How does capital structure affect firm performance? Recent evidence from Europe countries

Transcription

1 How does capital structure affect firm performance? Recent evidence from Europe countries Name: Guangchen Shen Supervisor: Prof. Marco Da Rin ANR number: Graduate Time: 1/09/2012 Department: Tilburg School of Economics and Management July, 2012 Abstract: The relationship between capital structure and firm performance has long been a topic in modern capital structure literature. This present paper empirically exams the effect of capital structure on firm s performance based on 2007 data from 4 big economics in Europe: Germany, France, Italy, and UK. The paper finds a negative relationship between firm s leverage and firm s performance and finds the relationship between capital structure and firm performance may be not linear in case of Germany and France. Key words: Europe, Capital Structure, Performance. 1

2 How does capital structure affect firm performance? Recent evidence from Europe countries 1. Introduction 1.1 Background Now it is already 55 years since Modigliani and Miller made their first paper of irrelevance theory regarding capital structure and firm performance. Lots of literatures have been made on this topic and lots of departures of the irrelevance theory have been found. For example, agency cost and information asymmetric problem will affect the way a firm chooses its capital structure, thus affecting the firm s performance. Being advantageous of nourish data of listed and unlisted companies in Europe, the purpose of this study is to take a close look at the theories of capital structure and firm performance. The paper focus on four largest economics in Europe and take Germany, France, Italy, UK as sample, and then test them empirically on a cross-sectional base. 1.2 Previous literature of capital structure and firm performance The Modigliani and Miller (1958), in their known capital structure irrelevance theory, claims that in an efficient market which has no tax, no transaction cost, no information asymmetry, the value of a firm is unaffected by how that firm is financed. MM theory predicts that there is no relationship between a firm s capital structure and its performance. The MM theory makes the core stone of the modern corporate finance. After the original paper in 1958, Modigliani and Miller (1963) states that, considering the effect of corporate tax and tax deduction, the firm s value will increase when firm takes on more debt and this increasing amount will be the value of tax shield. This means that firms will benefit from taking more leverage. However, the Modigliani-Miller theorem will lose its explaining power when the market is not efficient. The inefficient market concept is closer to reality, which has taken taxes, information asymmetry, transaction costs, bankruptcy costs, agency conflicts and other imperfect elements into considerations. Since then, the following literature is awash by various extensions of the Modigliani-Miller theory. Usually one of the imperfect elements mentioned above is chosen, and the author will test how this introduction of imperfect elements will affect the MM theory which is made on an efficient market assumption. And then a lot of departures from irrelevance theory are found and the modern capital structure theory is developing at the mean time. When considering the corporate income tax, there is a tax shield benefit, so according to Modigliani and Miller (1963), the firms should use as many debts as 2

3 possible. But more debt than necessary in a firm s capital structure is found in reality, and it is obvious that excessive debt will introduce risk into firms. Then the concept of bankruptcy cost is introduced as an offset effect to the benefit of using debt as tax shield. Kraus and Litzenberger (1973) considered a balance between the benefit of tax shield and the risk added from bankruptcy cost, so there will be an optimal capital structure, any departure from the optimal capital structure cannot maximum the value of the firm. This is the trade-off theory. Myers (1984) identifies the pecking order theory. Because of information asymmetry, managers will first use internal funds, then debt, then equity as their source of finance when making financing decisions. Jensen and Meckling (1976) identify agency cost. The agency cost theory suggests that due to the effect of the separation of control and ownership, the agency of a firm will not always work on the behalf of the shareholders. When firm raise debts, there will also be conflicts between shareholders and bondholders. The conflicts between shareholders and managers, as well as the conflicts between shareholders and bondholders will all raise cost in the firms operation, investing and financing activities. Agency cost theory plays an important role, according to theory, in how we understand the effect that leverage exerts on firm performance. The separation of control and ownership in a firm might make the manager to max his personal utility rather than that of shareholders. And this conflict between shareholders and managers cause agency cost. Under agency cost theory, when firms take a higher leverage, due to the rising potential bankruptcy probabilities, the managers are also faced with the risk of losing their positions. The managers then will act on the best interest of shareholders and managers will have less free cash flow to use for their individual interest. Less empire-building activities, less under-investment problems, better investment decisions are also expected. So it should be expected to observe a positive relationship between firm performance and firm leverage. However, when the rise of debt decreases the amount of money that mangers can use for their personal purposes, it also decreases the money that company can use for investment and raises the cost of outside financing. If the proportion of debt in capital structure increases above a certain level, the adding cost of debt includes a higher bankruptcy cost, higher financial distress problem and more conflict between shareholders and debt holders, thus damaging the firm performance. What s more, according to the pecking order theory, firms tend to use internal funds rather than debt in financing activities because the information asymmetry problems. So more profitable firms should choose less debt, a negative relationship between leverage and firm performance is expected under the pecking order theory. The empirical test result of the relation between firm s leverage and firm performance is mixed. Abor (2005) compares the capital structures of publicly quoted firms, large unquoted firms, and small and medium enterprises in Ghana and made research on the influence of capital structure on profitability of listed companies on the Ghana Stock 3

4 Exchange. The result suggests a positive relationship between firm s leverage and performance. Capon et al. (1990) conduct a meta-analysis of results from 320 published studies related to financial performance, and find a positive relationship between usage of leverage and the financial performance. Ari (2011) use eastern Asian companies as a sample and find a positive between firm s performance and the leverage. Zeitun and Tian (2007) use a sample consist of 167 Jordanian companies and the research shows a significantly negative relationship. Rajan and Zingales (1995), use data from G7 countries and find a negative relationship between firm leverage and firm performance. Onaolapo (2010) use data from Nigeria and find a significantly negative relationship between firm s debt ratio and a firm s ROE or ROA. Majumdar and Chhibber (1997), Fama and French (2002), Booth (2001) also claim negative relationship between financial leverage and performance. 1.3 Theories of Reverse Causality from Performance to Capital Structure This paper is about how capital structure affects firm performance, but there is also possibility that the firm s performance will affect the way that managers choose the capital structure of the firm. If performance can affect capital structure as well, then there will be a reverse causality problem. Berger and di Patti (2002), using data from the US banking industry, use a simultaneous-equation model which shows how performance can affect capital structure. They give two hypotheses regarding how performance can affect capital structure: the efficiency-risk hypothesis and the franchise-value hypothesis. The efficiency-risk hypothesis claims that higher profitability often reduces the bankruptcy cost of a firm. Because when a firm is performing well, the firm will usually have a high expected return. A high expected return can be seen as a substitute of equity, because they can both be used for deduction of potential portfolio risk of the firm. So according to the positive relationship between performance and expected return, and the substitute relationship between expected return and equity, a firm with better performance will tend to use less equity in its capital structure. This hypothesis suggests a positive relationship between a firm s leverage and its performance. The franchise-value hypothesis, on the other hand, takes a new look from the aspect of economic rent. According to the franchise-value theory, a better performance might produce economic rents for firms in future, and firms are willing to take a lower leverage to protect this franchise value. So when firms have better performance, they tend to maintain a lot of equity in their capital structure. Contrast to the efficiency-risk theory, franchise-value hypothesis indicates s a negative relationship between a firm s leverage and its performance. According to these two hypotheses, firm performance can affect its capital structure in two ways, and the two effects are opposite to each other. Berger and di Patti (2002) do not actually solve the reverse causality problem, however, they give a new theory about how firm performance can affect firm capital structure. Some previous literatures develop some innovative approach to solve the problem by 4

5 taking an exogenous shock which only affect the firm s leverage or the company s competing environment. For example, Zingales (1998) study the truck industry using deregulation as an outside shock. However, due to the limitation character of data, it is not possible to find an instrument variable in regression to solve the reverse causality. The paper stills needs further development in case of endogeneity problem. The paper is developed as follow. The second section will discuss the data, the third section discuss expectations and method. The fourth section will build up the methodology and analysis the result. The fifth section concludes. 2. Data and preliminary observations The paper tests the relationship between capital structure and firm performance using specific 2007 accounting data of companies in Germany, France, Italy and UK. The data is collected from Amadeus Dataset (Analyses Major Databases from European Sources). Amadeus Dataset has financial statement information of over 7 million European firms and includes listed and unlisted companies. The original data of this paper is composed of 942,337 German companies, 794,486 French companies, 934,832 Italian companies and 2,385,106 British companies. All subsamples include all listed and part of unlisted companies in each country. 2.1 Measurement According to previous literature, there are numerous ways to measure the performance of firms. 1) Use the information from financial statement (e.g. calculate the value of target firm s ROE). 2) Use book value with market information, Tobin s q is a frequently used method. 3) Stock market return. 4) Modeled method like Z-score method. Due to the lack of access to market information of the data from Amadeus Dataset, this paper will use the first method, ROE as a proxy for firm s performance measurement. In preview literature, Zeitun and Tian (2007), Lemmon and Roberts and Zender (2008) use ROE or ROA as a proxy for firm s performance measure. The paper uses debt to equity ratio as the proxy for leverage, log (Asset) as proxy for firm size and the previous year sales growth rate as a proxy for firm s growth. Tangible asset is measured by using total tangible fixed asset divided by total asset and the firm s tax burden is measured by calculating the ratio of tax of net profit. The Amadeus uses a 4-digit identifies to distinguish different industries, for simply, this paper takes only the first 2 digit as identification for a firm. All the definitions and construction of variables in regressions are listed in table 1 below: Table 1 Definition of variables The table illustrates the constructions of variables for analysis. All variables are calculated by the 5

6 2007 data from Amadeus Dataset and are calculated on an annual base. Variable Definitions Construction ROE Firm's performance Gross profit / Total equity LEV Leverage (Current liabilities + Non-current liabilities)/ Total equity LEVsquare Square of Leverage Square of LEV ASSET Total Asset Fixed Asset + Current Asset SIZE Size of firm Log (Asset) AGE Age in operation 2007 Birthday + 1 GROWTH Sales growth in previous year Sales of 2007/ Sales TAN Tangible asset ratio Tangible Fixed Asset/ Total Asset TAX Tax burden Tax / Gross profit INDdummy Industry identification The first two digit of nace_rev11 as industry identification. 2.2 Summary The summary of all the variables are shown in Table 2 below: Table 2 Summary of variables The sample is all firms in the merged Amadeus Data in The table gives information about mean, median and standard deviance (SD) of variables, as well as the summary statistics in subsamples: France, Germany, Italy and UK. The definitions of variables are provided in Table 1. Country ROE LEV SIZE AGE GROWTH TAN TAX FRANCE GERMANY ITALY UK Total Mean Median SD Mean Median SD Mean Median SD Mean , Median SD Mean , Median SD The mean ROE of the full sample is 0.76, which means that out of every 100 equity, 76 net profits is earned. Germany and France have a ROE of 0.5 and 0.27, respectively. Italy has a low ROE ratio, while the UK sample has a quite high mean ROE of 3.7. So in terms of ROE, the UK firms seem to be more efficient than firms from other countries. The four Europe country firms have an average debt to equity ratio of 9.7 and the specific number fluctuates within a small range among each country. Italy has the highest leverage ratio of 12.6 while France takes a small leverage on average of

7 What s more, the average size of firms of each country does not differ from each other a lot. The Germany firm has an average age of 15 years, more than the other 3 countries that have an average age of around 10 years operating history. All countries have an average sale growth 3% in the year of 2007, but German firms have a quicker sale growth of 23%, higher than France s 1% and Italy s 6%, while UK firms have a negative sale growth of -6%. The average tangibility asset ratio is 20%, UK and Italy firms have ratios more than 20%, Germany has a ratio of 18%, and France has a ratio of 14%. Considering the tax burden variable, it is easy to find that Italian firms are suffering from a much heavier tax burden than firms in other countries. On average the tax amount is even more than the amount of net profit. At first sight, there is a positive relationship between leverage and firm s ROE, but this is just obvious among the German, France and UK sample. The Italian firms take a rather high leverage, but the ROE is the lowest among the four countries. As for other variables, size seems to exert a negative effect on firm s performance, Growth has a positive relationship with ROE and the other relationships are not easy to tell straight from a simple descriptive statistics. In case of potential correlation problem among dependent variables, a correlation matrix is made and correlation between variables is shown in table 3. Table 3 Correlation matrix The sample is all firms in the merged Amadeus Data in The table presents correlation between each variable and the p-value of the correlation. For example, the correlation between ROE and LEV is -0.02, the p value equals The definitions of variables are provided in Table 1. ROE 1.00 ROE LEV SIZE AGE GROWTH TAN TAX LEV (0.00) SIZE (0.01) (0.00) AGE (0.08) (0.00) (0.00) GROWTH (0.00) (0.26) (0.00) (0.02) TAN (0.10) (0.04) (0.00) (0.00) (0.01) TAX (0.95) (0.86) (0.00) (0.15) (0.68) (0.45) (Coefficients in first line, p-value in second line) In the correlation matrix, we can observe a negative relationship between ROE and leverage, and it is significant. What s more, size has a negative correlation with ROE as well, but the correlation is quite small. There is no obvious correlation between 7

8 ROE and AGE, GROWTH, TAN and TAX because all these correlations are not significant. There is also small but significant positive correlation between leverage and AGE, SIZE, which means that the age and size of a company might have effect on the firm s leverage decision making. 3. Expectations and methodology 3.1 Expectations and hypotheses LEV: The agency theory predicts that, when firm uses more debt, the manager will face more risk of bankruptcy and then be more efficient, agency cost decreases and a better performance of company is expected. So under the agency theory, there should be a positive relationship between leverage and the firm s performance. When a firm is operating well, the potential bankruptcy risk is low and the firm can be able to use a heavier leverage. However, under pecking order theory, firms with better profitability will tend to use less debt. As the franchise-value hypothesis in Berger and di Patti (2002) goes, the firms may use equity to protect the rents or franchise value, they will still maintain equity when they are performing well. And when a firm is over-leveraged, additional cost of debt will damage the performance of company. So the relationship between leverage and firm performance is mixed. LEV square: Square of leverage is used to test if there is a non-linear relationship between firm s leverage and firm performance. SIZE: Size is an important determinant of a firm s performance. Larger firms are usually more diversified, better-managed and have a larger risk tolerance. Small firms, on the other hand, may find it more difficult to solve the information asymmetry problem and thus may appear to perform worse than big companies. Penrose (1959) argues that bigger company is easier to achieve economic of scale and then results in a better performance. The paper expects a positive relationship between firm size and firm performance. The following hypotheses will be tested: Hypotheses: There is a positive relationship between size and firm performance. AGE: When firms grow older, they are usually more experienced. During the growth, firms invest in research and development, store their human capital resource, and gradually discover what they are good at. Hopenhayn (1992) shows that older firms are expected to enjoy better performance. The paper expects a positive relationship between firm age and firm performance. The following hypotheses will be tested: Hypotheses: There is a positive relationship between age and firm performance. GROWTH: It is obvious that growth opportunity is important to a firm s performance. The paper uses sales growth as proxy. Brush (1999) argues that sales growth influence the firm s ability to catch opportunities of investment, the use of new technology and provides opportunities for economic of scale, thus benefiting the health of the whole company. The paper expects a positive relationship between firm sales growth rate and firm performance. The following hypotheses will be tested: Hypotheses: There is a positive relationship between sales growth rate and firm performance. TAN: Tangibility is also a consideration related to firm performance, when firm has more tangible asset, it is faced with less bankruptcy risk and is considered to have 8

9 more flexibility when making financing decisions. Murillo (2007) has a result of positive relation between firm tangibility asset ratio and firm performance. The paper expects a positive relationship between firm tangibility asset ratio and firm performance. The following hypotheses will be tested: Hypotheses: There is a positive relationship between tangibility asset ratio and firm performance. TAX: Tax levels and tax structure will all influence the profitability and performance of company. When a firm is performing well, it receives a better profit, which means the firm will have more profit before tax, so it will tend to pay more tax. So this paper expects a positive relationship between firm tax burden and firm performance. The following hypotheses will be tested: Hypotheses: There is a positive relationship between firm tax burden and firm performance. 3.2 Methodology This paper will use two OLS regressions to study the relationship between capital structure and firm performance. The first regression is set without square of leverage; the second regression is set with square of leverage to exam if there is a non-linear relationship between the dependent variable and the independent variable. All variables other than leverage are control variables which control for the characters of firms that may affect firm performance. Robustness check is added to both regressions in case of heterogeneity problem. All data and regressions are run by STATA. The two regression formulas are shown below: ROE = α + β 1 Lev + β 2 SIZE + β 3 AGE + β 4 GROWTH + β 5 INDdummy + β 6 TAN + β 7 TAX + β 8 DIV + e ROE = α + β 1 Lev + β 2 Lev 2 + β 3 SIZE + β 4 AGE + β 5 GROWTH + β 6 INDdummy + β 7 TAN + β 8 TAX + β 9 DIV + e 4. Result and Discussions Table 4 and table 5 below show the regression result. Table 4 Regression without square of leverage The sample is all firms in the merged Amadeus Data in The table presents the results of OLS regressions for the full sample and four subsamples: Germany, France, Italy and UK. The dependent variable is ROE and independent variables are LEV, SIZE, AGE, GROWTH, TAN, TAX and INDdummy. The results are composed of coefficients of regression, the t statistic and the product of coefficient and SD as a check for economic significance, the coefficients of industry dummies do not show in the table. For example, in full sample, the coefficients of LEV is -0.01, the t statistic of this coefficient is -1.16, a one standard deviation change in LEV is associated with a change in ROE. Robustness check is added to the regression. The definitions of variables are provided in Table 1. 9

10 Hypothesis Full Sample Germany France Italy UK LEV *** * (-1.16) (-0.67) (-3.60) (-2.15) (-0.68) [-7,26] [-5,11] [-8,25] [-7,04] [-1,87] SIZE * ** *** (-2.14) (0.71) (3.21) (1.44) (-4.58) [-1,04] [0,32] [0,27] [0,13] [-2,51] AGE *** (0.54) (-1.63) (-4.23) (-0.69) (1.15) [0,06] [-0,41] [-0,14] [-0,05] [0,15] GROWTH *** (1.87) (0.68) (-1.64) (-0.02) (5.14) [0,06] [0] [0] [0] [0,09] TAN (1.27) (-0.88) (-0.17) (-1.21) (0.19) [0,17] [-0,27] [-0,01] [-0,05] [0,07] TAX (0.23) (-0.59) (1.17) (1.94) (0.44) [0,04] [-0,24] [0,04] [0,14] [0,13] _cons 3.49 * *** (2.48) (0.62) (-1.96) (-1.59) (4.71) N R adj. R (Coefficients in first line, the product of coefficient and SD in brackets, t statistics in parentheses, * p < 0.05, ** p < 0.01, *** p < 0.001) Table 4 shows the regression result without square of leverage. It is apparent that in all full-sample and sub-samples, leverage shows a negative relationship with ROE. But this negative relationship is only significant in the case of France and Italy, while in the full sample case, it is only significant at a 12.3% level. There are three different explanations to the negative sign. The first one is pecking order theory, firms tend to use internal funds rather than debt in financing activities, so more profitable firms should choose less debt. The second explanation is that, the firm is over leveraged by the manager, thus hurting the company s performance. The third explanation is the franchise-value suggested by Berger and di Patti (2002): franchise value is associated with high efficiency from the possibility of liquidation, which means that equity serves as a substitute of firm performance. The negative relationship between leverage and ROE is consistent with the result of previous empirical literature of Rajan and Zingales (1995), Rajan and Zingales (1995) use the financial data of G7 countries of 1990s. As their cross-sectional regression shows, the leverage of 10

11 Germany, France, Italy and UK all has a negative relationship with profitability, but only the UK sample is significant at 10%. Compared to the result of Rajan and Zingales (1995), this paper finds a more significant relationship between France and Italy while UK is becoming less significant. However this may due to the fact that Rajan and Zingales (1995) use the data in 1990s and these two papers are using different proxies for profitability and leverage. Zeitun and Tian (2007), Majumdar and Chhibber (1997) and Booth (2001) found similar relationship between capital structure and firm performance in their findings. For size, the sign is positive in case of Germany, France and Italy but not significant. A positive sign is consistent the previous expectations, the bigger firms are expected to achieve better performance. However, in case of UK, size has a negative and significant effect on ROE of UK companies, small companies sometimes suffer less from agency problem and may have a more flexible structure to fit the change (Big companies are too big to change), the similar negative relationships are found by Yang and Chen (2009). Age works as a negative and significant determinant for ROE of Germany and France. This is in contract with the original expectations. It seems that older companies are not performing better than their young competitors. When firms grow older, they may become more inert and inflexibility. Barron et al. (1994) gives two explanations, the first one is because older firms do not fit well to changing environment, and second one is that older companies become ossified by accumulated routines and old structures. GROWTH has a positive effect on firm performance, extremely significant in case of UK. This finding indicates that previous sales growth serves as a quite good predictor of firm performance and work very well in case of UK. And this is basically in consistent with our previous expectations. Tangible asset ratio has a general positive but insignificant effect on ROE. Our original expectation is that companies with more tangible asset have more flexibility in investing, a better access to financing. This result confirms our previous expectations. The relationship of TAX and ROE is positive, only significant in case of companies in Italy. This is also in consistent with our previous expectations. What s more, this paper adds the product of coefficient and SD of variable to illustrate the economic significance of independent variable. The paper times the standard deviation of the independent variable with the coefficient of independent variable, this result shows the unit change of dependent variable due to one standard deviation change of independent variable. Take the independent variable of Germany as example, one standard deviation change in Germany firm s age will cause a change in the ROE of Germany firms. The mean of German firms ROE is 0.50, so the change is quite a lot compared to the mean and is obvious of economic significance. The papers find that, except for the cases of growth, all independent variables are economic significant. 11

12 Table 5 Regression with square of leverage The sample is all firms in the merged Amadeus Data in The table presents the results of OLS regressions for the full sample and four subsamples: Germany, France, Italy and UK. The dependent variable is ROE and independent variables are LEV, SIZE, AGE, GROWTH, TAN TAX and INDdummy. The results are composed of coefficients of regression, the t statistic and the product of coefficient and SD as a check for economic significance, the coefficients of industry dummies do not show in the table. For example, in full sample, the coefficients of LEV is -0.01, the t statistic of this coefficient is -1.22, a one standard deviation change in LEV is associated with a change in ROE. Robustness check is added to the regression. The definitions of variables are provided in Table 1. Hypothesis Full Sample Germany France Italy UK LEV * (-1.22) (1.36) (0.22) (-2.05) (-0.64) [-6,26] [12,81] [0,14] [-7,04] [-1,3] LEVsquare *** (-0.67) (-1.39) (18.04) (-0.00) (-1.45) [-31,29] [-29,9] [6,79] [-0,03] [-32,16] SIZE * *** (-2.11) (0.59) (-0.72) (1.55) (-4.50) [-1,02] [0,22] [-0,02] [0,13] [-2,47] AGE * (0.51) (-1.62) (-2.01) (-0.70) (1.07) [0,06] [-0,31] [-0,02] [-0,05] [0,14] GROWTH *** (1.87) (-1.45) (1.16) (-0.03) (5.14) [0,06] [0] [0] [0] [0,09] TAN * (1.24) (-2.47) (-1.05) (-1.29) (0.16) [0,17] [-0,58] [-0,02] [-0,05] [0,06] TAX * (0.21) (0.58) (1.97) (1.95) (0.43) [0,03] [0,22] [0,01] [0,14] [0,12] _cons 3.43 * ** *** (2.45) (1.00) (2.66) (-1.90) (4.65) N R adj. R (Coefficients in first line, the product of coefficient and SD in brackets, t statistics in parentheses, * p < 0.05, ** p < 0.01, *** p < 0.001) 12

13 Table 5 shows regression with square of leverage. After adding the square of leverage, the R square of regression of Italy and UK remain basically the same, but the explaining power of regression of Germany and France increase dramatically, R square of regression of Germany increase from 4.9% to 15.8%, France increases from 65.8% to 90.88%, and the sign of coefficient of leverage and ROE change. Although the results are not significant at high levels, there is some evidence that in case of Germany and France, the relationship between leverage and firm performance is not simply linear. 5. Conclusion This present paper empirically exams the effect of capital structure on firm s performance based on 2007 data from 4 big economics in Europe: Germany, France, Italy, and UK. Being advantageous in nourish data of nearly 400 thousand companies in these four economics with a lot of observations of unlisted companies, the cross-sectional regression finds a negative relationship between firm s leverage and firm s performance. This finding is quite similar compared to the findings of Rajan and Zingales (1995), who use the data of G7 countries in 1990s. The similarity may indicate that the relationship between capital structure and firm profitability in Europe developed countries does not change during the past decades. The negative relationship does not support the positive expectation made under agency cost theory, however, the finding is in line with the predication of pecking order theory and the franchise-value hypothesis. In the second regression with a new variable square of leverage, the regression result of Germany and France changes a lot. The explanation power of the second regression has a dramatic rise compared to previous regressions without square of leverage, and the sign of coefficients of leverage change. This indicates that in case of Germany and France, the relationship between leverage and firm performance might not be linear. The paper confirms the positive effect of potential growth ability and tangibility ratio on a firm s performance, which is in consistent with the findings with previous literatures. The age of a firm is found to be a negative determinant of firm performance, which is not in consistent with many findings of previous literature. The possible explanation could be older firms have slow reactions to the fast changing business environment, or with the passage of time, firms tend to get old routines and management structures which are not fit the nowadays business environment. This paper tries to make contributions in empirical test on relationship between capital structure and firms performance based on recent data in Europe countries. But due to limitations of data, the paper can only use basic accounting information of a company, so there is no market information used in the regressions. Regression result will be more solid if variable related to market information is introduced, for example Tobin s q as a proxy for firm performance. 13

14 What s more, reverse causality is only discussed by not solved in this paper. It could be capital structure that affects the performance, but there is also possibility that the performance of a firm can affect the decisions made by managers about how to construct a firm s leverage. Further development can be made through introducing instrument variable to rule out the possibility of such endogeneity problem. 14

15 Reference [1] Abor, "Managing foreign exchange risk among Ghanaian firms," Journal of Risk Finance, Emerald Group Publishing, vol. 6(4), pages [2] Ari Warokka, Juan Jose Duran Herrera, Haim Hilman Abdullah. (2011), East Asian Corporate Governance: A test of the relationship between capital structure and firm performance International Journal of Economics and Finance Studies. [3] Barron, D. N., West, E. and Hannan, M. T. (1994). A time to growth and a time to die: Growth and mortality of credit unions in New York, American Journal of Sociology 100(2), [4] Berger, A. N. & di Patti, E. B. (2002), "Capital structure and firm performance: a new approach to testing agency theory and an application to the banking industry," Finance and Economics Discussion Series , Board of Governors of the Federal Reserve System (U.S.). [5] Boodhoo, R (2009): Capital structure and ownership structure: a review of literature. The Journal of On line Education, January Edition, pp 1-8. [6] Brush (1999): the free cash flow hypotheses for sales growth and firm performance. Strat. Mgmt. J., 21: [7] Carpentier, V. (2006) 'Funding in Higher Education and Economic Growth in France and the United Kingdom, ', Higher Education Management and Policy 18(3), [8] Capon, N.; Farly, J.U.; and Hoenig, S.M A meta-analysis of financial performance. Management Science 16, [9] Devinaga Rasiah, Peong Kwee Kim. (2011), A Theoretical Review on the use of the Static Trade off Theory, the Pecking Order Theory and the Agency Cost Theory of Capital Structure International Research Journal of Finance and Economics ISSN Issue 63 (2011) [10] Eugene F. Fama & Kenneth R. French, "The Equity Premium," Journal of Finance, American Finance Association, vol. 57(2), pages , 04. [11] Hopenhayn, H. A Entry, exit and firm dynamics in long run equilibrium. Econometrica 60: [12] Jensen, M (1986): Agency costs of free cash flow, corporate finance and takeovers. American Economic Review, 76, pp [13] Jensen, M.J., and W.R. Meckling (1976), Theory of the Firm: Managerial Behavior, Agency Cost, and Ownership Structure, Journal of Financial Economics 3, [14] Kraus, A. and R.H. Litzenberger, "A State Preference Model of Optimal Financial Leverage," Journal of Finance, September 1973, pp [15] Majumdar, S.K and Chhibber, P (1999): Capital structure and performance: evidence from a transition economy on an aspect of corporate governance. Public Choice, 98, pp [16] Margaritis, D and Psillaki, M (2006): Capital structure and firms efficiency in New Zealand. Journal of Finance. [17] Modigliani, F and Miller, M (1958): The cost of capital, corporation finance and 15

16 the theory of investment. American Economic Review, 48, pp [18] Modigliani, F.; Miller, M. (1963). "Corporate income taxes and the cost of capital: a correction". American Economic Review 53 (3): [19] Myers C.S., S.N. Majluf, (1984) Corporate Financing and Investment Decisions When Firms Have Information That Investors Do Not Have, Journal of Financial Economics, Vol. 13, No. 2, pp [20] Murillo Campello, (2007) Asset Tangibility and Firm Performance under External Financing: Evidence from Product Markets [21] Nigel Driffield, Vidya Mahambare, Sarmistha Pal. (2006), How Does Ownership Structure Affect Capital Structure and Firm Performance? Recent Evidence from East Asia. [22] Noel Capon; John U. Farley; Scott Hoenig. (1990), Determinants of Financial Performance: A Meta-Analysis. Management Science, Vol. 36, No. 10, Focussed Issue on the State of the Art in Theory and Method in Strategy Research. (Oct., 1990), pp [23] Onaolapo, Adekunle A, Kajola, Sunday O. (2010), Capital Structure and Firm Performance: Evidence from Nigeria European Journal of Economics, Finance and Administrative Sciences ISSN Issue 25 (2010). [24] Penrose, E.T (1959): The Theory of the Growth of the Firm. New York, Wiley. [25] Pinegar, M and Wilbricht, L (1989): What managers think of capital structure theory: a survey. Financial Management, Winter, pp [26] Rajan, R., and Zingales, L. (1995), Is there an Optimal Capital Structure? Some Evidence from International Data, Journal of Finance, 50, [27] Shyan- Rong Chou, Chen-Hsun Lee. (2007), The research of the effects of capital structure on firm performance and evidence from the non-financial industry of Taiwan 50 and Taiwan mid-cap from 1987 to 2007 [28] Yang, Chih-Hai, and Ku-Hsieh Chen Are small firms less efficient? Small Business Economics 32 (4): [29] Zeitun, R and Tian, G.G (2007): Capital structure and corporate performance: evidence from Jordan. Australasian Accounting, Business & Finance Journal, 1(4), pp