• No results found

ASTER T HESIS M

N/A
N/A
Protected

Academic year: 2021

Share "ASTER T HESIS M"

Copied!
51
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

M

ASTER

T

HESIS

ABOUT

T

HE

I

MPACT OF VICARIOUS

L

EARNING FROM

N

EIGHBOURING

F

IRMS ON

F

OREIGN

D

IRECT

I

NVESTMENT

D

ECISIONS

MS

C

I

NTERNATIONAL

B

USINESS AND

M

ANAGEMENT

F

ACULTY OF

E

CONOMICS AND

B

USINESS

U

NIVERSITY OF

G

RONINGEN

N

AME

:

K

ATHARINA

K

ARALL

S

TUDENT

N

UMBER

:

3573575

D

ATE OF

S

UBMISSION

:

17

TH OF

J

UNE

,

2019

S

UPERVISOR

:

D

R

.

J.

C

ANELLO

(2)

A

BSTRACT

Making successful strategic decisions is a reoccurring and fundamental topic for organisations, regardless of their past experience and size. Hence, it is inevitable that firms have reliable external resources to base their decisions on. One possibility of an external resource is organisational learning, in particular vicarious learning. The focus of this study is on vicarious learning among geographically proximate located peers. To narrow the learning and mimicry process further down, an emphasis is placed on the strategic decision of initiating foreign direct investment. A sample of 13,273 Italian firms operating in the manufacturing sectors of textile, wearing apparel and leather products has been selected for the analysis. Their FDI decisions have been studied by using a logistic regression model. Furthermore, given the fact that small firms suffer from financial and managerial constraints as well as less foreign experience, a distinction is made between small and large firms. The findings of this study demonstrate that vicarious learning from nearby located peers hampers FDI decisions of large firms. However, the model indicates that the effect remains relevant for small firms, resulting from their several internationalisation constraints. Thus, the findings of this study contribute to current research of organisational learning, provide theoretical and practical implications, and offer some important limitations for future research on observing other organisations to make crucial strategic decisions.

Key words: Foreign Direct Investment (FDI), strategic decisions, organisational learning,

(3)

A

CKNOWLEDGMENT

(4)

T

ABLE OF

C

ONTENT ABSTRACT ... I ACKNOWLEDGMENT ... II LIST OF FIGURES ... V LIST OF TABLES ... V LIST OF ABBREVIATIONS ... V 1 INTRODUCTION ... 1 2 THEORETICAL FRAMEWORK ... 4 2.1 INTERNATIONALISATION THEORIES ... 4 2.2 INTERNATIONALISATION OBSTACLES ... 5 2.2.1 Internal Internationalisation Obstacles ... 5 2.2.2 External Internationalisation Barriers ... 6 2.3 SPATIAL SPILLOVERS ... 7 2.4 ORGANISATIONAL LEARNING ... 8 2.5 SMALL FIRMS AND THEIR LIABILITIES ... 11 2.6 CONCEPTUAL FRAMEWORK ... 14 3 RESEARCH DESIGN ... 14 3.1 DATA COLLECTION ... 15 3.2 SAMPLE ... 15 3.3 MEASUREMENTS ... 16 3.3.1 Dependent Variable ... 16 3.3.2 Independent Variable ... 16 3.3.3 Moderator Variable ... 16 3.3.4 Control Variables ... 17 3.4 METHODOLOGY ... 20 3.5 ROBUSTNESS TEST ... 21 4 RESULTS ... 21 4.1 DESCRIPTIVE STATISTICS ... 22 4.2 MULTICOLLINEARITY ... 23

4.3 RESULTS OF LOGISTIC REGRESSION ... 25

4.4 RESULTS OF ROBUSTNESS TEST ... 29

5 DISCUSSION ... 30

6 CONCLUSION ... 33

(5)

6.2 PRACTICAL IMPLICATIONS ... 35

6.3 LIMITATIONS AND FUTURE RESEARCH ... 35 7 REFERENCES ... 37 8 APPENDICES ... I

(6)

L

IST OF FIGURES

Figure 1: Conceptual Framework ... 14

L

IST OF TABLES Table 1: Descriptive Statistics ... 22

Table 2: Pearson correlations matrix ... 24

Table 3: Results of the binary logistic regression ... 26

Table 4: Output of the linear probability model ... 29

L

IST OF ABBREVIATIONS FDI Foreign Direct Investment MNE Multinational Enterprise

NACE Nomenclature statistique des activités économiques dans la Communauté européenne NUTS Nomenclature des unités territoriales statistiques

SME Small and medium-sized enterprise US United States

(7)

1 I

NTRODUCTION

It has long been recognised by international business scientists that internationalisation offers significant benefits and opportunities for companies. Anderson and Gatignon (1986), Buckley and Casson (1981), Hennart (1982), and Rugman (1981) have all argued that firms tend to locate themselves abroad to either exploit foreign advantages or to overcome domestic market imperfections. In addition to this early understanding of the importance of internationalisation, Blomström and Kokko (1998) have stated that investing in foreign markets commonly precipitates greater growth as opposed to solely operating within the domestic market, and this results in economies of scale. In addition, they contend that acquiring modern technology and gaining insights into management and marketing skills are further advantages of an internationalisation. Additional recent research from Blonigen (2005) supports these findings and demonstrates that firms opt to engage in foreign direct investment (henceforth FDI) because it provides intangible assets that can be regarded as public goods. A recent insight from Ning and Wang (2018) indicates that these assets give firms remarkable competitive advantages over others.

(8)

Spatial spillovers between firms in the same region are potential means for organisations to overcome these barriers. Unfortunately, the main literature of spatial spillovers in connection with internationalisation focuses on the relationship between the investing firm and the host market. Since the research of spatial spillovers concerning the relationship between home and host country is extensive, only some are mentioned here to give an impression how vast it is. For instance, Hassine, Boudier and Mathieu (2017) have delineated the possible spatial spillovers in their paper. Additionally, Driffield and Love (2007) mention the effects of FDI on the domestic productivity. Moreover, Albornoz, Cole, Elliot and Ercaloni (2009), Huber (2008) and Perkins and Neumayer (2009) further investigate the effects of internationalisation spillovers for the firm in the host country.

Spatial spillovers of organisations located in the same region are rarely mentioned in existing literature, but that would be a suitable solution to overcome the several hurdles that internationalisation creates. Therefore, since spillovers occurring in the home countries are less inspected, there is a gap in current research. This thesis addresses this gap by providing a different angle on spatial spillovers. It combines existing literature regarding spatial spillovers and interorganisational learning. More specifically, it addresses the gap by examining the role of vicarious learning among geographically proximate located firms in influencing internationalisation-related decisions. This is a reasonable approach because it builds upon existing research concerning network relationships. Several authors have already acknowledged that network relationships generally offer connections and opportunities (Coviello, & Munro, 1995; Coviello, & Munro 1997; Ellis, 2011; Johanson, & Vahlne, 2003). More importantly, they enable the access to necessary resources to internationalise (Zahra, Ireland, & Hitt, 2000), and provide information about how to compete successfully abroad (Sharma, & Blomstermo, 2003).

Additionally, the thesis verifies whether this process of relying on vicarious learning from nearby located organisations affects small and large firms equally. This is an important endeavour, since small firms suffer from financial and managerial constraints (Buckley, 1989) and thus face greater uncertainties (Mutinelli, & Piscitello, 1998).

(9)

on FDI decisions, this thesis contains a sample of 13,723 Italian firms in the manufacturing sectors of textile, wearing apparel and leather products and combines various firm-level and province-level factors. The data for this purpose was downloaded from ORBIS and the Italian Bureau of Statistics. While many recent studies of spatial spillovers have focused on the relationship between the host and home country, this thesis examines the spatial spillovers resulting from learning between organisations in the same region. A moderating effect of the firm size will further demonstrate whether businesses restricted by their size must particularly rely on vicarious learning. Employing a quantitative approach and conducting a logistic regression analysis, the underlying research question for this project is as follows:

“Does vicarious learning from neighbouring peers influence the FDI decision of manufacturing firms? Is this type of learning more relevant for smaller firms?”

As mentioned above, this study focuses on Italian manufacturing businesses in the sectors of textile, wearing apparel and leather production. The constantly changing consumer preferences and new adaptations in technology render it difficult for firms in the manufacturing sectors to remain competitive (Bell, Crick, & Young, 2004). Additionally, particularly the textile and clothing industry in the European Union is severely affected by the increased competition from overseas. Technical innovation, focusing on niche markets and outsourcing can provide possible solutions to tackle the fierce overseas competition (Taplin, 2006). All these measures result in serious and inevitable restructuring of the organisations, producing an enhanced need for internationalisation. Consequently, these sectors were selected to demonstrate that initiating FDI can be an effective means to survive the necessary restructuring due to the rigorous competition.

(10)

2 T

HEORETICAL

F

RAMEWORK

As a result of a constantly growing international environment, firms have to invest in domestic and foreign markets to stay competitive (Blomström, & Kokko, 1998). However, to make a decision regarding whether to internationalise or not, necessary know-how for successful internationalisation must be gained.If, due to unfamiliarity with the host context, this knowledge is not anchored in the company yet, uncertainty occurs. As a result of being unfamiliar with the host context, a firm faces higher burdens in estimating additional costs involved in doing business abroad, which can be summarised as the liability of foreignness (Zaheer, 1995). A reduction of perceived costs and risks can be achieved through vicarious learning, since firms are able to gain crucial insights by observing the actions from other firms in the field (Huber, 1991). Business actions by others in the same industry and country offer useful information about opportunities concerning growth and investment (Bastos, & Greve, 2003). When entering a foreign market, it is inevitable for firms to know about how to internationally manage distributing, marketing and servicing of their products (Blomström, & Kokko, 1998). Due to spatial spillovers and organisational learning among nearby businesses, other firms are able to overcome this, as well as many other hurdles. Successful prior actions of organisations with similar characteristics, such as being from the same home country are used by inexperienced companies to proceed in the same kind of business activities (Henisz, & Delios, 2001; Nelson, & Winter, 1982). For instance, Head, Ries and Swenson (1995), Shaver and Flyer (2000) and Chung and Song (2004) have demonstrated that this was true for Japanese firms who located their plants in the United States where other Japanese firms were already located. Hence, they followed the same decisions of companies that had certain attributes in common with them. However, negative externalities from fierce competition may outweigh the positive effects of co-locating with other firms (Zhu, Eden, Miller, Thomas, & Fields, 2012).

2.1 INTERNATIONALISATION THEORIES

In the last decades scientists have developed useful internationalisation theories for firms to rely on. Two widely used theories are the Uppsala model from Johanson and Vahlne and the OLI model from Dunning.

(11)

organisation’s lack of market knowledge and being at mercy of constant internal and external changes. Accordingly, new opportunities and obstacles are emerging simultaneously. This results in a slow and incremental internationalisation decision-making process where firms firstly expand to close-by locations before going further away, so they are able to avoid risks (Johanson, & Vahlne, 1977).

Dunning’s OLI model is a holistic framework which identifies and evaluates substantial factors that influence FDI decisions (Dunning, 1988). The abbreviation OLI stands for ownership, location and internalisation. Ownership refers to the competitive advantage organisations have over others. So, it implies that if a firm has better competitive advantages in relation to others, the more likely it is that the business will engage in production abroad. Location relates to the advantages of the region where the firm wants to expand to. Suggesting that immobile, natural or created endowments in the host country used together with the firm’s own competitive advantage will have a larger benefit than operating solely in the home market. Internalisation focuses on the evaluation of performing an activity internally or due to locational advantages through outsourcing (Dunning, 2000).

Even though these theories are strongly embedded and omnipresent in international business, some organisations still struggle to apply them accurately because they experience a lack of crucial market information and international experience, and both provoke uncertainty.

2.2 INTERNATIONALISATION OBSTACLES

2.2.1 Internal Internationalisation Obstacles

(12)

on the costs of doing business abroad too (Mutinelli, & Piscitello, 1998). Therefore, drastic consequences may occur when firms suffer from a lack of host country-related information and international experience. In addition to these uncertainty sources, there are more external factors affecting the internationalisation uncertainty.

2.2.2 External Internationalisation Barriers

Exchange rate effects are the first external factor that influences a company’s FDI decision (Blonigen, 2005). Froot and Stein (1991) and Klein and Rosengren (1994) have proved that a depreciation of a currency influences FDI decisions by a sample of U.S. businesses. Increased firm wealth is a result of a currency appropriation, since it provides the firm with low-cost funds in foreign countries that experience the depreciation of their currency (Blonigen, 2005). Conversely, when a firm’s FDI decision is motivated by the acquisition of assets which do not involve currency transactions, an exchange rate appreciation of that foreign currency lowers the price in the foreign currency but does not lower the nominal return (Blonigen, 1997). Secondly, taxes influence organisations’ FDI decisions, because higher taxes discourage FDI engagement. Taxes from the local and host country determine FDI decisions, so firms have to deal with double taxation, which further complicates expectations of the FDI. Thirdly, the institutional quality in the local and host country shape whether or not FDI will be successful, in particular if the host country is less developed. Trade protection is the fourth external factor influencing FDI, since a higher trade protection in the host country makes FDI more expensive and companies in this case will most likely substitute FDI with exports. Lastly, trade effects are an important external determinant for FDI decisions (Blonigen, 2005).

(13)

have to face uncertainties in internationalising. The precariousness in connection with initiating FDI boosts the perceived risks of the decision-makers. The chances are likely that inappropriate decisions incur or foreign relations are managed badly (Weichmann, & Pringle, 1979; Mariotti, & Piscitello, 1995). Managers’ perceived uncertainty and insecurities can hinder a successful internationalisation process. To overcome this, managers are obliged to obtain helpful and requisite know-how. Spatial spillovers are certainly one appropriate means to gain that kind of needed knowledge. Moreover, vicarious learning and mimetic isomorphism can applicable solutions too.

2.3 SPATIAL SPILLOVERS

As a result of locally bounded and territorially rooted interactions among actors, institutions and local economies, spatial spillovers occur (Capello, 2009). They are the transfer of essential knowledge from one firm to another. These are pure externalities where the receiver of spillovers benefits from the knowledge but does not compensate the sender for it. Spatial spillovers may arise out of the premise that knowledge, gained by one firm usually spreads within an area and does not belong to the organisation solely (Capello, 2009). The knowledge contained in spatial spillovers may spill to other firms in the same area through multiple transmission channels. Imitation effects, labour mobility and the creation of linkages are examples of the multiple transmission channels, and create backward and forward linkages between companies (Fatima, 2017).

(14)

organisations’ survival rates (Ingram, & Baum, 1997; Chuang, & Baum, 2003). They further have indicated that interfirm knowledge spillovers in connection with access to labour and suppliers are one reason for businesses to grow (Marshall, 1920; Chung, & Alcácer, 2002). Knowledge spillovers are measured by combining the link of the productivity growth of a firm and the innovative activity of other firms, but only under the premise that these two businesses have a functional relationship (Caragliu, & Nijkamp, 2012).

Spatial spillovers may be shared voluntarily or involuntarily. If they are voluntarily shared, the receiving company must be well-informed about organisational learning to successfully find and implement information about other firms’ strategies.

2.4 ORGANISATIONAL LEARNING

To get the most out of knowledge spillovers, organisational learning needs to be facilitated. Relevant and important sources of interorganisational imitation opportunities signal possibilities to grow (Bastos, & Greve, 2003). Generally, imitating business actions of organisations in the same sector is a usual business behaviour (Henisz, & Delios, 2001). By applying mimicry organisations are able to justify their decisions for themselves and crucial stakeholders (Guillén, 2002). It is an important ability for firms to evaluate and use outside knowledge, and to relate it with prior knowledge. At the easiest level, examples for prior knowledge are basic skills or sharing the same language. Without prior knowledge, the firm is not able to find, assimilate and apply new knowledge (Cohen, & Levinthal, 1990). The possession of valuable prior knowledge enables an organisation to be more creative and find linkages and associations with new know-how that was not considered before (Bradshaw, Langley, & Simon, 1983; Simon, 1985). Assimilating novel knowledge is particularly successful if the firm is boasted with a diversified background (Cohen, & Levinthal, 1990). Nevertheless, it is imperative to consider that evolving knowledge requires reasonable time and effort (Harlow, 1949), especially if learning takes place in novel situations. Moreover, spreading the knowledge to geographically ultimate located subsidiaries needs to be considered too (Cohen, & Levinthal, 1990).

(15)

of special interest, thus, there is a focus on the acquisition of knowledge. This concept includes five major dimensions.

(1) Congenital learning refers to the knowledge a firm has already before creating a new venue. It is a combination of knowledge inherited at a firm’s conception and additional know-how acquired prior the incorporation (Huber, 1991). Firms with fewer prior internationalisation experience benefit more from congenital learning (Bruneel, Yli-Renko, & Clarysse, 2010), since it allows to leapfrog steps of internationalisation and to expand rapidly and fearless (Casillas, Barbero, & Sapienza, 2015).

(2) Experiential learning happens consciously through direct experience, and unconsciously, through unintended by-products of operating with others. It is a crucial means to reduce perceived risks and to encourage expansions into new markets (Casillas et al., 2015). Firms concurrently learn how to refine routines and to discriminate useless ones (Levitt, & March, 1988). Since the past is not the best predictor of the future, there are some difficulties with experiential learning. Constant learning leads to inertia within the organisation and makes new experiments for successful learning rather improbable (Levitt, & March, 1988).

(3) Grafted learning is the knowledge acquisition of firms by hiring new and experienced members, or acquiring whole new organisations (Huber, 1991). When the hired manager or founder of a venture has substantial prior international experience, geographic diversity implies a more beneficial effect on learning (Yeoh, 2004). This does not guarantee that an organisation will experience more effective learning, since this rather depends on the complementarity of newly acquired and already existing knowledge. Benefits of gaining knowledge through grafted learning are higher awareness and confidence in decisions, and better knowledge of foreign markets (Casillas et al., 2015).

(4) Searching and noticing, this know-how and insight acquisition can be split in three forms: scanning, focused research, and performance monitoring (Huber, 1991). It includes gathering information about other countries, markets and institutions. The process of searching information is part of explicit and objective learning (Casillas et al., 2015). Baum, Li and Usher (2000) describe organisational search as a process where firms try to find and apply alternatives for current business actions within their external environment.

(16)

since these firms are in the same cultural and business environment, so they have common business practices and organisational structures. Therefore, this offers a simple means to imitate business behaviours (Stinchcombe, 1965; Guillén, 2001). Firms gain information through “consultants, professional meetings, trade shows, publications, vendors and suppliers, and, in less competitive environments, networks of professionals” (Huber, 1991: 96). Merely observing and imitating will not be rewarding in fast-changing and competitive business environments (Huber, 1991). However, following vicarious learning results in a reduction of costs, and involves tacit and explicit learning (Casillas et al., 2015). Firms learn by observing business behaviours of others who share similar attributes. For instance, one of these attributes may be the same home country, as they face similar obstacles when investing abroad. Due to this, it makes it easier for the learning firm to draw conclusions from previous actions among other firms (Cohen, & Levinthal, 1990). Moreover, vicarious experience enables firms to recognise challenges in advance and to prepare for possible difficulties. Hence, firms with vicarious learning experience are able to avoid similar mistakes (Kim, & Miner, 2007), and it allows them to imitate successful survival strategies (Terlaak, & Gong, 2008).

Vicarious learning is more or less related to isomorphism, because it similarly helps to increase know-how by imitating others. It is a constraining process where one firm is duplicating the actions of another firm who faces comparable environmental conditions. Hence, a firm adjusts and modifies its behaviour to be better compatible with environmental characteristics (Hawley, 1968). Hannan and Freeman (1977) have stated that firms learn about appropriate responses and adapt their business actions akin to others. In particular, mimetic isomorphism arises when firms experience uncertain situations, and as a remedy begin to imitate others (DiMaggio, & Powell, 1983). As mentioned in economic theory, information is always an appropriate answer to uncertainty (Arrow, 1972). Examples for uncertain circumstances are poorly understood technologies (March, & Olsen, 1976), unclear defined goals, or environmental uncertainty. In these situations, firms tend to try to model themselves on others and hunt for valuable solutions with little effort or expense (Cyert, & March, 1963). However, the degree of mimicry is not solely dependent on benefits the firm gains from other’s experience, it further depends on gains or losses other firms have experienced resulting from their behaviour (DiMaggio, & Powell, 1983).

(17)

1996). Managers devote increased attention to geographically proximate located firms because they are part of an indispensable informal network relationship. The information contained in informal networks flows through “local trade shows, conferences, seminars, communication with personnel from nearby research institutes, organised social activities, or from employees switching companies” (Fernhaber & Li, 2013:7). This information is considered to be more significant, because it is mainly of tacit nature (Fernhaber, & Li, 2013; Audretsch, 1998). The study of Birkinshaw and Hood (2000) has demonstrated that international opportunities and knowledge lie in the informal network relationships among neighbouring organisations. The most protruding benefit of these informal networks is that firms located nearby are more willing to offer resources than distant peers would be (Fernhaber, & Li, 2013).

For all these reasons, the first hypotheses of this thesis proposes that already ongoing FDI engagement among firms in the same Italian region has an influential impact on others to engage in it too. This is stated in

Hypothesis 1. The likelihood of engaging in FDI is higher when a larger number of local firms are already engaged in FDI in the focal firms’ home region.

Making satisfied decisions concerning internationalisation intentions entails various hurdles and barriers, and results in significant uncertainty (Mutinelli, & Piscitello, 1998). However, small firms already experience even greater constraints and liabilities during their usual way of doing business. Therefore, internationalising constitutes a more extensively perceived risk for small firms than it does for large firms, and this makes the role of learning particularly relevant for small firms.

2.5 SMALL FIRMS AND THEIR LIABILITIES

The importance of small and medium-sized enterprises (henceforth SMEs) is maybe overlooked for the global economy, but nearly 80 per cent of the global economic growth stems from SMEs (Jutla, Bodorik, & Dhaliqal, 2002). There are two general obstacles regarding internationalisation in comparison to larger firms. Namely, smaller firms face shortages of financial funds, and experience time constraints (Buckley, 1979). Therefore, this thesis placed a special emphasis on small firms and their impediments.

(18)

clothing, or furniture (White, 1983). Generally, SMEs restrict their businesses to the region of their location, but presently become more active in different regions (Matlay, Ruzzier, Hisrich, & Antoncic, 2006). Since small firms are restricted in their size, they have to face barriers which hinder their growth potential (Mutinelli, & Piscitello, 1998). Buckley (1989) has divided the broad barriers into two sub-categories. (1) Internal barriers, which refer to managerial constraints, such as a lack of time and managerial skills. The lack of sufficient time has a negative effect on FDI decisions too. Due to the time constraints of managers, short cuts in decision making processes are necessary, and there may not be adequate time to properly evaluate all possible strategies. But then, making a strategic decision about initiating FDI is usually a time intense endeavour and would require precise judgements. Likewise, small firms face a constraint in skilled managers, and this constitutes another severe impediment. It is rather improbable that small firms employ special executives to manage firms’ international operations, or to have a hierarchy of managers who evaluate crucial and complex decisions. (2) External barriers, which refer to unfavourable market conditions or institutional agreements (Buckley, 1989).

Since SMEs deteriorate from liability of smallness and constraints regarding finance and personnel, learning and mimicry is particularly relevant for them (Kale, & Arditi, 1998). Financial constraints refer to difficulties about getting external funds, resulting in growth obstacles (Buckley, 1989). Thus, their internationalisation success mainly depends on gaining the necessary knowledge, and overcoming the impediments (Bruneel et al., 2010; Sapienza, De Clercq, & Sandberg, 2005). Another constraint which especially smaller and less experienced firms face, are cost asymmetries when obtaining information (Mutinelli, & Piscitello, 1998). Aside from those liabilities, some small firms already experience difficulties in their respective market and fight for survival to keep their business position there (Singh, Garg, & Deshmukh, 2010). So, the international inexperience and their financial and managerial constraints result in additional situations of uncertainty (Mutinelli, & Piscitello, 1998).

(19)

Due to all their liabilities, small firms are mostly able to gain experience solely in their home country. This may be insufficient to manage operations successfully in foreign countries. Inappropriate decisions abroad are the result of international inexperience, stemming from an unawareness of foreign economic and cultural environments. As a result of the unfamiliarity with the host context, mistakes occur when negotiating with foreign suppliers, customers, banks or local authorities (Weichmann, & Pringle, 1979; Mariotti, & Piscitello, 1995). Furthermore, managers may wrongly estimate risks and expected returns of foreign operations (Davidson, 1980; Caves, 1982). One possible explanation for those mistakes may be the lack of sufficient time for superior time-intense management activities.

All the above explained constraints which SMEs undergo lead to a greater internationalisation risk. Another reason for that is the higher proportion of resources devoted to a single FDI, resulting in immense costs if the FDI fails (Buckley, 1989).

Albeit all the negative aspects internationalisation may bring about, there are still various crucial reasons why SMEs should still pursue FDI. Small firms need to prevail over those constraints and internationalise, because this strengthens their competitiveness, chance of survival (Zhou, & Wu, 2014; Lee, Kelley, Lee, Lee, & Lee, 2012; Xuemei, 2011), and positively affects their productivity (Coviello, McDougall, & Oviatt, 2011). Moreover, small firms gain growth and economies of scale, become stronger competitors to others, and experience a superior performance, and this is all achieved by internationalisation (Toulova, Votaupalova, & Kubickova, 2015; Coviello et al., 2011; Ruigrok, & Wagner, 2003).

Accordingly, if small firms take the risk and internationalise, the benefits will outweigh the perceived insecurities and uncertainties (Pangarkar, 2008). In spite of this, they still must be able to overcome the liabilities and gather necessary knowledge for a thriving FDI engagement. So, small businesses are decidedly dependent on vicarious learning among neighbouring peers. This leads to the second hypothesis of this study, which proposes a particular importance of vicarious learning among small firms.

(20)

2.6 CONCEPTUAL FRAMEWORK

The above argumentation and derivation of the hypotheses is figuratively summarised in the conceptual framework, depicted in figure 1. It is designed by the author for the purpose of the master thesis.

3 R

ESEARCH

D

ESIGN

The following section of the thesis describes the research design that investigates whether or not vicarious learning of organisations has an impact on FDI decisions. The main idea is to study FDI strategies by selecting a sample of organisations that are operating domestically in the manufacturing sectors of textile, wearing apparel and leather products. Likewise, the impact of firm-level and province-level factors influencing FDI decisions of organisations in 2017 is explored. The focus is on vicarious learning, proxied by the number of firms that already engaged in FDI in 2013. The two databases used for the empirical analysis are the number of companies not engaging in FDI in 2013, and the number of companies that either did or did not engage in FDI in 2017. These two databases were merged with the “BVD-ID-Number”, an individual identification number businesses. This provided the following information about the firms. Firstly, it shows how many businesses started FDI between 2013 and 2017. Secondly, it demonstrates the number of companies still not engaging in FDI in 2017. Thirdly, it indicates the firm-specific characteristics of the firms in 2013.

First of all, the sample and database used for this study is explained. Subsequently, the dependent and independent variables are described, followed by the moderator variable. Additionally, a clarification about the control variables is given. Lastly, the structure and reasoning of the applied methodology and the robustness test are justified.

Decision to engage in FDI

Firm size

Vicarious learning from neighbouring peers

Figure 1: Conceptual Framework

(21)

3.1 DATA COLLECTION

To be able to investigate the impact of vicarious learning among neighbouring firms on FDI decisions, a quantitative approach is selected. The ORBIS database was chosen to gather secondary data and also functions as the main source to derive information from. As the sample includes 2013, it was necessary to use the historical ORBIS database, in particular ORBIS Neo. ORBIS is widely utilised and offers information about approximately 300 million companies worldwide. Since ORBIS merges more than 160 different sources, it is possible to compare businesses globally and to gather extensive information about them (Bureau van Dijk, 2019). Additionally, the Italian Bureau of Statistics was used to collect province-level factors that influence FDI decisions.

3.2 SAMPLE

Italian manufacturing firms were selected for the sample. However, with the aim of getting significant results, this needs to be more specified. Hence, the sample includes Italian companies with a focus on the manufacturing sectors of (1) textiles, (2) wearing apparel and (3) leather and related products. These sectors were identified by using the NACE codes. This code classifies economic activities within the European Union (Eurostat, 2019a).

(22)

Information was not fully available for every control variable, so some companies had to be omitted. In total, there are 13,273 manufacturing companies in the sample.

3.3 MEASUREMENTS

This section explains the selected variables to explore the impact of vicarious learning among geographically proximate peers on FDI decisions. It begins with the dependent variable, followed by the independent variable, the moderator variable and lastly the control variables. An overview of these variables is depicted in Appendix A.

3.3.1 Dependent Variable

Whether or not a company initiated FDI in 2017 is this study’s dependent variable. This was measured by the presence of foreign subsidiaries, with at least 10% of direct ownership. This information was derived from ORBIS and was used as a binary variable in the statistical analysis. If firms engaged in FDI in 2017 it was coded as 1, and otherwise as 0.

3.3.2 Independent Variable

The independent variable is the number of firms already conducting FDI in 2013 within the separate Italian regions, and this is the proxy variable for organisations’ vicarious learning. To be able to establish this variable the two main databases explained above were used. By using the NUTS3 code, organisations already engaging in FDI in 2013 were summarised for each specific Italian region. This code is the nomenclature of territorial units for statistics. It is a hierarchical system and splits regions within a country up. It is ranged from NUTS1 until NUTS3, the higher the number, the more specific, and hence smaller is the region. Since the purpose is to study the impact of vicarious learning among close-by firms, the NUTS3 code was applicable for the merging (Eurostat, 2019b).

3.3.3 Moderator Variable

(23)

turnover than 10 Million euro are taken into consideration (Commission of Recommendation, 2003). On basis of this, a dummy variable for small firms (coded as 1) was created. Furthermore, Inkpen and Bearmish (1997) have argued that the size of a business influences the FDI decision, since it takes less effort for larger businesses to initiate FDI, due to their sufficient resources.

3.3.4 Control Variables

The control variables for the study are split in firm-level and province-level factors. The decision regarding whether or not to pursue FDI depends not solely on learning, but on other various reasons, for instance see Blonigen (2005). Therefore, it was crucial to take firm-level and province-firm-level factors into account. Firstly, an overview of the firm-firm-level variables is given, and secondly of the province-level factors.

3.3.4.1 Firm level

Firstly, for the firm-level, the company’s sector, age, total assets and the number of subsidiaries were selected to control for the study. This was all derived from the ORBIS database.

Sector

Relating to Goerzen and Sapp (2005) who controlled for the industry sector in their study, the first control variable of the present study is the sector of the manufacturing industry. There are three different manufacturing sectors chosen, and they need to be controlled for, because differences occur within sectors. The NACE code from ORBIS was taken to create dummy variables for the manufacturing sectors of textile (NACE code 13), wearing apparel (NACE code 14) and leather products (NACE code 15).

Firm Age

(24)

Total Assets

Inkpen and Beamish (1997) have stated that larger organisations have more resources, and sufficient financial resources are inevitable for firms to start FDI (Chang, & Rhee, 2011). Moreover, some studies (Nielsen, & Nielsen, 2011; Ferreira, da Salva Vicente, Borini, & de Almeida, 2017) have used the amount of total assets to determine the business size. So, this was also included in the control variables. This variable was transformed to a logarithm variable in Stata to limit the influence of potential outliers.

Number of Subsidiaries

It was further necessary to control a company’s number of subsidiaries. Previous research was extremely interested in the performance of subsidiaries, since they contribute to the overall business’ success (Andersson, Forsgren, & Holm, 2002; Feinberg, 2000; Venaik, Midgeley, & Devinney, 2005). Additionally, advanced knowledge and routines were found rather in subsidiaries than in firms’ headquarters (Frost, Birkinshaw, & Ensign, 2002). As a result, it was assumed that organisations with a greater number of subsidiaries have better knowledge and capabilities which can be applied. Hence this impacts the learning and imitation process and helps to conduct a FDI decision.

3.3.4.2 Province-level

The province-level data includes the rate of unemployment, the degree of openness, the ethnic diversity, the institutional quality, and the number of patents. The province-level data was derived from ISTAT, the Italian Bureau of Statistics, particularly for each of the NUTS3 regions in Italy.

Rate of Unemployment

The general economic state of a province has an impact on a firm’s decision whether FDI is feasible or not. Inspired by the study of Strat, Davidescu and Paul (2015), who measured the correlation between inward FDI and the unemployment rate of EU member states, this study controls for the unemployment rate of the divers Italian regions.

Degree of openness

(25)

correlation between the productivity level of a firm and the engagement in exports. When relating to the present study, a higher degree of openness results from more exports and this impacts FDI decisions.

Ethnic Diversity

Alesina and La Ferrera (2005) have demonstrated that due to ethnic diversity among employees the productivity of an organisation is boosted. This results from increased problem-solving skills and information flows. The present study assumes that a greater number of migrants in a province positively influences the vicarious learning and helps to make a decision regarding FDI. The variable was calculated by dividing the migrants of a province through the total residents.

Institutional Quality

The study of Buchanan, Le and Rishi (2012) has indicated that institutional quality is a crucial determinant for FDI. It is of special importance for developing countries, but not less relevant for developed countries. The results of that study have shown that good institutional quality has a positive effect on FDI (Buchanan et al., 2012). Institutional quality is different within regions, so it can influence the FDI decisions. Included in this variable are the political stability and the rate of corruption, and both are decisive for FDI decisions, because high corruption results in greater uncertainty within a region (Deseatnicov, & Akiba, 2016; Daniels, Raderbaugh, & Sullivan, 2002).

Number of Patents

The number of patents is another central factor which must be included in the control variables. When firms are geographically proximate located, knowledge flows from one organisation to another (Capello, 2009). A higher number of patents within a region signals a greater knowledge flow. That flow of knowledge, in the form of patents, is highly essential for FDI engagement (Shaver et al. 1997). So, in the present study, the number of patents within a region is an indicator of how much knowledge already exists and this is certainly affecting FDI decisions.

(26)

3.4 METHODOLOGY

The aim of this study is to figure out whether the learning and mimicry among close-by located firms impacts decision of others to engage in FDI too. To be able to test this, the dependent variable is the FDI decision. As it is used as a binary variable in the statistical analysis, it can be referred to as a dichotomous variable. The hypotheses are tested with a regression in the statistical software program Stata SE. Since the sample size is rather big, Stata SE is a perfect fit, because it is particularly designed for large datasets (Stata, 2019).

The usage of regressions is widely common in non-experimental social sciences (Lewis-Beck, Bryman, & Liao, 2004). In particular, this study follows a logistic regression model, because the dependent variable is a dummy variable (Caudill, 1988). Moreover, logistic regression models have been widely used for econometric applications in top journals (Horowitz, & Savin, 2001). It is important to mention that no statistical model is a completely true representation of reality, it is rather a useful representation of the reality (Chatterjee, & Simonoff, 2013).

When specifying an organisation’s FDI decision as a linear function of the independent variables, the following equation occurs:

!"#$ = &'+ )*$&+ + !,$ &.+ ()*$ 0 !1$) &3 + 4$ + 5$

Where FDI is the dependent variable, namely the positive or negative FDI decision of companies in 2017. VC is the main independent variable, so the proxy variable for firms performing vicarious learning. The coefficients of vicarious learning are depicted in &+. FS is

the moderator of this study, the firm size. &. contains the coefficients from the moderator which tests hypothesis 2. The used control variables are summarised in Z and 5$ is the error term of the analysis.

(27)

The multicollinearity of the logistic regression is tested with the variance inflation factor (henceforth VIF). The benchmark for the VIF value has to be between 0 and 10 (Field, 2009)

3.5 ROBUSTNESS TEST

A robustness test is carried out to validate the selected analytical method (Heyden, Questier, & Massart, 1998). Small changes in the analysis, for instance, using a different regression model, are used validate results (Mulholland, 1988). As mentioned above, the statistical analysis is conducted with logistic regression, so it is possible to use a linear probability model as a robustness test. Usually, these two models are chosen when the dependent variable is a dummy variable (Caudill, 1988), thus both are applicable to this study. Moreover, this study includes a large sample size and both, the logistic regression and the probability model are estimated by maximum likelihood, so both models are valid. Another similarity of these models is that both are symmetrical around zero and except for bigger trails in the logistic model, their shapes are quite akin (Horowitz, & Savin, 2001).

4 R

ESULTS

(28)

4.1 DESCRIPTIVE STATISTICS

Variable Mean SD Min Max

FDI Decision .021 .145 0 1

Firm Age 18.122 17.82 1 140

Total Assets (in th USD) 7197.698 61217.75 .066 4590000

Sector 13 .297 .456 0 1 Sector 14 .409 .492 0 1 Sector 15 .297 .457 0 1 Subsidiaries .412 1.661 0 78 Unemployment .106 .052 .044 .262 Institutional Quality .668 .216 0 1 Degree of Openness 315000 159000 4214.113 1100000

Patents on Total Residents 0 0 0 .002

Ethnic Diversity .089 .035 .014 .146

No. of Firms engaging in FDI 14.473 12.877 0 43

Small Firm .856 .351 0 1

Table 1: Descriptive Statistics

(29)

unemployment rate in 2013 was 26.2% (SD=5.2%). The ethnic diversity in the various Italian provinces ranged from 1.4% until 14.6% (SD=3.5%), with an average ethnic diversity of 8.9% in a province. The sample includes a mean of 85.6% (SD= 35.1%) small firms, so the vast majority of the organisations.

4.2 MULTICOLLINEARITY

(30)

Variables (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (1) FDI Decision 1.000 (2) Firm Age 0.079 1.000 (3) Total Assets 0.261 0.294 1.000 (4) Sector 0.003 -0.118 -0.078 1.000 (5) Subsidiaries 0.286 0.128 0.389 -0.002 1.000 (6) Unemployment -0.062 -0.135 -0.186 0.133 -0.069 1.000 (7) Institutional Quality 0.035 0.100 0.121 -0.152 0.064 -0.644 1.000 (8) Degree of Openness 0.074 0.096 0.174 -0.083 0.043 -0.601 0.359 1.000

(9) Patents on Total Residents 0.021 0.052 0.088 -0.007 0.045 -0.416 0.208 0.425 1.000

(10) Ethnic Diversity 0.052 0.100 0.169 -0.145 0.052 -0.830 0.583 0.541 0.324 1.000

(11) No. of firms engaging in FDI 0.054 0.098 0.172 -0.057 0.058 -0.448 0.351 0.419 -0.009 0.541 1.000 (12) Small Firm -0.271 -0.216 -0.620 0.022 -0.311 0.119 -0.087 -0.104 -0.045 -0.101 -0.117 1.000

(13) Learning * Size -0.092 -0.018 -0.152 -0.042 -0.120 -0.354 0.282 0.324 -0.007 0.442 0.789 0.386 1.000

(31)

Table 2 depicts the results of the Pearson correlation matrix. All correlations between the variables lay between -1 and 1, according to Field (2009) the following thresholds are taken into account for this study. If the value is between 0 – 0.1 there is no correlation between the variables, but a value between 0.1 – 0.3 depicts a medium correlation effect. A value between 0.3 – 0.5 is signalling a high correlation effect (Field, 2009). From the value 0.5 on, there is no general agreement on the threshold concerning various correlation levels. Some authors have applied a value of 0.7, whereas others find a value of 0.9 as a high correlation (Tabachnick, & Fidell, 2007). To avoid a too high correlation between the variables in this study, a benchmark of 0.7 is considered as high correlation.

The general conclusion from the Pearson correlation matrix is that there were some correlations, but only two correlations were considered high. When looking at the province-level control variables, a high negative correlation was found between ethnic diversity and unemployment rate (-0.830), and a significant positive correlation between the interaction term and the independent variable (0.789). When following the recommendation from Field (2009) high correlations were found between several variables. Despite that recommendation, and since there is no common agreement, this study considers it as medium correlations. In the province-level control variables there were a few medium correlations. There was a negative medium correlation between the institutional quality and the unemployment rate (-0.644), as well as between the degree of openness and the unemployment rate (-0.601). Moreover, ethnic diversity had a mediocre correlation with institutional quality (0.583) and degree of openness (0.541). Lastly, the independent variable was moderately correlated with ethnic diversity (0.541).

Additionally, the multicollinearity was tested with the VIF. Since all values were below the threshold of ten, as recommended by various authors (Neter, Wassermann, & Kutner, 1985; Kennedy, 1992; Studenmund, 1992) it is concluded that there are no multicollinearity issues in this study.

4.3 RESULTS OF LOGISTIC REGRESSION

Table 3 shows the results of the logistic regression. Firstly, an overview of the used models is given, and additionally a detailed description of the results.

(32)

interaction term between the independent variable and the moderator, and finally presented the full regression.

(1) (2) (3) (4) (5)

VARIABLES Model 1 Model 2 Model 3 Model 4 Model 5 Dependent Variable FDI Decision (1) yes (0) no

Control Variables

Firm Age 0.00179 0.000778 0.000836 7.53e-05 -0.000103 (0.00359) (0.00370) (0.00371) (0.00368) (0.00369) Total Assets 1.073*** 1.036*** 1.046*** 0.802*** 0.804*** (0.0521) (0.0529) (0.0534) (0.0724) (0.0725) Sector 14 0.367** 0.428*** 0.402** 0.408** 0.402** (0.162) (0.165) (0.165) (0.164) (0.164) Sector 15 0.519*** 0.569*** 0.600*** 0.525*** 0.524*** (0.168) (0.171) (0.171) (0.171) (0.171) Subsidiaries 0.0127 0.0225 0.0236 0.0400* 0.0403* (0.0197) (0.0206) (0.0207) (0.0207) (0.0206) Unemployment -10.34*** -11.75*** -11.81*** -11.62*** (4.010) (4.116) (4.100) (4.113) Institutional Quality -0.220 -0.202 -0.262 -0.282 (0.417) (0.414) (0.412) (0.414) Degree of Openness 1.79e-06*** 2.09e-06*** 2.05e-06*** 2.05e-06***

(5.01e-07) (5.16e-07) (5.16e-07) (5.20e-07) Patents on Total Residents -1.157 -2.008* -2.019* -2.086*

(998.4) (1.119) (1.111) (1.121) Ethnic Diversity -2.553 -1.320 -1.960 -2.044 (3.821) (3.831) (3.835) (3.848) Independent Variable Learning of Firm -0.0116** -0.0108* -0.0159** (0.00580) (0.00575) (0.00637) Moderator Variable Small Firm -1.002*** -1.386*** (0.213) (0.294) Interaction Term

Learning * Small Firm 0.0213**

(0.0107) Constant -14.48*** -13.30*** -13.21*** -10.17*** -10.08***

(0.549) (0.990) (0.991) (1.158) (1.160) Observations 13.273 13.273 13.273 13.273 13.273

* p<0.10, ** p<0.05, *** p<0.01 Standard error in parentheses

Table 3: Results of the logistic regression

(33)

likely to decide in favour of FDI. This finding is aligning with the research of Ouimet and Zarutskie (2014) and Brouthers and Brouthers (2003) who have stated that old-established companies have greater funds, and this results in a higher probability to engage in FDI. The positive sign of subsidiaries (!=0.0127) implies that possessing subsidiaries leads to a greater possibility for firms to engage in FDI.

Model 2 added the control variables of the province-level to the analysis. Here, both the unemployment rate (p≤0.01, !=-10.34) and the degree of openness (p≤0.01, !=1.79e-06) showed significance. The unemployment rate correlated negatively with the FDI decision, meaning that a decrease in unemployment positively influences companies’ FDI decisions. Whereas the degree of openness had a positive correlation, so depicting a higher possibility of FDI engagement when a region is exporting more. Total assets (p≤0.01, !=1.036) and the sectors 14 (p≤0.01, != 0.428) and 15 (p≤0.01, != 0.569) showed significance again. The correlation between total assets and FDI decision correlated positively and this implies that greater total assets lead to a higher probability of FDI engagement.

Model 3 further included the independent variable. The relationship between the independent and dependent variable was significant (p≤0.05, !=-0.0116), and supports the first hypothesis. So, it is true that vicarious learning from neighbouring peers has an impact on FDI decisions. Conversely to the anticipated assumption of this study, it showed a negative correlation with the dependent variable. Therefore, spatial spillovers from geographically proximate located peers have a negative impact on firms’ FDI decisions. This model does not include the moderator of being a small firm yet, thus the negative impact is true for large organisations solely. Except from this unanticipated result, the other variables are aligning with the second model, only the patents on total residents started to be significant from this model on (p≤0.10, !=-2.008). As that coefficient is negative, a higher number of patents decreases the probability of a company to engage in FDI.

(34)

may negatively affect large organisations’ FDI decisions. Besides this finding, subsidiaries were significant and had a positive value (p≤0.10, !=0.0400). Moreover, total assets were significant again (p≤0.01, !=0.802), inferring that an increase in total assets has a positive impact on a firm’s FDI decision. Also, the degree of openness depicted a positive and significant value again (p≤0.01, !=2.05e-06).

The interaction term between the independent variable and the moderator was included in the fifth model, and it finally represented the complete regression. The correlation of these variables was significant (p≤0.05) and had a positive coefficient (!=0.0213), consequently this supports the second hypothesis. Therefore, this analysis confirmed that vicarious learning is of particular relevance for small firms. The independent variable is negatively correlated with the FDI decision, but showed significance (p≤0.05, !=-0.0159). So, the positive coefficient of the interaction term implies that small companies benefit from vicarious learning. Similarly, to the fourth model, the subsidiaries were significant and positively correlated with the dependent variable (p≤0.10, !=0.0403). This indicates that for small firms, subsidiaries play a more decisive role than for larger firms. Align with the previous models, unemployment rate had explanatory power to the FDI decision and a negative coefficient (p≤0.01, !=-11.62). Hence, a reduction of the unemployment rate increases the probability for an organisation to engage in FDI. This result was also reflected in prior research where Strat et al. (2015) have demonstrated that there is a correlation between the unemployment rate in the EU and inward FDI. The same is true for the degree of openness, as this variable had again a positive coefficient and was significant (p≤0.01, !=2.05e-06). The independent variable, the moderator variable and the interaction term were significant in this model. Furthermore, the same control variables on the firm- and province-level as before. Astonishingly, firm age was negatively correlated with the dependent variable (!=-0.00013). Despite the small number, it still indicates that older businesses are less likely to engage in FDI, and this finding is contradictory with prior research (Ouimet, & Zarutskie, 2014; Brouthers, &Brouthers, 2003).

(35)

internationalisation. Moreover, the institutional quality showed no explanatory power to the dependent variable in all models and was also always negatively correlated. This result is inconsistent with prior research of several authors too (Deseatnicov, & Akiba, 2016; Daniels et al. 2002).

4.4 RESULTS OF ROBUSTNESS TEST

The results of the robustness test, where a linear probability model was applied are depicted in table 4.

(1) (2) (3) (4) (5)

VARIABLES Model 1 Model 2 Model 3 Model 4 Model 5 Dependent Variable FDI Decision (1) yes (0) no

Control Variables Firm Age 0.000180 -0.000320 -0.000296 -0.000571 -0.000613 (0.00177) (0.00181) (0.00181) (0.00182) (0.00182) Total Assets 0.502*** 0.490*** 0.492*** 0.393*** 0.393*** (0.0257) (0.0260) (0.0261) (0.0352) (0.0353) Sector 14 0.192** 0.214*** 0.208*** 0.207*** 0.205*** (0.0751) (0.0765) (0.0766) (0.0771) (0.0771) Sector 15 0.240*** 0.265*** 0.273*** 0.241*** 0.241*** (0.0780) (0.0795) (0.0798) (0.0807) (0.0807) Subsidiaries 0.0170* 0.0198** 0.0202** 0.0274*** 0.0277*** (0.00980) (0.00997) (0.00999) (0.0101) (0.0101) Unemployment -4.362** -4.750*** -4.845*** -4.869*** (1.727) (1.771) (1.785) (1.798) Institutional Quality -0.140 -0.132 -0.157 -0.164 (0.189) (0.188) (0.190) (0.191) Value Added Exports 9.03e-07*** 1.01e-06*** 1.01e-06*** 1.00e-06***

(2.37e-07) (2.49e-07) (2.49e-07) (2.52e-07) Patents on Total Residents -604.7 -874.2* -862.4* -897.1*

(469.0) (522.6) (522.2) (527.9) Ethnic Diversity -1.512 -1.103 -1.403 -1.464 (1.748) (1.767) (1.784) (1.793) Independent Variable Learning of Firm -0.00373 -0.00350 -0.00608* (0.00276) (0.00277) (0.00330) Moderator Variable Small Firm -0.369*** -0.490*** (0.0916) (0.123) Interaction Term

Learning * Small Firm 0.00675

(0.00460) Constant -6.982*** -6.475*** -6.450*** -5.211*** -5.151*** (0.261) (0.448) (0.450) (0.539) (0.543) Observations 13.273 13.273 13.273 13.273 13.273

* p<0.10, ** p<0.05, *** p<0.01 Standard error in parentheses

(36)

The sample depicted some differences in both analyses, but it still remained consistent and the validity was proved. In both analyses, total assets and sector showed the same level of significance. In the linear probability model, also the number of subsidiaries was significant in all five models, whereas it was only significant in model 4 and model 5 in the logistic regression. There were no indifferences in the province-level control variables, since unemployment and value added exports depicted the same levels of significance in both analyses. Nevertheless, the independent variable was different in both analyses. Whereas it was significant in the logistic regression, hence supporting hypothesis 1, it was only significant in model 5 in the linear probability model. Although there is a slight difference in regard to the independent variable, the moderator variable however remains unchanged in both analyses. The interaction term was significant in the logistic regression and supports hypothesis 2, but it had no explanatory value in the linear probability model and does therefore reject the second hypothesis.

5 D

ISCUSSION

The main objective of this thesis was to contribute to the study of vicarious learning of geographically proximate located firms. It has therefore emphasised the decision of engaging in FDI. Since it can be difficult for companies to make strategic decisions, it was assumed that in this case, inexperienced businesses prefer to rely on other, nearby-located peers. Since geographically proximate companies experience the same business environment, this assumption was made. This was expected to be true for all firms, regardless of their size. However, since small organisations have fewer financial and managerial resources, it is likely that these firms have to rely even more on learning from other local businesses.

(37)

uncertainty for organisations if they lack the required information about the host market. In support of this, Zaheer (1995) has stated that firms incur immense additional costs if they want to engage in business abroad and are unfamiliar with the host context. Consequently, the result of the logistic regression was astonishing.

However, the interaction term between the moderator and the independent variable had a positive correlation. Therefore, it was concluded that the negative correlation between vicarious learning and FDI decisions only applies to large businesses and does not affect small firms. Although this was not anticipated, this study illustrated that large businesses do not necessarily have to rely on vicarious learning from geographically proximate located peers. One explanation for this astounding result is that large firms are well aware of congestion effects. These effects surface when a high number of firms are operating in similar sectors agglomerate nearby (Pouder, & St. John, 1996). The present study has focused on companies in the manufacturing sectors of textile, wearing apparel and leather products, which are related sectors. Therefore, the benefits of locating near others are outweighed by congestions effects, resulting in a greater competition, as these firms all reach out for the same type of resources (Pouder, & St. John, 1996). Another possible reason why large firms do not engage in vicarious learning to make a decision about FDI engagement is that they already possess sufficient resources and do not dependent on imitating others. This is supported by prior research. Inkpen and Beamish (1997) contend that being a large business also indicates that they possess adequate resources. Ergo, large firms instead rely on their internal resources to make a decision about conducting FDI and do not imitate neighbouring peers. Moreover, Chang and Rhee (2011) have stated that decent financial resources are an indispensable requirement for FDI engagement. Since this study also demonstrates that greater total assets result in positive FDI decisions, it is assumed that large organisations obtain sufficient financial resources and are able to make the decision solely based on internal factors.

(38)

barriers and constraints, resulting in a more difficult internationalisation. Buckley (1989) has also recognised the restrictions of small firms. These various constraints and barriers exacerbate uncertainties in internationalising for small companies (Mutinelli, & Piscitello, 1998).

The correlation between the interaction term with the dependent variable was positive in the analysis. This indicates that vicarious learning of neighbouring peers is beneficial for small businesses. Therefore, small firms can overcome the several internationalisation barriers by applying vicarious learning. This decreases their perceived uncertainty and the unfamiliarity with the host context, and results in positive FDI decisions. So, this realisation of the thesis illustrates that small companies defeat their insecurities through vicarious learning from geographically proximate located businesses.

To conclude the findings from testing the two hypotheses, only small firms benefit from vicarious learning, and large firms have realised the existence of congestion effects and preferably rely on their internal resources. Aside from the surprising result of the first hypothesis, some control variables also yielded results contradictory to the assumptions made. Firstly, the firm age was positively correlated with FDI decisions in the first four models, which suggests that older organisations are more likely to engage in FDI. Recent research by Ouimet and Zarutskie (2014) and Brouthers and Brouthers (2003) has confirmed this. However, in the fifth model this relationship was suddenly negative, which indicates that older firms are less likely to internationalise. Model 5 included the interaction term between being a small business and vicarious learning from geographically proximate peers. Therefore, in this model it is more likely that younger organisations will engage in FDI. Consistently with this finding, the research of Fernhaber and Li (2013) also concluded that young ventures will prefer to use vicarious learning to make internationalising decisions.

(39)

decision-making. Conversely, the present study does not relate to this finding, as the correlation between the ethnic diversity and the FDI decision is negative in all models.

Another surprising effect was the institutional quality of a region, since it was insignificant in all models. This study anticipated that enhanced institutional quality results in positive FDI decisions, but no support was found. That finding contradicts to that of Buchanan et al. (2012), because they argue that the high institutional quality of a region is positively related to FDI engagement. Additionally, Deseatnicov and Akiba (2016) and Daniels et al. (2002) have determined that the political stability and the rate of corruption are decisive for firms when making a decision about FDI engagement. These two factors were included in the institutional quality for this study, which also contradicts recent research. Moreover, the subsidiaries are also insignificant in the first three models, which also deviates from prior research. Frost et al. (2002) revealed that most of a firm’s knowledge lies in the subsidiaries rather than in the headquarters. Therefore, the present study assumed that organisations with more subsidiaries possess greater knowledge about internationalisation and the host market, resulting in a positive FDI decision. However, the results of the analysis have proven this assumption wrong. So, the conclusion of this analysis for the number of subsidiaries is that it only has explanatory power for FDI decisions if small businesses are involved.

One control variable that always acted accordingly to the anticipation of this thesis is the rate of unemployment. The effect of the unemployment rate to the FDI decision was significant and negatively correlated in all models. Consequently, it coincides with the prediction of this study as well as with prior research (Strat et al., 2015). It suggests that a greater rate of unemployment impedes organisations’ decision to engage in FDI.

6 C

ONCLUSION

(40)

6.1 THEORETICAL IMPLICATIONS

The literature on organisational learning and small firms is already quite extensive. However, this study has yielded two theoretical implications which can contribute to the current literature. As was already noted by other authors, it is particularly burdensome for SMEs to survive and hold their business position (Singh et al., 2010).

This major impediment of SMEs leads to the first theoretical implication of this study. Prior research has stressed that small businesses suffer extensively from various liabilities and have offered solutions to this major problem, this study provides another suitable solution. For instance, Ramsden and Bennet (2005) suggest that SMEs should seek external help from advisers. This may be too costly for some small firms who already suffer from financial constraints. Consequently, examining the business practices of neighbouring peers could be a seamless and less expensive solution to overcome the typical impediments of small companies. Although the possibility of observing and learning from neighbouring peers to address compelling strategic decisions is effortless and easy to obtain, it has been somewhat overlooked in past research. It may be concluded that some constraints typical for small firms cannot be resolved through the observation and mimicry of neighbouring businesses. However, when small firms successfully implement business strategies of neighbouring firms, it will have positive effects on decisions regarding FDI engagement. Therefore, the chance of gaining insights into others’ business practices and gathering essential information does not have to be costly or difficult to achieve.

Referenties

GERELATEERDE DOCUMENTEN

This chapter describes how the Consumat framework is applied specifi- cally to a model of the personal vehicle market which has been given the name STECCAR (Simulating the Transition

The most important finding from the future prediction analysis is that although Random Forest, XGBoost, CatBoost are similar in performance and feature importance, their

Local ties enable mobility through local social capital especially in rural areas, which leads to an increased likelihood of migration for individuals with local ties living

A focus-plus- context approach is taken by rendering nearby building models based on LIDAR data in order to give better insight in the scale of the dataset, whilst maintaining

Where Zuse explores information processing theory as a way to give new insights in physics and where Schmidhuber (mostly) hypothetically explores what it would mean for a universe to

The first part of this overarching project consist of an information analysis regarding the people involved (stakeholders), work flows and creation and design of business

The contract type is performance based contract, where inspections preventive and corrective maintenance repairs are included.. The quay is an asset which needless to

Furthermore, a ‘toolbox’ was set up to support the students performing this assign- ment. This toolbox, which was digitally accessible, included a repository of relevant