Relationship between Business Cycles and Financial Criteria of Performance Appraisal in Companies Listed in Tehran Stock Exchange Naser Yazdanifar Master of Accounting (Corresponding Author) Department of Accounting, Science and Research Branch, Islamic Azad University, Yazd, Iran Jamal Barzegari Khanagha Department of Accounting, Faculty of Economic, Management and Accounting, Yazd University, Yazd, Iran Mohammad Reza Aslami Department of Accounting, Science and Research Branch, Islamic Azad University, Yazd, Iran Abstract The present study sets to investigate the relationship between business cycles and some financial criteria of performance appraisal in the companies listed in Tehran Stock Market (TSE). In the present study, business cycle is meant to involve economic expansion and depression. Financial criteria of performance appraisal entail such variables as earnings per share, price to earnings ratio, return on sales, return on assets, and return on equity. In order to test research hypotheses, the data issued by the companies listed in TSE was studied for the 1998-2009 period. Spearman test, linear regression and developed linear models were used to analyze the data. The results showed a significant correlation of business cycles with earnings per share, return on sales, return on assets and return on equity (P<0.05). However, there was no significant relationship between business cycles and price to earnings ratio (P<0.05). Keywords: BUSINESS CYCLES, ECONOMIC EXPANSION AND CONTRACTION, EARNINGS PER SHARE, PRICE TO EARNINGS RATIO, RETURN ON SALES, RETURN ON ASSETS, RETURN ON EQUITY Introduction The information provided by accounting system is the basis of future predictions. However, accounting data are not enough to conduct an accurate prediction of future data. Rather, other factors such as national economic conditions (i.e. expansion or recession) need to be addressed as well. Economic conditions may exert various effects on companies and their performance and may influence the behavior of accounting data. Business cycles refer to regular fluctuations in macroeconomic activities that are typically organized by business enterprises. A cycle begins with a period of economic expansion in economic activities and ends up with a depression period (contraction). In other words, every business cycle includes improvement, boom, downturn and recession (Burns and Mitchell, 1946). In the present study, a business cycle refers to economic expansion and contraction. COPY RIGHT 2013 Institute of Interdisciplinary Business Research 291
Theoretical foundations Unless it increases return on investors assets, understanding a business cycle will not benefit investors. Fatas (2002) contends that business cycles influence productivity, investment and research and development costs. These variables are the major factors affecting economic development. In this regard, a business cycle is an important factor of economic growth. Fluctuations in business cycles may alter the development process and create permanent effects in economy. Investment links business cycles and long-term economic growth. He contends that increased fluctuations and uncertainty in business cycles may decrease risks, which in turn decreases return on invested capital and economic growth in the long run. Since managers and investors pay greater attention to risk of capital loss during recession periods, which intensifies the rapid downturn of stock prices in capital markets, corporate performance may be sensitive to business cycles. This may bring about more careful earnings management. The study helps decision makers improve their predictions of corporate performance consistent with macroeconomic conditions. Therefore, the present findings may assist decision makers, investors, financial analysts and corporate managers. The study aims to address the question is there any relationship between business cycles and financial criteria of performance appraisal in companies? Literature review Namazi and Kermani (2008) investigated the effect of ownership structure on the performance of companies listed in TSE. They reported a significant negative correlation between institutional ownership and company performance but a significant positive correlation between corporate ownership and company performance. Managerial ownership exerted a significant negative effect on company performance. Pur Heidari and Alipur (2009) studied the behavior of accounting data based on business cycles in TSE. They also investigated the behavior of accounting data based on business cycles and particular properties of companies. The results showed a significant correlation between some accounting variables (i.e. sales growth and gross margin) and business cycles. The results showed no correlation between some variables (i.e. change in total assets) and business cycle. The results also revealed that the size of companies and cyclical/non-cyclical nature of companies (specific company features) influenced the association between accounting variables and business cycles. Pur Heidari and Forouzesh (2010) conducted a study to investigate the relationship between business cycles and profit management based on specific company features such as company size, price to earnings ratio and their cyclicality in TSE. The results showed that company managers did not pay attention to business cycles and GDP in decision-making. There was a significant negative correlation between profit management and the ratio of total debt to companies market value. They also reported a linear correlation between GDP and profit management. COPY RIGHT 2013 Institute of Interdisciplinary Business Research 292
Materials and method The research hypotheses are formulated as follows: 1. There is a significant correlation between business cycles and price to earnings ratio (P/E). 2. There is a significant correlation between business cycles and earnings per share (EPS). 3. There is a significant correlation between business cycles and return on sales (ROS). 4. There is a significant correlation between business cycles and return on assets (ROA). 5. There is a significant correlation between business cycles and return on earnings (ROE). The research hypotheses were developed so that the relationships could be investigated among variables, which required the analysis of year-company data over 12 fiscal years (1998-2009). Following initial collection of the data on each variable, they were analyzed in terms of normality and outliner. Considering the research questions, Spearman test, linear regression analysis and developed linear models were used to analyze the data. The thematic scope of the research includes the examination of the correlation between business cycles and performance in the companies listed in TSE. The temporal scope of the study includes a 5- year period from among the years 1998 through 2009. The spatial scope of the study encompasses all companies listed in TSE consistent with their defined characteristics. Population, sampling technique and sample size The population of the study consisted of all companies listed in TSE. Systematic elimination method was used to select the research sample. Accordingly, companies that met the following criteria were selected as the sample while the rest were excluded from the research: 1. Companies should not be involved in investment, insurance, leasing and banking as these activities assume a different nature. 2. Financial statements and respective annotations on the companies should be available for 12 successive years (from 1998 to 2009). 3. They should have their fiscal year end in March 20 in order to facilitate comparison and avoid heterogeneity. 4. They should not have altered their fiscal year over the research period (from 1998 to 2009). 5. In order to have a homogeneous sample, they should have been listed in TSE before 1998 and have begun trading their shares from the beginning of 1998. 6. They should not have stopped their trading in TSE from 1998 to 2009. In other words, they should have kept their stock active over the respective years. In case of any halt in stock trading, the interruption should not have lasted more than three months. COPY RIGHT 2013 Institute of Interdisciplinary Business Research 293
Definition of variables and their measurement techniques The research addresses eight dependent variables used in separate patterns. The independent variables include economic expansion and contraction within business cycles (BC). 1. Dependent variables 1.1 Return on assets (ROA) ROA is obtained via dividing net profit by total company assets. 1.2 Earnings per share (EPS) EPS is obtained by dividing after-tax net profit by the number of company shares. 1.3 Return on equity (ROE) ROE is obtained via dividing net profit by book value of equity at the beginning of the period. 1.4 Return on sales (ROS) ROS is obtained by dividing operating profit by sales revenue. 1.5 Price to earnings ratio (P/E) Price to earnings ratio is obtained by dividing stock prices by earnings per share. 2. Independent variables A business cycle is a type of regular fluctuation in macroeconomic activities, which is typically organized by business enterprises. Every business cycle comprises improvement, boom (expansion), downturn and recession stages (Burns and Mitchell, 1946). In the present study, a business cycle is meant to indicate economic expansion and contraction, as the independent variables, which are considered as virtual variables. 3. Auxiliary variable Size: natural logarithm of total company assets at the end of the year MTB: the ratio of market value to book value of equity (MTB) at the beginning of the year Financial leverage (LEV): a financial indicator defined as the ratio of total debts to total assets. Descriptive statistics Descriptive statistics was used to describe raw data, as illustrated in Table 1. As shown in the table, EPS and ROS obtained the highest and lowest mean scores, respectively. The median statistic shows that ranking data in an ascending order helps determine the status of every score against the middle score. Mode indicates what score has the highest frequency. COPY RIGHT 2013 Institute of Interdisciplinary Business Research 294
Standard deviation (SD) denotes the deviation of the data from the mean. A small SD score indicates slight deviation from the mean score while a large SD score denotes the broad dispersion of the data around the mean. After business cycles that assumed a 0/1 status, the company size and EPS had the smallest and largest SDs, respectively. Table 1. Descriptive statistics of research variables Variable No. Mean Median Mode SD Skewness Kurtosis P/E 876 17.085 4.840-0.810 435.75-13.118 413.78 EPS 876 1095.28 550.85-2080 3112.03 9.720 141.56 ROS 876 0.413 0.160 0.002 2.182 14.40 236.38 ROA 876 1.109 0.0925 0.001 22.642 29.387 867.59 ROE 876 9.281 0.352 0.000 81.306 11.949 156.765 BC 876 0.583 1.000 1.00 0.493-0.339-1.890 LEV 876 2.435 0.704 0.752 27.186 28.009 810.9 Size 876 12.481 12.487 12.782 1.784 0.173 0.766 MTB 876 65.049 1.924-1.130 760.089 18.121 353.78 Skewness shows asymmetrical distribution around mean. When skewness coefficient is negative, the distribution is leaned to the left. When skewness coefficient is positive, the distribution is leaned to the right. However, when skewness coefficient is zero, the distribution is symmetrical. With larger absolute value of skewness coefficient, the population will be more deviated from normal distribution. The skewness of price to earnings ratio and business cycles were negative (skewed left) while they were positive in other variables (skewed right). In certain situations where it is not possible to make decisions based on measures of central tendency and skewness, Kurtosis may be used as one of the parameters to compare the distribution of sample data with normal distribution. In terms of Kurtosis, there are three types of distributions: 1. Distributions that have greater variability comparing with normal distribution; that is, the distribution curve is shorter than the normal distribution. These distributions have negative Kurtosis coefficient. 2. Distributions that are longer than normal distribution. The soaring of these distributions refers to the fact that the data are centered around the mean. In other words, the data are less dispersed. These distributions have positive Kurtosis coefficient. 3. Distributions that have the same Kurtosis as normal distribution. These distributions have average Kurtosis. The Kurtosis coefficient equals zero in these distributions. COPY RIGHT 2013 Institute of Interdisciplinary Business Research 295
Only had business cycle a negative Kurtosis coefficient while other variables had a positive Kurtosis coefficient. Testing research hypotheses Testing H1 There is a significant correlation between business cycles (BC) and price to earnings ratio (P/E). In the first hypothesis, the relationship between business cycles and P/E ratio was test. To this end, H 0 and its corresponding hypothesis were stated as follows: H 0 (b 1 =0): There is no significant correlation between BC and P/E. H 1 (b 1 0): There is a significant correlation between BC and P/E. Table 2 illustrates R, adjusted R, standard error of the estimate and Durbin- Watson coefficient. R indicates the variability of dependent variable, which may be explained by regression analysis. The difference between R and adjusted R may relate to the sample size and number of variables. When the sample is small, adjusted R produces better interpretation. Nevertheless, with larger sample sizes, the two converge. Increased number of variables in the regression model may increase the value of R. In order to prevent bias in R, adjusted R statistic is used to eliminate problems in R (Momeni & Faal Qayomi, 2010). Standard error of the estimate measures the dispersion of points around the regression line (in the two-dimensional space). The greater the value of standard error of the estimate, the more dispersed the points will be around the regression line. The measurement unit for this index is the same as the one for the dependent variable (i.e. efficiency). Table 2. Summary of H1 regression model Model R R Adjusted R Standard error of the estimate Durbin- Watson coefficient Regression 0.276 0.076 0.072 0.961 1.979 As shown in the table, R and adjusted R are 0.076 and 0.072, respectively, indicating that regression or independent and control variables account for 7.2 percent of variations in the dependent variable. Table 6 illustrates the results of ANOVA to examine how certain is a linear correlation between the dependent and independent variables. In order to examine the linearity of regression, statistical hypotheses of the overall significance of regression model were formulated as follows: H 0 : There is no linear correlation between the two variables. H 1: There is a linear correlation between the two variables. As shown in Table 3, since the overall significance of regression model is smaller than 0.05, there is a linear correlation between the dependent and independent variables. In this table, the Regression row denotes variations in the dependent variable, which is determined by independent variables. The Residual row also indicates variations in the dependent variable, which is determined by other variables (random factors). COPY RIGHT 2013 Institute of Interdisciplinary Business Research 296
Table 3. ANOVA-for-regression results of the first hypothesis Model Sum of df Mean F Significance s Regression 66.120 4 16.530 17.892 0.000 Residual 804.703 871 0.924 Total 870.824 875 Following the examination of overall significance of regression model, we need to examine partial regression, which help either support or reject the research hypothesis. Table 4 illustrates partial regression. The coefficient table includes two types of : non-standardized (B) and standardized (Beta). In non-standardized, the variables are on different scales. However, in standardized, all variables are converted to a similar scale, which facilitates the comparison of different variables. Therefore, standardized are used to compare the effects of several independent variables on the same dependent variable. Table 4. Partial regression of the first hypothesis Model Acronym Non-standardized Standardized T Significance B Standard error Beta Constant C -0.071 0.053-1.340 0.180 Business BC 0.122 0.072 0.060 1.692 0.091 cycle Financial LEV -0.163 0.033-0.163-4.936 0.000 leverage Company Size 0.035 0.035 0.035 0.985 0.325 size Market MTB 0.214 0.034 0.215 6.342 0.000 value to book value of equity As shown in the table, the significance level of LEV and MTB (0.000) are smaller than 0.05 while the significance level of BC (0.091) and Size (0.325) are greater than 0.05. Therefore, the first research hypothesis (H1) is rejected. That is, there is no significant correlation between BC and P/E. The regression model may be written as follows: EPS = - 0.71 + 0.060 BC 0.163 LEV + 0.035 Size + 0.215 MTB Testing H2 There is a significant correlation between business cycles (BC) and earnings per share (EPS). In the second hypothesis, we addressed the relationship between BC and EPS. Accordingly, H 0 and its corresponding hypothesis were formulated as follows: H 0 (b 1 =0): There is no significant correlation between BC and EPS. COPY RIGHT 2013 Institute of Interdisciplinary Business Research 297
H 1 (b 1 0): There is a significant correlation between BC and EPS. Table 5 illustrates R, adjusted R, standard error of the estimate and Durbin- Watson coefficient of the second hypothesis. Table 5. Summary of H2 regression model Model R R Adjusted R Standard error of the estimate Durbin- Watson coefficient Regression 0.491 0.241 0.237 0.8713 1.790 As shown in the table, R is 0.241, suggesting that regression or independent and control variables account for about 24 percent of variations in the dependent variable. We need to examine the results of ANOVA and partial regression in order to test the linear regression model and the second hypothesis. Table 6. ANOVA-for-regression results of the second hypothesis Model Sum of df Mean F Significance s Regression 209.495 4 52.374 68.987 0.000 Residual 661.248 871 0.759 Total 870.743 875 As shown in the table, the overall significance of regression model (0.000) is smaller than 0.05. Thus, there is a linear correlation between the dependent and independent variables. We need to examine partial regression to test the second research hypothesis. Table 7. Partial regression of the second hypothesis Model Acronym Non-standardized Standardized T Significance B Standard error Beta Constant C 0.370 0.048 7.679 0.000 Business BC -0.635 0.065-0.314-9.694 0.000 cycle Financial LEV -0.156 0.030-0.156-5.215 0.000 leverage Company Size 0.060 0.032 0.060 1.876 0.061 size Market MTB 0.278 0.031 0.279 9.085 0.000 value to book value of equity The significance level of BC, LEV and MTB (0.000) indicates the significant relationship between these variables and EPS. Considering the of these variables in the regression model, there is a significant negative relationship of operating cycle and LEV with EPS while there is a significant positive correlation between MTB and EPS. Therefore, the COPY RIGHT 2013 Institute of Interdisciplinary Business Research 298
results support the second hypothesis indicating a significant correlation between BC and EPS. The regression may then be written as follows: Testing H3 EPS=0.370 0.314 BC 0.156 LEV + 0.060 Size + 0.279 MTB There is a significant correlation between business cycles (BC) and return on sales (ROS). In the third hypothesis, we examined the relationship between BC and ROS. Accordingly, H 0 and its corresponding hypothesis were formulated as follows: H 0 (b 1 =0): There is no significant correlation between BC and ROS. H 1 (b 1 0): There is a significant correlation between BC and ROS. Table 8 illustrates R, adjusted R, standard error of the estimate and Durbin- Watson coefficient of the third hypothesis. Table 8. Summary of H3 regression model Model R R Adjusted R Standard error of the estimate Durbin- Watson coefficient Regression 0.390 0.152 0.148 0.920 1.723 As illustrated in the table, R is 0.152, which indicates that regression or independent and control variables account for about 15 percent of variations in the dependent variable. We need to examine the results of ANOVA and partial regression in order to test the linear regression model and the third hypothesis. Table 9. ANOVA-for-regression results of the third hypothesis Model Sum of df Mean F Significance s Regression 132.004 4 33.001 38.960 0.000 Residual 737.784 871 0.847 Total 869.787 875 Since the overall significance of regression model (0.000) is smaller than 0.05, there is a linear correlation between the dependent and independent variables. We need to examine partial regression to test the third research hypothesis. COPY RIGHT 2013 Institute of Interdisciplinary Business Research 299
Table 10. Partial regression of the third hypothesis Model Acronym Non-standardized Standardized T Significance B Standard error Beta Constant C 0.212 0.051 4.166 0.000 Business BC -0.364 0.069-0.180-5.259 0.000 cycle Financial LEV -0.192 0.032-0.192-6.068 0.000 leverage Company Size 0.055 0.034 0.055 1.633 0.103 size Market MTB 0.242 0.032 0.242 7.466 0.000 value to book value of equity The significance level of operating cycle, LEV and MTB (0.000) indicates a significant correlation between these variables and ROS. Considering the of these variables in the regression model, there is a significant negative relationship of operating cycle and LEV with ROS while there is a significant positive correlation between MTB and EPS. Therefore, the third hypothesis is supported, indicating a significant relationship between BC and ROS. The regression model may be written as follows: Testing H4 ROS= 0.212 0.180 BC 0.192 LEV + 0.055 Size + 0.242 MTB There is a significant correlation between business cycles (BC) and return on assets (ROA). In the fourth hypothesis, we examined the relationship between BC and ROA. Accordingly, H 0 and its corresponding hypothesis were formulated as follows: H 0 (b 1 =0): There is no significant correlation between BC and ROA. H 1 (b 1 0): There is a significant correlation between BC and ROA. Table 11 illustrates R, adjusted R, standard error of the estimate and Durbin- Watson coefficient of the fourth hypothesis. Table 11. Summary of H4 regression model Model R R Adjusted R Standard error of the estimate Durbin- Watson coefficient Regression 0.488 0.239 0.235 0.872 1.927 As shown in the table, R is 0.239, indicating that regression or independent and control variables account for about 24 percent of variations in the dependent variable. We need to examine the results of ANOVA and partial regression in order to test the linear regression model and the fourth hypothesis. COPY RIGHT 2013 Institute of Interdisciplinary Business Research 300
Table 12. ANOVA-for-regression results of the fourth hypothesis Model Sum of df Mean F Significance s Regression 207.750 4 51.937 68.212 0.000 Residual 663.192 871 0.761 Total 870.942 875 As shown in Table 12, the overall significance of regression model (0.000) is smaller than 0.05. Thus, there is a linear correlation between the dependent and independent variables. We need to examine partial regression to test the fourth research hypothesis. Table 13. Partial regression of the fourth hypothesis Model Acronym Non-standardized Standardized T Significance B Standard error Beta Constant C 0.203 0.048 4.201 0.000 Business BC -0.348 0.066-0.172-5.300 0.000 cycle Financial LEV -0.133 0.030-0.133-4.421 0.000 leverage Company Size -0.187 0.032-0.187-5.866 0.000 size Market MTB 0.312 0.031 0.312 10.170 0.000 value to book value of equity There is a significant relationship of all independent and control variables of the regression model with ROA. Considering the of these variables in the regression model, there is a significant negative relationship of operating cycle, Size and LEV with ROA while there is a significant positive correlation between MTB and ROA. Therefore, the fourth hypothesis is supported, indicating that there is a significant correlation between BC and ROA. The regression model may be written as follows: Testing H5 ROA = 0.203 0.172 BC 0.133 LEV 0.187 Size + 0.312 MTB There is a significant correlation between business cycles (BC) and return on earnings (ROE). In the fifth hypothesis, we examined the relationship between BC and ROE. Accordingly, H 0 and its corresponding hypothesis were written as follows: H 0 (b 1 =0): There is no significant correlation between BC and ROE. H 1 (b 1 0): There is a significant correlation between BC and ROE. COPY RIGHT 2013 Institute of Interdisciplinary Business Research 301
Table 14 illustrates R, adjusted R, standard error of the estimate and Durbin- Watson coefficient of the fifth hypothesis. Table 14. Summary of H5 regression model Model R R Adjusted R Standard error of the estimate Durbin- Watson coefficient Regression 0.538 0.290 0.286 0.843 1.978 As shown in Table 14, R is 0.290, which denotes that regression or independent and control variables account for about 29 percent of variations in the dependent variable. We need to examine the results of ANOVA and partial regression in order to test the linear regression model and the fifth hypothesis. Table 15. ANOVA-for-regression results of the fifth hypothesis Model Sum of df Mean F Significance s Regression 252.395 4 63.099 88.729 0.000 Residual 619.404 871 0.711 Total 871.798 875 As illustrated in Table 15, the overall significance of regression model (0.000) is smaller than 0.05. Therefore, there is a linear correlation between the dependent and independent variables. We need to examine partial regression to test the fifth research hypothesis. Table 16. Partial regression of the fifth hypothesis Model Acronym Non-standardized Standardized T Significance B Standard error Beta Constant C 0.243 0.047 5.195 0.000 Business BC -0.416 0.063-0.205-6.559 0.000 cycle Financial LEV -0.017 0.029-0.017-0.594 0.553 leverage Company Size -0.007 0.031-0.007-0.222 0.824 size Market MTB 0.443 0.030 0.443 14.945 0.000 value to book value of equity The significance level of regression displays a significant relationship of BC and MTB with ROE (0.000). Considering the of these variables in the regression model, there is a significant negative relationship of operating cycle with ROE while there is a significant positive correlation between MTB and ROE. Therefore, the fifth hypothesis is COPY RIGHT 2013 Institute of Interdisciplinary Business Research 302
supported, indicating that there is a significant correlation between BC and ROE. The regression model may be written as follows: ROE= 0.243 0.205 BC 0.017 LEV 0.007 Size + 0.443 MTB The present findings revealed a significant correlation between some accounting variables and BC. The results showed a significant correlation of EPS, ROA, ROE and ROS with BC in economic expansion and contraction periods. However, there was no significant relationship between BC and P/E ratio that is a special feature of companies. In other words, P/E does not contribute to the effect of BC on accounting variables. COPY RIGHT 2013 Institute of Interdisciplinary Business Research 303
References 1 Jalali Naeeni, A. R. (1997). Investigating business cycles in Iran s economy. Tehran: Higher Institute for Research in Planning and Development. 2. Mokhtar Zadeh, F. (2008). Investigating the determinants of business cycles in Iran s economy based on the accounting approach of business cycle. M.S thesis, Shahid Beheshti University, Tehran, Iran. 3. Ebrahimi Kordlor, A., & Hasani Azar Dariani, E. (2008). Investigating profit management at the time of initial public offering in the companies listed in TSE. Accounting & Audit Studies,. 4. Mokhtar Zadeh, F. (2008). Investigating the determinants of business cycles in Iran s economy based on the accounting approach of business cycle. M.S thesis, Shahid Beheshti University, Tehran, Iran, 25-44. 5. Namazi, M.,& Kermani, E. (2008). Effect of capital structure on performance in the companies listed in TSE. Accounting & Audit Studies, 53, 83-100. 6. Heidari, M. (2002). Management performance appraisal in the companies listed in TSE: a comparison of CVA model and stock returns. M.S thesis, Tehran University, Tehran, Iran. 7. Rostami, N.D. (2005). Investigating the relationship between financial ratios and stock returns in the companies listed in TSE. 8. Novo, R.P. (1994). Financial management (Vol. I). Jahan Khani, A.,& Parsaian, A. (Trans.). Tehran: SAMT Publications. 9. Karami, Gh., Moradi, M. T., Moradi, F.,& Mosala Nezhad, A. (2006). Investigating linear and non-linear correlations between financial ratios and stock returns in TSE. 10. Mashayekhi, B. (2004). Accumulative information content of accounting items pertaining to prediction of stock returns. M.S thesis. 11. Mehrani, K., & Mehrani, S. (2003). The relationship between profitability ratios and stock returns in TSE. Accounting & Audit Studies, 33. 12. Noorvash, I.,& Gholam Zadeh, M. (2003). Investigating the behavior of accounting profit using Box-Jenkins time series approach. Accounting & Audit Studies, 10 (31). 13. Rahnamay Rood Poshti, F. (2006). Financial value creation. Tehran: Termeh Publications. 14. Ghalibaf Asl, H.,& Rezaie, F. (2007). Effect of the structure of board of directors on the performance in the companies listed in TSE. Financial Research, 9 (23), 33-48. 15. Mashayekh, B. (2004). Investigating the relationship between economic value added and financial ratios. PhD dissertation, Tehran University, Tehran, Iran. COPY RIGHT 2013 Institute of Interdisciplinary Business Research 304
16. Noorvash. I., Karami, Gh.,& Vafi Sani, J. (2009). The relationship between corporate governance mechanisms and agency costs in the companies listed in TSE. Accounting Research, 1 (1), 4-25. 17. Perez-Quiros, G., and Timmermann, A.(2007).Variations in the mean and volatility of stock returns around turning points of the business cycle.forecasting Volatility in the. 18. Powell, J. G., and Shi, J. Smith, T. and Whaley, R. (2009).Political regimes, business cycles, seasonalities, and returns.journal of Banking& Finance, Vol. 33, pp.1112 1128. COPY RIGHT 2013 Institute of Interdisciplinary Business Research 305