“Upper echelons’ personality traits and corporate earnings management in Nigeria”

ARTICLE INFO Taleatu Taofiki Akinwumi, Adetula Dorcas Titilayo and Iyoha Francis Odianonsen (2020). Upper echelons’ personality traits and corporate earnings management in Nigeria. Problems and Perspectives in Management, 18(2), 90101. doi:10.21511/ppm.18(2).2020.09 DOI http://dx.doi.org/10.21511/ppm.18(2).2020.09 RELEASED ON Monday, 04 May 2020 RECEIVED ON Saturday, 14 December 2019 ACCEPTED ON Monday, 24 February 2020


INTRODUCTION
Empirical research on earnings management behavior of failing firms developed within the last ten years and remained scanty (Dutsi & Rausch, 2016). However, the knowledge of earnings management behavior in periods before corporate failure could be of use to stakeholders such as analysts and regulators. For instance, the outcome of this kind of studies could provide relevant variables for predicting financial failure, which should be of interest to analysts and regulators. Specifically, Beneish, Press, and Vargus (2012) and Campa and Camacho-Minano (2014) documented that upward earnings management in failing firms could be included as a proxy in the bankruptcy prediction model. Fraudulent earnings management has been ascribed to the key figures such as CEOs and CFOs of the corporate organizations (Hambrick, 1990;Hambrick & Mason, 1984). For instance, CEOs and CFOs were implicated in Enron and WorldCom scandals in the United States. CFOs were implicated in Satyam and Peregrine scandals, while both CEO and CFO were implicated in Cadbury scandal in Nigeria. However, aggressive earnings management could be explored by the upper echelons in corporate entities to hide financial crises. In Nigeria, Assets Management Corporation of Nigeria (AMCON) was established in 2010 to mop up non-performing loans from the banking sector. Out of the list of one hundred chronic AMCON debt defaulters with a debt of 5 billion nairas and above, ninety-eight were non-listed companies (AMCON, 2016;AMCON, 2018). This group of debtors is described in this study as troubled companies. So, the current wave of bank loan defaulters in the Nigerian banking system afforded a unique opportunity to study earnings management in troubled non-listed companies.
Personality trait of the top management has been copiously treated as a relevant variable in psychology research. It has also been identified as relevant variables in financial reporting research (Plockinger, Aschauer, Hiebl, & Rohatschek, 2016). The typical personality traits in psychology are the Dark Triad traits, namely Machiavellian, narcissistic, and psychopathic. Narcissistic traits studies have been extended to corporate risk research (Chatterjee & Hambrick, 2007) and earnings management studies (Duchon & Drake, 2009;Hales, Hobson, & Resutek, 2012). Research outcomes of these previous studies suggested that higher levels of narcissism in CEOs and CFOs might result in higher earnings management. In other words, narcissistic managers have been linked with upward manipulations of corporate performance for personal or corporate gains. Machiavellian and psychopathic traits are less examined in accounting and finance research (Murphy, 2012). Consequently, the current study examined the effect of these three traits on earnings management in troubled non-listed companies in Nigeria.

LITERATURE REVIEW
Narcissism is the first component of the Dark Triad, and it refers to exaggerated feelings of oneself (Jones, 2013). The second component is psychopathy, which is characterized by callousness, fearlessness, and aggressiveness (Furnham, Richards, & Paulhus, 2013). The third component is Machiavellianism, which reflects that an individual is willing to manipulate or exploit others to achieve his or her goals (Miller & Lynam, 2012). These three traits are related to a certain degree and have recently become objects of research in accounting and finance.
Empirical studies on the relationship between personality traits and financial reporting outcomes are relatively scarce. Recently, research on the influence of the "Dark Triad" personality in psychology (psychopathy, narcissism, and Machiavellianism) on financial reporting outcomes is gaining momentum ( However, narcissism appears to be the most studied components of the Dark Triad in financial accounting research. Amernic and Craig (2010) argued that extremely narcissistic CEOs tend to make accounting choices that present a good picture of the company's financial sta-tus. Olsen et al. (2014) investigated the relationship between CEOs' narcissism and market performance. The study found that narcissistic CEOs are more likely to explore real earnings management rather than accrual-based earning management to increase reported earnings per share. Hales et al. (2012) examined the relationship between narcissistic trait and managerial reporting and revealed that narcissism induced participants to inflate reported performance. Ham, Lang, Seybert, and Wang (2014) utilize the size of CFO and CEO signatures as a proxy for narcissism to determine the influence of the executive's personality on financial reporting outcomes. The results revealed that CFO narcissism could predict accruals and real earnings management, while CEO narcissism was not found to be associated with any of these financial reporting outcomes.
Furthermore, Kim (2018) examined the influence of CEO narcissism on organizational performance and earnings management and found that CEO narcissism was strongly associated with firm performance, but was not correlated with earnings management. These results suggest that narcissism is not a sustainable leadership style and that the personal characteristics of managers are important in determining firm performance. In a related study, Alex et al. (2015) investigated the relationship between CEO narcissism and earnings manipulation. The study provided evidence that firms with more narcissistic CEOs were more likely to engage in the manipulation of accounts to present better financial performance. Empirical evidence has also linked top management narcissism with audit outcomes. An example is an experimental study conducted by Johnson, Kuhn, Apostolou, and Hassell (2013), where it was found that the client's narcissism has a significant positive effect on auditor's fraud risk assessment. Rijsenbilt and Commandeur (2013), in another study, examined the relationship between narcissism and fraud. The research outcome showed that narcissistic CEOs had a higher tendency to engage in fraud.
Some experimental studies have revealed that managers with stronger levels of the Dark Triad of personality engage more in aggressive reporting (Majors, 2016;Zhao, Zhang, & Zhu, 2016). Nevertheless, Murphy (2012) conducted an experimental study on Machiavellianism and rationalization of misreporting. The study observed that the misreported participants displayed negative emotions, but misreported Machiavellians felt significantly less guilty than others. This outcome suggests that Machiavellians tend to be involved in financial misreporting, such as earnings management, without feeling guilty for doing so. However, caution has to be exercised while interpreting these results. As a result of the small sample size that characterized the experimental study, it would be necessary to validate the outcomes by a study that would involve a larger sample size. Consequently, a survey study involving a larger sample size was explored in this paper. It was not only observed that these previous studies are foreign but also noted that psychopathic trait was scarcely explored in financial accounting and reporting research. Hence, in line with the geographical, methodological, and variable exclusion gaps identified in the previous study, the following hypothesis, stated in its null form, was formulated for the study: H0: Upper echelons' personality traits have no significant effect on earnings management in troubled non-listed companies in Nigeria.

METHODS
The survey research design was adopted in this study. This research design was adopted because the study explored direct means of measuring personality traits by obtaining responses from targeted participants. By exploring Slovin's (1960) sampling size formula, eighty (80) non-listed companies (Appendix A, Triad to provide a more uniform assessment and to trim down the total length of measuring items. The test consists of 27 items (9 for each of the components) that must be rated on how much the participants agree with them. The test is Internet-based and should not take most people for more than five minutes. However, the study adopted Jones and Paulhus (2014) research instrument that captured the three traits to measure the CFOs' personality traits.
Subsequently, two hundred and forty (240) copies of the questionnaire (Appendix A, Table A4) were administered on CFOs and other financial officers in the sampled companies (3 participants per company), two hundred and nineteen (219) copies were recovered. In contrast, two hundred and four (204) copies, representing 85% response rate, were found suitable and were used for data analysis. While one hundred (100) observations are considered good for multiple regression analysis, Loehlin (1992) where EM is earnings management, which is a function of two variables: CAM -change of accounting methods and REM -real earnings management; ,

AGE
, GENDER EDU and TENURE are age, gender, educational level, and tenure of office of the respondents, respectively, while PTRA is the personality trait of the respondents, which is a vector of three variables, namely NAR -narcissistic trait, PSY -psychopathic trait, and MAC -Machiavellian trait. Each of the traits was measured with 5 items on the questionnaire.
The composite reliability of the specific constructs was ascertained with confirmatory factor analysis. The factor loadings exceed the minimum value criterion of 0.70, composite reliability of each construct also surpasses 0.80, and the construct average variance extracted estimate (AVE) was above 0.50, the benchmark (Appendix A, Tables  A1-A3). Since all three conditions for assessing the convergent validity were met, the degree of fitness of the measurement model is regarded as valid. Moreover, the hypothesis of the study was tested with simple and multiple linear regression analyses using AMOS SPSS software package.

RESULTS
The results of the study consist of descriptive statistics and test of hypothesis. Descriptive statistics includes percentages, means, and standard deviations in respect to the variables of the study. The hypothesis of the study was tested with multiple regression analysis and structural equation modeling.

Descriptive statistics
Descriptive statistics is presented in

Test of hypothesis
The results of the multiple regression analysis are contained in Tables 2-4. The model summary depicted in Table 2 shows the level at which the variation in earnings management could be explained by the variation in upper echelons' Machiavellian, narcissistic, and psychopathic traits (Model 2a). These three personality traits had significant positive relationships with earnings management (R = 0.819). The R 2 value of 0.671 indicates that 67.1% of the variation in earnings management can be explained by the variation in three personality traits. Model 2b also reflected significant positive relationship between earnings management and personality trait as a combined variable of the three traits (R = .829), while 68.7% of the variation in earnings management could be explained by variation in personality trait as a combined variable (R 2 = .687). This result implies that a unit increase in upper echelons' personality traits will lead to an increase in earnings management (upward earnings management).

DISCUSSION
This study attempted to find out whether the CFOs exhibit some personality traits that could have a significant effect on earnings management. The outcomes in respect to Machiavellian trait (Mean = 3.0) and psychopathic trait (Mean = 2.6) agreed to a large extent with the norms specified by the study of Jones and Paulhus (2014) for Machiavellian trait (Mean = 3.1) and psychopathic trait (Mean = 2.4) respectively. However, the finding in relation to narcissistic trait (Mean = 3.7) displayed by the CFOs was contrary to the norm specified by Jones and Paulhus (2014) for narcissistic trait (Mean = 2.8). This outcome suggests that the CFOs exhibit strong narcissistic trait, which may have implications for financial reporting outcomes.
The hypothesis sought to investigate the effect of upper echelons' personality traits on earnings management. The alternative hypothesis was supported by the results of the linear and multiple regression analyses. That is, the study posited that upper echelons' personality traits have a signifi-

CONCLUSION
This study documented a high possibility of corporate earnings management through the change of accounting methods and real earnings management. It was observed further that personality traits, most especially the narcissistic trait, have a significant positive effect on corporate earnings management in the sampled firms. Hence, it was concluded that the upper echelons' personality traits have a significant effect on corporate accounting choices and real earnings management in troubled non-listed companies in Nigeria.
The study established upward earnings management by individuals with any of three personality traits examined in this study. Economic implications of these outcomes include misallocation of resources by investors and aggravation of corporate debt crisis. Information asymmetry between the management and the investors/creditors could mislead investors and creditors into the provision of capital to these troubled companies. Consequently, it is recommended that during professional selection of individuals into upper echelons' positions in a corporate organization, such individuals should be subjected to personality trait test to recommend the individuals that fall within the norms for the appointment. It is also recommended to include the elements of psychology in the curriculum of accounting education in Nigeria to inculcate acceptable leadership traits in learners. This study included only the CFOs as respondents; Board chairpersons, audit committee chairpersons, and CEOs may be considered in future studies.