Capital Structure and Earnings Management Practices: Empirical Analysis in Sub-Sahara Africa

: This study assesses the impact of capital structure (CS) on Earnings Management Practices (EMP) in selected firms in sub-Sahara Africa. EMP was proxied by real EMP using Rowchodhury's (2006) model and accrual EMP measured by Khothari et al . (2005) model. The study applied the Generalized Method of Moments (GMM) estimator to data collected from the financial statements of two hundred and seventy-six (276) firms purposively selected from Nigeria, Ghana, Kenya, Tanzania, South Africa, and Zimbabwe from 2010 to 2020, given 3,069 observations firm-years. The results show that firms in Kenya, Nigeria, and Tanzania partake in real EMP through the structure of their capital. However, firms in Ghana, South Africa


Introduction
The main goals of financial reporting are to give stakeholders information about an enterprise's performance and assist users of accounting information in making wise choices.Accounting earnings and their components are especially crucial for stakeholders who want to gauge a company's performance and forecast its operating cash flows in the future.In the corporate world, problems with corporate financial distress and outright business failure have been a major problem.Many well-known American companies, including Enron and WorldCom, have failed in the recent past (in the early 2000s) for a variety of reasons, including dishonest profit manipulations, dishonest accounting practices, and persistently subpar performance, among others [25,29,45].This has brought vital and functional changes to the costs of running organizations.In order to avoid breaching debt covenants or to keep relationships for future debt raising, managers have strong incentives to manage earnings, including increasing their bargaining power and securing more favorable terms in debt contract negotiations [35,36,47].Additionally, the disclosure of violations, whether required or voluntary, may increase the cost of debt, reduce the efficacy of investments, or result in an acceleration of payments and a reduction in credit lines.By using Earnings Management Practices (EMP), Directors of companies may be able to convince investors that the company has stronger earning potential, resulting in a rise in the share price.Similarly, EMP can be conducted to influence borrowing costs, which will eventually jeopardize the interests of shareholders.Hence, this study examined the effect of Capital Structure (CS) on EMP.
The effect of CS on EMP can be seen from different theoretical perspectives, with inclusive opinions from different scholars.Because empirical evidence is still inconclusive, this study shed more light on assessing the effect of CS on EMP.Despite many studies on EMP in Practices: Empirical Analysis in Sub-Sahara Africa developed countries, which have tried to carry out more literature on capital structure and EMP [5,[15][16][17], there is still limited research in sub-Saharan African countries.More so, previous academic studies reflected that capital structure and EMP, especially in sub-Saharan African countries, were largely based on accrual EMP [21,31,48].Nevertheless, as managers can use accrual and real EMP as substitutes, the influence of capital structure should be considered on both techniques rather than only accrual.In addition, most of these studies considered a maximum of three countries and made use of pooled ordinary least squares and panel data analysis [28,42].Further, using less than ten years of observations as done in these studies may not be enough to show a true picture of the influence of capital structure on EMP.It is believed that a new study on the influence of a firm's CS on EMP will contribute significantly to the existing literature.Therefore, for a period of eleven years, this research evaluated the influence of CS on accrual and real EMP among selected listed firms in six sub-Saharan African countries.

Capital Structure and Earnings Management Practices
The repetitive choice of an accounting policy or set of reporting guidelines is a component of EMP.When preparing and presenting financial statements for particular reasons, managers frequently use EMP.It aims to present a revenue flow with a smaller deviation from trend than would have otherwise been seen [12,13,48].According to Musa A. F. and Muhammad A. I [25], EMP occurs when the Directors use their discretion in reporting transaction financial statement in a favourable manner rather than the ways it actually occurs in order to mislead some stakeholders about the true performance of the firms.According to research by Leuz et al. [22], EMP refers to the insider manipulation of a company's reported economic success in order to deceive some stakeholders or affect contractual outcomes.In addition, Wallace D. N. and Pornsit J. [53], describe EMP as the practice of using the latitude provided by accounting principles to give managers the ability to affect reported earnings, causing reported revenue to be higher or lower than it otherwise would be.
In terms of managing earnings, capital structure choices may occur in two ways.In the first instance, the degree of debt may be utilized as a management decision-control tool.Higher debt levels limit the power of opportunistic managers by reducing the cash flow available for discretionary spending on irrelevant assets [11,14,29,52].An excess of debt compared to equity in a company also raises the risk of insolvency.Thus, it would stand to reason that companies with more leveraged capital arrangements would exhibit less real or discretionary EMP, which can be described as the reverse leverage.Similarly, CS choices, however, may work against investors' interests and in favor of managers' selfish behavior.In order to fulfill the requirements, set forth by the debt covenants, managers may falsify the financial accounts [33,38,46].As a result, companies that are getting close to breaking debt covenants manage profits more aggressively.In light of these points, it appears that sub-Saharan Africa is a region where empirical problems with the fostering or restricting effects of debt on EMP, also known as leverage and reverse leverage effects, exist.

Theoretical Review
Signaling Theory: Spence (1973) initiated the signaling theory, and Ross (1977) later developed it.The theory looked at how managers and investors interacted in an environment of asymmetric knowledge.Therefore, a firm shows some signals through EMP to assess itself to other firms or to the industry.In fact, there is a lot of incomplete information between different business stakeholders in market economies.Due to their access to confidential information that enables them to communicate with a variety of investors and market players, managers are therefore thought to be the most knowledgeable party regarding the firm's future prospects [43].The analytical framework for this study was based on signal theory because a company sends signals through its EPM to compare itself to other businesses in its industry or to other industries.Managers can use EMP to increase earnings reported and accurately report the financial performance of the business by having knowledge of the expectations and future prospects of the company [57].Managers are able to share and report on confidential information about the company's future performance thanks to these accounting practices.As a result, market standards and manager expectations are brought into alignment.

Capital Structure and EMP
In order to understand the effect of CS and EMP, there have been various studies conducted in Asia, Europe, Africa, and the Pacific.In Asia, Lastari N. and Aeni N. [20] opined that capital structure positively influences accrual EMP.In the same vein, Alomush et al., Khanh T. and Thu J., Nanik L. and Nur A. [2,16,26] observed a positive and significant influence of capital structure on accrual EMP in their empirical findings conducted in Jordan, Vietnam, and Indonesia, respectively.However, Hassan T. M. and Abdulrahaman A. [14] in Bangladesh discovered a negative and significant effect of capital CS on accrual EMP, as well as Neeraj et al., [27] in their studies conducted in India.In Europe, the study observed the work of Mercedes et al. [23] that was conduct in Spain, where the authors observed a negative and significant effect of CS on accrual EMP.Similarly, negative and significant effect of CS were revealed in the works of Doukakis L. and Ugrin et al. [8,52], conducted across firms in European countries.In international settings, Ningrun et al. [29] worked across 30 countries; Paolo S. and Laora M. [38] conducted across countries in Latin America, where the studies observed negative and significant impact of CS on accrual EMP.However, Lemma et al. [21] worked across 44 countries using panel regression for 748 non-financial firms and discovered a positive and significant influence of CS on accrual EMP.
In Africa, Nyatichi et al. and Saline et al. [30,42] found a negative and significant influence of CS on accrual EMP.In the same vein, Okafor et al. [32] in a study conducted in Nigeria found negative and significant influence of CS on accrual EMP.However, Ogiriki and Iweias [29] revealed a positive and significant impact of capital structure on accrual EMP in another study conducted among manufacturing firms in Nigeria.More so, the work of Nelson et al. [28], conducted among firms in Kenya and Tanzania, discovered that CS has a positive and significant effect on accrual EMP.In addition, Elkalla T. [9] found that the capital structure revealed positive and significant influence on accrual EMP in MENA countries.More so, Bassiouny et al., Swai J., Waweru N. and Riro G. [6,48,54] in their studies carried out in Kenya, Egypt, and East Africa, respectively, found that capital structure positively and significantly influences accrual EMP.Thus, the authors observed that the majority of studies reviewed came up with different views on how capital structure can influence EMP and considered less than six countries, particularly in Africa.The study hypothesized as follows: H 01 : Capital structure does not significantly drive accrual EMP of firms in sub-Sahara Africa.

Capital Structure and Real EMP
In Asia, Anabelen et al., Mohdsuffian et al., [4,24] in their studies carried out in Malaysia and Korea, respectively, that capital structure has positive and significant effect on real EMP.Nevertheless, Rajeevan S. and Ajward R. [39] in Sri Lanka discovered a negative and significant relationship between capital structure and real EMP.With regards to Europe, the authors observed the study of Owusu et al. [37] that showed a negative impact of CS on real EMP, while Lara et al. [19] discovered a positive and significant influence of capital structure on real EMP.In an international setting, Anagnostopoulou S. and Tsekrekos A. [5] carried out a study among eighteen countries around the world, and the work of Kim et al. [17] conducted in North America found a positive and significant influence of capital structure on real EMP.Though, Alsharairi M. and Salama A. [3] found a negative influence of capital structure on real EMP in a study conducted in North America.In Africa, Elkalla T. [9] found that the capital structure had a negative influence on real EMP.The authors observed that the majority of studies reviewed were carried out on accrual EMP, particularly in Africa.This study is considered both accrual and real EMP to fill identified research gap.Thus, the study hypothesized that: H 02 : Capital structure has no significant influence on real EMP of firms in sub-Sahara Africa.

Design and Data
The study used ex-post facto research design while the population of the research consists of listed companies in sub-Sahara Africa.Using multi stage sampling technique, Stratified sampling method was used to divide sub-Sahara Africa into 4 regions, while 3 regions and 6 countries (Eastern: Kenya and Tanzania; Southern: South Africa and Zimbabwe; Western: Nigeria and Ghana) were considered for the study.The Central Africa region was excluded because the Douala Stock Exchange was the only stock exchange that belonged to the African Securities Exchanges Association in the region but had just two non-financial listed companies, resulting in inconsistent data needed for the study.A purposeful sampling method was employed to select NFLF companies with the data needed for the study from 2010 to 2020.Therefore, 76 out of 166 firms (45.8%) in Nigeria, 12 out of 31 firms (38.7%) in Ghana, 127 out of 250 firms (50.8%) in South Africa, 31 out of 63 firms (49.2%) in Zimbabwe, 26 out of 64 firms (40.6%) in Kenya, and 7 out of 25 firms (28%) in Tanzania were chosen, giving a total of 279 firms in the six chosen Sub-Saharan African nations.Data collected were analyse using Generalized Method of Moments (GMM) estimator.

Model Specification
AEMP was measure by DA while NDA is calculated using Khothari et al. [18] model adapted from the work of Orazalin N. and Akhmetzhanov R. [34].

Descriptive Statistics
The findings revealed that Real Earnings Management Practices (REMP) for Kenya, Tanzania, and Ghana companies have positive mean values of 0.025, 0.065, and 0.068, respectively, according to the descriptive statistics.For companies in South Africa, Zimbabwe, and Nigeria, the mean numbers were all negative (-0.018, -0.037, and -0.023, respectively).This indicates that businesses in Kenya, Tanzania, and Ghana participated in REMP to a greater extent than companies in South Africa, Zimbabwe, and Nigeria.Additionally, it was discovered that firms in Kenya, Tanzania, and Ghana had larger standard deviations than those in South Africa, Zimbabwe, and Nigeria, which is a sign of higher REMP volatility.With the exception of South Africa, the mean values for Discretionary Accrual Earnings Management Practices (DAEMP) were positive for companies across all of the study countries.This suggests that companies in Kenya, Tanzania, Zimbabwe, Nigeria, and Ghana participated in more income-increasing AEMP, with respective mean values of (0.016, 0.028, 0.066, 0.017, and 0.005).Furthermore, it was discovered that companies in those countries had higher standard deviations than companies in South Africa, suggesting higher volatility.Kenya's (0.316), Zimbabwe's (0.352), and Nigeria's (0.303) mean values for capital structure are greater than Tanzania's (0.182), South Africa's (0.164), and Ghana's (0.164) mean values.With regards to capital structure, Kenya's (0.316), Zimbabwe's (0.352), and Nigeria's (0.303) mean values for capital structure are greater than Tanzania's (0.182), South Africa's (0.164), and Ghana's (0.179).This suggests that businesses in Kenya, Zimbabwe, and Nigeria depended more on debt funding than equity financing.With mean values ranging from 4.567 to 4.886, there is no discernible difference in firm size between the various nations, with the exception of South African firms, which have a mean value of 5.592, suggesting that South African firms are marginally larger than those in other countries.

Correlation Matrix
To examine the correlations between the model variables, a Pearson correlation matrix, as shown in Table 3, was calculated.The study found that the majority of variable correlations are low, with all ˂ coefficients being 0.8, which is the threshold for multi-collinearity problems indicated by earlier study by (Khanh & Thu, 2019).

Diagnostic of the Variables
As shown in Table 4, the VIF was employed in the study to assess for multicollinearity diagnostic.The capital structure in Tanzania had the highest calculated VIF number of 1.09, and with the average VIF value of 1.03 across all variables.All VIF values, however, were considerably below the cutoff of 5, indicating that multicollinearity between the study model variables is not a major issue.

Effect of Capital Structure on Accrual and Real Earnings Management Practices
Table 5 shows the effect of Capital Structure (CS) on accrual and real EMP.The coefficients REMP and DAEMP (1-δ) as observed in Table 4 which is speed of adjustment represent the coefficients of the lagged real EMP t-1 and accrual EMPt -1 , respectively .The coefficient (1-δ) is negative and significant for Kenya, Tanzania, Zimbabwe, as well as Ghana, while it is positive and significant for South Africa and Nigeria.The results shown that CS in firms in Nigeria (western region), Kenya and Tanzania (eastern region) (β= 1.201; 0.002; 0.031; p > |t| = 0.000; 0.004; 0.002 ˂0.05, respectively) positively and significantly influence real EMP.Though, CS in firms in Zimbabwe (β =-0.062, p ˂ >|t|=0.0000.05), South Africa (β =-0.005; p>|t|=0.029˂0.05), and Ghana (β =-0.014; p>|t|= ˂ 0.033 0.05) negatively and significantly influence real EMP.

Discussion of Findings
The positive and significant influence of Capital Structure (CS) on real EMP in firms in Nigeria, Kenya and Tanzania implies that firms in those countries engaged in REM through the structure of their capital.This result agrees with earlier findings carried out by [4,7,26], while differ with the results of [37,42].Nevertheless, the negative and significant effects of CS on real EMP in firms in Ghana, South Africa and Zimbabwe suggests that firms in those countries do not engage in real EMP through the structure of their capital.The negative effect agrees with the research findings of [3,9] but differs from [12,24].Furthermore, the positive and significant impact of CS on accrual EMP among firms in South Africa and Ghana suggests that firms in those countries manipulate earnings through accrual EMP.This supports the research findings of [20,31] and is different from the findings of [30,32].However, the negative and significant effect of CS on accrual EMP among firms in Kenya, Zimbabwe, and Nigeria implies that low-geared firms in these countries do more accrual EMP, while levered firms do not engage in a higher level of AEMP.The findings are at variance with the empirical studies carried out by [23,27], but agree with the empirical findings of [2,30].These results imply that, in a group of firms that finance their operations more through debt than equity, real EMP is more preferred than accrual EMP by exploiting accounting choices to twist firms' earnings.Although, in highly geared firms, the attention of capital providers could be drawn to checkmate's operation and limit the chance to use accrual EMP.Thus, firms might switch to real EMPs, which are masked as daily operational activities and are more difficult to discover.
In addition, Positive and significant influence of FS on real EMP in Nigeria, South Africa and Zimbabwe indicate that the bigger the firms in Zimbabwe, South Africa, and Nigeria, the more those firms engage in real EMP, while a negative effect among firms in Kenya and Ghana means that the bigger the firms in Kenya and Ghana, the lower the firms in those countries participated in REMP.More so, the positive and significant effect of CS on accrual EMP in firms in Kenya, South Africa, Nigeria as well as Zimbabwe implies that the bigger the firms in these countries, the higher those firms engagement in accrual EMP.However, the negative and significant effect of CS on accrual EMP Ghana signifies that the bigger the firms in Ghana, the fewer firms are engaged in accrual EMP.

Conclusion
The study examines the influence of CS on accrual and real EMP in sub-Sahara Africa.Kenya, Tanzania, Zimbabwe, South Africa, Nigeria, and Ghana were the six countries selected for the study.Using data compiled by the Machameratios database of 279 selected listed firms across six countries from 2010-2020.The study concludes that firms in selected countries in sub-Sahara Africa substitute real and accrual forms of EMP.Therefore, the study recommends that capital providers in Nigeria, Ghana, Kenya, Tanzania, South Africa and Zimbabwe should maintain sufficient attention to both real and accrual EMP for sustainable leveraging and the management of opportunistic selections of accounting choices, but increase the use of real EMP.The study has contributed to the existing literature of accounting and EMP by extending measures of the effect of capital structure on earnings management practices to real earnings management techniques across six sub-Sahara African countries, compared to two or three countries in previous analyses.Further research could consider more than six countries as well as add more variables as a control variable in sub-Sahara African countries rather than only firm size.

Table 1 .
Measurement of Study Variables.

Table 5 .
GMM Results for effect of Capital Structure on Accrual and Real EMP.