Linking corporate social responsibility, cooperation and innovation: the triple bottom line perspective

Purpose – Drawing on the resource-based view (RBV) and knowledge-based view (KBV) theories, this study contributes to deepen the knowledge that corporate social responsibility (CSR) exerts on firms ’ innovation, considering the role played by cooperation. The research also seeks to ascertain the factors that influence the development of business cooperation. Design/methodology/approach – The database used is the Community Innovation Survey (CIS, 2014) appliedintheEuropeanUnion(EU)duringthetimeperiod2012 – 2014.Asampleof7083Portuguesefirmswere analyzed through the partial least squares structural equation modeling (PLS-SEM). Findings – The results suggest that CSR positively relates with firms ’ innovation, and business cooperation partially mediates this relationship. The outcomes also reveal that investing in certain types of innovation activities increases the firms ’ willingness to cooperate. Originality/value – The findings contribute to encourage an open innovation strategy as an easy and effective way to cope with rapid trends and changes, since it demonstrates the complementary between innovation and cooperation, as sources of value creation. From a triple bottom line (TBL) perspective, it also highlights that CSR must include social, economic and environmental initiatives, and should be a part of the firms ’ innovation strategy. As a result, managers who intend to contribute for society in the long term should plan, monitor and manage all CSR dimensions.

In addition, the 2030 Agenda for Sustainable Development, adopted by the United Nations members in 2015, includes 17 Sustainable Development Goals (SDGs) to be accomplished by 2030.These SDGs set the agenda for trends in CSR, which will gain an important position in the upcoming years.Considering the 17 goals, firms around the world are seen as relevant drivers in ensuring the development of this agenda.Although the advances in our understanding of how CSR may enhance innovation-related capabilities (He & Shen, 2019), due to the focus on economic benefits, previous studies have failed to explore the consequences of CSR for environmental innovation (hereafter, eco-innovation) (Pan, Sinha, & Chen, 2020).This is extremely important since eco-innovation is more closely related to sustainable benefitsat least in the short-termthan to economic interests that firms pursue.Thus, we intend to develop theoretical and practical contributions grounded on the ambitious goals of the 2030 Agenda, particularly, at the 9th SDG -Industry, Innovation and Infrastructure.
In line with the above discussion, we pose the following research questions: What relationship really exists between CSR and innovation?Is innovation enhanced by cooperation?Are CSR practices explaining innovation through cooperation?Trying to answer these questions, our research aims to achieve a better understanding of the CSRinnovation link, considering the role played by cooperation.This paper also examines the factors that influence the firms' willingness to cooperate.Addressing the research purpose, the database used is the Community Innovation Survey (CIS, 2014).The empirical analysis is carried out on a sample of 7,083 Portuguese firms from manufacturing and service sectors, covering a three year-period.
Portugal is a small open economy characterized by its strong innovation index (European Innovation Scoreboard, 2020).Considering the innovation production in small-and mediumsized enterprises (SMEs), Portugal assumes a leadership position by presenting highest shares of innovative products and business processes.According to the European Innovation Scoreboard [1] (EIS, 2020), between 2012 and 2019, the Portuguese innovative performance grew more than 20% in three interfaces: (1) innovation-friendly environment (e.g.broadband penetration), (2) investment level (e.g.non-R&D innovation expenditure) and (3) finance and support (e.g.venture capital expenses).For all of these reasons, the Portuguese economy represents a relevant setting for this study.
By using the partial least squares structural equation modelling (PLS-SEM), the current study provides interesting evidence of CSR-innovation relationship, expanding literature on this topic.As managers from innovation-oriented firms are concerned with making choices The triple bottom line perspective that influence business prosperity, these findings may help to broaden the field of CSR research by identifying the factors that shape the firms' innovative orientation.
The remainder paper is structured as follows.After this introduction, a brief review of the relevant literature is presented.Next, we describe the methodology adopted for data analysis.Then, we present and discuss the research findings.The last section provides the study's main conclusions, implications and limitations, as well as potential lines for future research.

Conceptual background 2.1 Cooperation
The model of firms operating in isolation lost interest, and, currently, a new perspective has been developed, where the organizations acting jointly face greater possibilities of success (Bayona, Garc ıa-Marco, & Huerta, 2001).Accordingly, cooperation corresponds to "connections based on partnerships with external actors that can be defined as bilateral cooperative relations with environmental constituents" (Lee, Lee, & Pennings, 2001, p. 620).Although the firms have their own resources, they recognize the relevance of interacting with other stakeholders who have additional assets; thus, external contacts play a very important role in obtaining those assets and identifying entrepreneurial opportunities, since autonomous actions are embedded within higher interorganizational networks (Granovetter, 1985;Burt, 1992).
Therefore, the establishment of cooperative relationships with other firms or institutions is seen as an opportunity to access complementary resources for faster development of innovations, to enhance competitive advantage and to share costs and risks (Faria et al., 2010).In order to overcome certain economic or technological constraints, firms establish relationships based on cooperation that represents interdependent and common goals (Easton & Ara ujo, 1992).The neoclassical theory emphasizes the economic agent as an isolated individual, neglecting the relevance of social ties; nonetheless, the economic sociology introduced a new element to the analysis, considering that economic decisions are embedded in a broad system of social connections (Braga, Gonçalves, & Braga, 2016).
In interorganizational networks, the process of knowledge transfer can be conducted formally and informally (Powell, Koput, & Smith-Doerr, 1996).The formal (or business) cooperation is contract-based since the relationship is based on regulations, while informal (or social) cooperation is trust-based ensuring the sustainability of business relationship (Hakansson & Johanson, 1998).In this paper, the term cooperation is used to denote a set of connected actors, which may be either organizations or individuals (Coviello & Cox, 2006).More specifically, we focus on formal (or business) cooperation, including collaboration among several partners and interfirm alliances with firms of the same group, suppliers, customers, competitors, consultants, universities, government and research centers (Batas & Liu, 2013;DGEEC, 2016a).

Innovation
Over the last decades, innovation has been highlighted as a competitiveness factor since the firms, in order to accomplish their goals, face the challenge of adapting to the environment pressures.Several scholars (e.g.Schumpeter, 1942;Hipp & Grupp, 2005;Amara, Landry, & Doloreux, 2009) have tried to conceptualize and explain the innovation process.The construction of this concept constantly meets the theories of Schumpeter (1942) that defines innovation as the application of new ideas in the generation of products or processes.
Thus, innovation can be categorized in different typologies depending on the influence that will be placed at the organizational level.The Oslo Manual (OECD, 2005) defines four types of innovation: (1) product innovation concerns to the design and commercialization of new or improved products/services; (2) process innovation relates to the production of new equipment on production processes; (3) organizational innovation encompasses the introduction of a new structure inside the organization and (4) marketing innovation involves the implementation of new marketing practices (e.g.new approach to sales).
According to Flikkema, Jansen, and Van Der Sluis (2007), innovations can be classified as technological when they apply to products or processes and non-technological when referring to organizational or marketing aspects.Schumpeter (1934) distinguishes between five types of innovation, emphasizing that two of them correspond to technological innovations (i.e. the introduction of new products and processes), while the remaining are intertwined with nontechnological innovations (i.e. the creation of new organizational structures, opening to new markets and developing new sources of raw materials).Moreover, the term innovation can be employed in different contexts referring either to a process or an outcome (OECD, 2018).Following the Manual Oslo (OECD, 2018) guidelines, to avoid this confusion, we use innovation activities to denote the process whereas innovation is limited to the outcome: An innovation is the introduction of new or significantly improved product, process, organizational or marketing methods [. ..] innovation activities include the acquisition of machinery, equipment, buildings, software, and licenses, engineering and development work, feasibility studies, design, training, R&D and marketing when they are specially undertaken to develop and/or implement a product or process innovation (DGEEC, 2016a, p. 91).
Nowadays, innovation management has evolved, and firms are adopting business models based on eco-innovation (Valdez-Ju arez & Castillo-Vergara, 2021).Eco-innovation (also called environmental, sustainable or green) is a relevant approach for addressing environmental concerns, offering double externalities to effectively control pollution and resource use (Pan et al., 2020).Thus, this concept evolved from practices exclusively oriented to environmental damage reduction (Rennings, 2000) towards a more complex multidimensional level (Pereira, MacLennan, & Tiago, 2020).
In a broader scope, eco-innovation can be scaled from traditional types of innovations that occur at the product level, production process and organizational management (Kemp & Pearson, 2020).Drawing on the Oslo Manual definitions (OECD, 2005), Kemp and Pearson (2020) define an eco-innovation as the "production, assimilation or exploitation of a product, production process, service, management or business method that is novel to the organization [. ..]" resulting "in a reduction of environmental risk [. ..]" (p. 7).In other words, a sustainable innovation relates to the development of new/improved products, processes, organizational or marketing methods, focusing on environmental benefits (DGEEC, 2016a).

Corporate social responsibility
The firms' survival involves not only maximizing profits but also generating benefits for stakeholders and related communities (Y añez-Araque, Hern andez, Guti errez-Broncano, & Jim enez-Est evez, 2020).The interest on CSR adoption and its influence on businesses has increased considerably (Boesso, Favotto, & Michelon, 2015;Reverte, G omez-Melero, & Cegarra-Navarro, 2016).According to Perdomo and Escobar (2011), there is not a universal measurement of CSR; over the time, different approaches have emerged with several studies characterizing CSR as a multidimensional construct (e.g.Bocquet et al., 2013;Cegarra-The triple bottom line perspective  Carroll's (1979) research defined CSR as a construct with four dimensions: economic, legal, ethical and discretionary responsibilities.Analyzing 37 definitions of CSR, Dahlsrud (2008) concluded that the concept includes five dimensions: environmental, social, economic, stakeholders and voluntariness.The Organization of Economic Cooperation and Development (OECD, 2009) describes a responsible business conduct as "making a positive contribution to economic, environmental and social progress."This notion represents the widely acknowledged triple bottom line (TBL) perspective, unpacking CSR into social, environmental and economic dimensions (Bansal, 2005).
While institutional and stakeholders' theories are used to explore the antecedents of CSR, the resource-based view (RBV) is applied to evaluate the consequences of CSR (Pan et al., 2020).The RBV has become a central framework in strategic management (Barney, 1991), focusing on economic issues to explain sustained competitive advantage (Pan et al., 2020).The economic CSR relates to the shareholders' perspective, supporting the idea that the firms' interactions with suppliers, customers and other stakeholders influence sustainability and performance in the long term (Torugsa, O'Donohue, & Hecker, 2013).
However, the existing RBV suffers from some constraints by ignoring the interaction between the organization and its natural environment (Hart, 1995).Built on the RBV, the natural resource-based view (NRBV) describes how firms obtain competitive advantages that allow to sustain the earth's natural resources and ecosystems (Svensson et al., 2018).The environmental CSR emphasizes the introduction of specific actions that minimize the firms' environmental footprint, allowing an efficient use of the available resources (Orlitzky, Siegel, & Waldman, 2011).
Nevertheless, the RBV and the NRBV's do not consider the TBL's social dimension (Svensson et al., 2018).Addressing this gap, the social resource-based view (SRBV) tries to explain how social capabilities may complement the two other dimensions, extending the scope of analysis by including a broad range of economic, social and environmental stakeholders (Tate & Bals, 2018).Accordingly, all three dimensions should be linked to achieve shared TBL value since "the global economy serves society, which lies within Earth's  (Griggs et al., 2013, p. 306).In social CSR, workplace and community assume an important role (Torugsa et al., 2013), since "the health, safety and general well-being of employees" allow "firms to act as good citizens in the local economy" (European Commission, 2003, p. 5).

Research model and hypothesis development
The competitiveness in international markets has promoted the development of cooperation agreements (Freire, 2000).Considering business innovation, Lundvall and Nielsen (1999) confirmed that a strong knowledge-base and R&D investments are key to firms' success.These researchers also pointed that the reinforcement of employees' skills enhances firms' willingness to introduce innovations.The triple bottom line perspective Nevertheless, firms cannot be restricted only to their capabilities; they need to benefit from external resources, establishing relationships with other stakeholders and extracting from them the effects of interactive processes (Caloghirou, Kastelli, & Tsakanikas, 2004).Previous studies display that "if alliances are about knowledge acquisition, the number of alliances will be limited by its absorptive capacity" (Grant & Baden-Fuller, 2004, p. 78).Following this reasoning, Tsai (2009) underlined the relevance of such ability and the effectiveness of collaborative networks on recognizing new opportunities.
A firm's decision to cooperate is driven by the fact that cooperation is an effective way to improve firms' success (e.g.Becker & Dietz, 2004;Abramovsky, Kremp, L opez, Schmidt, & Simpson, 2009).The literature has been trying to isolate the factors that influence the decision to cooperate (Faria et al., 2010).In agreement with this research stream, R&D activities, human resources qualification, firm size and competitiveness have emerged as the most important antecedents of cooperative interactions (Abramovsky et al., 2009).
The institutions with certain internal competencies in R&D, as well as those that obtain technologies outside, are more likely to interact with other partners (Bayona et al., 2001).Caloghirou et al. (2004) also explored to what extent the firms' internal and external capabilities affected, or not, their innovation level.The results suggested that the improvement of R&D potential and the investments on human resources exert positive effects on developing new or improved products.
Hence, the literature highlights that the investments in innovation activities related to buildings, equipment, software and external knowledge are driving forces of business cooperation (e.g.Mansfield, 1988;Shields & Young, 1994;Weiss, 2003;Camacho & Rodr ıguez, 2005;Elche & Gonz alez, 2008).According to these scholars, firms that invest in R&D, in improving their structures and training their employees obtain different technological abilities and, consequently, have a greater ability to cooperate.Through previous studies, the following hypotheses are proposed: H1a-h.The investments in innovation activities (i.e.in-house R&D; external R&D; assets acquisition; knowledge acquisition; employees' qualification; introduction of innovations; design; other activities) are positively related to firms' willingness to cooperate.
The RBV describes the firm as a unit of resources that create competitive advantages and enhance long-term performance (Barney, 1991).Considering that CSR actions encourage the development of the firms' intangible resources, these practices allow the development of capabilities that lead to sustained competitive advantages (Gallego-Alvarez, 2011).However, the RBV is limited since it does not explain the differences between general resources and knowledge-based capabilities; thus, the analysis of CSR can be complemented by adding the knowledge-based view (KBV), which considers the firms as entities able to integrate and distribute knowledge (Grant, 1996).
The adoption of CSR is a driving force of network relationships and leads to the development of strong ties with firms' stakeholders (Jansen, Van Den Bosch, & Volberda, 2006).As socially responsible organizations, it is essential to recognize the relevance of each stakeholder and integrate this knowledge into their strategy (Gras-Gil, Manzano, & Fern andez, 2016).In this way, CSR practices bring external knowledge to the firm (Luo & Du, 2015), resulting in new ideas (Katila & Ahuja, 2002).Accordingly, CSR may contribute to the innovation strategy in three ways: (1) through the interactions with internal and external stakeholders; (2) identifying new business opportunities arising from social demands and environmental concerns and (3) creating better working conditions based on employees' confidence (Hern andez & S anchez, 2012).
Therefore, CSR practices have become a vital condition to firms' innovation and cooperation (Alarc on & S anchez, 2013).The establishment of cooperative relationships INMR 20,3 allows sharing resources, capabilities or activities that support the knowledge inflow ensuring business prosperity (Caro, Peñalver, & Nieto, 2011).In a complex and unpredictable environment, cooperation brings more flexibility (Briones Peñalver, Bernal Conesa, & Nieves Nieto, 2018).Consequently, the sources that determine cooperation do not reside exclusively in the firm but are also related to external contingencies (e.g.social and environmental aspects).Given this pattern, the following hypothesis is proposed: H2. CSR initiatives have a positive influence on firms' cooperation.
The research on firm's cooperation with other stakeholders, and its potential effect on innovation, is not new (Fernandes et al., 2017).Literature emphasizes that the organizations that do not cooperate, and do not formally or informally exchange knowledge, limit their capability to adapt to market uncertainties (Hanna & Walsh, 2008).Typically, the process of developing an innovation involves three strategies: (1) generate internal knowledge (make), (2) purchase it (buy) or (3) cooperate with other agents (Navarro, 2002).
The firms' decision to cooperate is driven by the fact that the actors' linkages are an efficient channel to produce product/services innovations (Becker & Dietz, 2004).However, several other factors can support such decision-share expenses and risks, exploit synergies, recognize new opportunities and benefit from government financial support (Becker & Dietz, 2004;Freel & Harrison, 2006).According to Kotler et al. (2000), an innovation structure must include the means for the systematic generation of new ideas to implement in renewed products.These ideas can come from internal, external and institutional sources that have a positive influence on innovation production (Braga & Braga, 2013;Bach, Lojpur, Pekovi c, & Stanov ci c, 2015).This is consistent with the open innovation paradigm popularized by Chesbrough (2003).Open innovation includes R&D externalization, outsourcing, interfirm collaboration and organization-environmental interaction (Lazzarotti, Manzini, Nosella, & Pellegrini, 2017).According to this approach, the development of innovative outputs is facilitated by the firms' openness towards external knowledge sources (Chesbrough, 2006;Ferreira & Teixeira, 2019).Empirical evidence shows that firms implementing open innovation need to interact with a complete network of suppliers, customers, high education institutes, competitors and research centers (Perkmann & Walsh, 2007;Gassmann, Enkel, & Chesbrough, 2010).Thus, several researchers (e.g.Koschatzky & Sternberg, 2000;Miotti & Sachwald, 2003;Becker & Dietz, 2004;Faems, Looy, & Debackere, 2005;Faria et al., 2010;Lewandowska et al., 2016) acknowledge that innovation is enhanced by cooperation (see Appendix 1).
A recent research stream, focusing on environmental issues, also underlines those cooperative mechanisms as relevant drivers of sustainable practices and eco-innovations (e.g.Scandelius & Cohen, 2016;Le on-Bravo et al., 2017;Garc es-Ayerbe et al., 2019;Pereira et al., 2020).The assumption that environmental issues do not represent the core competencies of most companies is unanimous (Horbach, Oltra, & Belin, 2013).In this context, the collaboration with external partners enables to access useful knowledge for developing ecoinnovations (Melander, 2018), by allowing the integration of sustainable aspects into product design (Juntunen, Halme, Korsunova, & Rajala, 2019).From a cost point of view, cooperation is essential to the eco-innovation process since generates economies of scale and promotes knowledge overflows (Fabrizi, Guarini, & Meliciani, 2018).Accordingly, we argue that cooperation with several partners has a leverage effect on innovative outcomes:

The triple bottom line perspective
Spanish firms, Garc ıa-Piqueres and Garc ıa-Ramos (2020) concluded that CSR positively influences product, process and organizational innovations.Bocquet, Le Bas, Mothe, and Poussing (2017) also confirmed this positive link, highlighting that innovation serves as a mediator between CSR and firm performance.Following the TBL's perspective, Bacinello et al. (2020) unpacked CSR into social, economic and environmental dimensions, identifying a positive influence on innovation outcomes.
On the other hand, some studies reported a mixed or even negative effect (e.g.Gallego-Alvarez, 2011; Bocquet et al., 2013;Costa et al., 2015;Mithani, 2017).According to Bocquet et al. (2013), proactive CSR is positively related to innovation, whereas reactive CSR has a negative effect.Likewise, Mithani (2017) found that ecological and social contributions weaken the effect of R&D, suggesting that "managerial attention to innovation can be undermined by a greater emphasis on social responsibility" (p.699).The current lack of consensus is also aggravated since previous research suggests that innovation can be antecedent, moderator or an outcome of CSR initiatives (Pan et al., 2020).
Despite the relationship between CSR and general innovation is well documented, a limited number of studies have analyzed eco-innovation resulting from CSR (Pan et al., 2020).With increasing environmental issues, firms are more concerned about future generations raising more investments to facilitate sustainable environmental innovations (Arag on-Correa, Hurtado-Torres, Sharma, & Garc ıa-Morales, 2008).Although the theoretical perspective considers that social and economic CSR may not directly relate to environmental issues, some scholars suggest that the three dimensions are interconnected (e.g.Bansal, 2005;Torugsa et al., 2013).
Social CSR provides complementary resources by increasing the commitment with environmental values and improving employees' skills for adopting sustainable activities (Graafland, Van de Ven, & Stoffele, 2003).In contrast to the literature revealing that firms targeting financial goals tend to be less focused on other nonmarket strategies (i.e.social and environmental initiatives) (Friedman, 2007), recent findings suggest that, in the long run, economic and environmental interests can coexist (Jamali & Karam, 2018;Vishwanathan, 2020).The emphasis on long-term economic performance can generate sufficient cash flow to environmental practices, fostering a stable relationship with stakeholders; this provides access to human capital and land that are required for eco-innovation development (Pan et al., 2020).Thus, we hypothesize as follows: H4a-c.CSR initiatives are positively related to firms' innovation (i.e.technological, nontechnological and eco-innovations).
Figure 3 presents a conceptual model based on hypothesis development.

Methodology 4.1 Data collection and sample
For this study, a secondary database was selected based on the Portuguese Community Innovation Survey.The CIS instrument provides useful information on how firms interrelate with the external environment in order to access powerful information for the development of new innovation projects.In doing so, firms might use external agents as information sources or engage in formal cooperation activities (Fernandes et al., 2017).Developed under the guidelines of the Oslo Manual, the survey is the main statistical instrument to monitor the Europe's progress in terms of innovation.Each member of the European Union (EU) performs, at the firm level, the usual consistency and logical tests, as well as corrections for potential bias (Faria et al., 2010).The CIS questionnaire and the methodology of analysis are harmonized across countries allowing the comparison of results in different EU members.

INMR 20,3
This survey aims to collect data on innovation understood in a broader perspective rather than exclusively examining the invention process.Thus, the questionnaire comprises a wide range of innovation activities going beyond R&D expenditures, personnel training, market analysis and trial production to include the introduction of innovative production processes and organizational changes (Faria et al., 2010).Following the Eurostat recommendations, the Portuguese version directly collects information on product, process, organizational, marketing and environmental innovations.The dataset analyzes the period between 2012 and 2014, contemplating firms with ten or more employees operating in different sectors (Table 1).The CIS questionnaire was available between 9th October 2014 and 8th June 2016  The triple bottom line perspective (DGEEC, 2016b).Based on census combination for large firms and random sampling for other groups, the survey consisted of 9,455 enterprises.In the corrected sample of 8,736 companies, 7,083 valid answers were considered (i.e.81% response rate) (DGEEC, 2016b).As shown in Table 2, most of sampled firms are from manufacturing sectors (manufacturers 5 4,088).Further, following the European recommendation (Decree-Law No. 98/2015), more than half are classified as SMEs (84.9%).
In order to examine the sample, we performed correlation analysis and t-tests.The different analyses were conducted using IBM SPSS Statistics software, version 25.Correlations are generally low to moderate given an indication that there is a low risk of facing collinearity issues on the partial regressions (see Appendix 3).It is noteworthy that there is a strong positive correlation between technological and nontechnological innovations (0.585 *** ), since they can be considered measuring the firms' innovation accomplishments.Addressing this issue, we analyzed the common method bias through a full collinearity assessment approach (Kock, 2015).According to the author, the occurrence of Variance Inflation Factor (VIF) values greater than 3.3 is an indication of pathological collinearity.As all VIF values are clearly below to 3.3, the entire dataset can be considered free of common method bias (see Appendix 4).Regarding sector differences, the t-tests revealed significant statistical differences for social (p < 0.001) and environmental (p < 0.001) CSR dimensions (Table 3).Manufacturing firms are more concerned about issues underlying the social (mean for manufacturing 5 2.90 vs mean for service 5 2.65) and environmental dimensions (mean for manufacturing 5 2.14 vs mean for service 5 1.92), compared with service firms.To uncover any significant differences between industry typology and cooperation, a t-test confirmed that there are not statistically differences in the firms' cooperation (t (3537) 5 À0.650, p 5 0.516) for manufacturing and service firms.Finally, concerning innovation differences, the t-test revealed statistical differences for product (p 5 0.004), process (p < 0.001), organizational (p 5 0.001), marketing (p < 0.001) and ecoinnovations (p < 0.001).Manufacturing firms are more concerned in developing technological (i.e.product and process) and eco-innovations, while service firms tend to focus on nontechnological innovations (i.e.organizational and marketing).

Measurement, scale development and analysis techniques
Following previous research (e.g.Costa et al., 2015;Cegarra-Navarro et al., 2016;Briones Peñalver, Bernal Conesa, & de Nieves Nieto, 2018;Bacinello et al., 2020), this study adopts the PLS-SEM to test the proposed model.According to Hair, Risher, Sarstedt, and Ringle (2019), this technique was primarily selected because (1) The structural model is complex and includes many constructs, indicators and relationships.
(2) The research is based on secondary data.
(3) This method works well with larger samples.
(4) Distribution issues are concerned, such as lack of normality.

The triple bottom line perspective
Considering the sample size (n 5 7,083), the variables included on the analysis do not follow a normal distribution.Thereby, the lack of distributional assumption was the main reason for choosing PLS-SEM (Hair, Sarstedt, Ringle, & Mena, 2012;Nitzl, 2016).It should be noted that, in a limited number of situations, nonnormal data may also influence the PLS-SEM results.The use of bias-corrected and accelerated (BCa) bootstrapping handles this issue, as it adjusts the confidence intervals for skewness (Efron, 1987).Following these guidelines, we employed the BCa bootstrapping in order to adjust the data for skewness and potential bias (Aguirre-Urreta & R€ onkk€ o, 2018).
The CIS questionnaire used in this study is divided into 13 sections.There are five sections accessing product, process, organizational, marketing and eco-innovations.To evaluate the firms' innovation, we adopted scale validated in previous research (e.g.OECD, 2005;Flikkema et al., 2007;Kemp & Pearson, 2020).With regards to cooperation, new variables were defined according to the type of partner and the geographic market in which it is located (Portugal, Europe, USA, China/India, other countries).CSR was operationalized using the TBL perspective: economic, social and environmental dimensions (Y añez-Araque et al., 2020).
The measurement of innovation activities is grounded in the theoretical definition of OECD ( 2005), which describes them as the process that generates innovation (e.g.R&D activities, assets and knowledge acquisition, among others).Following previous research (e.g.Baum, Locke, & Smith, 2001;Caloghirou et al., 2004;Tourigny & Le, 2004;Casanueva, Castro, & Gal an, 2013), this study also controls firm size, public financial support, firm internationalization and sector differences.Appendix 5 provides complete information regarding the variable's operationalization, as well as how they relate with the CIS questionnaire.

Findings
The reflective measurement model evaluation focuses on the constructs' convergent validity, internal consistency reliability and discriminant validity (Sarstedt, Ringle, & Hair, 2017).We found that the four measurement models meet the relevant assessment criteria (Table 4).More specifically, all outer loadings are above 0.60 indicating a sufficient level of reliability (Hair, Hult, Ringle, & Sarstedt, 2013).Further, the AVE values are higher than 0.50, providing support for the measures' convergent validity (Fornell & Larcker, 1981).Composite reliability (CR) ranges from 0.718 to 0.873, which is above the recommended value of 0.70 (Hair et al., 2019).Moreover, the Cronbach's alpha (CA) display values between 0.669 and 0.754, which is considered an acceptable measure (Hair et al., 2019).Finally, most of the constructs have ρA values greater than 0.707 (Dijkstra & Henseler, 2015), except for nontechnological innovation that is slightly lower (ρA 5 0.675).These results suggest that the constructs exhibit sufficient levels of internal consistency reliability.
We access the discriminant validity by using the heterotrait-monotrait ratio (HTMT) (Henseler, Ringle, & Sarstedt, 2015) (Table 5).All the results are below to the conservative threshold of 0.85 (Kline, 2011).Moreover, we undertake the bootstrapping procedure with 5,000 samples selecting the no sign option, BCa bootstrap confidence intervals and one-tailed testing at the 5% significance level.The outcomes reveal that none of the HTMT confidence intervals include the value of 1, which means that discriminant validity has been established between all the pair of constructs.Thus, the reflective model suggests that measures display satisfactory levels of validity and reliability, i.e. we can proceed with the structural model evaluation.
The structural model assessment involves the analysis of collinearity issues through VIF values (see Appendix 4).As all VIF are below to the recommended value of 5 (Hair, Hult, Ringle, & Sarstedt, 2017), we conclude that multicollinearity is not a problem in our data.The path coefficients range between À0.159 and 0.176 with different significance levels (Table 6).We also found that the model explains 17.9% of cooperation, 16.4% of technological innovation, 14.3% of nontechnological innovation and 17.9% of eco-innovation.However, the constructs explained variance decreases when is adjusted for the number of variables in the model.The overall approximate model fits (SRMR) are below to the recommended value of 0.08 (Henseler et al., 2014), being smaller than their corresponding 95 and 99% quantile (Henseler, Hubona, & Ray, 2016), which suggests the existence of a good fit.

Discussion
The results support some of the research hypotheses (Table 6).With regards to Hypothesis 1, the relationship between innovation activities and the establishment of cooperative agreements is partially confirmed.More specifically, the findings reveal that firms investing in asset acquisition (H1c: β 5 0.049; p < 0.001), knowledge acquisition (H1d: β 5 0.058; p < 0.01), introduction of innovations (H1f: β 5 0.075; p < 0.001) and other activities (H1h: β 5 0.080; p < 0.001) display a higher willingness to cooperate.With a higher investment in innovation activities, managers are more oriented to the firms' interests,  catering its values and commitments in strategic decision-making, becoming more receptive to knowledge exchange with different stakeholders.
Thus, we found support for the relevant role of assets acquisition, innovation introduction, knowledge acquisition and other-related activities (e.g.feasibility studies, testing, tooling up and industrial engineering) as driving forces of business cooperation.These outcomes not only validate the theoretical predictions but also conform to many earlier studies (e.g.Mansfield, 1988;Shields & Young, 1994;Weiss, 2003;Camacho & Rodr ıguez, 2005;Elche & Gonz alez, 2008), highlighting the important role of innovation activities in shaping firms' strategies.
Moreover, Hypothesis 2 -CSR initiatives have a positive influence on the firms' cooperationis also supported (H2: β 5 0.176; p < 0.001).When the firm increases its commitment with CSR (i.e.economic, social and environmental initiatives), it becomes more able to develop network interactions.As socially responsible firms, they recognize the relevance of each partner for knowledge acquisition (Gras-Gil et al., 2016), which is consistent with the KBV perspective (Grant, 1996).The TBL's of CSR allows the integration of new stakeholders with the ability to influence the firms' strategic decision.We conclude that the focus on CSR enhances the managers' openness to welcome new ideas and managerial approaches, supporting the development of an atmosphere that promotes business collaboration.Through constant interactions with diversified partners (e.g.customers, suppliers, universities, research centers), firms can internalize the shared knowledge to achieve more returns that noncooperative counterparts (Liao & Long, 2019).
A recent research stream also highlights that CSR actions are closely related to environmental social governance (ESG) mechanisms (Ruan & Liu, 2021).ESG includes the focus on environmental concerns (e.g.climate change), social responsibility (e.g.human rights) and corporate governance (e.g.shareholder protection) (Lagasio & Cucari, 2019).The firm's stakeholdersshareholders, investors, government and regulatory agencieshave shown increased interest in CSR and ESG issues.Among the existing studies, a lot have revealed that ESG reduce corporate risk-taking behavior (Di Tommaso & Thornton, 2020), allowing more stability and elasticity in terms of CSR operations (Almeyda & Darmansya, 2019).In a broader scope, CSR and ESG commitments function as a necessary stimulus for enterprises undertake cooperative relationships by reducing the firm's systematic risk (Zhao et al., 2018) and increasing the quality of management practices (Ling, Forrest, Lynch, & Fox, 2007).
Likewise, Hypothesis 3cooperation with different partners positively influences the firm's innovationis validated.Our study shows that cooperation with different partners leads to higher levels of technological innovation (H3a: β 5 0.044; p < 0.01), nontechnological innovation (H3b: β 5 0.073; p < 0.001) and eco-innovation (H3c: β 5 0.065; p < 0.005).Network relationships allow resource and know-how interchange, reducing the costs of innovation process.Further, network members tend to lead firms' managers into an aspirational and risky growth path.These stakeholders have greater awareness of market opportunities and make better risk assessment.Thereby, firms with high level of collaboration are more likely to encourage an open innovation environment (Yun, Zhao, Jung, & Yigitcanlar, 2020), which can inspire shared knowledge to be turned into creative results (Li, Ma, Liu, & Liang, 2020).Our findings support the open innovation paradigm (Chesbrough, 2003) by emphasizing the vital role of business collaboration for the development of innovative outputs (e.g.Becker & Dietz, 2004;Faria et al., 2010;Lewandowska et al., 2016;Garc es-Ayerbe et al., 2019;Pereira et al., 2020).
Moreover, our study provides an additional insight.The outcomes indicate that CSR initiatives have a direct and significant impact on innovation (Table 6) but also and indirect, mediated effect through cooperation (Table 7).Given this pattern, we conclude that the research model is partially mediated (direct effect of CSR on innovation plus an indirect effect through cooperation), i.e. a combination of CSR and cooperation helps to improve firms' innovation.Such implication confirms the Hern andez and S anchez (2012) findings, proposing that CSR enhances innovation through the development of interactions with internal (e.g. business group), external (e.g.suppliers, customers, competitors) and institutional stakeholders (e.g.universities, research centers).
Concerning to control variables (Table 6), the most of coefficients are statistically significant in the expected direction.Tables 8 and 9 summarize the results of each research hypotheses.Of the four hypotheses initially formulated, three are supported by data collected (i.e.H2; H3a-c, H4a-c).However, Hypothesis 1the investments in innovation activities are positively related to the firm's willingness to cooperateis partially validated since only a few innovation activities explain interfirm collaboration (Table 9).

Conclusions
This article focuses on the effect of CSR on the firms' innovation, while exploring the mediating role of cooperation.We examined how the three dimensions of CSR affect innovation, which was classified into technological, nontechnological and environmental.The results suggest that CSR is positively related to all innovation types, and the development of cooperative relationships partially mediates this relationship.Moreover, we found that the investment in innovation activities (e.g.knowledge acquisition, innovation introduction, assets acquisition and other-related activities) tend to influence the firms' decision to cooperate.H1a.The investments in in-house R&D are positively related to the firm's willingness to cooperate (2) Although the lack of consensus on defining CSR, there is a research stream highlighting its multidimensional nature (e.g.Guerrero-Villegas et al., 2018; Garc ıa-Piqueres & Garc ıa-Ramos, 2020).Nonetheless, the previous literature has failed on using the three CSR dimensions (Gallagher, Hrivnak, Valcea, Mahoney, & LaWong, 2018), being mostly focused on limited indicators and/or a singular variable (Anser, Zhang, & Kanwal, 2018).From a TBL perspective, our research unpacked CSR into economic, social and environmental dimensions, concluding that CSR has a positive effect on innovation.This theoretical unpacking, through the lens of RBV, SRBV and NRBV, combined with its multidimensionality, may contribute to solve the inconsistencies reported on the CSR-innovation relationship.Therefore, this study supports the adoption of TBL's framework, which is linked with the SDGs of the 2030 Agenda for Sustainable Development, designed to address social, economic and environmental challenges faced by the planet.
(3) We add to the CSR-innovation literature by exploring the mediating role of cooperation.Anchored on the KBV framework, the results reveal that firms, simultaneously, oriented to CSR actions and cooperation agreements are able to increase their innovation level, which is a vital condition for growth and sustained competitive advantage.
(4) The integration of socially responsible actions not only results in an ethical and moral positioning but also allows the generation of intangible resources with a high strategic value, such as external cooperation and business innovation.

Managerial implications
First, firms need to devote resources for an open innovation strategy that allows to cope with trends and changes of the dynamic markets.Managers should consider the vital role of the relationships established with different stakeholders to explore new opportunities and maintain their competitive advantage.Second, considering that CSR contributes to innovation, this is a necessary stimulus for firms undertake CSR initiatives.This is a business vision that involves more than publicly demonstrate that organizations are socially responsible.SME owners may stop thinking that CSR is only applicable on larger firms who have the financial and human resources, breaking with the idea that, in smaller businesses, the costs of adopting CSR will exceed the benefits.Such a vision is an opportunity for all firms, regardless their size, to adapt on new circumstances (established in part by the 2030 Agenda) and to move from traditional management of CSR to its effective integration in firms' strategy.Therefore, true CSR must include social, economic and environmental initiatives and should be a part of the innovation strategy.As a result, managers who intend to contribute for society in the long term should plan, monitor and manage all CSR dimensions.Third, cooperation is the way by which CSR positively impacts firms' innovativeness, i.e. managers aiming to increase innovation should focus on CSR enhanced by network interactions.In other words, cooperation should be considered a mediator to achieve greater innovative outcomes.Finally, firms must be aware of the ESG role on their future sustainability.Around the world, there has been a proliferation of several reports aiming to incentive enterprises to improve their ESG performance.The European Directive 2014/95/EU encourages all firms to disclose in their annual reports nonfinancial information, namely (1) policies and outcomes regarding environmental, social and employee issues, (2) respect for the human rights, (3) anticorruption measures and (4) cultural diversity among workers.Indeed, after the appearance of COVID-19 epidemic, global investors' attention to CSR, ESG and sustainable investing have risen to a higher degree.Thus, managers should carefully consider the possibility of disclosing this type of information.In doing so, they will guarantee INMR 20,3 the transparency of their operations avoiding to harm corporate reputation in the short-term and ensuring business prosperity in the long-term.

Limitations and future research
This research has some limitations that must be considered when interpreting the results.The first limitation is the cross-sectional nature of the data that does not allow evaluating the firms' evolution in terms of innovation.Besides, the investigation is limited by the time span considered (three-year period).Longitudinal analysis would be important to understand the influence of CSR and cooperation on innovation outcomes.Second, this study is limited in scope since we only tested a sample of Portuguese firms.Although the results can be generalized into a limited extent for smaller, open and relatively well-developed economies, future research should consider evidence from different countries in order to validate our findings in other contexts.Third, the database used has two mainly boundaries: (1) the CIS data are usually available for the community a lot of time after being collected and (2) the metrics used to operationalize CSR and innovation are restrained by the survey.In order to provide updated contributions, future studies could apply a questionnaire to the firms' top management team.Through a primary database, it would be possible to overcome the main constraints of CIS ( 2014).In addition, the analysis of effects related to regulation and market trends is a fruitful direction for new studies.Beise and Rennings (2005) found that ecologically sustainable products (e.g.green electricity) must comply with foreign market and regulative norms if producers intend to be internationally successful.This means that for introducing an eco-innovation in the external environment, both elementsinternational market and regulationshould be aligned and be coherent worldwide.A plausible avenue for future research could explore the effect of regulation systems on the mechanism through which CSR dimensions translate the effect of eco-innovation on the firm internationalization patterns (in terms of speed of internationalization).Note(s): Based on the Eurostat guidelines (2020), firms were classified as manufacturing enterprises (12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32)(33)(35)(36)(37)(38) and service enterprises (42)(43)(46)(47)(48)(49)(50)(51)(52)(53)(58)(59)(60)(61)(62)(63)(64)(65)(66)(67)(68)(69)(70)(71)(72)(73)(74)(75)86) Table A5.

Social
Resource-Based View (SRBV) -Main scope in which a firm operates.-Certifications, awards or distinctions.-Communication channels currently used.-Collaboration with other entities in the field of CSR.-Breakdown of employees by gender, nationality and disabilities.-Training expenses.-List of expenses to improve health and safety.-Number of community projects supported.-Results of customer satisfaction surveys.to the community.-Total sum of all taxes paid.-Reserves.-Own endowment funds.Natural Resource-Based View (NRBV) -Annual electricity consumption.-Annual water consumption.-Use of recycled materials.-Waste recycled.-Corporate environmental initiatives (e.g., energy plan, luminaire control, waste treatment).
investments in external R&D are positively related to the firm's willingness to cooperate Not supported H1c.The investments in assets acquisition are positively related to the firm's willingness to cooperate Supported H1d.The investments in knowledge acquisition are positively related to the firm's willingness to cooperate Supported H1e.The investments in employees' qualification are positively related to the firm's willingness to cooperate Not supported H1f.The investments in innovations introduction are positively related to the firm's willingness to cooperate

Table 1 .
Sample distribution by NACE codes Note(s): 736 of sampled firms (10.4%) did not provide information about the number of employees Average extracted variance; CR 5 Composite reliability; CA 5 Cronbach's alpha t-value thresholds at one-tailed test of alpha Path coefficients significant at p-values: þ p < 0.05; **p < 0.005; ***p < 0.001.The values in the brackets represent t-values.t-value thresholds at one-tailed test of alpha 5 0.05 and 5,000 resamples: t (0.05; 4,999) 5 1.645; t (0.01, 4,999) 5 2.327; t (0.005, 4,999) 5 2.576; t (0.001; 4,999) 5 3.091 Innovation corresponds to a systematic phenomenon in which integrative learning and cooperative entrepreneurship are fundamental.Therefore, firms investing in innovation activities are more likely to cooperate with diversified actors, in order to share knowledge and risks.
Note1.The European Innovation Scoreboard provides a comparative assessment of research and innovation performance in the EU members and other related countries.This report allows policymakers to assess relative strengthens and weaknesses of national innovation systems, track progress and identify priority areas to boost innovation (EIS, 2020).