The testing of the modified jones model based on earnings management behavior of listed firms in Vietnam

The research result showed that there are two factors affected by Modified Jones Model based on earnings management behavior with significance level 0.05.

n trị - Kinh nghiệm quốc tế thực trạng ë ViÖt Nam THE TESTING OF THE MODIFIED JONES MODEL BASED ON EARNINGS MANAGEMENT BEHAVIOR OF LISTED FIRMS IN VIETNAM #Nguyen Anh Hien Saigon University Abstract: Nowadays, researchers have come up with different models to identify earnings management behavior by accrual earnings The objective of the study is to test the Modified Jones Model based on earnings management behavior of listed firms in Vietnam The official quantitative studied with a sample of 380 listed firms in Vietnam Sample had 200 firms of HCM City Stock Exchange (HOSE) and 180 firms of Hanoi Stock Exchange (HNX) The data collected from 2013 to 2014 and random sampling technique The Data used and analyzed for multilinear regression The research result showed that there are two factors affected by Modified Jones Model based on earnings management behavior with significance level 0.05 Keywords: Earnings management, behavior, Modified Jones Model and listed firms Introduction In recent years, many listed firms have changed their business results before and after the audit that has caused much concern for investors and management agencies At the request of the stock exchange, these firms have written explanations but generally they have not relieved the anxiety of investors There is so much the information on the financial statements, profit information that is most interested investors and information for managers are tend to impact on the most important Besides, many investors always want to know how their money used and the results of their business However, the separation between owners (investors) and managers in the joint stock model has led to conflicts of interest between owners and managers While the owner wants to know a lot of information related to the business results, the manager tends to limit the information provided or provide information of low reliability Investors as well as researchers are almost passive in the face of information provided by listed firms in Vietnam Above-mentioned thing, many researchers have looked at modeling to identify and measureearnings management behavior In the world, many studies have used Modified Jones Model based on earnings management behavior of listed firms, but is this model really suitable for use in the stock market in Vietnam? The results of this research helps not only stock market managers but also investors who apply the research results for improving information policy on the stock market development in the future Literature review Previous research was based on the accrual basis accounting to divide the profits of a firm into two categories such as cash earnings and accrued profits (accrual earnings) In particular, the net profit earned in cash is generated from the revenue and expenses in cash that the firm has collected and spent in the period[1] Particularly, firm had accrued profits, the profits shall be calculated in the firm's period but not yet collected in cash such as the 55 n trÞ - Kinh nghiƯm qc tÕ thực trạng Việt Nam turnover amounts sold to customers after subtracting non-payment expenses; expenses payable; prepaid expenses; provisions for asset depreciation, fixed asset depreciation in the period Managers are often unable to make corrective actions on monetary gains[2] However, firm had accrued benefits, managers can adjust using accrual basis accounting In theory, the accrual basis is to ensure that the firm accurately reflects the revenue and expenses in the period to ensure profit reflected honestly and reasonably In practice, this principle can be used to make profit-taking behavior such as recognizing revenue when sales or service provision is not available or the firm had been using the provisioning, cost of expenses, allocation of prepaid expenses to increase or decrease expenses during the period[3] Accruals include all adjustments that allow a business to change from a cash basis to an accrual basis - whether this means allocations, provisions or changes in accounting methods Changes in working capital also form part of accruals The following equation demonstrates the calculation: Total accruals (TA) = earnings - cash flow (from operations) However, it is not possible to use the TA variable to measure profitability directly because of the accruals appropriate to the actual business situation of the firm Besides, accounts are receivable bad debts reserve, stock price decrease reserves, payable expenses are made according to regulations Therefore, the researchers separated the TA into two components: - Non-discretionary accruals (NDAs) are accruals made in accordance with the rules and regulations of the accounting standards and system - Discretionary accruals (DA) or abnormal accruals are the accruals made by the manager for the purpose of adjusting the profit for the period[4] Therefore, the new DA is the appropriate gauge to assess the profitability of business managers In principle, when a project is positive, it is likely that the manager exaggerates profits when a project is negative, it is possible for the manager to hide the profit Thus, the DA shows that there is a subjective adjustment from the manager to the profit of the firm[5] The trick for researchers is to identify the discretionary component of accruals It is difficult to this because non-discretionary and discretionary components of accruals cannot be observed directly, so it is necessary to develop methods for estimating the discretionary accruals What researchers want to know can be shown as: Discretionary accruals (DA) + Non-discretionary accruals (NDA) = Earnings - Cash Flow (from operations), so we have formula of DA following: DA = TA - NDA Therefore, the researcher has to determine the DA and we must estimate the NDA Jones (1991) presented several models of NDA estimation as follows: Determine: Non-discretionary accruals (NDA) according to the following equation: NDAt/At-1 = α/At-1 + β1∆REVt/At-1 + β2PPEt/At-1 Note: 56 n trị - Kinh nghiệm quốc tế thực tr¹ng ë ViƯt Nam NDAt: Non-discretionary accruals variable can not be adjusted year t At-1: Total assets at the end of year t-1 ∆REVt: Net sales revenue change in year t PPEt: Cost of fixed assets of year t ∆ = change operator In the formula above, all variables of the equation are divided by At-1 (the end-ofyear assets t-1) to minimize the risk due to heterogeneous variances[6] α, β1, β2 are the parameters estimated by the least squares (OLS) of the coefficients a1, a2, a3 in the following model: TAt/At-1 = a1/At-1 + a2∆REVt/At-1 + a3PPEt /At-1 + εt The residual ε in the above model represents the undefined variable, including the autoregressive variable (DAt) After estimating the Non-discretionary accruals (NDA), from the equation: DAt = TAt - NDAt We have: DAt/At-1 = TAt/At-1 - NDAt/At-1 From this, the accounting variable is determined as follows: DAt/At-1 = TAt/At-1 - a1/At-1 - a2∆REVt/At-1 - a3PPEt/At-1 In the original model, Jones (1991) introduced the regression equation with only two variables, REV and PPE Jones argues that the value of ∆REV represents the net sales revenue of a firm in the accounting period which reflects the business situation and environment and is an objective item not taken advantage by the regulator[7] Firm is to adjust the profit in the period The value of PPE represents the entity's internal resources in generating revenue and the depreciation expense that is a large amount of unplanned expense that affects the total profit in the period However, in his study, the author also mentioned that the limitations of the model are that when selecting REV to make the study, it is possible that net sales revenue is also affected by the revenue recorded This may be a shortfall of the business Thus, improved the Jones model (1991) by adding a change in the receivables account (∆REC) to the model[7] The Jones model is improved as follows: NDAt/At-1 = α/At-1 + β1(∆REVt - ∆RECt)/At-1 + β2PPEt/At-1 Inside: ∆RECt is the change in the customer receivables account Methods of research The researcher has quantitative research: The sample of 380 firms was selected randomly from listed firms on both HOSE and HNX There are 200 firms listed on the HNX, 180 firms listed on the HOSE Selected firms not include businesses in the banking, securities or insurance sectors Because these firms have business characteristics, accounting regime is different from the normal business enterprises With a sample rate of 56%, the representativeness and reliability of the study were ensured Data of 380 firms are collected in 57 n trị - Kinh nghiệm quốc tế thực trạng ViÖt Nam 2013 and 2014 The researcher had used the SPSS software; this quantitative method was implemented in two steps: - The researcher estimates the coefficients of the model In this step, use the Ordinary Least Square (OLS) method to estimate the coefficients in each model [8] - The researcher had four tests for the multiple regression function were performed, including partial correlation of regression coefficients, determine the level of fit of the model, test the multi-collinear phenomenon, test the variance of the variable change (using Spearman test) The researcher had to analyze the magnitude of multicollinearity by considering the size of the variance inflation factor (VIF) [8] A rule of thumb is that if VIF >10 then multicollinearity is high The researcher has the following model: TAt/At-1 = a1/At-1 + a2(∆REVt - ∆RECt)/At-1 + a3PPEt /At-1 + εt The variables in the model are determined as follows: TAt = Earnings - Cash Flow (from operations) At-1: Total assets at the end of year t-1 ∆REVt: Net sales revenue change in year t ∆RECt is the change in the customer receivables account REVREC = ∆REVt - ∆RECt PPEt: Cost of fixed assets of year t Research results 4.1 The testing of the modelcoefficients Table 1: Coefficient of model (2013) Unstand Coefficients Model B Std Error (Constant) ,173 ,084 REVREC ,669 ,079 PPE -,133 ,129 Standar dized Coe t Sig Beta 95,0% Confidence Interval for B Lower Bound Upper Bound Collinearity Statistics Correlations Zeroorder Partial Part Tol VIF 2,051 ,041 ,007 ,339 ,400 8,474 ,000 ,513 ,824 ,402 ,400 ,400 ,999 1,001 -,048 -1,027 ,002 -,387 ,121 -,063 -,053 -,048 ,999 1,001 (Source: The researcher’ processing SPSS 22.0) Table showed that column “Sig (2-tailed)” < 0.05 with significance level 0.05 This showed the testing of the Modified Jones Model based on earnings management behavior of listed firms in Vietnam with significance level 0.05 based on 380 samples in 2013 58 n trị - Kinh nghiệm quốc tế thùc tr¹ng ë ViƯt Nam Table 2: Coefficient of model (2014) Model Unstand Coefficients B Std Error (Constant) ,091 ,013 REVREC ,022 ,060 PPE ,032 ,020 Standar dized Coe t Sig Beta 95,0% Confidence Interval for B Lower Bound Upper Bound Collinearity Statistics Correlations Zeroorder Partial Part Tol VIF 6,847 ,000 ,065 ,118 ,019 ,363 ,000 -,096 ,139 ,025 ,019 ,019 ,994 1,006 ,083 1,603 ,000 -,007 ,070 ,084 ,083 ,083 ,994 1,006 (Source: The researcher’ processing SPSS 22.0) Table showed that column “Sig (2-tailed)” < 0.05 with significance level 0.05 This showed the testing of the Modified Jones Model based on earnings management behavior of listed firms in Vietnam with significance level 0.05 based on 380 samples in 2014 4.2 The testing of the model hypotheses The researcher had been testing model results and the researcher’ collecting table 1, table and processing SPSS 22.0 bellowed: Table 3: The testing of the model hypotheses Testing type 2013 model 2014 model REVREC variable: Sig = 0,000; REVREC variable: Sig = 0,000, PPE variable: Sig = 0,002 PPE variable: Sig = 0,000 R2 = 34,9%, Sig = 0.003 R2 = 46,3%, Sig = 0.003 The variance inflation factor (VIF) VIF< 2, No Multicollinearity VIF < 2, No Multicollinearity Heteroscedasticity REVREC variable: Sig = 0,207 REVREC variable: Sig = 0,346 PPE variable: Sig = 0,316 PPE variable: Sig = 0,760 Coefficient of regression Testing F (Source: The researcher’ collecting table 1, table and processing SPSS 22.0) - Table showed that the results of the regression coefficient testing: With the data set for 2013, the REVREC variable and the PPE variable have a sig value: Sig = 0,000 and Sig = 0.002 respectively (2013) Besides, Table showed that the results of the regression coefficient testing: With the data set for 2014, the REVREC variable and the PPE variable have a sig value: Sig = 0.000 and Sig = 0.000 (2014) This shows that both variables are statistically significant in the model - Table showed that the results of the regression coefficient testing of F value: R square (R2) of the model: R2 of 2013 and 2014 respectively: R square (R2) = 34.9% and R square (R2) = 46.3% Besides, table showed that that the two independent variables explained the dependent variable at the rate of 34.9 % (2013) with significance level 0.05 based on 380 samples in 2013 and 46.3% (2014) with significance level 0.05 based on 380 samples in 2014 59 n trị - Kinh nghiệm quốc tế thực trạng ViÖt Nam - Table showed that there is not Multi-collinearity: The variance inflation factor (VIF) in both years is less than 10 Thus, the independent variables are not correlated - Table showed that there is no heteroscedasticity: With the results of Spearman test, the researcher saw that the mean of REVREC in 2013 and 2014 are Sig = 0.207 and Sig = 0.346 respectively The significance level of the PPE variable is 0.316 and 0.760 respectively, all of variables are greater than 0.05 The above-mentioned tests, it is possible to conclude that the Modified-Jones model is in line with the real situation in Vietnam, the REVREC and PPE variables have a significant effected This finding is consistent with many other studies in the world as it is used in the study of earnings management behavior.This is a science evidentforthis research that helps not only stock market managers but also investors who apply the research results for improving information policy on the stock market development in the future Conclusions The quality of information on financial statements of the listed firms in Vietnam that is a concern of managers and investors It is one of the most important reasons affecting the quality of financial statements of listed firms Investors need to identify earnings management behavior Researchers are unable to access the reality of businesses that often offer models to identify these behaviors The study tested the Modified Jones Model based on earnings management behavior of listed firms in Vietnam The official quantitative studied with a sample of 380 listed firms included: 200 firms of HOSE and 180 firms of HNX The data collected from 2013 to 2014 listed firms in Vietnam The Data used and analyzed for multilinear regression The research result show that the Modified-Jones model is in line with the real situation in Vietnam to identify trend in earnings management behavior of listed firms.‡ -References [1] [2] [3] [4] [5] [6] [7] [8] Peasnell, K Pope, P & Young, S., " Detecting Earnings Management using Cross-Sectional Abnormal Accruals Models,"Accounting and Business Research, vol 30, no 4, pp 313-326, 2000 Boynton, E., C., Dobbins, S., P and Plesko, A., G., "Earnings management and corporate alternative minimum tax,"Journal of accounting Research, vol 30, pp 131-153, 1992 Dechow, R., Sloan, G., and Sweeney, A., P., "Detecting earnings management,"The Accounting Review, vol 70, no 2, pp 193-225, 1995 Kothari, Leone, Wasley, "Performance matched discretionary accrual measure,"Journal of accounting and economics, vol 39, pp 163-197, 2005 Guay, W., S.P Kothari, & R.L Watts, "A Market-Based Evaluation of Discretionary Accruals Model,"Journal of Accounting Research, vol 34, pp 83-105, 1996 Healy, P & Wahlen, J.M., " A Review of the Earnings Management Literature and its Implications for Standard Setting,"Accounting Horizon, vol 13, no 4, pp 365-384, 1999 Jones, J., "Earnings Management During Import Relief Investigations,"Journal of Accounting Research, vol 29, pp 193228, 1991 Hoang Trong and Chu Nguyen Mong Ngoc, Analysis of research data with SPSS, HCMC: Hong Duc Publishing House, 2008 -  60 ... reality of businesses that often offer models to identify these behaviors The study tested the Modified Jones Model based on earnings management behavior of listed firms in Vietnam The official... information policy on the stock market development in the future Conclusions The quality of information on financial statements of the listed firms in Vietnam that is a concern of managers and investors... management behavior of listed firms in Vietnam with significance level 0.05 based on 380 samples in 2014 4.2 The testing of the model hypotheses The researcher had been testing model results and the