The Impact of Financial Analysts on Earnings Management : Empirical evidence from Swedish listed companies

(1)

The Impact of Financial Analysts

on Earnings Management

MASTER THESIS WITHIN: Business Administration NUMBER OF CREDITS: 30 ECTS

PROGRAMME OF STUDY: Civilekonom AUTHORS: Nicholas Morgan & Tim Roth

(2)

Master Thesis in Business Administration

Title: The Impact of Financial Analysts on Earnings Management Authors: Nicholas Morgan & Tim Roth

Tutor: Argyris Argyrou Date: 2020-05-18

Keywords: Accrual-based earnings management, analyst coverage, analysts forecast, discretionary accruals, earnings management, real activities, real earnings management.

Abstract

Background: Existing literature (e.g., Burgstahler & Eames, 2006; Yu, 2008) is inconsistent

whether forecast consensus of earnings pressure managers to engage in earnings management or whether the additional layer of monitoring frightens managers from manipulating earnings. The study conjectures that analyst coverage has a negative impact on the occurrence of earnings management.

Purpose: The purpose of the study is to investigate whether companies listed on the Swedish

stock exchange engage in earnings management by using discretionary accruals or real activities but also to identify whether analyst coverage has an impact on managers decision to engage in earnings management.

Method: The study investigates 568 company-year observations derived from 71 companies

between the time period 2009-2016. The study uses the modified Jones model by Kothari, Leone and Wasley (2005) and Kim, Park and Wier (2012) to estimate discretionary accruals, and Roychowdhury´s (2006) model to estimate real activities. To distinguish whether analyst coverage has an impact of earnings management, the study divides the sample into two groups depending on the number of financial analysts who monitors each company.

Conclusion: The study provides empirical evidence of earnings management on the Swedish

stock exchange and that analyst coverage has a negative impact on managers decision to engage in earnings management. However, the study concludes that earnings management still occurs regardless of analyst coverage and that managers prefer real activities over discretionary accruals since it is more difficult to detect. From a societal perspective this could raise concerns, suggesting that additional gatekeepers or standards should be implemented to reduce managers ability to mislead stakeholders.

(3)

Acknowledgement

The authors of this thesis would like to express their gratefulness and appreciation to everyone who has contributed and inspired with their expertise and suggestions along the way towards a complete thesis.

The authors would like to especially thank their supervisor Argyris Argyrou for his encouragement, flexibility, guidance and wide expertise within the accounting field. Argyris has expressed his kindness and flexibility to adjust his schedule to the authors’ but also to the situation that Covid-19 pandemic has caused.

Last, but absolutely not least, the authors would like to thank each other for an excellent and rewarding collaboration, resulting in an experience and a complete thesis to be proud of.

Jönköping International Business School May 2020

Nicholas Morgan Tim Roth

(4)

List of equations

Equation 1 – Discretionary accruals... 30

Equation 2 – Total accruals ... 30

Equation 3 – Industry-specific parameters ... 31

Equation 4 – Non-discretionary accruals ... 31

Equation 5 – Real earnings management ... 33

Equation 6 – Normal levels of cash flow from operations ... 34

Equation 7 – Abnormal levels of cash flow from operations ... 34

Equation 8 – Normal levels of production costs ... 35

Equation 9 – Abnormal levels of production costs ... 35

Equation 10 – Normal levels of discretionary expenses ... 36

Equation 11 – Abnormal levels of discretionary expenses ... 37

Equation 12 – T–statistics ... 38

Equation 13 – Normal levels of cost of goods sold... 72

Equation 14 – Normal inventory growth... 72

List of figures

Figure 1 – A hypothetical value function ... 16

List of tables

Table 1 – Sample size... 25

Table 2 – Industries ... 26

Table 3 – Descriptive statistics variables ... 27

Table 4 – Descriptive statistic for estimating mean coefficients ... 42

Table 5 – Model coefficients for N_CFO, N_DISEXP, N_PROD and TA ... 43

Table 6 – Summary statistics of output for DA and REM ... 44

Table 7 – Correlation Matrix ... 46

Table 8 – Descriptive statistics for analyst coverage ... 47

Table 9 – T-statistic for analyst coverage ... 48

Table 10 – T-statistic for DA and REM group specific ... 49

Table 11 – Summary statistics of output for DA including normal-values ... 73

(6)

1 Introduction

In this section the study introduces earnings management and provides a background of how managers can influence reported earnings through real activities and accrual-based earnings management.

1.1 Background

Earnings management occurs when managers use their discretion to actively influence earnings by utilizing the flexibility of accounting choices or through operational activities. Managers have many incentives to engage in earnings management, but this study solely focuses on capital market motivations. This incentive stems from managers desire to reach earnings targets forecasted by financial analysts but also expectations from investors. Dechow (1994) explains that investors and analysts apply reported earnings, rather than current cash, as a gauge to estimate future cash flows. This is strengthened by Graham, Harvey and Rajgopal (2005) who found in their study that CFOs believe that earnings, not cash flow, are the key metric considered by financial analysts to forecast future performance. This means that earnings act as a powerful indicator for firm valuation but if earnings are manipulated, stakeholders could take decisions based on false economic premises.

Managers have two means for manipulating earnings, accrual-based earnings management and real earnings management. Manipulating earnings by revaluating transactions in accordance to regulations, standards and principals are known as accrual-based earnings management. These activities are characterised by changing accounting methods which do not have a direct impact on cash flow (Gunny, 2010). Examples of accrual-based earnings management includes delaying asset write-offs or through reversals of accruals known as cookie-jar reserves (Levitt, 1998). By contrast, real earnings management is characterized by manipulating earnings through operating activities that affects cash flows. An example of real earnings management is reduction of discretionary expenses such as research and development (R&D) to report lower costs resulting in higher reported earnings (Roychowdhury, 2006).

(7)

Earnings management has become a highly visible field in recent years due to the impact of a series of major accounting scandals (Yu, 2008). Enron is one of the most well-known bankruptcy in the world. The top management team of the company created special purpose entities to hide losses and liabilities that concealed the unpleasant truth (Elias, 2004). As Enron declared bankruptcy and collapsed, investors lost millions, employees lost their jobs and the public lost confidence in the process of financial reporting. Corporate scandals such as Enron and WorldCom led to the enactment of Sarbanes Oxley Act (SOX) in 2002 whose main purpose is to restore confidence in financial reporting and to protect investors interests (Cohen, Dey, & Lys, 2008).

Likewise, in Europe, the International Accounting Standard Board (IASB) introduced International Financial Reporting Standards (IFRS) in 2005. The adoption of IFRS has led to stricter disclosure policies, changes of acceptable accounting choices and other relevant accounting policies with the purpose of harmonizing and improving the accounting quality globally (Ferentinou & Anagnostopoulou, 2016). Favourable accounting choices acceptable under local legislation may no longer be applicable for companies to apply. However, to enhance transparency, all listed companies within EU are since 2005, required to prepare financial statements of higher quality and with greater value-relevance for external parties in accordance to IFRS (Regulation (EC) 1606/2002).

Aggressive financial reporting by managers have forced regulators and standards setters to combat managers opportunistic behaviour by changing regulations within accounting. Cohen et al. (2008) reveals in their study that accrual-based earnings management has declined after the implementation of SOX while real earnings management has increased. The reason behind the switch is due to the tightening of accounting standards by regulators resulting in fewer options for managers to apply accrual-based earnings management. A practical explanation behind the increase is connected to the difficulties for external actors to identify activities associated to real earnings management.

However, regulators and standard setters are not the only gatekeepers against earnings management, financial analysts could also have a significant impact on earnings management. Dyck, Morse and Zingales (2010) investigated fraud detection in the US

(8)

fraudulent financial reporting in their process of estimating future accounting figures. Sun and Liu (2011) regards analysts as an important intermediary between the capital market and companies since analysts collect, analyse and interpret information of a company which they later convey to the capital market. The conveyed information could decrease the information asymmetry between the top management team and owners which subsequently could decrease agency cost and earnings management (Jensen & Meckling, 1976).

By contrast, several previous studies have identified that financial analysts put pressure on managers to reach forecasted earnings thresholds (e.g., Graham et al., 2005; Zang, 2012). This pressure could force managers to engage in earnings management in order to meet or beat analysts’ forecasts (Graham et al., 2005). The ambiguity between existing literature advocates for two potential outcomes, the pressure effect and the monitoring effect.

1.2 Problem

The importance of meeting forecasted earnings imposes pressure on managers to undertake decisions which could have unfavourable consequences for the stakeholders. As a result, earnings management has become a highly researched topic in the accounting field. Despite the large amount of research within earnings management, the topic is still a fruitful area to investigate since prior literature provides inconclusive results leading to a gap between theory and practice.

Existing literature has paid little attention to answer if forecasted earnings by financial analysts have an impact on managers decision to engage in earnings management. The majority of prior literature measures accrual-based earnings management and thus misses an important cornerstone of earnings management, real earnings management (Fields, Lys, & Vincent, 2001). For example, Yu (2008) found a negative correlation between analyst coverage and accrual-based earnings management but does not incorporate real earnings management. Another problem with existing literature is that the majority selects companies that have barely reached forecasted earnings, suspect companies, when measuring financial analysts’ impact on managers decision to engage in earnings management (e.g., Burgstahler & Dichev, 1997; Roychowdhury, 2006; Zang, 2012). This

(9)

means that existing literature excludes companies that cannot meet earnings targets with earnings management activities but still manipulate the results downwards to simplify the process of meeting future earnings targets.

The purpose of earnings management is to mislead stakeholders about the underlying economic performance (Healy & Wahlen 1999). Agency, signaling, and prospect theory could help explain why managers want to mislead stakeholders and how financial analysts affects managers decision to engage in earnings management. Agency and signaling theory are both concerned with information asymmetry, and prior research has documented that an increased analyst coverage will decrease information asymmetry between owners and managers (Yu, 2008). As a result, managers will manipulate earnings less since the risk of being detected increases.

Prospect theory posits that decision-makers acquire value from gains and losses and assess options with respect to a reference point and not wealth or welfare (Kahneman & Tversky, 1979). Earnings targets could be considered as a natural reference point for managers and theoretically managers will do anything to meet or beat earnings targets regardless of information asymmetry as they are loss-averse (Zhang, Bartol, Smith, Pfarrer, & Khanin, 2008).

1.3 Purpose & research question

The purpose of the study is twofold. First, the study investigates whether, companies listed on the Swedish stock exchange engage in earnings management by using discretionary accruals or real activities. Further, the study examines whether analyst coverage has an impact on managers decision to engage in earnings management. The motivation behind the study stems from the inconclusive results in existing theory and literature so far presented.

(10)

2 Literature review

The purpose of this section is to present findings from existing literature within earnings management. The two major classifications of earnings management, accrual-based and real earnings management are introduced, followed by a review of financial analysts' impact on managers decision to engage in earnings management. The section ends with presenting the theoretical framework and hypotheses development.

2.1 Evidence on earnings management

Earnings management is defined as: “Earnings management occurs when managers use judgement 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.368).

Earnings management has received a lot of attention due to major corporate scandals such as Enron and WorldCom where accounting activities have damaged a company severely or in extreme cases led to its downfall. Regardless of the severe consequences that earnings management could lead to, evidence from a study by Graham et al. (2005) shows that managers feel that they choose the “lesser evil” when they engage in earnings management. The activities or accounting choices helps them to meet or beat financial analysts’ forecasted earnings which subsequently decreases the volatility of the share-price.

Failure to meet or beat financial analysts’ forecasted earnings, does not only create bad publicity of the company, it also creates direct monetary losses for the company and its shareholders in terms of downward price revisions (Brown, Hagerman, Griffin, & Zmijewski, 1987). Skinner and Sloan (2002) propose that there is an asymmetry between market response and earnings surprises. They highlight that the magnitude in a price response of a negative earnings surprise is significantly larger and more severe compared to the price response of a positive earnings surprise.

(11)

Managers have at least two options to meet or beat analysts’ forecasted earnings. They can either apply the traditional mechanism: earnings management which includes accrual-based and real earnings management, or they can spend resources on guiding analysts´ expectations downwards to avoid optimistic forecast (Matsumoto, 2002). Either direction taken entails costs. Guiding analysts is a current process that most likely leads to beneficial and achievable targets in the future. However, revising current expectations downwards could lead to negative stock price responses in the short run. The timing of revising is important. If it is conducted too early, the value of the stock may be lower for an extended period of time. Even though the risk of guiding analysts is high, the magnitude of cost is greater for actual negative earnings surprises than negative forecast revisions (Matsumoto, 2002).

Healy and Wahlen (1999) describes three main incentives to why managers engage in earnings management activities. First, Regulatory Motivations. Managers have incentives to manipulate earnings due to regulatory and anti-trust motivations. For example, managers that seeks governmental subsidies or protection, could have incentives to appear less profitable. Second, Contracting Motivations. To meet lending and compensation contracts, managers have incentives to engage in earnings management. For example, when dividends cannot be paid to the shareholders without violating the lending agreement, managers may manipulate earnings in order to satisfy both the banks and shareholders.

The study builds on the last incentive by Healy and Wahlen (1999), Capital Market Motivations. Healy and Wahlen (1999) describes that managers have incentives to manage earnings to meet financial analysts' consensus regarding forecasted earnings but also to meet investors' expectations which positively affect short-term share price. Thus, managers have many incentives to engage in earnings management activities. Some with more ethical intentions where managers act in the best interest of the current shareholders while other incentives include a more selfish and opportunistic aspect. However, one aspect is clear, managers must believe that stakeholders are unable to identify activities associated with earnings management or that the process of detecting is too costly to

(12)

The ethical dilemma of earnings management is highlighted by the former SEC Chairman Levitt, who argues that managers utilize the flexibility of accounting choices within financial reporting to affect earnings and describes it as “accounting hocus-pocus" (Levitt, 1998, p.14). The former chairman further explains that a serious issue of mistrust could arise between owners and managers as a consequence of earnings management. On the other hand, Gunny (2010) identified that earnings management activities could have beneficial outcomes for the company and its stakeholders. Evidence from her study suggests that there is a positive correlation between companies that engage in real earnings management to meet or beat earnings targets and future performance. However, future performance is measured by return on assets, which inevitably increases as earnings increases.

By contrast, existing literature have revealed that real earnings management may have a negative effect on future performance and cash flows (Ewert & Wagenhofer, 2005; Roychowdhury, 2006; Badertscher, 2011). Long-term firm value could, for example, be negatively affected by reductions in discretionary expenses (R&D) as a company might lose its competitive advantage against competitors if it does not develop new improved strategies or products. Interestingly, Graham et al. (2005) found that managers (78%) are willing to sacrifice long-term economic value to reduce potential short-term losses that otherwise a negative earnings surprise could create. Regardless if earnings management could be considered unethical and affect future performance negatively, evidence shows that managers are short-term oriented and believe that the benefits of manipulating earnings exceeds the costs (Graham et al., 2005).

2.2 Accrual-based earnings management

The methodology of manipulating earnings by changing accounting methods is known as accrual-based earnings management. The process consists of revaluating transactions with the objective to mislead stakeholders by illustrating that the company met its benchmark and is performing well (Healy & Wahlen, 1999). Managers are often required to rely on their own judgement when they prepare financial statements as local accounting legislation and IFRS sometimes provide several accounting methods which includes different valuation options. The flexibility of accounting choices could create costs for

(13)

stakeholders as the flexibility allows managers to behave opportunistically (Fields et al., 2001). For example, depreciation can be valued according to the straight-line method, the double decline balance, sum of years digits or units of production. The flexibility creates an option for managers to choose a method that can affect the desirable direction of earnings without any impact on the statement of cash flow. However, the flexibility of choice is only available a limited amount of times which may result in a nasty surprise for stakeholders once managers change back to the original valuation and the financial reality is revealed.

Levitt (1998) mentions a few popular ways of misleading stakeholders through accrual-based earnings management, for example, big bath and cookie jar reserves. In bad years where poor results are inevitable, managers could take a big bath. This includes manipulating the results downwards and thus, under-report earnings (make results appear even worse) which subsequently simplifies the process of meeting future earnings forecasts and make future results appear enhanced (Kirschenheiter & Melumad, 2002). On the other hand, cookie jar reserves allow managers to meet or beat earnings targets by manipulating results both downwards and upwards. In a year with weak financial performance the cookie jar reserve could be utilized to boost earnings by reducing expenses through reversals of accruals and reserves, given that earnings targets are within reasonable limits. Adversely, in a year with strong financial performance, the cookie jar reserve could be utilized to reduce earnings by making large accruing expenses and overstating reserves (Kokoszka, 2003). However, existing literature (Dechow, Sloan, & Sweeney, 1995; Dechow, Kothari, & Watts, 1998; Barth, Cram, & Nelson, 2001) suggests that current and past performance are a precondition in order for managers to manipulate the results upwards through accruals-based earnings management.

Existing literature explains that accrual-based earnings management is more costly in the short-run since the probability of being scrutinized and detected by external actors is higher compared to real earnings management. However, by incorporating the long-run perspective, Moradi, Salehi and Zamanirad (2015) argues that the process of applying accrual-based earnings management is less costly for both owners and managers since the

(14)

process does not affect cash flow, nor future operating performance. Additionally, managers have the ability to affect payable bonuses through accrual-based earnings management since discretionary accruals can be influenced after the fiscal year but before the earnings announcement. Thereby, accrual-based earnings management is an option for managers to manipulate earnings without long-term operating consequences.

Compensation plans were initially thought to align managers interests with owners, resulting in a beneficial relationship between the two parties with a lower agency cost of monitoring managers. Nonetheless, managers incentives of manipulating reported earnings could illustrate that their interest of achieving current period financial reporting targets is greater than receiving a larger bonus in the future, indicating that managers are short-term oriented and incentivized by continuous improvements of self-interest (Moradi et al., 2015). The process of accrual-based earnings management is limited by available accounting choices implemented by IFRS and national legislation. At some point, manipulating earnings through accrual-based earnings management becomes inefficient because changes would not lead to the desired earnings. Hence, managers face no other alternative than manipulate earnings through real earnings management which has an impact on future cash flows and could impose long-term consequences.

2.3 Real earnings management

Accrual-based earnings management has received a lot of attention in the past but recent studies have also provided evidence of the existence of real earnings management (Bartov 1993; Bens, Nagar, & Wong, 2002; Roychowdhury 2006; Zang 2012). Dechow and Sloan (1991) found evidence that managers reduce R&D expenses near the end of the fiscal year in order to accelerate short-term earnings. Research conducted by Baber, Fairfield and Haggard (1991) and Bushee (1998) display consistent evidence of reductions of discretionary expenses to meet or beat forecasted earnings. A more recent study conducted by Roychowdhury (2006) reveals that stakeholders are misled to believe that performance targets are met by normal business operations while the reported earnings is actually achieved by real activities manipulation.

(15)

Roychowdhury (2006) detects that managers manipulate earnings by three major real activities. First, managers increase sales by offering large discounts or generous credit terms (cash flow from operations). Second, managers deviate from market demand and overproduce goods, resulting in higher overhead cost for inventory but a lower cost of goods sold (production cost). Lastly, to improve margins, managers reduce costs related to marketing, maintenance and R&D (discretionary expenses). The common denominator to why expenses related to R&D are often manipulated is because the invested capital does not generate immediate revenues or income which results in managers taking precautionary decisions whether it is worth investing in or not.

Graham et al. (2005) concluded from their survey that managers prefer real earnings management over accruals-based earnings management as a method to manage reported earnings with the purpose of meeting earnings targets forecasted by financial analysts. Moreover, the survey reveals that 80 % of the managers would make reductions in advertising and R&D but also that 55,3% of the survey participants would delay or reject a project in order to meet earnings targets. The results imply that managers have a greater incentive to meet short-term targets and would therefore sacrifice long-term values which a project may generate for the company in the future.

There are at least two explanations for why managers have greater incentives to manage earnings through real activities than discretionary accruals (Roychowdhury, 2006). First, the activities associated with real earnings management does not significantly deviate from normal business practices, indicating that the probability of being detected by either an auditor or a regulator is lower than managing earnings by accrual-based activities (Roychowdhury, 2006). Second, managing earnings by solely applying different accounting choices (accruals) alone is risky due to the probability of being scrutinized and detected. The benefits associated with real activities is confirmed by Zang (2012) who emphasizes that managers prefer applying real earnings management over accrual-based earnings management. However, evidence from her study suggest that there is a relative trade-off between accrual-based and real earnings management based on its relative costs and that the two strategies of managing earnings have a direct substitution.

(16)

earnings management is present, apply both methods in order to meet or beat earnings targets due to the timing of applying them.

At some point companies may face no other alternative than to report losses in their annual reports. Managers may have applied real activities to manage earnings but still experience a shortfall between actual earnings and forecasted targets. Even though real activities would be able to close the gap, the process can only be applied during the fiscal period (Cohen et al., 2008). Likewise, accrual-based earnings management which can be applied subsequent to the fiscal year, may be an unavailable option since it has already been utilized to the maximum level (Cohen et al., 2008).

2.4 Analyst coverage

Signals of accounting information normally travels from the company to individuals, who based on the information, undertake certain decisions. Sometimes the information channel is disrupted by a third party, financial analysts who validates the information published by managers and republish it as analytical forecast to the market. Financial analysts could have a significant impact on earnings management for several reasons. Yu (2008) describes that unlike managers main purpose of protecting current shareholders interest, financial analysts are expected to provide financial information to a broader audience including current and potential shareholders but also other market participants. Yu (2008) further explains that their resources in terms of education, relationships to managers and their knowledge of the industry, means that analysts possess important characteristics of monitoring and scrutinizing financial information. Since their process is conducted on a regular basis, irregularities in financial statements and management behaviour are most likely quickly identified and questioned.

Yu (2008) found in her study that earnings management is less frequent in firms with high analyst coverage, indicating that analysts play an important intermediary role between investors and managers. Healy and Palepu (2001) found that financial analysts collect private information which could help owners detect managerial misconduct. This indicates that besides auditors, financial analysts play an important role as external monitors and provide an additional layer of scrutiny on the financial reporting process

(17)

which decreases the information asymmetry and potentially managers tendency to engage in earnings management.

On the other hand, several prior studies have documented the importance of meeting forecasted earnings created by financial analysts suggesting that managers are pressurized to engage in earnings management (e.g., Degeorge, Patel, & Zeckhauser, 1999; Graham et al., 2005; Roychowdhury, 2006; Gunny, 2010; Zang, 2012). In a survey by Graham et al. (2005), 90% of 401 managers describes that financial analysts and institutional investors are key groups of estimating a company’s share price. Degeorge et al. (1999) found that managers attach great importance of reaching financial analysts forecast consensus and that managers will manipulate earnings upwards if they slightly miss the earnings target. However, managers do not only have incentives to manage earnings upwards. In a year with strong financial performance, managers have incentives to manage earnings downwards and thus, hide earnings for future years in order to hold forthcoming earnings forecasts steady (Graham et al., 2005).

Interestingly, financial analysts themselves are subjected to internal and external pressures which could; influence their incentives, diminish their role as external monitors and their ability to guard against earnings management activities (Yu, 2008). The forecasts and recommendations could be influenced by important clients, the employer and competitors. For example, existing literature (Michaely & Womack, 1999; Dechow, Hutton, & Sloan, 2000) found that financial analysts are less inclined to share unfavourable opinions regarding forecasted earnings of companies heavily invested in by their clients. To act in best interest of their clients, analysts may even try to convince managers to manipulate earnings in order to meet preannounced forecasted earnings. This kind of behaviour illustrate that analysts have incentives to satisfy their own interest by satisfying their clients.

Thus, in line with Yu (2008), the study identifies two potential outcomes of analyst coverage: first, increased analyst coverage will decrease managers tendency to engage in earnings management due to the decreased information asymmetry. Second, increased

(18)

2.5 Theoretical framework

In this section the study introduces agency theory, signaling theory and prospect theory. The purpose is to explain why managers want to mislead stakeholders and how financial analysts affects managers decision to engage in earnings management from a theoretical perspective.

2.5.1 Agency theory

In larger companies, where ownership and management are separate, information asymmetry creates (agency) costs for owners to monitor managers. The agency theory explains the problems that arise in the relationship between owners and managers. Jensen and Meckling (1976) describes that owners (principals) assign authority to managers (agents), entrusting them to run the daily operations without being continuously monitored. Nevertheless, assuming that both parties of the relationship are utility maximisers, there is a risk that the agents will undertake decisions in their own self-interest (Jensen & Meckling, 1976).

Actions that managers undertake to meet or beat earnings targets could misinform stakeholders regarding the company’s financial position and firm value. This could, for example, make current and potential shareholders take decisions based on false economic premises, thereby earnings management is an agency cost (Xie, Davidson, & DaDalt, 2003; Zahra, Priem, & Rasheed, 2005). Contrariwise, Dechow, Sloan and Sweeney (1996) argues that firm value will be reduced on the stock market if earnings management is suspected. Thus, managers have incentives to hide certain activities to not damage the company in the short run, but current and potential owners remain misled about the economic situation.

Eisenhardt (1989) explains that one problem with agency relationship is that the principals do not always know when their interest is put aside as information asymmetry exits between the two parties. However, monitoring and controlling that the agents are behaving in line with the implicit contract is problematic or expensive for principals to solve. Another problem is that agents and principals have different attitudes towards risk which implies that agents could take actions that the principals do not approve

(19)

(Eisenhardt, 1989). Financial analysts could reduce the information asymmetry between owners and managers which subsequently leads to lower engagement in earnings management since managers incentive of misleading shareholders diminishes.

Alternatively, owners can mitigate mentioned problems by spending resources on bonding and monitoring activities which are expenses related to agency costs. Spending resources on agency costs could motivate agents to act in the best interest of the principals (Jensen & Meckling, 1976). For example, managers might be offered stock options and thereby become an owner of the company (Jensen & Meckling, 1976). Spending resources on agency cost could be appropriate for companies with low analyst coverage to reduce potential high levels of information asymmetry. Contrary, companies monitored by more financial analysts are expected to have lower levels of earnings management activity since the degree of information asymmetry should be lower when adding an additional external actor that monitors the company.

Agency theory is relevant for the study since it explains the agency relationship and the consequences of information asymmetry. If managers believe that they possess information different to the market and that shareholders are unable to identify activities related to earnings management, managers will manipulate earnings. When financial analysts collect, interprets and distribute financial information to the capital market, information asymmetry decreases which favours shareholders.

2.5.2 Signaling theory

Managing earnings in order to meet earnings target, both at company and market level, sends signals to stakeholders. The achieved accounting figures in the financial statements may send signals to the market that the firm is operating as normal or even better compared to previous years. This indicates that short-term targets are met and that managers have fulfilled their duty of managing the company. However, managers may possess information regarding long-term issues that is not yet available to stakeholders. Financial analysts could reduce the information asymmetry by sending signals of information that reveals future consequences.

(20)

The signaling theory is concerned with reducing information asymmetry between two parties (Spence, 2002). The sender of the information (company) must decide whether and how to share information to its stakeholders. Thresholds and other financial figures in the financial statements are information that managers can influence before earnings of the fiscal year is announced. Normally, the transformation of information goes through a tangible information channel (statements) which means that no physical contact between the sender and the receiver exist (Spence, 2002). This may lead to a different interpretation of the available information by the receiver compared to the initial thoughts published by the sender. To solve the issue of information asymmetry, accounting standards and legislations requires managers to publish descriptive information as notes in the financial statements.

Companies, governments and individuals make decisions based on available information. Stiglitz (2002, p.469) clarifies that information asymmetry exist when “different people know different things”. Information of private nature creates asymmetries between those who have the information available compared to those who would make better decisions if they had it. Sun, Salama, Hussainey and Habbash (2010) argues that managers could exploit existing information asymmetry and solely report favourable information in accordance with market expectations. They further explain that managers have incentives to disclose information that send positive signals to the market in order to attract current and potential shareholders but also to strengthen the corporate image. The information asymmetry creates options for managers to choose what will be publicly available and if the gap is big enough, managers may have the ability to hide future economic difficulties for a longer period than expected but only until reserves are no longer available. Thus, without information asymmetry, managers would have no incentives to engage in earnings management activities because stakeholders would possess the same information leading to a higher risk of being detected.

The different actors within signaling theory explains why the theory is relevant for the study. Most of the time, the company (managers) is viewed as the sender of information and stakeholders as receivers. This may create incentives for managers to engage in earnings management activities since they still have some control of the information within the financial statements. Nevertheless, auditors, financial analysts and media also

(21)

play an important role as intermediates and becomes important senders of information, which not only is received by stakeholders but also by companies in return. However, by adding more actors who are involved in monitoring the company´s financial information, the fewer the options are from hiding (unpleasant) information due to the decrease in information asymmetry.

2.5.3 Prospect Theory

An alternative theory that could help explain why managers undertake certain decisions such as earnings management is the prospect theory by Kahneman and Tversky (1979). The prospect theory posits that decision-makers acquire value from gains and losses and assess options with respect to a reference point and not wealth or welfare. A key tenet behind the theory is that individual’s value functions are generally convex in losses and concave in gains which gave rise to the S-shaped curve (Kahneman & Tversky, 1979). Thus, value functions around the reference points is steepest but also steeper in a loss direction rather than a gain direction (see figure 1 below). This indicates that decision-makers tend to be loss averse rather than wealth maximizers (Zhang et al., 2008)

(22)

Kahneman and Tversky (1979) explain that the psychological response of perceptual and sensory dimensions often shares the same property which is a concave function of the size of physical and monetary changes. For example, individuals regard the difference between a loss of 100 and 200 as greater than the difference between a loss of 1.100 and 1.200 (Kahneman & Tversky, 1979).

Analyst consensus forecast of earnings could be considered a natural reference point for managers since stakeholders build up their expectations based on estimates from financial analysts. Managers have incentives to decrease losses close to the reference point to influence the value perceived by stakeholders whereas an attempt to decrease losses when the actual outcome is far from the reference point is meaningless. Hence, the cost of mitigating losses becomes higher than the benefits (Burgstahler & Dichev, 1997).

Prospect theory is relevant for the study since it illustrates that decision-makers (managers) are short-term oriented and loss averse. Managers are seeking for instant response and benefits in line with possessed benchmarks within the fiscal year. As described under section 2.4 analyst coverage, managers may manipulate earnings since they are pressurized by stakeholders and especially financial analysts, to meet predetermined targets but also because they have a self-interest of maximizing compensation. Decisions of certain activities related to real earnings management in line with prospect theory, will most likely influence future earnings negatively. However, since managers are appointed and compensated for managing shareholders investments, they have incentives to fulfill investors' expectations by fulfilling the short-term targets at the cost of sacrificing long-term value.

(23)

2.6 Hypotheses development

Earnings management has been thoroughly investigated by existing literature where both accrual-based and real earnings management have been present (e.g. Yu, 2008; Gunny 2010; Zang, 2012). As the social reality changes, some people are harmed while others benefit from the change. Financial analysts are important actors in the capital market by monitoring companies’ annual report and distributing information to a broader audience (Yu, 2008). In addition to accountants, financial analysts could limit managers ability to engage in earnings management since forecasted financial information becomes publicly available which benefits stakeholders. The additional information stakeholders receive from forecasts decreases the information asymmetry which imposes difficulties for managers to hide financial information related to earnings management. From a theoretical perspective, agency theory suggests that managers have incentives to engage in earnings management when information asymmetry exists. By contrast, prospect theory posits that managers will manipulate earnings if they see the possibilities to meet or beat earnings targets when the benefits exceed the cost of manipulating.

Based on existing literature two potential outcomes of analyst coverage impact on managers decision to engage in earnings management are identified. First, increased analyst coverage will decrease managers tendency to engage in earnings management due to the decreased information asymmetry. Second, increased analyst coverage will increase managers tendency to engage in earnings management due to the pressure effect. However, there is a discrepancy between previous literature on how they measure analyst coverage impact on earnings management. For example, Zang (2012) used suspect companies, in other words, companies with obvious incentives to manipulate earnings. She found evidence of managers being pressurized by financial analysts to reach forecasted earnings through the usage of accrual-based and real earnings management. On the other hand, Yu (2008) uses the entire sample and not only suspect companies. He found evidence of managers being frightened by an increase in analyst coverage suggesting that analyst coverage is negatively correlated with earnings management.

(24)

In line with Yu (2008), the study uses the entire sample size but divides the sample into two groups depending on the number of financial analysts that monitors each company. The two groups (A and B) are the dependent variables, proxied by accrual-based and real earnings management whereas analyst coverage is the independent variable. The study conjectures that an increase in analyst coverage will decrease managers engagement in earnings management. Hence, the study expects group B to manipulate earnings less since they are monitored by more analysts. The discussion leads to the following hypotheses.

𝐻_01𝐴: 𝜇(𝐴𝑉_𝐷𝐴)_𝐴− 𝜇(𝐴𝑉_𝐷𝐴)_𝐵 = 0

𝐻1𝐴: 𝜇(𝐴𝑉_𝐷𝐴)𝐴− 𝜇(𝐴𝑉_𝐷𝐴)𝐵 ≠ 0

H01A: There is no difference in mean absolute values of accrual-based earnings

management between group A and B.

𝐻_01𝐵: 𝜇(𝐴𝑉_𝑅𝐸𝑀)_𝐴− 𝜇(𝐴𝑉_𝑅𝐸𝑀)_𝐵 = 0

𝐻1𝐵: 𝜇(𝐴𝑉_𝑅𝐸𝑀)𝐴− 𝜇(𝐴𝑉_𝑅𝐸𝑀)𝐵 ≠ 0

H01B: There is no difference in mean absolute values of real earnings management

between group A and B.

Further, the study investigates the relative trade-off between companies’ usage of discretionary accruals and real earnings management. Lately, researchers have identified a preference of manipulating earnings through real activities rather than discretionary accruals and there are several valid explanations behind this. First, accounting standards and legislations are thoroughly updated, limiting managers options to manipulate earnings through discretionary accruals (Cohen et al., 2008). Second, the probability of being detected is lower for real activities than for discretionary accruals (Roychowdhury, 2006). Therefore, the study conjectures that managers may manipulate earnings through real earnings management rather than accrual-based earnings management. Hence, the study expects both group A and B to have higher levels of real earnings management than accrual-based earnings management. The discussion leads to the following hypotheses.

(25)

𝐻02𝐴: 𝜇(𝐴𝑉_𝐷𝐴)𝐴− 𝜇(𝐴𝑉_𝑅𝐸𝑀)𝐴 = 0

𝐻2𝐴: 𝜇(𝐴𝑉_𝐷𝐴)𝐴− 𝜇(𝐴𝑉_𝑅𝐸𝑀)𝐴 ≠ 0

H02A: There is no difference between mean absolute values of discretionary accruals and

absolute values of real earnings management in group A.

𝐻_02𝐵: 𝜇(𝐴𝑉_𝐷𝐴)_𝐵− 𝜇(𝐴𝑉_𝑅𝐸𝑀)_𝐵 = 0

𝐻_2𝐵: 𝜇(𝐴𝑉_𝐷𝐴)_𝐵− 𝜇(𝐴𝑉_𝑅𝐸𝑀)_𝐵 ≠ 0

H02B: There is no difference between mean absolute values of discretionary accruals and

(26)

3 Methodology

In this section, the study first introduces the research methodology and the process of selecting companies. Further, variables of interest in terms of data selection are presented, followed by measurement, variables and equations for accrual-based earnings management, real earnings management and analyst coverage impact on earnings management. Lastly, an evaluation of the method is discussed.

3.1 Research methodology

3.1.1 Philosophical view of science and scientific reasoning

In order to fulfil the purpose of the study and determine the effect analyst coverage has on earnings management, the study applies a natural scientific epistemological position acknowledged as a positivistic approach. The idea of positivism is that the nature of knowledge exists and that it can be measured by objective methods rather than subjective judgement which means that the social phenomena is independent of people (Bryman & Bell, 2015)

An alternative philosophy to positivism is interpretivism. This position criticizes the application of objective methods and advocates for subjective reasoning. Instead of natural science, interpretivism builds on the ontological position constructionism which means that the social phenomena are accomplished by people (Bryman & Bell, 2015). Thus, the societal reality focuses on what people are experiencing, thinking and feeling and not empirical data.

Depending on whether a researcher sees the reality as natural science or social science, there are two approaches a researcher may take, deductive reasoning or inductive reasoning (Bryman & Bell, 2015). This study applies deductive reasoning which means that the process starts from studying existing literature and theories within the earnings management field to generate hypotheses and predictions. Once hypotheses are formed, the study collects data to test the hypotheses which are either accepted or rejected. The final step of the deductive approach loops back to existing theories which are then compared with the findings. The deductive theory is an objective method to demonstrate causality based on natural science and does not normally include human aspects.

(27)

By contrast, inductive reasoning aims to increase general understanding of social knowledge. This approach is connected to interpretivism and entails a subjective reasoning which is fulfilled by people. The process starts by collecting rich data from qualitative research designs such as interviews. The findings are based on peoples’ interests, experiences, thoughts and feelings related to a societal reality. Once the data is collected, the findings generate a theory based on the researcher’s interpretation and belief

3.1.2 Research Strategy

According to Bryman and Bell (2015) a positivistic view and a deductive scientific reasoning often leads to a quantitative research strategy; this is also the case for this study. A quantitative research method is characterized by collecting numerical data and often leads to complex models (Smith, 2011). By examining previous literature, a clear dominance of a quantitative rather than a qualitative research strategy is identified (e.g., Roychowdhury, 2006; Yu, 2008; Cohen and Zarowin 2010; Gunny, 2010; Kim et al., 2012). The primary strengths of a quantitative research method are the ability to collect data rapidly and economically. Also, statistical analysis of a large observation has the ability to identify discrepancies which may be of interest for policy decisions (Easterby-Smith, Thorpe & Jackson, 2015).

However, a quantitative research strategy is not without flaws. Saunders, Lewis and Thornhill (2009) argues that a quantitative method requires a large amount of data in order to make the results generalizable on a company-level. Easterby-Smith et al. (2015), argues that the method has difficulties in generating theories since the process focuses on what is or what has been. Further, they argue that this imposes challenges for policy-makers to interpret what actions should be implemented.

If the purpose of the study would have been to investigate why managers engage in earnings management and how analyst coverage impacts their decisions, then a qualitative approach would have been appropriate to follow. A qualitative research design

(28)

size where societal reality is investigated in depth. A common structure is to interview people with the aim to understand and interpret how the societal reality is determined by people rather than from an objective point of view (Easterby-Smith et al., 2015).

The study could be expanded by using additional collection methods to triangulate the results and enhance the validity of the study. For example, interviews or surveys could be conducted to complement the results and be able to explain why Swedish companies engage in earnings management. However, the probability of receiving trustworthy and accurate data would most likely have been rather low. Most managers would be unwilling to admit that they engage in earnings management, nor would they be keen to explain why. The reason for this relates to the purpose of earnings management, namely to mislead stakeholders (Healy & Wahlen, 1999).

3.2 Empirical research design and method

The research design is contingent on several factors including data, purpose of the study and research question (Smith, 2011). However, normal levels of accruals and real activities are usually estimated through a time-series or a cross-sectional research design which limits the study’s options in terms of viable research designs (Defond & Jiambalvo, 1994). Both research designs imply collection of data for several observations at a specific time to identify patterns and correlations between a set of variables (Bryman & Bell, 2015).

The choice of research design stems from existing research within the earnings management field where the cross-sectional design is overrepresented (e.g., Roychowdhury, 2006; Yu, 2008; Sun & Liu, 2016). Defond and Jiambalvo (1994) explain that the cross-sectional models estimate industry-specific coefficients for every year and thereby avoid stable coefficients for the period which a time-series analysis would have generated.

The study uses archival sources of secondary data. Numerical data in terms of financial information is collected through Thomson Reuters DataStream and consists of annual reports from companies constituting to the sample size. For a quantitative research it is

(29)

important to collect data for a large sample size so that either causality or correlation could be identified. Using archival sources have several advantages. First, the reliability and quality of the secondary data is considered to be high because annual reports are accepted during annual general meetings and these documents are non-editable once published. Second, the process of collecting data is less time-consuming and less expensive compared to collecting primary data. This enables more time to focus on analyzing the results which is possible when relying of publicly available data of high quality.

3.2.1 Dependent, independent and proxy variables

The study divides company-year observations into two groups (A and B). Group A consist of companies monitored by 1-4 financial analysts while group B consist of companies monitored by 5-30 financial analysts. In this study group A and B are the dependent variables. Analyst coverage is the independent variable since the study conjectures that financial analysts have an impact on managers decisions to engage in earnings management and that the number of financial analysts monitoring a company annually has a negative effect on the magnitude of earnings management. The two elements of earnings management, discretionary accruals and real activities, are used as proxies for the dependent variables (group A and B).

3.2.2 Sample selection

This study selects companies that are listed on the Swedish stock exchange, Nasdaq OMX Stockholm for the time period 2009 to 2016. Nasdaq (2020) has divided the 385 listed companies on the stock exchange into three subclassification; Small-, Mid- and Large Cap. Small Cap consists of companies with a share value up to EUR 150 million. Companies with a share value exceeding EUR 150 million but below EUR 1 billion are classified as Mid Cap. At the top of the hierarchy, Large Cap is categorized as companies with a share value over EUR 1 billion (Nasdaq, 2020).

(30)

The study excludes 314 companies, as shown in table 1, from the final sample size due to the following reasons. First, the study excludes 173 companies since they are not listed on the stock market from January 2009. Second, it excludes 55 financial institutions due to their inherent differences in reporting regulations where accounting rules such as the accrual process are different compared to those of other listed companies (Yu, 2008; Sun & Liu 2016). Third, companies with less than three years of required financial data to calculate real earnings management, accrual-based earnings management and financial analyst impact on earnings management are excluded. The study excludes 29 companies within this category. Including companies with less than three years of required financial data would not provide a complete picture of neither a company’s earnings management activities nor financial analysts impact on earnings management since it would be hard to differentiate normal values from abnormal values.

Table 1 – Sample size

Nasdaq OMX Stockholm Number of companies

Nasdaq OMX Stockholm complete 385 Excluded companies

Finance sector -55

Not listed from 1/1 2009 -173 Secondary listed on Nordic Stockholm -7 Not calendar year as fiscal year -14 Amounts not in SEK -6 Incomplete information -29 Excluded due to business classification -3 Missing analysts forecast -23 Incomplete full analysts forecast -4 Total number of excluded 314

Unique companies 71

Company-year observations 568

Please see appendix B for more information

The study utilizes Thomson Reuters Business Economic Sector classification to divide the companies into different industries, see table 2 for industry classification. Industries with seven or less companies are excluded from the study, similarly to Cohen et al. (2008) and Sun and Liu (2016) who excluded industries with eight or less companies. This reduces a single company’s effect on the specific industry’s estimated normal value. Thus, companies within the energy, telecommunication services and utilities sector are

(31)

excluded. The final sample consist of five industries with 71 unique companies and a total of 568 company-year observations.

Table 2 – Industries

Industry code Economic sector Number of companies

51 Basic Materials 8

52 Industrials 22

53-54 Consumer Goods & Services 21

56 Healthcare 7

57 Technology 13

Total 71

Please see appendix B for more information.

3.2.3 Data collection

The study collects financial statements for all the companies included in the final sample size for the years 2009 to 2016 from Thompson Reuters DataStream. As discussed under sample selection, a longer period is preferable in order to estimate earnings management activities but also to identify patterns and deviations of annual financial information between years. The study selects annual data instead of quarterly since managers try to avoid reporting annual losses. Roychowdhury (2006) argues that companies are more likely to report losses in their interim reports due to seasonality and that stakeholders regard annual losses more seriously.

The study defines analyst coverage as the number of financial analysts that monitors a company. To measure analyst coverage impact on managers decision to engage in earnings management, the study collects annual data for analyst coverage from Institutional Brokers Estimates Systems (I/B/E/S) accessed through Thomson Reuters Eikon.

(32)

3.2.4 Variables

The study collects annual data for the variables presented in table 3. The study interprets discretionary expenses as the total of sales, general and administrative expenses (SG&A) since not all companies spend resources on research and development (R&D), nor is marketing expenses separately disclosed in the income statement. Therefore, where R&D costs are activated, the study aggregates the cost to SG&A expenses. All numbers in table 3 are expressed in thousand SEK.

Table 3 – Descriptive statistics variables

Variable Mean Median Std Dev Percentiles

25% 75%

Cash 2840885 244650 9734177 99356 1334750

Cost of goods sold 17979472 2813052 35311148 568250 17815000

Current asset 12967250 1425000 31729919 412963 11117500 Current liabilities 9036383 927920 22504736 209855 7410250 Depreciation 945297 105861 2311900 26056 799250 Equity 9675278 1318861 21535316 379968 9368975 IBEXI 1204010 188100 2636421 29860 1168250 Inventories 3949927 449450 7830981 28296 3566750 Market value 19882931 3517730 40851209 923160 17542543

Net cash flow 2058709 281157 4191752 59225 2047000

Net sales 24765918 3891700 49130892 815988 26766500 PPE 11797380 1017289 27897970 79530 9025000 Receivables 5738242 642930 14921682 189235 3434250 R&D 1497068 99000 5198083 26304 4117000 Revenue 24847186 3891700 49258552 815988 26766500 SG&A 4349595 856600 9455659 278939 4044500 Total assets 25406454 2789150 56080116 730300 28242000

(33)

3.3 Measurements of earnings management

The study examines the two earnings management methods, accrual-based and real earnings management with the purpose of capturing all activities related to manipulated earnings. Fields et al. (2001) explains the importance of incorporating both activities and describes that the complete effect of earnings management activities cannot be identified by only including one earnings management method.

The study applies lagged total asset with the purpose of comparing dependent variables within earnings management. Lagged variables are displayed by dividing the variable of interest with (At-1) which means that lagged total asset is company specific. Another

widely used variable is the scaled intercept b1(1/At-1). Roychowdhury (2006) explains

that the scaled intercept generally is included when estimating non-discretionary accruals and that it is used to avoid a false correlation between scaled cash flow from operations and scaled sales.

In addition to signed values, the study utilizes absolute values of discretionary accruals and real earnings management to reduce the adverse effect between positive and negative values, since earnings management could include both increasing and income-decreasing activities (Warfield, Wild, & Wild, 1995; Klein, 2002). Following Roychowdhury (2006), the study estimates all earnings management proxies through a cross-sectional regression for every year and industry. Mean values of coefficients for each industry are obtained by SPSS and are utilized to estimate normal levels for each industry. For each sample, the industry’s normal level is subtracted from the actual company level and the difference is defined as the residual (abnormal values). If the residual is greater or smaller than zero, then earnings management is present.

(34)

3.4 Proxies – Accrual-based earnings management

Accrual-based earnings management is easier to detect compared to real earnings management since managers are required to disclose changes of accounting choices in the notes of the financial statements. These notes must include the impact and deviation between pre and post of the revaluation of the line item.

Total accruals can be decomposed into two groups, non-discretionary and discretionary accruals (Darmawan, Sutrisno, & Mardiati, 2019). A major objective of the discretionary models (e.g., the modified Jones model) is to filter out non-discretionary accruals from total accruals (Kothari et al., 2005). Non-discretionary accruals are mandatory accruals related to normal economic conditions of the firm that are activated but not yet realized/paid or vice versa. This means that managers are not able to affect non-discretionary accruals. On the other hand, non-discretionary accruals are accruals that are not dependent on a contract between the company and a contractor, indicating that managers have the ability and authority to change the policy or nature of the activated expense in the financial statements (Darmawan et al., 2019).

To identify if managers have utilized accounting choices to influence reported earnings, the study applies discretionary accruals (DA) as a proxy for accrual-based earnings management. Discretionary accruals are measured by a modified Jones (1991) model and are calculated every year for all industries (see appendix B for industry classification). Thus, the study partly controls for changes in economic conditions in different industries that influence total accruals (Defond & Jiambalvo, 1994; Kasznik, 1999). Compared to the original Jones (1991) model, the study captures changes in accounts receivable which is essential to include since traded accounts receivable is an independent variable of discretionary accruals that managers have the flexibility to manipulate (Kothari et al., 2005; Darmawan et al., 2019). Additionally, the study includes lagged total assets and income before extraordinary items which increases the accuracy of estimating discretionary accruals (Kim et al., 2012).

(35)

The study excludes non-discretionary accruals due to managers inability to influence prearranged accruals in the short run. As a result, non-discretionary accruals are subtracted from total accruals. If there is a residual, managers are manipulating earnings through discretionary accruals. The study begins with presenting equation 1, discretionary accruals. The equation is derived from equation 2 and 4.

𝐷𝐴𝑡 𝐴_𝑡−1= 𝑇𝐴𝑡 𝐴_𝑡−1− 𝑁𝐷𝐴𝑡 𝐴_𝑡−1 Equation 1

Where t denotes fiscal year. DA = Discretionary accruals TA = Total accruals

NDA = Non-discretionary accruals. At-1 = Lagged total assets

When discretionary accruals are positive (negative), it means that the company’s total accruals are larger (smaller) than non-discretionary accruals indicating that the company is manipulating earnings upwards (downwards) since they have a higher level of total accruals than the specific industry’s estimated normal value (Cohen & Zarowin, 2010).

First, the study calculates total accruals for every company and year using company specific figures. In line with Kothari et al. (2005), the following equation is applied to calculate total accruals:

𝑇𝐴_𝑡 𝐴_𝑡−1=

(∆𝐶𝐴_𝑡− ∆𝐶𝐿_𝑡− ∆𝐶𝐴𝑆𝐻_𝑡− 𝐷𝐸𝑃_𝑡) 𝐴_𝑡−1

Equation 2

Where t denotes fiscal year. TA = Total accruals

∆CA = Change in current assets

∆CL = Change in current liabilities excluding the portion of long-term debt. ∆CASH = Change in cash and cash equivalents

The Impact of Financial Analysts on Earnings Management : Empirical evidence from Swedish listed companies