Pre-Liberalization Foundations and the FDI-Based Internationalization of SMEs from Emerging Markets

Research on the internationalization of small and medium enterprises based in emerging markets (EM-SMEs) is gaining momentum, yet, less is known about the specific factors that deter EM-SMEs’ from internationalizing via foreign direct investment (FDI). In this paper, using institutional and organizational imprinting perspectives, we argue that EM-SMEs founded in the era prior to market liberalization are less likely to internationalize via FDI than those founded during or after market liberalization. We also argue that this effect is moderated by EM-SMEs’ size and ownership dispersion. Our data used to test our hypotheses is based on 2277 SMEs from 14 emerging markets. Overall, we contribute to an improved understanding of the factors that determine the FDI-based internationalization of SMEs from emerging markets.


Introduction
Recent studies are increasingly focusing on the distinctive factors that determine the internationalization of small-and medium-sized enterprises based in emerging markets (EM-SMEs) (Deng & Zhang, 2018;Fabian et al., 2009;Javalgi & Todd, 2011;Musteen et al., 2014;Narooz & Child, 2017;Radulovich et al., 2018;Wu & Zhao, 2015). However, most prior studies on EM-SMEs' internationalization have focused on their exporting behavior, rather than on their foreign direct investments (FDI) 1 3 (Deng & Zhang, 2018;Jean & Kim, 2020;Kim & Hemmert, 2016;Manolova et al., 2010;Tiessen et al., 2001). Engaging in FDI requires greater resource commitment as well as risk-taking propensity, and in general, due to SMEs' resource and capability constraints, exporting has remained a key mode of their internationalization. However increasingly, SMEsnot limited to those from emerging economies -are also engaging in FDI in a more accelerated fashion (Huett et al., 2014;Jansson & Sandberg, 2008;Stoian et al., 2018). Yet, as a recent literature review highlights (Laufs & Schwens, 2014), research on SMEs' FDI is still in its infancy, and especially, little is known about the unique barriers faced by SMEs from emerging markets in engaging in FDI (Qiao et al., 2020).
Prior studies have acknowledged that EM-SMEs, as compared to their developedcountry counterparts, lack sophisticated resources and capabilities to a greater extent (Javalgi & Todd, 2011;Narooz & Child, 2017). This causes EM-SMEs to suffer from relatively greater 'liabilities of smallness' (Ko & Liu, 2017;Lefebvre, 2020), which may further deter their internationalization prospects. At the same time, EM-SMEs also face 'liabilities of origin' (or liabilities of 'home') (Fiaschi et al., 2017;Marano et al., 2017) caused by the institutional voids in emerging markets that result in a negative perception of emerging-market firms and their products and services in international markets (Agnihotri & Bhattacharya, 2019). As such, due to these additional liabilities generated by EM-SME's home-institutions, the barriers for internationalization for EM-SMEs are higher. In this situation, undertaking FDI may be riskier for EM-SMEs. Nonetheless, FDI is regarded as an important internationalization strategy for EM-SMEs to catch-up with their developed-country counterparts on technological fronts, as well as to escape from the potentially rising domestic costs and other institutional challenges in emerging economies (Cuervo-Cazurra, 2016;Qiao et al., 2020;UNCTAD, 2019).
Due to the importance of home-institutional factors in the internationalization of emerging-market-firms (mentioned above), we first suggest that research on EM-SMEs' FDI can further benefit by taking into account the institutional conditions under which the SMEs were founded, especially whether they were founded before or after the market-liberalization era in emerging markets. Founding conditions, in general, form an important issue in entrepreneurship research (Bamford et al., 2000;Eisenhardt & Schoonhoven, 1990;Geroski et al., 2010). During the 1980s and 1990s, many emerging economies across the world liberalized their markets by reducing tariffs on imports, by promoting exports and foreign investment and by privatizing state-owned enterprises; and all of this has been of significant importance to the development of SMEs in such countries (Gupta et al., 2004;Yamak & Üsdiken, 2006). Prior studies have also emphasized how institutional changes following liberalization have impacted firms' internationalization, however, these studies have focused on large, group-affiliated and state-owned firms (Chittoor et al., 2009;Dau, 2012Dau, , 2013, and lesser so on SMEs. Institutions refer to the set of rules and regulations, the level of bureaucracy, the mechanism for accessing markets, the judiciary, government policies, contracting regimes, and the degree of stability that influence business activity (Cuervo-Cazurra, 2008;Denzau & North, 1994;Kimberly, 1979). Organizations develop distinct resources and capabilities in response to the external institutional characteristics Pre-liberalization Foundations and the FDI-Based… surrounding them, and these capabilities form the basis of their competitive advantage (Oliver, 1997). External institutions surrounding firms at the time of their founding often leave a permanent imprint on them, making them difficult to change when new or different institutions are introduced in the future, since the founding period is such a sensitive time (Stinchcombe, 1965). In line with this, we expect that EM-SMEs founded in the pre-market-liberalization era would vary in their FDI behavior as compared to those founded in the later periods. This is because, before the market-liberalization era, institutions favoring protectionism in most emerging markets would have deeply impacted SMEs; and these would persist well-beyond the market-liberalization period (Maksimov et al., 2017). In contrast, during and after the liberalization period, governments in emerging markets encouraged competitive learning and risk-taking (Anand et al., 2006;Raithatha & Popli, 2022); and these institutional ideologies form the basis of SMEs' resources and capabilities founded in the post-liberalization period. In this respect, our first research question is: To what extent do emerging market SMEs founded prior to the market liberalization era differ in their FDI as compared to those founded in the post liberalization era?
Second, prior research has argued that a number of factors can condition this founding institutional effect, and these primarily revolve around the growth and governance characteristics of firms (Cheng & Yu, 2008;Lu, 2002;Maksimov et al., 2017). It is argued that the founding institutional effect is greater during the developmental stages of the firm, and can fade away as firms grow (Freeman et al., 1983) and based on how they are governed (Marquis, 2003). Based on this, a notable variable that we suggest is SMEs' size. This is because, when SMEs grow (in terms of increased sales and performance) their scope of learning and risk-taking improves, and they can buffer themselves from the imprinted institutional effects at the time of founding (Bhagat et al., 2015;Ng et al., 2013). Likewise, in terms of SMEs' governance, we suggest that ownership dispersion can have a moderating effect because a larger number of shareholders on the board can increase groupthink (Johnson et al., 1993), which can reduce the imprinted mental models of pre-liberalization founding (as compared to that in concentrated ownership) and potentially increase the risk taking propensity of firms (Eddleston et al., 2008). Based on these arguments, our second research question is: To what extent do SMEs' size and ownership dispersion moderate pre-liberalization founding effect on emerging market SMEs' FDI?
Our primary contribution lies in examining the FDI-based internationalization behavior of EM-SMEs. As previously emphasized, there is limited, yet growing research on SMEs' FDI in general, as most prior studies have focused on SMEs' exporting behavior (Laufs & Schwens, 2014). Recent studies on the FDI patterns of SMEs have been, so far, limited to SMEs from developed countries such as Japan (Lu & Beamish, 2001;Urata & Kawai, 2000), Germany (Huett et al., 2014), Sweden (Jansson & Sandberg, 2008) and Spain (Shin et al., 2017). These studies have focused on the economic characteristics of the 'host' markets and on the unique resources and learning capabilities of SMEs (Zahra, 2005). Urata and Kawai (2000), for instance, suggest that attractive markets in developed countries and the availability of lower labor-cost resources in developing countries were the key motivations for Japanese SMEs' FDI. Likewise, Huett et al. (2014) suggest that knowledge intensive SMEs are more likely to enter developed countries, whereas internationally experienced SMEs are more likely to enter developing countries via FDI. Studies on international new ventures and born globals (Øyna et al., 2018) also provide valuable insights in this direction. E.g.  find that that new ventures (based in Spain) with greater international market orientation, and those with greater marketing capabilities  are more likely to engage in higher commitment foreign entry modes. We take this research forward by focusing on EM-SMEs, as they face unique 'home' institutional and resource-based challenges, and given that emerging markets have undergone significant institutional changes following the market-liberalization period. As such, theoretically, we take the discussion forward from the 'host' to the 'home' based factors that determine SMEs' FDI in general.
In doing so, we also contribute the emerging literature on the unique barriers faced by EM-SMEs in undertaking FDI. A limited number of studies have also examined how communist institutions have impacted the FDI of firms from specific home-countries such as China (Marquis & Qiao, 2018) and Poland (Ciszewska-Mlinaric et al., 2018); however, unlike ours, these studies focus on large firms and on firms from single countries, while there is limited research focusing on SMEs in this context (Qiao et al., 2020). Our study makes an empirical contribution to these studies by re-examining the institutional effect in the context of SMEs, as well as by: (1) examining important firm-level contextual factors that moderate the founding institutional effect; and (2) by focusing on a larger set of home-countries and contributing to the greater generalizability of findings.
Our paper is structured as follows. In the following sections, we first present our theoretical background and formulate our hypotheses. We then describe our data and present our findings. Finally, we discuss our results and conclude paper by highlighting our contributions and limitations and by suggesting worthwhile avenues for future research.

Institutions, Resources and Barriers to Internationalization of EM-SMEs
The institutional theory has been an important lens in the analysis of emergingmarket (EM) based firms in general (Cuervo-Cazurra, 2008;Cuervo-Cazurra et al., 2018a, 2018bCuervo-Cazurra et al., 2018a, 2018bLuo & Zhang, 2016). As firms are embedded in their home-institutions, they develop routines and practices that reflect those of their home-institutions (Cuervo-Cazurra, 2008). As such, they face liabilities of foreignness -the costs of adjusting to new institutions when investing abroad. Prior studies suggest that, in addition to the liabilities of foreignness, EM firms in particular (including EM-SMEs) suffer from liabilities of 'home' (Stevens & Shenkar, 2012). This is because, emerging markets have weaker formal institutions, which leads to poorer governance structures and business ethics among firms based in such countries (Agnihotri & Bhattacharya, 2019;Fiaschi et al., 2017;Marano et al., 2017;Tashman et al., 2019). These liabilities create additional barriers for EM firms' internationalization. Due to this, many EM firms tend to invest in developed countries to escape from these challenges at home (Barnard & Luiz, 2018;Cuervo-Cazurra, 2016;Jormanainen & Koveshnikov, 2012).
EM-SMEs also lack strong technological capabilities as well as human resources to compete on equal terms with large enterprises and with SMEs from developed countries (Almor & Hashai, 2004;Laufs & Schwens, 2014). Home-institutional factors, as described above, complement these resource constraints. The information asymmetries in emerging markets' arising from institutional voids, for instance, reduce the utility of SMEs' managerial knowledge, and this subsequently deters SME internationalization (Narooz & Child, 2017). Studies also highlight how institutional changes following market-liberalization, such as technological upgrading and greater capital availability, have benefited firms to internationalize (Chittoor et al., 2009;Dau, 2012Dau, , 2013Popli et al., 2017). These studies have, however, focused on large firms, and only a limited number of empirical studies confirm that institutional developments following market-liberalization are also beneficial for SMEs to internationalize (Cieslik & Kaciak, 2009;Ciszewska-Mlinaric et al., 2018;Shapero & Sokol, 1982). An alternative set of studies suggest that despite the market-liberalization in emerging economies, the weak institutional systems with greater levels of corruption, instability and bureaucracy, have impacted SME internationalization negatively (Shirokova & McDougall-Covin, 2012). These factors create barriers for EM-SMEs to obtain reliable information and knowledge from their home-institutional sources that may be useful in undertaking internationalization. Tovstiga et al. (2004), for instance, suggest that institutional voids in regards to intellectual property protection, international trade regulations and warranties, and customs and tariffs also result in obstacles for SMEs to internationalize.
Overall, prior studies have shown that there are various barriers to internationalization for EM-SMEs. Differences in home-institutional settings form an important basis of the differences in the ways EM-SMEs internationalize as compared to SMEs from developed countries. In this context, research could benefit from a comprehensive analysis that takes into account both institutional and resource barriers, and we suggest that a combination of institutional and resource-based theory can provide a solid theoretical framework for following such a path, which we undertake within our research.

Pre-Liberalization Institutions and EM-SMEs FDI-Based Internationalization
We rely on organizational imprinting (Stinchcombe, 1965) mechanisms within institutional theory to develop our arguments on the relationship between pre-liberalization founding and the FDI based internationalization of SMEs from emerging economies. Organizational imprinting has been an important lens in understanding SMEs' growth in general (Bamford et al., 2000;Mathias et al., 2015). It is argued that external institutional conditions at the time of organizations' founding impact their future actions (Kriauciunas & Kale, 2006;Wei, 2017). During founding stages, characteristics of the external environment get 'stamped' onto organizational behavior, and these characteristics persist despite subsequent future changes in the business environment (Eisenhardt & Schoonhoven, 1990;Marquis & Tilcsik, 2013;Shinkle & Kriauciunas, 2012). Such an imprint occurs because when organizations face common conditions of uncertainty arising from the external environment, they are likely to develop common models in interpreting the expectations laid by the external environment and in taking actions in response (Denzau & North, 1994). Such actions and strategies developed in response are regarded as the basis of competitive advantage by firms.
Market (or economic) liberalization in most emerging markets has been characterized by a transition from protectionist trade and investment policies towards a freemarket-based business environment. Although emerging countries have differed in their implementation of this transition, most countries across the world have adopted a standard set of reforms (Önis, 1992). For instance in Central and Eastern Europe, the fall of communism led many countries (including Hungary, Poland and Czech Republic -then Czechoslovakia) to eliminate tariff and non-tariff barriers (Cieślik & Hagemejer, 2011). Similar policies were employed in other emerging countries such as India (Raithatha & Popli, 2022), Russia (Fish & Choudhry, 2007), China (Woo, 1994), Turkey (Aricanli and Rodrik (1990), as well as several Latin American countries such as Brazil and Mexico (Anand et al., 2006). Economic liberalization has also focused on the development of a positive mentality about internationalization among SMEs (Yamak & Üsdiken, 2006). This has been facilitated through improved availability of finance, reduced tax and the development of export processing zones where SMEs are encouraged to engage in export-focused manufacturing (Bustos, 2011). Greater levels of imports and FDI in these countries also provided opportunity for technological learning for SMEs by collaborating with large foreign multinational enterprises (Prashantham & Birkinshaw, 2008). Government control over business activity was significantly reduced and this encouraged enterprises to compete with other firms by engaging in risk-taking and innovation (Raithatha & Popli, 2022).
However, in the pre-liberalization era in most emerging markets, first, the institutional mechanisms of resource access by enterprises were strongly influenced by a high level of government control and intervention (Kriauciunas & Kale, 2006). For example, in India, the 'license raj' during the pre-liberalization era between 1956 to 1980 meant that smaller enterprises were required to obtain a license from the government to undertake any production beyond their licensed capacity; however, large firms were allowed to pre-empt this requirement (Majumdar, 2004). SMEs, in particular, had very limited access to financial and knowledge-based resources to undertake expansion. Such an environment placed little emphasis on competition (Anand et al., 2006), and therefore, SMEs in particular, had less incentives in improving their product (or service) quality, or their innovativeness (Raithatha & Popli, 2022). Importantly, pre-liberalization regimes discouraged the risk-taking propensity of SMEs by micromanaging them -for example, by imposing fines and penalties to those who did not align their activities and strategies with the ideals of these regimes (Banalieva et al., 2018;Han et al., 2014;Roth & Banalieva, 2016).
Based on institutional imprinting logics, the effect of such regimes would have had a deep impact on the business models of SMEs founded during this period in as much as their propensity to undertake FDI-based internationalization is concerned. This is because, as discussed previously, FDI-based internationalization is a risky strategy for EM-SMEs due to their relatively scarce resources as well as due to their liabilities of origin. Institutional factors during pre-liberalization would complement the risk perception of SME managers, and are likely to leave a long-lasting imprint on the mental models and strategies developed in SMEs during this time. As such, due to the institutional effect of such conditions at the time of founding, we expect EM-SMEs founded in the pre-liberalization era to be less likely to develop the innovative technologies and the risk-taking propensity important for engaging in FDIbased internationalization, even after market-liberalization. Due to this, we expect SMEs founded prior to the market liberalization era to be less likely to engage in FDI-based internationalization in the future than those founded in the pre-liberalization period.
Hypothesis 1: EM-SMEs founded in the pre-liberalization era are less likely to engage in FDI than those founded in market-liberalization period.

The Moderating Effect of SMEs' Size
According to both institutional and organizational imprinting perspectives, the effect of the external environment engraved on firms' behavior varies by the growth of the organization (Felin & Zenger, 2009;Hsu & Lim, 2013). Despite SMEs being smaller organizations by themselves (as compared to large-sized firms), their scope of learning and risk-taking propensity varies significantly in terms of their size and growth (Dimitratos et al., 2003). An increase in SMEs' size reflects the possession of greater resources and enables them to get easier access to credit, which can be utilized in exploring riskier strategies such as diversification and innovation (Beck & Demirguc-Kunt, 2006). Increased size also reflects a change in their boundary conditions and learning capabilities, in that, larger SMEs tend to have a larger network of partners, suppliers and customers as compared to smaller SMEs (Dickson et al., 2006). Due to their relatively wider business connections and networks, larger SMEs are better equipped in terms of both knowledge and other financial resources, and can reduce their dependence on regulatory institutional actors in emerging economies (Haveman, 1993). Due to this, we expect that the institutional effect of preliberalization founding would be greater among SMEs that have remained smaller over time, than among larger SMEs. Thus, we propose the following hypothesis: Hypothesis 2: The effect of pre-market liberalization founding on EM-SMEs' FDI-based internationalization reduces with the increasing size of the SME.

The Moderating Effect of SMEs' Ownership Dispersion
Finally, we argue that the effect of external institutions on firms can also be moderated by SMEs' ownership-dispersion characteristics. Firms with more dispersed ownership are characterized by a greater number and diversity of shareholders or board members (Jacoby & Zheng, 2010;Zheng & Li, 2008). Greater diversity and size at the shareholder-level leads to more discussion and argumentation in strategic decision-making, and this can reduce the impact of institutionally imprinted behavior on individual owners or board members (Eddleston et al., 2008). Likewise, a greater number of shareholders allows managers to take decisions more independently, to take more risks, and to engage in more innovativeness (Maksimov et al., 2017). Conversely, firms with more concentrated ownership, such as those having single or few shareholders, tend to reflect a greater institutional effect of the external environment and would have a greater scope of directing the management based on their imprinted mental models. Once again, we expect this logic to apply in relative terms for SMEs with varied levels of ownership dispersion. Thus, if an SME was founded in the pre-liberalization era, greater levels of dispersion in SME ownership will make the SME lesser susceptible to the founding institutional effect of such a regime, and lead to a greater level of adaptability when the market-liberalizes, subsequently enabling the SME to engage to a greater extent in FDI-based internationalization. Based on this, we propose the following hypothesis: Hypothesis 3: The effect of pre-market liberalization founding on EM-SMEs' FDI-based internationalization reduces with ownership dispersion of the SME.

Data Collection and Sample
Our data has been collected from Bureau Van Dijk's Orbis database. Orbis provides financial and other market information of millions of companies on a worldwide basis, and has been used in a variety of studies, including on SMEs (Muñoz-Garcia & Vila, 2019; Valtakoski & Witell, 2018). Within Orbis we searched for firms based in all the emerging market countries listed in the International Monetary Fund (IMF), having less than 250 employees, and having at least one foreign subsidiary with a majority (i.e. at least 51%) shareholding. Orbis does not provide information about foreign subsidiaries created by firms on a yearly basis -the information is provided based on the latest year as reported by companies. This resulted in a sample of firms covering the years 2015-2019. Searching for firms in a single year would have drastically reduced our sample. Thus, our sampling strategy is based on our research questions and data availability. We wanted to focus on the FDI-based international expansion of EM-SMEs and that is why we searched for EM-SMEs with at least one foreign subsidiary during this period. We also excluded firms with zero foreign subsidiaries as including such firms would have led to a larger sample but also would have led to very low effect-sizes in regressions. We used the European definition of SMEs for SME selection as this is most popular in prior studies (Bacon & Hoque, 2005). The definition of SME differs in various parts of the world (Storey, 1994) based on size, shape and capital employed. In Europe, SMEs are defined based on employee numbers such as micro (0-9), small (10-99) and medium (100-250). Based on this sampling strategy and after accounting for missing values in our data, we were left with a total sample of 2277 SMEs. Although our sample appears small, it is also indicative of the fact that only a few SMEs engage in FDI, which is the primary basis of our sampling strategy. The final sample of SMEs was based in 14 emerging market countries. Table 1 provides a description of our sample based on the SMEs' home country.
Based on Table 1, we do recognize that SMEs from some emerging countries are less represented than others. This could be due to a number of reasons. European emerging countries, for instance, are represented more as our selection of SMEs as based on the European definition. In other countries that are represented less, SMEs rarely existed prior to the pre-liberalization era, and found it hard to survive as they were not represented in the national planning system, with priority given to state-owned firms and large, group affiliated firms. At the same time, our sample also reflects the nature of engagement in FDI by EM-SMEs. FDI by SMEs from Latin America and Asia continues to be much less as compared to that of SMEs from emerging European countries. Regional investment regulations may also be an important factor affecting the sample. As such, we did not want to exclude countries with smaller SME numbers to preserve this heterogeneity, and we explore the aforesaid relationships in our analysis.

Measures
Our key dependent variable is SMEs' FDI-based internationalization which is measured by the ratio of foreign subsidiaries to total subsidiaries (RFSTS). Such scope metrics of internationalization have been used in various prior studies (Delios & Beamish, 1999;Kedia & Mozumdar, 2003;Lu & Beamish, 2001;Morck & Yeung, 1991;Pantzalis, 2001;Zahra et al., 2000). While various  Boratav et al. (1996) South Africa 12 1990 Nowak and Ricci (2006) measures of internationalization exist -such as export revenues and ratio of foreign to total sales, we suggest that our measure, which is in line with our research objectives, enables us to effectively measure the extent of emerging market SMEs' greater commitment to foreign markets. As such, our measure allows us to understand the extent to which the firm is likely to set up a subsidiary in a foreign market versus in its home market.
Our key independent variable is pre-liberalization founding. Different emerging market countries have had different years in which the market liberalization era started. As seen in Table 1, we relied on academic literature to find out the year in which market-liberalization occurred in the respective emerging-market country. Based on this year, we operationalized our independent variable using a dummy variable that took the value 1 if the SME was founded prior to the market-liberalization year, and 0 otherwise. Such a mechanism of measuring the imprinting effect of home institutions (based on firms' founding year and country-of-origin) is commonly used in prior studies (Ciszewska-Mlinaric et al., 2018;Shinkle & Kriauciunas, 2012;Shirodkar et al., 2017;Vidaver-Cohen et al., 2015). However understandably, this is an imperfect measure given that both pre-and post-liberalization periods are long and can vary across countries. We deal with this issue in our robustness tests and post-hoc analyses.
Based on previous literature, we incorporated several control variables, some of which are used to test our moderating effects. The data for these variables were collected from Orbis and for the same year as the SME invested in FDI. First, we measured SMEs size (our first moderator) by the total assets, following prior studies (Shen et al., 2009;Sogorb-Mira, 2005). We also control for SMEs ownership dispersion (our second moderator), which is measured by the number of shareholders, following past studies that have used this measure (Jacoby & Zheng, 2010;Zheng & Li, 2008).
Among other control variables, we first control for gearing or leverage, which forms an important firm-specific attribute in internationalization (Agyei-Boapeah, 2015;Pisani et al., 2017). We also control for the profitability of the SME measured by return on equity (ROE), which is also an important attribute in this context (Majocchi & Zucchella, 2003). We also control for the independence level of SMEs. Prior research suggests that SMEs could vary in their independence levels in terms of their ways of financing and based on informal associations with other large firms (Russo & Perrini, 2010). We calculated this based on the independence level of the firm obtained from Orbis. Orbis rates the independence level based on 'A' (for highly independent), 'B+', 'B' 'C' up to D (highly dependent). We convert this into a scale variable with 0 for D, through 4 for A. We control for the industry sector and measure it by a dummy variable that takes the value 1 for manufacturing and 0 for other sectors. We also control for a dummy variable that indicates whether the SME is based in the European Union i.e. EU (1) or not (0). This is because, as seen in our Table 1, nearly 50% of SMEs in our sample come from emerging/transition economies within the EU. As the EU allows special regional benefits for investment, this is likely to have an impact on SMEs' internationalization (Manolova et al., 2010). Likewise, we also included dummy variables for other regions, including the Americas, Africa, Middle-East and Asia depending on the SME's home country. Finally,

Pre-liberalization Foundations and the FDI-Based…
we include year dummies for the years 2015-2018 as control variables, as the SMElevel data we obtain from Orbis is spread across these years. Table 2 provides the descriptive statistics, means and correlations between the variables. As we do not see any high correlations between the independent and control variables, we expect that there were no multicollinearity problems at this point, however we also verified this by checking for the VIF (variable inflation factor) in our regressions (Schroeder et al., 1990). Table 3 provides the regression results with FDI-based internationalization -measured by ratio of foreign subsidiaries to total subsidiaries (RFSTS), as the dependent variable.

Descriptive Statistics and Regression Results
In Model 0, we include the control variables only. Then in Model 1, we include our independent variablepre-liberalization founding of the home country. The model shows that pre-liberalization founding is significantly and positively associated with RFSTS (β = − 0.052, p = 0.02), suggesting that being founded in the preliberalization era reduces the probability of investing overseas (vis-à-vis domestically) by 5.2% units. This supports our Hypothesis 1. The size of the effect, is however, small, and this is consistent with some prior studies (Shinkle & Kriauciunas, 2012). The effect size may depend on the timing of pre-liberalization founding as well as level of institutional change brought about as a result of market-liberalization. In many countries within our sample (especially, European countries), the change in institutions following liberalization could be argued to be relatively of much smaller magnitude than that in other emerging markets (such as Peru). In Peru, for instance, the pre-liberalization era was characterized by several years of military rule, severe macro-economic issues such as high inequality and hyperinflation, and despite the move towards market-liberalization advocated by the government, the working-class and various labor groups were extremely skeptical about the new institutional changes (Stokes, 1996). In contrast, the magnitude of institutional change between pre and market liberalization in Asian emerging countries such as Malaysia and Indonesia was not similar to Peru, and they went through a much more gradual process of market liberalization (Ritchie, 2005). Due to this, the effect of pre-market liberalization founding among SMEs from these countries could also be expected to be smaller.
In Model 2, we introduce the interaction term Pre-liberalization founding × Size to test the moderating effect of SME size. The results from this model show that the interaction term is significantly and negatively associated with FDI-based internationalization, however, the co-efficient is smaller (β = − 3.187E−07, p = 0.019). This shows that larger SMEs have a lesser pre-liberalization founding effect of market liberalization than smaller SMEs insofar as the impact on RFSTS is concerned, supporting our hypothesis 2; however, the effect-size is very marginal. To verify this result further, we adopted the method suggested by Aiken et al. (1991) and      Table 3 Regression results

Pre-liberalization Foundations and the FDI-Based…
Standard errors in parentheses ***p < 0.01, **p < 0.05, *p < 0.10 Table 3 (continued) calculated the High Size (one standard deviation above the mean of Size) and Low Size (one standard deviation below the mean of Size). We then created two separate interaction terms Pre-liberalization founding × High Size and Pre-liberalization founding × Low Size and ran two separate regressions to examine the effects at these levels of Size. For High Size, the coefficient of pre-liberalization founding changes to 0.167 (p = 0.082) and for Low Size, the coefficient of pre-liberalization founding changes to − 0.287 (p = 0.005). 1 This further confirms that, at low SME size, the negative effect of pre-liberalization founding on the propensity to carry out FDIbased internationalization is much stronger than at high SME size.
In Model 3, we introduce the interaction term Pre-liberalization founding × Shareholders to test the moderating effect of SMEs' ownership dispersion. The model shows that the interaction term is negative but not significantly associated with internationalization (β = − 0.001, p = 0.788). Thus, our hypothesis 3 is not supported by our data. An explanation of this unexpected finding could be related to alternative explanations about the complexities relating to ownership dispersion that we do not account for. Johnson et al. (1993), for example, argue that ownership dispersion increases groupthink and inertia which can inhibit risk-taking, innovation and organizational change processes (Barkema & Shvyrkov, 2007). Likewise, Wei (2017) argues that founding conditions of the firm impact the appointment of board members and its financing options -e.g. Chinese SOEs founded after the marketliberalization period tended to avoid the appointment of state-connected board members. Thus, there may be a complex relationship between pre-liberalization founding, ownership dispersion and internationalization than we might have expected.
Our control variables also provide some interesting findings. SME size showed a significant and negative effect in EM-SMEs' FDI-based internationalization, suggesting that large-sized SMEs are more likely to invest in the domestic emerging market (vis-à-vis, foreign markets); however, as the β indicates, the effect is marginal in all the models. The number of shareholders also has a significant negative impact on RFSTS, thus suggesting that EM-SMEs with concentrated (or dominant) ownership structures are more likely to internationalize through FDI. Prior studies on the impact of ownership dispersion on internationalization tend to expect a curvilinear relationship between ownership and internationalization (Liu et al., 2011;Oesterle et al., 2013), however, unlike ours, these studies have not focused on EM-SMEs. Gearing is significantly and positively associated with internationalization, however, the effect is marginal. This is consistent with some prior studies on firms from emerging markets (Gaur et al., 2014). The independence level of SMEs is also strongly and significantly associated with their FDI-based internationalization. In emerging markets, many large firms spin-off SMEs to take advantage of the special government incentives provided to SMEs (Filatotchev et al., 2009), however such SMEs are financially dependent on their large-size parent -and we find such SMEs would be less likely to invest in foreign markets. Finally, EU-based, American and Middle-East and Asian SMEs are more likely to internationalize via FDI. This could Pre-liberalization Foundations and the FDI-Based… be attributed to the greater ease of doing business within the regional market, as facilitated by their respective institutional arrangements (Manolova et al., 2010).

Robustness Tests
We conducted a number of robustness tests to verify our results. First, we used an alternative measure of pre-liberalization founding for which we subtracted the SMEs' founding year from the market-liberalization year of the SMEs' home country. This not only allowed us to use a continuous variable to measure the effect of pre-liberalization founding but also to account for the age of the SME. This is because, market-liberalization is a lengthy process and firms founded in the years closer to the market liberalization period may have a reduced pre-liberalization founding effect as compared to those founded in the earlier times. We re-ran our analyses using this alternative measure. Table 4 provides the results. All our results remained consistent with our main results. In fact, the significance of the pre-liberalization founding effect is stronger using this alternative measure, as seen in models 4-6.
Second, as our sample consisted of different emerging market countries, as an additional robustness test, we replaced our regional dummy variables (used as control variables) with country-level dummy variables. We re-ran all the regressions and yet, our results remained intact. Third and finally, we also used a different measure of SME-size (i.e. operating revenue) and re-ran all the regressions. This also provided us with similar results. 2

Post-hoc Analyses
In addition to our robustness tests, we also conducted some post-hoc analyses. First, due to the heterogeneity of the pre-liberalization era across various emerging markets, SMEs in different countries, SMEs founded in different countries might have been impacted differently in relation to the (imprinting) mechanisms leading to their future FDI-based internationalization. To investigate this, we split our sample based on the countries which provided a large enough sample size in terms of the number of SMEs and re-ran the regressions. These countries included Brazil, Czech Republic, Hungary, India, Poland and Russia (based on Table 1). The results are presented in Table 5 -in models 7 through 12. We found that the negative effect of pre-liberalization founding on the FDI based internationalization of EM SMEs holds true for the cases of Brazil, Hungary, India and Russia. For Czech Republic and Poland, the effect is negative (similar to our overall finding) but not significant. The effect sizes are also vary slightly across countries -e.g. the pre-liberalization founding effect was highest in Brazil. The country-specific effects also showed some interesting results in terms of the role of industry. E.g. while manufacturing SMEs showed Table 4 Regression results with continuous variable for pre-liberalization founding Standard errors in parentheses ***p < 0.01, **p < 0.05, *p < 0. Pre-liberalization Foundations and the FDI-Based… a positive effect of FDI in the case of most countries, SMEs in the finance industry showed a negative effect in the case of India; and SMEs in the information industry showed a positive effect in the case of Brazil and Poland. We suggest that future research can investigate these country-specific effects using larger samples. Second, recognizing the effect of industry, we also tested the moderating effect of the different industries in our sample, but this was not significant. 3 Among the firm-level factors, we tested for the moderating effect of firm-performance (ROE). We found that with greater ROE-based performance, the pre-liberalization founding effect was weakened. Please refer to model 13 in Table 6 for the results. Similar to the effect of firm-size, one may argue that superior performance may also increase the risk-taking propensity of EM SMEs and may reduce the pre-liberalization imprinting effect. However again, this warrants further research and we encourage future research to take this up.

Theoretical Implications
A large number of studies in international business are focusing on the 'institutional' determinants of emerging-market firms' internationalization (Buckley et al., 2015;Hong et al., 2015;Ramamurti, 2012). However, while large firms (such as state-owned enterprises and firms affiliated to business groups) are well-financed and better supported by the home government to internationalize via FDI (Chittoor et al., 2015;Zhou, 2018), the extent of such support is limited for SMEs. Particularly, government-incentives to EM-SMEs may be limited to exporting, although, even this has been argued to be inadequate (Pradhan & Sahu, 2008;Tinits & Fey, 2022). In the light of this, our study is the first that focuses on the factors that determine the FDI-based internationalization of EM-SMEs. FDI presents the next-stage of commitment to internationalization according to process-based models of firms' internationalization behavior, as there is both a greater level of risk and resource commitment required to engage in FDI (Johanson & Vahlne, 2009). Many prior studies on SMEs' FDI have also been based on SMEs from developed countries (e.g. Arranz et al., 2016;Dimitratos et al., 2010;Hutchinson et al., 2006), yet increasingly, studies are focusing on the distinctive factors that impact EM-SMEs' internationalization (e.g. Nakos et al., 2019;Prashantham, 2011;Radulovich et al., 2018), albeit these studies only examine the exporting behavior of EM-SMEs. Indeed, home-institutional characteristics play an important role in SMEs' internationalization from these contexts (Narooz & Child, 2017).
Based on the above, our study was motivated by the research question as to how the institutional imprinting effect of pre-market-liberalization would have an impact on the FDI-based internationalization of EM-SMEs. Using insights from institutional and resource-based theories, and based on recent works on Pre-liberalization Foundations and the FDI-Based… SME-internationalization, we argued that EM-SMEs founded before the marketliberalization period develop negative mental models about internationalization, as compared to those founded in the market-liberalization period. Whereas the pre-liberalization era was characterized by several restrictions on SMEs reducing their risk taking propensity (Banalieva et al., 2018;Han et al., 2014;Roth & Banalieva, 2016), the market-liberalization period has facilitated international collaborations, new forms of human and financial capital, and thus plays an important role in the diffusion of technologies that were needed by SMEs to enhance their market and technological capabilities. The market liberalization period also promoted entrepreneurship, allowed for skill-development in manufacturing and exporting, and enhanced the overall innovative capacity of SMEs (Maksimov et al., 2017). As new skills and managerial knowledge were released into the SMEs sector during and after market liberalization, this increased their likelihood of FDI-based internationalization (Blalock & Gertler, 2008;Blalock & Simon, 2009;Görg & Strobl, 2005;Haskel et al., 2007;Keller & Yeaple, 2009). Our empirical results provide support to 2 out of 3 of our hypotheses. First, regarding the effect of pre-liberalization founding on SMEs' internationalization behavior, we had argued that EM-SMEs founded before market-liberalization era would be less likely to internationalize via FDI, as compared to EM-SMEs after liberalization. Our results (see Table 3, Model 1) support this argument. Overall, we suggest that the effect of pre-liberalization founding forms an important variable in analyzing the extent to which emerging market SMEs will engage in FDI in foreign markets. We also attempt to disentangle some of the complexities associated with the effect of pre-liberalization founding on EM-SMEs' FDI. In our second hypothesis, we argued that EM-SMEs' size would moderate the relationship between preliberalization founding and EM-SMEs' FDI. This is because larger SMEs would have more resources to buffer themselves from the effect of local institutions (Maksimov et al., 2017). Thus the pre-liberalization founding effect would be greater among smaller SMEs, who depend on the government and other regulatory authorities to a greater extent to access vital resources. Our results support our Hypothesis 2 (see Table 3, Model 2). Finally, in our Hypothesis 3, we argued that SMEs' level of ownership dispersion would moderate on the relationship between pre-liberalization founding and their FDI-based internationalization. This is because with greater ownership dispersion, the risk taking propensity of managers would increase, creating an inertia against the effect of founding-institutions. However, our results do not support this. Yet, overall, via our arguments, we theoretically contribute to the institutional theory by advancing the conditions under which founding effects can be moderated.

Managerial Implications
Several implications for managers of SMEs from emerging economies can be derived from the study. First, our study implies that despite institutional changes in emerging economies that are supportive of market liberalization, SMEs founded in the era prior to liberalization may not be able to easily adapt to these changes and will thus miss opportunities from foreign expansion. This would particularly be the case for smaller SMEs. Managers of SMEs must therefore identify practices and blueprints that are imprinted from founding conditions and prioritize efforts to break away from such practices (Shinkle & Kriauciunas, 2012). In particular, as SMEs from emerging economies suffer from greater competitive-effects and constraints in internationalizing as compared to SMEs from developed countries, breaking away from the effect of pre-market liberalization founding is crucial for success in internationalization.

Limitations and Future Research
Like all academic studies, our study also bears some limitations, providing worthwhile avenues of further research in both theoretical and empirical terms. First, although we have provided a strong theoretical basis for the impact of pre-liberalization founding on EM-SMEs' FDI based internationalization, we were unable to explore the complexities in this relationship due to the limitations of our data. The impact of external institutions at the time of founding can be moderated by various factors such as founders' demographics (Marquis & Qiao, 2018), their levels of future learning and unlearning from other internationalization modes (e.g. exporting), subnational characteristics of the location where the SME was founded, and so on. We encourage future researchers to account for these complexities. Secondly, we look at the SMEs' data based on their current information (i.e. years 2015-2019), and this means that several SMEs that existed during the pre-market-liberalization era might have ceased to survive in the current time. Our sample selection leads to the exclusion of these SMEs. Future research could possibly address this. Third, although we include 14 emerging market countries in our sample, our sample is unbalanced in terms of the number of SMEs from each country. Particularly, there are very few SMEs from some countries such as Chile, Colombia, Indonesia, Mexico, Philippines, South Africa and Turkey. Future research could benefit from a greater representation of these countries. Yet, despite these limitations, we believe that we contribute in worthwhile ways to the so-far limited research on the FDIbased internationalization of SMEs from emerging markets. Data Availability Our data has been collected from Bureau Van Dijk's Orbis database. Due to the proprietary nature of this database, the data used for this study cannot be made publicly available.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.