Introduction
Small and medium-sized enterprises (SMEs) are significant drivers of employment and innovation, often considered the backbone of many economies (Amini, Reference Amini2004; Paul, Parthasarathy, & Gupta, Reference Paul, Parthasarathy and Gupta2017; World Bank, 2015). Many of them play a critical role in sustaining the global supply chain through their exporting activities (Tan, Smith, & Saad, Reference Tan, Smith and Saad2006). Exporting is typically the initial and necessary step in the internationalization process for firms (Johanson & Vahlne, Reference Johanson and Vahlne1977, Reference Johanson and Vahlne2009), and remains the primary mode of internationalization, especially for SMEs (Paul et al., Reference Paul, Parthasarathy and Gupta2017). SMEs tend to choose exporting as a foreign market entry mode because it is less risky and costly compared to equity modes of entry, given their limited internal resources and capabilities (Lee, Jiménez, & Devinney, Reference Lee, Jiménez and Devinney2020; Nakos & Brouthers, Reference Nakos and Brouthers2002).
Against this backdrop, available studies have demonstrated that overseas network ties, as external sources of resources and information, are positively related to the international performance of exporting SMEs (e.g., Ge & Wang, Reference Ge and Wang2013; Li, Reference Li2020; Musteen, Francis, & Datta, Reference Musteen, Francis and Datta2010). Social networks serve as vital channels for accessing overseas information, knowledge, and know-how, providing a cost-effective alternative to extensive research typically undertaken by larger firms (Musteen et al., Reference Musteen, Francis and Datta2010). Moreover, unlike firms that enter markets through equity modes, which benefit from control and coordination afforded by ownership and direct investment (Meyer, Estrin, Bhaumik, & Peng, Reference Meyer, Estrin, Bhaumik and Peng2009), exporting firms rely more heavily on the benefits provided by network ties. These benefits help exporting SMEs overcome the inherent disadvantages of limited resources in the global market (Child et al., Reference Child, Narooz, Hsieh, Elbanna, Karmowska, Marinova, Puthusserry, Tsai and Zhang2022) and enhance their international performance, including levels of internationalization (e.g., Li, Reference Li2020), as well as revenue and profitability (e.g., Ge & Wang, Reference Ge and Wang2013; Musteen et al., Reference Musteen, Francis and Datta2010).
Existing studies are helpful in our understanding of the relationship between overseas network ties and the international performance of exporting SMEs. However, there remain two notable deficiencies. First, they have largely focused on the direct relationship between network ties and international performance (e.g., Ge & Wang, Reference Ge and Wang2013; Li, Reference Li2020; Musteen et al., Reference Musteen, Francis and Datta2010), while overlooking a key underlying mechanism, that is, the relationship between network ties and organizational learning (Stoian, Rialp, & Dimitratos, Reference Stoian, Rialp and Dimitratos2017). A firm’s long-term success depends on organizational learning, namely, its capability of exploitative learning to ensure its current viability and exploratory learning to ensure future viability (He & Wong, Reference He and Wong2004; Levinthal & March, Reference Levinthal and March1993; March, Reference March1991). Exploitation involves the refinement and incremental extension of current products and market capabilities, while exploration is centered on developing new products and market capabilities (March, Reference March1991; Voss & Voss, Reference Voss and Voss2013). Both are essential for sustained organizational performance (He & Wong, Reference He and Wong2004), particularly for exporting SMEs (Dasi, Iborra, & Safon, Reference Dasi, Iborra and Safon2015; Karami & Tang, Reference Karami and Tang2019), navigating the turbulent global market. The existing literature has highlighted the role of social networks in learning, emphasizing their importance as the channels for accessing and acquiring valuable information, knowledge, and other resources for both exploiting existing capabilities and exploring new possibilities (Inkpen & Tsang, Reference Inkpen and Tsang2005; Jansen, Van Den Bosch, & Volberda, Reference Jansen, Van Den Bosch and Volberda2006; Junni, Sarala, Tarba, Liu, & Cooper, Reference Junni, Sarala, Tarba, Liu and Cooper2015; Liu, Chen, & Kittilaksanawong, Reference Liu, Chen and Kittilaksanawong2013; Phelps, Heidl, & Wadhwa, Reference Phelps, Heidl and Wadhwa2012). The international business literature, particularly the social network view of internationalization, has also long emphasized that network relationships offer potential for learning and creating new knowledge, which underpins a firm’s internationalization and success in the global market (Johanson & Vahlne, Reference Johanson and Vahlne1977, Reference Johanson and Vahlne2009). However, such research in the context of firm internationalization is scarce, with rare exceptions such as Patel, Fernhaber, McDougall-Covin, and van der Have (Reference Patel, Fernhaber, McDougall-Covin and van der Have2014), who examined the balance of local and foreign networks for innovation in international markets but did not distinguish between exploitative and exploratory learning. In summary, existing studies have not explored the relationship between different network ties and both forms of organizational learning for exporting firms.
Second, earlier literature tended to view social networks as unidimensional. Increasing research has focused on the specific attributes of different types of network ties and their heterogeneous effects (e.g., Chung, Yen, & Wang, Reference Chung, Yen and Wang2020; Heirati & O’Cass, Reference Heirati and O’Cass2016; Mahmood, Zhu, & Zajac, Reference Mahmood, Zhu and Zajac2011; Yiu, Lau, & Bruton, Reference Yiu, Lau and Bruton2007). Available research reveals that exporting firms often rely on two types of overseas networks for their business, namely overseas business ties and ethnic ties (Lee et al., Reference Lee, Jiménez and Devinney2020). Overseas business ties refer to connections among parties formed in business transactions (Yiu et al., Reference Yiu, Lau and Bruton2007) – specifically for exporting SMEs, network ties with overseas buyers and marketing channels, as these firms are typically the suppliers in a global supply chain (Dyer & Nobeoka, Reference Dyer and Nobeoka2000; Liu, Deligonul, Cavusgil, & Chiou, Reference Liu, Deligonul, Cavusgil and Chiou2021; McEvily & Marcus, Reference McEvily and Marcus2005). By contrast, overseas ethnic ties are network relationships with overseas individuals, organizations, and communities based on mutual ethnic origins (Almeida, Phene, & Li, Reference Almeida, Phene and Li2015; Kim, Choi, Jean, & Cadogan, Reference Kim, Choi, Jean and Cadogan2020; Prashantham, Dhanaraj, & Kumar, Reference Prashantham, Dhanaraj and Kumar2015). While the importance of business ties has long been recognized, a growing body of literature has highlighted the role of ethnic ties in cross-border information and knowledge exchange, particularly for firms operating in international markets. Specifically, extant studies have underscored the importance of ethnic ties in facilitating location choice (Jean, Tan, & Sinkovics, Reference Jean, Tan and Sinkovics2011; Zaheer, Lamin, & Subramani, Reference Zaheer, Lamin and Subramani2009) and knowledge integration across borders (Marino, Mudambi, Perri, & Scalera, Reference Marino, Mudambi, Perri and Scalera2020), and in enhancing international performance (Jean et al., Reference Jean, Tan and Sinkovics2011; Li, Reference Li2020). Ethnic ties differ in nature and functions from other types of network ties, such as business ties and political ties, as they are based on mutual ethnic origins rather than interests. However, there has been scant research on the distinctness of ethnic and nonethnic ties, particularly in the context of exporting SMEs. One exception is the recent work by Li (Reference Li2020), who examined the different roles of ethnic and nonethnic ties in the international performance of returnee entrepreneurial firms.
Organizational learning theory suggests that exploitation and exploration are fundamentally different learning approaches, each requiring different resources (March, Reference March1991; Raisch & Birkinshaw, Reference Raisch and Birkinshaw2008). Moreover, social network theory posits that the structural and relational characteristics of an entity’s network ties can influence its access to information, knowledge, and support (Ahuja, Reference Ahuja2000; Granovetter, Reference Granovetter1973; Uzzi, Reference Uzzi1997; Zaheer & Bell, Reference Zaheer and Bell2005). This suggests that different types of network ties may be associated differently with the two distinct forms of organizational learning, a question that remains insufficiently explored, both in SME contexts and within the broader scope of organizational learning literature. Therefore, an important, yet unanswered question arises: how are overseas business and ethnic ties associated with the exploitative and exploratory learning of exporting SMEs, respectively?
This question is especially pressing for exporting SMEs from emerging economies, including China, which not only face resource constraints but also lack international experience due to their newness to the international market. Available research shows that for firms from emerging economies such as India (Almeida et al., Reference Almeida, Phene and Li2015; Prashantham et al., Reference Prashantham, Dhanaraj and Kumar2015; Zaheer et al., Reference Zaheer, Lamin and Subramani2009), Czech (Musteen et al., Reference Musteen, Francis and Datta2010), and Turkey (Kurt, Sinkovics, Sinkovics, & Yamin, Reference Kurt, Sinkovics, Sinkovics and Yamin2020), overseas network ties, in particular ethnic ties, are crucial for overcoming such limitations. This is especially true for firms from China, given the prevalence and the increasing significance of guanxi and thus their tendency to rely on guanxi to get things done (Bian, Reference Bian2018; Chen, Chen, & Huang, Reference Chen, Chen and Huang2013), particularly when navigating increasingly challenging international markets. Notably, China is also the largest exporting country worldwide, with SMEs contributing about 50% of export value (People’s Daily, 2022). SMEs in China are also surprisingly innovative, accounting for more than 70% of technological innovations (Xinhuanet, 2023). Without a further understanding of how different types of overseas ties, namely business and ethnic ties, are related to different forms of learning, that is, exploratory and exploitative learning, exporting SMEs from emerging economies like China, already constrained by limited resources and experience, may end up worse off. They may risk either misusing valuable ties or engaging in ineffective learning and losing out in the highly competitive international market.
Drawing from social network theory and organizational learning, we propose that overseas business ties are more positively related to exploitative learning than exploratory learning, whereas overseas ethnic ties are more positively related to exploratory learning than exploitative learning. This is because overseas business ties, often confined within industry networks and contractual in nature, may facilitate within-network learning and enhance current operations but limit radical innovation and entry into unrelated markets (Adler & Kwon, Reference Adler and Kwon2002; Dyer & Nobeoka, Reference Dyer and Nobeoka2000; Lew, Sinkovics, & Kuivalainen, Reference Lew, Sinkovics and Kuivalainen2013). In contrast, ethnic ties, based on shared ethnic backgrounds rather than formal contracts, typically span beyond industry boundaries. These ties are structurally open and relationally more engaged, often characterized by a high degree of trust and mutual understanding (Almeida et al., Reference Almeida, Phene and Li2015; Batjargal, Hitt, Tsui, Arregle, Webb, & Miller, Reference Batjargal, Hitt, Tsui, Arregle, Webb and Miller2013). The expansive nature of ethnic ties, particularly guanxi in the Chinese context, facilitates access to a diverse array of information and knowledge, encourages the open exchange of innovative ideas, and fosters a supportive network for exploring novel opportunities (Ramasamy, Goh, & Yeung, Reference Ramasamy, Goh and Yeung2006).
Furthermore, we examine how resource conditions, in terms of slack, moderate the relationships between two types of overseas ties and exploitative and exploratory learning. The effectiveness of leveraging external acquisition of knowledge and information requires internal resources (Raisch, Birkinshaw, Probst, & Tushman, Reference Raisch, Birkinshaw, Probst and Tushman2009). According to Penrose’s (Reference Penrose1959) theory of the growth of the firm, organizational slack – excessive resources beyond current operations – plays a critical role in supporting organizational growth and adaptation. These resources can be redeployed to promote growth in both existing and new areas (Mawdsley & Somaya, Reference Mawdsley and Somaya2021; Sakhartov & Folta, Reference Sakhartov and Folta2014) and enable firms to adjust established routines and capabilities (Sui & Baum, Reference Sui and Baum2014). Given the importance of slack resources, existing studies have explored the moderating effect of slack resources on various relationships, such as between exploratory learning and managers’ internationalization intentionality (Dasi et al., Reference Dasi, Iborra and Safon2015), creative work density and innovation performance (Chen & Huang, Reference Chen and Huang2010), Chief Executive Officer (CEO) succession and post-succession change in the degree of internationalization (Lin & Liu, Reference Lin and Liu2012), and innovation aspiration and the number of R&D alliance (Tyler & Caner, Reference Tyler and Caner2016). However, a gap remains in understanding how these resources moderate the relationship between network ties and organizational learning. Given the critical role of slack resources in directly impacting organizational learning (Voss, Sirdeshmukh, & Voss, Reference Voss, Sirdeshmukh and Voss2008) and enabling specific relationships as mentioned above, it is crucial to investigate its moderating effect on the relationships between network ties and different types of learning, particularly in the context of resource-constrained exporting SMEs (Lubatkin, Simsek, Yan, & Veiga, Reference Lubatkin, Simsek, Yan and Veiga2006).
We test our hypotheses on survey data from 172 exporting SMEs in China collected through an interview-administered questionnaire survey. We first employ ordinary least squares (OLS) to estimate the equations and find support for our hypotheses that overseas business ties are positively related to exploitative learning, whereas overseas ethnic ties are positively related to exploratory learning. Furthermore, slack resources strengthen the positive relationship between overseas ethnic ties and exploratory learning, although the moderating effect is insignificant for the relationship between business ties and exploitative learning. We also conduct a series of robustness tests, including using two-stage least squares (2SLS) with an instrumental variable, and obtain consistent results.
Our study seeks to make several contributions to the literature. First, we contribute to the literature on the role of social networks in the internationalization of SMEs by specifically examining the relationship between network ties and exploitative and exploratory learning, which in turn enhances firm performance. While existing literature has demonstrated a positive relationship between network ties and the performance of exporting SMEs (e.g., Ge & Wang, Reference Ge and Wang2013; Li, Reference Li2020; Musteen et al., Reference Musteen, Francis and Datta2010), it often overlooks the specific learning mechanism involved. Stoian et al. (Reference Stoian, Rialp and Dimitratos2017) attempted to examine the bridging role of knowledge in the direct relationship between interorganizational networks and international performance of SMEs; however, they did not focus on exporting SMEs, nor did they distinguish different types of social networks. Our research focuses on the relationships between network ties and different types of organizational learning, enriching our understanding of how social networks support the international performance of exporting SMEs. Second, we contribute to the broader discourse on social networks and organizational learning. Moving beyond the general assertion that networks facilitate learning (e.g., Almeida et al., Reference Almeida, Phene and Li2015; Atuahene-Gima & Murray, Reference Atuahene-Gima and Murray2007), our research delves into how two distinct types of overseas network ties are related to exploitative and exploratory learning differently, thus refining our understanding of the nuanced roles these ties play in organizational learning of exporting SMEs. Third, we add to the literature on the role of slack resources by revealing that the moderating role of slack resources is not uniform across different network ties and learning types in the context of SMEs; specifically, we demonstrate that slack resources may facilitate exploratory learning through ethnic ties, but this effect does not extend to exploitative learning through business ties. Fourth, we enrich the burgeoning literature on ethnic ties (e.g., Jean et al., Reference Jean, Tan and Sinkovics2011; Li, Reference Li2020) by distinguishing these from nonethnic ties, particularly business ties, in the context of exporting SMEs.
Theoretical Background and Hypotheses Development
Social Networks and Organizational Learning
Organizational learning can be both exploratory and exploitative. As defined earlier, exploratory learning emphasizes introducing radical or experimental changes with uncertain outcomes, whereas exploitative learning focuses on incremental modifications and refinement with predictable outcomes (March, Reference March1991). Both types of learning are important for firms to maintain and enhance competitiveness (He & Wong, Reference He and Wong2004; Jansen et al., Reference Jansen, Van Den Bosch and Volberda2006), particularly for SMEs to compete effectively and achieve superior performance in the international market (Cui, Walsh, & Zou, Reference Cui, Walsh and Zou2014). However, exploitation and exploration represent fundamentally different organizational learning approaches that involve distinct strategies and resources (March, Reference March1991; Raisch & Birkinshaw, Reference Raisch and Birkinshaw2008). Exploitative learning focuses on deepening existing knowledge bases and optimizing current operations, while exploratory learning necessitates a newer and more diverse array of information and knowledge that is essential for developing new products, services, and markets (Raisch & Birkinshaw, Reference Raisch and Birkinshaw2008).
Social networks are widely acknowledged as vital sources of organizational learning (Dyer & Nobeoka, Reference Dyer and Nobeoka2000; Levinthal & March, Reference Levinthal and March1993). Network ties are particularly critical for exporting SMEs, which typically suffer from internal resource constraints, including limited experience, skills, time, capital, and information for competing internationally (Lindsay, Rod, & Ashill, Reference Lindsay, Rod and Ashill2017; Manolova, Manev, & Gyoshev, Reference Manolova, Manev and Gyoshev2010; Musteen et al., Reference Musteen, Francis and Datta2010; Paul et al., Reference Paul, Parthasarathy and Gupta2017). These ties provide rich information, knowledge, and relational and other resources necessary for exploiting existing capabilities and exploring new possibilities, thereby fostering competitive advantages (Inkpen & Tsang, Reference Inkpen and Tsang2005; Liu et al., Reference Liu, Chen and Kittilaksanawong2013).
Furthermore, social network theory illuminates how network characteristics affect knowledge acquisition. Structurally, a closed network, typically characterized by high interconnectedness, may provide homogeneous and redundant information (Uzzi, Reference Uzzi1997), while an open network may connect individuals from different groups and clusters, thus giving access to diversified or new information and resources (Burt, Reference Burt1992). Relationally, contract-based and ethnicity-based network ties exhibit different characteristics (Blomqvist, Hurmelinna, & Seppänen, Reference Blomqvist, Hurmelinna and Seppänen2005; Blomqvist, Hurmelinna-Laukkanen, Nummela, & Saarenketo, Reference Blomqvist, Hurmelinna-Laukkanen, Nummela and Saarenketo2008), thus shaping access to different types of resources and opportunities (Uzzi, Reference Uzzi1997; Zaheer & Bell, Reference Zaheer and Bell2005). The structural and relational differences in social networks and the resulting differences in the types of knowledge and resources provided may influence firms’ engagement in exploitation and exploration differently. For example, Atuahene-Gima and Murray (Reference Atuahene-Gima and Murray2007) find that while intra-industry managerial ties, likely characterized as a closed network, are positively related to exploitative learning and negatively to exploratory learning, extra-industry managerial ties (a more open network) are just the opposite.
Building on this foundation, our study delves into how different types of overseas network ties relate to exploitative and exploratory learning in exporting SMEs. Specifically, we argue that overseas business ties and ethnic ties are differently associated with the two distinct types of learning in exporting SMEs. Exporting firms in China often engage with both types of ties – overseas business ties refer to relationships with overseas trade businesses, and overseas ethnic ties with ethnic Chinese communities (Zhou, Wu, & Luo, Reference Zhou, Wu and Luo2007). They are embedded in business ties, particularly supplier-buyer relationships, which facilitate knowledge transfer and sharing between firms and their customers and marketing distributors within the networks (Sharma, Kumar, Yan, Borah, & Adhikary, Reference Sharma, Kumar, Yan, Borah and Adhikary2019; Zhou, Zhang, Sheng, Xie, & Bao, Reference Zhou, Zhang, Sheng, Xie and Bao2014). They are also often connected to overseas Chinese communities such as Chinese enterprises and organizations, and overseas alumni and countrymen (Lew, Khan, Nicholson, & He, Reference Lew, Khan, Nicholson and He2016). Existing studies suggest that ethnic ties of top managers facilitate resource flows and enhance firms’ ability to learn by accessing and leveraging such resources (Lew et al., Reference Lew, Khan, Nicholson and He2016; Zaheer et al., Reference Zaheer, Lamin and Subramani2009).
While the extant literature suggests that both types of network ties are conducive to organizational learning, it remains unclear whether they relate to exploitative and exploratory learning differently. Based on social network theory, we argue that the informal, ethnic-based nature of overseas ethnic ties (Almeida et al., Reference Almeida, Phene and Li2015) can provide access to a diverse array of knowledge and information and other resources (Batjargal et al., Reference Batjargal, Hitt, Tsui, Arregle, Webb and Miller2013) conducive to exploratory learning. Conversely, contract-based business ties within supplier-buyer networks tend to focus on industry-specific, homogeneous information, aiming to enhance existing operations rather than exploring new opportunities (Choi & Kim, Reference Choi and Kim2008; Dong, Zeng, & Su, Reference Dong, Zeng and Su2019; Dyer & Nobeoka, Reference Dyer and Nobeoka2000).
Overseas Business Ties and Exploitative vs. Exploratory Learning
Exporting SMEs’ overseas business ties may involve both vertical ties with upstream and downstream partners and horizontal ties with firms in the same sub-sector, offering distinct types of knowledge and fulfilling different roles. Due to resource constraints and the need for effective interactions with business ties, SMEs are more inclined to engage with vertical rather than horizontal ties. Vertical ties provide complementary resources such as raw materials, components, knowledge, advanced technology, and market access, which are crucial for the operation of exporting SMEs in the global market (Choi & Kim, Reference Choi and Kim2008; Dong et al., Reference Dong, Zeng and Su2019;). In contrast, horizontal ties may involve direct competition and the exchange of redundant information and knowledge. For instance, knowledge sharing with horizontal ties may lead to imitation and the development of competing products (Mesquita & Lazzarini, Reference Mesquita and Lazzarini2008). This study specifically examines the vertical business ties with overseas customers and marketing channels, which are instrumental in supporting the international expansion of SMEs, particularly from emerging economies, which often face significant barriers due to their inherent disadvantages. In this context, the network relationships between small and medium-sized suppliers in emerging economies and their overseas buyers have integrated firms into the global supply chain (Zhou et al., Reference Zhou, Wu and Luo2007).
We argue that knowledge sourcing through overseas business ties is more likely to be exploitative rather than exploratory due to their structural and relational characteristics. Structurally, vertical overseas business ties form a closed network within the industry or supply chain. These tightly knit networks facilitate SMEs’ opportunities for exploitation but limit exploratory learning. These networks, often in the same or related industries, lead to significant overlap and redundancy in the information exchanged (Dong et al., Reference Dong, Zeng and Su2019; Sharma et al., Reference Sharma, Kumar, Yan, Borah and Adhikary2019). Firms within such a network can benefit from within-network learning, which is critical for effective coordination along the supply chain (Dyer & Nobeoka, Reference Dyer and Nobeoka2000). For instance, Zhang and Cantwell (Reference Zhang and Cantwell2011) suggest that vertical networks promote knowledge-exploitation type innovation due to intra-industry supply chain linkage. Similarly, Toyota’s network is engineered to exploit its production know-how and existing knowledge that resides within its suppliers (Dyer & Nobeoka, Reference Dyer and Nobeoka2000). Thus, these vertical networks are more suited for exploiting and diffusing existing knowledge rather than exploring new knowledge (Batjargal et al., Reference Batjargal, Hitt, Tsui, Arregle, Webb and Miller2013; Zhang & Cantwell, Reference Zhang and Cantwell2011).
Relationally, business ties are often contract-based, which may hinder exploratory innovation activities and are negatively related to firms’ exploratory capabilities (Lew et al., Reference Lew, Sinkovics and Kuivalainen2013). This is because firms tend to focus on fulfilling the contract, that is, conforming to existing demands and improving current processes (Chen, Liu, Wei, & Gu, Reference Chen, Liu, Wei and Gu2018; Zhang, Lyles, & Wu, Reference Zhang, Lyles and Wu2020). Such relationships make firms emphasize the exploitation of current capabilities involving the incremental development of the firm’s current products and markets (Lew et al., Reference Lew, Sinkovics and Kuivalainen2013). Therefore, often contract-based overseas business ties enhance exploitative learning but reduce the possibility of exploratory learning.
Additionally, exporting SMEs are rarely motivated to risk their business ties for exploratory learning largely due to the high value and dependence they place on these relationships (Dong et al., Reference Dong, Zeng and Su2019). Business ties, particularly those with customers and marketing channels, provide a predictable and steady stream of revenue and reduce marketing expenditures (Sirdeshmukh, Singh, & Sabol, Reference Sirdeshmukh, Singh and Sabol2002). The development and maintenance of such ties requires not only significant investment of tangible resources such as financial and human resources, but also the cultivation of reputation and trust (Adler & Kwon, Reference Adler and Kwon2002), which are costly and time-consuming to build. Given these investments, the inherent risks associated with exploratory learning, characterized by its experimental nature and uncertain outcomes, become particularly daunting for exporting SMEs. Exploratory learning may result in operational and reputational failures (Voss et al., Reference Voss, Sirdeshmukh and Voss2008), which may damage the carefully constructed business ties that SMEs rely on for their international operations and lead to potential losses in revenue, market position, and erosion of trust with important business partners. Therefore, firms are not motivated to risk their business ties to engage in exploratory learning (Adler & Kwon, Reference Adler and Kwon2002) and will exercise caution in pursuing exploratory activities when their engagement with business partners is high. Instead, they will try to protect such hard-earned, precious social capital by focusing on exploitative learning, which involves improvement in ongoing value-creating activities (Christensen & Bower, Reference Christensen and Bower1996) and current products and services (Voss et al., Reference Voss, Sirdeshmukh and Voss2008), and can be easily communicated to and accepted, or even welcomed, by existing business ties. In other words, overseas business ties are more effective in promoting exploitative learning.
In summary, the stronger an exporting SME’s overseas business ties are, the more likely the firm will engage in exploitative learning rather than exploratory learning. Thus, we posit that:
Hypothesis 1 (H1): Exporting SMEs’ overseas business ties are more positively related to their exploitative learning than their exploratory learning.
Overseas Ethnic Ties and Exploitative vs. Exploratory Learning
In contrast, overseas ethnic ties are more likely to foster exploratory rather than exploitative learning. These ties, grounded in shared ethnic origins, provide a distinct structural and relational framework compared to business ties. Structurally, ethnic ties often extend beyond within-industry business networks, connecting individuals or organizations across diverse industries, professions, and geographic regions. This broader reach facilitates access to varied and novel information, making these ties particularly conducive to exploratory learning. Relationally, these ties are characterized by trust and mutual understanding rooted in shared cultural and social backgrounds, which can accelerate the formation and deepen the engagement, unlike the formal and often contractual nature of business ties.
First, ethnic social networks may provide more diversified sources of knowledge and information conducive to exploratory learning. While overseas business ties typically connect within the global supply chain, overseas ethnic ties are more likely to form social bonds that transcend the supply chain and reach into various sectors. This structural characteristic of ethnic ties facilitates connections among members across sectors and professions, fostering a broad network of contacts. For instance, individuals within the overseas Chinese community, coming from different professions and industries, may introduce new customers and opportunities to exporting SMEs, enabling them to tap into resources beyond their current capabilities (Batjargal et al., Reference Batjargal, Hitt, Tsui, Arregle, Webb and Miller2013; Stam & Elfring, Reference Stam and Elfring2008). These resources can include private information about new sales opportunities, diverse tangible resources, and brokerage connections with distant networks such as customers, suppliers, and governments (Batjargal et al., Reference Batjargal, Hitt, Tsui, Arregle, Webb and Miller2013; Lew et al., Reference Lew, Khan, Nicholson and He2016; Stam & Elfring, Reference Stam and Elfring2008). Leveraging the information and resources sourced from ethnic ties, SMEs can significantly expand their knowledge and resource base, facilitating the acquisition of new capabilities that surpass the demands of existing customers and markets (Batjargal et al., Reference Batjargal, Hitt, Tsui, Arregle, Webb and Miller2013).
Second, while business ties may demotivate SMEs from engaging in exploratory capability building due to the difficulty of forming and the risk of losing such ties, ethnic ties offer them the opportunities, freedom, and autonomy to explore new avenues. Exporting firms may find it easier and quicker to build overseas ethnic ties thanks to common cultural norms and shared value systems (Almeida et al., Reference Almeida, Phene and Li2015). Particularly in the Chinese context, the concept of guanxi, which emphasizes personal contacts and connections over formal, contractual relationships (Jean et al., Reference Jean, Tan and Sinkovics2011; Li, Reference Li2020), plays a critical role in these interactions. At the heart of guanxi is trust and commitment to mutual benefits, which encourages a more open exchange of ideas and information. This foundation of trust also means that partners are more likely to share risky, untested ideas with the potential for innovation, without fear of opportunism or exploitation. Furthermore, guanxi involves an emotional bond, which enhances motivation to support each other’s success and share risks (Ramasamy et al., Reference Ramasamy, Goh and Yeung2006). Thus, guanxi fosters a supportive network for SMEs, which is crucial for them to access broader information and resources and explore new opportunities beyond their immediate operational needs.
The discussion above points to our second hypothesis:
Hypothesis 2 (H2): Exporting SMEs’ overseas ethnic ties are more positively related to their exploratory learning than their exploitative learning.
Slack Resources as a Moderator
Social networks provide access to external information and knowledge, yet the availability of internal resources, such as slack resources, is important for realizing their potential in enabling organizational learning (Raisch & Birkinshaw, Reference Raisch and Birkinshaw2008; Raisch et al., 2009). Slack resources, long understood as a buffer for organizational operations, are vital for facilitating learning and acting as a contingent factor (Cyert & March, Reference Cyert and March1963; Raisch & Birkinshaw, Reference Raisch and Birkinshaw2008). Defined as ‘the pool of resources in an organization that is in excess of the minimum necessary to produce a given level of organizational output’ (Nohria & Gulati, Reference Nohria and Gulati1996), slack resources have been widely recognized for their role in fostering learning and innovation (Cyert & March, Reference Cyert and March1963; Greve, Reference Greve2003; Kavusan & Frankort, Reference Kavusan and Frankort2019; Levinthal & March, Reference Levinthal and March1993; Pitelis, Reference Pitelis2007; Tyler & Caner, Reference Tyler and Caner2016). This body of literature underscores slack resources as a critical internal contingency that influences firms’ decisions related to strategic directions, such as innovation, and ultimately their performance. Firms with ample slack resources are better positioned to engage in various types of learning as they have the resources to redeploy toward specific learning processes.
Given the critical importance of slack resources, examining their moderating role in the relationships between different types of network ties and the two types of learning becomes particularly relevant for exporting SMEs. These firms often operate under resource constraints, which may affect their learning capabilities. For instance, Lubatkin et al. (Reference Lubatkin, Simsek, Yan and Veiga2006: 647) note that SMEs ‘lack the amount of slack resources and the kind of hierarchical administration systems that can help or impede larger firms in managing their contradictory knowledge processes’, suggesting that this limitation may hinder their ability to utilize external knowledge acquired through network ties to engage in exploitative and exploratory learning. While overseas business and ethnic ties provide external resources for exporting SMEs to participate in both types of learning, firms need the necessary internal resources to implement such learning actions. The low marginal cost of using excessive resources motivates and enables firms to engage in changes by redeploying the available resources (Pitelis, Reference Pitelis2007). These changes can include both exploitation, such as refining current products, solving current problems, and other activities involving predictable outcomes, and exploration, such as developing new products, experimenting, and other activities with less predictable outcomes (March, Reference March1991). Since all these activities demand a variety of internal resources (Sui & Baum, Reference Sui and Baum2014), organizational learning facilitated by network ties is inherently dependent on the availability of sufficient slack resources.
Although slack can take different forms (financial, human resources, production, etc.), our interest lies in the resources relevant to exploitative and exploratory learning. In this respect, we follow Sui and Baum (Reference Sui and Baum2014) to use labor productivity as an indicator of slack resources. High productivity implies that firms are equipped with a pool of ample resources, such as production and technological capacities, R&D facilities and personnel, and managerial resources, which are indispensable during organizational learning processes. Managers in more productive firms are often ambitious and have the entrepreneurial ambition for firm growth. Thus, they are motivated to engage in organizational learning to ensure that growth occurs (Mishina, Pollock, & Porac, Reference Mishina, Pollock and Porac2004). Firms with high productivity are more likely to possess the resources desired for exploitative and exploratory learning. These resources may strengthen the positive relationships between the two types of network ties and exploratory and exploitative learning.
First, firms with high slack have abundant production capacities (Sui & Baum, Reference Sui and Baum2014), which enables them to materialize what they learn from either business or ethnic ties by tapping into new markets or extending their current product lines (Phan & Fan, Reference Phan and Fan2007). Second, better R&D facilities and knowledge-based resources, including R&D excellence and technical skills, can enhance exploiting current products and markets and exploring new markets and products informed by both ties. Third, highly productive firms are more likely to be abundant in specific tacit human resources, particularly R&D personnel, who are especially crucial for performing exploitative and exploratory learning (Lecuona & Reitzig, Reference Lecuona and Reitzig2014). Fourth, firms with high productivity are also typically more efficient in management and thus possess excess managerial resources (Villar, Dasi, & Botella-Andreu, Reference Villar, Dasi and Botella-Andreu2018), such as managers’ time and efforts, which allow them to engage in exploitative and exploratory learning stemming from the ties.
In summary, the role of organizational slack is critical in determining how effectively exporting SMEs can utilize their network ties for organizational learning. While network ties provide essential external resources for both exploitative and exploratory learning, the capability of SMEs to harness these opportunities is contingent upon the availability of internal slack resources. Slack resources, represented by labor productivity, allow firms to redeploy resources efficiently and engage in learning activities. Therefore, for exporting SMEs from emerging economies, which typically face more challenges in the global market due to their relatively small size (Mudambi & Zahra, Reference Mudambi and Zahra2007) and resource constraints (Dimitriadis, Reference Dimitriadis2021; Sui & Baum, Reference Sui and Baum2014), slack resources strengthen the relationship between overseas ties and a specific type of learning by providing them with necessary contingent resources. Accordingly, we posit that:
Hypothesis 3 (H3): Slack resources strengthen the positive relationship between overseas business ties and exploitative learning.
Hypothesis 4 (H4): Slack resources strengthen the positive relationship between overseas ethnic ties and exploratory learning.
Methods
Data
We test our hypotheses using survey data on exporting SMEs in China. As latecomers in the global innovation arena, Chinese exporting SMEs continuously face challenges from intensified competition and shifting customer demands. They are pressured to upgrade within global value chains to enhance their exploitative and exploratory learning capabilities, thereby increasing the added value acquired from chain operations and enhancing their bargaining power against foreign customers. Therefore, China’s export sectors provide an ideal setting to test our hypotheses.
We followed the standard procedures to develop the questionnaire. We started by developing an English version of the questionnaire, which was then translated into Chinese and subsequently back-translated into English by two independent translators. After comparing and revising the versions to ensure conceptual and item equivalence, we further revised some of the questionnaire items based on a series of in-depth interviews with senior executives to enhance their completeness and clarity.
We chose to collect data through interviewer-administered questionnaire surveys at the 117th Canton Import and Export Trade Fair (referred to as Canton Fair thereafter), known as the ‘barometer’ of China’s foreign trade situation and the ‘weathervane’ for export. As the largest trade fair of its kind worldwide, it attracts the highest number of commercial visitors and participating companies. Annually, over 20,000 domestic companies participated, with over 90% being SMEsFootnote 1 from 16 sectors (see Table 1 for details). Established in 1957, the Canton Fair has become an important platform for Chinese firms to exhibit their products and technology to global buyers. Firms take the opportunity to showcase their products and negotiate with potential foreign buyers. The Canton Fair, therefore, provides an excellent opportunity to collect first-hand data from fair participants as their participating teams are mostly composed of senior executives, including sales, marketing, and foreign trade directors or even general managing directors, particularly for SMEs. Furthermore, the seniority of the interviewees, combined with responsibilities for exporting and other international activities in their firms, provides them with insight into their firms’ development, especially firms’ product designs, market development, and overseas network relationships, thereby ensuring the quality of the data collected.
Sample information and demographic profile of the trade fair
Table 1 Long description
The table compares the percentage of firms in various sectors between a sample and the overall population at a trade fair. In Phase 1, electronics and electric household appliances have a similar representation in both the sample and population. Phase 2 shows consumer goods are more prevalent in the sample than in the population. Phase 3 highlights a notable disparity in textiles and garments, with the sample having a much lower percentage than the population. Other sectors like medicine and health products are more represented in the sample than in the population. These differences suggest variations in sector representation between the sample and the overall population.
We employed stratified random sampling by sector to ensure the representativeness and reliability of the sample.Footnote 2 We employed and trained 16 research assistants, each responsible for one of the 16 sectors, to avoid cross-survey error. If the firm’s representatives refused to participate, we approached the exhibitor next to them. There was no survey bias arising from these firms as they declined before seeing the questionnaire. Our sampling resulted in 253 firms being invited to participate, of which 229 accepted, yielding a response rate of 90.5%. After dropping observations with excessive missing values and removing firms with over 1,000 employeesFootnote 3, we have a final sample of 172 firms. The average number of employees was 188, indicating that most firms were small- or medium-sized. Europe was the largest overseas market for the sample firms, with 70 firms reporting it as their primary market, followed by North America with 53 firms, Southeast Asia with 21, Africa with 12, South America with 9, and 7 firms exporting primarily to other regions. The geographic distribution of the firms is in line with the overall distribution of exporting firms, with 43 firms from Guangdong province (24.29%), followed by Zhejiang (39, 22.03%), Jiangsu (25, 14.12%), Shandong (19, 10.73%), and Fujian (15, 8.47%). Firms from these provinces are generally among the most export-oriented enterprises in China, representing regions where the export value exceeds 100 billion dollars.Footnote 4 Firms from the remaining 29 provinces accounted for 20% of our sample. We also verified the industry distribution of the sample firms. Table 1 compares the industry distribution of all firms at the Fair and those in our sample, and the results indicate that the distribution is largely consistent, confirming the representativeness of our sample.
To minimize the possibility of common method bias, we implemented several safeguards. We assured participants of the survey’s anonymity and intentionally included some interference in the items. Furthermore, we employed statistical techniques to control for common method bias, which are detailed in the next section. Additionally, we examine whether respondents who completed the questionnaire differ from those who did not, that is, item nonresponse bias, using t-test (Loosveldt, Pickery, & Billiet, Reference Loosveldt, Pickery and Billiet2002). Analysis of key firm characteristics (size, age, ownership, and export intensity) showed no significant item nonresponse bias. This finding suggests that item nonresponse is minimal in our sample, enhancing the reliability of our study results. This study has been made publicly available (see the Data Availability Statement for access details).
Variables and Measures
Our dependent variables are exploratory and exploitative learning. We adopt a composite approach to measuring exploratory and exploitative learning, each using four items drawn from Atuahene-Gima and Murray (Reference Atuahene-Gima and Murray2007). The four items for exploratory learning assess the extent to which firms focus on entering new product-market domains when collecting and acquiring information, knowledge, and ideas. Conversely, the items for exploitative learning gauge the extent to which firms focus on improving the efficiency of current products or markets. These measures, capturing the essence of exploring new possibilities and exploiting existing certainties, parallel those developed by Bierly and Daly (Reference Bierly and Daly2007) and He and Wong (Reference He and Wong2004), and are widely used in organizational learning literature (e.g., Wei, Yi, & Guo, Reference Wei, Yi and Guo2014).
We developed a measurement scale of overseas business and ethnic ties that assesses the frequency of interactions and communications between firm executives and their overseas ties on a five-point scale (1 = very infrequently, 5 = very frequently). This scale follows the empirical work on business ties (Peng & Luo, Reference Peng and Luo2000; Sheng, Zhou, & Li, Reference Sheng, Zhou and Li2011; Yiu et al., Reference Yiu, Lau and Bruton2007) and overseas ethnic ties (Lin, Zheng, Lu, Liu, & Wright, Reference Lin, Zheng, Lu, Liu and Wright2019), as well as theoretical discussions on the internationalization of SMEs (Martineau & Pastoriza, Reference Martineau and Pastoriza2016).Footnote 5 We chose to measure interaction frequency because social networks are dependent on and maintained by individual actions (Pinho & Prange, Reference Pinho and Prange2016). Our measure of network ties, which includes two items each, covers the breadth of their conceptual domains and aligns with the literature. For instance, our measure of ethnic ties includes interactions with overseas alumni, countrymen, and ethnic associations, consistent with the measure used by Lin et al. (Reference Lin, Zheng, Lu, Liu and Wright2019). Overseas Chinese, particularly the newer generation, maintain broad, multinational, and geographically dispersed networks across industries (Hou, Ye, Chen, & Fang, Reference Hou, Ye, Chen and Fang2018), which contributes to a more diversified background of ethnic ties than business ties. Specifically, overseas Chinese individuals and organizations, such as associations for overseas Chinese in various countries and cities, have played a significant role in firm internationalization, not only by helping Chinese firms go global but also by providing resources in various industries for their sustained development (Overseas Chinese Affairs Office, 2006). However, there is still a potential issue with the measures of overseas ethnic and business ties as they may influence each other or have confounding factors.Footnote 6 To prevent respondents from confusing two distinct types of relationships – overseas ethnic ties and overseas business ties – we used specific instructions to prime their focus at the start of each survey section. This approach was developed after pre-testing showed that some respondents struggled to distinguish them. For the section on overseas ethnic ties, the instructions emphasized: ‘The following questions are about your personal or your firm’s connections with individuals or organizations of shared ethnic origin overseas, such as relatives, friends, or business associates who are part of your ethnic community’. For the section on overseas business ties, the instructions stated: ‘The following questions are about your firm’s formal relationships with its key overseas business partners, such as customers, distributors, and suppliers’. This framing, combined with the item wordings, naturally directed respondents to categorize their ties according to the primary basis of the relationship without explicit prompting from the researcher. To further address this concern, we use the 2SLS with the famous hometowns of overseas Chinese as the instrumental variable to re-estimate the equations in our robustness tests and have obtained similar results, which we will discuss later.
Our moderating variable is slack resources. As slack can manifest in various ways, it is challenging to measure directly (Nohria & Gulati, Reference Nohria and Gulati1996). Recognizing this, researchers have developed various measures of slack resources, typically focusing on the conditions in which slack resources are likely to be available or their antecedents as indicators (Nohria & Gulati, Reference Nohria and Gulati1996). For example, the current ratio has been used as an indicator of financial slack, and the number of skilled labor relative to sales as an indicator of human resource slack (Paeleman, Fuss, & Vanacker, Reference Paeleman, Fuss and Vanacker2017). In this study, we follow Sui and Baum (Reference Sui and Baum2014) to use labor productivity as a proxy for organizational slack. This is in line with our research focus on factors related to exploitative and exploratory learning. Slack is generated by high performance (Greve, Reference Greve2003). Labor productivity, defined as the output per labor, is an indicator of a firm’s performance and a precursor to slack. As discussed earlier, higher productivity implies the availability of superior technological and managerial resources, indicating surplus resources that exceed immediate operational demands. This additional capacity, or slack, can be allocated toward learning activities, specifically enhancing the relationship between overseas ties and learning.
We control for an array of variables including firm age, firm size (the total number of a firm’s employees), firm performance (subjective evaluation on return on assets, ROA, on a scale of 1–5), export intensity (the ratio of a firm’s total exports relative to its total sales), ownership (captured by two dummy variables- ‘state-owned’ and ‘foreign’ enterprisesFootnote 7), strategic flexibility (measured based on six items, Sanchez, Reference Sanchez1995), and market and technological turbulence (measured based on six items, Jaworski & Kohli, Reference Jaworski and Kohli1993). We also include province and industry fixed effects. We rescale the variables of firm age, firm size, exporting intensity, and slack resources by taking the natural logarithm to account for the skewed distribution of the variables.
All measurement items for the latent variables are rated on a five-point Likert scale. Table 2 presents the results of reliability and convergent validity tests. As shown in Table 2, the factor loadings for all items of exploitative learning, business ties, and ethnic ties are above the recommended cutoff point of 0.70 (Hair, Black, Babin, & Anderson, Reference Hair, Black, Babin and Anderson2018).Footnote 8 The factor loadings for three items of exploratory learning are above 0.80 whereas the fourth item loaded at 0.50, which is below the cutoff point of 0.70. We retain this item for theoretical reasons to capture the full dimensions of exploratory learning but obtain substantially the same results when excluding this item.Footnote 9 The factor loadings for one item of strategic flexibility and one item of market and technological turbulence are below 0.70 but they are close (0.67). Again, we retain these items for theoretical reasons. The Cronbach’s 
$\alpha $ and composite reliability of all six constructs are also above the recommended cutoff point of 0.70 (Hair et al., Reference Hair, Black, Babin and Anderson2018). We also computed the average variance extracted (AVE), and it exceeds the usually recommended cutoff value of 0.50 (Fornell & Larcker, Reference Fornell and Larcker1981). Thus, we can infer a satisfactory level of reliability and convergent validity.
Reliability and validity
Table 2 Long description
The table measures reliability and validity across various learning and strategic categories, including exploitative and exploratory learning, overseas ethnic ties, strategic flexibility, and market and technological turbulence. Exploitative learning exhibits high composite reliability at 0.90 and a strong average variance extracted at 0.69, suggesting consistent measurement. In contrast, exploratory learning has a lower average variance extracted of 0.61, indicating less consistency in measurement. Overseas ethnic ties show high reliability with a Cronbach's alpha of 0.86 and an average variance extracted of 0.80, suggesting strong internal consistency. Strategic flexibility and market turbulence both have composite reliability values of 0.90, indicating robust measurement across these categories. Factor loadings vary, with exploratory learning showing a wide range from 0.50 to 0.87, highlighting variability in how well individual items measure the underlying construct. Overall, the table provides insights into the reliability and validity of different strategic and learning approaches, with exploitative learning demonstrating the most consistent measurement.
For discriminant validity, as shown in Tables 2 and 3, the square root of the AVE is above the highest correlation coefficient of the construct with other variables in the analysis, meeting the criterion for discriminant validity defined by Fornell and Larcker (Reference Fornell and Larcker1981). Furthermore, we adopt the newly developed Heterotrait–Monotrait ratio (HTMT) (Henseler, Ringle, & Sarstedt, Reference Henseler, Ringle and Sarstedt2015) to evaluate discriminant validity, with results reported in Table 4. If the HTMT of all latent variables is below the predefined threshold of 0.85 (Kline, Reference Kline2011), then discriminant validity is inferred. Results reported in Table 4 confirm that the discriminant validity of the measures is satisfactory, with all ratios below 0.85.
Descriptive statistics
Table 3 Long description
The table presents descriptive statistics for various firm characteristics, including age, size, performance, and strategic factors. Firm size has the highest mean value of 4.79, indicating larger firms in the sample, while SOE and Foreign firm have the lowest mean of 0.03, suggesting few firms are state-owned or foreign. Export intensity shows a significant negative correlation with firm size, suggesting larger firms may have lower export intensity. Overseas business ties are positively correlated with slack resources, indicating firms with more business ties may have more resources available. The data highlights significant relationships at the 10% level, emphasizing the importance of strategic flexibility and market turbulence in firm performance. Interpretations should consider the sample size of 172 and the statistical significance of coefficients.
Notes: N = 172. Coefficient first, then the p-value below it. Bold coefficients indicate statistical significance at the 10% level.
Results of discriminant validity (HTMT)
Table 4 Long description
The table measures discriminant validity among various variables, focusing on the relationships between exploitative learning, exploratory learning, overseas ethnic ties, overseas business ties, and strategic flexibility. Exploratory learning shows a moderate correlation with exploitative learning at 0.508. Overseas ethnic ties have a low correlation with exploitative learning and exploratory learning, with values of 0.057 and 0.344 respectively. Overseas business ties correlate moderately with overseas ethnic ties at 0.373. Strategic flexibility shows low to moderate correlations with other variables, with the highest being 0.406 with market and technological turbulence. These correlations suggest varying degrees of association among the variables, with some showing stronger relationships than others.
Table 3 presents the descriptive statistics for all variables. Consistent with the notion that exploration involves more risks and uncertainties than exploitation, firms tend to be more active in exploitation (mean = 3.99) than exploration (mean = 3.35). Also, exporting SMEs established closer relationships with business partners than with ethnic communities. Overseas business ties correlate positively to exploitative learning (
$r = 0.21,{\text{ }}\,p \lt 0.01$) while overseas ethnic ties correlate positively to exploratory learning (
$r = 0.30,{\text{ }}\,p \lt 0.01$), providing initial evidence supporting H1 and H2.
Common Method Bias
Since data on the latent variables were collected through a single respondent questionnaire survey, we conducted Harman’s one-factor analysis to check the common method bias. If the common method is a significant concern, a single factor will dominate the results of factor analysis (Podsakoff & Organ, Reference Podsakoff and Organ1986). Our principal component analysis extracted four distinct factors, with the largest factor accounting for 32.9% of the variance. This is below the recommended threshold of 50% for indicating significant common method bias (Podsakoff & Organ, Reference Podsakoff and Organ1986). Additionally, we incorporate various objective variables into our analysis, such as firm age, firm size, and export share, to further minimize the effect of common method bias. Thus, we are confident that common method bias does not compromise the validity of our findings in this study.
Results
We first test our hypotheses using OLS, as it is robust and easy to interpret in examining linear relationships (Hair et al., Reference Hair, Black, Babin and Anderson2018). While there is no consensus on the optimum sample size for OLS, our sample size of 172 is considered adequate for estimation, as suggested by Green (Reference Green1991), and is comparable to those typically used in studies on network ties, which are usually survey-based (e.g., Acosta, Crespo, & Agudo, Reference Acosta, Crespo and Agudo2018; Liu et al., Reference Liu, Chen and Kittilaksanawong2013; Musteen et al., Reference Musteen, Francis and Datta2010; Pinho & Prange, Reference Pinho and Prange2016; Tang, Reference Tang2011; Yiu et al., Reference Yiu, Lau and Bruton2007; Zhang, Ma, Wang, Li, & Huo, Reference Zhang, Ma, Wang, Li and Huo2016). Our examination of the variance inflation factors, which range from 1.31 to 1.91, suggests that there are no serious multicollinearity problems (Neter, Wasserman, & Kutner, Reference Neter, Wasserman and Kutner1985). We also conduct the Breusch–Pagan test to address the potential problem of heteroskedasticity; the results suggest that heteroskedasticity is not present in our models, with the chi-square values of 0.34 (
$p = {\text{ }}$0.561) and 0.08 (
$p = {\text{ }}$0.782) for Model 4 and Model 8, respectively. To further address the inefficiency concerns associated with heteroskedasticity, we apply robust standard errors to correct for this in our estimations. We follow a stepwise approach to include the variables in the equations, and the results suggest that 
${R^2}$ increasements are statistically significant, indicating that the models are more significant when including the hypothesized variables.
The results of OLS are reported in Table 5. Before turning to the results of hypothesis testing, it is worth noting that our results suggest that strategic flexibility and innovative environment are both significantly related to organizational learning, which is consistent with the literature (Jaworski & Kohli, Reference Jaworski and Kohli1993; Sanchez, Reference Sanchez1995; Zhou & Wu, Reference Zhou and Wu2010). It is also important to note that the coefficients of some organizational characteristics, such as size and age, are insignificant, which could be attributed to the controversial effects of these variables, as discussed in previous studies (e.g., Jansen et al., Reference Jansen, Van Den Bosch and Volberda2006; Zhou & Wu, Reference Zhou and Wu2010).
Overseas ties, slack resources, and exploitative and exploratory learning
Table 5 Long description
The table examines the impact of various factors on exploitative and exploratory learning across different models. Market and technological turbulence consistently has a significant positive effect on both types of learning, with coefficients above 0.34 and p-values less than 0.01 in all models. Strategic flexibility also positively influences learning, with significant coefficients in most models. Firm performance shows a positive effect on exploitative learning, particularly in Model 3, while overseas ethnic ties and business ties have mixed effects depending on the learning type. Slack resources and their interaction with overseas ties show varied impacts, indicating complex relationships. The table includes fixed effects for location and industry, and the number of observations varies between models.
*** Notes: p < 0.01, **p < 0.05, *p < 0.1. Coefficient first, standard error next to it in parentheses, then the p-value below it.
Models 1 and 5 include the control variables only. Model 2 introduces the variables of the two types of overseas ties. We find that overseas business ties are positively related to exploitative learning (
$\beta = 0.134,{\text{ }}p = 0.017$). Surprisingly, ethnic ties are negatively related to exploitative learning (
$\beta = - 0.070,{\text{ }}p = 0.092$) although the economic size of the coefficient is very small, and it is only marginally significant. The results of Model 6 suggest that there is a positive relationship between overseas ethnic ties and exploratory learning (
$\beta = 0.121,{\text{ }}p = 0.014$) but no evidence of the relationship between business ties and exploratory learning (
$\beta = - 0.087,{\text{ }}p = 0.187$).
H1 posited that overseas business ties would be more positively related to exploitative learning than to exploratory learning. Beyond the direct effects reported above, which are consistent with this view, we conducted significance tests for the difference between coefficients. Given that our model involves two different dependent variables (exploitative learning and exploratory learning), we used a seemingly unrelated regression model (Zellner, Reference Zellner1962) to determine if the coefficient estimates for the same independent variable across the two regression equations were statistically different from one another (see Table 6). The results show that the effect of overseas business ties on exploitative learning (β = 0.134, p = 0.007) was significantly greater than its effect on exploratory learning (β = −0.087, p = −0.203), and this difference was statistically significant (Δβ = 0.221, χ2 = 10.71, p < 0.01). Thus, H1 was strongly supported. Similarly, H2 proposed that overseas ethnic ties would be more positively related to exploratory learning than to exploitative learning. The difference test (Table 6) confirmed that the effect of overseas ethnic ties on exploratory learning (β = 0.121, p = 0.006) was significantly greater than its negative effect on exploitative learning (β = −0.070, p = 0.058), and this difference was significant (Δβ = 0.191, χ2 = 14.45, p < 0.001). Therefore, H2 was also supported.
Tests of coefficient differences for overseas ties on organizational learning: Seemingly unrelated regression (SUR) results
Table 6 Long description
The table presents results from a seemingly unrelated regression analysis examining the impact of various factors on exploitative and exploratory organizational learning. Market and technological turbulence has a strong positive effect on both learning types, with high statistical significance. Strategic flexibility also positively influences both learning types, though to a lesser extent. Overseas ethnic ties positively affect exploratory learning, while overseas business ties enhance exploitative learning. Firm performance shows a marginal positive effect on exploitative learning. Other factors like firm age, size, and ownership types have less pronounced effects. The analysis includes location and industry fixed effects, and the number of observations is consistent across models.
*** Notes: p < 0.01, **p < 0.05, *p < 0.1. Coefficient first, standard error next to it in parentheses, then the p-value below it.
We then examine the moderating effect of slack resources. In Models 3 and 7 of Table 5, we add the variable of slack resources to the equations and find that it is significantly positively related to both exploitative and exploratory learning (
$\beta = 0.223,{\text{ }}p = 0.065;{\text{ }}\beta = 0.293,{\text{ }}p = 0.056$). This demonstrates the role of organizational slack in facilitating both types of learning, which is consistent with our previous discussion and affirms its role as a moderator.
In Model 4 of Table 5, the coefficient of the interaction term of business ties and slack is negative, although insignificant (
$\beta = - 0.122,{\text{ }}p = 0.184$), suggesting that H3 is not supported, and the coefficient of the interaction term of ethnic ties and slack is positive and insignificant (
$\beta = 0.103,{\text{ }}p = 0.243$). In Model 8, the coefficient of the interaction term of ethnic ties and slack is positive and significant (
$\beta = 0.216,{\text{ }}p = 0.051$), indicating that slack positively moderates the relationship between overseas ethnic ties and exploratory learning. In contrast, the coefficient of the interaction term of business ties and slack is positive but insignificant (
$\beta = 0.030,{\text{ }}p = 0.789$). Thus, H4 is supported. The result of the moderating effect of slack on the relationship between ethnic ties and exploratory learning is graphically represented in Figure 1. The graph shows that the relationship between overseas ethnic ties and exploratory learning strengthens with more abundant slack resources, as indicated by a steeper slope when the level of organizational slack is higher. A further analysis of the effect size is shown in Figure 2, which plots the marginal effect of overseas ethnic ties on exploratory learning along with the 95% confidence bands over the relevant values of the moderating variable slack. An inspection of Figure 2 reveals that overseas ethnic ties have a statistically significant positive effect on exploratory learning over most of the sample values of slack (lg_slack is from 0.01 to 3.58, see Table 3). For example, when lg_slack is equal to approximately 3.5, that is, slack being 33.115, the marginal effect of overseas ethnic ties on exploratory learning is equal to approximately 0.77 percentage points: using Table 5 values, this is calculated from Model 8 as 0.77 = 0.010 +(0.216 × 3.5). Since the confidence interval bands cross 0 for values of lg_slack left of A (about 0.45), the interaction effect is significant in the case of values of lg_slack right of A (slack >1.57). In a word, H4 is at least partially supported again.
Moderating effect of slack resources on the relationship between overseas ethnic ties and exploratory learning (95% confidence intervals)
Figure 1 Long description
A line graph showing the relationship between exploratory learning and ethnic ties. The x-axis is labeled 'Ethnic ties' ranging from 1.5 to 4.5 and the y-axis is labeled 'Exploratory learning' ranging from 2.5 to 4.5. Two curves are plotted: one for lower slack resources and one for higher slack resources, both marked with triangles. The curve for higher slack resources shows a steeper increase compared to the curve for lower slack resources. Shaded areas around the curves represent confidence intervals.
Marginal effect of overseas ethnic ties on exploratory learning based upon Model 8 in Table 5
Figure 2 Long description
A graph showing the marginal effect of overseas ethnic ties on exploratory learning plotted against lgslack. The x-axis is labeled 'lgslack' ranging from 0 to 4. The y-axis is labeled 'Marginal effect of overseas ethnic ties on exploratory learning' ranging from negative 1.5 to 1.5. A red line with data points shows a positive trend. Two green lines represent confidence intervals. A vertical line labeled 'A' is drawn at approximately 0.45 on the x-axis.
We conduct a number of robustness checks. First, as discussed earlier, overseas business and ethnic ties may influence each other and could have confounding factors, both of which may bias the results. To address this issue, we employ 2SLS, using the dummy variable for famous ancestral hometowns of overseas Chinese (‘qiaoxiang’) as the instrumental variable for overseas ethnic ties, and re-estimate the equations.Footnote 10 The choice of qiaoxiang as an instrumental variable is based on its strong historical and sociocultural relevance to the formation of overseas ethnic ties. Residents in qiaoxiang are more likely to have overseas ethnic ties through kinship and cultural ties, overseas associations of fellow countrymen, as well as government agencies responsible for maintaining and promoting such ties (Cheng & Ngok, Reference Cheng, Ngok, Douw, Huang and Godley2013). This makes qiaoxiang a potent predictor of overseas ethnic ties. Moreover, qiaoxiang satisfies the exogeneity requirement as its relationship with the two types of learning is only through the channel of the instrumented variable, that is, overseas ethnic ties. On the other hand, qiaoxiang is not necessarily related to overseas business ties. For instance, while firms in coastal cities are typically more engaged in international trade and possess broader business ties, not all coastal cities are qiaoxiang.
Our empirical tests confirm the appropriateness of qiaoxiang as the instrumental variable. The results are reported in Table 7. In the first stage of our 2SLS analysis, we find that qiaoxiang is significantly related to ethnic ties (F-statistics being 15.26). We also test its relationships with overseas business ties and the two types of learning, and do not find any significant relationships. Upon estimation of 2SLS models, the core variable of overseas ethnic ties remains statistically significant in Model 3, where it shows a positive effect (β = 0.470, p = 0.059). This confirms the robustness of our primary results.
Results of two-stage least squares (2SLS) with famous hometowns of overseas Chinese (‘qiaoxiang’) as instrumental variables
Table 7 Long description
The table presents results from a two-stage least squares analysis, focusing on the impact of various firm characteristics on exploitative and exploratory learning. Market and technological turbulence shows a strong positive effect on both learning types, with high statistical significance. Firm size negatively influences exploratory learning, but this effect is less statistically significant. Estimated overseas ethnic ties have a positive impact on exploitative learning and a less significant positive effect on exploratory learning. The hometown of overseas Chinese is a significant factor in the first stage, indicating its importance as an instrumental variable. Observations and R-squared values are consistent across models, suggesting stable results. Interpretations should consider the varying significance levels of the coefficients.
Notes: The Model 1 presents the results from the first stage of 2SLS, the subsequent columns report the results from the second stage of 2SLS, using the estimated overseas ethnic ties from the first stage as the independent variable, respectively;
*** p < 0.01, **p < 0.05, *p < 0.1. Coefficient first, standard error next to it in parentheses, then the p-value below it.
Second, we use the Heckman two-stage estimation techniques to address potential selection bias. Based on the assumption that respondents with excessive missing values are more similar to nonrespondents than to respondents with completed questionnaires (Loosveldt et al., Reference Loosveldt, Pickery and Billiet2002), and in order to test selection or partial nonresponses bias, we perform a null model of Probit on the full sample in the first stage and included the inverse Mills ratio obtained from the first stage in the second stage. The results, reported in Table 8, are similar to those reported in Table 5, suggesting no selection bias in the sample. Third, to rule out the possibility that the results are driven by control variables, we repeat the regressions by omitting different control variables in separate iterations and obtain largely consistent results. Fourth, we are able to use an alternative measure of firm performance in our regressions, and again, the results are similar.Footnote 11
Results of the second stage of Heckman model
Table 8 Long description
The table presents the results of the second stage of the Heckman model, analyzing the effects of various firm characteristics on exploitative and exploratory learning. Market and technological turbulence has a consistently strong positive effect across all models, indicating its significant role in both learning types. Strategic flexibility also shows positive impacts, particularly in models 1, 2, 4, 5, and 6, suggesting its importance in adaptive strategies. Firm performance is notably significant in models 3 and 4 for exploitative learning, while overseas ethnic ties and business ties have mixed effects across models. The presence of location and industry fixed effects is consistent, and the R-squared values indicate moderate explanatory power. Interpretations should consider the varying significance levels and potential interactions between variables.
*** Notes: p < 0.01, **p < 0.05, *p < 0.1. Coefficient first, standard error next to it in parentheses, then the p-value below it.
Discussion
This article aims to examine the relationship between overseas business and ethnic ties and exploitative and exploratory learning, as well as the moderating effect of slack resources on these relationships, within the context of exporting SMEs from China. While it is well-documented that network ties can enhance organizational learning, the specific relationship between different types of external network ties and different types of organizational learning (exploitation vs. exploration), along with the boundary conditions of internal resources, remains less understood. Our study reveals that overseas business ties are positively related to exploitative learning, whereas overseas ethnic ties are positively related to exploratory learning. Our results contrast with Li’s work (Reference Li2020) on the internationalization of returnee entrepreneurial firms, which suggests that ethnic ties often provide homogeneous information and business ties deliver diversified information. We observe that ethnic ties are more likely to provide access to diversified information and resources, thereby promoting exploratory learning, whereas business ties tend to supply information and knowledge focused on refining existing businesses and operations, aligning more with exploitative learning. This divergence may arise from the different measures and operational contexts of returnee entrepreneurs versus exporting SMEs. Specifically, Li’s (Reference Li2020) measure of overseas ethnic and nonethnic ties includes social interactions with mentors, colleagues, customers, and business associates, as well as professional interactions at conferences, which may limit the scope of their ethnic ties. This is understandable, as returnee entrepreneurs’ overseas network ties are typically established during their time of studying or working abroad and may confine their ethnic connections to familiar or similar ethnic communities. In contrast, managers of exporting SMEs may prioritize access to diverse networks and information channels by connecting with a broader spectrum of overseas ethnic communities to navigate complex international markets and foster innovation.
Moreover, our results confirm the hypothesis that organizational slack strengthens the positive relationship between overseas ethnic ties and exploratory learning. However, the moderating effect of slack resources on the relationship between overseas business ties and exploitative learning is found to be insignificant. This may be due to two reasons. First, business networks tend to be more closed and homogeneous (Dyer & Nobeoka, Reference Dyer and Nobeoka2000; Sharma et al., Reference Sharma, Kumar, Yan, Borah and Adhikary2019) and contract-based, where resources and processes are often predefined by contracts or network norms, leaving less scope for slack to enhance refinement beyond established mechanisms (Levinthal & March, Reference Levinthal and March1993; March, Reference March1991). Consequently, there is limited scope to strengthen the relationship between business ties and exploitative learning, even with the availability of slack resources. This finding is in line with the existing literature suggesting that such business networks have more confined boundaries (Burt & Burzynska, Reference Burt and Burzynska2017). Second, our measurement of slack in terms of labor productivity may not capture the type of slack that is most relevant to or needed for business-tie-driven exploitative learning. Research shows that different types of slack resources operate through contingent pathways (Freitas, Serra, Guerrazzi, Scazziota, & Scafuto, Reference Freitas, Serra, Guerrazzi, Scazziota and Scafuto2025). Nevertheless, we can conclude that resources play a more important role in supporting the relationship between ethnic ties and exploration. With available slack resources, firms are more inclined to redeploy them toward exploration driven by ethnic ties rather than toward exploitation driven by business ties.
Our results also suggest that ethnic ties are negatively related to exploitative learning, and the results are consistent across all models (Models 2–4 in Table 5). While the economic size of the coefficients is small, this trend may indicate the diverting effect of ethnic ties on firms’ focus. Specifically, the presence of intensified ethnic ties may steer firms away from concentrating on optimizing existing operations toward exploring new products and markets, thereby reducing the level of exploitative learning within the firms.
Theoretical Contributions
We make contributions to several strands of the literature. First, we advance our understanding of the role of social networks in the organizational learning of exporting SMEs, particularly by distinguishing different types of network ties and distinct types of organizational learning. While exploitation and exploration are crucial for firm survival and performance, achieving both can be challenging, especially for SMEs constrained by limited resources. SMEs use network ties to access information, knowledge, and resources needed for exploiting existing competencies and exploring new capabilities (Heirati & O’Cass, Reference Heirati and O’Cass2016; Sharma et al., Reference Sharma, Kumar, Yan, Borah and Adhikary2019). Previous studies have primarily focused on the direct relationship between social networks and SME internationalization (Ge & Wang, Reference Ge and Wang2013; Musteen et al., Reference Musteen, Francis and Datta2010), leaving a mechanistic gap regarding why network ties are beneficial for SME internationalization. Our study addresses the role of social networks in organizational learning, which can be an important source of competitive advantage for firms seeking internationalization. In doing so, we complement and enrich the existing literature on social networks and SME internationalization (e.g., Karami & Tang, Reference Karami and Tang2019; Lindsay et al., Reference Lindsay, Rod and Ashill2017; Musteen et al., Reference Musteen, Francis and Datta2010), thereby enhancing our understanding of how social networks may facilitate the international performance of exporting SMEs through learning.
Moreover, our study provides evidence that certain network ties may facilitate one side of organizational learning over another and identifies exactly which type of network ties is related to which type of organizational learning. Given the resource constraints of SMEs, it is particularly important to understand the role of different ties and their relationships with different types of learning in order to allocate limited resources more effectively to achieve organizational goals. Thus, our research sheds light on new avenues that may help SMEs shape their organizational learning, and potentially their internationalization and performance, through strategic planning of their networking activities.
Second, alongside the first contribution, we also contribute to the general literature on social networks and organizational learning by distinguishing the effects of two different types of overseas ties. According to social network theory, network ties can be categorized in a variety of ways. Recent literature has examined the differences in the nature and impacts between horizontal and vertical business group networks (Zhang & Cantwell, Reference Zhang and Cantwell2011), business and political or institutional ties (Heirati & O’Cass, Reference Heirati and O’Cass2016; Li et al., Reference Li, Chen, Liu and Peng2014; Peng & Luo, Reference Peng and Luo2000; Yiu et al., Reference Yiu, Lau and Bruton2007; Zhang et al., Reference Zhang, Ma, Wang, Li and Huo2016), inter-personal and inter-organizational ties (Eberhard & Craig, Reference Eberhard and Craig2013; Ge & Wang, Reference Ge and Wang2013; Mahmood et al., Reference Mahmood, Zhu and Zajac2011; Manolova et al., Reference Manolova, Manev and Gyoshev2010). Expanding this line of research, we differentiate two types of overseas ties, that is, business ties and ethnic ties, and examine how the structural and relational characteristics of these network ties determine their relationships with different types of organizational learning. Structurally, overseas business ties are more confined to closed networks within a specific industry or supply chain, thus more related to exploitative learning, while overseas ethnic ties, which connect members from diversified industries, professions, and geographical areas, are more related to exploratory learning. Relationally, business ties are often contractual and focused on immediate operational needs, which can constrain risk-taking and innovation, leading firms to engage predominantly in exploitative learning. In contrast, ethnic ties, often characterized by trust and shared cultural values, provide access to a more diversified range of information and resources, which can motivate firms to pursue exploratory learning. Our study thus complements and enriches the literature on social networks and organizational learning by showing how different types of network ties in a global marketplace are related to distinct types of learning. Our results highlight the need for further attention to the differentiated values of different types of social networks in learning processes in future research.
Third, we contribute to the literature on the contingency role of slack resources in organizational learning. While the enabling role of organizational slack in organizational learning is well-recognized, its significance becomes critical in the context of exporting SMEs, given their inherent resource constraints. Moreover, we examine whether slack resources moderate the relationship between different types of network ties and the two types of learning uniformly. Our results offer new insights into the nuanced nature of this moderating effect by revealing that it varies across different network types and learning types. Specifically, we show that the availability of slack is an important condition for ethnic ties and exploratory learning but not for business ties and exploitative learning. As discussed above, this distinction can be attributed to the limited scope of business ties, as suggested by Burt and Burzynska (Reference Burt and Burzynska2017). In contrast, the contents of ethnic ties are more diversified and thus can have a stronger positive relationship with exploratory learning when firms have the necessary resources to redeploy. This result confirms that organizational slack provides conditions and capabilities for SMEs to carry out exploratory learning more easily when acquiring knowledge and information from ethnic ties.
Fourth, our study contributes to the growing literature on ethnic ties. While the role of business ties has long been recognized, the importance of ethnic ties has only begun to attract attention more recently. Existing literature on ethnic ties has largely focused on how they facilitate knowledge flow across borders (Choudhury & Kim, Reference Choudhury and Kim2019; Marino, Mudambi, Perri, & Scalera, Reference Marino, Mudambi, Perri and Scalera2020) and thus influence firms’ strategic decisions in the global market (Lew et al., Reference Lew, Sinkovics and Kuivalainen2013; Li, Reference Li2020). Li (Reference Li2020) suggests that ethnic ties often provide homogeneous information. By contrast, we argue and provide evidence that ethnic ties promote exploratory learning for exporting SMEs, very likely by providing access to diversified information and resources. Our study, therefore, enriches the literature on ethnic ties, specifically on the relationship between ethnic ties and different types of learning, and provides a new arena in this line of research. Moreover, while our research is set in China, the findings may be applicable to other emerging economies where social networks, and particularly ethnic ties, are playing a significant role in firm internationalization. For instance, while guanxi is emphasized in the Chinese context, the importance of wasta (Arab) (Berger, Silbiger, Herstein, & Barnes, Reference Berger, Silbiger, Herstein and Barnes2015; Khakhar & Rammal, Reference Khakhar and Rammal2013), blat (Russia) (Chimenson, Tung, Panibratov, & Fang, Reference Chimenson, Tung, Panibratov and Fang2022; Smirnova, Reference Smirnova2020), yongo (South Korea) (Horak & Klein, Reference Horak and Klein2016), and jeitinho (Brazil) (Amado & Brasil, Reference Amado and Brasil1991; Braga, Oliva, Kubo, McKenna, Richardson, & Wales, Reference Braga, Oliva, Kubo, McKenna, Richardson and Wales2021) has also been examined in the literature. However, caution is necessary to generalize the results due to potential cultural differences, as discussed later.
Managerial Implications
Our findings offer several key managerial implications. First, exporting SMEs with limited resources should prioritize networking based on their primary learning goals. Rather than intensifying both business and ethnic ties simultaneously, managers should focus on the tie that best supports their immediate strategic need. Although building and maintaining these ties requires investment, they serve as crucial external resources and are generally less costly and more feasible compared with other competency-building strategies such as vertical integration (Dong et al., Reference Dong, Zeng and Su2019). For exploitation and incremental improvement, such as refining existing products or securing stable revenue streams in current international markets, exporting SMEs should prioritize deepening key overseas business ties (e.g., with major customers and distributors), which are readily leveraged even with minimal slack due to their structured nature. Conversely, for exploration and new opportunities, such as entering new geographic markets or developing innovative products, a conscious investment in nurturing overseas ethnic ties is crucial, as they provide the diverse information and brokerage connections vital for such discovery.
Second, managers can operationalize this strategic prioritization through low-cost, high-impact actions tailored to each tie. To leverage ethnic ties for exploration without significant expenditure, SMEs can actively participate in online ethnic community forums, alumni groups, or industry-specific social media groups (e.g., WeChat, LinkedIn) to tap into novel ideas. Attending cultural or business association events, even virtually, serves as another low-cost method to build connections and access diverse information. Conversely, to deepen business ties for exploitation, managers should focus on relational investments that build trust and reliability, such as providing exceptional customer service and proactively sharing market insights with key partners. This approach enhances the value of existing ties without necessitating large financial outlays, making the strategy both effective and feasible for resource-constrained SMEs.
Third, our findings confirm that slack resources are a critical catalyst for exploratory learning via ethnic ties, significantly amplifying their benefits. Managers should therefore create and allocate slack strategically. SMEs can achieve this by excelling in exploitative learning through business ties, which involve deepening relationships with key partners to secure stable revenue and operational efficiencies. The predictable resources generated from this exploitation should then be deliberately channeled to fund exploratory activities informed by ethnic ties. This creates a self-reinforcing cycle: using the reliable outputs of business networks to finance the transformative potential of ethnic networks, thereby directly leveraging the catalytic role of slack that our study identifies.
Limitations and Future Research Directions
This study is subject to several limitations. First, while our survey data enables the collection of specific data, such as ethnic ties, which would not be available from archival data, the cross-sectional nature of our data constrains our ability to establish causality. The absence of temporal variation in our single-period dataset inherently restricts our ability to control for time-specific confounders (e.g., annual policy shifts or macroeconomic fluctuations) through year fixed effects. Similar to other studies using cross-sectional data on exploitation and exploration (e.g., Cao, Gedajlovic, & Zhang, Reference Cao, Gedajlovic and Zhang2009), we assume that firms’ exploitation and exploration are stable over time. Future research should aim to construct longitudinal datasets with richer data that can examine the dynamic relationship between overseas ties, organizational learning, and long-term firm performance over time. Second, our study is contextualized in China, a typical emerging economy. However, it is important to recognize the substantial heterogeneities among emerging economies. For example, while existing literature emphasizes the significance of network ties – known as wasta, blat, yongo, and jentinho in different emerging economy settings – for the success of firms from these countries (Horak, Afiouni, Bian, Ledeneva, & Fey, Reference Horak, Afiouni, Bian, Ledeneva and Fey2020), particularly for firm internationalization, it is also recognized that cultural factors play a significant role in how social networks may influence organizational strategies and performance. Therefore, caution should be exercised when generalizing our results to other emerging economies. Future research based on a cross-country sample may provide deeper insights by examining these relationships within diverse cultural settings. Third, in this study, we examine the moderating role of slack resources using labor productivity as an indicator. While productivity is an appropriate indicator of slack (Sui & Baum, Reference Sui and Baum2014), particularly in our context of exporting SMEs in China with exploitative and exploratory as dependent variables, future research employing other measures of slack and distinguishing different types of slack may yield additional insights. Fourth, our measure of overseas ethnic ties and business ties is based on two items each. Future studies that incorporate more items would be able to better capture the dimensions of these network ties.
Data availability statement
The datasets generated during the current study are available in the Open Science Framework repository: Yuan, J. (2025, December 23). Overseas ties and exporting SME learning. Retrieved from osf.io/m7vs9
Acknowledgements
We gratefully acknowledge the financial support provided by the National Natural Science Foundation of China (Grant No. 72474055), Project of Guangdong Philosophy and Social Sciences (Grant No. GD23XGL064), 2024 Annual Special Commission Project of Guangdong Philosophy and Social Sciences Innovation Project (Grant No. GD24WTCXGC12), and Guangdong Province Key Laboratory of Humanities and Social Sciences for General Universities ‘Global Supply Chain Big Data Laboratory’ (Grant No. 2022WSYS005). We are grateful to the anonymous reviewers for their careful reading and valuable feedback. Their insightful critiques and detailed suggestions have greatly enhanced the rigor and clarity of this paper.
Xiaoying Li was a scholar of international business who previously worked at the University of Essex. Her research focused on international business and related areas, contributing to the understanding of global business dynamics. She was widely respected for her academic work and dedication to advancing knowledge in her field.
Jing Yuan (200711651@oamail.gdufs.edu.cn) is an associate professor at Guangdong University of Foreign Studies, holding a PhD in Management and serving as a master’s supervisor and visiting scholar at Lancaster University, UK. Her research explores the international business of Chinese firms, global value chains, and global supply chains. Through empirical analysis and case studies, she examines how firms from emerging economies integrate into and reshape global networks. Her work has appeared in top Chinese journals such as Management World.
Qihai Huang (q.huang@hud.ac.uk) is Strategic Lead/Associate Dean for Research, School of Business, Education and Law, and Professor of Entrepreneurship and Management at the University of Huddersfield. His research focuses on entrepreneurship, innovation, and wellbeing, with a particular interest in emerging markets. He has published widely in leading academic journals and actively contributes to advancing management scholarship. He is committed to fostering academic excellence and supporting the development of future business leaders through research, teaching, and leadership.
