Contents lists available at ScienceDirect Technological Forecasting & Social Change journal homepage: www.elsevier.com/locate/techfore The trajectory of the ability to innovate and the financial performance of the Brazilian industry☆,☆☆ David Ferreira Lopes Santosa,*, Leonardo Fernando Cruz Bassob, Herbert Kimurac a São Paulo State University (Unesp), School of Agricultural and Veterinarian Sciences, Jaboticabal, SP CEP 14.884-990, Brazil b Mackenzie Presbyterian University, School of Applied Social Sciences, Sao Paulo, SP CEP 01302-907, Brazil c University of Brasilia, School of Business, Accounting and Economics, Brasilia, DF CEP 70910-900, Brazil A R T I C L E I N F O Keywords: Innovative capacity Business competitiveness Technological strategies Emerging markets Profitability JEL classification: L60 M21 O14 O25 O32 O33 A B S T R A C T This research analyzes the cumulative trajectory of Brazilian industry's ability to innovate and the impact of this resource on firms' financial performance. From a broad base of data taken at the firm level, a cross-sectional analysis and a longitudinal analysis were combined, through structural equation modelling, in the construction of the trajectory of resource innovation with the combined use of the following techniques: a multilevel model, latent trajectory analysis, and an autoregressive model. The empirical model shows that the ability to innovate consists of factors that are associated with internal, external, and human resources. The influence on financial performance is positive and significant when the analysis involves the long term. The autoregressive effect of the ability to innovate in time is not significant, suggesting that the innovation process is cumulative, interactive, and nonlinear. These results are relevant to emerging countries that require continued public policies and a greater intensity of business investment in the innovation process, aiming at the longevity of companies. 1. Introduction Recent studies in innovation have sought to evaluate the cumulative trajectory of this organizational ability, whose bases turn towards the construction of technological paradigms, regimes, and standards (Castellacci, 2008; Figueiredo, 2010; Forés and Camisón, 2016; Sundbo and Gallouj, 2000). In parallel to academic studies, analyses by market professionals and public managers focus on understanding how to foster and develop strategies that are oriented towards innovation, whose results con- tribute to the growth of companies and the development of countries (Damanpour et al., 2009; Hu, 2014, 2008; Kostopoulos et al., 2011; Lancker et al., 2016; Samara et al., 2012). However, as a result of its analytical complexity and its evolutionary nature as a process of knowledge accumulation and refinement, managing the ability to in- novate is presented as one of the main challenges of organizational studies (Bessant, 2008; Cantwell and Fai, 1999; Dewangan and Godse, 2014; François et al., 2002; Kash and Rycroft, 2002; Olaru and Purchase, 2015; Rousseau et al., 2016; Silverberg and Verspagen, 2005). The difficulty in modelling innovation as an organizational resource is expressed in the controversial results regarding its influence on the business and economic performance of firms (Cozzarin, 2004; Gunday et al., 2011; Jiménez-Jiménez and Sanz-Valle, 2011; Kim et al., 2016; Liao and Rice, 2010; Lööf and Heshmati, 2006, 2002; Prajogo, 2016). More specifically, this scenario is relevant to emerging markets, whose insertion into global value chains depends on the degree of in- novation and competitiveness of their companies, which brings with them institutional, environmental, and social demands that must be overcome (Castellacci and Natera, 2016; Paunov, 2012; Wu et al., 2016; Xie et al., 2016). As one of the major emerging economies, in real terms, Brazil ex- perienced an increase of 135% in investments in innovation between 2000 and 2013 (MCTI, 2016). In 2013, the expenditure on science, technology, and innovation was 1.24% of the gross domestic product (GDP) compared to an average of 1.9% for the top 40 countries with spending on innovation (UNDP, 2013). However, investments in in- novation have not yet had a significant impact on the country. From http://dx.doi.org/10.1016/j.techfore.2017.09.027 Received 17 February 2017; Received in revised form 21 September 2017; Accepted 26 September 2017 ☆ This document was a collaborative effort. ☆☆ The authors gratefully acknowledge support from FAPESP (2013/12483-6), CAPES (Coordination for the Improvement of Higher Education Personnel) (20132053/ 33024014019P9), CNPq (National Council for Scientific and Technological Development) (443450/2014-5 and 310666/2016-3), and MackPesquisa (1562 NUCOI-FMP-030/2013). * Corresponding author. E-mail addresses: david.lopes@fcav.unesp.br (D.F.L. Santos), leonardofernando.basso@mackenzie.br (L.F.C. Basso), herbertkimura@unb.br (H. Kimura). Technological Forecasting & Social Change 127 (2018) 258–270 Available online 10 October 2017 0040-1625/ © 2017 Elsevier Inc. All rights reserved. T http://www.sciencedirect.com/science/journal/00401625 https://www.elsevier.com/locate/techfore http://dx.doi.org/10.1016/j.techfore.2017.09.027 http://dx.doi.org/10.1016/j.techfore.2017.09.027 mailto:david.lopes@fcav.unesp.br mailto:leonardofernando.basso@mackenzie.br mailto:herbertkimura@unb.br https://doi.org/10.1016/j.techfore.2017.09.027 http://crossmark.crossref.org/dialog/?doi=10.1016/j.techfore.2017.09.027&domain=pdf 2000 to 2012, for example, Brazil ranked only 70th in the Global In- novation Index 2015 (WIPO, 2015) and 74th in the Global Competi- tiveness Index (84th in the Innovation sub-item) (WEF, 2015), with a participation in international trade of only 1.22% in 2015 (WTO, 2015). Given the need to analyze investments in the formation of the ability to innovate with the financial performance of companies in emerging economies, the following question motivates this study: How does the cumulative process of the ability to innovate influence the financial performance of Brazilian industrial companies? Despite the limitations of research on this subject for the country (de Guimarães et al., 2016; Figueiredo, 2010; Santos et al., 2014), the Brazilian case can serve as a benchmark for other emerging countries, notably in Latin America, which demand higher investments in in- novation to increase their competitiveness and economic development. The specific condition of the Brazilian case, which may serve as a comparison for other emerging countries, has been characterized by: low intensity of business investments in R &D (Cyrino et al., 2017), emphasis on directing investments to acquire machinery and equipment (Frank et al., 2016); lack of insertion of researchers, with master and doctorate degrees, in companies (Santos et al., 2014); little interaction between companies and universities and research institutes to generate innovation (de Moraes Silva et al., 2017). As the largest volume of expenditures on science, technology and innovation conducted by the country comes from the government and the interaction among triple helix agents (university, private sector and government) is limited, the aggregate results of innovation are modest (Cyrino et al., 2017). This environment may be similar to the context of many developing coun- tries, which strive to get a competitive advantage from innovation. The differences in the technological trajectories of each country or sector are viewed as one of the variables that explain the heterogeneity of the results of innovation, even among developed countries (Atalay et al., 2013; Castellacci and Natera, 2016; Jackson et al., 2016; Samara et al., 2012; Wu et al., 2016). However, there is still no empirical model that shows the cumulative process of innovation and the effects on fi- nancial performance (Liao and Rice, 2010; Olaru and Purchase, 2015). Thus, with the aim of analyzing the cumulative trajectory of in- novation and its impact on financial performance, this study proposes an exploratory model of the cumulative process of business innovation based on investments in innovation by Brazilian industry between 2000 and 2011. This article is organized as follows. The next section provides the theoretical foundations that justify the model structure and allow the results to be discussed. The third section presents the survey and con- struction process of the variables. The fourth section presents the ana- lysis of the results, comparing them with the literature. Finally, the fifth section discusses the implications of the study for the theory of in- novation and the development of public policies and business strate- gies. 2. Theoretical framework The trajectory of innovation was formalized by Pavitt (1984), who showed differences in innovation investments and in innovation results among different sectors. The understanding of the differences in efforts and innovation results for each sector brought new prospects for the economic evaluation of innovation. Several quantitative studies whose data for analysis are aggregated by countries have been developed (Castellacci and Natera, 2016; Hatzikian, 2013; Hinloopen, 2003; Kirner et al., 2009; Solow, 1957). The heterogeneous results confirm the difficulties in relating in- novation investments to financial performance (Castellacci, 2008; Kleinschmidt, 1991; Liao and Rice, 2010; Santos et al., 2014). Crossan and Apaydin (2009) explain these differences by means of a framework that is divided into the determinants and the dimensions of innovation, with the determinants being distributed into group, organizational, and process levels and dimensions being stratified into process and results. The methodological limitation in understanding the extent of the innovation process and the differences in its results in terms of sectorial characteristics is one of the current challenges of this area of knowledge (Atalay et al., 2013; Ryu and Lee, 2016). For instance, Weber and Schaper-Rinkel (2017) investigate quality innovation and performance in Swiss hospitals, Wang et al. (2015) analyze open innovation and performance in high tech companies. The ability to innovate is an organizational resource that is con- tinuous and non-discrete in nature; moreover, it has cumulative char- acteristics because its formation requires the development of knowl- edge in an interactive and recursive manner and is dependent on the social networks in which each firm is inserted (Dewangan and Godse, 2014; Figueiredo, 2010; Hatzikian, 2013; Kash and Rycroft, 2002; Lancker et al., 2016; Sundbo and Gallouj, 2000). Thus, there is a learning curve in the innovation process, which means that the results of this resource are best observed in the financial performance of companies (Cimoli and Dosi, 1995; Linton and Walsh, 2004; Teece, 2010; Velu, 2015). The formation of the ability to innovate requires different types of investments (machinery, knowledge, people, materials, etc.) (Figueiredo, 2010; Urgal et al., 2011); manifests itself at all levels of the organization (strategic, tactical, and operational) (Teece, 2010); has results that are expressed in different ways (financial, market, opera- tional) (Dewangan and Godse, 2014) and that can be compared in different ways (business, market, country, world) (Crossan and Apaydin, 2009); and must be analyzed considering the time variable, given that there is a cumulative factor of innovation efforts (Cantwell and Fai, 1999). Internal research and development (R & D) is the classic variable for measuring investments in innovation and a source of mandatory ability to innovate for companies that take offensive and defensive strategies of innovation (Baumann and Kritikos, 2016; Bäck and Kohtamäki, 2015; Colombo and Rabbiosi, 2014; Freeman and Soete, 1997; Howell, 2016; Hung and Chou, 2013; Kim et al., 2016). However, the ability to in- novate is not restricted to these investments (Hatzikian, 2013; Lööf and Heshmati, 2006). A second variable that is widely used to measure innovation is human capital, which is scaled in different ways (e.g., the number of engineers and/or technicians, length of experience, the institutional education level (doctorate, master's degree, bachelor's degree), the number of people dedicated to R &D) (Hatzikian, 2013; Kim et al., 2016; Lööf and Heshmati, 2006). Innovation, as an intangible resource, comprised by different in- ternal and external sources to the organization, as well as, a result of a cumulative process of knowledge, allows the study of the different strategies and results, since the activities that constitute innovation are heterogeneous and exploited differently by firms (Tavassoli and Karlsson, 2015). In this context, companies present different levels of innovation, reflecting their absorptive capacity and the persistence of innovation results, e.g., Bartoloni and Baussola (2017) and Maslach (2015). These concepts have been explored in the recent literature, and one of the central axes is the cumulative process of knowledge and learning from internal and external sources that provide better innovation and en- trepreneurial performance (Lewin et al., 2011; Rangus and Slavec, 2017; Tavassoli and Karlsson, 2015; Tsai, 2001). Our study shows that linking innovation with the financial perfor- mance of companies requires the need to understand innovation as an organizational resource that brings together tangible and intangible elements, internal and external, to companies. This argument is com- patible with other studies (Lichtenthaler, 2016; Wang et al., 2017). However, differently from the studied literature, we present the cap- ability to innovate as a cumulative resource, i.e., the investments and the structure directed to innovation in the past contribute to the con- stitution of the resource innovation in the present. The contribution derives not only from an additive process of investments but also D.F.L. Santos et al. Technological Forecasting & Social Change 127 (2018) 258–270 259 through organizational learning capabilities (Bartoloni and Baussola, 2017). Thus, the capability to innovate is an organizational resource with cumulative nature of investments in the dimensions that constitute it, such as human capital, internal capital and relational capital. The management of the ability to innovate as a complex firm re- source that involves internal and external sources was also explored by Wang et al. (2017). The authors highlight the importance of managers in fostering a style of innovation-driven leadership as a way to exploit available resources that, in fact, contribute to value creation. In addi- tion, Lichtenthaler (2016) shows that innovation, as a resource that reflects the absorptive capacity of companies, involves multiple internal and external variables, and has effects on financial performance that occur over time. which is in keeping with the approach used in this study. Other internal factors that, according to the literature, are re- sponsible for the ability to innovate or generate innovations include the following: training (Liao and Rice, 2010); project development; mar- keting for the introduction of a new product (Kim et al., 2016); orga- nizational changes (Hu, 2014) and the acquisition of machinery and equipment (Hashi and Stojčić, 2013; Lööf and Heshmati, 2006; Urgal et al., 2011). External factors for innovation are more recent in empirical studies that discuss the impact of innovation on financial performance (Hu, 2014; Kostopoulos et al., 2011). This situation stems from, among other factors, the increased complexity and specialization of knowledge (Cruz-González et al., 2015; Hung and Chou, 2013; Ritala et al., 2015; Ye et al., 2016). Investments in the development of external R & D and the acquisition of external knowledge have become relevant for firms' ability to innovate (Bäck and Kohtamäki, 2015; Cruz-González et al., 2015; Wang and Wang, 2012; Ye et al., 2016). Regardless of the configurations and levels of investment in in- novation, all strategies that are associated with this resource should be directed to the creation of value that, in turn, can occur immediately or be reflected only in the long run (Bernardo, 2014; Lancker et al., 2016). The metrics for discriminating financial and economic results are more consolidated than the innovation indicators. However, associating investments in innovation with the financial performance of companies is still a controversial topic with regard to the methods and period of analysis, particularly in emerging countries (Atalay et al., 2013; Santos et al., 2014). Nevertheless, the availability of information and the re- liability of the verification process of primary data at the firm level are important elements for the validity of the empirical model (Klingenberg et al., 2013). The following are the main metrics that are used to analyze the financial performance of innovation resources: Return on Assets (ROA) and Return on Investment (ROI) (Klingenberg et al., 2013; Liao and Rice, 2010; Wang and Wang, 2012); Return on Equity (ROE): (Liao and Rice, 2010); Return on Sales (ROS): (Kostopoulos et al., 2011); and the EBTIDA margin (Klingenberg et al., 2013). Considering the theoretical discussion, Fig. 1 illustrates the con- ceptual model that is proposed for analyzing the influence of the cu- mulative trajectory of the ability to innovate on the financial perfor- mance of companies. The model presents innovation as a resource, i.e., an (intangible) asset of the organization that is oriented towards the generation of economic results from the combination of different fac- tors, segregated here into human capital, internal capital, and relational capital (Santos et al., 2014; Sundbo and Gallouj, 2000; Urgal et al., 2011). 3. Materials and methods 3.1. Materials The database was built with two of Brazil's major sets of business information, the Annual Industrial Survey (PIA) and the Technological Innovation Survey (PINTEC), between 2000 and 2012. Both surveys are managed by the Brazilian Institute of Geography and Statistics (IBGE). The radical of the codes of the National Register of Legal Entities (CNPJ) was used at the enterprise level to combine the two databases. The PIA is an annual survey that comprises more than 56,000 Fig. 1. Conceptual mdel for the cumulative trajectory of innovation. D.F.L. Santos et al. Technological Forecasting & Social Change 127 (2018) 258–270 260 industrial companies in Brazil with more than 30 employees or rev- enues exceeding US$ 3.2 million, and it has been conducted since 1996. First published in 2000, the PINTEC is a triennial survey; its ques- tionnaire follows the Oslo Manual, and the structure is similar to that of the Community Innovation Survey (CIS). The last published edition of the PINTEC (2011) covered 116,632 industries with more than 10 employees, and therefore, the sample is wider than that of the PIA. Considering the availability of the database, five cross-sectional samples were constructed, i.e., 2000/2001, 2003/2004, 2005/2006, 2008/2009, and 2011/2012, in addition to a sixth sample with data from all years for only the companies that presented results in all edi- tions of the PINTEC, thus constituting a balanced panel. After the combination and validation of the samples, the empirical study involved an analysis in a restricted environment of IBGE in the city of Rio de Janeiro called the ‘secret room’. The results of the analysis are available only with the approval of IBGE technicians who assess the confidentiality and security of the information of the respondent com- panies. Table 1 discriminates the 25 industrial sectors of the cross-sectional samples and the longitudinal sample of companies between the 2000 and 2011. The variations between years are due to the difference in the scope of the studies, the opening and closing of companies, and movements of corporate demerger, and the merger and acquisition of companies. The reduction in the companies in the longitudinal sample is a reflex of the intersection between the cross-sectional periods. The first seven sectors account for over 50% of the companies in the sample, but the distribution is not concentrated, which depicts Brazil's heterogeneous industrial profile. It was not possible to extract de- scriptive results per sector due to IBGE's requirement to restrict the results to sectors with at least 100 companies in all samples. Despite the differences among sectors, there is empirical evidence that there is a common technological trajectory for Brazilian industry based on the similarities in the external conditions for all companies and the dominant market perspective focused on the domestic market, which is still very protected by public policies (Castellacci and Natera, 2016). The operationalization of the variables of interest occurred as shown in Table 2, and all variables were standardized by subtracting the values observed by the sample mean and dividing by the sample standard deviation. 3.2. Method Factor analysis was used as a method of verifying the inter- dependence between the variables and their congregation in common factors, as represented by the latent variables HC, IC, and RC of the theoretical model in Fig. 1. The dimensions that are derived from factor analysis are vectors that minimize the distances between the individual values and express similarities (Dray, 2008). The matrix model follows Eq. (1) (Audigier et al., 2014). − = +μ ϵX ΛF (1) where (X−μ)is the difference between a real-valued vector (X= [X1,X2,⋯ ,Xn]T) and a mean vector (μ=[μ1,μ2,⋯ ,μn]T), and the var- iance-covariance matrix is positive. Xi represents measured variables depicted in Table 2, corresponding to dimensions of human capital, internal capital and relational capital. The observable variables can be expressed as a linear function of p latent variables or factors, where p< n. The latent variables Fi constitute a vector F=[F1,F2,⋯ ,Fp]T, uncorrelated with the error term ϵ, related to the corresponding ob- servable variable Xi. In the context of this study, we aim to analyze whether the structure of investments in innovation, taking into account the measured variables, reflects the latent dimensions studied in the literature. Following the evaluation of the factor analysis, based on structural equation modelling, a combination of three techniques was used: i) a latent trajectory, (ii) an autoregressive model (Delsing and Oud, 2008; Geiser et al., 2013), and (iii) a multilevel structure (Yuan and Bentler, 2007). This combination of techniques allows modelling, through time, innovation as a construct that cannot be directly observed by a single measured variable. Traditional multivariate regression analysis, Table 1 Sectorial stratification of the study sample. Sectors 2000/2011 2000/2001 2003/2004 2005/2006 2008/2009 2011/2012 Mean Manufacture of food products and beverages 250 1020 1012 1339 1420 1302 1219 Manufacture of machinery and equipment 126 547 506 656 941 945 719 Manufacture of chemicals and pharmaceutical chemicals 109 535 468 600 730 685 604 Manufacture of textiles 105 404 353 445 506 433 428 Manufacture of furniture and wood products 80 551 489 640 684 643 601 Manufacture of leather goods and footwear 67 377 355 498 465 397 418 Manufacture and assembly of motor vehicles, trailers, and bodies 65 278 262 364 515 433 370 Manufacture of office machines, computers, and communication equipment 63 160 144 194 356 283 227 Manufacture of metal products, excluding machinery and equipment 62 451 439 578 757 719 589 Manufacture of clothing items and accessories 55 453 405 642 570 523 519 Manufacture of rubber items and plastic materials 55 479 426 576 751 678 582 Manufacture of pulp, paper, and paper products 52 236 207 252 358 285 268 Manufacture of machinery, appliances, and electrical equipment 49 224 214 264 371 346 284 Manufacture of non-metallic mineral products 46 385 325 482 530 511 447 Metallurgy 42 195 193 250 329 270 247 Manufacture of wood products 34 316 270 369 332 257 309 Manufacture of coke, petroleum refining, development of nuclear fuels, 34 100 114 107 152 116 118 and alcohol production Manufacture of other transport equipment 21 55 84 105 96 96 87 Extraction of non-metallic minerals 13 122 106 141 153 184 141 Editing, printing, and reproduction of recorded media 11 254 203 279 128 147 202 Other sectors 1 9 5 7 13 16 10 Coal extraction 0 10 0 0 6 10 5 Oil extraction and related services 0 0 15 21 5 12 11 Extraction of metallic minerals 0 27 20 22 26 14 22 Manufacture of tobacco products 0 19 23 24 25 22 23 Total 1340 7207 6638 8855 10,219 9327 8449 D.F.L. Santos et al. Technological Forecasting & Social Change 127 (2018) 258–270 261 regardless of using cross section, longitudinal or panel data, would require a single observed variable to represent a proxy for the concept of innovation. This limitation has also been discussed in other studies, such as Cammarano et al. (2017), Luo et al. (2017)and Stefan and Bengtsson (2017). Due to the complexity of the innovation concept, SEM would allow considering innovation as a latent construct asso- ciated with various variables. Although regression with panel data could be used for analyzing cross-section and temporal observations, unfortunately in our case, the number of firms is high, but the number of years is small. In addition, the Brazilian Institute of Geography and Statistics does not conduct and publish PINTEC and PIA surveys following a constant frequency, which is critical for the use of traditional panel data techniques. We therefore chose longitudinal SEM, which makes possible a representation of in- novation as a non-observable latent variable and allows a framework to analyze relationships between constructs over time. A latent trajectory model identifies the dynamics of a construct in different moments in time due to changes in the variables that give rise to it, according to Eq. (2) (Geiser et al., 2013). = + +y ξ ζ ϵi t i t i t i t, , , , (2) where y is the ability to innovate (estimated construct), i refers to ob- servations, and t=0,1,2,⋯ ,s is the point in time when the values for each variable are presented. It is important to show that ξ characterizes the observation unit that generates the indicator for the observed variable. ζ is a specific residual indicator of the latent variable that characterizes the interaction between the observation units and the time when they were included. In the study we conducted, we consider t as a consecutive integer number related to the order of the PINTEC release dates, since IBGE does not provide yearly data. This strategy aims to show evidence of the effect on the construct, ability to innovate, of investments realized in different moments of the time (0,1,2,⋯ ,s), since the time of observations imust be less than the time of y. Finally, ϵ is the residual error term. The association of the latent trajectory with the autoregressive model is due to the theoretical assumption that the latent variables are hierarchically arranged in time (Cantwell and Fai, 1999; Castellacci, 2008). The possibility of associating the autoregressive model with the trajectory of the latent variable is predicted by Bollen and Curran (2004), according to Eq. (3). = + + + − − y α β ρ yΛ ϵi t i t i t t i t i t, 2 , 1 , 1 , (3) The extension for a bivariate (two latent variables) may be defined by Eqs. (4) and (5)(Bollen and Curran, 2004; Delsing and Oud, 2008). = + + + + − − − − y α β ρ y ρ xΛ ϵi t yi yt yi y y i t y x i t yi t, 2 , 1 , 1 ,t t t t1 1 (4) = + + + +− −− − x α β ρ x ρ yΛ ϵi t xi xt xi x x i t x y i t xi t, 2 , 1 , 1 ,t t t t1 1 (5) As assumptions, the error terms should have a zero mean and cannot be correlated with the other variables (Bollen and Curran, 2004). Be- cause of the possibility of multiple variables, Bollen and Curran (2004) postulate a simple matrix expression with two equations that describe the autoregressive model with a latent trajectory (Eqs. (6) and (7)). = + +η μ ζBi η ii (6) =o Pi ηi (7) The expression presupposes a structural relationship between the variables; thus, ηi is the vector that gathers the observable variables, the intercepts, and the covariances; μ is the vector of the means or inter- cepts; B is the matrix coefficient resulting from the coefficients of the relationships established in the vector ηi; and ζi is the residual vector. Table 2 Variables used in this research. Reflective latent variables Description of the observable variables Calculation Source Human Capital (HC) Doctorate (PhD) (v8) Doctors Total employees PINTEC Master's (Ma) (v9) Masters Total employees PINTEC Bachelor's (Bach) (v10) Bachelors Total employees PINTEC Technicians (Tec) (v11) Technicians Total employees PINTEC Internal Capital (IC) (v1) Internal R & D (IRD) Internal R & D Revenue PINTEC Acquisition of Machinery and Equipment (AME) (v2) Investment in Mac.and Equip Revenue PINTEC Training (TR) (v3) Training Revenue PINTEC Introduction of Technological Innovation (ITI) (v4) Invest.Introd.Innovation Revenue PINTEC Industrial Projects (IP) (v5) Invest.in Industrial Projects Revenue PINTEC External Capital (RC) External R & D (ERD) (v6) Invest.in External R & D Revenue PINTEC Acquisition of Other Knowledge (AOK) (v7) Invest.in Other Knowled. Revenue PINTEC Formative latent variable Description of observable variables Calculation Source Financial performance (FP) ROA (v12) Operating Net Profit. Total Investment PIA Cash Generation (CG) (v13) EBTIDA Revenue PIA Table 3 Descriptive results of the samples in the study (%). Variables 2000/2001 2003/2004 2005/2006 2008/2009 2011/2012 Longitudinal x s x s x s x s x s x x IRD 0.74 14.31 0.57 9.28 0.53 11.33 0.00 0.00 0.60 7.43 0.58 3.55 ERD 0.07 0.59 0.05 0.55 0.03 0.43 −0.00 0.00 0.06 0.74 0.06 0.50 AOK 0.14 1.62 0.10 2.27 0.44 23.84 0.00 0.00 0.20 3.14 0.13 1.26 AME 4.35 62.23 2.17 13.2 3.09 63.42 0.00 0.02 3.80 102.5 1.43 8.66 TR 0.15 1.68 0.08 1.06 0.15 4.85 0.00 0.00 0.07 0.94 0.05 0.31 ITI 0.27 4.29 0.20 4.46 0.24 5.32 0.00 0.00 0.17 3.53 0.16 0.76 IP 0.56 5.88 0.55 5.03 0.36 2.37 0.00 0.01 0.31 5.11 0.29 1.83 PhD 0.03 0.42 0.02 0.25 0.03 0.53 0.02 0.21 0.04 0.62 0.03 0.31 Ma 0.06 0.50 0.07 0.81 0.06 0.60 0.07 1.06 0.07 1.25 0.08 0.45 Bach 0.60 2.19 0.57 3.00 0.52 2.71 0.41 2.78 0.70 4.66 0.72 2.28 Tec. 0.60 2.13 0.48 2.68 0.40 1.98 0.26 2.09 0.35 2.91 0.49 1.90 CG 2.65 32.03 55.61 85.53 2.66 65.37 1.29 166.37 5.89 43.86 17.45 32.40 ROA − 2.86 674.14 584.29 29.756 −23.15 2.787. 3.46 22.610 1.16 9.179 38.84 480.06 Note: x is the average of the variables; s is the standard deviation of the variables. D.F.L. Santos et al. Technological Forecasting & Social Change 127 (2018) 258–270 262 Eq. (7) allows the observed variables (Oi) to be separated from the la- tent variables (Bollen and Curran, 2004). Finally, this autoregressive structure of the latent variables is pos- tulated in a multilevel structure of latent variables, according to Eqs. (8), (9), and (10) (Little, 2013). = + +y ηT Λ Θitc tc tc oc itc (8) = = +E y μ( ) T Λ Aitc y tc tc tctc (9) = ′ Σ Λ Ψ Λ Θtc tc tc tc itc (10) where i refers to the values of each individual company; t is the time in which each of the individual values is taken; c refers to the structure at the level in which the variables are grouped; y are the coefficients of the variables; μy is the average of the vectors; T is the vector of the vari- ables' mean; A is the vector of the latent variables' mean; Σ is the im- plicit variance-covariance matrix of the model; η is the vector that gathers the estimators of the latent variables; Λ is the matrix of the factor loadings of the estimators of the relations between constructs; Θ is the matrix of residual variances; Ψ is the matrix of covariances and variances between the constructs; and Λ′ is the transposed matrix Λ (Little, 2013). Thus, all relationships anticipated in the theoretical model, in- cluding the observed variables, are set for this multilevel model with an autoregressive latent trajectory to be calculated by structural equation modelling. The two main tests for evaluating the adjustment of the model are RMSEA and CFI (Little, 2013; Yuan and Bentler, 2007). The use of SEM dealing with innovation as a latent resource, with a relevant time path in a multilevel structure, aims to mimic the way the ability to innovate develops within the firm. In this context, the method allows the estimation of the impact of investments in innovation as a multidimensional construct, on financial performance. The auto- regressive model explores the effects of investments in innovation throughout the years, taking into account, the cumulative characteristic of these innovation efforts as suggested by Cantwell and Fai (1999). 4. Results and discussions Table 3 shows the descriptive results of the variables used for the cross-sectional and longitudinal samples that compose this study. It is noted that for all variables, there is a high dispersion of results, which shows relevant differences in the investment strategies in in- novation of Brazilian companies and in the composition of the structure of human capital. A high dispersion in the metrics that are associated with financial performance is also identified. It is verified in the 2008/ 2009 sample that there is a reduction in the average investment in in- novation. This fact has also been evidenced by (Paunov, 2012). The manner in which investments in innovation are organized in Brazilian films' ability to innovate can be identified by the factor ana- lysis presented in Table 4. For the first and last samples, the model adjusted better with six factors and the others with five; the total explained variance exceeded 70%. For parsimony, Table 4 presents only the factors that gathered more than one variable with a coefficient above 0.4. The tests that were anticipated for the KMO and Bartlett's sphericity factor analysis allow the reliability of the grouping of factors to be accepted. The grouping of the factors does not reflect a standard structure of innovation over time, as initially expected by the theoretical model and provided in the study by Santos et al. (2014). However, the relevant variables interspersed over time and indicate that during the period, there can be a dynamic in their management and, indeed, different forms of configuration. It is emphasized that the variables that were anticipated for fi- nancial performance showed a higher correlation between them than the factors that gather the innovation variables. These results suggest that the construct of financial performance is formed by the constructs Ta bl e 4 C on so lid at io n of th e Fa ct or ia l M at ri ce s of th e R es ea rc h. M at ri x 20 00 /2 00 1 M at ri x 20 03 /2 00 4 M at ri x 20 05 /2 00 6 M at ri x 20 08 /2 00 9 M at ri x 20 11 /2 01 2 1 2 3 4 1 2 3 1 2 3 1 2 3 1 2 TR 0. 87 8 IR D 0. 93 6 A O K 0. 94 4 M A 0. 89 6 IT I 0. 87 4 A O K 0. 84 5 IT I 0. 88 4 IT I 0. 93 3 TE C 0. 88 6 IR D 0. 81 3 A M E 0. 79 6 TE C 0. 59 3 A M E 0. 81 7 BA C H 0. 88 1 IP 0. 78 1 R O A A M E TR 0. 63 5 IT I 0. 90 5 TR 0. 64 3 IT I 0. 93 3 R O A M A 0. 86 7 TR 0. 82 6 BA C H 0. 91 9 IP 0. 92 8 PH D 0. 74 8 BA C H 0. 79 3 IP 0. 85 3 M A 0. 88 IR D 0. 93 9 PD E 0. 68 6 PH D 0. 68 9 IR D 0. 55 3 0. 66 1 TE C 0. 63 7 PH D 0. 93 5 BA C H 0. 63 3 R O A A O K 0. 63 6 R O A BA C H 0. 82 5 M A 0. 56 C G A M E C G TE C 0. 78 5 A O K 0. 91 7 IR D 0. 64 3 C G PH D M A 0. 62 9 IP 0. 55 1 0. 71 IP 0. 58 5 R O A ER D ER D TR 0. 61 9 0. 67 8 TE C ER D A O K C G C G ER D PH D A M E K M O = 0. 53 9 K M O = 0. 61 7 K M O = 0. 61 6 K M O = 0. 64 1 K M O = 0. 62 7 Ba rt le tt 38 ,0 33 .2 89 (p -v al ue = 00 0) Ba rt le tt 71 ,3 91 .3 (p -v al ue = 00 0) Ba rt le tt 61 ,3 71 .7 (p -v al ue = 0. 00 ) Ba rt le tt 45 ,4 36 .6 81 Ba rt le tt 66 ,3 04 . (p -v al ue = 0. 00 0) (p -v al ue = 0. 00 ) D.F.L. Santos et al. Technological Forecasting & Social Change 127 (2018) 258–270 263 of innovation and non-reflexive. The cross-sectional analysis of the influence of innovation on fi- nancial performance was conducted with the aid of five structural equation models, whose consolidated results are shown in Table 5. It is verified that except for AME in 2011, all observable variables were significant and positive in the formation of the latent variables of the model (HC, IC, and RC). These three latent variables were sig- nificant and had a positive influence on the construct of ability to in- novate, demonstrating, for the first time for the Brazilian reality, this structure for this organizational resource, innovation. It is also em- phasized that the covariances among the latent variables of HC, IC, and RC were significant and positive in almost every year. However, the adjustment tests do not allow the reliability of the model to be con- firmed with the available data, despite the improved testing over time. In the last year of analysis, RMSEA was 0.11, which is a value close to the maximum acceptable limit of 0.10. The construction of the trajectory model prioritized an evolutionary analysis of the structure dedicated to innovation, from 2000 to 2011, and the financial results were updated considering the year following the last aggregated information. Fig. 2 shows the complete studied model. In the construction of this model, there are a total of 159 coefficients, 125 covariances, and 85 variances, and 125 parameters are estimated for a total of 1340 companies, repeated at five periods, re- sulting in 2086 degrees of freedom. Table 6 shows the estimated coefficients of the regressions. The RMSEA of the independent model was 0.105, which allows the model to be accepted at the limit of approximately 90%. However, the standardized values of the regression variables of the ability to innovate on financial performance were at the value of 1.0, and this value is considered to be a transgressor in structural equation modelling (Little, 2013). It is noteworthy that for the entire period, the influence of resource innovation on financial performance is positive and significant, except for the 2008/2009. It is observed that the influence of the ability to innovate (ABIN) in 2011 is greater than the rest. All independent variables were significant for the construction of the first-level latent variables in the entire period (HC, RC, IC). However, it is noted that there is significance for the construct of re- source innovation only for 2000 and that IC is significant in explaining the ABIN in 2003 and 2005. The autoregressive process of resource innovation presented posi- tive estimators throughout the period but without significance. All covariances between the observable variables and the first-level latent variables were positive and significant. Given this indication, the analysis of the model was conducted without the autoregressive process of the ABIN, considering only the multilevel process and the trajectory of the latent variable. Fig. 3 shows the structural equation modelling and the influence on financial per- formance for 2012 only. The RMSEA presented the statistic of 0.111 slightly below the pre- vious model, which keeps a fragile adjustment. However, considering the exploratory nature of this study and the complexity of the model, which involves a number of variables (138), estimators (59), and cov- ariances (125), the reliability of 88% is judged as acceptable. Table 7 shows the results of the estimators of the model with the trajectory of the ABIN. All variables that compose the first-level latent variables were sig- nificant and had estimates with positive coefficients. Similar to the other model, the ABIN construct showed a significant and positive impact on financial performance. Moreover, in this model, the ABIN also impacted CG, which in turn also had a positive and significant influence on financial performance. Among the first- and second-level latent variables, it is observed that with the exception of the IC and RC of 2008 and 2011, all were significant in the composition of the ABIN, and the positive and sig- nificant influence of the HC on the ABIN stands out throughout the period. All covariances between provided in the model were positive and significant. Thus, the results suggest that initially, the cumulative pro- cess of the ABIN does not necessarily occur in a causal relationship but rather in an aggregation of competences in the constitution of this in- tangible asset. The factor analysis and the significance of the variables in the cross- sectional models for the formation of the latent variables validate a Table 5 Aggregated results of the cross-sectional samples of the PINTEC. Relations between 2000/2001 2003/2004 2005/2006 2008/2009 2011/2012 Estimate Estimate Estimate Estimate Estimate Constructs ABIN HC 0.087*** 0.181*** 0.62*** 0.353*** 0.067*** ABIN IC 0.725*** 0.671*** 0.733*** 0.684*** 0.962*** ABIN RC 0.362*** 0.374*** 0.016*** 0.311*** 0.149*** FP ABIN 0.44*** 0.002 0.085*** −0.166*** 0.008 PHD HC 0.465*** 0.545*** 0.294*** 0.408*** 0.155*** MA HC 0.073*** 0.317*** 0.972*** 0.878*** 0.612*** BACH HC 0.361*** 0.647*** 0.38*** 0.769*** 1.353*** TEC HC 0.564*** 0.407*** 0.026*** 0.498*** 0.005*** IRD IC 0.643*** 0.936*** 0.604*** 0.888*** 0.844*** AME IC 0.332*** 0.383*** 0.016*** 0.122*** 0.001 TR IC 0.155*** 0.753*** 0.533*** 0.739*** 0.551*** ITI IC 0.949*** 0.953*** 0.956*** 0.436*** 0.891*** IP IC 0.823*** 0.68*** 0.093*** 0.833*** 0.745*** ERD RC 0.465*** 0.545*** 0.294*** 0.408*** 0.155*** AOK RC 0.073*** 0.317*** 0.972*** 0.878*** 0.612*** ROA FP 0.906*** 0.143*** 0.358*** 0.064*** 1.352*** EBTIDAM FP 0.155*** 0.387*** 0.96*** 0.563*** 0.044*** Result of the covariances of the latent variables HC < –> IC 0.117*** 0.461*** 0.026** 0.163*** 0.194*** RC < –> IC −0.144*** 0.123*** 0.996*** 0.436*** 0.073*** RC < –> HC 0.161*** 0.75*** 0.011 0.065*** 0.048*** Adjustment tests RMSEA 0.2 0.19 0.137 0.138 0.11 CFI 0.09 0.561 0.563 0.622 0.751 *** Significance: 1%. D.F.L. Santos et al. Technological Forecasting & Social Change 127 (2018) 258–270 264 structure of the ABIN of companies based on HC, represented by the profile of the firms' professionals dedicated to R &D, the investments in the ABIN made internally, and investments in the construction and acquisition of knowledge in networks. The literature occasionally ex- plores these investments in a stratified manner; in other cases, the sources of information are taken through surveys. In this study, it was possible to show an integrated structure for firms' ABIN with the dis- crimination of all variables. The literature points to the obstacles in analyzing business innova- tion, due to the difficulty in measuring associated variables and the complexity of its relationships with other variables. For Shafique (2012), innovation as an organizational resource can be considered a black box. In our study, we take into account constructs that are related to innovation and discussed in the literature. Results of the study contribute to the understanding of the construct ABIN, strengthening the structural concept from the constructs related to internal, external and human resources (IC, RC, HC, respectively), and extending results from other studies, e.g., Santos et al. (2014). It is important to highlight that, in some studies, e.g., Crossan and Apaydin (2009), these resources are individually analyzed regardless of whether the approach focused on their impact on innovation or financial per- formance. The correlations between these different resources depicted in Figs. 2 and 3 indicate that the complexity of investments in business innovation should take in account not only individual elements but a combination of them. The structure of IC as a resource that brings to- gether investments in Internal R & D (Howell, 2016), training (Liao and Rice, 2010), machine acquisition (Hashi and Stojčić, 2013), introduc- tion of technological innovation and industrial projects, extends the discussion of the results achieved by specific research on these vari- ables, e.g., Frank et al. (2016). The results from the study also indicate the importance of the in- teraction of external and internal knowledge with innovation (de Moraes Silva et al., 2017; Wang et al., 2012). Such evidence was al- ready discussed in the literature, but the relationship between external sources to innovation has been analyzed separately from that of internal resources to the innovation capability (Ye et al., 2016). The significance of the multilevel structure of ABIN strengthens the conceptual proposal regarding the intangibility of the resource in- novation, how dynamic it is and how it influences financial perfor- mance (Bernardo, 2014). Considering the difficulty in specifically structuring the innovation resource or the frequent arguments pointing out that innovation is constituted by multiple sources and depends on a cumulative process, our study helps better understand the complex Fig. 2. Innovation management model with trajectory. D.F.L. Santos et al. Technological Forecasting & Social Change 127 (2018) 258–270 265 relationships, taking advantage of an empirical study using a large database, which is updated from time to time. In the longitudinal sample, all observable variables that underpin the model showed significant and positive covariances during the analyzed period, except for the variable Bach. This situation confirms the importance of continued investment in the sources of the ABIN over the course of time. Still in the analysis of covariance, it is emphasized that the corre- lations between first-level constructs were significant and positive, which confirms the importance of integration between the various sources of the ABIN. Therefore, the construct of innovation presupposes that its constituent variables are understood in an integrated manner. The ABIN structured in an autoregressive form did not show sig- nificant coefficients regarding its confirmatory factor structure, except for 2000 and the IC in two other periods. Despite the heterogeneity of the sample, it is verified that the innovation process does not require a static structure of investments in the sources of resource innovation but rather a dynamic view. It is noteworthy that HC was the only significant and positive finding in the formation of the ABIN in the analyzed period. IC and RC showed variations in the confirmatory structure of the ABIN, and IC had a negative influence on the structure in 2000/2001 and 2011/2012, whereas RC was positive across the sample, except for 2008. These results suggest that the growth in resources for innovation and the policies established in Brazil in this period allowed the growth in companies' ABIN by 2011, whose negative result can express the limitation in making effective the expenditure on innovation along with business performance, or the negative marginal growth in the learning curve, after the initial trajectory of the ABIN is established. The lower representativeness of the expenditure with RC in the early years of the project confirms the dependence of the innovation process provided on its internal R & D structure and the fact that Brazilian industry is positioned on the international scene with an in- tensive strategy to scale and/or depend on providers. 5. Final considerations This study advanced the discussion of the modelling of organiza- tions' ABIN by proposing a multi-level structured model through latent variable analysis and in an autoregressive manner. The prospection of the model was based on real information on investments in the innovation process from a broad base of Brazilian companies. However, the study presents the conjectural limitation of the country regarding the heterogeneity of the companies, a fact that complicates a statistical analysis of all of the information gathered. It is Table 6 Results of the model with the trajectory and autoregression of the latent variable of in- novation. Variables Variables Coefficients ABIN_00 HC_00 0.312*** ABIN_00 IC_00 0.0283*** ABIN_00 RC_00 0.140*** ABIN_03 RC_03 −0.036 ABIN_03 HC_03 −0.013 ABIN_03 IC_03 0.105* ABIN_03 ABIN_01 0.048 ABIN_05 IC_05 0.114** ABIN_05 RC_05 −0.021 ABIN_05 ABIN_03 0.063 ABIN_05 HC_05 −0.008 ABIN_08 ABIN_05 0.272 ABIN_08 RC_08 0.084 ABIN_08 HC_08 0.078 ABIN_08 IC_08 0.189 ABIN_11 ABIN_08 0.251 ABIN_11 IC_11 −0.025 ABIN_11 RC_11 0.068 ABIN_11 HC_11 −0.026 FP_04 ABIN_03 1.000*** FP_06 ABIN_05 1.000*** FP_09 ABIN_08 1.000 FP_12 ABIN_11 1.000*** FP_01 ABIN_00 1.000*** IP_03 IC_03 0.825*** ITI_03 IC_03 0.378*** TR_03 IC_03 0.394*** AME_03 IC_03 0.348*** IRD_03 IC_03 0.161*** IP_05 IC_05 0.707*** ITI_05 IC_05 0.475*** TR_05 IC_05 0.371*** AME_05 IC_05 0.363*** IRD_05 IC_05 0.181*** IP_08 IC_08 0.565*** ITI_08 IC_08 0.347*** TR_08 IC_08 0.264*** AME_08 IC_08 0.258*** IRD_08 IC_08 0.125*** PHD_03 HC_03 0.825*** MA_03 HC_03 0.378*** BACH_03 HC_03 0.394*** TEC_03 HC_03 0.348*** PHD_05 HC_05 0.825*** MA_05 HC_05 0.378*** BACH_05 HC_05 0.394*** TEC_05 HC_05 0.348*** PHD_08 HC_08 0.825*** MA_08 HC_08 0.378*** BACH_08 HC_08 0.394*** TEC_08 HC_08 0.348*** IP_11 IC_11 0.825*** ITI_11 IC_11 0.378*** TR_11 IC_11 0.394*** AME_11 IC_11 0.348*** IRD_11 IC_11 0.161*** PHD_11 HC_11 0.825*** MA_11 HC_11 0.378*** BACH_11 HC_11 0.394*** TEC_11 HC_11 0.348*** PHD_00 HC_00 0.825*** MA_00 HC_00 0.378*** BACH_00 HC_00 0.394*** TEC_00 HC_00 0.348*** IP_00 IC_00 0.825*** ITI_00 IC_00 0.378*** TR_00 IC_00 0.394*** AME_00 IC_00 0.348*** IRD_00 IC_00 0.161*** ERD_00 RC_00 0.825*** AOK_00 RC_00 0.378*** ERD_03 RC_03 0.690*** AOK_03 RC_03 0.293*** Table 6 (continued) Variables Variables Coefficients ERD_08 RC_08 0.825*** AOK_08 RC_08 0.378*** ERD_05 RC_05 0.690*** AOK_05 RC_05 0.293*** ERD_11 RC_11 0.364*** AOK_11 RC_11 0.120*** ROA_03 FP_04 0.586*** CG_03 FP_04 0.231*** ROA_05 FP_06 0.930*** CG_05 FP_06 0.337*** ROA_08 FP_09 0.086*** CG_08 FP_09 0.028*** ROA_11 FP_12 1.307*** CG_11 FP_120 0.213*** ROA_01 FP_01 0.733*** CG_01 FP_01 0.289*** * Significance: 10%. ** Significance: 5%. *** Significance: 1%. D.F.L. Santos et al. Technological Forecasting & Social Change 127 (2018) 258–270 266 emphasized that this limitation is a fact that does not occur in most studies on structural equation modelling that use a scale score obtained through a questionnaire. The access to a data set that includes several in different periods allowed the analysis of relationships between innovation and perfor- mance along time. By using longitudinal SEM, we could study complex and latent constructs. However, the study has some limitations. Endogeneity problems due to omitted variables and simultaneous causality can impact results. For instance, other variables can influence performance as well as performance itself can influence the availability of resources related to innovation. Using variables in different periods can reduce endogeneity issues, but not eliminate them. Further studies could explore panel data regressions using instrumental variables to take into account omitted variables. Regardless of the limitations of the proposed empirical model, it was possible to validate more clearly a framework for the establishment of the ABIN based on the following types of capital: human, internal, and relational. However, the results show that these capitals are not isolated features but rather are investments that are significantly correlated and that congregate tangible and intangible assets. Thus, it is noted that the assumption of the stratified analysis of these resources in financial performance will be limited. It was also found that the cumulative process of innovation occurs in a combined mode over the course of time, not necessarily in a linear fashion. Public policies that promote innovation are important for the growth of IC and RC, but in the Brazil case, maintaining the partici- pation of researchers in companies and in numbers lower than those in developed countries can contribute to the absence of a greater ABIN, given that researchers drive companies towards technological para- digms. Environmental conditions, particularly the 2008/2009 crisis, have affected the investments of Brazilian companies in innovation, and this situation requires care by the public entity to ensure the existence of a financial flow for innovation processes because future recovery may require a new learning curve, depending on the speed at which in- novation and technological standards are changed. In this context, there are also the managerial implications of this study, with the following standing out: i The management of innovation should include an integrated struc- ture of internal and external sources aimed at the ABIN and a structure of researchers oriented towards innovation, which con- tributes to the company's strategy; and ii Innovation and, in effect, the generation of results are a cumulative process of knowledge and therefore must be preserved in the busi- ness model, aiming at the generation of competitive advantage. Thus, there is no ABIN with specific projects or only specific man- agement practices. Our study may have some implications for public policies. For Fig. 3. Innovation management model with trajectory without the autoregressive process. D.F.L. Santos et al. Technological Forecasting & Social Change 127 (2018) 258–270 267 instance, government could encourage private investment in innova- tions by relaxing trade barriers, guaranteeing stability in credit lines for innovation, and fostering the development of capital markets that fi- nance innovation projects. In addition, government funding to start-ups and innovative companies should be accompanied by the development of an infrastructure that could allow for a more efficient interaction between relational and intellectual capital. Regulatory changes in the relationship between public universities and the private sector should allow greater integration of academic researchers and business en- trepreneurs. Although some initiatives regarding business incubators and science parks in Brazil aim to promote a stronger interaction within the triple helix framework, interestingly, most professors at public universities are not allowed to directly participate in or manage private business. The limitations of this study include the size and access policies of the research data. Regarding the database, due to constraints in the availability of data from industries with few companies, the study could not generate sector-specific results. A sectorial assessment would greatly contribute to the breakdown of the results, given the hetero- geneity of the different segments in Brazil. The inclusion of new companies in the model, or even the analysis of transversal samples, shows that new observations in the PINTEC/PIA will possibly contribute to a better analysis of the model adjustments. In addition, exploring the establishment of the ability to innovate con- tinues to be a frontier field for the better definition and taxonomies of the strategies used by companies and sectors. References Atalay, M., Anafarta, N., Sarvan, F., 2013. The relationship between innovation and firm performance: an empirical evidence from Turkish automotive supplier industry. Procedia. Soc. Behav. Sci. 75 (1), 226–235 apr. Audigier, V., Husson, F., Josse, J., 2014. A principal component method to impute missing values for mixed data. ADAC 10 (1), 5–26 dec. Bartoloni, E., Baussola, M., 2017. Driving business performance: innovation com- plementarities and persistence patterns. Ind. Innov. 1–21. http://dx.doi.org/10. 1080/13662716.2017.1327843. jun. Baumann, J., Kritikos, A.S., 2016. The link between r & d, innovation and productivity: are micro firms different? Res. Policy 45 (6), 1263–1274. http://dx.doi.org/10.1016/ j.respol.2016.03.008. jul. Bernardo, M., 2014. Integration of management systems as an innovation: a proposal for a new model. J. Clean. Prod. 82 (1), 132–142 nov. Bessant, J., 2008. Dealing with discontinuous innovation: the European experience. Int. J. Technol. Manag. 42 (1/2), 36. Bollen, K.A., Curran, P.J., 2004. Autoregressive latent trajectory (ALT) models a synthesis of two traditions. Sociol. Methods Res. 32 (3), 336–383 feb. Bäck, I., Kohtamäki, M., 2015. Boundaries of r & d collaboration. Technovation 45-46, 15–28. http://dx.doi.org/10.1016/j.technovation.2015.07.002. nov. Cammarano, A., Michelino, F., Lamberti, E., Caputo, M., 2017. Accumulated stock of knowledge and current search practices: the impact on patent quality. Technol. Forecast. Soc. Chang. 120, 204–222. http://dx.doi.org/10.1016/j.techfore.2016.12. 019. jul. Cantwell, J., Fai, F., 1999. Firms as the source of innovation and growth: the evolution of technological competence. J. Evol. Econ. 9 (3), 331–366 aug. Castellacci, F., 2008. Technological paradigms, regimes and trajectories: manufacturing and service industries in a new taxonomy of sectoral patterns of innovation. Res. Policy 37 (6-7), 978–994. http://dx.doi.org/10.1016/j.respol.2008.03.011. jul. Castellacci, F., Natera, J.M., 2016. Innovation, absorptive capacity and growth hetero- geneity: development paths in latin America 1970–2010. Struct. Chang. Econ. Dyn. 37, 27–42. http://dx.doi.org/10.1016/j.strueco.2015.11.002. jun. Cimoli, M., Dosi, G., 1995. Technological paradigms, patterns of learning and develop- ment: an introductory roadmap. J. Evol. Econ. 5 (3), 243–268 sep. Colombo, M.G., Rabbiosi, L., 2014. Technological similarity, post-acquisition r & d re- organization, and innovation performance in horizontal acquisitions. Res. Policy 43 (6), 1039–1054. http://dx.doi.org/10.1016/j.respol.2014.01.013. jul. Cozzarin, B.P., 2004. Innovation quality and manufacturing firms performance in Canada. Table 7 Results of the model with the trajectory of the ABIN latent variable. Variables Variables Coefficients ABIN IC_03 0.355*** ABIN IC_05 0.324*** ABIN IC_08 0.085 ABIN HC_03 0.103*** ABIN HC_05 0.103*** ABIN HC_08 0.103*** ABIN IC_11 −0.062 ABIN HC_11 0.103*** ABIN HC_00 0.103*** ABIN IC_00 −0.062 ABIN RC_00 0.284*** ABIN RC_03 0.085*** ABIN RC_05 0.085*** ABIN RC_08 −0.062 ABIN RC_11 0.031 CG ABIN 0.540*** Performance ABIN 0.351*** Performance CG 0.766*** IP_03 IC_03 0.792*** ITI_03 IC_03 0.470*** TR_03 IC_03 0.462*** AME_03 IC_03 0.393*** IRD_03 IC_03 0.136*** IP_05 IC_05 0.707*** ITI_05 IC_05 0.521*** TR_05 IC_05 0.437*** AME_05 IC_05 0.416*** IRD_05 IC_05 0.153*** IP_08 IC_08 0.610*** ITI_08 IC_08 0.425*** TR_08 IC_08 0.350*** AME_08 IC_08 0.332*** IRD_08 IC_08 0.119*** PHD_03 HC_03 0.792*** MA_03 HC_03 0.470*** BACH_03 HC_03 0.462*** TEC_03 HC_03 0.393*** PHD_05 HC_05 0.792*** MA_05 HC_05 0.470*** BACH_05 HC_05 0.462*** TEC_05 HC_05 0.393*** PHD_08 HC_08 0.792*** MA_08 HC_08 0.470*** BACH_08 HC_08 0.462*** TEC_08 HC_08 0.393*** IP_11 IC_11 0.792*** ITI_11 IC_11 0.470*** TR_11 IC_11 0.462*** AME_11 IC_11 0.393*** IRD_11 IC_11 0.136*** PHD_11 HC_11 0.792*** MA_11 HC_11 0.470*** BACH_11 HC_11 0.462*** TEC_11 HC_11 0.393*** PHD_00 HC_00 0.792*** MA_00 HC_00 0.470*** BACH_00 HC_00 0.462*** TEC_00 HC_00 0.393*** IP_00 IC_00 0.792*** ITI_00 IC_00 0.470*** TR_00 IC_00 0.462*** MA_00 IC_00 0.393*** IRD_00 IC_00 0.136*** PDExt_00 RC_00 0.792*** AOK_00 RC_00 0.470*** ERD_03 RC_03 0.683*** AOK_03 RC_03 0.333*** ERD_08 RC_08 0.792*** AOK_08 RC_08 0.470*** ERD_05 RC_05 0.683*** AOK_05 RC_05 0.333*** ERD_11 RC_11 0.272*** AOK_11 RC_11 0.102*** ROA_11 Performance 0.458*** Table 7 (continued) Variables Variables Coefficients GA_11 Performance 0.300*** ROS_11 CG 0.724*** EBTIDAM_11 CG 0.539*** *** Significance lesser than 1%. D.F.L. Santos et al. Technological Forecasting & Social Change 127 (2018) 258–270 268 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0005 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0005 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0005 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0010 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0010 http://dx.doi.org/10.1080/13662716.2017.1327843 http://dx.doi.org/10.1080/13662716.2017.1327843 http://dx.doi.org/10.1016/j.respol.2016.03.008 http://dx.doi.org/10.1016/j.respol.2016.03.008 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0025 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0025 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0030 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0030 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0035 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0035 http://dx.doi.org/10.1016/j.technovation.2015.07.002 http://dx.doi.org/10.1016/j.techfore.2016.12.019 http://dx.doi.org/10.1016/j.techfore.2016.12.019 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0050 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0050 http://dx.doi.org/10.1016/j.respol.2008.03.011 http://dx.doi.org/10.1016/j.strueco.2015.11.002 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0065 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0065 http://dx.doi.org/10.1016/j.respol.2014.01.013 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0075 Econ. Innov. New Technol. 13 (3), 199–216 apr. Crossan, M.M., Apaydin, M., 2009. A multi-dimensional framework of organizational innovation: a systematic review of the literature. J. Manag. Stud. 47 (6), 1154–1191 sep. Cruz-González, J., López-Sáez, P., Navas-López, J.E., Delgado-Verde, M., 2015. Open search strategies and firm performance: the different moderating role of technolo- gical environmental dynamism. Technovation 35, 32–45. http://dx.doi.org/10.1016/ j.technovation.2014.09.001. jan. Cyrino, A.B., Parente, R., Dunlap, D., de Góes, B.B., 2017. A critical assessment of Brazilian manufacturing competitiveness in foreign markets. Compet. Rev. 27 (3), 253–274. http://dx.doi.org/10.1108/cr-08-2016-0046. may. Damanpour, F., Walker, R.M., Avellaneda, C.N., 2009. Combinative effects of innovation types and organizational performance: a longitudinal study of service organizations. J. Manag. Stud. 46 (4), 650–675 jun. de Guimarães, J.C.F., Severo, E.A., Dorion, E.C.H., Coallier, F., Olea, P.M., 2016. The use of organisational resources for product innovation and organisational performance: a survey of the Brazilian furniture industry. Int. J. Prod. Econ. 180, 135–147. http://dx. doi.org/10.1016/j.ijpe.2016.07.018. oct. de Moraes Silva, D.R., Furtado, A.T., Vonortas, N.S., 2017. University-industry r & d co- operation in Brazil: a sectoral approach. J. Technol. Transf. http://dx.doi.org/10. 1007/s10961-017-9566-z. mar. Delsing, M.J.M.H., Oud, J.H.L., 2008. Analyzing reciprocal relationships by means of the continuous-time autoregressive latent trajectory model. Statistica Neerlandica 62 (1), 58–82 jan. Dewangan, V., Godse, M., 2014. Towards a holistic enterprise innovation performance measurement system. Technovation 34 (9), 536–545 sep. Dray, S., 2008. On the number of principal components: a test of dimensionality based on measurements of similarity between matrices. Comput. Stat. Data Anal. 52 (4), 2228–2237 jan. Figueiredo, P.N., 2010. Discontinuous innovation capability accumulation in latecomer natural resource-processing firms. Technol. Forecast. Soc. Chang. 77 (7), 1090–1108 sep. Forés, B., Camisón, C., 2016. Does incremental and radical innovation performance de- pend on different types of knowledge accumulation capabilities and organizational size? J. Bus. Res. 69 (2), 831–848 feb. François, J., Favre, F., Negassi, S., 2002. Competence and organization: two drivers of innovation. Econ. Innov. New Technol. 11 (3), 249–270 jan. Frank, A.G., Cortimiglia, M.N., Ribeiro, J.L.D., de Oliveira, L.S., 2016. The effect of in- novation activities on innovation outputs in the Brazilian industry: market-orienta- tion vs. technology-acquisition strategies. Res. Policy 45 (3), 577–592. http://dx.doi. org/10.1016/j.respol.2015.11.011. apr. Freeman, C., Soete, L., 1997. The Economics of Industrial Innovation. Pinter, London. Geiser, C., Bishop, J., Lockhart, G., Shiffman, S., Grenard, J.L., 2013. Analyzing latent state-trait and multiple-indicator latent growth curve models as multilevel structural equation models. Front. Psychol. 4 (1). Gunday, G., Ulusoy, G., Kilic, K., Alpkan, L., 2011. Effects of innovation types on firm performance. Int. J. Prod. Econ. 133 (2), 662–676. http://dx.doi.org/10.1016/j.ijpe. 2011.05.014. oct. Hashi, I., Stojčić, N., 2013. The impact of innovation activities on firm performance using a multi-stage model: evidence from the community innovation survey 4. Res. Policy 42 (2), 353–366 mar. Hatzikian, Y., 2013. Exploring the link between innovation and firm performance. J. Knowl. Econ. 6 (4), 749–768 mar. Hinloopen, J., 2003. Innovation performance across Europe. Econ. Innov. New Technol. 12 (2), 145–161 mar. Howell, A., 2016. Firm r & d, innovation and easing financial constraints in China: does corporate tax reform matter? Res. Policy 45 (10), 1996–2007. http://dx.doi.org/10. 1016/j.respol.2016.07.002. dec. Hu, B., 2014. Linking business models with technological innovation performance through organizational learning. Eur. Manag. J. 32 (4), 587–595 aug. Hu, M.-C., 2008. Knowledge flows and innovation capability: the patenting trajectory of Taiwan thin film transistor-liquid crystal display industry. Technol. Forecast. Soc. Chang. 75 (9), 1423–1438 nov. Hung, K.-P., Chou, C., 2013. The impact of open innovation on firm performance: the moderating effects of internal r & d and environmental turbulence. Technovation 33 (10-11), 368–380 oct. Jackson, S.A., Gopalakrishna-Remani, V., Mishra, R., Napier, R., 2016. Examining the impact of design for environment and the mediating effect of quality management innovation on firm performance. Int. J. Prod. Econ. 173 (1), 142–152 mar. Jiménez-Jiménez, D., Sanz-Valle, R., 2011. Innovation, organizational learning, and performance. J. Bus. Res. 64 (4), 408–417 apr. Kash, D.E., Rycroft, R., 2002. Emerging patterns of complex technological innovation. Technol. Forecast. Soc. Chang. 69 (6), 581–606 jul. Kim, B., Kim, E., Miller, D.J., Mahoney, J.T., 2016. The impact of the timing of patents on innovation performance. Res. Policy 45 (4), 914–928 may. Kirner, E., Kinkel, S., Jaeger, A., 2009. Innovation paths and the innovation performance of low-technology firms—an empirical analysis of German industry. Res. Policy 38 (3), 447–458. http://dx.doi.org/10.1016/j.respol.2008.10.011. apr. Kleinschmidt, E., 1991. The impact of product innovativeness on performance. J. Prod. Innov. Manag. 8 (4), 240–251 dec. Klingenberg, B., Timberlake, R., Geurts, T.G., Brown, R.J., 2013. The relationship of operational innovation and financial performance—a critical perspective. Int. J. Prod. Econ. 142 (2), 317–323 apr. Kostopoulos, K., Papalexandris, A., Papachroni, M., Ioannou, G., 2011. Absorptive ca- pacity, innovation, and financial performance. J. Bus. Res. 64 (12), 1335–1343 dec. Lancker, J.V., Mondelaers, K., Wauters, E., Huylenbroeck, G.V., 2016. The organizational innovation system: a systemic framework for radical innovation at the organizational level. Technovation 52-53, 40–50. http://dx.doi.org/10.1016/j.technovation.2015. 11.008. jun. Lewin, A.Y., Massini, S., Peeters, C., 2011. Microfoundations of internal and external absorptive capacity routines. Organ. Sci. 22 (1), 81–98. http://dx.doi.org/10.1287/ orsc.1100.0525. feb. Liao, T.-S., Rice, J., 2010. Innovation investments, market engagement and financial performance: a study among australian manufacturing SMEs. Res. Policy 39 (1), 117–125 feb. Lichtenthaler, U., 2016. Absorptive capacity and firm performance: an integrative fra- mework of benefits and downsides. Tech. Anal. Strat. Manag. 28 (6), 664–676. http://dx.doi.org/10.1080/09537325.2015.1131258. jan. Linton, J.D., Walsh, S.T., 2004. Integrating innovation and learning curve theory: an enabler for moving nanotechnologies and other emerging process technologies into production. R D Manag. 34 (5), 517–526. http://dx.doi.org/10.1111/j.1467-9310. 2004.00359.x. nov. Little, T.D., 2013. Longitudinal Structural Equation Modeling. The Guilford Press, New York. Lööf, H., Heshmati, A., 2006. On the relationship between innovation and performance: a sensitivity analysis. Econ. Innov. New Technol. 15 (4-5), 317–344 jun. Luo, B.N., Lui, S.S., Kim, Y., 2017. Revisiting the relationship between knowledge search breadth and firm innovation. Manag. Decis. 55 (1), 2–14. http://dx.doi.org/10.1108/ md-07-2015-0327. feb. Lööf, H., Heshmati, A., 2002. Knowledge capital and performance heterogeneity: Int. J. Prod. Econ. 76 (1), 61–85. http://dx.doi.org/10.1016/s0925-5273(01)00147-5. mar. Maslach, D., 2015. Change and persistence with failed technological innovation. Strateg. Manag. J. 37 (4), 714–723. http://dx.doi.org/10.1002/smj.2358. jan. MCTI, 2016. Consolidated indicators. In: Tech. rep. Brazilian Ministry of Science, Technology, Innovations and Communications. www.mct.gov.br/index.php/ content/view/2068.html. Olaru, D., Purchase, S., 2015. Innovation network trajectories: the role of time and his- tory. J. Bus. Ind. Mark. 30 (3/4), 342–353 may. Paunov, C., 2012. The global crisis and firms' investments in innovation. Res. Policy 41 (1), 24–35 feb. Pavitt, K., 1984. Sectoral patterns of technical change: towards a taxonomy and a theory. Res. Policy 13 (6), 343–373 dec. Prajogo, D.I., 2016. The strategic fit between innovation strategies and business en- vironment in delivering business performance. Int. J. Prod. Econ. 171, 241–249. http://dx.doi.org/10.1016/j.ijpe.2015.07.037. jan. Rangus, K., Slavec, A., 2017. The interplay of decentralization, employee involvement and absorptive capacity on firms innovation and business performance. Technol. Forecast. Soc. Chang. 120, 195–203. http://dx.doi.org/10.1016/j.techfore.2016.12. 017. jul. Ritala, P., Olander, H., Michailova, S., Husted, K., 2015. Knowledge sharing, knowledge leaking and relative innovation performance: an empirical study. Technovation 35, 22–31. http://dx.doi.org/10.1016/j.technovation.2014.07.011. jan. Rousseau, M.B., Mathias, B.D., Madden, L.T., Crook, T.R., 2016. Innovation, firm per- formance, and appropriation: a meta-analysis. Int. J. Innov. Manag. 20 (3), 1650033 apr. Ryu, H.-S., Lee, J.-N., 2016. Innovation patterns and their effects on firm performance. Serv. Ind. J. 36 (3-4), 81–101 mar. Samara, E., Georgiadis, P., Bakouros, I., 2012. The impact of innovation policies on the performance of national innovation systems: a system dynamics analysis. Technovation 32 (11), 624–638 nov. Santos, D.F.L., Basso, L.F.C., Kimura, H., Kayo, E.K., 2014. Innovation efforts and per- formances of Brazilian firms. J. Bus. Res. 67 (4), 527–535 apr. Shafique, M., 2012. Thinking inside the box? Intellectual structure of the knowledge base of innovation research (1988–2008). Strateg. Manag. J. 34 (1), 62–93 jul. Silverberg, G., Verspagen, B., 2005. A percolation model of innovation in complex technology spaces. J. Econ. Dyn. Control. 29 (1-2), 225–244 jan. Solow, R.M., 1957. Technical change and the aggregate production function. Rev. Econ. Stat. 39 (3), 312–320. Stefan, I., Bengtsson, L., 2017. Unravelling appropriability mechanisms and openness depth effects on firm performance across stages in the innovation process. Technol. Forecast. Soc. Chang. 120, 252–260. http://dx.doi.org/10.1016/j.techfore.2017.03. 014. jul. Sundbo, J., Gallouj, F., 2000. Innovation as a loosely coupled system in services. Int. J. Serv. Technol. Manag. 1 (1), 15. Tavassoli, S., Karlsson, C., 2015. Persistence of various types of innovation analyzed and explained. Res. Policy 44 (10), 1887–1901. http://dx.doi.org/10.1016/j.respol.2015. 06.001. dec. Teece, D.J., 2010. Business models, business strategy and innovation. Long Range Plann. 43 (2-3), 172–194 apr. Tsai, W., 2001. Knowledge transfer in intraorganizatiional networks: effect of network position and absorptive capacity on business unit innovation and performance. Acad. Manage. J. 44 (5), 996–1004. http://dx.doi.org/10.2307/3069443. oct. UNDP, 2013. Human development report 2013. In: Tech. rep. United Nations Development Programme, Washington, DC Communications Development Incorporated. Urgal, B., Quintás, M.Á., Tomé, R.A., 2011. Conocimiento tecnológico, capacidad de innovación y desempeño innovador: el rol moderador del ambiente interno de la empresa. Cuadernos de Economí,a y Dirección de la Empresa 14 (1), 53–66 jan. Velu, C., 2015. Business model innovation and third-party alliance on the survival of new firms. Technovation 35, 1–11. http://dx.doi.org/10.1016/j.technovation.2014.09. 007. jan. Wang, C.-H., Chang, C.-H., Shen, G.C., 2015. The effect of inbound open innovation on D.F.L. Santos et al. Technological Forecasting & Social Change 127 (2018) 258–270 269 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0075 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0080 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0080 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0080 http://dx.doi.org/10.1016/j.technovation.2014.09.001 http://dx.doi.org/10.1016/j.technovation.2014.09.001 http://dx.doi.org/10.1108/cr-08-2016-0046 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0095 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0095 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0095 http://dx.doi.org/10.1016/j.ijpe.2016.07.018 http://dx.doi.org/10.1016/j.ijpe.2016.07.018 http://dx.doi.org/10.1007/s10961-017-9566-z http://dx.doi.org/10.1007/s10961-017-9566-z http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0110 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0110 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0110 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0115 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0115 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0120 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0120 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0120 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0125 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0125 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0125 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0130 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0130 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0130 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0135 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0135 http://dx.doi.org/10.1016/j.respol.2015.11.011 http://dx.doi.org/10.1016/j.respol.2015.11.011 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0145 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0150 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0150 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0150 http://dx.doi.org/10.1016/j.ijpe.2011.05.014 http://dx.doi.org/10.1016/j.ijpe.2011.05.014 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0160 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0160 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0160 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0165 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0165 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0170 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0170 http://dx.doi.org/10.1016/j.respol.2016.07.002 http://dx.doi.org/10.1016/j.respol.2016.07.002 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0180 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0180 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0185 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0185 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0185 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0190 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0190 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0190 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0195 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0195 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0195 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0200 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0200 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0205 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0205 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0210 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0210 http://dx.doi.org/10.1016/j.respol.2008.10.011 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0220 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0220 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0225 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0225 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0225 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0230 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0230 http://dx.doi.org/10.1016/j.technovation.2015.11.008 http://dx.doi.org/10.1016/j.technovation.2015.11.008 http://dx.doi.org/10.1287/orsc.1100.0525 http://dx.doi.org/10.1287/orsc.1100.0525 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0245 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0245 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0245 http://dx.doi.org/10.1080/09537325.2015.1131258 http://dx.doi.org/10.1111/j.1467-9310.2004.00359.x http://dx.doi.org/10.1111/j.1467-9310.2004.00359.x http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0260 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0260 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0265 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0265 http://dx.doi.org/10.1108/md-07-2015-0327 http://dx.doi.org/10.1108/md-07-2015-0327 http://dx.doi.org/10.1016/s0925-5273(01)00147-5 http://dx.doi.org/10.1002/smj.2358 http://www.mct.gov.br/index.php/content/view/2068.html http://www.mct.gov.br/index.php/content/view/2068.html http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0290 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0290 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0295 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0295 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0300 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0300 http://dx.doi.org/10.1016/j.ijpe.2015.07.037 http://dx.doi.org/10.1016/j.techfore.2016.12.017 http://dx.doi.org/10.1016/j.techfore.2016.12.017 http://dx.doi.org/10.1016/j.technovation.2014.07.011 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0320 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0320 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0320 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0325 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0325 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0330 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0330 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0330 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0335 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0335 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0340 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0340 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0345 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0345 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0350 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0350 http://dx.doi.org/10.1016/j.techfore.2017.03.014 http://dx.doi.org/10.1016/j.techfore.2017.03.014 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0360 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0360 http://dx.doi.org/10.1016/j.respol.2015.06.001 http://dx.doi.org/10.1016/j.respol.2015.06.001 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0370 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0370 http://dx.doi.org/10.2307/3069443 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0380 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0380 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0380 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0385 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0385 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0385 http://dx.doi.org/10.1016/j.technovation.2014.09.007 http://dx.doi.org/10.1016/j.technovation.2014.09.007 firm performance: evidence from high-tech industry. Technol. Forecast. Soc. Chang. 99, 222–230. http://dx.doi.org/10.1016/j.techfore.2015.07.006. oct. Wang, Y., Li, J., Furman, J.L., 2017. Firm performance and state innovation funding: evidence from China's innofund program. Res. Policy 46 (6), 1142–1161. http://dx. doi.org/10.1016/j.respol.2017.05.001. jul. Wang, Y., Zhou, Z., Li-Ying, J., 2012. The impact of licensed-knowledge attributes on the innovation performance of licensee firms: evidence from the chinese electronic in- dustry. J. Technol. Transf. 38 (5), 699–715 may. Wang, Z., Wang, N., 2012. Knowledge sharing, innovation and firm performance. Expert Syst. Appl. 39 (10), 8899–8908 aug. Weber, K.M., Schaper-Rinkel, P., 2017. European sectoral innovation foresight: identi- fying emerging cross-sectoral patterns and policy issues. Technol. Forecast. Soc. Chang. 115, 240–250. http://dx.doi.org/10.1016/j.techfore.2016.09.007. feb. WEF, 2015. The global competitiveness report 2015-2016. In: Tech. rep. World Economic Forum, Geneva. WIPO, C.U.I., 2015. The global innovation index 2015: Effective innovation policies for development. In: Tech. rep. World Intellectual Property Organization. WTO, 2015. International trade statistics 2015. In: Tech. rep. World Trade Organization, Geneva. Wu, J., Wang, C., Hong, J., Piperopoulos, P., Zhuo, S., 2016. Internationalization and innovation performance of emerging market enterprises: the role of host-country institutional development. J. World Bus. 51 (2), 251–263 feb. Xie, X., Huo, J., Qi, G., Zhu, K.X., 2016. Green process innovation and financial perfor- mance in emerging economies: moderating effects of absorptive capacity and green subsidies. IEEE Trans. Eng. Manag. 63 (1), 101–112 feb. Ye, J., Hao, B., Patel, P.C., 2016. Orchestrating heterogeneous knowledge: the effects of internal and external knowledge heterogeneity on innovation performance. IEEE Trans. Eng. Manag. 63 (2), 165–176 may. Yuan, K.-H., Bentler, P.M., 2007. Multilevel covariance structure analysis by fitting multiple single level models. Sociol. Methodol. 37 (1), 53–82 dec. Dr. David Ferreira Lopes Santos is an assistant professor at the Department of Economics, Management and Education at São Paulo State University. He holds PhD degree in Administration from the Mackenzie Presbyterian University. He is recently conducting additional research lines related to firms' innovation dynamic and innovation habitats. His research interests focus on innovation, eco-innovation and business per- formance assessment. Since May 2015 he is the coordinator of the postgraduate program in Administration at São Paulo State University. He has published in a range of academic journals including Journal of Business Research, Brazilian Journal of Science and Technology, Regional and Sectoral Economic Studies and International Journal of Business and Emerging Markets. Dr. Leonardo Fernando Cruz Basso was born in Araraquara (Brazil) and graduated in Mechanical Engineering from the Technological Institute of Aeronautics (ITA- Sao Jose dos Campos, Brazil-1974), MA and PhD in Economics - New School for Social Research (New York-1979–1984) He attended a post-doctoral program at the University of Bielefeld (Germany). Full professor at the Department of Economics- Mackenzie Presbyterian University (São Paulo, Brazil). Researcher of the National Council for Scientific and Technological Development (CNPq/Brazil), with experience in economics and business, with emphasis in the following areas: value creation, open economy mac- roeconomics, valuation, innovation, intangible assets, exchange rate and intellectual ca- pital. Dr. Herbert Kimura is a full professor at the Department of Business Management at the University of Brasilia and researcher of the National Council for Scientific and Technological Development (CNPq/Brazil). He holds PhD degrees in Statistics and Business from the University of Sao Paulo. His research interests focus on innovation, performance assessment and risk management. He has published in a range of academic journals including Journal of Business Research, Journal of Banking and Finance, Journal of Financial Stability, Numerical Linear Algebra with Application and International Review of Financial Analysis. Dr. Kimura was the manager of the University of Brasilia's Science Park and Incubator Program from August/2013 to January/2017. D.F.L. Santos et al. Technological Forecasting & Social Change 127 (2018) 258–270 270 http://dx.doi.org/10.1016/j.techfore.2015.07.006 http://dx.doi.org/10.1016/j.respol.2017.05.001 http://dx.doi.org/10.1016/j.respol.2017.05.001 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0405 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0405 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0405 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0410 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0410 http://dx.doi.org/10.1016/j.techfore.2016.09.007 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0420 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0420 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0425 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0425 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0430 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0430 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0435 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0435 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0435 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0440 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0440 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0440 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0445 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0445 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0445 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0450 http://refhub.elsevier.com/S0040-1625(17)30215-9/rf0450 The trajectory of the ability to innovate and the financial performance of the Brazilian industry Introduction Theoretical framework Materials and methods Materials Method Results and discussions Final considerations References