• unlimited access with print and download
    $ 37 00
  • read full document, no print or download, expires after 72 hours
    $ 4 99
More info
Unlimited access including download and printing, plus availability for reading and annotating in your in your Udini library.
  • Access to this article in your Udini library for 72 hours from purchase.
  • The article will not be available for download or print.
  • Upgrade to the full version of this document at a reduced price.
  • Your trial access payment is credited when purchasing the full version.
Buy
Continue searching

Essays on borrowing, monitoring, and the cost of debt

Dissertation
Author: Sangshin Pae
Abstract:
The two essays in this dissertation study issues related to debt contracting. The first essay examines whether aspects of debt contracting have been affected by the provisions of the Sarbanes-Oxley Act of 2002 that increased monitoring of management's activities by independent directors, auditors and regulators. Using a sample of 4,610 new private debt contracts, I document a significant decrease in the cost of debt after the implementation of Sarbanes-Oxley, after controlling for other influencing factors. I also show that small firms and growth firms experienced a relatively greater reduction in the cost of debt with the increase in monitoring, consistent with greater levels of information asymmetry or uncertainty associated with these firms and thus with Sarbanes-Oxley having a greater impact on them. Overall, the findings in this study suggest that increased monitoring induced by Sarbanes-Oxley had a significant impact on contracts in the private debt market. The second essay examines the relationship between borrower's choice of debt financing (public versus private) and its choice of earnings management tools (accrual-based versus real activities). Based on a sample of public and private debt issuers from 1992 through 2002, I document that public debt issuers increase their accruals prior to issue then decrease their accruals subsequent to issue year while private issuers do not show specific pattern. In addition, I find evidence that both public and private debt issuers engage in real earnings management. However, the results suggest that private debt issuers manipulate earnings via real actions more heavily than public debt issuers. Overall, the findings in this study suggest that public debt issuers prefer to manipulate earnings through income increasing accruals while private debt issuers prefer to manage earnings through real activities.

iv TABLE OF CONTENTS Page LIST OF TABLES ................................................................................................. v LIST OF FIGURES ...............................................................................................vi ABSTRACT ......................................................................................................... vii CHAPTER 1. INTRODUCTION ............................................................................ 1 CHAPTER 2. THE IMPACT OF THE SARBANES-OXLEY ACT ON PRIVATE DEBT CONTRACTING ......................................................................................... 3 2.1. Introduction ................................................................................................. 3 2.2. Hypotheses Development and Empirical Models ....................................... 6 2.3. Sample selection ...................................................................................... 15 2.4. Descriptive statistics ................................................................................. 16 2.5. Results...................................................................................................... 18 2.5.1. Correlation Analysis ............................................................................ 18 2.5.2. Logistic and Multiple Regression Results ........................................... 19 2.5.3. Instrumental Variable (IV) Approach ................................................... 22 2.6. Conclusion ................................................................................................ 25 CHAPTER 3. EARNINGS MANAGEMENT AND THE ISSUANCE OF PRIVATE AND PUBLIC DEBT ........................................................................................... 27 3.1. Introduction ............................................................................................... 27 3.2. Literature Review ...................................................................................... 30 3.2.1. Earnings Management ....................................................................... 30 3.2.2. Private versus Public Debt .................................................................. 33 3.3. Hypotheses Development ......................................................................... 35 3.4. Sample selection ...................................................................................... 40 3.5. Research Design ...................................................................................... 41 3.5.1. Accrual-Based Measure ..................................................................... 41 3.5.2. Real Activities Measure ...................................................................... 42 3.5.3. Cross-Sectional Regression Analysis ................................................. 43 3.6. Results...................................................................................................... 46 3.6.1. Descriptive Statistics .......................................................................... 46 3.6.2. Multiple Regression Results ............................................................... 47 3.6.3. Sensitivity Analysis ............................................................................. 49 3.7. Conclusion ................................................................................................ 51 CHAPTER 4. CONCLUSION .............................................................................. 52 LIST OF REFERENCES .................................................................................... 76 VITA ................................................................................................................... 83

v LIST OF TABLES Table Page Table 1 Sample selection ................................................................................... 58 Table 2.1 Descriptive statistics ........................................................................... 59 Table 2.2 Difference in characteristics between pre-SOX and post-SOX ........... 60 Table 3. Pearson correlation matrix .................................................................... 61 Table 4 Logit regression results .......................................................................... 62 Table 5 Multiple regression results ..................................................................... 63 Table 6 Multiple regression results (firms that issued both pre-SOX and post- SOX period) ........................................................................................................ 64 Table 7 Multiple regression results (small firms) ................................................. 65 Table 8 Instrumental Variable approach ............................................................. 66 Table 9.1 Mean and median discretionary total accruals around public issues .. 67 Table 9.2 Mean and median discretionary total accruals around private issues . 67 Table 10.1 Sample distribution by year .............................................................. 68 Table 10.2 Descriptive characteristics ............................................................... 68 Table 10.3 Pearson correlation matrix ................................................................ 68 Table 11 Levels of discretionary accruals ........................................................... 69 Table 12.1 Levels of abnormal CFO ................................................................... 70 Table 12.2 Levels of abnormal discretionary expenses ...................................... 70 Table 12.3 Levels of abnormal production costs ................................................ 71 Table 13.1 Changes in discretionary total accruals (public debt sample) ........... 72 Table 13.2 Changes in discretionary total accruals (private debt sample) .......... 72 Table 14.1 Changes in abnormal CFO (public debt sample) .............................. 73 Table 14.2 Changes in abnormal CFO (private debt sample) ............................ 73 Table 15.1 Changes in abnormal discretionary expenses (public debt sample) . 74 Table 15.2 Changes in abnormal discretionary expenses (private debt sample) 74 Table 16.1 Changes in abnormal production costs (public debt sample) ........... 75 Table 16.2 Changes in abnormal production costs (private debt sample) .......... 75

vi LIST OF FIGURES Figure Page Figure 1 Association among monitoring, debt covenants, and cost of debt ........ 54 Figure 2 Dow Jones Industrial Average during sample period ........................... 55 Figure 3 5-year Treasury bond rate during sample period .................................. 55 Figure 4.1 Discretionary total accruals of public debt issuers surrounding the issue year ..................................................................................................... 56 Figure 4.2 Discretionary total accruals of private debt issuers surrounding the issue year ..................................................................................................... 56 Figure 5 Time distribution ................................................................................... 57

vii ABSTRACT Pae, Sangshin. Ph.D. Purdue University, August, 2008. Essays on Borrowing, Monitoring, and the Cost of Debt. Major Professors: Mark Bagnoli, Susan Watts.

The two essays in this dissertation study issues related to debt contracting. The first essay examines whether aspects of debt contracting have been affected by the provisions of the Sarbanes-Oxley Act of 2002 that increased monitoring of management’s activities by independent directors, auditors and regulators. Using a sample of 4,610 new private debt contracts, I document a significant decrease in the cost of debt after the implementation of Sarbanes-Oxley, after controlling for other influencing factors. I also show that small firms and growth firms experienced a relatively greater reduction in the cost of debt with the increase in monitoring, consistent with greater levels of information asymmetry or uncertainty associated with these firms and thus with Sarbanes-Oxley having a greater impact on them. Overall, the findings in this study suggest that increased monitoring induced by Sarbanes-Oxley had a significant impact on contracts in the private debt market. The second essay examines the relationship between borrower’s choice of debt financing (public versus private) and its choice of earnings management tools (accrual-based versus real activities). Based on a sample of public and private debt issuers from 1992 through 2002, I document that public debt issuers increase their accruals prior to issue then decrease their accruals subsequent to issue year while private issuers do not show specific pattern. In addition, I find evidence that both public and private debt issuers engage in real earnings management. However, the results suggest that private debt issuers manipulate earnings via real actions more heavily than public debt issuers. Overall, the findings in this study suggest that public debt issuers prefer to manipulate earnings through income increasing accruals while private debt issuers prefer to manage earnings through real activities.

1 CHAPTER 1. INTRODUCTION Debt financing is a predominant source of external funding for U.S. public companies. For instance, in 2005 the total new capital raised in the private debt market was $1,500 billion and in the public debt market was about $700 billion. Given the economic importance of debt markets, however, previous accounting studies have relatively paid less attention. This dissertation studies issues pertaining to debt contracting and debt market.

The first essay studies association among three devices that debtholders use to mitigate agency costs and to protect themselves from managerial opportunistic behavior. This essay examines whether an association exists among monitoring, debt covenants, and the cost of debt and whether it has been affected by certain provisions of the Sarbanes-Oxley Act of 2002. In addition, this essay examines whether the increased monitoring induced by Sarbanes-Oxley Act differentially affected high information-asymmetry firms such as small and growth firms.

The second essay studies the motivation of earnings management prior to the issuance of debt. This essay examines whether firms engage in earnings manipulation through income increasing discretionary accruals or by real economic actions prior to the issuance of debt. In addition, this essay examines whether firms financing from public debt market manipulate their earnings differently from firms financing from private debt market.

2 Overall, the research problem of this dissertation range from the agency costs of debt to earnings management. In the rest of this dissertation, each essay is presented as a separate chapter.

3 CHAPTER 2. THE IMPACT OF THE SARBANES-OXLEY ACT ON PRIVATE DEBT CONTRACTING 2.1. Introduction

The purpose of this paper is to investigate whether an association exists among monitoring, debt covenants, and the cost of debt 1 and whether it has been affected by certain provisions of the Sarbanes-Oxley Act of 2002 (hereafter SOX). I also examine whether the increased monitoring that resulted from SOX differentially affected firms with certain characteristics, such as size or presence of growth options.

The agency costs of debt arise because of conflicts between shareholders and debtholders and are mainly due to asset substitution and underinvestment problems which occur after the debt contract is finalized. Jensen and Meckling (1976) posit that shareholders have incentives to invest in high variance projects, i.e., risky and high expected return projects, at the expense of debtholders (asset substitution). Alternatively, Myers (1977) argues that shareholders have incentive to underinvest in positive net present value (NPV) projects because the positive expected NPV fails to cover previously promised debt repayments (underinvestment). To mitigate shareholder-bondholder agency costs, bondholders use three devices – monitoring, debt covenants, and cost of debt – to protect themselves from managerial opportunism. Thus, an exogenously mandated increase in monitoring should produce less reliance on the other two methods, debt covenants and cost of debt, for reducing agency costs.

1 I use “cost of debt” and “interest rate” interchangeably in the paper.

4 Figure 1 shows three methods that lower the agency costs of debt and the relationships between the methods. In this paper, I examine whether the event of increased regulatory monitoring produced less reliance on debt covenants, the cost of debt or both. To accomplish my objective, I compare the number of debt covenants and the cost of debt for private debt contracts during the pre-SOX and post-SOX periods. Using data on private debt issued during the 1999-2005 time period (specifically a sample of 4,610 facilities), I empirically examine the association among monitoring, debt covenants, and the cost of debt. I focus on private debt because, by its nature, private debt is more likely to have restrictive covenants than public debt (Kwan and Carleton, 2004). 2 Due to the large number of bondholders involved in a public debt issue, renegotiating a debt contract following a debt covenant violation can be costly and difficult. Thus, private debt contracts generally contain a greater number of debt covenants than public debt contracts making it easier to examine the association among monitoring, debt covenants, and cost of debt in private debt contracts.

In the analysis that follows, I rely on the definition of monitoring in Jensen and Meckling (1976) 3 and the specific provisions in SOX that increase the monitoring of public companies, especially Title II “Auditor Independence,” Title III “Corporate Responsibility,” and Title IV “Enhanced Financial Disclosures.” 4

2 Although I examine private debt contracts, the companies in my sample are publicly traded firms and so their corporate governance structure and internal control systems are affected by the implementation of Sarbanes-Oxley Act. Therefore, their private debt contracts will be affected by SOX. 3 Jensen and Meckling (1976) define monitoring as more than just measuring or observing the behavior of the agent. It includes efforts on the part of the principal to “control” the behavior of the agent through various activities. I define “monitoring” in terms of broad measure which includes all internal and external monitoring that affects agent’s behavior to reduce his or her discretion. 4 The major accounting, auditing, and corporate governance provisions in SOX prohibit accounting firms from performing certain non-audit services for their audit clients, require audit committee members to be independent of management, mandate CEO and CFO certification of financial statements.

5 Previous studies on debt contracts document only the association between monitoring and debt covenants (Black et al., 2004) or the association between debt covenants and the cost of debt (Beatty et al., 2002). Black et al. (2004) find a negative relationship between monitoring and the frequency of debt covenants. In particular, they document decreases in the use of debt covenants during the periods of increased monitoring. In addition, Beatty et al. (2002) find a negative relationship between the inclusion of debt covenants and the cost of debt. They argue that managers are willing to bear higher interest rates to retain accounting flexibility. I extend the work in these two studies by evaluating a broader set of variables that are used to mitigate agency costs between shareholders and debtholders. Specifically, I hypothesize that during the period of increased monitoring induced by SOX (i.e., post-SOX period), the number of debt covenants or the cost of debt will decrease in private debt contracts due to interaction effects among monitoring, debt covenants, and cost of debt. In addition, I hypothesize that small firms and growth firms will demonstrate a much greater impact from the implementation of SOX because the Act will reduce conflicts between shareholders and debtholders more for high information- asymmetry firms.

The empirical results are generally consistent with my hypothesis. I find a statistically significant decrease in the cost of debt after the implementation of SOX. However, I do not find evidence that the usage of debt covenants decreased during the post-SOX period. The results indicate that exogeneously mandated monitoring produces less reliance on the cost of debt but not on debt covenants. In addition, I find evidence that, on average, small firms and growth firms are more influenced by increased monitoring due to SOX. These results are consistent with my hypothesis that firms with higher information asymmetry were more influenced by SOX.

6 The results demonstrated in this paper make a number of contributions to our understanding of debt contracting. First, this paper examines the association among all three devices that are used to minimize the conflicts between shareholders and debtholders. Prior research has only found evidence of an association between monitoring and debt covenants (Black et al., 2004) and an association between debt covenants and the cost of debt (Beatty et al., 2002). This paper is the first to study the relationship among all three methods. Second, while most previous studies related to SOX focused on stock price reactions and corporate governance issues 5 , this paper addresses the impact of the SOX on debt contracts. Finally, this paper provides a foundation for future work on the relations between monitoring, debt covenants, and the cost of debt.

The remainder of the paper is organized as follows. Section 2.2 develops the hypotheses and presents the empirical models used to investigate the relation between monitoring, debt covenants, and the cost of debt. Section 2.3 describes the sample, data sources, and variable measurements. Section 2.4 provides descriptive statistics. Section 2.5 presents correlation analysis and my primary results. It also examines a potential endogeneity problem and presents instrumental variable approach results. Finally, in section 2.6, I conclude. 2.2. Hypotheses Development and Empirical Models

In this section, I develop each of my four principal hypotheses based on the impact of the increased monitoring imposed by SOX and present the regression models used in empirical tests. The first two hypotheses consider the association among the three most important components of debt contracts: monitoring, debt covenants, and the cost of debt. The next two hypotheses consider whether the

5 See Berger et al. (2004), Jain and Rezaee (2004), Li et al. (2004), and Zhang (2005).

7 increased monitoring due to SOX had different effects based on the size of the firm or its growth options.

Prior empirical work on debt contracts shows that when the level of regulatory monitoring increases, banks in their role as borrowers reduce their use of debt covenants intended to reduce agency costs (Black et al., 2004). They argue that regulatory monitoring and debt covenants both strive to limit a bank’s default risk, and since regulatory monitoring cannot be controlled by bank shareholders, they try to minimize the agency costs by substituting monitoring through debt covenants where debt covenants and regulatory monitoring intersect. SOX’s provisions force increased levels of inside and outside monitoring on firms. For example, Section 202 requires that all auditing services and all permitted non- auditing services to be pre-approved by the client company’s independent audit committee, Section 302 requires each public company’s CEO and CFO to certify that they have reviewed the quarterly and annual reports their companies file with the SEC, and Section 403 requires most transactions by insiders to be electronically filed with the SEC within two business days. In the presence of this additional monitoring, lenders and borrowers may be able to reduce the number and type of covenants employed. I examine whether this substitution effect between monitoring and debt covenants can be generalized by expanding the sample of firms to include other industries besides banks, as examined in Black et al. (2004).

In addition, I develop a second hypothesis based on prior research in managerial opportunism. Jensen and Meckling (1976) suggest that covenants are included in debt contracts as a strategy for restricting managerial opportunism. They argue that by agreeing to restrict future opportunistic behavior, borrowers can reduce their current borrowing costs. Thus, the borrower faces a trade-off between retaining the possibility of future opportunistic behavior and obtaining a lower

8 interest rate. Beatty et al. (2002) find evidence that borrowers are willing to pay substantially higher interest rates to retain accounting flexibility that may help them avoid covenant violations.

Black et al. (2004) only investigate the association between monitoring and debt covenants. While prior study (Beatty et al., 2002) suggests that the cost of debt and debt covenants could be substitutes, I hypothesize that there is also a substitution effect between monitoring and the cost of debt. For example, if the cost of debt increases and the number of debt covenants decreases at the same time, we cannot attribute the decreased number of debt covenants solely to the effect of increased monitoring. 6 Thus, I expect increased monitoring to affect debt covenants, the cost of debt, or both. Because of its increased monitoring requirements, SOX should have altered the use of debt covenants and the cost of debt in post-SOX debt contracts. This leads to my first and second hypothesis (in alternative form):

H1: The number of debt covenants is likely to decrease, ceteris paribus, during the post-SOX period.

H2: The cost of debt is likely to decrease, ceteris paribus, during the post-SOX period. I test these hypotheses by estimating (1) a logistic regression model and (2) a multiple regression model in which either debt covenants or spread is the

6 Like most previous studies on debt covenants, I do not measure the tightness of debt covenants due to the cost of accessing actual debt covenant information. Prior studies use proxies such as debt-equity ratio (Duke and Hunt, 1990; Press and Weintrop, 1990; DeFond and Jiambalvo, 1991), direct measurements of covenant slack (Dichev and Skinner, 2002), or number of covenants (Black et al., 2004; Begley and Feltham, 1999) for measuring the tightness of debt covenants. Begley and Feltham (1999) find evidence that the existence and tightness of the covenants are highly positively correlated.

9 dependent variable. Ten covariates are included to control for other potentially relevant explanatory factors, discussed below. The model is as follows:

tititititi0 GROWTHVARLEVSIZESOXlogit(1]Prob[COV ,5,4,3,2,1 D D D D D D

tititititi SPREADCOLAMOUNTMATRATING ,10,9,8,7,6 DDDDD 

) ,11 ti TBOND D

(1) titititititi GROWTHVARLEVSIZESOXSPREAD ,5,4,3,2,10, EEEEEE 

tititititi COVCOLAMOUNTMATRATING ,10,9,8,7,6 E E E E E

ti TBOND ,11 E

(2)

where i refers to the facility 7 and t refers to the time of issuance. Variable definitions are as follows:

- COV is a dummy variable that is equal to 1 if the debt covenant is included in debt contract, zero otherwise; - SOX is an indicator variable, which is a proxy for monitoring that is set equal to 1 if the sample period is in the post-SOX period, zero if the sample period is in the pre-SOX period; - SIZE is the natural logarithm of the total assets of firm at quarter-end; - LEV is the ratio of long-term debt to total assets at quarter-end; - VAR is 5 years earnings variability prior to the debt contract, computed as the standard deviation of firm i’s net income before extraordinary items (scaled by total assets) measured over rolling five year windows; - GROWTH is the market-to-book ratio, computed as market value of equity divided by total book value of equity; - AMOUNT is the total amount of money (scaled by total assets) that the firm borrowed in debt contract;

7 A facility is a tranche of a private debt offer.

10 - MAT is the stated maturity calculated in years; - RATING is the Moody’s senior debt ratings information on each loan facility, ratings of Aaa, Aa, A, Baa, Ba, B, and lower than B represents 1 through 7, respectively; - COL is a dummy variable that is set equal to 1 if the loan is secured, zero otherwise; - SPREAD is the basis point spread over LIBOR 8 inclusive of all fees, which is a proxy for cost of debt. In general, this spread is fixed over the life of the loan; - TBOND is a 5 year Treasury bond rate.

The dummy variable, denoted SOX, is set equal to one for the post-SOX period and zero for the pre-SOX period to determine whether there were any changes in debt contracts before and after the implementation of the Act. I expect that this SOX variable, which is the main variable of interest, would be significantly negative due to the additional monitoring requirements associated with SOX.

As a control variable, I use natural log of total assets, denoted as SIZE, to proxy for the size of the firm. Small firms are generally viewed as more risky to creditors, so I expect this SIZE variable to have a negative relationship with the dependent variables. In addition, large firms may be better able to manage risk, so firm size may affect interest rate choice.

Interest rates on loans are extremely sensitive to default risk. As a proxy for the loan’s credit risk, I use Moody’s senior debt ratings information on each loan facility, as provided in the Dealscan database, denoted by RATING. I code RATING into 1 through 7 to represent the ratings of Aaa, Aa, A, Baa, Ba, B, and

8 London Inter-Bank Offer Rate (LIBOR) is the interest rate that the banks charge each other for loans. LIBOR is the primary benchmark for interest rates around the world. Interest rate contracts are generally settled based on LIBOR.

11 lower than B. Higher default risk (a higher value of debt rating) is related to higher interest rate and more debt covenants; therefore, I expect a positive relation between RATING and dependent variables.

Debt contracts are also affected by the leverage of the firm. Capital Structure Theory suggests that at relatively low debt levels, the probability of bankruptcy and financial distress is low and the benefit from debt outweighs the cost. However, as the debt level increases, the possibility of financial distress also increases, so the benefit from debt financing may be more than offset by the financial distress costs. Thus, the interest rate does not have a linear relationship with the firm’s leverage, making it hard to predict whether leverage will have a positive or negative relationship with dependent variables. This difficulty is compounded because different firms and industries have different optimal levels of leverage. The variable LEV is the ratio of long-term debt over total capital (debt plus equity), and I do not predict a sign for 3 D or 3 E .

I use the standard deviation of the firm’s return on assets (scaled by total assets) over the past five years, denoted by VAR, rather than the stock price variations to measure the variability of the firm’s performance. Because debtholders usually look at the firm’s earnings stability rather than the stock price movements, earnings variability is assumed to be a better proxy for measuring risk. Therefore, I expect VAR to have a positive relationship with dependent variables. Other control variables are debt characteristic variables such as the cost of debt, total amount of debt, maturity and whether the loan has collateral. Since debt covenants and cost of debt are substitutes (that is, when debt covenants increase, interest rates decrease and vice versa), I expect SPREAD to have a negative relationship with COV. The debt contract is also affected by its maturity, the total amount of the loan and whether it is secured by collateral. I expect MAT to have a positive relationship with dependent variables because longer maturity

12 debt is riskier farther in the future because of increased uncertainty of repayment. Collateral allows the lender to recover, at least partially, the principal. When the borrower fails to make a promised payment, the lender can sell the collateral, thus reducing the likelihood or amount of loss on debt. Therefore, I expect a negative relationship between COL and SPREAD because if the debt contract is secured by collateral, the cost of debt should be lower. 9

In addition, a firm’s growth opportunities are reflected in the market-to-book ratio. Growth firms have more intangible assets whose valuation depends heavily on future profitability. Therefore, I expect higher market-to-book ratio (i.e., growth firms), denoted by GROWTH, to have a positive relationship with SPREAD. To control for the effects of extreme values, I remove those observations that have negative or zero book value.

During my sample period, economic environments were unstable 10 , and to capture the macro-economic factors, I obtained five-year Treasury bond rates from the United States Department of the Treasury (http://www.treasury.gov). Figure 2 shows the Dow Jones Industrial Average (DJIA) and figure 3 shows five- year Treasury bond rates during my sample period.

Since SOX contains substantive reforms with respect to financial reporting, we would expect it to have a significant impact on reducing information asymmetry. In particular, firms that significantly manage earnings should be impacted differently than firms that do not manage earnings. In the bond market, as in the stock market, the risk premium is different depending on the level of information

9 However, a number of prior studies have found a positive relation between the existence of collateral and the cost of debt (Berger and Udell, 1990; John et al., 2003). Berger and Udell (1990) and John et al. (2003) suggest that lower quality firms are required to use collateral when issuing debt, while higher quality firms are able to issue without it. 10 The Enron scandal was revealed in late 2001.

13 asymmetry. The level of information asymmetry for small firms is more likely to be reduced than for large firms as a result of SOX. Thus, the relative reduction in the risk premium on debt contracts is likely to be greater for smaller firms, and thus, I expect a greater reduction in the cost of debt issued during the post-SOX period for smaller firms.

As just noted, prior studies show that the level of information asymmetry differs between small firms and large firms. For example, Lakonishok and Lee (2001) find that insider purchases in smaller firms predict future returns, but this predictive power does not hold for larger firms. Similarly, Finnerty (1976) and Seyhun (1986) find insider profits are larger for smaller firms. If firm size is proxy for information asymmetry and information asymmetry is greater in smaller firms, then the results of Lakonishok and Lee (2001), Seyhun (1986), and Finnerty (1976) suggest insider profits are larger when greater information asymmetry is present. A recent paper by Bharath et al. (2006) also documents that small borrowers have greater information asymmetries. In addition, Dixon et al. (2006) posit that because small businesses are likely to be less diversified and less able to leverage economies of scale or to access capital markets, the cost of complying with a particular regulation may be different for smaller and larger firms. Thus, my third hypothesis is (in alternative form):

H3: Small firms are more likely to have a greater relative reduction in the cost of debt after the passage of SOX than larger firms.

Barclay and Smith (1995) suggest that a firm’s future investment opportunities may be viewed as options whose value depends on the likelihood that the firm will exercise the options optimally. Therefore, the contracting costs due to

14 underinvestment 11 and asset substitution 12 are higher for firms with more growth options because the conflict between shareholders and bondholders over the exercise of the options is greater. Shareholders of high-growth firms can more easily substitute riskier projects for less risky ones and are also more susceptible to foregoing positive NPV projects if the gains accrue predominantly to the bondholders. That is, management may invest in high-risk negative NPV projects that increase the value of the equity but decrease the value of the debt. Consistent with my third hypothesis, if we expect the Sarbanes-Oxley Act to have a significant impact on reducing information asymmetry, then it should have a greater impact on firms that have more growth options because the conflict between shareholders and bondholders is greater for these firms. Thus, when the information asymmetry is reduced due to Sarbanes-Oxley, the risk premium of debt contracts will decrease. This leads to my fourth hypothesis (in alternative form):

Full document contains 98 pages
Abstract: The two essays in this dissertation study issues related to debt contracting. The first essay examines whether aspects of debt contracting have been affected by the provisions of the Sarbanes-Oxley Act of 2002 that increased monitoring of management's activities by independent directors, auditors and regulators. Using a sample of 4,610 new private debt contracts, I document a significant decrease in the cost of debt after the implementation of Sarbanes-Oxley, after controlling for other influencing factors. I also show that small firms and growth firms experienced a relatively greater reduction in the cost of debt with the increase in monitoring, consistent with greater levels of information asymmetry or uncertainty associated with these firms and thus with Sarbanes-Oxley having a greater impact on them. Overall, the findings in this study suggest that increased monitoring induced by Sarbanes-Oxley had a significant impact on contracts in the private debt market. The second essay examines the relationship between borrower's choice of debt financing (public versus private) and its choice of earnings management tools (accrual-based versus real activities). Based on a sample of public and private debt issuers from 1992 through 2002, I document that public debt issuers increase their accruals prior to issue then decrease their accruals subsequent to issue year while private issuers do not show specific pattern. In addition, I find evidence that both public and private debt issuers engage in real earnings management. However, the results suggest that private debt issuers manipulate earnings via real actions more heavily than public debt issuers. Overall, the findings in this study suggest that public debt issuers prefer to manipulate earnings through income increasing accruals while private debt issuers prefer to manage earnings through real activities.