The effect of accounting conservatism on the efficiency of debt contracting : a study on the effects of accounting conservatism on the renegotiation of debt agreements

(1)

Amsterdam Business School

The effect of accounting conservatism on the efficiency of debt

contracting

A study on the effects of accounting conservatism on the renegotiation of debt agreements.

Name: Ernst de Lange Student number: 10262040 Supervisor: S.W. Bissessur Date: 16 June 2016

Word count: 15233

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

(2)

Statement of Originality

This document is written by student Ernst de Lange 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

During recent years accounting conservatism has come under increased scrutiny. Though research shows many advantages of accounting conservatism, the FASB has removed accounting conservatism as a fundamental principal of accounting. This study examines whether a higher C_Score, as a proxy for accounting conservatism, has an effect on the occurrence of a decrease commitment amendment, as a proxy for debt contract renegotiation. I employ a large sample of syndicated loans and their respective

amendments, and match these to firm-year accounting conservatism scores. My results show that the likelihood of a decrease commitment amendment is not significantly related to the firm-year C_Scores. The results are robust to using different measures of accounting conservatism. I conclude with a call for additional research on debt renegotiation measures and negative effects of accounting conservatism.

Acknowledgement: I express my gratitude to Dr. Sanjay Bissessur, who provided excellent

(4)

1. Introduction

1.1 Background

Accounting conservatism is widely regarded as bad news being recognized more quickly compared to good news and that the financial statements contain a downward bias of accounting net asset value (Ruch & Taylor, 2015). Accounting conservatism (hereafter: conservatism) is seen as one of the fundamental characteristics of accounting, often being referred to as a principle of accounting (Basu, 1997).

Most literature has viewed conservatism in a positive light. Conservatism is seen to bring many benefits to several users of financial statements. For instance, interest rates for companies that exert a lot of conservatism are lower compared to peers that are less conservative in their financial reporting (Zhang J. , 2008). Also, conservatism is seen as relieving some asymmetry problems. In a debt agreement, lenders have high downward potential but no upward potential. Lenders are compensated with a fixed amount of money in return for the capital they provided to the borrower. When a borrower experiences financial difficulties, the chance of bankruptcy increases for the borrower, thus lowering the expected payoff to the lender. By exerting conservatism potential negative performance shocks for the borrower are communicated with more timeliness compared to positive performance shocks (Ruch & Taylor, 2015). Shareholders can also benefit from

conservatism. Conservatism restricts managers in their ability to use discretion to change accounting data, thus reducing information asymmetry and the associated losses. The effect of this is increased firm value, benefiting shareholders (Lafond & Watts, 2008). Another study has shown that there is a significant relation between conservatism and stock performance during the 2008 financial crisis (Francis, Hasan, & Wu, 2013).

Despite the seemingly positive opinion in much of the research, conservatism has recently come under increased scrutiny. Conservatism was previously seen as a

fundamental accounting concept that was used to prevent financial statements from being too optimistic about a firms prospects (Ruch & Taylor, 2015). The following quotes taken from FASB statements show the change in the standard setter’s opinion on conservatism:

“Conservatism in financial reporting should no longer connote deliberate, consistent understatement of net assets and profits. The Board emphasizes that point because

(6)

virtue. That notion became deeply ingrained and is still in evidence despite efforts over the past 40 years to change it. The convention of conservatism, which was once commonly expressed in the admonition to “anticipate no profits but anticipate all losses,” developed during a time when balance sheets were considered the primary (and often only) financial statement, and details of profits or other operating results were rarely provided outside business enterprises.” (Financial Accounting Standards Board, 1980)

“Chapter 3 does not include prudence or conservatism as an aspect of faithful representation because including either would be inconsistent with neutrality. Some

respondents to the Discussion Paper and Exposure Draft disagreed with that view. They said that the framework should include conservatism, prudence, or both. They said that bias should not always be assumed to be undesirable, especially in circumstances when bias, in their view, produces information that is more relevant to some users.” (Financial Accounting

Standards Board, 2010)

The two previous quotes are from so called Statements of Financial Accounting Concepts. In these statements the FASB gives guidance on how different qualitative features and purposes of financial accounting should be interpreted (Financial Accounting Standards Board, 2016). These statements show that, even though conservatism is often seen as a fundamental concept in accounting, over time conservatism has come under increased criticism. The FASB feels that conservatism could introduce an unwanted bias in financial statements which could negatively affect the neutrality of said financial statements (Ruch & Taylor, 2015). Because of this potential bias the FASB chose to leave out conservatism as a fundamental principle in their latest statement. It was this criticism of conservatism that provides the motivation for conducting this study.

As previously mentioned conservatism has recently been debated and its effect on the relevance of the financial statements is being questioned. Despite the earlier mentioned benefits of conservatism in the case of a debt arrangement, a potential negative effect of conservatism has been theorized (Gigler, Kanodia, Sapra, & Venugopalan, 2009). Gigler et al. (2009) theorize that optimal debt contracting minimizes the cost of two separate errors. The first error relates to false alarms, the violation of debt covenants and renegotiation of the debt because of conservative accounting or even rejecting debt agreements before they are entered into. The second error relates to being too optimistic in financial statements which, as previously mentioned, causes a higher chance of default on debt if the real situation for a

(7)

company is less positive than the accounting shows (Gigler, Kanodia, Sapra, & Venugopalan, 2009). The idea behind the theorized effect behind the first type errors is that too much conservatism makes the financial accounting position of a company that is presented as being significantly worse compared to its real position. This negative perspective on

performance may induce lenders to renegotiate debt contracts. This effect could affect debt agreements that are in reality still profitable (Gigler, Kanodia, Sapra, & Venugopalan, 2009).

In this paper, I hypothesize that firms who use more accounting conservatism in their financial statements experience more debt renegotiations. From the regression on a sample of 14,557 firm-year observations I find no evidence of a relationship between accounting conservatism and the renegotiation of debt contracts.

This paper has several contributions to both academic research and practice. Firstly, this paper benefits the growing literature on the effects of accounting conservatism. My results show that in accordance with most previous literature, accounting conservatism benefits seem to outweigh its weaknesses. Also, the research on measures of debt contract renegotiation is limited. During the process of writing this thesis I faced problems when attempting to find empirical measures of debt contract renegotiation. Hopefully, this paper will motivate other researchers to expand the literature on debt renegotiation measures.

There are several contributions this paper makes to different aspects of accounting practice. First and foremost, this paper was motivated by statements from standard setters and regulators who voiced the concern that accounting conservatism creates an unwanted negative bias in financial statements. The lack of evidence on one of the possible arguments against accounting conservatism shows the standard setters and regulators that using this specific effect of accounting conservatism as an argument against conservatism is unfunded. Secondly, this study contributes to practice by giving preparers evidence that accounting conservatism does not affect the efficiency of debt contracting in a way, such as suggested by Gigler et al. (2009). Thus, preparers should use caution when interpreting the statements about unwanted biases made by regulators and standard setters. Lastly, the lack of a

relationship could be an indication that lenders take accounting conservatism in

consideration when determining the appropriate action after a borrower violates their debt covenants.

The paper continues as follows. Section 2 contains a literature review of agency theory, accounting conservatism and debt contracting, and furthermore develops the

(8)

hypothesis for this study. Section 3 will discuss the research methodology, including the sample selection, operationalization of variables, and control variables. Section 4 contains the results of the study. Section 5 discusses the results and provides a conclusion to this study.

(9)

2. Literature review & hypothesis development

2.1 Agency theory

This thesis finds its foundations in the well-documented and often-studied agency theory. Agency theory was first explored in the 1960s and 1970s, as part of the theory that describes how individuals deal with risk sharing (Eisenhardt, 1989). Agency theory

transformed the view of an organization to that of a set of contracts, which determine how actors divide costs and rewards (Jensen & Meckling, 1976). All actors are assumed to have certain traits. For instance, they are assumed to be self-interested, risk averse and bounded rational. The theory’s most quoted example is that of the owners or shareholders of a company, called the principal, transferring decision rights to managers, referred to as agents (Eisenhardt, 1989). Owners can decide to transfer decision rights for a variety of reasons. The transfer of these rights leads to a situation where the managers are able to decide matters which can affect the welfare of the principal. These types of relationships are referred to as contracts. The managers and owners engage in a contract to benefit both parties. The manager agrees to perform a service on behalf of the principal and the contracts are set up to provide the most benefit. However, it is highly unlikely that the contracts can be set up in a way that guarantees that the agent will always make their decisions to benefit the principal. It is widely believed that a more realistic setting is one where both the principal and the agent are attempting to maximize their personal utility (Jensen & Meckling, 1976). This more realistic view on the ownership and managing of firms leads to several problems. Agency theory not only applies to manager-shareholder

relationships, but can be applied to numerous organizational situations. In essence agency theory can be used for all situations where there is a principal who has delegated a certain task or decision right to an agent but who have goals and risk attitudes that differ from one another (Eisenhardt, 1989). One of these situations is determining how to finance an organization (Eisenhardt, 1989; Jensen & Meckling, 1976).

2.1.1 Agency problems

Problems in the agency theory setting generally arise because of differing interests for the principal and the agent and the division of labor (Jensen & Meckling, 1976; Eisenhardt, 1989). The first problem that arises in agency theory settings relates to the amount of effort the agent will put in and the reward he receives for his efforts. This so called agency

(10)

problem is a two part problem. The first issue is that the principal and the agent have different preferences when it comes to the outcome of the contract. The fact that it is often difficult if not impossible for the principal to fully monitor the agent’s efforts is the second factor to contribute to the problem (Eisenhardt, 1989). The monitoring problem stems from the information asymmetry that is present in principal-agent relationships. Agents usually have additional information about their efforts and the effects of their efforts compared to principals (Lafond & Watts, 2008). The principal has no way to verify that the agent acted in the interest of the principal (Eisenhardt, 1989). When the interests of principals and agents differ, conflicts of interest exist because the principal and agent have differing payoffs for different situations. The agent is usually bound to receive a certain reward for performing at a certain level. This reward can be fixed or variable but the reward is usually capped at some level of performance. If the agent is trying to maximize his own payoff he will halt his efforts after this performance level is reached. If the agent were to continue to exert effort after this level the agent would receive the same level of reward for a higher effort. Provided the agent sees his effort as costs of gaining rewards, he will not put in more effort than needed to reach the required performance (Jensen & Meckling, 1976). The second problem that arises in agency theory is the risk sharing problem when the risk appetite or tolerance of agents and principals differ (Eisenhardt, 1989). This problem can be prevalent when for instance, the agent is close to achieving their performance target at the end of the year. The agent may feel an incentive to take excessive risks to achieve his performance target. If the agent does not take the excessive risks he will most likely fail to achieve his target anyhow and thus he might take additional risks, even though these risky projects or decisions could harm the principal’s welfare. Specific agency problems related to private debt will be discussed in later sections and form the focal point for this study.

2.1.2 Positivist agency theory

Agency theory has been developed from two perspectives. The first perspective is the positivist agency theory and the second is the principal-agent perspective. The two

perspectives share their core principles but differ on other facets, like style, mathematical thoroughness, and dependent variable (Eisenhardt, 1989).

Positivist agency theory is mainly occupied with reaching solutions for the standard agency problem in the owner-manager relationship. This focus is most clearly described by

(11)

two propositions that are found in the positivist agency theory field. The first proposition relates to the focus of contracts. Contracts can be focused on behavior or outcomes. Contracts based on behavior lead to lower freedom and the agent will focus mostly on showing the required behavior but is not particularly focused on the outcome of his or her behavior. This leads to differing goals for the principal and the agent. The principal is

attempting to reach certain outcomes while the agent is less concerned about the outcome but more about the behavior. Outcome based contracts are thought to align the interests of the principal and the agent better than behavior focused contracts. By basing a reward for a manager on the revenue he achieves the contract will benefit the principal better compared to a reward based on how the manager behaves in achieving outcomes. Outcome based contracts are thought to be an efficient measure to prevent the agent from being optimistic about their own performance. Secondly, the positivist agency theory proposes that

information is a commodity that can prevent the agent from being unjustly optimistic about his or her performance (Eisenhardt, 1989). By obtaining more information about the agent’s behavior, the principal reduces the information asymmetry between the agent and the principal (Lafond & Watts, 2008).

2.1.3 Principal-agent study

Unlike the positivist agency theory, principal-agent studies focus more on a holistic theory of agency that can be used in a wide variety of situation, ranging from shareholder-manager to insurer-insured relationships (Harris & Raviv, 1978). Compared to the positivist agency theory, principal-agent research has had less attention from researchers. Another difference is that while positivist research is trying to find solutions for agency problems, principal-agent studies are more concerned with finding which solutions are more effective

(Eisenhardt, 1989). The principal-agent literature refers to two specific problems that arise when the principal and agent have differing goals and the principal is unable to verify the behavior of the agent. The first issue is the issue of moral hazard. Moral hazard refers to the situation where the agent does not put in any effort because he wishes to minimize his efforts and maximize his utility, the agent is also aware that the principal cannot verify the effort he put in. The second issue that is referred to in the literature is the issue of adverse selection. This situation exists because information asymmetry exists at the inception of the contract. An example is a newly hired employee who most likely has more information

(12)

about his ability than his new employer and can benefit greatly from this informational advantage (Armstrong, Guay, & Weber, 2010). Both the positivist and the principal-agent theory are involved in finding measures that relieve agency problems, some of these measures will be discussed in the next section.

2.1.4 Measures to reduce agency problems

In order to motivate the agent to act in the best interest of the principal, there are a number of measures to limit the behavior that can be detrimental to the welfare of the principal. One of these measures is corporate governance (Ahmed & Duellman, Accounting

conservatism and board of director characteristics: An empirical analysis, 2007). Corporate governance entails all measures that are used by a company to guarantee that the assets a company holds are used in an efficient way, which will provide capital providers with a healthy return on the capital they provide. This prevents managers from using assets to their personal advantage or the advantage of a third party. Thus, corporate governance relieves agency problems through better monitoring of management (Lara, Osma, & Penalva, 2009). As previously mentioned, agency theory is not only applicable in

shareholder-manager situations but can be applied in many different situations. This also applies to corporate governance. Corporate governance can be used as broadly as needed in order to better align the interest of principals and agents (Armstrong, Guay, & Weber, 2010).

Accounting conservatism is another measure to align the behavior of the agent and the goals of the principal (Lara, Osma, & Penalva, 2009). Conservatism limits manager’s opportunistic behavior by restricting payments to themselves or third parties and reduces the problems that stem from investing decisions, increases the efficiency of contracts involving debt, increasing monitoring which reduces information asymmetry and reducing litigation costs (Watts, 2003a).

2.2 Conservatism

The notion of conservatism is widely described as recognizing losses and expenses immediately and delaying the recognition of gains and revenues until they can be determined with a greater level of certainty compared to the losses and expenses. This difference of timing in recognizing amounts leads to net assets being understated (Basu, 1997; Lafond & Watts, 2008; Khan & Watts, 2009; Gigler et al., 2009; Ahmed & Duellman,

(13)

2007). This definition of conservatism stems from the study by Basu in which an attempt was made to measure the level of conservatism by measuring how much faster the losses and expenses (referred to as ‘bad news’) are recognized compared to gains and revenues (referred to as ‘good news’). The paper explains conservatism with the example of

depreciating an operating asset. If the remaining estimate of useful life changes the asset’s present value changes. Depending on the direction of the change the treatment differs because of conservatism. If the asset’s useful life increases because the asset is expected to function at the desired level for a longer time period, the gain is spread out over the

remaining life of the asset. The amount that is to be depreciated is now divided by a longer time period so each year the company would recognize a lower depreciation amount which results in higher year profits. This could be classified as good news. Due to conservatism the firm cannot elect to recognize the difference between the old and new depreciation

expenses for the remaining years in the current years. This would give managers a great way to increase current year performance when they desire higher profits in the current year. Also, since this change in estimated useful life was unexpected (i.e. it was not expected at the purchase of the asset) the expectation might have only changed temporarily and could return to the old estimate in the near future. On the other hand we have a situation where an asset is expected to be usable for a shorter time period than was initially expected. This situation is referred to as a bad news situation. Conservatism dictates that in this situation the firm should take a one-time loss on the asset to compensate for the shorter life

expectancy. Because of this difference in treatment, earnings respond to bad news in a timelier manner compared to good news (Basu, 1997). An example of conservatism in a non-financial situation is when a mother tells her child to bring a jacket outside because it might rain later, just in case it might rain later.

Another interesting finding in the study by Basu is that the market reacts more strongly to unexpected good news compared to bad news. The explanation behind this is that investors in the market know that companies are often conservative and thus when a company reports unexpected good news this probably has a more significant meaning than unexpected bad news (Basu, 1997). This finding could possibly be an argument against finding a relationship between conservatism and the renegotiation of debt. One could assume that the lenders know that companies are often conservative and take this into account when evaluating debt agreements.

(14)

2.2.1 Unconditional conservatism

Research has often attempted to split conservatism into unconditional and conditional conservatism.

Unconditional conservatism are accounting policies that have been adopted previous to news events that warrant the use of conservatism. In other words, conservatism that was agreed upon at the start of recording accounting data and policies that are being

implemented regardless of unexpected news (Beaver & Ryan, 2005). Examples can be found in corporate policy, for example companies might prefer to immediately expense all R&D expenses because the outcomes of the research can be uncertain (Gigler, Kanodia, Sapra, & Venugopalan, 2009). They can also be found in regulation or standard setting principles. For instance, revenues should only be recorded when it is highly likely that these funds will flow to the company. Much of the issues that the FASB has with conservatism most likely stems from unconditional conservatism because this introduces a systematic bias in accounting data (Beaver & Ryan, 2005). Unconditional conservatism is sometimes thought of as having no effect on economic outcomes because seeing as how it is systematically applied, users of financial statements can quite easily reverse these changes in the accounting data to arrive at values that more accurately reflect the true values. However, as many studies show; users of financial statements have only limited abilities to uncover these policies and often myopically focus on bottom-line data because of their limited focus (Ball & Brown, 1968; Zhang Y., 2008). Studies on the subject of accounting conservatism specifically also agree with the view that unconditional conservatism can affect accounting information and should not be ignored (Gigler, Kanodia, Sapra, & Venugopalan, 2009).

2.2.2 Conditional conservatism

Conditional conservatism is associated with choices that are being made following

unexpected news. The previously mentioned example of a change in the estimate of useful life for an asset is a good example of this (Basu, 1997). When the news reaches

management about a decrease of remaining useful life they have to decide for that specific situation how they would like to recognize this potential loss. The context of the news will determine the appropriate course of action for the managers. If the news is for instance given by only one report in a series of multiple reports that could possibly contradict this report management could choose not to immediately react by taking a loss on the asset but

(15)

rather wait until the remaining useful life can be determined. Being conditionally

conservative means that when negative unexpected news reaches management, a loss is immediately taken to reflect the change in circumstances. At the same time when

unexpected positive news is brought to attention, a lot more verification is needed before a gain can be recognized. Examples of conditional conservatism include the writing down of inventory to reflect lower market pricing, impairment of operating assets to reflect a faster than expected change in technology or the write down of goodwill to reflect lower acquired brand values (Beaver & Ryan, 2005).

2.2.3 Role of conservatism

Conservatism serves many roles. The most common explanations include: reducing

information asymmetry problems in contracting, lowering the chance and costs of litigation, lowering taxation, and adhering to regulation (Watts, 2003a).

Conservatism is often said to have an impact on contracting efficiencies, which are influenced by information asymmetry. What is meant by this is that firms often have lower information compared to specific actors within the firm, which can for instance create moral hazard problems. A much used example of this is when a firm employs a conservative

outlook on their results to prevent managers from using their discretion to increase their performance based on these results. Managers feel the incentive to deliver high results because often they are rewarded and penalized based on their results (Watts, 2003a). Another potential explanation for conservatism could lie in the realm of litigation. Shareholder litigation is a serious problem, especially in countries that adopt a shareholder model and often have a common-law system. An example could the United States of America (Ball, Robin, & Wu, 2003). When firms are conservative they understate their net assets, by doing this they reduce the costs that they potentially face when they are sued as well as lowering the chance of being sued (Watts, 2003a). That potential litigation costs of shareholder lawsuits can be high is clear from the many cases in which companies that faced these lawsuits were convicted of paying large fines (Frayter, 2008; Fisher, 2015; Wallace 2015). When firms have been conservative the effects of negative news is likely to be less (Basu, 1997). This results in a lower chance of litigation and lower expected costs of litigation (Watts, 2003a).

(16)

assets and reduce their current year tax earnings. By doing this they reduce the effective tax rate that companies have to pay. By being conservative it is also possible to delay the

payment of due taxes. If all losses are taken immediately but gains are deferred to future periods the taxes that have to be paid on the gains are delayed to a later period. This is another explanation for conservatism (Watts, 2003a).

Lastly, conservatism could also be promoted by regulatory and standard setting authorities. These authorities have a potential to face criticism when they allow companies to systematically overstate they results. The authorities would prefer to minimize these political costs and hence they could benefit from promoting conservatism (Watts, 2003a). Recently conservatism has been under increased scrutiny from regulators and standard setters, who believe conservatism may bias financial statements (Financial Accounting Standards Board, 2010).

2.2.4 Users of financial statements affected by conservatism

Different users of financial statements are likely to be affected by conservatism in different ways. The following section describes the different users of financial statements that are expected to be affected by conservatism.

The first set of users are related to the equity market. Users include investors and analysts. Equity market users are likely to require information that is relevant to their decision making. For investors and analysts this means that the data in the financial

statements is a true representation of the underlying value of the firm. When financial data has a strong relation to market performance equity users are likely to benefit. The two previously mentioned forms of conservatism affect the relation in different ways. Firstly, conditional conservatism benefits the relation when news is bad and leads to a lower relation when a firm receives good news (Ruch & Taylor, 2015). This is most easily

understood because unconditional conservatism increases the recognition speed of losses and increases the recognition speed of gains (Basu, 1997). When a firm recognizes losses quicker they are more responsive to market changes, increasing the degree to which financial data reports an accurate value of the underlying firm value (Ruch & Taylor, 2015). Secondly, unconditional conservatism generally leads to a lower association between true firm value and the value reported in financial statements. By indiscriminately expensing all R&D expenses the firm completely ignores the fact that some of the R&D could lead to new

(17)

technological insights that could benefit the firm financially in the future (Ruch & Taylor, 2015).

The second group of users affected are corporate governance users. The main focus of corporate governance users is to use financial data to avoid agency costs caused by observing managers due to information asymmetry. By being conservative firms can prevent that managers are compensated too much for performance they did not really achieve from doing a good job but rather from using their discretion to artificially inflate their

performance. Conservative estimates of performance also motivates managers to work hard to achieve sustainable performance. Managers feel the pressure to make investment

choices that will result in verifiable positive results and are encouraged to drop bad projects (Ruch & Taylor, 2015).

The last group consists of debt market users. The users in this group can be either lenders or borrowers. This group has similar requirements as equity market users although in a slightly different form. Relevance to their decision making is the main aim of

information for this group like equity market users. Providing conservative estimates is thought to benefit these users because it lowers the chance they invest in companies that will later on default on their loans (Ruch & Taylor, 2015). Lenders usually only get a fixed amount of interest and repayment of the initial loan regardless of how good the company they provided loans to performs. Hence, they care about getting their interest payments and their initial investments back at the end of the loan period. Their main concern is that the company does not unexpectedly defaults on the loan. Because of this they are

interested in the lower bound of the value or a conservative estimate of the company that they provide capital to (Watts, 2003a). Since lenders are interested in the absolute lower bound of the value their informational needs are met when a company uses unconditional conservatism. By providing lenders with enough assurance that even when everything goes wrong the lenders can still count on their required payments, lenders charge lower interest rates. By being conservative debt agreements are said to be more efficient (Ruch & Taylor, 2015).

2.3 Debt contracting

Capital for companies arises from two distinct sources, equity is the first source and debt is the second. By incurring more debt a company increases its leverage. By increasing this

(18)

leverage shareholders can potentially increase the upward potential payoff on their investments while limiting their downward potential to their initial investment. When a company choses to engage in this form of capital they are referred to as the borrower in the debt agreement. The capital providers who provide the funds in return for a fixed amount of interest and repayment of the initial loan are referred to as the lenders in debt agreements. As opposed to the previously mentioned limit on the downward potential for shareholders from the borrower, lenders have a limited upward potential. This is one of the reasons why lenders require interest on the initial loan to compensate them for this additional risk they incur. Increasing the interest rate compensates the lender for potential agency problems. Interest rates can increase or decrease depending on the assumed risk the lender accepts (Bharath, Sunder, & Sunder, 2008). Lenders decide on debt contracts based on information. This information often comes from financial statements and is more relevant to lenders when it increases their ability to make informed decisions (Ruch & Taylor, 2015).

When acquiring debt capital, borrowers usually have two options at their disposal. The first is public debt and the second type of debt is classified as private debt. Debt is the primary way of external funding in the U.S., with the market for syndicated loans being roughly $1.500 billion in 2005. The corporate bond market was valued at roughly $700 billion during the same year (Bharath, Sunder, & Sunder, 2008). The different types of debt will be discussed in the following section.

2.3.1 Public and private debt

Public debt is mostly made up of corporate bonds. These bonds are sold at an arm’s length transaction from the corporation to the lender. The corporation announces their wish to attract more capital by selling bonds on the open market. They usually offer an interest rate and capital providers can choose to buy the bonds if they feel the interest properly

compensates them for the perceived risk on the bond. If the perceived risk is considered too high, the lenders will not purchase the bonds and the borrower would have to increase the interest rate on the bonds until they are able to attract the desired level of capital. Public lenders differ from private lenders in that they usually will not directly contact the borrower to discuss the terms of the debt agreement. These lenders are often bond holders who will only have the option to buy the bonds or to decide to not lend their capital to the borrower. This in turn will determine the price for the bond. The price is the only way to negotiate on

(19)

the bonds for the lenders (Bharath, Sunder, & Sunder, 2008).

Private debt is debt where the lenders are made up of one or a small group of capital providers. Private lenders directly lend money to a borrower and engage in direct

negotiations between the lender and the borrower. Examples include bank loans and syndicated loans. The borrower will have direct contact with a bank in order to secure a loan. They will receive a proposal, with the terms and conditions of the loan, from the bank and most likely reply with a counter offer. The negotiations on private debt can include: Interest, maturity, debt covenants, and collateral among others. Through negotiation they will eventually reach a consensus or not (Bharath, Sunder, & Sunder, 2008). As mentioned before, loans can be between a borrower and one or multiple lenders. Loans that are agreed upon with multiple lenders and one borrower are called syndicated loans. The market for syndicated loans makes up the majority of the financing market for U.S. markets and its total value had increased from $8 billion in 1991 to $144.6 billion for the year 2003 (Wittenberg-Moerman, The role of information asymmetry and financial reporting quality in debt trading: Evidence from the secondary loan market, 2008). In 2005 it is estimated that the total syndicated loan market had a value of $1.500 billion (Bharath, Sunder, & Sunder, 2008). Private lenders have more information due to their direct communication and involvement with the borrower (Bhattacharya & Chisea, 1995).

As one can imagine the power to negotiate the terms for the debt agreement differs between the two types of debt agreements. These differences have been confirmed by several studies. Private lenders are better able to negotiate terms compared to public lenders, who only have the option to collectively decide not to lend any money. By rejecting or accepting the bond interest rate and maturity the lenders determine the market price for the bonds. This will in turn increase the interest rate the borrower has to pay on the bonds until an efficient point is reached where the bonds are purchased by the public lenders (Bharath, Sunder, & Sunder, 2008; Berlin, & Mester, 1992). In public debt agreements there is no direct negotiation between borrowers and lenders (Bharath, Sunder, & Sunder, 2008). With private debt agreements, lenders are able to communicate directly with the borrower and propose the terms and conditions under which they would accept the loan agreement. Lenders in private debt agreements have better information processing capabilities due to their increased involvement and focus on the borrower compared to bond holders (Bharath, Sunder, & Sunder, 2008). Also, lenders in private debt receive private information in the

(20)

form of financial disclosures, covenant compliance, amendment desires, financial forecasts, and acquisition or disposition plans (Wittenberg-Moerman, The role of information

asymmetry and financial reporting quality in debt trading: Evidence from the secondary loan market, 2008).

2.3.2 Debt contracting efficiency

As mentioned previously, debt agreements include an interest rate because lenders require to be compensated partially for the information asymmetry they accept when entering into a debt agreement (Bharath, Sunder, & Sunder, 2008). The information asymmetry present in debt agreements leads to specific agency problems between borrowers and lenders (Ahmed, Billings, Morton, & Stanford-Harris, 2002; Jensen & Meckling, 1976;

Gopalakrishnan & Parkash, 1995). An often quoted measure for the efficiency of debt contracts is the interest spread. This is the premium the lenders require above the risk free interest rate. When the information asymmetry is reduced, interest rates are also reduced. The costs of acquiring the capital go down for the borrower. This is often seen as an

increase of the efficiency of debt contracts (Ruch & Taylor, 2015). Some studies have challenged this assertion. One study in particular states that the efficiency of debt

contracting is not only dependent on a reduction of the borrowing costs, as measured by a reduction of interest rates. Instead, it claims efficient debt contracts reduce the costs of two different errors. The first error relates to costs that arise because of unjustifiable optimism in financial statements. This causes lenders to get a biased positive view of the borrower. This can cause lenders to enter into loans that are in reality not profitable to the lenders because of a higher underlying default rate than initially expected (Gigler, Kanodia, Sapra, & Venugopalan, 2009). The costs of this type of error can be reduced by being conservative in the financial statements, an effect which is confirmed by many studies (Zhang J., 2008; Ahmed, Billings, Morton, & Stanford-Harris, 2002). The second type error relates to costs that are incurred when capital providers falsely think that debt agreements are unprofitable due to an increased default risk. These false alarms are the other side of the conservatism coin as proposed by Gigler et al. The efficiency of debt contracting is only enhanced when conservatism decreases the total costs of the two types of errors (Gigler, Kanodia, Sapra, & Venugopalan, 2009). Regulators have also echoed the idea that conservatism could

(21)

decrease the efficiency of debt contracts due to the bias it introduces in financial information (Barker, 2015).

2.3.3 Debt renegotiation

Public lenders usually buy corporate bonds at a set interest percentage. This percentage is thought to compensate them for the risk of default on the loan and to compensate them fairly for providing the capital (Bharath, Sunder, & Sunder, 2008).

As mentioned before, negotiations on debt contracts in a private debt setting can determine the terms and conditions on a great variety of factors. When private lenders are unsure that borrowers will remain in the same financial condition that they are in at the conception of the loan, they can choose to include debt covenants in the debt agreement. Covenants are used to reduce information asymmetry problems that often plague debt agreements. Debt covenants transfer decision rights to debt holders when they are violated. When a borrower violates a debt covenant, the covenant often determines that the lender will be allocated the decision to terminate or continue the debt agreement (Gigler, Kanodia, Sapra, & Venugopalan, 2009). This reallocation of decision rights gives the lender the power to enter into renegotiations in order to reassess the profitability of the debt agreement and adjust the terms of the loan in order to make it profitable if the current profitability is questioned. These adjustments can take many forms (Dichev & Skinner, 2002; Chen & Wei, 1993; Gopalakrishnan & Parkash, 1995). The study by Gopalakrishnan & Parkash (1995) has identified six possible renegotiation responses to debt covenant violations. First, the lender can choose to terminate the loan. Second, they can demand immediate repayment of the loan. Third, lenders can opt to increase the collateral required from the borrower. Fourth, lenders can increase the interest rate. Fifth, they can impose additional constraints in the form of more covenants. Sixth and lastly, they can choose to waive the violation and allow the loan to continue as it was before the violation. The different types of debt renegotiation can thus be an indication that decision rights were transferred to the lenders.

2.4 Prior literature

Prior literature has devoted considerable effort to studying conservatism. However, debt renegotiation and its link with conservatism has not been studied in great detail yet. This study is, to my knowledge, the first study that will attempt to study the link between

(22)

this study of great contribution. Since the previous research on the subject is limited, this study provides a starting point upon which to base further research. Previous research has expanded the knowledge of the relationship between conservatism and information asymmetry in debt contracts. From these studies, knowledge arises that can be applied to the renegotiation of debt agreements. An overview of this research will be given here. One researcher that has focused on the information asymmetry in debt contracts and the link with accounting conservatism is Regina Wittenberg-Moerman. In one of her papers she focuses on the question which company and debt agreement characteristics increase or decrease the information asymmetry in private debt trading. She hypothesizes that certain firm and debt agreement characteristics will lead to a higher bid-ask spread in private debt agreements which is a proxy for higher information asymmetry. The firm and loan characteristics that she identifies and indeed finds evidence for that they decrease the bid-ask spread are when a firm reports publicly, the availability of a credit rating on the loan, for profit borrowers, more reputable arrangers. In addition she finds that conservatism is associated with a lower bid-ask spread. Revolving loans, distressed loans and loans that institutional investors have issued are associated with higher bid-ask spreads (Wittenberg-Moerman, The role of information asymmetry and financial reporting quality in debt trading: Evidence from the secondary loan market, 2008).

In another paper she hypothesizes and finds that previous bid-ask spreads on syndicated loans have a positive impact on current bid-ask spreads. Also, she finds that longer maturity syndicated loans have a positive relation with bid-ask spreads. This further indicates the influence of information asymmetry on debt agreements

(Wittenberg-Moerman, The Impact of Information Asymmetry on Debt Pricing and Maturity, 2005). The role of conservatism on bid-ask spreads is also found in the opposite direction. One study found that when banks are more conditionally conservative they charge a higher bid-ask spreads on the loans they enter into. This implies that banks who are more cautious will charge their borrowers higher rates on their loans compared to less cautious banks (Lim, Lee, Kausar, & Walker, 2014).

In their study Ahmed et al. (2002) examine the relationship between conservative accounting, bondholder-shareholder conflicts and reduction of debt costs. The find that conservative accounting can mitigate bondholder-shareholder conflicts over dividend payouts. Secondly, they find that conservative accounting is associated with a reduction in

(23)

debt costs.

The contracting benefits of conservatism are also discussed in a paper by Zhang J. (2008). In this paper Zhang J. examines whether conservatism benefits both lenders and borrowers in debt agreements. The paper concludes that indeed both parties can benefit from accounting conservatism. Zhang J. also finds that firms that are more conservative will violate debt covenants more often, in the case of a negative price shock. This is reflected in lower initial interest rates for borrowers and with timelier signaling of default risks through debt covenant violations for lenders (Zhang J. , 2008).

Accounting conservatism has often been seen as the result of debtholders’ demands. Using an experiment where fiduciary duties for managers of near insolvent firms were extended to creditors via a court ruling, it is shown that firms did increase the conservative nature of their accounting. This suggests that debtholder demands do play a key role in determining the level of accounting conservatism (Aier, Chen, & Pevzner, 2014).

Confirmation for this is also found in another study which examined the occurrence of conservative debt contract modifications and the relationship with agency costs. What they find is that when other demand sources of conservative accounting are decreased and agency costs are higher, conservative contract modifications have a higher occurrence. Their results confirm the role of debtholder demands in determining the level of conservatism in firms (Beatty, Weber, & Yu, 2008).

Debt covenants are often seen as determining the transfer of decision rights from shareholders to lenders when the financial status of a company reaches a certain low level. This restricts managers’ opportunistic behavior and limits their ability to extrapolate lenders wealth. According to Nikolaev (2010), who studied public debt agreements. When debt agreements rely on covenants to a greater extent, the demand for conservative accounting is increased. Lenders can only rely on debt covenants to function properly as a signaling device when losses are recognized in a timely fashion. This can be seen as the demand from debtholders to apply prudence in the accounting of the borrower (Nikolaev, 2010).

According to one study, the efficiency of debt contracts is increased when a debt contract includes some covenant that allows for liquidation of the contract if the accounting gives a sufficiently negative signal. In the same article it is revealed that inefficient

liquidation decisions can be made because the information available to the lender does not perfectly reveal the true state of the company due to noise. In the model described in the

(24)

study, the costs of renegotiation matter. When renegotiation of debt contracts is costless, there will always be efficient liquidation decisions. However, when there are costs involved with renegotiation, the accounting information is also relevant. The paper further provides a tradeoff that will determine the efficiency of the debt contract. An efficient debt contract will reduce the total costs of renegotiating the contract but achieving efficient liquidation decisions and the cost of inefficient liquidation decisions but not incurring the renegotiation costs. The author finds that when the liquidation value of a contract increases, there is a stronger demand for more conservative accounting. This is because when there is a larger benefit from efficiently liquidating the contract in comparison to the loss from inefficient liquidation of the contract (Li, 2013).

The Gigler et al. (2009) paper that motivated this study reflects the previously mentioned inefficient decisions due to noise. The paper has provided a theoretical

framework of why one might expect a link between conservatism and the renegotiation of debt. Often, conservatism is seen as increasing the effectiveness of debt contracts because it is widely documented that applying conservative accounting decreases interest rates (Ahmed, Billings, Morton, & Stanford-Harris, 2002). As previously mentioned, the Gigler et al. (2009) paper challenges this assumption by developing a model where the efficiency of debt contracts is not measured simply by interest rates, but rather where efficient debt contracts reduce the costs of two distinct errors. The theoretical framework they provide leads to a conclusion that is quite different from the conclusions about the benefits arising from accounting conservatism in debt agreements in the majority of research on the

subject. They arrive at the conclusion that conservatism can actually decrease the efficiency of debt contracts (Gigler, Kanodia, Sapra, & Venugopalan, 2009).

(25)

2.4 Hypothesis development

Based on the previous discussion it seems that there are a lot of benefits from using conservatism in debt contracts. Interest rates can be lowered, which naturally benefits borrowers who are able to receive financing at lower costs (Watts, 2003a). Also, the assurance provided to lenders by valuing the company at the lower bound in the financial statements gives lenders a better picture of chances of default (Ruch & Taylor, 2015). However, in the face of all the criticism that conservatism has received it seems likely that there has to be a potential downside to using conservatism in financial statements. As mentioned earlier the reason often given is that conservatism distorts the unbiased data in the financial statements (Financial Accounting Standards Board, 2010). It seems though that much of the reasoning is mostly concerned with normative argumentation. The FASB uses the argumentation that accounting should represent a true value of the firm, but not much has been researched on the potential practical downsides of using conservatism.

One study that mentions one of these potential practical downsides is a study by Gigler et al. which theorizes that conservatism has the potential to increase the occurrence of debt agreements being prematurely called. The study claims that being conservative in financial statements will increase both the frequency of low performance results and changes the informational content of these results. This can increase the information asymmetry between the borrower and the lender in the debt agreement. Also they argue that conservatism will affect the best fitting debt covenants and that debt contract

efficiency is not necessarily measured by interest rates (Gigler, Kanodia, Sapra, & Venugopalan, 2009). The paper provides an informative example of an exam that helps clarify their arguments. The example concerns the grading of the exam. The paper assumes that grades are a way of communicating whether the student taking the exam is of high or low ability. This is useful to outsiders like employers who rely on education to sort potential employees. This selection reduces the adverse selection problem present in assessing potential employees. It states there are two options; either the exam is graded conservative or liberal. By scoring the students that took the exam in a conservative way, it will be harder for low ability students to achieve a good grade. At the same time it will also have an effect on high ability students who are now more likely to receive a lower grade. This makes students that score high on the exam to very likely be of high ability. In conjunction, some of the high ability students will score lower which will make outsiders question their ability

(26)

because low ability students will generally score lower. This could potentially affect the ability of the exam to signal information about students’ abilities. When the exam is graded in a more liberal way the problem is reversed and some low ability students will perform quite well which also distorts the informational content (Gigler, Kanodia, Sapra, & Venugopalan, 2009).

It further argues that decision rights should only be transferred from the firm to debt holders, due to the violation of debt covenants, when it is estimated that the firm will perform sufficiently bad in the future. They argue that due to conservatism changing the relation between financial statements and the estimated future performance of the firm, changes in conservatism will result in changes in debt covenants. The argument is that these changes in debt covenants can nog always change at the exact amount needed to offset the change in conservatism. This leads to the paper questioning whether more conservatism leads to more efficient debt agreements or not. The paper disputes previous research which claims that interest rates are a good measure of debt contracting efficiency, because as the above argumentation states, debt holders will compensate these lower interest rates with more decision rights transferred to the debt holders in the form of debt covenants (Gigler, Kanodia, Sapra, & Venugopalan, 2009).

The paper proposes a different measure of efficiency which they see as more suited. The optimal debt agreement will minimize the costs from making wrong investment

decisions because of too much conservatism, which will lead to false alarms, and being too optimistic. These false alarms happen when a firm is so conservative that the debt holder will doubt its financial stability while in truth the company is still financially healthy. The second cost is the costs of firms being too optimistic, causing debt holders to be willing to lend them money while in reality the firm is not in a financially healthy enough state for these debt agreements. Debt holders can request to gain decision rights on whether or not to continue the debt agreement when the firm signals a significantly negative indicator. By violating this debt covenant the borrower must allow the debt holder to choose to either continue the debt agreement or to change the initial contract, by either prematurely calling the debt or renegotiating the terms (Gigler, Kanodia, Sapra, & Venugopalan, 2009). The renegotiation of debt agreements is also found in other studies (Dichev & Skinner, 2002). Seeing as how conservatism can change the signal that accounting data gives there is reason to believe that being sufficiently conservative can, given the previous examples, alter the

(27)

signal in such a way that to debt holders it seems that the expected outcome of the debt agreement is sufficiently diminished, by violating the debt covenant, that they decide that the terms of the debt agreement have to be renegotiated. The real financial status of the borrower however should not change due to these accounting choices so these debt agreements in reality could still be profitable to the lenders (Gigler, Kanodia, Sapra, & Venugopalan, 2009).

Despite previous evidence on the benefits of conservatism (Francis, Hasan, & Wu, 2013; Zhang J. , 2008; Ahmed & Duellman, Accounting conservatism and board of director characteristics: An empirical analysis, 2007; Ahmed, Billings, Morton, & Stanford-Harris, 2002), the Gigler et al. (2009) study provides a compelling conceptual idea as to how conservatism could also negatively affect debt contracting efficiency. The study proposes that profitable debt agreements may be renegotiated because of the conservatism exercised by the borrower in the debt agreement. This study will examine this possible relationship. Therefore, the hypothesis is as follows:

H1: There is a higher occurrence of renegotiation of debt agreements for borrowers that employ a higher level of accounting conservatism.

(28)

3. Research methodology

3.1 Sample

In order to be able to test the hypothesis, I require a sample of debt contracts, including data on their possible renegotiation. I will also require the matching conservatism proxies for these firms for the periods in which these firms experienced a debt contract

renegotiation, together with the conservatism proxies for the periods where the firms did not experience a debt renegotiation and several control variables. This will enable me to compare the two situations. The conservatism proxy chosen in this study is calculated using data from the Compustat database. The required data is present only in the North America Capital IQ Compustat database. This restricts the sample to US and Canadian firms. To match the observations in the Thomson One sample, years 1985-2015 are collected. The total sample of firm year observations is 376,264. 33,613 observations are deleted from the sample for having a financial industry format. Duplicates for gvkey and fyear are deleted. From this sample, observations with insufficient data to calculate the required variables are dropped. By doing this, 222,098 observations remain. While calculating the conservatism proxy, the data is trimmed by dropping the top and bottom 1% of observations based on firms’ earnings, returns, sizes, market-to-book ratios, leverages, ROAs, and net cash flows from operating activities. This reduces the suitable sample to 117,676 observations. The calculated variables are entered into a regression that will be explained in the following section, this results in 116,788 C_Scores, 57,480 NOPA scores, and 77,714 SKEW scores. The sample for the debt contract renegotiation sample is collected from the Thomson One database. The sample consists of all syndicated loans for US or Canadian firms for the fiscal years 1985-2015. Syndicated loans make up the majority of capital provision in the United States and continues to grow. Syndicated loans also imply informed parties who process the public information to their best ability and supplementary information is available about syndicated loan trading compared to other forms of debt (Wittenberg-Moerman, 2008). The broad period is chosen because preliminary research shows that suitable observations are not be in abundance. The total sample consists of 126,211 observations. After deleting observations with missing date data 114,696 observations remain. The 1985 starting year is chosen because that year is the year the Thomson One started recording the suitable data. The data on debt renegotiation is then matched to the Compustat data using CUSIP codes

(29)

and firm years. This results in 23,378 matched observations. Finally, duplicates are removed when a firm has experienced multiple loan amendments in one year. I keep one firm year observation for each firm in each year. The DC indicator variable is equal to 1 when the firm has experienced one or more decrease commitment amendments. When a firm has

experienced multiple loan amendments, none of which includes a decrease commitment amendment, the DC indicator variable is equal to 0. Removing the duplicates reduces the sample to 14,557 observations. The sample is summarized in table 1. The firm year observations where a debt renegotiation, described in a following section, is experienced are given an indicator variable 1. This indicator variable is matched to the conservatism proxy score for the identical firm year.

Table 1 Sample Selection

Compustat Thomson One Deleted observations Compustat 376,264

Observations with sufficient data

222,098 154,166

After trimming the data 117,676 104,422

Thomson One 126,211

Observations with sufficient data

114,696 11,515

Combined sample 23,378 23,378

After dropping duplicates 14,557 14,557 8,821

Final sample 14,557 14,557

3.2 Conservatism operationalization

Through extensive research on the subject of conservatism, proxies have been developed that in a large sense represent the degree of conservatism in financial statements. Many papers have considered the importance of the study conducted by Basu (1997). The definition of conservatism used, being the asymmetric timeliness of bad news vs. good news, is being used in most papers that came after it (Ruch & Taylor, 2015; Khan & Watts, 2009; Penman & Zhang, 2002). Measuring conservatism in its pure form is near impossible because it requires extensive knowledge of the company and all transactions that require companies to make decisions with regard to how conservative they wish to approach their business. As such, many studies have developed proxies for conservatism.

(30)

One of the most promising measures for conservatism is the so called C_Score (Francis, Hasan, & Wu, 2013; Khan & Watts, 2009). This measure is based on several features. The C_Score is based on natural logarithm of firm size, market-to-book ratio and the firm’s leverage (Khan & Watts, 2009). These characteristics are readily available in the Compustat database and therefore seems a logical choice. Furthermore, studies confirm that these characteristics vary empirically and theoretically with levels of conservatism (Lafond & Watts, 2008).

The C_Score is based on the measure of asymmetric timeliness from Basu (1997). This paper specified the regression as follows:

𝜒 = 𝛽1 + 𝛽2 ∗ 𝐷 + 𝛽3 ∗ 𝑅 + 𝛽4 ∗ 𝐷 ∗ 𝑅 + 𝜀

The 𝜒 being earnings, D being the dummy variable equal to 1 when earnings<0, R being the returns (which measures news) and 𝜀 being the residual.

The C_Score calculates the timeliness of the negative news a firm experiences. It is calculated as follows:

𝐶_{𝑆𝑐𝑜𝑟𝑒} = 𝜆1 + 𝜆2 ∗ 𝑆𝑖𝑧𝑒 + 𝜆3 ∗ (𝑀

𝐵) + 𝜆4 ∗ 𝐿𝑒𝑣

The G_Score, which measures the good news timeliness is measured as followed: 𝐺_{𝑆𝑐𝑜𝑟𝑒} = 𝜇1 + 𝜇1 ∗ 𝑆𝑖𝑧𝑒 + 𝜇1 ∗ (𝑀

𝐵) + 𝜇1 ∗ 𝐿𝑒𝑣

Taking the formulas of the C_Score and the G_score and applying these to the Basu formula gives: 𝜒 = 𝛽1 + 𝛽2 ∗ 𝐷 + 𝑅 ∗ (𝜇1 + 𝜇1 ∗ 𝑆𝑖𝑧𝑒 + 𝜇1 ∗ (𝑀 𝐵) + 𝜇1 ∗ 𝐿𝑒𝑣) + 𝐷 ∗ 𝑅 ∗ (𝜆1 + 𝜆2 ∗ 𝑆𝑖𝑧𝑒 + 𝜆3 ∗ (𝑀 𝐵) + 𝜆4 ∗ 𝐿𝑒𝑣) + (𝛿1 ∗ 𝑆𝑖𝑧𝑒 + 𝛿2 ∗ ( 𝑀 𝐵) + 𝛿3 ∗ 𝐿𝑒𝑣 + 𝛿4 ∗ 𝐷 ∗ 𝑆𝑖𝑧𝑒 + 𝛿5 ∗ 𝐷 ∗ (𝑀 𝐵) + 𝛿6 ∗ 𝐷 ∗ ( 𝑀 𝐵)) + 𝜀

This allows us to extract the corresponding 𝜆1, 𝜆2, 𝜆3 and 𝜆4 values needed to calculate the C_Score.

In order to increase the robustness of this study, I employ a second measure of conservatism. For this purpose I follow previous studies and use the market-to-book ratio as a proxy for conservatism (Beatty, Weber, & Yu, 2008; Ahmed, Billings, Morton, & Stanford-Harris, 2002; Beaver & Ryan, 2005). I follow the method of Khan and Watts (2009) to calculate the market-to-book ratio. This is calculated as the market value of equity divided

(31)

by common equity.

Further increasing the robustness of this study, I employ two additional measures of conservatism found in Francis et al. (2013) the measures I employ are the non-operating accruals and the skewness of earnings.

Using the non-operating accruals as a measure for conservatism is broadly used in the literature (Zhang J. , 2008; Francis, Hasan, & Wu, 2013). The measure for non-operating accruals, NOPA, is equal to the non-operating accruals divided by the lagged total assets. This is multiplied by -1, as to create consistency in the conservatism measures. By

multiplying by -1, a higher NOPA is associated with higher conservatism.

Using the skewness of earnings as a measure for conservatism is valid. This is because when earnings are skewed negatively, the recognition of bad news requires lower verification compared to good news (Zhang J. , 2008). The skewness of earnings is calculated using the three previous firm years. In order to control for firm performance variation, the skewness of earnings is divided by the skewness of cash flows plus 1, which similarly is calculated on a three year basis. Similar to NOPA, the result is multiplied by -1 so a higher SKEW is associated with higher levels of conservatism.

3.3 Renegotiation of debt contracts

Debt contracts are negotiated at the inception of the contract. After certain events, like debt covenant violations, debt contracts may be renegotiated. As previously explained, debt renegotiation exists in multiple forms (Gopalakrishnan & Parkash, 1995; Chen & Wei, 1993). This study will focus on renegotiations that are thought to have the strongest impact. The relationship between conservatism and repayment of loans is the one that is studied most heavily in this thesis (Gigler, Kanodia, Sapra, & Venugopalan, 2009). The renegotiation of debt contracts is found in the Thomson One syndicated loans database. Thomson One refers to debt renegotiation as amendments. The previously mentioned possible responses to debt covenant violation give us an indication of what type of amendment would be suitable for this study. The covenants that are based on debt-to-equity ratios and tangible net worth have the highest chance to cause a default on debt.

Several renegotiation options are possible after a debt covenant violation. Waiving the violation is the most likely of options. This might not seem like it would be of any cost to the creditor or the borrower, but studies have shown that waivers are thought to be costly

(32)

by creditors (Chen & Wei, 1993). The second most likely response is placing additional constraints on the borrower. An increase of interest rates is named the third most likely response. Increasing the collateral is the subsequent most likely option. The least likely response to a debt covenant violation is found to be termination or immediate repayment of the loan (Gopalakrishnan & Parkash, 1995).

The amendment type that is most similar to what is described in the Gigler et al. (2009) paper, being the repayment of a loan, is the decrease commitment amendment and using this amendment is in line with the previously mentioned goal of finding renegotiations that have a high influence on the borrower. The firms that experienced a decrease

commitment are given a dichotomous variable 1 for the respective year of the amendment. The findings in the Gopalakrishnan & Parkash (1995) study are in accordance with the respective amendments found in the Thomson One database. In the final sample, the number of amendments that represent repayment of (part of) a loan is a very small

percentage of the total sample. If a statistical relationship is found, it could be an indication that the other possible responses also happen on a greater occasion for more conservative firms. The definition of an amendment in the Thomson One database does not imply that it is guaranteed whether the amendment is the result of renegotiation or if it is the result of one-sided lender decisions. The conceptual idea of this uncertainty is not unfunded. Previous studies have found that creditors have waived part of a loan to a borrower in the time of financial difficulty, this prevents the borrower from filing for bankruptcy (Bester, 1994). However, since debt renegotiation is a private affair between a creditor and a borrower, the motivations for the renegotiation will remain unknown. The only way to find the real reasons behind renegotiation would be an in depth case study including interviews with the different actors. This could be an interesting follow-up research if the results from this study indicate an influence of conservatism on debt renegotiation.

3.4 Control variables

In determining control variables, this study looks at previous studies who included control variables that have had an influence on renegotiation of debt contracts.

Following Beatty et al. (2008), I include several control variables that are likely to affect contract renegotiations. I include maturity of the loan in years, because a longer maturity will lead to a higher chance of covenant violation. The loan divided by total assets is included because firms will likely face stricter covenants when the loan is a bigger portion

The effect of accounting conservatism on the efficiency of debt contracting : a study on the effects of accounting conservatism on the renegotiation of debt agreements