Subnational institutional configurations and international expansion of SMEs in emerging economies

In contrast to institutional studies based on comparisons between countries and based on a deterministic perspective that analyses the influence of institutions on the performance of the individual in emerging economies, the present research looks to the most recent versions of neo institutionalism and to the literature on entrepreneurship and international business in order to study (1) the existence of subnational institutional differences in emerging economies and (2) whether such differences relate to the international expansion of SMEs located in such territories. This study takes evidence from a sample of 296 entrepreneurs who founded their SMEs in the Federal State of Sergipe (Brazil) and analyses from a comparative perspective of six tourist municipalities located in that state. The results allow us to identify the existence of differences in the regulative, normative, and cognitive institutional dimensions between subnational territories and how such differences relate to the international expansion of SMEs located in each subnational territory.


Summary highlight
Keywords Cognitive institutions · Regulative institutions · Normative institutions · International expansion · Entrepreneurship · Emerging countries Contributions: In contrast to institutional studies based on comparisons between countries and on a deterministic perspective that analyses the influence of institutions on the behaviour of the individuals and companies in emerging economies, the present research looks to the most recent versions of neo institutionalism and compares subnational territories to provide evidence of the impact that institutions have on the international expansion of SMEs managed by entrepreneurs. Furthermore, unlike other studies that analyse the individual effects of regulative, normative, and cognitive institutional dimensions, this work offers an analysis of their joint effect on the internationalisation of SMEs.
Research purpose: This research examines the existence of subnational institutional differences in emerging economies, as well as the question of whether such differences relate to the international expansion of SMEs located in such territories.
Theoretical framework: Institutional theory provides the main theoretical bases for grounding the study, which also draws on the literature on entrepreneurship and international business.
Basic methods and information/data: This study takes evidence from 296 entrepreneurs who are founders, owners, and managers of SMEs in six tourist municipalities located in the Federal State of Sergipe (Brazil). Data was collected through a structured questionnaire during face-to-face interviews addressed by six trained Brazilian interviewers. Data was analysed from a comparative perspective for the six tourist municipalities.
Results/findings: Results indicate that the six territories compared show significant differences in regulative, normative, and cognitive aspects of institutions and that such differences between territories relate to the international expansion of SMEs located in such territories. In addition, this work finds that the effects of each institutional dimension on the SMEs' international expansion are contingent on the JEL B52 · F23 · L26 1 3 Subnational institutional configurations and international… development and content of the other institutional dimensions with which they interact in various configurations.

Limitations (if there are any):
The study takes evidence from SMEs located in a single country, a single federal state and a single industry. In addition, the research is based on a sample of companies registered in a public census that was not complete, so younger companies that are not yet registered or those from the informal sectorwhich are not unusual in emerging countries-have not been considered.
Theoretical implications and recommendations for further research: The present work offers some contributions to the literature on the intersection between institutions, entrepreneurship, and internationalisation of SMEs as it provides evidence on the existence and relevance of subnational institutional configurations within emerging economies, as well as on the relationship between such configurations and the international expansion of the SMEs managed by entrepreneurs and located in those areas. In particular, the study finds that culture, which is a highly relevant institution in emerging countries with a strong impact on entrepreneurial activity, is insufficient by itself (that is, in the absence of regulative and cognitive institutions that support international activity) to promote the international expansion of entrepreneurs.
Practical implications and recommendations: Results may be useful for entrepreneurs located in emerging economies who are interested in international expansion. They should be aware of the existing institutions in the territories in which they are located and how such institutions can condition their international aspirations; considering moving to other territories within the country with more favourable institutions for internationalisation, as well as engaging in bottom-up initiatives to modify institutions through business associations, are some of the possible options for them to contemplate.
Policy implications and recommendations: First, policymakers should consider that the formulation of specific policies and single support programmes, as well as their general application to the whole of a country, may be insufficient to promote the international expansion of SMEs if such policies and programmes do not consider the institutional configuration that prevails in each subnational territory. Second, to promote the international expansion of SMEs in their countries, policymakers must at least guarantee (1) the existence of public resources available to the entrepreneur-e.g. state aid, regional funds, information offices that address issues of relevance to the expansion of ventures-(2) the availability of training opportunities to create a knowledgeable population to manage these activities for the firm, and (3) a culture conducive to successful and expansive entrepreneurship, consistent with the local idiosyncrasies of each territory.
Future research suggestions: Other decisions relevant to understanding differences in the internationalisation strategy implemented by SMEs located in subnational territories, such as location choice, entry mode, speed and scale of operations abroad and international performance, and how they are conditioned by subnational institutions should be studied.

Introduction
The increasing interest in the study of internationalised firms from emerging countries in the 2000s (e.g. Demirbag et al. 2009;Luo and Tung 2007;Pananond 2007;Yiu et al. 2007) led to increasing attention being subsequently paid specifically to small and medium-sized enterprises-SMEs-(e.g. Adomako et al. 2020;Cardoza et al. 2016;Ketkar and Acs 2013). This type of firm is the research focus of this work because its growth and internationalisation represent an entrepreneurial activity of high relevance in the promotion of economic development and the generation of wealth in emerging and developing countries (Chandra et al. 2020). Nevertheless, there is still a lack of awareness regarding the factors that condition the decision to expand internationally among the entrepreneurs who manage them (Cardoza et al. 2015).
Some factors considered in the literature that is specifically interested in finding the drivers of firms' international expansion in emerging economies are the firm-specific ownership advantages (Cheng and Yu 2008), the local geographical resources (Gilmore et al. 2003), the economic openness (Luo and Tung 2007), and the institutional environment (Chan et al. 2008;Demirbag et al. 2009). However, criticisms have emerged regarding the suitability of some of these factors. For example, approaches based on resources may be suitable for countries with stable institutions but may not be suitable for emerging economies (Peng et al. 2008), where low munificence of resources exists (Yiu et al. 2007) and ownership advantages are difficult to reach (Yan et al. 2018;Meyer and Nguyen 2005;Pananond 2007). As an alternative, authors have increasingly focused on institutional theory for grounding their studies (e.g. Cardoza et al. 2015;Cheng and Yu 2008;Luo and Zhang 2016;Yan et al. 2018).
Accordingly, this paper looks to institutional theory to find the antecedents of the decision made by SMEs based in emerging economies to expand internationally. Previous works have so far considered the institutional framework as representing stable and fixed conditions (Peng et al. 2008), encompassing regulative (e.g., laws), normative (e.g. cultural values), and cognitive (e.g. socially admitted business practices) institutional dimensions (Scott 1995(Scott , 2014. In particular, they have analysed the effect of such institutional dimensions on international business activity (e.g. Demirbag et al. 2007;Meyer et al. 2009), particularly in the case of SMEs (e.g. Cardoza et al. 2015;Deng and Zhang 2018;Wu and Deng 2020). In these studies, it has been common to omit the analysis of some institutional dimensions (Trevino et al. 2008) and, in the majority of cases, these have been considered analytically differentiated aspects of institutions (e.g. Adomako et al. 2020). However, recent versions of neo institutionalism highlight that aspects of institutions are in practice interconnected (Szyliowicz and Galvin 2010); indeed historical evidence shows their joint effect as complements or substitutes (e.g. Park 2021). Thus, it is necessary to develop new research that considers the interactions that give rise to the institutional configurations that facilitate the international expansion of SMEs.
Likewise, previous studies have been carried out from a macro perspective focused on national comparisons (e.g. Adomako et al. 2020;Fuentelsaz et al. 2019;Puffer et al. 2016) and have overlooked other levels of institutional influence (Donnelly and Manolova 2020). This contrasts with the fact that each local institutional territory has a degree of idiosyncrasy (North 1990(North , 2005, thus institutional differences likely exist at the provincial and municipal levels within a country (Meyer and Nguyen 2005), which is considered by recent versions of neo institutionalism (Alvesson and Spicer 2019). In this respect, the institutional peculiarities of a location-e.g. tax rates, level of training of the local population, etc.-can affect not only the business decisions made by the SMEs located in those areas (Bjørnskov and Foss 2016;Chen et al. 2020) but also the effectiveness of the policies and programmes implemented by their governments to encourage a specific business practice or strategy (Cantwell et al. 2010;North 2005;Rodríguez Pose 2013). In this regard, Donnelly and Manolova (2020) call our attention to the lack of knowledge about the specific levels of institutional influence (e.g. supranational, national, subnational) on international business decisions, and Chandra et al. (2020) remark on the need for further studies that examine how institutions interact within countries to facilitate the internationalisation of SMEs.
On the basis of the above, and in line with the most recent versions of neo institutionalism, this study aims to answer the following research question: what subnational institutional configurations favour the international activity of the SMEs located in them? For this purpose, we first study the existence of subnational institutional differences within an emerging country and later whether such differences relate to the international expansion of SMEs located in such territories. We take evidence from entrepreneurs located in the Federal State of Sergipe, Brazil. This country has emerged as one with a significant capacity to expand internationally, due mainly to the notable role played by its multinationals (Abreu Campanario et al. 2012) but also because of the growing internationalisation of its SMEs (Cyrino et al. 2010).
In addition, and given that Jackson and Deeg (2008) warn that each institutional framework can provide different strengths and weaknesses to different industries, it seems advisable to choose a specific industry in which to carry out the research. In accordance with this, the present work focuses on the tourism industry as this is a key strategic axis for the socioeconomic development of emerging economies (Liu and Wall 2006). The tourism industry provides a valid way to increase the foreign expansion of a country's businesses, and not only through the activity of its multinationals-e.g. large hotel chains, airlines-but also through its SMEs-e.g. firms that organise leisure and recreational activities such as golf, scuba diving, deep-sea fishing; consultancy in the areas of architecture, urbanism and design specialising in hotels, etc. (López de Ávila 2013). However, entrepreneurship research has been overlooked in the tourism industry (Altin et al. 2017). Furthermore, the international expansion of SMEs in this industry has not sufficiently attracted the research interest of academics in the field of international business.
We structure our manuscript as follows. We first present the theoretical foundations of the work that are mainly based on institutional theory, and we develop hypotheses and propositions. Second, we describe the methodology of the empirical study carried out. Third, we report findings in two subsections, the first related to the identification of subnational institutional configurations and the second focused on the analysis of the relationships between such configurations and the internationalisation of SMEs located in each territory under study. Fourth, we discuss results, and finally, we end the manuscript with a section of conclusions.

Theoretical foundations
In the search for the conditions that favour the international expansion of firms in emerging economies, previous literature has considered traditional drivers such as firm-specific ownership advantages (Cheng and Yu 2008), local geographical resources (Gilmore et al. 2003), economic openness (Luo and Tung 2007), and the institutional environment and its level of development (Chan et al. 2008;Demirbag et al. 2009). With respect to ownership advantages, its relevance in analysing the international expansion of firms from emerging economies has not been exempt from criticism due to (1) the difficulties faced by these firms in developing such assets-e.g. based on technology- (Yan et al. 2018) and (2) the existence of alternative sources of advantage in such economic contexts-e.g. ability to engage in networks- (Meyer and Nguyen 2005;Pananond 2007;Yiu et al. 2007). Peng et al. (2008) warn that the approach based on resources emerged in the competitive scenario of the USA, where there is a stable and market-oriented institutional framework, which is not the reality in emerging countries. For Yiu et al. (2007), the low munificence of resources existing in emerging economies calls for a review of this theoretical perspective to explain the international expansion of firms.
On the other hand, in the work of Rodrik et al. (2004), the three remaining determinants were compared, and it was found that once the institutional variables are introduced into the models, local geographic resources and economic openness lose their significance or, if anything, maintain a weak effect. Along this line, Valliere and Peterson (2009) did not find any effect of incoming direct investment flows on the growth of emerging economies. These results can be partly explained because although the positive influence of foreign capital investment on the internationalisation of local firms has been demonstrated due to several factors-e.g. transfer of financial resources and the 'spillover effect' (Cantwell et al. 2010), increased business opportunities and business relationships (De Clercq et al. 2008), etc.-this effect is not always direct. Obstacles such as the limited absorption capacity of local businesses or the incompatibility between the knowledge transferred by the multinationals and local practices and culture have been argued (Cantwell et al. 2010). Additionally, it has been shown that the spillover effect generated by the investment made by foreign multinationals only occurs when they interact with local firms (Damijan et al. 2003), a practice that only reaches a small group of SMEs in emerging economies.
As a result, various research works have found that the internationalisation of firms from emerging economies does not respond to the investment development path proposed by Dunning (1981) and based on the reception of incoming foreign capital (Felzensztein et al. 2022;Goldstein and Pusterla 2010;Luo and Tung 2007), placing the focus on institutional theory as an alternative and suitable perspective for carrying out the research (Cardoza et al. 2015;Cheng and Yu 2008;Luo and Zhang 2016;Yan et al. 2018). The present work follows the same theoretical perspective.
Within the literature on international business, the institutional theory is mainly based on economic and sociological institutionalisms (Peng et al. 2017). From the economic vision, institutions are described as the 'rules of the game' in an organisational field (North 1990(North , 2005 that determine the feasibility of participating in economic activity while facing a level of transaction costs associated with the deficiencies in the institutional environment (Carney et al. 2009). The role of 'players' corresponds to organisations and entrepreneurs (North 1990). Accordingly, the type of organisation that is created (North 1990) and the strategies implemented (Bjørnskov and Foss 2016), international expansion among them, will be conditioned by the opportunities provided by the institutional framework (Meyer and Nguyen 2005;Simón-Moya et al. 2014). Following North (1990), it is possible to distinguish two types of institutions: formal-e.g. laws, regulations, and contracts that facilitate economic exchange-and informal-e.g. social conventions and values that affect the interpretation of and compliance with formal regulations (Meyer and Nguyen 2005). Both formal and informal institutions collectively generate institutional quality in the territories (Park 2021).
According to the sociological vision of institutions, three dimensions-or aspects of institutions-are identified (Scott 1995(Scott , 2014. The regulative dimension includes the laws and policies formulated by the government in order to promote certain types of behaviour and restrict others (Scott 1995). The normative dimension involves values and social norms about admissible human behaviour, including the choice of entrepreneurship as a professional career (Lim et al. 2016). These institutions not only define socially acceptable objectives-e.g. obtain benefits-but also the appropriate ways to seek their attainment (Huang and Sternquist 2007). Finally, the cognitive dimension-coined cognitive-cultural dimension by Scott (2014)encompasses the knowledge shared by the organisations and individuals of a given territory. This dimension includes, among others, the decisions that have been successfully applied by organisations (Lu 2002) and so are not questioned (Scott 1995) by managers when facing new decisions (Lim et al. 2016;Lu 2002). According to Stephan et al. (2015), a kind of parallelism exists between formal/informal institutions as considered by North (1990North ( , 2005 and regulative/cognitive-normative institutional dimensions conceived by Scott (1995). Specifically, some authors relate the regulative dimension of institutions to the country's overall governance structure and law (formal institutions), and the normative and cognitive dimensions to practices applied in the business contexts that support informality (informal institutions) (e.g. Donnelly and Manolova 2020;Ketkar and Acs 2013;Stephan et al. 2015).
Focusing on the sociological tradition of institutional theory, authors have underscored how the view of institutions has evolved (Peng et al. 2017). In particular, while the early versions of the theory considered the existence of stability, as well as that organisational dynamics, took place as a result of macro forces that caused isomorphism, such perspectives started to change in the 1990s (Alvesson and Spicer 2019). In particular, the most recent versions of neo institutionalism proposed by 1 3 Alvesson and Spicer (2019) among others, advocate (1) focusing on change more than on stability; (2) contemplating variability more than uniformity and macro patterns; and (3) examining several institutions rather than a single institution. These most recent versions of neo institutionalism ground the present work.

Institutions and internationalisation of SMEs in emerging countries
Previous literature has described emerging economies as '[…] those countries whose national economies have grown rapidly, in which the economic sectors have carried out and are carrying out relevant and continuous structural changes, and whose markets are promising despite their volatility and weak legal systems' (Luo and Tung 2007, 483). These economies differ from developed ones regarding the effect of institutions on firms' international business strategies, which is largely greater in the emerging contexts (Peng et al. 2008). Indeed, in emerging economies, SMEs willing to internationalise face problems that are different from those SMEs located in developed countries (Chandra et al., 2020), and they hardly ever undertake international expansion just to exploit the resources they have, given that they often do not have ownership advantages compared to their foreign counterparts (Cheng and Yu 2008). As an alternative, they will adopt the internationalisation strategy based on the local institutional framework (Yan et al. 2018), while seeking an opportunity to increase their assets as a result of the external expansion (Yiu et al. 2007).
This approach is consistent with institutionalism, which suggests that the adoption of a particular strategy will be influenced by the institutional framework (North 1990(North , 2005Scott 1995). In particular, SMEs will adopt those strategies for which they find institutional support (Adomako et al. 2020). Concerning international business decisions, firms must consider both home and host countries' institutions (Peng et al. 2008), but it is the home institutional framework that managers mostly take into consideration when making decisions (Chan and Makino 2007).
Institutions have been analysed by the literature on entrepreneurship (e.g. Lim et al. 2016;Simón-Moya et al. 2014;Stenholm et al. 2013;Valdez and Richardson 2013) and international business (e.g. Demirbag et al. 2007;Chan and Makino 2007;Ketkar and Acs 2013;Meyer et al. 2009). Consistent with early versions of institutionalism, they have mostly been studied through macro-level proxy variables and considered single factors whose effect is analytically estimated in isolation (Szyliowicz and Galvin 2010). For example, Deng and Zhang (2018) and Wu and Deng (2020) focus on the influence of the regulative institutional dimension on the internationalisation of SMEs from China, Demirbag et al. (2007) analyse the regulative and normative aspects of institutions, and Chan and Makino (2007) study the influence of cognitive and regulative dimensions on the decisions related to the level of ownership of subsidiaries abroad.
In the context of emerging economies, previous studies have found weak institutional conditions (e.g. Gil-Barragan et al. 2020), which hinder economic development (De Clercq et al. 2010); they have also shown several effects of institutions on firms' international decisions (e.g. Chandra et al. 2020;Doh et al. 2017;Donnelly and Manolova 2020). Firstly, in emerging economies, the regulative dimension is Subnational institutional configurations and international… characterised by limited legislation that results in a legal vacuum (Puffer et al. 2010;Yan et al. 2018), bureaucracy (Chandra et al. 2020;Ghalwash et al. 2017), unforeseen, inconsistent and vaguely defined governmental policies and programmes (Demirbag et al. 2007;Ferreira et al. 2011;Puffer et al. 2010), and, in general, institutional voids (Doh et al. 2017;Puffer et al. 2016). This is especially observable at the subnational level, where greater discretion by the local authorities is frequent (Meyer and Nguyen 2005). Other weaknesses include the limited information about the competition in the market (Acs et al. 2008;Fornes et al. 2021;Chandra et al. 2020) provided by statistical agencies and information offices. Likewise, these economies usually lack incentives that stimulate entrepreneurship and business growth (De Clercq et al. 2010;Meyer et al. 2009) or such incentives are unclear or even inappropriate due to the lack of experience in policymaking by governments (Yan et al. 2018). In contrast, when the regulative dimension of institutions provides stable support for business expansion-e.g. effective and stable stimuli to boost and carry out internationalisation-entrepreneurs can trust them and undertake internationalisation because they do not feel that they will waste the resources of their SMEs due to the inconsistencies of the regulations (Volchek et al. 2015). In addition, if managers can find information offices in their territory that provide relevant data and advise on crucial issues for SMEs to internationalise (Hessels and Terjesen 2010)-e.g. required legal procedures (Trevino et al. 2008)-public aid and incentives encouraging international expansion (Luo and Tung 2007;Meyer and Nguyen 2005), and an offer of training that allows the updating of the employees to the needs of the industry in which the firm operates (Domadenik et al. 2008), they will be in a position to undertake international expansion with their SMEs. According to Meyer and Nguyen (2005), although such regulation comes from a central government and is uniform for the whole country, the effectiveness with which it is implemented in each municipality generates differences at the subnational level. Therefore, the perception of entrepreneurs in each locality on the regulative dimension of institutions may depend on the attention and help they receive from an information office, the support and speed with which they can process a request for assistance, or the possibility their employees will receive training offered by the public administration at centres, among other things.
Secondly, in emerging economies, the normative dimension of institutions is usually highly developed because it includes informal values and beliefs that tend to proliferate to compensate for the lack of a stable regulative dimension (Fornes et al. 2021;Puffer et al. 2010;Yan et al. 2018). Such normative aspects of institutions often differ among subnational territories due to the existence of strong local cultures and traditions (Cantwell et al. 2010;Meyer and Nguyen 2005;Rodríguez Pose 2013). These informal normative institutions encompass social values about admissible human behaviour and socially acceptable business objectives-e.g. power, status, as well as adequate ways of seeking their attainment-e.g. licit, regulated businesses, ethical relationships with government, etc. (Huang and Sternquist 2007). Some authors have found in emerging economies a permissive cultural environment towards corruption (Ferreira et al. 2011;Puffer et al. 2010Puffer et al. , 2016Trevino et al. 2008), irregularities in the formalisation of the business start-up and firms' subsequent activities, or hostility towards those who make their fortunes based on their investments (Busenitz et al. 2000), all of which generate uncertainty (Tracey and Phillips 2011) and harm SMEs' international expansion-e.g. in terms of internationalisation speed and results (Gil-Barragan et al. 2020). By contrast, entrepreneurs will feel encouraged to undertake international expansion when they perceive that those who are successful in business enjoy respect and status in the locality; that SMEs' success and international expansion are considered honourable and praised and such success is spread by the media; when successful entrepreneurship is the subject of great joy in private conversations, and that, in general, successful and expansive business practice is valued. In addition, the fact of having been socialised among entrepreneurs with similar objectives will legitimise such decisions (Powel and DiMaggio 1991).
Thirdly, emerging economies typically lack a large number of big companies and internationalised firms, since it is SMEs that proliferate; in addition, some SMEs operate in the informal sector. For this reason, these economies have less quality business knowledge accumulated on the basis of their firms' experience-e.g. knowledge of host markets, how to deal effectively with foreign customers-which limits the possibilities of other firms to internationalise (Chandra et al. 2020;Luo and Tung 2007). In this respect, in contexts of risk, such as the decision to internationalise, managers try to reduce the uncertainty of decision-making by imitating the successful decisions frequently made by previous firms (Scott 1995). Business habits, under this approach, offer a form of nonreflective, self-sustainable behaviour that arises as a result of actions repeated over time, which are commonly used where conditions of uncertainty exist (Lu 2002). This lack of practical knowledge of firms' experience of internationalisation is coupled with the lack of education in business in many cases. Specifically, in emerging economies, there often exists a lack of training courses and university studies in business administration (Ghalwash et al. 2017), which generates quite an absence of theoretical guides for actions on how to approach the process of business expansion. In this respect, Trevino et al. (2008) warn that education in an emerging economy is an adequate proxy for the quality of the workforce and the degree of openness of an economy to international business, and Simón-Moya et al. (2014) highlight the role of knowledge in contributing towards the internationalisation of new ventures. On this basis, we can say that in those localities where entrepreneurs perceive the existence of relevant knowledge (practical and/or theoretical) for undertaking international expansion, which can be accessed from other firms or captured through the hiring of available skilled personnel, they will feel they are in a better position to undertake international expansion. This is so because, in the internationalisation process of SMEs, knowledge of other markets plays a key role as it can decrease the levels of uncertainty in those foreign markets (Pino et al. 2021).
Finally, given that these institutional dimensions are interconnected and highly interdependent (Gries and Naude 2011; Szyliowicz and Galvin 2010), Jackson and Deeg (2008) warn that their one-dimensional study is inadequate since such analysis does not allow the examination of the effect that the interaction among institutions can have on business internationalisation. In this regard, Gries and Naude (2011) assert that the institutional dimensions reinforce each other to maintain the institutional status quo. For example, informal institutions can act as a resilient force against legislative change because, in the decision-making processes, individuals take into consideration both enacted laws and social norms and shared cognitions (North 2005). Thus, one would expect formal institutions to fail to achieve their intended effect if informal institutions cease to be congruent with the formal ones (Cantwell et al. 2010). However, their joint effect is increased if these, even with different contents, are oriented in the same direction and so complement each other.

Subnational institutional configurations and internationalisation of SMEs in emerging countries
Researchers highlight the existence of institutional differences between territories of the same country (e.g. Chen et al. 2020;Meyer and Nguyen 2005). Indeed, the most recent versions of neo institutionalism highlight local institutional idiosyncrasies (North 1990(North , 2005, proposing that the institutional peculiarities of a place-e.g. tax rates, the educational level of the local population, cultural norms, etc.-may not only affect the business decisions adopted by the entrepreneurs located in these areas but also the effectiveness of the policies implemented by the governments to encourage a specific business practice or strategy (Cantwell et al. 2010;North 2005;Rodríguez Pose 2013).
In the context of the developed economies, and concerning the regulative dimension of institutions, subnational institutional plurality is common and derives from administrative decentralisation, with governments at different levels that legislate in their areas of administrative competence. In emerging economies, however, administrative centralisation is common, although there is also a space for local singularity in such formal institutions due to the unequal effectiveness with which local authorities implement national regulations (Meyer and Nguyen 2005), or the zeal with which they pursue their fulfilment. Additionally, informal institutions, whether we look at normative or cognitive aspects of institutions (Stephan et al. 2015), also differ between territories since they arise and are consolidated from the shared experience and common learning that takes place between individuals located in a given area, who consequently face similar events and contexts (North 1990). As a result, the expectation is that each territory within a country will have its own institutional configuration, characterised by a combination of formal (regulative) and informal (normative and cognitive) institutions.

H1.
There are differences at the subnational level as far as the regulative, normative, and cognitive dimensions of institutions are concerned, with each territorial unit having its own institutional configuration.
As stated by North (1990), and later empirically corroborated by Stephan et al. 2015, formal and informal institutions act together in configurations to shape individuals' actions. Relatedly, Park (2021) reminds us that historical evidence demonstrates the interactions between formal and informal institutions, either as complements or substitutes. In addition, concerning institutional dimensions and in line with the precepts of the most recent versions of neo institutionalism, Szyliowicz and Galvin (2010) highlight that they are interconnected, and Jackson and Deeg (2008) assert that interactions between dimensions must be considered because a single dimension, individually considered, can have different and even opposite effects on the behaviour of both individuals and businesses. The nature of the effects of an institutional dimension will depend on the interaction with the other institutional dimensions in a particular configuration, as shown by Fuentelsaz et al. (2019) for regulative and normative aspects of institutions at a national level. This may be the reason why regulatory changes enacted by central governments have diffuse effects on firms located in different territories of the same country, or sometimes do not correspond with what was expected (Dunning and Lundan 2008). For example, several authors have pointed out that macro-level regulatory changes are inadequate if they are inconsistent with the local institutional configuration (Cantwell et al. 2010;North 2005). Given the existence of different institutional configurations within a country, the changes undertaken by the central government should be based on the correct combination of formal and informal institutions (Rodríguez Pose 2013).
For example, a territory with an institutional configuration based on regulative, normative, and cognitive aspects of institutions favourable to SME international expansion, which complement each other, could be expected to generate entrepreneurial opportunities perceptible to the entrepreneurs, who, as a result, will feel highly encouraged to undertake international expansion. However, given that emerging economies are characterised by limited and weak regulation and that the cognitive aspects of institutions are often of low qualitative value in terms of business knowledge for internationalisation, as already argued, the normative dimension can acquire greater relevance in providing legitimacy to the business decisions related to international expansion if it drives and legitimises such entrepreneurial action. Therefore, this second institutional configuration could potentially also lead to the international expansion of SMEs. Given the lack of previous literature that identifies the institutional configurations that drive the international expansion of SMEs, we establish, as a proposal, the following: P1. The greater the institutional complementarity favourable to expansive business activity in a specific territorial institutional configuration, the greater will be the international expansion of SMEs in that territory.

Population, fieldwork, and sample
The population under study includes entrepreneurs who manage businesses located in the six tourist municipalities of Sergipe (Brazil)-specifically, Aracaju, Sâo Cristobâo, Pirambu, Larajeinras, Estancia, and Canindé-who fulfilled the following preconditions: they are founders, owners, and managers of an SME operating in a tourism-related sector-i.e. hospitality, catering, transport, recreational activities, tour operators, retail trade, and consultancy services. The condition of the founder, owner, and manager of the SME represents a relevant requirement in the present study in order to make possible the analysis (on a territorial level) of the potential relationships between "institutions in a specific territory as perceived by entrepreneurs" and "the international expansion of the SMEs that they manage".
In the absence of a registry of such entrepreneurs in the municipalities under study, we resorted-as an approximation-to the registration of firms. The census provided by the Ministério do Desenvolvimento, Indústria e Comércio Exterior, Governo do Estado de Sergipe (Secretary of State for Economic Development and Science and Technology, Commercial Board of the State of Sergipe) was used and, according to the list provided, there were 1520 firms operating in the tourism industry in the six municipalities analysed. However, this census did not offer a reliable calculation of the firms due to two basic problems: the lack of an up-to-date registry that included business 'ups and downs', and the existence of a high level of informal entrepreneurship-that is, of active but not legally constituted businesses. In addition, on the lists consulted, there was no information-or it was unreliable-regarding the size of the firms or the subsector in which they operated (most of these fields were blank in the census). All these circumstances prevented the proper determination of the population size and, consequently, the definition of a sample size following statistical criteria. As an alternative, a fixed quota system was established and it was guaranteed that each municipality had a balanced representation in the sample (between 15 and 20%), thus enabling the subsequent comparative analysis.
We obtained the information through a structured questionnaire written in Portuguese (the official language of Brazil). To carry out the fieldwork, the SEBRAE institution (Serviço Brasileiro de Apolo as Micro e Pequenas Empresas) collaborated, selecting six interviewers and participating in their training (e.g. regarding the terms and the meaning of the items in the questionnaire) and coordination. Because SEBRAE is a well-recognised organisation that provides support to micro and small businesses in Sergipe, the entrepreneurs invited to participate in the study were very favourable towards collaborating in the data collection phase. As a result of this willingness, abstentions, or refusals were rather low, and we reached a satisfactory participation level of entrepreneurs in each municipality (response rate higher than 90%).
Specifically, we obtained a sample of 296 entrepreneurs (sampling error of ± 3.69% with respect to the 1520 tourism firms including the census of the Junta Comercial do Estado de Sergipe) with the following territorial distribution: São Cristovão, 17%; Aracaju, 23%; Pirambu, 14%; Laranjeiras, 14%; Stay, 15%; and Canindé, 19%. On average, the SMEs managed by their founders and participating in the study had 9.89 years of life and had a staff of 3.3 full-time employees. They operated in the following tourism subsectors: accommodation (19.26%); restoration (29.39%); transport (14.53%); recreational activities (12.16%); intermediary agents (3.04%); commercial (20.27%); and others (1.35%). Finally, 89 of the 296 SMEs in the sample were internationalised. Following Yiu et al. (2007) and Zahra et al. (2000), this variable was measured using a 7-point Likert scale, 7 indicating that the SME is actively engaged in international expansion according to its entrepreneur. Therefore, the variable indicates the intensity of the international business activities undertaken by the SME from the perspective of the entrepreneur managing it. Although methodologists have traditionally advised the use of measures based on multiple items, this point of view has been modified (Loo 2002). In fact, in the literature on international business, the use of a single item is frequent for the dependent variable (e.g. Hessels and Terjesen 2010;Ogbuehi and Longfellow 1994).

Institutional dimensions
The information regarding regulative, normative, and cognitive institutional dimensions was collected through 22 items that measured aspects of institutions on a 7-point Likert scale (where 7 means that the entrepreneur positively values the existence of such aspects of institutions in their municipality). Scales were built in coherence with the precepts of neo institutionalism. Specifically, in the design of the regulative and cognitive scales, we took into account Busenitz et al.'s (2000) scale, successfully replicated by Manolova et al. (2008) in three emerging economies, although their scales were designed to measure the institutional effect on the entrepreneurial orientation of the students. In addition, the scale proposed by Tominc and Rebernik (2007) was used as a reference to build the measurement of the normative institutional dimension and included cultural norms with a proven effect on high-growth business aspirations. The existence of the three institutional dimensions in our scale was confirmed through a factor analysis of principal components with Promax rotation (KMO = 0.763, χ2 = 1329.385***). The Promax rotation method was used because the present study is based on theoretical approaches that suggest the existence of relationships between institutional dimensions (Szyliowicz and Galvin 2010). Such relationships, if any exist, are hence incorporated into the estimates used in this research later on. Table 1 shows details of the scales, validity, and reliability assessments. Referring to validity, the assessment included several indicators. Specifically, we first considered the variance explained by the full model (61.536%) and by each factor in the factor analysis. Second, and following Hair et al. (1995), we verified that convergent validity was high through three tests: (1) the factor loadings on all items in the factor analyses are equal and above 0.6; (2) the average variances extracted (AVE) of all the constructs are equal and above 0.5; and (3) the composite reliability (CR) is over 0.7. Third, we assessed discriminant validity through the square roots of the AVE following Fornell and Larcker's (1981) criterion. Specifically, the square root of AVE for institutional dimensions is superior to the correlations between them, which implies discriminant validity is good (please see correlations in Table 2). Concerning the reliability of the scale, internal consistency was checked through Cronbach's alpha for all the items (alpha = 0.797) and each institutional dimension (see Table 1). As the scale for each institutional dimension is made up of less than 10 items, alphas equal to and above 0.6 can be considered acceptable (Loewenthal 2015).

Data analysis
Data analysis was carried out in three phases. First, an analysis of correlations was carried out in order to examine the possibility of multicollinearity. Likewise, and given that the research uses a single source of data-the SMEs' entrepreneur-and is cross-sectional, which could lead to a common method of variance, analyses were carried out in order to inform on this risk. For this purpose, the survey-taker was asked to ensure that the participants understood all the questions and, during the empirical study, the respondents were guaranteed total anonymity. In addition, after the construction of the database, a Harman factor test was carried out following the procedure indicated by other authors (Koropp et al. 2013;Li et al. 2007). The tests performed show that there are four factors in the data with eigenvalues greater than 1, regardless of whether a non-rotated factor analysis of principal components is used (total explained variance = 63.47%), a principal component analysis with varimax rotation (total explained variance = 63.47%), or a factor analysis of main axes with varimax rotation (total explained variance = 53.19%). The first extracted factor explains 26.99%, 19.93%, and 17.42% of the total variance, respectively. Therefore, the Harman test failed to identify a single or main factor underlying all the variables included in the present work, so there is no evidence of the existence of a common method of variance. Second, a mean difference test was carried out to determine the existence of statistically significant differences among the six municipalities analysed and compared in this study in terms of independent and dependent variables (institutional dimensions and the international expansion of SMEs, respectively). To determine whether there were statistically significant differences between the means, the F test from the one-factor ANOVA was run by using standardised values for each institutional dimension obtained from the factor analysis. To make comparisons between variables possible, we also used standardised values for the variable 'SMEs international expansion'. And third, a qualitative comparative analysis of the institutional configurations in each municipality and the international expansion of SMEs located in such territories was carried out. Results Table 2 shows the correlations among the variables under study, taking the participant entrepreneur as the unit of analysis. With respect to the multicollinearity of the data, the general rule is that the correlation should not be higher than 0.75. In our sample, the highest correlation index for the independent variables is observed between the regulative and cognitive institutional dimensions (0.433). These results suggest the absence of multicollinearity. In addition, it should be noted that the correlations identified show significant associations between the international expansion of the SMEs and the cognitive, regulative, and normative aspects of institutions that promote such expansion. Table 3 shows the mean difference test carried out to assess the existence of statistically significant differences among municipalities in terms of the institutional dimensions. Results indicate that the territories compared have significant differences in regulative, normative, and cognitive aspects of institutions. Firstly, while Aracaju and Estância present the most favourable regulative institutional dimensions for businesses according to the entrepreneurs located in these municipalities, Pirambu and Canindé stand out for having the worst. The existence of these differences is relevant because they affect institutional aspects such as the existence of state aid and regional funds, information offices on issues of relevance to the operation and expansion of the business, or the training offered in professions required for the development of the tourism SME, among others.

Subnational institutional configurations
Secondly, the results also find differences concerning the normative institutional dimension, which encompasses cultural aspects related to entrepreneurship such as the high level of respect for people who succeed with their own business, the existence of frequent conversations and news in local media concerning successful entrepreneurs, and the acceptance of female entrepreneurs. Looking at the mean comparison, we find that the territories of Laranjeiras and Estância stand out for having the most favourable normative institutional dimension, compared to Aracaju and São Cristovão, which are characterised as having those that provide a lower level of support for the development of successful and productive entrepreneurship. However, and thirdly, it is precisely these two municipalities-Aracaju and São Cristovão-that have the most suitable cognitive institutional dimension for facilitating the growth and expansion of the SMEs located in them. Cognitive aspects of institutions include the existence in the municipality of people with university degrees, computer skills, managerial experience, and experience in the industry in which the SME operates. Canindé and Pirambu, on the contrary, rank the worst concerning this institutional dimension.
Considering these results globally, we can affirm that each municipality has its own configuration or combination of institutional dimensions, in such a way that aspects of regulative, normative, and cognitive institutions are favourable to firms expanding at different degrees, such differences being perceived by the entrepreneurs who have their tourism SMEs located in them. Thus, results provide support for H1. Table 4 shows the mean difference test for municipalities of Sergipe with respect to the international expansion of SMEs. The test performed finds statistically significant differences between these territories. This result guarantees the existence of an adequate context in which to carry out the qualitative comparative analysis of institutional configurations and its relationship with the international expansion of SMEs put forward in proposition P1. Specifically, the average values for each municipality, shown in Tables 3 and 4, allow such an analysis to be carried out.

Subnational institutional configurations and internationalisation of the tourism SME: A territorial comparative analysis
Focusing on the three municipalities where the international expansion of SMEs is higher, Estância, Canindé, and Aracaju, we can observe the different institutional configurations that characterise each territory. In Estância, where the international expansion of SMEs is highest, the regulative and normative institutional dimensions are favourable to firms' expansive activity, both ranking second in the rankings of the six municipalities analysed, while the cognitive dimension ranks third. Thus, we found coherent formal and informal institutions that complement each other and support internationalisation.
Aracaju, being the third-ranked municipality in terms of the international expansion of SMEs, stands for having the most favourable regulative and cognitive aspects of institutions for expansive business activity, compared to the other municipalities-although its normative dimension ranks it at fifth. In this case, the formal regulative and informal cognitive institutions are the ones that are complementary

Subnational institutional configurations and international…
in supporting internationalisation; such a great and harmonised effect can in some way make up for the lack of a favourable culture (informal normative institutions) and allow Aracaju to be located in third place in the rankings of SME international expansion, albeit with great distance from Estância. Canindé, however, being the second municipality in terms of the international expansion of SMEs, ranks third regarding the normative dimension of institutions and the fifth/ sixth concerning the regulative and cognitive dimensions, respectively; so, complementarity in the suitability of formal and informal institutions to boost expansive business activity was not found, and normative aspects of institutions only moderately support expansive business activity: Canindé is much closer to Aracaju (the third) than to Estância (the first).
On the other extreme, Laranjeiras, São Cristovão, and Pirambu are the municipalities where the international expansion of SMEs is the lowest, and these municipalities are characterised by the absence of complementary formal and informal institutions that favour expansive business activity. São Cristovão and Pirambu have regulative and normative institutional dimensions that rank low in the rankings of the six municipalities, and the cognitive aspects of institutions are especially low in the case of Pirambu (it ranks fifth); in Laranjeiras the normative dimension is the most favourable for expansive business activity, but the regulative and cognitive dimensions are low and occupy the third and fourth position in the rankings.
All considered, the results seem to be in line with proposition P1, but they also offer some findings that invite a deeper discussion. We offer such a discussion below.

Discussion
Although institutional theory suggests that no form of government is superior to any other, generally speaking, our work shows that the lack of an institutional configuration that has regulative, normative and cognitive dimensions that support expansive business activity results in a lower level of international expansion of SMEs. The case of the São Cristovão and Pirambu municipalities, which both occupy the second worst place due to the low level of international expansion of their SMEs, reflects this finding (Fig. 1). Both municipalities suffer from low values in regard to regulative, normative and cognitive institutional dimensions. In these cases, we could point to the existence of a certain institutional coherence, although it does not appear to be conducive to the international expansion of SMEs.
Likewise, Cantwell et al. (2010) warn that there is a possibility that institutions will fail to achieve their desired effect-in this case, the internationalisation of SMEs-if those that are informal-i.e. normative and cognitive-cease to be congruent with those that are formal-i.e., regulative. This approach is evident in this study when making subnational territorial comparisons. The comparative analysis shows that in those municipalities where there is a dissociation between regulative and normative institutional dimensions, the international expansion of SMEs is much lower. An example of this is represented by the case of Laranjeiras, which is the municipality with the least amount of internationalisation of the SMEs participating in the study. Although it is the place with the strongest normative aspects of institutions, it lacks regulative and cognitive ones.
In Aracaju, meanwhile, despite the existence of regulative and cognitive aspects of institutions that are highly favourable to business expansion, the fact that these are not supported by the normative ones leads to a diminished-although not insignificant-international expansion of SMEs. On the opposite end of the spectrum, Estância stands out as being the territory in which the greatest degree of coherence between the regulative and normative aspects of institutions exists, giving rise to the highest levels of international expansion registered at the municipal level. This is despite the fact that its cognitive institutional dimension is not especially favourable to international activity. Canindé represents an exceptional case, with an alternative institutional configuration that gives rise to a certain level of internationalisation of SMEs, occupying the second place in the rankings of this study, although it is closer to the group of municipalities with the low international expansion than to Estância. Regarding its institutional configuration, it is based on the normative aspects of institutions that are moderately favourable to expansive business activity, albeit with low levels of the other institutional dimensions, which invites us to examine additional arguments that explain this particular level of international expansion of its SMEs. To this end, it seems appropriate to consider variables that describe the socioeconomic situation of the municipality in case they are helpful in contextualising and interpreting the results. Table 5 collects such secondary information. With respect to the export sector, there is no aggregate information on foreign direct 1 . The figures for the export of services do not include those strictly related to tourism, although there are subsectors related to this activity, such as transportation and personal, cultural, and recreational services. After examining several databases, we requested the information directly from Banco do Brasil, where we were informed that they did not perform such calculations for the municipal level.

3
Subnational institutional configurations and international…

Subnational institutional configurations and international…
Considering such contextual data, Canindé stands out for having an extensive territory (the largest one), although it is sparsely populated and very rural (it is the most rural of the six under study and is in second position regarding agricultural GVA). Despite enjoying the second highest GDP in absolute terms and the highest GDP per capita compared to the other municipalities analysed, it is also the municipality that has the greatest proportion of the population living in extreme poverty (21.74) and a relatively low rate of human development (0.567). Canindé, on the other hand, has few registered firms. These circumstances allow us to suggest the existence of an economy based on large firms that generate GDP and that, despite their business practices and experience being impervious to the general population of entrepreneurs (cognitive aspects of institutions are evaluated as less favourable), they somehow manage to spread and boost the international expansion of some SMEs. This is also encouraged by relatively favourable social values existing in the municipality. In addition, and considering Wu and Deng's (2020) conclusions, it could be possible that some SMEs try to fill institutional voids in their municipality by escaping to other countries without abandoning operations in their location at home; in other words, as they face a weak institutional framework in their municipality that limits opportunities to reach growth or even survive, SMEs look to internationalise. Certainly, this is an option mainly suitable for large firms, but SMEs nurtured with the necessary resources could also do that. In this respect, Cardoza et al. (2015) found that in the emerging economy of China, some SMEs expand internationally even when perceiving poor regulatory frameworks and weak support systems from the government and that SMEs are basing the international expansion on their capabilities rather than on support from the government.
On the other hand, with regard to Estância, it should be noted that the highest position concerning SME international expansion is consistent with the levels recorded for the export sector in this municipality. Specifically, a large number of exports are recorded, accumulating 49.85% of total exports made from Sergipe, as counted by the Brazilian Institute of Geography and Statistics of Brazil. In addition, imports, counted as the second most relevant activity, are thriving according to the set of economic indicators. This situation is similar to Aracaju, which occupies the third highest position both in the rankings of the internationalisation of SMEs and the exports registered in this municipality. According to socioeconomic data, this municipality stands out for being the biggest in terms of the total population, people enrolled in secondary school, the number of registered firms, and imports of US$ FOB. All of this can justify that, in this municipality, the cognitive aspects of institutions are the best assessed by entrepreneurs.
In the case of Laranjeiras, a municipality with weak regulative and cognitive institutional dimensions that favour expansive business activity, but the most favourable normative aspects of institutions, an incongruence has been identified between the international expansion of SMEs established in the municipality (the lowest) and registered export sector activity according to the secondary data. This incongruence may be related to the economic structure of this municipality and the industry from which we took the empirical evidence (i.e. tourism). Indeed, although Laranjeiras has important tourist attractions and is being promoted as a destination, boasting the second most important historical city in Brazil and qualified as a 'monument city' by the government and as a 'historic city' by the National Historic Patrimony, this sector has still 1 3 not achieved much relevance in the local economy. Specifically, its main economic activities at present are agriculture and industry. In the agricultural sector, sugarcane cultivation predominates-a product widely exported from Brazil-and in the industrial sector, the production of cement, which is exported to more than 20 countries in America, Africa, and Europe, is paramount. This circumstance may justify the unequal relevance found in our work regarding the exports registered according to FOBderived from the agricultural and industrial activities-and the international expansion of the SMEs that operate in the tourism industry found in this municipality.
Concerning the São Cristovão and Pirambu municipalities, which according to our study occupy the second worst position concerning the international expansion of their SMEs, congruence with the secondary data of registered exports has been identified, as exports are almost nonexistent in these municipalities. In addition, both territories share the lowest GDP current prices and lowest GVA industrial current prices. All of this seems to be coherent with their institutional configurations (the regulative, normative and cognitive institutional dimensions are to a low degree considered favourable to firms). Pirambu is also the smallest municipality according to the total population and the number of registered firms, as well as the territory with the lowest number of people enrolled in secondary school, which is coherent with the low evaluation of the cognitive aspects of institutions by entrepreneurs located in this municipality.

Conclusion
The present study has been based on the most recent versions of neo institutional theory and the literature on entrepreneurship and international business in order to contribute to the existing gaps in the literature on SME internationalisation in emerging economies (e.g. Chandra et al., 2020;Deng and Zhang 2018). Specifically, we answer the call concerning the need for further studies that identify an environment conducive to the internationalisation of SMEs in emerging economies (Deng and Zhang 2018), taking into consideration the interactions between formal and informal institutions at the subnational level (Chandra et al. 2020). In our answer to this call, we go beyond the study of single institutional dimensions and examine configurations of the regulative, normative, and cognitive dimensions of institutions; we also go beyond national comparisons and focus on a specific emerging economy, Brazil, in order to study and compare institutions at the subnational level. Thus, as highlighted by Pino et al. (2021), since previous studies on the internationalisation of SMEs in emerging economies mainly focuses on China or India, our study is of relevance concerning the advancement of knowledge specific to Latin America. In addition, as the few works carried out on Latin American SMEs have a narrow focus and address scattered areas (Cardoza et al. 2016;Felzensztein et al. 2022;Fornes et al. 2021), the present work may offer some relevant contributions as it considers the three dimensions of institutions and their combination in configurations, as detailed below.
First, we found that the effects of each institutional dimension on the international expansion of SMEs are contingent on the development and content of the other institutional dimensions with which it interacts. This configuration effect shows that the impact of the institutional dimensions on business internationalisation is not individualised, there being different combinations with an unequal degree of association with the levels of internationalisation. Specifically, and considering the results globally, our work illustrates how the coherence between the formal institutions-i.e., regulative aspects of institutions such as legislation approved by the governmental bodies that support expansive business activity-and the informal ones-i.e. normative aspects of institutions such as cultural values favourable to a high rate of growth in entrepreneurship-is necessary for boosting the international activity of SMEs located in each territory. In addition, we found that the relationship of these configurations with the internationalisation of SMEs may be conditioned by other socioeconomic variables, such as the industries that predominate in the territory or the factual location of large firms. Thus, unlike a good portion of the previous literature that analyses the individual effects of institutional dimensions on SME internationalisation, providing new, albeit partial, insights (e.g. Adomako et al. 2020), our work offers a more complete picture of how institutions operate to create an environment conducive to SME internationalisation.
Second, and compared to the previous literature, which has been mostly dominated by comparisons between countries, hence overlooking other levels of institutional influence (Donnelly and Manolova 2020), our study reveals the importance of the institutional configuration of territories at a lower level, within an emerging country, through its relationship with the international expansion of the SMEs located in those areas. These intra-country configurations were identified in the present study from the perceptions of entrepreneurs on the regulative, normative, and cognitive aspects of institutions; they also affect their decisions concerning internationalisation. Therefore, our study suggests the need to consider the sub-national level of institutional influence in the attempts to generate institutional environments conducive to SME internationalisation.
From the findings of this paper, practical implications can be derived. First, the formulation of specific policies and single support programmes, and their general application to the whole of a country, may be insufficient to promote the international expansion of SMEs if such policies and programmes do not consider the institutional configuration that prevails in each subnational territory. Our work also suggests that to promote the international expansion of SMEs in emerging economies, the following must be guaranteed: (1) the existence of public resources-e.g. state aid, regional funds, information offices that address issues of relevance to the expansion of ventures-available to the entrepreneur, (2) the availability of training opportunities to create a knowledgeable population to manage these activities for the firm, and (3) the promotion of a culture conducive to successful and expansive entrepreneurship and consistent with the local idiosyncrasies of each territory. Although culture, as an informal institution, is highly relevant in emerging economies, and previous studies highlight its strong impact on entrepreneurial activity in general, when addressing issues related to international expansion, our work finds that by itself-that is, in the absence of a regulative and cognitive framework that supports international activity-is insufficient. In addition, our results may be useful for entrepreneurs located in emerging economies who are interested in international expansion. They should be aware of the existing institutions in the territories in which they are located and how such institutions can condition their international aspirations; considering moving to other territories within the country with more favourable institutions for internationalisation, as well as engaging in bottom-up initiatives to modify institutions through business associations, are some of the possible options for them to contemplate. This last suggestion is coherent with Peng et al.'s (2017) view, who consider that in the current dynamic and complex environment, firms, and governments must not only adapt to the current institutional framework but also seek changes in institutions.
Finally, the generalisation of the results of this study is subject to some limitations. The first is in the study sample, since all SMEs come from a single country, a single federal state and a single industry. Although the variables country, state and industry are controlled and, therefore, make the analysis of the differences between the perceptions that each entrepreneur has about the institutions possible, the question arises as to whether the conclusions reached are relevant only in the context analysed. In addition, our research is based on a sample of SMEs registered in a public census that is not complete, so younger SMEs that are not yet registered or those from the informal sector have not been considered. Additionally, our work is oriented towards understanding the decision by the SME entrepreneur to internationalise, but other issues relevant to the understanding of the phenomenon, for example concerning the decision-making process of the international strategy (e.g. location choice, entry mode, etc.) and the subsequent international operations (e.g. the international performance of SMEs) have not been studied. Therefore, the literature would benefit from the efforts made in future research to correct these limitations.