns to the investments than secondary buyouts do. All things being equal this makes secondary buyouts a less attractive choice of exit for private equity firms. Despite the evidence provided in the literature secondary buyouts are immensely popular. Today one in seven deals in the United States involve a portfolio company that is sold on from one private equity firm to another. This is the highest frequency since 2006 and secondary buyouts thereby account for a third of buyout volume in the United States in 2014 according to data from Preqin (Goldfarb, 2014). In Europe, secondary buyouts is an equivalently popular choice of exit for private equity firms; between 2007 and 2013 the average share of secondary buyouts amounted to 26.2 % 2 of all exited buyouts over the period, measured by the value (EVCA, 2014). A paradox arises, however, from the high frequency of secondary buyouts since researchers and the business press continuously have expressed their skepticism about secondary buyouts in terms of value creation potential. Researchers argue that the value creation potential from operations is limited, as the first private equity owner have realized the easiest improvements, with the highest impact already and only leaves the remnants to the secondary owner (Achleitner and Figge, 2012). They further argue that secondary buyouts may take on excessive financial risk in order to exploit attractive debt market conditions (Freelink and Volosovych, 2012). A final criticism posted by scholars is that secondary buyouts are too expensive due to the timing- and negotiation skills of the selling private equity (Achleitner and Figge, 2012). With statements like the ones found below, the business press shares the skepticism of academic scholars. Once a business has been spruced up by one owner, there should be less value to be created by the next (The Economist, 2010). 2 The 26.2 % is based on the aggregated deal value of the exited buyouts. When looking at the number of portfolio companies exited through secondary buyouts the average frequency is reduced to 11.2 % of all exited buyouts, i.e. the numbers are comparable between Europe and the US. 5

8 ( ) After all, since a goal of private equity firms is to trim costs from their companies before selling them, how much more fat can a second private equity owner cut? Secondary buyouts feed the criticism that private equity firms are more focused on fees and financial engineering than operational improvements (Flaherty and Davies, 2007). In sum, this means that secondary buyouts have become a significant part of the private equity industry, but at the same time secondary buyouts generate major doubts about their justification. Wright et al. (2009) have highlighted secondary buyouts as one of the areas where further research on performance between first and second buyouts is required. Based on the situation with ambiguity from the controversy between the popularity of secondary buyouts and the widespread skepticism, we are motivated to investigate the following research question. Research Question: What are the differences in operational value creation in portfolio companies in SBOs, compared to LBOs in the Nordics? 1.1. Scope and delimitations Throughout this thesis there will be a general focus on operational performance and value creation with a specific emphasis on secondary buyouts (SBOs) compared to the equivalent leveraged buyout (LBO). The return on invested capital (ROIC) serves as a framework since it is an overall indicator of operational performance (Petersen and Plenborg, p. 94). The measure of ROIC itself will not be included for evaluation. Instead, the analysis will involve financial measures that are components of ROIC when breaking it down. The scope of the analysis is bounded by the performance of the private equity owned portfolio company and there will therefore not be focused on fund level performance. When covering the academic literature we will address the theoretical arguments of others concerning the role of 6

9 debt and capital structure in leveraged buyouts. Debt and capital structure will however not be a part of the analysis in this thesis, as the focus solely will be on operational performance. We are examining Nordic portfolio companies, which everywhere mentioned in this thesis refers to Denmark, Norway and Sweden. Investigating the Nordics is natural due to comparability and because these buyout markets has a significant size that is growing. These characteristics make it an ideal source for studying secondary buyouts (Wang, 2012) Definitions and abbreviations Throughout this thesis a number of abbreviations and technical terms are applied. Abbreviations will be stated in brackets and used whenever convenient going forward. Technical terms and references to names in figures are shown in italics. We use the terms leveraged buyout, LBO, primary buyout, primary owner and first round buyout interchangeably as a reference to the first private equity ownership period. In the rest of this chapter, and chapter 2 and 3, leveraged buyout will refer to the concept in general. From chapter 4 and onwards the leveraged buyout term will be used as a reference for the transaction before the SBO. Based on the context of the application it should be clear whether LBO refer to the first transaction or the concept. Finally, we use SBO, secondary buyout, secondary owner and second round buyout interchangeably for describing second consecutive private equity ownership Thesis structure The remainder of the thesis is structured as follows. Chapter 2 contains a conceptual introduction to private equity in general as well as a short overview of the recent development of the buyout markets in Europe and the Nordics. Chapter 3 provides a literature review on value creation in leveraged buyouts in general and with a certain focus on SBOs. Chapter 4 is a brief introduction to the ROIC framework used for the forthcoming hypotheses development, data collection and analysis. Furthermore, we provide an outline of the study design applied. In chapter 5 the hypotheses are developed based on what was covered chapter 3 and 4. Chapter 6 serves a comprehensive explanation of the sampling process, the data collection and the statistical method applied. In chapter 7 we provide descriptive statistics on the sample and engage in preliminary analysis. Hereafter we report the results from testing the hypotheses and analyze on the statistically significant results. Chapter 8 contains a discussion of the results reported in chapter 7. 7

10 Furthermore, implications of the findings on SBOs compared to LBOs is discussed in the context of operational performance based on the 7 preceding chapters. Finally, in chapter 8, we also engage in a discussion regarding limitations and biases of the method used. Ultimately, chapter 9 is used to sum up and draw conclusions as well as outlining what we see as future research proposal based on our findings. 8

11 2. About Private Equity and SBOs This chapter serves the purpose of describing private equity as business model. In order to do so the first section will place private equity relative to alternative investments. Hereafter, a thorough introduction to private equity investments is given. To facilitate this the section is split into the role of the limited partners and the general partners, respectively, as well as a review of the options of exiting the investment. Finally, a snapshot of the activity in the European buyout is provided to show trends in the overall market and the sub- market for SBOs. Moreover, the Nordic buyout market is placed in the overall picture, as this is a specific part of the research question and hence the scope of the analysis Private equity as an asset class Private equity investments are considered relatively risky investments compared to investments in publicly traded equity because investors in private equity are subject to illiquidity and long- term commitment. Due to these risk- properties of private equity, investors should expect a higher return from private equity investments compared to, for example, investments in quoted equity (Gilligan and Wright, 2010). Investors in quoted equity rely on publicly available information to purchase tradable shares in the market. In practice, this means that investors act on the basis of imperfect information and low control, but at the same time high liquidity. Most investors in quoted equity do not have control over the company in which they have invested, but they are compensated for that by being able to easily sell their shares in the market (Gilligan and Wright, 2010). This is very different for private equity investments. First, investors commit their capital for a period of 10 years and, secondly, they cannot generally sell only a portion of their investments and therefore rely on a sale of the entire target company. However, investors in private equity are compensated for the illiquidity and long- term commitment by stronger control and private information (Gilligan and Wright, 2010). 9

12 2.2. Private equity investments Leveraged buyout transactions represent the late stage of the private equity category, mature and stable firms, while venture capital represents the early stage, business start- ups, of the private equity category (Berg and Gottschalg, 2003; Kaplan and Schoar, 2005). Another common distinction of leverage buyout transactions is between Management Buyouts (MBOs) and Management Buy- ins (MBIs). In a typical MBO the current management team seeks support from outside investors of both debt and equity capital in order to acquire the majority control of the company they are managing from its current owners. In a MBI, an external management team seeks to take control of a company with support from outside investors (Loose, 2005). Private equity investments are carried out through so- called limited partnerships in which the private equity firm serves as the general partner and where the limited partner is the sum of investors (Kaplan and Schoar, 2005). In a typical leveraged buyout transaction, a target company is acquired by a private equity firm using a relatively small portion of equity, 10 to 40 percent, and a relatively large portion of debt financing, percent, in order to acquire majority control of an existing or mature firm (Kaplan and Strömberg, 2009; Wright et al., 2009). Wright et al. (2009) describe leveraged buyouts in the following way: Buyouts involve private equity firms, backed by borrowings, acquiring a significant or majority stake in an existing business. Private equity firms are typically active investors with access to comprehensive and timely information. In contrast to investors in public companies, they take board seats and specify detailed contractual restrictions on the behavior of management. They also benefit generally both from more detailed pre- purchase due diligence and from full, timely information on the current trading of the business in which they invest. Lenders to leverage buyouts also typically specify and closely monitor detailed loan covenants. (Wright et al., 2009, p. 3) One or more financial institutions usually provide the debt used in leveraged buyouts, while the private equity firm will invest the capital committed by the limited partners as an equity investment in order to cover the rest of the acquisition price (Kaplan and Strömberg, 2009). The management team of the target company is usually required to co- invest by participating in the 10

13 equity investment, although their investments often only equals a small fraction of the total equity investment (Kaplan and Strömberg, 2009). The respective private equity firms, i.e. the fund managers, responsible for making the investments in target companies are being compensated in two ways. First, they earn management fees, which correspond to usually 1-2 percent of the total capital committed to the fund. Management fees also include a percentage of capital employed this means that the fund managers earn a percentage of the respective investments made during the acquisition phase (Kaplan and Strömberg, 2009). Secondly, fund managers receive carried interest, which is an incentive scheme that kicks in only if the fund managers have generated profits above a threshold, generally corresponding to 8 percent return on capital. Carried interest generally equals 20 percent of all profits above the threshold (Kaplan and Strömberg, 2009). In that regard, it is common that the fund managers also invest themselves alongside the investors in order to align interests. From the realm of mergers and acquisitions (M&A), leveraged buyout transactions can be compared to unrelated acquisitions because private equity firms usually manage their target companies completely independent from each other (Berg and Gottschalg, 2003; Baker and Montgomery, 1994). This means that leveraged buyout are not motivated by synergies - the potential advantages of integrating one company into to another company - but merely by the potential to increase the value of the respective target companies as a stand- alone company beyond the acquisition price (Berg and Gottschalg, 2003; Gilligan and Wright, 2010) The limited partner Investors in private equity, who are generally institutional investors and high net worth individuals, initially make commitments to invest a certain amount of capital in a fund, and then pay in their capital when they are required to do so by the private equity firm in order to finance the acquisition of the respective target companies (Gilligan and Wright, 2010). After committing their capital to the fund, the limited partners have little influence on how the private equity firm deploys the capital as long as it is within the fond agreement (Kaplan and Strömberg, 2009; Gilligan and Wright, 2010). 11

14 The general partner Baker and Montgomery (1994) compare conglomerates and LBO associations, i.e. private equity firms, with respect to organizational form. In that regard, they find that private equity firms have flat hierarchical structures with a small number of partners, associates and back office positions. Also, they argue that private equity firms are efficient organizations because they apply limited reporting lines; all of the non- partner professionals report to all of the partners as a group. Baker and Montgomery (1994) argue that both organizations are diversified, but that they differ in their acquisition- and divestiture strategies. Conglomerates are strategic buyers because their acquisitions often involve companies that complement their existing business with the potential of achieving scale- and scope economies, whereas private equity firms do not attempt to foster synergies across their acquisition. Private equity firms do not care so much about resource relatedness and strategic fit, but merely about a set on generic criterions and financial benchmarks. With regards to divestitures, Baker and Montgomery (1994) argue that conglomerates are unlikely to sell their acquired units as long as they meet the profit hurdle rate of the respective conglomerates whereas private equity firms, in contrast, explicitly pursue a buy and sell strategy because they must return cash to their investors. Another dimension that distinguishes private equity firms, from for example traditional portfolio investors, is that private equity firms engage in active ownership. Jensen (1989) describes private equity firms as: investors who hold large equity or debt positions, sit on boards of directors, monitor and sometimes dismiss management, are involved with the long- term strategic direction of the companies they invest in, and sometimes manage the companies themselves (Jensen, 1989, p. 6). In terms of active ownership, Berg and Gottschalg (2003) suggest that private equity firms improve the operational efficiency, and thereby create value, of the portfolio company by engaging in two strategies, which they refer to as corporate refocusing and buy- and- build strategies. 12

15 The former, corporate refocusing, often involves reductions in the complexity of the portfolio company by restructuring around its core business. Berg and Gottschalg (2003) specifically discuss reductions in the degree of diversity and inefficiencies among different product lines. In other words, the private equity firm focuses on allocating resources towards the activities that are core to the competitive advantage of the portfolio company. The latter, a buy- and- build strategy, is characterized by an initial investment in a target company that is not performing to its abilities. Then, the private equity firm will in order to change the performance of the target company supply the company with cash, leverage and know- how in order to make add- on acquisitions to the portfolio company over time as well as organic growth. According to Berg and Gottschalg (2003) private equity firms engage in buy- and- build strategies in order to drive consolidation in a particular market segment and in order to achieve a dominant market position. Brigl et al. (2012) suggest further ways in which private equity firms can create operational value by focusing on four broad areas of the business for the portfolio company. In total, these four areas present 25 ways in which operational value can be achieved. This involves improvements in the financial structure of the company, a focus on the bottom line of the company or, lastly, improvements in the top line of the company. These means of operational value creation are illustrated in the below figure 1. 13

16 Figure 1: Four areas where portfolio companies can create operational value Source: Brigl et al. (2012) Private equity funds and fundraising Private equity firms raise capital from investor through private equity funds. These funds are typically structured as finite- life vehicles and therefore have a fixed investment period, of usually 10 years, during which the private equity firm must engage in fundraising, acquisitions of target companies, value creation and, finally, divestments (Gilligan and Wright, 2010; Berg and Gottschalg, 2003). Thus, private equity firms generally have five years to invest the fund s capital and additionally five years to return capital to the limited partners (Kaplan and Strömberg, 2009). As the private equity business model is built upon buying and selling companies, private equity firms need to raise new funds every three to five years in order to facilitate a sufficient deal flow. In that regard, private equity firms depend on a good track record for both timely and successful exits, and past performance measured by Internal Rate of Return (IRR) because these factors are important with regards to enhancing the reputation of the respective private equity firms which increases the probability of raising new funds (Kaplan and Schoar, 2005). 14

17 Gompers and Lerner (1998) discuss the determinants of capital inflows into the private equity industry at the aggregate level. They find that changes in macroeconomic factors are positively related to capital inflows into private equity. Among those factors are overall economic growth, the past performance of the private equity industry, and changes in capital gains tax rates. In addition, they also argue that firm- level performance and firm- level reputation affect fundraising. Hence, private equity firms with large equity stakes in companies, which have recently been IPOed, are more likely to raise lager funds in the future. Also the reputation, measured by the age and size of the private equity firm, influence its ability to raise new capital. Kaplan and Schoar (2005) analyze the correlation between track record performance and capital inflows of private equity firms at the fund level. They show that the best performing private equity firms are more likely to raise follow- on funds, and larger funds. In particular, they show that capital inflows are positively related to past performance. Interestingly, they also find that return persistence exists in both ends of the performance distribution. This means that poorly performing private equity firms continue to perform poorly, while successful private equity firms continue to perform well. Kaplan and Schoar (2005) argue that their results can be explained by access to proprietary deal flow, which means that the best performing private equity firms have access to superior deals. They also argue that their results are due to the heterogeneity of the quality of general partners. Because general partners engage in active ownership, return persistence, can be explained by scarcity of high- quality general partners The acquisition phase The acquisition phase does not only involve the actual buyout transactions, but also investment sourcing, due diligence and negotiations. With regards to investment sourcing, private equity firms rely on screening the market for potential target companies which offer the value creation potential high enough to meet the return criteria, measured by IRR, demanded by investors (Loose, 2005). Private equity firms invest considerable time and effort in building relationships, which may turn out to give privileged access to attractive deals, the proprietary deal flow, that generally offer higher returns because the attention from competing buyers is avoided (Gilligan and Wright, 2010; Loose, 2005). 15

18 It is also during the acquisition phase that the private equity firm starts to build an understanding of the company through the due diligence process. Private equity firms will also at this relatively early stage develop a suggested business plan for the company, which also includes plans with regards to the preferred exit choice in the future. In addition, private equity firms will at this point make important decisions regarding the structure of the transaction (e.g. the split between debt and equity), the distribution of management equity stakes and the design of incentive systems (Berg and Gottschalg, 2003; Loose, 2005). Private equity firms also engage in negotiations about the acquisition price of the target company. The acquisition price is a critical factor because it will influence the return on the investment once the target company is exited (Berg and Gottschalg, 2003). In general, deals are preferably privately negotiated in order to avoid the attention of competing buyers, which will drive up the acquisition price (Loose, 2005). Gompers and Lerner (2000) study how the inflow of capital to private equity affects the valuation of target companies. They show that there is a strong positive relationship between capital inflows and the valuation of private equity investments. In order to explain this relationship, Gompers and Lerner (2000) argue that because private equity partnership agreements specify clearly that funds can only invest in private equity, the demand for private equity investments results in more competition and thereby higher valuations. Thus, rising private equity valuations are not driven by better investment prospects, but merely by the demand for private equity investments (Gompers and Lerner, 2000) Value creation During the holding period private equity firms will begin to actively manage the investments completed during the acquisition phase. This means that the private equity firms will seek to implement the strategic-, organizational- and operational changes formulated in the suggested business plan (Berg and Gottschalg, 2003). In practice, the implementation of the business plan is more of an iterative, than linear, process where the business plan is changed and revised continuously during the holding period (Berg and Gottschalg, 2003). 16

19 Private equity firms will seek to exert their influence during the holding period through their representation on the board of directors of the respective portfolio company. Private equity boards are often engaging in the formulation of strategy and are often serving as a source of strategic initiatives and ideas, for example on M&A. Private equity boards are also often assuming the role of stimulating the management team of the target company to think more broadly and creatively about new business opportunities. In that regard, the top three priorities of private equity boards are value creation, exit strategy and strategic initiatives, including M&A (Acharya et al., 2008) Divestment phase The divestment phase terminates the entire investment, and is very important because the exit choice, and the timing, ultimately determines the return that the private equity firms will achieve on their investment (Berg and Gottschalg, 2003). The divestment phase is also very important because successful exits enhance the reputation of the respective private equity firms, which is important for their ability to raise future funds The exit choice Private equity investments are often realized through initial public offering (IPOs), trade sales or via secondary buyouts. Initial public offerings are often considered the preferred exit choice of private equity firms because it enhances the reputation of the respective private equity firms, which is important in relation to their ability to raise future funds (Sousa, 2010). IPOs are arguably also preferred because these tend to result in the highest valuation of the company (Nikoskelainen and Wright, 2007; Sousa, 2010). However, IPOs are by nature subject to potentially high uncertainty in the stock market, which make this exit choice difficult to manage. Furthermore, private equity firms cannot, due to lockup clauses, realize the entire investment in the company at the IPO, which means that the respective private equity firms will continue their relationships with the company (Sousa, 2010). Private equity firms can also exit their investments in a trade sale to a strategic buyer. These buyers are often larger firms operating in a comparable industry to the target firm who intends to vertically- or horizontally integrate (Sousa, 2010). This exit choice is preferable because private 17

20 equity firms can realize the entire investment and thereby end their relationship with the portfolio company. Finally, private equity firms can exit their investments through a sale of the company to another private equity firm. A secondary buyout is preferable, like trade sales, because private equity firms can realize the entire investment and thereby end their relationship with the company (Sousa, 2010) The European and Nordic private equity market at a glance As touched upon in the introduction the global financial crisis did not gone unnoticed in the private equity industry. The most recent yearbook of private equity data published by the European Venture and Capital Association (EVCA) in 2014 shows the characteristics and trends of the private equity market since 2007 and hence prior to the outburst of the financial crisis. In this section we will cover the major trends in private equity activity in the buyout market with a distinct focus on SBOs and the Nordics. Figure 2: The development of the overall buyout market in Europe against the development in SBO activity 3 Value of deals Total Value of all buyouts and SBOs in Europe ( ) Deal value total ( thousands) Deal value SBOs ( thousands) Value of deals SBOs Source: Authors contribution. Data: EVCA (2014) 3 Value of all buyouts is based on investment activity data. Value of SBOs is based on divestment activity data. 18

21 Figure 2 shows that the overall buyout market in Europe peaked in 2007/08 in aggregated deal values. The numbers of firms invested in, not present in the figure, fluctuated between 4,800 and 5,600 over the period with 2009 being the year with the fewest investments in parity with the aggregated value (EVCA, 2014). The trend in the amount invested in SBOs appears to follow the buyout market in general except for 2013 where the SBO activity increases in a decreasing market. Between 2007 and 2013 SBOs has accounted for an average of 26.2 % of the exits of portfolio companies measured by deal value and is together with trade sale the most prevalent route of exit based on the value of these types of deals (EVCA, 2014). Figure 3: Private equity activity measured by value of investments in the Nordics compared to the largest markets in Europe as percentages of the total European market NORDICS PRIVATE EQUITY ACTIVITY RELATIVE TO THE LARGEST MARKETS IN EUROPE Nordics United Kingdom France Germany 14,96% 17,96% 12,44% 17,40% 16,71% 12,93% 28,68% 24,60% 20,72% 11,68% 14,86% 18,03% 13,79% 15,85% 21,42% 14,35% 18,03% 30,34% 22,84% 27,28% 26,92% 8,73% 8,66% 9,46% 12,02% 11,46% 11,85% 12,07% Source: Authors contribution. Data: EVCA (2014) The chart in figure 3 shows the activity in Denmark, Norway and Sweden grouped as the Nordics in comparison with the three largest buyout markets in Europe as a percentage of the total European buyout market. The percentage calculations are based on aggregated deal values of the domestic country of the portfolio company. The numbers show that whereas the German, French and UK market shares are more or less in parity with the beginning of the period the Nordic market has grown relative to others. Looking at the number of portfolio companies invested in, Germany is 19

22 the largest market over France, UK and the Nordics over the entire period. For the market share, measured by number of companies invested in, the Nordics have between 11 % and 16 % of the total European market over the period. A chart showing the shares based on number of companies traded can be found in the figure 11 in appendix 1. Figure 4: Private equity activity measured by value of investments in Denmark, Norway and Sweden between 2007 and ,00% 60,00% 50,00% 40,00% 30,00% 20,00% 10,00% Private equity in the Nordics by the value of pormolio companies traded 0,00% Denmark Norway Sweden Source: Authors contribution. Data: EVCA (2014) Figure 4 presents the activity in the Nordic countries relative to each other measured by the aggregated deal value of the investments. Looking at the chart it is evident that Sweden historically has been the most developed private equity market of the three, quantified by the level of activity. The activity in 2013 is however an exception for this pattern. Furthermore, it can be seen that Demark and Norway are comparable in terms of the value of portfolio companies traded in most years. The pattern of Sweden being the most active buyout market, Norway the second most active in front of Denmark is evident when looking at the number of buyouts per year. By doing so the relative sizes seem stable with Sweden accounting for approximately 60%, Norway approximately 25% and Denmark the remaining 15 %. A chart similar to figure 4, but based on the number of portfolio companies invested in, is provided in the figure 12 in appendix 1. 20

23 The market data show that the overall European buyout market has been picking up after the financial crises, although without reaching the levels of the buyouts in 2007 and SBOs have been a popular choice of exit over the entire period covered, and for 2013 the value of SBOs traded increased in a slightly decreasing buyout market stressing the central role that SBOs play. Market data furthermore show that the Nordics as a group has increased its share of the total market continuously over the period opposed to the three countries with the largest buyout markets in Europe. In 2013, these countries had a market share more or less similar to where they started. Zooming in on the Nordics it was evident that Sweden is the most developed private equity market with close to 50 % or more of the deal flow if disregarding

24 3. Literature Review Following the increased frequency of LBOs and private equity transactions throughout the 1980s, Jensen (1989) stated that a new organizational form had emerged and that it should be encouraged due to its ability to make significant gains in operating efficiency, employee productivity and value for the owners. The success of the private ownership was credited to the mitigation of misaligned interest between managers and owners by the use of a new model of general management. The model comprised of highly leveraged financial structures, performance driven payment, and substantial equity investment by the management. This chapter serves as an overview of central and recent literature on LBOs in general and SBOs specifically. First, we present the three main levers of value creation in LBOs. Next, we shed light on how SBOs distinguish from LBOs and explains some of the motives that drive private equity players to engage in SBOs. Lastly, we highlight empirical findings on changes in operational performance post LBOs generally and present results from scholars examining SBOs in various ways. The theoretical explanation of value creation in LBOs also applies in order to explain the value creation potential in SBOs because an SBO is simply a second LBO. Hence, the value creation levers should logically be the same (Freelink and Volosovych, 2012). In that regard, the literature on LBOs commonly recognizes three key value creation levers in LBO transactions, namely, financial engineering, governance engineering and operational engineering as suggested by Kaplan and Strömberg (2009). Other recently active scholars have commonly referred to these value creation levers as operational performance, leverage and pricing (Evers and Hege, 2012; Achleitner and Figge, 2012; Freelink and Volosovych, 2012). The concept of these categorical levers will be outlined in the following Financial Engineering Financial engineering is related to finding the optimal capital structure of the company and exploiting the benefits of leverage the best way possible, i.e. where the point where the marginal cost of debt just offset the marginal benefits of debt (Jensen, 1986). There are at least two ways in 22

25 which the use of leverage, the borrowing that is done in connection with the transaction, affects the portfolio company in private equity in terms of increasing value. First, and most evident, leverage has a potential effect on the value of the portfolio company due to the tax deductibility of interest paid on debt (Kaplan and Strömberg, 2009; Kaplan, 1989b). Kaplan (1989b) finds that the benefit varies between 4 and 40 percentage of a firm s value can be attributed the interest tax shield depending on the underlying assumptions. Kaplan and Strömberg (2009) claims that the value possible to generate from interest tax deductions has decreased for portfolio companies compared to the 1980s, as the corporate tax and the level of leverage in transactions has decreased in the 1990s and 2000s. Secondly, there is a less direct effect from the increase in interest payments. This makes the free cash flow of the company a scarce resource for the management and hence incentivizes them to allocate capital optimally (Jensen, 1986). High level of debts is a mechanism that forces management to focus on cash flow generating activities only rather than spending money on empire- building projects with limited returns or organizational inefficiencies (Jensen, 1989). Thus the disciplining effect caused by high levels of debt serves as an important tool for the owners to control management. Besides its benefits, leverage comes at a price as it increases the likelihood of financial distress (Kaplan and Strömberg, 2009). Further cautionary note with regards to leverage is placed by Rappaport (1990) who argues that borrowing per se creates no value other than tax benefits. Value comes from the operational efficiencies debt inspires (p.102) Governance Engineering Agency theory suggests that there is an inherent conflict between the owners and the managers of the corporation, and the agency theory is therefore concerned with the contractual problems that arise when the principal (the owner) hires an agent (manager) to perform a task on behalf of the principal. This relationship often involves delegating some degree of decision- making authority to the agent, and because both parties seek to maximize their own utility, agency theory suggests that the agent does not always act in the best interest of the principal. In that situation the principal can either introduce monitoring or incentives to resolve the conflict of diverging 23

26 interests. However, these actions are costly to the principal and those costs are known as agency costs (Meckling and Jensen, 1976). In terms of agency costs, Jensen (1989) discusses how LBOs provide the portfolio companies with strong prerequisites for improving operational performance. First, he argues that LBOs reduce agency costs as the use of high leverage mitigates the conflict of interests between the owners and the managers of the portfolio company in regards to managing the free cash flow. High leverage increases the pressure on managers to allocate the free cash flow more efficiently because the periodic interest- and principal payments must be honored. Secondly, Jensen (1989) argues that LBOs reduce agency costs by aligning interests, between the owners and the managers, through managerial equity ownership and incentives schemes. In private equity driven LBOs management is required to make a significant co- investment in the company (Kaplan and Strömberg, 2009). This means that the managers of the portfolio company are invited to participate in the potential upside of the company, which makes it suboptimal for managers to focus on short- term performance at the expense of long- term value creation (Jensen, 1989). Finally, Jensen (1989) argues that LBOs reduce agency costs through effective monitoring. Investors and creditors are in a privileged position to closely monitor the decisions and strategies put forward by the managers of the portfolio company through their representation on the board of directors. Governance engineering do not per se have a direct effect on the cash flow of the portfolio company and hence value creation, but arguably serve to indirectly enhance the cash flow of the company (Berg and Gottschalg, 2003). Scholars have contributed their empirical results on performance improvements in LBOs to reductions in agency costs, change in ownership and improved incentive alignment (e.g. Kaplan, 1989a; Bull, 1989; Smith, 1990). Wright et al. (2009) questions the value creation proposition based on solving in the agency problem in SBOs by posting the argument that a resolution to agency problems is expected to generate a one- off change in performance of the portfolio company. This implies that such improvements are not easily available to the follow- on private equity firm in SBOs if the primary owner succeeded in aligning the interests. However, Figge (2012) acknowledges that agency problems can be further reduced in SBOs given that I) the management team learns to better understand the mechanisms of the management package, or II) given that the management team 24

27 increases their financial commitment to the management package, or III) given an increase in the number of participant in the management package. Thus, according to Figge (2012) incentive alignment can be improved in SBOs under these circumstances Operational Engineering Private equity firms use their operating expertise and industry knowledge to identify attractive investments in which they can improve the operating performance and cash flow of the respective portfolio companies (Kaplan and Strömberg, 2009). Loos (2005) recognizes that about two thirds of value creation in leveraged buyout transactions is derived from improvements in operational performance during the holding period and Guo et al. (2011) find that operating performance is one of the key value creation drivers in private equity transactions. Operational engineering are commonly recognized as those changes that have a direct effect on the earnings and eventually the cash flow of the portfolio company (Berg and Gottschalg, 2003). In particular Berg and Gottschalg (2003) categorize operational engineering into cost optimization and margin improvements, reducing the capital requirements and increasing the utilization of assets. Under these categories opportunities related to sales growth, cost optimization and margin improvements, reductions in capital requirements, removal of managerial inefficiencies, improved asset efficiency, synergies from acquisitions, are the primary ones (Berg and Gottschalg, 2003; Loose, 2005; Kaplan and Strömberg, 2009). This view is shared with Gou et al. (2011) who states that ( ) firm value will increase if there are firm- specific improvements in operating performance; for instance, sample firms may improve profitability, eliminate unproductive assets, use remaining assets (including working capital) more efficiently, or make value- increasing acquisitions. The value creation driven by operational performance improvements in SBOs have been faced by skepticism and challenged in the literature: ( ) The operational value creation potential in SBOs is thought to be limited, since the first private equity sponsor will already have realized the low- hanging fruit, that is, the easily realized value creation measures with the largest impact. (Achleitner and Figge, 2012, p.2). 25

28 However, Achleitner and Figge (2012) examine operational performance in various types of buyouts and finds that SBOs offer potential for operational performance improvements similar to other types of buyouts Differences between LBOs and SBOs As described throughout the literature, consensus is that LBOs increases value of companies, particularly in terms of improvement of operational performance (Kaplan and Strömberg, 2009). However, as touched upon briefly in the previous section this is not necessarily the case for SBOs. The recent literature on SBO is characterized by the notion of conventional wisdom, which refers to the perceived limited value creation potential for SBOs relative to buyout transactions sourced elsewhere (Sousa, 2010; Achleitner et al., 2012; Bonini, 2012; Wang, 2012; Achleitner and Figge, 2012; Evers and Hege, 2012). The skepticism among scholars is based on the logic that private equity firms who are similar in their business model, relies on the same sources of value creation and hence the value creation potential has thus probably already been exhausted by the selling private equity firm (Achleitner et al., 2012). In that regard, Cumming and MacIntosh (2003) argue that private equity firms will only sell a portfolio company once the expected marginal return of value creation through their effort and investments is lower than the marginal cost represented by that very effort and investment. This, arguably, makes it questionable why follow- on private equity firms should consider SBOs to be attractive investments in terms of value creation potential. The absence of evident opportunities should thus result in no or limited interest in making investments in a SBO (Freelink and Volosovych, 2012). In order to find an explanation of the paradox of SBOs, the rest of this section will provide some of the theoretical arguments, alternative to value creation, suggested as the drivers of SBOs in the literature Theoretical drivers of SBOs In the literature on SBOs different hypotheses have been proposed and tested in order to explain both the sellers and buyers structural- and opportunistic motivations for engaging in SBOs. In the following we will go through some of the central theories proposed in the most recent literature. A review of the actual findings will be presented in the third section. 26

29 Motivation for engaging in SBOs Window of opportunity The window of opportunity hypothesis states that private equity firms seek to take advantage of the macroeconomic trends in the debt- and equity market in order to achieve the highest possible price at the time of the exit of the LBO. In that regard, SBOs are arguably positively related to cheap debt financing conditions caused by liquid debt markets and negatively related to hot IPO markets (Sousa, 2010; Evers and Hege, 2012). The use of higher leverage in SBOs can arguably be explained by lower information asymmetry in SBOs compared to LBOs. In other words, the lending banks know the portfolio company and its management team better after the first- round buyout and are therefore more willing to provide debt financing as they know that the portfolio company was able to cover its debt obligations in the first- round buyout (Evers and Hege, 2012). The window of opportunity hypothesis is supported by the empirical findings of several scholars (Wang, 2012; Bonini, 2012; Evers and Hege, 2012; Achleitner and Figge, 2012) who show that SBOs are driven by macroeconomic debt- and equity market conditions. Skilled seller According to the skilled seller hypothesis private equity firms use their negotiation experience and market timing skills to sell their portfolio companies at a higher multiple than the acquisition multiple. This makes it harder for the second private equity owner to make a good deal in terms of IRR if the entry multiple gets too high (Evers and Hege, 2012). Other scholars find that SBOs are more expensive than LBOs, which supports the skilled seller hypothesis (Wang, 2012; Achtleitner and Figge, 2012). Structural incentives The structure hypothesis as suggested by Sousa (2010) implies that private equity firms choose SBOs as an exit choice in order to signal to investors, the limited partners (Sousa, 2010). Since private equity funds are structured as limited- life vehicles, private equity firms have strong incentives to avoid long holding periods because a good track record, measured by IRR, is very important to the reputation of the respective private equity firm and its future ability to raise new funds (Sousa, 2010; Evers and Hege, 2012). In that regard, it has been argued that transactions, 27

30 where both the seller and the buyer are private equity firms, are quick and efficient deals because both parties are professional (Degeorge et al., 2013). Specialization Sousa (2010) argues through the specialization hypothesis that private equity firms have different focus in terms of the investments, e.g. private equity firms focusing on portfolio companies in the early stages opposed to others focusing on developed portfolio companies. This implies that a second private equity firm with a completely different focus can benefit the portfolio company in terms of value creation. The first private equity firm is thus interested to pass on the portfolio company as their abilities have been deployed fully, and value has been maximized given their abilities. Acharaya et al. (2012) finds evidence for specialization for buyouts in general. They show that human capital factors such as abilities and experience from operations possessed by the private equity professionals can improve the companies acquired by the private equity firm. Efficiency gains Wang (2012) formulates the efficiency gains hypothesis, which is related to the value creation argumentation. Specifically, it states that even though there might be efficiency gains in the first LBO there could not still be value to capture for the second private equity owner. It is argued, that this for instance could be the case, if the first private equity firm exited before all benefits were exploited or if the second private equity firm have a different set of skill and hence be able improve the operational efficiency further as suggested by Sousa (2010) above. Dry powder Degeorge et al. (2013) argues that dry powder, that is the difference between the assets under management and the capital invested by the private equity firms, drives the number of SBOs. For the buying private equity fund, the dry powder late in the investment period of the fund cycle increases the number of SBOs (Figge, 2012). 28

31 Go- for- broke Axelson et al. (2009) presents the go- for- broke hypothesis stating that the financial structure of private equity funds incentivizes the general partners with untapped funds to take bad deals rather than not investing the funds. Degeorge et al. (2012) argues that in today s private equity industry there is ample possibilities to exert a go- for- broke behavior through SBOs with all the companies held by private equity firms. Collusion Wang (2012) argues that SBOs might be driven by a collusion motive, which means that private equity firms are trading portfolio companies among each other because opacity and lack of regulations creates a favorable environment for trading these assets. Alternatively, the collusion motive implies that private equity firms are trading portfolio companies in order to help each other out, as the seller need to demonstrate successful exits towards its investors and the buyer are under pressure to invest its capital Empirical findings This section contains a review of some of the empirical studies on operational performance in various types of LBOs. In the literature review, in the first section of this chapter, it was argued that there might still be value creation potential left from enhancing the operational performance of the company in SBOs. Furthermore as suggested by Kaplan and Strömberg (2009) operational improvements has become the primary source of value creation for private equity investors. After highlighting empirical findings on operational performance in LBOs, a review on the empirical findings from the recent SBO studies will be provided to round off the literature review Empirical findings on operational performance in leveraged buyouts In the comprehensive academic literature on LBOs there have been numerous studies to examine the operational performance and efficiency of firms in private equity transactions in various ways. In that regard, the studies presented in the table 1 below are merely a representative selection of what has been written in studies that deals with operational performance improvements of first 29

32 round buyouts. Studies that examine first round buyouts relative to secondary buyout or secondary buyouts only, will be discussed in greater details in the next section. Table 1: Previous findings on operational performance changes in LBOs Paper Kaplan (1989a) Findings EBIT/sales and EBIT/total assets increases approx. 20 % more, compared to industry peers, in the first three years after the buyout. Bull (1989) Comparisons of pre and post buyout performance are made for the same entity in terms of the ratio of company performance for seven accounting variables. In that regard, post buyout performance is found to be superior to pre buyout performance. Smith (1990) Post- buyout operating returns increases compared to industry peers. Working capital management is tightened after the buyout and working capital turnover increases by approx. 19 % compared to industry peers. Muscarella and Vetsuypens (1990) Accounting measures of performance show significant improvements in profitability, which is derived from the firms ability to reduce costs. Guo et al. (2011) Post- buyout gains in operating performance are either comparable to or slightly exceed those observed for benchmark firms. Gaspar (2012) Shows significant small increases in ROIC and EBITDA margin in the post- buy out years. In terms of capital utilization there is evidence for tighter control of working capital whereas results are mixed for capital turnover and asset utilization. 30

33 3.4. Empirical studies on SBOs The relatively young and concentrated empirical body of literature on SBOs can be divided into three categories, as suggested by Evers and Hege (2012). The first category focuses on the drivers of SBOs while comparing them to other exit strategies (Sousa, 2010; Achleitner et al., 2012). The second category examines the differences in performance and value creation between LBOs and SBOs (Achleitner and Figge, 2012). Meanwhile the third category merely is a combination of the two previous categories (Bonini, 2012; Wang, 2012; Evers and Hege, 2012; Freelink and Volosovych, 2012). In addition, Figge (2012) suggests that the literature can be divided according to the level of analysis employed in the respective studies, and he particularly distinguishes between the fund level and the transaction level. Where fund level studies focus on the structure- and performance of private equity funds, deal level studies focus on deal returns and the drivers of those deals. The literature on SBOs, which we cover belongs to the latter, namely deal level studies. Sousa (2010) studies a sample of private equity backed companies that have been exited through SBOs, IPO or trade sale in Europe from 2000 and up to The support for the structure hypothesis is mixed, which states that private equity firms chose SBOs in order to avoid long holding periods. The window of opportunity hypothesis, where private equity firms taking advantage of favorable capital market conditions, is however confirmed. In addition to this, both the specialization hypothesis and the monitoring hypothesis are rejected (Sousa, 2010). These hypotheses states that a follow- on private equity firms can increase the performance of the portfolio company because they are more specialized and profitable companies needs less control, respectively. Achleitner et al. (2012) analyze the characteristics of SBOs as an exit strategy compared to other exit strategies. They use a data set including both North America and European exits with a total of 1,112 observations between 1995 and Firstly, they argue that SBOs are equally attractive as LBOs when considering the return to the selling private equity firm. In other words, they argue that private equity firms do not have a clear pecking order with regards to exit strategy. Secondly, they find evidence that SBOs are more likely to occur when the lending capacity of the portfolio firm is high, when the liquidity of debt markets is high, and when the undrawn capital commitment in private equity industry is high. 31

34 Achleitner and Figge (2012) analyze the value creation profile of SBOs, relative to LBOs, based on a sample of worldwide buyouts, including 448 SBOs, in the period between 1990 and First and foremost, they examine the value creation from operational improvements, such as revenue growth, EBITDA growth and EBITDA margin. They find no evidence that SBOs offers lower operational value creation potential, or returns, than LBOs. Interestingly, Achleitner and Figge show that SBOs use percent more leverage, measured by debt/ebitda, than LBOs which they explain by reduced information asymmetry in financial buyouts. Finally, consistent with Wang (2012) and Sousa (2010) they find that SBOs are more expensive than LBOs, which they argue can be attributed to the seller s market timing- and negotiations skills. Bonini (2012) examines the effects on operating performance of SBOs and the determinant of SBO activity. The study is based on a sample of European leveraged buyout transactions from 1998 to 2008, and data is collected at firm level for 163 portfolio companies for two consecutive LBO acquisitions during the same period from 1998 to In that regard, Bonini concludes that first round LBOs generate significant operational improvements and efficiency gains. SBOs, however, do not seem to provide significant improvements in operational performance. Interestingly, Bonini is able to show that for all measures of operating performance the first round LBOs result in a steep, one- off increase, which is smaller or absent in SBOs and he therefore rejects the hypothesis that operating value creation is the main driver behind SBOs. In addition, it is shown that an important driver of SBOs is the flipping hypothesis, which state that private equity firms take advantage of favorable market multiples and favorable financing terms. Wang (2012) studies both the potential explanations for SBO activity and the performance of SBOs. The study focuses on the United Kingdom with 908 deals between year 1997 and year Specifically, Wang (2012) analyzes the role of liquidity- based market timing, efficiency gains and collusion. The results show that SBOs are largely driven by favorable capital market conditions, which is in line with the conclusion of Sousa (2010). Wang (2012) also argues that SBOs are more expensive than LBOs and offer limited efficiency gains. Finally, no support is found for the collusion hypothesis, which states that private equity firms trade companies between themselves in order to manipulate their returns. Evers and Hege (2012) study the French post- crisis leveraged buyout market using two data sets. One of their samples comprises 438 LBOs from 2007 to 2011, and the other sample includes 32

35 139 exits also during the period from 2007 to They conclude that SBOs are not only driven by macroeconomic factors, but also by the value creation potential that they offer. Particularly, they find that private equity firms tend to focus on driving sales of smaller companies in LBOs, and on improving margins of larger companies in SBOs. When they examine the drivers behind SBOs as an exit choice they find that undrawn capital commitments, and pressure to realize investments are important drivers alongside macroeconomic conditions. Interestingly, Evers and Hege (2012) do not find support for the use of higher leverage in SBOs compared to LBOs, which is contrary to the other current studies. Freelink and Volosovych (2012) study a sample of 101 SBOs of UK portfolio companies between 1998 and The study stands out by examining the operating performance of the portfolio company over the complete holding period from investment to exit. Freelink and Volosovych (2012) examine which levers create value in secondary buyouts and to what extent. The characteristics of their sample show that portfolio companies increase the level of debt in the SBOs. Furthermore, it is shown that holding periods are shorter and deal values higher in SBOs. There is empirical evidence supporting that operating performance decreases in the SBOs. We can summarize the current literature according to the consensus that exists about SBOs. First, SBOs are found to be driven by favorable debt- and equity market conditions and, secondly, SBOs use higher leverage than LBOs. Third, SBOs are more expensive than LBOs and, finally, there is some degree of inconsistency in the empirical results regarding operational performance of portfolio companies in SBOs. We can also briefly summarize the literature according to the geographic focus of empirical papers. The geographical dimensions are Europe (Sousa, 2010; Achleitner et al., 2012; Bonini, 2012), United Kingdom (Wang, 2012; Freelink and Volosovych, 2012), France (Evers and Hege, 2012), North America (Achleitner et al., 2012) and Worldwide (Achleitner and Figge, 2012). 33

36 4. Framework and study design In this chapter we will introduce ROIC as a tool and use it for developing a set of testable hypothesis regarding operational performance and value creation in SBOs. Secondly, we will go through the study design, which we have deployed for the data collection and analysis Return on invested capital When assessing the performance of operations in companies using accounting data, ROIC is the overall measure for how profitable operations are and a better metric than other return measures as it focuses solely on operations (Koller et al., 2010, p. 164). In practice RIOC indicates how big a return a business is able to generate per unit invested in its operations. Using ROIC for analyzing operational performance has the advantage that it incorporates both margin improvements and capital efficiency into the performance metric (Gaspar, 2012). In that regard, decomposing ROIC into profit margin and turnover rate of invested capital can be used to explain whether profitability is driven by optimization of the revenue and cost relation of improved utilization of the net operating assets of the company. The profit margin and turnover rate of invested capital can be further be broken into its components as shown in the ROIC tree in figure 5. We rely on the breakdown of ROIC because this allows us to formulate a set of testable hypothesis, in a structured way, for value creation in the companies backed by private equity firms. We do not attempt to measure and evaluate ROIC for two reasons. First, because we use measures of the components that goes into ROIC and, secondly, because ROIC should always be evaluated in comparison to the respective companies cost of capital (WACC) or alternatively by benchmarking ROIC against competitors (Petersen and Plenborg, 2012, pp ). That is not the purpose of our thesis and it is therefore considered to be outside the scope of our research questions. 34

37 Figure 5: ROIC tree ROIC Profit margin Turnover rate of invested capital Revenue Producqon costs Non- current assets Operaqng Working Capital Selling, General & Administraqve costs Amorqzaqon & depreciaqon Source: Authors creation following Petersen and Plenborg (2012, p. 94) 4.2. Study design To examine the changes in operational performance we have chosen an approach where two consecutive private equity ownership periods of the same portfolio company are examined. Comparability of the ownership periods is strived to ensure that the second private equity firm to own the portfolio company has exited 4 the investment, in line with the first private equity firm. This gives three points of references in order to measure the performance of both the LBO and the SBO and to compare them. As shown in figure 6 the first reference point is the entry year of the first private equity firm into the investment. The next reference point in order to measure the changes in the operating performance is the year when the first private equity firms exits and the second private equity firm enters the investment. Finally, the last reference point is when the second private equity firm exits the investment. 4 For the exit of the second private equity firm this requirement is relaxed slightly. Hence, the decision rule applied is that portfolio companies held by the company above an empirically based threshold will be included along with the exited observations. A comprehensive argumentation of the sample creation is provided in the chapter on data collection. 35

38 Figure 6: Illustration of study design Source: Authors contribution The use of reference points and the matching portfolio companies in the study design applied in this thesis differs from the ones applied in the existing literature in a couple of ways. Previous studies have either been looking at operational performance changes in buyouts in general including SBOs but not SBOs solely, e.g. Gaspar (2012) or that the SBOs are examined relative to LBOs without matching them one- to- one, meaning that the number of LBOs examined are higher than SBOs as there are more LBOs than SBOs to evaluate, e.g. Wang (2012). In comparison, the observations examined in this thesis are only SBOs with a corresponding LBO, which causes the number of LBOs and SBOs to be the same. Another parameter where this thesis stands out from the majority of other studies examining operational performance, in buyouts in general and SBOs, specifically, is on the time horizon over which the performance is evaluated. The common way to examine the performance changes in the existing literature is to look at performance changes in shorter periods around the year of the buyout. The length of the periods varies from 1 year to 5 years, for instance change from 2 years before the buyout to the year before the buyout and from the year before the buyout to 4 years after the buyout (Kaplan 1989a; Gaspar, 2012). In the study design deployed in this thesis, we compare ownership periods between two reference points regardless of their length. There are studies in the academic literature that examines the full period of ownership (Gou et al., 2011; Freelink and Volosovych, 2012). This method is similar to the one applied for both periods we compare in the study design. 36

39 5. Hypothesis development In the literature review provided earlier, we explained how recent empirical papers on secondary buyouts are characterized by conventional wisdom; namely that the first private equity firm to own an asset has already exploited that asset for easiest accessible improvements on operational performance, and hence left the portfolio company with limited or no potential for further improvements. In that regard, we now develop four hypotheses inspired by the literature on operational performance enhancement in buyouts and based on the operational performance framework illustrated by the ROIC tree. The hypotheses are thus formulated based on the operational value creation potential in secondary buyouts compared to the first buyout in order to contribute to the current academic debate. We start out by the topline in the financial profit and loss statement. Here Brigl et al. (2012) and Kengelbach et al. (2013) argue that revenue growth is essential to operational value creation in leveraged buyout transactions in general. However, while scholars such as Gaspar (2012) and Kaplan (1989a) commonly acknowledge that revenue growth increases during the first round buyout, there is yet no consensus regarding the sales growth potential for secondary buyouts. Evers and Hege (2012) argue that LBOs offer higher sales growth than SBOs while Achtleitner and Figge (2012) find that sales growth in SBOs is equal to sales growth in LBOs. Finally, Wang (2012) contributes to the ambiguity by showing that topline growth increases more during the SBO, relative to the LBO. We follow Wang (2012), in developing our hypothesis for revenue growth, because private equity firms in general do not operate with standardized approaches to topline- initiatives (Brigl et al., 2012). This arguably means that the growth potential of the portfolio company has not been fully exploited during the first round buyout. This has lead us to suggest the following hypothesis Growth hypothesis: Compared to LBOs, revenue growth is higher in SBOs. Nor is there any consensus yet in the empirical literature on secondary buyouts when it comes to improving the profitability of the portfolio company. Several scholars have argued that profit margin and EBIT to assets improves during the first round buyout (e.g. Kaplan, 1989a; Bull, 1989; 37

40 Muscarella and Vetsuypens, 1990; Gaspar, 2012). These improvements are partially based on reduction of overhead costs as suggested by Phan and Hill (1995) and growth in revenue as covered above. Logically, however, the largest improvements in terms of overhead optimizations are expected to be realized in the first period. In the context of secondary buyouts, Achtleitner and Figge (2012) argue that margin improvements in SBOs are equal to margin improvements in LBOs whereas Evers and Hege (2012) find that margin improvements are higher in SBOs, compared to LBOs. Interestingly, Bonini (2012) finds abnormal margins during the LBO, but he also shows that margins revert to industry averages during the secondary buyout. Hence, we propose the following hypothesis for changes in profitability in secondary buyouts Profitability hypothesis: Compared to LBOs, margin improvements are lower in SBOs. Private equity firms are also expected to be able to create operational value by focusing on the financial structure of the portfolio company (Brigl et al., 2012). This typically implies a focus on engaging in optimization of working capital, which includes inventory, receivables and payables management (Brigl et al., 2012). For first round buyouts, Smith (1990) finds that the ratio sales to WC increases post buyout around the buyout year and Gaspar (2012) show that WC to sales is reduced during leveraged buyouts. Both studies are thus finding evidence for improvement in improved working capital management. However, only Evers and Hege (2012) have studied WC improvements in a SBO context. They find that there are no differences in WC improvements between the LBO and SBO. Moreover, the first private equity firm is expected to professionalize the working capital management, leaving incremental improvements for in the portfolio company for the second round. Therefore, we test the following hypothesis Working capital management hypothesis: Compared to LBOs, reductions in working capital are lower in SBOs. Private equity firms also engage in fixed asset optimization when working with the financial structure of the portfolio company. This includes decisions regarding capacity, leasing, financing, and improved asset utilization (Brigl et al., 2012). The most often suggested way to improve asset 38

41 utilization is to eliminate unproductive assets or sell off assets that are more valuable outside the company (Phan and Hill, 1995; Gou et al., 2011). In that regard, Gaspar (2012) finds that operating assets to sales increases during the first round buyout and we therefore suggest that Fixed asset utilization hypothesis: Compared to LBOs, reductions in capital tied in PP&E are lower in SBOs. 39

42 6. Data collection and applied method A main issue when studying private equity is the availability of data and disclosure of information. Privately held companies do not have the same obligations of sharing information as publicly traded companies (Sousa, 2010). Researchers measuring performance normally prefer market- based measures to accounting figures (Bull, 1989). Since our sample is on private companies, market data is naturally limited, and hence we use information and data available from annual reports and databases. Both the selection of the sample and the actual collection of the data have been subject to a number of decisions and the following will provide a comprehensive description of the processes. In the remainder of this chapter, we present the data collected and the choices made regarding the method applied. First, we will outline the criterions for the observations that go into the final sample. Hereafter, we serve an explanation for the data collected in the various observations Sample selection In this section it is explained, step- by- step, how we got from our gross sample to the final sample, which we will use for our analysis in chapter seven. A waterfall showing the steps in the sample selection process is illustrated in figure 7. The initial sample was comprised by data from two M&A databases, Mergermarket and Zephyr published by Bereau van Dijk, commonly used by academic scholars (Bonini, 2012; Wang, 2012; Freelink and Volosovych, 2012). To limit the search three criterions on, I) geographic scope, II) time period and III) deal characteristics, were set up 5. The Mergermarket database, which has been our primary source of transactions, showed a search result of 227 transactions. We used Zephyr for crosschecking and validation of data as it has a good coverage of European deals (Wang, 2012). Filtering for the same criterions in Zephyr as in the Mergermarket database resulted in 250 observations. These transactions defined the initial sample. To get from the initial sample to private equity driven SBOs we applied the following definition SBOs are leveraged buyouts in which both the buyer and the seller are private equity firms. 5 Criterions for the search were deals covering Denmark, Norway and Sweden in a period covering 1998 to April 10 th 2014, and finally the deal had to be a secondary buyout. 40

43 Private equity firms are characterized by investing in private companies, performing active ownership, and exerting relatively short periods of ownership (Bennedsen et al., 2008). A commonality for the two data sources applied is that the definitions of secondary buyouts and private equity sponsors are rather broad. Observations that violated the above definitions were thus discarded. In addition, we removed some of the observations that were characterized by majority ownership stake by others than the private equity firms, such as MBOs or MBIs, following the definition of Evers and Hege (2012). The application of a tighter definition of SBOs reduced the sample from 250 observations to 153. The major difference between the existing literature and our sample is that we consider observations where the SBOs are exited. Similar to the study design of this thesis a few academic scholars have examined portfolio companies owned by private equity firms in two consecutive periods (Bonini, 2012; Wang, 2012). Furthermore previous studies evaluate operational performance and value creation by comparing changes in years around the transaction year (e.g. Kaplan 1989a; Wang 2012) or by looking at changes from the beginning of the holding to the end of it (Gou et al., 2011; Freelink and Volosovych, 2012). This thesis compares two full private equity ownership periods in which the private equity firm, theoretically, will be able to enhance operational performance throughout the entire holding period. To get from private equity driven SBOs to SBOs exited we thus discarded all observations where the second private equity firm to own the portfolio company was still involved. Applying this criterion reduced the sample considerably from 153 to 42 observations. 41

44 Figure 7: The process from initial to final sample. Sample Selection Process Sample Size Initial sample Private Equity driven SBOs SBOs exited SBOs exited or SBOs with a holding period above 4 years 74-2 Sample before data collection 72-9 Final sample 63 Source: Authors contribution We consider a sample of 42 observations in a sampling process not yet finalized to be potentially critical. Hence for the next step we decided to increase the sample by adding back observations that had not yet been exited. The approach of using unrealized deals to increase the sample size follows the method applied by Achleitner and Figge (2012) who also examine changes in operational performance in private equity owned companies. To get to SBOs exited or with a holding period above 4 years we utilized the 42 exited SBOs already identified. From those we computed a median (average) holding period of 4 years (4.5 years). We applied the median as the criterion for not exited SBOs to add back, as it mitigates the effect of outliers compared to the average. Applying a holding period of 4 years is a realistic proxy and supported by research conducted by the private equity intelligence site Preqin, showing average holding periods of 42

45 approximately 4 years since 2006, and by large Nordic private equity firms such as EQT and Nordic Capital who expects holding periods in a range around 4 years for their investments 6. In terms of holding periods of SBOs specifically, Freelink and Volosovych (2012) and Achleitner and Figge (2012) examine samples with median holding periods of 3 years and 3.3 years, respectively. Adding the not exited SBOs with a holding period of 4 years or more to the SBOs exited makes a preliminary sample of SBOs exited or with a holding period above 4 years containing 74 observations. We removed two observations in order to get to the sample before data collection. For those observations the year of entry was equal to the year of exit 7. This makes it practically impossible to measure any effect from the change in ownership. Furthermore, we argue that changes in the operational performance or the potential hereof are not the drivers of those deals. This brings the sample before data collection down to 72 observations. The final adjustment made to reach the final sample was made after actually collecting the data input for the variables for the analysis. We found that in nine of the observations accounting data was not sufficiently available for the three accounting years needed in order to compute the variables 8. Discarding the aforementioned observations left us with a final sample size of 63 observations. A list of the portfolio companies in the sample can be found in appendix 2 in the tables 21 and 22. In addition to the company name, the list further contains information on company IDs, the origin country of the companies and the years of the reference points we used to evaluate the operational performance of the companies. 6 ( and ( 7 The two observations are Nedermand Holding (SE) and Papyrus (SE). 8 The observations removed due to problems with the data material are Gislaved Folie AB (SE ), Modul- System HH AB (SE ), Findus AB (SE ), Tritech Teknik AB (SE ), Carpark AB (SE ), Perstorp Group (SE ; SE ), Constructor Group AS (NO ), EET Nordic Group A/S (DK ) and Green House of Scandinavia A/S (DK ) 43

46 6.2. Data collection While the previous section served the purpose of explaining how we ended out with the observations to go into the final sample, this section will provide an outline of how the individual observations have been treated Specifying the relevant years of the observations The list of SBOs obtained through the Mergermarket database includes the date of completion for the transactions. The information on the actual date of exit of the first private equity firm and entry of the second private equity firm was used to apply a simple rule making it possible for us to use data for the year of the transaction. The rule state that if the transaction is completed in the first half of the year, i.e. before July 1 st, we have used the accounting data from the accounting year before the transaction, T- 1, and correspondingly we have used the accounting year of the transaction, T, if the completion date is in the second half of the year. Kaplan (1989a) and others are omitting the year of the transaction because they can be difficult to interpret. Our motive, however, for applying the above method is that we are not looking at performance around the buyout year but between two ownership periods. The timing difference between the transaction and the accounting period is thus six months at maximum and hence the numbers should be at least as- or more accurate compared to omitting the entire year. Bonini (2012) who are looking at consecutive buyouts between private equity firms presents an opposing view to the one of Kaplan (1989a) as he argues that: ( ) the extant literature on LBO performance provides strong evidence that most of the performance change is achieved in during the first 2 years, including the acquisition year ( ). (Bonini, 2012, pp ). Furthermore, according to Kaplan and Strömberg (2009) and Gou et al. (2011) evidence shows that the first two years of ownership, including the year of the transaction, are important in terms of changes in operational performance. In addition to that, private equity firms often operate with 100- days planning, which suggest that important actions are taken early and hence the first accounting year can be relevant when investigating the operational performance. A discussion of this method and implications hereof is covered later. For the entry of the first private equity firm and the exit of the second private equity firms we have collected data on the transaction year from the websites of the private equity firm or 44

47 websites monitoring private equity such as unquote.com or argentum.no. In some cases however, we use data of an adjacent year to the transaction year as a proxy in order to be able to use the observation. There are three specific cases where a proxy is applied, I) if the observation is one of the not- exited observations, we use data of the most recent annual report for the exit data, II) if the financial statements of the transaction year diverge from twelve months we use data of the an adjacent year and III) if the observation is exited in 2013 or 2014 we use the data of the most recent year available. The rationales behind using proxies in the above three types of cases is that for recent observations we cannot access better data than the latest published. For observations where we do not have detailed information on the time of the transaction an adjacent year is a better alternative than to remove the observation as it can be equally accurate as using the transaction year. The relevant observations exposed to the use of proxies are marked by an asterisk in the relevant years in table 21 and table 22 in appendix Accounting data In order to get the right data for studying operational performance we used the official sources of annual reports and accounting data in the Denmark, Sweden and Norway to determine and validate the relevant companies and their company IDs 9. The actual data have been collected from hard copies of the annual reports of the companies, Amadeus via Orbis 10 published by Bureau van Dijk, which is a commercial database with accounting data for companies globally, and Proff.no, a Norwegian website that offers financial statement data from the official source, Brønnøysundregistrene. Preferably we would have relied on one source of data for the collection. However, we experienced in line with Wang (2012) that databases such as Amadeus does not necessarily provide complete coverage. Thus, the use of different sources for data collection has merely been a matter of accessibility. We have kept the number of sources at a minimum to ensure simplicity, transparency and quality. The company IDs for the financial statements used is listed in table 21 and table 22 in appendix 2. 9 Erhvervs- og selskabsstyrelsen ( and Navne og numre in DK ( Bolagsverket in SE ( and Brønnøysundregistrene in NO ( 10 For all Orbis data the Global standard format has been used. It has the advantage of a standardized template that makes it easier to compare across accounting principles. 45

48 In table 21 and table 22 in appendix 2 it is furthermore specified whether the data collected is from consolidated or unconsolidated financial statements for the specific observations. The existing literature on operational performance in buyouts is ambiguous in terms of whether to use consolidated or unconsolidated financial statements. Whereas Wang (2012) and Freelink and Volosovych (2012) are stressing the importance of solely using consolidated data, Sousa (2010) advocates for using unconsolidated financial statements over consolidated if these are available. Gaspar (2012) uses both consolidated and unconsolidated financial statements but controls for the effect hereof, whereas other studies using accounting data to examine changes in operating performance are not discussing this matter (e.g. Kaplan, 1989a; Bonini, 2012). We have used the consolidated financial statements if these were accessible over the three years we have been looking at. However, in some cases consolidated statements have either not been available or have not contained any operating figures in the income statement. In these cases we have used the unconsolidated financial statements for the company in which the data on operations of the company was reported. The method of using different types of statements and implications hereof will be discussed later Measuring performance According to Barber and Lyon (1996) the task of measuring performance and evaluating it has three critical elements. These elements are, using the right measures, applying a benchmark to control for exogenous effects and choosing the appropriate statistical test given your research design. In this section we will cover the first two elements thoroughly, whereas the latter will be discussed more comprehensively when outlining the statistical approach of the thesis. In the literature on operational performance in buyouts, the earnings before interest and tax (EBIT) or the earnings before depreciation, amortization, interest and tax (EBITDA) is commonly used as the main measures. In this thesis we follow the literature and uses EBIT (e.g. Kaplan, 1989a; Bonini, 2012; Wang, 2012; Freelink and Volosovych, 2012). As noted we alternatively could have used EBITDA, which is equally popular in the literature (e.g. Achleitner and Figge, 2012; Wang, 2012). EBITDA has some useful propositions in terms of approximating cash flows, which ultimately determines the value of the company. It can however be problematic to assess the profitability using EBITDA as it includes the revenue from using the resources of the company but at the same time neglects to consider the cost of resources used in the operations by omitting the 46

49 depreciation and amortization (Petersen and Plenborg, 2012, pp ). In addition to the caveats that EBITDA impose, access to depreciation and amortization details were not explicit for all observations across the data sources we used, which ultimately made us leave out EBITDA in the analysis Computation of performance measures To make proper comparison of operational performance across observations or time the measure must be scaled (Barber and Lyon, 1996). Using financial ratios makes it possible to compare across multiple periods (Petersen and Plenborg, 2012, p. 63). In the existing literature performance measures have been scaled with different figures such as total assets, operating assets and sales (e.g. Kaplan, 1989a; Smith, 1989; Wang, 2012). We will primarily be using operating revenue as the measure we scale by. This has certain advantages over e.g. total assets, as total assets do not only relate to the operations. Furthermore, using operating revenue ensures that both figures are from the income statement when looking at EBIT and hence potentially better matched (Barber and Lyon, 1996). In terms of disadvantages of using the EBIT to operating revenue, as a measure of operational performance, is that it does not take into account the productivity of assets, e.g. the firm s ability to increase operating income without changing the operating assets in place (Barber and Lyon, 1996). To accommodate this, changes in EBIT will be evaluated by scaling it to total operating assets in addition to revenue. For other measures of performance we will take into account the utilization of the current and non- current operating assets. Scaling the variables creates financial ratios, which is generally preferred over absolute values. The use of percentages or ratios can however be problematic if the measure is negative. A negative ratio can lead to results being nonsensical when looking at percentage changes over a period. This forces the researcher to discard the relevant observations, which ultimately reduces the statistical power and imposes a bias as it for instance ignores bad performing companies in the case of a negative EBIT (Barber and Lyon, 1996). In the collection of data, we have not disregarded negative values, as this would reduce the size of the sample critically. Instead of leaving out these observations we have looked at the absolute changes between the ratios of the financial performance from the entry to the exit causing the change in operating performance in the LBO and the SBO to be measured in 47

50 percentage points. When comparing the SBO and the corresponding LBO, we subtract percentage points from each other implying that the measure of these changes are differences between percentage points. This method has the benefit that it does not force us to disregard observations and reducing the statistical power as well as imposing bias from leaving out variables with negative EBIT or negative working capital Variables In order to test the hypotheses presented earlier we have computed variables on the basis of profitability and operating performance. ROIC can be considered as the overall profitability for operations and can be decomposed into profit margin and turnover rate of invested capital 11 (Petersen and Plenborg, 2012, p.94). In addition ROIC is also the accounting equivalent of IRR (Petersen and Plenborg, 2012, pp ). IRR is a key performance indicator of private equity investments, which makes it highly relevant to be looking at the components of ROIC in this context. The variables we use to test for changes in operational performance between the first- and second private equity owner are the accounting figures that go into the calculation of the profit margin and the turnover rate of invested capital. Furthermore, academic scholars studying changes in operational performance associated with buyouts largely inspire the variables we use. In the following we go through the variables for the analysis Growth Growth in revenue is measured by using the cumulated annual growth rate (CAGR) 12 in revenue. This variable is included to show whether the revenue of the portfolio company is growing relatively more over the holding period in the first- or second private equity ownership period. The revenue is linked to the competitive edge of the portfolio company and is determined by the size of the market, the market share of the company as well as the price and product mix of the company (Petersen and Plenborg, 2012, p.94). The market share can be grown organically by expansion or inorganically by acquisitions of companies to add- on to the portfolio company as well as mergers. In the academic literature, on SBOs focusing on operational performance, growth in 11 Profit margin = (EBIT/Net revenues) x 100. Turnover rate of invested capital = (Net revenue/invested capital) x 100 (Petersen and Plenborg, 2012, pp ) 12 Formula used for CAGR in revenue = (Revenue in the year of entry/revenue in the year of exit)^(1/number of years in the holding period)

51 revenue is often included in the analysis (e.g. Achleitner and Figge, 2012; Evers and Hege, 2012; Wang, 2012). Topline growth is reflected in ROIC through the profit margin. As mentioned above the growth is measured in percentage over the holding period of the LBO and the SBO, which implies that the difference between the two ownership periods is in percentage points Profitability. As stated above we will be evaluating the changes in EBIT in two ways. First, we will investigate the profit margin, which is the EBIT to revenue ratio, and the earnings component of ROIC when decomposed. This variable is interesting, as it takes into account both the revenue and the cost of doing operations such as the cost of goods sold, marketing, distribution, administration as well as depreciation and amortization (Petersen and Plenborg, 2012, p. 94). This variable is thus included to show if there is a change in the portfolio companies ability to optimize the cost side of the business without affecting revenue over the holding periods and increase the overall profitability. The profit margin is used by most of the academic papers that examines operational performance in buyouts (e.g. Kaplan, 1989a; Bonini, 2012; Wang, 2012; Freelink and Volosovych, 2012). The second measure of changes in EBIT, we examine, is the return on operating assets, which is calculated as the EBIT to total operating asset ratio. The return on assets is commonly used as a measure of change in profitability (Kaplan, 1989a; Wang, 2012). We use operating assets 13 to scale by, inspired by Smith (1990), as this is considered to be the assets that drive EBIT (Petersen and Plenborg, 2012, p.25). The changes in profitability ratios are measured by percentage point in the LBO and the SBO. This is due to the potential of negative EBIT figures. This implies that the difference in operating performance changes between the two types of buyouts is presented in difference between percentage points Working capital management. Working capital 14 is included to analyze how much the portfolio companies optimize the capital required to run the business in the LBO versus the SBO. The working capital is measured relative to revenue following Gaspar (2012), which implies that the lower the ratio, the more the portfolio company has optimized working capital. The change in working capital is often mentioned as a 13 Operating assets are here computed as total assets less financial assets. 14 We have computed working capital as current assets current liabilities. 49

52 lever of operational performance in the literature however omitted in the actual analysis (e.g. Gou et al., 2011; Freelink and Volosovych, 2012). A few studies are including working capital. Smith (1990) studies changes in the working capital turnover in buyouts and Evers and Hege (2012) also investigate changes in working capital. The measure of changes in working capital relative to revenue is chosen because looking at the working capital turnover like Smith (1990) imposes a challenge when facing negative working capital. A negative working capital can be a good sign for a company if it is caused by the firm s ability to tighten the receivables period, extending the payables period and keeping inventories down. However the interpretation of the working capital turnover is that the higher the turnover, the more efficiently the working capital is managed. Hence using the measure with negative working capital will result in nonsensical results. Working capital is the current part of the invested capital, which is the asset component of ROIC, decomposed. The changes in working capital management are measured by percentage point in the LBO and the SBO, as working capital can obtain negative values. The way of measuring implies that the difference in working capital to revenue changes between the SBO and the LBO is stated in difference between percentage points Fixed asset utilization. The measurement of the fixed asset utilization is property, plant and equipment (PP&E) to revenue and follows Gaspar (2012). The variable is included to examine if there are difference between the changes in the utilization of fixed assets over the holding periods in the LBO and SBO, respectively. Wang (2012) also include fixed assets but uses it as a size measure. Whereas looking at working capital was the current asset part of invested capital, fixed assets can be considered as the non- current part of invested capital (Petersen and Plenborg, 2012, p.94). The changes in the PP&E to revenue ratios are measured by percentage point in the LBO and the SBO. The variable is not taking on negative values but is measured in this way for consistency in the measurements for the ratios used in the analysis. As for the other variables, this means that the difference in fixed asset utilization between the two types of buyouts is presented in difference between percentage points. 50

53 Adjusting for the industry effects The purpose of this thesis is to study whether performance of portfolio companies are significantly better or worse in one period compared to the other. To assess this, it is important to ensure comparability over the observations, which can be done by controlling with an industry benchmark. Operating performance measures are based on accounting numbers and are generally evaluated relative to an industry benchmark. (Barber and Lyon, 1996, p. 360). By matching the companies in the study with a benchmark, we adjust for exogenous effects such as the economic environment and industry specific trends (Barber and Lyon, 1996). There are numerous ways of adjustments applied in the literature. To make the adjustment for industry effect we have applied an approach similar to Smith (1990) and used averages of accounting figures for the firm specific industries in the respective countries to compute adjustments. The industries were identified by 4- or 3- digit NACE codes 15 for the company across all observations, obtained through the Amadeus database. Hereafter, calculations of industry adjustments specific to Denmark, Norway and Sweden in the years we evaluate the operational performance were made using the same approach to computations as we did for the variables. In computing the industry adjustments we ignored all companies in the given industry where data was not applicable or had no revenue as we assumed that these were non- operating firms. For industry adjustments prior to 2004 we have used Danish data for industry adjustments across Denmark, Norway and Sweden because we only had access to Danish data through Amadeus for the period prior to Statistical Method In this section we, first, discuss our choice of the particular statistical significance test applied and, secondly, the hypothesis testing will be outlined mathematically. 15 NACE is a European industry classification system, where the first four digits are shared in all European countries. The NACE codes are equivalent to industry classification systems such as SIC and NAICS ( 51

54 Choice of significance test We use the nonparametric Wilcoxon signed- rank test 16 following other scholars who also studied the operational performance of private equity owned companies (e.g. Kaplan, 1989a; Gou et al., 2011; Smith, 1990; Wang, 2012). Furthermore, Barber and Lyon (1996) argue that the Wilcoxon signed- rank test is more powerful than the paired- sample t- test when examining measures of operating performance because these often include extreme observation. Test assumptions Agresti and Franklin (2009) describe the assumptions of the Wilcoxon signed- rank test, in the following way, which we will discuss below in relation to our study Random sample of matched pairs for which the differences of observations have a symmetric population distribution and can be ranked (Agresti and Franklin, 2009, p. 753) Random sampling is required in order to obtain a sample that represents the total population of secondary buyouts in the Nordics. In order to take this into account, we include secondary buyouts from Sweden, Norway and Denmark from diverse industries and over a long period of time. We also include secondary buyouts involving the management team backed by private equity firms, and we do not distinguish between successful and unsuccessful secondary buyouts. Secondly, the Wilcoxon signed- rank test is appropriate to use when analyzing matched pairs, i.e. the same participants, under two different conditions. We are thus comparing the respective target companies in our sample using repeated measures for the first round buyout and the secondary buyout, respectively. In addition, because we use relative measures for operational performance it means that the differences observed for our repeated measures, under the LBO and the SBO, can be ranked which is also an assumption of the Wilcoxon signed- rank test. 16 Nonparametric significance tests do not assume a particular form of distribution, such as the normal distribution (Agresti and Franklin, 2009, p. 734). We tested our data and concluded that it is not approximately normally distributed, which made us use the Wilcoxon signed rank test in line with similar studies. 52

55 Thirdly, the differences of observations must have a symmetric population distribution. In appendix 3 we show that our data meets this assumption by providing histograms, for each measure of operational performance, which show the distribution of the differences of observations 17. Hypotheses The Wilcoxon signed- rank test assumes the following two- sided hypotheses 18 Franklin, 2009, p. 753). H 0 : Population median of difference scores is zero H a : Population median of difference scores in not zero (Agresti and Hypotheses testing In order to explain what goes into the hypotheses testing this subsection will outline the mathematical steps required by the study design to get to the numbers tested, following the approach of Gaspar (2012). We let y!,! denote the measure of performance of the j portfolio company in the LBOs. The time of the measure is indicated by t, where t Entry, Exit. Likewise the measure of performance for the j portfolio company in the SBO is represented by x!,!. The changes in the performance measure are thus given by y! = y!,!"#$ y!,!"#$% x! = x!,!"#$ x!,!"#$% To test for the difference in changes over the two private equity ownership periods the Wilcoxon signed rank test, tests if the median of differences is significantly different from zero. To be able to test for this the differences must be defined and computed as represented by z! = x! y! 17 The histograms shown in appendix 3 are for the full sample and are merely included for illustrative purpose. 18 The algorithms used by SPSS to compute the test statistic and p- value, amongst others, can be found in appendix 3 53

56 To control for industry specific and economic trends we subtract the corresponding benchmark!"#$%"& denoted y!,!!"#$%"& for the LBO and equivalently x!,! for the SBO. This gives the following equations for the changes in the LBO, y!"#$%&'"!"#$%"&!"#$%"&! = y!,!"#$ y!,!"#$ y!,!"#$% y!,!"#$% Similarly, the changes in the SBO can be represented by x!"#$%&'"!"#$%"&! = x!,!"#$ x!,!"#$!"#$%"& x!,!"#$% x!,!"#$% These equations expresses the changes in the performance measure during the ownership periods and to test whether there are differences between the changes in the two periods the using the above notation means that z!!"#$%&'" = x!!"#$%&'" y!!"#$%&'! The above equation describes the difference between the changes in the operational performance over the two periods. Applying the notation defined above for testing implies that H! = median! z!!"#$%&'" = 0 H! = median! z!!"!"#$%& 0 The null hypothesis, H 0, state that changes in the operational performance of the representative, median, SBO is not different from the changes in operational performance of the median LBO. The alternative hypothesis, H A, is that the changes in operational performance between the two periods are different. 54

57 Testing in practice In order to carry out the testing of our hypotheses, we use IBM SPSS Statistics 21 for which access is provided by Copenhagen Business School. The results of the hypotheses tested will be reported in chapter 7. We inform that output from SPSS, which we have not included in the tables provided in chapter 7, will be available in appendix 4. We engage in a discussion of the results in chapter 8. 55

58 7. Results and analysis In this chapter we present the results on changes in operating performance between the LBOs and SBOs. First, we present descriptive statistics on the sample and the variables that go into the analysis. Next, we test for changes in operating performance on the full sample as well as different subsamples Descriptive Statistics The following presents tables and figures to provide an overview of the sample given its characteristics. Table 2: Distribution of the sample by country of the portfolio company and the ownership status of the SBO The table shows the number of firms in the sample divided by country and ownership status as absolute numbers and percentages of the full sample. Denmark Norway Sweden Total Exited (54%) Not Exited (46%) Total 9 (14%) 18 (29%) 36 (57%) 63 (100%) Table 2 presents the sample of portfolio companies across country of headquarter location as well as ownership status of the SBO. The sample comprises 63 firms with characteristics to qualify for the sample as covered in the chapter six above. As argued earlier Sweden is by far the country with the most developed private equity industry, which is also reflected in this sample with 57% of the observations being Swedish companies. The rest of the sample entails 29% Norwegian companies and 14% Danish. Across countries we see that the not exited SBOs in the sample are spread almost evenly if the size of the respective buyout markets is taken into account. In addition, none of the countries have a predominance of not exited SBOs. 56

59 Table 3: Timing of the transactions in the sample The table shows the number of transactions in a given year by the nature of the transactions. For LBO- and SBO transactions the total number is 63 whereas the SBOs exited sums to a total of 34. Year of Transaction LBO transactions SBO transactions SBOs exited Table 3 exhibits the distribution of the transactions of the firms in the sample over time. As the SBO transactions is a function of the LBO transactions and SBOs exited is a function of SBO transactions there is a natural pattern in the numbers. There are however indications of buyouts happening in waves, which can be seen by the number of LBO transactions in the periods and as well as the number of SBO transactions prior to the financial crisis, between

60 Figure 8: Exit route for SBOs The diagram presents choice of exit for the second private equity firm to own the portfolio company. The total number of exited SBOs in the sample is 34 firms. 1; 3% Exit Route for SBOs 12; 35% 19; 56% Private equity IPO Trade sale Bankruptcy 2; 6% Figure 8 shows the exits routes chosen by the second private equity firm to own the portfolio company. Interestingly the preferred route of exit in the sample is to a third private equity firm, known as a tertiary buyout, with a frequency of 56% of the observations. The second largest source of exit is trade sales meanwhile IPOs are almost as rare as bankruptcies in the sample. This distribution in the choice of exit is in line with the pattern in data covering the entire buyout market (EVCA, 2014). Kengelbach et al. (2013) who have a majority of tertiary buyouts in their sample discuss tertiary buyouts as a viable exit strategy and finds that the return on the tertiaries is close to the ones of trade sales. In contrary, Nikoskelainen and Wright (2007) argue that exits by IPO and trade sales offer higher returns than secondary buyouts. The reason for the low number of IPOs, among SBOs in the sample, might therefore be due to cold IPO markets or because these specific secondary buyouts are less well suited for IPOs. 58

61 Figure 9: Comparison of geographic origin in LBO versus SBO The chart shows the geographic origin of the private equity firm in LBOs and SBOs, respectively. Origin of the Private Equity Firms Nordic Not Nordic LBO SBO Figure 9 presents the geographic location of the private equity firm in the LBO and SBO, respectively. As it can be seen the Nordic private equity firms outnumber the Not Nordic for both types of buyouts. The differences are however less evident for the SBOs. Nordic private equity firms having a superior knowledge about good investment opportunities locally in the first buyout round might explain the development in the geographic origin of the private equity firms. After being owned by a private equity firm already, the awareness of the company grows. This might be the explanation of the increase in numbers of Not Nordic private equity firms to acquire Nordic companies in SBOs. 59

62 Figure 10: Geographic origin of the private equity firm in the LBO and the SBO. The diagram shows the geographic origin of the private equity firm tracked over both the LBO and the SBO to see if the geographic base of the private equity firm changes. Change in the Origin of the Private Equity Firms 13, 21% 21; 33% 10; 16% 19; 30% Nordic; Nordic Nordic; Not Nordic Not Nordic; Not Nordic Not Nordic; Nordic Figure 10 shows the geographic origin of the private equity firms in both the LBO and the SBO, i.e. if the private equity firms to own a portfolio company are Nordic in both the LBO and the SBO or Not Nordic in both the LBO and in the SBO. Likewise, it can be seen whether the geographic origin changes between the LBO and the SBO, i.e. if the private equity firms are Nordic in the LBO and Not Nordic in the SBO and vice versa. In line with the conclusion from figure 9, the most frequent combination, 33 % of the sample, is that the private firms are Nordic based in both the LBO and the SBO. Interestingly, the diagram shows that the number of firms where the geographic origin changes between the LBO and the SBO, i.e. Nordic to Not Nordic or Not Nordic to Nordic, is greater than the number of firms where the origin of the private equity firm do not change. The distribution is however 51 % of the firms experience a change in the origin of the private equity firm whereas for the remaining 49 % of the observations the firms continue to be owned by a Nordic based private equity company or a Not Nordic private equity firm. 60

63 Table 4: Distribution of sample by industry The table shows the industries represented in the sample. 19 NACE Rev. 2 industry classification # of observations % of total sample Manufacturing 24 38% Construction 1 2% Wholesale and retail trade 11 17% Transportation and storage 1 2% Information and communication 3 5% Financial and insuarance activities 2 3% Professional, scientific and technical activities 12 19% Administrative and support service activities 6 10% Human health and social activities 1 2% Other service activities 2 3% Total % Table 4 presents the distribution of the industries in which the firms in the sample operate. Manufacturing is the most frequent industry among the firms in the sample accounting for 38% of the total. Wholesale and Professional, Scientific and Technical activities accounts for 17% and 19%, respectively, while the rest of the companies in the sample belongs to industries covering the last 26%. This composition is similar to the ones of other including this information (Smith, 1990; Wang, 2012; Gaspar, 2012). 19 The categories are based on NACE rev. 2 industry classification and are adopted from Eurostat ( RA /EN/KS- RA EN.PDF p. 57) 61

64 Table 5: Summary statistics for median, mean and Std. Dev of firm characteristics at the points of reference for the full sample The table presents the firm characteristics at the time of the transactions. Revenue and total assets are in millions DKK. The rest of the measures are ratios. Number of obs. (N =63) Entry of LBO Exit of LBO/Entry of SBO Exit of SBO Median Mean Std. Dev. Median Mean Std. Dev. Median Mean Std. Revenue (MDKK) Total assets (MDKK) EBIT/Revenue EBIT/Total operating assets Working capital/revenue PP&E/Revenue Dev. Table 5 is an overview of the sample at the three reference points where the performance is evaluated. The revenue and the total assets are size measures and the figures shows expectedly an upward trend in the size of the portfolio companies over time for both the median and mean of the companies. The remaining figures in the table are measures of the ratios we use in the hypotheses. 62

65 Table 6: Summary statistics for median, mean and Std. Dev of changes in the ownership periods of the firms in the full sample The table presents sample characteristics for changes in the operational performance in LBOs and SBOs over their holding periods. The change in CAGR in revenue is measured in percentage points whereas the other measures are changes between percentage points. Number of obs. (N =63) LBO SBO Median Mean Std. Dev Median Mean Std. Dev Change in CAGR in revenue Change in EBIT/Revenue Change in EBIT/Total operating assets Change in Working Capital/Revenue Change in PP&E/Revenue Table 6 shows the summary statistics for changes in the operational performance in LBOs and SBOs, respectively. It can be seen that both the median portfolio company and the mean of portfolio companies increases revenue over the period, however the increases appears to be bigger in the first period compared to the second period. Studies investigating SBOs have shown that there is no difference in sales growth between the LBO and the SBO (Achleitner and Figge, 2012; Kengelbach et al., 2013). The standard deviations show that variations are smaller in the SBO than in the LBO. In terms of profitability, the summary statistics also suggest that the profit margin and the return on operating assets increases more in the LBO than in the SBO. The utilization of working capital presents ambiguous figures for the sample firms, as the median change is negative for both types of buyout whereas it is positive for the LBO and negative for the SBO, when looking at the mean changes. The lower the number, the better the portfolio companies have been at optimizing working capital relative to revenue. The same interpretation holds for PP&E to revenue. The summary statistics indicate that the portfolio companies are better at utilizing fixed assets in the LBO when compared to the SBOs. Whether these differences are statistically significant will be examined in the section below. 63

66 7.2. Test results In this section we provide the results of the statistical tests conducted. First, we handle the results without adjusting our data for industry specific effects. We are aware of the limitations of these results; namely that they may reflect the economic development or other particularities in the respective industries of the portfolio companies rather than the change in private equity ownership. Secondly, we provide the results for our total sample after adjusting for industry specific effects. After that point, we rely only on adjusted data in the remainder of the testing. We round off this chapter by presenting test results for subsamples based on different contingencies given by the nature of our sample. For the results provided in the remainder of this chapter we underline that for growth and profitability measures, a positive value in median of all differences indicates that the SBO is increasing the particular performance measure more than the corresponding LBO. For working capital management and fixed asset utilization, a negative value in median of all differences means that the SBO is decreasing the particular performance measure more than the corresponding LBO. Note that for working capital management and fixed asset utilization a decrease in the performance measure implies an improved performance. The test results will be presented in tables throughout the section and we interpret and comment on significant results. In chapter 8 we engage in a more thorough discussion of our findings. 64

67 Total sample Table 7: Changes in operational performance for the full sample of unadjusted data The table reports median percentage point changes in CAGR in revenue. The remaining measures of the changes in operational performance are median difference between percentage points. Significance levels are based on two- tailed Wilcoxon Signed rank test. *, ** and *** marks tests that are significantly different from zero at 10%, 5% and 1% level, respectively. For each matched pair of responses, Wilcoxon measures the difference between the responses. The numbers in the hard brackets shows the proportion of positive differences around the median of zero. Operational Performance Measures. Full sample (N= 63). Unadjusted data. Median of all Test Proportion of differences Statistic P- value positive differences Description Growth CAGR in Revenue ** [0.38] Profitability EBIT/Revenue [0.43] EBIT/Total Operating Assets [0.44] Working Capital Management Working Capital/Revenue * [0.44] Fixed Asset Utilization PP&E/Revenue [0.54] Growth and working capital management Based on the Wilcoxon signed rank test, we find support for rejecting the null hypothesis for CAGR in revenue at a 5 % significance level. The median difference of means that the second private equity firm experiences lower growth rates in revenue during the holding period compared to the primary owner. Looking at the working capital to revenue ratio, we find support for rejecting the null hypothesis at a 10 % significance level. The median difference of means that secondary owner is reducing working capital to revenue relative more than the first owner. 65

68 Table 8: Changes in operational performance for the full sample of industry adjusted data The table reports median percentage point changes in CAGR in revenue. The remaining measures of the changes in operational performance are median difference between percentage points. All the measures are adjusted for industry specific effects. Significance levels are based on two- tailed Wilcoxon Signed rank test. *, ** and *** marks tests that are significantly different from zero at 10%, 5% and 1% level, respectively. For each matched pair of responses, Wilcoxon measures the difference between the responses. The numbers in the hard brackets shows the proportion of positive differences around the median of zero. Operational Performance Measures. Full sample (N= 63). Median of all differences Test Statistic P- value Proportion of positive differences Description Growth CAGR in Revenue [0.48] Profitability EBIT/Revenue EBIT/Total Operating Assets [0.46] [0.48] Working Capital Management Working Capital/Revenue * [0.41] Fixed Asset Utilization PP&E/Revenue [0.48] Working capital management Looking at the working capital to revenue ratio, we find support for rejecting the null hypothesis at a 10 % significance level. The median difference of shows that the secondary owner seems to be superior in reducing working capital to revenue, compared to the first private equity owner. Sub- conclusion for full sample Above we provided the results for our four hypotheses using both unadjusted data and adjusted data for the full sample of 63 SBOs and corresponding LBOs. For the unadjusted data, statistically significant results were found for CAGR in revenue being different between the two periods at a 5 % significance level. The working capital to revenue ratio was significant at a 10 % significance level. However, after adjusting for industry specific effects we found statistically significant results only for the working capital to revenue ratio at a 10 % significance level. 66

69 In the remainder of this analysis, we rely only on adjusted data because these are more reliable for showing the effects of the changes in private equity ownership over time Subsamples In this subsection we divide the total sample into relevant subsamples in order to investigate if this will change the results for our hypotheses, compared to our results for the total sample. For the total sample adjusted for industry effects we found statistically significant results only for the working capital to revenue ratio, why we are motivated to explore if this changes when splitting the total sample according to specific contingencies. In order to do so, we investigate the following subsamples 1) The country of the portfolio companies Given our focus on the Nordics we divide the total sample into subsamples with companies from Sweden, Norway and Denmark. This subsample is particularly relevant because the buyout market in Sweden is arguably more developed, as argued previously, than the markets in Norway and Denmark. Correspondingly, portfolio companies from Sweden account for 57 % of the total observations. 2) The country of the private equity firm Inspired by Gaspar (2012) who divides his sample according to deals made by domestic funds vs. deals made by foreign funds, we divide our sample based on secondary buyout transactions where the geographic region of the private equity owner changes. Based on our focus on the Nordics we define geographic regions as Nordic and not Nordic, respectively. 3) The industry of the portfolio company Dividing the total sample according to the industry of the portfolio companies is particularly relevant because we assume that different types of firms has different prerequisites for enhancing operational performance. Furthermore, 38 % of the portfolio companies operate in manufacturing, which makes a split natural. Gaspar (2012) also 67

70 divides his sample into targets operating in the industrial sector vs. targets operating in the service sector. To serve our purpose we use two categories being manufacturing and non- manufacturing. 4) The form of the financial statements used to collect data In the academic literature there is an ongoing discussion about the use of consolidated vs. unconsolidated financial statements (e.g. Wang, 2012; Sousa, 2012; Freelink and Volosovych, 2012). We therefore divide the sample into transactions for which consolidated and unconsolidated financial statements were used, respectively. This approach is in line with how Gaspar (2012) addresses this issue. The subsample is therefore primarily related to our method rather than practical explanations of SBOs. 5) Realized deals and unrealized deals As the sample consists of both 45 % unrealized- and 55 % realized deals, we split the sample based on these characteristics. This is similar to Achleitner and Figge (2012) who use a sample consisting of 47 % unrealized deals. The subsample is included as a mean of testing our method more than providing practical explanations of SBOs. 68

71 The country of the portfolio company Table 9: Changes in operational performance for Swedish portfolio companies The table reports median percentage point changes in CAGR in revenue. The remaining measures of the changes in operational performance are median difference between percentage points. All the measures are adjusted for industry specific effects. Significance levels are based on two- tailed Wilcoxon Signed rank test. *, ** and *** marks tests that are significantly different from zero at 10%, 5% and 1% level, respectively. For each matched pair of responses, Wilcoxon measures the difference between the responses. The numbers in the hard brackets shows the proportion of positive differences around the median of zero. Operational Performance Measures. Swedish portfolio companies (N= 36). Median of all Test Proportion of differences Statistic P- value positive differences Description Growth CAGR in Revenue [0.44] Profitability EBIT/Revenue [0.39] EBIT/Total Operating Assets [0.36] Working Capital Management Working Capital/Revenue [0.47] Fixed Asset Utilization PP&E/Revenue [0.44] 69

72 Table 10: Changes in operational performance for Norwegian portfolio companies The table reports median percentage point changes in CAGR in revenue. The remaining measures of the changes in operational performance are median difference between percentage points. All the measures are adjusted for industry specific effects. Significance levels are based on two- tailed Wilcoxon Signed rank test. *, ** and *** marks tests that are significantly different from zero at 10%, 5% and 1% level, respectively. For each matched pair of responses, Wilcoxon measures the difference between the responses. The numbers in the hard brackets shows the proportion of positive differences around the median of zero. Operational Performance Measures. Norwegian portfolio companies (N= 18). Median of all differences Test Statistic P- value Proportion of positive differences Description Growth CAGR in Revenue [0.50] Profitability EBIT/Revenue EBIT/Total Operating Assets [0.56] [0.67] Working Capital Management Working Capital/Revenue ** [0.33] Fixed Asset Utilization PP&E/Revenue [0.44] Working capital Management For the working capital to revenue ratio, we find support for rejecting the null hypothesis at a 5 % significance level. The median of all differences of implies that the second private equity firm to own a Norwegian portfolio company is able to reduce working capital to revenue more, compared to the first private equity firm. 70

73 Table 11: Changes in operational performance for Danish portfolio companies The table reports median percentage point changes in CAGR in revenue. The remaining measures of the changes in operational performance are median difference between percentage points. All the measures are adjusted for industry specific effects. Significance levels are based on two- tailed Wilcoxon Signed rank test. *, ** and *** marks tests that are significantly different from zero at 10%, 5% and 1% level, respectively. For each matched pair of responses, Wilcoxon measures the difference between the responses. The numbers in the hard brackets shows the proportion of positive differences around the median of zero. Operational Performance Measures. Danish portfolio companies (N= 9). Median of all differences Test Statistic P- value Proportion of positive differences Description Growth CAGR in Revenue [0.56] Profitability EBIT/Revenue EBIT/Total Operating Assets [0.56] [0.56] Working Capital Management Working Capital/Revenue [0.33] Fixed Asset Utilization PP&E/Revenue [0.67] Sub- conclusion: The country of the portfolio company Above we showed the results for the subsamples of Swedish-, Norwegian-, and Danish portfolio companies. We did not find any statistically significant results for Swedish- and Danish portfolio companies. For the subsample of Norwegian portfolio companies, we found support for rejecting the null hypothesis for the working capital to revenue ratio at a 5 % significance level. This result indicated that the working capital management improved more in the SBO compared to the LBO. 71

74 The country of the private equity firm Table 12: Changes in operational performance for portfolio companies where the origin of the private equity owner changes The table reports median percentage point changes in CAGR in revenue. The remaining measures of the changes in operational performance are median difference between percentage points. All the measures are adjusted for industry specific effects. Significance levels are based on two- tailed Wilcoxon Signed rank test. *, ** and *** marks tests that are significantly different from zero at 10%, 5% and 1% level, respectively. For each matched pair of responses, Wilcoxon measures the difference between the responses. The numbers in the hard brackets shows the proportion of positive differences around the median of zero. Operational Performance Measures. Change in regional origin of the private equity firm (N=32). Median of all Test Proportion of differences Statistic P- value positive differences Description Growth CAGR in Revenue [0.44] Profitability EBIT/Revenue [0.47] EBIT/Total Operating Assets [0.50] Working Capital Management Working Capital/Revenue [0.34] Fixed Asset Utilization PP&E/Revenue [0.44] 72

75 Table 13: Changes in operational performance for portfolio companies where the origin of the private equity owner persists The table reports median percentage point changes in CAGR in revenue. The remaining measures of the changes in operational performance are median difference between percentage points. All the measures are adjusted for industry specific effects. Significance levels are based on two- tailed Wilcoxon Signed rank test. *, ** and *** marks tests that are significantly different from zero at 10%, 5% and 1% level, respectively. For each matched pair of responses, Wilcoxon measures the difference between the responses. The numbers in the hard brackets shows the proportion of positive differences around the median of zero. Operational Performance Measures. No change in regional origin of the private equity firm (N=31). Median of all Test Proportion of differences Statistic P- value positive differences Description Growth CAGR in Revenue [0.52] Profitability EBIT/Revenue [0.45] EBIT/Total Operating Assets [0.45] Working Capital Management Working Capital/Revenue [0.48] Fixed Asset Utilization PP&E/Revenue [0.52] Sub- conclusion: The country of the private equity firm We cannot reject the null hypothesis for any of the measures when dividing our total sample according to whether or not the geographic region of the private equity firm changes between the LBO and the SBO. Note, however, that the p- values for working capital to revenue are relatively low for both subsamples; and 0.196, respectively. 73

76 The industry of the portfolio company Table 14: Changes in operational performance for portfolio companies in the manufacturing industry The table reports median percentage point changes in CAGR in revenue. The remaining measures of the changes in operational performance are median difference between percentage points. All the measures are adjusted for industry specific effects. Significance levels are based on two- tailed Wilcoxon Signed rank test. *, ** and *** marks tests that are significantly different from zero at 10%, 5% and 1% level, respectively. For each matched pair of responses, Wilcoxon measures the difference between the responses. The numbers in the hard brackets shows the proportion of positive differences around the median of zero. Operational Performance Measures. Manufacturing firms (N= 24). Adjusted data. Median of all Test Proportion of differences Statistic P- value positive differences Description Growth CAGR in Revenue [0.63] Profitability EBIT/Revenue [0.42] EBIT/Total Operating Assets [0.42] Working Capital Management Working Capital/Revenue ** [0.33] Fixed Asset Utilization PP&E/Revenue [0.63] Working capital management The working capital to revenue ratio is significant at a 5 % level. The median of all differences of shows that the secondary owner has optimized the ratio more than the primary owner has. 74

77 Table 15: Changes in operational performance for portfolio companies not in the manufacturing industry The table reports median percentage point changes in CAGR in revenue. The remaining measures of the changes in operational performance are median difference between percentage points. All the measures are adjusted for industry specific effects. Significance levels are based on two- tailed Wilcoxon Signed rank test. *, ** and *** marks tests that are significantly different from zero at 10%, 5% and 1% level, respectively. For each matched pair of responses, Wilcoxon measures the difference between the responses. The numbers in the hard brackets shows the proportion of positive differences around the median of zero. Operational Performance Measures. Non- manufacturing firms (N= 39). Description Growth Median of all differences Test Statistic P- value Proportion of positive differences CAGR in Revenue * [0.38] Profitability EBIT/Revenue EBIT/Total Operating Assets Working Capital Management [0.49] [0.51] Working Capital/Revenue [0.46] Fixed Asset Utilization PP&E/Revenue [0.38] Growth The results for non- manufacturing companies show a negative median of all differences of for CAGR in revenue, which is statistically significant at a 10 % significant level. This means that the primary owner grows revenue relatively more than the secondary owner over the holding period. Sub- conclusion: The industry of the portfolio company For the subsample composed of manufacturing firms, we found statistically significant results for the working capital to revenue ratio at a 5 % significance level. We also found statistically significant results for CAGR in revenue for non- manufacturing firms at a 10 % significance level. 75

78 The form of the financial statements used to collect data Table 16: Changes in operational performance for portfolio companies where data is obtained from consolidated financial statements The table reports median percentage point changes in CAGR in revenue. The remaining measures of the changes in operational performance are median difference between percentage points. All the measures are adjusted for industry specific effects. Significance levels are based on two- tailed Wilcoxon Signed rank test. *, ** and *** marks tests that are significantly different from zero at 10%, 5% and 1% level, respectively. For each matched pair of responses, Wilcoxon measures the difference between the responses. The numbers in the hard brackets shows the proportion of positive differences around the median of zero. Operational Performance Measures. Consolidated financial statements (N= 33). Median of all differences Test Statistic P- value Prop. of positive differences Description Growth CAGR in Revenue [0.48] Profitability EBIT/Revenue EBIT/Total Operating Assets [0.52] [0.55] Working Capital Management Working Capital/Revenue [0.42] Fixed Asset Utilization PP&E/Revenue [0.48] 76

79 Table 17: Changes in operational performance for portfolio companies where data is obtained from unconsolidated financial statements The table reports median percentage point changes in CAGR in revenue. The remaining measures of the changes in operational performance are median difference between percentage points. All the measures are adjusted for industry specific effects. Significance levels are based on two- tailed Wilcoxon Signed rank test. *, ** and *** marks tests that are significantly different from zero at 10%, 5% and 1% level, respectively. For each matched pair of responses, Wilcoxon measures the difference between the responses. The numbers in the hard brackets shows the proportion of positive differences around the median of zero. Operational Performance Measures. Unconsolidated financial statements (N= 30). Median of all differences Test Statistic P- value Proportion of positive differences Description Growth CAGR in Revenue [0.47] Profitability EBIT/Revenue EBIT/Total Operating Assets [0.40] [0.40] Working Capital Management Working Capital/Revenue * [0.40] Fixed Asset Utilization PP&E/Revenue [0.47] Working capital management We can reject the null hypothesis for the working capital ratio at a 10% significance level. The median of all differences of implies that the secondary owner is better at optimizing the working capital to revenue ratio relative to the primary owner. 77

80 Realized deals and unrealized deals Table 18: Changes in operational performance for portfolio companies where the second private equity firm has exited The table reports median percentage point changes in CAGR in revenue. The remaining measures of the changes in operational performance are median difference between percentage points. All the measures are adjusted for industry specific effects. Significance levels are based on two- tailed Wilcoxon Signed rank test. *, ** and *** marks tests that are significantly different from zero at 10%, 5% and 1% level, respectively. For each matched pair of responses, Wilcoxon measures the difference between the responses. The numbers in the hard brackets shows the proportion of positive differences around the median of zero. Operational Performance Measures. Realized deals (N= 34). Median of all differences Test Statistic P- value Proportion of positive differences Description Growth CAGR in Revenue [0.53] Profitability EBIT/Revenue EBIT/Total Operating Assets [0.50] [0.47] Working Capital Management Working Capital/Revenue [0.47] Fixed Asset Utilization PP&E/Revenue [0.56] 78

81 Table 19: Changes in operational performance for portfolio companies where the second private equity firm has not exited The table reports median percentage point changes in CAGR in revenue. The remaining measures of the changes in operational performance are median difference between percentage points. All the measures are adjusted for industry specific effects. Significance levels are based on two- tailed Wilcoxon Signed rank test. *, ** and *** marks tests that are significantly different from zero at 10%, 5% and 1% level, respectively. For each matched pair of responses, Wilcoxon measures the difference between the responses. The numbers in the hard brackets shows the proportion of positive differences around the median of zero. Operational Performance Measures. Unrealized deals (N= 29). Median of all differences Test Statistic P- value Proportion of positive differences Description Growth CAGR in Revenue [0.41] Profitability EBIT/Revenue EBIT/Total Operating Assets [0.41] [0.48] Working Capital Management Working Capital/Revenue * [0.34] Fixed Asset Utilization PP&E/Revenue [0.38] Working capital management We note that the working capital to revenue is significant at a 10 % significance level. The negative sign of the median of all differences suggest that the secondary owner has optimized the working capital to revenue more than the first private equity owner. 79

82 8. Discussion In the previous chapter we tested four hypotheses regarding the value creation potential of companies backed by private equity firms for the second, consecutive time. This chapter contributes with a discussion of the results presented in the previous chapter. Furthermore, we reconsider and discuss our decisions with regards to the method applied Prerequisites for the discussion In the following discussion we will focus our attention towards the statistically significant results found in the previous chapter. We consider our discussion to be partially explorative in nature, which means that we intend to point out areas of future interest to both private equity practitioners and academic scholars based on our empirical results. We do not engage in an elaborate discussion of statistically insignificant results because this would be merely speculation. However, we do start out with a brief discussion of how to interpret our statistically insignificant results Interpreting statistically insignificant results Although some of our results are statistically insignificant, they still have an interesting interpretation that serves to expand our general knowledge on secondary buyouts. Based on conventional wisdom, academic scholars and the business press are generally concerned about the perceived limited opportunities for operational value creation improvements in secondary buyouts. However, the lack of significance shown in our results suggest that SBOs are not inferior deals compared to LBOs in terms of operational value creation measured by revenue growth, profitability and fixed asset utilization. In other words, we cannot reject that SBOs are similar to LBOs. We must be aware that the reason for our statistically insignificant results can be two- fold. First, it can be because SBOs are similar to LBOs in line with the argument above but, secondly, it can also be due to lack of statistical power. The particular discussion of statistical power will be picked up in part two of the discussion when we reconsider our method. 80

83 8.3. Fundamental reasoning A fundamental principle when dealing with corporate mergers and acquisitions is the question of who is the best owner of a particular asset. Arguably the best owners are those companies whose distinctive characteristics enable them to create more value in a given business than other potential owners could, which is captured well in the following: If a board and management team want to create the most value for their own shareholders, they must be clear about how their company will add more value to a business than other potential owners can. If that isn t the case, the company might best serve the shareholders interests by selling the business or by not buying it in the first place (Dobbs et al., 2009). The best owner principle is also relevant to apply in connection with the discussion of the value creation potential in secondary buyouts. Private equity firms must reconsider their ownership of the respective portfolio companies over time, just like corporate executives must reevaluate their portfolio of corporate assets over time. Hence, private equity firms are expected to engage in sourcing of deals for which they could be the best owner and, similarly, exit portfolio companies for which they are no longer the best owner. Hence, it is particularly important to understand that the best ownership is not permanent or static but rather can change over the life cycle of a business (Dobbs et al., 2009) Superior working capital management For the working capital to revenue ratio the median difference is negative, which implies that the median sample company in secondary buyouts is able to improve the working capital to revenue ratio more, relative to the median sample company in first buyout round. These results are interesting, first, because they are statistically significant, but also because they suggest the opposite of our initial expectations. We expected the primary owner to reap the biggest improvements in working capital to revenue, but our results show that it is actually the secondary owner who improves working capital to revenue the most. 81

84 In practice this means that our results may help to explain the popularity of secondary buyouts among private equity practitioners. Furthermore, our results also suggest that academic scholars should perhaps reconsider their skepticism towards SBOs. This is especially the case when further taking into consideration that we cannot reject that operational improvements of the LBO and the SBO is similar for the other measures we investigate. In the following discussion we will therefore talk more about the implications of our results for both private equity practitioners and academic scholars. Improvements in working capital have been the focus of empirical papers on buyouts in general. For instance, Smith (1990) finds that sales to working capital increases around the buyout year and Gaspar (2012) show that working capital to sales is reduced during leveraged buyouts. However, only Evers and Hege (2012) have studied working capital improvements particularly for SBOs and they find that there are no differences in working capital improvements between the LBO and SBO. We therefore consider our results as a valuable contribution to the current debate about SBOs, and furthermore suggest that more research on working capital improvements in SBOs is needed in order to reach a consensus in the literature. We are actually surprised to find that the secondary owner is able to improve working capital to revenue more than the primary owner because we know that private equity firms, also during the first round buyout, are expected to focus on the financial structure of the portfolio company which typically implies that they will engage in optimization of working capital (Smith, 1990; Gou et al., 2011; Gaspar, 2012). In particular, these optimizations include inventory, receivables and payables management (Brigl et al., 2012) why we suggest that scholars in the future could build on our empirical findings by I) studying working capital to revenue improvements in SBOs and II) by examining more closely which of these areas of working capital management private equity firms focus on. However, based on the current state of the young and concentrated body of literature on secondary buyouts we argue that the most reasonable theoretical argument for our empirical 82

85 results is the structure hypothesis (Sousa, 2010) or the forced seller hypothesis as suggested by Evers and Hege (2012). Because private equity funds are structured as limited- life vehicles, private equity firms have strong incentives to avoid long holding periods as a good track record, measured by IRR, is very important to the reputation of the respective private equity firm and its future ability to raise new funds (Sousa, 2010; Evers and Hege, 2012). In this particular context, it implies that primary owner chooses to exit a portfolio company early, in order to signal to its investors, and thereby pass on the portfolio company to another private equity firm before they themselves have finished exploiting improvements from a tighter working capital management and growing revenue. In contrary, Cumming and MacIntosh (2003) has argued that private equity firms will only sell a portfolio company once the expected marginal return of value creation through their own effort and investments is lower than the marginal cost represented by that very effort and investment. Theoretically speaking this argument makes it questionable why follow- on private equity firms should consider SBOs to be attractive investments in terms of value creation potential. However, given our findings and the popularity of SBOs among practitioners, it seems that Cumming and MacIntosh (2003) s argument is merely theoretical and does not take into account the structural incentives of private equity firms to exit a portfolio company early. Another theoretical argument that seems questionable, given our findings, is the go- for- broke hypothesis (Axelson et al., 2009; Degeorge et al., 2012) which implies that private equity firms are willing to invest in SBOs, although they are perceived less attractive, as the alternative of not investing the committed capital sends an undesired signal to the investors. Ultimately, this suggests that SBOs are second- hand deals, but our findings and the lasting popularity of SBOs seem to suggest otherwise. We have already pointed out that more research on the operational value creation potential in SBOs, and particularly on working capital management, is needed in order to improve the academic knowledge on secondary buyouts, as the literature is still young and concentrated. For private equity in general, our results suggest that SBOs are arguably not inferior deals compared to LBOs why we should expect to see more SBOs in the years to come. First, because 83

86 SBO activity is a function of LBO activity, which has been catching up again in recent years after the financial crisis. Secondly, the number of SBOs will presumably stay high because private equity firms are becoming more experience with SBO deals and therefore continue to improve their skills in sourcing and optimizing these investments, making the performance of SBOs better Potential industry differences The results of the tests on the subsamples of industries show statistically significance for working capital to revenue for manufacturing firms. This is interesting as the negative sign suggest that the working capital management is better in the secondary buyout compared to the primary buyout. For the non- manufacturing firms there is no support for working capital management being different between the LBO and the SBO. Furthermore, when examining the CAGR in revenue for non- manufacturing companies the results are significant and negative. This implies that the revenue growth is higher in the portfolio company over the holding period in the LBO compared to the SBO. For the portfolio companies in the manufacturing industry we cannot reject that the difference is zero between the LBO and the SBO, although with a p- value of and a positive sign it points towards the revenue growth being higher in the secondary buyout compared to the primary buyout. The results obtained for the sample split by industry are interesting and makes us speculate, that the manufacturing industry is the better option for private equity firms investing in SBOs in the Nordics. Conversely, private equity firms should be more skeptical when considering a SBO of a Nordic non- manufacturing firm in terms of making operational value creation. The overall implication is that private equity professionals and scholars probably should think about the importance of the industry in which the respective portfolio companies operate when engaging in SBOs. In that regard, the composition of the sample that we are looking at shows apparent evidence that manufacturing companies are a popular type of company with 24 out of 63 sample companies operating in manufacturing, which clearly constitutes the biggest concentration of companies within a particular industry in the sample. 84

87 8.5. Limitation and biases of the method applied The objective of this is section is to discuss the implication of the method applied. By identifying and discussing major discrepancies between the methods we use in this thesis and the methods applied in the existing literature, we assess the impact of the choices made Sample period The buyouts included in the sample have taken place between 1995 and 2014, with the first SBOs completed in 2001, which makes the period covered by the sample quite a long. From academic literature, we know that private equity transactions are cyclical and appears in so- called buyout waves (Kaplan and Strömberg, 2009). The fact that buyouts are happening in waves also explains the content covered by the literature on private equity and buyouts. There is a comprehensive and older literature on leveraged buyouts in general, inspired by the earlier waves, whereas the literature on SBOs is sparse and recent. An implication of this is that consensus on evidence made by scholars investigating the performance of SBOs is not yet well developed and hence the expectation formation with regards to the performance of SBOs is not uniform. In relation to this, Wright et al. (2009) highlighted the need for investigation of whether SBOs are any different from LBOs. Since then the number of empirical papers published and working papers on SBOs has increased significantly although the amount of literature still is relatively limited. Looking at the timing of transactions in the sample, which are presented as a part of the descriptive statistics previously shown in table 3, the cyclicality is evident despite the number of observations. For the first buyout round 59 of the transactions takes place in two short periods between 1997 and 2001 and 2003 and The secondary buyout are naturally a function of the timing of the LBOs also in this setting, however there are evidence of clustering of deals between 2004 and 2008 where 47 out of the 63 transactions take place. The cyclicality can arguably be contributed to the overall economic conditions. We argue, that macroeconomic chocks from the burst of the dot- com bubble in 2000/01 and the global credit- and financial crisis from 2007/08 causes the slowdown in 2001 to 2003 and from 2008 and onwards in the sample. Covering a period with several rising- and declining buyout waves may impose bias in the results, given the assumption that portfolio companies in times with high activity finds it easier to divest the portfolio company compared to periods with slowing activity. In terms of our results this may affect CAGR in revenue, since this decreases as time passes by, all things been equal. If a 85

88 private equity firm finds it difficult selling a portfolio company due to a decline in the buyout activity or cold IPO markets the holding period can be extended beyond the optimal, causing a lower CAGR in revenue. The other measures we have covered, profitability, working capital management, and fixed asset utilization can also be affected by the overall economic conditions at the time of the transactions. We argue, however, that taking into account the industry adjustments smoothens such effects out. Additionally, these measures are not directly affected by the time in the computations. The overall implication of using a period covering several buyout waves and different business cycles is that most transactions are conducted in clusters. We argue that mainly CAGR in revenue is affected and that this effect can be present for both the LBO and the SBO, which ultimately may cause some netting out in the bigger picture when making comparisons Length of periods evaluated In order to compare the changes in operational performance between the LBO and the SBO we have evaluated changes over the period from the entry to the exit of the private equity firm. This approach is similar to the one applied by Gou et al. (2011) who evaluates the performance over the full period in addition to the shorter periods. Likewise, Freelink and Volosovych (2012) examine the changes in operating performance over the full period. Bonini (2012) also contributes to the methodological discussion concerning the performance window. Ideally, we should try and collect data on as many fiscal years as possible following the buyout. (Bonini, 2012, p.12). Hence, based on the methods of these scholars we perceive full holding periods as a valid way to go. In contrary, Wang (2012) advocates for an event window of three years around the buyout year to capture the effect of changes and highlights the importance of keeping the event window shorter in order to avoid noise from longer holding periods. Looking at changes over full holding periods imply that the mere effect of the buyout in the important first couple of years is not isolated. We have chosen to put emphasis on the fact that the private equity firm is able to influence the operations of the portfolio company throughout the entire holding period in line with Freelink and Volosovych (2012). Undoubtedly, there is a trade- off between capturing the effect of the activities over the entire holding period and avoidance of capturing noise that cannot be contributed to the private equity firm, when setting the duration of 86

89 the event window. Furthermore, comparing varying length of holding periods implies that changes might be differing only because of the length of the period covered. First, if the period is short, the changes might not be fully reflected in the measures, as the effect has not kicked in yet. On the other hand, a long period might be affected by noise from other activities than the key initiatives driven by the investing private equity firm. We believe however that the varying lengths reflect the private equity firm s window of actions and hence the potential for changing the operational performance Consecutive buyouts Looking at a sample consisting of the same portfolio companies in the first buyout and the second buyout follows the approach of Bonini (2012). The only difference is that the sampling approach of Bonini (2012) also includes a few tertiary buyouts, while we are only looking at secondary buyouts. The majority of the literature examining secondary buyouts is not specifying strict rules for which observations to include, such as sequential buyouts. Thus, the number of LBOs is outnumbering the SBOs in these studies due to the nature of the buyouts examined (Wang, 2012). Other studies that are looking at buyouts in general are not spending much effort on the linkage between first- round buyouts and secondary buyouts (e.g. Gaspar, 2012; Gou et al., 2011). We acknowledge that the method of applying sequentially buyouts by private equity firms may cause some inappropriate properties such as reducing the sample size and imposing bias towards the LBOs being more successful. Using consecutive buyouts can result in survivorship bias from the LBOs in our sample, as LBOs performing badly and ultimately faces bankruptcy does not end as SBOs. In the majority of the sample SBOs are not exited through bankruptcy and hence share this characteristic with the LBO implying no bias. We are however including not exited SBOs, which might cause observations where the performance is so poor that it affects the overall picture of SBOs to be included. We have added- back observations of not exited SBOs with a ownership of four years or more based on the historical median holding period of SBOs in our sample. While we find this approach well founded we stress the potential bias posed by an unintentional long holding period, as discussed in relation to the cyclicality of the buyout activity above, possibly can be reflected in a method using historical numbers. 87

90 Another limitation created by looking at consecutive buyouts is that the LBOs only qualify for the sample if the exit is to another private equity firm and therefore the performance of these LBOs, relative to other LBOs, are unknown. In other words, there might be a bias from the LBOs that go into the sample, since we are not aware of their performance relative other LBOs in the setup we have applied. In turn, the strictness of the study design caused by the consecutive buyouts has the advantage that there is an easy comparison across the sample making it possible to assess the difference in performance changes over the two periods Use of the transaction year Another relevant discussion is whether it is appropriate to use the year of the transaction when evaluating operating performance measures. Kaplan (1989a) advocates for not using accounting data for the year of the buyout for two reasons. First, it includes both pre- and post- buyout performance and secondly there might be fees related to the buyout reducing operating income and write- ups of inventory increasing the assets. Given the widespread method of omitting the transaction year in the literature, it would also have been relevant for us to complete this study applying the years around the transactions in order to see if this alters the results. The rule applied for deciding what accounting year to use, discussed in chapter six, has however resulted in usage of data from the year before the buyout for a number of observations in line with existing literature. Everything else being equal this reduces the potential bias caused by using the year of the transaction. The data collected is for a single year and thus the balance sheet data is single figures and not averages between ultimo and primo numbers. This approach was chosen for three reasons. First, we wished to have the most recent and accurate numbers. Secondly, the purpose of this thesis is not to do financial statements analysis and hence we wanted to keep the complexity down. Third, having to include another annual report would have caused the sample size to drop due to availability of financial statements, especially for the earlier LBOs. This approach implies that extraordinary circumstance happening in one year can affect the overall inference, we are however less concerned about this matter because we are looking at several observations simultaneously, which is believed to reduce a potential bias from outliers. 88

91 The use of the transaction year has the advantage that the timespan where the operational performance changes are measured is more accurate and hence includes potential contributions from this period also. In terms of cons, using the year of transaction might impose bias on the EBIT measures lowering it by the cost of the transaction being ascribed to the portfolio company and on the asset base and working capital from write- ups of assets as suggested by Kaplan (1989a) Accounting data used Our data collection follows Gaspar (2012) who is also collecting data from both consolidated and unconsolidated financial statements. The use of unconsolidated financial statements leads to inclusion of minority interests and interests in the group company. These interests should ideally be adjusted for in the balance sheet as they are not a part of the operations of the firm, and affects the profit or loss on an after tax basis (Petersen and Plenborg, 2012, p. 78). Using financial data from the Amadeus database however made this adjustment difficult, as the level of detail for the observations included was not sufficiently high for us to adjust the current assets and current liabilities for these effects. Given these limitations we decided not to adjust working capital across the sample regardless of the data source. This decision was made to ensure consistency in the data collected. The interests in the group company often are present on both sides of the balance sheets, which may cause some equalization effect in the working capital. We are however aware that this potentially has caused some bias in the computations of working capital and total operating assets. Similar to Gaspar (2012) we test the performance changes for the consolidated- and unconsolidated financial statements, respectively. We use tests of the subsamples to reassure that the results and the conclusions are not driven by the source of the data collected. In our tests we find small discrepancies between the results regarding the sign of profitability, however not significant. Apart from that, the results are comparable across the two subsamples and the full sample. Gaspar (2012) argue that use of unconsolidated statements may lead to biases as the private equity firm buying the portfolio company often reorganizes the operating assets in different holding layers, which makes the continuous analysis of performance changes difficult. One of the primary arguments for including unconsolidated financial statements in our thesis is to reach a sufficient number of observations in order to make proper statistical inference. Having 89

92 relied on observations of firms with consolidated financial statements only would have caused the sample size to be reduced critically. Another valuable property by not excluding observations based on their type of financial statement is that restricting the sample to include only firms with consolidated financial statements may leave the sample non- representative of the buyout market (Wang, 2012). In sum, there are pros and cons for using unconsolidated financial statements to continuous analysis of performance changes and hence one would preferably use only consolidated statement if these were available in large scale and for a representative range of the buyout market in the Nordics Measuring changes As covered in chapter six, we are ultimately measuring the performance changes in profit margin, return of operating assets, working capital management and fixed asset utilization as differences between percentage points. We are aware that this method is underestimating or overestimating the changes achieved compared to a method where we would have examined percentage changes. For example if a company has increased its EBIT- margin from 5 % to 10 % over the period, it is a 100 % increase, whereas if the company has increased the EBIT- margin from 10 % to 15 % corresponds to an increase of 50 %. In our method both of these changes is evaluated as an increase of 5 percentage points and hence a difference of zero between the two, understating the performance of the first- relative to second period. Given our study design, we are looking at paired data and we are thus comparing the performance of a company to its own performance in another period. This implies that the comparisons are made between comparable companies, which reduce the bias relative to a situation where the comparison was made across firms and industries. As mentioned earlier, we use percentage point changes in order to not discard observations with negative values and lowering the statistical power of the testing. This however means that we have been aware of the practical interpretation of the results found in the analysis when making comparisons to others Statistical power In order to ensure proper statistical inferences, in practice, a sample size of 30 observations is preferable (Agresti and Franklin, 2009, p. 328). We are therefore comfortable about our results for the total sample composed of 63 secondary buyouts, although we would have increased the 90

93 sample size if possible. However, as we argued earlier our sample has been limited, with regards to size, by our focus on the Nordics and our focus on consecutive buyouts. We also tested our hypotheses on a number of subsamples for which the number of observations naturally decreased. With respect to our subsamples we are therefore aware of the potential challenge of making correct statistical interferences due to the limited subsample sizes. However, these subsamples naturally reflect the size of the secondary buyout market in the Nordics why we found it relevant to include the subsamples, and provide the results obtained, while keeping in mind the potential limitations of these results. Furthermore, because we consciously intended to point out areas of future interest to both private equity practitioners and academic scholars it is very important to highlight our findings on these subsamples. 91

94 9. Conclusion Our master thesis has been inspired by the recent popularity of secondary buyouts among practitioners and the opposite view held by academic scholars who argue that secondary buyouts offer limited potential for operational value creation. The body of literature on secondary buyouts is relatively young and concentrated why we identified ample opportunities to contribute to this debate. In order to examine the operational value creation potential of secondary buyouts, we identified a hand- collected sample of 63 secondary buyouts with portfolio companies only from Sweden, Norway and Denmark. In addition to our focus on the Nordics, our sample is also novel because we only included sample companies that have experienced two consecutive buyouts by private equity firms. In terms of measuring operational performance of the respective portfolio companies, we decomposed ROIC into its components of the profit margin and the turnover rate of invested capital, which allowed us to develop four testable hypotheses. Furthermore, in order to isolate the effects of the change in private equity ownership on operational value creation, we adjusted our data for industry specific effects related to the respective industries of the sample companies. Table 20: Summary of hypotheses and results Hypotheses Growth hypothesis Compared to LBOs, revenue growth is higher in SBOs Support No Profitability hypothesis Compared to LBOs, margin improvements are lower in SBOs. No Working capital management hypothesis Compared to LBOs, reductions in working capital are lower in SBOs No Fixed asset utilization hypothesis Compared to LBOs, reductions in capital tied in PP&E are lower in SBOs No 92

95 With regards to revenue growth, profitability and fixed asset utilization we cannot reject that SBOs are similar to LBOs, which suggest that secondary buyouts are not inferior deals in terms of operational performance improvements, compared to first round buyouts. Moreover, we found statistically significant results, which show that SBOs offer higher working capital improvements than LBOs measured by working capital to revenue. In our discussion we devoted our attention primarily towards our findings on working capital management because they are statistically significant, and since they suggest the opposite of our initial expectations. Based on the literature on both LBOs and SBOs we expected the primary owner to achieve the biggest improvements in working capital to revenue, but our results show that it is actually the secondary owner who improves working capital to revenue the most. In that regard, we engaged in a discussed of how the structural incentives of private equity firms can explain the logic of our results. Because private equity funds are structured as limited- life vehicles, private equity firms have strong incentives to avoid long holding periods. This arguably implies that the primary owner may choose to exit a portfolio company early and thereby pass on the portfolio company to another private equity firm before they themselves have finished exploiting operational improvements Future research and implications Based on our empirical findings, which overall suggest that secondary buyouts are not inferior deals in terms of operational performance improvements, we suggest that more research on secondary buyouts is desirable in order to expand the young and concentrated body of literature on SBOs. In other words, our findings suggest that academic scholars should strive to expand their general knowledge on secondary buyouts. Our results on working capital management also suggest that more research on working capital to revenue is desirable. In particular we leave open the discussion of which components of working capital private equity firms are focusing on; are secondary owners perhaps superior in optimizing inventories or managing receivables and payables? When looking at SBOs in the Nordics we found patterns suggesting that portfolio companies in the manufacturing industry seem to be better suited for SBOs than companies in other industries, in 93

96 terms of operational value creation. Hence, our results makes it relevant for future studies to examine which specific characteristics of manufacturing firms that appear to make them more attractive SBO targets? For private equity in general, and for private equity managers, our results suggest that SBOs are arguably not inferior deals compared to first round buyouts. We therefore expect to see more SBOs in the future. We know that SBO activity is a function of LBO activity, but the number of SBOs will arguably stay high because private equity managers are becoming more experience with SBO deals. And since private equity managers will continue to improve their expertise within SBOs, we expect the performance of SBOs to improve in the future. 94

97 References Acharya, Viral and Kehoe, Conor and Reyner, Michael (2008). The voice of experience: Public versus private equity. Insights and Publications by McKinsey and Company (December 2008). Available at: versus_private_equity. Retrieved at 10. April Acharya, Viral and Gottschalg, Oliver and Hagn, Moritz and Kehoe, Conor (2011). Corporate Governance and Value Creation: Evidence from Private Equity. Available at: og.pdf. Retrieved 21. April Achleitner, Ann- Kristin and Christian Figge (2012). Private Equity Lemons? Evidence on Value Creation in Secondary Buyouts. European Financial Management, volume 20, issue 2., March Pp Achleitner, Ann- Kristin, Oliver Bauer, Christian Figge, and Eva Lutz (2012). Exit of last resort? Empirical evidence on the returns and drivers of secondary buyouts as private equity exit channel. Available at: Retrieved 10. June Agresti, Alan and Franklin, Chris (2009). Statistics - the art and science of learning from data. Second edition. Pearson Education, Inc. ISNB- 10: Axelson, Ulf, Per Strömberg, and Michael S. Weisbach (2007). Why are buyouts levered: The financial structure of private equity funds. NBER Working Paper Series, Working Paper Available at: Retrieved 23. April. Baker, Georg and Cynthia Montgomery (1994). Conglomorates and LBO Associations: A Comparison of Organizational Forms, Working Paper. Available at: Retrieved 13. June

98 Barber, Brad M. and John D. Lyon (1996). Detecting abnormal operating performance: The empirical power and specification of test statistics. Journal of Financial Economics, Vol. 41, pp Bennedsen, Morten, Steen Thomsen, Søren Bo Nielsen, Jakob Bundgaard, Kasper M. Nielsen, and Thomas Poulsen (2008). Private Equity i Danmark. Centre for Economic and Business Research. Available at: advisory.com/wp- content/uploads/2011/12/cebr.pdf. Retrieved 15. June Berg, Achim and Oliver Gottschalg (2003). Understanding Value Generation in Buyouts. Working Paper. Working Paper series Insead, Version 1.0 as of April 4, Available at: Retrieved 18. March Bergström, Clas, Michael Grubb and Sara Jonsson (2007). The Operating Impact of Buyouts in Sweden. A Study of Value Creation. The Journal of Private Equity, Winter 2007, Vol. 11, No.1, pp Bonini, Stefano (2012). Secondary Buyouts. Working Paper Series, Working paper number 441, Version: May Available at: Retrieved 17. March Brigl, Michael, Peter Nowotnik, Krista Pelisari, John Rose, and Paul Zwillenberg (2012). The 2012 Private Equity Report Engaging for Growth. Available at: Retrieved 30. April Bull, Ivan (1989). Financial Performance of Leveraged Buyouts: An Empirical Analysis. Journal of Business Venturing 4., pp Cumming, Douglas J. and Jeffrey G. MacIntosh (2003). Venture- Capital Exits in Canada and the United States. University of Toronto Law Journal, Vol. 53, pp Degeorge, Francois, Jens Martin, and Ludovic Phalippou (2013). Agency costs and investor returns in private equity: Consequences for secondary buyouts. ECGI - Finance Working Paper No. 96

99 384 and Swiss Finance Institute Research Paper No Available at: Retrieved 25. March Degeorge, Francois, Jens Martin, and Ludovic Phalippou (2012). Is the rise of secondary buyouts good news for investors?. Available at: rge_ger.pdf. Retrieved 25. March Dobbs, Richard and Huyett, Bill and Koller, Tim (2009). Are you still the best owner of your assets. Insights and Publications by McKinsey and Company (November 2009). Available at: ssets. Retrieved at 10. April Evers, Francois and Ulrich Hege (2012). Value Creation and Drivers of Secondary Buyouts in France. Available at: Retrieved 17. March EVCA (2014). The EVCA Yearbook 2013 European Private Equity Activity. Published May 6 th 2014 by EVCA Research. Available at: data/annual- activity- statistics/. Retrieved 17. March Flaherty, Michael and Davies, Megan (2007). Private equity firms losing LBO exit strategy. Reuters (September 2007). Available at: column- deal- sept- idusn Retrieved at 10. April Figge, Christian A. M. (2012). Essays on Secondary Buyouts. Dissertation of Technische Universität München. Availble at: Retrieved 1. July 2014 Freelink, Stephan and Vadym Volosovych (2012). Do Secondary Buyouts Create Value? Evidence from the United Kingdom. Available at: 97

100 Reading/papers/EFMA2013_0009_fullpaper.pdf. Retrieved 16. March Gaspar, José- Miguel (2012). The performance of French LBO Firms: New data and new results, Finance, vol. 33, no. 2, 2012, pp Gilligan, John and Mike Wright (2010). Private Equity Demystified An Explanatory Guide, 2 nd edition. Available at: Retrieved 21. April Goldfarb, Jeffrey (2014). A private equity perennial that may flourish. The New York Times, Deal Book (May 2014). Available at: private- equity- perennial- that- may- flourish/?_php=true&_type=blogs&_r=0. Retrieved at 10. April Gompers, Paul A. and Josh Lerner (1998). What Drives Venture Capital Fundraising?. Brookings Papers: Microeconomics, pp Gompers, Paul and Josh Lerner (2000). Money Chasing Deals? The impact of fund inflows on private equity valuations. Journal of Financial Economics, Vol. 55, pp Gou, Shourun, Edith S. Hotchkiss and Weihong Song (2011). Do buyouts (Still) Create Value?. The Journal of Finance, Vol. 66, No. 2, April, pp Jensen, Michael C. (1986). Agency Costs of Free Cash Flow, Corporate Finance, and Takeovers. American Economic Review, May, Vol. 76, No. 2, pp Jensen, Michael C. (1989). Eclipse of the Public Corporation, Harvard Business Review, September- October, Revised Kaplan, Steven (1989a). The Effect of Management Buyouts on Operating Performance and Value. Journal of Financial Economics, 24, pp Kaplan, Steven (1989b). Management Buyouts: Evidence on Taxes as a Source of Value. The Journal of Finance, Vol. 44, No. 3, pp

101 Kaplan, Steven and Antoinette Schoar (2005). Private Equity Perfomance: Returns, Persistence and Capital Flows. The Journal of Finance, Vol. 60, Issue 4, August 2005, pp Kaplan, N. Steven and Per Strömberg (2008). Leveraged Buyouts and Private Equity. Journal of Economic Perspectives, Volume 23(1), pp Kengelbach, Jens, Michael Brigl, Falk Beilesch, Bernhard Schwetzler, and Alexander Knauer (2013). Take a Second Look at Secondaries. Owners That Raise Their Value- Creation Game Can Excel, Boston Consulting Group and HHL Leipzig Graduate School of Management, Available at: t_secondaries_feb_2013_tcm pdf. Retrieved 26. March Koller, Tim and, Goedhart, Marc and Wessels, David (2010). Valuation measuring and managing the value of companies. Fifth Edition. Copyright by McKinsey and Company. Published by John Wiley and Sons, INC. ISBN: Lichtenberg, Frank R. and Donal Siegel (1990). The Effects of Leveraged Buyouts on Productivity and Related Aspects of Firm Behavior. Journal of Financial Economics, 27, pp Loos, Nicolaus (2005). Value Creation in Leveraged Buyouts. Dissertation of the University of St. Gallen. Available at: Retrieved 7. June Meckling, William H. and Michael C. Jensen (1976). Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure. Journal of Financial Economics, October, 1976, Vol. 3, No. 4, pp Muscarella, Chris J. and Michael R. Vetsuypens (1990). Efficiency and Organizational Structure: A Study of Reverse LBOs. The Journal of Finance, Vol. 45, No. 5, pp Nikoskelainen, Erkki and Mike Wright (2007). The impact of corporate governance mechanisms on value increase in leveraged buyouts. Journal of Corporate Finance, 13 (2007), pp

102 Petersen, Christian V. and Thomas Plenborg (2012). Financial statement analysis, Valuation, Credit Analysis, Executive compensation, publishing: Harlow, Pearson, ISBN: Phan, H. Phillip and Charles W. L. Hill (1995). Organizational restructuring and Economic performance in leveraged buyouts: an ex post study, Academy of Management Journal, Vol. 38, No.3 (June 1995), pp Rappaport, Alfred (1990). The Staying Power of the Public Corporation. Harvard Business Review, January- February, 1990, pp Sousa, Miguel (2010). Why do Private Equity firms sell to each other?. Available at: _FMA.pdf. Retrieved 17. March Smith, Abbie J (1990). Corporate Ownership Structure and Performance. Journal of Financial Economics, 27, pp The Economist (2010). Circular logic: Private equity companies look to each other to solve their problems. The Economist (February, 2010). Available at: Retrieved at 10. April Wang, Yingdi (2012). Secondary Buyouts: Why buy and at what price?. Journal of Corporate Finance, 18 (2012), pp Wright, Mike, John Gilligan, and Kevin Amess (2009). The Economic Impact of Private Equity: What we know and what we would like to know. Venture Capital, Vol.11, No. 1, January 2009, pp

103 Appendix 1 Nordic private equity activity in numbers of portfolio companies traded Figure 11: Distribution of portfolio companies in the Nordics The chart shows the relative distribution of private equity deals in the Nordics. 80,00% 60,00% 40,00% 20,00% Private equity in the Nordics by the number of pormolio companies traded 0,00% Source: Authors contribution. Data: EVCA (2014) Denmark Norway Sweden Figure 12: Nordic private equity activity development in the European market against the largest buyout markets The chart illustrates the number of portfolio companies traded as a percentage of the total European market NORDICS PRIVATE EQUITY ACTIVITY RELATIVE TO THE LARGEST MARKETS IN EUROPE BY NUMBER OF COMPANIES Nordics United Kingdom France Germany 21,61% 24,11% 25,54% 27,09% 26,33% 25,35% 25,86% 13,38% 18,17% 15,97% 15,54% 16,26% 16,03% 14,86% 15,96% 17,73% 13,92% 14,29% 13,32% 13,01% 12,26% 15,62% 10,92% 12,81% 13,39% 12,54% 13,48% 12,36% Source: Authors contribution. Data: EVCA (2014) 101

104 Appendix 2 Lists of company and deal specific information Table 21: Exited observations of the sample The table contains the exited SBOs and the relevant information of the observations Obs. # Portfolio Company Company ID Country of the portfolio company Status of SBO Entry PE1 Exit PE1/ Entry PE2 Exit PE2 Cons. annual reports 1 Ahlsell Sverige AB SE Sweden Exited No Alignment Systems AB SE ; SE Anticimex AB SE ; SE digit NACE Code Sweden Exited 2002* Yes 7022 Sweden Exited Yes Com Hem AB SE Sweden Exited No Dometic International AB SE ; SE Flexlink AB SE ; SE Sweden Exited Yes 7010 Sweden Exited Yes Global Blue SE Sweden Exited No Hilding Anders International AB 9 HMS Industrial Networks AB SE Sweden Exited No 3103 SE Sweden Exited No JD Stenqvist AB SE Sweden Exited No Jetpak Group AB SE Sweden Exited Yes Molnlycke Health Care AB SE Sweden Exited Yes NVS Installation AB SE Sweden Exited Yes Pelly Intressenter AB SE ; SE Sweden Exited * Yes Phadia AB SE Sweden Exited 2005* No Plastal Group AB SE Sweden Exited * No Thule AB SE Sweden Exited No Tolerans AB SE ; SE Xlent Consulting Group Sweden Exited * No 2899 SE Sweden Exited Yes A/S Cimbria DK Denmark Exited Yes Aalborg Industries A/S DK Denmark Exited Yes Contex Holding A/S DK Denmark Exited No Logstor A/S DK Denmark Exited No Schades A/S DK ; DK Takeda Pharmaceuticals International GmbH Denmark Exited Yes 4676 DK Denmark Exited No Aibel AS NO Norway Exited No

105 27 Handicare AS NO Norway Exited Yes Kongsberg Automotive Holding ASA NO Norway Exited Yes Roxar ASA NO Norway Exited Yes SecuriNet AS NO Norway Exited No StormGeo AS NO Norway Exited * Yes Via Travel Group ASA NO Norway Exited No Viking SeaTech NO Norway Exited No Wonderland AS NO Norway Exited No

106 Table 22: Not exited observations of the sample The table contains the not exited SBOs and the relevant information of the observations Obs. # Portfolio Company Company ID Country of the portfolio company 35 Alimak Hek Group AB SE ; SE Sweden Status of SBO Not exited 36 Almondy AB SE Sweden Not exited 37 Attendo AB SE ; SE Aura Light International AB 39 Coor Service Management AB Sweden Not exited SE Sweden Not exited SE ; SE Sweden Not exited 40 DIAB Group AB SE Sweden Not exited 41 Euroflorist Sverige AB SE Sweden Not exited 42 GCE Holding AB SE ; SE Inflight Service Europe AB SE ; SE Sweden Sweden Not exited Not exited 44 Inwido AB SE Sweden Not exited 45 Kwintet AB SE Sweden Not exited 46 Lekolar AB SE Sweden Not exited 47 LGT Logistics SE Sweden Not exited 48 Multicom Security AB SE Sweden Not exited 49 Q- MATIC Sweden AB SE Sweden Not exited 50 Semantix AB SE ; SE Sweden Not exited 51 Textilia AB SE Sweden Not exited 52 Dansk Industri Syndikat A/S DK Denmark Not exited 53 Icopal a/s DK Denmark Not exited 54 Glud & Marstrand A/S DK Denmark Not exited 55 Saferoad NO ; SE Norway Not exited 56 Get AS NO Norway Not exited 57 Helly Hansen AS NO Norway Not exited 58 InfoCare ASA NO Norway Not exited 59 Jernia AS NO Norway Not exited Entry PE1 Exit PE1/ Entry PE2 Exit PE2 Cons. annual reports 4- digit NACE Code * Yes * Yes * Yes * Yes * Yes * Yes * No * Yes * Yes * No * Yes * No * No * Yes * Yes * No * No * No * Yes * Yes * Yes * No * * No * Yes * Yes

107 60 Jøtul AS NO ; NO Norway Not exited 61 Lindorff Group AB NO Norway Not exited 62 Nopco Paper Technology AS NO Norway Not exited 63 Plantasjen ASA NO Norway Not exited * Yes * No * No * No

108 Appendix 3 Test assumptions and algorithms Distribution of differences around the median of zero. The differences are of observations, for each measure of operational performance, for the total sample industry adjusted. 106