The relationship between diversification
and firm performance in the car
manufacturing industry
By
Stephan Krahl
MSc. International Business and Management
University of Groningen
Student number: S2519232
Date:
20. June 2014
Word count:
18.627
INTRODUCTION 5 1.
THEORETICAL BACKGROUND 8
2.
2.1. Product diversification and performance 10
2.1.1. Resource-based perspective 11
2.1.2. Transaction economies perspective 12
2.1.3. Organizational learning perspective 13
2.1.4. Finance perspective 14
2.1.5. Market power 15
2.1.6. Empirical evidence 15
2.1.1. Theoretical conclusion and hypothesis 17
2.2. Related and unrelated diversification 18
2.3. International diversification and performance 21
2.3.1. Economies of scale 22
2.3.2. Resource-based perspective 23
2.3.1. Transaction cost economies 23
2.3.1. Organizational learning perspective 24
2.3.2. Risk reduction 25
2.3.1. Market power 25
2.3.2. Flexibility 26
2.3.3. Empirical evidence 26
2.3.4. Theoretical conclusion and hypothesis 27
2.4. Technological diversification and performance 27
2.4.1. Theoretical arguments 28
2.4.2. Empirical evidence 29
2.4.3. Theoretical conclusion and hypothesis 29
2.5. Interaction effects between different dimensions of diversification 30
2.6. Summery and assessment of theoretical arguments 32
METHODOLOGY 34
3.
3.1. Data and sample 34
3.4. The Model 41
RESULTS 42
4.
4.1. Diversification and Profitability (model 1 and 2) 42
4.1.1. Model 1 43
4.1.2. Model 2 44
4.1.3. Summary of model 1 and 2 46
4.2. Diversification and Growth (model 3) 46
4.3. Interaction effects 48
4.4. Robustness checks 49
DISCUSSION AND MANAGERIAL IMPLICATIONS 53
5.
5.1. Diversification and profitability 53
5.1. Diversification and growth 56
5.2. Interactions effects 57
CONCLUSION AND LIMITATIONS 59
Abstract
In the field of diversification, there has been a broad amount of contrary studies that often failed to provide a comprehensive framework and a suitable variety of measures to obtain comparable and reliable results. Furthermore, there has been no recent study that focused on the specificities of the automotive industry in terms of diversification. This thesis addresses these research gaps, by including three distinct dimensions of diversification in the car manufacturing industry and by analyzing their impact on performance in terms of both, profitability and growth. The empirical findings show that product diversification does not influence performance, while technological diversification has a negative effect on profitability, but not on growth. Moreover, international diversification was found to be negatively associated with sales growth, but not with profitability. Additionally, it was demonstrated that a combination of international and product diversification also negatively affects sales growth. Since technological diversification does not interact with any other dimension, the hypothesized optimal sequence of first implementing technology-based, then product-based and finally international diversification, was not confirmed. However, another interaction effect between the international distribution of subsidiaries and the international distribution of sales was identified and proven to negatively affect sales growth.
Acknowledgement
Introduction
1.
Diversification is an important part of international business strategy, industrial organization and financial management, especially when it comes to multinational enterprises (Lampel & Giachetti 2013; Purkayastha, Manolova & Edelman 2012; Geringer et. al. 1989) Therefore it appears to be crucial to understand the relationship between diversification and firm performance (Wiersema & Bowen 2011). Although there has been a wide range of literature examining the impact of diversification on firm performance neither the empirical results nor the theoretical discussion are conclusive (Bausch & Pils 2009; Palich, Cardinal & Miller 2000; Datta et. al. 1991; Vasudevan & Varadarajan 1989). One reason for this confusion could be the great number of other factors that are suggested to moderate the relationship between diversification and firm performance. Some of them are R&D-intensity (Garcia-Vega 2006; Stimpert & Duhaime 1997), firm size (Chang & Wang 2007; Lang & Stulz 1993), IT-intensity (Ravichandran et. al. 2009), capital-intensity (Bettis 1981), institutional context (Purkayastha, Manolova & Edelman 2012; Geringer, Tallman & Olsen 2000) and market structure (Adner & Zemsky 2012; Christensen & Montgomery 1981). In order to reduce this complexity and the impact of unknown or not measurable factors it appears to be logical to choose a homogenous population of firms, which operate in the same market and have access to similar resources. This argument is supported by Datta et. al. (1991), who stated that the inconclusive findings may be caused by the cross-sectional data sets, which were mostly used in previous studies. For that reason, this thesis examines the impact of diversification on performance specifically within the car manufacturing industry. Furthermore, the empirical and theoretical analysis aims to provide a broad evaluation of diversification strategies in this context. This shall help managers and researcher to better understand the effects of diversification on automotive firms and the potential influence of other related factors.
(Belis-Bergouignan et. al. 2010). Therefore automotive firms appear to be a good population to observe differences in diversification strategies, while facing a low risk of bias caused by heterogeneity. Lampel & Giachetti (2013) already used a very similar sample to examine the impact of international diversification of manufacturing operations on firm performance in the automotive industry. However this paper focused only on the diversity of manufacturing activities across geographic regions and did not capture diversification in terms of international sales, business segments or R&D. Therefore, the objective of this thesis is to contribute to the previous literature by developing and testing a more comprehensive model of diversification, specifically tailored to the car industry.
One reason why it is interesting to investigate this field of research within the context of car manufacturers is demonstrated in the case study of BMW by Gassmann et. al. (2010). The paper showed the benefits, which car manufacturers can gain from searching for technologies beyond industry boundaries and thereby illustrated the relevance of diversification in the automotive industry. When BMW decided to develop a new computer –based car control concept (later named iDrive) in the late 1990s it appeared to be that the required technology could neither be found within the organization itself nor within the automotive supplier network. Therefore BMW started searching for a partner outside the industry and found a small Silicon Valley based company, which was focused on joysticks and medical equipment. This horizontal exploration strategy of BMW paid off and lead to a successful cross-industry alliance. Referring to this case study, it can be argued that technological diversity across industries triggers radical innovations and thereby not just prevents negative lock-in effects, but also improves firm performance.
diversification is specifically important in this thesis, since innovativeness is essential to gain a competitive advantage in the automotive industry (Ili, Albers & Miller 2010). The recent trends in electric vehicle technology (Wells & Nieuwenhuis 2012) and innovations in electronic control systems (Gassmann et. al. 2010) are further examples that confirm this statement. In other words, it appears to be specifically suitable to consider diversity across technological fields additionally to diversity across product segments and geographic segments, in order to capture diversification in the car manufacturing industry. In order to address this literature gap, this thesis provides a broader model of diversification, which can detect further potential interaction effects among the three dimensions of diversification.
However, it appears to be ambiguous how to measure product and international diversification. While previous studies often focused on either the concentration of sales or direct investments (Lampel & Giachetti 2013; Shaver 2011; Datta et. al.), this thesis considers both perspectives by developing two distinct measures for both product and international diversity. Therewith this thesis follows the approach of Oh & Rugman (2012) towards different global strategies. Thus, the empirical model of this thesis is not just able to detect interaction effects among different dimensions of diversification, but also between sales-based and direct investment-based diversification within one dimension. Furthermore, a dummy variable for brand diversity was developed, in order to capture this specific characteristic of some automotive groups. For example, Volkswagen covers multiple types of vehicles by owning MAN, Volkswagen, Skoda, Lamborghini, Audi, SEAT, Bentley, Bugatti, Scania and Volkswagen Commercial vehicle (MarketLine Industry Profil: Car Manufacturing 2013). This raises the question whether automotive corporations with several brands achieve higher performance than single brand firms.
evaluate current diversification strategies within this context and provides a basis for further implications for decision-making. Moreover this approach follows the suggestion of Purkayastha, Manolova & Edelman (2012), who argued that since the relationship between diversification and performance highly depends on the context in which the research is conducted, empirical studies should focus on one industry to obtain more reliable results. Aside from the empirical contribution of this thesis, the theoretical section gives a detailed overview of arguments from different perspectives related to firm-level diversification. These aspects are critically evaluated and applied to the specific context of the car manufacturing industry. This theoretical contribution aims to provide a better understanding of the empirical results and helps managers and researcher to evaluate different diversification strategies within the car manufacturing industry.
In section 2.1 to 2.4, previous literature is reviewed and a hypothesis for each dimension of diversification and its relationship to performance is defined. In order to reduce the scope and complexity of this thesis, the hypotheses only relate to the three diversification dimensions and performance and are not specified for each single measure. Subsequently, in section 2.5, two hypotheses for potential interaction effects between different dimensions of diversity are added. Furthermore, section 3 explains the methodology before the results of the empirical analysis are illustrated in section 4. Consequently, section 5 discusses the findings of this thesis and outlines managerial implications. Lastly, the conclusions as well as the limitations of this study are drawn and implications for further literature are suggested.
Theoretical background
2.
the type of diversification strategy and the mode of diversification as three distinct research directions and suggested a combination of these approaches for future literature.
technological diversity, since companies in the car manufacturing industry face high pressure to innovate due to demanding customers, intense competition and environmental protection guidelines (Ili, Albers & Miller 2010). In order to address these industry-specific challenges Gassmann et. al. (2010) suggested that car makers should try to acquire knowledge in non-traditional networks of R&D suppliers, which do not particularly belong to the automotive industry. Another argument to include technological diversification as a third dimension along with product and international diversification are potential interactions between international and technological portfolios as a result of the recent trend of electric vehicles. Wells & Nieuwenhuis (2012) found that regionally diverging demand and differential opportunities lead to distinct regional development strategies of electric car technology, which suggests that internationalization might require technological diversity. Furthermore it was demonstrated that the recent technological trends in the automotive industry demand a diverse patent portfolio, emphasizing the importance of this dimension (Oltra & Saint Jean 2009). Therefore it seems to be suitable to examine the effect of diversification on performance in the automotive industry based on a technological, an international and a product portfolio dimension.
In section 2.1-2.4 the theoretical and empirical link between the three dimensions of diversification and firm performance, as suggested by previous literature, is discussed. Subsequently, the effects of interaction terms between different dimensions of diversification toward performance are examined from a theoretical point of view. Lastly, each argument is critically assessed in the context of the car manufacturing industry, in order to develop a specific theoretical framework for this thesis.
2.1.Product diversification and performance
The theoretical connection between product diversification and firm performance is mostly explained by the resource-based perspective (Geringer, Tallman & Olsen 2000; Tallman 1996) and transaction cost economies (Teece 1986; Tallman 1996, Jones & Hill 1988). Additionally, the organizational learning perspective (Prahalad & Bettis 1986) and finance literature (Lang & Stulz 1993; Williamson 1975; Higgins & Schall 1975; Lawellen 1971) are both related to this field.
2.1.1. Resource-based perspective
of a diversified product portfolio is only feasible when all business segments can be managed under a common strategic paradigm.
Since the sample of this thesis consists only of publicly listed automotive companies it can be assumed that the managerial capabilities of each firm are sufficient to efficiently allocate resources across business units. Furthermore it is stated that synergies are particularly strong when intangible assets can be shared among product segments (Allee 2008; Chang 1995). This aspect indicates that the resource-based argument plays an important role in the case of this thesis, since supplier networks and brand names highly influence performance in the car manufacturing industry (Dyer & Nobeoak 2002; Corsten, Gruen & Peyinghaus 2011; MarketLine Industry Profil: Automotive Manufacturing 2013).
2.1.2. Transaction economies perspective
diversification risky. Generally, the previous literature agrees that coordinating costs negatively impact the relationship between product diversification and firm performance (Ravichandran et. al. 2009; Palich, Cardinal & Miller 2000). Therefore, Rumelt (1982) detected a balance between economies of scope and diseconomies of organizational scale, which combines the resource-based and the transaction cost economies perspective. The empirical research of Grant, Jammine & Thomas (1988) supports this argument by demonstrating that there is a positive relationship between product diversity and profitability, which eventually becomes negative, when complexity reaches a certain level. In the car making industry, coordinating costs between different business units can be expected to be low due to the high independence among different entities. This observation can be illustrated by the cases of Toyota and Volkswagen, which both embody multiple autonomous brands (MarketLine Industry Profil: Car Manufacturing 2013). However, this high degree of independence might also lead to interest divergence and thereby cause agency costs (Nayyar 1992). Another characteristic of the car making industry is the high complexity of its supply chain, which was shown to increase transaction costs (Choi & Krause 2006). From these observations it can be reasoned that transaction costs economies have a greatly negative impact on the relationship between product diversification and firm performance in the context of this thesis.
2.1.3. Organizational learning perspective
(1996), who illustrated the dynamic learning possibilities of entries into new business segments.
However due to the high independence of business unites in the car manufacturing industry (MarketLine Industry Profil: Car Manufacturing 2013), learning spillovers can be expected to be relatively low. On the other side, the overlap with different other industries, like electronics or Information Technology (Gassmann et. al. 2010), might indicate a high degree of relatedness among several diversified activities. According to multiple studies this factor triggers a positive relationship between product diversification and performance (Boschma et. al. 2012; Ravichandran et. al. 2009; Chang & Wang 2007; Dubofsky 1987). Therefore, a small, but considerable impact of the organization learning perspective can be predicted in the case of this thesis.
2.1.4. Finance perspective
Moreover, there are several scholars, which apply a finance perspective to explain the relationship between product diversification and firm performance (Singh, Mathur & Gleason 2004; Lawellen 1971; Higgins & Schall 1975). It is argued that internal markets for capital, human and other resources are more efficient than external markets, because they can exploit superior information flows and controls (Williamson 1992, Servaes 1996). Thus, diversified firms have better opportunities to allocate their capital among business segments, in order to exploit temporary growth opportunities, which may provide them an advantage over single-product firms (Lang & Stulz 1993). Furthermore the finance literature argues that firms can reduce their risk by investing into business activities, which are not correlated to their other businesses in terms of cash stream and stability (Lawellen 1971, Higgins & Schall 1975). This effect might provide diversified organizations an enhanced debt capacity, which enables them to build a greater tax shield and thereby increase the company`s value (Singh, Mathur & Gleason 2004). This proposition is also supported by Rumelt (1982), who states that portfolio diversification can decrease non-systematic risk and thereby reduce capital costs.
shown in appendix B. Therefore, a high correlation between these activities can be assumed, which leads to a low reduction of systematic risk. Thus, the financial perspective is most likely not contributing to the benefits of product diversification in the context of this analysis. However, a small effect can be expected for the measure of brand diversification, since multi-brand corporations do not rely heavily on one specific group of customers and thereby are less affected by demand fluctuation (Morgan, Lopo & Lego 2009).
2.1.5. Market power
Another argument that indicates a positive relationship between product diversification and firm performance is made by Caves (1981). It is argued that diversified organizations have better opportunities to exploit economies of scale, cross-subsidization, predatory pricing, reciprocity in buying and selling, and creating entry barriers. This enables them to gain larger market shares and higher market power compared to not-diversified firms. However, the bargaining power of suppliers and buyers in the car manufacturing industry, as well as the threat of new entrants and substitutes is suggested to be moderate (MarketLine Industry Profil: Car Manufacturing 2013), which indicates a low importance of market power in the case of the car making industry.
2.1.6. Empirical evidence
also illustrate the importance of interaction effects between different types of diversification. Furthermore this curvilinear relationship is in line with the findings of Grant, Jammine & Thomas (1988).
On the other side there has been empirical evidence that product diversification may negatively impact firm performance. For example the study of Tongli, Ping & Chee Chiu (2005) found that the measures for return on assets, share price and Tobin`s Q of Singapore based firms are all negatively influenced by the degree of diversification across different product segments. Another paper by Stimpert & Duhaime (1997) stated that product diversification reduces R&D spendings, which consequently causes decreasing capital investments and business-unit effectiveness. Although these results do not necessarily indicate a negative impact on performance, the variables certainly impact the competitive positions of the company. The empirical analysis of Amit & Livnat (1988) also found that product diversification is negatively associated with profitability. However, this study additionally showed that diversified firms achieve more stable cash flows and an increased level of leverage, which indicates lower operating risk.
diversification strategy puzzling. This shows how difficult it is to estimate the relationship between diversification and performance, despite of the broad amount of research.
Another issue of previous studies is the high complexity of strategic decision making. For example, changes in a firm`s diversification strategy might for example be facilitated by opportunities or uncertainty (Hoskisson & Hitt 1990). Another suggestion is that companies might face a tradeoff between product diversification and other strategic approaches (Geringer, Beamish & d Costa 1989). Moreover, it is stated by Hoskisson & Hitt (1990) that adopting a diversification strategy is often based on a firm`s resource endowment and different unknown incentives.
This thesis aims to reduce these potential problems through the very homogenous selection of firms. This strategy furthermore justifies that no control variables or moderators are included in the model. It can be expected that most of these factors suggested by previous literature, like capital intensity (Bettis 1981; Datta et. al. 1991), firm size (Elango, Ma & Pope 2009; Lang & Stulz 1993) and R&D intensity (Chang & Wang 2007; Ravichandran et al. 2009) are widely homogenous across the sample of this thesis. Nevertheless, the robustness of the models against these assumptions is tested in section 4.4. Another contribution of this thesis is the distinction between three different measures to capture the dimension of product diversification. This addresses the measurement issue and the lack of comparability among previous studies (Bausch & Pils 2009; Datta et al. 1993). While the first measure refers to the distribution of direct investment across different industry, the second measure is an indication for the distribution of sales across different industries. Additionally, a dummy variable is included to capture whether the firm sell vehicles under one or multiple brand names.
2.1.1. Theoretical conclusion and hypothesis
Zenka 2011), which means that single company`s behavior will not be affected greatly by home country-specific characteristics.
Combining these different theoretical aspects and perspectives, it remains ambiguous whether product diversification has a negative or a positive impact on firm performance. From the literature review and the specific characteristics of the car manufacturing industry, it can be concluded that economies of scope (Tanriverdi & Venkatraman 2005; Nayyar 1992) are the main benefits associated with product diversification in this context, while the advantages of market power, resource allocation (Prahalad & Bettis 1986) and learning spillovers (Ruigrok & Wagner 2003) only play a moderate role. However, it was also emphasized that transaction costs are expected to equalize most of these gains (Ravichandran et. al. 2009; Jones & Hill 1988). The importance of each theoretical argument in the context of this thesis is further assessed in section 2.6 and in appendix A. In conclusion, it is hypothesized that product diversification has a weakly positive effect on performance in the case of this thesis
H1: Product diversification positively affects firm performance in the car manufacturing
industry
2.2.Related and unrelated diversification
suggestion is in line with Jones & Hill (1988), who explained that related businesses can realize economies of scope by sharing inputs. However, it is also stated that this diversification strategy will not reduce total risk, since the earnings of both business units are correlated (Michel & Shaked 1984). Furthermore it remains unclear how and what components of the products have to be related, in order to generate benefits. While several studies refer to manufacturing relatedness (Farjoun 1998), R&D relatedness or technological relatedness (Cassiman et. al. 2005), Tanriverdi (2005) argued that cross-business knowledge about products, customers and management practices is the crucial factor for sub-additive values. This suggestion is in line with the findings of Neffke & Henning (2013), which show that firms prefer to diversify into skill-related industries due to the importance of human capital. Another important aspect of related diversification strategies are the costs of coordination and motivation caused by information asymmetries and interest divergence, which often make its implementation difficult (Nayyar 1992). Unrelated diversification, on the other hand, is unlikely to produce synergies, but may significantly reduce a company`s total risk (Michel & Shaked 1984). This statement is supported by Jones & Hill (1988), who further explained that this strategy leads to lower capital costs and more efficient capital allocation. Additionally, it is stated by Pehrsson (2006) that unrelated diversification can produce greater organizational learning outcomes, while it is s also associated with high costs of organization due to the complexity of portfolio management. Thus, theoretical arguments are not conclusive on whether related or unrelated diversification leads to higher performance.
suggested that a moderate level of relatedness facilitates the best performance outcomes. (Pehrsson 2006; Palich, Cardinal & Miller 2000)
Another problem that emerges especially with the approach of related and unrelated diversification is the ambiguous causality between diversity and performance (Bausch and Pils 2009; Grant, Jammine & Thomas 1988; Rumelt 1974). It is stated that firms with current poor performance rather tend to engage in unrelated business activities in order to exploit opportunities in other industries with better profit perspectives. Highly profitable companies, on the other hand, have an incentive to further invest in activities within their current industry instead of diversifying into unrelated and eventually less profitable segments (Lang & Stulz 1993; Gort, Grabowski & McGuckin 1985). This suggestion is in line with the findings of Christensen & Montgomery (1981), which indicate that firms operating in markets with low-growth opportunities prefer to invest into unrelated markets. Furthermore, Park (2002) provided empirical evidence that firms within highly profitable industries rather tend to choose acquisitions in related business fields.
These theoretical and empirical arguments demonstrate how problematic the approach of unrelated and related diversification strategies is. This is why this thesis measures diversification without this distinction or Rumelt’s (1974, 1982) categorization. Another reason for this decision is that Kim, Hwang & Burgers (1993) showed that both related and unrelated diversification is associated with a superior return-risk-ratio, which indicates that both strategies have a similar effect. Moreover, Nayyar (1992) explained that it is very difficult to distinguish between related and unrelated diversification accurately, despite of the broad amount of studies. In order to avoid such problems this thesis focuses only on the degree of diversification without any categorization of considering relatedness.
Car Manufacturing 2013), a distinction between related and unrelated diversification would not make sense in the context of this thesis. Furthermore, it has to be noted that it is likely that the product diversification, measured in this thesis, mostly consists of related diversification. The reason is the sample selection, which exclusively contains automotive companies and no large groups or parent companies. The Tata group, for example, is highly unrelated diversified into several industries (Khanna, Palepu, & Bullock 2008; Khanna & Palepu 1997). However, with Tata motors, only the automotive affiliate of the group is included into the sample, in order to prevent the industry-specific performance effects of other affiliates of the group.
It can be concluded that the theoretical arguments do not differ between related and unrelated diversification, but the magnitude of each single aspects might vary. When moving from related to unrelated diversification, the benefits of synergies decrease, while opportunities for risk reduction and more efficient resource allocation arise (Tanriverdi 2005; Jones & Hill 1988; Michel & Shaked 1984). Thus, it is plausible that unrelated diversification is emphasized in developing environments, in order to equalize high uncertainty, while related diversification is emphasized in developed environments where firms face more intensive competition (Purkayastha, Manolova & Edelman 2012; Wright et. al. 2005; Khanna & Palepu 1997). However, it appears to be reasonable for this thesis to neglect this distinction, since the analysis is conduct within a very globalized industry (MarketLine Industry Profil: Car Manufacturing 2013; Zapata & Nieuwenhuis 2010) and the sample exclusively consists of firms operating within the automotive industry.
2.3.International diversification and performance
by measuring international diversification with the geographic distribution of sales as well as with the geographic distribution of subsidiaries. Moreover, it appears to be essential for theory and praxis to examine the way in which internalization strategies impact firm-level performance. The reviews of previous literature by Hennart (2011), Glaum & Oesterle (2007) and Wiersema & Bowen (2011) illustrated the contrary results of conceptual as well as empirical papers. Furthermore these studies indicate that the findings about the relationship between international diversification and firm performance depend on the firm`s context and on how the research is conducted. Therefore, it is also very difficult to predict the outcomes of international diversification in the car manufacturing, which makes this dimension an interesting part of this thesis. In the following sections, the main benefits and drawbacks of international diversification are summarized and applied to the context of the car manufacturing industry.
2.3.1. Economies of scale
might in fact benefit from additional economies of scale when they expand internationally. Furthermore, the geographic sales distribution in the automotive industry changes quickly due to the emergence of new markets (Solvell 2012; Pavlinek & Zenka 2011; McKinsey&Company: Advanced Industries 2013), which indicates that opportunities for economics of scale rather lie in global markets than in domestic demand. Moreover it can be argued that the car manufacturing industry is globally convergent and thereby provides good conditions to realize cross-national economies of scale. This conclusion can be drawn from recent trends of modularity (Ro, Liker & Fixson 2007; Pandremenos, et. al. 2009), which require a certain degree of product standardization (Jacobs, Vickery & Droge 2007). From these arguments it can be reasoned that economies of scale are important benefits of international diversification in the automotive industry.
2.3.2. Resource-based perspective
Similarly, it can be stated from a resource-based perspective that Multinationals can leverage firm-specific resources across geographic regions (Wiersema & Bowen 2011; Hitt, Hoskisson & Kim 1997; Teece 1982). For instance, Porter (1985) argued that international scope enables organizations to exploit interrelatedness between different areas. In general, intangible assets, like brand names and networks, can easier be leveraged across different markets than tangible assets, because they do not depreciate in value when they are transferred (Allee 2008; Chang 1995). Therefore, intangible assets provide great synergy opportunities to international diversified organizations (Purkayastha, Manolova & Edelman 2012). Since brands and networks appear to be a very important competitive factors in the car manufacturing industry (MarketLine Industry Profil: Car Manufacturing 2013; Sturgeon, Van Biesebroeck & Gereffi 2008), this aspect of international diversification might be an essential argument in the context of this thesis. It can be argued that these strategic opportunities of international diversified corporations give them a competitive advantage over domestic firms and ultimately lead to superior performance (Geringer, Tallman & Olsen 2000; Hitt, Hoskisson & Kim 1997).
2.3.1. Transaction cost economies
the proposition that governance costs increase when companies diversify geographically, since international scope requires more information processing and a larger administrative organization (Wiersema & Bowen 2011; Geringer, Tallman & Olsen 2000; Chang 1995). For instance, Porter (1985) stated that due to its high complexity international diversification results in increasing coordination costs caused by distribution costs, trade barriers, logistical costs, cultural diversity and country differences. In the case of the car manufacturing industry, trade barriers and local subsidization are essential, as documented in the example of China (Gan 2003). This indicates a high relevance of transaction cost economics in the context of this thesis.
Another transaction cost related argument is based on the liability of newness, which emerges when firms conduct foreign direct investments. It refers to costs, associated with installing new facilities, staffing and the implementation of management and network systems (Lu & Beamish 2004). Contrary to these suggestions it can be argued that internationalization reduces transaction costs, since the coordination of activities and resources can be more efficiently conducted within a hierarchical organization than via the market (Hennart 2011; Chang 1995).
2.3.1. Organizational learning perspective
2.3.2. Risk reduction
Kim, Hwang & Burgers (1993) showed empirically that another benefit of international diversification is the possibility of total corporate risk reduction. These findings are in line with Hisey & Caves (1985), who identified the possibility of risk-spreading as one of the main motives to adopt an international diversification strategy. The advantage of risk reduction has further been pointed out by Wiersema & Bowen (2011) and Geringer, Tallman & Olsen (2000). The literature provides three explanations for this connection: first, internationally diversified firms can better counter aggressive actions of competitors. Second, multinationality helps companies to respond to changes in countries’ interest rates, wage rates and factor prices. And third, global market diversification reduces the impact of supply and demand fluctuation on a firm`s cash flows (Kim, Hwang & Burgers 1993). However, the aspect of total corporate risk reduction appears to be of minor importance for this thesis, since risk is not included as a dependent variable. However, a possible connection to firm performance can be drawn from a financial perspective by reasoning that lower corporate risk is associated with lower funding costs (Stulz 1999, Grosse 1992) and thereby gives internationally diversified companies a competitive advantage. Since the automotive industry is highly correlated with economic development, it could be argued that car makers greatly benefit from diversification across geographic regions, because this strategy reduces their dependence on single economies and makes them less vulnerable to demand fluctuations.
2.3.1. Market power
2.3.2. Flexibility
Another aspect, which indicates a positive relationship between international diversification and performance, is the flexibility that comes with multinationality. This argument is based on the work of Kogut (2002; 1985), who pointed out that companies, which operate in multiple markets, can exploit real exchange rate variances and arbitrage opportunities. This refers to allocating production activities, shifting taxable income (Taylor & Richardson 2013) and exploiting imperfect financial markets (Grosse 1992) and information advantages. This proposition was also supported by Glaum & Oesterle (2007), who explained that market imperfection might result from economic, political, legal and cultural differences among countries and provides international organization arbitrage opportunities. Williamson (1975) even argued that internationalization prevents the misallocation of resources.
2.3.3. Empirical evidence
existing literature by including two different measures for international diversification, which cover both, geographic sales distribution and foreign direct investments.
2.3.4. Theoretical conclusion and hypothesis
From a theoretical point of view it was shown that the main benefits of international diversification for automotive firms lie in the opportunities of leveraging market power and intangible assets (Allee 2008; Crocioni 2008; Tallman & Li 1996). However, it appears to be likely that these gains are exceeded by the increasing transaction costs that are associated with geographic scope (Wiersema & Bowen 2011; Hennart 2011). For this reason a small negative effect of international diversification on firm performance can be expected in the case of this thesis. A further illustration of all theoretical arguments associated with this dimension is presented in appendix A.
H2: International diversification negatively affects firm performance in the car
manufacturing industry
2.4.Technological diversification and performance
industry (Ettlie & Pavlou 2006; Wells & Nieuwenhuis 2012), it can be reasoned that technological diversification is an important dimension in the case of this thesis.
2.4.1. Theoretical arguments
technological diversification ultimately improves performance by enhancing innovation outcomes.
Because of the high level innovativeness in the automotive industry (Ili, Albers & Miller 2010) it can be assumed that technological spillovers within the patent portfolio as well as R&D-efficiency and risk might be important factors for competitive advantages in the case of this thesis.
2.4.2. Empirical evidence
Miller (2006), for instance found a positive correlated between technology-based diversification and market-based firm performance, while controlling for capital-intensity, R&D-intensity and endogeneity effects. Leten, Belderbos & Van Looy (2007), on the other hand identified an inverted U-shaped relationship between technological diversification and technological performance. This finding is in line with the results of the study of Huang & Chen (2010), which also indicates an inverted U-shaped connection between technological diversity and innovation performance. This relationship can be explained by increasing coordination and integration costs, which occur when firms spread their R&D investments across different technical fields. Another empirical contribution has been made by Lin, Chen & Wu (2006), who suggested that technology-based firms should focus on one specific technological field and concentrate their R&D activities within their core business activities. From these partly contrary results it can be concluded that the relationship between technological diversification and firm performance might depend on the firms` context or additional unknown factors, which makes it difficult to predict the outcomes of the analysis with regard to the patent distribution variable.
2.4.3. Theoretical conclusion and hypothesis
Breschi, Lissoni & Malerba 2003) are predicted to overwhelm the additional coordinating costs in the case of the automotive industry, a positive link to firm performance is hypothesized for this variable. The relevance of the theoretical arguments related to this dimensions are further illustrated in appendix A.
H3: Technological diversification positively affects firm performance in the car
manufacturing industry
2.5.Interaction effects between different dimensions of diversification
When comparing the theoretical arguments, which explain the relationship between the three different dimensions of diversification and firm-level performance, several similarities occur (see appendix A). This indicates that the costs and benefits of different dimensions of diversification are to some extend interrelated to each other. This suggestion is in line with several papers illustrating moderating and interaction effect in this field of research. (Geringer, Tallman & Olsen 2000; Hitt, Hoskisson & Kim 1997; Kumar 2009; Garcia-Vega 2005)
Therefore, a positive moderating effect of product diversification when examining the relationship between international diversification and performance can be expected to be discovered in the empirical analysis.
The interrelatedness between technological and product diversification is mostly described from a theoretical point of view. For example, Lin, Chen & Wu (2006) stated that a diversified R&D portfolio enables firms to develop a wider range of products. This aspect is in line with Granstrand (1998), who explained that technological diversification can be considered as input-based, while business diversification refers to the output of R&D. From this assumption it can be concluded that the ability to become a multiproduct firm is to some extend determined by the possession of diversified technological input factors. Similarly, Garcia-Vega (2005) pointed out that technology-based diversification and innovations positively affect product diversity. It can be argued that these aspects also apply in the case of the car manufacturing industry by referring to the development of different engine technologies. The study of Chan (2007) stated that firms, which engaged in different types of engine technologies, like electrical engine, gasoline engine, alternative fuel engine and diesel engine were most successful in developing hybrid vehicles. This enables car makers to diversify their product portfolio in terms of energy sources, which appears to be of increasing importance in this industry (Tate, Harpster & Savagian 2008). Thus, it can be concluded that technological diversification can provide car manufacturing firms great opportunities to diversify their product portfolio. From this proposition it can be hypothesized that the interaction effect of technological and product diversification positively affects firm performance.
H4: The interaction term between technological and product diversification positively
affects firm performance in the car manufacturing industry
H5: The interaction term between product and international diversification positively
affects firm performance in the car manufacturing industry
of distinct products and to engage in different industry activities (Garcia-Vega 2005). Furthermore, product diversification in terms of different brands and affiliate activities supports the internationalization process and thereby positively affects geographic diversification (Becker-Ritterspach & Bruche 2012; Khanna, Palepu, & Bullock 2008). By considering these three dimensions of diversification this thesis found that there might be a preferable strategic order of implementing different types of diversification in the car manufacturing industry. International diversification is expected to be more valuable when the firm previously obtained a diverse product portfolio and product diversification is predicted to produce better performance outcomes when the firm previously obtained a diverse technological portfolio. Therefore, the suggested sequence of strategy implementation would be to first create a broad technological portfolio, then follow a product diversification strategy and lastly expand to diverse geographic regions. The successful implementation of each step supports the subsequent step.
2.6.Summery and assessment of theoretical arguments
In section 2.1 to 2.5 the theoretical arguments, explaining the relationship between the three dimensions of diversification and firm performance, are described and applied to the context of this thesis. This section shortly summarizes this theoretical contribution, which is further illustrated in appendix A. The consequently emerging hypotheses are summarized in Figure 1.
Figure 1: Conceptual model and Hypothesis
International diversification benefits are mostly driven by the ability of multinational car manufacturers to leverage intangible asset and to exploit market power (Allee 2008; Crocioni 2008). The role of economies of scale (Hult 2011), resource flexibility (Kogut 2002) and risk reduction (Kim, Hwang & Burgers 1993) is of minor importance in the context of this thesis, while organizational learning spillovers (Zahra, Ireland & Hitt 2000) appear to be completely irrelevant. Despite the great number of advantages of geographic diversification, the negative impact of transaction cost appears to be the most important argument, as shown in appendix A. Therefore a weak negative connection between international diversification and performance is predicted.
Furthermore, it is hypothesized in section 2.5 that the interaction effects between technological diversity and product diversification, as well between product diversity and international diversification positively influence firm performance (Becker-Ritterspach & Bruche 2012; Khanna, Palepu, & Bullock 2008; Lin, Chen & Wu 2006). Since there has been no recent study that considers all three dimensions of diversification simultaneously, additional relevant interactions that have not been detected by previous research might occur in this thesis. Therefore, the analysis section tests for further possible interaction term.
Methodology
3.
This thesis aims to measure the relationship between diversification and firm performance within the car manufacturing industry. In order to obtain more reliable and detailed results, three dimensions for diversification with individual measures were distinguished. In the following sections it is described how the data was collected and under what criteria the companies were selected for the sample. Afterwards, it is illustrated how exactly each dimension is captured and which method was chosen to calculate diversification. Instead of combining all dimensions, its effects on firm performance were examined individually through a linear multiple-regression analysis, in order to discover potential interaction effects and to facilitate more detailed results. This method is in line with the main literature streams (Kumar 2009; Geringer, Tallman & Olsen 2000; Chen, Yang & Lin 2013). In section 3.3, it is explained how firm performance, which is captured by three dependent variables, is defined and measured for the analysis.
In order to understand the choice of this methodological approach, it has to be noted that this thesis aims to go beyond the defined hypotheses by testing for unpredicted interaction effects and multiple distinct measures for both diversification and performance.
3.1.Data and sample
and performance (Lampel & Giachetti 2013; Chakrabarti, Singh & Mahmood 2007). First, all companies with the US Standard Industry Classification code 3711, which stands for “motor vehicles and passenger car bodies” were filtered. This method to create industry-specific samples was also applied in previous studies (Elango & Sambharya 2004). Secondly, only very large sized, publicly listed and active firms with recent available detailed financial data were selected. Lastly several companies, which appeared to be automotive suppliers instead of car manufacturers had to be deleted from the list. Ultimately, this method creates a very homogenous sample of directly competing firms and thereby eliminates potential industry-specific effects. Additionally, a similar population was already used by Lampel & Giachetti (2013) to examine the relationship between diversification of international manufacturing operations and firm performance in the car making industry. Oltra & Jean (2009) is another study that focused on the car manufacturing industry by choosing eleven companies that overlap with the sample of this thesis. This provides further justification for this very specific sample selection. One aspect that makes these firms very interesting in terms of diversification is the differences in ownership and corporate structure, as well as the diverse product lines across global car companies (Lampel & Giachetti 2013). For example, Volkswagen operates in four business segments: passenger cars and light commercial vehicles, financial services, trucks and buses and power engineering. Especially financial services are a business segment that many car makers include in their portfolio (MarketLine Industry Profil: Car Manufacturing 2013). Ownership and corporate structure refer to the multiplicity of brands that some automotive groups embody. Often, a brand is represented by a highly autonomous business unit or company that independently decides about products and market activities (Lampel & Giachetti 2013). This consequently leads to a high degree of product and international diversification of some automotive corporations and brings up the questions how these differences impact firm performance. Thus, it can be concluded that the sample of this thesis appears to be a highly relevant to the field of diversification research.
difference between the measures of the independent and the dependent variables. However, it is stated in the literature that diversification is not likely to change over a small amount of time (Leten, Belderbos & Van Looy 2007; Garcia-Vega 2005), which reduces this methodological issue and validates the approach.
3.2.Independent variables
All independent variables are measures of the three dimensions of diversification. Product and international diversity are both measured by a sales-based and a subsidiary-based variable, while technological diversification is only measured with patent data. The sales-based measures capture the concentration of a company`s sales across different product segments or geographic regions, while the subsidiary-based measures refer to the concentration of subsidiaries` turnover across different industries or countries. Additionally, the thesis introduces a brand-based dummy variable, in order to capture this specific type of product diversification within the car industry. A list of all measures, including short descriptions and abbreviations is provided in appendix C.
Previous literature suggested two calculation formulas to measure diversification: the Herfindahl-index and the entropie-measure (Bauch & Pils 2009). In the case of product diversification, stands for the share of either sales or turnover within the business or product segment i. For international diversification, equals the share of sales or turnover within the geographic region i. And for technological diversification, is the share of the number of patents that a firm owns within the patent class i (Jacquemin & Berry 1979). The number of segments or classes is always captured in n. While the Herfindahl-Index decreases in value when the degree of diversification becomes higher, the Entropy measure increases with lower levels of concentration. In the case of only one segment or class, which is the highest possible concentration, the Herfindahl-Index is equal to 1, while the Entropy measure is equal to 0. When the concentration decreases the entropy measure rises, while the Herfindahl-Index becomes smaller (Acar & Sankaran 1999).
Herfindahl-Index: H = ∑
Both of these concentration measures are objective and highly applicable to capture industry specialization (Acar & Sankaran 1999). However, many studies prefer the entropy-measure and argue that it might be more decomposable, which enables the researcher to better distinguish between related and unrelated diversification (Berry 1972; Palepu 1985). Since it was proved by Acar & Sankaran (1999) that the Herfindahl-Index is also decomposable and this thesis does not distinguish between related and unrelated diversification, this argument seems to be irrelevant. Another criterion to choose between these methods is the comparability among the different measures. Since the Herfindahl-Index can only adopt values between 0 and 1, while the entropy-measure can theoretically grow to a very large number, it appears to be more suitable to select the Herfindahl-Index in the context of this thesis (Acar & Sankaran 1999). This approach, furthermore, finds support in previous diversification literature (Elango, Ma & Pope 2008; Lin, Chen & Wu 2006; Singh, Mathur, Gleason 2004; Bühner 1987).
3.2.1. Product diversification
The dimension of product diversification refers to the distribution of a firm`s activities across different related or unrelated business segments (Palepu 1985; Rumelt 1974). As demonstrated in the literature review, this dimension is highly complex and there has been no generally accepted approach to conceptualize it (Bausch & Pils 2009). For this reason, this thesis distinguishes between three different measures of product diversification to build a reliable and comprehensive model.
(Bausch & Pils 2009; Datte et al. 1991). As stated in appendix C, this measure is abbreviated as PROSUB.
The second measure is based on 2013 sales information and the distinction of business lines published on Orbis database. Although the classification of these business lines is individually undertaken by each firm, this information has the advantage of capturing the distribution of the total sales across all companies` business units. This variable aims to complement the first measure to obtain a more complete indication of product diversification. As stated in appendix C, this measure is abbreviated as PROSEG.
The third measure is a dummy variable, which is equal to 0 when the company sells vehicles under only one brand name, like KIA Motors, and is equal to 1 when the company sells vehicles under multiple brand names, like Volkswagen. Generally, brand names appear to be very important in the automotive industry, which is why this additional variable was developed. The information about major brand names is partly available on Orbis database and was completed by using recent annual reports. This measure is specifically designed for the car manufacturing industry and has not been applied before in this way. However, it is stated by several papers that intangible assets, like brands, play an essential role when it comes to diversification, since they can easily generate synergies (Hitt et. al. 2006; Tanriverdi & Venkatraman 2005; Chang 1995). This justifies the inclusion of this measure as a third proxy to complement sales-based and subsidiary-based product diversification. As stated in appendix C, this measure is abbreviated as PROB.
3.2.2. International diversification
The second measure uses the distribution of sales across different geographic segments of the year 2013. As described above the Herfindahl-Index is applied to calculate the concentration of sales, in order to generate objective values. Although the distinction of geographic segments has been undertaken by each company individually, the classifications are very similar so that it was possible to standardize the information. After this aggregation, the sales for five geographic segments, which were Africa, Australia, Europe, North America and South America were obtained for each company and inserted to the Herfindahl-Index. The subsequently calculated concentration values provide an objective indication for the degree of international diversification of sales. As stated in appendix C, this measure is abbreviated as INTSEG.
By using these two measures for international diversification, this thesis covers firm activities in terms of sales as well as foreign direct investments and thereby provides a comprehensive picture of this dimension. This also addresses the issue of different and incomparable measurement approaches of previous literature, which was suggested to cause the contrary empirical results of prior studies (Glaum & Oesterle 2007; Hennart 2011; Wiersema & Bowen 2011). There is another difference between the two measures: while the first, sales-based, measure considers only five geographic regions, the second, subsidiary-based, measure distinguishes regions on a country-level. It can be stated that the combination of these two classifications of regional segments adds additional robustness to the analysis.
3.2.3. Technological diversification
technological diversification. As stated in appendix C, this measure is abbreviated as TECH.
3.3.Dependent variables 3.3.1. Profitability
The dependent variable is firm-level performance and is also withdrawn from Orbis database. Previous literature suggested either a market-based or an accounting-based approach to measure performance in the context of diversification (Bausch & Pils 2009). While market-based measures reflect the future performance expectations of investors, accounting-based measures refer to actual historical results (Dubofsky & Varadarajan 1987). Therefore several studies measured performance with return on assets and Tobin`s Q in order to include both types of measures (Chen, Yang & Lin 2013; Ravichandran et. al. 2009; Tongli, Ping & Chiu 2005). However, most research in the field of diversification focused on accounting-based performance in terms of return on assets and return on equity (Qian 2008; Elango 2008; Buhner 1987). Furthermore it is stated by Holzmann, Copeland & Hayya (1975) that mangers make strategic decisions towards diversification based on financial statements, which justifies the application of accounting-based measures in the case of this thesis. For this reason, performance is measured by two profitability ratios: net return on assets using P/L before tax (ROA) and return on equity using net income (ROE). In order to prevent the influence of performance fluctuation, the average profitability ratios between 2010 and 2012 were calculated for both measures and then included into the model as the dependent variables. Due to a lack of available data on Orbis the profitability measures of the year 2013 are not considered.
3.3.2. Growth
inclusion of growth measures as dependent variables contributes to the integrity of this thesis. Another argument that justifies this approach is the finding that growth is an important part of the strategy of multinational corporations and is furthermore determined by a firm`s international portfolio (Nachum & Song 2011). Especially in the automobile industry, market share, which is obviously determined by firm growth, plays an important role (Train & Winston 2007). Thus, it appears to be essential to add growth as a dependent variable additionally to profitability. In this thesis, growth is calculated by the percental increase or decrease of total sales from 2004 to 2012 and represented by the variable GROWTH.
3.4.The Model
The analytical model aims to evaluate if and how firm-level accounting based profitability and growth performance can be explained by the three dimensions of diversification in the car manufacturing industry. Another objective of the empirical part of this thesis is to detect potential interaction effects between the dimensions of diversification. Thus, a multiple regression analysis is the best statistical tool to examine the data. This approach to investigate the relationship between diversification and performance has been widely accepted in prior studies (Thomas 2006; Tongli, Ping & Chiu 2005; Geringer, Tallman & Olsen 2000).
Results
4.
The results section first explains model 1 and 2, which analyze the relationship between diversification and profitability, in terms of return on assets and return on equity. Subsequently, the results of model 3, which examines the impact of diversification on sales growth, are described in section 4.2. Afterwards, the influence of selected interaction terms between different measures of diversification on firm performance is revealed in section 4.3. Lastly the robustness checks are explained in section 4.4. An overview of all significant results and the findings of the hypothesis tests is provided in appendix F.
A list of abbreviations including brief descriptions for all variables is provided in appendix C. Importantly, it has to be noted that all diversification measures were calculated with the Herfindahl-Index, which increases, when the degree of diversification decreases. Thus, a positive correlation between the independent and dependent variables implies a negative relationship between diversification and performance and vice versa. The only exception is the dummy variable, which is equal to 1 when a firm sells vehicles under several brand names and equal to 0, when it sells vehicles under only one brand name. Thus the value for this variable is higher when the company is diversified across different brands.
4.1.Diversification and Profitability (model 1 and 2)
Table1: Descriptive statistics of variables (Profitability) Mean Standard deviation N ROA 14,603 38,250 30 ROE 3,356 8,461 30 PROSEG 0,789 0,228 30 PROSUB 0,514 0,277 30 PROB 0,400 0,498 30 INTSEG 0,671 0,272 30 INTSUB 0,539 0,312 30 TECH 0,442 0,169 30 4.1.1. Model 1
The first model analyzes the explanatory power of the diversification measures towards the variance of return on assets (ROA). Before the results can be interpreted, the assumptions of the model have to be tested. The correlation matrix in table 2 shows that only few independent variables have a correlation value higher than 0.5, but all values are still below the threshold of 0.7 suggesting that there is no multi-colliearity problem. Furthermore the tolerance values in table 3 are not close to 0, which indicates that each variable has its own unique variance that cannot be explained by other independent variables. Also the relatively low VIF values confirm the assumption of no multi-colliearity. Moreover the model was tested for linearity, normality and potential outliers as shown in appendix D.
shows that none of the independent variables has a significant effect on return on assets, since the p-values are all above 0.05. Thus, the first model has no predictive power towards the dependent variable, return on assets.
Table 2: Correlation matrix (Return on assets)
Model 1 ROA PROSEG PROSUB PROB INTSEG INTSUB TECH
ROA 1,000 PROSEG -,138 1,000 PROSUB -,192 ,190 1,000 PROB ,136 -,505 -,279 1,000 INTSEG -,008 ,321 ,383 -,592 1,000 INTSUB -,106 ,397 ,512 -,357 ,547 1,000 TECH ,240 ,015 ,119 -,306 ,298 ,296 1,000
Table 3: Linear regression results (Return on assets)
Model 1 Beta p-value Tolerance VIF
Constant 0,656 PROSEG 0,011 0,964 0,642 1,558 PROSUB -0,164 0,478 0,711 1,406 PROB 0,253 0,361 0,498 2,008 INTSEG 0,176 0,520 0,509 1,964 INTSUB -0,128 0,638 0,513 1,951 TECH 0,322 0,145 0,808 1,238 4.1.2. Model 2
Compared to model 1, the measure of technology-based diversification (TECH) shows a relatively low, but again positive correlation with the dependent variable, illustrated in table 4. Also similar to the first model is that the international diversification measures (INTSUB, INTSEG) have almost no correlation with the dependent variable, while the product diversification measures (PROSUB, PROSEG) show a small negative correlation. In line with the findings of model 1, table 5 illustrates that none of the independent variables have a significant effect on return on equity, since all p-values are above 0.05. Therefore, this model has no explanatory power towards return on equity.
Table 4: Correlation matrix (Return on equity)
Model 2 ROE PROSEG PROSUB PROB INTSEG INTSUB TECH
ROE 1,000 PROSEG -0,142 1,000 PROSUB -0,145 0,190 1,000 PROB 0,085 -0,505 -0,279 1,000 INTSEG 0,028 0,321 0,383 -0,592 1,000 INTSUB -0,027 0,397 0,512 -0,357 0,547 1,000 TECH 0,108 0,015 0,119 -0,306 0,298 0,296 1,000
Table 5: Linear regression results (Return on equity) Model 2 Beta p-value Tolerance VIF
4.1.3. Summary of model 1 and 2
From model 1 and 2, it can be concluded that none of the diversification measures contribute any predictive power, when either return on assets or return on equity is inserted as the dependent variable to capture profitability. While international diversification does not seem to have any correlation with profitability, the product diversification measures show a small negative correlation with the dependent variable, indicating that this dimension is positively associated with profitability. The measure for technological diversification is positively correlated with both return on assets and return on equity, which implies a negative influence of this dimension on profitability.
4.2.Diversification and Growth (model 3)
The relationship between diversification and firm growth is examined by conducting a linear regression analysis with all six independent variables of diversification and one dependent variable for growth. Table 6 shows the descriptive statistics of these variables. The sample was reduces to 27 companies, since three firms have no sales and turnover information available on Orbis for the year 2004.
Table 6: Descriptive statistics of variables (Growth)
Mean Standard error N
GROWTH 4,421 15,549 27 PROSEG 0,787 0,223 27 PROSUB 0,482 0,270 27 PROB 0,444 0,506 27 INTSEG 0,656 0,280 27 INTSUB 0,525 0,309 27 TECH 0,443 0,176 27
