Model for accounting information valuation, using multiple linear regression.
The major problem of intangible assets is based on the impossibility of being assessed. In time many valuation models --e.g. Direct Intellectual Capital Methods, Market Capitalization Methods, Return on Assets Methods, Scorecard Methods (Luthy, 1998)--, have been issued, but unfortunately, most of them can not generate comparability of data obtained after evaluation of different organizations. The main arguments regarding the need for a new model for assessing intangible assets may be the difficulty and high costs involved in measuring them, the uncertainty of results and the impossibility of recording social phenomena with scientific precision. Since now, there is no reference of the existence of a model using intangible assets not recognized in traditional financial statements in Romania.
2. CREATING A MODEL FOR PROFITABILITY MEASUREMENT USING MULTIPLE REGRESSION
In order to achieve an econometric model to determine companies profitability by using intangible assets not recorded in traditional financial statements we introduced in SSPS the responses of knowledge-based organizations managers and of the Top 100 listed organizations on BSE to the next question: Could you give grades from 1 to 10 for each item listed below, according to its importance for the success of the company where you work? (items: knowledge and skills of human capital, relations with customers, suppliers relations, company image, customers loyalty, alliances, partnerships, etc. organizational culture, professional skills of employees, work experience, employees loyalty, employees satisfaction, employees education, employees creativity, corporate reputation) and, also, the data from the 2009 financial statements (profit, total assets, current liabilities, outstanding payments, long-term debt, debt ratio, liquidity, solvency, capital, permanent capital, current assets, turnover, economic return, return on equity). The question is part of a questionnaire sent on 01.07.2010, by e-mail to 94 companies (26 knowledge-based organizations and 68 companies from Top 100). We received responses from 52 companies from Top 100 and 21 knowledge-based organizations listed on the BSE. We have chosen as a dependent variable the economic rate of return (ROA).
The standardization has been used because the analyzed variables had very different values; after that the variables are measured on the same scale (standard errors and their averages are between 0 and 1). Since values were standardized, free term is zero and does not appear in equation (Niculescu-Aron, 2007).
We kept in model only items that have significance level (Sig.) less than 0.05 (Ho, 2006), which shows that between independent variables (suppliers relations--Sig = 0.000; company image--Sig = 0.001; employees satisfaction--Sig = 0.000; corporate reputation--Sig = 0.033; liquidity--Sig = 0.000; debt ratio--Sig = 0.000; shareholder's equity Sig = 0.000) and the dependent variable (ROA) is a strong connection (Table 1).
First we tested if there were aberrant differences between values of the same variable of the analyzed population. From the figure no. 1 it can be seen that were seven cases (5, 7, 17, 35, 43, 44 and 53) with extreme values which could distort the results of multiple regression. After we had eliminated them, the values followed a normal distribution (Figure no. 2).
[FIGURE 1 OMITTED]
[FIGURE 2 OMITTED]
Next, we tested if were violated other assumptions, as normality, linearity and homeoscedascticity through evaluation of residuals mathematical diagrams (residuals scatter plots). The diagram (Figure no. 3) does not show abnormal vales (have not appeared the box Casewise Diagnostics, showing deviations from normal values).
[FIGURE 3 OMITTED]
We also tested the multicollinearity. One of the diagnostic tools can be found in the Tolerance and Variance Inflation Factor (VIF) (Meyers et al., 2006). The tolerance of all predictors is bigger than 0.01 and VIF values are lower than 10, so multicollinearity is not a problem (Table no. 2).
Table no. 3 provides a variety of measures assessing the success of the model in predicting the dependent variable. The adjusted [R.sub.2] is 0.639, so about 63.9% of the criterion variable's variance is explained by regression model.
Table no. 4 provides a summary of the analysis of variance regression. Because F (8.49) = 27.165, p < 0.00, we can say that there is a significant relationship between the independent variables and the dependent variable.
The coefficients (table no. 2) are significant predictors because sig < 0.05 (Pecican, 2007).
Multiple regression is defined by the equation (Meyers et al., 2006):
[Y.sub.pred] = [[beta].sub.1] [x.sub.z1] + [[beta].sub.2] [x.sub.z2] + .... [[beta].sub.n] [x.sub.zn]
[Y.sub.pred-] dependent variable
[[beta].sub.1], [[beta].sub.2] ... [[beta].sub.n]--regression coefficients
[x.sub.z1], [x.sub.z2] ... [x.sub.zn-] independent variables
In our case the regression line is: ROA = (-1.127) * suppliers relations + 1.063 * company image + 0.906 * employees satisfaction + (-0.678) * reputation of the organization + (-0.789) * liquidity + (-1.371) * debt ratio + 1.563 * shareholders' equity.
ROA can be explained by the indicators calculated and is influenced by elements from the traditional financial statements (liquidity, debt ratio, shareholders' equity), but also by intangible elements unrecorded in traditional financial statements (suppliers relations, company image, employee satisfaction, corporate reputation). Also, one can say that the model shows that 63.9% of observations are explained using the developed model. The unexplained observations can be justified by using a small number of observations (21 knowledge-based listed on BSE), due to the lack of companies that can be classified as knowledge-based organizations listed on BSE and also by managers' reluctance of providing information on questionnaires.
Ho, R. (2006). Handbook of Univariate and Multivariate Data Analysis and Interpretation with SPSS, Taylor & Francis Group, ISBN 978-1584886020, Florida
Luthy, H. (1998). Intellectual Capital and Its Measurement, http://www.3.bus.osaka-cu.ac.jp/apira98/archives/htmls/25.htm, Accessed on: 2009-06-10
Meyers, L. S.; Gamst, G. & Guarino, A. J. (2006). Applied Multivariate Research. Design and Interpretation, Sage Publication, ISBN 978-1412904124, London
Niculescu-Aron, G. (2007). Metode econometrice pentru afaceri, ASE Publishing, ISBN 978-973-594-996-9, Bucharest
Pecican, E. S. (2007) Econometrie pentru ... economisti, Publishing Economica, ISBN 978-973-709-340-0, Bucharest
***(2011)http://www.bvb.ro/ListedCompanies/SocietatiMain.a spx, Accessed on:2010-07-30
Tab. 1. Coefficients Unstandardized Standardized Coefficients Coefficients Std. Model B Error Beta t Sig. suppliers relations -4.361 1.020 -1.127 -4.277 .000 company image 4.177 1.202 1.063 3.474 .001 employees satisfaction 3.862 1.054 .906 3.663 .000 corporate reputation -2.585 1.197 -.678 -2.160 .033 liquidity -.136 17.000 -.789 -7.979 .000 debt ratio -.656 34.000 -1.371 -19.111 .000 shareholders' equity 2287E-8 .000 1.563 9.353 .000 Tab. 2. Coefficients Unstandardized Collinearity Coefficients Statistics B Std. Model B Frror t Sig. Tolerance VIF suppliers -.153 .018 -8.725 .000 .749 1.336 relations company -.679 .051 -13.326 .000 .722 1.386 image employees 1.826E-8 .000 10.415 .000 .928 1.078 satisfaction corporate -4.003 1.572 -2.546 .013 .718 1.393 reputation liquidity 4.029 1.978 2.037 .046 .563 1.776 debt ratio 4.119 1.841 2.237 .029 .251 3.991 shareholders' -4.996 2.182 -2.289 .025 .688 1.453 equity Tab. 3. Model Summary Adjusted R Std. Error of the Model R R Square (d) Square Estimate 1 .815 (a) .663 .639 6.39679 Tab. 4. ANOVA Sum of Mean Model Squares df Square F Sig. 1 Regression 10003.964 7 1111.552 27.165 0,000 Residual 5073.942 59 40.919 Total 15077.906 (b) 66
|Printer friendly Cite/link Email Feedback|
|Author:||Radneantu, Nicoleta; Stan, Elena Roxana; Gabroveanu, Emilia|
|Publication:||Annals of DAAAM & Proceedings|
|Date:||Jan 1, 2011|
|Previous Article:||Adaptive control of a three-linkmanipulator at end-effector movement on optimal trajectories.|
|Next Article:||Analysis of the territorial concentration of tourist accommodation made by tourist destinations from the point of view of sustainable development of...|