Corporate governance and investment efficiency in publicly traded firms

(1)

Corporate Governance and Investment Efficiency in

Publicly Traded Firms

Name: Daqi Shen

Student number: 11617810 Thesis supervisor: Peter Kroos Date: June 24, 2018

Word count: 13,352

MSc Accountancy & Control, specialization Control

(2)

Statement of Originality

This document is written by student Daqi Shen 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 and its references have been used in creating it.

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

(3)

Abstract

Most previous literature studies determinants of corporate investment efficiency in terms of external factors. This thesis aims at examining whether factors in internal firm level as manifested in corporate governance quality, are associated with investment efficiency, by empirical analysis of publicly held firms incorporated in the U.S.

In this thesis, corporate governance quality is categorized into board governance, ownership structure, and management compensation, with the purpose of a systematic study about their functions for investment efficiency. Empirical results have found that the fraction of independent directors and the scale of management compensation is positively associated with investment efficiency, while negative function is reflected between board size and investment efficiency. In addition, the fraction of shares held by the largest shareholders turns out to have a U-shaped relation with investment efficiency. These findings suggest a significant correlation between corporate governance mechanisms and firm investment decisions, which may shed some light on further research regarding the effects of other firm-level elements on investment efficiency.

(4)

Ownership structure has been studied for decades since it is supposed to influence management control and firm performance. Traditionally, separation of ownership and control is advocated among large firms owing to the benefits brought by specialization of management and risk bearing effect on agents (Fama and Jensen, 1983). Moreover, as referred to the context where firms are publicly held, the risk diversification of investors (mainly outside shareholders), turns out to be an advantage as well. However, side effect occurs simultaneously, as classified by agency problem, where in modern publicly held firms, well-informed inside managers are self-interested and may in their decision making depart from what is required to maximize shareholder returns at the expense of less well-informed corporate outsiders e.g. investors (Berle and Means, 1932; Jensen and Meckling, 1976). To address this agency problem, financial reporting is regarded as an effective remedy (partially) with its function of mitigating information asymmetries as well as establishing a more transparent information environment through periodic disclosure from insiders to outsiders (contracting parties) (Armstrong et al. 2010). In addition to this mechanism, corporate governance mechanisms from firm inside function to align manager performance to the best interest of shareholders. Shleifer and Vishny (1997) studied and compared prior papers on corporate governance, suggesting that investor rights under the protection of laws is one of the essential elements adopted by corporate governance to alleviate agency problems, and compared to this, concentrated ownership seems to be a more universally accepted approach to guarantee investor rights. Recent research paper suggests compensation contracts signing can incentivize managers to take actions that maximize shareholders’ interest (Murphy, 1999). Most prior work concerning the effects of corporate governance quality focuses on e.g., the level of CEO compensation. For instance, Core et al. (1999) studied how board and ownership structure impacts the extent to which managers have access to extracting rent and compensation in excess of their deserved wage rate implied by economic determinants. Not so much research up to now has examined the actual decision-making activities of CEOs and specifically their investment decisions. Nonetheless, in modern corporations, one of the most relevant tasks of CEOs are deciding about investments through which they can generate considerable value

(7)

(Bertrand and Schoar, 2003). The benchmark here is investment efficiency, where any deviations generated represent either over- or under-investment. Based on this finding, my study plans to examine whether corporate governance mechanisms influence managerial performance in the context of investment efficiency.

1.2 Research question

This thesis sets out to examine to what extent CEOs and executive managers invest in accordance with efficient investment and more specifically, whether corporate governance quality is associated with efficient investment. The following research question will be investigated:

“Are corporate governance mechanisms associated with investment efficiency?”

The assumption behind this research question is that deviations in corporate governance mechanisms represent deviations from the optimal level of corporate governance which are supposed to subsequently impact investment efficiency. However, in some cases corporate governance is considered to be determined by agency problems in the sense that low quality corporate governance is the rational response to the absence of agency problems. Under this circumstance, variation in corporate governance mechanisms do not capture variations in quality, which is not within the scope of this thesis.

1.3 Motivation and contribution

This thesis contributes to the prior literature work in corporate governance field. A large stream of literature examined several proxies for the quality of corporate governance, and the impact of corporate governance on a wide array of outcome variables such as executive compensation level or more specifically defined as excess compensation. Core et al. (1999) suggested a negative function between governance structures and agency problems, where executives consume more compensation in firms with greater agency problems and weaker corporate governance. Other papers verified this relation through empirical studies in the context of different countries and regions (Conyon, 1997; Ozkan, 2007; Weintrop et al., 2007). I contribute to this stream of literature by means of examining the relation between corporate governance and investment decisions, and more specifically, whether corporate governance mechanisms are

(8)

associated with management investment decisions as measured by the indicator of investment efficiency.

My second contribution will be to the papers focusing on investment efficiency. Prior literature mainly looked at the impact of external governance factors on investment efficiency. Biddle et al. (2009) and Verdi (2006) examined the relation between financial reporting quality and over- / under-investment which is determined by one mechanism as manifested by the reduction of frictions. A limited number of papers investigated the consequences of firm-level factors on investment efficiency. Within these papers, monetary incentives are mostly researched, such as firm bonus plans, stock option, incentive compensation package, etc. (Dyl, 1988; Yermack, 1995; Murphy, 1999; Bebchuk and Fried, 2003). Richardson (2006) studied certain governance structures that have a positive relation with over-investment of free cash flow and at last called for a further research on abnormal investment under the existing framework but in a more general context as well. Based on this reason, I provide insights to this literature through examining the influence of internal factors related with corporate governance on investment efficiency.

Third and final, this study also features a societal contribution. Scientific research in business tries to explain executive decision making at firms. Especially of interest here are the investment decisions that executive managers make, and the extent to which these decisions are aligned with the interest of shareholders. Deviations from efficient investments may be beneficial for executive managers due to the maximization of self-interest but come at the expense of shareholders and can be considered a cost of the separation of ownership and control. Information about the extent that corporate governance can mitigate deviations from investment efficiency therefore is of interest for shareholders and researchers and reduces one of the costs of the widespread model of publicly held firms.

1.4 Structure

This paper is structured as follows. Section two is composed with literature review as well as hypothesis development. In section three, the research methodology is described, which includes sample design, model specification, and variables selection. Section four demonstrates findings of the regression model and hypotheses. Discussion and conclusion will be demonstrated in section five.

(9)

2. LITERATURE REVIEW, THEORY AND HYPOTHESIS DEVELOPMENT

2.1 Literature on basic theory 2.1.1 Agency theory

The development of modern corporate governance is the result of a rise of agency problems. Berle and Means (1932) are the first to raise agency theory, which puts forward a phenomenon that accompanied by a rapid emergence of corporations in large sizes, the concentration of economic power becomes more and more frequent, which gives birth to a powerful class of professional managers. They tend to make use of professional expertise to isolate themselves from supervisory mechanisms by shareholders and the public. Jensen and Meckling (1976) developed a framework of corporate governance from perspective of agency theory, defining agency relationship as a contractual connection, in which professional agents contract with the principal with the purpose of executing firm management and operation in exchange of a deserved remuneration. During the period of contract, different motivations and other emerging factors caused by information asymmetry and incomplete contracting theory (Grossman and Hart, 1986) may lead these agents to deviate from their contracted initial goals of maximization of shareholders and the firm (as self-interest is dominant among all motivations). To effectively monitor agent behaviors and reduce deviations, agency costs inevitably occur as a result of this contractual relationship between firm owners and the management.

With respect to principal-agent relationship, mixed interpretations make up of previous literature. Ross (1973) firstly defined and examined the agency relationship that arises between two or more parties, where one party manages decision problems with professionalism on behalf of the other party, and the work derived characteristics of non-Pareto efficiency concerning the solution to problems of the principal. Holmstrom (1979) examined the effects of imperfect information on principal-agent relationship as a consequence of moral hazard, which suggests that contracts of both parties have the space to be improved via any imperfect information about the actions or state of nature. Stiglitz (1988) argued the relationship between agents and the principal is substantially a compensation contract designed by the principal (owner) to impel the agent to take actions in congruence with their desired interest. Overall, the main characteristics of the different models described above is that the principal hires agents to perform a

(10)

related task on their behalf, and the former one acts as the residual claimant. Interests between the principal and the agent may not be perfectly aligned in the sense that self-interested agents may prefer activities (e.g., consuming perks) not in the best interest of the principal.

2.1.2 Theory of information asymmetry

Underlying agency theory models is the concept of asymmetric information under which local agents have informational advantages (mostly in the form of private information) relative to the principal. Traditionally, when involved in a market with complete information, participants are aware of all the knowledge necessary and essential for business transactions. According to Neoclassical Economics, perfect competition does not embrace information failure, consequently everyone is able to engage in business transactions under the circumstance of economic equilibrium. However, it is not the case in real life, where uncertainties and risks exist in business transaction process and consequently participants in different backgrounds can only obtain information in specific or related fields. Owing to information asymmetry, the party possessing more information is more prone to taking advantageat the expense of the other party with less knowledge. In firm level, the theory is interpreted as owners possess limited information about managers’ performance in firm decision-making and operation, but the management has more complete information regarding their own performance, which potentially gives rise to information asymmetry between two parties. Agency problems are of high possibility to arise because the agent will take advantage of information for self-interest maximization.

Akerlof (1970) firstly promoted one aspect of the information asymmetry theory indicated by “The Market for Lemons”, which refers to a market where asymmetric information dominates business transactions. For most buyers, the only way to protect their interests is to try to lower product price through communication and bargaining, and to make assumptions that most products feature fair average quality (FAQ) in the market. Consequently, products with low quality will gradually take the place of goods with relatively high quality, which is due to the fact that sellers own more information about products compared with buyers, and eventually bad products drive out good because of adverse selection. Later, Spence (1978) suggested the theory of job signaling by arguments and discussions in the dimension of job markets, where most employers

(11)

are not able to confirm job applicants’ productive capabilities at the moment of recruitment, and thus they evaluate applicants through conditional probability featured in combination of signals and indices. Then Stiglitz (1990) promoted models of signaling and screening from the perspective of game theory in the context of trading, which derived a more systematic theory of information asymmetry. The work of model integration by Stiglitz (1990) reveals an allocation of two categories within a time horizon, under which the first dimension is before trading which induces adverse selection, and the second dimension is after trading which leads to moral hazard. The phenomenon will distort the proper functioning of markets as well as market information.

The mixed research work all indicates a common characteristic of the theory of information asymmetry that imbalance of information possessed among two or more parties will lead to inefficient outcomes, and specifically parties with insufficient information will be worse off.

2.2 Investment efficiency 2.2.1 Definition

The interpretation of investment efficiency is more feasible within the category of economics. According to the theory of Pareto efficiency, under the circumstance of Pareto optimality, the reallocation process cannot make any participating individual better off without making at least one individual worse off. As a reference, Pareto optimality can then be further developed and applicated in the field of investment efficiency interpretation among public companies. When getting to the bottom of investment efficiency in nature, it reflects the Pareto efficiency of capital allocation, which aims at allocating limited resources to departments or projects with potentially highest marginal efficiency. If investing activities are instrumental in maximization of firm and shareholder interest, then these investments are supposed to be efficient. On the other hands, investments with returns higher or lower than this level are both considered to be inefficient in the sense that they have consumed improper levels of matching capital. Conceptually, the literature of Biddle et al. (2009) provided relatively the most precise definition with regard to investment efficiency, which refers to a firm accepting investment of projects characterizing positive net present value (NPV) in the context of inexistence of market frictions such as agency costs, and the first project

(12)

considered and selected is the one with the highest NPV; the second project characterizes NPV of the second highest, and so on. Investment efficiency can then be classified by two dimensions, where under-investment represents the miss of investment opportunities which tend to generate positive NPV in the absence of adverse selection, and over-investment represents investment in projects with negative and zero NPV.

2.2.2 Principal-agency relationship and non-efficiency investment

Berle and Means (1932) brought forward the agency theory, which marks the commencement of research in the context of modern corporate governance. Based on previous literature review concerning principal-agent problems, and the research of investment efficiency above, potential consequences brought by investment efficiency can then be categorized into two conflicting dimensions.

2.2.2.1 Conflicts between shareholders and management

The hypothesis of free cash flow (FCF) posited by Jensen (1986) reveals that the tendency of profit and firm size expansion by management is due to the personal psychology of pursuing management compensation comparatively higher than average market level. In this case, once the firm has acquired considerable free cash flow, the management will be impelled by the motivation mentioned before, and consequently managers become more prone to expanding investment scale frequently through taking projects that turn out to produce negative net present value in the long term but bring considerable value in first several years (short term), which refers to firm over-investment. Murphy (1985) extended the scope by proposing a hypothesis of the construction of an “empire” by management, and the benefits of which incentivize managers to dominate increasing resources which exceed the corresponding scale of internal incentive compensation in the way of frequent investment in new projects. Narayanan (1985) supported this finding by investigating the preference of the management in investing activities and revealed that the management is more inclined to invest in projects featuring short payback period even though these projects may have negative net present value in the long term, since managerial reputation can be built rapidly through considerable accumulation of value. With regard to this phenomenon, Scharfstein and Stein (1990) raised a concern by considering a situation

(13)

that some executives in upper level may therefore have an aversion to long-term as well as forward-looking investment in fear of failure and reputation damage. These selections will bring more incentives to place the allocation of free cash flow to shareholders in a sub-optimal position, which gives rise to conflicts between shareholders and agents with different interests and purposes.

2.2.2.2 Conflicts between large shareholders (controllability of blockholders) and small / minority shareholders

Holmen and Hogfeldt (2004) studied initial public offerings in juristic and financial fields and found the inclination of large shareholders in decision making. The power of controllability in decision-making motivates them to adopt investment of projects catering to their own interests but at the expense of small and minority shareholders, and the phenomenon of which can be stated as “tunneling”. Under-investment exists in this type of conflicts as well, because corporate free cash flow which is originally supposed to be used in projects investment, can be utilized by large shareholders with certain controllability to invest in their own financial companies, and those shareholders can thus obtain considerable personal gain by allocating residual value equivalent to bank deposit interest to the firm. Moreover, when the return due to self-interest maximization is higher than the yield allocated to the firm after investment, it is possible for large shareholders to miss projects with positive net present value, leading to under-investment.

2.2.3 Information asymmetry and inefficient investment

“The Market for Lemons” demonstrated by Akerlof (1970) commences the research of information asymmetry. Narayanan (1988) suggested that under the circumstance of information asymmetry, the inability of outside investors to obtain essential information which is equivalent to that of inside management turns out to be obstacles when they try to identify whether firm over-investment is caused by investment in projects with negative net present value, which violates the principle of maximization of shareholder interest. A study in stock market is more prominent in such kind of problems. According to Myers and Majluf (1984), the theory of pecking order suggests that internal financing is firstly advocated in decision-making of firm investment, and afterwards bonds and stock issuance are to be considered. It is suggested that when the management and

(14)

existing shareholders are in congruence with the same interest, the management will choose to issue stock at the moment when it is overvalued. On the other hand, adverse selection caused by information asymmetry may motivate investors (in other words, potential shareholders of the firm) to pressure the management to discount new shares, which has negative impact on fund raising and subsequent firm investment efficiency. Narayanan (1988), and Heinkel and Zechner (1990) put forward another case in stock market, where over-investment takes place in the case that information asymmetry exists in all-equity financing. This is due to the fact that the stock market values investment in new projects according to their average value, while market equilibrium involves a number of projects with various NPVs. Hence, the sale of overvalued stocks may cover the loss of investment in projects with negative NPV, which ultimately brings benefits to the firm as a whole. So once stock price is higher than the rational level, over-investment in projects with negative net present value arises as well.

2.3 Corporate Governance

Corporate governance exercises rules and practices to bring monitoring effects and influence the procedure of decision-making by managers when there exists separation of ownership and control (Larcker et al., 2005). Corporate governance mechanisms guarantee the providers of capital to the firm to get required return (Shleifer and Vishny, 1997). The objective of corporate governance is to pursue the balance among different stakeholders, under which the balance between the proxy and principal plays the most important role. Since corporate governance represents a complicated and multi-faceted construct, Larcker et al. (2005) developed substantive measures for corporate governance from a set of indicators. In consideration of this thesis’ research focus, several relative mechanisms are screened as follows.

2.3.1 Board governance

Board of directors is composed of several individuals who establish policies and make decisions concerning major company issues on behalf of shareholders. Besides, managerial behaviors are frequently monitored by the board. In most cases, the board is categorized by inside directors and outside directors. Inside directors consist of CEO who may also be the chairman of the board in some cases, other executives of the corporation and large shareholders with influential voting rights. Outside directors, also

(15)

frequently known as independent directors, represents board members who do not have material relationship with the corporation, and most of whom have professional expertise in certain specific backgrounds. Prior to the introduction of Sarbanes–Oxley Act (SOX), board size, composition and director characteristics were not formally addressed by the securities law (Karmel, 1984). However, considering the effectiveness of board functions, firms especially those trade stock publicly, have long been encouraged to set up board of directors as well as increase independent directors, which is evidenced by Linck et al. (2007) who found that board independence enhanced substantially in post-SOX period. Generally, the board of directors exercises two functions: monitoring and advising (Adams and Ferreira, 2007). The monitoring activities require directors to investigate and monitor executives’ behavior which may not align with or even harm firm and shareholder interest. The advising activities provide opportunities for directors from different technical backgrounds to offer suggestions and expertise regarding good decision-making in accordance with firm and shareholder interest. Overall, the design of the board should be customized to maximize shareholder wealth given the firm’s characteristics, after taking into consideration of the balance between costs and benefits that monitoring and advising functions will generate (Linck et al., 2007).

2.3.2 Blockholders

A blockholder is the owner who possesses large fraction of company shares, which provides them with influential voting rights as a reward. Though there is no specific number of shares that has been agreed on to define a blockholder, in this context the assumption of blockholders is the owners who have top several (one to five, accordingly) largest fractions of corporate stock. A high level of ownership concentration is encouraged when the firm pursues efficiency of decision-making process, since diversity and conflicts decrease in accordance with the decline in number of stakeholders. Furthermore, more blockholders are considered to function better monitors, because their fractions of shares make up of the majority of corporate stock, so investments in monitoring pay off for them instead for small shareholders. But it should be taken into account as well that ownership structure with overconcentrated level will lead to opportunities for blockholders to behave in collusion with the management through making decisions not in congruence with firm interest.

(16)

2.3.3 Incentive compensation

Incentive compensation is the supplementary compensation on top of base salary, which is measured by managerial performance and bears the responsibility of aligning managerial behavior to the interest of shareholders and firm. It is worth noting that monetary incentives may motivate employees to achieve short-term goals, yet it lacks persistence as a long-term motivator.

The description of the above three dimensions of corporate governance mechanisms may shed some light on some critical factors of the successful application of internal corporate governance. First, an effective corporate governance structure, including the interaction among board of directors and other participating parties within an organization, is of vital importance. Board of directors is one of the key elements in determining the effectiveness of corporate supervision mechanism (Fama and Jensen, 1983), therefore its composition and quality are supposed to function properly which should be in accordance with firm characteristics. Baysinger and Butler (1985) suggested that board composition, in terms of outside or independent directors, shows a mild but lagged effect on firm performance. A dysfunctional governance structure with characteristics such as CEO duality, lower fraction of independent directors, etc., is suggested to have a close relation with decreased firm performance (Daily and Dalton, 1992). Findings by Millstein and MacAvoy (1998) revealed a positive association between involvement of active board of directors and returns of investors in publicly traded firms. Second, an appropriate balance of shares fraction between blockholders and small shareholders should be taken into consideration. In firms with high level of concentrated ownership structures, blockholders are motived to maximize self-interest at the expense of small and minority shareholders owing to their influential voting power. Under this circumstance, conflicts between blockholders and small / minority shareholders take the place of the principal-agent problem. In order to address this modern corporate governing issue, blockholders are motivated by different approaches to play an active role in firm management to earn the trust of small investors and policy makers (Holderness, 2003). The third point is the role that incentive compensation plays in modern corporation. An appropriate design of compensation contract is able to significantly enhance the level of managerial productivity as well as to reduce agency costs thanks to a reduction in monitoring costs (Murphy, 1986). To solve the problem

(17)

that incentive compensation lacks long-term motivations, the work of Mehran (1995) implied that the form rather than the level of compensation is what incentivizes executives to continuously generate firm value. Besides, compensation with equity-based is suggested to be introduced and adopted in firms with more outside board directors.

2.4 Hypothesis development

The complicated governance structures in modern corporation involves participation of multi-parties, both internally and externally (Larcker et al., 2005). The internal governing factors indicated before turn out to be checks and balances among shareholders, board of directors, and firm executives.

2.4.1 Impact of board governance on investment efficiency

The board of directors owns a proportion of decision rights on material firm activities such as the appointment of management, decision of significant investment projects, merger and acquisition matters, etc. Generally, the optimal balance of a board size is 8 to 9 individuals, accordingly. Conversely, a board size comprising more than 10 individuals generates side effects such as a rising cost due to coordination and communication among various members, which may eventually exceed the benefits that the supervision mechanism will bring (Lipton and Lorsch, 1992). Jensen (1993) supported the finding that a board with relatively smaller size contributes to efficient communication, which potentially mitigates agency problems. The comparison between independent and executive directors implied a positive relation with the fraction of non-executive directors and effectiveness of investor protection (Byrd and Hickman, 1992; Brickley et al., 1994). Brickley and James (1987) found a negative association between CEO consumption and the fraction of independent directors, which is evidenced by the specialism of independent directors in different backgrounds and fields, which gives them opportunities to express relative opinions regarding investment related strategies, while few evidences are found concerning their professional knowledge in remission of under-investment. Based on previous literature, the following hypotheses are proposed:

H1a: The size of board of directors has negative relation with corporate investment efficiency.

(18)

H1b: The fraction of independent directors is positively associated with corporate investment efficiency.

2.4.2 Impact of ownership structure on investment efficiency

From the perspective of investors, concentrated ownership structure is one of the most common mechanisms adopted by corporate governance since there is a homologous relation between cash flow and investing activities. And simultaneously, firm and shareholder interests are protected through the accompanying effect of agency cost reduction (Shleifer and Vishny, 1997). La Porta et al. (1999) compared the ownership concentration of firms with sizes of top 10 which are incorporated in 49 countries respectively, and demonstrated a trend of following up with a more concentrated ownership structure in modern corporations. Recent literature indicated that in firms which have several or more blockholders, their influential power of voting rights awarded by share fraction is sufficient enough in decreasing management private returns as well as protecting interests of small and minority shareholders (Bennedsen and Wolfenzon, 2000; Francis and Hege, 2001; Maury and Pajuste, 2005; Pagano and Roell, 1998). Nonetheless, when the fraction of shares owned by blockholders reaches a certain threshold, blockholders’ influential controllability in decision-making and management appointment and removal begins to adversely create potential opportunities of collusion between blockholders and the management. In the meanwhile, the entrenchment effect aroused by holding large fractions of shares will decrease the supervising effect of investment efficiency related with managerial performance, which adversely acts on the pursuing process of firm objectives (Shleifer and Vishny, 1997). This kind of phenomenon will soon be vanishing once the fractions of blockholders’ shares increase and exceed a threshold, in which they become the majority of the entire firm’s interest as reflected in a major ownership share. Under this circumstance, the benefits brought by inefficient investment with blockholders’ personal preferences are mostly offset by the negative effects on interest in firm level. Ultimately, when the percentage of shares held by blockholders indefinitely approaches 100%, the preference of risk aversion again reverses the relation based on the premise that owners are rational individuals (Kapopoulos and Lazaretou, 2007). Therefore, the hypothesis in this section will be:

(19)

corporate investment efficiency.

2.4.3 Impact of incentive compensation on investment efficiency

Compensation contracts have been proved to be a practical approach to reduce agency costs and increase executive efficiency (Murphy, 1986). Information asymmetry occurs in public companies due to the impact brought by both internal and external uncertainties. To reduce agency costs which are resulted by moral hazard and adverse selection by the agent, compensation contracts related with managerial performance are frequently encouraged to align managerial behavior with firm interest, which in many cases are expressed as firm performance. While one of the most relevant tasks of executives are deciding about investments through which they can generate considerable value (Bertrand and Schoar, 2003), executives are more prone to reducing inefficient investment for the sake of self-interest. Simultaneously, a high level of managerial compensation is able to decrease agency costs which tend to reduce firm free cash flow, as discovered by Lewellen et al. (1987), who also suggested high remuneration rate shows close relation with high dividend payout rate. In this way, the possibility of over-investment due to the fact that firms own considerable free cash flow can be lowered. Based on the inference, the last hypothesis regarding incentive compensation will be:

H3: Incentive compensation is positively associated with corporate investment efficiency.

(20)

3. RESEARCH DESIGN AND METHODOLOGY

3.1 Sample selection

The sample of the thesis consists of 775 observations of 155 publicly held companies listed in the U.S. stock market (the majority of which are listed in NYSE and NASDAQ) over the period from 2012 to 2016. The consideration of the time-period chosen embraces both the impacts of timeliness and minimization of volatility effect caused by the financial crisis of 2008. Several relevant variables, such as firm growth opportunity (Growth) as expressed in the growth rate of operating income, involve the data with the time span of two consecutive fiscal years. Therefore, I expanded the scope of data gathering by including the data in year 2011, with the aim of achieving a more complete interpretation of variables. The preliminary data collection in terms of control variables (most of which are reporting items in financial statements) and information about incentive compensation can be obtained from financial statement data in CRSP and S&P Compustat annual database (with both active and inactive securities). The information of ownership structure and board details can be derived from Institutional Shareholder Services (ISS) and Orbis database. The entire sample includes securities which are either active or inactive in database.

The screening process includes two steps mainly. First, according to Global Industry Classification Standard (GICS), the classification of industries consists of 11 sectors and subsequent 24 industry groups. Sectors of financial institutions and real estates (sector codes of 40 and 60) are excluded from the sample in view of their characteristics of high market uncertainty and volatility of risk and return in relevant industries. In the next step, data shown as abnormal information is excluded, which is defined as insufficient information, obviously incorrect data under visual inspection (e.g. fraction of independent directors in board which exceeds 100%) and information provided in inconsecutive years.

3.2 Model description and variable definitions 3.2.1 The model of investment efficiency

According to prior literature on investment efficiency and the context of this thesis, the original model developed by Richardson (2006) in the angle of free cash flow, which characterizes the quantification of inefficient investment level, has long been widely

(21)

accepted and applied in literature work (Biddle et al., 2009). According to Richardson (2006), investment expenditure is divided into two parts: the fraction of investment expenditure on asset maintenance (e.g. depreciation), and the residual fraction of investment expenditure which is supposed to support new investments in the following year, and the new investment in the following year can then be accordingly sorted into over-investment or under-investment (abnormal investment) in projects with negative net present value and expected investment in projects. The fitted value in the regression is supposed to be the portion of the expected level of new investment, while the residual value of the regression (unexplained part) is therefore able to directly reflect the level of inefficient investment and more specifically the degree of over-investment and under-investment in the presence of positive and negative sign, respectively. Considering the empirical analysis of this paper, only the absolute residual value is required to reflect the exact level of inefficient investment. In combination with the work of Hubbard (1998), which studies investment decisions in corporate level, the model of Richardson (2006) can be spread as follows.

Model 1a

𝐼𝑁𝐸𝑊,𝑡 = 𝛼 + 𝛽1𝑉 𝑃⁄ 𝑡−1+ 𝛽2𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝑡−1+ 𝛽3𝐶𝑎𝑠ℎ𝑡−1+ 𝛽4𝐴𝑔𝑒𝑡−1+ 𝛽5𝑆𝑖𝑧𝑒𝑡−1 + 𝛽6𝑆𝑡𝑜𝑐𝑘 𝑅𝑒𝑡𝑢𝑟𝑛𝑠𝑡−1+ 𝛽7𝐼𝑁𝐸𝑊,𝑡−1+ Σ𝑌𝑒𝑎𝑟 𝐼𝑛𝑑𝑖𝑐𝑎𝑡𝑜𝑟 + Σ𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 𝐼𝑛𝑑𝑖𝑐𝑎𝑡𝑜𝑟

In this model, V/P, Leverage, Cash, Age, Size, Stock Returns, and INEW refer to the

control variables for firm new investments in year t-1, and Σ𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 𝐼𝑛𝑑𝑖𝑐𝑎𝑡𝑜𝑟 , Σ𝑌𝑒𝑎𝑟 𝐼𝑛𝑑𝑖𝑐𝑎𝑡𝑜𝑟 refer to two dummy variables in relation with industry membership and year, respectively.

According to Richardson (2006) model, the thesis is intended to build four separate multiple linear regression models (MLR) with the purpose of testing the relationship between the level of firm investment efficiency and relative board governance quality (H1a and H1b), ownership structure quality (H2) and the scale of management incentive compensation (H3). On the other hand, the consideration of four separate multiple linear regression models involves an effective minimization of the possibility of collinearity problems in statistics as well. In this case, the model of Richardson (2006) (model 1a) can be revised and extended with several relevant control variables, which

(22)

is demonstrated as follows.

Model 1b

𝐼𝑁𝑉𝑖,𝑡= α + 𝛽1𝐺𝑟𝑜𝑤𝑡ℎ𝑖,𝑡−1+ 𝛽2𝐿𝑒𝑣𝑖,𝑡−1+ 𝛽3𝐶𝑎𝑠ℎ𝑖,𝑡−1+ 𝛽4𝑆𝑖𝑧𝑒𝑖,𝑡−1+ 𝛽5𝐴𝑔𝑒𝑖,𝑡−1 + 𝛽6𝐼𝑁𝑉𝑖,𝑡−1+ Σ𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 𝐼𝑛𝑑𝑖𝑐𝑎𝑡𝑜𝑟 + Σ𝑌𝑒𝑎𝑟 𝐼𝑛𝑑𝑖𝑐𝑎𝑡𝑜𝑟 + 𝜀𝑖.𝑡

Where the residual of 𝜀_𝑖.𝑡 estimates in the model is defined as the level of corporate inefficient investment, according to Richardson (2006). A positive number represents a larger scale of investment compared with the expected, which is supposed to be over-investment; a negative number refers to a smaller scale of investment compared with the expected, which is supposed to be under-investment. The deviation of absolute residual value from zero reflects the level of inefficient investment.

Variables involved in model 1b will be interpreted in detail in the following section.

3.2.2 Definition and interpretation of variables

To test the four hypotheses regarding the relationship between investment efficiency and corporate governance quality, the metrics model is supposed to embrace relevant control variables as the determinants of investing activities in firm level. According to previous literature work on firm investment, variables such as leverage, firm cash holdings, firm age, firm size, and the level of new investment in prior period have been studied to reflect a high relevance with firm investment decisions (Barro, 1990; Bates, 2005; Hubbard, 1998; Lamont, 2000). During the process of studying the impacts that independent variables may have on the level of inefficient investment, lagging effect in dimension of years is of considerable importance, so data of the variables involved are aligned in terms of prior period (t-1).

Expenditure on new investment in prior period (INVi, t-1)

The ratio of new investment to total assets is regarded as a significant performance index for firm investing activities (Hubbard, 1998), which to some extent dominates many other factors. The management normally adjusts current year’s investment scale on the basis of last year’s investment performance to avoid unexpected fluctuation both internally and externally. This thesis adopts the approach of the aggregation of cash used for acquisition of intangible assets, property, plant and equipment (PPE), and

(23)

increase in investments, divided by total assets at the beginning of current fiscal year, as the indicator of expenditure on new investment in each period.

Firm growth opportunity (Growth)

Operating income represents the accounting figure reported in financial statements, which records the profit realized from firm’s investment in operating activities. In general, most firms generate income and value through sales of products and services, so the growth of operating income therefore can be of use in reflecting the promotion of sales, as well as firm growth. A steady increase in operating income indicates the competitiveness of a firm in certain markets, which motivates the management to enlarge the scale of investment as reflected in market share. In this thesis, the ratio of changes in operating income between two consecutive fiscal years divided by prior year’s operating income is derived as the interpretation of firm growth opportunity.

Debt to assets ratio (Lev)

Debt to assets ratio in this thesis is supposed to be in the form of total liability divided by total assets, which refers to a manifestation of firm capital structure in the dimension of debt. The impact of leverage level on investment scale is mainly due to the information asymmetry between corporate creditors and shareholders, and the consequent adverse selection and moral hazard. Since corporate creditors possess right of recourse with regard to the residual interest, which is of high relationship with managerial performance, the monitoring effect brought by these creditors can to some extent restrain managers’ behavior, which may ultimately have an impact on corporate investment decisions.

Cash holdings (Cash)

Firm cash holdings are the assets in the form of ready cash, which is frequently used as an indicator to reflect firm liquidity. Firm cash holdings play the role of an emergency reserve once the firm encounters emergencies or involves operating / investing activities which characterize cash outflow. Apart from daily general expenses, cash outflow in investment activities can reflect a high tendency for the management to enlarge the scale of investments. In this thesis, the definition of cash holdings is represented as the ratio of cash to current year’s total assets.

(24)

Firm size (Size)

While small business appreciates more controllability on firm strategy and daily activities, firms with relatively larger sizes are generally considered to have a longer period of decision-making process. However, large firms are supposed to possess more cash holdings and investment opportunities brought by their influential power in markets, which contributes to a high possibility of investment scale expansion and moreover over-investment. This thesis adopts the natural logarithm of total asset at each year end to denote firm size.

Firm age (Age)

The research of firm dynamics aspects suggests that with the increase of firm age and size, firm growth opportunity as manifested in investment scale, will gradually decrease (Evans, 1987). Firms in early stages are more prone to growing rapidly because of their few and straightforward strategies, practically concentrated controllability over the structure and the motivation of market occupancy in advance. When a firm comes into its mature stage, increasing aspects of the entire firm, such as balance among stakeholders with different interests, coordination of different strategies, etc. all require to be taken into account, which incentivizes the management to be inclined to more careful decisions related with firm investment. Traditionally, firm age is expressed as the period from the year of incorporation to current fiscal year, which is applicable to nearly all business nature. In this thesis, since the research objects refer to the characteristics of publicly held companies listed in the U.S. stock market, the firm age is consequently redefined as the years over the period from the fiscal year of initial public offerings to current fiscal year.

Industry and year dummy variables

Apart from the control variables mentioned above, dummy variables of industry membership and year are included in this model with the purpose of controlling the impact of industrial and year factors on firm new investment.

Table 1 below reports the summary of variables and their interpretation in the form of mathematics.

(25)

Table 1 Summary of variables

Category Variable Variable

Abbreviation Variable Interpretation

Investment efficiency

Inefficient investment in terms of the level of new

investment

INVi, t

(Cash for acquisition of intangible assets + property, plant and equipment + increase in investments) / total assets

Board of directors

Board size BDS Number of board members

Independent

directors ID

Proportion of independent (outside) directors in board Ownership

structure

Ownership

concentration OC

Proportion of the shares held by the largest blockholder

Management compensation

Incentive

compensation IC

Natural logarithm of the sum of top three annual management compensation

Control variables

Firm growth

opportunity Growth

(Current period’s operating income – last period’s operating income) / last period’s operating income

Debt to assets

ratio Lev

Current period’s total liability / current period’s total assets Firm cash

holdings Cash

Current period’s cash / current period’s total assets

Firm size Size Natural logarithm of current

period’s total assets

Listed years Age

Years over the period from the fiscal year of corporate initial public offerings to current fiscal year

New investment

in prior period INVi, t-1

(Cash used for acquisition of intangible assets + property, plant and equipment (PP&E) + increase in investments) / total assets

(26)

Industry dummy

variable ∑Industry indicator

9 industry sectors with 8 industry dummy variables

Year dummy

variable ∑Year indicator

5 years of sample with 4 year dummy variables

3.2.3 Specification of regression models between corporate governance quality and investment efficiency

Model 2a and model 2b represents two regression models which is intended to test the relationship between investment efficiency as manifested in the level of inefficient investment (INV) and board size (BDS) as well as the fraction of independent directors in board (ID). Model 2c demonstrates the regression model testing the relationship between investment efficiency in terms of the level of inefficient investment (INV) and the fraction of shares held by the largest blockholder (OC). It should be noted that to test the U-shaped relationship, a new variable of OC2 _{has been introduced. Model 2d}

demonstrates the regression model testing the relationship between investment efficiency in terms of the level of inefficient investment (INV) and the scale of management incentive compensation (IC).

Model 2a 𝐼𝑁𝑉𝑖,𝑡= α + 𝛽0𝐵𝐷𝑆𝑖,𝑡−1+ 𝛽1𝐺𝑟𝑜𝑤𝑡ℎ𝑖,𝑡−1+ 𝛽2𝐿𝑒𝑣𝑖,𝑡−1+ 𝛽3𝐶𝑎𝑠ℎ𝑖,𝑡−1+ 𝛽4𝑆𝑖𝑧𝑒𝑖,𝑡−1 + 𝛽5𝐴𝑔𝑒𝑖,𝑡−1+ 𝛽6𝐼𝑁𝑉𝑖,𝑡−1+ Σ𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 𝐼𝑛𝑑𝑖𝑐𝑎𝑡𝑜𝑟 + Σ𝑌𝑒𝑎𝑟 𝐼𝑛𝑑𝑖𝑐𝑎𝑡𝑜𝑟 + 𝜀𝑖.𝑡 Model 2b 𝐼𝑁𝑉𝑖,𝑡= α + 𝛽0𝐼𝐷𝑖,𝑡−1+ 𝛽1𝐺𝑟𝑜𝑤𝑡ℎ𝑖,𝑡−1+ 𝛽2𝐿𝑒𝑣𝑖,𝑡−1+ 𝛽3𝐶𝑎𝑠ℎ𝑖,𝑡−1+ 𝛽4𝑆𝑖𝑧𝑒𝑖,𝑡−1+ 𝛽5𝐴𝑔𝑒𝑖,𝑡−1 + 𝛽6𝐼𝑁𝑉𝑖,𝑡−1+ Σ𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 𝐼𝑛𝑑𝑖𝑐𝑎𝑡𝑜𝑟 + Σ𝑌𝑒𝑎𝑟 𝐼𝑛𝑑𝑖𝑐𝑎𝑡𝑜𝑟 + 𝜀𝑖.𝑡 Model 2c 𝐼𝑁𝑉𝑖,𝑡= α + 𝛽0𝑂𝐶𝑖,𝑡−1+ 𝛽1𝑂𝐶𝑖,𝑡−12 + 𝛽2𝐺𝑟𝑜𝑤𝑡ℎ𝑖,𝑡−1+ 𝛽3𝐿𝑒𝑣𝑖,𝑡−1+ 𝛽4𝐶𝑎𝑠ℎ𝑖,𝑡−1+ 𝛽5𝑆𝑖𝑧𝑒𝑖,𝑡−1 + 𝛽6𝐴𝑔𝑒𝑖,𝑡−1+ 𝛽7𝐼𝑁𝑉𝑖,𝑡−1+ Σ𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 𝐼𝑛𝑑𝑖𝑐𝑎𝑡𝑜𝑟 + Σ𝑌𝑒𝑎𝑟 𝐼𝑛𝑑𝑖𝑐𝑎𝑡𝑜𝑟 + 𝜀𝑖.𝑡 Model 2d 𝐼𝑁𝑉𝑖,𝑡= α + 𝛽0𝐼𝐶𝑖,𝑡−1+ 𝛽1𝐺𝑟𝑜𝑤𝑡ℎ𝑖,𝑡−1+ 𝛽2𝐿𝑒𝑣𝑖,𝑡−1+ 𝛽3𝐶𝑎𝑠ℎ𝑖,𝑡−1+ 𝛽4𝑆𝑖𝑧𝑒𝑖,𝑡−1+ 𝛽5𝐴𝑔𝑒𝑖,𝑡−1 + 𝛽6𝐼𝑁𝑉𝑖,𝑡−1+ Σ𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 𝐼𝑛𝑑𝑖𝑐𝑎𝑡𝑜𝑟 + Σ𝑌𝑒𝑎𝑟 𝐼𝑛𝑑𝑖𝑐𝑎𝑡𝑜𝑟 + 𝜀𝑖.𝑡

(27)

4. EMPIRICAL RESULTS

4.1 Descriptive results

Panel A of table 2 shows descriptive statistics on control variables of the revised Richardson (2006) model for firm investment decisions within the sample. In the sample with a size of totally 775 observations (155 firms), the level of inefficient investment as manifested in the ratio of yearly expenditure on new investment in comparison with total assets, shows a mean value of 0.12 (median = 0.06), with a maximum value of 0.32 and a minimum value of 0.02, which reflects a relatively huge distribution among the sample firms. The finding suggests that a large number of publicly traded firms in the U.S. stock market seem to be involved in the common problem of inefficient investment. From the perspective of external factors, it can possibly be explained by the data collection process which features a wide coverage of industrial sectors within the sample, and consequently firms in different industries may encounter unique local investment climate. Firm growth opportunity as manifested in the growth rate of operating income, shows a mean value of 49.08% (median = 6.554%), revealing a tendency of aggressive market expansion with respect to the sample firms. Firm leverage ratio (debt to assets ratio) shows a mean value of 0.53 (median = 0.54), reflecting a relatively high debt to assets ratio within the sample firms. This can probably become one of the potential determinants that prevent firms from further new investment, and the work of Fazzari et al. (1988) supports this finding by stressing that firm leverage significantly restricts the scale of additional funds raised by the firm in the dimension of investment. The finding of firm cash holdings, granted as a percentage of total assets, reveals a mean value of 11.91% (median = 9.25%), which suggests an overall loose liquidity of the sample firms. Firm ages in the form of years over the period from the fiscal year of firm initial public offerings to current fiscal year in the sample, shows an average value of approximately 20 years (median = 20 years), which represents a mature stage in which most sample firms get involved. Firm size refers to the natural logarithm of firm total assets, which averages at 8.41 (median = 8.33), with a maximum value of 9.91 and a minimum value of 7.00. The difference of firm sizes among the sample is fairly small.

(28)

Table 2 - Panel A

Descriptive statistics of variables in revised Richardson (2006) model

Variable Mean Std. Dev. 10% 25% 50% 75% 90%

INVt 0.123533 0.171813 0.016273 0.030557 0.058332 0.136954 0.320833 INVt-1 0.129116 0.180042 0.017188 0.031374 0.058976 0.144615 0.330952 Leverage 0.530726 0.223942 0.255857 0.3854 0.538243 0.640915 0.77663 Cash 0.119081 0.100922 0.014575 0.042048 0.092474 0.17028 0.262306 Size 8.412088 1.170123 6.995919 7.636575 8.332092 9.102143 9.908177 Age 20 6.970777 13 16 20 23 27 Growth 0.490781 10.94905 -0.38321 -0.0877 0.065424 0.198814 0.47721

Panel B of table 2 demonstrates the descriptive statistics results related with specific independent variables. From the perspective of board size, the sample firm board sizes (BDS) average approximately 9 to 10 members (median = 9 members), with a maximum value of 12 members and a minimum value of 7 members, which reflects an ordinary pattern of how firms set up the board of directors in terms of board size in the past few decades. This is consistent with the literature work of Lipton and Lorsch (1992) and Jensen (1993), which suggests that a board with smaller size may to some extent increase the efficiency of decision making. The fraction of independent directors in board (ID) shows a mean value of 80.84% (median = 83.33%), which is supposed to be a relatively high proportion. Besides, the entire sample firms explicitly present the finding that the sample of independent directors as a percentage of total board members has a maximum value of 90%, and a minimum value of 66.67%, which can shed some light on the global trend of the introduction of independent directors in board. The ownership concentration (OC) as manifested in fractions of shares held by the largest blockholder averages 10.55% (median = 9.1%), indicating that firms listed in the U.S. stock market are more likely to follow a dispersed ownership structure. However, it should also be noted that the gap between the maximum value (15.1%) and minimum value (6%) is relatively large, which can potentially be attributed to a number of factors, such as the number of shares issued, firm nature, and firms’ own understanding of blockholders, etc. Lastly, when it comes to management incentive compensation, the natural logarithm of the aggregation of top three annual management compensations (IC) averages 16.41 (median = 16.41), with a moderate difference between the maximum value (17.13) and minimum value (15.68).

(29)

Table 2 - Panel B

Descriptive statistics of independent variables in model 2a, 2b, 2c, and 2d

Variable Mean Std. Dev. 10% 25% 50% 75% 90%

INVt 0.123533 0.171813 0.016273 0.030557 0.058332 0.136954 0.320833

BDS 9.428387 1.846071 7 8 9 10 12

ID 0.808409 0.09735 0.666667 0.75 0.833333 0.888889 0.9

OC 0.10553 0.063037 0.06 0.0717 0.091 0.122 0.151

IC 16.41398 0.619919 15.68042 16.01755 16.40857 16.77044 17.13356

Table 3 reports the Pearson’s correlation coefficients (lower-triangular cells) and Spearman’s rank correlation (upper-triangular cells) between relevant variables. From the perspective of Pearson’s correlation test, the correlation coefficient is explicit in that the level of new investment in current period (t) has a positive correlation with the level of new investment in prior period (t-1), firm cash holdings (Cash), and firm size (Size). In contrast, a significant negative correlation between the level of new investment in current period and firm leverage (Leverage) in the form of debt to assets ratio, reveals that a decrease in debt compared with total assets can potentially shift firm’s strategic focus from debt redemptions to the propensity of seeking growth opportunities in terms of new investment. The correlation coefficient of the level of new investment in current period and prior period is 0.573, which passes the significant test (p < 0.01), and all the correlations of other variables reveal weaker results than that of the level of new investment in prior period, suggesting that the level of new investment in prior period may possibly have the most impact on the level of new investment in current period, and the management tends to determine new investment scale on the basis of last year’s investment performance.

On the other hand, a significant positive correlation is also explicit between the level of new investment in current period and the number of board members, the fraction of shares held by the largest blockholder, and the size of management compensation. The fraction of independent directors shows a significant negative correlation with the level of new investment in current period. The upper-triangular cells of Spearman’s rank correlation correspondingly support the similar results of correlation except the variable of firm size, which turns out to have no significant correlation. However, the result of Spearman regarding the variable of firm growth opportunity shows a significant correlation with the level of new investment in current period, which verifies the argument of a link between the growth of operating income and the level of investment

(30)

scale. The correlation between the level of new investment in current period and firm cash holdings is positive, but relatively weak.

(31)

Table 3

Correlation Coefficient

INVt INVt-1 BDS ID OC IC Growth Leverage Cash Size Age

INVt 0.664*** 0.113*** -0.160*** 0.121*** 0.042 0.112*** -0.179*** 0.068* 0.050 -0.026 INVt-1 0.573*** 0.060* -0.129*** 0.044 0.108*** 0.114*** -0.180*** -0.031 0.087** -0.013 BDS 0.151*** 0.063* 0.028 0.042 0.272*** 0.075** 0.240*** -0.227*** 0.524*** 0.120*** ID -0.097*** -0.079** -0.078** -0.150*** 0.151*** 0.007 0.253*** -0.011 0.097*** 0.015 OC 0.083** 0.025 0.123*** -0.204*** -0.177*** -0.157*** -0.067* 0.008 -0.095*** -0.004 IC 0.012 0.117*** 0.234*** 0.135*** -0.074** 0.115*** 0.186*** 0.015 0.567*** 0.075** Growth 0.043 0.040 -0.010 0.027 -0.013 0.026 -0.067* 0.026 0.052 -0.002 Leverage -0.171*** -0.199*** 0.181*** 0.171*** -0.046 0.163*** 0.017 -0.322*** 0.330*** 0.106*** Cash 0.090** -0.039 -0.174*** 0.042 0.016 0.004 0.064* -0.255*** -0.266*** -0.193*** Size 0.145*** 0.211*** 0.495*** 0.068* -0.007 0.539*** -0.013 0.248*** -0.280*** 0.229*** Age 0.048 0.064* 0.160*** 0.020 -0.001 0.081** -0.035 0.053 -0.179*** 0.274***

Lower-triangular cells report Pearson's correlation coefficients, upper-triangular cells are Spearman’s rank correlation *** p<0.01, ** p<0.05, * p<0.1

(32)

4.2 Results of hypothesis tests

4.2.1 Empirical test of revised Richardson model

Table 4 with panel A demonstrates the summarized regression statistics for model 1b. In general, the results with a sample of 775 observations over the period from 2012 to 2016 reveal a significant relationship between investment efficiency and corporate governance quality as categorized into three dominating dimensions: board governance, ownership structure and management compensation. From table 4, it is found that the correlation between the expenditure on new investment in prior period and the one in current period is the strongest among all control variables, with the coefficient of approximately 0.5 among four models, and it passes the test of significance (p<0.01). Firm growth opportunity as manifested in growth rate of operating income has a positive relationship with firm new investment expenditure, but the coefficients do not pass the test in all of four models, which indicates that the expenditure on new investment may not have strong connection with investment opportunity in the sense that investment decision making do not always embrace investment opportunities. The finding can indirectly reflect the possibility of inefficient investment in the sample firms. Firm leverage ratio and firm age both show negative relationship with new investment expenditure but are not significant in all four models, which indicates that the level of debt and firm listed years have a limited impact on firm new investment in the U.S. stock market. The potential interpretation of this finding is that the U.S. financial market features a mature system of debt financing business with diverse channels and relatively low financing costs. Moreover, firms possess limited advantages in relation with their position in terms of history in stock markets. Firm cash holdings show a positive significant correlation with firm new investment expenditures, which is in accord with the findings of previous literature that investment expenditure has a positive relationship with cash flow (Hubbard, 1988). The variables of firm size show discrepant results in the context of four models, and only the variable in model 4d shows positive significant relationship with the expenditure on new investment.

(33)

Table 4 - Panel A

Summarized model 1b OLS regression results

Variable INVt (M. 2a) INVt (M. 2b) INVt (M. 2c) INVt (M. 2d)

INVt-1 0.51531*** 0.51337*** 0.50948*** 0.51997*** (16.911) (16.606) (16.635) (16.974) Growth 0.00018 0.00017 0.00019 0.00020 (0.402) (0.376) (0.419) (0.429) Leverage -0.02477 -0.01654 -0.02597 -0.01885 (-0.983) (-0.640) (-1.027) (-0.742) Cash 0.18318*** 0.18318*** 0.18244*** 0.21103*** (3.233) (3.191) (3.206) (3.656) Size -0.00196 0.00954* 0.01060** 0.02150*** (-0.353) (1.867) (2.089) (3.551) Age 0.00063 0.00050 0.00047 0.00027 (0.772) (0.606) (0.570) (0.332)

Year control control control control

Industry control control control control

Constant -0.06087 0.06347 -0.10041* 0.49396*** (-1.250) (1.028) (-1.950) (3.416) Observations 775 775 775 775 R-squared 0.378 0.363 0.375 0.371 R2_-adjusted _0.362 _0.347 _0.358 _0.355 F-value 24.11 22.65 22.60 23.42 t-statistics in parentheses *** p<0.01, ** p<0.05, * p<0.1

4.2.2 Empirical test of relation between inefficient investment and corporate governance quality

Panel B of table 4 represents the OLS regression results of the four models which are supposed to test four hypotheses respectively. Five independent variables (the fifth variable of OC2_{was included to test the U-shaped relation of ownership concentration)}

all show relatively significant correlation with firm investment efficiency, which are in congruence with the outcomes of Spearmen’s rank correlation.

With respect to model 2a (H1a), the result shows a positive significant relation between board size and inefficient investment. The coefficient of 0.01567 (p<0.01) indicates that the number of board members has great impact on firm investment efficiency. A board

(34)

with a smaller size can both effectively and efficiently improve investment efficiency and mitigate agency costs. Conversely, a board with a larger size will bring potential problems such as an increased agency cost and the inefficiency of decision making among board members, which adversely acts on investment efficiency. Results from descriptive statistics also demonstrate a mainstream of firms in the U.S. concerning the establishment of board of directors, in which most sample firms have board members with no more than 10 individuals. These results verify hypothesis 1a and supports the previous literature work about the optimal number of board members (Lipton and Lorsch, 1992, Jensen, 1993).

Model 2b aims to test whether the fraction of independent directors has a mitigating effect on inefficient investment (H1b). The negative coefficient of -0.11685 (p<0.05) suggests a high fraction of independent directors of board does help to reduce the level of inefficient investment, which in other words increases investment efficiency level. The appointment and removal of independent directors should be isolated from the influence of firm owners and related stakeholders to guarantee a proper functioning of independent directors. Generally, there are two vital functions of independent directors: restriction of managerial power and keeping a balance among different stakeholders. When the management shows a tendency of inefficient investment because of self-interest maximization, the professional expertise and extensive business experience of independent directors will play a role in evaluation and judgement of managerial behavior. As a result, more independent directors in the board is supposed to increase more mitigating effects on inefficient investment.

The third hypothesis (model 2c) proposes an inverted U-shaped relation between ownership concentration and inefficient investment. The influential voting power of blockholders awarded by their large fractions of shares gives them a relatively high level of controllability, which creates potential opportunities for blockholders to appoint and be in collusion with the management to invest in projects with personal interest but at the expense of firm and small / minority shareholders’ interest. When the fraction of shares held by blockholders reaches a high level which refers to a dominating interest of the entire firm, the consequence brought by blockholders’ investment decisions will probably act on both the firm and blockholders themselves. Hence blockholders will then be inclined to efficient investment which is in congruence with firm interests. The result of OLS regression statistics is consistent with this

Corporate governance and investment efficiency in publicly traded firms

Corporate Governance and Investment Efficiency in

Publicly Traded Firms

CONTENTS