Antecedents and Outcomes of Network Involvement in the Internationalization Process: A Case of SMEs from the USA, China, and Russia.

The internationalization processes of small and medium-sized enterprises (SMEs) differ across economic and cultural environments. Creating an operationalizable and predictive framework to explain them has long been a challenging research task. This task is particularly relevant in an era of reconfiguration of globalization, which directly affects the small business sector. This study proposes a model which includes networking, attitudinal, cultural, and environmental factors as antecedents of the degree of SME internationalization. We collected data and tested this model in three distinct cultural environments: the USA, China, and Russia. The results suggest that attitudinal characteristics of managers, such as global mindset and relationship commitment, condition SMEs network involvement. The influence of network involvement on the degree of SMEs internationalization becomes stronger with increasing environmental turbulence. This study’s findings have practical implications for businesses operating in different countries, as well as governmental organizations and educational institutions.


Introduction
In the last few decades, the internationalization of small and medium-sized enterprises (SMEs) has been a hot topic among scholars (Coviello & Munro, 1995;Etemad & Wright, 2003;Matlay et al., 2006). The main reason for this interest is the role that SMEs play in the world economy, contributing up to 60% of the workforce and 40% of GDP (Bell, 2015), and the multiplicity of the pathways and antecedents of their internationalization. The current tectonic changes in the world economy will likely attract further interest in the internationalization of SMEs on the part of both scholars and practitioners.
Major research questions regarding SME internationalization in the following years will likely be focused on the degree to which globalization reconfiguration will affect the small business sector. Some researchers point to the beginning of a virtual revolution that will lead to comprehensive de-globalization (Sułkowski, 2020) or a new disruptive paradigm in globalization (Díaz et al., 2020). Others argue that new patterns of world trade will be introduced, which will make some businesses, industries, and economies winners and some losers (Barua, 2020).
In the current environment, a host of research questions are particularly relevant. Will the effect of the global crisis on SME internationalization be critical and long-term or relatively easy to overcome and short-term? Will it hamper the efforts of millions of SMEs to enter the global market or just modify the pathways of these efforts? Will SMEs become a factor further eroding the fabric of the world economy or, on the contrary, reinforce it by creating new non-traditional internationalization links? Finally, after the recession, will the small and medium-sized business sector move forward or fall behind other sectors in reinstating international business networks?
These questions and the necessity to provide answers have aroused a renewed research interest in the mechanisms of SME internationalization to explore the forces that drive this process, other than current market opportunities and immediate profit considerations (Alayo et al., 2019;Nummela et al., 2020;Supardi, 2020). These questions can be approached from various theoretical perspectives, one of which is a networking approach that attempts to establish the causality between the internal (domestic) and external (global) networks of which an enterprise can be a part (Galkina & Chetty, 2015;Hånell et al., 2018).
Considering the importance of the research topics mentioned above, this study focused on the networking mechanisms of SME internationalization and their antecedents in managerial practices. We aimed to contribute to creating a testable and quantifiable framework to explain networking in the context of SME internationalization, thus continuing the stream of research examining whether and how networks matter in enterprises' internationalization (Musteen et al., 2014;Sedziniauskiene et al., 2019;Senik et al., 2011). We used survey data collected in multiple stages from textile SMEs in three countries radically different economically and culturally: the USA, China, and Russia. This study also focused on a comparison between a developed (USA) and two emerging (China and Russia) economies.
The networking approach has its roots in social exchange theory, which views firms as parts of multiple internal (interpersonal) and external (interorganizational) networks (Chetty & Patterson, 2002). The basic conceptual component of the network model is firms' dependency upon resources possessed or controlled by other firms (Borgatti & Foster, 2003). The only way to gain access to these resources is to establish relationships with customers, suppliers, families, and friends within the network (Collinson & Houlden, 2005;Mort & Weerawardena, 2006).
The main benefit of networks in the process of internationalization is that they provide firms with market knowledge and help them identify new opportunities (Åkerman, 2015;Coviello & Munro, 1995;Tiwari & Korneliussen, 2018). Previous studies have established the relationship between a firm's networking and the marketing strategy that it pursues (Goldenberg et al., 2009;Stremersch et al., 2007). In the case of SMEs, these activities often lead to a reduced perception of the related risks and a stronger predisposition to commit resources to internationalization (Chetty & Patterson, 2002;Coviello & Munro, 1995). Networking relationships help small firms accelerate the internationalization process and gain the knowledge necessary for risk minimization (Holm et al., 1996).
Overall, the usefulness and applicability of the network approach to explaining the internationalization process of SMEs is well documented. However, it is not without limitations. A major limitation of the networking perspective noted by some studies is its explanatory rather than predictive nature (Malhotra et al., 2003). Nevertheless, it is one of the most dynamically developing schools of thought in internationalization research, especially related to the small business sector (Mort & Weerawardena, 2006). One of the most attractive elements of the network approach is that it can explain the non-rational and non-economic motives of small business owners and managers in-volved in the internationalization process (Chetty & Holm, 2000;Kalinic et al., 2014). However, Zahoor et al. (2020) identified gaps in internationalization research based on the networking perspective, noting that "we know little about how the quality and intensity of networks, as well as entrepreneurs' personality characteristics, can influence the effectual and causal approach to network building for internationalization" (p. 447). This study aimed to fill this gap.

Hypotheses Development
The theoretical framework upon which the hypothesized effects are based is social capital theory. Nahapiet and Ghoshal (1998) conceptualize social capital as resources that become available through networking. One of the key components of social capital is relational embeddedness characterized by trust (Musteen et al., 2014). Another core element is information benefit. A diversity of network links facilitates access to information (McKeever et al., 2014). Within this framework, we consider two basic variables that underlie these two core components of social capital. One is relationship commitment, which reflects the relational dimension of social capital acquired through network interactions (Rhinesmith, 1992), and the other is global mindset, which reflects the informational dimension (Nummela et al., 2003). The research model with the hypothesized relationships between the exogenous and endogenous variables is shown in Figure 1. Relationship commitment and global mindset are attitudinal variables and are derivatives of the personality traits of a business owner or manager. The choice of attitudinal constructs is consistent with the interactionist approach built on the tenets of social capital theory (Burt, 1997). This approach links personality characteristics to network dimensions such as size, functions, social role, and informational capabilities (Anderson et al., 2007). Global mindset is a person's predisposition to openness to the outside world conditions network diversity (i.e., network width) (Gustafsson et al., 2005), while relationship commitment contributes to the quality and functional role of the ties involved in it (i.e., network depth) (Ehret, 2004). Morgan and Hunt (1994) conceptualize relationship commitment as an understanding of the key importance of the relationship with an exchange partner and the related willingness to expend maximum effort on it. Ritter et al. (2004) distinguish between different relationship levels and posit that the network structure is conditioned by the strength and depth of these relationships. Mechanisms of distribution and balance of resources involved in a network are directly affected by relationship commitment. Galimberti and Wazlawick (2016) researched the influence of relationship clusters on software SMEs' networking and internationalization processes. Studies on the banking (Kassim & Abdulla, 2006) and manufacturing sectors (Braziotis & Tannock, 2011) have documented the empirical link between relationship commitment and networking. Thus, we propose the following hypothesis: H1: An SME's relationship commitment positively influences the degree of its network involvement.

Global Mindset
Kedia and Mukherji (1999) conceptualize global mindset as a sum of strategic and cultural factors on the individual and organization levels. It influences the size of a firm's network (Stam & Elfring, 2006), its limits and boundaries (Kyvik, 2011), and the ties needed to enter the global market (Oviatt & McDougall, 2005). Felício et al. (2016) documented the relationship between individual and corporate global mindsets and SMEs' internationalization process. Thus, our next hypothesis is as follows: H2: An SME's global mindset positively influences the degree of its network involvement.

Network Involvement
Network interactions create the conditions for leveraging resources through collaborations with external agents and provide a stimulus to search for new partners (del Carmen & Holgado, 2019; Maurer et al., 2011). The acquisition of social capital is accompanied by mimicking the networking patterns of other network participants (Aarstad et al., 2010). In the case of successful internationalization, this mimicry plays the role of an additional "pull" factor for network members to go global. This occurs when the network as a whole aims to enter new markets (e.g., a group of Taiwanese textile enterprises starting to export to China), and the networking collaboration mechanisms enhance the new market entry process for any given network member (Tellis et al., 2009). Even in the case of domestic networking relationships, smaller firms often form indirect links with a wider network of globally connected firms (Tiwari et al., 2016). Thus, we formulate the following hypothesis: H3A: An SME's internationalization is positively influenced by its degree of network involvement.

Mediator Effect
Previous studies have provided evidence in favor of the link between attitudinal characteristics of an SME's owner/manager and a firm's degree of internationalization (DOI) (Hutchinson et al., 2007;Kuemmerle, 2005). Based on this link, we argue that a firm's network involvement mediates the relationship commitment-DOI and global mindset-DOI relationships and suggest a model of mediating effects. A variable has a mediator function if it is both the result of the determinant factors and the antecedent of the result (Baron & Kenny, 1986). Thus, we propose two hypotheses particularly related to the network's mediator effect: H3B: Network involvement acts as a mediator between relationship commitment and SME internationalization.
H3C: Network involvement acts as a mediator between global mindset and SME internationalization.

Environmental Turbulence
Environmental conditions affect the collective and interorganizational actions of firms, especially SMEs (Ambroise et al., 2018;Rialp-Criado & Komochkova, 2017). Networking behavior is particularly prevalent in turbulent conditions (Coviello & McAuley, 1999). The social capital perspective explains the influence of environmental turbulence on the content and role of network relationships (Westhead et al., 2004). In turbulent conditions, networking interactions start playing a wider role in SMEs' decision-making process (Karami et al., 2020), and their importance as a source of capabilities, resources, and situation-specific knowledge increases (Kamasak et al., 2016). Therefore, we hypothesize a moderating effect as follows: H4: The link between an SME's degree of network involvement and its internationalization is stronger under a high level of environmental turbulence.

Moderating Role of Countries' Differences
The role of cultural differences in interorganizational networks has been researched extensively. Social capital such as "guanxi" (relationships maintained within a network in Chinese cultural contexts) has been considered in analyses of Chinese economic collaboration patterns (Nolan & Rowley, 2020;Wu & Wang, 2016). Recent studies on SME networks in Japan (Khare, 2012), India (Narasimhan et al., 2015), Brazil (Figal et al., 2015), Poland (Bylok et al., 2016), and Lithuania (Diskienė et al., 2018) have emphasized the influence of cultural factors. Evidence suggests that the content of network links and social capital associated with them differs between cultures. In societies with higher individualism scores (Hofstede, 1980), network links tend to be more formalized. Their role is usually limited to business functions and does not extend to personal life (Dodd & Patra, 2002). Conversely, in collectivistic cultures, the role of networks tends to be more comprehensive, as they provide not only business but also social support for their members (Nakata & Sivakumar, 2001). Therefore, we postulate that the role of networks in internationalization differs between highly collectivistic (China), moderately collectivistic (Russia), and individualistic (USA) cultures and propose the following hypothesis: H5: The link between an SME's network involvement and its degree of internationalization is stronger in China, followed by Russia, and weak in the USA.

Measurements
To measure the DOI, we used Sullivan's (1994Sullivan's ( , 1996 five-dimensional scale (percentages of foreign sales, assets, and overseas subsidiaries, managers' international experience, and number of psychic zones in which a company operates). For network involvement, we used Ostgaard and Birley's (1996) scale adapted from Birley et al. (1991).
The two attitudinal constructs have well-established measures: Kaufman et al. 's (2006) five-item scale for relationship commitment and Nummela et al. 's (2004) seven-item scale for global mindset. For environmental turbulence, we used Burton et al. 's (2002) three-item scale measuring predictability, complexity, and equivocality. Three control variables were measured: the number of a firm's employees, the time of its establishment, and the time of global market entry. The constructs' measurement instruments are presented in Table 1.

Sampling and Data Collection
According to Craig and Douglas (2000), a major sampling challenge in cross-cultural research is to maintain a balance between cross-national comparability and within-country representativeness. The major decision conditioned by this consideration is that the sample should represent enterprises from the same industry. We chose the Reliability and validity measures reported Firm internationalization 1) Three-year average of foreign sales as percentage of total sales (FSTS) 2) Three-year average of foreign assets as a percentage of total assets (FATA) 3. Overseas partners as percentage of total partners (OSTS) 4) Cumulative duration of firm managers' international assignments weighted by the reported total number of years of work experience of the management team (TMIE) 5) The dispersion of the firm's operations among the ten psychic zones of the world as identified by Ronen and Shenkar (1985) Express your agreement/disagreement with statements below: 1) Networking is the only way to achieve our growth objectives. 2) We will have to network in order to succeed in future. 3) It is important for our company to internationalize rapidly. 4) The company's management uses a lot of time for planning networking operations. 5) The growth we are aiming at can be achieved mainly through internationalization. 6) The founder/owner/management of the company is willing to take the company into international markets. 7) The company's management sees the whole world as one big marketplace. textile industry because in all three countries included in the study, it relies on SMEs to a high degree (Collinson & Houlden, 2005). Another important consideration is that the sample frames from all countries should be compatible. The selected sample frames were the Textile Yellow Pages for the USA, the Chamber of Commerce list of domestic manufacturers for China, and the list of Textile Trade Fair Participants for Russia, with 6,283, 17,565, and 1,621 firms, respectively. The data were collected via an internet survey of SME owners/managers in multiple stages between 2005 and 2018. The number of responses used in the study were 293, 244, and 287, respectively. The total sample size (824 respondents) fit the requirements of structural equation models. Summarized sample statistics are displayed in Table 2. 6. Analysis and Results

Data Pretest and Scale Purification
Based on the procedure proposed by Cadogan et al. (1999), pretesting was performed on a sample of 59 SMEs (14 from the USA, 22 from China, and 23 from Russia). We evaluated the wording of the questionnaire, cross-cultural appropriateness, and the dimensionality of constructs. The validity and reliability of the scales were pretested using Sin et al. 's (2005) methodology.

Cross-Cultural Stability, Reliability, and Validity
Reliability and discriminant and convergent validity issues were addressed using the same methodology as in the pretest stage. Cross-cultural stability was evaluated with factor analyses performed on the three countries' datasets separately. The reliability coefficients in each sample were higher than .70 for all multi-item variables (Nunnally, 1978).

Model Fit Indicators and Regression Estimates
Structural equation modeling was used as a hypothesis testing tool. The goodness of fit evaluation was based on Hair et al. 's (1992) cutoff criteria. The model showed good fit.

Mediation Effect
The mediating effect of network involvement in the relationship commitment-DOI and global mindset-DOI relationships (H3B and H3C) was tested based on Baron and Kenny's (1986) methodology. A significant relationship between global mindset and DOI was found, supporting H3B (β = .06, p < .05 with network involvement in the model and β = −.059, p = .248 without it). In contrast, the relationship commitment-DOI link was insignificant both with and without network involvement in the model (β = .037, p = .089 and β = −.002, p = .464, respectively), failing to support H3C (Figure 2

Moderating Effects
To test interaction effects (H4 and H5), we followed a procedure based on multiple group analysis and a nested goodness-of-fit strategy (Kline & Dunn, 2000). The turbulence summary score was converted to a grouping variable (high vs. low turbulence). The hierarchy of the models with increasing constraints on the number of invariant parameters (  significant in the high-turbulence group (β 2 = 1.14, p < .00) and insignificant in the low-turbulence group (β 1 = −.01, p < .75), suggesting a moderating effect, thus supporting H4 (Figure 3).
To test the significance of the interaction effect involving country variables, three groups of observations on the USA, China, and Russia subsamples were used. The difference between the CFI values did not meet the cutoff criterion of .01 (Δ CFI < .001; Table 3). Moreover, a path coefficient analysis for each of the three country subsamples revealed no substantial differences in cross-cultural model performance (Figure 4). Therefore, a moderating effect of the country variable (H5) was not confirmed, demonstrating the configural equivalence of the model across the three cultural subsamples.

Control Effects
The effects of three control variables (firms' age, size, and time of entry to a foreign market) were tested using Bettencourt et al. 's (2001) hierarchical regression method. Evidence of the unique variance beyond one explained by control variables was obtained (Table 4). 7. Discussion

Study Contribution
This study provides clear evidence of a positive link between the degree of network involvement and SMEs internationalization. We demonstrate that in studying the antecedents of SMEs gravitation to the world market, not only market conjuncture considerations but also networking and attitudinal factors such as global mindset and relationship commitment should be investigated. Accordingly, the intention to internationalize should be seen as conditioned not only by external (market) but also by internal (managerial and cultural) factors. Overall, our results are consistent with recent research showing that SME internationalization is a process influenced by multiple factors extending far beyond current market attractiveness ( Javalgi & Grossman, 2014;Schu & Morschett, 2017;Šarapovas et al., 2016). The main theoretical contribution of this study is that it takes another step in the direction of using the networking approach to explain SME internationalization. We attempted to investigate both outcomes and antecedents of SMEs' networking by incorporating it with attitudinal and environmental variables into a single framework. The proposed model was validated on a sample highly diversified culturally, historically, politically, and economically, including a developed economy (the USA) and two emerging economies, one of which (China) is the largest in the world and is based on the manufacturing sector, and the other is mid-range and based on exports of natural resources (Russia).
An unexpected result of this study was that we found no significant differences between the three countries in the strength of the network involvement-internationali-zation link. A possible explanation could be that cross-industry differences of this link prevail over cross-cultural differences-a phenomenon observed in other studies as well (e.g., Winsted, 1999)-and since all SMEs in the sample belonged in the same industry (textile), the link did not differ significantly between the three countries. To test this proposition, further comparative research with samples from different industries and countries is needed.

Study Limitations
The main limitation of this study is that only one industry was represented in the sample. Going forward, research should expand to other industries, first of all belonging to the knowledge-intensive cluster that is predisposed to and dependent upon networking relationships to a higher degree than the manufacturing sector (Coviello & McAuley, 1999;Galimberti & Zanella, 2019;Jain et al., 2015).
Another sampling limitation of this study is that it was based on samples from only three countries and therefore does not provide sufficient grounds for worldwide generalization. Further research should include more countries to test the model in more culturally, politically, and economically diverse environments. Analyses should also include countries representing different psychic zones as conceptualized by Sullivan (1994).

Conclusion
In a broad macroeconomic sense, this study's results provide insights into the mechanisms predisposing SMEs to internationalization under COVID-19 pandemic conditions as well. As a contributor to turbulence in the world economy, the pandemic can be considered a factor that either disrupts internationalization processes or lays the groundwork for a new digital globalization from which many SMEs can benefit (Schilirò, 2020).
Based on the study's results, one can speculate that the global economic crisis caused by the COVID-19 pandemic will not fundamentally affect the SME internationalization process in the long term but will become a powerful, albeit secondary (compared to the predisposition to network involvement), factor. Environmental turbulence acts as a moderator rather than a causal trigger of network involvement and internationalization. Boosting SMEs' performance through internationalization, given proper consideration and support by national governments, may become a major factor in stabilizing nations' socioeconomic situation after the pandemic, as well as one of the leverages of global economic recovery as a whole.