Private Equity backed IPOs - Performance, Corporate Governance & Management Control

0 University of Groningen

Faculty of Economics and Business

Private Equity backed IPOs -

Performance, Corporate Governance &

Management Control

Master Business Administration – Finance & O&MC

Coen H. des Tombe *

Master Thesis MSc BA Finance & O&MC

ABSTRACT

In this exploratory study we examine performance, corporate governance structures and management control processes of PE-backed IPOs in Europe. We consider a sample of 221 firms from 2003 to 2010 with a unique corporate governance dataset. In this study we perform a cross-section analysis by applying OLS, analysing how PE set corporate governance characteristics influence post-IPO performance. In addition, we analyse three in-depth case studies to explore how corporate governance and management control are interlinked and influence performance. Our results show an average long-run underperformance of PE-backed IPOs relative to their sector. In addition, we show that post-IPO ownership structures and board characteristics have a significant influence on performance. We find differences in performance and corporate governance characteristics between common law and civil law affiliated PE-backed firms. Moreover, we find significant differences in performance and effects of corporate governance characteristics between VC-backing and PE-backing. Furthermore, we identify that PE sponsors bring more value in the firm on the way in than on the way out. As a result, PE-backed firms in the long-run may remain with negative effects of the pre-IPO implemented strategies.

Keywords: PE-Backed IPOs, Performance, Corporate Governance, Management Control JEL-Codes: G14, G24, G31, G32 and G34

December, 2013

Dr. B. Qin

*

Coen Henri des Tombe (S1692585)

Master student at the Economics and Business Faculty of the University of Groningen

1

ACKNOWLEDGEMENTS

2

Table of content

1. Introduction 3

2. Literature review 5

2.1. Private equity performance 5

2.2. Corporate governance mechanisms in private equity 6

2.3. Management control processes in private equity 10

3. Research methodology 12

3.1 Quantitative research 13

3.1.1. Variables and measurement 13

3.1.2. The model 16

3.1.3. Data 16

3.1.4. Data processing 17

3.2 Qualitative research 18

3.2.1. Research design 19

4. Empirical results quantitative analysis: Performance and corporate governance 21

4.1. Univariate analysis 21

4.2. Bivariate analysis 22

4.3. Multivariate analysis 23

4.4. Summary quantitative analysis and robustness checks 29

5. Qualitative analysis: Corporate governance and management control 29

5.1. Case 1: Ziggo 30

5.2. Case 2: Wavin 33

5.3. Case 3: Stentys 36

6. Discussion 38

6.1. Performance 38

6.2. Control throughout the value step-up process 39

7. Conclusion 44

7.1. Conclusion 44

7.2. Limitations and further research 45

8. References 46

9. Appendices 50

Appendix I: Literature review 50

Appendix II: Research Methodology 53

Appendix III: Quantitative analysis 54

Appendix IV: Qualitative analysis 59

3

I. INTRODUCTION

The private equity (PE) sector is an important source of funds for start-up firms, firms in financial distress, middle market firms, and divisions of public firms seeking buyout financing (Jelic et al., 2005; Cumming et al., 2007). The sector has emerged spectacularly over the last 30 years both in the US and Europe (Wright et al., 2009; Gompers and Lerner, 2001; Wright et al., 2006). The first big wave of LBO transactions occurred in the mid-1980s, whereas the second and even larger wave took place between 2003–2007 (Kaplan and Strömberg, 2008; Wright et al., 2009). According to the ‘Bain Private Equity Report 2012’, private equity owned companies increase their value faster than their competitors. The rise of the private equity market poses critical research questions regarding how these transactions increase private and social value.

A common way for PE sponsors to exit their investments is through IPOs on capital markets (Lerner, 1994). Empirical evidence suggests that they choose this exit channel strategically and build up reputation primarily through successful IPOs (Levis, 2010). PE-backed firms that go public can also drive up firm performance by building relationships with top-tier financial institutions that at least partly mitigate informational asymmetries at the time of the IPO. Since PE sponsors tend to hold significant ownership and board positions (Bruton et al., 2010), and continue to be involved in the firm after going public (Cao, 2009), they might also be able to provide better access to capital in the post-IPO period. According to Bergström and Nilsson (2006), Katz (2009) and Levis (2010), PE involvement has a positive influence on the firm’s post-issue operating and market performance. Moreover, various studies show that specific corporate governance characteristics play an important role in listed firms and have influence on post-IPO performance.

Filatotchev and Wright (2005) find that examination of IPOs offers potential for more insightful analysis of corporate governance structures and management control processes since structures and controls of the firm at the moment of listing is clearer than at any point in the firm’s history. Consistently, Beatty and Zajac (1994) argue that studying IPO firms ‘provide a particularly clear test of the agency-based, contingency perspective’. Moreover, Von Eije et al. (2000) show significant changes in governance and management controls at the value step up of non PE-backed IPOs.

This study applies an explorative research approach in which we highlight and review theoretical starting points of research evidence on the effects of private equity on performance, corporate governance and management control processes of PE-backed firms in Europe. The aim is to: (i) analyse long-term market performance of PE-backed IPOs compared with their respective sector index; (ii) examine the influence of specific corporate governance structures on PE-backed IPO’s long-term performance; (iii) analyse in-depth how PE sponsors organise corporate governance structures and management control processes of PE-backed IPOs; (iv) create a complete view on how PE set corporate governance structures and management control processes affect firm performance. This can be translated into the following research question: How can corporate governance structures

and management control processes affect PE-backed IPO’s value step-up and long-term market performance?

4

Netherlands. They find that underpricing is affected positively in venture capital backed companies that exhibit large net earnings growth and negatively in venture capital backed companies that exhibit small net earnings growth.

In global literature, numerous studies evaluated the impact of PE involvement on post-IPO performance. However, modest research is performed on PE set corporate governance structures and their effects on performance, and barely research exists on management control processes and their effects on PE backed firm’s performance. The bulk of existing research is US- or UK- based and little work is conducted on European civil law countries. As corporate governance in Europe is different compared to the US and UK, European-based research in this area is scarce and differs across types of IPOs and performance estimations. Furthermore, most empirical research is based on quantitative datasets where they are unable to explore process-related aspects of private equity. There is a growing appreciation for qualitative studies in corporate governance research, providing an important contribution in understanding governance processes and mechanisms in private equity.

Therefore, this paper will explore and conceptualize performance, private equity set corporate governance mechanisms and control processes of European PE-backed IPOs through a quantitative and qualitative analysis. In total, 15 widely discussed independent variables are selected in order to answer the research question. These variables can be divided into four groups: (i) ownership identity; (ii) ownership structure; (iii) board characteristics and; (iv) leverage. The unique corporate governance dataset contains data manually collected from different sources. In the quantitative study, a cross-section analysis will be performed through OLS regressions. Secondly, a qualitative study will be performed through three in-depth case studies of PE- and VC- backed IPOs through interviews and extensive research on the firms, in order to analyse how corporate governance structures and management control processes in PE-backed firms are set by PE sponsors and in the end influence the firm’s performance.

Contributions of this thesis include: (i) combining an analysis of post-issue performance with an in-depth analysis of corporate governance structures and management control processes of PE-backed firms through a quantitative and qualitative study; (ii) based on existing research, proposing a model of the most important corporate governance variables which previously have only been researched individually; (iii) identifying differences in performance and corporate governance structures between common law UK and civil law Europe based PE-backed firms; (iv) identifying differences in corporate governance structures and management control between PE- and VC- backing in Europe and; (v) identifying the effects of the credit crisis on PE-backed IPOs.

5

II. LITERATURE REVIEW

In this section we explore and highlight important literature regarding the influence of PE-backing on performance, corporate governance structures and management control processes. We summarize the most important corporate governance related PE literature in table 1 of the appendix 1.

2.1. PRIVATE EQUITY PERFORMANCE

For the purpose of this study, a PE-backed firm is defined as a company where a PE sponsor has acquired a controlling interest backed by substantial borrowings, also called a leveraged buyout (LBO). Private equity involves investment in listed or unlisted companies and includes both early stage and later stage buyouts. Early stage buyouts, called venture capital (VC), are companies that receive start-up, development, or expansion backing. Later stage buyouts, called private equity (PE), are generally more mature companies, either listed or unlisted. In contrast to PE sponsors, VC sponsors usually have a minority interest and are prepared to invest for longer periods. After acquisition, sponsors become active investors taking board seats driving more efficient use of organizational resources, debt and other financial instruments (Citron et al., 2006).

PE sponsors exit their portfolio firms through either selling to financial or strategic buyers, or through capital markets by an IPO, called a PE-backed IPO (Rindermann, 2004). Empirical evidence shows that PE sponsors choose the exit channel strategically and build up reputation primarily through successful IPOs (Gompers, 1996). In addition, portfolio firms build relationships with top-tier financial institutions that mitigate informational asymmetries at the time of the IPO, reducing underpricing (Acharya et al., 2009).

Post-IPO, PE funds tend to retain share ownership and continue to be involved in the portfolio firm through taking board positions (Filatotchev et al., 2006; Bruton et al., 2010). In addition, sponsors are also able to provide better access to capital in the post-IPO period. Given these characteristics private equity involvement should have a positive influence on portfolio firms’ post-issue operating and market performance (Acharya et al., 2009). Nevertheless, according to Bergström and Nilsson (2006), long-run performance patterns differ across weighting methods and stock exchanges.

Levis (2010) shows outperformance of PE-backed IPOs in the UK over a 36-month period across different benchmarks and estimation procedures during 1992-2005. Consistently, Bergström and Nilsson (2006) and Bruton et al. (2010) find evidence that PE-backed IPOs in the UK outperform non-PE-backed IPOs across all different time horizons on an aggregated level. Meier and Möhlmann (2010) show that as the period of time increased, German PE-backed IPOs outperformed the market to a greater extent relative to the non-PE-backed sample.

6

2.2. CORPORATE GOVERNANCE MECHANISMS IN PRIVATE EQUITY 2.2.1 Agency relationships

The bulk of the private equity literature follows on the assumptions of the mainstream finance approach to corporate governance, specifically agency theory (Jensen, 1993). The theory refers to the problem of ensuring that managers really act as agents of owners maximizing shareholder value rather than pursuing their own interests, for instance pet projects and empire building. Changes in ownership structure, for instance an LBO, can align interests more effectively and are likely to optimize economic performance (Gordon and Roe, 2003; Rollinson and Dundon, 2007). Consistently, Masulis and Thomas (2008) find evidence that PE involvement help offset agency problems and reduce agency costs. Firstly, LBOs reduce manager’s discretion to misuse free cash flow by ensuring that they must repay debt service payments, emphasizing efficient operations and creating strong personal incentives to work hard to avoid bankruptcy. Secondly, agency cost reductions arise out of a strong realignment of managerial incentives, which focuses executives’ efforts more sharply on performance and value. Thirdly, agency costs are reduced by increased board monitoring as a result of much stronger financial incentives of directors and improved internal reporting.

2.2.2 Institutional context

Corporate governance characteristics of firms differ across the world and are rather country-specific than firm-specific. Aguilera and Jackson (2003) and La Porta et al. (2000) show differences in national institutions, in particular between Anglo-Saxon common law systems and European civil law systems, influencing the effectiveness of corporate governance on firm level. La Porta (2000) argues that investor rights are stronger in common law systems than in their civil law counterparts; the former encouraging practices that are more receptive to owner interests. Specifically, Anglo-Saxon companies consider shareholders as the only owners of the company whereas in Europe, besides shareholders, all stakeholders are taken into account including customers, suppliers and employees. Furthermore, in common law systems investors are willing to take more risks and use arms-length control mechanisms since they have more legal remedies, while in civil law environments fewer legal remedies are available (Bruton et al., 2010). Regarding PE-backed firms, Bruton et al. (2010) and Rindermann (2004) support the argument that institutional affiliation moderates the impact of PE-backed firm’s performance. In addition, Meuleman and Wright (2011) highlight differences in governance of VC-backed firms due to institutional affiliation. In particular, the US, UK, France and the Netherlands differ in monitoring, strategy and networking roles.

2.2.3. PE corporate governance mechanisms

Although there is growing evidence on the effects of governance structures of private equity and their portfolio firms in different countries, these insights have mainly been researched individually and have not been matched by findings into the international cross-border governance mechanisms and processes of private equity value creation. In this section, we emphasize key private equity governance mechanisms identified by existing literature.

Ownership identity

7

As mentioned earlier, private equity covers different types of investors, in particular early stage venture capital and later stage private equity investments, which may differentiate in governance mechanisms and their impact on agency problems (Osnabrugge, 2000). Portfolio firms of PE sponsors tend to be larger in enterprise value, use more debt and sponsors generally hold a majority stake controlling the board of directors. VC sponsors generally acquire a minority stake and invest mostly in technology industries (Katz, 2009). Furthermore, PE sponsor representatives tend to have a more financial or consulting background whereas VC managers are often successful entrepreneurs or possess specialized technology expertise (Fraser-Sampson, 2007; Klein and Zur, 2007). According to Bruton et al. (2010), VC-backed IPOs outperform PE-backed IPOs in the UK and France between 1996-2002. However, Levis (2010) shows that VC-backed firms outperform the market less than PE-backed firms in the UK between 1992-2005. Consistently, Katz (2009) shows that VC-PE-backed firms have less earnings quality than PE-backed firms in the US.

According to Cao and Lerner (2007) and Kaplan and Schoar (2005) large PE sponsors, those with greater assets under management (AuM), are expected to establish better governance practices due to sharing of best practices, exercising closer monitoring, and thus be associated with better earnings quality and performance in the post-IPO period (Katz, 2009; Ivanov et al., 2008). Kaplan and Schoar (2005) show that PE fund size, as a reputation proxy, has a positive influence on the IRR of portfolio firms in the US between 1980 and 2001. Furthermore, PE sponsor age is a second reputation proxy, where higher age is a sign of successful track record and buildup of assets under management over time. Acharya et al. (2009) show in their study of 66 PE-backed IPOs in the UK between 1996 and 2004 that PE-backed IPOs have greater operating outperformance the older the PE sponsors are. According to Masulis and Thomas (2008), potential agency costs arise out LBOs with multiple PE sponsors, also called club deals or syndication partners. Free riding by PE sponsors and disagreements in portfolio firm´s major policies may increase agency costs. However, Randoy et al. (2006) identify potential synergies between managers and multiple owners, referring to the resource dependency theory in which is emphasized that ownership characteristics might facilitate access to critical resources, such as capital, suppliers, or cooperative partners. Consistently, Archarya et al. (2009) find that club deals have a higher IRR than non-club deals.

Ownership structure

Ownership structure characteristics are referred to as how voting rights of the firm are divided (Mayer, 1992;) and can be a key to more effective corporate governance and shareholder value maximization (Shleifer and Vishny, 1986). In the context of PE-backed IPOs, although PE sponsors use an IPO as a way of exiting the firm, the involvement of PE sponsors is not completely terminated at the time of the IPO (Levis, 2010). Lockup agreements, performance incentives, and liquidity considerations often result in original PE sponsors retaining high, albeit reduced, holdings for a considerable period of time after flotation. This continuing involvement facilitates closer monitoring and reduces information asymmetries and potential conflicts with other stakeholders, leading to better operating performance and, depending upon initial valuations, market performance (Levis, 2010).

8

longer lockup agreements result in better post-issue short-term performance of PE-backed IPOs. However, Bruton et al. (2010) show negative effects from lockups on post-issue performance.

High share retention by PE sponsors is an important governance proxy that enhances post-issue performance and reduces the negative effects of the ‘IPO discount’ arising from agency conflicts (Bruton et al., 2010). According to Katz (2009), who examines 147 PE-backed firms, find that PE sponsors on average own more than 50 percent of total deal equity immediately after IPO, thus controlling the firm. The study shows that firms with majority ownership by PE sponsors experience better earnings quality than those with minority ownership by a PE sponsor. Consistently, Filatotchev

et al. (2006) find similar effects on post-issue stock performance in their study of VC-backed IPOs in

the UK.

Furthermore, Jain and Kini (1994) and Cao (2009) show that longer substantial PE share retention result in better post-issue operating performance. Consistently, Levis (2010) and Bruton et al. (2010) show that PE share retention has a positive influence on long-term stock performance of PE-backed IPOs in the UK. Specifically, according to Cao (2009), buyout sponsors maximize benefits of control through a staged exit strategy and are more likely to continue to hold equity in high-quality cash-rich firms, as they are reluctant to release a high proportion of the future cash flows to outside investors.

Supervisory board and executive board

PE sponsors control portfolio firms through sponsor representatives in the supervisory board. The board is seen as the solution to the problematic aspects of manager-shareholder interactions (Davila and Foster, 2007). Directors carry out a monitoring role over proper use of organizations’ cash flow, the firm’s response to takeover threats, executive and senior management hiring and compensation, and finally representation of the shareholders. Consistently, Erhardt et al. (2003) point out that supervisory boards are commonly the most influential actors determining strategy direction and decision making, inherent to their structural position. Furthermore, Archarya et al. (2009) analyses board effectiveness of PE-backed and non-PE-backed firms through 20 executive interviews, and highlight PE Boards’ aligned focus on value creation, their sharper clarity performance priorities, and the greater engagement and commitment of their Board members.

According to Archarya et al. (2009) and Cornelli and Karakas (2010), PE-backed firms have a smaller board size than non-PE-backed firms being more effective in decision making. Filatotcchev et al. (2006) find evidence that board size of portfolio firms in the UK are smaller relative to their sector, more productive, and have a positive influence on post-issue performance.

Private equity sponsors typically become active investors through taking board seats. Board share can be defined as the number of directors in the board who represent PE sponsors after an IPO (Bruton et

al., 2010). According to Katz (2009), PE sponsors in the US account for 38.35 percent of the board in

9

balance between collaboration and confrontation in addressing issues related to the ‘professionalization’ of portfolio companies is identified as an important area for future research. Gong and Wu (2011) add to evidence on the role of PE sponsors in executive replacement by studying a sample of 126 US LBOs. They find that 51 percent of CEOs are replaced within two years after LBO. Consistently, Archarya et al. (2009) show that sponsors are more likely to replace CEOs in companies with high agency costs and if pre-LBO performance is low, in order to align goals. Remarkably, boards of post-LBO companies are likely to replace entrenched CEOs indicating that types of investors may have a direct impact on other governance factors.

Incentives encompasses all financial rewards that are contingent on firm performance, such as executives' stock options, but also the change in value of the executives' equity ownership, which varies with firm performance. According to Filatotchev and Bishop (2002), in a study of 251 IPOs in the UK, executive compensation effects of PE-backed firms are different than non-PE firms. The CEO receives a lower total compensation and higher pay-for-performance. Consistently, Kaplan and Strömberg (2008) find evidence that PE-owned companies use much stronger incentives for their executives. However, they also show that compensation differences between PE-owned companies and public companies disappear over a very short period (one to two years) after the PE-backed IPO. Retained executive share ownership is widely acknowledged as a potential sign of the high quality of the firm (Certo et al., 2001). In addition, it is argued that retained ownership results in more alignment of goals and less separation of ownership and control. Jain and Kini (1994) find evidence that executive ownership retention in PE-backed firms has a positive influence on operating performance. Consistently, Filatotchev et al. (2006) show that on average 14 percent of the executives retain equity ownership in VC-backed IPOs in the UK. In addition, they found that it has a positive influence on firm performance. Furthermore, Archarya et al. (2009) identifies that increased executive co-ownership provides strong incentives for managers to improve operating performance as due to emotional involvement, executives may be more strongly interested in the success of the firm. However, a study of 147 PE-backed IPOs in the US by Katz (2009), found that CEO co-ownership had a negative effect on earnings quality. Furthermore, it is argued that equity ownership and stock option compensation have a negative relation in PE-backed firms, where co-ownership is experienced to have a more positive effect on post-issue performance of PE-backed IPOs than stock options (Beatty and Zajac, 1994).

Leverage – Performance and governance mechanism

The use of debt, also called ‘leverage’, increases the returns on equity invested, as debt has a limit to its return (its interest rate) whereas equity owns the amount that is ‘left over’ after everyone else has been paid. A PE transaction is always done with the least possible equity, complementing the remaining sum with leverage. Cao and Lerner (2009), Kaplan and Strömberg (2008), Meier and Meuleman (2010) show in their studies that leverage has a positive effect on stock performance. According to Acharya et al. (2009), outperformance of stock returns of portfolio firms can be attributed to leverage as it provides the ability to drive EBITDA growth ahead of their peers for similar levels of sales growth.

10

incentive alignment of management improving firm efficiency because it (i) makes stock much more sensitive to improvements in firm value; (ii) motivates managers to use firm cash conservatively; (iii) eliminate underutilized assets so as to minimize the risk of bankruptcy, financial distress, and the accompanying forced management turnover and; (iv) is considered as a bonding mechanism for the management (Masulis and Thomas, 2008). Rindermann (2004) and Denis and Denis (1995) show that leverage improves operating performance, inferring operational improvements due to management incentives and creditor monitoring.

2.3 MANAGEMENT CONTROL PROCESSES IN PRIVATE EQUITY

2.3.1.Management control

Corporate governance structures and management control are inextricably linked and changes in corporate governance mechanisms will usually have immediate effects on the effectiveness of management control (Merchant and Van der Stede, 2007, p.577). Corporate governance focuses on controlling the behaviours of top management and, through their direction, those of all the other employees in the firm. Management control takes the perspective of top management and asks what can be done to ensure the proper behaviours of employees in the organization. The controlling of top management is performed through the supervisory board whereas employees are controlled by top management.

Management control systems are defined as formal and informal mechanisms and processes used for measuring, controlling and managing performance, implementation of strategies and ultimately for achieving their overall objectives. Simons (1995) views management control systems essentially as a means to successfully to implement strategy.

Management control systems facilitate growth (Simons, 1995; Flamholtz and Randle, 2000). The need for these systems have been argued from agency and information-processing perspectives (Baiman, 1982; Narayanan and Davila, 1998). Firstly, as companies grow, direct observation of the agent’s effort becomes too costly and motivation and monitoring have to happen through appropriate systems. Secondly, company growth also affects the ability to move information to the right decision makers. Without formal systems, the number of interactions required to move information around the company increases exponentially with the number of employees and communication becomes too costly. For this reason, management control systems become a very important element in the process of becoming listed.

11

management control, highlighting the previous mentioned controls by Merchant and Van der Stede (2007).

According to Von Eije et al. (2000), important management control changes due to the value step-up process (IPO) include financial management and financial reporting, as they comprise changes in effectiveness and efficiency, planning and control, capital budgeting and internal communications (on company performance). Companies where the respondents reported changes in these organizational variables showed an above average long-term performance. In addition, IPOs appear to develop and apply more financial participation by employees than non-public firms (Von Eije et al., 2002). According to Filatotchev and Wright (2005), assessment of IPO firms offers in-depth revision of corporate governance structures and management controls since structures and controls of the firm at listing is likely clearer than at any point in the firm’s history.

2.3.2 Private equity and management control

Private equity literature that focuses on management control processes is very scarce. The majority of literature is focused on the impact of PE investors on performance of firms. What remains less researched is the relationship between private equity investors and the main stakeholders, such as employees, and the structure and processes involving LBOs (Wright and Van Aelst, 2009). Many studies argue that in order to get a better view on management control processes, these can only be analysed through time-consuming qualitative studies, in particular case studies. Single studies tried to form starting points or bridges between PE literature, corporate governance literature and management control literature.

Bloom et al. (2009) provide an examination of types of managerial practices through interviews in a sample of 222 PE-backed and non-PE-backed firms, particularly non-listed firms in the manufacturing sector in the US, Europe and Asia. They find that private equity owned firms are better managed than other types of firms. However, relative to listed firms with dispersed owners, the difference is not significant. In addition, looking at management practices in detail they find that PE-owned firms have strong people management practices (hiring, firing, pay and promotions), target management practices with tough targets integrated across the short- and long-run, well understood by the employees and linked to firm performance. In addition, regarding operational management practices, PE-owned firms are even better in the adoption of modern ‘lean manufacturing’ practices, using continuous improvements and a comprehensive performance documentation process.

Davila and Foster (2007) examine 78 early-stage start-up companies, PE-backed and non-PE-backed, through different methods. They indicate that the number of employees, presence of venture capital, international operations, and time to revenue are positively associated with the rate of adoption of management control systems. Key management control systems include financial planning and financial evaluation.

Wood and Wright (2009) and Wright et al. (2009) highlight in their literature reviews, based on US- and UK-based studies, there is little direct evidence of PE sponsor’s influence on managerial factors such as the experience and expertise of portfolio company lower management, employment and PE representatives involvement in monitoring the portfolio firm in the aftermarket performance. However, investments in communication and control systems facilitate a better monitoring by the management and facilitate the possibility to make rapid and better adjustments to the strategy if needed.

12

addition, evidence shows that compensation and debt differences between PE-owned companies and public companies disappear over a very short period (one to two years) after the PE-owned firm goes public. They raise questions about whether and how PE firms put in place incentives and controls they to create value.

Regarding employment, Goergen and Renneboog (2007) find that LBOs in the UK are accompanied by falls in employment without there being a corresponding increase in productivity and profitability. However this is difficult to measure since firms have divestures during the process. However, Amness

et al. (2008) find no effects of LBOs on employment in the UK. Gospel et al. (2008) extend this

analysis of private equity firms with distinctive governance mechanisms. They find substantial effects on employment using case studies from three European countries. Negative effects amongst others include employment reduction. However, they also show benefits in terms of work organisation and employee voice. Finally, they find that national systems of labour regulation do not seem to impact the scope of employment reductions but do have effect on employee voice.

Finally, the study by Archarya et al. (2009) discusses that management control mechanisms of PE-backed firms need to be further researched including: (i) the provision of focused incentives subjecting management to key performance indicators (KPIs); (ii) drafting of value-creation (‘100-day’) plans that serve as the initial blue-print of a company’s strategic and operational agenda; (iii) PE-backed firms employment of external support and; (iv) PE formal and informal interactions in the post-issue period.

III. RESEARCH STRATEGY AND METHODOLOGY

Cooper and Schindler (2006, p.139) give different ways for classifying research: an exploratory research and a descriptive approach. Generally both are used with a clear emphasize on one of them. Exploratory research helps in determining the best research design, data collection method and selection of subjects. The results usually are not particularly useful for decision-making but provide significant insight into a given situation. Descriptive studies test clearly stated hypotheses or answer investigative questions and are more formalized.

In addition, different research strategies can be used including quantitative research and qualitative research (Bryman and Bell, 2011, p.25) . Quantitative research can tell us "what", "how often" or "how many" specific corporate governance structures occur while qualitative research can give indications "why", "how" and "when" something occurs (Yin, 2009, p.6). Specifically, Bryman and Bell (2011, p.27) discuss dilemma´s in choosing research strategy, in particular distinguishing between quantitative and qualitative research, the former focussing on statements of reality through generalisation of research results, whereas the latter focuses on understanding phenomena and their complexity in their specific context.

In this explorative research we perform a quantitative study and qualitative study, both acting complementary, addressing two questions:

(a) Which corporate governance structures and management control processes do PE sponsors use in European PE-backed firms?

13

Previously, building on existing research, this study identified key corporate governance variables in private equity literature that operationalise the model for the quantitative analysis. Apart from the empirical research based on large datasets, qualitative, case study-based research is used to explore process-related aspects of private equity that empirical studies cannot capture because of their design. In this way, we provide a full understanding of governance mechanisms and management control processes in PE-backed firms.

Case study-based research is becoming increasingly important in PE literature. In addition, there is a growing appreciation of qualitative studies in corporate governance research in general since traditional, finance-driven empirical studies are unable to provide a full understanding of governance processes and mechanisms (Bacon et al., 2010).

This section is further structured as follows: (i) in the quantitative research strategy we explain the variables and measurement, the model, data and data processing; (ii) in the qualitative research strategy we discuss the case study methodology, case research design and data.

3.1. QUANTITATIVE RESEARCH

The key questions we want to answer with the quantitative analysis are: (i) do European PE-backed firms outperform the market and; (ii) how much of the performance generated by PE-backed firms comes from specific corporate governance structures. Building on existing PE literature, mostly US- and UK-based studies, discussed in the previous sections, a selection of key corporate governance variables effecting PE-backed firm’s performance are chosen.

3.1.1. Variables and measurement

Dependent variable – Post-issue long-run performance

In the analyses we use post-issue long-term stock performance, consistently with Bergström and Nilsson (2006), Rindermann (2004) and Filatochev et al. (2006). Stock performance is chosen over accounting measures as these can be subject to many forms of manipulation and are not perfectly correlated with real performance. For instance, the value of high-tech firms is related to growth opportunities and expectations of future profitability. Therefore it seems appropriate to analyse stock returns, whether the market recognizes the value-added potential of PE- or VC-backing in firms going public. Secondly, stock performance is chosen over profitability measures as these measures change following buyouts and IPOs, confounding value creation and value capture. For instance measures scaled by total assets increase substantial at IPO, imparting a downward bias. Measuring stock performance is based on the efficient market hypothesis stating that a firm’s stock price reflects all (new) information related to the value of a company. Academic literature mainly employs measurement periods of either 36 months or 60 months. We measure performance over 36 months due to data limitations and once many PE-backed stocks become delisted after 36 months. For robustness sake, we also employ 12-month and 24-month stock performance.

14

sector index or value-weighted MSCI European index computed over 12-, 24-, and 36-month time periods.

In order to calculate stock performance, academic literature mainly employs two methods to calculate long-horizon abnormal returns: cumulative abnormal returns (CARs) and buy-and-hold returns (BHRs). We use the cumulative abnormal returns method for comparison reasons (Ritter, 1991; Bergström and Nilsson, 2006; Mian and Rosenfeld, 2004). The initial returns are excluded from the calculations for two reasons: (i) not all investors are allocated shares in the IPO and; (ii) initial returns may incorporate effects that do not concern the true value of the issuing firm (Bergström and Nilsson, 2006). We calculate monthly abnormal returns in the event period:

bt it

it R R

AR   (1)

where ARi,t is the monthly abnormal return of IPO, Rit is the continuously compounded monthly stock

return and Rb,t is the continuously compounded monthly benchmark return over time period t. The monthly abnormal returns are then aggregated through time called cumulative abnormal returns, calculated according to the formula below:





AR

i

CAR

)

(

As mentioned earlier, there is a debate on the best way to measure an IPO’s long-term performance. Possibly, a different benchmark or method to calculate long-term performance could lead to different results.

Explanatory variables

In this analysis we use several corporate governance proxies based on mostly Anglo-Saxon studies which distinguished these to be of major influence and explore their effects on the performance measures. Most variables were previously analysed independently or a few together. We distinguish between four types of governance mechanisms that have major influence: (i) ownership identity, (ii) ownership structure; (iii) supervisory board and executive board characteristics; and (iv) leverage. Panel A of table 1 shows an overview of the corporate governance proxies analysed in this research, their definition and their expected influence on post-issue long-term performance.

Control variables

15 Table 1. Overview proxies, definition and expected outcome

Panel A: Explanatory variables

Variable Definition Expected

outcome

1.1. Venture capital Equal 1 if VC-backed, 0 otherwise +

1.2. Syndication partners The number of PE sponsors or VC sponsors in a portfolio firm that took part in the syndicate (excluding co-investors)

+/- 1.3. Sponsor reputation 1. Log age of PE sponsors or VC sponsors at time of IPO. Highest age is

taken if more than one sponsor was involved

2. Assets under management (AuM) of PE sponsors. Highest is taken if more than one sponsor was involved. Categorized: (i) 0<5bn; (ii) 5<15bn; (iii) 15<30bn and; (iv) 30bn<

+ +

2.1. Lockup periods Days sponsors and insiders are forbidden to sell any of their shares after IPO. Categorized (i) <90 days; (ii) 90<180 days; (iii) 180 days <

+/- 2.2. Majority ownership

dummy

A dummy that equals 1 if above 50 percent of shares are retained by sponsors at IPO, 0 otherwise

+ 2.3. PE share retention 1. The post-IPO PE share retention period

2. Average percentage of share retention during post-IPO holding period

+ 3.1. Board size The total number of directors on the board a year after IPO - 3.2. PE board share Ratio of the number of PE or VC representatives in the board divided by

total board size a year after IPO

+ 3.3. PE board activity The period PE representatives stay active in the board post-IPO + 3.4. Executive board

share retention

1. Percentage share retention immediately after IPO

2. A dummy that equals 1 if managers own shares, 0 otherwise

+/- +/- 3.5. Executive turnover Executive turnover during pre-IPO holding period +

4. Leverage ratio Average debt to equity ratio over post-IPO period +

Panel B: Control Variables

Variable Definition Expected

outcome

1. Firm size dummy Log of total assets at IPO +

2. Technology dummy Equals 1 if the IPO-company belongs to the technology sector based on industry codes, 0 otherwise

+ 3. Cold market/ crisis

dummy

A dummy that equals 1 if firms went public in the crisis period (2008-2010), 0 otherwise

+/-

4. UK dummy Equals 1 for UK and Ireland based IPOs, zero otherwise +

Notes (a): Category: (1) ownership identity variables; (2) ownership structure variables; (3) supervisory board and executive board variables; (4) leverage variable

Kaplan and Strömberg (2008) and Acharya et al. (2009) show counter-cyclicality in fundraising and performance, finding positive and stronger returns on average and robust during downturns in the UK. Nevertheless, the credit crisis may have different effects on PE-backed IPOs than other down cycles due to its scale and global impact. Finally, in this analysis we also control for effects related to institutional affiliation as discussed in the previous section. We create a dummy for common law country based IPOs, specifically the UK and Ireland (Rindermann, 2004; Bessler and Seim, 2011). Panel B of table 1 shows an overview of the control variables used in the quantitative analysis.

3.1.2. The model

16

OLS, the basis of the model is shown in equation 1 whereas a model with several explanatory variables is shown in equation 2:

          (1)

(2)

where y represents the dependent variable, the performance measure, is the constant and the coefficient of independent variable . Whereas most variables are based on US and UK literature, we explore which of those variables have significant influence in Europe through stepwise building. Variables are included one by one into the regressions, basing the candidate variables on a financial or economic theory based on the previous literature review.

After proposing the basis of the model, we use different interaction dummy variables in the model to detect interaction effects of UK institutional affiliation and venture-backing on firm performance. Effects of interaction dummy variables are the product of the original regressor of each explanatory variable and the binary UK or VC interaction dummy variable. Nevertheless, when applying interaction dummies, it is vital to use the intercept dummy in the regressions as well to avoid a bias in the estimates of interaction terms. Finally, we introduce intercept dummy variables to identify differences between civil law regions, specifically French, German, and Scandinavian civil law (Meuleman and Wright, 2011). To identify statistically significant differences between two groups, we conduct Student’s t-tests.

3.1.3. Data

The sample firms used in the quantitative analysis are a selection of PE-backed IPOs in Europe that completed an initial public offering on a major stock index from January 2003 to December 2010. The starting year 2003 is chosen as the beginning of a period of global economic growth, especially in the private equity sector (Bain, 2012) and because information regarding PE-backed IPOs in Europe is young and are mainly gathered from 2001. The initial sample of PE-backed IPOs that took place from 2003 to 2010 was screened for the following criteria: (i) the geographical location of the target firm is in developed Europe, in particular 17 countries including: Austria, Belgium, Germany, Denmark, Spain, Finland, France, United Kingdom, Greece, Ireland, Italy, Luxembourg, Netherlands, Norway, Portugal and Sweden, Switzerland; (ii) the target firm is listed in the EU as for instance European firms listed in the US need to fulfil the Sarbanes Oxley Act that has different rules than in the EU and; (iii) firms in the banking industry are excluded.

The corporate governance dataset is unique and manually collected from different sources. Firstly, to identify PE-backed IPOs, database ThomsonOne was used to identify PE-backed IPOs (n=295). However, the database was not reliable and had to be checked on the basis of information from the individual company prospectuses. In addition, Dealogic was used as a complementary database to form the final set of PE-backed IPOs. The databases also provide information including the identity of the PE sponsors, country, lockup periods, and pre-IPO holding period of the PE sponsor. Daily closing prices of stocks, market and sector indices and accounting data including leverage and total assets are gathered from Datastream. The closing prices, which are used to calculate the returns, are adjusted for dividends, stock splits, and new share issues.

17

investments. The final classification of VC- and PE-backed IPOs was completed on the basis of information from the individual company prospectuses. Data regarding retained ownership by PE sponsors are retrieved from Lion shares, a database that collects global equity ownership data. However, these were only available for currently listed firms. In addition, we manually collect IPO prospectuses and annual reports for information on the representation of the buyout group on the board of the firms and buyout organization’s and management’s ownership holdings.

From the original list of IPOs (n=295) over the period of 2003–2010, we excluded 67 firms after applying the selection criteria. An additional three firms were deleted due to the unavailability of data. Furthermore, four firms were omitted from the sample due to extreme outliers. The result is a final sample of 221 PE-backed IPOs for which we were able to gather all stock, IPO, PE sponsor and firm data. The smaller sample of 129 IPOs is used whenever IPO ownership and governance data are analyzed.

3.1.5. Data processing

Before running a multivariate regression analysis, variables were examined for problems related to their distribution and the assumptions of ordinary least square regression analysis (Brooks, 2008; Tabachnik and Fidell, 1996). An outlier analysis is essential as datasets may contain numbers not reflecting the real situation. An outlier analysis has taken place by drawing up histograms for each of the dependent and independent variables. Removing outliers from the original dataset, creates a more normal distribution and a better analyzable dataset.

Furthermore, variables are checked for normality of distributions through the Jarque-Bera test, which if significant and large, rejects the null-hypothesis and thus normality. Common methods resolving normality are taking the log, replacement by a dummy variable, trimming or winsorizing. Therefore we take the log of PE sponsor age and the log of total assets of the firms. In addition, we categorize PE assets under management and lockup periods. If a variable remained not normally distributed, normality was assumed, as the sample size is sufficient, according to the ‘central limit theorem’ (Brooks, 2008, p.164) which states that the sample mean converges to a normal distribution when sample sizes are sufficiently large.

Through stepwise building variables are included one by one into the regressions, basing the candidate variables on a financial or economic theory explained in the previous section, taking into account multicollinearity problems, and in the end leaving out several variables. Significant coefficients at

p=0.1 were used as a guide for adding variables as well as the Akaike information criterion. However,

a few variables also become significant after the inclusion of the intercept and interaction dummies. The significance of the regression models are tested with the Chow-test, comparing the impact of the pooled regressions by applying an F-test statistic. In forming the final models, we check the variance inflation factors for remaining multicollinearity. Finally, the models are checked for heteroscedasticity with the Breusch-Pagan-Godfrey test and if necessary tests were adjusted through White’s heteroscedasticity-consistent standard errors and covariance.

18

3.2. QUALITATIVE RESEARCH

As mentioned in the previous section, in this explorative research we firstly examine rather key PE common law corporate governance variables hold for European civil law countries through a quantitative analysis. The results are used as a basis for the qualitative research in which we further analyse in-depth processes of specific governance structures in European PE-backed or VC-backed IPOs. The research method followed in this qualitative research is the case study methodology.

3.2.1. Case study methodology

Case studies are an extension of empirical quantitative research. The justification for case-based research in the area of accounting and control systems is widely recognized (Ferreira and Otley, 2005; Scapens, 1990). Accounting and other controls cannot be fully understood in isolation and therefore a more contextual approach is required (Ferreira and Otley, 2005). According to Chenhall (2003), management control procedures function in complex organizational settings and the nature of controls requires their study within the particular settings in which they are used. Case studies seem to be an ideal vehicle for capturing these organizational activities in-depth in their organisational context as multiple data sources are used permitting a variety of factors to be captured in the descriptive material. In particular, to understand the context how management control systems function, phenomena on these aspects must be described in detail providing patterns of reporting relationships, decision making and control (Otley, 1999).

Case studies have been defined by Yin (2009, p.13) as ‘an empirical inquiry that investigates a contemporary phenomenon within its real life actual context’. The boundaries between phenomenon and context are unclear, resulting in more interesting variables than simply only measures, where multiple sources of evidence need to be used (Biemans and Van der Meer-Kooistra, 1994, p.54). Otley (1999) classifies case studies into four categories: (i) exploratory, explaining problems in the form of deductive (testing of hypotheses) or inductive work (making generalizations from observations resulting in theoretical statements); (ii) critical, criticising theory, proposing evidence that theory is wrong; (iii) illustrative, using existing theories to explain observations or (iv) accidental studies, when accidently gotten access to certain aspects or sources of an organization. In addition, Cooper and Schindler (2006, p.217) classify case studies basically in a descriptive or exploratory approach, often combined.

Furthermore, a common used research framework in management control literature is the extended performance and management control (PMC) framework by Ferreira and Otley (2005). The framework helps describing the structure and operation of management control systems (MCSs). In addition, it allows a speedy and comprehensive overview of many aspects and appreciation of the structure of control systems design in use. It describes the structure of the ‘package of controls’ that are being deployed by senior managers that are designed to help ensure that its strategies and plans are being effectively implemented in practice. The extended PMC framework is a combination of frameworks by Simons (1995) and Otley (1999).

3.2.2. Research design and data

19

extent that only the general aspects of the study design were determined in advance. This approach was considered particularly suitable given the general absence of prior knowledge regarding management control practices in private equity.

Specifically, outcomes of the quantitative analysis were used followed by the use of the extended Performance Management and Control (PMC) framework, a basis for further analyzing management control in the case studies. The framework is graphically shown in appendix II. The PMC framework specifically focuses on results controls. However we also highlight personnel controls and cultural controls were analysed, as these may also have influence.

The case studies are endeavoured to examine the functioning of corporate governance structures and management control systems in three medium to large (non-finance) companies located in Europe. The initial focus was on Dutch PE-backed firms. However, over the last 10 years only five PE-backed IPOs were performed in the Netherlands of which only one is currently listed. We extended our range to the Benelux and France. In addition, the selection process main goal was to obtain diversity so as to enrich the understanding of control practices in a variety of industries, sizes, ownership structures and management control systems design. For this reason, also following the results of the quantitative analysis, we consciously choose to select PE-backed as well as VC-backed firms.

Initially, a total of 8 PE-backed firms listed on a stock exchange in Europe were contacted through close contacts. In addition, due to outcomes of the quantitative analysis, a total of 3 VC-backed firms were also contacted through close contacts. Key-contacts were involved on a number of occasions and contacted using different means of communication (phone calls and e-mails). Initial contacts aimed to gain access to companies and further contacts were made to arrange interviews with key candidates. The goal was to meet either the CFO, controller or other top management, board of directors, or a sponsor representative. The interviewing process and visits to premises took place over a two month period.

As the research topic required candidates to be in the top level of the portfolio firm or a representative of a sponsor, three of the eleven contacts could provide us with the right candidates for interviews. The final case study sample exists of three firms from the Netherlands and France of which two Dutch PE-backed IPOs, Ziggo and Wavin, and one French VC-backed IPO, Stentys SA. The PE sample includes one service and one product firm while the VC sample includes a product firm. In all three case studies, interviews were conducted with executive and/or senior management. The following provides a short description of the firms:

(i) Wavin NV: was floated on the AEX stock index in July 2006 by PE sponsors AlpInvest and CVC. The firm engages in the manufacture and supply of plastic pipe systems and solutions for above and below ground applications primarily in Europe.

(ii) Ziggo NV: is a Dutch service provider of entertainment, information and communication through television, broadband internet and telephony services. Business to business customers use services such as telephony, broadband internet, television and data communication. Ziggo was founded in 2008 after a merger and in March 2012 it was floated to the market by PE sponsors Cinven and Warburg Pincus.

20

cardiologists. Stentys was founded in 2006 and is headquartered in Paris, France. The firm was floated to the stock exchange in 2010 by Soffinova and Omnes capital.

According to case study literature, it is important to verify information in case study research in multiple ways to improve research validity, called data triangulation (Yin, 2009, p.85), while others view validity and triangulation problematic as “it suggests certainty has been gained in the capture of an objective reality” (Ahrens and Chapman, 2006, p.834). In this case study research data triangulation was used to validate the research. Semi-structured interviews were the central part of the data collection process. Four interviews took place on companies’ premises and one interview took place by phone. Other material was collected and used at an earlier stage, including annual reports, website information and news. In addition, independent sources of information were surveyed for any relevant information, such as broker reports and online media news through news database Factiva and database Capital IQ. The different data collections per company are summarized in table 2.

The interviews were semi-structured without recording. The transcripts of these interviews were made right after the interview had taken place. The study included factual accuracy checking procedures where companies were provided with a preliminary version of each case study. These were sent to the interviewees who later reported back on the conclusions drawn that no major adjustments had to be made.

Table 2. Overview of data sources per company

Type Wavin NV Ziggo NV Stentys SA

Interview on-site 2 2 -

Phone interview - - 1

Internal documents Y Y N

Annual reports Y Y Y

Broker reports Y Y Y

Public available data Y Y Y

Bryman and Bell (2011, p.488) and Cooper and Schindler (2006, p.218) identify that case study research suffers from several limitations. Firstly, case studies lack generalization due to the nature of case studies. Secondly, they suffer from the problem of data availability and how data is collected, as firms have to be willing to cooperate. Nevertheless, these limitations need to be interpreted in the light of this exploratory study, making it otherwise impossible to provide deep understandings of the phenomena studied.

IV. EMPIRICAL RESULTS

21

4.1. UNIVARIATE ANALYSIS - DESCRIPTIVE STATISTICS

4.1.1. Post-issue market performance

Table 3 provides the absolute, market-adjusted and sector-adjusted performance of PE-backed IPOs in Europe. PE-backed IPOs in Europe show negative abnormal performance relative to the market and their sector through all three event periods. Portfolio firms in Europe do show positive median abnormal returns over a 12-month event period. There is a tendency of PE-backed IPOs to exhibit positive abnormal returns in the short-run and then decline in the long run. UK (common law) based firms perform worse than civil law based firms regardless of market or sector benchmark over all time horizons. The results for long-term under-performance of the entire sample are consistent with Bergström and Nilsson (2006) and Bjugert and Johanson (2004). Table 1 of appendix III shows the average abnormal performance per country and sector. Countries and industries show differing return patterns across different time horizons. Panel A shows that Austria, Denmark, Poland and France are the best performing countries over a 36-month event period while Norway and the Netherlands are the worst performing countries. In addition, panel B shows that best performing sectors are telecommunications, food and beverage, leisure and recreations, finance and insurance while retail and the oil and gas sector are the worst performing sectors.

Table 3. Average absolute, market-adjusted and sector-adjusted performance of PE-backed IPOs

IPO return (%) Market-adjusted (%) Sector-adjusted (%)

12- month 24- month 36- month 12- month 24- month 36- month 12- month 24- month 36- month Europe 14% 14% 11% -7% -23% -34% -6% -22% -33% (6%) (0%) (-17%) (2%) (-13%) (-19%) (0%) (-9%) (-20%)

Europe (Ex UK) 14% 13% 12% -4% -17% -20% -4% -17% -20%

(6%) (0%) (-17%) (6%) (-8%) (-7%) (1%) (-9%) (-7%)

UK 15% 15% 9% -12% -33% -58% -10% -32% -55%

(9%) (-4%) (-16%) (-4%) (-19%) (-27%) (-2%) (-16%) (-29%)

Notes: Average post-issue performance over a 12-, 24- and 36-month event period. Medians are reported in parentheses.

4.1.2. Firm and offering characteristics at the time of the IPO

A total of 329 sponsors were involved in the 221 PE- and VC-backed IPOs in Europe between 2003 and 2010. There was a total deal value of €30.4 billion (€60.2 billion including non-deal value) with a median deal size of €65 million in Europe and €58 million in the UK. Sponsors are involved with their portfolio firms for an average of 4 years before flotation, where in the UK this period is 4.3 years. Of the total sample, 35 percent are service firms. PE-backed IPOs are, on average, larger companies in terms of firm size (€950 million) than their VC-backed counterparts (over €200 million). PE-backed buyouts account for 74 percent in terms of volume and almost 80 percent in terms of the value raised at the IPO. Almost 80 percent of the IPOs involve only a single active sponsor (excluding co-investors), while the remaining 21 percent included multiple active sponsors (club deals), with a maximum of 4 sponsors. As of ultimo 2013, 42 percent of the total sample is delisted.

22

considerable difference in board share and board activity, the former 26 percent in Europe while 11 percent in the UK, the latter almost 2 years in Europe while 1 year in the UK.

Furthermore, panel B shows considerable differences across civil law regions in Europe: (i) lockup period; (ii) post-IPO holding period; (iii) PE share retention and; (iv) PE board activity. In particular, the Scandinavian civil law region shows a lower lockup period (a median of 0) and a higher degree of post-IPO holding period compared to the other civil law regions. In addition, the Scandinavian civil law region show long board activity compared to the other regions. Finally, the French civil law region shows almost twice the degree of share retention compared to the German civil law region.

Panel D shows differences between countries. Remarkable differences include: (i) Finland with a nearly twice as large pre-IPO holding period than the average of the sample; (ii) the Netherlands and Norway having almost twice as large post-IPO holding periods than the average of the sample; (iii) Spain, Switzerland and Finland show zero percent or very little post-issue share retention while Belgium shows the highest average share retention of 45 percent; (iv) Belgium and the Netherlands show the longest board activity and; (v) portfolio firms in Ireland have more than twice the degree of leverage than the average of the sample.

Finally, buyout and VC-backed firms show notable corporate governance differences in panel C, including: (i) VC-backed firms are younger and have lower assets under management; (iii) VCs show lower degree of lockup periods; (iii) VCs have a twice as small degree of post-issue share retention than buyouts; (iv) VCs show an average larger board size by one member while board share and board activity are similar and; (v) buyout firms have a larger degree of leverage than VC portfolio firms. Nevertheless, it is worth noting that unlike the US, where buyout and VC-backed firms are largely distinct, in Europe there is a major overlap among the sponsors in VC- and PE-backed IPOs, which is a possible explanation for the modest differences.

4.2. BIVARIATE ANALYSIS - CORRELATIONS

Table 4 of appendix III shows the correlation coefficients for the variables used in the regression models for the full sample. A bivariate analysis is done through Pearson correlations test, firstly examining correlations between the independent variables and performance measures and secondly, identifying highly correlated independent variables causing problems of multicollinearity.

Regarding correlations between the dependent and independent variables, ownership identity, in particular VC-backed portfolio firms, show a highly significant positive correlation with both performance measures. Secondly, ownership structure characteristics, in particular post-IPO holding period and PE share retention, have significant negative correlations with both performance measures while lockup period shows a significant negative correlation with the 12-month performance measure. Concerning board characteristics, board activity and board ownership have significant negative correlations with both performance measures. Finally, IPOs in the credit crisis have a significant negative influence on 12-month performance while firm size shows a significant positive correlation with long-term performance.

23

on economic theory highlighted in the literature review, in order to avoid multicollinearity problems with the board activity variable, still making it possible to add additional explanatory power to the models. Finally, board size shows high correlations with the firm size control variable, causing a possible problem of multicollinearity.

Remarkably, the VC variable shows significant negative correlations with PE assets under management, PE age, firm size, leverage and lockup periods. In addition, the UK variable shows significant correlations with board share and board activity. These (significant) correlations are not high causing multicollinearity problems, but can be taken into account through testing whether VC and UK interaction dummies increase the explanatory power of the regression models.

Although the correlation matrix identifies high correlations between the independent variables and performance measures, the regression analysis will decide whether these variables also show significant explanatory power when estimating the regression formula.

4.3. MULTIVARIATE ANALYSIS

To explore to what extent the observed market performance measures are related to the presence of specific private equity corporate governance variables, we estimate cross-sectional regressions with various explanatory and control variables, as outlined in section 3.1. We include an intercept dummy variable for the participation of UK-based firms to detect potential performance effects of institutional (common law) affiliation. In addition, we use different interaction dummy variables in the model specifications to reveal interaction effects of UK institutional affiliation and VC-backing. For robustness sake we also check interaction effects of the crisis dummy. The effect of interaction dummy variables are the product of the original regressor of each explanatory variable and the binary UK, VC or crisis interaction dummy variable.

4.3.1 Post-issue 12-month market performance

Table 4 shows the cross-sectional regression estimates of the relationship between the 12-month performance (sector-adjusted) and various explanatory corporate governance variables. Column 1 represents the first model with a final candidate set of six corporate governance variables and three control variables after adjusting for multicollinearity, having a significant explanatory power and adjusted R-squared of 0.321.

Firstly, we identify that firms with VC-backing perform significantly better than firms with PE-backing, consistent with the studies by Rindermann (2004), Bruton et al. (2010) and Brav and Gompers (2003), however, contrary to UK literature (Levis, 2010). In addition, it was assumed that lockup periods would have positive effects on post-issue performance as mentioned in US- and UK-based PE literature. However, the regression specifications show a highly significant negative effect. Generally, lockups are mostly required by banks to reduce risk and are used for creating confidence to investors.