The empirical study of the correlations between different impact factors and profitability has ... companies' performance on the base of correlation b...

0 downloads 0 Views 174KB Size

FACTORS INFLUENCING THE COMPANIES’ PROFITABILITY Camelia Burja1

ABSTRACT: The information about company performance, especially about its profitability, is useful in substantiating managerial decisions regarding potential changes in the economic resources that the company will be able to control in the future. This objective aims achieving superior economic results that will increase the company’s competitiveness and will satisfy the shareholders’ interests. The paper presents some company performance analysis models, which highlight the influencing factors. The models are based on regression analysis, and the obtained results emphasize the strong connection between the profitability of the analyzed company expresses through Return on assets and the management of available resources. Key words: financial statements, regression analysis, performance, resource management JEL codes: D24, M48

Introduction At microeconomic level, performance is the direct result of managing various economic resources and of their efficient use within operational, investment and financing activities. To optimize economic results, a special attention should be given to the proper grounding of managerial decisions. These should be based on complex information regarding the evolution of all types of activities within the company. A synthetic picture of the company’s financial position and its performance is found in the annual financial statements, which therefore become the main information sources that allow the qualitative analysis of how resources are used during the process of creating value. In order one company to run on a long-term performance way, it is needed to develop, implementation and maintaining the strategies, measures and coherent policies from economic and financial point of view, resulted from a good knowing of internal and external specific conditions in which the firm acts. The qualities of managerial options depend by the ability of identifying those elements that productively used could lead to increasing of the results and performance. The research objective of this paper is to investigate how economic performance is achieved by companies in the industry. To reach this goal, we believed that the most appropriate indicators that express the aspects related to economic development and performance growth of companies should be chosen among the relative profitability indicators. The empirical study of the correlations between different impact factors and profitability has been conducted by using the information taken from the annual financial reports of a company in the Romanian chemical industry for the period 1999-2009 and by using appropriate statistical techniques. Starting from the economic content of rate of return and the information provided by various financial indicators computed on the basis of financial statements, the regression analysis helped identify an econometric model of economic performance assessment expressed as Return on total assets. This reflects a combination of elements that explain and influence the evolution of companies’ return, such as: the financial result, the advantageous use of the financing structure, the size of the technical and productive infrastructure, the efficiency of current assets, etc. 1

„1 Decembrie 1918” University, Faculty of Science, Alba Iulia, Romania, e-mail: [email protected] 215

Annales Universitatis Apulensis Series Oeconomica, 13(2), 2011

The statistical tests performed on variables and on the overall model validate its accuracy and the opportunity of using it in the analysis of microeconomic performance and in substantiating decision-making processes related to resources’ management. Literature review The considerable numbers of studies that approach the performance issue at microeconomic level prove the special importance of financial management aspects, on the improvement of which depend the obtained results and the companies’ competiveness. In the case of economic agents, various methods may be used to study performance. One way to study company performance is regression analysis, which allows the modelling the functional form of dependence between various economic and financial indicators. Modelling economic performance aims to increase efficiency by improving interventions in an adaptivelearning cycle (Campbell et al., 2001). The indicators involved in the regression analysis of economic performance are numerous. Models developed to study the impact caused by the allocation and use of capital within the firm tie performance to the contribution of various resources to the increase of efficiency, expressed in terms of profitability (Dumbravă, 2010). Recent literature analyzes the profitability of companies from various countries and economy sectors through indicators like net operating profitability (NOP) ( Raheman et al., 2010), (Dong and Su., 2010), return on total assets (ROTA) (Deloof, 2003), (Padachi, 2006), return on invested capital (ROIC), return on assets (ROA) (Narware, 2010). In these cases, the elements considered by profitability analysis as independent variables are financial indicators that express the working capital. Profitability at microeconomic level has been studied depending also on indicators such as current ratio, liquid ratio, receivables turnover ratio and working capital to total asset (Singh and Pandey, 2008). Other studies consider performance assessment expressed by earnings before interests and taxes (EBIT) and the associated risk resulted from the influence of using a certain financing structure (Akintoye, 2008) or expressing it though economic value added (EVA), return on equity (ROE), operating profit margin (OPM), earnings per share etc (Ryan, 2008). For Romania, a few econometric performance analysis models have been used for companies listed on the Bucharest Stock Exchange. These emphasize the correlation between intangible assets and company performance expressed by annual average market price, price/earnings ratio and earnings per share (Purcărea and Stancu, 2011). Other models analyze companies’ performance on the base of correlation between net profit and cash-flow (Matis et al., 2010). Methodology for analysis the microeconomic performance To identify the functional form that describes how to increase economic performance of companies, it was started from the content of annual financial statements, of the possibilities to reflect the correlations between patrimonial elements and the conclusions of previous research. The technique used by this study is regression analysis. It is considered one of the most valuable methods of establishing a conditioning between various phenomena due to its high level of generality and applicability (Albright et al., 2006). In order to select variables and appropriate performance analysis models we used the financial statements of a company in the Romanian chemical industry, representative for this sector, for the period between 1999 and 2009. The financial analysis indicators through their content express and can characterize the modality of patrimonial resources management, the conformity with the principles of a balanced functioning, options and strategies financing, the efficiency of resources used etc. From these

216

Annales Universitatis Apulensis Series Oeconomica, 13(2), 2011

financial indicators we have selected the most representative ones that exert a very strong impact on the firms’ performance. In order to specify the analysis model, we used as exogenous variable a series of indicators of financial analysis computed mainly as rates of the balance sheet, such as Fixed Assets Ratio, Sales to Current Assets Ration, Sales to equity Ratio, Debt Ratio, Gross Margin Return on Inventory, Expenses Revenue Ratio and structure of financing sources (Financial Leverage Ratio). For the evaluation of profitability, Return on total assets (ROA) was used as a dependent variable. It is considered that it includes all the influences of the assets’ management and it is acknowledged as a key indicator of increasing company performance; it also defines their economic growth potential (Helfert, 2002). 1. The influence factors of economic performance – variables The Return on Assets (ROA) indicator expresses the company’s ability to generate profit as a consequence of the productive use of resources and of the efficient management, and it’s used as a dependent variable in the assessment of economic performance. It is computed as a ratio between Net Income and Total Assets (Burja, 2010). In following it is presented the economic significance and the calculus way for the selected variables in order to study their impact on the industrial companies’ performance. Fixed Assets Ratio (FAR) expresses the share of the assets that the company disposes of permanently for its activities and indicates the level of capital investment in the technical and productive infrastructure. A high level of this indicator means an active investment policy, but its growth over a certain level (50%) may lead to an efficient use of the working capital and it limits the ability to expand current activities.

Fixed Assets Ratio

Fixed assets Total assets

(1)

Debt Ratio (DAR) shows the extent to which the total assets of the company are funded by loans. A growth in dynamic ensures an increase in the amount of the business’s financing sources, but also leads to less autonomy and financial solvency. For this reason, it’s necessary to rationally and efficiently use this financing method.

Debt Ratio

Total Debt Total assets

(2)

A good view of the modality of business financing is provided by the indicator Financial Leverage Ratio (FLR). It can be expressed as a ratio between debts and own capitals. Achieving a optimum rapport of financing structure can ensure company’s investors by the perspective of a future development and implicitly, of the increasing of equities (Ryan, 2008).

Financial Leverage Ratio

Total Debt Total shareholder equity

(3)

Sales to Current Assets Ratio (SCAR) is expressed as a ratio between Net sales and Total current assets and shows the incomings of the company from the management of current assets. A high level of this indicator signals the existence of a working capital deficit. In dynamic, usually a 217

Annales Universitatis Apulensis Series Oeconomica, 13(2), 2011

decrease of the ratio means a narrowing down of the company’s activity, which slows its production, thus diminishing inventories and accounting receivables, which are related to the current activity.

Sales to Current Assets Ratio

Net Sales Total current Assets

(4)

Sales to equity ratio (SER) shows how well were used the own capitals for generating sales. In dynamic, an increasing of this indicator, generally suggests a positive aspect that reveals a better management of own capitals used in activity and a raise of their efficiency.

Sales to Equity Ratio

Net Sales Total shareholde rs equity

(5)

Gross Margin Return on Inventory (GMROI) indicates if the modality of inventory management generates profit. It is an important indicator for appreciating the inventory efficiency and the company’s performance.

Gross M arg in Re turn on Inventory

Gross M arg in Average Inventory

(6)

The indicator Expenses Revenue Ratio (ERR) connects expenses with revenue, and expresses the efficiency achieved by a company through minimize its costs. In dynamic, a decrease of this ratio indicates an improvement in resources management and economic performance increasing.

Expenses Re venue Ratio

Operating Costs Operating Income

(7)

Net Income (NI) is an absolute expression of return, which synthesizes all financial flows related to the consumption of production factors and to achieve revenues. Through their significance, the selected indicators and the independent variables express various aspects of efficient management of resources and they were used in modelling the performance for the analyzed company. 2. The statistic characterization of influence factors Table 1 presents the statistical elements characteristic to variables used in modelling the performance of an industrial company during the period 1999-2009. Table no. 1 Descriptive statistics of variables Variables (ratio) ROA FAR DAR FLR

Mean 0,068 0,379 0,414 0,766

Maximum

Minimum

0,128 0,466 0,551 1,233

-0,002 0,208 0,178 0,263

218

Standard Deviation 0,042 0,082 0,120 0,300

Annales Universitatis Apulensis Series Oeconomica, 13(2), 2011

SCAR 2,404 3,322 1,700 0,468 SER 2,628 3,850 1,856 0,488 GMROI 0,192 0,443 -0,008 0,130 ERR 0,079 0,104 0,050 0,015 NI (lei) 24338 55615 -1037 17715 Source: data computed on the basis of the company’s annual financial statements

During the analyzed period, the company has a capital investment level of 37.9%, with an average variation of 8.2%. This situation indicates a high level of investments, which practically contributed to doubling the technical and productive infrastructure of the company. Increasing the debt ratio from a minimum of 17.8% to a maximum value of 55.1% generated for the overall period a debt level that represents 41.4% of total employed assets. Sales to Current Assets Ratio had an average of 2.404, the standard deviation being 0,468. In terms of inventory, the company functioned with a return average of 19.2%, the minimum value being -0.8%, which was due to the loss recorded in 2006. The maximum return on inventory was 44.3%. The efficiency of own capital used by firm was 2.628 with a standard deviation of 0.488, meaning that the capitals was on economic circuit an average number of 139 days with a maximum of 214 days and a minimum of 109 days. The capital structure utilized in activity financing is considered normal, its level of 0.766 suggesting that in average terms, own capitals surpassed debts, the financial autonomy level being high. The Financial Leverage had a minimum value of 0.263 and a maximum of 1.233. The average profit during the analyzed period was 24338 lei, with an annual average variation of 17715 lei. The maximum profit resulted from the activity was 55615 lei, but there was also a loss of 1037 lei during one year. The analysis of the correlation coefficients between exogenous variables ( r ) xi / x j highlights the existence of a connection of medium intensity that is normal, considering the economic content of the indicators. 3.3. Models of the performance regression analysis One possibility for a more precise characterization of how the dependence relationship between the company’s return and the selected factors is represented by the estimation of the influence of exogenous variables based on multifactorial regression analysis. There were identified some econometric models using specialized soft Eviews 7.1. The general presentation form of the dependence between the endogenous variable (Y) and exogenous variable that influence it is (Anderson et al., 2007): Y 0 1 X 1 2 X 2 .... k X k

(1)

where: X1, X2,..Xk – independent variables; - residual variable (random), which expresses how much Y may vary due to factors that are not included in the model; 0 - regression constant (intercept) 1 , 2 ,... k - regression coefficients for independent variables (model parameters); k – number of independent variables; The acceptance of the regression equation for analysis is done after subjecting the identified model to a series of statistical tests regarding quality of coefficients, the significance level of the 219

Annales Universitatis Apulensis Series Oeconomica, 13(2), 2011

variables for the analysed phenomenon, the presence of multi-collinearity, the residual variable’s quality, the accuracy of the model, etc. A statistical test necessary for acceptance of the regression models concerns the quality of the regression parameters. Theirs stability depend by the correlation grade of the exogenous variables. In the case of existence of a strong correlation between them, practical it is not possible to determine the one independent variable’s own effect on the dependent variable. A simple possibility to find out the multicollinearity is the Klein criterion according to which, the two variables i and j are collinear if R 2y r 2 , where R 2y is the determination coefficient of the dependent xi / x j variable y. A test that more precisely points out those variables that diminish the estimation quality in the case of existence of the multicollinearity phenomenon, is the variation inflation factors (VIF) test provided by Eview. This presents a diagnosis of the variation level for the estimated coefficients if between exogenous variables exists collinearity, so that the situation is favourable for the low levels of VIF. The testing of the regression parameters is done with the t-Student statistic. If all values t 1 , t 2 ,.... t k t , k , n k 1 , it results that the coefficients are significant and variables influence the y phenomenon. The estimation quality have to be also appreciated, making an analysis of the quality of the residuals for not to be autocorrelated. For this analysis can be used Durbin-Watson, BreuschGodfrey Serial correlation LM test şi heteroskedasticity tests. In the case of Breusch-Godfrey Serial correlation LM test, the test decision is made comparing the calculated value of the LM statistic with the theoretical value of 2p statistic (p parameters of the model). A great probability of the LM statistic indicates that residuals are not correlated. There is the same interpretation, if LM statistic LM calc 2 , p , where is the significance level of the test (Voineagu et al., 2007). Utilization of the heteroskedasticity test is made for identifying the first level of autocorrelation of residuals. The decision of inexistence of it, is taken in the case LM calc 2 , ,k where k is the number of independent variables of the model; also the probability of LM statistic have to be high. Also, this test allows the analysis of homoskedasticity of residuals, or the existence of a constant variation in residuals values, only in a such case the estimated parameters being correct. The validation of correctness of the model in a whole is done on base of the F test. The test decision is Fcalc F ; k ; n k 1 . A measure of the model quality is the value of the R2 determination coefficient. As its value increases concomitantly with the number of variables introduced in the model, it is considered that R2 adjusted prevents this inconvenient, and appreciation of the model’s quality on its base is more correct (Anderson, 2007). 4. Results of the multifactor regression analysis In summary, table 2 presents the influence factors and the characteristics of the performance’s manifestation form expressed by ROA.

220

Annales Universitatis Apulensis Series Oeconomica, 13(2), 2011

Table no. 2 The factors’ influence over economic performance Regression coefficient Coefficients Model quality diagnostic (VIF) 0,111 (0,0033) R2 ajust 0,948 Fstat 46,585 (0,000) -0,410 (0,000) 1,6 0,098 (0,011) 1,2 0,000002 (0,000) 1,2 0,012 (0,233) 2,0

Model

Independent variables

M1

C FAR DAR NI SCAR

M2

C FAR DAR NI

0,127 (0,000) -0,371 (0,000) 0,088 (0,015) 0,000002 (0,000)

1,0 1,1 1,1

R2 ajust 0,942 Fstat 55,532 (0.000)

M3

C FAR NI FLR

0,136 (0,000) -0,362 (0,000) 0,000002 (0,000) 0,034 (0,025)

1,0 1,1 1,1

R2 ajust 0,934 Fstat 48,633 (0.000)

0,084 (0,015) -0,156 (0,002) 0,217 (0,000) 0,015 (0,021) -0,469 (0,032)

1,2 1,3 1,1 1,2

R2 ajust 0,972 Fstat 88,974 (0,000)

M4

C FAR GMROI SER ERR Source: Eviews 7.1 results

Model 1. The performance analysis model describes the dependency between ROA and FAR, DAR, NI and SCAR variables for i annual observations: ROAi 0 1FARi 2 DARi 3 NI i SCARi i

i 1,..., n

(2)

Because in the case of variables included in the model the determination coefficient 2 (the highest correlation coefficient is 0.517), it results that there is no R y 0.948 r 2 xi / x j multi-collinearity for variables and the model is not affected by this phenomenon. The variation inflation factor indicates a high level of stability for all parameters; their value may vary only between 1.2 and 2. The LM value for checking serial correlation and heteroskedasticity shows that the estimations are not contaminated by the influence of residual factors, these are not autocorrelated. Therefore, LM calc 0.238 02.05,5 11.7 and LM calc 4.481 02.05,4 9.488 for the first and respectively, for the second test. The F statistic for which Fcalc 46.585 F0.05,4,6 4.534 with a statistical probability of 0.0001 recommends the model as valid. This has a good determination ratio (0.948), which confirms that the regression equation expresses a high level of dependency of ROA by the considered variables. The FAR variable has a negative coefficient within the regression equation, which shows that increasing the share of fixed assets as a result of investing a part of the company’s capitals leads to increasing the total assets, thus on account of this factor, the return diminished by –0.41 (p221

Annales Universitatis Apulensis Series Oeconomica, 13(2), 2011

value 0.000). The size of the profit obtained during the period had only slightly influenced ROA. DAR had a favourable impact that led to a return increased by 0.098 for p-value 0.011, which shows that the borrowed amounts contributed to an increased performance. Sales to current assets ratio has also had a positive influence on profitability, which increased by 0.012 (p-value 0.233). Model 2. Due to the analysis of the model 1 put in evidence that SCAR variable has a significance level only of 76.7%, this variable would be excluded from the next regression models. The second model used as explicative variables the indicators FAR, DAR and NI. ROAi 0 1FARi 2 DARi 3 NI i i

i 1,..., n

The model presents a higher stability of estimators, these can vary by a maximum value of 1.1 so that it can say that practically there is no collinearity between variables. The senses of influences exerted by the analyzed variables in the model 1, namely FAR, DAR and NI are the same in this regression but their impacts is more reduced. The action exerted by the Fixed Assets Ratio on the profitability was -0.371 with 0.000 pvalue, the debts level influenced profitability with 0.088 (p-value is 0.015). Also, the influence of profit over evolution of ROA was very low. For the analyzed company this situation corresponding with assertion that Return on assets depends little by the profit obtained, being more influenced by the amount of assets managed (Padachi, 2006). The residuals are not correlated, probabilities of statistic for LM Serial Correlation and Heteroskedasticity

test

are

0.896,

respectively

0.648.

The

Fisher

statistic

for

which

Fcalc 55,532 F0.05,3,7 4,347 (p-value 0.000) indicates the fact that profitability can be explained in proportion of 94.2% through the considered variables in the model. Model 3. This model reveals the influence of a specific financial structure on the performance indicators (ROA).

ROAi 0 1FARi 2 NI i 3 FLR i

i 1,..., n

The model is not affected by the variables collinearity, the coefficients are stable because the VIF values are low (maxim 1.1) and the residuals do not present the first level autocorrelation LM calc 0.001 02.05,4 9,49 and they are homoskedastic LM calc 1.411 02.05,3 7,815 . The good level of F statistic, for which Fcalc 48,633 F0.05,3,7 4,347 suggests the validity of estimation. All these features of the model offered it a high grade of credibility so that, it can be considered that ROA performance indicator was modified in proportion of 0.934 due to the influence factors FAR, NI and FLR. The variable Fixed Assets Ratio influenced profitability with 36.2% (p-value 0.000) and the action exercised by the financial structure practiced was positive, namely 0.034 (p-value 0.025). Model 4. Another analysis model identified the influence of weight of fixed assets (FAR), inventories efficiency (GMROI), capitals efficiency (SER) and operating costs (ERR) on the ROA indicator. ROAi 0 1FARi 2 GMROI i 3 SER 4 ERR i

i 1,..., n

The stability of estimators is high, they can vary with a maximum 1.3 and this aspect indicates the absence of the multicollinearity phenomenon between variables. The same conclusions resulted from Klein criterion. 222

Annales Universitatis Apulensis Series Oeconomica, 13(2), 2011

Analysis of the quality for residual factors reveals that these are not correlated with a probability of 0.801 and are homoskedastic with a probability of 0.917. Significance verification of the determination coefficient using F test, for which Fcalc 88.974 F0.05,4,6 4,534 indicates the validity of model and that ROA indicator was influenced in proportion of 97.2% by the explicative variables of model. The weight of productive fixed assets influenced the profitability with -0.156 (p-value 0.002), the efficiency of inventories used (GMROI) contributed with 21.7% (p-value 0.000) to ROA, the contribution of efficiency of capitals (SER) was 0.015 (p-value 0.021) and Expenses Revenues Ratio had an impact of -0.469 with 0.032 p-value. Conclusions The results of the study show a strong dependent relationship between company performance and how the available resources are managed. For performance indicator Return on assets were identified some influence factors that through their common action can contribute to increasing or lowering of the profitability of the analyzed company. From the numerous combinations that can be made with these factors, using the multifactor regression analysis, were selected some models with more significance in their economic content and statistical characteristics. Among the factors with a good action on profitability were found the efficiency of inventories, debts level, financial leverage, efficiency of capitals. The positive impacts of them show also, some of the action ways in order to improve the performance. The proper organization of operating activities should be aimed at the efficient use of current assets, which usually have the highest share in total assets. The efficiency of utilization of current assets increases when the rotation of the component elements (inventories and receivables) speeds up so that overall result will be a higher earning. Using combined sources to fund activities and increase debt to a certain level that doesn’t affect the financial autonomy of the company is another way designed to increase the assets’ ability to generate profit. In the analyzed situation, action of the financial leverage was favourable and it acted in the sense of increasing the ROA, this aspect justifying the company’s financing strategy through increasing debts. The profitability of one company can increase also through acceleration of own capitals rotation, elements that can participate in this manner to many economic circuits, contributing in greater measure to value creation and profit. A significant impact on the profitability increasing exerted the actions of lowering the all operating expenses. Due to the indicator Expense Revenues Ratio, Return on assets considerably, increased. In all the case of the considered models, the impact of Fixed Assets Ratio was negative and led to lower return, which shows that the investments in the technical and productive infrastructure of the company have not yet generated sufficient positive effects. These may happen in the year to come. It can affirm that a better management of a company profitability implies adoption of some adequate strategies which can be identified through analysis of how were manifested the phenomena in their concrete microeconomic environment. The elements on which it can intervene for improving the performance are those with a high impact, and factors that influenced negative the profitability constitute some reserves of economic increasing in the future activity.

223

Annales Universitatis Apulensis Series Oeconomica, 13(2), 2011

References 1. Albright S.C., Winston W.L., Zappe C., 2006. Data Analysis and Decision Making, 3th edition, Thompson South-Western: Masion, USA 2. Akintoye I.R., 2008. Sensitivity of Performance to Capital Structure European Journal of Social Sciences, vol. 7 (1) 3. Anderson D.R., Sweeney D.J., Williams T.A., 2007. Statistiques pour l’economie et la gestion, Editions De Boeck, Bruxelles 4. Burja V., The analysis of the financial statements of companies, 2010. Aeternitas Publishing, Alba Iulia 5. Campbell B.M., Sayer J.A., Frost P., Vermeulen S. Ruiz Perez M., Cunningham A.B., Prabhu R., 2001. Assessing the performance of natural resource system, Conservation ecology, vol.5 6. Deloof M., 2003. Does Working Capital Management Affect Profitability of Belgian Firms?, Journal of Business&Accounting, vol.30, issue 3-4 7. Dong H.P., Su J., 2010. The relationship between working capital management and profitability: a Vietnam case, International Research Journal of Finance and Economics, issue 49 8. Dumbravă M., 2010. Analysis of the firm’s performance: methods and models, Economica Publihing House, Bucharest 9. Helfert E. A., 2002. Techniques of Financial Analysis: a Guide to Value Creation, McGrawHill Education 10. Matis D., Vladu A.B. , Negrea L., Sucala L., 2010. Jones, Dechow and Kasznik Models Significance in the Romanian Economic Environment, Annales Universitatis Apulensis , Oeconomica, vol. 12(1) 11. Narware P.C., 2010. Working Capital Management: The Effect of Market Valuation and Profitability in Malaysia, International Journal of Business and Management, vol. 5, issue 11 12. Padachi K., 2006. Trends in Working Capital Management and its Impact on Firms’ Performance: An Analysis of Mauritian Small Manufacturing Firms, International Review of Business Research Papers, vol.2, issue 2, pp. 45 -58 13. Purcărea I.I., Stancu I., 2008. The influence of R&D Policy on Performance of the Companies Listed with Bucharest Stoch Exchange (through Intangible Assets), Theoretical and Applied Economics, AGER, vol 11 14. Raheman A., Afza T., Qayyum A., Bodla M.A., 2010. Working Capital Management and Corporate Performance of Manufacturing Sector in Pakistan, International Research Journal of Finance and Economics, issue 47 15. Rayan K, 2008. Financial leverage and firm value, Gordon Institute of Business Science, University of Protoria 16. Singh J. P., Pandey S., 2008. Impact of working Capital Management in the Profitability of Hindalco Industries Limited. Icfai University Journal of Financial Economics, vol. 6(4), pp.62-69 17. Voineagu V., Ţiţan E., Şerban R., Ghiţă S., Todose D., Boboc C., Pele D., 2007. Theory and econometric practice, Meteor Press Publishing, Bucharest

224