University of Amsterdam
Faculty of Economics & Business
Msc. Accountancy & Control
The Influence of Firm Leverage on the Trade-Off
between Real and Accrual-Based Earnings
Management
Name:
Martijn Kasper
Student number: 10564241
Date:
17
thof August 2015
2
Abstract
This paper researches the influence of firm leverage on the trade-off between real and accrual-based earnings management. Prior literature shows that managers’ accounting choices are influenced by scrutiny from large institutional investors and debt covenants. Other research shows that managers trade-off different forms of earnings management based on their relative costliness. To measure accrual-based earnings management the Jones model (1991) was used, as used by Zang (2012). For measuring real earnings management the Roychowdhury model (2006) was used, as used in Zang (2012). For measuring the trade-off in earnings management the Zang model (2012) was used. The results neither show a significant relationship between firm leverage and real or accrual-based earnings management, nor between firm leverage and the trade-off in earnings management.
This document is written by [Student] Martijn Kasper, 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
Table of contents
2 Introduction 4
3 Literature Review and Hypotheses 6
3.1 Accrual-Based Earnings Management 6
3.2 Real Earnings Management 8
3.3 Tradeoff Between Accrual-Based and Real Earnings Management 9
3.4 Earnings Management and Firm Leverage 11
3.5 Hypothesis Development 12
4 Methodology 14
4.1 Sample Selection & Event Period 14
4.2 Firm Leverage and Accrual-Based Earnings Management 14 4.3 Firm leverage and real earnings management 15
4.4 Costs of earnings management 17
4.5 Firm leverage and the tradeoff between accrual-based and real earnings management 18
4.6 Control variables 19
5 Empirical results 20
5.1 Descriptive statistics 20
5.2 Research model 1: firm leverage and accrual-based earnings management 21
5.2.1 Correlation matrix analysis 21
5.2.2 Regression analysis and review of hypotheses 21 5.3 Research model 2: firm leverage and real earnings management 23
5.3.1 Correlation matrix analysis 23
5.3.2 Regression analysis and review of hypothesis 24 5.4 Research model 3: firm leverage and the tradeoff between accrual-based and real earnings management 25
5.4.1 Correlation matrix analysis 25
5.4.2 Regression analysis and review of hypothesis 26
7 Conclusion 28
4
2
Introduction
This thesis examines the effect of debt leverage on the tradeoff between accrual-based earnings management and real activities earnings management in the United States. A widely covered topic in accounting research, earnings management entails the actions that managers undertake to purposefully manipulate a company’s earnings numbers to match a pre-determined target.
Earnings management occurs when managers use judgment in financial reporting and in structuring transactions to alter financial reports to either mislead some stakeholders about the underlying economic performance of the company or to influence contractual outcomes that depend on reported accounting numbers. (Healy & Wahlen, 1999, p.365)
Early studies have found evidence of accrual-based earnings management, such as Healy (1985) who showed that managers use discretionary accruals to maximize the value of their bonus awards. He found evidence that managers use income-decreasing accruals when reaching the upper boundaries and income-increasing accruals when reaching the lower boundaries of their bonus schemes. Accrual-based earnings management focusses on how managers manipulate discretionary accruals. Roychowdhury (2006) however found evidence that managers also manipulate real activities, influencing the actual cash flows of a company, to avoid having to report losses. Real activities earnings management consists of managers deviating from normal business activities, with the goal of meeting certain earnings thresholds. (Roychowdhury, 2006, p. 336) Methods used for real earnings management are acceleration of sales, alterations in shipment schedules, delaying of research and development (R&D) and maintenance expenditures. (as cited in Roychowdhury, 2006, p. 336) Real earnings management, per definition, has negative implications for financial stakeholders, as it entails management making suboptimal decisions.
Managers have different motives to engage in earnings management. Healy & Wahlen (1999) researched that earnings management mostly occurs due to the influence of stock market perceptions, to increase management’s compensation, and to reduce the likelihood of violating lending agreements.
On earnings management to influence stock market perceptions, little evidence is available, as it is difficult to measure managers’ motives in general, but especially with respect to
5
influencing stock prices. Research does show that companies typically engage in more earnings management before IPO’s, suggesting a clear link between earnings management and boosting stock prices. (Taoh et al. 1998) Regarding the effect of earnings management on stock prices, Dechow et al. (1996) researched companies that are subject to SEC investigations for non-compliance with GAAP and found that cost of capital rises significantly for these companies, suggesting a negative relationship between earnings management and stock prices.
The subject of earnings management in relation to management compensation contracts is researched extensively. Healy (1985) found a clear link between managers’ reaching the lower or upper bound of their bonus plans and the extent to which they respectively inflate and deflate earnings. Not only are accounting policies related to bonus compensation contracts, but managers do also change their accounting procedures when their bonus compensation plans change.
Several researchers have tried to link earnings management to the violation of debt contracts, among which Healy & Palepu (1990), who found that companies close to violating their dividend constraints use dividend cuts to avoid violations. Sweeney (1992) however found that companies close to violating debt-contracts use income-increasing accounting methods to avoid violation. It thus appears that, although not very clear, a relationship exists between debt covenants and earnings management.
Despite the fact that an extensive literature on earnings management exists, it remains a widespread topic in accounting research. This is partly due to its elusive nature, as earnings management is generally only measured by deviations from industry averages, not fully respecting the wide range of other variables that might be accessory to this deviation. In addition, the wide array of implications of earnings management continues to be relevant for both standard setters and credit providers. Most importantly accrual-based influences financial reporting quality, which directly impairs the informativeness of financial statements. This is particularly important for public debt holders, as they primarily rely on the information provided in the financial statements for their decision making. Real earnings management has implications for both public and private debt holders, as management make suboptimal decisions to meet earnings thresholds. This means that real earnings management has, per definition, negative implications for financial stakeholders.
6
A recent development in the literature on earnings management, is the phenomenon of companies substituting one form of earnings management for the other, as external factors change. Cohen et al. (2002) showed that accrual-based earnings management increased steadily from 1987 to 2002, after which the Sarbanes-Oxley Act (SOX) was implemented. After the passage of SOX the level of accrual-based earnings management dropped quickly, being substituted by real earnings management, evidencing that managers tradeoff accrual-based earnings management for real earnings management. This relationships is interesting because it might allow real earnings management, which is very difficult to detect, to be predicted by accrual-based earnings management. This research is particularly interested in finding a relationship between the extent to which a firm is financed by external debt and the tradeoff in earnings management. Hence the research question: to what extent is the tradeoff between accrual-based earnings management and real earnings management influenced by firm-leverage?
The remainder of this paper proceeds as follows. Section 2 discusses the tradeoff in earnings management related to firm leverage and previous research. Section 3 discusses the hypotheses that are used in answering the research question. Subsequently, section 4 discusses the methodology, followed by descriptive statistics in section 5. In section 6 I present my results, after which I give a conclusion and opportunities for future research.
3
Literature Review and Hypotheses
3.1
Accrual-Based Earnings Management
Accruals are non-cash transactions. The International Accounting Standards Board (IASB) defines accrual-based accounting as follows:
“Accrual accounting depicts the effects of transactions and other events and circumstances
on a reporting entity’s economic resources and claims in the periods in which those effects occur, even if the resulting cash receipts and payments occur in a different period. This is important because information about a reporting entity’s economic resources and claims and changes in its economic resources and claims during a period provides a better basis for assessing the entity’s past and future performance than information solely about cash receipts and payments during that period.” (The Conceptual Framework for Financial Reporting 2012)
7
Whereas real earnings management has direct negative economic implications for a firm, accrual-based earnings management is related to accounting choices and is therefore negatively associated with accounting quality. Accruals can be divided in discretionary and nondiscretionary (Jones 1991). With discretionary accruals managers can use their discretion in choosing a period to which the expense is allocated. Nondiscretionary accruals are expensed in the period in which they occur.
One of the first to provide a positive theory on accrual-based earnings management were Zmijewski & Hagerman (1981). They found the first evidence that managers use their discretion, to make their accounting choices be in line with the firm’s strategy. Especially firm size showed to be strongly associated with a firm’s income strategy. Until then, most research assumed that accounting choices within a firm were made individually, and not as part of an overall strategy. Zmijewski & Hagerman showed that firms engage in both income-increasing and –decreasing strategies. Another study researching the effect of bonus contracts on accruals is that of Dechow & Sloan (1991). They found that CEO’s manage earnings performance in the last years of their employment, to accelerate earnings and hence their bonuses. Since CEO’s won’t be held accountable for negative developments happening after their employment, they don’t take into consideration the negative implications of accelerating earnings.
In addition to using earnings management for bonus contracts, managers also try to influence stock prices by using their discretion in accruals. Sloan (1996) researched to what extent investors can anticipate future earnings, by looking at current stock prices and future earnings. He divided earnings in cash-flow- and accrual-based, and showed that earnings concerning the latter are significantly less consistent than the former. He argues that the efficient market hypothesis –assuming that all available information is incorporated in stock prices– is incorrect. Sloan’s research provides evidence that abnormal accruals are not fully incorporated into stock prices, suggesting that accrual-based earnings management is often not detected, or taken into account by investors.
Xie (2001) extended this line of research by studying a comprehensive sample of 7,506 U.S. firms. She found that, when investors take accruals into consideration, they have a tendency to overestimate the persistence of positive accruals, causing them to overprice. Although it is
8
evidenced that abnormal accruals reverse over time (Ball & Shivakumar, 2005), investors do not really appear to understand this phenomenon. Hence, Xie’s research indicates that firms do reap benefits of earnings management in the form of higher share prices. Despite investors generally overpricing positive accruals, Teoh et al (1998a) found that firms that have abnormally high accruals in the three years after an IPO (initial public offerings) experience a decrease of up to 20 percent in their stock returns, compared to a sample of control firms. It thus appears that, although managers benefit from favorable returns during an IPO, investors anticipate a reverse in accruals and value shares accordingly. This notion is further evidenced by Dechow (1996). who found evidence that managers overstate their earnings to decrease cost of capital, but that cost of capital significantly increases after the revelation of earnings management by SEC (Securities and Exchange Commission). It thus appears that the implications of earnings management for both firms and investors are not yet fully understood.
3.2
Real Earnings Management
As mentioned before, the vast majority of the real earnings management literature has examined accrual-based earnings management. Until recently, the managing of real activities in order to meet earnings benchmarks was not well understood. Graham et al (2005) surveyed more than 400 executive managers, informing about their point of view regarding earnings management. More than 75% of the interviewed managers said that they would prefer taking real economic decisions that could possibly have long-term effects, than change their accounting choices to influence earnings. Showing its prevalence might be higher than previously thought, this study paved the way for further research on real earnings management. Real earnings management can be measured in several ways. Some examples include: cutting of R&D investments (Gunny 2010; Zang 2012; Roychowdhury 2006; Cohen et al 2008), overproducing to decrease COGS expenses (Gunny 2010; Roychowdhury 2006; Cohen et al 2008), cutting on advertising expenses (Zang 2012; Roychowdhury 2006; Cohen et al 2008)
Several other studies have tried to shed light on the reasons that managers engage in real earnings management. The most widely used way of engaging in real earnings management is by cutting R&D expenditures, which is described in a wide array of different researches. Mande et al (2000), found that in a sample of Japanese firms, managers cut R&D expenditures to boost short-term earnings. Using their research setting to compare between the 90’s Japanese financial crisis
9
and the time period in anticipation of the crisis, they showed that managers heavily cut on R&D expenses following the crisis, evidencing that these decisions were not in best interest of the company, but indeed to boost short-term earnings. Roychowdhury (2006) extended the notion of real earnings management, by including several other proxies, including price discounts, and other discretionary expenditures, such as advertising and maintenance. Through his research he showed that managers cut on discretionary expenditures and use overproduction to decrease COGS, in order to increase reported margins. The study also provides evidence suggesting that real earnings management is prevalent around zero-earnings benchmarks.
Gunny (2010) extends this line of research by showing that firms engaging in real earnings management are actually more likely to meet zero-earnings benchmarks. She argues that firms that find themselves just below zero-profit are incentivized to make economic decisions to just meet the zero-earnings benchmark. The reason for this is that analysts’ and investors’ negative response to a small loss is much greater than their positive response on a small profit (Gunny 2010).
3.3
Tradeoff Between Accrual-Based and Real Earnings Management
In recent years, the idea that accrual-based earnings management and real earnings management are used as a trade-off for the other has gained more and more support. The assumption behind this idea is that managers always have incentives to manage earnings, and will look for alternatives if their current earnings management strategies become too costly.
The first empirical evidence on this tradeoff comes from Cohen et al (2008), who used a large-sample research setting, set around the implementation of the Sarbanes-Oxley (SOX) act in 2002, to research if and how managers substitute one form of earnings management for the other. They argued and found that, since the implementation of SOX required companies to publish extensive financial information (making accrual-based earnings management more easily detectable), managers would engage less in accrual-based earnings management, but substitute this by manipulating earnings through real economic decisions. Although the authors state that the changes might not have been solely due to the implementation of SOX, the findings do suggest that external factors have an impact on the choice between accrual-based and real earnings management. The authors also argued that firms trade-off between real and accrual-based earnings management to avoid sudden negative earnings surprises. Increasing scrutiny over
10
managers –caused by strengthened regulations around earnings publications– cause them to change their earnings management strategies.
Cohen & Zarowin (2010) investigated the phenomenon in a different setting, namely around seasoned equity offerings (SEOs). In contrast to the findings of Cohen et al (2008), real earnings management declined in their sample, in favor of accrual-based earnings management. The increased importance of inflated earnings around SEOs causes manager to change to accrual-based earnings management, despite higher scrutiny over the financial statements by investors and analysts.
Zang (2012) shed more light on how the tradeoff is made. By arguing that, while real earnings management can only be exercised during the fiscal year, managers are still able to use their discretion in allocating accruals once the fiscal year has ended. Hence accrual-based earnings management decisions follow real earnings management decisions. She provides evidence that managers base their accrual-based earnings management on the outcomes of their real earnings management, using the two as substitutes. In addition, she shows that managers substitute both earnings managements based on their relative costs. If the environment a firm operates in is characterized by low competitiveness (proxied by market share), low financial health (proxied by financial ratios), high marginal tax rates, and high scrutiny from institutional investors, real earnings management is more costly as opposed to accrual-based earnings management, and managers are likely to substitute the former with the latter. On the other hand, if a firm’s accounting practices are more constrained by regulatory scrutiny, prior periods’ accrual manipulation, and shorter operating cycles, managers tend to engage more in accrual-based instead of real earnings management.
Cohen et al (2008) describe negative earnings surprises as the most important cost of accrual-based earnings management Since accruals reverse (Ball & Shivakumar, 2005), positive accruals are followed by negative accruals, reversing a company’s profit as well. Since shareholders don’t fully incorporate earnings management in their share prices (Sloan 1996), the risk of an unanticipated drop in earnings is a steep correction of the share price. The cost of real earnings management lies directly with the firm, as per definition it means a firm makes suboptimal decisions.
11
Based on prior literature, it appears that managers engage in both real and accrual-based earnings management for the same reason: managing earnings, in order to positively influence perceptions of stakeholders. An additional reason why managers engage in real earnings management is that it is harder to detect and therefore less susceptible to causing negative share price corrections (cohen et al 2008). Since their benefits are similar, managers use real and accrual-based earnings management as a tradeoff based on their relative costs and not on their benefits. These costs seem to change as external factors change, such as financial reporting requirements or shareholder composition.
3.4 Earnings Management and Firm Leverage
Widely researched, an important factor influencing earnings management is firm leverage. An influential study on this subject is from Bharath et al (2008) who researched a large sample of over 3,000 U.S. firms, over a time span of 15 years, on the influence of accounting quality on the borrower’s choice between public (bondholders) and private (debtholders) debt contracts. They found that companies who engage more in accrual-based earnings management prefer private debt over public debt, as private debtholders are usually given inside information to satisfy their risk averseness. Borrowing from private lenders mitigates the risk premium that would come with debt contracting if they were to borrow from public lenders, who solely rely on financial information that is made public by the company. This implies that managers do bring in line their accounting strategies with their borrowing strategies.
Additional support for the view that managers take into account external lenders’ demands and wishes comes from Healy & Palepu (1990). They researched if managers change their accounting methods after the implementation of dividend covenant restrictions by external debt-providers and find that, although US GAAP leaves them with a lot of discretion regarding the accounting for dividends, managers do not change their accounting methods to circumvent these covenants. This provides further evidence that debt holders do indeed have considerable scrutiny over managers and possibly their accounting choices. Bushee (1998) writes that with an increasing degree of institutional investors, incentives for the engagement in real earnings management drop. He measures this through investments in research and development to meet short-term earnings goals. According to Bushee, institutional investors reduce myopic behavior of managers, indicating that powerful stakeholders have a considerable amount of scrutiny over managers’
12
decisions. He provided evidence that the composition of a company’s debtholders and shareholders have large implications for company related decisions that managers make.
Furthermore, although earnings management seems to be occurring in many different industries and types of firms, its prevalence does appear to be mitigated by laws and regulations scrutiny of investors. First of all, despite their dispersed nature, the presence of public investors (bondholders) decreases earnings management. Burgstahler (2006) found that private firms exhibit higher levels of earnings management than publically listed firms. In addition, he finds that strong legal systems mitigate the prevalence of earnings management. This underpins the theory that firms comply to the capital market’s demand for informativeness of financial statements.
3.5
Hypothesis Development
Although the real reasons that managers use earnings management remain elusive, some reasons (for both accrual-based and real earnings management) include boosting share prices (by meeting or just beating earnings-benchmarks), meeting bonus contract covenants and avoiding meeting debt covenants.
Earnings management comes as well with costs. In addition to the costs mentioned by Zang (2012), accrual-based earnings management is characterized by a possible risk-premium demanded by shareholders, and –in case of earnings surprises– can cause a downward correction in a company’s share price. The costs of real earnings management are the negative consequences of cutting on R&D expenses and other discretional expenditures. Because real earnings management is harder to detect, its possible effect on a company’s share price is lower than that of accrual-based earnings management.
Since accrual-based earnings management is mainly costly to shareholders –as it decreases the quality of financial reporting, but does not change the actual economic and investment-related decisions that are made in a company– we expect that firms with high leverage (and hence a smaller relative dependency towards shareholders) are more incentivized to use accrual-based earnings management in favor of real earnings management. This thought is strengthened by the idea that private debtholders have private information over shareholders, and therefore pay less importance to accrual-based earnings management, as they don’t have to rely solely on financial statements. As a result I develop the following hypothesis:
13
H1: Firms with high leverage increase their accrual-based earnings management.
As I assume private debtholders have high scrutiny over managers in their accounting decisions, I believe that this also has an influence on the level of real earnings management present in a firm. As mentioned before, real earnings management has direct negative economic implications for financial stakeholders. Since private debtholders have private information that public debtholders don’t have, I expect the first to be better able to detect real earnings management and its negative economic consequences. I hence develop the following hypothesis:
H2: Firms with high leverage decrease their real earnings management activities.
In recent years several researches have provided evidence for a trade-off between accrual-based and real earnings management. Most of these researches have been conducted around the implementation of SOX in 2002, giving managers incentives to decrease their levels of accrual-based earnings management in favor of real earnings management. SOX requires increased disclosure of financial information regarding accrual-based earnings management (Cohen et al 2008), which came available to public debtholders. The higher availability of information increased the relative cost of accrual-based earnings management and caused managers to switch to real earnings management.
Since private debtholders have more inside information than public debtholders, I expect a change in firm leverage have an effect on the tradeoff between accrual-based and real earnings management. I therefore develop the following hypothesis:
H3: Firms with high leverage that show increased accrual-based earnings management use this as a substitute for their real earnings management activities.
14
4
Methodology
4.1 Sample Selection & Event Period
Previous researches regarding the tradeoff between accrual-based and real earnings management are set around a specific event, such as the implementation of SOX or SEOs. My research does not focus on such an event, but on the change in a firm’s leverage over time. As a result I use a time-series regression, between the years 2004 until 2014. I start my sample in 2004, because I assume that the disruptive effects of the implementation of SOX will have disappeared. Including earlier years would distort the effect of firm leverage, as SOX caused managers to substitute accrual-based for real earnings management (Cohen et al 2008).
I gather my sample from the Compustat Merged Database, since it allows me to acquire the statements of cash flows, which I use to calculate accruals. I only allow firms that have all the data available to calculate accruals cash-flows. According to Cohen et al (2008) this might cause a self-selecting bias, because the extent to which managers influence financial information might be correlated to their earnings management strategies. Nevertheless, they argue that this bias reduces variation in the metrics used, making the test more conservative. I follow their reasoning in my model. Following Zang (2012) I don’t include financial institutions in my sample, nor regulated industries. The sample gathered consists of 22,275 observations. For some variables less observations are taken into account, due to missing values.
The sample is not be limited to suspect firms, as is the case in Zang’s (2012) paper. The reason for this is that my paper includes the variable firm leverage and therefore does not focus on suspect firms specifically. In addition, in Zang’s method of selecting suspect firms independent variables are selected that mostly represent market responses to earnings. As my research has a large focus on debt, Zang’s variables would cause a bias towards public debtholders’ incentives to engage in earnings management.
4.2 Firm Leverage and Accrual-Based Earnings Management
In measuring firm leverage and accrual-based earnings management I use the Zang (2012) model. She uses an adjusted version of the Jones (1990) Model, in which the difference between actual and expected accruals are used to proxy for accrual-based earnings management. Expected accruals are calculated within industry averages. Using industry averages helps give a better reflection of total accruals, controlling for cross-industry variations.
15
Following Zang (2012), I calculate the normal level of accruals as follows:
𝐴𝑐𝑐𝑟𝑢𝑎𝑙𝑠𝑡 𝐴𝑡−1 = 𝛼0+ 𝛼1( 1 𝐴𝑡−1) + 𝛼2( 𝛥𝑆𝑡 𝐴𝑡−1) + 𝛼3( 𝑃𝑃𝐸𝑡 𝐴𝑡−1) + 𝜀𝑡 (1)
In this formula, Accrualst is earnings before extraordinary items and discontinued operations minus the operating cash flow reported in the statement of cash flow, in year t. PPEt is gross property, plant, and equipment. Equation 1 is measured using a time-series from 2004-2014, using the two-digit SIC industry grouping, to make the coefficient reflect industry-wide averages.
A is the total of assets. I will use the residual to proxy for accrual-based earnings management
(AMt)
Firm leverage (FL) is proxied as the debt-to-equity ratio of the firm, using long-term debt, as done in previous research (Dhaliwal 1980). Hence, the model describes the influence that firm leverage has on accrual-based earnings management (H1) is as follows:
𝐴𝑀𝑡 = 𝛼 + 𝛽1𝐹𝐿𝑡+ ∑ 𝛽𝑗,𝑡𝐶𝑜𝑛𝑡𝑟𝑜𝑙𝑠𝑗,𝑡+ 𝜀𝑡 (2)
4.3 Firm leverage and real earnings management
In measuring firm leverage and real earnings management I use the Zang (2012) Model. This model includes the following metrics to proxy for real earnings management: increasing earnings by reducing costs of goods sold by overproducing inventory (1), and cutting discretionary expenditures (2), including R&D, advertising, and selling, general and administrative expenditures (SG&A). Both metrics are calculated by abnormal values compared to industry averages.
Increasing earnings by reducing costs of goods sold by overproducing inventory is calculated as follows. First, the normal level of production is calculating following Roychowdhury (2006; as cited in Zang 2008), using the following formula:
𝑃𝑅𝑂𝐷𝑡 𝐴𝑡−1 = 𝛼0+ 𝛼1( 1 𝐴𝑡−1) + 𝛼2( 𝑆𝑡 𝐴𝑡−1) + 𝛼3( 𝛥𝑆𝑡 𝐴𝑡−1) + 𝛼4( 𝛥𝑆𝑡−1 𝐴𝑡−1) + 𝜀𝑡 (3)
16
In this model, PRODt is the total cost of goods sold for a particular firm for year t, plus the change in inventory over year t–1. A is the total of assets in year t–1. S equates to net sales in year
t. ΔS equates to the net sales in year t–1 compared to t.
The abnormal level of production costs (RMPROD) is estimated using εt, Where a higher value means a higher error means a higher level of abnormal level of production costs, indicating higher levels of earnings management, through a larger reduction of cost of goods sold (Zang 2012). Equation 3 is measured using a time-series from 2004-2014, using the two-digit SIC industry grouping, to make the coefficient reflect industry-wide averages (PRODt), and errors (εt). In measuring discretionary expenditures I also use a model developed by Roychowdhury (2006; as cited by Zang 2012). Normal levels of discretionary expenditures are measured using the following formula:
𝐷𝐼𝑆𝑋𝑡 𝐴𝑡−1 = 𝛼0+ 𝛼1( 1 𝐴𝑡−1) + 𝛼2( 𝑆𝑡−1 𝐴𝑡−1) + 𝜀𝑡 (4)
In this model, DISX equates to discretionary expenditures in year t, proxied by the sum of R&D, advertising, and SG&A expenditures. The abnormal levels of discretionary expenditures are measured using εt. To make sure that higher values indicate greater amounts of discretionary expenditures (RMDISX), I multiply the residuals by –1. Both discretionary expenditures and abnormal levels of production is added up to proxy for the levels of real earnings management present in a firm. Equation 4 is measured using a time-series from 2004-2014, using the two-digit SIC industry grouping, to make the coefficient reflect industry-wide averages (DISXt), and errors (εt).
Firm leverage (FL) is proxied as the debt-to-equity ratio of the firm, using long-term debt, as done in previous research (Dhaliwal 1980). Hence, the model describing the influence that firm leverage has on real earnings management (H2) is as follows:
17
4.4 Costs of earnings management
Both accrual-based and real earnings management come with costs. These costs are associated with their relative trade-off and thus function as a predictor for the way managers substitute one form of earnings management for the other (Zang 2012). Following Zang (2012), the costs of real earnings management are measured using the following variables: market share (as market share increases, real earnings management becomes less costly), a firm’s financial health (better financial health indicates lower costs for real earnings management), and finally marginal tax rates (the relationship p between marginal tax rates and real earnings management costs is positive). I do not include the cost of institutional ownership, as there is a direct relationship with firm leverage, possibly causing a bias. A firm’s market share is based on three-digit SIC codes using the Harris Method (1998; as cited in Zang 2012), measured at the beginning of the year. A firm’s financial health is measured using the modified Altman Z-score (1968; as cited in Zang 2012), measured at the beginning of the year:
𝑍𝑆𝐶𝑂𝑅𝐸𝑡= 0.3 ( 𝑁𝐼𝑡 𝐴𝑠𝑠𝑒𝑡𝑡) + 1.0 ( 𝑆𝑎𝑙𝑒𝑠𝑡 𝐴𝑠𝑠𝑒𝑡𝑡) + ( 𝑅𝑒𝑡𝑎𝑖𝑛𝑒𝑑 𝐸𝑎𝑟𝑛𝑖𝑛𝑔𝑠𝑡 𝐴𝑠𝑠𝑒𝑡𝑡 ) + ( 𝑊𝑜𝑟𝑘𝑖𝑛𝑔 𝐶𝑎𝑝𝑖𝑡𝑎𝑙𝑡 𝐴𝑠𝑠𝑒𝑡𝑡 ) + [𝑆𝑡𝑜𝑐𝑘 𝑃𝑟𝑖𝑐𝑒𝑡∗ 𝑆ℎ𝑎𝑟𝑒𝑠 𝑂𝑢𝑡𝑠𝑡𝑎𝑛𝑑𝑖𝑛𝑔𝑡 𝑇𝑜𝑡𝑎𝑙 𝐿𝑖𝑎𝑏𝑖𝑙𝑖𝑡𝑖𝑒𝑠𝑡 ]
The marginal tax rate is measured using the Graham Method (1996a; as cited in Zang 2012). Costs associated with accrual-based earnings management are divided into scrutiny of auditors and regulators and the flexibility within a firm’s accounting systems. Following Zang (2012), costs associated with scrutiny of auditors and regulators are proxied using big-4 auditor presence. Second proxy described by Zang (whether the observation is before or after the SOX era) is left out as my sample does not include the implementation of SOX.
Again following Zang (2012), proxies associated with flexibility within a firm’s accounting systems are accrual-based earnings management in previous years (due to its reversible nature. I.e. accrual-based earnings management in the current period is constrained by that of previous periods) and the length of a firm’s operating cycle. Following Zang (2012) accrual-based earnings management in previous periods is measured using the Barton and Simko’s Method (2002; as cited in Zang 2012), being proxied by Net Operating Assets (NOA). Changes in net operating
18
assets reflect abnormal accruals, making it a good proxy for accrual-based earnings management (Zang 2012). As with audit tenure, this variable is used as an indicator, equaling 1 if net operating assets at the beginning of the year divided by lagged sales are above the median of the corresponding industry year and 0 if they are below.1 Length of operating cycle is measured as
the days receivable plus the days inventory less the days payable, at the beginning of the year (Dechow 1994; as cited in Zang 2012).
4.5 Firm leverage and the tradeoff between accrual-based and real earnings
management
According to Zang (2012), managers trade-off real versus accrual-based earnings management, based on their relative costliness. She argues that real activities earnings management decisions are made during the fiscal year, since they rely on true economic decisions that cannot be altered after they occurred. Accrual-based earnings management on the other hand can be exercised after the fiscal year has ended, and before the financial statements are published. Zang (2012) also states that, although managers make real earnings management decisions based on their expected outcomes, these outcomes can never be fully anticipated. Hence, managers are faced with unexpected outcomes of real earnings management by the end of the fiscal year. Zang argues that managers increase their accrual-based earnings management, based on the unexpected real earnings management. If unexpected real earnings management is high, accrual-based earnings management is low, and vice versa.
Following Zang (2012), I use the following formula to estimate unexpected real earnings management:
𝑅𝑀𝑡 = 𝛽0+ ∑ 𝛽𝑘 1,𝑘𝐶𝑜𝑠𝑡 𝑜𝑓 𝑅𝑀𝑘,𝑡 + ∑ 𝛽𝑙 2,𝑙𝐶𝑜𝑠𝑡 𝑜𝑓 𝐴𝑀𝑙,𝑡+ ∑ 𝛽𝑚 3,𝑚𝐶𝑜𝑛𝑡𝑟𝑜𝑙𝑚,𝑡+ µ𝑡 (6)
RMt represents the levels of earnings management as calculated in formula 3 and 4. Their aggregated amount represents RMt. Both Cost of RM and Cost of AM are explained in the following section, as well as the control variables. For formula (6) β2 is expected to be negative, because in firms that use real earnings management (i.e. RM is positive), the costs of accrual-based earnings management are expected to be high. Consequently, B1 is expected to be low, as
19
the costs for using real earnings management are expected to be low in firms using real earnings management. The residual of formula (6) represents unexpected real earnings management (Unexpected RM).
In measuring the effect that firm leverage has on the trade-off between accrual-based and real earnings management, a moderated multiple regression is done using Unexpected RM as the explaining independent variable, FL as the moderating independent variable, and AM as the dependent variable. The formula is as follows:
𝐴𝑀𝑡= 𝛽0+ 𝛽1𝑈𝑛𝑒𝑥𝑝𝑒𝑐𝑡𝑒𝑑 𝑅𝑀𝑡+ 𝛽2𝐹𝐿𝑡+ 𝛽3(𝑈𝑛𝑒𝑥𝑝𝑒𝑐𝑡𝑒𝑑 𝑅𝑀𝑡∗ 𝐹𝐿𝑡) + 𝛽4𝐶𝑜𝑛𝑡𝑟𝑜𝑙𝑠𝑡+ 𝜀𝑡
(7)
When AM is positive I expect Unexpected RM to be negative and FL to be positive. This is due to the fact that I expect managers to engage in more accrual-based earnings management as firm leverage increases, and the substituting relationship between accrual-based earnings management and real earnings management causes real earnings management to drop. I expect the moderating variable (Unexpected RM * FL) to decrease with FL, as I expect that the decrease in real earnings management is stronger than the increase in accrual-based earnings management, due to the greater scrutiny caused by private debtholders.
4.6 Control variables
Following Zamri et al (2013), who researched the effect of firm leverage on earnings management, I include the following control variables for the formulas 2 and 5. Return on Assets (ROA) is used to control for the negative relationship between earnings management and firm performance. SIZE is used to control for the relationship between earnings management and its relative size. Finally, I control for big 4 auditor presence (Zamri et al 2013).
Following Zang (2012), I use ROA in formula 6, calculated by using net income for the rolling four quarters ending with the third of the year, to control for firm performance. I use market-to-book-ratio to control for a firm’s growth rate. In formula 7 I use firm size to control for the increased possibilities firms have for earnings management when they are larger. In addition, I use big 4 auditing to control for scrutiny of auditors.
20
5
Empirical results
5.1
Descriptive statistics
From the descriptive statistics in table 1 we can see that the firm leverage is relatively high, indicating a high debt to equity ratio. The high standard deviation suggests a large dispersion, possibly among different industries or company sizes. Accrual-based earnings management has a mean of 0.62, slightly higher than that reported by Zang (2012). Real earnings management has a mean of 0.56, and a standard deviation of 2.4, which is also slightly higher than that reported by Zang (2012). These differences can be possibly attributed to the difference in time period and the fact that levels of earnings management have been increasing over time.
Unexpected real earnings management has a mean of 13.8 and a relatively low standard deviation of 3.2. This indicates that across industries and different company sizes the extent to which the consequences of real earnings management are anticipated does not differ greatly. Remarkable is the negative return on assets of -2.8, which could be attributed to the large impact of the financial crisis. Big 4 presence has a mean of 0.47, indicating that 47% of all companies observed are audited by a big 4 firm. The market-to-book ratio has a mean of 76.3 which suggests that the book value of firms is considerably higher than the market value. This could again probably be attributed to the financial crisis and the drop in stock prices.
Table 1
Descriptive statistics
Variable Mean Std. Deviation
Firm Leverage 4.2 413
Accrual-based Earnings management 0.62 2.4
Real Earnings Management 0.56 3.7
Unexpected Real Earnings Management 13.8 3.2
Firm leverage & Unexpected Real earnings management -0.2 -0.4
Return on Assets -2.8 86.4
Size 14523 119246
Big 4 presence 0.47 0.5
21
5.2
Research model 1: firm leverage and accrual-based earnings management
5.2.1 Correlation matrix analysis
The correlation matrix for research model one is shown in table 2. The significance levels are derived from the corresponding coefficients. Correlations are considered significant if their corresponding P values are less than 0.1. For analytical purposes they are further divided into three different categories: <0.1, <0.05, and <0.01.
As can be seen from table 2, return on assets (roa) is significantly and inversely correlated to accrual-based earnings management (am), although the correlation is not very strong (0.111). This indicates that better performing companies tend to engage less in accrual-based earnings management. Against my expectations, big 4 auditor presence (aud) and accrual-based earnings management (am) are positively correlated (0.011). Firm leverage (fl) and total assets (ta) are negatively correlated (-0.43). This inverse relationship suggests that smaller companies are more highly leveraged than larger companies. Return on assets (roa) and total assets (ta) are both slightly correlated with big 4 presence (aud). This relationship can be attributed to the fact that both more profitable and larger companies have the financial means to hire big 4 firms.
Table 2
Pearson correlation matrix: firm leverage and accrual-based earnings management
am fl roa ta aud am 1 fl 0 1 roa -0.111 *** 0 1 ta 0.001 -0.043 *** 0.004 1 aud 0.011 * 0.008 * 0.04 *** 0.09 *** 1
* correlation is significant at 0,1 (2-tailed) ** correlation is significant at 0,05 (2-tailed) *** correlation is significant at 0,01 (2-tailed)
5.2.2 Regression analysis and review of hypotheses
This section is divided into the following sections. Firstly, the hypothesis associated with this research model is discussed, after which the dependent variable is discussed and under which conditions the hypothesis will be either supported or rejected. In addition, the variables that did not prove significant will be discussed in this section.
22
The dependent variable in the first research model is accrual-based earnings management (am) and the independent variable is firm leverage (fl). The control variables are big 4 presence (aud), total assets (ta) and return on assets (roa).
The hypothesis associated with this research model is: firms with high leverage increase their accrual-based earnings management. The hypothesis is supported if the variable firm leverage is significant with a P-value of <0.1 and if the coefficient is positive. Alternatively, the hypothesis is rejected if these criteria are not met. Table 3 shows the output of the regression analysis on firm leverage and its relationship to accrual-based earnings management. The positive coefficient of firm leverage (fl) has an insignificant P-value (0.997). as a result H1 is rejected and there is no evidence that firm leverage is significantly positively related to accrual-based earnings management.
The only variable that is significantly and negatively correlated with accrual-based earnings management is return on assets (roa) (P 0.000). This proves that better performing companies (as measured by return on assets) engage less in accrual-based earnings management. This relationship could be explained by the fact that as companies perform better, their incentives to engage in accrual-based earnings management disappear.
The directions of all variables are as predicted. Nevertheless, besides return on assets, none of the variables are significantly correlated with accrual-based earnings management, and as a result no conclusions can be drawn upon them. Most notable is the absence of a significant relationship between firm leverage and accrual-based earnings management. This suggests that levels of accrual-based earnings management are not influenced by the debt-to-equity ratio of a firm.
23
Table 3
Regression analysis: firm leverage and accrual-based earnings management Variables Predicted sign Value of coefficients P-value
fl + 1.302 0.997
roa - -0.267 0.000 ***
ta + 3.901 0.974
aud - -2.591 0.368
Dependent variable: am * correlation is significant at 0.1 (2-tailed) R2: 0.12 ** correlation is significant at 0.05 (2-tailed) R2 adjusted: 0.12 *** correlation is significant at 0.01 (2-tailed)
5.3
Research model 2: firm leverage and real earnings management
5.3.1 Correlation matrix analysis
The correlation matrix of the second research model (firm leverage and real earnings management) is shown in in table 4.
A positive and significant relationship exists between big 4 presence (aud) and real earnings management (rem), with a corresponding coefficient of 0.031. Although this relationship is relatively weak, it is significant with a P-value of <0.05. Surprisingly this implies that, within the sample, when companies are audited by a big 4 firm, they tend to engage in more real earnings management. This could possibly be due to the tradeoff between accrual-based and real earnings management. Real earnings management is more difficult to detect and managers might switch from accrual-based earnings management if scrutiny of auditors becomes higher. Apart from the ones already discussed in the previous section, all other relationships are insignificant.
Table 4
Pearson correlation matrix: firm leverage and real earnings management
rem fl roa Ta aud
rem 1
fl -0.001 1
roa 0.008 0 1
ta 0.008 -0.043 *** 0.004 1
aud 0.031 ** 0.008 ** 0.04 *** 0.09 *** 1
* correlation is significant at 0.1 (2-tailed) ** correlation is significant at 0.05 (2-tailed) *** correlation is significant at 0.01 (2-tailed)
24
5.3.2 Regression analysis and review of hypothesis
The dependent variable in the second research model is real earnings management (rm) and the independent variable is firm leverage (fl). The control variables are return on assets (roa), total assets (ta) and big 4 presence (aud).
The hypothesis (h2) associated with the second research model is: firms with high leverage decrease their real earnings management activities. This hypothesis is supported if the variable firm leverage is significant with a P-value of <0.1 and if its coefficient is positive. Alternatively, the hypothesis is rejected if one of these criteria is not met. Table 5 shows the output of the regression analysis on firm leverage and its relationship real earnings management. The positive coefficient of firm leverage (fl) has an insignificant P-value (0.947). as a result H2 is rejected and there appears to be no evidence that firm leverage is significantly positively related to real earnings management.
The only variable with a significant P-value (0.032) is big 4 presence (aud). Remarkable though is the positive direction of the coefficient, while a negative direction was predicted. This indicates that companies audited by big 4 firms engage more in real earnings management than companies that are audited by smaller firms. As described in the aforementioned section, this phenomenon is possibly due to the illusive nature of real earnings management, making it harder to detect. As audit quality increases with the presence of a big 4 auditor, managers might switch from accrual-based earnings management to real earnings management.
None of the other variables have a significant P-value (P-value <0.1), including the main independent variable firm leverage (fl). Although the direction of the coefficient (-4.210) is as predicted, the P-value (0.947) is not significant. Remarkably, the coefficient of return on assets (roa) is slightly positive, indicating that better performing companies (as measured by their return on assets) engage more in real earnings management. Nevertheless, this relationship is insignificant, indicating that this relationship is not causal and most likely due to circumstantial factors.
25
Table 5
Regression analysis
Variables Predicted sign Value of coefficients P-value
fl - -4.210 0.947
roa - 0.002 0.620
ta + 8.161 0.714
aud - 1.140 0.032 **
Dependent variable: rm * correlation is significant at 0.1 (2-tailed) R2: 0.001 ** correlation is significant at 0.05 (2-tailed) R2 adjusted: 0.000 *** correlation is significant at 0.01 (2-tailed)
5.4 Research model 3: firm leverage and the tradeoff between accrual-based
and real earnings management
5.4.1 Correlation matrix analysis
The correlation matrix for research model three is shown in table 7. Unexpected real earnings management (unexp rm) is significantly (P-value <0.01) and positively correlated to based earnings management (am) (0.129). This underpins the theory that managers use accrual-based earnings management to substitute real earnings management, accrual-based on real earnings management outcomes (unexpected real earnings management). Return on assets (roa) is also positively correlated to accrual-based earnings management (P-value <0.01), at 0.111. market-to-book ratio (mktb) is positively and significantly (P-value <0.01) and positively correlated to accrual-based earnings management (0.02). indicating that as a company’s book value increases, relative to its market value, accrual-based earnings management increases. This relationship could possibly be explained by managers’ efforts to boost stock prices by manipulating financial statements. Furthermore, big 4 presence (aud) is significantly (P-value <0.1) correlated to accrual-based earnings management (0.001).
Return on assets is positively and significantly (P-value <0.01) correlated to return on assets (0.124), indicating that the unexpected outcome of real earnings management increases as firm performance (measured by return on assets) increases. In addition, unexpected real earnings management is positively and significantly (P-value <0.01) correlated to both market-to-book ratio (0.167) and big-4 presence (0.23). Firm leverage is positively and significantly (P-value <0.01) correlated to total assets (0.43) and also positively and significantly (P-value <0.1) to big 4 presence. Return on assets and market-to-book ratio are positively and significantly (P-value
26
<0.01) correlated (0.089) and just as big 4 presence to return on assets (P-value <0.01) (0.04). market-to-book ratio and big-4 presence are positively and significantly (P-value <0.01) correlated (0.022). Finally, total assets is positively and significantly (P-value <0.01) correlated to big 4 presence. As most of these correlations do not add to answering the research question or the hypothesis, they are not discussed in further detail.
Table 6
Pearson correlation matrix: firm leverage and the tradeoff in earnings management
am unexp rm fl unexp*fl roa mktb ta aud
Am 1 unexp rm 0.129 *** 1 fl 0 -0.011 1 unexprem*fl -0.006 0.053*** 0.995 *** 1 Roa 0.111 *** 0.124 *** 0 -0.007 1 Mktb 0.02 *** 0.167 *** 0 0.012 0.089 *** 1 Ta 0.001 -0.016 0.43 *** 0.007 0.004 -0.003 1 Aud 0.011 * 0.23 *** 0.008 * 0.023 0.04 *** 0.022 *** 0.09 *** 1 * correlation is significant at 0.1 (2-tailed)
** correlation is significant at 0.05 (2-tailed) *** correlation is significant at 0.01 (2-tailed)
5.4.2 Regression analysis and review of hypothesis
The dependent variable in this third research model is accrual-based earnings management (am). The independent variables are unexpected real earnings management (unexp rm), the moderating variable firm leverage (fl), and finally the interacting variable (unexp rm*fl). The control variables are return on assets (roa), market-to-book ratio (mktb), total assets (ta) and big 4 presence (aud). The results of the regression analysis can be found in table 7.
The hypothesis (h3) associated with the third research model is: firms with high leverage that show increased accrual-based earnings management use this as a substitute for their real earnings management activities. This hypothesis is supported if the interacting variable (unexp rm*fl) is significant with a P-value of <0.1 and if its coefficient is positive. Alternatively, the hypothesis is rejected if one of these criteria is not met. The positive coefficient of the interacting variable (unexp rm*fl) has an insignificant P-value (0.578). as a result H3 is rejected and there
27
appears to be no evidence that firm leverage plays a moderating role in the tradeoff between accrual-based earnings management and unexpected real earnings management.
Unexpected real earnings management (unexp rm) has the highest coefficient (13.835) and there seems to be a significant causal relationship with accrual-based earnings management (P-value <0.01). This is consistent with Zang (2012), who showed that managers base levels of accrual-based earnings management on the unexpected outcomes of their real earnings management activities. Return on assets again seems to have a significant (P-value <0.01) and causal relationship with accrual-based earnings management, with a coefficient of 0.241. Finally, the market-to-book ratio has a significant (P-value <0.05) effect on accrual-based earnings management, with a coefficient of 0.026). However, the findings of these regression should be interpreted with caution, due to the high risk of multicollinearity in moderated regression analysis, leading to a higher standard error.
Table 7
Regression analysis
Variables Predicted sign Value of coefficients P-value
unexp rm - 13.835 0.000 *** fl - 0.094 0.576 unexp rm*fl + 0.222 0.579 Roa - 0.241 0.000 *** Mktb + 0.004 0.026 ** Ta + 2.355 0.627 Aud - 7.924 0.439
Dependent variable: am * correlation is significant at 0.1 (2-tailed) R2: 0.029 ** correlation is significant at 0.05 (2-tailed) R2 adjusted: 0.025 *** correlation is significant at 0.01 (2-tailed)
28
7
Conclusion
This paper aims to answer the research question: to what extent is the trade-off between accrual-based earnings management and real earnings management influenced by firm-leverage? The research can be divided into three topics: the relationship between a firm’s debt leverage and its accrual-based earnings management activities (1), the relationship between a firm’s debt leverage and its real earnings management activities (2), and finally, the effect of firm leverage on the way managers substitute one form of earnings management for the other (3).
In accounting research two main forms of earnings management exist: accrual-based earnings management and real earnings management. Recent papers have shown that managers trade-off one form of earnings management for the other. Real earnings management activities are restricted to the ending of the fiscal year, after which decisions cannot be altered. Only after closing of the fiscal year can managers often assess the effects of their real earnings management activities. A discrepancy between their planned earnings management activities and the actual outcome is defined as unexpected real earnings management. Accruals can be allocated to a fiscal year, after it has actually ended. As a result, managers use accrual-based earnings management to substitute for their real earnings management activities. Accrual-based and real earnings management have different implications for investors and debtors. Since real earnings management has direct negative implications for a company, increased scrutiny from debtors (for example through debt covenants) might cause managers to substitute real earnings management for accrual-based earnings management.
The results indicate that no direct relationship exists between both accrual-based and real earnings management and firm’s debt leverage. In addition, firm leverage does not seem to have a moderating effect on the way managers substitute one form of earnings management for the other. This suggests that managers do not change their earnings management strategies as their firm’s debt leverage increases.
This research has several limitations. First of all, the theory of this research assumes that managers trade-off earnings management incentivized by scrutiny from debtors, while the
29
prevailing method used to research the trade-off in earnings management assumes that managers trade-off earnings management based on the unexpected outcomes of their real earnings management activities. Secondly the use of moderated regression analysis brings about the issue of multicollinearity, which decreases the accuracy of my regression. Dealing with multicollinearity is beyond the scope of this paper.
I suggest that future research not only focus on debt-to-equity ratio, but search for other ways of measuring scrutiny from debtors. Examples are presence of debt covenants, or long-term debt as opposed to short-term debt. With different proxies for scrutiny from debtors, researchers might be better able to link firm leverage to the trade-off between accrual-based and real earnings management.
30
8
Bibliography
Healy, P. M., and Wahlen, J. 1999. A review of the earnings management literature and its implications for standard setting. Accounting Horizons 13 (4): 365–383.
Ball, R., and Shivakumar, L. 2005. Earnings quality in UK private firms: comparative loss recognition timeliness. Journal of Accounting and Economics 39 (1): 83-128.
Bharath, S. T. 2008. Accounting quality and debt contracting. Accounting Review 81 (1): 1-28. Burgstahler, D. C., Hail, L., Leuz, C. 2006. The Importance of Reporting Incentives: Earnings Management in European Private and Public Firms. International Journal of Accounting 81 (5): 983-1016.
Bushee, B. 1998. The influence of institutional investors on myopic R&D investment behavior.
The Accounting Review 73 (3): 305–333.
Cohen, D., and Zarowin., P. 2010. Accrual-based and real earnings management activities around seasoned equity offerings. Journal of Accounting and Economics 50 (1): 2–19.
Cohen, D., Dey, A., and Lys, T. 2008. Real and accrual-based earnings management in the pre- and post-Sarbanes-Oxley period. The Accounting Review 83 (3): 757–787.
International Accounting Standards Board. 2012. The Conceptual Framework of Financial Reporting. Accessed via www.iasplus.com/en/standards/other/framework, on Tuesday 12th of May 2015.
Dechow, P., and R. Sloan. 1991. Executive incentives and the horizon problem: An empirical investigation. Journal of Accounting and Economics 14 (1): 51–89.
Dechow, P., Hutton, A., and Sloan, R. 1996. Economic consequences of accounting for stock-based compensation. Journal of Accounting Research 34: 1-20
Dhaliwal, D. 1980. The Effect of the Firm's Capital Structure on the Choice of Accounting Methods. The Accounting Review 55 (1): 78-84.
Graham, J., Harvey, C., and Rajgopal, S. 2005. The economic implications of corporate financial reporting. Journal of Accounting and Economics 40 (1-3): 3–73.
Gunny, K. A. 2010. The relation between earnings management using real activities manipulation and future performance: Evidence from meeting earnings benchmarks. Contemporary Accounting
Research 27 (3): 855–888.
Healy, P., and Palepu, K. 1990. Effectiveness of accounting-based dividend covenants. Journal of
31
Healy, P. M. 1985. The effect of bonus schemes on accounting decisions. Journal of Accounting
and Economics 7 (1-3): 85–107.
Jones, J. 1991. Earnings management during import relief investigations. Journal of Accounting
Research 29 (2): 193–228.
Mande, V., File, R.G. et Kwak, W. 2000. Income smoothing and discretionary R&D expenditures of Japanese firms. Comptemporary accounting research 17 (2): 263-302.
Roychowdhury, S. 2006. Earnings management through real activities manipulation. Journal of Accounting and Economics 42 (3): 335–370.
Sloan, R. 1996. Do Stock Prices Fully Reflect Information in Accruals and Cash Flows about Future Earnings? The Accounting Review 71 (3): 289-315.
Sweeney, P. 1992. Debt-covenant violations and managers’ accounting responses. Journal of
Accounting and Economics 17: 281-308.
Teoh, S., Welch, J., and Wong, T. 1998. Earnings management and the underperformance of seasoned equity offerings. Journal of Financial Economics 50 (1): 63–99.
Xie, H. 2001. The Mispricing of Abnormal Accruals. The Accounting Review 76: 357-373.
Zamri, N., Rahman, R., and Mohd Isa, N. 2013. The impact of Leverage on Real Earnings Management. Procedia economics and finance 7: 86-95.
Zang, A. 2012. Evidence on the Trade-Off between Real Activities Manipulation and Accrual-Based Earnings Management. The Accounting Review 87 (2): 675-703.
Zmijewski, M., and Hagerman, R. 1981. An income strategy approach to the positive theory of accounting standard setting/choice. Journal of Accounting and Economics 3: 129-149.
