Social Capital and Innovation in East Asia

According to Coase (1960), social trust—generalized trust or trusting broad ranges of other people—can promote innovation because it reduces transaction costs among economic players. This argument is further elaborated by the school of social capital (Putnam 2000, Fukuyama 1995). In explaining the role of social trust, Fukuyama (1995) introduced the concept of high- and low-trust societies, distinguished by the extent of trust clusters. In high-trust environments, trust-based relationships exist between a large number of social actors so that individuals with a broad spectrum of backgrounds can socialize and cooperate with one another. On the other hand, in low-trust milieus, trust remains in families and friends and therefore interactions and cooperation with unrelated people are limited. As generalized trust with strangers (high-trust cluster, i.e., social trust) can facilitate sharing valuable information and knowledge with a wide range of entrepreneurs and economic actors, it plays an important role in innovation through learning spillovers in society (Welter 2012). Therefore, high-trust economies are likely to foster innovative entrepreneurship and economic growth.

However, when one considers high-performing East Asian economies, this presumed link between social trust and innovation becomes puzzling. East Asian economies—especially the People's Republic of China (PRC); Taipei,China; and Hong Kong, China; as well as Singapore, which has an ethnic Chinese majority, but also the Republic of Korea to some degree—are often defined as having low-trust societies because personal ties based on family values tend to overshadow broader social networks and generalized trust (Fukuyama 1995). In spite of this, East Asian economies have achieved a high level of innovation which has contributed to their rapid growth in the last several decades. For instance, the Global Innovation Index (GII) (Cornell University, INSEAD, and WIPO 2014–2018), which evaluates inputs and outputs of innovation of countries and economies worldwide, ranks East Asian economies ahead of most others in terms of innovation in technological and knowledge development. Singapore was ranked 5th; the Republic of Korea 12th; Hong Kong, China 14th; and the PRC 17th out of 126 economies evaluated in 2018 (in addition, Japan was ranked 13th). Moreover, these economies are characterized as early adopters of innovation outcomes such as information technologies, online games, and new medical treatments.

This puzzle exhibited by these East Asian economies provides the motivation to reevaluate the relationship between social trust and innovation. The central question here is how the presumably low-trust East Asian economies have played lead roles in building innovation economies. One potential answer to this conundrum is that the relationship between social trust and innovation differs in East Asia compared to others, with social trust not necessarily playing an important role in innovation in the East Asian context.

Let us consider the East Asian economies’ development paths, where innovation was led by the states during the period of fast economic growth. The state played the role of an entrepreneur and participated in markets as the main actor (Gulati 1992, Moon and Prasad 1994). The state's role contrasts with the industrialization path taken by Western economies that was mainly led by individual entrepreneurs. Under the collective leadership of the state, the role of social trust in innovation can be different from the one observed in Western economies. Because of state-mediated cooperation among individual entrepreneurs and knowledge and technology spillovers in East Asia, the level of trust between individual entrepreneurs may not be as important as in Western economies. Instead, social norms and values can be more crucial determinants of innovation, such as the degree to which individual entrepreneurs share collective goals of development and accept the modes of development proposed by the state. Likewise, trust in institutions that set up common goals and coordinate economic activities can be an equally important factor in encouraging individual stakeholders to support innovation.

In this regard, shared social norms and institutional trust are proposed as the elements of social capital that play the key role of contributors to innovation in East Asia instead of social trust. With this proposition, this paper aims to empirically examine how social norms and institutional trust promote innovation in East Asia by borrowing concepts from social capital theory and disentangling dimensions of social capital to ascertain their roles in innovation. To do so, the study conducts multilevel analyses through which the relationships between different components of social capital and innovation are identified in the six high-performing East Asian economies (Hong Kong, China; Japan;¹ the PRC;² the Republic of Korea; Singapore;³ and Taipei,China). The findings of this paper suggest that social trust is not an important determinant of innovation in East Asia, while its positive effect on innovation is generally supported in other parts of the world. Instead of social trust, the findings highlight the importance of social norms and institutional trust, as hypothesized above. In particular, social norms that endorse the shared societal goal of growth primacy and competition- and performance-based incentives are the prime drivers of innovation in East Asia. These results contribute a new piece of evidence in explaining East Asian development by emphasizing competition-based innovation (instead of innovation based on cooperation as proxied by social trust).

The paper is organized as follows. Section II provides a literature review and presents testable hypotheses. In section III, social capital is decomposed into different dimensions and each component is explained. In section IV, the hypotheses are tested across the economies, followed by an analysis at the individual level presented and discussed in section V. Section VI recaptures the main findings and concludes the paper.

In the social capital literature, many studies emphasize social trust as an important determinant of innovation and growth (Akcomak and ter Weel 2009; Beugelsdijk and van Schaik 2005; Dakhli and De Clercq 2004; Hauser, Tappeiner, and Walde 2007; Horvath 2013; Knack and Keefer 1997; La Porta et al. 1997; Paldam and Svendsen 2000; and Zak and Knack 2001).⁴

Nonetheless, the East Asian experience presents an interesting case with counterevidence to the presumed positive relationship between social trust and innovation, as their economies grew fast despite lower levels of social trust. This observation offers an alternative view on the role of social capital. In particular, social trust may not always be the necessary condition to achieve innovation, but other types of social capital may emerge as more important and take over the role of social trust when social trust endowments are not readily available in society (Horvath 2013). In fact, substantial empirical evidence is provided in the literature that supports the role of other types of social capital in innovation. In a cross-country study, Doh and Acs (2010) show that civic norms and institutional trust promote associated entrepreneurial activities. In another study on firm-level behaviors, Landry, Amara, and Lamari (2002) propose stakeholders’ participation and reciprocal norms as crucial social capital assets that determine firms’ decision to innovate. In addition, Yoon et al. (2015) suggest shared norms as a type of cognitive social capital that vitalized entrepreneurial efforts in the Republic of Korea and Taipei,China. Another study on Spain by Molina-Morales and Martínez‐Fernández (2010) also endorses the importance of shared orientations in shaping policy makers’ and entrepreneurs’ perceptions toward innovation.

As summarized above, the literature gives a special emphasis on social norms because innovation can be facilitated when society shares supportive norms and values. Such social norms are probably more important in East Asian economies, considering that they have upgraded their economic structures from agriculture-based economies to high technology-based innovation economies within only a few decades under the leadership of the states. Compared to private-led Western industrialization where trust between individual entrepreneurs and investors were an important determinant in forming a cooperation for innovation, the essential elements that supported public-led innovation in East Asia are arguably different. For economic players in East Asia, it is more crucial to share goals set by the state and demonstrate collective will toward the goals in order to participate in entrepreneurial projects sponsored by the state. Hence, social norms of embracing developmental goals can play a deciding role in stimulating innovation. Similarly, state-led innovation requires citizens to endorse and support the legitimacy and leadership of the state as the main economic stakeholder, and therefore the importance of trust in formal institutions overshadows that of social trust among individuals in the East Asian economies.

With these arguments in mind, social norms and institutional trust that reflect societal recognition of state leadership and common goals are hypothesized as essential elements of social capital that drive innovation in East Asia. Accordingly, this paper proposes and empirically tests the following hypotheses:

• Hypothesis 1. In high-performing East Asian economies, social trust has no positive effect on innovation.

• Hypothesis 2. In high-performing East Asian economies, social norms that support the shared goal of development have a positive effect on innovation.

• Hypothesis 3. In high-performing East Asian economies, trust in formal institutions has a positive effect on innovation.

To test for the hypotheses articulated above, social capital is decomposed into different dimensions, distinguishing between social trust, norms, institutional trust, and others, so that the net effect of each element can be identified. This paper follows the measurement approach suggested by Scrivens and Smith (2013) in a study done for the Organisation for Economic Co-operation and Development (OECD). According to Scrivens and Smith (2013), social capital consists of four essential dimensions: trust, personal and social networks, cooperative norms and values, and civic engagement. This classification seizes the concept of social capital defined by Putnam (1993) and Coleman (1990) who proposed social networks and shared norms as underlying values of creating social trust and citizens’ participation. Adopted by these two representative scholars of the social capital school, this concept is widely recognized and used in empirical studies on social capital.

With this in mind, the empirical model of this paper is designed to account for each effect of social norms, civil participation, trust, and networks by employing data from the 4th–6th waves of the World Values Survey (World Values Survey Association 2005–2014). A special emphasis is placed on social and civic norms shared by members of society that support collective goals and values toward development, as hypothesized above. Accordingly, the following questions in the survey are selected to measure such norms: (i) whether one supports the shared societal goal of economic growth (growth primacy, answered on a scale of 0 to 1); (ii) how much one embraces the values of competition and incentive orientation as modes of achieving goals (competition orientation and acceptance of inequality, respectively; scale of 1 to 10); (iii) reciprocity for others in society (scale of 1 to 6); and (iv) tolerance toward social minorities (scale of 1 to 10). Components (i) and (ii) represent social norms, while (iii) and (iv) reflect civic-mindedness.

Trust, another important dimension of social capital, is further subdivided into social, personal, and institutional trust. In this analysis, institutional trust is hypothesized as a crucial determinant of innovation in East Asia because it can be more relevant for state-led innovation than trust among individuals. Institutional trust is measured by degrees of trusting various types of formal institutions that were asked in the World Values Survey: (i) trust in courts (answered on a scale of 1 to 4); (ii) trust in parliaments (scale of 1 to 4); and (iii) trust in fairness of the rules (scale of 1 to 10). Distinguished from institutional trust, social trust is measured by using the question on how much one trusts other people in general (scale of 1 to 4). This question captures the level of trust in broadly defined groups of people. This type of trust can be an asset in forming cooperation with many other individuals for economic and entrepreneurial activities. In contrast, personal trust is evaluated through a question on how much one trusts family members, close friends, and relatives (scale of 1 to 4). Personal trust refers to a limited cluster of people who can be available for cooperation, which is different from social trust.

The other decomposed dimension of social capital is networks, which accounts for the scope of connectivity among individuals. Networks can function as a tool of imposing informal sanctions and encouraging members to participate in common activities. Similar to distinguishing social and personal trust, networks are also separated based on their extent, with each network's outreach involving different people and participatory causes: social networks for civil participation (humanitarian networks) and personal networks for private interests (hobby networks). The degree of one's participation in each type of network is measured by the frequency of participation in the respective network on a 3-point scale from 0 (no participation) to 2 (very frequent participation).

In total, social capital is decomposed into 12 different dimensions. The binary correlations across the different elements of social capital are all positive varying from 0.07 to 0.68 (see Appendix 2). These results show that the decomposed elements of social capital share the same direction, but each variable includes an independent component to a considerable degree (between 0.32 and 0.93). This finding verifies the empirical justification of distinguishing different dimensions of social capital and identifying each effect. The descriptive statistics of the social capital variables can be found in Appendix 1, and the utilization of the variables in the empirical model is discussed in more detail in section V.

Before closely examining the relationship between social capital and innovation in East Asia at the individual level, an aggregate analysis is conducted at the level of each national economy to show whether the effects of social capital are different between East Asia and other parts of the world. If so, the evidence can be used to support a specific role of social capital in East Asian economic development.

The aggregate analysis employs a reduced form of using one representative measurement for each dimension of social capital so that links between social capital and innovation can be presented without multicollinearity. Accordingly, the following variables are encompassed in the model: social trust (trust in other people in general), social networks (participation in humanitarian organizations), institutional trust (trust in the parliament), the shared norm of growth primacy (endorsing growth as the most important societal goal), and tolerance toward minorities (accepting homosexuals as neighbors). The decomposition analysis model in section V includes the full set of 12 social capital variables in order to disentangle the net effect of each element on innovation.

Besides social capital variables, income (gross domestic product per capita) and population variables enter the model as control variables as they represent important conditions in each economy that influence innovation. In addition, time-invariant unobserved heterogeneity in each economy (denoted as $αi$⁠) is accounted for in this model (fixed effects are applied when the dependent variable is the logarithmic (log) number of patents and random effects when the GII is used as the dependent variable due to limited variations in the GII scores). The rest of the unobserved effects is addressed as an error term (⁠$uit$⁠), and time effects are controlled for by using year dummies (⁠$t$⁠).

As for the dependent variables, two innovation measurements are used in this model. The first is the (log) number of patent applications of an economy in a given year. Patents are considered an important indicator of innovation outputs, and the number of patents is the most frequently used measurement of innovation in the literature (Knack and Keefer 1997, Zak and Knack 2001). Second, the knowledge and technology outputs of the GII are taken as an alternative measurement. This index evaluates multifaceted dimensions of knowledge and technology including knowledge creation (e.g., patent applications); impact (e.g., high technology outputs); and diffusion (e.g., intellectual property receipts). This measurement is chosen because it focuses on outputs of high technology-based innovation that are relevant for the scope of this paper.

The period of investigation is 2005–2017 when the dependent variable is (log) patent applications, because data on social capital variables are available in the World Values Survey for this period. When the GII serves as the dependent variable, the analysis includes the period 2014–2017 due to the availability of the index. The model is estimated using a linear estimation technique.

In Table 1, the results are presented separately for the global sample of countries (up to 83 countries that were surveyed in the World Values Survey, excluding the six East Asian economies) and six East Asian high-performing economies (Hong Kong, China; Japan; the PRC; the Republic of Korea; Singapore; and Taipei,China). The comparison between the global and East Asian samples enables us to identify whether the effect of social trust differs in East Asia and whether social norms and institutional trust play an important role in promoting innovation.

First, in the global sample, it is evident that social trust has a positive effect on innovation (both patent applications and the GII). Without considering the other components of social capital (columns 1 and 3), increasing social trust by 1 percentage point increases the number of patent applications by 5.2% and the GII score by 0.2 percentage point. When social networks, institutional trust, the norm of growth primacy, and tolerance are all included in the model (columns 2 and 4), the effect of social trust decreases but remains positive and significant. Given that accounting for other compounding effects of social capital reduces the effect of social trust by 17%–60%, 40%–83% of the total effect of social capital can be attributed to social trust in the global sample.

However, when the sample is limited to the six East Asian economies, the results show quite a different outlook. Without controlling for the other components of social capital, social trust has no effect on innovation (column 5), in contrast to social trust's positive effect in the global sample.⁵ Moreover, its effect on East Asian innovation becomes negative—although marginally significant at the 10% level—after accounting for the effects of the other social capital components (column 6). Instead, social and civic norms, networks, and institutional trust promote innovation in East Asia, as all of these effects become positive and significant.

This comparison suggests that, as expected, the role of social trust is different in the East Asian economies than in the rest of the world. The findings presented in Table 1 support hypotheses 1, 2, and 3 in that the role of social norms and institutional trust dominates that of social trust in this region. In the following section, these hypotheses are further investigated by using decomposed data at the individual level in the six East Asian economies.

In this section, the relationships between innovation and social capital in the East Asian economies are further disentangled by employing the full set of social capital variables. To this end, a decomposition analysis that exploits individual variations of the variables is conducted for two reasons. First, this analysis can reveal how individual endowments of social capital influence one's attitudes toward innovation, with the evidence contributing to explanations of behavioral consequences of social capital. Second, the application of individual data enables the use of finer sets of social capital and innovation measurements so that a more precise link between each element of social capital and innovation can be established. In this decomposition analysis, the 6th wave of the World Values Survey 2010–2014 is used because it provides the largest set of social capital and innovation measurements.

Accordingly, 12 variables which mirror various aspects of social capital are incorporated in the model (instead of the five variables that appeared in the parsimonious model in the aggregate analysis). As presented in section III, the 12 social capital variables represent social, personal, and institutional trust, social and civic norms, and social and personal networks. The vector of social capital consists of the following variables:

In this model, individual attitudes toward innovation are used as the dependent variables. These variables reveal the extent to which individuals value innovation. Six measurements that evaluate the degree of importance and acceptance of the particular aspects of innovation are selected to compose the set of dependent variables as presented below:

In equation 2, $k$ denotes each of the 12 social capital variables, $j$ is each of the nine demographic variables (⁠$x$⁠), $m$ is each of the six economy dummies, and $i$ represents each individual (up to 7,462 in total). This model is estimated by using two regression methods: ordered probit, given the ordered structures of the dependent variables; and ordinary least squares (OLS) to compute marginal effects. The results of the linear and nonlinear estimations are presented and compared in Table 2.

In addition, the potential endogeneity of the model is addressed by using an instrumental variable (IV) approach, for which the instrument for each social capital variable is the average value of the respective social capital variable of the same demographic group that shares the same gender and age cohort with individual $i$ in the same economy (naturally excluding individual $i$ in computing the average value of the group). The exclusion criteria of the chosen instruments are tested by two-stage least squares estimations and the results are presented in Appendix 3 (the first-stage results for the explanatory power of the instruments and the Hansen tests for the exogeneity of the instruments). The IV estimations produce results that are qualitatively similar to those of the OLS and ordered probit methods presented in Table 2. Hence, I will focus on the findings in Table 2 in the following section.

The results presented in Table 2 show the effects of the 12 social capital variables in the six East Asian economies. The table highlights the importance of shared social norms and institutional trust over social trust. Note that when the dependent variable is “too much dependence on science,” the effects are expected to stand in opposition to those in the other models because this variable reflects negative attitudes toward innovation.

First, the decomposition analysis shows no positive effect of social trust on innovation in general. It even constrains individuals from having positive attitudes toward technology and science. The only exception is its positive effect on promoting new and creative ideas. Instead of social trust, personal trust plays a more positive role in innovation, improving one's support for technology and science. However, personal trust negatively affects one's attitudes toward new ideas and creativity, which is exactly the opposite effect of social trust. This finding infers a region-specific aspect of East Asian development, in that innovation has been facilitated through cooperation between personally connected individuals and thus personal connectivity is an important driving force. However, personal trust can also be a hindrance to new ideas and creativity, while social trust is their promoter. These mixed findings complicate the role of individual trust in East Asia.

In contrast to the mixed results of trust in individuals, trust in formal institutions has a more robustly positive effect on innovation. All of the three types of trust in formal institutions stimulate individuals’ positive attitudes toward innovation—the effects of trust in courts and fairness of the rules are significant in four out of the six models and the effect of trust in parliaments is significant in three models. A 10 percentage point increase in trust in courts, parliaments, and fairness of the rules boosts one's support for innovation by 9–19.2 percentage points, 12.7–21.4 percentage points, and 2.8–8.8 percentage points, respectively.

More importantly, shared social norms and values produce the most prominent effects of determining one's attitudes toward innovation. Particularly, the norms of growth primacy and competition orientation have robustly positive, significant effects in all models. Agreeing that economic growth is the main goal of the economy (growth primacy) increases the level of positive attitudes on innovation by 1.5–2.4 percentage points and reduces negative perceptions by 1 percentage point. Increasing the degree of accepting the norm of competition by 10 percentage points enhances one's innovative attitudes by 3.7–14.5 percentage points. In addition, the acceptance of economic inequality has a positive effect in five out of the six models—increasing the acceptance level of this norm by 10 percentage points improves one's supportive attitudes toward innovation by 3.1–7.2 percentage points.

These findings disclose the importance of the shared norm of stimulating growth through competition and individual efforts in East Asia. The role of social norms is further evident in the positive effect of reciprocity, which is significant in four of the six models. Increasing the degree of reciprocity for others in society by 10 percentage points raises one's positive attitudes toward innovation by 5–34.2 percentage points. The positive role of social norms found in this analysis provides an indication that East Asian innovation and growth have been achieved through the collective will of the society and shared norms toward development. On the other hand, the role of networks is limited in explaining innovation in this region. The effects of social and personal networks are either insignificant or negative. The only exception is the positive effect of social networks on new ideas and creativity—a 10 percentage point increase in social network participation increases one's support for new ideas and creativity by 11.2 percentage points.

Regarding the effects of individual traits, gender is an important determinant that shapes one's attitudes toward innovation. Women tend to support innovation less than men, and this negative effect remains significant with a gender gap of 0.5–2 percentage points against women after controlling for other individual differences in characteristics. This gender difference in attitudes toward innovation may be explained by the limited role women could play in phases of economic development, which results in fewer opportunities to contribute to innovation (this is probably true not only in East Asia but also worldwide). In contrast to the gender effect, the effects of income and education are generally insignificant in most specifications. Nonetheless, there is some evidence that higher income levels and university education enhance individuals’ support for new ideas and creativity and for science and technology. This finding implies that innovation is an outcome of resourceful environments.

The results above highlight shared social norms and institutional trust as crucial determinants of innovation in East Asia. This common finding in the six economies is further disentangled in this section by identifying economy-specific heterogeneity in the effects of social capital. Accordingly, the sample is divided by economy in order to implement within-economy analyses. Table 3 presents the results estimated by OLS. Overall, there is considerable heterogeneity across the six economies in the effects of the different types of social norms and institutional trust.