1
“How do fraud schemes develop over time regarding the fraud
amount and its major players?”
Master Thesis AccountancyArjan Edel S2350459 Ohmstraat 11-1 1098 ST, Amsterdam T: 0614051491 E: a.edel.1@student.rug.nl Word count: 11158 Supervisor: Dr. K. Linke
2 Abstract
In this study we examine the development of the players and the fraud amount in financial statement fraud schemes. The basis for our study is the theory of Palmer and Maher (2006). They argue that a fraud scheme consists of several stages. Namely the initiation stage, the proliferation stage, and the institutionalization stage. We found that the number of people increases in the proliferation stage compared to the initiation stage, that the CEO is the initiator of the fraud, and that financial middle management is recruited during the proliferation stage. Furthermore we examined the development of the fraud amount and found evidence for an increasing fraud amount during the stages of Palmer and Maher (2006).
3
Table of Contents
1 Introduction ... 4 2 Theoretical Framework and Hypotheses Development ... 6 2.1 Theory ... 6 2.2 Hypotheses development ... 8 3 Research Method ... 15 3.1 Data Source ... 15 3.2 Data ... 16 3.3 Variables ... 18 3.4 Control variables ... 20 4 Results ... 22 4.1 Descriptive statistics ... 22 4.2 Results statistical tests ... 24 5 Discussion and Conclusion ... 29 6 References ... 31 7 Appendix ... 36
4
1
Introduction
Corporate fraud has a substantial impact on society. Sutherland (1949), a pioneer in white-collar fraud literature, exposes that white-white-collar crime has an even deeper impact on society than street crime because it corrodes trust in authorities and institutions. Nowadays, white-collar crime is still an issue in the corporate landscape. This statement is supported by the research of the Association of Certified Fraud Examiners ACFE in 2016. They state that in particular financial statement fraud has the largest impact compared to other types of corporate fraud. The case of Toshiba1, a fraud scheme from 2008 to 2014, is an example of a fraudulent financial reporting case. Top Executives of Toshiba systematically pushed the targets of the several departments. Managers of the divisions reported incorrectly a profit of 1.2 billion instead of the losses they were subjected to. The fraud had a deep impact on Toshiba, 16 executives left the company or were fired and it had a negative effect of 32% on Toshibas stock price in 20162. This case is just an example; together with many fraudulent financial reporting scandals in the 21st century (Enron, Ahold e.g.) is the motivation for research to investigate the topic of fraudulent financial reporting.
According to Anand, Dacina and Murphy (2015), most scientific research in the field of fraudulent financial reporting is based on the theory of the Fraud Triangle. However, the fraud triangle is not sufficient to understand fraudulent financial reporting. Namely, the fraud triangle focuses on the factors affecting individual fraud perpetrators, views fraud as static instead of a process and is not sufficient in order to understand the collusion aspect of fraud (Dorminey, Fleming & Kranacher, 2010; Morales, Gendron & Guénin-paracini, 2014; Free, 2017). In contrast to the fraud triangle, there are articles that noted that fraud not must be seen as a static act. For example, Ashforth and Anand (2003) developed a theory in which they explain that corruption in organizations goes through several stages and becomes normalized, so which indicates that fraud is a process. Since the literature regarding the fraud process is scarce and there is few empirical evidence for the fraud process, we examine the fraud process in this study. An aspect of the fraud process is the so-called slippery slope effect. In non-accounting literature, the phenomenon of the slippery slope effect is researched. Studies of Gino and Bazerman (2009) and Welsh, Snyder, Christian and Ordonez
1
https://www.japantimes.co.jp/news/2015/09/18/business/corporate-business/pressure-to-show-a-profit-led-to-toshibas-accounting-scandal/#.W78TR2gzaUk
2 https://etfdailynews.com/2016/12/28/toshibas-accounting-scandal-tanks-stock-by-32/
5 (2015) concluded that small ethical transgressions can lead to accepting more unethical behavior from others and that small ethical transgressions lead to bigger ethical transgressions over time. In the field of fraudulent financial reporting, empirical evidence of the fraud process and the slippery slope is scarce. The study of Schrand and Zechman (2012) found that financial statement fraud often starts with an optimistically unintentional misstatement, which could lead to intentional misreporting in order to fix the overstatement, through the slippery slope effect. Thus, the slippery slope is related to a growing fraud amount (Reckers & Samuelson, 2016). The report of the ACFE (2016) examined the fraud amount in fraud schemes in general. They found that the longer the fraud exists, the higher the loss of the fraud. The results of Schrand and Zechman (2012) and ACFE (2016) indicate that the fraud amount will increase during the fraud. However, the development of the fraud amount in fraudulent financial reporting is not empirically researched in prior literature yet, therefore we examine the fraud amount during the fraud process.
The second subject of our study is the involvement of organizational players in a fraudulent financial reporting scheme. Beasley, Hermanson and Carcello (2010), Trompeter, Carpenter, Desai, Jones and Riley Jr. (2013), Free (2017) and Albrecht, Holland and Malagueno (2015) examined the collusion aspect of fraud. These articles concluded that financial statement fraud is mostly perpetrated in teams and that multiple organizational players collude in order to override internal controls that are supposed to prevent financial statement fraud. These articles did not examine the development of such a “team” during the fraud process. The development of the number of players in a fraud scheme is interesting. Namely, the ACFE (2018) concluded that there is a positive relation between the number of fraud perpetrators and the loss that is associated with the fraud scheme. Because of the relation between the number of people involved and the loss of the fraud, together with the lack of research in the number of people involved, we examine the development of the number of people involved during financial statement fraud schemes. Additionally, Beasley et al. (2010) examined what players are likely to be involved. But, this research did not mention the development of the involvement of players during the fraud scheme. Therefore, we extend this article and examine whether the CEO is the initiator of the fraud and to what extent financial middle managers are recruited, in order to gain evidence for the players involved in collusion.
As stated in the subsequent paragraphs, this study is about the process of fraudulent financial reporting in organizations. We examine the development of a fraudulent financial reporting
6 on several aspects. We do this in an empirical way by analyzing litigation releases of the SEC. In this way, we broaden the scientific knowledge about the relatively new and scarce literature of the fraud process.
Our research question is the following:
“How do fraud schemes develop over time regarding the fraud amount and its major players?”
In the next sections of this research, we describe our theoretical framework followed by the hypotheses. Then we describe the methodology, the results of our statistical tests and in the end, we present our conclusions, limitations, and recommendations for further research.
2 Theoretical Framework and Hypotheses Development
2.1 Theory
Fraudulent financial reporting is defined by the research of Beasley et al. (2010 p.11) as “the intentional material misstatement of financial statements or financial disclosures or the perpetration of an illegal act that has a material direct effect on the financial statements or financial disclosures”. Literature close to our subject investigates determinants affecting the occurrence and risk of fraudulent financial reporting. Hogan, Rezaee, Riley and Velury (2008) summarize this literature in their review commissioned by the audit section of the American Accounting Association (AAA). The relevant subjects in the fraudulent financial reporting literature are according to Hogan et al. (2008) incentives/pressures for committing fraud, opportunities, the role of the auditor in reducing the opportunity, the detection of fraud by auditors and the impact of fraud auditing standards. Trompeter, Carpenter, Desai, Jones and Riley Jr. (2013), reviewed and summarized literature in the field of fraudulent financial reporting. They extend the work of Hogan et al. (2008) by using recent literature and include literature outside the field of accounting. The basis for both Hogan et al. (2008) and Trompeter et al. (2013) is that the literature is divided into the three factors of the fraud triangle, composed by (Cressey, 1950) namely opportunity, rationalization, and incentives.
7 There is evidence in the recent literature for the applicability of the fraud triangle in fraudulent financial reporting cases (Yung, 2009; Bell & Carcello, 2000; Raezee, 2005). However, according to (Free & Murphy, 2015), the applicability of the fraud triangle in our research area is questionable. This is because the fraud triangle focuses on the factors affecting individual fraud perpetrators and view fraud as static instead of a process (Morales, Gendron & Guénin-paracini, 2014; Free, 2017). Since we are interested in the development of a fraud scheme over time, we use a different theory as a basis for our research.
There are some theories that describe how corporate corruption develops over time in an organization, namely Brief, Burkram and Duterich (2001) and Ashforth and Anand (2003). They both approach fraud as a process that develops over time and where multiple organizational players will be involved. According to Palmer and Maher (2006), these studies can be implicitly combined in one model that describes the way corporate fraud develops during stages. The theory of Palmer and Maher (2006) consists of the following stages: initiation stage, proliferation stage, institutionalization stage and socialization stage. This theory is applicable to our research because it clearly separates the fraud scheme in different stages. We use this theory because it guides us to the development of the process of a fraudulent financial reporting scheme. On the basis of the stages of Palmer and Maher (2006) we build our hypotheses. The stages are briefly discussed in the next paragraph.
The initiation stage is the moment in the fraud scheme where aggressive reporting turns into fraudulent financial reporting. An employee of the top management (CEO, CFO and COO (Beasley et al., 2010)) of the organization mostly perpetrates this first step in a fraud scheme. The proliferation stage is the phase in the fraud process where the higher manager starts recruiting lower employees to co-offend in the fraud. In this stage, according to this theory, the number of people involved in the fraud will increase. The next phase is the institutionalization phase, where the fraud is embedded in the processes and structures of the organization. The last phase in a fraud scheme is the socialization stage, where newcomers in the organization automatically and mindlessly will be involved in the fraud scheme. As if these stages are met, the fraud is normalized in the organization.
Brief et al. (2001), Ashforth and Anand (2003) and therefore Palmer and Maher (2006), all indicate that fraud is subject to the slippery slope effect. According to Welsh et al. (2015), the slippery slope is defined as the process in which small ethical violations can lead to
8 bigger violations over time by means of the slippery slope of unethical behavior. The relation between fraudulent financial reporting and the slippery slope effect is assumed in the articles of Schrand and Zechman (2012), Free and Murphy (2015) Suh, Sweeney, Linke and Wall, (2018). When the slippery slope effect is applied to fraudulent financial reporting, an increasing fraud amount or an “increasing degrees of ethical grayness” (Reckers & Samuelson, 2016) could occur. Additionally, according to the study of Gino and Bazerman (2009), the slippery slope effect also occurs in the fact that people accept unethical behavior that gradually develops rather than unethical behavior occurs abruptly. In other words, through the slippery slope effect, people accept increasing unethical behavior from others. Concluding, on the basis of Ashfort and Anand (2003) and Brief et al. (2001) and Palmer and Maher (2006), we view fraud as an act that doesn’t remain stable, but as a process that evolves and develops over time through the previously mentioned stages. Furthermore, the slippery slope effect is applicable to the perpetration of fraudulent financial reporting (Palmer & Maher, 2016). In the next section, we build our hypotheses on the basis of the theory of Palmer & Maher (2006). We examine the development of the number of people involved in each stage, the involvement of the CEO as an initiator, the recruitment of financial middle management in the proliferation stage and the development of the fraud amount during the stages.
2.2 Hypotheses development
Literature in the field of corporate corruption and financial statement fraud describes that most financial statement frauds are perpetrated by more than one person or is perpetrated in groups (Albrecht et al., 2015). Therefore, Ramamoorti (2008) view fraud as a “team sport”. Through colluding, these teams or groups could override internal controls as segregation of duties (Ramamoorti, 2008). In cases where more people engaged in the fraud scheme, the fraud amount was significantly higher than when fraud is perpetrated by an individual (ACFE, 2018). Furthermore, the conclusion of a fraud report of KPMG in 2018 is that the relative portion of collusion compared to individual perpetrators increases3. In the category of fraudulent financial reporting, Beasley, Hermanson and Carcello (1999) found that in 83 percent of the fraud cases, the CFO and CEO collude. Concluding, on the basis of these articles, we expect that financial statement fraud is perpetrated in groups.
9 In the first hypotheses, we zoom in on the first two stages of a fraud scheme, namely the initiation and the proliferation stage (Palmer & Maher, 2006). The initiation stage is the period where the first corrupt decision or act is perpetrated. Additionally, the initiation stage is the period in the fraud scheme where aggressive financial reporting becomes fraudulent financial reporting (Palmer & Maher, 2006). Subsequent to the initiation stage, the proliferation stage is the stage where new employees are recruited to co-offend in the fraud scheme and is typically the stage where the number of people involved in the fraud scheme increases. The reason for expanding the fraud scheme in terms of the number of players is, that it is necessary in order to override controls (Ramamoorti, 2008) and without collusion it is more likely that the fraud will be revealed (Tillman & Indergraad, 2007). Therefore, the initiator of the fraud will recruit new organizational players. The recruitment of new employees in a fraud scheme is often a top-down process, organizational players in the top of the organization recruit lower-level employees in order to co-offend in fraudulent financial reporting (Palmer & Maher, 2016). The leaders of an organization are important role models for other organizational players (Ashforth & Anand, 2003). Therefore, the role of power could play an important role in recruiting individuals for engaging in the fraud process (Albrecht et al., 2015). Power is used for exerting pressure on individuals to co-offend in fraudulent financial reporting (Albrecht et al., 2015).
We propose that in line with Ramamoorti (2008), that fraudulent financial reporting is a team sport. According to Palmer and Maher (2006), the recruitment of new employees during the fraud starts in the proliferation stage, after the first corrupt act is perpetrated in the initiation stage. Therefore, we expect that the number of people involved in the proliferation stage increases compared to the number of people involved during the initiation stage.
H1: In the proliferation stage, the number of people involved in the fraud will increase compared to the initiation stage
Subsequent to the initiation stage and the proliferation stage, the institutionalization stage starts (Palmer & Maher, 2006). In the first hypothesis we argue that in the proliferation stage the employees are recruited. Subsequent to the recruitment of new perpetrators, the fraud will be institutionalized (Ashforth & Anand, 2003). The institutionalization consists in our case of embedding the fraud in structures and processes of the organization, and mindlessness executing the fraud. So, the fraud perpetrated by the recruited employees will become routine
10 and habit (Ashforth & Anand, 2003). We expect that during the institutionalization stage it is likely that there are no additional players recruited during the fraud. Taking in mind that the fraud is routinized, new employees could disturb this process. Furthermore, new players could possibly be whistleblowers that internally or externally report the fraud scheme to the organization or to the SEC (Lee & Xiao, 2018)
Therefore, our second hypothesis is the following:
H2: In the institutionalization stage, the number of people involved in the fraud remains stable
According to the theory of Palmer and Maher (2006) and the article of Albrecht et al. (2015), a fraud scheme will mostly start at the top of the organization and will expand during the fraud scheme to lower functions. Top functions that are likely to engage are named in the research of Beasley et al. (2010). These functions are the CEO, CFO and the COO. So, the likeliness of the involvement of top functions fraudulent financial reporting is clear. However, the research of Beasley et al. (2010) made no distinction in which function is related to the initiation of the fraud scheme. On the basis of the following paragraphs, we expect that the CEO is the initiator of fraudulent financial reporting in organizations. We use the pressure or incentive factor and the opportunity factor of the fraud triangle of Cressey (1950) as the basis for our argumentation.
There are several reasons why a CEO could perceive pressure that could lead to engaging in fraudulent financial reporting practices. These reasons include pressure that arises from analysts and pressure from shareholders or the board of directors. According to (Albrecht et al., 2015; Hogan et al., 2008; Rezaee, 2005) pressure that arises from meeting or beating analysts’ forecasts could influence the decision of top management, and therefore the CEO, to engage in fraudulent financial reporting practices. And indeed, Beasley et al. (2010) indicate this in their research as one of the reasons. Not only pressure from analysts could make the CEO decide to engage in fraudulent financial reporting. Also, pressure from shareholders and/or the board of directors could be a reason to decide to engage in fraudulent financial reporting. Since the function of the board of directors is monitoring the CEO, they could exert pressure on the CEO in order to meet the desired earnings targets (Schnatterly, Gangloff & Tuske, 2018).
11 So, pressure from analysts and shareholders are factors that contribute to the decision to engage in fraudulent financial reporting. But, the CEO could also have incentives that contribute to the decision to engage in fraudulent financial reporting. Namely, according to Hogan et al. (2008), compensation plans can lead to misbehavior of the CEO because of the incentives it arises. For example, several studies find that CEOs who are compensated with stock options have incentives to increase stock price in order to increase their compensation. Additionally, Efendi, Srivastava and Swanson (2007) found that firms with a CEO with "in-the-money" stock options are more likely to misstate the financial statements. Burns and Kedia (2006) confirm this finding, they found that CEO compensation affects aggressive accounting principles that lead to restatements, in particular when the CEO has stock options that are linked to sensitive stock prices. Concluding, compensation plans could lead to incentives for the CEO to engage in fraudulent financial reporting.
Without the opportunity for committing fraud, the most motivated CEO is not able to commit fraudulent acts (O’Connor et al., 2006). The opportunity for a CEO for perpetrating fraud is dependent on the CEOs power. The more powerful the CEO is, the more the opportunity arises for committing fraud by overruling the board or override organizational controls (Schnatterly et al., 2018). Another factor contributing to the opportunity for the CEO is the information asymmetry that arises from the moral hazard problem. Moral hazard occurs when managers, in this case, the CEO act in a way against the interest of the principal (the shareholder) because of the fact that the behavior of the CEO is hard to observe (O’Connor, Priem & Coombs, 2006). Thus, when the CEO perceives pressure or has incentives for engaging in fraudulent financial reporting, the opportunity for arises because of the moral hazard problem.
To conclude, pressure from both analysts and the board of directors; incentives from compensation plans and opportunity due to the CEO his power and the moral hazard problem, we expect that the CEO is the initiator of the fraud. Therefore, our third hypothesis is the following:
H3: The initiator of the fraud is the CEO of the organization
According to ACFE (2018), financial middle managers are the second group of fraud perpetrators. Furthermore, according to Beasley et al. (2010), controllers are the third named perpetrator of fraudulent financial reporting. Since fraudulent financial reporting is typically a
12 top-down type of fraud, the fraud scheme will be extended to organizational players lower in the hierarchical structure of the organization during the proliferation stage (Palmer & Maher, 2006). The role of the controller and therefore financial middle managers in organizations is to maintain the accuracy of the financial statements (Linke, 2012). The hierarchical position of financial middle managers is under top management, especially under the CFO (Linke, 2012). When top management decides to engage in fraudulently financial reporting practices, they could expose pressure to the financial middle managers in order to commit fraudulent acts (Clinard, 1983). Also, Davis, DeZoort, and Kopp (2006) found that pressure from higher management could lead to inappropriate budget decisions of management accountants. Furthermore, these statements are confirmed by Brickey (2003 p.375), who concludes that mid-level managers “are most likely to be “hands-on” when it comes to implementing the fraud.”
The case in the Article of Albrecht et al. (2015), describes this situation applied to a fraudulent financial reporting case. Top management exerted pressure on the CFO in order to do “everything that is necessary” to meet the disclosed numbers. In other words, the perpetrating of fraud is allowed. The CFO started recruiting financial middle management (the VP of accounting, the VP of financial reporting, the director of financial reporting and, the controller) in order to execute the fraud (Albrecht et al., 2015). The CFO decided to do this because all these people were involved in the financial reporting process. Therefore, it was necessary to recruit financial middle management in order to execute the fraud scheme (Albrecht et al., 2015).
To conclude, financial middle management is under the responsibility of top management, fulfills the functions with influence on financial reporting and could be exposed to pressure from higher functions. Therefore, the choice of top management to recruit financial middle management in their fraud scheme is a logical one. Since the fraud expands top-down during the proliferation stage we expect that financial middle management is recruited and involved in the proliferation stage.
H4: Financial middle management is recruited during the proliferation stage
In the fifth and the sixth hypothesis, we investigate the fraud amount in the light of the process theory of Palmer and Maher (2006). Beasley et al. (2010) examined in their research the fraud amount of 300 companies. They made a distinction in their results in the total
13 cumulative fraud amount and the fraud amount per fraud technique. Furthermore, they concluded that the development of a fraud scheme is not isolated to a single fiscal period, but typically lasts two fiscal years. However, Beasley et al. (2010) did not examine the development of the fraud amount during the fraud period. Gaining insight into the development of the fraud amount is evident. Namely, consequences of financial statement fraud are severe because the financial statements are no longer reliable, and therefore decisions of investors are based on incorrect data (ACFE, 2018). Furthermore, declines in stock value, delisting from the stock exchange or filing for bankruptcy, are possible consequences when financial statement fraud is revealed (Rezaee, 2005). Since the consequences of financial statement fraud are severe, it is evident that we gain insight in the development of the fraud amount. This might be useful to emphasize the urgency for prevention or detection mechanisms in organizations.
In the initiation stage, the initiator of the fraud perpetrates its first fraudulent act by crossing the line from aggressive reporting to fraudulent financial reporting (Palmer & Maher, 2006). A mechanism that is related to such an escalation is the slippery slope effect (Schrand & Zechman, 2012; Free & Murphy, 2015; Suh et al., 2018). The slippery slope is in the literature defined as the fact that a small unethical act could lead to bigger unethical acts over time (Reckers and Samuelson, 2016; Brown, Rennekamp, Zeybert & Zhu, 2014). If the fraud scheme lasts longer, the fraud scheme will expand by recruiting new perpetrators in the proliferation stage (Palmer & Maher, 2006). Recruiting of new perpetrator leads to a bigger team, as stated in hypothesis two. This provides the opportunity for collusion in order to override anti-fraud controls (Ramamoorti, 2008). Since anti-fraud controls are mostly based on the principles of segregation of duties and independent checks, a team of people could have the opportunity to circumvent these controls (ACFE, 2018). Therefore, we expect that the fraud amount increases during the proliferation stage compared to the initiation stage. H5: During the proliferation stage, the fraud amount increases compared to the initiation stage
To summarize, during the initiation stage, the first fraudulent act is perpetrated. During the proliferation stage, the team for further perpetrating the fraud is formed (Palmer & Maher, 2006). In the institutionalization stage, the fraud is embedded in the organizational structures and processes and finally becomes routinized (Palmer & Maher, 2006).
14 During the institutionalization stage, an organization might become dependent on fraudulent acts (Ashforth & Anand, 2003). If budgetary goals are met by perpetrating fraudulent acts, the organization could be obliged to perpetrate another fraudulent act in order to meet the subsequent budgetary goals (Ashforth & Anand, 2003). Moreover, the organizations’ dependency could also be caused by the obligation to fix the perpetrated fraud. In the article of Schrand and Zechman (2016) an example of such a dependency can be found. Namely, they found through analyzing Accounting and Enforcement releases of the SEC that initial misstatements of fraudulent financial reporting are relatively small; in the subsequent periods of the fraud scheme, the expected earnings in order to fix the fraud are relatively poor, which leads to an escalation of the misstatements. So, in order to cover the reversal for the subsequent periods, it is necessary to perpetrate new fraudulent acts. The obligation to perpetrate fraudulent acts might lead to an institutionalized and repeatedly executed task (Ashforth & Anand, 2003). Therefore, the fraudulent practices might “become embedded in the ongoing routines of the organization, a deviant subculture tends to emerge to normalize corruption.” (Ashforth & Anand, 2003 p.9). Finally, the fraudulent act might become routinized and habitual. Since multiple actors are involved in the financial reporting process (Albrecht et al., 2015), fraudulent tasks could be divided over multiple subunits (Argote & Ingram, 2000). The tasks might become separated in specialized tasks, in which individuals might not know how their acts contribute to fraudulent acts (Ashforth & Anand, 2003). Therefore, the act is transformed into a routine and a mechanical and highly programmed operation. (Kelman, 1973 p.46).
To conclude, an organization might become dependent on the perpetration of fraudulent acts. This could lead to repeating, specialized tasks, divided over the subunits of the organization. The fraud will be embedded in the structures and processes and finally becomes routinized. We expect that in the institutionalization stage, the fraud amount increases compared to the proliferation stage.
H6: During the institutionalization stage, the fraud amount increases compared to the proliferation stage
15
3 Research Method
This section gives insight into our research methodology including, the used data source, samples, and variables. We used a quantitative research method in order to examine the fraud process. This provides us with the opportunity to examine the fraud process on a broader scale than examining the fraud process in a qualitative manner due to time constraints and the difficulty of contacting fraud perpetrators. Moreover, our hypotheses are more suited for quantitative research than for qualitative research since we are examining the number of people involved, actors involved and the fraud amount.
3.1 Data Source
In order to gather data for our research, we selected the USA as the jurisdiction for our research. The USA has compared to other jurisdictions more cases of fraudulent financial reporting in the last decades. Furthermore, the governing body responsible for enforcing the federal securities law, the Securities and Exchange Commission, discloses data regarding to fraud cases on their website. This makes the data publicly available and therefore easily accessible for our research. Therefore, we chose the USA as jurisdiction and we use as Beasley et al. (1999; 2010) the database of the SEC in order to collect data.
16 The SEC is the regulator in the United States in the field. Their mission is to “protect investors, maintain fair, orderly, and efficient markets, and facilitate capital information.”4 In this role, the SEC is charged with the enforcement and prosecution of financial reporting fraud. The enforcement division of the SEC investigates possible fraudulent financial reporting cases and prosecutes the suspects in these cases at federal courts. These cases are reported and disclosed in the publicly available litigation releases in the database on the website of the SEC (https://www.sec.gov/litigation/litreleases.shtml). We analyze and use these litigation releases (i.e. complaints) for collecting our data. An advantage of these complaints is that they consist of a comprehensive, complete and detailed description of the prosecuted actors, the fraudulent acts, the used fraud techniques, the fraud amount and the development of the fraud. Besides the advantages of the litigation releases of the SEC, there some disadvantages to notice. According to Beasley et al. (2010), the cases of the SEC are not randomly selected and could be selected on the enforcement strategies and the probability of successful finding fraudulent financial reporting. Furthermore, people and organizations in the complaints are not convicted for financial reporting fraud. They were only prosecuted; therefore, financial fraudulent reporting in some complaints is not proven. Taken the advantages and disadvantages in consideration, the SEC database is the most suitable for our research, also for the reason that there are no better sources available (Beasley et al., 2010).
3.2 Data
In order to collect data for testing our hypothesis, we selected cases of financial statement fraud from the period starting 2000 and ending 2015. We chose for 2000 as a starting point because most complaints before 2000 are not available in the database of the SEC. This is because of the fact that complaints were obtained as hard copies (Linke, 2012). We use 2015 as the end of the period because there is a lag in timing from the moment a fraud is perpetrated, the discovery of the fraud, the investigation of the fraud by the SEC and the issuance of the litigation release by the SEC. For example, the case of Osiris Therapeutics, INC started in 2014, ended in 2015 and is issued in the SEC database in 2017. Therefore, we cannot use more recent cases. We started with 233 selected cases. In order to select only financial reporting fraud cases, the cases that did not comply with the following definition were excluded from our sample. "Financial statement fraud is the intentional material misstatement of financial statements or financial disclosures or the perpetration of an illegal
17 act that has a material direct effect on the financial statements or financial disclosures” (Beasley et al., 2010 p.7). Additionally, the cases in which only the organization was prosecuted are also excluded from our sample. Therefore, we excluded 47 cases, resulting in a sample size of 186 complaints from the period 2000 to 2015.
In addition, not all the complaints in our sample met the requirements of the variables in our hypotheses. So we had to exclude several cases per hypotheses. For hypotheses one and two we excluded the cases in which the fraud scheme lasted no longer than one quarter of a year. In 15 complaints this was the case. Furthermore, for hypotheses two we excluded the cases in which there was no institutionalization stage. Therefore, we have for respectively hypothesis one and hypothesis two 171 and 143 applicable cases. For hypothesis four we excluded the cases in which there was one quarter. For hypothesis four we have 171 applicable cases. For hypothesis five we excluded 15 cases in which the fraud lasted no longer than one quarter. We excluded 32 in which no clear fraud amount was named in the complaint. Furthermore we exclude 69 cases in which the fraud amount was not clearly divided between the initiation stage and the proliferation stage. This left us with 70 cases in total to test hypothesis five. For hypothesis six we excluded 15 cases in which there was one quarter, 28 cases with missing values, 68 cases in which there was no split up of fraud amount and 29 cases in which there was no institutionalization stage. Thus, we selected 46 cases for hypothesis six. For a comprehensive view of the excluded cases, see table 3.1.
Table 3.1 H1 H2 H3 H4 H5 H6 Cases 186 186 186 186 186 186 One quarter 15 15 15 15 15 Missing value 32 28 No split up of fraud amount 69 68 No institutionalization stage 28 29 Cases 171 143 186 171 70 46
18 The selected complaints are analyzed by a group of five researchers. We divided the complaints between these people. Then each group member checked the applicability of the selected complaints on the type of fraud (financial statement fraud or not) and if the complaint was attached. The cases that were doubtful regarding the compliance with the definition of financial statement fraud, where checked by the other group members and if necessary excluded. In order to be uniform, we collaboratively determined the requirements for each variable, stated in section 3.3, and manually collected them.
3.3 Variables
In this section, we first explain how we determined the stages described in the theory section. Next, our variables are described per hypotheses, together with a description of our control variables.
Stages
As stated in our theory section, we determined four stages in the fraud process, namely the initiation stage, the proliferation stage, the institutionalization stage, and the socialization stage. For the purpose of our research we make no difference between the institutionalization stage and the socialization stage. The reason for this is that there is no clear distinction regarding timespan between the stages. According to the research of Beasley et al. (2010), a typical fraudulent financial reporting scheme has an average duration of 31.4 months and a median duration of 24 months. Because of the fact that the initiation stage is just one act that has to be perpetrated, we assume that the initiation stage is during first quarter of the fraud scheme. When the initiation stage ends, the proliferation stage will start in the second quarter of the fraud scheme. We assume that this stage will end at quarter three because in this stage top management requests employees for perpetrating fraud. There must a small period between the request and the involvement of the new players in order to continue the fraud scheme. After the first act, new employees are involved and according to (Palmer & Maher, 2006) institutionalization starts. Subsequently, the institutionalization stage starts and will end until the fraud is stopped or discovered.
19 Hypothesis one and two
In our first hypothesis, we test whether the number of people in the proliferation stage compared to the number of people in the initiation stage. The variables we use are the number of people in the initiation stage (NMBR_INI) and the number of people in the proliferation stage (NMBR_PRO). We counted the number of people according to the stages in each complaint. In our second hypothesis, we test whether the number of people involved in the fraud remains stable. The variables we use are NMBR_PRO and number of people in the institutionalization stage (NMBR_INST). We counted the number of people named according to the applicable stage in each complaint.
Hypothesis three
In our third hypothesis, we test whether the CEO is the initiator of the fraud scheme (CEO_INI). We determined the CEO as the initiator of the fraud if he is the only perpetrator of the fraud in the initiation stage. In cases there was more than one perpetrator in the initiation stage, we determined the CEO as the initiator if this is explicitly mentioned. CEO_INI is a dichotomous variable where 1 = CEO is the initiator, and 0 = CEO is not the initiator.
Hypotheses four
In our fourth hypothesis, we test whether financial middle management is recruited during the proliferation stage. According (Linke, 2012), job titles that fulfill the criteria of financial middle managers are Controller, Chief Accounting Officer, Vice president of Accounting and Vice President of financial reporting. Therefore, these functions are considered as financial middle managers. First, we determined whether the financial middle manager is present in the initiation stage (FMM_INI), then we determined whether the financial middle manager is present in the proliferation stage (FMM_PRO). FMM_INI and FMM_PRO are measured as a dummy variable where 1 is involved in the stage, and 0 is not involved in the stage.
Hypotheses five and six
In hypotheses five and six we investigate the development of the fraud amount during the fraud scheme. In hypotheses five we test whether the fraud amount increases during the initiation stage and the proliferation stage. The variables we use are the fraud amount in the initiation stage (LOGAMNT_INI) and the fraud amount in the proliferation stage
20 (LOGAMNT_PRO). For hypothesis six we use LOGAMNT_PRO and the fraud amount in the institutionalization stage (LOGAMNT_INST). We collected the fraud amount or the restatement as named in the complaints. The fraud amount is expressed in dollars.
3.4 Control variables
Our first control variable is Big 4 auditor (CTRBIG4). Auditors have an important role in reducing the opportunity to commit fraudulent financial reporting (Hogan et al., 2008). There is evidence in research that hiring a BIG 4 auditor increases the quality of earnings. When an organization hires a Big 4 auditor, the quality of the audit increases compared to hiring a non-big 4 auditor due to the non-bigger size of the Big 4 audit firms (DeAngelo, 1981). Francis and Wang (2008) extended this research by testing for endogeneity. They also concluded that hiring a Big 4 auditor enhances the audit quality. As we take a look at auditor and fraud firms, according to Farber (2005), fraud firms less often hiring big 4 firm auditors than non-fraud firms. For these reasons, we expect that having a Big 4 auditor have its impact on non-fraud. Therefore, we control for this variable. CTRBIG4 is founded in the litigation release of the SEC or in the 10k reports found in the EDGAR database of the SEC. We measure Big 4 auditor as a dummy variable. Where 1 is the presence of a Big 4 auditor, and 0 is the non-presence of a Big 4 auditor.
Our second control variable is Firm size (CTRFIRMSZ). The article of Beasley et al. (1999) states that smaller firms are at higher risk for fraud. Furthermore, small films have worse internal control systems that prevent the organization from fraud (Beasley et. al., 1999). Furthermore, according to Beasley et al. (1999), the size of the firm may influence the size of the fraud. We measure the total assets of the firm in dollars (Beasley et al., 1999; Beasley et al., 2010;). Total assets are derived from the Compustat database.
Our third control variable is the financial crisis (CTRFINCR). 2008 was the years of the global financial crisis (Ivashina & Scharfstein, 2010). We control for this variable because of its impact on financial decisions on organizations worldwide, regardless of industry and firm size. Constrained firms during the financial crisis made other choices regarding investments, cash, hiring new employees and debt (Campello, Graham, & Harvey, 2010). Furthermore, financially distressed firms are more likely to show fraudulently behavior regarding to overstating earnings (Rosner, 2003). For these reasons, we are controlling for the financial
21 crisis. The financial crisis is measured as a dummy variable, in which the fraud took place in the financial crisis is 1, and the fraud took not place in the financial crisis is 0.
Table 4.1
Summary of Variables
Variable Abbreviation Type Test
Number of people in initiation stage
NMBR_INI Numerical Wilcoxon Signed Rank Test
Number of people in proliferation stage
NMBR_PRO Numerical Wilcoxon Signed Rank Test
Number of people in institutionalization stage
NMBR_INST Numerical Wilcoxon Signed Rank Test
CEO is the initiator CEO_INI Dummy Chi-square test Financial middle
management in initiation stage
FMM_INI Dummy Wilcoxon Signed Rank Test
Financial middle management in proliferation stage
FMM_PRO Dummy Wilcoxon Signed Rank Test
Fraud amount in initiation stage
LOGAMNT_INI Numerical Dependent t-test Fraud amount in
proliferation stage
LOGAMNT_PRO Numerical Dependent t-test Fraud amount in
institutionalization stage
LOGAMNT_INST Numerical Dependent t-test
Firm size CTRFIRMS Numerical
Big 4 auditor CTRBIG4 Dummy
Financial Crisis CTRFINCR Dummy
22
4 Results
In this section we present the results of the statistical tests we performed in order to test our hypotheses. In the first paragraph the descriptive statistics are presented, followed by the results of the test and lastly the control variables are presented.
4.1 Descriptive statistics
Table 4.1
Descriptive Statistics Numerical Data
N Minimum Maximum Mean Std. Deviation H1 and H2 NMBR_INI 171 1 6 1.98 1.027 NMBR_PRO 171 1 10 2.56 1.393 NMBR_INST 143 1 10 3.04 1.883 H5 and H6 LOGAMNT_INI 70 4.608 8.919 6.422 0.884 LOGAMNT_PRO 71 5.281 9.302 6.896 0.797 LOGAMNT_INST 46 4.913 9.648 7.041 0.840 Control variables CTRFIRMS 186 CTRLBIG4 186 CTRFINCR 186
Table 4.1 shows the summary of the descriptive statistics of our numerical variables, namely the variables of hypothesis one, two, five and six. As stated in the methodology section we excluded cases for both hypotheses one and two and hypotheses five and six (See table 3.1). Therefore, the number of observations for NMBR_INI and NMBR_PRO is 171, and for NMBR_INST 143. The variable NMBR_INI ranges from 1 people involved to 6 people involved, has a mean of 1.96 and a standard deviation of 1.040. The variable NMBR_PRO has a mean of 2.56 and ranges from 1 people involved to 10 people involved and has a standard deviation of 1.393. The variable NMBR_INST has a mean of 3.04, ranges from 1 people involved to 10 people involved and has a standard deviation of 1.869.
We excluded also cases for hypothesis five and hypothesis six (Table 3.3). Therefore the N of LOGAMNT_INI, LOGAMNT_PRO and LOGAMNT_INST are respectively 70, 71 and, 46.
23 The Variable LOGAMNT_INI has a mean of .422, ranges from 4.608 to 8.919 and has a standard deviation of 0.884. The Variable LOGAMNT_PRO has a mean of 6.896, ranges from 5.281 to 9.302 and has a standard deviation of 0.797. The Variable LOGAMNT_INST has a mean of 7.041, ranges from 4.913 to 9.468 and has a standard deviation of 0.840.
Table 4.2
Descriptive statistics dummy variables
CEO_INI 0 89 1 97 Total 186 FMM_PRO 0 1 Total FMM_INI 0 124 13 137 1 1 33 34 Total 125 46 171
Table 4.2 shows the descriptive statistics of hypothesis three and four. For hypothesis three, the total number of observations is 186 (n = 186). The CEO is 89 times not the initiator of the fraud and 97 the initiator of the fraud. For hypothesis four, the total number of observations is 171. Financial middle management is 124 times not observed in the initiation stage, and 13 times not in the initiation stage while observed in the proliferation stage. Financial middle management is three times observed in the initiation stage while 33 times in the proliferation stage.
24
4.2 Results statistical tests
Hypotheses one and two
Hypothesis one and two test the number of people during the initiation stage, proliferation stage and the institutionalization stage. The dependent t-test can be used to test the difference in the means of the variables NMBR_INI, NMBR_PRO and NMBR_INST. However, the assumption of normality is not met (Table 4.3). Namely, for all the variables of hypotheses one and two, the Kolmogorov-Smirnov test is significant. Therefore, we used the Wilcoxon signed rank test. The assumptions for dealing with a continuous variable, independent observations and a non-normal distribution are met.
Table 4.3
Tests of Normality H1 and H2
Kolmogorov-Smirnov Shapiro-Wilk
Statistic df Sig. Statistic df Sig.
NMBR_INI .235 184 .000 .810 184 .000
NMBR_PRO .191 184 .000 .899 184 .000
NMBR_INST .153 184 .000 .881 184 .000
The results of the Wilcoxon Signed Rank test for hypotheses one and two are presented in table 4.4. The results for hypothesis one indicate that the number of people was higher in the proliferation stage (M = 0) compared to the initiation stage (M = 63.), Z = -7.196, p < .001. Therefore, we can accept hypothesis one, the number of people increases in the proliferation stage compared to the initiation stage. The results for hypothesis two indicate that the number of people was higher in the institutionalization stage (M = 0) compared to the proliferation stage (M = 17.50), Z = -5.204, p < 0.001. Thus, we reject our expectations in hypothesis two, the number of people does not remain stable in the institutionalization stage compared to the proliferation stage.
25 Table 4.4
Wilcoxon Signed Rank Test
N Z Sig. Negative ranks (Mean) Positive ranks (Mean) NMBR_PRO – NMBR_INI 171 -7,196 .000* 0 32 NMBR_INST – NMBR_PRO 143 -5.204 .000* 0 17.50
Both tests are based on negative ranks *Significant at the p < 0.01 level
Hypothesis Three
For hypotheses three, we performed the Chi-square test. In table 4.5 are the results presented of hypothesis three. 89 times we observed that the CEO was not the initiator, 97 times we observed that the CEO was the initiator. The result of the chi-square test shows that there is no significant between proportions of initiation and not initiation. Therefore, the groups are equal (χ2 (1) = 0.344 p = .557). However, in our test we made no difference whether the CEO was involved in the fraud or not. Therefore, we composed a new variable in which we tested whether the CEO is the initiator when he is involved in the fraud. We composed a dichotomous variable (CEO_INV_INI) where 0 is CEO involved but not the initiator, and 1 is CEO involved and also the initiator. The results of the test are presented in table 4.5. 5 times the CEO was observed as involved but not the initiator of the fraud. The CEO was 97 times involved and also the initiator of the fraud. The result of the chi-square test shows that when the CEO is involved, he is associated with initiating the fraud (χ2 (1) = 82.980 p < .001). Therefore we can conclude that the CEO is the initiator of the fraud, thus we accept hypothesis three, with the note that we only tested in cases where the CEO is involved.
26 Table 4.5 Chi-square test (1) CEO_INI Observations 0 89 1 97 Total 186 df Sig. Chi-Square .344a 1 .557 Chi-square test (2) CEO_INV_INI Observations 0 5 1 97 102 Value df Sig. Chi-square 82.980a 1 .000*
*Significant at the p < 0.01 level
Hypothesis Four
In hypothesis four we tested whether financial middle management is recruited during the proliferation stage. We tested the relation of two variables namely FMM_INI and FMM_PRO. The results presented in table 4.7 indicate that based on negative ranks that FMM_PRO (M = 8.50) increases compared to FMM_INI (M = 8.50), Z = 2,500, p < 0.05. We can conclude that financial middle management is recruited during the proliferation stage. Thus, we can accept hypothesis four.
27 Table 4.6
Wilcoxon Signed Rank Test
FMM_PRO – FMM_INI Z Sig. Negative ranks
(Mean)
Positive ranks (Mean)
2.500 0.012* 8.50 8.50
The tests is based on negative ranks *Significant at p < 0.05
Hypotheses five and six
For hypothesis five and six we examined the development of the fraud amount during the initiation stage, the proliferation stage and the institutionalization stage. First we reduced skew by performing a log transformation. Then, a dependent t-test is performed in order to test the difference in means of the variables. The normality assumption regarding to the Shapiro-Wilk test are met, except for the LOGAMNT_INI variable. However, according to the Kolmogorov-Smirnov test all variables are normally distributed (including LOGAMNT_INI). Since the other assumptions (continuous variable, matched pairs and no significant outliers) are met.
Table 4.7
Tests of Normality H5 and H6
Kolmogorov-Smirnov Shapiro-Wilk
Statistic df Sig. Statistic df Sig.
LOGAMNT_INI .119 45 .113 .945 45 .034*
LOGAMNT_PRO .116 45 .113 .960 45 .119
LOGAMNT_INST .111 45 .200 .968 45 .250
*Significant at the p < 0.05 level
The test is conducted to compare the fraud amount in the initiation stage and the fraud amount in the proliferation stage. As presented in table 4.8, the results of hypothesis five
28 indicate that there is a significant difference in the means of LOGAMNT_INI (M = 6.422, SE = .106) and LOGAMNT_PRO (M =6.900 SE = .096), t (69) = -8.71, p < 0.001. Therefore our results suggest that the fraud amount in the proliferation stage increases compared to the initiation stage. Thus, we can accept hypothesis five. The results of hypothesis six also indicate that there is a significant difference in the means of LOGAMNT_PRO (M = 6.748 SE = .114) and LOGAMNT_INST (M = 7.040 SE = .124), t(45) = -4,66, p < 0.001. Therefore our result suggests that the fraud amount in the institutionalization stage increases compared to the proliferation stage. Thus, we can accept hypothesis six.
Table 4.8
Paired samples statistics
Observations Mean Std. deviation Std. Error mean
LOGAMNT_INI 70 6.422 .884 .106
LOGAMNT_PRO 70 6.900 .802 .096
LOGAMNT_PRO 46 6.748 .802 .114
LOGAMNT_INST 46 7.040 .840 .124
Dependent t-test
Observations Mean Std. deviation T df. Sig. LOGAMNT_PRO – LOGAMNT_INI 70 -.478 .459 -8.709 69 .000* LOGAMNT_INS – LOGAMNT_PRO 46 -.293 .426 -4.660 45 .000*
*Significant at the p < 0.01 level
Control Variables
For every significant relationship we test whether the control variables have influence on the dependent variable. For the numerical variables hypotheses one, two, three and four we performed an ordinary regression. For the dummy variable hypothesis four we performed a logistic regression. Table 7.1 (appendix) shows the results of the regressions regarding the control variables. CTRFIRMS, CTRBIG4 and CTRFINCR have no significant influence on the variables of hypotheses one to four. For hypotheses five and six, CTRFIRMS seems to have an influence on LOGAMNT_INI and LOGAMNT_PRO. This can be explained by the fact that CTRFIRMS is a proxy of the total assets, and therefore could have an influence on
29 the fraud amount. Additionally, we composed four new variables for hypothesis one, two, five and six. We calculated the difference between the variables NMBR_INI and NMBR_PRO; NMBR_PRO and NMBR_INST; LOGAMNT_INI and LOGAMNT_PRO and LOGAMNT_PRO and LOGAMNT_INST. The new variables are DIF_NMBR_INI_PRO, DIF_NMBR_PRO_INST, DIF_AMNT_INI_PRO and, DIF_AMNT_PRO_INST. As stated in table 7.2 (appendix), only between DIF_NMBR_INI_PRO and the control variable CTRBIG4 seems to be a significant relation.
5 Discussion and Conclusion
In this study, we examined the development of the fraud regarding the number of people involved, its major players and the fraud amount. The research question was “How do fraud schemes develop over time regarding the fraud amount and its major players?” The basis for our study was the theory of Palmer and Maher (2006), in which financial statement fraud is classified in several stages. These stagers are the initiation stage, proliferation stage, and the institutionalization stage. We examined the development of the number of people involved in these stages and we concluded that the number of people involved is significantly higher in the proliferation stage compared to the initiation stage. This was in line with our expectations in hypotheses one. The number of people involved in the institutionalization increased compared to the proliferation stage; this was not line with our expectations described in hypotheses two. The theory of Ashforth and Anand (2003)could explain the outcome of this result. They argue that during a fraud scheme there could be newcomers in the process who will be socialized and will be part of the fraud scheme. This could explain the fact that the number of people also increased in the institutionalization stage compared to the proliferation stage. Furthermore, in hypothesis three we examined whether the CEO is the initiator of the fraud. We could not conclude that the CEO is always the initiator of the fraud, however, if the CEO is involved, it is likely that he is the initiator of the fraud. Furthermore, we accepted hypothesis four, we can conclude that financial middle management is likely to be recruited in the proliferation stage. In hypotheses five and six we examined the development of the fraud amount during a financial statement fraud scheme. We concluded that the fraud amount increases in the proliferation stage compared to the institutionalization stage. Moreover, the fraud amount also increases in the institutionalization stage compared to the proliferation stage. For these reasons, we accepted both hypotheses five and six.
30 Despite the fact we found significant results, our study is subject to some limitations. First, our variables are based on hand-collected data. Which means that collecting the data is subject to the judgment of the collectors which could lead to less reliable data. Secondly, the time span of our stages is based on assumptions since no statistical evidence was available. Therefore, there could be a bias in the determination of the timespan of each stage what makes our results less reliable. We must better understand how long it exactly takes to recruit the team needed to perpetuate the fraud and how long it takes to fully implement the fraud in the institutionalization stage. Therefore, we recommend examining the duration of each stage in further research. The data we collected was from the SEC in the USA. As mentioned in the methodology section this data is subject to the selection criteria of the SEC, so the data is not randomly selected. Moreover, financial statement fraud is not proven in some cases. A possible consequence is that our data is not fully extracted from financial reporting fraud cases. Additionally, our data is collected from complaints in the U.S.A. It is possible that the results of our study are therefore not generalizable for other countries. The last limitation is the number of cases that were applicable to our hypotheses regarding the fraud amount (hypotheses five and six). Since the fraud amount was not clearly described in some complaints of the SEC, we only had 70 and 46 applicable cases for hypothesis five and six, respectively. Therefore it is questionable whether these results are generalizable.
We tried to show insight into the process of fraudulent financial reporting. But, there are some subjects that were not in our scope and are possibly interesting for further research. The scope of our major players was the involvement of the CEO and financial middle management. As stated in Beasley et al. (1999; 2010;) and the ACFE reports of 2016 and 2018, there might be more players involved. For example, the CFO, other C-level managers or salespeople could also get involved in a financial statement fraud scheme. It is interesting for further research to examine to what extent these functions are involved, and whether they initiate financial statement fraud, or in what stage they are recruited. Additionally, also the use of different fraud techniques related to stages, functions or fraud amount could be an interesting field of research since Beasley et al. (2010) found that there are several fraud techniques. For example, improper revenue recognition or the overstatement of existing assets.
The results of our study have some implications for practical purposes. Insight is gained on the development of the fraud process of fraudulent financial reporting regarding its major
31 players and the fraud amount. Our results of hypothesis one and two could gain insight for the enforcement division of the SEC. They could better understand how a fraudulent financial reporting scheme develops over time regarding the number of people involved. Additionally, the increasing fraud amount during the stages shows the importance of detecting fraudulent financial reporting in the beginning phase of the fraud scheme. In this way audit committees and auditors might prevent the organization for escalating the fraud. Lastly, we concluded that when the CEO is involved in the fraud, it is likely that he is the initiator of the fraud. So, for policy makers this could be useful in order to develop litigation or guidelines for controls preventing fraudulent financial reporting.
6 References
Albrecht, C., Holland, D., Malagueño, R., Dolan, S., & Tzafrir, S. (2015). The role of power in financial statement fraud schemes. Journal of Business Ethics, 131(4), 803-813. Anand, V., Dacin, M. T., & Murphy, P. R. (2015). The continued need for diversity in fraud
research. Journal of Business Ethics, 131(4), 751-755.
Argote, L., & Ingram, P. (2000). Knowledge transfer: A basis for competitive advantage in firms. Organizational behavior and human decision processes, 82(1), 150-169.
Albrecht, C., Holland, D., Malagueño, R., Dolan, S., & Tzafrir, S. (2015). The role of power in financial statement fraud schemes. Journal of Business Ethics, 131(4), 803-813. Anand, V., Dacin, M. T., & Murphy, P. R. (2015). The continued need for diversity in fraud
research. Journal of Business Ethics, 131(4), 751-755.
Argote, L., & Ingram, P. (2000). Knowledge transfer: A basis for competitive advantage in firms. Organizational behavior and human decision processes, 82(1), 150-169. Ashforth, B., Anand, V. 2003. The Normalization of Corruption in Organizations. Research
in organizational Behaviour, 25: 1-52.
Association of Certified Fraud Examiners. (2016). Report To The Nations. 2018 Global Study on Occupational Fraud And Abuse (pp. 1-80). ACFE.
Association of Certified Fraud Examiners. (2018). Report To The Nations. 2018 Global Study on Occupational Fraud And Abuse (pp. 1-80). ACFE.
Bandura, A. (1999). Moral disengagement in the perpetration of inhumanities. Personality and Social Psychology Review, 3: 193–209.
32 Beasley, M. S., Carcello, J. V., & Hermanson, D, R. (1999). Fraudulent financial reporting
1987–1997. An analysis of U.S. public companies. Durham: Committee of Sponsoring Organizations of the Treadway Commission (COSO).
Beasley, M. S., Carcello, J. V., Hermanson, D. R., & Neal, T. L. (2010). Fraudulent financial reporting 1998–2007: An analysis of U.S. public companies. Durham: Committee of Sponsoring Organizations of the Treadway Commission (COSO).
Bell, T. B., & Carcello, J. V. (2000). A decision aid for assessing the likelihood of fraudulent financial reporting. Auditing: A Journal of Practice & Theory, 19(1), 169-184.
Brown, T. J. (2014). Advantageous comparison and rationalization of earnings management. Journal of Accounting Research, 52(4), 849-876.
Brown, T., Rennekamp, K. M., Seybert, N., & Zhu, W. (2014). Who stands at the top and bottom of the slippery slope?.
Brickey, K. F. (2003). From Enron to WorldCom and beyond: Life and crime after Sarbanes- Oxley. Wash. ULQ, 81 - 357.
Brief, A. P., Buttram, R. T., & Dukerich, J. M. (2001). Collective corruption in the corporate world: Toward a process model. Groups at work: Theory and research, 471 - 499. Burns, N., & Kedia, S. (2006). The impact of performance-based compensation on
misreporting. Journal of financial economics, 79(1), 35-67.
Campello, M., Graham, J. R., & Harvey, C. R. (2010). The real effects of financial
constraints: Evidence from a financial crisis. Journal of financial Economics, 97(3), 470-487.
Cressey, D. R. (1950). The criminal violation of financial trust. American Sociological
Review, 15(6), 738-743.
Davis, S., DeZoort, F. T., & Kopp, L. S. (2006). The effect of obedience pressure and
perceived responsibility on management accountants' creation of budgetary slack. Behavioral Research in accounting, 18(1), 19-35.
DeAngelo, L. E. (1981). Auditor size and audit quality. Journal of accounting and
economics, 3(3), 183-199.
Dorminey, J. W., Fleming, A. S., Kranacher, M. J., & Riley Jr, R. A. (2010). Beyond the fraud triangle. The CPA Journal, 80(7), 17.
Efendi, J., Srivastava, A., & Swanson, E. P. (2007). Why do corporate managers misstate financial statements? The role of option compensation and other factors. Journal of
33 Eshleman, J. D., & Guo, P. (2014). Do Big 4 auditors provide higher audit quality after
controlling for the endogenous choice of auditor?. Auditing: A Journal of Practice &
Theory, 33(4), 197-219.
Farber, D. B. (2005). Restoring trust after fraud: Does corporate governance matter?. The
Accounting Review, 80(2), 539-561.
Free, C., & Murphy, P. R. (2015). The ties that bind: The decision to co-offend in fraud. Contemporary Accounting Research, 32(1), 18-54.
Gino, F., & Bazerman, M. H. (2009). When misconduct goes unnoticed: The acceptability of gradual erosion in others’ unethical behavior. Journal of experimental Social
psychology, 45(4), 708-719.
Hogan, C. E., Rezaee, Z., Riley Jr, R. A., & Velury, U. K. (2008). Financial statement fraud: Insights from the academic literature. Auditing: A Journal of Practice &
Theory, 27(2), 231-252.
Huang, H. W., Rose-Green, E., & Lee, C. C. (2012). CEO age and financial reporting quality. Accounting Horizons, 26(4), 725-740.
Hunton, J. E., Hoitash, R., & Thibodeau, J. C. (2011). Retracted: the relationship between perceived tone at the top and earnings quality. Contemporary Accounting
Research, 28(4), 1190-1224.
Ivashina, V., & Scharfstein, D. (2010). Bank lending during the financial crisis of 2008. Journal of Financial economics, 97(3), 319-338.
Jensen, M. C., & Meckling, W. H. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of financial economics, 3(4), 305-360.
Kelman, H. G. (1973). Violence without moral restraint: Reflections on the dehumanization of victims and victimizers. Journal of social issues, 29(4), 25-61.
Martin Dougall, S. (2019). Fraud Barometer: October 2016 – September 2017. Retrieved from https://home.kpmg/au/en/home/insights/2018/02/fraud-barometer-october-2016-september-2017.html
Lee, G., & Xiao, X. (2018). Whistleblowing on accounting-related misconduct: A synthesis of the literature. Journal of Accounting Literature, 41, 22-46.
Linke, K. G. B. (2012). Finance middle managers and financial statement fraud. Groningen: University of Groningen, SOM research school.
34 Morales, J., Gendron, Y., & Guénin-Paracini, H. (2014). The construction of the risky
individual and vigilant organization: A genealogy of the fraud triangle. Accounting,
Organizations and Society, 39(3), 170-194.
Ndofor, H. A., Wesley, C., & Priem, R. L. (2015). Providing CEOs with opportunities to cheat: The effects of complexity-based information asymmetries on financial reporting fraud. Journal of Management, 41(6), 1774-1797.
O'Connor Jr, J. P., Priem, R. L., Coombs, J. E., & Gilley, K. M. (2006). Do CEO stock options prevent or promote fraudulent financial reporting?. Academy of Management
Journal, 49(3), 483-500.
Palmer, D., & Maher, M. W. (2006). Developing the process model of collective corruption.
Research in Organizatinal Behavior, 28(4), 107-135.
Rezaee, Z. (2005). Causes, consequences, and deterence of financial statement fraud. Critical
Perspectives on Accounting, 16(3), 277-298.
Ramamoorti, S. (2008). The psychology and sociology of fraud: Integrating the behavioral sciences component into fraud and forensic accounting curricula. Issues in Accounting
Education, 23(4), 521-533.
Reckers, P., & Samuelson, M. (2016). Toward resolving the debate surrounding slippery slope versus licensing behavior: The importance of individual differences in accounting ethical decision making. Advances in accounting, 34, 1-16.
Rosner, R. L. (2003). Earnings manipulation in failing firms. Contemporary Accounting
Research, 20(2), 361-408.
Schnatterly, K., Gangloff, K. A., & Tuschke, A. (2018). CEO Wrongdoing: A Review of Pressure, Opportunity, and Rationalization. Journal of Management, 44(6), 2405-2432.
Tillman, R., & Indergaard, M. (2007). Control overrides in financial statement fraud. Report to the Institute for Fraud Prevention.
Suh, I., Sweeney, J. T., Linke, K., & Wall, J. M. (2017). Boiling the Frog Slowly: The Immersion of C-Suite Financial Executives into Fraud. Journal of Business Ethics, 1-29.
Sutherland, E. H., Geis, G., & Goff, C. (1983). White collar crime: The uncut version (Vol. 58). New Haven, CT: Yale University Press.
