How do independent directors moderate the relation between blockholders and cost of debt : evidence from U.S. S&P 500 blockholding firms

(1)

How do independent directors moderate the relation between

blockholders and cost of debt: evidence from U.S. S&P 500

blockholding firms.

Name: Zhang Zhongyi Student number: 11325518

Thesis supervisor: Prof. V. O’ Connell Date: 23-06-2017

Word count: 13073

MSc Accountancy & Control, specialization: Accountancy Faculty of Economics and Business, University of Amsterdam

(2)

Statement of Originality

This document is written by student Zhongyi Zhang 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

This paper investigates effect of concentrated ownership and independent directors on cost of borrowing, and cost of borrowing is seen as a measurement of corporate governance quality. Specifically, it examines the how board independence moderates the relationship between blockholding percentage and cost of debt. Using total percentage of shares hold by blockholders (shareholders who hold more than 5% of total shares) and percentage of independent directors on board data in Standard & Poor’s (S&P) 500 corporations during 1996 – 1998 period, I find the following results: (1) Evidence suggests that compared to total blockholding, outside blockholdings are more closely related to lower cost of long-term borrowing. (2) This paper fails to find a statistically significant relationship between percentage of independent directors on board and cost of debt financing. (3) The positive effect of outside blockholders on interest spreads in long-term debt is impaired after introducing board independence as a moderator. (4) Additional sensitivity testing indicates that higher level of board

independence in large firms is associated with higher cost of long-term borrowing, this result further challenges the effectiveness of independent directors’ monitoring on governance.

Key words: blockholders, independent directors, corporate governance, and cost of debt

(4)

Acknowledgement

First of all, I would like to express my gratitude to my supervisor, Prof. Vincent O’Connell. He is always patient to help and giving timely respond to my emails. Although we cannot meet due to geographical long distance, his creative and comprehensive advice constantly steer me in the right direction.

I would also like to thank Dr. Pouyan Ghazizadeh, who is the lecturer in Financial Accounting Research course. He gave me inspirations and suggestions on my research question when I was writing my proposal.

My appreciation also extends to my family and my friends in China. They always support me and encourage me when I encounter challenges during my master study.

(5)

This study is motivated by recent theoretical and empirical studies on blockholders: when large shareholders exert monitoring in corporate governance, they can enhance governance quality in two mechanisms - “exit” and “voice”. “Voice”, which also refers to “active monitoring hypothesis”, implies that blockholders have incentives to monitor and then enjoy the benefits of intervention and the voting power (Shleifer and

Vishny,1986). The greater ownership they have, the more likely they are going to exert intervention directly. “Exit” also refers to “passive monitoring hypothesis”, it implies that blockholders may sell their stocks when they don’t want to initiate corrective action in the face of bad firm performance (Coffee, 1991). We can unite the two monitoring hypotheses into one - “shared benefits hypothesis”, which assumes that all other shareholders are able to share the benefits of efficient monitoring exerted by

blockholders. Consistent with this hypothesis, this paper suggests that higher percentage of blockholders ownership in total shares is associated with higher quality of corporate governance. However, an inherent free-riding issue is involved in “shared benefits hypothesis”: both mechanisms “exit” and “voice” assume that blockholders have private information on the firm, thus less informed stakeholders have incentives to free-ride on blockholders’ monitoring. Instead of improving corporate governance, blockholders can be detrimental to governance mechanisms by extracting private benefits of control or pursuing other private objectives. “Private benefits hypothesis” assumes that,

blockholders may tunnel corporate resources away from the firms to pursue private benefits (Edmans, 2014). Large shareholders have controlling power on management, thus they are able to manipulate the extent of disclosure to maximize private benefits. This kind of actions may be hidden in the information disclosed and is associated with poor corporate governance.

The blockholders are owners of the firms who gain large stakes, their existence is expected to alleviate agency problem to some extent. However, there is another institution- board of directors, which is especially set to solve the inherent shareholder-manager conflicts. In 2003, the world largest stock exchange NYSE amended its regulations, required that all the listed corporations to have more than 50% of independent directors on their boards. Later in 2005, SEC also amended its rules on

(7)

board independence in mutual funds. The successive changes on regulations indicate that the authorities may perceive “maintaining outsiders like independent directors on the board” to be an effective tool to curtail agency problem since outsiders are believed to be independent and effective in monitoring. Prior literature (Bhojraj and Sengupta, 2003; Fama and Jensen, 1983a; Weisbach,1988) also provides ample empirical evidence on the necessity to have an outside - controlled board. Recent studies always include ownership structure and board independence as corporate governance attributes, the two parties are both expected to solve agency problem, thus free-riding problem of monitoring is evident. Besides, ownership structure and board independence are interrelated (Bushee et al., 2004), it is interesting to study how independent directors’ participation in governance influence blockholders’ decision: whether they are going to enhance or weaken corporate governance with their private information.

Blockholders and independent directors both are assumed to exert actions and threats to managers, but their activities are not easily observed by public and minority shareholders. For example, Edmans (2014) points out that informal managerial actions like writing letters, also referred as “jawboning”, is an economical choice for

blockholders to intervene. But these actions can hardly be detected by other

stakeholders. To solve this problem, this paper adopts an indirect method to capture corporate governance by measuring cost of debt financing, which is shown to be

negatively related to governance level. Debtholders have no effective control on the use of the funds that they provide, they highly depend on the corporate governance quality when they charge the risk premium. Piot and Missonier-Piera (2007) studied French non-financial listed companies and arrived at the conclusion that corporate governance quality has a significant reducing effect on the cost of debt. Bhojraj and Sengupta (2003) used institutional blockholder as the key attribute of corporate governance, and

discussed the relationship between corporate governance and bond rating and yield in the U.S. during 1991 – 1996. They found that effective corporate governance is linked to higher bond rating and lower bond yields.

In this study, I restrict my samples to S&P 500 blockholding firms, during the time period of 1996-1998. It is before 2003, when NYSE requires all the U.S. listed firms to have a board of directors with more than 50% of independent directors. During 1996 to 1998, firms can autonomously choose board independence level which helps to filter out the influence of rigidly meeting regulatory requirements.

(8)

This paper makes three main contributions to existing corporate governance studies. First of all, it examines total blockholding and outside blockholding separately in

blockholding - cost of debt relationship, and finds that independence of blockholders plays a significant role on their monitoring decisions. Outside blockholders are perceived by creditors to be more effective in monitoring than insiders, thus creditors charge lower credit risk premium when they observe a high outside blockholding percentage. Secondly, I examine the effect of board independence on cost of borrowing before the outside –controlled boards are made mandatory in the U.S. stock markets. Even when firms can autonomously decide the percentage of outsiders on board, the relationship between level of board independence and cost of debt is not statistically significant, which suggests that from creditors’ perspective, independent directors may not be effective in supervision. This finding is inconsistent with results from most corporate governance studies, which indicate high percentage of independent directors on board contributes to higher level of corporate governance and thus leads to lower cost of debts. And after introducing the board independence as a moderating variable in blockholding - cost of debt relationship, the free-riding problem becomes severe when blockholders and independent directors are both aiming to monitor managers. Lastly, evidence from additional sensitivity testing further questions the function of

independent directors on governance.

The reminder of this paper is organized as follow. Section 2 outlines the literature review and the hypothesis development. Section 3 describes the sample selection

process and shows the process how control variables are chosen. Section 4 describes the data, outlines the methodology and results analysis, it also includes a firm size

sensitivity testing. The conclusions are shown in the last section. 2. Literature review and hypothesis development

2.1Terminology and Underlying theory 2.1.1Blockholders:

Until now, no unambiguous definition of a blockholder is acknowledged by

researchers in this field (Edmans, 2014). From the monitoring function perspective, any investor who has incentive to monitor managers’ actions can be classified as a

blockholder (Edmans, 2014). Generally speaking, prior empirical studies count

(9)

however, there is no underlying theory supporting this definition. The U.S. researchers commonly use this 5% holding definition in their studies since it is compulsory for investors to fill the Schedule 13 (specifically 13 F) when they cross a 5% threshold. However, the rigid 5% definition has a drawback: some shareholders may deliberately control their ownership to be just lower than the threshold, they are influential yet not counted as blockholders. “Form over substance” problem arises from this definition.

Tran (2014: 186) makes a specific definition of four types of blockholders as follow: (1) family, (2) financial, (3) corporate and (4) miscellaneous (group of managerial and governmental ownerships). And he points out that in large firms, block ownerships held by corporations or financial institutions are associated with lower interest spreads (Tran, 2014). And Dlugosz et al. (2004) group blockholders into two broad types: insiders and outsiders. The inside blockholders are (1) officers, (2) directors, (3) affiliated entities and (4) ESOP(1). They are either directly or indirectly related to firm operation, thus these insiders have access to private managerial information. The remaining

shareholders who hold more than 5% equity stakes but are not included in the any of the prior four types are classified as outside blockholders. In their study, they find a

statistically significant positive relationship between total percentage of outside blockholders and firm value.  

2.1.2Independent directors

Board of directors is an economic institution, and it is developed in modern corporations as a strong control device to deal with shareholder-manager conflicts (inherent agency problems). The main objective of directors is to monitor top managers, and they are motivated by market pressures and concern for reputation (Hermalin and Weisbach, 1991). Standard & Poor (2003) suggests that apart from ownership structure and quality of financial reporting, board composition is an important dimension to capture corporate governance quality. Maintaining high percentage of independent directors on the board is a strong signal of credible corporate governance mechanism. Nowadays, after authorities (like NYSE and SEC) put more and more restrictions on the board independence in public firms, the preliminary economic function of board of directors may have changed: some firms appoint independent directors to enhance

1_{ESOP - employee share ownership plans is a commonly used motivational tool, it gives stocks to employees to} closely related their interest to firm’s profit (Pugh et al., 2000)

(10)

corporate governance while others do so just to satisfy regulatory requirements. While after 2003, all the listed firms are with an outside-controlled board of directors, thus the investors may not be able to tell the quality of corporate governance merely based on the board independence information.

2.1.3 Agency theory

The separation of ownership and control leads to an unsolvable problem – inadequate incentives of managers (agent) to maximize the firm value (interest of shareholders). This problem originates from asymmetric information and it results in high cost of capital in the financial market. Besides the agency problem between

managers and shareholders, the agency problem of shareholders and bondholders is also unavoidable according to “wealth transfer hypothesis”. It indicates that when

shareholders motivate managers to engage in risky investments, bondholders bear a disproportionate share of the failures (Ashbaugh-Skaife et al., 2006). And there may exist a potential misalignment of interest between shareholders and bondholders (Jensen and Meckling, 1976).

Involving blockholders in monitoring or controlling activities may be useful in mitigating the inherent agency problem, their interventions in governance can help to curb managerial discretions which potentially cause such problems (Shleifer and Vishny, 1986; Huddart, 1993; Noe, 2002). Thus information transparency is enhanced and may lead to lower cost of capital which is beneficial to both the agent and shareholders.

Blockholders may not be directly involved in the company’s daily management. However, compared to small shareholders who can exert little influence on the company, blockholders have private information and can influence the corporate governance in several channels. The existence of these blockholders may ease or worsen the agency problem in some ways. This paper attempts to explore the blockholders influence on cost of debt through their influence on corporate governance.

2.2 Corporate governance and cost of debt

Corporate governance is difficult to be captured quantitatively, some researchers attempt to construct a corporate governance index to measure quality of corporate governance directly. Gompers et al. (2001) build a Governance Index and find that it is highly correlated with firm value, returns on equity and agency costs. This index is used as a key indicator in later corporate governance studies (Villalonga and Amit, 2006;

(11)

Bebcuk et al., 2009 and Klapper and Love 2002). In this paper, I am going to capture quality of governance from creditors’ perspective. Creditors base their judgement of default risk mainly on two things: information risk and monitoring costs, both of them are reduced in an enhanced governance environment.

The creditors charge the price of debt on the basis of their expectation of default probability. Literatures show that default risk and quality of corporate governance are highly correlated. In the survey of Coombes and Watson (2000), over 80% of investors said that they would pay a premium for well-governed firms. Effective corporate governance may reduce the risk premium demanded by creditors in two ways: (1) reducing agency risk. Managers may act to fulfill their private goals, and creditors generally expect that managers take actions which deviate from the corporate benefit. Thus creditors think it is reasonable to add a risk premium on their required price. If corporate governance is effective, internal regulations and stakeholders’ supervisions will contribute to detect and correct managers’ self - serving actions, then reduce the agency risk. Li et al. (2014) expect that if governance mechanisms reduce agency risk, the firms with stronger governance should have higher bond rating and lower yield. They also find empirical evidence to support their hypotheses.

(2) reducing information risk. Managers are the agents who really run the business, they have private information that creditors do not have. Dechow et al. (1996) point out that managers may be influential in financial reporting quality, they may issue misleading financial statements to hide negative news. If the corporate governance is in good condition, the reliability and usefulness of information disclosed may be enhanced, thus creditors are expected to charge less risk premium. Sengupta (1998) found a positive (negative) association between the quality of corporate disclosure and bond ratings (yields), suggesting that governance mechanisms can affect cost of debt through the reduced information risk. The negative relationship between information risk and cost of borrowing is confirmed by Pittman and Fortin (2004). They investigated on the effect of retaining reputable auditors (which is seen by public as guarantee of financial

statement reliability that can reduce information risk) and they found that enhanced quality of financial reporting induces lower cost of debt.

Monitoring cost is another crucial factor that creditors will consider when they provide funds to listed firm. Creditors may conduct monitoring in two ways: Direct monitoring and indirect (delegated) monitoring, both method are costly and strenuous

(12)

for them as outsiders. Direct monitoring is always observed as adding additional debt covenants in public bond contract. The monitoring cost increases with the complexity of covenant structure. However, most of the time, since bondholders are dispersed, the costs that individual creditors need to pay will definitely outweigh the benefits they can get through direct monitoring. Then creditors are more likely to delegate monitoring to other stakeholders. In a good corporate governance environment, the delegation can be more effective since creditors can free-ride on board monitoring and blockholders’ supervision. Li at el. (2014) find that the association between covenants in public bond contracts and borrowers’ corporate governance quality is less pronounced when

bondholders can rely on senior bank creditors for monitoring. 2.3 Blockholders and Corporate Governance

According to the definition of blockholders, they are shareholders with more than 5% of equity in a firm. They provide financial capital, but the operating activities are

performed by agents. Compared to minor shareholders, blockholders are more

concerned about whether capitals are used in an effective way to maximize their welfare (Mikkelson and Ruback, 1985). Blockholders are perceived to be influential

stakeholders, their voting power and exiting market option (blockholders can sell their stocks and exit the market) can keep managers away from self-serving activities thus enhance corporate governance.

2.3.1 Blockholders’ positive effect on corporate governance

Edmans summarizes the studies on blockholders up to 2013. He points out that there are two channels through which blockholders can affect the corporate governance in a good way. As an investor, blockholder have strong incentives to monitor managers and exert direct interventions when corporate governance is poor, this channel is referred to as “voice”. The other channel is through “exit”, which represents the potential threat or the real action of selling stock in order to hurt managers. The two mechanisms both assume that the blockholders have private information on V* (firm value with intervention or long-run fundamental value), but they have two distinct basic assumptions: In “voice” channel, blockholders are expected to intervene in management directly because they have a significantly large stakes. Blockholders have access to managers’ private information and they are motivated to supervise managers’ major decisions. Thus if managers are destroying firm value for their personal purposes,

(13)

blockholders are expected to take actions to intervene and to improve companies’ corporate governance. However, in the exit mechanism, blockholders can simply sell the stocks when the manager destroys V (firm value without intervention), and drive the stock price down towards V* through intervention. Both “voice” and “exit” channels assume blockholders are making use of their private information and influential power in a way which contributes to a well-functioning governance system (Shleifer and Vishny, 1997).

Results of prior researches are found to support “shared- benefit hypothesis”, which assumes blockholders engage in supervision and enhance corporate governance.

Graham, Harvey, and Rajgopal conducted a survey in 2005, and they find that over 75% of CEO are myopic when they make decisions on long-term investments. Top managers are prone to engage in earnings management activities since they are under the pressure of short-term earnings target. However, the problem of myopic management investment can be alleviated by blockholders’ monitoring. To make sure management are investing in aligned with shareholders’ interest, blockholders are motivated to collect private information about long-term fundamental values of the stock and they can intervene if they detect earnings management activities which deteriorate governance. Ajinkya et al (1999) document that higher level of institutional blockholding in public firms is related to better corporate disclosure condition perceived by financial analysts. Edmans (2009) points out that in the face of poor earnings performance, less informed stakeholders (like creditors, minority shareholders and potential investors) can make judgement based on blockholders’ informative trading actions. The uninformed parties expect that blockholders have access to private information. When the firm is going through a low earning period, firms are expected to be engaging in desirable long-term investment if blockholders do not sell their stocks, and the minor shareholders will not exit the market. Otherwise, less informed shareholders may infer severe agency problem inside the firm so they will follow blockholders if blockholders sell their stocks. Thus blockholders’ trading is a good way to communicate private information to the outside, it reduces information risk and contributes to transparent corporate governance. Accessing private information through non-public channels is not economical for creditors, however, the observable trading actions provide an inexpensive way to help creditors charge risk premiums. Besides the informative trading performed by blockholders, the exit threats

(14)

on managers can also reduce information risk. Exit threat motivates managers to provide higher quality financial reporting (Dou et al., 2016). Nevertheless, some

literatures hold different view of managers’ reaction to blockholders’ monitoring. Being aware of blockholders’ intervention, managers may be less motivated to search potential positive NPV projects because they are worried about their desirable projects are

opposed by influential blockholders (Burkart et al.,1998). 2.3.2 Blockholders’ negative effect on corporate governance

Instead of improving corporate governance, blockholders may do harm to

governance by extracting private benefits of control or pursuing other private objectives. There are three main negative effects of blockholders’ private- benefit seeking:

deterring transparent information disclosure, encroaching welfare upon the minority shareholders, and being myopic in making investment decisions.

Bushee and Goodman (2007) find that substantial change in institutional ownership is closely related to informed trading. In most cases, blockholders can take advantage of their private information and trade stocks when the shares are overvalued. On average, the stocks trades by blockholders are priced higher than the stocks traded by other minority shareholders who act after blockholders (Barclay and Holderness,1989). Blockholders are reluctant to share private information and thus they deter information disclosure which worsen corporate governance. Fan and Wong (2002) provide

additional evidence on blockholders’ negative effect on information quality, they find that concentrated ownership contributes to less useful earnings information disclosed by firms. For rent-seeking purposes, blockholders are likely to disallow private information disclosure to other investors.

To pursue private benefits, blockholders have chance to tunnel corporate resources away from corporations for their personal use. The tunneling activities include

transferring assets to their own businesses and encouraging managers to choose suboptimal projects where blockholders can get kickbacks. The actions may be hidden in the information disclosed, thus it is costly for minority shareholders to detect, and they can hardly intervene. Besides, the tunneling transactions are detrimental to firm value. Blockholders pursue their proprietary benefits at the expense of minority

(15)

shared with minority investors, in most cases, minority investors have no choice but follow blockholders’ trading and trade at a lower price than blockholders’.

Connelly et al. (2010) point out in their paper that, the managers are not the only stakeholder who may be myopic in choosing investment projects, blockholders also have incentives to encourage short-term projects instead of exert “voice” when managers have myopic tendencies. In this case, blockholders are expected to gain short-term profit by trading in large volumes, they may vote against the long-term projects that help the firms to gain competitive advantage. Chen et al. (2007) provide evidence that unstable institutional shareholders are speculating in shares, they trade frequently and care more about short-run returns.

Studies have not reach an agreement on which effect of blockholders on corporate governance prevails, and the evidence is mixed. This paper is going to investigate how creditors perceive the influence of concentrated ownership. We assume that, creditors are willing to charge less when they provide fund to firms with high corporate

governance quality. If the blockholders are perceived to devote themselves in

monitoring and improving financial disclosure rather than pursuing private benefits, the higher percentage of blockholders should be related to lower cost of borrowing.

There are various measurements of blockholding within a firm, the total numbers of blockholders, the aggregation of percentage of top five blockholdings, and total

percentage of blockholdings. Among these measurements, the total block size is regarded to be more relevant and commonly used in concentrated ownership research (Bushee and Goodman, 2007; and Parrino et al., 2003).

The preceding discussion motivates my first testable hypothesis:

Hypothesis 1: The higher the level of blockholder ownership, the higher the quality of corporate governance, which leads to lower cost of borrowing.

2.4 Independent directors and Corporate Governance

Managers operate but do not own the firms, this agency problem can rationalize managers’ incentive to consume perks, to overinvest and to engage in risky investments (Ertugrul and Hegde, 2008). All these actions are detrimental to corporate governance. The inherent agency problem can be alleviated by creating a board of directors to monitor managers. Inside the board, independent directors attract more attentions than insiders: they are expected to be more conscientious because they bear high reputation

(16)

cost (Fama and Jensen,1983b). Chiang and He (2010) points out that independent directors should maintain their independency to make sure that their supervision is not influenced by other stakeholders (like managers, blockholders or other directors).

Empirical studies have demonstrated many ways in which independent directors can improve governance quality. Lin (1996) finds that independent directors are active in major restructuring transactions within a firm, their participation will restrict managers’ motivation to expropriate firm resources thus protect shareholders’ welfare. Weisbach (1988) and Masulis & Mobbs (2013) indicate that an outside-controlled board is more likely to force poorly performing Chief Executive Officers to leave. And Core et al. (1999) also suggest that firms with higher level of independent directors on board face fewer CEO overcompensation issues. Managerial opportunism problem can be

alleviated by independent directors. Beasley (1996) points out a negative relationship between board independence and frequency of financial reporting misstatements. Dimitropoulos and Asteriou (2010) examine Greek listed companies and get the similar results: they document that usefulness of earnings in financial reporting is associated with high percentage of independent directors. The findings show that independent directors can exert effective supervision on corporate governance by reducing

information risk. When creditors can delegate monitoring tasks to independent directors, they may charge lower interest spreads.

The regulatory reforms in U.S. stock markets provide further evidence on how independent directors can enhance corporate governance. All the listed firms are required to have more than 50% independent directors on the board after 2003. Chen et al. (2011) compare noncompliant firms(2) and compliant firms, and find that poor information environment induces noncompliant firms to participate in earnings management activities. Armstrong et al. (2014) study the information asymmetry problem in noncompliant firms and find that financial information disclosure is substantially influenced after the regulatory shock of increased board independence. Firms improve their corporate information transparency, which may reduce information risk of creditors and enable borrowers to borrow at a lower cost. Guo and Masulis (2014) also examine the influence of amended regulations on corporate governance, they

observe an increased CEO turnover in noncompliant firms after listed firms are forced to raise level of board independence.

(17)

There are other literatures which fail to detect any significant relationship between governance quality and fractions of independent directors [Hermalin and Weisbach (2003); Baysinger and Butler (1985); and Bhagat & Black (2000)]. These studies question the usefulness of directors’ role in representing shareholders and monitoring managers. Independent directors are perceived to represent firm owners while they are not firm owners. Like other outsiders, it is costly for independent directors to acquire private and decision-relevant information. Core et al. (1999) suggest that researchers might have misplaced emphasis on outside-controlled board of directors. They challenge the effectiveness of outside directors, saying that outside directors are no more effective than insider directors. Harris & Raviv (2006) even argue that insiders with private benefits are investing more wisely than uninformed outside directors.

From creditors’ perspective, independent directors are generally perceived to be beneficial to governance. Ertugrul and Hegde (2008) examine the relation between stock option compensation for outside directors and corporate bond yields, they find that corporations with higher percentage of “independent directors’ stock and option compensation / total compensation” incur lower cost of capital financing. Using S&P 500 firm samples, Anderson et al. (2004) document a negative relationship between board independence and bond yield spread. Creditors typically requires the board to provide a credible audited financial statement in their lending agreements (Daley and Vigeland, 1983). The reliability of financial reporting is associated with the complexity of bond covenants. And board structure can help creditor evaluate information risk related to financial statements.

In line with major findings of prior research, this paper assumes that independent directors can influence corporate governance in a positive way: by reducing information risk and exerting effective monitoring. This argument motivates my second hypothesis as follow:

Hypothesis 2: The higher the percentage of independent members on the board, the higher quality of corporate governance, which leads to lower cost of borrowing.

2.5 The moderating effect of independent directors.

Ashbaugh-Skaife et al. (2006) indicate that prior studies typically focus on one attribute of corporate governance. However, they may ignore endogenous issue when

(18)

multiple governance characteristics are functioning at the same time. In three recent surveys conducted by Coombes and Watson (2000), more than 75% of investors said they are referring to board composition when they make major investment decisions. Bushee et al. (2004) also suggests that ownership structure can influence board independence

The “private- benefit hypothesis” assumes blockholders could make use of private information and controlling power, and they have incentives to harm other stakeholders for their private benefits. If “private-benefit hypothesis” prevails over “share-benefits hypothesis”, then high percentage of blockholders within a firm may worsen

governance and then result in higher cost of debt. As independent directors are outsiders aiming to monitor management, it is highly probable that independent directors can detect blockholders’ tunneling actions when they exert monitoring on managers. By introducing board independence as a moderator, we can observe whether blockholders will exert positive influence on corporate governance when they face more stringent external supervision.

However, the “share-benefit hypothesis” suggests that blockholders and board of directors are performing similar monitoring tasks. Then inherent free - riding problem can be detrimental to governance. If more than one party is involved in the same task, each party may view their contribution on monitoring as being reduced (Harris and Raviv,2006). This argument suggests that blockholders will spare less efforts on costly monitoring when they can delegate their jobs to independent directors. However, problem arises when blockholders enjoy more benefits than independent directors from high corporate governance quality. After weighing the cost and benefit, independent directors have less incentives to exercise influential monitoring. Thus I expect the positive effect of blockholders on corporate governance is impaired after I introduce a moderating variable (percentage of independent directors on board) into the relationship between blockholders and cost of debt.

The two competing views result in divergent outcomes after the moderating effect of board independence is included in the relationship between blockholders and cost of borrowings. Consistent with my first hypothesis, I expect the positive effect of

blockholders is more obvious even after introducing a moderating variable (board independence), which leads to my third hypothesis:

(19)

Hypothesis 3: Higher board independence enhances the creditors’ perception on blockholders’ positive effects on corporate governance, and thus leads to lower cost of borrowing.

3. Methodology 3.1 Sample selection

In this section, I will illustrate the detailed sample selection procedure, it starts with merging the two key database (percentage of blockholders total ownership and

percentage of independent directors on board) during fiscal year 1996-1998 which gets 1172 observations at the first stage. Further selection process excludes 441 items from initial pool of samples, and my empirical study is based on the remaining 731

observations from S&P 500 firms.

Since blockholders are the key factor in my hypotheses, this section starts with a discussion of the blockholders database provided by Dlugosz et al. (2004) which aimed at fixing the mistakes and biases (overlaps and preference shares) of ownership

information in Compact Disclosure and providing a less noisy database for later studies. Dlugosz et al. (2004) combined IRRC database with Compact Disclosure, since IRRC provides data of S&P 500 firms whose corporate governance data are available and quite complete. Their database excluded all multiple-class companies in Compact Disclosure, which contains mistakes that are difficult to fix. As Demsetz & Villalonga (2001) indicated, the ownership structure variable cannot be treated endogenously, since shareholders are made multi-dimensional. Dlugosz’s study divides the blockholders into 5 different types. To make it simpler, I only adopt “inside or outsider” criterion in classification. The insiders refer to officers, affiliated entity, directors and ESOP. The outsider is defined as other shareholders with more than 5% shares who are not classifies as inside blockholders.

Since this paper use Dlugosz et al (2004) database, I follow the definition of “blockholder” in their paper- “shareholders with more than 5% shares”. There is an alternative measurement - “the necessity to fill a Schedule 13 (specifically 13 F) in the U.S”, however measurement of Dlugosz et al (2004) is more accurate since they collect blockholding data directly from corporation annual proxy database, thus this paper

(20)

directly uses their blockholding database. I choose firm year from 1996 to 1998 to establish the panel data, this time period selection is made for two reasons: one reason is due to the fact that bond financing volume soared in this period, from 651 billion dollars to 1001 billion dollars (Bhojraj and Sengupta, 2003), almost doubled in three years which makes it interesting to study the change of cost of borrowing; another reason is under consideration of the blockholder database availability – a less noisy database can be found in “WRDS- Blockholders” database.

The board independence data are derived from RiskMetrics - Directors Legacy. The percentage of independent directors of the board data is from Board Affiliation. I use tickers of Blockholder database of that firm year to match Blockholder and Independent Director data, and nearly all the board affiliation of firms in Blockholder database can be found.

The dependent variable - cost of borrowing and all the control variables are from Compustat, (North America- Fundamental Annual). I choose the U.S. firms for two reasons: one is the data availability, it is found in a study (Holderness, 2009) that more than 90% of U.S. listed firms consist of at least one blockholder; the other reason is the strong shareholder protection regulations and their stringent enforcement in the U.S. (Harford 2006). These regulations can reduce agency cost to some extent, and make the regression results more evident.

The details about the variable definition are listed in Table 1. Insert Table 1 here

After the initial merge of blockholders, RiskMetrics and Compustat database, I exclude the missing and zero value blockholding sets, and the get 391 observations in 1996, 340 in 1997 and 441 in 1998. As the dependent variable – “cost of average short-term and long-term borrowing” is an indirect measurement, I adopt Byun (2007) trimming procedure and exclude the observations which fall in 1st or 99th of the population of leverage, “cost of borrowing” and “LTcod(3_{)”. Besides, I manually}

exclude the items with missing components in Z-score, and firm age under 1 year. Before running the panel data regression, I also found four years with reoccurring ticker

(21)

items in the same year, I drop the ones with zero leverage. After the final trimming process, 768 firm-year observations are left.

Since the regulations on financial service industry are more stringent than that of non-financial companies, and also the borrowing decisions are affected by some

industry-specific factors like “deposit insurance” (Pittman and Fortin, 2004), the cost of borrowing in financial industry is not comparable to that in non-financial firms. In addition, there are some firm tickers reoccurring since these corporations are included both in non-financial industries and financial service industries. To make the results less noisy, I exclude all financial firms, insurance companies and real estate companies before I do the regressions. To make it clearer, I exclude firms with SIC code ranging from 6022-6200, 6312-6400 and 6500-6799 (Pittman and Fortin, 2004). It further limits sample firms to 731 firm- years.

3.2 Control variables • Z-score

Z-score was initially developed by Altman in 1968 to measure the bankruptcy risk of firms. This model combines 5 ratios allocated with different weights and it is widely used in later interest rate studies and demonstrated to have high accuracy in predicting financial strength of a firm. The 5 ratios are selected from 22 potential indicators (Altman, 2014). From liquidity point of view, “working capital/total assets” is empirically demonstrated to excel quick ratio and current ratio. As for profitability indicator, “retained earnings/total assets” is adopted due to its cumulative predictive nature, and “EBIT/ total assets” is chosen to filter out tax and leverage influence. The leverage ratio definition is not uniformed, the most commonly used are “debt to equity ratio”, “debt ratio (total debts to total assets)” and “equity ratio (total equity to total assets)”. This paper adopts the “debt to equity ratio” defined in Petersen and Rajan (1995) to measure insolvency, and Altman finally adds “sales/ total assets” measuring the operating efficiency to Z-score.

I first bring Z-score as a control variable after weighing the fitness of various proxies of financial strength and default risk. It is better than standalone ratios like quick ratio or return on equity in prediction. I assume creditors based their interest

(22)

premium charges on the comprehensive measure of liquidity, insolvency and profitability. When borrowers are less likely to go bankruptcy or to be in financial distress, the creditors are more ensuredto get their principle and interest back on time. According to Z-score interpretation by Altman, Z-score is adversely related to

bankruptcy risk. And creditors charge interest premium based on default risk (which is highly related to bankruptcy risk) of the borrower, thus I assume that Z-score is

adversely related to cost of borrowing. • Tangibility

Tangible assets can be used as collaterals when corporations need external financing, prior literature states that asset tangibility is a significant determination of firm’s credit risk and credit constraints (Almeida and Campello, 2007). When firms are nearly to go bankruptcy, higher tangibility indicates higher probability for creditors to recapture debt principles in bankruptcy liquidation procedure. Besides, low level of tangibility always exacerbates the shortage of external funds condition, and drives firms to tap higher cost of debt from new creditors. Thus it is often argued that tangibility and cost of borrowing should be negatively related. Nevertheless, views on this relationship are mixed, some researchers argue that main credit providers like banks may require higher level of leverage in screening riskier borrowers (Morsman, 1986), loan spreads may be higher in firms with more PPE(4). In this paper, I assume there is an adverse relationship between asset tangibility and cost of borrowing.

• Market to book ratio

Market to book ratio which is also referred as P/B ratio is acknowledged to be a good indicator of firm’s growth opportunity. By using the credit spread and controlling other variables, Chen and Zhao (2006) empirically show that high P/B ratio can lead to lower cost of borrowing and have access to more funds. More and better growth opportunities mean high gearing, which will lead the firms to pursue positive NPV investment opportunities (Rajan and Zingales 1995, Myers 1977). Pursuing positive NPV investment rather than private interest of managers is consistent with shareholder’s interests, it to some extent alleviate agency problem and make the firm with higher

(23)

growth potential more attractive to creditors. I expect that there is a negative relationship between P/B ratio and cost of borrowing.

P/B ratio can be captured in two ways, either total assets or net assets method. In this paper, I am adopting the way Hovakimian et al. (2001) used, which is defined as “(market value of equity + book value of debt) / total assets”, the total assets method. • Leverage

In the first control variable of Z-score, leverage is taken as a part of the multiple discriminate model. This paper is going to include leverage as a standalone control variable. From the Jensen and Meckling (1976) point of view, agency cost increases with the leverage, thus leads to higher cost of borrowing. This view is challenged by Green and Talmor (1986), arguing that shareholders in highly leveraged firms are inclined to avoid negative NPV investments. Although the opinions towards relation between leverage and cost of borrowing are mixed, this paper assumes that there exists a positive relationship, since we build our basic hypothesis on agency theory. • Board size

Board structure is generally discussed in two aspects: board independence and board size. I already bring independence as a moderating variable, thus I include board size as a control variable. Prior literatures do not reach an agreement on the impact of board size on board effectiveness. The larger board size can exert powerful monitoring on the one hand, and may lead to rigid and time-consuming decision making procedures on the other hand (Harford et al. 2006). Based on Irish companies, O’Connell and Cramer (2010) demonstrated that board size and firm performance are inversely related, which supports prior research on U.S. samples. It is clear that there exists a tradeoff between monitoring and efficiency. My prediction of the direction of relationship between board size and cost of borrowing is negative.

• Firm age

One of the important determinant of interest premium charge is based on firm reputation, the longer time firms stay active in stock market, the more likely for firms to accumulate reputation through reliable credit records. Diamond (1989) declares that

(24)

loan financing cost may decrease over time during lives of public firms. This view is supported by Pittman and Fortin (2004) in their study of how retaining Big Six impact cost of external financing during the first nine years after firms go public. The positive impact of retaining reputable auditors is impaired over time, since firm age is enough to provide lengthy credit history. This paper predicts an inverse relationship between firm age and cost of debt.

The data description can be found in Table 2 Insert Table 2 here

The samples are selected ranging from 1996 to1998. The mean of cost of borrowing (8%) is slightly lower than that of LTcod (11%), this variation is due to the lack of long-term debt expense data in Compustat, thus I used “interest and related expense total” as the numerator for both ratios. The mean of total blockholders is 26.9%, higher than that of outsideBH (19.7%). This results shows that among the corporations with concentrated ownership, most of the aggregated blockholders have more than 20% of the total shares- much higher than the blockholding definition (5%), the blockholders are less likely to be insiders and shareholders at the same time. Descriptive data of board independence indicates that even without mandatory regulation, most of the listed firms have an independent board, the mean and median are both above 60%. As for the bankruptcy proxy of Z-score, which related closely to the default risk. Both mean and median cross the threshold of assumed healthy firm (Z-score > 2.675).

3.3 Models formulation

The research model designs build on the prior work of Bhojraj and Sengupta (2003) searching for the relationships of ownership structure, board structure and bond yields and Pittman and Fortin’s (2004) work on how retaining Big six auditor impacts cost of borrowing. Bhojraj and Sengupta (2003) suggest that more effective corporate

governance mechanisms (aggregated effects of blockholders and independent directors) are linked to lower bond yields. They use bond spreads (spread over U.S. treasury bond interest) to be the dependent variable. In Pittman and Fortin study, they bring in an indirect measurement of cost of debt as the dependent variable. Due to data availability and feasibility, this paper uses the indirect measurement: average interest rates of total

(25)

short-term and log-term borrowing, and makes some modifications by providing an alternative: only test on long-term debt.

Before studying the moderating effect of board independence, it is better to figure out the standalone effects of blockholding and board independence respectively on cost of borrowing.

To test the first hypothesis of the Blockholder percentage effect on cost of borrowing, I include all the control variables I mentioned together with blockholding percentages and adopt “Model 1” in below:

Model 1:

Cost of borrowing_- = α + β₁BLOCK_-+ +β₆BDSIZE_-+ β_<LEV_-+ β_>Z − score_-+ β_BTANGI_-+β_GP/B_-+ β_JAGE_-+ ε

For the BLOCK variable, I am going to test total blockholder percentage and outside blockholder percentage respectively, both of them and the six control variables are going to be regressed on two measurements of cost of borrowing, the “interest cost on average short-term and long-term debt” and “cost of debt on long-term borrowing”.

The results of Model 1 regression are presented in Table 2 Insert Table 2 here

Based on the regression results in Table 2 (detailed illustration in Section 4), I choose to adopt “cost of long-term debt” as cost of borrowing proxy and “outside BH” as blockholding proxy for the next two models.

To test the sole effect of board independence on firm’s cost of long-term debt, I introduce “model 2” in below, testing Hypothesis 2:

Model 2:

Cost of borrowing_- = α + β₁INDB_-+ +β₆BDSIZE_-+ β_<LEV_-+ β_>Z − score_-+ β_BTANGI_-++β_GP/B_-+ β_JAGE_-+ ε

The final step in my main empirical tests is to combine outside blockholders and board independence together, I will test my samples based on the following model.

(26)

Model 3:

Cost of borrowing_- = α + β₁BLOK_-+ β₆INDB_-+ β_< BLOK_-, INDB_- + β_>BDSIZE_-+ β_BLEV_-+ β_GZ − score_-+ β_JTANGI_-+β_MP/B_-+ β_NAGE_-+ ε

Table 4 lists pooled OLS regression results of model 2 & model 3. Insert Table 4 here

4. Empirical results and interpretation

4.1 Results of corporate governance proxy on cost of borrowing

The pooled OLS statistical results of model 1 are shown in Table 3. Based on the significance indications, I found that compared to total blockholding of the firm, the outside blockholders are more influential to the cost of long-term debt borrowing, and OLS regression shows that the relationship is negative and statistically significant at 5% level (t= -2.09). From corporate governance quality point of view, the result indicates that the effect of “shared benefit hypothesis” prevails over that of “private benefit hypothesis” when firms with higher percentage outside blockholders raise fund by issuing long-term debt. Nevertheless, the result reverses when I regress total blockholders on cost of total borrowing, the correlation is marginally positively significant with a t-value of 1.73. And the total blockholding percentage failed to be found any significance linked to cost of long-term debt. To rationalize the conflicting results, we have to be clear that keeping other influencing conditions the same, short – term borrowing is always less economical than long-term. Thus the cost of total

borrowing reflects a mixed effect after combining short-term and long-term borrowings together. And it is generally accepted that it is the top managers instead of shareholders that are making major financing choices. In prior studies of corporate governance and choice of maturity structure of corporate debt, it is found that managerial stock owners are inclined to borrowing in short-terms (Datta et al, 2005), since managerial(insider) ownership to some extent better aligns the interests of managers and shareholders. The marginally positive relation between total blockholders and cost of borrowing may have resulted from the influence of inside blockholders on cost of short-term borrowing, which leads to higher cost of capital in terms of indirect and mixed effect measuring of debt price.

(27)

To filter out the influence of indirect measurement of cost of total debt, I exclude the short-term borrowing and inside blockholding at the same time, I test only on the outsiders to alleviate the mixed effects of blockholders with different interests. Then I regress outside BH together with other control variables on “LTcod”, I assume outside blockholders have fewer incentives and opportunities to tunnel corporate resources for private use compared to insiders. And the regression results confirmed my assumption of coefficient direction: pooled OLS regression shows that the relationship between percentage of outside blockholders and cost of long-term borrowing is negative (coefficient=-0.068), and statistically significant at 5% level (t = 2.09).

Board of director independence is always taken as a strong indicator of corporate governance quality or taken as part of multivariable governance grading. And several empirical studies based on some major capital markets in the world have demonstrated that firms with higher independency board can borrow at lower cost, face less strict debt covenants and also rated higher in credit rating. Nevertheless, this paper fails to find significant evidence of positive effect of independent directors on cost of borrowing, tested both on cost of average short-term and long-term borrowing and cost of long-term debt.

The moderating effect of board independence on the relationship between

blockholding and cost of debt also fails to find empirically significant support. And the significant correlation between outside blockholders and LTcod is impaired when I introduce the moderating variable of ID(5). This result can be explained from several aspects:

First, the independent board directors are different from outside blockholders in nature. Blockholders provide capital to corporation, and their funds are managed by agents. Compared to insiders who actually engaged in managing the corporations, outside blockholders’ information sources are usually limited to financial statements, stock price fluctuations or major officer turnover. They have motivation to monitor whether their funds are misused or transferred and the agency problem is more intense than that between managers and independent directors. Since independent directors are not necessarily investing in the corporations they are providing service, in some cases,

(28)

they are offered the position due to relatively high reputation. So the true impact that independent directors made to corporate governance is limited since they have fewer incentives to monitor how funds are used than fund providers (blockholders).

Secondly, outside directors are proven to be ineffective in corporate governance in prior literature. The impaired effectiveness may due to cultural consideration

(Mace,1986), since directors are offered jobs by managers, they are reluctant to act against managers when they actually implement monitoring actions. According to Harris and Raviv (2006) study on board composition, the benefits of having inside directors sometimes outweigh the costs. Inside directors may not try to maximize the shareholders benefit due to private consideration, however, their access to inside information can help them manage the firms more efficiently. Thus contradicting the conventional views, they argue that sometimes it is optimal to have insider-controlled board. This argument is backed by evidence found in many surveys which fails to find significant relationship between board independence and firm performance. Due to the fact that insiders are able to do their jobs well and improve operating efficiency using private management information, the underlying theory of agency problem is

challenged and setting an independent board seems unnecessary.

Thirdly, the t-value of [outsideBH, ID] is 1.55, while t value of outisdeBH is only 0.71 in our model 3. This impairment can be explained by outside blockholders and independent directors’ motivation to free-ride on each other. The outsideBH and ID are both outsiders, prior research (Bushee et al., 2004) documented that these two variables are correlated to some extent. And either party has incentive to free-ride on the other’s monitoring. After adding ID as a moderating variable, the positive effect of outside blockholders on corporate governance may be undermined owing to the delegation problem. However, ID’s influence on corporate governance and operating efficiency is muted, indicating that the delegation turns out to be ineffective, which explains the insignificant t-test of outside blockholders on LTcod.

The signs of control variables are mostly consistent with my predictions. Apart from Z-score and firm age, other control variables are empirically significant. The insignificant result of Z-score can be explained by the sample characteristics. All the blockholding firms are derived from S&P 500, both mean and median of their Z-scores

(29)

do not fall in the range which indicate high possibility of insolvency. Above the threshold of 2.675, the z-score do not show strong significance in the regression. 4.2 Firm size sensitivity test

This section extends the main research field by addressing the potential differential impacts of corporate governance proxy on cost of debt according to divergent firm size. This sensitivity test is inspired by O’Connell and Cramer (2010) and Byun (2007). There are several ways to measure firm size value, Carey et al. (1993) uses “ln (1 + total assets)” in measurement, while I adopt “log (firm’s total assets)”. The classification of large and small firms is based on the median of this proxy in sample population.

However, the large firm grouping and the small counterpart are not tested

separately, I first add firm size proxy as a control variable into the moderating model 3, test on total populations to see the sole effect of firm size on cost of borrowing, then a dummy variable D which substitutes the original firm size proxy is introduced in model 5 to test differences between large and small firms.

It is generally perceived that large firm can borrow at lower cost. The reduced cost can result from high quality of corporate governance in large firms or other factors like economies of scale and external public monitoring. Literature has not reached a

conclusion so far on whether firm size is playing a positive or negative role in corporate governance. Our main research aim is to study how two corporate governance attributes impact the perceived default risk by creditors. In this part, I am going to integrate the generally acknowledged influential variable- “firm size” into my prior model – Model 3 to test firm size’s impact in blockholding firms.

From corporate governance perspective, both outside blockholders and independent directors are supposed to influence corporate governance effectiveness through

monitoring (Tanaka, 2014). Fama (1985) indicates that compared to small firms, large firms are under even more scrutinized external monitors. In Holderness (2009) study of the diffuse ownership in United States, he claims that apart from firm age, firm size is most negatively related to concentrated ownership. As individual blockholders are less likely to get in control when firms become larger and their wealth is constraint. As for literatures on board of directors, Lehn’s research in 2009 indicates that firm size can

(30)

influence board composition, and Finkelstein and Hambrick (1996) suggest that in small firms, directors are more likely to influence the strategic direction. Then I can infer that this kind of dominant impact of directors may be impaired to some extent in large firms when other parties (like managers, shareholders and creditors) can also voice for their interests and protect the quality of corporate governance from detrimental tunneling intentions.

From dependent variable perspective, Fama and Jensen (1983a, b) and Carey et al. (1993) both find that external financing is a substantial expense of contracting costs, the large firms can enjoy the benefit of economic scale, while the related debt incurred cost can be a burden for small firms. Lin et al. (2001) also build their research on the

assumption that larger firms are confronted with less information asymmetries when they raise capital by issuing debt. This assumption is supported by the fact that large corporations which are not new in stock market maintain traceable credit records and they are frequently targeted by financial analysts and public attention. In this case, information risk of creditors is reduced in large firms, which enables large firms to borrow with lower debt cost compared to relatively smaller firms. Thus large firms are perceived to be less risky from creditors’ point of view, thus they may face lower external capital costs.

Due to the fact that firm size is likely to be associated with both corporate

governance variables and dependent variables, the following two models are introduced to examine the potential effects.

To test whether firm size is making a difference merely on cost of borrowing, I adopt the first sensitivity test model by including firm size proxy as an independent variable in the Model 3, and Model 4 is shown as follow:

Model 4:

Cost of Debt_- = α + β₁BLOCK_-+ β₆INDB_-+ β_< BLOCK_-, INDB_- + β_>FSIZE_-+ β_BZ − score_-+ β_GTANGI_-+ 𝛽_J𝑃/𝐵_S+βMLEV-+ βNAGE-+ β1TBDSIZE-+ ε

This regression is run on total sample without discrimination of large or small firms. And the results are shown in Table 4, while they do not indicate a significant role that firm size is playing in the moderation model (with a t-value of 0.76). Besides, the

(31)

moderating effect between independent directors and outside blockholders is still not distinct in Model 4.

To further extend the analysis, I test total sample on Model 5. I introduce a dummy variable D to test the potential differences between large and small firms (defined in the following paragraphs), I set D = 1 if the firm size is large and D = 0 when firm size is small, D measures the dummy intercept in the regression results. Two new variables BLOCK*D and INDB*D are introduced to test if there is a difference between large and small firms in corporate governance attributes’ impact on firms’ cost of long-term debt.

Model 5:

Cost of Debt_- = α + 𝛽₁BLOCK_-+ 𝛽₆INDB_-+ 𝛽_< 𝐵𝐿𝑂𝐶𝐾_S, 𝐷_S + 𝛽_> 𝐼𝑁𝐷𝐵_S, 𝐷_S + 𝛽B𝐷S+ 𝛽G𝑍 − 𝑠𝑐𝑜𝑟𝑒S + 𝛽J𝑇𝐴𝑁𝐺𝐼S+ 𝛽M𝑃/𝐵S + 𝛽N𝐿𝐸𝑉S+ 𝛽1T𝐴𝐺𝐸S+𝛽11𝐵𝐷𝑆𝐼𝑍𝐸S+𝜀

The dummy variable is adopted to differentiate large firms from small firm. To start with, the population is ranked in a descending order, and median value is used as

criterion. The firm- years with a firm size proxy higher than zero and lower than median are marked as 0, and those observations with higher than median value are marked as 1. Among the 731 observations, 366 firm-years (top 50% items) are included in large firms grouping and the remaining 365 observations are put in “small firms” counterpart.

The two subgroups main characteristics are summarized as follow:

large firms grouping small firms grouping

Median Mean Sdv Median Mean Sdv

LTcod 0.090 0.125 0.137 0.085 0.112 0.118

outsideBH 0.140 0.178 0.147 0.201 0.215 0.134

ID 0.667 0.627 0.192 0.625 0.596 0.184

Firm size 3.669 3.742 0.394 2.811 2.782 0.343

The regression results of Model 4 and Model 5 are shown in Table 4, together with ID and its moderating results.

(32)

After interacting a dummy variable, the strongly significant positive relationship between ID and cost of long-term debt is observed for the first time (coefficient = 0.074, t = 2.07). And after the interaction between Dummy and ID, a significant role the firm size is playing in the correlation is observed. With a coefficient of 0.12 and t-value of 2.46, I can infer that creditor may perceive independent directors’ positive impact on corporate governance in big firms to be insignificant. The higher level of board

independence, the more interest premiums are going to be charged in large firms. This results seem to contradict prior research results which claim that large firms should be equipped with stronger corporate governance system and high percentage of

independent directors in the board is a strong indicator of effective corporate

governance. The strong positive correlation in Model 5 together with the insignificant moderating effect of independent directors in Model 3 may indicate that the existence and percentage of independent directors cannot be seen as an important proxy of corporate governances.

The interaction of BH and D fails to find a strong influential impact that big firms make(t=0.67), though BH and LTcod are still significantly negatively correlated (t=1.92). As I have demonstrated already with model 1, outside blockholders do enhance corporate governance via “shared-benefit channel”, and large firms with amplified governance quality should face lower cost of borrowing. However, the result in Model 5 indicates that outside blockholders are playing a monitoring role in either large or small firms without discrimination on firm size. Their independence and bonded interests may be the major factors that outside blockholders can exert effective influence on corporate governance and it also explain why creditors allocate monitoring role to blockholders when creditors have little motivations to do so, but firm size makes no difference in the regressions.

Overall, the additional sensitivity test of firm size on the two presumed corporate governance attributes imply that the negatively significant correlation between BH and Ltcod is hardly altered with different firm sizes. And the sensitivity test further question whether independent directors should be seen as an influential protector of shareholders’ welfare.

(33)

Overall result summary

Model 1 ( Hypothesis 1)

Percentage of outside blockolders are significantly negatively related to cost of long-term debt borrowing, while the relation between total blockholding and LTcod is not statistically

significant. Compared to inside blockholders, outside blockholders are assumed to induce lower cost of external capital through their positive impact on corporate governance

The relationship between percentage of independent directors on board and cost of debt measurements (cost of total borrowing and cost of long-term debt) is not statistically significant.

The moderating effect of independent directors on the relation between outside blockholders and cost of long-term borrowing is not empirically significant.

Model 4

( sensitivity test)

Sole effect of firm size on cost of borrowing is empirically insignificant.

Model 5

( sensitivity test) Firm size does not make significant impact on the relationship between blockholders and cost of debt. While in large firms, the percentage of independent directors are positively related to long-term borrowing, indicating that creditors perceive large firms with more independent directors to be in higher level of default risk. This regression results together with Model 2 results cast doubt on whether “board independency” can be taken as a corporate governance quality proxy.

5. Conclusions

This study aims to investigate the potential relationship between cost of borrowing and two commonly used corporate governance attributes- blockholders’ ownership and board independence. I restrict samples within time period of 1996 to 1998 and study S&P 500 blockholding firms. Instead of using a direct measurement Governance- Index to capture level of corporate governance quality, I bring a proxy of creditors’ evaluation on default risk – “cost of borrowing” as a dependent variable in my three main models. The core findings are as follow: (1) The independency of blockholders play an

important role in their monitoring decisions. Outside blockholders have less access to private information, thus they have fewer chances to tunnel corporate resources away from the firms for private use. Thus the statistically negative association between outside blockholding and cost of long-term borrowing indicates that creditors may

How do independent directors moderate the relation between blockholders and cost of debt : evidence from U.S. S&P 500 blockholding firms