Internal Corporate Governance Practices and Choice of External Auditor in Nigeria: A Logistic Regression Analysis

The research provided empirical evidence on how internal corporate governance practices were proxied by board size, board independence, and share ownership by institutional investors with firm size and leverage as control variables. It might influence the choice of the external auditor, (a dummy proxied with the Big4 versus non-Big4 audit firms dichotomy). The sample was composed of 27 purposely selected quoted non-financial firms spread across 10 sectors on the Nigerian Stock Exchange (NSE). There were 189 firms/year of dataset observations. These secondary panel data were sourced mainly from selected firms’ annual reports and accounts from 2011 to 2017. Moreover, descriptive analysis and test of mean difference were conducted, while the panel logistic regression was adopted as the estimation method. The test of mean difference reveals that many firms with larger board size, board independence, and considerably higher institutional investors engage Big4 auditors. Meanwhile, firms with higher leverage employ non-Big4 auditors. The results from the multivariate analysis show that key determinants of the choice of external auditors are board independence and firm size. This suggests that firms have a higher propensity of choosing a Big4 audit firm as the number of independent board members, as well as their increase in sizes. These findings are mostly consistent with previous studies.


INTRODUCTION
The audit of corporate entities in Nigeria is statutory. Hence, it is compulsory. Consequently, the appointment of an external auditor, who is engaged to audit these entities, is provided for in the Companies and Allied Matters Act (CAMA) Cap C20, Laws of Federal Republic of Nigeria, 2004. However, there are no regulations that explicitly state the auditor type, rather and at best, what is provided for CAMA are the required qualifications of corporate auditors. Even then, the law only lists those that are disqualified from acting as companies' auditors, and nothing more. This leaves the choice of external auditors to the discretion of members or directors, as the case may be, though it is guided by the practicality of the engagement and ethics. This is unlike the public sector, in which the Auditor-General for the Federation (AGF) compiles a list of acceptable auditors and sends it to government ministries, departments, and agencies from which they can select. The Nigerian audit market is composed of large, medium, and small audit firms. It is divided between the Big4 and non-Big4 professional practicing firms. The Big4 is composed of Pricewaterhouse Coopers (PwC), Ernst and Young (EY), KPMG, and Akintola Williams Delloite (AWD). The last one is the outcome of a merger between April 1999 and May 2004 between Akintola Williams & Co. and Deloitte & Touche. Although the appointment of the external auditor is simple, the specific choice of auditor-type is ultimately engaged. It may be associated with the interplay among key stakeholders (investors, directors, regulators, and others) who typically constitute the corporate governance mechanisms within corporate entities (Soyemi, 2015b).
Corporate governance, as a composite concept has evolved into the fore-front of hot and consistent research by researchers spanning over three decades. According to Kumar and Zattoni (2014), it has grown initially from addressing basic issues around governance to develop a conceptual agenda to address these issues to appraise the function and usefulness of those charged with governance of the entity. Nevertheless, it is also noteworthy to identify Wong (2016), who posited the place of institutional settings, which also shaped the landscape of corporate governance structure in every nation. The political undertone, usually associated with appointment and selection procedures, especially politically connected firms, may also play a major role, especially in a country like Nigeria (Guedhami, Pittman, & Saffar, 2014). This is not within the scope of this research.
The choice of the Big4 auditors by most firms may not be unconnected with the consensus among practicing accountants. They tend to equate firms with good audit quality. This view is not in any way significantly different among researchers (Libby, Rennekamp, & Seybert, 2015;Quick, Schenk, Schmidt, & Towara, 2018). This places both theoretical and practical significance on the choice of external auditors. Stemming from this, a sizeable number of them are mostly engaged by firms to conduct their annual audit (Soyemi, 2014;Soyemi, 2015a;Soyemi, 2015b), unlike the non-Big4 firms. This trend may not also be isolated from the position. A trade-off always exists as the engagement of these high-quality professional firms or otherwise to advance good corporate governance practices or prolong apparent gains from relatively weak corporate governance practices (Lin & Liu, 2009). It may be true, especially for developing economies. For example, Nigeria, which the corporate governance landscape is far from global best practices, is very weak for now. The aftermath of the financial crisis in the 2000s has renewed the efforts towards how firms are governed. Auditors move away from the provision of non-core audit services (Albring, Robinson, & Robinson, 2014). Hitherto, in Nigeria, most firms are closely connected to family lineage by not leaving out firms in financial services. This situation is similar to emerging economies generally associated with weaker legal environment (He, Rui, Zheng, & Zhu, 2014). The minority shareholders are barely protected, and close family ties are prevalent. This is acknowledged as applicable, prior to now, in Taiwan (Cho & Wu, 2014;Hsu, Lin, & Tsao, 2018), Malaysia (Husnin, Nawawi, & Puteh Salin, 2016), Bangladesh (Khan, Muttakin, & Siddiqui, 2015), India (Kulkani & Maniam, 2014), and Indonesia (Darmadi, 2016).
The role of external auditors is essentially to reduce information asymmetry emanating from agency theory. It states that the separation of ownership from management may lead to a conflict of interest between these two major groups. The financial reports prepared by management may be misleading with the presence of errors and material misstatements. Therefore, auditors are engaged to critically report on the financial statements and feedback on their truth and fairness. They also see whether they have been prepared in all materials in compliance with applicable financial framework and legislations. Arising from the role of external auditors in providing credibility to the financial reports prepared by those charged with governance of the entity (management), there is the possibility that a complementary (Hsu, Lin, & Tsao, 2018) versus substitute (Nasrudin, Mohamed, & Shafie, 2017) in which association exists between auditor-type and internal corporate governance practices. Both tendencies have been documented and supported by Cho and Wu (2014). These two alternative possibilities are brought to reality through various research. Lin and Liu (2009) showed a complementary relationship in their study of the effect of internal corporate governance practices on auditor choice. They used ownership concentration, size of the supervisory board, and Chief Executive Officer (CEO) duality. They found that firms with a larger controlling shareholder, smaller board size, and CEO/ Chairman position held by one person are less likely to hire high-quality auditor (using top10 versus non-top10 auditors dichotomy). This result suggested that firms with weak corporate governance mechanisms were inclined to engaging low-quality auditors.
Similarly, Karaibrahimoğlu (2013) examined firm-level corporate governance influences on the choice of the independent external auditor in Turkish environments. Corporate governance practices were managed with board independence, the board size, CEO duality, size of the audit committee, and institutional ownership in controlling firm size, leverage, and profitability. Using an empirical model based on multinomial logit and panel regression analysis with 805 datasets for five years from 2005 to 2009, the researcher revealed a significantly positive influence of board size, ownership concentration, and firm size with a negative and statistically negative tendency with board independent. Overall, the selected firmlevel corporate governance practices were accounted for only 22,4% of the variations in the auditor choice. Furthermore, Ianniello, Mainardi, and Rossi (2013) also conducted an empirical inquiry on how internal corporate governance, represented by board independence, CEO duality, and board size, affected the choice of external auditor among sampled Italian listed companies with 667 firms/year observations during 2007-2010. Although findings from their study displayed positive and significant influence with the board and firm size, negative and significant effect was shown by CEO duality with the choice of the Big4 professional firms. Overall, the selected internal corporate governance practices were accounted for only 19,3% of the variations in the auditor choice.
Moreover, Olowookere and Inneh (2016) examined effective determinants of auditor choice between 50 quoted manufacturing companies in Nigeria. Unlike previous research, this research adopted the use of primary data gathered through a questionnaire. Those were administered on selected shareholders of a few manufacturing firms. Rather than corporate governance issues, auditor/auditee attributes were studied based on Oxera Consulting Ltd (2006). It includes technical skill, specific-sector expertise, international coverage, management preference for the specific auditor, long-term relationship with the incumbent auditor, the reputation of audit firm with investors, reputation of audit firm with corporate broker, and reputation of audit firm with other external advisers using the common Big4 or non-Big4 dichotomy as the dependent variable. With a very low-level pseudo coefficient of determination (R 2 ) of 6,20%, only international coverage and long-term relationship with incumbent auditors depicted positive and negative significant relationship with the choice of external auditors. The others were not effective factors to be considered in the selection of external auditors. While contributing to the auditor choice debate, Aslan and Aslanertik (2017) also examined both firm (corporate governance inclusive) and IPO attributes, which were capable of influencing the choice of highquality auditors with 745 firms/year observations in Turkey. The IPO attributes were unique here. Corporate governance factor displayed positive and significant effects on the selection of external auditors for both firm and IPO age models with McFadden R 2 of 21% and 22%, respectively.
A similar study in internal corporate governance mechanisms and decisions to auditor choice in Malaysia is conducted by Nasrudin, Mohamed, and Shafie (2017). They analyzed the public entities companies listed on the mainboard from 2006 to 2015. Using a logit regression model, the researchers revealed that companies with less concentrated ownership structure, larger membership of audit committee and board, and lower independent directors were most likely to engage the services of the Big4, that was high-quality auditors. Then, Dwekat, Mardawi, and Abdeljawad (2018) investigated the impact of corporate governance mechanisms on auditor choice among 46 listed Palestinian corporations covering 2013-2015. Their findings were that firms with high ownership concentration, the larger board size, subsistence of audit committee, and large firm size tend to engage high-quality auditors and otherwise with higher gearing levels. This implies that quoted Palestinian firms saw the hiring of high-quality auditors as complementary to further reinforce the good corporate governance practices.
The researcher is motivated by the need to join the research in this area among emerging economies. Similar to others, the Nigerian environment provides a rich institutional background in the study of corporate governance practices as it relates to auditor choice. This research provides additional information on the influence of internal corporate governance practices on the choice of the external auditor. Thus, it contributes to the literature in Nigeria. The Nigerian context provides a rich background in the area of auditing and assurance, as the country is composed of varieties of the small, medium, and large professional practicing firms, both indigenous and otherwise. While it is expected that this current study provides plausible empirical results to decisions regarding audit choice, especially in an emerging economy like Nigeria, the dynamism of the corporate governance landscape is also revealed. It provides public company boards, policymakers, and stakeholders with verifiable explanations to ensure that informed decisions are made.

METHODS
The sample for this research consists of 27 quoted non-financial firms, spread across 10 sectors, excluding the financial sector on the Nigerian Stock Exchange (NSE) (see Table 1). These firms are selected using a stratified random sampling technique with each sector representing each stratum. The sample period is a 7-year covering from 2011 to 2017, amounting to 189 firm-year observations. The description of the variables used in this research alongside their measurement is as shown in Table 2. The varying sources from previous researchers are as well given.    (2012) Big4 it = β 0 + β 1 BS it + β 2 BI it + β 3 IO it + β 4 lnTA it + β 5 LEV it + ɛ it (1) The description and measurement for both the dependent and independent variables are as shown in Table 2. However, β 0 , β 1 -β 5 , and ɛ are the model intercept, regression coefficients, and error term, respectively. The panel logistic regression is the estimation technique.

RESULTS AND DISCUSSIONS
The distribution of the auditor type is presented in Figure 1. As depicted in Figure 1, about 78,84% of listed firms in Nigeria use the Big4 auditors (PwC, EY, KPMG, or AWD). Meanwhile, 21,16% of listed firms use non-Big4. This suggests that most listed firms in Nigeria engage the Big4 auditors. It applies to most Nigerian firms, as documented by Soyemi (2015a). Figure 2 presents the distribution of these auditor types across listed companies in the NSE. From Figure 2, most of the Big4 auditor account for at least 11,1% of the market. Moreover, non-Big4 accounts for less than 4% of the market. These two pictorial presentations are reinforcing in themselves and reflecting the true representation of the audit market in Nigeria. For preliminary and prior to regression analysis, Table 3 shows the summary descriptive statistics for key variables considered in this research. The statistics presented are mean, median, standard deviation (SD), minimum value, maximum values, skewness (skew), kurtosis (Kurt), and Jarque-Bera. On the average, about 78,8% is higher for Italian firms (-87%) (Allegrini & Greco (2013) and lower for Palestinian corporations (75%) and Malaysian firms (-62%) (Nasrudin et al., 2017). Moreover, it is much lower for Turkish firms (45%) of listed companies using the Big4 auditors. Board size ranges from 4 to 19 members, with a mean and standard deviation of 10 members and 3 members, respectively. This is similar to what is obtained in Italian firms and Palestinian corporations, as stated by Ianniello et al. (2013) and Dwekat et al. (2018). Similarly, the number of independent board members ranges from 3 to 16 persons. On average, the listed companies have 7 independent board members. It implies that 70% of board members in the sampled firms are independent (very low among Italian firms asmuch (37,5%) and Turkish firms (3,6%), higher than Palestinian corporations (92%), and similar with the developed countries of US and UK as reported in Allegrini and Grego (2013)). Furthermore, about 83% of firms have institutional investors in their ownership structure. It implies a concentrated ownership pattern. It is similar to the practices in developing countries, such as Turkish firms (81%) and Palestinian corporations (65%). However, it is different to a more dispersed ownership structure in the US and UK. Then, the values for skewness are moderately skewed. It indicates that the data gathered are nearly symmetrical. The positive sign is noticeable in all the kurtosis values showing a t-distribution with a relatively sharper peak than the normal distribution. Overall, the data follows normal distribution as the values for the Jacque-Bera are significant, except for institutional ownership.
The control variables (firm size and leverage) are explained in Table 3. The mean value of firm size, proxied with the natural logarithm of total assets, is 16,819 ranging from 13,102 to 20,762 with a standard deviation of 1,80. Similarly, the average value for the debt-equity ratio is 63,747, with a standard deviation of 25,376. These values range from 6,344 to 179,392.   Table 4 depicts the results of the t-test. It is a parametric test used to see whether internal corporate governance practices differ in the two groups of companies that engage the Big4 and non-Big4 audit firms. It also compares the key firm characteristics by Big4 and non-Big4 audit firms.
From the result, most firms with larger board size and board independence use Big4 auditors. However, there is no significant difference between firms that use Big4 and those that use non-Big4 auditors. On the other hand, firms with significantly higher institutional investors employ big4 auditors. Moreover, in line with expectation, the larger firms (firm size) use more Big4 auditors compared to non-Big4 auditors. This may be a result of high audit quality that is usually associated with the Big4 audit firms, typically rooted in previous research. Aside, the complexity involved in the conduct of the audit of large firms, it is characterized by heavy transactions, deployment of ICT, large branch networks, and others.
On the contrary, firms with higher leverage use more non-Big4 auditors compared to Big4 auditors. This result indicates the existence of both complementary and substitute role of engaging the Big4. It is complementary for firms with a large number of institutional investors and a larger size. However, it is a substitute for firms with higher gearing ratios. Hence, they select non-Big4 professional firms. Table 5 shows the correlation coefficients and their association between the variables used in the model. Aside from being used as the strength of linear association, the correlation results are preliminary to confirm the presence of multicollinearity between the explanatory variables. The panel logistic regression result is presented in Table 6. Both FEM and REM are similar in coefficients and significance. From the results, the key determinants of audit quality are board independence (BI) and firm size (LnTA). Firms have a higher probability of choosing a Big4 auditor as the number of their independent board members increases. It suggests that firms with higher independent board members prefer employing a Big4 auditor compared to a non-Big4 auditor. Similarly, firm size has a positive impact on audit quality. Specifically, larger firms have a higher probability of choosing a Big4 auditor compared to non-Big4 auditors. These findings are in line with that of Karaibrahimoğlu (2013), Ianniello et al. (2013), Matonti, Tucker, and Tommasetti (2016), Aslan and Aslanertik (2017), and Dwekat et al. (2018). Karaibrahimoğlu (2013) showed a positive and significant relationship between firm size. However, it is negative for board independence and choice of Big4 rather than a positive and significant influence, as noted in this research. Ianniello et al. (2013) observed a positive and significant association between firm size and auditor choice and board size. This result is similar to the findings in this research. However, the findings of Houqe et al. (2015) and Matonti et al. (2016) are contradicted with this research. Furthermore, the findings are in line with Dwekat et al. (2018). They showed a positive and significant effect between firm size and auditor choice. On the other hand, other corporate governance measures, the board size, institutional ownership, and leverage are not significant determinants of audit quality in Nigeria. It suggests that the board size, the proportion of institutional ownership, and the amount of leverage do not affect the choice of auditor (Big4 or non-Big4) in Nigeria.

CONCLUSIONS
The findings of this research clearly show that board independence and firm size are the most important factors. Those are considered and given priority in the selection of external auditors in Nigeria. This is not only plausible but also in line with the previous findings from previous researches in this area. A notable research and policy implication is the renewed interest and calls on emerging economies to see their corporate governance landscape as a great influencing factor in the selection of auditors.
It consequently affects financial reporting quality. Therefore, it reduces the incidence of information asymmetric among firms. The role of the Big4 auditors in the Nigerian context is complementary. It gives kudos to the emerging corporate governance landscape in this regard. However, concerning board size, institutional ownership, and leverage are not effective determinants of external auditor choice. It may be due to the weaknesses in these governance practices. This is not only applicable to the Nigerian environment but also generally among developing nations that are usually associated with emerging corporate governance practices.
Building on previous researchers, this research has contributed to improving the scarcity of studies in this area. Hence, the researcher adds the literature on auditor choice, especially among emerging countries. In addition, it is about the first empirical study that uses robust logistic regression analysis on secondary source data in Nigeria. Consistent with extant literature in this area, this research uses the typical Big4 versus non-Big4 dichotomy to proxy auditor choice. It is binary, which is associated with its limitations in terms of methodology, unlike the continuous variable like industry specialization.
Consequently, future research may focus on the use of a continuous variable to proxy auditor choice, alongside with a significant increase in the sample size, inclusive of firms in the financial services sector, and coupled with elongated years of coverage. Therefore, it will be improved on the firm/year data observations. The use of system equations and dynamic models as the estimation technique is also suggested in future research in this area.