University of Groningen
Faculty of Economics and Business International Business & Management 2014/2015
Nettelbosje 2 9747 AE Groningen The Netherlands
How do governments treat foreign subsidiaries? The
role of institutions between firms and host countries.
March, 2015
MSc Thesis (EBM719A20) IB&M Roel Freriks - s2354799
r.freriks@student.rug.nl Parallelweg 49
9717 KS Groningen +31630582217
2
How do governments treat foreign subsidiaries? The role of institutions
between firms and host countries.
MSc Thesis by R. Freriks – Master program of International Business & Management Faculty of Economics and Business, University of Groningen
March 2015
ABSTRACT.
MNEs operating internationally are actively participating in an international business environment. When doing business in a foreign country it is argued that these foreign firms face an additional cost, which is often referred to as a liability of foreignness. One important source of this is the differential treatment by governments. In this thesis, together with the theory of economic nationalism, the treatment of foreign firms by host country government has been investigated. The role of institutional and regulative foundations have also been taken into account here. Results indicate no significant results regarding differences in respect to a foreign-domestic support gap. However, descriptive results do imply some differences are in place, especially on different country levels. The role of institutions (and regulations) is of huge significance here.
3
CONTENT.
ABSTRACT. ... 2
List of figures and tables. ... 5
List of abbreviations. ... 6
1. INTRODUCTION. ... 7
2. BACKGROUND AND LITERATURE REVIEW. ... 9
2.1. CDBA and LOF. ... 9
2.1.1. Defining CDBA. ... 9
2.1.2. Defining LOF. ... 10
2.2. Economic nationalism as a source of LOF. ... 11
2.2.1. Defining economic nationalism. ... 12
2.2.2. Criticism on economic nationalism. ... 13
2.3. Distance as a source of LOF. ... 14
2.3.1. Geographic distance. ... 14
2.3.2. Cultural distance. ... 15
2.3.3. Institutional distance. ... 16
2.3.4. Regulative distance. ... 18
3. THEORY AND HYPOTHESES. ... 19
3.1. Formulation of hypotheses. ... 19
3.2. Summary of research question and hypotheses. ... 23
4. DATA AND METHOD. ... 24
4.1. Relevant sample size and countries. ... 24
4.2. Relevant variables. ... 25
4.2.1. Dependent variable. ... 25
4.2.2. Independent variables. ... 26
4
4.2.4. Overview of variables. ... 28
5. EMPIRICAL RESULTS. ... 29
5.1. Introduction of the empirics. ... 29
5.2. Descriptive statistics. ... 30
5.3. Baseline statistical results. ... 32
5.3.1. Basic model and control variables. ... 32
5.3.2. Independent variables. ... 32
5.3.3. Interaction effects. ... 33
5.3.4. Analysis on country level. ... 36
5.3.5. Implications for hypotheses. ... 40
5.4. Robustness and Extensions. ... 42
6. DISCUSSION & CONCLUSION. ... 44
6.1. Main findings and theoretical implications. ... 44
6.2. Limitations and future research. ... 46
5
List of figures and tables.
Figure 1 The IBE within an overarching global meta-environment. p.8
Figure 2 Economic nationalism as viewed by Pickel (2003). p.14
Figure 3 Conceptual model. p. 23
Table 1 Respondents (firms) per country. p. 25
Table 2 Overview of variables used. p. 28
Table 3 Descriptive statistics relevant variables. p. 30
Table 4 Breakdown of answers on domestic firm and foreign firm level. p. 31
Table 5 Baseline results regression analysis. p. 35
Table 6a Results means analysis on country level. p. 37
Table 6b Results regression analysis on country level. p. 38
Appendices
Table 7 Breakdown descriptive statistics of all variables. p. 55
Table 8 Robustness check main multiple regression analysis;
different dependent variable. p. 56
Table 9 Robustness check country level; different dependent variable. p. 57
6
List of abbreviations.
BEEPS Business Environment and Enterprise Performance Survey
BES Business Environment Survey
CDBA Costs of doing business abroad
EBRD The European Bank for Reconstruction and Development
FSAs Firm specific advantages
HIID Harvard Institute for International Development
IBE International business environment
LOF Liability of foreignness
MNE(s) Multinational enterprise(s)
NICs Newly industrialized countries
7
1. INTRODUCTION.
Multinational enterprises (MNEs) are actively participating in an international business environment (IBE). Sethi & Guisinger (2002) define this IBE as an unstructured combination
of various elements that differ along several dimensions. These dimensions can for example be geographic, social, political or economic; and are of a direct impact on MNEs operating in the particular environment. This is illustrated in figure 1 (page 8). Furthermore, these dimensions are multifariously entangled and even go beyond political boundaries most of the time. Some elements, such as the geographic dimension, will remain moderately stable over time, whilst others, like the economic dimension (e.g. fluctuating exchange rates), are more volatile and complex (Sethi & Guisinger, 2002).
This thesis is about the treatment of firms by governments, where the distinction between whether a firm is domestic or foreign owned is important. When doing business in a different country, MNEs face (additional) costs, and a liability of foreignness (LOF) in particular. Coping with host governments is one aspect of this; it is one the key components of LOF (Zaheer, 1995). Therefore, it could be argued that governments react differently towards foreign subsidiaries as opposed to domestic firms. These interventions by governments are, however, not a new phenomenon. Especially during the financial crisis that started around 2008, political and economic interventions by governments were becoming more prominent and visible. Governments intervened in a number of ways. For example, governments encouraged more consumption by consumers, protected financial systems and markets by saving or nationalizing banks, or took on substantial debts (Clift & Woll, 2012). As will be shown in the literature review, LOF is a steadily researched topic amongst scholars – but very few authors look further than linking LOF to just firm performance. In this thesis the focus will be on looking if there is a relation between LOF and the way firms are treated by governments in an international setting. Also the role of institutional and regulatory distance will be analysed to see if treatment by governments is affected by this.
8
institutional distance is related to this and if it, in turn, affects how governments treat foreign subsidiaries operating in their country (or IBE).
Figure 1: The IBE within an overarching global meta-environment (taken from Sethi & Guisinger, 2002).
The above information translates into the following research question:
“How do governments treat foreign subsidiaries? What is the role of institutions between firms and host country?”
Based on literature and existing data an answer will be provided to the research question. The data will be taken from Business Environment and Enterprise Performance
Survey (BEEPS) and will be analysed through SPSS (Statistical Package for the Social
Sciences).
9
2. BACKGROUND AND LITERATURE REVIEW.
In this section the theoretical background and literature review linked to the research question will be presented. Firstly, the concepts and definitions of various variables are given. These variables are costs of doing abroad (CDBA) and LOF in section 2.1, economic nationalism in section 2.2 and institutions in section 2.3. The research question fundamentally is a combination of LOF and economic nationalism. As will be explained below, LOF is a broader concept and will be used as an introduction to economic nationalism. Lastly, the third variable that is incorporated in the research question is the level of institutions. This will be closely related to (institutional) distance. As will be discussed, there are several forms of distance. 2.1. CDBA and LOF.
MNEs that engage in international business differentiate themselves from enterprises that stay domestic. MNEs do this by seeking to originate and exploit certain substantial competitive advantages by using assets and resources in more than one country. Despite the competitive advantages this can give, additional costs and substantial risks are faced when engaging in international business (McDougall, Oviatt & Shrader, 2003; Oviatt & McDougall, 1994). These additional costs can be classified as the costs of doing business abroad. The CDBA can however be seen as a broader concept and several authors (Eden & Miller, 2001; Kostova & Zaheer, 1999; Zaheer, 1995) identify liability of foreignness (LOF) as one of the critical components of CDBA. This has generated debate amongst scholars, because some authors (Luo & Mezias, 2002) believe that CDBA equalize LOF, whereas others (Sethi & Guisinger, 2002; Zaheer, 1995; Kostova & Zaheer, 1999), as mentioned above, claim LOF is a critical, but different part of CDBA (Sethi & Judge, 2009).
2.1.1. Defining CDBA.
10
hand, CDBA derives from an economic perspective towards MNE theory. LOF, on the other hand, is based more so on the sociological and institutional analysis. In other words: “While the costs of doing business abroad focus on market-driven economic costs, I see the liability of foreignness as focusing on the more social costs of access and acceptance” (Zaheer, 2002: p352).
2.1.2. Defining LOF.
In this thesis the definition by Zaheer (1995) will be used. Zaheer defines LOF as “the costs of doing business abroad that result in a competitive disadvantage for a MNE subunit ...broadly defined as all additional costs a firm operating in a market overseas incurs that a local firm would not incur” (Zaheer, 1995: p342-343). This is exactly what differentiates MNEs that are doing business abroad from enterprises that stay domestic. If MNEs did not face these additional costs or disadvantages it could be argued that it would be just business instead of international business. There clearly are challenges and barriers to overcome when engaging in business abroad, as can be seen in the very extensive related to this. Different scholars (Petersen & Pedersen, 2002; Luo, Shenkar & Nyaw, 2002; Hennart, Roehl & Zeng, 2002; Sethi & Guisinger, 2002) have analysed this dyad through different perspectives; this for example being organizational learning, socio-economic theory, evolutionary dynamics, resource-based view respectively (Sethi & Judge, 2009).
Zaheer (1995, page 343) argues that there are four different sources of liability of foreignness:
- Spatial distance, such as transportation, traveling, coordination costs; - Unfamiliarity with the local environment;
- Home-country environments that create costs;
- And lastly; differential treatment by the host country environment/government.
This thesis will focus on the latter, the differential treatment of foreign subsidiaries by the government of host countries, as it connects the research question to the term of economic nationalism. This will be elaborated upon in section 2.2.
11
socials costs as “those actions of business firms which have harmful effects on others.” These social costs arise from the unfamiliarity, relational and discriminatory hazards that foreign firms face which domestic firms do not. Such costs are inherently due to uncertainty and are likely to persist over time. It is argued that these hazards are driven by the concept of
institutional distance (Eden & Miller, 2004). This will be further elaborated upon in section
2.3. Social costs and economic costs are likely to be greater in markets where there is more uncertainty about the other party (Temouri & Jones, 2014). London & Hart (2004) argue that firms without a capacity to appreciate and create a social value may struggle to overcome the LOF.
As a concluding remark, LOF is the accumulated result of all the various elements in the direct business environment of a firm throughout its entire foreign operations. The associated costs are therefore incurred not only during the entry stage into a foreign market, but may persevere during the entire presence of the firm’s operations in the host country (Sethi & Guisinger, 2002). LOF does not seem to be related to firm age, but instead to firm experience of doing business in a certain foreign country (Eden & Miller, 2004). LOF can persist also with experienced firms if the managers do not actively engage in local host environment learning (Petersen & Pedersen, 2002).
2.2. Economic nationalism as a source of LOF.
12
2.2.1. Defining economic nationalism.
To fully understand the concept of economic nationalism it is useful to firstly understand what is meant by the term ‘nation’. Anderson (2006) defines a nation as “an imagined community –
and imagined as both inherently limited and sovereign” (Anderson, 2006: p7). To explain this
into further detail: Nations are considered imagined, because inhabitants of a nation, either big or small, will never know most of the other people living inside the same nation (Anderson, 2006). Yet everyone in the nation still identifies themselves with each other for example culturally, linguistically or with sports celebrations. Nations are limited in the sense that every nation has boundaries, essentially separating itself from other nations (Anderson, 2006).
Economic nationalism according to Fölster (2009) is “protectionism for their own
country’s benefit and intervention in their domestic economy to pursue own interests”
(Fölster, 2009: p14). This implies that every nation is self-interested and thus own interests and opportunities are more important than the interests of any other nation. Implied here is the notion that own interests should be protected and safeguarded against at any cost. This extent of nationalism can vary between countries, within countries itself, and also over time (Dogan, 1994). Transaction costs, taxes, commissions and costs of collecting information are examples of explicit costs that are associated with doing business, investing abroad and thus economic nationalism. Next to these explicit costs, firms or investors also face implicit costs. These are costs that are the result of disadvantages in relation to domestic firms. Both explicit and implicit costs have a negative effect on net returns (Ahearne, Griever & Warnock, 2004).
13
nationalism. Economic nationalism is seen as a dynamic concept (Zhang & He, 2014). An important notion of this nationalism is that the actions taken by nations are to accomplish and endure sovereignty and economic development in the home environment (Gellner, 1983; Gilpin, 1987; Miller, 2000). Governments want to serve the national interest (Zhang & He, 2014) and economic nationalism is a crucial part of the available (in)formal institutions in a nation (Beland & Lecours, 2005). It also has an effect on enabling or holding back foreign direct investment as imposed by foreign investors (Dikova, Sahib & Van Witteloostuijn, 2010).
2.2.2. Criticism on economic nationalism.
However, there is also criticism on economic nationalism. It was argued earlier that nations are considered imagined (Anderson, 2006). Solt (2011) elaborates upon this notion, by arguing that nationalism therefore does not exist - it is a myth. This myth entails that individual persons are meant to be together in a united and similar community which is comprised and epitomized by an own state (Solt, 2011). Helleiner (2002), for example, argues that economic nationalism is ambiguous and obscure. Economic nationalism is also described as “the notion that a nation requires a strong, independent economy to liberate itself and compete with other nations” (Shechter, 2008, page 571). More criticism comes from Pickel (2003), who claims that economic nationalism is outdated and a too narrow economical doctrine. Pickel (2003) makes several comments. The first one is that there are different varieties and forms of economic nationalism; one of these is economic liberalism, which actually can be viewed as sort of the opposite of economic nationalism. Secondly, economic nationalism is a specific part of nationalism theory overall and economic theory is insufficient for studying this. Thirdly, “nationalism as economic ideology and as political action occurs
within particular national economic, political and cultural systems. Economic nationalism as an idea or policy and cannot be explained or evaluated in general terms” (Pickel, 2003, page
14 Figure 2: Economic nationalism as viewed by Pickel (2003).
Taking the criticism into account the conclusion can be logically made that some form of economic nationalism exists. This implies that some governments can have a bias for domestic firms and can choose to treat foreign subsidiary differently. It can even be the case that some foreign subsidiaries are also given preferential treatment compared to other foreign subsidiaries. This would then be some form of selective economic nationalism.
2.3. Distance as a source of LOF.
The premise that different types of distance can negatively influence cross-border economic activities has been widely discussed in international trade and international business studies. This section will describe three different kinds of distances. These are geographical distance, cultural distance and institutional distance.
2.3.1. Geographic distance.
15
Instead of measuring geographic distance by the actual amount of kilometres it is best measured by the ease and time it costs of relocating. Locations with a direct air infrastructure link and great physical distance can for example still be easier to reach than locations that are in closer proximity but have little to no infrastructure available (Holmstrom, Conchúir, Ågerfalk & Fitzgerald, 2006). Geographic distance according to Arora & Fosfuri (2000) is a component of cultural distance and is strongly correlated to it (Cantwell, Dunning & Lundan, 2010), meaning that the larger the geographical distance is, the larger the cultural distance is. This excludes, for instance, colonial history, as Australia has a relatively small cultural distance compared to England (United Kingdom), whilst the geographic distance is considerably large.
2.3.2. Cultural distance.
Cultural distance reviews differences in cultures of the home country environment and host countries where firms can set up subsidiaries (Johanson & Vahlne, 1977; Kogut & Singh, 1988). Probably the most known definition put forward by Hofstede (1980), who distinguishes four dimensions: power distance; uncertainty avoidance; masculinity versus feminism; and individualism versus collectivism. Despite being one of the foundations of defining cultural distance, Hofstede’s work has been criticised. Cultural distance encompasses differences in the most dominant norms and values in a country that determines the culture. Norms can be defined as the behaviour and customs and, together with prominent values, this establishes a culture. Kogut & Singh (1988) have argued that this differs per country. These norms, values and practices are incorporated in the prevalent aspects of a country’s culture. These aspects can, for example, be languages, political systems, living habits, religions, routines, and business practices (Benito & Gripsrud, 1992; Li & Guisinger, 1992; Mariotti & Piscitello, 1995; Morosini, Shane, & Singh, 1998; Shenkar, 2001; Em, 2011). Cultural distance increases complexity and uncertainty. It also increases costs related to training, monitoring, control (Tihanyi, Griffith, & Russell, 2005), and communication (Vachani, 1991) and thus has an effect on foreign firms operating in a culturally distant environment. It is more problematic to comprehend business partners and to detect market opportunities when facing higher cultural distance (Simonin, 1999; Em, 2011).
16
and cognitive distance. The debate is about the role of culture in an institution-based view of global strategy. How can all the kind of different dimensions of distance be measured and is there a correct, applicable unit of analysis? How do all these difference matter for strategy? Institutional measures may better explain differences in firm performance than cultural distance measures (Peng & Pleggenkuhle-Miles, 2009).
2.3.3. Institutional distance.
In section 2.1 it was argued that LOF is focusing more on social costs of access and acceptance (Zaheer, 2002). It was also suggested that these social costs originate from the different hazards (these being unfamiliarity, discriminatory and relational) that foreign firms face, as opposed to domestic firms. Such costs are inherently due to uncertainty and are likely to persist over time. These hazards are driven by the concept of institutional distance, distance here means when said settings in other countries differ from the home environment. (Eden & Miller, 2004).
To elaborate upon these hazards; Unfamiliarity costs result from a foreign firm’s lack of knowledge of, or experience in, the host country. This comes down to domestic firms having an advantage over foreign firms when operating in the home country. It is important to stress that this inexperience appears to have nothing to do with firm age. Discrimination
hazards arise from the discriminatory treatment by the host government in a variety of ways
ranging from discriminatory taxation to discriminatory procurement. It can also surface in discriminatory treatment by customers in the host country who may prefer a local product out of nationalistic reasons or dislike products from a foreign country for historical reasons. To a great extent, the discrimination hazards arise from the “legitimacy deficit” (Schmidt & Sofka, 2009) faced by a firm in a foreign country. Relational hazards relate to the higher costs that a foreign firm would incur with respect to both internal organization and external market transactions (Auger, Devinney, Louviere & Burke, 2010).
17
provides an alternative explanation for the (social) behaviour of MNEs and gives stability and meaning (Scott, 1995).
This construct of institutional distance has been linked to the following two aspects of MNE operations, namely:
- The establishment of legitimacy in the host country (Kostova & Zaheer, 1999)
- The transfer of strategic orientations and organizational practices from the parent firm to the foreign subsidiary (Kostova, 1999).
As derived from this theoretical concept, institutional distance basically is the extent of similarity (or dissimilarity) between the regulatory, cognitive, and normative institutions between countries (Scott, 1995; Kostova, 1996; Eden & Miller, 2004). Regulative in this context means the rules and laws that are present in a country to warrant and guarantee stability and structure. Normative relates to social values, customs, culture and norms, whereas cognitive refers to cognitive structures that are deemed normal and taken for granted (Scott, 1995; Yiu & Makino, 2002). In this thesis the focus will be on the regulative pillar. It should be noted that cognitive distance has consequences for decision making and problem solving, but is as opposite to other types of distance on an individual level (Nooteboom, 2000; Em, 2011). However, this is beyond the scope of this thesis and shall therefore not be included in any hypotheses. Normative distance is also less relevant to the main research question.
18
2.3.4. Regulative distance.
Firms are embedded in political environments (Zukin & DiMaggio, 1990), which is why the regulatory dimension explains how LOF exists in capital markets. This dimension is built from the official laws, rules and policies that structure and support the participants and structure transactions in product and capital markets (Bell, Filatotchev & Rasheed, 2012). The regulatory dimension of a host country can have an impact on a foreign firm that is present in that specific country, depending on whether this distance (difference in level of regulatory dimension between host and home country) is large or small. In essence, the regulatory dimension can provide support to firms, for example government regulations and laws that give structure to competition within a certain industry (Barnett & Carroll, 1995).
19
3. THEORY AND HYPOTHESES.
This section formulates the hypotheses concerning the differential treatment experienced by multinational enterprises (MNEs) in their host countries. The key theory and hypotheses are visually summarized in section 3.2, in the conceptual model. The function of this conceptual model is to explain the main elements to be studied (key factors, concepts or variables) and the presumed relationships among them (Miles and Huberman, 1994).
Following the background discussion in the previous section and as mentioned before, the focus is on one of the four sources of LOF, as identified by Zaheer (1995), namely differential treatment by government (see also, Hymer 1960). However, in existing theory, as outlined in the literature review, the reasoning behind LOF is not researched in depth, as LOF is viewed as something that is just there. At the simplest level, LOF is the costs of engaging in international business. In principle, both the home and the host country governments have possibilities to put MNEs at a disadvantage compared to domestic firms (Hymer, 1960). However, the focus in this thesis is on the way host-country governments treat MNE subsidiaries originating from other countries. One of the reasons for this is that home-country governments can prevent MNEs from using some of their firm specific advantages (FSAs), particularly when they involve technologies that are considered of strategic-military importance (Hymer, 1960). The reason for this behaviour can be one of self-interest, because they want to provide protection for domestic firms from losing market share or competitive advantage. This translates back to the theory of economic nationalism, as discussed. However, it should be noted that this could also be the other way around. If countries lack a clear abundance in for example an industry sector or a specific type of technology that is important for an industry, then they may welcome foreign MNE subsidiaries and treat them the same as domestic counterparts. That said, this thesis will focus on differential treatment. This shall be elaborated upon in hypothesis 1 and 2.
3.1. Formulation of hypotheses.
20
Hypothesis 1. On average, host-country governments provide less support to
foreign firms than to domestic firms.
As just indicated, it is important to stress that this hypothesis is very general and does not consider any further contingencies that are likely to affect how governments treat foreign subsidiaries. These contingencies will be considered in the forthcoming hypotheses. A first contingency to take into account is that not all governments may be a source of LOF to the exact same extent, implying that there are different ways of looking towards foreign MNE subsidiaries. Indeed, many governments could be rather supportive towards foreign MNEs, by acknowledging them as an important source of knowledge or as beneficial foreign capital inflow. Hence, these foreign MNE subsidiaries are seen as a beneficial and advantageous entity which could adjust the behaviour of the government towards them accordingly. The generally positive experiences with the East-Asian newly industrialized countries (NICs) such as South Korea; Hong Kong; Singapore; Taiwan, supports this perspective. All these countries have benefited tremendously from foreign MNEs outsourcing part of their value-added activities to these countries (e.g., Hobday, 1995). Hence, the following closely-related set of hypotheses is proposed:
Hypothesis 2a. Some host-country governments are more supportive of foreign firms
than other governments are.
Hypothesis 2b. The difference in government support for foreign firms vis-à-vis
domestic firms (i.e. the foreign-domestic the support gap) varies across countries.
Hypothesis 2a captures the general idea and assumption that not all governments treat foreign firms alike. However, as this thesis is concerned with exploring the idea that governments are a source of LOF, hypothesis 2b is more relevant. This is because the entire idea and concept of LOF is that foreign firms are at a disadvantage compared to domestic firms. Hypothesis 2b therefore encompasses the idea that, in terms of government support, this gap is not the same in all host countries. Hypothesis 2b directly compares domestic firms with foreign firms whereas hypothesis 2a compares government support on a country level.
21
crucial for the future growth and development of China. On the contrary, in other industries domestic Chinese firms should be enabled to develop themselves, without foreign firms snatching up resources such as experienced production employees. Indeed, this provides another link to the theory of economic nationalism, where behaviour by governments can put foreign firms at a disadvantage compared to domestic firms. Another example could be that due to an abundance of a specific resource or strength of a specific sector, governments may be more discriminative towards foreign firms operating in these same environments. This behaviour could for example derive from protectionism (or in a broader term; economic nationalism) towards domestic firms. Therefore the question can be raised if certain industries are indeed more important to governments than other industries, which may result in a differential treatment towards foreign firms to favour domestic firms. Again, a contingency such as this one suggests differences in governmental support for foreign firms and, particularly, differences in the foreign-domestic support gap, this time across industries. The accompanying hypothesis therefore reads the following:
Hypothesis 3. The foreign-domestic support gap varies between manufacturing and
service industries.
22
lack of data – which will be explained in the limitations in section 6.2. The accompanying hypothesis is stated as follows, where again attention has been limited to the foreign-domestic gap in government support:
Hypothesis 4. The foreign-domestic support gap increases when the institutional
foundation of a host-country is higher.
As argued, institutional distance is considered important, so therefore to analyse this deeper it makes sense to also formulate a hypothesis about the regulative distance. Regulative distance is as discussed, part of the institutional distance. Incorporating regulations in the analysis will elaborate the role of institutions. By doing so, the analysis regarding institutional distance will be explored in depth and, as such, government behaviour can hopefully be better understood. The regulatory distance hypothesis will try to describe the ease of accessing government information, such as time dealing with government and the effect of government interventions on policies. By imposing policies or engage in interventions, governments may try to structure the market in the way they would like to see it. The question is; how much influence do governments actually have? All governments will have to comply by international laws and rules; else they may risk sanctions by international organizations and/or other countries. Nonetheless, certain policies can be constructed in such a way to make it significantly harder for foreign firms to operate in a specific (host) market. I argue that the predictability of changes in government policy-making also plays a role here. This is especially true for countries that are developing, or in a transition from communism to open market and may experience unpredictable changes as a result. Apart from interventions, policies and the level of predictability, another contingency here is time dealing with government. How helpful are governments towards firms from a different country operating in their business environment? Does this reflect in the way of how predictable changes in regulations are, or how easy it is to obtain information regarding regulations and laws? In short, do these firms have easy access to government resources and information? Or does it cost these firms more time to adequately get the information they need and do governments play a role in this? To what extent do governments involve with foreign MNE subsidiaries by imposing interventions on policies and influencing time dealing?
Therefore the final hypothesis is formulated as:
Hypothesis 5. The foreign-domestic support gap increases when the regulative
23
Together, these five sets of hypotheses collectively encompass the research question in a direct and specific way. This is neglecting the fact that, as mentioned earlier, separate hypotheses about cross-industry differences and cultural/geographic distance are not taken into account, to save space and to remain true to the core of this thesis.
3.2. Summary of research question and hypotheses.
Figure 3: Conceptual model
24
4. DATA AND METHOD.
Section 4 describes how the hypotheses will be analysed with the use of available data. This research uses a deductive approach and started with theory based on its closed research question. A closed research question starts with thinking up a theory about the research question. This, in turn, is narrowed down to specific hypotheses that can be tested through observations or data. This serves to confirm or dismantle an existing theory. Based on the deduction perspective, a closed research question is answered based on an idea that is worked out before the actual observations or data analysis takes place.1 To test the hypotheses data from BEEPS 1999 will be analysed with the use of SPSS. General information about BEEPS can be found in appendix 1.
4.1. Relevant sample size and countries.
To be able to provide an answer to the research question, decisions have to be made regarding which variables are most relevant for each hypothesis. The initial sample size of the BEEPS 1999 database was 4104 firms in total. As not all necessary data is available for every firm and country, the sample size needs to be narrowed down. Furthermore, to generate a more cohesive sample all Eastern-Asian/Eurasian countries are filtered out of the dataset. This consequently has an effect on the total amount of responses. Ultimately, after the filtering out of the above mentioned countries and variables, this leave a final sample size of 2988 firms, covering 19 countries.
The analysed countries are: Albania; Bosnia and Herzegovina; Bulgaria; Belarus; Croatia; Czech Republic; Estonia; Hungary; Latvia; Lithuania; (FYR) Macedonia; Moldova; Poland; Romania; Russia; Slovakia; Slovenia; Turkey and Ukraine.
The majority of this geographic area can be seen as a transition area, which means that most countries are economies that changed from centrally planned economies to free market economies. This is especially relevant to take into account since the database used is from 1999. This transition process lead to economic liberalization and therefore an elimination of imposed trade barriers. A breakdown of the countries and amount of domestic and foreign firms per country is visualized in table 1.
25
Table 1: Respondents (firms) per country.
Country Frequency Domestic Firms Foreign Firms Percent
Albania 119 106 13 4.0
Bosnia and Herzegovina 177 168 9 5.9
Bulgaria 117 109 8 3.9 Belarus 128 118 10 4.3 Croatia 122 117 5 4.1 Czech Republic 122 95 27 4.1 Estonia 104 91 13 3.5 Hungary 126 105 21 4.2 Latvia 145 118 27 4.9 Lithuania 101 98 3 3.4 Moldova 129 114 15 4.3 Poland 221 198 23 7.4 Romania 114 99 15 3.8 Russia 516 495 21 17.3 Slovakia 120 113 7 4.0 Slovenia 121 108 13 4.0 Turkey 142 136 6 4.8 Ukraine 239 220 19 8.0 FYR Macedonia 125 114 11 4.2 Totals 2988 2722 266 100% 4.2. Relevant variables.
This section will describe the dependent and independent variables that are used to answer the research question of this thesis. Control variables are also explained here.
4.2.1. Dependent variable.
Helpfulness of government
26
4.2.2. Independent variables.
There are four independent variables that can be distinguished. Control of firm
To be able to analyse the various hypotheses, domestic and foreign firms have to be distinguished from each other. To do so, a dummy variable is created in order to label and differentiate the control of the firm. This dummy variable is created out of several other variables deriving from the BEEPS dataset, where respondents could indicate whether their firm is foreign owned or domestically owned. Consequently, this transforms this particular variable into a dichotomous nominal independent variable. The aim of this independent variable is to analyse whether there are differences in treatment by governments between domestic and foreign firms or not. This can therefore be seen as the main independent variable in this analysis.
Type of industry
A second independent variable is the particular industry the firm operates in. The BEEPS dataset has a variable where firms can indicate in which industry they operate in, with a total of 14 different industries. A dummy variable is created and these 14 industries are divided into two categories; manufacturing and services. The manufacturing category encompasses industries like farming, mining, repair and construction. The service category encompasses industries such as financial services, personal services, business services, retail and transportation.
Levels of institutions (institutional distance)
27
Likewise, the second variable is “Rating of the efficiency of government in delivering services”. Here, the possible answers range from “very efficient (1)” to “very inefficient (5)”. The corresponding dummy variable transforms these answers into “efficient” (1) and “inefficient” (2). This variable tells something about the strength of institutions, since when a government is very efficient it will be beneficial for firms operating in that particular country and vice versa. This variable together with the first variable should complement the analysis of whether the level of institutions plays a role in the helpfulness of governments.
Level of regulations (regulative distance)
This last independent variable is very similar to the level of institutions; as explained in the theory it is essentially a part of the level of institutions. Variables used in the dataset are “Information on the laws and regulations affecting my firm is easy to obtain” and “Interpretations of regulations affecting my firm are consistent and predictable”. Both variables have answers based on a 5-point scale, ranging from “strongly agree” (1) to “strongly disagree” (5). Both variables have been transformed into a separate dummy variable with two possible answers; “agree” (1) and “disagree” (2).
4.2.3. Control variables.
Two control variables, firm age and firm size, are added to test whether this has an effect on average treatment by governments.
Firm age
The first control variable will be firm age. A new dummy variable is computed to determine the firm age of the companies. One question in the dataset is regarding the years the firm was found (variable q6yr). Since the data is from 1999, the age is determined simply by subtracting 1999 with the value of q6yr. The firms are furthermore divided in two groups, those with less than 10 year existence are classified as “young firms” (1) and those older than 10 years are classified as “mature firms” (2). The purpose of this is to test the effect of firm age on how firms are being treated by (host)governments.
Firm size (amount of employees)
28
“medium/large” (50 or more employees). The purpose of this control variable is to test if firm size plays a role in how firms are being treated by (host) governments.
4.2.4. Overview of variables.
An overview of all relevant variables is given in table 2.
Table 2: Overview of variables used.
Name and type Variable Description Type of data
DV #1a q20a (BEEPS) Helpfulness central / national
government Ordinal, but tested as continuous
IV A1 (BEEPS) Country of respondent. Categorical
IDV #1 Control of firm Control of firm, domestic or foreign?
Dummy variable.
0 = domestic; 1 = foreign.
IDV #2 Industry Main area of activity, industry Dummy variable.
1 = manufacturing; 2 = services.
IDV #3a InfoObtain Information on the laws and
regulations affecting my firm is easy to obtain
Dummy variable of q15 (BEEPS). 1 = agree;2=disagree
IDV #4a Predictability
How predictable are changes in rules laws or regulations affecting your firm
Dummy variable of q36 (BEEPS). 1 = predictable; 2=unpredictable
CV #1 Firm age
Age of the firm based on 1999 minus the answer given on what year the firm was found.
Dummy variable. 1 = Young <10 years; 2 = Mature ≥ 10 years.
CV #2 Firm size Amount of full-time employees. Dummy variable. 1=small;
2=medium/large.
Extensions
DV #1b q21a (BEEPS) Helpfulness central / national
government Ordinal, but tested as continuous
IDV #3b Interpretations
Interpretations of regulations affecting my firm are consistent and predictable
Dummy variable of q16a (BEEPS). 1 = agree;2=disagree
IDV #4b Effectiveness Rating of the efficiency of government in delivering services
Dummy variable of q48b (BEEPS). 1 = efficient; 2 = inefficient.
DV: dependent variable; IV: independent variable; IDV: independent dummy variable; CV: control variable
29
5. EMPIRICAL RESULTS.
Section five presents the empirical results of the analysis. Before that will be done, an explanation on how the hypotheses were tested is provided – since there are some substantial factors that have to be taken into account.
5.1. Introduction of the empirics.
The empirical analysis will firstly provide a brief descriptive analysis. The descriptive analysis is added to provide more information and to complement non-significant results for some of the hypotheses. For the statistical tests the hypotheses will be tested by a multiple linear regression. This however requires some explanation. As described, there is one dependent variable, classified as ordinal data. The main independent dummy variable (control of firm) is classified as dichotomous nominal data. Logically the relationship between these variables can hence be tested with an ANOVA one-sided test. A one-way analysis of variance (ANOVA) is used when there is a categorical independent variable (two or more categories) and a normally distributed interval dependent variable. However, since regression is a more suitable test for this type of analysis in this particular field, an exception will be made. By analysing the dependent variable as though it is a continuous variable, a linear multiple regression analysis can still be performed. Despite this implying a minor limitation, in practice the norm is to treat variables like ‘helpfulness of government’ as a continuous variables, as this does not introduce serious biases.
30 5.2. Descriptive statistics.
Two tables are presented in this section. Table 3 provides a breakdown of descriptive statistics of the dependent variable, independent variables and control variables. Reported are the frequency, percentage, means, standard deviation (SD) and minimum and maximum value of the variable in the dataset. Also included are the frequencies and percentages of the possible answers for each variable. Table 4 will do the same. But to complement the descriptive analysis, table 4 will provide a breakdown of all variables on domestic and foreign firm level.
Table 3: Descriptive statistics relevant variables
Frequency % Means SD Min Max
Helpfulness central/national government (DV) 3.75 1.157 1 5
Helpful 463 15.5%
Neutral 846 28.3%
Unhelpful 1679 56.2%
Control of Firm (IV) .0890 .28482 0 1
Domestic 2722 91.1%
Foreign 266 8.9%
Industry (IV) 1.4759 .49950 1 2
Manufacturing 1566 52.4%
Services 1422 47.6%
Predictability of changes (IV) 1.7175 .45027 1 2
Predictable 844 28.2%
Unpredictable 2144 71.8%
Government effectiveness (IV) 1.6827 .46549 1 2
Efficient 948 31.7%
Inefficient 2040 68.3%
Information on laws and regulations affecting firm is easy to obtain (IV)
1.2979 .45739 1 2
Agree 2098 70.2%
Disagree 890 29.8%
Interpretations of regulations affecting firm are consistent and predictable (IV)
1.4796 .49967 1 2 Agree 1555 52% Disagree 1433 48% Firm age (CV) 1.2155 .41126 1 2 Young 2344 78.4% Mature 644 21.6% Firm size (CV) 1.5184 .49974 1 2 Small 1439 48.2% Medium/Large 1549 51.8%
31
Table 4: Breakdown of answers on domestic firm and foreign firm level.
Domestic firms Foreign firms
Frequency % Frequency %
Helpfulness central/national government (DV)
Helpful 434 15.9% 29 10.9% Neutral 743 27.3% 103 38.7% Unhelpful 1545 56.8% 134 50.4% Industry (IV) Manufacturing 1444 53% 122 45.9% Services 1278 47% 144 54.1%
Predictability of changes (IV)
Predictable 768 28.2% 76 28.6%
Unpredictable 1954 71.8% 190 71.4%
Government effectiveness (IV)
Efficient 868 31.9% 80 30.1%
Inefficient 1854 68.1% 186 69.9%
Information on laws and regulations affecting firm is easy to obtain (IV)
Agree 1900 69.8% 198 74.4%
Disagree 822 30.2% 68 25.6%
Interpretations of regulations affecting firm are consistent and predictable (IV)
Agree 1433 52.6% 122 45.9% Disagree 1289 47.4% 144 54.1% Firm age (CV) Young 2111 77.6% 233 87.6% Mature 611 22.4% 33 12.4% Firm size (CV) Small 1321 48.5% 118 44.4% Medium/Large 1401 51.5% 148 55.6%
Key conclusions based on this table are that compared to domestic firms relatively less foreign firms classify the government as helpful. Surprisingly, also relatively less foreign firms rate the host-government as unhelpful compared to their domestic counterparts. This implies that governments on average are slightly less helpful towards domestic firms.
Regarding the control variables, the majority of both domestic (77.6%) and foreign (87.6%) firms are young (<10 years) whereas firm size roughly has an equal distribution.
32 5.3. Baseline statistical results.
The statistical analysis is performed in the form of a multiple linear regressions. Baseline results can be found in table 5. As can be seen six different models are tested. R, R², F-test scores and significance of every model are reported in this table as well. It can be concluded that every model seems to be significant (p < 0.001). Despite the relatively low scores for R and R², meaning that a relatively low amount of variance is explained in the outcome variable, the scores are increasing in every model, indicating a slightly better fit. Every model will be explained in more detail below.
5.3.1. Basic model and control variables.
The first model is the basic model, which is basically the foundation of all models and represents the relationship between the dependent variable (helpfulness of central/national government) and the main independent variable (control of firm). This reflects on hypothesis 1. Control variables are also added in this first model.
Results show a slight negative effect on the control of the firm (IDV #1), albeit this is not significant (β = -.086, p > 0.10). Thus, it can be concluded that whether a firm is domestic or foreign has no statistically significant effect on the helpfulness of the government. Both control variables firm age and firm size are strongly significant (p < 0.001) and both seem to have a negative relationship on the dependent variable: β = -.392 for firm age and β = -.330 for firm size. This implies that the older and the larger a firm is, the more helpful governments are expected to be.2
5.3.2. Independent variables.
In model 2 the industry variable is included (IDV #2), which reflects on hypothesis 3. Again, control of firm is insignificant (β = -076, p > 0.1). Industry however seems to be significant (p < 0.05) and the coefficients report a negative relationship (β = -.134). This implies that the helpfulness of governments decreases for firms operating in the service industry, when compared to firms operating in the manufacturing industry. The control variables firm age and size appear to remain negatively significant (p < 0.001), with values of β = -.402 and -.372 respectively. The implication for this is explained in model 1.
2 Helpfulness is measured on a 5-point scale, with 1 being most helpful and 5 most unhelpful. A negative
33
Model 3 then adds the two variables that are an indication of the level of institutions, which reflects on hypothesis 4. Here, both variables are strongly significant (p < 0.001). The first variable (IDV #3a), the predictability of changes in rules/laws that affect the firm, has a positive β-value of .322. The second variable (IDV #3b); the rating of the efficiency of government in delivering services, has a positive β-value of .668. This implies that an increase of either independent variable would increase the dependent variable. Meaning that a higher level of institutions in a country leads to governments on average being more helpful towards firms. Control of firm remains insignificant (p > 0.10). Industry in this model is slightly less significant, but the p-value is still lower than 0.10, with a negative coefficient of -.08. Both control variables firm age and firm size remain statistically significant (p < 0.001) and again both have negative coefficients of β = -.300 and β = -.288 respectively.
Model 4 incorporates the final independent variables (for hypothesis 5) and is similar to model 3. These new independent variables are an indication of the regulatory environment instead of the broader institutional level. Both variables are strongly significant (p < 0.001). The first variable (IDV #4a), “Information on the laws and regulations affecting my firm is easy to obtain”, has a positive effect on the dependent variable (β =.241). The second variable (#IDV #4b), “interpretations of regulations affecting my firm are consistent and predictable” is also positively related (β = .178). This implies that an increase of either independent variables would increase the dependent variable, meaning that a higher level of regulative environment leads to governments being more helpful. For the other independent variables; control of firm remains insignificant (β = -.083, p > 0.1). Industry remains significant and negatively related (β = -.087, p > 0.05). The independent variables for the levels of institutions remain strongly significant in model 4 as well (p < 0.001). Both remain positively related to the dependent variable (β = .601 and β = .241 respectively). Again, the control variables firm age and firm size remain statistically significant (p < 0.001) and both have negative coefficients of β = -.282 and -.253 respectively.
5.3.3. Interaction effects.
34
of activity, institutional distance, and regulatory distance on helpfulness of government are moderated by whether the firm is foreign or domestic. So therefore all other independent variables are combined with the main independent variable, which is control of the firm. As can be seen in model 6 of table 5, all results of interaction effects are insignificant in this model (p > 0.1). Therefore the main effects of the independent variables in this model are more relevant and thus these results are reported below instead.
35 Table 5: Baseline results regression analysis
Model 1
Basic model + control variables
Model 2 + Industry Model 3 + Institutional distance Model 4 + Regulatory distance Model 5 +Interaction effects
β t Sig. β t Sig. β t Sig. β t Sig. β t Sig.
Constant 4.732 (.080) 59.346 .000 5.003 (.119) 41.969 .000 2.999 (.159) 18.836 .000 2.547 (.167) 15.274 .000 2.552 (.170) 15.020 .000 Firm age (CV) -.392 (.053) -7.423 .000*** -.402 (.053) -7.598 .000*** -.300 (.050) -1.214 .000*** -.282 (.050) -5.652 .000*** -.283 (.050) -5.672 .000*** Firm size (CV) -.330 (.043) -7.609 .000*** -.372 (.045) -8.183 .000*** -.288 (.043) -5.960 .000*** -253 (.043) -5.875 .000*** -.250 (.043) -5.791 .000*** IDV #1 -.086 (.073) -1,182 .237 -.076 (.073) -1.039 .299 -.084 (.069) -6.670 .225 -.083 (.068) -1.212 .226 -.271 (.395) -.685 .494 IDV #2 -.134 (.044) -3.050 .002** -.081 (.042) -1.941 .052* -.087 (.041) -2.115 .034** -.088 (.043) -2.057 .040** IDV #3a .322 (.044) 7.262 .000*** .276 (.044) 6.227 .000*** .268 (.046) 5.885 .000*** IDV #3b .668 (.043) 15.461 .000*** .601 (.043) 13.825 .000*** .594 (.045) 13.324 .000*** IDV #4a .241 (.047) 5.136 .000*** .251 (.048) 5.205 .000*** IDV #4b. .178 (.043) 4.183 .000*** .182 (.044) 4.126 .000*** IDV #1 x IDV #2 (.138) .001 .004 .997
IDV #1 x IDV #3a .051
(.066)
.780 .436
IDV #1 x IDV #3b .049
(.071)
.691 .490
IDV #1 x IDV #4a -.060
(.055)
-1.087 .277
IDV #1 x IDV #4b -.015
(.056)
-.275 .783
Additional statistics Model 1 Model 2 Model 3 Model 4 Model 5
R .228 .234 .391 .415 .416
Adj. R² .051 .054 .151 .170 .173
F 54.604 43.393 89.855 77.418 47.790
Sig. .000 .000 .000 .000 .000
(Standard errors in parentheses); * p < 0.10; ** p < 0.05; *** p < 0.001.
IDV #1: Control of Firm IDV #2: Industry
IDV #3a: Predictability of changes in rules/laws that affect the firm IDV #3b: Government effectiveness
IDV #4a: Information on laws and regulations affecting firm is easy to obtain
IDV #4b:Interpretations of regulations affecting firm are consistent and predictable
36
5.3.4. Analysis on country level.
As indicated before, hypotheses 2a and 2b are tested separately from the other hypotheses. Because of the amount of countries, a total of 19 separate regression tests per type of government have been performed. Important to note here is that in the whole sample size there are only 266 foreign firms distinguished. Consequently, these 266 firms are spread over the 19 countries, which results in only a few foreign firms per country. This imposes a limitation on the analysis, since it is less accurate.
Two tables are important for the analysis on country level. Table 6a presents a basic descriptive analysis based on means of the dependent variable. The results of the regression analysis are reported in table 6b.
Results descriptive and mean analysis
Before analysing the variables and values by a statistical test it may be insightful to compare the means per country.
The means are reported in table 6a under “means”. As can be seen, in the following eight countries domestic firms classify their government as more unhelpful compared to
foreign firms (mean domestic firms > mean foreign firms): Albania; Croatia; Czech Republic;
Estonia; Poland; Romania; Turkey; FYR Macedonia. Vice versa, in the following 11 countries domestic firms classify their government as more helpful than foreign firms: Bosnia and Herzegovina; Bulgaria; Belarus; Hungary; Latvia; Lithuania; Moldova; Russia; Slovakia; Slovenia; Ukraine.
37 Table 6a: Results means analysis on country level.
Country N Sample Means Country N Sample Means
Albania 106 Domestic 3.50 Moldova 114 Domestic 4.12
13 Foreign 3.15 15 Foreign 4.33
119 Total 3.46 129 Total 4.15
Bosnia and Herzegovina
168 Domestic 3.20 Poland 198 Domestic 3.37
9 Foreign 3.78 23 Foreign 3.26
177 Total 3.23 221 Total 3.36
Bulgaria 109 Domestic 4.06 Romania 99 Domestic 3.69
8 Foreign 4.13 15 Foreign 3.20
117 Total 4.06 114 Total 3.62
Belarus 118 Domestic 3.64 Russia 495 Domestic 4.10
10 Foreign 3.70 21 Foreign 4.24
128 Total 3.65 516 Total 4.11
Croatia 117 Domestic 3.17 Slovakia 113 Domestic 3.35
5 Foreign 3.00 7 Foreign 3.57
122 Total 3.16 120 Total 3.36
Czech Republic 95 Domestic 4.03 Slovenia 108 Domestic 3.56
27 Foreign 3.74 13 Foreign 3.62
122 Total 3.97 121 Total 3.57
Estonia 91 Domestic 3.10 Turkey 136 Domestic 3.67
13 Foreign 2.85 6 Foreign 3.00
104 Total 3.07 142 Total 3.64
Hungary 105 Domestic 3.13 Ukraine 220 Domestic 4.23
21 Foreign 3.14 19 Foreign 4.53
126 Total 3.13 239 Total 4.26
Latvia 118 Domestic 3.75 FYR Macedonia 114 Domestic 4.18
27 Foreign 3.89 11 Foreign 4.09
145 Total 3.78 125 Total 4.18
Lithuania 98 Domestic 4.31 Moldova 114 Domestic 4.12
3 Foreign 5.00 15 Foreign 4.33
38
Table 6b: Regression analysis on country level. Model 1 Basic model
+ control variables
Model 2 + Countries
Model 3 + Interaction effects
β t Sig. β t Sig. β t Sig.
Constant 4.722 59.557 .000 5.025 56.155 .000 5.024 55.247 .000
Firm age (CV) -.387 -7.351 .000*** -.242 -4.464 .000*** -.249 -4.549 .000***
Firm size (CV) -.333 -7.673 .000*** -.406 -9.501 .000*** -.402 -9.307 .000***
Albania -.753 -6.923 .000*** -.718 -6.259 .000***
Bosnia & Herz. -.939 -9.862 .000*** -.956 -9.769 .000***
Bulgaria -.074 -.673 .501 -.095 -.835 .404 Belarus -.317 -2.973 .003** -.295 -2.657 .008** Croatia -.762 -6.897 .000*** -.743 -6.562 .000*** Czech Republic -.243 -2.260 .024** -.216 -1.796 .073* Estonia -1.023 -8.946 .000*** -.997 -8.205 .000*** Hungary -1.064 -10,005 .000*** -1.075 -9.332 .000*** Latvia -.341 -3.409 .001** -.372 -3.407 .001** Lithuania .034 .294 .769 .012 .105 .917 Moldova .069 .660 .510 .073 .658 .511 Poland -.713 -8.240 .000*** -.699 -7.710 .000*** Romania -.568 -5.140 .000*** -.507 -4.303 .000*** Slovenia -.814 -7.538 .000*** -.822 -7.388 .000*** Slovakia -.435 -4.017 .000*** -.431 -3.785 .000*** Turkey -.335 -3.187 .001** -.305 -2.837 .005** Ukraine .108 1.294 .195 .094 1.085 .278 FYR Macedonia -.006 -.059 .953 .015 .133 .894 IDV #1 x ALB -.284 -.908 .364 IDV #1 x BH .450 1.233 .218 IDV #1 x BUL .374 .956 .339 IDV #1 x BEL -.217 -.617 .537 IDV #1 x HRV -.308 -.631 .528 IDV #1 x CZE -,105 -.452 .651 IDV #1 x EST -.182 -.576 .564 IDV #1 x HUN .091 .357 .721 IDV #1 x LVA .184 .808 .419 IDV #1 x LIT .883 1.414 .158 IDV #1 x MOL -.005 -.018 .986 IDV #1 x POL -.092 -.394 .694 IDV #1 x ROM -.435 -1.473 .141 IDV #1 x RUS .073 .308 .758 IDV #1 x SLV .193 .465 .642 IDV #1 x SLK .010 .031 .975 IDV #1 x TUR -.541 -1.217 .224 IDV #1 x UKR .221 .867 .386 IDV #1 x FYRM -.187 -.557 .578
Additionalstatistics Model 1 Model 2 Model 3
R .227 .401 .406
Adj. R² .051 .156 .154
F 81.197 28.509 14.910
Sig. .000 .000 .000
39 Results statistical analysis
The most important conclusion from above table is that the main independent variable reflecting the control of the firm being domestic or foreign, is insignificant (p > 0.10) in all three models. Besides that, the control variables firm age and firm size are significant (p < 0.001) and negatively related to the dependent variable in all models. This implies that on average in every analysed country, the older and the larger a firm is, the more helpful governments are expected to be.
Model 2 shows that almost all countries (except Bulgaria and FYR Macedonia) are significant (p < 0.10) and negatively related to the dependent variable. Lithuania, Moldova and the Ukraine are the only three countries with a positive coefficient, although these results are all insignificant (p > 0.10). These results imply that on average, in countries with significant results that the helpfulness of government is considered as unhelpful. This varies slightly per country.
40
5.3.5. Implications for hypotheses.
The above described results have thus consequences for the formulated hypotheses. Hypothesis 1 stated that host-country governments on average will provide less support to foreign firms than to domestic firms. The coefficient reports a slightly negative relationship, meaning that when a firm is foreign the government can indeed be expected to be less helpful. However, as the result is not significant, hypothesis 1 is not supported.
Hypothesis 2a and 2b then argued that some host-country governments are more supportive of foreign firms than other governments are (2a). And that the difference in government support for foreign firms vis-à-vis domestic firms (i.e. the foreign-domestic the support gap) varies across countries (2b). It already is explained that these hypotheses were tested separately from the other hypotheses. Descriptive statistics imply that there indeed seems to be difference in government support across countries, as can be seen through the various reported means. The statistical tests however provide no significant results and henceforth both hypothesis 2a and 2b are not supported.
The effect of industry differences were analysed in hypothesis 3. The hypothesis states that the foreign-domestic support gap varies between manufacturing and service industries. According to the significant test results there indeed seems to be an effect on the type of industry. However, as control of firm is insignificant no statistical proof is available that supports the notion that the foreign-domestic support gap varies between industries. Consequently, this means that no support is found for hypothesis 3.
Hypothesis 4 argued that the foreign-domestic support gap increases when the
institutional foundations of a host-country are higher. It seems to be the case that governments
indeed are more helpful towards firms when the institutions in a country are predictable and efficient. However, hence no significant relationship is present on whether a firm is domestic or foreign, hypothesis 4 can therefore not be supported.
41
found that the foreign-domestic support gap indeed changes. Therefore, no support is found for hypothesis 5.
42 5.4. Robustness and Extensions.
In the dataset not only the helpfulness of central and national governments was measured, but also the helpfulness of local and regional governments. Therefore, as an extension and robustness check, the helpfulness of local and regional governments was added to the statistical testing. It can be argued that central and national governments can be characterized as the most important level of government. Generally the central government will overlap with local and regional government. Nevertheless, local and regional government may give a different or complementary result. This is especially relevant for firms located in an important specific region, for example in an area with a lot of industrial activity. Results can be found in table 8, appendix 3. Conclusions are that regarding the control variables firm age and firm size the results are approximately the same. In all models both control variables remain significant (p < 0.001) and shows a negative relationship with the dependent variable.
Control of firm has an even higher level p-values in all models compared to the original regression analysis and is therefore even less significant. Industry on the other hand has roughly the same outcome. Industry is significant (p < 0.10 for model 3 and p < 0.05 for the other models) and has a slight negative relationship on the dependent variable. The variables reflecting institutional distance and regulative distance are, just like the original regression, significant and positively related. Lastly, also in this regression model interaction effects are not significant with helpfulness of local/regional governments as dependent variable.
A second extension regarding institutional distance (hypothesis 4) and regulatory distance (hypothesis 5) was to analyse a second variable for both hypotheses. These are incorporated in the original analysis and in the robustness check with the different dependent variable.
43
44
6. DISCUSSION & CONCLUSION.
The hypotheses were tested by using a sample of 2988 firms from 19 countries in Central and Eastern Europe, taken from the BEEPS dataset of 1999. All these countries can be classified as transition economies (especially at the time of measurement).
6.1. Main findings and theoretical implications.
For hypothesis 1 the results of the regression analysis imply that host-country governments on average do not provide less support to foreign firms than to domestic firms. However, based on the descriptive analysis it appears that both domestic and foreign firms receive little support from governments. It is important to note that this does not deny the fact that foreign firms face a liability of foreignness, it merely does not confirm it. It appears that both domestic and foreign firms classify governments as being unhelpful. As noted in the literature review, it is already proven by scholars (i.e. Dinc & Erel, 2013) that economic nationalism in Europe exists. Dinc & Erel (2013) further implied that level of affinity for a foreign country and weak governments influences the level of economic nationalism, with the latter being most relevant for the transitional European countries analysed in this thesis.
