Innovation systems and entrepreneurial ecosystems: Implications for policy and practice in Latin America

Innovation System (IS)

ISs, alongside the concept of industrial policy, draw from a long line of international contributors, including Alexander Hamilton (1791) in the USA and Friedrich List (1841/1928) in Germany, and more recently, Briton Chris Freeman and Swedish-Dane Bengt-Åke Lundvall, who are recognised as co-founders of the concept of national IS (NIS) in the late 1980s (Lundvall, 2007).

Through his research on the ‘national business of innovation’ developed by Whitley (1994) in the UK, Lundvall argued that ‘theoretical work in academia as well as innovative efforts in firms often follow parallel tracks and that quite similar new ideas, concepts, and products emerge almost simultaneously’ (60). At the time, he differentiated the business-systems approach, which emphasised the ‘combination of elements’, into different patterns, while the IS approach emphasised the ‘process where agents interact’.

Since then, the concept of IS has become popular among policymakers and international entities such as the Organization for Economic Cooperation and Development. However, even though the IS approach is recognised as an important paradigm, only six out of the European Union’s 28 countries have developed system-oriented innovation policy evaluation practices (Borrás and Laatsit, 2019).

An extensive IS literature is characterised by distinctions between a narrow and a broad view of innovation. As Johnson suggest, as the term ‘innovation’ expanded to ISs, some definitions focused too narrowly on science-based innovation, while others took a broader range of experience-based innovation into account. The narrow view refers to high-technology products and processes of research and development (R&D). Scholars of the narrow view (e.g. Nelson, 1993) emphasise the relationships between firms and research institutions. Some innovation scholars have identified a bias among scholars and policymakers that consider innovation processes largely as aspects connected to formal processes of R&D, especially in science-based industries (Jensen et al., 2007).

The broad view of IS has been championed by Schumpeter (1912), particularly in his latter writings (1942, 1949) and by subsequent neo-Schumpeterian scholars and includes ‘all parts and aspects of the economic structure and the institutional set-up affecting learning as well as searching and exploring’ (Lundvall, 1992: 12).

This broad view has been the genesis of a more distinctive, NIS perspective, defined as ‘an open, evolving and complex system that encompasses relationships within and between organisations, institutions and socioeconomic structures which determine the rate and direction of innovation and competence-building emanating from processes of science-based and experience-based learning’.

This broader definition includes user–producer interaction and interactive learning processes taking place inside firms, including ‘learning by doing’, and not only technology-science R&D (Jensen et al., 2007; Lundvall et al., 2011). This distinction is highlighted in the seminal paper by Jensen et al. (2007). On one hand, their work identifies modes of innovation (Science, Technology, and Innovation, STI-mode) that emphasise the promotion of R&D and create access to explicit codified knowledge; and on the other hand, the Doing, Using, and Interacting (DUI) mode of innovation that emphasises learning based on tacit knowledge.

Since the 1980s, much of the IS literature has focused on the economic dimensions of innovation, with further attempts to establish the central role of institutions, social dimensions of the systems, and the context of historical path-dependencies of countries and regions (e.g. Freeman, 2002; Lundvall, 1992). A regional perspective on innovation (RIS), popularised by Cooke (1992) and others, places emphasis on economic and social interactions among agents, spanning the public and private sectors to engender and diffuse innovation within regions embedded in wider national and global systems (Asheim et al., 2011). Knowledge is seen as central to RIS, with the geographical dimensions of knowledge transfer identified as a key variable shaping regional innovation performance. As Asheim et al. (2011) argue, the long-term development of regions depends on their ability to diversify into new applications and new sectors while building on their current knowledge base and competences.

Learning processes are fundamental to the broader view of IS, with innovation relying on a systemic transformational process in order to produce economic benefits. However, this transformation cannot be achieved without the combination of existing and new learning processes to produce knowledge as the main element towards building technical and economical capacity into the system (Cassiolato et al., 2013; Lundvall, 1992). Transformation also benefits from knowledge spillovers arising from interactions among established firms, research institutions, and new ventures that stimulate innovation and provide competitive advantage (Audretsch and Lehmann, 2005).

A broader view of IS could also include social as well as technological innovations. Lundvall (1988) proposed a ‘welfare-industrial complex oriented towards the fulfilment of social needs in relation to energy, housing, environment, transport, and the health service [which with the support of the state could] be as effective as warfare-industrial complexes in provoking new technology’ (366). Lundvall’s notion that efforts on social innovations are important for the wealth of nations gained particular prominence in the Scandinavian countries.

Despite its broad appeal, the IS concept has been criticised for lack of a validated, unifying theory or commonly accepted definition. A number of scholars have criticised the blurriness and constant change of the IS concept, and it has been even criticised as a buzzword (Doloreux and Parto, 2005). Much of the empirical work on RIS has been based on well-functioning, successful regional economies, and on innovation in high-technology sectors (Asheim et al., 2011). Such studies often fail to account for differences that exist within the various types of organisations, the variety of knowledge links and interactive learning within these organisations, and the significant differences between the institutional and social contexts in different regions (Doloreux and Porto Gomez, 2017).

The conditions that enable growth to accrue in regions where innovation occurs have also been neglected in the RIS literature, where it is more commonly assumed that the conditions conducive to innovation will automatically lead to growth. Doloreux and Porto Gomez (2017) call for further study to determine the extent to which local innovation is associated with regional economic growth and the transformation and mutations of RISs.

Finally, Altenburg (2009) identifies the many challenges facing IS in underdeveloped economies in suggesting a more effective model. He argues that ‘while neoclassical economists mainly blame labour market segregation and administrative entry barriers, structuralists emphasise multiple market failures in financial, educational and information markets as the main reason’ (49).

Overall, Altenburg considers that the main challenge ‘is not primarily to develop “new to the world” innovations, but the development and broad dissemination of affordable and adapted technologies’. Altenburg calls for further attention and efforts on developing inclusive innovation and its diffusion – innovation that focuses on issues such as access to water and food, promoting interaction among firms that often have a lack of trust among them, etc. A ‘broadening out’ of the IS concept, to include education and financing, can be observed in China (Lin, 2014), and China’s rise as an economic power has made its modus operandi attractive to many developing countries in Latin America.

Entrepreneurial Ecosystem (EE)

EE have attracted increasing attention from scholars and policymakers in recent years (Alvaleden and Boschma, 2017), with the concept drawing from a number of different models and theories (Scaringella and Radziwon, 2018). The notion of ‘ecosystem’ draws on a branch of the biological sciences, which investigate the relationship between organisms and their environment (Aldrich, 1990). Shumpeter (1949) earlier highlighted the importance of environment and contributions by multiple actors to the entrepreneurial process, where ‘environment, public authority, corporate action and individual initiative must have co-operated’ (264).

Malecki (2018) places the beginnings of EE in the early 1990s, identifying contributions from Moore and his ‘business ecosystem’ ideas, Van de Ven’s ‘infrastructure for entrepreneurship’, ‘high‐technology entrepreneurship’, and Spilling’s concept of a local ‘entrepreneurial system’. Malecki (2018) shows that usage of the term EE has overtaken other concepts, such as environments for entrepreneurship, which also highlight the mechanisms, institutions, networks, and cultures that support entrepreneurs.

The EE concept is characterised by different elements or attributes supporting entrepreneurial activity. Bruns et al. (2017) define an EE as a multidimensional set of interacting factors that moderate the effect of entrepreneurial activity on economic growth. Spigel (2017) focuses on elements that develop simultaneously and reinforce each other: ‘an ecosystem’s attributes are sustained and reproduced through their relationships with other attributes’ (8). These relationships may have different densities; in a low-density EE, for instance, one attribute may be more dominant and drive others, such as a strong local market supporting the inflow of opportunities, whereas in a high-density EE, the elements support each other in a more balanced and stronger manner. Isenberg (2011) indicates that elements of the EE interact in complex and specific ways leading to unique configurations of different EEs.

The EE literature identifies the crucial role of local networks that enable entrepreneurs to get access to resources that include knowledge, finance, and human capital (Stuart and Sorenson, 2005). Spigel (2017) refers to the ‘presence of social networks that connect entrepreneurs, advisors, investors, and workers, and that allow the free flow of knowledge and skills’ (8). Similarly, Feld (2012) refers to the importance of interactions in a successful start-up community and a high network density among actors and groups of actors, where everyone is willing to contribute to the ecosystem.

Institutions are regarded as crucial ingredients for entrepreneurial networks (Johannisson et al., 2002), and institutions may enable or constrain interactions across individuals, firms, and other organisations (Huggins et al., 2012). Institutions may be more supportive in some regions than in others, and therefore may lead to different rates and types of entrepreneurial activities and different regional development paths (Gertler, 2010).

Cooke (2016) suggests that EEs are not only advantaged by their networking propensity but also by their regional ‘communities of practice’ and the free goods of interactive learning from ecosystem peers whether firms or intermediaries. Cooke also suggests that EEs offer the potential for the accumulation of social value and associated economic efficiencies that can be superior to the traditional hegemonic model of individualistic ‘property rights’ entrepreneurship.

The EE literature highlights the critical importance of place, where different interconnected actors and factors such as human capital, networks, and institutions are organised spatially and geographically concentrated (Johannisson, 2000; Qian et al., 2013). Studies on EE have often examined successful high-tech places in different countries to determine elements contributing to their success (Feld, 2012; Mack and Mayer, 2016; Spigel, 2017). Literature also suggests a link between EE and clusters in providing opportunities for entrepreneurship, such as a specialised labour market and geographically localised knowledge spillovers (Delgado et al., 2010). Some scholars emphasise the importance of global connections between different EEs (Malecki, 2011), while others have emphasised the role of multinationals in attracting skilled labour into a region (Mason and Brown, 2014).

Despite its recent popularity, a number of weaknesses have been identified with the EE concept, limiting its application to policy, practice, and theory (Alvedalen and Boschma, 2017; Autio et al., 2014; Spigel, 2017; Stam, 2015). Malecki (2018) cautions that ecosystems are defined in very different ways, at different scales, and with different research designs and data, while Ahokangas et al. (2018) identify the challenge of figuring out where an ecosystem starts and where it ends. Motoyama and Watkins (2014) point out the problem of identifying ‘system elements’ without giving proper attention to the connections between them, and treating all elements as equally important. Similarly, Alvedalen and Boschma (2017) identify gaps in understanding how different EE sub-networks connect to each other, to what extent they overlap, and the implication of overlap for the performance of individual entrepreneurs and of the EE as a whole.

EE literature also suggests that EEs, including successful ones, underestimate the influence of entrepreneurial traits and tend to discriminate against women regarding access to opportunities or to capital (McAdam et al., 2019). Other scholars suggest a bias towards high-impact, ambitious entrepreneurship found in developed economies that may neglect low or non-growth entrepreneurship and self-employment initiatives more common in less developed economies, like in Latin America (Acs et al., 2014; Stam, 2015; Szerb et al., 2019).

While differences in levels of entrepreneurship are well recognised across regions, use of the EE concept to explain such differences is subject to debate. Successful EE requires places where vested interests are not well represented or unable to dominate local institutions (Alvedalen and Boschma, 2017; Boschma, 2017). Cooke (2019) identifies a number of ‘downsides’ of entrepreneurship, such as vested interests, fraud, corruption, and multi-client incompetence that may result in a ‘bad’ EE. Earlier, Baumol suggested that entrepreneurship can sometimes be unproductive or even destructive, as it can be associated with rent-seeking enterprises, often based on corruption schemes, as well as enterprises based on the destruction of the environment.

The EE literature has been further criticised for applying a static framework that describes relations in EE without considering their evolution over time (Alvedalen and Boschma, 2017). Different elements of EEs will shift in importance and in their connections as they evolve (Mack and Mayer, 2016). Institutions also tend to be treated as static entities, which ignores how institutions emerge, change, and may lose their relevance (Hwang and Powell, 2005). Sotarauta and Pulkkinen (2011) identify the need to analyse the encounters of entrepreneurs and institutions to better understand the ways in which actors may change the very institutions that govern their own activities, how institutional transformation occurs in some regions but not in others, and the implications of institutional transformation.

Various scholars call for an evolutionary perspective that would allow for comparing different EEs over time and identifying the mechanisms, connections, and activities that are more or less important to how they perform (Alvedalen and Boschma, 2017; Cooke, 2016; Mason and Brown, 2014). At the same time, Bruns et al. (2017) suggest that the ‘quality’ of an EE, given its multidimensionality, is difficult to measure directly.

Finally, Stam and Spigel suggest that EE is still in search of a clear analytical framework that makes explicit key actors and distinguishes cause from effect in determining critical success factors. Spigel (2017: 49) cautions that EE should be considered more of a ‘conceptual umbrella’ for the things, within a region, that promote entrepreneurship and new venture growth rather than a cohesive theory.

Comparing the IS and EE concepts

A central question arising from the above discussion is the extent of similarities and differences between IS and EE concepts and their significance. Various scholars suggest that the IS literature, with its primary focus on organisations and institutions, has not made explicit links with entrepreneurship (Acs et al., 2014; Landström et al., 2015). Moreover, a system approach to entrepreneurship has not been taken up systematically in the IS literature (Qian et al., 2013). As stated by Acs et al. (2014): ‘it is perhaps a little surprising, if not even ironic, that although the National Innovation System literature was heavily influenced by the Schumpeterian tradition, the entrepreneur remained conspicuously absent in this literature’ (477–478). This view is made explicit in the definition of National System of Entrepreneurship as ‘…the dynamic, institutionally embedded interaction between entrepreneurial attitudes, ability, and aspirations, by individuals, which drives the allocation of resources through the creation and operation of new ventures’ (Acs et al., 2014: 479).

Spigel and Harrison (2018) suggest that the EE concept is superior to IS in analysing new venture creation and thereby promoting local and regional development, highlighting key differences between the concepts in the role of the state, accessing resources, the role of knowledge, key actors, and industry. These authors also suggest that the role of knowledge in IS appears to be more limited, e.g. confined to ‘knowledge producers like universities or research labs’. However, while the state is characterised as the main IS actor and the entrepreneur as the main EE actor (e.g. Stam and Spigel, 2016), this distinction may be relevant for an RIS perspective (e.g. Cooke et al., 1997) or some ‘narrow’ models of IS, but it does not appear to match much of the IS literature.

Further, while a common theme among contributing perspectives of EE is the emphasis on environment, the importance of place, and the dynamics of context in understanding socioeconomic activity, this is also highlighted in the IS literature. IS scholars recognise the geographical and cultural closeness necessary to facilitate interaction in addition to a common cultural and educational background, but they tend to recognise the firm rather than the individual entrepreneur as the force behind innovation, and even placed state entities, including universities and research labs, as secondary actors (Jensen et al., 2007; Lundvall, 1988, 1992; Lundvall et al., 2011; Nelson, 2013).

Taking a systemic view, the EE literature tends to move in the direction of the IS literature that investigates how networks of actors are involved in the generation, diffusion, and use of innovations, and how institutions influence these interaction patterns (Qian et al., 2013). Entrepreneurship is more likely to resemble a network and not an individual. As Schumpeter (1949) concluded: ‘In many cases (…) it is difficult or even impossible to name an individual that acts as the entrepreneur’ (261).

Knowledge is identified as important to both IS and EE perspectives. The IS approach is characterised by technical innovation and the production, transfer, and exploitation of knowledge, being it new or incremental from prior discoveries. Knowledge is naturally inscribed in the centre of all R&D, institutional development, market conditions, channels of distribution, and in a broader sense, in all the operation and dynamics of the system (Velasco and Gregson, 2012). At the same time, knowledge is inherent within EEs, which are characterised by the free flow and exchange of knowledge and interactive learning between entrepreneurs, advisors, investors, etc. (Collinson and Gregson, 2003; Cooke, 2016; Huggins et al., 2012).

Although different types of knowledge may be more or less prevalent within an IS or EE, learning is also important to both. As Lastres et al. (2003) states, ‘…the emphasis is in the capacity to learn (as well as to forget) and to innovate as being crucial to the productivity and competitiveness of economic agents, rather than on the capacity to acquire and use new technical means’ (3). From a policy perspective, this suggests that if a territorial entity (e.g. region or country) does not have the capabilities to absorb, appropriate, reproduce, and introduce learning into the productive sector to generate competitiveness in their industries, then IS (or EE) policies may not be successful, as market dynamics make it more difficult to compete in global markets (Lastres et al., 2003; Velasco and Gregson, 2012).

As discussed earlier, while both IS and EE concepts appear to lack a validated, unifying theory or commonly accepted definition, some scholars are comfortable with these loosely defined concepts. One of the most prominent IS scholars, Lundvall (2007), even criticised ‘attempts to make the [IS] concept “more rigorous” through organizing the definition and analysis around a list of “functions”, “factors” and “activities”’ (1). Attempts to classify IS and EE into different theoretical models actually reveal similarities between the two concepts, which will now be discussed.

Based on Schumpeter’s contributions, Freeman proposed categorising his definition of entrepreneur into two different theoretical models (Andersen, 2009) known as Mark I and Mark II, a categorisation popularised by early 20th century British military tanks. Mark I considers established firms to be limited in order to focus on the single entrepreneur who is coming up with a relatively new product/service and creating a new firm. Mark I focuses on the single entrepreneur who develops an innovation by creating a new firm or business, which is the most commonly accepted definition of entrepreneurship (e.g. Gartner et al., 2004; Shane, 2003). Mark II, on the other hand, focuses on established firms. Schumpeter argues that innovation is ‘not any more embodied typically in new firms, but goes on, within the big units now existing [within large companies]’ (70), where these firms perform repeated incremental and radical innovations in which the intrapreneur plays a key role. In addition to Schumpeter, other scholars have emphasised the importance of recognising that most innovation comes from established companies, not new entrepreneurs (Drucker, 2014; Lundvall et al., 2002). Intrapreneurship, or Mark II entrepreneurship, can also be related to oligopoly dynamics (Andersen, 2013). Research also shows that only a small proportion of start-ups and young businesses are responsible for economic growth, job creation, or increased productivity (Acs et al., 2016; Stam, 2015).

Following these categories, Andersen (2009) also differentiates between Mark III, which considers entrepreneurship in the public sector and within public institutions, and Mark IV, which broadly considers entrepreneurship occurring within the entire system or environment. The concept of entrepreneurship in Mark III and IV borders on and diffuses with the concept of the ecosystem, thereby entering the realms of EE research (Andersen, 2013).

Figure 1 summarises the previous discussion, which identifies the three different classifications of entrepreneurship (Mark I, II, III). This cumulative entrepreneurial activity, when combined with the study of entrepreneurship as a system (Mark IV), relates to two approaches to IS. The first approach identifies IS in a narrow sense, which relates to innovation processes, new products and services, and processes of R&D. The second approach identifies IS in a broad sense, which considers relationships within and between organisations, institutions, and socioeconomic structures which determine the rate and direction of innovation and competence-building.

OPEN IN VIEWER

Figure 1.

Intersection between entrepreneurship and EE and IS. Source: Authors’ elaboration. EE: entrepreneurial ecosystem; IS: innovation system.

Jensen et al. (2007) found empirical evidence that firms using mixed strategies that combine a strong version of the narrow STI-mode with a strong version of the broader DUI-mode are those that excel in product innovation. In other words, the knowledge created in universities and research labs would adhere to the STI-mode of innovation, but the IS literature also acknowledges the DUI-mode, which shares many of the traits of mainstream EE research.

In practice, much activity within an EE will be less than high growth, innovation-intensive venturing that produces disproportionate economic outcomes (Acs et al., 2014; Szerb et al., 2019), which is often equated with hitting the STI jackpot (Cooke, 2016). Instead, it will be incremental, problem-solving DUI innovation that accounts for the greater part of social value arising from an EE setting (Cooke, 2016).

Background

Efforts of designing and implementing policies for scientific and technological development are not new in Latin America (Cimoli et al., 2006) and a bias towards high-tech, science-based innovation (STI-mode of innovation) or a narrow IS approach has historical roots in a number of Latin American countries. Arocena and Sutz (2000) identified almost two decades ago that the IS concept in Latin America is weak, mainly due to the historical neglect of technical innovation, resulting in ‘holes’ in the institutional fabric and the absence of a rich, diversified, and specific institutional setting directed to support innovation at all levels.

Many Latin American countries during the second half of the 20th century attempted to create their national systems of science, technology, and innovation (ST&I), following a more practical than theoretical perspective, i.e. evidenced by a strong belief that scientific progress would automatically turn into technological innovation (Cimoli et al., 2006) and the assumption that knowledge was a public good, i.e. non-rival and non-excludable in consumption (Velasco and Gregson, 2012). Government and public agencies became natural knowledge providers, with innovation expected to develop and flow from government and public institutions (supply-side) to production (demand-side). Public funds were the major source of ST&I financing, with approximately 80% of regional ST&I expenditure in the 1980s and 1990s financed by the State (ECLAC, 2004).

The public sector and public research institutions also set ST&I priorities and resource allocations. According to Cimoli, et al. (2006), ST&I institutions were run under a hierarchical, non-flexible, and pyramidal managing style that made it difficult, if not impossible, to cope with and respond to the dynamics of private sector knowledge and technological requirements. Imports of capital goods and technology licensing were seen as the basic sources of technological upgrading, with R&D activities mainly carried out by large public enterprises operating in strategic sectors such as transport and telecommunications, and by public research institutes and universities active in selective industries such as energy, agriculture, mining, forestry, and aeronautics (Cimoli et al., 2006; ECLAC, 2004). Innovation related institutions came to be regarded as ‘markets’ for trading or exchanging information more than as part of an articulated and flexible system through which know-how, codified and non-codified knowledge embodied in routines, production processes, or research results are transferred (Fagiolo and Dosi, 2003).

From the mid-1990s, policies to support the creation of research consortia, science parks, and public–private partnership for research were designed, although seldom implemented, in almost all Latin American countries (Bielschowsky, 2016). This shift in the ST&I policy discourse, from market incentives towards the role of networking and linkages between public and private agents, echoed the political practices of the north. The relevance of networks increased as NISs faced global incentives and pressures in the development of local organisation and production processes (Cimoli et al., 2006; 2016).

Although these policies attempted to stimulate interaction and cooperation between public and private sectors, it reflected a neutral ‘soft’ policy approach, with limited industrial prioritisation and missing linkages between and across industry sectors (Cimoli et al., 2006). On average, the Latin American production pattern induced private sector and enterprises to express what amounted to a low demand for knowledge on the one hand, and on the other, led domestic agents to mostly seek outwardly oriented linkages, privileging foreign companies and research laboratories that already have a sound reputation and worldwide recognised experience in effective and efficient ST&I efforts. Thus, a mismatch ensued between demand-side needs and supply-side offerings, hampering the impact of ST&I policies (Cimoli et al., 2006, 2016). This mismatch included knowledge and products to generate innovation, as pure supply-side or demand-side oriented technology policies are clearly not enough to stimulate innovation where production and management are increasingly carried out within networks (ECLAC, 2004).

A global perspective on supply–demand relationships and international market fluctuations needs to be considered in the construction of effective national policy frameworks. Next, it is important to use global dynamics to nurture local ones (Velasco and Gregson, 2012). In this sense, contextual policies should be carried out in all levels of a national system in order to have innovation, ‘… a government policy for glocalisation of technology is necessary in order to achieve the objective of industrialization; and this must be done at sub-national, national and supranational levels’.

Popular ideas on the ‘narrow sense’ of innovation continued to be adopted across Latin America. Paradoxically, when the IS concept was first developed in the 1980s by Lundvall, Nelson, and their collaborators, it was referred to system promotion as a suggestion for what third-world countries could do to develop their economies (Lundvall et al., 2002). However, social and political problems faced by a number of Latin American countries have hampered the achievement of explicit goals set out by government ST&I agencies. As a result, many countries had industrial sectors with low levels of technological culture, which translated into low levels of investment in R&D and ST&I activities, and a diversification of policies which had earlier failed to tackle the needs of ST&I activities according to internal and external supply and demand (Alcorta and Peres, 1998).

The result has been Latin America falling behind the technological frontier both in terms of technological specialisation and productivity growth. Evidence has shown that overall, Latin American countries concentrate on stagnant technologies, compared to Asian countries, which make greater accomplishments in dynamic technologies (Urraca-Ruiz, 2019). In contrast, Asian economies like Korea and Taiwan, followed by Malaysia, Thailand, and Indonesia, were successful in creating expanding capabilities in technology-intensive industries and production stages, combining selective import substitution policies with aggressive yet gradual export-oriented strategies (Cimoli et al., 2006). In catching up, some of these countries shifted to intensive patterns based on learning and knowledge, with enterprises adopting more capital-intensive processes and increased efficiency in production through incremental innovations and imitative practices.

Overall in Latin America, production structure transformation and technology policy have resulted in limited pervasiveness and diffusion of knowledge and intangibles in regional production systems, and the existence of pockets of technological excellence with a generalised low-tech and low-skilled jobs production apparatus and limited demand for advanced knowledge (Cimoli et al., 2006, 2016). Latin American IS is also characterised by weak links between the academic and research sector and the industry sector, directly affecting critical knowledge flows and spillovers required to successfully innovate (Lastres et al., 2003; Velasco and Gregson, 2012). This reduced capability to absorb, appropriate, reproduce, and introduce learning into the productive sector is translated into late insertion of technology and high costs to acquire infrastructure, and, as a result, high uncertainty about the costs and benefits of research and its outputs and low appropriation of new knowledge by the private sector that fuels industry competitiveness.

Turning to EEs, the EE concept remains undeveloped in a Latin American context with very few studies on entrepreneurship or specifically on EE, compared with IS-related studies (Alvarez and Grazzi, 2018). As suggested earlier, successful EE requires places where vested interests, corruption, and fraud are not well represented (Alvedalen and Boschma, 2017; Boschma, 2017; Cooke, 2016). Unfortunately, such schemes based on political clientelism and even organised crime have a higher rate in Latin America. Local as well as transnational elites are able to obtain political and economic power and dominate institutional settings and entrepreneurial activities (Franz, 2018).

While social and political problems have hampered the achievement of IS policy goals across many Latin American countries, there is a growing recognition of the potential of social innovation and entrepreneurship in Latin America (Domanski et al., 2017). Public transportation in Latin America could greatly benefit from innovation and new entrepreneurial approaches to urban planning. Public transportation is used by 39% of citizens in Latin America on a daily basis, compared to 23% of Europeans or less than 10% of North Americans (Daude et al., 2017). However, local planning often caters to the interest of the elites though land allocation designed to obtain real estate gains based on corruption schemes, while transportation is designed in favour of the more affluent car users (Freire-Gibb and Tapia Carrillo, 2019).

The same holds true for education, which is recognised as an important factor to enhance economic development, and while most children in Latin American countries are now enrolled in education institutions, the quality of education for the majority is very low, with large class sizes per teacher and few resources (Ferreyra et al., 2017). At the level of higher education, vocational education is also of very low quality and the public universities, which are more accessible for the majority, tend to have very little autonomy and often controlled by the interests of the elites.

Latin America also has a significant ‘informal economy’, estimated to encompass nearly 50% of jobs and 30% of GDP in Latin America, which is a larger proportion of the economy than that of sub-Saharan Africa (Casabón, 2017). Salinas et al. (2018) suggest that informal entrepreneurial activity is more abundant in Latin American countries that have lower tax morale. A number of innovation scholars identify the need for more effort to transform the historical, so-called informal economy into a more formal economy in an inclusive way so that it is further beneficial to society (e.g. Altenburg, 2009; Lundvall, 1988; Neuwirth, 2012).

The 2018 Global Entrepreneurship Monitor (GEM) report, which examines entrepreneurship in 54 economies, found that Latin America has the weakest institutional framework conditions to support entrepreneurship compared to the other four regions: Africa, Asia/Oceania, Europe, and North America (Bosma and Kelly, 2018). Framework conditions are crucial in determining the technological behaviour of firms through networks; enterprises never act alone and perform in a setting where they are expected to interact and maintain channels of communications on a continual basis with other economic and non-economic agents like other enterprises, universities, public institutions, and non-governmental and civil society organisations.

At the same time, there are positive signs for entrepreneurship in Latin America. The 2018 GEM report found that Latin America stands out for its low fear of failure rate, with less than one-third of those seeing entrepreneurial opportunities stating that fear of failure would prevent them from starting a business. Some countries, such as Chile, are also identified with high early stage entrepreneurial activity, with one-fourth of the population engaged in starting or running a new business. Having said that, researchers have also pointed out that the ‘excessive political and economic centralisation’ around the capital of Chile could hinder the EEs of the nation (Espinoza et al., 2019).

Aboal and Veneri (2016), in their study of entrepreneurs across nine Latin American countries, found a surprising level of potential entrepreneurs working in larger firms. The authors suggest policy actions that could impact their predisposition to become entrepreneurs and create new firms that include reducing the risk of failure.

Overall, evidence suggests that entrepreneurship has high potential in Latin America, but multiple challenges need to be addressed in establishing supportive EEs. Recent research suggests that for regions with a poor EE, policymakers may foster regional performance by redirecting resources to innovation-driven entrepreneurship (Szerb et al., 2019), given evidence that it is associated with superior regional performance. The authors of the study suggest that innovation-driven entrepreneurship may act as a substitute for the shortage of an appropriate EE. However, these findings are based on a comparison of EU regions, and a policy to direct resources to favour high-potential entrepreneurship may not be an appropriate recommendation for Latin America or other less developed regions.

Discussion of IS and EE concepts for Latin America

Previous discussion has suggested that IS and EE ‘systems’ are enabled or limited by particular capabilities within a defined territory (region, country, etc.) arising from configurations and combinations of different system components. The configuration and extent of external linkages beyond the territory will contribute to this uniqueness. This suggests that there is no EE or IS model for Latin American policymakers to emulate, since there is a large variety of regional landscapes, resulting from different national and local histories, institutional and social contexts, political logics and functions. Policy goals, instruments, and capabilities need to be tailored to country and regional specific contexts.

Latin America has different factors and attributes that challenge the application of a narrow view of IS, which emphasises high-technology products and processes of R&D. These factors include infrastructure deficiencies, weak and politically influenced institutions, and low levels of articulation and engagement between the main productive actors of the IS. Unlike most Western economies, where institutions play a critical role in providing economic stability and regulatory boundaries, such institutions are under much more political influence in most Latin American economies. Therefore, changes in the central government can have an immediate effect on the stability of IS (or EE-related) policies. Policy failure may result due to the great level of detail being generated when trying to regulate the insertion of different policies into the national system (Alcorta and Peres, 1998; Prebisch, 2016).

At the same time, the advantages of a broad view of IS for Latin America should not be overstated, as it is important to acknowledge that some sub-national areas in Latin America are very distinctive – culturally, democratically, and even economically – while others are less so, having weak cultural markers, centralised administration, and disarticulated economic activities (neither specialised nor diversified) (Cooke, 2016). It is also important to recognise relevant differences across cities in a region. For example, in the case of Ecuador, weak local institutions are the main reason for the city region of Guayaquil’s relative economic and demographic decline in relation to the city region of Quito. More inclusive local governance is a key to more equitable development, which can go in hand with a more effective and inclusive IS (Freire-Gibb and Tapia Carrillo, 2019). Adopting a broader view of IS to include social as well as technological innovation (Altenburg, 2009; Lundvall, 1988), and of EE, to include social value accumulation as an alternative to profit-seeking entrepreneurship (Cooke, 2016) appear highly relevant for Latin America.

As suggested earlier, EEs will be affected by elements that are defined within a territorial IS system. In the case of Latin America, EEs will be affected by limited demand for and diffusion of knowledge and intangibles in regional production systems and by the co-existence of islands of technological excellence with a generalised low-tech and low-skilled jobs production apparatus (Cimoli et al., 2006; 2016). This is because entrepreneurial opportunities available in different sectors depend on their respective sectorial dynamics and reflect the learning processes associated with the spread of technological paradigms.

It is evident that entrepreneurship per se has a number of challenges in Latin America, which lacks supportive framework conditions and where some entrepreneurial endeavours are bad for society and often involve corruption. Furthermore, much entrepreneurial activity is blurred with the informal economy, and as found in other places. The entrepreneurship literature suggests that entrepreneurship is more likely to resemble a network and not an individual (Schumpeter, 1949). However, there is a common perception in Latin America of the lone, hard-working entrepreneur who becomes rich, with evidence that Latin America has the lowest level of trust in public institutions and trust in peers compared to other regions globally (Halpern, 2005).

Overall, much of the EE literature has many limitations for Latin America, as it takes for granted features of an advanced economy, while the IS literature favours studies of well-functioning economies and of innovation in technology-based sectors. These factors are not typical in most of Latin America, which is characterised by high levels of poverty, inequality, and large informal economies. The EE concept is also based primarily on ‘opportunity-driven’ entrepreneurs, which, with a few metropolitan exceptions, this is not widespread in Latin America. Higher rates of ‘necessity-driven’ entrepreneurship are identified in Latin America compared to the rest of the world (Acs and Amorós, 2008), with necessity-driven entrepreneurs less inclined to start growth-oriented businesses (Bosma and Kelly, 2018).

While vested interests, the dominance of elites, and social inequality remain challenges to functional and effective EEs, the GEM study and other research findings suggest positive signs for entrepreneurship in Latin American and identify particular framework conditions that can improve support for entrepreneurial activity. Recent research identifying the potential of institutional entrepreneurship in Latin America (e.g. Mark II component of an EE) suggests further attention at the policy level for creating and adapting institutions and framework conditions to leverage intrapreneurial talent and provide more opportunities for new venturing and knowledge spillovers.