Firm-Specific versus Industry-Specific Factors
and Firm Performance:
A Comparison of European Firms
June 2009
Author: Research Supervisor:
J.E. Knoop Dr. G.J. Lanjouw
s1257471 Faculty of Economics
Nieuwstraat 89, 9724 KJ Landleven 5, 9747 AD
Pagina | 1
Abstract
Pagina | 2
1.
Introduction
Performance and economic downturn are key subjects in the newspapers today. In order to have an influence on performance, especially when performance decreases, it is helpful to know as a manager what kind of variables are the prime determinants of firm performance. Especially when the economy is cooling down firms need to know what variables to focus on in order to turn the process of decreasing performance around. Since the world has gone into a credit crunch the subject of performance drivers has become even more important than it has ever been. Complete industries are failing when it comes to performance, banks are collapsing, auto industry is suffering big time. Enough reason to re-investigate the drivers of performance and to find out whether the shift of focus towards firm-level resources as explanators of firm performance, two decades ago, is still justified by statistical evidence from recent years. This research will, therefore, investigate the effect of firm-specific versus industry-specific factors on firm performance within 206 European BIK-code industries.
Pagina | 3
2.
Literature Review
2.1. Explaining performance variation
There are two ways in which performance differences are and have been explained. They are either explained by industry effects or they are subscribed to firm effects. This field of research is characterized by a long debate between various scholars about whether industry-level factors or firm-industry-level factors are the prime determinants of performance. Over the last two decades there has been a major shift in consensus regarding the variables that explain performance variation. The focus during these two decades has shifted from industry-specific to firm-specific factors (Hoopes et al, 2003). In order to understand these two different views on the determinants of performance we will first set out the two primary views on performance drivers.
2.2. Structure-Conduct-Performance paradigm
Pagina | 4 model which represents the SCP paradigm. In this model Porter stresses the importance of the competitive forces of an industry in explaining performance variation.
Nguyen, Seror and Devinney (1990) argued in favour of a positive relationship between industry concentration and average industry profitability. Furthermore, they found a clear synergistic effect in their research on Canadian manufacturing firms. Scherer and Porter (both 1980) and others used various characteristics in describing industry structure. They used, for instance, characteristics like the number of buyers and sellers, variations in their size, product differentiation, regulations, barriers to entry and exit, and so on. In a research done by Kunkel (1991) major theoretical and empirical works within industrial organization economics were reviewed in order to find out which industry structural elements are the most important factors for influencing industry and firm performance. Kunkel concluded from this research that the most important industry structural elements are life cycle stage, industry concentration, entry barriers, and product differentiation. From these sets of industry structure variables this research will choose a representative set of variables according to which this research are going to test performance variations found in the dataset.
Gronhaug and Fredriksen (1988) found a reverse effect of what this research is trying to investigate, namely that higher market shares and elimination of competitors can be caused by improved firm performance. This research will try to investigate whether performance variations can be explained by variations in industry structure variables.
2.3. Resource-based view
Pagina | 6
3.
Aim of the Research
Pagina | 7
4.
Model Variables and Hypotheses
4.1. Introduction
Pagina | 8 industries. These industries are either service industries or manufacturing industries. Based on this research focus may be reallocated towards other variables determining firm performance or this research may support earlier findings providing more consensus in this field of research.
4.2. Performance
This research is going to explain potential variations in performance within a large number of industries. In order to do this, measures and information about performance of the chosen industries are needed. Therefore, this research, first of all, investigates what previous research used as measures for firm performance. In the study of Hawawini et. al. (2003) a statement was found that in the past research on firm and industry effects, firm performance measures have always been based upon return on assets. Besides Hawawini et al. (2003), a wide variety of other researches (Schmalensee, 1985; Wernerfelt and Montgomery, 1988; Rumelt, 1991; Roguebert et. al., 1996; McGahan and Porter, 1997; Chang and Singh, 2000) have been testing the effect of firm specific versus industry factors in relation to performance variation and have measured the effects of internal and external effects on firm performance, for instance profitability or return on assets (Galbreath and Galvin, 2008).
Pagina | 9 4.3. Industry Concentration
An industry typically consists of a given number of competitors with each its own market share. The variable industry concentration refers to this number of competitors within an industry. In this research, industry concentration will be measured according to the Herfindahl index. The Herfindahl index is a statistical measure of industry concentration developed independently by A.O. Hirschman (1945) and A.C. Herfindahl (1950). The Herfindahl index (also called the Hirschman-Herfindahl index) is calculated by summing the squared market shares of each individual competitor within the industry (Equation 1).
Equation 1: Formula used to measure the Herfindahl index.
Pagina | 10 The following hypothesis relates industry concentration and firm performance:
Hypothesis 1: There is a positive relationship between industry concentration and firm
performance.
Firm performance is expected to increase due to an increase in industry concentration, because when industry concentration increases this means that fewer organization are serving the same market. Competition among the individual firms within the market falls and individual market shares will rise. Therefore, the firms that are left in the market will perform better because of less competition.
4.4. Industry Type
Pagina | 11 Since there has been a decided shift away from a manufacturing-dominated economy towards a service-dominated economy, consequently the next hypothesis is:
Hypothesis 2: Service-oriented industries perform systematically better than
manufacturing-oriented industries.
Since there has been such a strong shift towards service-oriented industries one may predict that there is a positive relationship between service-oriented industries and firm performance and that this relationship is significantly stronger than the relationship between manufacturing-oriented industries and firm performance. This might be concluded from the witnessed shift that can be explained as being most likely caused by economic reasons (for instance that service firms are better performing than manufacturing firms).
4.5. Capital-to-Sales Ratio
The first firm-level variable that this research will focus on is the capital-to-sales ratio. This ratio is an indication of the firm’s ability to finance additional sales without incurring additional debt and is measured by dividing the firm’s fixed assets by the firm’s sales revenue (Equation 2).
Equation 2: Formula used to measure capital-to-sales ratio.
Pagina | 12 Besides that, the capital-to-sales ratio can also create barriers to entry (Oustapassidis, 1997). Kwoka and Ravenscraft (1986) found a significant positive effect of capital-to-sales ratio on profitability. Furthermore, the higher the capital intensity the greater the positive effect on profit margins of the firm (Hay and Morris, 1991). Therefore, the following hypothesis can be constructed:
Hypothesis 3: There is a positive relationship between the capital-to-sales ratio of a firm and
its performance.
When capital intensity is high, this means that a firm needs to invest more capital to increase their sales. Therefore, also the entry into the market requires a lot of capital and, hence, entry costs are high. Besides that, high capital intensity is also an advantage for the individual firm since it allows this firm to increase their sales more easily than a firm with a lower capital intensity (this firm needs to invest more). Conclusion, high capital-to-sales ratios leads to less direct competition and, therefore, better performance.
4.6. Leverage
The variable leverage shows the efficiency of the use of borrowed capital. In this research leverage will be measured by using the debt ratio which indicates what proportion of a companies’ assets is financed by debt. The debt ratio gives an idea about the leverage of the company and the potential risks that the company is facing in terms of its debt-load. The debt ratio is measured by dividing the total amount of liabilities by the total amount of assets (Equation 3).
Pagina | 13 If leverage means greater risk and greater risk in turn means a greater profitability than the coefficient of the debt ratio is positive, that would mean that more leveraged firms (firms that have more liabilities) are more profitable, all else equal (Martin, 1993). Empirical studies, however, found a negative effect of leverage on profitability which means that the borrowed capital has either not effectively been used or the costs of borrowing were simply higher than the benefits of the relevant investment. Another possibility is that the firms who finance their investments from their retained profits are more profitable than the firms that borrow capital for their investments. For instance, Shepherd (1994) concluded that leverage has no significant separate role in explaining and/or influencing profitability. Nevertheless, this research hypothesizes:
Hypothesis 4: A positive relationship exists between leverage and firm performance.
Pagina | 14
5.
Sample and Methods
5.1. Introduction
Galbreath and Galvin (2008) already mentioned in their article that the vast majority of the researches comes from the United States and have also investigated datasets that originated from the United States. Very little is known about results on firm specific versus industry factors and their relationship towards firm performance outside of the United States. Galbreath and Galvin (2008) have, therefore, focused on a large number of Australian firms and have recommended further research to focus on either undeveloped or developing countries. Since it is hard to get good information inside developing countries this research will focus on data from European industries while investigating the drivers of performance. Also this research tries to compare findings from Galbreath and Galvin (2008), which based its findings on series of questionnaires, with data based on economic figures.
5.2. Sample Collection
This research will use a dataset containing 206 three-digit BIK-code industries from which this research will randomly select 5 firms per industry and use data for each of these firms from the year 2006. These 5 firms per industry will be selected using the random selection tool available in Microsoft Excel. This tool randomly selects 5 firms out of the entire industry. Randomly selecting 5 firms per industry gives a better representation of the industry than simply picking the 5 largest companies from each industry which may cause biased results.
Pagina | 15 subscription and within this database can be searched on industry classification to get a dataset from each BIK-code industry chosen. From these results we will take the largest sample as possible including all active organizations with reported data for the entire time-series. In order to get performance results for all our organizations within our sample, this research will delete organizations from the dataset that are missing information in order to get most reliable results from the dataset. Missing information for a few of the organizations is offset by keeping the sample as large as possible by including all organizations that do fulfil the full information requirement. Missing information is possible, for instance, because the data-source does not provide enough information on the smaller organizations. Including these smaller organizations with missing information may jeopardize the aim of this research to get a representative test on the performances of the BIK-code industries. Besides that, the industries will still provide this research with enough reliable data to conduct a representative test on the relationship between firm-specific versus industry factors and firm performance.
This research will limit itself to the year 2006 because there is not enough information available for all organizations in the electronic database ‘Amadeus’ about return on assets from before 2006 and there is not yet enough information for all organizations about the year 2007 and onwards. Data from the year 2006 should provide enough performance information to conduct a reliable test on the relationship between firm-level versus industry structure characteristics and firm performance. Therefore, our data-sample will be limited to only 1 year. This year should provide this research with enough information to investigate the stated relationships for the 206 three-digit BIK-code industries chosen. The data which this research is going to need, can be divided into three types namely performance data, industry structure data and firm-level data.
• Performance Data
Pagina | 16 As in previous research, performance will be determined by return on assets (Hawawini, Subramanian and Verdin, 2003). This research will use a data-sample that contains reliable performance data from the year 2006 and it will only use an accounting based measure of performance, because not all organizations within the European BIK-code industries are represented in financial datasets. Therefore, it would be very difficult to find any financial (value-based) measures of performance for every organization in every BIK-code industry within our data-sample. Nevertheless, most researches done in the past only used the accounting based measure of performance, because in the past there was even less information about value-based measures of performance (Hawawini, Subramanian and Verdin, 2003). This is why this research will limit itself to explaining variations in the accounting based measure of return on assets. In a comparable research done by Uri (1988), the performance measure has not been corrected for inflation. In this research data was obtained from one unique source with similar industry codes as the BIK-codes that this research is using. Since all variables are recorded in the same currency and because this research limits itself to observations from one year inflation effects will not be present.
• Industry Structure Data
Pagina | 17 variations and try to find any relationship with the chosen industry structure characteristics and compare these results with findings on firm-specific variables. Again data will be used for the year 2006.
• Firm-level Data
Lastly, this research needs information on two measures of firm-specific data: capital-to-sales ratio and leverage. Each variable has been used before as measures of firm-specific
characteristics and, therefore, they function as good measures for firm-level data in this research. Both variables can be calculated using data that can be abstracted from the
electronic database called ‘Amadeus’. Again the same dataset can be used since all other data in this research is abstracted from the same source. Since within this research data was obtained from one unique source with similar industry codes as the BIK-codes that this research is using, all variables are recorded in the same currency and because this research limits itself to observations from one year inflation effects will not be present. Whenever there is missing information for any organization in one of the BIK-code industries this
organization will be deleted from the dataset to prevent the occurrence of any bias in the dataset. Data concerning the chosen firm-specific characteristics will be gathered for the year 2006.
5.3. Analysis
Pagina | 18 firm performance. In her research she used both industry- and firm-specific variables in explaining firm performance. All regressions done in this research are done with use of the fixed effects model.
Since the industries used in this research are clearly defined by means of three-digit BIK-code and selection based on their main activity, there will be no difficulties with common firm membership. Since this research analyses variations in performance between various organisations based on figures from one year, it is not possible to use the fixed effects model. Instead, this research will use cross-section analysis for which a regression will be made using data from one year in order to estimate the effect of firm-level versus industry structure variables on firm performance. To check whether one of the chosen variables is endogenous a Granger Causality (Granger, 1969) test or a Hausman (Hausman 1978)test will be performed. If one of the variables appears to be endogenous this research will try to tackle the problem of endogeneity.
Using the cross-section analysis this research needs to do a series of diagnostic and robustness checks in order to justify the regression specification. First of all, all the underlying assumptions for empirical research need to be checked. These underlying assumptions include: normality, homoskedasticity, linearity and the absence of correlated errors. Missing information is deleted from the dataset and outliers have also been removed. By removing all missing information and outliers from the 1030 firms 91 have been removed. This results in a dataset containing 939 firms with data with normally distributed observations. In order to improve the data to satisfy the assumption of normality, some of the variables have been transformed. Therefore, the proposed model specification will be:
Perf
i = α + β1LOG(Conc)i + β2(Ind) i + β3LOG(Cap) i + β4 ARCSIN(Lev) i + ê i Equation 4: Economic model used in this research.
Pagina | 19 (Herfindahl, 1950) which is the sum of the squared market shares within each industry (LOG(Conc)
i). The interpretation of the variable industry concentration has changed due to
the LOG-transformation; a 1 percent change in industry concentration is associated with a change in performance of 0.01* β
1. (Ind)i is the type of industry specified for each firm i
measured as being either service-dominated or manufacturing-dominated, and the third control variable is the logarithm of the capital-to-sales ratio for each firm i (LOG(Cap)
i). Also
for the variable capital-to-sales ratio the interpretation has changed due to the LOG-transformation; a 1 percent change in capital-to-sales ratio is, similar to industry concentration, associated with a change in performance of 0.01* β
3. The fourth control
variable is the leverage of each firm measured as the ARCSIN of the debt-ratio (ARCSIN(Lev)
i). The ARCSIN-transformation is far more difficult to interpret than the
Pagina | 20
6.
Results
As described above, from the original dataset all organizations with missing data and all outliers have been removed. And, secondly, also the data have been normalized by transforming some of the variables resulting in the variables described in Table 1. Table 1 shows the descriptive statistics for all the variables concluded in this research.
Table 1. Descriptive Statistics
N = 939
Minimum Maximum Mean Std Deviation
ROA 2006 -19,77 39,20 7,0795 9,47564
LOG_IND_CONC -6,61 -,95 -3,6879 1,16146
IND TYPE ,00 1,00 ,5996 ,49025
LOG_CAP_SAL -6,82 3,93 -1,4749 1,62570
ARCSINLEV 12,65 90,00 51,4336 16,19929
6.1. Diagnostic and Robustness Checks
Pagina | 21 Table 2. Pairwise Granger Causality Test
N=939 Lags: 8
Null Hypothesis: ROA_2006 LOG_IND_ CONC
IND TYPE LOG_CAP_ SAL ARCSINLEV ROA_2006 does not Granger Cause F-stat Prob X X 1.59970 0.1207 1.80837 0.0719 1.03796 0.4054 0.58841 0.7880 LOG_IND_CONC does not Granger Cause F-stat Prob 0.79074 0.6109 X X 0.90000 0.5157 1.00814 0.4280 1.77898 0.0775 IND TYPE does not Granger Cause F-stat Prob 1.33526 0.2220 0.43592 0.8998 X X 1.29421 0.2427 0.72567 0.6690 LOG_CAP_SAL does not Granger Cause F-stat Prob 3.13878 0.0016 0.71257 0.6806 1.42852 0.1803 X X 0.68775 0.7027 ARCSINLEV does not Granger Cause F-stat Prob 0.52875 0.8354 2.34029 0.0172 2.13735 0.0301 1.22873 0.2787 X X
Table 3. Durbin-Watson Test
N=939 Dependent Variable: ROA_2006 Method: Least Squares
Coefficient Std. Error t-Statistic Prob.
Pagina | 22 Consequently, this research also needs to check for heteroskedasticity. This can be done by using the White Heteroskedasticity test. Results from the White heteroskedasticity test can be found in Table 4. The observed R-squared value exceeds the critical value of the chi-square distribution for the 5-percent confidence interval with 5 degrees of freedom. Therefore, the dataset is characterized as being heteroskedastic. This means that the standard errors need to be corrected for heteroskedasticity (White/Newey-West) in order to get reliable results from the regression model.
Table 4. White Heteroskedasticity Test
N=939
Dependent Variable: RESID^2 Method: Least Squares
F-statistic 2.664095 Prob. F(13,890) 0.0011
Observation*R-squared 33.88858 Prob. Chi-Square(13) 0.0013 Scaled explained SS 61.59116 Prob. Chi-Square(13) 0.0000
Pagina | 23 is that there is no problem with multicollinearity since TOL > .10 and VIF < 10 for all independent variables in the model. Table 6 shows the relevant TOL and VIF figures on which the conclusion about multicollinearity can be based.
Table 5. Correlations
N=939
ROA_2006 LOG_IND_ CONC
IND TYPE LOG_CAP_ SAL ARCSINLEV ROA_2006 Pearson Correlation Sig. (2-tailed) 1 . ,063 ,053 ,008 ,796 -,221(**) ,000 -,235(**) ,000 LOG_IND_CONC Pearson Correlation Sig. (2-tailed) ,063 ,053 1 . ,178(**) ,000 ,046 ,160 -,116(**) ,000
IND TYPE Pearson Correlation Sig. (2-tailed) ,008 ,796 ,178(**) ,000 1 . -,003 ,924 -,073(*) ,026 LOG_CAP_SAL Pearson Correlation Sig. (2-tailed) -,221(**) ,000 ,046 ,160 -,003 ,924 1 . -,308(**) ,000 ARCSINLEV Pearson Correlation Sig. (2-tailed) -,235(**) ,000 -,116(**) ,000 -,073(*) ,026 -,308(**) ,000 1 .
* Correlation is significant at the 0.05 level (2-tailed). ** Correlation is significant at the 0.01 level (2-tailed).
Table 6. Tolerance and VIF Statistics
Pagina | 24 At last, this research will perform a Ramsey test to check for model misspecification. The Ramsey test tests whether non-linear combinations of the estimated values help explain the exogenous variable. If these non-linear combinations have any power in explaining the exogenous variable, then the model is mis-specified (Ramsey, 1969). Table 7 shows the output for the Ramsey Test, and it shows that the measured F-value does not exceed the critical value for the F-distribution which is 2.9957. Therefore, the null hypothesis that the coefficients on the added variables are jointly zero is accepted at the 5% level. This means that in order to describe the complete variation in the dependent variable and ascribe it to independent variables the model does not have to be improved. Also, the model does seem to be significant because the probability of the F-score is 0,0017. Nevertheless, it is possible to improve the regression since the constructed model does only explain 17% of the variance in the dependent variable and there seem to be various other factors explaining firm performance that are not included in this model. Since the model constructed suffices for this research, that tries to answer the question whether industry or firm-specific variables are more important in describing firm performance, improving the model is not necessary. Therefore, this research will continue using this model and process the results that can be found using this dataset.
Table 7. Ramsey RESET Test
N=939 Lag: 9 F-statistic 2.982647 Prob. F (9,925) 0.0017 Log likelihood ratio 26.86219 Prob. Chi-Square (9) 0.0015
Dependent Variable: ROA_2006 Method: Least Squares
White Heteroskedasticity-Consistent Standard Errors & Covariance
R-squared 0.176077 Adjusted
R-squared
Pagina | 25 6.2. Regression Results
Regression of the proposed model shows very significant results that should be able to answer the general question posed in this research. Although the model does only explain 17 percent of the measured variance in firm performance, it does show significant difference between the group of industry variables and the group of firm-level variables in explaining firm performance. Table 8 summarizes the results of the multiple regression test performed on the selected sample of 939 organizations for the year 2006. The results from the multiple regression test show support for the resource-based view. As well the effect of the capital-to-sales ratio (LOG(Cap)) as the effect of leverage (ARCSIN(Lev)) on firm performance (ROA) is strongly significant. The SCP-paradigm is not supported by the findings of this research. Neither industry concentration (LOG(Conc)) nor the type of industry (Ind) has a significant influence on the firm performance (ROA) in 2006.
Table 8. Regression Results
N=939
Dependent Variable: ROA_2006 Method: Least Squares
White Heteroskedasticity-Consistent Standard Errors & Covariance
Coefficient Std. Error t-Statistic Prob.
C 15.86817 1.490516 10.64609 0.0000 LOG_IND_CONC 0.360344 0.266191 1.353702 0.1762 IND TYPE -0.474450 0.588854 -0.805717 0.4206 LOG_CAP_SAL -1.894561 0.190544 -9.942883 0.0000 ARCSINLEV -0.193835 0.018995 -10.20457 0.0000
Pagina | 26 Secondly, we will check whether the one-sided causal relationship of ROA_2006 on IND TYPE has any effect on the coefficient of IND TYPE.
Table 9. Regression Results LOG_IND_CONC
N=939
Dependent Variable: ROA_2006 Method: Least Squares
White Heteroskedasticity-Consistent Standard Errors & Covariance
Coefficient Std. Error t-Statistic Prob.
C 8.980905 1.118071 8.032496 0.0000
LOG_IND_CONC 0.515595 0.281780 1.829780 0.0676
Pagina | 27 Table 10. Group Statistics / Independent Samples T-test
N=939
Dependent Variable: ROA_2006 IND
TYPE
N Mean Std. Deviation Std. Error Mean
0 376 6.9818 9.01072 0.46469
1 563 7.1447 9.78120 0.41223
t-test for Equality of Means
t df Sig. (2-tailed) Mean Difference Std. Error Difference Equal variances assumed -,258 937 0,796 -,16292 0,63141 Equal variances not assumed -,262 847,312 0,793 -,16292 0,62119
Pagina | 28
7.
Discussion
7.1. Introduction
In this chapter the findings from the regression model will be used to draw conclusions necessary to answer the general question posed in this research. The findings will be compared with the expectations earlier mentioned in this research. Where the results differ from the expectations, potential explanations for these contradicting findings will be proposed and discussed. Furthermore, this research will provide useful information for the discussion on drivers of firm performance. Besides that, this chapter will also get into the limitations of this research and provide recommendations for future research.
7.2. Summary of the Results
The constructed regression model used in this research provides a set of findings according to which this research will make some recommendations. First of all, the results found in this regression will be summarized. The first two regressors used in the model are industry-specific variables from which the first variable, LOG_IND_CONC, is found to have an insignificant positive effect on ROA_2006. The second industry-specific variable is IND_TYPE which has an insignificant negative effect on ROA_2006. From the two firm-specific variables both the first variable, LOG_CAP_SAL, and the second variable, ARCSINLEV, have a significant negative effect on ROA_2006. Conclusion of this summary is that the two industry-specific variables show an insignificant effect towards ROA_2006, contradicting the SCP-paradigm mentioned earlier, and the two firm-specific variables are found to have a significant effect on ROA_2006, supporting the resource-based view that has been introduced earlier in this research.
7.3. Discussing the Results
Pagina | 29 (LOG_IND_CONC) does show a positive effect on ROA_2006 but this effect is found to be insignificant and also the effect of IND_TYPE is insignificant but, in contrast to LOG_IND_CONC, negative. An explanation for the fact that LOG_IND_CONC does not seem to have a significant effect on ROA_2006 may be that there is strong anti-monopoly and anti-trust policy conducted by the European Commission. Whenever the European Commission suspects some form of trust within the market than they will impose disciplinary action. That may be an explanation why the effect of variance in industry concentration is minimal. On the other hand, the findings from this research are different than the findings from Nguyen and Devinney (1990), which may be explained by the fact that Nguyen and Devinney used a market-based measure of performance whereas this research uses a book-based measure of performance. The conducted research shows that IND TYPE also has an insignificant effect on ROA_2006. The expectation that service-oriented industries would have significantly more effect on firm performance is not found in this research. This is in contrast with findings from Hufbauer and Warren (1999) and Wölfl (2005) which reported that service-oriented industries lead to significantly more GDP growth and employment in developed countries. A possible explanation for this might be that the dataset used in this research includes countries from the former Eastern Bloc in which the costs of manufacturing are significantly lower than in the developed countries. This may be the cause that the effect of the shift towards service-oriented industries within the developed countries is offset by the increase in manufacturing-oriented industries in the countries from the former Eastern Bloc.
Pagina | 30 significant positive effect of capital-to-sales ratio on profitability, and Hay and Morris (1991), which states that the higher the capital intensity the greater the positive effect on profit margins of the firm. The findings for leverage are also in line with findings from earlier studies. Although Martin (1993) stated that “if leverage means greater risk and greater risk in turn means a greater profitability than the coefficient of the debt ratio is positive, that would mean that more leveraged firms (firms that have more liabilities) are more profitable, all else equal”, this research has found a significant negative effect of leverage on firm performance. This is in line with earlier empirical studies that found a negative effect of leverage on profitability which means that the borrowed capital has either not effectively been used or the costs of borrowing were simply higher than the benefits of the relevant investment. Another possibility is that the firms who finance their investments from their retained profits are more profitable than the firms that borrow capital for their investments. Secondly, the capital-to-sales ratio was also expected to have a positive effect on firm performance. Kwoka and Ravenscraft (1986) found a positive significant effect of the capital-to-sales ratio on profitability. Oustapassidis (1997) found in his study on the performance of strategic groups in the Greek dairy industry two different significant effects of the ratio of capital over sales on profitability. Oustapassidis found a significant negative effect of the capital-to-sales ratio on profitability for advertised firms and a significant positive effect of the capital-to-sales ratio on profitability for unadvertised firms. A potential explanation for the fact that the effect found in this research is also negative may be that the database used for this research consists of the 250.000 largest firms in Europe. These firms can be considered to be firms with branded products and, therefore, these firms can be classified as being advertised firms. This may be an important reason why this dataset shows an identical result for the effect of the capital-to-sales ratio on performance.
7.4. Recommendations for Future Research
Pagina | 32
References:
Andrews, K.R. 1971. The Concept of Corporate Strategy. Homewood, IL: Dow-Jones Irwin.
Bain, Joe S. 1959. Industrial Organization. Wiley, New York.
Bain, Joe S. & Qualls, P. David 1987. Industrial Organization: A Treatise. JAI Press, Greenwich, CT.
Barney, Jay B. 1986. Strategic Factor Markets: Expectations, Luck, and Business Strategy. Management Science, 32(10): 1231-1241.
Barney, Jay B. 1991. Firm Resources and Sustained Competitive Advantage. Journal of Management, 17(1): 99-120.
Bharadwaj, Sundar G. & Varadarajan, P. Rajan & Fahy, John 1993. Sustainable
Competitive Advantage in Service Industries: A Conceptual Model and Research Propositions. Journal of Marketing, 57(4): 83-99.
Camelo-Ordaz, C. & Martín-Alcazar, F. & Valle-Cabrera, R. 2003. Intangible
Resources and Strategic Orientation of Companies: An Analysis in the Spanish Context. Journal of Business Research, 56(2): 95-103.
Chang, Sea-Jin & Singh, Harbir 2000. Corporate Industry Effects on Business Unit Competitive Position. Strategic Management Journal, 21: 739-752.
Galbreath, Jeremy & Galvin, Peter 2008. Firm Factors, Industry Structure and
Pagina | 33 Gronhaug, Kjell, & Fredriksen Tor 1988. Concentration Ratios, Strategy and
Performance: The Case of the Norwegian Telecommunications Industry. Managerial and Decision Economics, 9(4): 257-262.
Hawawini, Gabriel, & Subramanian, Venkat, & Verdin, Paul, 2003. Is Performance
Driven by Industry- or Firm-specific Factors? A New Look at the Evidence. Strategic Management Journal, 24: 1-16.
Hatten, Kenneth J. & Schendel, Dan E. & Cooper, Arnold C. 1978. A Strategic Model
of the US Brewing Industry: 1952-1971. Academy of Management Journal, 4: 592-610.
Hay, Donald Andrew & Morris, Derek J. 1979. Industrial Economics: Theory and Evidence. Oxford University Press, Oxford.
Hay, Donald Andrew & Morris, Derek J. 1991. Industrial Economics and Organization, Theory and Evidence. Oxford University Press, Oxford.
Herfindahl, Orris C. 1950. Concentration in the Steel Industry. Unpublished Ph. D. dissertation, Columbia University, New York.
Hirschmann, Albert O. 1945. National Power and the Structure of Foreign Trade. Berkeley and Los Angeles, University of California Press, 1: 78.
Hoopes, David G., Hadsen, Tammy L., Walker, Gordon. 2003. Guest Editors'
Introduction to the Special Issue: Why is There a Resource-Based View? Toward a Theory of Competitive Heterogeneity. Strategic Management Journal, 24(10): 889-902.
Hufbauer, Gary & Warren, Tony 1999. The Globalization of Services: What Has
Happened? What Are the Implications? Working paper, Institute for
Pagina | 34 Kunkel, S.W. 1991. The Impact of Strategy and Industry Structure on New Venture
Performance. Unpublished Doctoral Dissertation, University of Georgia
Kwoka Jr., John E. & Ravenscraft, David J. 1986. Coöperation versus Rivalry: Price-cost Margins by Line of Business. Economica, 53(211): 351-363.
Martin, Stephen 1993. Advanced Industrial Economics. Blackwell, Oxford.
Mason, E.S. 1939. Price and Production Policies of Large Scale Enterprises. American Economic Review, Supplement 29: 61-74.
Mauri, Alfredo J., & Michaels, Max P. 1998. Firm and Industry Effects Within
Strategic Management: An Empirical Examination. Strategic Management Journal, 19(3): 211-219.
McGahan, Anita M. & Porter, Michael E. 1997. How Much Does Industry Matter, Really? Strategic Management Journal, 18: 15-30.
McGahan, Anita M. & Porter, Michael E. 2005. Comment on ‘Industry, Corporate and
Business-segment Effect and Business Performance: a Non-parametric Approach’ by Ruefli and Wiggins. Strategic Management Journal, 26(9): 873-880.
Nelson, Richard R. 1991. Why Do Firms Differ and How Does It Matter? Strategic Management Journal, 12: 61-74.
Nguyen, The Hiep, Serror, Ann, & Devinney, Timothy M. 1990.
Diversification Strategy and Performance in Canadian Manufacturing Firms. Strategic Management Journal, 11(5): 411-418.
Pagina | 35 Penrose, Edith T. 1959. The Theory of the Growth of the Firm. Oxford: Basil Backwell.
Porter, Michael E. 1979. The Structure Within Industries and Companies Performance. Review of Economics and Statistics, 61(2): 214-227.
Porter, Michael E. 1980. Competitive Strategy. Free Press, New York.
Rumelt, Richard P. 1991. How Much Does Industry Matter? Strategic Management Journal, 12(3): 167-185.
Scherer, F.M. 1980. Industrial Market Structure and Economic Performance. Houghton-Mifflin, Boston, MA
Schmalensee, Richard 1985. Do Markets Differ Much? American Economic Review, 75(3): 341-351.
Selznick, P. 1957. Leadership in Administration. Evanston, IL: Row Peterson.
Shepherd, William G. 1994. The Economics of Industrial Organization. Prentice-Hall International Editions, Englewood Cliffs, NJ.
Swartz, Teresa A. & Bowen, David E. & Brown, Stephen W. 1992. Fifteen Years after
Breaking Free: Services Then, Now and Beyond. In: Swartz, Teresa A. & Bowen, David E. & Brown, Stephen W. 1992. Advances in Services Marketing and Management: Research and Practice, Volume 1. Greenwich, CT, JAI Press.
Uri, Noel D. 1988. A Re-examination of the Relationship Between Industry Structure and Economic Performance. Applied Economics, 20(10): 1383-1400.
Pagina | 36 Weiss, L. 1974. The Concentration-profits Relationship and Anti-trust. Industrial
Concentration: The New Learning, Boston, MA: Little, Brown & Co., 55-97.
Wernerfelt, Birger 1984. A Resource-based View of the Firm. Strategic Management Journal, 5(2): 171-180.
Wernerfelt, Birger & Montgomery, Cynthia A. 1988. Tobin’s Q and the Importance of Focus in Firm Performance. American Economic Review, 78(1): 246-250.
Pagina | 37
Data References:
Performance Data:
Amadeus
http://www.rug.nl/bibliotheek/catalogibestanden/elekbestanden/gamma/index?lang=nl
University library subscription
Industry Structure Data: Amadeus
http://www.rug.nl/bibliotheek/catalogibestanden/elekbestanden/gamma/index?lang=nl
University library subscription
Firm-level Data: Amadeus
http://www.rug.nl/bibliotheek/catalogibestanden/elekbestanden/gamma/index?lang=nl
