Leverage as an instrument of control, mitigation of agency costs in privately held companies in the Netherlands

(1)

Amsterdam Business School

Leverage as an instrument of control, mitigation of agency costs in

privately held companies in the Netherlands

Name: Allard Ligteringen

Student number: 10467947

Date: May 26

th

₂₀₁₅

Paper: Master Thesis

1

st

_Supervisor:

_{prof. dr. L.R.T. van der Goot}

2

nd

_Supervisor:

_{Dr. Ir. S.P. van Triest}

Word count: 16,120

MSc Accountancy & Control, specialization Control

(2)

This document is written by Allard Ligteringen who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text have been used creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for its contents.

(3)

Abstract

The cost of the the separation of ownership and management, because of financing a firm by equity, is described by the Agency Theory. When a Private Equity firm acquires a majority of the shares of the subsidiary (the portfolio firm), the characteristics of the conflict between owner and management change, and this conflict can be better described and studied as a management control problem, rather than an Agency Cost problem. The management of the Private Equity firm controls a majority, or all, of the shares in the portfolio firm, and can thus appoint and dismiss managers and make, or influence, decisions on how the portfolio firm is operated. Nevertheless, important information regarding the operations, markets and finances of the portfolio firm, remains available only at the level of the management of the portfolio firm.

The aim of this study, is to find if the decisions that are made by the management of the PE-firm, on the degree to which the portfolio firm is leveraged, are influenced by the knowledge that it has about the portfolio firm.

A final sample of sixty Dutch portfolio firms was constructed. Data was collected on the debt-ratios of these firms and on three proxies for the information that is available to the management of the PE-firm. A statistical regression test yielded a low adjusted R2_{and an F-value just above the}

ten-percent confidence level. A significant, negative correlation was found between the number of portfolio firms that are held concurrently by a PE-firm, and the extent to which the portfolio firms are leveraged.

(4)

1 Introduction

In his seminal paper, Jensen (1989) argues that Private Equity (PE) is an effective new business model to mitigate the agency costs that result from the separation of ownership and management in publicly traded firms. The combination of high leverage, the involvement of an active investor and the alignment of the interests of management and owner, by granting management a significant equity stake, is expected to reduce agency costs and to result in the maximization of wealth.

Of course, a wider context exists outside of the Agency Theory. The extensive use of leverage increases the risk of financial distress and corporate failure. The control that a single, powerful investor has over all of the assets of the firm, has implications for all the stakeholders in the firm. The financial problems that the Dutch retailer Vroom & Dreesmann (owned by the PE-firm Sun Capital) is recently facing, illustrate that PE can also be studied from a Legitimacy Theory or Stakeholder Theory perspective.

The central theory in this study, however, is the Agency Theory. The aim of this study is to describe the relationship between corporate governance measures and the financial structures of companies (as part of those corporate governance structures) that are owned by PE-firms. The main corporate governance structures that are used by PE-firms, as described by Jensen (1989), exists of: (i) active monitoring of the management of the subsidiary (portfolio firm), (ii) alignment of the interests of the owner and management by having management invest own wealth in portfolio firm stock and (iii) an extensive use of leverage.

The relation between the way in which the portfolio firm is financed and corporate governance is here studied mainly within this structure. In theory, the choice between leverage and equity should be a neutral one, as both are an equation of risk and rewards. In practice, this choice is often approached in a different way (Axelson et al, 2009, p. 1549).

The primary research question is: are the choices that the general partners of a PE-firm make when they decide on how the portfolio firm is financed influenced by the state of other corporate governance measures, or are these choices primarily based on the price and the availability of debt, the marginal tax rate and the maximum debt that the portfolio firm can bear without an excessive risk of financial distress? The secondary research question is: what is the nature of the relationship between the use of leverage and other corporate governance variables, if this relationship exists?

This study is quantitative in nature and the sampled firms are all based in the Netherlands. In order to be able to study the described relationships, three proxies are used that are associated with corporate governance and agency conflicts between the owner (the PE-firm) and the management of the portfolio firm. These proxies are (i) Firm Size, measured by the natural logarithm of annual sales, (ii) Board Composition, measured by the percentage of outside directors in the board and (iii) Span of Control, measured by the number of Dutch portfolio firms that are owned concurrently by a PE-firm in the reporting year of the sampled portfolio firm.

Financial statements are collected from various sources (mainly from the ‘Kamer van Koophandel’ [the Dutch Chambers of Commerce (KvK)], which are used to calculate the debt-ratios of the sampled portfolio firms. Data on the proxies are collected from the same financial statements, if possible, or from the internet and surveys by telephone and e-mail. A statistical regression test is used to infer if a significant part of the variability in debt-ratios of these companies is explained by the scores of the associated companies on these proxies.

This study contributes to the existing literature in the following ways: First, the relationship between leverage and corporate governance has not been described in this way in the existing literature. Dey (2008) researches a relation between leverage and control within organizations. She uses leverage as one of several proxies for agency conflicts and finds that organizations with greater agency conflicts, have better corporate governance structures in place. The relation between debt and control is here thus indirect, and not investigated in a PE-context. Second, in many countries (especially the USA), financial data from privately owned firms is not available, or difficult to acquire (Bratton, p. 523; Cumming and Walz, 2009, p. 6; Metrick and Yasuda 2011, p. 622; Kaplan and Strömberg, 2009, p. 132). In the Netherlands, most firms have a legal obligation to submit basic financial data at the KvK. This makes it possible to study data that is obfuscated in many other countries. Third, no known study focuses solely on Dutch firms in a PE-context.

(6)

Studying PE in a Dutch context might also be interesting because of the particularly bad reputation that the PE business model has in Dutch discourse. Typically, PE-firms are referred to as a swarm of locusts, which sucks companies dry and then moves on. The disastrous buyout in 2004 of the high-profile Dutch media corporation PCM by the PE-firm Apex, and the excellent book ‘de Geldpers’ that was written about this episode by the Dutch journalist Joost Ramaer, are perhaps at the root of this disposition towards PE in the Netherlands. The recent publicity around the financial distress at retailer Vroom & Dreesmann, does probably not help to create a more positive attitude towards the PE sector. The remainder of this study is organized as follows: section two includes an overview of the relevant literature and the hypothesis development. In section three are described the research methodology, the proxy definitions and the data collection and sampling process. The quantitative description of the collected data and the statistical analysis is described in section four. Section five includes a summary and the findings and the conclusions and limitations of this research.

2 Existing literature and hypothesis development

2.1 Private Equity, typology, history, characteristics and definitions

2.1.1 Typology

PE is a term that is used in many contexts, and that applies to a wide variety of investors (Morrell and Clark, 2010; Wright et al. 2009). In the literature on the subject, Venture Capital is generally described as a form of PE. A distinction can be made between investors in early stage companies (Venture Capital) and later stage companies (Wright et al. 2009, p. 357). Morrell and Clark (2010, p. 251) offer a typology of PE that is based on the characteristics of an investment transaction. The transaction is placed in a matrix, where the y-axis measures the complexity from simple to complex. Complexity is here a measure of the number of actors and interrelations that are involved in the transaction. The x-axis is a measure of the control that the investor has over the assets of the investee. The extremes of the axis are indicated as Venture (investments in non-controlling shares of immature companies) and Buyout (investments in controlling shares of mature firms or divisions). Based on this classification, four kinds of Private Equity are identified: (i) Venture Capital, (ii) Corporate Venture Capital, (iii) mid-size buyout and (iv) large buyout.

Another typology of PE is provided by Hoskisson et al (2013). In this study, PE-firms are placed in a matrix that consists of the dimensions Financial Structure Emphasis and Portfolio Firm Scope. The Financial Structure Emphasis variable is measured by the leverage that is applied to the investee. This is a reflection of the investment horizon of the PE-firm, where more leverage is indicative of a short-term focus. The Portfolio Firms Scope dimension is measured by the level of industry diversification in the portfolio firms. A focus on investments in a specific industry (e.g. media) allows a PE-firm to increase its specific knowledge and thus to guide the investee more effectively. With a more diversified portfolio, industry-specific knowledge is more difficult to acquire, and the PE-firm is more likely to apply financial leverage as a means to increase firm value. The distinct types of PE-firms that are derived from this matrix are: (i) Short-Term Efficiency Niche Players. They have focused portfolios that are highly leveraged, (ii) Niche Players with Long-Term Equity Positions. These PE-firms finance their acquisitions with equity, but they also have portfolios that are focused on a limited number of industries, (iii) Diversified Players with focused groups of portfolio firms. These PE-firms hold diversified portfolios that are mostly equity financed, (iv) Short-Term Diversified Efficiency-Oriented Players. The diversified portfolio of these PE-firms are mostly financed with debt (Hoskisson et al, 2013, p. 25-26).

The PE-firm generally is compensated in three ways: (i) The general partner of the investment fund earns a management fee that is based on the committed capital of the fund. (ii) The general partner earns a percentage of the profits of the fund and (iii) some general partners charge deal and monitoring fees to the firms that they invest in (Kaplan and Strömberg, 2009, p. 123-124).

Between the moment of the introduction and the liquidation of the fund, typically no new participants or withdrawals are allowed (Achleitner et al., 2010; Axelson et al., 2009, p. 1550).

(7)

2.1.2 Historic context

Regarding the history of PE, two distinct waves of investment activity are described in the literature. The first wave emerged during the 1980’s. The transactions were characterized by acquisitions of majority control in mature companies and the use of high degrees of leverage. Also stock incentive instruments were applied to align the interests of managers and owners. The peak of the first wave was reached in 1988 with a buyout volume of USD 60 billion (Kaplan and Stein, 1993, p. 313). In the later stages of the first wave, the excessive use of junk-bonds and other unsecured loans increased the risk for the investors in these transactions.

The second wave originated around the year 2000 and it was cut short by the financial crisis of 2007 and 2008. An additional motivation behind the buyout transactions in this second wave was the increased costs that come with being listed at a stock market that were caused by new legislation like the Sarbanes-Oxley Act (SOX). PE-firms in this second wave also focused more on operational improvements (Hoskisson et al., 2013).

Several classifications of PE transactions exist. The most commonly used are: insider [Management Buyout (MBO)], our outsider [Management Buy-in (MBI)] driven transactions, or transactions that involve whole companies, or divisions of companies (carve-outs). In general, PE transactions are found to generate significant financial returns for investors (Cumming et al., 2007, p. 440; Nikoskelainen et al., 2007).

2.1.3 General characteristics

General characteristics of PE that are found in the literature are: the division between limited partners (the investors, e.g. pension funds) and general partners (the firm managers), large capital resources, the small size of PE-firms (measured by its number of employees), compared to the size of the firms that they own (Kaplan and Strömberg, 2009, p. 123). The duration of the existence of specific investment funds is around ten years. The illiquidity of the investment during the existence of the fund is a factor of significance, for which suitable compensation needs to be offered (Palepu, 1990, p. 248). Metrick and Yasuda (2011, p. 621) name four key characteristics of PE: (i) a PE fund is a financial intermediary between the investor and the portfolio company, (ii) a PE fund invests only in private companies, which means that after the transaction the shares in the company are not traded at any stock market, (iii) a PE fund takes and active role in monitoring and helping the companies in its portfolio and (iv) a PE fund’s primary goal is to maximize its return by selling the company.

2.2 Private Equity, agency costs and corporate governance

2.2.1 Agency costs

The Agency Theory quantifies and describes the costs that come with the separation of ownership and management. Agency costs caused by self-interested managerial behavior can include empire building, consumption of corporate resources as perquisites, refusal by management to take on optimal levels of risk and manipulating financial figures to optimize compensation (Dey, 2008, p.1145). This conflict of interests also exists between PE-firms and the management of their portfolio firms. In this relation, the PE-firm can be regarded as the ultimate block holder of shares (Bratton, 2008, p. 510).

From an Agency Theory perspective, leverage can be used to control cash flows and to discipline management to repay free cash flows to investors. Free cash flow is excess of that required to fund all projects that have positive NPV’s when discounted at the relevant cost of capital (Jensen, 1986, p. 323). Conflicts of interest between shareholders and managers over payout policies are especially severe when the organization generates substantial free cash flows (Jensen, 1986; Dey, 2008, p. 1162).

Leverage is one of the instruments that are commonly used in firms with a high risk of agency conflicts (Dey 2008, p. 1161; Jensen, 1986). Managers in general are inclined to add to the resources under their control, rather than repay cash flows to shareholders (D’Mello and Miranda, 2009). Large free cash flows help them to do so. Because primary buyout candidates are firms with substantial free cash flows (Jensen, 1986, p. 325), it follows that the agency costs faced by PE-firms are initially high. The use of leverage to mitigate agency problems has been described in various studies (Jensen, 1986; D’Mello and Miranda, 2009). While the issuing of debt can be regarded as an effective instrument to

(8)

mitigate agency costs, (Bratton, 2008, p. 520) finds that its use by PE-firms is mostly ignored and that buyouts are driven by economics of leverage rather than cost reduction. Cumming et al (2007, p. 441) find that existing literature indicates that PE-firms create shareholder wealth mainly by making use of under valuations of pre-buyout companies, by increasing the effect of tax shields and by incentive realignment, rather than by achieving operational efficiencies.

Conversely, Acharya et al. (2012) find that specifically large and mature PE-firms do create operational efficiencies within their portfolio firms, but in this study, the effect of the use of leverage on operational performance is mostly left out of the equation. According to the definitions that are used in this study, a mature PE-firm operates for at least five years. A large PE-firm has assets under management of at least USD 300 million (Acharya et al., 2012, p. 369-370).

The size of the acquired firm, as well as additional acquisitions by the PE-firm during the period that the firm was held (buy-and-build strategy), were found to be positively correlated to increases in firm value (Nikoskelainen et al., 2007). This is mainly caused by the robustness that larger companies have in a contracting economy. Also larger companies tend to operate in more diverse markets.

An agency conflict of a different nature exists between the PE-firm and its investors (Axelson et al, 2009), but this conflict will remain outside the scope of this study.

When the corporate structure of a PE-firm and its holdings is compared to that of a diversified firm with large business units, the following similarities can be observed: (i) a holding company, where decisions on financial planning and strategy are made, that is small in comparison to the size of its business units (Kaplan and Strömberg, 2009), (ii) a certain degree of autonomy that management has at the level of the business unit (or portfolio firm) level and (iii) an information asymmetry between the management of the business unit (or portfolio firm) and the management of the holding company (or PE-firm) regarding the business units operations, products and markets. Just as PE portfolio firms, diversified firms tend to be leveraged to greater extent than single line firms (Berger and Ofek, 1995, p. 41).

As far as these comparisons can be acknowledged, the specific agency costs that are inherent to diversified firms, have to be also considered in a PE-context. Martin et al (2013) study the effects of geographic diversification on the corporate valuation of Bank Holding Companies in an Agency Theory context. Geographic diversification is measured across different states in the USA. It is found that geographic diversification makes it more difficult for outside investors to exert effective corporate control and that as a result, agency problems increase and corporate valuation decreases (Martin et al, 2013, p. 1818). If this effect is observable in relation to geographic diversification within one country (the USA), it will very likely also exist in an international context.

In diversified firms, resources are often not allocated in a way that would benefit the firms as a whole to the best effect. Two major causes are identified: First, because the managers of the business units have knowledge that is only available to them, and not to the top management, they can influence the resource allocation in a way the benefits the business unit, at the expense of the firm. Second, because the interests of top management are different than those of the business unit managers, the latter have an incentive not to share all the information they have with the top management. Both information asymmetry and divergence of interests between the top management and the business unit management lead to increased agency costs (Harris et al. 1982, p. 605).

In general, diversification is found to reduce firm value, mainly due to poor allocation of resources at the business unit level. This effect is less strong when the business units operate within the same, industry (Berger and Ofek, 1995, p. 60).

This comparison of the structure of a PE-firm and its holdings with a diversified firm, illustrates that the nature of the agency conflict in this study differs from the typical agency conflict that is described in the literature. The position of the management of the PE-firm, which owns 100 percent of the portfolio firm’s shares, is very different from that of a large and diversified group of shareholders, as described by the Agency Theory. In this respect, the agency conflict that is described in this study, has many of the characteristics of a management control problem. From this point of view, it can be questioned to what extent leverage can be used by the management of the PE-firm, as an instrument of management control. Also, in this respect, it is relevant to consider if the debt is provided by the PE-firm, or by a third party. To the extent that the debt is provided by a third party, control over free cash flows of the portfolio firm by the management of the PE-firm is diminished. The available literature on this subject cannot be effectively applied, mainly because no valuation of the PE-firm is available.

(9)

2.2.2 Corporate governance

Short et al. (2005) name two broad dimensions of corporate governance. The first dimension concerns the accountability of management to shareholders. The second dimension relates to mechanisms that must assure a maximum increase of wealth by managerial actions and decisions. The costs and benefits of enforcing accountability should be balanced against the effects on wealth creation.

The prevailing opinion in the existing literature is that the effectiveness of a board is determined by its size, composition and independence. A high percentage of non-executive directors is associated positively with monitoring and accountability, but negatively with enterprise performance variables as innovation, venturing and renewal factors (Short et al., 2005, p. 338).

Corporate governance structures can be divided into two categories, internal and external. Internal governance structures can be split into monitoring related or incentive related (Weir et al., 2005, p. 912). The main internal corporate governance structures reviewed in literature that are applied by PE-firms, are: (i) the use of leverage, (ii) active monitoring of the investee and (iii) concentration of equity under management (Wright et al 2009; Hutton and Majadillas, 2014).

A relationship between firm valuation and the implementation of good corporate governance structures was found in multiple studies, although the cause and effect relationship is more difficult to confirm (Beiner et al., 2006, p. 255). More specific, a positive correlation between managerial ownership of company shares and firm valuation was found (Beiner et al., 2006, p. 266). Beiner et al. (2006, p. 277) also suppose substitution effects among corporate governance mechanisms, where a selection is made among different mechanisms that serve the same purpose. The non-executive director has an important role in the corporate governance framework as both a strategic advisor and a corporate watchdog. Opinions vary on the question if these two functions can be effectively performed by one person (Short et al., 1999, p. 339).

Under Dutch law, it is mandatory for some firms to appoint a supervisory board. This law applies to larger legal entities with a limited liability (“structuurvennootschappen”). This applies to entities for which the following is true over a period of three consecutive years: (i) the firm has at least16 million Euro in capital and retained earnings, (ii) a works council is appointed, and (iii) at least one hundred people are employed in the Netherlands. A minimum of three members are appointed in the supervisory board (Dutch Civil Code, Book 2, Article 268).

There are indications in existing literature that salaries are increased, and bonus plans expanded and made available to a larger group of employees in the post-buyout firm. This specifically applies when divisions are bought from larger entities, and become stand-alone companies (Baker and Wruck, 1989, p. 176).

2.2.3 Leverage

There is much evidence that indicates that corporations that are owned by PE-firms, are financed differently than family-owned of publicly traded companies. These companies are in many cases burdened with high debt-ratios, and with debt comes the obligation to make interest payments. Dutch PE-owned companies, like HEMA, disclose that the large interest payments that need to be made, impair profitability and make it difficult to pursue certain profitable strategies. Child-care organization Estro went bankrupt, in part because of the high interest payments that it needed to make. That raises the question: which purposes are served with these capital structures, and whose purposes are they? The main factors that determine the financial structure of a firm have been researched and described at length. Because in most tax systems interest expenses are deductible from corporate taxable income, the incentive to finance with debt, rather than with equity, is dependent on the marginal tax rate (Modigliani and Miller, 1963; Feld et al., 2013). This is generally referred to in corporate finance literature as the Tax Shield.

Specific for way that PE-firms finance their acquisitions, is the extensive use of leverage. Similar to investing in the stock market with borrowed money, financing a firm with only limited amounts of equity, promises a greater reward when the exit value of the firm exceeds the purchase value, compared to an investment that is financed with 100% equity (Bratton, 2008, p. 520 – 521). The extent to which PE-firms leverage their acquisitions is dependent on the price and availability of debt. When debt is available at low price in large quantities, debt-ratios in PE-owned companies rise. Conversely, debt-ratios in publicly traded companies are mostly dependent on firm

(10)

characteristics, where large and stable cash flows and high profits make it more profitable to finance with debt because of the reduced risk of financial distress and the effect of the Tax Shield. The effect of firm characteristics on the debt-ratios of PE-owned companies is almost completely absent (Axelson et al., 2013). Other studies find less of a distinction between determinants of corporate finance in publicly traded and PE-owned companies.

It has been inferred, that PE-firms specifically target companies for acquisition that are most suited for leverage increases. The criteria that are used are: a large free cash flow, which reduces the risk of financial distress (Achleitner et al. 2010, p. 807) and low growth expectations (Achleitner et al. 2010, p. 807; Jensen, 1986). The combination of large free cash flows and limited investment opportunities increases the risk that management will be inclined to make investments that have a negative Net Present Value, or spend on perks. Low growth opportunities are also among the investment criteria that PE-firms use, because they reduce the information asymmetry between owner and management (Jensen, 1986).

In highly leveraged organizations, the holders of equity may have an incentive to take actions for the benefit of themselves, at the expense of the debt holders, such as dividend payments, or making all or nothing investments. Typically, actions like these are prohibited by detailed covenants with the debt holders. These covenants can cover sales of assets, changes in corporate structure, issuance of additional debt, payment of cash dividends, et cetera (Baker and Wruck, 1989, p. 171).

In as much as the PE-firm and its portfolio firms resemble a diversified firm, a more extensive use of leverage can be expected when the number of portfolio firms increases. More diversification offers increased benefits from the use of leverage, because of the higher debt-capacity and the increased effectiveness of the tax shield (Berger and Ofek, 1995, p. 41). Firm size is also positively related to leverage. Larger firms generally have less volatile earnings, a reduced risk of financial distress and therefore a better access to capital (Fama and French, 2002, p. 29).

2.2.4 Monitoring

Many indicators of agency conflicts are associated with the way the management of a firm is structured. PE-firms use these indicators when they evaluate agency cost reductions that they can achieve, when they acquire a firm. From an Agency Theory perspective, boards of directors are put in place to monitor managers of behalf of the shareholders (Lynall et al., 2003, p. 417; Braun and Latham, 2009). Variables with implication for corporate governance are: board independence (the degree to which board members are dependent on the current CEO or organization), CEO-duality (the CEO also is chairman of the board), lack of expertise, et cetera.

Besides Agency Theory, many other theories are applied in the studies that relate to boards of directors. Lynall et al. (2003) also apply, among others, Resource Dependence theory in their study of changes in corporate boards of directors. This theory views the role of the board of directors as a provider of the needs of the company. Four specific needs of companies are identified: (i) advice and counsel; (ii) legitimacy; (iii) channels for communicating information between the company and external organizations and (iv) assistance in obtaining resources or commitments from important elements outside the company (Lynall et al., 2003, p. 418). The main findings of the study are that corporate board structures are path-dependent, which means that they are mostly derive from structures that originate in the past, and for that reason, are not always suited in for the current environment that the firm operates in.

The finding that a larger board is related to an increase in firm valuation (Beiner et al., 2006, p. 277) can also be explained from a resource dependence theory perspective.

In the study of Braun and Latham (2009), also both Agency Theory and resource dependence theory are applied. The nature of board restructurings is studied, following a buyout. Three variables are measured, the first of which is board size. Larger boards are less effective in supervising the CEO (Jensen, 1989), but from a resource dependence perspective, a large board is more effective in providing resources for the firm. The second variable is CEO-duality. From an Agency Theory perspective, the combined role of CEO and chairman implies that the effectiveness of the supervision of the CEO is reduced. This can result in increased agency costs. From a resource dependence theory perspective, the combined role of CEO and Chairman can be seen as effective, because it communicates strong leadership to the external environment. The third variable is board composition and expertise. From an Agency Theory perspective, outside directors are seen as an effective in monitoring and reducing

(11)

managerial self-interested behavior (Arondo, 2005; Rosenstein and Wyatt, 1990; Morck et al., 1988, p. 307). From a resource dependence theory perspective, insiders offer firm-specific knowledge and outsiders provide access to critical resources. Changes in board composition are measured, following a PE buyout. One of the main findings of the study, is that board characteristics are related to the probability of a buyout. Companies that are subject to a buyout are found to have smaller boards, more board independence and fewer insiders. This can be explained from a resource dependence theory perspective, assuming that the board size is increased and the right mix of inside and outside expertise is created after the buyout. Companies that are subject to a bout are also found to have a higher degree of CEO-duality, which can be explained from and Agency Theory perspective, assuming that the roles of CEO and Chairman are separated after the buyout.

Increased levels of CEO-duality in firms going private, are confirmed in a study in the UK, but no evidence is found regarding difference in non-executive (outsider) director representation, in firms that are taken private, compared to firms that remain quoted (Weir et al., 2005, p. 911). These different findings can possibly be explained from the fact that the study of Weir et al (2005) does not focus exclusively on firms that are acquired by PE-firms. Portfolio firms have more formal meetings per year than comparable quoted firms and that CEO’s of these firms are more frequently replaced (Kaplan and Strömberg, 2009, p. 131).

2.2.5 Concentration of equity under management

The costs that come with the separation of ownership and management are at the basis of the Agency Theory. Aligning the interests of owners and managers via managerial share ownership seems to be the perfect solution to reduce agency costs. Weir et al. (2005, p. 938) find CEO shareholdings to be of more value in achieving post-buyout gains than leverage. Nikoskelainen et al. (2007, p. 529) find evidence that managerial equity holdings are related to increases in enterprise value and equity value.

In publicly traded companies, management can be directly monitored with company stock in two ways: (i) if the firm operates poorly, the company can be the target of a takeover and the manager is at risk of losing his position. (ii) The performance of the stock can be aligned with the remunerations of the manager. These specific monitoring benefits are lost when the stock loses its liquidity due to the buyout. The value of the portfolio firm is only determined when the PE-firm divests its interest in it; no reliable value is available in the interim. The risk of a buyout also does not exist (Holmström and Tirole, 1993, p. 679).

Nevertheless, PE-firms make extensive use of company stock to align the interests of the management of the portfolio firm with their own. One of the main investment criteria that PE-firms use is low managerial ownership of company shares, as this separation of ownership and management is a cause for agency conflicts. PE-firms seek to create value by aligning the interests of owner and managers, and thus by mitigating agency costs (Achleitner et al. 2010).

Firms that have a dominant shareholder are less likely to be acquired by a PE-firm, because the dominant shareholder is motivated to invest in high quality corporate governance structures and keep agency costs under control. Also, the dominant shareholder is often able to extract private benefits from the company, and is only willing to sell the shares at a premium in order to be compensated for the loss of these benefits. Specifically in family owned companies, the dominant shareholder often extracts private benefits at the expense of the remaining shareholders. The value that can be created by a PE buyout is thus expected to be lower and the probability of a transaction is reduced. The likelihood of a buyout is increased when the dominant shareholder is found to lack the necessary expertise to keep agency costs under control (Achleitner et al., 2013, p. 75).

The opinions in existing literature vary on optimum percentage of company shares that management should own. Halpern et al., (2009) find that companies are more likely to be targeted for a buyout when managerial ownership of company stock are either very limited or very significant. Low managerial ownership leaves much of the agency conflict unresolved (Achleitner et al., 2013, p. 77). High managerial ownership can result in management entrenchment. The substantial voting power that comes with the ownership of a large fraction of firm shares assures managers of continued employment at attractive terms. However, management entrenchment can also result from tenure, involvement in the founding of the company or personality. Whatever the cause, management entrenchment leads to a reduced effectiveness of the checks and controls on the actions of managers (Morck et al., 1988, p. 294).

(12)

hypothesis perspective when managerial ownership rises from 0 to 5 percent, a negative effect when managerial ownership is increased up to 25 percent, and a positive effect beyond 25 percent (Morck et al., 1988, p. 311).

Apart from entrenchment, high personal investment of management in firm shares can also result in loss of shareholder value, because it makes managers more risk-averse (Halpern et al., 1999, p. 283). The typical solution for this problem that is used by PE-firms, is to have management invest a substantial amount of their own wealth in the company. In this way, management faces both a downside and an upside effect. Furthermore, because shares in companies that are owned by PE-firms cannot be freely traded, and are thus illiquid, the incentive of management to manipulate short-term profits is greatly reduced. (Kaplan and Strömberg, 2009).

A case study related to the buyout in 1986 of a US manufacturing company also documents that the specific demand from the new owners, that managers buy equity with their own, not the firm’s money (Baker and Wruck, 1989, p. 173). The recorded effects of this managerial ownership are (i) a focus of management on the need to reduce the risk of a default on debt. Such a default, even if it would not result in the liquidation of the company, would dilute the ownership of management by a debt-for-equity conversion and (ii), even if no immediate risk of a default exists, management is motivated to increase firm value and avoid the risk of a default in the long term (Baker and Wruck, 1989, p. 175).

The effect of ownership by management of company shares is greatly reduced when the investment by management in the post-buyout firm is less than what was divested at the moment the company was de-listed. The reasons for this, are that the reduced investment in the post-buyout firm can be regarded by the manager as a cheap call option, which motivates excessive risk taking. Also, when great financial wealth is achieved at the moment that the firm was de-listed, the financial future of the manager becomes more detached from the success of the company (Hutton and Majadillas, 2014).

Evidence suggests that buyouts generate in general large increases in wealth for the pre-buyout shareholders (Palepu, 1990, p. 248). Companies that are taken private are likely to have higher CEO shareholdings (Weir et al., 2005). This gives more weight to the effects on agency costs of management divestment of shares. The adverse effect of management divestment of shares is also found by Kaplan and Stein (1993). Cases where management is found to own less shares after the buyout that they did before, are associated with a significantly higher probability of default (Kaplan and Stein, 2005, p. 354).

The amount that management invests is typically small compared to the total investment that is made in the firm, but significant in relation to the personal wealth of the individual manager (Kaplan and Strömberg, 2009, p. 125). The median CEO is found to own 3 percent of the company shares, the median management team owns 15 percent (Kaplan and Strömberg, 2009, p 131).

2.3 Hypotheses development

The use of leverage comes with an increased risk of financial distress, or even corporate failure (Kaplan and Strömberg, 2009). This makes it relevant to study the variables that influence the extent to which firms are leveraged.

Leverage is regarded in literature as a proxy for agency conflicts, because owner-managers have an incentive to invest in high-risk projects, in order to transfer wealth from creditors to shareholders (Dey, 2008, p. 1161). Specifically in a PE-context, leverage is also regarded as a corporate governance instrument, which is used to discipline management to operate the company in a way that generates cash (Baker and Wruck, 1989, p. 167, Jensen, 1986). Leverage also serves as an instrument to maximize pay-outs to investors, by making use of the tax shield.

The aim of this study is to research the relationship between agency conflicts, corporate governance structures and debt-ratios of portfolio firms. This relationship can work in two directions. If debt and corporate governance are both used to mitigate agency costs, it can be supposed that they are used as substitutes for one another. A PE-firm that has a weak position to exert control over management of its portfolio firm, might be inclined to raise the debt ratio, in order to minimize the opportunities for management to invest in unprofitable projects, and thus force management to disgorge cash flows to the owners (Jensen, 1986, p. 324).

However, in this scenario, the combination of weak control and high leverage will increase costs in the form of a higher risk of insolvency (Kaplan and Strömberg, 2009, p. 129). Therefore, a second relationship is considered, where it is supposed that control measures and leverage are used as

(13)

complements. To mitigate the risk of insolvency, the PE-firm might use leverage only in relation to the knowledge it has about the investee. More direct control will lead to more knowledge, and this could lead to a debt ratio that is closer to the maximum level of debt that the firm can bear, without an increased risk of insolvency. This second point of view was put forward by a Senior Vice President of Marlin Equity during a short interview.

The hypothesis that is tested in this paper is called the Knowledge Hypothesis. This hypothesis supposes that the general partners of a PE-firm will seek to leverage the portfolio firm close to the maximum percentage that it can bear, without taking on an excessive risk of financial distress. The main motivations for this strategy lie in the increased effectiveness of the corporate tax shield (Modigliani and Miller, 1963; Feld et al., 2013), the increased control over the free cash flows of the portfolio firm (Jensen, 1986; D’Mello and Miranda, 2009) and the maximization of profits at the moment that the portfolio firm is divested for a higher price than the acquisition price (Bratton, 2008). In order to be able to properly quantify the risk of financial distress, specific knowledge about the portfolio firm’s operations and finances are required. It is thus supposed that a positive correlation exists between variables that indicate an increased knowledge at the level of the PE-firm, about a portfolio firm, and the debt-ratio of that portfolio firm. This hypothesis will be falsified by negative correlations between measurements of variables that indicate increased knowledge about portfolio firms and the related debt-ratios of those portfolio firms, or by low and multi-directional correlations.

The proxies that are developed to study the described relationships, and the methods that are used to collect and validate the data are presented in section 3.

3 Research methodology

3.1 Introduction

The aim of the research process is to enable a quantitative study into the relationship between the perceived risk of agency conflicts and information asymmetries, from the point of view of the management of the PE-firm, and the debt-ratios of the related portfolio firms. The initial phase of the research process can be divided into two main parts. The first part consists of the formulation of the relevant variables that can be used for the quantitative analysis. The second part consists of the data collection process that leads to an a-selective sample of sufficient size.

3.2 Research variables

3.2.1 Proxies for the probability of agency conflicts

In order to be able to determine the probability of agency conflicts and information asymmetries between the management of the portfolio firm and its owners, based on the described literature, three proxies are formulated. These are:

(i) Board Composition, the percentage of non-executive directors, out of the total number of directors. Fernandez and Arondo (2005) found that the proportion of outside members of the board can be used as a proxy for control over managerial actions. Rosenstein and Wyatt (1990) found positive stock price effects when a new outsider director was appointed. In general, non-executive directors are associated with monitoring and accountability (Short et al., 2005, p. 338). Morck et al. (1988, p. 307) remark that the role of outside directors is particularly related to the fiduciary duties of management towards shareholders. In relation to the Knowledge Hypothesis, a higher score on the proxy Board Composition, indicating a larger percentage of outsiders in the board of directors, is expected to be associated with reduced firm-specific knowledge and, consequently, with a low debt-ratio. The correlation between debt-ratio and the proxy-score is expected to be negative.

(ii) Firm Size, measured as the natural logarithm of annual sales in Euro’s. One of the proxies for agency conflicts that is used by Dey (2008, p. 1160), is firm size. Larger firms are more difficult to govern, and this increases the risk of agency conflicts. The annual sales are derived from published financial statements, or from financial statements that are purchased

(14)

from the KvK. A high and positive correlation can be expected for the scores on the proxies Board Composition and Firm Size. Boone et al. (2007, p. 90) found firm size to be a strong determinant of the percentage of independent board members. They concluded that a one-standard-deviation increase in firm size predicts an increase in the percentage of independent board members of 1.8. In relation to the Knowledge Hypothesis, a higher score on the proxy Firm Size, will be associated with reduced knowledge about the firm’s operations and finances, because more information is required on large firms than on small firms. The debt-ratio of the portfolio firm is expected to be low. The correlation between debt-ratio and the proxy-score is expected to be negative.

(iii) Span of Control, a count of the total number of portfolio firms that are owned concurrently in the Netherlands by the PE-firm in the sampled financial reporting year of the portfolio firm. The structure of a PE-firm that owns several portfolio-firms has many similarities with a diversified firm that has several business units. Diversified firms are faced with agency problems that are caused by geographic diversification (Goetz et al., 2013), asymmetric information distributions between top management and business unit management (Harris et al., 1982) and poor resource allocations (Berger and Ofek, 1993). In order to be able to allocate the number of concurrently owned portfolio firms, a timeline is created per PE-firm, which includes the acquisitions and divestments of companies in the Netherlands. For this, information is used from the websites of the PE-firms, and from the NVP-file. In relation to the Knowledge Hypothesis, a higher score on the proxy Span of Control, will be associated with geographic diversification effects (Martin et al., 2013) and information asymmetries (Haris et al., 1982) between the general partners of the PE-firm and the portfolio firm managers. Therefore, knowledge about each individual portfolio firm is expected to be impaired, and the debt-ratio of the portfolio firm is expected to be low. The correlation between debt-ratio and the proxy-score is expected to be negative.

Besides data relating to the three proxies, data has been collected on the debt-ratios of the portfolio firms, the percentage of the shares in the portfolio firm that is owned by the firm, the type of PE-firm that is involved, the year of investment and the year of divestment. Only datasets that contain complete and reliable information on all of these variables are included in the sample. The following is a detailed description of all the variables that are used for the research:

(i) The proxy Board Composition. The units of measurement are (1) the percentage of non-executive directors in a one-tier board of directors, or, (2) the number of members in the supervisory board as a percentage of the combined total of the number of members in the board of directors and the supervisory board. Because the two-tier board model is more commonly used in the Netherlands, in most cases, the second equation applies. As members of the board of directors are counted all Chief Executive Officers (CEO, CFO, CIO, COO, et cetera) and all Directors and Vice Presidents. In the cases where individuals, or a group of individuals are referred to, on the portfolio firm website, in the annual financial report, or in other reliable sources, such as het ‘Financieele Dagblad’, as (member of) the board of directors, all these individuals are counted as a member of the board of directors, regardless of their title or function. As non-executive directors, or as members of the supervisory board, are counted all individuals that are referred to as such, on the portfolio firm website, in official publications, or in other reliable sources. In the cases where the portfolio firm (also) has an advisory board, the members of the advisory board are not counted as non-executive directors.

(ii) The proxy Firm Size. The unit of measurement is the natural logarithm of the portfolio firm’s annual sales in Euro’s. As a reliable measurement of annual sales are regarded amounts that are referred to as such in audited financial statements that are disclosed by the portfolio firm, either on the website, or via the KvK, In the cases where the amounts are published in a currency other than Euro, the amounts are converted to Euro against the exchange rate of the last date of the applicable reporting year. The historic exchange rates are derived from the OANDA website: http://www.oanda.com/currency/converter/.

(15)

(iii) The proxy Span of Control. The unit of measurement is the total number of portfolio firms that are owned or controlled concurrently in the Netherlands by the PE-firm in the sampled financial reporting year of the portfolio firm. In order to be counted, it must at the minimum be very likely that the PE-firm has actual control over the portfolio firm. A minority share without an indication of control over the portfolio firm’s management will not be counted. The portfolio firm must be a legal entity under the Dutch law, most likely a ‘Besloten Vennootschap’ (BV) or a ‘Naamloze Vennootschap’ (NV). It must also be probable that the seat of business of the portfolio firm is in the Netherlands. Business units in the Netherlands of firms that are managed from abroad are not counted.

(iv) Portfolio firm debt-ratios. The unit of measurement is the percentage of total liabilities out of the balance sheet total. In annual financial reports, the consolidated balance sheet is typically divided in the sub-sections Total Assets and Total Liabilities. In these cases the debt-ratio is calculated as Total Liabilities / Balance Sheet Total. The financial statements that are derived from the KvK have a structure that specifies the following types of liabilities: Accruals, Subordinated Debt, Long Term Debt, Short Term Debt Accounts Payable and Shares Owned by Third Parties. All the types of debt, with the exception of Shares Owned by Third Parties are included, in order to calculate the debt-ratio. Shares Owned by Third Parties is not counted as debt, because the amount is not subject to re-payments or interest re-payments.

3.3 Data and sample selection

3.3.1 General assumptions and approach

There are currently around 1,400 Dutch firms that are PE-owned [source: Nederlandse Vereniging voor Participatiemaatschappijen (NVP)]. The goal of the data collection and sampling process is to construct an a-selective sample of sufficient size that contains records of complete and reliable data on all of the variables that are described in section 3.2. It is of particular importance that all of the collected data are related to the same year.

Another consideration is that, since the dependent variable that is studied is the debt-ratio of portfolio firms, and since the use of debt is influenced by interest rates (Axelson et al., 2013), the financial data that is collected should be derived from as short a range of years as can be accomplished, in order not to introduce new independent variables into the analysis.

In order to be able to generalize about debt policies of PE-firms, the sample also should include a substantial fraction of the PE-firms that are involved. Following the findings of Acharya et al. (2012), primarily acquisitions by PE-firms that are large and mature are selected for sampling. Specifically large and mature PE-firms focus on creating value by reducing agency costs, (Acharya et al., 2012). This suggests, that if the use of debt is regarded as an effective instrument to mitigate agency costs, PE-firms can be to make use of it. Including primarily acquisitions of these PE-firms in the sample, will make it more probable that a relation between debt-ratio’s and the probability of agency conflicts can be established.

The sample is not limited to any specific size; as many records as possible are included in it. The size of the sample is only limited by the availability of information, and the limited time that is available to collect the data. As a starting point of the data collection process, the table from the NVP-website was used. This can be found at: http://www.nvp.nl/pagina/overzicht_portefeuillebedrijven/.

The population has been defined as: all Dutch companies for which a non-bank-related PE-firm, or a consortium of PE-firms, has a majority, or a controlling minority, of the shares in the years since 2007.

A distinction is made between PE-firms and Venture Capital firms, because Venture Capital invests mostly in early stage companies that generate limited cash flows. Therefore, Venture Capital firms make little use of leverage (Wright et al, 2009, p. 357). Venture Capital firms do not typically obtain majority control of the company that they invest in (Kaplan and Strömberg, 2009, p. 121).

The focus of the data collection process, was to include mostly acquisitions from large and mature PE-firms, based on the findings of Acharia et al. (2012). Mature, later stage portfolio firms were selected for sampling (Wright et al. 2009).

(16)

Per the typology of Morrell and Clark (2010), transactions are included where the PE-firm has control over the assets of the portfolio firm (buyout) and the PE-firms involved would be classified as either mid-size buyout and or large buyout. Per the definitions of Hoskisson et al. (2013), the sampled PE-firms are most likely classified as either Short-Term Efficiency Niche Players or Short-Term Diversified Efficiency-Oriented Players, based on the typical investment term of tem years and the increased levels of leverage that is associated with a short investment horizon.

3.3.2 Selection of portfolio firms

The focus of the data collection process was on the portfolio firms that were owned in the research period by PE-firms with the right characteristics. From the NVP-table, those portfolio firms were selected that were owned within the research period. Some of the large and mature PE-firms could be identified at an early stage, based on references in ‘het Financieele Dagblad’ or other sources. The size and quantity of the acquisitions also proved to be a good measure of the classification of the PE-firm. These early indications of the size and maturity were then confirmed on the website of the PE-firm. In almost every case, a detailed history of the firm was provided there, with an overview of current and past transactions.

The large majority of investment firms, and their related portfolio firms, which were included in the NVP-table, were excluded from the sample because they were related to banks or because they were classified as Venture Capital firms, or in the absence of a clear indication that they were not Venture Capital firms.

The investment firms that are related to banks could in most cases be identified by the name of the fund or firm, e.g. ‘ABN AMRO Participaties’, ‘Friesland Bank Investments’ or ‘Van Lanschot Participaties’. The Venture Capital firms were primarily eliminated by reviewing all firms or funds with the word ‘venture’ in the name.

Regarding the control that the general partners of the PE-firm have over the management of the portfolio firm, initially the indications in the NVP-table were followed, in those cases where it was indicated that a minority share in the portfolio firm was owned by the PE-firm. Also, a confirmation of the actual acquisition of the portfolio firm was sought from any source other than the NVP-table. These confirmations could be found in annual financial reports or on the websites of the PE-firms. There were some instances, where acquisitions were included in the NVP-table, which could not be confirmed by any secondary source.

By this process of elimination, a selection was made of portfolio firms that were owned and controlled by PE-firms of the researched type, in the research period.

3.3.3 Collection of financial data

Both of the variables debt-ratio and annual sales were derived from disclosed financial statements. Whenever possible, published corporate financial statements were used. In most cases however, purchases of financial statements from portfolio firms were made from the KvK. The availability of usable financial statements was limited by that fact that in some cases no financial statements were available for the portfolio firm, or that no profit & loss statement was included in the financial statement. The most likely reason for this, is that the financials of the portfolio firm were consolidated with those of another entity outside the Netherlands.

Several times, it was noticeable that financial statements were published by the portfolio firm up to the year of the acquisition and no longer from that year on. In some of the instances where the profit & loss statement was missing in the KvK statement, the annual sales amount could be retrieved from another source, in most cases the website of the PE-firm.

In order to make inferences about debt-ratios, it was necessary to include in the sample those entities that actually contained the debt, as well as the assets. In many cases, financial statements for several legal entities were available with names similar to the name given in the NVP-table, or on the PE-firm website. Which of the financial statements should be included in the sample, was only obvious after the purchase of the financial statement had been made. A good rule to go by, proved to be to select a Dutch NV over a BV, and to select a Holding company, when available. A first glance at the statement would tell if the correct statement had been acquired. In several instances, more than one financial statement had to be purchased from the KvK.

(17)

The financial statements were collected from the following years: 2008 (2), 2009, (3), 2010 (3), 2011 (4), 2012 (16), 2013 (32).

In the cases where financial statements were not available at the KvK of on the portfolio firm website, no inquiries were made by phone or by e-mail. The probability that the company would provide the information in this way was regarded as very low. Ultimately, more financial statements were collected than could be used for the research. Around twenty statements were collected or purchased, that could not be completed with data on other variables, and that could, for that reason, not be included in the sample.

3.3.4 Collection of data regarding Board Composition and Span of Control

Collecting data relating to the proxy Board Composition, proved to be more difficult than anticipated. The assumption beforehand was that this information would not be regarded as confidential, and that members of boards of directors and supervisory boards can expected to be proud to be publicly associated with these accomplished positions. The few annual financial statements that could be collected contained all the required data regarding the board of directors and the supervisory board, without exception. In some instances, information on Board Composition was provided on the company website, including title, function, nationality and date of appointment.

For the company supervisory board, this information was given much less frequently. This could of course indicate that no supervisory board existed, but this needed to be confirmed. In some cases, references were made to the supervisory board, without any additional specific information, which at least confirmed the existence of a supervisory board. In rare cases, company brochures or reports (e.g. CRI reports), provided the required information. Searches on Linked-in also proved effective in some cases. Later in the data collection process, inquiries to the availability of Board Composition information were made first, before financial statements were purchased from the KvK, so as to avoid that resources were wasted to collect data that could not be used of the research. Where information regarding the board of directors or the supervisory board could not be obtained via the internet, or other sources, inquiries were made by telephone and by e-mail. The survey process was limited in scope, only about twenty companies were approached, and it yielded only three useful responses. The low rate of response was probably in part due to the fact that the inquiries were made in December and January, when due to year-end pressures, the willingness to cooperate is generally low.

The survey process was structured as follows: initial telephone calls were made to those individuals that probably would be able to provide the information over the telephone. In most cases these were secretaries to the board of directors. A quick introduction was made and the purpose of the inquiry was given. When no specific name was available, a request was made to be connected to the Public Relations department. In two instances, the information was provided directly over the telephone. In the other cases, the reception desk either refused to connect the call, or the call was connected but the information was not provided, or a request was received to send the inquiry by mail. From the e-mails that were sent, one refusal and one detailed reply were received. No further inquiries were made.

Data regarding the Span of Control proxy could be collected in almost every case from the websites of the PE-firms. Detailed portfolios were provided there, including the year of investment, the year of divestment and the percentage of the company owned. In many cases, selections per region or per country could be made. In some cases, the acquisition of a Dutch company was reported in the NVP table, with no mention of it on the PE-firm website. Conversely, some acquisitions were disclosed on the PE-firm website, which were not included in the NVP-table. In those instances, the data from the PE-firm websites were used in the sample.

3.3.5 Collection of various other data

In the sample, it is recorded in which year the portfolio firm was acquired by the PE-firm. The unit of measurement is any year, within the research period, in which the PE-firm owns and controls the portfolio firm, with the exclusion of the years in which the acquisition and the divestment took place. The research period is specified based on the following criteria: (1) the portfolio firm is acquired in or after 2007, or (2) the portfolio firm is divested after 2007 and (3) the portfolio firm was divested within twelve years of the year of the acquisition or, when the portfolio firm has not been divested, the year of

(18)

the acquisition is within twelve years of the current year. The years of the acquisition and the divestment are from the sample excluded because it can be assumed that in these years, material changes in the capital structure of the portfolio firm are implemented. Including financial ratios from these years in the sample introduces the risk that the financial structure that the PE-firm intends to implement, and which is the subject of this study, is not captured. An exception is made for the financial ratios that are derived from the acquisition year, provided that a material change in the financial structure of the portfolio firm is recorded, compared to the years prior to the acquisition. In that case, it would be reasonable to assume that the new financial structure is implemented by the end of the reporting year in which the acquisition took place. Example: Hitec Power Protection B.V. was acquired by Egeria in 2012. There are no financial statements available for 2013 or 2014. The debt-ratio at 31 December 2012 is calculated at 76.5%. In 2011 and 2010 the debt ratio was 64.1% and 63.7% respectively. Based on the difference in the debt-ratios of 2012, compared to those in the preceding years, it seems reasonable to assume that a new financial structure was put in place in 2012 by the PE-firm.

Before the portfolio firm was included in the sample, the control that the general partners of the PE-firm have over the actions of the management of the portfolio firm was confirmed. The unit of measurement is the percentage of shares owned by the PE-firm, or a consortium of PE-firms. Besides this objective criterion, a few more subjective criteria are used such as representation of the PE-firm in the supervisory board or the board of directors of the portfolio firm and the distribution of the shares that are not owned among the remaining shareholders. All considered, it must at the minimum be very likely that general partners of the PE-firm have control over the actions of the management of the portfolio firm to the extent that they are able to decide how the portfolio firm is financed and operated. The aim of the data collection process was focussed on collecting data on portfolio firms which are, or were, owned by large and mature PE-firms. The units of measurement are the number of years that the PE-firm has operated and, the assets that are managed by the PE-firm. According to the definitions that are used in the study of Acharya et al (2012), a large an mature PE-firm operates for at least five years and has assets upwards of USD 300 million. These data were mostly collected from the websites of the PE-firms.

3.3.6 General remarks regarding the sample

The sample that was constructed in this way is useful, because it is of sufficient size in relation to the population, specifically relating to that fraction of the population of portfolio firms that is owned by large and mature PE-firms, and it is a-selective, because each portfolio firm for which a complete data set can be collected, had an equal chance of being selected.

On the other hand, it can be argued that the sample is not a-selective, because the portfolio firms for which it proved to be impossible to construct a complete dataset, might share features that are relevant for the research. In this way, a selection bias may have been introduced into the sample. Even though Boards of director structures have a tendency to remain unchanged over time due to path-dependency (Lynall et al., 2003), combining the most recent financial statements with data collected on boards seems prudent. An overview of the sampling process is provided in table 1.

(19)

Table 1: Sample selection

Base sample Criteria Excluded from

sampling

Available for sampling

(1) All Dutch firms that are, or have been owned by a PE-firm during the research period source: NVP-table

Acquired in or after 2007

Divested after 2007

Held for a maximum of 12 years after the year of acquisition

Dutch firm

1,805

(2) Excluded all acquisitions by bank-related

Investment firms

Website references

Name of the bank included in PE-firm name

182 1,623

(3) Excluded all acquisitions from firms other than

PE-firms (Venture Capital)

179 1,444

(4) Excluded all acquisitions for which the PE-firm, Or the consortium of PE-firms, does not have a Controlling share

212 1,232

(5) Excluded firms that are not owned by PE-firms that are classified as large and mature

989 243

(6) Excluded firms for which it is not possible to collect all the data required that is related to each of the three proxies or the debt-ratio

148 95

(7) Excluded all firms for which the acquisition was not confirmed from a secondary source

Reference on the PE-firm website

Reference in annual report

43 52

(8) Excluded one outlier Debt-ratio > 3x Standard deviation from the mean

1 51

(9) Included portfolio firms from PE-firms that are not classified as large and mature

(9) 60

4 Statistical analysis

4.1 Sample data characteristics

In order to be able to study the research question, quantitative data has been collected on portfolio firms and on the related PE-firms. Sixty complete data sets were included in the sample. One record was not included in the sample, because the debt-ratio was more than three standard deviations from the mean, and it was for that reason classified as an outlier. It is remarkable that no less than five of the sampled

Leverage as an instrument of control, mitigation of agency costs in privately held companies in the Netherlands