Firm profitability and the size of the subsidiaries
network: Theory and evidence from the Real
Estate and Food-Beverages industries.
Master’s Thesis
University of Groningen
International Economics and Business
Bore Pupulkovski (1497588)
Master’s Thesis Supervisor: Dr. G. De Jong
Table of contents
1. INTRODUCTION 3
2. THEORETICAL FOUNDATIONS AND HYPOTHESIS 7 1.1 Introduction 7 2.2 OLI Paradigm 9 2.3 Criticism and extensions of OLI paradigm 10 2.4 The resources-based view theory 12 2.5 Network theory 13 2.6 Control variables 14 2.7 Theoretical model 16
3. METHODS 18
3.1 Data collection and sample 18 3.2 Measurements and controls 20 3.3 Econometric Techniques 23
4. RESULTS 26
4.1 Descriptive statistics 26 4.2 Comparison of the profitability results 27 4.3 Regression analysis results 28 5. SUMMARY AND CONCLUSION 33 5.1. Conclusions 33 5.2 Limitations and further research 34
List of references 36
1. INTRODUCTION
The importance of networking is relevant in every sector of business life. Some are born with networks while others develop networks during their lifetimes. It appears that every company is part of a network and the standard of living is becoming greatly influenced by the quality of business networks. For an illustration why networking is becoming more important let us consider the study from UNCTAD, (2000). This study claims that the main driving force towards the current trend of economic globalization has been a sudden increase in the number of cross-border inter-firm agreements. That means that networking has become one of the major organizational forms to come into view in the past decade, and is therefore interesting for research. Furthermore, many firms have entered into strategic alliances forming networks, the management of which has become a critical issue for MNE managers. Finally, the literature on strategic management recognizes the importance of networking and its implications for many of the core strategic management fields. Thus, the influence of networks on individual firm performance has become important for both managers and scientists, as elaborated by Dyer and Singh, (1998); Gulati et al., (2000) and Koka and Prescott, (2002). Therefore, the main research question in this study is: what the value of a subsidiaries network is for the focal firm? I intend to answer this research question by performing two different, yet related analyses. The first analysis, will involve comparison of firms with and without network on their profitability. The second analysis deals with estimating the impact of the subsidiaries network on the focal firm profitability.
However, perfect competition rarely exists in real markets. Also, firms are not all equally productive and they classify themselves in productivity levels as stated by Helpman et al. (2004). The different productivity levels are the result of certain firms possessing specific forms of competitive advantage. The more valuable the competitive advantage of the firm, the higher will be the profitability and the likelihood of outperforming the competitors, due to the maintenance of above-average profits in the long run. This micro-economics theory implies that firms constantly search for competitive advantage that will enable them to stay profitable. Competitive advantage for a firm may be the ownership of any valuable resource or capability.
Competition among firms has intensified in recent times as a result of globalization and easier communication. This urges firms to join a race to search for competitive advantage. One way of gaining competitive advantage is to own natural resources such as oil or gas. Another way is to acquire the source of competitive advantage by obtaining competitive knowledge or information before the other companies. The process of obtaining information and knowledge is most important and interesting for our research. For example, Wheelwright and Clark (1992) point out that the endurance of organizational vitality depends partly on the possibilities of imitating firms’ competitive advantage and their development of new processes and products. This may be seen as learning or obtaining knowledge from other firms. Therefore, a firm’s long-term success is derived from its ability to renew its skills at lower costs and over less time than its competitors, as stated by Prahalad and Hamel (1990).
Research in the field of networking has become more interesting in recent times. Its implications have also become recognized in other fields of strategic research. This may be illustrated by the fact that several research streams incorporate the effect of networking in their analysis of firms. More precisely, strategic literature has in the past typically viewed firms as autonomous entities striving to find competitive advantage from either external industry sources (e.g., Porter, 1980), or from internal resources and capabilities (e.g., Barney, 1991). In the more recent literature, research has shifted from viewing firms as autonomous entities to researching the connections that those firms have with each other, thus the firm network. For example, Gulati, Nohria, and Zaheer (2000: 203) suggested that the performance of a firm is greatly influenced by the network that the firm is embedded in and that firms can be better understood by examining their network relationships. Because of the increased interest in the network research this study intends to add value to the current research by investigating the benefits of the size of a subsidiaries network.
Subsidiaries can occupy different positions in their relations with the parent company. Rugman and Verbeke (2001) have studied the value-creating activities of the subsidiaries of MNE’s and they found out that subsidiaries may be associated with more than one value-creating activity at the same time. Therefore, they conclude that identifying the ‘role’ of the subsidiary in the network is becoming less relevant. This means that sharing knowledge and the process of creating competitive advantage may occur in all subsidiaries of the firm independently of the position of the subsidiary. For this reason, this study researches the impact of the size of the subsidiaries network, without distinguishing the positions of the subsidiaries. Previous literature has dealt with different aspects of networks. However, in general the relationship between the size of a subsidiaries network and its impact on focal firm performance to my knowledge has not this far been investigated.
price above its costs and/or to lessen average costs. The former situation will represent the firm’s monopoly power or differentiated positioning, while the latter would reflect the firm’s costs efficiency. For any of the previously described effects to take place, the firm must have specific competitive advantage that will enable it to increase the profit margin to a higher level than its competitors. It was previously indicated that networking may play one of the crucial roles in the process of creating competitive advantage. Therefore, this study utilizes profit margin as a measurement of profitability of the firm.
Previous similar studies examining the impact on firms’ profitability or performance have mainly utilized other measures that reflect the overall performance of the firm. Among the most common performance measurements are ‘return on assets’ and ‘return on equity’. These measurements capture the overall performance of the firm while the profit margin, as explained in the previous paragraph, captures the value of the competitive advantage of the firm. Therefore, the impact of the size of a subsidiaries network on the profitability of the firm measured as profit margin will be investigated.
2. THEORETICAL FOUNDATIONS AND HYPOTHESIS
2.1 Introduction
Gulati, Nohria and Zaheer (2000) identify 5 fields in strategic research where the incorporation of network analysis carries additional value and explanatory power. The first such field is industry analysis. The application of network analysis to industry analysis can give a better understanding of the profitability of firms within an industry. The second field in which networking will bring additional value is explaining the position of the firm in the industry. In contrast to earlier studies, when the position of the firm in the industry was determined on the basis of market share, size and productivity, a better understanding of the firm’s position may now be obtained by including network analysis. Third, networks might play a crucial role in understanding the dynamics of an industry. This is because the previous networking experience of a firm will influence its future decisions in choosing a new partner for a strategic alliance. Fourth, implementing network theory in transaction costs theory may facilitate explanation of why firms engage in networking. More precisely, it is easier and more cost effective to engage in new networks for firms that already have a network and are experienced in networking, than for firms that are engaging in a network for the first time. The last field of research is resources-based theory. In this field, a network may be seen as a source of competitive advantage. Furthermore, every network has its unique characteristics that carry competitive advantages to the firms embedded in it, as pointed out by Gulati, Nohria and Zaheer, (2000).At this point may be worth mentioning that in previous literature networks have been defined differently. Therefore it is useful to state the definition that will be used in the current study. In this study the size of subsidiaries network is defined as
number of subsidiaries that are owned by the focal firm with any percentage of ownership up to fourth level of indirect ownership. The measurement of the size of
subsidiaries network (discussed in 2.6) elaborates more explicitly on this definition.
industry can no longer be meaningfully analyzed without considering the implications of networking. Recent changes due to inter-firm linkages in the US automobile industry include fewer suppliers, greater supplier involvement in design processes and long-term relationships. These intensified linkages have improved the overall competitiveness of the US automobile industry as stated by Dyer (1996), and Gulati and Lawrence (1999). Other examples of networking and their strategic importance for firms in the industry have been discussed; for examples see Womack, Jones and Roos, (1990). Therefore, current business analysis can not be imagined without an awareness of the strategic networks in which firms are embedded.
Among the subjects most usually studied in the field of network literature are specific parts of networks such as strategic supplier networks, (Jarillo, 1988; Dyer and Singh 1998), or specific aspects of a network such as learning processes in strategic alliances (Hamel, Doz, and Prahalad, 1989), or benefits from building trust among alliances, (Gulati, 1995; Zaheer and Venkatraman, 1995). A firm that is owned by another firm, thus a subsidiary, has the capacity to share knowledge created by the parent company or by the various units in the parent company network, as stated by Bartlett and Ghoshal (1989). Moreover, some researchers address aspects of the whole network, such as network diversification (see for example Goerzen and Beamish, 2005). In summary, the growing literature on networking is focused on specific features of the network, namely: (1) parts of the network, such as supplier or customer networks; (2) aspects of the network, such as trustworthiness, sharing of information and creation of knowledge; (3) the effects of being a subsidiary of another firm or (4) diversification of the whole network. The increasing body of literature in the field of networking highlights the attractiveness of network studies and points out the importance of condensing and focusing research in this field, (see Gulati, Nohria and Zaheer, 2000). Therefore, this study intends to identify another aspect of importance for understanding the profitability of firms.
example, comparing the productivity levels of firms from different countries (e.g. Keay, 2000). Also, a comparison of the economic performances of firms that are domestically and foreign owned has been made (see Bellak, 2001). Beside the increased interest regarding strategic networking in the literature, in general the profitability of firms with and without a network of subsidiaries has not been compared to my knowledge.
The second aim is to find empirical evidence for the impact of the size of a subsidiaries network on the focal firm profitability. Previous literature has examined similar, but still different aspects. For example, Goerzen and Beamish (2005) elaborate on the diversification of the network and performance of the firm, using network size as a control variable. This study examines 580 Japanese MNC’s with 13,529 subsidiaries in terms of the diversity of the network and performance. The study of Goerzen and Beamish differs significantly from the current study in two main aspects. The first aspect is that the current study examines the exact relationship between the size of a subsidiaries network and profitability assuming an inverse-U-shaped relationship (discussed in section 2.4), while the study of Goerzen and Beamish suggests a positive relationship. The second aspect is that the current study utilizes profit margin as a measurement of profitability, while the study of Goerzen and Beamish utilizes other measurements of performance. The relationship between subsidiary network size and profitability, as examined in the current study, in my knowledge has generally not been examined previously.
The implications of network size and its impact on firms’ profitability may be interpreted as another theoretical dimension in network research and as such a fruitful venue for further researchers. Evidence of the presence of value for firms with a network relative to those without will have practical implications and recommendations for autonomous firms that are independently striving for profitability. Finally, estimating the impact of network size on profitability will help managers in making strategic decisions.
2.2 OLI Paradigm
comparative advantages, defined as Ownership, Location and Internationalization (OLI). A firm’s Ownership advantage could be a product, production process, such as a patent, or trade secret to which other firms don’t have access. The Location advantage means that the foreign market must have characteristics that make it profitable to produce the product in the foreign country righter than production at home, such as low tariffs, transport cost, cheap factor prices, access to customers. Finally the firm must have motives to establish a production abroad rather than license the foreign firm to produce the product or use the process. These motives are referred to as Internationalization advantage. Therefore, Dunning with the OLI paradigm classifies firms into three types according to their competitive advantage. That is the firms with the best competitive advantage can undertake FDI and thus develop a subsidiaries network in foreign countries. The firms that have less competitive advantage are incapable of undertaking FDI but still capable of exporting to foreign countries. Finally, the firms with the lowest competitive advantage serve only the home market.
Dunning explains clearly how the level of engagement of firms in foreign markets, and thus with a developing network of subsidiaries, depends strongly on the possession of valuable competitive advantage. This means that the competitive advantage of firms is positively correlated with the establishment of a subsidiaries network. When the OLI paradigm was first established, Dunning assumed that the competitive advantage of the firms originates from sources within the firm, therefore ignoring the possibility of creating competitive advantage from possession of a network. Even so, the OLI paradigm is well known for explaining the behavior of firms with respect to internationalization. In contrast, the present study aims to prove that firms possessing a network are capable of creating competitive advantage. Therefore, it is worth mentioning that Dunning (1995) elaborates on how in the last two decades capitalism has changed drastically from its original definition, and goes on to elaborate on criticism and extensions of the OLI paradigm.
2.3 Criticism and extensions of OLI paradigm
age of alliance capitalism where networks of strategic alliances are playing a significant role in the process of explaining company endurance and profitability. More precisely, Dunning explains how the old concept of the OLI paradigm should be extended to include the aspect of networks. More precisely, firms should not be analyzed only internally as was the case earlier, but the networking of the firms should also be taken into account. The following citation illustrates the need to make changes to the OLI paradigm to incorporate the importance of the subsidiaries network:
“The second revered concept that is now under scrutiny is that the resources and competences of wealth-creating institutions are largely independent of each other; and that individual enterprises are best able to advance their economics objectives, and those of society, by competition righter than cooperation... Although, for more than a century, scholars have acknowledged that the behavior of firms may be influenced by the actions of their competitors (Cournot, 1851), while Marshall (1920) was one of the first economists to recognize that the spatial clustering or agglomeration of firms, with related interests might yield agglomerative economies and an industrial atmosphere, external to the individual firm, but internal to the cluster.
… The OLI configuration determining trans-border activities is being increasingly affected by the collaborative production and transactional arrangements between firms; and this need to be systematically incorporated into the eclectic paradigm.”(Citation: Dunning, 1995, page 2)
This modification of the OLI paradigm implicitly points out the importance of firm networking as a source of creating competitive advantage. Therefore, with this criticism and proposed extensions Dunning recognizes that competitive advantage should not only be considered to arise from internal sources but also from the network of the firm.
remaining firms are classified according to productivity levels as follows: the least productive choose to serve only the home market, the more productive firms serve the home and foreign markets with exports, while the most productive firms serve the home and foreign markets with FDI. With his study Helpman points out that firms that have subsidiaries network as a result of the FDI activities, are the most productive ones.
Moreover, networks act as a source of knowledge and information, as stated by Dyer and Singh (1998) and Gulati (1999). Therefore, the network may be thought of as a unique and irreplaceable, valuable resource or a source of competitive advantage for the firm.
The main points of these studies can be summarized as follows. First, the OLI paradigm should be extended to recognize the network as a source of creating competitive advantage. Second, the model of Helpman explains how the less productive firms are threatened with being forced to leave the market while the most productive firms are those with a network. Third, networks may be a source of competitive advantage for the firm. Combining these points, one may derive the conclusion that firms with networks are more profitable than firms without networks. To test for this empirically I established the first hypothesis.
H1: Firms with a network of subsidiaries have a higher
profitability than firms without a network of subsidiaries.
2.4 The resources-based view theory
recent times. Davis and Greve (1997) and Palmer et al. (1995) indicate that networks enable firms to exchange new forms of practice among themselves faster, and also facilitate the transmission of information.
On one hand, possessing a larger network of subsidiaries will mean more sources of information for the focal firm and therefore potentially more benefits. On the other hand, managing a firm’s network involves managerial time and requires effort for the application of appropriate governance mechanisms, developing inter-firm knowledge, sharing routines, and making appropriate relation-specific investments (Dyer and Singh, 1998). As such, the network can be described as an inimitable and irreplaceable resource of the firms, but at the same time as a possible constraint, as also elaborated by Gulati, Nohria, Zaheer (2000). For this reason the size of the network of subsidiaries may positively influence the profitability of the firm but may also have drawbacks due to the demands placed on managing the network. The network theory (described in section 2.5) elaborates more explicitly on this relationship.
2.5 Network theory
Current literature suggests three aspects where firms may gain economic benefits from efficient inter-firm linkages (Burt, 1992). The first is increased access to information, considering that networks provide a larger extent of information compared to what an individual firm possesses alone. The second aspect is that information may be obtained earlier in comparison with an individual firm, which yields competitive advantage for the firm that gains the information earlier. The third benefit is that the interests of the focal firm are presented to third parties in a positive light. Because of these reasons it has been suggested that firms with an effective set of linkages to other firms will have reliable contacts in more than one location, where useful information may surface (Burt, 1992). Therefore, owning a larger network of subsidiaries will provide more benefits to the focal firm.
expected that positive effects of increasing the network size diminish and eventually turn into negative effects at larger network size levels, when the marginal benefits are overwhelmed by the marginal costs of increased managerial effort. For this reason, the inverse-U-shaped relation between the size of the subsidiaries network and the profitability is based on the allocation of managerial time and organizational effort. Bearing in mind that there might be an inflection point after which the marginal costs of increasing the network size are higher than the marginal benefits, and thus that the overall effect of increasing network size is negative for the focal firm, an important question is: why might firms go beyond the threshold network size levels? According to Harbison and Pekar (1998), the dynamic new environment consisting of the globalization of markets has increased competition. This intensifying competition among firms has forced top managers to search for new capabilities, and to become involved in more strategic alliances with various firms in the hope of ensuring that at least some of them will yield strong positive results. Bearing in mind these arguments, the second hypothesis is as follows:
H2: The size of the subsidiaries network will have an
inverse-U-shaped relationship with focal firm profitability.
2.6 Control variables
The model that will be used to test the second hypothesis assumes that the size of the subsidiaries network influences the profit margin of the firm. However, there are other factors that also influence the profit margin. In order to properly estimate the effect of the size of the subsidiaries network, these factors should be taken into account. Therefore, external factors that may influence the profit margin independently of the network size are used in the model as control variables. These control variables are detailed below.
when a firm has existed for a long time, this implies that it performs in a healthy fashion, since it has operated for many years.
The size of the focal firm is important. In general, larger firms are associated with
better performance because of the economics of scale and scope. Another aspect is that firm size may be proven to advance performance through facilitating access to lower cost of capital, while at the same time lowering risk (Chang and Thomas, 1989). Therefore, it is important to include the firm size as a control variable.
Debt-to-equity ratio, or financial leverage, is widely used in prior research, because it
contributes to the risk-return outcomes (Buhner, 1987). Jensen (1989) argues that capital structure, defined as the debt-to-equity ratio, affects firm performance. This is because risk might be decreased with an optimum level of debt-to-equity. Therefore, the profits may be directly influenced by the debt and equity structure of the firm. Thus, debt-to-equity is an important control variable.
Intangible assets of the focal firm. Current network research has been criticized for
not taking into account proprietary assets (see Dess et al., 1995). Proprietary assets seem to be an important variable when investigating the profitability of a firm, because they may directly represent competitive advantage of the firm. For example, firms that gain more knowledge from R&D or possess brand reputation from marketing investments are capable of asking higher prices for their products. Prior research (e.g., Caves, 1996) utilizes proprietary assets as R&D Intensity and Marketing Intensity taken as a share of the R&D and Marketing expenses from the operating revenue. Both expenditures contribute to building the intangible assets of the firm. Therefore, in this study intangible assets are used as a control variable.
Cash flow of the focal firm is a measure of the ability of the firm to generate cash
flows in the future. Alternatively, it may be seen as a measure of free cash flow (Doukas
et al., 1999) which is directly influencing profitability. Therefore, cash flow is an
important control variable.
Therefore, the benefit of creating competitive advantage applies to both the subsidiary and the parent firm. As explained so far, this study examines the impact of the size of the network of subsidiaries that the company possesses. Still, some of the firms in the sample that own a network of subsidiaries may be themselves owned by another firm. Therefore, ownership dummies will be used to control whether the focal firm of the network is owned by another firm. Graph 1 in appendix A defines and explains what is the network of subsidiaries, investigated in this study and what is the ownership of the focal firm, that is used as a control in this study. Ownership will be split into three levels:
1. Firms that are less than 25% owned by another firm will be controlled with
dummy ownership <25%.
2. Firms that are between 25% and 50% owned by another firm will be controlled with dummy ownership 25%-50%.
3. Firms that are more than 50% owned will be controlled with dummy
ownership >50%.
For this purpose 3 dummies will be used. The rest of the firms are not owned by another firm.
2.7 Theoretical model
After explaining the theory, building the hypotheses and elaborating on the control variables, the model that will be used for this study is now formulated:
Profit margin =β0 + β1(number of subsidiaries) + β2(number of
subsidiaries)2 + β
3log(employees) + β4(age) + β5(debt-to-equity) + β6(intangible assets) + β7(cash flow) + β8(dummy ownership >25%)
+ β9(dummy ownership 25%-50%) + β10(dummy ownership > 50%) +
(ε)
β0 is a constant that theoretically represents what the profit margin will be if all the
other variables are 0. The βι (i=1 to 10) coefficients represent estimates that show the
direction and intensity of the impact of the consequent variable on the dependent variable
3. METHODS
3.1 Data collection and sample
Secondary data from the database AMADEUS were used in the collection of data for the empirical analysis of the hypothesis. The AMADEUS database provides data on European companies. The sample for empirical analysis was created via several steps. Firstly, the two-digit UK SIC (2003) codes for the real estate and the food and beverages industries were selected. Burt (1992) has shown that industries that occupy structural holes enjoy greater returns. As defined by Burt certain industry is rich with structural holes, to the extend that the producers are more concentrated (or less in number) while the suppliers and customers are many and disorganized. Moreover Burt (1992) has shown that services industries are more abundant with structural holes than manufacturing industries. This implies that different industries are capable of benefiting from the network to a different extent. This study therefore, analyses two industries with fundamental differences. More precisely, the food and beverages industry manufactures fast moving consumer goods, while the real estate industry provides services in durable goods. Moreover Burt distinguishes the real estate industry as industry with many producers, but yet as industry that retains extremely high profit margins. Therefore, comparison of the results from the current study to the findings of Burt might have additional value. Another point to bear in mind might be that the food and beverages industry is constrained by the concentration of its customers, since most of the largest supermarkets are organized in powerful chains. After selecting for this 2 industries data was downloaded
some of the data and were thus deleted from this sample. Regression analyses were performed in SPSS and in EVIEWS. Using both software packages allowed different estimations to be used. For example, ordinary least squares regression is available in SPSS, while EVIEWS has an option to apply the White technique to control the heteroskedasticity (discussed in 3.3). For the final sample, firms that have more than 100 subsidiaries (which are more than 10 standard deviations, see table 7 and 8, Appendix B) in their network were considered as outliers and were deleted. Therefore, data of 3396 firms in the real estate industry and 4986 firms in the food and beverages industry remained.
It may be worth explaining the selection procedure for some of the variables. Primarily, the variable number of subsidiaries accounts for all the subsidiaries that are partly or wholly owned by the focal firm. One may argue that subsidiaries of which only a small percentage are owned should not have been included in the network. Conversely, social network theory argues that weak ties may have strength (Granovetter, 1983). More precisely, Granovetter points out that a social system that does not have weak ties will be fragmented and incoherent. Also, new ideas will spread slowly and scientific endeavors will be handicapped. This is because the weak ties may always be a source of fresh ideas. The same will apply to firms. More precisely, owning a small stake in a remote firm, without much co-operation, may be a source of different ideas and thus beneficial. Therefore, subsidiaries that have weak ties with the focal firm are also important for generating ideas.
less significant in most cases, selecting the 6th or higher levels might lead to overestimation of the number of subsidiaries that are in the network of the focal firm. That is because the benefits from the 6th level subsidiary will be less effective. Therefore, an arbitrary decision was made to choose the 4th level of indirect ownership of the subsidiaries which are accounted in the network of the focal firm. Also, it may be worth mentioning that the observed firms can not be a fully owned subsidiary from another company. This means that if A owns B with 100% and B owns C with some percentage than A will have 2 subsidiaries (this are B and C), but B will not be listed as company with one subsidiary (that is C) because B is fully owned by A and is thus subsidiary itself. Finally, the AMADEUS database has a so-called ‘independence indicator’, that is created from the databases. It indicates whether the firm is owned by another firm, giving different levels of ownership. This indicator is used to create the dummy variables that control for the independence level of the firm. The rest of the variables were selected for the year 2004 in the Euro currency.
3.2 Measurements and controls
Table 1 Description, measurement, type, and expected sign of variablesa
Variable Measurement Type
variable
Expected sign Profit margin
((Operating profit before tax + Financial Profit)/ (operating
revenue))*100
Dependent
variable / Size of subsidiaries
network Number of subsidiaries
Independent
variable Positive+
Size of subsidiaries
network2 (Number of subsidiaries)2
Independent
variable Negative- Size of the focal firm Log(number of employees) variable Control Positive+ Age of the focal firm 2004-(date of incorporation) Control
variable Positive+ Debt-to-equity of the focal
firm
((Non current liabilities + loan)/ (shareholder founds)) * 100
Control
variable Negative- Intangible assets of the
focal firm
Intangible fixed assets in thousands Euro
Control
variable Positive+ Cash flow of the focal firm (Profit + depreciation) in
thousands Euro
Control
variable Positive+ Ownership dummy below
25%
Dummy (value 1 if less than 25% of the focal firm is owned
by another firm)
Control
variable Positive+ Ownership dummy
between 25% and 50%
Dummy (value 1 if between 25% and 50% of the focal firm
is owned by another firm)
Control
variable Positive+ Ownership dummy above
50%
Dummy (value 1 if more than 50% of the focal firm is owned
by another firm)
Control
variable Positive+
a Size of subsidiaries network2- indicates squared term Source: The author
may be used to test for SCP relationships in empirical studies when one controls for differences in capital intensity. Therefore, profit margin is an appropriate measurement for testing the relationship between profitability and subsidiaries network size.
The most commonly used measurements of firms’ profitability or performance in the current literature are excess market value, average market value or Tobin’s q. Excess market value is the ratio of the market value plus the book value of the debt minus total assets, all divided by net total sales (Allen and Pantzalis. 1996). Average market value is the market value divided by total assets, and Tobin’s q is the ratio of market value to book value. Using such measurements may facilitate a comparison of the results and adds to the robustness of the study. However, the calculation of such measurements requires knowledge of the market value, which is not available for many firms in the sample. This is because lots of these firms are not publicly traded and thus the market value is not available. Alternatively, other accounting measurements, such as return on equity and return on assets, are often criticized as being unreliable. This is because of varying accounting principles in different countries. Managers may also manipulate these ratios for a variety of reasons, such as the avoidance of corporate or personal taxes, especially in the case of small firms (see Sapienza, Smith, and Gannon, 1988; Powell and Dent-Micallef, 1997).
As mentioned in section 2.7, the variable number of subsidiaries is added a second time in the model for estimating the inverse-U-shaped relationship. The mechanism behind the econometric technique will now be explained. Adding the square of the term
number of subsidiariesyields a model in which the response of profitability to the size of the subsidiaries network depends on the level of the number of subsidiaries. More precisely, for lower levels the impact of the number of subsidiaries may be positive, while for higher levels it may be negative. In the formula below the variable profit
margin will be denoted as (PM) and the variable number of subsidiaries will be written as (subs). Assuming that all the other variables in the model are held constant, the response
of the expected E (PM) to subs is given by the first derivative of the regression equation with respect to subs. Therefore:
When subs increases by one unit and the rest of the variables are held constant, the expected value of PM will increase by β1 + 2β2 (subs). In order to determine the
anticipated signs of β1 and β2, we will assume that subsidiaries increase performance,
thus it is expected that β1 > 0. Furthermore, to achieve diminishing and negative returns
after a certain level, the response of performance increase to the number of subsidiaries must decline and become negative as subs increases. Therefore, it is expected that β2 < 0
3.3 Econometric Techniques
This section explains the statistical methods that will be used in order to empirically test the hypothesis. The first hypothesis is based on a simple comparison of the means of the two independent samples, while for the second hypothesis a multiple regression analysis will be used.
Comparison of means of the profitability of firms with a subsidiaries network and firms without a subsidiaries network
In order to estimate whether firms that own subsidiaries have higher profitability than those without subsidiaries, the means from the two independent samples were compared. Firms were sorted into two groups: 1. the firms that have a minimum of two subsidiaries and 2. the firms that do not have any subsidiaries. Firms with one subsidiary were excluded from both samples, because such firms do not belong to either of the two groups (owning one subsidiary is not a network, but also cannot be classed as being without a network). The SPSS software was used to calculate the means of the profitability and the standard deviations in the two separate samples. Whether the means difference is significant was estimated using the t-test. The t-test technique takes into account the standard deviations of the two independent samples and the number of observations.
Multiple regression analysis and assumptions
MR1. The value of y for each value of x is Y = b0+ b1 x +b2 x2....+εt
MR2. The average value of the random error e is E (ε) = 0
MR3. The covariance between any pair of random errors is cov (εi, εj) = 0
MR4. The values of xtk are not random and are not highly correlated linear functions
of the other explanatory variables.
MR5. The variance of the random error e is var (ε) and is constant
The first assumption, MR1, assumes that there is a linear relationship between the dependent and the explanatory variables. Here, b0 is the intercept, and b1, b2 … bn are the
slopes of the function. Finally, εt represents the error term consisting of all the factors
influencing Y other than the specified exogenous variables included in the model.
MR2 implies that the error ε is a random variable with an average value of zero. This means that the linear model is a single line that minimizes the value of the sum of all the error terms. In other words, the average error term should have a mean of zero as all the negative and positive error terms cancel each other out.
According to MR3, the errors are not autocorrelated. For cross-sectional data, the randomness of the sample implies that the error terms for different observations would be uncorrelated. Autocorrelation mostly appears as a problem in time series analysis. However, sometimes this assumption might be violated in cross-section data and it will thus be formally tested, using the Durbin-Watson test.
The multicolinearity, MR4, is an assumption that the independent variables are not highly correlated. If this assumption is not fulfilled, the estimates may seriously deviate from the real values. Normal procedure to test for multicolinearity is to make a bivariate correlating table with pairs of variables, where correlation above 0.8 indicates a problem. In addition, the “variance inflation factor” (VIF) will be calculated, in order to establish whether the estimations of the parameters are seriously inflated by correlation. Hair et. al. (2006) argues that a VIF above 10 will mean the estimators are inflated.
but it is no longer consistent and as a result the regression output, in terms of test statistics, can no longer be considered reliable. This is so because the variance, which is at the heart of these statistics, is no longer constant and will hence be misleading. For normal cross-sectional data, the distribution of the error term (ε) should graphically show a constant distribution in relation to the function Y = b0 + b1 X +…+ ε. Any pattern,
when plotting the residuals against the predicted and other individual independent variables, will show a heteroskedasticity problem. The sample used in the current study exhibits a degree of heteroskedasticity. Therefore, some of the available procedures for correcting for heteroskedasticity will now be explained, with an emphasis on what is most appropriate for the current sample.
4. RESULTS
4.1 Descriptive statistics
The firms in both industries vary from very large to very small. Furthermore, the age of the firms varies from old to relatively young. The high frequency of the variables that is not necessarily concentrated around the mean value is responsible for the absence of a normal distribution of the variables. Another noticeable feature is that in the real estate industry firms have on average a much higher profit margin of 11 compared to 3 in the food and beverages industry. The mean of the number of subsidiaries is 3.9 in the real estate industry while in the food and beverages industry it is 2.5. Another variable that has a higher mean in the real estate industry is the debt-to-equity ratio. Regarding the other variables, size, age, and intangible assets are noticeably higher in the food and beverages industry. Cash flows are slightly smaller in the real estate industry. The descriptive statistics are presented in appendix B, tables 7 and 8, for both industries.
Bivariate statistics
The pair-wise correlations of the variables are presented in the table below. The highest correlation in the real estate industry is 0.42, showing that in general the variables are little correlated, implying that there is no potential problem of multicolinearity.
Table 2 Bivariate Correlations in the Real estate industry
Profit
margin subsidiaries Number of Age Log(employees) Debt-to-equity Intangible assets flows Cash
Profit margin 1 Number of subsidiaries 0.065 1 (0)*** Age -0.033 -0.003 1 (0.052)* (0.842) Log(employees) -0.206 0.19 0.154 1 (0)*** (0)*** (0)*** Debt-to-equity -0.135 -0.073 -0.058 -0.091 1 (0)*** (0)*** (0.001)*** (0)*** Intangible assets -0.009 0.076 -0.016 0.017 -0.01 1 (0.588) (0)*** (0.348) (0.328) (0.538) . Cash flows 0.122 0.183 0.036 0.168 -0.048 0.423 1 (0)*** (0)*** (0.034)** (0)*** (0.004)*** (0)***
*significant at 10%; **significant at 5%; ***significant at 1% Significance levels are in parenthesis
In the Food end beverages industry the highest correlation is 0.76 for cash flows and intangible assets and 0.49 for cash flow and number of subsidiaries. This may signal a potential problem of multicolinearity. Therefore, an additional test with the VIF was applied, the results of which are presented in appendix C, table 9. Also the Durbin-Watson test for autocorrelation is presented in appendix C table 9 and is conclusive for no autocorrelation.
Table 3 Bivariate Correlations in food and beverages industry
Profit margin
Number of
subsidiaries Age Log(employees)
Debt-to-equity Intangible assets Cash flows Profit margin 1 Number of subsidiaries 0.074 1 (0)*** Age 0.04 0.093 1 (0.004)*** (0)*** Log(employees) 0.006 0.357 -0.092 1 (0.66) (0)*** (0)*** Debt-to-equity -0.137 -0.014 0.016 -0.04 1 (0)*** (0.322) (0.254) (0.005)*** . Intangible assets 0.041 0.311 0.013 0.183 0.001 1 (0.004)*** (0)*** (0.346) (0)*** (0.965) . Cash flows 0.133 0.489 0.05 0.316 -0.016 0.764 1 (0)*** (0)*** (0)*** (0)*** (0.25) (0)*** . *significant at 10%; **significant at 5%; ***significant at 1%
Significance levels are in parenthesis Source: The author’s calculations
4.2 Comparison of the profitability results
The results from the empirical analysis comparison of the profit margins of the firms with and without a subsidiaries network are presented below.
Table 4 Comparison of means of profit margins for firms with and without subsidiaries
Industry Firms with subsidiaries network
Firms without subsidiaries network
Difference in
profit margin t-value Real Estate
14.68 8.69 5.99 7.12***
Food and
Beverages 3.48 2.76 0.73 2.79***
As can be seen from the table, the profit margin for firms with a subsidiaries network is higher than that of firms without a subsidiaries network. The difference between the means is much larger in the real estate industry (5.99) than in the food and beverages industry (0.73). Both industries show a significantly higher profit margin for firms with a subsidiaries network at the 1% level. This provides support for hypothesis one.
4.3 Regression analysis results
For the second hypothesis a multiple regression analysis was performed. The results from this analysis are given below. The first table shows the results from the real estate industry.
Table 5 Results from the regression analysis for the Real Estate Industry a
Dependent variable Profit margin Profit margin Profit margin
Constant 24.0251 23.1839 21.1282 (20.648)*** (19.622)*** (15.999)*** Subsidiaries 0.4725 0.4034 (4.802)*** (4.061)*** Subsidiaries2 -0.0051 -0.0043 (-3.358)*** (-2.825)***
Age of the focal firm -0.0071 -0.0033 -0.00046
(-0.641) (-0.301) (-0.040) Size of the focal firm -3.3483 -3.5482 -3.5503
(-13.016)*** (-14.175)*** (-14.200)*** Debt-to-equity of the focal firm -0.0040 -0.0039 -0.0039
(-8.844)*** (-8.402)*** (-8.505)*** Intangible assets of the focal firm -0.0186 -0.0186 -0.0183
(-1.323) (-1.493) (-1.474) Cash flows of the focal firm 0.1610 0.148 0.145
(3.560)*** (3.491)*** (3.470)***
Ownership dummy below 25% 2.0779
(1.668)*
Ownership dummy between 25% and 50% 3.6944
(2.757)***
Ownership dummy above 50% 3.2598
(3.621)***
Adj. R-squared 0.094 0.103 0.106
Durbin-Watson 1.904 2.012 2.016
F-value 71.608 56.498 41.224
N-number of observations 3396 3396 3396
a For calculations the White covariance estimator was used *significant at 10%; **significant at 5%; ***significant at 1% t- values are in parenthesis
It is apparent that only the variables age and intangible assets are not significant. The rest of the variables are significant at the 1% level. The Adjusted R-square is 0.104, which means that 10.4% of the variation of the profit margins can be explained using the dependent and control variables. The variables that measure the effect of the size of the subsidiaries network are as expected and confirm the inverse-U-shaped relationship as hypothesized. The following table presents the results for the food and beverages industry.
Table 6 Results from the regression analysis for the Food and Beverages Industry a
Dependent variable Profit margin Profit margin Profit margin
Constant 4.6652 4.8952 4.3069 (7.755)*** (7.841)*** (6.687)*** Subsidiaries 0.1019 0.0993 (2.297)** (2.214)** Subsidiaries2 -0.0017 -0.0017 (-2.049)** (-1.956)*
Age of the focal firm 0.0056 0.0047 0.0047
(1.958)** (1.659)* (1.658)*
Size of the focal firm -0.3362 -0.4088 -0.4302
(-2.980)*** (-3.424)*** (-3.600)***
Debt-to-equity of the focal firm -0.0025 -0.0024 -0.0024
(-5.964)*** (-5.946)*** (-6.024)***
Intangible assets of the focal firm -0.0094 -0.0093 -0.0092
(-2.642)*** (-2.776)*** (-2.773)*** Cash flows of the focal firm 0.0462 0.0478 0.0470
(3.867)*** (4.117)*** (4.067)***
Ownership dummy below 25% 0.3108
(0.633)
Ownership dummy between 25% and 50% 0.6817
(1.988)**
Ownership dummy above 50% 1.0492
(3.549)***
Adj. R-squared 0.047 0.049 0.051
Durbin-Watson 1.865559 1.471 1.474
F-value 50.26300 37.357 27.677
N-number of observations 4986 4986 4986
a For calculations the White covariance estimator was used *significant at 10%; **significant at 5%; ***significant at 1% t- values are in parenthesis
Source: The author’s calculations
R-square is 0.05 which means that only 5% of the variation in profit margin can be explained by the empirical model in the food and beverages industry. The variables for the size of the subsidiaries network are significant at 5 % and 10% levels. They also support the second hypothesis.
It is apparent for both industries that the estimated coefficients that determine the influence of the network size on the profitability of the company decline when controlling for whether the firm is owned by another firm. In other words, the coefficient for the variable subsidiaries is lower in the second model where dummy variables are used than in the model without dummy variables.
The difference of the impact of network size on profitability between the two industries is also considerable. In the food and beverages industry there is a positive effect of 0.099, which for the real estate industry is 0.403. There is a negative effect from the large size of the network of 0.0017 for the food and beverages industry, which for the real estate industry is 0.0043. The following figure helps to visualize the meaning of these values. -8 -6 -4 -2 0 2 4 6 8 10 12 1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 91 96 Number of Subsidiaries Im p ac t o n Pr o fit M ar g in Real Estate Food & Beverages
The figure clearly shows the inverse-U-shaped relationship between the network size and the profitability in the two industries. It is clear that the firms in the real estate industry derive more benefit from networking than firms in the food industry. The same argument is provided by Burt (1992), which argues that some industries obtain higher benefits from their networks than others. In the food and beverages industry, marginal benefits increase up to the 30th subsidiary where the impact reaches a maximum of 1.48, above which the impact decreases, while in the real estate industry marginal benefits are positive as far as the 47th subsidiary where the impact reaches a maximum of 9.55. In the food and beverages industry the impact becomes negative after the 60th subsidiary, which occurs in the real estate industry after the 95th subsidiary. Considering the mean values of the network size in the two industries in the current sample (see tables 7 and 8 in appendix B, the value for food and beverages is 2.5 with a standard deviation of 6.4, and 3.9 with a standard deviation of 8.6 for real estate), one may easily conclude that only a very small number of firms have such a large network size that results in a negative impact on profitability. On the contrary, the average size of the network is quite small compared to the point at which maximum benefits are obtained.
Additionally, the dummies for ownership may be important in providing an explanation since they carry impact for the profitability of the firm. In the real estate industry all three dummies are significant. This means that at any level of ownership, the firm as a subsidiary receives benefits from the parent company. In the food and beverages industry the dummy for ownership less than 25% is not significant. This means that there is no empirical evidence that firms that are less than 25% owned obtain benefits from the parent company. There is significant evidence that firms that are more than 25% owned extract benefits. Another important characteristic that may be derived from the dummy variables is that the impact on the real estate industry of being a subsidiary is much higher compared to that in the food and beverages industry. This is more evidence that the real estate industry benefits more from this aspect of networking than the food and beverages industry.
5. SUMMARY AND CONCLUSION
5.1. Conclusions
The effect of the network on the profitability of firms is of growing interest in the current literature. Research is increasingly focusing on many aspects of the network and on its effects on the firms embedded in the network. This study attempts to explore beyond the network, since it compares the profitability of firms with a network relative to those without a network. The results confirm that firms without networks are on average less profitable within an industry. Close consideration is needed in interpreting the results from the first empirical testing. On one hand, the comparison is based on a simple comparison of independent samples using a means statistical technique, which is not sufficient to confirm that networks are crucial for firms to achieve higher profitability. On the other hand the second analysis contributes the robustness of the first analysis. In previous literature it has also been suggested that firms not participating in a network may be barred from new opportunities that arise in the network. This is because the network does not provide information to firms that are not embedded in it. One example may be the study of Westney (1993), which has shown that the R&D subsidiaries of U.S. companies located in Japan have often performed ineffectively because they were locked out of the local networks that tied Japanese R&D labs to suppliers and customers. She suggests that firms should overcome this problem by forming a partnership with a local firm rather than operating alone.
than the marginal costs involved in their managerial efforts. Second, for researchers the effect of the size of the subsidiaries network may add new aspect for research.
The different impact of the network size in the two industries in the current study is also an interesting point. Previous literature, such as Burt (1992), has shown that services industry benefit more from networking than the manufacturing industry. Burt also points out that besides the higher competition in the real estate industries the profit margins remain high. The findings in this study seem complementary with those from Burt because of two main things. Firstly, the real estate industry had significantly larger benefits from networking. This one supports the argument of Burt of the higher profit margins in service industries. Secondly, this study shows that the size of subsidiaries networks in the real estate industry is almost twice as high as the size of subsidiaries networks in the food and beverages industry. This one may be considered as explanatory argument for the higher profit margins, besides the large number of competitors in the real estate industry. More precisely, even though there are many competitors in the real estate industry this study shows that they are interconnected to much higher extend than firms in the food end beverages industry.
However, the effect of the size of the subsidiaries network has not been researched and it may be another aspect to which researchers and managers should pay attention. The difference between the two industries is obvious from the empirical results in this study. Furthermore, networks prove to be beneficial for any firm embedded in the network as a subsidiary. Besides showing strategic managers the importance of the network, this study aims to highlight that the property of network size should be taken into consideration.
5.2 Limitations and further research
investigated in further research into aspects of the size of the network. It may be interesting to study to what level of indirect ownership there is a significant benefit for the focal firm. Similarly, it may be interesting to investigate whether the results are strengthening when attributing higher weight to the subsidiaries that are closer to the focal firm relative to the more distant subsidiaries, which are at more distant levels in the indirect chain of ownership. The question of whether different ownership levels of the subsidiaries influences the benefits of the focal firm should also be investigated.
List of References:
Allen L, Pantzalis C. (1996) “Valuation of the Operating Flexibility of Multinational
Corporations”, Journal of International Business Studies, Vol. 27, 1996
Barney J. (1991). “Firm resources and sustained competitive advantage”, Journal of
Management, 17(1), pp. 99–120.
Bartlett C, Ghoshal S. (1989). “Managing across Borders: The Transnational
Solution”, Harvard Business School Press: Boston, MA.
Bellak C. (2001) “How domestic and foreign firms matter and why does it matter?”
Word count: 11.665
Bernhard L, Philipp JS, John W. (2005), “Volatility Modeling – From ARMA to
ARCH”.
Buhner R. (1987). “Assessing international diversification of West German
corporations”, Strategic Management Journal 8(1): 25–37.
Burt R. (1992). “Structural Holes: The Social Structure of Competition”, Harvard
University Press: Cambridge, MA.
Caves R. (1996). “Multinational Enterprise and Economic Analysis”, (2nd edition). Cambridge University Press: New York.
Chang Y, Thomas H. (1989). “The impact of diversification strategy on risk–return
performance”, Strategic Management Journal 10(3): 271–284.
Collins NR, Preston LE. (1969). “Price–cost margins and industry structure”, Review
of Economics and Statistics 51: 226–242.
Cooper DR, and Schindler PS, (2006), “Business Research Methods”, 9th international
edition, McGraw-Hill, pp. 403-404
Cournot AA. (trans. N. T. Bacon) (1851). “Researches into mathematical principles of
the theory of wealth”, New York: Macmillan
Cubbing J, Geroski P. (1987). “The convergence of profits in the long run: inter-firm
and inter-industry comparisons”, Journal of Industrial Economics 35: 427–442.
Davis G, Greve H. (1997). “Corporate elite networks and governance changes in the
Dess G, Gupta A, Hennart J-F, Hill C. (1995). “Conducting and integrating strategy
research at the international, corporate, and business levels: issues and directions”,
Journal of Management 21: 357–393
Doukas J, Pantzalis C, Kim S. (1999) “Intangible assets and the network structure on
MNC’s” Journal of international financial management and accounting 10:1
Dunning, JH. (1995), “Reappraising the Eclectic Paradigm in An Age Of Alliance
Capitalism”, Journal of International Business Studies, Vol. 26 Issue 3. p 461-491
Dunning, JH (1977), “Trade, Location of Economic Activity and MNE: A search for an
Eclectic Approach.” The international Allocation of Economic Activity. London:
Macmillan, pp. 395-418
Dyer J. (1996). “Specialized supplier networks as a source of competitive advantage:
Evidence from the auto industry”, Strategic Management Journal, 17(4), pp. 271–
291.
Dyer J, Singh H. (1998). “The relational view: Cooperative strategies and sources of
interorganizational competitive advantage“, Academy of Management Review,
23(4), pp. 660–679. 2
Goerzen A, Beamish P. (2005), “The effects of alliance network Diversity on
Multinational Enterprise Performance,” Strategic Management Journal, Strat.
Mgmt. J., 26: 333–354
Granovetter M. (1983), “The strength of weak ties: A network theory revisited”
Sociological Theory, Volume 1, 201-233.
Gulati, R. (1995). “Social structure and alliance formation pattern: A longitudinal
analysis”, Administrative Science Quarterly, 40, pp. 619–642.
Gulati, R. (1999). “Network location and learning: The influence of network resources
and firm capabilities on alliance formation”, Strategic Management Journal, 20(5),
pp. 397–420
Gulati R, Gargiulo M. (1999). “Where do interorganizational networks come from?”
American Journal of Sociology, 104(5), pp. 1439–1493.
Gulati R, Lawrence P. (1999). “Organizing vertical networks: A design perspective”,
Gulati R, Nohria N, Zaheer A. (2000). “Strategic networks”, Strategic Management
Journal, Special Issue 21(3): 203–215.
Hair JF, Anderson RE, Tatham RL, Black WC. (2006), “Multivariate Data Analysis”
Sixth edition, Upper Saddle River, NJ: Pearson Prentice Hall.
Hamel G, Doz Y. L, Prahalad C. K. (1989). “Collaborate with your competitors—and
win”, Harvard Business Review, 67(1), pp. 133–139.
Harbison J, Pekar P. (1998). “Smart Alliances: A Practical Guide to Repeatable
Success”, Jossey-Bass: San Francisco, CA.
Helpman E, Melitz JM, Yeaple RS. (2004), “Export Versus FDI with Heterogeneous
Firms”, the American Economic Review, Vol. 94 NO. 1
Jarillo JC. (1988). “On strategic networks”, Strategic Management Journal, 9(1), pp.
31–41.
Jensen M. (1989). “Eclipse of the public corporation”, Harvard Business Review 67(5):
61–74.
Juster T, Stafford F. (1991). “The allocation of time: empirical findings, behavioral
models, and problems of measurement”, Journal of Economic Literature 29: 471–
522.
Keay, I. (2000), “Canadian manufacturers’ relative productivity performance,
1907-1990”, Canadian Journal of Economics 33 (4), 1049-1068.
Koka B, Prescott J. (2002). “Strategic alliances as social capital: a multidimensional
view”, Strategic Management Journal 23(9): 795–816.
Marshall A. (1920). “Principles of economics”, London Macmillan
Martin S. (1993). “Industrial Economics: Economic Analysis and Public Policy”, (2nd
edition), Prentice-Hall: Englewood Cliffs, NJ.
McEvily B, Zaheer A. (1999). “Bridging ties: A source of firm heterogeneity in
competitive capabilities”, Strategic Management Journal, 20(12), pp. 1133–1158.
Palmer D, Barber B, Zhou X, Soysal Y. (1995). “The friendly and the predatory
acquisition of large U.S. corporations in the 1960s: the other contested terrain”,
American Sociological Review 60: 469–499.
Pennings J, Barkema H, Douma S. (1994). “Organizational learning and
Peteraf MA. (1993). “The cornerstones of competitive advantage: A resource-based
view”, Strategic Management Journal, 14(3), pp. 179–191.
Porter ME. (1980). “Competitive Strategy”, Free Press, New York.
Powell TC, Dent-Micallef A. (1997). “Information technology as competitive
advantage: the role of human, business, and technology resources”, Strategic
Management Journal 18(5): 375–405.
Prahalad C, Hamel G. (1990). “The core competence of the corporation”, Harvard
Business Review 68(3): 79–91.
Rugman AM, Verbeke A. (2001), “Subsidiary-specific advantages in multinational
enterprises” Strategic Management Journal 22:
UNCTAD. (2000). “World Investment Report”, United Nations: New York.1
Uri N. (1988). “A re-examination of the relationship between industry structure and
economic performance”, Applied Economics 20: 1383–1400.
Sapienza HJ, Smith KG, Gannon MJ. (1988). “Using subjective evaluations of
organizational performance in small business research”, American Journal of
Small Business, Winter: 45–53
Spanos Y, Zaralis G. and Lioukas S. (2003). “Strategy and Industry Effects on
Profitability: Evidence from Greece”
Westney DE. (1993). “Cross-Pacific internationalization of R&D by U.S. and Japanese
firms”, R&D Management, 23(2), pp. 171–181.
Wheelwright S, Clark K. (1992). “Revolutionizing Product Development”, Maxwell
MacMillan Canada: Toronto.
White H. (1980) “A heteroskedasticity-consistent covariance matrix estimator and a
direct test for heteroskedasticity”, Economics 48 (1980), pp. 817–838.
Womack JP, Jones DT, Roos D. (1990). “The Machine That Changed World”, Harper
Perennial, New York.
Zaheer A, Venkatraman N. (1995). “Relational governance as an inter-organizational
strategy: An empirical test of the role of trust in economic exchange”, Strategic
