Public finance and SME foreign market entry mode choice: The moderating effect of slack resources

(1)

Master Thesis (IB&M)

Public finance and SME foreign market entry mode choice:

The moderating effect of slack resources

University of Groningen

Faculty of Economic and Business

Supervisor: Dr. A. Kuiken Co-Assesor: Dr. M.C. Sestu Submitted by: Alicia Schrader

Student number: 4083814

(2)

Schrader, A. (2020) 2

Acknowledgements

First, I would like to dedicate my gratitude for the supportive manner of my supervisor Dr. Andrea Kuiken (FEB, University of Groningen). Her door was always open whenever I ran into a trouble spot or had a question about my research or writing. She consistently allowed this paper to be my own work but steered me in the right the direction whenever she thought I needed it. Despite the difficult circumstances due to the COVID-19 situation, she managed to provide me full assistance and supervision throughout the process of writing this thesis. I highly appreciate all the virtual meetings and communication.

Second, I must express my very profound gratitude to my parents and my sisters for providing me with unfailing support and continuous encouragement throughout my years of study and especially through the process of researching and writing this thesis. This accomplishment would not have been possible without them. Thank you.

(3)

Schrader, Alicia 3

Abstract

Despite the increasing importance of small-medium sized enterprises (SMEs) and their internationalization in terms of contribution to the domestic economy, most of the internationalization literature in focuses solely on multinational enterprises (MNEs). This study contributes to the literature of SMEs and entry mode by providing novel insights and empirical evidence for the effect public finance in the home country has on SMEs’ equity entry mode choice and how this effect is moderated by the SMEs’ level of financial and human slack resources.

By drawing on the Resource Based View (RBV), the premise of this study is that SMEs’ resource limitations play a major role in the entry mode decision, and argues that public finance positively affects the likelihood of SMEs to opt for a wholly-owned subsidiary (WOS) rather than a joint venture (JV). It is further proposed that this effect is negatively moderated by the degree of financial slack resources and positively moderated by the degree of human slack resources the SME possesses.

The empirical analysis of 264 foreign entries of Swedish SMEs in the period of 2007-2017, however, revealed some surprising results. While public finance was indeed found to increase SMEs’ likelihood to establish WOSs rather than JVs by mitigating the liabilities of smallness inherent to SMEs, this effect is negatively moderated by the degree of human slack. Financial slack was not found to have any significant moderating effect in this regard.

Key words: entry mode, SMEs, home-country, slack resources, resource-based view, joint venture, wholly-owned subsidiary

(4)

Schrader, Alicia 4

1. Introduction

In 1948, one of the economically most important trade agreements, the General Agreement on Tariffs and Trade (GATT), was signed (McKenzie, 2008). The subsequent reduction of trade and investment barriers led to the emergence of new large markets and a global boom in foreign direct investment (FDI) (Kokko, 2006). In fact, in the period from 1990 to 2000, the global stock of FDI increased from 8.3% to 17.5% (UNCTAD, 2002). Four years after the observation of this trend, the 2nd OECD Conference of ministers responsible for SMEs took place in Turkey. The corresponding OECD report provides evidence that SMEs remain relatively under-represented in international markets compared to their contribution to local and national economies. This is surprising, considering that SMEs represent the majority of firms in most countries. For instance, in Sweden SMEs account for 99% of the enterprises (OECD, 2019a). They play a crucial role in the economic growth of their home countries (Musso & Francioni, 2014). Indeed, the OECD reports that SMEs typically account for around 50% of GDP, and 60% of employment in national or local economies. However, they represent only 20% - 40% of exports, and even less of international investment. (OECD, 2018b).

When firms decide to invest abroad, it is not solely of importance to decide on what market to enter, but also on how to enter it (Lee & Lieberman, 2010). In the International Business literature, international entry mode is listed as the third most examined research field (Werner, 2002).

Because of the entry mode’s possible negative effect on the firm’s performance, the importance of making the right decision is high (Lu & Beamish, 2001; Nakos & Brouthers, 2002). Once a firm has decided on an entry mode, it can become costly and time-consuming to make changes to that decision (Kumar & Subramaniam, 1997).

Following the increasing number of SMEs operating in international markets (Nummela et al., 2006; OECD, 2019b; UNCTAD, 1993), many scholars have realized the long-term strategic importance of entry mode decisions especially for SMEs (Burgel & Murray, 2000) and conducted research on the topic (e.g., Coviello & Mc Auley, 1999; Laufs & Schwens, 2014).

(7)

Schrader, Alicia 7

highlights the ownership mode as one of the strategically most important decisions to take (Dikova & van Witteloostujin, 2007). The ownership decision in equity modes refers to the decision whether the firm wants to enter the foreign market with a partner through a JV, or alone by establishing a WOS (Pan & Tse, 2000).

The choice of ownership mode involves different levels of resource commitment. Considering the nature and characteristics of SMEs, the ownership mode choice becomes particularly challenging for these types of firms (Agndal & Chetty, 2007; Bruneel & De Cock, 2016). In this regard, research frequently denotes SMEs’ limitations in terms of resources owing to their small size (e.g., Burgel & Murray, 2000; George et al., 2005). It is often argued that due to the resource constraints, SMEs are prone to choose entry modes involving relatively low resource commitment such as JVs (e.g., Agarwal & Ramaswami, 1992; Brouthers et al., 2003; Buckley & Casson 1976; Campa & Guillén, 1999; Caves & Mehra 1986; Cho 1985; Hutchinson et al., 2006; Kimura, 1989; Rialp et al., 2002; Ripollés et al., 2012; Stopford & Wells, 1972; Terpstra & Yu 1988; Trevino & Grosse, 2002; Yu, 1990; Yu & Ito 1988). Nevertheless, scholars also report SMEs engaging in higher commitment modes (e.g., Brouthers & Nakos, 2004; Buckley, 1989; Kohn, 1997; Maekelburger et al., 2012; Pavitt et al., 1987; Yap & Souder, 1994).

To contribute to resolving inconsistent findings, scholars increasingly started examining different determinants of SMEs’ choice of foreign market entry mode (e.g., Brouthers & Nakos, 2004; Cheng, 2008; López-Duarte & Vidal-Súarez, 2010; Musso & Francioni, 2012; 2014; Nakos & Brouthers, 2002; Schwens et al., 2018).

After reviewing the current state of the literature of SMEs and entry mode, Laufs and Schwens (2014) found that research would benefit from further examination of how contextual dimensions influence the FDI entry mode decisions of SMEs. Notably, taking into account the specific SME characteristics in terms of resource limitations, there is a dearth of literature examining to what extent SMEs’ home market might be a source for additional resources (Laufs & Schwens, 2014). Scholars mention that SMEs, due to their limited resources, may benefit from home-country support when they pursue FDI (Aharoni et al., 1981; Laufs & Schwens, 2014; Torres et al., 2016). Indeed, a recent OECD paper suggests SMEs’ need for access to financial sources to secure the full development of their potential economic growth (OECD, 2019a).

(8)

Schrader, Alicia 8

governments encouraging their firms to internationalize (Cui & Jiang, 2010; Luo et al., 2010; Marinova et al., 2012; UNCTAD, 2006; Wang et al., 2012; Yamakawa et al., 2008). In fact, a recent paper of the OECD (2019a) revealed that in most of the OECD countries, SME finance is an important component of the policy agenda. Nevertheless, the relationship between home-country support and SMEs’ FDI entry mode decisions remains poorly understood (Brewer, 1993; Dominguez, 2018; Shapiro & Globerman, 2003; UNCTAD, 2001). Laufs and Schwens (2014) particularly highlight that there is a complete dearth of studies examining how public finance in the domestic market impacts SMEs’ foreign market entry mode choice.

However, scholars frequently suggest to take both, internal and external drivers for entry mode, coherently into consideration (e.g., Brouthers, 2002; Nisar et al., 2012; Paul et al., 2017) by incorporating the firm as well as its context (He et al., 2013).

While often statistical methods are used to address this endogeneity, these can be seen as “crutches or substitutes for critical thinking” (Reeb et al., 2012, p. 216). Accordingly, Bruneel (2016) claims that research yet fails to theoretically address endogeneity by incorporating both, internal and external factors, to explain SMEs’ entry mode decision-making process.

(9)

Schrader, Alicia 9

on SMEs’ entry mode decision would, therefore, constitute a benefit and advance for the state of literature.

Concluding the above, this study aims to fill the presented literature gap by analyzing the effect of public finance provided by SMEs’ home countries on the entry mode decision and how this effect is moderated by the SMEs’ availability of financial and human slack resources.

Hence, this thesis adopts the following research question:

How does public finance affect SMEs’ entry mode decisions and how is this effect moderated by the SMEs’ level of financial and human slack resources?

To answer the research question, this study builds on the RBV, which is one of the most commonly used theoretical frameworks in international research (Peng, 2001). According to the RBV, firms achieve competitive advantages due to their unique, valuable, and hard to imitate resources and capabilities (Barney, 1991; Wernerfelt, 1984). The paper’s main premise is that such resources not only enable SMEs to commit the necessary resources for establishing full ownership mode but also lead to a preference for WOS that may better protect these valuable resources. Likewise, it differentiates between the dissimilar considerations of externally obtained financial resources through public finance and internally available financial slack resources and how these may vary as drivers for SMEs’ entry mode.

This study hypothesizes that public finance is able to compensate for SMEs’ resource constraints and thus increases the likelihood of SMEs to opt for a WOS rather than a JV. In addition to that, this study will build up a theoretical argumentation on why the likelihood of a WOS is expected to be negatively moderated by the firm’s internal financial slack resources and positively moderated by the human slack resources.

Taking into consideration the aforementioned gap in the literature of SME and entry mode, by executing a quantitative analysis of Swedish SMEs’ foreign market entries to test its theoretical hypotheses, this study’s contribution is twofold.

(10)

Schrader, Alicia 10

Second, it contributes to the research on SME internationalization (e.g., Brouthers et al., 2009; Lu & Beamish, 2001; 2006) by enhancing a more comprehensive understanding how SMEs’ specific characteristics, particularly their financial and human resources (Brouthers & Nakos, 2004; Nakos & Brouthers, 2002), play a role in the context of their entry mode decisions. Examining external and internal determinants in SMEs’ entry mode decision jointly will provide a more exhaustive picture of SME entry mode decisions and its boundary conditions. Foremost, the contribution to this research is built on the emphasis of slack resources rather than absolute resources (Kiss et al., 2017).

Firm-level data from 264 Swedish SMEs’ foreign market entries in the period of 2007-2017 will be collected from the Zephyr and Orbis database. Combined with data from SME direct governmental loans provided by the Swedish government in this period retrieved from a recent OECD report, this thesis will conduct a binary logistic regression to test the hypotheses.

The results show that public finance indeed increases the SMEs’ probability to opt for a WOS rather than a JV. However, with respect to the moderating effects of the financial and human slack, this paper reveals some surprising findings. Contrary to my predictions, internal financial slack did not significantly affect the influence of public finance on the SMEs’ entry mode decision. Moreover, human slack was identified as a significant moderator, however, decreasing the likelihood of SMEs to opt for WOS.

The remainder of this thesis is organized as follows. In chapter 2, a brief literature review on SMEs and entry modes will be provided, while chapter 3 will present the formulation of specific hypotheses related to the research question. Subsequently, in chapter 4, the data, and the sample analytical method used to test the hypotheses will be described. Thereafter, chapters 5 and 6 will present and discuss the results. Lastly, chapter 7 limitations and implications of this study will be depicted.

2. Literature Review

(11)

Schrader, Alicia 11

literature concerning the home-country effect and how it is relevant in the SME entry mode context. Finally, this chapter will present the main findings of the literature with respect to financial and human slack resources in the context of SMEs and entry mode.

2.1 Resource-based view

The RBV’s main maxim is that a firm’s core resources and competencies ensure its long-term competitive advantage (Collis, 1991; Hamel & Prahalad, 1990; Prahalad & Bettis, 1986; Stalk et al., 1992). These resources and capabilities are considered imperfectly mobile across firms (Barney, 1991). Considering that FDI involves the transfer of considerable amounts of resources, the RBV can help to explain the entry mode choice and understand the entry mode characteristics that support this choice (Meyer, Estrin &Peng, 2009; Meyer, Wright & Pruthi, 2009). The RBV of internationalization argues that the consideration of all available resources and capabilities of a firm as well as the environmental realities are the basis for major decisions, such as entry mode choices (Bell et al., 1998; Grant, 1991).

By drawing on arguments from the RBV in the context of FDI, a firm’s internationalization can be viewed as “the process of mobilizing, accumulating, and developing resource stocks for international activities” (Ahokangas, 1998, p. 3.1). In other words, the unique and interdependent resources a firm possesses contribute to its internationalization process and enable a more efficient pursuit of individual strategic aims (Ainuddin et al., 2007). Here, both, the amount of tangible and intangible resources can act as a firm’s strength or weakness (Wernerfelt, 1984).

2.2. Entry modes

2.2.1. Ownership mode in entry mode

(12)

Schrader, Alicia 12

When it comes to deciding on how to enter a foreign market via FDI, important strategical decisions will have to be taken. One important decision is the ownership mode, which refers to the decision whether to enter the foreign market alone (establishing a WOS) or to involve a local partner by establishing a JV (Brouthers & Hennart, 2007; Dikova & van Witteloostuijn, 2007). Classifying entry modes in terms of their level of ownership and distinguishing between full ownership mode (WOS) and shared ownership mode (JV) is a common practice in the research field of entry mode decisions (e.g., Brouthers & Hennart, 2007; Kaynak et al., 2007; Sestu et al., 2018; Tsang, 2005; Woodcock et al., 1994).

Among the vast array of alternative modes available, JVs and WOSs represent two primary and largely competing strategic options that a firm can choose from. The differences in the level of ownership, and thus the control and possible profit returns gained in the organization through the respective entry mode binds the acquiring firm to accept the different degrees of risk exposure and resource commitment (Agarwal & Ramaswami, 1992; Chan, 1995; Hill et al., 1990; Polat, 2007). For this reason, scholars, typically classify JVs and WOSs along a continuum from minimum to maximum control, risk, commitment, and potential return (e.g. Anderson & Gatignon, 1986; Hill et al., 1990).

2.2.2. Joint Ventures

On one side of the spectrum, a JV represents a strategic option used by firms to enter a foreign market to jointly produce and sell products or services with a partner (Kotler & Armstrong, 1994). In this case, the investing firm agrees upon sharing both commitment but also control with the partner firm in order to gain access to the partner’s knowledge or resources (Anderson & Gatignon, 1986; Dominguez, 2012).

It is argued that through a JV, firms save on the costs of creating strategic resources internally or acquiring them externally (Chen & Hennart, 2002) which implies the reduced resource commitment associated with this ownership mode.

If one takes into consideration that in a shared ownership mode, each partner perceives control rights proportional to its equity investment (Mantecon et al., 2016), one can suggest that with the equity decreasing, the control over the foreign operation also decreases, however, at the same time the risks increase. The risks associated with this type of foreign market entry are often related to possible conflicts with and opportunistic behaviors of the JV partner (Chen & Hennart, 2002).

(13)

Schrader, Alicia 13

2.2.3. Wholly-owned subsidiaries

On the other side of the spectrum, a WOS refers to the entry option in which the firm undertakes the investment entirely by itself and hence retains whole ownership of its subsidiary (Raff et al., 2009). Accordingly, the eminent difference between a WOS and a JV is related to the resources involved in the transaction. While in a JV the partner firms share and provide access to a rash of their internal firm resources, by establishing a WOS the firm fully relies on its own assets (Raff et al., 2006; 2009). In other words, a firm opting for a WOS needs to commit more resources. However, equity modes that require high resource commitment also provide the acquiring firm benefits related to higher levels of control and higher profit returns (Canabal & White, 2008; Chan, 1995; Dong et al., 2008).

Establishing a subsidiary entirely by itself prevents the expanding firm from fearing opportunistic behavior of a partner, which is why it is frequently argued that WOSs offer better protection of the firm’s resources and entail lower risk associated with the exploitation of its valuable resources (Bloodgood et al., 1996).

2.2.4. Commitment, risk, control and profit return in JVs and WOSs

As can be concluded from the aforementioned 2 sections, the two different ownership modes are highly distinct from each other in terms of the costs (risks and resource commitment) and the benefits (control and profit return) that are involved in the transaction (Sharma & Erramilli, 2004). Table 1 summarizes the aforementioned evaluation of both entry modes, JVs and WOSs, along the continuum of the 4 cost aspects. This portrayal illustrates how JVs and WOSs represent two highly opposing options. Depending on the firm’s perception and assessment of the advantages and disadvantages entailed in each mode, it prefers to opt for either of the two options. Against the background of this framework, this study is built upon the emphasis that there is a general preference of firms to opt for a WOS (Mantecon et al., 2016), in case this option becomes feasible.

Table 1 Framework entry mode (JV vs WOS)

Resource

commitment Risks Control Profit return

JV low high low low

(14)

Schrader, Alicia 14

2.3. SMEs & entry modes

For SMEs, the entry mode decision is a particularly critical step. Scholars often argue that SMEs feature specific characteristics that influence their internationalization strategies (Cheng, 2008; Pinho, 2007; Laufs & Schwens, 2014).

Among these, SMEs’ resource availability is notably often associated with the foreign market entry mode choice as it determines the level of commitment the SME can make in the foreign market (Erramilli & D’Souza, 1993; Laufs & Schwens, 2014; Lin & Ho, 2019; Lu & Beamish, 2001; Oviatt & McDougall, 1994; Westhead et al., 2001).

Firm size is widely recognized as an appropriate proxy for firm resources (Ali & Camp, 1992; Bonaccorsi, 1992). In this regard, a large body of research has focused on the difference between SMEs and larger firms in terms of their resource availability (e.g., Erramilli & D’Souza, 1993;1995) and examined how this affects their internationalization behavior in general (e.g., Hessels & Parker, 2013; Hessels, 2008). At this juncture, firm-level resources were identified as an important antecedent of entry mode choice (Delios & Beamish, 1999).

The ability to execute higher ownership modes depends on the firm’s capability and predisposition to commit resources (Brouthers & Hennart, 2007; Buckley, 1989; Calof, 1994; Cui & Jiang, 2009; Johanson & Vahlne, 1977). Aulakh and Kotabe (1997), for example, highlight that firms may be restricted in their ownership decisions and structural capabilities to operate in foreign markets when their resources are limited. Ripollés et al. (2012) show that resource constraints can limit SMEs’ ability to commit strongly to a foreign market by choosing high-commitment entry modes.

In this context, scholars specifically point to SMEs’ limits in terms of financial and human resources (Fink et al., 2008; Jansson & Sandberg, 2008). The disadvantages SMEs face due to their lack of internal resources is often referred to as “liability of smallness” (Aldrich & Auster, 1986; Maekelburger et al., 2012).

(15)

Schrader, Alicia 15

mode choice (e.g., Brouthers, 2013; Nisar et al., 2012). Particularly SMEs, due to the frequently mentioned liability of smallness (e.g., Bonaccorsi, 1992; Erramilli & D’Souza, 1993; 1995), might face the need to access external resources and augment their internal resource stock in order to be able to engage in foreign operations (Lindsay et al., 2017; Madhok, 1998).

Following his classification of internal and external influences, Root (1994), for example, refers to (1) home country factors as a major external influence and (2) company resource-commitment factors as a major internal influence on foreign entry modes.

2.4. Home-country factors

With regard to the home-country, Root (1994) claims that environmental factors in the firm’s country of origin influence the entry mode choice to foreign markets (Mayrhofer, 2004).

Harzing (2002), for example, shows that firms with different nationalities vary regarding their entry strategies. More specifically, research conducted by Kojima (1978) and Lecraw (1983) shows that tendencies in entry mode choices of Japanese companies and U.S. firms clearly differ. While Japanese firms tend to prefer to form JVs, U.S. firms seem to prefer to establish WOSs. These findings of national differences in entry mode decisions support the suggestion that home-country factors may affect firms’ entry mode decisions (Erramilli, 1996; Pan, 1997).

According to Wan and Hoskisson (2003), the home-country environment frames a firm’s institutional context. Its infrastructure displays an important determinant for firms’ internationalization due to its ability to influence the company’s opportunities to invest abroad. Indeed, entry mode studies report on the role of institutional factors, regarding these as an important antecedent to entry mode choice (e.g., Brouthers, 2002; Ojala & Tyrväinen, 2007; 2008;).

Laanti et al. (2017) indicate that particularly SMEs may depend on the home-country environment. For example, scholars suggest that home-country institutions, such as government assistance, are particularly important to SMEs’ internationalization process (Crick & Lindsay, 2015; Durmuşoğlu et al., 2012).

(16)

Schrader, Alicia 16

internationalization by compensating for their lack of resources and other difficulties associated with operating abroad (Guler & Guillén, 2010).

Concluding the above, scholars agree that home-country institutions may affect firms’ entry mode decisions and a substantial amount of research confirms that particularly for SMEs the institutional environment matters as to their opportunities to internationalize.

However, in the specific context of SMEs and entry mode, Laufs and Schwens (2014) point out the lack of research conducted to better understand the influence home governments have on SMEs’ entry mode decision. Particularly as to how the government might represent an important source of external resources for SMEs. The authors specifically point out a complete absence of empirical analyses of the influence of public subsidies on SMEs’ FDI entry mode choice (Laufs & Schwens, 2014). To the best of my knowledge, after reviewing the SME entry mode literature, this gap remains and a study empirically examining how SME public finance influences SMEs’ entry mode decision would highly contribute to filling this gap.

Yet, this gap is surprising considering that a recent paper of the OECD (2019a) reveals that in most of the OECD countries SME finance became an important component of the policy agenda and that many governments tend to develop new initiatives. Moreover, financial incentives are regarded as one of the most important and helpful governmental initiatives in terms of support for their internationalization by SMEs (Sauvant et al., 2014). Financial incentives are defined as direct financial advantages to home-country firms investing abroad(UNCTAD, 2001).

2.5. Company resource-commitment factors

(17)

Schrader, Alicia 17

However, SMEs are also frequently shown to be capable of establishing full ownership modes (e.g., Brouthers & Nakos, 2004; Maekelburger et al., 2012), which questions the explanatory power of absolute levels of resources in the entry mode context and calls for consideration of other measures of resources.

For example, scholars have started turning their attention to slack-related issues in the context of internationalizing firms (e.g., Lin et al., 2009; Sui & Baum, 2014; Tseng et al., 2007). It is frequently argued that slack resources rather than absolute resources are able to explain strategic behavior in the context of internationalization (Cyert & March, 1963; George, 2005; Penrose, 1959).

Slack resources in general are referred to as “stocks of assets that are owned or controlled by the firm” (Amit & Schoemaker, 1993, pp.35). Prior research of internationalization predominantly focused on financial and human slack resources.

Financial slack resources are referred to as financial resources that are “in excess of what is needed for a firm to meet its current commitments and support current sales levels” (Mishina et al., 2004, p. 1183). They represent unabsorbed slack, in other words, resources that are currently uncommitted and readily available for redeployment within a firm (e.g., Bradley et al., 2011; George, 2005).

Human slack resources, on the other hand, refer to the number of employees in excess of a firm’s operational needs (Bourgeois, 1981; Mishina et al., 2004).

Research building on the RBV highlights that in contrast to absolute resources, the deployment of slack resources does not interfere with the firm’s daily operations. Unlike the possession of absolute resources that reflect the firm’s day-to-day operations, resource surplus remains after investing the resources required to keep the current level of activity. Therefore, slack resources are more directly tied to a firm’s ability to pursue expansion or growth strategies, such as internationalization (Cyert & March, 1963; Penrose, 1959).

For example, Cheng and Lin (2012) argue that only by means of slack resources, the firm is able to experiment with non-daily-business activities, such as foreign market expansion. Similarly, Penrose (1959) states that the availability of growth opportunities is an indication of the firm’s unused productive resources.

(18)

Schrader, Alicia 18

Second, opponents of slack resources view them as bringing inefficiency, inhibiting organizational growth, innovation, and risk-taking, by causing agents to indulge in satisficing, politicking, turf protection, and self-serving behavior to preserve the status quo (Cyert & March, 1963; Daniel et al., 2004; Jensen, 1986).

In both cases, it is claimed that slack resources are decisive for the firm’s internationalization strategies. As this type of resource is not needed for the current operations, it may provide a theoretically more justifiable basis for explaining decisions related to growth strategies than absolute resources do. For this reason, scholars increasingly started to consider slack resources in the SME internationalization context as well (e.g., Dasi et al, 2015; Kiss et al., 2017; Sui et al., 2014). This study will tie up to this trend and incorporate the impact of slack resources on SMEs’ entry mode decisions.

Examining the moderating effect of SMEs’ financial and human slack resources on the relationship between public finance and SMEs’ entry mode decision represents a response to call for future research by Laufs and Schwens (2014) with respect to a more comprehensive consideration of resources in the context of SMEs’ entry mode decision. More specifically, it remains unclear how SMEs are able to overcome the barriers related to the constraints they face in this context. As the authors specifically highlight the absence of empirical evidence for the effect of public finance on SMEs’ entry mode decision, this study’s purpose is to bring novel insights by integrating internal factors, namely SMEs’ financial and human slack resources, to that relation. Foremost, by considering new measures of firm resources, slack resources, this study aims at enhancing the comprehension of the joint effect of external and internal resources in SMEs’ entry mode decisions.

3. Hypotheses Development

Drawing on arguments from the RBV (Cyert & March, 1963; Penrose, 1959), this study will first focus on the effect of public finance on the entry mode decisions of SMEs and establish hypotheses as to the relationship between them. In the following section, the moderating effect of financial and human slack resources on the aforementioned relationship will be hypothesized.

3.1. The main effect of Public Finance on entry mode decision

(19)

Schrader, Alicia 19

When internal financial resources are missing, firms may be forced to access financial resources externally. However, firms that face constraints as to the availability of internal financial resources, such as SMEs (Bannó et al., 2014), are associated with supplementary obstacles with respect to the access of external financial resources.

They face difficulties in obtaining the debt for internationalization activities, which in turn leads to higher costs related to their internationalization (Cooper et al., 1994).

More specifically, due to the presence of agency problems, that are particularly present in the case of small firms (Myers, 1997; Stiglitz & Weiss, 1981), lenders feel forced to rely on collateral when borrowing debt to this type of firm. Typically, a firm’s capital is used as collateral for debt (Etemad, 2004). Hence, SMEs that generally possess less capital, lack the required collateral (European Commission, 2007b) and are accordingly faced with high barriers for accessing external sources of finance, such as lending, credits, and private debt (Chittenden et al., 1996; Myers, 1997). However, this debt is necessary for any type of firm to engage in foreign operations (Bannock & Peacock, 1989).

Consequently, SMEs willing to invest in foreign markets rely on access to other, less costly forms of external finance, such as public finance.

Public finance provides several advantages for SMEs related to the access of finance and thus, barriers in the context of the entry of foreign markets. In contrast to traditional external debt, Bannó et al. (2014) indicate that governmental SME finance provides SMEs access to capital at a lower cost while no guarantees are required. Furthermore, Feldman and Kelley (2006) found that receiving a governmental grant increases firms’ access to funding from other external sources which implies that the effect of public finance on SMEs’ finance is twofold.

A firm willing to engage in higher FDI modes needs to commit more financial resources. The level of commitment directly corresponds to the level of ownership of the entry mode (Etemad, 2004). In other words, full ownership modes, thus WOSs, require higher financial resources compared to shared ownership modes, thus JVs. By financially compensating for missing resources (Bannó & Sgobbi, 2010), public finance has the power to encourage SMEs to use full rather than shared ownership modes in their foreign entry. SMEs in possession of the required financial resources for WOSs do not rely on sharing the financial commitment with JV partners, and may prefer the full ownership mode. This mode provides firms with full control over the foreign operations, reduces the risks associated with the behavior of JV partners and is associated with higher profit return (see Table 1).

(20)

Schrader, Alicia 20

sources of external finance. In turn, the availability of financial resources increases SMEs’ probability to enter foreign markets with WOSs and decreases the probability to opt for a JV. Subsequently, this study hypothesizes the following:

H1: Public finance increases the probability of SMEs to opt for a WOS rather than a JV.

3.2. The moderating Effect of Financial slack

Even though SMEs are generally associated with resources limitation, they may still vary in their individual levels of initial internal financial resources. This raises the question of how internal financial resources contribute to the aforementioned relationship?

Here, levels of slack resources, rather than absolute resources, act as determinants of investment decisions in foreign markets as these represent a buffer in outputs and are thus the ones used by the SME to launch internationalization (Mishina et al., 2004). A firm may, for instance, possess high levels of absolute financial resources but the deployment is largely budgeted for daily business activities. In this case, the level of financial slack is low. Generally, the level of financial slack is highly associated with SME’s decision as to the commitment in FDI.

For example, George (2005) argues that financial slack resources ease the firm’s capital restrictions. In other words, SMEs possessing a substantial amount of financial slack face fewer barriers accessing other sources of external finance, such as lending, credits, and private debt, and subsequently rely less on the access to public finance provide by the government.

However, in contrast to public finance, other forms of external finance entail high costs of capital deployment (Buch et al., 2010). Traditional forms of external finance imply dividend payout on equity or interest payment on debt, leaving SMEs with pressures to meet the debt liabilities and to stay profitable. These pressures often impede firms to initiate new growth activities (Tseng et al., 2007), such as FDI.

(21)

Schrader, Alicia 21

Accordingly, I argue that higher degrees of financial slack decrease SMEs’ chances to receive public subsidies which in turn increases their need to access more costly traditional external finances. In turn, the positive effect of public finance on SMEs’ likelihood to opt for a WOS as foreign entry mode is eased. Yet, I argue it increases SMEs’ likelihood to choose JV to spread financial commitment and risks to be better be able to deal with the debt liabilities they are likely to face. Therefore, this study hypothesizes the following:

H2: The degree of financial slack negatively influences the effect public finance has on SMEs’ probability to opt for a WOS rather than a JV.

3.3. The moderating Effect of Human slack

Resources, however, should not be generalized but regarded separately. Not only financial resources but also human resources are crucial for firms’ internationalization. Human slack resources represent a widely accepted preconditioned firm capability to recognize and exploit business opportunities (Cerrato & Piva, 2010).

They refer to the number of employees the firm has in excess of its operational needs (Bourgeois, 1981; Mishina et al., 2004). For this reason, they provide the SME with a buffer in outputs and are thus the ones used by the SME to engage in foreign market expansion. Generally, when a firm is willing to increase its commitment in foreign markets, it can be inferred from this that the number of employees within the firm engaged in managing these international activities also increases (e.g. contact with clients and suppliers, management of commercial and productive subsidiaries, etc.) (Cerrato & Piva, 2010).

Moreover, following the reasoning of Calof (1994), a lack of human resources available to send abroad implies a burden for firms to set up businesses in foreign markets by themselves. Consequently, firms lacking human resource slack may be forced to establish JVs in order to access the required human resources from the JV partner.

Under the assumption that WOSs present the preferred entry mode due to the lower risks and higher profits this mode entails compared to JVs, I argue that SMEs being granted the external financial resources by the government and at the same time possessing the human slack necessary for full ownership entry modes are more likely to opt for a WOS than SMEs that lack this human slack.

(22)

Schrader, Alicia 22

qualitative human slack in form specialized knowledge and experience entailed in the human capital to manage the foreign entry.

The latter is an important aspect that refers to the consideration of absorbed human slack resources as firms’ intangible assets established through accumulated human capital which is considered highly specialized, skilled, and rare (Mishina et al., 2004). Especially in foreign operations, human capital is regarded a scare strategic resource as it counts in highly valuable specialized knowledge and experience of the workforce in the foreign market (Brambilla et al., 2012; Gomez-Mejia, 1988; Love & Roper, 2015; Tookey, 1964) that is difficult to imitate (Dutta et al., 2016; Zhang et al., 2018). Therefore, this type of resource requires special protection (Barney, 1991). Taking up the aforementioned risks that are involved in JVs (see Table 1), this entry mode option becomes even less desirable. Due to the risks related to the possible opportunistic behavior of the JV partner, SMEs possessing higher levels of unique human slack may show a higher preference for WOSs since this entry mode typically provides better protection of those scare resources.

To sum up, taking into account the availability of financial resources in form of public finance that is required for establishing a WOS, human slack further enhances SMEs’ likelihood to opt for a WOS rather than a JV because

(1) the feasibility of the WOS option increases due to the availability of necessary human resources for the full ownership mode, and

(2) the preference for a WOS increases in order to protect the human resources.

For this reason, a positive moderating effect of human slack on the relationship between public finance and entry mode is expected and the following is hypothesized:

H3: The degree of human slack positively influences the effect public finance has on SMEs’ probability to opt for a WOS rather than a JV.

(23)

Schrader, Alicia 23

Figure 1 Conceptual Framework

4. Methodology

4.1 Data and sample selection

This study aims at providing empirical evidence for the afore developed hypotheses based on theoretical arguments from the RBV found in the state of the literature of SMEs and entry modes. A single-country cross-sectional data analysis of 264 foreign entries by Swedish SMEs between 2007 and 2017 will be conducted. While a multi-country approach would be highly beneficial, the OECD highlights that due to differences in definition and coverage for the SME public finance indicators, comparability across countries is hampered. However, at the individual country level, the scoreboard represents a useful data source to compare SMEs’ access to finance over time and the changing conditions for SME financing (OECD, 2019a).

Sweden as a sample country is particularly interesting because SMEs dominate the Swedish business landscape by covering 99% of all the limited liability companies and accounting for 60% of the Swedish employment (OECD, 2019a).

Additionally, the Scoreboard for Sweden from the OECD report provides rich data in terms of public support for SMEs’ finance.

The firm-level data was taken from 2 different Bureau van Dijk databases: Zephyr for the foreign entry data and Orbis for the respective firm-specific data. Using Zephyr, I collected data for transactions made by Swedish SMEs with non-Swedish companies, that is, a company with its headquarter located outside Sweden.

With respect to the classification of SMEs, this study used the European Union definition. While the range of the number of employees of a SME is between 10 and 249, the turnover should be above EUR 2 million and not exceed EUR 50 million (OECD, 2005).

Financial slack

Ownership mode Public SME finance

(24)

Schrader, Alicia 24

This is not only the most commonly used definition in management studies (Claessens et al., 2000; Hall et al., 2009; Musteen et al., 2010), in the context of this study it is also the most feasible one as the data for the independent variable has been collected from an OECD report. In this regard, verifying that the sample firms, whose foreign entries were operated in different years, accord with the definition of SMEs, I ensured that pre-deal firm size information was taken.

Furthermore, the selection of the sample FDI transactions was based on 2 criteria: (1) the Swedish SME does not have an initial stake in the foreign company, and (2) the selected deals permit the acquiring firm to take control of at least 10% of the stake (Cuyper et al., 2015), as an investment below 10% (OECD, 2008) might be considered a portfolio investment rather than a FDI (López-Duarte & Vidal-Suárez, 2013).

During the sample selection process, I made sure no company was involved in several deals within the selected period to ensure the independence of the observations.

According to a rule of thumb suggested by Peduzzi (1996), the minimum sample size should be estimated as follows:

𝑛 =(10𝑥 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒𝑠) 0.50

Based on the equation above my sample should include at least 180 observations. Thus, the current sample consisting of 264 deals provides a sufficient amount to work with.

The resulting sample is unbalanced with respect to the number of deals each year (2007-2017) by ranging from 17 to 34 deals per year (see Appendix 1).

4.2 Measurements

4.2.1. The dependent variable

In this study, the entry mode decision, more precisely the ownership mode of the FDI, represents the dependent variable. To maintain homogeneity with the majority of empirical studies in the entry mode literature, the study distinguishes the entry modes in terms of full and shared ownership at a threshold of 95% ownership (e.g., Cui & Jiang, 2009; Hennart, 1991; Makino & Neupert, 2000; Yiu & Makino, 2002).

(25)

Schrader, Alicia 25

Accordingly, a JV is established when the company owns more than 10% but less than 95% of the foreign unit’s equity and is coded = ‘‘0’’, while a WOS is established when the company owns more than 95% of the investment and is coded= ‘‘1’’.

4.2.2. The independent variable

A recent OECD paper provides country-level data for a range of government policy instruments to foster SME access to finance, such as government loan guarantees, direct lending to SMEs, subsidized interest rates, and SME Banks (OECD, 2019a). It was found that guarantees remain the most widely used tool to strengthen SME access to traditional bank financing (OECD, 2019a). This study will use data from the OECD report and measure public finance in terms of yearly provided “direct government loans” particularly for SMEs by the Swedish government for 2007-2017 (OECD, 2019a).

In the OECD report, the sum of SME direct governmental loans is reported in the Swedish currency. In order to maintain consistency as for the currency within the study, the amount was converted to EUR using the exchange rate of the 31st of December of the respective year. 4.2.3. The moderator

This study includes 2 moderating variables, the firm’s financial slack resources, and the firm’s human slack resources. With respect to the calculation of the financial slack, this study follows the most commonly used operationalization of high-discretion financial slack, the current ratio (Bradley et al., 2011). The human resource slack is calculated as the ratio of employees to total sales, similar to the approach of previous studies (Kiss et al., 2017).

𝐹𝑖𝑛𝑎𝑛𝑐𝑖𝑎𝑙 𝑟𝑒𝑠𝑜𝑢𝑟𝑐𝑒 𝑠𝑙𝑎𝑐𝑘 = 𝐶𝑢𝑟𝑟𝑒𝑛𝑡 𝐴𝑠𝑠𝑒𝑡𝑠 𝐶𝑢𝑟𝑟𝑒𝑛𝑡 𝐿𝑖𝑎𝑏𝑖𝑙𝑖𝑡𝑖𝑒𝑠

𝐻𝑢𝑚𝑎𝑛 𝑟𝑒𝑠𝑜𝑢𝑟𝑐𝑒 𝑠𝑙𝑎𝑐𝑘 = 𝐹𝑖𝑟𝑚 𝑒𝑚𝑝𝑙𝑜𝑦𝑒𝑒𝑠 𝐹𝑖𝑟𝑚 𝑠𝑎𝑙𝑒𝑠

4.2.4. Control variables

(26)

Schrader, Alicia 26

Industry

The firm’s industry type is frequently identified as a determinant for entry mode decisions (e.g., Brouthers & Brouthers, 2003; Erramilli & D’Souza, 1993; Erramilli & Rao, 1993; Ripollés et al., 2012; Shrader et al., 2000). There seem to be differences between service and manufacturing firms in choosing a suitable foreign market entry mode (Blomstermo et al., 2006; Erramilli, 1989; 1991; 1992; Erramilli & D’Souza, 1993; Erramilli & Rao, 1993). This study used the NACE codes for industry classification provided by the Orbis database to classify the firms’ industry. A dummy control variable “industry type” was established with 1= if the acquiring company is in the manufacturing industry, and 0= otherwise.

Firm age

Consistent with prior SME internationalization studies (e.g., Lu & Beamish, 2001), this thesis also controlled for firm age. Firm age is an important control variable that needs to be established since it is a well-accepted proxy for experience and extend of network and legitimacy (Ouimet & Zarutskie, 2014). Younger firms may prefer lower ownership to better deal with the risks, to gain access to resources and knowledge from local partners, and benefit from lower entry costs (Musteen et al., 2009). Firm age was calculated as the difference between the founding year and the year of the foreign entry.

Firm size

Studies examining SME internationalization frequently point out to the importance of controlling for firm size (e.g., Lu & Beamish, 2001) as it was shown to influence firms’ entry mode decision (Contractor, 1984; Zahra et al., 2000). Lager companies typically prefer full control entry modes (Sanchez-Peinado et al., 2007). In line with approaches by other researchers, firm size is measured by the total amount of employees of the parent company (Freeman et al., 1983). Following previous studies (e.g., Dai et al., 2014; Zahra et al., 1997), the size variable was log-transformed to help alleviate skewness.

Firm performance

(27)

Schrader, Alicia 27

I ensured that the information taken was pre-deal data to ensure that the effect is incorporated accordingly.

𝑅𝑂𝐴 = 𝑁𝑒𝑡 𝑖𝑛𝑐𝑜𝑚𝑒 𝑇𝑜𝑡𝑎𝑙 𝑎𝑠𝑠𝑒𝑡𝑠

GDP growth

To control for economic factors of the Swedish home-country that differ between the years involved in the cross-sectional data set and that might influence the firms’ entry mode decision, the study also includes a control variable for GDP growth. Previous studies (e.g., Barry et al., 2003; Dunning et al., 2001; Liu et al., 2005) indicate a positive relationship between OFDI and GDP. It is argued that a stronger home economic environment enhances a firm’s international competitiveness (Deng & Yang, 2015). GDP growth was measured by the annual Swedish growth rate of GDP of the respective deal year to capture the effect on the ownership decision. Data for this variable was obtained from the World Bank database.

Years

Lastly, in order to control for the years that are covered in this study (2007-2017), a clustered robust standard errors logistics regression model was applied.

4.3. Analytical Procedure and Statistical Model

Because of the dichotomous nature of the dependent variable, the statistical method that was used to test the hypothesized relationship is a binary logistic regression (Canabal & White, 2008). This is in line with prior studies examining entry mode decisions (e.g., Chen & Hennart, 2004; Slangen & Hennart, 2008; Wang & Schaan, 2008). As mentioned before, to account for potential bias due to the different years used in this cross-sectional dataset, the study opted for a model with clustered robust standard errors, which allows to account for potential correlation between the residuals (Nichols & Schaffer, 2007).

(28)

Schrader, Alicia 28

relationship with the predictor variables (Bewick et al., 2005). In other words, it computes the probability of the selected response (WOS) as a function of the values of the explanatory variables (Sarkar et al., 2011).

Hence, the following estimation is the baseline of this study’s approach:

Pr (WOS=1) = α0 + ß1 Public Finance+ ß2 Financial Slack + ß3 Human Slack + ß4

Pub_Fin#Fin_Slack+ ß5 Pub_Fin#Hum_Slack + ß6 Firm size+ ß7 Firm age + ß8

ROA + ß8 GDP growth + ß9 industry + Year effect + 𝜀

With α0 representing the constant term, ßs the coefficients of interest, and 𝜀 the error term. Based on the formula above, a positive coefficient implies that SMEs tend to favor WOSs, whereas a negative coefficient indicates that they are more likely to choose a JV.

In order to enhance interpretability, I standardized the predictor variables public finance, financial slack, and human slack. The three variables are measured on different scales which inhibits the comparison of the effects. Standardization denotes that the variables are rescaled to have the mean of zero and a standard deviation of one. Due to the rescaling, the effects of the variables can be better compared with each other (Bring, 1994).

The STATA 16 program is used to run the logistic regression. However, this is tied to fulfilling certain assumptions further explained hereinafter.

4.4. Evaluation of the Method Assumption

First, a logistic regression assumes a linear relationship between any continuous independent variables and the logit transformation of the dependent variable (Peng et al., 2002). Therefore, the model was checked for the presence of a specification error (Bewick et al., 2005). For this purpose, I conducted a linktest on the full model. The linktest is a STATA tool that helps to identify whether the model is properly specified (e.g. whether relevant variables are omitted or the link function is not correctly specified). The results (see Appendix 2) indicate that the model is well specified (the prediction squared is insignificant), which means that no important variables in the model are omitted or irrelevant variables are included in the model (Torres-Reyna, 2007).

(29)

Schrader, Alicia 29

the null hypothesis (i.e. the observed and expected proportions are the same across all groups) is not rejected. However, the results of the Hosmer–Lemeshow should be interpreted with caution as the sample size is smaller than 400 (Bewick et al., 2005). After checking the Nagelkerke R2 across all specified models, the full model (Model (6)) revealed the best model fit, however, the value of 0.129 of this measure is not specifically high either.

Third, the data must not show multicollinearity. To check for assumption, the Pearson Correlation Matrix (see Table 2) gives some initial important indications. As suggested by Cohen (1988), coefficient values below 0.3 indicate acceptable small pairwise correlations between variables. Table 2 reveals that the pairwise correlations between human slack and firm size, between financial slack and firm age, as well as between human slack and financial slack require some attention.

For this reason, I checked for the variance inflation factor (VIF) of the variables. The VIFs measure the extent to which the variances of the coefficients estimated in a regression are inflated when compared to the cases in which the independent variables are not linearly related (Midi et al., 2010). High VIF values indicate the presence of multicollinearity. The Collinearity Diagnostics in Stata showed that all the VIFs are below 1.46 with a mean VIF of 1.17 (see Appendix 3) which is below the recommended threshold of 10 (e.g., Dormann et al., 2013; Hair et al., 2001; Neter et al., 1985). For this reason, no problems with multicollinearity were detected.

Fourth, there should be no significant outliers, high leverage points, or highly influential points. A good way of looking at them is to graph the standardized Pearson residuals, deviance residuals, and leverage residuals against the predicted probabilities. The scatter plots in the appendices 4, 5, and 6 show the results of the analysis. After some observations (such as 61 and 211) appeared to represent influential observations, I compared the logistic regression results of the model including the observations and the model excluding the observations. However, the exclusion of the observations had an impact on neither the fit statistics of the model nor the specific parameter estimates. For this reason, all 264 observations were included in the model.

(30)

Schrader, Alicia 30

5. Analysis and Results

5.1. Descriptive Statistics and Correlation

Table 1 provides the descriptive statistics and the pairwise correlations for each of the variables used in the study. The original sample consisting of 264 foreign entries of Swedish firms contains 224 WOSs (84.85%), and 40 JVs (15.15%).

The firm age mean of roughly 30 indicates that the average firm age in this sample is 30, including SMEs between the age of 2 and 175 at the point of the transaction. Considering that firm size was log-transformed in this study, direct interpretation of the mean is rather difficult. Since the industry dummy was coded 1 for belonging to the manufacturing industry, the mean value of 0.398 of the variable indicates that 39.8% of the transaction were operated by acquirers from the manufacturing industry. GDP growth in the 10 years was 1.972 on average. The negative minimum value of -4.236 may be explained by the global financial crisis (2007/2008) which caused highly negative GDP growth rates in Sweden in the years 2008 and 2009. The mean ROA of the sample firms is 0.050.

As for the main predictors of interest, the mean public finance in the 10 years is -0.006, the mean financial slack is 2.30e-08 and the mean human slack is -0.0005 in the present sample. The standard deviation of all three means is close to 1 as a result of the standardization of the variables.

(31)

Schrader, Alicia 31

Table 2 Descriptive Statistics and Pearson Correlation Matrix

5.2. Results of the Regression Analysis

The results of the regression analysis are shown in Table 3. Six different models were specified. In contrast to linear regressions, the R2 is not accepted in logistic regressions because it shows

(32)

Schrader, Alicia 32

the variance extraction by the independent variable (Hu et al., 2006). In logistic regressions, the estimates are maximum likelihood estimates that are not calculated to minimize the variance but are split into two categories. For this reason, to evaluate the goodness-of- fit a pseudo R2 has been developed. The Nagelkerke R2 is a common proxy used to compare the goodness-of-fit of different models in a logistic regression (Hu et al., 2006). As addressed earlier, this study compared the results of the different models specified and reports the values of the Nagelkerke R2 (see Table 3). While the values are not particularly high in either of the models, when comparing the values for the six models, one can see that the values increased from 0.084 in Model (1) to 0.129 in Model (6), which indicates an improved fit of the full model compared to the previous models. An R2 of 0.129 practically means that the model explains 12.9.% of the variability in the scores (Freese et al., 2006).

The regression coefficients that can be retrieved from Table 3 express the relative importance of each of the independent variables in standardized terms (Menard, 2011). In other words, they estimate the impact of the explanatory variables on the probability that the foreign unit is wholly-owned by the Swedish SME while holding the other variables constant. A positive coefficient indicates that the corresponding independent variable tends to increase the probability that a WOS was chosen, while a negative coefficient indicates that the independent variable tends to increase the probability of a JV arrangement.

Model (1) includes the five control variables. The results in Table 3 show that, with ROA being an exception, the variables indicate a positive beta. However, GDP growth is the only significant estimate at the 0.1% level (ß=0.157). The other control variables show insignificant results.

Model (2) introduces the two moderating variables in order to be able to separately test the moderating effect later with the introduction of the interaction terms in Model (4), (5), and (6). With the introduction of the moderators in a separate model, the main effect of them was tested. The results in Table 3 show that both, financial slack (ß= .177) and human slack (ß=-0.229) are insignificant predictors for the probability of an SME to opt for a WOS (both p-values are above 0.05).

(33)

Schrader, Alicia 33

level. For this reason, Hypothesis 1 can be confirmed: the greater the public finance, the higher the probability of SMEs to opt for a WOS rather than a JV.

The marginal effects of public finance were plotted and can be depicted in Figure 2. Marginal effects can be a useful tool in logistic (nonlinear) regressions to make the effects of variables more intuitively meaningful (Williams, 2012). They measure the effect a change in one of the regressors has on the conditional mean of the dependent variable (Cameron & Trivedi, 2010). To illustrate the marginal effects of public finance, I calculated the marginal effects of representative values (MERs) (Williams, 2012). With the MERs, I set a range for the public finance measure and calculate the predictive probabilities for the values along that range. This means that at each stage the effect of the predictor on the probability of Y is estimated while holding all other predictors at its mean (Williams, 2012). Figure 2 below depicts the plotted results of the calculation.

Figure 2 Predictive Margins of Public Finance on the entry mode=”1” (WOS)

From the plotted margins we can see that at the lowest level of public finance in the Swedish sample the probability of WOS is at around 78%, while at the highest levels of public finance it is at around 87%. For detailed margins see Appendix 7.

.75 .8 .85 .9 Pr(W O S)

low medium high

(34)

Schrader, Alicia 34

Model (4) was specified to test the first moderating effect of financial slack on the relationship between public finance and the entry mode. For this reason, at this point, the interaction term of financial slack and public finance was introduced. Hypothesis 2 proposed that financial slack negatively moderates the relationship between public finance and the probability of SMEs to opt for a WOS. While the ß-coefficient of the interaction term is positive (.863), it is statistically insignificant (p>0.05). This denotes that Hypothesis 2 does not find support.

Model (5) was established to test Hypothesis 3. The results of the introduction of the interaction term of public finance and human slack reveal a negative moderating effect of human slack on the relationship between public finance and entry mode. The interaction term’s negative ß-coefficient (-0.253) is significant at the 0.1% level. Hence, while a positive moderating effect of human slack was predicted, the results show that higher levels of human slack mitigate the positive effect of public finance on the likelihood of a WOS.

For this reason, Hypothesis 3 must be rejected. Instead, the results indicate that the greater the human slack resources, the more negative the relationship between public finance and the probability of a WOS becomes.

However, the interpretation of interaction terms, respectively their effect, is less intuitive compared to covariates. The issue with the specification of the marginal effects of interaction terms is that the value of the interaction term cannot change independently of the values of the two component terms. Hence, one cannot estimate a separate effect for the interaction (Williams, 2012). However, one can calculate the average marginal effects of one of its components. In order to shed more light on the moderating effect of human slack, I further analyzed its marginal effects. As financial slack was not found a significant moderator, I will refrain from further analysis of this effect.

(35)

Schrader, Alicia 35

Figure 2 Average marginal effects of human slack on entry mode

Model (6), the full model, incorporates both interaction terms. It can be observed that the effect of the main predictor public finance remains positive (ß= 0.374) and significant at a 0.1% level. The moderating effect of financial slack remains insignificant and the negative moderating effect of human slack (ß= -0.274) remains negative at a 1% significance level. No significant changes in the other variables can be observed.

-. 1 5 -. 1 -. 0 5 0 .05 Ef fe ct s o n Pr(W O S)

low medium high

(36)

Schrader, Alicia 36

Table 3 Results for binary logistic regression model: WOS (coefficients)

(37)

Schrader, Alicia 37

5.3. Robustness Tests

In order to ensure the robustness of the results, I further conducted 3 robustness tests.

First, to test that the results do not change with the statistical tool used, I conducted a probit regression with the same models. While the logit model uses the cumulative distribution function of the logistic distribution, the probit model uses the cumulative distribution function of the standard normal distribution (Hoetker, 2007). Probit is thus an alternative to my original approach and is also frequently used by researchers for binary dependent variables (e.g., Lin et al., 2010; Mahagaonkar, 2008). For detailed results of the probit regression see Appendix 9. Overall, the probit reveals somewhat equal results to the logit along all models which supports the robustness of my model.

Second, given that the real effects of the financial crisis manifested strongly in 2009 and 2010 in European (and Sweden) trade activity, observations from these specific years might bias the results. This study tried to address such issues by including robust standard errors for the year fixed effects in the logistic regression. However, as a robustness test, I also removed firm-year observations from 2009 and 2010 and rechecked the estimations (e.g., Paeleman et al., 2019). In accordance with the original results, the new results reveal a positive effect of public finance on the entry mode and a negative moderating effect of human slack on that relationship. Unsurprisingly, GDP is not a significant predictor of the entry mode anymore. Overall, the results indicate the robustness of my model. Results of the full model can be found in Appendix 10.

Third, previous studies examining the effect of financial slack have used different indicators to measure financial slack. Rafailov (2017) points out that besides the one used in this study, scholars have also used the level of working capital to measure financial slack. More specifically, by means of the robustness test for this study, the ratio of working capital by sales (WCS) will substitute the prior financial slack measure (Bourgeois, 1981).

The estimation is the following:

𝑊𝐶𝑆 =LMNNOPQ RSSOQSHLMNNOPQ TURVUTUQUOS QWQRT SRTOS

(38)

Schrader, Alicia 38

6. Discussion and Conclusion

SMEs’ liability of smallness is often argued to restrict their degree of freedom in decisions with respect to growth options in internationalization (Aldrich & Auster, 1986; Maekelburger et al., 2012). However, we know little about how external resource augmentation through the home government may increase SMEs’ propensity to engage in higher commitment modes of FDI. In order to concurrently examine how SMEs’ entry mode decision is influenced by the level of external resources available while taking into consideration their internal resources, this study explored the relationship between public finance for SMEs and their entry mode decision in terms of ownership mode (JV vs. WOS) and how this relationship is moderated by the degree of financial and human slack resources available.

By drawing on arguments from the RBV, I tested the elaborated predictions on a sample of 264 foreign entries of Swedish SMEs in the period of 2007-2017. According to the results, public finance has the power to influence SMEs’ entry mode decision, in a way that it increases the probability to opt for a WOS rather than a JV. Human slack moderates this relationship, however, contrary to what was proposed, in a negative way, whereas financial slack does not moderate this relation indicated by the insignificant results. Considering the theoretical gap identified in the progress of the literature review of SMEs and entry mode decisions, the results of this study advance extant research and reveal several remarkable insights that entail novel practical implications.

The positive influence of public finance on the probability of SMEs to enter the foreign market via WOS rather than via JV emphasizes the context-dependency of SMEs’ entry mode choice. Scholars usually argue that firms lacking the necessary resources tend to choose lower investment modes (Paul al., 2017), such as JVs that provide access to the partner’s resources. However, once the WOS becomes a feasible entry mode option, it is the preferred option due to overweighting benefits in terms of control, better protection of valuable firm resources and capabilities (Erramilli & Rao, 1993), and higher profit returns (Brouthers, 2002).

Public finance and SME foreign market entry mode choice: The moderating effect of slack resources

Master Thesis (IB&M)