Master’s Thesis
Msc International Economics & Business Msc International Business & Management
Firm-level research does not confirm
prosperity of offshoring.
A firm-level research that provides insight in the financial performance of firms that offshore
business functions and firms that do not.
Supervisor: Prof. dr. van Ees Co-assessor: Prof. dr. de Haan
Name student: Donnie Temmink Student number: s1878247 E-mail: temmink.d@gmail.com
2 July 2015
3
Preface
This master’s thesis was written to fulfill the master studies International Economics and Business and International Business and Management of the University of Groningen. During my master studies I developed an interest in the fact that the production of goods and services is increasingly organized in ‘global value chains’. After reading the existing literature on this topic I wondered if firms that source internationally actually improved their financial performance whit this strategy. This interest resulted eventually in the research question of my master’s thesis.
I would like to thank my supervisor from the University of Groningen, prof. dr. van Ees, for his critical note, valuable feedback and good guidance. I also would like to thank Statistics Netherlands for allowing me to use their unique datasets for my thesis. As well for the workspace that was created and the support I received during my three months as an intern. In particular I would like to thank my supervisor from Statistics Netherlands, Dr. Oscar Lemmers, for his support and valuable advice.
4
Abstract
This research seeks to investigate if firms that have offshored business have a better financial
performance than firms that remained to perform all business functions from home base. To test if this is the case, official datasets of Statistics Netherlands are used. Offshoring is a concept term used to indicate firms that relocate business functions to a foreign destination. In this research a distinction is made between firms that internationally outsourced supporting or core business functions and firms that internationally insourced business functions. After controlling for firm-specific characteristics, the results showed that firms which offshored business functions do not perform better than firms that remained to perform all business functions from home base. This was mostly due to the fact that firms which internationally insourced business functions have a substantial lower financial performance compared to firms that internationally outsourced business functions. In the latter group, the firms that outsourced supportive business functions had a higher return on value added than firms that
outsourced core business functions.
5
Table of contents
1.Introduction ... 6
2.Literature review and hypotheses ... 9
2.1 Offshoring ... 9
2.2 Governance mode ... 11
2.2.1 International insourcing ... 12
2.2.2 International outsourcing ... 13
2.3 Financial performance ... 14
3. Data and Methods... 16
3.2 Variables ... 18 3.2.1 Dependent variables ... 18 3.2.2 Independent variables ... 20 3.2.3 Control variables ... 20 3.3 Models ... 23 3.4 Methodology ... 24 4. Empirical Results ... 24 4.1 Descriptive analyses ... 25 4.2 Explanatory analyses ... 25 4.3 Robustness checks ... 29 4.4 Discussion ... 30
5. Conclusions and future research ... 31
6. References ... 34
6
1.Introduction
Should a firm transfer business functions currently performed at home base to a foreign location?1 Executives are increasingly dealing with such questions about spatial reorganization of value chains. This relocation of business functions, driven by globalization, has led to a vast increase of international trade in intermediate goods and services (Miroudot et al., 2009). For instance in the Netherlands, the import value of intermediate goods between 2003-2013 nearly doubled (Statistics Netherlands, 2015). These numbers indicate that firms are increasingly and intensively operating in global oriented value chains (Gerrefi, 2005; Lewin & Volberda, 2011; Timmer et al., 2013). According to the theory of comparative advantage of the classical economist Ricardo,
specialization takes place on the product level between countries. Because every country produces the products for which it has a comparative advantage, international trade takes place. The current
fragmentation of the value chain is somewhat an extension of this principle, but then at a task or function level.2 Every firm performs an activity in the global value chain where it has an absolute advantage (Baldwin, 2006). A substantial difference between current fragmentation and the theory of Ricardo is that in the Ricardian theory transportation and transactions cost are not taken into account. For firms that consider to internationally reallocate business functions these costs are however of great importance. Excessive transaction costs can harm the profitability of a firm and therefore is an
important reason why some firms are still reluctant to transfer business functions to foreign locations (Williamson, 1975).3
Historically, firms performed sequential production stages in close proximity to each other, mostly in the same facility or factory (Baldwin & Evenett, 2012). Around the 1980s a trend started of companies geographically relocating their value chains. This means that firms relocated business functions to foreign locations that were previously produced from home base (Grossman & Helpman, 2011; Roza et al., 2011). This relocation was primarily driven by the vast wage difference between the peripheral south and the industrial north.4 Before the 1980s it was hard to exploit these wage
differences because of coordination constraints. Due to a lack of good communication technology it was very costly to coordinate activities at distance. However, by technological advances in the ICT sector, telecommunication became more reliable, cheaper and widespread (Jones & Kierzkowski, 1990; Baldwin & Evenett, 2006). This made it technically possible for firms to control and monitor activities abroad. Consequently, firms started to relocate activities to low labor cost countries with the
1
According to Porter (1990), home base is the nation where most of the competitive advantages of firms are created and sustained. The home base is the location where in general most advanced skills, productive jobs and core technologies are located. In this research the home base county will be the Netherlands, because this is the location from where activities are offshored.
2
Fragmentation was introduced by Jones and Kierzkowski (1990) and describes the process of geographical dispersion of locations in which the production process and service links take place.
3
Transaction costs are the costs associated with economic exchange, including the cost of overcoming market-imperfections (Williamson, 1975).
4
7 intention to cut labor costs (Baldwin, 2006; Grossman & Helpman, 2001; Nielsen & Tilewska, 2011).5 The process of transferring business functions across borders is called offshoring (Lewin & Volberda, 2011; Manning et al., 2008). In the beginning of the 1980s the motivation of firms to offshore business functions was mainly driven by cost advantages. Recently, however, firms have started to offshore business functions based on other motivations. Firms nowadays also offshore business functions to get access to qualified personnel (Kedia & Lahiri, 2007; Manning et al., 2008), increase market knowledge (Bertrand, 2011) or enhance flexibility (Farrel, 2005). This makes offshoring a strategic instrument for firms to achieve a wide variety of objectives (Hutzschenreuter et al., 2011; Manning et al., 2008). According to Roza, Bosch and Volberda (2011), these different motives for offshoring can be categorized in three groups: cost, resource and entrepreneurial drivers. The desirable effect of these motives are respectively to decrease costs, improve the efficiency of operations or actualize new business opportunities.
It has been argued that offshoring could be an important driver for firms to remain competitive (Katobe et al., 2007; Farrell, 2005). Through offshoring, firms can make use of the comparative advantages some locations provide (Mudambi, 2008; Porter, 1990; Kogut, 1985). And although these potential advantages of offshoring are widely discussed, there is a lack of empirical evidence if firms that offshore business functions actually exploit these advantages and so improve their financial performance. In this paper, we will address this issue by making use of data provided by Statistics Netherlands (CBS). This dataset originated after a survey conducted among Dutch firms in 2011, where firms were asked if and how they performed activities abroad that were previously performed from home base. By using this dataset we examine if Dutch firms that offshored business functions perform better financially than firms that did not offshore. To investigate this relationship the following research question is constructed.
"Do firms that offshored business functions have a better financial performance than firms that perform all business functions from home base?"
Offshoring is a concept referring to firms that source products or services from abroad, that were previously sourced domestically. This can be conducted by either buying the product from a foreign supplier or by performing the activity in-house, but abroad. These two forms can be defined as ‘international insourcing’ and ‘international outsourcing’, see figure 1.The decision to either
internationally insource or outsource an activity can be driven by different motivations. An important determinant of this decision is defined in the theory of transaction cost economics (TCE) (Williamson, 1975, 1985). TCE emphasizes that the ‘make or buy’ decision firms are facing is a balancing act between the investment costs to perform the activity in-house weighted against the possible cost of
5
8 market friction when buying it on the market (Ellram et al., 2013; Mol et al., 2005). Reasons why firms choose to make the activity in-house is often due to intellectual protection constraints, the preference of high managerial control and prevention of hold-up problems (Roza et al., 2011; Gereffi et al., 2005). The choice of firms having to buy the business function is mostly driven by low
switching cost to new partners, no specific investment costs and the possibility to buy the newest services and products currently available on the market (Gereffi et al., 2005; Roza et al., 2011). These characteristics of both offshore modes are widely discussed in current literature, however there is a lack of empirical evidence what the effect of both offshore modes is on the financial performance of a firm. Therefore in this research the financial performance of firms that internationally insourced or outsourced business functions will be compared.
Figure 1: Types of sourcing
The tremendous growth in intermediate trade, due to offshoring, has not led simultaneously to an increase in literature about the effect of offshoring on the financial performance of firms. There is abundant theoretical literature exploring the potential benefits and harms of offshoring, however there is a lack of empirical evidence on the actual financial consequences. Till now, just a few researchers addressed this relationship between offshoring and financial performance (Kotabe & Mol, 2009; Mol et al., 2005; Görg and Hanly, 2003). However, these papers focused solely on firms that international outsourced business functions and show ambiguous results. In contrast, this research also takes into account firms that internationally insourced activities. Therefore, to the best of our knowledge, this paper is first in providing an comprehensive overview about how firms that offshored business functions actually perform. Hereby, we make a comparison with firms that remained to perform all business functions from home base. Furthermore, this research makes a distinction in which business functions that are internationally outsourced. Where current literature looks at internationally
outsourcing in general, this research splits internationally outsourcing in outsourcing core or
supportive business functions. By splitting this group of firms that internationally outsourced business functions, results found by previous research can possibly be further explained.
9 theoretical background that led to the argumentation of different hypotheses. Section three covers the data and methodology part. Section four will comprise the empirical results, robustness check and a discussion regarding the results. Finally, section five will provide an answer to the research questions and will elaborate upon its implications for further research.
2.Literature review and hypotheses
In this section the literature about international sourcing will be discussed. Also the variables and the development towards the hypotheses will be explained.
2.1 Offshoring
According to Baldwin (2006), globalization emerged in two unbundlings. The first unbundling started halfway the 19th century and was caused by the discovery of electricity and steam. These inventions decreased the costs of transporting products and people vastly and enabled geographical separation of production and consumption. In combination with cost differentials and economies of scale, it became genuinely feasible for firms to separate these two functions (Baldwin & Evenett, 2014). The first unbundling was followed by the second unbundling around the 1980’s. This second unbundling is characterized by the geographical fragmentation of the value chain. Because the term value chain is of pivotal importance within this study, the term is explained by the definition used by Kaplinsky and Morris (2001): “a value chain is the full range of activities which are required to bring a product or service from conception, through the different phases of production, delivery to final consumer and final disposal after use." The reason why business functions in value chains were previously clustered was mainly because this kept transaction costs low. Before the 1980s it was difficult to coordinate activities abroad. However, when telecommunication became widely available across the globe due to ICT developments, coordination at distance became technically possible. The incentive to reallocate business functions was further reinforced by trade liberalization (Baldwin & Evenett, 2014). These developments decreased transactions costs to control foreign operations and resulted in firms starting to offshore business functions (Manning et al., 2008). In the beginning firms especially offshored business functions that were related to the manufacturing stage. Cost differentials could be exploited due to the abundance of low labor cost countries in the world. Mr. Stan Shih, founder of the Taiwanese technology firm ACER, has illustrated this development with a concept called the ‘smile curve’ (Baldwin & Evenett, 2015). This ‘smile curve’, displayed in figure 2 (see appendix), asserts that the most value is added in the product related support activities: product concept, design and R&D, sales, marketing and after sales services. Manufacturing, especially final assembly, has become a stage were low value added activities are performed. Prior to the second unbundling this ‘smile curve’ was flatter, as in those times manufacturing jobs were able to support higher value added and were well paid (Baldwin & Evenett, 2014).
10 offshoring is not solely used to establish cost benefits anymore. In addition to cost benefits, firms start to offshore business functions to increase knowledge of foreign markets, getting access to resources or enhance flexibility (Bertrand, 2011; Farrell, 2005; Manning et al., 2008; Katobe et al., 2009).
Furthermore, offshoring increases export sales, especially in the markets where firms have sourced their intermediate products (Bertrand, 2011). Another advantage of offshoring is that it provides firms the opportunity to offer customers 24 hours service support, because they can have employees in different time zones (Siem & Rather, 2013).
Besides the advantages of offshoring, there is some uncertainty about managing activities abroad (Hutzenschreuter et al., 2011). This uncertainty is based on operating in an unknown foreign environment, with different institutions (Scutcliff & Zaheer, 1998). As a result, transaction costs usually increase when firms internationalize their sourcing process, due higher governance and control costs (Williamson, 1975; Coase, 1937). If these additional transaction costs are not compensated by the potential benefits, there is no incentive to offshore for firms in the first place.
Lately, there also have been firms that decided to reshore activities (Ellram et al., 2013; Gray, 2013). Reshoring, sometimes referred to as “backshoring” or “reversed offshoring”, is the relocation of previously offshored activities back to the home base location of the firm (Ellram, 2013). There are different reasons for firms to reshore activities, for example the development in technology of
machines. Machines can replace labor and thereby makes the comparative advantage of cheap labor in a foreign country redundant. Therefore, firms can decide to reshore production phases close to their customers again, to reduce transportation costs (Neil, 2013). There also have been firms that offshored business functions, but did not sufficiently take into account the working conditions of the employees abroad. This led to a rise of criticism in the media against how firms govern their foreign suppliers (Lock & Romis, 2007; Lund-Thomsen & Nadvi, 2010). However, the percentage of firms actually reshoring activities is limited. A survey among European manufacturing firms indicates that only 4% of the firms have reshored activities in 2012 (Dachs & Zinker, 2013).67 However, for every firm that has reshored activities, more than three firms offshored activities to foreign locations (Dachs & Zinker, 2013). These numbers show that on average fragmentation continues and the potential financial benefits of offshoring outweigh the benefits of reshoring. This could be explained by the belief that offshoring is a strategic tool to achieve a wide range of goals. Besides, there also have been further developments in information and communication technology the last decades. These
developments welcomed a further decrease in transaction costs for firms to control foreign operations (Baldwin, 2006; Ford, 2011). This makes it possible for firms to exploit comparative advantages of countries and regions (Kogut, 1985) Furthermore, the developments in ICT makes it easier for firms to
6
The European Manufacturing Survey (EMS) is organized by a consortium of research institutes and universities co-ordinated by the Fraunhofer Institute for Systems and Innovation Research ISI and takes place every three years. More than 3,500 firms in 13 countries participated in the latest EMS survey in 2012.
7
11 find specialized suppliers abroad. These suppliers can potentially deliver a higher quality or can produce at lower cost (Chan & Kumar, 2007). Moreover, Firms that offshored supportive business functions, that are not part of their main business, can increasingly focus on sustaining and improving their core competences. This specialization is important in current markets, because competition is increasingly focused on task level (Baldwin, 2006). To summarize, due to lower transactions costs, better possibilities to find foreign suppliers and increased competition on task-level, the potential of offshoring has increased. Offshoring seems to be the formula for firms to remain competitive, improve financial performance and secure sustainable existence for now and in the future. Therefore, our first hypotheses postulates that firms which offshored business functions have a better financial
performance than firms that did not.8
Hypothesis 1: Firms that offshored business functions have a better financial performance than firms
which remained to perform all business function from home base.
2.2 Governance mode
Lead firms that operate in internationally fragmented value chains can govern these chains in different ways. A lead firm is a firms that is the initiator of new or existing products. In addition the lead firm controls the geographical dispersion of their production networks (Sturgeon, 2001). Gerreffi (2005) conceptualized different governance modes for lead firms to monitor global activities based on explicit coordination and power asymmetry. This conceptualization is based on ascribing different values to three variables: complexity of information, the ability to codify information and the
competence of the supplier. This led to five generic governance modes situated between, on one end of the spectrum, hierarchy, and on the other end, arm’s length. Between those two extremes, three
intermediate governance modes are constructed, respectively, modular, relational and captive. These intermediate modes of value chain governance comprise network relationships (Gereffi, 2005) .In this research a distinction will be made between two different governance modes, depending on whether or not the supplier vertically integrated with the lead firm. These two forms will respectively be called international outsourcing and international insourcing (Nielsen & Tilewska, 2011). Compared to the governance modes conceptualized by Gerreffi (2005), hierarchy is comparable to international insourcing and the other four modes fall under the umbrella of international outsourcing. Sourcing, included in both terms, describes the phenomenon of firms managing the flow of services, finished products or components used to serve foreign and domestic markets (Kotabe et al, 2007).
Firms are heterogeneous and significantly different in competences, strategies and capabilities (Barney, 1991). Therefore, firms that source internationally can use different governance modes
8
12 (Quinn & Hillmer, 1994). These different perspectives on boundary decisions are mainly determined by firm-specific knowledge (Connor & Prahalad, 1996), strengths and weaknesses of the firm and the attributes of the transactions (Lieblein & Miller, 2002). These attributes describe the characteristics of a certain product or service (Lieblein & Miller, 2002). If transactions concern time sensitive, non-standard products or focus on integral product architectures, coordination costs often increase (Baldwin & Clark, 2000). In these cases, firms in general will choose more frequently for an integrated governance mode, like hierarchy (Gerreffi, 2005).
A governance mode is selected with certain expectations, however, managers should take into account that both forms have their cons and pros, and it is hard to enjoy all preferred goals by one specific governance mode (Hutzschenreuter et al. 2011). The characteristics of both governance modes, international insourcing and international outsourcing, are explained further in the following sections.
2.2.1 International insourcing
In the theory of Gereffi (2005) international insourcing is governed by hierarchy, decisions are made by the lead firm and flow from headquarter to foreign subsidiaries. This form provides the highest managerial control for firms and is preferable when complexity of transaction is high and when the ability to codify transaction and capabilities in the supply base are low (Gereffi, 2005). A vertically integrated firm can be established by acquiring an existing supplier abroad or setting up a business function, though either a greenfield or brownfield investment (Hutzschenreuter et al., 2011; Johnsen, 2005).9 Firms that internationally insourced business functions have in general lower
transaction cost compared to firms that internationally outsourced business functions, however, on the other side, higher production costs (D’Avani & Ravenscraft, 1994). Lower transaction costs are achieved because firms can align their interest with foreign affiliates, and in this way reduce their dependency on external suppliers, which prevent potential hold-up problems (Pfeffer & Salancik, 1978; Williamson, 1985). A more hierarchical structure aligns interest of both parties which leads to a more efficient sequential production and lower transaction costs (Williamson, 1975).
Although insourcing has its benefits, production costs are often higher compared to external suppliers, due to a lack of competition. External suppliers provide the benefit of specialization and can realize lower production costs (Lieblein & Miller, 2003). However, in some cases a firm is forced to develop a task or business function in-house, because no suppliers are available on the market that delivers a desired product or service (Gerreffi, 2005). Williamson (1998) argues that the firms decision about an applied governance mode in the past, affects future governance choices. Firms that chose to internalize business functions in the past are more likely to remain integrated in the future. This indicates that the decision of firms to govern foreign activities by a chosen governance mode is partly
9
13 determined by so called ‘path dependence’. This means that international insourcing entails more irreversible investment cost than firms that make use of international outsourcing. Insourcing makes firms more rigid compared to outsourcing (Leiblein & Miller, 2003). International insourcing therefore bears also a certain amount of risk. Because in markets that are highly dynamic and exposed to volatile demands, large investments can lose their value quickly (Betis et al., 1992).
2.2.2 International outsourcing
International outsourcing is carried out when a firm transfer business functions to an external supplier abroad, that were previously performed in-house (Siems & Ratner, 2003; Grossman & Helpman, 2001) The external supplier is not part of the same enterprise group as the lead firm (Hutzschenreuter et al., 2011). International outsourcing can take place through different governance modes, namely market based, modular and relational and captive. The similarity between these modes is that the lead firm has no direct control over its supplier (Lieblein & Miller, 2003).
A reason for firms to refrain from outsourcing can be due to sensitivity of sharing information, which is especially the case with R&D related activities. Firms can also be reluctant to outsourcing because they do not want to increase their coordination costs, become dependent on an outside party or lose any competences (Bertrand, 2011). Outsourcing can have a negative effect on organizational innovation and control over the activities of the firm. It can also increase transaction costs
significantly, due to uncertainty and contract enforcement (Gilly & Rasheed, 2000). Especially in countries with underdeveloped institutions, contract enforcement can be a protracting process. In these countries, outsourcing may have a negative effect on the financial performance of a firm. As a result, firms in these situations will preferable choose for vertical integration (Bettis et al., 1992).
Nevertheless, there is a trend of suppliers increasing their capabilities, hence a shift to more intermediate governance modes within global value chains (Gerreffi, 2005). Suppliers which improve their capabilities, provide opportunities for lead firms to choose the most efficient and best qualitative suppliers available on the market (Gerreffi, 2005; Dess et al., 1995). Compared to a vertically
14 However, a governance decision is firm specific, recent developments are especially
beneficial for firms that internationally outsourced business functions. International outsourcing provides the opportunity for lead firms to make use of supplier specialization, improve their flexibility and realize cost benefits. Therefore, we hypothesize that firms that internationally outsourced activities have a better financial performance than firms that internationally insourced business functions.
Hypothesis 2: Firms that internationally outsourced business functions have a better financial
performance than firms which internationally insourced business functions
A firm can internationally outsource different business functions, for instance supporting or core activities.10 When lead firms outsource supporting business functions to an outside supplier, they can focus more on activities related to their core competences (Quinn, 1992). Summarized by Siems & Ratner (2003), as a firm to remain competitive you should: “Do what you do best and outsource the rest.” By contrast, according to the ‘smile curve’ (Baldwin & Evenett, 2015), lead firms should outsource their core business functions to shift focus to high value added activities. Both theories sound plausible. However, there is no empirical evidence that supports the theory about the ‘smile curve’ (Baldwin, 2006). Neither is there any empirical evidence that supports the theory that firms should focus on their core business functions and outsource supportive business functions (Siems & Ratner, 2003).
The datasets used for this research contain information about which business function are outsourced by firms. This makes it possible to test which kind of firms perform better, firms that outsourced core or supportive business functions. Therefore this research is the first that can provide some insight in which theory outweighs the other. However, this does not mean that firms should always outsource supportive or core business functions, because this decision is also firm specific, based on transaction costs. However, this research does provide an overview of how firms generally perform, after outsourcing core or supportive business functions, controlled for some firm specific characteristics.
Hypothesis 3a: Firms that internationally outsourced supportive business functions have a better
financial performance than firms which internationally outsourced core business functions.
Hypothesis 3b: Firms that internationally outsourced core business functions have a better financial
performance than firms which internationally outsourced supporting business functions.
2.3 Financial performance
In this study the relationship between offshoring and financial performance is addressed. This
10
15 financial performance of an enterprise is based on its profitability (Cowen et al., 1997). Where a vast amount of literature focusses on the antecedents of offshoring (Lahiri & Kedia, 2011; Lewin & Volberda, 2011; Weidenbaum, 2005), there is a limited amount of literature about the consequences or financial implications of offshoring. Firms that offshore business function will have the expectation that this will improve their performance. The question is whether this is indeed the case.
Till now there have been a few researchers which addressed the relationship between
offshoring and financial performance (Görg & Hanley, 2004; Mol et al., 2005; Kotabe & Mol, 2009; Bertrand, 2011). Görg and Hanely (2004) investigated the relation between outsourcing and the ratio of net profits over total output at the plant level. They concluded that this depends on the
characteristics of the plant and their interactions; especially the offshoring of large plants has a positive effect on financial performance. These benefits mainly holds for manufacturing firms, but are less clear for service firms. Mol, Tulder and Beije (2005) investigated the relationship between international outsourcing and financial performance. In their research, they aggregated several financial and market performance indicators to one performance variable. Hereby financial performance indicators were return on sales (ROS) and return on investment (ROI) and market performance indicators were market share and sales growth.11 They conclude that there is neither a performance enhancing nor worsening effect on manufacturing firms that internationally outsourced business functions. In their research only manufacturing firms that internationally outsourced business functions were taken into account.
In the research of Kotabe and Mol (2009), where the effect of outsourcing on financial performance is investigated, return on value added (ROVA) was taken as an indicator of firm performance. The authors conclude that the relationship between the overall outsourcing level and firm performance is negatively curvilinear in nature.
In the paper of Bertrand (2011) the performance measure used is export sales, where a positive relation is found between international outsourcing and export sales. When looking at the current literature on offshoring, it is notable that it is exclusively focused on internationally outsourcing and that internationally insourcing is underexposed. Therefore there is little known about the effects on financial performance on firms which internationally insourced business functions.
In this research two indicators for financial performance are used, ROVA and ROS. The choice for these indicators are based on the following reasons. Firstly, both indicators are used in other papers where the relation between offshoring and financial performance was investigated (Mol et al., 2005; Kotabe & Mol, 2009) Secondly, both indicators are ratios, making it possible to compare firms. Thirdly, these measures are easy to calculate and to interpret. Further, the first measure, return on sales, provides a good aggregated view on the accounting performance of a firm. The second
measure, return on value added, is less aggregated and an insightful tool that shows how much profit a
11
16 firm actually made in relation to the value it added. This latter measure is increasingly important in the literature about offshoring, which addresses that firms should especially focus on the high ‘value added’ business functions (Baldwin, 2006) In section 3.21. these two financial measures are further defined and operationalized .
3. Data and Methods
This section will firstly describe the creation of the dataset, then provide an overview of the variables used in the analysis, and finally give a formulation and explanation of the empirical model.
3.1 Data collection and sample
This research investigates if any differences exist regarding financial performance between firms that offshored business function and firms that remained performing business functions from home base. Therefore firm level data is required. Firstly, to find financial differences between firms that offshore and firms that do not. Secondly, to have the possibility to control for firm specific characteristics (Hijzen et al., 2006). However, this micro level data about offshoring and financial performance is firm sensitive information and therefore not available in public databases. This lack of pooled data about firms decisions to offshore can be an important reason why, till now, there has been published so little about the financial consequences of offshoring. However, there are to my
knowledge two institutions that own firm level information regarding offshoring decisions. These are the Offshoring Research Network12 (ORN) and Statistics Netherlands.13 Statistics Netherlands (CBS) is the official statistical institute of the Netherlands and possess data of a big pool of variables from a large amount of firms. Fortunately, a request to get access to their unique datasets was granted. A caveat was that the required data could only be accessed through computers at the office from Statistics Netherlands. Therefore, for this research, I spend three months at Statistics Netherlands in Heerlen.
The firm specific variables within Statistics Netherlands are to be found in different datasets and therefore had to be combined. The main dataset, used in this research, contains information if firms perform any business functions abroad that were previously performed from home base. This main dataset is complemented by several other datasets, that contained information about other firm specific characteristics like financial figures, type of ownership, amount of employees and industry.
This main dataset originated from a survey conducted in 2012 among enterprises operating in the Netherlands.14 The idea behind the survey originated on an European level. Due to discussions about the impact and magnitude of globalization, some National Statistical Offices within the ESS in
12
The researchers who investigated the relation between international outsourcing and financial performance (Mol et al., 2005; Katobe & Mol, 2009), made use of the datasets from the Offshoring Research Network.
13
Also some national statistics institutions in Europe have micro data regarding offshoring (see footnote 16), however due to confidentiality of this firm level data, it is difficult as a foreigner to get access to this data
14
17 cooperation with Eurostat initiated this Global Value Chain survey (henceforth GVC).15 The objective of this survey was to monitor the effect of globalization on businesses by providing new statistical evidence (Eurostat, 2012). In all participating countries similar surveys were conducted by the concerned National Statistical Offices. However, in this paper only the GVC survey conducted by Statistics Netherlands will be used and discussed.
The data from the GVC survey contains comprehensive information about the offshoring decisions of firms. Questions included in the survey are for example: does the enterprise makes use of offshoring, which business functions are offshored and how are business functions are governed. To be included in the survey, enterprises needed to have at least 100 employees and had to be active in the non-financial economy.16 The non-financial economy, also called business economy, is a combination of sectors B-N, excluding sector K (Eurostat, Nace rev. 2). This classification of economic activities by industry is determined by Eurostat and is displayed in table 3.1 (see appendix). In the Netherlands there were 4560 firms in 2011 that met the requirements for the GVC survey. To limit the
administrative burden, but to create substantial confidence intervals, the survey was send to 2182 firms (Smeets & Staats, 2012).17 Eventually, 1307 firms returned the survey, equal to a response rate of 62,7% from the sample.
The GVC survey does not contain any financial performance indicators of firms like net profits, total turnover or costs of sales. However, this financial information is important for the construction of the variables used in this research. Therefore this information is obtained from the Production Statistics of Statistics Netherlands. These production statistics contain a wide pool of variables of about 80.000 enterprises registered in the Netherlands.18 Firms that receive the annual survey about their production figures are legally obliged to respond, and to respond truthfully. Therefore the response rate of this annual survey is very high, what makes the retrieved data highly representative. Up to and including the year 2013 these Productions Statistics are available within
Statistics Netherlands.
Eventually, the GVC dataset and the Production Statistics datasets are merged through a unique identifier, a so called enterprise id (BEID). In the first instance only the productions statistics of 2011 will be used in this research. However, to conduct robustness checks, the Production Statistics of the years 2012 and 2013 will also be used (see chapter 4.2). There was not a complete coverage between the GVC survey and the Productions Statistics of each year. This lack of complete coverage can be due to different reasons, for example, firms that have gone bankrupt or removed business
15
The European Statistical System (ESS) is an institution with the objection to provide comparable statistics at an European level. Eurostat is an European institution with the main objective to process en publish statistical information at EU level. The countries that cooperated in this project regarding global value chains were: Czech Republic, Denmark, Finland, Ireland, the Netherlands, Norway, Portugal and Slovenia
16
The financial economy consist of monetary financial institutions, insurance institutions, pension funds and remainder financial institutions (www.CBS.nl/methoden/begrippen).
17 By creating this sample, the imputed weighted response rate was 67%, eventually this was a little lower, namely 62,7%. 18
18 functions out of the Netherlands. If the financial figures of firms were not available, these firms were dropped, to obtain a balanced dataset. Likewise, all the foundations and government institutions in the dataset were dropped. These foundations, consisting of charities and local broadcasting stations, do in general not have as their main objective to make a profit, therefore this would not be a reliable control group.19 Thereby, the final sample used in this research consists out of 1156 firms.20 It is important to note that none of the enterprise names can or will be published within this research, due to the sensitivity of the data. Therefore only aggregated results will be presented.
3.2 Variables
3.2.1 Dependent variables
In this study two financial dependent variables are used, return on sales (ROS) and return on value added (ROVA). These financial ratios are indicators about the profitability of a firm (Cowen et al., 1997). Both ratios are multiplied by a hundred so they can be interpreted as percentages. The two financial measures are constructed for the years 2011, 2012 and 2013.
The first financial performance indicator that will be discussed is return on sales, or so called operating profit margin, and is calculated by the following formula:
* 100%
EBIT is an abbreviation for ‘earnings before interest and taxes’ and is calculated as operating income plus non-financial income. The advantage of using EBIT is that it nulls the effect of different tax rates and capital structures used by different enterprises (Goldstein et al., 2001). Therefore it is easier to make a reliable comparison between the financial performances of enterprises. The denominator, total turnover, is the total amount of revenues a firm receives from its business activities. Because ROS is a ratio, it is a convenient tool to compare firms among each other. However, ROS is a highly aggregated performance indicator (Banker et al., 1996). Therefore, this financial measure is less useful when explaining causes of low or high performance. In that extent this measure can be even somewhat deceitfully, because profit margins in some industries are in general lower compared to other industries. This does not necessarily means that these industries are less profitable, because relative ROS margin can be low, but the absolute profits can be high. A good example of such sector is the manufacturing industry. In this industry, firms are dealing with very high quantities but low margins. Comparing the ROS of firms active in the manufacturing industry, with the professional services industry, where margins are in general higher, would not be a fair comparison. Therefore ROS is an
19
Ditto counts for government institutions, these institutions will not have as their main objective to make a profit. Therefore, these institutions are also excluded from the file. An overview of legal forms included and excluded in the dataset can be found in table 3.1 (see appendix)
20
19 adequate tool to compare firms within the same industry, but not to compare firms active in different industries. To control for this, industry is added as a control variable in our model (see chapter 3.2.3). Furthermore, in the research of Mol, Tulder and Beije (2005), where the relationship between financial performance and offshoring is investigated, ROS was also used as a determinant of financial
performance. The authors conclude that no performance enhancing effects was observed by firms that internationally outsourced business functions. Important to note is that this research made use of data from 2001. Due to developments, as discussed in section 2.2.1, we expect that this effect nowadays is observable. Katobe & Mol (2009) state that ROS is not an appropriate measure in their research because it carries a consistent bias towards a negative relationship with outsourcing. However, this is only the case when looking at a panel dataset, where the transition of firms that offshored business functions is investigated. In their research there is not a comparison made between firms that offshored business functions and firms that did not. Because this research uses a cross section dataset, where firms that offshored business functions are compared with firms that did not, ROS is an appropriate measure.
The second financial measure used in this paper to provide an indication about the profitability of a firm is return on value added (ROVA). This measure is calculated by the following formula:
ROVA
* 100%
In principle these two financial variables are comparable, both are ratios and make use of EBIT in the nominator and total turnover in the denominator. However, by calculating the ROVA, all the business functions that a firms purchased from an external supplier, so called external sourcing costst, are subtracted from the total turnover. The remainder is the so called ‘value-added’ of a company. This value added consists partly out of costs made for the production of the product or service. These are mainly labour and capital costs. The remaining part of the value added is the premium for the firm, or so called profit.
20 added to the product or service. And in particular, by looking at the ROVA, how much profit it made over this ‘value added’
3.2.2 Independent variables
The independent variable in the first regression is offshoring. This is a dichotomous variable, that takes the value of 1 if an enterprise make use of (an) internal or external supplier(s) abroad and 0 otherwise. The amount of firms in the sample which actually made use of offshoring in the year 2011 is 365, from a total of 1156 firms. Important to note is the survey does not provide any insight to what extent the firms make use of offshoring, regarding magnitude, but only that business functions that were previously performed from home base are now sourced from abroad. Because the variables are dichotomous, a distinction regarding the degree of offshoring business functions cannot be made. However, it is possible to compare the means of firms that use offshoring as an strategic instrument and firms that do not. And this is the specific question addressed in the hypothesis. Neither is specified from which year onwards firms started to source business functions from abroad.
For the second hypothesis a distinction under the umbrella of offshoring is made between international insourcing and international outsourcing. A categorical variable is created, where the variable takes the value of 0 if no activities are offshored, 1 if activities are international insourced, 2 if activities are outsourced and 3 if activities are internationally insourced and outsourced. This last mode where firms internationally insourced and outsourced business functions will be called hybrid sourcing firms. This category has to be included otherwise a clear comparison between firms which solely insourced or outsourced cannot be made. The distribution of which governance mode was used by firms that offshored can be found in table 3.3 (see appendix).
In the third hypothesis we further zoom in on firms that internationally outsourced business functions. Therefore, a distinction is made between firms that internationally outsourced core or supporting business functions. Core functions are defined as the production of goods and services for the market. Supporting activities are specified as: distribution and logistics, marketing, ICT services, administrative services, R&D activities and additional supporting activities. A categorical variable is created, where the variable takes the value of 0 if all business functions are performed from home base, 1 if business functions are international insourced, 2 if business functions are hybrid sourced, 3 if supporting business functions are internationally outsourced, 4 if core business functions are
international outsourced and 5 if both core and supporting business functions are internationally outsourced. How many firms fall into each category is tabulated in table 3.3 (see appendix).
3.2.3 Control variables
21 an effect on the coefficient nor on the explicability of the model. Therefore, trade status is further excluded from the regressions as control variable.
Firm size: There are several measures that can indicate firm size. Some examples are total turnover,
employment or total assets. These different indicators are all valid measurement and often highly intercorrelated (Shalit & Sankaru, 1977). Because we focus on value chains in this research it would be particularly interesting to relate firm size with the amount of ‘value added’. This value added is partly determined by the allocated production factors, labour and capital (Solow, 1956). These production factors are used as inputs to produce the output, or value added, of the firm. Therefore, total amount of employees is used in this research as an indicator of firm size. The most adequate measure of labour is the total hours worked or the amount of full time equivalents (FTE’s).
Unfortunate, this data is not available in the Production Statistics of Statistics Netherlands. This input would have been difficult to capture when looking at other indicators of firms size like total turnover or total assets.
In most literature where firm size is used as an control variable, the natural logarithm of the size is taken. However, because our dependent variable is financial performance this would assume that there is a linear or exponential relationship between firm size and financial performance. Partly this is confirmed by the existing literature, larger firms can create scale and scope advantages, what could improve efficiency and so financial performance (Majundar, 1997). However on the other hand, larger firms are less nimble to respond on changing market conditions, what could harm their
competitiveness. Small firms on the contrary have the possibility to adapt fast to new demands by having a more flexible managerial organization (Audretsch & Siegfried, 1993).
For the firms included in the sample for this research there was neither a linear nor an exponential relationship found between firm size and financial performance. The largest firms in the sample, with 2000 employees or more, had lower financial ratios than firms with less than 2000 employees. To control if this was due to a sample error, the relationship between firm size and financial performance was examined for all the firm included in the Production Statistics.21 Here a similar results were found and therefore we use a categorical variable to control for firm size. Statistics Netherlands make use of official size classes, as displayed in table 3.4 (see appendix). To make it comprehensible model we aggregated this into three groups. The first group, with the smallest sized firms, consists out of firms with employees up to 250 employees. This is the official Dutch maximum to fall into the group of small and medium enterprise (Statistics Netherlands, 2013).22 Above the 250 employees a distinction is made between two groups, a group with less than 2000 employees and a
21
The total population in the Production Statistics of firms with 100 employees or more is 4560. In this population a similar relation between firm size and financial performance is found as in the dataset used for this paper.
22 According to the official categorization from Statistics Netherlands a small or medium enterprise (SME) is a firm with less
22 group with 2000 employees or more. The new firm size categorization is displayed in table 3.4 (see appendix)
Labour productivity: A high labour productivity has a positive effect on the financial performance
of a firm (Youndt et al., 1996). Labour productivity is calculated as the value added per person employed. In contrary to firm size, there is in general a linear or exponential relationship between labour productivity and financial performance. Therefore, to make a comprehensible model, the natural logarithm of labour productivity is used.
Foreign control: The fact if a firms is foreign controlled is determined by the Ultimate Controlling
Institutional Unit (UCI) . The UCI is defined as an enterprise or institutional unit, which has the ultimate control and is not controlled by another institutional unit (Eurostat, 2015). To specify, foreign controlled is not the same as foreign owned. An enterprise can be owned by a fund in a tax heaven, but is not controlled from this location. The effect of foreign ownership on financial performance is somewhat ambiguous. However, empirical evidence exist that especially firms controlled by large foreign enterprises have a higher financial performance than domestic controlled firms (Barboasa & Louri, 2005). Foreign ownership is a dichotomous variable that takes the value of 1 if a firms is foreign controlled and 0 otherwise. The amount of firms in the sample that are foreign controlled is 337.
Industry: The industry in which an enterprise operates, is an important firm characteristic. As
discussed in chapter 3.2.1, margins are often also industry dependent. Besides, regarding the state of the market, some industries flourish while others are in decline. Therefore it is important to include industry as a control variable in this research.
Statistics Netherlands allocates to every firms a five digit code according to which industry the firms is active.23 The global value chain survey was conducted among firms operating in the business economy, in the sectors B-N. These classification is conceived by Eurostat and can be found in table 3.5 (See appendix). In this table is displayed how many firms from the sample are active in each industry. Because the response of companies in some sectors was very low, this created collinearity, see table 3.6.1 Therefore, to prevent collinearity, the sector ‘electricity, gas and steam supply’ are combined with the sector ‘water supply’ to a new industry: ‘energy’. The sectors ‘mining and quarrying’ and ‘real estate activities’, had so few observations, that firms in these industries were dropped from the sample. This generated a new industry distribution which can be found in table 3.5 (see appendix).24
23 This five digit numbers is called within the National Statistics the Standaard Bedrijfsindeling (SBI). The first two digits
corresponds to the international categorization of industries devised by the OESO. This industry categorization is called the Nomenclature generale des Activites economiques dans les Communautes europeennes (NACE).
24
23
3.3 Models
To test the hypotheses three regression equations will be used. The abbreviation and function of the variables is explained in table 3.5 and the equations are displayed below:
Hypothesis 1
ROVA2011 = α + β1offshoring + β2firmsize + β3 productivity + β4 industry + β5 foreigncontrolled + e
ROS2011 = α + β1offshoring + β2firmsize + β3 productivity + β4 industrty + β5 foreigncontrolled + e
Hypothesis 2
ROVA2011 = α + β1insourcing + β2 outsourcing + β3 hybridsourcing + β4size + β5 productivity + β6
industry + β7 foreigncontrolled + e
ROS2011 = α + β1insourcing + β2 outsourcing + β3 hybridsourcing + β4size + β5 productivity + β6
industry + β7 foreigncontrolled + e
Hypothesis 3
ROVA2011 = α + β1 insourcing + β2 hybridsourcing + β3 supportbusiness + β4 corebusiness + β5
hybridbusiness + β6size + β7 productivity + β8 industry + β9 foreigncontrolled + e
ROS2011 = α + β1 insourcing + β2 hybridsourcing + β3 supportbusiness + β4 corebusiness + β5
hybridbusiness + β6size + β7 productivity + β8 industry + β9 foreigncontrolled + e
Where ‘α’ refers to the intercept and ‘e’ to the error term that accounts for the discrepancy between the predicted outcome and the actual observed outcome.
Table 3.6 Variables included in the model
Dependent variables Label Function
Return on Sales 2011 ROS11 Ratio, calculated as EBIT by total turnover
Return on Value added 2011 ROVA11 Ratio calculated as EBIT by total turnover-external sourcing costs Independent variables
Offshoring offshoring Dummy variable indicating if a firms offshored business functions *
International insourcing insourcing Dummy variable indicating if a firms internationally insourced business functions *
Hybrid sourcing hybridsourcing Dummy variable indicating if a firms internationally insourced and outsourced business functions * International outsourcing outsourcing Dummy variable indicating if a firms internationally outsourced business functions *
Outsourcing core business function(s) corebusiness Dummy variable indicating if a firms internationally outsourced core business functions * Outsourcing supporing business function(s) supportbusiness Dummy variable indicating if a firms internationally outsourced supporting business functions * Outsourcing hybrid businesss functions hybridbusiness Dummy variable indicating if a firms internationally outsourced supporting and core business functions *
Control variables
Firm size size
Categorical variable take the value of 0 if employees <250, 1 if 250<employees<2000 and 2 if employees>2000
Labour productivity productivity Natural logarithm of productivity, which is calculated as added value per employer Industry industry Categorical variable which included 9 different industries, see table 3.5 Foreign controlled foreingcontrolled Dummy variable indicating if a firms is foreign controlled
24
3.4 Methodology
To test whether differences exist regarding financial performance between firms that offshored business functions and firms that remained to perform business functions form home base, ordinary least squares (OLS) regressions are used. When conducting OLS regressions, observations are pooled together, without making any provision for individual or time differences. Because this is a cross section dataset this is not a problem, because there is no time dimension. Note that OLS only provides a statistical relation between the variables, but this does not have to indicate a causal relation between the variables. Therefor using only data, and no economic theory, this test cannot demonstrate that firms which offshore business functions are performing better due to the decision to offshore business functions. This can be due to two reasons. Firstly, there can be another reason why firms that
offshored activities perform better, a so called confounding factor, but is not captured in this model. Secondly, it can also be that the relation between firms offshoring activities and a better financial performance is vice versa. Thus, firms that have better financial performances are more likely to offshore business functions.
However, the combination of a statistical relation, together with the theoretical background used for the construction of the hypotheses, is a strong indication for a causal relation.25
Before being able to make any inferences about the data, some tests need to made, to check whether the model is appropriate. Firstly Breusch-Pragans tests are conducted, to test if
heteroskedasticity is present. These tests are conducted for both dependent variables for the year 2011 and tabulated in table 3.7.1 and 3.7.2 (see Appendix). Because the results show that heteroskedasticity is present, we correct for this by making use in the regressions of robust standard errors.26 To check for multicollinearity a VIF test is conducted. With low VIF values, the results indicate that there is no multicollinearity present among the explanatory and control variables in the sample (table 3.6.2, see appendix).
4. Empirical Results
In this section the results of the analyses are presented. This sector is structured as follows. First an overview of the summary statistics will be provided. Hereafter the unconditional means will be compared according to the constructed hypothesis by making use of several t-tests. Finally, OLS
25
To identify this causal relation, initially we planned to make a panel dataset and hereby compare the marginal differences regarding financial performance between firms that offshored activities and firms which kept performing business functions from home base. These marginal differences can be found by using propensity score matching (PSM). PSM is a statistical tool, that matches two almost identical firms by looking at a wide scale of variables like industry, turnover and firm size. If done so, this method attempts to measure the effect of a certain intervention, policy or treatment, in this case offshoring. Because all the other characteristics are almost identical, the difference in financial performance is then probably explained by offshoring. Unfortunate, it was not clearly stated in the data when enterprises exactly offshored activities. Due to the lack of a clear starting point it is hard to measure the actual difference between these coupled firms. The sample was also too small to find a large pool of coupled enterprises with comparable characteristics. Therefore we were not able to use this technique in this study.
26
25 regressions will compare the conditional means, by controlling for firm specific characteristics. The analyses are conducted with the statistical software of Stata 12 at the office of Statistics Netherland in Heerlen.
4.1 Descriptive analyses
An overview of the summary statistics is tabulated in table 4.1.The summary statistics show the central tendency by providing the mean, median and the standard deviation of the variables. Due to the sensitivity of the micro-data, maximum and minimum cannot be presented. If this information would be published, it would make it possible for third parties to derive firm sensitive information on individual enterprises, which is forbidden by Dutch law on Statistics Netherlands.
In this research we try to investigate if differences between financial performance of firms exist, depending on if and in how they offshored business functions. To get this insight, the summary statistics are further specified according to explanatory variables. Therefore the mean and median of every specific governance mode, regarding the two financial indicators, is calculated and tabulated in table 4.2 (see appendix). The summary statistics show that the difference between firms that did offshore business functions and firms that did not is negligible. From these firms which offshored business functions, firms that internationally outsourced business functions have a substantially higher average financial performance compared to firms which internationally insourced business functions. In this pool of firms that international outsourced activities, firms which internationally outsourced supportive business functions have a higher ROVA and ROS compared to firms which internationally outsourced core business functions. Important to note is that these summary statistics neither indicate any correlation nor control for any firm specific characteristics, this will be done in the exploratory
analyses (see section 4.2).
To get a first insight in the dynamics of the variables in the sample we used a correlation matrix. A correlation matrix is a useful tool to measure the statistical relationship between two continuous variables. This sample however also includes dichotomous variables and therefore these variables are excluded from the correlation matrix in table 4.3 (see appendix). The correlation matrix shows that the two financial variables ROS and ROVA are strongly positive correlated. Furthermore, labour productivity has a positive correlation with both financial indicators, as was assumed in chapter 3.2.3.
4.2 Explanatory analyses
26 remained to perform business functions from home base (H1), firms that internationally insourced business functions and firms that internationally outsourced business functions (H2) and firms that internationally outsourced supporting business functions and firms that internationally outsourced core business functions (H3). In all three hypothesis we make a statistical examination of two population means, using a t-test. If one group has a significant higher mean over the other this is indicated with the stars. The result of these t-tests are tabulated in tables 4.5.
Table 4.5 t-tests
Dependent variable Home Base Offshoring Insourcing Outsourcing Hybridsourcing Coresourcing
N = 740 N = 352 N = 160 N = 133 N = 52 N = 42 ROS 2011 5,429 4,697 3,837 6,382 * 3,235 6,174 (0,336) (0,559) (0,849) (0,914) (1,135) (1,315) ROVA 2011 9,155 9,915 8,327 13,007 * 6,254 16,995 *** (0,499) (0,857) (1,224) (1,483) (1,828) (2,524) Standard deviations in parentheses.
Significance levels of 10%, 5% and 1% indicated by *, ** and *** respectively.
Regarding the first hypothesis, no significant difference was found regarding financial performance between firms that offshored business functions and firms that remained to perform business functions from home base. The result of the second t-test displays that firms which internationally outsourced business functions have a ROS and ROVA that is significantly higher than firms which internationally insourced activities. The last t-test shows that the ROVA of firms that internationally outsourced supporting activities was significantly higher than the mean of firms that internationally outsourced core activities.
These results do not convincingly underpin that the made assumptions in the hypotheses, that one group performs better than another, are true. These results make it questionable if the usage of on-tailed significance levels is legitimate. Therefore to not jump into conclusion, in the OLS regression, two tailed significance levels will be used.27 In the conducted t-tests it is not possible to control for any firm specific characteristics. In the OLS regression, tabulated in table 4.8, these control variables are included.
27
27 The results presented in table 4.8 reports the relation between the independent variables and the dependent variables ROS (1) and ROVA (2). Important to note is that the coefficients of these variables are added to the constant but do not have an effect on the steepness of the OLS line.
Furthermore, the coefficients displayed in table 4.8 are compared to firms which remained to perform business functions from home base, which were taken as a base in these regressions.
The results show that firms which offshored business functions do not have a significant better performance compared to firms that remain to perform business functions from home base. In contrast, firm that perform activities abroad perform no better, even worse, compared to firm that remained to perform business functions from home base. Firms that offshored have in general a return on sales that is 1.7 percentage point lower than the firms that remained to perform business functions from home base. This was not expected by the first hypothesis, where was presumed that firms that offshore activities perform better than firms that perform activities from home base. Therefore the first hypothesis is rejected.
In the second part of the analyses, firms that offshored business functions were divided into three different groups, insourcing, outsourcing and hybrid sourcing. Especially firms that make use of insourcing and hybrid sourcing perform worse than firms which remained to perform business functions form home base. The coefficients of insourcing and hybrid sourcing are negative regarding the constant of firms that performed activities from home base. Firms that internationally outsourced business functions do not have a significant better nor worse effect compared to firms that perform Table 4.8 OLS regression
(1) (2) (1) (2) (1) (2)
ROS11 ROVA11 ROS11 ROVA11 ROS11 ROVA11
Offshoring -1.701** -1.136 (-2.65) (-1.18) Insourcing -2.425** -2.366 -2.405** -2.407 (-2.70) (-1.88) (-2.68) (-1.91) Hybrid sourcing -3.767** -4.916** -3.739** -4.906** (-2.86) (-2.66) (-2.84) (-2.66) Outsourcing 0.0567 1.859 (0.06) (1.28) Core outsourcing -2.422* -3.796* (-2.20) (-2.09) Supporting outsourcing 0.208 6.359** (0.17) (2.72) Hybrid outsourcing 3.256 4.719 (1.74) (1.70) constant -14.05*** -17.86*** -14.65*** -18.93*** -14.04*** -17.65*** (-5.07) (-4.49) (-5.26) (-4.78) (-5.09) (-4.48) N 1090 1057 1090 1057 1090 1057 adj-R2 0.113 0.110 0.120 0.121 0.128 0.135 t statistics in parentheses * p<0.05, **p<0.01, *** p<0.001
28 activities from home base. The second hypothesis presumed that firms that internationally outsourced business functions perform better than firms that internationally insourced business functions.
However table 4.8 only shows the relationship of both governance mode compared to firms that firms remained to perform business functions from home base. By making use of the contrast function it is possible to compare the conditional mean of two categories, by changing the category which is taken as the base for the regression. Therefore to compare firms that internationally insourced business function and firms that internationally outsourced business functions, firms that internationally insourced business functions are taken as the base in the OLS regression, see table 4.9.
Table 4.9 Contrast function
Outsourcing vs Insourcing Contrast coëfficiënt Standard error
ROVA 2011 4,225* 1,769 (2.39) ROS 2011 2,45* 1,161 (2.11) t statistics in parentheses * p<0.05, **p<0.01, *** p<0.001
These results show that firms which internationally outsourced business functions have a ROS that is in general 2.45 percent point higher, and a ROVA that is 4.22 percent point higher compared to firms that internationally insourced business functions. Therefore the second hypothesis is confirmed.
Finally, in the last analysis we compare the financial performance of firms that internationally outsourced core business functions with firm that outsourced supportive business functions. Table 4.8 shows that firms which internationally outsourced core business functions have a lower financial performance compared to firms which firms which remained to perform activities from home base. Firms that internationally outsourced supportive business functions do not have a significant higher ROS compared with firms that remain to perform business functions from home base. Firms that internationally outsourced supporting business functions do have a higher ROVA than firms which remained to perform business functions from home base. In table 4.10 the conditional means are compared of firms that internationally outsourced core business functions and firms that
internationally outsourced supporting business functions.
Table 4.10 Contrast function
Supporting vs Core business functions Contrast coëfficiënt Standard error
ROVA 2011 10,154*** 2,848 (3.56) ROS 2011 2,63 1,569 (1.68) t statistics in parentheses * p<0.05, **p<0.01, *** p<0.001
