“Capital structure and practices of accrual-based earnings management among non-financial Vietnamese listed firms”

This paper’s primary goal is to examine the influence of a firm’s capital structures on practices of accrual-based earnings management by doing empirical research covering 51 non-financial Vietnamese listed companies during a period from 2013 to 2022. To estimate accrual-based earnings management practices, the modified Dechow and Dichev (2002) model was mobilized. Then, a regression between earnings management estimated values based on accruals and a group of capital structure variables and control variables that are hypothesized to influence earnings management practices is performed. The feasible generalized least square model is used to address econometric issues. Empirical results reveal that activities for managing accrual-based earnings indicate a considerable adverse influence from institutional ownership. However, other hypothesized variables that are management ownership, ownership concentration, foreign ownership, and leverage do not have a determinant sign as expected. ROA, one of five control factors, has a favorable impact on accrual-based earnings management practices, whereas company size has a negative impact. The study provides useful information to investors and stakeholders for their making investment decisions in Vietnam. The empirical findings are also based for recommendations to control earnings management practices at Vietnamese listed enterprises to enhance accounting information quality, thus contributing to the sustainable development of the Vietnam Stock Exchange.


INTRODUCTION
The stated financial performance results of a firm are one of the most crucial components of its financial statements and have a great influence on the choices made by external users.Many important economic decisions such as raising capital, loan covenants, and executive team compensation are largely based on operational results publicly disclosed in firms' financial reports.Given the advantages resulted from information asymmetry, managers tend to manipulate actual operating results to maximize their own interests (Healy & Wahlen, 1999).Numerous academics and practitioners are drawn to the subject since it is widely known and regularly used in the empirical literature in terms of earnings management (Ghazali et al., 2015).Even though there are certain instruments implemented to ensure and enhance the transparent and reliable level of corporate financial reports, the possibility of corporation to intervene financial reports through earnings management practices still occurs since these practices are within the alternatives allowed by the accounting principles and standards (Degeorge et al., 1999;Rahman & Ali, 2006).This makes earnings management practices different from fraud as no violation against the accounting principle framework took place.However, these practices still result in improper financial information about the corporation.
Previous studies acknowledge the impact of internal and external monitoring methods on limiting earnings manipulation behavior (González & García-Meca, 2014;Ghazali et al., 2015).For the internal monitoring mechanisms, a more responsive system of a firm's corporate governance can limit the opportunities toward earnings management practices (González & García-Meca, 2014;Mensah & Boachie, 2023).For the external ones, among others, creditors' monitoring can prevent managers from inflating earnings (Ghazali et al., 2015).
In Vietnam, many listed companies recently reported different earnings before and after the audit of financial statements.Most notably, Dabaco announced a profit after tax of 150 billion dong in its 2022 financial reports.After the audit, the reported number has been reduced by 145 billion dong.The enterprise has explained that the cause of this is due to a change in estimating the completion of construction works.Danh Khoi Group reported a profit of 6.3 billion dong after tax in 2022 in its own reports.After the audit, it changed to a loss of nearly 73 billion dong.The company explained that external auditors asked to record a provision for bad debts of 92.5 billion dong.Another example is Nam Kim Steel who also increased his loss nearly doubled after the audit, recording a loss of 124.68 billion dong instead of a loss of 66.71 billion dong previously recorded, mainly due to the inflation in cost of sales resulting from provision for inventory devaluation (An, 2023).
Given the critical and emerging concerns related to reported earning quality in the country, this paper was written to look into the effects of capital structures proxied by ownership variables and leverage ratio on practices of earnings management among Vietnamese non-financial listed firms.

LITERATURE REVIEW AND HYPOTHESES DEVELOPMENT
Although earnings management (EM) is a widely researched term and draws attention of numerous researchers throughout the world, there is no unified concept of this act given by researchers (Beneish, 2001).Schipper (1989) described EM as the actions taken by management to generate the required amount of reported earnings with the goal of attaining some types of personal benefits within the established framework of generally accepted accounting principles.Healy and Wahlen (1999) indicate that EM happens when management team intentionally employs discretion to adjust financial statements to either mislead external users about the business's actual financial performance or influence the results of agreements based on financial statements.EM is management efforts to misstate the reported incomes in the ways of either magnifying or reducing reported accounting figures to antici-pated level (Akers et al., 2007).These definitions mentioned above have something in common with each other.Therefore, it can be stated that EM is the behavior of managers who driven by getting their own intended goals use accounting techniques within the prescribed principles, standards, and regulations to misinform external party users about the company's trustworthy financial situation.
In the agency theory, Jensen and Meckling (1976) highlighted that the split of management and ownership would create more conflicts, costs, and information asymmetry, which would stimulate managers to behave opportunistically (Fama & Jensen, 1983).As a result, efficient monitoring systems are required to minimize such agency conflicts and costs.An effective monitoring mechanism will help optimize the credibility and reliability of financial statements and which will reduce managers' capacity to manage profits (Klein, 2002).In light of the agency theory, it can be inferred that capital structure is one of critical factors that can mitigate EM practices.Concerning the impact of management ownership on managing earnings, according to Warfield et al. (1995), managers who own a small percentage of a company's shares have a strong motivation to manipulate accounting data to safeguard their meagre interests.On the contrary, when managerial ownership is higher and when management interests are closely associated with those of the corporation as a whole, it is expected that the divergences of interests between managers and stockholders reduce significantly.Studies by Klein (2002) and Ali et al. (2008) Bushee, 1998).However, institutional investors, on the other side, can also increase management incentives to engage in manipulating earnings.According to Pound (1988), institutional investors can collude with firm management.Institutional investors cannot play their supervisory responsibilities and vote against management as that can affect their existing and future business connections (Porter, 1992).Furthermore, if organizational owners focus on temporary financial results instead of long-term ones, they will not pay attention to management oversight activities (Bushee, 1998).
The concentration of ownership is scaled by the percentage of owned shares, usually from five percent, possessed by the investors.Debt-to-asset ratio is a leverage ratio that indicates how much debts in financing the company's assets.A higher debt ratio results in a higher risk for investors and creditors due to a higher dependence level of the corporation on its creditors and the higher debt burden.According to some earlier studies, managers of heavily leveraged businesses have considerable motivation to fudge profit to slack off on loan requirements (Jiang et al., 2008) After the reviews of prior background studies mentioned above, it is acknowledged that there is currently not much research in Vietnam on the influence of capital structures on EM practices and the prior research conclusions remain contradictory.
Therefore, the purpose of this study is to (re)examine the impact of firm capital structures proxied by ownership variables and leverage ratio on EM practices of Vietnamese listed companies over the course of 10 years, from 2013 to 2022.
The following hypotheses are developed and tested in the study based on evaluating pertinent papers as previously provided above: H1: There exists a negative impact of management ownership on EM.
H2: There exists a negative impact of institutional ownership on EM.
H3: There exists a negative impact of blockholders' ownership on EM.
H4: There exists a negative impact of foreign investors' ownership on EM.
H5: There exists a negative impact of leverage on EM.

METHOD
Quantitative method is used for this study.

Dependent variable and earnings management estimation model
The model of detecting EM through accrual variables was built by Jones (1991).The total accruals include discretionary accruals (DA) and non-discretionary accruals (NDA).EM is done through accrual accounting variables, which can be adjusted due to the flexibility of accounting principles.The total accruals (TA it ) is measured by taking net income after tax substract operating activities' cash flows of company i for the year t.
To determine the DA, it is necessary to estimate the NDA.There are several models that estimate NDA.
The Jones model (1991) determines the variable (NDA) according to the following equation: , where А i(t-1) : Total assets of company i for the year t-1; ΔREV it : Change in net revenue of company i for the year t with respect to year t -1; PPE it : Gross tangible fixed assets of company i for the year t; β 0 , β 1 , β 3 : the parameters estimated by the method of OLS of the coefficients α 0 , α 1 , α 2 in the model represented below: .
The Dechow et al. equation (1995) adds the increase and decrease variable of customer receivables into the equation to eliminate the effect of accrual sales due to increase in customer accounts receivable during the period.The model is:  To find out the most suitable model to estimate EM, this study runs regression of all four models indicated above and the results show that the modified Dechow and Dichev (2002) equation provides the highest adjusted-R 2 with 78.14%.This signifies that 78.14% of the variance in the dependent variable is explained by the independent factors.So that the modified model of Dechow and Dichev (2002) is the most suitable model for the observations in this study.Finally, relying on the modified model of Dechow and Dichev (2002), EM is estimated by taken the absolute value of DA it when DA it = TA it -NDA it .

Independent variables and control variables
The study comprises five independent variables and five control variables.Their names, measurement and expected sign are all presented in Table 1.

RESULTS
As indicated in Table 2, the average value of absolute EM is 0.0461, the minimum value is 0.000026, and the maximum value is 0.282, with a standard deviation of 0.0426.The average value of EM demonstrates that a large number of companies in the studied sample are managing earnings to achieve their goals in financial reports.The description of five independent variables shows that the minimum values are 0 for IO, FO, MO, and CO except 0.031 for LEV, while the maximum values are all above 0.90 ranging from 0.911 to 0.991, and the average values are between the range from 0.126 to 0.532 with standard deviations between 0.173 and 0.296.For ROA, the minimum value is -0.322, and the maximum value is 0.465; the average value is 0.077, indicating a moderate profitability level during the specified time frame.For BSIZE and FSIZE, the mean values are 5.92 and 11.72, respectively.For the dummy variable (AUD), the mean value is 0.31 which implies that one third of companies in the research sample have financial reports audited by Big4 auditing firms.The mean value of CF is 0.073 and its standard deviation is 0.130.FSIZE variables.The correlation coefficients indicate that the study model does not face severe multicollinearity problems as none of the correlation coefficients is above 0.8 (Gujarati, 2004;Hair et al., 2006).However, to test for the likelihood of multicollinearity issues between the independent and control variables, the variance inflation factors (VIF) and tolerance ratio (1/VIF) computations were still performed (Hair et al., 2006).As indicated in Table 4, the values of VIF are less than 10 and the values of tolerance ratio (1/VIF) are larger than 0.10 for the study model.This demonstrates that the study's model explaining capability is not affected by multicollinearity (Hair et al., 2006).The three statistical models, namely pooled OLS, FEM, and REM, are run, aiming to choose the most suitable model used in the research.The significant value (p-value) of FEM is 0.000 so that FEM is more appropriate than OLS.The Hausman's test result is 0.000 so that FEM is the better one com-pared to REM.Auto correlation and heteroskedasticity are also existed, so the FGLS model is run to fix the econometric problems.
As noticed in Table 5, the FGLS model's p-value is 0.0000 that is significant at the 5% level; therefore, it can be said that the dependent and independent variables indeed have a statistically significant relationship, so the model is appropriate, and the results are reliable.Among five independent variables, only IO has the p-value of 0.048, which is statistically significant at 5% level and the coefficient is -0.0001853.Hence, the hypothesis H2 is accepted, which means that institutional ownership has an unfavorable and statistically significant effect on EM.On the contrary, the p-value of other four independent variables are all above 5% level of significance; therefore, the hypotheses H1, H3, H4, and H5 are rejected.Two out of five control variables show the different significant impact on EM.While ROA positively impacts EM, FSIZE helps reduce EM practices as it signals a negative effect.The influences of other control variables are insignificant as the p-values are more than 5% significant level.

DISCUSSION
The regression model on the impact of capital structures on EM practices of Vietnamese listed companies is as follows: 0.0671214 0.0001853 0.0023048 0.0038921 .

CONCLUSION
Based on the sample of 51 Vietnamese non-financial listed from 2013 to 2022, the paper examines the influence of firms' capital structures on EM.The findings suggest that greater institutional ownership would lessen EM practices, while other capital structure variables have no impact on EM activities.
The study contributes to extent of the literature and empirical evidence about the capital structures' impacts on EM in Vietnamese listed firms.The findings can be considered as reference for companies' stakeholders, such as the State Securities Commission of Vietnam (SSC), creditors, investors, and shareholders to find out the existence of EM and the factors that affect EM practices.Therefore, regulations can be built up to restrain EM and contribute to higher quality of a firm's financial reporting.
This study cannot avoid some constraints.First, the small number of firm-year observations could impact the findings.This limitation may result from the small number of Vietnamese listed firms.Second, due to the lack of information disclosure, there are some other capital structures that cannot be includ- (2002) propose a new measurement model for earnings quality.According to the authors, accruals and cash flows (CFO) are correlated with earnings quality, in which the former adjust the latter over time.However, in the opinion of McNichols (2002), when calculating discretionary accruals, the Dechow and Dichev (2002) model does not account for the effect of long-term accruals.Therefore, McNichols (2002) expands the Dechow and Dichev (2002) equation by adding the changes in revenues and in tangible fixed assets as additional explanatory variables in estimating discretionary accruals.Hence, the Dechow and Dichev (2002) model is modified as follows: (Khanna & Palepu, 2000)s to exercise controls on company management with the aims of ensuring a better rate of return on their overseas investments.Furthermore, because of possessing up-to-date and superior management skills and techniques, foreign investors may exercise better monitoring activities over managers in developing economies(Khanna & Palepu, 2000).
(Ali et al., 2008;Alves, 2012)t al., 2002;Zhong et al., 2007)ers' ownership,Dechow et al. (1996)andYeo et al. (2002)indicate that the control of a company's blockholders is the same as that of institutional investors on earning manipulation practices.This indicates that blockholders' monitoring and controlling help reduce opportunistic behavior, thereby limiting earnings management practices.In comparison with minority shareholders, majority parties are interested in and play an active role in supervising the company to protect their large proportion of interests(Yeo et al., 2002;Gabrielsen et al., 2002;Zhong et al., 2007).Therefore, minority investors have an incentive to free-ride in overseeing management activities(Ali et al., 2008;Alves, 2012).However, as the degree of concentration ownership is high, it leads to problems of agency(Boubakri etal., 2005; Nguyen et al., 2021).In this context, blockholders may manage earnings to attain their own interests, causing damage to the minority investors' interests.Kim and Yoon (2008) document the positive impact of ownership concentration level on earnings managing practices.In contrast, according to Sharma and Kuang (2014) and Al-Fayoumi et al. (2010), ownership concentration has no bearing on the amount of earnings management.Foreign ownership is another critical determinant that expects to have an impact on earnings management.Foreign investors, when compared to lo-Investment Management and Financial Innovations, Volume 20, Issue 4, 2023 http://dx.doi.org/10.21511/imfi.20(4).2023.16cal investors, , 2007).Foreign ownership has been found to render an unfavorable influence on practices of earnings management in several nations, such as Vietnam (Nguyen et al., 2021), Japan (Guo et al., 2015), and Malaysia (Shayan-Nia et al., 2017).

Table 2 .
Descriptive statistics Source: Results from data analysis.

Table 3 .
Correlation matrixSource: Results from data analysis.

Table 4 .
Variance inflation factor and tolerance valuesSource: Results from data analysis.

Table 5 .
(Maswadeh, 2018)2021) the FGLSSource: Results from data analysis.showanadverseeffect of management ownership toward EM, and agrees with Habbash (2010) who does not find any association between the two variables.The finding about the insignificant influence of ownership concentration on EM activities disagrees with the results ofNguyen et al. (2021)andMaswadeh (2018), which show either a significant positive impact(Nguyen et al., 2021)or a significant negative impact(Maswadeh, 2018)of this variable on EM.For leverage ratio, the finding of this study is incompatible with earlier research performed byNguyen etal.(2021),Hoang et al. (2014), and Maswadeh (2018) who find a positive effect of leverage ratio variable on EM.Third, ROA is positively associated with EM.This outcome is inconsistent with the finding of Nguyen et al. (2021), which shows a negative impact of ROA on EM, and agrees with previous research that indicated that highly profitable firms are more ready to exercise earnings adjustment (Dechow & Dichev, 2002).Four, FSIZE has shown a negative impact on EM, the result agrees with the findings in the research conducted by Warfied et al. (1995), Nguyen et al. (2021), and Maswadeh (2018) who show that large firms demonstrate low level of incentives to manipulate earnings because these firms disclose not only mandatory information but also voluntary information and are constantly monitored by numerous stakeholders, such as state agencies, analysts, and investors.Fifth, other control variables show insignificant impacts on EM activities: audit quality (AUD), board size (BSIZE), and cash flow ratio (CF).