Diverse knowledge for diverse innovation: Evidence from Chilean firms

Lauterbach, Rodolfo Article Diverse knowledge for diverse innovation: Evidence from Chilean firms Estudios de Economía Provided in Cooperation with: Department of Economics, University of Chile Suggested Citation: Lauterbach, Rodolfo (2024) : Diverse knowledge for diverse innovation: Evidence from Chilean firms, Estudios de Economía, ISSN 0718-5286, Universidad de Chile, Departamento de Economía, Santiago de Chile, Vol. 51, Iss. 1, pp. 85-115 This Version is available at: https://hdl.handle.net/10419/312791 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-nc-sa/4.0/ 85 Estudios de Economía, Vol.51 - Nº 1, Junio 2024. Págs 85-115 Diverse knowledge for diverse innovation; evidence from Chilean firms Conocimientos diversos para innovaciones diversas, evidencia de firmas chilenas. RODOLFO LAUTERBACH* Abstract Using the Chilean Innovation Survey for 2019-2020, this work studies the effects of different knowledge sources on a range of innovation outputs. Findings reveal distinct impacts of sourcing information from competitors, customers, and government agencies on product, process, marketing, organizational, and social innovation outputs. Information from customers has a positive effect on overall innovation. Social innovation is positively influenced by information sourced from government agencies. These findings contribute to the understanding of how different knowledge sources shape innovation outputs on developing countries. They provide valuable insights for firms, policymakers, and researchers seeking to enhance innovation capabilities and inform evidence-based policies. Key words: Innovation output, Diverse knowledge sources, Chilean Innovation survey, Binary instrumental variable model. JEL Classification: O31, O32, D22. *Universidad de Las Américas Chile, Facultad de Ingeniería y Negocios, Sede Providencia, Manuel Montt 948, Santiago, Chile. [email protected] Received: March, 2023 Accepted: April, 2024 86 Estudios de Economía, Vol.51 - Nº 1 Resumen Utilizando la Encuesta de Innovación de Chile para 2019-2020, este trabajo estudia los efectos de diferentes fuentes de conocimiento en una variedad de resultados de innovación. Los hallazgos revelan distintos impactos de la obtención de información de competidores, clientes y agencias gubernamentales en los resultados de innovación social, organizacional, de marketing, de procesos y de productos. La información de clientes tiene un efecto positivo en la innovación general. La innovación social se ve influenciada positivamente por la información procedente de agencias gubernamentales. Estos hallazgos contribuyen a comprender cómo las diferentes fuentes de conocimiento dan forma a los resultados de la innovación en los países en desarrollo. Proporcionan información valiosa para empresas, formuladores de políticas e investigadores que buscan mejorar las capacidades de innovación e informar políticas basadas en evidencia. Palabras clave: Resultados de innovación, Diversas fuentes de información, Encuesta de innovación chilena, Modelo de variable instrumental binaria. Clasificación JEL: O31, O32, D22 1. INTRODUCTION Innovation has long been recognized as a crucial driver of economic growth and competitiveness. As societies and economies become increasingly complex and interconnected, the ability of firms to adapt and innovate becomes ever more essential. Understanding the factors that contribute to successful innovation is therefore of paramount importance for policymakers, researchers, and business leaders alike. It has been widely acknowledged that firms need to look beyond their internal resources and tap into external knowledge to foster innovation. However, the specific mechanisms through which diverse knowledge sources influence innovation outcomes require further investigation. Over the years, scholars have made significant progress in developing models to comprehend the dynamics of innovation. Data collected through innovation surveys has played a pivotal role in unraveling the causality behind innovation success. By examining various firm-level variables, such as research and development (R&D) expenditure, human capital, sales, and total employees, researchers have sought to identify the determinants of innovation output. However, a fundamental question remains: Are there specific variables that have distinct impacts on specific types of innovation outputs? To shed light on 87 Diverse knowledge for diverse innovation; evidence from Chilean... / Rodolfo Lauterbach this matter and disentangle the intricate causal relationships behind innovation success, it is crucial to develop robust models that consider bidirectional effects. The more we understand which factors contribute to specific types of innovation, the better equipped we are to formulate effective government policies that promote desirable outcomes for local economies, particularly in developing countries. While previous studies have shed light on the effects of knowledge sources on innovation performance in various contexts, there is a need to explore these issues within the unique context of Chilean firms. Chile is a dynamic and emerging economy that has made significant efforts to foster innovation and entrepreneurship. Therefore, examining the role of diverse knowledge in Chilean firms’ innovation outputs can provide a broader understanding of innovation dynamics in emerging economies. Building upon the seminal work of Cohen and Levinthal (1990), this paper focuses on the firm’s capacity to acquire and utilize information from diverse sources as a critical determinant of innovation. Recognizing the importance of addressing endogeneity concerns, our approach draws inspiration from the work by Crepon, Duguet, and Mairesse (1998). The primary objective of this paper is to estimate an empirical model that reveals the causal relationships between different types of information sources and various forms of innovation output. To accomplish this, we leverage reliable innovation survey data collected in Chile during the period of 2019-2020. Our model considers the evolution of empirical research on the determinants of innovation output and employs instrumental variables to estimate a binary treatment model with idiosyncratic average effects. Our findings demonstrate that the utilization of diverse innovation information sources has varying impacts on different types of innovation outputs, each with its unique magnitude. Notably, while information sourced from customers positively influences most types of innovation, we found no discernible effect from information obtained from competitors. Furthermore, government information emerges as a particularly valuable resource, benefiting social innovation significantly while also exhibiting positive effects on process and organizational innovations. By shedding light on the intricate relationships between information sources and innovation outputs, this study provides valuable insights for policymakers, researchers, and firms seeking to enhance their innovation capabilities. The empirical evidence presented herein serves as a foundation for evidence-based policy recommendations aimed at fostering specific types of innovation that can drive the local economies of developing countries forward. Overall, this research contributes to the existing literature by offering a comprehensive analysis of the role of diverse knowledge in driving diverse 88 Estudios de Economía, Vol.51 - Nº 1 innovation outcomes. By highlighting the nuanced relationships between information sources and innovation outputs, we aim to stimulate further research and inform strategic decision-making processes in both the public and private sectors. The rest of this paper is structured as follows. Section 2 presents a literature review with previous findings in the topic of this work. Section 3 proposes a theoretical model by which the variables are related. Section 4 presents the database while section 5 presents the empirical strategy. Results are discussed on section 6 and section 7 concludes with a discussion about the value of our findings. 2. PREVIOUS LITERATURE RESEARCH ON INNOVATION PERFORMANCE This literature review firstly highlights the importance of investigating innovation performance and its determinants in various contexts. Understanding the factors that contribute to successful innovation outcomes is crucial for firms and policymakers alike. By reviewing previous studies on innovation performance, this paper aims to contribute to the existing body of knowledge by examining the role of diverse knowledge in fostering innovation, specifically focusing on evidence from Chilean firms. By examining the literature on innovation performance, this paper aims to discuss the importance of external knowledge, ownership structure, organizational practices, sectoral differences, customer participation, and the effectiveness of different knowledge sources in driving innovation. Understanding these factors can help firms and policymakers develop strategies and policies that promote innovation and enhance overall economic performance. Numerous studies have focused on investigating innovation performance and its determinants. Crepon, Duguet, and Mairesse (1998) developed a model that established a framework for exploring the causation of innovation output and productivity growth by linking innovation survey variables. Building on this model, subsequent research has further examined the relationship between innovation survey variables and innovation output. The importance of external knowledge for innovation has been emphasized in various studies. Sofka and Grimpe (2010) argued that firms should develop strategies to leverage external information, and the success of this strategy significantly influences innovation outcomes. They demonstrated that combining in-house R&D investments with a market-oriented search strategy enhances the effectiveness of innovation efforts. 89 Diverse knowledge for diverse innovation; evidence from Chilean... / Rodolfo Lauterbach Ownership structure has also been identified as a factor influencing innovation performance. Choi, Lee, and Williams (2011) found that firms with foreign ownership have a higher probability of successful innovation. Their study, conducted on Chinese firms, revealed that foreign ownership and affiliation with a business group strongly influence the volume of patent registrations. This suggests that ownership structure plays a vital role in determining innovation outcomes. Organizational practices have been recognized as crucial factors for innovation success. Mol and Birkinshaw (2014) highlighted the significance of certain organizational practices in fostering innovation. They emphasized the role of external involvement in the innovation management process, which not only provides direct input from external change agents but also brings prior external experience as an internal agent of change. Analyzing sectoral differences in innovation outcomes is also important. Castellacci (2008) presented a sectoral taxonomy that integrated manufacturing and service industries within a comprehensive framework. This approach underscored the increasing importance of vertical linkages and inter-sectoral knowledge exchanges between these interconnected branches of the economy. Božić and Mohnen (2016) conducted a quantitative analysis using Croatian Community Innovation Survey data and found that while there are some differences, service and manufacturing SMEs share similar determinants of innovation activities. However, service SMEs rely more on acquired knowledge compared to their manufacturing counterparts. The relationship between customer participation and innovation performance has been explored in several studies. Chang and Taylor (2016) conducted a meta-analysis that examined the effects of contextual factors on the relationship between customer participation and new product development performance. Their analysis revealed that involving customers in the ideation and launch stages of new product development improves new product financial performance directly, as well as indirectly through accelerated time to market. However, customer participation in the development phase slows down time to market, leading to a deterioration in new product financial performance. The study by Anzola-Román, Bayona-Sáez, and García-Marco (2018) investigated the influence of internal and externally sourced innovation practices on the likelihood of achieving product and process innovations. Their findings indicated positive effects of internal R&D and externally sourced innovation practices, as well as a positive influence of organizational innovation on the realization of technological innovations. Understanding the most effective sources of innovative ideas remains a significant challenge in technological innovation management. Criscuolo et al. (2018) examined the effectiveness of different combinations of knowledge 90 Estudios de Economía, Vol.51 - Nº 1 sources for achieving innovative performance. Their study, based on a largescale sample of UK firms, revealed important differences between product and process innovation, with broader knowledge searches associated with the former. THE MANAGEMENT OF INNOVATION AND FIRM PERFORMANCE Innovation is widely recognized as a crucial driver of firm success, contributing to competitive advantage, market growth, and long-term sustainability. As the business landscape becomes increasingly dynamic and complex, organizations must continuously adapt and innovate to stay ahead. Consequently, understanding firm management factors that influence innovation performance has become a topic of great interest for researchers and practitioners alike. Cohen and Levinthal (1990) highlighted the concept of absorptive capacity, which refers to a firm’s ability to acquire, assimilate, and utilize external knowledge to foster innovation. They emphasized that prior knowledge and experiences significantly influence a firm’s absorptive capacity. This perspective underscores the importance of leveraging diverse knowledge sources and learning from external information to enhance innovation capabilities. By exploring the relationship between diverse knowledge and innovation outcomes, valuable insights can be gained into how firms can effectively tap into a range of knowledge domains. While much of the existing literature has primarily focused on product and process innovation, there is a growing recognition of other dimensions of innovation that extend beyond tangible outputs. These dimensions include management, organizational, and social innovations, which encompass novel practices, structures, and techniques that advance organizational goals. OECD/Eurostat (2018) proposed a comprehensive framework encompassing these various innovation types. Acknowledging and exploring these diverse dimensions of innovation contribute to a more comprehensive understanding of the innovation process and its impact on firm performance. Chen, Wang, and Huang (2019) investigated the relationship between organizational innovation and technological innovation capabilities, exploring their impact on firm performance. Through structural equation modeling, their study revealed that innovation capabilities partially mediate the link between organizational innovation and firm performance. Furthermore, effective innovation management practices play a vital role in realizing the full potential of innovation. Birkinshaw and Mol (2008) identified four key processes—motivation, invention, implementation, and theorization and labeling—that collectively shape management innovation. By examining the roles of change agents within and outside the organization, valuable in- 91 Diverse knowledge for diverse innovation; evidence from Chilean... / Rodolfo Lauterbach sights can be gained into how innovation management practices can be optimized to maximize the benefits derived from innovation efforts. However, despite the recognized importance of innovation and its multidimensional nature, challenges persist in realizing significant economic returns from innovation. Teece (1986) highlighted that profits often accrue to complementary asset owners, customers, and imitators rather than to the original developers of intellectual property. This raises important questions regarding the alignment of innovation strategies with appropriate management practices to ensure that firms capture and capitalize on the economic benefits of their innovative endeavors. Given the multifaceted and ongoing nature of innovation, it is essential to delve into the literature to gain a comprehensive understanding of the relationship between diverse knowledge and diverse innovation outcomes. By exploring the interplay between absorptive capacity and different dimensions of innovation, in the context of effective innovation management practices, this study aims to provide evidence on the relationship between diverse knowledge and diverse innovation outcomes among Chilean firms. Through this investigation, valuable insights can be obtained to inform firms’ innovation strategies and enhance their ability to drive successful innovation outcomes while realizing economic returns. INFORMATION SOURCES AND INNOVATION The study of information sources and their impact on firm-level innovation performance is highly motivated by the recognition of innovation as a critical driver of firm success. In today’s competitive business environment, firms are constantly seeking ways to improve their innovation capabilities and outcomes. Understanding the role of information sources in this process is essential for firms aiming to leverage knowledge effectively and achieve sustainable innovation performance. Previous research has shed light on the influence of different types of information sources on innovation. Arvanitis, Lokshin, Mohnen, and Wörter (2013) conducted a study based on panels of Dutch and Swiss innovating firms, finding that both “buying” and “cooperating” have a positive effect on innovation. However, simultaneous utilization of these information sources does not necessarily lead to higher innovation performance. Pejić Bach et al. (2015) emphasized the catalytic role of information sources in innovation improvement, utilizing CIS data from Croatia, France, and the Netherlands. Their findings indicated that internal sources, customers, suppliers, and universities are important information sources for both internal and external R&D activities across the three countries. Interestingly, firms from the Netherlands exhibit different 92 Estudios de Economía, Vol.51 - Nº 1 patterns in utilizing information sources, relying more on competitors compared to firms from Croatia and France. Additionally, government information sources had a relatively smaller impact on firms’ innovation performance. The distinction between internal and external sources of information has been explored in relation to the generation of product and process innovation. Gómez, Salazar, and Vargas (2016) examined the usage of internal and external sources of information by Spanish firms, including customers, suppliers, competitors, consultants, and universities. They found that the importance of external sources of information varies depending on the type of innovation considered. For process innovation, firms mainly rely on suppliers, while for product innovation, the main contribution comes from customers. Damanpour, Sanchez-Henriquez, and Chiu (2018) investigated the dual role of internal and external sources of knowledge and information in the adoption of managerial innovations. Their findings indicated that internal implementation actions have a stronger effect than external implementation actions in influencing innovation adoption. Dotzel and Faggian (2019) analyzed the relationship between external knowledge sourcing and various innovation outcomes in rural and urban establishments in the U.S. Their results suggested that external knowledge sourcing specifically promotes product, process, and green innovation in U.S. firms. They also highlighted the potential importance of knowledge sourcing from non-local organizations, particularly in supporting innovation in rural markets compared to urban markets. Furthermore, the literature has explored the effects of different combinations of knowledge sources on innovation output. Basit and Medase (2019a) highlighted the positive link between knowledge diversity and firm-level innovation performance, emphasizing the importance of knowledge from customers in the private and public sectors, as well as knowledge from competitors. Basit (2021) extended this research by examining the impact of external knowledge sources on the willingness of small and medium-sized enterprises (SMEs) to introduce organizational innovation, revealing the greater importance of external knowledge for small firms and their propensity to utilize diverse sets of external knowledge. By delving into the literature on information sources and innovation, it becomes evident that diverse knowledge utilization plays a vital role in driving firm-level innovation performance. The interplay between different types of information sources, whether originating from paid deals or cooperation agreements, and whether derived from internal or external agents, offers valuable insights for firms aiming to enhance their innovation capabilities and achieve superior innovation outcomes. Therefore, this study seeks to contribute to the existing body of knowledge by examining the relationship between diverse knowledge sources and diverse innovation outcomes within the context of 99 Diverse knowledge for diverse innovation; evidence from Chilean... / Rodolfo Lauterbach TABLE 2 DESCRIPTION OF VARIABLES Variable Name Type Description Any innovation Dummy 1 if the firm introduced any innovation in 2019-2020 and 0 otherwise Product innovation Dummy 1 if the firm introduced new or significantly improved product or service in 2019-2020 and 0 otherwise Process innovation Dummy 1 if the firm introduced new or significantly improved operational processes in 2019-2020 and 0 otherwise Marketing innovation Dummy 1 if the firm introduced marketing innovation (i.e. significant modification in design or packaging of goods or services) in 2019-2020 and 0 otherwise Organizational innovation Dummy 1 if the firm introduced organizational innovation (i.e. new business practices for organizing procedures) in 2019-2020 and 0 otherwise Social innovation Dummy 1 if the firm introduced social innovation in 2019-2020 (i.e. sustainable innovation) and 0 otherwise Internal R&D Dummy 1 if the firm carried out internal R&D activities Other Innovative Investments Dummy 1 if the firm carried out investments Source of knowledge from competitors Dummy 1 if the firm get information source for new ideas in current innovation projects from the competitors in 2019–2020 and 0 otherwise Source of knowledge from the customers Dummy 1 if the firm get information source for new ideas in current innovation projects through the customers sector in 2019–2020 and 0 otherwise Source of knowledge from the government Dummy 1 if the firm get information source for new ideas from interaction with government agencies in 2019-2020 and 0 otherwise Graduate employees Continuous standardized to 0-1 Number of graduate employees (professional, master or PhD) to the total number of employees in 2020. Employment log The log of the number of employees as a measure of firm size High Exports Dummy 1 if exports were higher than USD$500,000 over 2019-2020 and 0 otherwise. Public Funding Dummy 1 if firm received innovation funding from the public sector in 2019-2020 and 0 otherwise Ac tA ct ii 11 3 ,,…Dummy 1 if firm belongs to specific sector and 0 otherwise Source: Variables defined based on data from the Chilean Innovation Survey 2019-2020. 100 Estudios de Economía, Vol.51 - Nº 1 TABLE 3 CORRELATION MATRIX 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 1. Any Innovation 1,00 2. Product Innovation 0,66 1,00 3. Process Innovation 0,93 0,51 1,00 4. Marketing Innovation 0,50 0,37 0,53 1,00 5. Organizational Innovation 0,62 0,36 0,66 0,52 1,00 6. Social Innovation 0,32 0,31 0,28 0,22 0,24 1,00 7. R&D Activity 0,52 0,48 0,46 0,30 0,32 0,31 1,00 8. Other Innovative Investments 0,76 0,53 0,72 0,40 0,47 0,29 0,46 1,00 9. Source Competitors 0,12 0,12 0,13 0,11 0,10 0,09 0,17 0,14 1,00 10. Source Clients 0,54 0,54 0,49 0,41 0,37 0,27 0,49 0,59 0,20 1,00 11. Source Government 0,25 0,23 0,22 0,16 0,18 0,24 0,31 0,26 0,31 0,29 1,00 12. Highly Educated 0,12 0,14 0,11 0,07 0,09 0,08 0,12 0,09 0,04 0,10 0,08 1,00 13. Log Total Employees 0,19 0,13 0,17 0,08 0,14 0,09 0,22 0,18 0,07 0,13 0,08 -0,20 1,00 14. Exports USD$500.000+ 0,12 0,10 0,08 0,06 0,07 0,06 0,19 0,10 0,05 0,10 0,07 0,00 0,33 1,00 15. Public Inn. Funding 0,22 0,22 0,19 0,10 0,15 0,17 0,32 0,23 0,07 0,19 0,24 0,09 0,04 0,05 1,00 Source: Own calculations based on the Chilean Innovation Survey 2019-2020. 101 Diverse knowledge for diverse innovation; evidence from Chilean... / Rodolfo Lauterbach 5. EMPIRICAL STRATEGY To empirically estimate the theoretical model, our first step is to examine the presence of endogeneity related to a selectivity problem. The literature provides several compelling reasons why innovation could also influence R&D, which have been well-documented. Mansfield (1969) presented one of the earliest works on this relationship, arguing that successful innovation increases a firm’s technological opportunities, making further innovation efforts more likely. Another argument for the impact of innovation on R&D is the difficulty firms may face in obtaining funding for innovation projects from external sources due to their inherent riskiness (Peters, 2009). If successful innovations lead to increased profitability and access to external funding, firms are more likely to engage in further R&D. Furthermore, the relationship between innovation, exporting, and R&D has been discussed as a bidirectional force by Harris and Moffat (2011). Some studies have emphasized the persistence of innovation and its positive impact on subsequent R&D investment. Geroski et al. (1997) and Malerba and Orsenigo (1999) have also explained the mechanism through which innovation influences R&D. From the literature, it can be concluded that the determinants of R&D expenditure for an individual firm are not completely independent of the firm’s probability of innovating. Innovating firms allocate resources to R&D to achieve innovations, while non-innovating firms may invest in R&D to enhance their absorptive capacities. Additionally, the variables that explain R&D may differ depending on whether the firm is innovating or not. Hence, there is a selectivity problem. To address the selectivity problem, one perspective is to consider innovation as an auto-selection process. Expected R&D investment depends on the firm’s innovation status, making the selectivity problem more complex than a simple sample selectivity bias. Kriaa and Karray (2010) suggest that one approach to solving this problem is to limit observed heterogeneity between firms while also controlling for unobserved heterogeneity. Other researchers have used an approach based on Heckman (1979) to address selectivity problems in this model. Following Basit and Medase (2019b), this work adopts an instrumental variable (IV) binary treatment model with a selection equation based on a set of instruments as the empirical methodology. The econometric model aims to study the relationship between firm-level innovation, human capital, internal R&D activities, and sources of knowledge flows. Given the binary nature of the endogenous and instrumental variables, this study employs an IV binary treatment model. The estimation method is a 102 Estudios de Economía, Vol.51 - Nº 1 two-stage Heckman binary treatment model. This empirical setup allows us to address potential endogeneity problems. The binary treatment model used in this research has been thoroughly explained by Wooldridge (2010) and has been employed by authors such as Basit (2021) and Cerulli (2012). The twostage Heckman binary treatment model with heterogeneous treatment response helps to address the endogeneity issues that arise in this context, where the relationship between innovation output and performance differs between firms investing in R&D and those that do not. The specification of the instrumental variable model is as follows: (2) ywxwxewe e x        00 01 0 Where we assume that observable and unobservable heterogeneity are not the same, so ee 10   . Following the principle of the two-stage sample selection estimation of Heckman (1979), we assume that on a binary treatment model we can still observe normality of the error term. This way we use a general model firstly specifying a fundamental regression. (3) yx ii i   1 Where selection implies that the dependent variable is known under the condition that z ii   20 Where   12 12 001~,,~,, ,NNand Corr    . And if we could assume that  0 , we could ignore the selection problem. The strategy then implies the estimation of two equations, the main equation with innovation output as dependent variable, and a selection equation with R&D dummy as a dependent variable. Innovation output is a set of dummy variables that can describe each type of innovation separately with: (4) INNxRD ii ii     12 & where: xInnAct InfInf InfHCEmp Ac tA ct iiii ii ii i n   ,,,,,,,, 12 31 , and: (5) RD xz ii ii &   112 where: zHExpPublicFunding ii i   ,, Where InnActi is a dummy with value 1 if the firm has spent on any of the other non-R&D activities: machinery, knowledge acquisition, training. Infi 1 103 Diverse knowledge for diverse innovation; evidence from Chilean... / Rodolfo Lauterbach is a dummy indicating whether the source of ideas for innovation developed with information from competitors, Infi 2 is a dummy indicating whether the source of ideas for innovation developed with information from customers, Infi 3 is a dummy of whether the ideas for innovations came from government sources. HCi is human capital intensity measured as highly educated employees divided by total employees, Empi is the log of the number of employees as a measure of firm size. Ac tA ct ii n1,,… are economic sector dummies. For the first stage equation the instrumented variable RD i & is a dummy equal to 1 if the firm has done R&D investment. We used two instruments that are statistically valid with significantly higher correlation to the instrumented variable compared to the endogenous variable1. The instruments are HExp i that is a dummy indicating whether exports were higher than USD$500,000, and PublicFundingi, that is a dummy equal to 1 if the firm received any kind of public funding for innovation during the period 2019-2020, and 0 otherwise. The instruments were chosen considering that both, access to exporting markets and access innovation public funding, are expected to have a greater impact over R&D efforts compared to innovation outputs because the latter result from a more complex knowledge generation processes that is affected by innovation efforts, information sources and firms’ human capital. 6. MAIN RESULTS The findings of this paper are presented in this section. The estimation method begins with a set of preliminary binary probit regressions using 5,519 observations. This step is taken before considering any endogeneity problems. Table 4 displays plausible results that align with the theoretical model. All types of innovation outputs considered in the model are positively and significantly influenced by both R&D investment and other innovative investments. The proportion of employees with a professional title or higher level of education also has a positive and significant impact on innovation output in all regressions. Firm size, measured as the logarithm of total employees, consistently shows a positive parameter in all regressions, although its impact appears to be lower compared to the other variables. Furthermore, firm size has a significant impact on process, organizational, and social innovation, but its significance is not observed in the case of product and marketing innovation. This preliminary result suggests that smaller firms may have the ability to achieve these types of innovation output without facing clear disadvantages due to their size. 1 See the correlation details on table 3 104 Estudios de Economía, Vol.51 - Nº 1 Based on these initial results, there is evidence that information from competitors may have very little or no impact on all types of innovation output. This result could be due to biases caused by endogeneity problems. It could also be attributed to the fact that a relatively small percentage of firms utilize information from competitors. Regarding sourcing innovation information from customers, the results indicate that it is an important and significant variable that positively affects all types of innovation outputs. This finding suggests that firms attach greater importance to customer feedback, indicating that customer-oriented firms are more likely to succeed in their innovative endeavors. This observation aligns with recent management literature that emphasizes the importance of focusing business models on customers. The regressions also reveal that sourcing information from government agencies is associated with specific types of innovation outputs. The results propose that government information has a significant impact only in the case of social innovations. However, following our empirical strategy and in line with previous literature2 on the estimation of innovation determinants, Table 5 examines the same question using a Heckman two-stage binary instrumental variable treatment model. This estimation method has been employed in other papers, including Basit and Medase (2019a, 2019b). The binary selection variable is R&D activity, and we use dummy variables as instruments to indicate whether exports exceed USD$500,000 and whether any public funding for innovation was received. Both instruments exhibit considerably higher correlation with the R&D activity dummy compared to innovation output variables. Like on the previous regressions, control variables for economic sector are included but not reported in the table. First stage results are reported on the first column. Note that the first-stage equation is the same for all six innovation equations. We find that both instruments, exports and public funding have a positive and significant impact on R&D efforts at the firm level. The results of the following columns suggest that innovative investments other than R&D is also a significant determinant of all types of innovation. The previous finding indicating a low and insignificant impact of information from competitors on innovation output is also supported by these results. Information from clients as a source for innovation ideas has a positive and significant effect in all cases except for social innovation, where government agencies emerge as the only important and significant information source. Government information also has a positive and significant impact on process and organizational innovation, albeit with smaller parameter sizes. The positive effect of the proportion of highly educated employees on different types of innovation 2 The argumento f why R&D should be considered endogenous on an Innovation equation is particularly well explaiden in the work by Crepon, Duguet and Mairesse (1998). 105 Diverse knowledge for diverse innovation; evidence from Chilean... / Rodolfo Lauterbach output is observed, although the effect is smaller than what was observed in the previous table and is not significant in the case of social innovation. The results also demonstrate that the logarithm of the total number of employees has a positive and significant impact on innovation outputs, except in the case of product and marketing innovation, which is consistent with the findings from the previous table. For each of the second stage equations, the two-stage Heckman model estimates rho (actually, the inverse hyperbolic tangent of rho) that represents the correlation of the residuals in the two equations. Additionally, it presents the estimation of sigma (actually, the log of sigma) which represents the standard error of the residuals of the second stage equation. Lambda is rho*sigma, which is found to be significant on all but one of the equations which suggests that the estimation of R&D in the first equation is relevant for the estimation of the second stage equations for all kinds of innovation output except organizational innovation. These results are relevant because, based on a previously validated empirical strategy3 that takes endogeneity into account, they explain the importance of different information sources for the several distinct types of innovation outputs. The results show some similarity with previous works4 regarding the importance of customer information for innovation output but also differ finding that for the case Chilean firms the importance of information from competitors is not an important determinant for innovation output. It could be the case that this result is observed because Chilean firms have a very low probability of sourcing innovation information from competitors, and hence there is not enough variation to find a significant parameter. In fact, only 0.4% of firms declared to have used information from competitors as a source of innovation ideas. Additionally, it could also be the case that the low use of competitor information is the result of low trust or higher levels of secrecy among industry-level competitors. In any case, this result calls for further research that can dig into industry-level information flows to explain this low frequency and low impact firm competitor relation. But even though information from competitors has not proven to be relevant for innovation output, we found that innovation output among Chilean firms is driven to a large extent by market orientation, particularly by sourcing information from customers. In the line with the findings by Anzola-Román (2018), this result is important from a managerial point of view because it shows that considering customer data is an important driver of innovation success. Additionally, from the public policy perspective, this result implies the opportunity of developing public instruments to promote customer-firm inter3 Heckman (1979) 4 Basit and Medase (2019a, 2019b). 106 Estudios de Economía, Vol.51 - Nº 1 actions such as experimental fairs or targeted consumer surveys. But the most relevant and novel result found on this work is related to the estimation of social innovation determinants. The work by Tortia et al (2020) discussed how social innovation interplays with entrepreneurship in public and private institutions. Social innovations imply the achievement of results that benefit socially vulnerable groups or the enviroment, it should be financially sustainable, and it functions based on the use of new approaches and ideas to solve a particular social problem. We find that in terms of information sources, social innovation is mainly driven by information from government institutions, while customer and competitor sources are not relevant when the full model is estimated. This result, if confirmed by further research, could have important public policy implications. The work by Mulgan (2007) discussed that social innovation often involves universities, government agencies and private companies working together. He also showed that social innovation is more related to the combination of knowledge from different actors rather than the advancement of new technologies at the individual organizational level. Particularly, considering that social innovation has a large positive externality component, our results suggest that public funding instruments to promote collaboration with government institutions could help promote innovations that have the highest social value. 107 Diverse knowledge for diverse innovation; evidence from Chilean... / Rodolfo Lauterbach TABLE 4 PRELIMINARY BINARY PROBIT REGRESSION Any Innovation Product Innovation Process Innovation Marketing Innovation Organizational Innovation Social Innovation R&D Activity 1.313 0.863 0.798 0.412 0.406 0.666 (0.106)** (0.081)** (0.086)** (0.089)** (0.080)** (0.100)** Innovation Activity 2.397 1.183 2.125 1.062 1.240 0.806 (0.078)** (0.071)** (0.069)** (0.080)** (0.069)** (0.097)** Source Competitors -0.403 -0.381 0.216 0.089 -0.111 -0.349 (0.495) (0.283) (0.405) (0.282) (0.268) (0.301) Source Clients 0.586 0.868 0.296 0.676 0.378 0.296 (0.120)** (0.084)** (0.094)** (0.090)** (0.083)** (0.105)** Source Government 0.152 0.048 -0.024 0.105 0.151 0.573 (0.214) (0.143) (0.163) (0.149) (0.137) (0.146)** Highly Educated 0.412 0.439 0.390 0.307 0.383 0.349 (0.095)** (0.109)** (0.092)** (0.118)** (0.101)** (0.147)* Total Employees 0.062 0.029 0.059 0.010 0.077 0.066 (0.017)** (0.019) (0.016)** (0.021) (0.018)** (0.025)** Constant -1.760 -2.097 -1.768 -1.818 -2.244 -2.093 (0.171)** (0.197)** (0.168)** (0.183)** (0.195)** (0.203)** N 5519 5519 5519 5519 5519 5519 * p<0.05; ** p<0.01 Source: Own calculations based on the Chilean Innovation Survey 2019-2020. 108 Estudios de Economía, Vol.51 - Nº 1 TABLE 5 BINARY TWO-STAGE HECKMAN INSTRUMENTAL VARIABLE REGRESSION 1st Stage R&D Activity Any Innovation Product Innovation Process Innovation Marketing Innovation Organizational Innovation Social Innovation Innovation Activity 0.941 0.758 0.212 0.731 0.165 0.279 0.048 (0.074)** (0.014)** (0.013)** (0.015)** (0.012)** (0.014)** (0.009)** Source Competitors 0.117 -0.054 -0.084 0.068 0.060 0.016 -0.037 (0.323) (0.059) (0.051) (0.061) (0.046) (0.054) (0.035) Source Clients 0.997 0.250 0.271 0.218 0.230 0.129 0.015 (0.083)** (0.021)** (0.019)** (0.022)** (0.017)** (0.020)** (0.013) Source Government 0.575 0.143 0.012 0.113 0.044 0.058 0.138 (0.145)** (0.031)** (0.027) (0.032)** (0.024) (0.029)* (0.019)** Highly Educated 0.792 0.093 0.040 0.091 0.037 0.043 0.008 (0.123)** (0.014)** (0.012)** (0.014)** (0.011)** (0.013)** (0.008) Total Employees 0.203 0.019 -0.000 0.019 0.003 0.010 0.000 (0.021)** (0.003)** (0.002) (0.003)** (0.002) (0.002)** (0.002) Exports USD$500.000+ 0.348 (0.098)** Public Inn. Funding 1.080 (0.116)** R&D Activity -0.177 0.295 -0.272 -0.041 0.051 0.198 (0.041)** (0.039)** (0.043)** (0.035) (0.042) (0.027)** Constant -3.162 0.002 0.019 -0.003 0.044 -0.017 0.052 (0.239)** (0.024) (0.021) (0.026) (0.019)* (0.023) (0.015)** 5519 5519 5519 5519 5519 5519 5519 lambda 0.265 0.022 0.277 0.063 0.026 -0.055 (0.022)** (0.021)** (0.023)** (0.019)** (0.023) (0.015)** Rho 0.962 -0.096 0.962 0.281 0.054 -0.395 Sigma 0.275 0.236 0.287 0.215 0.252 0.165 * p<0.05; ** p<0.01 Source: Own calculations based on the Chilean Innovation Survey 2019-2020. 115 Diverse knowledge for diverse innovation; evidence from Chilean... / Rodolfo Lauterbach Teece, D. J. (1986). “Profiting from technological innovation: Implications for integration, collaboration, licensing and public policy.” Research Policy, 15(6), 285-305. Tortia, Ermanno C., Florence Degavre and Simone Poledrini (2020) Why are social enterprises good candidates for social innovation? Looking for personal and institutional drivers of innovation. © 2020 The Authors Annals of Public and Cooperative Economics © 2020 EMF Tsai, K.H. (2009), “Collaborative networks and product innovation performance: toward a contingency perspective”, Research Policy, Vol. 38 No. 5, pp. 765-778. Vega-Jurado, J., Gutiérrez-Gracia, A. and Fernández-de-Lucio, I. (2009), “Does external knowledge sourcing matter for innovation? Evidence from the Spanish manufacturing industry”, Industrial and Corporate Change, Vol. 18 No. 4, pp. 637-670.